AGUILERA y Col - Food Engineering Interfaces PDF

Food Engineering Series
For other titles published in this series, go to

www.springer.com/series/5996
.
José Miguel Aguilera Gustavo V. Barbosa-
l
Cánovas Ricardo Simpson Jorge

l l
Welti-Chanes Daniela Bermúdez-Aguirre

l
Editors
Food Engineering Interfaces

Editors
Prof. José Miguel Aguilera Dr. Gustavo V. Barbosa-Cánovas
Pontificia Universidad Católica de Chile Washington State University
Dept. de Ingenierı́a Quı́mica y Dept. Biological Systems
Bioproceso Engineering L.J. Smith Hall 220
Vicuña Mackenna 4860 Pullman, WA 99164-6120
Santiago USA
Chile barbosa@wsu.edu
jmaguile@ing.puc.cl
Dr. Jorge Welti-Chanes
Dr. Ricardo Simpson Director de Post y Pregrado
Universidad Técnica Federico División de Biotecnologı́a y Alimentos
Santa Marı́a Instituto Tecnológico y de Estudios
Depto. Procesos Quı́micos, Superiores de Monterrey
Biotecnológicos y Ambientales Av. Eugenio Garza Sada 2501 Sur
Av. España 1680 Col. Tecnológico
Valparaı́so 64849 Monterrey
Chile N.L. México
ricardo.simpson@usm.cl jwelti@itesm.mx
Dr. Daniela Bermúdez-Aguirre

Washington State University
Dept. Biological Systems Engineering
213 LJ Smith Hall
99164-6120 Pullman
Washington
USA
daniela@wsu.edu
ISSN 1571-0297
ISBN 978-1-4419-7474-7 e-ISBN 978-1-4419-7475-4
DOI 10.1007/978-1-4419-7475-4
Springer New York Heidelberg Dordrecht London
# Springer Science+Business Media, LLC 2011
All rights reserved. This work may not be translated or copied in whole or in part without the written
permission of the publisher (Springer ScienceþBusiness Media, LLC, 233 Spring Street, New York,
NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in
connection with any form of information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed is forbidden.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they
are not identified as such, is not to be taken as an expression of opinion as to whether or not they are
subject to proprietary rights.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Preface
The articles included in Food Engineering Interfaces are based on the presentations
made by invited speakers of the 10th International Congress on Engineering and
Food (ICEF 10) held in Viña del Mar, Chile, in April 2008. All the chapters were
significantly upgraded from the original presentations and were later peer-reviewed
and copyedited. This book was conceived well ahead of the event in order to make
sure that readers would have the opportunity to get a taste of the challenges that lie
ahead. At the same time, efforts were made to cover, with a high degree of detail,
specific topics that are very relevant at the present time. As a result, this book is an
excellent addition to the literature as the topics are well blended and clearly expand
old food engineering boundaries.
The book is divided into five parts: selected topics in food engineering, advances
in food process engineering, water management in food, food microstructure, and
food packaging. All the 28 chapters have been written by renowned professionals
working in food engineering and related disciplines. The first chapter of the book
deals with the history and future of food engineering, and the remaining chapters in
Part I deal with topics such as microbial risk assessment using new engineering
tools, the development of eco-indicators to monitor the environmental impact on
selected food industries, the mathematical simulation of gastric digestion, the
engineering challenges posed by incorporating fiber in selected foods, the applications
of computational fluid dynamics (CFD) in food processing, food safety engineering,
food engineering economics, and the systemic approach for curriculum develop-
ment. Part II has two comprehensive overviews, one on advanced thermal proces-
sing and the other on nonthermal processing, as well as chapters dealing with the
optimization of thermal processes, the sterilization of foodstuff by pressure-assisted
thermal processing (PATS), and the extraction of essential oils and nutraceuticals
by supercritical fluid extraction. Part III has a very good selection of chapters,
including an update on glass transition in foods, caking in food powders, the
rehydration modeling of food particulate systems, and the identification of drying
zones in a laboratory spray-dryer. Part IV includes a chapter that envisions how
food microstructure studies will help the development of healthy foods that pro-
mote well-being and pleasure, as well as a chapter that deals with the analysis of
food microstructures by synchrotron X-ray-computed tomography. Finally, Part
V covers, among other topics, the utilization of edible coatings as food safety
v
vi Preface
and quality enhancers, physical properties of gel-based edible films, and packag-
ing materials based on renewable resources.
We truly hope that this book, with its visionary approach, will be a valuable
addition to the food engineering literature and will promote interest in food
engineering research, development, and implementation.
José M. Aguilera
Gustavo V. Barbosa-Cánovas
Ricardo Simpson
Jorge Welti-Chanes
Daniela Bermúdez-Aguirre
Acknowledgments
The editors would like to express their gratitude and appreciation to Sharon
Himsl, Publication Coordinator, Center for Nonthermal Processing of Food,
Washington State University, for her professionalism and dedication in facilitating
the steps needed to complete this challenging project. She kept track of all the
manuscripts and had them copyedited and effectively interacted with the editors,
authors, and the staff at Springer.
vii
.
Contents
Part I Selected Topics in Food Engineering
1 The Beginning, Current, and Future of Food Engineering:

A Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Dennis R. Heldman and Daryl B. Lund
2 Advances in 3D Numerical Simulation of Viscous

and Viscoelastic Mixing Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Kiran V. Vyakaranam and Jozef L. Kokini
3 CFD: An Innovative and Effective Design Tool

for the Food Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Tomás Norton and Da-Wen Sun
4 Incorporation of Fibers in Foods: A Food Engineering Challenge . . 69

Madhuvanti Kale, Dhananjay Pai, Bruce Hamaker,
and Osvaldo Campanella
5 Gastric Digestion of Foods: Mathematical Modeling of Flow

Field in a Human Stomach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Samrendra Singh and R. Paul Singh
6 State of the Art in Immobilized/Encapsulated Cell Technology

in Fermentation Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Viktor A. Nedović, Verica Manojlović, Branko Bugarski,
and Ronnie Willaert
7 Multifactorial Assessment of Microbial Risks in Foods: Merging

Engineering, Science, and Social Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . 147
Valerie Davidson, Juliana Ruzante, and Carlos Daza Donoso
ix
x Contents
8 Development of Eco-efficiency Indicators to Assess

the Environmental Performance of the Canadian Food
and Beverage Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Michèle Marcotte, Yves Arcand, Dominique Maxime,
and Denyse Landry
9 Food Process Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

George Saravacos and Zacharias Maroulis
10 Systemic Approach to Curriculum Design and Development . . . . . . . 237

Inés Ecima, Maurcio Pardo, and Gloria González-Mariño
Part II Advances in Food Process Engineering
11 Innovations in Thermal Treatment of Food . . . . . . . . . . . . . . . . . . . . . . . . . . 247

Arthur Teixeira
12 Optimization of Food Thermal Processing: Sterilization

Stage and Plant Production Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Ricardo Simpson and Alik Abakarov
13 Recent Advances in Emerging Nonthermal Technologies . . . . . . . . . . . 285

Daniela Bermúdez-Aguirre and Gustavo V. Barbosa-Cánovas
14 High-Pressure-Induced Effects on Bacterial Spores, Vegetative

Microorganisms, and Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Dietrich Knorr, Kai Reineke, Alexander Mathys, Volker Heinz,
and Roman Buckow
15 High Pressure Sterilization of Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341

Hosahalli Ramaswamy
16 Bioseparation of Nutraceuticals Using Supercritical

Carbon Dioxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Feral Temelli and Bernhard Seifried
17 Mass Transfer and Equilibrium Parameters on High-Pressure

CO2 Extraction of Plant Essential Oils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
José M. del Valle, Juan C. de la Fuente, Edgar Uquiche, Carsten Zetzl,
and Gerd Brunner
Contents xi
Part III Water Management in Food
18 Glass Transitions: Opportunities and Challenges . . . . . . . . . . . . . . . . . . . . 473

Yrjö H. Roos and Nattiga Silalai
19 Caking of Water-Soluble Amorphous and Crystalline

Food Powders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
Stefan Palzer and Karl Sommer
20 Effective Drying Zones and Nonlinear Dynamics

in a Laboratory Spray Dryer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515
Ulises Ramón Morales-Durán, Liliana Alamilla-Beltrán, Humberto
Hernández-Sánchez, Jose Jorge Chanona-Pérez, Antonio Ruperto
Jiménez-Aparicio, and Gustavo Fidel Gutiérrez-López
21 Rehydration Modeling of Food Particulates Utilizing

Principles of Water Transport in Porous Media . . . . . . . . . . . . . . . . . . . . . 535
I. Sam Saguy, Oranit Troygot, Alejandro Marabi, and Rony Wallach
22 Responses of Living Organisms to Freezing and Drying: Potential

Applications in Food Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
Marı́a del Pilar Buera
Part IV Food Microstructure
23 Food Microstructures for Health, Well-being, and Pleasure . . . . . . . . 577

José Miguel Aguilera
24 Fruit Microstructure Evaluation Using Synchrotron X-Ray

Computed Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
Pieter Verboven, Quang Tri Ho, Els Herremans,
Hibru Kelemu Mebatsion, Bart Nicolaı̈, Greet Kerckhofs,
Martine Wevers, and Peter Cloetens
25 Multifractal Characterization of Apple Pore and Ham

Fat-Connective Tissue Size Distributions Using Image Analysis . . . . 599
Fernando Mendoza, Nektarios Valous, Adriana Delgado,
and Da-Wen Sun
Part V Food Packaging
26 New Packaging Materials Based on Renewable Resources:

Properties, Applications, and Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 619
Stéphane Guilbert, Carole Guillaume, and Nathalie Gontard
xii Contents
27 Edible Coatings to Improve Food Quality and Safety . . . . . . . . . . . . . . . 631

Noemı́ Zaritzky
28 Physical Properties of Edible Gelatin Films Colored

with Chlorophyllide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 661
Paulo J.A. Sobral, Rosemary A. Carvalho, and Carmen S. Fávaro-Trindade
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
Contributors
Alik Abakarov Departamento de Ingenierı́a Quı́mica y Ambiental, Universidad

Técnica Federico Santa Marı́a, P.O. Box 110-V, Valparaı́so, Chile
José Miguel Aguilera Department of Chemical and Bioprocess Engineering,

Pontificia Universidad Católica de Chile, Santiago, Chile
Liliana Alamilla-Beltrán Departamento de Graduados e Investigación en

Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional,
Carpio y Plan de Ayala, s/n CP, 11340 México, DF, Mexico
Yves Arcand Agriculture and Agri-Food Canada’s Food Research and Develop-
ment Centre, 3600 Casavant Blud West, St. Hyacinthe, Quebec, Canada
Gustavo V. Barbosa-Cánovas Center for Nonthermal Processing of Food,

Washington State University, Pullman, WA, USA
Daniela Bermúdez-Aguirre Center for Nonthermal Processing of Food,

Washington State University, Pullman, WA, USA
Gerd Brunner Thermische Verfahrenstechnik, Technische Universität Hamburg-

Harburg (TUHH), Harburg, Germany
Roman Buckow CSIRO Food and Nutritional Sciences, 671 Sneydes Road,
VIC-3030 Werribee, Australia
Branko Bugarski Department of Chemical Engineering, University of Belgrade,

Karnegijeva 4, 11000 Belgrade, Serbia
Osvaldo Campanella Agricultural and Biological Engineering Department, Purdue

University, 47907 West Lafayette, IN, USA
Rosemary A. Carvalho Department of Food Engineering, University of São

Paulo, Pirassununga (SP), Brazil
xiii
xiv Contributors
Jose Jorge Chanona-Pérez Departamento de Graduados e Investigación en

Peter Cloetens European Synchrotron Radiation Facility, 6 Rue Jules Horowitz,

BP 220, 38043 Grenoble Cedex, France
Valerie Davidson School of Engineering, University of Guelph, N1G 2W1

Guelph, Ontario, Canada
Juan C. de la Fuente Departamento de Procesos Quı́micos, Biotecnológicos y

Ambientales, Universidad Técnica Federico Santa Marı́a, Valparaı́so, Chile
Adriana Delgado FRCFT Group, Biosystems Engineering, UCD Agriculture &

Food Science Centre, University College Dublin, Belfield, Dublin 4, Ireland
José M. del Valle Departamento de Ingenierı́a Quı́mica y Bioprocesos, Pontificia

Universidad Católica (PUC) de Chile, Santiago, Chile
Carlos Daza Donoso School of Engineering, University of Guelph, N1G 2W1

Guelph, Ontario, Canada
Inés Ecima Ingenierı́a de Producción Agroindustrial, La Sabana University,

Bogotá, Colombia
Carmen S. Fávaro-Trindade Department of Food Engineering, University of São

Paulo, Pirassununga (SP), Brazil
Nathalie Gontard Joint Research Unit, Agropolymers Engineering and Emerging

Technologies, Montpellier SupAgro, INRA, UM II, CIRAD, 34060 Montpellier,
France
Gloria González-Mariño Biosciences Doctoral Program, La Sabana University,

Bogotá, Colombia
Stéphane Guilbert Joint Research Unit, Agropolymers Engineering and

Emerging Technologies, Montpellier SupAgro, INRA, UM II, CIRAD, 34060
Montpellier, France
Carole Guillaume Joint Research Unit, Agropolymers Engineering and Emerging

Technologies, Montpellier SupAgro, INRA, UM II, CIRAD, 34060 Montpellier,
France
Gustavo Fidel Gutiérrez-López Departamento de Graduados e Investigación en

Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacio-
nal, Carpio y Plan de Ayala, s/n CP, 11340 México, DF, Mexico
Contributors xv
Bruce Hamaker Department of Food Science, Purdue University, 47907 West

Lafayette, IN, USA
Volker Heinz German Institute of Food Technology, Prof.-von-Klitzing-Str. 7,

D-49601 Quakenbrueck, Germany
Dennis Heldman Heldman Associates, 5224 Kings Mills Rd; #314, 45040 Mason,
OH, USA
Humberto Hernández-Sánchez Departamento de Graduados e Investigación en

Els Herremans Division BIOSYST-MeBioS, K.U.Leuven, W. de Croylaan 42,

Box 2428, BE-3001 Leuven, Belgium
Quang Tri Ho Division BIOSYST-MeBioS, K.U.Leuven, W. de Croylaan 42,

Antonio Ruperto Jiménez-Aparicio Centro de Desarrollo de Productos Bióticos

del Instituto Politécnico Nacional, CEPROBI-IPN, Yautepec, Morelos, Mexico
Madhuvanti Kale Department of Food Science, Purdue University, 47907 West

Lafayette, IN, USA
Greet Kerckhofs Research group of Materials Performance and Non-destructive

Evaluation, Katholieke Universiteit Leuven, Kasteelpark Arenberg 44, BE-3001
Leuven, Belgium
Dietrich Knorr Department of Food Biotechnology and Food Process

Engineering, Berlin University of Technology, Koenigin-Luise-Str. 22, D-14195
Berlin, Germany
Jozef L. Kokini Department of Food Science and Human Nutrition 2 College of

ACES, University of Illinois at Urbana Champaign, 61801 Urbana, IL, USA
Denyse Landry Agriculture and Agri-Food Canada Head Quarter, 1341 Baseline
Road, Tower 5, Floor 2, Room 126, Ottawa, ON, K1A OC5, Canada
Daryl B. Lund Department of Food Science, University of Wisconsin-Madison.

1605 Linden Dr., 53706 Madison, WI, USA
Verica Manojlović Department of Chemical Engineering, University of

Belgrade, Karnegijeva 4, 11000 Belgrade, Serbia
xvi Contributors
Alejandro Marabı́ Institute of Biochemistry, Food Science and Nutrition, Robert

H. Smith Faculty of Agricultural, Food and Environment, The Hebrew University
of Jerusalem, P.O. Box 12, 76100 Rehovot, Israel
Michèle Marcotte Agriculture and Agri-Food Canada’s Eastern Cereal and Oilseed
Research Centre, 960 Carling Avenue, KW Neatby, Room 1093, Ottawa, ON, k1A
OC6, Canada
Zacharias Maroulis School of Chemical Engineering, National Technical

University of Athens, 15780 Athens, Greece
Alexander Mathys Department of Food Biotechnology and Food Process

Engineering, Berlin University of Technology, Koenigin-Luise-Str. 22, D-14195
Berlin, Germany
Dominique Maxime Agriculture and Agri-Food Canada’s Food Research and

Development Centre, 3600 Casavant Blud West, St. Hyacinthe, Quebec, Canada
Hibru Kelemu Mebatsion Division BIOSYST-MeBioS, K.U.Leuven,

W. de Croylaan 42, Box 2428, BE-3001 Leuven, Belgium
Fernando Mendoza FRCFT Group, Biosystems Engineering, UCD Agriculture &

Food Science Centre, University College Dublin, D4 Belfield, Dublin, Ireland
Ulises Ramón Morales-Durán Departamento de Graduados e Investigación en

Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico
Nacional, Carpio y Plan de Ayala, s/n CP, 11340 México, DF, Mexico
Viktor A. Nedović Department of Food Technology and Biochemistry, University

of Belgrade, Nemanjina 6, P.O. Box 127, 11081 Belgrade-Zemun, Serbia
Bart Nicolaı̈ Division BIOSYST-MeBioS, K.U.Leuven, W. de Croylaan 42,

Tomás Norton Food Refrigeration and Computerised Food Technology (FRCFT),

University College Dublin, National University of Ireland, Agriculture & Food
Science Centre, Belfield, Dublin 4, Ireland
Dhananjay Pai Whistler Center for Carbohydrate Research, Purdue University,

47907 West Lafayette, IN, USA
Stefan Palzer Nestlé Research Centre Lausanne, Vers-Chez-Les-Blanc, 1000

Lausanne 26, Switzerland
Maurcio Pardo Ingenierı́a de Producción Agroindustrial, La Sabana University,

Bogotá, Colombia
Contributors xvii
Marı́a del Pilar Buera Departamentos de Industrias y de Quı́mica Orgánica,

Facultad de Ciencias Exactas y Naturales, 1428 Ciudad de Buenos Aires, Argentina
Hosahalli Ramaswamy Department of Food Science, McGill University, Ste

Anne de Bellevue, Montreal, Quebec, Canada
Kai Reineke Department of Food Biotechnology and Food Process Engineering,

Berlin University of Technology, Koenigin-Luise-Str. 22, D-14195 Berlin,
Germany
Yrjö H. Roos Department of Food and Nutritional Sciences, University College

Cork, Cork, Ireland
Juliana Ruzante Joint Institute for Food Safety and Applied Nutrition, University
of Maryland, 20742 College Park, MD, USA
I. Sam Saguy Institute of Biochemistry, Food Science and Nutrition, Robert H.

Smith Faculty of Agricultural, Food and Environment, The Hebrew University of
Jerusalem, POB 12, 76100 Rehovot, Israel
George Saravacos School of Chemical Engineering, National Technical Univer-

sity of Athens, 15780 Athens, Greece
Bernhard Seifried Department of Agricultural, Food and Nutritional Science,

University of Alberta, T6G 2P5 Edmonton, Alberta, Canada
Nattiga Silalai Department of Food and Nutritional Sciences, University College

Cork, Cork, Ireland
Ricardo Simpson Departamento de Ingenierı́a Quı́mica y Ambiental, Universidad

Técnica Federico Santa Marı́a, P.O. Box 110-V, Valparaı́so, Chile
R. Paul Singh Department of Biological and Agricultural Engineering, University

of California, 95616 Davis, CA, USA
Samrendra Singh Department of Biological and Agricultural Engineering, Uni-

versity of California, 95616 Davis, CA, USA
Paulo J.A. Sobral Department of Food Engineering, University of São Paulo,

Pirassununga (SP), Brazil
Karl Sommer Department of Process Engineering of Disperse Systems, Technical

University of Munich, Am Forum 2, 85354 Freising, Germany
xviii Contributors
Da-Wen Sun Food Refrigeration and Computerised Food Technology (FRCFT),

University College Dublin, National University of Ireland, Agriculture & Food
Science Centre, Belfield, Dublin 4, Ireland
Arthur Teixeira Agricultural and Biological Engineering Department, University

of Florida, 32611-0570 Gainesville, FL, USA
Feral Temelli Department of Agricultural, Food and Nutritional Science, Univer-

sity of Alberta, T6G 2P5 Edmonton, Alberta, Canada
Oranit Troygot Institute of Biochemistry, Food Science and Nutrition, Robert H.

Smith Faculty of Agricultural, Food and Environment, The Hebrew University of
Jerusalem, P.O. Box 12, 76100 Rehovot, Israel
Edgar Uquiche Departamento de Ingenierı́a Quı́mica, Universidad de La Fron-

tera, Temuco, Chile
Nektarios Valous FRCFT Group, Biosystems Engineering, UCD Agriculture &

Food Science Centre, University College Dublin, Belfield, Dublin 4, Ireland
Pieter Verboven Division BIOSYST-MeBioS, K.U.Leuven, W. de Croylaan 42,

Kiran V. Vyakaranam Department of Food Science, Rutgers University, 08901

New Brunswick, NJ, USA
Rony Wallach Department of Soil and Water Sciences, Robert H. Smith Faculty
of Agricultural, Food and Environment, The Hebrew University of Jerusalem,
P.O. Box 12, 76100 Rehovot, Israel
Martine Wevers Research group of Materials Performance and Non-destructive

Evaluation, Katholieke Universiteit Leuven, Kasteelpark Arenberg 44, BE-3001
Leuven, Belgium
Ronnie Willaert Structural Biology Brussels, Vrije Universiteit Brussel, Flanders

Institute for Biotechnology, Pleinlaan 2, B-1050 Brussels, Belgium
Noemı́ Zaritzky Centro de Investigación y Desarrollo en Criotecnologı́a de Ali-

mentos (CIDCA), Universidad Nacional de La Plata (UNLP) – CONICET La Plata
and Depto de Ingenierı́a Quı́mica, UNLP, La Plata, Argentina
Carsten Zetzl Thermische Verfahrenstechnik, Technische Universität Hamburg-

Harburg (TUHH), Harburg, Germany
.
Part I
Selected Topics in Food Engineering
.
Chapter 1
The Beginning, Current, and Future of Food
Engineering: A Perspective
Dennis R. Heldman and Daryl B. Lund
1.1 Introduction
Food engineering is a field of study that has emerged in rather recent history, but
is based on concepts that have evolved over a significant period of time. In general,
food engineering is the interpretation and application of engineering principles and
concepts to any aspect of food manufacturing and operations, including the con-
struction of the facilities for these operations. With this broad interpretation, almost
any of the more traditional fields of engineering (electrical, mechanical, civil,
chemical, industrial, and agricultural) would contribute to the successful operation
of the facilities and processes dealing with food. Recent focus in food engineering
has been on processes occurring within manufacturing operations, specifically on
the influence of unique characteristics and properties of foods and ingredients in
these processes.
The purpose of this paper is to review the history and current status of food
engineering, and to provide a framework for discussion about its future. In addition
to identifying some of the founding fathers and giants in food engineering, this
review will include the development of educational programs and the role of
food engineering within those programs. Attention will be given to evolution of
research and the potential of food engineering research in the future. Finally, there
will be an attempt to evaluate the impact of public sector research on research
conducted within the food industry, as well as applications of research within the
industrial sector.
D.R. Heldman (*)

Heldman Associates, 5224 Kings Mills Rd; #314, 45040 Mason, OH, USA
e-mail: drheldman@earthlink.net
D.B. Lund
Department of Food Science, University of Wisconsin-Madison. 1605 Linden Dr., 53706 Madison,
WI, USA
e-mail: dlund@cals.wisc.edu
J.M. Aguilera et al. (eds.), Food Engineering Interfaces, Food Engineering Series, 3
DOI 10.1007/978-1-4419-7475-4_1, # Springer ScienceþBusiness Media, LLC 2011
4 D.R. Heldman and D.B. Lund
1.2 Scope of Food Engineering
As suggested, the term “food engineering” has many interpretations and the scope of
this term has evolved over time. In addition, there are varied interpretations and
meanings in different regions of the world. It is likely our attempt at a definition will
fall short and not capture all of the various interpretations. However, the broadest
interpretation of food engineering occurs in the industrial sector, where it is gener-
ally accepted on a global basis. This general interpretation includes applications of
engineering in any aspect of production, handling, storage, processing, packaging,
and distribution of food. This is certainly the interpretation accepted by the founding
fathers of the International Congress of Engineering and Food (ICEF), who did not
use “food engineering” in the title of the congress. Obviously, this broad interpreta-
tion is inclusive of the expertise associated with most of the engineering fields.
In development of food engineering curricula, a more narrow interpretation of
food engineering has been applied within education institutions, so as to define the
uniqueness of the discipline in teaching and research. As might be expected, these
interpretations focus more on basic concepts and principles, and the impact of the
product being handled on the application of engineering principles, with special
emphasis on the word “food.”
Undergraduate food engineering teaching programs incorporate both basic engi-
neering principles and the unique chemical and microbiological characteristics of
foods. These programs tend to integrate basic engineering, chemistry, and microbi-
ology into process designs. It is clear that in most regions of the world, these
programs have evolved at the interface between science and engineering. In the
United States, two types of programs have developed in parallel: some with a clear
alignment with engineering curricula (usually in chemical or agricultural engineer-
ing), and many with food science programs (associated with colleges of agricultural
sciences). Graduate programs and research in food engineering have evolved with
an evident focus on basic principles, mathematical models, and process simulations.
In summary, food engineering provides an essential link between engineering
and the food sciences. The dimensions of this interface present unique challenges in
education, research, and applications, and will continue to provide professionals
associated with food engineering with opportunities in the future and a continuing
evolution of scope.
1.3 Definitions of Food Engineering
There are many definitions of food engineering, and not one is universally accepted.
One of the early definitions appears in the preface of the book Elements of Food
Engineering (1952) by Parker, Harvey, and Stateler. Parker proposed that “Food
Engineering is concerned with the design, construction, and operation of industrial
processes and plants in which intentional and controlled changes in food materials
1 The Beginning, Current, and Future of Food Engineering: A Perspective 5
are performed with due consideration to all economic aspects considered.” Later in
1963, Charm suggested that the purpose of food engineering is “to illustrate the
common relationship between basic engineering principles and the fundamentals of
food processing.” Finally, in 1966, Earle defined food engineering as “the study of
the processes that transform raw materials into finished products, or preserves foods
so they can be kept for longer periods.” There are obvious similarities and differ-
ences in these early definitions. All three indicate that food engineering is the
application of engineering to manufacturing and preservation of foods, but there
are distinct differences, from more emphasis on basic principles to greater emphasis
on applications of engineering in practice.
A more current definition of food engineering has been posted online (Wikipedia
2010; http://en.wikipedia.org/wiki/Food_engineering) as follows: “Food engineer-
ing refers to the engineering aspects of food production and processing. Food
engineering includes, but is not limited to, the application of agricultural engineer-
ing and chemical engineering principles to food materials.” However, given the
evolution of research and industrial applications, the following definition should
also apply, now and in the near future: Food engineering is both the identification
and creation of the physical principles associated with foods and ingredients, and
the applications of the principles to the handling, storage, processing, packaging
and distribution of consumer food products. This definition attempts to include both
the interests associated with engineering research and the applications of outcomes
from research to ultimate applications in processes associated with the food indus-
try. Undoubtedly, a better appreciation of the diversity of this unique field of study
solicits more discussion of the subject.
1.4 Origins of Food Engineering
There are clear relationships between the origins of food engineering and food
preservation processes. Design and prediction of the effect of preservation properties
on food requires quantification and modeling. One of the most visible applications can
be found in the book Sterilization in Food Technology by Ball and Olson (1957). This
book provides numerous examples of applications of mathematics to heat transfer to
foods in containers and the description of the kinetics of microbial destruction.
Historically, food preservation evolved from an art to a science. Early preservation
techniques (drying, salting, fermentation, and cooling) were based on “trial and error.”
Later, technologies such as canning and freezing were based on observations that
provided the basis for science-based improvements in the processes. Many of these
early scientific findings occurred in the basic sciences of chemistry, biology, and
physics. One of the more significant outcomes from Louis Pasteur (1860) in response
to food spoilage challenges was to provide the foundation for bacteriology and
microbiology. Other contributions from basic science and mathematics were signifi-
cant, such as the influence of temperature on reactions (Arrhenius 1889) and modeling
heat and mass transfer (Carslaw and Jaeger 1946; Crank 1956).
There are several other early and significant contributions that have become the
foundation of food engineering. The role of Nicholas Appert in demonstrating thermal
processing of foods in the early 1800s is recognized as the basis for many current
shelf-stable foods. The invention and development of mechanical refrigeration by
von Linde (1896) was significant, followed by Birdseye (1930) who established the
basis of the modern frozen food industry. Although dehydration has been recognized
as a food preservation process for centuries, the contributions of von Loesecke (1943)
advanced the process in a quantitative manner. These three processes (thermal
processing, food cooling and freezing, and food dehydration) continue to be the
most visible processes in the research literature on food engineering.
1.5 Evolution of Food Engineering
In general, food engineering was started by engineers who happened to have

some acquaintance with the food industry. However, there were other engineers
and scientists who appreciated the application of engineering principles to
something as complex as food. At the risk of omission, here are some of the
giants in food engineering and others who have made significant contributions
to food engineering (in no particular order): Bird, Stewart and Lightfoot,
Kessler, Thijssen, Leniger, Cheftel, King, Karel, Plank, Kuprianoff, Diendorfer,
Humphrey, Hallstrom, Heldman, Labuza, Berk, Mizrahi, Morgan, Pflug, Earle,
Goldblith, Proctor, Hayakawa, Ball, Zahradnick, Saravacos, Brody, Duckworth,
Cheftel (J.C.), Loncin, Saguy, Ohlsson, Pham, Teixeira, Sastry, Hall, Farrall,
Harper (J.C.), Nickerson, Mannheim, Harper (J.M.), Merson, Lund, Knorr,
Singh (R.P.), Schluender, Schwartzberg, Bimbenet, and Lovric.
In addition to people, there are several events and developments that have
contributed to the evolution of food engineering in the past 50 years. For profes-
sionals involved in teaching, there are several textbooks, including the following
key books:
1952 – Elements of Food Engineering by Milton E. Parker, Ellery H. Harvey, and
E. S. Stateler
1963 – Fundamentals of Food Engineering by Stanley E. Charm
1966 – Unit Operations in Food Processing by Richard Earle
1975 – Food Process Engineering by H.A. Leniger and W.A. Beverloo
1975 – Food Process Engineering by Dennis R. Heldman
1976 – Elements of Food Engineering by John C. Harper
1979 – Food Engineering; Principles and Selected Applications by Marcel Loncin
and Larry Merson
1980 – Fundamentals of Food Process Engineering by Romeo Toledo
1984 – Introduction to Food Engineering by R. Paul Singh and Dennis R. Heldman
Although other textbooks have been published in this recent time period, these
nine demonstrate the evolution of food engineering as a field of study. Parker et al.
(1952) placed significant emphasis on the description of manufacturing operations

and the role of engineering in all aspects of the manufacturing facility. Charm
(1963) introduced mathematics and an emphasis on process design and unit opera-
tions. Earle’s book (1966) was ideal for undergraduate students pursuing a degree in
food science, and the objectives in Harper’s book (1976) were similar. Books by
Leninger and Beverloo (1975) and Heldman (1975) were designed for students with
a significant background in engineering concepts, and emphasized advanced analy-
sis of unit operations. The book by Loncin and Merson (1979) provided even more
emphasis on mathematical analysis of unit operations, and was suited to introduc-
tory graduate students in some dimension of food engineering. Both Toledo (1980)
and Singh and Heldman (1984) developed textbooks for undergraduate students in
food science with specific attention to topics identified in guidelines for food
science curricula recommended by the Institute of Food Technologists.
During the past 25 years, several of these pioneering textbooks have been
published in later editions. In most cases, emphasis has been on new information
and technologies, as well as on advancement in tools available for conducting
analysis. There has been a continuous trend toward more mathematical description
of processes, and computer simulation of operations within the food industry.
Further, there has been an obvious emphasis on the impact of processes on product
quality attributes, including nutrients, as well as changes in process design to
enhance the efficiency of processes.
Another measure of the evolution and development of food engineering as a field
of study and research is the holding of national and international meetings that
focus on food engineering. Some of the key meetings and events that have occurred
during the last 50 years include:
1972 – Food Engineering Conference; organized by H.A. Leninger and Hans

Thijssen, hosted at Agricultural University, Wageningen, The Netherlands.
1976 – International Congress on Engineering and Food (ICEF1); 1st Congress
was chaired by Dr. Joe Clayton, held in Boston, Massachusetts. The 11th
Congress will be hosted in Athens, Greece on May 22–26, 2011.
1976 – Food Engineering Division of IFT; a petition signed by several food
engineering members of IFT was approved by the Executive Committee; the
Division continues to organize symposia and research programming at annual
IFT meetings.
1981 – International Conference in Fouling and Cleaning; conference was
organized by Bengt Hallstom, Lund University, and Daryl Lund, University
of Wisconsin; subsequent conferences were held in 1985 (USA) and 1989
(Germany); now part of ICEF.
1991 – Conference of Food Engineering (CoFE); first of nine conferences
organized by Dr. Martin Okos, held in Chicago. The 11th conference is being
planned for 2011.
1996 – CIBIA; 1st Congreso IberoAmericano de Ingenieria de Alimentos (CIBIA)
was hosted in Campinas, Brazil. The 6th Congreso was hosted in Bogata,
Columbia, in 2009.
2007 – European Workshop on Food Engineering and Technology; first of series of

workshops held with focus on applications of food engineering research.
Although many other meetings related to food engineering have occurred, the
above list illustrates the continuing development of research programs in the past
35 years.
The historical evolution of food engineering can also be illustrated by the
publication of research literature. Specific journals for publication of current food
engineering research have been created with some of the more visible journals and
reference books, such as:
1970 – Journal of Food Science; first issue published, which included a section
titled “Applied Science and Engineering by the Institute of Food Technologists.”
1976 – Journal of Food Process Engineering; first issue published by Food &
Nutrition, Press.
1982 – Journal of Food Engineering; first issue published by Academic Press, Inc.
1992 – Handbook of Food Engineering; first edition published by Marcel Dekker,
Inc.
1997 – Handbook of Food Engineering Practice; published by Academic Press, Inc.
2003 – Encyclopedia of Agriculture, Food and Biological Engineering; first edition
published by Marcel Dekker, Inc.
Information about food engineering has been published in many other journals
and references, but the above list illustrates the evolution within a relatively short
time period.
All of the developments described have occurred during a period of significant
change in the food industry and related industries. Bruin and Jongen (2001)
suggested the following categories of significant change:
1. Mergers and acquisitions of companies in the food and related industries. These
changes have caused significant reorganization within all companies, impacting
the role of food engineers in the industry.
2. Expectations for innovation. Changes in consumer expectations have increased
in frequency, forcing the identification and implementation of new, innovative
processes.
3. Changes in industry expectations. Previously, expectations of the food industry
were to deliver an abundant and safe supply of food to consumers. These
expectations have evolved to include not only more abundant and safe foods
but also foods with greater convenience, and eventually, foods with enhanced
influence on health and wellness.
4. Advances in computer technology. These advances have provided the tools to
accelerate the responsiveness of engineers to all types of challenges.
These changes have influenced the environment of food engineering activities,

and have had an impact on the expectations of food engineers in the industry. Many
of these changes also will be incorporated into future educational programs for food
engineers. Thus, a challenge to educators will be to incorporate these changes into

food engineering programs.
1.6 Evolution in Food Engineering Research
There have been significant advancements in food engineering research over the
past 50 years. Although the research has been conducted over a broad range of
applications, the unique focus of research that has evolved is the integration of
reaction kinetics with transport phenomenon to accomplish process design. The
research model concept is illustrated in Fig. 1.1.
The concept described by the model encourages research in three distinct areas:
kinetic parameters, transport phenomenon, and process design. Research on kinetic
models has expanded the availability of kinetic parameters and enhanced the investi-
gation of appropriate models used for food products. Transport phenomenon research
has improved our understanding of heat and mass transfer in foods, with specific
attention to the unique properties of food products and the mathematic models needed
to describe the phenomenon in a food system. Process design research integrates the
kinetic models with appropriate transport phenomenon models to allow prediction of
a parameter to quantify the quality of the food product. In general, the concept is ideal
for optimization of the process through maximizing the product quality attributes,
while identifying the best combination of process parameters.
1.6.1 Kinetic Models
The development of kinetic models for food systems is still evolving. A typical
model used to describe changes in a food component or attribute is expressed as:
dA=dt ¼ kA (1.1)
Food Engineering Research

Research Model –
Kinetic Transport
Models Phenomenon
Process
Design
Fig. 1.1 A food process

engineering research model Product Quality
where A ¼ the concentration or intensity of component, and k ¼ the first-order rate

constant.
There has been a significant increase in the measurement and publication of
rate constants for an array of product component changes as a function of process
or storage parameters, as illustrated by Villota and Hawkes (2007). Traditionally,
the same model has been used to describe changes in microbial populations during
preservation processes; the rate constants have been assembled in several loca-
tions, including the IFT Task Force Report (2000). These rate constants are
evaluated and expressed as a function of parameters such as temperature, water
activity, pressure, pH, etc. The evolution of more complex models has been
explored by von Bockel (2008), with specific attention to variations from first-
order kinetics.
1.6.2 Transport Phenomenon
Significant attention has been given to the development of models to describe heat,
mass, and momentum transfer in food systems, expressed as follows:
@N=@t ¼ at2 N (1.2)
where N ¼ intensity of a process parameter, and a ¼ appropriate property of food

structure.
Expressions of this type are used to predict distribution histories of the preser-
vation process parameter. Typical parameters include temperature, moisture con-
tent, pressure, and other parameters used to accomplish reductions in microbial
populations. Over the past 30–40 years, the availability of quantitative physical
properties data has increased significantly (Rao et al. 2005). In addition, there has
been increased emphasis on the development of prediction models for properties
based on product composition.
1.6.3 Process Design
Integration of the appropriate kinetic model with transport phenomenon expres-

sions leads to process design. In general, the outcomes are product quality attri-
butes, predicted as follows:
ð
A¼ kðN) dt (1.3)
where A ¼ the intensity or concentration of a quality attribute.

The focus of process design is on the product, with impact of the process on
reduction of the microbial population and/or the retention of a sensitive product
quality attribute. Process design provides the opportunity to consider multiple
attributes as long as the kinetic parameters and models are available. An even
more important dimension of this approach is the potential for process optimization,
which is the identification of the process parameters to achieve maximum retention
of product quality attributes, while ensuring the desired reduction in microbial
population for safety or shelf-life extension.
1.7 Contributions of Food Engineering Research
The process design approach has contributed to many positive products manufac-
tured by the food industry. These contributions can be expressed in several different
ways. In general, process design facilitates the quantitative prediction of process
outcomes, while evaluating a range of process parameters without conducting time-
consuming and costly experimental trials.
1.7.1 Safe and Wholesome Foods
Process design has had significant impact on the safety and wholesomeness of food
products. The development of food canning by Nicholas Appert (1750–1841) took
14 years. Using the current process design approaches, processes can be developed
within hours. These approaches ensure microbiological safety of products, along
with the capability to maximize the retention of product quality attributes.
A current example of process design is illustrated in Fig. 1.2. Knorr (2006) has
demonstrated that through measuring and understanding the kinetics of Escherichia
coli population reductions at high pressures, the retention of product quality
attributes in sausage are evident. These same approaches have been used to ensure
maximum retention of nutrients in a food product using high temperatures to ensure
microbiological safety and shelf-life of the product.
1.7.2 Affordable Food Supply
Food engineering research has contributed to keeping quantities of food available at

a modest cost to the consumer. These contributions have been accomplished
through overall improvements in the efficiency of food manufacturing. These
improvements have occurred through a focus on specific processes, as well as
an analysis of processing operations. Many of the improvements in efficiency
have occurred through increased capacity of individual pieces of equipment and
Of E. Coli in Raw Sausage
E. coli
0
30 °C
–1
Log Reduction [–]
–2
–3 5000 bar
–4
–5 6000 bar
–6
0 20 40 60 80 100 120
Time [s]
Food Biotechnology and Food Process Engineering
Fig. 1.2 Process design for pathogen reduction of Escherichia coli in sausage while retaining
product quality attributes (Knorr 2006)
throughput of processing lines. Another contributor to a low-cost food supply has

been the development of processes that allow foods to be transported for long
distances without noticeable reductions in product quality attributes.
An example of process improvements for shelf-stable foods has been aseptic
processing and packaging. This continuous process has been demonstrated as an
excellent alternative to the more traditional batch retort process. As illustrated in
Fig. 1.3, the process involves continuous processing of both product and packaging,
before filling the product into the package. As illustrated by Bruin and Jongen
(2001), improvements in the process have continued with capabilities to accom-
plish thermal processing of food particles within the carrier liquid.
1.7.3 Convenient Food Products
Process design has contributed to the development of convenient foods in a

significant way. Examples include products with reduced preparation times, pro-
ducts with improved quality attributes, products with extended shelf-life, and a
variety of shelf-stable products.
Aseptic Processing and Packaging
• Key Process Components
Package Material Package

or Container Sterilization
Raw Product Aseptic Final

Product Preservation Filling Product
Fig. 1.3 The aseptic processing and packaging concept in an aseptic environment
Freeze Drying
compressor
door
refrigeration
heated shelves vacuum
coils
pump
© 2002 HowStuffWorks
Fig. 1.4 The freeze-drying process
Freeze-dried foods are examples of products convenient for consumers. As

illustrated in Fig. 1.4, the freeze-drying process produces high-quality dry foods
by removing moisture from the product through sublimation at low temperature,
and preserving the product’s quality attributes. The freeze-drying process is not a
new one, since the first application was to produce freeze-dried coffee in 1938. The
process gained significant visibility in the 1960s during development of low-mass
foods for space flights. Applications of the process to foods continue as efforts to
reduce the costs associated with the process are evaluated.
1.7.4 Product Quality Improvements
Process design has contributed to food quality improvements in many different

ways. These improvements have occurred as a result of new and unique ingredients
manufactured using new processes created through process design. Product flavors
are being enhanced with new technologies to preserve volatile flavors, with specific
attention to encapsulation processes. Based on new physical properties knowledge,
process design has created an array of new and improved product textures. Many of
the recent improvements in food product quality can be traced to the “food stability
map” (Fig. 1.5) as suggested by Labuza et al. (1970).
The concept of water activity has provided a unique basis for product and
process development. Many new relationships between reaction rates in foods
and water activity have guided the development of foods with extended shelf-life
and unique quality attributes. These early developments have led to the more recent
concepts of glass transition as illustrated in Fig. 1.6 (Roos and Karel 1991). The
more sophisticated relationships between temperature and moisture content within
a food provide even more opportunities for process development now and in the
future. These state diagrams provide the ideal basis for evaluation of new ingre-
dients and their impacts on response of a product to a process, and optimization of
product quality attributes as a function of process parameters. These concepts will
provide the basis for food engineering research and applications for many years into
the future. All of these developments are based on the unique relationships between
the process and the ingredients of the food products.
Physical change region

LIP
ID
RELATIVE REACTION RATE
OX
ING
MOISTURE CONTENT
IDA
WN
TY
TIO
RO
I
TIV
H
N
TH
B
OWT
AC
IC
ROW
H
AT
TIC
WT
D GR
YM
MA
ST G
RO
NZ
ZY
MOL
IA G
N-E
YEA
EN
NO
TER
M
ER
BAC
TH
SO
MOISTURE CONTENT I
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

WATER ACTIVITY
Fig. 1.5 The food stability map (Figure redone by T.P. Labuza based on Labuza et al. 1970)
Fig. 1.6 The “state diagram” concept (Roos and Karel 1991)
1.7.5 Innovative Food Products
The development of many innovative new products has resulted from process
design. As new process technologies are identified and applied to food
manufacturing the opportunities for new and different products evolve. Similar
opportunities occur as new packaging and packaging systems are developed. In
many situations, these new process technologies are important to the enhancement
of product quality.
An excellent example of a process technology with significant impact on food
product development is extrusion. Extrusion is a process developed in the polymer
industry, but it has been adapted to a range of applications in the food industry. The
process combines the parameters of pressure, temperature, and time to produce
products with unique characteristics and properties. By considering the significant
range for each parameter, and the combinations of parameters for each one, the
opportunities for product attributes are nearly unlimited.
The opportunities for future developments in food extrusion are emphasized by
Bruin and Jongen (2001). By superimposing a typical cycle during an extrusion
process on a state diagram (Fig. 1.7), the depth of understanding achieved from
recent process design research is recognized.
Extrusion path in state

diagram
• Path with steps
160
140 1. Wetting
Free flow
120
2. Heating and
100
mixing
80
T [C]
60 3. Expansion
40
Rubber 4. Drying
20
Glass
0 5. Cooling
–20 0 10 20
–40
Moisture [%]
Fig. 1.7 The extrusion process on a state diagram (Bruin and Jongen 2001)
As is evident in Fig. 1.7, the steps associated with the extrusion process shift the
product ingredients from a glass state to a rubber state, followed by elevation of
temperature before expansion and drying back to a glass state. The final step of
cooling returns the product to the same composition and temperature, but the final
product has completely different properties and characteristics. Obviously, this
understanding of the process and the product creates an unlimited array of oppor-
tunities for new and innovative products.
1.8 The Future of Food Engineering
The future of food engineering is very bright. It is evident that food engineering
education will continue to evolve for both undergraduate and graduate students.
It will continue to be an attractive field of study because of the unique character-
istics of foods and ingredients, and the challenges associated with applications of
engineering concepts and principles. The education of future students will be driven
by the new information created by current and future research.
One of the currently evolving areas of research is nanoscale science and the
translation of outcomes into applications in food systems. Bruin and Jongen (2001)
have suggested that nanoscale adds another scale of consideration to existing scales
that range from molecular scale to supply-chain scale. Although the visible applica-
tions are still evolving, there seem to be several potential outcomes. The shelf-life of
foods is likely to be increased through a combination of nano-sensors used to detect
the onset of product deterioration, followed immediately by intervention to prevent
this deterioration. Through the study and understanding of food properties at the
nanoscale, the texture of food products can be improved and new product textures
are likely to evolve. Many of these improvements are likely to be the result of the
creation of nano-structured particles and films. The encapsulation of flavors within
nanoscale particles should ensure that food products retain optimum flavor intensity
through processing steps and for longer periods of time during storage and distribu-
tion. Similar opportunities seem logical for bioactive compounds. Knowledge about
packaging materials at the nano-scale should lead to the development of new and
improved packaging materials for food products.
Food engineering will contribute to the goals of improved health and wellness of
consumers through development of functional foods. The application of engineer-
ing concepts and principles to the metabolism of food will provide insights on the
product and process development cycle. The potential for incorporating bioactive
compounds into food products in a manner that ensures delivery of the compound to
the appropriate site within the body during metabolism of the food is achievable
through food engineering research.
An evolving challenge to food manufacturing and distribution is the constraint of
sustainability. Food engineering should contribute to sustainability in many differ-
ent ways. The basic concepts of material and energy balances will become standard
tools in evaluation of all scales of operation from the point of raw material
production to delivery of the product to the consumer and beyond. The transitions
of new information from process design to commercial manufacturing operations
will continue to be a significant challenge. These transitions must be accomplished
in a more efficient manner.
1.9 Summary
Food engineering has a relatively brief history as a field of study. Educational

programs for undergraduate and graduate studies have evolved over the past
50 years, with somewhat different approaches globally. Many engineering impacts
on the food industry have a longer history, as indicated by the origin of several
preservation processes, such as drying, canning, and refrigeration. More recently,
food engineering research has brought a quantitative dimension to both product and
process development. The future of food engineering research will continue to
focus on technology transfer and attempts to accelerate the transition of new
processes from the laboratory bench to operations scale in manufacturing. Future
research must be directed toward meeting the expectations of the consumer for safe
and convenient foods that will contribute to a healthy lifestyle for extended periods
of time.
References
Arrhenius S (1889) Z Phyzik Chem 4:226, Title not available

Ball CO, Olson FCW (1957) Sterilization in food technology. McGraw-Hill, New York
Birdseye C (1930) Production of quick-frozen fish. US Patent #1773070
Bruin S, Jongen RG (2001) Food process engineering: the past 25 years and challenges ahead.
The Food Engineering Division Lecture. Institute of Food Technologists Annual Meeting,
New Orleans, LA
Carslaw HS, Jaeger JC (1946) Heat of conduction in solids. Oxford University Press, London
Charm SE (1963) The fundamentals of food engineering. The AVI Publishing Co., Westport, CT
Crank J (1956) The mathematics of diffusion. Oxford University Press, London
Earle RL (1966) Unit operations in food processing. Pergamon, London
Harper JC (1976) Elements of food engineering. The AVI Publishing Co., Westport, CT
Heldman DR (1975) Food process engineering. The AVI Publishing Co., Westport, CT
IFT Task Force Report (2000) Kinetics of microbial inactivation for alternative food processing
technologies. JFS Special Supplement. pp 1–108
Knorr D (2006) Processing concepts for non-thermal modification of foods. 13th World Food
Congress of Food Science & Technology. Nantes, France. Sept. 17–21
Labuza TP, Tannenbaum SR, Karel M (1970) Water content and stability of low moisture and
intermediate moisture foods. Food Technol 24:543–550
Leninger HA, Beverloo WA (1975) Food process engineering. Reidel, Dordrecht, The Netherlands
Loncin M, Merson RL (1979) Food engineering. Principles and selected application. Academic,
New York
Parker ME, Harvey EH, Stateler ES (1952) Elements of food engineering. Reinhold, New York
Rao MA, Rizvi SSH, Datta AK (2005) Engineering properties of foods, 3rd edn. CRC Press/Taylor
& Francis, Boca Raton, FL
Roos YH, Karel M (1991) Applying state diagrams to food processing and development. Food
Technol 45(12):68–71, 107
Singh RP, Heldman DR (1984) Introduction to food engineering. Academic, Orlando, FL
Toledo RT (1980) Fundamentals of food process engineering. The AVI Publishing Co., Westport,
CT
Villota Ricardo, Hawkes James G (2007) Reaction kinetics in food systems. In: Heldman Dennis
R, Lund Daryl B (eds) Handbook of food engineering, 2nd edn. CRC Press/Taylor & Francis,
Boca Raton, FL
von Bockel MA (2008) Kinetic modeling of reactions in foods. CRC Press/Taylor & Francis, Boca
Raton, FL
von Linde C (1896) Process and apparatus for liquefying gases or gaseous mixtures, and for
producing cold, more particularly applicable for separating oxygen from atmospheric air.
German Patent # GB189512528
von Loesecke HW (1943) Drying and dehydration of foods. Reinhold, New York
Chapter 2
Advances in 3D Numerical Simulation of Viscous
and Viscoelastic Mixing Flows
Kiran V. Vyakaranam and Jozef L. Kokini
2.1 Introduction
Mixing processes in the food industry often involve highly viscous and viscoelastic
fluids like wheat flour dough, pastes, batters, and syrups. The design of mixing
equipment, whether batch or continuous, is aimed at achieving a well-mixed and
blended product with a consistent rheological character. Design of mixing equipment
also involves devising guidelines for scale-up of mixing devices and their compari-
son in terms of mixing efficiency, especially when batch mixers are to be replaced by
continuous mixing equipment. A well-designed mixing process could result in the
blending of ingredients, improvement of rate of heat transfer, facilitation of chemical
reactions, creation of structure, addition of energy to create or break molecular
bonds, etc. In order to evaluate the efficiency of a mixer design in performing
these operations, a kinematic analysis of the process can be made wherein the mixing
mechanisms are characterized as “dispersive” or “non-dispersive” (or extensive)
(Wang and Manas-Zloczower 2001). Dispersive mixing is quantified by elongational
flow and shear stress, which aid in the breakup of cohesive particles like droplets,
bubbles, or solid agglomerates. Non-dispersive mixing includes the spatial separa-
tion and rearrangement of an initially cohesive cluster of particles (distributive
mixing) or the laminar stretching and folding of the material elements, resulting in
a homogeneous distribution (Meijer and Janssen 1994).
Mixing devices currently used in process industries are either batch or continu-
ous. These devices vary in the design of the impellers, kneading blocks, and/or
static mixing elements, depending on the material used, the throughput require-
ments, and the specific aim of the mixing process, for example, the stretching and
folding of material or dispersion of ingredients. Examples of batch mixers include
K.V. Vyakaranam
Department of Food Science, Rutgers University, New Brunswick, NJ 08901, USA
J.L. Kokini (*)
Department of Food Science and Human Nutrition, College of ACES, University of Illinois at
Urbana Champaign, Urbana, IL 61801, USA
e-mail: kokini@uiuc.edu
20 K.V. Vyakaranam and J.L. Kokini
the stirred tank reactors with Rushton impellers and dough kneaders like the
Brabender farinograph and Banbury mixers, while examples of continuous mixers
include the twin-screw/single-screw extruders and static mixers fitted with varying
mixing elements (e.g., Kenics mixer). While the evaluation of flow and subsequent
mixing parameters can be done using visualization techniques like particle image
velocimetry (PIV) and laser doppler anemometry (LDA), it is often not practical to
optimize the design of such large mixing devices (i.e., to industrial scale) through
experimental trial-and-error analysis. Numerical simulation techniques thus have
proven to be a viable, nonintrusive, and cost-effective alternative tool that can
optimize the design of complex mixer geometries through analysis of the flow field
and mixing parameters.
Research involving numerical simulation of mixing flows has been under steady
development, ranging in the beginning from the simplest 2D problems to more
realistic 3D mixer geometries. In this chapter we initially discuss the basic ideas
behind numerical simulation of mixing flows, using the finite element method
(FEM) and the governing equations and theoretical measures of the mixing process.
A review is then presented of the recent work done in analyzing flow and mixing of
viscous and viscoelastic fluids in various batch and continuous mixing geometries
using FEM simulations.
2.2 Theoretical Measures of Mixing
There are several measures used to quantify distributive and dispersive mixing, as
well as laminar stretching and folding of the material. During mixing of two
components (distributive mixing) the homogeneity of the mixture can be quantified
through the scale of segregation, S(t), which is a measure of the binomial distribu-
tion of the components and is calculated as
ðz
SðtÞ ¼ Rðr; tÞdr (2.1)
0
where Rðjr jÞ is a correlation coefficient that gives the probability of “M” pairs
of material points in the mixer separated by a distance jr j having the same
concentration,
P
M
ðc0 j cÞ ðc00 j cÞ
j¼1
Rðjr jÞ ¼ (2.2)
MS2
The parameter S(t) gives an indication of the average size of the segregated
regions but cannot detect local defects in the flow.
2 Advances in 3D Numerical Simulation of Viscous and Viscoelastic Mixing Flows 21
When a cluster of material points is distributed, the difference between the actual
distribution and the ideal distribution of the material points is called the “pair-wise
distribution index” or the “cluster distribution index” e, and is calculated as
Ð
1 2
cðrÞ cðrÞideal dr
0
e¼ Ð
1 2 ; (2.3)
cðrÞideal dr
0
where c(r) is the coefficient of the probability density function and the value of index
e varies from 0 (ideal distribution) to 1 (no distribution) (Connelly and Kokini 2007).
The efficiency of a mixing flow to stretch and fold the material can be studied
using a kinematic approach wherein the deformation of infinitesimal material lines
and surface elements is tracked (Ottino 1989).
If the motion of the fluid is described by X ¼ wðX,t), and the deformation of an
infinitesimal material line by dx¼FdX, then the length of stretch l can be defined
in terms of the strain as
jdxj
l ¼ lim (2.4)
jdxj!0 jdXj
and a local instantaneous efficiency of mixing given by
D ln l=Dt
el ðX;M;tÞ ¼ 1 (2.5)
ðD : DÞ2
where M ¼ jdX dx 1/2

j and D represents the rate of strain tensor with a magnitude (D:D) .
If the material is an incompressible Newtonian fluid, the local instantaneous
efficiency will be the fraction of dissipated energy used to stretch the material,
ranging between 1 and 1. The time averaged efficiency hel i given in (2.6)
provides information on the nature of reorientation (stretching and folding) of the
flow. Flows with no reorientation will have hel i decaying with time as t1, whereas
hel i tending toward a constant value would indicate a flow with strong reorientation
(Ottino 1989).
ðt
1
hel iðX; M; tÞ ¼ el ðX; M; t0 Þdt0 (2.6)
t
0
Similar efficiency in stretching of an infinitesimal area element can also be

evaluated. In a chaotic time-periodic flow, the arithmetic and geometric means of
the length of stretch, l, grow exponentially and can be represented as

l a eYn ; hli a eLn (2.7)
where n is the number of periods of revolutions and Y and L are the topological
entropy exponent and Lyapunov exponent, respectively. While topological
entropy is a measure of the mixing rate in chaotic regions of flow, the Lyapunov
exponent measures the rate of elongation or stretching (Muzzio et al. 2000; Zalc
et al. 2002a).
Dispersive mixing involves breakup of agglomerates and drops in flow,
caused by stresses large enough to overcome the cohesive or interfacial forces
that tend to keep the agglomerate or the drop intact. The mechanical stress
required for breakup depends on the type of flow, with pure elongation or
irrotational flows being more effective than flows with a rotational component
(Grace 1982; Bentley and Leal 1986). A dispersive “mixing index” lMZ, which
quantifies the relative strength of the pure elongational flow component, can be
defined as
jDj
lMZ ¼ (2.8)
jDj þ jOj
where D and V are the rate of deformation and vorticity tensors, respectively (Yang
and Manas-Zloczower 1992). lMZ ranges from 0 for pure rotation to 0.5 for simple
shear, and to 1.0 for pure elongation. The Manas-Zloczower mixing index
(as defined above) is not frame-invariant. However, several other frame-invariant
flow-type indices were found to be computationally difficult to evaluate due to the
requirement of higher-mesh densities in the simulation; and provided information
similar to that of the non-frame-invariant index (Wang and Manas-Zloczower 2001;
Connelly 2004).
2.3 Governing Equations for Calculation of Flow
The velocity and pressure distributions for an incompressible and isothermal fluid
flow are calculated from the Navier–Stokes equations of mass and momentum
conservation:
rv¼0 (2.9)

@v
r s þ rf¼r þ v:rv (2.10)
@t
where r is the fluid density and f is the external body force per unit mass.
The stress tensor s given in (2.10) incorporates the isotropic pressure (P) and
extra stress tensor (T), defined as
s ¼ PI þ T (2.11)
An additional energy conservation equation has to be solved in the case of a non-

isothermal flow to obtain the temperature distribution:

@T
rCðTÞ þ v rT ¼ T : rv þ r r q (2.12)
@t
where C(T) is the dependence of heat capacity on temperature, r is the volumetric

heat source, q is the heat flux, and T:rv represents the viscous heating.
The extra stress tensor of (2.11) for a generalized Newtonian fluid in an iso-
thermal flow is given by
T¼ 2mD (2.13)
where D is the rate of deformation tensor and m(_g) is the viscosity function of the
local shear rate, g_ , as given in (2.14).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
g_ ¼ 2trðD2 Þ (2.14)
The simplest form of the viscosity function is in the case of a Newtonian fluid,
when it reduces to a constant value (m0) called Newtonian or zero-shear-rate
viscosity. Examples of shear-dependent viscosity models used for polymeric
liquids include the power-law model and the modified cross model (Prakash
1996); m can be calculated respectively as
m ¼ mð_gÞn1 (2.15)
m ¼ m0 ð1 þ kg_ Þn1 (2.16)
For viscoelastic fluids, the extra stress tensor T is divided into a viscoelastic part
(T1) and a purely viscous part (T2), given by
dT1
AðT1 ;lt Þ:T1 þ lt ¼ 21 D; T2 ¼ 22 D (2.17)
dt
where the relaxation time lt, the viscosity factor 1, and the function A (T1, lt)
depend on the specific viscoelastic model used. The common differential viscoelas-
tic models used to describe polymeric food materials are the Oldroyd-B, Maxwell,
White-Metzner, Phan-Thien-Tanner, and Giesekus models (Connelly 2004).
2.4 Numerical Methods for Simulation of Mixing Flows
For a given system of fluid and mixing process, the solution of the above set of
equations (consisting of conservation of mass and momentum), the energy equation
(for non-isothermal problems), and the appropriate constitutive equations can be
obtained by several numerical techniques. The FEM is the most commonly used
technique for numerical simulation of viscous mixing flows, even though other
methods like the Finite Difference Method and the Finite Volume Method have
been used (Dhanasekharan and Kokini 2003; Connelly and Kokini 2003; Heniche
and Tanguy 2008).
Simulating a mixing flow using the FEM technique involves three steps:
l Construction of flow volume (or domain of flow problem) into a mesh made up
of several finite elements or sub-domains
l Derivation of algebraic equations relating the physical quantities between the
element nodes
l Solving whole flow domain: assembling equation parts using continuity and/or
balancing the physical quantities across the elements (Reddy 2006)
The mesh generation step requires the construction of the mixing flow domain
geometry using a network of linear triangular and quadrilateral elements (2D) or
hexahedral, prism, tetrahedron, and wedge elements (3D). Several Computational
Fluid Dynamics (CFD) software suites include mesh generation software, for
example the Gambit (ANSYS Inc., Lebanon, NH) mesh generator. While main-
taining the coarseness of the mesh is important in keeping the computational costs
low, a higher density of finer mesh elements is needed in areas involving high
gradients of flow properties. Hence, it is important to strike a balance with the
coarseness/fineness of the mesh depending on the available computational capabil-
ities. Additionally, for a higher accuracy and quicker convergence of the solution,
the transition from coarser to finer mesh element regions has to be smooth and the
individual elements as close as possible to ideal shapes, for example square/
equilateral shapes (Connelly 2004; Heniche and Tanguy 2008). In the next step,
the governing equations of motion are discretized using a weighted residual method
and approximation of the flow variables for each element. For a generalized
Newtonian fluid, the velocity, pressure, and stress fields (in case of viscoelastic
flows) are approximated using the following equations
X
vh ¼ Vi c i (2.18)
X
ph ¼ p i pi (2.19)
X
Th ¼ Ti fi (2.20)
where ci, pI, and ’i are the finite element basis functions and Vi, pi, and Ti are the
nodal variables.
There are several methods available to discretize the governing equations, for
example the Galerkin, the pressure stabilized Petrov-Galerkin, and the Galerkin
least-square methods. In the Galerkin method, assuming the inertia terms to be
negligible, the set of finite element equations in the flow domain O are formulated as
ð
pk ½r:va dO ¼ 0 (2.21)
O
ð a
ð
Dv
cj r f þ rcTj : pa I þ 22 Da þ Ta1 dO ¼ cj s:nds (2.22)
Dt
O O
ð
dT1
fi gðT1 Þ:Ta1 þl 21 D dO ¼ 0
a
(2.23)
dt
O
For differential viscoelastic models, this system of equations can be solved by

using a coupled method where the extra stress tensor is the primary variable along
the pressure and velocity fields; the Newton–Raphson technique can be used to
solve for the variables. In this method, one usually encounters a large number of
unknowns and high computational costs. Another method often used is the
decoupled method where the viscoelastic stress tensor is computed separately
from the flow kinematics; the method is iterated using the Picard scheme. The
accuracy and stability of Galerkin formulations deteriorates in viscoelastic flow
problems as the elasticity number increases; thus further stabilization techniques
such as streamlined upwind (SU) and elastic-viscous-stress splitting (EVSS) are
needed. The reader is directed to literature by Connelly (2004) and references
therein for a detailed review of these techniques.
Several commercial FEM simulation packages are available, such as FIDAP
(ANSYS, Inc.), AcuSolve (Acusim Software), Polyflow (ANSYS, Inc.), and
Poly3D (Rheosoft). In particular, the Polyflow suite, which includes a mesh gener-
ator (Gambit), an FEM solver (Polyflow), and a post-processor (Fluent-Post/Field-
view/CFX Post), has been extensively used by our group to simulate viscous and
viscoelastic flows in mixers and extruders.
Solving for the flow in a mixing device using a CFD simulation package would
include the following steps:
l Building mixer geometry and converting into FEM mesh using a mesh generator
(e.g., Gambit)
l Defining flow boundaries, material properties, numerical parameters, and
operating conditions
l Solving of velocity profiles using an FEM solver (e.g., Polyflow)
l Calculating various mixing measures from velocity data using a post processor
(e.g., Fluent-Post/Fieldview/CFX-Post)
Simulation of mixers with geometry that varies continuously with time, as in the
case of moving impellers or screw/paddle elements, requires special meshing
techniques. One approach that can be applied to a single impeller or a single
screw mixer is the “rotating reference frame technique” in which the impeller/
screw becomes the fixed reference frame with the barrel rotating around it. This
enables the flow domain mesh to be fixed in time. The velocities are then
transformed back to the inertial reference frame. In asymmetrical mixer geo-

metries, as in the case of an eccentrically placed impeller or a twin screw
continuous mixer design, a “mesh superposition technique” (MST) can be
used in which the impeller or the screw/paddle elements are meshed separately
from the flow domain and then superimposed (Connelly and Kokini 2006a).
Here, the equation of motion is modified by the introduction of a penalty term
(H) to distinguish the solid (impeller or paddle) mesh elements from the fluid
elements:
H ðvvÞ þ ð1 HÞðrP þ r:T þ rg raÞ ¼ 0 (2.24)
The penalty term H is essentially a step function and is set to a value of 1 for all
nodes on solid moving parts, and a value of 0 for all nodes in the fluid volume. The
continuity equation is also modified with a relative compression factor b to account
for conservation of mass in the regions of flow domain covered by the moving solid
elements:
b
r:v þ DP ¼ 0 (2.25)
m
2.5 3D Numerical Simulation of Model Mixing Geometries
In this section we present a review of studies evaluating the efficiency of distribu-

tive and dispersive mixing in model batch and continuous mixing geometries using
numerical simulation techniques. These studies examined the effects of mixer
operating parameters like screw design, screw speed, and material rheology on
the variations in flow and mixing profiles. Recent studies in the area of numerical
simulation and CFD of mixing processes can be broadly classified based on the
mixing geometries investigated, for example, stirred tank reactors (Zalc et al. 2001,
2002a; Alvarez-Hernández et al. 2002; Rivera et al. 2004, 2006; Barailler et al.
2006; Iranshahi et al. 2006, 2007), dough kneaders (Jongen 2000; Jongen et al.
2003; Connelly and Kokini 2004, 2006a,b, 2007), static mixers (Rauline et al. 2000;
Zalc et al. 2002b, 2003; Heniche et al. 2005), and continuous mixers/extruders
(Dhanasekharan and Kokini 2000, 2003; Wang and Manas-Zloczower 2001).
2.5.1 Stirred Tank Reactors/Batch Mixers
Mixing of viscous liquids in stirred tank reactors and impeller-based batch mixers is
widely employed in many process industries where effective distribution of addi-
tives is of great importance. Impeller design and speed of the impeller (as quantified
a
0.24 m
b
0.36 m
0.08 m
0.06 m 0.08 m
Fig. 2.1 Examples of impeller designs in batch mixers: (a) Three-Rushton turbine (Zalc et al.
2001); (b) Paravisc impeller (Iranshahi et al. 2006); (c) Rushton turbine with coaxial anchor
(Rivera et al. 2006)
by Reynold’s number, Re) are the major operating variables that can be tweaked to
attain optimum mixing conditions and power efficiency. Figure 2.1 shows a few of
the impeller designs used in stirred tank reactors and impeller-based mixers. The
geometrically simple and symmetrical construction of the impellers, coupled with a
relatively large flow volume in the reactor/vessel (usually cylindrical), makes these
mixers an excellent choice for use of CFD in the calculation and visualization of the
various mixing measures as well as their validation using imaging techniques like
the PIV.
Laminar mixing in a three-Rushton turbine stirred tank reactor was studied
using the ORAC CFD package (Dantec Dynamics, Mahwah, NJ) (Zalc et al.
2001, 2002a) (Fig. 2.1a). The flow measurements were validated with PIV
experiments using planar laser-induced fluorescence. In Fig. 2.2 the simulation
results are in excellent agreement with the experiments in revealing the size and
location of poorly mixed regions in the mixer. Local mixing efficiency can be
quantified and visualized by computing the stretching value (l). The stretching
value l is calculated from deformation of infinitesimal vectors in the flow, which
usually takes place at an exponential rate in chaotic flow regions as compared to a
linear rate in non-chaotic flows. The contour maps of the logarithm of stretching
shown in Fig. 2.3 reveal the spatial heterogeneity of stretching when mixing at
various impeller speeds after 20 revolutions. Knowledge of the distribution
patterns of stretching could be valuable in deciding the injection point for
additives to achieve optimal distribution (Zalc et al. 2002a).
a b
Fig. 2.2 Comparison of

experimental and simulated
results showing excellent
agreement in the mixing
patterns in a three-Rushton
turbine impeller mixer after
600 revolutions at
(a) Re ¼ 20; and (b) Re ¼ 40
(Zalc et al. 2002a)
Fig. 2.3 Contour plots of ln(l) reveal the spatial heterogeneity of stretching at various impeller
speeds: (a) Re ¼ 20; (b) Re ¼ 40; (c) Re ¼ 160 (Zalc et al. 2002a)
In order for a mixing process to deliver optimum distributive and dispersive

mixing, the entire set of operating conditions needs to be considered, that is, the
impeller design, impeller speed, fluid rheology, and the injection point for addi-
tives. While comparing the mixing performance of an Ekato Paravisc impeller with
a Double Helical Ribbon (DHR) impeller (Fig. 2.2b), Iranshahi et al. (2006) showed
that the Paravisc impeller was more sensitive to the injection point chosen for a
tracer in the effective distribution of the tracer particles in a Newtonian fluid
(Fig. 2.4).
An important criterion for many mixing processes is the balance between
distributive and dispersive mixing. Rivera et al. (2006) defined a parameter that
is a ratio between two dimensionless quantities, the “head number” Nh (represent-
ing the shearing ability of the impeller) and the “flow number” Nq (representing
the pumping ability of the impeller). The effect of an anchor co-rotating and
counter-rotating with a Rushton turbine (Fig. 2.2c) was studied for a Newtonian
and a non-Newtonian fluid by simulating the flow and mixing process, using
POLY3DTM (Rheosoft, Inc.) finite element software. It can be seen from Table 2.1
that the co-rotating anchor was more effective in combined distribution and
dispersion, while the counter-rotating anchor was poor at distribution. The poor
distributive mixing with the counter-rotating mode can be visualized in Fig. 2.5,
which shows the intensity of segregation of tracer particles (a measure of homo-
geneity of the mixing) at two different mixing times. Rivera et al. (2006) refer to
this figure as “tracer dispersion,” whereas the actual process measured by the
Fig. 2.4 Intensity of segregation vs. time used for different Re and points of injection of tracer
(Iranshahi et al. 2006)
Table 2.1 Comparison of shearing and pumping abilities of the coaxial mixer for different
rotating modes and fluid rheology (Rivera et al. 2006)
Operating conditions Nq Nk Nh/Nq
Co-rotating Newtonian fluid 0.917 1.546 1.685
Rushton impeller only Newtonian fluid 0.658 1.474 2.477
Counter-rotating Newtonian fluid 0.756 1.631 1.948
Co-rotating non-Newtonian fluid 0.761 0.865 1.136
Rushton impeller only non-Newtonian Fluid 0.452 1.082 3.222
Counter-rotating non-Newtonian fluid 0.608 1.457 1.778
intensity of segregation is distributive mixing. Dispersive mixing in our definition

relates to the breakup (e.g., of tracer drops) caused by the shearing action of the
impellers. The rheology of the fluid also affected the efficiency of the mixing process
in that the shear thinning of the viscosity resulted in smaller well-mixed areas.
2.5.2 Dough Mixers and Kneaders
Dough kneaders and mixers are designed for highly viscous doughs and batters that
require actions that include pushing portions of the material through other portions,
elevating, dropping and rotating the material, and cutting or dividing the material
a b
Z Y Z Y
X X X X
Y Z Y Z
c d
Z Y Z Y
X X X X
Y Z Y Z
Fig. 2.5 Distributive mixing in a Rushton turbine impeller with a coaxial anchor: (a) co-rotating
after 15 s; (b) co-rotating after 150 s; (c) counter-rotating after 15 s; (d) counter-rotating after 150 s
(Rivera et al. 2006)
(Connelly 2004). Most dough mixers are hence built to operate horizontally, and
typically use co-rotating twin blades or arms like roller bars, single and twin sigma
blades, open paddle four-way blades, double-arm blades, and spindle geometries.
However, several smaller batch dough kneaders with Z blades, sigma blades, or
rotating arms are used for empirical dough testing in laboratories. Examples include
the plastograph (also called a farinograph) with two counter-rotating but non-
intermeshing blades inside a closed cavity; the do-corder with two intermeshing
blades co-rotating in a bowl, with limited clearance at the bowl wall; the planetary
mixer (or mixograph) with mixing caused by the planetary motion of a rotating arm
attached with a pair of vertical pins on either end; and the Eberhart spiral mixer with
a spiral rod rotating in a bowl (Jongen et al. 2003) (Fig. 2.6).
Research aimed at optimizing batch kneaders involves experimental methods
relating the operating parameters of these mixers to the dough development,
rheology, and sensory quality of the product sample. Only a few studies were
able to gather specific information on the flow profiles and mixing parameters
due to the inherent difficulties with experiments (Prakash et al. 1999; Prakash and
Kokini 2000). However, due to the recent advances in the design and simulation
capabilities of various CFD software packages, specifically those aimed at viscous
and viscoelastic flows, it is now possible to obtain (in a nonintrusive way) detailed
information on flow and mixing parameters that includes distribution, dispersion,
and laminar stretching in dough mixing.
a b L
H ω
ωL ω2
Rb ω
Rb H
Ro Ro
c d
Fixed Pin
ω2
Rb
ωL
R
Dough Volume
L
L2 Moving Pin
R1
ω2
Ro
ω2
d
ω1
Fig. 2.6 Dough kneaders of various geometries: (a) Farinograph; (b) Do-corder; (c) Mixograph;
(d) Spiral mixer (Jongen et al. 2003)
Jongen et al. (2003) used a virtual FEM to study the deformation of viscous pastes
in four different batch mixing geometries (Fig. 2.6). The technique involved super-
imposing the moving elements on the flow domain FEM mesh with the FIDAP
(Ansys Inc., Lebanon, NH) FEM software. In order to quantify the deformation of the
material, a flow parameter R2 was defined as the ratio of the second invariants of the
shear rate and vorticity rate tensors. The normalized value of R2 was calculated as:
1 R2
D¼ (2.26)
1 þ R2
A comparison of the cumulative time-averaged frequency function for the three

types of mixers (Fig. 2.7) showed that the spiral mixer had the most rotational flow;
the do-corder the most elongational flow character; and the planetary mixer and
plastograph the highest (approximately) shear to the material. An investigation of
the same parameter at different operating conditions in the plastograph showed that,
while rotational speed of the blades and gap clearance did not affect the type of
deformation (value of D), a change in the blade design helped reduce the rotational
character of the mixing profile.
A more in-depth and detailed analysis of the 3D flow and mixing profiles in the
Brabender® farinograph was performed by Connelly and Kokini (2006a, b) using
MST in which the moving sigma blades and the barrel volume were separately
meshed and superimposed. An FEM along with the MST was implemented in the
a 1 b 1
0.9 Plastograph Reference
Do-corder Slow speed
Cumulated frequency F
Cumulated frequency F
0.8 Planctary 0.8 Large gap
0.7 Spiral Half blade
0.6 0.6
0.5
0.4 0.4
0.3
0.2 0.2
0.1
0 0
–1 –0.8 –0.6 –0.4 –0.2 0 0.2 0.4 0.6 0.8 1 –1 –0.8 –0.6 –0.4 –0.2 0 0.2 0.4 0.6 0.8 1
Deformation type D Deformation type D
Fig. 2.7 (a) Cumulative distribution of mean flow-type parameter D for various kneader geome-
tries; (b) cumulative distribution of mean flow-type parameter D for a farinograph at different
blade configurations and speeds (Jongen et al. 2003)
Polyflow® CFD software (ANSYS Inc., Lebanon, NH). Figure 2.8 shows the effect
of fluid rheology on the velocity contour maps on the vertical center plane between
the two blades. Increased shear thinning of the fluid caused a decrease in the
velocity of the fluid right above the blades. The dispersive mixing ability was
calculated as defined by (2.8). A histogram of the distribution of the dispersive
mixing index in the central planes between the farinograph blades for the three
fluids shows a decrease in the elongational flow regions with increasing shear
thinning. The paths of material points can be calculated from the velocity profiles,
which can then be used to evaluate the distributive and stretching efficiency
parameters such as the cluster distribution index, scale of segregation, and the
mean length of stretch. The density of probability of the length of stretch experi-
enced by 10,000 infinitesimal material lines in the farinograph (Fig. 2.9) for three
blade revolutions showed a gradual and steady increase in the amount of material
undergoing effective stretching over time.
Limited work has been done on studying the effect of a fluid’s viscoelasticity on
flow and mixing profiles. Connelly and Kokini (2003) simulated the flow of a Phan-
Thien Tanner fluid model in a 2D single screw mixer using a rotating reference
frame technique; they compared different methods of handling the instabilities in
dealing with a differential viscoelastic fluid model. It was found that the effect of
viscoelasticity (an increase in relaxation times) was to create asymmetry in the
velocity and pressure profiles and to reduce the effects of shear thinning on the
pressure and stresses (Fig. 2.10).
2.5.3 Continuous Mixers and Extruders
Simulation of continuous mixer geometries requires increased computational costs

due to the complexity and lack of symmetry in the geometry. The most common
continuous screw-type mixer and extruder design used in the food industry is the
a Corn syrup Velocity magnitude x-velocity component

20 10
18 8
16 6
14 4
12 2
10 0
8 –2
6 –4
4 –6
2 y x –8 y x
0
z –10
z
b 2% CMC
20 10
18 8
16 6
14 4
12 2
10 0
8 –2
6 –4
4 –6
2 y x –8 y x
–10
0 z z
c 0.11% Carbopol
20 10
18 8
16 6
14 4
12 2
10 0
8 –2
6 –4
4 –6
2
y x –8 y x
0
z –10 z
Fig. 2.8 Effect of fluid rheology on the velocity profiles in the vertical center plane of a
farinograph (Connelly and Kokini 2006a)
two-lobed twin-screw co-rotating design, for example the Readco® continuous

mixer shown in Fig. 2.11. The design consists of a pair of co-rotating shafts fitted
with a series of conveying screws, two- or three-lobed kneading blocks that can be
staggered followed by discharge screw elements. There have been several analyti-
cal, experimental, and numerical studies conducted to decipher the flow and mixing
patterns in both the conveying screw regions and the kneading region. As a first step
0.50
0.45 1 blade
2 blade
0.40 3 blade
Density of Probability 0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
–6.0 –3.0 0.0 3.0 6.0 9.0 12.0 15.0 18.0 21.0
In (stretch)
Fig. 2.9 Density of probability for length of stretch ln(l) experienced by 10,000 infinitesimal
material lines in a farinograph (Connelly and Kokini 2006b)
3e+04 a b
1.7e+04
4e+03
–9e+03
–2.2e+04
–3.5e+04
–4.8e+04
–6.1e+04
–7.4e+04
–8.7e+04
–1e+05
Fig. 2.10 Shear stress (g/cm2) contour maps for fluid with relaxation times of (a) 0 s and (b) 100 s
in a 2D single-paddle mixer (Connelly and Kokini 2003)
to studying the twin-screw continuous mixer design, Connelly and Kokini (2007)
used numerical simulations to compare mixing of a Carreau model fluid in both a
2D single-paddle mixer and a 2D twin-paddle mixer. The presence of a second
paddle element in the twin-screw geometry significantly improved both dispersive
a b c
Y Y Y
Z X Z X Z X
Fig. 2.11 Kneading elements of the Readco® twin-screw processor (Readco Inc., York, PA)
arranged in (a) FLAT, (b) 45F, and (c) 45R configurations (Vyakaranam and Kokini 2008)
a b
1 1
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
Y Y
0.1 0.1
Z X Z X
0 0
1
c 1
d
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 Y 0.2 Y
0.1 0.1
Z X Z X
0 0
Fig. 2.12 Mixing index (lMZ) contours for the (a) single-screw and twin-screw mixers after
rotation of the paddles at (b) 45 , (c) 67.5 , and (d) 90 (Connelly and Kokini 2007)
and distributive mixing. Figures 2.12 and 2.13 shows the dispersive mixing index
and the cluster distribution index in both mixers. While the twin-screw geometry
shows an increase in the area and magnitude of elongational flow, the single-screw
geometry showed a cyclic cluster distribution index, suggesting that the material
points were unable to leave the streamlines, leading to poor distributive mixing.
While these simulations were useful in showing the positive effect of a twin-
paddle element on the mixing, 3D simulations of the full-length mixer are needed
for a realistic simulation of continuous mixing, which involves axial flow. The
study of continuous mixers is mainly focused on evaluating the effects of screw
speed, stagger angle, and thickness of the paddle elements on the dispersive and
distributive mixing measures, and comparing the performance to a batch mixer.
1.00
Twin Screw-b1
0.90 Single Screw-b1
Twin Screw-b2
0.80 Single Screw-b2
Cluster Distribution Index 0.70

0.60
0.50
0.40
0.30
0.20
0.10
0.00
0 60 120 180 240 300 360 420 480 540 600
Time (s)
Fig. 2.13 Comparison of cluster distribution index for ten revolutions in the single-screw and
twin-screw mixers with different locations for the initial cluster in the flow domain: b1 (center
location) and b2 (leftmost location) (Connelly and Kokini 2007)
For example, consider the 3D numerical simulation of the mixing of a generalized

Newtonian fluid in a full-length kneading section of the Readco® continuous
processor with nine paddle pairs. The flow data were evaluated for every 10 of
movement by the paddles for ten rotations at three different screw speeds (1, 50, and
100 RPM); next the various distributive and dispersive mixing measures were
evaluated. The paddle elements were staggered forward and reversed at an angle
of 45 (Fig. 2.11).
Figure 2.14 shows the effects of screw speed and stagger angle on the time-
averaged efficiencies of the three continuous mixer geometries in comparison
with the farinograph. The time averaged instantaneous mixing efficiency
decreased with increasing screw speed, with highest efficiency at the lowest
screw speed of 1 RPM, which has been attributed to higher residence time of
the material at lower screw speed resulting in greater stretching and folding of the
material. The FLAT (or neutral, with no element stagger) configuration showed
the highest efficiency (Ashokan 2008).
Mixing efficiency is also affected by the individual paddle element width and the
increase in the number of gaps between the elements, as a result of increasing the
number of elements. Figure 2.15 shows the area stretch (as a measure of distributive
mixing) of different kneading element widths for two-lobe kneading elements.
The smaller the disc width, the better was the area stretch; it was hypothesized
that the increase in the number of gaps caused an increase in the area stretch
(Ishikawa et al. 2001).
Yoshinaga et al. (2000) studied the isothermal flow of a Carreau model non-
Newtonian fluid in the kneading section of a twin-screw extruder using FEM
a 0.30
b 0.18
Mean Time Averaged Efficiency
Mean Time Averaged Efficiency

0.16
0.25
0.14
0.20 0.12
0.10
0.15
0.08
0.10 0.06
100 rpm 0.04 Flat
0.05 50 rpm 45F
1 rpm 0.02 45R
Batch Mixer Batch Mixer
0.00 0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 0 1 2 3 4 5 6
Rotations Time (s)
Fig. 2.14 Effect of (a) screw speed and (b) screw configuration on time-averaged mixing
efficiency of the Readco® twin-screw continuous mixer and the farinograph batch mixer (Ashokan
2008)
0.007
NNN
0.006
NNW
Distribution function
0.005
NWW
0.004
0.003
Fig. 2.15 Distribution of the 0.002

area stretch in mixing region
of a co-rotating twin-screw 0.001
extruder for different
kneading disc widths (i.e., 0
NNN, NNW, and NWW) 0 100 200 300 400 500
(Ishikawa et al. 2001) In (Area stretch)
simulations. The shafts were fitted with five pairs of three-lobed mixing elements
configured at different stagger angles. Distributive mixing was measured by resi-
dence time distributions and the minimum distance between mass-less particle
markers after mixing for a given time. The simulations were performed as a quasi-
steady state wherein the velocity profiles were calculated for every 3 of rotation of
the screw elements. Both the residence time distributions and the mean nearest
distance between the markers showed the neutral stagger configuration to be the
most beneficial for distributive mixing (Fig. 2.16).
Dispersive mixing can be analyzed by evaluating the dispersive mixing index
and corresponding shear stresses (or shear rates, if a Newtonian fluid). Figure 2.17
shows the contour maps of shear rate and mixing index (lMZ) at the 1st, 4th, and 8th
a b
1.2 1.6
Cumulative distribution function (–)
Mean Nearest Distance between

1.0 R30
L30
Markers (mm)
0.8 1.2
0.6 N
0.4 0.8
R30
0.2 L30
Average Residence
N
Time = 4.8 s
0.0 0.4
0 10 20 0.0 0.5 1.0 1.5 2.0
Time [ s] Time [ s]
Fig. 2.16 (a) Residence time distribution and (b) mean nearest distance between markers in
mixing region of a co-rotating extruder fitted with three-lobed kneading discs (Yoshinaga et al.
2000)
paddle elements of the FLAT configuration for a Newtonian fluid mixed at 75 RPM.
The elongational flow, indicated by the red shade, is the highest in the center region
of the mixer (around the 4th paddle element), while highest shear rates are found in
the intermeshing regions. The elongational flow is known to cause a breakup of
agglomerates and drops in flow; the corresponding shear rate can be made dimen-
sionless by using the capillary number (Ca). The Ca can be defined as the ratio of
the viscous forces acting to deform the drop to the interfacial forces resisting
deformation and acting to restore its spherical shape. At high enough Ca, the
drop deformation becomes unsteady, leading to breakup. The equation for Ca is
given as:
m_g
Ca ¼ r (2.27)
s
where m is the viscosity of the continuous medium (food matrix), g_ is the local
shear rate, r is the bubble radius, and s is the surface tension of the food matrix
(Meijer and Janssen 1994; Risso 2000). The effect of paddle element stagger on
the distribution of Ca and lMZ for air bubbles in a Newtonian corn syrup is shown
in Fig. 2.18 for the FLAT, 45F (with a 45 forward element stagger), and the 45R
(with a 45 reverse element stagger) configurations. It can be seen that the density
of points in the mixer with Ca and lMZ, which are high enough for breakup,
decreases with a reverse stagger angle of the paddle elements, suggesting higher
breakup in the FLAT configuration. The effects of stagger angle and screw speed
on the shear stresses (or Ca) and distribution of flow type give useful information
that aids in the design of mixers with better dispersive abilities.
Numerical simulation of the flow of fluids modeled by a viscoelastic constitutive
equation in a complex 3D geometry is a nontrivial task. For example, the MST
Shear Rates at P1 Mixing Index at P1

1.52e+02 1.00e+00
1.37e+02 9.00e–01
1.22e+02 8.00e–01
1.06e+02 7.00e–01
9.12e+01 6.00e–01
7.60e+01 5.00e–01
6.08e+01 4.00e–01
4.56e+01 3.00e–01
3.04e+01 Y 2.00e–01 Y
1.52e+01 1.00e–01
Z X Z X
0.00e+00 0.00e+00
POLYFLOW Time Index: 0.80000001 POLYFLOW Time Index: 0.80000001
Contours of ca0.02 Aug 24, 2006 Contours of Mix Aug 24, 2006
FLUENT/Post 1.2 FLUENT/Post 1.2

1.59e+02 1.00e+00
1.43e+02 9.00e–01
1.27e+02 8.00e–01
1.11e+02 7.00e–01
9.55e+01 6.00e–01
7.96e+01 5.00e–01
6.37e+01 4.00e–01
4.78e+01 3.00e–01
3.18e+01 2.00e–01
Y Y
1.59e+01 1.00e–01
Z X Z X
0.00e+00 0.00e+00

1.40e+02 1.00e+00
1.26e+02 9.00e–01
1.12e+02 8.00e–01
9.78e+01 7.00e–01
8.39e+01 6.00e–01
6.99e+01 5.00e–01
5.59e+01 4.00e–01
4.19e+01 3.00e–01
2.80e+01 2.00e–01
Y Y
1.40e+01 1.00e–01
0.00e+00 Z X 0.00e+00 Z X
Fig. 2.17 Distribution of shear rate (_g) and mixing index (lMZ) at different locations in the twin-
screw mixer at a screw speed of 75 RPM (Vyakaranam and Kokini 2008)
technique in the POLYFLOW FEM solver does not support the use of a viscoelastic
constitutive equation; therefore a new technique is needed to model the fluid flow.
One such technique is the Pseudo Steady State (PSS) technique in which the flow is
modeled as a steady state with the viscoelastic constitutive equation for a snapshot
of the moving geometry; the velocity, pressure, and stress results are then used as
the input for successive time steps (Ishikawa et al. 2001). Preliminary demonstra-
tion of the technique using a 2D Readco® geometry was successful in obtaining the
velocity and shear rate profiles at low Deborah numbers and small rotations of the
paddle elements (Ashokan 2008). Further work is needed to extend the simulation
a Capillary Numbers for Mixing Index > 0.8 – Flat

configuration, 75 rpm - Bubble Diameter 400 micron
60
p1-Flat
Capillary Number, Ca
50 p4-Flat
p8-Flat
40
Ca-cr
30
20
10
0
0.8 0.82 0.84 0.86 0.88 0.9 0.92 0.94
Mixing Index
b Capillary Numbers for Mixing Index > 0.8 – 45F

60
45F-p8
50 45F-p4
45F-p1
40
Ca-cr
30
20
10
0
0.8 0.82 0.84 0.86 0.88 0.9 0.92 0.94
Mixing Index
c Capillary Numbers for Mixing Index > 0.8 – 45R

60
50 45R-p8
45R-p1
40
45R-p4
30 Ca-cr
20
10
0
0.8 0.85 0.9 0.95 1
Mixing Index
Fig. 2.18 Distribution of Ca and lMZ values for all calculated points in mixing region of the
Readco® twin-screw mixer at different screw configurations (Vyakaranam et al. 2009)
of viscoelastic simulations in full 3D geometry at higher Deborah numbers and

larger rotations of the paddles.
2.6 Conclusions
Recent studies in numerical simulation of viscous flow and mixing in the laminar regime
have shown the FEM to be a cost-effective and nonintrusive technique for analyzing and
visualizing the distributive and dispersive mixing ability in batch and continuous mixing
devices of various geometries. A review of recent work analyzing the mixing of viscous
and viscoelastic fluids using numerical simulations has been presented. Stirred tank
reactors and batch mixers of varying geometries have been successfully employed to
demonstrate the use of numerical simulations for visualization and analysis of viscous
and viscoelastic mixing in 3D. While the simulation of flow and mixing for fluids of
relatively simple rheological character (viscous Newtonian and non-Newtonian) has
been done for complex geometries at a 3D level, there is a need for further work in
simulating the mixing of more complex viscoelastic fluid models in full-scale continu-
ous mixer geometries.
Acknowledgments The authors would like to thank Dr. Nesli Sozer for help with proofreading
the document.
References
Alvarez-Hernández MM, Shinbrot T, Zalc J, Muzzio FJ (2002) Practical chaotic mixing. Chem
Eng Sci 57(17):3749–3753
Ashokan BK (2008) Developing methods for design and analysis of continuous mixers through 3D
numerical simulation of flow and mixing, Ph.D. thesis. Food Science, Rutgers University
Barailler F, Heniche M, Tanguy PA (2006) CFD analysis of a rotor-stator mixer with viscous
fluids. Chem Eng Sci 61(9):2888–2894
Bentley BJ, Leal LG (1986) An experimental investigation of drop deformation in steady, two-
dimensional linear flows. J Fluid Mech 167:241–283
Connelly RK (2004) Numerical simulation and validation of the mixing of dough-like materials
in model batch and continuous dough mixers, Ph.D. thesis: 430. Food Science, Rutgers
University
Connelly RK, Kokini JL (2003) 2-D numerical simulation of differential viscoelastic fluids in a
single-screw continuous mixer: application of viscoelastic finite element methods. Adv Polym
22(1):22–41
Connelly RK, Kokini JL (2004) The effect of shear thinning and differential viscoelasticity on
mixing in a model 2D mixer as determined using FEM with particle tracking. J Non-Newtonian
Fluid Mech 123(1):1–17
Connelly RK, Kokini J (2006a) 3D numerical simulation of the flow of viscous Newtonian and
shear thinning fluids in a twin sigma blade mixer. Adv Polym Technol 25(3):182–194
Connelly RK, Kokini J (2006b) Mixing simulation of a viscous Newtonian liquid in a twin sigma
blade mixer. AIChE J 52(10):3383–3393
Connelly RK, Kokini JL (2007) Examination of the mixing ability of single and twin screw
mixers using 2D finite element method simulation with particle tracking. J Food Eng 79
(3):956–969
Dhanasekharan M, Kokini JL (2000) Viscoelastic flow modeling in the extrusion of a dough-like
fluid. J Food Proc Eng 23(3):237–247
Dhanasekharan M, Kokini J (2003) Design and scaling of wheat dough extrusion by numerical
simulation of flow and heat transfer. J Food Eng 60(4):421–430
Grace HP (1982) Dispersion phenomena in high viscosity immiscible fluid systems and applica-
tion of static mixers as dispersion devices in such systems. Chem Eng Commun 14:225
Heniche M, Tanguy PA (2008) Finite element modeling of viscous mixing: a review. Chem
Product Process Model 3(1):55
Heniche M, Tanguy PA, Reeder MF, Fasano JB (2005) Numerical investigation of blade shape in
static mixing. AIChE J 51(1):44–58
Iranshahi A, Heniche M, Bertrand F, Tanguy PA (2006) Numerical investigation of the mixing
efficiency of the Ekato Paravisc impeller. Chem Eng Sci 61(8):2609–2617
Iranshahi A, Devals C, Heniche M, Fradette L, Tanguy PA, Takenaka K (2007) Hydrodynamics
characterization of the Maxblend impeller. Chem Eng Sci 62(14):3641–3653
Ishikawa T, Kihara S, Funatsu K (2001) 3-D non-isothermal flow field analysis and mixing
performance evaluation of kneading blocks in a co-rotating twin screw extruder. Polym Eng
Sci 41(5):840–849
Jongen T (2000) Characterization of batch mixers using numerical flow simulations. AIChE J 46
(11):2140–2150
Jongen TRG, Bruschke MV, Dekker JG (2003) Analysis of dough kneaders using numerical flow
simulations. Cereal Chem 80(4):383–389
Meijer MEH, Janssen JMH (1994) Mixing of immiscible liquids. In: Manas-Zloczower I, Tadmor Z
(eds) Mixing and compounding of polymers: theory and practice. Carl Hanser Verlag, New
York
Muzzio FJ, Alvarez MM, Cerbelli S, Giona M, Adrover A (2000) The intermaterial area density
generated by time- and spatially periodic 2D chaotic flows. Chem Eng Sci 55(8):1497–1508
Ottino JM (1989) The kinematics of mixing: stretching, chaos and transport. Press Syndicate of
University of Cambridge, Cambridge
Prakash S (1996) Characterization of shear rate distribution in a model mixer using laser doppler
anemometry. Ph.D. thesis: 315. Food Science, Rutgers University
Prakash S, Kokini JL (2000) Estimation and prediction of shear rate distribution as a model mixer.
J Food Eng 44(3):135–148
Prakash S, Karwe MV, Kokini JL (1999) Measurement of velocity distribution in the Brabender
farinograph as a model mixer, using laser-doppler anemometry. J Food Process Eng 22
(6):435–454
Rauline D, Le Blévec JM, Bousquet J, Tanguy PA (2000) A comparative assessment of the
performance of the Kenics and SMX static mixers. Chem Eng Res Des 78(3):389–396
Reddy JN (2006) An introduction to the finite element method. Tata McGraw-Hill, New Delhi
Risso F (2000) The mechanisms of deformation and breakup of drops and bubbles. Multiphase Sci
Technol 12:1–50
Rivera C, Heniche M, Ascanio G, Tanguy P (2004) A virtual finite element model for centered and
eccentric mixer configurations. Comp Chem Eng 28(12):2459–2468
Rivera C, Foucault S, Heniche M, Espinosa-Solares T, Tanguy PA (2006) Mixing analysis in a
coaxial mixer. Chem Eng Sci 61(9):2895–2907
Vyakaranam KV, Kokini J (2008) Study of the dynamics and size distributions of air bubbles
during mixing in a continuous food mixer. In: Campbell G, Scanlon M, Pyle L (eds) Bubbles in
food 2: novelty, health and luxury. AACC, St. Paul
Vyakaranam KV, Evans ME, Ashokan BK, Kokini JL (2009) Evaluation of mixing and air bubble
dispersion in viscous liquids using numerical simulations. In: Cullen PJ (ed) Food mixing:
principles and applications. Wiley-Blackwell, Oxford
Wang W, Manas-Zloczower I (2001) Temporal distributions: the basis for the development of
mixing indexes for scale-up of polymer processing equipment. Polym Eng Sci 41
(6):1068–1077
Yang H-H, Manas-Zloczower I (1992) Flow field analysis of the kneading disc region in a
co-rotating twin screw extruder. Polym Eng Sci 32(19):1411–1417
Yoshinaga M, Katsuki S, Miyazaki M, Liu L, Kihara S, Funatsu K (2000) Mixing mechanism of
three-tip kneading block in twin screw extruders. Polym Eng Sci 40(1):168–178
Zalc JM, Alvarez MM, Muzzio FJ, Arik BE (2001) Extensive validation of computed laminar flow
in a stirred tank with three Rushton turbines. AIChE J 47(10):2144–2154
Zalc JM, Szalai ES, Alvarez MM, Muzzio FJ (2002a) Using CFD to understand chaotic mixing in
laminar stirred tanks. AIChE J 48(10):2124–2134
Zalc JM, Szalai ES, Jaffer S, Muzzio FJ (2002b) Characterization of flow and mixing in an SMX
static mixer. AIChE J 48(3):427–436
Zalc JM, Szalai ES, Muzzio FJ (2003) Mixing dynamics in the SMX static mixer as a function of
injection location and flow ratio. Polym Eng Sci 43(4):875–890
Chapter 3
CFD: An Innovative and Effective Design
Tool for the Food Industry
Tomás Norton and Da-Wen Sun
3.1 Introduction
Fluids are ubiquitous in nature. In fact, many if not most industrial food processes
require work done on or by a fluid-based system during product development.
During the design and optimization of food processes, engineers must regularly
tackle problems that involve not only fluids but also complex phenomena such as
heat and mass transfer, phase change, and chemical reactions. Forming a compre-
hensive understanding of food-processing systems is not easy because the above
phenomena generally progress as invisible to the eye, and thus often their dynamics
cannot be easily realized or quantified. Moreover, when presented with complex
flow fields, which are governed by nonlinear dynamics, measurements alone may
not be sufficient for accurate analysis, as not only can equipment be intrusive but
too many measurement probes may also be required to fully diagnose a problem.
Computational fluid dynamics (CFD) is mainly concerned with the numerical
solution of the partial differential equations governing transport of mass, momen-
tum, and energy in moving fluids. However, most importantly, many processes
involving the motion of fluids and heat transfer, mass transfer, and chemical
reactions can be accurately described with CFD. Currently CFD is being univer-
sally used as an effective tool to investigate the spatiotemporal dynamics of many
interesting processes used by the food and beverage industry such as mixing,
drying, cooking, sterilization, chilling, and cold storage (Sun 2007). As a result,
the application and development of CFD for food processing is now proving to play
a fundamental role in problem solving for both industry and research, and its use by
the scientific community has grown exponentially within the last number of years
(Norton and Sun 2007). CFD has many advantages over traditional design and
analysis techniques, for example:
T. Norton and D.-W. Sun (*)

Food Refrigeration and Computerised Food Technology (FRCFT), University College Dublin,
National University of Ireland, Agriculture & Food Science Centre, Belfield, Dublin 4, Ireland
e-mail: dawen.sun@ucd.ie
46 T. Norton and D.-W. Sun
l Because food processes can be examined in a virtual environment, designs can

be achieved without encountering health and safety concerns or scaling issues
associated with physical prototyping, etc.
l Many designs of a single system/process can be explored on the computer
with virtual prototypes, thereby reducing wasteful trial and error techniques in
industry.
l Parametric studies can be easily completed in order to come up with design
solutions for a new geometry at a fraction of the cost of manufacturing the
various design options.
The fundamentals of CFD applications, its general applications in the food
industry, and the existing limitations and challenges that face current users of this
technology have been reviewed by some authors (Xia and Sun 2002; Norton and
Sun 2007). In light of this technology’s rapid development over the last number of
years, and the engineering challenges that face the food industry, this chapter gives
a comprehensive account of the latest advances made through successful CFD
applications in research. Firstly, the equations governing the physical mechanisms
encountered when modeling different processes in the food industry will be dis-
cussed, followed by an overview of how they are solved in a CFD environment. An
account of CFD modeling studies conducted on emerging food-processing techno-
logies will then be highlighted. Finally, an overview will be given on the challenges
still encountered in CFD modeling of food processes.
3.2 Modeling Food Processes: Solving Governing Partial

Differential Equations
For all food-processing systems, partial differential equations (PDEs) can describe
the intensity and spatiotemporal migration of thermodynamic and fluid-dynamic
phenomena. In order to develop and optimize food-processing systems, engineers
need to form a complete picture of the various mechanisms working within. Also, to
avoid the expense and time associated with physical experimentation, such insight
demands solutions to PDEs that describe the system. However, whether or not such
solutions are achievable is dependent on the physics involved, the level of precision
associated with the analysis tools at hand, and the amount of computing power
available to the engineer. In fact, because PDEs are generally by their very nature
time-consuming to solve, they require substantial computer resources; therefore
adequate solutions have not always been attainable with prevailing computer
power. In the past this meant that PDEs needed to be simplified so that they
could be manageably solved by hand, and assumptions impacting the quality of a
solution could easily be introduced. Furthermore, localized phenomena such as
mixing, fouling, cleaning, etc. cannot be accurately determined by simple models;
numerical analysis has always been required to establish the distribution of vari-
ables, that is, temperature, fluid velocity, and concentration in such systems.
3 CFD: An Innovative and Effective Design Tool for the Food Industry 47
Modern CFD codes have been developed using numerical algorithms that solve
the nonlinear PDEs governing fluid flow, heat transfer, and many other physical
phenomena. With this myriad of capabilities, CFD techniques can be used to build
distributed parameter models that are spatially and temporally representative of
food-processing systems, thereby allowing solutions with high levels of physical
realism. The accuracy of a CFD simulation is a function of many parameters,
including the level of empiricism involved, that is, via turbulence models or
additional physical models; the assumptions involved, that is, Boussinesq versus
the ideal gas approximation for inclusion of buoyancy effects; the simplification of
both geometry and boundary conditions to reduce processing time; and whether
processes modifying the physical mechanisms are included, that is, chemical
kinetics. As with other modeling techniques, the greater the number of approxima-
tions, the less accurate the CFD solution will be (Verboven et al. 2004). However,
if used correctly, CFD provides an understanding of the physics of a system in
greater detail through nonintrusive flow, thermal, and concentration field predic-
tions. Furthermore, after many years of development, CFD codes can now solve
advanced problems related to combined convection, radiation and conduction heat
transfer, flows in porous media, multiphase flows, and issues involving chemical
kinetics.
Many applications of CFD in modern food processing already exist, such as for
ovens, refrigerators, heat exchangers, in-place cleaning operations, spray dryers,
biosensors, and pasteurization techniques, among others. CFD simulation has made
giant leaps in realism with the yearly experience gained from each specialization in
tandem with the continual development of commercial CFD codes. For example, in
recent designs of multi-deck chilled cabinets, CFD simulations have been used to
visualize the impact of various design constraints without the need for physical
prototyping (Foster et al. 2005). Further simulations have revealed the impact of
these refrigeration systems on the occupant comfort in retail stores. In the advance-
ment of oven technology, CFD analyses have worked in conjunction with other
mathematical models of food quality attributes to optimize variations in the weight
loss and crust color of bread products prior to baking (Zhou and Therdthai 2007).
Also, in modern operations like air impingement ovens, CFD has been employed so
that the interaction of the jet flow pattern with the product can be understood;
spatially dependent heat and mass transfer coefficients can be predicted; and
equipment design parameters can be optimized (Kocer et al. 2007; Olsson and
Tr€agårdh 2007). In recent times, three-dimensional CFD modeling of plate heat
exchangers (PHEs) has allowed the milk fouling process to be simulated based on
both hydrodynamic and thermodynamic principles. This fouling model permits
assessing the influence of corrugation shapes and orientations on the PHE perfor-
mance (Jun and Puri 2006). The current standard for CFD modeling of spray dryers
has also raised the value of simulations by extending the basic models of flow
patterns and particle trajectories to sub-models such as drying models, kinetic
models of thermal reactions, sub-models describing product stickiness, and
agglomeration models (Straatsma et al. 2007). In other areas, CFD has recently
been used to model:
l Pasteurization of intact eggs (Denys et al. 2007); this application formerly had
little information concerning temperatures and process times required.
l Application of biosensors in the food industry (Verboven et al. 2007a); this
modeling allows the optimization of biosensor design in terms of fluid flow,
mass transfer, and chemical kinetics.
l Tea fermentation and infusion (Lian 2007); CFD is used to predict the interplay
between the heat transfer and enzymatic reactions during tea fermentation;
optimal process conditions for desired flavor generation and functional proper-
ties can also be obtained.
All of these applications are good examples of CFD’s ability to resolve problems in
both novel and conventional food-processing operations, even in systems for which
there is little prior knowledge. Also, since CFD can compute variables such as fluid
velocity, pressure, and temperature at many thousands of locations within a virtual
geometry, the solutions can be comprehensive, both spatially and temporally. The
various PDEs that govern the phenomena encountered in food-processing systems are
introduced next. The governing equations of fluid flow and heat transfer can be
considered as mathematical formulations of the conservation laws of fluid mechanics.
When applied to a fluid continuum, these conservation laws relate the rate of change of
a desired fluid property to external forces. The following sections describe the
governing equations for these various phenomena in Cartesian coordinates.
3.2.1 Conservation of Mass
The mathematical formulation of the law of conservation of mass is also known as

the continuity equation. For a fluid element the solution for each requires a separate
equation. The conservation of mass states that the net mass flowing through all the
surfaces of a fluid element must be equal to the rate of accumulation in the element.
For incompressible fluids, which are commonly encountered in the food industry,
this means that the quantity of mass entering a fluid element must balance exactly
with that leaving
r:~
v¼0 (3.1)
where ~
v consists of components of vi .
3.2.2 Conservation of Momentum
The conservation of momentum for a fluid element is also termed Newton’s second
law of motion, which states that the sum of the external forces acting on the fluid
particle is equal to its rate of change of linear momentum:
@vi
r þ r~ vi ¼ rp þ mr2~
v:r~ vi þ rg (3.2)
@t
This mathematical formulation of fluid motion has been used for almost two
centuries, since the emergence of three very important scientists in the field of fluid
mechanics, namely, Leonhard Euler, Claude-Louis Navier, and George Gabriel
Stokes. Leonhard Euler (1707–1783), a Swiss mathematician and physicist, for-
mulated the Euler equations, which describe the motion of an inviscid fluid based
on the conservation laws of fluid mechanics. The French engineer and physicist,
Claude-Louis Navier (1785–1836), and Irish mathematician and physicist, George
Gabriel Stokes (1819–1903), later introduced viscous transport into the Euler
equations by relating the stress tensor to fluid motion. The resulting set of equations
is now termed the Navier–Stokes equation for Newtonian fluids; it has formed the
basis of modern-day CFD.
3.2.3 Conservation of Energy
The mathematical formulation of the conservation of energy is also called the first
law of thermodynamics, which states that the rate of change in energy of a fluid
particle is equal to the heat addition and the work done on the particle, assuming
that the thermophysical properties are not temperature-dependent:
@~
vi
rcp þ~
v:rT ¼ lr2 T þ sT (3.3)
@t
As many food-processing operations involve heat transfer, and because the

properties of fluids can be temperature-dependent, (3.3) is often coupled with the
Navier–Stokes equation. When conjugate heat transfer is under investigation it is
important to maintain continuity of thermal exchange across the fluid–solid inter-
face, and the transport of heat in a solid structure should also be considered in CFD
simulations (Verboven et al. 2004). The Fourier equation, which governs heat
transfer in an isotropic solid, can be written as
@~
vi
rcp ¼ lr2 T þ sT (3.4)
@t
For a conjugate heat transfer situation, where evaporation at the food surface is
considered and where the heat transfer coefficient is known, the boundary condition
for (3.4) may be written as follows:
@T
h Tbf Ts þ es Tbf4 Ts4 ¼ l (3.5)
@n
where e is the emission factor coefficient and s is the Stefan–Boltzmann constant.

Heat transported by radiation and convection from air to food raises the sample
temperature (Aversa et al. 2007). The solution of (3.3) can be used for food surfaces
to calculate the local heat transfer coefficients as given below:

l@T
@n surface
h¼ (3.6)
Tbf Ts
Equation (3.6) can be used provided the surface temperature is assumed inde-
pendent of the coefficient during calculations.
3.3 Modeling Properties of Fluids
3.3.1 Density
In many food-processing operations, variables such as temperature, concentration,

and fluid velocity are functions of density variations caused by the heating and
cooling of fluids. Therefore, the density of the fluid must be accurately represented
in CFD computations. Since many fluid flows encountered in food-processing
applications can be regarded as incompressible, there are two means of accounting
for density variations, namely, the Boussinesq approximation and the ideal gas
equation (Ferziger and Peric 2002). The Boussinesq approximation has been used
successfully in many CFD applications (Abdul Ghani et al. 2001):
r ¼ rref ½1 bðT Tref Þ (3.7)
The approximation assumes that the density differentials of the flow are only
required in the buoyancy term of the momentum equations. In addition, a linear
relationship between temperature and density, with all other extensive fluid proper-
ties being constant, is also assumed. This relationship only considers a single-
component fluid medium; however, by using Taylor’s expansion theorem, the
density variation for a multicomponent fluid medium can also be derived.
For large temperature differences, the fluid flow becomes compressible with a
strong coupling between the continuity, the momentum, and the energy equations
through the equation of state; its properties (viscosity, heat conductivity) also vary
with the temperature, making the Boussinesq flow approximation inappropriate and
inaccurate (Ferziger and Peric 2002). Therefore, in such cases another method of
achieving the coupling of the temperature and velocity fields is necessary. This can
be done by expressing the density difference by means of the ideal gas equation:
pref M
r¼ (3.8)
RT
This method can model density variations in weakly compressible flows, mean-
ing that the density of the fluid is dependent on temperature and composition but
small pressure fluctuations have no influence.
3.3.2 Viscosity
Viscosity is an important fluid property that must be accurately represented in a

CFD model, as it causes the resistance experienced by a fluid when flowing through
any geometry. Viscosity is accurately quantified via the relationship between the
shear stress in a fluid to the rate of deformation of the fluid.
Newtonian fluids are those that have a linear relationship between stress and rate
of deformation, with the proportionality constant, which is actually viscosity. Any
fluid that does not obey the Newtonian relationship between the shear stress and
shear rate is called a non-Newtonian fluid. The shear stress in a Newtonian fluid
is represented by the second term on the right-hand side of (3.1). Many food-
processing media have non-Newtonian characteristics and the shear-thinning or
shear-thickening behavior of these fluids greatly affects their thermal–hydraulic
performance (Fernandes et al. 2006). In recent years, CFD has provided better
understanding of the mixing, heating, cooling, and transport processes of non-
Newtonian substances. Of the several constitutive formulas that describe the rheo-
logical behavior of substances, which include the Newtonian, power law, Bingham,
and Herschel Bulkley models, the power law is the most commonly used model in
food engineering applications (Welti-Chanes et al. 2005). However, there are some
circumstances where modeling the viscosity can be avoided, as low velocities
permit the non-Newtonian fluid to be considered Newtonian, for example, as
shown by Abdul Ghani et al. (2001).
3.4 Modeling Particular Flow Regimes
3.4.1 Turbulent Flows
Turbulent flows are often encountered in the food industry owing to high flow rates
and heat transfer interactions. Currently, even though the Navier–Stokes equations
can be solved directly for laminar flows, it is not possible to solve the exact fluid
motion in the Kolmogorov microscales associated with engineering flow regimes;
thus turbulence requires modeling. For this reason, turbulence models are being
developed yearly, with prediction accuracy undergoing great improvement over the
last few years for a great many flow regimes. However, none of the existing
turbulence models are complete; that is, their prediction performance is highly
reliant on turbulent flow conditions and geometry. Without a complete turbulence
model capable of predicting the average field of all turbulent flows, the present
understanding of turbulence phenomena will reduce the generality of solutions. In
the following sections, some of the best performing turbulence models are discussed.
3.4.1.1 Large Eddy Simulations
Large eddy simulation (LES) forms a solution given the fact that large turbulent
eddies are highly anisotropic and dependent on both the mean velocity gradients and
geometry of the flow domain. This is done mathematically by separating the velocity
field into a resolved and sub-grid part. The resolved part of the field represents the
large eddies, while the sub-grid part of the velocity represents the small scales whose
effect on the resolved field is included through the sub-grid scale model. With the
advent of more powerful computers, LES offers an accurate means of computing
turbulent flow. However, the lengthy time involved in arriving at a solution means
that it is an expensive technique, and consequently, applications of LES for CFD
calculations of food processing are still uncommon (Turnbull and Thompson 2005).
3.4.1.2 Reynolds Averaged Navier–Stokes
When variables are influenced by turbulence, engineers are generally content with a
statistical probability that processing variables within the flow regime (e.g., veloc-
ity, temperature, and concentration) will lie within a certain range of values.
Therefore, the predictions afforded by the Reynolds-averaged Navier–Stokes equa-
tions (RANS), which determine the effect of turbulence on the mean flow field
through time averaging, are in many cases sufficient. By averaging in this way, the
stochastic properties of turbulent flow are essentially ignored along with six addi-
tional stresses (Reynolds stresses) emerging, which need to be modeled by a
physically well-posed equation system to obtain an accurate closure of the equation
system. The following paragraphs describe the common techniques used.
Reynolds Stress Models
The Reynolds stress closure model (RSM) generally consists of transport equations
for the Reynolds stresses – three transport equations for the turbulent fluxes of each
scalar property and one transport equation for the dissipation rate of turbulence
energy. RSMs have exhibited far superior predictions for flows in confined
spaces where adverse pressure gradients occur. Terms accounting for anisotropic
turbulence, which are included in the transport equations for the Reynolds
stresses, means that these models provide a rigorous approach to solving complex
engineering flows. However, storage and execution time can be expensive for three-
dimensional flows. Moreover, convergence of the RSMs has been reported to be
quite poor in the literature for many flow configurations.
Turbulent Viscosity Models
The turbulent viscosity hypothesis (Boussinesq relationship) states that an increase

in turbulence can be represented by an increase in the effective fluid viscosity. The
Reynolds stresses are proportional to the mean velocity gradients with the effective
viscosity representing the proportionality constant (Ferziger and Peric 2002). For a
k–e type turbulence model turbulence viscosity can be represented as follows:
k2
mt ¼ rCm (3.9)
e
For the k–o type turbulence without the low Reynolds number modifications, the
turbulence viscosity can be represented by (Wilcox 1993):
k
mt ¼ r (3.10)
o
This hypothesis forms the foundation for many of today’s most widely used
turbulence models, ranging from simple models based on empirical relationships to
variants of the two-equation k–e model, which describes turbulence viscosity
through turbulence production and destruction (Versteeg and Malalsekeera 1995).
All turbulence viscosity models have relative merit with respect to simulating food-
processing operations.
The Standard k–e Model
The standard k–e model (Launder and Spalding 1974), which is based on the
transport equations for the turbulent kinetic energy k and its dissipation rate e, is
semi-empirical and assumes isotropic turbulence. Although it has been successful
in numerous applications and is still considered an industrial standard, the standard
k–e model is limited in some respects. A major weakness of this model is that it
assumes an equilibrium condition for turbulence, that is, the turbulent energy
generated by the large eddies is distributed equally throughout the energy spectrum.
However, energy transfer in turbulent regimes is not automatic and a considerable
length of time may exist between the production and the dissipation of turbulence.
The RNG k–e Model
The renormalization group (RNG) k–e model (Choundhury 1993) is similar in form
to the standard k–e model, but owing to the RNG methods from which it has been
analytically derived, it includes additional terms for dissipation rate development
and different constants from those in the standard k–e model. As a result, the
solution accuracy of highly strained flows has been significantly improved. The
calculation of the turbulent viscosity also takes into account the low Reynolds
number if such a condition is encountered in a simulation. The effect of swirl on
turbulence is included in the k–e RNG model, thereby enhancing accuracy for
recirculating flows.
The Realizable k–e Model
In the realizable k–e model (Shih et al. 1995) Cm is expressed as a function of mean
flow and turbulence properties, instead of being assumed constant, as in the case
with the standard k–e model. As a result, it satisfies certain mathematical con-
straints on the Reynolds stress tensor that are consistent with the physics of
turbulent flows (e.g., normal Reynolds stress terms must always be positive).
Also, a new model for the dissipation rate is used.
The k–o Model
The k–o model is based on modeled transport equations, which are solved for the
turbulent kinetic energy k and the specific dissipation rate o, that is, the dissipation
rate per unit turbulent kinetic. An advantage that the k–o model has over the k–e
model is that its performance is improved for boundary layers under adverse
pressure gradients, as the model can be applied to the wall boundary, without
using empirical log-law wall functions. A modification was then made to the linear
constitutive equation of the k–o model to account for the principal turbulence shear
stress. This model is called the shear-stress transport (SST) k–o model; it provides
enhanced resolution of the boundary layer in viscous flows (Menter 1994).
3.4.2 Flows Containing Different Phases
Multicomponent flows occur where the species are mixed at molecular level, that is,
they share the same pressure, velocity, and temperature fields. Alongside global
mass continuity the transport equation for the species provides the means for
updating the field of mass fractions, defining the mixture composition, via the
following equation:
@eXk
þ r ðe~
vi Xk Þ ¼ r ðDk rXk Þ þ Sk (3.11)
@t
However, as the relative size of the species becomes bigger there is deviation
from the uniform flow field, and hence multiphase modeling must take over the
multicomponent modeling. In multiphase flows more than one immiscible fluid is
present in the flow. In CFD modeling, the term multiphase denotes a fluid system
that contains fluids of different physical properties, and does not necessarily refer to
the state of matter, that is, whether it is solid, liquid, or gas. In multiphase systems,
multiple fluids are mixed at a macroscopic level, where the mixing scale is larger
than the molecular scale. In general, two-phase fluid systems are encountered in the
food industry; however, the same concepts for their modeling apply to systems
having many more phases. Multiphase flows can be broadly placed into the
following categories:
1. Continuous; all the fluids involved in the flow are continuous such as liquid–liquid
or liquid–gas flows where a free and distinct interface exists between fluids, for
example, bubbly fluids or immiscible fluids.
2. Continuous–dispersed; namely, liquid–solid flows that contain a dispersed phase
within a primary continuous phase, for example, particle-laden flows such as
chunky soups, spray dryers, and fluidized beds.
In multiphase flows, each phase competes for the same volume, resulting in
interactions between phases due to their proximity. Owing to the abrupt difference
between the properties of the various phases, mass, momentum, and energy are
exchanged between the phases. Such exchanges need to be accounted for in a CFD
model, and in many cases these interactions are complex and specific to the type of
flow being modeled.
Multiphase modeling has to provide models for tracking phases and/or predict-
ing their distribution in space and time. Three approaches are generally used: the
volume of fluid model (VOF), Eulerian–Eulerian model, and Lagrangian-Eulerian
model, and are briefly discussed in the following sections.
3.4.2.1 Volume of Fluid Model
This simple multiphase model is well suited to simulate flows of immiscible fluids
on numerical meshes that are capable of solving the interface between the mixture’s
phases. In other words, the VOF is a volume fraction tracking technique, which is
highly effective when the shape of the interface between phases is important.
Moreover, by using the VOF technique no additional modeling of interphase
interactions is required, as all phases are assumed to share the same velocity,
pressure, and temperature fields. The VOF model can calculate large-scale break-
ups and agglomerations such as sloshing effects caused by a moving water tank. If
the breakup is too small to calculate, that is, during the formation of air bubbles in
water or water droplets in air, then very high mesh refinement is required for
accurate predictions.
3.4.2.2 Eulerian–Eulerian Model
Models simulating the distribution of phases via the Eulerian–Eulerian concept

look at the presence and transport of phases in space, rather than tracking individual
particles of phases. In the Eulerian approach, the phases are treated as interacting
and interpenetrating continua (Nijdam et al. 2006). The phases share the same
volume, penetrate each other in space, and exchange mass, momentum, and energy
with each other. Each phase is described by its distinctive physical properties and
has its own velocity, pressure, concentration, and temperature field. Coupling of the
phases is achieved through pressure and interphase exchange coefficients.
The Eulerian–Eulerian model is applicable for continuous–dispersed as well as
continuous–continuous systems. For continuous–dispersed systems, the velocity of
each phase is computed using the Navier–Stokes equations. The dispersed phase
may be in the form of particles, drops, or bubbles. The forces acting on the
dispersed phase are modeled using empirical correlations and included as part of
the interphase transfer terms. Drag, lift, gravity, buoyancy, and virtual mass effects
are some of the forces that may be acting on the dispersed phase. These forces are
computed for an individual particle and then scaled by the local volume fraction to
account for multiple particles.
3.4.2.3 Lagrangian–Eulerian Model
This model is applicable to modeling continuous–dispersed systems and is very

often referred to as a discrete particle model or particle transport model. The
primary phase is continuous, whereas the secondary phase is discrete and may be
composed of particles, drops, or bubbles. The continuous phase flow is computed
by solving the Navier–Stokes equations, and using their solutions in the flow
trajectories, as well as the heat and mass transfer to and from the discrete phase,
which is computed by solving the appropriate ordinary differential equations. The
dispersed phase is represented by tracking a small number of representative particle
streams; coupling between the continuous and discrete phases is achieved by
including appropriate interaction terms in the equation set. Lagrangian–Eulerian
modeling works best when the motion of particles is dominated by their interaction
with the continuous phase rather than with each other, that is, when the dispersed
phases are dilute. Moreover, this model is mainly applicable for low volume
fractions of the dispersed phase, as the volume displacement caused by the discrete
phase is not taken into account.
3.4.3 Modeling Flows through Porous Media
Many large-scale processes in the food industry may have the potential to be grid
point demanding in CFD models owing to the complex geometry of the modeled
structures. For example, to predict the detailed transfer processes within a cold
store containing stacked foods one must mesh all associated geometry with a
complex unstructured or body-fitted system, which is a highly arduous, and in
many cases, inaccessible task. In any case, both computational power and CFD
algorithms have not yet reached such levels of maturity wherein these types of
computations can be achieved. Therefore, other methods must be used to exploit
the physical relationships that exist on a macroscopic level and sufficiently
represent the dynamic flow effects representative of the modeled material.
The porous media assumption relates the effects of particle size and shape,
alignment with airflow, and void fraction to the pressure drop over the modeled
products. This method basically applies Darcy’s law to a porous media by
relating the velocity drop through the pores to the pressure drop over the
material. An extension of this law to account for most commonly encountered
nonlinear relationships between pressure drop and velocity is represented by the
Darcy–Forcheimer equation:
@p m
¼ v þ rC2 v2 (3.12)
@x K
Equation (3.12) represents the most common relationship used to describe

pressure drop across packed beds. In the CFD model this equation is added as an
additional sink term to the momentum equation. The general relationships
employed to determine both the permeability and inertial loss coefficient can be
obtained by inference from the Ergun equation. However, considerable information
regarding the detailed flow and transfer processes taking place within the stacked
material is lost in this type of modeling strategy. Therefore, before modeling a
porous media one must ensure that the parameters in the momentum source terms
represent the physical media as closely as possible.
3.5 Numerical Methods Used by CFD Code Developers
CFD code developers have a choice of many different numerical techniques to

discretize the transport equations. The most important of these include finite
difference, finite elements, and finite volume. The finite difference technique is
the oldest one used, and many examples of its application in the food industry exist.
However, due to difficulties in coping with irregular geometry, finite difference is
not commercially implemented. Furthermore, the current trend of commercial CFD
coding is aimed toward developing unstructured meshing technology capable of
handling the complex three-dimensional geometries encountered in industry.
Therefore, the prospects of finite difference being used in industrial CFD applica-
tions seem limited.
Finite element methods have historically been used in structural analysis where
the equilibrium of the solution must be satisfied at the node of each element. Nicolaı̈
et al. (2001) provided a short introduction on using the finite elements method in
conduction heat transfer modeling, but it will not be discussed here. It is sufficient
to say that as a result of the weighting functions used in this method, obtaining a
three-dimensional CFD solution with a large number of cells is impractical at
present. Therefore, finite elements are not generally used by commercial CFD
developers, especially since many of these CFD codes are marketed toward solving
aerodynamic problems. Nonetheless, finite elements methods have been used in the
modeling of electromagnetic heating in microwave ovens (Verboven et al. 2007b);
vacuum microwave drying (Ressing et al. 2007); radio frequency heating of food;
and conduction and mass transport during drying (Aversa et al. 2007). Therefore, it
would seem the finite elements method is amenable to the modeling of novel
thermal processes if the details of fluid flow do not need explicit quantification.
With finite volume techniques the integral transport equations governing the
physical process are expressed in conservation form (divergence of fluxes); the
volume integrals are then converted to surface integrals using Gauss’s diver-
gence theorem. This is a direct extension of the control volume analysis that
many engineers use in thermodynamics and heat transfer applications, etc., so it
can be easily interpreted. Thus, expressing the equation system through finite
volumes forms a physically intuitive method of achieving a systematic account
of the changes in mass, momentum, and energy, as fluid crosses the boundaries
of discrete spatial volumes within the computational domain. Also, finite
volume techniques yield algebraic equations that promote solver robustness,
adding further reasons to why many commercial developers implement this
technique.
3.6 Applications of CFD in Food Industry
3.6.1 Sterilization
3.6.1.1 Canned Foods
Sterilization is a conventional thermal process that can be modeled with CFD. In

this process rapid and uniform heating is desirable to achieve a predetermined level
of sterility with minimum destruction in the color, texture, and nutrients of food
products (Tattiyakul et al. 2001). Siriwattanayotin et al. (2006) used CFD to
investigate sterilization value (F0) calculation methods, and concluded that when
F0 was determined using the “Thermal Death Time” (TDT) approach, the process
time to achieve the desired temperature in a sugar solution was underestimated if
the surrounding temperature was lower than the reference value.
Canned viscous liquid foods such as soup, carboxyl-methyl cellulose (CMC), or
corn starch undergoing sterilization have been simulated with CFD. In most cases
where natural convection occurs, fluid velocities and shear rates are rather low and
thus non-Newtonian fluids can be assumed as Newtonian (Abdul Ghani et al. 2003).
Neglecting heat generation due to viscous dissipation is another assumption that is
generally made. In such simulations, CFD has shown the transient nature of the
slowest heated zone (SHZ), and has illustrated the large amount of time needed for
heat to be transferred throughout food, as well as the sharp heterogeneity in the

temperature profile when the process is static.
CFD has also been used to study the effect of container shape on the efficiency of
the sterilization process (Varma and Kannan 2005, 2006). Conical vessels pointing
upward were found to reach appropriate sterilization temperature the quickest
(Varma and Kannan 2005). Full cylindrical geometries performed best when
sterilized in a horizontal position (Varma and Kannan 2006).
3.6.1.2 Pouched Foods
In recent years, CFD has provided a rigorous analysis of the sterilization of three-
dimensional pouches containing liquid foods (Abdul Ghani et al. 2002; Abdul
Ghani and Farid 2006). Coupling first-order bacteria and vitamin inactivation
models with the fluid flow has allowed prediction of the transient temperature,
velocity, and concentration profiles of both the bacteria and ascorbic acid during
natural convection. The concentrations of bacteria and ascorbic acid after heat
treatment of pouches filled with the liquid food were measured, and close agree-
ment was found with the numerical predictions. The SHZ was found to migrate
during sterilization until eventually resting in a position at a distance about 30%
from the top of the pouch. As expected, the bacterial and ascorbic acid destruction
was seen to depend on both the temperature distribution and flow pattern.
3.6.2 Pasteurization
CFD has been used to predict the transient temperature and velocity profiles during
pasteurization of intact eggs (Denys et al. 2007). Owing to its ability to account for
complex geometries, heterogeneous initial temperature distributions, transient
boundary conditions, and nonlinear thermophysical properties, CFD has permitted
a comprehensive understanding of this thermal process. Such analysis has allowed
the gap in knowledge of this area to be filled, because up to recently little informa-
tion was available on the correct processing temperatures and times for safe
pasteurization, without loss of functional properties. In the series of studies pub-
lished on this topic by Denys et al. (2007), a procedure to determine the surface heat
transfer coefficient using CFD simulations of eggs filled with a conductive material
of known thermal properties was first developed, after which conductive and
convective heating processes in the egg were modeled (Denys et al. 2007). This
revealed that, similar to the phenomena noted by Abdul Ghani et al. (1999) for
canned foods, the cold spot moved during the process toward the bottom of the egg.
The location of the cold zone in the yolk was predicted to lie below its geometrical
center, even when the yolk was positioned at the top of the egg. It was concluded
that no convective heating takes place in the egg yolk during processing.
3.6.3 Aseptic Processing
3.6.3.1 Plate Heat Exchangers for Milk Processing
In the dairy industry it is essential to heat-treat milk products in a continuous

process, as on the one hand, it is necessary to promote microbial safety and increase
the shelf life of milk. However, on the other hand, efficient plant processes are also
desirable, but the adverse influences of heat on the sensory and nutritional proper-
ties of the final milk product act as a hindrance to efficient thermal processing.
Many CFD studies of PHEs exist, and have presented different techniques for
geometry optimization, for example, corrugation shape or the optimization of other
process parameters such as inlet and outlet positions and PHE–product temperature
differences (Grijspeerdt et al. 2007; Jun and Puri 2005). Grijspeerdt et al. (2007)
investigated the effect of large temperature differences between the product and
PHE, and noted that the larger the difference, the greater is the opportunity
for fouling. While fouling has been studied by the same group of authors, Jun and
Puri (2005) were the first to couple a fouling model with a three-dimensional
thermal–hydraulic model with CFD simulations. In this study, Jun and Puri inves-
tigated the influence of various PHE designs on fouling rates, comparing those
typically used in the dairy industry with those used in the automobile industry.
3.6.3.2 Plate Heat Exchangers for Yoghurt Processing
Fernandes et al. (2005, 2006) studied the cooling of stirred yoghurt in PHEs with
CFD simulations in order to investigate the thermal–hydraulic phenomena. They
modeled the rheological behavior of yoghurt via a Herschel–Bulkley model. In
addition to accounting for this rheological behavior, they also provided a high level
of precision in the PHE geometrical design and the imposed boundary conditions.
During the course of these studies, it was found that due to the higher Prandtl
numbers and shear thinning effects of the yoghurt, the Nusselt numbers of the fully
developed flows were more than ten times higher than those of water. This result
presented a substantial thermal–hydraulic performance enhancement in comparison
with that from Newtonian fluids. Furthermore, it was shown that PHEs with high
corrugation angles may provide better opportunities for the gel structure breakdown
desired during the production stage of stirred yoghurt.
3.6.4 Drying
3.6.4.1 Fluidized Beds
Because of the complex interactions that occur during fluidized bed drying, empir-
ical correlations are only valid for a certain range of conditions, and CFD simula-
tions have been the only means of providing accurate information on the flow
phenomena. However, similar to other drying applications, difficulties exist in mod-

eling the interactions between the solid and liquid phases, as well as limitations in
computing power, and consequently only a limited number of CFD simulations exist.
The Eulerian–Eulerian approach offers the most efficient way of representing
the two phases in this type of system, considering the present level of computer
power available and the large number of granules (grain) in a typical system;
consequentially its use has been preferred over discrete methods by some
researchers. Szafran and Kmiec (2004), via the multi-fluid granular kinetic
model of Gidaspow et al. (1992) and the k–e model, used this approach in their
CFD simulations of the transport mechanisms in the spouted bed.
A major difficulty encountered in modeling spouted bed dryers has been the
excessive computing times required to simulate only a fraction of the drying
process. For example, owing to the small time steps required to resolve the instabi-
lities in the flow regime, it would take a two-dimensional CFD simulation almost
1 year to simulate 1 h of drying (Szafran and Kmiec 2004). Thus, at present CFD can
only be used as a tool that permits a deeper understanding of the flow patterns and
their effects on the drying kinetics rather than for design and optimization.
3.6.4.2 Spray Drying
Spray drying is another traditional drying technique and is used to derive powders
from products, with its main objective being to create a product that is easy to store,
handle, and transport. CFD has been a necessary requisite for accurate spray
dryer modeling, and has been employed for over 10 years now (Langrish and
Fletcher 2003).
One of the big difficulties when using CFD software packages in spray dryer
modeling is that owing to the presence of both solid and fluid, the mass transport
limitations within a droplet cannot be easily taken into account, and therefore sub-
models must be included to do so; for accurate solutions these must be used
alongside many other sub-models that account for other phenomenological aspects
(Straatsma et al. 2007). In a recent study by Straatsma et al. (2007), sub-models for
mass transport, inter-particle collision, agglomeration, thermal reactions, and stick-
iness were implemented with an Lagrangian-Eulerian model of an industrial dryer.
The CFD simulations allowed the authors (Straatsma et al. 2007) to assess the
agglomeration size of the particles and the stickiness of the particles colliding with
each other and the wall, and as a consequence allowed the fouling liability of the
dryer to be evaluated.
3.6.4.3 Forced Convection Drying
Because of the complex geometry usually encountered in drying applications,

theoretical studies are often not applicable; to obtain the spatial distribution
of transfer coefficients with reasonable accuracy it is necessary to solve the
Navier–Stokes equations in the product surroundings. From the distribution of

transfer coefficients the correct temperature and moisture profiles within the pro-
duct can be predicted, so that the drying process can be optimized. Kaya et al. (2007)
used this approach with CFD to determine the transfer coefficients; then the heat
and mass transfer within the food was simulated with an external program. How-
ever, it is also possible to do this within the CFD package by determining the heat
transfer coefficient, and the mass transfer with the Lewis relation, once a transient
solution is performed.
3.6.5 Cooking
3.6.5.1 Natural Convection Ovens
Electric ovens are commonly used household appliances that rely on conjugate
thermal exchange to produce the desired cooking effect in a foodstuff. For that
reason, CFD is an appropriate tool to quantify the internal thermal field. A thorough
investigation into the thermal profile of an electrical oven, operated under both broil
and bake modes, was completed by Mistry et al. (2006). The solution first obtained
from the steady-state analysis yielded a flow field, which opposed that evident from
experimental observation. This was addressed by imposing an artificial, that is, a
“numerical,” vent suction pressure, the value of which was tweaked until thermal
field predictions corresponded with experimental measurements. Full cycling
times, employing intermittent ON/OFF operation of heaters, were also simulated
for both the broil and bake cycles. From the comparison of predictions, the broil
cycle was confirmed to be less efficient, with a notable heterogeneity in temperature
profile, owing to temperature stratification; this underscored the fact that the main
thermal exchange in this cycle was due to radiation.
3.6.5.2 Forced Convection Ovens
Application of CFD in jet impingement oven systems provides detailed understand-

ing of the effect of different oven geometries as well as object geometries on the
system performance. A full three-dimensional CFD model of a multiple jet impin-
gement oven was developed by Kocer et al. (2007). Convection and conduction
heat transfers were coupled. However, as the thermal exchange was a result of jet
impingement only, radiation was ignored with a small compromise in accuracy.
Moisture transfer was not considered. Kocer et al. (2007) then determined a
correlation for the average Nusselt number in terms of Reynolds number for
multiple jets impinging on the surface of a cylindrical model cookie, which
indicated the strong dependence of surface heat transfer coefficient on velocity of
the jet.
3.6.5.3 Baking Ovens
Recent CFD modeling studies of the bread-baking process have looked at the two-
dimensional physical representation, coupling convection, and radiative heat trans-
fer via the discrete ordinate (DO) model (Wong et al. 2007a). Moreover, the density,
heat capacity, and thermal conductivity were allowed to vary with temperature.
However, some discrepancies between predictions and measurements of the actual
baking process were found, especially for those comparisons made at the dough
center, which were probably caused by no modeling of the moisture transport in the
dough and evaporation kinetics. Moreover, the confining effect afforded by the two-
dimensional model was seen to cause lack of correspondence in the validation study.
3.7 Challenges in Use of CFD in the Food Industry
3.7.1 Improving the Efficiency of the Solution Process
Insight into the numerical abilities of CFD packages is important if one needs to
solve the problems of excessive computing times. Taking the parallelization fea-
tures of commercial CFD codes as an example, these can allow a solution to be
formed quicker, via domain decomposition, as long as the computing power is
available and Lagrangian particle tracking is not employed. Alongside this, the
solving techniques employed in commercial CFD codes have also been found to
play a major role in efficiency. Fletcher et al. (2006) noted how segregated solvers
and coupled solvers can bring different attributes to solution progression, and found
that owing to the reduced levels of “random noise” introduced, the coupled solver
permitted a high level of control over the solution process, allowing efficient and
accurate predictions of the transient evolution of the flow instability in a spray
dryer, when compared to the segregated solver.
Simplifying the geometrical representation of CFD models can also cut down on
both pre-processing and solving time. The two-dimensional modeling technique
assumes that the length of a system is much greater than its other two dimensions,
and that the process flow is normal up to this length. As the effects of the confining
geometry are essentially disregarded, accurate judgment of whether the process is
amenable to the two-dimensional assumption is required.
3.7.2 CFD to Control Food Processes
All processes in the food industry are performed under controlled conditions.
Unfortunately, due to the nonlinearity of the transport phenomena, CFD techniques
are not yet amenable to the online control of thermal processes; reduced order
models, which use statistical data to manipulate the process variables via controlled
inputs, are more appropriate. However, this does not mean that the actions of a
control system cannot be modeled by CFD. Wong et al. (2007b) were the first to
implement a control system within a CFD model to simulate its performance. Such
abilities undoubtedly provide benefits during the pre-design or optimization stages
of system development.
3.7.3 Turbulence
One of the main issues facing the food industry over the last two decades is the fact
that most turbulence models have been shown to be application-specific. At the
present time, there are many turbulence models available; however, until a com-
plete turbulence model capable of predicting the average field of all turbulent flows
is developed the CFD optimization of many thermal processes will be hampered.
The reason is that in every application many different turbulence models must be
applied until the one that gives the best predictions is found. The closest to the
complete turbulence model thus far is the LES, which uses the instantaneous
Navier–Stokes equations to model large-scale eddies, with smaller scales solved
with a sub-grid model. However, using the LES model demands large amounts of
computer resources, which may not be presently achievable.
3.7.4 Need for Sensitivity Analysis
In CFD simulations, the boundary conditions must be adequately matched to the

physical parameters of the process, with the precision of similarity being
conditioned by the mechanism under study and the level of accuracy required.
Even when this is done, the CFD solution still may not be a correct physical
representation of the physical system. This was shown by Mistry et al. (2006),
who found that an artificial pressure differential was required to predict the correct
flow patterns in an oven heated by natural convection. Such results suggest the
importance of sensitivity analysis studies alongside experimental measurements in
the early stages of model development. Sensitivity analyses are also necessary for
turbulence model specification, or turbulence model tuning via inlet conditions, and
for CFD model simplification.
3.8 Conclusions
CFD has played an active part in the design of food operation processes for over a
decade now. In recent years simulations have reached higher levels of sophistica-
tion, as application-specific models can be incorporated into the software with ease
via user-defined files. The importance of maintaining a high level of accuracy via
circumspect choices made during model development is evident from the reviewed
studies, as many of these studies provide detailed validation exercises. Undoubt-
edly, with current computing power progressing unrelentingly, it is conceivable that
CFD will continue to provide explanations for transport phenomena, leading to
better design of processes in the food industry.
3.9 Nomenclature
Cm Empirical turbulence model constant

cp Specific heat capacity (W kg1 K1)
D Diffusion coefficient
E Mole fraction
g Acceleration due to gravity (m s2)
K von Karman constant
M Molecular weight (kg kmol1)
Nu Nusselt number
p Pressure (Pa)
R Gas constant (J kmol1 K1)
Re Reynolds number
K Turbulent kinetic energy (m2 s2)
sT Thermal sink or source (W m3)
sc Concentration sink or source (mol m3)
T Temperature (K)
Tu Turbulence intensity (%)
t Time (s)
U Velocity component (m s1)
~
vi Velocity component (m s1)
x Cartesian coordinates (m)
Xk Volume fraction of k
3.9.1 Greek Letters
r Density (kg m3)

m Dynamic viscosity (kg m1 s1)
b Thermal expansion coefficient (K1)
l Thermal conductivity (W m1 K1)
e Turbulent dissipation rate (m2 s3)
’ Transported quantity
G Diffusion coefficient of transported variable
# Diffusivity of the mass component in the fluid (m2 s1)

o Specific dissipation (s1)
mt Turbulent viscosity (kg m1 s1)
3.9.2 Subscripts
i Cartesian coordinate index

bf Bulk fluid
s Surface
V Mesh element volume
A Area of mesh element
m Mean
t Turbulent
k kth phase
References
Abdul Ghani AG, Farid MM (2006) Using the computational fluid dynamics to analyze the thermal
sterilization of solid–liquid food mixture in cans. Innovative Food Science & Emerging Tech-
nologies 7(1–2):55–61
Abdul Ghani AG, Farid MM, Chen XD (2002) Theoretical and experimental investigation of the
thermal inactivation of Bacillus stearothermophilus in food pouches. J Food Eng 51(3):221–228
Abdul Ghani AG, Farid MM, Chen XD, Richards P (1999) An investigation of deactivation of
bacteria in a canned liquid food during sterilization using computational fluid dynamics (CFD).
J Food Eng 42:207–214
Abdul Ghani AG, Farid MM, Chen XD, Richards P (2001) Thermal sterilization of canned food in
a 3-D pouch using computational fluid dynamics. J Food Eng 48:147–156
Abdul Ghani AG, Farid MM, Zarrouk SJ (2003) The effect of can rotation on sterilization of liquid
food using computational fluid dynamics. J Food Eng 57:9–16
Aversa M, Curcio S, Calabro V, Iorio G (2007) An analysis of the transport phenomena occurring
during food drying process. J Food Eng 78(3):922–932
Choundhury D (1993) Introduction to the renormalization group method and turbulence model-
ling. Fluent Inc. Technical Memorandum TM-107
Denys S, Pieters J, Dewettinck K (2007) CFD analysis of thermal processing of eggs, chapter 14.
In: Da-Wen Sun (ed) Computational fluid dynamics in food processing. CRC Press, Boca
Raton, FL, pp 347–381
Fernandes CS, Dias R, Nóbrega JM, Afonso IM, Melo LF, Maia JM (2005) Simulation of stirred
yoghurt processing in plate heat exchangers. J Food Eng 69:281–290
Fernandes CS, Dias RP, Nobrega JM, Afonso IM, Melo LF, Maia JM (2006) Thermal behaviour of
stirred yoghurt during cooling in plate heat exchangers. J Food Eng 69:281–290
Ferziger JH, Peric M (2002) Computational methods for fluid dynamics. Springer-Verlag, Berlin/
Heidleberg, pp 1–100
Fletcher DF, Guo B, Harvie DJE, Langrish TAG, Nijdam JJ, Williams J (2006) What is important
in the simulation of spray dryer performance and how do current CFD models perform? Appl
Math Model 30(11):1281–1292
Foster AM, Madge M, Evans JA (2005) The use of CFD to improve the performance of a chilled
multideck retail display cabinet. Int J Refrig 28:698–705
Gidaspow D, Bezburuah R, Ding J (1992) Hydrodynamics of circulating fluidized beds, kinetic
theory approach. Fluidization VII, Proceedings of the 7th Engineering foundation conference
on fluidization, Brisbane, Australia, pp 75–82
Grijspeerdt K, Vucinic D, Lacor C (2007) CFD modelling of the hydrodynamics of plate heat
exchangers for milk processing, chapter 19. In: Da-Wen Sun (ed) Computational fluid dyna-
mics in food processing. CRC Press, Boca Raton, FL, pp 403–417
Jun S, Puri VM (2005) 3D milk-fouling model of plate heat exchangers using computational fluid
dynamics. Int J Dairy Technol 58(4):214–224
Jun S, Puri VM (2006) A 2D dynamic model for fouling performance of plate heat exchangers.
J Food Eng 75(3):364–374
Kaya A, Aydin O, Dincer I (2007) Numerical modeling of forced-convection drying of cylindrical
moist objects. Numer Heat Transf A-Appl 51(9):843–854
Kocer D, Nitin N, Karwe MV (2007) Applications of CFD in Jet impingement oven, chapter 19.
In: Da-Wen Sun (ed) Computational fluid dynamics in food processing. CRC Press, Boca
Raton, FL, pp 469–487
Langrish TAG, Fletcher DF (2003) Prospects for the modelling and design of spray dryers in the
21st century. Drying Technol 21(2):197–215
Launder BE, Spalding DB (1974) The numerical computation of turbulent flows. Comput Meth
Appl Mech Eng 3:269–289
Lian G (2007) CFD simulation of multi-physical multi-(bio) chemical interactions of tea fermen-
tation and infusion, chapter 24. In: Da-Wen Sun (ed) Computational fluid dynamics in food
processing. CRC Press, Boca Raton, FL, pp 478–505
Menter FR (1994) Two-equation eddy-viscosity turbulence models for engineering applications.
AIAA J 32(8):1598–1604
Mistry H, Ganapathi-subbu, Dey S, Bishnoi P, Castillo JL (2006) Modeling of transient natural
convection heat transfer in electric ovens. Appl Therm Eng 26(17–18):2448–2456
Nicolaı̈ BM, Verboven P, Scheerlinck N (2001) Modelling and simulation of thermal processes,
Chapter 6. In: Philip Richardson (ed) Thermal technologies in food processing. Woodhead
Publishing, Cambridge/England, pp 91–109
Nijdam JJ, Guo BY, Fletcher DF, Langrish TAG (2006) Lagrangian and Eulerian models for
simulating turbulent dispersion and coalescence of droplets within a spray. Appl Math Model
30(11):1196–1211
Norton T, Sun D-W (2007) An overview of CFD applications in the food industry, chapter 1. In:
Da-Wen Sun (ed) Computational fluid dynamics in food processing. CRC Press, Boca Raton,
FL, pp 1–43
Olsson E, Tr€agårdh C (2007) CFD modelling of jet impingement during heating and cooling of
foods, chapter 20. In: Da-Wen Sun (ed) Computational fluid dynamics in food processing. CRC
Press, Boca Raton, FL, pp 478–505
Ressing H, Ressing M, Durance T (2007) Modelling the mechanisms of dough puffing during
vacuum microwave drying using the finite element method. J Food Eng 82(4):498–508
Shih TH, Liou WW, Shabbir A, Zhu J (1995) A new k–e eddy viscosity model for high Reynolds
number turbulent flows—model development and validation. Comput Fluids 24(3):227–238
Siriwattanayotin S, Yoovidhya T, Meepadung T, Ruenglertpanyakul W (2006) Simulation of
sterilization of canned liquid food using sucrose degradation as an indicator. J Food Eng
73:307–312
Straatsma J, Verdurmen REM, Verscheuren M, Gunsing M, de Jong P (2007) CFD simulation of
spray drying of food products, Chapter 10. In: Da-Wen Sun (ed) Computational fluid dynamics
in food processing. CRC Press, Boca Raton, FL, pp 249–287
Sun D-W (ed) (2007) Computational fluid dynamics in food processing. CRC Press, Boca Raton, FL
Szafran RG, Kmiec A (2004) CFD modeling of heat and mass transfer in a spouted bed dryer. Ind
Eng Chem Res 43(4):1113–1124
Tattiyakul J, Rao MA, Datta AK (2001) Simulation of heat transfer to a canned corn starch
dispersion subjected to axial rotation. Chem Eng Process 40:391–399
Turnbull J, Thompson CP (2005) Transient averaging to combine large eddy simulation with
Reynolds averaged Navier-Stokes simulations. Comput Chem Eng 29:379–392
Varma MN, Kannan A (2005) Enhanced food sterilization through inclination of the container
walls and geometry modifications. Int J Heat Mass Transfer 48:3753–3762
Varma MN, Kannan A (2006) CFD studies on natural convective heating of canned food in conical
and cylindrical containers. J Food Eng 77(4):1027–1036
Verboven P, Atalay YT, Vermeir SE, Nicolaı̈ BM, Lammertyn J (2007a) CFD design and
optimization of biosensors for the food industry, chapter 26. In: Da-Wen Sun (ed) Computa-
tional fluid dynamics in food processing. CRC Press, Boca Raton, FL, pp 478–505
Verboven P, de Baerdemaeker J, Nicolaı̈ BM (2004) Using computational fluid dynamics to
optimise thermal processes, chapter 4. In: Philip Richardson (ed) Improving the thermal
processing of foods. Woodhead Publishing, Cambridge/England, pp 82–102
Verboven P, Datta AK, Nicolaı̈ BM (2007b) Computation of airflow effects in microwave and
combination heating, Chapter 12. In: Da-Wen Sun (ed) Computational fluid dynamics in food
processing. CRC Press, Boca Raton, FL, pp 313–331
Versteeg HK, Malalsekeera W (1995) An introduction to computational fluid dynamics. Longman,
Harlow, UK, pp 1–100
Welti-Chanes J, Vergara-Balderas F, Bermúdez-Aguirre D (2005) Transport phenomena in food
engineering: basic concepts and advances. J Food Eng 67:113–128
Wilcox DC (1993) Comparison of 2-equation turbulence models for boundary-layer flows with
pressure gradient. AIAA J 31(8):1414–1421
Wong SY, Zhou WB, Hua JS (2007a) CFD modeling of an industrial continuous bread-baking
process involving U-movement. J Food Eng 78(3):888–896
Wong SY, Zhou WB, Hua JS (2007b) Designing process controller for a continuous bread baking
process based on CFD modelling. J Food Eng 81(3):523–534
Xia B, Sun D-W (2002) Applications of computational fluid dynamics (CFD) in the food industry:
a review. Comput Electron Agric 34:5–24
Zhou W, Therdthai N (2007) Three-dimensional CFD modelling of a continuous baking process,
chapter 11. In: Da-Wen Sun (ed) Computational fluid dynamics in food processing. CRC Press,
Boca Raton, FL, pp 287–213
Chapter 4
Incorporation of Fibers in Foods: A Food
Engineering Challenge
Madhuvanti Kale, Dhananjay Pai, Bruce Hamaker, and Osvaldo Campanella
4.1 Introduction
Dietary fiber is an essential part of the human diet. It performs several functions in
the body, such as water binding, cholesterol and fat binding, attenuation of blood
glucose levels, preventing constipation, and facilitating good colonic health. How-
ever, dietary fiber often plays a poor functional role in processed foods, impairing
processing operations and greatly compromising the sensory quality of the food. As
a result, processed foods are typically made from ingredients having low fiber
content, such as refined flours. Diets that predominantly consist of such processed
foods are therefore seriously lacking in this valuable nutritional constituent. The
alarming growth in the prevalence of diseases such as obesity, coronary heart
disease, diverticular disease, and colorectal cancer has been associated at least in
part with low fiber intake. This conclusion calls for urgent action by the food
industry to produce fiber-enriched processed foods that will be a good source of
fiber, while at the same time being acceptable to the consumer.
The importance of dietary fiber and potential fiber sources as well as food
vehicles for fortification will be briefly discussed in this chapter. Emphasis will
be given to corn bran as a fiber source, as it is a cheap and abundant by-product of
the corn milling industry. The food processing and product quality problems
associated with the addition of corn bran in convenience foods such as extruded
breakfast cereals and snacks will be discussed, with some strategies for fiber
modification to overcome these problems. In addition, some analytical techniques
M. Kale and B. Hamaker,

Department of Food Science, Purdue University, West Lafayette, IN, USA
and
Whistler Center for Carbohydrate Research, Purdue University, West Lafayette, IN, USA
D. Pai and O. Campanella (*)
Whistler Center for Carbohydrate Research, Purdue University, West Lafayette, IN, USA
and
Agricultural and Biological Engineering Department, Purdue University, West Lafayette, IN, USA
e-mail: campa@purdue.edu.
70 M. Kale et al.
to assess fiber functionality will be described. Finally, a process dealing with alkali
treatment of corn bran to improve fiber functionality during extrusion processing
will be presented as a case for fiber incorporation in extruded foods. The rheological
behavior of fiber-enriched foods and its role in the extrusion process will also be
discussed.
4.1.1 Dietary Fiber
Dietary fiber typically refers to those carbohydrates from plant foods that are not
digested by enzymes in the small intestine, including non-starch polysaccharides
and lignin (USDA 2005), and also resistant starch. Dietary fiber can be soluble or
insoluble. Soluble fiber typically undergoes fermentation by colonic microorganisms,
whereas insoluble dietary fiber passes through the body largely unchanged. Incor-
poration of insoluble fiber at high levels into processed foods often causes problems
due to poor functional properties, such as severe reduction of expansion in extruded
foods (Breen et al. 1977; Mottur and Glass 1985; Hseih et al. 1989, 1991; Lue et al.
1991; Onwulata et al. 2001; Moraru and Kokini 2003). Soluble fiber, on the other
hand, tends to compromise the sensory quality of the food, and may hamper
functionality as well (Colliopoulos 1991; Ringe and Stoll 1991; Foehse et al.
1992; van Lengerich and Larson 2000).
4.1.2 Importance of Dietary Fiber
The Food and Nutrition Board of the Institute of Medicine (National Academy of
Sciences 2002) has set recommendations for fiber intake at 38 and 25 g/day (14 g/
1,000 kcal) for men and women, respectively. These levels are based on the
minimum intake required to reduce the risk of coronary heart disease and colon
cancer, prevent constipation and diverticular disease, provide energy for colonic
bacteria, attenuate blood glucose and lipid levels, and provide foods that may
contribute to satiety. On average, people consume about one half of these recom-
mended levels. Thus, there is an obvious need to increase dietary fiber consumption
among the general public. One way to achieve this goal is through fiber incorpora-
tion in commonly consumed foods, for instance, breakfast cereals, ready-to-eat
(RTE) puffed snacks, yogurt, and baked products.
4.1.3 Sources of Fiber
Most plant foods contain some amount of dietary fiber. However, much of this fiber
is removed during processing and pre-processing operations. A common example
4 Incorporation of Fibers in Foods: A Food Engineering Challenge 71
of pre-processing is the refining of cereal flour to remove bran, which improves the
functionality of the flour, while producing a fiber-rich by-product which can be a
potentially valuable source of fiber for fortification. There are other soluble fiber
sources, such as psyllium and inulin, which in fact show good processing function-
ality in extruded foods because they do not reduce their expansion (Foehse et al.
1992). However, their significantly higher cost and tendency to dissolve rapidly in
water, saliva, or milk and to produce gummy or slimy texture limit the level at
which they can actually be utilized in these products.
Cereal brans represent a valuable class of by-products of the milling industry
that can serve as a source of dietary fiber. Many countries are major producers and
processors of corn and, consequently, corn bran is one of the most abundant and
cheap resources for fiber fortification. Therefore, the focus of this chapter will be on
corn bran as a source of fiber. The discussion will apply, in the most part, to other
fiber sources as well.
4.1.4 Fortification of Foods with Fiber
The level of added fiber and the amount of fiber per serving are technologically and
nutritionally significant quantities to be decided by the food manufacturer. From a
practical standpoint, the level of addition is dictated by the occurrence and extent of
processing difficulties and the effect of the fiber on the sensory qualities of fiber-
enriched products, while the amount of fiber per serving decides the kind of labeling
declaration that can be made, more specifically if it can be labeled as a “good,”
“very good,” or “excellent” source of dietary fiber.
Several foods can be considered as candidates for fortification with fiber, the
notable categories being extruded foods, baked foods, beverages, and dairy pro-
ducts such as yogurts; the latest reports even suggest the development of high-fiber
drinks (Decker 2007). This chapter focuses on fiber processing and fiber function-
ality enhancement for extrusion processing, as incorporation of high levels of fiber
is arguably most challenging in extruded foods.
4.2 Processing and Chemical Evaluation of Fiber-Enriched

Foods and Corn Fibers
4.2.1 Extrusion Processing
Extrusion is a processing tool that can be used to incorporate fibers in highly

consumed foods like breakfast cereals and snacks. Extrusion technology has had
an important role in the food industry development as it is an efficient and cost-
effective manufacturing process. In a way, it can be considered as a process
72 M. Kale et al.
intensification tool, a term used to categorize processes that result in better products
and processes that are safer, cleaner, smaller, and cheaper. Extrusion cooking
technologies are used for cereal and protein processing in the food, pet food, and
feed sectors. Raw material properties are critical in determining the magnitude of
variables used in the process such as pressure, temperature, moisture, feed, as well
as providing structure to the final product (Guy 2001). The most commonly used
materials are corn and wheat flours or meals. Other starch sources such as rice,
potato, rye, and barley flours are also used for extrusion purposes, as well as
globular proteins such as those present in soybean, which provide structure to the
food product when used at high concentrations larger than 40% (Guy 2001; Moraru
and Kokini 2003). While proteins denature at the high temperature existing in the
extruders, fibers were reported to be unaffected by the heat and shear applied during
extrusion processing (Guy 2001).
Moraru and Kokini (2003) have extensively described the mechanism of extru-
date expansion. During expansion, there is a rapid decrease in the moisture and
temperature of the product, which causes a dramatic change in the extrudate
rheology and transport properties (Fan et al. 1994). Several researchers (Lai and
Kokini 1990; Kokini et al. 1992; Fan et al. 1994; Della Valle et al. 1996, Li et al.
2004) have hypothesized that the rheological properties of the polymeric matrix just
before exiting the extruder (known as melt) have an important role in expansion,
since they determine the resistance to the growth of the air bubbles due to the
pressure difference between the inside and the outside of the bubble.
Both raw material properties and operating parameters affect the rheological
properties of the melt, and have a large influence on the product expansion. The film
of biopolymers surrounding the bubbles must flow easily in the bubble walls to
allow the bubbles to expand, as the superheated water is released very quickly at
atmospheric pressure once the product exits the extruder die. The resistance to this
flow is governed by the shear viscosity. Expansion seems to be favored at certain
optimum shear viscosities depending on the material, since it has been observed
that excessively high or low viscosities impede expansion (Fan et al. 1994; Della
Valle et al. 1996; Li et al. 2004; Moraru and Kokini 2003; Kokini et al. 1992). Melt
shear viscosity is affected by biopolymer structure, die dimensions (L/D), screw
speed, temperature, and moisture. Concerning the biomolecular structure, several
researchers have reported differences in extrusion expansion and melt rheology due
to differences in amylose/amylopectin content (Guy 2001; Chinnaswamy and
Hanna 1988; Della Valle et al. 1996; Kokini et al. 1992; Lai and Kokini 1990;
Li et al. 2004).
The mechanism of the expansion process can be described as consisting of
several stages starting from a nucleation stage, which generates a two phase
structure, where tiny gas bubbles are surrounded by polymeric films (bubble
wall). As bubbles grow, these films are stretched to a point where they are no
longer able to withstand the stretching forces (Gendron and Daigneault 2000; Guy
2001). When the film becomes thin, shear effects become less important and the
deformation flow of relevance is that of an extensional flow. Consequently, exten-
sional properties of the melt, which measure the resistance of the film to the
stretching deformation, become the control for the extent of bubble growth. Even
though the relevance of extensional properties and biaxial stretching have been
recognized as one of the main modes of deformation in food extrusion (foam
extrusion), few rheological studies using this type of flow have been performed,
mainly due to the difficulty of finding reliable techniques to measure biaxial
extensional properties of extruded melts.
One aspect that has been investigated in the area of plastic extrusion, but not in
foods, is the effect of the material molecular structure on expansion and extensional
properties of the melt. For plastics it has been recognized that extensional properties
of the melt are very sensitive to macromolecular structural differences (Hingmann
and Marczinke 1994; Gendron and Daigneault 2000). In this sense, structural
aspects of food polysaccharides and their potential effects on melt rheology are
discussed in this chapter.
4.2.2 Challenges in Corn Fiber Processing
It has been reported that the presence of a high proportion of fiber in product
formulation results in the reduction of expansion during extrusion, which in turn
yields products that are dense, tough, and non-crispy (Lue et al. 1991). Mendonca
et al. (2000) reported that co-extrusion of corn bran and cornmeal resulted in less
radial expansion and products with undesirable textural characteristics when the
amount of corn bran was increased. Garber et al. (1997) investigated the influence
of corn bran particle size on extrusion expansion and found that radial expansion
decreased with increasing particle size, producing very hard extrudates of poor
textural properties.
Moraru and Kokini (2003) hypothesized that above a critical concentration,
fibers may disrupt the continuous structure of the melt, thus impeding its elastic
deformation during expansion. Since fibers are not much affected by the harsh
environment existing inside the extruder and largely retain their macrostructure
(Guy 2001), their physical presence in the extrudate bubble walls will tend to
reduce the expansion because they disrupt the bubble wall film when their struc-
tures penetrate it. In addition, it has been hypothesized that, even at small concen-
trations, the structural anisotropy caused by the aligning of long and stiff fiber
molecules in the extruder flow direction increases the extensional viscosity of the
melt film, thus lowering the radial expansion of the product.
4.2.3 Chemistry of Corn Fiber
Corn bran, like any other food material, is a complex system. It consists of thick-
walled cells of pericarp, germ and aleurone layer, and a residual endosperm tissue
of the corn kernel (Saulnier et al. 1995). Corn bran contains about 80–90% dietary
74 M. Kale et al.
fiber (Shelton and Lee 2000). The major polysaccharides present in cereal brans are
cellulose and hemicelluloses. Cellulose is a polymer made of b (1 ! 4)-linked
glucose units. Hemicelluloses are classically defined as polysaccharides that are
extractable from higher plants by aqueous alkaline solutions (Schulze 1891). It is
now known that there are at least four major types of hemicelluloses, namely
xylans, mannans, mixed linkage glucans, and xyloglucans (Ebringerova et al.
2005). In addition to cellulose and heteroxylans, primary cell walls of cereals
consist of certain b-D-glucans, xyloglucans, aromatic substances (such as phenolic
acids), and structural proteins are also present (Saulnier et al. 1995; Carpita 1996).
The major heteroxylans in cereals are arabinoxylans, with some amount of glu-
curonic acid present. Corn bran arabinoxylans, for instance, comprise about 60% of
total bran weight and contain some glucuronic acid and trace amounts of D- and
L-galactose that are extracted from crude corn fiber under alkaline conditions.
Wheat bran contains approximately 45% total dietary fiber that is composed of
44% hemicellulose, 32% cellulose, 5% lignin, and 4% insoluble pectins (Claye
et al. 1996). Saulnier and Thibault (1999) suggested, based on their studies on corn
bran, that the corn cell wall is a three-dimensional network of cellulose bundles
embedded in a network of highly cross-linked arabinoxylans (Fig. 4.1).
Arabinoxylans are composed of a b (1 ! 4)-linked xylan backbone, with
arabinosyl residues attached as short branches. Some of the arabinosyl residues
are esterified with ferulic acid, which is a phenolic acid (Doner and Hicks 1997).
Figure 4.2a shows a schematic structure of a feruloylated arabinoxylan. Cross-
linking of arabinoxylan chains through diferulic acid bridges has been identified
(Saulnier et al. 1995). This is shown in the schematic Fig. 4.1, and in closer detail in
Fig. 4.2b. Each corn bran heteroxylan contains about 75 ferulic acid esters and is
connected by about 30 diferulic cross-links (Saulnier and Thibault 1999).
Fig. 4.1 Schematic of cereal cell wall structure (Adapted from Saulnier and Thibault 1999)
a
H O
H
H
H
O H O
H H O
H O H H
OH H HO H O
OH
H H HO H
H
H
O H O
OH H OH OH
H O H HO H H
H
OH H H O
HO OH
O H H HO H
H
MeO H OH O H
O HO O
H O H
H O H
HO H
OH H O H O
H H H O
H O H H
OH H OH OH H HO OH H O
H H HO H
H
O
OH H OH OH
H O H HO
OHH
O H
MeO H OH
O
b
HO
H O
H
H
H
O
H O
H H O
H O H H
OH H HO H O
OH
H H H O H
H H
O H
OH H OH OH O
H O H HO H
H
H
OH H H O
HO OH
O H H HO H
H
MeO H O H O
O OH
H O H
H O H
HO H
OH H O H O
H H H O
H O
H H
OH H OH OH H HO OH H O
H H O H
H H H
O H
HO H HO OH H OH OH O
H O H H O H
H H
H HO OH H OH
O HO
H H H
O H
H MeO H
HO O OH
O H
OH
H H H
H O H HO HO
O OH H HO
H OH
H H
H H H O H
O HO HO
O H H O
H O OMe
H
H O H H H
O HO H HO O
H OH
H H H
O H O
H
O H
O OH
H
H
H O H
O H HO HO H O OMe
OH
H H
H OH H O
O HO OH H HO
H
H H
H H H O H
O HO
O H
O
H H
H O H H
O H HO
OH
H H
O H
H
Fig. 4.2 (a) Schematic of feruloylated arabinoxylan molecule. (b) Schematic of two arabinoxylan
chains cross-linked by diferulate bridges
The presence of these diferulic bridges and the cross-linking of arabinoxylan

chains severely impair the functionality of corn fiber. Also, the presence of crystal-
line cellulose microfibrils greatly compromises functionality. Thus, strategies for
improving the functionality of bran should be essentially aimed at breaking apart
this network of cross-links to aid the dispersion of the fiber material in the food
system. Some strategies for modification of fibers to be able to incorporate them in
large quantities into processed foods are discussed below.
76 M. Kale et al.
4.2.4 Strategies for Modification of Corn Fiber
Fiber modification strategies generally fall into three categories: chemical, physical,
and enzymatic. The ultimate aim of all of these methods is to alter the molecular and/
or supramolecular structure of the fiber so that its functionality can be improved.
4.2.4.1 Chemical Modification
The chemical modifications of interest here are those intended to increase function-
ality through the selective breakage of ferulic acid cross-links, as well as other
hydrolysis reactions to reduce molecular size. Ferulic acid cross-link removal and
other main chain hydrolysis effects are achieved by the addition of either an acid or
a base. This aids in the dispersal of bran polysaccharides, can decrease polymer
chain length, and may open up junction zones to increase water-binding sites and
solubility. These treatments additionally allow for greater functionality because
they loosen the tightly packed fiber polymeric chains. Treatment with hydrogen
peroxide has also been used to remove lignin (Gelroth and Ranhotra 2001) and to
bleach the undesirable brown color (Doner and Johnston 2001).
4.2.4.2 Physical Modification
Physical treatments involve high shear, high temperature, extrusion, and corn bran
grinding. Particle size reduction has been studied as a possible method for fiber
incorporation into an extruded corn-based product (Blake 2006). However, it was
found that in the range studied, particle size does not have much effect on the
properties of the extrudate, notably expansion.
High shear may decrease polymer chain length and increases the ability of a fiber
to associate with other food components. Shear has been used to create a patented
dietary fiber (Z-TrimTM) in USDA laboratories that is used as fat mimetics. On the
other hand, high temperature processing results in swelling of the cell wall material
and improvement in solubilization of hemicellulose and insoluble pectins (Ng et al.
1999).
The use of extrusion as a processing tool for physical pretreatment offers advan-
tages involving a combination of both shear and high temperature, which results in
dispersion of the dietary fiber produced by the depolymerization and disruption of its
crystalline structure within the cell wall (Ning et al. 1991; Ralet et al. 1993).
4.2.4.3 Enzymatic Modification
A review on enzymatic modification of dietary fibers (Arrigoni 2001) describes two

approaches to dietary fiber modification by enzymatic means. The first is to increase
dietary fiber content in a food by removing digestible compounds; the second is to

use enzymes to alter the fiber fraction itself. The second approach, which is more
pertinent to fiber fortification of foods, is aimed at the modification of the fiber. This
approach requires a mixture of cellulose, hemicellulose, and pectinases. Enzymes
can also be used to modify purified fiber fractions, such as to hydrolyze polysac-
charides in order reduce molecular weight and possibly improve functionality.
4.3 Techniques to Assess Fiber Chemistry and Fiber-Enriched

Foods
4.3.1 Rheological Characterization
The rheological properties of fiber directly affect its behavior in a processing

system. This brief discussion focuses on the assessment of these rheological proper-
ties, using instruments that may either simulate the actual processing conditions or
measure properties that will affect flow behavior during processing.
4.3.1.1 Solution Rheology
Polysaccharides produce solutions of high viscosity at low concentrations. The

viscosity can be measured using a rotational rheometer. These data may provide
qualitative information about the molecular weight and branching of the fiber
molecules.
4.3.1.2 Capillary Rheometry
A capillary rheometer is a pressure-driven rheometer. Gravity, compressed gas, or a

motor-driven piston can be used to generate pressure on the test fluid in a reservoir
(Macosko 1994). The fluid flows through a capillary die of radius R and length L.
Pressure drop and flow rate through this die are used to determine apparent viscosity
of the fluid (Malkin and Isayev 2006; Macosko 1994).
Irrespective of the type of fluid, for a cylindrical capillary, the wall shear stress is
given by
DPR
sw ¼ (4.1)
2L
where DP is the pressure causing the flow, and R and L are the radius and length
of the capillary die, respectively. Foods undergoing extrusion are generally
78 M. Kale et al.
biopolymers that exhibit non-Newtonian behavior. The wall shear rate in a cylin-
drical capillary is calculated by Malkin and Isayev (2006) as

dlog_g0
g_ w ¼ g_ 0 3þ (4.2)
dlogsw
The bracketed term in the above equation is the correction for non-Newtonian
fluid and g_ 0 is the average shear rate in the capillary, and is given by
Q
g_ 0 ¼ (4.3)
pR3
where Q is the flow rate through the capillary. The flow curves obtained from the
test can be used to measure the apparent viscosities at different shear rates and thus
calculate the rheological parameters (i.e., power law) of non-Newtonian fluids.
The advantages in using a capillary rheometer is that it is relatively inexpensive
to build and simple to operate. For viscous polymeric melts, capillary rheometry
appears to be one of the few satisfactory means of obtaining data at shear rates
greater than 10 s1 (Macosko 1994). Capillary rheometers can also reduce solvent
evaporation more efficiently than rotational geometry instruments, which have
larger exposed free surfaces. These instruments are also one of the simplest extru-
sions and die flow simulators (Macosko 1994). The disadvantages are that a large
amount of sample is needed for the test and the instrument can measure only steady
shear properties. The reader is encouraged to refer to Malkin and Isayev (2006) and
Macosko (1994) for details on working principles and different tests that can be
conducted using capillary rheometry.
4.3.1.3 Lubricated Squeezing Flow
Entrance pressure drop measured from the flow into a converging die has been used
to evaluate the extensional viscosity of cornmeal dough (Bhattacharya et al. 1994;
Padmanabhan and Bhattacharya 1993; Seethamraju and Bhattacharya 1994).
Although this may be the best method for approximating extrusion conditions,
the amount of sample needed for each trial is very high.
The lubricated squeezing flow technique is a simple method to measure exten-
sional viscosity of dough-like samples. The material whose extensional viscosity is
to be measured is placed between two lubricated plates and compressed at either a
constant velocity or a constant strain rate. The squeezing force is used to calculate
the extensional viscosity. If the sample fills the plates completely, i.e., sample and
plate radius is the same during the test, then the method is known as the constant
area method. If the sample radius is less than the plate radius, it is called constant
volume method.
Extensional strain is calculated using the following equation:

1 V
e¼ t (4.4)
2 HðtÞ
For the constant area method and constant crosshead velocity, the extensional
viscosity in lubricated squeezing flow is given by the following equation (Campa-
nella and Peleg 2002):
2FðtÞHðtÞ
mb ¼ (4.5)
pR2 V
where mb is the extensional viscosity, F(t) is the force at time t after commencement
of the experiment, H(t) is the momentary sample height, R is the radius of the
sample, which is the same as the radius of the plate (constant area method), and V is
the crosshead velocity or velocity of deformation.
At a constant strain rate e_ , extensional viscosity is given by
2FðtÞ
mb ¼ (4.6)
pR2 e_
Apart from the simplicity, this method offers a practical way to avoid the
extensive structural disruption that occurs at the narrow space of the die. By
using lubricated squeezing flow, this damage can be almost completely eliminated
and the specimen can be tested as practically intact by carefully placing the material
on the lower plate with a wide spatula or spoon (Suwonsichon and Peleg 1999;
Corradini et al. 2000). For details and variations of the technique for use in food,
please refer to the review by Campanella and Peleg (2002).
4.3.2 Structural Characterization
As described earlier, the molecular and structural characteristics of fiber may play a
significant role in its functionality during processing and also in the human body.
This discussion will focus on a few techniques commonly used for structural
characterization of polysaccharides in general, and arabinoxylans, which are the
major polysaccharides in corn fiber, in particular. The specific characteristics that
appear to be more significant for fiber functionality are chemical nature, constituent
sugars, linkage pattern, molecular weight, molecular size, degree of branching, and
branching pattern.
4.3.2.1 Chemical Analysis
The chemical constituents of cereal brans have been discussed in some detail above.
The chemical characterization of fiber is a fundamental consideration in its
80 M. Kale et al.
evaluation for fiber fortification. As mentioned before, cellulose, hemicelluloses,

and lignins are the principal constituents of brans. Some structural proteins and
phenolic acids which are components of cell walls are also present. Invariably, bran
has some starch associated with it. Also, there is some lipid constituent particularly
associated to the germ that can be mixed with the bran during the milling operation.
Thus, cereal bran represents a complex system of starch and non-starch polysac-
charides (arabinoxylans), proteins, and lipids. Fiber modification, particularly
chemical modification as discussed before, is often aimed at enriching a particular
fraction which has good functionality and nutritional properties. Thus, the chemical
analysis of the fiber is arguably one of the most important parts of its characteriza-
tion. Various techniques available for monosaccharide analysis and to determine
the amounts of lipids, protein, and phenolic acids in the fiber can be used to
chemically characterize the fibers. A detailed description of each technique is
beyond the scope of this chapter and the reader is referred to specific works on
each technique for further detail.
Monosaccharide Analysis
The major polysaccharides of cereal cell walls are cellulose, glucuronoarabinox-

ylans, b-D-glucans and some other glycans such as glucomannans (Carpita 1996).
Thus, the various monosaccharides present in the bran are D-glucose, D-xylose,
L-arabinose, D-galactose, glucuronic acid, and D-mannose. Monosaccharide analy-
sis is often carried out using chromatography techniques, both gas chromatography
(GC) and high-pressure liquid chromatography (HPLC). Trifluoroacetic acid is
used to hydrolyze the polysaccharide. This can be followed by HPLC analysis of
the monosaccharides. For GC analysis, the monosaccharides have to be derived to
make them volatile. A common derivation technique is conversion of the sugars
into their alditol acetates (Englyst et al. 1982).
Protein Estimation
Protein content can be measured using the Bradford assay or, more commonly, the
bicinchoninic acid (BCA) method, both of which are colorimetric methods. The
Kjeldahl and Dumas methods are based on nitrogen content estimation and are less
common owing to their tedious nature.
Phenolic Acid Analysis
There are several techniques for identification and quantification of phenolic acids.
Ferulic acid, being the major phenolic acid of cereal brans, is of relevance to the
incorporation of fibers in food products. Barberousse et al. (2008) reviewed various

techniques for quantification of ferulic acid, including ultraviolet (UV) spectrome-
try, fluorescence spectroscopy, and HPLC with UV detection.
Lipid Analysis
A normal phase HPLC technique with evaporative light-scattering detection was

developed for non-polar lipid analysis by Moreau et al. (1996). This technique was
recently used by Yadav et al. (2007) for quantification of non-polar lipids in corn
fiber gum extracted from wet-milled corn bran using alkaline hydrogen peroxide.
4.3.2.2 Spectroscopy Techniques
Nuclear magnetic resonance (NMR) and Fourier transform infrared spectroscopy

(FT-IR) are the most commonly used techniques to evaluate the fine structure of
polysaccharides. 13C-NMR spectroscopy has been widely used to determine the
linkage pattern, relative contents of mono- or di-substituted sugar residues, and to
elucidate the relative contents of different monosaccharides present (Izydorczyk
and Biliaderis 1995). NMR spectroscopy of the polysaccharide itself cannot pro-
vide information about the sequence of monosaccharide residues. Instead, break-
down of the chain with site-specific enzymes can be used to degrade the polymer
backbone followed by a study of the structures of formed oligomers. Kacurakova
et al. (2000) studied the structure of plant cell wall polysaccharides, including
pectins and hemicelluloses using FT-IR. They suggest that the FT-IR spectrum in
the 1,200–800 cm1 region can be used to identify polysaccharides and report the
absorption maxima for the studied polysaccharides. Robert et al. (2005) analyzed
the FT-IR spectrum of wheat endosperm arabinoxylans and performed principal
component analysis of some model mixtures of polysaccharides to determine the
relative proportions of each in the mixture. Kacurakova and Wilson (2001)
reviewed FT-IR spectroscopy techniques for various carbohydrates, including cell
wall polysaccharides.
4.3.2.3 Thermal Analysis
One thermal property of polysaccharides that is most relevant to its functionality

during processing is the glass transition temperature and its variation with moisture
content. Thermal decomposition onset temperatures may also be significant,
depending on the kind of processing. These properties can be studied using differ-
ential scanning calorimetry (DSC), dynamic mechanical analysis (DMA), and
thermo gravimetric analysis (TGA) (Abiad et al. 2009).
82 M. Kale et al.
4.3.2.4 High-Pressure Size Exclusion Chromatography Techniques
High-pressure size exclusion chromatography (HPSEC), also known as gel perme-

ation chromatography, is commonly used to separate polymers based on their size.
HPSEC can be used as a method to estimate molecular weight by calibrating the
instrument with standards of known molecular weight. The elution time of the
sample can therefore be related to its molecular weight. However, this method of
estimation may not be accurate for fibers as differences in degree of branching and
in size may cause polymers of the same molecular weight to elute at different times.
Thus, for absolute characterization of molecular weight, an analysis technique such
as light scattering coupled with HPSEC is often required.
4.3.2.5 Light Scattering
Light scattering is a powerful tool for analysis of macromolecules in solution. It

has been extensively applied in the world of synthetic polymers for determination
of molar mass, size, degree of branching, and calculation of parameters which are
very relevant to the functionality of the polymers during processing, such as
branching ratio and persistence length of the polymer. Although its potential
application to food polymers is equally significant, it could be somewhat compli-
cated by the inherent polydispersity and diverse nature of food macromolecular
components. For details regarding the theory of light scattering, the reader is
referred to the seminal works of Debye (1944), Zimm (1948a, b, c), and Wyatt
(1993).
The results presented below focus on static light scattering for the analysis of
polysaccharides. Since static light scattering measures the average molecular
weight and size, it is often desirable to fractionate the molecules based on size
prior to the light-scattering analysis. This is most commonly established by HPSEC,
although other liquid chromatography techniques may also be used.
HPSEC-MALS
HPSEC separates molecules based on their size, with the larger molecules eluting
first from the column followed by the smaller molecules. When an MALS
instrument is used as an SEC detector, a concentration detector such as a refractive
index detector or a UV detector (for molecules which absorb light in the UV
range, such as proteins) is often coupled to it. This enables calculation of concen-
tration for each data slice, which makes it possible to calculate the molecular
weight (Mw) and the root mean square (rms) radius (also known as radius of
gyration of the molecule, rg) for each individual data slice, as well as the average
across a peak.
Conformation Plots and Branching Analysis
Knowing the average molecular weight (Mw) and rms radius/radius of gyration (rg)
across a peak, a log–log plot of rg versus Mw, called a conformation plot, can be
used to obtain information about the conformation of the molecule in solution. For
instance, for a solid sphere, the radius is proportional to the cube root of molar
mass; thus the conformation plot will have a slope of 0.33 (Wyatt 1993). For rigid
rods, the slope will be 1, while random coils should give a slope of 0.5–0.6. In
general spherical-like molecules in solution are associated with a large degree of
branching. Thus, these plots can be used to compare the degree of branching
between different fiber polymers; lower slope of the conformation plot implies a
greater degree of branching, which, as discussed, may have an important effect on
the expansion of fiber-enriched foods. Another method of determining the degree of
branching of a polymer is the use of the radius method described by Zimm and
Stockmayer (1949). In this method, the branching ratio of a polymeric molecule is
simply defined as the ratio of rms radius of the polymer to that of a linear polymer
that has the same composition and molecular weight. This method has been used by
Grcev et al. (2004) to find the degree of branching of polyvinyl acetate. Its use for
natural polymers such as polysaccharides may be limited due to the difficulty in
obtaining linear and branched polymers of the same molecular weight. So, for these
cases, the conformation plots serve as a simple method to qualitatively evaluate
branching.
Other methods that can be used include spectroscopy techniques such as NMR,
mass spectrometry, and indirect techniques such as GC to determine individual
sugar residues if the branches and backbone are known to be made of different
sugars. Some data showing the use of light scattering to determine molecular
weight, size, and degree of branching of alkali-soluble corn arabinoxylans are
presented and discussed in a later section.
4.3.2.6 Fermentation Profiling
The fermentation profile of fiber within the colon is also an important aspect to be
considered due to the implications of the role of fiber on human health. In particular,
the type and amount of short chain fatty acids produced, as well as the completeness
and rapidity of fermentation are all important functional parameters to consider. In
vitro fermentation has been used to characterize the pH, gas production, and short-
chain fatty acid production by the polysaccharide in the colon. Studies point to a
desirability for fibers with controlled fermentation to distribute beneficial fermenta-
tion products into the distal colon, while providing some bulking effect for regularity
(Rose et al. 2007). Adiotomre et al. (1990) suggested in vitro techniques, dialysis,
and fermentation to screen polysaccharides for their impact on nutrient absorption,
sterol metabolism, bulking, and fermentation.
84 M. Kale et al.
4.4 Structure and Functionality of Alkali-Treated

Corn Arabinoxylans
A chemical method to improve the functionality of corn bran so it may be

incorporated at high levels in an extrusion process is described below.
4.4.1 Description of Chemical Treatment
Corn bran was mixed with 5% NaOH in a 1:10 ratio for 2 h at room temperature to
hydrolyze ferulic acid ester cross-linkages following the procedure described by
Doner et al. (2000). The mixture was then neutralized with hydrochloric acid, and
95% ethanol was added to the neutralized extract in a ratio of 2:1. The precipitate,
termed “alkali-treated bran” (ATB), was dried at 50 C, and milled using a CyclotecTM
cyclone mill equipped with a screen of mesh size 40.
Separation of the alkali-soluble bran (ASB) hemicellulosic fraction from the
alkaline-insoluble fraction was done by centrifuging the alkaline mixture (before
the neutralization step to produce ATB) at 11,159 g for 8 min. The supernatant was
neutralized with HCl and 80% ethanol was added to the supernatant in a ratio of 2:1
to obtain the ASB. The precipitate was dried at 50 C and milled using a cyclone
mill equipped with a screen of mesh size 40.
After alkali treatment, there was a significant reduction in insoluble fiber, from
79–52%, and an increase in soluble fiber from 1.6–30% (Blake 2006). The main
dietary fiber component of the ASB fraction contains 64% of soluble dietary fiber,
which is the branched hemicellulose B. Alkali treatment of corn bran disrupts its
supramolecular structure producing dietary fiber that can be separated into various
components, which are hemicelluloses A, B, and an alkali-insoluble fraction that
consists of cellulose and cross-linked hemicellulose (hemicellulose C), and starch.
4.4.2 Rheological Properties and Extrusion Expansion

of Modified Fibers Mixed with Cornmeal
4.4.2.1 Extrusion Trials
Each fiber, that is, ATB, ASB, and unmodified bran (UMB), was mixed with
cornmeal in a ratio so as to produce a mixture containing 26% total dietary fiber.
The details of the process and extrusion equipment used can be found in Pai et al.
(2009). Sectional expansion index (SEI) was measured in each case. SEI is defined
as the squared ratio of average diameter of the extrudate to that of the die diameter
(Alvarez-Martinez et al. 1988). The SEI of each extrudate is listed in Table 4.1
below. The SEI of extrudates containing 26% ASB was equivalent to that of the
Table 4.1 Sectional Sample SEI

expansion indices of
Cornmeal 27
cornmeal extruded with
Cornmeal + UMB 11
various fibers
Cornmeal + ATB 21
Cornmeal + ASF 27
Fig. 4.3 Extrudates obtained from mixing cornmeal with unmodified corn bran and alkali-
modified corn fiber
cornmeal with no fiber added. Incorporation of both the modified bran fractions
ASB and ATB resulted in much higher radial expansion as compared to the UMB.
The difference in expansion is clearly visible in Fig. 4.3, with the alkali modified
fiber giving radial expansion similar to that of the control (cornmeal only).
4.4.2.2 Capillary Rheometry
In order to study the effect of incorporation of these fibers on the melt shear
rheology, a twin-bore capillary rheometer (Rosand RH 2,000 Capillary Rheometer,
Malvern Instruments, Southborough, MA) was used. These experiments were
performed to characterize the shear rheology of the melt at conditions similar to
those of the extrusion operation. Due to operational limitations of the capillary
rheometer, samples with 20% total fiber and 25% moisture were prepared. Shear
rates in a range of 20–800s1 for extrusion operations have been reported by several
researchers (McMaster et al. 1987; Senouci and Smith 1988; Lai and Kokini 1990;
Padmanabhan and Bhattacharya 1993; Seethamraju and Bhattacharya 1994;
Della Valle et al. 1996; Drozdek and Faller 2002; Sandoval and Barreiro 2007).
86 M. Kale et al.
Fig. 4.4 Apparent shear viscosity versus shear rate data from capillary rheometry of cornmeal
with various fibers (shear rate range: 100–200 s1)
Steady shear viscosities were measured at shear rates of 100, 150, 175, and 200 s1.
No end correction was applied due to the high sample requirement. A long die was
used (L/D ¼ 8) in order to minimize the contribution of the nonlinear portion of the
pressure drop (Byler and Kwei 1971).
Addition of ATB at the 20% level resulted in a significant increase in shear
viscosity, while a decrease was noted with the addition of ASB. However, by the
addition of UMB to the cornmeal, the shear viscosity increased substantially
compared to the control. Figure 4.4 shows the differences in melt shear rheology
based on addition of fibers. As shown by the linearity of the plot in logarithmic
coordinates, all samples showed a pseudoplastic behavior and obeyed the power
law model in the experimental shear rate range tested.
4.4.2.3 Extensional Rheology
Lubricated squeezing flow between two plates was chosen to characterize the
extensional viscosity of the mixed fiber samples. Huang and Kokini (1993) and
Wikstr€om and Bohlin (1999) used this technique for characterizing the extensional
viscosity of wheat doughs.
In this work, it was necessary to add gluten at an 8.3% level to produce a
cohesive dough. Thus, the mixture of cornmeal and gluten served as a base to
observe the effect of the different fiber components on the extensional rheology of
the dough. The samples prepared in this way consisted of cornmeal mixed with
Fig. 4.5 Extensional viscosity versus strain data obtained from lubricated squeezing flow rheo-
metry of cornmeal with various fibers
gluten and each of the fiber components (ATB, ASB, and UMB), with a resultant
composition of approximately 20% total dietary fiber, 25% cornmeal, 8.3% gluten,
and water making up the rest of the sample. Lubricated squeezing flow experiments
were conducted on a Sintech 1/G Universal Testing Machine (MTS, Eden Prairie,
MN). Samples of initial height 10 mm were compressed between Teflon plates
(25.4 mm in diameter) which were lubricated with silicone oil. Experiments were
conducted with the constant area method with crosshead velocities of 10 and
20 mm/min until the point where the force reached a value of 100 lbf. Results
showed that the addition of gluten served to improve the squeezing flow testing and
its reproducibility without qualitatively changing the behavior of the samples. All
samples were shown to exhibit strain thinning behavior. Differences in extensional
rheology were observed with the type of fiber used. Cornmeal with ATB showed
the highest extensional viscosity, while addition of UMB resulted in an extensional
viscosity higher than that for cornmeal with no added fiber, but ASB addition
resulted in an extensional viscosity slightly lower than that of the control cornmeal
(Fig. 4.5).
4.4.2.4 Effect of Rheology on the Expansion Process
Once bubble growth is initiated, the bubbles have to overcome the resistance
created by the shear viscosity in order to grow. After the bubbles have grown
sufficiently, the melt film between the bubbles undergoes biaxial extension (Gendron
and Daigneault 2000; Steffe 1996). Hence, consideration of extensional viscosity
becomes an essential part in the prediction of extrusion expansion from the melt
rheology. Once a certain normal stress in the melt film has been exceeded due to
88 M. Kale et al.
further bubble expansion, the melt film ruptures, causing bubble collapse. For the
bubbles to remain intact, the melt should have an extensional viscosity high enough
to withstand the extensional stress caused by bubble expansion.
Excluding the sample of cornmeal with ASB, it was observed that shear viscosity
increased for cornmeal, cornmeal with ATB, and cornmeal with UMB, whereas
extensional viscosity increased for cornmeal, cornmeal with UMB, and cornmeal
with ATB. Addition of both UMB and ATB to cornmeal increased shear and
extensional viscosities when compared to cornmeal alone, thus increasing overall
resistance to expansion during the two stages of bubble growth, that is, when shear
and extensional flows, respectively, were the prevalent flows, which caused the
observed lower expansion. However, comparing the rheology of cornmeal with
UMB to that of cornmeal with ATB, it can be concluded that an increase in
extensional and decrease in shear viscosities resulting from alkali treatment
resulted in an increased extrusion expansion. This trend suggests that a relatively
high extensional viscosity and a low shear viscosity appear to be the rheological
parameters that cause good expansion.
Although cornmeal with ASB had a significantly lower shear viscosity compared
to cornmeal alone, both samples showed a similar radial and longitudinal expan-
sion, suggesting that the shear viscosity could not be the sole rheological parameter
governing the extrudate expansion and that extensional viscosity also has a role to
play in expansion. The similarity in expansion can be explained by the effect of
rheology on the expansion mechanism. Sun (2004) and Stange and Munstedt (2006)
showed that higher shear viscosity causes a slower diffusion of the gas in polymers
resulting in better foam expansion, which would explain the better expansion
observed on extrudates prepared with cornmeal alone. Although the extensional
viscosity of the ASB sample was lower than that for the cornmeal control, the
similar expansion probably cannot be explained by the reduced extensional viscosity
of the sample with ASB compared to cornmeal alone. In the present study, it was
observed that extrudates with bran had a larger number of bubbles compared to the
cornmeal control (not shown). The presence of a high amount of bubbles of large
sizes may have increased the heat transfer rate of the extrudate containing ASB as
compared to cornmeal. This effect would cause a more rapid heat and moisture loss,
which could have promoted the attainment of the glass transition sooner than the
control sample, thus preventing further bubble growth. The combination of reasons
presented above probably explains why cornmeal with ASB has the same expansion
as the control samples.
4.4.2.5 Possible Effects of Structure and Composition of Fibers on Rheology
An increase in the soluble arabinoxylan content in the ATB fiber compared to UMB
caused a decrease in shear viscosity and an increase in extensional viscosity. On the
other hand, ASB (highest soluble fiber content) exhibited both lower shear and
extensional viscosities as compared to the control with no fiber. This finding
suggests that the interaction of ASB with starch is different than the interactions of
UMB or ATB with starch.
The major component of ASB is hemicellulose B (64%), which is a branched
arabinoxylan (Whistler 1993). Due to its high degree of branching, it is postulated
that ASB is able to interact with starch differently than the other two fibers. An
enhanced functionally that seems to result from the molecular dispersal is produced
by the chemical treatment. The favorable interaction between starch and the ASB
fraction would in turn result in a rheological behavior of the fiber in the melt similar
to that of starch, that is, without disrupting the melt structure and resulting in a melt
rheology which favors bubble growth and expansion. A similar phenomenon has
been observed during the extrusion of synthetic polymers, where the use of highly
branched polymers such as low-density polyethylene (LDPE) results in higher
expansion than linear high-density polyethylene (HDPE) (Gendron and Daigneault
2000). Li et al. (2004) reported that higher amylopectin content leads to lower melt
shear viscosity, resulting in higher expansion. Thus, the highly branched arabinoxylan
fraction of hemicellulose B in the ASB fiber would lead to higher expansion.
Conversely, the main components of the water-insoluble portion of the ATB fiber
consist of cross-linked insoluble hemicellulose, such as hemicellulose C (15% w/w
of bran, linear xylan polysaccharide with few arabinose branches), and cellulose
(16% w/w) (Sugawara et al. 1994). UMB is composed of 79% insoluble dietary
fiber as compared to ATB, which has 52% (Blake 2006). The high proportion of
cross-linked and linear polymers that compose UMB are likely to break the
continuity of the melt forming a multiphase melt, whereas ATB may form a melt
that is less discontinuous, thus favoring some expansion.
4.4.3 Effect of Branching on Rheology of Alkali-Soluble

Corn Arabinoxylans
Physico-chemical characterization and rheological characterization are comple-

mentary techniques for assessment of fiber quality. While most of the rheological
measurements can be used to simulate actual processing conditions and directly
assess the behavior of the fiber under these conditions, there are other rheological
methods aimed to be related with structural parameters, which can be used as tools
for fiber quality assessment. In order to illustrate the use of both of these techniques,
differences in structural and rheological properties of some alkali-soluble corn bran
fractions are discussed.
Alkali treatment of bran was carried out as described previously and the alkaline
supernatant was separated from the insoluble fraction. The alkaline supernatant was
acidified to precipitate the alkali soluble and separate the water-insoluble fraction.
Ethanol was then added to this solution so that the solution contained 20% by
volume of ethanol. The fraction that precipitated out was separated and freeze-dried
and regarded as the 0–20% fraction. The ethanol concentration was then increased
90 M. Kale et al.
to 40% and the fraction called the 20–40% fraction was isolated. Similarly, the
40–60% and 60–80% fractions were collected. This fractionation was carried out
for the purpose of obtaining fractions homogeneous in terms of molecular char-
acteristics of the polysaccharides, with lesser variability in their structures. It was
observed that most of the polysaccharides precipitated in the 20–40% and 40–60%
fractions. The shear and extensional viscosities in doughs containing these fibers
and their solutions in water were measured as follows.
4.4.3.1 Melt Shear Rheology Using Capillary Rheometry
Melt shear rheology was measured using the same methodology as described in the
capillary rheometry section. Samples were prepared by mixing cornmeal, fiber, and
water such that each sample contained 25% fiber and 50% cornmeal. Steady shear
experiments were conducted at 120 C, which was the extrusion temperature for
trials conducted in our laboratory (Blake 2006). Steady shear viscosities were
measured at shear rates of 75, 100, 150, 175, and 200 s1. The variation of shear
viscosity with shear rate is shown in Fig. 4.6. It is clear that addition of the fiber
fractions 20–40% and 40–60% reduces the shear viscosity of the melt. There is also
some difference between the measured shear viscosities. The samples containing
the 20–40% fraction show a slightly higher viscosity compared to those with the
40–60% fraction, but addition of both fiber fractions significantly lowers the
viscosity compared to cornmeal alone.
Corn meal
Corn meal + 20–40% fraction
Corn meal + 40–60% fraction
Shear viscosity (Pa.s)
1000
60 80 100 120 140 160 180 200 220

Shear rate (1/s)
Fig. 4.6 Apparent shear viscosity as a function of shear rate for cornmeal mixed with the 20–40%
and 40–60% fractions of ASB (range: 75–200 s1)
4.4.3.2 Extensional Rheology Using Lubricated Squeezing Flow
The effect of individual fiber fractions on extensional viscosity of dough containing

these fibers along with the control samples (no fiber) was characterized by the
lubricated squeezing flow technique. Samples were prepared with cornmeal mixed
with each of the fiber components, that is, 20–40% and 40–60% ASB fractions,
resulting in a final fiber composition of 25%.
The constant area method was used at constant biaxial strain rates of 0.1 s1 and
1
1 s until the total biaxial strain reached a value of 90% Cauchy strain. Wei et al.
(2007) reported that the dependence of extensional viscosity on strain is probably
the result of orientation developed in the polymer molecules as the fluid undergoes
extensional deformation. Cornmeal with no fiber showed a higher extensional
viscosity compared to samples with added fiber. In order to distinguish the rheological
behavior of the two fiber fractions, results for cornmeal are excluded, but its
elongational viscosity is slightly higher than the elongational viscosity of the
fiber-enriched samples. At a lower strain rate of 0.1 s1 differences in extensional
viscosity between the 40–60% and 20–40% fractions are quite significant (Fig. 4.7).
This difference is minimal at 1 s1, which indicated that there could be fine
structural differences between the two fractions that were apparent only at low
strain rates. Moreover, although the 40–60% fraction samples exhibited a lower
extensional viscosity at 0.1 s1, that is, a lower resistance against deformation, they
showed a lower decrease in extensional viscosity when the strain rate was increased
from 0.1 to 1 s1, as compared to samples with 20–40% fraction. It was hypothe-
sized that the likely higher branching of the 40–60% fraction was responsible for
this behavior.
Corn meal + 20–40% –0.1s–1

Extensional Viscosity (Pa.s)
Corn meal + 40–60% –0.1s–1
Corn meal + 20–40% –1s–1
100000
Corn meal + 40–60% –1s–1
0.5 1.0 1.5 2.0 2.5

Cauchy Strain
Fig. 4.7 Extensional viscosity as a function of Cauchy strain for cornmeal mixed with 20–40%
and 40–60% fractions of ASB
92 M. Kale et al.
4.4.3.3 Solution Shear Rheology Using Rotational Rheometer
The 20–40% and 40–60% fractions were dissolved in water to make 5% solutions
which were placed in concentric cylinder geometry in a rheometer (ARG2 from TA
Instruments, Newcastle, DE) and viscosity was determined at shear rates varying
from 1 to 200 s1. Both samples showed shear thinning behavior for shear rates less
than 20 s1. Between 20 and 200 s1 the samples showed a Newtonian behavior
(Fig. 4.8).
The 40–60% fraction had a lower solution viscosity compared to the 20–40%
fraction, but similar molecular weight distribution profiles (results from HPSEC
not shown). This suggests that the 40–60% fraction has a smaller hydrodynamic
radius and thus a more highly branched structure. It is thus hypothesized that
there are fine structural differences between the two fractions of alkali-soluble
corn bran, as seen from the difference in their behavior in these rheological
tests.
4.4.3.4 Branching Analysis Using HPSEC-MALS
In order to confirm the existence of structural differences as suggested by rheologi-

cal properties, the two fractions were analyzed by HPSEC-MALS and conforma-
tion plots were made as described above. The slope of the conformation plot for the
40–60% fraction was of 0.39, which is slightly lower than that for the 20–40%
fraction (0.44), as can be seen in Fig. 4.9. These results confirm the qualitative
0.042
0.040
20-40%
40-60%
0.038
Shear viscosity (Pa.s)
0.036
0.034
0.032
0.030
0.028
0.026
0 50 100 150 200

Shear rate (1/s)
Fig. 4.8 Comparison of solution (5%) shear rheology between 20–40% and 40–60% fractions for
shear rates between 1 and 200 s1
conformation plot
Fraction 40- 60%.vaf (0.39±0.02) Fraction 20-40%.vaf (0.44±0.01)
rms radius (nm)
100.0
1.0×106 1.0×107
molar mass (g/mol)
Fig. 4.9 Conformation plots for 20–40% and 40–60% fractions
information obtained from the rheology tests, which indicated that the 40–60%
fraction has a more branched structure than the 20–40% fraction. Thus, while
rheological characterization can be used to hypothesize about differences in fine
structure of the polymers, branching analysis using, for example, HPSEC-MALS
must be used to confirm differences in branching pattern between the polysaccha-
ride polymers.
4.4.3.5 Possible Implications in Extrusion
Irrespective of the strain rates, it was found that the control cornmeal with no added
fiber exhibited the highest extensional viscosity among these tests. Addition of
fibers in all forms decreased shear and extensional viscosities of the cornmeal fiber
composites. It was also noted that the presence of modified fibers allowed the
sample to be squeezed to extremely high strains without evidence of fracture
along the lateral surface of the cylindrical sample. The different behaviors can be
observed by comparing Fig. 4.10a (cornmeal) and b (composite of cornmeal and
25% ASB), where the latter showed no fracture and a homogeneous deformation.
Homogeneous deformation during extensional flow is a characteristic of branched
polymers (Stange and Munstedt 2006). Homogeneous deformation prevents the
growth of inhomogeneities that can result in sample breakage or cause the inho-
mogeneities to grow in a restrained manner. It could very well be expected
that these fibers will have a favorable impact on extrusion expansion due to
homogeneous deformation.
94 M. Kale et al.
Fig. 4.10 Pictures of samples after lubricated squeezing flow test: (a) cornmeal, (b) cornmeal
with 20–40/40–60% fractions of ASB. Cornmeal shows extensive fracture, while samples with
ASB fractions are squeezed without fracture
4.5 Conclusion
Dietary fiber is an important part of the diet, with significant nutritional and
physiological benefits. However, poor functionality during processing makes pro-
duction of high-fiber foods a challenge. Some strategies for fiber modification,
including physical, chemical, and enzymatic methods, may be employed to
improve this functionality and make addition of high levels of fiber into processed
foods possible. Several sources of fiber are available to the food industry, cereal
brans being arguably among the cheapest and most abundant. Physico-chemical and
rheological characterization techniques can be successfully employed for the
assessment of fiber quality for potential incorporation into different foods. Some
rheological tests are designed to simulate actual processing conditions, while others
give information about the structure of the polymer. Complementary information
on these structures can be obtained by physical and chemical techniques such as
light scattering and monosaccharide analysis. Together, all of these techniques,
along with modification strategies, can be employed to overcome the engineering
challenge of fiber incorporation and aid the food industry in producing better
products that will help to improve public health.
Acknowledgments The authors wish to acknowledge Midwest Advanced Food Manufacturing

Alliance (MAFMA) and Dr. Kevin Schilling of Grain Processing Corporation for his continued
support during this research.
References
Abiad MG, Carvajal MT, Campanella OH (2009) Review on methods and theories to describe the
glass transition phenomenon: applications in food and pharmaceutical products. Food Eng Rev
1:105–132
Adiotomre J, Eastwood MA, Edwards CA, Brydon WG (1990) Dietary fiber – in vitro methods
that anticipate nutrition and metabolic-activity in humans. Am J Clin Nutr 52:128–134
Alvarez-Martinez L, Kondury K, Harper J (1988) A general model for expansion of extruded
products. J Food Sci 53:609–615
Arrigoni E (2001) Enzymatic modification of dietary fiber sources. In: Cho S, Dreher M (eds)
Handbook of dietary fiber. Marcel Dekker, New York, pp 206–208
Barberousse H, Roiseux O, Robert C, Paquot M, Deroanne C, Blecker C (2008) Analytical
methodologies for quantification of ferulic acid and its oligomers. J Sci Food Agric
88:1494–1511
Bhattacharya M, Seethamraju K, Padmanabhan M (1994) Entrance pressure-drop studies of
cornmeal dough during extrusion-cooking. Polym Eng Sci 34:1187–1195
Blake O (2006) Effect of molecular and supramolecular characteristics of select dietary fibers on
extrusion expansion. Ph.D. thesis, Purdue University, Indiana, USA
Breen M, Seyam A, Banasik O (1977) The effect of mill by-products and soy protein on the
physical characteristics of expanded snack foods. Cereal Chem 54:728–736
Byler L, Kwei T (1971) Flow behavior of polyethylene melts containing dissolved gases. J Polym
Sci 35:165–176
Campanella O, Peleg M (2002) Squeezing flow viscometry for nonelastic semiliquid foods: theory
and applications. Crit Rev Food Sci Nutr 42:241–264
Carpita N (1996) Structure and biogenesis of the cell walls of grasses. Annu Rev Plant Phys
47:445–476
Chinnaswamy R, Hanna MA (1988) Relationship between amylose content and extrusion-expansion
properties of corn starches. Cereal Chem 65:138–143
Claye S, Idouraine A, Weber C (1996) Extraction and fractionation of insoluble fiber from five
fiber sources. Food Chem (UK) 57:305–310
Colliopoulos J (1991) Psyllium mucilloid fiber food products. US Patent 5,009,916
Corradini M, Stern V, Suwonsichon T, Peleg M (2000) Squeezing flow of semi-liquid foods
between parallel Teflon® coated plates. Rheol Acta 39:452–460
Debye P (1944) Light scattering in solutions. J Appl Phys 15:338–342
Decker K (2007) Fiber goes with the flow. Food Product Design. http://www.foodproductdesign.
com/articles/beverages/7a1applications.html
Della Valle G, Collona P, Patria A, Vergnes B (1996) Influence of amylose content on the viscous
behavior of low hydrated molten starches. J Rheol 40:347–362
Doner LW, Hicks KB (1997) Isolation of hemicellulose from corn fiber by alkaline hydrogen
peroxide extraction. Cereal Chem 74:176–181
Doner L, Johnston D (2001) Isolation and characterization of cellulose/arabinoxylan residual
mixtures from corn fiber gum processes. Cereal Chem 78:200–204
Doner L, Sweeney G, Hicks K (2000) Isolation of hemicellulose from corn fiber. US Patent
6,147,206
Drozdek K, Faller J (2002) Use of a dual orifice die for on-line extruder measurement of flow
behavior index in starchy foods. J Food Eng 55:79–88
Ebringerova A, Hromadkova Z, Heinze T (2005) Hemicellulose. Adv Polym Sci 186:1–67
Englyst H, Wiggins HS, Cummings JH (1982) Determination of the non-starch polysaccharides in
plant foods by gas-liquid-chromatography of constituent sugars as alditol acetates. Analyst
107:307–318
Fan JT, Mitchell JR, Blanshard JMV (1994) A computer-simulation of the dynamics of bubble-
growth and shrinkage during extrudate expansion. J Food Eng 23:337–356
Foehse KB, Efstathiou JD, Stoll JR (1992) High soluble fiber barley expanded cereal and method
of preparation. US Patent 5,151,283
Garber BW, Hsieh F, Huff HE (1997) Influence of particle size on the twin-screw extrusion of corn
meal. Cereal Chem 74:656–661
Gelroth J, Ranhotra G (2001) Food uses of fiber. In: Cho S, Dreher M (eds) Handbook of dietary
fiber. Marcel Decker, New York, pp 435–452
96 M. Kale et al.
Gendron R, Daigneault L (2000) Rheology of thermoplastic foam extrusion process. In: Lee S (ed)
Foam extrusion. Technomic Publishing Co, Lancaster, PA, pp 35–80
Grcev S, Schoenmakers P, Iedema P (2004) Determination of molecular weight and size distribution
and branching characteristics of PVAc by means of size exclusion chromatography/multi- angle
laser light scattering (SEC/MALLS). Polymer 45:39–48
Guy R (2001) Extrusion cooking. Technologies and applications. Woodhead Publishing Limited,
Cambridge/England
Hingmann R, Marczinke BL (1994) Shear and elongational flow properties of polypropylene
melts. J Rheol 38:573–587
Hseih F, Mulvaney S, Huff S, Lue S, Brent J (1989) Effect of dietary fiber and screw speed on
some extrusion processing and product variables. Lebensmittel-Wissenshaft und-Technologie
22:204–207
Hseih F, Huff H, Lue S, Stringer L (1991) Twin-screw extrusion of sugar beet and corn meal.
Lebensmittel-Wissenshaft und-Technologie 24:495–500
Huang H, Kokini J (1993) Measurement of biaxial extensional viscosity of wheat flour doughs.
J Rheol 37:879–891
Izydorczyk M, Biliaderis C (1995) Cereal arabinoxylans: advances in structure and physicochemical
properties. Carbohydr Polym 28:33–48
Kacurakova M, Wilson R (2001) Developments in mid-infrared FT-IR spectroscopy of selected
carbohydrates. Carbohydr Polym 44:291–303
Kacurakova M, Capek P, Sasinkova V, Wellner N, Ebringerova A (2000) FT-IR study of plant cell
wall model compounds: pectic polysachharides and hemicelluloses. Carbohydr Polym 43:95–203
Kokini JL, Chang C, Lai L (1992) The role of rheological properties on extrudate expansion.
In: Kokini J, Ho C, Karwe M (eds) Food extrusion science and technology. Marcel Dekker,
New York, pp 631–653
Lai L, Kokini J (1990) The effect of extrusion operating conditions on the on-line apparent
viscosity of 98% amylopectin (Amioca) and 70% amylose (Hylon 7) corn starches during
extrusion. J Rheol 34:1245–1266
Li P, Campanella O, Hardacre A (2004) Using an in-line slit-die viscometer to study the effects of
extrusion parameters on corn melt rheology. Cereal Chem 81:70–76
Lue S, Hsieh F, Huff H (1991) Extrusion cooking of corn meal and sugar beet fiber: effects on
expansion properties, starch gelatinization, and dietary fiber content. Cereal Chem 68:227–234
Macosko C (1994) Shear rheometry: Pressure driven flows. In: Macosko C (ed) Rheology:
Principles, measurements and applications. Wiley-VCH, New York, pp 237–283
Malkin A, Isayev A (2006) Rheology: concepts, methods and applications. ChemTec Publishing. Online
version: http://knovel.com/web/portal/browse/display?_EXT_KNOVEL_DISPLAY_bookid¼2189
andVerticalID¼0
McMaster T, Senouci A, Smith A (1987) Measurement of rheological and ultrasonic properties of
food and synthetic polymer melts. Rheologica Acta 26:308–315
Mendonca S, Grossmann MVE, Verhe R (2000) Corn bran as a fiber source in expanded snacks.
Lebensmittel-Wissenschaft Und-technologie-Food Sci Technol 33:2–8
Moraru CI, Kokini JL (2003) Nucleation and expansion during extrusion and microwave heating.
Comprehensive Reviews in Food Science and Food Safety 2:120–138
Moreau RA, Powell MJ, Hicks KB (1996) Extraction and quantitative analysis of oil from
commercial corn fiber. J Agric Food Chem 44:2149–2154
Mottur G, Glass R (1985) Reduced calorie puffed snack food products. US Patent 4,517,204
National Academy of Sciences, Institute of Medicine (2002) Appendix E: dietary intake data from the
continuing survey of food intakes by individuals (1994–1996, 1998) In: Dietary reference intakes
for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and amino acids (Macronu-
trients) pp 801–819. Available at: http://www.nap.edu/books/ 0309085373/html/265.html
Ng A, Lecain S, Parker ML, Smith AC, Waldron KW (1999) Modification of cell-wall polymers of
onion waste III. Effect of extrusion-cooking on cell-wall material of outer fleshy tissues.
Carbohydr Polym 39:341–349
Ning L, Villota R, Artz W (1991) Modification of corn fiber through chemical treatments in
combination with twin-screw extrusion. Cereal Chem 68:632–636
Onwulata C, Konstance R, Smith P, Holsinger V (2001) Co-extrusion of dietary fiber and milk
proteins in expanded corn products. Lebensmittel-Wissenshaft und-Technologie 34:424–429
Padmanabhan M, Bhattacharya M (1993) Effect of extrusion processing history on the rheology of
cornmeal. J Food Eng 18:335–349
Pai D, Blake O, Hamaker B, Campanella O (2009) Importance of extensional rheological proper-
ties on fiber-enriched corn extrudates. J Cereal Sci 50:227–234
Ralet M, Della Valle G, Thibault J (1993) Raw and extruded fibre from pea hulls. I: composition
and physico-chemical properties. Carbohydr Polym 20:17–23
Ringe M, Stoll J (1991) R-T-E cereal with psyllium. US Patent 5,026,689
Robert P, Marquis M, Barron C, Guillon F, Saulnier L (2005) FT-IR investigation of cell wall
polysaccharides from cereal grains: arabinoxylan infrared assignment. J Agric Food Chem
53:7014–7018
Rose DJ, Demeo MT, Keshavarzian A, Hamaker BR (2007) Influence of dietary fiber on inflamma-
tory bowel disease and colon cancer: importance of fermentation pattern. Nutr Rev 65:51–62
Sandoval A, Barreiro J (2007) Off-line capillary rheometry of corn starch: effects of temperature,
moisture content and shear rate. Lebensmittel-Wisenshaft und- Technologie 40:43–48
Saulnier L, Thibault J (1999) Ferulic acid and diferulic acids as components of sugar beet pectins
and maize bran heteroxylans. J Sci Food Agric 79:396–402
Saulnier L, Vigouroux J, Thibault J (1995) Isolation and partial characterization of feruloylated
oligosaccharides from maize bran. Carbohydr Res 272:241–253
Schulze E (1891) Information regarding chemical composition of plant cell membrane. Ber Dtsch
Chem Ges 24:2277–2287
Seethamraju K, Bhattacharya M (1994) Effect of ingredients on the rheological properties of
extruded corn meal. J Rheol 38:1029–1044
Senouci A, Smith A (1988) An experimental study of food melt rheology. I. Shear viscosity using a
slit die viscometer and a capillary rheometer. Rheologica Acta 27:546–554
Shelton D, Lee W (2000) Cereal carbohydrates. In: Kulp K, Ponte J (eds) Handbook of cereal
science and technology. Marcel Dekker, New York, pp 385–416
Stange J, Munstedt H (2006) Rheological properties and foaming behavior of polypropylenes with
different molecular structures. J Rheol 50:907–923
Steffe J (1996) Rheological methods in food process engineering, 2nd edn. Freeman Press, East
Lansing, MI
Sugawara M, Tetsuya S, Totsuka A, Takeuchi M, Ueki K (1994) Composition of corn hull dietary
fiber. Starch-Staerke 46:335–337
Sun S (2004) Physical chemistry of macromolecules: Basic principles and issues, 2nd edn. Wiley,
New York
Suwonsichon T, Peleg M (1999) Imperfect squeezing flow viscometry of mustards with suspended
particulates. J Food Eng 39:217–226
USDA (2005) Backgrounder 2005 dietary guidelines advisory committee report [Online]. Avail-
able at United States Department of Agriculture: http://www.health.gov/dietaryguidelines/
dga2005/Backgrounder.html
van Lengerich B, Larson M (2000) Cereal products with inulin and methods of preparation.
US Patent 6,149,965
Wei X, Collier J, Petrovan S (2007) Shear and elongational rheology of polyethylenes with
different molecular characteristics. II. Elongational rheology. J Appl Polym Sci
104:1184–1194
Whistler R (1993) Hemicelluloses. In: Whistler R, BeMiller J (eds) Industrial gums, 3rd edn.
Academic, San Diego, CA
Wikstr€om K, Bohlin L (1999) Extensional flow studies of wheat flour dough. II. Experimental
method for measurements in constant extension rate squeezing flow and application to flours
varying in breadmaking performance. J Cereal Sci 29:227–234
98 M. Kale et al.
Wyatt PJ (1993) Light-scattering and the absolute characterization of macromolecules. Anal Chim
Acta 272:1–40
Yadav MP, Moreau RA, Hicks KB (2007) Phenolic acids, lipids, and proteins associated with
purified corn fiber arabinoxylans. J Agric Food Chem 55:943–947
Zimm BH (1948a) The dependence of the scattering of light on angle and concentration in linear
polymer solutions. J Phys Colloid Chem 52:260–267
Zimm BH (1948b) The scattering of light and the radial distribution function of high polymer
solutions. J Chem Phys 16:1093–1099
Zimm BH (1948c) Apparatus and methods for measurement and interpretation of the angular
variation of light scattering - preliminary results on polystyrene solutions. J Chem Phys
16:1099–1116
Zimm BH, Stockmayer WH (1949) The dimensions of chain molecules containing branches and
rings. J Chem Phys 17:1301–1314
Chapter 5
Gastric Digestion of Foods: Mathematical
Modeling of Flow Field in a Human Stomach
Samrendra Singh and R. Paul Singh
5.1 Introduction
Knowledge of the disintegration of solid foods in a human stomach is essential to

assess the bioavailability of nutrients in the gastrointestinal (GI) tract. Some of the
key factors impacting gastric digestion are stomach physiology including composi-
tion and rheological properties of gastric contents, stomach wall motility in fed/fasted
states, and hydrodynamics and mechanical forces that act on the ingested food. To
advance our current understanding of the digestion process, a quantitative description
of the flow field in the human stomach is necessary. The overall goal of this study was
to develop a mathematical model of the flow field in a human stomach and to use the
model to investigate the influence of stomach wall motility on the flow behavior.
5.2 Fluid Flow in a Human Stomach
The human stomach consists of two parts: the proximal stomach, which can be
divided into the fundus and the orad corpus (one third of corpus); and the
distal stomach, which includes the remaining corpus, the antrum, and the pylorus
(Szurszewski 1981) (Fig. 5.1). The stomach plays three major roles in the digestion
process: storing of high-density food, mixing, and emptying the chyme (Forte 1996;
Schubert and Makhlouf 1992; Soll and Berglindh 1994). The stomach reduces the
size of food particles and fat globules and disperses them in the gastric juice. It
also controls the pH, viscosity, and density of the gastric content with the help of
digestive secretions. After the ingestion of food, powerful contraction waves
originating from the mid-stomach move toward the pylorus at a speed of
2.5 mm/s. The depth of the contraction waves is shallow at the origin and increases
S. Singh and R.P. Singh (*)

Department of Biological and Agricultural Engineering, University of California, Davis, CA
95616, USA
e-mail: rpsingh@ucdavis.edu
100 S. Singh and R.P. Singh
Esophagus Fundus
Proximal
stomach
Corpus
Duodenum Antrum
Pylorus
d
A Distal stomach
Pyloric valve
Antral Contraction Wave (ACW)
Fig. 5.1 Schematic diagram of a human stomach (A wave amplitude, d width of wave) (not drawn
to scale)
as they approach the pylorus where they end. The occlusion diameter becomes
narrower during its travel and at the end the gastric content is squeezed back
through the narrow opening creating a retropulsive jet. These antral contraction
waves (ACWs) generate forces and fluid motions that break down and mix the
gastric content (Horowitz et al. 1994; Kelly 1980; Macagno and Christensen 1981;
Pallotta et al. 1998; Indireshkumar et al. 2000; Pal et al. 2004).
The size and shape of the stomach vary from one individual to another. It
depends upon many factors including age, food habits, posture, and the interval
since eating. The stomach walls are very flexible and the size of the stomach can
vary from a few hundred milliliters (when empty) to 2 L. The shape of the stomach,
it can be assumed, is somewhat like a “J.”
The food has a significant effect on the properties of the contraction waves.
The length of propagation, the amplitude of contractions, and the time duration the
stomach takes to generate the contraction waves are affected by the volume and
chemical and physical properties of the meal (Code and Carlson 1968; Meyer 1991;
Mayer 1994; Camilleri and Prather 1993). A large meal volume induces powerful
phasic contractions that originate higher up in the stomach as compared to small
volumes. A highly viscous food results in shallow contractions and a diluted or
aqueous meal generates deep contractions (Carlson et al. 1966). The deforming wall
boundaries of the stomach fill with a complex mixture of fluid and the food particulates
generate a complex and unsteady flow field. There are some distinctive features of the
flow field inside the stomach such as retropulsive jet and recirculating eddy structures.
As the ACW moves closer to the pylorus, the space between the contraction
wave and the pyloric valve decreases. The gastric content trapped inside this
5 Gastric Digestion of Foods: Mathematical Modeling of Flow Field 101
decreasing space is pushed back into the proximal part of the stomach, forming a
jet-like flow. It is also known as retropulsive jet. Retropulsive jet has been reported
by several authors (Carlson et al. 1966; Keinke et al. 1984; Issa et al. 1994; Pallotta
et al. 1998; Pal et al. 2004). Retropulsive flow results in an increase in the flow
velocity. The interaction of retropulsive jet with a relatively stagnant fluid inside the
stomach develops a high shear field at the boundaries of the jet. The food particu-
lates carried along with the retropulsive jet are subjected to this high shear. Issa
et al. (1994) visualized the movement of gastric content using a high-speed echo
planner magnetic resonance imaging (MRI) technique. They observed the retro-
pulsive motion with an increase in activities in the distal part of the stomach.
Pallotta et al. (1998) studied the retropulsive flow with the contraction pattern in
the antral region and in the closing and opening of the pyloric valve. They discussed
the role of retropulsive jet in mixing and transport of gastric contents. Pal et al.
(2004) observed retropulsive flow in their simulation model of the stomach. They
also reported the contribution of this motion in the mixing of the gastric content.
Pal et al. (2004) modeled the flow and mixing in human stomach using a two-
dimensional numerical model. They assumed that two-dimensional models have
the same quantitative behavior as that of three-dimensional models in axisymmetric
geometries. They demonstrated two important flow phenomenon taking place
inside the stomach and were able to show the retropulsive jet also reported by
many researchers earlier. Their simulation showed an increase in flow velocity up
to 7.5 mm/s. The retropulsive jet disperses the particles along the longitudinal axis
of the stomach. They also observed a second flow pattern (flow eddies) between any
two ACWs. These eddies are important for lifting the food particles from the
stomach wall and transporting them toward the center of the stomach. Although
retropulsive jets have been reported by many researchers, eddies between the
contraction waves were reported for the first time by this research group. Neverthe-
less, the human stomach is not axisymmetric in nature, especially considering the
effect of gravitational force on the lower part of the stomach. Therefore, it is
necessary to model the stomach as a three-dimensional geometry.
Further, the flow inside a human stomach is peristaltic in nature, driven by
deforming boundaries of the stomach muscles. In order to solve the peristaltic flow
in a complex system, such as a human stomach, one has to start with the peristaltic
flow in simpler geometries. There are several publications dealing with analytical as
well as numerical solutions of peristaltic motion in tubular geometry. Burns and
Parkes (1967) expanded the stream function as a Fourier series and used wall
boundary conditions to evaluate the coefficients of the Fourier series. They were
able to obtain a reasonable estimate of the flux through the tube under peristaltic
deformation of the walls as described by sinusoidal waves. Shapiro et al. (1969)
studied the flow properties of viscous incompressible fluid in a tube with deforming
walls. They solved the flow equations to obtain the parabolic velocity profile along
the cross section of the tube under the deforming wall according to (5.1):
u 1 dP 2
¼ 1 h r2 (5.1)
c 4mc dx
where u is the fluid velocity in the wave frame of reference, dP=dx is the
pressure gradient in x direction, and r is the radial position from the axis of
the tube. Zien and Ostrach (1970) analyzed peristaltic flow in two-dimensional
geometry due to the traveling wave motion of the confining walls. They solved
the stream function (c) in the form of an expansion with power of (a/l). Li
(1970) applied the analysis carried out by Zien and Ostrach (1970) for an
axisymmetric geometry, which represents many of the biological systems in a
more realistic way. He obtained the mean values and distribution of fluid
velocity and pressure in terms of amplitude, frequency, and wavelength of the
sinusoidal wave. Shen (1976) reduced the three-dimensional Navier–Stokes
equations to a series of linear boundary value problems with Laplace and
biharmonic operators.
Food particulates upon breakdown either dissolve in the gastric juice (polysac-
charides) or disperse as a suspension. This contributes to change in viscosity of the
gastric fluid. Many food fluids are non-Newtonian in nature with flow character-
istics that depend upon the shear rate. Many researchers have reported the viscosity
values of the stomach and small intestine digesta, but mostly at one shear rate value
as observed by Dikeman and Fahey (2006). Additionally, the flow dynamics of the
gastric fluid depends upon its viscosity. Overall, viscosity of the gastric fluid is
influenced by the composition of the food material and the secretions through
stomach walls after food intake. Moreover, in solid foods the viscosity changes as
the food particles disintegrate.
5.3 Procedures in Modeling
5.3.1 Stomach Geometry
We used the geometrical dimensions of a human stomach as defined by Pal et al.

(2004). Stomach volume was assumed to be 500 mL. The grids were constructed
with the help of GAMBIT, a three-dimensional mesh creation software. The
stomach was constructed with a series of circular rings in varying diameters
across the length. The rings were used as a base frame and joined together to form
the three-dimensional shape of the stomach. To save on computational time, the
stomach geometry was split in half along the plane of symmetry (Fig. 5.2b),
whereupon only half the stomach was used to compute the flow field. Hexagonal
grids were generated using the Hex–Cooper algorithm. A total of 281,602 cells
were used to construct the grids of stomach geometry. The centerline of the
stomach geometry was represented by a third-order polynomial equation given
by (5.2):
y ¼ 0:0067x3 0:0624x2 þ 0:3436x 5:8323 ðunits in mmÞ (5.2)

a b
Non-deforming wall
Deforming wall
Plane of Symmetry
Fig. 5.2 (a) Schematic diagram of circular rings constructing stomach geometry; (b) boundary
conditions used in FLUENT simulation
5.3.2 Deformation of Stomach Using Dynamic Meshing
The grids of the stomach geometry were deformed with every time step to represent
a real contracting stomach. An ACW started from near the middle of the stomach
(14.4 cm from the pyloric valve) and traveled toward the pylorus at a speed of
2.4 mm/s. The lifetime of an ACW lasted for 60 s. Any two ACWs were separated
by a 20 s time gap, making its frequency 3/min. The shape of an ACW was taken as
the sixth power of the sinusoidal function. The amplitude of an ACW varied along
its travel length from its origin to the pylorus. The occlusion diameter ratio was
defined as the ratio of diameter of the contraction ring (e) to the original diameter of
the section before contraction (D). The occlusion diameter ratio (e/D) decreased
linearly from 1.0 to 0.6 during the first 17.5 s and then remained constant at 0.6 for
the next 16.0 s. The occlusion diameter ratio decreased linearly from 0.6 to 0.1 for
23.5 s and then increased to 1.0 at the end of the wave’s travel. The stomach
geometry repeated itself after a 20 s time interval.
An imaginary line was assumed along the center of the stomach, connecting the
center of all the circular rings used to construct the stomach geometry. Deformation
in the stomach grid was performed by moving the nodes of the grid closer
(contraction) or farther (expansion) from the imaginary centerline of the stomach
(Fig. 5.3). Radial distance of the nodes (vertices of the computational elements)
from the imaginary centerline was increased or decreased to create the desired
volumetric deformations. The amount of displacement of a given node (Dri) at a
given time was the function of its distance from the centerline of the stomach (xi)
and the center of the ACW (xw) (Fig. 5.4a; (5.3) and (5.4)).
" 6 #
e p d jxw xi j
Dri ¼ b 1 sin if j xw x i j d (5.3)
D 2 d
Dri ¼ 0 if jxw xi j > d (5.4)

Fig. 5.3 Schematic diagram

showing movement of a node
during deformation
Imaginary centerline of the stomach
Location of a node at time t
} Dr
r
Node location at time t + Dt
Stomach wall
Dr2 Dr5
Dr3 Drw Dr4
Antral Contraction Wave

(ACW)
x1 x2 x3 xw x4 x5 x6
Imaginary centerline of the stomach
Fig. 5.4 (a) Schematic diagram describing sinusoidal deformation procedure along stomach wall;
(b) grids of stomach geometry with ACW modified by flow time
where Dri is the displacement of the ith node along the radial distance (m), b is the
maximum depth (amplitude) of the contraction wave at a given time (m), De is
the occlusion diameter ratio, d is half of the width of the contraction wave (m), xi
is the projection of the ith node on the imaginary stomach centerline (m), and xw is
the position of the center of the contraction wave (m).
At any given time all ACWs were perpendicular to the imaginary centerline.
Grids were denser near the lower (deformation) region and coarser in the upper
region of the stomach (Fig. 5.4b). The grid size was selected such that the
stomach wall was smooth at the peak of the ACWs. A relatively larger grid
size resulted in a non-smooth surface near the ACWs and a smaller grid size
increased the computational time. The grid size for this study was taken as
2.0 mm.
Fluid flow in a system is governed by three basic equations: the continuity,

conservation of momentum, and conservation of energy equations. Since this model
did not involve any heat transfer, only the first two equations were considered:
Continuity : r:U ¼ 0 (5.5)
@U 1
Momentum : þ ðU:DÞU ¼ Dp þ uf r2 U þ g (5.6)
dt rf
where U is the velocity component (m/s), p is the pressure (Pa), vf is the kinematics
viscosity of the fluid, and rf is the density of the fluid (kg/m3).
The above flow equations were solved with the help of a CFD solver (FLUENT
6.3.26). There were a number of assumptions made to simplify the model. The
gastric fluid was assumed to behave like a Newtonian fluid at a very low Reynolds
number. The stomach did not empty the gastric content for the duration of the
simulation. Physical properties of the gastric fluid remained unchanged with time.
The chemical reactions between the gastric fluid and food were not accounted for in
the model. The temperature of the system remained constant at 37 C.
5.4 Results and Discussions
5.4.1 Validation of a Modeled Flow Field inside a Circular Tube
A flow field inside a tubular geometry was simulated using FLUENT. Dimensions
of the tube were taken from that defined by Shapiro et al. (1969). Water was used as
the fluid inside the tube. The boundary wall of the tube was deformed as a
sinusoidal wave (Fig. 5.5). Flow in a tube is axisymmetric in nature; therefore, it
was sufficient to solve the flow field in a two-dimensional plane with a width equal
to the radius of the tube. The computational grid was created using GAMBIT.
c
b
a h
Line 1
Line 2
Y
X
Fig. 5.5 Peristaltic flow in a circular tube

0.025
Fluent Modeled
0.02
Shapiro et al (1969)
0.015
Axial Velocity (m/s)
0.01
Along line 2
0.005
0
Along line 1
–0.005
–0.01
–0.015
–0.02
–0.025
0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014
Axial Distance (m)
Fig. 5.6 Comparison between velocity profiles obtained by FLUENT simulation and analytical
model by Shapiro et al. (1969)
A user-defined function was written in C language to deform the nodes of the

computational grid given by (5.7):
2p
h ¼ a þ b sin ðX ctÞ (5.7)
l
Velocity profiles obtained in the simulation were compared to the analytical

equation given by Shapiro et al. (1969) along two lines, as shown in Fig. 5.6. There
was an excellent agreement between the predicted and the analytical results. This
confirmed the capability of this CFD model to handle peristaltic flow with deform-
ing walls. A correlation between the two results showed slopes of 0.996 (R2 ¼ 1)
and 0.998 (R2 ¼ 0.99). This validation supports the use of FLUENT software in
handling peristaltic types of flow under deforming walls.
5.4.2 Pressure Validation
The pressure field simulated by the stomach model was compared to the data
available in literature. For gastric juice viscosity 0.1 (kg/m.s) and density 1,000 kg/
m3, the pressure fields simulated by this stomach model and the model from Pal et al.
(2004) are shown in Fig. 5.7. The pressure contours in the antral region are
comparable and the maximum pressure is on the order of 0.1 mmHg. However,
the pressure between the most distal and the middle ACWs did not agree in two
cases. The Pal model showed uniform pressure in this region, whereas our model
showed that pressure in the region between the ACWs decreased in steps. The flow
a b
p (mmHg)
0.10
0.09
0.07
0.06
0.04
0.03
0.01
0
Fig. 5.7 (a) Pressure field predicted by Pal et al. (2004); (b) pressure field obtained from our
model (density ¼ 1,000 kg/m3, viscosity ¼ 1,000 cP)
inside the stomach was a pressure-driven flow caused by deforming walls and
decreasing volumes between the ACWs. Thus, the volume between two consecu-
tive ACWs is constantly decreasing, and therefore, the pressure between two
consecutive ACWs always increases with time until the volume completely col-
lapses in the pylorus.
5.4.3 Flow Field Inside the Stomach
Simulated unsteady flow patterns are shown in Fig. 5.9 for fluid density 1,000 kg/m3,
viscosity 1 cP, and ACW lifetime (TL) of 60 s; these settings were used for most of
the results unless otherwise specified. The lifetime of an ACW was defined as the
time interval between its creation and end.
The velocity contours shown in Fig. 5.9 represent the flow field beginning at
time N.TL, where N is an integer. Most of the flow activity takes place in the distal
part of the stomach as compared to the proximal part where flow was minimal. As
the ACW travels closer to the pyloric valve, the fluid trapped between the contrac-
tion wave and pyloric region is forced behind the enclosure, creating a retropulsive
jet. Figure 5.8 shows the volume between the two consecutive ACWs and the cross-
sectional area of an ACW as a function of time. The area marked by a circle in
Fig. 5.8 corresponds to the time of high intensity of the retropulsive jets. A
combination of high rate of volume change (steeper slope) and the small cross-
sectional area results in high-intensity retropulsive jets. The jet length increases as
the ACW moves closer to the pyloric valve. The maximum fluid velocity is
achieved just before the ACW collapses completely near the pyloric valve
(Fig. 5.9). The maximum fluid velocity observed was 30 mm/s. Minimum fluid
velocity is observed when only two ACWs are present and a third one has just
started to grow (Fig. 5.9a, f).
90 0.0045
Cross sectional area of the ACW(m2)

80 0.004
Volume between two ACWs (ml)

Volume
Area
70 0.0035
60 0.003
50 0.0025
40 0.002
30 0.0015
20 0.001
10 0.0005
0 0
0 10 20 30 40 50 60
Time (s)
Fig. 5.8 Graph showing change in volume between two successive ACWs and change in cross-
sectional area of an ACW with time
a b c
N.TL sec N.TL + 4 sec N.TL + 8 sec
d e f
N.TL + 12 sec N.TL + 16 sec N.TL + 20 sec
Fig. 5.9 Velocity contours in stomach as ACWs move toward the pylorus
There was an additional flow pattern besides the retropulsive jet, wherein flow
recirculation was observed between two ACWs. This recirculation was mainly due
to the wall motion of the contraction waves and was assisted by shearing flow of the
retrograde jet. The maximum velocity in these eddying flow structures was found to
2.08e–02
1.97e–02
1.87e–02
1.76e–02
1.66e–02 m/s
1.56e–02
1.45e–02
1.35e–02
1.25e–02
1.14e–02
1.04e–02
9.34e–03 Retrograde Jet
8.31e–03
7.27e–03
6.23e–03
5.19e–03
4.15e–03
3.11e–03
2.08e–03 Y
1.04e–03 Z X
3.93e–07 Flow recirculation
Fig. 5.10 Velocity vectors (m/s) of flow in a human stomach at TL(N + 1/3) s
be on the order of 7.5 mm/s. The role of these flow recirculation patterns and
retropulsive jets in mixing gastric content has been discussed previously by Pal
et al. (2004). The recirculating flow scrapes any food particles adhering to the
stomach walls and pushes them toward the retropulsive jet region.
In Fig. 5.10 it can be observed that the retrograde jet is the main flow event
taking place inside the stomach. Also, the flow direction below the retrograde jet
and between two contraction waves is toward the pyloric valve. The flows in this
region and the retrograde jet are in opposite directions from each other. The
retrograde jet carries the gastric content away from the lower stomach, whereas
the flow recirculation moves part of the gastric content back toward the lower
stomach region.
5.4.4 The Effect of Viscosity of Gastric Fluid
At the start of the digestion process the gastric fluid is less viscous (close to water).
The viscosity of the gastric fluid gradually increases as the food particulates
disintegrate and dissolve into the gastric juices. The viscosity plays a significant
role in flow dissipation of gastric flow as well as in the drag force on the food
particulates. The effect of viscosity on the gastric flow was observed in this study at
five levels: 1, 5, 10, 20, and 50 cP.
The length of the retropulsive jet in this study is defined as the distance from the
origin of the jet to the point where jet velocity drops to 90% of its original velocity.
It can be observed that the length of the retropulsive jet decreases as the gastric juice
becomes increasingly viscous (Figs. 5.11 and 5.12). The energy loss due to viscous
dissipation increases with the viscosity of the gastric juice and the jet length
becomes shorter. The role of retropulsive jets is mainly for shearing and mixing
of the food particulates. Therefore as viscosity increases, mixing of the food
particulates is expected to decrease.
0.025 Effect of viscosity on retropulsive jet
m = 1 cp
0.02 m = 5 cp
m = 10 cp
Velocity magnitude (m/s)
m = 20 cp
0.015 m = 50 cp
Centerline of the stomach

0.01
0.005
0
0 2 4 6 8 10 12 14
Distance from pyloric valve along the centerline of the stomach (cm)
Fig. 5.11 Velocity profiles along centerline of stomach at time TL(N + 1/3) s. r ¼ 1,000 kg/m3
Effect of viscosity on the retropulsive jet length

12
10
y = 9.6099x–0.424
R2 = 0.9761
8
Jet length (mm)
0
0 10 20 30 40 50 60
Viscosity (cP)
Fig. 5.12 Retropulsive jet length decreases as viscosity of gastric fluid increases. r ¼ 1,000 kg/m3
5.4.5 The Effect of Density of Gastric Fluid
The effect of density of the gastric fluid was studied at three levels: 1,000, 1,050,
and 1,100 kg/m3. Small differences were noticed near the regions of the contraction
waves. Beyond that there was no notable difference in the flow field due to change
in the density of gastric fluid (Fig. 5.13). The maximum velocity of the retropulsive
jet remained the same for all three densities. Thus, it may be concluded that the
density of the gastric fluid does not have any significant effect on the flow field;
however, it is likely to play a major role in the flow dynamics of the food
particulates. The density difference between the gastric fluid and the food material
determines the flow path of a food particulate and the frequency at which the particu-
late is exposed to the antral grinding. A food particulate that is heavier than the gastric
fluid is likely to settle at the bottom of the stomach, whereas a lighter food particulate
would float along the surface of the gastric fluid.
0.02
r = 1000 kg/m3
r = 1050 kg/m3
r = 1100 kg/m3
0.015
0.01
0.005
0
0 2 4 6 8 10 12 14
Distance from the pyloric valve along the centerline of the stomach (cm)
Fig. 5.13 Velocity profiles along centerline of stomach at time t + 10 s for gastric fluid viscosity
of 1 cP
5.4.6 The Effect of ACW Speed
Three levels of wave speed (mild, normal, and high) were selected to study their
effect on the flow field inside the model stomach. The speed of contraction waves
was changed by altering their lifetime. The total lifetime (TL) of an ACW for a
normal speed was 60 s, whereas for mild and high speeds the lifetimes were 66
and 54 s, respectively
The velocity profiles along the centerline of the stomach were compared to
assess any changes due to the speed of the ACWs. Figure 5.14 shows velocity
magnitude along the stomach centerline when the most distal (leftmost) ACW
was 7.2 cm away from the pyloric valve. The velocity of the gastric fluid
increased with the speed of ACW. The maximum velocity magnitude increased
by 6.45% (19.7–20.97 mm/s) at a 10% increase in ACW speed (2.4–2.68 mm/s).
The length of the retropulsive jet also increased from 10.52 to 11.71 mm (11.3%
increment). The flow pattern before the start and near the end of the jet remained
similar. The flow recirculation near the ACW also intensified as the ACW
traveled faster.
Increasing the wave speed (decreasing TL) increases the fluid velocity
(Figs. 5.14 and 5.15) and the length of the retropulsive jet increases with the
wave speed. There was an increase in the size of eddies due to flow recirculation.
It was also observed that the relationship between maximum velocity and lifetime
of an ACW was linear for the range considered in this study (Fig. 5.16). The
retropulsive jets in the stomach are generated by continuously decreasing the
Effect of ACW speed

0.025
0.02
High Speed ACW

Normal ACW
Low speed ACW
0.015
0.01
0.005
0
0 2 4 6 8 10 12 14
Distance from the pyloric valve along the stomach centerline (cm)
Fig. 5.14 Effect of ACW speed on flow profile along centerline of stomach at time TL(N + 1/3) s.
Density ¼ 1,000 kg/m3, viscosity ¼ 1 cP, high-intensity ACW
2.00e-02 scale representing velocity magnitude 2.00e-02 2.00e-02

1.90e-02 1.90e-02 1.90e-02
1.80e-02 1.80e-02 1.80e-02
1.70e-02 Low speed ACWs 1.70e-02 Normal speed ACWs 1.70e-02
1.60e-02 TL = 66 s 1.60e-02 1.60e-02
1.50e-02 1.50e-02 TL = 60 s 1.50e-02 High speed ACWs
1.40e-02 1.40e-02 1.40e-02
1.30e-02 1.30e-02 1.30e-02 TL = 54 s
1.20e-02 1.20e-02 1.20e-02
1.10e-02 1.10e-02 1.10e-02
1.00e-02 1.00e-02 1.00e-02
9.00e-03 9.00e-03 9.00e-03
8.00e-03 8.00e-03 8.00e-03
7.00e-03 7.00e-03 7.00e-03
6.00e-03 6.00e-03 6.00e-03
5.00e-03 5.00e-03 5.00e-03
4.00e-03 4.00e-03 4.00e-03
3.00e-03 3.00e-03 3.00e-03
2.00e-03 2.00e-03 2.00e-03
1.00e-03 1.00e-03 1.00e-03
0.00e+00 0.00e+00 0.00e+00
Fig. 5.15 Effect of speed of ACW on flow profile inside stomach at time TL(N + 1/3) s. Density
of gastric juice ¼ 1,000 kg/m3, viscosity ¼ 1 cP
0.025
Maximum velocity, Vmax (m/s)

0.023
y = – 0.000476x + 0.048176
0.021
0.019
0.017
0.015
48 54 60 66 72
Life time of an ACW, TL (s)
Fig. 5.16 Maximum velocity magnitude in stomach at time TL(N + 1/3) s, as a function of speed
of ACW. Density of gastric juice ¼ 1,000 kg/m3, viscosity ¼ 1 cP
volume between two ACWs. The rate at which the volume between two ACWs
collapses is directly proportional to the speed of the ACWs. Therefore, a higher
ACW speed means a faster collapsing volume between two ACWs, which results in
stronger retropulsive jets.
5.4.7 The Effect of Depth of Contraction
The initial phase of gastric motility consists of mild contraction waves increasing
slowly to high-intensity contraction waves over time. The mild waves represent the
initial quiescence phase when the food is entering or has just entered the stomach.
After the quiescence period the high-intensity contraction waves emerge, during
which time most of the mixing and disintegration activities take place (Coupe et al.
1991). Two contraction profiles of ACWs were used to study the effect of the
contraction depth on the flow field. The change in nondimensional occlusion
diameter (e/D) with time was varied, as described by (5.8) (high-intensity waves)
and (5.9) (mild waves).
Occlusion diameter ratio profile for high-intensity waves (Pal et al. 2004):
8
>
> 0:5 þ 0:5ð17:5Þ=17:5 0:0 t 17:5
e <
0:5 17:5 t 33:5
¼ (5.8)
D >
> 0:1 þ 0:4ð57 tÞ=23:5 33:5 t 57:0
:
0:1 þ 0:9ðt 57Þ=3:0 57:0 t 60:
Occlusion diameter ratio profile for mild waves (included in this study):
8
>
> 0:75 þ 0:25ð17:5 tÞ=17:5 0:0 t 17:5
e <
0:75 17:5 t 33:5
¼ (5.9)
D >
> 0:35 þ 0:4ð57 tÞ=23:5 33:5 t 57:0
:
0:35 þ 0:65ðt 57Þ=3:0 57:0 t 60:0
The maximum velocity of retropulsive jet was higher for the high-intensity
waves compared to the mild waves (Figs. 5.17 and 5.18). This was due to the
bigger occlusion diameter (larger opening) of the mild waves. For a given flow rate
the velocity was inversely proportional to the cross-sectional area of the flow path.
Moreover, in the case of mild waves, the rate at which the volume between any two
neighboring ACWs decreased was less compared to the high-intensity waves. The
flow rate of the gastric fluid was directly related to the rate of change in volume
0.02
velocity magnitude (m/s)
0.015
Mild waves
High intensity waves
0.01
0.005
0
0 2 4 6 8 10 12 14
Distance from the pyloric valve (cm)
Fig. 5.17 Velocity profiles along centerline of stomach for mild and high-intensity waves at
TL(N + 1/3) s
a b
2.00e-02 2.00e-02
1.90e-02 1.90e-02
1.80e-02 1.80e-02
1.70e-02 1.70e-02
1.60e-02 1.60e-02
1.50e-02 1.50e-02
1.40e-02 1.40e-02
1.30e-02 1.30e-02
1.20e-02 1.20e-02
1.10e-02 1.10e-02
1.00e-02 1.00e-02
9.00e-03 9.00e-03
8.00e-03 8.00e-03
7.00e-03 7.00e-03
6.00e-03 6.00e-03
5.00e-03 5.00e-03
4.00e-03 4.00e-03
3.00e-03 Y 3.00e-03
2.00e-03 2.00e-03
1.00e-03 Z X 1.00e-03
0.00e+00 0.00e+00
Fig. 5.18 Velocity contour for (a) mild and (b) high-intensity ACWs at time TL(N + 1/3) s
between two ACWs. Therefore, the combination of larger occlusion diameter and
less flow rate resulted in a drop in the retropulsive jet velocity. The maximum
velocity observed for mild ACWs was 5.9 mm/s, which is 70% less than for the
high-intensity waves. A lower velocity of retropulsive jet (for mild ACWs) results
in poor mixing and shearing, and the retropulsive jet may not be able to carry the
larger food particulates for long distances, resulting in poor mixing. Further, the
eddy flow near the ACWs becomes milder with contraction depth.
5.5 Conclusions
This study was focused on modeling the flow field and the disintegration kinetics of
the food particles inside the human stomach. The effect of the viscosity of gastric
juice, density of the juice, speed, and intensity of ACWs on the flow field was
studied. Fluid flow inside a human stomach driven by peristaltic deformation of the
stomach walls was simulated using FLUENT software, while dynamics meshing
was used to model the contracting shape of the stomach. The flow field was solved
assuming that the gastric juice follows a laminar flow rather than a non-Newtonian
flow at a very low Reynolds number, as found in the stomach. Retropulsive and
recirculatory flow structures were observed inside the simulated flow field of the
stomach. The size of the retropulsive jet is a function of the viscosity of the gastric
juice and characteristics of the ACWs. The length of the retropulsive jet decreased
with increasing viscosity or decreasing speed/depth of the ACWs. Recirculatory
flow regions were observed near the ACWs.
5.6 Suggestions for Future Work
This model has a number of limitations that can be pursued as future modifications.
This research was limited by information available in medical literature on the
human stomach, but with advancements in measurement techniques and methods,
the model can be modified in the future to predict the flow field and disintegration
kinetics more realistically. It was assumed that the flow inside the stomach is
Newtonian in nature, but there are many reports suggesting that the gastric content
behaves like a non-Newtonian fluid. The non-Newtonian properties of the gastric
content can be measured as a function of food properties (size, chemical composi-
tions, etc.) and the fluid’s concentration. A physical model of the human stomach
can be used for validation purposes, since one of the major limitations of this study
was lack of validation methods. In the future the flow field could be validated with
this physical model, more closely representing a stomach with peristaltic flow. This
physical model should also be capable of simulating food disintegration.
Furthermore, the current model does not consider the gastric emptying during
the digestion/disintegration process. The size, shape, and flow field of the stomach
change as the gastric contents are emptied and the stomach becomes increasingly
smaller. This model assumes that the depth and speed of the ACWs remain the same
throughout the digestion process. Four digestion cycle stages, wherein the proper-
ties of the ACWs change, can be added to the model. However, at the current
computational power available with computers, it could take an extremely long
time to complete such a simulation, but with advancements in computational speed
or application of parallel computing simulation, time may become more reasonable.
References
Burns JC, Parkes T (1967) Peristaltic motion. J Fluid Mech 29:731–743

Camilleri M, Prather CM (1993) Gastric motor physiology and motor disorders. In: Feldmann M,
Scharschmidt BF, Sleisenger MH (eds) Sleisenger and Fordtran’s gastrointestinal and liver
disease, 6th edn. W.B. Saunders, Philadelphia, pp 572–86
Carlson HC, Code CF, Nelson RA (1966) Motor action of the canine gastroduodenal junction; a
cineradigraphic, pressure and electric study. Am J Dig Dis 11:155–72
Code CF, Carlson HC (1968) Motor activity of the stomach. In: Code CF (ed) Handbook
of physiology, alimentary canal, vol IV. American Physiological Society, Washington, DC,
pp 1903–1916
Coupe AJ, Davis SS, Wilding IR (1991) Variation in gastrointestinal transit of pharmaceutical
dosage forms in healthy subjects. Pharm Res 8:360–364
Dikeman CL, Fahey GC (2006) Viscosity as related to dietary fiber: a review. Crit Rev Food Sci
Nutr 46(8):649–63
Forte JC (1996) Gastric function. In: Greger R, Windhorst U (eds) Comprehensive human
physiology: from cellular mechanism to integration. Springer-Verlag, Berlin, p 1253
Horowitz M, Dent J, Fraser R, Sun W, Hebbard G (1994) Role and integration of mechanisms
controlling gastric emptying. Dig Dis Sci 39(12):7S–13S
Indireshkumar K, Brasseur JG, Faas H (2000) Relative contributions of pressure pump and
peristaltic pump to gastric emptying. Am J Physiol Gastrointest Liver Physiol 278:G604–16
Issa B, Freeman A, Boulby P, Wright J, Gowland P, Bowtell R, Spiller R, Mansfield P (1994)
Gastric motility by tagged EPI. Magma, magnetic resonance materials in physics. Biol Med
2(3):295–298
Keinke O, Schemann M, Ehrlein HJ (1984) Mechanical factors regulating gastric emptying of
viscous nutrient meals in dogs. Q J Exp Physiol 69:781–95
Kelly KA (1980) Review: Gastric emptying of liquids and solids: roles of proximal and distal
stomach. Am J Physiol 239:G71–G76
Li C (1970) Peristaltic transport in circular cylindrical tubes. J Biomech 3(5):513–523
Macagno EO, Christensen J (1981) Fluid mechanics of gastrointestinal flow. In: Johnson LR (ed)
Physiology of the gastrointestinal tract. Raven, New York, pp 335–58
Mayer EA (1994) The physiology of gastric storage and emptying. In: Johnson L (ed) Physiology
of the gastrointestinal tract, 3rd edn. Raven, New York, pp 929–76
Meyer JH (1991) The physiology of gastric motility and gastric emptying. In: Yamada T (ed)
Textbook of gastroenterology. J.B. Lippincott, Philadelphia, pp 137–57
Pal A, Indireshkumar K, Schwizer W, Abrahamsson B, Fried M, Brasseur JG (2004) Gastric
flow and mixing studied using computer simulation. Proc Roy Soc B: Biol Sci 271
(1557):2587–2594
Pallotta N, Corazziari E, Scopinaro F, Bonino R, Schillaci O, Vignoni A, Mangano M, Torsoli A
(1998) Noninvasive estimate of bile flux through the gallbladder in humans. Am J Gastro-
enterol 93:1877–85
Schubert ML, Makhlouf GM (1992) Neural, hormonal, and paracrine regulation of gastrin and acid
secretion. Yale J Biol Med 65(6):553–560
Shapiro AH, Jaffrin MY, Weinberg SL (1969) Peristaltic pumping with long wavelengths at low
Reynolds number. J Fluid Mech 37:799–825
Shen MC (1976) Asymptotic theory for peristaltic transport in a tube of arbitrary cross section.
Phys Fluids 19(2):213–218
Soll AH, Berglindh T (1994) Receptors regulating acid secretory function. In: Johnson LR, Alpers
DH, Christensen J, Jacobson ED, Walsh JH (eds) Physiology of the gastrointestinal tract,
3rd edn. Raven, New York, pp 1139–1158
Szurszewski JH (1981) Electrical basis for gastrointestinal motility. In: Johnson LR (ed) Physio-
logy of the gastrointestinal tract. Raven, New York, pp 1435–1466
Zien TF, Ostrach S (1970) A long wave approximation to peristaltic motion. J Biomech
3(1):63–75
.
Chapter 6
State of the Art in Immobilized/Encapsulated
Cell Technology in Fermentation Processes
Viktor A. Nedović, Verica Manojlović, Branko Bugarski, and Ronnie Willaert
6.1 Introduction
The process of sugar conversion from wort or malt into alcohol, carbon dioxide, and
other components catabolized by yeast enzymes is called the alcohol fermentation
process. In beverage production, it is of great importance to achieve a particular
balance between different secondary metabolites. High productivity is another
demand of the beverage industry. Immobilization of cells provides high cell
densities leading to higher volumetric productivities, and as a consequence, reduces
essential bioreactor sizes (decreased capital costs) and shortens residence times.
Immobilized cell technology (ICT) coupled with continuous mode of fermentation
offers additional benefits, like ease of biomass separation and recovery, simplifica-
tion of process design, lower risk of microbial contamination of the pitching yeast
population, greater efficiency in utilization of carbohydrates, and better use of
equipment and potential savings. However, continuous fermentation processes
have not been commercially successful due to many practical problems, such as
increased risk of contamination not only during fermentation but also during
storage of wort in supplementary holding tanks, which are usually required for
batches upstream and downstream fermentation processes; in addition, there are
variations in beverage flavor and poor understanding of the fermentation kinetics
V.A. Nedović (*)

Department of Food Technology and Biochemistry, University of Belgrade, Nemanjina 6,
P.O. Box 127, 11081 Belgrade-Zemun, Serbia
e-mail: vnedovic@agrif.bg.ac.rs
V. Manojlović and B. Bugarski
Department of Chemical Engineering, University of Belgrade, Karnegijeva 4, 11000 Belgrade,
Serbia
e-mail: vmanojlovic@tmf.bg.ac.rs; branko@tmf.bg.ac.rs
R. Willaert
Structural Biology Brussels, Vrije Universiteit Brussel, Flanders Institute for Biotechnology,
Pleinlaan 2, B-1050 Brussels, Belgium
e-mail: Ronnie.Willaert@vub.ac.be
120 V.A. Nedović et al.
under continuous conditions. Over the last 30 years, ICT for alcoholic beverage
production has been extensively investigated and some systems have already
reached commercial exploitation. Intensification of a particular fermentation pro-
cess using ICT can generally be industrialized if the acquired new characteristics
result in a more economic system and the new technology can be readily scaled up.
ICT processes have been designed for different stages in the beer fermentation
process: wort acidification, primary fermentation, and bioflavoring during second-
ary fermentation; these fermentation processes are used in the production of
alcohol-free or low-alcohol beers (Brányik et al. 2005; Nedovic et al. 2005a;
Willaert and Nedovic 2006), as well as wine (Divies and Cachon 2005) and cider
(Nedovic et al. 2000; Durieux et al. 2005). The most challenging complex applica-
tion in fermentation processes is the combined main (ethanol fermentation) and
secondary fermentation (maturation) processes.
Traditional beer fermentation technology uses freely suspended yeast cells to
ferment wort in a non-stirred batch reactor. The traditional primary fermentation for
lager beer takes approximately 7 days with a subsequent secondary fermentation
(maturation) of several weeks. The resulting beer has a well-balanced flavor profile.
Nowadays, large breweries use a selected specific yeast strain and elevated tem-
peratures to accelerate production. This enables the production of finished lager
beer in 12–15 days. ICT is able to produce lager beer in a much shorter time period
(usually 1–3 days). A major difficulty is to achieve the correct balance of sensory
compounds to create an acceptable flavor profile in such a short time frame. ICT for
beer production can only be introduced successfully on an industrial scale if the
flavor profile can be controlled and fine-tuned.
Cider and wine production also involves complex processes that imply transfor-
mation of apple juice in the case of cider, or grape juice in the case of wine, by
activity of both yeast and lactic acid bacteria (LAB) to accomplish alcoholic and
malolactic fermentations (MLFs). The traditional process consists of natural fer-
mentation via autochthon yeasts and bacteria associated with the fruit or the cellar
equipment. This natural process is very unpredictable in terms of desirable flavor
compounds formation. The development of starter cultures enabled the use of
selected strains and the control of cider production to achieve high and uniform
quality, through several successive steps: pretreatment, alcoholic fermentation of
sugars into ethanol proceeded by yeast strains, and malolactic fermentation (MLF),
that is, bacterial conversion of L-malic into L-lactic acid and carbon dioxide (needed
to reduce acidity). Spontaneous MLF of cider begins within a few hours if the
temperature of the juice rises above 10 C. This process is usually very slow. It
requires 2–3 weeks to accomplish the main fermentation and several months for the
maturation. There is a risk of spontaneous fermentation by indigenous microbial
flora and it is difficult to control the flavor formation. The initiation of MLF appears
to be the main limiting factor in cider and wine production. MLF can occur several
weeks after alcoholic fermentation but there is no guarantee it will occur, which
is an unfavorable milieu for growth of microorganisms (ethanol > 10%;
pH < 3.0–3.5; temperature < 15 C). ICT offers a new alternative to better control
of the microbiology that defines the final product. In addition, this new technology
6 State of the Art in Immobilized/Encapsulated Cell Technology 121
significantly reduces consumption of time, facilitating MLF simultaneously with

alcoholic fermentation or at the end of this process.
Key parameters of this technology are the selection of carrier materials and the
method of immobilization together with the bioreactor design. Determination of
these parameters is directed by operational conditions such as temperature, pH,
substrate composition, and fluid dynamics, wherein special attention should be paid
to mass transfer properties since limited nutrient supply can result in changes in
yeast metabolism, leading to inadequate flavor of the final product.
6.2 Carrier Selection and Design
Cell immobilization can be classified into four categories based on the mechanism
of cell localization and the nature of support material: (i) attachment to the support
surface, which can be spontaneous or induced by linking agents; (ii) entrapment
within a porous matrix; (iii) containment behind or within a barrier; and (iv) self-
aggregation, naturally or artificially induced. Various supports and immobilization
techniques have been proposed and tested for application in brewing and wine- and
cider-making. Those that fulfill the following prerequisites are preferable:
l High surface-to-volume ratio of the immobilization support to achieve high cell
loading capacity
l Simple procedure and non-harsh conditions under way during immobilization
l Mechanical stability (compression, abrasion) and chemical stability of the
immobilization support
l Sterilization capability and regeneration of the immobilization support
l Cost-effectiveness of the support and immobilization process
l Suitability for conventional reactor systems
l Acceptance of immobilization support by consumers and avoidance of negative
effects on final product (e.g., off-flavor formations)
l Retention of immobilized cell viability
l Avoidance of negative effects of cell immobilization on biological and meta-
bolic activity of immobilized cells
l Easy separation of carriers with immobilized cells from media
l Wide choice of yeast
l Compounds approved for food applications
Table 6.1 Summarizes most of the carrier materials and bioreactors used in
fermentation processes for alcoholic beverages.
6.2.1 Immobilization on Solid Carrier Surfaces
Cell immobilization by adsorption to a support material is a very popular method,

because it is simple, easy to carry out, cheap, and fast. Microorganisms adsorb
Table 6.1 Carrier materials and reactor types for selected fermentation processes using immobi-
lized cells
Carrier material Reactor type Type of Product Reference
fermentation
Apple pieces Fixed-bed AF Wine Kourkoutas et al. 2001, 2002
g-Alumina Fixed-bed AF Wine Bakoyianis et al. 1997; Loukatos et al.
2000
Ca-alginate beads Fixed-bed AF Beer Ryder and Masschelein 1985; White
and Portno 1979; Onaka et al.
1985; Ryder and Masschelein 1985;
Ca-alginate beads Gas-lift AF Beer Nedovic et al. 1993, 1996, 2004, 2005a
Ca-alginate beads AF Sparkling Fumi et al. 1987; Fumi et al. 1988
wine
Ca-alginate beads Fixed-bed AF Wine Ferraro et al. 2000
Ca-alginate beads Fixed-bed AF and Cider Simon et al. 1996
MLF
Ca-alginate beads Fixed-bed MLF Cider Cabranes et al. 1998
Ca-alginate beads Fixed-bed AF and Cider Nedovic et al. 2000
MLF
Ca-alginate beads Shaken flasks MLF Cider Herrero et al. 2001
Ca-alginate beads Fixed-bed Maturation Beer Shindo et al. 1994
Ca-alginate beads Gas-lift AF Beer Smogrovicová et al. 1997;
Smogrovicová and D€omény 1999
Ca-alginate beads Shaken flasks MLF Wine Kosseva et al. 1998
Ca-alginate beads Shaken flasks MLF Wine Kosseva and Kennedy 2004
k-Carrageenan Gas-lift AF Beer Mensour et al. 1996, Mensour et al.
beads 1997; Decamps et al. 2004
Ceramic beads Fixed-bed AF Beer Inoue 1995
Corncobs Gas-lift AF Beer Brányik et al. 2006
Chitosan Shaken flasks MLF Wine Kosseva et al. 1998
Chitosan Fluidized-bed AF Beer Unemoto et al. 1998; Maeba et al. 2000
DEAE-cellulose Fixed-bed AF Beer Kronl€of et al. 1989; Andersen et al. 1999
DEAE-cellulose Fixed-bed Maturation Beer Pajunen and Gr€onqvist 1994
DEAE-cellulose Fixed-bed Acidified Pittner et al. 1993
wort
DEAE-cellulose Fixed-bed MLF Wine Maicas, Pardo, and Ferrer 2001
DEAE-cellulose Fixed-bed Alcohol- Collin et al. 1991; Lommi 1990
free
beer
Delignified Fixed-bed AF Wine Bardi and Koutinas 1994; Iconomou
cellulosic et al. 1996; Iconomopoulou
material et al. 2003
Delignified Fixed-bed MLF Wine Agouridis et al. 2005
cellulosic
material
Gluten pellets Fixed-bed AF Beer Bardi et al. 1997
Gluten pellets Gas-lift AF Beer Manojlovic et al. 2008
(external-
loop)
Gluten pellets Fixed-bed AF Wine Bardi et al. 1996; Iconomopoulou
et al. 2002
Gluten pellets Fixed-bed and AF Wine Sipsas et al. 2009
MFBT
Kieselguhr Fixed-bed AF Beer Narziss and Hellich 1971; Moll et al.
(diatomaceous 1973; Virkaj€arvi and Pohjala 2000
earth)
Kissiris Fixed-bed AF Wine Bakoyianis et al. 1992
“in the bottle” Maturation Beer Lemonnier and Duteurtre 1989;
(continued)
Table 6.1 (continued)

Carrier material Reactor type Type of Product Reference
fermentation
Microfiltration
membranes
Microfiltration Membrane MLF Cider Lovitt et al. 2006
ceramic reactor
membranes
Orange pieces Fixed-bed AF Wine Plessas et al. 2007
Pear pieces Fixed-bed AF Wine Mallios et al. 2004
PVA beads Gas-lift AF Beer Smogrovicová et al. 2001
PVA beads Fixed-bed Maturation Beer Smogrovicová et al. 2001
PVA beads Champagne AF Champagne Martynenko et al. 2004
bottles
PVA/alginate MFBT AF Beer Manojlovic et al. 2007
beads
PVA Lentikats® Gas-lift AF Beer Smogrovicová et al. 2001;
Bezbradica et al. 2007
Polyvinyl chloride Fixed-bed AF Beer Moll et al. 1973
granules
Porous glass beads Fixed-bed AF Beer Virkaj€arvi and Kr€onlof 1998;
Virkaj€arvi and Pohjala 2000
Porous glass beads Fixed-bed Maturation Beer Linko et al. 1993; Aivasidis 1996
Porous glass beads Fixed-bed Alcohol- Aivasidis et al. 1991
free
beer
Quince pieces Fixed-bed AF Wine Kourkoutas et al. 2003
Raisin berries Fixed-bed AF Wine Tsakiris et al. 2004a, 2004b
Self-aggregation Stirred-tank AF Beer Coutts 1957; Linko et al. 1997
using super- reactors
flocculent
yeast
Self-aggregation Erlenmeyer AF Wine Peinado et al. 2006; Peinado
using yeast flasks et al. 2005
biocapsules
Silicon carbide Monolith AF Beer Van De Winkel et al. 1993;
rods reactor Andries et al. 1996
Silicon carbide Monolith Alcohol- Van De Winkel et al. 1991
rods reactor free
beer
Spent grains Gas-lift AF Beer Brányik et al. 2002, 2004
Spent grains Fixed-bed AF Beer Kopsahelis et al. 2007
Spent grains Fixed-bed AF Wine Mallouchos et al. 2007
Sponge-like Fixed-bed AF and Cider Scott and O’Reilly 1996
material MLF
Stainless-steel Gas-lift AF Beer Verbelen et al. 2006
fiber cloth
Stainless-steel Fluidized-bed AF Beer Cross and Mavituna 1987
wire spheres
Wood chips Fixed-bed AF Beer Pajunen et al. 2001
(Aspen)
Wood chips Fixed-bed AF Beer Linko et al. 1997; Kronl€of and
(Beech) Virkaj€arvi 1999
Watermelon Fixed-bed AF Wine Veeranjaneya Reddy et al. 2008
pieces
PVA polyvinyl alcohol, AF alcoholic fermentation, MLF malolactic fermentation, MFBT multi-
stage fixed-bed tower
spontaneously on a wide variety of organic and inorganic support materials.

Binding of cells occurs through interactions such as Van der Waals forces, ionic
bonds, hydrogen bridges, or covalent interactions. Microbial cells exhibit a dipolar
character and behave as cations or anions depending on the cell type and environ-
mental conditions, such as pH of the solution. The thickness of cell film usually
ranges from one layer of cells to 1 mm or more. The strength with which the cells
are bonded to the carriers as well as the depth of the biofilm varies from one system
to another. Cell detachment and relocation readily occur, followed by establishment
of equilibrium between adsorbed and freely suspended cells. Various rigid organic
and inorganic support materials have been used in alcohol fermentation processes.
Inorganic materials are cheap and abundant. Among the inorganic types, porous
glass beads have been used successfully for primary beer fermentation (Tata et al.
1999) and beer maturation (Yamauchi et al. 1995), kissiris (a porous volcanic
mineral found in Greece, similar to granite, containing 70% SiO2,13% Al2O3, and
other inorganic oxides) for wine (Kana et al. 1989), ethanol production (Bakoyianis
et al. 1992), g-alumina for wine-making (Kana et al. 1989; Loukatos et al. 2000),
and stainless-steel wire spheres for ethanol production (Bekers et al. 1999). Various
organic materials are suitable for immobilization in beverage production, such as
diethylaminoethyl (DEAE) cellulose, delignified cellulosic material, wood, saw-
dust, delignified sawdust, gluten pellets, and spent grains, aimed at applications in
packed-bed reactors. A micrograph of yeast cells immobilized on wood chips is
shown in Fig. 6.1. Solid materials like glass and cellulose have been treated with
polycations, chitosan, or other chemicals to obtain preformed carriers with
enhanced adsorption ability (Norton and D’Amore 1994). In recent years, special
attention has been paid to usage of fruit pieces, since they are of food-grade purity,
and are easily accepted by consumers. Apple (Kourkoutas et al. 2001, 2002), quince
(Kourkoutas et al. 2003), pear (Mallios et al. 2004), raisin berries (Tsakiris et al.
2004a, 2004b), grape skin (Mallouchos et al. 2002), orange (Plessas et al. 2007),
Fig. 6.1 Scanning electron microscope (SEM) photo of yeast cells immobilized on wood chips
and watermelon (Veeranjaneya Reddy et al. 2008) have been used so far as support
materials for cells involved in fermentation processes.
6.2.2 Entrapment Within Porous Matrix
Entrapment involves containment of living cells within a network, which permits

the diffusion of substrates and products, thereby making possible the growth and
maintenance of active cells. Natural polysaccharides (e.g., alginate, pectate, carra-
geenan, chitosan, agar, polygalacturonic acid), synthetic polymers (polyvinyl chlo-
ride, polyvinyl alcohol (PVA) lens-shaped Lentikats, polyacrylamide), and proteins
(gelatin, collagen) can be gelled into hydrophilic matrices under mild conditions,
thus allowing cell entrapment with minimal loss of viability. Very high biomass
loadings can be achieved, since gel systems are characterized by very high poros-
ities (95–98%) and the cells are protected from fluid shear. Cell growth in the
porous matrix depends on diffusion limitations imposed by the porosity of the
matrix and the available interstitial space, which decreases in time with cell
propagation and the accumulation of biomass. Available literature shows that the
effective oxygen penetration rate varies in the range 0.08–010 mm in carrageenan
beads (Huang et al. 1990) and 0.1–0.15 mm in alginate beads (Ogbonna et al. 1991).
Cells that are growing can cause stress expressed in volumetric deformation and
also partial disintegration of the hydrogel. It opens a new space for cell growth
inside the matrix. This particular part of the process causes the mechanical trans-
formation of the network. There are only a few reports on the relaxation effects of
hydrogel caused by cell growth (a recent one is from Pajic-Lijakovic et al. 2007).
Gels are mostly used in the form of spherical beads with diameters ranging from
about 0.3 to 5 mm. Smaller particles show better mass transfer properties of
nutrients and metabolic products. Moreover, reduction in bead size lowers the
shear forces and may increase their long-term stability. However, small beads
have larger surface-to-volume ratio compared to big particles and therefore can
be more easily harmed by swelling or by exposure to oppositely charged ions
(Strand et al. 2002). In addition, smaller spheres are more fragile to internal stresses
caused by cell proliferation and expansion of cell colonies. Numerous techniques
for bead production have been developed up to now and are available to achieve the
desired size of capsules (Nedovic and Willaert 2004; Prusse et al. 2008). The use of
synthetic hydrogels may allow design of a carrier/matrix with preferred character-
istics. LentiKats® (specially designed particles made from PVA) have a specific
lenticular shape (Fig. 6.2) due to which they combine the advantages of small (good
diffusion properties) and large (easy retention and removal) beads. Production of
LentiKats® particles is based on the usage of the specially designed LentiKats®
Printer in lab and industrial scales. LentiKats® particles with immobilized yeast
have been successfully used for beer fermentation performed in a gas-lift bioreactor
(Bezbradica et al. 2007) and in cider production (Durieux et al. 2002). The design of
synthetic materials that could balance the opposite demands of high open-pore
Fig. 6.2 LentiKats®

hydrogel particle based on
polyvinyl alcohol (PVA) 3 – 4 mm
200 – 400 µm
structure, and at the same time, provide good protection to cells against washout is a
challenge for researchers involved in polymer science. One such attempt was the
design of a synthetic double-layer hydrogel, where the core was made of hydro-
xyethylcellulose cryogel and then coated with a layer of poly(ethylene oxide)
(Manojlovic et al. 2009).
A disadvantage of gels is the limited mechanical stability under conditions of rapid
cell growth, excessive CO2 production, or prolonged exposure to phosphates during
the maturation process. Several methods have been proposed for reinforcement of gel
structures. For example, alginate gel can be strengthened by reaction with polyethy-
leneimine, gluteraldehyde cross-linking, addition of silica, genepin, and PVA, or by
partial drying of the gel (Willaert and Baron 1996). The major drawback in these
systems is mass transfer limitation. However, understanding of mass transfer phenom-
ena within entrapment matrices may allow one to simultaneously provide different
conditions at the carrier surface and in the interior, which could be attractive for co-
immobilization of different cell types performing consecutive processes. For example,
gels with varying degrees of anisotropy, with respect to polymer concentration, can be
formed by controlling the kinetics of the gel formation. Simply by adjusting the
concentration of alginate and the cross-linking ions, the distribution of the polymer
in the gel can be controlled; alginate beads with a capsular structure have been made
without adding polycations or any other non-gelling polymer (Thu et al. 2000).
Another way is to create an external layer of another polymer around the hydrogel
core. Double-layer beads solve the problem of escaping of cells out of beads when the
system also contains (besides immobilized cells) free ones. However, microencapsu-
lation is generally too expensive to be used in the beverage industry.
Porous preformed supports can be inoculated directly from the bulk medium. In
these systems, cells are not completely separated from the effluent, similarly as in the
adsorption method. Cell immobilization occurs by attachment to the internal sur-
faces, self-aggregation, and retention in dead-end pockets within the material (Baron
and Willaert 2004). Ideally, the colonized porous particles should retain some void
spaces for flow so that mass transport of substrates and products can be achieved by
both molecular diffusion and convection. Consequently, mass transport limitations
are less stringent under optimal conditions as compared to gel entrapment methods.
However, when high cell densities are reached, convection is no longer possible and
the particles behave as dense cell agglomerates with high diffusion limitations.
As compared to gel particles, preformed carriers provide better mechanical proper-

ties and higher resistances to compression and disintegration.
6.2.3 Cell Aggregation
Cell immobilization by self-aggregation is based on formation of cell clumps or

floccules, which can be naturally occurring as in the case of flocculent yeast strains,
or induced by addition of artificial flocculating agents or cross-linkers. It is the
simplest and the least expensive immobilization method. However, interactions
among cells are not easily controlled and cell aggregates are very sensitive to
conditions in fermentors, including pH, dissolved oxygen, and medium composi-
tion. The flocculation of Saccharomyces cerevisiae is determined by the adhesin
protein family. The adhesin proteins are encoded by the genes FLO1, FLO5, FLO9,
and FLO10 (Verstrepen et al. 2003a, 2004). These proteins are called flocculins
(Caro et al. 1997) because they promote cell–cell adhesion forming multicellular
clumps that settle out of solution. The structure, location, and activity of flocculins
are well described in a recent report from Van Mulders et al. (2009). The ethanol
productivity achieved by flocculated cells was almost double that of a freely
suspended yeast cell system (Xu et al. 2005). It was found that the floc size
distribution influenced the effectiveness of glucose uptake and ethanol production
(Ge et al. 2006). An interesting approach has been recently proposed by Peinado
and coworkers (2005, 2006): a filamentous fungus and a flor yeast under adequate
conditions form a cluster that looks like a hollow biocapsule; here mycelium creates
walls around the biocapsule and the yeast is entrapped in an inner space. The yeast
biocapsules were successfully used in must (wine) fermentation.
6.2.4 Containment Behind a Membrane Barrier
Cell immobilization behind or within a porous barrier includes systems with cells
contained in a compartment separated by a preformed membrane such as hollow
fiber and flat membrane modules. Micromembrane technology, like microencapsu-
lation, is generally too expensive to be used in beverage production. Moreover,
mass transfer limitations are relatively high (Lebeau et al. 1997), and membrane
biofouling caused by cell growth often occurs (Gryta 2002).
6.3 Bioreactor Design
Selecting the appropriate reactor type or configuration for an immobilized cell

system is related to a number of important factors, such as carrier design, supply
and removal of gases and solutes in the liquid phase, as well as removal of excess
biomass formed, investment and operation costs, operation mode, maintaining

sterile conditions, heat and mass transfer rates, and others. In fermentors, immobi-
lized cells can be either mixed with suspended carriers or localized on carrier
particles/surfaces, which are then fixed or in movement. The operation mode of
immobilized cell reactors can be batch, fed-batch, or continuous. Continuous
operation eliminates the unproductive time in batch and fed-batch processes asso-
ciated with filling, emptying, cleaning and disinfection/sterilization, and start-up
phase of the fermentation. Therefore, it provides a higher productivity compared to
the “old-fashion” batch method of processing. With respect to sterility, reactors that
can be directly inoculated with cells or cell-aggregates (e.g., membrane modules or
reactors packed with preformed porous carriers) are more desirable compared to
reactors that require transfer of a biocatalyst from the immobilization equipment to
the reactor (e.g., fermentors using cells entrapped in gel systems). The bioreactor
should be designed and the hydrodynamic conditions optimized to provide easy
access to nutrient media, optimum mass transfer from flowing media to the support
interior, controlled yeast growth, low shear experienced by cells, simple scale-up,
controlled oxygenation, complete attenuation and desired flavor profile, consistent
product quality, and low risk of contamination.
Most of the studies have been on packed-bed (fixed-bed) bioreactors. Packed-
bed bioreactors are characterized by a simple design consisting of a single column
packed with biocatalysts. The liquid flow is close to the plug flow regime and
causes low shear rates (Obradovic et al. 2004). The main disadvantages of packed-
bed fermentors are high mass transfer restrictions, accumulation of carbon diox-
ide, non-uniform temperature profiles, flow channeling, and stagnant zones.
Therefore, the first trials conducted on using a packed-bed configuration for
primary beer fermentation on an industrial scale gave unsatisfactory results with
respect to product quality. Afterwards, packed-bed reactors were selected for
production of alcohol-free or low-alcohol beers and for enhanced flavor matura-
tion using immobilized cells. In these applications, conditions are anaerobic and
yeast growth is limited. Immobilization of cells can be by adsorption (e.g., DEAE-
cellulose beads) or by a combination of adsorption and entrapment (e.g., porous
glass beads). These carrier materials need to be mechanically strong to withstand
the high pressures in packed-bed reactors. However, the use of mechanically week
materials (e.g., hydrogels) can be limited to lower bed heights and liquid flow
rates due to possible compression of beads. In order to eliminate some of the
drawbacks of the packed-bed configuration, a modification of a packed-bed
fermentor, that is, the “multistage fixed-bed tower” (MFBT), has been proposed
for beer production (Manojlovic et al. 2007) and wine-making (Sipsas et al. 2009).
It consists of a vertical cylindrical tank with five packed sections containing
freeze-dried immobilized cells. A relatively small (5000–10000 L) MFBT biore-
actor (Fig. 6.3) (Koutinas et al. 1997; Loukatos et al. 2000) was proposed for
industrialization of immobilized cells in wine-making. Handling of the support at
this scale could be performed without any problems, and cell immobilization
could be carried out in the bioreactor. The application of the MFBT bioreactor
on an industrial scale eliminates insufficient mass transfer and enables support
Fig. 6.3 Multistage fixed- Beverage

bed tower (MFBT) bioreactor
(Adapted from Sipsas et al.
2009)
Multi-stages
sustaining the
immobilized
biocatalyst
Feed
division, especially when mechanically unstable supports are used to minimize

high pressure effects, which may result in support destruction and reduction of
fermentation activity (Kourkoutas et al. 2009). Experiments concerning long-term
storage of the immobilized biocatalysts (Kourkoutas et al. 2003) are very
promising, since the preparation of new biocatalysts, emptying and filling of the
bioreactor, could be avoided when industrial production is halted. Taking into
consideration the above discussion of technical problems, the scale-up of the
proposed technology seems feasible.
Different approaches for the adaptation of bioreactors containing immobilized
cells were investigated in order to correct the final beverage quality. One was to
establish biocatalyst movement or circulation for the purpose of speeding up the
transfer of nutrients and metabolic products through the fermenting medium, as in
fluidized-bed, stirred-tank, and gas-lift bioreactors. In the fluidized-bed bioreactors,
particles with immobilized cells are fluidized in the liquid up-flow, while gas can be
optionally supplied. As a consequence of particle fluidization, moderate local
mixing is established, which provides better mass and heat distribution with more
uniform liquid flow throughout the reactor volume, as compared to packed-bed
reactors. It is difficult to maintain low-density particles in fluidization and to
prevent their washout. Particle movements and collisions in the fluidized state
result in moderate shear stresses and abrasion, creating a need for relatively
mechanically stable supports (Obradovic et al. 2004). The scaling up of fluidized-
bed bioreactors addresses problems due to the difficulties in controlling the bed
expansion and may encounter hydrodynamic problems. In stirred tank reactors high
aeration resulted in a less balanced aroma profile of the final product. Beers
produced in fluidized and stirred-tank fermentors had high concentrations of dia-
cetyl and low concentrations of higher alcohols and esters (Okabe et al. 1992;
Mensour et al. 1997).
Gas-lift reactors are especially attractive since they apply pneumatic agitation
with no mechanical devices. They are based on liquid circulation, which can be
effectively tuned to achieve an adequate flow regime and optimal external mass
transfer. This bioreactor concept was introduced in beer fermentation studies by a
Serbian group in 1993 (Nedovic et al. 1993). Internal loop configuration (Fig. 6.4)
has been investigated in lab- and pilot-scale production mainly for beer fermenta-
tion (Nedovic et al. 1993, 2004, 2005a; Mensour et al. 1997), while the external
loop design (Fig. 6.5) has been recently tested for alcoholic fermentation in lab-
scale beer production (Manojlovic et al. 2008). Efficient mixing and low shear rates
make gas-lift reactors suitable for all types of low-density immobilization materials
(Mensour et al. 1997; Obradovic et al. 2004).
The design of membrane reactors is relatively complex and expensive, mainly
due to the high cost of the membrane material. Membrane reactors provide simul-
taneous bioconversion and product separation. A special design of a multichannel
loop bioreactor has been developed by the Belgian company, Meura (Tournai), for
production of lager, ale, and acidified wort (Masschelein et al. 1994). Yeast cells are
Gas Flow Outlet
Bubble
Effluent
Internal
tube (riser Down-comer section
section)
Biocatalyst
Outer
thermal
jacket
Feed
Gas distributor
Gas Flow Inlet
Fig. 6.4 Gas-lift bioreactor internal loop configuration with an immobilized biocatalyst (Adopted
from Nedovic et al. 1993)
Gas Flow Inlet
Effluent Down-comer
section
Riser
section
Outer
thermal Immobilized
jacket biocatalyst
Bubble
Gas
Gas Flow Inlet distributor
Fig. 6.5 Gas-lift bioreactor external loop configuration with an immobilized biocatalyst (Adapted
from Manojlovic et al. 2008)
immobilized in porous sintered silicon carbide rods perforated with 19 or 37

channels for fluid flow. This immobilization method can be regarded as contain-
ment behind a preformed barrier, and as entrapment in a porous preformed support.
Continuous beer fermentation technology using yeast flocculation and cell recy-
cling has been successfully exploited over almost 40 years by Dominion Breweries
in New Zealand (Coutts 1957; Van de Winkel and De Vuyst 1997).
Selected yeasts entrapped in micro-filtration membranes have been developed
and used in wine production. On-market available “Millispark” cartridges (Millipore)
were used for secondary fermentation of sparkling wine in bottles (Lemonnier and
Duteurtre 1989). In dry wine production, a single-vessel membrane bioreactor was
found unsuitable for continuous fermentation, as high levels of unfermented sugars
were reported (Takaya et al. 2002). However, a double-vessel continuous mem-
brane configuration resulted in a sugar content lower than 4 g/L, which was
considered satisfactory for dry wine-making. Additionally, wine productivity was
28 times higher compared to traditional batch systems.
In a recent study, the approach of splitting cell propagation and cell maturation
was applied in a pilot-scale membrane bioreactor with ceramic membrane modules
to perform MLF of media containing ethanol (Lovitt et al. 2006). Herein, the
overall productivity of both process rate and longevity was successfully increased.
6.4 Impact of Immobilization on Flavor Formation
Although ICT offers a number of benefits, it has so far found limited application in
the fermentation industry. A major difficulty is to achieve the correct balance of
volatile compounds to create an acceptable flavor profile of alcoholic beverages.
Immobilized cells appear to have modified physiology compared to the physiology
of free cells. The nutrient uptake and synthesis patterns of metabolites such as fusel
alcohols, esters, and carbonyl compounds change upon immobilization; the follow-
ing paragraphs describe the impact of immobilization on flavor formation.
6.4.1 Influence of ICT on Higher Alcohol Production
Higher alcohols (also called “fusel alcohols”) are produced by yeast cells and
represent the major fraction of the volatile compounds. Higher alcohols can be
classified as aliphatic [n-propanol, isobutanol, 2-methyl butanol (or active amyl
alcohol), 3-methyl butanol (or isoamyl alcohol)], and aromatic (2-phenyl ethanol,
tyrosol, tryptophol). Aliphatic higher alcohols contribute to the “alcoholic” or
“solvent” aroma of a beverage, and produce a warm mouthfeel. The aromatic
alcohol 2-phenyl ethanol has a sweet scent and is a positive contribution to the
aroma, whereas the aroma of tyrosol and tryptophol are undesirable. Higher alco-
hols are synthesized by yeast during fermentation via the catabolic (Ehrlich) and
anabolic pathway (amino acid metabolism) (Ehrlich 1904).
Catabolism of the branched-chain amino acids (leucine, valine, and isoleucine),
aromatic amino acids (phenylalanine, tyrosine, and trytophan), and sulfur-contain-
ing amino acid (methionine) leads to the formation of fusel acids and fusel alcohols.
Firstly, the yeast cells use amino acids from the wort to produce the corresponding
a-keto acids via a transamination reaction. The excess oxoacids are subsequently
decarboxylated into aldehydes and further reduced (by alcohol dehydrogenase) to
higher alcohols. The simplified Ehrlich pathway is shown in Fig. 6.6. The genes
encoding each step of the process are quoted in a recent review by Hazelwood et al.
(2008).
In the anabolic pathway, the higher alcohols are synthesized from a-keto acids
during the synthesis of amino acids from the carbohydrate source. Both pathways
may take place during the same fermentation in the traditional batch process with a
switch from the degradative route to the biosynthetic route, occurring when the
amino acids in the substrate have been metabolized or missed. The pathway choice
depends on the individual higher alcohol and on the level of available amino acids.
The importance of the anabolic pathway increases during the later stage of a
conventional batch fermentation as wort amino acids are depleted, as well as in
cider production where the apple juice contains only small amounts of amino acids.
Conditions that promote yeast cell growth – such as high levels of nutrients
(amino acids, oxygen, lipids, zinc), increased temperature, and agitation – stimulate
Transamination amino acid

RCH(NH2)COOH
R’COCOOH
R’CH(NH2)COOH a-keto acid

RCOCOOH
Decarboxylation
CO2
fusel aldehyde
Reduction
RCHO Oxidation
NADH, H+ NAD+
NAD+ NADH, H+
fusel acid in fusel alcohol
Export
RCOOH RCH2OH
ATP
ADP
fusel acid out
Fig. 6.6 The Ehrlich pathway (Adapted from Hazelwood et al. 2008)
the production of higher alcohols (Landaud et al. 2001). On the other hand,
conditions that restrict yeast growth – such as lower temperature and higher
(CO2) pressure – reduce the extent of higher alcohol production (Renger et al.
1992).
In immobilized systems with enhanced or similar free amino nitrogen (FAN)
uptake levels, the formation of higher alcohols was higher or equal to batch systems
(Shen et al. 2003). A decrease of higher alcohol production in beer upon cell
immobilization, compared to free-cell fermentation, has been frequently reported
and nicely summarized by Willaert and Nedovic (2006). This decrease has been
attributed to the limited cellular growth in immobilized cell systems, leading to
poor nitrogen removal. Similarly, in the case of cider production, in a continuous
fermentation system with Saccharomyces bayanus co-immobilized with Oenococcus
oeni in alginate beads, production of fusel alcohols was several times lower
compared to synthesis during batch fermentation process with suspended cells
(Nedovic et al. 2000). The anabolic flux limitation of yeast cells in the pseudo-
stationary phase was proposed to justify the lower concentration of fusel alcohols. It
was also found that the behavior of cells adsorbed on the carrier surface was similar
to that of free cells, but significantly different from entrapped cells (Smogrovicová
and D€omeny 1999). New technologies have introduced some new inclusion car-
riers, with adjusted shape and size to overcome internal mass transfer restrictions,
which give similar higher alcohol concentrations, compared to a conventional pro-
cess (Nedovic et al. 2005b).
It has been demonstrated that mass (i.e., amino acids) transfer rates in the
fermenting medium or, in other words, the external mass transfer properties, also
influence higher alcohol synthesis. Thus, in fluidized-bed and gas-lift bioreactors the
rate of amino acid uptake increased with the superficial velocity of the fluid (Cop
et al. 1989; Masschelein et al. 1994; Nedovic et al. 1996; Aivasidis et al. 1991).
6.4.2 Ester Production in ICT Systems
Esters constitute a major group of desirable flavor compounds. Among the esters
formed, the most significant in fermented beverages are ethyl acetate (fruity,
solvent-like), isoamyl acetate (pear drops), isobutyl acetate (banana-like), ethyl
hexanoate (apple-like), and 2-phenyl acetate (honey, fruity, flowery). They are
formed by yeast during fermentation in a reaction between the alcohols, fatty
acids, co-enzyme A (CoASH), and an ester synthesizing enzyme. Actually, the
formation of esters occurs in two steps: (1) fatty acids that have undergone a
previous activation by CoASH form acyl-CoA and (2) alcohols become esterified
by reacting with acyl-CoA to the corresponding ester under the action of alcohol
acetyl transferase (Peddie 1990). Because ethanol is the dominant alcohol in fer-
menting beverages, ethyl acetate (produced from acetyl-CoA and ethanol) is the
dominant ester. It has been shown that the main factor controlling ester biosynthesis
is the expression level of the ATF1 gene, which encodes alcohol acetyl transferase I
(Lilly et al. 2000; Verstrepen et al. 2003b). ATF1 gene expression is repressed by
oxygen and unsaturated fatty acids (Fujii et al. 1997; Fujiwara et al. 1998). The ester
production rate is influenced by many factors, such as temperature, specific growth
rate, pitching rate, top pressure, oxygen availability, as well as fermenting medium
composition (Willaert and Nedovic 2006; Verbelen et al. 2009).
In some immobilized processes low ester concentrations are found, while in
others ester synthesis is increased upon cell immobilization. Low ester content is
related to the low cellular metabolic activities in these systems. In a study of
continuous fermentation in cider, concentration of isoamylacetate was two times
lower compared to concentration achieved in a control batch fermentation process
with suspended cells, as a result of isoamylalcohol availability (Nedovic et al.
2000). On the other hand, due to mass transfer limitations, oxygen concentration
in an immobilization matrix is low, causing reduced cellular growth, so that the
cellular acetyl-CoA pool is more available for ester synthesis instead of channeling
for fatty acid biosynthesis. Thus, the anaerobic conditions and the absence of
substantial levels of unsaturated fatty acids limit cell growth during production
and stimulate formation of acetate esters. For example, this occurred during the
production of alcohol-free beer in a packed-bed reactor with surface-attached cells
on DEAE-cellulose beads (Van Iersel et al. 1999). In another study, a 22% increase
in ester concentration upon cell immobilization on stainless-steel fiber cloth was
explained by a significant rise in the expression level of AFT1 in the immobilized
cells, leading to enhanced ester concentrations in the final fermented product (Shen
et al. 2003).
6.4.3 Carbonyl Compounds Production in ICT Systems
The most important carbonyl compounds formed in beverage fermentation are

acetaldehydes, diacetyl, and 2,3-pentanedione. Aldehydes, having very low flavor
thresholds, tend to be considered as off-flavors (e.g., acetaldehyde causes a green
leaf-like flavor in beer). As intermediates in the formation of ethanol and higher
alcohols from amino acids and sugar, the conditions favoring alcohol production
also generate the formation of small quantities of aldehydes. These may be excreted
but can be reabsorbed and reduced by yeast to the corresponding alcohol during the
later stages of fermentation (acetaldehyde is normally reduced to ethanol). The
most extensively studied carbonyl compound is diacetyl, which makes an important
contribution to the flavor of cider, red wine, beer, and some distilled products such
as whisky and rum. Although its presence may contribute to the correct flavor,
especially in cider and red wine, excessive production can lead to off-flavors,
particularly in the case of beer. Diacetyl and 2,3-pentanedione are side products
of amino acid synthesis in yeast. They are produced by the spontaneous oxidative
decarboxylation of the corresponding acetohydroxy acids, a-acetolactate, and
a-acetohydroxybutyrate, which are metabolite intermediates of the common bio-
synthetic pathways of valine and isoleucine. Acetohydroxy acids are firstly excreted
from the yeast cells to the surrounding medium where they are transformed
chemically into diacetyl and 2,3-pentanedione. The formation and subsequent
reassimilation of actohydroxy acids by the yeast and degradation of diacetyl are
shown schematically in Fig. 6.7. Due to the coupling of acetohydroxy acids with the
anabolic metabolism of yeast, they are produced only in the earlier phases of batch
fermentation. In the later stages of fermentation, actively metabolizing yeast cells
are able to reduce diacetyl and 2,3-pentanedione to acetoin and butane-2,3-dione,
and 2,3-butanediol, respectively. Hence, the balance between rate of formation and
rate of degradation determines the final concentrations of diacetyl and 2,3-penta-
nedione in beverages. In cider, in addition to the formation of vicinal diketones by
yeast, a part of diacetyl present is also produced by bacteria from pyruvic acid.
Thus, Oenococcus oeni is able to produce diacetyl directly by the activity of the
diacetyl synthetase without any excretion of precursors in the fermenting medium.
In the case of diacetyl content in beer, different phenomena upon immobilization
have been reported. In most cases, the production of diacetyl by immobilized cells
is much higher than for free cells. In addition, the production of vicinal diketones
can be controlled by the initial yeast cell concentration in Ca-alginate beads
(Shindo et al. 1994). This has been explained by an increased expression of the
acetohydroxy acid synthetase gene during the growth of the yeast cells in the carrier
(Shindo et al. 1994). In a recirculation bioreactor system with continuous sugar
feed, the concentration of 2,3-pentanedione was two to four times larger than the
diacetyl concentration due to a more intensive pentanedione pathway (Pajunen
et al. 2001). In an air-lift reactor with spent grains as the immobilization matrix,
the total diacetyl concentration largely varied depending on the operational condi-
tions and decreased with increasing aeration and temperature (Brányik et al. 2004).
Extract
Fermenting medium
Glucose
EMP
pathway
Pyruvate Acetohydroxy acid
syntethase
a-acetolactate a-acetolactate
SLOW
Valine CHEMICAL
OXIDATIVE
Diacetyl Diacetyl
diacetyl
Yeast reductase
cell Acetoin
butanediol
dehydrogenase
2,3-butanediol
Fig. 6.7 Schematic presentation showing diacetyl formation, reassimilation, and removal
(Adapted from Willaert and Nedovic 2006)
After a maturation period of 10 days at 4 C, the concentration of diacetyl was

reduced below its flavor threshold (Brányik et al. 2002). The concentrations of
vicinal diketones were around 20 times higher in a continuous fermentation of cider
with yeast immobilized in Lentikats at a sugar attenuation of 95% compared with
those obtained in a batch process (Durieux et al. 2002). In another study with
alginate beads used for cider production in the same continuous system, diacetyl
concentration was increased two times compared to concentration achieved in a
fermentation process with suspended cells (Nedovic et al. 2000). The larger con-
centration of diacetyl in the immobilized systems was explained by the diffusion
mass transfer effect that prevents transfer of diacetyl from the medium to the
immobilized yeast after chemical oxidative decarboxylation of a-acetolactate in
the medium. Addition of the missing enzyme a-acetolactate decarboxylase (com-
mercially available) to the wort may also lead to increased formation of diacetyl
(Hanneman 2002).
The drawback in using alginate is biomass leakage due to local overpressure in
the beads generated by carbon dioxide production. Nedovic and coworkers (2004,
2005a) suggested that free yeast issued from the continuous reactor could be used
for diacetyl uptake in a maturation tank. Optimization of the operational parameters
in a gas-lift bioreactor with alginate microbeads as yeast carriers can provide low
concentrations of diacetyl. One way to decrease the diacetyl level in a continuous
system is to prolong the residence time during fermentation and/or maturation

process (Andersen et al. 1999; Nedovic et al. 2000). Also the use of genetically
modified yeast enables reduction of the diacetyl level (Hammond 1995).
6.4.4 Secondary Fermentation Using ICT
The maturation of green beer is needed primarily to reduce the level of diacetyl (an
unwanted aroma compound in beer). This vicinal diketone has a very low threshold
(0.08–0.15 ppm) in beer (Wainwright 1973). The traditional maturation process
lasts for 3–4 weeks at a low temperature and low yeast concentration. However,
using ICT, this period could be reduced to 2 h. So far, two continuous maturation
systems have been implemented industrially. The first one is a packed-bed bioreac-
tor with DEAE-cellulose granules (later replaced by cheaper aspen wood chips)
used at Sinebrychoff Brewery (Finland); it has a capacity of 1 million hectoliters/
year (Yamauchi et al. 1995; Virkaj€arvi 2002). Another system was developed by
Alfa Laval and Schott Engineering (Mensour et al. 1997) based on porous
glass beads (Dillenhofer and Ronn 1996). The Alfa Laval system for secondary
fermentation of beer is shown in Fig. 6.8. This system has been implemented in
several breweries in Finland, Belgium, and Germany. The German company
Brau & Brunnen purchased and installed a 30,000 hL/year pilot-scale Alfa Laval
Main
fermentation
Green
beer
Maturation Beer chiller

Separator column
Yeast Heat exchanger
with
immobilized yeast
Yeast Accelerated thermal acetolactate
Cooling
removal conversion to diacetyl
Fig. 6.8 Process flow sheet for secondary fermentation of beer using the Alfa Laval system
maturation system in 1996 (Mensour et al. 1997). The same system was imple-
mented in a medium-sized German brewery as well (Sch€aff/Treuchtlingen) (Back
et al. 1998). The beers obtained overall yielded good analytical and sensorial
results.
6.4.5 Malolactic Fermentation in ICT Systems
MLF is a very important process in the maturation of alcoholic beverages, particu-

larly wines and ciders. The fermentation is catalyzed by a number of LAB that can
gain a competitive advantage by metabolizing malic acid to lactic acid. MLF is
generally recognized as an important manufacturing step: (1) it reduces the acidity
of wine or cider, (2) it stabilizes the product with respect to microbial spoilages
through the bacteriostatic effect of the lactic acid produced and consumption of
residual substrates, and (3) it contributes to the flavor complexity of wine/cider by
producing compounds such as acetaldehyde, acetic acid, ethyl acetate, ethyl lactate,
diacetyl, acetoin, and 2,3-butanediol. The MLF process, therefore, not only affects
the flavor of the beverage, but also stimulates growth or enhances resistance to the
extreme environment found in wines and ciders, for example, low pH and high
alcohol concentration. Whether MLF should be encouraged or discouraged in wine-
making depends on the quality of ripe grapes and the desired level of certain
flavorful by-products.
The growth of LAB, that is, O. oeni and Lactobacillis brevis, has been investi-
gated in this context. The initiation of MLF appears to be the main limiting factor in
cider and wine production. Several strategies have been suggested to sustain and
accomplish MLF in wine and cider: the use of enzymatic reactors, recombinant
yeast strains, and cell-recycle bioreactors (Durieux et al. 2005). The immobilization
of LAB for controlling MLF provides increased tolerance of malolactic bacteria
and acceleration of the MLF process. Different immobilization supports have been
used so far to immobilize O. oeni, such as calcium alginate, kappa-carrageenan,
cellulose sponge, and polyacrylamide (reviewed by Kourkoutas et al. 2009).
The immobilization of O. oeni enabled accomplishment of MLF simultaneously
with the alcoholic fermentation or at the end of this fermentation. The alcoholic
fermentation was preceded either by free (Cabranes et al. 1998) or co-immobilized
yeast cells in batch (Scott and O’Reilly 1996) and continuous (Nedovic et al. 2000)
bioreactor systems. Yeast growth and ethanol production were not affected by the
presence of immobilized O. oeni. The final organoleptic profile of cider depended
on the fermentation temperature and the carrier used.
Thus, in the case of alginate beads, malic acid was fully metabolized at 18 C,
while a residual amount of this acid was detected at 12 C (Cabranes et al. 1998).
Acetic acid produced at the end of fermentation at 18 C was doubled compared to
that obtained by free O. oeni. In another study, using Lentikats as an alternative to
alginate beads in a continuous system, the largest malic acid attenuation was
achieved at temperatures between 25 C and 30 C (Durieux et al. 2000). Surprisingly,
the malic acid conversion was possible even at very acidic pH (down to pH 2.3),
while with free cells the MLF did not occur below 3.9. A modification of the cell
physiology and the immobilized cell microenvironment, characterized by pH
gradient inside the matrix, was proposed to explain the improved performance of
O. oeni at acidic pH by respectively allowing generation of enough adenosine
triphosphate (ATP) to maintain cytoplasmic pH without any perturbation of the
MLF and by restoring favorable pH in the direct environment of the cells. In
continuous systems, deacidification levels can be easily adjusted as a function of
the residence time (Nedovic et al. 2000). In those systems, the production of soft or
dry cider is possible by controlling the feeding flow rates.
6.5 Conclusion
In this chapter, an overview of ICT applications in fermentation processes aimed at

alcoholic beverage production was presented. The objective was to analyze and
assess data on the impact of immobilization technologies on viable microbial cells
in the alcoholic and malolactic fermentation of beer, wine, and cider. ICT is well
established for flavor maturation and the production of alcohol-free and low-
alcohol beer on an industrial scale. In addition, several primary beer and wine
fermentation processes based on ICT have been developed on a pilot and an
industrial scale. However, the issues of mass transfer limitations and process
control still need to be resolved in order to obtain a beverage of consistent quality.
Selecting a suitable carrier and bioreactor system is a challenge and many issues
should be taken into account, such as maintenance of cell viability during produc-
tion, product quality, safety and stability during processing and storage, and
investment and operating costs. In particular, assessment of industrial feasibility
of the immobilization fermentation technology is mandatory for cost-effective,
large-scale applications.
References
Agouridis N, Bekatorou A, Nigam P, Kanellaki M (2005) Malolactic fermentation in wine with

Lactobacillus casei cells immobilized on delignified cellulosic material. J Agric Food Chem
53(7):2546–2551
Aivasidis A (1996) Another look at immobilized yeast systems. Cerevisia 21(1):27–32
Aivasidis A, Wandrey C, Eils HG, Katzke M (1991). Continuous fermentation of alcohol-free beer
with immobilized yeast cells in fluidized bed reactors. Proc. 23rd EBC Cong., pp 569–576
Andersen K, Bergin J, Ranta B, Viljava T (1999). New process for the continuous fermentation of
beer. Proc. 27th Eur. Brew. Conv. Cong. EBC, pp 771–778
Andries M, Van Beveren PC, Goffin O, Masschelein CA (1996) Design and application of an
immobilized loop bioreactor for continuous beer fermentation. In: Wijffels RH, Buitelaar RM,
Bucke C, Tramper J (eds) Immobilized cells: basics and applications. Elsevier, Amsterdam,
pp 672–678
Back W, Krottenthaler M, Braun T (1998) Investigations into continuous beer maturation.

Brauwelt International 3:222–226
Bakoyianis V, Kanellaki M, Kaliafas A, Koutinas AA (1992) Low temperature wine making by
immobilized cells on mineral kissiris. J Agric Food Chem 40:1293–1296
Bakoyianis V, Koutinas AA, Agelopoulos K, Kanellaki M (1997) Comparative study of kissiris,
g-alumina, and calcium alginate as supports of cells for batch and continuous wine-making at
low temperatures. J Agric Food Chem 45:4884–4888
Bardi EP, Bakoyianis V, Koutinas AA, Kanellaki M (1996) Room temperature and low
temperature wine making using yeast immobilized on gluten pellets. Process Biochem
31:425–430
Bardi EP, Koutinas AA (1994) Immobilization of yeast on delignified cellulosic material for room
temperature and low-temperature wine making. J Agric Food Chem 42:221–226
Bardi E, Koutinas AA, Kanellaki M (1997) Room and low temperature brewing with yeast
immobilized on gluten pellets. Process Biochem 32:691–696
Baron GV, Willaert RG (2004) Cell immobilisation in pre-formed porous matrices. In: Nedovic V,
Willaert R (eds) Fundamentals of cell immobilisation biotechnology. Springer, Dordrecht, The
Netherlands, pp 229–244
Bekers M, Ventina E, Karsakevich A, Vina A, Rapoport A, Upite D, Kaminska E, Linda R (1999)
Attachment of yeast to modified stainless steel wire spheres, growth of cells and ethanol
production. Process Biochem 35:523–530
Bezbradica D, Obradovic B, Leskosek-Cukalovic I, Bugarski B, Nedovic V (2007) Immobili-
zation of yeast cells in PVA particles for beer fermentation. Process Biochem 42
(9):1348–1351
Brányik T, Silva DP, Vicente AA, Lehnert R, Almeida e Silva JB, Dostálek P, Teixeira JA
(2006) Continuous immobilized yeast reactor system for complete beer fermentation
using spent grains and corncobs as carrier materials. J Ind Microbiol Biotechnol
33:1010–1018
Brányik T, Vicente AA, Cruz JMM, Teixeira JA (2002) Continuous primary beer fermentation
with brewing yeast immobilized on spent grains. J Inst Brew 108:410–415
Brányik T, Vicente AA, Cruz JMM, Teixeira JA (2004) Continuous primary fermentation of beer
with yeast immobilized on spent grains – the effect of operational conditions. J Am Soc Brew
Chem 62:29–34
Brányik T, Vicente AA, Dostálek P, Teixeira JA (2005) Continuous beer fermentation using
immobilized yeast cell bioreactor systems. Biotechnol Prog 21(3):653–663
Cabranes C, Moreno J, Mangas JJ (1998) Cider production with immobilized Leuconostoc oenos.
J Inst Brew 104:127–130
Caro LH, Tettelin H, Vossen JH, Ram AF, van den Ende H, Klis FM (1997) In silicio identification
of glycosyl-phosphatidylinositol-anchored plasma-membrane and cell wall proteins of
Saccharomyces cerevisiae. Yeast 13:1477–1489
Collin S, Montesinos M, Meersman E, Swinkels W, Dufour JP (1991) Yeast dehydrogenase
activities in relation to carbonyl compounds removal from wort and beer. Proc. Eur. Brew.
Conv. Cong., pp 409–416
Cop J, Dyon D, Iserentant D, Masschelein CA (1989). Reactor design optimization with a view to
the improvement of amino acid utilization and flavor development of calcium alginate
entrapped brewing yeast fermentations. Proc. 22nd EBC Cong., pp 315–322
Coutts MW (1957) A continuous process for the production of beer. UK Patent 872,391400
Cross PA, Mavituna F (1987) Yeast retention fermentors for beer production. Proc. 4th Eur. Cong.
Biotechnol. Amsterdam, pp 199–200
Decamps C, Norton S, Poncelet D, Neufeld RJ (2004) Continuous pilot plant-scale immobilization
of yeast in k-carrageenan gel beads. AlChE J 50:1599–1605
Dillenhofer W, Ronn D (1996) Secondary fermentation of beer with immobilized yeast. Brauwelt
International 14:344–346
Divies C, Cachon R (2005) Wine production by immobilized cell systems. In: Willaert R, Nedovic
V (eds) Applications of cell immobilisation biotechnology. Springer, Dordrecht, The Netherlands,
pp 285–293
Durieux A, Bodo E, Nedovic V, Simon JP (2002) Effect of yeast and Oenococcus oeni immobi-
lisation on the formation of flavour components for cider production. Proc. International
workshop bioencapsulation X: cell physiology and interactions of biomaterials and matrices.
Prague, Czech Republic, pp 54–57
Durieux A, Nikolay X, Simon JP (2000) Continuous malolactic fermentation by Oeconoccocus
oeni entrapped in Lentikats. Biotechnol Lett 22:1679–1684
Durieux A, Nicolay X, Simon J-P (2005) Application of immobilisation technology to cider
production: a review. In: Willaert R, Nedovic V (eds) Applications of cell immobilisation
biotechnology. Springer, Dordrecht, The Netherlands, pp 275–284
Ehrlich F (1904) Uber das nat€ urliche isomere des leucins. Berichte der Deutschen Chemisten
Gesellschaft 37:1809–1840
Ferraro L, Fatichenti F, Ciani M (2000) Pilot scale vinification process using immobilized Candida
stellata cells and Saccharomyces cerevisiae. Process Biochem 35:1125–1129
Fujii T, Kobayashi O, Yoshimoto H, Furukawa S, Tamai Y (1997) Effect of aeration and
unsaturated fatty acids on expression of Saccharomyces cerevisiae alcohol acetyltransferase
gene. Appl Environ Microbiol 63:910–915
Fujiwara D, Yoshimoto H, Sone H, Harashima S, Tamai Y (1998) Transcriptional co-regulation of
Saccharomyces cerevisiae alcohol acetyltransferase gene ATF1 and D-9 fatty acid desaturase
gene, OLE1 by unsaturated fatty acids. Yeast 14:711–721
Fumi MD, Trioli G, Colagrande O (1987) Immobilization of Saccharomyces cerevisiae in calcium
alginate for sparkling wine processes. Biotechnol Lett 9:339–342
Fumi MD, Trioli G, Colombi MG, Colagrande O (1988) Immobilization of Saccharomyces
cerevisiae in calcium alginate gel and its application to bottle-fermented sparkling wine
production. Am J Enol Viticult 39:267–272
Ge XM, Zhang L, Bai FW (2006) Impacts of yeast floc size distributions on their observed rates for
substrate uptake and product formation. Enzyme Microb Technol 39:289–295
Gryta M (2002) The assessment of microorganism growth in the membrane distillation system.
Desalination 142:79–88
Hammond JRM (1995) Genetically-modified brewing yeast for the 21st century. Progress to date.
Yeast 11:1613–1627
Hanneman W (2002) Reducing beer maturation time and retaining quality. Mast Brew Assoc Am
Techn Quart 39(3):149–155
Hazelwood LA, Daran J-M, van Maris AJA, Pronk JT, Dickinson JR (2008) The Ehrlich pathway
for fusel alcohol production: a century of research on Saccharomyces cerevisiae metabolism.
Appl Environ Microbiol 74(8):2259–2266
Herrero M, Laca A, Garcia LA, Dı́az M (2001) Controlled malolactic fermentation in cider using
Oenococcus oeni immobilized in alginate beads and comparison with free cell fermentation.
Enzyme Microb Technol 28:35–41
Huang J, Hooijmans CM, Briasco CA, Geraats SGM, Luyben KCAM, Thomas D, Barbotin JN
(1990) Effect of free-cell growth parameters on oxygen concentration profiles in gel-immobilized
recombinant Escherichia coli. Appl Microbiol Biotechnol 33:619–623
Iconomopoulou M, Kanellaki M, Soupioni M, Koutinas AA (2003) Effect of freeze-dried cells on
delignified cellulosic material in low-temperature wine making. Appl Biochem Biotechnol
104:23–36
Iconomopoulou M, Psarianos K, Kanellaki M, Koutinas AA (2002) Low temperature and ambient
temperature wine making using freeze-dried immobilized cells on gluten pellets. Process
Biochem 37:707–717
Iconomou L, Kanellaki M, Voliotis S, Agelopoulos K, Koutinas AA (1996) Continuous wine
making by delignified cellulosic materials supported biocatalyst. An attractive process for
industrial applications. Appl Biochem Biotechnol 60:303–313
Inoue T (1995) Development of a two-stage immobilized yeast fermentation system for continu-
ous beer brewing. Proc. Eur. Brew. Conv. Cong., pp 25–36
Kana K, Kanellaki M, Papadimitriou A, Psarianos C, Koutinas AA (1989) Immobilization of
Saccharomyces cerevisiae on g-alumina pellets and its ethanol production in glucose and raisin
extract fermentation. J Ferment Bioeng 68:213–215
Kopsahelis N, Kanellaki M, Bekatorou A (2007) Low temperature brewing using cells immobi-
lized on brewer’s spent grains. Food Chem 104(2):480–488
Kosseva M, Beschkov V, Kennedy JF, Lloyd LL (1998) Malolactic fermentation in Chardonnay
wine by immobilized Lactobacillus casei cells. Process Biochem 33:793–797
Kosseva MR, Kennedy JF (2004) Encapsulated lactic acid bacteria for control of malolactic
fermentation in wine. Artif Cells Blood Substit Immobil Biotechnol 32:55–65
Kourkoutas Y, Koutinas AA, Kanellaki M, Banat IM, Marchant R (2002) Continuous wine
fermentation using a psychrophilic yeast immobilized on apple cuts at different temperatures.
Food Microbiol 19:127–134
Kourkoutas Y, Komaitis M, Koutinas AA, Kaliafas A, Kanellaki M, Marchant R, Banat IM (2003)
Wine production using yeast immobilized on quince biocatalyst at temperatures between 30
and 0 C. Food Chem 82:353–360
Kourkoutas Y, Komaitis M, Koutinas AA, Kanellaki M (2001) Wine production using yeast
immobilized on apple pieces at low and room temperatures. J Agric Food Chem 49:1417–1425
Kourkoutas Y, Manojlovic V, Nedovic V (2009) Immobilisation of microbial cells for alcoholic
and malolactic fermentation of wine and cider. In: Zuidam N-J, Nedovic VA (eds) Encapsula-
tion technologies for food active ingredients and food processing. Springer, Dordrecht, The
Netherlands, pp 327–345
Koutinas AA, Bakoyianis V, Argiriou T, Kanellaki M, Voliotis S (1997) A qualitative outline to
industrialize alcohol production by catalytic multistage fixed bed tower (MFBT) bioreactor.
Appl Biochem Biotechnol 66:121–131
Kronl€of J, H€ark€onen T, Hartwall P, Home S, Linko M (1989) Main fermentation with immobilized
yeast. Proc. 22nd Eur. Brew. Conv., Zurich, pp 355–362
Kronl€of J, Virkaj€arvi I (1999). Primary fermentation with immobilized yeast. Proc. Eur. Brew.
Conv. Cong., pp 761–770
Landaud S, Latrille E, Corrieu G (2001) Top pressure and temperature control the fusel alcohol/
ester ratio through yeast growth in beer fermentation. J Inst Brew 107:107–117
Lebeau T, Jouenne T, Junter GA (1997) Simultaneous fermentation of glucose and xylose by pure
and mixed cultures of Saccharomyces cerevisiae and Candida shehatae immobilized in a two-
chambered bioreactor. Enzyme Microb Technol 21:265–272
Lemonnier J, Duteurtre B (1989) Un progress important pour le champagne et lens vins “methode
traditionnelle”. Rev Fr Cenol 121:15–26
Lilly M, Lambrechts MG, Pretorius IS (2000) Effect of increased yeast alcohol acetyltransferase
activity on flavor profiles of wine and distillates. Appl Environ Microbiol 66:744–753
Linko M, Suihko M-L, Kronl€ of J, Home S (1993) Use of brewer’s yeast expressing a-acetolactate
decarboxylase in conventional and immobilized fermentations. Mast Brew Assoc Am Techn
Quart 30:93–97
Linko M, Virkaj€arvi I, Pohjala N, Lindborg K, Kronl€ of J, Pajunen E (1997) Main fermentation
with immobilized yeast – a breakthrough? Proc. 26th Eur. Brew. Conv. Maastricht, pp 385–394
Lommi H (1990) Immobilized yeast for maturation and alcohol-free beer. Brew Dist Int 5:22–23
Loukatos P, Kiaris M, Ligas I, Bourgos G, Kanellaki M, Komaitis M, Koutinas AA (2000)
Continuous wine making by g-alumina-supported biocatalyst. Quality of the wine and
distillates. Appl Biochem Biotechnol 89:1–13
Lovitt R, Jung I, Jones M (2006) The performance of the membrane bioreactor for the malolactic
fermentation of media containing ethanol. Desalination 199:435–437
Maeba H, Unemoto S, Sato M, Shinotsuka K (2000) Primary fermentation with immobilized yeast
in porous chitosan beads. Pilot scale trial. Proc. 26th Conv. Inst. Brew. Aus. N.Z. Sec.
Singapore, pp 82–86
Maicas S, Pardo I, Ferrer S (2001) The potential of positively charged cellulose sponge for
malolactic fermentation of wine using Oenococcus oeni. Enzyme Microb Technol 28:415–419
Mallios P, Kourkoutas Y, Iconomopoulou M, Koutinas AA, Psarianos C, Marchant R, Banat IM
(2004) Low temperature wine-making using yeast immobilized on pear pieces. J Sci Food
Agric 84:1615–1623
Mallouchos A, Loukatos P, Bekatorou A, Koutinas A, Komaitis M (2007) Ambient and low
temperature wine-making by immobilized cells on brewer’s spent grains: Effect on volatile
composition. Food Chem 104(3):918–927
Mallouchos A, Reppa P, Aggelis G, Koutinas AA, Kanellaki M, Komaitis M (2002) Grape skins as
a natural support for yeast immobilization. Biotechnol Lett 24:1331–1335
Manojlovic V, Agouridis N, Kopsahelis N, Kanellaki M, Bugarski B, Nedovic V (2008) Brewing
by immobilized freeze dried cells in a novel gas flow bioreactor. Proc. 2008 Joint Central
Europ. Cong. Food, 6th Croat. Cong. Food Technol. Biotechnol. Nutrit. Cavtat, Croatia 2,
pp 327–334
Manojlovic V, Nedovic V, Bugarski B, Winkelhausen E, Velickova E, Petrov P, Ivan B, and
Tsvetanov C (2009). Immobilized yeast cells in double-layer hydrogel carriers for beer
production. Proc. COST Spring workshop on bioencapsulation, Luxembourg, p 112
Manojlovic V, Sipsas V, Agouridis N, Bugarski B, Leskosek-Cukalovic I, Kanellaki M, and
Nedovic V (2007). Beer fermentation by immobilized yeast in PVA/alginate beads using a
catalytic multistage fixed bed tower bioreactor. Proc. 5th Int. Cong. Food Technol. Thessaloniki,
Greece, pp 219–222
Martynenko NN, Gracheva IM, Sarishvili NG, Zubov AL, El’Registan GI, Lozinsky VI (2004)
Immobilization of champagne yeasts by inclusion into cryogels of polyvinyl alcohol: Means of
preventing cell release from the carrier matrix. Appl Biochem Microbiol 40:158–164
Masschelein CA, Ryder DS, Simon J-P (1994) Immobilized cell technology in beer production.
Crit Rev Biotechnol 14:155–177
Mensour N, Margaritis A, Briens CL, Pilkington H, Russell I (1996) Applications of immobilized
yeast cells in the brewing industry. In: Wijffels RH, Buitelaar RM, Bucke C, Tramper J (eds)
Immobilized cells: basics and applications. Elsevier, Amsterdam, pp 661–671
Mensour N, Margaritis A, Briens CL, Pilkington H, Russell I (1997) New developments in the
brewing industry using immobilised yeast cell bioreactor systems. J Inst Brew
103:363–370
Moll M, Durand G, Blachere H (1973) Continuous production of fermented liquids. French Patent
73/23397. US Patent 4009286
Narziss L, Hellich P (1971) Ein Beitrag zur wesentlichen Beschleunigung der G€arung und Reifung
des Bieres. Brauwelt 111:1491–1500
Nedovic V, Bezbradica D, Obradovic B, Leskosek-Cukalovic I, Bugarski B (2004). Primary beer
fermentation by PVA-immobilized brewing yeast in a gas-lift bioreactor. World Brew. Cong.
2004 CD Rom Proc. San Diego CA, pp O-63
Nedovic V, Cukalovic IL, Bezbradica D, Obradovic B, Bugarski B (2005b) New porous matrices
and procedures for yeast cell immobilisation for primary beer fermentation. Proc. 30th Eur.
Brew. Conv. Prague, pp 401–413
Nedovic V, Durieux A, Van Nederveide L, Rosseels P, Vandegans J, Plaisant AM, Simon J-P
(2000) Continuous cider fermentation with co-immobilized yeast and Leuconostoc oenos cells.
Enzyme Microb Technol 26:834–839
Nedovic V, Obradovic B, Vunjak-Novakovic G, Leskosek-Cukalovic I (1993) Kinetics of beer
fermentation with immobilized yeast cells in an internal-loop air-lift bioreactor. Chem Indus
47:168–172
Nedovic V, Vunjak-Novakovic G, Leskosek-Cukalovic I, Cutkovic M (1996) A study on consi-
derable accelerated fermentation of beer using an airlift bioreactor with calcium alginate
entrapped yeast cells. Proc 5th World Cong Chem Eng 2:474–479
Nedovic V, Willaert R (eds) (2004) Fundamentals of cell immobilisation biotechnology. Kluwer,
Dordrecht, The Netherlands
Nedovic V, Willaert R, Leskosek-Cukalovic I, Obradovic B, Bugarski B (2005b) Beer production

using immobilized cells. In: Nedovic V, Willaert R (eds) Applications of cell immobilisation
biotechnology. Springer, Dordrecht, The Netherlands, pp 259–273
Norton S, D’Amore T (1994) Physiological effects of yeast cell immobilization applications for
brewing. Enzyme Microb Technol 16:365–375
Obradovic B, Nedovic V, Bugarski B, Willaert RG, Vunjak-Novakovic G (2004) In: Nedovic V,
Willaert RG (eds) Fundamentals of cell immobilisation biotechnology, vol 8a, Focus on
biotechnology. Kluwer, Dordrecht, pp 411–436
Ogbonna JC, Matsumura M, Kataoka H (1991) Effective oxygenation of immobilized cells
through reduction in bead diameters: a review. Process Biochem 26:109–121
Okabe M, Katoh M, Furugoori F, Yoshida M, Mitsui S (1992) Growth and fermentation
characteristics of bottom brewer’s yeast under mechanical stirring. J Ferment Bioeng 73
(2):148–152
Onaka T, Nakanishi K, Inoue T, Kubo S (1985) Beer brewing with immobilized yeast. Nat Bio/
Technol 3:467–470
Pajic-Lijakovic I, Plavsic M, Nedovic V, Bugarski B (2007) Ca-alginate hydrogel mechanical
transformations – the influence of yeast cell growth dynamics. J Microencapsul 24(5):410–429
Pajunen E, Gr€onqvist A (1994) Immobilized yeast fermenters for continuous lager beer matura-
tion. Proc. 23rd Conv. Inst. Brew. Aus. N.Z. Sec., Sydney, pp 101–103
Pajunen E, Tapani K, Berg H, Ranta B, Bergin J, Lommi H, Viljava T (2001) Controlled beer
fermentation with continuous on-stage immobilized yeast reactor. Proc 28th EBC Cong
49:1–12
Peddie HAB (1990) Ester formation in brewery fermentations. J Inst Brew 96:327–331
Peinado RA, Moreno JJ, Maestre O, Mauricio JC (2005) Use of a novel immobilization yeast
system for winemaking. Biotechnol Lett 27:1421–1424
Peinado RA, Moreno JJ, Villalba JM, González-Reyes JA, Ortega JM, Mauricio JC (2006) Yeast
biocapsules: a new immobilization method and their applications. Enzyme Microb Technol
40:79–84
Pittner H, Back W, Swinkels W, Meersman E, Van Dieren B, Lomni H (1993) Continuous
production of acidified wort for alcohol-free-beer with immobilized lactic acid bacteria.
Proc. Eur. Brew. Conv. Cong., pp 323–329
Plessas S, Bekatorou A, Koutinas AA, Soupioni M, Banat IM, Marchant R (2007) Use of
Saccharomyces cerevisiae cells immobilized on orange peel as biocatalyst for alcoholic
fermentation. Bioresour Technol 98(4):860–865
Prusse U, Bilancetti L, Bučko M, Bugarski B, Bukowski J, Gemeiner P, Lewinska D, Manojlovic
V, Massart B, Nastruzzi C, Nedovic V, Poncelet D, Siebenhaar S, Tobler L, Tosi A, Vikartovská
A, Vorlop K-D (2008) Comparison of different technologies for alginate beads production.
Chem Pap 62(4):364–374
Renger RS, Vanhateren SH, Luyben K (1992) The formation of esters and higher alcohols during
brewery fermentation – the effect of carbon-dioxide pressure. J Inst Brew 98:509–513
Ryder DS, Masschelein CA (1985) The growth process of brewing yeast and the biotechnological
challenge. J Am Soc Brew Chem 43(2):66–75
Scott JA, O’Reilly AM (1996) Co-immobilization of selected yeast and bacteria for controlled
flavour development in an alcoholic cider beverage. Process Biochem 31(2):111–117
Shen H-Y, Moonjai N, Verstrepen KJ, Delvaux FR (2003) Impact of attachment immobilization
on yeast physiology and fermentation performance. J Am Soc Brew Chem 61(2):79–87
Shindo S, Sahara H, Koshino S (1994) Suppression of a-acetolactate formation in brewing with
immobilized yeast. J Inst Brew 100:69–72
Simon JP, Durieux A, Pinnel V, Garré V, Vandegans J, Rosseels P, Godan N, Plaisant AM,
Defroyennes J-P, Foroni G (1996) Organoleptic profiles of different ciders after continuous
fermentation (encapsulated living cells) versus batch fermentation (free cells). In: Wijffels RH,
Buitelaar RH, Bucke C, Tramper J (eds) Immobilized cells: basics and applications. Elsevier,
Amsterdam, pp 615–621
Sipsas V, Kolokythas G, Kourkoutas Y, Plessas S, Nedovic VA, Kanellaki M (2009) Comparative

study of batch and continuous multi-stage fixed-bed tower (MFBT) bioreactor during wine-
making using freeze-dried immobilized cells. J Food Eng 90:495–503
Smogrovicová D, D€omény Z (1999) Beer volatile by-product formation at different fermentation
temperature using immobilized yeasts. Process Biochem 34:785–794
Smogrovicová D, D€omény Z, Gemeiner P, Malovı́ková A, Sturdı́k E (1997) Reactors for the
continuous primary beer fermentation using immobilised yeast. Biotechnol Tech
11:261–264
Smogrovicová D, D€omeny Z (1999) Beer volatile by-product formation at different fermentation
temperature using immobilized yeasts. Process Biochem 34:785–794
Smogrovicová D, D€omény Z, Navrátil M, Dvorák P (2001) Continuous beer fermentation using
polyvinyl alcohol entrapped yeast. Proc Eur Brew Conv Cong 50:1–9
Strand BL, Gaserod B, Kulseng B, Espevik T, Skjak-Braek GJ (2002) Alginate-polylysine-
alginate microcapsules: effect of size-reduction on capsule properties. J Microencapsul
19:615–630
Takaya M, Matsumoto N, Yanase H (2002) Characterization of membrane bioreactor for dry wine
production. J Biosci Bioeng 93:240–244
Tata M, Bower P, Bromberg S, Duncombe D, Fehring J, Lau V, Ryder D, Stassi P (1999)
Immobilized yeast bioreactor systems for continuous beer fermentation. Biotechnol Prog
15:105–113
Thu B, Gaserod O, Paus D, Mikkelsen A, Skjak-Braek G, Toffanin R, Vittur F, Rizzo R (2000)
Inhomogeneous alginate gel spheres: An assessment of the polymer gradients by synchrotron
radiation-induced x-ray emission, magnetic resonance microimaging, and mathematical
modeling Biopolymers 53:60–71
Tsakiris A, Bekatorou A, Psarianos C, Koutinas AA, Marchant R, Banat IM (2004a) Immobili-
zation of yeast on dried raisin berries for use in dry white wine-making. Food Chem
87:11–15
Tsakiris A, Sipsas V, Bekatorou A, Mallouchos A, Koutinas AA (2004b) Red wine making by
immobilized cells and influence on volatile composition. J Agric Food Chem 53:1357–1363
Unemoto S, Mitani Y, Shinotsuka K (1998) Primary fermentation with immobilized yeast in a
fluidized bed reactor. Mast Brew Assoc Am Techn Quart 35:58–61
Van De Winkel L, De Vuyst L (1997) Immobilized yeast cell systems in today’s breweries and
tomorrow’s. Cerevisia 22(1):27–31
Van De Winkel L, Van Beveren PC, Borremans E, Goossens E, Masschelein CA (1993) High
performance immobilized yeast reactor design for continuous beer fermentation. Proc. 24th
Eur. Brew. Conv. Congr., pp 307–314
Van De Winkel L, Van Beveren PC, Masschelein CA (1991) The application of an immobilized
yeast loop reactor to the continuous production of alcohol-free beer. Proc. Eur. Brew. Conv.
Cong., pp 307–314
Van Iersel MFM, Van Dieren B, Rombouts FM, Abee T (1999) Flavor formation and cell
physiology during the production of alcohol-free beer with immobilized Saccharomyces
cerevisiae. Enzyme Microb Technol 24:407–411
Van Mulders SE, Christianen E, Saerens SM, Daenen L, Verbelen PJ, Willaert R, Verstrepen KJ,
Delvaux FR (2009) Phenotypic diversity of Flo protein family-mediated adhesion in Saccha-
romyces cerevisiae. FEMS Yeast Res 9(2):178–190
Veeranjaneya Reddy L, Harish Kumar Raddy Y, Prasanna Anjaneya Reddy L, Vijaya Sarathy
Reddy O (2008) Wine production by novel yeast biocatalyst prepared by immobilization on
watermelon (Citrullus vulgaris) ring pieces and characterization of volatile compounds.
Process Biochem 43:748–752
Verbelen PJ, De Schutter DP, Delvaux F, Verstrepen KJ, Delvaux FR (2006) Immobilized yeast
cell systems for continuous fermentation applications. Biotechnol Lett 28:1515–1525
Verbelen P, Nedovic VA, Manojlovic V, Delvaux F, Leskosek-Cukalovic I, Bugarski B, Willaert
R (2009) Bioprocess intensification of beer fermentation using immobilised cells. In: Zuidam
N-J, Nedovic VA (eds) Encapsulation technologies for food active ingredients and food
processing. Springer, Dordrecht, The Netherlands, pp 303–327
Verstrepen K, Derdelinckx G, Verachtert H, Delvaux FR (2003a) Yeast flocculation: what brewers
should know. Appl Microbiol Biotechnol 61:197–203
Verstrepen KJ, Moonjai N, Derdelinckx G, Dufour J-P, Winderickx J, Thevelein JM, Pretorius IS,
Delvaux FR (2003b) Genetic regulation of ester synthesis in brewer’s yeast: new facts, insights
and implications for the brewer. In: Brewing yeast fermentation performance, vol 2, 2nd edn.
Blackwell Science, Oxford, pp 234–248
Verstrepen KJ, Reynolds TB, Fink GR (2004) Origins of variation in the fungal cell surface.
Nat Rev Microbiol 2:533–540
Virkaj€arvi I, Kr€onlof J (1998) Long-term stability of immobilized yeast columns in primary
fermentation. J Am Soc Brew Chem 56:70–75
Virkaj€arvi I, Pohjala N (2000) Primary fermentation with immobilized yeast: effects of carrier
materials on the flavour of the beer. J Inst Brew 106:311–318
Virkaj€arvi I, Vainikka M, Virtanen H, Home S (2002) Productivity of immobilized yeast reactors
with very-high-gravity worts. J Am Soc Brew Chem 60(4):188–197
Wainwright T (1973) Diacetyl - a review. J Inst Brew 79:451–470
White FH, Portno AD (1979). The influence of wort composition on beer ester levels. Proc. Eur.
Brew. Conv. Cong., pp 447–460
Willaert R, Baron GV (1996) Gel entrapment and micro-encapsulation: methods, applications and
engineering principles. Rev Chem Eng 12:1–205
Willaert R, Nedovic V (2006) Primary beer fermentation by immobilised yeast – a review in
flavour formation and control strategies. J Chem Technol Biotechnol 81:1353–1367
Xu TJ, Zhao XQ, Bai FW (2005) Continuous ethanol production using self-flocculating yeast in a
cascade of fermentors. Enzyme Microb Technol 37:634–640
Yamauchi Y, Okamoto T, Murayama H, Kajino K, Amikura T, Hiratsu H, Nagara A, Kamiya T,
Inoue T (1995) Rapid maturation of beer using an immobilized yeast bioreactor. 1. Heat
conversion of a-acetolactate. J Biotechnol 38:101–108
Chapter 7
Multifactorial Assessment of Microbial Risks
in Foods: Merging Engineering, Science,
and Social Dimensions
Valerie Davidson, Juliana Ruzante, and Carlos Daza Donoso
7.1 Introduction and Context
The tenth International Congress on Engineering and Food (ICEF 10, Viña del Mar,
Chile, April 2008) broadly explored “the new agenda of innovation and achieve-
ment in food and engineering.” A common theme across many presentations was
the need for food engineers to work across interfaces between engineering and
science disciplines. Martin Cole, Director of the National Center for Food Safety
and Technology (USA), recognized this as a “crucial role” for food engineers in his
plenary lecture on emerging challenges in food safety. He pointed out that to be
successful in the current global food environment, “it will be necessary for the food
engineer to move deeper into the basic sciences,” and that in the area of food safety,
the food engineer will need “to use risk-driven studies to develop, improve and
optimize food products and processes” (Cole and Rodriguez 2008).
Under the ICEF 10 conference theme of “Food Processing” our research group
presented a body of work titled “Multifactorial Risk Prioritization Framework for
Food-borne Pathogens” (Davidson et al. 2008). This framework systematically
compiles information and characterizes risks due to microbial hazards in foods
based on four factors: public health, market value, consumer perceptions of risk,
and social concerns. The framework includes specific tools to assist risk managers
in comparing and ranking risks as a basis for setting priorities and strategic alloca-
tion of resources for risk assessment needs and interventions to mitigate risks. The
Multifactorial Risk Prioritization Framework is positioned at the interface between
risk management and risk assessment and serves as an important connection
between these essential elements of food safety.
V. Davidson (*) and C.D. Donoso

School of Engineering, University of Guelph, Guelph, ON N1G 2W1, Canada
e-mail: vdavidso@uoguelph.ca
J. Ruzante
Joint Institute for Food Safety and Applied Nutrition, University of Maryland, College Park,
MD 20742, USA
148 V. Davidson et al.
In this chapter, we outline the components of the multifactorial framework and

the tools to assist risk managers. The discussion highlights the multidisciplinary
nature of the risk framework and the ways that food engineers can contribute to the
development and implementation of food safety strategies. We also offer some
thoughts on the challenges of multidisciplinary research as well as the implications
for training food engineers.
7.2 Risk Management Frameworks for Food Safety
A generic representation of a risk management cycle, fully integrated with risk

assessment, is shown in Fig. 7.1. Risk management is an iterative process, beginning
with the need to evaluate and rank risks within an appropriate context (Block A).
Typically, microbial risk assessment (Block B) is focused on public health and
carried out by experts who have specialized scientific and technical knowledge and
are at an arm’s-length from the risk management process. Their knowledge is
derived from observations, forecasts, and expert judgments. These various inputs
need to be summarized in formats that allow decision-makers to compare risks
across multiple dimensions, with an understanding of any limitations, including
a. Risk Evaluation
Identify food safety
problems
Establish risk profile
Rank risks
d. Monitoring b. Risk Assessment

and Review
Results monitoring Public health
Stakeholder
communication
Process verification
Data administration
c. Risk Management
Options assessment
and interventions
Fig. 7.1 Generic risk management cycle

7 Multifactorial Assessment of Microbial Risks in Foods 149
uncertainty and variability, in the data and opinions. Based on scientifically sound
risk assessment, risk managers make decisions regarding allocation of resources to
implement interventions and mitigate serious risks (Block C). The final step
includes routine surveillance and data gathering (Block D) in order to review
specific interventions.
The joint FAO/WHO report, “Principles and guidelines for incorporating micro-
biological risk assessment in the development of food safety standards” (Food and
Agriculture Organization of the United Nations and the World Health Organization
2002), defines a comprehensive context for risk management related to food safety.
It recognizes that risk management must rest on the foundations of scientifically
sound risk assessment and should consider areas of health impact (e.g., adverse
health effects, duration of illness, severity and impairment of life quality), eco-
nomic and market impact (e.g., economic burden and facilitation of fair trade),
impact on consumer behavior, and social/ethical concerns (susceptibility of
exposed population, nutritional status, social status).
A comprehensive review (Henson et al. 2007) identified a number of frame-
works for prioritization of food-borne microbial risks. Impact on human health was
the primary basis for ranking risks but some frameworks attempted to incorporate
additional risk factors. In the United States, the Food Safety Research Consortium
developed the Food-borne Illness Risk Ranking Model (FIRRM), which evaluates
multiple aspects of public health, including incidence of illness, outcome severities,
monetary (e.g., cost of illness) and non-monetary factors (quality-adjusted life
years) (Batz et al. 2004). New Zealand’s risk ranking tool includes public health
measures that are weighted by severity in the general public and subpopulations; it
acknowledges the economic implications of food-borne pathogens on trade (New
Zealand Food Safety Authority 2008). A recent FAO/WHO expert meeting defined
six criteria for ranking microbial risks in fresh fruits and vegetables and three
criteria are associated with market and economic factors: size of production in a
global context, diversity and complexity of the production chain and industry, and
extent of international trade and economic impact (Food and Agriculture Organi-
zation of the United Nations and the World Health Organization 2008). Finally, the
European Commission (EC) has put forward an integrated framework that high-
lights the importance of balancing the scientific, economic, social, and cultural
aspects of risks and benefits as well as risk–benefit distribution within society
(European Commission 2002). The EC framework is intended to provide guidelines
for harmonizing risk management strategies in the European Union.
7.3 Multifactorial Risk Prioritization Framework
The Multifactorial Risk Prioritization Framework recognizes the need to evaluate

multiple risk factors in an integrated format so that risk managers can compare and
rank risks – at the level of pathogen–food pairs, across pathogens, and across food
a. Risk Evaluation
Identify food safety
problems
Establish risk profile
Rank risks
d. Monitoring and Risk Prioritization b. Multifactorial

Review Framework Assessment
Results monitoring 1. Public health Risk assessment
Stakeholder 2. Market assessment Market assessment
communication 3. Consumer perception &
acceptance of risk Consumer assessment
Process verification 4. Social sensitivity (perception & acceptance)
Data administration Social sensitivity
c. Risk Management
Options assessment
and interventions
Fig. 7.2 Multifactorial risk prioritization framework
categories. Figure 7.2 shows the integration of the Multifactorial Risk Prioritization
Framework within a risk management cycle. The framework is intended to assist
risk managers and provide tools to support decision-making. Risk prioritization or
ranking at the evaluation stage (Block A) is based on multifactor risk profiles
extracted from the framework knowledge base; this ranking is the basis for allocat-
ing resources for development of adequate knowledge, which is carried out through
multidimensional assessments of high-priority hazards (Block B). Ranking is also a
key component in selecting appropriate interventions (Block C). As shown in
Fig. 7.2, the framework addresses the need for an integrated information system
that provides inputs for evaluation and assessment of interventions, and can be
updated by knowledge developed through risk assessments, consumer behavior,
market-level studies, and surveillance systems.
Given the nature and complexity of risk associated with microbial hazards in
foods, one of the first challenges in developing a framework is to establish an
appropriate context for comparing different pathogens and food sources. The
Multifactorial Risk Prioritization Framework integrates information from four
major areas identified in the FAO/WHO guidelines: public health, market-level,
consumer behavior, and social factor assessments. Risk assessments for each area
are complex and multidimensional. A detailed discussion of possible risk measures
for each factor in the framework is presented in Henson et al. (2007) and the
rationale for choosing particular risk measures is given in Ruzante et al. (2009).
In the discussion that follows, the multifactor risk measures are outlined in terms of
underlying models, data requirements, and engineering analysis; decision tools are
also demonstrated with examples of risk profiles based on Canadian data.
7.3.1 Defining Risk Factors
7.3.1.1 Public Health Assessment
Clearly public health impact is a primary consideration for risk management

decisions related to food safety. Potential outcomes include morbidity (immediate
illness with varying degrees of severity and longer-term sequelae) and mortality
(fatalities). Since health outcomes vary for different food-borne pathogens, a
number of aggregate measures are used to summarize burden and cost of illness.
In the current framework, we use two dimensions to characterize public health
impact: cost of illness and disability-adjusted life years. Cost of illness (COI)
reflects medical care costs, productivity costs due to time lost from paid employ-
ment, and economic value of premature death. Disability-adjusted life years
(DALY) is a summary measure of population health and combines morbidity and
mortality measurements in a single parameter, as shown in the following equations:
DALY ¼ Years of life lossðYLLÞ þ Years lived with disabilityðYLDÞ;
where
X
9
YLL ¼ dj ej ðj ¼ age group; d ¼ # deaths; e ¼ expected lifespanÞ;
j¼1
and
X
YLD ¼ nl tl wl ðl ¼ health outcome; n ¼ # of cases;
l
t ¼ duration of the illness; w ¼ diability=severity weightÞ:
The COI approach can mask differential effects across subpopulations since it
does not consider productivity costs outside the paid labor force (e.g., young
children and the elderly are not included). Although DALY is used by the WHO
in global assessments of the burden of disease, the disability weights are subjective
valuations of the time lived in nonfatal health status (Murray 1994; Murray and
Acharya 1997). Since both COI and DALY have strengths and weaknesses, we
include both measures in the framework to provide complementary depictions of
the public health impact.
The underlying models for COI and DALY are developed by economists and
scientists with specialized knowledge of public health. However, implementation
requires careful systems analysis to ensure consistency of data and analysis across
diverse health outcomes, and to define limitations in the calculated values due to
uncertainty and variability in model inputs. As food engineers, we work with
economists, epidemiologists, microbiologists, veterinarians and food scientists to
understand the issues related to underreporting of food-borne illness, attribution of
illness to particular food categories, and other challenges in building credible
models. At present, we are using existing knowledge of uncertainty and variability
in model parameters to estimate uncertainty intervals for DALY and COI values in
the framework. We are also working to develop better knowledge for the Canadian
context in areas such as attribution of illness to particular food categories.
7.3.1.2 Market-Level Assessment
The economic losses arising from food-borne pathogen incidence and prominent
outbreaks can be large – for individual groups or firms and, in some cases, for an
entire industry sector. Recently a Canadian food processor agreed to pay up to $27
million to settle class action lawsuits initiated after a national listeriosis outbreak
linked to 20 deaths (Canadian Broadcasting Corporation 2008). Another recent
outbreak of Salmonella saitpaul associated with tomatoes and peppers caused
losses of approximately $100 million for the US tomato industry due to lower
prices and reduced demand (Western Farm Press 2008). These examples underscore
the importance of considering economic impacts that could arise from food-borne
pathogen outbreaks or unacceptable levels of incidence in the risk management
process. The goal of the market-level dimension is to review a priori the potential
economic impacts of pathogen incidence or outbreaks at an overall market level. At
this point, we do not attempt to estimate the likelihood of an economically signifi-
cant outbreak or unacceptable incidence levels; however, we recognize the value of
developing estimates of likelihood for the framework in the future.
Henson et al. (2007) presented a comprehensive list of market-level intelligence
that would be helpful in evaluating economic impacts; this list is summarized in
Table 7.1. This broad set of inputs spanning the entire food supply chain, from
production to retail, accounts for the economic value of imports and exports. In
developing some of this information for the Canadian context, we encountered
challenges in obtaining complete information for particular food groups and indus-
try subsectors. For example, information on wages and number of employees by
sectors was not available in Canada. Also, public information on the value of sales
is generally limited to cash receipts at the farm gate and comparable information for
processed foods is only available through special reports (e.g., reports produced by
government agencies such as Statistics Canada and Agriculture and Agri-food
Canada) or market surveys prepared by market consultants.
Figure 7.3 shows an example of market-level information extracted from the
framework database for Canadian chicken products. The market summary includes
descriptive and quantitative information that provides context and consistent measures
for decision-makers to use in comparing risks across pathogen–food combinations.
Table 7.1 Suggested elements for market-level assessment (Henson et al. 2007)
Subsector Domestic market International market
Retail level l Value of sales ($) l Value of trade ($ and % of value of domestic
production)
l Volume of sales (#) l Volume of trade (#)
l Employment (#) l Change in consumption (%)
l Wages ($) l Impact on market access (index)
l Change in l Export competitiveness (index)
consumption (%) l Size of global market (% of global production)

l Number of trade partners (#)
l Trade concentration ratio (index)
Processing–distributing– l Value of sales ($) l Value of trade ($ and % of value of domestic

wholesaling level production)

Farm level l Value of sales ($) l Value of trade ($ and % of value of domestic
production)

At present we use a simple aggregate measure as a basis to compare market risk

across different food categories. This is an estimate of the economic importance of
domestic market activities (on an annual basis) calculated as the total value at retail
plus the value of exports minus the value of imports. This measure captures the size
of economic activity in the domestic market that is potentially at risk due to
incidence and outbreaks.
The market-level analysis is a critical component of the risk context and must be
integrated into the food safety framework. As mentioned, we would like to include
estimates of the likelihood of outbreaks in a future version of this framework. This
will require a multidisciplinary approach, and food engineers’ knowledge of pro-
cessing operations and systems-based failure analysis will be important contribu-
tions to this enhancement of the framework.
7.3.1.3 Consumer Assessment
The consumer behavior dimension of the multifactorial risk framework recognizes

that consumers may well be willing to accept a food-safety risk if the perceived risk
Fig. 7.3 Summary of market-level information for Salmonellosis – chicken

is low or if the perceived benefits arising from the consumption of a particular food
offset perceptions of ill consequences. The consumers’ perception and acceptance
of risk arising from food-borne pathogens can translate into shifts in market demand
for particular food products. For example, the Guelph Food panel (a large-scale
panel of consumers dedicated to food research) reported the following changes in
consumer behavior after a recent listeriosis outbreak in Canada (University of
Guelph 2008):
l Thirty-nine percent would not consume ready-to-eat meats at home (up from 6%
reported before recall).
l Fifty-six percent would not consume ready-to-eat meat products in fast-food
outlets or restaurants (up from 9% reported before recall).
l Thirty percent stopped buying ready-to-eat meats from Canada.
Consumers may switch from certain products perceived as “less safe” to those
perceived as “more safe,” although such perceptions may bear only a loose rela-
tionship with scientific assessments of risks to human health. Furthermore, con-
sumer perceptions of risks associated with food are an important determinant in the
confidence the general public has in the security of food systems and in related
systems of public regulation and oversight. Relatively small but highly visible
outbreaks of disease can have a profound impact on the trust that consumers have
in food producers, manufacturers and distributors, or government regulators.
Measurement of risk acceptability and perception is complex. There is a consid-
erable amount of literature and survey-based research on how consumers rank
different risks (Fife-Schaw and Rowe 1996; Fischoff et al. 1978; Frewer et al.
1994, 1997, 1998a, 1998b). However, as Frewer et al. (1998a) point out, food is not
something to be avoided. Thus, consumers may view food-related risks differently
from avoidable risks.
In the current framework, we use five scales to assess consumers’ perception and
acceptance of risks associated with a food–pathogen combination:
1. Degree to which risk is perceived as uncontrollable by consumers
2. Degree to which risk is perceived as unknown to the individual
3. Degree to which risk is perceived as unknown to scientists
4. Degree to which risk is perceived as involuntary
5. Degree to which consumers perceive health outcome as severe
Each criterion is scored on a three-level nominal scale (low, medium, high) by a
consumer panel. To date, a small Delphi panel has provided opinions for the six
case studies based on their knowledge of consumer behavior. The overall measure
used to represent consumers is the sum of the average scores for each criterion,
normalized on a 0–1 scale (i.e., sum is divided by 5).
Some of the researchers in our team recently conducted a consumer survey to
explore perceptions and acceptance of risks related to pathogen–food combinations.
Preliminary results need to be validated with a follow-up survey; however, initial
results showed that consumers are concerned about risks perceived as uncontrolla-
ble (e.g., through their own preparations prior to consumption, by process steps in
manufacturing) and where the outcome is severe (e.g., chronic health effects,
death). There is also concern about risks in foods given to young children or
other vulnerable groups.
7.3.1.4 Social Factor Assessment
As noted by Caswell (2005), there needs to be “a recognition that risk analysis,

especially risk assessment, are tools for making decisions that better reflect
society’s values.” Impacts on vulnerable groups such as young children and the
elderly are given careful consideration in risk management decisions because
society feels an obligation to protect these groups. We have incorporated a social
sensitivity factor into the risk prioritization framework to account for the potential
of wider social consequences, beyond what is accounted for in public health and
market-level dimensions. The social sensitivity factor is based on two sub-criteria:
consumers (e.g., subpopulations such as pregnant women) and firms (e.g., niche
industries unique to certain regions).
At the consumer level, the sensitivity measure does not relate to the potentially
greater individual risk that consumer subgroups may face from a particular
pathogen–food combination, since this is incorporated into the public health factor.
Societal concerns may relate, for example, to the more limited ability of some
individuals to take self-protective actions or may be based in altruism. Similarly,
on the food supply side, such concerns do not reflect the unfavorable economic
impact on individual firms or an industry per se because this is included in the
market-level factor. Instead this measure reflects sensitivity to the role that such
enterprises play in rural or economically vulnerable areas and their contribution to the
historical and/or social fabric of society. The social sensitivity factor is perhaps more
controversial than the other factors since it is dependent upon individual views and
ethical positions. However, these considerations need to be incorporated in an explicit
rather than ad hoc manner. At this point, the social sensitivity factor is a binary flag
(0 ¼ no concern; 1 ¼ concern) for each of the two sub-criteria.
7.3.2 Information Cards
The first tool that has been developed in the Multifactorial Risk Prioritization
Framework is a set of concise summaries or information cards for each of the
four risk factors. For example, the information card in Fig. 7.4 shows the public
health impact for a specific food–pathogen combination: Salmonellosis in chicken.
These information cards provide contextual information, such as health outcomes
and market characteristics (Fig. 7.3), as well as quantitative indicators. The data
required to calculate the measures for each of the four risk factors are stored in
electronic format and are updated at intervals appropriate for the information
source. For example, Statistics Canada data are available on an annual basis but
Fig. 7.4 Summary of public health information for Salmonellosis – chicken

information from the scientific literature may not be revised as frequently. In cases
where data are updated on an annual basis, values are shown for the most recent
year as well as a 3-year average based on annual values to reduce the effect of
unusual events on decisions. At this point, the values for all risk factors are shown
as single values or point estimates, without explanation of uncertainty. We recog-
nize the need to quantify the uncertainty due to variability and, in some cases,
limited knowledge or data in each of the risk measures. This analysis is in progress.
7.3.3 Multifactorial Risk Prioritization
As outlined in the preceding sections, one or two aggregate measures are defined for
each of the four areas of risk assessment: public health (DALY and COI values),
market-level (economic value of domestic market activities), consumer (aggregate
measure of perception and acceptance of risk), and social sensitivity (consumer and
firm). While we recognize that other measures could be used, we consider these six
measures to be comprehensive and consistent with the principles outlined by FAO/
WHO. Using our framework, multidimensional risk profiles can be created for
pathogen–food combinations, food categories, and different pathogens. In all cases,
a ranking process must be able to compare each one on the basis of multiple criteria.
Multi-criteria decision analysis (MCDA) techniques are structured, consistent,
and transparent, and therefore helpful in dealing with large amounts of complex
information. We have explored a number of MCDA tools for use in the prioriti-
zation framework (Daza Donoso 2008; Ruzante et al. 2009). Daza Donoso (2008)
compared a number of MCDA techniques on the basis of their expandability,
ability to incorporate uncertainty and variability, and interactions with decision-
makers (e.g., level of knowledge, inputs and time requirements, analytical skills).
On this basis, outranking methods had definite advantages over techniques such as
the multi-attribute utility theory and the analytic hierarchy process. Outranking
methods are based on pair-wise comparisons of alternatives. For each criterion,
alternative “a2” is compared to alternative “a1” using a preference or outranking
relation defined in an appropriate scale for the criterion. This permits the use of
ordinal scales and binary values for risk criteria. Furthermore, the outranking
relations can be constructed to reflect uncertainty and varying degrees of prefer-
ence (weak to strong) or higher ranking. Figure 7.5 shows an outranking relation
for the criterion related to consumer risk perception and acceptance of risk. Two
discrimination thresholds are defined: indifference (r) and strict preference (s).
When two alternatives (a1 and a2) are compared, there is insufficient evidence
to rank a2 higher when the difference in the criterion scores is less than 0.05
(indifference threshold r). This threshold value is defined specifically for the
criterion related to consumer perception and acceptance of risk; it is based on our
estimate of the consistency in opinions from our Delphi panel. For score differ-
ences in the interval between 0.05 and 0.33, there is weak but increasing
preference (or higher priority) for a2; above 0.33 (strict preference threshold s)
Indifference Weak preference Strict preference for a2

for a2
Difference in consumer perception and acceptance scores (a2 – a1)
Fig. 7.5 Outranking relation for consumer perception and acceptance of risk criterion
a2 clearly ranks as a higher risk than a1. The characteristics of the outranking
relations (thresholds and transitions between thresholds) can be used to reflect
some aspects of uncertainty and variability in the risk measures.
Daza Donoso et al. (2008) used two outranking methods, ELECTRE III (Elimi-
nation et Choix Traduisant la Réalité) and PROMETHEE (Preference Ranking
Organization Method for Enrichment Evaluations) to rank six pathogen–food
combinations. The multidimensional risk profiles for the six examples are summar-
ized in Table 7.2. The ELECTRE III method developed by Roy (1978) is available
in Windows-compatible software from the LAMSADE research group (Laboratoire
d’analyse et modélisation pour l’aide à décision 1994). PROMETHEE is an out-
ranking method developed by Brans and colleagues (Brans and Vincke 1985; Brans
et al. 1986) and implemented in Decision Lab software (Visual Decision Inc.,
Montreal, QC, Canada). The same discrimination thresholds were defined for
PROMETHEE and ELECTRE III ranking relations as shown in Table 7.3.
Although the ELECTRE III method allowed for a “veto” effect, this component
was not used in ranking the six pathogen–food combinations since it was assumed
that all six pairs posed risk, regardless of how much lower its performance in any
one criterion may be. The final rankings were based on the equal weightings for the
four risk factors (Table 7.2).
As shown in Table 7.4, Escherichia coli O157 in beef is ranked as the highest
priority in both outranking methods. At the number two rank, there are differences
depending on the ranking method. In ELECTRE III, two pathogen–food
Table 7.2 Risk profiles for six pathogen–food combinations (Canadian data)
Pathogen–food Public health Market Consumer perception Social sensitivity
combination DALYa COIb Domestic and acceptance of risk Consumer Firm
sizec
(years) (CAN (CAN$ Normalized average Flag Flag
$ 106) 106)
Campylobacter/ 808 79.8 5472 0.3 0 0
chicken
Salmonella/chicken 449 79.4 5472 0.25 0 0
Salmonella/spinach 1 0.2 118 0.5 0 0
Escherichia coli 3 0.5 118 0.8 1 0
O157/spinach
E. coli O157/beef 260 40.2 5264 0.6 1 0
Listeria 58 12.7 974 0.6 1 1
monocytogenes/
ready-to-eat meats
Criterion weights 0.125 0.125 0.25 0.25 0.125 0.125
a
DALY – Disability-adjusted life years
b
COI – Cost of illness
c
Domestic market size (annual basis) ¼ value of retail sales + value of exports – value of imports
Table 7.3 Characteristics of outranking relations in ELECTRE III and PROMETHEE analysis
Criterion Discrimination thresholds Transition between
Indifference Strict thresholds
preference
DALY (years) 10 78 Linear
COI (CAN$ 106) 0.1 9 Linear
Economic value of domestic market 100 1000 Linear
activities (CAN$ 106)
Consumer risk perception and 0.05 0.33 Linear
acceptance of risk
Social sensitivity – consumer 0 1 Not applicable
Social sensitivity – firm 0 1 Not applicable
Table 7.4 Ranking of six pathogen–food combinations

Rank position ELECTRE III PROMETHEE I PROMETHEE II
Partial ranking Complete ranking
1 Escherichia coli O157/beef E. coli O157/beef E. coli O157/beef
2 Campylobacter/chicken Campylobacter/chicken Listeria monocytogenes/
L. monocytogenes/RTEm L. monocytogenes/RTEm RTEm
3 Salmonella/chicken Salmonella/chicken Campylobacter/chicken
E. coli O157/spinach
4 Salmonella/spinach E. coli O157/spinach Salmonella/chicken
5 Salmonella/spinach E. coli O157/spinach
6 Salmonella/spinach
combinations – Campylobacter in chicken and Listeria monocytogenes in ready-to-

eat meats – are considered to be of equivalent risk priority but are actually
incomparable. The basis for this “incomparability” can be understood by compar-
ing the risk profiles in Table 7.2. Campylobacter in chicken has high scores in the
factors related to public health and market-level impact but much lower scores
(relative to the other five pairs) in consumer perception and acceptance of risk and
social sensitivity. L. monocytogenes in ready-to-eat meats is the highest risk case in
terms of social sensitivity factors and higher than Campylobacter in chicken in
terms of consumer perception and acceptance of risk. The PROMETHEE I ranking
is referred to as a partial ranking and positions the same two pathogen–food
combinations at number two. In the partial ranking procedure, the six pairs are
ordered based on the degree to which each pair outranks the other pairs (positive
preference flows) as well as the degree to which each pair is outranked by the others
(negative preference flows). Incomparability in PROMETHEE I arises when there
is an inconsistency in the ordering based on the two analyses – the two pairs change
positions in the positive and negative preference orders. The PROMETHEE II
ranking is referred to as a complete ranking and is based on the net preference
flow (positive preference flow – negative preference flow). It may be easier to
interpret because there are no incomparable cases. However, the complete rankings
can mask diversity in the risk profiles that should be appreciated by risk managers in
making their decisions.
Overall, this is a limited data set but it is useful in demonstrating some of the
limitations and the value of MCDA methods in ranking microbial risks in the food
system. Future work includes refinement of ranked lists to reflect uncertainty in
ranking and feasibility of interventions to reduce risks.
7.4 Food Engineering at Risk Management/Risk Assessment

Interface: Challenges and Implications for Training
As noted in the introduction, the Multifactorial Risk Prioritization Framework has

been developed by a multidisciplinary research group. The members of the group as
well as their primary discipline areas and affiliations are listed below:
Sven Anders Agricultural Economics University of Alberta

Julie Caswell Agricultural Economics University of Massachusetts – Amherst
John Cranfield Agricultural Economics University of Guelph
Valerie Davidson Food Science/Engineering University of Guelph
Carlos Daza Donoso Food Science/Engineering University of Guelph
Jeff Farber Food Microbiology Health Canada
Aamir Fazil Risk Assessment Public Health Agency of Canada
Spencer Henson Agricultural Economics University of Guelph
Shannon Majowicz Epidemiology Public Health Agency of Canada
Juliana Ruzante Veterinary Medicine University of Maryland
Claudia Schmidt Agricultural Economics University of Guelph
In addition, we have consulted widely to address specific data areas (e.g., food
attribution and market information) and to gather expert knowledge needed for
health outcome and economic models. Not surprisingly there have been challenges
as well as rewards. In keeping with the overall theme of “Food Engineering at the
Interfaces” we have included a short reflection on our experiences in multidisci-
plinary research.
Food safety efforts at the interface of risk management and risk assessment
clearly must rely on science- and systems-based analyses. Many disciplines
(e.g., epidemiology, microbiology, food science, economics, food engineering)
bring important but specialized knowledge to this interface. It is our observa-
tion that language is one of the first challenges in a multidisciplinary group.
Each discipline has its own unique vocabulary and terms, but it is the deeper
use of language – the way we develop our ideas, frame research questions,
and present analysis – that is probably more challenging. Key to success is
the willingness of the experts to communicate and explain their knowledge at
an appropriate level as well as the willingness of “outsiders” to admit any
incomplete understanding and to ask questions. In other words, everyone must
reach a point of humility and this is not necessarily easy for researchers in
any discipline.
Another challenge is the range of methods required in this research. Food
engineers are trained in quantitative analysis and predictive modeling but are less
familiar with qualitative methods and survey tools. These are important techniques
in areas related to consumer behavior as well as in some data gathering. Some
information cannot be obtained by direct observation and must be developed by
soliciting experts’ opinions. Food engineers need to recognize the need for addi-
tional research tools and to value the contributions of all research methods equally.
Adjectives like “hard” and “soft” should be avoided when discussing methods and
results, because they are often used erroneously to convey perceptions that certain
methods lack scientific rigor.
Finally, we were fortunate that there was a funding opportunity through the
Natural Sciences and Engineering Research Council (NSERC), a national funding
agency for science and engineering research in Canada, which established food
safety as a priority area in the Strategic Grants Program. The NSERC funding was
critical in moving the framework from a conceptual stage to the concrete mea-
sures and six case studies presented here. The funding was a unique opportunity
to bring this multidisciplinary research group together. Several components of the
work (e.g., economics, consumer behavior) fall under the jurisdiction of the Social
Sciences and Humanities Research Council and there are very few programs to
support research at the interface of two research councils.
Multidisciplinary research is an excellent training experience for graduate stu-
dents. A number of graduate students in engineering are currently involved in
developing different aspects of the framework (knowledge base, multi-criteria
decision, analysis tools). Their knowledge and skills are important contributions
to the framework; they are also learning new areas of public health, risk manage-
ment, and operations research.
7.5 Concluding Statements
The Multifactorial Risk Prioritization Framework addresses the need to base risk
management decisions on scientific analysis across a comprehensive set of criteria,
and to communicate such decisions among varied stakeholders in a clear and
transparent way. It is not intended to replace risk managers but to provide context
and tools for strategic planning of research needs and interventions that reduce risk.
The Multifactorial Risk Prioritization Framework has been developed and
implemented by a multidisciplinary group that includes food engineers. Systems
analysis, modeling, and quantitative methods are necessary components of risk
analysis, but they are not sufficient on their own. Food engineers must work
collaboratively with scientists (e.g., economists, epidemiologists, microbiologists,
risk analysts) as well as decision-makers (e.g., risk managers, policy makers) to
develop effective strategies for setting food-safety priorities. The present frame-
work is an excellent example of the potential to develop a more comprehensive
solution by working together rather than in isolated disciplines.
Acknowledgments The financial support of the Natural Sciences and Engineering Research
Council to carry out this work is gratefully acknowledged.
References
Batz MB, Hoffmann S, Krupnick AJ, Morris JG, Sherman DM, Taylor MR, Tick JS (2004)
Identifying the most significant microbiological food-born hazards to public health: a new
risk ranking model. Food Safety Research Consortium Discussion Paper Series No. 1. Avail-
able at: http://www.thefsrc.org/Discussion%20Papers/FRSC-DP-01.pdf (accessed January 30,
2009)
Brans JP, Vincke P (1985) A preference ranking organization method: the PROMETHEE method
for MCDM. Manage Sci 31(6):647–656
Brans JP, Vincke P, Mareschal B (1986) How to select and how to rank projects: The PRO-
METHEE Method. Eur J Oper Res 24:228–238
Canadian Broadcasting Corporation (2008) Maple Leaf settles class action listeriosis lawsuits for
$27 M. Available at: www.cbc.ca/canada/story/2008/12/18/listeriosis-settlement.html?ref ¼
rss (accessed January 29, 2009)
Caswell JA (2005) Opportunities for Risk Reduction: An Economist’s Perspective. In: Hoffmann
SA, Taylor MR (eds) Toward Safer Food Perspectives on Risk and Priority Setting. RFF Press,
Washington, DC, p 276
Cole M, Rodriguez A (2008) Emerging Challenges in Food Safety and the Role of Food Engineers,
Plenary Lecture at 10th International Congress on Engineering and Food, Viña del Mar, Chile,
April 20–24
Davidson VJ, Ruzante J, Fazil A, Caswell J, Cranfield J, Henson S (2008) Multifactorial risk
prioritization framework for food-borne pathogens. 10th International Congress on Engineer-
ing and Food, Viña del Mar, Chile, April 20–24
Daza Donoso CA (2008) Application of multicriteria decision analysis tools to the prioritization of
microbial hazards in food systems. M.Sc. thesis, Faculty of Graduate Studies, University of
Guelph
Daza Donoso CA, Davidson VJ, Ruzante JM, Fazil A (2008) Prioritization of microbial risks in
foods using multi-criteria decision analysis tools, Canadian Society of Chemical Engineering
Conference, October 19–22, Ottawa, Ontario
European Commission (2002) Risk assessment of foodborne bacterial pathogens: quantitative
methodology relevant for human exposure assessment. Health and Consumer Protection
Directorate-General Directorate C – Scientific Opinions, C1 – Follow-Up and Dissemination
of Scientific Opinions, Preliminary Report. Available at: http://europa.eu.int/comm/food/fs/sc/
ssc/out252_en.pdf (accessed January 30, 2009)
Fife-Schaw CR, Rowe G (1996) Public perceptions of everyday food hazards: a psychometric
study. Risk Anal 16(4):487–500
Fischoff B, Slovic P, Lichtenstein S, Read S, Combs B (1978) How safe is safe enough? A
psychometric study of attitudes towards technological risks and benefits. Policy Sci
9(2):127–152
Food and Agriculture Organization of the United Nations and World Health Organization (2002)
Principles and guidelines for incorporating microbiological risk assessment in the development
of food safety standards and related texts. Kiel, Germany. Report of a Joint FAO/WHO
consultation, March 18-xx Available at: http://www.fao.org/docrep/006/y4302e/y4302e00.
HTM (accessed January 31, 2009)
Food and Agriculture Organization of the United Nations (FAO) and World Health Organization
(WHO) (2008) Microbial hazards in fresh fruits and vegetables (pre-publication version).
Available at: http://www.fao.org/ag/agn/agns/files/ffv_2007_Final.pdf (accessed January 30,
2009)
Frewer LJ, Howard C, Shepherd R (1997) Public concerns about general and specific applications
of genetic engineering: risk, benefit and ethics. Sci Technol Hum Values 22:98–124
Frewer LJ, Howard C, Hedderley D, Shepherd R (1998a) Methodological approaches to assessing
risk perceptions associated with food-related hazards. Risk Anal 18:95–102
Frewer LJ, Howard C, Shepherd R (1998b) Understanding public attitudes to technology. J Risk
Res 1(3):221–235
Frewer LJ, Shepherd R, Sparks P (1994) Biotechnology and food production: knowledge and
perceived risk. Br Food J 96:26–32
Henson S, Caswell J, Cranfield J, Fazil A, Davidson VJ, Anders S, Schmidt C (2007) A multi-
factorial risk prioritization framework for food-borne pathogens. Available at: Social Science
Research Network: http://ssrn.com/abstract ¼ 989768
Laboratoire d’analyse et modélisation pour l’aide à décision (LAMSADE) (1994) ELECTRE III/
IV Software, Université Paris-Dauphine, PARIS CEDEX 16, France
Murray CJL (1994) Quantifying the burden of disease: the technical basis for disability-adjusted
life years. Bull World Health Organ 72(3):429–445
Murray CJL, Acharya AK (1997) Understanding DALYs. J Health Econ 16:703–730
New Zealand Food Safety Authority (2008) Food safety in New Zealand: application of a risk
management framework. Available at: http://www.nzfsa.govt.nz/about-us/risk-management-
framework/page-03.htm (accessed January 30, 2009)
Roy B (1978) ELECTRE III: Un algorithme de classements fondé sur une représentation floue des
préférences en présence de critères multiples. Cahiers du Centre d’Etudes de Recherche
Opérationnelle 20(1):3–24
Ruzante JM, Davidson VJ, Fazil A, Cranfield JAL, Henson SJ, Caswell JA, Anders SM, Schmidt C,
Farber J (2009) A Multifactorial risk prioritization framework for foodborne pathogens,
accepted risk analysis, July. doi:10.1111/j.1539–6924.2009.01278.x
University of Guelph (2008) Consumers changed habits following listeriosis outbreak, Study finds,
December 03 news release. Available at: http://www.uoguelph.ca/news/2008/12/post_157.
html (accessed November 18, 2009)
Western Farm Press (2008) Fresh tomato industry shaken by FDA salmonella links, seeks answers.
Available at: http://westernfarmpress.com/vegetables/fresh-tomato-0821/
Chapter 8
Development of Eco-efficiency Indicators
to Assess the Environmental Performance
of the Canadian Food and Beverage Industry
Michèle Marcotte, Yves Arcand, Dominique Maxime, and Denyse Landry
8.1 Introduction
Agriculture and Agri-Food Canada (AAFC) is committed to adopting sustainability

as a key principle and has been working on the development of agri-environmental
indicators (AEI) to assist the agricultural sector in assessing its impact on environ-
mental health. Since 2003 and for a period of 5 years, through the implementation
of the Agriculture Policy Framework (APF) (Agriculture and Agri-Food Canada
2004), agri-environmental indicators have been developed specifically addressing
environmental issues at the Canadian Food and Beverage Industry (FBI) level to
provide a more comprehensive assessment of the impact of the whole food value
chain on the environment (NAHARP 2004). This chapter presents the general
approach being used in the development of science-based agri-environment indi-
cators specific to the Canadian Food and Beverage Industry, their calculation using
statistical survey data, and a first report on the results.
8.2 Overview of the Canadian FBI
The Canadian food and beverage industry (FBI) consists of approximately 6,700
establishments involved in the transformation of raw agricultural commodities into
semi-prepared and consumer-ready food and beverage products. The FBI is important
M. Marcotte (*)
Agriculture and Agri-Food Canada’s Eastern Cereal and Oilseed Research Centre, 960 Carling
Avenue, KW Neatby, Room 1093, Ottawa, ON, K1A 0C6, Canada
e-mail: Michele.marcotte@agr.gc.ca
Y. Arcand and D. Maxime
Agriculture and Agri-Food Canada’s Food Research and Development Centre, 3600 Casavant
Blud West, St. Hyacinthe, QC, Canada
D. Landry
Agriculture and Agri-Food Canada Head Quarter, 1341 Baseline Road, Touer 5, Floor 2, Room
126 Ottawa, ON, K1A OC5, Canada
166 M. Marcotte et al.
to the Canadian economy, being the largest manufacturing employer (296,000

Canadians), especially in many rural and agricultural areas across Canada. Over 90%
of FBI establishments have fewer than 100 employees – 45% have fewer than five
employees. It is the second largest manufacturing industry in Canada and accounts for
14% of total manufacturing shipments and 2% of the national Gross Domestic Product
(GDP). The FBI produces shipments worth $87 billion. Furthermore, it supplies
approximately 77% of all processed food and beverage products available in Canada.
The food and beverage processing industry (FBI), which is classified as a
manufacturing industry, is a major intermediary in the food chain. The FBI is the
critical link in the “farm to fork” value chain that transforms agricultural products for
consumers. It is the pathway of almost half of Canada’s raw primary agricultural
output, and around 70% of FBI’s inputs come directly or indirectly from agricultural
production or fisheries. These inputs are processed into a wide range of food products
prior to shipping to domestic (78%) or international consumers (22%). It is an
important industry in all provinces and generally ranks among the top three
manufacturing sectors in terms of shipments and jobs. Figure 8.1 shows the distribu-
tion of food and beverage manufacturing establishments in each province. Establish-
ments in Ontario and Quebec accounted for 56% of the total number of establishments.
Ontario accounts for 40% of shipments, Quebec for 22% and Alberta for 13%.
According to the North American Industry Classification System (NAICS)
(Statistics Canada 2004), the FBI is divided into sub-sectors according to simila-
rities in their input materials and the production processes involved. The food
processing sub-sectors are: animal food, grain and oilseed, sugar and confectionery,
fruit, vegetable and specialty food, dairy products, meat products, seafood products,
bakeries and tortilla, and other food. Beverage processing sub-sectors are: soft drink
and ice manufacturing, breweries, wineries, and distilleries.
The food and beverage industry uses a wide range of technologies to achieve
two primary objectives: (1) to carry out the desired processing (e.g., bread
9 1 6
(13%) (8.5%) (4%) (3.7%)
(32%) (25%)
(6.4%)
Fig. 8.1 Distribution of the Canadian food and beverage industry (Statistics Canada 2002a)
8 Development of Eco-efficiency Indicators to Assess the Environmental Performance 167
production) and (2) to stabilize foods and beverages so they will have a longer
shelf life (e.g., milk pasteurization). The manufacturing steps and processes for
most foods and beverages are well known and usually fall into the following
categories:
l Preparing raw materials (washing, cutting, mixing, homogenization)
l Utilizing heat (sterilization, pasteurization)
l Utilizing cold (refrigeration, freezing)
l Removing water (drying, evaporation, pressing, filtration)
l Modulating product composition (pH, salts, sugars, preservatives, smoking
fermentation)
l Modulating product environment (dissolved oxygen, modified or controlled
atmosphere, active packaging)
l Separating/concentrating the components of agricultural products (extraction,
membrane, distillation)
As in all other sectors, food and beverage processing plants are required to meet
various environmental performance standards, which may be critical for competing
on the world market. In manufacturing food products, the food and beverage
industry uses a significant amount of resources (raw agricultural products or
ingredients, energy, water). It also generates gaseous emissions, liquid wastes and
solid organic residues. While most packaging waste is generated at the consumption
level, almost all of it enters the system at the processing stage. Five issues (or
environmental loads) have been identified for the development of eco-efficiency
indicators in the food and beverage processing industry:
l Energy use
l Greenhouse gas (GHG) generation
l Water use
l Waste water production (e.g., effluents)
l Packaging waste generation
The indicators are based on the concept of eco-efficiency, which is a widely
recognized concept in the manufacturing industry and is often used to help
companies characterize and meet both environmental and economic objectives
(Verfaillie and Bidwell 2000) Eco-efficiency is defined as a process during which
goods or services of greater value or greater quantity are produced using fewer raw
materials, and less water and energy, thereby reducing natural resource depletion
and pollution (NRTEE 2001). The indicators essentially compare the environmen-
tal factors or “loads” to the quantity of products manufactured. While this actually
provides an intensity rating (which is the inverse of efficiency), the use of a
common denominator (physical production unit or value) facilitates a comparison
within each sub-sector for each of the five issues of interest. The indicator concepts
are evaluated and validated using data from past surveys (e.g., Annual Survey of
Manufactures for 2002; Statistics Canada 2002a), and the Annual Industrial
Consumption of Energy Survey for 2002 (Statistics Canada 2002b) and Industrial
Water Use Survey (Statistics Canada 2008d).
8.3 Energy Consumption and Greenhouse Gas Emissions
The food and beverage industries (FBI) consume a significant amount of energy for
processing most of what Canadians eat and drink. This energy consumption has a
significant impact on the environment, whether direct (impacts generated onsite) or
indirect (impacts generated during energy production, and its conversion and
transportation to site where consumed). For example, a direct impact would be
when natural gas is burnt onsite in an industrial oven, or in the case of an indirect
impact, when the company uses an electric oven with electricity coming from a coal
burning power plant. The burning of fossil fuels results in the emission of green-
house gases (GHG) and other residues.
According to statistics on energy consumption in the Canadian industry, pub-
lished by the Canadian Industrial Energy End-use Data and Analysis Centre1
(CIEEDAC 2008), the Canadian FBI consumed around 100,700 TJ (terajoules or
millions of megajoules) in 2002, which is about 4% of the energy consumption of
all Canadian manufacturing sectors. This is enough energy to power approximately
2.5 million Canadian households for 1 full year. The left side of Table 8.1 presents
Table 8.1 Energy use and greenhouse gas emissions from manufacturing industries and agricul-
ture in Canada in 2002
Energy use GHG emissions
Terajoules %a Thousands/tons %a
CO2 equivalent
Food manufacturingb 88,765 3.5 3,477 3.3
Beverage manufacturingb 11,975 0.5 517 0.5
Total, Food and Beverage manufacturing (FBI)b 100,740 4 3,994 3.8
Pulp and paper manufacturingb 830,779 33 9,888 9.5
Total, manufacturing industriesb 2,515,928 100 103,911 100
Agriculturec 205,655 – 52,000 –
Canadac, d 9,669,768 – 720,000 –
a
Since data sources and accounting methods differ, precise comparisons are not recommended
between manufacturing and other data, and are presented for information only. CIEEDAC data is
deemed more comprehensive and includes energy and GHG emissions from waste, biomass, etc.,
which can influence significantly total energy use in some manufacturing sectors
b
Manufacturing data from CIEEDAC (2008), both for energy and GHG
c
Energy use data (Statistics Canada 2003, 2004, 2007); GHG emission data from Environment
Canada’s GHG national inventory reports (Environment Canada 2007; Environment Canada
2004) GHG for Agriculture excludes combustion-related emissions, which are a small percentage
of total agriculture emissions
d
Canada energy use is the total net energy supply in Canada
1
CIEEDAC has developed and maintains a comprehensive database on energy, and GHG from
information supplied by Natural Resources Canada, and is produced by Statistics Canada, among
others. It should be noted there are some discrepancies in official publications by Statistics Canada
or Environment Canada (for more information on issues concerning energy accounting and GHG
calculations, see Nyboer 2008a, b).
details on energy use in the FBI as compared to other manufacturing industries, as

well as energy used by the agricultural sector and the total net energy supplied in
Canada.
In 2002, around two thirds of FBI energy needs were supplied by natural gas, a
non-renewable, fossil fuel resource, and more than 20% by electricity (CIEEDAC
2008), although this figure may vary significantly depending on the activity sector
within the FBI. The FBI sub-sectors can be distinguished by the raw materials they
process into food products; the FBI is a very diverse industry, manufacturing a
broad range of different products using specific processes and technologies with
various energy needs.
Energy is an important input cost for the FBI, typically ranking third after raw
materials and labor costs. Energy accounts for a significant share of FBI production
costs, up to 16% for the corn milling industry, around 10% for rendering and meat
processing, rice and malt processing, distillery and brewery sub-sectors (Navarri
et al. 2001). Furthermore, it is estimated that 40% of the value of processed food is
added through energy-intensive manufacturing (US Environmental Protection
Agency 2007), such as process heating and cooling systems. Figure 8.2 shows the
typical energy uses in a FBI plant. It is important to note that most of the energy
demand is necessary to maintain food safety. Thus, FBI is very sensitive to rises in
energy prices, especially since profit margins are often low, and increased energy
Building heating,
Energy for processes
ventilation and lighting
Recycled energy (combustible by-products, heat)
Recycled energy
Steam, Energy use in
heat processes
7.5% 56.9%
Energy Concentration
Generation of Energy conversion Crystallization
energy/services distribution 71.6% Drying/evaporation
100% 88.9% For heating 53% Distillation
For refrigeration and Freezing
Generated steam Steam system
Purchased electricity Electrical cabling cooling 6% Melting
Purchased fuels Fuel system Motors 11.8% Mixing
In-plant Grinding
transportation 0.7% Packaging
Other 0.3% Energy storage
11.1% 9.8% Waste handling
14.4%
Energy losses Energy losses ?
Equipment inefficiency
Combustion losses System, valve, trap (motors, drag, heat Losses through heat
in boilers losses dissipation) dissipation, exhaust
gases, by-products
Fig. 8.2 Diagram of typical energy flows in a food manufacturing plant (Adapted from US
Department of energy 2004)
efficiency, rather than just energy reduction, is the main pathway to a sustainable
food industry.
Energy sources are numerous but most are of the non-renewable type, whether
they are fossil fuels or electricity produced using coal, heavy fuel or nuclear sources.
Apart from the fact that the price of all fossil fuels is skyrocketing and its extraction
and refining generate different types of pollution, one of the most concerning side
effects of using energy coming directly or indirectly from fossil fuels is the genera-
tion of greenhouse gases (GHG) that contribute to global warming.
There are typically three main sources of direct – onsite – GHG emissions in the
FBI.2 Fossil fuel combustion (e.g., in boilers and ovens) is the major source,
accounting for up to 90% of total emissions in a plant heavily relying on these
types of energy, emitting mainly CO2 (and other pollutants contributing to GHG,
such as N2O; or not contributing, such as soot particles). Another source is
refrigeration and freezing units using hydrofluorocarbons (HFCs), which can leak
during a system’s lifespan. Despite the low volume of such leaks, their global
warming potential (GWP) is a hundred to a thousand times higher than that of CO2,
meaning the release into the atmosphere of a few kilograms of most refrigerants
may have a similar contribution to global warming as a ton of CO2. Thus, they may
account for a significant share of total GHG emissions in plants with large needs
for cold processing or storage (e.g., frozen foods, dairy products, meat, and
seafood). A third source of emissions is closely related to biomass and organic
wastes, and can be either solid or liquid. A plant’s wastewater treatment system
using anaerobic digestion emits methane, which should be inventoried as a GHG
(because 1 kg of methane has a similar effect as 21 kg of CO2), if not captured to
fire in a boiler (methane when burned is chemically transformed into CO2 with a
GWP of 1). In addition, solid biomass (e.g., spent grains from distilleries, brew-
eries residues, agriculture wastes) may be land filled, composted (onsite or offsite)
or burned by some FBI plants for energy valorization, all of these emitting various
levels of GHG.3
Unfortunately, the two latter sources of emissions (refrigerants and organic
wastes) are still difficult to estimate accurately at the sectoral or FBI level, as
reliable statistics, indeed aggregate, are available only for fossil fuel related GHG
emissions. At FBI production sites in 2002 and 2005, there were 3,994 and 4,020 kt
of CO2-equivalent GHG produced, respectively (CIEEDAC 2008), that is, 3.7% of
total Canadian industry emissions.
Demand for energy in the FBI is expected to grow due to increased demand for
shelf-stable products, individual ready-to-serve or quick-to-prepare meals, and
processed fresh food. This makes energy a particularly complex challenge for the
FBI given that the industry is often reluctant to changes because of its desire to
2
Note that the consumption of electricity does not emit GHG directly.
3
According to GHG inventory and reporting standards from the Intergovernmental Panel on
Climate Change, GHGs from biomass combustion do not have to be reported in national inventory
since the biomass is carbon neutral, provided it is consumed in a sustainable way (i.e., at a rate not
higher than one needed for its renewal).
maintain high product quality and its obligation to ensure product safety. The
underlying intent of the following eco-efficiency indicators is to provide more
comprehensive – disaggregated – insight on energy use and consequent GHG
emissions through reporting of sub-sector and regional specificities related to
activity levels. They also aim to follow over time energy efficiency improvements,
as well as shifts to energy sources emitting less GHG.
The two indicators were developed as a way to provide establishments with a
measure of their “environmental impact” on the quantity of energy used (ECI), as
well as a measure of the effect of their GHG management practices (GHGEI).
By comparing with different agreed to targets (which should be related to the
sustainable capacity of the natural habitat), these indicators could also give local,
provincial, and national authorities a more adequate picture of the total energy
demand/GHG generation as well as the relative impact of each stakeholder, helping
governments take appropriate actions to sustain human and economic growth.
8.3.1 The Indicators: Energy Consumption Intensity

and Greenhouse Gas Emission Intensity
In order to provide information about these environmental concerns, two indicators

have been developed: the energy consumption intensity (ECI) indicator and the
greenhouse gas emission intensity (GHGEI) indicator, the former expressing the
ratio of the amount of energy used per dollar of manufactured goods produced4
(MJ/$) and the latter expressing the ratio of the amount of GHG emitted (expressed
in CO2-equivalent) per dollar of manufactured goods produced (kg CO2-e/$),
respectively:
Total energy consumed Total GHG emissions

ECI ¼ GHGEI ¼
Value of production Value of production
These intensity indicators reflect how much energy is used (ECI) and the volume
of GHG emissions produced (GHGEI) per dollar of product sold, and are inversely
proportional to efficiency performances: the larger the indicator, the lower the
efficiency and the environmental performance.
Since the FBI manufactures a broad range of highly diversified products (often
tightly linked to regional agricultural production characteristics), and because the
industry involves various scales of production (from small facilities to very large
4
As defined by Statistics Canada in the Annual Survey of Manufactures (ASM) protocol: “The
production is measured by the value of shipments of goods of own manufacture, that is the selling
value of goods made by reporting establishments, excluding transfers into inventory and consign-
ment sales, shipping charges by common or contract carriers, discounts and returns, federal and
provincial sales taxes and excise duties and taxes, sales of goods purchased for resale.”
processing plants), the indicators should reflect these peculiarities. Therefore, in

order to maintain consistency within food industry sectors and to make valid
regional comparisons, establishments were grouped according to similar character-
istics: same sector of activity, same region, and same size as determined by number
of employees. Table 8.2 details these characteristics.
Both indicators use data from the Statistics Canada 2002a Annual Survey of
Manufactures (ASM5). For each establishment included in the calculations, the
value of production (i.e., value of products sold) and the purchase cost of each
form of energy used are needed. The amount of each form of energy used is
calculated, using appropriate regional energy prices, and converted into a com-
mon unit of energy (megajoule, MJ) using Statistics Canada’s 2002 energy
conversion factors (2004). GHG emissions (CO2, CH4, and N2O) in CO2 equiva-
lents are also calculated for each form of energy consumed, using Environment
Canada (2004a) GHG emission factors and the Intergovernmental Panel on
Climate Change (IPCC 1996) global warming potentials (GWP). HFCs were
not included here, as the ASM does not provide sufficient data to quantify these
refrigerants.
The ECI (or GHGEI) indicator is calculated for each group according to two
complementary methods (Fig. 8.3):
l The first method calculates a global indicator value for the group, using the sum
of all energy used by all establishments included in the group divided by the total
sales value of the whole group. This process gives a good picture of the group as
a whole (i.e., considers the group as one unique establishment). Unfortunately,
this method does not provide information on the performance of any individual
plant compared to the group.
l The second method calculates a median indicator range for the group. First, a
single indicator is calculated for all establishments, which are then ranked by
increasing value. Forty percent of the establishments with the lowest values are
tagged as “better than average,” while 40% with the highest values are tagged as
“worse than average.” The remaining values are tagged as “on average,” and the
lowest and highest values of this “on average” group represent the median
indicator range. This second method provides an indication of the range of
efficiencies within the groupings and is useful for comparison to the global
indicator, because each establishment has the same importance in the calculation
regardless of size. It also involves calculations particularly suited to the presence
of outliers in the groups (e.g., invalid or unrepresentative single indicators).
Lastly, the median is a more robust statistic than the mean in the case of small
groups.
5
More details about ASM method can be found at Statistics Canada Web site. http://www.statcan.
ca/cgi-bin/imdb/p2SV.pl?Function ¼ getSurvey&SDDS ¼ 2103&lang ¼ fr&db ¼ imdb&dbg
¼ f&adm ¼ 8&dis ¼ 2
Table 8.2 Parameters for plant grouping and indicator reporting for energy and greenhouse gas
indicators
Sector (o) and sub-sector (l) short Activity sector (o) and sub-sector (l) definition and NAICSa code
name
○ Grain and oilseed ○ Grain and oilseed milling (3112)
l Flour l Flour milling (311211)
l Malt l Rice milling and malt manufacturing (311214)
l Oilseed l Oilseed processing (311224)

l Breakfast l Breakfast cereal manufacturing (311230)
○ Sugar and conf. ○ Sugar and confectionery product manufacturing (3113)

l Sugar l Sugar manufacturing (311310)
l Cacao conf. l Chocolate and confectionery manufacturing from cacao beans
(311320)
l
Chocolate conf. l Confectionery manufacturing from purchased chocolate (311330)
l
Candy conf. l Non-chocolate confectionery (311340)
○ F&V ○ Fruit and vegetable preserving (3114)

l Frozen F&V l Frozen food (311410)
l Other F&V l Fruit and vegetable canning, pickling and drying (311420)
○ Dairy ○ Dairy product manufacturing (3115)

l Milk l Fluid milk (311511)
l Other dairy l Butter, cheese and dry and condensed dairy product manufacturing
(311515)
l Ice cream l Ice cream and frozen dessert manufacturing (311520)
○ Meat ○ Meat product manufacturing (3116)

l Red meat slaughter l Animal (except poultry) slaughter (311611)
l Red meat l Rendering and meat processing from carcasses (311614)
l Poultry l Poultry processing (311615)
○ Seafood ○ Seafood product preparation and packaging (3117)

○ Bakeries ○ Bakeries and tortilla manufacturing (3118)
l Retail bakeries l Retail bakeries (311811)
l Com. bakeries l Commercial bakeries and frozen bakery product manufacturing
(311814)
l
Cookies l Cookie and cracker manufacturing (311821)
l
Flour mixes l Flour mix and dough manufacturing from purchased flour (311822)
l Pasta l Dry pasta manufacturing (311823)
○ Beverage ○ Beverage manufacturing (3121)

l Soft drinks l Soft drink and ice manufacturing (312110)
l Breweries l Breweries (312120)
l Wineries l Wineries (312130)
l Distilleries l Distilleries (312140)
Region short name Geographical region

○ CA ○ Canada
l AT l Atlantic (Newfoundland and Labrador, Nova Scotia, Prince Edward
Island, New Brunswick)

l QC l Quebec
l ON l Ontario
l PR l Prairies (Manitoba, Saskatchewan, Alberta)
l BC l British Columbia
Size short name Establishment size category

○ All ○ All sizes
l Small l Up to 49 employees
l Medium l 50–99 employees

l Large l 100–199 employees
l Very large l 200 or more employees
a
North American Industry Classification System
174
Median range indicator Global indicator ( ) is a mean value of single plant indicators influenced by each plant’s
represents a range of intensity for production sales. It is a weighted mean.
mid-performing plants in groupings
Global indicator value is always within the distribution of signle plant indicator values. Plants with larger sales
drive the global indicator towards their own indicator value, thus influencing the positioning of global indicator
Range is directly calculated by keeping the
relative to median range.
central porting of the distribution of individual
plant values (*) (see graph below)
Case 1 (see Case 1 graph): Case 2 (see Case 2 graph): Case 3 (see Case 3 graph):
This method gives the same importance to
each plant, whatever its market share •Few plants with large portion of sales •Few plants with large portion of sale -“Worse than average” plants
(production sales) have individual indicators that are highly have individual indicators that are (taken as a group) have a similar
“worse than average” slightly “worse than average” market share and similar individual
•Many plants cumulating a large portion •Production sales are more evenly indicator values as “Better than
of sales have individual indicators that shared between plants but still favors average” ones
are “worse than average” “worse than average” plants
Reverse cases (global indicator below median range) occur if sales are
mainly due to “better than average” performing plants.
Plant A
“Worse than average”
performing plants (40%
of plants)
Mid-performing plants
(20% of plants)
“Better than average”
Indicator value (Intensity)

performing plants (40%

of plants)
Case 1 Case 2 Case 3

Example of Median range showing spreading of
individual plant indicator values Likely interpretations of global indicator and median range relative positioning
(assuming mean value is close to median value of singke plant indicators)
Fig. 8.3 How to read and interpret indicators data

M. Marcotte et al.
8.3.2 Results and Interpretation: ECI and GHGEI
This report provides the national and provincial results of both indicators for the
year 2002 for various manufacturing sectors of the FBI. The main objective of
reporting these indicators is to follow over time how a given sector is increasing or
reducing its intensity (i.e., increasing or reducing its pressure on resource depletion
or GHG production). Consequently, the present report does not discuss any timeline
trend but sets the baseline. Thus, the following discussion aims rather at interpreting
these 2002 indicator results by focusing on the inter-sectoral and inter-regional
distinctiveness of the Canadian FBI, while avoiding as much as possible direct
comparisons of the indicators. The main goal here is to provide objective informa-
tion that enlightens future interpretation of indicator trends.
Results will first present the effect of different employee sizes on each sector for
ECI and GHGEI at the national level. Second, the difference between sub-sectors
will be presented at the national level. Finally, sectors will be compared per region.
All graphs will present the different groupings on the X-axis and the value of the
indicator on the Y-axis. Additional comparisons at the regional level will be possible
by using Figs. 8.4 and 8.5, respectively, to present the importance of each sector (in
terms of sales) in each province and the proportion of each energy source used
in each region.
Unless otherwise indicated, the ECI indicator is expressed in megajoules per
dollar in 2002 (MJ/$), and the GHGEI indicator in kilograms of CO2 equivalent per
dollar in 2002 (kg CO2e/$). Establishments’ groupings are reported using the short
4.0
Median Range (regardless of the size)
Median Range
3.5
Global indicator (regardless of the size)
Global indicator
3.0
2.5
ECI (MJ/$)
2.0
1.5
1.0
0.5
0.0
All
Medium
Large
All
Very large
All
Small
Medium
Large
Very large
All
Small
Very large
All
Small
Medium
Large
Very large
All
Small
Medium
Large
Very large
All
Small
Medium
Large
Very large
All
Very large
Grain & Sugar & F&V Dairy Meat Seafood Bakeries Beverage
Oilseed Conf.
Fig. 8.4 Energy consumption intensity (ECI) as a function of activity sector and size calculated
for year 2002
0.28
0.26 Median range (whole sector)
Median range
0.24
Global indicator (whole sector)
0.22
0.20
GHGEI (kg CO2e/$)
0.18
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
Whole sector
Flour
Malt
Oilseed
Breakfast
Whole sector
Sugar
Cacao conf.
Chocolate conf.
Candy conf.
Whole sector
Frozen F&V
Other F&V
Whole sector
Milk
Other dairy
Ice cream
Whole sector
Red meat slaughter
Red meat
Poultry
Seafood
Whole sector
Retail bakeries
Com. bakeries
Cookies
Flour mix
Pasta
Whole sector
Soft drink
Breweries
Wineries
Distilleries
Grain & Oilseed Sugar & Conf. F&V Dairy Meat Bakeries Beverage
Fig. 8.5 Greenhouse gas emissions intensity (GHGEI) as a function of activity sector and sub-
sector calculated for year 2002. There is no sub-sector in the seafood sector
names listed in Table 8.2. Data from some sub-sectors and some plant sizes could
not be published due to confidentiality, while the “animal food” sector and “other
food” manufacturing sectors are not covered within this indicator set. In the
following, the median range indicator will be referred to as median range, median,
or typical representative plant indicator.
8.3.2.1 National Results and Interpretation: ECI and GHGEI
Figure 8.4 provides an overall picture of ECI indicator values across Canada for all
sectors reported, with or without regard to size of establishments.
ECI, Sectoral Features Regardless of Size of Establishments
Looking first at the median range indicators without regard to size of establishments
(“blue-filled bars” in Fig. 8.4), two sets of sectors stand out when it comes to energy
eco-efficiency. In the sugar and confectionery, dairy, meat, and seafood sectors a
typical representative plant shows an ECI of around 0.75 MJ/$, while other sectors
have a median range ECI of around 1.5 MJ/$. The higher values for grain and
oilseed, fruit and vegetable (F&V), bakeries, and beverage sectors are hardly
surprising, given that energy makes up a high proportion of the production cost in
these sectors, which generally make extensive use of energy-intensive operations,
such as evaporation, concentration and drying, cooking and baking, and process
heating and freezing for the fruit and vegetable sector.
Looking now at sector global indicators (“black dots” in Fig. 8.4), the observed
values reveal that the grain and oilseed, and sugar and confectionery sectors, and to
a lesser extent the fruit and vegetable sector, have far higher global intensity (two to
four times higher than those of dairy, meat and seafood sectors). Comparing global
with median range calculation methods for all three sectors (i.e., grain and oilseed,
sugar and confectionary, and fruit and vegetable) shows that global intensity is
significantly higher than each sector’s own median range intensity, especially for
sugar and confectionery (global intensity is 136% higher than upper boundary of
median range) and grain and oilseed (64% higher).
The situation described above highlights the fact that the biggest players (in
terms of sales) within these sectors are energy-intensive plants, although not
enough to shift up the median range of the whole sector. This particular situation
reveals that within a group the “worse than average” category is accountable for a
lot more than 40% of sales (even though it accounts for exactly 40% of plants), thus
shifting the global indicator far above the median range indicator. Conversely, if
most sales are made by “better than average” plants, the global indicator will then
be shifted below the median range indicator. Moreover, if the proportion of “better
than average” and “worse than average” sales are approximately equal, then the
global indicator will stand very close to, if not within, the median range indicator.
The situation where the global indicator is far above the median range reveals a
potential gain in energy efficiency for the major players of these sectors.
It is worth noting, however, that this may be due to a number of reasons. First,
larger plants (in term of sales) within a sector are generally more automated than
smaller plants and therefore may require more energy to operate their higher capital
intensity operations, be it for production (machine feeding, conveyors, etc.) or
sanitization/food safety and quality standards to satisfy stringent buyer require-
ments. Second, larger plants often sell products that are more commoditized, less
differentiated, and therefore of lower value per volume sold. For example, con-
sumers are often willing to pay a premium for hand-made specialty bread from a
local bakery instead of “industrial” bread from one of the big Canadian bakeries.
The fact this bread is hand-made may surprisingly reduce the quantity of energy
used in production since some operations (e.g., machine feeding, emptying and
cleaning, packaging, etc.) are not automated, whereas specialty bread would bring
up the price, tending to reduce ECI value of the local bakery.
The evidence of few plants having low energy efficiency within a sector, while
representing a high share of sales, is also observed to a lower extent in the fruit and
vegetable sector, and to a much lower extent in the dairy and meat sectors. This is
not the case however in the seafood, bakery or beverage sectors. However, it is
worth noting there is no sector where the major players (in terms of production)
have a “better than average” ECI, which would have shifted the global indicator
below the median range.
ECI, Sectoral Features with Regard to Establishment Size
The effect of establishment size is also reported for all FBI sectors (white bars and
red dots, Fig. 8.4). This graph allows easy identification of employment size
categories where significant improvements can be achieved, at least by some of
the larger selling processors within most size groups. For instance, in the fruit and
vegetable sector, the median range ECI is similar for “small,” “medium” and
“large” plants but significantly higher for “very large” ones. However, this trend
is not confirmed by the corresponding global indicator because of the medium size
category, which is the only one in the sector to show a global ECI (red dot) higher
than the median range (white bar). Consequently, the big sellers of medium size
plants should be investigated to see why they are using more energy than their
peers. Furthermore, any plant in the very large category should also be investigated
for energy efficiency improvements in order to achieve a typical intensity in
line with other categories (i.e., between 1.0 and 1.6 MJ/$, a kind of “average”
benchmark).
Within the meat sector, the median range is not dependent on plant size and has a
value close to 0.75 MJ/$. However, all but the very large size categories include
plants with potential energy efficiency improvements because the biggest sellers in
each group size have global indicators significantly above their respective median
range. Similar behavior of the median range indicator prevails in the seafood sector.
As for the global indicator, the very large category only depicts this situation,
though a slight gain only in efficiency seems achievable. Thus, very similar values
for all median ranges in this sector, together with several global indicators ranking
inside – or very close to – the median range, might depict a situation where similar
performance is already achieved within most of the sector, probably because similar
best practices have been globally deployed. This situation is also observed within
the beverage sector (although only one size category is provided), and partly within
the fruit and vegetable sector. The dairy sector, as well as the grain and oilseed, and
sugar and confectionery sectors, show that some plants could achieve improve-
ments whatever the size category reported.
The case of the bakeries sector is singular in this study, in the sense that it is the
only sector to display a similar trend for both median range and global calculation
methods. Both methods show that smaller plants are more efficient energy users
(excluding the very large category). But a finer analysis shows a tendency towards
medium to very large establishments with a high share of sales to a downward shift
of global ECI towards the lower boundary of the median range. This means that
these plants tend to perform better than a typical plant in the category; and that they
could have implemented eco-efficiency measures or better management practices.
Lastly, the commonly held belief is undermined in that the biggest establish-
ments are more eco-efficient because of their capacity (financial resources, know-
how) to be proactive in environmental management. Indeed, whatever the sectors
displayed in Fig. 8.4, the ECI of a typical representative plant does not decrease
clearly or significantly for those with more employees. At least this can be observed
through the global ECI trend (e.g., meat sector from large plants, and bakeries
sector from very large plants), but as already explained only a few plants in the size
category are included.
ECI, Sub-sectoral Features Regardless of the Size of Establishments
A study of the effect of sub-sectors on ECI was done and the following conclusions
surfaced.6 Within three of the eight sectors studied (grain and oilseed, sugar and
confectionary, and beverage sectors), there are significant differences between sub-
sector ECI values. This may have an impact on large data spreading in a sector and
influence sector conclusions. The malt and the oilseed sub-sectors have the highest
ECIs within the grain and oilseed sector,7 about twice as high as that of flour and
breakfast sub-sectors. Sugar refining plants are the driver of the whole sugar and
confectionery sector’s energy intensity; whereas cacao and chocolate confectionery
sub-sectors show a very low intensity (as low as a typical dairy, meat, or seafood
plant). There are few differences within the dairy sector, as well as the meat sector.
Retail and commercial bakeries are slightly more energy-intensive than cookie
and cracker manufacturing, flour mix and dough manufacturing, and dry pasta
manufacturing sub-sectors within the bakeries sector. Within the beverage sector,
winery and distillery sub-sectors clearly stand out, the first being the sector’s least
intensive energy user, and the second posting a very large median range, which may
be explained by a large spreading of single plant ECIs.
Greenhouse Gas Emission Intensity
The total amount of GHG a plant can emit depends on the quantity of energy it
consumes as well as on the type(s) of energy used. Obviously, some energy types
are cleaner than others: consumption of electricity does not emit GHG at all, while
some fossil fuels emit less GHG during combustion than others to provide the same
amount of useful energy (e.g., natural gas emits around 1.5 times less GHG than
light fuel oil to provide 1 MJ of steam in a boiler). Thus, for a given quantity of total
energy consumed, the mix of energy types involved will influence the amount of
GHG emitted.
Figure 8.5 provides Canada-wide details on the GHGEI median range indicator
at the sub-sector level, as well as global and median range indicator at the sector
level (global indicator at sub-sector level could not be published for confidentiality
reasons). With regard to national GHGEI results at sector level (Fig. 8.5), displayed
patterns are fairly similar to those of ECIs (Fig. 8.4). It means that the amount of
6
See technical supplement for detailed results.
7
Data on two other subsectors of the grain and oilseed sector, i.e., wet corn milling (or ‘corn
milling’) and fat and oil refining and blending, cannot be published because of confidentiality.
GHG produced per quantity of energy used was not very different from one FBI
sector to another in 2002. Hence, analysis and arguments provided above for ECIs
should hold true for GHGEIs, and therefore will not be repeated here.
As an adjunct to the GHGEI indicator, calculation of sector ratio GHGEI/ECI for
the global indicator (amount of GHG emitted for each MJ of energy consumed)
provides useful information complementary to the whole sector GHGEI indicator
(Fig. 8.5), allowing sector comparisons independent of the dollar value of sales.
Furthermore, it provides an indicator on how clean the energy consumed within a
sector is, or how much effort this sector is putting to reduce GHG emissions once
energy needs have been optimized. Improvements are mainly technological and can
be looked for in different areas, for example, process changes in electric techno-
logies, and shifts in energy supply toward cleaner energy or energy mix, etc. Sector
results of this calculation are as follows, in increasing order of GHG emitted per
unit of energy consumed (g CO2e/MJ): bakeries (36.7), dairy (37.6), meat (37.7),
seafood (39.6), grain and oilseed (40.8), sugar and confectionery (42.5), beverage
(43.1), and fruit and vegetable (45.0).
It is worth noting that differences between sectors are slight; bakeries compared
to fruits and vegetables show both a 10% difference with respect to mean of dataset
(41 10%). Thus, no sector clearly distinguishes itself, whether as a “cleaner
energy” user or not. This is somewhat surprising since one would have expected
sectors heavily using electric processes (e.g., cooling, refrigeration, and freezing) to
clearly stand out. Several factors are responsible for these similar results. First,
the expected difference is likely to stand out more clearly at the sub-sector level
(e.g., frozen food requires mostly electricity while fruit and vegetable canning uses
more fossil fuel, yet both are in same sector). Second, any sector has to comply with
food safety standardized procedures requiring pasteurization or sterilization pro-
cesses and/or strict sanitation processes that might involve a significant amount of
thermal energy from fossil fuels. Third, plants relying on one unique energy source –
be it clean or not – are now scarce. Fourth, regional distribution of the sector
displayed in Fig. 8.6 reveals that Alberta, Ontario and Quebec are each contributing
approximately 30% to national production. Figure 8.7 shows that their respective
energy grid is very different. This tends to level out the national ratio. Lastly, it
should be noted, specifically for the seafood sector, that fuel consumed by fishing
vessels is often accounted for in the manufacturing process because fish processing
generally starts on board.8 It tends to counterbalance the electric process energy
used at the processing plant for refrigeration and freezing operations.
The seafood sector, while the most GHG-efficient of sectors, is not the cleanest
energy user, which is due to particularly marked regional discrepancies related to
both the kind of seafood processed and energy source availability.
8
National data on energy consumption and GHG emission from seafood sector in 2002 are
available from CIEEDAC and give a ratio of 20.2 g CO2e/MJ, well below the value calculated
here, whereas other available sector ratios are in accordance. It is likely that CIEEDAC raw data
does not account for fishing vessel fuel consumed.
Relative contribution by province to Grain & Oilseed

Canadian production by food and Sugar & Confectionery
beverage manufacturing
Fruit & Vegetable
sectors, 2002.
Dairy
Meat
Seafood
Bakeries
Beverage
Other
70%
28%
64%
31%
Fig. 8.6 Overview of provincial characteristics of Canadian food and beverage industry produc-
tion (Statistics Canada, Annual Survey of Manufactures 2002a). Contribution is measured in terms
of sales. Each sector representation is not necessarily equal to 100% across Canada due to
confidentiality of data in certain provinces; data below 4% is not represented. Atlantic Provinces
data is grouped together. It excludes animal feed. “Other” stands for snack foods, coffee and tea,
flavoring syrup and concentrate and all other manufacturing sectors
The fruit and vegetable sector is the worst performer; the sugar and confection-
ery sector and the beverage sector are both high ranking, which is less of a surprise
given the high GHGEIs that can be reached in the sugar sub-sector and the
distilleries sub-sector, respectively. As noted, regional particularities (also dis-
cussed later) induce these results.
The grain and oilseed sector, while the least GHG-efficient (and least energy-
efficient) of sectors, is mid-ranking in the GHGEI/ECI ratio. This may be explained
by a significant amount of energy consumed by the sector in flour milling and
breakfast cereal manufacturing sub-sectors in the form of electricity or low-emitting
process energy, as in the bakeries sector. The bakeries sector, which as mentioned is
an intensive energy user and GHG emitter sector, shows the lowest GHGEI/ECI
ratio of the FBI. Energy in this sector is mainly required for baking processes in
well-controlled gas or combined-energy ovens, which emit lower quantities of
GHG. The bakeries sector is a rather “dry” sector and uses only a small amount
Atlantic Quebec
Confidential (Diesel Heavy Fuel Oil
fuel oil, light fuel oil Natural Gas 0.7%
and kerosene, Still 52.5%
gas and petroleum Electricity Other*
coke, LPG and gas 31.9% 0.1%
plant NGL, Coal) Confidential (Diesel
24.5% fuel oil, light fuel oil
and kerosene, Still
gas and petroleum
coke, LPG and gas
plant NGL, Coal, Coke
Electricity and coke oven gas)
Natural Gas 38.3% 8.4%
Heavy Fuel Oil 11.8%
31.8% (*) «Other» includes steam and waste
fuels from the cement industry
Ontario Prairies
Heavy Fuel Oil

0.2%
Other*
Natural Gas Other* Natural Gas 5.5%
1.7% Confidential (Diesel 67.0% Confidential (Diesel
65.6%
fuel oil, light fuel oil
and kerosene, Still
and kerosene, Still
gas and petroleum
gas and petroleum
coke, LPG and gas
coke, LPG and gas
plant NGL, Coal,
plant NGL, Coal,
Coke and coke
Coke and coke
Electricity oven gas)
Electricity oven gas)
21.1% 6.4%
24.7% 7.8%
(*) «Other» includes steam and waste (*) «Other» includes steam and waste
fuels from the cement industry fuels from the cement industry
British Columbia and Territories Alberta

Natural Gas Other*
72.0% 7.6%
Confidential (Diesel Confidential (Diesel
fuel oil, light fuel oil and kerosene, Still
Natural Gas and kerosene, Still gas and petroleum
coke, LPG and gas
76.1% gas and petroleum plant NGL, Coal,
coke, LPG and gas Coke and coke
oven gas)
plant NGL, Coal, Electricity 4.8%
15.6%
Coke and coke
oven gas) (*) «Other» includes steam and waste
fuels from the cement industry
6.8%
Electricity
17.2% Manitoba
Natural Gas Confidential (Diesel
61.1% fuel oil, light fuel oil
and kerosene, Still
gas and petroleum
coke, LPG and gas
Saskatchewan plant NGL, Coal,
Coke and coke
oven gas)
Natural Gas
4.7%
25.1%
Confidential (Diesel
fuel oil, light fuel oil Electricity
and kerosene, Still 34.2%
gas and petroleum
coke, LPG and gas
plant NGL, Coal)
33.3%
Electricity
41.5%
Fig. 8.7 Share in each region and province of energy sources used in 2002 by manufacturing
industries in the category, “other manufacturing” (Natural Resources Canada 2007)
of steam; other energy usages in this sector are supplied by electricity, which is not
directly emitting GHG.
8.3.2.2 Provincial Results and Interpretation: ECI and GHGEI
Provincial indicators are available at the sector level only and cannot be broken
down by size of employment. Not all sectors are necessarily reported for each
region, because of a very low number of plants within a specific sector in a given
region (e.g., seafood sector in Prairie Provinces) and/or to respect the confidential-
ity of responding plants under the Statistics Act (e.g., fruit and vegetable sector in
the Atlantic region).
Energy Consumption Intensity
Figure 8.8 illustrates the regional effect on the global and median ECI indicators.
The main regional features are presented, often with reference to national results
available.
The global ECI peak of 3.8 MJ/$ for the sugar and confectionery sector in British
Columbia (red dot) is twice as high as the sector scores in Ontario. This result is
symptomatic of a sector with very energy-intensive activities and is delivered by
only a few big plants. As already mentioned, the sugar refinery sub-sector is
responsible for a high global ECI within this sector: a typical Canadian sugar
refinery can display intensity as high as 4.2 MJ/$, whereas typical confectionery
Canadian plant intensity does not exceed 1.2 MJ/$ (data not shown). However, it
should be noted that a typical British Columbian plant within the sugar and
confectionery sector (white bar) displays the lowest Canadian ECI in this sector.
Moreover, the British Columbian global ECI indicator in the meat sector is the
highest in Canada. It indicates that energy efficiency improvements could be
achieved by local meat plants with a large share of sales, reaching values close to –
or below – 1 MJ/$. These plants are not numerous because the provincial median
4.0
Median Range
3.5 Global indicator
3.0
2.5
ECI (MJ/$)
2.0
1.5
1.0
0.5
0.0
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Grain & Oilseed
Meat
Grain & Oilseed
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Grain & Oilseed
F&V
Dairy
Meat
Bakeries
Meat
Seafood
BC PR ON QC AT
Fig. 8.8 Energy consumption intensity (ECI) as a function of regions and activity sector calcu-
lated for year 2002
range ECI is typical of the national values. The same conclusions prevail for the
bakeries sector as well as the fruit and vegetable sector, compared to provinces like
Ontario and Quebec, even though the latter sector is well performing compared to
the national values. Lastly, the dairy sector and the seafood sector are both well
performing in British Columbia.
The Prairies’ grain and oilseed sector’s performance is at a level similar to that of
Ontario and Quebec, and of Canada as a whole. The meat sector in the Prairies is
one of the main food sectors in these provinces; Alberta is one of Canada’s leaders
in red meat processing. However, the Prairies’ meat sector displays the highest
typical plant ECI across Canada, twice as high as that of Ontario, despite a more
energy efficient production by some of the largest processors (in terms of sales); this
is illustrated by the global ECI as close to the median range bottom boundary.
The good performance of the Ontario sugar and confectionery sector (global
indicator), as compared to British Columbia level, is explained by a higher diversi-
fication of the sector, with a larger share of low energy-intensive activities such as
confectionery manufacturing. Establishments in the fruit and vegetable sector typi-
cally reflect higher intensity in Ontario than in British Columbia or Quebec. The
global ECI of the sector is nevertheless lower than the median range (though close
to), which means that a major part of the fruit and vegetable processing in Ontario is
performed by plants slightly more energy-efficient than a typical representative one.
Lastly, this province is among the most eco-efficient regarding energy use in the
meat (lowest median range across Canada), seafood, and bakeries sectors.
Quebec performs relatively well in the grain and oilseed and the fruit and
vegetable sectors since the province shows the lowest global ECI across Canada
for both sectors.9 For the former though, some higher intensity plants could still
make improvements, whereas the fruit and vegetable sector reveals no “atypical”
eco-efficiency plant (global ECI stands within the median range). Quebec’s energy
intensity (either global or typical) in the meat sector stands at the national level,
quite similar to Ontario and Atlantic Provinces. Lastly, the dairy sector in Quebec
shows the highest global ECI across Canada, significantly above the intensity
achieved by a typical representative plant of the sector in the province, or in Ontario
or British Columbia, or Canada as a whole.
The Atlantic Provinces’ food industry is highly oriented towards seafood, fruit
and vegetable processing, confectionery, and beverage. Most of Canadian seafood
production and processors are located in these provinces (Fig. 8.6), which conse-
quently “shape” the Canadian seafood sector indicators. British Columbia and the
Atlantic region show quite similar values, a bit higher than that of Ontario, although
this latter is hardly comparable since, as a non-coastal province, its processing
activities are limited.
9
Global ECI at national level for fruit and vegetable sector (2 MJ/$) is higher than any provincial
value of the sector, which means at least one of the non-reported provincial global ECIs (i.e., from
Prairies or Atlantic Provinces) is at least worth 2 MJ/$.
Lastly, it is worth noting that the peculiarity of the large spreading of the ECI
median range in the distilleries sub-sector in Canada as a whole (data not shown,
but relative spreading is similar to that of GHGEI shown in Fig. 8.4) is not observed
in Ontario (data not shown). Ontario’s median range ECI stands between 0.56 and
1.09 MJ/$. It is likely that the cause of this observed national spreading comes from
another province or region.
Greenhouse Gas Emission Intensity
As explained earlier, the energy mix (i.e., proportion of each energy type) influ-
ences the total amount of GHG a plant will emit. Even though each establishment is
free to choose the energy type(s) to use, it still depends on the local availability and
price from suppliers. Figure 8.6 shows the different provincial energy mixes
consumed in 2002 for an aggregate group of manufacturing industries, including
the FBI. Consequently, the GHGEI indicator can be discussed more easily and in
more detail using this regional analysis.
Figure 8.9 illustrates the regional GHGEI indicator results per sector. This figure
shows a computation of the regional global indicator ratio GHGEI/ECI – the
amount of GHG emitted for each MJ of energy consumed – to provide insight on
“how clean” the energy consumed is and the effort made to shift to technologies or
processes using electricity or low GHG-emitting energies.
0.20
0.06
Median Range BC PR ON QC AT
0.18 Global indicator 0.05

GHGEI/ECI (kg CO2e/MJ)
0.04
0.16
0.03
0.14 0.02
GHGEI (kg CO2e/$)
0.01
0.12
0.00
Grain & Sugar & F&V Dairy Meat Seafood Bakeries
Oilseed Conf.
0.10
0.08
0.06
0.04
0.02
0.00
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Grain & Oilseed
Meat
Grain & Oilseed
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Grain & Oilseed
F&V
Dairy
Meat
Bakeries
Meat
Seafood
BC PR ON QC AT
Fig. 8.9 Greenhouse gas emissions intensity (GHGEI) as a function of region and activity sector
calculated for year 2002. Global indicators’ ratio GHGEI/ECI (kg CO2 equivalent of GHG emitted
per MJ of energy consumed) as a function of region
British Columbia seems to have adopted cleaner energy mixes or processes in

the fruit and vegetable, dairy, seafood, and bakeries sectors, as shown by its lower
GHGEI/ECI ratio in these sectors as compared to other provinces. This is not due to
the availability of a particular energy type, as in Fig. 8.6 showing the Prairies,
Ontario and Quebec to have similar or cleaner energy mixes. Plants in British
Columbia from the above-mentioned sectors have probably been proactive in
choosing lower emission energy mixes. Such a shift is not always possible, espe-
cially in older plants where the production and processes are not easily modified to
take a different energy source. This is the case obviously in the sugar and confec-
tionery sector, for which British Columbia displays a high GHGEI/ECI ratio (along
with a huge ECI). In fact, with 47 g of CO2-equivalent emitted per MJ used, it is the
highest value of the province and the third highest value nationwide. Comparison of
the global and median range calculating methods reveals that a few large plants are
responsible for these “below average” performances, especially when pointing out
that the median range of both ECI and GHGEI are among the lowest sector values
attained in Canada. It is very likely a large part of British Columbia’s activity in this
sector is oriented towards sugar manufacturing, which needs extensive amounts of
fossil fuel energy for process operations that cannot be handled economically by
electricity. On the other hand, a very small part of activity in this sector comes from
other sub-sectors (e.g., confectionery manufacturing activities) that are very effi-
cient users of “cleaner” energy sources.
The Atlantic Provinces emit around 30% more GHGs per unit of energy con-
sumed than any other province in the meat sector and twice as much as British
Columbia in the seafood sector. The marked dependence of this region on GHG-
intensive sources of energy, such as heavy fuel oil and refined petroleum products,
is the reason for such difference. As a result, the Atlantic region, which ranked first
in terms of energy-related eco-efficiency (ECI), is last when it comes to GHG
emissions. In the seafood sector, however, Ontario cannot directly compare because
of its activities mainly orientated towards secondary processing operations involv-
ing more fossil fuel thermal energy.
Between the extreme positions of the Atlantic region and British Columbia, with
regard to the GHGEI/ECI ratio, the Prairies, Ontario and Quebec globally rank in
the order of decreasing GHG emission per unit of energy consumed. One exception
is Quebec being over Ontario in the dairy sector, also showing a higher median
range GHGEI. This is likely due to the more diversified and structurally different
dairy industry in Quebec, with more small factories and more atypical processes
(e.g., cheese specialties), implying less rationalized operations (e.g., cleaning and
disinfection), which tends to emit more GHG. The difference observed between
Quebec and Ontario for ECI in the fruit and vegetable sector, Quebec being more
energy-efficient, is still larger when it comes to the GHGEI. This holds true also for
the grain and oilseed sectors in the same two provinces.
Although data cannot be released, the fruit and vegetable sector has significant
activities in all Canadian regions (Fig. 8.6). Given the national GHGEI/ECI ratio of
45 g CO2e/MJ (data not shown), it is likely that the Prairies’ and/or the Atlantic
region’s GHGEI/ECI values are significantly above the national average.
Limitations: ECI and GHGEI
The proposed indicators have several limitations, most resulting from the desire to
provide indicators that report sub-sector, regional and size specificities. First, both
indicators are calculated per value of product manufactured, instead of per volume
as initially intended (Marcotte et al. 2005) due to the limited amount of volume data
in the Statistics Canada ASM database. Second, these new indicators from the
Agriculture and Agri-Food Canada’s series of agri-environmental indicators were
computed for a single benchmark year, 2002, which does not allow for year-to-year
comparison.
8.3.3 Response Options: ECI and GHGEI
As this series of energy and GHG indicators for the FBI is the first of its kind, there
are no benchmark values to compare the indicators with a previous situation or to
assess the effectiveness of any measures already implemented by the FBI or various
governmental institutions. However, the fact remains that measures to improve
energy efficiency and the on-site production of secondary energy (such as steam)
will help reduce both energy consumption intensity and GHG emissions. The
challenge the FBI faces is to reduce its current demand for energy by improving
its efficiency through the adoption of best practices10 without, however,
compromising hygiene or food safety procedures. Improved efficiency would result
in a reduction in the intensity indicators measured here. Companies can take a large
variety of possible measures, the costs of which vary. Low-cost measures, first and
foremost the measurement of consumption and heat flows, and management of the
procedures that require most energy will allow companies to react swiftly to any
problems and to avoid waste. As seen in Fig. 8.2, energy losses are at all operational
stages and account for a large share of the initial energy used, although some of
these losses are physically unavoidable. A more comprehensive analysis of opera-
tional procedures may reveal opportunities requiring varying levels of investment.
By using the process integration approach, for example, a plant’s energy and water
flows can be studied jointly, which is particularly relevant to the FBI, where these
two resources are often closely linked. For example, in the dairy products group,
60% of the water consumed is used for energy-related operations (e.g., to cool or
generate steam). The results of a process integration analysis generally provide a
series of options for optimizing the flow of heat and water in a plant, as a result of
which the best ways to recycle and save energy and water can be identified. Such an
analysis may, for example, confirm the benefits of using an energy cogeneration
system to simultaneously produce mechanical – usually transformed into electricity –
and thermal energy (cogeneration), and also cooling energy (trigeneration).
10
Energetics Inc. and E3M Inc. (2004).
8.4 Water Intake and Water Discharge
The FBI is known for its need for exceptionally large amounts of water, in both
volume and quality, as it uses water both as an ingredient and to carry out numerous
processing operations. Indeed, water is used at almost all stages of processing: as a
heat transfer medium (e.g., as hot water for blanching or steam for heating), as a
carrier (e.g., for transportation of fragile products on a production line), or for
washing, rinsing, cleaning and sanitizing operations. However, water needs vary
significantly in both quality and quantity depending on the kind of products
manufactured and on the process implemented to achieve the desired transforma-
tion. Water quality is of paramount importance for meeting food hygiene and safety
standards when there is a chance of direct or indirect contact with food (Maxime
et al. 2005).
It is estimated that in Canada the FBI withdrew 1,500 million m3 of water in
2005, i.e., almost 20% of the total amount withdrawn by Canada’s manufacturing
industries (Fig. 8.10) and close to 3% of total water intake in Canada (Environment
Canada 2008a), which is equivalent to 600,000 Olympic pools. Establishments
obtained half of their water requirements from public water suppliers and supplied
the other half themselves from surface water systems (e.g., lakes, rivers), ground
water systems (e.g., wells, springs) or even tide water bodies (e.g., estuaries,
bays, oceans), each source needing its own water treatment according to the quality
Petroleum and Chemicals,

coal, 4.7% 6.8% * Others: aggregation of low or confidential data from the following manufacturing sectors:
plastic and rubber products, textile mills, textile products, wood products, fabricated metals,
Others *, transportation equipment, and non-metallic minerals.
14.8%
Primary metals,
20.6%
Total 2005
Manufactures Food,
7778,9 Mm3 17.6%
Beverage and
Paper, 33.4%
tobacco, 2.1%
Millions of cubic metres
0 500 1000 1500 2000 2500
Food 1366.8
Beverage and tobacco 160.6
Plastic and rubber x

Textile mills 19.8
Textile products x
Wood 124.2
2598.3
Paper
Primary metals 1606.2
Fabricated metals x
Transportation equipment 48.7
Non-metallic minerals x
Petroleum and coal 364.8
Chemicals 532.5 X: confidential data
Fig. 8.10 Water intake volumes by the main manufacturing sectors in Canada and share of total
manufactures’ intake calculated for year 2005
of the water withdrawn and based on the required quality standards at each stage
of the process. Of the FBI’s intake volume, 4% was recirculated or reused in
process or cooling systems and 77% was discharged after use as wastewater,
either to public utilities or directly back to the environment, generally after onsite
treatment. The remaining 19% was either incorporated into finished products,
evaporated during processing operations, or processed as part of wastewater
sludge and solid wastes.
The main pollutants of the FBI industry are biodegradable materials and resi-
duals from cleaning agents with no hazardous or acute human toxic pollutants,
although ecotoxicity of some effluents might be of concern in fragile environments
(e.g., a bay or an estuary receiving effluents from fish processing plants [Environ-
ment Canada 2008b]). Prior to discharging wastewater directly to the environment
or to a public sewer, the raw wastewater must be treated to convert these pollutants
into non-polluting chemicals or to abate their load according to regulations. Some
large plants operate their own wastewater treatment system (Fig. 8.11). However,
the build-up of some pollutants and the wastewater treatment plant’s capacity to
properly treat the pollutant loads of the effluents received, particularly during
pollution peaks, may pose ecological concerns in the receiving waters. Chlorides
(e.g., discharged from food salting processes and water softener regeneration)
become toxic to aquatic life at high concentrations; phosphorus and nitrogen
compounds (e.g., from sanitation chemicals) induce the aquatic system’s eutrophi-
cation. A wide unknown also exists regarding the build-up and toxicity in the
environment of residual pesticides (e.g., from fruit and vegetable sector), antibio-
tics, growth hormones, and pathogenic organisms (e.g., from meat and seafood
sectors). It is also worth mentioning that processors often overuse or overdose
sanitation chemicals to ascertain their food safety security margin, thus increasing
the toxicity level of chemicals in wastewater.
Incorporation
into
Evaporation products Recycling
Direct
discharge **
Public system
BOD direct
discharge
INITIAL WATER
WWTS* Discharge to
TREATMENT USE
public sewer
Independent system
BOD sewer
discharge
Recycling
FBI Plant
Boundaries
* Wastewater treatment system (if applicable) Note: losses may also occur at various stages
** Discharge subjected to regulatory acceptance
Fig. 8.11 Primary water and wastewater flow in a food processing plant
The FBI significantly depends on – and therefore contributes to – demand for

water resources, which can have direct and indirect impacts on their availability
and quality, as well as on the ecosystems and economies that depend on them.
In fact, depletion of high-quality water reserves in some parts of the country has
already pushed up industrial water supply costs and placed additional pressure
on public water utilities to find new supply sources (Environment Canada
2004b). The geographic concentration of FBI groups and, for some sectors,
the marked seasonality of their withdrawals (e.g., fruit and vegetable processing
reaches a peak just after the harvest season) increase pressure on the environ-
ment locally and seasonally. Part of the problem with the food-processing
industry’s use and discharge of large amounts of water is that it is often located
in rural areas in which the municipal water treatment systems (i.e., drinking
and wastewater systems) are designed to serve small populations. As a result,
one medium-sized plant can have a major effect on local water supply and
surface water quality. A large food processing plant will typically use more
than 4,000 m3of drinking water per day, which is the equivalent of supplying
12,000 residents.
8.4.1 The Indicators: Water Intake Intensity and Water Discharge

Intensity
Two indicators were developed: water intake intensity (WII) indicator and water
discharge intensity (WDI) indicator, the former expressing the ratio of the amount
of water used by the plant per dollar of manufactured goods produced11 and the
latter expressing the ratio of the amount of water discharged per dollar of manu-
factured goods produced (both in L/$):
Volume of water withdrawn Volume of water discharged

WII ¼
Value of production WDI ¼ Value of production
The WII indicator measures the volume of water withdrawn from the environ-
ment to meet the production needs of the FBI plant, while the WDI indicator
measures the volume of water returned to the environment through the plant’s
different liquid waste streams. Theoretically, the WDI reflects the maximum
amount of water withdrawn in excess. Moreover, this “auxiliary” indicator is a
mandatory step in quantifying the pollution load emitted from the FBI into the
11
The production is measured by the value of shipments of goods of own manufacture, that is,
the selling value of goods made by reporting establishments, excluding transfers into inventory
and consignment sales, shipping charges by common or contract carriers, discounts and returns,
federal and provincial sales taxes and excise duties and taxes, sales of goods purchased for
resale.
public sewage system or directly into the environment (these pollution load indi-
cators will be released in the next report). These intensity indicators are inversely
proportional to efficiency performances: the larger the indicator, the lower the
efficiency and the environmental performance.
The two indicators were developed as a way to provide establishments with a
measure of their “environmental impact” on the water resource as well as a measure
of the effect of their water management practices on the sustainable capacity of the
natural habitat; these indicators could also give local, provincial, and national
authorities a more adequate picture of the total demand/impact on the water
resource as well as on the relative impact of each stakeholder. These indicators
could also be used to foresee the capacity of a region to accept more development,
helping governments take appropriate actions to sustain human and economic
growth.
tightly linked to regional agricultural production characteristics), the indicators
should reflect these peculiarities. Therefore, in order to maintain consistency within
food industry sectors and to make valid regional comparisons, establishments were
stratified into groups having similar characteristics: same sector of activity and
same region. Table 8.3 details these characteristics.
The method employed to calculate both indicators uses data gathered from two
Statistics Canada surveys, the 2005 Industrial Water Survey12 and the Annual
Survey of Manufactures and Logging (ASML).13 Both indicators are calculated
for each grouping by dividing the total of all water withdrawn (or discharged,
respectively) by the sum of the sales value for the whole group. This process gives a
good picture of the group as a whole (considering the group as one unique
establishment) but does not provide information on the performance of any indi-
vidual plant compared to the group.
The proposed indicators have several limitations, most of them resulting from
the desire to provide indicators that report sub-sector and regional specificities.
First, both indicators are calculated per value of product manufactured, instead of
per amount of product manufactured, which would have been a better eco-effi-
ciency indicator (Marcotte et al. 2005) since there were too few available and
reliable data in the Statistics Canada ASML database. Second, these new indicators
(from Agriculture and Agri-Food Canada’s series of agri-environmental indicators)
were computed for a single benchmark year, 2005, which does not allow for year to
year comparison.
12
For more details on the survey, go to http://www.statcan.ca/cgi-bin/imdb/p2SV.pl?Function ¼
getSurvey&SDDS ¼ 5120&lang ¼ fr&db ¼ IMDB&dbg ¼ f&adm ¼ 8&dis ¼ 2
13
For more details on the survey, go to http://www.statcan.ca/cgi-bin/imdb/p2SV.pl?Function ¼
getSurvey&SDDS ¼ 2103&lang ¼ fr&db ¼ IMDB&dbg ¼ f&adm ¼ 8&dis ¼ 2
Table 8.3 Parameters for plant grouping and indicator reporting for water indicators
Sector (o) and sub-sector (l) Activity sector (o) and sub-sector (l) definition and NAICSa code
short name
○ Grain and Oilseed ○ Grain and oilseed milling (3112)
l Flour l Flour milling (311211)
l Malt l Rice milling and malt manufacturing (311214)
l Oilseed l Oilseed processing (311224)
l Breakfast l Breakfast cereal manufacturing (311230)
○ Sugar and Conf. ○ Sugar and confectionery product manufacturing (3113)

l Sugar l Sugar manufacturing (311310)
l Cacao conf. l Chocolate and confectionery manufacturing from cacao
beans (311320)
l Chocolate conf. l Confectionery manufacturing from purchased chocolate
(311330)
l Candy conf. l Non-chocolate confectionery (311340)
○ Dairy ○ Dairy product manufacturing (3115)

l Milk l Fluid milk (311511)
l Other dairy l Butter, cheese and dry and condensed dairy product
manufacturing (311515)
l Ice cream l Ice cream and frozen dessert manufacturing (311520)
○ Meat ○ Meat product manufacturing (3116), made of

l Red meat slaughter l Animal (except poultry) slaughter (311611)
l Red meat l Rendering and meat processing from carcasses (311614)
l Poultry l Poultry processing (311615)
○ Seafood ○ Seafood product preparation and packaging (3117)

○ Bakeries ○ Bakeries and tortilla manufacturing (3118)
l Retail bakeries l Retail bakeries (311811)
l Com. bakeries l Commercial bakeries and frozen bakery product
manufacturing (311814)
l Cookies l Cookie and cracker manufacturing (311821)
l Flour mixes l Flour mix and dough manufacturing from purchased flour
(311822)
l Pasta l Dry pasta manufacturing (311823)
○ Beverage ○ Beverage manufacturing (3121)

l Soft drink l Soft drink and ice manufacturing (312110)
l Breweries l Breweries (312120)
l Wineries l Wineries (312130)
l Distilleries l Distilleries (312140)
Region short name Geographical region

○ CA ○ Canada
l AT l Atlantic (Newfoundland and Labrador, Nova Scotia, Prince
Edward Island, New Brunswick)

l QC l Quebec
l ON l Ontario
l PR l Prairies (Manitoba, Saskatchewan, Alberta)
l BC l British Columbia
a
North American Industry Classification System
8.4.2 Results and Interpretation: WII and WDI
This report presents the national and provincial results of both indicators for the
year 2005 for various manufacturing sectors of the FBI. The main objective of these
indicators is to follow over time how a given sector is increasing or reducing its
intensity (i.e., increasing or reducing pressure on resource depletion or, indirectly,
on aquatic ecosystem pollution); however, the present report does not discuss any
timeline trend but sets the baseline. Thus, the following discussion aims rather at
interpreting the results of the 2005 indicators through focusing on the inter-sector
and inter-regional distinctiveness of the Canadian FBI, while avoiding as much as
possible direct comparisons of the indicators. The main goal here is to provide
objective information that enlightens future interpretation of the indicators’ trends.
Data from the sugar and confectionery sector (except for Ontario), the fruit and
vegetable sector, and partly, the bakeries sector, could not be published due to
confidentiality, while the animal food sector and the other food manufacturing
sectors were not covered within this indicator set.
8.4.2.1 National Results and Interpretation: WII and WDI
Figure 8.12 provides an overall picture of both indicator values across Canada.
The first thing that can be noted is the great variability from one industry group to
the next in terms of both intake and discharge intensity. The meat sector and the
18
16
14
12
10
l/$
0
Grain & Dairy Meat Seafood Bakeries Beverage
Oilseed
WII WDI X: confidential data
Fig. 8.12 Water intake intensity (WII) and Water Discharge Intensity (WDI) as a function of FBI
sectors calculated for year 2005
seafood sector are at two extremes, with WII national values ranging from 3 to
17 L/$ of product sold, respectively.
Because of the important availability of marine water, the seafood industry has
been reputed for its extensive use of water and ice14 for fish conservation, handling,
rinsing and defrosting, and for general hygiene and washing of equipment. However,
uncontrolled procedures are still widely used in this sector (Tchoukanova et al. 2003;
Fisheries and Oceans Canada 2003; European Commission 2006). Even though a
large amount of water is used in this sector for hygiene and process purposes (i.e.,
WII is high), the potential for improvement is also important (WDI is also high) and
could be easily achieved through dry procedures and better water management to
save water, especially during fish washing and area clean up, without compromising
safety and quality standards. This could beneficially lead to wastewater volume and
pollution content reduction, especially because the seafood industry effluents are
typically rich in organic matter, suspended solids, nitrogen and phosphorus nutrients,
and sanitation chemicals that can affect the health of the receiving environment
(Environment Canada 2008b; Gonzalez and Poirier 2003). However, wastewater
from processing facilities is generally untreated except for fine screening before
discharge from the plant, and dissolved pollutants are thus rejected into the receiving
environment (e.g., in harbor’s seawater). Moreover, the legislation, whether federal
or provincial, remains mainly based on old regulations and guidelines from the 1970s.
Since that time, little has been done to assess the impact or to identify suitable
emission standards, whereas other FBI sectors have fallen under increasingly strict
controls (Tchoukanova et al. 2003; AMEC Earth and Environmental Ltd 2003;
Fisheries and Oceans Canada 2003).
The meat sector is rather similar to the seafood sector in the sense that large
amounts of water are required for carcass washing and surface cleaning opera-
tions. This results in large volumes of wastewater rich in nitrogenous organic
matter and suspended solids. However, WII and WDI indicators are much lower
for the meat sector than the seafood sector for several reasons. The coastal
location of the majority of establishments in the seafood sector are in the
Atlantic Provinces and British Columbia, where fewer incentives exist to limit
water intake and especially to recycle water as compared to the meat sector in
regions where water is less abundant. This relative scarcity has led to more
advanced implementation of best management practices in the meat sector (e.g.,
air instead of water thawing, dry cleaning methods), as well as water recycling
opportunities owing to a more widespread use of onsite advanced wastewater
treatment systems. The provincial section below discusses these points in greater
detail.
Intake intensity is also high in the grain and oilseed sector (15 L/$), as this
industry uses a great deal of water and steam during extraction processes (e.g., wet
dehulling, wet milling, malt steeping, oil refining, and deodorization). The sector
14
Seafood sector’s plants can use seawater, either directly in some processes where there is no
contact with food, or after adequate treatment where contact is possible.
also shows that 22% of water withdrawn is not discharged, the result of (WII-WDI)/
WII. Water is actually consumed to some extent through evaporation15 during
dehydration and drying operations in flour processing, malting processes, and
breakfast cereal manufacturing.
The dairy products sector, bakeries and tortillas sector, and beverage sector all
show intermediate WII values of 11, 9, and 12 L/$, respectively. The bakeries
sector includes manufacturing activities requiring that water be mixed with flour or
semolina to form dough. This is the main use of water, along with water usage for
equipment cleaning. A large amount of added water is evaporated during baking,
cooking and drying, and thus it is not discharged as wastewater, which comes
mainly from cleaning operations. A very sensible sector regarding microbiological
food safety, like the meat and seafood sectors, the dairy sector needs water for
cleaning and disinfecting all equipment used in contact with milk or dairy products
in the process lines, requiring large quantities of water whatever the dairy product
manufactured. Cleaning and disinfection wastewaters are highly concentrated in
dissolved organic substances and minerals, such as phosphorus, nitrogen and
chloride, which contain cleaning agent residuals as well. As in the meat sector,
dairy plants generally treat wastewater onsite to accommodate the contaminant’s
threshold in public sewage systems. It should be noted however that the dairy sector
has long been a fore-runner within the FBI in the area of water management and
potential savings through membrane separation and cleaning-in-place technologies
development, allowing them to reduce the strength of effluents, to recycle water and
chemicals, and to recover valuable milk ingredients.
The beverage sector exhibits different behavior regarding water usage because
in this case, water is the main ingredient in most of the finished products. That could
explain the large difference between WII and WDI (12 vs 8 L/$, respectively):
almost one third of water withdrawn is not discharged, but consumed to some extent
through incorporation into products such as bottled water, soft drinks, beer and
other alcoholic beverages. Another significant usage of water in this sector is in the
cleaning and rinsing of containers and equipment.
The seafood sector’s high WII and WDI values signify that it has a higher
environmental impact, thus there is far greater potential for improvement in the
seafood sector than the meat sector. Or, said differently, for the same quantity of
water withdrawn/used, an average meat plant would sell six times more products (in
value) than a seafood plant.
15
Even though evaporation is the obvious main cause for significant water consumption in the
sector, this consumption figure is calculated from a volume balance and accounts also for any
water loss (leaks, incidental overflows, etc.) and/or water added into finished product. Further-
more, leaks of water are unfortunately endemic to any industrial activity using water, which should
be controlled to the extents possible.
8.4.2.2 Provincial Results and Interpretation: WII and WDI
Due to the methodology used to survey establishments and calculate the WII and
WDI indicators, it is impossible to report on more regional groupings than those
displayed in Fig. 8.13. Thus, missing sectors in the figure does not mean that an
industry is not present in the region.
As can be seen, regions can be put in perspective for the seafood sector and meat
product sector only. In the seafood sector, for example,16 there is great variability
between British Columbia (WII ¼ 7 L/$), the Atlantic Provinces (WII ¼ 26 L/$)
and Ontario (WII ¼ 0.7 L/$), with a difference of a factor of four between the first
two regions, and a factor of 40 between the latter two. Actually, the Ontario seafood
industry is quite different from that of British Columbia and Atlantic Provinces,
where WII and WDI indicators cannot be blindly compared.
Fish landings in Ontario are restricted and limited to freshwater fish, and imports
dominate in quantity (Fig. 8.14). It is therefore not a surprise to see lower intensity
of water intake and discharge from Ontario establishments since their main activity
is essentially secondary processing, such as canning, and they are generally tech-
nologically more advanced in streamlining of operations in this respect. In addition
(also true for Prairies), it is reasonable to believe that a portion of imports to Ontario
28
26
24
22
20
18
16
l/$
14
12
10
8
6
4
2
0
Meat
Seafood
Bakeries
Meat
Sugar &
Conf.
Dairy
Meat
Seafood
Meat
Meat
Seafood
BC PR ON QC AT
WII WDI X: confidential data
Fig. 8.13 Water intake intensity (WII) and Water discharge intensity (WDI) as a function of
region and sector calculated for year 2005
16
Aquaculture is not considered a manufacturing group. Any industrial processing activity that
might be carried out on an aqua farm is therefore not taken into account in this study.
400
Groundfish
350 Pelagic & other finfish
Freshwater fish
Shellfish
300
Other
Imports
250
200
150
100
50
0
BC PR ON QC AT
Fig. 8.14 Seafood commercial landings and imports in 2005 (Fisheries and Oceans Canada 2007)
include entirely or partially processed products, leading to lower water intake and
discharge intensity indicators. The case of British Columbia is a happy medium
between the Atlantic Provinces and Ontario. Since British Columbia and the
Atlantic Provinces are the two regions processing significant quantities of landed
fish, they remain comparable to some extent. In this respect, the fourfold difference
between their indicators suggests that the dominant shellfish activity is the driver of
high intensity in the Atlantic region. This activity often involves use of seawater
(Fisheries and Oceans Canada 2003) that can be discharged back to the sea (which
is not allowed in British Columbia), thus implying the high discharge rate observed
in Atlantic region as compared to British Columbia (97.2% vs 89%, respectively).
Conversely, British Columbia’s seafood product industry focuses on the freezing,
canning and secondary processing of salmon, bottom-feeding fish such as halibut,
tuna and other fish, and roe (British Columbia Ministry of the Environment 2007).
With production ranging from the landing to advanced processing of seafood, all of
which require water (brine, canning juices, use of steam during canning operations,
etc), it is not surprising that British Columbia has a lower discharge rate, in other
words a higher water consumption rate in comparison to the Atlantic region. In
contrast, Ontario, which imports most of its production, uses only a few water-
intensive operations. Finally, one also needs to be reminded that regarding water
and wastewater, local regulatory differences with respect to intake volume, dis-
charge pollution load, and volume are also important drivers of an establishment’s
behavior, whatever the characteristics of the production.
It should be noted that Canada’s average is pushed up significantly by the
Atlantic Provinces’ intensity indicators since these provinces account for a large
share (70%) of seafood production in Canada (Statistics Canada 2006).
The other sector that can be analyzed at the regional level is the meat sector. In
this sector, there are two regional groups: British Columbia and Atlantic Provinces
on one side, with WII indicators between 5 and 7 L/$, and Ontario, Prairies and
Quebec regions on the other side, with intensity between 2 and 3 L/$, which are two
to three times lower than indicators of the first group. The two groups actually differ
substantially because of the level of industrialization in industry and their qualita-
tive production differences.
British Columbia and the Atlantic Provinces have numerous small plants pro-
cessing different products (beef, hog and poultry), which account for only 6% and
3% of national manufacturing shipments, respectively. As a consequence, most of
the Canadian meat processing activities are performed in other regions (typically
the Prairies, Ontario and Quebec), which have the largest and most specialized
facilities. Because of their size, these facilities were historically the first to face
local issues concerning very large water usage (and increasing cost), and effluent
limits requirements. Onsite wastewater treatment was necessary for most of the
facilities, which also had to absorb the associated cost. Consequently, they had to
globally implement rationalized procedures aimed at limiting water usage, espe-
cially because the generalized implementation of HACCP program17 tends to
require greater use of water for sanitation purposes.
Here, too, Canada’s national intensity (2.8 and 2.5 L/$ for WII and WDI,
respectively) are similar to the more eco-efficient group (Quebec, Ontario and the
Prairies), which can be explained by their share of Canada’s meat production (91%)
(Statistics Canada 2006).
8.4.3 Response Options: WII and WDI
The challenge FBI faces is to reduce current demand for water by improving its
efficiency through adoption of the best practices, without compromising the quality
of its products, hygiene or food safety. Improved water use efficiency would result in
a reduction of the water intake intensity indicator measured here, and would translate
into a decrease of water resource depletion, as well as the cost associated with water
utility. Consequently, the volume of water discharged and associated pollution would
be reduced, leading to a reduction of the water discharge intensity indicator and
discharge cost. Ultimately, each plant should work in “pseudo closed loop,” which
means withdrawing only water added to the final product and recycling all other
17
HACCP: Hazard Analysis and Critical Control Point. HACCP is an internationally recognized
preventative system designed to detect potential hazards before they occur, and to implement
control measures to reduce or eliminate the likelihood of their occurrence at each step in a process,
and in all ingredients and packaging. All meat products must have a HACCP plan and if a new
meat item is produced it cannot be marketed until a HACCP plan for that production process is
developed. Canadian law requires that all federally inspected meat processing facilities (who are
eligible to export) develop HACCP systems for their beef product lines. Today, food plants
worldwide are using HACCP in a wide range of food manufacturing settings. HACCP is thus
not limited to meat products.
water 100%; this includes water discharged in the wastewater, through the chimneys
(during baking, cooking or drying) or water lost in solid wastes. Of course, this is not
economically feasible but there are a large variety of possible measures that compa-
nies can take towards that goal, the cost of which varies greatly (European Commis-
sion 2006).
Low cost measures, such as simply measuring discharge and managing procedures
that require large quantities of water, will allow companies to react swiftly to any
problems and to avoid waste. Optimizing cleaning and disinfection procedures and
cycles and using control valves can also be quick cost-effective measures, particu-
larly for those sub-sectors urged by stronger food safety regulations on cleaning
efficiency. Furthermore, more advanced technologies for filling and cleaning equip-
ment, albeit more expensive, would allow establishments to win on two fronts: first
by simultaneously reducing water requirements and the volume of effluents, and
second, making it possible to recover usable raw materials from the latter. A more
comprehensive analysis of operational procedures may reveal opportunities requiring
varying levels of investment. For example, by using the process integration approach,
energy and water flows can be studied simultaneously across the plant. This is
particularly relevant to the FBI, where these two resources are often closely linked.
A process integration analysis aims to optimize interactions between the different
components of the production chain rather than improving each component individ-
ually. Given constraints such as water and effluent quality standards, food safety and
quality, process and financial options, the best strategy for water (and energy) usage,
recycling and discharge within a plant can be identified, leading to savings in both
resources and production costs (Gonzalez and Poirier 2003).
8.5 Packaging Use
Food packaging fulfils several functions. It protects products to ensure they are
preserved from the moment they are packaged until their consumption, thereby
preventing any physical damage, spoilage or sensory attribute degradation during
transportation, storage and handling along the supply chain. Packaging is also the
conveyor of consumer and manufacturer information and is a marketing tool for the
brand owner. Usually one can distinguish between primary, secondary and tertiary
packaging as follows. Primary packaging is in direct contact with the food product
(e.g., the plastic liner inside a breakfast cereal box, or a glass beer bottle). Secondary
packaging exists mainly to facilitate consumer use. It may be, for example, the folding
carton of a breakfast cereal box, the boxboard of a case of beer bottles, or the shrink
plastic film used in packing water bottles together. Tertiary packaging is for products
packaged together as a lot and is used to facilitate their transfer from the manufacturing
plant to the warehouse, distribution center, and retail store (e.g., a larger boxboard
containing 12 cereal boxes and the pallet containing eight of these boxboards). Thus,
primary and secondary packaging is generally discarded by the consumer, while
tertiary packaging is managed by wholesalers and/or retailers.
Although very useful, the use of packaging generates important environmental

impacts. First in line, the manufacturing of packaging uses important resources
such as iron, aluminum and silica (for metal and glass containers), trees (for paper
and cardboard), and oil (for plastics) as well as energy for their processing.
Secondly, primary and secondary packages become waste at the consumer level,
where they must be collected and transported to a site and either eliminated or
sorted for reuse or recycling. The same problem exists for tertiary packaging
although this is managed at the industry level.
Management of packaging wastes became a major local issue in the late 1980s
when it was realized that close to 80% of packaging used in Canada was going to
disposal (CCME 1990), and that in 1989 an estimated 30% (by weight) of munici-
pal wastes going to the landfill consisted of disposed packaging (CCME 1996).
Hence, the creation of “alternative solutions” to divert packaging from landfills
came into being, such as local selective collection, deposit-refund containers,
recyclable material sorting plants, waste exchange systems for recycled material,
plants dedicated to specific recycled materials (e.g., aluminum, plastics, etc.), new
waste reclamation routes (particularly for plastics), and energy recovery through
waste incineration. However, these options are still not available everywhere in
Canada. For example, EPS (expanded polystyrene), the most voluminous domestic
packaging waste, can be recycled (plastic with logo #6) but it is not accepted in blue
boxes in most parts of Canada simply because the price of the virgin material is
lower than its recycling cost. Valorization of packaging material raises a serious
challenge due to the diversity of material (Table 8.4) and the many players and
responsibilities involved (by designers, manufacturers, distributors, wholesalers,
retailers, consumers, and local authorities), resulting in a large proportion of
packaging waste still being land filled or incinerated.
These environmental concerns should be put into perspective along with recent
and future societal trends. Canadian households generate more waste than ever (an
increase of 11% between 2002 and 2006). Unfortunately, despite all the efforts put
into the valorization programs, land filled and incinerated wastes have increased by
5.2% during the same period (Statistics Canada 2008a). Societal trends including
decreasing household size, growing demand for products packaged in smaller or
individual portions, and the increasingly technical nature of newer packaging
(designed to better maintain food quality, safety and marketing appeal) are resulting
in increased packaging waste quantity per capita. Figure 8.15 provides a breakdown
of household packaging waste generation in Ontario in 2005. Assuming this Ontario
figure can be applied to Canada as a whole,18 it can be estimated that about 2.3
million tons of household packaging waste were generated in the country in 2005.
As food packaging accounts for around 60% of total packaging (Dworkin 2006),
18
National data is extrapolated from Ontario because its data is the most complete within major
provinces. With similar data in Quebec (queryRecyc-Québec 2007) and B.C., extrapolation to
national values should not be far from the real amount of food waste generated by an average
Canadian home.
Table 8.4 Environmental issues and cost of main food packaging materials (Marsh and Bugusu
2007)
Material Environmental issues Cost
Advantages Disadvantages
Glass l Reusable l Heavy and bulky to l Low cost material
l Recyclable transport but somewhat

l Often contains costly to
recycled content transport
Aluminum l Recyclable l No disadvantages in rigid l Relatively
form expensive but

l Lightweight l Separation difficulties in value encourages
l Economic laminated form recycling
incentive to
recycle
Tinplate l Recyclable l Heavier than aluminum l Cheaper than
l Magnetic, thus aluminum

easily separated
Tin-free steel l Recyclable l Heavier than aluminum l Cheaper than
l Magnetic, thus tinplate

easily separated
Polyolefins (e.g., l Recyclablea l Easily recycled in semi- l Low cost
polyethylene, l High energy source rigid form but

polypropylene) for incineration identification and
separation more difficult
for films
Polyesters (PET, l Recyclablea, b l Easily recycled in rigid l Inexpensive but
PETE, form but identification higher cost

polycarbonates, and separation more among plastics
and polyethylene difficult for films
naphthalates)
Polyvinyl chloride l Recyclablea l Contains chlorine l Inexpensive
(PVC) l Requires separating from
other waste
Polyvinylidene l Recyclablea l Contains chlorine l Inexpensive but
chloride l Requires separating from higher cost

other waste among plastics
Polystyrene (PS) l Recyclablea l Requires separating from l Inexpensive
other waste
Polyamide l Recyclablea l Requires separating from l Inexpensive but
other waste higher cost

among plastics
Ethylene vinyl l Recyclablea l Requires separating from l Inexpensive when
alcohol (EVOH) other waste used as thin film

Polylactic acid l Recyclablea, c l Requires separating from l Relatively
(PLA) other waste expensive

Laminates/ l Often allows for l Layer separation is l Relatively
coextrusions source reduction required expensive but

(plastic and cost- effective for
plastic /or foil or purpose used
paper)
(continued)

Material Environmental issues Cost
Advantages Disadvantages
Paper and l Made from l Low cost
paperboard renewable
resources
l Recyclableb
a
All thermoplastics are technically recyclable and can be recycled at production site, contributing
to lower cost. As inexpensive materials, post-consumer recycling competes with the ease of
separating and cleaning these materials
b
Recycled extensively for non-food product uses
c
Can be broken down to monomer level and reprocessed
Old Corrugated Containers

Glass packaging Paper based Gabletop
23% packaging Paper Laminants
41% Aseptic Containers
Old Boxboard
PET bottles (#1)
HDPE bottles (#2)
Residential Plastic Film
Aluminum packaging wastes Plastic Laminants
packaging generated in Ontario Polystyrene (#6)
3% 857 047 tonnes Other Plastics
(2005) F&B Steel Cans
Steel
packaging Aerosols
7% Paint Cans
F&B Aluminum Cans
Other Aluminum Packaging
Plastic F&B Flint Glass
packaging F&B Coloured Glass
26%
Fig. 8.15 Household packaging waste generation in Ontario for the year 2005 (Stewardship
Ontario 2006, 2007)
food packaging waste would tally to approximately 1.4 million tons or 42 kg per
inhabitant.19
Packaging waste is on the rise. Hence, management costs supported by local
governments have increased dramatically in the last decade, especially for collec-
tion and transportation (Statistics Canada 2008c). Despite numerous valorization
programs, packaging today still accounts for a significant 15% share of materials
19
However, this value represents only part of the grand total since industrial, commercial and
institutional sector waste is not considered here. This latter waste comes from food consumption
outside the home, in workplaces and food services and tertiary packaging waste from processors,
wholesalers and retailers. For example, in the United States in 1999, 27% of residual beverage
containers were generated outside the home (R.W. Beck Inc. 2002).
disposed of in landfills (Downham 2008). Part of the solution lies in the reduction of
packaging. With the main objective of decreasing the costs associated with pack-
aging and transportation, while maintaining the quality of products sold, food
processors have significantly reduced the amount of material per standard container
in the last few decades (Refreshments Canada 2008; Marsh and Bugusu 2007).
However, the reduction of FBI packaging waste is a complex issue. Managing
packaging-related environmental issues cannot be limited to reducing packaging
quantity only, because trying to decrease the latter may lead to food losses,
producing a larger negative impact than the gain provided by reduction in packag-
ing (Erl€ov et al. 2000). The real challenge lies rather in choosing packaging
materials and designs with a smaller environmental impact, taking into account
the manufacturing steps of each material, the minimum requirements of the pro-
duct–packaging–distribution system, and impact that packaging generates once it
becomes waste.
8.5.1 The Indicator: PUI
In order to provide information about the packaging materials purchase, to fulfill

the production needs of FBI plants, the packaging use intensity indicator (PUI) was
developed. The PUI expresses the ratio of the value of packaging and containers
purchased per dollar of manufactured goods produced20 ($/$):
Value of packaging and containers purchased

PUI ¼
Value of production
The proposed indicator was developed as a way to provide establishments with a

measure of their “environmental impact” on the quantity of packaging put on the
market. By comparing with different territories (regions/countries), this indicator
may also give local, provincial, and national authorities a more adequate picture of
the total amount (financial value) of each type of packaging generated, as well as
the relative impact of each stakeholder, helping governments take appropriate
actions to sustain human and economic growth.
tightly linked to regional agricultural production characteristics), and because the
industry involves various scales of production (from small facilities to very large
processing plants), the indicator should reflect these peculiarities. Therefore, to
maintain consistency within food industry sectors and to make valid regional
20
The production is measured by the value of shipments of goods of own manufacture, that is, the
selling value of goods made by reporting establishments, excluding transfers into inventory and
consignment sales, shipping charges by common or contract carriers, discounts and returns, federal
and provincial sales taxes and excise duties and taxes, and sales of goods purchased for resale.
comparisons, establishments were grouped so as to have similar characteristics

(Table 8.2) (i.e., same sector of activity, same region, and same size as determined
by number of employees). The PUI is calculated for each group according to two
complementary methods, as described earlier for ECI and GHGEI. This indicator
uses data from Statistics Canada’s 2002 Annual Survey of Manufactures (ASM21).
For each establishment included in calculations, the value of production (i.e., value
of products sold) and the purchase cost of each packaging or containers are needed.
8.5.2 Results and Interpretation: PUI
This report presents national and provincial results of the PUI for the year 2002 in
various manufacturing sectors of the FBI. The main objective of indicator reporting
is to follow over time how a given sector is increasing or reducing its intensity (i.e.,
increasing or reducing its pressure on materials requirement); however, the present
report does not discuss any timeline trend but sets the baseline. Thus, the following
discussion aims rather at interpreting this 2002 indicator results by focusing on the
inter-sector and inter-regional distinctiveness of the Canadian FBI, while avoiding
as much as possible direct comparisons of the indicators. The main goal here is to
provide objective information that enlightens future interpretation of PUI trends.
Results will first show the difference between sectors and sub-sectors for PUI at
the national level. Second, the effect of different employee sizes will be briefly
presented for a few sectors and sub-sectors, at the national level. Finally, sectors
will be compared per region. All graphs will present the different groupings on the
X- axis and the value of the indicator on the Y-axis.
Unless otherwise stated, the indicator is expressed either in dollars per dollar
($/$) or dollars per one hundred dollars ($/$100). Establishment groupings are
reported using the short names listed in Table 8.1. Data from some sub-sectors
and some plant sizes could not be published due to confidentiality, while the
animal food sector and the other food manufacturing sectors are not covered
within this indicator set.
8.5.2.1 National Results and Interpretation: PUI
Figure 8.16 provides an overview of PUI values across Canada for all sectors and
sub-sectors reported, without regard to size of establishment. Looking first at the
indicator’s median range at the sector level (blue-filled bars in Fig. 8.16), three sets
of industry groups stand out when it comes to “typical plant” intensity: some show
21
More details about the ASM method can be found at the Statistics Canada Web site. http://
www.statcan.ca/cgi-bin/imdb/p2SV.pl?Function ¼ getSurvey&SDDS ¼ 2103&lang ¼ fr&db ¼
imdb&dbg ¼ f&adm ¼ 8&dis ¼ 2
0.28
0.26 Median range (whole sector)
Median range
0.24
Global indicator (whole sector)
0.22
0.20
19.3%
0.18
0.16
PUI ($/$)
0.14
14.2%
0.12
0.10 9.5%
0.08 8.4%
5.4%
0.06
4.7% 4.1%
0.04
0.02 3.2%
0.00
Whole sector
Flour
Malt
Oilseed
Breakfast
Whole sector
Sugar
Cacao conf.
Chocolate conf.
Candy conf.
Whole sector
Frozen F&V
Other F&V
Whole sector
Milk
Other dairy
Ice cream
Whole sector
Red meat slaughter
Red meat
Poultry
Seafood
Whole sector
Retail bakeries
Com. bakeries
Cookies
Flour mix
Pasta
Whole sector
Soft drink
Breweries
Wineries
Distilleries
Grain & Oilseed Sugar & Conf. F&V Dairy Meat Bakeries Beverage
Fig. 8.16 Packaging use intensity (PUI in $/$) as a function of activity sector and sub-sector for
year 2002. Percentage data expresses PUI global values as $/100$. Global indicator is not available
at sub-sector detail
intensity as low as a few percent (less than 5%); some sectors show intensity higher
than 15%; “in-between” sectors with PUI from 6 to 10% also exist.
The high values of the fruit and vegetable sector and the beverage sector (up to
six or seven times higher than least intense sectors) can be explained on the sole
basis of containers usually used in both industry groups. The fruit and vegetable
sector uses, for all production of non-frozen products (approximately half of total
shipments), mainly metal containers (steel cans) or glass containers, which are
costly compared to non complex plastic packaging. Fresh juices from fruits or
vegetables, also sold by this group, are often packaged (especially for retail sales, in
volumes under 2 L) in sophisticated composite packaging such as multilayered
laminated cartons, including aluminum and/or plastic films (gable top and aseptic
cartons), whose price compared to the price of the final product is higher than for
plastic bottles and jugs. It should be noted here that the fruit and vegetable sector
shows the peculiarity of a global PUI below the median range PUI, meaning that
few of the larger establishments in terms of production sales spend less for
packaging materials than typical plants in the sector.
The beverage manufacturing sector also uses glass (bottles) or metal (aluminum
cans, kegs) for all alcoholic beverages, as well as for numerous other beverages not
packaged in plastic bottles, thus inducing a share of packaging cost to production
cost as high as almost 20%.
A closer look at the sub-sector level within the fruit and vegetable and the
beverage sectors (white-filled bars in Fig. 8.15) reveals high intensity in the fruit
and vegetable canning, pickling, and drying sub-sectors (labeled “Other F&V”).
This industry uses a lot of cans and glass jars; its indicator is higher than that of the
frozen fruit and vegetable industry, for which the primary packaging is mainly
flexible plastics or boards. It can be more difficult to interpret the beverage sector
results at the sub-sector level due to product-based intrinsic economic differences
between non-alcoholic (soft drink) and alcoholic industries (breweries, wineries
and distilleries), as well as among alcoholic beverage industries. However, it should
be noted that the unit cost of multi-layer and multi-material containers for fruit
beverages (gable top and aseptic multi-layered cartons, Table 8.4), may be the
driver behind the high intensity of the soft drink industry, together with lower
economic value of soft drinks and bottled water as compared to alcoholic beve-
rages. Furthermore, the deposit system implemented in all provinces for reusable
glass beer bottles (Comeau 2005) significantly reduces the cost of packaging
purchased by breweries.
Among the least intensive sectors, i.e., the grain and oilseed, dairy, meat, and
seafood sectors, meat product manufacturing showed little variability within its
industries. It is worth pointing out that meat and fish products sold fresh at the retail
level (mainly on foam polystyrene trays covered with plastic shrink film) are
generally packaged by retailers themselves, and consequently such packaging is
not accounted for by the indicator.
The grain and oilseed sector, on the contrary, shows significant variability in the
PUI. The oilseed processing sub-sector is the least intensive, eight times less than the
breakfast cereal sub-sector. The latter provides the marketplace mainly household
retail products packaged in small units, within a plastic liner bag inside a boxboard.
With a typical value (i.e., median range) of packaging, expenses around 3–4% of
finished product, the dairy sector as a whole shows a low intensity. However, it is
one of the few food sectors (together with the bakeries and tortilla manufacturing
sector) for which the global PUI value is significantly higher than the median range
(at least +40%). This can be explained by this sector’s make up of numerous
establishments with large numbers of shipments and high intensity simultaneously.
Given the high concentration ratio in this sector, where only a few companies – that
is, a limited number of very productive establishments – dominate the market, this
indicator shows that these establishments could improve their eco-efficiency and
shift intensity down to a lower median level. In industries within the dairy sector,
better performance of the other dairy products sub-sector (butter, cheese, and dry
and condensed dairy products manufacturing) can be seen. This can be explained by
the very high sales value of this kind of processed food compared to that of the fluid
milk sub-sector or ice cream and frozen dessert sub-sector, despite the packaging
quantities (and cost) involved in the selling of highly diversified products packaged
in small units such as yogurts, cream and cheese.
The bakeries sector shows an intermediate intensity level. The cookie and
cracker sub-sector stands out from the others with its high intensity (typical plant
values are about 13%) without pulling the whole sector up to its level; however,
since the median range of the latter still stays between 6% and 7% and the global
indicator is slightly higher than 8%. The cookie and cracker sub-sector is quite
similar to the breakfast cereal sub-sector within the grain and oilseed sector. It is
mainly a retail products industry, often using multiple packaging (e.g., cookies are
packaged in trays and/or individual plastic bags, and over-packed within folding
cartons) in comparison with bakery products from commercial or retail bakeries
sub-sectors, which use more basic packaging.
The sugar and confectionery sector belongs also to the group of “intermediate-
intensity” sectors. All sub-sectors within this industry behave similarly, except for
the sugar manufacturing sub-sector. It is not surprising to find the lowest intensity
for this latter one, given the simplicity and low cost of packaging materials used for
powdered sugar (paper pouches, plastic bags, and paper bags for larger quantities),
and sugar syrups, which are essentially wholesale marketed in plastic pails or
drums, moreover generally reused. On the contrary, chocolate and confectionery
products are mainly shipped individually packaged, involving more advanced
costly packaging materials.
There is no clear trend regarding the influence of establishment size (in terms of
number of employees) on the PUI. At the sector level, only the dairy sector displays
a clear and significant influence, displayed through both global and median range
indicators (results available in technical supplement of the report). Surprisingly, the
global PUI increases with plant size from 3.5$/100$ for small plants up to 6.8$/100$
for very large plants. This trend does not hold true at the sub-sector level. Other sub-
sectors show slight trends but none are significant.
8.5.2.2 Provincial Results and Interpretation
Details by region (province or group of provinces) on the PUI are presented in

Fig. 8.17 (FBI regional sub-sector detail is not available due to confidential data).
British Columbia stands among the least intense region for the sugar and
confectionery sector, which is explained by the predominance of the sugar
manufacturing sub-sector within this industry in this province. Conversely, the
province shows a significantly higher PUI than any other region in the seafood
sector. The global PUI is 50% higher there than in Quebec, and more than three
times higher than in Ontario. Moreover, a typical plant of the sector is at least three
times more intensive than a typical plant in any other region. British Columbia’s
seafood industry activity is largely oriented towards salmon canning, which means
that a large amount of its packaging purchases are expensive steel or aluminum
cans. The seafood industries in the Atlantic Provinces and in Quebec are mostly
focused on shellfish and seafood activities involving more basic and cheaper
packaging materials, together with far higher selling prices.
In the beverage sector, British Columbia stands out as the most intensive
province. This may be due to the peculiarity of that province to package beer22
22
The breweries subsector accounted for 44% of beverage sector sales of manufactured goods in
British Columbia, in 2002 (Statistics Canada 2008b).
0.30
Median Range
Global indicator
0.25
0.20
PUI ($/$)
0.15
0.10
0.05
0.00
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Beverage
Grain & Oilseed
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Beverage
Grain & Oilseed
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Beverage
Grain & Oilseed
Sugar & Conf.
F&V
Dairy
Meat
Seafood
Bakeries
Beverage
F&V
Dairy
Meat
Seafood
Bakeries
Beverage
BC PR ON QC AT
Fig. 8.17 Packaging use intensity (PUI in $/$) as a function of activity sector and region
calculated for year 2002
mainly in cans or draught kegs instead of refillable glass bottles as other regions
do, thus increasing packaging cost.23 In 2002, use of bottles represented only
26.1% of the beer sold in the province, whereas bottled beer represented 47.5%
in the Prairies, 76.9% in the Atlantic Provinces, 77.2% in Ontario, 85.1% in
Quebec, and in Canada averaged 68.1% overall (Brewers Association of Canada
2007).
The Prairies, like British Columbia, show a low PUI in the sugar and confec-
tionery sector. It is also the lowest intensive region in the meat sector, one of the
main economic drivers of the FBI in the Prairies. This is also true in the fruit and
vegetable sector, where the region compares with the Atlantic Provinces in
global PUI (8$/100$). However, good economic performance is common to
only a few of the larger selling plants since – as revealed through the median
range indicator – a typical mid-performing plant should score higher, between 11
and 20$/100$, as in the Prairies’ fruit and vegetable sector. It is likely that larger
productive plants can benefit from cheaper packaging materials24; the same
reasoning may hold for the sugar and confectionery, meat, and seafood sectors.
The reverse is observed in the grain and oilseed sector, another important FBI
23
Glass bottles are reused by breweries (no need to buy a new bottle for each portion sold) contrary
to aluminum cans, which are recycled (a new can is needed for each portion sold).
24
The F&V sector in the Prairies is dominated by a few large plants selling uncooked French fries
in large plastic bags to restaurants.
sector in the Prairies, where the higher positioning of the global indicator may
reveal that few of the largest selling plants are more intensive than other mid-
performing plants.
Ontario is a leading province in almost all sectors of Canadian FBI, both in
terms of number of establishments and production. It is thus not surprising to
see national sector values presented as being close to that of the province,
whether for the global or the median range PUI. The main exception is in the
seafood sector, for which Ontarian establishments are quite different from
processing plants in coastal regions and deal mainly with imported materials
for secondary processing. This is true in the Prairies too, where similar PUI
values are observed.
Other exceptions, though to a lower extent, are in the meat and the fruit and
vegetable sectors where the dominance of Ontario is less marked because
of the share in production with other provinces. In the meat sector, the Prairies
and Quebec are also big players, and variability among PUIs is slight, although
Ontario’s larger processors seem to show a higher intensity.
In the fruit and vegetable sector, Ontario (along with Quebec) is the most intensive
buyer of packaging materials. Mid-performing plants can spend up to 25$ for packag-
ing per 100$ of production, which is substantial. As seen previously, such a high
intensity level is characteristic of the canning and pickling sub-sector, and Ontario’s
fruit and vegetable industry was mainly engaged (75%) in this sub-sector activity in
2002 (Statistics Canada 2008b). This was also the case for Quebec (78%).
Figure 8.14 also confirms the previous observation of the effect of economic
concentration in the dairy sector. It is not surprising to see that the global PUI is
higher than the median range, especially for Quebec and Ontario. In 2002, both
provinces together accounted for more than 74% of Canadian dairy product
manufacturing shipments (Statistics Canada 2006), and almost 60% of all establish-
ments. The concentration effect is particularly marked in this sector, with three
companies owning 75% of the market through 15% of all establishments (Agricul-
ture and Agri-Food Canada 2005). These highly productive establishments influ-
ence the global PUI as a result. However, this analysis needs to take into account
several particularities in Quebec, such as a fluid milk market less oriented than
elsewhere to plastic pouches (which are cheaper than milk cartons), and a cheese
market open to specialty goods packed in more sophisticated and costly packaging
materials (although these products are also more expensive than traditional North
American cheeses).
The Atlantic Provinces show the lowest global PUI in the fruit and vegetable
sector, slightly below 8$/100$. However, unlike the Prairies, their median range
stays in line with this value and displays a narrow range. It is thus likely that all
plants in the region behave similarly in this sector concerning packaging require-
ments, provided their production is qualitatively comparable. This region might
also serve as a benchmark for other provinces in this sector, provided production
patterns are similar. Unfortunately, the lack of disaggregated data about the relative
importance of both fruit and vegetable sub-sectors in the Atlantic Provinces does
not allow such inferences.
8.5.3 Limitations: PUI
The proposed indicators have several limitations, most of them resulting from the
desire to provide indicators that report sub-sector, regional and size specificities.
First, the indicator is calculated per value of product manufactured, instead of per
volume of product manufactured as initially intended, which would have been a
better eco-efficiency indicator (Arcand et al. 2005) due to the limited amount of
volume data in the Statistics Canada ASM database. Second, this new indicator from
Agriculture and Agri-Food Canada’s series of agri-environmental indicators was
computed for a single benchmark year, 2002, which does not allow for year to year
comparison. Lastly, another significant limitation of the PUI resides in its environ-
mental significance, since, at this stage of development, it does not report packaging
quantities, nor does it weight each type of packaging accounted for, due to an
environmental impact factor. As mentioned above, it is through such weighting that
the environmental issues of food packaging can truly be addressed, and future
enhancements are expected using this approach.
8.5.4 Response Options: PUI
Since this is the first time the PUI has been calculated, we still do not have any
references for comparison purposes and are thus unable to assess the benefit of
any actions previously implemented, whether by industry or government. The
industry has an important role to play in reducing the environmental impact of
packaging. The 4R (Reduction, Reuse, Recycling, Recovery) waste management
hierarchy is the rule of thumb in any waste minimization strategy.25 As already
mentioned, the industry can act by reducing the quantity of materials needed to
pack a given quantity of finished product. However, such a source reduction
approach is limited by the minimal protection required by the product and the
distribution system, a threshold that cannot be crossed. Another source reduction
option is to reduce over-packing (secondary packaging) to the bare minimum by
redesigning the packaging. New designs may also offer the opportunity to choose
materials with a lower environmental impact: either the same materials with a
higher recycled content or those coming from cleaner processes (interaction with
the supplier is thus of prime importance), or new materials with a smaller
ecological footprint (calculated over whole life cycle, taking into account end-
of-life options). The industry may be influenced in this way by downstream
market drivers from distributors and consumers who – thanks to their purchasing
power – may request low ecological footprints. This phenomenon is already being
observed. A significant amount of packaging waste also comes from imported
25
A short list of generic beneficial operating or management practices is also provided in the
appendix section of the Web version of this document.
food products, over which the industry has no influence, unless it is the actual
importer who can in this case negotiate with suppliers, e.g., for more environ-
mental-friendly packaging. However, governments can play a role by requiring
transparent information about the packaging materials used and their ecological
footprint. Governments also should ensure that these materials, once they become
waste, are efficiently managed through locally available reuse or recycling chan-
nels in order to limit land filling. A principle already implemented in a few
provinces, referred to as the extended producer responsibility, is also a policy tool
that may raise the food and beverage industry’s awareness of packaging. This
principle should ideally be based on criteria arising from more thorough life cycle
assessments. As a supporting stakeholder, governments could play a more active
role in setting up and centralizing Canadian life-cycle databases, to then be made
publicly available to assist decision makers in the industry. The industry is not
solely responsible and should not be expected to meet all the constraints alone.
The consumer is the last decision-maker when purchasing and is also the “local”
waste generator. One of the mandates of governments and local authorities is to
inform consumers and to support education of the public about packaging wastes.
One step further could be to guide the public about alternative consumer choices,
through information based on impacts embedded in the packaging of products.
8.6 Conclusions and Recommendations
In summary, the Canadian FBI sector uses a significant amount of energy: close to
4% of the energy used by the Canadian manufacturing industry as a whole. The
consumption of this energy is responsible for most of the FBI’s greenhouse gas
(GHG) emissions. The grain and oilseed milling sector, and the sugar and confec-
tionery products manufacturing sector are among the most energy-intensive users.
Indicators also reveal that the largest companies in terms of sales in these groups are,
in general, “worse than average” performers, regardless of their location in Canada.
Within these sectors, three sub-sectors – rice milling and malt manufacturing, oilseed
processing, and sugar manufacturing (particularly in British Columbia) – drive up
their respective group’s energy intensity values. The same is true of the distilleries
within the beverage sector. Least energy-intensive sectors are seafood, meat, and
dairy products manufacturing. Greenhouse gas emission intensity indicators are
generally consistent with energy consumption. However, regional differences exist
in use of petroleum products, the combustion of which emits large quantities of
GHGs. In 2005, the FBI was responsible for nearly 20% of the total water used by all
Canadian manufacturing industries. Of this water intake (i.e., volume of water
withdrawn), the FBI discharges 77% and recirculates 4%. The seafood product
preparation and packaging industry is the most intense water withdrawing industry
among the FBI in Canada, with a national average value of 17 L of water withdrawn
per dollar of product sold, although regional values vary from 1 L/$ in Ontario to
26 L/$ in the Atlantic Provinces. Nevertheless, it is also the sector that discharges the
most water (96% is withdrawal). In contrast, the beverage manufacturing industry

group is the most intense “consumer” of water (difference is between withdrawn and
discharged water) with discharges accounting for two-thirds of the water withdrawn,
most of it being incorporated into finished products. The meat product
manufacturing industry group is the least intense water-withdrawing group (3 L/$),
withdrawing six times less than the seafood product group and almost four times less
than the dairy product manufacturing group (11 L/$). Within the meat sector,
geography makes a clear difference, with provinces known for livestock production
(the Prairies, Ontario and Quebec) performing much better (below 3 L/$) than
British Columbia (7 L/$) and the Atlantic Provinces (5 L/$). The FBI transforms
raw agricultural commodities, or ingredients, into semi-prepared and consumer-
ready food and beverage products that have to be packaged adequately to ensure
they can reach consumers without losing their physical and hygienic integrity and
quality attributes, as well as to convey consumer and manufacturer information. The
Packaging Use Indicator (PUI) first results reported here are sectorial measures for
2002 and provide a baseline for future reference. The highest PUI values (greater
than 15$ of packaging per 100$ of product) are produced by the fruit and vegetable
and the beverage manufacturing sectors, which are typically high consumers of such
packaging as metal cans (iron and aluminum) and glass containers. Conversely,
animal materials processing groups (meat, fish and seafood, and dairy products)
show the lowest PUI values for reasons such as the economic value of their
production (e.g., highly processed dairy products), the small quantity of packaging
required (e.g., unprocessed meat products, fish and seafood), or even geographical
characteristics (e.g., fish and seafood in Atlantic region and Quebec).
The indicators proposed here could be optimized to become more robust,
have better coverage, represent relevant environmental issues and, lastly, facili-
tate response options. ECI and GHGEI indicators at the sub-sector level with
both global and median range results could be released. The ECI indicator could
be calculated by adding a weight for each source of energy used, computing an
environmental pressure factor taking into consideration the total direct and
indirect impacts of a source and geographic location. This would highlight
any shift made to more sustainable energy use, in particular, renewable ones.
An ECI indicator calculated to break down the energy consumption according to
main uses in a plant would be a significant improvement. Causes of low
efficiencies could then more easily be identified, and therefore eco-efficiency
measures and best management practices could be more straightforwardly
deployed.
Improvements could also include the development of purely physical indicators
independent of financial data, for example, L of water used per kg of product sold
instead of per $. These indicators would be independent of price fluctuation, from
region to region, or from high quality vs. low quality products sold. This raises a
delicate issue because the data from ASML does not provide “per kg” data,
although the survey is an ideal source of information, because of its frequency
and coverage of the FBI. Another source of physical data (i.e., all outputs sold by
each establishment expressed in kg per year) is therefore required, unless the ASML
is adapted accordingly, which could serve many other national purposes (e.g.,
national environment accounts managed by Statistics Canada). One other option
is the generation of a tool to gather and aggregate establishments’ physical data and
to calculate sector and regional indicators, which at the same time could allow
establishments to compute their own eco-efficiency indicators. This would allow
the measurement of the quality of the water discharged by, for example, quantifying
the pollution loads (organic load using biochemical oxygen demand) (BOD), as
described by Maxime et al. (2006), as well as measuring nitrogen, phosphorus,
specific chemicals, microbial loads, and toxicity among other).
Taking a life cycle approach as a basis for the analysis of environmental issues of
packaging is gradually achieving consensus within policies both abroad26 and in
Canada,27 as well as within the scientific community. From a broader perspective,
many urban communities now assess and choose solid waste management plans on
the basis of life cycle-based comparison studies. It is well known that such analysis is
very time-consuming in the data inventory phase and thus requires a significant
investment, but it becomes very cost-effective afterwards and will pay off during
subsequent similar analyses. Furthermore, it is not limited to environmental assess-
ments and can also be used for cost analyses of the whole life cycle of a packaged
product. The current indicator should eventually be improved to allow for a break-
down by types of packaging. Simultaneously, it should weight every type of packag-
ing material put on the market by an environmental pressure factor embedding the
total impacts – direct and indirect – of each type, also taking into account geographi-
cal location, as the life cycle approach allows.
Lastly, the adoption of a comprehensive, open approach to environmental pres-
sures and natural resource use along the food chain, from production to final
consumption and even to end of life cycle (including waste management and
biomass recycling at farms, establishments, and consumer and post-consumer
levels), would be another useful research stream to precisely identify the causes
and division of responsibilities, as well as the effects and response options (avoiding
collateral damages). The national environment accounts mentioned above and life-
cycle analyses are in line with this. Additional national research initiatives in these
areas would contribute towards achieving a sustainable Canadian agri-food system.
In practice, the food and beverage industry’s environmental performance is
influenced by an individual plant’s business practices, including internal policies,
management system and staff awareness, and manufacturing processes. Some of
the available business practice options can be grouped under the label “best
operating practices” (BOPs), such as those recently described (European Commission
2006). Three criteria are generally used to prioritize and classify these practices: the
26
For example, in Australia with the National Packaging Covenant (http://www.packagingcovenant.
org.au) and Europe (European Union 2004).
27
For instance, the Canadian Council of Ministers of the Environment’s Extended Producer
Responsibility Task Group is developing a Canada-wide strategy for sustainable packaging.
Aspects and stakeholders involved are considered in the life cycle. (http://www.ccme.ca/
ourwork/waste.html)
investment cost to adopt, quantification of the anticipated environmental gains

(made apparent by the indicators), and the return-on-investment period. Once
calculated, the indicators will shed light on the extent to which the best operating
practices are being implemented and, in a sense, will quantify a company’s efforts
towards environmental sustainability (Richard 2003; Industry Canada 2001)
Acknowledgments The authors would like to thank the Manufacturing, Construction and Energy
Statistics Division of Statistics Canada, Ottawa, and in particular Daniel Scott, André Gravelle and
Francine Rouleau. Many thanks are addressed to Andy Shinnan, François Soulard, Joe St. Lawrence
and Martin Lemire, of the Environment Accounts and Statistics Division, Statistics Canada, Ottawa.
Dr. André Talbot of the Aquatic Ecosystem Protection Research Division, Environment Canada,
Montreal, is kindly acknowledged for much advice provided. The authors would like to thank Isabelle
Vézina, a master student in environment at the Université de Sherbrooke, for her dedicated work at the
Food Research and Development Centre – St Hyacinthe, from April to September, 2006. Finally,
many thanks are being addressed to Mathieu Guillemette from Éco Entreprises Québec.
References
Agriculture and Agri-Food Canada (AAFC) (2004) Agricultural Policy Framework. http://www.
agr.gc.ca/cb/apf/index_e.php. Accessed 12 October 2004
Agriculture and Agri-Food Canada (AAFC) (2005) Canadian Dairy Industry Profile. Ottawa, Ontario:
AAFC (August 2005). http://www4.agr.gc.ca/resources/prod/doc/dairy/pdf/dairyprofile.pdf
AMEC Earth & Environmental Ltd. (2003) Management of Wastes from Atlantic Seafood
Processing Operations. Final report submitted to the National program of action, Atlantic
regional team, Environment Canada Atlantic Region. http://www.dal.ca/aczisc/nparpt.pdf
(visited September 12, 2008). Other related documents (including this one) also available
online: Canada. Fisheries and Oceans Canada (DFO), Gulf Home; Seafood Plant Effluents –
Turning Problems into Profits; http://www.glf.dfo-mpo.gc.ca/os/effluents/index-e.php [last
updated January 15, 2007; visited September 12, 2008]
Arcand Y, Maxime D, and Marcotte M (2005) Organic solid residues and packaging wastes,
Ch. 30. In: Lefebvre A, Eilers W, Chunn B (eds) Environmental sustainability of Canadian
agriculture: agri-environmental indicator report series – report #2 (eds). Agriculture and Agri-
food Canada, Ottawa, Ontario, pp 184–187. http://dsp-psd.pwgsc.gc.ca/Collection/
A22–201–2005E.pdf
Brewers Association of Canada (2007) 2007 Annual Statistical Bulletin. http://www.brewers.ca/
UserFiles/Documents/pdfs/eng/statistics/asb/2007/Brewers_ASB_2007_Eng.pdf
British Columbia Ministry of Environment, Oceans and Marine Fisheries Branch (2007) 2006
British Columbia Seafood Industry Year in Review. October 22, 2007. The Ministry, Victoria,
BC. http://www.env.gov.bc.ca/omfd/reports/YIR-2006.pdf [visited November 14, 2007]
CCME (Canadian Council of Ministers of the Environment) (1990) National Packaging Protocol.
http://www.ccme.ca/assets/pdf/napp_e.pdf
CCME (Canadian Council of Ministers of the Environment) (1996) National Packaging Survey-
The 1996 Results. http://www.ccme.ca/assets/pdf/packaging_srvy_96.pdf
CIEEDAC (Canadian Industrial Energy End-use Data and Analysis Centre) (2008) Data report
created online July 30, 2008, from CIEEDAC database, covers annual energy consumption,
CO2, CH4 and N2O emissions, cumulative GHG emissions, production data. Simon Fraser
University, Burnaby, British Columbia. http://www.cieedac.sfu.ca
Comeau M (2005) La gestion des contenants de boissons au Québec. Évaluation du marché et
analyse des options de récupération. Recyc-Québec, Québec, Quebec (La Société québécoise
de récupération et de recyclage), September 1, 2005. http://www.recyc-quebec.gouv.qc.ca/

upload/publications/consigne/EtudeContBoissons.pdf; http://oee.nrcan.gc.ca/corporate/statistics/
neud/dpa/comprehensive_tables/index.cfm?attr ¼ 0
Downham J (2008) Presentation on packaging issues at the City of Toronto Packaging Waste
Reduction Forum, September 10, 2008, Toronto, Ontario. http://www.toronto.ca/garbage/
packagingforum/pdf/1_downham_pac_na08_cot_jd_speech_6up.pdf
Dworkin L (2006) Presentation on the packaging industry in Canada, Section 4.2 in Marbek
Resource Consultants Ltd., National Extended Producer Responsibility (EPR) Workshop: final
report, pp 11–12. Canadian Council of Ministers of the Environment (CCME), Ottawa,
Ontario, 2007. (PN 1376) http://www.ccme.ca/assets/pdf/epr_wkshp_rpt_1376_e.pdf
Energetics Inc. and E3M Inc. (2004) Energy use, loss and opportunities analysis: US
manufacturing and mining. Prepared by energetics inc. and E3M inc. for US Department of
Energy (DOE), energy efficiency and renewable energy industrial technologies program.
Columbia, Maryland and North Potomac, Maryland: December 2004. http://www1.eere.
energy.gov/industry/energy_systems/pdfs/energy_use_loss_opportunities_analysis.pdf
Environment Canada (2004a) Greenhouse gas division. Canada’s greenhouse gas inventory,
1990–2002. Environment Canada, Ottawa, August 2004. http://www.ec.gc.ca/pdb/ghg/inventor-
y_report/1990_02_report/toc_e.cfm; http://www.ec.gc.ca/pdb/ghg/inventory_report/
1990_02_report/errata_e.cfm; http://www.ec.gc.ca/pdb/ghg/ghg_home_e.cfm
Environment Canada (2004b) Threats to Water Availability in Canada. National Water Research
Institute (NWRI), Scientific Assessment Report Series No. 3 and ACSD Science Assessment
Series No. 1. Environment Canada, NWRI, Burlington (Ont.). http://www.nwri.ca/threats2full/
intro-e.html
Environment Canada (2007) National inventory report 1990–2005: Greenhouse gas sources and
sinks in Canada. Greenhouse gases division, No de cat.: En81–4/2005E. http://www.ec.gc.ca/
pdb/ghg/inventory_e.cfmCanada
Environment Canada (2008) Water use in Canada, 2005. http://www.ec.gc.ca/water/en/manage/
use/e_wuse.htm [visited September 12, 2008]
Environment Canada, Atlantic Region (2008) Report on the characterization and toxicity
testing of fish processing plant effluent in Canada. Toxic Chemicals Updates Newsletter –
Atlantic Canada Edition, 10(1), January 2008. ISSN 1206–5455. http://atlantic-web1.ns.ec.gc.
ca/epb/newsletters/toxchem/Default.asp?lang ¼ En&n ¼ 3F836B38–1 [visited September 12,
2008]
Erl€ov L, L€ofgren C, S€
orås A (2000) Packaging – a tool for the prevention of environmental impact.
Foundation Packforsk, Kista, Sweden, June 2000 (Report No. 194). http://www.stfi-packforsk.
se/upload/Gamla_PF-sidor/REPORT_194.PDF
European Commission (2006) Integrated Pollution Prevention and Control – Reference document
on best available techniques in the food, drink and milk industries. European Commission,
Directorate-General, Joint Research Centre Institute for Prospective Technological Studies
(Seville, Spain), European Integrated Pollution Prevention and Control Bureau. http://eippcb.
jrc.ec.europa.eu/pages/locatemr2.cfm?file ¼ fdm_bref_0806.pdf&name ¼ Food,%20drink%
20and%20milk%20processes&info ¼ info/fdm.htm
European Union (2004) Directive 2004/12/EC of the European parliament and of the council of 11
February 2004 amending directive 94/62/EC on packaging and packaging waste – statement by
the council, the commission and the European parliament. official journal of the European
Union L 47, 18/02/2004, pp 26–32. Published online edition: http://eur-lex.europa.eu/smartapi/
cgi/sga_doc?smartapi!celexplus!prod!DocNumber&lg ¼ en&type_doc ¼ Directive&an_doc
¼ 2004&nu_doc ¼ 12; [see also: ECOLAS – PIRA (2005). Study on the implementation of
directive 94/62/EC on packaging and packaging waste and options to strengthen prevention
and re-use of packaging. Report 03/07884–AL; http://ec.europa.eu/environment/waste/studies/
packaging/050224_final_report.pdf]
Fisheries and Oceans Canada (DFO), Gulf Home (2003) Fish plant effluents: a workshop on
sustainability. Canadian industry report of fisheries and aquatic sciences. In: Morry CJ,
Chadwick M, Courtenay S, Mallet P (eds), p 271. http://www.glf.dfo-mpo.gc.ca/os/effluents/

workshop-report-e.pdf; other related documents, including this one, available online: Canada.
Fisheries and Oceans Canada (DFO), Gulf Home. Seafood plant effluents – turning problems
into profits. http://www.glf.dfo-mpo.gc.ca/os/effluents/index-e.php [last updated January 15,
2007; visited September 12, 2008]
Fisheries and Oceans Canada (DFO), Statistical Services Unit (2007) DFO’s statistical services.
DFO, Ottawa, policy sector, the unit, March 14, 2007. http://www.dfo-mpo.gc.ca/communic/
Statistics/main_e.htm [visited November 15, 2007]
Gonzalez M, and Poirier S (2003) Minimiser l’impact environnemental des usines de transfor-
mation de produits marins. Alerte Technologique 1(1):8–10. Shippagan, Institut de recherche
sur les zones côtières Inc., New Brunswick, Canada. http://www.umcs.ca/Institut/AT001.pdf
[visited September 12, 2008]
Industry Canada (2001) Three steps to eco-efficiency. Ottawa. http://www.ic.gc.ca/eic/site/ee-ee.
nsf/vwapj/finaltool.PDF/$FILE/finaltool.PDF
IPCC (Intergovernmental Panel on Climate Change) (1996) Climate change 1995. The science of
climate change. Contribution of working group 1 to the second assessment report of the
intergovernmental panel on climate change. In: Houghton JT, Meira Filho LG, Callender
BA, Harris N, Kattenberg A, Maskell K (eds) Cambridge University Press, Cambridge, UK.
http://www.ipcc.ch/ipccreports/sar/wg_I/ipcc_sar_wg_I_full_report.pdf
Marcotte M, Maxime D, and Arcand Y (2005) Energy use and greenhouse gas emissions, ch. 28.
In: Lefebvre A, Eilers W, Chunn B (eds) Environmental sustainability of Canadian agriculture:
agri-environmental indicator report series – report #2. Agriculture and Agri-food Canada,
Ottawa, pp 176–179
Marsh K, Bugusu B (2007) Food packaging – roles, materials, and environmental issues. J Food
Sci 72(3):R39–R55. http://webprod.ift.org/NR/rdonlyres/C3FC4F7C-BE99–4124-BA67-A5-
C3A77D1B05/0/FoodPkgEnviron.pdf
Maxime D, Arcand Y, Marcotte M (2005) Water use and effluent generation, ch. 29. In: Lefebvre
A, Eilers W, Chunn B (eds) Environmental sustainability of Canadian agriculture: agri-
environmental indicator report series – report #2. Agriculture and Agri-food Canada, Ottawa,
pp 180–183. http://dsp-psd.pwgsc.gc.ca/Collection/A22–201–2005E.pdf
Maxime D, Marcotte M, Arcand Y (2006) Development of eco-efficiency indicators for the
Canadian food and beverage industry. J Cleaner Prod 14(6–7):636–648
NAHARP (National Agri-Environmental Health Analysis and Reporting Program) (2004) Agri-
culture and agri-food Canada. http://www.agr.gc.ca/env/naharp-pnarsa/index_e.php; see also
NAHARP summary at http://www.agr.gc.ca/env/naharppnarsa/pdf/naharp-pnarsa_e.pdf
[accessed: 12 October 2004]
Natural resources Canada, office of energy efficiency (2007) Comprehensive energy use database,
Table 10: other manufacturing secondary energy use and GHG emissions. Office, Ottawa,
December 12, 2007
Navarri P, Legault A, Bédard S (2001) Process integration. Presentation at managing waste and
wastewater in the food processing industry seminar, organized by Manitoba Hydro. November
21, 2001, Winnipeg, Manitoba. Varennes, Natural Resources Canada, CANMET energy
technology centre, Quebec
NRTEE (National Round Table on the Environment and the Economy) (2001) Eco-efficiency
indicators. Workbook. Calculating eco-efficiency indicators. A workbook for industry. Prepared
by indeco strategic consulting, inc. & Carole Burnham consulting. Renouf Ltd. Ottawa, ON
Nyboer J (2008a) Development of energy intensity indicators for Canadian industry, 1990–2006.
Simon Fraser University, Burnaby BC, Canadian Industrial End-use Energy Data and Analysis
Centre (CIEEDAC) [prepared for Canadian industry program for energy conservation and
natural resources Canada]. http://www.cieedac.sfu.ca/
Nyboer J (2008b) Development of greenhouse gas intensity indicators for Canadian industry,
1990–2005. Simon Fraser University, Burnaby BC, Canadian Industrial End-use Energy Data
and Analysis Centre (CIEEDAC) [prepared for Environment Canada]. http://www.cieedac.sfu.ca/
Recyc-Québec (2007) Bilan 2006 sur la gestion des matières résiduelles au Québec. La Société
québécoise de récupération et de recyclage (Recyc-Québec), Québec, Quebec. http://www.
recyc-quebec.gouv.qc.ca/upload/Publications/Bilan2006.pdf
Refreshments Canada (2008) Beverages & the environment: about beverage packaging. Refresh-
ments Canada. Online promotional literature from beverage trade association, Toronto,
Ontario. Visited December 27, 2008. http://www.refreshments.ca/en/environment/beverages_
environment.asp
Richard F (ed) (2003) Waste management guide for small and medium enterprises. Canadian
version. Éditions Ruffec and NI environnement, Montreal. http://www.qc.ec.gc.ca/dpe/
Anglais/dpe_main_en.asp?innov_guide_mat_residuelles; http://www.qc.ec.gc.ca/dpe//
Publication/Mat
RW Beck Inc (2002) Understanding beverage container recycling: a value chain assessment
prepared for the multi-stakeholder recovery project. Businesses and Environmentalists Allied
for Recycling (BEAR), Santa Monica, California, January 16, 2002. http://www.globalgreen.
org/bear/Projects/FinalReport.pdf
Statistics Canada (2002a) Annual survey of manufactures. http://www.statcan.gc.ca
Statistics Canada (2002b) Annual industrial consumption of energy. http://statcan.gc.ca
Statistics Canada (2003) Report on energy supply-demand in Canada, 2002. Statistics Canada,
Ottawa, October 2003 (Catalog no. 357–003-XIB, annual publication). http://www.statcan.ca/
english/freepub/57–003-XIB/0000257–003-XIB.pdf
Ottawa, November 2004 (Catalog no. 57–003-XIB, annual publication). http://www.statcan.
ca/english/freepub/57–003-XIB/0000357–003-XIB.pdf
Statistics Canada (2006) CANSIM Table No 301–0006. Principal statistics for manufacturing
industries, by North American Industry Classification System (NAICS), annual. [124134 series].
Statistics Canada, Ottawa, November 30, 2006. http://cansim2.statcan.ca/cgi-win/cnsmcgi.
exe?Lang ¼ E&RootDir ¼ CII/&ResultTemplate ¼ CII/CII___&Array_Pick ¼ 1&ArrayId
¼ 3010006 [visited December 10, 2007]
Ottawa, June 2007 (Catalog no. 57–003-XIE, annual publication). http://www.statcan.ca/bsolc/
english/bsolc?catno ¼ 57–003-X; http://www.statcan.ca/english/freepub/57–003-XIE/57–003-
XIE2005000.htm; http://www.statcan.ca/english/freepub/57–003-XIE/57–003-XIE2005000.pdf
Statistics Canada (2008a) EnviroStats, 2(3):18 (Table 3) (Catalog no. 16–002-X). http://www.
statcan.ca/english/freepub/16–002-XIE/16–002-XIE2008003.pdf
Statistics Canada (2008b) CANSIM Table #301–0003: annual survey of manufactures (ASM),
principal statistics by North American Industry Classification System (NAICS). Statistics
Canada, Ottawa, Ontario. http://cansim2.statcan.ca/cgi-win/cnsmcgi.exe?Lang ¼ E&RootDir
¼ CII/&ResultTemplate ¼ CII/CII___&Array_Pick ¼ 1&ArrayId ¼ 3010003 [last modifi-
cation Apr. 25, 2008; visited 2008/11/04]
Statistics Canada (2008c) CANSIM Table #153–0045: local government characteristics of the
waste management industry. Statistics Canada, Ottawa, Ontario. http://www.statcan.ca [last
modification August 14, 2008; visited 2008/11/18]
Statistics Canada (2008d) Industrial water use, 2005 (Catalog no. 16–401-XIE). March 2008,
Statistics Canada, Ottawa, Ontario. http://www.statcan.ca/bsolc/english/bsolc?catno ¼ 16–401-X
Stewardship Ontario (2006) Ontario blue box markets overview: 2003/2004 blue box materials
generated and municipally marketed, mass balance report. Stewardship Ontario, Toronto,
Ontario, March 2006. http://webservices.siriusweblabs.com/dotconnector/files/domain4116/
2003–2004%20Mass%20Balance%20Report%20-%20final.pdf
Stewardship Ontario (2007) Ontario generation and recovery of blue box materials by demo-
graphic type (based on waste composition study results). Stewardship Ontario, Toronto,
Ontario. http://www.stewardshipontario.ca/pdf/eefund/gen_and_rec_rates.pdf
Tchoukanova N, Gonzalez M, Poirier S (2003) Best management practices: marine products
processing. Prepared for fisheries and oceans Canada – Gulf region. Fisheries and Marine
Products Division of the Coastal Zones Research Institute inc., Shippagan, New Brunswick,
Canada. http://www.glf.dfo-mpo.gc.ca/os/effluents/bmp-e.pdf. Also available in html (revised
version) at: http://www.glf.dfo-mpo.gc.ca/os/effluents/bmp-e.php.
US Department of energy (2004) Energy efficiency and renewable energy. Industrial technologies
program, industrial energy systems: food and beverage energy footprint. http://www.eere.
energy.gov/industry/energy_systems/footprints.html [content updated 18 August 2004;
accessed 17 September 2004]
US Environmental Protection Agency (2007) Energy trends in selected manufacturing sectors:
opportunities and challenges for environmentally preferable energy outcomes, final report.
Prepared by ICF international for US Environmental Protection Agency (EPA)’s office of policy,
economics, and innovation, sector strategic division, Washington, DC. March 2007. http://www.
epa.gov/sectors/energy/index.html; http://www.epa.gov/sectors/pdf/energy/report.pdf
Verfaillie HA, Bidwell R (2000) Measuring eco-efficiency. A guide to reporting company
performance. World Business Council for Sustainable Development (WBCSD), Geneva,
Switzerland. 38 pp. http://www.wbcsd.org
Chapter 9
Food Process Economics
George Saravacos and Zacharias Maroulis
9.1 Importance of Economics in Food Processing
Food processing is a large industry consisting of several small or medium size

plants and a few large installations. It produces a variety of food products that
satisfy the nutritional and sensory needs of consumers. Economics plays a very
important role in the efficient design of food processes and the production of
sufficient quantities of food products for an ever expanding world population
(Maroulis and Saravacos 2007).
The design and optimization of food processes is based on the principles of food
science and the established techniques of chemical process design, and simplified
computer spreadsheet techniques have been applied to the design of various food
processes (Maroulis and Saravacos 2003). For example, heat and mass transfer food
processes, such as heating, cooling, freezing, evaporation, dehydration, extraction,
and membrane separations, can be designed and optimized using transport proper-
ties and other engineering data of foods, as published in the literature (Saravacos
and Maroulis 2001; Rao et al. 2005; Rahman 1995). On the other hand, mechanical
processes involving solid and semi-solid food materials, such as grinding,
mixing, and mechanical separations or forming, are designed empirically using
specialized processing equipment provided by equipment manufacturers and sup-
pliers (Saravacos and Kostaropoulos 2002).
Food plant design is the integration of process design and process engineering
economics (Lopez-Gomez and Barbosa-Cánovas 2005). Food processing plants are
characterized by strict hygienic (sanitary) and food safety requirements, which
should be satisfied in any economic analysis. However, recent worldwide demand
for fresh processed foods at affordable prices stresses the importance of process
economics.
G. Saravacos (*) and Z. Maroulis

School of Chemical Engineering, National Technical University of Athens, 15780 Athens, Greece
e-mail: gsaravac@otenet.gr, maroulis@mail.ntua.gr
220 G. Saravacos and Z. Maroulis
9.2 Process Engineering Economics
Process engineering economics is based on the estimation of capital and operating

costs, on which plant profitability can be estimated. It is applied in designing
new processing plants, and in process and product improvement of existing instal-
lations (Clark 1997; Holland et al. 1997; Peters et al. 2003; Couper 2003). Safety
and environmental requirements of the process should be considered.
9.2.1 Capital Cost
The capital cost is based on the equipment cost (Ceq), which is estimated from
material and energy balances on the process flowsheet. Cost of food processing
equipment is estimated from equipment size (surface area, volume, weight) using
empirical correlations or the suppliers’ data (Maroulis and Saravacos 2007). Sizing
of food processing equipment is based on material and energy balances of the
process flowsheet, using simplified design equations and engineering property data
(Saravacos and Kostaropoulos 2002). Empirical correlations (diagrams) of cost
versus equipment size (Guthrie charts) can be used for engineering calculations.
Cost quotations of specialized food processing equipment can be obtained from
equipment suppliers.
Typical cost: capacity diagrams of food processing equipment include fluid
transport-power (kW); vessel-volume (m3); conveyor belts-surface area (m2);
heat exchanger-surface area (m2); dryers-surface area (m2); filters-surface area
(m2); size reduction-capacity (kg/s); mechanical processing-capacity (kg/s); and
utilities-kW (Maroulis and Saravacos 2007). Increases in cost of processing equip-
ment (inflation) over the years are taken into account using the Marshall and Swift
(M&S) cost index, published periodically in the journal Chemical Engineering,
e.g., M&S Index: (1990) ¼ 915, (2005) ¼ 1,261.
9.2.2 Operating Cost
The operating cost includes the direct cost of raw and packaging materials, labor,
and utilities, and various indirect costs such as maintenance, quality assurance, and
office personnel.
9.2.2.1 Raw Food Materials and Packaging Materials
The cost of raw food materials is estimated with data obtained from US government
organizations such as the Bureau of Labor Statistics (http://www.bls.gov) and the
9 Food Process Economics 221
National Agricultural Statistics Services (http://www.nass.usda.gov). Typical farm

prices in the US (2003) for food materials are ($/kg): wheat 0.13, corn 0.09,
potatoes 0.14, milk 0.27, beef 1.61, chicken 0.72, hogs 0.78, store-bought fresh
tomatoes 0.76, tomatoes for processing 0.10, oranges 0.10, grapes 0.40, peaches
0.40, and apricots 0.65. Both retail and farm prices, or retail to farm price ratios (r),
should be considered. Typical (r) values in the US are: eggs 1.7, milk 4, sugar 5,
wheat flour 5.5, margarine 10, potato chips 15, bread 25, corn flakes 28, and corn
syrup 35.
The cost of packaging materials may run from relatively low in bulk packing of
fruit products in plastic-lined drums to very high in plastic cups for dairy products
(Maroulis and Saravacos 2007). Intermediate costs include using metallic cans, glass
bottles, and paper/plastic bags. The packaging cost per unit mass of food products ($/
kg) increases substantially as the package size is reduced (e.g., plastic cups 0.80, glass
bottles 0.20, metallic cans 0.12, plastic bags 0.10, and 208 L plastic drums 0.09).
9.2.2.2 Labor
Labor requirements include production and nonproduction workers, and amount to

about 20% of the manufacturing cost. The process labor required can be estimated
from the process flowsheet and the material and energy balances. The supporting
labor is usually estimated using empirical correction factors and the total personnel
is approximately equal to three times the process labor. A labor cost model (as in
following example) has been proposed for estimating the total labor cost, using
correction factors for the country (i.e., cultural) effect, skilled personnel, supervi-
sion, social benefits, and overtime work.
The annual operating time (hours) is estimated as the product of (hours per
shift) (shifts per day) (days per week) (weeks per year). Food plants can
be operated for one or more seasons. The annual operating time for one season
can vary from 500 to 2,100 h, with workers employed 12.5 weeks a year and
5–7 days a week, 1–3 shifts a day. A two-season operation (25 weeks) amounts
to 1,000–4,200 h a year, while an all-year operation (50 weeks) amounts
to 2,000–8,400 h a year. Food preservation plants usually operate 2–4 months
per year, due to seasonal availability of agricultural raw materials. Some food
manufacturing plants may operate 12 months a year, e.g., bakeries. Food ingre-
dients plants may operate several months a year, depending on the availability of
raw materials. Bulk storage of large volumes of raw materials (e.g., wheat, corn,
and oilseeds) near a food plant can extend its operating time substantially.
The manpower requirements of a food processing plant can be estimated
analytically by summation of the requirements for each processing equipment
or operation (Maroulis and Saravacos 2007). Approximate personnel require-
ments for various food processing plants can also be estimated from the following
empirical equation developed by regression analysis from published data
(Bartholomai 1987).
M ¼ 10F2=3 (9.1)
where M is the manpower (number of personnel) and F is the plant product capacity (t/h).
The cost of labor primarily depends on the country and the degree of specializa-
tion. Labor rates are found in government publications such as the US Bureau of
Labor Statistics. The average hourly rate for production workers in the US was $15
in 2008. Higher rates are found in some European countries, such as Germany and
Norway, but substantially lower rates are observed in the rest of the world, with
very low rates in underdeveloped countries. The cost of skilled operators, mechan-
ics, technicians, foremen, and plant managers is normally estimated as 2, 3, 3.5, 4,
and 5 times, respectively, the cost of unskilled labor.
9.2.2.3 Utilities
The main public utilities used in food processing plants are energy and water. Energy
includes fuel (oil, gas) and electricity, and it is conveniently expressed in kWh/kg
product (1 kWh ¼ 3.6 MJ). Water is used as process water, and in cleaning opera-
tions, cooling, refrigeration, and steam generation; it is also involved in wastewater
effluents. The cost of fuel ($/kWh) is a function of the cost of crude oil ($/barrel
(bbl) ¼ 159 L oil). The cost of crude oil is affected strongly by international political
crises and wars (e.g., increasing from $60 to $100/bbl between the years 2007 and
2009). An empirical model was developed for the estimation of various energy forms
from the price of crude oil (Maroulis and Saravacos 2007). Typical diagrams of
energy cost of crude oil at $100/bbl and $60/bbl are shown in Figs. 9.1 and 9.2.
Typical energy cost (US 2008) follows: electricity – $0.10/kWh; steam – 10 bar
$08/kWh ($40/t); fuel oil – $1.30/L; and natural gas – $0.80/m3. The cost of food
processing water (cleaning of or adding to food products) is about $0.50/m3, while
boiler water is more expensive ($2.50/m3) due to its pretreatment. Waste disposal
costs are about $50/t, while wastewater treatment (primary and secondary) costs
about $1.00/t.
9.2.3 Process Profitability
9.2.3.1 Capital Cost
The total capital cost CT is covered partly by the investor’s own capital CO and
partly by a bank loan CL:
CT ¼ CO þ CL (9.2)
The loan is expressed as a fraction (leverage, l) of the total capital (l ¼ CL/C T).
A typical leverage in food processing is l ¼ 0.50. The total capital cost CT invested
0.20
Refrigeration
Thermal systems
0.15
Purchased electricity
Utility cost ($/kWh)
Self-generated electricity
0.10
Steam
0.05
Crude oil cost ($/bbl) = 100.0
Cooling water
0.00
–200 –100 0 100 200 300 400 500 600 700 800
Temperature (°C)
Utility Description $/kWh $/Common unit

Electricity Purchased 0.117
Self-generated 0.120
Crude 100.0 $/bbl
Fuels Fueloil 0.096 0.96 $/L
Natural gas 0.64 $/m3
Coal 65.0 $/t
Thermal systems 300°C 0.116
400°C 0.125
600°C 0.148
Steam 40 bar 0.105 52.5 $/t
10 bar 0.076 37.8 $/t
5 bar 0.061 30.4 $/t
1 bar 0.027 13.3 $/t
Cooling water tower 0.012 0.271 $/m3
river or sea 0.231 $/m3
well 0.312 $/m3
Refrigeration 5°C 0.081
–20°C 0.102
–50°C 0.150
Fig. 9.1 Energy cost of oil priced at $100/bbl
in a processing plant consists primarily of the fixed capital cost CF and, secondly,
the working capital Cw:
0.20
Refrigeration
0.15
Utility cost ($/kWh)
Purchased electricity
0.10
Thermal systems
Self-generated electricity
0.05
Steam
Crude oil cost ($/bbl) = 60.0
Cooling water
0.00
–200 –100 0 100 200 300 400 500 600 700 800
Temperature (°C)
Utility Description $/kWh $/Common unit

Electricity Purchased 0.103
Self-generated 0.078
Crude 60.0 $/bbl
Fuels Fuel oil 0.062 0.63 $/L
Natural gas 0.42 $/m3
Coal 65.0 $/t
Thermal systems 300°C 0.076
400°C 0.081
600°C 0.096
Steam 40 bar 0.068 34.2 $/t
10 bar 0.049 24.6 $/t
5 bar 0.040 19.8 $/t
1 bar 0.017 8.7 $/t
Cooling water tower 0.010 0.238 $/m3
river or sea 0.203 $/m3
well 0.274 $/m3
Refrigeration 5°C 0.072
–20°C 0.089
–50°C 0.132
Fig. 9.2 Energy cost of oil priced at $ 60/bbl

CT ¼ CF þ Cw (9.3)
The fixed capital cost CF includes the cost of processing equipment Ceq, instal-
lation, piping, instrumentation and control, electrical installations, buildings, land
and site improvements, office facilities, engineering, and contingency. The working
capital Cw consists of the total amount of money invested in raw materials and
supplies in stock, finished and semi-finished products, accounts receivable and
payable, and cash kept on hand. The working capital amounts to 15–50% of the
total cost. In general, food manufacturing plants require higher proportions of
working capital than food preservation operations. The fixed cost CF is estimated
from the equipment cost, using the empirical equation,
CF ¼ fL Ceq (9.4)
where fL is the Lang factor. The empirical Lang factor fL in older plants (Bartholomai,
1987) is approximately fL ¼ 2, but in recent installations, using more instrumenta-
tion, process control and computers, values are fL ¼ 3 for main plants and fL ¼ 4 for
grassroots plants (including off-site facilities) (Maroulis and Saravacos 2007).
9.2.3.2 Manufacturing Cost
The manufacturing cost CM of a processing plant consists primarily of the variable

cost of raw and packaging materials, labor, and utilities. It also includes the fixed
cost of maintenance, insurance, taxes and royalties, and the indirect cost of sales
and general expenses. The total manufacturing cost CMT or total annualized cost
TAC includes the effect of capital cost, as given by the following equation:
CMT ¼ CM þ eCT (9.5)
where CT is the total capital cost, and e is the capital recovery factor, as calculated
from the following equation:
.h i .h i
e ¼ i 1 ð1 þ iÞN ¼ ið1 þ iÞN ð1 þ iÞN 1 (9.6)
The capital recovery factor is a function of the discount (interest), rate i, and the
lifetime of the investment N (years).
9.2.3.3 Discounted Cash Flow
The annual gross profit before taxes PG is the sales income S minus the
manufacturing cost CM:
P G ¼ S CM (9.7)
The cumulated cash flow (CCF) or annual cash flow is defined by the simplified
equation,
CCF ¼ CO þ N P (9.8)
where CO is personal capital and P the annual profit after taxes and loan payment,
assumed to be constant over the operating time N (years). The investor’s own capital is
the total capital minus the bank loan. The discounted cash flow or the net present value
(NPV) is lower than the CCF, because it includes the interest rate i of the money:
NPV ¼ CO þ P=e (9.9)
where e is the capital recovery factor, defined in (9.5). In the simplified (9.7) and
(9.8), the annual profit after taxes and loan payment P is assumed to be constant
over the years. When P varies over time, the following summation equations,
similar to (9.8) and (9.9), are used to estimate CCF and NPV:
X
N
CCF ¼ CO þ Pn (9.10)
n¼1
X
N
Pn
NPV ¼ CO þ (9.11)
n¼1
ð1 þ iÞn
9.2.3.4 Measures of Plant Profitability
The following measures are used to evaluate plant profitability:

(a) Cash: CCF (non-discounted) and NPV (discounted)
(b) Time: simple payback period, SPB ¼ CT/P (non-discounted)
Discounted payback period, DPB ¼ ln((1 iSPB)1)/ln(1 þ i)
(c) Rate: return on investment, ROI ¼ 1/SPB ¼ P/CT (non-discounted)
Internal rate of return, IRR: ROI ¼ IRR/(1 (1 þ IRR)N) (discounted)
The results of profitability analysis are presented in diagrams of CCF and NPV
versus time (years), and break-even analysis of (annual income or sales) and
(manufacturing cost or expenses) versus annual operating time. The latter diagram
shows the optimum operating time for maximum profit for the specific food plant
(Maroulis and Saravacos 2007). The values of SPB and DPB can be estimated from
the cash flow diagrams as the intersection of the CCF and NPV curves with the time
axis, respectively.
9.3 Food Processing Plants
For the purpose of economic analysis, food processing plants can be categorized as
follows: food preservation, food manufacturing, and food ingredients plants. Food
engineering is mainly concerned with food preservation and food manufacturing
plants, while food ingredients plants are more related to chemical engineering.
Application examples of some typical plants are presented in this section from
process data in the literature (Maroulis and Saravacos 2007).
9.3.1 Food Preservation Plants
Food preservation processes include concentration, dehydration, freezing, and

thermal processing (pasteurization, canning). Most plants are located near the
production of raw materials (fruits and vegetables, animal products) and usually
operate only for one season (e.g., 2–3 months a year). Application examples of two
typical food plants are presented in this section, from process data in the literature
(Maroulis and Saravacos 2007).
The economics of an orange juice concentrate (OJC) is summarized here as an
application example. Figure 9.3 shows a block diagram of the process with material
and energy balances based on 1 kg of 65% TS (total solids) OJC, using literature
data on fruit composition and process technology. The requirement for raw material
is 10.35 kg oranges of 13.6% TS. Figure 9.3 shows the material and energy balances
of the orange juice concentrate OJC processing plant.
The extracted juice is concentrated from 12% to 65% TS, using a four-effect
falling film evaporator. The OJC is packed in plastic drums and stored at 0 C. The
orange peels are used to produce press oil and then dehydrated in an air dryer to
produce animal feed with 10% moisture. The theoretical energy requirements for
the process are 8 MJ/kg of product. Assuming a thermal efficiency of 75%, the
actual energy requirement is about 10.7 MJ/kg or 3 kWh/kg of product.
The plant operates seasonally when ripe oranges are available for processing,
which is 16 weeks (or about 4 months) a year. The annual operating time is assumed
to be 1,280 h/y and the annual production 1,280 t/year OJC (16 weeks/year 5
days/week 2 shifts/day 8 h/shift). Table 9.1 shows the annual material bal-
ances for the orange juice concentrate plant based on a production rate of 1 t/h OJC.
The OJC product is packed in 5,565 plastic drums of 230 kg capacity, amounting
to 5,565 0.23 ¼ 1,280 t/year OJC. The required processing equipment, includ-
ing cleaning of oranges, juice extraction and finishing, juice evaporation, aseptic
packing, and peel drying, were sized and priced following simplified engineering
procedures (Saravacos and Kostaropoulos 2002). The equipment cost was $1.62
million. Table 9.2 shows the capital cost components of the OJC plant. The working
capital is assumed to be 25% of the fixed.
Fig. 9.3 Block process diagram of the OJC plant

Table 9.1 Annual material Oranges 13,890 t/year

balances of the OJC plant OJC 1,280 t/year
Dried peels 1,150 t/year
Peel oil 38 t/year
Table 9.2 Capital cost of Process equipment, Ceq 1.62 M$

the OJC plant Lang factor, fL 3
Fixed capital, CF 4.86 M$
Working capital, CW 1.21 M$
Total capital, CT 6.07 M$
Table 9.3 Annual operating Raw materials (0.12$/kg oranges) 1.67 M$

cost of the OJC plant Labor 0.72 M$
Packaging materials (20$/drum) 0.11 M$
Utilities 0.65 M$
Waste treatment 0.07 M$
Variable manufacturing, CMV 3.22 M$
Fixed manufacturing, CMF 0.49 M$
Overheads, Cover 0.23 M$
Manufacturing cost, CM 3.94 M$
Capital recovery factor, e 0.083 (i ¼ 0.07, N ¼ 27)
Capital charge, eCT 0.50 M$
Total annualized cost, CM + eCT 4.44 M$
The annual operating cost of the OJC plant was calculated from material balances,
cost data, and empirical correlations, as shown in Table 9.3 (2006 prices). It is assumed
that the discount (interest) rate is i ¼ 0.07 and the lifetime of the plant is N ¼ 7 years.
Table 9.4 summarizes the plant profitability data for the OJC plant. The inves-
tor’s own capital C0 is assumed to be 50% of the total CT. The capital return ratio is
defined as CRR ¼ NPV/C0.
The annual sales income is based on an OJC product price of $3.60/kg, which it
is assumed includes the income from the dried peels and peel oil. Figure 9.4 shows
the cost components of the orange juice concentrate. The raw material (oranges)
is the main cost item (36%), followed by labor (20%) and utilities (15%). Food
manufacturing (equipment) cost is about 19% of the total, while bulk packaging in
large containers is a relatively low cost. Overheads and capital charge amount to
18% of the total cost.
The relatively high cost of utilities in the OJC plant is due to the removal of large
amounts of water by evaporation and dehydration, requiring large amounts of steam
and fuel gas. Figure 9.5 shows the cash flow, the cumulative cash flow (CCF), and
the net present value (NPV) of the OJC plant as a function of operating time N
(years). The values of depreciation period (ND ¼ 7 years), loan period (NL ¼ 15
years), salvage period (NS ¼ 20 years), and project lifetime (NE ¼ 27 years) are
also shown. The ROI value is estimated from the relation ROI ¼ 1/SPB.
Table 9.4 Plant profitability Annual sales income, S 4.61 M$/year

of the OJC plant Manufacturing cost, CM 3.93 M$/year
Gross profit, PG 0.68 M$/year
Net present value, NPV 1.30 M$
Own capital cost, C0 3.01 M$
Capital return ratio, CRR 0.43
Internal Rate of Return, IRR 0.12
Return On Investment, ROI 0.16
Simple payback period, SPB 6.3 years
Discounted payback period, DPB 10 years
4.00
3.50
Capital Charge
Fixed
3.00 Overheads
Manufacturing
Unit Cost ($/kg)
2.50
Labor
2.00 Waste
Treatment
Utilities
1.50
Packaging
1.00
Raw Materials
0.50
0.00
Operating Cost
Fig. 9.4 Operating cost of the OJC plant
Figure 9.6 shows a break-even analysis of the OJC plant. For the process system
analyzed here, the optimum operating time (maximum annual profit) is about
1,000 h/year.
9.3.2 Food Manufacturing Plants
Food manufacturing plants typically produce several food products packaged in

small consumer units, using agricultural raw materials or semi-finished food
Fig. 9.5 Cash flow and NPV of the OJC plant
materials and food ingredients. The plants are preferably located close to large
consumer centers, operate several months a year, and require more labor and
packaging materials than food preservation plants. Food manufacturing plants are
characterized by strict hygienic and food safety regulations, due to the sensitivity of
food products to microbial and chemical spoilage. Thus, they require special quality
control and compliance with government and international regulations. Products
10
1 shift 2 shifts 3 shifts + weekends
8
Annual income/outcome (M$/y)
Sales
4 Manufacturing cost
2
Profit
0
0 500 1000 1500 2000 2500 3000 3500
Annual operating time (h/y)
Fig. 9.6 Beak-even analysis of the OJC plan
produced in food manufacturing plants include cereal, dairy, fruit and vegetable,
animal, and prepared food products. It should be noted that these plants may use
some food preservation technology during processing and storage.
As a typical application example, the economics of yogurt manufacture in
producing consumer packages is discussed in this section. Figure 9.7 shows the
process block diagram and the material and energy balances for the manufacture of
dairy yogurt.
The material and energy balances in this example are based on the raw materials,
1,000 kg of milk and 40 kg of nonfat dry milk (NFDM). The yogurt plant operates
48 weeks a year, 5 days a week, with two shifts per day and 8-h shifts, amounting to
3,840 h/year. Table 9.5 shows the annual material balances for the yogurt
manufacturing plant, producing 4,000 t/year yogurt.
The yogurt product is packaged in plastic cups of 0.250 kg capacity, resulting in
15,976,000 cups/year. The theoretical energy requirement for the process is
0.73 MJ/kg product. Assuming a thermal efficiency of 75%, the actual energy
requirement is about 1 MJ/kg or 0.3 kWh/kg product. The required processing
equipment, including for milk storage, mixing, homogenization, heat treatment,
culture inoculation, and aseptic packaging, was sized and priced, following simpli-
fied engineering procedures (Saravacos and Kostaropoulos 2002). The equipment
cost was $2.49 million. Table 9.6 shows the capital cost components of the yogurt
plant. The working capital is assumed to be 25% of the sales.
The annual operating cost of the yogurt plant was calculated from material
balances, cost data, and empirical correlations, as shown in Table 9.7 (2006 prices).
Fig. 9.7 Process block diagram of the yogurt manufacturing plant
Table 9.5 Annual material Material product rate t/year

balances of the yogurt Milk 3,840
manufacturing plant NFDM 156.3
Stabilizer 76.8
Fat removed 76.8
Yogurt 3,994
Table 9.6 Capital cost of the Process equipment, Ceq 2.49 M$

yogurt manufacturing plant Lang factor, fL 3
Fixed capital, CF 7.47 M$
Working capital, CW 5.05 M$
Total capital, CT 12.52 M$
Table 9.7 Annual operating Raw materials (0.35 $/kg milk) 7.68 M$
cost of the yogurt Labor 3.24 M$
manufacturing plant Packaging materials (0.04$/piece) 3.84 M$
Utilities 0.32 M$
Variable manufacturing, CMV 15.09 M$
Fixed manufacturing, CMF 0.75 M$
Overheads, Cover 1.01 M$
Manufacturing, cost CM 16.85 M$
Capital recovery factor, e 0.083 (i ¼ 07, N ¼ 27)
Capital charge, eCT 1.04 M$
Total annualized cost CM + eCT 17.89 M$
Table 9.8 Plant profitability Annual sales income, S 20.18 M$/year

of the yogurt manufacturing Manufacturing cost, CM 16.84 M$/year
plant Gross profit, PG 3.34 M$/year
Net present value, NPV 16.99 M$
Own capital cost, C0 6.26 M$
Capital return ratio, CRR 2.72
Internal Rate of Return, IRR 0.32
Return On investment, ROI 0.40
Simple payback period, SPB 2.5 years
Discounted payback period, DPB 3 years
It is assumed that the discount (interest) rate is i ¼ 0.07 and the lifetime of the plant
is N ¼ 27 years.
Table 9.8 summarizes the plant profitability data for the yogurt plant. The
investor’s own capital C0 is assumed to be 50% of the total CT.
The annual sales income is based on a yogurt product price of $1.25/kg.
Figure 9.8 shows the cost components of the yogurt manufacturing plant. The
raw material (milk) is the major cost item (43%), followed by packaging materials
(23%), labor (18%), and processing equipment (5%). Utilities, overheads, and
capital charge amount to about 11% of the total cost.
The cumulative cash flow (CCF) and the net present value (NPV) plots of the
yogurt plant are similar to the diagrams of the OJC plant. Break-even analysis of the
yogurt plant resulted in a diagram similar to Fig. 9.6, indicating a higher optimum
operation time (maximum annual profit) of about 2,500 h/year.
Fig. 9.8 Operating cost of the yogurt manufacturing plant
9.3.3 Food Ingredients Plants
Food ingredients plants utilize agricultural and natural raw materials to recover
valuable components, which are then used in food manufacturing. Typical products
are wheat flour, vegetable oils, starch, pectin, sugar, and proteins. Special food
components used in smaller quantities include flavorings, vitamins, coloring mate-
rials, sweeteners, preservatives, and antioxidants.
9.4 Conclusions
Applying the principles of process economics to food processing can lead to useful
information on the economics of food processes. Simplified procedures based on
material and energy balances, sizing of process equipment, and cost data on raw
materials, labor, and utilities result in new food process designs and economic
evaluation of existing processing operations. Raw materials, followed by labor, are
the major cost items in most food production processes. Packaging cost also may be
important for some food manufactured products. Food products manufactured from
raw food materials and food ingredients are generally more profitable than products
of food preservation plants.
References
Bartholomai A (1987) Food factories – processes, equipment, costs. VCH Publishers, Weinheim
Clark JP (1997) Cost and profitability estimation. In: Valentas KL, Rotstein E, Singh RP (eds)
Handbook of food engineering practice. CRC Press, New York, pp 537–557
Couper JR (2003) Process engineering economics. Marcel Dekker, New York
Holland FA, Wilkinson JK (1997) Perry’s chemical engineers’ handbook. In: Perry RH, Green
DW, Maloney JO (eds) Process economics, 7th edn. McGraw-Hill, New York, pp 9.1–9.63
Lopez-Gomez A, Barbosa-Cánovas G (2005) Food plant design. Taylor & Francis, New York
Maroulis ZB, Saravacos GD (2003) Food process design. Marcel Dekker, New York
Maroulis ZB, Saravacos GD (2007) Food plant economics. CRC Press, Boca Raton, FL
Peters SM, Timmerhaus KD, West RE (2003) Plant design and economics for chemical engineers,
5th edn. McGraw-Hill, New York
Rahman S (1995) Food properties handbook. CRC Press, New York
Rao MA, Rizvi SSH, Datta AK (2005) Engineering properties of foods, 3rd edn. Taylor & Francis,
New York
Saravacos GD, Maroulis ZB (2001) Transport properties of foods. Marcel Dekker, New York
Saravacos GD, Kostaropoulos AE (2002) Handbook of food processing equipment. Kluwer/
Plenum, New York
Chapter 10
Systemic Approach to Curriculum Design
and Development
Inés Ecima, Maurcio Pardo, and Gloria González-Mariño
10.1 Introduction
Engineering leaders and educators consider innovation to be an important topic in

any program related to the training of engineers (Apelian 2007; National Academy of
Engineering 2004). The development of traditional skills and knowledge are not the
most important goals of an engineering curriculum nowadays, and due to the
influence of technology and globalization, traditional engineering skills are now
considered a commodity. Identifying this trend, Thomas Friedman (2005) described
the modern world as “flat,” trying to illustrate the great impact of technology
and globalization on the economy of modern society. To make matters worse for
new engineering graduates, today’s computers are capable of doing almost every-
thing engineers have done up until now. As a consequence, it is important to reorient
engineering education in such a way that engineers are prepared to understand the
societal context of their work from both a local and global perspective. Innovation
and creativity also should be coupled with the engineer’s ability to gather informa-
tion, analyze it, make decisions, and take the right course of action.
Food engineering is a branch of engineering education that has inadvertently
addressed the need for innovation in meeting the demands of technological and
scientific advances in the new century. There are many challenges involved during
processing in how to preserve the original quality of food and its active ingredients,
which are considered good for humans. As a result, customization of food products
is actually one of the strongest trends in the food industry today (Higgins 2007;
Coulston et al. 2003). Thus, collaboration among food engineers, nutritionists,
medical doctors, microbiologists and microelectronic engineers, together with
other experts, has become an essential element in dealing with these challenges.
I. Ecima and M. Pardo

Ingenierı́a de Producción Agroindustrial, La Sabana University, Bogotá, Colombia
G. González-Mariño (*)
Biosciences Doctoral Program, La Sabana University, Bogotá, Colombia
e-mail: gloria.gonzalez@unisabana.edu.co
238 I. Ecima et al.
For one, it is necessary that engineers be capable of producing and offering

customized products at fair prices to the market. This customization should not only
attend to fair pricing and the consumers’ tastes but to their health needs as well.
Food engineers must be able to integrate new technologies into traditional proces-
sing lines in order to offer products nearer to consumer needs. They must also have
an ample view of the food sector and be capable of relating people and procedures
that affect the production and quality of food, by undertaking joint actions, making
faster decisions, and sharing their knowledge and experience with conforming
professional communities and peers located around the world.
To achieve this goal, engineering students should have flexible educational
programs that encourage a systemic approach to the food industry, programs that
allow them to travel the world and remain in permanent contact with their collea-
gues. These programs must also be developed according to international standards
so students are allowed to travel without interrupting their studies. Nowadays the
trend is to solve problems and make decisions from a broad perspective instead of
concentrating on a specific discipline. Generally, however, this does not mean that
the depth of understanding a specific scientific field is lost; rather it means that the
engineering professional is able to decide which basic science principles are
actually needed to solve a particular problem.
10.2 Systems Theory and Thinking
Systemic or systems theory has been taken into account by numerous scientific
fields including the field of education (Cabrera and Colosi 2008; Kim and Senge
1994; Ison 2008; Senge 1990). Although there are many conflicting claims that
need to be reconciled around this topic, such discussion is out of the scope of this
chapter. Therefore the classical definition proposed by Ludwig von Bertalanffy in
1968 is being used as reference:
Classical sciences tried to isolate the elements of the observed universe. . . Now we have
learned that for an understanding of not only the elements, but their interrelations as well are
required. . . It is necessary to study not only [the] parts and processes in isolation, but also to
solve the problems [organization and order] resulting from dynamic interaction of parts, and
making the behavior of the parts different when studied in isolation or within the whole. . .
General system theory, therefore, is a general science of wholeness. . . The meaning of the
somewhat mystical expression ‘The whole is more than the sum of its parts’ is simply that
constitutive characteristics are not explainable from the characteristics of the isolated parts.
The characteristics of the complex, therefore, appear as ‘new’ or ‘emergent’. . ..
Additionally, a system can be defined as a dynamic and complex whole interacting

as a unit, located within an environment in which energy, material, and information
flow between the different elements that compose the system and between the
elements and their surrounding environments. If the set of courses, methods, people,
and tools involved in a learning-teaching interaction are considered as a system, they
will show certain characteristics as summarized in Table 10.1.
10 Systemic Approach to Curriculum Design and Development 239
Table 10.1 Comparison between the characteristics of a basic system and those of the curriculum
as a system
System Curriculum
Dynamic and complex whole Dynamic and complex academic, social, technical,
interacting as a unit scientific and empiric knowledge interacting as a
unit
Energy, material, and information Topics, issues, theories, information, people,
flowing between the different feelings, perceptions, sciences, personal skills,
elements that compose the system and hidden things flowing between students,
teachers, parents, industry, among others
Community located within an Academic community immersed in a context
environment
Energy, material, and information Knowledge and agents interacting within the
flowing from and to the academic community and national and
surrounding environment international peer communities
Seeks equilibrium but can exhibit Seeks articulation but has to be flexible. Loop
oscillating, chaotic or exponential planning and developing-assessing changes in
behavior every course
According to O’Connor and McDermott (1998), systems thinking (i.e., the

practice of systems theory) comprises the whole and the parts as well as the links
among parts. Furthermore, it studies the whole in order to understand the parts. It is
a pattern of thinking that is not disciplinary in scope and can act as a bridge between
the physical, natural, and social sciences. Four universal conceptual patterns have
been observed in every system and can be used as tools to apply systems thinking in
any situation, as proposed by Cabrera et al. (2008):
1. System definition
2. Distinction from others
3. Perspective or frame of reference
4. Relationships between systems and parts
10.2.1 Designing Curriculum Using Systems Thinking
If the four universal principles just described are followed in the design of a
curriculum, the first step should be to define the framework and the system, its
parts and its limits. Therefore, a definition of the curriculum from the point of view
of the institution is needed. Additionally, the boundaries of this system should be
identified to differentiate it from others. Finally, relationships should be identified.
Table 10.2 describes the different aspects that should be taken into account in
defining the curriculum. Readers should note that even though Table 10.2 depicts a
step by step methodology this is far from a real procedure. Many steps need to be
carried out in parallel and many others will require revision after defining the whole
curriculum. Figure 10.1 gives a closer visual description of the methodology used
for systems thinking.
240 I. Ecima et al.
Table 10.2 Some steps taken into account during curriculum design
Activity Systems thinking principles used
Define program (field and duration) Perspective, System definition, and distinction
and level (undergraduate or
postgraduate)
Find similar programs (locally and System distinction
abroad)
Identify educational trends nationally System distinction
and internationally
Identify skills of student entering the Relationships (between system and environment)
program
Identify skills and competencies Relationships (between system and environment)
demanded by society
Identify institution’s educational Perspective and system definition (parts)
purposes
Define problematic and nuclear topics System definition (parts)
to be taught
Define courses System definition (parts)
Relationships (between parts of system)
Define learning-teaching strategy System definition (parts)
Identify institutional capacities System definition (parts)
(teachers, buildings, equipment,
software)
Define assessment strategy System definition (parts)
10.2.2 Designing Curriculum for Undergraduate Courses
Educational institutions are guided by educational goals, so the first step in designing
a curriculum is to establish the goals. At the School of Engineering at Universidad de
La Sabana, the main educational goal at the undergraduate level is to teach students
how to design and manage their own projects when they become professionals.
To achieve this goal, an interdisciplinary curriculum is designed that takes into
account the basic sciences, such as physics, chemistry, and biology, with mathe-
matics as the common language, which allows the engineers to represent different
phenomena. Study of the basic sciences should include the following objectives:
l Describe and study a phenomenon from the perspective of each basic science
using mathematical language
l Identify the principles, laws, and theories that explain the natural phenomenon
l Understand the phenomenon as a whole from different points of view
Additionally, graduates are expected to have the following competencies as a
result of their educational program. They should be able to:
l Adopt systems thinking as an attitude
l Identify food systems and their parts at different levels, such as tissues,
processed products, and supply chains
ENVIRONMENT
Similar Programs.
National and
International
Experience
Industrial Educational systems.
needs and National and
International
opportunities de Tec
ve hn
lo ic
pm al University
Educatio
en na l n
gie
s
t tio Re
ate
s
rses Pr earc
Pr
titu
Cou
Str
o h
Re fess ers,
oje
s or
gro earc s,
In s
dolo and
up h
s
gie
ct
s
Me ware
Infrastru
tho
Soft
Prospectus
cture
R
el wee
be
at n
t
io u
ns n
ls
skil
hi its
t’s
ps
den
Stu
Institutional Formative
Purposes
le ific
Pe
e
ow nt
dg
op
kn cie
le
S
Society needs and

Scientific development challenges
Fig. 10.1 Closer visual description used for curriculum design under systems thinking methodology
l Design processes taking into account the interactions between food materials
and other components of the food chain
l Manage food handling and processing projects by considering all factors of the
network
During this learning – teaching period, the educator should also assist students in
finding the linkages between the concepts and reality.
10.2.3 Designing Curriculum for a Master’s Degree Program
At this level the educational goal is to train students to be professionals, as capable

of solving an industrial problem in an innovative way. To innovate means to
perform a creative solution in order to obtain a new product. In the end, these
242 I. Ecima et al.
Fig. 10.2 Biosciences

curriculum development
Core Biosciences:
Biology Principles:
Functionality
Biochemistry
Physics
- Genetic
Molecular information flow
Chemistry Biology -Bioenergetics and
Transportation
-Intra and inter
Mathematics Genetics cellular
communication
professionals who come from different disciplines should have in common their
capacity to solve problems using systems thinking and integrating the different
sciences. The training process has the following purposes:
l Develop the student’s ability to solve problems using an integration of sciences
approach.
l Develop the student’s creativity and design capacity.
l Develop the student’s ability in mathematical modeling.
In the case of the Master’s program in Process Design and Management, the
former training purposes are not only competencies to develop but are also specific
courses wherein students can find practical situations to develop these skills.
10.2.4 Designing Curriculum for a Doctoral Program
At the doctoral level, the educational goal is to train researchers via different
disciplines. For example, in the biosciences program, the goal is to guide students
during the study of biological elements (molecules) and their vital functions, in such
a way, that they are able to discover new and different uses for these molecules. To
achieve this goal, the biosciences course uses a systems thinking approach to review
different functions such as transport, signaling, and reproduction in cells. This
systems thinking approach is supported by a permanent analysis-synthesis activity
using concepts from molecular biology, biochemistry, and genetics (Fig. 10.2).
10.3 Conclusion
The four universal principles of systems thinking have been used at Universidad de
la Sabana in curriculum design for both undergraduate and postgraduate levels.
Additionally, systems thinking has been introduced as an educational goal in order
to give engineers more skills in defining systems (their parts and interactions),
which will serve as a tool in problem solving and decision taking.
References
Apelian D (2007) The engineering profession in the 21st century – educational needs and societal
challenges facing the profession. AFS Hoyt Memorial Lecture. American Foundry Society.
International Journal of Metalcasting Fall 07
Bertalanffy L von (1968) General system theory: foundations, development, applications. Braziller,
New York
Cabrera D, Colosi L (2008) Distinctions, systems, relationships, and perspectives (DSRP):
a theory of thinking and of things. Eval Program Plann 31:311–334
Cabrera D, Colosi L, Lobdell C (2008) Systems Thinking. J Eval Prog Plan 31:299–310
Coulston AM, Feeney MJ, Hoolihan LE (2003) The challenge to customize. J Am Diet Assoc
103(4):443–444
Friedman TL (2005) The world is flat: a brief history of the twenty-first century. Farrar, Straus &
Giroux, New York
Higgins KT (2007) Meeting the challenges of customized manufacturing. BNP Media. Food
Engineering online magazine: http://www.foodengineeringmag.com/
Ison RL (2008) Systems thinking and practice for action research. In: Reason P, Bradbury H (eds)
The Sage handbook of action research participative inquiry and practice. Sage, London,
pp 139–158. ISBN 1-4129-2029-9, 978-1-4129-2029-2
Kim DH, Senge P (1994) Putting Systems thinking into practice. Syst Dyn Rev 10(2–3):277–290
National Academy of Engineering (2004) Educating the engineer of 2020: Adapting engineering.
Education to the New Century. Available for free online: http://fermat.nap.edu/catalog/11338.
html
O’Connor J, McDermott I (1998) The art of systems thinking. Thorson, London
Senge P (1990) The fifth discipline. Doubleday Dell Publishing, New York
.
Part II
Advances in Food Process Engineering
.
Chapter 11
Innovations in Thermal Treatment of Food
Arthur Teixeira
11.1 Introduction
Thermal processing (heat treatment) for sterilization of shelf-stable foods has been
one of the most widely used methods of food preservation during the twentieth
century, and has contributed significantly to the nutritional well-being of much of
the world’s population. For most solid and semi-solid foods, thermal processing is
accomplished after the product has been filled and hermetically sealed in airtight
containers. Thermal processing consists of heating food containers in pressurized
retorts at specified temperatures for prescribed lengths of time. These process times
are calculated on the basis of achieving sufficient bacterial inactivation in each
container to comply with public health standards and to ensure that the probability
of spoilage will be less than some minimum. Many liquid food products, such as
milk or fruit juices, are pasteurized or sterilized by heat treatments applied before
filling and sealing into packages. These heat treatments are accomplished by
pumping the liquid product through a series of heat exchangers and hold tubes
that deliver a high temperature-short time (HTST) heat treatment for pasteurization,
or an ultra-high temperature (UHT) short time treatment for sterilization.
This chapter presents a review of recent innovations that have been made in the
field of thermal processing for food preservation, and attempts to foresee the impact
these innovations may have on the marketplace. Three categories of innovation are
addressed: (i) improved methods in estimating kinetic parameters for establishing
optimum process conditions, (ii) a review of new equipment systems for retorting
and materials handling to improve cookroom operations, and (iii) new and novel
packaging systems, and their potential impact on markets for shelf-stable foods.
A. Teixeira
Agricultural and Biological Engineering Department, University of Florida, Gainesville,
FL 32611–0570, USA
e-mail: atex@ufl.edu
248 A. Teixeira
11.2 Microbial Kinetics for Process Calculations
The manner in which populations of microorganisms decrease in response to lethal

heat treatments is fundamental to the engineering design of thermal inactivation
processes (pasteurization and sterilization) and important in the food, pharmaceutical,
and bioprocess industries. In order to determine the optimum process conditions
needed to achieve desired results, the effect of these conditions on rates (kinetics) of
population decrease needs to be characterized and modeled mathematically. This
chapter section will focus on the use of new methods for estimating kinetic para-
meters more accurately. Use of more accurate parameters in appropriate mathemati-
cal models will enhance the accuracy of these models. When properly developed and
validated, these models can predict the extent to which a population of viable spores
will be reduced in response to a lethal heat treatment at specified temperature and
time. These models are essential tools in establishing process conditions, i.e., time
and temperature, needed to achieve a specified level of bacterial lethality.
Thermal inactivation of bacteria generally follows first-order kinetics; it can be
described by logarithmic reduction in the number of bacterial spores with time during
exposure to a constant lethal temperature. The logarithm of number of bacterial
spores when plotted against time normally produces a straight line known as a
survivor curve. The slope of this curve will give the first order rate constant at this
temperature. Alternatively, the reciprocal of this rate, known as decimal reduction
time, D, is expressed as time in minutes to achieve one log cycle reduction in spore
population when subjected to a specified constant lethal temperature.
The temperature dependency of the rate constant is also an exponential function
that can be described by the Arrhenius equation. This is the equation used for a
straight line on a semi-log plot when the natural log of the rate constant is plotted
against the reciprocal of absolute temperature. Likewise, the temperature depen-
dency of the decimal reduction time, D, is also logarithmic (straight line on a semi-
log plot of D versus temperature within the lethal range), and can be expressed as
the temperature difference, Z, required for the curve to transverse one log cycle.
Thermal death-time kinetic parameters “D” and “Z” are widely used in the food
science community, while the first order rate constant and the Arrhenius relation-
ship are most often used in the chemical and biochemical communities. In either
case, these parameters must be estimated as accurately as possible to calculate
optimum thermal process conditions that ensure public safety with maximum product
quality.
Traditional methods for estimating these kinetic parameters consist of plotting a
series of survivor curves at different lethal temperatures, and measuring the slope or
D-value from each curve. Data for these curves are obtained from carefully executed
microbiology laboratory procedures. Typically, a series of small glass vials contain-
ing a known high concentration of viable spores are immersed simultaneously into a
heated oil bath maintained at a known constant lethal temperature. At predetermined
time intervals one of the vials is quickly removed and immediately cooled until all
vials have been removed. Thus, each vial contains spores exposed to a lethal
11 Innovations in Thermal Treatment of Food 249
temperature for a different length of time. The contents of each vial are plated onto
appropriate growth media and incubated for subsequent enumeration of survivors.
The logarithm of number of survivors from each vial is plotted against time to
produce the log-linear survivor curve, and the D-value at that temperature is taken
from the slope of that curve.
The same experiment is repeated at different constant lethal temperatures in
order to obtain D-values at different temperatures, from which the temperature
dependency factor, Z-value, can be determined. These methods are prone to error
because of heat transfer limitations causing thermal lag in the temperature response
in each vial. The spores themselves do not rise instantly to the oil bath temperature,
nor do they cool instantly in response to the quenching temperature. In those cases
where the kinetics are needed for mesophylic spores in a liquid product, errors can
be minimized by use of a three-neck flask containing the liquid product heated to a
known constant lethal temperature. One neck is used to accommodate a thermometer
for temperature measurement, a second neck holds an inoculation needle through a
rubber stopper, and the third neck holds a number of extraction needles for
removing and quenching samples of spore suspension at predetermined time inter-
vals (Rodriguez et al 1988). Although this method minimizes errors from lag in
thermal response, it is still prone to errors in timing, particularly when samples must
be extracted within a few seconds of each other at higher temperatures where the
lethal rate is very high.
An entirely new and different experimental approach for estimating kinetic
parameters in liquid products was first developed by Swartzel (1984) and called
the Equivalent Point Method (EQM). It was later modified by Welt et al (1997), and
called the Paired Equivalent Isothermal Exposure (PEIE) method. These methods do
away completely with the use of constant temperature baths (isothermal experi-
ments). Instead, a batch of liquid product inoculated to a known high initial concen-
tration of spores is “processed” through an HTST or a UHT heat exchanger-hold
tube system that delivers a precisely known dynamic temperature-time profile
experienced by the inoculated product. Samples of the “processed” product are
plated, incubated, and enumerated to determine the number of survivors, thus giving
the extent of reaction accomplished. By knowing the extent of reaction from two
different sets of process conditions and the precise dynamic temperature-time
profiles of each, a nearly “true” estimate of kinetic parameters can be determined.
This was confirmed by work carried out at the University of Florida, in which
kinetic parameters for thermal inactivation of Escherichia coli in orange juice were
estimated using the new PEIE method with a Microthermics® HTST Lab 25 pasteur-
izer, as well as the traditional isothermal bath method with the three-neck flask
(Moody 2003). D-values obtained from both methods at different temperatures are
shown in Table 11.1. For the two temperatures used in common by both methods
(58 and 60 C), D-values estimated by the new PEIE method were lower than those
estimated by the traditional isothermal bath method. Lower D-values mean faster
kinetics and would justify shorter process times if known to be true. In order to
determine which method produced estimates closest to the “truth,” inoculated sam-
ples of orange juice were processed through the HTST pasteurizer at temperatures
250 A. Teixeira
Table 11.1 Comparison of kinetic parameters (D- and Z-values), estimated using traditional
isothermal and new PEIE methods (Reproduced from Moody 2003. With permission)
Temperature (C) Isothermal (Three-neck flask) Dynamic (PEIE)
D-value (s) Standard Deviation D-value (s) Standard Deviation
52 353 39.08
55 148 2.18
58 34.7 2.27 29.8 3.3
60 18 1.52 13.27 1.54
62 6.93 0.47
z-value (C) 5.99 6.16
Table 11.2 Results of validation experiments comparing the number of survivors measured
experimentally with those predicted using traditional and new PEIE methods (Reproduced from
Moody 2003. With permission)
Experiment Hold Tube Survivors (cfu)
Time (sec) Temp. (C) Initial (cfu) PEIE Isothermal Experimental
Predicted Predicted
I 15 65 5.4 l08 3.5 l03 5.0 l04 5.2 l03
2.8 103
II 10 65 5.4 10 8
2.0 10 8.6 10
3 4
l.0 103
1.3 103
above the range used to estimate the parameters (65 C). Two different processes were
used, one with a hold tube residence time of 15 s, and 10 s for the other, in order to
obtain distinctly different numbers of survivors from each process. Processed sam-
ples were plated, incubated, and enumerated to determine the number of survivors
from each process. These experimental results were compared with numbers of
survivors predicted by the mathematical model using both sets of parameters.
Model-predicted results using parameters estimated by the new PEIE and
traditional isothermal bath method are compared with experimental results in
Table 11.2. The number of survivors predicted by the model, using parameters
estimated by the PEIE method, agrees very closely with those measured experi-
mentally (within a fraction of one log cycle). In contrast, the number predicted by
the model using parameters estimated by the traditional isothermal bath method
differed from those measured experimentally by one and two log cycles for
processes I and II, respectively. The significance of these differences can best
be appreciated when the two sets of parameters are used to calculate the process
time needed to accomplish a 6-log cycle reduction in the E. coli population in
orange juice at a typical pasteurization temperature of 65 C. At this temperature,
the D-value estimated by the traditional isothermal bath method would be 1.8 s
leading to a hold tube residence time of 1.8 (6) ¼ 11 s. In contrast, the D-value
estimated by the new PEIE method would be 1.3 s leading to a hold tube residence
time of 1.3 (6) ¼ 8 s. This 30% reduction in process time is made possible by
adopting better methods for estimating kinetic parameters more accurately.
11.3 Retort Equipment Systems in Cookroom Operations
The preponderance of thermally processed shelf-stable foods in the marketplace

today continues with products processed in batch retorts that must be repeatedly
loaded and unloaded between each process cycle throughout the workday. In years
past, this was performed by teams of workers with the help of chain hoists and rail
carts moving from retort to retort around the cookroom floor, while retort operators
kept vigilance as time keepers, temperature monitors and record keepers. Recent
innovations in retort control and materials handling systems have essentially
revolutionized traditional cookroom operations. Today’s modern retorts such as
that shown in Fig. 11.1 are equipped with sophisticated computerized electronic
control systems that can be remotely monitored from a control room by a single
operator who may be responsible for an entire battery of retorts. These control
systems are capable of operating each retort through its entire process cycle, while
controlling and recording temperatures and pressures, as well as monitoring process
conditions at all critical control points. At the end of each process cycle, a complete
set of batch records is provided for compliance with record keeping requirements of
the FDA Low-acid Canned Food regulations.
The introduction of automation and robotics for the materials handling opera-
tions on the cookroom floor has perhaps had the greatest impact in reducing the cost
of manufacturing thermally processed products. An automated batch retort system
in a modern cookroom today consists of a battery of retorts laid out in a row on the
cookroom floor to accommodate automated loading and unloading (Fig. 11.2). Both
track-guided and trackless systems are available for this purpose. In track-guided
Fig. 11.1 Modern Retort with pressure vessel, field devices, and process control system (Photo
courtesy of JBT FoodTech, formerly FMC FoodTech, Madera, CA)
252 A. Teixeira
Fig. 11.2 Modern cook room showing battery of retorts arranged for automated loading and
unloading (Photo courtesy of JBT FoodTech, formerly FMC FoodTech Madera, CA)
Fig. 11.3 Track-guided automated batch retort system (ABRS) (Courtesy of Allpax Products,
Inc., Covington, LA)
systems, a rail cart transfers crates or baskets of products from the loading stations
to the retorts, and from the retorts to the unloading stations automatically on a rail
track that allows the cart to move in a transverse direction along the cookroom floor
until it is aligned with the target retort (Fig. 11.3). Once the cart is aligned with the
retort, the loaded baskets or crates are automatically transferred from the cart into
the retort for loading operations, or from the retort onto the cart for unloading
operations. Trackless systems such as that illustrated in Fig. 11.4 work much the
Fig. 11.4 Trackless system layout for automated guided vehicles (AGV) (Photo courtesy of JBT
FoodTech, formerly FMC FoodTech, Madera, CA)
same way, except that the rails and rail carts are replaced by automated guided
vehicles (AGV) that move about the cookroom floor controlled by electronic
guidance systems. These systems offer the advantage of keeping the cookroom
floor free of rails or tracks that could impede safe movement of workers on the floor.
A close-up view of an AGV that has just been loaded or is about to unload a retort is
shown in Fig. 11.5, and an automated materials handling system for retort basket,
rack, or tray loading is shown in Fig. 11.6.
11.4 Flexible Retortable Packages
Perhaps the most intriguing innovation in thermal processing in recent years has
been the introduction of flexible retortable pouches and semi-rigid trays and bowls in
the marketplace for shelf-stable canned foods (Fig. 11.7). The relatively thin profiles
254 A. Teixeira
Fig. 11.5 Automated Guided Vehicle in process of retort loading/unloading (Photo courtesy of
JBT FoodTech, formerly FMC FoodTech, Madera, CA)
Fig. 11.6 Automated materials handling system for retort basket, rack, or tray loading (Courtesy
of Allpax Products, Inc., Covington, LA)
Fig. 11.7 New flexible and semi-rigid retortable packaging systems
offered by these flexible packages allows them to be heated more rapidly than metal
cans or glass jar counterparts, often resulting in better product quality. Moreover,
they offer the consumer a variety of convenient features. Flexible pouches carry less
weight, occupy less space, and are easy to open with a simple pair of scissors. Semi-
rigid trays and bowls can be fashioned as attractive serving dishes that are micro-
waveable, and can be placed on the dinner table ready to serve directly from the
microwave oven. However, the flexible and semi-rigid properties that make these
attractive features possible pose technical challenges in the retort processing of such
packages. These flexible packages lack the strength of traditional metal cans and
glass jars, and are incapable of withstanding the pressure gradients normally experi-
enced by cans and jars during typical retort process cycles.
Successful retorting of such flexible or semi-rigid packages requires that they be
processed under carefully controlled dynamic pressure excursions (profiles) to prevent
package expansion during processing. When not controlled properly, this expansion
can result in permanent distortion or deformation of the final package after retorting.
Design and control of these complex pressure profiles is accomplished with the
simultaneous use of both pressure and deflection detectors during a heat penetration
test. The electronic deflection detectors continuously monitor the package expansion
or contraction, sending a signal to the on-line retort control system, which adjusts
overriding air pressure as needed to counter the detected expansion or contraction,
thus resulting in a dynamic pressure profile ideally suited for that product.
The need for these dynamic pressure excursions during processing requires that
retort pressure be controlled independently from temperature. This means that pure
256 A. Teixeira
Fig. 11.8 Modern steam–air mixture retort with programmable pressure control (Courtesy of
Allpax Products, Inc., Covington, LA)
saturated steam can no longer be used alone as the heating medium in the retort and
that controllable overriding air pressure must be included during heating. However,
the introduction of air into the retort during heating with steam poses the risk of
insulating air pockets forming around packages, which could leave them under-
processed. The desire to meet the processing needs of these flexible packaging
systems has created a market for new retort designs and control systems that are
capable of delivering these complex process conditions. Essentially, all designs
accomplish the dynamic pressure profiles with overriding air pressure, but they
eliminate the problem of insulating air pockets in different ways. Some retorts heat
with steam–air mixtures that are kept well mixed with the use of strong fans within
the retort (Fig. 11.8). Other retort designs feature water immersion, water spray, or
water cascade as a means of contacting the packages with the heat exchange
medium during processing. In addition to independent pressure control for these
flexible packages, their lack of strength and rigidity also requires special racking
designs and systems in the materials handling of these packages to safely support
them within the retort (Fig. 11.9).
11.5 Market Implications
Prepared foods in microwavable retortable dinner trays and lunch bowls have
been gaining in popularity because of their convenience, attractiveness, and
quality relative to their traditional “canned” food counterparts in metal cans and
Fig. 11.9 Custom-designed racking systems for retortable flexible pouches (Courtesy of Allpax
Products, Inc., Covington, LA)
glass jars. These novel packages were made possible by the development of food-
grade polymer films with high oxygen-barrier properties capable of withstanding
the high temperatures and pressures during the necessary heat sterilization pro-
cess in pressurized retorts. The attractiveness of these packages and the fact that
they are microwavable make them well suited for ready-to-eat dinner menu items.
These packages have gained particular interest with those wishing to market
258 A. Teixeira
Fig. 11.10 Prototype of shelf-stable prepared meal ready-to-eat in retortable microwavable semi-
rigid tray with transparent lid stock (Reproduced from Rich 2007 with permission)
shelf-stable prepared foods that can be featured as natural, organic, ethnic,

vegetarian, or cultural such as “Ayurvedic.” Ethnic food dishes typical of Latino,
Mediterranean, Arabic, Indian, and Oriental cultures are becoming increasingly
popular worldwide.
Research at the University of Florida has been underway to assist a small start-up
company in developing a line of ethnic menu items in retortable semi-rigid trays
(Rich 2007). The company has been operating restaurants featuring authentic
cultural food dishes appropriate for strict vegetarian diets and the Ayurvedic life
styles in India and Southeast Asia. It now wishes to distribute these menu items in
retail markets as pre-packaged, shelf-stable, ready-to-eat meals so consumers can
enjoy them conveniently in their homes (Fig. 11.10). This technology will enable
manufacturers of shelf-stable food products to introduce into the marketplace
a wider variety of fully-prepared ready-to-eat convenience food items.
References
Moody V (2003) Thermal inactivation kinetics of Escherichia coli and Alicyclobacillus acidoter-
restris in orange juice. Dissertation, Agricultural and Biological Engineering Department,
presented to Graduate School of University of Florida
Rich EC (2007) Heat penetration studies on shelf-stable ethnic foods in retortable trays. Project
report toward Engineering degree, Agricultural and Biological Engineering Department,
Institute of Food and Agricultural Sciences (IFAS), University of Florida
Rodriguez AC, Teixeira AA, Smerage GH, Busta FF (1988) Kinetic effects of lethal temperatures
on population dynamics of bacterial spores. T ASAE 31(5):1594–1601, 1606
Swartzel KR (1984) A continuous flow procedure for reaction kinetic data generation. J food sci 49
(3):803–806
Welt BA, Teixeira AA, Balaban MO, Smerage GH (1997) Iterative method for kinetic parameter
estimation from dynamic thermal treatments. J Food Sci 62(1):8–14
.
Chapter 12
Optimization of Food Thermal Processing:
Sterilization Stage and Plant Production
Scheduling
Ricardo Simpson and Alik Abakarov
12.1 Introduction
A large number of real-life, decision-making problems in food sciences and engi-

neering can be formulated as a problem of optimization of various real-valued
functions (objective functions) with real-valued or discrete decision variables with
specific constraints, where the global optimum corresponds to the best solution of
the initial problem. Depending on the features of the decision variables, optimiza-
tion problems can be continuous, integer (discrete) or mixed; and depending on
the features of the objective functions and constraints, they can be linear or non-
linear with many local solutions. Thus, each optimization problem can require any
number of optimization techniques, which could guarantee finding an optimal
solution.
Two optimization problems in thermal process calculation are considered in this
chapter. First, there is the problem of thermal process calculations where process
times at specified retort temperatures are calculated to achieve safe levels of
microbial inactivation (lethality). Therefore, the accuracy of methods used for
this purpose is important to food science and engineering professionals working
R. Simpson (*)
Departamento de Ingenierı́a Quı́mica y Ambiental, Universidad Técnica Federico Santa Marı́a,
P.O. Box 110-V, Valparaı́so, Chile
and
Centro Regional de Estudios en Alimentos Saludables, Blanco 1623 Room 1402, Valparaı́so,
Chile
e-mail: ricardo.simpson@usm.cl
A. Abakarov
Departamento de Ingenierı́a Quı́mica y Ambiental, Universidad Técnica Federico Santa Marı́a,
P.O. Box 110-V, Valparaı́so, Chile
e-mail: alik.abakarov@usm.cl
262 R. Simpson and A. Abakarov
in the field (Holdsworth 1997). The second problem consists of determining an

optimized scheduling at a food cannery using several autoclaves of different
capacities to sterilize given amounts of different canned food products with specific
quality requirements.
12.1.1 Thermal Process Calculation
Optimization of thermal sterilization is an optimal control problem, the solution to

which requires searching for the best retort temperature as a function of process time.
Banga et al. (1991) showed that the optimal control problem can be transformed into
a nonlinear programming (NLP) problem, and in most cases the NLP problem
becomes a multi-modal optimization problem with several types of constraints.
These types of optimization problems make use of classical deterministic optimiza-
tion methods within a local search domain, such as Hooke-Jeeves, Nelder-Mead, and
Quasi-Newton methods (Himmelblau 1972), and are frequently limited in their
effectiveness. To ensure a global solution to these problems, it would be more
suitable to use global optimization methods. Considerable work has been reported
in the literature showing that variable retort temperature (VRT) processing can be
used to marginally improve the quality of canned food, and alternatively, to reduce
the sterilization process time, in comparison to traditional constant retort temperature
(CRT) processing (Banga et al. 1991; Banga and Alonso 2003; Teixeria et al. 1975;
Almonacid-Merino et al. 1993; Simpson et al. 2008; Abakarov et al. 2009).
The usefulness and advantages of some global optimization algorithms based
on the utilization of Gaussian probability distribution for VRT thermal process-
ing optimization were presented by Banga and Casares (1987), Banga et al. (1991,
2003, 2005), Banga and Seider (1996), Simpson et al. (2008), and Abakarov
et al. (2009).
Other global optimization algorithms, such as genetic algorithms (GA), were also
successfully implemented for VRT thermal processing by Chen and Ramaswamy
(2002). This chapter explains the implementation of a global stochastic method,
an adaptive random search based on logistic curve utilization for finding the
optimum VRT in thermal processing of conduction heated canned food. The specific
objectives were the following:
l Explore the use of a cubic spline approximation for optimum dynamic tempera-
ture profiles to simplify the problem by reducing the number of variables
(dimensional space of random search)
l Search for the optimum variable retort temperature profile to maximize retention
of a specified quality factor (thiamine) within the constraint of assuring mini-
mum required target lethality
l Search for the optimum variable retort temperature profile to minimize process
time within the constraints of assuring both minimum required target lethality
and quality retention
12 Optimization of Food Thermal Processing 263
12.1.2 Optimal Scheduling for Food Canneries
In the majority of small- to medium-sized food canneries, retorting is carried out in

a battery of retorts as a batch process (Norback and Rattunde 1991). In such
canneries, the unloading and reloading operations for each retort are also labor-
intensive. Therefore, a well-designed and -managed plant should be utilized to
optimize the whole sterilization process. In other words, it is necessary to develop a
suitable mathematical model for operation of the whole plant and to find the optimal
values of the decision variables. The result of such a model involves the quantities
of each product to be loaded into the autoclaves in each batch, and the optimal
solution that gives optimum scheduling. One interesting and practical objective
function is to find out the minimum plant operation time for specific amounts of
different products (Simpson and Abakarov 2009).
In general, “scheduling” is understood as the process of managing given re-
sources across a variety of possible tasks in order to optimize the objective
function (criterion). Many scheduling models can be mathematically described by
mixed-integer linear programming (MILP). Examples of sequential MILP short-
term scheduling models can be found in Méndez and Cerda (2000, 2002),
Harjunkoski and Grossmann (2002), and Castro and Grossmann (2006). The
following researchers have presented sequence-based MILP models for multipro-
duct batch plants: Jung et al. (1994), Moon et al. (1996), Castro and Grossmann
(2005), Floudas and Lin (2004), Gupta and Karimi (2003), Maravelias (2006),
Mendez et al. (2001), (2006), Ha et al. (2006), and Liu and Karimi (2008). An
interesting application of MILP planning for a petrochemical plant was described
by Hui and Natori (1996). Erdirik-Dogan and Grossmann (2007) presented a multi-
period mixed-integer linear programming model for the simultaneous planning and
scheduling of a single-stage, multi-product, continuous plant with parallel units.
In research by Doganis and Sarimveis (2007, 2009), the MILP model was devel-
oped to target optimal production scheduling for a single yogurt production line.
This model took into account all standard constraints encountered in production
scheduling (material balances, inventory limitations, machinery capacity, labor
shifts, and manpower restrictions). Simpson and Abakarov (2009) proposed a
particular MILP model for optimal scheduling of food canneries, addressing the
problem of sterilizing given quantities of different canned food products within a
minimum plant operation time, with specific quality requirements, in a set number
of autoclaves of the same capacity. However, the model although relevant, cannot
be implemented in most canning plants because the autoclaves are of different
capacities.
In this chapter, a mathematical model for thermal process scheduling at food
canneries with autoclaves of different capacities is described. The resulting model
is based on mixed-integer linear programming and simultaneous sterilization. Since
the initial version of the model developed here was non-linear due to its non-linear
(minimax) objective function, the necessary modifications were made to transform
it into a MILP model.
12.2 Methodology
12.2.1 Sterilization Stage Optimization
In this example, a cylindrical container with radius R and height 2L was used. The
mathematical model describing heat conduction in this particular case is a mixed
boundary problem (Teixeira et al. 1969), as follows:
2
@T @ T 1 @T @ 2 T
¼a þ þ (12.1)
@t @r 2 r @r @z2
where T is temperature, t is time, r and z are radial and vertical locations within the
container, and alpha (a) is thermal diffusivity of the product.
There were the following initial and boundary conditions (according to symmetry):
TðR; z; tÞ ¼ Trt ðtÞ
Tðr; L; tÞ ¼ Trt ðtÞ
@T
ð0; z; tÞ ¼ 0
@r
@T
ðr; 0; tÞ ¼ 0
@z
Tðr; z; 0Þ ¼ Tin
where Trt ðtÞ, t 2 ð0 : tf Þ is retort temperature as a function of time, and Tin is initial
temperature at t ¼ 0.
The first objective was to find the retort function for Problem 1: Trt ðtÞ,
Tlow Trt ðtÞ Thigh , where the final quality retention CðtÞ is maximized, while
the final process lethality Fd0 is held to a specified minimum. A second objective
was also to find the retort function for Problem 1: Trt ðtÞ, Tlow Trt ðtÞ Thigh , but
where the final process time tf is minimized subject to the same lethality require-
ment above, while the quality retention must not fall beneath some specified
minimum.
The lethality constraint can be specified as follows:
1. F0 ðtf Þ Fd0 ; where Fd0 is the final required lethality, and is calculated as a
function of time and temperature at the critical point (cold spot) according to
the following equation:
ðt
ðTTref Þ
F0 ðtÞ ¼ 10 z dt (12.2)
0
where T is temperature at the critical point or cold spot, normally at the geometric
center of the container (in the case of conduction-heated canned foods).
Quality retention, on the other hand, is greatly affected by the non-uniform
temperature distribution existing at any point in time from the heated boundary
to the cool center cold spot, and must be integrated in space over the volume of
the container, as well as over time. To accomplish this integration over both
space and time, the following approach was used:
2. Cðtf Þ Cd , where Cd is the desired volume-average final quality retention value
and calculated as follows:
2 3
ðL ðR ðt ðTTref Þ
2 ln 10
CðtÞ ¼ C0 2 exp4 10 z 5 dr dz (12.3)
LR Dref
0 0 0
12.2.1.1 Penalty Functions
To handle constraints 1 and 2, the following penalty functions were used:
X
tf

P1 ¼ A F0 d F0 ðtÞ þ jF0 d F0 ðtÞj
t¼0
X
tf
P2 ¼ A Cd CðtÞ þ jCd CðtÞj
t¼0
where A is a sufficiently large number. Use of this kind of penalty function will lead
very quickly to finding X0 X, where all given constraints are satisfied when the
random search is implemented.
12.2.1.2 Process Optimization and Computer Simulation
In general, function Trt ðtÞ over t 2 ð0 : tf Þ can be parameterized using Np points; during
each time interval t0k ¼ ðtk ; tkþ1 Þ, k 2 0 : ðNp 1Þ the value of Trt ðt0k Þ remains constant
(uk ) (Teixeria et al. 1975; Banga et al. 1991, 2003, 2005). However in this case, the use
of cubic spline in approaching global optimization problems with random search
techniques can produce superior results over discrete step-wise functions (Simpson
et al. 2008), mainly because the cubic spline approximation allows reducing signifi-
cantly the number of decision variables and therefore the random search.
Natural cubic splines are used for creating a smooth function that can fill in the
gaps between data points in a step-wise discrete function to approximate a smooth
trend. The philosophy in splining is to use low order polynomials to interpolate
from grid point to grid point. This is ideally suited when one has control of the grid
locations and the values of data being interpolated, as in the work presented here.
Therefore, approximation by the cubic spline was only utilized in this work to
obtain optimal VRT profiles.
The following penalty function was used simultaneously with the cubic spline
approximation in the search process
in order
to hold the autoclave temperature
profile Trt ðtÞ in the given range Tlow ; Thigh :
X
tf

P3 ¼ A jTlow Trt ðtÞj þ jThigh Trt ðtÞj ðTlow Trt ðtÞÞ
t¼0
where A is a sufficiently large number.

In general, the objective function used by the search procedure can be presented as:
Fðu1 ; u2 ; . . . ; uNP 1 ; tf Þ ¼ tf þ P1 þ P2 þ P3 ; tf 2 ½tleft; tright
where ui , i 2 1 : ðNp 1Þ are the control variables, and tleft and tright could be
obtained from the following expressions:
ð
tright
ðTlow Tref Þ
Fd0 ¼ 10 z dt (12.4)
0
tð
left
ðThigh Tref Þ
Fd0 ¼ 10 z dt (12.5)
0
12.2.1.3 Adaptive Random Search Method
In this method, let Ii X be a perspective interval for variable xi ; i 2 1 : n, and let

2q be the width of each perspective interval. Let I be a Cartesian product of sets,
Ii ; i 2 1 : n, and for the random search, let I be a perspective sub-domain with
center point x0i ; i 2 1 : n.
The pedestal distribution is utilized in the adaptive random search as a probability
distribution Ps ðxs1 Þ. After each calculation of objective function the pedestal
distribution of x ¼ hx1 ; x2 ; . . . ; xn i is modified so that the probability of finding the
optimal value of the objective function is increased.
The example in Fig. 12.1 (a, b, c) shows a two-dimensional pedestal frequency
distribution transformation during the whole random search process. Figure 12.1a
shows the pedestal frequency distribution used for the first algorithms iteration.
Because there is no information about the optimization problem being solved, any
point of the domain X could be chosen with the same probability.
During the search process, the random search generates random vector values of
x0 ; x1 ; . . . ; xs , calculates objective function, accumulates information about the
solved problem, and transforms the pedestal frequency distribution according to the
computations done. Transformations consist of reducing the deviation of the pedestal
Fig. 12.1 Two-dimensional pedestal frequency distribution transformations occurring during the
whole random search process
distribution (or perspective sub-domain I X) around the mean (or center point,
x0i ; i 2 1 : n), which is the current best solution: x0 2 X; f ðx0 Þ< f ðx j Þ; 8j 2 1 : s.
Figure 12.1b shows that the pedestal frequency distribution could be obtained in the
middle of the search process, and now the points of domain X (in terms of probability
density) can be divided into two non-overlapping subsets; h and H are the proba-
bility densities of the non-perspective and the perspective sub-domains, correspond-
ingly. Figure 12.1c shows the pedestal frequency distribution for the final algorithm
iteration, transformed into the d function, and it is assumed for any point x0 of the
final perspective domain I, the following condition holds: jf ½x0 f ð½x Þj e.
The adaptive random method procedure could be described as follows:
1. Set the total number of random search iteration Ns ; set the center point
x0i ; i 2 1 : n as the center point of initial perspective sub-domain I; and set
iteration counter s ¼ 1.
2. Generate a new vector x j 2 X from the current pedestal probability distribution:

Pj ðx j1 ; x0 Þ
3. Compute the value of objective function F j ¼ Fðx j Þ, and by using the formula:
n o
j j
Fmin ¼ min F j ; Fmin
Calculate a minimal value of objective function in step: j, j 2 1 : Ns .

j j1
4. If Fmin <Fmin , set the center point x0 ¼ x j .
5. If j < Ns , then go to step 2.
To modify the pedestal probability distribution (step 2), the following formulas
are used:
l Calculate the probability density h j for step j, j 2 1 : Ns as follows:
1 pj
hj ¼
1 vj
where v j is volume of the perspective domain of random search, which is in step

j (v0 ¼ 1), and p j is the assumed in step j probability, where the optimal point
belongs to the current perspective domain I.
l To calculate current volume v j and assumed probability p j , the basic random
search method uses the following expressions:
1
q0 ¼
2
1
q jþ1 ¼ q j ð2eÞNs
n
v j ¼ 2q j ; (12.6)
where e is given accuracy;
8
>
> s j ðpmin 1Þ
< þ 1; if 0 v j vmin;
pj ¼ smin
>
> v ð1 pmin Þ pmin vmin
j
: þ ; if vmin v j 1
1 vmin 1 vmin
where pmin and qmin (vmin ¼ ðqmin Þn ) are two heuristic parameters of adaptive
random search, which allow tuning the algorithm to different types of problems
as well as humans can.
In this chapter, a modification of the adaptive random search was used for
thermal process calculation. The proposed modification utilizes the logistic curve
as a law to reduce the q value during the whole search process, in other words, as a
law to reduce the volume of perspective sub-domain I from v0 ¼ 1 to vNs ¼ ð2eÞn :
The logistic equation was first published by Verhulst (1845) as a model of

population growth. The continuous version of a logistic model is as follows:
dx
¼ mxð1 xÞ (12.7)
dt
where m is the Malthusian parameter (rate of maximum population growth). The
solution of (12.7) could be written as:
1
xðtÞ ¼ ; (12.8)
1 þ a em t
where a and m are the parameters of the logistic curve. The function of xðtÞ is also
called the sigmoid function (Krose 1993).
Figure 12.2 displays two curves; the interrupted curve (dashed line) shows a
relationship between volume v j and iteration number j in a basic version of the random
search, and is obtained with (12.6); the solid curve is the possible logistic curve,
utilized as the same relationship in the proposed modification and obtained in (12.8).
A wide set of experiments and practice implementations of the basic random search
algorithm have been conducted, to show the effectiveness of the adaptive algorithm
and to make recommendations for choosing heuristic parameter values (Abakarov and
Sushkov 2002, 2005; Simpson et al. 2008). However, we propose the following:
l Reduce the local nature of the basic random search algorithm, because after 25%
iterations, volume of the perspective sub-domain was reduced by more than 50%,
which is not well-grounded in the case of multimodal optimization (Fig. 12.2).
Fig. 12.2 Curves showing the relationship between volume vj and iteration number j for the basic
version of adaptive random search, obtained by (12.6), and for the modified version, (12.8)
l The nature of the logistic curve as a law of perspective sub-domain is reduced, as

it is soundly related to behavior of the random search seemingly needed in
general, namely, in the case of any a priori information about optimization
problem, which is not available.
The whole random search process based on logistic curve utilization can be
divided into three stages (Fig. 12.2) and interpreted as follows:
Stage I. There is no apriori information about the solved problem, so the slow
reduction of volume of the perspective sub-domain is preferable at the beginning
of the search process, which allows accumulating primary information about the
location of the global solution inside domain X.
Stage II. Based on primary information accumulated in Stage I, we can assume the
global solution belongs to the perspective sub-domain and accelerates the volume
reduction process of the perspective sub-domain around its center point.
Stage III. To ensure the global solution is computed with accuracy e, the process of
reducing the volume of the perspective sub-domain should be slowed-down.
12.2.2 Canning Plant Scheduling
The simultaneous sterilization approach was developed especially for small can-
neries with few retorts, as they frequently process small batches of different products
in various container sizes that require different process times and retort temperatures
(Simpson 2005). The proposed approach takes advantage of the fact that, for any
given product and container size, there exist many alternative combinations of retort
temperatures (above the lethal range) and corresponding processing times that
deliver necessary lethality (F0 value). In practice, the following two F0 values are
considered for each product sterilized: Fmin 0 and Fmax
0 . These values are product-
related, but in general the value Fmin
0 is chosen according to safety criterion and Fmax
0
according to quality criterion (resulting in a safe product with required quality). All
combinations of retort temperature and processing time correspond to the same F0
value and are called isolethal or equivalent lethality processes (Holdsworth and
Simpson 2007). Figure 12.3 shows two equivalent lethality curves corresponding to
0
the values Fmin0 and Fmax
0 . The region R between the two curves contains all
combinations of retort temperatures and processing times sufficient for sterilization
of a selected product. We will call this region the “permissible region.”
One important aspect of the simultaneous sterilization approach is that the
difference in the absolute level of quality retention is relatively small over a
practical range of isolethal process conditions. The relative insensitivity of quality
over a range of different isolethal process conditions opens the door to maximizing
output from a fixed number of retorts for different product and container sizes. The
following procedure (I) can be utilized to obtain equivalent lethality processes (in
permissible region) for a given product and its container size:
Tmin Tmax
160
F0min
140
F0max
120
100
Time (min)
80
60
40
20
0
106 108 110 112 114 116 118 120 122 124 126
Temperature (°C)
0
Fig. 12.3 Two equivalent lethality curves corresponding to values Fmin
0 and Fmax
0 . The region R
between the curves contains all combinations of retort temperatures and processing times suffi-
cient for sterilization of a selected product
Procedure (I):
1. For a given product and its can size, choose Fmin 0 and Fmax
0 values.
2. Choose an interval ½Tmin ; Tmax , where Tmin and Tmax are the minimum and
maximum retort temperatures, respectively.
3. Choose the discretization points Ti ; i 2 1 : N, in ½Tmin ; Tmax .
4. For each discretization point, Ti ; i 2 1 : N, and lethality values, Fmin
0 and Fmax
0 ,
obtain isolethal processes. (Heat penetration tests can be conducted on each
product or computer simulations to establish processing times ti ; i 2 1 : N, at
retort temperatures Ti ; i 2 1 : N, to achieve the necessary target lethality values).
5. To obtain a set of two continuous curves per product, fit the discrete values
ti ; i 2 1 : N, using the cubic spline procedure. The region restricted by the two
continuous curves, and by the left and right bounds of retort temperatures Tmin
and Tmax , is the permissible region for the given product (Fig. 12.3). Thus, the
two strictly decreasing continuous functions, which are iso-lethality curves, are
obtained for the product (Fig. 12.3) as follows:
m; M : ½Tmin ; Tmax ! ð0; þ1Þ
where mðTÞ MðTÞ for each T 2 ½Tmin ; Tmax and the minimum and maximum
times needed to process the product at temperature T are denoted by mðTÞ and
MðTÞ, respectively.
Figure 12.4 shows the results of utilizing procedure (I) for three different pro-
ducts: A, B and C. (The permissible regions for each product, RA , RB and RC , were
obtained). The intersection RABC of the permissible regions RA , RB and RC indicates
processes that are sufficient for simultaneous sterilization of products A, B, and C.
Fig. 12.4 The intersection RABC of the permissible regions RA , RB and RC indicates processes that
are sufficient for simultaneous sterilization of products A, B and C
Any simultaneous sterilization possibility can be presented as the following:
hv; t; T i
where v ¼ ðv1 ; v2 ; . . . ; vn Þ is a simultaneous sterilization vector, vk 2 f 0; 1g,

k 2 1 : n, and

1; if product k can be sterilized

vk ¼
0; otherwise
and t; T are the necessary time and retort temperature, respectively, used to process
a subset of products determined by vector v.
It can be assumed the simultaneous sterilization vector v1 is dominated by the
other simultaneous sterilization vector v2 if the following two conditions hold:

8j 2 1 : n v1 ½ j ¼ 1 ) v2 ½ j ¼ 1
t1 bt2
where t1 and t2 are the times necessary to process the subsets of the products
determined by the vectors v1 and v2 , respectively.
This fact can be denoted as v1 > v2 . Obviously, only non-dominated vectors
should be used in the sterilization process; otherwise, a sterilization vector exists
that could be replaced by another vector, which would mean it is possible to
sterilize at least the same subset of products in reduced process time.
12.2.2.1 Problem Definition
The optimization problem in this research consists of finding a given quantity of

each product, aj , j 2 1 : n, a subset of simultaneous sterilization vectors, V opt VA ,
and its assignment among a given number of autoclaves of different capacities, ck ,
k 2 1 : s, such that each product is sterilized completely within the minimum plant
operation time.
To solve the optimization problem in the present research, the following data
was generated:
l Number of sterilization products: n
l Quantity for each product: aj ; j 2 1 : n
l Number of sterilization vectors: m
l Set of non-dominated sterilization vectors: VA ¼ fvi g, vij 2 f0; 1g, i 2 1 : m,
j21:n
l Set of sterilization times: TA ¼ fti g; i 2 1 : m
l Number of autoclaves: s
l Capacity of autoclave: ck , k 2 1 : s
12.2.2.2 Mathematical Model Description
Two types of decision variables were used in the mathematical model:

l Integer decision variables:

1; if vector vi 2 V is used for sterilization process in the autoclaves;

uki ¼
0; otherwise
l Continuous decision variables: xkij , i 2 1 : m; j 2 1 : n; k 2 1 : s, corresponding

to the quantity of product Pj loaded into autoclave k utilizing sterilization vector i
The objective function to minimize plant operation time can be written as
follows:
( )
X
m
max uki ti ! min (12.9)
k u
i¼1

m
P
where the goal is to minimize the maximum time max uki ti spent by one
k i¼1
of the autoclaves kin processing its part of the product and thus minimizing plant
operation time.
Given all products must be completely sterilized, the following constraints
should be considered:
X
s X
m
xkij ¼ aj ; 8j 2 1 : n (12.10)
k¼1 i¼1
For all chosen simultaneous sterilization vectors vi , i 2 1 : M, and all given

autoclave capacities, ck , k 2 1 : s, the amount of product in each batch loaded
into autoclave k should be less than its capacity ck . These constraints can be
written as:
X
N
xkij uki ck ; 8i 2 1 : m; 8k 2 1 : s (12.11)
j¼1
Since the above mathematical model is non-linear, due to its non-linear (minimax)
objective function, the following modifications were made in order to transform
it into an equivalent MILP model. The objective function (2) was replaced by the
following objective function:
MinMax ! min
where MinMax 2 ð0; þ1Þ, and the following constraints for each of the given
autoclaves k 2 1 : s, were added to constraints in (12.10) and (12.11):
X
m
uki ti MinMax; k21:s
i¼1
Thus, the minimization of plant operation time by MILP can be written in terms
of an objective function:
MinMax ! min
subject to:
X
m
uki ti MinMax; k21:s
i¼1
s X
X m
xkij ¼ aj ; 8j 2 1 : n
k¼1 i¼1
X
N
xkij uki ck ; 8i 2 1 : m; 8k 2 1 : s
j¼1
The obtained MILP model is the preferred choice over the model with a non-
linear objective function, for two reasons. First, there are algorithms, which gua-
rantee finding a global solution to the linear programming problems (Taha 2006).
Second, efficient computer tools have been developed based on such algorithms
(Robert Fourer, “Software Survey: Linear Programming” 2009).
12.3 Results and Discussion
12.3.1 Thermal Process Calculation
The following data in Table 12.1 was used in the optimization of the thermal
processing of canned foods.
12.3.1.1 Maximizing Quality Retention Problem
The next numerical experiment consists of searching for the optimum variable
retort temperature profile, to maximize retention of a specified quality factor
(thiamine) within the constraint of assuring minimum required target lethality and
by variable retort temperature profile utilization, within the upper and lower limits
of retort temperature in the range of this study (110–140
C). In this example, a final
target lethality of Fd0 ¼ 8 min was chosen (typical for many canned foods).
The results obtained from a basic random search algorithm, and its modifica-
tion, are shown in Fig. 12.5 (Simpson et al. 2008) and Fig. 12.6, respectively.
Utilizing the proposed modification, the optimal solution computed before was
improved in terms of process time (89 against 91 min) and number of objective
function computations (600 against 1,000 calculations). The final lethality F0 was
equal to 8.001.
12.3.1.2 Minimization Process Time Problem
This numerical experiment deals with searching for the optimum variable retort
temperature profile to minimize process time within the constraints of assuring both
minimum required target lethality and quality retention. In this case, the search
routine was restricted in two ways, first to satisfy the lethality constraint of
Fd0 ¼ 8min, as before, and second, that the quality constraint thiamine retention
could not fall below 50% (CðtÞ > 0:5).
Table 12.1 Parameters Can radius (m) 0.04375

utilized in thermal process Can height (m) 0.1160
simulation study (From Thermal diffusivity 1.5443 107
Garcı́a et al. 2005) a (m2s1)
T0 (
C) 71.11
Microorganism Bacillus stearothermophilus
zM; ref (
C) 10
TM; ref (
C) 121.11
Nutrient Thiamine
zN; ref (
C) 25.56
Dref ðsÞ 10716.0
TN; ref (
C) 121.11
Fig. 12.5 Optimum VRT profile for maximum thiamine retention (55%) from a basic random
search algorithm using cubic spline approximation and 1,000 iterations
140
120
Retort Temperature (°C)
100
80
60
40
20
0
0 10 20 30 40 50 60 70 80 90 100 110 120
Process Time (min)
Fig. 12.6 Optimum VRT profile for maximum thiamine retention (55%) from a modified random
search algorithm using cubic spline approximation and 600 iterations
The results obtained from a basic random search algorithm, and its modification,
are shown in Fig. 12.7 (Simpson et al. 2008) and Fig. 12.8, respectively. Utilizing
the proposed modification, the optimal solution computed before was improved in
terms of Process Time (67 against 68 min.) and the number of objective function
computations (600 against 1,000 calculations). The final thiamine retention Cðtf Þ
was equal to 0.503 (50%).
140
120
100
80
60
40
20
0
0 10 20 30 40 50 60 70 80 90 100
Process Time (min)
Fig. 12.7 Optimum VRT profile for minimum thiamine retention (50%) from a basic random
search algorithm using cubic spline approximation and 1,000 iterations
140
120
100
80
60
40
20
0
0 10 20 30 40 50 60 70 80 90 100
Process Time (min)
Fig. 12.8 Optimum VRT profile for minimum thiamine retention (50%) from a basic random
search algorithm using cubic spline approximation and 600 iterations
12.3.1.3 Canning Plant Optimization
To demonstrate the ability of the proposed approach, several plant production

problems were solved. These problems differed according to the capacities of the
autoclaves used in the plant. As shown in Table 12.2, 16 combinations of products
and can sizes were selected for this study.
A computer program developed by the authors was used to generate a set VA0 of
79 non-dominated simultaneous sterilization vectors and a set of times, TA0 , for
temperatures ranging from Tmin ¼ 110
C to Tmax ¼ 124
C. Values 0.1 and 0.5 were
used as precision parameters of procedure (II) Dt and DT , respectively.
Table 12.3 shows all simultaneous sterilization vectors computed at temperature
112
C utilizing procedure (II). The following common data was used for problems 1–4:
l Number of combinations of products and can sizes: 16
l Quantity for each product: see Table 12.4
l Number of sterilization vectors: 79
l Set of non-dominated sterilization vectors: VA0
l Set of sterilization time: TA0
Table 12.2 Combinations of products and can sizes used in this study
Can size
Product 211 400 300 407 307 409 307 113 401 411
Number of combinationsa
Asparagus 1 2 3 4 5
Corn 6 7 8 9
Green beans 10 11 12 13
Peas 14 15 16
a
Number of product and can size combinations
Table 12.3 Sterilization vectors computed at temperature 112

C
Vector Products Time (min)
1a 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 60.06
2 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 97.05
3 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 104.84
4 0 0 1 0 0 1 0 0 0 1 0 0 0 0 1 0 61.99
5 0 0 1 0 0 1 0 1 0 1 0 0 0 0 1 0 62.14
6 1 0 1 0 0 1 0 1 0 1 0 0 0 0 1 0 63.35
7 1 0 1 1 0 1 0 1 0 1 0 0 0 0 1 0 63.54
8 1 0 1 1 0 1 0 1 0 1 0 0 0 1 1 0 65.06
9 1 0 1 1 0 1 0 1 0 1 0 1 0 1 1 0 67.23
10 1 1 1 1 0 1 0 1 0 1 0 1 0 1 1 0 70.81
11 0 1 0 0 1 0 1 0 1 0 0 0 0 0 0 1 86.07
12 0 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 79.78
13 1 1 1 1 0 1 0 1 0 1 0 1 0 1 1 1 71.25
14 1 1 1 1 0 1 0 1 1 1 0 1 0 1 1 1 73.25
a
Row refers to the 16 product/can size combinations used in study
Table 12.4 Quantities for the 16 product/can size combinations

Combinations of products and can sizes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Quantitya (volume in 7 13 4 16 6 17 18 5 8 11 2 14 10 12 19 9
thousands of liters)
a
Product quantities were generated randomly in accordance with uniform probability distribution
from interval ½1; 20
Specific data for Problem 1:

Number of autoclaves: 1
Capacity of autoclave c1 ¼ 20; 000 L
(Results for Problem 1 are presented in Table 12.5).
The MILP model proposed in this research always gave plant operation times
lower than or equal to those of the non-simultaneous sterilization operation, but the
difference between the schedules depended on given data, including the
Table 12.5 Results obtained for Problem 1

Optimal plant operation time
Non-simultaneous sterilization case: 332.5 min
Simultaneous sterilization case: 251.42 min
Optimal solution: simultaneous sterilization case
Autoclave Simultaneous sterilization vectors utilized
c1 ¼ 20000L v29 ; v44 ; v50 ; v51 ; v54 ; v55 ; v59 ; v68 ; v70 ; v78
Number of vector Quantities for each can size and product combination Time (min)
loaded into autoclave (volume in thousands of liters)
1a 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
29 0 0 0 0 0 0 0 0 0 0 2 0 10 0 0 0 69.18
v
v44 0 0 4 0 0 0 0 5 0 11 0 0 0 0 0 0 18.13
v50 0 1 0 0 0 0 0 0 0 0 0 10 0 0 0 9 23.76
v51 0 0 0 0 0 0 0 0 0 0 0 4 0 12 0 0 20.14
v54 7 0 0 13 0 0 0 0 0 0 0 0 0 0 0 0 16.44
v55 0 12 0 0 0 0 0 0 8 0 0 0 0 0 0 0 21.63
v59 0 0 0 3 0 17 0 0 0 0 0 0 0 0 0 0 16.14
v68 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 30.0
v70 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 22.0
v78 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 0 14.0
a

c1 ¼ 20000L v56 ; v58 ; v59 ; v61 ; v68 ; v69 ; v70
c2 ¼ 15000L v29 ; v55 ; v62 ; v75
1a 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
v29 0 0 0 0 0 0 0 0 0 0 2 0 10 0 0 0 69.18
v55 0 7 0 0 0 0 0 0 8 0 0 0 0 0 0 0 21.63
v56 0 0 4 0 0 0 0 0 0 11 0 0 0 0 0 0 13.62
v58 7 0 0 1 0 0 0 0 0 0 0 0 0 12 0 0 17.14
v59 0 0 0 15 0 0 0 0 0 0 0 0 0 0 5 0 16.14
v61 0 0 0 0 0 1 0 5 0 0 0 0 0 0 14 0 15.6
v62 0 6 0 0 0 0 0 0 0 0 0 0 0 0 0 9 21.02
v68 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 30.0
v69 0 0 0 0 0 16 0 0 0 0 0 0 0 0 0 0 15.0
v70 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 22.0
v75 0 0 0 0 0 0 0 0 0 0 0 14 0 0 0 0 18.0
a

c1 ¼ 20000L v55 ; v56 ; v58 ; v59 ; v61 ; v70 ; v75 ; v78
c2 ¼ 10000L v62 ; v68 ; v74 ; v76
1a 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
55 0 12 0 0 0 0 0 0 8 0 0 0 0 0 0 0 21.63
v
v56 0 0 4 0 0 0 0 0 0 11 0 0 0 0 0 0 13.62
v58 7 0 0 0 0 0 0 0 0 0 0 0 0 12 0 0 17.14
v59 0 0 0 16 0 2 0 0 0 0 0 0 0 0 0 0 16.14
v61 0 0 0 0 0 15 0 5 0 0 0 0 0 0 0 0 15.6
v62 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 9 21.02
v68 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 30.0
v70 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 22.0
v74 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 48.0
v75 0 0 0 0 0 0 0 0 0 0 0 14 0 0 0 0 18.0
v76 0 0 0 0 0 0 0 0 0 0 0 0 10 0 0 0 39.0
v78 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 0 14.0
a

Autoclave Simultaneous sterilization vectors
c1 ¼ 20000L v54 ; v55 ; v58 ; v61 ; v70
c2 ¼ 15000L v29 ; v51
c3 ¼ 10000L v56 ; v57 ; v62 ; v68 ; v78
Number of vector Quantities for each can size and product combination loaded into Time (min)
autoclave (volume in thousands of liters)
1a 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
29 0 0 0 0 0 0 0 0 0 0 2 0 10 0 0 0 69.18
v
v51 0 0 0 0 0 0 0 0 0 0 0 14 0 1 0 0 20.14
v54 7 0 0 7 0 6 0 0 0 0 0 0 0 0 0 0 16.44
v55 0 12 0 0 0 0 0 0 8 0 0 0 0 0 0 0 21.63
v56 0 0 4 0 0 0 0 0 0 6 0 0 0 0 0 0 13.62
v57 0 0 0 0 0 0 0 5 0 5 0 0 0 0 0 0 14.23
v58 0 0 0 9 0 0 0 0 0 0 0 0 0 11 0 0 17.14
v61 0 0 0 0 0 11 0 0 0 0 0 0 0 0 9 0 15.6
v62 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 9 21.02
v68 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0 30.0
v70 0 0 0 0 0 0 18 0 0 0 0 0 0 0 0 0 22.0
v78 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10 0 14.0
a
Row refers to the 16 product/can size combinations used in the study
simultaneous sterilization vectors, quantity of products considered, and capacity of

autoclaves. According to the test problems (1–4), reduction in plant operation time
between simultaneous and non-simultaneous operations was in the range 20–25%.
In addition, this reduction was linked to the number of autoclaves; a larger number
resulted in less reduction in plant operation time.
All numerical results for problems 1–4 were obtained with online mixed-integer
linear programming solvers of COIN-OR (Computational Infrastructure for Opera-
tions Research, http://www.coin-or.org) using AMPL (a modeling language for
mathematical programming, http://www.ampl.com) input.
12.4 Conclusions
Two optimization problems in thermal process calculation were solved in this

chapter. The results drawn from this work are as follows:
1. Global optimization using random search techniques can be used effectively to
search for optimum variable retort temperature processes that will either:
l Maximize quality retention subject to achieving minimum target lethality, or
l Minimize process time to reach target lethality subject to holding quality
retention to a specified minimum.
2. Use of cubic spline in approaching global optimization problems via random

search techniques can produce superior results over discrete step-wise functions
in cases where optimal VRT profiles are expected to be a smooth curve; so in
other words the optimal VRT profile can be better approximated by implemen-
tation of cubic spline as opposed to increasing the number of discretization
points in the domain ½t0 ; tf .
Also, this research proposed a mathematical model for optimized scheduling at
food canning plants where autoclaves of different capacities are used to sterilize
given amounts of various canned food products with specific quality require-
ments. This approach is of special relevance to small- and medium-sized can-
neries, which normally work with many products at the same time. Depending on
the situation, the proposed methodology for developing the MILP model may be
used to address other types of optimization problems arising in food processing
plants.
Acknowledgements The authors Ricardo Simpson and Alik Abakarov are grateful for the
financial support provided by CONICYT through the FONDECYT project number 1090628.
References
Abakarov A, Sushkov Yu (2002) The statistical research of random search. Mathematical models.
Theory and application. Saint-Petersburg State University, Saint-Petersburg, pp 70–101
Abakarov A, Sushkov Y (2005) The algorithm of adaptive random search for discrete-continuous
optimization. In: Proceedings of the 5th Saint Petersburg Workshop on Simulation, St. Petersburg
State University, St. Petersburg, June 26–July 2, 2005, pp 11–17
Abakarov A, Sushkov Yu, Almonacid S, Simpson R (2009) Thermal processing optimization
through a modified adaptive random search. J Food Eng
Almonacid-Merino SF, Simpson R, Torres JA (1993) Time variable retort temperature profiles for
cylindrical cans: batch process time, energy consumption, and quality retention model. J Food
Process Eng 16(4):171–187
Banga JR, Casares JJ (1987) ICRS: application to a wastewater treatment plant model. In: The
Institution of Chemical Engineers (ed) Process optimisation (IChemE Symposium Series No.
100). Pergamon, Oxford, pp 183–192
Banga JR, Seider WD (1996) Global optimization of chemical processes using stochastic algo-
rithms. In: Floudas CA, Pardalos PM (eds) State of the art in global optimization: computa-
tional methods and applications. Kluwer, Dordrecht, The Netherlands, pp 563–583
Banga JR, Perez-Martin RI, Gallardo JM, Casares JJ (1991) Optimization of thermal processing of
conduction-heated canned foods: study of several objective functions. J Food Eng 14:25–51
Banga JR, Balsa-Canto E, Moles CG, Alonso AA (2003) Improving food processing using modern
optimization methods. Trends Food Sci Technol 14(4):131–144
Banga JR, Balsa-Canto E, Moles CG, Alonso AA (2005) Dynamic optimization of bioprocesses:
efficient and robust numerical strategies. J Biotechnol 117(4):407–419
Castro PM, Grossmann IE (2005) New continuous-time MILP model for the short-term scheduling
of multi-stage batch plants. Ind Eng Chem Res 44(24):9175–9190
Castro PM, Grossmann IE (2006) An efficient MILP model for the short-term scheduling of single
stage batch plants. Comput Chem Eng 30:1003–1018
Chen CR, Ramaswamy HS (2002) Modeling and optimization of variable retort temperature
(VRT) thermal processing using coupled neural networks and genetic algorithms. J Food
Eng 53(3):209–220
Doganis P, Sarimveis H (2007) Optimal scheduling in a yogurt production line based on mixed
integer linear programming. J Food Eng 80:445–453
Doganis P, Sarimveis H (2009) Mixed integer linear programming scheduling in the food industry.
In: Ferruh E (ed) Optimization in food engineering. CRC Press, Boca Raton, FL, p 800
Erdirik-Dogan M, Grossmann I (2007) Simultaneous planning and scheduling of single-stage
multi-product continuous plants with parallel lines. Comput Chem Eng 32(11):2626–2642
Floudas CA, Lin X (2004) Continuous-time versus discrete-time approaches for scheduling of
chemical processes: a review. Comput Chem Eng 28(11):2109–2129
Fourer R (2009) Software Survey: Linear Programming. OR/MS Today 36(3):46–55
Garcı́a MSG, Balsa-Canto E, Alonso AA, Banga JR (2005) Computing optimal operating policies
for the food industry. J Food Eng 74(1):13–23
Gupta S, Karimi IA (2003) Animproved MILP formulation for scheduling multi-product, multi-
stage batch plants. Ind Eng Chem Res 42(11):2365–2380
Ha JH, Chang HK, Lee ES, Lee IB, Lee BS, Yi G (2006) Inter-stage storage tank operation
strategies in the production scheduling of multi-product batch processes. Comput Chem Eng
24:1633–1640
Harjunkoski I, Grossmann IE (2002) Decomposition techniques for multistage scheduling pro-
blems using mixed-integer and constraint programming methods. Comput Chem Eng 26:1533
Himmelblau DM (1972) Applied nonlinear programming. McGraw-Hill, New York
Holdsworth SD (1997) Thermal processing of packaged foods. Blackie Academic & Professional,
London
Holdsworth SD, Simpson R (2007) Thermal processing of packaged foods, 2nd edn. Springer,
New York, p 412
Hui C-W, Natori Y (1996) An industrial application using mixed-integer programming technique:
a multi-period utility system model. Comput Chem Eng S20:S1577–S1582
Jung JH, Lee HK, Lee IB (1994) Completion times and optimal scheduling for serial multi-
product processes with transfer and set-up times in zero-wait policy. Comput Chem Eng
18:537–544
Krose Ben JA (1993) Fundamentals. In: Patrick van der Smagt (ed). An introduction to networks
neural, 5th edn. The University of Amsterdam, Amsterdam, The Netherlands
Liu Y, Karimi I (2008) Scheduling multistage batch plants with parallel units and no interstage
storage. Comput Chem Eng 32:671–693
Maravelias CT (2006) A decomposition framework for the scheduling of single- and multi-stage
processes. Comput Chem Eng 30:407–420
Méndez C, Cerda J (2000) Optimal scheduling of a resource-constrained multiproduct batch plant
supplying intermediates to nearby end-product facilities. Comput Chem Eng 24:369
Méndez C, Cerda J (2002) An efficient MILP continuous-time formulation for short-term sche-
duling of multiproduct continuous facilities. Comput Chem Eng 26:687
Mendez CA, Henning GP, Cerda J (2001) An MILP continuous time approach to short-term
scheduling of resource-constrained multi-stage flowshop batch facilities. Comput Chem Eng
25:701–711
Mendez CA, Cerda J, Grossmann IE, Harjunkoski I, Fahl M (2006) State-of-the-art review of
optimization methods for short-term scheduling of batch processes. Comput Chem Eng 30
(6–7):913–946
Moon S, Park S, Lee WK (1996) New MILP models for scheduling of multi-product batch plants
under Zero-Wait policy. Ind Eng Chem Res 35:3458–3469
Norback J, Rattunde M (1991) Production planning when batching is part of the manufacturing
sequence. J Food Proc Eng 14:107–123
Simpson R (2005) Generation of isolethal processes and implementation of simultaneous sterili-
zation utilizing the revisited general method. J Food Eng 67(1–2):71–79
Simpson R, Abakarov A (2009) Optimal scheduling of canned food plants including simultaneous
sterilization. J Food Eng 90:53–59
Simpson R, Abakarov A, Teixeira A (2008) Variable retort temperature optimization using
adaptive random search techniques. J Food Control 19(11):1023–1032
Taha H (2006) Operations research: an introduction, 8th edn. Prentice Hall, New Jersey (USA)
Teixeira AA, Dixon JR, Zahradnik JW, Zinsmeister GE (1969) Computer optimization of nutrient
retention in thermal processing of conduction heated foods. Food Technol 23(6):137–142
Teixeria AA, Zinsmeister GE, Zahradnik JW (1975) Computer simulation of variable retort
control and container geometry as a possible means of improving thiamine retention in
thermally-processed foods. J Food Sci 40(3):656–659
Verhulst P-F (1845) Recherches mathématiques sur la loi d’accroissement de la population. Nouv
Mém de l’Academie Royale des Sci Et Belles-Lettres de Bruxelles 18:1–41
Chapter 13
Recent Advances in Emerging Nonthermal
Technologies
Daniela Bermúdez-Aguirre and Gustavo V. Barbosa-Cánovas
13.1 Introduction
Thermal treatment and reduction of water activity in food have been two of the
most used techniques to process and preserve food. Pasteurization and sterilization,
the main thermal processes used around the world to inactivate pathogenic bacteria,
reduce spoilage microorganisms and enzyme activity, inactivate spores (using
sterilization), and extend the shelf life of the product. Drying is one of the most
effective techniques used to reduce the water activity in the product and as a result
the growth of microorganisms is retarded. However, the use of high temperature in
these processes has an undesirable effect on the quality of the final product. While
heat is responsible for achieving microbial and enzyme inactivation in the product,
heat also affects the sensorial quality (color, taste, texture, flavor) and nutrient
content of the product and promotes undesirable chemical reactions. In addition to
using heat, for some products chemical compounds can be added to preserve the
product and provide better stability during storage, such as artificial colorants and
preservatives. Most of the time, the final product does not represent the original
characteristics of the fresh product because of the changes in its properties during
processing. Thus, the search for new alternatives in food processing and preserva-
tion has become a priority of food scientists, not only to provide a better quality
product, but also to satisfy the needs and preferences of the consumer, and further to
fulfill regulations regarding food safety and to offer new alternatives in the food
market. These new alternatives include the use of other preservation factors, as
opposed to heat, that are able to inactivate microorganisms and enzymes and
provide better stability to the product with only minor changes in the overall quality
of the food.
D. Bermúdez-Aguirre (*) and G.V. Barbosa-Cánovas

Center for Nonthermal Processing of Food, Washington State University, Pullman, WA, USA
e-mail: daniela@wsu.edu; barbosa@wsu.edu
286 D. Bermúdez-Aguirre and G.V. Barbosa-Cánovas
13.1.1 Consumer Trends
Consumers are becoming more demanding about what to expect after paying for a
food product. In the past, consumers would buy a food product simply to satisfy a
primary need, i.e., to satisfy their hunger. Today, consumers expect additional
benefits from the same products, and some of these benefits include (among others)
fortification with extra nutrients, enhancement of characteristics like flavor and
texture, food free of chemicals or preservatives, and appealing product appearance;
in other words, consumers want to eat fresh-like products but also want extended
shelf life. The trend toward “clean” labels (with or without chemicals) is one of the
most cited around the world (Clark 2002); from 2008 to 2009 alone there was an
increase in the percentage of consumers who considered the absence of chemical
additives and preservatives in products to be a very important characteristic,
products free of chemical additives (from 37% in 2008 to 44% in 2009) and those
free of preservatives (from 28% in 2008 to 34% in 2009), according to the latest
survey (Sloan 2010). However, consumers have also developed preferences regard-
ing the technology used to process foods, as well as the origin and possible
modification of products (e.g., genetically modified foods); sometimes the con-
sumer does not easily accept novel technologies for processing food (Nielsen et al.
2009). For example, irradiated food is not well accepted in specific regions of the
world, where some consumers believe that food becomes radioactive with this
processing technology; in some countries the term “irradiation” has been changed
to “electric pasteurization.”
To gain acceptance of a new technology, food scientists must provide consu-
mers with enough information about its benefits. Sometimes the educational
background of consumers is important in achieving a full understanding of the
technology; other factors to consider are age, gender and geographical region
(Nielsen et al. 2009).
Generally, when higher quality is offered to consumers they are willing to pay a
higher price if the product is attractive to them (Lelieveld 2005). One interesting
study on the consumer’s acceptance level in the use of two novel nonthermal
technologies, high hydrostatic pressure and pulsed electric fields (PEF), provided
some facts on what consumers think is positive or negative about these technolo-
gies. Consumers preferred both nonthermal technologies because of the products’
high quality attributes (taste, nutritional content); on a negative note, they did not
like the higher price, the lack of information about the technology, and the extended
shelf life of products (Nielsen et al. 2009). This last point is really interesting
because these particular consumers are looking for food products with fresh-like
characteristics but also believe that if a product has a longer shelf life than
conventionally treated ones, the product will have lower quality. That fact is
opposite to the goal set for specific foods such as military and space rations in
which the basic requirement is a food with at least a 3-year shelf life. Maybe the
consumer’s lack of information is what leads to rejection of the novel technologies,
since high pressure and pulsed electric fields can produce safe products with long
13 Recent Advances in Emerging Nonthermal Technologies 287
shelf life while maintaining fresh-like characteristics during storage. On the hand,
participants in the study were positive towards these new technologies to some
degree because both are environmentally friendly and no preservatives are added to
the food during processing (Nielsen et al. 2009).
13.2 Emerging Technologies
It has been shown that there has been an important increase in the research of novel
technologies, due to a joint effort by academia, industry and government (Clark
2008). The area commonly referred to as emerging technologies is very wide-
ranging with two main trends in research, that of thermal and nonthermal novel
technologies.
Although novel thermal technologies such as microwave, ohmic heating, radio
frequency and inductive heating use heat as an inactivation tool, the application
and generation of heat is different from conventional thermal treatment. Proces-
sing times in these thermal technologies are very short, and lower temperatures
are used, which allows having a safe product but with minimal change in quality.
Another important development in the thermal processing area is the use of
pouches instead of cans. Layers of polyethylene or polypropylene, adhesive,
aluminum foil, polyester and nylon are constructed together to contain and
sterilize the food product (Clark 2002). These pouches are convenient because
of better heat penetration in food, lighter weight, feasibility to label them directly
on the surface and higher production rate. In the last several years, use of
polymeric trays for thermal processing (conventional or novel processes) has
become more common.
13.2.1 Nonthermal Technologies
Several nonthermal technologies have been explored recently, from high hydro-
static pressure to cold plasma. Some physical and chemical preservation factors
have been combined in an intelligent way to successfully inactivate bacteria and
extend the shelf life of the product, as shown in Table 13.1. During research of these
novel technologies, knowledge about changes in the product, such as protein
modification, delivered new ingredients for novel food products. The novel non-
thermal technologies offer not only opportunities to explore other inactivation
effects on microorganisms, but also to develop new processes and products through
specific modification of certain food components, as shown in Table 13.2. In some
of these technologies, the use of mild thermal treatment has been useful to enhance
the effects of the technology by providing better characteristics in the final product;
however, the name “nonthermal” is used because the main effects in the product
and microorganisms are the result of pressure, electricity, light or sound, among
Table 13.1 Examples of microbial inactivation using selected nonthermal technologies
288
Microorganism Technology Food product Tested conditions Results Reference

Listeria monocytogenes LM High hydrostatic Food matrix (soybean 448 MPa, 41 C, 6 log reduction; soybean Gao et al. 2007
54004 pressure protein, sucrose and 11 min protein, sucrose and pH
bean oil) had significant effect on
reduction
Escherichia coli ATCC Pulsed electric fields Simulated milk ultra 9.5–43 kV/cm, Up to 1.2 log reduction Alkhafaji and Farid
25922 filtrated 26.4 ms, 25 C 2008
Escherichia coli ATCC Ultrasound Broth, model orange juice, 37.5 mm,20 kHz, Up to 6 log reduction; Patil et al. 2009
25922 model apple juice 30 C, 15 min higher inactivation when
cells were exposed to
previous acid stress
conditions
Hiochi bacteria Bacteriocin Raw sake Bacteriocin from 3 log reduction using 18–35 Taniguchi et al. 2009
(Lactobacillus species) Lactococcus U/ml of C101910 and
lactis subsp. 5.6 U/ml of NBRC
lactis C 12007
101910 and
NBRC
12007, pH
4.5, 4 C, 24 h
Escherichia coli type 1 Cold plasma Pericarp of mango 16 kV, 30 kHz 3 log reduction Perni et al. 2008b
Pseudomonas aeruginosa, High pressure Microbiological media 35 C, 10 MPa, 6 log reduction Furukawa et al. 2009
Aeromonas hydrophila, Carbon dioxide 1 min
Salmonella enteritidis,
Salmonella
typhimurium, Yersinia
enterocolitica,
Escherichia coli,
Staphylococcus aureus,
Listeria monocytogenes
D. Bermúdez-Aguirre and G.V. Barbosa-Cánovas
13
Table 13.2 Examples of changes or improvements in food quality/characteristics using nonthermal technologies
Characteristic Technology Food product Tested conditions Results Reference
Activity of bovine cathepsin High hydrostatic Bovine meat 0.1–650 MPa, More than 50% reduction of Buckow et al. 2010
D (meat tenderization) Pressure 20–75 C enzymatic activity using
100–400 MPa
Peroxidase (POD) and Pulsed electric fields Apple juice 40 kV/cm, Reduction of 71% PPO and Riener et al. 2008
polyphenoloxidase 100 ms, 50 C 68% POD
(PPO) activity
Tartaric esters extraction Ultrasound assisted Red grape marc Not mentioned Yield increased 16–23% using Vilkhu et al. 2008
extraction (UAE) ultrasound
Chlorophylls Pulsed electric fields Spinach puree 60 kV/cm Increased green color because Soliva-Fortuny et al. 2009
of microbial and
Recent Advances in Emerging Nonthermal Technologies
enzymatic inactivation
Texture Pressure assisted Carrot, zucchini, 600 MPa, 105 C Better texture and retained Nguyen et al. 2010
thermal apricot, red color compared to using
processing radish and jicama only high pressure or
thermal treatment
289
other. In this chapter, some of the most known nonthermal technologies such as
high hydrostatic pressure and pulsed electric fields, and some of the newest
technologies such as cold plasma will be addressed.
13.3 High Hydrostatic Pressure
Even though high hydrostatic pressure is currently used in the food industry to
process and preserve specific food items, there is still important research being
conducted. Because of the benefits observed in food products using high pressure,
in addition to microbial inactivation and enzymatic stability, many researchers are
devoted to studying other effects of high hydrostatic pressure on foods. To mention
just one example, when pressurized products were offered to panelists along with
untreated samples, the panelists could not distinguish between the treated and
untreated samples (Torres and Velazquez 2005). Also, research in the development
of new products using the particular characteristics of pressure in treating some
food components is a new and fascinating area that is being explored around the
world. Many products have been tested under high pressure conditions, such as
dairy (milk, yogurt, cheese, ice cream); fruits and vegetables in different presenta-
tions such as purée, juices, jams or chunks; meat (turkey, beef, pork, poultry),
seafood; grains; eggs; ethnic products; and specific food components such as starch
or protein, to mention just a few examples.
13.3.1 Effects on Microorganisms, Enzymes and Food

Components
Although high pressure technology was used to inactivate microorganisms in milk

in 1899 by Hite, it was not until 1990 when Japanese researchers looked again at
this technology and started to use it in food processing. The first commercial
pressurized food products were launched into the market in 1990 using high
pressure as a microbial and enzymatic inactivation tool (Rastogi et al. 2007).
Today the number of pressurized food products extends around the world, including
a number of vegetable and animal food products; in some of these products more
preservation factors, in addition to pressure, are included.
Figure 13.1 provides examples of specific food products treated with high
pressure, demonstrating how this technology achieves better quality and longer
storage. For example, Fig. 13.2a shows the extension of shelf life for macaroni salad
with vegetables, a product currently sold at the deli counter of supermarkets. After
pressurizing the product, the growth of mesophiles and lactic acid bacteria was
delayed considerably, resulting in a stable product for more than 50 days. Similar
cases have been shown in other products such as artichoke dip, guacamole,
b 10
a 8
9
8
7
Quality Index
6 7
Log CFU/g
5 6
4 5
3 4
2 3
1 2
0 1
0 10 20 30 40 50 1 2 3 4 5 6 7 8 9 10 11
Days of storage Shelf-life (weeks)
APC Non-HPP APC HPP Lactis Non HPP Lactis HPP Standard Process w/preservatives HPP
c 50 15 0.6
Vitamin C (mg/100 ml)
Folic acid (mg/100 ml)

40
Niacin (mg/100 ml)

10 0.4
30
20
5 0.2
10
0 0 0
0 400 480 540 0 400 480 540 0 400 480 540
Pressure (Mpa) Pressure (Mpa) Pressure (Mpa)
Processtime: 1 minute, Processtemperature(IT): 20°C Processtime: 1 minute, Processtemperature(IT): 20°C Processtime: 1 minute, Processtemperature(IT): 20°C
d 9
8
7
Log (CFU/g)
6
5
4
3
2
1
0
0 10 20 30 40 50 60
Days after HPP
Control-E.coli O157:H7 HPP-E.coli O157:H7 Control-Salmonella
HPP-Salmonella Control-LM HPP-LM
Fig. 13.1 Examples of the use of high pressure for some food products: (a) Shelf life extension of
macaroni salad with vegetables (pH 4.83); (b) quality extension (texture, color and nutritional
content) of RTE meats (sliced cooked ham); (c) nutrient retention (vitamin C, folic acid and
niacin) in fresh orange juice processed with high pressure; (d) inactivation of pathogenic micro-
organisms in strawberry and banana smoothie (Adapted from Avure Technologies, Inc. 2010)
hummus, dips and salsas. In Fig. 13.1b, the graph shows how the quality index of
some RTE meats is kept almost constant during storage after pressurization. This
example shows that the overall quality (texture, color and nutritional content) of
sliced cooked ham after 10 weeks of storage; the quality of the pressurized product
is the same as ham just after processing. The comparison in the plot of the common
and traditional ham processed with chemicals and additives shows the advantages
of pressurizing the product in addition to having a preservative-free product.
Figure 13.1c shows the nutrient content in orange juice after pressurizing the
juice at room temperature (20 C) and inactivating the microorganisms in the
product; the content of vitamin C, folic acid and niacin remained the same as that
in the fresh product even when using the highest pressure (540 MPa) for only 1 min.
Finally, in Fig. 13.1d, the inactivation of pathogenic bacteria such as Escherichia
coli, Salmonella and Listeria monocytogenes is shown in a popular beverage
a Model QFP 2L-700

Characteristics:
Lab scale
2 liters
Maximum pressure vessel: 100000 psi (689 MPa)
Maximum temperature: 90˚C
b Model QFP 35L-600

Characteristics:
Suitable for food product development
35 liters
c Model QFP 320L-400

Characteristics:
Suitable for seafood processing
320 liters
Maximum pressure vessel: 22000 – 58000 psi (150 - 400 MPa)
Temperature range: 4- 35˚C
d Model QFP 350L-600

Characteristics:
World’s most successful horizontal HPP system
Highest throughput worldwide
350 liters
Model QFP 100L–600

Charateristics:
Highest available throughput for small, medium and
specialty food processors
100 liters
Maximum pressure vessel: 87000 psi (600 Mpa)
Fig. 13.2 Examples of current high pressure processing equipment (Avure Technologies, Inc.
2010)
(strawberry and banana smoothie). Bacteria were successfully inactivated using

high pressure and after more than 50 days there was no recovery of any bacteria
(Avure Technologies, Inc. 2010).
After many studies on microbial inactivation using high pressure, sometimes in
combination with temperature, new preservation factors are now being tested in
combination to achieve higher inactivation and/or to preserve the original compo-
sition of some food products. An interesting example is the use of food additives
commonly used in some processed food items that were used in a combination with
high pressure treatments to explore the inactivation of Salmonella enteritidis.
Around 30 food additives were tested in combination with pressure ranging from
100 to 400 MPa, at 25 C, and holding times from 30 to 180 min. All the additives
had important synergistic effects on inactivation at 1% concentration; however,
citric acid, adipic acid, C8-sugarester, C10-sugarester, tannin, nisin, wasabi extract,
e-polylysine and protamine showed the strongest inactivation of cells (Ogihara
et al. 2009). However, the number of preservation factors used in combination
with high pressure is not limited; Rastogi et al. (2007) mention some other nonther-
mal technologies used in combination with pressure with positive results, such as
gamma irradiation, alternating current, ultrasound, carbon dioxide, and argon and
microbial peptides or bacteriocins.
The effect of high pressure on some components of foods such as proteins has
been used to improve quality parameters in target food products. For example, the
use of high pressure on meat can extend the shelf life of the product and tenderize
and soften its texture, producing a higher quality food because of the modification
of myofibrillar proteins and enzymatic (mainly proteolytic) systems (Buckow et al.
2010). Also, these changes in proteins have been used successfully in fish meat; in
carpaccio and carpaccio-like products, high pressure offers an alternative to
“processing” the product but still maintains its raw appearance. However, because
of the high pressure processing, these fish products have better texture, they are
microbiologically safer and have a longer shelf life. Gómez-Estaca et al. (2009)
showed results in these fish products using certain fish varieties such as salmon,
tuna and cod, with positive important results during the sensorial evaluation.
Another example of the use of high pressure in product development is the
production of high energy density (HED) foods. High pressure is responsible for the
modification of proteins as mentioned above, but also for the gelatinization of
starch. One of the main requirements in producing food for young children is to
meet the nutritional requirements of this age group to permit correct growth and
development. Sometimes the food offered at home does not contain enough nutri-
ents or adequate consistency. However, the use of amylolytic lactic acid bacteria
such as Lactobacillus plantarum A6, can change the rheological properties of
cereal-based children’s foods when using a pretreatment such as high pressure
homogenization, which gelatinizes starch, with further fermentation produced by
lactobacillus. This combination of processes offers an HED product that is suitable
for young children’s consumption (Nguyen et al. 2007).
An extensive research program on the quality of fruits and vegetables treated by
high pressure has been explored in the last few years. Currently, microbial quality is
not the only important parameter to consider during processing. More quality
parameters are under consideration during experimentation in addition to using
novel food products. For example, an apple-broccoli juice functional food was
tested at 500 MPa and 10 min; in addition to showing excellent microbial quality
for more than 30 days of refrigerated conditions (no presence of coliforms, yeasts,
molds, or salmonella), this product also showed similar concentration of sulforaphane
(a nutritional component), anti-mutagenic activity, and sensorial quality as a frozen
juice (Houŝka et al. 2006).
One market that has been explored with successful results in high pressure
processing is the seafood product market. High pressure by itself has been impor-
tant in improving the quality of some seafood products such as surimi, for which
there have been interesting reported results (Tabilo-Munizaga and Barbosa-Cánovas
2004, 2005). One of the most important discoveries using this technology was in the
processing of shelled seafood; high pressure can open the shells, “extract” the meat
and increase the yielding of the process, in addition to inactivating microorganisms.
All of these benefits in addition to the reduction of labor costs and increase of
product shelf life have been reported, not only for oysters (as shown in the past), but
also as shown in current studies for lobsters, clams and other fresh products (Avure
Technologies, Inc. 2010).
13.3.2 Advances in High Pressure Processing
In the last few years, high hydrostatic pressure has shown important advances. The
fact that it is one of the few nonthermal technologies commercially available today
is because of the extensive research that has provided information on this techno-
logy. Although the number of commercial products treated with high pressure was
small in the beginning, the demand for these products has increased considerably.
The most known case is with Avomex Inc. This company started in 1996 as a small
facility processing avocado paste (guacamole), which under pressure keeps its
freshness (Torres and Velazquez 2005). Worldwide, the number of facilities has
increased in addition to the number of products processed by high pressure, not only
those processed by Avomex.
Several facts have been noticed during the processing of foods using this
technology, such as compression heating. Although high pressure is considered a
nonthermal technology, in recent years there has been much research devoted to
studying the effect of temperature during the compression taking place in the high
pressure process. This change in temperature is reversible and after completion of
pressurization, the temperature returns to the starting temperature. The increase of
temperature depends on certain factors such as pressure medium and vessel,
pressurization rate, initial temperature and food composition (e.g., higher fat
content and higher temperature increase) (Patazca et al. 2007). Considering the
pressure medium, some studies have shown that three of the most common pressure
transmitting fluids used in high pressure processing have different compression
heating values, the highest being for ethanol, followed by ethylene glycol and the
last being water. Also, an important fact is that when the initial temperature is high,
the compression heating is increased (Buzrul et al. 2008). This last fact is important,
for example in microbial inactivation, because in some cases preheating of the food
is required to kill target bacteria in the product; thus, increase in temperature
because of compression during pressurization can enhance the inactivation treat-
ment (Wilson et al. 2008).
With regard to equipment, currently there are a good number of high pressure
systems available worldwide, not only at lab scale, but also for commercial
applications. Figure 13.3 presents a few examples of the high pressure systems
available at Avure Technologies, Inc. (2010), which offers options to the food
processor, such as the first unit (a) for lab research that has a volume capacity of
only 2 L and has been designed to research very high pressures (690 MPa) and
high temperatures (90 C). The second option (b) is more suitable for pilot plant
scale or product development and has a larger capacity (35 L). The third option
(c) represents one of the most innovative systems in food processing in the last
few years; this equipment is being used in the food industry to open the shells of
seafood to release the meat. This unit (Fig. 13.3c) shows the smallest equipment
available for this purpose (320 L) and can process up to 2,300 kg of seafood/h with a
processing time of 3.5 min. The biggest unit for this purpose (not shown here) has a
volume of up to 687 L, which represents a production of 500 kg per cycle, and each
pressure cycle takes approximately 3.7 min. Finally, in Fig. 13.3d the largest high
pressure equipment is shown; this device has a volume of 350 L, operates at
600 MPa, and can be used for different applications such as the pasteurization of
avocado sauce, salsa, and ready-to-eat meat, among others (Avure Technologies,
Inc. 2010). Other companies around the world that manufacture high pressure systems
for food processing are Hyperbaric NC (Spain), Engineered Pressure Systems Inter-
national (EPSI) (Belgium), Kobe Steel (Japan), Stansted Fluid Power, Ltd. (UK),
Resato International (The Netherlands), UNIPRESS (Poland), ACB Pressure System-
Alstom Hyperbar (France), and UHDE (Germany).
13.3.3 Pressure Assisted Thermal Sterilization (PATS)
One of the most significant advances in high pressure technology is the approval by
the Food and Drug Administration of the process called pressure assisted thermal
sterilization (PATS), which is basically the combination of pressure (600 MPa) and
selected temperature to achieve sterilization patterns in low-acid food products
(NCFST 2009). This approval, released in February 2009, opens a world of
opportunities in the development of other food products using high hydrostatic
pressure and also in transporting food products to remote places and storage for
longer periods of time. Just a few years ago, inactivation of spores seemed to be a
challenge for the technology, and researchers even were planning to use extremely
high pressure (GPa) to achieve inactivation, which was not feasible from the
c d
Ground Treatment
Electrode Area (Gap)
High Voltage
Electrode
Insulator Treatment
Ground Area (Gap)
Electrode
Fluid Flow
Fig. 13.3 Examples of pulsed electric field devices: (a) PEF 25 kW PowerMod®; (b) PEF
industrial system to process 1,000–5,000 L/h; (c) layout of PEF treatment chamber; (d) set up of
two PEF treatment chambers connected together (DTI 2010)
economic point of view (Torres and Velazquez 2005). Today, it is possible to

achieve inactivation of spores using the intelligent combination of pressure, tem-
perature and time without exceeding the values of those parameters. For example,
the combination of pressure (0.1–1,400 MPa), temperature (70–120 C) and holding
times (less than 8 min) were required to achieve a good degree of inactivation
(almost 6 log) of endospores of Clostridium botulinum and Bacillus amyloliquefa-
ciens using a buffer solution as the medium of treatment; however, in some samples
there was a minor number of remaining spores (Margosch et al. 2006). Spores are
very resistant organisms to most of the tested inactivation factors, such as heat,
mainly because of their composition. Clostridium botulinum spores have been
inactivated in some food matrixes using high pressure and high temperature, but
the mechanism of inactivation is not fully understood; meanwhile, for Bacillus
spores there are two possible theories about inactivation using high pressure, such
as the release of Ca-DPA (dipicolinic acid), related to germination processes or the
formation of pores in the membranes. Also, there is evidence that the presence of
minerals and increase of temperature weaken spore resistance to high pressure
treatment (Wilson et al. 2008).
13.3.4 Future of High Hydrostatic Pressure
High pressure has shown important results in the last few years, and with the
approval of PATS a new huge market is now available to commercialize new
products. These products will change the marketing of low acid foods indeed, not
only with the feasibility of having a better quality product with longer shelf life, but
also because of the kind of packaging used for these products. Commercialization
of PATS treated products could be achieved around the world, allowing the free
interchange of ethnic food products to remote places on the planet. Also, these
PATS food items can be used for the military, space missions, and humanitarian
purposes. Development of equipment is growing, together with the development of
products in different parts of the world, and there are increasingly more companies
offering several options for food processing. Food scientists must now focus on the
development of novel, innovative and safe products using this technology, which
holds promise for an exciting future.
13.4 Pulsed Electric Fields
Another nonthermal technology that has shown more advances in research and
could be the next window for commercial application is pulsed electric fields (PEF)
technology. This nonthermal technology, which involves the application of an
electric field applied in a pulsed way into a food placed between two electrodes,
has been extensively researched around the world. PEF technology was first used to
process liquid foods such as in model systems of juices and milk, followed by
further applications to real food items. Currently, another area of PEF research is
dedicated to studying the application of this technology to solid food products,
which has shown important benefits. Some advances in this technology are pre-
sented in the next paragraphs, showing the advantages of processing foods with
electric fields, as well as developments in the manufacturing of treatment chambers
and some of the innovative products treated with PEF.
13.4.1 Effects on Microorganisms, Enzymes and Food

Components
As in the case of high hydrostatic pressure, pulsed electric fields technology is being
widely explored in the juice industry. Most of the research conducted in the last few
years has been related to the application of PEF for microbial and enzymatic
inactivation in fruit and vegetable juices (Riener et al. 2008; Aguiló-Aguayo
et al. 2008a, b, 2009a, b; Evrendilek et al. 2008; Odriozola-Serrano et al. 2008;
Oms-Oliu et al. 2009; Odriozola-Serrano et al. 2009; Martı́nez-Viedma et al. 2009).
First, the inactivation of microorganisms continues to be studied under PEF
treatments using different approaches to confirm the cell death. Spores of Penicil-
lium expansum were studied after processing cherry juice, peach and apricot nectars
under PEF treatment from very low strengths (13 kV/cm) to higher strengths
(34 kV/cm) using 218 ms as the longest processing time. Results showed positively
that as the electric fields and treatment times were increased, spore germination was
completely inhibited (Evrendilek et al. 2008).
Important results have been observed in enzymatic inactivation of juices; some
enzymes that are resistant to thermal pasteurization and responsible for the reduction
of quality have been successfully inactivated using PEF. For example, in strawberry
juice, lipoxygenase (LOX) was inactivated after processing the juice at 35 kV/cm by
1,000 ms, showing a residual activity of 65% and 70% depending on the mode of PEF
operation (monopolar or bipolar). Meanwhile, a monopolar mode was more effective
than bipolar for reduction in the activity of another important enzyme in strawberry
juice, b-glucosidase (73.2%), making it possible to achieve a more stable juice during
storage (Aguiló-Aguayo et al. 2008a). PEF strawberry treated juice also showed flavor
stability during storage because of the slow development of unpleasant compounds,
compared to the thermal treated samples (Aguiló-Aguayo et al. 2009a). Similar
studies using PEF showed that the color and concentration of 5-hydroxymethyl
furfural (HMF) in strawberry juice, which is responsible for the development of
browning, were positively affected after processing (Aguiló-Aguayo et al. 2009b).
Few changes in color were observed for strawberry, tomato and watermelon juices
after processing at 35 kV/cm, 40 C and 1,000 ms, and lower concentration of HMF
was detected after PEF processing. In the case of watermelon juice treated under
PEF, lycopene, vitamin C and antioxidant capacity were evaluated after processing.
Lycopene retention and antioxidants in watermelon juice capacity showed the best
results after processing with PEF; however, for vitamin C, when the same treatment
was applied, a reduction in vitamin C content up to 72% was shown (Oms-Oliu et al.
2009).
Tomato juice is another food product that is of high importance in the food industry.
After processing with thermal treatment it shows an important decrease in quality
attributes. The search for new alternatives to process this product has led food
scientists to explore the use of PEF for tomato juice. One of the most important
enzymes in tomato juice is peroxidase (POD), which generates some quality problems
during storage. This enzyme was totally inactivated in the juice using 35 kV/cm,
2,000 ms (7 ms bipolar pulses) at 200 Hz; very low residual activities (around 8%) were
found under similar processing conditions (Aguiló-Aguayo et al. 2008b). In addition,
after treatment, tomato juice showed an enhancement of some carotenoids such as
lycopene, b-carotene and phytofluene, in making the color of the juice very red. Also,
the evaluation of phenolic compounds, pH and soluble solids remained similar to the
fresh product. Only a minor decrease in the health-related compounds was observed
after processing, but important losses were observed during storage, with the excep-
tion of b-carotene, phytoene and caffeic acid (Odriozola-Serrano et al. 2009).
Furthermore, the combination of PEF with other preservation factors such as
bacteriocins has been used recently to inactivate target microorganisms such as
lactic acid bacteria. Martı́nez Viedma et al. (2009) successfully studied the combi-
nation of PEF treatment at 35 kV/cm, 1,000 ms and enterocin AS-48 (2 mg/mL) in
apple juice to inactivate Lactobacillus diolivorans 29. The juice was stable
for 15 days at refrigerated and room temperature without the presence of this
microorganism.
13.4.2 Recent Advances in PEF Processing
Since the first experiments conducted with PEF, the treatment chamber has been the
key part of the system, in which inactivation of microorganisms and enzymes takes
place when the liquid is passed through the chamber containing the electrodes.
Treatment chambers can be classified in accordance with the operation mode, in
static and continuous. Among the first chambers, some of those that have been used
for food processing are the U-shaped, parallel plate, disk-shaped, wire cylinder, rod-
rod and the sealed static treatment chamber. For continuous processing, the most
common devices include the co-axial and co-field treatment chambers (Huang and
Wang 2009). In Fig. 13.2, images of current PEF systems, which now have a more
compact design than previous older units, are presented. Figure 13.2 a shows the
PEF PowerMod 25 kW system, which is suitable for use in pilot plant scale and has
the feasibility to work with one or two chambers together (Fig. 13.2d), allowing
recirculation of the fluid between each cycle. The maximum output voltage is 35 kV,
delivering monopolar pulses, and the maximum flow rate recommended is 10 L/min.
In Fig. 13.2c the layout of the co-field treatment chamber is shown, showing that the
treatment chamber in Fig. 13.2d and the one installed in the PowerMod 25 kW unit
have a gap distance of 0.65 cm and an electrode diameter of 0.5 cm. Finally,
Fig. 13.2b presents the first commercial PEF unit; designed by Diversified Techno-
logies, Inc. for use at Ohio State University, it has the capacity to process
1,000–5,000 L/h but can also be scaled-up to process up to 50,000 L/h (DTI 2010).
However, several issues need to be resolved to enhance and optimize the
treatment during processing. For example, the possibility of electrical arcing
because of the presence of bubbles in the food product is a big problem during
PEF processing, although this may have been fixed by using some equipment under
vacuum conditions before starting the process or installing sensors in specific parts
of the system to detect bubbles and automatically stop the process. Erosion of
electrodes because of strong processing conditions during PEF is another issue that
is currently under study and optimization by PEF manufacturers. When an electrical
current is discharged into the electrodes and a fluid passes through the chamber,
electrochemical reactions take place; as a result some metal particles from the
electrodes are released to the medium and particles from the food are deposited onto
the electrode surface (Mastwijk 2006). Some available options to minimize the
erosion of electrodes include using a specific type of pulse during processing or
using stronger and durable material such as platinum, gold and metal oxides; also,
the application of conductive polymer coatings onto the electrode can reduce the
erosion problems (Góngora-Nieto et al. 2002). Most of the electrodes currently used
around the world are made of stainless steel, and some have coatings consisting in
part of the strongest materials, such as platinum; however, these coatings are not
always strong enough when there are arcing problems during operation, which can
erode the electrode surface.
13.4.3 PEF Extraction
Several years ago, PEF technology was recognized as a new processing tool for
liquid food products; however, in the last few years PEF has become more impor-
tant in extraction processes in the food industry and for treatment of solid foods.
PEF has been applied, for example, in processing alfalfa mash to extract the juice
from the product. Even though other technologies have been applied to maximize
the yield of extraction, the juice was never totally extracted. PEF, however, was
applied to alfalfa mash with the highest electric field at 2.5 kV/cm and obtained a
maximum amount of juice of 13.88 g per 40 g of alfalfa (Gachovska et al. 2009).
PEF was able to damage the tissue of alfalfa and extract the juice of the vegetable,
although according to the authors, the capacitance of the discharge capacitor should
be higher than 1 mF to optimize the process.
Another example is the extraction of betanine extraction from red beet roots,
which was faster and with higher yield (90% of extraction) after PEF treatment than
the conventional treatment (López et al. 2009a). Also, it is known that for extraction
purposes, the intensity of the treatment does not need to be so high; lower electric
fields are needed, but longer processing times are required to achieve good extrac-
tion levels. For example, for betanine, the electric field was 7 kV/cm, 5 pulses (2 ms)
and 300 min (López et al. 2009a); these processing conditions allowed extracting
the content from the tissue, but did not support microbial inactivation, which
requires electric fields higher than 20 kV/cm in combination with a specific
temperature and pulse width conditions. A similar example is the extraction of
sucrose from sugar beet, in which an electric field of 7 kV/cm at 40 C and 60 min
was able to achieve up to 80% of sucrose extraction (López et al. 2009b). Also, the
extraction of anthocyanins and phenols from red grapes has been widely used in the
wine industry to reduce the time of maceration and to increase the extraction of
phenolic compounds (Puértolas et al. 2010; López et al. 2008a, b). In addition to
these extraction processes for wine making, PEF is also able to inactivate spoilage
yeast and bacteria in the wine. Dekkera anomala, D. bruxellensis, Lactobacillus
hilgardii, and L. plantarum are some examples of bacteria that have been success-
fully inactivated by PEF in must and wine; more than 99.9% of spoilage bacteria in
both products can be eliminated (Puértolas et al. 2009).
Some studies using PEF showed an interesting behavior in potato starch after
processing. The starch granules were damaged because of the intensity of the
electric field. Pieces of the granules after processing showed some structures that
were similar to gel and a decrease in the gelatinization temperatures as the PEF
treatment became more intense (Han et al. 2009).
13.4.4 Future of Pulsed Electric Fields
Indeed, the next technology to be launched in the market of nonthermal processing

will be pulsed electric fields. The number of patents using pulsed electric fields
technology has increased in the last few years, from very basic treatment chambers
to more complex systems (Barbosa-Cánovas and Sepúlveda 2005). In 2007, Gene-
sis, the first company in the United States to use pulsed electric fields to process
fruit juice in the Pacific Northwest, offered PEF treated juices of outstanding
quality. (Because of issues not related to the technology, this company is no longer
using PEF). The price of a PEF treated product compared with a conventional
thermal treated product is not too different; when Genesis products were in the
market, the price of their juice was comparable to that of fortified juices using
botanical ingredients, extra nutrient content, and similar characteristics in other
novel products. Lelieveld (2005) agreed with the idea that the relative cost of PEF is
not high and there is an interesting comparison in cost between thermal processing
and PEF processing starting from the initial investment. Hoogland and de Haan
(2007) presented a study in which there is an increase in cost in the initial
investment (0.8 Eurocent/L) and energy (0.2 Eurocents/L) of a PEF system com-
pared with traditional thermal processing equipment, but there is a reduction in cost
of cleaning (0.05 Eurocents/L), maintenance (0.05 Eurocent/L), and downtime
(2.5 Eurocent/L), having a final balance of 1.5 Eurocent/L favorable to PEF
equipment.
Furthermore, in 2010, at least two important companies reportedly will start
processing food products with PEF in Europe (Kempkes 2009). Some of the
European food companies currently researching and working with PEF are in The
Netherlands, Switzerland and France; the basis of their research is mainly focused
on milk. Most of the work to scale-up the PEF system has been completed, and now
some minor issues remain before pursuing the industrial use of the technology.
Although right now the technology in Europe is being used for pasteurization of
products such as juice or milk, its use could be extended in the coming years to a
number of products, not only liquids. PEF is now under research to treat some solid
food products to improve their quality for extraction purposes or microbial inacti-
vation. However, most of this information is still under research and to date few
results have been published. Nevertheless, use of pulsed electric fields has a
promising future in Europe and could be approved by regulatory agencies soon.
13.5 Ultrasound
Ultrasound is another nonthermal technology under research in the last few years,
as a possible tool to inactivate microorganisms and enzymes. While the lethal
effects of ultrasound in microorganisms have long been known, in the last 2 decades
there has been a strong research emphasis at several universities and research
centers around the world regarding the use of ultrasound for bacteria inactivation.
Positive results have been found when ultrasound is used in combination with
temperature (thermo-sonication) or pressure (mano-sonication) or even both
(mano-thermo-sonication) in the inactivation of pathogenic bacteria and spoilage
microorganisms. Also, the use of temperature and ultrasound together have been
successfully used to reduce the enzymatic activity in some target products such as
juices, providing better stability during storage. One of the most explored products
in the last 5 years has been milk under sonication, which shows positive results in
pasteurization standards, better homogenization and color, as well as new physical
properties for the development of dairy products.
13.5.1 Extraction
Another interesting application of ultrasound is the extraction of components from

food matrixes. For example, the extraction of polyphenols is of great interest in the
food industry, and these compounds are retained in apple pomace during juice
production. Polyphenols are very important as preservatives and antioxidants, in
addition to their importance in other industries. Ultrasound was used to study the
extraction of these compounds from apple pomace and to improve the typical
maceration extraction process (Virot et al. 2009). Results showed that ultrasound
increased extraction yield up to 20% and reduced the time considerably. This fact
can be attributed to the series of explosions and implosions generated by cavitation
during sonication, which produces higher pressure and temperature and disrupts the
tissue and vegetable walls of the apple, thus more efficiently extracting the com-
pounds. Similar results were observed in phenolic extraction from citrus peel, in
which compounds such as caffeic, p-coumaric, ferulic, sinapic, protocatechuic,
p-hydroxybenzoic and vanillic acid were studied under sonication using 40 C as
the highest temperature. Again, the extraction yielding phenolic compounds was
higher when time and temperature were increased; however, temperature was an
important factor in the stability of the phenolic compounds and had to be controlled
in each particular case (Ma et al. 2009).
Ultrasound, also called ultrasound-assisted extraction (UAE), has been used to
extract phospholipids from palm-pressed fiber (Chua et al. 2009). Typical extrac-
tion of phospholipids from palm oil takes a considerable amount of time and
requires the use of solvents. These phospholipids are currently used in the food
industry as emulsifiers and are of great importance in specific products. The need to
have a higher yield in extraction in addition to a faster and solvent free method is
highly desirable. In the study conducted by Chua et al. (2009), the extraction of
phosphatidylethanolamine and phosphatidylcholine was enhanced with the use of
ultrasound; a simpler process with a higher yield and shorter processing times was
achieved with this nonthermal technology.
Other applications of UAE include the extraction of herbal extracts such as
fennel, hops, marigold, mint, geniposide, carnosic acid from rosemary, almond oils,
ginseng saponins, ginger, soy protein and soy isoflavones, as well as polyphenols,
amino acid and caffeine from green tea and pyrethrines from flowers. Some
bioactive compounds such as beta-carotene, other polyphenols, and gingerol have
been successfully extracted with UAE from carrots, red grape marc, black tea, apple
and ginger (Vilkhu et al. 2008).
13.5.2 Recent Advances in Ultrasound
As mentioned before, ultrasound has been used to inactivate bacteria and enzymes,
and some of the most studied products have been fruit juices because of the easy
application of ultrasound to these products. Favorable results have been observed in
microbial inactivation and stability during storage, and minor changes have been
observed; however, there is still a lack of information regarding other components
after processing, for example nutritional content in juices. One of the most impor-
tant components in red grape juices and wines is anthocyanins because of their
excellent antioxidant activity; however, most of the anthocyanins are unstable
during processing. Ultrasound was used to study anthocyanin content after proces-
sing and encouraging results were observed. Higher retention was observed in
anthocyanin content after processing; the three major anthocyanins in red grape
juice, cyaniding-3-O-glocosides, malvanidin-3-O-glucosides and delphinidin-3-O-
glucosides, were retained respectively in the juice 97.5, 48.2 and 80.9% after
sonication (Tiwari et al. 2010). Also, non-enzymatic browning and ascorbic acid
degradation were studied in sonicated orange juice. Low temperature (5 C) and
intermediate amplitudes (42.7 mm, 20 kHz) were the best conditions for achieving
less degradation of ascorbic acid and less browning (Valdramidis et al. 2010).
These are good examples of the use of ultrasound for juice processing; sonication
inactivates bacteria and enzymes, retains higher concentration of nutrient com-
pounds and maintains a product with fresh-like characteristics.
Other potential advances in ultrasound technology are related to extraction

processes. Many extraction studies are currently being conducted, but in addition
to the extraction process, ultrasound is able to encapsulate the extracted compound
through covalent bonding and microsphere formation. This research area, which is
very new, mixes different frequencies during operation and uses different com-
pounds to extract and encapsulate target compounds (Vilkhu et al. 2008). Very few
experiments have been conducted in this area, but by applying the basic principles
of ultrasound and the appropriate combinations of processing parameters, food
components and other processing variables, this process has huge potential in
food applications and other related industries.
Some research has been started in the last few years related to the sonochemistry
in certain foods products to study the reactions that ultrasound generates in food
during processing. While some of these reactions could produce toxic compounds,
the intelligent selection of frequency and other processing variables could be useful
in reducing the amount of toxic compounds (if any) and achieving similar results in
terms of inactivation. Also, these chemical reactions can be used to generate new
compounds in foods with specific purposes such as modification of proteins (Jambrak
et al. 2009), hydroxylation of phenolic compounds to enhance their antioxidant
properties, and so on (Ashokkumar et al. 2008).
Ultrasound comprises a very wide area of research that has been explored deeply
in the last few years, and currently presents a number of challenges and has raised
some other issues that need to be resolved in areas such as equipment manufacturing.
Many researchers around the world are now looking at ultrasound not only as a
technology to kill bacteria and produce safe products, they are also looking at how to
use the properties of this technology to improve the quality of food and to find new
alternatives to making a stable product and having new and improved food ingre-
dients. Ultrasound is currently used with many other processes such as emulsifica-
tion, homogenization, crystallization, dewatering, degassing, defoaming, particle
size reduction and viscosity alteration (Patist and Bates 2008). Most ultrasound
research equipment is on a lab and pilot plant scale, but further scale-up technology
to the industrial level is highly desirable and possibly will occur once the majority of
small technical difficulties in lab and pilot plant devices have been solved.
13.6 Cold Plasma
When an electric field is applied to a gas, molecules are ionized and gas plasma is
produced. The electrons, which are usually bound to molecules of gas, start to separate
from the gas; the resultant ionized gas is known as discharge plasma (Perni et al.
2008b). Some authors consider plasma as a fourth state of matter (Selcuk et al. 2008)
and that the mix of free radicals, excited species (atoms and molecules), free electrons
and ions are responsible for microbial inactivation. Atomic oxygen has been found in
plasma and is considered to be one of the most lethal components for inactivating
microorganisms (Perni et al. 2008b). One of the main advantages of plasma in food
applications is that this technology can be used at low temperatures, for example, as
low as ambient temperature (Perni et al. 2008a); that allows microbial inactivation
because of the generation of free radicals, resulting in a food product free of any
residues and that retains the fresh-like characteristics of the food.
13.6.1 Processing Conditions
Studies related to microbial inactivation in food using cold plasma are very scarce and
can be considered as a first-generation technology in food research (Niemira and Sites
2008). As a new technology, the interest in this nonthermal option is rising and some
attempts to inactivate bacteria are under development. Plasma basically consists of a
corona discharge from an electrode; it has been used mainly on food surfaces for
disinfection (Perni et al. 2008a). Corona discharge has a very high temperature, on the
order of 1,000 K, but the heat located in the electrode is quickly dissipated. In the past
the use of vacuum conditions were applied during the generation of plasma to maintain
low temperature; today, cold plasma is generated at voltages below those used for
corona discharge effects at higher frequencies than those for dielectric-barrier dis-
charge (Perni et al. 2008a). A schematic view of a cold plasma device is shown in
Fig. 13.4; the system basically consists of a pair of electrodes, one grounded and the
other powered, and a ceramic tube. The gas inlet is set at one of the extremes in the
ceramic sheath; once the gas comes into contact with the electric discharge, the plasma
is generated on the surface of the sample. Today, cold plasma is tested for disinfection
Gas inlet
Ceramic
tube
Gas
mixture Powered
electrode
Plasma
Fig. 13.4 Layout of cold

plasma device (Adapted from
Perni et al. 2008b; Deng et al. Ground electrode
2006)
of food surfaces only; the microbial inactivation produced is attributed to the forma-
tion of free radicals in the plasma environment. Three main categories for plasma
generation can be cited: electrode contact, direct treatment and remote treatment
(Niemira and Sites 2008). The main difference between these approaches is whether
or not the plasma is in contact with the food item.
13.6.2 Effect on Microorganisms
The effect of cold plasma on microorganisms is related mainly to the generation of

certain chemical substances and their ability to injure and kill bacteria, yeast, molds
and other organisms. These chemicals are dissociated species of oxygen such as
ozone, atomic oxygen, hydroxyl, nitric oxide, super oxide radicals and other free
radicals (Selcuk et al. 2008).
Some cold plasma studies in food research are related to the inactivation of
Listeria innocua and Escherichia coli in apples, Salmonella enteritidis in canta-
loupe, Listeria monocytogenes in lettuce (Niemira and Sites 2008), and Aspergillus
spp. and Penicillum spp. in several types of seeds (tomato, wheat, bean, chick pea,
soy bean, barley, oat, rye, lentil and corn) (Selcuk et al. 2008).
Recent studies in microbial inactivation of Escherichia coli cells after plasma
treatment show important changes in their structure and breakdown of the cell
membrane, with loss of cellular components from the main cell (Critzer et al. 2007).
The main lethal effects have been attributed to the presence of oxygen atoms,
metastable oxygen molecules, and OH radicals (Yu et al. 2006) because they have
high oxidization potential in the cells. One theory about microbial inactivation is
related to the oxidation of amino acids and nucleic acids in cells as well as the
effects in lipids of the membrane (Critzer et al. 2007).
13.7 Dense Phase Carbon Dioxide
Another nonthermal technology that has been under research in the last 20 years is
dense phase carbon dioxide (DPCD), also called high pressure carbon dioxide
(HPCD); this technology allows microbial inactivation in foods and has seen a big
peak in research in the last 5 years (Garcia-Gonzalez et al. 2007). In this nonthermal
technology, carbon dioxide at supercritical conditions (temperature and pressure) is
applied to foods in the batch, semi-batch or continuous operation mode (Garcia-
Gonzalez et al. 2009); the pressurized CO2 diffuses through the material, dissolving
the cell components, generating microbial inactivation (Ferrentino et al. 2010a). The
critical conditions of CO2 (Tc ¼ 31 C, Pc ¼ 7.35 MPa, and up) are used for solid
and liquid foods. CO2 has the capability of diffusing throughout solid foods in a gas
form and dissolving materials similar to a liquid (Garcia-Gonzalez et al. 2007). The
critical conditions of CO2 are shown in the thermodynamic diagram in Fig. 13.5.
Fig. 13.5 Thermodynamic diagram (temperature–Pressure) for CO2 showing its critical conditions
As in other novel technologies, the mechanism of inactivation of microorga-

nisms is still unclear, but there are theories suggesting important changes in the pH
of the cell with subsequent damage in the cell membrane, and an unbalance of
electrolytes in the inner cell content with the final removal of vital components in
the cell and membrane (Garcia-Gonzalez et al. 2009).
In a study conducted by Garcia-Gonzalez et al. (2009) with a number of bacteria
and selected processing conditions of DPCD, results showed the resistance of
microorganisms to this technology, the Gram negative bacteria followed by Gram
positive bacteria and yeasts, with spores being the most resistant. To achieve spore
inactivation using DPCD, the use of moderate temperatures (60 C and higher) in
combination with this technology is required to obtain a good reduction degree
(Furukawa et al. 2009; Garcia-Gonzalez et al. 2007). Some of the factors that affect
microbial inactivation using DPCD are the pH of the medium and the water activity
(aw). When the pH of the medium is low, the microbial inactivation with DPCD is
higher, but with a decrease in aw the cells become more resistant. According to the
type of solute, cells were weakened with the presence of NaCl; meanwhile sucrose
and glycerol protected the cells (Garcia-Gonzalez et al. 2009). Other tested micro-
organisms using DPCD were Saccharomyces cerevisiae (Ferrentino et al. 2010b;
Parton et al. 2007a), Pichia awry (Parton et al. 2007b), Pseudomonas aeruginosa,
Aeromonas hydrophila, Salmonella enteritidis, S. typhimurium, Yersinia enteroco-
litica, Escherichia coli, Staphylococcus aureus, Listeria monocytogenes (Furukawa
et al. 2009), Bacillus subtilis (Parton et al. 2007a), Leuconostoc dextranicum (Lin
et al. 1993), and Enterococcus faecalis (Debs-Louka et al. 1999), among others.
Few studies have been conducted using DPCD to analyze other food properties.
Khorshid et al. (2007) used DPCD to precipitate protein from soy. Results of this
experiment showed that supercritical CO2 was useful in isolating pure protein,
having a higher recovery than with traditional methods; it also saved energy. Few
applications of DPCD for extraction purposes have been reported in food systems
and there have been few studies of this novel technology for microbial inactivation
in solid foods because of the possibility of altering the sensorial properties (Garcia-
Gonzalez et al. 2007).
Indeed, as in other technologies, DPCD is a wide research field with many
possibilities to explore. Research currently being conducted is focused on designing
the technology and improving the reactions to its use in processing of foods, as well
as some studies related to the solubility conditions of CO2 in water and other
complex systems at specific conditions (Parton et al. 2007b, b; Ferrentino et al.
2010b). A number of patents have been produced in the past few years regarding
DPCD equipment for pasteurization, especially in the United States, Germany and
Japan (Garcia-Gonzalez et al. 2007). Based on the current results of microbial
inactivation and quality characteristics of food, processing conditions have been
modified to achieve specific objectives.
13.8 Other Novel Nonthermal Technologies for Food Processing
Other nonthermal technologies that have been under research in the last few years
are ultraviolet (UV) and the use of certain compounds with antimicrobial activity
such as bacteriocins. In the case of ultraviolet technology, most of the applications
have been related more to the application of UV light on the surface of fresh cuts of
fruits and vegetables than in liquid products. For example, ultraviolet was applied to
fresh cuts of watermelon before packaging, followed by evaluation of the character-
istics of the fruit, such as CO2 and C2H4 production, microbial counts (mesophilic,
psychrophilic and enterobacteria), color parameters, nutrient compounds such as
lycopene, polyphenols and ascorbic acid and sensorial attributes. Results showed
that ultraviolet treatment can extend the shelf life of cut watermelon with minimal
changes in the quality characteristics (Artés-Hernández et al. 2010). Similar studies
were conducted with apple slices, using a pretreatment of hot water for blanching
and then dipping the slices into solutions with ascorbic acid and calcium chloride.
Browning and the growth of bacteria were delayed in samples treated with UV light
(Gómez et al. 2009). Other studies included fresh cut fruits such as pineapple,
guava, and banana (Alothman et al. 2009), and vegetables such as onion, escarole,
carrot and spinach (Selma et al. 2008).
In the field of bacteriocins, probably the most studied microorganism using these
compounds is Listeria spp., with bacteriocins found in seafood (Pinto et al. 2009); in
dairy products using peptide cerein 8A (Bizani et al. 2008); and in soft cheese using
E. faecium WHE 81 (Izquierdo et al. 2009). Even though a good number of bacter-
iocins have been isolated and used against specific target microorganisms in other
products, the only bacteriocin that has been approved for use in the food industry
(specifically for dairy products) is nisin, in specific concentrations. The use of
bacteriocins around the world is different in each country and legislation allows the
use of bacteriocins in specific products, which does not permit generalizations regard-
ing concentrations and types of food. However, research in this area is ongoing.
The list of other nonthermal technologies under research is still long; techno-
logies currently being tested in lab facilities include ozone by itself, ozone com-
bined with ultraviolet or electrolyzed water, the use of organic acids, supercritical
water, natural antimicrobials derived from species, and a combination of such with
most of the well established nonthermal technologies.
13.8.1 Future of Other Novel Technologies
Today, hundreds of food products have been evaluated in relation to at least one
nonthermal technology, depending on the product and the objective of the process.
Table 13.3 provides a brief summary of some of the tested technologies according
to the kind of product and the achievements accomplished for each one. These
newest technologies are being researched by food scientists for use in food proces-
sing and are indeed showing encouraged results.
Ultrasound was first used in the early 1900s, but interest was lost in the following
years; now it is becoming popular again in food processing. This technology has been
shown to be important not only for microbial inactivation, but also for reduction of
enzymatic activity to provide food products with longer stability. In addition to its
inactivation effects, ultrasound is able to reduce processing times considerably
because of the two in one effect, for example in milk processing (pasteurization and
Table 13.3 Examples of nonthermal technologies suitable for processing specific foods
Nonthermal Type of product Examples Results
technology
High hydrostatic Liquids Juices Pasteurization standards
pressure
Liquid-solids Eggs Sterilization
Gel-like Jams, jellies Pasteurization
Solids Cheese, meat Improved quality characteristics
(e.g., texture)
Pulsed electric fields Liquids Juices, milk Pasteurization standards
Ultrasound Liquids Juices, milk Pasteurization standards
Dense phase carbon Liquids Juices Pasteurization standards
dioxide
Ultraviolet Liquids Water, juices Pasteurization standards
Solids Fresh-cuts Disinfection
Cold plasma Solids Nuts, fresh-cuts Disinfection
Bacteriocin Liquids and solids Dairy products Delay of spoilage degree
homogenization in one step), product development (emulsification and reduction of

enzyme activity), cheese-making (increase of yielding, reduction of time for curdling
formation, and milk pasteurization), among others. This technology offers many
possibilities for use in the food industry. However, more research must be conducted,
mainly concerning the toxicological aspects of foods following ultrasound treatment.
As for the technologies ultraviolet light and ozone, both are currently used to
some degree in the industry, mainly for drinkable water. The feasibility of ultravio-
let radiation being transmitted through transparent water is the main characteristic
that allows its use for this product. To the contrary, the limitation of ultraviolet light
into opaque fluids such as milk is an issue, because the presence of fat globules and
casein micelles in milk prevents the radiation from being transmitted correctly.
Ultraviolet light has been tested to disinfect the surface of certain food products
with positive results; the main problem is when the product has a very porous
surface, since bacteria can hide inside the porous surface and thus cannot be reached
by ultraviolet radiation. Nevertheless, ultraviolet light is still being researched for
some of these products, and food scientists are trying to combine technologies or
find alternatives to process specific foods, such as the case of milk, using turbulent
flows to allow better contact between ultraviolet light and the milk.
Although plasma is probably the newest nonthermal technology for food proces-
sing, its research is becoming very important. Lately, conferences formerly devoted
to the use of plasma in general are now offering specific sessions related to the use
of plasma in food processing. Current research being conducted on plasma could
possibly be setting up the basic knowledge needed to understand how certain foods
react with plasma. This research represents a huge field of opportunities for study
not only in the area of microbial inactivation with plasma, but all aspects related to
the toxicology of the generated compounds at specific processing conditions,
including those generated into the food after processing.
13.9 Modeling of Microbial Inactivation in Nonthermal

Technologies
One common denominator in the nonthermal technologies is the inactivation patterns

of bacteria, spores and enzymes, which in most of the cases do not follow first order
kinetics. Much research has been conducted in the study of inactivation curves after
processing samples with high pressure, pulsed electric fields, ultrasound or ultravio-
let, to mention just a few. However, there are a number of reasons (or artifacts)
explaining why microbial inactivation does not follow a first order kinetics; some of
these are related with to experimental procedures, the mix of microorganism strains,
clumps of bacteria, protective effects of the media, or a non-homogeneous applica-
tion of the preservation factor (McKellar and Lu 2004). Some of the most common
shapes in microbial inactivation that do not follow first order kinetics include upward
or downward concavities and sigmoid curves (Peleg 2000). Most of these concavities
and specific shapes in microbial inactivation are called “shoulders” and “tails” in food
engineering and these are shown in Fig. 13.6, representing a typical inactivation
curve in nonthermal technologies. Here, the need to find alternative mathematical
approaches to fit and model the survivor data is a challenge for food scientists.
Survivor data allows a comparison of the inactivation degree under different proces-
sing conditions in order to find an equivalent inactivation for thermal processing.
Some of the mathematical models that have been used for fitting the survivor
data after using nonthermal technologies are the Weibullian equation, Fermi equa-
tion, Gompertz model, Baranyi model, Bigelow model, Peleg’s equation and others,
as presented in Table 13.4. For example, for pulsed electric fields inactivation the
three models used the most to fit the survival curve relating the electric field
strength and treatment time are the Huelsheger model, Peleg’s model and the
Weibull distribution (Fox 2007); all of these models are shown in Table 13.4.
A typical dose–response curve shows the resistance of a specific microorganisms
against a lethal agent; this agent can be any preservation factor such as chemicals or
nonthermal technologies. When the dose–response curve shows symmetry, one of
the options is to fit the data to the Fermi equation (Peleg et al. 1997):
1
SðXÞ ¼ (13.1)
1 þ exp XX
a
c
where S(X) is the fraction of survivors, X and Xc are the dose of the lethal agent, Xc
being the point of the curve when there is an inflection point or when the population
becomes 50% of the initial and a is the steepness of S(X) around Xc. It should be
noted that in some of the references cited in Table 13.4, where this equation was
used to fit the data after different processing technologies, the dose (X) also means
the degree of pressure applied, the intensity of the electric field, the concentration of
antimicrobial, or the intensity of ultrasound wave, among others.
Fig. 13.6 Example of an inactivation curve showing three common regions in nonthermal
processing: shoulders, linear region and tails
Table 13.4 Example of mathematical approaches to modeling microbial inactivation using nonthermal technologies
312
Nonthermal technology Example of inactivation Mathematical models Definition of terms Reference

Combination of nonthermal Inactivation of Listeria monocytogenes in ð1F1 Þð1þeb2 tL Þ Y: Log10 count of bacteria Buchanan et al.

hurdles (temperature, pH, model systems, showing significant tailing, 1þeb2 ðttL Þ at time t 1994
sodium chloride) perhaps because of mixed population Logistic-based equation Y0: Log10 at time zero
t: time
tL: duration of lag period
prior to initiation of
inactivation
F1: fraction of initial
population in major
group
b1: 2.3/D1 ¼ inactivation
rate for major group
b2: 2.3/D2 ¼ inactivation
rate for minor group
log N BM BðtMÞ
High pressure Inactivation of Listeria innocua in broth ¼ Cee Cee N/N0 ¼ Survival fraction Saucedo-Reyes
N0
Modified Gompertz equation B: relative death rate et al. 2009
qB þ ð1 qB ÞekmaxðtBðtÞÞ C: difference in value
Baranyi model between upper and
lower asymptote
M: maximum death rate
time
t: treatment time
qB: tailing ratio
kmax: maximum death rate

1
1n
High pressure thermal Inactivation curves of endospores of log N ðn1Þ N: number of survivors at Margosch et al.
sterilization (PATS) Clostridium botulinum and Bacillus N0 ¼ log 1 þ k N0 t ðn 1Þ time t 2006
amyloliquefaciens showing tailing nth order kinetics model N0: initial spore count
t: time
k: rate constant
n: reaction order
High pressure thermal Inactivation of Bacillus amyloliquefaciens log N=ðN1 0"m Þ ¼ ½ðbtÞ"n N: number of survivors at Rajan et al. 2006
sterilization (PATS) in egg patty mince time t
Weibull based equation N0: initial spore count
t: time
b, n: shape factors

13
High pressure in combination Inactivation of Zygosaccharomyces bailii lnðkÞ ¼ a1 þ bT1 P Pref þ gT1 ðP Pref Þ2 k: kinetic rate constant Reyns et al. 2000
with temperature in juices h

i P: pressure
dT1 ðT Tref Þ eT1 T ln TTref þ Tref
T: temperature
þ Tm ðP Pref ÞðT Tref Þ a, b, g, d, e, m: kinetic
Hawley equation parameters
lnðkÞ ¼ a2 þ b2 ðP Pref Þ
þg2 ðP Pref Þ2 þ d2 ðT Tref Þ
þe2 ðT Tref Þ2 þ m2 ðP Pref ÞðT Tref Þ
Quadratic equation
Pulsed electric fields Inactivation of Escherichia coli in juices logðSÞ ¼ Dt S: survival fraction at Rodrigo et al. 2003
following a non-linear trend time t
Bigelow model D: Decimal reduction time
lnðSÞ ¼ bðln t ln½ðtc ÞÞ b: regression coefficient
H€ ulsheger model t: treatment time
n
lsðSÞ ¼ at tc: critical treatment time
Weibull distribution function a: scale factor
n: shape factor
t p
Pulsed electric fields Inactivation of Lactobacillus plantarum Log10 NN0t ¼ d
Nt: microbial population at Gómez et al. 2005
in buffer and apple juice showing time t
concave upward curves Weibull distribution N0: microbial population
at time zero
t: treatment time
d: first decimal reduction

time
p: shape parameter
Pulsed electric fields Inactivation of Escherichia coli in buffer SðtÞ ¼ pek1 t þ ð1 pÞek2 t S(t): fraction of survivors Alvarez et al. 2003
Exponential based model t: treatment time
1
CFUðtÞ ¼ CFUð0Þ 1 þ e tm p: fraction of survivors in
s2
Sigmoidal equation population 1
1 t n
log10 SðtÞ ¼ 2:303 b
(1-p): fraction of survivors
in population 2
Weibull based model k1: specific death rate of
log10 SðtÞ ¼ aLnð1 þ ctÞ subpopulation 1
Empirical equation k2: specific death rate of
subpopulation2
313
CFU: concentration of
survivors
m: peak of PEF resistance
(continued)
314
Nonthermal technology Example of inactivation Mathematical models Definition of terms Reference

s: proportional parameter
b, n: shape parameters
a, c: parameters for
characteristics
Pulsed electric fields Inactivation of Escherichia coli and SðtÞ ¼ ekT ðEÞt s(t): fraction of survivors Amiali et al. 2007
Ea
Salmonella enteritidis in egg yolk using kT ðEÞ ¼ k0 e RT t: treatment duration
temperature and PEF Arrhenius models kT(E): kinetic rate constant
T: treatment temperature
Ea: activation energy
R: gas constant
k0: constant
Pulsed electric fields Inactivation of Lactobacillus plantarum in 100
h i s: percentage of survivors Rodrigo et al. 2001
s¼
EVc ðnÞ
orange-carrot juice 1þexp kðnÞ E: field strength
Peleg model Vc(n): critical value of E
(survival level 50%)
k(n): parameter indicating
steepness of curve
Pulsed electric fields Dose–response models based on sigmoid X1 X1/X0: survival fraction Lado and Yousef
X0 ¼ 1þe 1EEc
ð k0 Þ
curves E: electric field 2002
Fermi’s equation EC: critical electric field,
to reduce population
to 50%
k’: slope of steepest
segment of curve
1
t n
Ultrasound Inactivation of Escherichia coli in apple cider Log NN0 ¼ 2:303 b
Nt: microbial population at Ugarte-Romero
time t et al. 2006
Weibull based model N0: microbial population
at time zero
t: treatment time

b, n: shape factors
Ultrasound Inactivation of Listeria innocua in milk log NN0t ¼ b tn Nt: microbial population at Bermúdez-Aguirre
showing a non-linear trend time t et al. 2009
Weibull

distribution N0: microbial
n population at time zero
log NN0t ¼ k1 tn1 k2 t 2

t: treatment time
13
Sigmoid or four-parameter model b: rate parameter

n: shape parameter
k1, k2, n1, n2: constants
Ultrasound Listeria monocytogenes under manosonication logDms ¼ logD0 0:0091 ðA 62Þ Dms ¼ decimal reduction Pagán et al. 1999
and manothermosonication time
logDms ¼ logD0 0:0026 D0 ¼ control decimal
p þ 2:2x106 p2 reduction time
A: wave amplitude
p: static relative pressure
Ultrasound Escherichia coli under sonication in water k m1 N: microbial population at Ince and Belen
lnN ¼ N0 þ mþ1 t
time t 2001
N0: microbial population
at time zero
k: rate coefficient
h i m: empirical constant
Ultrasound Saccharomyces cerevisiae inactivated with CFUt ¼ CeeðAþBtÞ CeeðAÞ Log[CFU(t)/CFU(0)]: log Guerrero et al. 2005
0
log CFU
ultrasound and low weight chitosan in surviving fraction
broth Modified Gompertz equation A, B, C: different regions

of curve
Cold plasma Inactivation of Escherichia coli exposed on 1þCc ð0Þ N(t): cell concentration at Yu et al. 2006
NðtÞ ¼ N0 ek max t 1þCc ð0Þek max t
surface of membrane filters time t
N(0): initial cell
concentration
kmax: maximum
inactivation constant
Cc(0): initial concentration
of critical component

Dense phase carbon dioxide Inactivation of Salmonella typhimurium in y ¼ Aexp exp kdmA e ðg tÞ þ 1 A(Nmin/N0): lower Liao et al. 2010
carrot juice Gompertz equation asymptote value for
time approaching
infinity
A kdm: inactivation rate
y¼ 4kdm
1þexp½ A ðgtÞþ2
g: duration of lag phase
Modified logistic equation t: treatment time

1
Ultraviolet light and ozone Inactivation of Escherichia coli and t b N/N0: survivors fraction Bialka et al. 2008
log10 NN0 ¼ 2:303 a
Salmonella enterica in raspberries and t: treatment time
315
strawberries Weibull based equation a: characteristic time

b: shape parameter
Another equation commonly used to fit the survival data is the Weibullian
model; this model has been modified in specific cases according to the experimental
procedures and results. Basically, the Weibullian equation represents the distribu-
tion of the lethal dose applied to the microorganism (Peleg et al. 1997). The most
basic form of this mathematical model is:
log N
¼ b tn (13.2)
N0
where N is the microbial load at a specific time, N0 is the microbial count at time zero,
t is the treatment time and b and n are called shape factors. The parameter b is related
to the velocity of inactivation, while n is the parameter that measures and determines
the curve’s concavity. According to the value of n, it is possible to determine the
concavity of the curve. When n < 1 indicates an upward concavity, then the shape of
the inactivation curve shows the presence of “tailings”; meanwhile, when n > 1 there
is a downward concavity showing “shoulders” (Bermúdez-Aguirre et al. 2009).
The Weibullian model has been used for isothermal survival curves, but when
the temperature or the lethal agent is different during the whole process (i.e., non-
isothermal), then each part of the curve becomes a function of the lethal agent. For
example, if the inactivation generated with heat and temperature is not constant,
then the Weibullian equation needs to be modified to be a function of momentary
temperature, i.e., T(t), and transformed into a differential equation (Peleg et al.
2003) as follows:
n[ T(t) ] - 1
dLog10 SðtÞ Log10 SðtÞ n[ T(t) ]
¼ b½TðtÞn½TðtÞ (13.3)
dt b½TðtÞ
where S(t) represents the ratio N/N0.

Other specific cases have been addressed depending on the microorganism,
processing conditions and/or technology. For example, some studies have been
conducted on spore inactivation, which differs from vegetative cell inactivation
because of the intrinsic properties of the spores. It has been shown that when a
specific preservation factor is applied to spores, the inactivation is not feasible in
the first minutes of treatment, even though an activation of spores or germination
has been reported. In this case, the survival curve shows some concavities, which in
most cases represents a “shoulder.” Corradini and Peleg (2003) studied this specific
case and mentioned an “activation shoulder” that can show an upward concavity,
linear trend or downward concavity in the first part of the inactivation curve. This
“shoulder” generated in the curve because of the spore activation can be described
using three mathematical approaches, as suggested by Corradini and Peleg (2003):
t
Y1 ðtÞ ¼ (13.4)
k1 þ k2 t
where k1 and k2 are constants, t is the treatment time and Y1 the survival fraction
Y2 ðtÞ ¼ 1 lnf1 þ exp½bðt tc Þgm (13.5)
where b, m and tc are constants. Combining both equations, the general model can
be written as:

NðtÞ t½1 lnf1 þ exp½bðt tc Þgm
log ¼ (13.6)
N0 k 1 þ k2 t
Other interesting mathematical approaches have been reported by Peleg and

Penchina (2000) where the lethal agent against the microorganism shows a varia-
tion in intensity during treatment (heating, cooling, oscillating changes in agent,
concentration, etc.) and where some of the “artifacts” during processing (and others
mentioned above) are shown in the inactivation curve (i.e., mixed population,
effects of the medium, processing conditions) and need to be fitted (Peleg and
Cole 1998).
13.10 Final Remarks
This chapter has given a brief update on the development of nonthermal techno-
logies under research. Most of these technologies today are showing interesting
results not only in microbial inactivation or reduced enzymatic activity, but also in
providing food scientists with new input about product development, including
some fascinating new tools for specific operations. These discoveries were made by
food scientists trying to inactivate bacteria and enzymes in different foods; after
comparing the processed product with the control or untreated sample, they
observed specific changes in characteristics such as better color, homogeneity,
retention of nutrient or aroma compounds, higher efficiency of specific processes
or longer stability in microbial and enzymatic activity.
Right now some of these technologies are being used in combination to
enhance specific effects and to provide foods with outstanding quality. Also,
consumers are becoming more aware of what foods they are buying at the time
of purchase and of the benefits these novel technologies and novel products can
offer.
The current challenge of food scientists is to keep searching within these
technologies for the solutions to food problems related to spoilage, nutrient degra-
dation and damage to sensorial attributes after processing, but also to explore these
new tools in order to offer the consumer new and high quality products at an
affordable price, as well as to extend the shelf-life of products to allow interchange
around the world. Furthermore, the perception of the consumer regarding long shelf
life and freshness of the product could be improved by providing complete infor-
mation about these technologies.
References
Aguiló-Aguayo I, Oms-Oliu G, Soliva-Fortuny R, Martı́n-Belloso O (2009a) Flavour retention and

related enzyme activities during storage of strawberry juices processed by high-intensity
pulsed electric fields or heat. Food Chem 116:59–65
Aguiló-Aguayo I, Soliva-Fortuny R, Martı́n-Belloso O (2009b) Avoiding non-enzymatic brown-
ing by high-intensity pulsed electric fields in strawberry, tomato and watermelon juices. J Food
Eng 92:37–43
Aguiló-Aguayo I, Sobrino-López A, Soliva-Fortuny R, Martı́n-Belloso O (2008a) Influence of
high intensity pulsed electric field processing on lipoxygenase and b-glucosidase activities in
strawberry juice. Innov Food Sci Emerg Technol 9:455–462
Aguiló-Aguayo I, Odriozola-Serrano I, Quintão-Teixeira J, Martı́n-Belloso O (2008b) Inactivation
of tomato juice peroxidase by high-intensity pulsed electric fields as affected by process
conditions. Food Chem 107:949–955
Alkhafaji S, Farid M (2008) Modeling the inactivation of Escherichia coli ATCC 25922 using
pulsed electric field. Innov Food Sci Emerg Technol 9:448–454
Alothman M, Bhat R, Karim AA (2009) UV-radiation induced changes of antioxidant capacity of
fresh-cut tropical fruits. Innov Food Sci emerg Technol 10:512–516
Alvarez I, Virto R, Raso J, Condón S (2003) Comparing predicting models for the Escherichia coli
inactivation by pulsed electric fields. Innov Food Sci Emerg Technol 4:195–202
Amiali M, Ngadi MO, Smith JP, Raghavan GSV (2007) Synergistic effect of temperature and
pulsed electric field on inactivation of Escherichia coli O157:H7 and Salmonella enteritidis in
liquid egg yolk. J Food Eng 79:689–694
Artés-Hernández F, Robles PA, Gómez PA, Tomas-Callejas A, Artés F (2010) Low UV-C
illumination for keeping overall quality of fresh cut watermelon. Postharvest Biol Technol
55:114–120
Ashokkumar M, Sunartio D, Kentish S, Mawson R, Simons L, Vilkhu K, Versteeg C (2008)
Modification of food ingredients by ultrasound to improve functionality: a preliminary study
on a model system. Innov Food Sci Emerg Technol 9:155–160
Avure Technologies. Inc (2010) Kent, WA (Personal communication)
Barbosa-Cánovas GV, Sepúlveda D (2005) Present status and the future of PEF technology. In:
Barbosa-Cánovas GV, Tapia MS, Cano MP (eds) Novel food processing technologies. CRC
Press, Boca Raton, FL, pp 1–44
Bermúdez-Aguirre D, Corradini MG, Mawson R, Barbosa-Cánovas GV (2009) Modeling the
inactivation of Listeria innocua in raw whole milk treated under thermo-sonication. Innov
Food Sci Emerg Technol 10:172–178
Bialka KL, Demirci A, Puri VM (2008) Modeling the inactivation of Escherichia coli O157:H7
and Salmonella enterica on raspberries and strawberries resulting from exposure to ozone or
pulsed UV-light. J Food Eng 85:444–449
Bizani D, Morrissy JAC, Dominguez APM, Brandelli A (2008) Inhibition of Listeria monocyto-
genes in dairy products using the bacteriocin-like peptide cerein 8A. Int J Food Microbiol
121:229–233
Buchanan RL, Golden MH, Whiting RC, Phillips JG, Smith JL (1994) Non-thermal inactivation
models for Listeria monocytogenes. J Food Sci 59(1):179–188
Buckow R, Truong BQ, Versteeg C (2010) Bovine cathepsin D activity under high pressure. Food
Chem 120:474–481
Buzrul S, Alpas H, Largeteau A, Bozoglu F, Demazeau G (2008) Compression heating of selected
pressure transmitting fluids and liquid foods during high hydrostatic pressure treatment. J Food
Eng 85:466–472
Chua SC, Tan CP, Mirhosseini H, Lai OM, Long K, Baharin BS (2009) Optimization of ultrasound
extraction condition of phospholipids from palm-pressed fiber. J Food Eng 92:403–409
Clark JP (2002) Thermal and nonthermal processing. Food Technol 56(12):63–64
Clark JP (2008) Variety, novelty characterize processing and nonthermal papers. Food Technol
62(5):116–121
Corradini MG, Peleg M (2003) A theoretical note on estimating the number of recoverable spores
from survival curves having an “activation shoulder”. Food Res Int 36:1007–1013
Critzer FJ, Kelly-Wintenberg K, South SL, Golden DA (2007) Atmospheric plasma inactivation of
foodborne pathogens on fresh produce surfaces. J Food Protect 70:2290–2296
Debs-Louka E, Louka N, Abraham G, Chabot V, Allaf K (1999) Effect of compressed carbon
dioxide on microbial cell viability. Appl Environ Microbiol 65:626–631
Deng X, Shi J, Kong MG (2006) Physical mechanisms of inactivation of Bacillus subtilis spores
using cold atmospheric plasmas. IEEE Trans Plasma Sci 34(4):1310–1316
DTI (2010) Diversified Technologies, Inc. Bedford, Massachusetts (Personal communication)
Evrendilek GA, Tok FM, Soylu EM, Soylu S (2008) Inactivation of Penicillum expansum in sour
cherry juice, peach and apricot nectars by pulsed electric fields. Food Microbiol 25:662–667
Ferrentino G, Balaban MO, Ferrari G, Poletto M (2010a) Food treatment with high pressure carbon
dioxide: Saccharomyces cerevisiae inactivation kinetics expressed as a function of CO2
solubility. J Supercrit Fluid 52(1):151–160
Ferrentino G, Barletta D, Balaban MO, Ferrari G, Poletto M (2010b) Measurement and prediction
of CO2 solubility in sodium phosphate monobasic solutions for food treatment with high
pressure carbon dioxide. J Supercrit Fluid 52(1):142–150
Fox MB (2007) Microbial inactivation kinetics of pulsed electric field treatment. In: Lelieveld
HLM, Notermans S, de Haan SWH (eds) Food preservation by pulsed electric fields. Wood-
head Publishing Limited, Cambridge, England, pp 127–137
Furukawa S, Watanabe T, Koyama T, Hirata J, Narisawa N, Ogihara H, Yamasaki M (2009)
Inactivation of food poisoning bacteria and Geobacillus stearothermophilus spores by high
pressure carbon dioxide treatment. Food Control 20:53–58
Gachovska TK, Adedeji AA, Ngadi MO (2009) Influence of pulsed electric field energy on the
damage degree in alfalfa tissue. J Food Eng 95:558–563
Gao YL, Ju XR, Ding W (2007) A predictive model for the influence of food components on
survival of Listeria monocytogenes LM 54004 under high hydrostatic pressure and mild heat
conditions. Int J Food Microbiol 117:287–294
Garcia-Gonzalez L, Geeraerd AH, Elst K, Van Ginneken L, Van Impe JF, Devlieghere F (2009)
Influence of type of microorganism, food ingredients and food properties on high-pressure
carbon dioxide inactivation of microorganisms. Int J Food Microbiol 129:253–263
Garcia-Gonzalez L, Geeraerd AH, Spilimbergo S, Elst K, Van Ginneken L, Debevere J, Van Impe
JF, Devlieghere F (2007) High pressure carbon dioxide inactivation of microorganisms in
foods: the past, the present and the future. Int J Food Microbiol 117:1–28
Gómez N, Garcı́a D, Alvarez I, Raso J, Condón S (2005) A model describing the kinetics of
inactivation of Lactobacillus plantarum in a buffer system of different pH and in orange and
apple juice. J Food Eng 70:7–14
Gómez PL, Alzamora SM, Castro MA, Salvatori DM (2009) Effect of ultraviolet-C light dose on
quality of cut apple: microorganism, color and compression behavior. J Food Eng.
doi:10.1016/j.jfoodeng.2009.12.008
Gómez-Estaca J, López-Caballero ME, Gómez-Guillén MC, López de Lacey A, Montero P (2009)
High pressure technology as a tool to obtain high quality carpaccio and carpaccio-like
products from fish. Innov Food Sci Emerg Technol 10:148–154
Góngora-Nieto M, Sepúlveda DR, Pedrow P, Barbosa-Cánovas GV, Swanson BG (2002) Food
processing by pulsed electric fields: treatment delivery, inactivation level, and regulatory
aspects. Lebensm-WissTechnol 35:375–388
Guerrero S, Tognon M, Alzamora SM (2005) Response of Saccharomyces cerevisiae to the
combined action of ultrasound and low weight chitosan. Food Control 16:131–139
Han Z, Zeng XA, Yu SJ, Zhang BS, Chen XD (2009) Effects of pulsed electric fields (PEF)
treatment on physicochemical properties of potato starch. Innov Food Sci Emerg Technol
10:481–485
Hoogland H, de Haan W (2007) Economic aspects of pulsed electric field treatment of food. In:
Lelieveld HLM, Notermans S, de Haan SWH (eds) Food preservation by pulsed electric fields.
Woodhead Publishing Limited, Cambridge/England, pp 257–265
Houŝka M, Strohalm J, Kocurová K, Totuŝek J, Lefnerová D, Třı́ska J, Vrchotová N, Fiedlerová V,
Holasova M, Gabrovská D, Paulı́čková I (2006) High pressure and foods-fruit and vegetables
juices. J Food Eng 77:386–398
Huang K, Wang J (2009) Designs of pulsed electric fields treatment chambers for liquid foods
pasteurization process: a review. J Food Eng 95:227–239
Ince NH, Belen R (2001) Aqueous phase disinfection with power ultrasound: process kinetics and
effect of solid catalysts. Environ Sci Technol 35(9):1885–1888
Izquierdo E, Marchioni E, Aoude-Werner D, Hasselmann C, Ennahar S (2009) Smearing of soft
cheese with Enteroccus faecium WHE 81. A multi-bacteriocin producer against Listeria
monocytogenes. Food Microbiol 26:16–20
Jambrak AR, Lelas V, Mason TJ, Kreŝič G, Badanjak M (2009) Physical properties of ultrasound
treated soy proteins. J Food Eng 93:386–393
Kempkes M (2009) Personal communication. Pullman, Washington
Khorshid N, Hossain MdM, Farid MM (2007) Precipitation of food protein using high pressure
carbon dioxide. J Food Eng 79:1214–1220
Lado BH, Yousef AE (2002) Alternative food-preservation technologies: efficacy and mechan-
isms. Microb Infect 4:433–440
Lelieveld H (2005) PEF–A Food Industry’s View. In: Barbosa-Cánovas GV, Tapia MS, Cano MP
(eds) Novel food processing technologies. CRC Press, Boca Raton, FL, pp 145–156
Liao H, Kong X, Zhang Z, Liao X, Hu X (2010) Modeling the inactivation of Salmonella
typhimurium by dense phase carbon dioxide in carrot juice. Food Microbiol 27:94–100
Lin HM, Yang ZY, Chen LF (1993) Inactivation of Leuconostoc dextranicum with carbon dioxide
under pressure. Chem Eng J Biochem Eng J 52:B29–B34
López N, Puértolas E, Condón S, Álvarez I, Raso J (2008a) Application of pulsed electric fields for
improving the maceration process during vinification of red wine: influence of grape variety.
Eur Food Res Technol 227:1099–1107
López N, Puértolas E, Condón S, Álvarez I, Raso J (2008b) Effects of pulsed electric fields on the
extraction of phenolic compounds during the fermentation of most of Tempranillo grapes.
Innov Food Sci Emerg Technol 9:477–482
López N, Puértolas E, Condón S, Raso J, Alvarez I (2009a) Enhancement of the extraction of
betanine from red beetroot by pulsed electric fields. J Food Eng 90:60–66
López N, Puértolas E, Condón S, Raso J, Alvarez I (2009b) Enhancement of the solid-liquid
extraction of sucrose from sugar beet (Beta vulgaris) by pulsed electric fields. LWT–Food Sci
Technol 42:1674–1680
Ma YQ, Chen JC, Liu DH, Ye XQ (2009) Simultaneous extraction of phenolic compounds of
citrus peel extracts: effect of ultrasound. Ultrason Sonochem 16:57–62
Margosch D, Ehrmann MA, Buckow R, Heinz V, Vogel RF, G€anzle MG (2006) High-pressure-
mediated survival of Clostridium botulinum and Bacillus amyloliquefaciens endospores at
high temperature. Appl Environ Microbiol 72:3476–3481
Martı́nez Viedma P, Abriouel H, Sobrino-López A, Ben Omar N, Lucas López R, Valdivia E,
Martı́n-Belloso O, Gálvez A (2009) Effect of enterocin AS-48 in combination with high-
intensity pulsed electric field treatment against the spoilage bacterium Lactobacillus diolivor-
ans in apple juice. Food Microbiol 26:491–496
Mastwijk H (2006) Pulsed power systems for application of pulsed electric fields in the food
industry. In: Raso J, Heinz V (eds) Pulsed electric fields technology for the food industry.
Springer, New York, pp 223–238
McKellar RC, Lu X (2004) Primary models. In: McKellar RC, Lu X (eds) Modeling microbial
responses in food. CRC Press, Boca Raton, FL, pp 21–62
National Center for Food Safety and Technology (2009) NFSCT receives regulatory acceptance of
novel food sterilization process. Press release, February 27, 2009. Summit-Argo, IL
Nguyen LT, Tay A, Balasubramaniam VM, Legan JD, Turek EJ, Gupta R (2010) Evaluating the
impact of thermal and pressure treatment in preserving textural quality of selected foods.
LWT–Food Sci Technol 43:525–534
Nguyen TTT, Guyot JP, Icard-Vernière C, Rochette I, Loiseau G (2007) Effect of high pressure
homogenization on the capacity of Lactobacillus plantarum A6 to ferment rice/soybean
slurries to prepare high energy density complementary food. Food Chem 102:1288–1295
Nielsen HB, Sonne AM, Grunert KG, Banati D, Pollák-Tóth A, Lakner Z, Olsen NV, Žontar TP,
Peterman M (2009) Consumer perception of the use of high-pressure processing and pulsed
electric field technologies in food production. Appetite 52:115–126
Niemira BA, Sites J (2008) Cold plasma inactivates Salmonella Stanley and Escherichia coli
O157:H7 inoculated on golden delicious apples. J Food Protect 71(7):1357–1365
Odriozola-Serrano I, Soliva-Fortuny R, Gimeno-Añó V, Martı́n-Belloso O (2008) Modeling
changes in health-related compounds of tomato juice treated by high-intensity pulsed electric
fields. J Food Eng 89:210–216
Odriozola-Serrano I, Soliva-Fortuny R, Hernández-Jover T, Martı́n-Belloso O (2009) Carotenoid
and phenolic profile of tomato juices processed by high intensity pulsed electric fields
compared with conventional thermal treatments. Food Chem 112:258–266
Ogihara H, Yatuzuka M, Horie N, Furukawa S, Yamasaki M (2009) Synergistic effect of high
hydrostatic pressure treatment and food additives on the inactivation of Salmonella enteritidis.
Food Control 20:963–966
Oms-Oliu G, Odriozola-Serrano I, Soliva-Fortuny R, Martı́n-Belloso O (2009) Effects of high-
intensity pulsed electric field processing conditions on lycopene, vitamin C and antioxidant
capacity of watermelon juice. Food Chem 115:1312–1319
Pagán R, Mañas P, Raso J, Condón S (1999) Bacterial resistance to ultrasonic waves under
pressure (manosonication) and lethal (manothermosonication) temperatures. Appl Environ
Microbiol 65(1):297–300
Parton T, Bertucco A, Elvassore N, Grimolizzi L (2007a) A continuous plant for food preservation
by high pressure CO2. J Food Eng 79:1410–1417
Parton T, Elvassore N, Bertucco A, Bertoloni G (2007b) High pressure CO2 inactivation of food: a
multi-batch reactor system for inactivation kinetic determination. J Supercrit Fluid 40:490–496
Patazca E, Koutchma T, Balasubramaniam VM (2007) Quasi-adiabatic temperature increase
during high pressure processing of selected foods. J Food Eng 80:199–205
Patil S, Bourke P, Kelly B, Frı́as JM, Cullen PJ (2009) The effects of acid adaptation on
Escherichia coli inactivation using power ultrasound. Innov Food Sci Emerg Technol
10:486–490
Patist A, Bates D (2008) Ultrasonic innovations in the food industry: from the laboratory to
commercial applications. Innov Food Sci Emerg Technol 9:147–154
Peleg M (2000) Microbial survival curves – the reality of flat “shoulders” and absolute thermal
death times. Food Res Int 33:531–538
Peleg M, Cole M (1998) Reinterpretation of microbial survival curves. Crit Rev Food Sci
38(5):353–380
Peleg, M., Normand, M.D., and Campanella, O.H (2003) Estimating microbial inactivation
parameters from survival curves obtained under varying conditions – the linear case. Bull
Math Biol 65:219–234
Peleg M, Normand MD, Damrau E (1997) Mathematical interpretation of dose-response curves.
Bull Math Biol 59(4):747–761
Peleg M, Penchina CM (2000) Modeling microbial survival during exposure to a lethal agent with
varying intensity. Crit Rev Food Sci 40(2):159–172
Perni S, Liu DW, Shama G, Kong MG (2008a) Cold atmospheric plasma decontamination of the
pericarps of fruit. J Food Protect 71(2):302–308
Perni S, Shama G, Kong MG (2008b) Cold atmospheric plasma disinfection of cut fruit surfaces
contaminated with migrating microorganisms. J Food Protect 71(8):1619–1625
Pinto AL, Fernandes M, Pinto C, Albano H, Castilho F, Teixeira P, Gibbs P (2009) Characteriza-
tion of anti-Listeria bactericoins isolated from shellfish: potential antimicrobials to control
non-fermented seafood. Int J Food Microbiol 129:50–58
Puértolas E, López N, Condón S, Raso J, Alvarez I (2009) Pulsed electric fields inactivation of
wine spoilage yeast and bacteria. Int J Food Microbiol 130:49–55
Puértolas E, López N, Saldaña G, Alvarez I, Raso J (2010) Evaluation of phenolic extraction
during fermentation of red grapes treated by a continuous pulsed electric fields process at pilot-
plant scale. J Food Eng. doi:10.1016/j.foodeng.2009.12.017
Rajan S, Ahn J, Balasubramaniam VM, Yousef AE (2006) Combined pressure-thermal inactiva-
tion kinetics of Bacillus amyloliquefaciens spores in egg patty mince. J Food Protect
69(4):853–860
Rastogi NK, Raghavarao SMS, Balasubramaniam VM, Niranjan K, Knorr D (2007) Opportunities
and challenges in high pressure processing of foods. Crit Rev Food Sci 47:69–112
Reyns KMFA, Soontjens CCF, Cornelis K, Weemaes CA, Hendrickx ME, Michiels CW (2000)
Kinetic analysis and modeling of combined high-pressure-temperature inactivation of the yeast
Zygosaccharomyces bailii. Int J Food Microbiol 56:199–210
Riener J, Noci F, Cronin DA, Morgan DJ, Lyng JG (2008) Combined effect of temperature and
pulsed electric fields on apple juice peroxidase and polyphenoloxidase inactivation. Food
Chem 109:402–407
Rodrigo D, Barbosa-Cánovas GV, Martı́nez A, Rodrigo M (2003) Weibull distribution function on
an empirical mathematical model for inactivation of Escherichia coli by Pulsed electric fields.
J Food Protect 66(6):1007–1012
Rodrigo D, Martı́nez A, Harte F, Barbosa-Cánovas GV, Rodrigo M (2001) Study of inactivation of
Lactobacillus plantarum in orange-carrot juice by means of pulse electric fields: comparison of
inactivation kinetics models. J Food Protect 64(2):259–263
Saucedo-Reyes D, Marco-Celdrán A, Pina-Pérez MC, Rodrigo D, Martı́nez-López A (2009)
Modeling survival of High Hydrostatic Pressure treated stationary- and exponential- phase
Listeria innocua cells. Innov Food Sci Emerg Technol 10:135–141
Selcuk M, Oksuz L, Basaran P (2008) Decontamination of grains and legumes infected with
Aspergillus spp. and Penicillum spp. by cold plasma treatment. Bioresource Technol
99:5104–5109
Selma MV, Allende A, López-Gálvez F, Conesa MA, Gil MI (2008) Disinfection potential of
ozone, ultraviolet-C and their combination in wash water for the fresh cut vegetable industry.
Sloan E (2010) What, when, and where America Eats. Food Technol 64(1):34–40
Soliva-Fortuny R, Balasa A, Knorr D, Martı́n-Belloso O (2009) Effects of pulsed electric fields on
bioactive compounds in foods: a review. Trends Food Sci Technol 20:544–556
Tabilo-Munizaga G, Barbosa-Cánovas GV (2004) Color and textural parameters of pressurized
and heat-treated surimi gels as affected by potato starch and egg white. Food Res Int
37(8):767–775
Tabilo-Munizaga G, Barbosa-Cánovas GV (2005) Pressurized and heat-treated surimi gels as
affected by potato starch and egg white:microstructure and water-holding capacity. Lebensm-
WissTechnol 38(1):47–57
Taniguchi M, Ishiyama Y, Takata T, Nakanishi T, Kaneoke M, Watanabe K, Yanagida F, Chen Y,
Kouya T, Tanaka T (2009) Growth-inhibition of hiochi bacteria in namazake (raw sake) by
bacteriocins from lactic acid bacteria. J Biosci Bioeng. doi:10.1016/j.jbiosc.2009.11.015
Tiwari BK, Patras A, Brunton N, Cullen PJ, ÓDonell CP (2010) Effect of ultrasound processing in
anthocyanins and color or red grape juice. Ultrason Sonochem 17:598–604
Torres JA, Velazquez G (2005) Commercial opportunities and research challenges in the high
pressure processing of foods. J Food Eng 67:95–112
Ugarte-Romero E, Feng H, Martin SE, Cadwallader KR, Robinson SJ (2006) Inactivation of
Escherichia coli with power ultrasound in apple cider. J Food Sci 71(2):E102–E108
Valdramidis VP, Cullen PJ, Tiwari BK, ÓDonnell CP (2010) Quantitative modeling approaches
for ascorbic acid degradation and non-enzymatic browning of orange juice during ultrasound
processing. J Food Eng 96:449–454
Vilkhu K, Mawson R, Simons L, Bates D (2008) Applications and opportunities for ultrasound
assisted extraction in the food industry – a review. Innov Food Sci Emerg Technol 9:161–169
Virot M, Tomao V, Le Bourvellec C, Renard CMCG, Chemat F (2009) Towards the industrial
production of antioxidants from food processing by-products with ultrasound-assisted extrac-
tion. Ultrason Sonochem. doi:10.1016/j.ultsonch.2009.10.015
Wilson DR, Dabrowski L, Stringer S, Moezelaar R, Brocklehurst TF (2008) High pressure in
combination with elevated temperature as a method for the sterilization of food. Trends Food
Sci Technol 19:289–299
Yu H, Perni S, Shi JJ, Wang DZ, Kong MG, Shama G (2006) Effects of cell surface loading and
phase of growth in cold atmospheric gas plasma inactivation of Escherichia coli K12. J Appl
Microbiol 101:1323–1330
.
Chapter 14
High-Pressure-Induced Effects on Bacterial
Spores, Vegetative Microorganisms,
and Enzymes
Dietrich Knorr, Kai Reineke, Alexander Mathys, Volker Heinz,

and Roman Buckow
14.1 Introduction
Pressures currently used in the food industry range from tens of MPa in common
homogenizers or supercritical fluid extractors to up to 400 or 800 MPa in ultra high
pressure homogenizers or high pressure (HP) pasteurization units, respectively.
Laboratory sized HP research units can reach up to 1,400 MPa and temperatures
up to 200 C (Reineke et al. 2008). These HP units are principally used for the
inactivation of vegetative microorganisms to extend shelf life of the treated food.
However, there are numerous other interesting food applications for HP such as
food structure engineering (Knorr 2002; Rumpold 2005; Diels and Michiels 2006;
Knorr et al. 2006; Sharma and Yadav 2008), enhanced food quality (Ludikhuyze
et al. 2002b; Trejo Araya et al. 2007), stress response utilization (Ananta and Knorr
2004; Bothun et al. 2004; Kato et al. 2007; Pavlovic et al. 2008), or control of
bioconversion reactions (Knorr et al. 2006; Picard et al. 2006).
D. Knorr (*) and K. Reineke

Department of Food Biotechnology and Food Process Engineering, Berlin University of
Technology, Koenigin-Luise-Str. 22, D-14195 Berlin, Germany
e-mail: dietrich.knorr@tu-berlin.de, k.reineke@tu-berlin.de
A. Mathys
Department of Food Biotechnology and Food Process Engineering, Berlin University of
Technology, Koenigin-Luise-Str. 22, D-14195 Berlin, Germany
and
Nestlé Research Center, Food Science and Technology Department, Vers-chez-les-Blanc, P.O.
Box 44, CH-1000 Lausanne 26, Switzerland
e-mail: Alexander.Mathys@rdls.nestle.com
V. Heinz
German Institute of Food Technology, Prof.-von-Klitzing-Str. 7, D-49601 Quakenbrueck,
Germany
e-mail: V.Heinz@dil-ev.de
R. Buckow
CSIRO Food and Nutritional Sciences, 671 Sneydes Road, Werribee VIC-3030, Australia
e-mail: Roman.Buckow@csiro.au
326 D. Knorr et al.
The majority of isostatic HP application in the food industry is the pasteurization

of foods at ambient temperature. This principle was first reported by Hite (1899),
who achieved an extended shelf life of bovine milk after HP treatment. Further-
more, Bridgman (1914) reported that HP could coagulate egg albumin producing
gels different from those obtained by heat coagulation. In addition, he presented an
extensive dataset for the phase diagram of pure water under pressure (Bridgman
1912). Since the early 1980s, growing consumer demand for minimally processed,
fresh-like, safe, and high quality food products has triggered research efforts in the
field of alternative food processing technologies (Hendrickx and Knorr 2002). The
application of HP has been evaluated as a promising food processing alternative to
classical heat treatment technologies in several studies (Heinz and Knorr 1998) and
consequently, a first industrial HP application for the commercial preservation of
food was installed in Japan in 1991 (Yaldagard et al. 2008). Because of further
extensive research, HP technology was extended to a broad range of products and
the number of industrial HP systems for food pasteurization has steadily increased
during the past 10 years.
At present, there are more than 156 industrial scale installations with a maximum
volume of 687 L in use worldwide with an annual production of 300,000 t pasteurized
products (Tonello Samson 2010, NC Hyperbaric, Spain, personal communication).
In order to inactivate bacterial and fungal spores, a combination of HP and
elevated initial temperature (>80 C) seems to be a promising improvement to
traditional heat sterilization. Combined HP and heat treatments can result in a
sterilized food product with reduced thermal load and consequently higher nutri-
tional quality and functionalities (Heinz and Knorr 2002).
Extensive research in this field was carried out during the last decade by various
research institutes such as the University of Connecticut, the Technische Uni-
versit€at Berlin, the University of Delaware, CSIRO Food and Nutritional Science
Australia, the Ohio State University, and the National Center for Food Safety and
Technology. These research efforts enabled the Food and Drug Administration
(FDA) of the United States to certify a pressure assisted thermal sterilization
(PATS) process in February 2009. However, this promising sterilization technology
has not been implemented on an industrial scale, yet.
14.2 High Pressure Thermal Sterilization
The advantage of isostatic HP is the instantaneous and uniform application of pressure

to the product without any delay. During pressure build-up, adiabatic heating triggers a
thermo-physical effect leading to a temperature increase of the pressure transmitting
medium and the product. Ambient preheating combined with the compression heating
during pressure build-up allows the food to reach sterilization conditions within a short
time. During the dwell time, the denaturation of proteins (e.g., enzymes) and a
significant reduction of microorganisms were observed, without affecting molecular
bonds or inhibiting some chemical reactions. For example, Maillard reactions, which
14 High-Pressure-Induced Effects on Bacterial Spores 327
can produce off-flavors or the destruction of vitamins can be reduced under HP

conditions (Heremans 2002; Indrawati et al. 2005).
In addition to enhanced product quality and functionality, the HP sterilization
process needs to ensure the microbial safety. Currently, the commercial sterility of
low-acid (pH > 4.5) canned foods is often realized by conventional heat proces-
sing, in order to achieve a non-refrigerated, shelf-stable product. Sterilization of
food products includes the inactivation of yeasts, fungi, viruses, vegetative micro-
bial cells, and spores. Bacterial endospores require special attention, because of
their high resistance to wet and dry heat, freezing and thawing, chemical agents,
disinfectants, irradiation, and pressure (Setlow 2003). This high resistance is
attributed to a number of unique features of the dormant spore: the low spore
core water content, in which water is supposed to be free but highly viscous
and incorporated into a three-dimensional molecular matrix in dormant spores
(De Vries 2006); the thick peptidoglycan layer called the cortex; the low perme-
ability of the inner spore membrane to hydrophilic molecules; the high level of
dipicolinic acid (DPA) as well as high levels of minerals (e.g., Ca2þ) (Setlow 2003).
Long treatment times required to achieve “commercial” sterility cause unwanted
chemical and physical changes of the food. Consequently, intensive research on the
HP-combined thermal inactivation of microbial spores has been carried out in the
past. During the last several decades, two different approaches for the inactivation
of bacterial spores were investigated. One is the pressure-induced (150 MPa < p
< 400 MPa) germination of bacterial endospores (Gould and Sale 1970) at ambient
temperature, with a subsequent mild thermal treatment to inactivate the germinated
and hence temperature sensitive spores (Heinz and Knorr 1998; Wuytack et al.
1998; Paidhungat et al. 2002). A disadvantage of this procedure is, that spores in the
super dormant state are not activated by pressure (Margosch et al. 2004) and even
repeated cycles of pressure and heat treatment were applied (Hayakawa et al.
1994b; Meyer 2000), but failed to achieve commercial food sterility. Another
approach to HP food sterilization is the combination of HP with high temperatures.
This combination of HP at elevated temperatures (70–100 C) inactivates most of
the spores that normally survive heat pasteurization processes (Ananta et al. 2001;
Ardia et al. 2004b; Margosch et al. 2006; Mathys et al. 2009).
14.2.1 Development and Application of Temperature

Controlled Spore Inactivation
In contrast to conventional heat processing, where convective heating causes signif-

icant temperature gradients in the product, all volume elements in an HP vessel are
subjected to the same pressure at the same time due to the isostatic principle of
pressure transmission (Ting et al. 2002; Rastogi et al. 2007; Heinz et al. 2009).
Since pressure and temperature are closely connected physical parameters,
the thermodynamic effect of the adiabatic heating, which occurs during the com-
pression and decompression of the treated food and the pressure transmitting media,
328 D. Knorr et al.
has to be taken into account. Following the first and the second law of thermody-
namics and a rearrangement of the Maxwell equations, heating during the compres-
sion and cooling during the decompression can be described as a function of
thermo-physical properties of the compressible product (Perry 1984; Ardia et al.
2004; Reineke et al. 2008). This quasi adiabatic heating or cooling occurs instantly.
Hence, pressure-induced temperature changes are predictable and relatively
uniform within the food product. Thus, compression heating and decompression
cooling may be utilized to achieve a rapid and uniform heating and cooling of the
product. Typically, an ideal adiabatic process does not occur in practical applica-
tions, but the extent of temperature increase could be estimated within 3–9 C per
100 MPa depending on the thermo-physical properties of the food (Ting et al.
2002). Based on the assumption that the heat resistance of bacterial spores under HP
conditions is similar to or lower than at atmospheric pressure, the adiabatic heat of
compression could be used to reduce the heating and cooling times of a thermal
sterilization process (De Heij et al. 2003).
High pressure thermal sterilization (HPTS) has not yet been successfully applied
in the food industry, possibly because of the lack of knowledge on the involved
inactivation mechanisms of bacterial spores. Hence, new methodologies for a
detailed investigation of the heat and pressure effects on bacterial spores are
required. Reineke et al. (2008) described an innovative micro HP unit that is
capable of reaching isothermal and isobaric conditions in the range of 50–130 C
and up to 1,400 MPa within a few seconds (Fig. 14.1). This is advantageous for the
investigation of inactivation kinetics, e.g., of bacterial spores or enzymes, under
isothermal/isobaric conditions, which could possibly allow a deeper insight into the
underlying inactivation mechanisms. The high compression rates were achieved by
Fig. 14.1 Stansted Mini Foodlab FBG 5620 high pressure unit with heating/cooling system
combining a gas-filled pressure accumulator and an internal intensifier, which enabled

a maximum compression rate of 250 MPa s1. The installed heating/cooling system
allowed a maximum heating rate of 6 K s1 and was controlled by a proportional-
integral-derivate (PID) controller, which was implemented into the control software.
The control software facilitated a nearly ideal adiabatic pressure build-up phase and
isothermal dwell times by calculation of the adiabatic heating of water at the
corresponding pressure and preheating conditions. Inactivation kinetics of Geobacil-
lus stearothermophilus spores, a commercial sterilization indicator, were investigated
in the pressure-temperature range of 600–1200 MPa and 90–120 C in a pressure
stable ACES-buffer solution (pH 7) (Mathys 2008). Under dynamic pressure-tem-
perature conditions during the compression and decompression phase (1 s dwell
time), a 1–3 log10 reduction of viable spores was observed. During isothermal and
isobaric dwell times, microbial inactivation followed a nearly linear log10-inactiva-
tion behavior. This dataset covers the combined effects of pressure and temperature
on the inactivation of Geobacillus stearothermophilus spores and provides the basis
for generating a p-T diagram for spore inactivation with isorate lines for a fixed spore
inactivation. The nth-order approach has proven to be suitable to describe inactiva-
tion kinetics with tailing, which is often found for the inactivation of vegetative
microorganisms, bacterial spores, and enzymes during HP treatment.
Based on data obtained for isobaric and isothermal treatment conditions, an
empirical model, which is an analog to the nth-order approach for chemical degra-
dation reactions, was used:
N 1
ðn1Þ
log ¼ log 1 þ k N0 t ðn 1Þ (14.1)
N0 1 n
where N and N0 denotes the spore counts (CFU mL1) at time t and t0, respectively;
t ¼ time, k ¼ inactivation rate constant, and n ¼ reaction order. Analysis of the
inactivation kinetics revealed that all experimental data could be fitted best with a
fixed reaction order of n ¼ 1.05.
To obtain a functional relationship of the rate constant k with pressure and
temperature, empirical equations (e.g., Eq. 14.2) have often been suggested for
bacterial spore inactivation data (Ardia et al. 2004; Margosch et al. 2006):
lnðkÞ ¼ A þ Bðp p0 Þ þ CðT T0 Þ þ Dðp p0 Þ2 þ EðT T0 Þ2

(14.2)
þ Fðp p0 ÞðT T0 Þ þ higher order terms
Values for the parameters A-J for Geobacillus stearothermophilus spores

(Mathys 2008) were estimated as:
A ¼ 96:71; B ¼ 0:02; C ¼ 2:65; D ¼ 2:92; E ¼ 0:02; F ¼ 9:45 104 ;

G ¼ 1:25 108 ; H ¼ 9:33 105 ; I ¼ 3:87 106 ; J ¼ 6:52 108
330 D. Knorr et al.
Using Eq. (14.2), pressure-temperature isorate lines of spore inactivation can be

calculated for any pressure dwell time and facilitate the presentation of the survival
data in a pressure and temperature diagram. Based on the assumption of a nearly
linear log10 reduction under isothermal/isobaric conditions, in addition to the nth
order (n ¼ 1.05), a traditional first-order inactivation (n ¼ 1) was modeled in the
p-T landscape (Fig. 14.2a). This first order p-T diagram allows the calculation of
increased log10 reduction, just by multiplying the isorate lines by the same value.
Inactivation studies showed the highest heat sensitivity of G. stearothermophilus
spores at approximately 700 MPa, whereas higher pressures caused a stabilization
of the spores against inactivation at the same treatment temperature (Fig. 14.2a).
Similar results were found for Bacillus amyloliquefaciens (Margosch et al. 2006;
Rajan et al. 2006) and other strains of G. stearothermophilus spores (Ardia 2004).
Ardia (2004) investigated the impact of the dissociation equilibrium shift in
ACES and PBS buffer on the inactivation of bacterial spores and concluded that the
equilibrium shift is not the main source of the observed heat stabilization of
G. stearothermophilus spores under pressure. The author compared the behavior
of cortex lytic enzymes (CLEs) in the spore coat under HP with the behavior of
bacterial spores. CLEs catalyze the cortex break down and therefore are responsible
for the access of water to the spore core. It was assumed, that the CLEs are
inactivated in a certain range of pressure-temperature conditions. At this stage,
only a further temperature increase accelerates spore inactivation, and consequently
a further pressure increase would have no effect. With inactive CLEs, the spore
cortex can only be rehydrated by ultra high pressure, and hence pressures higher
than 800 MPa could squeeze water inside of the spore cortex (Ardia 2004) causing
an pressure generated increased heat sensitivity of the spore. On the contrary,
Margosch et al. (2006) proposed that bacterial spores are very complex, and that
a 1200
b
4 log10
Lethal effects (T, t=const.)
1000
VI
Pressure [MPa]
800 Σ Effects
IC
Cortex hydrolysis
+
1 min
full core hyreation
600 2 min via
3 min active CLEs Partial core
400 4 min rehydration
5 min without CLEs
8 min
200 10 min 10 5 1 Opening
DPA channels
0
0 20 40 60 80 100 120 140 400 600 800 1000 1200 1400
Temperature [°C] Pressure [MPa]
Fig. 14.2 (a) Isorate lines for a 4 log10 inactivation of Geobacillus stearothermophilus spores in
pressure stable 0.05 M ACES buffer (pH 7) after 1–10 min treatment at isothermal/isobaric
conditions with N (1 s) as initial population N0. Data for thermal inactivation are shown as arrows
with holding times in minutes. Dashed lines indicate the adiabatic heating of water. (b) Suggested
summed lethal effects at different pressure levels with two generated lethal effect distributions,
A (green dashes) and B (blue dots), showing different germination P reactions at given process
intensity. All lethal effects resulted in the cumulative distribution ( Effects) with a sensitive and
stabilized zone (Mathys et al. 2009)
a single protein or enzyme can only be responsible for stabilization under pressure,
if it is required for a vital function of the vegetative cell (e.g., as a structural or
protective component).
Based on Setlow’s (2003) germination model for Bacillus subtilis spores, an
extended inactivation mechanism of G. stearothermophilus spores at different
pressure-temperature conditions was proposed by Mathys et al. (2009). The mech-
anism includes a stable and sensitive domain for the spores in the pressure and
temperature landscape (Fig. 14.2). The germination mechanism is initiated by an
opening of the dipicolinic acid (DPA) channels at pressures higher than 500 MPa,
which enables spore germination in the absence of nutrients (Paidhungat et al.
2002). At these HP conditions, the spore germination is limited in buffer solutions.
The key factor for full cortex hydrolysis and core hydration is the activation of the
main CLEs. These enzymes probably show different activities or stabilities under
specific pressure and temperature conditions, generating different pathways of
inactivation. Heinz and Knorr (1998) assumed that pressure and temperature
initially triggers the CLEs, but inactivates the same enzymes at a later stage.
Hence, the CLE activity is dependent on pressure, temperature, and time. At ultra
HPs a partial core rehydration without any CLE activity could induce an additional
lethal effect (Setlow 2003). Summarizing all competing reactions, a cumulative
lethal
P effect distribution as a function of the applied pressure level can be generated
( Effects, Fig. 14.2b). However, the mechanisms of spore inactivation at HPTS
are still not fully understood, and further research is needed to identify the compo-
nents in bacterial endospores, that are most pressure sensitive and hence provide a
fast and complete inactivation by HPTS.
14.2.2 Industrial Relevance and Applications
Currently, industrial HPTS equipment exists with volumes up to 150 L, 700 MPa
maximum working pressure, and initial temperatures of up to 95 C (Heinz 2010).
The industrial relevance of this study can be exemplified by implementing the
generated data into process charts of an industrial scale HPTS unit (Fig. 14.3). One
of the few existing pilot systems is the Flow Pressure Systems QUINTUS (Food
Press Type; 35 L, 600 MPa) sterilization machine (Avure Technologies, Kent, WA,
USA), which is described elsewhere in the literature (Knoerzer et al. 2007). In
Fig. 14.3a and 14.3b (location two is near top closure; Knoerzer et al. 2007), two
different pressure and temperature profiles from an HPTS process with (10 L) and
without (35 L) a polytetrafluoroethylene (PTFE) carrier are shown. Temperature
decrease and inhomogeneities during the dwell time can be reduced, when an
insulating carrier is used (Juliano et al. 2009). Alternative possibilities to ensure a
homogenous temperature during dwell time would be the application of an internal
heater or heating of the vessel wall to an appropriate temperature level, that
minimizes the temperature gradient between pressure vessel and treated product
(Ardia 2004). In Fig. 14.3b the target inactivation level of 4 log10 can be varied,
332 D. Knorr et al.
a 1200 140 b 1200

Pressure PTFE insulated –4 log10 54 3 2 1
1000 non-insulated 130 1000
Temperature [°C]
Pressure [MPa]
Pressure [MPa]
VI
E
800 120 800
IC
600 110 600
400 100 400
200 90 200 10 5 1
0 80 0
0 1 2 3 4 5 6 7 8 0 20 40 60 80 100 120 140
Treatment time [min] Temperature [°C]
c 8
Σ F-value at 600 MPa [min]
7 PTFE insulated
non-insulated
6
5
4
3 12 D
2 8D
1 4D
0
0 1 2 3 4 5 6 7 8
Treatment time [min]
Fig. 14.3 Industrial process analysis of a vertical 35 L vessel (--, black) and with a 10 L insulated
carrier (–, red), using an F-value (c) (14.3), Tref ¼ 121.1 C, z600MPa ¼ 35.36 C, D121.1 C ¼ 13.52
s at 600 MPa in 0.05 M ACES buffer (pH 7), and pressure holding times T(t)p. Pressure and
temperature profiles (a) were taken from the literature (Knoerzer et al. 2007). The desired
inactivation level of (b) can be varied by multiplying the dwell times of the isorate lines with
the same value (Mathys et al. 2009)
because of a first-order kinetic approach. Process conditions were as follows: 90 C

initial temperature, 600 MPa final process pressure, and 285 s dwell time with water
as pressure transmitting medium (Knoerzer et al. 2007). Using a modified F121.1 C -
value concept for the thermal inactivation of G. stearothermophilus spores
(Eq. 14.3) in Fig. 14.3c,
ðt
TðtÞp Tref
dt¼Dref log10 N
F¼
Z600MPa N0
10 (14.3)
0
where reference temperature Tref is 121.1 C, T(t)p is the dwell time; z600MPa
¼ 35.36 C, D121.1 C ¼ 13.52 s at 600 MPa; N0 is the initial count, and N is the
survival count. Both processes, with or without insulating carrier, can be adequately
analyzed and compared.
In thermal processes the F-value is the accumulated lethality expressed as an
equivalent time at a specific reference temperature (Tref) with a specific z-value.
The F0 concept is accepted for the thermal sterilization in the food industry for the
case where Tref is 121.1 C, the D121.1 C – value is 13.8 s (for Clostridium botulinum),
and z-value is approximately 10 C (Stone et al. 2009). In contrast to conventional
thermal processes at ambient pressure with an expected z0.1MPa ¼ 10 C for spore
inactivation, the z-value at 600 MPa (z600MPa= 35.36 C for G. sterothermophilus)
was higher (Mathys et al. 2009), resulting in a lower temperature dependence of the
inactivation reaction rate. Because of the pressure dependence of the z-value, pre-
heating as well as compression and decompression phases could not be included in
the calculation of the spore inactivation level of G. sterothermophilus. However, the
additional processing time can be considered a safety factor. Using a PTFE carrier,
commercial sterility (>12 D) could be achieved within 3 min dwell time at 600 MPa
and 90 C initial temperature (Mathys 2008). An alternative process for a HPTS unit
with a 55 L horizontal vessel, 700 MPa working pressure, and 95 C initial tempera-
ture was also calculated by Mathys et al. (2008) (data not shown). He concluded that
more than 12 log10 inactivation of G. stearothermophilus spores can be achieved in
this machine without any insulation within 3 min at 700 MPa.
14.3 HP Effects on Vegetative Microorganisms and Enzymes
The specific effects of pressure on vegetative microorganisms are complex and

cannot be evaluated separate from simultaneous occurring heat effects. Primarily,
the lethal effects of HP on vegetative microorganisms are attributed to the inacti-
vation of essential cell enzymes and the rupture of the cell membrane (Ardia 2004;
Ananta 2005). In the course of identifying mechanisms behind this pressure-
induced inactivation, it was found that flow cytometry is a powerful tool to gain
insight into the states and mechanisms of cell damage of pressurized vegetative
microorganisms and bacterial spores (Ananta 2005; Black et al. 2006; Mathys
et al. 2007). However, pressure treatments do not necessarily weaken biological
cells. At low pressure levels, stabilization of microbial cells was observed; for
example, pressure-induced thermo-tolerance of lactic acid bacteria occurs after
HP treatment between 100 and 200 MPa (Ananta and Knorr 2003). As a result of
this phenomenon, pressure-induced stress response was regarded as a promising
processing option, such as pretreatment of bacteria before spray drying or freez-
ing for the purpose of starter culture production.
Microbial inactivation of more than five log cycles in food products has been
reported by several authors to occur at pressures between 300 and 800 MPa
(Henrickx and Knorr 2002; Ananta et al. 2005; Lori et al. 2007). The special
shape of isorate lines in the HP-temperature landscape (Fig. 14.4), especially for
the inactivation of microorganisms, shows synergisms between pressure and tem-
perature. This behavior is typical for vegetative cells, but was also observed for
bacterial spores (Heinz and Knorr 2002; Ardia 2004; Margosch et al. 2006), viruses
(Isbarn et al. 2007; Buckow et al. 2008), and proteins (Heinz and Kortschack 2002;
Smeller 2002) (Fig. 14.4). Increasing the process temperature allows the decrease
334 D. Knorr et al.
Fig. 14.4 Pressure-temperature isorate diagram showing a 5 log10 reduction of microorganisms,

prions, and viruses after a 4 min isothermal/isobaric treatment, with dashed lines (--) indicating
adiabatic heating of water; for H7N7, surrogate for bird flu virus H5N1 (in chicken meat slurry)
(Isbarn et al. 2007), Listeria monocytogenes 75903 (in ham slurry) (Lori 2008), prions-PrPSC (in
raw meat) (Heinz and Kortschack 2002), and Clostridium botulinum spores (in TRIS buffer pH
5.15) (Margosch et al. 2006). The blue line represents the freezing line between liquid water and
the different pressure-dependent ice modifications
of the applied pressure to achieve similar inactivation rates. However, unwanted

reactions also occur at moderate temperatures and elevated pressures, for example
the activation or insufficient inactivation of quality degrading enzymes, have to be
taken into account in these cases.
According to Smelt et al. (2001), the HP induced effects, resulting in the death of
vegetative cells, can be summarized as follows:
– Proteins and enzymes: HP induces unfolding of globular proteins. It is assumed
that the combined, complete or partial inactivation of numerous enzymes and
metabolic pathways leads to the inability to proliferate or cell death, respec-
tively (Bunthof 2002).
– Membranes: Membrane damage is considered as one of the key events related to
microbial cell death. Membranes undergo phase transitions and solidify under
pressure and perturbations are promoted (Schlueter et al. 2004). In addition,
pressure can lead to the detachment or inactivation of membrane proteins
(Ulmer et al. 2002).
– Ribosomes: The disintegration of ribosomes in their subunits is promoted by
pressure and may cause cell death (Niven et al. 1999).
– pH: The maintenance of intracellular pH is crucial for the survival of microbial
cells. Some authors attribute cell death predominantly to intracellular pH
changes, which are related to the inactivation of enzymes controlling the cell
acidity (Molina-Gutierrez et al. 2002).
Given that HP is regarded as a mild process, the primary structure of proteins is

not affected. However, pressure can have a deep impact on hydrophobic interac-
tions, which stabilize the quaternary, tertiary, and secondary structure through
reversible or irreversible unfolding, respectively. Generally, pressure induces
reversible changes in proteins and enzymes between 100 and 300 MPa near room
temperature, whereas pressure above 400 MPa can lead to an irreversible unfolding
of a protein and thus an inactivation of enzymes. Pressure also favors the unfolding
of protein chains as well as a dissociation of oligomeric proteins (Tauscher 1995;
Buckow 2006).
Due to conformational changes, unfolding of an enzyme can alter its functional-
ity, resulting in a decreased or increased catalytic activity or change in substrate
specificity (Ludikhuyze et al. 2002a; Buckow and Heinz 2008). The pressure
stability of enzymes can vary drastically, ranging from pressure sensitive enzymes,
such as phosphohexoseisomerase from bovine milk (p < 400 MPa) (Rademacher
and Hinrichs 2006), to extreme pressure resistant enzymes like, peroxidise from
horseradish (Smeller 2002). However, categorization of enzymes as a result of their
pressure stability is not appropriate, since there is a structural manifoldness among
enzymes catalyzing the same reaction (Buckow and Heinz 2008).
Figure 14.5a depicts the inactivation of 90% of different polyphenol oxidases
(PPO) after 10 min isothermal/isobaric treatment. The pressure-temperature resis-
tance of enzymes show a significant dependency on matrix conditions, such as the
pH-value (Zipp and Kauzmann 1973; Weemaes et al. 1997; Riahi and Ramaswamy
2004) or the presence of different ions (Buckow et al. 2007a). Furthermore, even
Fig. 14.5 Pressure-temperature isorate diagram for: (a) 90% inactivation of apple polyphenol
oxidase after 10-min isothermal/isobaric treatment (Buckow et al. 2009), avocado (Weemaes et al.
1998), white grapes (Rapeanu et al. 2005), and strawberry (Dalmadi et al. 2006); (b) 95%
inactivation of b-amylase (Heinz et al. 2005), b-glucanase (barley malt) (Buckow et al. 2005b),
and b-glucanase (B. subtilis) (Buckow et al. 2007b) in ACES buffer (pH 5.6; 0.1 M), a-amylase in
ACES buffer (pH 5.6; 0.1 M, containing 90 mM NaCl and 3.8 mM CaCl2) (Buckow et al. 2007a),
and glucoamylase isoenzymes GA1 and GA2 in ACES buffer (pH 4.5; 0.1 M) (Buckow et al.
2005a) after 30-min isothermal/isobaric treatment
336 D. Knorr et al.
isoforms of an enzyme from the same origin (e.g., glucoamlyse GA1 and GA2, as
shown in Fig. 14.5b) can vary in their physical stability up to several hundred MPa
(Buckow et al. 2005a; Buckow 2006; Rodrigo et al. 2006).
Pressure and temperature often act antagonistically on protein systems in the
high temperature domain (Fig. 14.5), which can result in an enhanced thermo-
stability of enzymes at specific pressures (Heremans and Smeller 1998; Lori et al.
2007). Such stabilization of enzymes occurs when the volume difference between
the folded and unfolded state of the protein is positive, which might be due to the
promoted formation of non-covalent bonds under pressure conditions.
14.4 Outlook, Needs, and Challenges
The continuously increasing HP research of the last decades has already generated
an impressive number of commercially available HP processed, high quality pro-
ducts. Besides the “cold” pasteurization, which mainly inactivates vegetative
microorganisms, there is still a lack of knowledge regarding the process conditions
that are necessary to inactivate pressure-resistant bacterial endospores. Stable
matrix (e.g., pressure stable buffer solutions) as well as defined treatment condi-
tions (e.g., isothermal and isobaric conditions during dwell time) are essential, to
generate reliable models, that predict microbial and enzyme inactivation and to
obtain a better understanding of the underlying mechanisms (Mathys et al. 2009).
This approach, in combination with a suitable pressure and temperature resistant
surrogate, will allow a clearer insight into the mechanisms leading to spore inacti-
vation under pressure and would consequently be the next step to successfully
introduce HPTS in the food industry. The reduction of relatively high processing
costs of HPTS; the investigation of the temperature evolution and distribution in
the pressure chamber during a pressure cycle; and the simulation of the behavior of
HP-treated biomaterials will be the research challenges of the near future (Delgado
et al. 2008).
Isostatic HP can also be used to generate new functional features, such as specific
textures or health promoting properties to develop tailor-made foods. A very
promising field for future research is the use of HP to modulate microbial fermenta-
tions or enzymatic bioconversions. HP might also influence the biosynthetic path-
ways of foods from plant or animal origin as well as of food-related microorganisms,
which could lead to the formation of product variations with unique properties
(Aertsen et al. 2009). A better understanding of the mechanisms underlying pressure
and temperature stability could also enable the development of HP-resistant enzymes.
Despite extensive research in the HP area, there is still a lack of data regarding the
behavior of nutrients, flavors, and allergens during the storage of HP-treated foods.
Acknowledgment The authors Dietrich Knorr and Volker Heinz acknowledge the support of part
of their work through the Marcel Loncin Research Prize.
References
Aertsen A, Meersman F, Hendrickx M, Michiels C (2009) Biotechnology under high pressure:

applications and implications. Trends Biotechnol 27(7):434–441
Ananta E (2005) Impact of environmental factors on vitality and stability and high pressure
pretreatment on stress tolerance of Lactobacillus rhamnosus GG (ATCC 53103) during
spray drying. Thesis, Technische Universit€at Berlin, Berlin
Ananta E, Heinz V, Knorr D (2005) Assessment of high pressure induced damage on Lactobacillus
rhamnosus GG by flow cytometry. Food Microbiol 21(5):567–577
Ananta E, Heinz V, Schl€ uter O, Knorr D (2001) Kinetic studies on high-pressure inactivation of
Bacillus stearothermophilus spores suspended in food matrices. Innovative Food Sci Emerg
Technol 2:261–272
Ananta E, Knorr D (2003) Pressure-induced thermotolerance of Lactobacillus rhamnosus GG.
Food Res Int 36(9–10):991–997
Ananta E, Knorr D (2004) Evidence on the role of protein biosynthesis in the induction of heat
tolerance of Lactobacillus rhamnosus GG by pressure pre-treatment. Int J Food Microbiol
96:307–313
Ardia A (2004) Process considerations on the application of high pressure treatment at elevated
temperature levels for food preservation. Ph.D. thesis, Berlin University of Technology, Berlin, p 94
Ardia A, Knorr D, Heinz V (2004a) Adiabatic heat modelling for pressure build-up during high-
pressure treatment in liquid-food processing. Food Bioprod Process 82(C1):89–95
Ardia A, Knorr D, Ferrari G, Heinz V (2004b) Kinetic studies on combined high-pressure and
temperature inactivation of Alicyclobacillus acidoterrestris spores in orange juice. Appl
Biotechnol Food Sci Policy 1(3):169–173
Black EP, Wei J, Atluri S, Cortezzo D, Koziol-Dube K, Hoover D, Setlow P (2006) Analysis of
factors influencing the rate of germination of spores of Bacillus subtilis by very high pressure.
J Appl Microbiol 102(1):65–76
Bothun GD, Knutson BL, Berberich JA, Strobel H, Nokes S (2004) Metabolic selectivity and
growth of Clostridium thermocellum in continuous culture under elevated hydrostatic pressure.
Appl Microbiol Biotechnol 65(2):149–157
Bridgman PS (1914) The coagulation of albumen by pressure. J Biol Chem 19:511–512
Bridgman PW (1912) Water, in the liquid and five solid forms, under pressure. Proc Am Acad Arts
Sci 47:441–558
Buckow R (2006) Pressure and temperature effects on the enzymatic conversion of biopolymers.
Thesis, Technsiche Universit€at Berlin, Berlin
Buckow R, Heinz V (2008) High pressure processing – a database of kinetic information. Chem
Ing Tech 80(8):1081–1095
Buckow R, Heinz V, Knorr D (2005a) Two fractional model for evaluating the activity of
glucoamylase from Aspergillus niger under combined pressure and temperature conditions.
Food Bioprod Process 83(C3):220–228
Buckow R, Heinz V, Knorr D (2005b) Effect of high hydrostatic pressure-temperature combina-
tions on the activity of b-glucanase from barley malt. J Inst Brew 111(3):282–289
Buckow R, Isbarn S, Knorr D, Heinz V, Lehmacher A (2008) Predictive model for inactivation of
feline calicivirus, a norovirus surrogate, by heat and high hydrostatic pressure. Appl Environ
Microbiol 74:1030–1038
Buckow R, Weiss U, Heinz V, Knorr D (2007a) Stability and catalytic activity of a-amylase from
barley malt at different pressure-temperature conditions. Biotechnol Bioeng 97:1–11
Buckow R, Weiss U, Knorr D (2007b) Combined pressure and temperature effects on the catalytic
activity of cellulase from Bacillus subtilis. In: Proceedings of the 4th International Conference
on High Pressure Bioscience and Biotechnology, Tsukuba, Japan
Buckow R, Weiss U, Knorr D (2009) Inactivation kinetics of apple polyphenol oxidase in different
pressure-temperature domains. Innovative Food Sci Emerg Technol 10(4):441–448
338 D. Knorr et al.
Bunthof CJ (2002) Flow cytometry, fluorescent probes, and flashing bacteria. Ph.D. thesis,
Wageningen University, Wageningen, The Netherlands, p 160
Dalmadi I, Rapeanu G, Van Loey A, Smout C, Hendrickx M (2006) Characterization and
inactivation by thermal and pressure processing of strawberry (fragaria ananassa) polyphenol
oxidase: a kinetic study. J Food Biochem 30(1):56–76
De Heij W, van Schepdael L, Moezelaar R, Hoogland H, Matser A, Van den Berg R (2003) High
pressure sterilization: maximizing the benefits of adiabatic heating. Food Technol 57:37–41
De Vries Y (2006) Bacillus cereus spore formation, structure, and germination. Ph.D. thesis,
Wageningen Universiteit, Wageningen
Delgado A, Rauh C, Kowalczyk W, Baars A (2008) Review of modeling and simulation of
high pressure treatment of materials of biological origin. Trends Food Sci Technol 19:
329–336
Diels AMJ, Michiels CW (2006) High-pressure homogenization as a non-thermal technique for
the inactivation of microorganisms. Crit Rev Microbiol 32(4):201–216
Gould GW, Sale AJH (1970) Initiation of germination of bacterial spores by hydrostatic pressure.
J Gen Microbiol 60:335
Hayakawa K, Kanno T, Yoshiyama K, Fujio Y (1994) Oscillatory compared with continuous high
pressure sterilization of Bacillus stearothermophilus spores. J Food Sci 59:164–167
Heinz (2010) German Institute of Food Technology, Quakenbrueck, Germany (personal commu-
nication)
Heinz V, Buckow R, Knorr D (2005) Catalytic activity of b-amylase from barley in different
pressure/temperature domains. Biotechnol Prog 21:1632–1638
Heinz V, Knoch A, Lickert T (2009) Product innovation by high pressure processing. New Food
2:43–44
Heinz V, Knorr D (1998) High pressure germination and inactivation kinetics of bacterial spores.
In: Isaacs NS (ed) High pressure food science, bioscience and chemistry. The Royal Society of
Chemistry, Cambridge, pp 435–441
Heinz V, Knorr D (2002) Effects of high pressure on spores. In: Hendrickx MEG, Knorr D (eds)
Ultra high pressure treatments of foods. Kluwer/Plenum, New York, pp 77–114
Heinz V, Kortschack F (2002) Method for modifying the protein structure of PrPsc in a targeted
manner. Germany, patent, (WO 02/49460). In: Hendrickx M, Knorr D (eds) Ultra high pressure
treatment of foods. Kluwer/Plenum, New York
Heremans K (2002) The effects of high pressure on biomaterials. In: Hendrickx M, Knorr D (eds)
Ultra high pressure treatments of foods. Kluwer/Plenum, New York, pp 23–52
Heremans K, Smeller L (1998) Protein structure and dynamics at high pressure. Biochim Biophys
Acta 1386:353–370
Hite BH (1899) The effect of pressure in the preservation of milk – a preliminary report. West
Virginia Agr Exp Stat Bull 58:15–35
Indrawati I, Van Loey A, Hendrickx M (2005) Pressure and temperature stability of 5-methylte-
trahydrofolic acid: a kinetic study. J Agric Food Chem 53(8):3081–3087
Isbarn S, Buckow R, Himmelreich A, Lehmacher A, Heinz V (2007) Inactivation of avian
influenza virus by heat and high hydrostatic pressure. J Food Prot 70(3):667–673
Juliano P, Knoerzer K, Fryer PJ, Versteeg C (2009) C. botulinum inactivation kinetics implemen-
ted in a computational model of a high-pressure sterilization process. Biotechnol Prog
25(1):163–175
Kato N, Sato T, Kato C, Yajima M, Sugiyama J, Kanda T, Mizuno M, Nozaki K, Yamanaka S,
Amano Y (2007) Viability and cellulose synthesizing ability of Gluconacetobacter xylinus
cells under high-hydrostatic pressure. Extremophiles 11(5):693–698
Knoerzer K, Juliano P, Gladman S, Versteeg C (2007) A computational model for temperature and
sterility distributions in a pilot-scale high pressure high-temperature process. AlChE J
53:2996–3010
Knorr D (2002) High pressure processing for preservation, modification and transformation of
foods. High Pressure Res 22:595–599
Knorr D, Heinz V, Buckow R (2006) High pressure application for food biopolymers. Biochim
Biophys Acta 1764:619–631
Lori S (2008) Untersuchungen zur Eignung der nicht-thermischen Druckinaktivierung pathogener
Bakterien in Lebensmitteln. Thesis, Hamburg University, Hamburg
Lori S, Buckow R, Heinz V, Lehmacher A (2007) Predictive model for inactivation of Campylo-
bacter spp. by heat and high hydrostatic pressure. J Food Prot 70(9):2023–2029
Ludikhuyze L, Van Loey A, Indrawati, Denys S, Hendrickx M (2002a) Effects of high pressure on
enzymes related to food quality. In: Hendrickx M, Knorr D (eds) Ultra high pressure treatments
of food. Kluwer/Plenum, New York, pp 115–166
Ludikhuyze L, Van Loey A, Indrawati, Hendrickx M (2002b) High pressure processing of fruit and
vegetables. In: Jongen W (ed) Fruit and vegetable processing: improving quality. Woodhead
Publishing, Cambridge, pp 346–362
Margosch D, Ehrmann MA, Buckow R, Heinz V, Vogel R, G€anzle M (2006) High-pressure-
mediated survival of Clostridium botulinum and Bacillus amyloliquefaciens endospores at high
temperature. Appl Environ Microbiol 72(5):3476–3481
Margosch D, G€anzle MG, Ehrmann MA, Vogel RF (2004) Pressure inactivation of Bacillus
endospores. Appl Environ Microbiol 70(12):7321–7328
Mathys A (2008) Inactivation mechanisms of Geobacillus and Bacillus spores during high
pressure thermal sterilization. Ph.D. thesis, Technische Universit€at Berlin, Berlin
Mathys A, Chapman B, Bull M, Heinz V, Knorr D (2007) Flow cytometric assessment of Bacillus
spore response to high pressure and heat. Innovative Food Sci Emerg Technol 8:519–527
Mathys A, Kallmeyer R, Heinz V, Knorr D (2008) Impact of dissociation equilibrium shift on
bacterial spore inactivation by heat and pressure. Food Control 19(12):1165–1173
Mathys A, Reineke K, Heinz V, Knorr D (2009) High pressure sterilization - development and
application of temperature controlled spore inactivation studies. High Pressure Res 29(1):3–7
Meyer RS (2000) Ultra high pressure, high temperature food preservation process. US patent
6,017,572
Molina-Gutierrez A, Stippl V, Delgado A, G€anzle M, Vogel R (2002) In Situ determination of the
intracellular pH of Lactococcus lactis and Lactobacillus plantarum during pressure treatment.
Appl Environ Microbiol 68(9):4399–4406
Niven GW, Miles CA, Mackey BM (1999) The effects of hydrostatic pressure on ribosome
conformation in Escherichia coli: an In vivo study using differential scanning calorimetry.
Microbiology 145(2):419–425
Paidhungat M, Setlow B, Daniels W, Hoover D, Papafragkou E, Setlow P (2002) Mechanisms of
induction of germination of Bacillus subtilis spores by high pressure. Appl Environ Microbiol
68(6):3172–3175
Pavlovic M, Hoermann S, Vogel R, Ehrmann MA (2008) Characterisation of a Piezotolerant
Mutant of Lactobacillus sanfranciscensis. Z Naturforsch 63b:791–797
Perry RH (1984) Perry´s chemical engineers’ handbook. McGraw-Hill, New York
Picard A, Daniel I, Montagnac G, Oger P (2006) In situ monitoring by quantitative Raman
spectroscopy of alcoholic fermentation by Saccharomyces cerevisiae under high pressure.
Extremophiles 11(3):445–452
Rademacher B, Hinrichs J (2006) Effects of high pressure treatment on indigenous enzymes in bovine
milk: reaction kinetics, inactivation and potential application. Int Dairy J 16(6):655–661
Rajan S, Ahn J, Balasubramaniam VM, Yousef AE (2006) Combined pressure-thermal inactiva-
tion kinetics of Bacillus amyloliquefaciens spores in egg patty mince. J Food Prot 69(4):
853–860
Rapeanu G, Van Loey A, Smout C, Hendrickx M (2005) Thermal and high-pressure inactivation
kinetics of polyphenol oxidase in victoria grape must. J Agric Food Chem 58(8):2988–2994
Rastogi NK, Raghavarao KSM, Balasubramaniam VM, Niranjan K, Knorr D (2007) Opportunities
and challenges in high pressure processing of foods. Crit Rev Food Sci Nutr 47:1–44
Reineke K, Mathys A, Knorr D (2008) Temperature control for high pressure processes up to
1400 MPa. J Phys Conf Ser 121:142012–142016
340 D. Knorr et al.
Riahi E, Ramaswamy HS (2004) High pressure inactivation kinetics of amylase in apple juice.
J Food Eng 64:151–160
Rodrigo D, Cortés C, Clynen C, Schoofs L, Van Loey A, Hendrickx M (2006) Thermal and high-
pressure stability of purified polygalacturonase and pectinmethylesterase from four different
tomato processing varieties. Food Res Int 39(4):440–448
Rumpold BA (2005) Impact of high hydrostatic pressure on wheat, tapioca, and potato starches.
Thesis, Berlin Technical University, Berlin, p 120
Schlueter O, Benet GU, Heinz V, Knorr D (2004) Metastable states of water and ice during
pressure supported freezing of potato tissue. Biotechnol Prog 20:799–810
Setlow P (2003) Spore germination. Curr Opin Microbiol 6:550–556
Sharma A, Yadav BS (2008) Resistant starch: physiological roles and food applications. Food Rev
Int 24(2):193–234
Smeller L (2002) Pressure-temperature phase diagram of biomolecules. Biochim Biophys Acta
1595:11–29
Smelt JP, Hellemons JC, Patterson M (2001) Effects of high pressure on vegetative microorgan-
isms. In: Hendrickx M, Knorr D (eds) Ultra high pressure treatments of foods. Kluwer,
New York, pp 55–76
Stone G, Chapman B, Lovell D (2009) Development of a log-quadratic model to describe
microbial inactivation, illustrated by thermal inactivation of Clostridium botulinum. Appl
Environ Microbiol 75(22):6998–7005
Tauscher B (1995) Pasteurization of food by hydrostatic high pressure: chemical aspects. Lebensm
Unters Forsch 200:3–13
Ting E, Balasubramaniam VM, Raghubeer E (2002) Determining thermal effects in high pressure
processing. J Food Technol 56(2):31–35
Trejo Araya XI, Hendrickx M, Verlinden BE, Van Buggenhout S, Smale N, Stewart C, Mawson A
(2007) Understanding texture changes of high pressure processed fresh carrots: a microstruc-
tural and biochemical approach. J Food Eng 80(3):873–884
Ulmer HM, Herberhold H, Fahsel S, G€anzle M, Winter R, Vogel R (2002) Effects of pressure-
induced membrane phase transitions on inactivation of HorA, an ATP-dependent multidrug
resistance transporter, in Lactobacillus plantarum. Appl Environ Microbiol 68:1088–1095
Weemaes CA, De Cordt SV, Ludikhuyze LR, Van den Brock I, Hendrickx M, Tobbuck P (1997)
Influence of pH, benzoic acid, EDTA, and glutathione on the pressure and/or temperature
inactivation kinetics of mushroom polyphenoloxidase. Biotechnol Prog 13:25–32
Weemaes CA, Ludikhuyze L, Van den Broeck I, Hendrickx M (1998) Effect of pH on pressure and
thermal inactivation of Avocado polyphenol oxidase: a kinetic study. J Agric Food Chem
46(7):2785–2792
Wuytack EY, Boven S, Michiels CW (1998) Comparative study of pressure-induced germination
of Bacillus subtilis spores at low and high pressures. Appl Environ Microbiol 64(9):3220–3224
Yaldagard M, Mortazavi SA, Tabatabaie F (2008) The principles of ultra high pressure technology
and its application in food processing/preservation: a review of microbiological and quality
aspects. Afr J Biotechnol 7:2739–2767
Zipp A, Kauzmann W (1973) Pressure denaturation of metmyoglobin. Biochemistry 12(21):
4217–4228
Chapter 15
High Pressure Sterilization of Foods
Hosahalli Ramaswamy
15.1 Introduction
Thermal processing for achieving commercial sterilization involves heating of

foods in hermetically sealed containers for specified time-temperature combina-
tions to eliminate the microbial pathogens that endanger public health, along
with microorganisms and enzymes that deteriorate food during storage. As
originally formulated, the aim of thermal processing was to produce a safe and
shelf-stable food product. Today, however, the consumer demands food products
that are more fresh-like with high-quality, high nutritive value, and end use
convenience. In addition, food processors look for more energy-efficient, cost-
effective, and high-capacity processing technologies. Many processing alterna-
tives including high temperature-short time (HTST) processing, aseptic proces-
sing and packaging, thin profile packaging and processing, and agitation
processing, have primarily evolved to minimize the severity of heat treatment
and promote product quality. Aseptic processing, microwave, radio frequency,
and ohmic heating techniques have gained attention as alternative and noncon-
ventional rapid processing techniques.
To meet consumer demands, there has been an increasing interest in the use of
high hydrostatic pressure processing as a nonthermal food preservation technique.
High pressure processing (HPP) is a method of food processing where food is
subjected to elevated pressures (often in excess, 80,000 lb/in.2 or approximately
5,500 atm or 550 MPa), with or without the addition of heat, to achieve microbial
inactivation or to alter the food quality attributes. It is also referred by various
similar names: high pressure (HP), ultra-high pressure (UHP), high hydrostatic
pressure (HHP) processing, etc. Most vegetative bacteria can be killed at pressures
at or around 550 MPa. Microbial spores, however, are more resistant and require
much higher pressures as well as higher processing temperatures. HPP retains food
quality, maintains natural freshness, and extends microbiological shelf life.
H. Ramaswamy
Department of Food Science, McGill University, Ste Anne de Bellevue, Quebec, Canada
e-mail: hosahalli.ramaswamy@mcgill.ca
342 H. Ramaswamy
Over the past decades, high pressure processing has emerged as a commercial
alternative to traditional thermal processing methods for many foods, e.g., jams,
fruit juices, guacamole, oysters, ready to eat meats, etc. Its primary advantage is
that it can inactivate microorganisms and enzymes at substantially lower treatment
temperatures (as compared to conventional thermal processing) that results in
processed foods possessing sensory and nutrient qualities closely resembling the
original, fresh, or raw product. Generally, high pressure processing as a nonthermal
food processing technique is widely applied in pasteurization of food to extend
food shelf life and preserve high qualities, such as natural color, flavor, and
nutrients. However, several resistant microorganisms, especially spores, and some
enzymes can survive during HP pasteurization. Therefore, HP processed foods
should be kept at refrigerated conditions to prevent spoilage and quality change
(Patterson 2005).
HPP has already become a commercially implemented technology worldwide,
spreading from its origins in Japan, followed by USA, Mexico, Europe, and now
Canada, with worldwide take-up increasing almost exponentially since 2000
(Norton and Sun 2008). High pressure research and development in different
disciplines within the food industry has been reviewed by several researchers
(Norton and Sun 2008; Rastogi et al. 2007; Toepfl et al. 2006; Torres and Velazquez
2005; San Martin-Gonzalez et al. 2002).
15.2 High Pressure Pasteurization
High pressure processing has been commercialized for a variety of acid and
acidified food products. For low acid foods, however, it has been used only as a
temporary measure of extending shelf-life under refrigerated storage conditions. It
has also been used for several other purposes including control of some pathogens
and viruses, for inducing functional changes, as well as improving the nutritional
and sensory quality of foods. Novelties in freezing and thawing application have
been achieved through the use of pressure shift freezing and thawing.
Pasteurization by HPP can be carried out at pressures in the range 400–600 MPa
at relatively moderate (20–50 C) or even at refrigerated temperatures. Under such
conditions, HPP can be effective in inactivating most vegetative pathogens and
spoilage microorganisms. High pressure processing is, as yet, considered to be a
novel process for the production of low acid foods with regulatory bodies.
Extensive research has been carried out in the author’s laboratory on HP
processing of several foods for pasteurization purposes: milk, meat, pork, fish,
orange juice, mango juice, apple juice, etc. (Mussa et al. 1999a, b; Basak et al.
2002; Riahi et al. 2003; Ramaswamy et al. 2003, 2008, 2009; Shao et al. 2006; Gill
and Ramaswamy 2008). These studies have aimed at generating the necessary
kinetic data useful in establishing HP processes for different foods, and then
verifying their validity through inoculated pack/challenge studies with pathogens.
Typically, these studies have shown a dual phase destruction kinetics for most
15 High Pressure Sterilization of Foods 343
pathogens. The first phase was generally characterized as pulse effect (PE), indicat-
ing the destruction to be essentially a pressurization-depressurization effect with no
hold time (depends on the pressure level) because the pressure is released immedi-
ately after the pressurization process. While pressure come-up and release duration
and profiles would undoubtedly contribute to some destruction based on the pres-
sure destruction rates, the measured PE were found to be significantly higher than
those that could be computed from the accumulated destruction during come-up
and come-down periods (based on the D and z values). Following the PE, the
destruction patterns in these studies have generally been found to be well described
by the log-linear model following the first order rate kinetics. Typical values of PE
and the HP destruction kinetic parameters, D and z for selected microbial pathogens
in different food media are shown in Tables 15.1 and 15.2.
Table 15.1 Pressure pulse effect (no-holding time) for selected pathogens in different foods
Microorganism Food base Pulse effect (log value) References
Escherichia coli (O17:H7) Apple juice (pH 3.5) 3.95 at 30 C, 400 MPa Riahi et al. (2003)
E. coli (O157:H7) Cheese 0.3 (400 MPa) Shao et al. (2006)
Listeria monocytogenes Cheese 0.25 (400 MPa) Shao et al. (2006)
L. monocytogenes Milk 1.3 (400 MPa) Mussa et al. (1999a)
L. monocytogenes Pork 1.5 (400 MPa) Mussa et al. (1999b)
Table 15.2 High pressure destruction kinetics in the context of pasteurization

Name D value (min) (MPa) Z value References
Escherichia coli O157:H7 in
Apple juice 14.3 (10 C, 400 MPa) 654 Riahi et al. (2003)
3.22 (35 C, 400 MPa) 137 Riahi et al. (2003)
Fish slurry 3.19 (400 MPa) 185 Ramaswamy et al. (2008)
Cheese 2.0 (400 MPa) 128 Shao et al. (2006)
Listeria monocytogenes in
Milk 10.1 at 350 MPa 266 Mussa et al. (1999a)
Pork 3.52 at 400 MPa 163 Mussa et al. (1999b)
Fish 49 at 400 MPa 103 Ramaswamy et al. (2008)
Cheese 1.4 at 350 MPa 82 Shao et al. (2006)
15.3 High Pressure Sterilization
HPP for producing shelf-stable low acid foods is still a topic of considerable
controversy. Processors worldwide are waiting for the full regulatory approval of
HP sterilization of low acid foods. HPP has the potential to produce better quality
foods than possible from the use of processing novelties such as microwave, RF, or
Ohmic heating techniques in combination with aseptic processing. This is because
344 H. Ramaswamy
HPP allows the product temperature to be increased very rapidly (due to adiabatic
heating) from around 90–100 C to the sterilization zone (120–130 C) (come-up
time) and bringing it back to nearly the same state almost instantaneously
by depressurization. The process can be formulated either as a pressure-assisted
thermal sterilization (PATS) or temperature-assisted pressure sterilization (TAPS).
Either way, it represents a considerable deviation from conventional thermal pro-
cessing because of lack of kinetic data on destruction of spoilage and pathogenic
microorganisms under processing conditions. A better understanding of the inacti-
vation kinetics of pathogens and/or their surrogates under high temperature high
pressure processing conditions is the key for the success of HPP and for FDA
approval. The critical factors in HPP include pressure, pressure holding time, time
to achieve treatment pressure, depressurization time, treatment temperature (includ-
ing adiabatic heating), product initial temperature, vessel temperature distribution
during the pressure treatment, product pH, product composition, product water
activity, packaging material integrity, and any concurrent processing aids. Although
HP processing-related research work has increased tremendously in the last decade,
there is still a serious lack of specific data in this area to permit establishment of a
reliable process.
As described earlier, HPP of low acid foods can either be treated as a PATS or
TAPS process. For PATS, achieving designated process lethality for commercial
sterilization becomes the main issue, which can be easily achieved, and such a
process would most likely meet the regulatory approval standards. The process
would rely on the effective use of compression heating of the food achieved during
the pressurization process. Starting at an appropriate preset initial temperature,
process temperatures can easily be brought to levels used for HTST levels practiced
in commercial sterilization applications. Thus, holding the product at such pres-
sures for a specified time would easily result in the accumulation of the desired
target process lethality. The process generally involves a come-up time during
which pressure and temperature continuously increase to set levels. During the
hold, pressure generally remains constant, while the temperature could drop due to
heat loss from the product to equipment structure. It is necessary to address this
issue in order to provide appropriate insulation for the product to prevent large
temperature drop and to appropriately take the temperature drop in to consideration
in establishing the process. In the second approach (TAPS), additional destruction
caused by the pressure at the elevated temperatures is taken into account. It is
recognized that unless the process pressure is 600–800 MPa range along with
elevated process temperatures in the 80–120 C range, it is not possible to kill the
microbial spores and hence not possible to achieve commercial sterility. However,
at such pressure levels, microbial destruction at the process temperatures could
occur much more rapidly than at the same temperature under conventional retort
processing conditions. Thus, the TAPS process could benefit from accelerated
destruction kinetics and could potentially result in an effective short time process
providing similar quality advantages as the PATS process. However, this process
would generally be considered a “novel” process and would need additional data on
pressure destruction kinetics on both spoilage and pathogenic microbial spores at
the prevailing processing conditions. While the PATS process could be more easily
cleared by regulatory agencies along the guidelines used for traditional thermal
processing, the TAPS process would require demonstration of accelerated spore
destruction kinetics of pathogenic and spoilage bacterial spores.
15.4 Compression Heating
Compression heating is an unavoidable consequence in high pressure processing.

When the food and medium are subjected to HP processing, their volume is reduced
and pressure is simultaneously elevated, both of which contribute to an increase in
temperature of the food and pressurization medium as a result of the adiabatic
compression. The temperature rise depends on the nature of the food and the
medium, and is also dependent on the initial temperature and pressure levels
involved. Extensive research has been carried out in this area and data have been
compiled on the compression heating behavior of various foods and pressurization
media. It is necessary to use such data to determine the temperatures of both the
medium and the product following pressurization.
Only the pressurization medium and the packaged food (including the packaging
material) are subjected to such compression heating. The material used for making
the pressure vessel will not undergo such adiabatic compression and will not be
subjected to such temperature increases. Thus, the steel chamber acts as a heat sink
for the temperature elevated food product and pressurization medium unless com-
pensated by an internal temperature heating system or adequate insulation at the
contact surface. Typical changes in the temperature of the product in insulated and
uninsulated products are shown in Fig. 15.1.
Tinsulated product
Temperature or pressure
TUninsulated product
TPressure medium
Pressure
Time
Fig. 15.1 Typical compression heating temperature rise in the pressurization medium and pro-
duct, and the subsequent heat loss during pressure hold time
346 H. Ramaswamy
It is necessary to evaluate the related time-temperature distribution and compute

the process time based on the cold spot location of the product. The holding time
needs to be adjusted to bring the accumulated lethality at the coldest point in the
product to the target levels (e.g., Fo value of 5 min). This process development would
normally follow the guidelines for thermal process establishments which may include
temperature distribution in the pressure vessel and heat penetration data gathering.
In the TAPS process temperature stability is equally important, since it can affect the
spore inactivation kinetics. This aspect is discussed in the next section.
The compression heating or the adiabatic temperature rise is dependent on the type
of pressurization liquid and food components since the pressure compressibility of
each material or constituent is different. Further, it also depends on the pressure level
and initial temperature, generally increasing with an increase in both initial tempera-
ture and pressure, and often demonstrating a synergy between the two. Some typical
values of adiabatic temperature that rise per 100 MPa pressure rise at an initial
temperature of 25 C are given below (Rasanayagam et al. 2003):
Orange juice, tomato salsa, skim milk, salmon fish: 2.6–3.0 C
Carbohydrates: 2.6–3.6 C
Proteins: 2.7–3.3 C
Mayonnaise: 5.0–7.2 C
Extracted beef fat: 6.2–8.3 C
Soybean oil: 6.2–9.1 C
Olive oil: 6.3–8.7 C
Shao et al. (2010) expressed the adiabatic temperature rise (DTP) as a quadratic
function of pressure and initial temperature:
DTP ¼ 3:06 þ 0:0224Ti þ 0:0423P þ 4:49 104 Ti 2 þ 1:31

104 Ti P 1:24 105 P2 (15.1)
(R2 ¼ 0.999, n ¼ 50, SE ¼ 0.20 C, P < 0.05 for all items) or a function of
pressure and a desired HP treatment temperature:
DTP ¼ 2:21 þ 0:0328TP þ 0:0381P þ 2:82 104 TP 2 þ 8:64

105 TP P 1:25 105 P2 (15.2)
(R2 ¼ 0.999, n ¼ 50, SE ¼ 0.18 C, P-value <0.05 for all items).
15.5 Spore Inactivation Studies
Commercially sterile low-acid food (LCAF) products are commonly produced by

thermal processing, and they are based on a process which should be able to achieve
a 12 decimal reduction (D value) in a Clostridium botulinum spore population.
Once the degree of sterility is achieved, the process needs to be fine-tuned with
respect to spoilage causing bacterial spores so that shelf stability can be assured.
High pressure processing when combined with moderately elevated temperatures
can inactivate highly resistant bacterial spores. Traditionally, bacterial spores such
as C. sporogenes, B. sterothermophilus, C. liquifaciencs, etc., with high heat resis-
tance, have been used for establishing the process for spoilage control and for testing
the efficacy of the commercial sterilization. The obvious choices for HP research
were to test if these were more pressure resistant as well. Hence, much of the work on
bacterial spores has been concentrated on the pressure destruction kinetics of such
spores in various food matrices (Patazca et al. 2006; Shao et al. 2010; Zhu et al.
2008).
There have been several reports on the pressure destruction of bacterial spores.
Rovere et al. (1996) reported D values of 3.5, 3.2 min for 600, and 700 MPa at
100 C, respectively, for C. sporogenes 7,955 spores in meat broth. Mills et al.
(1998) found that there is no inactivation of C. sporogenes spores with HP treat-
ments at even 600 MPa for 30 min when tested at 20 C and concluded that these
spores could not be inactivated by pressure alone. Meyer et al. (2000) reported that
a two-cycle treatment with initial temperature of 90 C combined with 690 MPa for
1 min achieved sterility in macaroni and cheese with 106/g of C. sporogenes spores.
Reddy et al. (1999, 2003, 2006) studied the effect of HP treatment on spores of
C. botulinum type A, B, and E at moderately elevated temperatures and found the
Type A spores to be more resistant. They reported that only 3 log cycle reductions
were achieved at 827 MPa and 75 C pressure conditions. Margosch et al. (2004a, b,
2006) reported that proteolytic TMW 2.357 (C. botulinum Type B) and TMW 2.479
(Bacillus amyloliquefaciens) exhibited a greater resistance to pressure than other
bacterial spores (Bacillus spp. and C. botulinum spp.). Koutchma et al. (2005), in
their pressure destruction kinetic study of C. sporogenes PA3679 spore in phos-
phate buffer, reported pressure (ZP) and temperature sensitivity (ZT) values of
23.7 C and 1,500 MPa, respectively. Ahn and Balasubramaniam (2007b)) observed
pressure-assisted thermal processing (PATP) at 700 MPa and 121 C for 1 min
inactivated up to 7–8 log reduction for C. sporogenes spores. In another paper (Ahn
et al. 2007a), it was reported that spore clumps formed during the PATP may lead to
an increase in pressure-thermal resistance, and that multiple-pulsed pressurization
can be more effective in inactivating bacterial spores. These studies in general
suggest that spore inactivation kinetics is likely to depend on the type of substrate
(food medium or matrix) in which they are pressure treated.
Recent studies at McGill (Shao 2008) have concentrated on spore (surrogate
and pathogenic) inactivation kinetics using high pressure. Destruction kinetics
tests were carried out with two strains of C. sporogenes (11437, 7955) and Geo-
bacillus stearothermophilus 10,149 spores suspended in milk at 700–900 MPa
and 70–100 C. The survival counts were found to well fit the first order linear
models. The D values of C. sporogenes 11,437 varied from 0.73 min at 900 MPa/
100 C to 17.0 min at 700 MPa/80 C, while they ranged from 6.0 to 833 min
at 80–100 C under thermal processing conditions. The D values associated
with C. sporogenes 7,955 spores were higher than for C. sporogenes 11437 and
348 H. Ramaswamy
110
Chamber pressure
100 800
90
Temperature (C)
Pressure (MPa)
600
80
70 Sample temperature 400
60
200
50
40 0
0 10 20 30
Time (min)
Fig. 15.2 Typical pressure and sample temperature profiles during HP treatment at 700 MPa and
80 C for 24 min
varied from 1.3 min at 900 MPa/100 C to 38.2 min at 700 MPa/80 C treatments,
and from 12.1 to 156 min at 80–100 C during thermal treatments. Typical temper-
ature stability of test samples during the pressure treatment is shown in Fig. 15.2.
The D values of G. stearothermophilus 10149 spores varied from 0.6 min at
900 MPa/90 C to 20.9 min at 500 MPa/70 C treatments with 6.3 to 49.4 min for
thermal treatments at 110–120 C. Hence, C. sporogenes 7955 spores were the
most resistant among those studied. The HP destruction kinetics of C. sporogenes
7955 spores were also studied in salmon and were lower than in milk (Shao 2008).
15.5.1 Clostridum botulinum Studies
C. botulinum is an anaerobic, mesophilic, spore-forming pathogen that poses public

health risk in low acid foods. Most commercial processes for LACF are based on
giving a process lethality of least 5 min or longer (which is > 20D of the target
C. botulinum spores). C. botulinum is a pathogen capable of producing severe
neurotoxins, and therefore is rarely used in process development studies. Instead,
surrogate spores of similar resistance are often employed. C. sporogenes, a typical
thermally resistant, mesophilic spore-former, is commonly used when inoculated
pack verification of thermal processes are warranted. While destruction kinetics of
bacterial spores have been widely studied under thermal processing conditions for
almost 200 years, similar work in the area of HP processing are in their infancy.
In the most recent studies at McGill (Shao 2008), 12 C. botulinum Group I
strains (62A IB1-B, CK2-A, MRB, Langeland, A6, GA0108BEC, PA9508B,
13983B, H461297F, GA0101AJO, HO9504A) were subjected to combinations of
high pressures (800 and 900 MPa) and elevated temperatures (90 and 100 C). The
treatment holding time varied from 0.5 to 15 min. Sample tubes were kept in a POM
plastic thick wall insulator, preheated to initial temperature, and placed into a
pressure chamber prior to HP treatments for controlling process temperature at
the desired condition. The survival counts showed that higher pressure and
temperature combinations always accelerated the inactivation of the spores. Strain
62A was completely inactivated by these combinations. Strains PA9608B,
HO9504A, and CK2-A were found to be of higher pressure resistance than the
rest 12 strains. The log reductions of these three strains were 1.63, 3.33, and 4.48
log units for 900 MPa/100 C/3 min and the estimated D values were 1.8, 0.88, and
0.66 min, respectively, at 900 MPa/100 C treatment. The strain PA9805B produced
the most pressure-resistant spores. By studying pressure destruction effects in milk
and buffer (substrate) on spore resistance under high pressure (700–900 MPa,
100–110 C) it was found that PA9508B spores had a higher resistance in milk
than in the buffer. In further studies, more detailed evaluation of pressure destruc-
tion of the PA9805B strain of C. botulinum spores were evaluated in milk. At
900 MPa, the associated D values were 14.5, 1.8, and 0.35 min at 90, 100, and
110 C, respectively. The ZT values were 11.2, 12.3, 12.4 C at 700, 800, 900 MPa,
respectively, increasing with pressure and with higher value than the thermal z
value 7.8 C. The ZP values were 470, 630, and 800 MPa at 90, 100, and 110 C,
respectively, increasing with temperature. By comparison of the ZP and ZT, it
appeared that the spore was relatively more sensitive to temperature than to
pressure. D value trends demonstrated that the pressure inactivation effect steadily
decreased as the temperature increased and that at temperatures beyond 115 C, heat
was essentially the principle mode of spore destruction.
Overall, spore inactivation studies have demonstrated several important find-
ings. The nonpathogenic C. sporogenes 7955 spore was the most resistant surro-
gate, but the pathogenic C. botulinum PA9508B spore was even more resistant. D
values associated with HP at elevated temperatures were higher than under conven-
tional thermal treatments, and hence provide accelerated destruction kinetics at
least for the nonpathogenic spores and better spoilage control. However, from a
safety point of view, conventional thermal sterility requirements would still persist
even under HP processing conditions. Milk, as a low acid food medium, provided
more resistance for HP destruction than fish.
References
Ahn J, Balasubramaniam V, Yousef A (2007) Inactivation kinetics of selected aerobic and

anaerobic bacterial spores by pressure-assisted thermal processing. Int J Food Microbiol
113:321–329
Ahn J, Balasubramaniam V (2007) Effects of inoculum level and pressure pulse on the inactiva-
tion of Clostridium sporogenes spores by pressure-assisted thermal processing. J Microbiol
Biotechnol 17(4):616–623
Basak S, Ramaswamy H, Piette G (2002) High pressure destruction kinetics of Leucononostoc
mesenteroides and Sachharomyces cerevisiae in single strength and concentrated orange juice.
Innovative Food Sci Emerg Technol 3(Sept):223–245
Gill A, Ramaswamy H (2008) Application of high pressure processing to kill Escherichia coli
O157 in ready-to-eat Meats. J Food Prot 71(11):2182–2189
Koutchma T, Guo B, Patazca E, Parisi B (2005) High pressure-high temperature sterilization: from
kinetic to analysis to process verification. J Food Process Eng 28:610–629
350 H. Ramaswamy
Margosh D, Ehrmann M, Buckow R, Heinz V, Vogel R, Ganzle M (2006) High-pressure-mediated

survival of Clostridium botulinum and Bacillus amyloliquefaciens endospores at high temper-
ature. Appl Environ Microbiol 72:3476–3481
Margosh D, Ehrmann M, Ganzle M, Vogel R (2004a) Comparison of pressure and heat resistance
of Clostridium botulinum and other endospores in mashed carrots. J Food Prot 67
(11):2530–2537
Margosh D, Ganzle M, Ehrmann M, Vogel R (2004b) Pressure inactivation of Bacillus endo-
spores. Appl Environ Microbiol 70(12):7321–7328
Meyer R, Cooper K, Knorr D, Lelieveld H (2000) High-pressure sterilization of foods. Food
Technol 54(11):67–72
Mills G, Earnshaw R, Patterson M (1998) Effects of high hydrostatic pressure on Clostridium
sporogenes spores. Lett Appl Microbiol 26:227–230
Mussa D, Ramaswamy H, Smith J (1999a) High pressure destruction kinetics of Listeria mono-
cytogenes in milk. Food Res Int 31(5):343–350
Mussa D, Ramaswamy H, Smith J (1999b) Ultra high pressure destruction kinetics of Listeria
monocytogenes in pork. J Food Protect 62(1):165–170
Norton T, Sun D-W (2008) Recent advances in the use of high pressure as an effective processing
technique in the food industry. Food Bioprocess Technol 1:2–34
Patazca E, Koutchma T, Ramaswamy H (2006) Inactivation kinetics of Geobacillus stearother-
mophilus spores in water using high-pressure processing at elevated temperatures. J Food Sci
71(3):M110–M116
Patterson M (2005) A review: microbiology of pressure-treated foods. J Appl Microbiol
98:1400–1409
Ramaswamy H, Jin H, Zhu S (2009) Effects of fat, casein and lactose on high-pressure destruction
of Escherichia coli K12 (ATCC-29055) in milk. Food Bioprod Process 87(C1):1–6
Ramaswamy H, Riahi E, Idziak E (2003) High-pressure destruction kinetics of E-coli (29055) in
apple juice. J Food Sci 68(5):1750–1756
Ramaswamy H, Zaman S, Smith J (2008) High pressure destruction kinetics of Escherichia coli
(O157:H7) and Listeria monocytogenes (Scott A) in a fish slurry. J Food Eng 87(1):99–106
Rasanayagam V, Balasubramaniam V, Ting E, Sizer C, Anderson C, Bush C (2003) Com-
pression heating of selected fatty food substances during high pressure processing. J Food
Sci 68(1):254–259
Rastogi N, Raghavarao K, Balasubramaniam V, Niranjan K, Knorr D (2007) Opportunities and
challenges in high pressure processing of foods. Crit Rev Food Sci Nutr 47(1):69–112
Reddy N, Solomon H, Fingerhut G, Balasubramaniam V, Rhodehamel E (1999) Inactivation of
Clostridium botulinum types A and B spores by high-pressure processing. IFT Annual Meeting:
Book of Abstracts
Reddy N, Solomon H, Tetzloff R, Rhodehamel E (2003) Inactivation of Clostridium botulinum
type A spores by high pressure processing at elevated temperatures. J Food Prot 66
(8):1402–1407
Reddy N, Tetzloff R, Solomon H, Larkin J (2006) Inactivation of Clostridium botulinum non-
proteolytic type B spores by high pressure processing at moderate to elevated high temperature.
Innovative Food Sci Emerg Technol 7:169–175
Riahi E, Ramaswamy H, Smith J (2003) High pressure destruction kinetics of E. coli (O157:H7) in
apple juice. 2003 CSAE/SCGR Meeting, Macdonald Campus of McGill University Ste, Anne
de Bellevue, Quebec, 6–9 July 2003
Rovere P, Carpi G, Dall’Aglio G, Gola S, Maggi A, Miglioli L, Scaramuzza N (1996) High-
pressure heat treatments: evaluation of the sterilizing effect and of thermal damage. Ind
Conserve 71:473–483
San Martin M, Barbosa-Canovas G, Swanson B (2002) Food processing by high hydrostatic
pressure. Crit Rev Food Sci Nutr 42(6):627–645
Shao Y (2008) High pressure destruction kinetics of bacterial spores in low acid foods. Ph.D.
thesis, McGill University
Shao Y, Ramaswamy H, Zhu S (2006) High pressure destruction kinetics of spoilage and
pathogenic bacteria in raw milk cheese. J Food Process Eng 30(3):357–374
Shao Y, Zhu S, Ramaswamy H, Marcotte M (2010) Compression heating and temperature control
for high pressure destruction of bacterial spores: an experimental method for kinetics evalua-
tion. Food Bioprocess Technol 3:71–78. doi:10.1007/s11947-008-0057-y
Toepfl S, Mathys A, Heinz V, Knorr D (2006) Potential of high hydrostatic pressure and pulsed
electric fields for energy efficient and environmentally friendly food processing. Food Rev Int
22(4):405–423
Torres J, Velazquez G (2005) Commercial opportunities and research challenges in the high
pressure processing of foods. J Food Eng 67(1–2):95–112
Zhu S, Naim F, Marcotte M, Ramaswamy H, Shao Y (2008) High-pressure destruction kinetics of
Clostridium sporogenes spores in ground beef at elevated temperatures. Int J Food Microbiol
126(1–2):86–92
.
Chapter 16
Bioseparation of Nutraceuticals Using
Supercritical Carbon Dioxide
Feral Temelli and Bernhard Seifried
16.1 Introduction
The use of supercritical carbon dioxide for bioseparation of nutraceuticals has been
growing rapidly over the past decade in response to increasing consumer demand
for “natural” products in the functional foods and nutraceuticals markets. Carbon
dioxide at temperature and pressure conditions above its critical point (31.1 C,
74 bar) is referred to as supercritical carbon dioxide (SC–CO2), which is a dense
fluid with physical properties in between those of a gas and a liquid. Thus, SC–CO2
can be used as a solvent similar to other organic solvents for the extraction of
nutraceuticals. The main distinction is that upon depressurization following extrac-
tion SC–CO2 becomes a gas and separates readily from the extract, eliminating the
need for extra heat treatment, which is necessary for the removal of organic solvents
from the extracts. As well, processing can be carried out at just above ambient
temperatures, minimizing degradation of heat labile bioactive compounds. Further-
more, CO2, being the second least expensive solvent after water, is nontoxic,
nonflammable, readily available, inexpensive, and can be recycled relatively easily.
Thus, CO2 has been the solvent of choice for the processing of nutraceuticals. In
addition, in the CO2 environment, undesirable oxidation reactions are minimized in
the absence of oxygen, maintaining the integrity of bioactives. Based on all these
advantages, SC–CO2 processing of nutraceuticals has reached commercialization
and numerous plants have been built around the world over the last decade. Other
fluids, such as propane, dimethylether, and others are being investigated for the
extraction and fractionation of biomaterials; however, this chapter will focus
mainly on CO2 due to the advantages listed above.
Because of the nonpolar nature of CO2 it is selected for nonpolar solutes and
requires the addition of a polar co-solvent such as ethanol to solubilize polar
solutes. In terms of nutraceuticals, protein- and carbohydrate-based bioactives are
F. Temelli (*) and B. Seifried

Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, AB,
Canada T6G 2P5
e-mail: feral.temelli@ualberta.ca
354 F. Temelli and B. Seifried
not soluble in SC–CO2 to any appreciable extent; therefore, this chapter will focus
on lipid-based nutraceuticals as well as on minor components or phytochemicals.
Even though SC–CO2 extraction has become more mainstream further processing
in terms of fractionation requires additional development. Natural materials are
very complex and each matrix needs to be studied separately to optimize various
processing parameters on a case-by-case basis. Fundamental data needed for
process design and optimization are missing in many cases. Therefore, our under-
standing of the fundamentals of various separation processes is limited. The
objective of this chapter is to provide a review of the fundamentals and separation
processes, especially the extraction and fractionation of major classes of nutraceu-
ticals that are typically processed by SC–CO2 (i.e., lipid-based nutraceuticals,
carotenoids, and phytochemicals), and to highlight the challenges and provide
some insight into the future outlook of such processes.
16.2 Fundamentals
A pure component when heated above its critical temperature (Tc) and pressurized
above its critical pressure (Pc) is called a supercritical fluid. In the supercritical fluid
region the phase boundary between the liquid and vapor phases disappears, thus, the
two phases having identical density become indistinguishable and form what is
often referred to as “dense gas” (Brunner 1994). Above the critical temperature a
pure gas cannot be liquefied even at very high pressures, as illustrated in Fig. 16.1
for CO2. Depending on molecular size, polarity, and intermolecular hydrogen
bonding, the critical temperatures of pure substances vary over a wide range, with
water having a particularly high critical temperature (Tc ¼ 374.2 C) and pressure
(Pc ¼ 220 bar), while that for SC–CO2 is considered to be moderate (Tc ¼ 31.1 C,
Pc ¼ 74 bar) (McHugh and Krukonis 1986).
Fig. 16.1 Schematic phase diagram for pure carbon dioxide showing solid, liquid, gas and
supercritical fluid regions, triple point (Tp), and critical point (Cp)
16 Bioseparation of Nutraceuticals Using Supercritical Carbon Dioxide 355
16.2.1 Physical and Transport Properties
In Fig. 16.1, there is no sudden change in component properties when crossing the
“dashed lines” from the liquid or gas region into the supercritical fluid region. With
the exception of the critical point, the variation in fluid properties is monotonous.
The values for density, viscosity, and diffusivity of supercritical fluids are between
those of gases and liquids (Table 16.1). However, at the critical point, some
physical properties such as heat capacity and thermal conductivity exhibit a maxi-
mum. Gas-like diffusivity and viscosity, low surface tension, and liquid-like den-
sity, together with the tunable solvent power of supercritical fluids, are particularly
advantageous for processes involving mass transfer, such as extraction and frac-
tionation. Close to the critical point the density of supercritical fluids is highly
dependent on pressure and temperature, where the isothermal compressibility of
CO2 tends to infinity, whereas at higher pressures the influence of temperature is
less pronounced. The pressure dependence of density is illustrated in Fig. 16.2,
where the slope of the isotherms close to the critical point located in the supercriti-
cal region is steeper than at higher pressures and temperatures. Therefore, the
density and subsequently the solvent strength of a supercritical fluid are adjustable
by modest changes in pressure and temperature, which are utilized in solubility-
based separation processes. Additionally, the knowledge of transport properties
is crucial for optimal design of separation processes. Transport properties of
Table 16.1 Physical Physical property Gas SCF Liquid

properties of supercritical
Density (g/cm3) 0.001 0.2–1.0 0.6–1.6
fluids (SCF)
Viscosity (mPa.s) 0.0001 0.001 0.01
Diffusivity (cm2/s) 0.1 0.001 0.00001
Fig. 16.2 Density versus pressure at various temperatures for carbon dioxide (Data from NIST
2009)
supercritical fluids, such as viscosity, diffusivity, and thermal conductivity are

influenced by pressure, temperature, and concentrations of solute or co-solvent.
The dynamic viscosity of neat CO2 increases with pressure and decreases with
temperature, as illustrated in Fig. 16.3 (Fenghour et al. 1998; NIST 2009). It is
challenging to measure the viscosity of supercritical fluids close to the critical point
due to the highly compressible nature of the fluid phase. Therefore, there
are discrepancies between results obtained by different experimental methods
(Fenghour et al. 1998). For example, experimental data for the dynamic viscosity
of SC–CO2 determined using a capillary instrument, suggesting that the viscosity
increased substantially along isotherms in the vicinity of the critical point (Michels
et al. 1957), were found to be inaccurate when compared to other methods (Kestin
et al. 1964). It has been shown in later measurements carried out using an oscillating
disc viscometer that there is only a mild divergence (<1%) of the viscosity
isotherms in the vicinity of the critical point (Kestin et al. 1964). The oscillating
disc viscometer is better suited for measurements close to the critical point, whereas
compressibility of SC–CO2 with capillary instruments can lead to errors since such
instruments require an appreciable pressure drop across the capillary. The viscosity
of binary or multicomponent mixtures of SC–CO2 involving solute and cosolvent
mixtures is even more complex to measure and only limited data are available in the
literature (Yener et al. 1998; Tuan et al. 1999). Using a capillary viscometer, Yener
et al. (1998) found that the viscosity of SC–CO2 saturated with methyl oleate at
maximum concentrations of 4–5 wt% increased by 15–20%, compared to neat
SC–CO2 at 137 bar and 50 C. The viscosity increase for SC–CO2 mixed with lipids
was found to be linear with increasing lipid concentration. The mixture viscosity of
SC–CO2 with various cosolvents at concentrations ranging from 1 to 5 mol% was
measured using a falling weight viscometer, indicating that the fluid viscosity was
increased by the cosolvent addition, depending on the size, polarity, and concentra-
tion of cosolvent molecules (Tilly et al. 1994). Therefore, during separation pro-
cesses, changes in the concentration of solutes or cosolvent can have a pronounced
effect on the viscosity of the supercritical phase, thereby impacting fluid flow,
diffusivity, thickness of boundary layer and, finally, the mass transfer rate.
120
305 K
100
315 K
Viscosity [mPa*s]
80 325 K
335 K
60
40 380 K
Fig. 16.3 Viscosity versus 20

pressure at various
temperatures for carbon 0
dioxide (Data from NIST 0 100 200 300 400
2009) Pressure [bar]
The diffusion coefficient being influenced by viscosity is equally important for

separation processes involving mass transfer. Diffusion in supercritical fluids was
reviewed by Liong et al. (1991), including a description of experimental methods
and a discussion of various factors influencing diffusion coefficients. Diffusion
coefficients of solutes in supercritical fluids are affected by numerous factors, such
as temperature, pressure, solute and cosolvent concentrations, density, viscosity, as
well as molar volume, molecular weight, structure, and polarity of the solute.
Similar to viscosity, interactions between solute, solvent, and cosolvent can affect
the diffusion coefficient. An excellent review of binary diffusion coefficients at
infinite dilution in supercritical fluids, as well as graphical correlations and trends,
are given by Suárez et al. (1998). The following general trends can be observed:
diffusion coefficients in supercritical fluids increase with increasing temperature
and decreasing pressure, due to reduced density and viscosity, leading to decreased
collisions and increased mean free paths of the solute. Additionally, smaller
molecules diffuse faster than larger ones and diffusion coefficients of polar solutes
seem to be more affected by temperature changes than those of nonpolar or low
polarity substances, indicating that interactions between solute and solvent need to
be considered.
Majority of the data available in the literature regarding diffusion coefficients of
solutes in SC–CO2 has been reported for infinite dilution. However, the reality in
some separation processes is far from that, and thus diffusion coefficients deter-
mined at infinite dilution have limited applicability and the concentration depen-
dence of diffusion coefficients has to be taken into account (Raspo et al. 2008).
Likewise, since most experimental data have been determined for infinite dilution,
most of the subsequent correlations and models developed to predict diffusion
coefficients are also valid for infinite dilution. For some nutraceutical solutes of
very low solubility, such as b-carotene, the diffusion coefficient at infinite dilution
may be applicable. However, for cases where solubility is high the influence of
concentration has to be taken into account. This can be accomplished by using a
modified form of the Darken equation (Darken 1948), which takes into account the
solute concentration and a thermodynamic factor that can be predicted by an
equation of state approach (Higashi et al. 1999).
Diffusion coefficients of unsaturated fatty acid methyl esters (FAME) with
carbon chain lengths ranging from C16 to C24 in SC–CO2 were reported by
Funazukuri et al. (1991). Additionally, the authors tested several correlations for
calculating the binary diffusion coefficient and proposed a new correlation based on
the Schmidt number. Diffusion coefficients of lipids, including oleic, linoleic and
linolenic acids, methyl and ethyl oleate and di- and trilinolein, were determined by
Rezaei and Temelli (2000) based on the peak broadening technique using super-
critical fluid chromatography; the diffusion coefficients decreased in the following
order: fatty acid esters > fatty acid > triglycerides, indicating that both size and
polarity play a role. The effects of molecular weight and degree of unsaturation of
lipids on infinite dilution binary diffusion coefficients in SC–CO2 were assessed by
Funazukuri et al. (2004a), showing that for C18 fatty acids with increasing number of
double bonds, the diffusion coefficient decreased moderately in the case of methyl
and ethyl esters as well as triglycerides. However, for free fatty acids (C18 and C20)
the trend was the opposite, which may be explained by a stronger effect of the
carboxyl group on diffusivity than that of double bonds. A relatively simple
hydrodynamic equation relating the binary diffusion coefficient at infinite dilution
to viscosity and temperature of the supercritical fluid, together with two solute
dependant parameters fitted to experimental data, was successfully applied to
lipids and other solutes, including docosahexaenoic acid, eicosapentaenoic acid,
a-tocopherol, and b-carotene (Funazukuri et al. 2004b). Under conditions away
from the binary mixture critical point, where diffusion coefficients tend to zero, a
correlation for calculating the binary diffusion coefficient of various biomaterials,
such as tocopherols and triglycerides, can be used with average deviations of 10%
compared to literature data (Catchpole and King 1994). The correlation requires the
solvent molecular weight, reduced temperature, density, solute molecular weight,
and an estimate of the solute critical volume as input data.
Knowledge of thermal conductivity of the supercritical fluid phase is important
for separation processes involving heat transfer, where separation is brought about
by a change in temperature, thereby affecting density and solubility. Thermal
conductivity for pure CO2 as a function of pressure and temperature is illustrated
in Fig. 16.4 (Vesovic et al. 1990; NIST 2009).
Data on transport properties of supercritical fluid systems are essential for
designing processing equipment. However, they are challenging to measure and
quantify accurately, to reflect real processing situations, where often multicompo-
nent mixtures at considerable concentrations are encountered rather than dilute
binary mixtures. Numerous attempts have been made to develop models and
correlations to accurately predict transport properties. However, the accuracy of
experimental data still outperforms most correlations and models, which indicates
that more research is needed to fully understand all factors impacting transport
properties.
0.14
0.12 305 K
Therm. Cond. [W/m*K]
315 K
0.1
325 K
0.08
335 K
0.06
380 K
0.04
0.02
0
0 100 200 300 400
Pressure [bar]
Fig. 16.4 Thermal conductivity versus pressure at various temperatures for carbon dioxide (Data
from NIST 2009)
16.2.2 Solubility Behavior
Solubility describes the equilibrium between a solute and a solvent and is a key
aspect of separations involving supercritical fluids, similar to conventional solvent
processing. Solubility is the maximum amount of a solute that can be solubilized in
a solvent at a given temperature and pressure, and is typically reported in terms of
mole fraction (i.e., moles of solute per mole of solvent).
16.2.2.1 Factors Affecting Solubility in Supercritical Fluids
Solubility of a solute in SC–CO2 is highly dependent on temperature and pressure,

which influence CO2 density and subsequently solvent power. The solubility of a
substance in a supercritical fluid depends on the interactions between the solute and
solvent. Increasing the pressure leads to liquid-like density of the supercritical fluid,
thus increasing the probability of interactions between the solute and solvent,
including dispersion, polar and hydrogen-bonding (McHugh and Krukonis 1986).
Consequently, solubility increases dramatically with pressure. Increasing tempera-
ture leads to a decrease in density, which is more pronounced at pressure levels
close to the critical point. However, temperature not only affects density of the
solvent but it also leads to an increase in the vapor or sublimation pressure of the
solute. Therefore, the impact of temperature on solubility depends on both effects.
A temperature increase usually leads to a decrease in solubility at low pressures due
to the stronger effect on density, whereas at higher pressures an increase in
temperature leads to increased solubility. This results in the well-known crossover
phenomenon for solubility isotherms. The pressure, above which the effect of
temperature on vapor or sublimation pressure prevails, is called crossover pressure.
The solubility isotherms converge with increasing pressure and intersect at the
crossover pressure, as illustrated in Fig. 16.5 for caffeine (Li et al. 1991), which is a
key ingredient of energy drinks, a product of increasing popularity. Separation of
solutes can be realized in processes taking advantage of the crossover pressure,
which in most cases is a pressure region rather than a specific pressure (Chimowitz
and Pennisi 1986). Below the crossover pressure, a solute can be precipitated by
increasing the temperature at isobaric conditions. This behavior, where solubility
decreases with increasing temperature is also referred to as retrograde condensa-
tion, which can be of advantage for certain separation processes. In terms of
nutraceuticals, heat sensitivity of the target compound would dictate how much
temperature can be increased to achieve retrograde condensation.
Solute properties, especially molecular weight, polarity, and vapor (or sublima-
tion) pressure, also influence solubility in SC–CO2. The solubility of substances in
SC–CO2 is affected by solute-solvent as well as solute-solute interactions, such as
hydrogen bonding. Because of the nonpolar nature of CO2, the solubility of
nonpolar components is usually higher than that for polar components with a
similar molecular weight. An increase in the molecular size of a solute decreases
600
y * 106 [molsolute/molsolvent]
500
313.15 K
400 333.15 K
353.15 K
300 368.15 K
200
100
0
0 50 100 150 200 250 300 350
Pressure [bar]
Fig. 16.5 Solubility isotherms and crossover pressure for caffeine (Solubility data from Li
et al. 1991)
the solubility in the supercritical fluid. Therefore, nonpolar solutes of low molecular
weight and high vapor pressure are preferentially solubilized in SC–CO2 at rela-
tively low density conditions, and higher density conditions are needed for larger,
slightly polar and less volatile solutes. Thus, a few rules of thumb were established
for extractability of natural substances by Stahl and coworkers (Stahl and Schilz
1976; Stahl and Quirin 1983):
“1) hydrocarbons and other lipophilic organic compounds of relatively low molecular mass
and polarity are easily extractable;
2) the introduction of polar functional groups, hydroxyl or carboxyl groups render the
extraction more difficult or impossible;
3) sugars and amino acids cannot be extracted up to 500 bar;
4) fractionation effects are possible if there are marked differences in mass, vapor pressure,
or polarity of the constituents of a mixture.”
Besides temperature and pressure, the solvent power of a supercritical fluid can be
adjusted by adding a cosolvent exhibiting interaction, such as hydrogen bonding,
charge transfer complex formation, and dipole-dipole coupling between solute and
cosolvent molecules (Ekart et al. 1993). A cosolvent can also interact with the
supercritical solvent, which can in turn affect the solubility of a solute (Ekart et al.
1993). Density of a supercritical fluid solution is increased by the addition of a
cosolvent, thereby affecting the solubility of a solute beneficially. In this context, it
is noteworthy to mention the so-called cosolvent effect (also referred to as the
entrainer effect) (Walsh et al. 1987), which refers to the dramatic increase in both
solubility and selectivity when certain cosolvents are added to a supercritical fluid
(Brunner and Peter 1982; Brunner 1983; Van Alsten and Eckert 1993; Ruckenstein
and Shulgin 2001; Ruckenstein and Shulgin 2002). The cosolvent effect can lead to
an increase in the solubility of a solute in a supercritical fluid by up to several hundred
percent (Schmitt and Reid 1986), which for some systems is more than what can be
achieved by a pressure increase of several hundred bars (Dobbs et al. 1986).
However, interactions between different solutes in a mixture can also cause a

decrease in solubility, as observed for the solubility of capsaicin, which was lower
in the presence of b-carotene (Skerget and Knez 1997). Recently, Nobre et al. (2009)
reported the solubility of a mixture of bixin and b-carotene (1:1, w:w) in SC–CO2 at
40 and 60 C and pressures of up to 350 bar and showed a substantial solubility
enhancement of up to about 265% for bixin in the presence of b-carotene. Solubilities
of solid mixtures in SC–CO2, including a discussion of the observed increase and
decrease of solubility, are reviewed by Lucien and Foster (2000). A beneficial
cosolvent effect can be exploited for improving separation processes.
The proper selection of a cosolvent can aid in separation processes by improving
selectivity and solubility. For example, the addition of up to 10% (vol%) ethanol
increased the solubility of gallic acid from 0.005 mg/kgCO2 in pure CO2 to 7.48 mg/
kgCO2 at 200 bar and 40 C (Murga et al. 2000). Furthermore, the solubility behavior
of ternary systems of lipids, cosolvents, and SC–CO2 is affected by cosolvent
€ undag and Temelli 2005; G€
addition (G€uçl€u-Ust€ € undag and Temelli 2006).
uçl€u-Ust€
The addition of cosolvents can increase as well as decrease selectivity in fraction-
ation processes of lipids using SC–CO2, if there are specific interactions between
the solutes of interest and cosolvent. For example, the addition of ethanol can
improve the deacidification of oils, as demonstrated for the separation of free fatty
acids (FFA) from palm oil (Ooi et al. 1996) or FFA and peroxides from Orange
Roughy fish oil (Catchpole et al. 2000). However, there are cases where selectivity
decreases upon the addition of a cosolvent, such as in the fractionation of shark liver
oil, where the separation of squalene from di- and triglycerides decreased with
increasing ethanol content (Catchpole et al. 2000).
16.2.2.2 Solubility Determination and Correlation
Methods to determine the phase equilibrium or solubility of a solute in a supercritical

fluid can be divided into two classes depending on how the concentration in the
phases is determined, namely analytical (direct sampling methods) and synthetic
(indirect methods) (Dohrn and Brunner 1995). Several reviews discussing the pros
and cons of the various experimental methods can be found elsewhere (Deiters and
Schneider 1986; Dohrn and Brunner 1995; Nagahama 1996). However, some chal-
lenges that can be encountered when measuring solubilities are highlighted here and
discussed briefly, using the case of b-carotene as an example. The literature data for
the solubility of b-carotene in SC–CO2 are very scattered with differences ranging
over more than an order of magnitude (G€ uçl€ € undag and Temelli 2004; de la
u-Ust€
Fuente et al. 2006). These discrepancies can be attributed to differences in the purity
of the solutes as well as the limitations of the experimental techniques used. Impu-
€ undag
rities in the sample can enhance or reduce the solubility of a solute (G€uçl€u-Ust€
and Temelli 2000), due to impurities acting like a cosolvent. Furthermore, b-carotene
can oxidize and isomerize easily during experiments or when exposed to air, which
requires special precautions during sample preparation and measurements (Cocero
et al. 2000). The crystalline nature of b-carotene affects solubility; as shown by
Sakaki (1992), the solubility of crystalline b-carotene was lower compared to that
of amorphous b-carotene, which may be expected due to the higher heat of fusion
for crystalline b-carotene. In addition to issues related to purity or cystallinity of
b-carotene, the different experimental methods applied for solubility determination
come with certain limitations and can lead to errors as well. For example, in systems
using a dynamic method reaching true equilibrium may be challenging. In static
systems or recirculation systems, attainment of equilibrium may be monitored by
adequate sensors to ascertain changes in the solute-SC–CO2 mixture over time. After
ensuring proper equilibration one has to watch out for another source of error caused
by the sampling/quantification of the solute in the supercritical phase. During off-line
sampling from a static equilibrium cell a small volume of equilibrated fluid phase is
removed, which can lead to changes in pressure and/or temperature, thereby disturb-
ing equilibrium. Furthermore, loss of solute by precipitation in the valves or sampling
lines can occur. To minimize disturbance of the equilibrium inside a static equilib-
rium cell, variable volume cells have been developed, which compensate for the
sample volume removed. On-line sampling allows solute quantification prior to the
depressurization step by means of chromatographic or spectrophotometric systems
attached to the extraction cell. However, the use of UV detectors can lead to errors
due to the potential saturation of the sensor (Mendes et al. 1999). Another method
based on a quartz crystal microbalance (QCM) was used by Saldaña et al. (2006) to
measure the solubility of b-carotene in situ, which eliminates the need for sampling.
However, the QCM technique is very sensitive to environmental factors and loading
the crystal uniformly can be challenging. In summary, solubility measurements of
unstable substances having a relatively low solubility such as carotenoids in SC–CO2
are challenging and great care must be taken to avoid the aforementioned errors and
pitfalls.
There are also numerous studies where solubility for a component is reported
based on dynamic extraction of a complex plant matrix. In this case, the “apparent
solubility” is not only based on thermodynamic solubility but also on the interactions
of the solute with the solid matrix, as well as the potential cosolvent effects of the
other components present. This approach may be more representative of the complex
multicomponent systems under consideration for process development purposes;
however, care must be taken to ensure equilibrium criteria are met. These include
having sufficient solute present to saturate the CO2, using low enough flow rates to
allow sufficient residence time, and evaluating the slope of the initial linear portion of
the extraction curve, which corresponds to the solubility-controlled region, rather
than reporting the yield obtained on a single point on the extraction curve. Compari-
son of b-carotene solubility in the binary system of pure b-carotene þ SC–CO2 to
the apparent solubility obtained in the multicomponent carrot system revealed a 5–10
fold reduction in the multicomponent system under similar temperature and pressure
conditions mainly due to interactions of b-carotene with the solid matrix (Saldaña
et al. 2006).
Bibliographic summaries covering high pressure phase equilibrium data
published up to 2002 are available in the literature (Hicks 1978; Knapp et al.
1981; Fornari et al. 1990; Dohrn and Brunner 1995; Christov and Dohrn 2002).
Furthermore, solubility data as well as correlations for numerous less volatile

substances and high boiling substances can be found (Bartle et al. 1991; Higashi
et al. 2001). An excellent collection of solubility data for more than 780 solutes
including liquids, solids, polymers, foods, drugs, nutraceuticals, dyes, pesticides,
and metal complexes has been compiled by Gupta and Shim (2007). Solubility of
numerous nutraceutical compounds in SC–CO2 has been investigated using pure
components and mostly employing static methods. Diaz-Reinoso et al. (2006)
compiled a list of pure phenolic components with antioxidant activity for which
solubility has been reported.
Accuracy and precision of measured solubility data are believed to be superior to
data predicted by correlations or models based on group contribution methods or
equations of state (EoS) in many cases. For example, the performance of six
different cubic EoS to predict the solubility of cholesterol and b-carotene in
SC–CO2 and ethane was assessed by Hartono et al. (2001). The study revealed
deviations by several orders of magnitude between the experimental data and
predicted solubility values in cases of the Van der Waals and Redlich Kwong
EoS, whereas the Mohsen-Nia–Moddaress–Mansoori EoS delivered acceptable
results (Hartono et al. 2001). However, the use of EoS requires the knowledge
of parameters, which in many cases are not experimentally accessible, such as
critical pressure and temperature. Since those parameters cannot be measured
experimentally due to decomposition of the solute, as in the case of cholesterol
and b-carotene, the critical parameters need to be estimated by group contribution
methods, which add uncertainty to the calculations (Somayajulu 1989). Further-
more, experimental data are needed to adjust interaction parameters in those EoS
models. The solubility of a mixture of bixin and b-carotene in SC–CO2, taking into
account the interaction between the solutes, was described reasonably well with a
model based on the Peng-Robinson EoS, with deviations from experimental data of
about 18.2 and 44.6% for b-carotene and bixin, respectively (Nobre et al. 2009).
Nevertheless, empirical, semi-empirical, and theoretical correlations to model
solubility data for binary supercritical fluid systems (Chrastil 1982; Ziger and
Eckert 1983; Schmitt and Reid 1985) can be used in certain cases to predict
solubility data based on available experimental data. Unlike approaches utilizing
EoS, which require properties of solutes, such as critical properties, that are often
difficult or impossible to determine experimentally, these empirical, semi-empirical,
and theoretical correlations are mostly based on accessible density and solubility
data. If the properties required for modeling are not available, the application of
property estimation methods is required, leading to uncertainties in the prediction
of solubility data. A well known density-based correlation for solubility data was
developed by Chrastil (1982) that relates the solubility S [g/L] of a solute to the
solvent density r [g/L] and temperature T [K]:
S ¼ rk eða=T þbÞ (16.1)
where k is the association constant, and a and b are defined as

DH
a¼ (16.2)
R

MGk
b ¼ ln þq (16.3)
MA þ k MG
with the ideal gas constant R, the sum of heat of solvation and vaporization DH, a
constant q, and the molecular weights of solute MA, and solvent MG. Chrastil’s
model is based on the assumption that a solute molecule (A) associates with k
molecules of the supercritical solvent (G) forming a solvate complex (AGk), which
is in equilibrium with the supercritical fluid. Plotting ln(S) versus ln(r) results in a
straight line with a slope of k and intercept of (a/T þ b). For solutes with a low
solubility, the solubility S can be converted from [g/L] into a mol fraction y in
[molsolute/molsolvent] by multiplying equation (1) with r*MA/MG, which leads to:
MA ða=T þbÞ
y ¼ rkþ1 e (16.4)
MG
Plotting ln(y) versus ln(r) also results in a straight line, now with a slope of
(k þ 1), as illustrated in Fig. 16.6 for the solubility data of caffeine in SC–CO2 (Li
et al. 1991).
The density-based correlation methods for solubility data work reasonably well
for nonpolar solutes and to a lesser extent for polar solutes at pressure levels
corresponding to high supercritical fluid densities (Gurdial et al. 1989). However,
when applied over the entire supercritical fluid density range to systems with higher
concentrations or polar solutes, simple relationships between the solute solubility
and solvent density exhibit a weakness, because the complex nature of solute-
solvent and solute-solute interactions cannot be described by these correlations.
–6
313.15 K
–8
ln y [molsolute /molCO2]
333.15 K
353.15 K
–10 368.15 K
–12
–14
–16
4.5 5 5.5 6 6.5 7
ln r [g/L]
Fig. 16.6 Logarithmic plot of caffeine solubility y [molsolute/molCO2] versus CO2 density [g/L]
(Solubility data from Li et al. 1991)
Chrastil’s model was applied to correlate literature data for solubility of lipids, such
as fatty acids, mono-, di-, and triglycerides, and fatty acid esters as well as minor
lipid components, such as b-carotene, a-tocopherol, stigmasterol, and squalene in
€ undag and Temelli 2000; G€
SC–CO2 (G€uçl€u-Ust€ € undag and Temelli 2004).
uçl€u-Ust€
Solubility data of oleic acid, b-carotene, and capsaicin in SC–CO2 were also
correlated using Chrastil’s model (Skerget et al. 1995), resulting in deviations
between experimental and correlated solubility values of up to about 30% for
both oleic acid and b-carotene. Furthermore, the authors found that the association
constant k differs between liquid CO2 and SC–CO2, which indicates that the same
model constants cannot be used for both cases.
16.3 Separation Processes
16.3.1 Extraction
Over the years, SC–CO2 extraction of a very large number of plant materials from
all over the world has been reported. Even though there are similarities regarding
the standard extraction parameters, every system is unique in terms of the specific
chemical composition and the interactions between the components. The discussion
here focuses on the extraction of lipid-based nutraceuticals (especially specialty oils
and carotenoids) and phytochemicals.
Similar to conventional extraction techniques, SC–CO2 extraction of targeted
components from a plant matrix is dictated by several parameters, including pre-
treatment of plant material, particle size, temperature, pressure, time, solvent flow
rate, and solvent-to-feed ratio. These parameters impact the efficiency of extraction,
which is determined in terms of yield and recovery of the targeted components.
Yield is the amount of total extract obtained per unit mass of starting feed material
(i.e., g extract/g feed), whereas recovery is the percentage of targeted component
originally present in the feed material recovered in the extract (i.e., [g target
compound in extract/g target compound in feed]*100). Unfortunately, there is
confusion in a number of reports presented in the literature where these basic
parameters are not reported properly.
Pretreatment of plant material may involve drying and/or grinding. Some prac-
tical considerations for SC–CO2 extraction of plant materials with respect to sample
preparation, selection of modifiers, collection methods, on-line coupling techni-
ques, means for avoiding mechanical problems, and approaches to optimization of
supercritical fluid extraction (SFE) conditions are summarized in a review by Lang
and Wai (2001). In general, plant matrices have a high moisture content, which may
be a detriment if the targeted components are nonpolar such as lipids. Since CO2
is a nonpolar solvent, the presence of high levels of moisture can act as a barrier
for CO2 and lipids to diffuse through the solid matrix. On the other hand, if the
targeted components are polar, such as phenolics, then water can act as a cosolvent
and enhance the recovery of these compounds. Selectivity of SC–CO2 extraction is

influenced by the moisture content of the plant matrix as well. For example,
presoaking the feed material with water is beneficial for the selective extraction
of alkaloids, such as caffeine from coffee beans (Peker et al. 1992). Furthermore,
some polar components of essential oils, such as thymol, linalool, and terpinene are
better solubilized in SC–CO2, when the moisture level of the samples is increased
(Stahl and Gerard 1985; Leeke et al. 2002). As well, the presence of some water
may be beneficial in terms of maintaining the permeability of cell membranes.
Therefore, there is an optimal level of moisture depending on the plant matrix and
the targeted bioactives. It should not be overlooked that water is also solubilized by
SC–CO2, despite at low levels (about 0.0067 mole fraction at 200 bar and 50 C)
(Wiebe 1941; King et al. 1992), and will be coextracted. A high level of water in
extracted lipids is not desirable since it will impact product quality negatively,
necessitating additional fractionation of the extract to remove water. Moisture level
of the feed material can be adjusted by various drying techniques such as air drying,
vacuum drying, sun drying, and freeze drying; however, care must be taken to
minimize any degradation of the bioactives due to exposure to high temperatures
over extended periods of time. Needless to say, cost is another consideration and
expensive techniques like freeze drying may not be feasible as a pretreatment prior
to SC–CO2 extraction.
It is important to understand the cellular structure of the raw material and the
location of the targeted components prior to use in extraction. For example, oil in
nuts is easily accessible by diffusion into the cellulosic structure, whereas oil in
whole seeds is not accessible due to the presence of an almost impermeable seed
coat (Sovova et al. 1994; Reverchon et al. 2000). Therefore, grinding of the plant
material to the optimal particle size to break up impermeable structures and to
expose the solute to solvent is essential for enhanced extraction efficiency. Depen-
ding on the grinding equipment used, again it is critical to avoid overheating of the
plant material and to minimize degradation of bioactives during grinding. Smaller
particle size is desirable to increase contact surface area with the solvent and to
minimize the path length that bioactives have to diffuse through to reach the bulk
phase. However, too small particles may lead to bed compaction and channeling
effects, which are not desirable.
Extraction temperature and pressure are critical parameters for optimal yield
since they dictate the solubility of bioactives in SC–CO2 as discussed above. Good
understanding of the solubility behavior of the targeted bioactive in SC–CO2 is
essential to optimize temperature and pressure for maximum solubility during
extraction.
Extraction time, solvent flow rate, and solvent-to-feed ratio are the other para-
meters that have to be optimized. Solvent-to-feed ratio and extraction time need to
be minimized for cost savings; however, sufficient residence time and amount of
solvent have to be provided to maximize bioactive recovery. The extraction kinetics
proceed through constant-rate and declining-rate periods. The freely available
solute on the particle surface is solubilized rapidly during the constant-rate or fast
extraction period, where the extraction rate is limited by equilibrium solubility.
This is represented by the initial linear portion of a typical extraction curve. Once
the solute on the particle surface is depleted, the extraction rate declines or the slow
extraction period begins, where SC–CO2 has to diffuse into the particles, solubilize
the solute, and SC–CO2 þ solute has to diffuse out to the bulk phase. This is a slow
process driven by the concentration gradient of the solute, and thus the extraction
curve approaches a constant value asymptotically. The extraction time should be
optimized to ensure the constant-rate period is completed since the yield increases
only marginally during the declining-rate period. The duration of constant-rate
period is shorter for smaller particle size due to increased contact surface area as
well as for temperature and pressure conditions corresponding to high solubility.
16.3.1.1 Lipid-Based Nutraceuticals
Specialty oils. Specialty oils have received growing interest over the past decade
mainly because of the high concentrations of polyunsaturated fatty acids (PUFA)
within their fatty acid profile, as well as the fat-soluble bioactive minor components
they contain, such as tocols (tocopherols and tocotrienols), carotenoids, sterols, and
squalene. Extensive research has demonstrated the health benefits of these compo-
nents, including beneficial effects against heart disease, cancer, and other degener-
ative diseases, as highlighted by Temelli et al. (2008) recently.
Specialty oils include nut oils (almond, hazelnut, peanut, pecan, pistachio, and
walnut), seed oils (borage, flax, evening primrose, grape, pumpkin, and rosehip),
cereal oils (amaranth, rice bran, oat, and wheat germ), and fruit and vegetable oils
(buriti fruit, carrot, olive, and tomato). Conventional methods of extracting these
oils involve mechanical pressing and/or solvent extraction. Mechanical pressing
at temperatures below 60 C (also referred to as cold pressing) is used exten-
sively; however, this technique suffers from the fact that residual oil in the press
cake is quite high (~5%), leading to low oil recoveries. Solvent extraction is
followed by heat treatment for solvent recovery, which compromise oil quality.
Therefore, SC–CO2 extraction of specialty oils gives a superior product with high
efficiencies. An extensive review of the subject by Temelli et al. (2008) gave a
detailed discussion of the effect of different extraction parameters on recovery as
well as the composition and quality of the extracted oils from nuts, cereals, seeds,
and fruits and vegetables in comparison to those obtained using traditional
techniques.
In general, temperatures of 35–80 C and pressures of up to 690 bar have been
used for SC–CO2 extraction of specialty oils, resulting in oil recoveries of greater
than 95%, as in the case of hazelnut (Bernardo-Gil et al. 2002), peanut (Goodrum
and Kilgo 1987), walnut (Oliveira et al. 2002), evening primrose (Favati et al.
1991), grape seed (Cao and Ito 2003), rice bran (Shen et al. 1996), and wheat germ
(Taniguchi et al. 1985) oils. The color of the extracted oils varied with extrac-
tion conditions, depending on the extent of extraction of pigments (Palazoglu
and Balaban 1998). Compositional analysis of all the specialty oils demonstrated
very high levels of oleic (C18:1), linoleic (C18:2), and linolenic (C18:3) acids,
accounting for more than 90% of the total fatty acids for the seed oils (Temelli et al.
2008). The tocopherol contents of walnut (Oliveira et al. 2002) and wheat germ
(Gomez and de la Ossa 2000) oils extracted with SC–CO2 were higher than those
extracted with hexane. The SC–CO2-extracted oil was clearer, thus requiring less
refining compared to hexane-extracted oil (Bernardo-Gil et al. 2002; Oliveira et al.
2002; Lopes and Bernardo-Gil 2005). Squalene was recovered from rice bran (Kim
et al. 1999), and Amaranthus grain (He et al. 2003) using SC–CO2 but to a much
lesser extent compared to that obtained by using solvent extraction.
Even though the majority of the studies used neat SC–CO2, several researchers
added ethanol (up to 10%, wt%, or mol%) as a cosolvent into SC–CO2 and achieved
higher yields. For example, the extraction yields of pistachio (Palazoglu and
Balaban 1998) and sesame (Odabasi and Balaban 2002) oils increased substantially
upon ethanol addition over that obtained by neat SC–CO2. On the other hand, the
phosphorus content of oat oil increased to 80 ppm, probably due to the extraction of
polar phospholipids in the presence of ethanol (Fors and Eriksson 1990). Although
ethanol addition may offer some advantages in terms of increased yield, it negates
the major advantage of SC–CO2 extraction of avoiding the use of organic solvents
and heat treatments since ethanol removal from the extract and the residual meal
requires the use of heat. Unfortunately, this aspect is usually overlooked in many
studies.
Carotenoids. Carotenoid pigments responsible for the bright yellow, orange, and
red colors of various plant materials are comprised of two main classes, carotenes
and xanthophylls, which are C40 polyunsaturated hydrocarbons and their oxyge-
nated derivatives, respectively. Because of the provitamin A activity of b-carotene
and the potent antioxidant activity of lycopene associated with reducing the risk of
prostate cancer (FDA 2005), numerous studies have focused on their extraction
from carrots and tomatoes, respectively, using SC–CO2. Use of by-products of
conventional processing industries (i.e., carrot cubes, carrot juice, tomato juice,
tomato paste, etc.) as the raw material for the recovery of these valuable carotenoids
is very attractive. In addition, other sources such as marigold and algae are also
receiving increasing attention as raw materials for carotenoids. Extraction of
carotenoids as well as oleoresins and capsaicinoids using SC–CO2 from various
sources is summarized in Table 16.2.
It is necessary to dry carrots and tomatoes prior to extraction due to their high
moisture content (80% and 95%, respectively) and samples are mostly freeze dried
for research purposes. The SC–CO2 extraction yields of a- and b-carotene increased
with decreasing moisture content of carrot feed material, whereas the yield for
lutein was higher at higher moisture levels since water can act as a cosolvent for the
relatively polar oxygenated lutein (Sun and Temelli 2006). From tomatoes contain-
ing 50–60% moisture, only trace levels of lycopene were extracted (Vasapollo et al.
2004).
As expected, lycopene extraction yield increased with pressure at 66 C and
with temperature at a pressure level of 450 bar, which is above the crossover
pressure of isotherms (Vasapollo et al. 2004). Lycopene recovery from tomato
skin reached 96% at 110 C in 40 min (Ollanketo et al. 2001); however, caution
Table 16.2 Extraction of carotenoids and other compounds from tomato, carrot, paprika, apricot, marigold, and algae
16
Feed material/characteristics Pressure (bar) Temperature ( C) Co-solvent Yielda/Recoveryb of target compound Reference
Tomato Lycopene b-Carotene
Skin þ seed 138–483 32–86 No 61b (Rozzi et al.
2002)
Skin þ seed 250, 300 60, 80 No 80b* 88b* (Sabio et al.
2003)
Skin þ pulp 77–281 40 No 42b 98b (Gomez-Prieto
et al. 2003)
Skin 200–500 40–100 No 94b (Topal et al.
2006)
Skin þ seed 460 80 No 31.4a (Vagi et al.
2007)
Dried powdered skin 400 60–110 Acetone, n/a highest recovery (Ollanketo et al.
dichloromethane, with acetone as 2001)
hexane, co-solvent
methanol,
ethanol, water;
added to sample
Skin þ seed 172–276 40–80 Chloroform, hexane 64a**, 84b** 35a**, 93b** (Cadoni et al.
(1 mL); added to 1999)
sample
Saponified tomato 370–530 47–63 EtOH (16%) 27.4a (Huang et al.
powder 2008)
Bioseparation of Nutraceuticals Using Supercritical Carbon Dioxide
Skin þ seed 250–350 45–75 EtOH, water, olive 73.3b (10% (Shi et al. 2009)
oil as single, EtOH þ 10%
binary, ternary olive oil)
modifier: 0, 5 or
10%
added to sample
Dried paste 200–300 35–65 No 20b 40b (Baysal et al.
2000)
EtOH; 5, 10, 15% 50b (5% EtOH) 50b (5% EtOH)
369
with SC–CO2
(continued)
370
Dried tomato 335, 450 45, 66 No 10b (Vasapollo
et al. 2004)
450 66 Hazelnut oil 30b
10% w/w added to
sample
Carrot b-Carotene
Freeze dried 120–327 40, 50 No 47.8b (Saldaña et al.
2006)
Pulp press cake 207, 276, 345 40, 55, 70 EtOH; 0, 5, 10% w/w 99.5b (Vega et al.
with SC–CO2 1996)
Particles 276–551 40, 55, 70 No 63b* (Sun and
Temelli
2006)
Moisture: 0.8–85% Canola oil; 2.5, 5% 147b* **
Particle sizes: 0.25–2 mm w/w with
SC–CO2
Paprika/red pepper Carotenoids Oleoresin (capsaicin)
Dried powder 300–500 60–80 No 62b (Ambrogi et al.
Capsicum annuum linear stage: 2002)
b-carotene þ free
xanthophylls
nonlinear stage:
xanthophyll esters
red pepper 150–230 40 No 5200a (oleoresin) (Duarte et al.
Capsicum frutescens 252a (capsaicinoids) 2004)
Paprika ground 450 50 No 90b (Nagy and
Capsicum annuum Simandi
Moisture: 5–30% 2008)
Oil cont. 5–27% w/w
(continued)
F. Temelli and B. Seifried
16
Chili powder CO2 45–75 Methanol, acetic 76a (capsaicinoids) (Peusch et al.
Capsicum frutescens Density: acid and water; (co-solvent water) 1997)
0.75–0.9 20 mL added to
g/mL 70 and 250 mg
Paprika powder sample
Capsicum annuum
Red pepper 320–540 40 No Red pigments: 88b (Uquiche et al.
Capsicum annuum Yellow pigments: 55b 2004)
Pelletized flakes moisture:
4% w/w
Apricot b-Carotene
Pomace 133–473 43–77 EtOH; 2–28% v/v 9.8a (Sanal et al.
with SC–CO2 2005)
Marigold Terpene/Carotenoid Oleoresin
Dried flowers 120, 150, 200 26, 30, 40 No 2,100a (Campos et al.
2005)
Flowers 120–200 20–40 No 3,540a (Danielski et al.
2007)
Dried flowers 300, 500, 689 50 EtOH; 0–20% v/v Faradiol esters 8,300a (Baumann et al.
with SC–CO2 650a 2004)
Dried petals 145–455 33.2–71.8 soybean oil; lutein 96,200a (Ma et al. 2008)
0–12.1% w/w
with SC–CO2
Bioseparation of Nutraceuticals Using Supercritical Carbon Dioxide
Algae Total Carotenoids Astaxanthinc

Canthaxanthind
Dried powder 300–500 40–80 no 22,800ac (Thana et al.
Haematococcus pluvialis 2008)
Freeze dried powder 200–500 40, 50, 60 EtOH; 5% mol 963a (D. salina)
(continued)
371
372
Nannochloropsis gaditana 186a (Synechococcus) (Macias-
Synechococcus sp. Sanchez
Dunaliella Salina 289a et al. 2008)
(Nannochloropsis)
Freeze dried, crushed 150–350 40, 55 No 6ac (Mendes et al.
Chlorella vulgaris 24ad 1995)
Dried powder 300–500 50–80 Soybean oil, olive 25bc (pure CO2) (Krichnavaruk
Haematococcus pluvialis oil, EtOH; 36bc (soybean oil) et al. 2008)
0–12% v/v 51bc (olive oil)
d.m. dry matter
*Skin only
**No cosolvent
***Compared to soxhlet
a
Yield (mg/100 g feed)
b
Recovery (%) [100 (g solute in extract/g solute in feed)]
c
Astaxanthin
d
Canthaxanthin
F. Temelli and B. Seifried
should be exercised in processing at such high temperatures to avoid any degrada-

tion of lycopene.
Because of the very low solubilities of b-carotene and lycopene in SC–CO2,
cosolvents have been incorporated in an attempt to increase extraction yields of these
carotenoids. Several of the studies mixed the cosolvent with the feed material prior to
extraction (Ollanketo et al. 2001; Vasapollo et al. 2004; Shi et al. 2009). The problem
with this approach is that the cosolvent concentration would change throughout the
extraction period as SC–CO2 flows through the extraction cell and solubilizes the
cosolvent together with the carotenoids. Therefore, it is desirable to mix the cosolvent
continuously into SC–CO2 to maintain a constant concentration. On the other hand,
Ollanketo et al. (2001) tested cosolvents like acetone, ethanol, methanol, hexane,
dichloromethane, and water and all except water increased lycopene recovery. Even
though such cosolvents can be used for analytical purposes, they would not be accept-
able for large-scale processing due to the demand for “natural” products, especially in
the nutraceutical market. Since carotenoids are fat-soluble pigments, vegetable oils were
also used as a cosolvent in SC–CO2 by mixing the oil with feed material (Vasapollo et al.
2004; Shi et al. 2009) and by continuous addition to SC–CO2 (Sun and Temelli 2006).
Vasapollo et al. (2004) used hazelnut oil for lycopene extraction and achieved a
lycopene recovery of 30% in 8 h. Upon continuous addition of canola oil into
SC–CO2 for the extraction of carotenoids from carrots, Sun and Temelli (2006) obtained
more than a two‐fold increase in yield at 288.0–846.7 mg/g and 333.8–900.0 mg/g feed
for a- and b-carotene, respectively. The major advantage of using vegetable oil as a
cosolvent is the elimination of organic solvents and the fact that vegetable oil enriched in
carotenoids can be used in a variety of applications without further treatment.
16.3.1.2 Phytochemicals
Research outcomes demonstrating the health benefits associated with various

phytochemicals have triggered the growing demand for “natural” antioxidants
and other plant extracts of different bioactivities, which in turn resulted in a
tremendous level of research activity focusing on the extraction, characterization,
and functionality determination for a range of plant materials from all over the
world. A comprehensive compilation of such works is beyond the scope of this
chapter; however, Diaz-Reinoso et al. (2006) have done an excellent job in such
an attempt.
Plant extracts are complex mixtures comprised of different classes of compounds,
including essential oils, esters, terpenes, fatty acids, waxes, resins, pigments, pheno-
lics, etc. In general, these compounds are present in relatively low concentrations,
where the total extract yield may be <5% (Diaz-Reinoso et al. 2006). Essential oils in
general are more volatile and can be easily extracted at relatively low pressures of
less than 150 bar at temperatures of 40–50 C using SC–CO2 (Reverchon 1997).
However, obtaining a total extract would require the use of higher pressures (up to
550–600 bar). Selective extraction of different groups of compounds can be
achieved by proper optimization of extraction conditions as well as the separation

conditions as discussed later.
The plant materials used for SC–CO2 extractions encompass herbs, spices,
aromatic, and medicinal plants. These plants are especially rich in bioactives with
antioxidant activities as well as other physiological benefits including antimicro-
bial, antiviral, anti-inflammatory and antimutagenic activities (Moure et al. 2001;
Sacchetti et al. 2005; Srinivasan 2005). Previous reviews focused on the
SC–CO2 extraction of bioactives from traditional Chinese medicinal plants (Li
2008), Latin American plants (Meireles 2008), and a variety of spices (Mukho-
padhyay 2008). In addition, by-products of the conventional processing indus-
tries are receiving growing attention for the recovery of such bioactives as
reviewed by Herrero et al. (2006). For example, grape seeds and skins as by-
products of the juice and wine industries are being used for the recovery of
catechins and other phenolics for their antioxidant activity (Palma et al. 1999;
Murga et al. 2000). Supercritical fluid extraction (SFE) of plant materials using
SC–CO2 is advantageous compared to conventional methods, such as hydro-
distillation or organic solvent extraction using acetone, methanol, ethanol, hexane
or others because generally the extracts obtained with SC–CO2 are more potent.
For example, SC–CO2 extracts of marjoram were more potent in inhibiting the
growth of fungi and bacteria than those obtained using ethanol (Vagi et al. 2005).
16.3.2 Fractionation
It may be desirable to further fractionate an extract in order to obtain the targeted

bioactives in a more concentrated form. The supercritical technology offers flexi-
bility in this regard as different fractionation protocols can be employed. Such
protocols can be categorized into three approaches requiring different types of
equipment.
1. Fractional extraction: Fractions can be collected as a function of time throughout
an extraction. In addition, extraction conditions can be programmed to change
over time. For example, pressure can be kept low initially to extract low
molecular weight compounds, and increased to a higher level once the low
molecular weight compounds are depleted to extract the higher molecular
weight compounds. As well, neat CO2 can be used first to extract neutral
compounds, followed by the injection of a polar cosolvent to recover more
polar components in the second fraction. One extractor and one separator are
sufficient for fractional extraction.
2. Fractional separation: In this case, multiple separators are necessary to collect
different fractions of an extract. Temperature and pressure conditions
corresponding to a very high solvent density are used to recover a total extract
followed by adjusting the temperature/pressure conditions in the separators for a
stage-wise drop of solvent density so that the highest molecular weight/lowest
volatility fraction falls out of solution first and is collected in the first separator,
followed by collection of lower molecular weight/higher volatility fractions in
the subsequent separators with a further drop in density in each separator.
3. Column separation: continuous feed of a liquid mixture is what distinguishes a
column operation from the first two cases, wherein a batch of solid material is
used as the feed. These are usually packed columns with independently heated
and controlled sections, designed to create a thermal gradient along the column
to take advantage of retrograde condensation so that an internal reflux can be
generated for increased efficiency of separation. Countercurrent contacting of
the liquid feed with SC–CO2 results in the separation of the liquid feed into
extract and raffinate streams coming out of the top and the bottom of the column,
respectively. Brunner (2009) provided an overview of the design considerations
for such systems. The following sections will focus on the different applications
of these three approaches to various nutraceuticals.
There are additional recent developments in the area of coupling membrane
technologies with supercritical technology, which may be the fourth approach in
fractionation to achieve difficult separations. It may be possible to separate those
components with similar solubilities in SC–CO2 by attaching a membrane unit after
an extraction or a fractionation unit. For example, squalene, which is receiving
increasing attention as a nutraceutical, and oleic acid have similar solubilities in
SC–CO2 due to the nonpolar nature of squalene, although squalene (C30) is a much
larger molecule than oleic acid (C18) (G€ uçl€ € undag and Temelli 2004). How-
u-Ust€
ever, such a large molecular weight difference can be an advantage for membrane
separation. With the use of an appropriate membrane these two lipids solubilized in
SC–CO2 can be separated under supercritical conditions (Ruivo et al. 2008).
16.3.2.1 Lipid-Based Nutraceuticals
Polyunsaturated fatty acids. Fractionation of polyunsaturated fatty acids (PUFA)

obtained from marine sources is the main focus of this section even though
processing of specialty oils containing g-linolenic acid (GLA or C18:3o6) or
a-linolenic acid (ALA or C18:3o3) using near critical fluids, such as CO2, propane,
and dimethylether, has also been investigated (Catchpole et al. 2009). Processing of
fish oils using supercritical fluids has been reviewed recently (Eltringham and
Catchpole 2008).
Ever since the early epidemiological studies showing a correlation between high
fish consumption and decreased fatal heart attacks (Strøm and Jensen 1951) and
heart diseases (Dyerberg et al. 1975), there has been growing interest in fish oils and
additional research has demonstrated the importance of long chain polyunsaturated
fatty acids (LC-PUFA) in terms of the health benefits of fish oil (Horrocks and Yeo
1999). The two most important LC-PUFA in fish oil, namely eicosapentaenoic acid
(C20:5o3, EPA) and docosahexaenoic acid (C22:6o3, DHA) belong to the family
of omega-3 fatty acids, where the first double bond begins with the third carbon
atom from the methyl end of the molecule. Noteworthy, omega-3 fatty acids such as
EPA and DHA obtained from fish oils seem to be superior in reducing the rates of
all-cause mortality, cardiac, and sudden death, and possibly stroke compared to
ALA found in vegetable oils (Sanderson et al. 2002; Wang et al. 2006). Omega-3
PUFA obtained from marine sources is mainly consumed in the form of either
triglycerides (TG) or fatty acid ethyl esters (FAEE). However, there are controver-
sial results in studies concerning the absorption of EPA in these two forms. Early
studies claimed that absorption of EPA and DHA when ingested as FAEE was
lower than that as TG (El Boustani et al. 1987; Lawson and Hughes 1988), whereas
in later studies the absorption levels were found to be similar (Nordoy et al. 1991;
Krokan et al. 1993). Alternative sources for EPA and DHA besides fish oil include
microbial or single cell oil, with some microbial strains (Crypthecodinium cohnii)
producing oils containing up to 50% DHA of the total fatty acids (Ratledge 2004;
Ward and Singh 2005). The EPA and DHA contents of fish oils depend on
numerous factors, such as species, origin, catch season, and feed of the fish, with
EPA and DHA levels of up to about 20 and 35%, respectively (Gruger et al. 1964;
Özogul and Özogul 2007; Visentainer et al. 2007).
In order to be suitable for human consumption crude fish oil needs to be refined
and purified. Refining processes of fish oils aim mainly at removing undesirable
compounds such as free fatty acids, off aromas, and peroxides. Additionally, high
value compounds such as squalene, diacyl glyceryl ethers (DAGE), vitamin A, and
vitamin D are targets of refining processes as well (Eltringham and Catchpole
2008). Fish oils in their TG form with EPA and DHA concentrations of up to
300 mg/g can be obtained by winterization, blending, solvent crystallization, and/or
vacuum distillation (Ackman 1988). For applications requiring higher EPA and
DHA levels, fish oil TG are hydrolyzed to liberate the fatty acids, which can then be
converted into fatty acid ethyl esters (FAEE) or methyl esters (FAME). The esters
are subsequently fractionated by various methods to obtain higher concentrates of
EPA and DHA (Shahidi and Wanasundara 1998) by means of vacuum or short path
distillation (Ackman et al. 1973; Breivik et al. 1997), solvent crystallization (Lee
2003), or urea complexation (Senanayake and Shahidi 2000; Gámez-Meza et al.
2003; Zuta et al. 2003). These conventional methods often require the use of
flammable and toxic organic solvents or elevated process temperatures, which
can lead to polymerization and degradation of thermally labile PUFA (Staby and
Mollerup 1993; Lee and Foglia 2001). Selective enzymatic hydrolysis of fish oil
TGs using fatty acid-specific lipases can be used as well for the separation and
concentration of EPA and DHA from other fatty acids present in fish oils (Linder
et al. 2002; Ramı́rez Fajardo et al. 2006).
Because of the drawbacks of the various aforementioned conventional techni-
ques, processing of heat sensitive materials containing PUFA using SC–CO2 offers
numerous advantages (Mishra et al. 1993). Fractionation of fish oil FAEE using
SC–CO2 has been investigated by numerous research groups. In an attempt to
fractionate fish oil FAEE in semi-continuous equipment, Eisenbach (1984) used a
packed column equipped at the top with a “hot finger” held at 90 C to generate a
reflux to take advantage of retrograde condensation. This method was suitable for
separating fractions of different carbon lengths but showed limited ability in

separating EPA and DHA from other C20 and C22 fatty acids, respectively. Nilsson
et al. (1988) used a packed column equipped with individually controlled heaters to
create temperature zones, ranging from room temperature at the bottom of the
column up to 100 C at the top. By applying incremental pressure programming to
the column, it was possible to reduce the required temperature gradient for the
fractionation (Nilsson et al. 1989).
More recently research activities focused mostly on continuous fractionation
of fish oil esters using countercurrent columns. Continuous fractionation of fish
oil FAEE was studied using a pilot plant countercurrent packed column (Fleck
et al. 1998; Riha and Brunner 2000) to separate low molecular-weight compo-
nents (LMC; C14 to C18) from high molecular-weight components (HMC; C20 to
C22), obtaining concentrations for HMC in the raffinate of greater than 95 wt% at
a recovery of HMC of greater than 95%. Continuous fractionation of squalene
from shark liver oil in a countercurrent packed column resulted in squalene with
up to 99% purity (Catchpole et al. 1997). Semi-continuous separation of a
commercial mixture of fish oil FAEE containing 64% EPA + DHA was investi-
gated by Perretti et al. (2007) using a column filled with Raschig rings, employ-
ing three temperature zones along the column held at constant temperatures of
40 C, 50 C and 60 C from bottom to top. In this study, 450 g of FAEE mixture
were loaded onto the column (L ¼ 3 m, ID ¼ 3 cm) at three different column
heights and subsequently fractionated at pressures ranging from 100 to 300 bar
and CO2 flow rates ranging from 2.5 to 10 kg/h over a duration of 2 h. By
fractionating the FAEE mixture at 150 bar with a CO2 flow rate of 5 kg/h, the
total content of EPA and DHA esters was increased from 64% in the feed mixture
to about 82% in the raffinate (fractions collected from the bottom of the column).
However, the yield of raffinate was less than 10% and the EPA/DHA ratio
changed from 1.61 for the initial mixture to 0.65 for the raffinate due to EPA
being extracted preferentially over DHA from the column.
Simulation of a countercurrent column was carried out by Gironi and Maschietti
(2006) after studying the fractionation of a natural mixture of fish oil FAEE by
means of a semi-continuous single stage process. The results obtained by the single-
stage process showed that in order to achieve the best compromise between
selectivity and solubility, the operating conditions resulting in a SC–CO2 density
between 570 and 595 kg/m3 should be selected at relatively high temperatures. For
example, at 70 C and 167 bar, an oil with more than 80% of EPA and DHA ethyl
esters was obtained in the process, with a recovery of about 40% in the raffinate.
With the results obtained in their semi-continuous fractionation experiments as well
as available pilot plant data (Riha and Brunner 2000), Gironi and Maschietti (2006)
satisfactorily modeled a continuous fractionation process, which enabled them to
study various process parameters, including the number of theoretical stages, reflux
ratio, and solvent-to-feed ratio. Based on the simulation of the continuous counter-
current process, it seems feasible to obtain a raffinate with 95% (wt%) of C20 þ C22
FAEE, together with 95% recovery of these valuable compounds, by operating
a multistage column (from 11 to 30 stages) with a reflux ratio of 2.5–5.2 and
a solvent-to-feed ratio in the range of 90–150. The results of such simulations based
on mass balances and equilibrium calculations are useful to better understand the
influence of various parameters on the theoretical performance of a column.
However, in order to calculate the required height and diameter of a packed
countercurrent column, mass transfer, fluid flow and flooding behavior considera-
tions should be taken into account, thus requiring more sophisticated calculations,
involving properties such as density, viscosity, and diffusivity of both the liquid and
gaseous phases as well as interfacial tension, which are often challenging to
measure or accurately predict (Blaha-Schnabel et al. 1996; Stockfleth and Brunner
2001; Ruivo et al. 2002; Martı́n and Cocero 2007; Brunner 2009). Therefore, more
research is needed, evaluating both the experimental and theoretical aspects in
order to obtain reliable data and to develop further the correlations and models
used for calculating such complex fractionation processes.
Another novel approach different from the ones described above in order to
concentrate EPA and DHA in the form of free fatty acids (FA), or FAEE, is the
combination of urea fractionation and supercritical processing. The first steps in
the development of the process include the use of urea dissolved in hot ethanol and
mixing it with FA or FAEE, thereby forming a complex with their long hydrocar-
bon chain, which depends on temperature and degree of unsaturation (Eltringham
and Catchpole 2008). Then, separation is achieved by cooling until precipitation of
urea complexes occurs. However, a potentially better way was suggested, where
SC–CO2 containing fatty acids or esters is passed through a bed of urea, which leads
to complexation of the monounsaturated and saturated fatty acids, thereby separat-
ing them from the polyunsaturates (Catchpole et al. 2009). Furthermore, the
fractionation process was improved by combining urea fractionation as described
above using ethanol with supercritical antisolvent fractionation (SAFT) (Catchpole
et al. 2002). In the combined process, urea complex formation takes place in
ethanol, followed by separation of the urea complex using a filter and feeding the
filtrate containing fatty acids or ethyl esters, solvent (ethanol and water), and
dissolved urea into a pressurized apparatus together with CO2. By mixing the
solution with CO2, the antisolvent effect of CO2 leads to precipitation of urea and
water, which are separated in the first separator, while fatty acids and ethanol can be
recovered separately in subsequent separators through sequential pressure reduc-
tion steps. With this patented method, it seems possible to obtain PUFA concen-
trates containing over 90% o-3 FA.
Tocopherols and phytosterols. Tocopherols and tocotrienols, which belong to
the family of Vitamin E compounds, have antioxidant activity and are used
extensively in the food, cosmetic, and pharmaceutical industries. Deodorizer distil-
late (DOD), a by-product of the vegetable oil refining process, is rich in tocopherols.
Conventional methods to isolate tocopherols by means of vacuum or molecular
distillation can lead to degradation of these thermolabile compounds. Supercritical
fluids could therefore offer an alternative to conventional methods, by taking
advantage of mild processing conditions. Separation of tocopherols using
SC–CO2 was studied in a countercurrent packed column (Fang et al. 2007). The
feed material was soybean DOD, which had been subjected to methyl esterification
and methylation to convert all free fatty acids and triglycerides into FAME, thereby
improving their solubility in SC–CO2. This pretreatment step facilitated the
removal of most sterols by cooling and crystallization, due to the low solubility
of sterols in FAME. The fractionation process consisted of two steps, where the first
step was to feed the pretreated DOD into a fractionation column and to extract the
FAME, while concentrating the tocopherols in the raffinate. The second step was to
increase the pressure to extract the tocopherols from the impurities remaining in the
column. To find the optimum conditions for the first step, the pretreated DOD
containing about 70 wt% FAME and 10–15 wt% tocopherols was fed continuously
into a packed column operated at 140, 160, and 180 bar with a linear temperature
gradient along the column ranging from 40 C at the bottom to 75 C at the top. Fang
et al. (2007) showed that operating the column at 180 bar resulted in the highest
extraction yield for FAME of over 70% requiring the shortest time among all
conditions investigated. However, the higher pressure also led to a threefold
increase in tocopherols being extracted from the column compared to that at
140 bar, which is clearly a consequence of higher tocopherol solubility at higher
pressure. Furthermore, increasing the pressure caused a decrease in the selectivity,
which is disadvantageous for the separation. The composition of tocopherols in the
fractions obtained from the column was influenced by pressure as well, with mainly
a-tocopherol being extracted at the low pressure conditions, whereas at 180 bar all
four isomers were found in the fractions in proportions similar to that of the feed
material. Thus, a pressure of 160 bar was selected to further optimize the feed
location, temperature gradient, and solvent-to-feed ratio. The second step of this
patented process was found to work best at a final pressure of 200 bar, which
resulted in a high tocopherol content (>50%) and tocopherol recovery (about 80%)
(Fang et al. 2007).
Fractionation of canola DOD was investigated by G€uçl€u-Ust€ € undag and Temelli
(2007) to enrich sterols and tocopherols using a packed column in a semi-continu-
ous process. The effects of pressure, temperature, and temperature gradient along
the column on the extraction yield and selectivity of the fractionation were studied.
Highest extract yields were obtained at the highest pressure (250 bar) and lowest
temperature (70 C) studied. However, the tocopherol concentration of the fractions
was higher at 250 bar and 100 C, with the content of tocopherol in the fractions
collected being lower than that of the feed material. While enrichment of tocophe-
rols in both the collected extract fractions and the column residue was less effective,
the sterol content increased by about four‐fold from 11% in the feed to 40.4%
(determined as % GC area) in the column residue after fractionation of canola DOD
using SC–CO2 at 250 bar and a temperature gradient of 70–100 C along the column
€ undag and Temelli 2007). In another study, olive pomace, a byproduct
(G€uçl€u-Ust€
of olive oil production, was extracted with SC–CO2 using a batch extractor operated
at 350 bar and 50 C, followed by two depressurization steps into subsequent
separators (Ibanez et al. 2000). The first separator in line after the extractor operated
at pressures ranging from 100 to 200 bar and various temperatures between 40 C
and 60 C contained major proportions of triglycerides, waxes, and sterols, whereas
the tocopherols were enriched in the second separator set at 10 bar and 25 C.
Phase equilibria measurements with soy DOD and crude palm oil (CPO) using a
static analytical method were carried out to assess the feasibility of fractionation of
tocochromanols (tocopherols and tocotrienols), sterols, and squalene by better
understanding their distribution coefficients (Gast et al. 2005). The distribution
coefficient Ki, describing the ratio of concentrations between the gaseous phase and
liquid phase (Ki ¼ yi/xi), is a measure to assess the feasibility of separation in a
fractionation column. A distribution coefficient of Ki 1 occurs when the con-
centrations of a component in liquid and gaseous phases are nearly equal, which
means that separation in a fractionation column is impossible. Substances that are
enriched in the gaseous phase, as evidenced by a Ki > 1, can be collected in the top
fraction, whereas substances with a Ki < 1 can be withdrawn with the liquid
raffinate at the bottom of the column. The distribution coefficients for sterols and
tocopherols in soy DOD investigated at 200–260 bar and 80 C were in the range of
about 0.35–0.4 and 0.75–1 for sterols and tocopherols, respectively, indicating that
the sterols stay preferably in the liquid phase, whereas the separation of tocopherols
is more challenging. The distribution coefficient for squalene in soy DOD was
above 3.8, indicating that enrichment in the gaseous fractions collected at the top of
the fractionation column was feasible. On the other hand, phase equilibrium studies
with CPO at 20–300 bar and 67 C revealed distribution coefficients for triglycer-
ides, tocochromanols, and FFA in the range of 0.6–0.8, 2.5–4.5 and 4.5–6.5,
respectively. Therefore, FFA and tocochromanols can be separated easily from
the CPO triglycerides in a fractionation column. The results obtained in the phase
equilibrium studies were reflected in the subsequent column fractionation experi-
ments, where in the case of soy DOD, a top product rich in squalene (18 wt%) and a
bottom product enriched in sterols (>50 wt%) were obtained (Gast et al. 2005). In
order to assess the feasibility of a separation process, phase equilibrium studies to
obtain distribution coefficients are very important and may require investigations
on a case by case basis since they depend largely on the nature and composition of
the starting feed material. Once distribution coefficients are available for a given
system the separation process can be reasonably well described by short-cut
methods to calculate the optimum solvent-to-feed ratio, reflux ratio, and the number
of theoretical stages.
Carotenoids. Fractional extraction of paprika using SC–CO2 while monitoring
the carotenoid concentration in the supercritical phase by means of a near infrared
visible detector was studied to optimize the operating conditions (Ambrogi et al.
2002). The influence of pressure on extraction performance was strong, with extract
and carotenoid recovery being faster at 500 bar compared to 300 bar. Additionally,
the amount of CO2 required to recover most of the carotenoids was three times more
at low pressure than higher pressure (300 bar). Ambrogi et al. (2002) also compared
the extraction performance of the pigments from paprika powder to that from
paprika oleoresins, with the powders resulting in a slower extraction due to the
effect of the cell structure retaining the pigments. Fractionation of carotenoids was
achieved by collecting fractions over time, where two stages of extraction were
identified during the experiments, namely the linear and nonlinear stages. In the first
stage, the extract was rich in free xanthophylls and b-carotene (about 50% w/w),
whereas in the second stage the concentration of xanthophyll esters was increased
in the extract. A further development of the aforementioned fractional extraction
process was to use a combined extraction-adsorption process for separating the red
pigments (mainly esterified carotenoids) from the yellow pigments (mainly free
carotenoids) using silica gel as adsorbent (Ambrogi et al. 2003). The authors noted
that b-carotene (free carotenoid) and cryptoxanthin (monoester) were poorly
adsorbed onto the silica gel, whereas capsanthin (diester) was efficiently retained.
However, recovery of the adsorbed carotenoids from silica gel using SC–CO2 may
be challenging, thus other organic adsorbants such as cyclodextrin or cellulose may
be promising candidates for further investigations.
16.3.2.2 Phytochemicals
Phytochemicals such as flavonoids, stilbenoids, essential oils and compounds with

antioxidative and antimicrobial activity are often heat-sensitive compounds. There-
fore, extraction and subsequent fractionation using SC–CO2 have been studied for
many phytochemicals. One possible pathway for fractionation using SC–CO2 is the
fractional separation, where an extract obtained at high pressures is subsequently
depressurized and/or subjected to temperature changes in a cascade of separators to
collect various fractions based on solubility. Flavonoids, such as those found in
propolis are nearly insoluble in pure SC–CO2. However, propolis can be readily
dissolved in ethanol, to form a tincture that contains up to 40 wt% propolis, which
can be further fractionated using SC–CO2 as both an antisolvent and a solvent
(Catchpole et al. 2004). In this process, the propolis tincture is mixed with SC–CO2,
which causes the antisolvent effect to precipitate high molecular mass components
and to dissolve ethanol together with soluble components, such as flavonoids. This
extract is then further separated in two steps by depressurization in two separators,
where a concentrated flavonoid fraction with a concentration of 20–35 wt% is
collected as the primary product, and an essential oil/ethanol fraction is produced
as a secondary product.
Phytochemicals found in rosemary and sage, including components that have
antioxidant activity such as carnosic acid, carnosol, rosmarinic acid, rosmanol,
epirosmanol, and isorosmanol, are used as natural ingredients in food, drug, cos-
metic, and medicinal applications. Sage herb (Salvia officinalis L.) was extracted
with SC–CO2 at 250–350 bar and 100 C using ethanol as cosolvent (Dauksas et al.
2001). The extract was subsequently fractionated in separators by stepwise depres-
surization and temperature change to obtain fractions of varying antioxidant activ-
ity. The antioxidant activity of the fractions depended largely on the pressure and
temperature settings in the separators.
Essential oils comprised of hydrocarbon monoterpenes, oxygenated monoter-
penes, hydrocarbon sesquiterpenes, and oxygenated sesquiterpenes can be reco-
vered by hydro-distillation from seeds, roots, flowers, herbs, and leaves. However,
this simple process suffers from several drawbacks, including thermal degradation,
hydrolysis, and loss of water‐soluble compounds, which are detrimental to the
quality of the final product. Therefore, extraction of essential oils using SC–CO2
technology offers great advantages, which also allows relatively simple fraction-
ation of extracts (Reverchon 1997; Reverchon and De Marco 2006). Isolating
essential oils using SC–CO2 involves extraction of plant matter at relatively mild
conditions at about 90–100 bar and 40–50 C, where most essential oil components
are soluble, followed by fractional separation. When extracting essential oils from
plant materials, the cuticular or leaf waxes located at the surface of vegetable matter
are coextracted (Reverchon 1992) even though the waxes have a relatively low
solubility in SC–CO2 compared to essential oils. Thus, to separate the essential oil
from the waxes, the extraction is carried out at 90 bar and 40 C, followed by a first
separation step in which the extract is cooled to 0 C, leading to the precipitation of
waxes. The second separation step can then be brought about by depressurization of
the remaining extract at 20 bar and by heating to 15 C, thereby recovering the
essential oils. Essential oil extraction and fractionation has been successfully
performed with basil, rosemary, marjoram, and numerous other plant materials on
a laboratory and industrial scale, as reviewed by Reverchon and De Marco (2006).
Another approach to fractionate essential oils is fractional extraction, where differ-
ent fractions are extracted by changing the polarity of the SC–CO2 throughout an
extraction process by adding a cosolvent. In this manner, grape seeds were
extracted in a two-step process, in which the seeds were subjected to pure
SC–CO2 extraction first, followed by extraction using SC–CO2 with added metha-
nol to change the polarity of SC–CO2 (Palma et al. 1999). The first fraction
contained mostly nonpolar compounds, including fatty acids, aliphatic aldehydes,
and sterols, whereas the second fraction using the cosolvent had phenolic com-
pounds, mainly catechin, epicatechin, and gallic acid. The first fraction showed
strong activity against several human pathogens, which was likely due to the sterols
according to Palma et al. (1999).
16.4 Commercialization and Future Outlook
Today’s health conscious consumers are increasingly demanding “natural” pro-

ducts when it comes to the functional foods and nutraceuticals marketplace. The
use of organic solvents and the safety of residual solvents in the final products are
being questioned. Government regulations are becoming stricter concerning the
use of organic solvents both from the residual and environmental perspectives.
Therefore, SC–CO2 technology has been experiencing tremendous growth at a
commercial scale around the globe, targeting various nutraceutical applications.
SC–CO2-extracted products are being promoted as “natural extract” products with
marketing advantages over other products extracted using organic solvents. As
well, considering the environmental concerns, “green” technologies like super-
critical processing are being evaluated more favorably.
Thermal or oxidative degradation of nutraceutical compounds is minimized
throughout SC–CO2 processing. However, care should be taken to maintain the
high quality during postprocess handling of the products. Further research is needed
to show the stability or changes in bioactivity of nutraceuticals during conventional
and SC–CO2 processes to clearly demonstrate the advantages of SC–CO2 techno-
logy for nutraceuticals processing, which then should be communicated to con-
sumers. Another benefit of supercritical technology over conventional solvent
extracts is longer shelf life due to inactivation of microorganisms and spores
throughout the high pressure treatment and the depressurization step. Such
SC–CO2 treatment is also being looked at as a sterilization technique for various
applications, including pharmaceuticals and medical devices (Foster et al. 2003).
Another major advantage of SC–CO2 technology is its flexibility and the possi-
bility of developing novel processes by coupling different unit operations. For
example, extraction operation can be combined with column fractionation or mem-
brane separation such that a crude extract can be obtained and refined or further
fractionated to isolate a bioactive component in two steps. Such approaches are
opening new opportunities for novel technology development. In addition, there are
tremendous developments in the area of particle formation using SC–CO2 technol-
ogies (Jung and Perrut 2001; Weidner 2009), which have opened up new opportu-
nities for the design of delivery systems for bioactive components. Even though
some of these approaches have been commercialized for various pharmaceuticals,
nutraceutical applications are very limited. These technologies can be classified
into three major categories, where CO2 can be used as: (1) solvent (rapid expansion
of supercritical solutions, RESS), (2) gas to saturate a solution (particles from gas-
saturated solutions, PGSS), or (3) antisolvent (gas anti-solvent, GAS; supercritical
fluid anti-solvent, SAS, and others). Weidner (2009) reviewed such techniques and
their food applications. In addition, Weidner (2009) also described a concentrated
powder form (CPF), where a liquid ingredient is contacted with a pressurized gas
and sprayed through a nozzle onto a solid carrier material. Coupling these techni-
ques with SC–CO2 extraction/fractionation would allow the recovery of a sensitive
bioactive from a plant matrix and encapsulation in a protective coating with
minimal degradation. Micro- and nano-sized particles of different morphologies
obtained through supercritical techniques can offer functionalities not possible with
conventional technologies, such as targeted delivery and controlled release
mechanisms; however, more research is needed to understand the relationships
between processing parameters and final ingredient functionality.
The feasibility of achieving bioseparations using SC–CO2 technology has been
demonstrated for various nutraceutical applications discussed in this chapter,
including extraction of bioactives from a variety of plant materials and fractionation
for further concentration of specific compounds. A large number of studies have
been carried out in laboratory and pilot plant scale. Some applications have already
reached commercial scale, such as extraction of specialty oils and isolation of
tocopherols from deodorizer distillate using a fractionation column. As in any
process, detailed economic feasibility evaluation is essential for the success of the
different approaches discussed for the variety of nutraceutical applications, and cost
evaluation should be done on a case by case basis. In general, it is common belief
that supercritical fluid technology is expensive and should be restricted to only
high-value end products. This is mainly due to the need for high capital investment
for high pressure equipment. However, it has been demonstrated that it is only a
matter of scale of operations. At large enough plant capacities, the viability of
supercritical processing becomes equivalent to conventional processes even for
low-value end products (Perrut 2000; Brunner 2005). Perrut (2000) reported the
linear relationship (on a log-log scale) between price index and plant capacity over
a range of lab to pilot and production units. Brunner (2009) also demonstrated the
substantial drop in cost for continuous supercritical operations and provided cost
comparisons for different column operations. The operating cost is generally lower
than that for conventional solvent extraction or high vacuum operations. As well,
compared to traditional solvent extraction, the overall process becomes simpler
because expensive solvent evaporators and meal desolventizers are not needed.
Thus, supercritical processing is more favorable especially when the rising cost of
energy is taken into account. Rosa and Meireles (2005) developed a rapid method to
estimate manufacturing cost by taking into account raw material, operational labor,
utilities, waste treatment, and investment. Coupling of supercritical extraction or
fractionation with a continuous cross-flow membrane system for the separation of
solutes from SC–CO2 without depressurization shows great promise in terms of
substantial savings in energy costs associated with the recompression of CO2.
Considering all the advantages discussed, the future of supercritical technology
for separations involving nutraceuticals is bright based on all the know-how gained
over the past 2 decades in this field.
Acknowledgements We would like to express our gratitude to the Natural Sciences and Engi-
neering Research Council of Canada for financial support of our research program on supercritical
fluid technology and Alberta Ingenuity for scholarship support to B. Seifried.
References
Ackman RG (1988) Oils and fats group international lecture – the year of the fish oils. Chem Ind
7:139–145
Ackman RG, Ke PJ, Jangaard PM (1973) Fractional vacuum distillation of Herring oil methyl-
esters. J Am Oil Chem Soc 50:1–8
Ambrogi A, Cardarelli DA, Eggers R (2002) Fractional extraction of paprika using supercritical
carbon dioxide and on-line determination of carotenoids. J Food Sci 67:3236–3241
Ambrogi A, Cardarelli DA, Eggers R (2003) Separation of natural colorants using a combined high
pressure extraction-adsorption process. Lat Am Appl Res 33:323–326
Bartle KD, Clifford AA, Jafar SA, Shilstone GF (1991) Solubilities of solids and liquids of low
volatility in supercritical carbon-dioxide. J Phys Chem Ref Data 20:713–756
Baumann D, Adler S, Gr€ uner S, Otto F, Weinreich B, Hamburger M (2004) Supercritical carbon
dioxide extraction of marigold at high pressures: Comparison of analytical and pilot-scale
extraction. Phytochem Anal 15:226–230
Baysal T, Ersus S, Starmans DAJ (2000) Supercritical CO2 extraction of b-carotene and lycopene
from tomato paste waste. J Agric Food Chem 48:5507–5511
Bernardo-Gil MG, Grenha J, Santos J, Cardoso P (2002) Supercritical fluid extraction and
characterisation of oil from hazelnut. Eur J Lipid Sci Technol 104:402–409
Blaha-Schnabel A, Beyer A, Czech B, Jakob H, Schiemann H, Weidner E, Peter S (1996)

Influence of interfacial tension and viscosity on the behavior of a packed column in near-
critical fluid extraction. Chem Eng Commun 146:13–31
Breivik H, Haraldsson GG, Kristinsson B (1997) Preparation of highly purified concentrates of
eicosapentaenoic acid and docosahexaenoic acid. J Am Oil Chem Soc 74:1425–1429
Brunner G (1983) Selectivity of supercritical compounds and entrainers with respect to model
substances. Fluid Phase Equilib 10:289–298
Brunner G (1994) Gas extraction: an introduction to fundamentals of supercritical fluids and the
application to separation processes. Springer, New York
Brunner G (2005) Supercritical fluids: technology and application to food processing. J Food Eng
67:21–33
Brunner G (2009) Counter-current separations. J Supercrit Fluids 47:574–582
Brunner G, Peter S (1982) On the solubility of glycerides and fatty acids in compressed gases in
the presence of an entrainer. Sep Sci Technol 17:199–214
Cadoni E, Rita De Giorgi M, Medda E, Poma G (1999) Supercritical CO2 extraction of lycopene
and b-carotene from ripe tomatoes. Dyes Pigm 44:27–32
Campos L, Michielin EMZ, Danielski L, Ferreira SR (2005) Experimental data and modeling the
supercritical fluid extraction of marigold (Calendula officinalis) oleoresin. J Supercrit Fluids
34:163–170
Cao X, Ito Y (2003) Supercritical fluid extraction of grape seed oil and subsequent separation of
free fatty acids by high-speed counter-current chromatography. J Chromatogr A 1021:117–124
Catchpole OJ, King MB (1994) Measurement and correlation of binary diffusion coefficients in
near critical fluids. Ind Eng Chem Res 33:1828–1837
Catchpole OJ, Von Kamp JC, Grey JB (1997) Extraction of squalene from shark liver oil in a
packed column using supercritical carbon dioxide. Ind Eng Chem Res 36:4318–4324
Catchpole OJ, Grey JB, Noermark KA (2000) Fractionation of fish oils using supercritical CO2 and
CO2 plus ethanol mixtures. J Supercrit Fluids 19:25–37
Catchpole OJ, MacKenzie AN, Grey JB (2002) Improvements in or relating to separation
technology. NZ518504, WO03089399
Catchpole OJ, Grey JB, Mitchell KA, Lan JS (2004) Supercritical antisolvent fractionation of
propolis tincture. J Supercrit Fluids 29:97–106
Catchpole OJ, Tallon SJ, Eltringham WE, Grey JB, Fenton KA, Vagi EM, Vyssotski MV,
MacKenzie AN, Ryan J, Zhu Y (2009) The extraction and fractionation of specialty lipids
using near critical fluids. J Supercrit Fluids 47:591–597
Chimowitz EH, Pennisi KJ (1986) Process synthesis concepts for supercritical gas extraction in the
crossover region. AlChE J 32:1665–1676
Chrastil J (1982) Solubility of solids and liquids in supercritical gases. J Phys Chem 86:3016–3021
Christov M, Dohrn R (2002) High-pressure fluid phase equilibria: experimental methods and
systems investigated (1994–1999). Fluid Phase Equilib 202:153–218
Cocero MJ, González S, Perez S, Alonso E (2000) Supercritical extraction of unsaturated products
Degradation of b-carotene in supercritical extraction processes. J Supercrit Fluids 19:39–44
Danielski L, Campos L, Bresciani L, Hense H, Yunes R, Ferreira S (2007) Marigold (Calendula
officinalis L.) oleoresin: solubility in SC–CO2 and composition profile. Chem Eng Process
46:99–106
Darken LS (1948) Diffusion, mobility and their interrelation through free energy in binary metallic
systems. Trans Am Inst Min Metall Eng 175:184–201
Dauksas E, Venskutonis PR, Povilaityte V, Sivik B (2001) Rapid screening of antioxidant activity
of sage (Salvia officinalis L.) extracts obtained by supercritical carbon dioxide at different
extraction conditions. Nahrung 45:338–341
de la Fuente JC, Oyarzún B, Quezadaa N, del Valle JM (2006) Solubility of carotenoid pigments
(lycopene and astaxanthin) in supercritical carbon dioxide. Fluid Phase Equilib 247:90–95
Deiters UK, Schneider GM (1986) High pressure phase equilibria: experimental methods. Fluid
Phase Equilib 29:145–160
Diaz-Reinoso B, Moure A, Dominguez H, Parajo J (2006) Supercritical CO2 extraction and

purification of compounds with antioxidant activity. J Agric Food Chem 54:2441–2469
Dobbs JM, Wong JM, Johnston KP (1986) Nonpolar co-solvents for solubility enhancement in
supercritical fluid carbon-dioxide. J Chem Eng Data 31:303–308
Dohrn R, Brunner G (1995) High-pressure fluid-phase equilibria: experimental methods and
systems investigated (1988–1993). Fluid Phase Equilib 106:213–282
Duarte C, Moldao-Martins M, Gouveia AF, Beirão da Costa S, Leitão AE, Bernardo-Gil MG
(2004) Supercritical fluid extraction of red pepper (Capsicum frutescens L.). J Supercrit Fluids
30:155–161
Dyerberg J, Bang HO, Hjorne N (1975) Fatty acid composition of the plasma lipids in Greenland
Eskimos. Am J Clin Nutr 28:958–966
Eisenbach W (1984) Supercritical fluid extraction: a film demonstration. Phys Chem Chem Phys
88:882–887
Ekart MP, Bennett KL, Ekart SM, Gurdial GS, Liotta CL, Eckert CA (1993) Co-solvent interac-
tions in supercritical fluid solutions. AlChE J 39:235–248
El Boustani S, Colette C, Monnier L, Descomps B, de Paulet AC, Mendy F (1987) Enteral
absorption in man of eicosapentaenoic acid in different chemical forms. Lipids 22:711–714
Eltringham W, Catchpole OJ (2008) Processing of fish oils by supercritical fluids. In: Martinez JL
(ed) Supercritical fluid extraction of nutraceuticals and bioactive compounds. CRC Press/
Taylor & Francis, Boca Raton, FL
Fang T, Goto M, Wang X, Ding X, Geng J, Sasaki M, Hirose T (2007) Separation of natural
tocopherols from soybean oil byproduct with supercritical carbon dioxide. J Supercrit Fluids
40:50–58
Favati F, King JW, Mazzanti M (1991) Supercritical carbon dioxide extraction of evening
primrose oil. J Am Oil Chem Soc 68:422–427
FDA (2005) Qualified health claims: letter regarding tomatoes and prostate cancer (Lycopene
health claim coalition; Docket no. 2004Q-0201). http://www.fda.gov/food/dietarysupple-
ments/consumerInformation/ucm072767.html.Retrieved 1 June 2009
Fenghour A, Wakeham WA, Vesovic V (1998) The viscosity of carbon dioxide. J Phys Chem Ref
Data 27:31–39
Fleck U, Tiegs C, Brunner G (1998) Fractionation of fatty acid ethyl esters by supercritical CO2: high
separation efficiency using an automated countercurrent column. J Supercrit Fluids 14:67–74
Fornari RE, Alessi P, Kikic L (1990) High pressure fluid phase equilibria: experimental methods
and systems investigated (1978–1987). Fluid Phase Equilib 57:1–33
Fors SM, Eriksson CE (1990) Characterization of oils extracted from oats by supercritical carbon-
dioxide. LWT Food Sci Technol 23:390–395
Foster N, Mammucari R, Dehghani F, Barrett A, Bezanehtak K, Coen E, Combes G, Meure L, Ng
A, Regtop HL, Tandya A (2003) Processing pharmaceutical compounds using dense gas
technology. Ind Eng Chem Res 42:6476–6493
Funazukuri T, Hachisu S, Wakao N (1991) Measurements of binary diffusion coefficients of
C16-C24 unsaturated fatty acid methyl esters in supercritical carbon dioxide. Ind Eng Chem Res
30:1323–1329
Funazukuri T, Kong CY, Kagei S (2004a) Effects of molecular weight and degree of unsaturation
on binary diffusion coefficients for lipids in supercritical carbon dioxide. Fluid Phase Equilib
219:67–73
Funazukuri T, Kong CY, Kagei S (2004b) Impulse response techniques to measure binary
diffusion coefficients under supercritical conditions. J Chromatogr A 1037:411–429
Gámez-Meza N, Noriega-Rodrı́guez JA, Medina-Juárez LA, Ortega-Garcı́a J, Monroy-Rivera FJ,
Toro-Vázquez HS, Garcı́a HS, Angulo-Guerrero O (2003) Concentration of eicosapentaenoic
acid and docosahexaenoic acid from fish oil by hydrolysis and urea complexation. Food Res Int
36:721–727
Gast K, Jungfer M, Saure C, Brunner G (2005) Purification of tocochromanols from edible oil.
J Supercrit Fluids 34:17–25
Gironi F, Maschietti M (2006) Separation of fish oils ethyl esters by means of supercritical carbon
dioxide: thermodynamic analysis and process modelling. Chem Eng Sci 61:5114–5126
Gomez AM, de la Ossa EM (2000) Quality of wheat germ oil extracted by liquid and supercritical
carbon dioxide. J Am Oil Chem Soc 77:969–974
Gomez-Prieto MS, Caja MM, Herraiz M, Santa-Maria G (2003) Supercritical fluid extraction of
all-trans-lycopene from tomato. J Agric Food Chem 51:3–7
Goodrum JW, Kilgo MB (1987) Peanut oil extraction with SC–CO2 – solubility and kinetic
functions. T ASAE 30:1865–1868
Gruger EH, Nelson RW, Stansby ME (1964) Fatty acid composition of oils from 21 species of
marine fish, freshwater fish and shellfish. J Am Oil Chem Soc 41:662–667
€ undag Ö, Temelli F (2000) Correlating the solubility behavior of fatty acids, mono-, di-,
G€uçl€u-Ust€
and triglycerides, and fatty acid esters in supercritical carbon dioxide. Ind Eng Chem Res
39:4756–4766
€ undag Ö, Temelli F (2004) Correlating the solubility behavior of minor lipid compo-
nents in supercritical carbon dioxide. J Supercrit Fluids 31:235–253
€ undag Ö, Temelli F (2005) Solubility behavior of ternary systems of lipids, co-solvents
and supercritical carbon dioxide and processing aspects. J Supercrit Fluids 36:1–15
€ undag Ö, Temelli F (2006) Solubility behavior of ternary systems of lipids in supercriti-
cal carbon dioxide. J Supercrit Fluids 38:275–288
€ undag Ö, Temelli F (2007) Column fractionation of canola oil deodorizer distillate using
supercritical carbon dioxide. J Am Oil Chem Soc 84:953–961
Gupta RB, Shim J-J (2007) Solubility in Supercritical Carbon Dioxide. CRC Press, Boca
Raton, FL
Gurdial GS, Wells PA, Foster NR, Chaplin RP (1989) The role of polarity in correlations of solid-
supercritical fluid phase systems. J Supercrit Fluids 2:85–96
Hartono R, Mansoori GA, Suwono A (2001) Prediction of solubility of biomolecules in supercriti-
cal solvents. Chem Eng Sci 56:6949–6958
He HP, Corke H, Cai JG (2003) Supercritical carbon dioxide extraction of oil and squalene from
Amaranthus grain. J Agric Food Chem 51:7921–7925
Herrero M, Cifuentes A, Ibanez E (2006) Sub- and supercritical fluid extraction of functional
ingredients from different natural sources: Plants, food-by-products, algae and microalgae – A
review. Food Chem 98:136–148
Hicks CP (1978) A bibliography of thermodynamic quantities for binary fluid mixtures. In:
McGlashan ML (ed) Chemical thermodynamics. Chemical Society, London
Higashi H, Iwai Y, Nakamura Y, Yamamoto S, Arai Y (1999) Correlation of diffusion coefficients
for naphthalene and dimethylnaphthalene isomers in supercritical carbon dioxide. Fluid Phase
Equilib 166:101–110
Higashi H, Iwai Y, Arai Y (2001) Solubilities and diffusion coefficients of high boiling compounds
in supercritical carbon dioxide. Chem Eng Sci 56:3027–3044
Horrocks LA, Yeo YK (1999) Health benefits of docosahexaenoic acid (DHA). Pharmacol Res
40:211–225
Huang W, Li Z, Niu H, Li D, Zhang J (2008) Optimization of operating parameters for supercriti-
cal carbon dioxide extraction of lycopene by response surface methodology. J Food Eng
89:298–302
Ibanez E, Palacios J, Senorans FJ, Santa-Maria G, Tabera J, Reglero G (2000) Isolation and
separation of tocopherols from olive by-products with supercritical fluids. J Am Oil Chem Soc
77:187–190
Jung J, Perrut M (2001) Particle design using supercritical fluids: literature and patent survey.
Kestin J, Zien TF, Whitelaw JH (1964) Viscosity of carbon dioxide in neighbourhood of critical
point. Physica 30:161
Kim HJ, Lee SB, Park KA, Hong IK (1999) Characterization of extraction and separation of rice
bran oil rich in EFA using SFE process. Sep Purif Technol 15:1–8
King MB, Mubarak A, Kim JD, Bott TR (1992) The mutual solubilities of water with supercritical
and liquid carbon dioxides. J Supercrit Fluids 5:296–302
Knapp H, D€oring R, Oellrich L, Pl€ ocker U, Prausnitz JM (1981) Vapor-Liquid equilibria for
mixtures of low-boiling substances DECHEMA Chem. Data series VI
Krichnavaruk, Shotipruk A, Goto M, Pavasant P (2008) Supercritical carbon dioxide extraction of
astaxanthin from Haematococcus pluvialis with vegetable oils as co-solvent. Bioresour Tech-
nol 99:5556–5560
Krokan HE, Bjerve KS, Mørk E (1993) The enteral bioavailability of eicosapentaenoic acid and
docosahexaenoic acid is as good from ethyl esters as from glyceryl esters in spite of lower
hydrolytic rates by pancreatic lipase in vitro. Biochim Biophys Acta 1168:59–67
Lang Q, Wai CM (2001) Supercritical fluid extraction in herbal and natural product studies –
a practical review. Talanta 53:771–782
Lawson LD, Hughes BG (1988) Human absorption of fish oil fatty acids as triacylglycerols, free
acids, or ethyl esters. Biochem Biophys Res Commun 152:328–335
Lee SK (2003) Method for isolating high-purified unsaturated fatty acids using crystallization.
US Patent 6,664,405
Lee KT, Foglia TA (2001) Fractionation of menhaden oil and partially hydrogenated menhaden
oil: characterization of triacylglycerol fractions. J Am Oil Chem Soc 78:297–303
Leeke G, Gaspar F, Santos R (2002) Influence of water on the extraction of essential oils from a
model herb using supercritical carbon dioxide. Ind Eng Chem Res 41:2033–2039
Li S (2008) Application of supercritical fluids in traditional Chinese medicines and natural
products. In: Martinez JL (ed) Supercritical fluid extraction of nutraceuticals and bioactive
compounds. CRC Press/Taylor & Francis, Boca Raton, FL, pp 215–242
Li S, Varadarajan GS, Hartland S (1991) Solubilities of theobromine and caffeine in supercritical
carbon dioxide: Correlation with density-based models. Fluid Phase Equilib 68:263–280
Linder M, Matouba E, Fanni J, Parmentier M (2002) Enrichment of salmon oil with n-3 PUFA by
lipolysis, filtration and enzymatic re-esterification. Eur J Lipid Sci Technol 104:455–462
Liong KK, Wells PA, Foster NR (1991) Diffusion in supercritical fluids. J Supercrit Fluids
4:91–108
Lopes IMG, Bernardo-Gil MG (2005) Characterisation of acorn oils extracted by hexane and by
supercritical carbon dioxide. Eur J Lipid Sci Technol 107:12–19
Lucien FP, Foster NR (2000) Solubilities of solid mixtures in supercritical carbon dioxide: a
review. J Supercrit Fluids 17:111–134
Ma Q, Xu X, Gao Y, Wang Q, Zhao J (2008) Optimisation of supercritical carbon dioxide
extraction of lutein esters from marigold (Tagetes erect L.) with soybean oil as a co-solvent.
Int J Food Sci Technol 43:1763–1769
Macias-Sanchez MD, Serrano CM, Rodriguez M, de la Ossa EM, Lubian LM, Montero O (2008)
Extraction of carotenoids and chlorophyll from microalgae with supercritical carbon dioxide
and ethanol as co-solvent. J Sep Sci 31:1352–1362
Martı́n A, Cocero MJ (2007) Mathematical modeling of the fractionation of liquids with supercrit-
ical CO2 in a countercurrent packed column. J Supercrit Fluids 39:304–314
McHugh MA, Krukonis VJ (1986) Supercritical fluid extraction: principles and practice. Butter-
worths, Boston, MA
Meireles MAA (2008) Extraction of bioactive compounds from Latin American plants. In:
Martinez JL (ed) Supercritical fluid extraction of nutraceuticals and bioactive compounds.
CRC Press/Taylor & Francis, Boca Raton, FL, pp 243–274
Mendes RL, Fernandes HL, Coelho JP, Reis EC, Cabral JMS, Novais JM, Palavra AF (1995)
Supercritical CO2 extraction of carotenoids and other lipids from Chlorella vulgaris. Food
Chem 53:99–103
Mendes RL, Nobre BP, Coelho JP, Palavra AF (1999) Solubility of b-carotene in supercritical
carbon dioxide and ethane. J Supercrit Fluids 16:99–106
Michels A, Botzen A, Schuurman W (1957) The viscosity of carbon dioxide between 0 C and
75 C and at pressures up to 2000 atmospheres. Physica 23:95–102
Mishra VK, Temelli F, Ooraikul B (1993) Extraction and purification of omega-3-fatty-acids with
an emphasis on supercritical-fluid extraction – a review. Food Res Int 26:217–226
Moure A, Cruz JM, Franco D, Manuel Domı́nguez J, Sineiro J, Domı́nguez H, Núñez MJ, Carlos
Parajó J (2001) Natural antioxidants from residual sources. Food Chem 72:145–171
Mukhopadhyay M (2008) Processing spices using supercritical fluids. In: Martinez JL (ed)
Supercritical fluid extraction of nutraceuticals and bioactive compounds. CRC Press/Taylor &
Francis, Boca Raton, FL, pp 337–366
Murga R, Ruiz R, Beltran S, Cabezas JL (2000) Extraction of natural complex phenols and tannins
from grape seeds by using supercritical mixtures of carbon dioxide and alcohol. J Agric Food
Chem 48:3408–3412
Nagahama K (1996) VLE measurements at elevated pressures for process development. Fluid
Phase Equilib 116:361–372
Nagy B, Simandi B (2008) Effects of particle size distribution, moisture content, and initial oil
content on the supercritical fluid extraction of paprika. J Supercrit Fluids 46:293–298
Nilsson WB, Gauglitz EJ Jr, Hudson JK, Stout VF, Spinelli J (1988) Fractionation of menhaden oil
ethyl esters using supercritical fluid CO2. J Am Oil Chem Soc 65:109–117
Nilsson WB, Gauglitz EJ, Hudson JK (1989) Supercritical fluid fractionation of fish oil esters
using incremental pressure programming and a temperature gradient. J Am Oil Chem Soc
66:1596–1600
NIST (2009) Isothermal properties for carbon dioxide. http://webbook.nist.gov. Retrieved 10 June
2009
Nobre BP, Mendes RL, Queiroz EM, Pessoa FP, Coelho JP, Palavra AF (2009) Calculation of
solubilities for systems containing multiple non-volatile solutes and supercritical carbon
dioxide. Ind Eng Chem Res 48:1551–1555
Nordoy A, Barstad L, Connor WE, Hatcher L (1991) Absorption of the n-3 eicosapentaenoic
and docosahexaenoic acids as ethyl esters and triglycerides by humans. Am J Clin Nutr 53:
1185–1190
Odabasi AZ, Balaban MO (2002) Supercritical CO2 extraction of sesame oil from raw seeds.
J Food Sci Technol 39:496–501
Oliveira R, Rodrigues MF, Bernardo-Gil MG (2002) Characterization and supercritical carbon
dioxide extraction of walnut oil. J Am Oil Chem Soc 79:225–230
Ollanketo M, Hartonen K, Riekkola ML, Holm Y, Hiltunen R (2001) Supercritical carbon dioxide
extraction of lycopene in tomato skins. Eur Food Res Technol 212:561–565
Ooi CK, Bhaskar A, Yener MS, Tuan DQ, Hsu J, Rizvi SSH (1996) Continuous supercritical
carbon dioxide processing of palm oil. J Am Oil Chem Soc 73:233–237
Özogul Y, Özogul F (2007) Fatty acid profiles of commercially important fish species from the
Mediterranean, Aegean and Black Seas. Food Chem 100:1634–1638
Palazoglu TK, Balaban MO (1998) Supercritical CO2 extraction of lipids from roasted pistachio
nuts. T ASAE 41:679–684
Palma M, Taylor LT, Varela RM, Cutler SJ, Cutler HG (1999) Fractional extraction of compounds
from grape seeds by supercritical fluid extraction and analysis for antimicrobial and agrochem-
ical activities. J Agric Food Chem 47:5044–5048
Peker H, Srinivasan MP, Smith JM, McCoy BJ (1992) Caffeine extraction rates from coffee beans
with supercritical carbon dioxide. AlChE J 38:761–770
Perretti G, Motori A, Bravi E, Favati F, Montanari L, Fantozzi P (2007) Supercritical carbon
dioxide fractionation of fish oil fatty acid ethyl esters. J Supercrit Fluids 40:349–353
Perrut M (2000) Supercritical fluid applications: industrial developments and economic issues. Ind
Eng Chem Res 39:4531–4535
Peusch M, M€uller-Seitz E, Petz M, M€ uller A, Anklam E (1997) Extraction of capsaicinoids from
chillies (Capsicum frutescens L.) and paprika (Capsicum annuum L.) using supercritical fluids
and organic solvents. Z. Lebensm. Unters. Forsch. Eur Food Res Technol 204:351–355
Ramı́rez Fajardo A, Esteban Cerdán L, Robles Medina A, Muñı́o Martı́nez MM, Hita Peña E,
Molina Grima E (2006) Concentration of eicosapentaenoic acid by selective esterification
using lipases. J Am Oil Chem Soc 83:215–221
Raspo I, Nicolas C, Neau E, Meradji S (2008) Diffusion coefficients of solids in supercritical
carbon dioxide: modelling of near critical behaviour. Fluid Phase Equilib 263:214–222
Ratledge C (2004) Fatty acid biosynthesis in microorganisms being used for Single Cell Oil
production. Biochimie 86:807–815
Reverchon E (1992) Fractional separation of SCF extracts from marjoram leaves: mass transfer
and optimization. J Supercrit Fluids 5:256–261
Reverchon E (1997) Supercritical fluid extraction and fractionation of essential oils and related
products. J Supercrit Fluids 10:1–37
Reverchon E, De Marco I (2006) Supercritical fluid extraction and fractionation of natural matter.
Reverchon E, Kaziunas A, Marrone C (2000) Supercritical CO2 extraction of hiprose seed oil:
experiments and mathematical modelling. Chem Eng Sci 55:2195–2201
Rezaei KA, Temelli F (2000) Using supercritical fluid chromatography to determine diffusion
coefficients of lipids in supercritical CO2. J Supercrit Fluids 17:35–44
Riha V, Brunner G (2000) Separation of fish oil ethyl esters with supercritical carbon dioxide.
Rosa PTV, Meireles MAA (2005) Rapid estimation of the manufacturing cost of extracts obtained
by supercritical fluid extraction. J Food Eng 67:235–240
Rozzi NL, Singh RK, Vierling RA, Watkins BA (2002) Supercritical fluid extraction of lycopene
from tomato processing byproducts. J Agric Food Chem 50:2638–2643
Ruckenstein E, Shulgin I (2001) Entrainer effect in supercritical mixtures. Fluid Phase Equilib
180:345–359
Ruckenstein E, Shulgin I (2002) The solubility of solids in mixtures composed of a supercritical
fluid and an entrainer. Fluid Phase Equilib 200:53–67
Ruivo R, Cebola MJ, Simoes PC, Nunes da Ponte M (2002) Fractionation of edible oil model
mixtures by supercritical carbon dioxide in a packed column. 2. A mass-transfer study. Ind Eng
Chem Res 41:2305–2315
Ruivo R, Couto R, SimÃmes PC (2008) Supercritical carbon dioxide fractionation of the model
mixture squalene/oleic acid in a membrane contactor. Sep Purif Technol 59:231–237
Sabio E, Lozano M, Montero De Espinosa V, Mendes RL, Pereira AP, Palavra AF, Coelho JA
(2003) Lycopene and b-carotene extraction from tomato processing waste using supercritical
CO2. Ind Eng Chem Res 42:6641–6646
Sacchetti G, Maietti S, Muzzoli M, Scaglianti M, Manfredini S, Radice M, Bruni R (2005)
Comparative evaluation of 11 essential oils of different origin as functional antioxidants,
antiradicals and antimicrobials in foods. Food Chem 91:621–632
Sakaki K (1992) Solubility of b-carotene in dense carbon-dioxide and nitrous-oxide from 308 to
323 K and from 9.6 to 30 MPa. J Chem Eng Data 37:249–251
Saldaña MDA, Sun L, Guigard SE, Temelli F (2006) Comparison of the solubility of b-carotene in
supercritical CO2 based on a binary and a multicomponent complex system. J Supercrit Fluids
37:342–349
Sanal IS, Bayraktar E, Mehmetoglu UU, Calimli A (2005) Determination of optimum conditions
for SC-(CO2 plus ethanol) extraction of b-carotene from apricot pomace using response surface
methodology. J Supercrit Fluids 34:331–338
Sanderson P, Finnegan YE, Williams CM, Calder PC, Burdge GC, Wootton SA, Griffin BA,
Millward DJ, Pegge NC, Bemelmans WJE (2002) UK Food Standards Agency a-linolenic acid
workshop report. Br J Nutr 88:573–579
Schmitt WJ, Reid RC (1985) The influence of the solvent gas on solubility and selectivity in
supercritical extraction. In: Penninger JML, Radosz M, McHugh MA, Krukonis VJ (eds)
Supercritical Fluid Technology. Elsevier, New York, pp 123–147
Schmitt WJ, Reid RC (1986) The use of entrainers in modifying the solubility of phenanthrene and
benzoic acid in supercritical carbon dioxide and ethane. Fluid Phase Equilib 32:77–99
Senanayake S, Shahidi F (2000) Concentration of docosahexaenoic acid (DHA) from algal oil via
urea complexation. J Food Lipids 7:51–61
Shahidi F, Wanasundara UN (1998) Omega-3 fatty acid concentrates: nutritional aspects and
production technologies. Trends Food Sci Technol 9:230–240
Shen Z, Palmer MV, Ting SST, Fairclough RJ (1996) Pilot scale extraction of rice bran oil with
dense carbon dioxide. J Agric Food Chem 44:3033–3039
Shi J, Yi C, Xue SJ, Jiang Y, Ma Y, Li D (2009) Effects of modifiers on the profile of lycopene
extracted from tomato skins by supercritical CO2. J Food Eng 93:431–436
Skerget M, Knez Z (1997) Solubility of binary solid mixture b-carotene-capsaicin in dense CO2.
J Agric Food Chem 45:2066–2069
Skerget M, Knez A, Habulin M (1995) Solubility of b-carotene and oleic acid in dense CO2 and
data correlation by a density based model. Fluid Phase Equilib 109:131–138
Somayajulu GR (1989) Estimation procedures for critical constants. J Chem Eng Data 34:106–120
Sovova H, Kucera J, Jez J (1994) Rate of the vegetable oil extraction with supercritical CO2. 2.
Extraction of grape oil. Chem Eng Sci 49:415–420
Srinivasan K (2005) Spices as influencers of body metabolism: an overview of three decades of
research. Food Res Int 38:77–86
Staby A, Mollerup J (1993) Separation of constituents of fish oil using supercritical fluids: a review
of experimental solubility, extraction, and chromatographic data. Fluid Phase Equilib
91:349–386
Stahl E, Gerard D (1985) Solubility behaviour and fractionation of essential oils in dense carbon
dioxide. Perfum Flavor 10(2):29–30, 32, 34–37
Stahl E, Quirin KW (1983) Dense gas extraction on a laboratory scale: a survey of some recent
results. Fluid Phase Equilib 10:269–278
Stahl HCE, Schilz W (1976) Extraktion mit € uberkritischen Gasen in direkter Kopplung mit der
D€unnschicht-Chromatographie Anwendungsm€ oglichkeiten auf dem Naturstoffgebiet. Chem
Ing Tech 48:773–778
Stockfleth R, Brunner G (2001) Holdup, pressure drop, and flooding in packed countercurrent
columns for the gas extraction. Ind Eng Chem Res 40:347–356
Strøm A, Jensen RA (1951) Mortality from circulatory diseases in Norway 1940–1945. Lancet
1:126–129
Suárez JJ, Medina I, Bueno JL (1998) Diffusion coefficients in supercritical fluids: available data
and graphical correlations. Fluid Phase Equilib 153:167–212
Sun M, Temelli F (2006) Supercritical carbon dioxide extraction of carotenoids from carrot using
canola oil as a continuous co-solvent. J Supercrit Fluids 37:397–408
Taniguchi M, Tsuji T, Shibata M, Kobayashi T (1985) Extraction of oils from wheat-germ with
supercritical carbon-dioxide. Agric Biol Chem 49:2367–2372
Temelli F, Saldaña MDA, Moquin PHL, Sun M (2008) Supercritical fluid extraction of specialty
oils. In: Martinez JL (ed) Supercritical fluid extraction of nutraceuticals and bioactive com-
pounds. CRC Press/Taylor & Francis, Boca Raton, FL, pp 51–101
Thana P, Machmudah S, Goto M, Sasaki M, Pavasant P, Shotipruk A (2008) Response surface
methodology to supercritical carbon dioxide extraction of astaxanthin from Haematococcus
pluvialis. Bioresour Technol 99:3110–3115
Tilly KD, Foster NR, Macnaughton SJ, Tomasko DL (1994) Viscosity correlations for binary
supercritical fluids. Ind Eng Chem Res 33:681–688
Topal U, Sasaki M, Goto M, Hayakawa K (2006) Extraction of lycopene from tomato skin with
supercritical carbon dioxide: Effect of operating conditions and solubility analysis. J Agric
Food Chem 54:5604–5610
Tuan DQ, Zollweg JA, Harriott P, Rizvi SSH (1999) Measurement and modeling of viscosity of
supercritical carbon dioxide/biomaterial(s) mixtures. Ind Eng Chem Res 38:2129–2136
Uquiche E, del Valle JM, Ortiz J (2004) Supercritical carbon dioxide extraction of red pepper
(Capsicum annuum L.) oleoresin. J Food Eng 65:55–66
Vagi E, Simandi B, Suhajda A, Hethelyi E (2005) Essential oil composition and antimicrobial
activity of Origanum majorana L. extracts obtained with ethyl alcohol and supercritical carbon
dioxide. Food Res Int 38:51–57
Vagi E, Simandi B, Vasarhelyine KP, Daood H, Kery A, Doleschall F, Nagy B (2007) Supercriti-
cal carbon dioxide extraction of carotenoids, tocopherols and sitosterols from industrial tomato
by-products. J Supercrit Fluids 40:218–226
Van Alsten JG, Eckert CA (1993) Effect of entrainers and of solute size and polarity in supercriti-
cal fluid solutions. J Chem Eng Data 38:605–610
Vasapollo G, Longo L, Rescio L, Ciurlia L (2004) Innovative supercritical CO2 extraction of
lycopene from tomato in the presence of vegetable oil as co-solvent. J Supercrit Fluids
29:87–96
Vega PJ, Balaban MO, Sims CA, O’Keefe SF, Cornell JA (1996) Supercritical carbon dioxide
extraction efficiency for carotenes from carrots by RSM. J Food Sci 61:757–759
Vesovic V, Wakeham WA, Olchowy GA, Sengers JV, Watson JTR, Millat J (1990) The transport-
properties of carbon-dioxide. J Phys Chem Ref Data 19:763–808
Visentainer JV, D’Addio Noffs M, De Oliveira Carvalho P, De Almeida VV, De Oliveira CC, De
Souza NE (2007) Lipid content and fatty acid composition of 15 marine fish species from the
southeast coast of Brazil. J Am Oil Chem Soc 84:543–547
Walsh JM, Ikonomou GD, Donohue MD (1987) Supercritical phase-behavior – the entrainer
effect. Fluid Phase Equilib 33:295–314
Wang C, Harris WS, Chung M, Lichtenstein AH, Balk EM, Kupelnick B, Jordan HS, Lau J (2006)
n-3 Fatty acids from fish or fish-oil supplements, but not a-linolenic acid, benefit cardiovascu-
lar disease outcomes in primary- and secondary-prevention studies: a systematic review. Am J
Clin Nutr 84:5–17
Ward OP, Singh A (2005) Omega-3/6 fatty acids: alternative sources of production. Process
Biochem 40:3627–3652
Weidner E (2009) High pressure micronization for food applications. J Supercrit Fluids
47:556–565
Wiebe R (1941) The binary system carbon dioxide-water under pressure. Chem Rev 29:475–481
Yener ME, Kashulines P, Rizvi SSH, Harriott P (1998) Viscosity measurement and modeling of
lipid supercritical carbon dioxide mixtures. J Supercrit Fluids 11:151–162
Ziger DH, Eckert CA (1983) Correlation and prediction of solid-supercritical fluid phase equili-
bria. Ind Eng Chem Process Des Dev 22:582–588
Zuta CP, Simpson BK, Chan HM, Phillips L (2003) Concentrating PUFA from mackerel proces-
sing waste. J Am Oil Chem Soc 80:933–936
Chapter 17
Mass Transfer and Equilibrium Parameters
on High-Pressure CO2 Extraction of Plant
Essential Oils
José M. del Valle, Juan C. de la Fuente, Edgar Uquiche,

Carsten Zetzl, and Gerd Brunner
17.1 Introduction
The production of plant extracts is currently limited by safety and regulatory

constraints on the concentration of toxic residues of organic solvents such as
n-hexane or methanol (Sanders 1993). Carbon dioxide (CO2) is an excellent alterna-
tive to these organic solvents due to its inertness, non-toxicity, non-flammability, and
low cost (Brunner 1994; del Valle and Aguilera 1999). The recommended temper-
ature in so-called supercritical fluid (SCF) extraction (SCFE) processes is slightly
above the critical temperature (Tc) of the solvent so that a near-environmental
temperature is applied when using CO2 (Tc ¼ 304.2 K) in SCFE, thus reducing
the requirement of energy for separation as well as the thermal damage of labile
bioactive compounds (Brunner 1994). Furthermore, objectionable solvent traces
are eliminated from the treated substrate and the extract because, being a gas
under normal environmental conditions, CO2 is easily and fully removed following
SCFE.
We would like to dedicate this chapter to the memory of our coauthors and friends Damian
Cardarelli (1964–2007) and Miguel Mattea (1955–2007) who died tragically as a result of injuries
sustained during a fire accident on December 5, 2007, in the pilot plant of the Facultad de
Ingenierı́a in the Universidad Nacional de Rio Cuarto (Córdoba, Argentina). We met them either
briefly (JMdV) or a long time ago (JdlF), but we were touched by their intellectual capacity,
perseverance, camaraderie, and warmness.
J.M. del Valle (*)
Departamento de Ingenierı́a Quı́mica y Bioprocesos, Pontificia Universidad Católica (PUC) de
Chile, Avda. Vicuña Mackenna 4860, Macul, Santiago, Chile
e-mail: delvalle@ing.puc.cl
J.C. de la Fuente
Departamento de Procesos Quı́micos, Biotecnológicos y Ambientales, Universidad Técnica
Federico Santa Marı́a, Valparaı́so, Chile
E. Uquiche
Departamento de Ingenierı́a Quı́mica, Universidad de La Frontera, Temuco, Chile
C. Zetzl and G. Brunner
Thermische Verfahrenstechnik, Technische Universit€at Hamburg-Harburg (TUHH), Harburg,
Germany
394 J.M. del Valle et al.
Because of their high compressibility, particularly in the vicinity of the critical

point, SCFs exhibit large variations in physical properties, including near-liquid
solvent power and near-gas transport properties, and it is possible to take advantage
of these improved properties to devise improved (medium-to-high yield, high
selectivity, fast) extraction processes (Brunner 1994; del Valle and Aguilera
1999). These advantages of SCFs extend to near-critical liquids and gases. Thus,
in this work the use of both supercritical CO2 and near-critical CO2 as extraction
solvents will be covered; they will be referred to as high-pressure CO2. The
extraction process with high-pressure CO2 will be referred as SCFE.
The typical quality fluctuations of vegetable substrates and difficulties in obtain-
ing the standardized information required for designing selective SCFE processes
make the industrial application of CO2 as an extraction solvent for plant materials
difficult (Zetzl et al. 2003). In this chapter an attempt is made to bridge the gap
between scientific research and industrial application of SCFE for obtaining rele-
vant plant extracts. Specifically, the focus will be on the SCFE of plant essential oils
using high-pressure CO2.
Two well-established SCFE applications in the food industry are the decaffeina-
tion of coffee (Lack and Seidlitz 1993) and the recovery of bitter compounds from
hops to be used in improved beer-producing processes (Hubert and Vitzthum 1978;
Gardner 1993). However, in this work it was felt that, viewed from a broader
chemical perspective, the two most important types of components extracted
from plant materials using high-pressure CO2 are fatty oils (triglycerides) and
essential oils (del Valle et al. 2005a). In the specific case of plant essential oils,
SCFE allows a higher yield to be attained in a shorter time as compared to steam
distillation, hydro-distillation, and conventional solvent extraction, while simulta-
neously avoiding thermal and hydrolytic degradation of labile compounds, as well
as toxic solvent residues in the product (Stahl et al. 1988; Moyler 1993; Reverchon
1997; Mukhopadhyay 2000; Reverchon and De Marco 2006, 2008; Quirin and
Gerard 2007).
Authors del Valle and de la Fuente (2006) described SCFE of fatty oils from
seeds, including relevant mass transfer and equilibrium parameters. Thus, the
purpose of this chapter is to provide similar information for the industrially
important case of SCFE of plant essential oils. Because the solubility of essential
oils in high-pressure CO2 is great, their selective extraction from a plant requires
low-to-medium densities to avoid contamination of the extract with heavier and/or
more polar compounds such as, e.g., fatty oils and carotenoids. Thus, for the
purpose of this chapter, plant essential oils are defined as those terpenes that can
be extracted with high-pressure CO2 at 15 MPa or less.
17.1.1 Chemistry and Localization of Essential Oils
Essential oils are complex mixtures of many volatile compounds that are responsi-
ble for the aroma of herbs and spices. As a group, essential oils do not share
17 Mass Transfer and Equilibrium Parameters on High‐Pressure 395
common chemical properties beyond conveying the characteristic aroma of the herb
or spice and consequently are used for flavoring foods, drinks, perfumes, cosmetics,
incense, and bath and house-cleaning products. For applications in foods, a spice is
a seed, fruit, root, bark, or leaf that is dried, ground (usually), and used in small
amounts as a preservative against deleterious or harmful microorganisms, or as an
additive to impart flavor or color to the food. Herbs differ from spices in that they
are leafy, green plant parts that are usually chopped into smaller pieces and used in
a fresh (undried) condition as food preservatives or flavor additives.
Brielmann et al. (2006) reviewed the chemistry of plant essential oils. The most
volatile components in essential oils are terpenes, which are secondary plant
metabolites derived from isoprene (2-methyl-1,3-butadiene), a 5-carbon unsatu-
rated hydrocarbon molecule. Terpenes include the 2-isoprene (or 10-carbon) mono-
terpene hydrocarbons that typically fit a C10H16 molecular formula (e.g., p-cymene,
limonene, a-pinene), the 3-isoprene (or 15-carbon) sesquiterpene hydrocarbons that
typically fit a C15H24 molecular formula (e.g., b-caryophyllene, a-humulene), and
oxygen-containing derivatives of monoterpene and sesquiterpene hydrocarbons
(the so-called oxygenated monoterpenes and oxygenated sesquiterpenes, respec-
tively) such as acetates (e.g., linalyl acetate, farnesyl acetate), alcohols (e.g.,
b-citronellol, geraniol, farnesol, linalool, menthol, patchoulol, verbenol), aldehydes
(e.g., p-anisaldehyde, citral), ketones (e.g., camphor, carvone, fenchone), phenols
(e.g., carvacrol, eugenol, thymol), and oxides (e.g., artemisinin, 1,8-cineole),
among others. Among these compounds the oxygenated monoterpenes are espe-
cially important because they are responsible for the characteristic aroma of the
herb or spice.
Plant essential oils are encapsulated in specialized secretory structures made of
high-molecular-weight nonvolatile waxes (CnH2n+2) and other fatty compounds that
protect them against evaporative losses and deleterious oxidative and/or hydrolytic
reactions by atmospheric oxygen and water. Essential oils are secreted into
specialized structures that can be located in either the surface (the so-called
glandular trichomes or glands) or below the outer surface of the plant material
(secretory ducts and secretory cavities) depending on the plant family and species
(Denny 1991; Zizovic et al. 2007c; Stamenić et al. 2008). The leaves, terminal
shoots, and flowers of aromatic herbs of the Lamiaceae family, such as basil (Ocimum
basilicum), lavender (Lavandula angustifolia), marjoram (Origanum majorama),
oregano (Origanum vulgare), pennyroyal (Mentha pulegium), peppermint (Mentha
piperita), rosemary (Rosmarinus officinalis), sage (Salvia officinalis), spearmint
(Mentha spicata), thyme (Thymus vulgaris), and wild thyme (Thymus serpyllum),
produce superficial oils that are stored in abundant secretory cells called glandular
trichomes or glands (Zizovic et al. 2005, 2007c; Stamenić et al. 2008).
Unlike the aromatic herbs of the Lamiaceae family that secrete oils in superfi-
cial glands, herbs and spices of the Apiaceae and Asteraceae families secrete oils
into subcutaneous ducts (Zizovic et al. 2007a,b,c; Stamenić et al. 2008). These
ducts are elongated cavities that can branch out to create a network of
interconnected pores extending from the roots, through the stems, and to the
leaves, flowers, and fruits of the plants. In this chapter the Apiaceae family is
represented by anise (Pimpinella anisum), caraway (Carum carvi), celery (Apium

graveolens), fennel (Foeniculum vulgare), and parsley (Petroselinum crispum),
whereas the Asteraceae family is represented by candeia (Eremanthus erythro-
pappus), carqueja (Baccharis trimera), chamomile (Matricaria recutita), and
marigold (Calendula officinalis).
Other forms of subcutaneous oils are accumulated in secretory cavities, which
are spherical structures lined with essential-oil-producing ephitelium cells (Zizovic
et al. 2007c; Stamenić et al. 2008). Secretory cavities can be found in both aerial
and underground parts of many plants including roots (valerian, Valeriana offici-
nalis, Valerianaceae), rhizomes (ginger, Zingiber officinale, Zingiberaceae), leaves
(alecrim pimenta, Lippia sidoides, Verbenaceae; boldo, Peumus boldus, Monimia-
ceae; cinnamon of Cunhã, Croton zehntneri, Euphorbiaceae; eucalyptus, Eucalyp-
tus globulus, Myrtaceae; ho-sho, Cinnamomum camphora, Lauraceae), flower buds
(clove, Syzygium aromaticum, Myrtaceae) and cones (hop, Humulus lupulus, Can-
nabaceae), fruit peels (orange, Citrus sinensis, Rutaceae); black pepper, Piper
nigrum, Piperaceae), and seeds (nutmeg, Myristica fragrans, Myristicaceae).
17.1.2 Organization of Chapter
Figure 17.1 represents the SCFE of a solid substrate in a packed bed. As will be
discussed in Sect. 17.2, the process can be modeled using a differential mass
balance equation (Fig. 17.1a) coupled with a mass transfer rate equation. During
SCFE of a pre-treated plant material, the partition of solutes between the solid and
fluid phases (K), the actual solubility of the solutes in the solvent (Csat), and various
resistances to mass transfer play important roles in extraction rates. The effective
diffusivity in the solid matrix (De), external mass transfer coefficient (kf), and axial
dispersion coefficient (Dax) all influence the concentration gradient-driven mass
transfer rates within the solid, the stationary SCF film surrounding each particle,
and along the bed, respectively (Fig. 17.1b).
Figure 17.2 presents an integral extraction plot of solute yield versus specific
solvent consumption resulting from the SCFE of the solid substrate in a packed bed.
Although there might be a positive effect of superficial solvent velocity on extrac-
tion rates, integral extraction curves generally collapse to a single line, at least
initially, with a slope that represents the so-called operational solubility (Cfo)
(Fig. 17.2, zone I). Scientific research on SCFE of plant essential oils, aimed at
the industrial application of the process, should result in mass transfer (kf, Dax, De)
and phase equilibrium or pseudo-equilibrium data (K, Csat, Cfo) that can be applied
for process design purposes.
In Sect. 17.2, mathematical models are presented and discussed for the SCFE of
essential oils from a packed bed of plant material, followed by a presentation of the
external (kf) and internal (De) mass transfer parameters that have been fitted in the
literature to cumulative extraction plots of essential oil yield versus time or specific
Fig. 17.1 Conceptual model of supercritical fluid extraction of solid substrates in packed beds: (a)
differential mass balance along packed bed; and (b) mass transfer phenomena within solid particle
(substrate), between particle and CO2 (solvent) phase, and in CO2 phase along bed
Fig. 17.2 Supercritical fluid

extraction curve showing
solute yield versus specific
solvent consumption. Both
the recovered solute and
amount of CO2 passed
through the packed bed are
expressed in a common base,
the amount of substrate
loaded into the extraction
vessel
solvent consumption (Sect. 17.3). Next, reported “operational” solubility (Cfo)

values are presented and compared to the “thermodynamic” or true solubility
(Csat) values of plant essential oil components in high-pressure CO2 in binary,
ternary, and more complex systems (Sect. 17.4). Some concluding remarks are
made in Sect. 17.5.
17.2 Mass Transfer Models
This section presents a mathematical model that can be used as a reference to

discuss the effect of mass transfer and equilibrium parameters on the SCFE of plant
essential oils using high-pressure CO2 as the solvent (Sect. 17.2.1). The so-called
diffusion model assumes internal mass transfer by diffusion within solid particles,
external mass transfer by convection through a static film of SCF next to the
particles, and axial dispersion of dissolved solute in the SCF phase along the bed.
This reference model also assumes a packed bed of spherical particles, an essential
oil that can be treated as a single substance (pseudo-solute), and a constant partition
coefficient for this pseudo-solute between the solid substrate and high-pressure
CO2. Discussion follows covering the effect on mass transfer rates of particle shape,
localization of the solute in specialized structures within the solid matrix, actual
composition of the essential oil, and its partition between the solid matrix and the
CO2 that limit the application of the diffusion model (Sect. 2.2). Finally, this section
concludes with a discussion of alternative internal mass transfer mechanisms that
can compliment diffusion, thus affecting the rate at which essential oils migrate
through the solid substrate (Sect. 2.3).
17.2.1 Diffusion Model
The reference model used in this section will be the diffusion model of Goodarznia
and Eikani (1998). The assumptions of this diffusion model are as follows: (1) the
substrate particles are spherical (diameter dp, radius R ¼ dp/2) and homogeneous;
(2) the physical properties of the extract (the essential oil) are those of a represen-
tative pseudo-solute; (3) the physical properties of the SCF and the substrate remain
unchanged during extraction; (4) the extract partitions between the solid and SCF
phases according to a constant coefficient (K) that is concentration-independent;
and (5) the solute disperses axially in the SCF as a result of irregularities in the
packing of the substrate in the bed and concentration-gradient-driven diffusion of
the solute along the packed bed. As a result of these assumptions, solute concentra-
tion in the solid matrix (Cs) depends on the radial position within the particle (r), the
axial position along the bed (z), and the extraction time (t); whereas solute concen-
tration in the SCF (Cf) depends only on z and t. The assumption of constant physical
properties is valid when the density and viscosity of the SCF depend only on the
extraction temperature and pressure, and are unaffected by the dissolved pseudo-
solute. Because of that, the variations in temperature and pressure should be
negligible within the bed, and the concentration of essential oils in the loaded
SCF phase should be small (selected process conditions should limit the solubility
and/or availability of the solute). On the other hand, the assumption of constant
physical properties of the solid phase is valid if the substrate remains unaffected
(does not swell or shrink) as a result of CO2 adsorption and solute removal, and this
results in a constant bed porosity (e) and, consequently, in a constant interstitial
velocity of the SCF in the packed bed (u ¼ U/e). These assumptions are valid in the
selective extraction of plant essential oils with high-pressure CO2 for several
reasons. The availability of solute is small because the content of essential oils in
a typical herb or spice is limited to a few percent or below. Recommended values of
extraction temperature and pressure (e.g., 323 K and 9 MPa) (Reverchon 1997)
place a limit on the solubility of the essential oil in high-pressure CO2, so as to
increase the selectivity of the process. A near-environmental extraction temperature
is selected to reduce thermal damage of labile compounds as well as energy
requirements of the process, which reduces the exchange of heat with the environ-
ment and minimizes radial temperature gradients in the extraction vessel. Finally,
as for other internally controlled mass transfer processes, small values of superficial
solvent velocity (1 U 5 mm/s) (Eggers 1996) are recommended to improve
economics, and these small velocities reduce losses of energy of the SCF as it
passes through the bed of packed substrate, and minimize axial pressures gradients
in the extraction vessel.
Equation 17.1 is a differential mass balance for the SCF surrounding the
particles of substrate in the packed bed. J, the so-called source-and-transfer term
(Zizovic et al. 2007c; Stamenić et al. 2008), is the flux of solute that is transferred
from the solid to the SCF, and can be estimated using (17.2). On the other hand, the
diffusion of the solute through the solid particles can be estimated using (17.3).
@Cf @Cf @ 2 Cf 6 ð 1 eÞ
þu Dax 2 ¼ J (17.1)
@t @z @z dp e

@Cs
J ¼ De (17.2)
@r R

@Cs De @ 2 @Cs
¼ 2 r (17.3)
@t r @r @r
Goodarznia and Eikani (1998) assumed constant solute concentrations in the

solid matrix (initial solute content Cso) and the SCF phase (initial solute content
Cfo) in the bed initially (17.4a and 17.4b, respectively), a symmetry condition for
the removal of solute from the particles (17.4c), continuity in the flux of solute
leaving a solid particle and entering the SCF film around it (17.4d), and the
Danckwerts conditions for axial dispersion in packed beds (17.4e and 17.4f).
Cs ¼ Cso ðt ¼ 0; 0 r RÞ (17.4a)
Cf ¼ Cfo ðt ¼ 0; 0 z HÞ (17.4b)

@Cs
¼0 ðt 0; 0 z H Þ (17.4c)
@r 0

@Cs C s jR
De ¼ kf Cf ðt 0; 0 z H Þ (17.4d)
@r R K
@Cf
uCf Dax ¼0 ðt 0; z ¼ 0Þ (17.4e)
@z
@Cf
¼ 0 ðt 0; z ¼ HÞ (17.4f)
@z
Examination of (17.1)–(17.4) suggests that variations in Cs and Cf as a function
of r, z, and t depend on the initial essential oil content in the substrate (Cso), the
geometry of the packed bed (diameter, DE; height, H; porosity, e) and the particles
(diameter, dp), and the conditions of the SCF (interstitial velocity, u; temperature, T;
and pressure, P).
A closer examination of (17.1)–(17.4) also suggests that the extraction rate and
yield depend on both kinetic (or mass transfer) parameters and equilibrium (or
solubility) parameters. The former category includes an internal mass transfer
coefficient (the effective diffusivity of the extract in the solid matrix, De), an
external mass transfer coefficient (the SCF film coefficient, kf), and an axial
dispersion coefficient (Dax) (Fig. 17.1b). Phase equilibrium parameters include
the solubility of the essential oils in high-pressure CO2 at T and P (Csat), and their
partition between the solid matrix and the SCF phase (K). Border condition 17.4b
implicitly requires extraction to be preceded by a static period (unaccounted for by
the model) so as to equilibrate the vessel to the required process temperature and
pressure conditions, and to dissolve free solute in the solid phase. The SCF phase
becomes saturated (initial solute concentration Csat) only if there is enough free
solute in the substrate; when there is less solute, the concentration Cfo is determined
by the ratio between the total amount of free solute and the total void volume
occupied by the SCF in the bed.
Some of the simplifications of the diffusion model applied in the literature
include the following (Table 17.1): neglecting the effect of axial dispersion in
mass transfer (model Diff/PF of Araus et al. 2009); treating the packed bed as a
perfectly mixed extraction vessel (model Diff-Sph/PM of Reverchon et al. 1993a;
models Diff-Slab/PM and Diff-Slab/IC of Gaspar et al. 2003; model Diff-Sph/IC of
Campos et al. 2005); and neglecting the external resistance to mass transfer (model
Diff-Slab/IC of Gaspar et al. 2003; model Diff-Sph/IC of Campos et al. 2005).
The diffusion model adopted in this chapter as the reference model (this section)
makes several simplifying assumptions about the particle geometry (spherical), the
Table 17.1 Summary of mathematical models used in the literature for high-pressure CO2 extraction of plant essential oils in packed beds
17
Mass transfer model Substrate Hydrodynamics in External mass Axial dispersion Sorption/
particlesa packed bedb transfer coefficientd “operational”
coefficientc solubility modele
Diffusion (Diff or D) models
Diff/ADPF (Goodarznia and Eikani 1998) Sphere Axially dispersed plug Literature Literature Linear isotherm
flow correlation correlation
Diff/PF (Araus et al. 2009; Infinite slab Plug flow Literature Neglected Linear isotherm
Uquiche et al. submitted) correlation
Diff-Sph/PM (Reverchon et al. 1993a) Sphere Perfect mixing Literature Neglected Neglected
correlation
Diff-Slab/PM (Gaspar et al. 2003) Infinite slab Perfect mixing Literature Neglected Neglected
correlation
Diff-Sph/IC (Campos et al. 2005) Sphere Perfect mixing Neglected Neglected Neglected
Diff-Slab/IC (Gaspar et al. 2003; Infinite slab Perfect mixing Neglected Neglected Neglected
Campos et al. 2005)
Shrinking-Core (SC) models
SC/ADPF (Spricigo et al. 2001; Porous Axially dispersed plug Literature Literature Limited by saturation
Machmudah et al. 2006) sphere flow correlation correlation
SC/ADPF (Steffani et al. 2006) Porous Axially dispersed plug Fitted to data Literature Limited by saturation
sphere flow correlation
SC/PF (Akgun et al. 2000; Porous Plug flow Literature Neglected Limited by saturation
Mass Transfer and Equilibrium Parameters on High‐Pressure
Germain et al. 2005) sphere correlation

SC/PF (Germain et al. 2005) Porous Plug flow Fitted to data Neglected Limited by saturation
sphere
Desorption-Dissolution-Diffusion
(DDD or D3) models
DDD/ADPF/BET (Ruetsch et al. 2003) Porous Axially dispersed plug Literature Literature BET isotherm
sphere flow correlation correlation
DDD/ADPF/Lang (Daghero et al. 2004) Porous Axially dispersed plug Literature Literature Langmuir isotherm
DDD/ADPF (Salimi et al. 2008) Porous Axially dispersed plug Literature Literature Several isotherms
401

(continued)
402
Mass transfer model Substrate Hydrodynamics in External mass Axial dispersion Sorption/
DDD/PM (Kim and Hong 2002) Porous Perfect mixing Fitted to data Neglected Linear isotherm,
sphere K=1
Intact-and-Broken-Cell (IBC) models
IBC-Diff (Kim and Hong 2002) Sphere Plug flow Neglected Literature Neglected
correlation
IBC/PF/PCPR (Machmudah et al. 2006; Unaccounted Plug flow Fitted to first Neglected PCPR isotherm
Sovová 2005; for stage
Langa et al. 2009)
IBC/PF (Louli et al. 2004) Unaccounted Plug flow Fitted to first Neglected Limited by saturation
for stage
IBC/PFNA (Sovová et al. 1994a) Unaccounted PF with no Fitted to first Neglected Decreasing solubility
for accumulation stage
Sovová (Campos et al. 2005;Louli et al. 2004; Unaccounted PF with no Fitted to first Neglected Limited by saturation
Mira et al. (1996, 1999); Papamichail et al. 2000; for accumulation stage
Povh et al. 2001; Ferreira and Meireles 2002;
Sousa et al. 2002; Martı́nez et al. 2003;
Rodrigues et al. 2003; Sousa et al. 2005; Vargas et al.
2006; Martı́nez et al. 2007; Bensebia et al. 2009)
Microscale (mS) models
mS/SGl (Zizovic et al. 2005; 2007c; Surface glands Axially Literature Literature Limited by saturation
Stamenić et al. 2008) dispersed correlation correlation
plug flow
mS/SDuct (Zizovic et al. 2007b, c; Secretory ducts Axially Literature Literature Limited by saturation
Stamenić et al. 2008) dispersed correlation correlation
plug flow
mS/SCav (Zizovic et al. 2007a, c; Stamenić et al. 2008) Secretory Axially Literature Literature Limited by saturation
cavities dispersed correlation correlation
plug flow
J.M. del Valle et al.
17
Diffusion models using Linear Driving Force (LDF)

LDF/ADPF (Reverchon and Marrone 1997) Unaccounted Axially Hidden in kg Literature Linear isotherm
for dispersed chart
plug flow
LDF/ADPF (Reis-Vasco et al. 2000) Unaccounted Axially Hidden in kg Fitted to data Linear isotherm
for dispersed
plug flow
LDF/PF/CDIC (Coelho et al. 1997) Unaccounted Plug flow Hidden in Neglected Limited by saturation
for variable kg
LDF-Sph/PF (Esquivel et al. 1996) Sphere Plug flow Hidden in kg Neglected Limited by saturation
LDF/PF (Reverchon et al. 1999) Unaccounted Plug flow Hidden in kg Neglected Linear isotherm
for
LDF/IMTC (Catchpole et al. 1996b) Sphere Plug flow Neglected Neglected Linear isotherm,
K«1
LDF-Slab/PMMS (Reverchon 1996) Infinite slab
Perfectly mixed Hidden in kg Neglected Linear isotherm
multistages
LDF/PMMS (Esquivel et al. 1996) Unaccounted Perfectly mixed Hidden in kg Neglected Linear isotherm
for multistages
LDF/UENA (Louli et al. 2004; Papamichail et al. 2000) Sphere Uniform Hidden in kg Neglected PCPR isotherm
extraction
NA
R-SO (Louli et al. 2004; Papamichail et al. 2000; Vargas et al. Sphere Uniform Neglected Neglected Neglected
2006; Reverchon et al. 1995a; Zekovic et al. 2001; Pfaf- extraction

Šovljanski et al. 2005) NA
DDD models using LDF
LDF-D3/DB/BET (Goto et al. 1998) Infinite slab Differential bed Literature Neglected BET isotherm
correlation
LDF-D3-Sph/DB (Perakis et al. 2005) Sphere Differential bed Literature Neglected Linear isotherm
correlation
LDF-D3-Slab/DB (Sousa et al. 2005; Goto et al. 1993) Infinite slab Differential bed Literature Neglected Linear isotherm
correlation
(continued)
403
404
Mass transfer model Substrate Hydrodynamics inExternal mass Axial dispersion Sorption/
Equilibrium Desorption (ED), Internal-Mass-Transfer-Control (IMTC), and External-Mass-Transfer-Control (EMTC) models
ED (Reis-Vasco et al. 2000) Unaccounted Plug flow Neglected Fitted to data Linear isotherm
for
IMTC (Ferreira et al. 1999) Sphere Perfect mixing Neglected Neglected Limited by
saturation
EMTC (Ferreira et al. 1999; Kotnik et al. 2007) Sphere Simplified diff. mass Fitted to first Neglected Limited by
balance stage saturation
a
The particles were treated as solid or porous spheres (Sph) or infinite slabs (Slab); or else their shape as an inner structure was unaccounted for
b
Flow conditions in extraction vessel are assumed to be axially dispersed plug flow (ADPF), plug flow (PF), perfect mixing (PM), or perfectly mixed
multistages (PMMS). For simplification, the packed bed was treated as a differential bed (DB), and the differential mass balance assumed plug flow with no
accumulation (PFNA) or uniform extraction with no accumulation (UENA)
c
The external mass transfer and was estimated from correlations in the literature, then fitted to data, or neglected. The internal-mass-transfer-control (IMTC)
models neglected the external resistance to mass transfer. The external mass transfer coefficient was sometimes hidden in a global mass transfer coefficient
(kg), either implicitly or explicitly; and in some cases (Reis-Vasco et al. 2000; Coelho et al. 1997; Ferreira et al. 1999; Kotnik et al. 2007) kg was assumed to be
an internal mass transfer coefficient (in IMTC models), and in others (Reverchon and Marrone 1997; Ferreira et al. 1999) (external-mass-transfer-control or
EMTC models), an external mass transfer coefficient
d
The axial dispersion coefficient was also estimated from correlations or charts in the literature, fitted to data, or neglected
e
The equilibrium concentration of essential oils in the SCF phase was assumed to be limited by saturation (solubility), or to depend on the residual
concentration of the oils in the solid substrate by a sorption isotherm given by a linear model, BET model, a Langmuir (Lang) model, or Perrut-Clavier-
Poletto-Reverchon (PCPR) model of Perrut et al. (1997), among others
solid matrix (homogeneous), the localization of the solute in the solid matrix
(homogeneous), the composition of the solute (fully characterized by a pseudo-
solute), the partition of the solute between the solid matrix and the CO2 (constant
and independent of solute concentration), and the mass transfer mechanism for the
extraction process (diffusion). As discussed in the following sections, some of these
simplifying assumptions should be avoided to improve the physical picture of the
extraction process.
17.2.2 Limitations of the Diffusion Model
When extracting herbs and spices it is important to consider the geometry of the
tissue following application of pretreatments aimed at increasing the speed and/or
final yield of the process. Mild pretreatments, such as coarse milling, are typically
applied prior to SCFE to take full advantage of the natural barriers within the plant
material to selectively extract the essential oils. Coarse milling of leaves and similar
plant parts results in large particles having a slab rather than a spherical geometry as
assumed in the diffusion model, so that the fitting of mathematical models to the
data improves when using the diffusion equation for an infinite slab instead of
(17.3) (Goto et al. 1993, 1998; Reverchon 1996; Gaspar et al. 2003; Araus et al.
2009). St€uber et al. (1997) derived analytical solutions for other regular geometries
such as disks and cylinders having different aspect ratios by combining the analytic
solutions for basic geometries such as infinite cylinders and infinite slabs using
superposition theorems. However, the model for spherical particles still can be used
to represent mass transfer from particles of various shapes (including disks and
cylinders with different aspect ratios) if a characteristic dimension is computed as
three times the volume-to-surface area ratio for a representative particle (dp/2, in the
case of a sphere) (Ma and Evans 1968).
When extracting herbs and spices it is important to consider the microstructure
of the native tissue. Unlike the assumptions in Sect. 17.2.1, herbs and spices are
heterogeneous when observed under the microscope. This is important to consider
because SCFE of plant materials depends, among other factors, on the location of
the solute within the plant tissue, with essential oils being encapsulated in isolated
glands, secretory cells, or cavities, or in interconnected pore networks (Zizovic
et al. 2007c; Stamenić et al. 2008). Furthermore, the physical properties of the solid
may change during extraction as a result of impregnation of CO2 and removal of
essential oils (Eggers 1996).
Also, unlike the assumptions in Sect. 17.2.1, essential oils are not single com-
pounds but mixtures of many different volatile terpenoids. During SCFE of herbs
and spices, the waxy constituents of the specialized encapsulating structures of the
essential oils are dissolved so that CO2 extracts also include waxes and other
compounds besides terpenoids. Thus, the components of CO2 extracts of herbs
and spices can be grouped as monoterpene hydrocarbons, oxygenated monoter-
penes, sesquiterpene hydrocarbons, oxygenated sesquiterpenes, waxes, and other
substrate specific families of compounds such as gingerols in ginger and phenyl-

propanoids in parsley. In this work, for the purpose of estimating the value of
physical properties of CO2 extracts, the authors selected representative compounds
of some of these groups, assigned the properties of representative compounds to
whole groups, and represented extracts as pseudo-solutes whose properties were
estimated as the weighted average of the properties of the representative com-
pounds in each considered group.
The assumption of a constant coefficient for the partition of the solute (linear
sorption isotherm) between the solid substrate and the high-pressure CO2 is invalid
when a fraction of the solute in the substrate, e.g., contained in broken cells or
cavities, is freely available to the CO2, so that its concentration in the fluid phase is
determined by availability or solubility constraints. It is also invalid in the final
stages of the extraction process, when all remaining solute is strongly bound to the
solid substrate (del Valle and de la Fuente 2006). Consequently, Araus et al. (2009)
claimed the necessity of relaxing the assumption of a constant partition coefficient
of the pseudo-solute between the solid substrate and the CO2 to improve modeling
of SCFE of plant essential oils.
The internal mass transfer mechanism may be different from that stated in
Sect. 17.2.1, since solute desorption from the solid, solubilization in the SCF,
and/or migration by diffusion through the pores of the solid matrix may control
mass transfer in the solid. The subject of the internal mass transfer mechanism is
analyzed in the following section.
17.2.3 Alternative Internal Mass Transfer Mechanisms
Considering plant tissue as a multiphase material constituted of interconnected or

isolated cells and a network of fully or partially interconnected pores (Zizovic et al.
2007c; Stamenić et al. 2008), and how its microstructure is affected by milling and
other pretreatments applied prior to SCFE, alternative mass transfer models picture
pretreated herbs and spices as porous solid matrices constituted of intact and broken
cells (Sovová 1994, 2005; Zizovic et al. 2007c; Stamenić et al. 2008). del Valle and
de la Fuente (2006) reviewed most of these models and others with alternative
internal mass transfer mechanisms in detail.
When the solid matrix where mass transfer takes place is assumed to be a
network of interconnected pores, two important models that can be applied are
the so-called shrinking-core (SC) and desorption-dissolution-diffusion (DDD)
models. The shrinking-core hypothesis assumes that the solute is retained in the
pores of the solid matrix by mechanical or capillary forces, so that pore space is
divided into an inner core filled with condensed solute and an outer region contain-
ing a solution of the solute in high-pressure CO2 that are separated by a moving
boundary. Roy et al. (1996) first applied the shrinking-core hypothesis to model
SCFE of ginger essential and fatty oils at >15 MPa (data not included), and there
are several examples in the literature on this hypothesis to model the extraction of
plant essential oils. Of these models (Table 17.1), that of Machmudah et al. (2006)
is different in that they assumed that the extract is a mixture of two pseudo-
components, namely the monoterpene and sesquiterpene hydrocarbons, on the
one hand, and their oxygenated derivatives (oxygenated monoterpenes and sesqui-
terpenes), on the other hand (Sect. 17.3.2), that are extracted separately at rates
defined by two independent values of De. Reverchon (1997) questioned the validity
of the shrinking-core hypothesis to model SCFE of substrates containing little
solute bound to the solid matrix, as it could be the case for essential oils in most
herbs and spices, where the driving force for the extraction depends little on the
solubility of the essential oil in high-pressure CO2 under extraction conditions.
On the other hand, the desorption-dissolution-diffusion hypothesis assumes that the
solute is partially adsorbed on the solid matrix within the pores, so that a fraction of
the solute is adsorbed on the solid matrix and the rest is dissolved in the SCF phase
within the inner pores of the solid, which are related by an equilibrium sorption
isotherm (see Table 17.1 for a summary of DDD models applied in literature). Goto
et al. (1993) assumed a linear sorption isotherm (or a constant equilibrium-partition
coefficient K), Ruetsch et al. (2003) applied a Brunauer-Emmet-Teller (or BET)
isotherm, Daghero et al. (2004) applied a Langmuir isotherm for solute concentra-
tions below the saturation concentration in the high-pressure CO2, and Salimi et al.
(2008) compared a linear isotherm with several other sorption models including
those of Langmuir, Freundlich, and Langmuir Freunlich. Kim and Hong (2002) did
not differentiate between the solute adsorbed on the solid and dissolved in the gas
phase within the pores, which implicitly implied a unitary equilibrium-partition
coefficient (K ¼ 1).
Sovová (1994) proposed the hypothesis of intact and broken cells (IBC), which
assumes that as a result of a mild pretreatment such as size reduction, particles of
milled plant material have broken cells on the surface, and intact cells in the
interior. An external (convective) mass transfer coefficient controls transfer of
solute from the broken cells to the SCF in the bed, whereas an internal (diffusive)
mass transfer coefficient controls transfer of solute from the intact cells (see
Table 17.1 for a summary of IBC models applied in literature). Reverchon et al.
(1999) considered the simultaneous extraction of free solute and bound solute using
separate coefficients for external mass transfer and internal mass transfer, respec-
tively, and separate equilibrium-partition coefficients for free solute between the
broken cells and the SCF in the bed, and between the bound solute in intact cells and
the SCF in the bed. The model (LDF/PF) of Reverchon et al. (1999) (Table 17.1) is
a simplification of this general IBC model, which they applied SCFE of lipids from
milled fennel seeds; the simplifying assumption was that there is no free essential
oil in milled fennel seeds and that all was bound to the solid matrix. Unlike
Reverchon et al. (1999), Sovová (2005) and Machmudah et al. (2006) assumed
that the driving force for the extraction of free essential oil is the difference between
an equilibrium concentration of the oil in high-pressure CO2 that depends on the oil
concentration in broken cells according to the so-called PCPR isotherm of Perrut
et al. (1997), as described in Sect. 17.4.4, and its concentration in the SCF phase,
whereas the driving force for extraction of bound essential oil is the difference in oil
concentration between the intact and broken cells (model IBC/PF/PCPR in

Table 17.1).
More advanced models differentiate the effect of the pretreatment on the sub-
strate depending on the localization of the essential oil in specialized structures in
the herb or spice, which are ruptured (surface glands or SGl, inner secretory ducts or
SDuct, inner secretory cavities or SCav) during milling, or burst (surface glands) as
a result of swelling of the glands by dissolution of high-pressure CO2 in the gland
contents (essential oils) (Zizovic et al. 2005, 2007a,b,c; Stamenić et al. 2008).
Table 17.1 classifies these so-called Micro-Scale (mS) models, which are further
described in Sect. 17.3.3.
Table 17.1 also summarizes several simplifications of the basic mass balance
and rate equations of mass transfer models used in the literature to simulate SCFE
curves of plant essential oils. Several of these simplifications pertain to the differ-
ential mass balance equation. When assuming a so-called differential bed (DB),
which is valid if the height of the packed bed is comparable with the inner diameter
of the extraction vessel, the second term on the right of the differential mass balance
equation (17.1) is replaced by an expression reflecting a linear variation in the
concentration of the solute in the SCF with the axial position along the bed (Goto
et al. 1993, 1998; Perakis et al. 2005). Alternative simplifications to the differential
mass balance equation include assuming that there is no accumulation (NA) of
solute in the SCF phase, or neglecting the accumulation or first term on the right of
(17.1) (Sovová 1994; Papamichail et al. 2000; Reverchon and Sesti Osséo 1994a);
assuming a uniform extraction (UA) along the bed, or replacing the second term on
the right of (17.1) by an expression reflecting a linear variation in solute concentra-
tion in SCF with the axial position along the bed, which is equivalent to assuming a
“differential” bed (Papamichail et al. 2000; Reverchon and Sesti Osséo 1994a); and
assuming a plug flow (PF), and/or neglecting the axial dispersion in the SCF phase
or the third term on the on the right of (17.1). As summarized in Table 17.1,
assuming the axially dispersed plug flow (ADPF) pattern (Reverchon and Marrone
1997; Goodarznia and Eikani 1998; Reis-Vasco et al. 2000; Spricigo et al. 2001;
Ruetsch et al. 2003; Daghero et al. 2004; Zizovic et al. 2005, 2007b, c; Machmudah
et al. 2006; Steffani et al. 2006; Salimi et al. 2008; Stamenić et al. 2008) that is
implicit in the differential mass balance equation (17.1) presented in Sect. 2.1 is
more the exception than the norm. Some authors (Reverchon et al. 1993a;
Reverchon 1996; Kim and Hong 2002; Gaspar et al. 2003; Campos et al. 2005;
Kotnik et al. 2007) circumvent the differential mass balance (17.1) treating the
packed bed as a vessel with perfect mixing (PM), and others (Esquivel et al. 1996;
Reverchon 1996; Sovová 2005) treating it as perfectly mixed multistages (PMMS)
or a series of perfectly agitated mixing vessels. The perfectly mixed vessel or “hot-
ball”-type model, so called because of the similitude in the mass transfer process
with the cooling of a hot ball of a solid material in a fluid (Reverchon 1997), or
Crank model, so called to honor the author of an influential book on diffusion in
solids (Crank 1975), fail to account for the effect of the increase in solute concen-
tration in the SCF phase along the bed in decreasing the rate of mass transfer.
Models assuming several (n) perfectly mixed vessels in series approach plug flow
as the value of n increases, with the n being an indirect measurement of the axial
dispersion along the packed bed; the treatment of a packed bed as a series of
discrete stages does not comply with the physical situation but is commonly applied
in other separation processes in packed beds such as chromatography (Martin and
Synge 1941).
Besides or instead of the differential mass balance equation (17.1), some models
listed in Table 17.1 simplify the so-called source-and-transfer J term (17.2) that
describes the rate of transfer of essential oil between the herb or spice and the high-
pressure CO2. The linear driving force (LDF) approximation for mass transfer from
the substrate to the SCF (see examples of this type of simplifying assumption in
Table 17.1), which is valid when the residual solute concentration profile in the
partially extracted solid substrate, Cs(r), is approximately parabolic, uses a global
driving force and a single global mass transfer coefficient. The global driving force
for mass transfer equals the difference between the average residual concentration
of solute in the solid, corrected by the equilibrium-partition coefficient ðCs =K Þ and
the concentration of solute in the SCF (Cf). The global mass transfer coefficient (kg,
(17.5), Goto et al. 1993), on the other hand, accounts for both the internal resistance
to mass transfer (related to an internal mass transfer coefficient, ki) and external
resistance to mass transfer (related to the external mass transfer coefficient, kf):
kf
kg ¼ (17.5)
1 þ Bix
where x, a particle-geometry parameter, equals 10 for a sphere and equals 6 for a

thin slab (for which the characteristic dimension dp corresponds to its thickness); Bi
is the dimensionless Biot number (17.6):
k f dp
Bi ¼ (17.6)
De
When both the inner resistance to mass transfer in the plant material and the
external resistance in the SCF film surrounding the particles can be neglected, it is
unnecessary to include the so-called source and transfer J term in the mass balance
and rate equations (17.1–17.4) because under these conditions Cf and Cs are related
by equilibrium. Reis-Vasco et al. (2000) applied this simplifying assumption to the
initial stages of the high-pressure CO2 extraction of pennyroyal essential oil.
The most simple models in Table 17.1 correspond to simple steady-state approx-
imations that assume the rate of extraction is defined by a single mass transfer
controlling resistance, which switches from external control (Ferreira et al. 1999;
Kotnik et al. 2007) to internal control (Kotnik et al. 2007) in a given transition time.
An alternative way to account for the transition from external control to internal
control in a single model is to use a concentration-dependent global mass transfer
coefficient when using the linear driving force approximation to mass transfer, as
done by Coelho et al. (1997). It is important to point out that some models in
Table 17.1 refer to authors who adopted them for SCFE of plant essential oils
instead of the original authors who applied them for alternative SCFE applications.
Specifically, these models are the Diff-Sph/IC model of Crank (1975), the EMTC
model of Brunner (1984), the IMTC model of Hong et al. (1990), the LDF/PF/CDIC
model of Cygnarowicz-Provost (1996), the LDF-Sph/PF model of Catchpole et al.
(1994), Sovová’s (1994) model, and the LDF-D3-Sph/DB model of Skerget and
Knez (2001).
Selected models that are highlighted in Table 17.1 are widely used because they
have relatively simple analytical solutions and are used to best-fit model parameters
with ease. The solution of the so-called Reverchon–Sesti Ossèo (R-SO) model was
applied by Reverchon and Sesti Osséo (1994a), Reverchon et al. (1995a), Papamichail
et al. (2000), Zekovic et al. (2001), Louli et al. (2004), Pfaf-Šovljanski et al. (2005),
and Vargas et al. (2006). On the other hand, the solution of Sovová’s (1994) model
was applied by Mira et al. (1996, 1999), Papamichail et al. (2000), Povh et al.
(2001), Ferreira and Meireles (2002), Sousa et al. (2002, 2005), Martı́nez et al.
(2003, 2007), Rodrigues et al. (2003), Louli et al. (2004), Campos et al. (2005),
Perakis et al. (2005), Vargas et al. (2006), and Bensebia et al. (2009). Sovová’s
model (1994) considers SCFE as a two-stage process, where the mass transfer
coefficient in the first (convection-controlled) stage is kf, whereas the mass transfer
coefficient in the second (internal-diffusion-controlled) stage decreases proportion-
ally to the difference between the solubility of the essential oil in high-pressure CO2
and its actual concentration in the SCF phase in the bed. Povh et al. (2001)
described a procedure to estimate the parameters of Sovova’s model based on the
fitting of an integral extraction plot of solute yield versus specific solvent consump-
tion (Fig. 17.2) to a spline with three straight lines representing successively the
constant, falling, and diffusion-controlled extraction rate periods. Sovová et al.
(1994a) applied an improved version of Sovová’s model (model ICB/ADPF in
Table 17.1) in two aspects: they did not neglect the accumulation of essential oil in
the SCF phase and they considered that the extraction was limited by a decreasing
solubility (LDS) or that the saturation solubility of the essential oil in high-pressure
CO2 decreased during extraction due to the progressive enrichment of the oil
remaining in the partially extracted substrate in less volatile compounds. Authors
using other models include Louli et al. (2004) (LDF/UENA model of Papamichail
et al. 2000), Sousa et al. (2005) (LDF-D3-Slab/DB model of Goto et al. 1993),
Campos et al. (2005) (Diff-Slab/IC model of Gaspar et al. 2003), Zizovic et al.
(2007c) (mS/SGl model of Zizovic et al. 2005; mS/SCav model of Zizovic et al.
2007a; mS/SDuct model of Zizovic et al. 2007b), Stamenić et al. (2008) (mS/SCav
model of Zizovic et al. 2007a; mS/SDuct model of Zizovic et al. 2007b), Langa et al.
(2009) (IBC/PF/PCPR model of Sovová 2005), and Uquiche et al. (submitted)
(Diff/PF model of Araus et al. 2009).
Although most authors of models in Table 17.1 applied the models to their own
data, there are some exceptions. Goodarznia and Eikani (1998) modeled the data of
Reverchon et al. (1993a) on SCFE of essential oils from basil, marjoram, and
rosemary, as well as the data of Sovová et al. (1994a) on SCFE of caraway essential
oil. Germain et al. (2005) also modeled the data of Sovová et al. (1994a) on SCFE
of caraway essential oils. Zizovic et al. (2007b) modeled data of Coelho et al.
(2003) on SCFE of fennel fruit oil. Zizovic et al. (2007c) modeled literature data on
SCFE of essential oils from orange peel (Mira et al. 1996), ginger rhizome (Roy
et al. 1996), clove bud (Reverchon and Marrone 1997), and eucalyptus leaf (Della
Porta et al. 1999). Stamenić et al. (2008) modeled the data of Machmudah et al.
(2006) on SCFE of nutmeg essential oil. Finally, Araus et al. (2009) modeled
literature data on SCFE of essential oils from sage (Reverchon 1996), lavender
(Akgun et al. 2000), oregano (Gaspar 2002), pennyroyal (Reis-Vasco et al. 2000),
and chamomile (Povh et al. 2001).
17.3 Kinetic Parameters of CO2 Extraction of Essential Oils
Table 17.2 summarizes the conditions for the experimental studies on kinetics of
mass transfer during SCFE of essential oils from herbs and spices reviewed in this
chapter. Although most authors in Table 17.2 modeled the results of their own high-
pressure CO2 extraction work, there are some exceptions, including the data of Roy
et al. (1996) on extraction of ginger essential oils at 313 K and 10.8 MPa, the data of
Della Porta et al. (1999) on extraction of eucalyptus essential oil at 323 K and
9 MPa, the data of Coelho et al. (2003) on extraction of fennel essential oil at 313 K
and 9 MPa, and the data of Gaspar (2002) on extraction of oregano essential oil
under selected conditions (310 K and 8 MPa or 320 K and 20 MPa).
In this work, analysis required the estimation of the physical properties of the
loaded CO2 phase produced during the extraction of essential oils. For that purpose,
the authors neglected the changes in physical properties associated with the disso-
lution of essential oils in the solvent under the assayed conditions, and estimated the
density (r) and viscosity (m) of the loaded high-pressure CO2 as a function of the
extraction temperature and pressure using the NIST (2000) database for pure CO2
(del Valle and de la Fuente 2006). On the other hand, D12 was estimated using the
equation of Catchpole and King (1994), which requires reduced temperature (Tr ¼
T/Tc) and reduced density (rr ¼ r/rc, where rc ¼ 467.6 kg/m3 is the critical
density of CO2) of the SCF, and the molecular weight (MW2) and critical volume
(Vc2) of the pseudo-solute. As informed in Sect. 17.2.2, for the purpose of estimat-
ing MW2 and Vc2, representative compounds in families of compounds were
selected, such as monoterpene (MT) hydrocarbons, oxygenated monoterpene
(OMT) compounds, sesquiterpene (ST) hydrocarbons, oxygenated sesquiterpene
(OST) compounds, waxes, and some plant-specific compounds; properties of rep-
resentative compounds to whole families were assigned; and extracts were consid-
ered as pseudo-solutes whose properties were estimated using Kay’s rule as the
weighted average of the properties of the representative compounds in each family
(Poling et al. 2000). The composition of the essential oils was taken from the
literature and was typically determined by gas chromatography analysis of the
steam distillate, hydro-distillate, or CO2-extract of the herb or spice. For calcula-
tions in this current work, only those families representing more than 1% of the
Table 17.2 Summary of extraction conditions of selected mass transfer studies on high-pressure CO2 extraction of plant essential oils in packed beds
412
Substrate Particle diameter Temperature Pressure Superficial velocity Extractor volume L/D ratio of
(dp, mm) (T, K) (P, MPa) (U, mm/s) (V, cm3) extractor (–)
Alecrim pimenta (Sousa et al. 2002) 0.38 298 6.7 0.12 220 27.8
Aniseed (Rodrigues et al. 2003) 0.50 303 8, 10, 14, 0.014–0.025 418 2.03
18
Basil (Reverchon et al. 1993a) 0.17 313 10 0.23 400 3.06
Black pepper (Ferreira et al. 1999; Ferreira 0.080, 0.11 303, 313, 323 15, 20 0.018–0.20 26 29.1
and Meireles 2002)
Black pepper (Perakis et al. 2005) 0.18 313, 323 9, 10, 15 0.20–0.59 751 5.45
Boldo (Uquiche et al. submitted) 0.39, 0.40, 2.36 313 10 0.123 5 2.32
Caraway (Sovová et al. 1994a) 0.38 296, 313 9, 10 0.028–0.048 150 5.30
Carqueja (Vargas et al. 2006) 0.50 313, 323, 9 0.35 4 4.09
333, 343
Celery (Papamichail et al. 2000) 0.21, 0.49 318, 328 10, 15 0.167, 0.456 400 2.90
Chamomile (Povh et al. 2001) 0.30 303, 313 10, 12, 0.030–0.087 204 4.18
16, 20
Chamomile (Kotnik et al. 2007) 0.11 303, 313 10, 15 0.17–0.19 55 4.18
Cinnamon of cunha (Sousa et al. 2005) 0.52 288 6.7 0.065, 0.068 222 27.7
Clove (Reverchon and Marrone 1997) 0.37 323 9 0.244–0.488 400 3.06
Clove (Ruetsch et al. 2003) 0.79 323 9, 12 0.38, 0.76 1,500 2.62
Clove (Daghero et al. 2004) 0.79 323 9, 12 0.38, 1.52 1,500 2.62
Clove (Martı́nez et al. 2007) 0.86 308 10.0 0.039, 0.11 6, 133, 280 0.98, 1.05, 2.20
Eucalyptus (Della Porta et al. 1999) 0.37 323 9.0 0.514 400 6.00
Fennel (Reverchon et al. 1999) 0.37 323 9 0.20–0.61 400 3.06
Ginger (Roy et al. 1996) 0.35 313 11 0.74 2.2 6.67
Ginger (Martı́nez et al. 2003) 1.02 293, 303, 313 15, 20 0.091–0.12 150 13.3
Hop (Pfaf-Šovljanski et al. 2005) 0.488 313 15 0.050 200 3.98
Ho-sho (Steffani et al. 2006) 0.37, 0.50, 1.00 313, 323, 333 8, 9, 10 0.18–0.71 8 7.31
Lavender (Reverchon et al. 1995a) 1.67 321 9 0.31 400 3.06
Lavender (Akgun et al. 2000) 1.20 308, 313, 323 8, 10, 12, 0.44–1.4 39 50.0
14
Marigold (Campos et al. 2005) 0.62 313 12, 15 0.068, 0.202 139 19.1
17
Marjoram (Reverchon et al. 1993a) 0.15 313 10.0 0.23 400 3.06
Nutmeg (Spricigo et al. 2001) 0.30, 0.68, 1.45 296 9 0.032–0.053 35 4.76
Nutmeg (Machmudah et al. 2006) 0.56, 0.69, 2.12 313, 318, 323 1.0, 1.5 0.011, 0.068 500 1.86
Orange (Mira et al. (1996) 0.30, 0.50, 1.50, 323 15 0.084, 0.585 300 2.30
7.50
Orange (Mira et al. 1999) 0.30, 0.50, 1.50, 323 15 0.084, 0.585 300 2.30
7.50
Oregano (Esquivel et al. 1996) 1.10 298, 313 7, 10, 15 0.50–0.63 30 3.99
Oregano (Gaspar 2002) 0.36 310, 320 8, 20 0.22, 0.088 196 2.00
Oregano (Gaspar et al. 2003) 0.33, 0.36, 0.70, 300, 310, 320 7, 8, 10, 0.017–0.058 319 0.47
1.55 15, 20
Oregano (Uquiche et al. submitted) 0.35, 0.36, 2.36 313 10 0.123 5 2.32
Parsley (Louli et al. 2004) 0.29, 0.50 308, 318 10, 15 0.16–0.45 185, 209 1.34, 1.52
Pennyroyal (Reis-Vasco et al. 2000) 0.30, 0.50, 0.70 323 10 0.48–0.97 247–379 3.28–5.02
Peppermint (Goto et al. 1993) 0.12 313, 333, 353 8,8, 14.7, 0.099–0.43 21 2.17
19.6
Rosemary (Reverchon et al. 1993a) 0.23 313 10 0.23 400 3.06
Rosemary (Coelho et al. 1997) 0.72, 1.33 308, 313 10, 12.5, 0.11–0.43 5 12.9
20
Rosemary (Bensebia et al. (2009) 0.44 308, 313 10, 12, 15, 0.25–0.69 125 13.4
18
Sage (Catchpole et al. 1996b) 0.50–1.53 291 7 0.19–0.48 2,959 2.41
Sage (Reverchon 1996) 0.25–3.10 323 9 0.85–1.51 400 5.94

Sage (Langa et al. 2009) 0.3, 0.5, 0.8 313, 323 9, 10 0.15–0.34 1,000 6.00
Spearmint (Kim and Hong 2002) 0.30 312, 322 6.9, 8.5, 0.017–0.068 46 3.54
10.3
Thyme (Zekovic et al. 2001) 3.32, 1.46, 0.70 313 10 0.063 200 3.98
Valerian (Zizovic et al. 2007a) 0.40, 0.65, 0.90 313, 323 10, 15 0.072, 0.147 150 2.35
Valerian (Salimi et al. 2008) 0.59 310 17 0.233 10 12.9
Vetiver (Martı́nez et al. 2007) 0.12 313 20 0.054, 0.150 9, 187 1.45, 1.47
Wild thyme (Stamenić et al. 2008) 0.700 323 10 0.147 150 2.35
413
whole essential oil were considered, and the most concentrated four families in
those cases where five or more were represented in excess of 1% each. Table 17.3
summarizes representative compounds of up to four families for essential oils in
Table 17.2. As an example, the lavender essential oil of Reverchon et al. (1995a)
has 3.33% MTs (main component, myrcene, representing 35.7% of all MTs), 87.9%
OMTs (main component, linalyl acetate, representing 39.4% of all OMTs), 6.14%
STs (main component, b-farnesene, representing 36.3 of all STs), and 2.63% OSTs
(main component, bisabolol, representing 79.5% of all MTs). The values of Vc were
estimated for the representative compounds in the various families using Joback’s
modification of the Lydersen’s group-contribution method (Poling et al. 2000).
Table 17.4 summarizes the estimations of the molecular weight (MW2) and
critical volume (Vc2) of the pseudo-solutes representing the essential oils of the
different herbs and spices reported in Table 17.3. It is clear that the properties of the
pseudo-solutes vary between substrates, as expected, but also for a single substrate
because of differences in the substrates or extracts. Indeed, the essential oils exhibit
differences due to typical variations in biological samples associated with genetic
and processing (e.g., harvest time, drying treatment, and storage condition) factors
(Zetzl et al. 2003). In addition, steam distillates, hydrodistillates, and CO2 extracts
from the same substrate exhibit differences due to thermal and/or oxidative degra-
dation of labile components during distillation, or solubilization of additional
compounds in CO2 with increased solvent power (Moyler 1993; Reverchon
1997). Figure 17.3 shows that the composition of sage essential oils extracted
with high-pressure CO2 at 313 K and 9 MPa, or 323 K and 10 MPa, changes
depending on process conditions and extraction time.
17.3.1 Axial Dispersion Coefficient
Authors del Valle and de la Fuente (2006) showed consistency in the reported
literature values of axial dispersion coefficients for flow of high-pressure CO2 in a
packed bed, when presenting the experimental values in a dimensionless plot of
Dax/D12 versus Pep where Pep is the Peclet number for the particle (17.7), which
in turn corresponds to the product of the dimensionless numbers of Reynolds
(Re, 17.8) and Schmidt (Sc, 17.9).
U dp
Pep ð¼ Re ScÞ ¼ (17.7)
D12
r U dp
Re ¼ (17.8)
m
m
Sc ¼ (17.9)
r D12
Table 17.3 Summary of compositions of plant essential oils from high-pressure CO2 extraction studies in Table 17.2. Besides the name of the main
17
component, for each of the up to 4 fractions the compound type,* percent of fraction, and percent of the main component in the fraction are indicated within
parenthesis
Plant material Fraction 1 Fraction 2 Fraction 3 Fraction 4
Basil (Reverchon and Sesti Osséo Estragole (OMT/89/54) a-trans Bergamotene T Cadinol (ST/2.2/47) –
1994b) (ST/9.0/46)
Lavender I (Reverchon et al. Myrcene (MT/3.33/35.7) Linalyl acetate (OMT/ cis-b-Farnesene (ST/6.14/36.3) a-Bisabolol (ST/2.63/79.5)
1995a) 87.9/39.4)
Lavender II (Akg€ un et al. 2000) Camphor (OMT/56.9) Fenchone (OMT/43.1) – –
Marjoram (Jimenez-Carmona et Sabinene (MT/17.7/42.3) cis-Sabinene hydrate b-Caryophyllene (ST/3.46/100) –
al. 1999) (OMT/78.9/78.6)
Oregano (Gaspar 2002) g-Terpinene (MT/6.78/ Thymol (OMT/85.8/42.4) b-Caryophyllene (ST/7.47/100) –
79.7)
Pennyroyal (Aghel et al. 2004) Limonene (MT/14.6/100) Pulegone (OMT/85.4/ – –
60.9)
Peppermint (Roy et al. 1996) Limonene (MT/2.86/ Menthol (OMT/92.8/ Germacrene (ST/4.35/57.2) –
45.0) 74.5)
Rosemary (Coelho et al. 1997) Limonene (MT/26.0/ Camphor (OMT/65.1/ a-Humulene (ST/7.23/44.1) Caryophyllene oxide (OST/
47.6) 57.6) 1.60/100)
Sage I (Reverchon et al. 1995b) b-Pinene (MT/11.6/21.1) 1,8-Cineole (OMT/70.8/ b-Caryophyllene (ST/14.5/48.8) Manool (OST/3.19/56.1)
76.8)
Sage II (Langa et al. 2009) (313 K/ Myrcene (MT/9.08/35.9) Camphor (OMT/80.4/ b-Caryophyllene (ST/6.19/29.2) Viridoflorol (OST/4.35/39.8)
9 MPa) 62.5)
Sage 2 (Langa et al. 2009) (323 K/ Myrcene (MT/11.4/39.1) Camphor (OMT/79.9/ b-Caryophyllene (ST/5.79/30.4) Caryophyllene oxide (OST/
10 MPa) 69.0) 2.88/15.5)
Spearmint (Özer et al. 1996) Limonene (MT/6.42/ Carvone (OMT/90.6/ b-Bourbonene (ST/2.95/78.6) –
81.9) 89.5)
Thyme (Zekovic et al. 2001) Thymol (OMT/19.6/4.9) b-Caryophyllene (ST/ Tetradecane (HC/78.2/67.1) –
2.25/100)
Wild thyme (Sefidkon et al. 2004) g-Terpinene (MT/52.2/ Thymol (OMT/32.9/58.1) b-Caryophyllene (ST/14.6/42) Spathulenol (OST/0.31/33.3)
Before flowering 42.8)
(continued)
415
416

Wild thyme (Sefidkon et al. 2004) g-Terpinene (MT/60.20/Thymol (OMT/32.09/ Germacrene-D (ST/6.80/85.0) Caryophyllene oxide (OST/
After flowering 42.7) 66.1) 0.91/87.5)
Anise (Rodrı́gues et al. 2003) Anethole (OMT/92.0/ g-Himachalene (ST/2.27/ Isoeugenyl 2-methyl-butyrate
(303 K/8 MPa) 97.7) 100) (OST/5.76/100)
(303 K/10 MPa) 98.3) 100) (ST/2.72/100)
(303 K/14 MPa) 98.2) 100) (OST/5.50/100)
Caraway (Sovová et al. 1994a) Limonene (MT/39.8) Carvone (OMT/60.2) – –
Celery (Mišić et al. 2008) Limonene (MT/28.9/ Sedanenolide (OMT/ b-Selinene (ST/4.89/73.5) Carotol (OST/2.22/78.3)
91.7) 64.0/77.7)
Fennel (Simandi et al. 1999) Limonene (MT/9.37/ Anethole (OMT/90.6/ – –
39.6) 72.2)
Parsley (Louli et al. 2004) Myristicin (PhPro/93.4/
a-Pinene (MT/6.59/49.1) – –
(10 MPa) 57.4)
Parsley (Louli et al. 2004) a-Pinene (MT/2.40/55.8) Myristicin (MT/2.40/ – –
(15 MPa) 55.8)
Carqueja (Simões-Pires et al. b-Pinene (MT/3.42/75.0) Carquejyl acetate (OMT/ Ledol (OST/53.9/38.6) –
2005) 42.7/90.7)
Chamomile I (Povh et al. 2001) b-Farnesene (ST/39.6/ a-Bisabolol (OST/60.4/ – –
86.3) 47.0)
Chamomile II (Kotnik et al. 2007) Matricine (OST/100) – – –
(303 K/10 MPa)
Chamomile II (Kotnik et al. 2007) Matricine (OST/96.5) – – –
(303 K/15 MPa)
(313 K/10 MPa)
(313 K/15 MPa)
Marigold (Danielski et al. 2007) Tetradecanoic acid Octacosane (HC/94.6/ Cholest-4-en-3-one-14-methyl –
(313 K/12 MPa) (LOHC/2.46/100) 94.6) (HOHC/2.94/40.7)
17
Marigold (Danielski et al. 2007) Octacosane (HC/98.7/ Taraxasterol (H-OHC/ – –

(313 K/15 MPa) 43.1) 1.35/34.2)
Alecrim pimenta (Sousa et al. p-Cymene (MT/3.22/ Thymol (OMT/74.8/67.9) b-Caryophyllene (ST/20.0/72.2) Caryophyllene oxide (OST/
2002) 45.0) 1.97/100)
Black pepper I (Ferreira et al. Limonene (MT/66.2/ b-Caryophyllene (ST/ – –
1999) 30.1) 33.8/21.8)
Black pepper II (Ferreira et al. Limonene (MT/1.00/ b-Caryophyllene (ST/ – –
1999) 57.8) 99.0/70.9)
Black pepper III (Perakis et al. Limonene (MT/31.2/ Linalool (OMT/1.57/ b-Caryophyllene (ST/34.2/35.5) Caryophyllene oxide (OST/
2005) 29.8) 51.3) 33.0/51.5)
Boldo (Miraldi et al. 1996) p-Cymene (MT/9.28/ Ascaridole (OMT/81.75/ Guaizulene (ST/8.96/100) –
94.7) 26.5)
Cinnamon of Cunha (Sousa et al. Anethole (OMT/96.5/ Germacrene – –
2005) 95.8) (ST/3.51/35.7)
Clove (Della Porta et al. 1998) Eugenol (OMT/86.3/ b-Caryophyllene (ST/ – –
77.5) 13.7/82.2)
Eucalyptus (Della Porta et al. a-Pinene (MT/12.2/86.1) 1,8-Cineole (OMT/67.1/ Aromadendrene (ST/12.3/65.3) Guaiol (OST/8.40/50.0)
1999) 93.3)
Ginger (Martı́nez et al. 2003) Zingiberene (OMT/9.77/ a-Zingiberene (ST/45.6/ Farnesol (OST/3.73/30.3) 6-Gingerol (Ging/40.9/28.9)
56.7) 35.6)
Hop (Pfaf-Šovljanski et al. 2005) a-Humulene (ST/13.5/ Isohumulone (a-acid/ Lupulone (b-acid/73.5/53) –
82.4) 13.0/100)
Ho-sho (Bakkali et al. 2005) Sabinene (MT/21.7/35.5) 1,8-Cineole (OMT/75.0/ Viridiflorol (OST/3.29/100) –
80.0)
Nutmeg I (Spricigo et al. 1999) Sabinene (MT/78.8/51.9) Myristicin (OMT/21.2/ – –
36.8)
Nutmeg II (Machmudah et al. 3-Cyclohexene-1-ol b-Phellandrene (MT/ Myristicin (OMT/36.2/89.9) Copaene (ST/1.83/68.9)
2006) (LOHC/15.7/26.4) 46.3/42.9)
Orange (Budich and Brunner Limonene (MT/98.3/ Linalool (OMT/1.75/ – –
1999) 97.6) 28.8)
Valerian I (Zizovic et al. 2007a) Borneol (OMT/15.9/36.8) Valerena-4,7(11)-diene Valerenal (OST/67.4/19.3) (E)-Valerenyl isovalerate (H-
(313 K/10 MPa) (ST/14.1/25.7) OHC/2.68/70.0)
417
(continued)
418

Valerian II (Zizovic et al. 2007a)
Isovaleric acid (LOHC/ Bornyl acetate (OMT/ Valerena-4,7(11)-diene (ST/ Valerianol (OST/66.2/19.7)
(313 K/10 MPa) 6.01/100) 13.9/59.3) 13.8/17.2)
Isovaleric acid (LOHC/ Bornyl acetate (OMT/ b-Bisabolene (ST/13.2/17.7) Valerianol (OST/68.4/16.6)
(313 K/15 MPa) 7.67/100) 10.7/61.6)
(323 K/10 MPa) 4.37/100) 12.1/58.9) 13.3/17.1)
(323 K/15 MPa) 3.57/100) 10.7/59.3) 13.3/17.1)
Valerian III (Zizovic et al. 2007a)
Isovaleric acid (LOHC/ Bornyl acetate (OMT/ d-Elemene (ST/28.6/26.9) Valerenal (OST/45.0/34.9)
9.64/100) 16.7/52.9)
*MT indicates a monoterpene hydrocarbon; OMT, an oxygenated monoterpene; ST, a sesquiterpene hydrocarbon; OST, an oxygenated sesquiterpene; HC, an
hydrocarbon; LOHC, a light oxygenated hydrocarbon; HOHC, a heavy oxygenated hydrocarbon; PhPro, a phenyl propanoid; and Ging, a gingerol
Table 17.4 Summary of estimated molecular weights (MW) and critical volumes (Vc) of plant essential oils in Table 17.3
17
Plant material Scientific name Family Identified Molecular weight Critical volume
compounds (%) (MW, Da) (Vc, cm3/mol)
Basil Ocimum basilicum Lamiaceae 99.4 153.0 508.0
Lavender I Lavandula angustifolia Lamiaceae 100.1 194.5 682.6
Lavender II Lavandula angustifolia Lamiaceae 76.9 152.2 503.5
Marjoram Origanum majorama Lamiaceae 36.8 152.0 517.7
Oregano Origanum vulgare Lamiaceae 87.0 152.2 440.3
Pennyroyal Mentha pulegium Lamiaceae 100.0 149.7 504.1
Peppermint Mentha piperita Lamiaceae 99.3 157.2 527.2
Rosemary Rosmarinus officinalis Lamiaceae 100.0 151.7 523.6
Sage I Salvia officinalis Lamiaceae 100.0 160.0 546.0
Sage II (313 K/9 MPa) Salvia officinalis Lamiaceae 81.4 155.2 519.8
Sage II (323 K/10 MPa) Salvia officinalis Lamiaceae 87.1 153.8 153.8
Spearmint Mentha spicata Lamiaceae 100.0 150.4 508.5
Thyme Thymus vulgaris Lamiaceae 11.9 186.8 719.0
Wild thyme, before flowering Thymus serpyllum Lamiaceae 98.0 148.1 506.6
Wild thyme, after flowering Thymus serpyllum Lamiaceae 88.2 144.3 500.4
Anise (303 K/8 MPa) Pimpinella anisum Apiaceae 99.2 152.7 500.0
Caraway Carum carvi Apiaceae 98.0 144.3 505.2
Celery Apium graveolens Apiaceae 99.8 172.7 585.9

Fennel Foeniculum vulgare Apiaceae 100.0 147.0 487.7
Parsley (10 MPa) Petroselinum crispum Apiaceae 83.4 187.1 533.1
Parsley (15 MPa) Petroselinum crispum Apiaceae 96.2 193.0 552.1
Carqueja Baccharis trimera Asteraceae 83.7 204.3 678.6
Chamomile I Matricaria recutita Asteraceae 44.4 214.9 790.6
Chamomile II (303 K/10 MPa) Matricaria recutita Asteraceae 7.56 306.4 886.5
419
(continued)
420
Plant material Scientific name Family Identified Molecular weight Critical volume
compounds (%) (MW, Da) (Vc, cm3/mol)
Marigold (313 K/12 MPa) Calendula officinalis Asteraceae 30.9 387.9 1,565.1
Marigold (313 K/15 MPa) Calendula officinalis Asteraceae 28.2 395.2 1,601.4
Alecrim pimenta Lippia sidoides Verbenaceae 95.2 159.0 470.4
Black pepper I Piper nigrum Piperaceae 99.6 153.5 562.1
Black pepper II Piper nigrum Piperaceae 99.2 203.3 719.3
Black pepper III Piper nigrum Piperaceae 73.3 179.7 633.9
Boldo Peumus boldus Monimiaceae 98.2 166.6 525.1
Cinnamon of Cunha Croton zehntneri Euphorbiaceae 95.7 149.6 149.6
Clove Syzygium aromaticum Myrtaceae 98.5 168.8 454.7
Eucalyptus Eucalyptus globules Myrtaceae 100.0 155.0 528.6
Ginger Zingiber officinale Zingiberaceae 71.7 241.1 799.0
Hop Humulus lupulus Cannabaceae 93.3 364.5 1,224.8
Ho-sho Cinnamomum camphora Lauraceae 83.7 151.4 516.2
Nutmeg I Myristica fragrans Myristicaceae 89.6 145.2 498.9
Nutmeg II Myristica fragrans Myristicaceae 99.6 143.5 143.5
Orange Citrus sinensis Rutaceae 100 136.5 508.49
Valerian I (313 K/10 MPa) Valeriana officinalis Valerianaceae 83.5 204.6 712.1
Valerian II (313 K/10 MPa) Valeriana officinalis Valerianaceae 89.1 201.9 688.4
Valerian III Valeriana officinalis Valerianaceae 82.7 190.2 662.7
Fig. 17.3 Effect of process time and conditions on the composition of sage essential oils extracted
with high-pressure CO2: monoterpenes, oxygenated monoterpenes, sesqui-
terpenes, oxygenated sesquiterpenes, and total essential oil components extracted at
313 K and 9 MPa; and monoterpenes, oxygenated monoterpenes, sesquiter-
penes, oxygenated sesquiterpenes, and total essential oil components extracted at
323 K and 10 MPa (Adapted from Langa et al. 2009)
In a related work, del Valle and Catchpole (2005) correlated literature data for
the axial dispersion of high-pressure CO2 in packed beds (Tan and Liou 1989;
Catchpole et al. 1996a; Funazukuri et al. 1998; Yu 1998; Ghoreishi and Akgerman
2004). Typical measurements in these studies were conducted by injecting a pulse
of solute (gaseous methane, volatile acetone or hexachlorobenzene, or low-volatility

benzoic acid, oleic acid, or squalene) into the high-pressure CO2 stream before a
bed (0.4 D 50.8 mm) packed with different materials (sand particles, glass
beads, or steel beads ranging in size from dp ¼ 0.05 to dp ¼ 3.18 mm), and
evaluating the dispersion (concentration profile) of the solute in the stream leaving
the bed. Different flow directions (horizontal, upward, downward) and regimes
(from molecular-transport-controlled, Ped ~ 9.1 103, to convection-controlled
flow conditions, Ped ~ 1.2 103) were assayed in these studies.
Figure 17.4 summarizes the results of the correlation study of del Valle and
Catchpole (2005). When dispersion is controlled by molecular transport (Ped 1)
the ratio Dax/D12 has a nearly constant value (slightly below 1 in this particular
case), as expected (Dullien 1992). On the other hand, under convention-controlled
conditions (Ped > 1), the ratio Dax/D12 increases exponentially with Ped, which
results in a nearly straight line having a slope between 1 and 2 when presenting the
data in a log–log plot, as also expected (Dullien 1992). Data scattering for large
values of Ped is partially explained by the scale of the measurements; data obtained
using large vessels exhibited a larger axial dispersion coefficient than data obtained
using small tubes (del Valle and Catchpole 2005). Dispersion typically increases as
the length of the bed (L) increases due to the increase in the residence time in the
packed bed (Han et al. 1985). Dispersion also increases as the diameter of the bed
(DE) decreases because of the pronounced decrease in local porosity in the radial
direction from the wall of the tube or vessel when the ratio DE/dp decreases, which
increases concentration-gradient-driven radial diffusion and consequently increases
Fig. 17.4 Dimensionless plot of the axial dispersion coefficient (Dax) in a packed bed operating
under a high-pressure CO2 and flow regime. The independent variable is the ratio Dax/D12 [where
D12 is a binary diffusion coefficient of the solute (component 2) in CO2 (component 1)] and the
dependent variable is the Peclet number for the particle (17.5)
axial dispersion of the solute (Fahien and Smith 1955). It is apparent for the results
in Fig. 17.4 that the effect of an increase in L/dp predominates over the effect of an
increase in D/dp, so that the axial dispersion is larger in large vessels than small
tubes (for purposes here, L/dp 250 in small packed beds). Thus, del Valle and
Catchpole (2005) selected Ped and the ratio L/dp as the independent variables in
their correlation for Dax/D12 (17.10).
2
Dax 0:530 Pep
¼ 0:540 þ : (17.10)
D12 Pep
1 þ 42:8 L d
=p
Figure 17.4 reports experimental measurements of axial dispersion using closed

circles for small packed beds (experimentally 89 L/dp 234) and open circles
for large beds (295 L/dp 2,320), and includes the limit between the two
regions predicted by (17.10) for L/dp ¼ 250 in the form of a segmented line.
In this work, the authors believe that a precise estimation of the value of the axial
dispersion coefficient is less important than determining whether axial dispersion
phenomena affects SCFE of plant essential oils to such an extent that the term with
Dax must be included in the differential mass balance equation (last term on the left
of (17.1)). This determination is important because axial dispersion phenomena is
typically neglected to simplify the fitting of model parameters. del Valle and de la
Fuente (2006) analyzed the effect of dispersion in mass transfer for SCFE of
oilseeds under typical industrial conditions based on the claim of Goto et al.
(1996) that it can be disregarded in mass transfer models when the value of the
Peclet number for the packed bed (PeL, 17.11) is above ca. 100.
UL
PeL ¼ (17.11)
Dax e
Recommended conditions for SCFE of plant essential oils are 323 K and 9 MPa
(Reverchon 1997), for which the following values of physical properties were
estimated in this work: r ¼ 285 kg/m3; m ¼ 2.47 105 Pa s; D12 ffi 3.00 108
m2/s; and, Sc ¼ 2.86. Conditions favoring axial dispersion in the extraction
vessel include a large superficial solvent velocity (e.g., U ¼ 5 mm/s) (Eggers
1996), a large particle size, a small bed voidage, and a tall vessel (see 17.10 and
17.11). Eggers (1996) mentioned that the aspect ratio (L/D) of a typical extraction
vessel is between 4 and 6, while Reverchon (1997) specified that an industrial SCFE
facility for plant essential oils (Essences, Salerno, Italy) used vessels of 0.3 m3, so
in this work a packed bed L ¼ 2.40 m tall and D ¼ 40 cm wide was selected for
calculations (L/D ¼ 6, V ¼ 0.3 m3). According to the analysis, dispersive effects
can be neglected (PeL 100) for small particles of dp 0.71 mm, but for larger
particles, it is necessary to reduce the bed voidage from e 0.59 for dp ¼ 1.0 mm
to e 0.34 for dp ¼ 1.5 mm.
As shown in Table 17.1, most authors neglect the contribution of axial dispersion
in their SCFE models for plant essential oils from solid substrates in packed beds.
There are exceptions (studies where the axial dispersion coefficient is estimated
using a literature correlation) including Ruetsch et al. (2003) and Daghero et al.
(2004), who used the correlation of Wakao and Kaguei (1982); Goodarznia and
Eikani (1998), Spricigo et al. (2001), Gaspar et al. (2003), Steffani et al. (2006),
Zizovic at al. (2005, 2007a,b,c), Salimi et al. (2008), and Stamenić et al. (2008),
who used the correlation of Tan and Liou (1989); and Machmudah et al. (2006),
who used the correlation of Funazukuri et al. (1998). On the other hand, Reis-Vasco
et al. (2000) used Dax as a fitting parameter for their experimental data, but their
best-fit values decreased when the superficial solvent velocity of the CO2 increased
(which was not as expected), and were 31–138 times larger than predicted using
(17.10).
17.3.2 External Mass Transfer Coefficient
There are two values of kf (expressed in, e.g., m/s) and kfap (expressed in, e.g., s1)
reported in the literature as external mass transfer coefficients. Thus, in this work all
values were recalculated first in similar units. To do this, the specific surface (ap), or
total particle surface (Sp) per unit volume of the packed bed [(1 e) Vp], was
estimated using (17.12):
c
ap ¼ : (17.12)
ð 1 e Þ dp
In (17.12), c is a factor that depends on the geometry of the particles: c tends to

2 for very thin slabs (approaching a so-called infinite slab geometry), for which dp is
the thickness of the slab; and c ¼ 6 for spheres. The void fraction in the packed bed
(e) in (17.12) ranged from 0.24 for milled black pepper (Ferreira et al. 1999) to
0.901 for rapidly decompressed oregano (Uquiche et al., submitted). We assumed a
typical value e ¼ 0.6 in absence of a reported value (Mira et al. 1996; Papamichail
et al. 2000; Louli et al. 2004; Zizovic et al. 2007a).
Figure 17.5 summarizes corrected values of kf from the literature in a dimen-
sionless plot of Sh/Sc1/3 versus Re, where Sh is the dimensionless Sherwood number
defined in (17.13):
k f dp
Sh ¼ (17.13)
D12
Figure 17.5 also includes reference lines corresponding to correlations in the

literature for the external mass transfer coefficient in packed beds operating with
liquids, gases, and SCFs. The correlations include the equation of Wakao and
Kaguei (1982) for mass transfer in packed beds with low-pressure liquids and
Fig. 17.5 Dimensionless plot of Sh Sc1/3 versus Re [where Sh, Sc, and Re are the dimensionless
Sherwood (17.11), Schmidt (17.7), and Reynolds (17.6) numbers] for literature values of the
external mass transfer coefficient (kf) in SCFE of plant essential oils as a function of the flow
regime in the packed bed. Plotted values were estimated from best-fitting parameters reported by
the identified authors using the following models (model names reported in Table 17.1): EMTC
model applied by Ferreira et al. (1999) and Kotnik et al. (2007); R-SO model applied by
Papamichail et al. (2000) and Louli et al. (2004); LDF/UENA model applied by Papamichail
et al. (2000) and Louli et al. (2004); LDF/PF/CDIC model applied by Coelho et al. (1997);
LDF-Sph/PF model applied by Esquivel et al. (1996); LDF/ADPF model applied by Reverchon
and Marrone (1997); Sovova’s model applied by Mira et al. (1996, 1999), Papamichail et al.
(2000), Povh et al. (2001), Ferreira and Meireles (2002), Sousa et al. (2002, 2005), Martı́nez et al.
(2003, 2007), Rodrigues et al. (2003), Louli et al. (2004), Campos et al. (2005), Perakis et al.
(2005), Vargas et al. (2006), and Bensebia et al. (2009); IBC/PF model applied by Louli et al.
(2004); IBC/PFNA model applied by Sovová et al. (1994b); IBC/PF/PCPR model applied
by Sovová (2005), Machmudah et al. (2006), and Langa et al. (2009); DDD/PM model applied
by Kim and Hong (2002); SC/PF model applied by Germain et al. (2005); and SC/ADPF
model applied by Steffani et al. (2006)
gases (17.14), which is valid for 3 < Re < 3,000 and 0.5 < Sc < 104, and the
equation of Puiggené et al. (1997) for evaporation of 1,2-dichlorobenzene from
glass beads into a high-pressure CO2 stream (17.15), which is valid for 10 < Re < 100
and Sc < 10.
Sh ¼ 2 þ 1:1 Re0:6 Sc0:33 (17.14)
Sh ¼ 0:206 Re0:8 Sc0:33 (17.15)
The experimental values of Sc for the kinetic studies in Table 17.2 ranged from
2.05 to 17.0, and this result conditioned the selection of values for the two lines
corresponding to the correlation of Wakao and Kaguei (1982) in Fig. 17.5.

The values of mass transfer coefficients predicted by Wakao and Kaguei (1982)
for liquids and gases (17.14) are above those predicted by the literature correlations
for mass transfer in packed beds operating with SCFs under forced convection (del
Valle and de la Fuente 2006). Among the specific correlations for mass transfer in
packed beds operating with SCFs, the predictions made in the equation of King and
Catchpole (1993) are above those made in the equation of Tan et al. (1988), which
are in turn above those made in (17.15) (del Valle and de la Fuente 2006). Thus, the
expected estimated values of Sh/Sc1/3 derived from the studies in Table 17.2 should
be seen between the top and bottom lines in Fig. 17.5.
Figure 17.5 shows that experimental values of Sh/Sc1/3 for the SCFE of plant
essential oils exhibit considerable scattering (a span covering 3 orders of magnitude
was observed, e.g., with data of Kim and Hong 2002), and are generally smaller (up
to several orders of magnitude) than those predicted by the literature correlations.
Most best-fit values were below the predictions of (17.15) with the exception of
values of Papamichail et al. (2000) and Martı́nez et al. (2007), and a single value of
Mira et al. (1999) using Sovová’s model (Table 17.1), some values of Papamichail
et al. (2000) and Louli et al. (2004) using model LDF/UENA, values of Machmudah
et al. (2006) and Langa et al. (2009) using model IBC/PF/PCPR, and values of
Steffani et al. (2006) using model SC/ADPF (who used (17.14) to get first guess
values for kf). Smaller best-fit values than predictions of literature correlations are
probably due to a combination of several factors, including underestimation of the
contribution of internal (solid phase) mechanisms to the total resistance to mass
transfer, overestimation of the mass transfer area, underestimation of solvent flow
heterogeneity effects, and underestimation of natural convection effects.
Some kinetic models assume that the extraction rate is controlled by external
resistances to mass transfer (Ferreira et al. 1999; Kotnik et al. 2007). If this is not
the case, the best-fit value of kf would be smaller than in model systems without
internal resistance to mass transfer, as those used to obtain the two correlations
reported in Fig. 17.5, to compensate the contribution to the total resistance of the
solid matrix (del Valle and de la Fuente 2006). This will also occur if the internal
resistance to mass transfer is not neglected, but underestimated.
An overestimation of ap would also be compensated by an underestimation of
the value of kf in cases where an overall mass transfer coefficient (kfap) is used as a
fitting parameter for experimental data (del Valle and de la Fuente 2006). In the
work presented in this chapter the assumption is that the solid particles were
spherical ((17.12) for c ¼ 6), and although spheres have the smallest surface-
area-to-volume ratio among regular geometric shapes, not all of the external surface
of the particles is fully available for extraction due to tight packing in low-porosity
beds (Marrone et al. 1998) or agglomeration of solid particles (Štastová et al. 1996;
Eggers et al. 2000; del Valle et al. 2008).
Other causes of a reduction in the extraction rate and an associated reduction in
the best-fitting value of kf are irregularities in the packing of solid particles and other
nonidealities that change the interstitial velocity of the high-pressure CO2 across the
extraction vessel (del Valle et al. 2004). Indeed, it is possible to have high-velocity
zones near the wall of an extraction vessel coexisting with low-velocity zones close
to the axis of the vessel, with the end result of smaller extraction rates and smaller
values of kf in the axis than near the wall of the vessel (Sovová et al. 1994b; del
Valle et al. 2004). These changes in velocity can be explained by the radial
variations in bed porosity that result when packing comparatively large particles
in small-diameter vessels (i.e., when dp/D < 10 for mono-disperse particles) or
radial variations in solvent viscosity when heating the packed-bed contents by
conduction through the wall of the extraction vessel. Consequently with this
hypothesis Brunner (1994) reported that the residual content of theobromine in
ground cocoa seed shells increases when moving from the vessel wall toward the
center of the extraction vessel in large-scale extraction experiments, and that these
changes in residual solute content are more pronounced when the superficial
velocity of high-pressure CO2 decreases from 3.5 to <1.7 mm/s. The global effect
of these changes is a reduction in the average value of the external mass coefficient
(del Valle et al. 2004).
The last explanation for the differences between experimental and correlated
values of kf is neglecting the undesirable effects of natural convection on the
external mass transfer. The natural convection phenomenon is important in experi-
ments where high-pressure CO2 moves slowly upwards in a vertical extraction
vessel and against the gradient in density (Dr ¼ rsat r) that develops when
essential oils dissolve into the CO2 stream, a condition under which the loaded
CO2 phase moves down under the influence of the force of gravity (St€uber et al.
1996; Puiggené et al. 1997; Germain et al. 2005). Thus, the extraction rate in a
SCFE system using CO2 upflow conditions is smaller than in a system using
downflow conditions due to the negative influence of natural convection on mass
transfer opposed by gravity (Sovová et al. 1994a, b; St€uber et al. 1996; Germain
et al. 2005). This is important to consider in SCFE of plant essential oils because of
several factors favoring undesirable convection phenomena (Germain et al. 2005):
(1) the preference of solvent upflow conditions in industrial practice to improve
extraction by fluidization of small particles and avoidance of compaction of the
packed bed; (2) the use of near-critical conditions for extraction (Table 17.2); (3)
the large solubility of essential oil components in high-pressure CO2 under typical
extraction conditions (Sect. 17.4); and (4) the use of a small superficial solvent
velocity in experimental studies, as reported in the literature. Both correlations
presented in Fig. 17.5 apply under forced convection conditions, but not when
natural convection effects start to dominate as the ratio Gr/Re2 increases (>> 1),
where Gr is the dimensionless Grashof number defined in (17.16):
g dp3 r2 Dr
Gr ¼ : (17.16)
csat m2
Germain et al. (2005) showed that for the extraction of caraway essential oils
with high-pressure CO2 at 313 K and 9–10 MPa (Sc ¼ 2.13), another correlation of
Puiggené et al. (1997) that takes into account the effect of natural convection
phenomena predicts smaller values of the external mass transfer coefficient when
mass transfer is opposed by gravity (CO2 upflow conditions) than when it is aided
by gravity (CO2 downflow conditions), whereas there are virtually no differences in
Sh/Sc1/3 for a relatively large Re (e.g., Re 50). For a value of Re slightly below
(Re 10) the value of Sh/Sc1/3 drops pronouncedly; this drop is only slightly
dependent on the value of Gr (for 10 < Gr < 38,000), but shifts to smaller values
of Re as the value of Sc decreases (e.g., Re 2.5 for Sc ¼ 34.4). This helps to
explain some of the experimental results reported in Fig. 17.5.
In selected cases, the value of kf was not fitted to a cumulative extraction plot,
but instead adopted from a literature correlation for forced convection. Goto et al.
(1993, 1998), Spricigo et al. (2001), and Steffani et al. (2006) used the equation of
Wakao and Kaguei (1982). Several authors used the equation of Tan et al. (1988),
including Reverchon et al. (1993a), Reverchon (1996), Goodarznia and Eikani
(1998), Perakis et al. (2005), Zizovic et al. (2005, 2007a,b,c), Salimi et al. (2008),
and Stamenić et al. (2008). Catchpole et al. (1996b), Gaspar et al. (2003), Ruetsch
et al. (2003), Daghero et al. (2004), and Machmudah et al. (2006) used the equation
of King and Catchpole (1993). Araus et al. (2009) and Uquiche et al. (submitted)
used the equation of Puiggené et al. (1997). It is relevant to mention that in most of
these studies, the selected equations were applied for values of Re below the limit
where the natural convection effects must be taken into account. Two exceptions
where natural convection phenomenon was accounted for are the works of Akgun
et al. (2000), who used the equation of Lee and Holder (1995), and Germain et al.
(2005), who use an equation of Puiggené et al. (1997).
17.3.3 Effective Diffusivity in the Solid Matrix
To compare the best-fit internal mass transfer parameters in the reviewed studies in
this work, the value of a microstructural factor FM (17.17) that embodies all effects
of the substrate and its pretreatment on inner mass transfer was estimated (Aguilera
and Stanley 1999; del Valle and de la Fuente 2006; Araus et al. 2009; Uquiche et al.
submitted):
D12
FM ¼ : (17.17)
De
The term FM gives the retarding effect of the solid matrix on the mass transfer
rate by indicating how many times smaller the effective diffusivity of the essential
oil is in the treated plant material than its binary diffusion coefficient in high-
pressure CO2. It is convenient to use FM instead of De in modeling SCFE processes
because it is independent of extraction temperature and pressure, interstitial solvent
velocity, and substrate particle size (Araus et al. 2009).
There are some problems in expressing the internal mass transfer parameters as
reported in the literature in consistent units. For example, some contributions
describe the best-fit value of an internal mass transfer coefficient ki (expressed in,
e.g., m/s) (Reverchon et al. 1999; Reis-Vasco et al. 2000) or the product kiap
(expressed in, e.g., s1) (Sovová et al. 1994a; Esquivel et al. 1996; Mira et al.
1996, 1999; Coelho et al. 1997; Ferreira et al. 1999; Papamichail et al. 2000; Povh
et al. 2001; Ferreira and Meireles 2002; Sousa et al. 2002, 2005; Martı́nez et al.
2003, 2007; Rodrigues et al. 2003; Louli et al. 2004; Campos et al. 2005; Perakis
et al. 2005; Sovová 2005; Machmudah et al. 2006; Vargas et al. 2006; Bensebia
et al. 2009; Langa et al. 2009), and in these cases the effective diffusivity was
computed using the relationship between De and the reciprocal of kiap (or the
so-called internal diffusion time ti, s) proposed by Villermaux (1987) and reported
in (17.18):
k i dp
De ¼ (17.18)
ð 1 eÞ x
where the particle-geometry parameter x was defined in connection with (17.5) in

Sect. 2.3 (x ¼ 10 for a sphere, x ¼ 6 for a for thin slab).
When values of the effective diffusivity were reported, those cases were distin-
guished where they referred to the movement of the pseudo-solute in the solid (e.g.,
diffusion model), from those referring to movement in the fluid phase trapped in the
solid (e.g., DDD models, SC models). Under the latter conditions, the values of De
can be estimated according to (17.19) (del Valle and de la Fuente 2006):
De ¼ D0e K (17.19)
0
where De (the effective diffusivity of the pseudo-solute in the fluid trapped within
the pores of the solid matrix) is the reported value, and K, the partition of the
pseudo-solute between the SCF and the herb or spice which, in absence of a
reported value (Sovová et al. 1994a; Mira et al. 1996, 1999; Coelho et al. 1997;
Ferreira et al. 1999; Papamichail et al. 2000; Povh et al. 2001; Ferreira and Meireles
2002; Martı́nez et al. 2003, 2007; Rodrigues et al. 2003; Louli et al. 2004; Campos
et al. 2005; Perakis et al. 2005; Vargas et al. 2006; Bensebia et al. 2009), is
estimated using (17.20):
Cfo
K¼ (17.20)
Cso
Table 17.4 is an example of computed values of FM in the case of SCFE of

essential oils from oregano bracts, reporting the effect on FM of different indepen-
dent variables such as extraction temperature, extraction pressure, superficial
velocity of the CO2, sample pretreatment, and sample particle size. It is important
to note that in those experiments where Esquivel et al. (1996), Gaspar (2002) (data
modeled by Araus et al. 2009), and Gaspar et al. (2003) studied the effect of
extraction temperature and/or extraction pressure, they also changed the superficial
velocity of the CO2, because they kept the mass flow rate constant instead. Indeed,
the superficial velocity of the CO2 (U) depends not only on its mass flow rate (Q),
but also on its density (r) under process conditions, which is a strong function of the
extraction temperature and pressure, especially under near-critical conditions,
according to (17.21):
4Q
U¼ (17.21)
p D2E r
As proposed by Araus et al. (2009), it was expected that the values of FM are
dependent on sample pretreatment, but independent of process temperature, process
pressure, CO2 superficial velocity, and substrate particle size, but clearly this is not
the case (Table 17.4). Some of the differences can be imputed to typical variations
in biological materials associated with differences in genetic makeup, growing
environment, and harvest time (Zetzl et al. 2003). Although some differences
were expected between the samples assayed by Esquivel et al. (1996), Gaspar
(2002), Gaspar et al. (2003), and Uquiche et al. (submitted), which explains why
the differences were smaller between experiments in a single study than among the
four studies, the percent differences in estimated values of FM between experiments
in single studies were clearly still large. The very large difference in values of FM
between the sample with dp ¼ 1.55 mm (untreated sample) and all other samples in
the study of Gaspar et al. (2003) can be explained by the positive effect of sample
milling in reducing internal resistance to mass transfer. Uquiche et al. (submitted)
also imputed the differences in values of FM between their samples to microstruc-
tural differences caused by conventional milling, low-temperature milling, and
rapid decompression of oregano and provided microscopy evidence of these differ-
ences. In work presented in this chapter the composition of the essential oil sample
of oregano reported by Gaspar (2002) was used as representative of the extracts
obtained in all experiments (Table 17.5); however, this assumption may be errone-
ous because of the aforementioned variability exhibited by biologic materials, as
well as the variability in extract composition with process temperature and process
pressure (i.e., extracts obtained using higher density CO2 are expected to contain
heavier and more polar compounds than the sample analyzed by Gaspar). As it will
be analyzed later in this section, the errors introduced when not accounting for the
actual composition are not as large as those observed in Table 17.5 and, in addition,
differences were not expected in extract composition as used in experiments done
by Gaspar et al. (2003) to assess the effect on extraction kinetics of the superficial
solvent velocity, or the size of milled particles (Table 17.5). In this work, the
authors believe mathematical models adopted and their ability to fit the physical
picture of the actual extraction process explain the differences among estimated
values of FM reported in Table 17.5 to a large extent. For example, in the work of
Gaspar et al. (2003) average values of FM were 7.32 105 (range of value between
2.65 105 and 5.72 106) using model IBC-Diff (Table 17.5), 4.28 105
(range: 1.46 105 to 3.81 106) using model Diff-Slab/IC, and 4.19 105
Table 17.5 Values of microstructural factor for high-pressure CO2 extraction of essential oils
from oregano in selected studies from Tables 17.1 and 17.2 as a function of extraction conditions
(system temperature and pressure, superficial velocity) and sample pretreatment and particle size
Studied effect Independent variable Microstructural Reference
factor (FM, –)
Process conditionsa T ¼ 298 K/P ¼ 7 MPa 34,900 Esquivel et al.
T ¼ 313 K/P ¼ 10 MPa 3,600 (1996)
T ¼ 313 K/P ¼ 15 MPa 1,880
Process pressureb P ¼ 7 MPa 409,000 Gaspar et al.
P ¼ 8 MPa 313,000 (2003)
P ¼ 10 MPa 265,000
P ¼ 15 MPa 277,000
Process T ¼ 300 K 277,000 Gaspar et al.
temperaturec T ¼ 310 K 308,000 (2003)
T ¼ 320 K 390,000
Superficial solvent U ¼ 0.017 mm/s 593,000 Gaspar et al.
velocityd U ¼ 0.029 mm/s 542,000 (2003)
U ¼ 0.040 mm/s 519,000
U ¼ 0.052 mm/s 479,000
Sample particle dp ¼ 0.330 mm 440,000 Gaspar et al.
sizee dp ¼ 0.360 mm 440,000 (2003)
dp ¼ 0.700 mm 477,000
dp ¼ 1.550 mm 5,720,000
Sample Conventionally milled 2,650 Uquiche et al.
pretreatmentf (dp ¼ 0.354 mm) (submitted)
Low-temperature-milled 2,540
(dp ¼ 0.339 mm)
Rapidly decompressed 1,000
(dp ¼ 0.844 mm)
p ¼ 1.10 mm/Q ¼ 8.33 g CO2/min
a
b
dp ¼ 0.36 mm/T ¼ 300 K/Q ¼ 8.33 g CO2/min
c
dp ¼ 0.36 mm/P ¼ 15 MPa/Q ¼ 8.33 g CO2/min)
d
dp ¼ 0.36 mm/T ¼ 310 K/P ¼ 10 MPa
e
T ¼ 300 K/P ¼ 7 MPa/U ¼ 0.017 mm/s
f
T ¼ 313 K/P ¼ 10 MPa/U ¼ 1.234 mm/s
(range: 1.38 105 to 3.81 106) using model Diff-Slab/PM. Furthermore, values
of FM for a selected experiment of Gaspar et al. (2003) – dp ¼ 0.360 mm/T ¼ 300
K/P ¼ 8 MPa/Q ¼ 8.33 g CO2/min – differed by a factor of more than 10
depending on the model: 3.13 105 using model IBC-Diff (Gaspar et al. 2003)
and 2.64 105 using model Diff-PF (Araus et al. 2009).
Table 17.6 supports the claim that mathematical models and their ability to fit
the physical picture of the actual extraction process explain the differences among
estimated values of FM to a great extent in the case of four different substrates. For
basil, marjoram, and rosemary average values of FM were approximately 1.5 times
larger using model Diff-Sph/PM (Reverchon et al. 1993a) than model Diff-ADPF
(Goodarznia and Eikani 1998), and approximately 70 times larger using model
Diff-ADPF than model mS/SGl (Zizovic et al. 2005). For caraway, values of FM
were 20–110 times larger using model IBC/PFNA (Sovová et al. 1994a) than model
Table 17.6 Values of microstructural factor for high-pressure CO2 extraction of plant essential
oils in selected studies from Tables 17.1 and 17.2. All reported values for a single substrate and
extraction condition were based on single experiments from an original publication, and fitted by
individual authors to different models
Substrate Original Goodarznia and Zizovic et Sovová Contribution of
Study Eikani (1998) al. (2005) (2005) the authors
Basil (Reverchon 91,200 72,000 675 – –
et al. 1993a)
Marjoram (Reverchon 97,100 59,100 894 – –
et al. 1993a)
Rosemary (Reverchon 48,400 28,800 891 – –
et al. 1993a)
Caraway (Sovová et 2,180,000 20,000 – – 1,860 (Germain
al. (1994a), extraction et al. 2005)
at 313 K and 9 MPa)
Caraway (Sovová et al. 680,000 36,000 – – 3,620 (Germain
(1994a), extraction at et al. 2005)
313 K and 10 MPa)
Pennyroyal (Reis-Vasco 280 – – – 436 (Araus et al.
et al. (2000), 2009)
dp = 0.3 mm)
Pennyroyal (Reis-Vasco 168 – – 817 436 (Araus et al.
et al. (2000), 2009)
dp = 0.5 mm)
Pennyroyal (Reis-Vasco 120 – 1,278 – 436 (Araus et al.
et al. (2000), 2009)
dp = 0.7 mm)
Diff-ADPF (Goodarznia and Eikani 1998), and approximately 10 times larger using
model Diff-ADPF than model SC/PF (Germain et al. 2005). Finally, for pennyroyal
values of FM were 1.5–3.5 times larger using model Diff-PF (Araus et al. 2009) than
model LDF/ADPF (Reis-Vasco et al. 2000), and for selected experiments 6.8 times
larger using model IBC/PF/PCPR (Sovová 2005) or 10.7 times larger using model
mS/SGl (Zizovic et al. 2005) than using model LDF/ADPF (Reis-Vasco et al. 2000).
Table 17.7 reports the interval of FM-values estimated for all mass transfer
studies in Tables 17.1 and 17.2 that inform best-fit values of internal mass transfer
parameters, with the exception of studies in Tables 17.5 and 17.6. The results of the
single experiments were collated using the same material based on this chapter’s
hypothesis that the values of FM depend on the sample and its pretreatment, but do
not depend on the extraction temperature and pressure, CO2 superficial velocity,
and particle diameter (del Valle and de la Fuente 2006; Araus et al. 2009; Uquiche
et al. submitted). Explanations of the observed variations in the form of wide
intervals for the values of FM were advanced in the previous paragraphs in
explaining some differences in Tables 17.5 and 17.6.
According to (17.17), the validity of reported values of the microstructural
correction factor in Tables 17.5–17.7 depends partially on the values of the binary
diffusion coefficient (D12) estimated in this chapter using the equation of Catchpole
and King (1994). This equation was developed with a large database that did not
include typical components in plant essential oils. Table 17.8 displays estimated
Table 17.7 Summary of values of the microstructural factor for high-pressure CO2 extraction of
plant essential oils in studies from Tables 17.1 and 17.2 that are not included in Tables 17.4 or 17.5
Substrate Pretreatment Model namea Microstructural
factor (FM, –)
Black pepper (Ferreira and Meireles Milled Sovová (230–2.40) 107
2002)
Black pepper (Perakis et al. 2005) Milled Sovová (12.5–1.47) 103
Valerian II (Zizovic et al. 2007a) Milled mS/SCav (218–6.79) 105
Valerian I (Zizovic et al. 2007a) Milled mS/SCav (7.25–3.58) 106
Valerian III (Zizovic et al. 2007a) Milled mS/SCav (4.63–2.99) 106
Spearmint (Kim and Hong 2002) Milled DDD/PM (30.1–4.17) 104
Marigold (Zizovic et al. 2007a) Milled Sovová (68.2–4.06) 103
Marigold (Zizovic et al. 2007a) Milled Diff-Slab/PM (11.2–4.99) 104
Sage (Reverchon 1996) Milled LDF-Slab/ 5.54 104
PMMS
Sage (Araus et al. (2009) Milled Diff/PF (39.5–3.95) 103
Sage (Catchpole et al. 1996b) Chopped LDF/IMTC 2.52 103
Sage (Langa et al. 2009) Milled IBC/PF/PCPR 49.9–3.75
Rosemary (Coelho et al. 1997) Milled LDF/PF/CDIC (16.9–7.39) 102
Rosemary (Bensebia et al. 2009) Milled Sovová (4.34–1.26) 102
Hop (Pfaf-Šovljanski et al. 2005) Milled R-SO 3.20 104
Aniseed (Rodrigues et al. 2003) Milled Sovová (30.5–2.09) 103
Eucalyptus (Zizovic et al. 2007c) Milled mS/SCav 1.58 104
Chamomile (Povh et al. 2001) Milled Sovová (14.7–7.97) 103
Chamomile (Araus et al. 2009) Milled Diff/PF 1.08 104
Chamomile (Kotnik et al. 2007) Milled EMTC (6.65–3.53) 103
Thyme (Zekovic et al. 2001) Milled R-SO (13.9–1.2) 103
Orange peel (Mira et al. 1996) Milled Sovová (13.5–1.15) 103
Orange peel (Mira et al. (1999) Milled Sovová 4290–7.29
Orange peel (Zizovic et al. 2007c) Milled mS/SCav 1.68 102
Ginger (Martı́nez et al. 2003) Milled Sovová (11.1–2.21) 103
Alecrim pimenta (Sousa et al. 2002) Triturated Sovová 1.03 104
Ho-sho (Steffani et al. 2006) Milled SC/ADPF (83.1–4.44) 102
Parsley (Louli et al. 2004) Milled Sovová (7.53–2.23) 103
Parsley (Louli et al. 2004) Milled Sovová (7.5–2.54) 103
Lavender (Reverchon et al. 1995a) Milled R-SO 6.41 103
Lavender (Akgun et al. 2000) Manually crushed SC/PF (36.8–9.25) 102
Lavender (Araus et al. 2009)[7] Milled Diff/PF 2.03 102
Cinnamon of cunha (Sousa et al. Triturated Sovová (6.17–4.62) 103
2005)
Boldo (Uquiche et al. submitted) Rapidly Diff/PF 5.66 103
decompressed
Boldo (Uquiche et al. submitted) Conventionally Diff/PF 4.75 103
milled
Boldo (Uquiche et al. submitted) Low-temperature- Diff/PF 4.18 103
milled
Nutmeg (Spricigo et al. 2001) Milled SC/ADPF (56.5–3.39) 102
Nutmeg (Machmudah et al. 2006) Milled – –
Fennel (Reverchon et al. 1999) Milled LDF/PF (2.25–2.37) 102
Celery (Papamichail et al. 2000) Milled Sovová (15.1–2.04) 102
Celery (Zizovic et al. 2005) Milled mS/SDuct 1.28 103
(continued)

Substrate Pretreatment Model namea Microstructural
factor (FM, –)
Carqueja (Vargas et al. 2006) Milled Sovová (4.25–1.79) 102
Carqueja (Vargas et al. 2006) Milled R-SO (8.30–1.58) 102
Clove (Martı́nez et al. 2007) Milled Sovová 84.2–69.6
a
Model names are specified in Table 17.1
variations in the D12 of typical components in plant essential oil extracts in high-
pressure CO2 as a function of CO2 density and system temperature. Values of D12
decrease when decreasing the temperature (a 6.4% decrease from 333 to 313 K), or
increasing the density of the CO2 (a 60% decrease from 285 to 683 kg/m3), or
increasing the size of the solute. However, changing D12 to a considerable extent
demands extreme changes in extract composition; the reduction is limited to 25%
between the heaviest OST (the 17-C farnesyl acetate) and lightest MT (the 10-C
limonene), 40% between the wax (the 28-C n-octacosane) and limonene, and 56%
between the triglyceride (the 57-C triolein) and limonene. According to this analy-
sis, the discrepancies in D12 associated with the use of the equation of Catchpole
and King (1994) and the adoption of an erroneous pseudo-solute could result in
discrepancies that are substantially smaller than discrepancies in values of FM
reported in Tables 17.5–17.7.
Another factor that is partially responsible for errors in the estimation of FM,
according to (17.19), is the erroneous estimation of the partition K of the pseudo-
solute between the solid and the fluid. A discussion of errors in the estimation of K
can be found in Sect. 4.4.
There are some reports on the estimation of De for a porous solid as a function of
D12 and the inner porosity (ep) of the solid substrate (Goto et al. 1993; Ruetsch et al.
2003; Daghero et al. 2004; Perakis et al. 2005). According to the equation of Wakao
and Smith (1962), (17.22), FM can be estimated as follows:
1
FM ¼ (17.22)
e2p
Considering that in these studies, inner porosity values for the substrates ranged
from ep ¼ 0.487 for black pepper (Perakis et al. 2005) to ep ¼ 0.537 for peppermint
(Goto et al. 1993), the values of FM estimated using (17.22) ranged from 4.0 to
4.2. Thus, (17.22) produces smaller values of FM than those reported in
Tables 17.5–17.7. Ruetsch et al. (2003) and Daghero et al. (2004) did not measure
ep, but assumed a value ep ¼ 0.5. On the other hand, Perakis et al. (2005) estimated
the porosity from experimental values of the true density (rp) of ground particles of
black pepper, and the bulk density (rp) of a packed of the particles according to
(17.23):
rb
e¼1 (17.23)
rp
17
Table 17.8 Predicted binary diffusion coefficient D12 (m2/s 108) of selected components in plant essential oils (component 2) in supercritical CO2
(component 1) as a function of system temperature and CO2 density
Solute MW2 (Da) Vc2 (cm3/mol) r = 285.0 (kg/m3) r = 682.6 (kg/m3)
313 K 323 K 333 K 313 K 323 K 333 K
(8.1 MPa) (9.0 MPa) (9.9 MPa) (10.0 MPa) (12.8 MPa) (15.7 MPa)
Limonene 136.2 507.5 3.38 3.49 3.60 1.37 1.41 1.45
Linalool 154.3 565.5 3.22 3.32 3.42 1.30 1.34 1.38
Linalyl 196.3 682.5 2.95 3.04 3.13 1.19 1.23 1.26
acetate
b- 204.4 722.5 2.88 2.97 3.06 1.16 1.20 1.23
Caryophyllene
Farnesol 222.4 837.5 2.71 2.79 2.88 1.09 1.13 1.16
Farnesyl acetate 264.4 954.5 2.55 2.63 2.71 1.03 1.061 1.09
n-Octacosane 394.8 1,603.5 2.04 2.11 2.17 0.824 0.850 0.877

Triolein 885.5 3,233.5 1.49 1.54 1.59 0.603 0.622 0.642
435
The problem with using (17.21) to estimate the inner porosity of the substrate is
that e is the total porosity of the bed instead of ep, which is smaller. (17.24) relates
the inner porosity of the substrate (ep) with the total porosity (e) and the interparticle
porosity of the particles in the packed bed (eb):
e eb
ep ¼ (17.24)
1 eb
Considering that randomly packed spherical particles in a dense bed have a large
ratio, D/dp, for which the expected value of interparticle porosity is eb ¼ 0.36
(Carman 1937), it can be estimated using (17.24) that ep ¼ 0.198 for e ¼ 0.487,
and then using (17.22) that FM ¼ 25.4, which is still small but closer to the
experimental values reported in Tables 17.5–17.7.
This chapter’s analysis of the inner mass transfer within the solid substrate
up to this point has implicitly considered that the substrate is a homogeneous
material. Tables 17.5–17.7 include examples of materials being considered as
being composed of broken superficial cells and intact inner cells, and total mass
transfer in these heterogeneous materials considers separate extraction of the two
fractions according to different controlling mechanisms, respectively external
resistance to mass transfer, and internal (diffusive) resistance to mass transfer.
These models usually use the ratio of broken to intact cells or the fraction of free
solute (a) in the pretreated substrate as a fitting parameter; however, this ratio has
to agree with the microstructure of the biological substrate, particularly of the
specialized secretory structures where plant essential oils are encapsulated. For
a spherical particle obtained by milling of a parenquimatous tissue, a is given
by (17.25):

dc 3
a¼1 1 (17.25)
dp
where the thickness of a superficial layer of ruptured cells is closely related to the
diameter (dc) of the specialized secretory structure, and dp is the diameter of the
milled particle.
Figure 17.6 shows the effect of the ratio between cell diameter and particle
diameter (dc/dp) on the best-fit values of fractions of broken cells (a) in milled herbs
or spices containing superficial glands, including oregano bracts (Gaspar et al.
2003), pennyroyal leaves (Sovová 2005), rosemary leaves (Bensebia et al. 2009),
and sage leaves (Langa et al. 2009); secretary ducts, including caraway fruits
(Sovová et al. 1994a), celery seeds (Papamichail et al. 2000), chamomile flowers
(Povh et al. 2001), aniseed fruits (Rodrigues et al. 2003), parsley seeds (Louli et al.
2004), and marigold flowers (Campos et al. 2005); or secretory cavities, such as
orange peels (Mira et al. 1996, 1999). For calculations, the cell diameters (dc)
estimated from microphotographs in the literature were as follows: for oregano,
dc ¼ 78 mm (Svoboda and Svoboda 2000); for pennyroyal, dc ¼ 52.5 mm (Zizovic
et al. 2007c); for rosemary, dc ¼ 85 mm (Svoboda and Svoboda 2000); for sage,
Fig. 17.6 Effect of particle diameter on the best-fit values of the fraction of broken cells in milled
herbs or spices. Calculations were made from data of caraway (Sovová et al. 1994a), orange
peel (Mira et al. 1996), orange peel (Mira et al. 1999), celery (Papamichail et al. 2000),
chamomile (Povh et al. 2001), oregano (Gaspar et al. 2003), aniseed (Rodrigues et al. 2003),
marigold (Campos et al. 2005), pennyroyal (Sovová 2005), parsley (Louli et al. 2004),
rosemary (Bensebia et al. 2009), and sage (Langa et al. 2009)
dc ¼ 52 mm (Serrato-Valenti et al. 1997); for caraway, dc ¼ 123 mm (Svoboda and

Svoboda 2000); for celery, dc ¼ 35 mm (Stamenić et al. 2008); for chamomile,
dc ¼ 140 mm (Zizovic et al. 2007c); for aniseed, dc ¼ 108 mm (Zizovic et al.
2007c); for parsley, dc ¼ 50 mm (Podlaski et al. 2003); for marigold, dc ¼ 120 mm
(Zizovic et al. 2007c); and for orange peels, dc ¼ 500 mm (Svoboda and Svoboda
2000). The line in Fig. 17.6 corresponds to (17.25) which is only valid for size
reduction relying on impact or attrition mechanisms that fracture surface cells only.
Figure 17.6 shows that (17.25) only partially explains the best-fit values of broken
cells reported in the literature, possibly because of differences between the assumed
and actual particle shape, microstructure, and fracturing mechanism of the
specialized secretory structures in the plant materials.
Zizovic et al. (2005, 2007a, b, c) have contributed a series of studies on the
mathematical modeling of SCFE of plant essential oils at the micro-scale that
demonstrate the need to include a description of the encapsulating secretory
structures in plant tissue to improve the predictive capacity of the model. These
structures (glandular trichomes or glands, large or small secretory cavities, and
secretory ducts) (Sect. 17.1.1), are not equally affected by pre-treatment, nor by the
extraction process. For example, milling causes partial destruction of glandular
trichomes or glands in the surface of leaves, terminal shoots, and flowers of
Laminaceae herbs, and subsequent SCFE causes partial destruction of the
remaining glands by a mechanism involving diffusion of CO2 through the gland

wall, its dissolution into the essential oil, and swelling the CO2-saturated mixture to
the extent of rupturing the gland. Zizovic et al. (2005) obtained microscopical
evidence of the rupture of a fraction f ~ 0.35 of the glands during milling of
dried mentha leaves to a final particle diameter of 0.5–1 mm, as well as of the
rupture of about j ~ 0.39 of the glands after contacting unmilled leaves with high-
pressure CO2 at 313 K and 10 MPa for 1 h. Zizovic et al. (2005) hypothesized that
all glands cracked during the process as a result of swelling of the gland contents,
become disrupted upon saturation of the essential oil with the CO2 at a single time
that is determined by diffusion of CO2 through the gland wall, but later Stamenić
et al. (2008) proposed a cracking time distribution based on microscopical evidence
gathered by exposing wild thyme leaves to high-pressure CO2 at 313 K and 10 MPa
different times between 20 min and 100 min. These models were applied to basil,
rosemary, marjoram, and pennyroyal by Zizovic et al. (2005) and to wild thyme by
Stamenić et al. (2008).
Unlike superficial glands that can rupture upon swelling of the gland contents by
dissolution of CO2 into the essential oil, inner secretory cavities in plant tissue are
ruptured only as a result of a size-reduction pretreatment, with the probability of
rupture being dependent on the relative sizes of secretory cavities and milled
particles. Since the physical picture for the extraction of essential oil from secretory
cavities at the microscale is similar to that of intact and broken cells proposed by
Sovová (1994) (Sect. 17.2.3), the mathematical model in this particular case
considered simultaneous extraction of essential oils from disrupted cavities con-
trolled by an external mass transfer coefficient, and of essential oils from nondis-
rupted cavities controlled by a best-fit internal mass transfer coefficient. Zizovic
et al. (2007a) estimated the fraction of disrupted cavities in milled valerian roots as
the ratio between the cavity diameter (determined by microscopy) and the particle
diameter (determined by sieving). Zizovic et al. (2007a) estimated the mean
diffusion pathway in nondisrupted cells as half the difference between the particle
diameter and cavity diameter, but they did not explain the basis for this assumption.
This model was applied to valerian roots by Zizovic et al. (2007a), and to clove
buds, ginger rhizomes, and eucalyptus leaves by Zizovic et al. (2007c). Conse-
quently, Zizovic et al. (2007c) developed a separate model for SCFE of plant
materials with large secretory cavities, such as citrus peels, where most cavities
are cracked.
In the case of plant material with secretory ducts, the size-reduction pretreatment
opens each duct on two opposite sides of the particle. The essential oil in the duct
becomes saturated with CO2 during extraction and the swelling of the mixture
results in a superficial layer of oil embedding the whole particle. SCFE can be
pictured as a two-stage process. In the first stage, when the mass transfer area is the
external surface of the particle, the extraction of essential oil is defined by convec-
tion of the superficial oil to the CO2 (mass transfer coefficient from a literature
correlation). In the second stage, when the mass transfer area diminishes to that
fraction of the external surface of the particle covered with openings of secretory
ducts, the essential oil moves from a receding boundary in the pore to the pore
opening by unimpeded diffusion (defined by the binary diffusion coefficient of the

oil in CO2), and then by convection (the same mass transfer coefficient from the first
stage) from the pore opening to the bulk of the high-pressure CO2 phase; the
diffusion path for the essential oil in this second stage is half the particle diameter.
Zizovic et al. (2007b) proposed a fully predictive mathematical model for SCFE
plant material with secretory ducts, which used the diameter of the secretory ducts
and their superficial density (number of ducts per unit surface area) as microstruc-
tural parameters. The model was applied for fruits of celery (Stamenić et al. 2008),
and fruits of fennel, and flowers of chamomile and marigold (Zizovic et al. 2007b).
17.4 Phase Equilibrium Effects in Essential Oil Extraction,

Fractionation, and Recovery
In this section, the effect of phase equilibrium on the extraction, fractionation, and
recovery of plant essential oils using high-pressure CO2 is discussed. Firstly, the
variations in the solubility of essential oil components in high-pressure CO2 are
discussed as a function of system temperature and pressure (Sect. 17.4.1). Sec-
ondly, there is discussion of the selective removal of high-solubility MTs from the
remaining OMTs so as to produce citrus essential oils with an improved function-
ality (more intense aroma, less off-flavors, increased solubility in water) based on
high-pressure phase equilibrium data for ternary CO2 + MT + OMT and more
complex systems (Sect. 17.4.2). Next, the thermodynamic solubility in CO2 of
complex essential oil mixtures is compared with the “operational” solubility of
the CO2 extract of the corresponding herb or spice under comparable temperature
and pressure (Sect. 17.4.3). Finally, the effect of sorption phenomena on the
aforementioned “operational” solubility (Sect. 17.4.4) is discussed.
17.4.1 Solubility in CO2 of Essential Oil Components in Model

(Binary) Systems
There are many experimental studies about the solubility of selected essential oil
components in CO2 as a function of system temperature and pressure, as shown in
Table 17.9. The data include four different solute families and CO2 conditions that
span from subcooled liquid, to superheated vapor, and to SCF (equilibrium temp-
eratures of 296–373 K, and equilibrium pressures of 0.81–30.9 MPa). There is
abundant information in the literature about the solubility of limonene in high-
pressure CO2, but virtually none about other 10-carbon MT compounds, the only
exceptions being p-cymene and a-pinene (Table 17.9). On the other hand, there is
extensive information on the solubility of OMT compounds in high-pressure CO2,
including anisaldehyde, camphor, carvacrol, carvone, 1,8-cineole (eucalyptol),
Table 17.9 Summary of high-pressure equilibrium CO2 + plant essential oil component binary
systems. The essential oil components were classified as monoterpene hydrocarbons, oxygenated
monoterpenes, sesquiterpene hydrocarbons and oxygenated sesquiterpenes, and waxy compounds
Solute Data Temperature(s) Pressure(s) Solubility range
points or temperature or pressure (mg solute/g
range (K) range (MPa) CO2)
Monoterpenes
p-Cymene (Wagner and Pavlicek 16 313–323 3.855–9.829 6.73–41.7
1993)
Limonene (di Giacomo et al. 1989) 26 308, 315, 323 3.0–10.0 0.681–142
Limonene (Matos et al. 1989) 14 318, 323 8.6–9.8 12.4–106
Limonene (Marteau et al. 1995) 29 310, 320, 323 6.96–9.85 5.58–73.2
Limonene (Iwai et al. 1996) 15 313, 323, 333 3.94–10.26 3.41–24.3
Limonene (Akgun et al. 1999) 22 313, 323, 333 7.04–11–20 1.24–54.8
Limonene (Chang and Chen 1999) 28 314, 323, 333 0.83–11.00 1.55–47.2
Limonene (Vieira de Melo et al. 16 323, 333, 343 7.00–10.55 4.96–88.2
1999)
Limonene (Kim and Hong 1999) 11 312, 322 6.2–9.3 3.10–79.4
Limonene (Berna et al. 2000) 8 318 6.9–10.5 5.27–1,202
Limonene (Gamse and Marr 2000) 42 304, 314, 324 3.1–8.4 2.79–65.4
Limonene (Benvenuti and Gironi 8 315 3.05–8.50 5.89–18.4
2001)
Limonene (Leeke et al. 2001) 15 318, 323 7.32–10.05 2.79–81.3
Limonene (Francisco and Sivik 8 313, 333 8–25 12.4–89.2
2002)
Limonene (Fonseca et al. 2003) 4 323 8.3–9.5 130–244
a-Pinene (Pavlicek and Richter 72 313, 323, 328 3.25–9.75 3.60–54.3
1993)
a-Pinene (Richter and Sovová 1993) 24 296–335 6.14, 7.65 4.15–102
a-Pinene (Akgun et al. 1999) 19 313, 323, 333 7.15–10.93 5.27–53.5
a-Pinene (Francisco and Sivik 2002) 8 313, 333 8–25 9.32–82.7
Oxygenated monoterpenes
p-Anisaldehyde (Mukhopadhyay and 13 323, 373 5.5–13.7 0.619–43.9
De 1995)
Camphor (Akgun et al. 1999) 39 313, 323, 333 7.51–12.61 2.98–139
Camphor (Carvalho et al. 2006) 15 314, 324, 334, 8.65–15.71 32.5–275
354
Carvacrol (Leeke et al. 2001) 49 313, 323 7.70–30.85 9.59–740
Carvone (Kim and Hong 1999) 13 312, 322 6.2–9.6 0.307–59.0
Carvone (Gamse and Marr 2000) 31 304, 314, 324 2.0–10.0 0.683–50.9
1,8-Cineole (Matos et al. 1989) 15 318, 323 7.75–9.80 14.1–82.5
1,8-Cineole (Francisco and Sivik 8 313, 333 8.0–25.0 10.5–86.2
2002)
Citral (di Giacomo et al. 1989) 29 308, 315, 323 3.0–11.0 0.104–78.9
Citral (Marteau et al. 1995) 26 310, 320, 330 7.94–11.93 2.42–4.85
Citral (Benvenuti and Gironi 2001) 10 315 4.70–9.58 0.692–36.0
Citral (Fonseca et al. 2003) 4 323 8.7–10.3 32.5–143
b-Citronellol (Tufeu et al. 1993) 15 309, 316, 321 8.46–15.70 25.8–160
Eugenol (Cheng et al. 2000) 21 308, 318, 328 1.480–12.512 4.48–61.8
Fenchone (Akgun et al. 1999) 21 313, 323, 333 7.04–11.50 1.04–1.04
Geraniol (Tufeu et al. 1993) 9 309, 316, 321 10.82–16.27 37.9–120
Linalool (Chang and Chen 1999) 29 313, 323, 333 0.81–11.06 1.05–35.4
Linalool (Vieira de Melo et al. 1999) 3 323 7.49–8.37 1.75–4.56
Linalool (Berna et al. 2000) 13 318, 328 7.1–11.1 2.39–1,381
(continued)

Solute Data Temperature(s) Pressure(s) Solubility range
points or temperature or pressure (mg solute/g
range (K) range (MPa) CO2)
Linalool (Fonseca et al. 2003) 3 323 8.6–9.4 132–196
Linalool (Iwai et al. 1994) 14 313, 323, 333 4.00, 10.96 1.05–27.9
Linalool (Raeissi and Peters 2005) 12 319–348 9.355–13.705 71.5
Menthol (Carvalho et al. 2006) 12 323, 343 6.5–13.5 0.256–218
Menthol (Sovová and Jež 1994) 40 308, 313, 318, 6.52–11.56 0.782–47.1
323, 328
Menthol (Sovová et al. 2007) 47 303, 313, 323, 6.64–14.37 1.24–147
333
Thymol (Carvalho et al. 2006) 12 323, 343 6.5–13.5 19.6–497
Verbenol (Richter and Sovová 1993) 73 313, 323, 328 5.09–13.78 0.208–194
Sesquiterpenes and oxygenated sesquiterpenes
b-Caryophyllene (Michielin et al. 86 303, 313, 323, 4.5–19.2 18.7–522
2009) 333, 343
Artemisinin (Xing et al. 2003) 36 310, 318, 328, 10.0–27.0 0.628–17.1
338
Farnesol (Núñez et al. 2010) 58 313, 323, 333 9.7–26 0.657–9.67
a-Humulene (Michielin et al. 2009) 77 303, 313, 323, 4.7–19.2 18.7–521
333, 343
Patchoulol (Hybertson 2007) 8 313, 323 10.0–25.0 2.17–48.0
Typical wax component
Octacosane (Reverchon et al. 1993b) 28 308, 313, 318 8.0–27.5 0.110–3.50
Octacosane (Stassi and Schiraldi 3 308–318 25 0.610–3.36
1994)
Octacosane (Chandler et al. 1996) 14 308, 318 10.0–24.0 0.170–4.22
citral, b-citronellol, eugenol, fenchone, geraniol, linalool, menthol, thymol, and

verbenol (Table 17.9). There is less solubility information about the 15-carbon
terpene compounds, which is limited to artemisinin (OST), b-caryophyllene (ST),
farnesol (OST), a-humulene (ST), and patchoulol (OST). Finally, although there is
extensive information in the literature on the solubility of waxy compounds in high-
pressure CO2, Table 17.9 is limited to experimental works with octacosane. Litera-
ture on the solubility of waxy compounds includes n-alkanes having 9–36 carbon
atoms under typical CO2 extraction conditions, as well as long-chain n-alkanes
under typical gas-processing conditions (Stassi and Schiraldi 1994; Chandler et al.
1996; Reverchon et al. 1993b). Stahl et al. (1988) conducted a systematic study on
the solubility of essential oil components in high-pressure CO2 including data for
limonene, anethole (OMT), carvone, eugenol, b-caryophyllene, and valeranone
(OST). This chapter does not include the results of Stahl et al. (1988) in Table 17.9
since they reported trend lines instead of actual experimental data, so their results
are of less value for quantitative comparison purposes.
Solubility values of selected essential oil components in high-pressure CO2 vary
widely, as exemplified in Fig. 17.7 for the solubility of limonene in high-pressure
Fig. 17.7 Solubility of limonene in supercritical CO2 as a function of system pressure (approxi-
mately between 1 MPa and 10 MPa) as reported by di Giacomo et al. (1989) at 313 K, Matos
et al. (1989) at 318 K, Marteau et al. (1995) at 310 K, Iwai et al. (1996) at 313 K, Akgun
et al. (1999) at 313 K, Chang and Chen (1999) at 314 K, Kim and Hong (1999) at 312 K,
Berna et al. (2000) at 318 K, Gamse and Marr (2000) at 314 K, Benvenuti and Gironi (2001)
at 315 K, Leeke et al. (2001) at 318 K, and Francisco and Sivik (2002) at 313 K.
Lines represent the solubility isotherms at 311 and 318 K predicted using the
Peng-Robinson-Mathias-Copeman equation of state with the modified Huron-Vidal (MHV1) –
UNIFAC mixing rules and model parameters in database of PE 2000 (Pfohl et al. 2000), a
shareware software for modeling high-pressure phase equilibria
CO2 at approximately 313 K (310–318 K) and as a function of system pressure, due

to the inherent limitations of methodologies applied to assess phase equilibrium
(Raal and M€uhlbauer 1998). Limonene is the most abundant MT in plant essential
oils, representing >90% (w/w) of some of them, because it acts typically as a carrier
of other compounds (mainly OMTs) that impact more definitely on plant aroma.
The experimental methods can be broadly classified as analytic (if one or both
phases in equilibrium are sampled and analyzed to determine composition) or
synthetic (if the global composition of the system is predetermined, and the condi-
tions of the system are changed so as to reach a phase boundary). These broad
classes, in turn, can be implemented in static or dynamic equilibrium cells, depend-
ing on agitation. In dynamic cells, the time to reach equilibrium is shortened by
agitation of the cell contents, or recirculation of one or the two phases in the cell
(the so-called multi-pass dynamic systems, where sampling is done in recirculation
loops). Measurements using all methods are susceptible to error if true equilibrium
conditions are not reached, and no single method is more questionable in this regard
than the one-pass dynamic method (di Giacomo et al. 1989; Kim and Hong 1999;
Gamse and Marr 2000; Benvenuti and Gironi 2001; Fonseca et al. 2003). Observa-
tion of the cell contents through a window-cell helps to ascertain the quality of
stirring and the number and nature of phases in the system. In the case of analytical
methods, where system temperature and pressure are kept constant, sampling and
analysis of one phase as a function of equilibration time can help to ascertain
whether equilibrium conditions have been reached. Another problem with analytic
methods occurs when an incomplete picture of the equilibrium is achieved when
only the CO2-rich phase is sampled, and when CO2 transfers to the solute-rich phase
(Matos et al. 1989; Berna et al. 2000). A final problem when using an analytic
system is the disturbance of system conditions and the associated shift in equilib-
rium during sampling, which causes a drop in pressure (Iwai et al. 1996; Vieira de
Melo et al. 1999). Since these disturbances are unavoidable, users of analytical
methodologies try to minimize the negative effects of sampling by using large cells
(Akgun et al. 1999; Leeke et al. 2001), or by reducing the size of the sampled
aliquot by coupling the cell with a high-sensitivity instrument such as an infrared
absorption device (Marteau et al. 1995), a densitometer (Chang and Chen 1999), or
a chromatograph (Francisco and Sivik 2002).
The sampling problem inherent in analytic methods can be avoided by using
synthetic methods. Synthetic systems use a variable volume cell to change system
pressure while keeping the temperature and global composition constant. Raising the
pressure (reducing the inner volume of the cell) to make the CO2-rich phase collapse
to a single bubble (as is the case when the composition of the liquid phase equals the
global composition of the system) determines the bubble pressure at the test temper-
ature. Alternatively, decreasing the pressure (increasing the inner volume of the cell)
to cause the solute-rich phase collapse to a single drop (as is the case when the
composition of the vapor phase equals the global composition of the system) deter-
mines the dew point at the test temperature. A problem of the synthetic method is the
inherent difficulty in reaching and determining both bubble and dew point conditions
for binary and more complex systems (Marteau et al. 1995).
The solubility of limonene in supercritical CO2 at 313 K increases steadily with
system pressure at low pressures (P < 8 MPa) and then increases sharply at high
pressures (P > 8 MPa), with the transition between steady and sharp increase
occurring close to the critical point of the mixture (Fig. 17.7). The reported critical
point of CO2 + limonene mixtures at 313 K is 8.3 MPa (Matos et al. 1989) or
8.5 MPa (Tufeu et al. 1993), and under these conditions the composition of the
liquid phase coincides with that of the SCF phase (35.1 g limonene/kg CO2). The
binary CO2 + limonene system exhibits a temperature crossover at ~8 MPa (Akgun
et al. 1999; Berna et al. 2000). Thus, as the temperature increases isobarically,
solubility of limonene in CO2 decreases at P < 8 MPa due to reduction in the
density and solvent power of CO2, whereas solubility increases at P > 8 MPa due
to the rise in vapor pressure and volatility of limonene. There is a general consis-
tency between experimental solubility data for <8.5 MPa and for <80 g limonene/
kg CO2, but the data of Chang and Chen (1999) and Gamse and Marr (2000) are
~3 times and ~10 times, respectively, above the general trend. The results of
Francisco and Sivik (2002) are questionable because they report limited solubility
of limonene in CO2 under conditions where the two components are mutually
miscible (at 313 K and well above the critical pressure of the CO2 + limonene
mixture). The data are more diverse at pressures >8.5 MPa due to experimental
difficulties in measuring solubilities under conditions near the critical point of a
mixture.
Figure 17.7 also includes the solubility isotherms at 310 and 318 K of limonene
in high-pressure CO2 predicted using the computer program PE 2000 (Pfohl et al.
2000). PE 2000 models liquid–vapor equilibrium by searching the phase transition
corresponding to the bubble-point curve of the CO2 + limonene system. Phase
equilibrium was modeled using the modification of Mathias-Copeman of the Peng-
Robinson (PR) equation of state (EoS), or the so-called PR-Mathias-Copeman EoS
(Poling et al. 2000), and the first modification (or Modification 1, M1) of the Huron-
Vidal (HV) mixing rules with the activity coefficients estimated using UNIFAC, or
the so-called MHV1-UNIFAC mixing rules (Poling et al. 2000). The database of PE
2000 included all model parameters for the CO2 + limonene system. Figure 17.7
shows a reasonable agreement between the predictions of the model and the
experimental measurements, including the temperature cross-over at ~8 MPa.
This finding is important because the selected model is of a predictive nature, in
that the model parameters for the binary system are not estimated using phase
equilibrium data; this chapter will expand on the implications of this feature when
comparing the solubilities in high-pressure CO2 of other compounds in Table 17.9.
Table 17.10 compares the solubilities in high-pressure CO2 of selected solutes
from each component family in Table 17.9 to those of limonene, under selected
system conditions (313 K and 8 or 9–10 MPa). The density of the CO2 under the
selected system conditions bracket the interval proposed by Reverchon (1997) for
SCFE of plant essential oils (250–500 kg/m3). At the lower end of recommended
CO2 density, ~260 kg/m3 (at 313 K and 8 MPa), the solubility of limonene in high-
pressure CO2 is approximately 6 g/kg, which is about 50% lower than the solubility
of a-pinene reported by Akgun et al. (1999), probably because of the increased
volatility (larger vapor pressure) of a-pinene as compared to limonene at 313 K
(Table 17.10). The oxygen-bearing functional groups in OMTs increase their
molecular weight and polarity, and decrease their vapor pressure as compared to
MTs, which causes the solubility of OMTs in high-pressure CO2 at 313 K and
8 MPa to be about 50% of that of limonene. At the upper end of recommended CO2
density, ~410 kg/m3 (at 313 K and 9 MPa), the binary systems of MTs and CO2 are
above their critical points, thus MTs are fully miscible with CO2. Tufeu et al. (1993)
reported the critical points at 314 K of the binary CO2 + citral (8.76 MPa, 36.4 g
citral/kg CO2) and CO2 + linalool (8.66 MPa, 11.5 g linalool/kg CO2) systems, and
these critical point values bring into question the limited solubility of citral in high-
pressure CO2 at 313 K and 9 MPa as reported by Benvenuti and Gironi (2001)
(Table 17.10). The solubilities in high-pressure CO2 at 313 K and ~9 MPa of the
two OMTs in Table 17.10 (camphor and menthol) are approximately 1 order of
magnitude larger than the corresponding solubilities of the ST (b-caryophyllene)
and OSTs (artemisinin, farnesol, patchoulol), also included in Table 17.10, which
are in turn approximately 2 orders of magnitude above the corresponding solubility
17
Table 17.10 Solubilities of selected plant essential oil components in high-pressure CO2 at 313 K and relatively low (approximately 8 MPa) or relatively high
pressure (8.6–10 MPa). Essential oil component included representative monoterpene hydrocarbons (limonene and a-pinene) and oxygenated monoterpenes
(citral, linalool, and menthol), a sesquiterpene hydrocarbon (b-caryophyllene), an oxygenated sesquiterpene (farnesol), and a hydrocarbon (octacosane)
Compound Molecular Vapor pressure (Pv, Pa) Low-pressure High-pressure
weight Pressure (P, CO2 density Solubility Pressure CO2 Solubility
(MW, Da) MPa) (r, kg/m3) (Csat, mg/g (P, MPa) density (Csat, mg/g
CO2) (r, kg/m3) CO2)
Limonene (Benvenuti and 136.2 515 (Espinosa-Dı́az et al. 8.13 295.6 7.04 9.0 408.4 CMc
a
Gironi 2001) 1999)
a-Pinene (Richter and Sovová 136.2 1,440 (Richter and Sovová 7.74 252.7 60.1 9.0 408.4 CM
1993) 1993)
a-Pinene (Akgun et al. 1999) 136.2 1,440 (Richter and Sovová 7.74 252.7 60.1 – – –
1993)
Camphor (Akgun et al. 1999) 152.2 133 (Espinosa-Dı́az et al. 7.97 276.1 3.60 8.8 436.1 98.9
1999)b
Citral (Benvenuti and Gironi 152.2 30 (Stull 1947) 8.00 261.3 1.30 9.0 408.4 7.66
2001) a
Linalool (Iwai et al. 1994) 154.2 95 (Espinosa-Dı́az et al. 7.99 278.3 5.27 – – –
1999)
Menthol (Sovová and Jež 156.3 <133 (Stull 1947)b 7.86 264.3 3.31 9.02 497.9 50.8
1994)
b-Caryophyllene (Michielin 204.4 0.018 (The Good Scents – – – 8.6 375.9 18.65
et al. 2009) Company 2009a)
Farnesol (Núñez et al. 2010) 222.4 0.192 (The Good Scents – – – 9.73 609.3 4.29
Company 2009b)
Octacosane (Chandler et al. 394.3 ~7 10–6 (Chandler et al. – – – 10.0 631.7 0.170
1996) (1996)
a
At 315 K
b
Sublimation pressure
c
Completely miscible
445
of a typical wax (n-octacosane), 0.015 g/kg CO2, which is as expected because of

the differences in molecular weight (increasing) between OMT compounds, ST/
OST compounds, and waxes.
Figure 17.8 expands the results in Table 17.10 to show solubility isotherms of
selected compounds in a wider pressure range, together with the estimates of the
predictive model in this chapter. In order to use the PR-Mathias-Copeman EoS in
PE 2000, an adjustment was made of the so-called model parameters c1, c2, and c3,
to vapor pressure data for a-pinene (Daubert and Danner 1989), citral (Hall 2001),
linalool (Espinosa-Dı́az et al. 1999), b-caryophyllene (Helmig et al. 2003; The
Good Scents Company 2009a), farnesol (Helmig et al. 2003; The Good Scents
Company 2009b), and n-octacosane (Daubert and Danner 1989; Chandler et al.
1996). Other parameters required were the normal boiling point, critical tempera-
ture, critical pressure, and acentric factor for the pure compounds. In those cases
where no reliable values for these properties were available, values were estimated
using group contribution methods. The normal boiling point of n-octacosane was
Fig. 17.8 Solubilities of selected plant essential oil components in supercritical CO2 as a function
of system pressure (approximately between 1 MPa and 12 MPa), including limonene at
315.7 K; a-pinene at 313 K, data of Richter and Sovová (1993) and Akgun et al.
(1999); citral at 315 K, data of Benvenuti and Gironi (2001); linalool at 313 K, data of
Iwai et al. (1994) or Chang and Chen (1999); caryophyllene at 313 K, data of
Michielin et al. (2009); farnesol at 313 K, data of Núñez et al. (2010); and octacosane
at 313 K, data of Chandler et al. (1996). Lines represent the solubility isotherms at corresponding
temperatures predicted by the Peng-Robinson-Mathias-Copeman equation of state with the mod-
ified Huron-Vidal (MHV1) – UNIFAC mixing rules using PE 2000 (Pfohl et al. 2000). The
solubility isotherm for limonene in supercritical CO2 at 315.7 K predicted by PE 2000 is
included as a reference
estimated using Joback’s method (Poling et al. 2000). On the other hand, the critical
temperature and critical pressure of citral, linalool, b-caryophyllene, farnesol, and
n-octacosane were also estimated using Joback’s method (Poling et al. 2000),
whereas the acentric factor of these same solutes was estimated using the method
of Lee-Kesler (Poling et al. 2000). PE 2000 includes all of these group-contribution
algorithms for the estimation of physical properties. Figure 17.8 shows that the
predictive model implemented in PE 2000 gives values of solubility for typical
essential components in high-pressure CO2 under typical extraction conditions for
herbs and spices that are only qualitatively correct; thus, it can be only moderately
appropriate for the purpose of discussing equilibrium effects on mass transfer
kinetics as attempted in this chapter. An advantage of the method is that it is fully
predictive if a group contribution method is applied to estimate the effect of
temperature on the vapor or sublimation pressure of the solutes. Obviously, the
predictive capabilities of the model improve when including experimental instead
of estimated values of relevant physical properties, as suggested by the better fit of
the model to experimental solubility data for limonene (Fig. 17.7) than higher
molecular solutes such as STs, OSTs, and waxes (Fig. 17.8 compares the experi-
mental data and predicted solubility isotherms for b-caryophyllene, farnesol, and
n-octacosane).
Differences in solubility between solutes using data for binary CO2 + solute
systems suggest the possibility of selectively recovering a single solute or mixture
of solutes in a complex essential oil sample. A general application of this type of
fractionation process is the deterpenation of essential oils, which consists of the
selective elimination of oxidation-prone and water-immiscible MTs that mask the
characteristic aroma of OMTs in an herb or spice, and cause a haze in aqueous
essential oil solutions (Stahl et al. 1988; Temelli et al. 1990; Reverchon 1997).
Mukhopadhyay and De (1995) proposed the selective recovery of menthol from
peppermint oil based on the solubility isotherms at 323 and 343 K of the binary
CO2 + menthol (component 2) and CO2 + thymol (component 3) systems at pres-
sures ranging from 6.5 to 13.5 MPa. They computed the values of the separation
factor between menthol and thymol, R23 (17.26), as a function of system tempera-
ture and pressure,
y2
R23 ¼ ; (17.26)
y3
and found that 3.37 R23 6.12, which suggests that it is possible to enrich
menthol in the high-pressure CO2 phase. Mukhopadhyay and De (1995) neglected
the molecular interactions between menthol and thymol, which may affect the
separation factor, as exemplified next for the separation of limonene and linalool
using high-pressure CO2. It is also important to mention that using the separation
factor to draw conclusions about the selective recovery of a component in a binary
or multicomponent mixture has limitations in that the analysis should be made by
comparing of concentration ratios between the two components in two phases
(liquid, SCF) in equilibrium; if these concentration ratios coincide in the two
phases, then it is not possible to selectively recover one of the components using
high-pressure CO2 as the separating agent. Thus, the selectivity of the separation
between menthol and thymol, a23 (17.27), instead of the separation factor R23
should be computed, where:
y2 =y3
a23 ¼ (17.27)
x2 =x3
Chafer et al. (2001) used composition information from the vapor phase for the
ternary CO2 (1) + limonene (2) + linalool (3) system to estimate the separation
factor R23 so as to evaluate the possibility of recovering a linalool-enriched fraction.
They concluded that the deterpenation of linalool was possible based on R23 values
of ~2 for mixtures of limonene and linalool containing about 90–95% (mol/mol) of
limonene, but also observed that the values of R23 were approximately four times
higher when estimated on the basis of binary equilibrium data, which reveals a loss
of information when solute–solute interactions in the more complex systems are not
accounted for. Because of the requirement of ternary data to assess the fractionating
capabilities of high-pressure CO2, this subject is discussed next.
17.4.2 Essential Oil Fractionation in Model (Ternary) Systems

and Complex Mixtures
Two CO2-containing tertiary systems of practical importance that have been exten-
sively analyzed in the literature are CO2 + limonene + citral and CO2 + limo-
nene + linalool, because limonene and citral (a mixture of two isomers, geranial
and neral) are the key MT and OMT, respectively, in lemon essential oil (Gironi and
Maschietti 2008), and linalool replaces citral as the key OMT in orange essential oil
(Budich and Brunner 1999). High-pressure CO2 fractionation allows deterpenation
of citrus oils so as to improve their shelf life, solubility in water, and aroma
(Stahl et al. 1988; Temelli et al. 1990; Reverchon 1997), and the designing of
these two deterpenation processes demands phase equilibria data for the aforemen-
tioned model systems (Budich et al. 1999; Gironi and Maschietti 2008).
The phase equilibria of the ternary CO2 (1) + limonene (2) + citral (3) system
was studied by Benvenuti and Gironi (2001) at 315 K and 8.4 or 9.0 MPa, and by
Fonseca et al. (2003) at 323 K and 9.5, 9.7, or 10.3 MPa. Fonseca et al. (2003)
reported that the selectivity a23 for the separation between limonene and citral
varied between 1.72 and 2.00, which indicates that the vapor (or CO2-rich) phase is
enriched in limonene, whereas citral remains in the liquid (or essential oil-rich)
phase. This agrees with the vapor pressure ratio between limonene and citral, P2Sat/
P3Sat ¼ 17 (Benvenuti and Gironi 2001), which defines separation under low-
solubility (relatively small pressure) conditions. The selectivity a23 did not depend
on system pressure (9.5–10.3 MPa), nor the composition of the essential oil model
mixture (49–73% w/w limonene) (Fonseca et al. 2003). On the other hand, values of
a23 estimated by Benvenuti and Gironi (2001) at 315 K ranged from 1 to 62

depending on the limonene content in the CO2-rich phase and the system pressure
(limonene content in a CO2-free basis ranged 36–90% mol/mol at 8.4 MPa, and
26–74% mol/mol at 9 MPa).
The vapor–liquid equilibria of the CO2 (1) + limonene (2) + linalool (3) ternary
system has been studied by Morotomi et al. (1999) at 313 K and 6.9 MPa, 333 K and
6.9 MPa, or 333 K and 10.0 MPa, by Vieira de Melo et al. (1999) at 323 K and 7.54,
8.08, 8.76, or 8.90 MPa, and by Chafer et al. (2001) at 318 or 328 K, pressures
between 7 and 11 MPa, and mixtures having 40% or 60% (w/w) limonene in a CO2-
free basis. The selectivity for the separation between limonene and linalool is
a23 > 1.2 (Morotomi et al. 1999), which suggests the possibility of enriching
limonene in the vapor phase, thus making the isolation of linalool in the liquid
phase feasible, particularly in essential oil mixtures enriched in oxygenated com-
pounds. Vieira de Melo et al. (1999) reported that a23 decreases from 3.75 to 2.14 as
the pressure increases from 7.54 to 8.90 MPa, probably as a result of the increase in
the solvent power of high-pressure CO2 for OMTs. The apparent selectivity at
323 K and 8 MPa based on binary data (a23 ¼ 4.6) was larger than the true
selectivity based on ternary data (a23 ¼ 3.6), thus confirming the need for ternary
equilibrium data for the design of deterpenation process (Vieira de Melo et al.
1999). Chafer et al. (2001) reported the composition of the vapor phase only; thus, it
was not possible to estimate values of selectivity for the separation of limonene and
linalool based on their data.
The use of model systems with a limited number of components is appropriate
only as a rough estimate of the behavior of more complex natural mixtures.
Because of that, Budich and Brunner (1999) and Budich et al. (1999) recommended
the estimation of separation factors and other design parameters for the high-
pressure CO2 deterpenation process by using actual essential oil mixtures. Cold-
pressed orange oil is constituted by about 200 components, mainly MTs (~95% w/w)
and OMTs. Temelli et al. (1990) measured the solubility in CO2 of this oil at 313,
323, 333, and 343 K and 8.3, 9.7, 11.0, and 12.4 MPa using a one-pass dynamic
method. Experimental results confirmed that CO2 preferably solubilizes MTs so
that the selectivity a23 for the separation between limonene (2) and linalool (3), the
key components in the MT and OMT fractions, respectively, ranged from 1.10 to
2.83 within the experimental region. Under isothermal conditions, a23 reached a
maximum at 9.7 MPa and then decreased to a minimum at 12.4 MPa. Temelli et al.
(1990) also reported an unusual increment in solubility at 313 K and 12.4 MPa,
which they attributed to the formation of a liquid phase, a common occurrence in
systems of high-pressure CO2 and complex liquid mixtures.
Budich and Brunner (1999) studied the equilibrium of CO2 + orange peel oil at
323, 333, or 343 K and 8–13 MPa. For data analysis they assumed orange peel oil
as a binary mixture of 98.25% (w/w) terpenes (T, 96.7% w/w limonene) and
1.75% (w/w) oxygenated aroma compounds (A, 28.8% w/w linalool). The solu-
bility of the essential oil in high-pressure CO2 was similar to that of pure
limonene, probably because of the elevated content of limonene in orange peel
oil. The selectivity aTA ranged from 1.3 to 3.2, which suggests the possibility of
removing the MT fraction using high-pressure CO2. Under isothermal conditions,

aTA decreases as the pressure increases. For low-solubility values (<10 g extract/
kg CO2) aTA decreases with temperature under isobaric conditions because of a
more limited increase in the vapor pressure of the OMTs than MTs; for high-
solubility values (>50 g extract/kg CO2) aTA increases with temperature due to
the reduction in the density and solvent power of the CO2; in the intermediate
range (10–50 g extract/kg CO2) there is not a definite trend for the variations in
aTA with temperature.
Figure 17.9 compares the values of selectivity measured by Budich et al. (1999)
for the multicomponent CO2 + orange peel oil system at 323 and 333 K with those
reported for the CO2 + limonene + citral or CO2 + limonene + linalool model
systems. For all temperature-solubility pairs, a23 < aTA, e.g., at 323 K the values
of a23 reported by de Vieira de Melo et al. (1999) are 25% of the values of aTA,
whereas the values of a23 reported by de Fonseca et al. (2003) are ~40% of the
values of aTA, and at 333 K Morotomi et al. (1999) report values of a23 ranging
from 16% to 51% of corresponding values of aTA at 333 K. Budich et al. (1999)
reported that values of aTA at 333 K and 10 MPa decreased pronouncedly as the
content of terpenes in the liquid phase increased, or when replacing the real OMT
fraction (aTA ~ 2.2 for a 98.3% w/w MT content in the liquid phase) by pure
Fig. 17.9 Separation factors between terpenes and aroma (oxygenated) compounds in model and
real systems as a function of essential oil concentration in the CO2 phase: a23 for the separation at
323 K of limonene and linalool in a ternary CO2 + limonene + linalool system reported by
Vieira de Melo et al. (1999) and Morotomi et al. (1999); a23 for the
separation at 333 K of limonene and citral in a ternary CO2 + limonene + citral system (Fonseca
et al. 2003); and aTA for the separation at 323 K or 333 K between terpene and
aroma (oxygenated) compounds in orange peel oil (Budich et al. 1999)
linalool (aTA ~ 1.2). This result is consistent with the value a23 ~ 1.1 reported by
Morotomi et al. (1999) for the model limonene + linalool system containing 85%
w/w MT in the liquid phase. Based on data in Fig. 17.9, Budich et al. (1999)
recommended deterpenation of orange oil at 333 K, where both aTA (¼1.5) and
the solubility of the oil in high-pressure CO2 are high enough to make the process
economical.
17.4.3 Thermodynamic and Operational Solubility in the CO2

Extraction of Essential Oils
There have been few publications on high-pressure phase equilibria between CO2
and complex mixtures other than those of Temelli et al. (1990) with cold-pressed
orange oil, and those of Budich and Brunner (1999) with orange peel oil. Reported
equilibrium isotherms include those of clove bud oil at 303, 308, 313, 318, and
328 K for 5.8–10.8 MPa (Souza et al. 2004); of fennel seed oil at 303, 313, 323,
and 333 K for 4.74–21.0 MPa (Moura et al. 2005); of vetiver (Vetiveria zizanioides)
root oil at 303, 318, and 333 K for 5–30 MPa (Takeuchi et al. 2008); of candeia
(Eremanthus erythropappus) bark oil at 313, 323, and 333 K for 6.27–25.2 MPa
(Teixeira de Souza et al. 2008); and of priprioca (Cyperus articulatus) rhizome oil at
313, 323, and 333 K and 4.42–29.9 MPa (Moura et al. 2005). These measurements
are usually performed using synthetic methods (Sect. 17.4.1), and are difficult to set
up, in that bubble and dew or cloud points are not easily visually identified for
complex mixtures of CO2 and natural extracts. With the exception of vetiver
root and priprioca rhizome, these studies complement other studies on SCFE
(Table 17.1). Furthermore, besides assisting the analysis of the extraction process,
the results of these studies can help to optimize the condition of the separation step
of the entire SCFE process. Indeed, the high-pressure phase equilibrium for com-
plex CO2 + essential oil systems under typical separation conditions in a SCFE
plant (e.g., 273–288 K and 2–9 MPa, Reverchon and De Marco 2008) determines
the residual solute content in a recycled CO2 stream, which affects the extraction
rate (del Valle et al. 2004), as well as the residual content of CO2 in the extract,
which affected solvent losses during the process (Takeuchi et al. 2008).
Table 17.11 presents the solubilities of essential oil extracts of orange peel, clove
bud, fennel seed, and candeia bark under selected temperature and pressure condi-
tions, as taken from the publications of Budich and Brunner (1999), Souza et al.
(2004), Moura et al. (2005), and Teixeira de Souza et al. (2008), respectively.
Specifically referred to are the values from the P-y branch of isotherms in equilib-
rium diagrams under the conditions where only two phases (a CO2-rich vapor phase
and an essential oil-rich liquid phase) were at equilibrium, which forced the neglect
of many data points under conditions where the authors reported partial liquid
miscibility. Liquid immiscibility conditions in complex systems result from
Table 17.11 Comparison of estimated thermodynamic solubility and operational solubility for
selected studies on high-pressure CO2 extraction of plant essential oils
Extract Temperature Pressure Solubility (Csat,
(T, K) (P, MPa) mg solute/g CO2)
Orange peel oil
Limonene (Iwai et al. (1994) 313 7.17 6.2
Linalool (Iwai et al. 1994) 313 7.17 2.1
Model oil mixturea 313 7.17 6.0
Actual essential oil (Budich and Brunner 1999) 313 7.17 5.1
Actual essential oil (Budich and Brunner 1999) 313 15 CM*
Ibid. in the presence of substrate (Mira et al. 1996, 313 15 13.2–19.3
1999)
Clove oil
Eugenol (Cheng et al. 2000) 313 8.06 5.6
b-Caryophyllene (Stahl et al. 1988) 313 8.06 2.3
Model oil mixtureb 313 8.06 5.0
Actual essential oil (Souza et al. 2004) 313 8.06 11
Ibid. in the presence of substrate (Martı́nez et al. 308 10 230
2007)
Fennel oil
Limonene (Iwai et al. 1994) 313 8.02 9.6
Anethole (Stahl et al. 1988) 313 8.02 6
Model oil mixturec 313 8.02 5
Actual essential oil (Moura et al. 2005) 313 8.02 26
Ibid. in the presence of substrate (Reverchon et al. 313 9 2
1999)
Candeia oil
Actual essential oil (Teixeira de Souza et al. 2008) 313 6.17 111
Ibid. in the presence of substrate (Teixeira de 313 10 4.5
Souza et al. 2008)
a
Ideal solubility estimated neglecting interactions between solutes and assuming orange peel oil as
a mixture of 98% (w/w) limonene and 2% (w/w) linalool
b
Ideal solubility estimated neglecting interactions between solutes and assuming clove oil as a
mixture of 86% (w/w) eugenol and 14% (w/w) b-caryophyllene
c
Ideal solubility estimated neglecting interactions between solutes and assuming fennel oil as a
mixture of 9% w/w limonene and 91% w/w anethole
*
Complete miscibility between the essential oil and high-pressure CO2
interactions between minor components in the essential oil mixture, and those
interactions have large effects on the actual equilibrium. One such effect is a limited
dependence on temperature of the equilibrium concentration of the essential oil in
the high-pressure CO2-rich phase. Another effect is unusual variations in essential
oil solubility as a function of system temperature and pressure, such as the increase
in solubility of cold-pressed orange oil in high-pressure CO2, as reported by Temelli
et al. (1990) at 313 K and 12.4 MPa. Taking further advantage of equilibrium data
of essential oils in high-pressure CO2 collected up to now demands more detailed
experimental evidence.
The thermodynamic solubility values reported in Table 17.11 are different from
those expected based on measurements for model binary systems of CO2 and the
main MT, OMT, and ST/OST in the actual essential oils. The belief here is that this
finding is due to the effect of minor components in the essential oils which, as
mentioned before, interact with each other and the major components in the
mixture, thus strongly affecting the equilibrium, as exemplified in the following
paragraph.
In general, the solubility of a particular substance in high-pressure CO2,
such as MT or OMT, may be affected by the presence of other substances in the
natural product, such as a wax or a triglyceride, which may exhibit a higher or
lower solubility in the CO2. The example of a major MT and a wax is relevant
in the particular case of essential oils because, as previously mentioned, they are
usually encapsulated within specialized wax-made structures that serve a protec-
tive function in herbs (Gaspar et al. 2001). Sovová et al. (2001) reported high-
pressure phase equilibria data for a ternary CO2 + limonene + blackcurrant oil
system at 313 K and 8–12 MPa, and showed that the solubility of limonene in
high-pressure CO2 decreased in the presence of triglycerides for pressures up to
20% higher than the critical pressure of the CO2 + limonene binary mixture
(Pcm ¼ 8.44 MPa).
17.4.4 Operational Solubility and Sorption Phenomena

in the CO2 Extraction of Essential Oils
Figure 17.1b defines the “operational” solubility of plant essential oils as the con-
centration of the essential oil in high-pressure CO2 at the outlet of the extraction
vessel, provided that CO2 and the herb or spice reach equilibrium conditions as
the CO2 travels along the vessel, and there is some free solute in the substrate. The
initial slope of the cumulative plot of solute yield versus specific CO2 consump-
tion may represent the actual “operational” solubility only if extraction is pre-
ceded by an initial static period to achieve steady temperature and pressure
conditions and equilibration between the substrate and the high-pressure CO2.
Table 17.11 shows that the operational solubility (in the presence of the solid
substrate) of selected essential oils in high-pressure CO2 is smaller than the
thermodynamic solubility (in the absence of the solid substrate) under equivalent
conditions, probably due to binding of the essential oils to the solid matrix or an
insufficient amount of essential oils in the herb or spice to saturate the CO2 (del
Valle et al. 2005; Sovová 2005).
Figure 17.10 shows selected values of operational solubility Cfo as a function
of the ratio between Cso and r, together with trend lines that can be explained on
the basis of either limited availability of the essential oil in the solid matrix or a
sufficient amount to saturate the high-pressure CO2 phase in a case where the
essential oil is not bound to the solid matrix. If the porosity of an extraction
vessel with an inner volume V packed with a milled herb or spice of inner
porosity ep is e, the vessel will contain V [e + (1 e) ep] r of high-pressure
CO2 at system temperature and pressure (at those conditions the density of the
CO2 is r). If the substrate loaded in the extraction vessel (a total weight V (1 e)
Fig. 17.10 Best-fit values of operational solubility of essential oils in high-pressure CO2 as a
function of the initial essential oil content in the herb or spice. Plotted values include substrates
with light essential oils (vc 550 cm3/mol) extracted with low-density (r 650 kg/m3) CO2 –
caraway (Sovová et al. 2004), nutmeg (Machmudah et al. 2006), and sage (Langa et al.
2009) – or high-density (r > 650 kg/m3) CO2 – orange peel (Mira et al. 1996, 1999), nutmeg
(Spricigo et al. 2001; Machmudah et al. 2006), alecrim pimenta (Sousa et al. 2002), and
aniseed (Rodrigues et al. 2003), as well as substrates with heavy essential oils (vc > 550 cm3/mol)
extracted with low-density (r 650 kg/m3) CO2 – chamomile (Povh et al. 2001), carqueja
(Vargas et al. 2006), and valerian (Zizovic et al. 2007a) or high-density (r > 650 kg/m3) CO2
chamomile (Povh et al. 2001), ginger (Martı́nez et al. 2003), marigold (Campos et al.
2005), and valerian (Zizovic et al. 2007a)
ep rs, where rs is the true density of the solid matrix) initially contains a
concentration Cso of essential oil in a solute-free basis, in a situation where
there is not enough solute to saturate the CO2 following a static extraction period,
(17.28) defines the operational solubility:
1 rs
Cfo ¼ e
Cso (17.28)
1
ep 1e þ1 r
where the term e/(1 e) represents the ratio of interparticle void volume to apparent
particle volume. Equation 17.28 suggests that a plot of Cfo versus (Cso/r) (Fig. 17.10)
will result in a straight line with a slope m (17.29):
rs
m¼ e
(17.29)
1
ep 1e þ1
On the other hand, if there is enough solute to saturate the CO2 phase (Cso/r
large), Cfo will reach an asymptotic value Csat (the thermodynamic solubility of the
complex essential oil mixture in the CO2 under process conditions). It is difficult to
compute the slope m in the absence of precise measurements of the bed porosity (e),
the interparticle porosity (ep), and the true density of the matrix (rs), which may
vary depending on the substrate and its pretreatment, but the data in Fig. 17.10 were
plotted under the simplifying assumption that m changes little between substrates.
Regarding the horizontal asymptote, it is important to stress that Csat is a strong
function of the essential oil mixture and the system conditions characterized by the
temperature and density of the CO2 (Chrastil 1982).
Figure 17.10 presents selected data that follow the trend suggested by the
hypothesis of limited solute in a noninteracting solid matrix. A log–log plot was
made to allow a wide range of experimental values in a single plot, and under those
conditions a linear relationship such as (17.28) (power relationship with an expo-
nent one) follows a straight line with a unitary slope (such as the trend line
included in Fig. 17.10 for values of Cso/r below 0.05 g dm3). Fig. 17.10 includes
essential oils with relatively low values of critical volume (Vc 550 cm3/mol)
enriched in monoterpene hydrocarbons, oxygenated monoterpenes, and related
compounds such as those of aniseed, caraway, alecrim pimenta, nutmeg, orange
peel, and sage, and extracts with larger values of critical volume (Vc > 550 cm3/
mol) that are instead enriched in heavier sesquiterpenes, waxes, and related
compounds, such as the extracts of carqueja, chamomile, ginger, marigold, and
valerian. Apparently, heavy extracts and essential oils behave the same for low
solute contents (Cso/r 0.05 g dm3) in the solid matrix, as expected; only for
high solute contents (Cso/r > 0.05 g dm3) do the values for heavy extracts level
off to an apparent solubility of Csat ~ 10 g/kg. Another trend that is apparent in
Fig. 17.10 is that the values of Cfo for low CO2 density (r 650 kg/m3) tend to be
below the values Cfo for higher CO2 densities (r > 650 kg/m3), particularly in the
upper end of values of Cso/r, as expected for an increase in solubility with the
solvent power of CO2.
Table 17.12 complements Fig. 17.10, providing values of operational solubility
(Cfo) reported from studies in Table 17.2 reviewed in this chapter. The data in
Fig. 17.10 exhibit scattering because of variations in substrates and their pretreat-
ments, extraction temperatures, and CO2 densities that are not fully accounted for
in the plot, and some additional values in Table 17.12 are not fully consistent
with the aforementioned trends. Of all single-measurement Cfo values reported in
Table 17.12 (boldo, cinnamon of Cunha, fennel, oregano, and pennyroyal), only the
one for pennyroyal does not follow the general trends in Fig. 17.10. In the case of
rosemary, the data of Coelho et al. (1997) and Bensebia et al. (2009) are inconsis-
tent, and only the Cfo values of Coelho et al. (1997), higher, follow the trend lines in
Fig. 17.10. The data on clove by Ruetsch et al. (2003) and Martı́nez et al. (2007) are
outside (above and/or to the right) the upper limits selected for Fig. 17.10 and are
inconsistent; only the Cfo values of Martı́nez et al. (2007) follow the general trends
in Fig. 17.10. The data of Perakis et al. (2005) on black pepper, to the right of the
upper limit of Cso/r (> 0.05 g m3) in Fig. 17.10, are only slightly below the top
Table 17.12 Summary of operation solubility values in high-pressure CO2 extraction of plant
essential oils studies from Tables 17.1 and 17.2, as a function of the extraction conditions and the
initial solute content of the substrates
Substrate Temperature CO2 density Initial solute Operational
(T, K) (r, kg/m3) content (Cso, solubility
mg solute/g (Cfo, mg
solute-free) solute/g CO2)
Clove (Martı́nez et al. 2007) 306 713 157 230
Clove (Ruetsch et al. 2003) 323 581 212 34.0
Clove (Ruetsch et al. 2003) 323 288 212 2.50
Orange peel (Mira et al. 1996) 323 700 100.0 95.0
Orange peel (Mira et al. 1996) 323 700 45.0 8.00
Black pepper (Ferreira and 313 780 35.8 93.2
Meireles 2002)
Black pepper (Ferreira et al. 303, 323 698, 847 35.8 89.0–85.8
1999)
Black pepper (Ferreira et al. 303–323 698–847 14.7 35.3–24.2
1999)
Black pepper (Perakis et al. 313, 323 384–780 92.0–155 3.80–2.50
2005)
Caraway (Sovová et al. 1994a) 313 623 28.8 80.9
Caraway (Sovová et al. 1994a) 313 484 28.8 18.2
Nutmeg (Spricigo et al. 2001) 296 819 18.0–69.0 67.5
Nutmeg (Machmudah et al. 313–323 629–780 58.0 24.0–9.00
2006)
Parsley (Louli et al. 2004) 318 742 650 33.0
Parsley (Louli et al. 2004) 308, 318 498, 713 120 8.302.80
Cinnamon of Cunhã (Sousa et 288 851 38.5 28.3
al. 2005)
Aniseed (Rodrigues et al. 2003) 313 700–836 31.3–105 27.7–11.0
Alecrim pimenta (Sousa et al. 283–298 728–891 22.4–34.0 22.7–13.2
2002)
Carqueja (Vargas et al. 2006) 313–343 208–486 17.5–24.0 19.1–6.60
Celery (Papamichail et al. 348, 328 498–742 500 8.31–2.12
2000)
Celery (Papamichail et al. 318 498 62.0 2.12
2000)
Ginger (Martı́nez et al. 2003) 293–313 847–905 20.0–25.0 6.41–5.15
Chamomile (Povh et al. 2001) 303, 313 623–809 22.4–34.2 3.71–1.15
Valerian (Zizovic et al. 2007a) 313, 323 384–780 6.14–12.6 3.22–0.511
Marigold (Campos et al. 2005) 313 718, 780 14.2 2.80
Fennel (Reverchon et al. 1999) 323 288 18.3 2.00
Rosemary (Coelho et al. 1997) 308, 313 629–777 7.05 1.98–1.65
Rosemary (Bensebia et al. 313, 333 290–780 32.0 0.356–0.238
2009)
Boldo (Uquiche et al. 313 632 13.1 1.75
submitted)
Pennyroyal (Reis-Vasco et al. 323 384 25.3 1.16
2000)
Sage (Langa et al. 2009) 313, 323 384, 486 12.9–19.2 0.800–0.600
Lavender (Akgun et al. 2000) 323 220–670 15.3 0.418–0.234
Oregano (Uquiche et al. 313 632 6.02 0.390
submitted)
solubility Csat ~ 10 g/kg for heavy extracts. Finally, the data of operational
solubility Cfo of Ferreira et al. (1999) and Ferreira and Meireles (2002) for
black pepper are irregularly high, whereas the data of initial solute content Cso
of Papamichail et al. (2000) for celery and of Louli et al. (2004) for parsley are too
high, considering the typical amount of extractible compounds in those substrates
(Moyler 1993).
Goto et al. (1998) showed that the operational solubility of menthol (the main
component in the essential oil of mint leaves) is smaller than its thermodynamic
solubility in high-pressure CO2 at 313 K and 13.6 MPa, and suggested that the transfer
of menthol to the CO2 phase was limited by strong interactions between the essential
oil and solid matrix. They also claimed weaker interactions between n-triacontane
(the main component in the cuticular waxes of mint leaves) with the solid matrix
than between menthol and the solid matrix because the operational solubility of
n-triacontane was closer to its thermodynamic solubility than the operational solubility
of menthol. As shown in Sect. 17.4.3, solute–solute interactions between the
components of the mint leaf extract may be partially responsible for a reduction
in the apparent solubilities of menthol and n-triacontane in CO2 when they are a part
of a complex essential oil mixture as compared with their corresponding thermody-
namic solubilities in the binary CO2 + menthol or CO2 + n-triacontane systems,
but this does not negate the possibility of a reduction in their apparent solubility due
to some additional interactions between the solutes and the solid matrix.
The effect of solute binding by the solid matrix, which affects solute availability
in SCFE, can be accounted for by an equilibrium sorption isotherm that relates the
concentration of solute in the high-pressure CO2 phase with the residual content of
solute in the solid phase (the pretreated herb or spice) under equilibrium conditions.
Some authors hypothesize that the operational solubility Cfo depends on the sub-
strate and extraction conditions, but does not depend on the solute content in the
substrate, as reported in Table 17.12. Other authors hypothesize a constant partition
coefficient K (17.20) for essential oils between the high-pressure CO2 phase and the
solid phase, which may correspond to a linear sorption isotherm (17.30a), derived
from (17.20) or the initial slope of another sorption isotherm model for a low
essential oil content in the pretreated herb or spice, such as the Freundlich (17.30b),
Langmuir (17.30c), or Brunauer-Emmett-Teller (BET, (17.30d) models. Finally,
selected authors combine the possibility of a constant Csat for large concentrations
of essential oil in the solid matrix (Cso > Clim) and a constant partition coefficient
for smaller values of Cso ( Clim) using the so-called isotherm of Perrut et al.
(1997). Table 17.13 summarizes the values of the linear partition coefficient K (or
equivalent partition coefficient, (17.31)), as reported in studies from Table 17.2
reviewed in this chapter.
The mathematical models for the linear, Freundlich, Langmuir, and BET’s
sorption isotherms are, respectively, as follows:
Cf
Cs ¼ (17.30a)
K
Table 17.13 Summary of values of solute partition coefficients in high-pressure CO2 extraction of
plant essential oils in studies from Tables 17.1 and 17.2, as a function of the extraction conditions
and the initial solute content of the substrates
Substrate Temperature CO2 density Initial solute Solute partition
(T, K) (r, kg/m3) content (Cso, mg coefficient
solute/g solute- (K, –)
free)
Cinnamon of Cunhã (Sousa 288 851 38.5 2.38–2.00
et al. 2005)
Carqueja (Vargas et al. 2006) 313 486 20.4 0.813
Clove (Ruetsch et al. 2003) 323 288, 581 212.1 0.314
Clove (Daghero et al. 2004) 323 288, 581 212.1 0.022–0.008
Nutmeg (Machmudah et al. 313 780 58.0 0.300–0.150
2006)
Nutmeg (Machmudah et al. 323 700, 780 58.0 0.300–0.100
2006)
Nutmeg (Machmudah et al. 318 742 58.0 0.200
2006)
Boldo (Uquiche et al. 313 632 13.1 0.134
submitted)
Parsley (Louli et al. (2004) 308 713 120 0.067a
Parsley (Louli et al. (2004) 318 498, 742 650, 120 0.051–0.023a
Parsley (Louli et al. (2004) 308 713 63 0.0099b
Parsley (Louli et al. (2004) 318 498, 742 63, 450 0.0076–0.0038b
Celery (Papamichail et al. 328 654 500 0.0605a
2000)
Celery (Papamichail et al. 318 498, 742 63, 270 0.0585–0.0471a
2000)
Celery (Papamichail et al. 328 654 417 0.0046b
2000)
Celery (Papamichail et al. 318 498, 742 417, 476 0.0041–0.00001b
2000)
Oregano (Uquiche et al. 313 632 6.02 0.0647
submitted)
Peppermint (Goto et al. 1993) 313 445, 777 – 0.0506–0.0202
Sage (Langa et al. 2009) 313 486 17.3–19.2 0.047–0.042
Sage (Langa et al. 2009) 323 384 12.9 0.033
Black pepper (Perakis et al. 313 486, 629 93.0, 134.0 0.0090–0.0025
2005)
Black pepper (Perakis et al. 323 384 84.0 0.0063
2005)
Pennyroyal (Sovová 2005) 323 384 25.3 0.063
a
Sovová’s model, model R-SO
b
Model LDF-UENA
n
ðCf Þ
Cs ¼ (17.30b)
k
k Cm Cf
Cs ¼ (17.30c)
1 þ k Cf
k Cm Cf
Cs ¼ (17.30d)
ðCsat Cf Þ½Csat þ ðk 1ÞCf
where k and n are empirical sorption energy parameters, and Cm (the so-called
monolayer coverage of the solid surface) is the maximal amount of essential oil that
can hold the solid matrix. Ruetsch et al. (2003) and Daghero et al. (2004) arbitrarily
assumed that the monolayer coverage corresponded to the initial solute content
(Cm ¼ Cso) in clove. The sorption isotherm models (17.30a–17.30d) are written
with Cs instead of Cf as the independent variable because they are adapted from the
literature on adsorptive separations. In adsorptive separation processes, where the
solute in a fluid phase transfers to a solid phase, there is no upper limit to solute
concentration in the fluid phase (for all practical purposes, solute and mobile or
fluid phases are mutually miscible), and eventual saturation of the solid matrix
imposes an upper limit on the concentration of the solute in the solid for high
concentration in the fluid under equilibrium conditions. Depending on the herb or
spice and its pretreatment, this upper limit in concentration of solute in the solid
substrate is not reached when the concentration of the essential oil in the CO2 is
limited by its solubility under process conditions. Because of that, Ruetsch et al.
(2003), Daghero et al. (2004), and Salimi et al. (2008) imposed an upper limit
(Cf Csat) to the value of solute concentration in the CO2 phase for high concen-
trations of solute in the solid phase (Fig. 17.11). Equation 17.31 defines the limit
value of K for small value of Cs (Cs ! 0) for both Langmuir’s and BET’s sorption
isotherm models (Frendlich’s isotherm model predicts limit values, K ¼ 0 for
n > 1 and K ! 1 for n < 1):
1
K¼ (17.31)
k Cm
Observation did not reveal special trends in values of the linear partition
coefficients (Table 17.13) as a function of the substrate and its essential oil content,
the temperature of the system, or the density of the CO2 (plots not shown). Based on
(17.28) for a noninteracting solid matrix, for a packed bed with porosity e ¼ 0.6, a
solid substrate with true density rs ¼ 1,000 kg/m3, an inner porosity ep ¼ 0.1, and
high-pressure CO2 at 323 K and 9 MPa (r ¼ 287.5 kg/m3), the expected value of
the partition coefficient in a situation of limited essential oil content in the plant
material would be K ¼ 0.222, which is between upper and lower limit values
reported in Table 17.13. No evidence was found that K increases as the initial
essential oil content in the herb or spice increases. Also, no evidence was found that
Fig. 17.11 Sorption isotherm models for equilibrium partition of essential oil between high-
pressure CO2 and an herb or a spice as a function of the solute content in the substrate under
equilibrium conditions at constant system temperature and pressure
K decreases as the density of the CO2 increases, as expected (a comparison of

(17.28), (17.29), and (17.30a) suggests that K ¼ m/r for a situation where the
essential oil does not interact with the solid matrix).
17.5 Concluding Remarks
This chapter reviewed mass transfer and phase equilibrium parameters that can be
used to design industrial SCFE processes for plant essential oils. Relevant mass
transfer parameters include an axial dispersion coefficient (Dax) for the migration of
the solute in the SCF along the bed; an external mass transfer coefficient (kf) for its
movement through the stationary SCF film surrounding the solid particles; and an
effective diffusivity (De) for its movement through the solid matrix, which were
computed in the form of a so-called microstructural factor (FM). This review
suggests neglecting axial dispersion effects to simplify the mass transfer models.
Based on this review, it is recommended that SCFE experiments be carried out
under forced convection conditions, and that the external mass transfer coefficient
(kf) be estimated using a literature-based correlation between dimensionless vari-
ables valid for mass transfer in packed beds operating with SCFs. Best-fitting
usually provides underestimations of kf because of the underestimation of internal
resistances to mass transfer, overestimation of the specific surface of the solid
substrate, neglecting of solvent flow heterogeneity effects when using a small D/dp
ratio, and/or neglecting of natural convention effects when using low-Re flow
conditions.
The values of the microstructural factor for inner mass transfer in the herb or
spice estimated in this chapter ranged from FM ¼ 102 to FM ¼ 105, which sug-
gested pronounced limitations to mass transfer within the solid matrix in the high-
pressure CO2 extraction of plant essential oils. The estimated values of FM, unlike
those expected, depended on the system (temperature, pressure) conditions, the
superficial velocity of the CO2, and/or the particle size of the substrate. Further-
more, for equivalent experiments, the best-fit values of FM changed dramatically
depending on the applied mathematical model, which raised questions about the
validity of some of the hypotheses of these mass transfer models. To improve the
modeling of high-pressure CO2 extraction of plant essential oils, extraction experi-
ments should be complemented by measurements using microscopy and allied/
complementary techniques to fully characterize the pretreated solid matrix at a
relevant scale (Aguilera and Stanley 1999; Zizovic et al. 2005, 2007a, b, c;
Stamenić et al. 2008). Such measurements would result in microstructure– extract-
ability relationships, which could be taken advantage of to optimize the pretreat-
ment of the herb or spice samples prior to SCFE.
Regarding phase equilibrium data for designing industrial SCFE processes for
plant essential oils, the conclusion here is that their “operational” solubility in high-
pressure CO2 depends markedly on the availability of the solute (the complex
essential oil mixture) and its partition between the solid matrix (the herb or spice)
and the SCF. The reviewed literature included several phase equilibrium studies
using binary systems CO2 + pure essential oil (mainly MT and OMT) component,
few studies using ternary CO2 + limonene + citral/linalool systems, and limited
studies using CO2 and complex essential oil mixtures (low-pressure CO2 extracts of
herbs or spices). Further advancements in this field will require additional fluid
phase equilibrium measurements using binary mixtures of CO2 and ST or OST
compounds, CO2 + MT + OMT model ternary mixtures representing plant essen-
tial oils other than citrus oils, or CO2 + complex essential oil mixtures. Further-
more, given that the “operational” solubility of essential oils in high-pressure CO2
does not depend solely on thermodynamic solubility, quantifying the effect of the
availability and binding of the solute to the herb or spice demands additional
measurements of the solid-SCF phase equilibrium in addition to the aforementioned
fluid phase equilibrium data. Because there is no real evidence in the literature that
the solute partition between the solid substrate and the SCF is constant, sorption
isotherms should be experimentally measured instead of assuming a sorption
pattern and achieving best-fit model parameters as part of the data analysis process.
Acknowledgments The present work was funded by the Chilean agency Fondecyt (Regular
project 105–0675 and International Cooperation project 703–0033). We are indebted to Verónica
Glatzel (PUC) for recalculating from the literature some of the values of external mass transfer
coefficient (kf), and effective diffusivities (De) that we report in Sect. 17.3.2 and 17.3.3, respec-
tively; and to Gustavo Lozano (TUHH) for simulating the solubility isotherms for selected
essential oil components included in Figs. 17.7 and 17.8 using the predictive methodology
described in Sect. 17.4.1 in PE 2000.
References
Aghel N, Yamini Y, Hadjiakhoondi A, Pourmortazavi SM (2004) Supercritical carbon dioxide

extraction of Mentha pulegium L. essential oil. Talanta 62:407–411
Aguilera JM, Stanley DW (1999) Microstructural principles of food processing and engineering,
2nd edn. Aspen Publishers, Gaithersburg, MD
Akgun M, Akgun NA, Dincer S (1999) Phase behaviour of essential oil components in supercriti-
cal carbon dioxide. J Supercrit Fluids 15:117–125
Akgun M, Akgun NA, Dincer S (2000) Extraction and modeling of lavender flower essential oil
using supercritical carbon dioxide. Ind Eng Chem Res 39:473–477
Araus K, Uquiche E, del Valle JM (2009) Matrix effects in supercritical CO2 extraction of
essential oils from plant material. J Food Eng 92:438–447
Bakkali F, Averbeck S, Averbeck D, Zhiri A, Idaomar M (2005) Cytotoxicity and gene induction
by some essential oils in the yeast Saccharomyces cerevisiae. Mutat Res 585:1–13
Bensebia O, Barth D, Bensebia B, Dahmani A (2009) Supercritical CO2 extraction of rosemary:
Effect of extraction parameters and modelling. J Supercrit Fluids 49:161–166
Benvenuti F, Gironi F (2001) High-pressure equilibrium data in systems containing supercritical
carbon dioxide, limonene, and citral. J Chem Eng Data 46:795–799
Berna A, Chafer A, Monton JB (2000) Solubilities of essential oil components of orange in
supercritical carbon dioxide. J Chem Eng Data 45:724–727
Brielmann HL, Setzer WN, Kaufman PB, Kirakosyan A, Cseke LJ (2006) Phytochemicals: The
chemical components of plants. In: Cseke LJ, Kirakosyan A, Kaufman PB, Warber S, Duke JA,
Brielmann HL (eds) Natural products from plants, 2nd edn. CRC Press, Boca Raton, FL,
pp 1–49
Brunner G (1984) Mass transfer from solid material in gas extraction. Ber Bunsen Ges Phys Chem
88:887–891
Brunner G (1994) Gas extraction: an introduction to fundamentals of supercritical fluids and the
application to separation processes. Springer, New York
Budich M, Brunner G (1999) Vapor–liquid equilibrium data and flooding point measurements of
the mixture carbon dioxide + orange peel oil. Fluid Phase Equilib 158–160:759–773
Budich M, Heilig S, Wesse T, Leibk€ uchler V, Brunner G (1999) Countercurrent deterpenation of
citrus oils with supercritical CO2. J Supercrit Fluids 14:105–114
Campos LMAS, Michielin EMZ, Danielski L, Ferreira SRS (2005) Experimental data and
modeling the supercritical fluid extraction of marigold (Calendula officinalis) oleoresin. J
Supercrit Fluids 34:163–170
Carman PC (1937) Fluid flow through granular beds. Trans Inst Chem Eng 15:150–166
Carvalho RN Jr, Corazza ML, Cardozo-Filho L, Meireles MAA (2006) Phase equilibrium for
(camphor + CO2), (camphor + propane), and (camphor + CO2 + propane). J Chem Eng Data
51:997–1000
Catchpole OJ, King MB (1994) Measurement and correlation of binary diffusion coefficients in
near critical fluids. Ind Eng Chem Res 33:1828–1837
Catchpole OJ, Andrews EW, Toikka GN, Wilkinson GT (1994) Mathematical models for the
extraction of oils from plant matrices using near-critical solvent. In: Perrut M, Brunner G (eds)
Proceedings of the third symposium on supercritical fluids, vol 2. Institut National Polytechni-
que de Lorraine, Lorraine, France, pp 47–52
Catchpole OJ, Bernig R, Bott MB (1996a) Measurement and correlation of packed-bed axial
dispersion coefficients in supercritical carbon dioxide. Ind Eng Chem Res 35:824–828
Catchpole OJ, Grey JB, Smallfield BM (1996b) Near-critical extraction of sage, celery, and
coriander seed. J Supercrit Fluids 9:273–279
Chafer A, Berna A, Monton JB, Mulet A (2001) High pressure solubility data of the system
limonene plus linalool plus CO2. J Chem Eng Data 46:1145–1148
Chandler K, Pouillot FLL, Eckert CA (1996) Phase equilibria of alkanes in natural gas systems. 3.
Alkanes in carbon dioxide. J Chem Eng Data 41:6–10
Chang CMJ, Chen CC (1999) High-pressure densities and P–T–x–y diagrams for carbon dioxide +
linalool and carbon dioxide + limonene. Fluid Phase Equilib 163:119–126
Cheng K-W, Kuo S-J, Muoi Tang M, Chen Y-P (2000) Vapor–liquid equilibria at elevated
pressures of binary mixtures of carbon dioxide with methyl salicylate, eugenol, and diethyl
phthalate. J Supercrit Fluids 18:87–99
Chrastil J (1982) Solubility of solids and liquids in supercritical gases. J Phys Chem 86:
3016–3021
Coelho JAP, Mendes RL, Provost MC, Cabral JMS, Novais JM, Palavra AMF (1997) Supercritical
carbon dioxide extraction of volatile compounds from rosemary. In: Abraham MA, Sunol AK
(eds) Supercritical fluids. Extraction and pollution prevention. American Chemical Society,
Washington, DC, pp 101–109
Coelho JAP, Pereira AP, Mendes RL, Palavra AMF (2003) Supercritical carbon dioxide extraction
of Foeniculum vulgare volatile oil. Flavour Frag J 18:316–319
Crank J (1975) The mathematics of diffusion, 2nd edn. Oxford University Press, New York
Cygnarowicz-Provost M (1996) Design and economic analysis of supercritical fluid extraction
processes. In: King JW, List GR (eds) Supercritical fluid technology in oil and lipid chemistry.
AOCS Press, Champaign, IL, pp 155–179
Daghero J, Ruetsch L, Zacchi P, Mattea M (2004) Supercritical CO2 extraction of herbaceous
matrices. Pilot plant experiments and modeling. V Encuentro Brasileño sobre Fluidos Super-
crı́ticos (EBFS 2004), Florianopolis, Brasil
Danielski L, Campos LMAS, Bresciani LFV, Hense H, Yunes RA, Ferreira SRS (2007) Marigold
(Calendula officinalis L.) oleoresin: solubility in SC-CO2 and composition profile. Chem Eng
Process 46:99–106
Daubert TE, Danner RP (1989) Physical and thermodynamic properties of pure chemicals: data
compilation. Hemisphere Publishers, New York
del Valle JM, Aguilera JM (1999) Extracción con CO2 a alta presión. Fundamentos y aplicaciones
en la industria de alimentos. Food Sci Technol Int 5:1–24
del Valle JM, Catchpole OJ (2005) Transferencia de masa en lechos empacados operando con
fluidos supercrı́ticos. I. Correlación de coeficientes de dispersión axial. XVI Congreso Chileno
de Ingenierı́a Quı́mica, Pucón, Chile
del Valle JM, de la Fuente JC, Cardarelli DA (2005) Contributions to supercritical extraction of
vegetable substrates in Latin America. J Food Eng 67:35–57
del Valle JM, de la Fuente JC (2006) Supercritical CO2 extraction of oilseeds: Review of kinetic
and equilibrium models. CRC Crit Rev Food Sci Nutr 46:131–160
del Valle JM, Rivera O, Mattea M, Ruetsch L, Daghero J, Flores A (2004) Supercritical CO2
processing of pretreated rosehip seeds: Effect of process scale on oil extraction kinetics. J
Supercrit Fluids 31:159–174
del Valle JM, Mena C, Budinich M (2008) Extraction of garlic with supercritical CO2 and
conventional organic solvents. Braz J Chem Eng 25:535–542
Della Porta G, Taddeo R, D’urso E, Reverchon E (1998) Isolation of clove bud and star anise
essential oil by supercritical CO2 extraction. Lebensm Wiss Technol 31:454–460
Della Porta G, Porcedda S, Marongiu B, Reverchon E (1999) Isolation of eucalyptus oil by
supercritical fluid extraction. Flavour Frag J 14:214–218
Denny EFK (1991) Field distillation for herbaceous oils, 2nd edn. Denny, McKenzie Associates,
Lilydale (Tasmania), Australia
di Giacomo G, Brandani V, del Re G, Mucciante V (1989) Solubility of essential oil components in
compressed supercritical carbon dioxide. Fluid Phase Equilib 52:405–411
Dullien F (1992) Porous media. Fluid transport and pore structure, 2nd edn. Academic, San Diego,
CA
Eggers R (1996) Supercritical fluid extraction of oilseeds/lipids in natural products. In: King JW,
List GR (eds) Supercritical fluid technology in oil and lipid chemistry. AOCS Press, Cham-
paign, IL, pp 35–64
Eggers R, Ambrogi A, von Schnitzler J (2000) Special features of SFC solid extraction of
natural products: Deoling of wheat gluten and extraction of rose hip oil. Braz J Chem Eng
17:329–334
Espinosa-Dı́az MA, Guetachew T, Landy P, Jose J, Voilley A (1999) Experimental and estimated
saturated vapor pressure of aroma compounds. Fluid Phase Equilib 157:257–270
Esquivel MM, de Sousa CL, Ribeiro MA, Bernardo-Gil MG (1996) Mathematical models for
supercritical extraction of oregano “Origanum virens L.”. In: von Rohr PR, Trepp C (eds) High
pressure chemical engineering. Elsevier, Amsterdam, The Netherlands, pp 525–530
Fahien RW, Smith JM (1955) Mass transfer in packed beds. Am Inst Chem Eng J 1:28–37
Ferreira SRS, Meireles MAA (2002) Modeling the supercritical fluid extraction of black pepper
(Piper nigrum L.) essential oil. J Food Eng 54:263–269
Ferreira SRS, Nikolov ZL, Doraiswamy LK, Meireles MAA, Petentate A (1999) Supercritical
fluid extraction of black pepper (Piper nigrum L.) essential oil. J Supercrit Fluids 14:235–245
Fonseca J, Simoes PC, Nunes da Ponte M (2003) An apparatus for high-pressure VLE measure-
ments using a static mixer. Results for (CO2 + limonene + citral) and (CO2 + limonene +
linalool). J Supercrit Fluids 25:7–17
Francisco JD, Sivik B (2002) Solubility of three monoterpenes, their mixtures and eucalyptus leaf
oils in dense carbon dioxide. J Supercrit Fluids 23:11–19
Funazukuri T, Kong C, Kagei S (1998) Effective axial dispersion coefficients in packed beds under
supercritical conditions. J Supercrit Fluids 13:169–175
Gamse T, Marr R (2000) High-pressure phase equilibria of the binary systems carvone-carbon
dioxide and limonene-carbon dioxide at 303, 313 and 323 K. Fluid Phase Equilib 171:165–174
Gardner DS (1993) Commercial scale extraction of alpha acids and hop oils with compressed CO2.
In: King MB, Bott TR (eds) Extraction of natural products using near-critical solvents. Blackie
Academic & Professional, London, UK, pp 84–100
Gaspar F (2002) Extraction of essential oils and cuticular waxes with compressed CO2: Effect of
extraction pressure and temperature. Ind Eng Chem Res 41:2497–2503
Gaspar F, Santos R, King MB (2001) Disruption of glandular trichomes with compressed CO2:
Alternative matrix pre-treatment for CO2 extraction of essential oils. J Supercrit Fluids
21:11–22
Gaspar F, Lu T, Santos R, Al-Duri B (2003) Modelling the extraction of essential oils with
compressed carbon dioxide. J Supercrit Fluids 25:247–260
Germain JC, del Valle JM, de la Fuente JC (2005) Natural convection retards supercritical CO2
extraction of essential oils and lipids from vegetable substrates. Ind Eng Chem Res
44:2879–2886
Ghoreishi SM, Akgerman A (2004) Dispersion coefficients of supercritical fluid in fixed beds. Sep
Purif Technol 39:39–50
Gironi F, Maschietti M (2008) Continuous countercurrent deterpenation of lemon essential oil by
means of supercritical carbon dioxide: experimental data and process modeling. Chem Eng Sci
63:651–661
Goodarznia I, Eikani M (1998) Supercritical carbon dioxide extraction of essential oils: Modelling
and simulation. Chem Eng Sci 53:1387–1395
Goto M, Sato M, Hirose T (1993) Extraction of peppermint oil by supercritical carbon dioxide.
J Chem Eng Jpn 26:401–407
Goto M, Roy BC, Hirose T (1996) Shrinking-core leaching model for supercritical-fluid extrac-
tion. J Supercrit Fluids 9:128–133
Goto M, Roy B, Kodama A, Hirose T (1998) Modeling supercritical fluid extraction process
involving solute-solid interaction. J Chem Eng Jpn 32:171–177
Hall HK (ed) (2001) Landolt-B€ ornstein: Numerical data and functional relationships in science
and technology – New Series. Group 4: Physical chemistry, Vol. 20: Vapor pressure of
chemicals. Subvolume B: Vapor pressure and antoine constants for oxygen containing organic
compounds. Springer, Berlin, Germany
Han N-H, Bhakta J, Carbonell RG (1985) Longitudinal and lateral dispersion in packed beds:
effect of column length and particle size distribution. Am Inst Chem Eng J 31:277–288
Helmig D, Revermann T, Pollmann J, Kaltschmidt O, Jiménez-Hernández A, Bocquet F, David D
(2003) Calibration system and analytical considerations for quantitative sesquiterpene mea-
surements in air. J Chromatogr A 1002:193–211
Hong IK, Rho SW, Lee KS, Lee WH, Yoo KP (1990) Modeling of soybean oil bed extraction with
supercritical carbon dioxide. Korean J Chem Eng 7:40–46
Hubert P, Vitzthum OG (1978) Fluid extraction of hops, spices and tobacco with supercritical
gases. Angew Chem Int Ed Engl 17:710–715
Hybertson BM (2007) Solubility of the sesquiterpene alcohol patchoulol in supercritical carbon
dioxide. J Chem Eng Data 52:235–238
Iwai Y, Hosotani N, Morotomi T, Koga Y, Arai Y (1994) High-pressure vapor-liquid-equilibria for
carbon-dioxide plus linalool. J Chem Eng Data 39:900–902
Iwai Y, Morotomi T, Sakamoto K, Koga Y, Arai Y (1996) High-pressure vapor-liquid equilibria
for carbon dioxide plus limonene. J Chem Eng Data 41:951–952
Jimenez-Carmona MM, Ubera JL, Luque de Castro MD (1999) Comparison of continous subcriti-
cal water extraction and hydrodistillation of marjoram essential oil. J Chromatogr A
855:625–632
Kim KH, Hong J (1999) Equilibrium solubilities of spearmint oil components in supercritical
carbon dioxide. Fluid Phase Equilib 164:107–115
Kim KH, Hong J (2002) A mass transfer model for super- and near-critical CO2 extraction of
spearmint leaf oil. Sep Sci Technol 37:2271–2288
King MB, Catchpole O (1993) Physico-chemical data required for the design of near-critical fluid
extraction process. In: King MB, Bott TR (eds) Extraction of natural products using near-
critical solvents. Blackie Academic & Professional, London, UK, pp 184–231
Kotnik P, Skerget M, Knez Z (2007) Supercritical fluid extraction of chamomile flower heads:
Comparison with conventional extraction, kinetics and scale-up. J Supercrit Fluids 43:
192–198
Lack E, Seidlitz H (1993) Commercial scale decaffeination of coffee and tea using supercritical
CO2. In: King MB, Bott TR (eds) Extraction of natural products using near-critical solvents.
Blackie Academic & Professional, London, UK, pp 101–139
Langa E, Della Porta G, Palavra AMF, Urieta JS, Mainar A (2009) Supercritical fluid extraction of
Spanish sage essential oil: Optimization of the process parameters and modelling. J Supercrit
Fluids 49:174–181
Lee CH, Holder GD (1995) Use of supercritical fluid chromatography for obtaining mass
transfer coefficients in fluid-solid systems at supercritical conditions. Ind Eng Chem Res
34:906–914
Leeke GA, Santos R, King M (2001) Vapor-liquid equilibria for the carbon dioxide plus carvacrol
system at elevated pressures. J Chem Eng Data 46:541–545
Louli V, Folas G, Voutsas E, Magoulas K (2004) Extraction of parsley seed oil by supercritical
CO2. J Supercrit Fluids 30:163–174
Ma YH, Evans LB (1968) Transient diffusion from a well-stirred reservoir to a body of arbitrary
shape. J Am Inst Chem Eng 14:956–961
Machmudah S, Sulaswatty A, Sasaki M, Goto M, Hirose T (2006) Supercritical CO2 extraction of
nutmeg oil: Experiments and modeling. J Supercrit Fluids 39:30–39
Marrone C, Poletto M, Reverchon E, Stassi A (1998) Almond oil extraction by supercritical CO2:
Experiments and modelling. Chem Eng Sci 53:3711–3718
Marteau Ph, Obriot J, Tufeu R (1995) Experimental determination of vapor-liquid equilibria of
CO2 + limonene and CO2 + citral mixtures. J Supercrit Fluids 8:20–24
Martin AJP, Synge R (1941) A new form of chromatogram employing two liquid phases. 1. A
theory of chromatography. 2. Application to the micro-determination of the higher mono-
amino-acids in proteins. Biochem J 35:1358–1368
Martı́nez J, Monteiro AR, Rosa PTV, Marques MOM, Meireles M (2003) Multicomponent model
to describe extraction of ginger oleoresin with supercritical carbon dioxide. Ind Eng Chem Res
42:1057–1063
Martı́nez J, Rosa PTV, Meireles MAA (2007) Extraction of clove and vetiver oils with supercriti-
cal carbon dioxide: modeling and simulation. Open Chem Eng J 1:1–7
Matos HA, Gomes de Azevedo E, Simoes PC, Carrondo MT, Nunes da Ponte M (1989) Phase
equilibria of natural flavours and supercritical solvents. Fluid Phase Equilib 52:357–364
Michielin EMZ, Rosso SR, Franceschi E, Borges GR, Corazza ML, Oliveira JV, Ferreira SRS
(2009) High-pressure phase equilibrium data for systems with carbon dioxide, a-humulene and
trans-caryophyllene. J Chem Thermodyn 41:130–137
Mira B, Blasco M, Subirats S, Berna A (1996) Supercritical CO2 extraction of essential oils from
orange peel. J Supercrit Fluids 9:238–243
Mira B, Blasco M, Berna A, Subirats S (1999) Supercritical CO2 extraction of essential oil from
orange peel. Effect of operation conditions on the extract composition. J Supercrit Fluids
14:95–104
Miraldi E, Ferri S, Franchi GG, Giorgi G (1996) Peumus boldus essential oil: New constituents and
comparison of oils from leaves of different origin. Fitoterapia 67:227–230
Mišić D, Zizovic I, Stamenić M, Ašanin R, Ristić M, Petrović SD, Skala D (2008) Antimicrobial
activity of celery fruit isolates and SFE process modeling. Biochem Eng J 42:148–152
Morotomi T, Iwai Y, Yamaguchi H, Arai Y (1999) High-pressure vapor-liquid equilibria for
carbon dioxide plus limonene plus linalool. J Chem Eng Data 44:1370–1372
Moura LS, Corazza ML, Cardozo-Filho L, Meireles MAA (2005) Phase equilibrium measure-
ments for the system fennel (Foeniculum vulgare) extract + CO2. J Chem Eng Data
50:1657–1661
Moyler DA (1993) Extraction of flavours and fragrances with compressed CO2. In: King MB, Bott
TR (eds) Extraction of natural products using near-critical solvents. Blackie Academic &
Professional, London, UK, pp 140–183
Mukhopadhyay M (2000) Natural extracts using supercritical carbon dioxide. CRC Press, Boca
Raton, FL
Mukhopadhyay M, De SK (1995) Fluid-phase behavior of close molecular-weight fine chemicals
with supercritical carbon dioxide. J Chem Eng Data 40:909–913
NIST (2000) Fluid Thermodynamic and Transport Properties (version 5.0). http://www.nist.gov/
srd/nist23.htm
Núñez GA, del Valle JM, de la Fuente JC (2010) Solubilities in supercritical carbon dioxide of
(2E,6E)-3,7,11-trimethyldodeca-2,6,10-trien-1-ol (farnesol) and (2 S)-5,7-dihydroxy-2-
(4-hydroxyphenyl)chroman-4-one (naringenin). J Chem Eng Data 55:3863–3868
Özer EÖ, Platin S, Akman U, Hortaçsu Ö (1996) Supercritical carbon dioxide extraction of
spearmint oil from mint-plant leaves. Can J Chem Eng 74:920–928
Papamichail I, Louli V, Magoulas K (2000) Supercritical fluid extraction of celery seed oil.
Pavlicek J, Richter M (1993) High pressure vapour-liquid equilibrium in the carbon dioxide–
a-pinene system. Fluid Phase Equilib 90:125–133
Perakis C, Louli V, Magoulas K (2005) Supercritical fluid extraction of black pepper oil. J Food
Eng 71:386–393
Perrut M, Clavier JY, Poletto M, Reverchon E (1997) Mathematical modeling of sunflower seed
extraction by supercritical CO2. Ind Eng Chem Res 36:430–435
Pfaf-Šovljanski II, Grujić OS, Peruničić MB, Cvetković IM, Zeković Z (2005) Supercritical
carbon dioxide hop extraction. APTEFF 36:111–120
Pfohl O, Petkov S, Brunner G (2000) PE 2000: a powerful tool to correlate phase equilibria.
Herbert Utz Verlag, M€ unchen, Germany
Podlaski S, Chrobak Z, Wyszkowska Z (2003) The effect of parsley seed hydration treatment and
pelleting on seed vigour. Plant Soil Environ 49:114–118
Poling BE, Prausnitz JM, O’Connell J (2000) The properties of gases and liquids, 5th edn.
McGraw-Hill, New York
Povh NP, Marques MOM, Meireles MAA (2001) Supercritical CO2 extraction of essential oil and
oleoresin from chamomile (Chamomilla recutita L.). J Supercrit Fluids 21:245–256
Puiggené J, Larrayoz MA, Recasens F (1997) Free liquid-to-supercritical fluid mass transfer in
packed beds. Chem Eng Sci 52:195–212
Quirin K-W, Gerard D (2007) Supercritical fluid extraction (SFE). In: Ziegler H (ed) Flavourings:
production, composition, applications, regulations, 2nd edn. Wiley-VCH, Weinheim, Germany,
pp 49–65
Raal JD, M€uhlbauer AL (1998) Phase equilibria: measurement and computation. Taylor &
Francis, Washington, DC
Raeissi S, Peters CJ (2005) Experimental determination of high-pressure phase equilibria of the
ternary system carbon dioxide + limonene + linalool. J Supercrit Fluids 35:10–17
Reis-Vasco EMC, Coelho JAP, Palavra AMF, Marrone C, Reverchon E (2000) Mathematical
modelling and simulation of pennyroyal essential oil supercritical extraction. Chem Eng Sci
55:2917–2922
Reverchon E (1996) Mathematical modeling of supercritical extraction of sage oil. J Am Inst
Chem Eng 42:1765–1771
Reverchon E (1997) Supercritical fluid extraction and fractionation of essential oils and related
products. J Supercrit Fluids 10:1–37
Reverchon E, De Marco I (2006) Supercritical fluid extraction and fractionation of natural matter.
Reverchon E, De Marco I (2008) Essential oils extraction and fractionation using supercritical
fluids. In: Martı́nez JL (ed) Supercritical fluid extraction of nutraceuticals and bioactive
compounds. CRC Press, Boca Raton, FL, pp 305–335
Reverchon E, Marrone C (1997) Supercritical extraction of clove bud essential oil: Isolation and
mathematical modeling. Chem Eng Sci 52:3421–3428
Reverchon E, Sesti Osséo L (1994a) Modelling the supercritical extraction of basil oil. In: Perrut M,
Brunner G (eds) Proceedings of the third symposium on supercritical fluids, Vol. 2., pp 189–196
Reverchon E, Sesti Osséo L (1994b) Supercritical CO2 extraction of basil oil: Characterization of
products and process modeling. J Supercrit Fluids 7:185–190
Reverchon E, Donsi G, Sesti Osséo L (1993a) Modeling of supercritical fluid extraction from
herbaceous matrices. Ind Eng Chem Res 32:2721–2726
Reverchon E, Russo P, Stassi A (1993b) Solubilities of solid octacosane and triacontane in
supercritical carbon dioxide. J Chem Eng Data 38:458–460
Reverchon E, Della Porta G, Senatore F (1995a) Supercritical CO2 extraction and fractionation of
lavender essential oil and waxes. J Agr Food Chem 43:1654–1658
Reverchon E, Taddeo R, Della Porta G (1995b) Extraction of sage oil by supercritical CO2:
Influence of some process parameters. J Supercrit Fluids 8:302–309
Reverchon E, Daghero J, Marrone C, Mattea M, Poletto M (1999) Supercritical fractional
extraction of fennel seed oil and essential oil: Experiments and mathematical modeling. Ind
Eng Chem Res 38:3069–3075
Richter M, Sovová H (1993) The solubility of 2 monoterpenes in supercritical carbon-dioxide.
Fluid Phase Equilib 85:285–300
Rodrigues VM, Rosa PTV, Marques MOM, Petenate AJ, Meireles MAA (2003) Supercritical
extraction of essential oil from aniseed (Pimpinella anisum L.) using CO2: Solubility, kinetics,
and composition data. J Agric Food Chem 51:1518–1523
Roy BC, Goto M, Hirose T (1996) Extraction of ginger oil with supercritical carbon dioxide:
Experiments and modeling. Ind Eng Chem Res 35:607–612
Ruetsch L, Daghero J, Mattea M (2003) Supercritical extraction of solid matrices. Model formu-
lation and experiments. Lat Am Appl Res 33:103–107
Salimi A, Fatemi S, Zakizadeh Nei Nei H, Safaralie A (2008) Mathematical modeling of supercriti-
cal extraction of valerenic acid from Valeriana officinalis L. Chem Eng Technol 31:1470–1480
Sanders N (1993) Food legislation and the scope for increased use of near-critical fluid extraction
operations in the food, flavouring and pharmaceutical industries. In: King MB, Bott TR (eds)
Extraction of natural products using near-critical solvents. Blackie Academic & Professional,
London, UK, pp 34–49
Sefidkon F, Dabiri M, Mirmostafa SA (2004) The composition of Thymus serpyllum L. oil.
J Essent Oil Res 16:184–185
Serrato-Valenti G, Bisio A, Cornara L, Ciarallo G (1997) Structural and histochemical investiga-
tion of the glandular trichomes of Salvia aurea L. leaves, and chemical analysis of the essential
oil. Ann Bot 79:329–336
Simandi B, Deák A, Rónyai E, Yanxiang G, Veress T, Lemberkovics E, Then M, Saa-Kiss A,
Vámos-Falusi Z (1999) Supercritical carbon dioxide extraction and fractionation of fennel oil.
Simões-Pires CA, Debenedetti S, Spegazzini E, Mentz LA, Matzenbacher NI, Limberger RP,
Henriques A (2005) Investigation of the essential oil from eight species of Baccharis
belonging to sect. Caulopterae (Asteraceae, Astereae): A taxonomic approach. Plant Syst
Evol 253:23–32
Skerget M, Knez Z (2001) Modelling high pressure extraction processes. Comput Chem Eng
25:879–886
Sousa EMBD, Chiavone-Filho O, Moreno MT, Silva DN, Marques MOM, Meireles MAA (2002)
Experimental results for the extraction of essential oil from Lippia sidoides Cham. using
pressurized carbon dioxide. Braz J Chem Eng 19:229–241
Sousa EMBD, Martı́nez J, Chiavone-Filho O, Rosa PTV, Domingos T, Meireles MAA (2005)
Extraction of volatile oil from Croton zehntneri Pax et Hoff with pressurized CO2: Solubility,
composition and kinetics. J Food Eng 69:325–333
Souza AT, Corazza ML, Cardozo-Filho L, Guirardello R, Meireles MAA (2004) Phase equilib-
rium measurements for the system clove (Eugenia caryophyllus) oil + CO2. J Chem Eng Data
49:352–356
Sovová H (1994) Rate of the vegetable oil extraction with supercritical CO2.1. Modeling of
extraction curves. Chem Eng Sci 49:409–414
Sovová H (2005) Mathematical model for supercritical fluid extraction of natural products and
extraction curve evaluation. J Supercrit Fluids 33:35–52
Sovová H, Jež J (1994) Solubility of menthol in supercritical carbon-dioxide. J Chem Eng Data
39:840–841
Sovová H, Komers R, Kucera J, Jež J (1994a) Supercritical carbon dioxide extraction of caraway
essential oil. Chem Eng Sci 49:2499–2505
Sovová H, Kucera J, Jez J (1994b) Rate of vegetable oil extraction with supercritical CO2. 2.
Extraction of grape oil. Chem Eng Sci 49:415–420
Sovová H, Stateva RP, Galushko AA (2001) Essential oils from seeds: Solubility of limonene in
supercritical CO2 and how it is affected by fatty oil. J Supercrit Fluids 20:113–129
Sovová H, Stateva RP, Galushko AA (2007) High-pressure equilibrium of menthol + CO2.
Spricigo CB, Pinto LT, Bolzan A, Novais AF (1999) Extraction of essential oil and lipids from
nutmeg by liquid carbon dioxide. J Supercrit Fluids 15:253–259
Spricigo CB, Bolzan A, Pinto LT (2001) Mathematical modeling of nutmeg essential oil extraction
by liquid carbon dioxide. Lat Am Appl Res 31:397–401
Stahl E, Quirin K-W, Gerard D (1988) Dense gases for extraction and refining. Springer-Verlag,
Berlin, Germany
Stamenić M, Zizovic I, Orlović A, Skala D (2008) Mathematical modelling of essential SFE on the
micro-scale. Classification of plant material. J Supercrit Fluids 46:285–292
Stassi A, Schiraldi A (1994) Solubility of vegetable cuticular waxes in supercritical CO2 iso-
thermal calorimetry investigations. Thermochim Acta 246:417–425
Štastová J, Jež J, Bartlová M, Sovová H (1996) Rate of vegetable oil extraction with supercritical
CO2. 3. Extraction from sea buckthorn. Chem Eng Sci 51:4347–4352
Steffani E, Atti-Santos AC, Atti-Serafini L, Pinto LT (2006) Extraction of ho-sho (Cinnamomum
camphora Nees and Eberm var. Linaloolifera fujita) essential oil with supercritical CO2:
Experiments and modeling. Braz J Chem Eng 23:259–266
St€uber F, Vázquez AM, Larrayoz MA, Recasens F (1996) Supercritical fluid extraction of packed
beds: External mass transfer in upflow and downflow operation. Ind Eng Chem Res
35:3618–3628
St€uber F, Julien S, Recasens F (1997) Internal mass transfer in sintered metallic pellets filled with
supercritical fluid. Chem Eng Sci 52:3527–3542
Stull DR (1947) Vapor pressures of pure substances organic compounds. Ind Eng Chem
39:517–540
Svoboda K, Svoboda T (2000) Secretory structures of aromatic and medicinal plants. A review and
atlas of micrographs. Microscopix Publications, Knighton, UK
Takeuchi TM, Leal PF, Favareto R, Cardozo-Filho L, Corazza ML, Rosa PTV, Meireles MAA
(2008) Study of the phase equilibrium formed inside the flash tank used at the separation step of
a supercritical fluid extraction unit. J Supercrit Fluids 43:447–459
Tan CS, Liou DC (1989) Axial dispersion of supercritical carbon dioxide in packed beds. Ind Eng
Chem Res 28:1246–1250
Tan CS, Liang SK, Liou DC (1988) Fluid-solid mass transfer in a supercritical fluid extractor.
Chem Eng J 38:17–22
Teixeira de Souza A, Benazzia T, Boer Grings M, Cabral V, da Silva E, Cardozo-Filho L, Ceva
Antunes O (2008) Supercritical extraction process and phase equilibrium of candeia (Ere-
manthus erythropappus) oil using supercritical carbon dioxide. J Supercrit Fluids
47:182–187
Temelli F, O’Connell JP, Chen CS, Braddock RJ (1990) Thermodynamic analysis of supercritical
carbon dioxide extraction of terpenes from cold-pressed orange oil. Ind Eng Chem Res
29:618–624
The Good Scents Company (2009a) http://www.thegoodscentscompany.com/data/rw1060851.
html
The Good Scents Company (2009b) http://www.thegoodscentscompany.com/data/rw1005091.
html
Tufeu R, Subra P, Plateaux C (1993) Contribution to the experimental determination of the phase
diagrams of some (carbon dioxide + a terpene) mixtures. J Chem Thermodyn 25:1219–1228
Uquiche E, Huerta E, Sandoval A, del Valle JM. Effect of boldo (Peumus boldus M.) pretreatment
on the kinetics of supercritical CO2 essential oil extraction. J Food Eng (submitted)
Vargas RMF, Cassel E, Gomes GMF, Longhi LGS, Atti-Serafini L, Atti-Santos AC (2006)
Supercritical extraction of carqueja essential oil: Experiments and modeling. Braz J Chem
Eng 23:375–382
Vieira de Melo SAB, Pallado P, Guarise GB, Bertucco A (1999) High-pressure vapor-liquid
equilibrium data for binary and ternary systems formed by supercritical CO2, limonene and
linalool. Braz J Chem Eng 16:7–17
Villermaux J (1987) Chemical engineering approach to dynamic modelling of linear chromatog-
raphy: A flexible method for representing complex phenomena from simple concepts.
J Chromatogr A 406:11–26
Wagner Z, Pavlicek J (1993) Vapour-liquid equilibrium in the carbon dioxide: p-cymene system at
high pressure. Fluid Phase Equilib 90:135–141
Wakao N, Kaguei S (1982) Heat and mass transfer in packed beds. Gordon & Breach, New York
Wakao N, Smith JM (1962) Diffusion in catalyst pellets. Chem Eng Sci 17:825–834
Xing H, Yang Y, Su B, Huang M, Ren Q (2003) Solubility of artemisinin in supercritical carbon
dioxide. J Chem Eng Data 48:330–332
Yu D (1998) Solute pulse dispersion in soil columns: a comparison of supercritical CO2, gaseous
and aqueous systems. Ph.D. thesis, University of California at Los Angeles, CA
Zekovic Z, Lepojevic Z, Tolic A (2001) Modeling of the thyme-supercritical carbon dioxide

extraction system. I. The influence of carbon dioxide flow rate and grinding degree of thyme.
Sep Sci Technol 36:2459–3472
Zetzl C, Brunner G, Meireles MAA (2003) Standardized low-cost batch SFE-units for university
education and comparative research. In: Brunner G, Kikic I, Perrut M (eds) Proceedings of the
6th international symposium on supercritical fluids. Institut National Polytechnique de Lor-
raine, Lorraine, France, pp 577–585
Zizovic I, Stamenić M, Orlović A, Skala D (2005) Supercritical carbon dioxide essential oil
extraction of Lamiaceae family species: Mathematical modelling on the micro-scale and
process optimization. Chem Eng Sci 60:6747–6756
Zizovic I, Stamenić M, Ivanović J, Orlović A, Ristić M, Djordjević S, Petrović SD, Skala D
(2007a) Supercritical carbon dioxide extraction of sesquiterpenes from valerian root. J Super-
crit Fluids 43:249–258
Zizovic I, Stamenić M, Orlović A, Skala D (2007b) Supercritical carbon dioxide extraction of
essential oils from plants with secretory ducts: Mathematical modelling on the micro-scale.
Zizovic IT, Stamenić MD, Orlović AM, Skala DU (2007c) Supercritical carbon-dioxide extraction
of essential oils and mathematical modelling on the micro-scale. In: Berton LP (ed) Chemical
engineering research trends. Nova Science Publishers, New York, pp 221–249
Part III
Water Management in Food
.
Chapter 18
Glass Transitions: Opportunities and Challenges
Yrj€o H. Roos and Nattiga Silalai
18.1 Introduction
Glass transition is a well-known transformation of the solid-fluid states of noncrystalline

solids and liquids. In food systems the glass transition is often a property of
carbohydrates and proteins, as observed in numerous food materials (White and
Cakebread 1966; Slade and Levine 1995). Glass transition results in various
physicochemical and thermal phenomena associated with the solid–liquid transfor-
mation of a supercooled liquid. A number of noncrystalline polymers, as well as
sugar melts, exhibit glass transition at approximately 100–150 C below their
equilibrium melting temperature (Roos 1993). The glass transition of low water
and frozen foods has been recognized as one of the most important factors affecting
food properties, processing characteristics, and shelf life (Roos 1995a). Glass
transition contributes to numerous engineering properties of foods, their structure
and texture, and reaction kinetics. It is also useful in understanding sensory proper-
ties, including hardness, crispiness, softness, and flavor release (Slade and Levine
1995; Roos 1995a; Roudaut et al. 2004). Knowledge of molecular mobility and
changes in mechanical and physical properties over the glass transition can be used
to control the characteristics of food solids, for example, in drying, freeze-drying,
freezing, and extrusion. An overview of the importance of the glass transition in
food engineering, including selected food systems and processes, as well as chal-
lenges in using glass transition data in the control of food properties, food proces-
sing, and storage stability, will be highlighted in this review.
Y.H. Roos (*) and N. Silalai

School of Food and Nutritional Sciences, University College Cork, Cork, Ireland
e-mail: yrjo.roos@ucc.ie
474 Y.H. Roos and N. Silalai
18.1.1 Confectionary
The physical state of sugars is important in manufacturing and control of quality

changes during storage of confectionery. Formation of noncrystalline solid sugar
structures is the basic technology underpinning the manufacturing of hard sugar
candies, which are often produced from sugar melts or concentrates by rapidly
cooling the product after heating to above melting temperatures. High solids
concentration and rapid cooling enhance the formation of solid, transparent, and
brittle, glassy materials. White and Cakebread (1966) stated that food materials
containing amorphous sugars, such as hard candies, were stable at temperatures
below their glass transition. Cotton candy is another example of sugar glass
technology. Cotton candy is made by melting crystalline sucrose, which is spun
under rapid cooling and dehydration into an amorphous (glassy) solid. At tempera-
tures exceeding the glass transition temperature, Tg, of sucrose, the spun structure
collapses as a result of crystallization of the sugar. A hard lump of sucrose crystals
imbedded in a partially glassy structure may form, making the shelf life of the
product very short (Labuza and Labuza 2004).
18.1.2 Frozen Foods
Crystallization and recrystallization of amorphous food components are time-

dependent phase transitions. Crystallization and recrystallization processes are
affected by the physical state of food solids and may be controlled by the glass
transition. One example is lactose crystallization in frozen desserts and ice cream.
Such solute crystallization is a result of freeze concentration of the solutes and
subsequent supersaturation. At sufficiently low temperatures the solutes, along with
some unfrozen water, vitrify into the solid (glassy) state reducing diffusion and
causing crystallization processes to cease. Crystallization and recrystallization
processes, however, take place in frozen foods above the onset temperature of
ice melting in the maximally freeze-concentrated solute matrix (Roos and Karel
1991b, d; Roos 1995a). As lactose is one the least soluble sugars, lactose crystalli-
zation often occurs in ice cream produced without stabilizers. Lactose crystalliza-
tion causes an undesirable coarse and sandy mouthfeel, while ice recrystallization
may be observed from large, coarse ice crystals and flaky ice structure.
Various authors (White and Cakebread 1966; Livney et al. 1995; Roos 1995a)
have reported that rates of crystallization processes increase above the glass
transition. At temperatures approaching the glass transition, crystallization and
recrystallization rates reduce as a result of limited molecular mobility (Roudaut
et al. 2004) when the material is transformed into the solid (glassy) state. At
temperature below the glass transition, crystallization of food components as
well as their recrystallization is unlikely because of the solid-like structure
and absence of translational diffusion of molecules (Roos 1995a; Hartel 1996;
18 Glass Transitions: Opportunities and Challenges 475
Roudaut et al. 2004). Recrystallization phenomena are typical of frozen foods,

which often may be stored at fairly high temperatures or exposed to significant
temperature fluctuations after manufacturing. Recrystallization of ice is a time-
dependent process and generally characterized by an increase in the size of ice
crystals with storage time (Roos 1995a; Hagiwara et al. 2005). Roos (1995a) pointed
out that rates of ice formation and recrystallization could be controlled by using the
temperature difference of the Tg (referred to as TTg) as the rate-defining factor.
The recrystallization rate is strongly reduced below the glass transition of the freeze-
concentrated, unfrozen system (Carrington et al. 1996). Thus, the glass transition
temperatures, Tg, of the freeze-concentrated solutes and unfrozen water content at
temperatures during frozen storage are the most important composition-dependent
factors, which then can be used to control the extent of ice formation and ice
recrystallization in frozen foods at any given storage temperature.
18.1.3 Cereal Foods
Starch is a mix of carbohydrate polymers and the main component of cereal foods.
Starch is present also in legume seeds and tuber plants. The two starch components,
amylose and amylopectin, may exist in crystalline, partially crystalline, and amor-
phous states (Slade and Levine 1991). In native starches, amylose may exist as a
noncrystalline component but amylopectin often exhibits partial crystallinity (Roos
1995a). Water in amorphous parts of starch components acts as a plasticizer and
decreases the Tg. Glass transition, melting temperatures, and water content are the
most important parameters characterizing the state and physical properties of starch
components over a wide temperature range in cereals processing and product
storage.
Phase transitions associated with gelatinization and loss of native structure in
granular starches can define and explain differences in the physical properties of
starches and their behavior in food products (Lund 1989). Levine and Slade (1990)
pointed out that starch gelatinization is a nonequilibrium melting process that
occurs during the heating of starch in the presence of water. Retrogradation of
starch is a temperature-, water-content-, and time-dependent crystallization phe-
nomenon and occurs after cooling of gelatinized starch. The rate of retrogradation
depends on the presence and ratio of amylopectin and amylose, and the molecular
weight of these starch components (Roos 1995a; Jouppila and Roos 1997). After
cooling of gelatinized starch, amylose crystallization occurs rapidly, while crystal-
lization of amylopectin may occur during storage of gelatinized starch and cereal
foods (Roos 1995a; Ronda and Roos 2008).
Several studies (Le Meste et al. 1992; Jouppila and Roos 1997; Ronda and Roos
2008) indicated that the textural characteristics of cereal foods at different tem-
peratures and water contents could be explained in terms of temperature with
respect to Tg. Starch retrogradation, including a basic crystallization process of
gelatinized starch components, can be detected by DSC from observed physical
change in a starch gel or paste, and by X-ray diffraction methods (Roos 2007a).
Glass transition can control texture- and stability-related phenomena such as
gelatinization and retrogradation, which may proceed over the temperature range
Tg < T < Tm (Biliaderis 1992; Yoshimura et al. 1996). Crystallization in gelati-
nized starch (or starch retrogradation) and the loss of water during storage were
found to contribute to aging and bread staling above the glass transition (Roos et al.
1996; Champion et al. 2000; Ronda and Roos 2008). Moreover, crispness, which is
affected by water content and glass transition, is an essential factor in achieving
acceptable quality among numerous cereal and snack foods (Roudaut et al. 2002).
Many cereal-based foods, such as breakfast cereals, wafers, and biscuits, have a
crispy or crunchy texture when consumed at low water contents (below 10% water);
however, crispness may be lost as the water content increases due to water sorption
and subsequent water plasticization of amorphous structures (Nicholls et al. 1995;
Roos et al. 1998; Hochstetter et al. 2006). A critical water activity (aw) at which
crispness is lost was found to be specific according to each material. A change
resulting in loss of crispness often occurred around 0.35–0.50 aw (Roos 1993; Roos
et al. 1996). When the critical water content or water activity is exceeded, the Tg of
the material occurs below the ambient temperature (Roos 1993).
18.1.4 Food Powders and Dehydrated Foods
Food powders containing amorphous carbohydrates, such as lactose in dairy powders,

may show changes in physical properties, for example, crystallization, clumping,
stickiness, and caking during processing, handling, and storage (Levine and Slade
1986). Stickiness and caking phenomena are related to the collapse phenomena
occurring in amorphous solids, and often result from increased temperature or
exposure of powders to high-humidity conditions.
Caking of powders may be considered as a collapse phenomenon, which occurs
when particle surfaces in contact form permanent aggregates and harden causing a
loss of free-flowing properties in powder particles. Boonyai et al. (2004) explained
the difference between stickiness and caking of food powders. They stated that
stickiness is an instantaneous process but caking occurs over a time period. The
common cause of stickiness is plasticization of particle surfaces, which allows a
sufficient decrease of surface viscosity for the formation of liquid bridges between
particles (Downton et al. 1982; Roos 1995a). These phenomena in powders, with
amorphous components contributing to stickiness, can be characterized as time-
dependent surface flow properties that are controlled by properties of the amor-
phous components and their flow at temperatures above the glass transition (Roos
1995a). Several researchers (Roos and Karel 1991a, b; Chuy and Labuza 1994;
Roos 1995a; Bhandari and Howes 1999) have illustrated possible relationships
among stickiness, caking, and glass transition. The food polymer science approach
has increased understanding of the state of food materials and its use to control or
prevent undesirable physical changes, including powder stability (Bhandari and
Howes 1999). Amorphous solids exist in a supercooled liquid (rubbery state) above
the glass transition at which plasticization-dependent viscosity decreases occur
rapidly. Glass transition enhances molecular mobility and flow, leading to sticki-
ness and crystallization problems in low-water and frozen foods (Roudaut et al.
2004). Therefore, glass transition is an important and useful concept in observation
of parameters, such as temperature and water content for the control or reduction of
liquid-like properties of powder components during processing and storage.
Collapse during freeze-drying and collapse of freeze-dried matrices may
adversely affect material properties during dehydration and storage, respectively.
Collapse is a result of viscous flow of amorphous materials or their components,
causing loss of structure, reduction of pore size, and shrinkage. Such collapse is
associated with undesirable appearance and loss of texture, and volatile sub-
stances (Karel and Flink 1973; Levi and Karel 1995). It occurs above glass
transition, the rate of which depends on temperature and water content. Levi
and Karel (1995) found that rates of collapse were strongly dependent on the
temperature above the glass transition. Increasing water content depressed the Tg
and increased the T–Tg at the observed temperature. The flow and rate of collapse
above Tg were functions of temperature difference (TTg), which could be
modeled by the Williams-Landel-Ferry (WLF) relationship (Roos 1995a). This
suggested that the Tg was an applicable parameter for the prediction and control of
collapse during freeze-drying and storage of amorphous foods.
18.2 Glass Transition: Opportunities
An increase in molecular mobility above the glass transition enhances flow of

amorphous structures and results in time-dependent physical changes (Roos and
Karel 1991a; Bhandari et al. 1997; Bhandari and Howes 1999). This occurs above a
critical temperature, or plasticization level, which often refers to the glass transition
temperature, Tg, or water activity and water content, respectively (Roos 2007b).
Glassy states may form in numerous food processes involving cooling of highly
supercooled liquids or removal of water by dehydration or freezing (Fig. 18.1).
Glass formation or vitrification of food solids may occur in various glassy, non-
equilibrium (solid glassy state) solid structures. These may have different physical
appearances, such as freeze-dried or spray dried structures, and also may exhibit
varying thermodynamic states (Roos 2008).
18.2.1 Freezing and Freeze-Drying
Ice crystallization and collapse of structure in partially frozen systems are con-
trolled by the molecular mobility of food components in freeze-concentrated,
unfrozen water-solute systems (Levi and Karel 1995; Roos 1995a). These changes
are time-dependent and reduce in rate as the maximally freeze-concentrated state is
Equilibrium
Liquid
SOLUTION
Time-dependent
Phenomena
D
n
eh
io Pl
at
yd
as
n
ilz
tio
ra
tia
ub Non-equilibrium
ra
tio
l iz
So
tu
Liquid at
n
Sa
Equilibrium io
n
Solid
(Pressure) Cooling
CRYSTAL Crystallization RUBBER Heating GLASS
g
lin Non-equilibrium
He
oo Solid
Heating
Cooling
Sl
at
C
ow
in
id
g
g
tin
ap
Co
ea
R
ol
H
in
g
Equilibrium MELT
Liquid
Fig. 18.1 Changes in physical state of amorphous materials around glass transition and time-
dependent characteristics (Data are from Roos and Karel 1991a)
approached. The glassy structures of the unfrozen water-solids phases that are
formed in maximally freeze-concentrated systems support the structure against
flow and collapse, but the structure does not allow further ice formation (Roos
and Karel 1991d; Roos 2007a). Freezing and frozen storage conditions, however,
affect the size of ice crystals, which could probably be further manipulated by the
control of freeze-concentration and ice formation.
Ice formation in foods results in a freeze-concentration of solids and a gradually
decreasing freezing temperature as the solute concentration of the unfrozen phase
increases. The freeze-concentrated, amorphous unfrozen phase contains unfrozen
water and solids; it provides a continuous phase for the dispersed ice crystals and
possibly lipids or other nondissolved solid components (Roos 1995a). The glass
transition temperature of the maximally freeze-concentrated unfrozen phase (Tg’) is
independent of the solute concentration prior to freezing and often corresponds with a
solute concentration (Cg’) of 80% (w/w) (Roos et al. 1996). At temperatures below
the Tg’ the amorphous unfrozen phase is vitrified and exists as a glassy solid (Roos
and Karel 1991d), whereas ice dissolution at temperatures above the onset tempera-
ture of ice melting, Tm’, decreases the viscosity of the unfrozen phase, which shows
viscous flow under gravity. Hence, frozen systems at temperatures above the Tm’
cannot support the solid structure (Fig. 18.2) and show collapse or shrinkage (Bhandari
and Howes 1999; Le Meste et al. 2002; Alves-Filho and Roos 2006). Stability of
frozen foods can be accomplished by manipulation of food composition to increase
the Tm’ or by using storage temperatures lower than the Tm’ (Hartel 1996). How-
ever, compositional changes in food formulations are not always possible.
Fig. 18.2 Viscous flow of freeze-concentrated liquid phase in freeze-drying
The control of ice crystal growth in frozen foods and ice cream has been
postulated by the Tm’ control (Goff et al. 1993). They suggested that polysaccha-
rides provided resistance to thermal deformation and increased subzero viscosity
above the Tm’. Hence, ice crystal size and the rate of growth of ice crystals in frozen
foods and ice cream containing polysaccharides were smaller than in frozen foods
and ice cream without added polysaccharides. Ice formation ceases as a result of
kinetic limitations for crystal growth below Tg’; some frozen foods are stable
against deteriorative changes because of the glassy state of the unfrozen phase
(Levine and Slade 1986; Roos and Karel 1991d). However, some reactions, such as
oxidation of sensitive components may accelerate as a result of maximum freeze-
concentration of foods. Studies of frozen food stability are relatively few and
therefore it requires substantial further attention by researchers. It appears, how-
ever, that the control of the glass transition of the unfrozen phase and ice melting is
fundamental to freezing and frozen storage as well as freeze-drying.
18.2.2 Spray Drying
Particle stickiness and subsequent deposition of semi-dried particles on dryer

surfaces is regarded as one of the most typical problems in spray drying. During
drying of sugar-rich foods, such as fruit juices, honey, and some starch derivatives
(glucose syrup/maltodextrins with higher dextrose equivalent values), their struc-
tures may remain as syrup-like liquids; such particles stick on the drier surfaces
leading to lower product yields and operating problems (Bhandari et al. 1997).
In industrial applications, sticky-point temperature curves may be generated and
used to develop optimal drying operations to minimize stickiness problems.
Stickiness is a time-dependent phenomenon that is related to structural transfor-

mations and flow of amorphous solids around the glass transition (Roos and Karel
1991c; Roos 1995a). Measured Tg values have correlated well with sticky-point
measurements and the glass transition concept has provided a better fundamental
understanding and predictability of stickiness (Roos et al. 1996; Adhikari et al.
2001, 2005). Application of glass transition data is extremely useful in the design of
dehydration equipment to minimize the stickiness of sugar-rich food solids (Truong
et al. 2005a, 2005b). Truong et al. (2005a, b) suggested that the glass transition
approach could be used to reduce stickiness because the difference between the
outlet air temperature and the Tg of the solids of the final product (TTg) was a
measurable, stickiness controlling property of the particles in the spray-drying
process. The Tg was influenced by solids composition, which affects the physical
changes in foods (Ozkan et al. 2002; Fitzpatrick et al. 2007a, b; Nijidam and Langrish
2006; Haque and Roos 2006). Amorphous low-molecular weight substances such as
fructose and glucose have a low Tg. High-molecular weight components, such as
polysaccharides and proteins, increase the Tg of mixtures containing sugars (Roos
and Karel 1991c; Haque and Roos 2004; Shrestha et al. 2007); they also affect
the powder characteristics during spray drying as shown in (Fig. 18.3) (Haque and
Roos 2006).
Maltodextrins are widely used as food components to reduce stickiness in sugar-
rich foods because of their Tg and viscosity increasing property in mixtures with
sugars (Bhandari et al. 1997). Addition of low dextrose equivalent (DE) maltodex-
trins decreases stickiness and caking during drying and storage (Adhikari et al.
2004; Langrish et al. 2007), and provides an excellent opportunity to improve
processing and storage characteristics. The effect on mixture properties can be
predicted using Tg data. The stickiness behavior can be characterized by various
techniques (Lazar et al. 1956; Downton et al. 1982; Chuy and Labuza 1994;
Hennigs et al. 2001; Ozkan et al. 2002). The sticky points of powders were found
Fig. 18.3 Composition affects glass transition and dehydration characteristics using scanning
electron microscopy (SEM) (Further data can be found in Haque and Roos 2006)
to be associated with glass transition temperatures at 10–20 C above the Tg (Roos

and Karel 1991b; Roos 1995a; Hennigs et al. 2001; Ozmen and Langrish 2002;
Adhikari et al. 2005). This indicated that the glass transition is a useful parameter to
control dehydration and powder characteristics, including stickiness, caking, and
agglomeration during drying processes and product storage.
18.2.3 Extrusion
Extrusion has become a well-established and widely used food processing method.
Extrusion processes often involve conversion of solid ingredients to a viscous but
homogeneous mixture followed by formation of dense, solid, or expanded highly
porous structures. Biodegradable packaging and edible films may also be manufac-
tured using extrusion processes. The use of extrusion in food applications is often
based on empirical knowledge of solids behavior and structure formation. However,
the materials science understanding of foods and knowledge of their phase and state
transitions are fundamental in the control of extrusion processes and material
behavior in structure formation and storage (Slade and Levine 1991, 1995; Roos
1995a). This development also is an opportunity to improve extrusion processes and
product quality, particularly in the development of novel extrusion applications.
The formation of the structure of several extruded foods and low-moisture foods,
such as snack foods and breakfast cereals, is influenced by temperature and water
content. The glass transition can be used to control their plasticization, gelatinization,
and the glass formation. In plasticization, increasing the temperature or the water
content results in a decrease of the Tg. Water plasticization also controls modulus and
tensile strength of extruded films produced by extrusion (Garcia et al. 2004; Roos
1995a; Hochstetter et al. 2006). One of the desired characteristics of extruded snacks
and breakfast cereals is a crispy texture. At the expansion of plasticized foods, starchy
solids are rapidly dehydrated and the solids vitrify, i.e., the structure of the food solids
becomes a glass, forming thin membranes with brittle and crispy characteristics.
Details on plasticization and structure formation in extrusion provide new opportu-
nities for controlling the process and the texture and deterioration of extruded
products since water plasticization and oxygen permeability can be manipulated
during storage (Roos 1995a; Roos et al. 1996; Hochstetter et al. 2006). For example,
Garcia et al. (2004) used glass transition data to control the thermal and water
plasticization of extruded meat products and to achieve improved water permeability
properties of films.
18.2.4 Encapsulation
Encapsulation processes are used to entrap food ingredients, enzymes, cells, or

other components in structure-forming, encapsulant materials. Applications of
encapsulation have been increased in the food industry, since this process often
aims at improved protection from heat and loss, or otherwise stabilizes sensitive
components in food processing and storage. Dispersed droplets and particles in
encapsulant matrices show reduced reaction rates and are often protected from
surrounding reactants as a result of decreased diffusion through the glassy structure
below the glass transition (Roos 1995a). Both temperature and water plasticization
of encapsulant matrices may increase structural changes and release of encap-
sulated substances. Above the glass transition, the amorphous materials exhibit
viscous flow and structures of food materials often change, which can result in
collapse and release of encapsulated components.
The formulation of encapsulant matrices and control of glass formation may be
used to improve encapsulation and stability of food products. For example, highly
polymeric compounds such as proteins and hydrocolloids have high glass transition
temperatures. These high molecular weight miscible compounds with miscible
lower molecular weight components increase the glass transition temperature of
the mixture and stabilize the glassy state against higher temperatures and water
contents (Roos 2008). There is, however, very little information about the misci-
bility of carbohydrate polymers and proteins in food systems, as well as on how the
use of various processes affects glass formation. Food structures also have hetero-
geneities that can greatly affect encapsulation, so their role in stability and shelf-life
control needs to be investigated further.
18.3 Glass Transition: Challenges
Changes in physicochemical and physical properties of foods are likely to occur

around glass transition, the point at which amorphous solids convert to liquids and
exhibit viscous flow. The glass transition results in enhanced molecular mobility,
which is associated with decreasing relaxation times and a lower viscosity. There
have been numerous studies of enthalpy relaxations around the glass transition
(Haque et al. 2006). These relaxations are examples of time-dependent glass
formation and present challenges in the understanding of glassy structures formed
in food processing.
Food processes, such as freezing and freeze-drying, spray drying or extrusion,
may produce structures specific to the product composition, process, and processing
parameters. These factors, together with thermal and water effects in storage,
contribute to the properties and quality of food systems at the time of consumption.
Understanding the state of glassy food materials and their time-dependent char-
acteristics may require rigorous experimental studies, as such research would
clarify the role of glassy states in the control of reaction rates and stabilization of
highly sensitive food systems.
It is generally agreed that endothermic enthalpy relaxations increase with
increasing aging time and temperature below the glass transition (Haque et al.
2006). Relaxations may also indicate exothermic enthalphy changes as translational
mobility of molecules appears around the glass transition (Roos 1995a, 2008).
These relaxations show qualitative and quantitative information on the enthalpy
state of molecules within a glass. As translational mobility of molecules appears
around the glass transition materials may either require (endothermal relaxation) or
release (exothermal relaxation) heat as they respond to the increasing temperature.
Molecular mobility and the time-dependent nature of the glass transition can also be
observed from mechanical and dielectric properties of materials (Champion et al.
2000; Le Meste et al. 2002; Roudaut et al. 2004; Roos 2008). Both dielectric and
dynamic mechanical properties show the a-relaxation around the glass transition.
The a-relaxation can be observed using dielectric analysis (DEA) or dynamic-
mechanical analysis (DMA) (Laaksonen and Labuza 2001; Laaksonen and Roos
2001; Royall et al. 2005; Hochstetter et al. 2006; Roos 2008). The a-relaxation
temperatures obtained are frequency-dependent, indicating the nonequilibrium and
time-dependent nature of the system. These relaxations may be used to describe
mechanical properties of food materials, such as stickiness and caking of amor-
phous powders. Silalai et al. (2009a) found that the a-relaxation of dairy powders
shifted to higher temperatures with increasing frequencies of DEA and DMA
measurements (Fig. 18.4). The frequency-dependence of the a-relaxation followed
the Arrhenius relationship (Talja and Roos 2001), and a frequency corresponding to
the glass transition temperature could be identified. The glass transition measured
by DSC and the a-relaxation temperatures (Ta) determined by DMA and DEA
correlated closely with the sticky-point temperatures obtained using an empirical
measurement (Silalai et al. 2009b).
Solids composition also affects structural formation and the quality of foods
during processing and storage. For example, powders containing high amounts of
lactose exhibited high adhesion of powder particles with increasing temperature
and water content as a result of surface plasticization. Surface plasticization caused
Fig. 18.4 Time-dependent characteristics and relaxations of the amorphous state

a decrease in sticky-point temperatures (Silalai et al. 2009b). Although many

researchers (Roos and Karel 1991b; Chuy and Labuza 1994; Roos 2002) have
demonstrated that stickiness and caking of dairy powders are related to glass
transition, the time-dependent flow and contact time measurements still remain as
challenging areas in glass transition research. However, composition, plasticiza-
tion, and temperature are the main causes leading to a sufficient decrease in surface
viscosity and higher adhesion and cohesion of powder particles. It was also shown
by Downton et al. (1982) that powders were free-flowing particles at temperatures
below glass transition; they became plasticized at above these temperatures, at
decreasing viscosity in the range 107 to 109 Pa s.
Crystallization of amorphous components, particularly sugars in low-water and
frozen foods, and starch components in gelatinized starch-containing foods, is often
the most dramatic change in food systems. Numerous examples are available on
lactose crystallization in dairy powders (Jouppila et al. 1997) and ice cream (Hartel
1996), as well as on starch retrogradation, including amylopectin crystallization
(Jouppila et al. 1997). Crystallization is affected by water content and other
components in low-water and frozen foods. Haque and Roos (2004) found that
the Tg of spray-dried lactose–protein mixtures increased as a result of added protein
components. The rate of crystallization of amorphous lactose was also reduced in
spray-dried lactose–protein mixtures (Fig. 18.5). The data suggested that various
proteins may affect lactose crystallization properties differently; studies of crystal-
lization with salts showed that salts may further complicate crystallization (Omar
and Roos 2007). These findings with information on water plasticization and
component miscibility are fundamental in describing the spray-drying character-
istics and product stability of sugar-containing food solids. The practical applica-
tions of these data are numerous but may require a systematic materials science
approach in product formulation and equipment design.
(exo) TcrP
Heat flow (Wg–1)
Tg Lactose Tcr
Lactose/WPI (3:1)
Lactose/Album in (3:1)
Lactose/Gelatin (3:1)
Lactose/Na-caseinate (3:1)
10 30 50 70 90 110 130
Temperatue (°C)
Fig. 18.5 Glass transition and crystallization temperatures of amorphous lactose in the presence
of various proteins (From Haque and Roos 2004)
The effects of food components in mixtures on dielectric and mechanical proper-

ties around the glass transition have been reported in several studies (Kalichevsky
et al. 1993; Nilolaidis and Labuza 1996; Talja and Roos 2001; Laaksonen and
Labuza 2001; Laaksonen et al. 2002; Laaksonen and Roos 2001, 2003; Hashimoto
et al. 2003). According to Silalai et al. (2009a, 2009b), dairy powders with
increased protein content exhibited higher Tg values as measured by DSC. This
was in agreement with increased Ta values as determined by DEA and DMA. There
were also other changes in mechanical and dielectric relaxations. Dielectric and
mechanical properties of powders with high lactose contents changed considerably
around the glass transition as compared to powders with high protein contents
(Figs. 18.6 and 18.7).
In frozen and freeze-drying systems, solids composition affects the glass transition
of the maximally freeze-concentrated unfrozen phase, Tg’, and onset temperature of
ice melting, Tm’ (Roos and Karel 1991d; Goff et al. 1993). The a-relaxations of
frozen systems were also affected by composition and frequency, as observed from
dielectric and dynamic-mechanical relaxations (Laaksonen and Roos 2001; Laaksonen
et al. 2002). These data are of significant importance in the control of freezing and
freeze-drying properties of food and other biological materials. The challenges in
the control of freezing properties, frozen food stability, as well as freeze-drying can
be addressed by understanding the state and phase transitions of freeze-concen-
trated systems. For example, polymeric materials in mixtures containing sugar
lower the Tg’ but increase the Tm’. This increases the broadness of the transition
Fig. 18.6 Dielectric thermal analysis (DEA) thermograms showing the effect of carbohydrate–
protein composition on dielectric relaxations of nonfat milk solids at water activity of 0.332.
The carbohydrate–protein ratios for MPC-3, MPC-25, SMP, and MPC-55 were 15:74, 27:59,
40:48, and 56:32 (% of total solids, w/w), respectively
Fig. 18.7 Dynamic mechanical analysis (DMA) thermograms showing the effect of carbohydrate–
protein composition on dielectric relaxations of nonfat milk solids at water activity of 0.332. The
carbohydrate–protein ratios for MPC-3, MPC-25, SMP, and MPC-55 were 15:74, 27:59, 40:48,
and 56:32 (% of total solids, w/w), respectively
Fig. 18.8 Differential scanning calorimetry (DSC) thermograms showing glass transition and ice
melting in sucrose and protein–starch systems (Data are from Singh and Roos 2005)
temperature range and suggests that ice formation is affected by the composition
and diffusional characteristics of the unfrozen matrix. It seems that polymeric
compounds, such as gelatin and corn starch, as shown in (Fig. 18.8), inhibit ice
formation resulting in a higher unfrozen water content with corresponding lowering
of the Tg’ and increase of the Tm’ (onset of melting occurred at a higher temperature
corresponding to a lower ice content and higher melting temperature) (Singh and
Roos 2005). Although it appears that compositional changes affect food properties
in freeze-concentrated systems, their complex structure makes experiments and
collection of low temperature data complicated. However, further analysis of
heterogeneities and time-dependent changes in composition and microstructure in
frozen systems will be useful in the design of freezing processes and formulation of
frozen foods with improved quality and stability.
18.4 Conclusion
Significant progress has been made in understanding the role of glass transitions in
food materials during food processing and storage. Glass transitions in foods have
been shown to affect the flow properties of concentrated systems and, therefore,
challenges exist in processes such as dehydration, extrusion, and freezing. The glass
transitions and water plasticization behavior of food solids may affect structural
changes, crystallization processes, as well as deteriorative reactions, which often
limit the shelf life of low-water and frozen foods. Novel thermal analytical systems
and other techniques are crucial to further understanding the complex nature of food
systems, as well as the translation of knowledge of glass transition properties to the
benefit of processing equipment and food product design for meeting processing
needs and establishing the highest quality standards.
References
Adhikari BP, Howes T, Bhandari BR, Troung V (2001) Stickiness in foods: a review of mechan-
isms and test methods. Int J Food Prop 4(1):1–33
Adhikari B, Howes T, Bhandari BR, Troung V (2004) Effect of addition of maltodextrin on drying
kinetics and stickiness of sugar and acid-rich foods during convective drying: experiments and
modeling. J Food Eng 62:53–68
Adhikari B, Howes T, Lecomte D, Bhandari BR (2005) A glass transition temperature approach
for the prediction of the surface stickiness of a drying droplet during spray drying. Powder
Technol 149:168–179
Alves-Filho O, Roos YH (2006) Advances in multi-purpose drying operations with phase and state
transitions. Dry Technol 24:383–396
Bhandari BR, Datta N, Howes T (1997) Problems associated with spray drying of sugar-rich foods.
Dry Technol 15(2):671–684
Bhandari BR, Howes T (1999) Implication of glass transition for the drying and stability of dried
foods. J Food Eng 40:71–79
Biliaderis CG (1992) Structures and phase transitions of starch in food systems. Food Technol
46(6):98–109
Boonyai P, Bhandari BR, Howes T (2004) Stickiness measurement techniques for food powders: a
review. Powder Technol 145:34–46
Carrington AK, Goff HD, Stanley DW (1996) Structure and stability of the glassy state in rapidly
and slowly cooled carbohydrate solutions. Food Res Int 29(2):207–213
Champion D, Le Meste M, Simatos D (2000) Towards on improved understanding of glass
transition and relaxations in foods: molecular mobility in the glass transition range. Food Sci
Technol 11:41–55
Chuy LE, Labuza TP (1994) Caking and stickiness of dairy-based food powders as related to glass
transition. J Food Sci 59(1):43–46
Downton GE, Flores-Luna JL, King CJ (1982) Mechanism of stickiness in hygroscopic amorphous
powders. Indus Eng Chem Fund 21:447–451
Fitzpatrick JJ, Hodnett M, Twomey M, Cerqueira PSM, O’Flynn J, Roos YH (2007a) Glass
transition and the flowability and caking of powders containing amorphous lactose. Powder
Technol 178(2):119–28
Fitzpatrick JJ, Barry K, Cerqueira PSM, Iqbal T, O’Neill J, Roos YH (2007b) Effect of composi-
tion and storage conditions on the flow ability of dairy powders. Int Dairy J 17:383–392
Garcia RA, Onwulata CI, Ashby RD (2004) Water plasticization of extruded material made from
meat and bone meal and sodium caseinate. J Agr Food Chem 52:3776–3779
Goff HD, Caldwell KB, Stanley DW (1993) The influence of polysaccharides on the glass
transition in frozen sucrose solutions and ice cream. J Dairy Sci 76:1268–1277
Hagiwara T, Mao J, Suzuki T, Takai R (2005) Ice recrystallization in sucrose solutions stored in a
temperature range of 21 C to 50 C. Food Sci Technol Res 11(4):407–411
Haque K, Roos YH (2004) Water plasticization and crystallization of lactose in spray-dried
lactose/protein mixtures. J Food Sci 69(1):23–29
Haque K, Roos YH (2006) Differences in the physical state and thermal behaviour of spray-dried and
freeze-dried lactose and lactose/protein mixtures. Innovative Food Sci Emerg Technol 7:62–73
Haque K, Kawai K, Suzuki T (2006) Glass transition and enthalpy relaxation of amorphous lactose
glass. Carbohyd Res 341:1884–1889
Hartel RW (1996) Ice crystallization during manufacture of ice cream. Trends Food Sci Technol
7:315–321
Hashimoto T, Hagiwara T, Suzuki T, Takai R (2003) Study on glass transition Katsuobushi (boiled
and dried bonito fish stick) by differential scanning calorimetry and dynamic mechanical
analysis. Fisheries Sci 69:1290–1297
Hennigs C, Kockel TK, Langrish TAG (2001) New measurements of the sticky behavior of skim
milk powders. Dry Technol 19(3&4):471–484
Hochstetter A, Talja RA, Helen HJ, Hyvonen L, Jouppila K (2006) Properties of gluten-based
sheet produced by twin-screw extruder. Lebensm Wiss Technol 39:893–901
Jouppila K, Roos YH (1997) The physical state of amorphous corn starch and its impact on
crystallization. Carbohyd Polym 32:95–104
Jouppila K, Kansikas J, Roos YH (1997) Glass transition, water plasticization, and lactose
crystallization in skim milk powder. J Dairy Sci 80(12):3152–3160
Kalichevsky MT, Blanshard JMV, Tokarczuk PF (1993) Effect of water content and sugars on the
glass transition of casein and sodium caseinate. Int J Food Sci Tech 28:139–151
Karel M, Flink JM (1973) Influence of frozen state reactions on freeze-dried foods. J Agr Food
Chem 21:16–21
Laaksonen TJ, Labuza TP (2001) Effects of moisture, sucrose, NaCl, and arabinoxylan on
relaxation in wheat dough as measured by DMTA. Int J Food Prop 4(2):311–325
Laaksonen TJ, Roos YH (2001) Dielectric relaxations of frozen wheat doughs containing sucrose,
NaCl, ascorbic acid and their mixtures. J Cereal Sci 33:331–340
Laaksonen TJ, Roos YH (2003) Water sorption and dielectric relaxations of wheat doughs
(containing sucrose, NaCl, ascorbic acid, and their mixtures). J Cereal Sci 37:319–326
Laaksonen TJ, Jouppila K, Roos YH (2002) Effects of arabinoxylans on thermal behaviour of
frozen wheat doughs as measured by DSC, DMA, and DEA. J Food Sci 67(1):223–230
Labuza TP, Labuza PS (2004) Influence of temperature and relative humidity on the physical
states of cotton candy. J Food Process Pres 28(4):274–287
Langrish TAG, Chan WC, Kota K (2007) Comparison of maltodextrin and skim milk wall
deposition rates in a pilot-scale spray dryer. Powder Technol 179:84–89
Lazar ME, Brown AH, Smith GS, Wong FF, Linquist FE (1956) Experimental production of
tomato powder by spray drying. Food Technol 10:129–134
Le Meste M, Huang VT, Panama J, Anderson G, Lentz R (1992) Glass transition of bread. Cereal
Food World 37:164–267
Le Meste M, Champion D, Roudaut G, Blond G, Simatos D (2002) Glass transition and food
technology: a critical appraisal. J Food Sci 67(7):2444–2458
Levi G, Karel M (1995) Volumetric shrinkage (collapse) in freeze-dried carbohydrates above their
glass transition temperature. Food Res Int 28(2):145–151
Levine H, Slade L (1986) A polymer physicochemical approach to the study of commercial starch
hydrolysis products (SHP’s). Carbohyd Polym 6:213–244
Levine H, Slade L (1990) Influences of the glassy and rubbery states on the thermal, mechanical
and structural properties of doughs and baked products. In: Faridi H, Faubion JM (eds) Dough
rheology and baked product texture. AVI, New York, pp 157–330
Livney YD, Donhowe DP, Hartel RW (1995) Influence of temperature on crystallization of lactose
in ice-cream. Int J Food Sci Tech 30(3):311–320
Lund DB (1989) Starch gelatinization. In: Singh RP, Medina AG (eds) Food properties and
computer-aided engineering of food processing systems. Kluwer, The Netherlands, pp 299–311
Nicholls RJ, Appelqvist IAM, Davies AP, Ingman SJ, Lillford PJ (1995) Glass transitions and the
fracture behaviour of gluten and starches within the glassy state. J Cereal Sci 21:25–36
Nijidam JJ, Langrish TAG (2006) The effect of surface composition on the functional properties of
milk powders. J Food Eng 77:919–925
Nikolaidis A, Labuza TP (1996) Glass transition state diagram of a baked cracker and its
relationship to gluten. J Food Sci 61(4):803–806
Omar AME, Roos YH (2007) Glass transition and crystallization behaviour of freeze-dried
lactose-salt mixtures. Lebensm Wiss Technol 40(3):536–543
Ozkan N, Walisinghe N, Chen XD (2002) Characterization of stickiness and cake formation in
whole milk and skim milk powders. J Food Eng 55:293–303
Ozmen L, Langrish TAG (2002) Comparison of glass transition temperature and sticky point
temperature for skim milk powder. Dry Technol 20(6):1177–1192
Ronda F, Roos YH (2008) Gelatinization and freeze-concentration effects on recrystallization in
corn and potato starch gels. Carbohyd Res 343:903–911
Roos YH (1995a) Phase transitions in foods. Academic, San Diego
Roos YH (1995b) Characterization of food polymers using state diagrams. J Food Eng 24:339–360
Roos YH (2002) Importance of glass transition and water activity to spray drying and stability of
dairy powders. Lait 82:475–484
Roos YH (2004) Phase and state transitions in dehydration of biomaterials and foods. In:
Mujumdar AS (ed) Dehydration of products of biological origin. Science Publishers, Enfield,
NH, pp 3–22
Roos YH (2007a) Phase transitions and transformations in food systems. In: Heldman DR,
Lund DB (eds) Handbook of food engineering, 2nd edn. CRC Press, New York, pp 284–352
Roos YH (2007b) Water activity and the glass transition. In: Barbosa-Cánovas GV, Fontana AJ,
Schmidt SJ, Labuza TP (eds) Water activity in foods. Fundamentals and applications. IFT
Press/Blackwell, Ames, IA, pp 29–45
Roos YH (2008) The glassy state. In: Aguilera JM, Lifford PJ (eds) Food materials science.
Principles and practice. Springer, New York, pp 67–82
Roos YH, Karel M (1991a) Applying state diagrams to food processing and development. Food
Technol 12:66, 68–71, 107
Roos YH, Karel M (1991b) Plasticizing effect of water on thermal behavior and crystallization of
amorphous food models. J Food Sci 56(1):38–43
Roos YH, Karel M (1991c) Phase transitions of mixtures of amorphous polysaccharides and
sugars. Biotechnol Progr 7:49–53
Roos YH, Karel M (1991d) Phase transitions of amorphous sucrose and frozen sucrose solutions.
J Food Sci 56:266–267
Roos YH, Karel M, Kokini JL (1996) Glass transitions in low moisture and frozen foods: effect on
shelf life and quality. Food Technol 50:95–108
Roos YH, Roininen K, Jouppila K, Tuorila H (1998) Glass transition and water plasticization
effects on crispness of a snack food extrudate. Int J Food Prop 1:163–180
Rosenberg M, Sheu TY (1996) Microencapsulation of volatiles by spray drying in whey protein-
based wall systems. Int Dairy J 6(3):273–284
Roudaut G, Dacremont C, Vallès Pàmies B, Colas B, Le Meste M (2002) Crispness: a critical
review on sensory and material science approaches. Trends Food Sci Technol 13:217–227
Roudaut G, Simatos D, Champion D, Contreras-Lopez E, Le Meste M (2004) Molecular mobility
around the glass transition temperature: a mini review. Innovative Food Sci Emerg Technol
5:127–134
Royall PG, Huang C, Tang SJ, Duncan J, Van-de-Velde G, Brown MB (2005) The development of
DMA for the detection of amorphous content in pharmaceutical powdered materials. Int J
Pharm 301:181–191
Shrestha AK, Howes T, Adhikari BP, Wood BJ, Bhandar BR (2007) Effect of protein concentra-
tion on the surface composition, water sorption and glass transition temperature of spray-dried
skim milk powders. Food Chem 104:1436–1444
Silalai N, Hogan S, O’Callaghan D, Roos YH (2009a) Dielectric relaxations and stickiness of dairy
powders influenced by glass transition. In: Fisher P, Pollard M, Windhab EJ (eds) Proceedings
of the 5th International Symposium on Food Rheology and Structure – ISFRS 2009, June
15–18, 2009, Z€urich, Switzerland, pp 428–430
Silalai N, Hogan S, O’Callaghan D, Roos YH (2009b) The control of dairy powder characteristics
using alpha-relaxation and glass transition data. In: Devahastin S (ed) Drying 2009 – Proceed-
ings of the 6th Asia-Pacific Drying Conference (ADC2009), October 19–21, 2009, Bangkok,
Thailand, pp 61–64
Singh KJ, Roos YH (2005) Frozen state transitions of sucrose-protein cornstarch mixtures. J Food
Sci Technol 70(3):E198–204
Singh KJ, Roos YH (2006) State transitions and freeze concentration in trehalose-protein-corn-
starch mixtures. LWT 39:930–938
Slade L, Levine H (1991) Beyond water activity: Recent advances based on an alternative
approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30
(2–3):115–360
Slade L, Levine H (1995) Water and the glass transition-dependence of the glass transition on
composition and chemical structure: Special implications for flour functionality in cookie
baking. J Food Eng 24(4):431–509
Talja RA, Roos YH (2001) Phase and state transition effects on dielectric, mechanical, and thermal
properties of polyols. Thermochim Acta 380:109–121
Truong V, Bhandari BR, Howes T (2005a) Optimization of co-current spray drying process of
sugar-rich foods. Part I-Moisture and glass transition temperature profile during drying. J Food
Eng 71:55–65
Truong V, Bhandari BR, Howes T (2005b) Optimization of co-current spray drying process of
sugar-rich foods. Part II-Optimization of spray drying process based on glass transition
concept. J Food Eng 71:66–72
White GW, Cakebread SH (1966) The glassy state in certain sugar containing food products.
J Food Technol 1:73–82
Yoshimura M, Takaya T, Nishinari K (1996) Effects of konjacglucomannan on the gelatinization
and retrogradation of corn starch as determined by rheology and differential scanning calorim-
etry. J Agr Food Chem 44:2970–2976
Chapter 19
Caking of Water-Soluble Amorphous
and Crystalline Food Powders
Stefan Palzer and Karl Sommer
19.1 Introduction
Time consolidation during storage, often called caking, is a major problem when
handling water-soluble powders in a humid and hot environment. Caking requires
grinding to transform the powder cake into single particles or agglomerates. Some-
times even whole silos are blocked by caked powder and the powder can only
be removed manually using mining techniques. Furthermore, the quality of powder
a products is negatively affected by caking during storage or transport. Instead of a
free-flowing powder, large powder lumps are delivered to the consumer. To reduce
the processing costs caused by the required grinding or sieving and to increase the
quality of the end product, such caking has to be avoided.
Mainly powders that contain a major amount of water-soluble amorphous substances
are sensitive to caking. Examples of such amorphous substances are: powdered vegeta-
ble-, yeast-, and meat extracts, hydrolyzed fish proteins, and hydrolyzed plant proteins,
flavoring powders, various food additives such as starch hydrolysates, amorphous
lactose, amorphous dextrose, amorphous sucrose, maltose, and amorphous organic
acids such as citric- and ascorbic acid and spray-dried dairy powders (Peleg and
Mannheim 1977; Wallack and King 1988; Aguilera et al. 1993; Aguilera et al. 1995).
Occasionally caking of fine water-soluble crystalline powders is also observed.
Sodium chloride, crystalline citric acid, urea, and sucrose are examples of crystal-
line powders that consolidate during storage at high relative humidity of the
surrounding air. Figure 19.1 shows scanning electron microscopic pictures of
different caked powders containing major amounts of amorphous substances.
Some inter-particle bridges are highlighted by white circles.
S. Palzer (*)
Nestle Product Technology Centre York, Nestec York Ltd, P.O. Box 204/Haxby Road, York
YO91 1XY, UK
e-mail: stefan.palzer@rdyo.nestle.com
K. Sommer
Department of Process Engineering of Disperse Systems, Technical University of Munich, Am
Forum 2, 85354 Freising, Germany
492 S. Palzer and K. Sommer
80 μm 15 μm 30 μm
Spray-dried skim milk powder containing crystalline lactose Spray-dried whole milk powder
400 μm 50 μm 20 μm
Spray-dried tomato powder manufactured on a spray/belt drier Spray-dried dextrose syrup DE 21
Fig. 19.1 Scanning electron microscopic pictures of caked powders (bridges highlighted by white
circles)
As demonstrated in the current study, the chemical composition, supra-molecular

structure, and microstructure as well as external factors such as stress, humidity, and
temperature are decisive in determining the caking mechanism and the caking kinetic.
19.2 Scientific Background
19.2.1 Supra-molecular Structure and Material Properties
The increase in adhesion forces between single particles causing caking depends on
the material properties and the storage or processing conditions. The material
properties are a function of the environmental conditions. They also strongly
depend on the supra-molecular structure of the material. The relation between
supra-molecular structure and material properties has been discussed extensively
in previously published scientific articles (e.g., Palzer 2006, 2009a, b). Generally,
one can distinguish between polar and nonpolar materials with amorphous, crystal-
line, and semi-crystalline supra-molecular structures. Amorphous water-soluble
structures composed of polar molecules are able to incorporate significant amounts
of water. Depending on the vapor pressure or water activity they absorb water
molecules, which are then stored inside the free volume in the molecular matrix.
The absorbed water increases the mobility of the molecules composing the amorphous
19 Caking of Water-Soluble Amorphous and Crystalline Food Powders 493
matrix and thus decreases the viscosity of the amorphous material. Furthermore,
amorphous matrices are thermo-sensitive. Their viscosity decreases markedly with
increasing temperature. Accordingly, a decrease in viscosity, which might cause an
increase in adhesion forces, can be equally achieved by increasing temperature or
moisture content. Crystalline substances, which might contain crystal water, absorb
practically no water molecules until they dissolve, while exceeding their critical
relative humidity. Thus, below the critical relative humidity, variations in air
humidity do not affect the mechanical properties of the material. Furthermore,
temperature variations below the melting temperature do not change the mechani-
cal material properties of crystalline structures significantly. When exceeding the
melting temperature the material liquefies completely at constant temperature.
In addition delisquence phenomena might further increase the moisture sensitiv-
ity of particle mixtures.
19.2.2 Adhesion Mechanisms
Increasing adhesion forces between particles are a prerequisite for any caking
process. Adhesion can be either linked to the development of material bridges or
increasing of Van der Waals forces between neighboring particles. For most caking
processes the following adhesion mechanisms are relevant:
l Increasing Van der Waals forces due to visco-elastic or plastic deformation
l Liquid bridges with low viscosity containing a low concentration of mono-,
oligo- or polymers or ions
l Liquid bridges with medium or high viscosity containing a high concentration of
dissolved oligomers or polymers
l Amorphous solid bridges
l Crystalline solid bridges
There is no clear transition from high viscous liquid bridges to amorphous solid
bridges. Generally, an amorphous solid is discussed if the viscosity exceeds
1010–1011 Pas. Depending on the nature of the particles, these listed adhesion
mechanisms can contribute to increasing the adhesion force F holding neighboring
particles together. The tensile strength of manufactured agglomerates depends on
the strength of these adhesion forces acting between primary particles. According to
Rumpf (1970), the tensile strength of an agglomerate composed of spheres with a
diameter a can be estimated through (19.1):
1e p
st ¼ KF K (19.1)
p a2 e
where F is the adhesion force between two spheres, e is the porosity of the agglo-
merate, and K represents the coordination number of the primary particles of the
agglomerate. For spherical particles K is approximately equal to p/e (Rumpf 1970).
Adhesion between food particles can be linked to increasing Van der Waals
forces, which are electrostatic forces generated by temporary load shifts in neigh-
boring molecules. The Van der Waals force FvdW between two deformed particles
can be estimated according to Lifshitz (1956) and Hamaker (1937). Except for
plastically deformable particles, most caking processes are not caused by increasing
Van der Waals forces. However, plastic or visco-elastic deformation of particles
leading to increasing number of contact points, increasing contact area, and
decreasing distance between neighboring particle surfaces might enhance caking
caused by sintering of amorphous substances.
Caking might be also linked to dissolution and re-crystallization of crystalline
material. The required water can be liberated in a powder bulk by condensation of
humidity from the air or re-crystallization of humid amorphous material. If suffi-
cient water is liberated, liquid bridges between the particles are generated. Dissolu-
tion of low molecular weight substances in the liquid changes the viscosity of the
liquid bridges only slightly. Thus, the strength of liquid bridges containing mainly
dissolved low-molecular substances or a low concentration of polymers mainly
depends on capillary forces. The capillary pressure pc in a liquid bridge can be
calculated based on the surface tension g, the wetting angle Y, the radius s of
the meniscus of the liquid bridge, and the radius x of the cross-section area of the
bridge. Assuming that the particles are spherical, the tensile strength st of a humid
powder bulk or agglomerate, stabilized by liquid bridges exhibiting a low viscosity,
can be estimated according to Rumpf (1958) through (19.2):
1e 2g
st ¼ S C cos y (19.2)
e a
where S is the saturation of the porous agglomerate with liquid (ratio of liquid
volume versus total void volume within the agglomerate), C is a constant parameter
(monodisperse spheres: C ¼ 6) (Rumpf 1958), e is the porosity of the agglomerate,
g is the surface tension of the liquid, a is the diameter of the primary particles, and
Y represents the contact angle between the liquid and the solid substance.
The adhesion forces caused by aqueous liquid bridges containing dissolved
substances can be further increased by drying. Finally, the liquid bridge is trans-
formed into a solid bridge (Fig. 19.2). Depending on the drying velocity, solid
bridges with amorphous or crystalline structures are obtained.
Fig. 19.2 Formation of liquid and solid bridges between crystalline particles (e.g., due to storage
at varying relative humidity RH or delisquence)
Changes in the tensile strength st of agglomerates due to dehydration of liquid

bridges can be predicted according to Rumpf (1958) by using (19.3):
Vdiss
st ¼ ð 1 eÞ s s (19.3)
Vagglo
where Vdiss is the dissolved solid volume within the entire agglomerate, Vagglo is the
volume of the agglomerate, ss represents the tensile strength of the solid substance
building the bridge after drying, and e is the agglomerate porosity. The tensile
strength of liquid bridges containing polymer molecules also depends on the strain
rate applied during the measurement.
For many powders sintering is the main mechanism leading to time consolida-
tion of the powder bulk. Like low-viscosity liquid droplets, amorphous particles
tend to adopt a spherical shape. By bringing two amorphous particles into contact
with each other, they can be considered as one new single particle. To minimize the
free surface area of this newly created particle, molecules are transported to the
contact point between the two primary particles (Fig. 19.3). Such a process is called
sintering. The driving force behind the sinter process is the difference between the
capillary pressure at the contact point between the particles and the Laplace
pressure in the volume of the two initial particles.
Since the capillary pressure is directly linked to the vapor pressure in the
continuous gas phase surrounding the particles, a gradient of the vapor pressure
exists across the particle. Based on these local differences between capillary and
vapor pressure and depending on how big the substances vapor pressure of the
substance is, different molecular transport mechanisms are observed. The relevant
literature (Kuczynski 1949; Schatt 1992; Wagner 1997) distinguishes between
transport of molecules through the vapor phase, surface diffusion, volume diffu-
sion, and grain-boundary diffusion. Substances having a medium or high vapor
pressure can be transported via the gas phase surrounding the particles. The sinter-
ing substance evaporates at convex particle surfaces, which exhibit a high vapor
pressure. Following the evaporation, molecules are transported by diffusion
through the gas phase to the concave meniscus of the bridge between two particles.
Such concave structures are characterized by a lower vapor pressure. The material
condenses at the contact point between the particles and builds a sinter bridge
s
a particle diameter
x diameter of the sinter bridge
s radius of curvature between the particles
a x volume diffusion
pc1 vapour pressure above the particles
pc2 vapour pressure above the sinter bridge
surface diffusion
condensation sublimation
pc2 < 0 pc1 < 0
Fig. 19.3 Sintering mechanisms

between them. This evaporation and condensation of substances can contribute

significantly to the growth of the sinter bridge if the vapor pressure of the sintering
substance is high enough. Since the vapor pressure of most major food components
is comparably low, a molecular transport through the gas phase seems to be less
relevant for sintering of amorphous food particles. Molecules can also be trans-
ported by surface or volume diffusion. Such diffusion of molecules inside the
particles, sometimes also referred to as viscous flow, seems to be the most relevant
mechanism for sintering of water-soluble amorphous food particles.
The kinetic of such transport processes strongly depends on the diffusion
coefficient, which is a function of viscosity. The viscosity of amorphous solids
depends on the temperature as well as on the plasticizer content of the material.
Thus, sintering of amorphous water-soluble particles can be controlled by adjusting
the temperature and/or the water content of the system.
Different sinter phases are usually distinguished according to the observed
structural changes in the particle package (Fig. 19.4). In an initial phase, when
the sinter bridges are still very thin, the particles adhere to each other. Macroscop-
ically, a reduction of the powder’s flowability (often described as “stickiness”) is
observed. With advancing sinter time the diameter of the sinter bridge grows, which
leads to increasing adhesion forces between the particles. Open pores are gradually
eliminated and the particles lose their structural integrity. The powder and particle
structure collapses, while the system is progressively transformed from a powder
into a foam-like system where the air is enclosed in an amorphous melt.
For modeling the sintering of organic particles by viscous flow, Frenkel (1945)
published equation (19.4). Assuming the spherical particles have a diameter a, the
diameter of the sinter bridge x can be calculated depending on the sinter time t, the
surface tension g, and the viscosity Z:
x2 1 g t
¼ (19.4)
a 6 a Z
Alternatively, the sinter kinetic can be predicted according to Rumpf et al.

(1976). Assuming a punctual contact between the particles and neglecting changes
in particle geometry during sintering, they modeled the viscous flow based on the
Navier–Stokes equations, where Ft represents the force with which the particles are
pressed together.
Fig. 19.4 Different phases of the sintering process

x2
4 g 2 Ft t
¼ þ (19.5)
a 5 a 5 p a2 Z
Both equation (19.4) and (19.5) are only valid for the initial phase of the sinter
process. For nonspherical particles, the relevant radius a must be estimated based
on the curvature radius at the contact point between the particles. A prerequisite for
a sinter process is molecular contact between the particles. A moderate force (e.g.,
caused by weight of particles) pressing the particles together ensures a continuous
contact between the particles during sintering. Except for pressure agglomeration
the impact of the force Ft on the sinter kinetic can be neglected. Variations in surface
tension are mostly limited. Hence, viscosity (which might vary by magnitudes)
(Palzer 2006, 2009a) governs the sinter kinetic. As mentioned before, the viscosity
of water-soluble amorphous substances depends on their moisture content w and the
temperature T. Combining the Gordon-Taylor equation with the Williams-Landel-
Ferry equation (Palzer 2009a), and considering the fact that due to sorption or
desorption processes the humidity of the amorphous substance might change as a
function of time, equation (19.6) is obtained:
x2 ð
tmax CðTTg ðtÞÞ
4 g 2 Ft 1
¼ þ 10 BþðTTg ðtÞÞ
dt (19.6)
a 5 a 5 p a2 g
t¼0
Where tmax is the time available for the sinter process. By combining (19.1) with
(19.6) the tensile strength of sintered agglomerates can be estimated.
ð
tmax CðTTg ðtÞÞ
ð1=eÞ p 4 g 2 Ft
st ¼ ss ðT; w; oÞ þ Z1 10 BþðTTg ðtÞÞ dt (19.7)
e 5 a 5 p a2 g
t¼0
However, in practice it remains difficult to predict the tensile strength of sintered

agglomerates because of the geometrical diversity of primary particles and the
visco-elastic nature of the built sinter bridges. According to Downtown et al. (1982)
a significant adhesion force F between the particles is observed if the diameter ratio
x/a exceeds a value of 0.01. Wallack and King (1988) reported that a value x/a of
0.1 is required for adhesion.
19.2.3 Caking Processes
Several of the mentioned adhesion mechanisms can lead to a time consolidation of

powders (Tomas 1996). As demonstrated in the current study, mainly sintering is
responsible for the observed caking of amorphous food powders. For caking caused
by sintering of amorphous particles, several stages can be distinguished. In the
initial phase of the process the powder starts to become sticky and particles adhere
to each other. They form brittle lumps, and with progressing sintering, a mechani-
cally stable powder cake is obtained. During a later phase of sintering, particles
missing a stabilizing inner structure (built of non-dissolving substances) lose their
structure and shape. The powder structure collapses, open pores in the particle
package disappear, and finally a highly viscous, amorphous melt is obtained. The
shape of some particles (e.g., dehydrated vegetable or fruit pieces) is preserved by
an insoluble matrix (e.g., composed of cellulose, fat, or proteins) and, thus, no
structural collapse of such powders is observed.
When water-soluble amorphous substances are stored in a high-humidity
environment they absorb water and thus their viscosity decreases. Although
sintering is accelerated due to the decreasing viscosity, the tensile strength of
the growing sinter bridges decreases with increasing water content. Due to the
progressing absorption of water, the strength of the powder cake first increases
and then in a later phase of the process decreases again. At low moisture content,
or in the early phase of the sintering process, the diameter of the sinter bridge
limits the stability of the particle cake. The growing diameter of the sinter bridges
leads to an increasing strength of the cake. With progressing moisture absorption
the viscosity of the bridge decreases and thus the strength of the powder cake
decreases again. Finally, the powder structure collapses and a pasty mass is
obtained. The viscosity of amorphous substances is also reduced at elevated
temperatures, leading to an accelerated sintering of the particles and a steadily
increasing strength of the powder cake.
Crystalline water-soluble powders behave differently when exposed to high
humidity or temperature. If their critical humidity is temporarily exceeded they
will partially dissolve. The formed low viscosity liquid bridges are often rather
fragile. However, a stable powder cake can be formed if the absorbed water
evaporates again enabling a re-crystallization of the dissolved substance. Caking
of crystalline powders stored in closed containers can thus be induced by conden-
sation and evaporation of moisture due to temperature variations. The condensate
builds liquid bridges between the particles. After a short time span the liquid
forming these bridges is saturated with the dissolving substance. If the condensed
water is later evaporated at increasing temperature stable bridges between the
particles are formed.
19.2.4 Powder Flowability and Stress States
Caking is characterized by significantly decreasing powder flowability. Increasing

adhesion forces lead to decreasing powder flowability. Yielding of a powder bulk
strongly depends on such adhesion forces, and the magnitude and direction of the
different stresses to which it is exposed. The stress state in a powder can be
analyzed by establishing force balances around a prism-shaped powder element.
The powder element is exposed to normal stresses s and shear stresses t. The position
σx τxy
y
σy
α
τα
τyx
σα
Fig. 19.5 Normal and shear stresses acting on a prism-shaped powder element
of the plane separating the powder prism from the remaining powder volume is
characterized by the angle a. The normal and shear stresses acting on or in this
plane are sa and ta (Fig. 19.5).
A necessary condition for yielding of the powder is that the two shear stresses
acting on the powder prism are equal to each other (txy ¼ tyx). By establishing
the force balances for the prism-shaped powder element shown in Fig. 19.5, the
equations for calculating the shear stress ta and the normal stress sa acting in the
inclined shear plane can be deducted ((19.8) and (19.9)), where dh is the depth of
the element and dl represents the length of the shear plane.
sa dl dh ¼ sy sin a dh sin a dl þ sx cos a dh cos dl

þtyx cos a dh sin a dl þ txy sin a dh cos a dl
, sa ¼ sy sin2 a þ sx cos2 a þ 2 txy sin a cos a (19.8)
sx þ sy sx sy
, sa ¼ þ cos 2a þ txy sin 2a
2 2
ta dl dh ¼ sy dl sin a dh cos a sx dl cos a dh sin a

txy dl sin a dh sin a þ txy dl cos a dh cos a (19.9)
s x þ sy
, ta ¼ sin 2a þ txy cos 2a
2
There are two values for the angle a for which the shear stress ta in the shear
plane equals zero.

1 2 txy 1 2 txy p
a¼ arctan or a¼ arctan
2 sx sy 2 sx s y 2 (19.10)
) ta ¼ 0
The normal stresses in the shear plane corresponding to these two angles are
called main consolidation stress s1 and s2. Knowing s1, s2, and the inclination
angle a allows calculation of the normal stress and the shear stress acting on the
powder element.
s1 þ s2 s1 s2
sxy ¼ cos 2a
2 2 (19.11)
s 1 s2
txy ¼ sin 2a
2
By varying the inclination angle a, a number of corresponding s/t combinations

are obtained. The so-called Mohr’s circle can be constructed by plotting these s/t
pairs in a s/t diagram. The Mohr’s stress circle represents all possible stress states
of a bulk solid element.
If s2 equals zero, the mono-axial stress state is given. In the mono-axial stress
state, the bulk solid element is exposed to stress in only one direction. For the
mono-axial and the bi-axial stress states, the Mohr circles are shown in Fig. 19.6. In
the bi-axial stress state both main consolidation stresses are larger than zero.
Complementary to the stress state, in a confined powder volume the flow proper-
ties of the powder have to be analyzed. Only the combination of possible stress
states with the material properties will allow a prediction of the powder’s behavior.
The powder flowability is characterized by the combination of normal and shear
stress, which leads to yielding of the powder bulk. The powder is always pre-
sheared to provide a constant bulk density and then sheared under varying normal
loads to identify the shear stress required for yielding. The obtained s/t combina-
tions build the so-called yield locus, which can be plotted in a s/t diagram
(Fig. 19.7). According to the theory of Mohr and Coulomb the obtained s/t
combinations form a straight line in the s/t diagram running through the origin:
t ¼ tan j s þ C (19.12)
where je is the angle for stationary flow and j is the flow angle of incipient flow.
Following the theory of the mono-axial and the bi-axial stress states, two Mohr
circles can be constructed to which the yield locus is tangential. The pre-shear point
a b
t t
ta ta
sa sa s2
a F a Fx
s,t txy s,t
s1 s1
sy
a a
s s2 s Fy Fy
s1=st sx s1
s1 s1
sa s2
sa F Fx
ta ta
Fig. 19.6 Mohr’s circle for the mono-axial (a) and bi-axial (b) stress states
t t
stationary flow pre-shear point
incipient flow
yield locus
large Mohr´s circle
small Mohr´s circle
pre-shearing shearing pre-shearing shearing t s

unconfined yield strength σc
Fig. 19.7 Determining the yield locus based on results from shear tests
is located on the larger Mohr circle, which represents the bi-axial stress state. The
normal stress at which the smaller Mohr circle intersects with the vertical stress axis
is called unconfined yield strength sc. The unconfined yield strength is a measure of
the stability of the powder cake and can thus be used to quantify caking. The normal
stress at which the larger of the two Mohr circles intersects with the vertical stress
axis is called principle consolidation stress s1 (Fig. 19.7).
The flow factor ffc of a powder is calculated as the ratio between the principle
consolidation stress s1 and the unconfined yield stress sc. It characterizes the
powder’s flowability at the powder density achieved during pre-shearing.
s1
ffc ¼ (19.13)
sc
Based on the flow factor, Jenicke (1970) published the following scale for
grouping powders concerning their flowability.
ffc < 1 hardened
1 < ffc < 2 very cohesive
2 < ffc < 4 cohesive
4 < ffc < 10 easy flowing
ffc > 10 free flowing
19.3 Materials and Methods
19.3.1 Materials
The spray-dried dextrose syrup used in the reported trials has a dextrose equivalent
(DE) of 20–23, an initial water content of 4% (db) (water activity, aw ¼ 0.18), and a
mean particle diameter x50,3 of (150 25) mm. The yeast extract is characterized by
a mean particle diameter x50,3 of (101 34) mm. The sodium chloride has a mean
particle diameter x50,3 of (507 75) mm and a water content below 0.1% (db). The
sucrose used is purified (99% pure sucrose). The crystalline sucrose powder has a
particle diameter x50,3 of (643 66) mm and contains less than 0.1% water (db).
The spray-dried amorphous sucrose powder has a diameter x50,3 of (135 22) mm.
19.3.2 Method for Measuring Caking
The flowability of the powder sample is evaluated through ring-shear tests.

Figure 19.8a illustrates the principle of the used ring-shear tester and Fig. 19.8b
shows the shear-cells used for time consolidation trials.
For each test the sample is filled into a cell having the form of a ring. The cell
is then closed with a cover ring that has some baffles reaching into the powder.
This cover ring lying on the powder bulk and can be turned using a motor. The
torque required to turn the cover ring against the cell body is measured depending
on the normal stress acting on the ring cover. By taking the ring geometry into
consideration, the measured torque can be transformed into a shear stress. The
shear stress t is the stress required to initiate yielding the powder which is
exposed to a defined normal stress s. Each powder sample is sheared twice: by
pre-shearing, a constant density is assured for all samples, whereas subsequent
shearing delivers information about the powder’s flowability. For pre-shearing,
the sample is filled into the ring cell and the ring cover is loaded with a defined
weight exposing the powder to a defined normal stress. Then the ring cover is
turned until a constant torque is obtained. This torque can be transformed into the
shear stress required for stationary flow of the powder at a constant density. After
performing such pre-shearing the powder has a defined density. Now the normal
stress acting on the ring cover is reduced and the powder is sheared again. The
shear stress at which the powder starts yielding is measured. This procedure has
to be repeated by applying the same normal stress for pre-shearing and varying
the normal stress during shearing. The obtained s/t pairs are only valid
a b
Weights
normal load FN
Cell cover
powder sample Bar
shear cell
Screw
Rubber-Joint
weight
torque
Acryl Glass Cap
Shear Cell
Fig. 19.8 Principle of a ring shear tester (a) and cells used for time consolidation trials (b)
for the constant powder density achieved by the chosen normal stress used for
pre-shearing.
Time-consolidation trials using a ring shear tester are performed as follows: The
samples are filled in ring-shear cells. A pre-shearing of the samples provides a
defined powder density. Next, the shear cells are stored under controlled conditions
for a defined time. The cover ring of the cell lying on the powder carries a defined
weight during storage. Each cell has a housing in which the temperature and the
relative humidity of the air can be adjusted individually (Fig. 19.8b). Before and
after storage, the flowability of the powder sample stored in the ring cell is
measured in a shear tester (Fig. 19.8a). The observed increase in unconfined yield
strength and the calculated increase of the flow-factor are appropriate measures to
quantify the intensity of caking.
To quantify the caking intensity, the unconfined yield strength sc of the initial
sample and the stored powder has to be compared with each other. Either the
increase in unconfined yield strength sct or the decrease in the flow factor ffc can
be used for quantifying the degree of caking. Empirically, it has been found that
strong caking corresponds to an increase in unconfined yield strength of more than
500–1,000 Pa.
One of the major drawbacks of time consolidation experiments in ring-shear
cells is the hampered exchange of moisture between the air and the sample during
the consolidation phase. Contrary to such test conditions, in practice moisture
absorption and time consolidation are often parallel processes, which are difficult
to separate from each other.
19.4.1 Caking of Amorphous Water-Soluble Powders
In the current study caking of pure dextrose syrup (DE 21) was investigated. In the
first series of trials the sintering of two amorphous dextrose syrup particles (DE 21)
was monitored. Two particles were exposed at constant temperature of 25 C to
relative humidity of 70%. Both particles were placed on a silanized hydrophobic
glass plate and the relative humidity of the surrounding air was adjusted by using a
saturated salt solution. Figure 19.9 shows the water content of the particles depen-
dent on storage time and images of the growing sinter bridge developed between the
two particles.
The absorbed moisture leads to a decreasing viscosity of the amorphous matrix
and thus sintering is accelerated. A stable sinter bridge is obtained after 3–6 h of
storage. Finally, the two particles coalesce and form a single viscous droplet.
Such changes also happen in a bulk of amorphous water-soluble particles stored
at elevated humidity and temperature.
14
13
12 moisture content during microscops

moisture content w/% (wb)ll
storage at 30 °C / 70% rH heating + camera

11
10 T, aw
particles
9 saturated salt solution
8
silanised glass plate
7 progressing spheronisation of an agglomerate Experimental set-up for monitoring of

consisting of two primary particles particle spheronisation
6
4
0 5 10 15 20 25 30
time t/h
Fig. 19.9 Moisture content and visual appearance of two dextrose syrup particles (DE 21) stored
at 25 C and 70% RH as a function of storage time
250 µm 100 µm 60 µm
Initial dextrose syrup DE 21 Sintered dextrose syrup DE 21 Sintered dextrose syrup DE 21

T = 20°C, aw = 0.20, t=0 h T = 30°C, RH = 70%, t = 4 h T = 30°C, RH = 70%, t = 12 h
100 µm 200 µm 100 µm

Initial tomato powder Sintered tomato powder Sintered tomato powder
T = 20°C, aw = 0.15, t = 0 h T = 33°C, RH = 20%, t = 64 h T = 33°C, RH = 20%, t = 168 h
Fig. 19.10 Sintering of spray-dried tomato and spray-dried dextrose syrup (DE 21) powders
stored at 10–20 C above their glass transition temperatures
Figure 19.10 shows a dextrose syrup powder and a tomato powder, which were
both stored at 10–20 C above their glass transition temperatures. With progressing
storage time the dextrose syrup particles sinter together and finally the powder
completely loses its structure. The tomato powder particles also sinter together.
However, dehydrated cell walls, which are mainly composed of water insoluble
cellulose and hemicellulose, preserve their structure. Thus, no complete collapse of
the particle structure is observed.
As already discussed, the sinter bridge diameter can also be calculated using
equations (19.5) or (19.6). Figure 19.11 shows a comparison of the calculated sinter
bridge diameter (19.6) with the measured sinter bridge diameter between two
dextrose syrup particles stored at 30 C and 70% RH (see also Fig. 19.9).
The calculated kinetics of sintering is in good agreement with the measured
increase of the diameter of the sinter bridge. Despite the various assumptions and
simplifications made, equation (19.6) allows a satisfying prediction of the sinter
kinetic of amorphous water-soluble particles. Possibly, the accuracy of the predic-
tion can be further improved by taking into account a moisture gradient along the
particle diameter. Using the Navier–Stokes equations for a pair of sintering parti-
cles, the viscosity gradient corresponding to the moisture distribution inside the
particles has to be considered.
0.9
Ratio sinter bridge / particle diameter (x/a) / ..
0.8
0.7
0.6
0.5
Calculated x/a for a = 255 μm
0.4
Calculated x/a for a = 355 μm
0.3 Calculated x/a for a = 410 μm
Measured x/a for a = 260-280 μm

0.2
Contact radius Measured x/a for a = 350-360 μm
0.1 due to manually
initiated collision Measured x/a for a = 400-420 μm
0
0 20 40 60 80 100 120 140 160 180
time t / min
Fig. 19.11 Comparison between the calculated and measured sinter bridge/particle diameter ratio
for pairs of spray-dried dextrose syrup particles (DE 21) stored at 30 C and 75% RH (Tg ¼ 4 C;
gl,g ¼ 70 mNm, Z ¼ 106 Pas)
t = 0 min t = 2 min t = 4 min t = 6 min t = 8 min
Fig. 19.12 Collapse of a package of dextrose syrup particles exposed to increasing temperature
(DE 21; 10% moisture (wb); heated in an oil bath at oil temperature of 90 C)
2000
yeast extract aw=0.17 t=96h
1800 yeast extract aw=0.17 t=0h initial yield locus
1600
time consolidation yield locus
1400
Shear stress τ / Pa
1200
1000
800
600
400
200
0
0 500 1000 1500 2000 2500 3000
σc
Normal stress σ / Pa
σct
Fig. 19.13 Initial shear locus of yeast extract and shear locus of yeast extract stored for 5 days at
10 C above Tg (ring shear tester; T ¼ 25 C; aw ¼ 0.17, t ¼ 96 h; consolidation stress 1,200 Pa)
A number of macroscopically visible changes in the powder bulk are linked to

the described sinter processes. Figure 19.12 shows a glass tube filled with a dextrose
syrup powder (DE 21; water content 10%). The tube is heated in an oil bath at 90 C.
A loss of porosity and a shrinking of the powder package are observed. In the final
stage of the process the particle package collapses and a viscous melt is obtained.
The development of sinter bridges between the particles leads to a decreasing
flowability of the powder bulk. Figure 19.13 shows the shift of the yield locus and
the increase in unconfined yield strength sct of yeast extract with a water activity of
0.17, stored for 5 days at 10 C above its glass transition temperature. The yeast
extract shown in Fig. 19.13 exhibits a strong caking that corresponds with an
increase in unconfined yield strength of roughly 1,000 Pas. Similar results can be
obtained using dextrose syrup, maltodextrine powders, or other water-soluble
amorphous food powders.
It is necessary to predict such caking of amorphous powders in order to reduce
manufacturing costs and to avoid a decrease in quality of the final product.
Assuming a constant viscosity of the amorphous substance forming the bridge,
the unconfined yield strength should be a function of the average sinter bridge
diameter. In Fig. 19.14 the unconfined yield strength of spray-dried tomato powder,
dextrose syrup, and hydrolyzed whey permeate (Teunou and Fitzpatrick 1999a, b)
is plotted against the calculated squared ratio between the diameter of the sinter
bridge and the particle diameter (x/a)2. The ratio (x/a)2 is calculated through (19.6).
According to the results obtained, the unconfined yield strength increases with the
increasing ratio (x/a)2 more or less linearly for (x/a)2 values exceeding 0.05. The
observed onset of consolidation is in agreement with the studies of Downtown et al.
(1982), Wallack and King (1988), and Aguilera et al. (1993) who observed a caking
if the ratio between the radius of the sinter bridge and the particle diameter x/a
exceeds a value ranging from 0.01 to 0.1. According to (19.7) the unconfined yield
strength of the powder cake should theoretically increase linearly with (x/a)2. Such
correlation can be confirmed for values of (x/a)2 exceeding 0.05 (Fig. 19.14).
In the initial phase of the storage process the growth of the sinter bridge might be
delayed due to incomplete contact between the particles. The gap at the contact
points between single particles has to be bridged before the development of the
bridge is initiated. Such a delay of the sinter process might explain why a theoretical
ratio (x/a)2 of 0.05 has to be exceeded before a linear increase in tensile strength is
observed. Another explanation is that the initial powder is in disequilibrium with
50000
hydrolysed whey permeate
45000 [Teunou & Fitzpatrick, 1999]
dextrose syrup DE21
40000 spray/belt dried tomato powder
unconfined yield strength σc / Pa
spray dried tomato powder

35000
30000
Tomato powder; ring shear cell
25000 64 h; T-Tg=+10°C; (x/a)2=0.175
20000
15000
10000
5000
0 Tomato powder; ring shear cell

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 64 h; T-T =-10°C; (x/a)2=0.001
g
2
calculated ratio (x/a) / -
Fig. 19.14 Measured unconfined yield strength of tomato powder, dextrose syrup DE 21, and
hydrolyzed whey permeate plotted against the calculated ratio (x/a)2 (19.6)
the surrounding air and in the early phase of the experiments moisture still migrates
into the particles. Only the moisture content on the particle surface is used for
calculating the diameter of the sinter bridge, although the viscosity of the particle
inside is also relevant for sintering by viscous flow. Hence, a calculation based
purely on the initial moisture content of the particle surface at the beginning of the
process yields too large bridge diameters.
Apparently, the risk of caking can be predicted through the calculated sinter
bridge diameter. It is important to notice that equation (19.6) used for calculating
the theoretical ratio (x/a)2 is valid only for the initial phase of the sintering process.
Furthermore, dehydrated cell structures in the tomato powder particles bind the
strongly plasticized amorphous substance through capillary forces and some sub-
stances might crystallize. Thus, the availability of amorphous material (which
builds the sinter bridges) might be limited. Consequently, the real growth kinetic
of the sinter bridge can deviate significantly from the predicted growth rate.
Despite all the mentioned effects, it has been demonstrated that the risk of caking
of water-soluble amorphous powders can be predicted by calculating the theoretical
sinter bridge diameter. The influence of viscosity, which might vary by order of
magnitudes due to increasing moisture and/or increasing temperature, dominates by
far most of the other effects. The general conclusion is that caking can be expected
when the calculated (x/a)2 value exceeds 0.05.
19.4.2 Caking of Crystalline Powders
Crystalline particles show no time consolidation at constant moisture content. If the

relative humidity of the air exceeds temporarily, or locally, the critical humidity of
the crystalline substance, a partial superficial dissolution is observed. With decreas-
ing of the relative humidity, the generated liquid bridges dry out again. The
dissolved substance re-crystallizes and builds solid bridges between the particles.
Such changes might be observed especially when storing the powder in closed
storage containers, excluding a convective moisture transfer to the environment. To
estimate the risk of caking, it is sufficient to analyze the expected variations in
temperature, and to estimate the corresponding changes in relative humidity of the
air through the Mollier diagram for humid air. If the predicted maximum relative air
humidity inside the powder bulk exceeds the critical humidity of the crystalline
substance, the probability of caking is high.
19.4.3 Caking of Powder Mixes
Dehydrated food products are often mixes of different powdered components.

Caking of such powder mixes depends on the physico-chemical properties of
their components, the amount of each component, and the distribution of the different
components in the mix. There are very few scientific studies that investigate the caking
of complex powder mixes in view of the particle size distribution and the particle
shape of single components.
In the current study, the caking of binary mixes of amorphous spray-dried yeast
extract and amorphous spray-dried dextrose syrup powder (DE 21) was investigated
through time consolidation experiments performed in ring-shear cells. The two
powders are characterized by a similar particle size distribution. Both powder
components were stored for 3 days in the form of a thin powder layer at 25 C
and a relative humidity (RH) of 17 or 35%. Because the yeast extract powder starts
to cake during storage at a relative humidity of 35%, it was sieved again before
being mixed with a varying amount of dextrose syrup powder. The powder mixes,
characterized by a water activity of 0.17 or 0.35 and containing different amounts of
yeast extract and dextrose syrup, were filled into the shear cells, pre-sheared, and
stored at 25 C for 96 h under a consolidation (normal) stress of 1,100–1,300 Pa. At
relative humidity of 35% the dextrose syrup has a glass transition temperature of
50–60 C, while that of the yeast extract is roughly 15 C. Thus, it can be expected
that the yeast extract particles will be strongly plasticized as the dextrose syrup
remains glassy. Figure 19.15 shows the measured yield loci dependent on storage
time, humidity, and composition of the blend.
At water activity of 0.17 and corresponding relative humidity of 17%, the shear
stress required for yielding for all powder blends is comparably low. All dextrose
syrup/yeast extract blends are free-flowing. The flowability of powder blends
containing yeast extract, which was conditioned at relative humidity of 35%,
significantly decreases if the blend contains more than 25% yeast extract.
12000
0-50% dextrose syrup; aw=0.35; 96 h: 100% DS; aw=0.17; 0 h
11000 not measurable 75% DS; aw=0.17; 0 h
50% DS; aw=0.17; 0 h
10000
25% DS; aw=0.17; 0 h
75% dextrose syrup 0% DS; aw=0.17; 0 h
9000 aw=0.35; t=96 h 100% DS; aw=0.17; 96 h 50% yeast extract;
Shear stress τ / Pa
8000 75% DS; aw=0.17; 96 h t=0 h / aw=0.35

50% DS; aw=0.17; 96 h
7000 25% DS; aw=0.17; 96 h
0% DS; aw=0.17; 96 h
6000 50% DS; aw=0.35; 0 h
25% DS; aw=0.35; 0 h
5000 100% dextrose syrup 0% DS; aw=0.35; 0 h
aw=0.35; t=96 h 100% DS; aw=0.35; 96 h
4000 75% DS; aw=0.35; 96 h
0-100% dextrose syrup 25% yeast extract;
3000 0-75% dextrose syrup aw=0.17; t=0 t=0 h / aw=0.35
aw=0.17; t=96 h
2000
1000 0-100% dextrose syrup aw=0.17; t=0

0
0 500 1000 1500 2000 2500 3000
Normal stress σ / Pa 0% yeast extract;
t=0 h / aw=0.35
Fig. 19.15 Yield loci of different dextrose syrup/yeast extract blends at a water activity of 17% or
35% (storage time 0 h and 96 h; consolidation stress 1,100–1,300 Pa; bulk density 480–490 kg/m3)
To analyze the experimental data and to characterize the flowability of the

samples, the unconfined yield strength can be plotted against the corresponding
principle consolidation stress (Fig. 19.16). In addition, lines of constant flow factor
ffc (19.13) can be included in the established graph to characterize the flowability of
the samples. A flow factor smaller than 1 is equivalent to a hard powder cake.
A flow factor between 1 and 2 represents a very cohesive powder. In Fig. 19.16 the
unconfined yield strength of the non-stored powder mixes is plotted against their
principle consolidation stress.
Apparently, the flowability of pure dextrose syrup only slightly decreases if the
water activity is increased from 0.17 to 0.35. As already mentioned, the flowability
of powdered mixes containing at least 25% yeast extract significantly decreases if
the water activity of the mix is increased from 0.17 to 0.35. A hard powder cake is
obtained at a yeast extract concentration of more than 50%. At this concentration
the macroscopic behavior of the powder mix is governed only by the yeast extract.
At a yeast extract concentration of below 25%, the flowability of the entire powder
mixture is identical to that of pure dextrose syrup.
In Fig. 19.17 the unconfined yield strength of non-stored (t ¼ 0 h) powder
blends, with water activity of 0.35, is compared with the unconfined yield strength
of identical samples stored for 96 h.
The flowability of all powder mixes with water activity of 0.17 remains
nearly unchanged during storage. The same result is obtained for pure dextrose
syrup with water activity of 0.35. However, the flowability of mixes with water
activity of 0.35 and containing at least 25% yeast extract strongly decreases
during storage. Powder blends containing a higher amount of yeast extract are
6000 ffc=1 ffc=2 water activity aw = 0.17 / not stored

100% dextrose syrup initial
5000 75% dextrose syrup initial

Unconfined yield strength σc /Pa

4000

ffc=4
3000
water activity aw = 0.35 / not stored
100% dextrose syrup aw=0.35

2000
ffc=10
1000
0 0% dextrose syrup aw=0.35

0 2000 4000 6000 8000 10000 12000
Principle consolidation stress σ1 / Pa
Fig. 19.16 Unconfined yield strength dependent on the principle consolidation stress for different
non-stored dextrose syrup/yeast extract blends with different water activity (storage time 0 h;
pre-shear stress 1,100–1,300 Pa; powder density 480–490 kg/m3)
60000
0-50% yeast extract
96 h not measurable! water activity aw=0.35 / not stored

25% yeast extract
Unconfined yield strength σc /Pa
aw=0.35; T=25 °C; 96 h

40000

20000
water activity aw=0.35 / stored for 96 h
ffc=1 100% dextrose syrup aw=0.35 96 h

10000
75% dextrose syrup aw=0.35 96 h

0
0 2000 4000 6000 8000 10000 12000
Principle consolidation stress σ1 / Pa
Fig. 19.17 Unconfined yield strength dependent on the principle consolidation stress for different
stored dextrose syrup/yeast extract blends with water activity aw of 0.35 (storage time t ¼ 0 h and
t ¼ 96 h; pre-shear stress 1,100–1,300 Pa; powder density 480–490 kg/m3)
strongly plasticized and, thus, ring shear tests are not the appropriate method to
characterize their flowability. Hence, flowability of these powder mixes was
described only qualitatively.
In a second series of trials the caking of powder blends composed of amorphous
spray-dried dextrose syrup (DE 21; aw ¼ 0.17) and amorphous maltodextrine (DE 6;
aw ¼ 0.20) was investigated. Both spray-dried powders, which are mixes of differ-
ent dextrose oligomers, have a similar particle size distribution. Different blends of
the components of these two powders were stored in the form of a thin layer in open
tins at 30 C and 70% RH. The caking intensity of the different powder blends is
described using a predefined semi-quantitative scale. Contrary to the time consoli-
dation trials in ring-shear cells, an exchange of moisture between the powder layer
in the tins and the surrounding air is possible. At 30 C and relative humidity of 70%
the glass transition temperature Tg of the dextrose syrup is exceeded by 30 C,
whereas for the maltodextrine the difference between storage temperature T and Tg
is smaller than 5 C. Accordingly, it can be expected that only the dextrose syrup
particles will sinter together with solid surfaces or neighboring particles. The results
of the trials are summarized in Table 19.1.
Apparently, a concentration of 17–20% of the sintering component is required for
time consolidation of the entire powder blend (Palzer 2006). Assuming a systematic
mix of powder fractions, a similar size of the particles, and a cubic particle package,
one adhesive particle should be able to bind six surrounding non-adhesive particles.
This theoretical ratio (1:6) is equal to a mass fraction of 14%. Since the adhesive
particles are neither systematically distributed within the mix nor is an ideal cubic
particle package given, it is not surprising that the experimentally determined
Table 19.1 Caking of dextrose syrup DE 21/maltodextrine DE 6 blends stored at 30 C/70% RH
Concentration of 0 10 14 17 20 33 50 67 80 83 86 100
dextrose syrup/% (1:9) (1:6) (1:5) (1:4) (1:2) (1:1) (2:1) (4:1) (5:1) (6:1)
(DE 21: DE 6)
Storage time t ¼ 0 h – – – – – – – – – – – –
Storage time t ¼ 24 h – – – – + + + + + + + +
Storage time t ¼ 72 h – – – + + + + + + + + +
Storage time t ¼ 168 h – – – + + + + + + + + +
() free flowing; (+) caked
concentration of the adhesive particles required for caking is a few percentages

higher than the theoretical value.
It can be expected that the smaller the particles of one powder fraction,
the larger their influence on the macroscopic behavior of the mix. To avoid
caking of powder masses and to improve their flowability, 0.2–3.0% of a flow
agent (e.g., silicid acid or starch) is often added to powders that tend to exhibit
strong caking. The efficiency of such anti-caking agents can be improved by
reducing their mean particle diameter. However, it should be considered that the
anti-caking agents used in amorphous powders are only effective for a limited
storage time. With progressing time these small particles, acting as distance
pieces between the amorphous particles, are absorbed into the amorphous matrix
due to sinter processes.
Sometimes caking of blends of water-soluble crystalline and amorphous parti-
cles is observed at humidity/temperature combinations, in which case neither the
pure amorphous substance nor the pure crystalline substance is sufficiently plasti-
cized, and starts to dissolve. At the contact point between the crystals and amor-
phous particles a tertiary molecular mix composed of both substances and water is
formed. The sorption isotherm of this particle mix deviates significantly from the
sorption isotherm obtained by combining the isotherms of the pure substances
according to their percentage in the mix. Accordingly, a partial dissolution of the
substances is often observed at conditions in which the pure components of the
powder mix are stable. Bringing the particles of both components in contact with
each other increases their moisture sensitivity. This phenomena is sometimes also
refered to as delisquence. Thus caking is observed at a relative humidity value
which is significantly below the humidity, leading to time consolidation of the pure
components of the powder mix (Schreyer et al. 2008).
19.4.4 Influence of Re-crystallization of Amorphous

Substances on Caking
Crystallization of amorphous substances during storage influences caking because

only amorphous substances sinter together at the given low temperatures. Further-
more, crystallization processes affect caking due to the different moisture capacity
Table 19.2 Caking intensity and crystallinity of skim milk powder stored at 30 C/70% RH
Storage time t/h 0 2 4 6 8 10 12 14 18 20 22 24
Caking intensity – – – – (+) (+) (+) (+) (+) (+) (+) (+)
Crystallinity of lactose (% of total 8 52 74 84 86 87 87 96 97 100 100 100
lactose)
() free flowing; (+) slightly caked
between amorphous and crystalline matrices. The water capacity of crystalline

structures is limited to their crystal water. With some exceptions the amount of
moisture that can be bound in a crystalline structure is small. Hence, in most cases
water is liberated during crystallization of amorphous water-soluble substances.
The liberated humidity has an autocatalytic effect on the crystallization process and
it accelerates sintering of the remaining amorphous substance. On the other hand, it
transforms the amorphous material, which builds the sinter bridges between the
particles into stable crystalline structures. Taking these two effects into account it
can be expected that in the initial phase of the process crystallization accelerates
caking, whereas toward the end of crystallization a shortage of amorphous sub-
stance will limit the growth of the sinter bridges.
Caking and crystallization of amorphous lactose or spray-dried milk powder
containing amorphous lactose is a well-known phenomenon (Lloyd et al. 1996).
In this chapter, the impact of lactose crystallization on the caking of spray-dried
skim milk powder is discussed. The milk powder was stored at 30 C and relative
humidity of 70%. Table 19.2 shows the degree of crystallinity (quantified by light
microscopy with polarized light and image analysis) and the caking intensity
observed for different storage times.
Exposing the skim milk powder to the test climate, crystallization immediately
starts and the amount of available amorphous lactose decreases. In the mean time,
water is liberated due to the transformation of the hygroscopic amorphous lactose
into its crystalline state. After 6 h of storage, 84% of the total lactose has already
been crystallized. After storage for 8 h a caking of the powder is observed. During
the following hours the hardness of the obtained skim milk powder cake does not
increase further. This observation might be explained by a shortage in amorphous
lactose due to the progressing crystallization.
19.5 Summary and Conclusions
As shown in the current study caking of water-soluble powders is linked to either

sintering of the amorphous substances or partial dissolution and re-crystallization of
crystalline materials. The kinetic of caking of amorphous particles can be predicted
based on the calculated theoretical sinter bridge diameter. The risk of caking of
crystalline substances can be estimated by comparing the critical relative humidity
leading to dissolution with the relative humidity of the surrounding air. Caking of
binary mixes of water-soluble amorphous particles (similar size) is determined by
sintering of the most moisture sensitive component, i.e., if the amount of this
component exceeds 15–20% in the mix. However, due to interactions between the
different substances, time consolidation of mixes of water-soluble crystalline and
amorphous powders is sometimes observed at temperature/moisture conditions in
which the isolated pure components of the mix remain free flowing.
References
Aguilera J, Levi G, Karel M (1993) Effect of water content on the glass transition and caking of
fish protein hydrolysates. Biotechnol Prog 9:651–654
Aguilera J, Valle J, Karel M (1995) Caking phenomena in amorphous food powders. Trends Food
Sci Technol 6:149
Downtown G, Flores-Luna J, King C (1982) Mechanism of stickiness in hygroscopic, amorphous
powders. Ind Eng Chem Fundam 21:447–451
Frenkel J (1945) Viscous flow of crystalline bodies under action of surface tension. J Phys (USSR)
9(5):385–391
Hamaker H (1937) The London-Van der Waals attraction between spherical particles. Physica
4(10):1058–1072
Jenike, A.W. (1970) Storage and flow of solids. Bull. 123, Eng. Exp. Station, University of Utah,
Salt Lake City/US
Kuczynski G (1949) Self-diffusion in sintering of metallic particles. J Metal 1(2):169–178
Lifshitz E (1956) The theory of molecular attraction forces between solids. Soviet Phys JETP
2(1):73–83
Lloyd R, Chen X, Hargreaves J (1996) Glass transition and caking of spray-dried lactose. Journal
of Food Science Technology 31:305–311
PalzerS (2006) Influence of supra-molecular structure and storage conditions on the caking of
powders. Proceedings of the 5th World Congress on Particle Technology 2006, paper no. 231a
Palzer S (2009a) Influence of material properties on the agglomeration of water-soluble amor-
phous particles. Powder Technol 189:318–326
Palzer S (2010) Relating the supra-molecular structure of water-soluble food solids with their
material properties. Trends in Food Science and Technology 21(1):12–25
Peleg M, Mannheim C (1977) The mechanism of caking of powdered onion. J Food Proc
Preserv 1:3
Rumpf H (1958) Grundlagen und Methoden des Granulierens. Chem Ing Techn 30(3):144–158
Rumpf H (1970) Zur Theorie der Zugfestigkeit von Agglomeraten bei Kraft€ubertragung an
Kontaktpunkten. Chem Ing Techn 42(8):538–540
Rumpf H, Sommer K, Steier K (1976) Mechanismen der Haftkraftverst€arkung bei der Partikel-
haftung durch plastisches Verformen. Sintern und viskoelastisches Fliessen. Chem Ing Tech
48(4):300–307
Schatt W (1992) Sintervorg€ange – Grundlagen. VDI Verlag, D€usseldorf/G
Schreyer E, Palzer S, Sommer K (2008) Sorption, sintering and caking behaviour of binary
mixtures of amorphous and crystalline particles. Workshop on Particle & Powder Technology
as applied to foods, nutraceuticals and pharmaceuticals. Massy Paris/France
Teunou E, Fitzpatrick J (1999a) Characterization of food powder flowability. J Food Eng 39:31–37
Teunou E, Fitzpatrick J (1999b) Effect of relative humidity and temperature on food powder
flowability. J Food Eng 42:109–116
Tomas J (1996) Zum Verfestigungsprozeß von Sch€ uttg€
utern – Mikroprozesse, Kinetikmodelle und
Anwendungen. Sch€ uttgut 2(1):31–50
Wagner T (1997) Kaltspr€ uhen ein- und mehrphasiger Fl€ ussigkeiten. Ph.D. thesis, Food Process
Engineering, Department of Food Science, ETH Z€ urich/CH
Wallack D, King C (1988) Sticking and agglomeration of hygroscopic, amorphous carbohydrates
and food powders. Biotechnol Prog 4:31–35
Chapter 20
Effective Drying Zones and Nonlinear Dynamics
in a Laboratory Spray Dryer
Ulises Ramón Morales-Durán, Liliana Alamilla-Beltrán, Humberto

Hernández-Sánchez, Jose Jorge Chanona-Pérez, Antonio Ruperto
Jiménez-Aparicio, and Gustavo Fidel Gutiérrez-López
20.1 Introduction
Extensive experimental spray drying tests have been carried out by various
researchers to assess drying kinetics inside the chamber (Zbicinski et al. 2002).
At the beginning of this decade, about 20,000 spray dryers were used commercially
for processing agrochemical, biotechnological, chemical, pharmaceutical, and food
products. Evaporative capacities of spray drying equipment are in the range of
1–10 L/h for laboratory units and up to about 200 t/h for large-scale processing
(Mujumdar and Devahastin 2000). Use of inadequate drying systems and incorrect
drying conditions may lead to obtaining products in which added value and physical
and chemical properties do not meet the required specifications. Also, in some
cases, deposition of wet or rubbery product on the wall of the drying chamber may
occur, causing enormous economical losses (Goula and Adamopoulos 2004).
Selecting drying conditions, atomizing device, and size and geometry of the drying
chamber must take into account the desired characteristics of the product and, in
most cases, involve complex mass and heat transfer calculations. The most com-
mon food powders are baby foods, dairy products, proteins, coffee and tea extracts,
flavors and encapsulated fats (Chen and Patel 2008; Goula and Adamopoulos
2008). The amount of moisture removed from a droplet inside the drying chamber
depends on the mechanisms that govern the evaporation rate and the residence time
of the droplet in each zone of the equipment. This phenomenon can be described
using transport equations, CFD (Seydel et al. 2006; Zbicinski and Li 2006) and
nonlinear considerations (Van den Bleek et al. 2002).
U.R. Morales-Durán, L. Alamilla-Beltrán, H. Hernández-Sánchez, J.J. Chanona-Pérez,

and G.F. Gutiérrez-López (*)
Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas,
Instituto Politécnico Nacional, Carpioy Plan de Ayala, s/n CP. 11340, México DF, Edomex, Mexico
e-mail: gusfgl@gmail.com
A.R. Jiménez-Aparicio
Carretera Yautepec-Jojutla, Km. 6, calle CEPROBI No. 8, Col. San Isidro, Yautepec, Morelos,
C.P. 62731, México
516 U.R. Morales-Durán et al.
Some authors consider that atomization is a key stage in the process (Allen and
Bakker 1994; Furuta et al. 1994; Oakley 1994). Type of atomizer and feed partially
determine the kind of drying chamber to be used and the energy spent in atomizing.
Size distribution of the spray as well as particle trajectory and velocity inside the
chamber and powder collection system will also contribute to the overall quality of
the operation (Filoková and Mujumdar 1995; Chawla 1994). Moreover, in some
cases particles can produce turbulence, which in time can lead to different rates
of transport that may affect the quality of the product using the same equipment
(Van den Bleek et al. 2002).
Special attention should be given to drying times, trajectories of particles, and
the time that droplets take to reach the wall of the dryer. Chamber diameter and
volume play a key role regarding this issue (Gutiérrez et al. 1997, 1998).
Droplets of variable size and different morphology should reach the wall and
leave the dryer at the desired final moisture content. For certain foods, drops in
spray dryers often approach and reach the boiling temperature of the dissolvent
(water in most food related applications); vapor generation may cause particles
to grow in size and then collapse. Bubbles may form and collapse repeatedly
in a process known as morphological development (Hecht and King 2000).
The complexity of the spray drying process makes experimental study of mass
transfer from individual drops impractical, which has induced research on
predicting moisture content as well as on morphological characteristics of
particles during the process by suspending a drop of the material under study
in a fixed position (Adhikari et al. 2000; Ferrari et al. 1989). In this context,
digital image analysis can play an important role when evaluating changes in
particle size, formation of crust, shrinkage, inflation, blow-up, presence of holes,
etc. (Adhikari et al. 2000; Aguilera and Stanley 1999). Reports on the prediction
of various thermal and hydrodynamic conditions have made it possible to
distinguish among different phenomena associated with droplet morphology
and have allowed the recognition of periods and actual drying zones in drying
chambers (Dolinsky 2001).
Identification of drying zones inside the chamber and change of moisture content
during dehydration may help to establish relationships between dryer design and
product quality. In this work, effective drying zones evaluated as heat transfer units
in the chamber as well as the morphological development of the particle using
digital image analysis are reported for a two-fluid nozzle laboratory co-current
spray dryer. Also, computational fluid dynamics (CFD) has been used in this work
to evaluate hydrodynamics to assess the effects of particle-gas interactions in the
effective drying zones. This has allowed the establishment of correlations between
lumped models and nonlinear considerations of air-particle interactions, including
the evaluation of Lyapunov coefficients and formation of attractors for trajectories
(Van den Bleek et al. 2002). Computer vision systems may be applied to aid
discussion of the usefulness of traditional design calculations and findings derived
from intensive computing exercises, as well as from advanced theories on chaotic
behavior.
20 Effective Drying Zones and Nonlinear Dynamics in a Laboratory Spray Dryer 517
20.2 Materials and Methods
20.2.1 Testing Material
Testing material was, in all cases, 40% TS maltodextrin solution (20 ED) supplied
by Arancia Corn Products S.A. (Mexico). Experiments on trajectories of particles
were performed by using powdered maltodextrin (20 ED) supplied by the same
company.
20.2.2 Spray Dryer
The spray dryer used in this work was a two-fluid nozzle co-current laboratory
spray dryer equipped with a peristaltic pump for feed fine control and cyclone for
powder collection. The dryer is depicted in Fig. 20.1. The dimensions of the drying
chamber were diameter 0.38 m, height 0.60 m, and height of cone 0.40 m. Details of
design and construction of the experimental dryer are given elsewhere (Alamilla-
Beltrán et al. 2001). The evaporative capacity of the dryer is 0.8 kg of evaporated
4
5
6 8
9
2
16 11
10
12
1
14 13
15
Fig. 20.1 Diagram of spray dryer. (a) Heater; (b) temperature recorder; (c) thermocouples; (d) air
to nozzle; (e) manometer; (f) rotameter; (g) liquid feed; (h) feed pump; (i) pulse dampener; (j)
thermoanemometer; (k) cyclone; (l) fan; (m) motor; (n) powder collector; (o) supporting structure;
(p) drying chamber
water/h (similar to laboratory-scale dryers, e.g., SDMicroTM Spray Dryer, Niro;

Mini Spray Dryer B-290, B€ uchi).
Experiments were conducted under the following drying conditions: inlet/outlet
temperatures of 200/173 C (denoted as high-temperature drying), and 170/145 C
(denoted as low-temperature drying) and volumetric airflow of 75 m3/h. Feed rate
was, in all cases, 1.39 kg/h. The nozzle was located at 0.1 m from the top of the
chamber. A J-type thermocouple was used to measure the inlet air drying temperature.
At the bottom of the chamber the outlet air drying temperature was evaluated.
20.2.3 Moisture Content and Sampling of Material
Sampling of material inside the chamber was performed using a bayonet-type

sampler previously described (Alamilla-Beltrán et al. 2005). It was introduced to
the dryer through specially constructed tapped devices separated 0.10 m from each
other. A diagram of the dryer showing the measuring points is presented in
Fig. 20.2. Average moisture content was evaluated following AOAC 32.1.03
method (AOAC 1995) for the different heights of the chamber in which each
sample was collected. Sampling was performed once drying proceeded in a
constant regime.
0.38 m
B
Sampling points
0.6 m
C
D
0.1 m
E
Fig. 20.2 Diagram of the

longitudinal section of the
dryer showing sampling
points and chamber
dimensions
20.2.4 Evaluation of Air and Product Temperature
The air drying temperature was evaluated using thermocouples type J. One thermo-
couple was fitted at the air inlet. A second thermocouple was positioned at the air
outlet duct at the bottom of the chamber. This thermocouple was used to measure the
outlet air drying temperature. The thermocouples were adapted to a digital register,
Barnant (model 692). The wet bulb temperature was measured at the same points and
inside the chamber at the sampling points (Fig. 20.2) during the drying process. With
dry bulb temperature and wet bulb temperatures, the enthalpy was calculated.
The product temperature was measured using the thermocouple collocated at the
sampling point A.
20.2.5 Mean Particle Diameter During Drying
Particles obtained during drying experiments were observed using light micro-
scopy. The microscope (Zeiss Axiophot) was fitted with a 20 objective and blue
filter (total magnification 100 and 200, depending on drying experiment). The
illumination was provided by a light-field source. Images were obtained by means
of a digital camera (ZVS-47DE) with resolution of 640 480 pixels and captured
in a PC. KS400 ver. 3.01 software was used to process images and to determine the
diameter of individual particles. Mean diameter of particles at different heights in
the chamber was reported as the corresponding average value for powders collected
at the center of the spray-cone and at different heights along the drying chamber as
depicted in Fig. 20.2. Initial diameter of the droplet was measured using a particle
analyzer Malvern Series 2,600 at the feed rate and atomizing air pressure used in the
drying operation.
20.2.6 Measurement of the Mass Flow Rate
Mass flow rate of air (Gs) was evaluated by measuring the mean air velocity in the
air exhaust duct of the dryer with a hot-wire anemometer (Digital Thermoanem-
ometer TSI Inc., model 8330-M, USA). Mass flow rate of liquid (L) was evaluated
using a peristaltic pump (Masterflex-Cole Parmer Instrument, model 7521–40).
Spraying air was controlled by means of a rotameter (Blue White).
20.2.7 Effective Drying Height
In this work, a proposal for the calculation of the effective drying height based on
the application of transfer units was applied. This methodology considers that
the longitude along which one of the fluids travels through the dryer is formed
by the number of transfer units (NtOG) corresponding to a dimensionless function of

the difference of temperatures and the driving force for heat transfer. The height of
the transfer unit (HtOG) is a function of the conditions of flow.
The evaluation of the effective drying height was based on the consideration that
the drying process is similar to the humidification operation, with the exception of
the influence of the solids (Foust et al. 1993). This influence is important and in this
proposal it was considered that the drying air is humidified in adiabatic form during
the process. The flow diagram for the proposal is presented in Figs. 20.3 and 20.4.
Figure 20.3 illustrates calculations of the product temperature at different points
along the dryer, and Fig. 20.4 shows the evaluation of the effective drying height.
The number of transfer units is related to its capacity to transfer heat and may be
evaluated by using (20.1).
Ach v Tgamb Twamb Ql ρl
Gsa Y´
nH L1
Tg1
Gsaa
fs
Cs
Xi Ss
(GsaY'ai)+(Ss(X1-Xi+1))+(GsaaY'aai)
Yi+1 =
(Gsa+Gsaa)i+1
Ei = Ss(X1-Xi)
Tgi+1 Csi+1
Wri = (Li-Ss) –E i
TwaaO TgaaO Li = Ss + Wr i
Xw = 1-fs
Yaa
Tgaa
CpL i = Li (1.675+2.512(Xw))
Haa
GsaHai + LiCpLi (tLi) + Gsaa Haai - (Gsa + Gsaa)Hi+1

tLi+1 =
Li+1Cp Li+1
Fig. 20.3 Flow diagram of evaluation of product temperature

Tgi tLi
Gs' Yi Yi+1
Tgi+1 tLi+1
(Dti - Dti+1)
Dtm =
æ Dt ö
ln çç i ÷÷
è Dti+1 ø æ é Y + Yi ù ö
G = Gs' çç 1 + ê 1 ú ÷÷
è ë 2 ûø dc
qD GsCs (Tgi - Tgi+1)

Uo = =
Vch Δ tm Ach Lsp Δtm
0.67
Ua = 237 G
dC
Gsavg
G'SCs
Gsavg HtOG =
Gs' = Ua
Ach
Tgi - Tgi+1 DtG

NtOG = =
Dtm Dtm
Z=HtOGNtOG
Fig. 20.4 Flow diagram of evaluation of effective drying height
The height of the heat transfer unit depends on the volumetric coefficient of heat
transfer and the volumetric airflow. It can be calculated as follows (Treybal 1996):
Tgi Tgiþ1 DTg

NtOG ¼ ¼ (20.1)
Dtm DTm
In the case of rotational dryers (in which air flows through a mass of dispersed
solids), the volumetric coefficient of heat transfer in kJ/(h) (m3 of dryer) ( C) can be
calculated with (20.3). This equation is based on the flow mass of drying air by unit
of transversal area, and the diameter of the same one (Friedman and Marshall 1949;
McCormick 1962).
G0s Cs
HtOG ¼ (20.2)
Ua
An increase in the speed of the gas will increase the exposed surface. The contact
air-particle consists of a mass of humid material in dispersed phase in contact with
drying air in the continuous phase.
G0:67
Ua ¼ 237 (20.3)
dc
The number of transfer units and the length of a transfer unit are related by
means of (20.4) (Treybal 1996). This relationship allows determination of the value
of the effective drying height (Z) inside the drying chamber.
Z ¼ NtOG HtOG (20.4)
20.2.8 Computational Fluid Dynamics Simulation
Simulation of the drying process was carried out using CFX 10.0. Drying air
velocity field was simulated using the k–e (Southwell and Langrish 2000; Langrish
et al. 2004). The Lagrangian–Euler approach was selected for the resolution of the
discrete phase. Drag force was assessed using the Schiller–Nauman correlation.
Airflow and particles-air interactions were also evaluated.
20.2.9 Simulation of Airflow Profiles
To simulate the air-velocity of drying air, a three-dimensional representation of

the equipment was made. Domain was divided in a 138,717 elements. Boundary
conditions were established, with inlet air taken as 1.54 102 kg/s according to
experimental data. Atomizing air inlet was 116 m/s corresponding to 1.54 104 kg/s
evaluated by considering the flow divided by the area of annulus of the nozzle. At the
outlet zone of the dryer, a relative static pressure of 0 Pa was considered. A steady-state
simulation was carried out until reaching an average relative error of 104. Steady-
state simulation was used as the initial value of nonsteady-state flow for 10 s, using a
0.24 s resolution. The need for these sorts of experiments and CFD simulation has been
noted by Fletcher et al. (2006).
20.2.10 Nonlinear Dynamics of the System
Dynamics of a dispersed maltodextrin powdered phase in the chamber was eval-

uated under airflow drying conditions. Fifty grams per minute of powdered malto-
dextrin having a range of particle sizes of 5–45 mm was fed into the dryer.
A 200 mW laser beam was fitted to the dryer to axially illuminate the trajectories
of the particles. Images of the trajectories were taken using a webcam (iSlim 2020
AF. Genius, USA) every 0.24 s. Images were taken at 0.1 m and 0.3 m from the
atomizing nozzle in the axial direction, which also corresponded to air velocity-
measuring points. Crops of images were obtained and fractal dimension of texture
(FDT) was evaluated by using the SDBC plugging of ImageJ (Chen et al. 2003).
Series 2 min long were recorded to observe oscillations of the FDT values. Time
series so obtained were processed to derive the corresponding Lyapunov coeffi-
cients of the system and the phase-space of reflected light (Wolf et al. 1985). These
calculations could render useful information on the nonlinear (chaotic) behavior of
the system and allow the derivation of important conclusions on overall tendencies
of flow into the dryer by calculating Kolmogorov Entropy values.
Evaluation of the number of transfer units allowed recognition of the three different
stages during drying. The experimental evidence of water removed from the
product (moisture content), temperatures of the drying air and the product (inside
the aspersion cone), along with changes in particle size (shrink and expansion) have
been associated with the stages of drying. Two of these factors are related to water
evaporated and shrinkage and a third with particle growth.
Initially, a first zone called the first drying stage was observed, in which
elimination of water solely occurs without the generation of powder. It would be
possible to assume that the flow in drying zones is less mixed than for zone 3 (or
that a controlling flow induced by the nozzle is generated). This section coincides
with the characteristics of stages I and II as reported by Dolinsky (2001).
A second zone, called the second drying stage was observed in which the
formation of the shell of the particle generates a powdered product with almost the
equivalent of the final moisture of the solid. For the low-temperature operation, a
93% loss of moisture was reached, while for high-temperature drying, the loss of total
humidity was 97%. In both cases, the temperature of the air reached minimum value.
Also, shrinkage of the particle was observed along with an incipient expansion.
At 200 C, particles shrank, mainly within the spraying cone. This could not be
appreciated at 170 C, but it is probable that similar phenomena would be observed
between 0.2 m and 0.3 m below the nozzle. At the same time, the temperature of
the product increased from the adiabatic-saturation temperature until the boiling
temperature of water was reached. Likewise, in this stage, the volumetric coefficient
of heat transfer had the highest value, suggesting the existence of a pattern of mixed-
flow. This section agrees with the characteristics of the stage III reported by Dolinsky
(2001). A better description of this mixed-flow can be carried out by applying CFD
techniques such as those described in the Materials and Methods section.
After the drying stages, a third zone, denominated as the expansion stage was
detected. In this phase it was observed that after shrinkage the particle begins to
expand. The size of the particle increased twofold and occasionally presented
cracking, and in some cases, explosion of the material and rupture of the shell
were observed. Particles that exploded were, in general, hollow spheres presenting
thin crusts for high-temperature drying conditions and thick crusts for low-temperature
drying conditions. The presence of this stage is related to the temperature of the
drying air, the moisture of the particle, and the gradient of pressures between the
interior and the exterior of the droplet. In this zone, the loss of humidity was not
significant and it could be considered as not important for drying. Also, in this zone,
the temperature of the solid remained constant at its higher value. In this stage, the
volumetric coefficient of heat transfer presented its lower value. Stages IV and V as
reported by Dolinsky (2001) coincide with this expansion stage. The effective
height of drying represents the minimum longitude to carry out effective drying.
On the other hand, solids concentration of the feed, inlet air drying temperature,
mass flow of inlet product, mass flow of atomizing air, and type of feed can modify
the morphological development of the particle and, in consequence, the final
characteristics of the product. Changes in operating conditions will cause different
effective heights of drying. However, drying and expansion stages will be present
for this material. In Figs. 20.5 and 20.6, a pictorial representation of punctual
temperatures of air and images of the product are presented for both experimental
conditions. The shadowed areas represent each of the recognized drying stages.
In Fig. 20.7 a representation of nonsteady-state air patterns inside the chamber is
presented. No influence of the nozzle in this case was considered. It should be noted
that there is a recirculating airflow with a central downward jet similar to a flow in
an infinite plenum surrounded by an upward stream of air. The highest density of
upward flow vectors was found in the first and part of the second drying zone
corresponding to the effective drying zones evaluated earlier through the evaluation
250
200
Second Drying Stage
First Drying Stage
Temperature (°C)
150
Expansion
100
50
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Distance along the dryer (m)
Air drying Product
Fig. 20.5 Temperature profile of drying air and product inside the dryer and an image of the final
dried product, corresponding to the condition 170/145 C
250
200
Temperature (°C)
Second Drying Stage

First Drying Stage
150
Expansion
100
50
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Distance along the dryer (m)
Air drying Product
Fig. 20.6 Temperature profile of drying air and product in the dryer and an image of the final dried
product, corresponding to the condition 200/173 C
Velocity
(Vector 1)
1.000e+000
7.500e-001
5.000e-001
2.500e-001
0.000e+000
[m s∧-1]
Fig. 20.7 CFD representation of non steady-state air patterns inside the chamber with no
influence of the nozzle
of the transfer units, which may explain the intense drying in this region. In
Fig. 20.8, the representation of nonsteady-state air patterns with influence of the
nozzle is presented at different times of simulation. It is possible to observe that a
nonstable air pattern was found. However, in all cases, a denser zone of mixed
Fig. 20.8 Representation of

nonsteady-state air patterns Velocity Initial (0 s)
with influence of the nozzle (Contour 1)
presented at 0, 5, and 10 s of 5.000e-001
simulation 4.500e-001
4.000e-001
3.500e-001
3.000e-001
2.500e-001
2.000e-001
1.500e-001
1.000e-001
5.000e-002
0.000e+000
[m s∧-1]
Velocity 5 s
(Contour 1)
5.000e-001
4.500e-001
4.000e-001
3.500e-001
3.000e-001
2.500e-001
2.000e-001
1.500e-001
1.000e-001
5.000e-002
0.000e+000
[m s∧-1]
Velocity 10 s
(Contour 1)
5.000e-001
4.500e-001
4.000e-001
3.500e-001
3.000e-001
2.500e-001
2.000e-001
1.500e-001
1.000e-001
5.000e-002
0.000e+000
[m s∧-1]
(upward and downward) flow was found, which confirms that effective drying
zones are found in the first zone (first and second drying zones) of the chamber.
In Fig. 20.9, it is possible to observe the simulated trajectories of the particles
used as tracers with no influence of the atomizing nozzle. Most particles were
directed to the bottom of dryer since drag dominated the process. However, when
the nozzle was considered in the simulation (Fig. 20.10), an imbalance of trajec-
tories was caused and the nozzle governed hydrodynamics. This phenomenon was
validated experimentally by feeding particles to the drying chamber. Images of an
air-particles mix were taken by illuminating the field with the laser beam. Reflected
light was scattered and seemed to present irregular patterns that were evaluated by
calculating their fractal dimension at different times. Surface 3.D plots of gray-level
intensity are shown in Fig. 20.11; they show the heterogeneity of laser beam
reflection in the particles and are an indicator of oscillation of the air-particle
flow. FDT was presented in the form of phase-space graphs, and the formation of
attractors was evident. For 10 cm axial distance from the top, Lyapunov coefficients
in the range of 3.46–13.62 were obtained, while for 30 cm the coefficients were
MDX. Mean Particle Diameter

(Res PT for MDX)
4.253e-005
3.322e-005
2.392e-005
1.461e-005
5.298e-006
[m]
Fig. 20.9 Simulated trajectories of the particles used as tracers with no influence of the atomizing
nozzle; most particles were directed to the bottom of the dryer since drag dominated the process
Water. Mean Particle Diameter

(Res PT for Water)
8.888e-005
6.934e-005
4.979e-005
3.025e-005
1.070e-005
[m]
Fig. 20.10 Simulated trajectories of the particles used as tracers with the presence of the
atomizing nozzle showing an imbalance of trajectories
5.44–11.63. These results also confirmed that more turbulent zones are found
around the atomizer and within the first drying zone evaluated earlier through the
evaluation of the transfer units, which may explain the intense drying in this region.
These results also agree with those obtained by Huang et al. (2005) who used a k–e
model and found that the highest turbulence was around the atomizing device. In
Figs. 20.12 and 20.13, graphs on the oscillations of FDT and corresponding
Lyapunov coefficients are presented, as well as phase-space diagrams showing
the obtained attractors at 10 and 30 cm axial distance from the top of the dryer.
20.4 Conclusions
The application of the concept of the number of transfer units in spray drying
allowed the establishment of three stages related to the drying of the material. An
initial first drying stage was identified, in which the product temperature remained
Greyscale intensity Greyscale intensity
Y Y
X Axial 1, t= 0 s X Axial 3, t=0 S

Y Y
X Axial 1, t= 0.60 S X Axial 3, t= 0.72 S

Y
Y
X Axial 1, t= 1.20 S
Greyscale intensity
Greyscale intensity
Y
Y
Greyscale intensity
Greyscale intensity
Y Y
Fig. 20.11 Surface 3-D plots of gray-level intensity corresponding to laser beam reflection on
particles, showing heterogeneity of flow at: (a) 10 cm and (b) 30 cm axial distance from the nozzle
a
2.72
2.7
2.68
FDT
2.66
2.64
2.62
De-phased FDT
2.6
2.6 2.61 2.62 2.63 2.64 2.65 2.66 2.67 2.68 2.69 2.7
b
2.72
2.7
2.68
FDT
2.66
2.64
2.62
De-phased FDT
2.6
2.6 2.62 2.64 2.66 2.68 2.7 2.72
Fig. 20.12 Phase-space diagrams of FDT showing the presence of an attractor at: (a) 10 cm and (b)
30 cm axial distance from the nozzle. The 45 straight line shows the nearest neighboring points
constant and near to the adiabatic saturation temperature of the drying air. A second
drying stage was identified, which showed an increase in the product temperature
with the presence of dry powder. A third expansion stage was then identified in
which an increment in the size of the particle was observed. The first and second
drying stages represent the area in which an effective drying inside the spraying
chamber was carried out. The calculated values of the volumetric coefficient of heat
transfer allowed the assumption that in the second drying stage a pattern of mixed
a
2.72
2.7 Lyapunov = 3.46
2.68
FDT
2.66
Lyapunov = 13.62
2.64
2.62
2.6
0 0.5 1 1.5 2 2.5 3
Time (s)
b
2.72
Lyapunov = 5.44
2.7
2.68
FDT
2.66
2.64 Lyapunov = 11.63
2.62
2.6
0 0.5 1 1.5 2 2.5 3 3.5 4
Time(s)
Fig. 20.13 Oscillations of FDT and corresponding Lyapunov coefficients for: (a) 10 cm and
(b) 30 cm axial distance from the nozzle. Distance of text-box covers points used for calculation
flow was present. Drying zones presented according to CFD simulation (and
experimental validation) a high deal of air-particle recirculation phenomena. Non-
linear dynamics could characterize these recirculation zones by using Lyapunov
coefficients and by the presence of attractors related to fractal dimension of texture
of the reflected laser beam cropped images. Highest simulated and validated
turbulence was found in the upper part of the dryer. Steady and nonsteady-state
simulations allowed insight into the nonlinear properties of generated turbulence.
While the transfer units approach is useful for construction of lumped models, CFD
and experiments based on air-particle trajectories constitute a good approach for the
understanding of turbulence inside the dryer. Both approaches may be useful and
could complement each other for design purposes.
20.5 Symbols
Ach Chamber area (m)

CpL Heat capacity of liquid
Cs Heat capacity of moist air (kJ/kg C)
E Mass flow rate of water evaporated (kg of water/h)
Fs Fraction of solids
Gs Mass flow rate of air (kg/h)
Gs0 Mass flux of air (kg/h m2)
Gsm Average mass flow rate of air
H Enthalpy of air (kJ/kg of dry air)
HtOG Length of heat-transfer units (m)
L Mass flow rate of liquid (kg/h)
Lsp Distance between sampling points (m)
NtOG Number of heat-transfer units
Ql Volumetric flow rate of liquid (m3/h)
qD Rate of heat flow (kJ/h)
tL Temperature of the solution or dry solid ( C)
Tg Temperature of drying air ( C)
Tw Wet bulb temperature of drying air ( C)
Uo Volumetric heat transfer-coefficient (kJ/h m3 C)
Uom Average volumetric heat transfer-coefficient (kJ/h m3 C)
v Air velocity (m/s)
Vch Volume of cylindrical section of the spray chamber (m3)
vh Humid volume (m3/kg dry air)
Wr Water in the solution (kg of water/h)
X Moisture content (dry basis) (kg of water/kg of dry solid)
Xw Moisture content (wet basis) (kg of water/kg of solution)
Y Absolute humidity of air (kg of water/kg of dry air)
Z Effective length of drying, m
20.5.1 Greek Letters
Dt Temperature difference between air and dry solid or solution ( C)

DtG Change in gas temperature owing to heat transfer to solid only ( C)
rl Density of liquid (kg/m3)
Dtm Logarithmic-mean temperature difference ( C)
o Dry solid
20.5.2 Subscripts
1 At the inlet conditions of spray dryer

2 At the outlet conditions of spray dryer
aa Atomizing air
aaO Atomizing air at ambient conditions
amb Ambient
a Drying air
i Sampling point, inlet conditions of spray dryer
Acknowledgment Author U.R. Morales-Durán thanks Biotecsa-México for their support. The
authors also thank IPN-México and CONACYT-México (project 84287) for their financial
support.
References
Adhikari B, Howes T, Bhandari B, Truong V (2000) Experimental studies and kinetics of single
drop drying and their relevance in drying of sugar-rich foods: a review. Int J Food Prop
3(3):323–351
Aguilera J, Stanley D (1999) Simultaneous heat and mass transfer: Dehydration. In: Microstruc-
tural principles of food processing and engineering, 2nd edn. Aspen Publishers, New York,
pp 373–411
Alamilla-Beltrán L, Hernández-Parada A, Chanona-Pérez J, Jiménez-Aparicio A, Suárez-Fernández O,
Santiago-Pineda T, Gutierrez-Lopez G (2001) Design and performance of a spray dryer for
food processing. Proceedings of the 8th International Conference on Engineering and Food.
ICEF 8, Technomic Pub. Co, USA, pp 1151–1155
Alamilla-Beltrán L, Chanona-Pérez J, Jiménez-Aparicio A, Gutiérrez-López G (2005) Description
of morphological changes of particles along spray drying. J Food Eng 67(1–2):179–184
Allen R, Bakker H (1994) Spray dryer control-based on line particle size analysis. Trans IChemE
Part A: Chem Eng Res Design 72:251–254
AOAC (1995) Official methods of analysis of AOAC International, 16th edn. AOAC International,
USA, pp 31–32
Chawla J (1994) Effect of the droplet agglomeration on the design of spray dryer towers. Drying
Technol 12(6):1357–1365
Chen X, Patel K (2008) Manufacturing better quality food powders from spray drying and
subsequent treatments. Dry Technol 26(11):1313–1318
Chen W-S, Yuan S-Y, Hsieh C-M (2003) Two algorithms to estimate fractal dimension of gray-
level images. Opt Eng 42(8):2452–2464
Dolinsky A (2001) High-temperature spray drying. Dry Technol 19(5):785–806
Ferrari G, Meerdink G, Walstra P (1989) Drying kinetics for a single droplet of skim milk. J Food
Eng 10:215–230
Filoková I, Mujumdar A (1995) Industrial spray drying systems. In: Mujumdar A (ed) Handbook
of spray drying. Marcel Dekker, New York, pp 263–307
Fletcher D, Guo B, Harvie D, Langrish T, Nijdam J, Williams J (2006) What is important in the
simulation of spray dryer performance and how do current CFD models perform? Appl Math
Model 30(11):1281–1292
Foust A, Wenzel L, Clump C (1993) Principles of unit operations. CECSA, México
Friedman S, Marshall W Jr (1949) Studies in rotary drying. Part II – heat and mass transfer. Chem
Eng Prog 45(9):573–588
Furuta T, Hayashi H, Ohashis T (1994) Some criteria of spray drying design for food liquid. Dry
Technol 12(1–2):151–177
Goula A, Adamopoulos K (2004) Influence of spray drying conditions on residue accumulation:
Simulation using CFD. Dry Technol 22(5):1107–1128
Goula A, Adamopoulos K (2008) Effect of maltodextrin addition during spray drying of
tomato pulp in dehumidified air: I. Drying kinetics and product recovery. Dry Technol 26(6):
714–725
Gutiérrez G, Ordorica C, Osorio G, Hernández A, Patiño R, Jiménez A, Santiago T (1997)
ASCON-programa para el establecimiento de las condiciones de operación en secadores por
aspersión de disco rotatorio. In: Mulet A, Ordorica C, Benedito J (eds) Herramientas de cálculo
en ingenierı́a de alimentos III. Univ. Politécnica de Valencia, Instituto Politécnico Nacional,
España-México, pp 93–103
Gutiérrez G, Osorio G, Jiménez A, Pyle L (1998) An assessment of droplet-air contact and spray
drying performance in bioprocess engineering. In: Galindo E, Ramı́rez O (eds) Advances in
bioprocess engineering II. Kluwer, The Netherlands, pp 251–275
Hecht J, King C (2000) Spray drying: influence of developing drop morphology on drying rates
and retention of volatile substances. 1. Single-drop experiments. Indus Eng Chem Res
39:1756–1765
Huang L, Passos M, Kumar K, Mujumdar A (2005) A three-dimensional simulation of a spray
dryer fitted with a rotary atomizer. Dry Technol 23(9–11):1859–1873
Langrish T, Williams J, Fletcher D (2004) Simulation of the effects of inlet swirl on gas flow
patterns in a pilot-scale spray dryer. Chem Eng Res Des 82(7):821–833
McCormick P (1962) Gas velocity effects on heat transfer in direct heat rotary dryers. Chem Eng
Prog 58(6):57–61
Mujumdar A, Devahastin S (2000) Mujumdar’s practical guide to industrial drying. Exergex
Corporation, Canada
Oakley D (1994) Scale-up of spray dryers with the aid of computational fluid dynamics. Dry
Technol 12(1–2):217–233
Seydel P, Bl€omer J, Bertling J (2006) Modeling particle formation at spray drying using popula-
tion balances. Dry Technol 24(2):137–146
Southwell D, Langrish T (2000) Observations of flow patterns in a spray dryer. Dry Technol
18(3):661–685
Treybal R (1996) Mass transfer operations. McGraw-Hill, México
Van den Bleek C, Coppens M-O, Schouten J (2002) Application of chaos analysis to multiphase
reactors. Chem Eng Sci 57(22–23):4763–4778
Wolf A, Swift J, Swinney H, Vastano J (1985) Determining Lyapunov exponents from a time
series. Physica D 16(3):285–317
Zbicinski I, Li X (2006) Conditions for accurate CFD modeling of spray-drying process. Dry
Technol 24(9):1109–1114
Zbicinski I, Delag A, Strumillo C, Adamiec J (2002) Advanced experimental analysis of drying
kinetics in spray drying. Chem Eng J 28:207–216
Chapter 21
Rehydration Modeling of Food Particulates
Utilizing Principles of Water Transport
in Porous Media
I. Sam Saguy, Oranit Troygot, Alejandro Marabi, and Rony Wallach
21.1 Introduction
In the last decade, there has been a continuous rise in the demand for convenience
foods, including dehydrated products, mainly due to modern lifestyles (Marabi and
Saguy 2009; Saguy et al. 2007). This trend is accompanied by a decrease in the
ability, desire or time to prepare food and an increase in financial means, leading
consumers to choose foods that are readily available, convenient, and require only
minimal or no preparation before consumption (Tillotson 2003). The rehydration
of dried foods is a fundamental unit operation in the food industry. The quality of
rehydrated and reconstituted products is affected by the drying conditions and
rehydration processes utilized, ultimately influencing consumer acceptance. During
the drying process, physicochemical changes, including textural and structural
modifications, migration of solutes, loss of volatiles and nutrients occur in an
irreversible manner, and have an impact on the quality of the final product.
Therefore, the drying process needs to be understood and controlled, to create a
dried product with optimal nutritional, sensorial, and rehydration characteristics
(Saguy et al. 2007). In addition to medium uptake, leaching of solids from the food
product to the medium is another important aspect during rehydration. To address
I.S. Saguy (*) and O. Troygot

Institute of Biochemistry, Food Science and Nutrition, Robert H. Smith Faculty of Agricultural,
Food and Environment, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot, 76100, Israel
e-mail: ssaguy@agri.huji.ac.il
A. Marabi
Institute of Biochemistry, Food Science and Nutrition, Robert H. Smith Faculty of Agricultural,
Food and Environment, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot, 76100, Israel
and
Solid Products, Department of Food Science & Technology, Nestlé Research Center, Vers-Chez-
Les-Blanc CH-1000, Lausanne 26, Switzerland
R. Wallach
Department of Soil and Water Sciences, Robert H. Smith Faculty of Agricultural, Food and
Environment, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot, 76100, Israel
536 I.S. Saguy et al.
this phenomenon, nondissolvable solids were proposed (Marabi et al. 2004a) and
utilized in the determination of the rehydration ratio(RR).
21.2 Mathematical Modeling
In recent years, the food domain has experienced an encouraging transition from the
empirically-based to physically-based models. Nevertheless, most rehydration stud-
ies reported in the literature still include empirical and semiempirical models to
describe the mechanisms of liquid uptake, solids leaching, and the kinetics of the
processes (Marabi and Saguy 2009).
Mathematical modeling facilitates an understanding of process characteristics,
providing insight into the governing mechanisms taking place, and thus it could be
utilized to improve the process for better products. Liquid uptake is typically
modeled by applying a mechanistic approach or employing an empirical approach.
Fick’s second law of diffusion is a typical example of a mechanistic approach,
whereas the first-order model is merely a curve-fitting empirical model. The devel-
opment of empirical models requires considerably less effort and therefore, they are
utilized frequently. However, empirical models are limited and nontransferable
(Saguy et al. 2005b).
21.2.1 Empirical and Semiempirical Models
Five models that are often used (Saguy et al. 2007) include: the exponential model
(Misra and Brooker 1980), Peleg’s model (Cunningham et al. 2007; Garcia-Pascual
et al. 2006; Giraldo et al. 2006; Peleg 1988), first-order kinetics (Gowen et al. 2007;
Krokida and Philippopoulos 2005); the Weibull distribution function (Cunha et al.
1998a, b; Cunningham et al. 2007; Garcia-Pascual et al. 2006; Machado et al.
1997), and the normalized Weibull distribution function (Marabi et al. 2003,
2004a, b; Marabi and Saguy 2004; Marabi and Saguy 2005).
Among the empirical models, the Weibull distribution function is used frequently
and has recently been improved to describe the rehydration of dried foods. It is
important to note however, that all of these empirical models offer rather limited
insight into the fundamental principles involved, and in some cases, hinder under-
standing of the transport mechanism(s) (Marabi and Saguy 2009). Nevertheless, they
provide an excellent basis for curve-fitting and allow process representation as a
function of physical properties and rehydration conditions.
21.2.2 The Diffusion Model
The diffusion model is a combination of physical and empirical approaches,

founded on Fick’s first and second laws. Solving Fick’s second law requires making
21 Rehydration Modeling of Food Particulates Utilizing Principles of Water Transport 537
numerous assumptions and simplifications (Saguy et al. 2007). To overcome some

of these difficulties, the effective diffusion coefficient Deff is utilized, which is
derived from experimental data and is an apparent value that encompasses all
intrinsic hydraulic properties of the particles. The effective diffusivity is linked to
both the porosity and tortuosity. A typical Deff value for moisture in foods ranges
from 10–8 to 10–12 m2/s, generally being closer to 10–10 m2/s (Maroulis et al. 2001),
and is influenced by temperature, water content, pressure, physical properties, and
the dried food’s structure. However, it is well documented that some or most of the
above assumptions made to solve Fick’s laws are not valid. Moreover, mechanisms
of mass transfer other than molecular or Fickian diffusion may also occur. Some of
the drawbacks of the Fickian approach were recently discussed (Saguy et al. 2007).
21.3 Paradigm Shift: Capillary Flow in Porous Media
As most empirical models are primarily utilized to avoid more complex considera-
tions and complications such as changes occurring within the product, the actual
microstructure is not considered. The alternative is to apply a physically-based
approach utilizing the porous media where the actual microstructure and void
channels are important. This may sound straightforward, but it requires a paradigm
shift and an interdisciplinary approach is suggested (Saguy et al. 2007). The
approach is based on integrating the know-how developed in other domains such
as soil science to study the porous media and fluid transport in foods.
Rehydration is a very complex phenomenon involving different transport
mechanisms, including molecular diffusion, convection, hydraulic flow, and capil-
lary flow (Saravacos and Maroulis 2001). Liquid imbibition follows various
mechanisms occurring in tandem (Chiralt and Fito 2003; Marabi et al. 2003; Marabi
and Saguy 2005; Oliveira and Ilincanu 1999). However, the contribution of
mechanisms involving mass flux due to a temperature gradient (Soret effect) is
considered insignificant and therefore often is disregarded (Datta 2007a). Several
studies utilizing the Washburn equation to represent the movement of liquids
into porous food matrices were reviewed recently (Marabi and Saguy 2009;
Saguy et al. 2007).
It is however worth noting that the rehydration mechanism is no longer relying
solely on the common postulation that it is governed by Fickian diffusion. Other
mathematical models based on capillary flow in porous media are also utilized.
These models are derived from the well-known Lucas (1918) and Washburn (1921)
equations (also called the Lucas-Washburn or Washburn-Rideal equation). This
approach is based on utilization of several well-known derivations known as the
Laplace, Poiseuille, and Lucas-Washburn equations (Marabi and Saguy 2009).
After imbibition of water into dried porous foods it was shown that the process
followed the Washburn equation (Lee et al. 2005; Saguy et al. 2005a). However,
discrepancies related to the utilization of a single “effective” cylindrical capillary
radius and constant contact angle were also reported (Saguy et al. 2005a). An
additional factor that may be responsible for the inaccuracies encountered when
comparing experimental data with the Washburn equation may be related to the
tortuosity of the pores within the food sample. The pore network is often regarded
as a bundle of cylindrical and straight capillaries with a determined effective
radius (Saguy et al. 2007). Thus, the Washburn equation may be utilized in its
original form (Aguilera et al. 2004), or otherwise corrected with a tortuosity factor
(Carbonell et al. 2004).
Other studies also concluded that there is water movement during various food
processes. For instance, water transfer during vacuum osmotic drying was
described as a combination of traditional Fickian diffusion and vacuum capillary
flow. In this case, water transfer was closely related to the porosity of the fruit being
tested (Shi and Fito-Maupoey 1994). Radial NMR micro-imaging technique uti-
lized to study pasta rehydration indicated a non-Fickian process (Hills et al. 1996).
Capillary penetration or another fast transport mechanism occurring near the
interface in the rehydration of apples was also suggested (Salvatori et al. 1999).
21.4 Flow in Unsaturated Porous Media
21.4.1 Capillarity and Tension Head
Water that has entered a porous medium but has not drained out of the sample will be
retained in the pores by capillary forces or will surround the surface of the particles
by molecular forces of adhesion and cohesion. Therefore, a simple measurement of
water content in the medium is not sufficient to enumerate the complete status of the
medium’s water. While the quantity of water present in the medium is very impor-
tant, the potential or affinity with which the water is retained is perhaps more
important, mainly if the dynamics of this water is considered (Saguy et al. 2007).
This potential may be defined as the amount of work done or potential energy stored,
per unit volume, in moving a mass of water from the reference state (typically
chosen as pure free water). Matric potential could be considered as potential energy
per unit volume (J m3), which is also expressed as Pascal. This clarifies the use of
the terminology “pressure potential” by soil physicists referring to matric potential
as soil pressure, or if divided by bulk density, as pressure head (Wallach 2007).
Capillarity deals with both the macroscopic and statistical behavior of interfaces,
rather than their molecular structure. Capillarity is also referred to as surface
tension, g (work per unit area expressed as J m2). This phenomenon is extremely
important in water retention in porous media. Surface tension occurs at the mole-
cular level and involves two types of molecular forces: adhesive forces, which are
the attractive forces of molecules of dissimilar substances and cohesive forces,
which are the attractions between molecules in similar substances. Cohesive forces
decrease rapidly with distance and are the strongest in solids, less strong in liquids,
and the weakest in gases (Saguy et al. 2007).
As water rises in a capillary, the meniscus is spherical in shape and concave

upward. By letting r equal the tube radius, the excess pressure above the meniscus
compared to the pressure directly below can be described under various assump-
tions by 2g/r. As the pressure on the water surface outside the capillary tube is
atmospheric, the pressure in the liquid below the meniscus will be less than
the atmospheric pressure above the meniscus by 2g/r. This will force the fluid up
the tube until the hydrostatic pressure of the fluid column within the tube equals
the excess pressure of 2g/r. The total force (upward) supports the weight of the
fluid column to the height hc. The height of the capillary rise can be expressed as
(Laplace equation):
2g cos ya
hc ¼ (21.1)
rgr
where hc ¼ capillary rise (water tension head) (m); g ¼ gravity (m/s2); ya ¼ apparent
contact angle (o).
21.4.2 Water Retention Curve
The water characteristic curve of porous media is also known as the water retension
curve (RC), which describes the functional relationship between the volumetric
water content y and matric potential c under equilibrium conditions. The matric
potential is usually replaced by the pressure potential head h (m), which is the
energy per unit weight of water. As the water in the unsaturated porous media is at
subatmospheric pressure, the pressure potential head is commonly called “tension
head.” This curve is an important property related to the distribution of pore space
(sizes, interconnectedness), which is strongly affected by texture and structure, as
well as related factors including organic matter content. The RC indicates the
amount of water in the porous medium at a given tension head (Saguy et al.
2007). It is also a primary hydraulic property required for modeling water flow in
porous media. RCs are highly nonlinear functions and relatively difficult to obtain
accurately (Hillel 1998; Wallach 2007).
The traditional method of determining water retention involves establishing a
series of equilibria between water in the porous medium sample and a body of water
at known water tensions. The medium-water system is in hydraulic contact with the
body of water via a water-wetted porous plate or membrane. At each equilibrium,
the volumetric water content of the medium is determined and paired with a value
of the tension head (h), determined from the pressure in the body of water and
the gas phase pressure in the substrate. The data pair (y, h) forms one point on an
RC. A summary of methods frequently used in soil science to determine the RC can
be found in the literature (e.g., Klute 1986). It is worth noting that RC quantification
resembles sorption isotherm determination in foods (Fontana and Campbell 2007;
Saguy et al. 2007).
Measured values of water content and tension head (y -h) are often fragmentary
and usually based on relatively few measurements over the wetness range of
interest. For modeling and analysis purposes, and characterization and comparison
of different substrates and scenarios, it is essential to represent the RC in continuous
and parametric form. A parametric expression of an RC model should contain as
few parameters as possible, to simplify its estimation and to describe the behavior at
the limits (wet and dry ends), while closely fitting the nonlinear shape of the y -h data
(Saguy et al. 2007).
Many models have been suggested to describe RC in soil science (Hillel 1998;
Jury and Gardner 1991). The most frequently used are those of Brooks and Corey
(1966), denoted as B-C model, and van Genuchten (1980), denoted as VG model.
The parameters of the models are typically determined by curve-fitting of experi-
mental data. B-C model follows:
8 l
>
> y yr h
>
<y y ¼ S e ¼ h > hb
s r hb
>
> y yr
>
: ¼ S e ¼ 1 h hb
ys yr
(
KðhÞ ¼ Ks Se 2þ2=ðmv þtÞ h > hb
(21.2)
KðhÞ ¼ Ks h hb
where Se ¼ effective saturation (-); hb ¼ air entry value (m); l ¼ “pore size
index” empirically-determined parameter (-); y, yr, ys ¼ current, residual, and
saturated volumetric water contents, respectively (m3/m3); Ks ¼ saturated hydrau-
lic conductivity (m/s); mv ¼ empiric parameter (-).
The residual water content is somewhat arbitrarily defined as the water content at
which the corresponding hydraulic conductivity is essentially zero, but very often it
is used as an empirical constant when fitting hydraulic functions. As opposed to ys,
which has a clear physical significance, the meaning of yr and its estimation have
not yet been resolved (Wallach 2007).When yr ¼ 0 (m3/m3), Se approaches S. Note
that Se varies between zero and one.
The other common parametric model for relating water content to the tension
head was proposed by VG:
mv
y yr 1
Se ¼ (21.3)
ys yr 1 þ ðah hÞn
where ah (m1) and mv (-) are empirical parameters (determining the shape of the
RC; derived by curve-fitting techniques).
The value of the hydraulic conductivity of a saturated porous medium (Ks)
depends on the properties of the medium and the flowing fluid. This dependence
can be separated into two factors: fluidity f (defined as rg/; ¼ viscosity (Pa s))
and intrinsic permeability k (m2):
krg
Ks ¼ kf ¼ (21.4)

The intrinsic permeability of a medium is a function of pore structure and

geometry. Particles of smaller-sized individual grains have a larger specific surface
area, increasing the drag on water molecules flowing through the medium, which
results in a reduced intrinsic permeability and Ks.
21.4.3 Richards Equation, Boundary, and Initial Conditions
The Darcy law describes the flow equation in saturated media:
DH
J ¼ Ks (21.5)
Ds
where H ¼ the hydraulic head (m); s ¼ distance along a stream line in the flow
field (m); DH/Ds ¼ the hydraulic-head gradient along the stream line (-); and
J ¼ flux density or flow per unit area opposite the direction defined by hydraulic-
head gradient (m/s).
The Darcy law can be coupled with the conservation of mass principles to derive
a continuity equation. Using one horizontal dimension leads to:

@y @ @h
¼ KðhÞ (21.6)
@t @x @x
Equation (21.6) contains two unknowns, namely y and h; applying the chain rule
for the left hand side of (21.6) provides the h-based continuity equation:

@h @ @h
CðhÞ ¼ KðhÞ (21.7)
@t @x @x
where C(h) ¼ ∂y/∂h is the slope of the RC and is called the specific moisture
capacity (1/m).
Applying the chain rule for the derivative to the right hand side of (21.6)
provides the y-based continuity equation, which is a diffusion-type (nonlinear)
equation:

@y @ @y
¼ Dh ðyÞ (21.8)
@t @x @x
where Dh (y) ¼ K(h)∂h/∂y (m2/s) is denoted as hydraulic diffusivity and is the

ratio of the unsaturated hydraulic conductivity to the specific moisture capacity.
The advantage of the y-based form is that Dh (y) does not vary with y nearly as
much as K(h) varies with h. The last two equations, along with their various
alternative formulations, are known as the Richards equation.
21.4.4 Developing the Theory of Flow in Porous Media
To utilize the theory of flow in porous-media in rehydration of dry food particu-

lates, the RC of these particulates is needed. The B-C and VG and other functions
used to model the RCs have significant limitations at low-liquid saturations in that
they use a parameter called residual saturation, which is the minimum liquid
saturation calculated by these equations. At low water content, vapor diffusion is
predominantly driven by vapor pressure differences, which in turn is driven by
local vapor pressure lowering (related to capillary pressure through Kelvin’s
equation). Hence, vapor flow is significantly influenced by the RC at low water
contents.
The limitations of the two-phase characteristic curves in the low water content
region lead to errors in liquid and vapor transport, and in the prediction of the actual
water content in this region. This value is important in the transition from liquid–
solid sorption to vapor-solid sorption, which is typically at 4–6 monomolecular
layers, or liquid saturation of 0.11–0.16 (m3/m3) based on soil surface area of
25 m2/g (Webb 2000). At higher liquid saturations, adsorption is primarily via
liquid–solid processes. As the liquid saturation decreases, however, vapor-solid
sorption contributes to the adsorption process more and more and may dominate at
low values of liquid saturation.
Physically-based modification of the RC has been suggested by a number of
authors (e.g., Campbell and Shiozawa 1992; Morel-Seytoux and Nimmo 1999;
Ross et al. 1991; Rossi and Nimmo 1994). Campbell and Shiozawa (1992) observed
that the capillary pressure in the dry region is a linear function of liquid saturation
on a semilog plot; using this relationship, they added a VG water content relation-
ship to obtain the full capillary pressure curve, where the parameters in the VG
equation were refit to the data assuming zero liquid residual saturation. Subsequent
models have all used the linear relationship observed by Campbell and Shiozawa
(1992). The model proposed by Rossi and Nimmo (1994) is based on the B-C model
with zero residual saturation, which is equivalent to the expression of Campbell
(1974), with the dry region function proposed by Campbell and Shiozawa (1992) in
the low water content region. Morel-Seytoux and Nimmo (1999) slightly modified
the Rossi and Nimmo (1994) model. Webb (2000) presented a similar approach
to the dry region modification of Morel-Seytoux and Nimmo (1999) in that it
merges an existing RC with a dry region expression. However, the approach of
Morel-Seytoux and Nimmo (1999) requires the simultaneous solution of two
nonlinear equations, while the Webb (2000) approach only requires a few iterations
in a single equation.
Following these approaches, we suggested a new method to predict the RCs of

porous media (food particulates in particular) from measured water sorption iso-
therms. This method is based on a four-step procedure (Saguy et al. 2007):
1. Water sorption isotherm
The suggested method adopts the fundamental water isotherm approach to
describe the relationship between water content and water activity in the dry
zone. This selection is straightforward and takes into consideration the knowl-
edge bank on the relationship between food stability and water activity.
Although the utilization of glass transition theory could also be considered, in
this case the relationship between water and the gas phase is of interest. The
conversion from water activity to tension head is carried out by utilizing the
Kelvin equation:
RT
h¼ lnðaw Þ (21.9)
rw gMw
where R ¼ gas constant (J·K1·mol1); Mw ¼ molecular water mass (0.018 kg/mol);

T ¼ absolute temperature (K); rw ¼ water density (1,000 kg/m3); aw ¼ water
activity (-).
However, the theoretical aspects of the Kelvin equation at low water activities
are still under development and further studies are required. A typical water
sorption isotherm is depicted in (Fig. 21.1) for microcrystalline cellulose
(MCC).
2. Retention curve
Initially, the sorption isotherm data is converted to water retention data and
plotted as tension head vs. volumetric volume of water to comply with the RC
presentation. Note that the moisture content for each data point of the sorption
20
Experimental data
Water content (g/100 g DB)
16 GAB
12
0
0.0 0.2 0.4 0.6 0.8 1.0
Water activity
Fig. 21.1 Typical water sorption isotherm for microcrystalline cellulose at 25 C (Adapted from
Saguy et al. 2007)
1.E+7
Prediction
1.E+6
Sorption data
Water tension (cm)
1.E+5 Retention curve

Exponential sorption
1.E+4 curve
1.E+3
1.E+2
1.E+1
1.E+0
0.0 0.2 0.4 0.6 0.8 1.0
q (cm3 / cm3)
Fig. 21.2 Typical water sorption and water retention data for microcrystalline cellulose at 25 C
(prediction made using B-C model) (Adapted from Saguy et al. 2007)
isotherm (m, aw) is converted to (h, y) by utilizing the Kelvin equation (21.9),
transforming the moisture content to its volumetric value:
rb
y¼m (21.10)
rw
where m ¼ moisture content (kg H2O/kg dry solids; DB); rb ¼ bulk density (kg/m3).
Typical values are presented in (Fig. 21.2).
3. Prediction and validation of RC from aw data
The sorption isotherm data after transformation to the adequate RC scale was
utilized to derive the B-C model using the two aforementioned requirements,
namely continuity and smoothness at a point defined as y*, h*. This yielded two
equations from which y* and l were derived. Once y* is known, it can be placed
in one of the equations to derive l and the B-C function can be drawn. Note that
in our first approach, and for simplicity of treatment, it was assumed that the
sorption isotherm is a linear function on a semilogarithmic scale (e.g., Webb
2000). Another requirement for using the B-C function, when physical data is
absent, is to know the values of its parameters: hb, ys (saturated water content), yr
(residual water content), and l (pore index parameter). The first two parameters
can be experimentally measured (although adaptation to food systems is
required): yr is an approximated parameter while l is evaluated through model
equations. Applying this approach in a VG model (instead of B-C one) requires
an additional parameter, which adds to the complexity of the solution without
gaining any further conceptual insight into the overall approach, and as such is
not covered herewith.
Validation was then carried out, in which RC was measured using the hanging
column method (Klute 1986) and the fit obtained, as depicted in (Fig. 21.2).
Since methods used in this case for soils are probably difficult for implementa-
tion in most food systems, measurements of only a few points is recommended,
starting from the wet end of the curve.
4. Retention curve validation
Once the water RC is known, the unsaturated hydraulic conductivity function
(21.2) can be determined and the Richards equation (21.8) solved. As for the
saturated hydraulic conductivity, it is independently measured. Again, certain
adaptations are necessary for evaluating Ks in food systems. Both the measure-
ment and prediction of 10-cm column dry MCC rehydration are depicted in
(Fig. 21.3). Simulations were made using Hydrus 1-D software (a code for
simulating one-dimensional movement), based on the Richards equations for
water flow (Simunek et al. 2005) for the following boundary and initial condi-
tions: constant upper boundary flux and constant tension head at bottom bound-
ary set to zero. The initial moisture content was assumed as oven-dry. The
simulation predicted the cumulative bottom flux, was then multiplied with the
sample cross-section, yielding the volume of water absorbed by the sample.
5. Typical experimental data on foods
To demonstrate the applicability of the above four-step procedure, imbibition
tests were carried out using freeze-dried wheat groats (burghul). Simulations
(not showed) illustrated that although the B-C characteristic-curve based on
water absorption model yielded a shape similar to the experimental data for early
times (ca. 500 s), a significant discrepancy was clearly observed at relatively
long times. Rehydration times longer than ca. 1,000 s are quite lengthy, com-
pared to the shorter duration utilized for typical “instant” consumer products.
Conversely, the B-C model was quite accurate at shorter times (e.g., less than
300 s). Thus, these data suggest that prolonged food rehydration is more complex
and requires additional consideration. A similar conclusion was reached after
7
Cumulative bottom flux (cm)
4
Calculated
Calculated
3
Experiment 1 1
2
Experiment 2 2
1
0
0 60 120 180 240 300 360 420
Time (s)
Fig. 21.3 Typical rehydration data for microcrystalline cellulose at 25 C (Adapted from Saguy
et al. 2007)
repeating the long-time rehydration experiments with freeze-dried carrots (data

not shown).
Next, an empirical approach was applied using a “double” Weibull model to
account for diffusion and relaxation, as both could have a completely different
time scale:
n h io n h io
wt ¼ w11 1 exp ðt=a1 Þb1 þ w12 1 exp ðt=a2 Þb2 (21.11)
where wt, w11, and w12 ¼ the momentary weight, infinite-weight due to diffusion,
and relaxation, respectively (kg H2O/kg dry solids); t ¼ time (s), a and b ¼ scale
parameter (s), and Weibull shape parameter (–); 1 ¼ diffusion and 2 ¼ relaxation.
When the combined model was utilized, the fitted data showed an excellent fit
for short and long rehydration times. The model was used to simulate the
contribution of each individual mechanism, namely diffusion and relaxation.
As expected, relaxation lags diffusion markedly. This could explain some of the
literature data in which both diffusion and nondiffusion mechanisms were
reported.
The criterion utilized to differentiate between the diffusion and relaxation
mechanisms is derived from the value of the Weibull shape parameter. Values
of 0.6–0.8 were defined as capillary flow, while values above 1.0 were char-
acterized as relaxation (Marabi et al. 2003). Figure 21.4 shows the ratio between
relaxation and diffusion; it is clear that at short rehydration time the freeze-dried
product depicts that the diffusion mechanism is dominating; however, as time
progresses, relaxation also starts to play a significant role.
This data suggests that additional considerations are required especially for long
rehydration time during which water flow is not capillary-driven. It is suggested
that the porosity of a dried food product (e.g., freeze-dried carrot, freeze-dried
60
50
Relaxation/Diffusion
40
Ratio (%)
30
20
10
0
0 2000 4000 6000 8000
Time (s)
Fig. 21.4 Simulated ratios (“double” Weibull model) between water uptake due to diffusion and
relaxation (freeze-dried wheat groats, burghul, 25 C)
wheat groats) is hierarchical (i.e., made of pores: between- and within- the
particles. The former pores are inter-porosity, while the latter are intra-porosity,
which are typically dead-ended). While water flow is driven by capillarity in the
interpores, it is also driven by moisture content gradients in the intrapores. This
concept (also known as “mobile-immobile”) is currently under development
and preliminary results are encouraging supporting the overall hypothesis that
physically-based model for the simulations of porous foodstuffs rehydration is
feasible.
This important topic will become apparent once internal structure data are
available (see recommendations below).
21.4.5 Rehydration of Foods Using Porous Media: Additional

Considerations
The theory of capillary imbibition for modeling the rehydration of foods was
applied by several other groups. For instance, a capillary-flow approach was
utilized (Weerts et al. 2003a, b) to model the temperature and anisotropy effects
during the rehydration of tea leaves. The predicted values agreed well with the
experimental data derived from NMR measurements, leading to the conclusion that
the physically based constitutive relationships of water activity and hydraulic
conductivity could be utilized to overcome the simplification of modeling water
transport as a process governed by Fick’s laws. The approach could be extended to
include gravity and osmotic pressure effects and also coupled with heat and solute
transport in porous media for modeling heat, water, and chemical transport in
general hydration and drying operations of porous food materials.
Another study (Singh et al. 2004) noted that one of the salient features of fluid
transport through biological systems is the complex flow path presented by the
biopolymeric matrix, thus expanding the above approach. Similarly, it was recently
proposed that use of an “effective” single cylindrical capillary radius is the simplest
possible model to describe capillary penetration into a porous medium. However,
this model may be insufficient in many cases (Saguy et al. 2005a), especially
when a significant distribution of pore size exists (Marmur and Cohen 1997), as
previously shown for dried foods (Karathanos et al. 1996). More specifically, and
depending upon the type of food and its processing history, it may contain pores
ranging in radius from 0.1 to 300 mm, those in the 10–300 mm range being the most
common (Bell and Labuza 2000).
Although numerous models have been proposed and frequently applied, addi-
tional improvements are required before modeling of the process based on the real
mechanism(s) occurring becomes possible. Nevertheless, recent studies focused on
more fundamental physical-based approaches have taken on a central role. Hence,
significant progress is anticipated.
21.5 New Approaches and Other Advances
The rehydration of dried foods involves different physical mechanisms, among

them water imbibition, internal diffusion (in the solid and pores), convection and
diffusion at the surface and within large open pores, hydraulic flow, capillary flow,
and relaxation of the solid matrix (Marabi and Saguy 2009; Saguy et al. 2007).
Additionally, we have highlighted the utilization of a double Weibull model as well
as mobile-immobile model. A growing body of data as well as ours indicates that
simplification of the rehydration process may not be fully warranted, or in some
cases, might even result in misleading conclusions (Datta 2007a, b; Saguy et al.
2005b; Weerts et al. 2003a, b). This realization is relatively new and important
insights are coming to light from the merging of different scientific disciplines,
including food and soil sciences, biophysics, environmental sciences, and petro-
leum and chemical engineering. The ultimate goal of these interdisciplinary efforts
is to provide more advanced and accurate physical models that describe the
mechanism(s) taking place during the rehydration process (Saguy et al. 2007).
The models recently developed and successfully applied to the flow of liquids
into dried foods are based on theories usually termed as “capillary-flow approach”
and/or “flow in porous media.” The theories are already well developed in other
fields, as are the methodologies needed to generate the required experimental data.
However, there is still a considerable lack of data and, more importantly, standard
methods allowing collection of the relevant physical properties of the food matrix
and its interaction with the liquid. It is expected in the near future that collaborative
studies will provide the information needed to apply these theories.
The common starting point for developing models that account for fluid flow in
porous media is Darcy’s equation and its Navier–Stokes analog (Datta 2007a;
Saguy et al. 2005b; Weerts et al. 2003b). In addition, the main roles of the food
particles’ physical properties (e.g., pore-size distribution, heterogeneity, tortuosity),
the embedding liquid (e.g., density, dynamic viscosity), the external conditions
(e.g., temperature, pressure), and the interface (e.g., contact angle) are also consid-
ered in order to quantify the changes occurring during drying and to facilitate the
modeling of the rehydration process. These studies on fluid flow in porous media
have presented, in adequate detail, the development of the models, and the reader is
referred to them for more detailed information (Saguy et al. 2007).
Recent studies focusing on the various transport mechanisms in porous media
include:
l Molecular diffusion of gases, including water vapor: Darcy flow of gases resulting
from pressure and liquid flow due to gas and capillary pressures (Datta 2007a).
l Utilization of Lucas-Washburn equation in modeling rehydration of dried foods
(Lee et al. 2005; Machado et al. 1997; Saguy et al. 2005a).
l New methods and technologies to obtain experimental data (Datta 2007a, b;
Saguy et al. 2005b) including: (1) density, porosity, thermal conductivity, and
specific heat of solid material, (2) molecular, capillary, and effective diffusivities,
(3) moisture isotherms, (4) vapor and liquid permeability, and (5) water-potential
curves; data in the last two categories is either nonexistent for foods or very
difficult to obtain.
It is worth noting that new technology and utilization of X-ray MCT capabilities
are spreading. X-ray MCT is a powerful tool for detailed observations of micro-
structure of porous foods and is expected to become a pertinent method for obtain-
ing detailed information that can be further utilized in computational simulations
of the imbibition of liquids into porous media. Hence, it could open new avenues
for ultimately quantifying and modeling internal changes during both drying
and rehydration, thereby providing the necessary information for studying the
mechanism(s) involved, and enabling better control and quality improvements.
21.6 Research Needs
Future research needs to focus on the accomplishment of four main goals, as

recently summarized (Saguy et al. 2007): (1) expand the theory and approach
outline above to different aw, and RC models and foods, (2) apply “transport in
porous media” theories to model the fate of dissolved and retained substances in
food particulates, (3) develop physical-based models-multidisciplinary research
integrating/assimilating theories and knowledge from other fields, and (4) perform
in-situ quantification–development of new methods/techniques, for in-situ quanti-
fying of changes in food matrix microstructure, and water/vapor movement.
21.7 Conclusions
We have highlighted how porous media physics could be implemented for model-
ing the rehydration of foods. The main conclusions can be summarized as:
l Sorption isotherm data could be extended to a complete RC for use in a rehydra-
tion model.
l Flow in porous media theory, as demonstrated, is applicable to the rehydration of
food particulates.
l Moving from empirical models to physically-based ones for foods was demon-
strated and therefore, is recommended.
l Multidisciplinary collaboration is paramount and synergistic in opening new
avenues of research.
The field is entering a new era in which innovative approaches and multidisci-
plinary teams will provide fresh insights, resulting in new discoveries and a better
and deeper understanding of the phenomena. This will ultimately lead to much
better control of the structure and rehydration, leading to improved food products
with enhanced consumer appeal and acceptability.
References
Aguilera JM, Michel M, Mayor G (2004) Fat migration in chocolate: diffusion or capillary flow in
a particulate solid? A hypothesis paper. J Food Sci 69:R167–R174
Bell LN, Labuza TP (2000) Moisture sorption – practical aspects of isotherm measurement and
use. American Association of Cereal Chemists, St. Paul
Brooks RH, Corey AT (1966) Properties of porous media affecting fluid flow. J Irrig Drain E-Asce
2:61–68
Campbell GS (1974) A simple method for determining unsaturated conductivity from moisture
retention data. Soil Sci l17:311–314
Campbell GS, Shiozawa S (1992) Prediction of hydraulic properties of soils using particle-size
distribution and bulk density data. International Workshop on Indirect Methods for Estimating
the Hydraulic Properties of Unsaturated Soils. University of California, Riverside, CA
Carbonell S, Hey MJ, Mitchell JR, Roberts CJ, Hipkiss J, Vercauteren J (2004) Capillary flow and
rheology measurements on chocolate crum/sunflower oil mixtures. J Food Sci 69:E465–E470
Chiralt A, Fito P (2003) Transport mechanisms in osmotic dehydration: the role of the structure.
Food Sci Technol Int 9:179–186
Cunha LM, Oliveira FAR, Ilincanu LA (1998a) Application of the probabilistic Weibull distribu-
tion to rehydration kinetics: relationship between the model parameters and the underlying
physical mechanisms. In: Oliveira JC, Oliveira FAR (eds) Proceedings of the 3rd workshop of
the Copernicus project, Leuven, Belgium
Cunha LM, Oliveira FAR, Oliveira JC (1998b) Optimal experimental design for estimating the
kinetic parameters of processes described by the Weibull probability distribution function.
J Food Eng 37:175–191
Cunningham SE, McMinn WAM, Magee TRA, Richardson PS (2007) Modelling water absorption
of pasta during soaking. J Food Eng 82:600–607
Datta AK (2007a) Porous media approaches to studying simultaneous heat and mass transfer in
food processes. I: problem formulations. J Food Eng 80:90–95
Datta AK (2007b) Porous media approaches to studying simultaneous heat and mass transfer in
food processes. II: property data and representative results. J Food Eng 80:96–110
Fontana AJ Jr, Campbell GS (2007) Applications of water activity in non food systems. In:
Barbosa-Canovas GV, Fontana AJ Jr, Schmidt SJ, Labuza TP (eds). Water Activity in
Foods: Fundamentals and Applications. Chapter 14. IFT Press/Blackwell, Ames, IA
Garcia-Pascual P, Sanjuan N, Melis R, Mulet A (2006) Morchella esculenta (morel) rehydration
process modelling. J Food Eng 72:346–353
Giraldo G, Vazquez R, Martin-Esparza ME, Chiralt A (2006) Rehydration kinetics and soluble
solids lixiviation of candied mango fruit as affected by sucrose concentration. J Food Eng
77:825–834
Gowen A, Abu-Ghannam N, Frias J, Oliveira J (2007) Modelling the water absorption process in
chickpeas (Cicer arietinum L.) – the effect of blanching pre-treatment on water intake and
texture kinetics. J Food Eng 78:810–819
Hillel D (1998) Environmental soil physics. Academic, San Diego
Hills BP, Babonneau F, Quantin VM, Gaudet F, Belton PS (1996) Radial NMR microimaging
studies of the rehydration of extruded pasta. J Food Eng 27:71–86
Jury WA, Gardner WR (1991) Soil physics. Wiley, New York
Karathanos VT, Kanellopoulos NK, Belessiotis VG (1996) Development of porous structure
during air drying of agricultural plant products. J Food Eng 29:167–183
Klute A (1986) Water retention: Laboratory methods. In: Klute A (ed) Methods of soil analysis.
Part I, physical and mineralogical methods, 2nd edn. American Society of Agronomy, Madison
Krokida MK, Philippopoulos C (2005) Rehydration of dehydrated foods. Drying Technol
23:799–830
Lee KT, Farid M, Nguang SK (2005) The mathematical modelling of the rehydration character-
istics of fruits. J Food Eng 72:16–23
Lucas R (1918) Ueber das zeitgesetz des kapillar aufstiegs von Flussigkeiten. Kolloid-Zeitschrift
23:15–22
Machado MF, Oliveira FAR, Gekas V (1997) Modeling water uptake and soluble solids losses
by puffed breakfast cereal immersed in water or milk. In: Jowitt R (ed) Proceedings of the seventh
international congress on engineering and food – Part I. Sheffield Academic Press, Sheffield, UK
Marabi A, Dilak C, Shah J, Saguy IS (2004a) Kinetics of solids leaching during rehydration of
particulate dry vegetables. J Food Sci 69:FEP91–FEP96
Marabi A, Jacobson M, Livings S, Saguy IS (2004b) Effect of mixing and viscosity on rehydration
of dry food particulates. Eur Food Res Technol 218:339–344
Marabi A, Livings S, Jacobson M, Saguy IS (2003) Normalized Weibull distribution for modeling
rehydration of food particulates. Eur Food Res Technol 217:311–318
Marabi A, Saguy IS (2004) Effect of porosity on rehydration of dry food particulates. J Sci Food
Agr 84:1105–1110
Marabi A, Saguy IS (2005) Viscosity and starch particle size effects on rehydration of freeze-dried
carrots. J Sci Food Agr 85:700–706
Marabi A, Saguy IS (2009) Rehydration and Reconstitution of Foods. In: Ratti C (ed) Advances in
food dehydration. CRC Press, Boca Raton
Marmur A, Cohen RD (1997) Characterization of porous media by the kinetics of liquid penetra-
tion: the vertical capillaries model. J Colloid Interf Sci 189:299–304
Maroulis ZB, Saravacos GD, Panagiotou NM, Krokida MK (2001) Moisture diffusivity data
compilation for foodstuffs: effect of material moisture content and temperature. Int J Food
Prop 4:225–237
Misra MK, Brooker DB (1980) Thin-layer drying and rewetting equations for shelled yellow corn.
Trans ASAE 23:1254–1260
Morel-Seytoux HJ, Nimmo JR (1999) Soil water retention and maximum capillary drive from
saturation to oven dryness. Water Resour Res 35:2031–2041
Oliveira FAR, Ilincanu L (1999) Rehydration of dried plant tissues: Basic concepts and mathemati-
cal modeling. In: Oliveira FAR, Oliveira JC (eds) Processing foods. CRC Press, Boca Raton
Peleg M (1988) An empirical-model for the description of moisture sorption curves. J Food Sci
53:1216–1219
Ross PJ, Williams J, Bristow KL (1991) Equation for extending water-retention curves to dryness.
Soil Sci Soc Am J 55:923–927
Rossi C, Nimmo JR (1994) Modeling of soil water retention from saturation to oven dryness.
Water Resour Res 30:701–708
Saguy IS, Marabi A, Wallach R (2005a) Liquid imbibition during rehydration of dry porous foods.
Innov Food Sci Emerg Technol 6:37–43
Saguy IS, Marabi A, Wallach R (2005b) New approach to model rehydration of dry food
particulates utilizing principles of liquid transport in porous media. Trends Food Sci Technol
16:495–506
Saguy IS, Troygot O, Marabi A, Wallach R (2007) Rehydration modeling of food particulates
utilizing principles of water transport in porous media. Bangkok: 10th International Sympo-
sium on the Properties of Water (ISOPOW X). (Also Chapter 17. In: Reid DS, Sajjaanantakul T,
Lillford PJ, Charoenrein S (eds) Water properties in food, health, pharmaceutical and
biological systems: ISOPOW 10). Wiley-Blackwell, 2010)
Salvatori D, Andres A, Chiralt A, Fito P (1999) Osmotic dehydration progression in apple tissue II:
generalized equations for concentration prediction. J Food Eng 42:133–138
Saravacos GD, Maroulis ZB (2001) Transport of water in food materials. Transport properties of
foods. Marcel Dekker, New York
Shi XQ, Fito-Maupoey P (1994) Mass transfer in vacuum osmotic dehydration of fruits:
a mathematical model approach. Lebensm Wiss Technol 27:67–72
Simunek J, van Genuchten MTh, Sejna M (2005) The HYDRUS-1D Software package for
simulating the movement of water, heat, and multiple solutes in variably saturated media.
[Version 3.0] Department of Environmental Sciences, University of California, Riverside
Singh PP, Maier DE, Cushman JH, Haghighi K, Corvalan C (2004) Effect of viscoelastic relaxa-
tion on moisture transport in foods. Part I: Solution of general transport equation. J Math Biol
49:1–19
Tillotson JE (2003) Convenience foods. In: Trugo L, Finglas PM (eds) Encyclopedia food sciences
and nutrition, 2nd edn. Academic, Oxford
van Genuchten MTh (1980) A closed-form equation for predicting the hydraulic conductivity of
unsaturated soils. Soil Sci Soc Am J 44:892–898
Wallach R (2007) Physical characteristics of soilless media. In: Raviv M, Lieth JH (eds) Soilless
culture: theory and practice. Elsevier, London
Washburn EW (1921) The dynamics of capillary flow. Phys Rev 17:273–283
Webb SW (2000) A simple extension of two-phase characteristic curves to include the dry region.
Water Resour Res 36:1425–1430
Weerts AH, Lian G, Martin D (2003a) Modeling rehydration of porous biomaterials: anisotropy
effects. J Food Sci 68:937–942
Weerts AH, Lian G, Martin DR (2003b) Modeling the hydration of foodstuffs: temperature effects.
AIChE J 49:1334–1339
Chapter 22
Responses of Living Organisms to Freezing
and Drying: Potential Applications in Food
Technology
Marı́a del Pilar Buera
22.1 Introduction
Some living organisms can survive under extreme stresses by adapting to situations
that would otherwise be lethal. They develop various mechanisms for adaptation,
surviving adverse environmental conditions such as lack of water (cryptobiosis
or anhydrobiosis) and freezing temperatures (cryobiosis) (Carpenter et al. 1986;
Crowe et al. 1998, Watanabe et al. 2002). Since water is required to hydrate
molecules (e.g., folding of proteins into active molecules cannot proceed without
water), and their macromolecular structure and functionality are sensitively deter-
mined by their interactions with water, in many cases the stress is governed by
interference from interaction with biomolecules (Franks 1982, 1995). At cold
temperature, water freezes and forms disruptive ice crystals that can damage
cellular structures. In conditions of low humidity or high osmolality, water leaves
the cells, thus altering the hydration of macromolecules and their ability to partici-
pate in reactions necessary for life.
Various strategies can protect cells against extreme temperatures and dehydra-
tion, enabling their survival. Since these tolerant organisms face the same problems
of protection during food preservation as their important biomolecules do, by
studying the strategies used in nature, innovative procedures can be developed for
use in food technology. By analyzing and interpreting these mechanisms, systems
and processes can be engineered to preserve food quality during processing and to
extend product shelf life (Aguilera and Karel 1997).
M. del Pilar Buera

Departamentos de Industrias y de Quı́mica Orgánica, Ciudad Universitaria, 1428 Ciudad de
Buenos Aires, Argentina
e-mail: pilar@di.fcen.uba.ar
554 M. del Pilar Buera
22.2 Desiccation Strategies: Glass Formation and

Solute-Protecting Interactions in Anhydrobiotes
Glass formation is a natural mechanism for the preservation of complex anhydro-

biotic organisms, which can tolerate desiccation and survive extended periods of
time in the dry state (Crowe et al. 1998). In response to dehydration the cytoplasm
of desiccation-tolerant organisms forms glasses; these organisms contain large
amounts of soluble nonreducing sugars and their state diagrams resemble those of
simple sugar mixtures. Some of the organisms (e.g., Artemia salina, yeast cells, and
tardigrades) accumulate a, a –trehalose, while others (pollen, seeds and resurrec-
tion plants, e.g., Selaginella lepidophyl) accumulate sucrose and other nonreducing
b-furanosides (e.g., raffinose, stachyose, and verbascose). Other examples of living
systems that exploit the formation of glass to preserve life are some species of the
simple flatworm (Tunnacliffe and Lapinski 2003).
The insect Polypedilum vanderplanki Hint. is the largest multicellular animal
known to tolerate almost complete dehydration without deterioration (Watanabe
et al. 2002). Their anhydrobiotic larvae show extremely high thermal tolerance
(270 C to þ102 C) and can recover soon after prolonged dehydration up to
17 years (Hinton 1960). Watanabe et al. (2002) showed that rapid accumulation
of trehalose (up to 18% of dry body mass) plays a key role in the successful
induction of anhydrobiosis of these larvae and that it is possible to store their
individual organs, and cells from them, at room temperature under dehydrated
conditions.
Crowe et al. (1998) reported that vitrification of the structure is necessary to
improve enzyme and liposome stability, but specific hydrogen-bond interactions
between sugars and the biomaterial are also needed. Besides forming glasses, in
which kinetic restrictions for physico-chemical changes such as chemical reactions
and crystallization operate, sugars develop the ability to protect proteins and
membranes through hydrogen bond interactions with the active biomolecules.
Dried yeast cells have been observed to have glass transition temperatures (Tg)
and water sorption isotherms that are independent of the amount of trehalose.
However, their viability was dramatically dependent on the quantity of disaccharides
present (Cerrutti et al. 2000).
Desiccation-sensitive organisms, on the other hand, generally lose their viability
during drying at water contents at which the glassy state has not yet been formed
(Buitink and Leprince 2004). Besides, sugars, proteins (dehydrines), salts, and
amino acids are accumulated by organisms resistant to dehydration; they act in
different ways, stabilizing proteins and membranes, contributing to osmotic adjust-
ment, or act as free radical scavengers.
The potential applications of intracellular disaccharides in animal cell biopre-
servation are limited by the inability of mammalian cells to synthesize or actively
accumulate sugars such as trehalose (Holovati and Acker 2007). The use of lipo-
somes as a permeabilization strategy for the intracellular delivery of trehalose
has been investigated as a useful cell delivery vehicle for membrane-impermeant
22 Responses of Living Organisms to Freezing and Drying 555
biologically active compounds, including drugs, enzymes, hormones, and genetic

material, and also for improved cell recovery following low-temperature exposure
(Holovati and Acker 2007).
22.3 Survival in Frozen Environments: Managing the Kinetics

of Ice Nucleation or Ice Crystal Growth
Many organisms exist in habitats where temperatures fall below the freezing point
of water. Intracellular ice formation changes the chemical environments of bio-
molecules. Particularly, pH change and ionic force increase may be a cause of protein
and nucleic acid damage (Franks 1982). In parallel, injuries due to mechanical
changes affect the systems: large ice crystals disrupt tissue structure and thus
regulation of ice crystal growth may be important for the survival of these organ-
isms in winter. Some plants and animals have evolved to prevent the lethal effects
of ice crystal formation in cells. The mechanisms by which organisms survive
extreme low temperatures are (1) extracellular water crystallization (formation of
crystals outside cell membrane), (2) use of antifreeze proteins (AFPs) or by the
formation of a glass (Clark et al. 2007).
One of the most studied groups of organisms is the Collembola (arthropods or
springtails) (Clark et al. 2007). In many freeze-tolerant insects, very potent ice
nucleators are present in the hemolymph. Among vertebrates, hatchling painted
turtles (Chrysemys picta) provide another example of “natural freeze-tolerance”
(Packard and Packard 2004; Costanzo et al. 2008). In freeze-tolerant species,
proteinaceous ice nucleators (INAs) trigger extracellular freezing at high subzero
temperatures, either to provide cold protection from released heat of fusion or to
establish a protective extracellular freezing, which drives water out of the cells,
further decreasing the temperature at which intracellular ice forms (Costanzo et al.
2008).
Alternatively, some organisms will die if frozen and avoid freezing by main-
taining their body fluids in the liquid state even at extremely low subzero tempera-
tures. Freeze-avoiding species increase their supercooling potential by removing
ice nucleators and accumulating polyols. Terrestrial invertebrates and polar marine
fish stabilize their supercooled state by means of noncolligatively-acting AFPs.
Fish, insects, and some plants that live in Arctic regions have evolved to produce
AFPs, which inhibit the growth of ice crystals by adsorption to the ice surface
(Zachariassen and Kristiansen 2000). These organisms (specifically insects) also
have the mechanisms to inactivate INAs, minimizing the risk of inoculation by ice
and INAs. Antinucleating protein of bacterial origin inhibits the fluctuation of ice
nucleus formation via foreign particles (Kawahara 2002).
Many nonfreezing tolerant arthropods and insects use freeze avoidance, but
others, such as the Arctic springtail Onychiurus arcticus or some insects, use the
strategy of protective dehydration (Clark et al. 2007). In protective dehydration,
loss of water occurs across a diffusion gradient between the super-cooled cell fluids
and ice in its surroundings, such that freezing point depression always exceeds the
environmental temperature, and eventually the organisms lose sufficient water and
desiccate to ensure that a freezing event cannot occur. Studies have shown with
the springtail O. Arcticus or the freshwater snail, P. canaliculata, that exposure to
subzero temperatures and low water vapor pressure induces extensive dehydration
through a highly permeable cuticle; in combination with rapid synthesis and
accumulation of membrane/protein cryoprotectant trehalose and some low molec-
ular weight compounds, cold hardiness is enhanced (Clark et al. 2007; Matsukura
et al. 2008). As a result, water in the cells freezes at 40 C (Zachariassen et al.
2004).
Table 22.1 summarizes the main strategies that allow some living organisms
to survive adverse conditions, showing representative examples of each group. It
should be noted that some species can switch between freeze avoidance and freeze
tolerance, depending on prevailing physiological and environmental conditions
(Costanzo et al. 2008).
22.4 Some Theoretical Considerations
Before interpreting different strategies depicted in previous sections, some theoret-

ical concepts should be introduced. Removal of water, either by drying or freezing,
induces supersaturation of cytosolic components, leading to an increase in cohesive
forces between molecules and restriction of molecular mobility within cytoplasm.
Although both drying and freezing are considered to promote damage by hydric
stress, different protective mechanisms operate in each process (Carpenter et al.
1986).
The strategies for avoiding the above-mentioned injuries caused by dehydration
and freezing can be analyzed through simplified temperature-composition state
diagrams (shown schematically in Fig. 22.1), in arbitrary units, but they are
generically designed by approximating the composition of cell cytosol, similar to
that also observed for seeds and different tissues (Levine and Slade 1992; Buitink
and Leprince 2004; Espinosa et al. 2006; Matiacevich et al. 2006).
In such supplemented diagrams the curves corresponding to equilibrium condi-
tions (liquidus, Tm, and solubility, Ts curves) and Tg of systems are plotted as a
function of water content (w). Since different curves in Fig. 22.1 delimit regions
where main dynamic changes could happen as a consequence of phase/state
changes (Levine and Slade 1992), this kind of diagram can predict whether systems
are under thermodynamic or kinetic control, for given composition-temperature
conditions, provided thermal history of samples is known (Levine and Slade 1986;
Roos and Karel 1991a; Levine and Slade 1992). Of particular interest for analysis of
involved phenomena are the formation of glasses and freezing characteristics of
systems.
22
Table 22.1 The main strategies allowing some living organisms to survive adverse conditions; involved mechanisms and representative examples
Type of organism Mechanisms Action Involved solutes Examples
Anhydro-biotes Glass formation at Reduce molecular mobility, Protective non- Trehalose Artemia salina, yeast cells
low water content protective hydrogen reducing sugars (Saccharomyces cerevisiae),
bond interactions Proteins: dehydrines tardigrades, flatworm, insects
(larvae of chironomid
Polypedilum vanderplanki),
rotifers (Philodina roseola);
resurrection plants (Selaginella
spp.)
sucrose, raffinose, pollen, orthodox seeds, resurrection
stachyose plants (Selaginella lepidophyl)
Freeze-tolerant Promote extracellular Decrease supercooling, Adaptive INAs Insects, arthropods or springtails
water crystallization increase ice nucleation (snow flea), vertebrates
temperature, decrease (hatchling painted turtles)
freezing time
Freeze-avoiding Inhibit growth of ice Increase supercooling, AFPs (non-colligative) Polar marine fishes, terrestrial
Responses of Living Organisms to Freezing and Drying
crystals by decrease freezing invertebrates, insects,

adsorption to ice temperature arthropods, nematodes, plants,
surface fungi
Inactivation of ubiquos Inhibit fluctuation of ice Proteins, polymers Antarctic bacteria (Pseudomonas
INAs nucleus formation by a fluorescens)
foreign particle
Vitrification at high Reduce molecular mobility Low molecular weight compounds: polyols freshwater snail (P. canaliculata),
water content fly (Eurosta solidaginis), frog
(Rana sylvatica)
557
Ts
100
Temperature A
Tm
0
Tg’1
Tg’2
–100 Tg
B
w1’ w2’
0 0.25 0.5 0.75 1
w
Fig. 22.1 Supplemented state diagram showing the equilibrium curves (solubility, Ts; ice melting,
Tm) and nonequilibrium glass transition curve (Tg). The main expected changes are indicated, as
are the mechanisms for avoiding damage (see text)
Amorphous glasses are formed by a continuous process in which no interface or

solidification front is involved. Their structure is comparable to that of a liquid
(some short-range order is observed, but no long-range order), but their properties
are those of a solid. Glasses are characterized by a critical temperature Tg, above
which most physical changes (including solute crystallization) result from sharp
increase in molecular mobility, which occurs above their Tg (Levine and Slade
1986; Roos and Karel 1991a, b, c; Slade and Levine 1991). As shown in Fig. 22.1,
Tg of a material is a function of the relative proportion of its glass-forming
components and water content. Inclusion of Tg curve in supplemented state dia-
grams (corresponding to nonequilibrium transition) allows including the concept of
time in which a certain event will take place or changes, which will be kinetically
delayed/inhibited under certain conditions.
In region below Tm, as ice separates out during freezing, solute concentration of
unfrozen phase in contact with ice increases. The combined effects of lowering
temperature and increasing concentration impose molecular mobility restrictions on
systems. Concentration of the unfrozen phase increases as temperature decreases.
The point at which the liquidus curve, Tm, intercepts Tg curve defines the point
(Tg0 ; wg0 ), which represents the particular temperature/solute concentration below
which water crystallization is kinetically inhibited. wg0 represents the lowest mass
fraction of water that can be obtained by solute concentration in amorphous matrix
due to ice crystallization. Tg0 is the glass transition temperature of maximum ice-
concentrated solution (Fig. 22.1). It should be noted that material composition will
affect Tg0 and wg0 , thus affecting the amount of water remaining unfrozen in the
matrix (wg01 and wg02 in Fig. 22.1) and temperature range at which the process
should be performed (Fig. 22.1). Due to predominantly nonequilibrium conditions
of naturally-occurring frozen or dehydrated systems, they are subjected to dynamical
(time-dependent) changes. Thus, the extent of damage will be related to the
mentioned variables, to the rate and temperature of cooling, and should be taken
into account in the design of processes and system formulation.
Freezing is a process of ice crystallization from supercooled water. In this
process, water should undergo a stage of ice nucleation, followed by growth of
ice. Nucleation is the initial stage and one of the most important steps toward
creating ice. Without this step, ice would never occur in supercooled water.
Nucleation can be regarded as a kinetic process for ice nuclei to overcome a kinetics
barrier, the so-called nucleation barrier under a given thermodynamic driving force,
which is proportional to the supercooling.
Ice growth is thermodynamically favored in water at temperatures below 0 C.
However, ice nucleation is not kinetically favored in pure water and can remain
liquid up to nearly 40 C, the homogeneous nucleation temperature (Th), if free of
ice nucleating species. A supercooled system is that in which no crystallization has
occurred, even if it is below the liquidus curves (Ts and Tm in Fig. 22.1). Ice
nucleation at temperatures greater than Th is induced by heterogeneous INAs
(Wowk and Fahy 2002). Almost all organic and inorganic substances can catalyze
ice formation (i.e., serve as ice nuclei) at temperatures between 15 C and 40 C.
The degree of supercooling is the thermodynamic impulsive force for crystalli-
zation, defined as:
DT ¼ Tm T (22.1)
where T is actual subzero temperature and Tm is equilibrium melting point of ice.

The transient ice-like embryos formed by aggregation of water molecules is
subjected to continuous fluctuation in size due to incorporation of new molecules
and detachment of others. For a given temperature, there will be a critical radius
that defines the minimum size a nucleus can have to be a stable crystal. Heneghan
et al. (2002) defined the supercooling point, which is often approximated in
biological studies as the temperature at which, on an average, 50% of samples are
frozen. The supercooling capacity of an organism, also frequently mentioned in
literature, is thus inversely related to this supercooling point: the lower the super-
cooling point (i.e., the lower the ice crystallization temperature), the higher the
supercooling capacity of an organism. Supercooling is limited by heterogeneous
nucleation in the presence of solid impurities. Certain compounds can serve as
nuclei at temperatures as high as 6 C, and the wide variety of impurities present
in natural systems makes homogeneous nucleation impossible. Ice nucleation
preferentially occurs on these impurities, such as mold walls or cell membranes
(nonadaptive INAs); since they are points with high excess energy, they lower the
energy required to form the interface between the existing phase and new phase.
Removing heterogeneities is one effective way of decreasing temperature at which
ice forms, i.e., increasing difficulty of freezing. Some proteins have the function of
inhibiting nucleating activities of nonadaptive INAs. The critical radius decreases

with decreasing temperature below Tm, and the rate of nucleation would increase
with temperature below Tm. This effect is limited by decrease in molecular mobility
at lower temperatures; the curves showing actual variation of nucleation frequency
with temperature, crystal growth, and overall crystallization are shown as full lines
in Fig. 22.2a, b.
Once initiated, crystal growth is slowed and often terminated with increasing
viscosity (MacKenzie 1977). Because the free energy barrier of three-dimensional
a
+ INAS
Overall
crystallization
Rate
Tg Temperature Tm
b
AFP
Overall
crystallization
Rate
Tg Temperature Tm
Fig. 22.2 Schematic curves showing ice nucleation and crystal growth rates, and overall crystal-
lization rate ( full lines) and their potential modification (dotted lines) by the presence of (a) ice
nucleating agents (INAs) or (b)antifreeze proteins (AFPs)
nucleation is much higher than that of growth, in the case of ice crystals, growth
normally becomes much easier than nucleation: once ice nuclei are formed, the
rapid growth rate leads to instant freezing (Du et al. 2003). To control nucleation
systematically, the upper size needs to be controlled.
22.5 Involved Mechanisms
22.5.1 Avoidance of Solids Crystallization in Supercooled State
Anhydrobiotes are dehydrated systems that have the ability to form glasses and can
be located in the low w region of Fig. 22.1. Amorphous glasses are metastable
materials in a nonequilibrium state, as these systems are located below the satura-
tion curve, Ts, in the temperature-composition state diagram (Fig. 22.1), in which
the stable form is crystal. However, in the glassy state, most structural changes
occur very slowly and only small motions of molecules, mainly rotational motions
of side chains and vibrations, may occur (Levine and Slade 1986; Slade and Levine
1991). Thus, crystallization of amorphous solids is kinetically delayed if sample
remains in the glassy state, below Tg. Since direct interaction of sugars with
molecules imbedded in the glassy matrix will prevent protein denaturation and
lead to optimal preservation (Arakawa and Timasheff 1982), it is important that
sugars remain in a noncrystalline state, either glassy or supercooled. It should be
noted that when solute crystallization occurs, composition of the noncrystalline
phase changes, and often an increase of deteriorative reaction rates is observed
(Buera et al. 2005).
In the case of seeds, embryos and plant tissues, presence of several oligosac-
charides has been related to seed longevity due to protection given to proteins and
membranes, and to prevention of sucrose crystallization (Obendorf 1997; Murthy
et al. 2003).
The influence of different yeast (e.g., Saccharomyces cerevisiae) cellular frac-
tions was studied in an attempt to gain knowledge on the feasibility of trehalose
crystallization in yeast cells. Certain constituents of S. cerevisiae cells inhibited/
delayed trehalose crystallization upon humidification at high R.Hs. (Espinosa et al.
2006). This reveals that cellular structures naturally have components that act in
delaying or inhibiting sugar crystallization in the interior of cells. Although crystal-
lization of sugars in the cytoplasm has been considered as a cause of viability loss, no
evidence of sugar crystallization in cells or tissues was ever reported (Buitink and
Leprince 2004; Espinosa et al. 2006; Matiacevich et al. 2006). In fact, sugar
crystallization has not been evident in any published, seed, fruit, or tissue thermo-
gram during rewarming of systems from below to above glass transition temperature.
The effect of a mixture of several sugars on crystallization and stability of
embedded enzymes was analyzed in freeze-dried model systems. In trehalose-
lactose systems, the time to crystallization of lactose increased when trehalose
was added (Mazzobre et al. 2001). The addition of raffinose also had a retarding
effect on sucrose crystallization (Buera et al. 2005). The onset temperature for
trehalose or sucrose crystallization, determined by DSC, increased when raffinose
or lactose was added and was even avoided during the run if enough of the second
sugar was present (Mazzobre at al. 2001). Buera et al. (2005) showed that the
presence of proteins delayed sugar crystallization and, in parallel, sugars retarded
protein denaturation. Santagapita et al. (2008) have also reported delay of trehalose
crystallization with the presence of polymers or salts. Because of this, when simple
sugars or polyol matrices are designed to protect biomolecules, a common practice
in pharmaceuticals or food ingredient formulations, a second excipient is needed.
22.5.2 Increase of Extracellular Ice Nucleation Rate: INAs
Many biological systems promote extracellular heterogeneous nucleation by the

presence of a variety of INAs, which may be either incidental or adaptive. Inciden-
tal INAs (e.g., cell walls or dust particles) promote heterogeneous nucleation only
as a normally unwanted side effect. Adaptive INAs are generally lipoproteins of
about 30 kD; they reduce supercooling capacity by increasing supercooling tem-
perature to as little as 1 C, nucleating ice crystals between cells. The amino acids
within proteins form templates for ice, which serve as embryos for ice formation.
Among the most efficient INAs in nature are ice nucleating proteins found on the
surface of certain bacteria, such as Pseudomonas syringae, Erwinia herbicola, and
Pseudomonas fluorescens (Kawahara 2002), commonly found on plant leaves,
and other above-ground plant parts. Such proteins must assume a rigid, ice-like con-
formation larger than 10 nm and be able to aggregate. A thin layer of ice can always
form on the surface of an INA. However, this will not lead to spontaneous ice
growth unless the INA is of a certain critical size above which free ice growth will
occur. Consequently, a larger INA gives a smaller required supercooling and a
higher Tn. As they have the tendency to form protein aggregates of large spatial
extent (larger than 10 nm) (Kajava and Lindow 1993), they act as strong heteroge-
neous ice nuclei in dewdrops, and ice crystals thus formed grow and break plant
tissues, causing frost injury to host plants (Lindow 1983). The effects of INAs on
nucleation kinetics, and further crystal growth, are schematized in Fig. 22.2a. It can
be seen that supercooling point is decreased in the presence of INAs and conse-
quently, supercooling capacity of organisms is decreased.
In natural settings, inhibition of bacterial INA is of considerable interest for
environmental ice control, particularly in the prevention of frost damage to crops
(Wowk and Fahy 2002). Ingestion of these ice-nucleating bacteria caused an abrupt
decrease in supercooling capacity of the Colorado potato beetle (Leptinotarsa
decemlineata Say), a freeze-intolerant species that overwinters as adults in shallow,
terrestrial burrows (Costanzo et al. 1998). The ice-nucleating-active bacterium
Pseudomonas syringae was employed as pest control in grain silos, to increase
mortality of the granary weevil Sitophilus granarius (L.) and saw-toothed grain
beetle Oryzaephilus surinamensis (L.) (Mignon et al. 1998).
22.5.3 Inhibition of Ice Crystal Growth: AFP
Freeze avoidance is often achieved by a high capacity for supercooling, minimizing

the risk of inoculation by ice and INAs. At low temperatures, many plants and
microorganisms produce colligative protectants such as proline or sucrose. How-
ever, some organisms are also capable of producing AFPs and antifreeze glycopro-
teins (AFGPs), protecting them from the negative effects of freezing. AFPs and
AFGPs were first identified in 1970 in Antarctic fish varieties that can survive in
seawater colder than the freezing temperature of their blood; AFPs and AFGPs
were then found in microorganisms, insects, plants, and nematodes. Their molecu-
lar weight ranged between 2,600 and 33,000 Da. Although they were thought to act
by colligative effect, decreasing melting point of water, it was later observed that
they act by avoiding growth of ice crystals, having a low effect on colligative
properties.
AFPs act at concentrations in the order of 105 M and inhibit ice crystal growth
by adsorption-inhibition. The nuclear magnetic resonance, x-ray structure, and
many spectroscopic studies with AFPs have been helpful in determining the
structure-function relationship. Mutational studies have indicated the importance
of hydrophobic residues in ice binding (Venketesh and Dayananda 2008). Ice
crystals can grow along six a-axes, all in the same plane, or along the c-axis,
which is perpendicular to the plane of the six a-axes. Ice crystal growth at tem-
peratures close to Tm typically occurs along the a-axes, which accounts for the
typical hexagonal shape of snowflakes. The c-axis growth results in needle-like ice
crystals, which are potentially damaging (Davies and Hew 1990). The morphology
of ice crystals will depend on the planes in which growth inhibition is performed by
AFP. AFPs typically found in Arctic fish specifically adsorb to basal ice (a-axis),
growing planes and avoiding propagation, and behave as “structural ice inhibitor
agents” (Pertaya et al. 2007a, b); ice crystal growth occurs in bipyramidal shape. By
contrast, AFPs from Arctic insects typically inhibit c-axis growth, forming brown
flat discs.
All AFPs tested to date show both antifreeze activity and inhibition of ice
recrystallization, suggesting a common mechanism for these two effects mediated
through ice binding. To visualize the binding of AFP to ice, Pertaya et al. (2007a, b,
2008) labeled the spruce budworm AFP with enhanced green fluorescent protein
(GFP) and observed the AFP-GFP molecules directly on ice crystals using confocal
microscopy. Melt-growth-melt sequences in low concentrations of AFPs revealed
that the same general behavior of an apparently rotated crystal, observed in pure ice
under high pressure, is reproduced in ice under the influence of AFP at ambient
pressure and temperatures near 0 C.
While AFPs lower the freezing point, the melting point is unaltered. The
separation of melting and freezing temperature is usually referred to as thermal
hysteresis (TH), and temperature of ice growth is referred to as hysteresis freezing
point. The hysteresis is the result of an adsorption of AFPs to crystal surface. This
causes ice to grow as convex surface regions between adjacent adsorbed AFPs, thus
lowering temperature at which the crystal can visibly expand (Kristiansen and
Zachariassen 2005). Within the resulting TH gap, ice crystals appear to be kineti-
cally stable, neither growing nor melting. The level of TH is directly related to
intrinsic activity of specific AF(G)P and to their concentration. Results by Evans
et al. (2007) showed that when AF(G)P are dissolved in salt solutions, such as NaCl,
at concentrations they could encounter in nature, there is a synergistic effect on TH
that is positively related to salt concentration. This enhancement could have resulted
from the hydration shell of dissolved ions, which reduces freezable water. Alterna-
tively, salt could influence the hydration shell surrounding AF(G)P, increasing
protein surface area available to adsorb to ice/water interface (Evans et al. 2007).
Five AFP types have been isolated from fish, six from plants, and two from
insects; five bacteria produce AFPs and show no related structures, varying in
protein size and specific activity (Davies and Hew 1990; Brush et al. 1994; Davies
et al. 2002; Gilbert et al. 2004). The AFPs found in certain insects (which allow
them to survive winters at temperatures as low as 30 C) are up to 100 times more
powerful than similar proteins in fish (Graether et al. 2000). Although TH activity
of plant AFPs is low compared to that of fish and insects, their inhibition of ice
recrystallization is comparable or even greater (Smallwood et al. 1999). The effect
of AFP on kinetic control of ice crystallization is shown schematically in Fig. 22.2b
(dotted lines).
22.5.4 Vitrification at High Water Content (Liquid N and/ or with

Concentrated Solutes)
The alternative route to remove ice crystallization is rapid cooling into the glassy
state, or vitrification (line A–B in Fig. 22.1). Vitrification as a cryopreservation
method has many primary benefits, such as no ice crystal formation through
increased speed of temperature conduction, which provides a significant increase
in cooling rates. This can eliminate structural damage during freezing, but mass
transfer rates limit the sample size in which this can be achieved and thawing is
never rapid enough to prevent crystallization from causing extensive damage
(Lillford and Holt 2002). It should be noted that the maximum ice nucleation occurs
just above Tg and the maximum ice-crystal growth-rate occurs just below Tm
(Fig. 22.1). Thus, the many nuclei formed when cooling can cause massive ice
growth when rewarming. This leads to devitrification. The main concern with
cooling is the maximum nucleation range near Tg. The addition of a cryoprotectant
may allow cooling up to the Tg region without forming any nuclei, if cooling is
quick enough. If nuclei are formed, the rapid cooling and high viscosity near Tg in
the presence of cryoprotectant will not allow the nuclei to grow very much and thus
prevents them from being harmful. Increasing cryoprotectant concentrations lowers
both Th and Tm, but the effect is more dramatic on Th than Tm. Enough cryoprotec-
tant to lower Tm will lower temperature of nucleation to Tg, thereby delaying
nucleation. Northern frogs use glucose as a cryoprotectant (Mietchen et al. 2008).
When temperatures drop, the liver of these frogs produces large amounts of
glucose, which a special form of insulin allows to enter into their cells in large
quantities. Except for the heart and brain, much of the frog’s body freezes. The two
disaccharides that most protect proteins and cell membranes against chilling,
freezing, and dehydration are sucrose and trehalose.
Dissolved polymers form glasses readily once they are freeze-concentrated.
Below Tg, no further ice forms, and crystal growth rates and sintering of ice are
kinetically slowed to almost zero. The damage caused by the amount of ice present
and temperature at which it is stabilized against growth and/or recrystallization
should be balanced (Lillford and Holt 2002). Increasing speed of thermal conduc-
tion and decreasing concentration of cryoprotectant are challenges to overcome
with vitrification methods. The convenience of vitrification could push develop-
ment of this technique beyond most presently common clinical uses for embryos
and tissue preservation (Kattera and Chen 2006).
22.5.5 The Problem of Recalcitrant Seeds
Orthodox seeds undergo a programmed desiccation at termination of their develop-

ment. Vitrification of cytoplasm components in orthodox seeds is proposed to be
advantageous for germplasm stability, and their transitions may affect seed viability
(Walters 2004). The cytoplasmic glass of seeds is composed mainly of sugars
(sucrose and oligosaccharides comprise over 10–20% of dry weight), high molecu-
lar weight oligosaccharides, and proteins. As storage temperature or water content
increases, seeds undergo glass-to-liquid transition (Tg), resulting in an increase in
molecular mobility. As Tg is a function of water content, whether seed tissues are in
the glassy state or not, it will depend on both seed water content and storage
temperature (Buitink and Leprince 2004; Matiacevich et al. 2006). Investigations
have led to the suggestion that seeds induce glass formation as water content is
depressed, thus limiting deteriorative reactions (Murthy et al. 2003). The seed’s
tolerance to desiccation (orthodox) can be attributed to the presence of the soluble
carbohydrates raffinose and sucrose in the cytoplasmasm of embryo cells. In
contrast, recalcitrant seeds (Araucaria angustifolia, palm) do not tolerate a reduc-
tion in water below a relatively high level without loss of viability. Conventional
storage techniques are thus not applicable to these seeds and cryopreservation is the
only feasible alternative for their long-term storage. Panza et al. (2006) have
observed that at necessary water contents (or relative humidities (R.Hs.)) at
which embryos of recalcitrant Araucaria angustifolia seeds retain their viability (at
and above R.H. 85%), water freezes upon cooling at subzero temperatures. Their
high-freezable water content promotes injuries, and their conservation at low-
temperatures represents a challenge. Thus, recalcitrant seeds are good models for
analyzing the impact of freezing rate, storage time, and temperature on degree of
injury. If plant tissues are exposed to rapid cooling only minor damage occurs
(196 C). Seeds with lower amounts of frozen water would be more easily
cryopreserved. The preservation of recalcitrant seeds could be improved by vitri-
fication, either by dehydration, after imbibition of embryos with protecting agents,
or by freezing, avoiding ice formation, but this is an area of further research, of
interest in biodiversity preservation, for example.
22.6 Applications in Food Technology
The implications of glass formation and glass transitions in food technology have
been extensively analyzed since the pioneering work by Levine and Slade in the
1980s (Levine and Slade 1986, 1992), based on polymer science approaches.
Afterwards, many scientific publications evolved (Roos 1995). Parallel advances
in the area of preservation mechanisms of living organisms under extreme condi-
tions promoted the study of sugar properties in relation to their protective effects on
labile biomolecules (Arakawa and Timasheff 1982; Crowe et al. 1998; Clegg
2001). Fundamental research areas were also focused on dynamic aspects related
to glass-forming properties of sugars, and on molecular interactions that sugars are
capable of developing in hydrogen bonding.
According to the discussed aspects, possible strategies to avoid sugar crystalli-
zation for food and ingredient formulation in low water content systems may
include vitrification or sugar crystallization delay in supercooled media, which
can be achieved by combinations with biopolymers, salts, or other sugars.
Recently, Leinen and Labuza (2006) and Belcourt and Labuza 2007) successfully
suppressed sucrose recrystallization by addition of raffinose to the formulation
of cotton candy and soft cookies, respectively, which consequently improved
technological properties of products.
The efficiency of freezing and resulting food quality is affected by two important
factors: supercooling (cooling of liquid below its freezing point, without freezing)
and nucleation (initiation of crystallization of liquid water into solid ice) (Li and
Lee 1995). Both phenomena are modified by the presence of two kinds of proteins
showing opposite behavior (Hew and Yang 1992; Li and Sun 2002), which provide
enormous technological potential in industrial processes (Lillford and Holt 2002).
INAs, on the one hand, decrease supercooling and increase nucleation tempera-
ture (Tn), thus decreasing freezing time and forming a large number of ice crystals
with a dendritic structure (Feeney and Yeh 1998; Li and Lee 1995; Li and Sun
2002). AFPs, on the other hand, increase supercooling, generate very small ice
crystals, lower freezing temperature, and retard recrystallization in frozen storage.
Lillford and Holt (2002) and Li and Sun (2002) have reviewed the main
potential applications of INAs and AFPs. The beneficial effects of INAs are
manifested in decreasing energy cost and crystal size, favoring quality improve-
ments and increased shelf life. Bacterial ice nucleators were employed in the
freezing of meat (Payne et al. 1994), and in products difficult to concentrate or
freeze-dried products, such as fruit juices or pastes (Li and Lee 1995). The
management of ice nucleation and crystallization times can also be employed
for the development of new products. For example, freeze-texturing of protein
products may generate foods with special characteristics (Watanabe and Arai
1994).
AFPs behave both as antinucleators and as growth inhibitors, preventing hetero-
geneous nuclei from growing (Lillford and Holt 2002). One of the most interesting
applications is the performance of AFP after ice is nucleated or in conditions where
ice crystals would normally grow and sinter (Griffith and Ewart 1995). The presence
of AFPs in these products may inhibit ice crystal growth at fluctuating subzero
temperatures during storage or transportation, thus preserving food quality and
decreasing operative costs.
AFPs are natural products that form part of a normal diet through their occur-
rence at high concentrations in food fish and vegetables (Smallwood et al. 1999).
Simple natural extracts that are purified or partially purified are likely to meet
consumer acceptance. The AFPs from vegetables have a number of potential
applications. For example, Zhang et al. (2007a) improved the texture of frozen
dough and the effect of volatile compounds in crumb foods by the addition of a
concentrated carrot extract containing carrot AFP. AFP is efficient at relatively low
concentration, but the cost of extraction from natural sources restricts their use only
to applications with very high added value. Strategies for improving AFP produc-
tion from natural sources can be developed. Naturally-occurring “antifreeze” has
been found in carrots, winter wheat, and rye (Li and Lee 1995). Desjardins et al.
(2007) showed that the thermal regimen of wolffish can be manipulated to enhance
their tolerance to subzero temperatures and also their ability to produce AFP.
Abundant AFPs could also be generated as by-products in processing certain fish
species for food (Griffith and Ewart 1995).
Alternatively, AFP genes can also be introduced into microorganisms used to
produce frozen yogurt (Yeh et al. 2008). Recent advances in biotechnology and
control of heterologous gene expression may make them potentially accessible
in high quantities. The multiple forms of AFPs synthesized within each organ-
ism, and demonstrated possibilities for useful modification through protein
engineering, make it feasible to choose the most appropriate AFP, with a
suitable level of activity for addition to a particular product or expression in a
particular plant or animal (Griffith and Ewart 1995; Li and Lee 1995; Lillford
and Holt 2002). Further, the ability to modify the rate and shape of crystal
growth and to protect cellular membranes during lipid-phase transitions have
resulted in identification of a number of potential applications of AFGPs as food
additives, and in the cryopreservation and hypothermal storage of cells and
tissues (Harding et al. 2003).
22.7 Future Prospects
The study of delay/inhibition of sugar crystallization in supercooled liquids at low

water content may increase the range of applications of sugar as excipients for food
ingredient formulation for specific purposes. Currently, relative long-term storage
of living cells in the high water content range can be done only in cryoprotectant
solution at extremely low subzero temperatures, and most of the time media
formulations are performed on an empirical basis. A standardized vitrification
protocol for cryopreservation on a scientific basis has yet to be defined, taking
into account the many variables that can profoundly influence its effectiveness
and potential to improve survival rates of vitrified cells (e.g., type and concentration
of cryoprotectant; temperature of vitrification solution at exposure; duration of
exposure to final cryoprotectant before plunging into LN2).
Liquid-to-solid nucleation of a supercooled aqueous solution seems to be a
simple phenomenon, but it remains a poorly understood process of evolution of
a metastable state to its final equilibrium state. Inputs from fundamental research
in this area are also needed (Heneghan et al. 2002). Careful design of experiments
for measuring temperatures and times of nucleation and crystal growth should be
developed (Chapsky and Rubinsky 1997; Heneghan et al. 2002; Du et al. 2003;
Zachariassen et al. 2004; Yin et al. 2005).
The presence of ice surfaces was reported as a cause of freeze-induced perturba-
tions of protein native fold, but these interactions are poorly understood. Gabellieri
and Stambrini (2006) reported that a binding method using a fluorescence probe
may find practical utility in testing the effectiveness of various additives employed
in frozen or freeze-dried protein formulations. Research work in this area would be
helpful in the development of “ice managing” additives. Besides, the quantitative
relationships between the activity of INAs and AFPs and variable parameters that
affect the measurement of activity should be established to elucidate the mechanism
and organize the research findings.
Fundamental research also on freeze tolerance mechanisms may allow identification
of secondary metabolic products that prepare organisms for stress from ice formation,
osmotic dehydration or freeze concentration; further, molecules that have applications
in industrial settings could be produced in large quantities once their functions are
elucidated (Feeney and Yeh 1998; Lillford and Holt 2002). The effect of proteins and
polysaccharides involved in INA inactivation should also be further investigated to
explore their potential applications (Kawahara 2002; Wowk and Fahy 2007).
Although many advances have been achieved in the last 10 years, additional
research is still needed to define the relationship between molecular structure and
the function of INAs and AFPs. The developments of infrared thermography,
microscopic and magnetic resonance spectroscopy, and imaging techniques for
cryobiological investigations were demonstrated as powerful tools for interpreta-
tion of mechanisms of action for different AFPs (Wisniewski et al. 2001; Pertaya
et al. 2007a, b, 2008; Mietchen et al. 2008), and are expected to further contribute to
advances in these areas.
The availability of ice nucleators or antinucleators for food uses involves

development of purification protocols of potential AFPs and INAs and effective
ways for their characterization (Kawahara et al. 2006; Zhang et al. 2007b). Besides
being nontoxic, nonpathogenic, and environmentally safe, they should be palatable
(Lillford and Holt 2002). Sensory evaluation should thus be performed on pro-
cessed products with these additives before being commercialized.
The chemical synthesis of AFPs is an attractive alternative to their difficult
isolation and purification from natural sources, and this would permit quality control
and mass production (Matsumoto et al. 2006). The successful synthesis of small
AFGPs using solution methods and solid-phase chemistry provides the opportunity
to perform key structure-activity studies that would clarify important residues
and functional groups required for activity (Harding et al. 2003). The possibility of
production using different molecular biological techniques, which will help increase
the yield, is also dealt with (Venketesh and Dayananda 2008).
Improving the efficiency of AFPs by the presence of simple molecules normally
present in foods, such as electrolytes (Du and Liu 2008), provides another new
insight into the antifreeze mechanism of AFPs on ice crystallization.
AFPs can display both protective and cytotoxic actions, and both nucleation of ice
and inhibition of ice crystal growth; they have also been shown to either stabilize or
disrupt the membrane, depending on several factors (membrane composition, type,
and concentration of AFP) (Wang 2000; Tomczak et al. 2001). At certain levels,
AFP-ice interactions may induce the complexes to aggregate, promoting ice nucle-
ation, and loss of ability to inhibit recrystallization. Although at low concentrations
AFPs can be employed very effectively to maintain texture and flavor of frozen foods,
it should be noted that bipyramidal or spicular ice crystals, which grow in solutions
containing high concentrations of AFP, are so detrimental for cell survival that they
are proposed as an important tool in cryosurgery (Matsumoto et al. 2006). These
properties indicate that careful consideration of concentrations and type of AFP used
require further study in order to understand the mechanisms by which they occur and
their potential usefulness in food technology.
The use of gene transfer to generate food organisms that produce AFPs needs to be
carried out. Since one important consideration is that of consumer acceptance of
transgenic products, transfer of genes from INA+ bacteria to other microorganisms
commonly used in production of foods (Yeh et al., 2008), or AFP genes from cereal
grains (e.g., winter wheat or rye) to plants for production of more freezing-resistant
fruits, is expected to be better accepted by consumers than the use of fish AFPs or
bacterial INAs (Lillford and Holt 2002). The addition of INAs or AFPs to food
products could involve significant modification in formulation and handling proce-
dures. Thus, a further analysis of phase/state diagrams is necessary before large-scale
food production.
The relatively high cost of freezing processes currently used in the food industry
and the search for products with special structures could be the major impulse
toward new innovations in freezing techniques. Additional work to analyze cost-
benefit, balancing the energy savings, and product-quality improvements, is essen-
tial before final commercialization of application in the food industry (Lillford and
Holt 2002). The discussed aspects are examples of how the tools from cryobiology
may stimulate those innovations.
It is interesting to note that the different aspects referred to in this chapter show the
synergistic interaction of multidisciplinary interfaces of physics, chemistry, medicine,
biotechnology, food science and technology, pharmacy, and biology, which are faced
with preservation of living organisms such as embryos and cells, structures like
liposomes, as well as labile molecules, such as enzymes, antibodies, and hormones.
Acknowledgments The author acknowledges the invaluable help of her collaborators Drs.
Florencia Mazzobre, Carolina Schebor, Beatriz Elizalde, Silvia Matiacevich, Nuria Acevedo,
Patricio Santagapita, Abel Farroni, and Nora Gutiérrez. Financial support by CONICET–PIP
5977, ANPCyT–PICT 20545, and UBACYT X024 is also acknowledged.
References
Aguilera JM, Karel M (1997) Preservation of biological materials under desiccation. Crit Rev
Food Sci Nutr 37:287–309
Arakawa T, Timasheff SN (1982) Stabilization of protein structure by sugars. Biochemistry
21:6536–6544
Belcourt LA, Labuza TP (2007) Effect of raffinose on sucrose recrystallization and textural
changes in soft cookies. J Food Sci 72:C065–C071
Brush RA, Griffith M, Mlynarz A (1994) Characterization and quantification of intrinsic ice
nucleators in winter rye (Secale cereale) leaves. Plant Physiol 104:725–735
Buera P, Schebor C, Elizalde B (2005) Carbohydrate crystallization phenomena in dehydrated
food and ingredient formulations: involved factors, consequences and prevention. J Food Eng
67:157–165
Buitink J, Leprince O (2004) Glass formation in plant anhydrobiotes: survival in the dry state.
Cryobiology 48:215–228
Carpenter JF, Crowe JH, Arakawa T (1986) Comparison of solute-induced protein stabilization in
aqueous solution and in the frozen and dried states. J Dairy Sci 73:3627–3636
Cerrutti P, Segovia M, Galvagno M, Schebor C, Buera MP (2000) Commercial yeast stability
during dehydration and thermal treatment as affected by intracellular content and external
presence of trehalose. Appl Microbiol Biotechnol 54:575–580
Chapsky L, Rubinsky B (1997) Kinetics of antifreeze protein-induced ice growth inhibition. FEBS
Lett 412:241–244
Clark MS, Thorne MAS, Purać J, Grubor-Lajšić G, Kube M, Reinhardt R, Worland MR (2007)
Surviving extreme polar winters by desiccation: clues from Arctic springtail (Onychiurus
arcticus). BMC Genomics 8:475–467
Clegg JS (2001) Cryptobiosis: a peculiar state of biological organization. Comp Biochem Physiol
128B:613–624
Costanzo JP, Litzgus JD, Iverson JB, Lee RE (1998) Soil hydric characteristics and environmental
ice nuclei influence supercooling capacity of hatchling painted turtles Chrysemys picta. J Exp
Biol 201:3105–3112
Costanzo JP, Lee RE Jr, Ultsch GRJ (2008) Physiological ecology of overwintering in hatchling
turtles. J Exp Zool A Ecol Genet Physiol 309:297–379
Crowe JH, Carpenter JF, Crowe LM (1998) The Role of vitrification in anhidrobiosis. Annu Rev
Physiol 60:73–103
Davies PL, Hew CL (1990) Biochemistry of fish antifreeze proteins. FASEB 4:2460–2468
Davies PL, Baardsnes J, Kuiper MJ, Walker VK (2002) Structure and function of antifreeze
proteins. Phil Trans R Soc B 357:927–935
Desjardins M, Le François NR, Fletcher GL, Blier PU (2007) High antifreeze protein levels in
wolffish (Anarhichas lupus) make them an ideal candidate for culture in cold, potentially ice
laden waters. Aquaculture 272:667–674
Du N, Liu XY (2008) Enhanced antifreeze effect of antifreeze protein on ice nucleation by
electrolyte. Cryst Growth Des 8:3290–3294
Du N, Liu XY, Hew CL (2003) Ice nucleation inhibition: mechanism of antifreeze by antifreeze
protein. J Biol Chem 278:36000–36004
Espinosa L, Schebor C, Buera MP, Moreno S, Chirife J (2006) Inhibition of trehalose crystalliza-
tion by cytoplasmic yeast components. Cryobiology 52:157–160
Evans RP, Hobbs RS, Goddard SV, Fletcher GL (2007) The importance of dissolved salts to the
in vivo efficacy of antifreeze proteins. Comp Biochem Physiol A Mol Integr Physiol 148:556–561
Feeney RE, Yeh Y (1998) Antifreeze proteins: current status and possible food uses. Trends Food
Sci Technol 9:102–106
Franks F (1982) The properties of aqueous solutions at subzero temperatures. In: Franks F (ed)
Water, a comprehensive treatise, vol 7. Plenum, London
Franks F (1995) Biophysics and biochemistry at low temperatures. Cambridge University Press,
Cambridge
Gabellieri E, Strambini GB (2006) ANS fluorescence detects widespread perturbations of protein
tertiary structure in ice. Biophys J 90:3239–3245
Gilbert JA, Hill PJ, Dodd CER, Laybourn-Parry J (2004) Demonstration of antifreeze protein
activity in Antarctic lake bacteria. Microbiology 150:171–180
Graether SP, Kuiper MJ, Gagne SM, Walker VK, Jia Z, Sykes BD, Davies PL (2000) Beta-helix
structure and ice-binding properties of a hyperactive antifreeze protein from an insect. Nature
406:325–328
Griffith M, Ewart KV (1995) Antifreeze proteins and their potential use in frozen foods. Biotech-
nol Adv 13:375–402
Harding MM, Anderberg PI, Haymet AD (2003) “Antifreeze” glycoproteins from polar fish. Eur
J Biochem 270:1381–1392
Heneghan AF, Wilson PW, Haymet ADJ (2002) Heterogeneous nucleation of supercooled water,
and the effect of an added catalyst. Proc Natl Acad Sci USA 99:9631–9634
Hew CL, Yang DS (1992) Protein interaction with ice. Eur J Biochem 203:33–42
Hinton HE (1960) A fly larva that tolerates dehydration and temperatures of 270 C to +102 C.
Nature 188:336–337
Holovati JL, Acker JP (2007) Spectrophotometric measurement of intraliposomal trehalose.
Kajava AV, Lindow SE (1993) A model for the three dimensional structure of ice nucleating
proteins. J Mol Biol 232:709–717
Kattera S, Chen C (2006) Cryopreservation of embryos by vitrification: current development. Int
Surg 91:S55–S62
Kawahara H (2002) The structures and functions of ice crystal-controlling proteins from bacteria.
J Biosci Bioeng 94:492–496
Kawahara H, Kawakami H, Obata H (2006) Purification and characterization of antifreeze protein
from a mushroom Flammulina velutipes. Cryobiology 53:428
Kristiansen E, Zachariassen KE (2005) The mechanism by which fish antifreeze proteins cause
thermal hysteresis. Cryobiol 51:262–280
Leinen KM, Labuza TP (2006) Crystallization inhibition of an amorphous sucrose system using
raffinose. J Zhejiang Univ Sci 7(2):85–89
Levine H, Slade L (1986) A polymer physico-chemical approach to the study of commercial starch
hydrolysis products (SHPs). Carbohydr Polym 6:213–244
Levine H, Slade L (1992) Glass transitions in foods. In: Schwartzberg H, Hartel R (eds) Physical
chemistry of foods. Marcel Dekker, New York, pp 83–221
Li B, Sun D-W (2002) Novel methods for rapid freezing and thawing of foods–a review. J Food
Eng 54:175–182
Li J, Lee T-C (1995) Bacterial ice nucleation and its potential application in the food industry.
Trends Food Sci Technol 6:259–265
Lillford PJ, Holt CB (2002) In vitro uses of biological cryoprotectants. Philos Trans R Soc Lond B
Biol Sci 357:945–951
Lindow SE (1983) The role of bacterial ice nucleation in frost injury to plants. Annu Rev
Phytopathol 68:523–527
MacKenzie AP (1977) Non-equilibrium freezing behaviour of aqueous systems. Philos Trans R
Soc Lond B Biol Sci 278:167–189
Matiacevich SB, Castellión ML, Maldonado SB, Buera MP (2006) Water‐dependent thermal
transitions in quinoa embryos. Thermochimica Acta 448:117–122
Matsukura K, Tsumuki H, Izumi Y, Wada T (2008) Changes in chemical components in the
freshwater apple snail, Pomacea canaliculata (Gastropoda: Ampullariidae), in relation to
development of its cold hardiness. Cryobiol 56:131–137
Matsumoto S, Matsusita M, Morita T, Kamachi H, Tsukiyama S, Furukawa Y, Koshida S,
Tachibana Y, Nishimura S, Todo S (2006) Effects of synthetic antifreeze glycoprotein
analogue on islet cell survival and function during cryopreservation. Cryobiol 52:90–98
Mazzobre MF, Soto G, Aguilera JM, Buera MP (2001) Crystallization kinetics of lactose in
systems co-lyofilized with trehalose. Analysis by differential scanning calorimetry. Food Res
Int 34:903–911
Mietchen D, Manz D, Volke F, Storey K (2008) Vivo assessment of cold adaptation in insect
larvae by magnetic resonance imaging and magnetic resonance spectroscopy. PLoS ONE
3(12):e3826, 1–10
Mignon J, Haubruge E, Gaspar Ch (1998) Effect of ice-nucleating bacteria (Pseudomonas
syringae Van Hall) on insect susceptibility to sub-zero temperatures. J Stor Prod Res
34:81–86
Murthy UMN, Kumar PP, Sun WQ (2003) Mechanisms of seed aging under different storage
conditions for Vigna radiata (L.) Wilczek: lipid peroxidation, sugar hydrolysis, maillard
reactions and their relationship to glass state transition. J Exp Bot 54:1057–1067
Obendorf RL (1997) Oligosaccharides and galactosyl cyclitols in seed desiccation tolerance. Seed
Sci Res 7:63–74
Packard GC, Packard MJ (2004) To freeze or not to freeze: adaptations for overwintering by
hatchlings of the North American painted turtle. J Exp Biol 207:2897–2906
Panza V, Láinez V, Maldonado SB, Maroder H, Buera P (2006) Storage changes and subcellular
freezing injuries in recalcitrant Araucaria angustifolia embryos. In: Water properties of food,
pharmaceutical and biological materials, eds. M.P. Buera, J. Welti‐Chanes, P. Lillford and
H.R. Corti. Boca Raton, FL: CRC Press/Taylor & Francis
Payne SR, Stanford D, Harris A, Young OA (1994) The effects of antifreeze proteins on chilled
and frozen meat. Meat Sci 37:429–438
Pertaya N, Marshall CB, DiPrinzio CL, Wilen L, Thomson ES, Wettlaufer JS, Davies PL,
Braslavsky I (2007a) Fluorescence microscopy evidence for quasi-permanent attachment of
antifreeze proteins to ice surfaces. Biophys J 92:3663–3673
Pertaya N, Celik Y, DiPrinzio CL, Wettlaufer JS, Davies PL, Braslavsky I (2007b) Growth–melt
asymmetry in ice crystals under the influence of spruce budworm antifreeze protein. J Phys
Condens Matter 19:412101, 12 pp
Pertaya N, Marshall CB, Celik Y, Davies PL, Braslavsky I (2008) Direct visualization of spruce
budworm antifreeze protein interacting with ice crystals: basal plane affinity confers hyperac-
tivity. Biophys J 95:333–341
Roos YH (1995) Water and phase transitions. In: Steve Taylor (ed) Phase transitions in foods.
Academic, London, pp 73–107, chapter 4
Roos Y, Karel M (1991a) Applying state diagrams to food processing and development. Food
Technol 45:66–71
Roos Y, Karel M (1991b) Amorphous state and delays ice formation in sucrose solutions. Int
J Food Sci Technol 26:553–566
Roos Y, Karel M (1991c) Nonequilibrium ice formation in carbohydrate solutions. Cryo Lett
12:367–376
Santagapita PR, Gómez Brizuela L, Mazzobre MF, Villalonga Santana R, Corti HR, Buera MP
(2008) Structure/function relationships of several biopolymers as related to b-fructofuranosi-
dase stability in dehydrated systems. Biomacromol 9:741–747
Slade L, Levine H (1991) Beyond water activity: recent advances based on an alternative approach
to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–360
Smallwood M, Worrall D, Byass L, Elias L, Ashford D, Doucet CJ, Holt C, Telford J, Lillford P,
Bowles DJ (1999) Isolation and characterization of a novel antifreeze protein from carrot
(Daucus carota). Biochem J 340:385–391
Tomczak MM, Hincha DK, Estrada SD, Feeney RE, Crowe JH (2001) Antifreeze proteins
differentially affect model membranes during freezing. Biochem Biophys Acta (BBA) –
Biomembranes 1511:255–263
Tunnacliffe A, Lapinski J (2003) Resurrecting Van Leeuwenhoek’s rotifers: a reappraisal of the
role of disaccharides in anhydrobiosis. Phil Trans R Soc Lond B 358:1755–1771
Venketesh S, Dayananda C (2008) Properties, potentials, and prospects of antifreeze proteins.
Crit Rev Biotechnol 28:57–82
Walters C (2004) Temperature dependency of molecular mobility in preserved seeds. Biophys J
86:1253–1258
Wang J-H (2000) A comprehensive evaluation of the effects and mechanisms of antifreeze
proteins during low-temperature preservation. Cryobiology 41:1–9
Watanabe M, Arai S (1994) Bacterial ice-nucleation activity and its application to freeze concen-
tration of fresh foods for modification of their properties. J Food Eng 22:453–473
Watanabe M, Kikawada T, Minagawa N, Yukuhiro F, Okuda T (2002) Mechanism allowing an
insect to survive complete dehydration and extreme temperatures. J Exp Biol 205:2799–2802
Wisniewski M, Fuller M, Glenn DM, Palta J, Carter J, Gusta L, Griffith M, Duman J (2001)
Factors involved in ice nucleation and propagation in plants: an overview based on new
insights gained from the use of infrared thermography. J Agr Sci 14:41–47
Wowk B, Fahy GM (2002) Inhibition of bacterial ice nucleation by polyglycerol polymers.
Wowk B, Fahy GM (2007) Ice nucleation and growth in concentrated vitrification solutions.
Cryobiology 55:330
Yeh C-M, Huang X-H, Sue C-W (2008) Functional secretion of a type 1 antifreeze protein
analogue by optimization of promoter, signal peptide, prosequence, and terminator in
Lactococcus lactis. J Agric Food Chem 56:8442–8450
Yin LJK, Chen ML, Tzeng SS, Chiou T-K, Jiang ST (2005) Properties of extracellular
ice-nucleating substances from Pseudomonas fluorescens MACK-4 and its effect on the
freezing of some food materials. Fisheries Sci 71:941–947
Zachariassen KE, Kristiansen E, Pedersen SA, Hammel HT (2004) Ice nucleation in solutions and
freeze-avoiding insects – homogeneous or heterogeneous? Cryobiology 48:309–321
Zachariassen KE, Kristiansen E (2000) Ice nucleation and antinucleation in nature. Cryobiology
41:257–279
Zhang C, Zhang H, Wang L, Gao H, Guo XN, Yao HY (2007a) Improvement of texture properties
and flavor of frozen dough by carrot (Daucus carota) antifreeze protein supplementation.
Zhang C, Zhang H, Wang L, Zhang J, Yao H (2007b) Purification of antifreeze protein from wheat
bran (Triticum aestivum L.) based on its hydrophilicity and ice-binding capacity. J Agric Food
Chem 55:7654–7658
.
Part IV
Food Microstructure
.
Chapter 23
Food Microstructures for Health, Well-being,
and Pleasure
José Miguel Aguilera
23.1 Introduction
For centuries, mankind has converted natural materials into an edible state, since
apart from some fruits, vegetables, and nuts, most plant and animal tissues are not
easily eaten. As large numbers of people flocked into cities, massive food produc-
tion became necessary with the concomitant problems of transporting, storing,
processing, and distributing millions of unitized and safe products. Presently, the
packaged food industry is the world’s largest business, with an annual turnover of
almost $1.6 trillion, and approximately 70% of the food consumed in homes in
large urban centers is purchased in supermarkets. Moreover, consumer food pro-
ducts are now cheaper and more abundant than ever before in the past. In this sense,
food manufacturing is the most prominent application of bioprocess engineering.
Yet, food is not distributed equally and it is estimated that 17% of the total
population in the world does not have enough food (FAO 2005). For those who
barely surpass the survival level, cereals, legumes, and tubers provide well over
50% of all calories. Rice alone, which undergoes only basic processing, provides
20% of the world’s food calories. As a result of intensive processing, the developed
world spends more than double the kcal per nutritional kcal compared to the less-
developed countries (LDCs), most of which comes from nonrenewable energy
sources (as opposed to burned biomass) (Giampietro et al. 1994).
Foods are processed to make them more palatable and safe, to improve their
shelf-life and convenience, and to expand the variety of the culinary supply. As is
the case of many materials typical in our daily life, some of the most desirable
properties of foods are related to events that occur at length scales below 100 mm.
Thus, food processing has been defined as a controlled effort to preserve, destroy,
and transform the microstructure of foods (Aguilera and Stanley 1999). For
J.M. Aguilera
Department of Chemical and Bioprocess Engineering, Pontificia Universidad Católica de Chile,
Santiago, Chile
e-mail: jmaguile@ing.puc.cl
578 J.M. Aguilera
example, the structure of fruits and vegetables must be preserved after harvest so
that we can enjoy their natural flavors and textures, as is the case with muscle tissue
after slaughter. Wheat has to be ground into fine particles, a prerequisite to mixing
the flour with other ingredients. The structure of oilseeds and sugar beets needs to
be partly obliterated so that the extracted products, oil and sugar, respectively, are
easily recovered from plant cells with minimal downstream processing. But it is the
transformation and creation of safe, healthy, and palatable structures which is the
main thrust of the food processing industry of the future.
23.2 Foods Are Unique Materials
People depict the perception of foods in terms typical of the mechanics and flow of
materials (e.g., tough, soft or thick, sticky, thin, etc.) and scientists try to character-
ize them according to the principles of materials science (Aguilera and Lillford
2008). But when it comes to “design” foods, they exhibit their own uniqueness.
Foods are unique among materials in daily life in that they are ingested, and become
part of our body. (Note: foods are not the only products we put in our mouth and
swallow. Pharmaceutical tablets and capsules also find their way into our bodies;
however, they do not have to taste good or look good!). Foods are exceptional
among engineering materials in that they have to fail in the mouth and disintegrate
in the digestive tract. This immediately adds several extra dimensions to their
intrinsic properties as materials: foods have to be palatable, chewable, and degrad-
able. Imperfections produced during fabrication (e.g., gas bubbles, surface cracks,
etc.) that are intolerable in engineering materials are part of the intrinsic quality of
some foods. For example, Swiss cheese must possess well-developed round or
slightly oval-shaped “eyes” that are relatively uniform in size and distribution
(USDA 2001). Unlike the case of engineering materials, food properties have no
absolute value, but instead depend largely on several characteristics assigned by
consumers (i.e., age of product, physiological state, etc.), including the consumer’s
personal judgment, which is largely based on traditions and learning. Moreover,
some foods may elicit different physiological responses in different individuals, as
is the case of food allergies.
23.3 Food Structure Matters
In foods, structure matters because it is responsible for many of food’s desirable

properties, such as appearance, texture, flavor release, and nutrient bioavailability
(Aguilera 2005; Parada and Aguilera 2007). Processed foods (e.g., baked and
confectionery products, dried pasta, processed meats, etc.) are structured mixed
systems in which individual components (proteins, fats, carbohydrates, etc.) are
usually reassembled into colloidal, amorphous or crystalline states or separated into
23 Food Microstructures for Health, Well-being, and Pleasure 579
Food product physics

Colloidal science
Polymer science
“Nano” sciences Resolution
Detection of the eye
Chemistry
in the
Digesta mouth
Water Gluten Microbial
Cocoa
0.3 nm network
Powder
cells
FOOD PRODUCTS
particles particles
Fiber
Flavors
Lipid Crystals
micelles
CHO Ice crystals
Microbubbles in icecream Bubbles
polymers
Plant Micro
cell droplets Fat droplets
Proteins
Emulsifiers walls
Plant Grains
Network Particle cells
gels gels
Cooked
Casein Starch starch
Casein micelles
nanoparticles
1nm 10 nm 100 nm 1 μm 10 μm 100 μm 1 mm 10 mm
Fig. 23.1 The scale of some microstructural elements important in foods (approximate values)
concentrated phases after heating/cooling and the application of shear. By food

microstructure we understand the spatial arrangement of identifiable elements in a
food and their interactions (Aguilera and Stanley 1999). Typical microstructural
elements in foods are plant cell walls, starch granules, water and oil droplets, fibers,
fat and ice crystals, gas bubbles, etc. Most of these elements cannot be seen by the
naked eye but their abundance, size, and size distribution are easily recognized as
texture, which is perceived in the mouth (Fig. 23.1).
The design challenges to food fabricators have evolved and expanded through
time. Until the late 1980s, the main structural properties assessed by scientists were
those associated with the physical states of products as they moved from producers
to consumers along the supply chain. The emerging concerns of consumers for
health and well-being have brought about new structural properties targeted at the
brain-gut axis, which will increasingly be the discriminating attributes that deter-
mine success in the marketplace. For the first time in the history of food and
nutrition, we have the knowledge and tools to design food for health, well-being,
and pleasure.
23.4 Nature Is the Ultimate Provider of Food Structures
Palatable foods are easily recognized by the sensations experienced during masti-
cation. Foods produced by nature are generally organized hierarchically from
molecules into assemblies and organelles (e.g., protein bodies, fat deposits, cell
580 J.M. Aguilera
Myofibril Fiber Muscle Meat
Fig. 23.2 An example of hierarchical scales in muscle tissue
walls, etc.) that are later compartmentalized into cells (some in the form of fibers)
and tissues. “Natural” food structures may be classified into four broad categories:
(i) Fibrous structures assembled hierarchically from small molecules (glucose) or
macromolecules (proteins) into tissues for specific functionality (e.g., muscles) and
held together at different levels by specific interfacial interactions (Fig. 23.2); (ii)
fleshy materials from plants that are hierarchal composites of hydrated cells that
exhibit internal (turgor) pressure of about 1 MPa (10 atm) and are bonded together
at the cell walls (e.g., tubers, fruits, and vegetables); (iii) encapsulated embryos of
plants that contain a dispersion of starch, protein, and lipids assembled into discrete
packets (e.g., in grains and pulses); and (iv) a unique complex fluid called milk,
intended for nutrition of the young mammal containing several nutrients in a state
of dispersion.
Structures in nature have been regarded as the inspiration for the design of
better-fabricated foods. There is ample awareness among cooks and food technol-
ogists that destroying the structure of natural foods results in rapid and uncontrolled
deterioration. Nature stabilizes structures against unwanted reactions mainly by
three means:
– At the molecular level by complexing reactants into passive forms.
– By compartmentalizing and segregating substrates and catalysts. Take garlic,
for example. When the tissue is disrupted the odorless precursor is placed in
contact with an enzyme that triggers a series of reactions leading to diallyl
disulfide, the powerful component of garlic odor.
– In seeds, nature “freezes” biological activity at ambient temperature by immo-
bilization of the whole system into an amorphous matrix.
Copying nature, however, has not always resulted in good food analogs. Marga-
rine and other spreadable fats made from vegetable oils have a microstructure
radically different from that of the “natural product” butter. In margarine, crystal-
lized fat forms a fine network of interconnected platelets composed of single
crystals and “sheet-like” crystal aggregates that occlude an emulsion of water and
liquid oil (Heertje 1993); in butter, fat remains largely as fat globules.
23.5 Atomic Doping by Nature and the Color of Some Foods
The presence or absence of specific atoms does make a difference in how we enjoy
our foods, and nature does a fabulous job in doping complex molecules with
selected atoms for this purpose. Chlorophyll, the green pigment that we appreciate
in all green vegetables and fruits, is similar in structure to hemoglobin, which
provides the red coloring in meat (Fig. 23.3). The notable difference between these
two substances is that in the center of the chlorophyll molecule is a magnesium
atom, while in the center of hemoglobin is iron. During thermal processing the Mg
atom is displaced by two hydrogen atoms, resulting in the olive-brown pigment
pheophytin.
23.6 The Cow’s Udder: A Fantastic Microfluidic Device
It is remarkable that the wide assortment of dairy products is derived from struc-
tures formed by three types of building blocks: casein micelles (CM), globular
whey proteins, and fat globules (FG). Emulsions (cream), foams (whipped cream),
gels (yoghurt), a plastic solid (butter), and cheeses of many textures are the result of
interactions between these two types of elements.
CM and FG are assembled inside the mammary cells of the cow’s udder. Casein,
b-lactoglobulin, and a-lactalbumin are synthesized from amino acids at the poly-
somes (poly-ribosomes) on the rough endoplasmic reticulum (RER). Synthesized
proteins are transferred from the RER to the Golgi apparatus where they are
processed for transport out of the cell. Casein is secreted from the Golgi apparatus
H2C CH R
Me Et
N N
Mg
N N
Me
Me
H
H2C H
H O
CH2 CO2 Me
CO2
Me H Me H
a, R=CH3
b, R=CHO
Fig. 23.3 Structure of chlorophyll molecule that gives green color to vegetables. Note the Mg
atom at the center
582 J.M. Aguilera
Fig. 23.4 Scheme of a

mammary cell in cow’s
udder, indicating location of
synthesis, assembly, and
release of casein micelles and
fat globules into the lumen
as a micelle assembled from the subcasein molecules, calcium, and phosphorous

(Fig. 23.4). Milk fat triglycerides are synthesized on the smooth endoplasmic
reticulum. Numerous small lipid droplets fuse together as they move toward the
apical membrane, and a large lipid droplet forces out the membrane of the cell,
becomes surrounded by it and enters the lumen (Heid and Keenan 2005). Conse-
quently, each fat globule in milk is covered by a biological membrane, the fat
globule membrane (Fig. 23.4). Thus, the mammary cell is an amazing microdevice
for the assemblage and release of individual particles of colloidal size.
23.7 The Kinetics of Structure Formation: The Case

of Whipped Cream
In contrast to natural materials, the morphological features of structures in fabri-

cated foods are, to a certain extent, within our control. Origin of the many structures
of foods, even those made from a single raw material (e.g., cow’s milk), lies in
the ingredients mix, the interaction between structural elements, and the fact
that thermodynamic equilibrium is practically never achieved during processing.
Usually, metastable structures are attained because they are favored kinetically, i.e.,
the approach to equilibrium is slow. Physical and chemical changes occur simulta-
neously with the formation of the matrix. At a certain point during the development
of a particular structure a process of shape stabilization sets in, usually by vitrifica-
tion, high viscosity effects, partial crystallization, phase separation, and/or forma-
tion of a network.
Fat Native
globule 4 -10 nm
Fat globule
membrance Homo-
genized
Air
Casein Gel
network
Casein
micelle
Whipped
3.6 nm cream
Bubble
Globular Aggregate interface
proteins
Native Denatured
β lactoglobulin globular protein
nm μm mm
Fig. 23.5 The main building blocks for structure formation in milk are fat globules, casein, and
globular whey proteins (only b-lactoglobulin is shown). Structure formation leading to a stable
whipped cream proceeds to the right (larger microstructures). At the center is shown the associa-
tion of destabilized casein micelles into a gel network, the basis of cheese-making
Whipped cream is a foam, stabilized by particles located at the air/liquid interface

(Fig. 23.5). Bubble stabilization is mostly achieved by partly-solid fat globules (ca.
1–4 mm diameter, see below) rather than by surface active agents (as in beaten egg
whites). This is why cream that has less than 30% fat will not whip. Temperature
strongly influences the physical state (i.e., ratio of solid fat to liquid fat) and
consequently the properties of fat globules. To stabilize the foam, fat crystals must
collect at the bubble interface to impede coalescence. Foam formation and stabili-
zation can basically be explained as a multistage process, with some significant
differences between nonhomogenized (pasteurized cream) and ultra high tempera-
ture (UHT, held at 140 C for 4 s) homogenized creams (Kulozik 2008). These
differences originate from the size and state of fat globules in each particular system
prior to whipping (Fig. 23.5). In nonhomogenized cream, fat globules are sur-
rounded by the native fat-globule membrane, while in homogenized UHT cream
they are covered by a modified membrane of proteins (casein and globular proteins).
Also, in the former case fat globules are larger than in the latter (~4 vs ~1 mm).
23.8 Hierarchical Arrangements in Fats
Molecular transformations and structure-building properties of proteins and poly-

saccharides are more or less well understood. Denaturation, aggregation, and
gelation of properties of several proteins are documented, as well as the
584 J.M. Aguilera
Crystallite Lamellae TAGs
Cluster Floc Fat crystal network
Fig. 23.6 Structural hierarchy of fat crystal networks (from top right, counter clockwise)
(Adapted from Narine and Marangoni 1999)
gelatinization and gel formation of starches and other hydrocolloids. Hierarchical

structures also develop spontaneously in the case of fats that influence the func-
tional properties of products (spreadability, mouth-feel, etc.). Fats and oils com-
prise primarily triacylglycerides (TAGs). As the temperature of a molten fat (oil) is
reduced below melting temperature it becomes supersaturated, with regards to the
higher-melting fraction of TAGs, and crystallization starts to take place. It is typical
of fats to exhibit polymorphism, that is, they can crystallize into different forms.
Moreover, hierarchical structures are assembled that span from thin crystalline
lamellas to a network of crystal flocs (Fig. 23.6). This is important because the
rheological properties of products containing fats depend on the microstructures
and bonding mechanisms holding the units together at different levels (Narine and
Marangoni 1999).
Due to their amphiphilic nature, monoglycerides and phospholipids form various
self-assembly structures (association colloids) when dissolved in water, for exam-
ple, lamellar, micellar, cubic or hexagonal mesophases in the range of 10–100 nm.
In molecular self-assembly, the final structure is “encoded” in the properties of the
molecules rather than obtained through chemical reactions. Liquid crystalline
phases arranged as a multilamellar a-coagel can incorporate large amounts of
water between ordered bilayers, and have been used in the design of low-calorie
fat spreads (Heertje et al. 1998).
23.9 Microstructures for Health
A next generation of processed foods will be designed for the brain-body axis. As
such, they will be enjoyable to eat, will be nutritionally efficient, and will make us
feel better. For example, many of these foods will be structured in such a way as to
control the release of flavors or bioactive molecules, the rate of stomach emptying,
etc. (Norton et al. 2007). As we understand more about how food components
interact with the human genome and metabolism, new perspectives for the preven-
tion of nutrition-related diseases will become available (German et al. 2004). One
approach is personalized nutrition or matching metabolic needs with specific foods
and diets. The challenge will be to make foods that are appealing to the health-
conscious consumer, have the right proportion of beneficial ingredients, and are
pleasant to be eaten.
In the last 10 years, evidence has shown that food structure plays a key role in
digestion and absorption and that these processes are not as simple and direct as
previously thought. People have become more concerned about the fraction of specific
nutrients reaching systemic circulation (bioavailability) than with the total amount
present in our food before ingestion. Before absorption into the gut a nutrient must
become bioaccessible, meaning that it has to be at the molecular level, in a colloidal
state, or in a micellar state, in the case of hydrophobic compounds (Duchateau and
Klaffke 2008).
The structural organization of the food matrix as well as its nature have been
demonstrated to affect the slow degradation of starch in pasta (Fardet et al. 1998)
and the possible influence on the postprandial glycemic response (Ricciardi et al.
2003). Resistant or slowly digestible starches, as well as control of the degree of
starch gelatinization, may be used to modulate the glucose concentration in the
blood (Parada and Aguilera 2009). The rate and extent of lipid digestion depends on
the breakdown of the matrix in which the lipid phase is embedded. Matrices
occluding lipids may be designed by gelation, extrusion, or drying of biopolymers
and sugars (McClements et al. 2008). It has been suggested that the controlled
physiological response of lipid emulsions may be affected by proper design of the
adsorbed layer on lipid droplets and their size (Singh et al. 2009). The role of food
structure in the availability of several nutrients (carotenoids, xanthophylls, etc.) has
been recently reviewed by Parada and Aguilera (2007). As we increase our under-
standing of how food structures are formed during processing and cooking, and
broken down during digestion, the ability to design healthy products will be
enhanced (Leser et al. 2003).
To control the microstructure of processed foods, traditional unit operations
will have to be scaled down to the micron level and below. Current technologies
used by food processors are largely based on breaking down structure into smaller
elements by the application of mechanical energy to the bulk phase (e.g., grind-
ing, homogenization, spray drying). The ratio between the structure-forming
device (e.g., gap in the homogenizing valve) and the microstructural elements
formed (liquid droplets) may be of several orders of magnitude. In the future,
structure operations will be more like the cow’s udder where assembly and
release of microstructures will occur at the individual element’s level. One inte-
resting application is the use of membranes and microfluidic devices in the forma-
tion of multiple emulsions and foams (Skurtys and Aguilera 2008; Muschiolik
2007).
586 J.M. Aguilera
23.10 Microstructures for Pleasure
With the advent of gastronomy in the nineteenth century came the first restaurant.
Presently, annual sales in the restaurant business around the world are between $1
and 1.5 trillion dollars. At restaurants one can experience the delights of almost any
traditional and ethnic cuisine and the creativity of talented chefs for a price. This is
in opposition to the concept of cheap and standardized “fast food” and can only be
explained by the desires of people to experience the full pleasure of their senses.
Some modern chefs have broken with old culinary routines, designing dishes
based on sophisticated techniques and guided by scientific principles explored in
their own laboratories. Taking science into the chef’s domain has been called
molecular gastronomy (Vega and Ubbink 2008; This 2006). For example, food
science and gastronomy have merged in the kitchen in the preparation of artificial
caviar, cryogenic desserts, innovative gels, and “airs” or foams (3-D sauces),
among other creations. These structural creations are the envy of food technologists
who have had experience with these novelties but have not succeeded in placing
them on supermarket shelves due to the high costs involved and lack of demand.
Nonetheless, top chefs nowadays have gained the trust of people who have been
seduced by their creations, and they are probably the most innovative professionals
in the food industry. Playing gastronomic engineering in the chef’s kitchen does not
require sizing the market or scaling up processes, since the minimal portions
prepared in the kitchen are almost under laboratory conditions (Aguilera 2009).
Therefore, chefs and food scientists/engineers alike are called to unite and expand
the offer of good food and the appreciation for taste and quality.
23.11 Conclusions
Most of the desirable food properties depend on the way a food is structured, either
by nature or processing. Throughout time, millions of “popular scientists” have
experimented with foods in the kitchen, so it is not surprising that many creative
food structures have originated either by trial-and-error (or just error!) or by
serendipity. It is only in the last few decades that we have gained a basic under-
standing of how food microstructures are formed from the molecular level to that of
products. This has initiated the era of food materials science and the bottom-up
design of food products. New technologies, most notably the use of membranes and
microdevices, promise to bring the scale of fabrication closer to that of key
microstructural elements. On the demand side, increasing evidence linking diets
and some nontransmissible diseases (cardiovascular diseases and some cancers)
will open new opportunities to tailor products that may help prevent or ameliorate
the effects of these diseases. Novel ingredients and nutraceuticals that are used
in the food health and well-being market need to be structured in attractive formats.
In the coming years, structuring foods for the brain (pleasure), the gut (gut health),
and the cells (health and vitality) will become increasingly important as new
knowledge emerges from studies of neurophysiology, consumer preferences, nutri-
tion and digestion, nanoscience, and nutrigenomics as well as other –omics.
Acknowledgments Research has been partly funded by the Fondecyt project 1095199 and the
Marcel Loncin Award of IFT.
References
Aguilera JM (2005) Why food microstructure? J Food Eng 67:3–11

Aguilera JM (2009) Gastronomic engineering. In: Moskowitz H, Saguy IS, Straus T (eds) An
Integrated approach to new food product development. Taylor & Francis, Boca Raton, FL,
pp 317–328
Aguilera JM, Lillford PJ (2008) Food materials science: principles and practice. Springer,
New York
Aguilera JM, Stanley DW (1999) Microstructural principles of food processing and engineering,
2nd edn. Aspen, Gaitherburg, MD
Duchateau G, Klaffke W (2008) Product composition, structure and bioavailability. Food Biophys
3:207–212
FAO (2005) The state of food insecurity in the world 2005. Food and Agriculture Organization,
Rome, Italy
Fardet A, Hoebler C, Baldwin PM, Bouchet B, Gallant DJ, Barry J-L (1998) Involvement of the
protein network in the in vitro degradation of starch from spaghetti and lasagna: a microscopic
and enzymic study. J Cereal Sci 27:133–145
German JB, Yeretzian C, Watzke HJ (2004) Personalizing foods for health and preference. Food
Technol 58(12):26–31
Giampietro M, Bukkens SGF, Pimentel D (1994) Models of energy analysis to assess the
performance of food systems. Agric Syst 45:19–41
Heertje I (1993) Structure and function of food products. Food Struct 12:343–364
Heertje I, Roijers EC, Hendrickx HACM (1998) Liquid crystalline phases in the structuring of
food products. Lebems-Wiss u- Technol 31:387–396
Heid HW, Keenan TW (2005) Intracellular origin and secretion of milk fat globules. Eur J Cell
Biol 84:245–258
Kulozik U (2008) Structured dairy products: Processing and matrix design. In: Aguilera JM,
Lillford PJ (eds) Food materials science. Springer, New York, pp 439–473
Leser ME, Michel M, Watzke H (2003) Food goes nano–New horizons for food structure research.
In: Dickinson E, van Vliet T (eds) Food colloids: biopolymers and materials. Royal Society of
Chemistry, Oxford, pp 3–13
McClements DJ, Decker EA, Park Y, Weiss J (2008) Designing food structure to control stability,
digestion, release and absorption of lipophilic food components. Food Biophys 3:219–228
Muschiolik G (2007) Multiple emulsions for food use. Curr Opin Colloid Interf Sci 12:213–220
Narine SS, Marangoni AG (1999) Relating structure of fat crystal networks to mechanical proper-
ties: a review. Food Res Intern 12:227–248
Norton I, Moore S, Fryer P (2007) Understanding food structuring and breakdown: engineering
approaches to obesity. Obes Rev 8(suppl1):83–88
Parada J, Aguilera JM (2007) Food microstructure affects the bioavailability of several nutrients.
J Food Sci 72:R21–R32
Parada J, Aguilera JM (2009) In vitro digestibility and glycemic response of potato starch is related
to granule size and degree of gelatinization. J Food Sci 74:E34–E38
588 J.M. Aguilera
Riccardi G, Clemente G, Giacco R (2003) Glycemic index of local foods and diets: the Mediterra-
nean experience. Nutr Rev 61:56–60
Singh H, Ye A, Horne D (2009) Structuring food emulsions in the gastrointestinal tract to modify
lipid digestion. Progr Lipid Res 48:92–100
Skurtys O, Aguilera JM (2008) Applications of microfluidic devices in food engineering. Food
Biophys 3:1–15
This H (2006) Molecular gastronomy: exploring the science of flavor. Columbia University Press,
New York
Vega C, Ubbink J (2008) Molecular gastronomy: a food fad or science supporting innovative
cuisine? Trends Food Sci Technol 19:372–382
USDA (2001) United States standards for grades of Swiss cheese, emmentaler cheese. }58.2574
Specifications for US grades. The United States Department of Agriculture, Washington, DC
Chapter 24
Fruit Microstructure Evaluation Using
Synchrotron X-Ray Computed Tomography
Pieter Verboven, Quang Tri Ho, Els Herremans, Hibru Kelemu Mebatsion,
Bart Nicolaı̈, Greet Kerckhofs, Martine Wevers, and Peter Cloetens
24.1 Fruit Quality and Microstructure
Fruits and vegetables cover 16.6% of the EU production of agricultural food.1 Fruit
is after milk (but leading over cereals) the most important food item consumed in
the EU (over 400 g/capita/day). Yearly, the production amounts to 38.3 million tons
of fruit and 66 million tons of vegetables. Fresh apples and pears are products of
many countries in the EU, with main production in France, Spain, Italy, Germany,
Poland, Austria, Belgium, and The Netherlands. Exports (3€ billion in 2005) are
growing for both fruits and vegetables, but imports are more than four times higher
and are competing strongly in the home market.
Fruit represents an important category of biological food products with high
water content and cellular microstructure. Fruit’s visual, textural, and nutritional
qualities have direct economic impact for fresh consumption as well as processing.
A fruit consists of different tissue types such as the epidermis (with cuticle), cortex
parenchyma tissue, core tissue, and vascular tissue, each with a different micro-
structural composition. From macro- to nano-scale, the physical properties of fruit
are affected by type of tissue, geometric properties of cells, presence of an adhesive
middle lamella between individual cells, cellular water potential, mechanical prop-
erties of cell wall, presence of intercellular spaces, and subcellular features, such as
plasma-membrane, plasmodesmata, and aquaporins (Fig. 24.1). Degradation of
1
http://ec.europa.eu/agriculture/capreform/fruitveg/presentations/fresh_en.pdf
P. Verboven (*), Q.T. Ho,, E. Herremans,, H.K. Mebatsion, and B. Nicolaı̈,
Division BIOSYST-MeBioS, K.U. Leuven, W. de Croylaan 42, P.O. Box 2428, BE-3001 Leuven,
Belgium
e-mail: Pieter.verboven@biw.kuleuven.be
G. Kerckhofs and M. Wevers,
Research group of Materials Performance and Non-destructive Evaluation, Katholieke Universiteit
Leuven, Kasteelpark Arenberg 44, BE-3001 Leuven, Belgium
P. Cloetens
ID19 High-resolution Diffraction Topography Beamline, European Synchrotron Radiation Facility,
6 rue Jules Horowitz, BP220, 38043 Grenoble Cedex, France
590 P. Verboven et al.
Fig. 24.1 Some features of apple microstructure (left to right): cross-section of cell wall with
plasmodesmata; cross-section of cells with intercellular pores; and epidermis with wax layer
fruit’s microstructure after harvest quickly leads to interior quality defects. The
amount of loss due to internal disorders that develop during storage varies from
year to year, but for some cultivars this is always significant. Peak losses of fruit
have been recorded as high as 20–30% in some years and locations.
Several disorders may occur that reduce the commercial value of the fruit.
Pome fruit, for example, is often stored for up to 10 months at a low temperature
(typically around 0 C) in combination with reduced O2 and increased CO2 partial
pressure (so-called “Controlled Atmosphere (CA) storage”) to reduce their respi-
ration rate and, hence, extend their storage life. However, the optimal gas
composition is critical, as too low O2 partial pressure in combination with too
high CO2 partial pressure may lead to physiological disorders and off-flavors.
In fact, under suboptimal conditions some fruits may develop browning and core
breakdown. This storage disorder is characterized by the development of brown
internal tissue, which further develops into cavities so that the fruit can no longer
be commercialized. This disorder has been monitored nondestructively by means
of MRI and X-ray CT (Lammertyn et al. 2003). How it relates to microstructure
is unknown.
The incidence of senescent breakdown in apple with symptoms similar to
browning varies from year to year, apparently affected by preharvest growing
conditions. This disorder is related to the age of the fruit, and occurs more often
in large, over-mature apples. Late harvest, delay in cooling, and storage at tem-
peratures above those recommended favor the occurrence. Incidence of senescent
breakdown usually indicates that storage life has passed.
Water core is a disease that appears as hard glassy regions near the core in
apples. Apples that have been exposed to high temperatures and sunlight near
maturity are more susceptible to water core development. Water core does not
develop in storage and may even disappear when originally present in a mild form.
Bitter pit symptoms are mainly found in the cortex apple tissue near the skin and
first appear as soft, brown areas, which eventually become desiccated due to the
collapse of surrounding cells, forming a dry cavity or “pit.” The disorder is due to
preharvest factors including the fruit’s mineral status and the climate in the orchard.
The microstructure of fresh pome fruit (apple and pear) has recently been investi-
gated by means of light microscopy (Schotsmans et al. 2004), scanning electron
microscopy and confocal microscopy (Veraverbeke et al. 2001), and very recently,
by X-ray computed microtomography (Mendoza et al. 2007; Mebatsion et al. 2006;
Verboven et al. 2008).
24 Fruit Microstructure Evaluation Using Synchrotron X-Ray Computed Tomography 591
Internal disorders in fruit can cause extreme losses during the storage season.
Since fruits normally have to be cut to detect internal disorders, they are often only
observed upon quality inspection of the whole batch after shipping. This typically
leads to refusal and subsequent destruction of the whole batch, which may cause
large financial loss. By using new microstructure sensors and improved understand-
ing of the effect of microstructure on such disorders, their occurrence may be
detected earlier or even in advance, resulting in more high quality fruit.
24.2 X-Ray Computed Tomography
X-ray CT is a relatively new technique developed in the late 1970s that enables the
nondestructive visualization of the internal structure of objects. The first, mainly
medical CT scanners had a pixel resolution of about 1 mm. In the 1980s, after some
technological advances towards micro-focus X-ray sources and high-tech detection
systems, it was possible to develop a micro-CT (or mCT) system, which nowadays
has a pixel resolution of 1,000 times better than medical CT scanners. The tech-
nique of X-ray (micro)-CT is based on the interaction of X-rays with matter. When
X-rays pass through an object they are attenuated depending on the density and
atomic number of the object under investigation and the applied X-ray energy. By
using projection images obtained from different angles, a reconstruction can be
made of a virtual slice through the object. After different consecutive slices are
reconstructed, a 3-D visualization can be obtained. Next to mCT technology, new
in-situ stages (rotatable support platforms for samples) can be developed, which
increase the possibilities of the tomographic systems. An example of such in-situ
stages is the environmental stage (or cooling stage), which is an interesting tool in
the study of fresh food products.
As a noninvasive technique, mCT has been applied to the study of internal 3-D
structures for several food products, e.g., marshmallow, aerated chocolate, and
chocolate muffins (Lim and Barigou 2004). Kuroki et al. (2004) obtained 3-D
spatial information on gas-filled intercellular spaces in cucumber fruit. Babin
et al. (2005) studied the microstructure of cellular cereal products captured by
synchrotron radiation mCT. Leonard et al. (2008) used mCT on processed banana.
Synchrotron radiation mCT has been applied by K.U. Leuven (Katholieke Univer-
siteit Leuven) with success to more difficult products, such as apple and pear, with
resolutions below 1 mm (Mendoza et al. 2007; Verboven et al. 2008).
A new challenge in technology is the nano-CT system that has opened up a new
era in X-ray imaging with a spatial resolution in the range of hundreds of nano-
meters. Proceeding to submicron pixel sizes requires increased performance of the
X-ray source, rotation stage, and X-ray detector. The fact that the object can be
scanned under normal environmental conditions without any coating, vacuum
treatment, or other preparation techniques makes it an interesting tool as such, as
well as a reference source for interpreting microstructure measurements with other
methods.
24.3 Synchrotron X-Ray CT of Fruit Tissue
Visualization of microstructure by X-ray microtomography of plant tissues in their

natural state has been difficult because of the low contrast and limited resolution
(Westneat et al. 2003; Kuroki et al. 2004; Kim et al. 2006; Rau et al. 2006; Mendoza
et al. 2007). Important structural features such as small voids between cells,
vascular capillaries, or cell walls could therefore not be visualized, rendering
incorrect connectivity information (Fig. 24.2).
High resolution tomography using synchrotron radiation offers a means to
explore at submicrometer resolution 3-D fruit tissues with high water content in
their natural state. As explained by Salvo et al. (2003), synchrotron radiation X-ray
tomography has important advantages over the X-ray tube tomography of conven-
tional X-ray CT equipment. X-ray tube tomography produces a divergent beam;
therefore, the resolution is limited by the beam angle and required field of view. The
beam angle may also produce artifacts in the reconstructed images. The parallel
beam produced by synchrotron radiation, with good spatial coherence, makes a
quantitative reconstruction that is free of geometrical and beam hardening artifacts
possible. These conditions can only be achieved at large scale facilities with high
energy flux sources and long distances between the source and tomography set-up
(around 100 m). The facility at ESRF (Grenoble, France) was accessible to the
authors on the basis of a successful research proposal.
Phase contrast imaging with synchrotron X-rays has been developed for edge
enhancement on tomographs with low absorption mode contrast (Davis et al. 1995;
Cloetens et al. 1999). High resolution phase tomography of biological tissues at a
pixel size close to 1 mm has only recently been achieved (Plouraboue et al. 2004;
Thurner et al. 2005), but it required sample preparation to improve contrast in the
images. For in vivo observations, high resolution phase tomography has thus far
been applied to relatively dry or hard biological samples, such as plant seeds
(Stuppy et al. 2003; Cloetens et al. 2006), as well as wet soft samples more prone
to damage by X-rays (Cloetens et al. 2006; Verboven et al. 2008). Figure 24.2c
demonstrates the power of the technique for imaging cells, cell walls, and pores in
pear fruit tissues.
24.4 3-D Imaging of Fruit Microstructure
24.4.1 Fruit and Methods
Pears (cv. Conference) were harvested (September 13, 2006) at the experimental
station Fruitteeltcentrum (Rillaar, Belgium). Apples (cv. Jonagold) were picked
during the pre-climacteric stage (September 25, 2006) at another experimental station,
PCFruit (Velm, Belgium). Some fruits with “Early” and “Late” picking dates
were also harvested 14 days before and after the given dates, respectively. All fruits
Fig. 24.2 (a) Microfocus X-

ray CT of pear at 9.5 mm pixel
size, image size 256*256
pixels. (b) High resolution X-
ray CT of pear at 4.8 mm pixel
size, image size 256*256
pixels. (c) Synchrotron X-ray
CT (phase contrast) of pear at
0.7 mm pixel size, image size
256*256 pixels. Dark areas
indicate air-filled pores,
lighter colored zones are cells
were cooled and stored under controlled atmospheric conditions (2.5 kPa O2, 0.7 kPa
CO2 – 1 C for pear; 1 kPa O2, 2 kPa CO2 – 0.8 C for apple) up to time of experiments
(Nov. 3–6, 2006). Picking data and cooling procedures are optimal commercial
practices to preserve fruit quality during long-term storage. Cylindrical samples,
5 mm diameter and 1–2 cm length, were removed from the fruit tissues of apple
and pear using a cork bore in the radial direction on the equator of the fruit. The
samples were mounted in a polymethyl methacrylatetube and covered with polymer
foil to avoid dehydration. As stated previously, the experiments were conducted on
beamline ID19 at the ESRF, Grenoble, France i.e., a long (150 m) imaging beamline
where the spatial coherence of the beam is particularly large (transverse coherence
length in the order of 100 mm). The system provided a field of view of
1.43 1.43 mm2 and, at best, an image pixel size of 0.7 mm. 3-D stacks of
2,048 2,048 2,048 pixels were obtained. Volume renderings and quantitative
measurements on the sample were obtained by 3-D image segmentation and isosur-
face representations with Amira (Mercury Computer Systems, Chelmsford, MA), and
dedicated software written for virtual tissue generation (Mebatsion et al. 2009).
24.4.2 Microstructure of Apple and Pear Fruit
Figure 24.3 shows the microstructure of different tissues of Jonagold apple and
Conference pear fruit obtained by synchrotron radiation X-ray imaging. The dermal
tissue presents a dense assembly of cells with little or no voids. Both fruits are
aerated by voids in between the cells of the parenchyma that makes up the bulk of
the cortex tissue of the fruit. The vascular tissue in the mature fruit contains empty
xylem vessels surrounded by dense tissue without any air voids. The parenchyma in
the pear fruit consists of smaller cells and air voids compared to the apple.
The structure of the voids is significantly different in the two fruits. In apple, the
voids are the size of the cells, while in pear the voids are small channel-like
structures. The total fraction of voids is also significantly larger in the Jonagold
apple than in the Conference pear. Verboven et al. (2008) used the microstructure
characteristics presented in Fig. 24.3 to interpret the apparent gas exchange proper-
ties of the two fruits in relation to storability and internal disorders.
24.4.3 Multiscale Modeling
One purpose of microstructure data is in modeling. As there are no good methods at

present to measure in vivo internal concentrations of constituents in plant tissues, a
mathematical modeling approach provides an alternative to predicting these concen-
trations. Such a model could be used conveniently to perform in silico experiments
to evaluate, for example, the effect of changing storage or processing conditions on
the internal quality. The multiscale modeling paradigm (Ho et al. 2009) provides
Fig. 24.3 The 3-D

microstructure of apple cv.
Jonagold (left) versus pear cv.
Conference (right) obtained
from synchrotron radiation
X-ray tomography. Air-filled
pores are colored in blue,
cells in yellow, cell walls and
sclereids in brown. Details of
results are given in Verboven
et al. (2008)
a means to combining the relative simplicity of continuum-type models defined

at the macroscale level with the level of detail of models incorporating the
microscale features that need to be obtained from 3-D microstructural imaging.
To this end, X-ray images must be introduced into the models. This can be
conveniently done using solid modeling suitable for volume meshing with
elements to be used in solution methods such as the finite element method
(Mebatsion et al. 2009). Figure 24.4 shows a way to achieve such a microscale
model. The coordinates of individual cells in the X-ray images (Fig. 24.4a)
across a series of slices were estimated after the transformation of digital images
to representative polygons defined by points on the natural boundaries of the
cells. Once the 3-D coordinates of individual cells were established, ellipsoids
that fit these sets of points were generated using a Least Squared Fitted Ellipsoid
c Intra-cellular O2 Concentration [mol/m3] Max: 0.411
×10-4 0.4
2 0.35
1
x
0.3
0
0.25
1
0.2
0
3
0.15
2
z
y x 1
0.1
0
Fig. 24.4 Use of apple 3-D microstructure images for modeling: (a) Synchrotron radiation X-ray
image. (b) Ellipsoid-fitted solid model of X-ray image using method of Mebatsion et al. (2009) and
(c) Computed O2 profile in tissue using model of Ho et al. (2009)
algorithm developed by Mebatsion et al. (2008b). The model tissue geometry

(Fig. 24.4b) was generated from the ellipsoids, which were truncated when
neighboring volumes overlapped. As a result, as many truncated ellipsoids as
there are cellular images were generated, producing a virtual tissue representative
of the fruit microstructure. Figure 24.4c demonstrates the results of the gas
exchange model by Ho et al. (2009) as applied to the 3-D microstructure model
of apple. In a future step, such models could be used to evaluate the tissue
properties of fruit to evaluate the effect of storage and processing conditions.
24.5 Conclusions
Today, internal defects in fruits can only be detected when the fruit is cut. The
whole batch is then assigned as lost; economic risks are therefore high. Early and
nondestructive detection of interior defects and/or aspects that initiate these
defects can prevent severe economicloss. To that end, the role of the fruit
microstructure must be understood and measurement of relevant features online
during preservation and transformation of the fruit should be targeted. With
respect to the first objective, it was demonstrated that the method of synchrotron
radiation X-ray tomography provides a suitable tool to probe the microstructure
of in vivo tissues. To develop fast and cheap online measurement technologies
(e.g., X-ray radiography, nuclear magnetic resonance relaxometry and diffuso-
metry, diffuse spectroscopy, or optical coherence tomography) the presented
method serves as a reference.
Acknowledgments The K.U. Leuven Interfaculty Council for Development Co-operation (IRO),
the K.U. Leuven Research Council (project OT-08023), the Fund for Scientific Research (project
G.06.03.08), and the Institute for the Promotion of Innovation by Science and Technology in
Flanders (projects IWT-060720 and IWT-050633) are gratefully acknowledged for financial
support. Pieter Verboven is Fellow of the Industrial Research Fund of the K.U. Leuven. The
results were obtained with a beamtime project of the European Synchrotron Radiation Facility in
Grenoble, France (experiment MA222).
References
Babin P, Della Valle G, Dendievel R, Lassoued N, Salvo L (2005) Mechanical properties of bread
crumbs from tomography based finite element simulations. J Mater Sci 40:5867–5873
Cloetens P, Ludwig W, Baruchel J, Van Dyck D, Van Landuyt J, Guigay JP, Schlenker M (1999)
Holotomography: quantitative phase tomography with micrometer resolution using hard
synchrotron radiation x rays. Appl Phys Lett 75:2912–2914
Cloetens P, Mache R, Schlenker M, Lerbs-Mache S (2006) Quantitative phase tomography of
Arabidopsis seeds reveals intercellular void network. Proc Natl Acad Sci USA 103:14626–14630
Davis TJ, Gao D, Gureyev TE, Stevenson AW, Wilkins SW (1995) Phase contrast imaging of
weakly absorbing materials using hard X-rays. Nature 373:595–598
Ho Q, Verboven P, Mebatsion H, Verlinden B, Vandewalle S, Nicolaı̈ B (2009) Microscale

mechanisms of gas exchange in fruit tissue. New Phytol 182:163–174
Kim SA, Punshon A, Lanzirotti A, Li L, Alonso JM, Ecker JR, Kaplan J, Guerinot ML (2006)
Localization of iron in Arabidopsis seed requires the vacuolar membrane transporter VIT1.
Science 314:1295–1298
Kuroki S, Oshita S, Sotome I, Kawagoe Y, Seo Y (2004) Visualization of 3-D network of gas-filled
intercellular spaces in cucumber fruit after harvest. Postharvest Biol Technol 33:255–262
Lammertyn J, Dresselaers T, Van Hecke P, Jancsok P, Wevers M, Nicolaı̈ BM (2003) Magn Reson
Imaging 21:805–815
Léonard A, Blacher S, Nimmol C, Devahastin S (2008) Effect of far-infrared radiation assisted
drying on microstructure of banana slices: an illustrative use of X-ray microtomography in
microstructural evaluation of a food product. J Food Eng 85:154–162
Lim KS, Barigou M (2004) X-ray micro-computed tomography of aerated cellular food products.
Food Res Int 37:1001–1012
Mebatsion HK, Verboven P, Ho QT, Mendoza F, Verlinden BE, Nguyen TA, Nicolaı̈ BM (2006)
Modeling fruit microstructure using novel ellipse tessellation algorithm. CMES-Comp Model
Eng 14:1–14
Mebatsion H, Verboven P, Ho Q, Verlinden B, Nicolaı̈ B (2008a) Modelling fruit (micro)
structures, why and how? Trends Food Sci Technol 19:59–66
Mebatsion HK, Verboven P, Jancsók PT, Ho QT, Verlinden B, Nicolaı̈ BM (2008b) Modeling 3-D
fruit tissue microstructure using a novel ellipsoid tessellation algorithm. CMES-Comp Model
Eng 29:137–149
Mebatsion HK, Verboven P, Melese Endalew A, Billen J, Ho QT, Nicolaı̈ BM (2009) A novel
method for 3-D microstructure modeling of pome fruit tissue using synchrotron radiation
tomography images. J Food Eng 93(2):141–148
Mendoza F, Verboven P, Mebatsion HK, Kerckhofs G, Wevers M, Nicolaı̈ B (2007) Three-
dimensional pore space quantification of apple tissue using X-ray computed microtomography.
Planta 226:559–570
Plouraboue F, Cloetens P, Fonta C, Steyer A, Lauwers A, Marc-Vergnes J-P (2004) X-ray high-
resolution vascular network imaging. J Microsc–Oxford 215:139–148
Rau C, Robinson IK, Richter CP (2006) Visualizing soft tissue in the mammalian cochlea with
coherent hard X-rays. Microsc Res Tech 69:660–665
Salvo L, Cloetens P, Maire E, Zabler S, Blandin JJ, Buffière JY, Ludwig W, Boller E, Bellet D,
Josserond C (2003) X-ray micro-tomography an attractive characterisation technique in
materials science. Nucl Instrum Meth B 200:273–286
Schotsmans W, Verlinden BE, Lammertyn J, Nicolaı̈ BM (2004) The relationship between gas
transport properties and the histology of apple. J Sci Food Agric 84:1131–1140
Stuppy WH, Maisano JA, Colbert MW, Rudall PJ, Rowe TB (2003) Three-dimensional analysis of
plant structure using high-resolution X-ray computed tomography. Trends Plant Sci 8:2–6
Thurner P, Muller R, Raeber G, Sennhauser U, Hubbell J (2005) 3D Morphology of cell cultures:
a quantitative approach using micrometer synchrotron light tomography. Microsc Res Tech 66:
289–298
Veraverbeke EA, Van Bruaene N, Van Oostveldt P, Nicolaı̈ BM (2001) Non destructive analysis
of the wax layer of apple (Malus domestica Borkh.) by means of confocal laser scanning
microscopy. Planta 213:525–533
Verboven P, Kerckhofs G, Mebatsion H, Ho Q, Temst K, Wevers M, Cloetens P, Nicolaı̈ B (2008)
Three-dimensional gas exchange pathways in pome fruit characterized by synchrotron X-ray
computed tomography. Plant Physiol 147:518–527
Westneat MW, Betz O, Blob RW, Fezzaa K, Cooper WJ, Lee W-K (2003) Tracheal respiration in
insects visualized with synchrotron X-ray imaging. Science 299:558–560
Chapter 25
Multifractal Characterization of Apple Pore
and Ham Fat-Connective Tissue Size
Distributions Using Image Analysis
Fernando Mendoza, Nektarios Valous, Adriana Delgado, and Da-Wen Sun
25.1 Application of Fractal and Multifractal Analysis

to Biological Material
Fractal and multifractal concepts have been increasingly applied in various fields of
science to describe the complexity and self-similarity of nature. A fractal describes
a rough or fragmented geometric shape that can be subdivided into parts, each of
which is at least approximately a reduced-size copy of the whole. Fractal dimen-
sions offer a systematic approach to quantifying irregular patterns that contain an
internal structure repeated over a range of scales (Mandelbrot 1992). Estimations of
fractal dimension are based on the box-counting technique, which allows obtaining
the scaling properties of 2-D fractal objects (e.g., from binary images) by covering
the images with boxes size e and counting the number of boxes containing at least
one pixel representing the object or structure under study. However, the disadvan-
tage of the box-counting technique is that the process does not consider the amount
of mass inside a box and therefore is not able to resolve regions with high or low
density of mass. Biological materials are frequently complex in the distribution of
their components or structures, which could be of interest to characterize and to
quantify, and therefore, simple fractal dimension estimations may not be enough to
appropriately characterize these materials.
By contrast, multifractal methods are suitable for characterizing a complex spatial
arrangement of mass because they can resolve local densities (Vicsek 1992).
Multifractal formalisms involve decomposing self-similar measures into inter-
twined fractal sets, which are characterized by their singularity strength and fractal
dimension. Multifractal characterization does not require a single dimension, but
rather a sequence of generalized fractal dimensions. Thus, a combination of all the
fractal sets produces a multifractal spectrum that can characterize variability and
heterogeneity of the studied variables (Kravchenko et al. 1999). The advantage of
F. Mendoza (*), N. Valous, A. Delgado, and D.-W. Sun

FRCFT Group, Biosystems Engineering, UCD Agriculture & Food Science Centre, University
College Dublin, Belfield D4, Ireland
e-mail: fmendoza@ing.puc.cl; dawen.sun@ucd.ie
600 F. Mendoza et al.
the multifractal approach is that multifractal parameters can be independent of the

size of the studied objects (Cox and Wang 1993), and that no assumptions are
required about the data following any specific distribution (Scheuring and Riedi
1994). This type of analysis has been successfully applied to plant science, ecology,
and agronomy research for study of vegetation patterns (Scheuring and Riedi 1994),
zooplankton biomass (Pascual et al. 1995), root systems of legumes (Ketipearachchi
and Tatsumi 2000), spatial and temporal variability of residual soil NO3–N and corn
grain yield (Eghball et al. 2003), and soil structure under long-term wastewater
irrigation (Xiaoyan et al. 2007), among others. However, very little further infor-
mation exists about the multifractality in food samples at micro and macro levels
and their related properties.
The objective of this chapter is to give an overview of the multifractal theory as
applied to natural objects and systems, and to illustrate with two examples the
complete application of this approach for characterizing contrasting PSD in apple
tissue and FSD in cooked pork ham images. The identification of potential multi-
fractal parameters useful for quality characterization and classification of these
samples will be a final aim as well. PSD in apple tissue and FSD in pork hams are
the fundamental physical properties analyzed in assessing their quality. In apple
tissue, PSD is related to the mass-transport phenomena characteristics and com-
plexity of O2 and CO2 diffusivity, and in the case of hams, FSD is related to the
sensory properties such as texture, taste, quality of raw meat, and visual appearance.
In both food products, accurate representation of these microstructural properties is
needed for objective quality characterization and prediction during apple preserva-
tion and ham formulation.
25.2 Theory of MFA
Measurement of multifractals mainly involves measuring a statistical distribution,

which in turn yields useful information even if the underlying structure does not
show a self-similar or self-affine behavior (Plotnick et al. 1996). In a homogeneous
system, the number N of features of a certain size e varies (Chhabra and Jensen
1989; Everstz and Mandelbrot 1992; Vicsek 1992):
N ðeÞ a eD0 ; (25.1)
where the fractal dimension D0,
log N ðeÞ
D0 ¼ lim (25.2)
e!0 log 1e
can be measured by counting number N of boxes needed to cover the object under
investigation for increasing box size e and estimating the slope of a log-log plot
25 Multifractal Characterization of Apple Pore and Ham Fat-Connective 601
(i.e., using box-counting method). For heterogeneous or nonuniform systems, the

probability within the ith region Pi scales as:
Pi ðeÞ ¼ eai (25.3)
where ai is the Lipschitz-H€ older exponent characterizing density in the ith box
(Halsey et al. 1986). One technique to determine multifractal parameters is to cover
a measure with boxes size e. The number of boxes N(a) where the probability has
values in the interval (a, a + da) is found to scale as (Halsey et al. 1986; Chhabra
and Jensen 1989):
N ðaÞ a ef ðaÞ (25.4)
where f ðaÞ can be defined as the fractal dimension of the set of boxes with
singularities a. The exponent a can take on values from the interval (a1, a+1),
and f ðaÞ is usually a single-humped function with a maximum at df ðaðqÞÞ=
daðqÞ ¼ 0, where q ¼ 0, fmax is equal to the box-counting or D0 (Vicsek 1992;
Gouyet 1996). In practice, using the box-counting method, for every box i the
probability of a containing object, also called the partition function, is obtained for
different moments q, which can vary from 1 to +1. The partition function is
represented by m(q,e) and defined as (Chhabra and Jensen 1989):
X
NðeÞ
mðq; eÞ ¼ Pqi ðeÞ (25.5)
i¼1
Multifractal sets can also be characterized on the basis of the generalized

dimensions of the qth moment orders of a distribution, Dq (Hentchel and Procaccia
1983). Based on the work of Rényi (1995) they are defined as:
nP
ðeÞ
log Pqi
1 i¼1
Dq ¼ lim (25.6)
e!0 1 q log e
and when q ¼ 1, things become tricky and can be computed as:
ðeÞ
nP
Pi log Pi
i¼1
D1 ¼ lim (25.7)
e!0 log e
The generalized dimension Dq is a monotone decreasing function for all real

q values within the interval (1, +1). When q < 0, m emphasizes regions in the
distribution with less concentration of a measure, whereas the opposite is true for
q > 0 (Chhabra and Jensen 1989). This means that the sum in the numerator of
(25.10) is dominated by the highest values of Pi for q > 0, and the lowest values of
Pi for q < 0.
The partition function (a log-log plot of the quantity m(q,e) over e for different
q yields) scales as:
mðq; eÞ a etðqÞ (25.8a)
or
log mðq; eÞ
tq ¼ lim (25.8b)
e!0 log 1e
where tðqÞ is the mass or correlation exponent of the qth order defined as (Halsey
et al. 1986):
tðqÞ ¼ ðq 1ÞDq (25.9)
The connection between the power exponents f ðaÞ (25.4) and tðqÞ (25.6) is
made via the Legendre transformation (Callen 1985; Halsey et al. 1986; Chhabra
and Jensen 1989):
f ðaðqÞÞ ¼ qaðqÞ tðqÞ (25.10)
and
dtðqÞ
aðqÞ ¼ (25.11)
dq
The f ðaÞ spectrum and the generalized dimensions contain the same information,
both characterizing an interwoven ensemble of fractal dimensions f ðai Þ. In each of
the ith fractals, the observable Pi scales with the Lipschitz-H€older-exponent ai.
The generalized dimensions for q ¼ 0, 1, and 2 mathematically describe the
defined fractal dimensions known as D0, D1, and D2. D0 is also known as a capacity
dimension; it is independent of q and provides global (or average) information on
the system (Voss 1988). D1 is related to the information or Shannon entropy
(Shannon and Weaver 1949), and quantifies the degree of disorder present in a
distribution. According to Gouyet (1996), for a measure m 2 (0,1), the value of D1
is in the range of 0 < D1 < 1. A D1; a value close to 1.0 characterizes a system
uniformly distributed throughout all scales, whereas a D1 close to 0 reflects a subset
of the scale in which the irregularities are concentrated. D2 is mathematically
associated with the correlation function (Grassberger and Procaccia 1983) and
computes the correlation of measures contained in intervals of size e. The relation-
ship between D0, D1, and D2 is D2 D1 D0, where the equality D0 ¼ D1 ¼ D2
occurs only if the fractal is statistically or exactly self-similar and homogeneous
(i.e., monofractal) (Korvin 1992).
25.3 Multifractal Characterization of PSD in Fresh

and Frozen-Thawed Apple Tissue
An apple fruit is mainly composed of the fleshy tissue of parenchyma cells

permeated with vascular and intercellular air spaces (Esau 1977). The structural
geometry of these intercellular spaces, or porous media, plays a fundamental role
in governing the fluid and gas transport through its tissue (Celia et al. 1995;
Dražeta et al. 2004). In general, the gas-filled intercellular spaces in plant organs
are considered as the predominant pathways for gas transport through the plant and
are greatly related to the characteristics of gas exchange (Raven 1996; Kuroki et al.
2004). The volume of air increases during fruit growth and occupies a considerable
proportion of the fruit at harvest (Harker and Ferguson 1988; Yamaki and Ino
1992). This increase in air space is accompanied by a proportional decline in fruit
density, while the density of the fruit cells themselves remains roughly constant
(Westwood et al. 1967; Baoping 1999). In addition, the microstructure and fraction
of intercellular air differs between cultivars and also between positions in the
parenchyma tissue of the same fruit (Baumann and Henze 1983; Vincent 1989).
Larger fruits of the same cultivar have a higher proportion of air than smaller fruits
(Volz et al. 2004). Therefore, the characterization of these intercellular air spaces
and distribution based on microstructural properties and realistic percolation
models have important agricultural applications, since they are related to the
understanding of fruit physiology and postharvest quality of the fruit during
preservation. Fruits with greater fractional air volumes have been shown to be
softer (Yearsley et al. 1997a, b; Volz et al. 2004), or more mealy (Harker and Hallet
1992), and to have greater internal gas diffusion rates (Rajapakse et al. 1990; Ho
et al. 2006). Furthermore, the volume of these intercellular air spaces continues to
increase during storage, and therefore, its measurement can be used to define the
age of the fruit and also to characterize the effects of different storage conditions
on its quality (Khan and Vincent 1990; Harker et al. 1999). All these features could
be related to its known susceptibility to internal tissue browning and breakdown
(Cheng et al. 1998).
On the other hand, freezing is a well-known preservation method widely used in
the food industry. It involves lowering the product temperature generally to 18 C
or below. At temperatures lower than 10 C, few microorganisms can develop,
chemical reactions rates are greatly reduced, and cellular metabolic reactions are
also delayed (Delgado and Sun 2001). One of the key issues in maintaining the
shelf-life and other quality attributes of frozen food is ice crystallization (Kennedy
2003). The freezing process combines the favorable effect of low temperatures with
the conversion of water into ice. The water–ice transition has the advantage of
fixing the tissue structure and separating the water fraction in the form of ice
crystals in such a way that water is not available either as a solvent or a reactive
component (Delgado and Sun 2001). However, the size and location of the
ice crystals may damage cell membranes and break down the physical structure.
Thus, the cause of the undesirable physico-chemical modifications during freezing
is the crystallization of water and sometimes solutes (Martino et al. 1998; Delgado
and Sun 2001). Slow freezing generally leads to large ice crystals formed exclu-
sively on extracellular areas, which can damage cell structure and affect the thaw
behavior as well as have an effect on the sensory properties and nutritional value of
foodstuffs, while high freezing rates produce small crystals evenly distributed all
over the tissue. Therefore, extended research has been carried out to control the
crystal size. Conventional cooling methods, such as air blast, plate contact, circu-
lating brine, and liquid nitrogen (ordered in increasing values of the heat transfer
coefficient) are the most common methods used for food freezing (Sun 2001).
Among these traditional methods, the immersion freezing process offers numerous
advantages, e.g., high-heat-transfer coefficients, individualized freezing, good
product quality, energy savings, etc.
We believe that the spectrum of local fractal dimensions could be an indication
of geometric characteristics of PSD in fresh and frozen-thawed apples. This hope is
solidly based on the fact that the fractal approach provides multifractal outputs that
can be adjusted in order to fit multifractal spectra obtained from experimental data.
Furthermore, it is well known that deviation from local fractal dimensions is greater
for nonuniforms or natural plant structures such as pore distribution in apples;
therefore, local fractal dimensions measured within a small region may vary from
point to point, and the object can be characterized by a spectrum of fractal dimen-
sions. Multifractal analysis (MFA) may improve characterization and discrimina-
tion of changes with high precision in apple tissue microstructure due to processing
and preservation.
25.3.1 Experimental Procedure for Apple Tissue
Figure 25.1 illustrates the experimental procedure for apple sampling, the recon-
structed X-ray image acquisition projection, and the stack of cropped X-ray images
used for further processing and MFA. In this example, a Granny Smith apple
(Malus domestica) with ~8.5 cm dia. (86.9% w/w moisture and ~12 Brix) was
selected from a local market in Dublin (Ireland) and stored at 4 1 C for use later
that same day. This apple variety was chosen as a test material due to the uniformity
of its tissues and its structural stability in storage (Sterling 1968). In addition, the
Granny Smith variety was available all over Ireland that year at a fairly constant
quality. In spite of this, it is well known that high microstructural variability exists
among apples even coming from the same lot, and therefore for analysis, the
samples were extracted from the same apple. The apple cylinders (i.e., two fresh
and two frozen samples) measuring 1.8 cm dia. and 2.5 cm high were cut using a
cork borer in radial orientation from the middle parenchyma (~10 mm from the
skin; Fig. 25.1a) and the same spatial position around the fruit. What remained of
the fruit tissue was used for initial water content and soluble solids determination.
Frozen samples were obtained by immersing samples in a bath system filled with
a freezing solution of 50% ethylene glycol and 50% water in volume. All samples
(fresh and frozen) were loosely wrapped in tissue paper saturated with water to
prevent browning reactions, and kept together in a refrigerator for 24 h at 4 1 C
to achieve uniform initial temperature and thawing of the frozen samples, and until
apple cylinders were imaged. Moisture loss or water uptake during this period was
considered to have little effect.
Samples were scanned with an X-ray micro-CT scanner (SCANCO MEDICAL
AG, mCT-40, Bassersdorf, Switzerland) over the interval 0 –180 (using a 0.9 scan
step) at a linear resolution of 10 mm by pixel, operating voltage of 70 kV, current of
114 mA, and exposure time of 8.4 s. To account for the nonuniformities in the X-ray
beam and nonuniform response of the CCD detector, the raw images were corrected
for dark and white fields by averaging the flat field correction references collected at
the beginning of the experiment. Each sample was enclosed in a plastic tube to avoid
dehydration without any special preparation. The imaging process took approximately
25 min per sample.
The X-ray shadow projections of the 3-D object, digitalized as 2,048 2,048
pixels, were processed using a mathematical back-projection procedure to obtain
reconstructed cross-section images of linear attenuation coefficient values with 256
gradations (8-bit) (Fig. 25.1b). Thus, a stack of 200 cross-sections of grey images of
the object (considering apple tissue and image background), with a 10 mm interslice
distance, was obtained from each scanned apple sample. Thus, for each set of
radiographic images, to extract only the apple tissue image, a fixed area in the
midsection of the scanned region (1,024 1,024 pixels, equivalent to 104.9 cm2)
was cropped and then subjected to image analysis (Fig. 25.1c).
To obtain binary images for the consecutive MFA, the pores need to be identified
and segmented at a certain threshold. However, the distinctions among the voids
and solid phases in the radiographic images of apple tissue are frequently not sharp.
They do not show a bimodal distribution due to the amount of peak overlap in the
attenuation coefficient histogram (Mendoza et al. 2007); as a result, dedicated
segmentation algorithms need to be used to closely represent the pore structure in
the images. Therefore, partition of apple tissue images into pores and cellular
a b c
200
1024 pixels 4
Radial 3
orientation 2
1024 pixels
∅=1.8 cm 1
L = 2.5 cm
Fig. 25.1 Illustration of experimental procedure for MFA of a fresh apple sample: (a) Sampling in
radial orientation; (b) representative X-ray cross-sectional image reconstructed with 256 grada-
tions, showing the cropped region (1,024 1,024 pixels2; black regions represent pores; gray
regions represent cellular material); (c) stack of cropped radiographic images used for further
processing and MFA
material was performed using the kriging-based segmentation algorithm developed

by Oh and Lindquist (1999). The algorithm is a nonparametric formulation able to
analyze regions of uncertainty based on the estimation of spatial covariance of the
image in conjunction with the indicator kriging to determine object edges (Mardia
and Hainsworth 1988). More details about the scanning process and segmentation
algorithm of radiographic apple tissue images can be found in Mendoza et al. (2007).
25.3.2 Extracted Features and Multifractal Spectrum

Computation
The fractal properties were extracted from digitized binary images of fat-connec-
tive tissue structures; 200 images from each type of apple sample, fresh or frozen-
thawed, were evaluated. Parameters calculated from each multifractal spectrum
were: the Hausdorff dimension, f ðaÞ; the singularities of strength, a; and their
generalized fractal dimension, Dq; all were calculated in the range of moment
orders (q) between 10 and +10 taken at 0.1 lag increments. In addition, the bulk
porosity of each apple sample, representing the void space as a fraction of the total
volume, was calculated from the binary segmented images by a simple count of the
pixels in the pore space divided by the total number of pixels in the image. Since
the X-ray CT technique gave a stack of 200 images per sample, the average value of
the computed multifractal parameters and spectrums from each stack of images and
repetition were reported and used in further analyses.
To calculate the f ðaÞ-spectra, the method developed by Chhabra and Jensen
(1989) was implemented in Matlab v7.0 (MathWorks, Inc., USA). Figure 25.2
illustrates the multifractal theory applied to a binary image of fresh apple tissue
corresponding to the cropped image shown in Fig. 25.1b. Thus, images were
partitioned using the box-counting algorithm to estimate the probability of contain-
ing pores (voids) for each box size e, from 2 to 1,024 (i.e., in steps of 2k, 1 < k
< 10). From this information the partition functions and mass or correlation
exponent of the qth order tðqÞ were obtained, as shown in Fig. 25.2b, c. Doing
this for images that are 1,024 1,024 pixels avoids artifacts, which occur when
boxes do not entirely cover the image at the borders. A family of normalized
measures, mi(q,e), was constructed for positive and negative values of q covering
variable ranges in steps of 0.1:
Pqi ðeÞ
mi ðq; eÞ ¼ (25.12)
ðeÞ
NP
Pqi ðeÞ
i¼1
where Pi(e) is the fraction (or probability) of pores contained in each ith box size e.
Note that for any value of q, the normalized measures take values in the interval (0, 1).
The direct computation of f ðaÞ values was made using the simplified relations
proposed by Chhaabra et al. (1989), and Chhabra and Jensen (1989):
ðeÞ
NP
mðq; eÞ log½mi ðq; eÞ
i¼1
f ðqÞ ¼ lim (25.13)
e!0 log e
Similarly, values of aðqÞwere computed by evaluating:
NP
ðeÞ
mðq; eÞ log½Pi ðeÞ
i¼1
aðqÞ ¼ lim (25.14)
e!0 log e
For each q, values f ðqÞ and aðqÞwere obtained from the slope of plots of the
numerators in (25.13) and (25.14) versus log e over the entire range of e values
considered. The range of q values over which both functions were linear (Dq)
was selected considering the coefficients of determination (R2) of both fits
(Fig. 25.2d, e). The f ðqÞ and aðqÞ functions (Fig. 25.2f) obtained over a given
Dq were used to construct f ðaÞ-spectra as an implicit function of q and e. The
symmetry of multifractal spectra was evaluated by comparing the width of the
spectra from their center [a(0)] to a( j qi j ). Values of j qi j were the same in both
the positive and negative domains and equal to the smaller of the two defining
the Dq interval.
a b Partition fuction: Spiq, q → –10: 1 : +10

c t (q), Mass or correlation
p i in boxes, steps of 2k → 1< k th
exponent of the q order
60 –10
–8 15
45 –6
1024×1024 pixels
–4 10
30 –2
–1 5
LogΣpiq
15 0
t(q)
0 1 0
2 –5
–15 4
6 –10
–30 8
–45 10 –15
0.0 0.5 1.0 1.5 2.0 2.5 3.0 –10 –5 0 5 10
Log(ε) Moment, q
d e f
0 0 2.2 f (q )
Simi(q,e) log[mi(q,e)]
a(q )
Simi(q,e) log[pi(e)]
–1 –1 2.0
f (q ), a(q )
–2 –2 1.8
–3 –3
q = –1 q = –1 1.6
–4 q=0 –4 q=0
q=1 q=1 1.4
–5 q=2 –5 q=2
1.2
10 100 1000 10 100 1000 –10 –5 0 5 10
e e Moment, q
Fig. 25.2 Illustration of multifractal theory applied to a binary image of fresh apple tissue
(corresponding to cropped image in Fig. 25.1b): (a) Binary image (pores are represented by
white pixels); (b) partition functions calculated in range of moment orders (q) between 10 and
+10, but showing only lag increments of 2; (c) mass or correlation exponent of the qth order tðqÞ;
(d) and (e) Plots of (25.14) and (25.13), respectively, as a function of e; (f) Illustration of f(q) and
a(q) spectrums
On the other hand, generalized dimensions were obtained as the slope of the
partition function over box size, both taken as logarithms (25.8) and (25.9). This
method is known as the method of moments (Everstz and Mandelbrot 1992) or
Rényi spectrum (Rényi 1995), as Dq is estimated for every moment q.
25.3.3 Results of MFA for PSD in Apples
To give a better idea about the microstructure of apple tissue, Fig. 25.3 shows
representative radiographic images extracted from fresh and frozen-thawed apple
tissue. The samples are characterized by different porous structures, illustrating the
deleterious effect of freezing on the apple’s microstructure. In the frozen-thawed
sample, the cell disruption and tissue due to ice crystal formation is evident, wherein
the pores are larger and rounded compared to fresh apple tissue. Also, differences
(heterogeneity) in the bulk porosity distribution between cultivars are visually evident
(Table 25.1). The average porosity computed from two stacks of apple images
extracted from the same fruit (taken parallel to the medial axis of each fruit, 0 and
180 around the fruit; 200 images per stack, area per image ¼ 104.9 cm2) showed
values 14.1 0.8% for fresh and 20.7 1.4% for frozen-thawed tissue. ANOVA
analysis of the average porosity confirmed the differences (P-value <0.05) between
apple samples and revealed a higher variability between images from frozen-thawed
tissue than those computed from fresh tissue, as represented by their standard
deviations.
A crucial step in MFA is to determine the range of both e and the moments of
order q over which a multifractal method is applicable. This means determining the
range of e and q in which the numerators of (25.13) and (25.14) are linear functions
of log e. The range of q values was selected considering the coefficients of
determination (R2) of the fits (Fig. 25.2d, e). In this study, the general criterion
was to choose the range of moments with R2 equal or larger than 0.95. Thus, the
condition was met for a(q) in the range Dq of 10 to +10, and for f(q) in the range
Dq of 1.4 to +5.8, for both fresh and frozen-thawed samples.
The corresponding f ðaÞ-spectra and Rényi spectra (range of moment order q
between 10 and +10) for each apple type are shown in Fig. 25.4a, b, respec-
tively. The f ðaÞ-spectra shows the typical hump-shape observed for multifractal
objects, but there are distinct differences in shape and symmetry between the
fresh and frozen-thawed apple samples (Fig. 25.4a). The curvature and symmetry
Table 25.1 Selected multifractal parameters and standard deviations from the analysis of binary
images of pores for fresh and frozen-thawed apple tissue
Porosity (%) a(0) a(1) a(1) a(0) D0 D1 f[a(1)] D0 D1/D0
Fresh 14.1 0.8x 0.214 0.015x 0.602 0.084x 0.090 0.006x 0.190 0.015x 0.955 0.003x
Frozen 20.7 1.4y 0.229 0.016y 0.791 0.129y 0.092 0.008y 0.238 0.025y 0.954 0.003y
x–y, values with different letters within each column indicate significant differences among fresh
and frozen apples (p < 0.05)
Fig. 25.3 Reconstructed 3-D X-ray apple tissue images (2003 pixels3 8 mm3): (a) Fresh;
(b) frozen-thawed tissue (dark regions represent pores; clear regions represent cellular material).
The pore space in frozen-thawed tissue appears larger and apparently more heterogeneous in size
than in fresh apple tissue, evidencing the deleterious effect of freezing on the apple microstructure
Fig. 25.4 Multifractal spectrums of PSD in fresh and frozen-thawed apple tissues expressed by:
(a) f ðaÞ-spectra; (b) Rényi dimensions spectra, Dq
of the f ðaÞ-spectra provide information on the heterogeneity of a system, defined

by the diversity of scaling exponents needed to characterize it. A homogeneous
fractal exhibits a narrow f ðaÞ-spectra, whereas the opposite is true for a hetero-
geneous fractal. Thus, PSD in fresh apple tissue is more homogeneous than in
frozen-thawed tissue as revealed by their f ðaÞ-spectra in Fig. 25.4a.
On the other hand, the Rényi spectra or generalized dimensions of the long
transect are sigma-shaped curves with a clear asymmetry with respect to the cut
point and the vertical axis; there is much more curvature for negative values of q than
for positive values, which tend to be constant (Fig. 25.4b). More specifically, they
exhibited pronounced decreasing Dq values with increasing q, and showed clear
statistical differences between apple samples for Dq values with moments lower than
q ¼ 0. In this example, the evaluated range of Dq ran from D10 ¼ 3.140 0.086
to D+10 ¼ 1.664 0.022 for fresh apple tissue, and from D10 ¼ 3.339 0.122 to
D+10 ¼ 1.686 0.021 for frozen-thawed tissue. However, the width of the Dq-
spectra was greater in the frozen-thawed tissue, which indicated that the heteroge-
neity of PSDs in the parenchyma tissue was more apparent than in fresh apple, also
confirming that the visual differences appreciated in the reconstructed 3-D images in
Fig. 25.3.
In addition, heterogeneity can be assessed at q ¼ 0 by the magnitude of differ-
ences in the values D0 and að0Þ, or more generally, by the magnitude of changes
around D0 in both f ðaÞ and a axes. Table 25.1 shows selected multifractal para-
meters for characterizing the heterogeneity of fresh and frozen-thawed apple
samples. In the present example, except for D1/D0, the computed parameters
(a(0) a(1), a(1) a(0), D0 D1, and f[a(1)] D0) showed statistical dif-
ferences (P-value <0.05) between apple samples. In a symmetric spectrum, the
widths ranging from a(0) to a(jqij), and D0 to D(jqij) are expected to be similar and
smaller. In this sense, the results presented in Table 25.1 clearly confirm that PSD in
fresh apple tissue is more homogeneous than that of frozen-thawed tissue. This
tendency toward homogeneity implies that regions with high and low concentration
of mass (pores) scale similarly.
25.4 Multifractal Characterization of FSD for Two Qualities

of Presliced Pork Hams
In ham products, FSD is a fundamental physical property analyzed in assessing

product quality. Here, FSD is related to sensory properties such as texture, taste,
quality of the raw meat, and visual appearance. Therefore, accurate representation of
this microstructural property is needed for an objective quality characterization and
prediction during ham formulation. Moreover, ham slices in general have complex
and inhomogeneous colored surfaces, and textures do not contain any detectable
periodic or quasiperiodic structure. Instead, textures exhibit random but persistent
patterns that result in a cloud-like appearance (Mendoza et al. 2009). These inhomo-
geneities can be attributed mainly to formulation, presence of pores/defects and fat-
connective tissue, and color variations (Valous et al. 2009). Thus, for objective
characterization, image analysis techniques need to take into account the high varia-
bility in texture appearance. Consequently, fractal metrics and concepts have been
recently applied to investigate and describe the complexities and self-similarities of a
variety of ham surfaces (Mendoza et al. 2009; Valous et al. 2009).
Similar to the multifractal characterization of PSD in fresh and frozen-thawed apple
tissues (10 mm/pixel), we have presented an application of multifractal theory for
characterization of FSD in two qualities of presliced pork hams typically consumed in
Ireland. Here, the extracted multifractal parameters and spectrums were computed
using the same principles and conditions mentioned above; nevertheless, it is impor-
tant to note that the MFA in this case was tested for characterization of structures
segmented from ham images captured at the macroscale level (0.102 mm/pixel).
25.4.1 Experimental Procedure for Ham Samples
Two qualities of pork hams were manufactured at Dawn Farm Foods Co. (Kildare,
Ireland), using different muscle sections and various percentages of brine solutions
(wet curing by injection). Specifically, the high yield ham (low quality) had a 50%
brine injection and was made from pork muscle called Silverside PAD (biceps
femoris), cut to 100 mm width. The injected muscle was vacuum-tumbled at
1,500 rpm for 12 h and vacuum-filled into PVC casings, clipped, and cooked at
82 C to a core temperature of 72 C. The medium-yield ham (intermediate quality)
had a 30% brine injection and contained three leg muscles: Topside, Silverside, and
Knuckle. The injected muscle was vacuum-tumbled at 500 rpm for 5 h and vacuum-
filled into PVC casings, clipped, and cooked at 82 C to a core temperature of 72 C.
All pork ham samples were chilled to 4 C before slicing. Images were acquired
immediately after slicing (100 slices per type of ham quality).
To ensure reproducibility in the preprocessing and MFA of ham images, a color-
calibrated computer vision system (CVS) as described by Valous et al. (2009) was
used for image acquisition (spatial resolution 0.102 mm/pixel). Then, a polynomial
transform was used for signal calibration of the captured images (Valous et al.
2009), which mapped the raw RGB primary signals into the sRGB color standard
(IEC 1999). These color-calibrated images were used in the identification and
segmentation of fat-connective tissue structures. For analysis, due to the large
variation in size and shape between the two presliced ham qualities, the images
were subsequently cropped in the central region to produce 512 512 pixel images
(equivalent to 2,727.4 mm2). The software package MATLAB (MathWorks, USA)
was used for image processing and multifractal computations.
The fat-connective tissue segmentation was performed using the green (G) inten-
sity images (from sRGB), since visually this color channel better represented the
edges of the fat-connective tissue structures of the two evaluated hams. Thus, the
green (G) intensity image was preprocessed with a median filter (3 3) to remove
impulse noise. Then, a morphological closing filter (3 3) was applied to remove
pores; a decrease in the intensity variations of the background pixels (more “flat”
background in relation to fat-connective tissue) was also carried out. A hi-pass filter
(15 15) was applied for increasing sharpness and contrast. This operation intro-
duced some random noise, which was then removed using the median filtering
operation (3 3), while a subsequent histogram-based segmentation using a unique
threshold value of 200 produced the binary images (Valous et al. 2009).
Figure 25.5 shows representative images of the two evaluated pork ham qualities
as well as the experimental procedure for cropping and binarization of the fat-
connective tissue structures. The original ham images were cropped to obtain three
square regions (512 512 pixels2) from each slice; since not all ham slices had the
same spatial dimensions, this process allowed receiving representative information
from each quality sample, facilitating better scrutiny and interpretation, and also
keeping computation times manageable. Also, the depicted binary images confirm
the performance of the segmentation method used in this investigation.
Fig. 25.5 Illustration of experimental procedure for MFA using ham samples: (a) Representative
images of the two evaluated pork ham qualities; (b) cropped color regions (512 512 pixels2);
(c) corresponding binary images of segmented fat-connective tissue structures used for further
processing and MFA
25.4.2 Results of MFA for FSD in Hams
As observed in Fig. 25.5, the evaluated ham qualities are characterized by different
contents and distributions of fat-connective tissue structures; they also show a high
variability in the textural appearance of slices from the same quality type. Fat-
connective tissue estimations (average of 300 square images per quality type)
showed values for high yield of 6.2 2.1% and for medium yield 2.9 1.6%.
The statistical differences (P-value <0.05) were evident between these average
indices (Table 25.2).
The range of q values for ham samples, considering the coefficients of determi-
nation (R2) equal or larger than 0.95, for the plot of numerators (25.13) and (25.14)
versus log e, were for a(q) in the range Dq of 1.7 to +10 and for f (q) in the range
Dq of 1.6 to +3.1 for both ham qualities.
Figures 25.6a, b depict the average f ðaÞ-spectra and Rényi spectra for each ham
quality and their standard deviations for each point, respectively. The f ðaÞ-spectra
revealed that FSD in medium-yield hams is slightly more homogeneous than in
high-yield hams (Fig. 25.6a), as indicated by the shorter amplitude and symmetric
shape of the average f ðaÞ-spectra for the medium-yield ham. Similarly, the Rényi
spectra (Fig. 25.6b) exhibited pronounced decreased Dq values with increasing q,
and showed clear statistical differences between the ham qualities in the entire
range of Dq values (with q moments ranging from 10 to þ10). In this example, the
Table 25.2 Selected multifractal parameters and standard deviations from the analysis of binary
images of fat-connective tissue for the two qualities of presliced hams
Fat-connective a(0) a(1) a(1) a(0) D0 D1 f [a(1)] D0 D1/D0
tissue index (%)
High 6.23 2.07x 0.385 0.042x 0.616 0.069x 0.164 0.020x 0.267 0.023x 0.909 0.013x
Medium 2.89 1.56y 0.393 0.070x 0.470 0.119y 0.174 0.029y 0.233 0.054y 0.888 0.021y
x–y, values with different letters within each column indicate significant differences among ham
qualities (p < 0.05)
Fig. 25.6 Multifractal spectrums of FSD in two qualities of presliced hams expressed as: (a) f ðaÞ-
spectra; (b) Rényi dimensions spectra, Dq
evaluated range of Dq ran from D10 ¼ 2.920 0.142 to D+10 ¼ 1.382 0.107
for high-yield hams and D10 ¼ 2.429 0.226 to D+10 ¼ 1.326 0.140 for
medium-yield hams, with an average D0 dimension for high- and medium-yield
hams of 1.808 0.080 and 1.572 0.180, respectively. Nonetheless, the width of
Dq-spectra is greater in high-yield hams, which indicates that the heterogeneity of
FSDs in the ham matrix is more apparent than in medium-yield types. In general,
both multifractal spectrums allowed for a robust characterization of these two types
of ham. The selected multifractal parameters presented in Table 25.2 confirm that,
with exception of the difference a(0) a(1) and in spite of the high variability of
samples coming from the same quality, the computed parameters a(1) a(0),
D0 D1, f[a(1)] D0, and D1/D0 are all statistically different (P-value <0.05)
and therefore could be used as quality predictors of hams.
25.5 Conclusions and Outlook
In this chapter, multifractal spectrum and the generalized dimensions of apple pore
structure are computed with data obtained from X-ray images (10 mm/pixel) of
fresh and frozen-thawed apple samples (Granny Smith), as well as color images
(0.102 mm/pixel) of two different qualities of presliced pork hams typically
consumed in Ireland. The extracted multiscaling properties from binary images

(of pores or fat-connective tissue structures) allowed evaluating and revealing the
significance of the method in discrimination of morphology and distribution of
pores and fat-connective tissue structures in these particular samples.
Multifractal parameters were found to reflect the major aspects of variability in
the PSD of apple tissue and FSD of ham fat-connective tissue, providing a unique
quantitative characterization of the spatial distribution data. The irregularity and
complexity of apple pore structure and ham fat-connective tissue, together with its
scale-invariant features suggest that multifractal distribution is a suitable and
natural model.
The results of both examples have also demonstrated that MFA has significant
benefits for quantitative analysis of the complex microstructure of biological
materials from images, allowing conclusions to be drawn on the exact topography
of, for example, PSD and FSD of apple tissue and presliced cooked pork hams.
Potential multifractal parameters for characterizing contrasting PSD and FSD are D0,
D1, a(0) a(1), a(1) a(0), D0 D1, f[a(1)] D0, and D1/D0. Furthermore,
modeling apple and ham images with multifractal spectrums, such as f ðaÞ-spectra
and Dq-spectra, allows the capture of differentiating microstructural features and
makes possible the automatic characterization, not only between apple varieties and
ham types, but also between roughness degrees or image textures, due to the
capability of the multifractal parameters to resolve local densities from images.
Multifractal parameters are promising descriptors for type identification and quality
assessment of apple and ham samples. Moreover, this multiscaling procedure and
image analysis technique should provide valuable opportunities for further research
of other natural and processed foods.
Acknowledgements The authors gratefully acknowledge the Food Institutional Research

Measure (FIRM) strategic research initiative, as administered by the Irish Department of Agricul-
ture and Food, for their financial support.
References
Baoping J (1999) Nondestructive technology for fruits grading. In: Proceedings of 1999 interna-
tional conference on agricultural engineering, Beijing, China, pp IV127–IV133
Baumann H, Henze J (1983) Intercellular space volume of fruit. Acta Hort 138:107–111
Callen HB (1985) Thermodynamics and an introduction to thermostatistics, 2nd edn. Wiley,
New York
Celia M, Reeves P, Ferrand L (1995) Recent advances in pore scale models for multiphase flow in
porous media. Rev Geophys Suppl 33:1049–1057
Cheng Q, Banks NH, Nicholson SE, Kingsley AM, Mackay BR (1998) Effects of temperature
on gas exchange of “Braeburn” apples. NZ J Crop Hort Sci 26:299–306
Chhabra A, Jensen RV (1989) Direct determination of the f(alpha) singularity spectrum. Phys Rev
Lett 62(12):1327–1330
Chhaabra AB, Meneveu C, Jensen RV, Sreenivasan KR (1989) Direct determination of the
f(alpha) singularity spectrum and its application to fully developed turbulence. Phys Rev A
40:5284–5294
Cox LB, Wang JSY (1993) Fractal surfaces: measurements and applications in earth sciences.
Fractals 1:87–117
Delgado AE, Sun D-W (2001) Heat and mass transfer models for predicting freezing process – a
review. J Food Eng 47(3):157–174
Dražeta L, Lang A, Alistair JH, Richard KV, Paula EJ (2004) Air volume measurement of
“Braeburn” apple fruit. J Exp Bot 55:1061–1069
Eghball B, Schepers JS, Negahban M, Schlemmer MR (2003) Spatial and temporal variability of
soil nitrate and corn yield: multifractal analysis. Agron J 95:339–346
Esau K (1977) Anatomy of seed plants. Wiley, New York
Everstz CJG, Mandelbrot BB (1992) Multifractal measures. In: Peitgen H, J€urgens H, Saupe D
(eds) Chaos and fractals. Springer, Berlin, pp 922–953
Gouyet J-F (1996) Physics and fractal structures. Springer, New York
Grassberger P, Procaccia I (1983) Characterization of strange attractors. Phys Rev Lett 50(5):346–349
Halsey TC, Jensen MH, Kadanoff LP, Procaccia I, Shraiman BI (1986) Fractal measures and their
singularities: the characterization of strange sets. Phys Rev A 33(2):1141–1151
Harker FR, Ferguson IB (1988) Calcium ion transport across discs of the cortical flesh of apple
fruit in relation to fruit development. Physiol Plant 74:695–700
Harker FR, Hallet IC (1992) Physiological changes associated with development of mealiness of
apple fruit during cool storage. HortScience 27:1291–1294
Harker FR, Watkins CB, Brookfield PL, Miller MJ, Reid S, Jackson PJ, Bieleski RL, Bartley T
(1999) Maturity and regional influences on watercore development and its postharvest disap-
pearance in “Fuji” apples. J Am Soc Hortic Sci 124:166–172
Hentchel HGE, Procaccia I (1983) The infinite number of generalized dimensions of fractals and
strange attractors. Physica D 8:435–444
Ho QT, Verlinden BE, Verboven P, Nicolaı̈ BM (2006) Gas diffusion properties at different
positions in the pear. Postharvest Biol Technol 41:113–120
IEC (1999) IEC 61966–2–1: multimedia systems and equipment – colour measurements and
management – Part 2–1: colour management – default RGB color space – sRGB. International
Electrotechnical Commission (IEC), Geneva, Switzerland
Kennedy C (2003) Developments in freezing. In: Zeuthen P, Bøgh-Sørensen L (eds) Food
preservation techniques. CRC Press, Cambridge/England, pp 228–240
Ketipearachchi KW, Tatsumi J (2000) Local fractal dimensions and multifractal analysis of the
root system of legumes. Plant Prod Sci 3:289–295
Khan AA, Vincent JFV (1990) Anisotropy of apple parenchyma. J Sci Food Agric 52:455–466
Korvin G (1992) Fractals models in the earth sciences. Elsevier, Amsterdam, The Netherlands
Kravchenko AN, Boast CW, Bullock DG (1999) Multifractal analysis of soil variability. Agron J
91:1033–1041
Kuroki S, Oshita S, Sotome I, Kawagoe Y, Seo Y (2004) Visualization of 3-D network of gas-filled
intercellular spaces in cucumber fruit after harvest. Postharvest Biol Technol 33:255–262
Mandelbrot BB (1992) The fractal geometry of nature, 2nd edn. W.H. Freeman, New York
Mardia KV, Hainsworth TJ (1988) A spatial thresholding method for image segmentation. IEEE
Trans Pattern Anal Mach Intell 6:919–927
Martino MN, Otero L, Sanz PD, Zaritzky NE (1998) Size and location of ice crystals in pork
frozen by high-pressure-assisted freezing as compared to classical methods. Meat Sci
50(3):303–313
Mendoza F, Verboven P, Mebatsion HK, Kerckhofs G, Wevers M, Nicolaı̈ BM (2007) Three-
dimensional pore space quantification of apple tissue using x-ray computed microtomography.
Planta 226:559–570
Mendoza F, Valous NA, Allen P, Kenny TA, Ward P, Sun D-W (2009) Analysis and classifica-
tion of commercial ham slice images using directional fractal dimension features. Meat Sci
81(2):313–320
Oh W, Lindquist W (1999) Image thresholding by indicator kriging. IEEE Trans Pattern Anal
Mach Intell 21:590–602
Pascual M, Ascioti FA, Caswell H (1995) Intermittency in the plankton: a multifractal analysis of
zooplankton biomass variability. J Plankton Res 17:1209–1232
Plotnick RE, Gardner RH, Hargrove WW, Prestegaard K, Perlmutter M (1996) Lacunarity
analysis: a general technique for the analysis of spatial patterns. Phys Rev E 53(5):5461–5468
Rajapakse NC, Banks NH, Hewett EW, Cleland DJ (1990) Development of oxygen concentration
gradients in flesh tissues of bulky plant organs. J Am Soc Hortic Sci 115:793–797
Raven JA (1996) Into the voids: the distribution, function, development and maintenance of gas
spaces in plants. Ann Bot 78:137–142
Rényi A (1995) On a new axiomatic theory of probability. Acta Mathematica Hungarica VI
3–4:285–335
Scheuring I, Riedi RH (1994) Application of multifractals to the analysis of vegetation pattern.
J Veg Sci 5:489–496
Shannon CE, Weaver W (1949) The mathematical theory of communication. University of Illinois
Press, Chicago
Sterling C (1968) Effect of low temperature on structure and firmness of apple tissue. J Food Sci
33:577–580
Sun D-W (ed) (2001) Advances in food refrigeration. Leatherhead Publishing, LFRA Ltd, Surrey
Valous NA, Mendoza F, Sun D-W, Allen P (2009) Colour calibration of a laboratory computer
vision system for quality evaluation of pre-sliced hams. Meat Sci 81(1):132–141
Vicsek T (1992) Fractal growth phenomena, 2nd edn. Word Scientific Publishing, Singapore
Vincent JFV (1989) Relationships between density and stiffness of apple flesh. J Sci Food Agric
31:267–276
Volz RK, Harker FR, Hallet IC, Lang A (2004) Development of texture in apple fruit – a
biophysical perspective. Acta Hort 636:473–479
Voss RF (1988) Fractals in nature: from characterization to simulation. In: Peitgen H-O, Saupe D
(eds) The sciences of fractal images. Springer, New York, pp 21–69
Westwood MN, Batjer LP, Billingsley HD (1967) Cell size, cell number and fruit density of apples
as related to fruit size, position in the cluster and thinning method. Proc Am Soc Hort Sci
91:51–62
Xiaoyan G, Peiling Y, Shumei R, Yunkai L (2007) Multifractal analysis of soil structure under
long-term wastewater irrigation based on digital image technology. NZ J Agric Res
50:789–796
Yamaki S, Ino M (1992) Alteration of cellular compartmentation and membrane permeability to
sugars in immature and mature apple fruit. J Am Soc Hortic Sci 117:951–954
Yearsley CW, Banks NH, Ganesh S (1997a) Temperature effects on the internal lower oxygen
limits of apple fruit. Postharvest Biol Technol 11:73–83
Yearsley CW, Banks NH, Ganesh S (1997b) Effects of carbon dioxide on the internal lower
oxygen limits of apple fruit. Postharvest Biol Technol 12:1–13
Part V
Food Packaging
.
Chapter 26
New Packaging Materials Based on Renewable
Resources: Properties, Applications,
and Prospects
Stéphane Guilbert, Carole Guillaume, and Nathalie Gontard
26.1 Introduction
Since the 1980s, research programs on high-value materials based on renewable

resources have boomed as a result of a larger demand for packaging materials with
better environmental balance and/or new or original properties, such as biodegrad-
ability, versatility according to physico-chemical conditions, controlled release or
adsorption, etc. These new materials are designed to have a better environmental
impact than conventional plastics. They are considered “bio-plastics” (Guilbert
et al. 2005) even if the definition of this term is not clear: the polymer’s origin
might be a fossil or from renewable resources, its synthesis can be chemical or
naturally made, and its end-life, conventional/biodegradable or compostable. The
different ways to obtain bio-plastics (commercially available or under develop-
ment) are presented in Fig. 26.1.
Materials based on “extractible” agropolymers such as starch, cellulose, or
proteins have been obtained through either casting techniques or thermoplastic
processing. Their water and temperature sensitivity can be built on to react to
fresh food physiology and/or to trigger responses adaptive to environmental
change. These materials also exhibit interesting properties related to mass transport
characteristics: they generally have a high carbon dioxide/oxygen (CO2/O2) selec-
tivity ratio associated with high gas permeability. Such bio-plastics, which can be
combined with paper or bio-polyesters for mechanical resistance, are able to create
unique modified atmospheres favorable to the preservation of fresh fruits and
vegetables (Gontard and Guillaume 2009). Bio-plastics based on proteins are also
used as an inclusion matrix for a controlled release of volatile antimicrobial agents
to increase food shelf-life. Their release kinetics and thus antimicrobial efficiency
can be modulated thanks to physically induced crosslinking treatments or introduc-
tion of nanoparticules to suit to the food distribution chain (Guillaume et al. 2008;
S. Guilbert (*), C. Guillaume, and N. Gontard

Joint Research Unit, Agropolymers Engineering and Emerging Technologies, Montpellier
SupAgro, INRA, UM II, CIRAD, 34060 Montpellier, France
e-mail: guilbert@supagro.inra.fr
620 S. Guilbert et al.
Renewable raw material

Fossil Raw
material
Bio-refinery product / co-product
Agro-polymer Fermentation /
(extractible GM Green chemistry
polymer) Agro- “additived”
Mono/oligomers
polymer and “oxo-
Chemical modification degradable”
Microbial or catalyst
Bio-polyester
Bio-polyester biodegradable
Bio-polyester, long polyesters
chain starch... Bio-polyolefins...
Starch, protein or
cellulose based Blends, composites,
Nano-composites, multilayers...
Fig. 26.1 Actual bio-plastics, commercial or under development
Gontard and Guillaume 2009). Nonconventional extractible polyesters are also now
produced either by microorganisms or genetically modified (GM) plants (Fig. 26.1).
Apart from extractible polymers, bio-sourced plastic polymers are produced by
chemical synthesis of monomers obtained by fermentation of renewable substrates.
Today, they are bio-polyesters, bio-polyolefins, and bio-elastomers. Some are still
under development, and others, such as polylactic acid (PLA), are already com-
mercialized for the formulation of packaging materials (Fig. 26.1). Their properties
are identical or close to conventional material properties, with renewability as a
strong functional and commercial advantage. Some of these bio-sourced polymers
(e.g., bio-polyesters) are biodegradable or compostable (e.g., PLA).
Biodegradable fossil-based plastic materials (generally, polyesters) have been
developed in order to facilitate processability optimization and to improve mechan-
ical, barrier, optical, and biodegradability properties. Their combination with
extractible agropolymer materials, or bio-sourced polyester materials, or fibers, to
form blends or composite materials, has given good test results. This could be a
solution to optimize production costs and modulate the biodegradability kinetics
and mechanical and transport properties.
26.2 Bio-plastics: Where Are We Now?
Bio-plastics can be classified into three categories according to chemistry used to

synthesize the polymer (conventional/biosynthesis), origin (natural/fossil), and end-
life (conventional/biodegradation or composting). Reference to the term “bio” can
be assigned to either of these categories (Table 26.1). A fourth category called
“additived polymers” is sometimes cited as “degradable” or even “biodegradable”
26 New Packaging Materials Based on Renewable Resources 621
Table 26.1 The main groups of bio-plastics

Polymer group Origin Chemistry End life Examples
Nature-made Renewable Naturally Biodegradable or Starch, cellulose,
polymers engineered compostable proteins,
PHAs
Chemically Renewable Conventional Biodegradable or PLA
synthesized synthesis compostable
polymers Fossil Conventional Biodegradable or PCL, PBAT,
synthesis compostable PBSA
Renewable Conventional Not biodegradable Bio-sourced
synthesis polyolefins,
PE, PP
(Fig. 26.1). It refers to conventional polymers (e.g., polyethylene [PE] and polypro-
pylene [PP]) mixed with 1–5% additives, which are supposed to promote the
degradation of the entire material. Additives are organic materials such as starch,
cellulose or ethylene vinyl alcohol (called “Additived with organic materials”), or
transition metals (e.g., cobalt, zinc, manganese) that catalyze oxidation and chain
scission (called “oxo-degradable”). In the first case, only a small portion of the
material biodegrades; the material is considered “fragmentable.” In the second case,
the pro-oxidation catalysts promote polymer chain scission when the material is
exposed to an abiotic environment (temperature, O2, UV, etc.). Both types of
additived polymers do not meet the standard specifications for compostable or
biodegradable plastics (ASTM D64000 or D6868). In addition, the environmental
consequences of their usage are largely unknown (e.g., ecotoxicity or accumulation
of nano-size products of chain scission). Therefore, additive polymers should
definitively be excluded from the bio-plastics family.
The first group of bio-plastics refers to “nature-made polymers” that have been
naturally engineered by vegetal or microbial cells (i.e., crop plants, industrial
microorganisms), which are then extracted, purified, and eventually modified.
These polymers are renewable. The main benefits are high yields from natural
fabrication and excellent carbon and energy balance. In addition, functional proper-
ties (e.g., mass transport properties) are often original and far from conventional
plastic properties, thus paving the way for new applications such as active or
intelligent materials (see Sect. 26.3). The formulation of bio-plastic materials
based on extractible agropolymers implies the use of polyesters, polysaccharides
or proteins. These polymers have the capability to form a matrix with sufficient
cohesion and continuity, a crystalline or amorphous continuous structure, under
certain conditions. Starch is the most commonly used agricultural raw material, and
is inexpensive, widely available, and relatively easy to handle. “All-starch” bio-
plastics are made from thermoplastic starch and are formed with standard techniques
(as used for synthetic polymer films) such as extrusion or injection molding. The use
of thermoplastic proteins was also investigated (Gontard and Guilbert 1994; Guilbert
et al. 2006), but commercial applications are still expected. Among the proteins, milk
proteins (casein, whey proteins), soya proteins, and cereal proteins (wheat gluten,
zein, etc.) have been more extensively studied (Guilbert and Cuq 2002).
The glass transition temperature and melting temperature of polysaccharides and

proteins are very high (above degradation temperature); thus, the use of plasticizers,
such as water, glycerol or lactic acid is required. One of the inconveniences using
water-soluble plasticizers is the migration towards and contact with the product
(food), leading to either a loss of material properties or pollution of the product.
This type of material based on a continuous hydrocolloid matrix is generally not
very resistant to water, as moisture-barrier properties are poor. In some cases, water
solubility or sensitivity to water is a functional advantage, e.g., in the formulation of
soluble sachets. For the majority of usages, reducing the plasticizer’s migration and
the improvement of water resistance and water-barrier properties are of prime
importance. Chemical modifications, such as grafting with hydrophobic com-
pounds or usage of specific additives (cross-linking agents or nonwater-soluble
plasticizers), adapted to the polymer structure have been proposed but with poor
success. In most cases, gains in water or mechanical resistance are counterbalanced
by a strong loss in elongation at strength. Regarding these developments, internal
plasticization through chemical grafting or design of copolymer materials (e.g.,
protein/polyesters) is the most promising (Auvergne et al. 2008). Another interest-
ing development is the formation of nano-reinforced materials, for example with
exfoliated nanoclays.
Commercial water-resistant, starch-based bio-plastics are produced by using fine
molecular blends of biodegradable synthetic polymers (see below) and starch. These
materials are made with gelatinized starch (up to 60%) and hydrophilic (e.g.,
ethylene vinyl alcohol copolymer [EVOH]) or hydrophobic synthetic biodegradable
polymers (e.g., polycaprolactone [PCL], or polybutylene adipateterephthalate
[PBAT], known as “Ecoflex®”) and compatibility agents (Fritz et al. 1994). The
polyesters form the continuous phase leading to materials presenting a relative water
resistance and acceptable barrier and mechanical properties. The most important
starch-/polyester-based materials on the market are those proposed by Novamont as
Materbi® or by Sphere as Bioplast.® Actual production of these starch/polyester
blends is estimated between 50 and 80 kT/year with a price between 1.2 and 3.0€/kg
according to the grade considered.
Microbial polyesters (e.g., polyhydroxyalkanoates [PHAs]) are excreted or stored
by microorganisms cultivated on starch hydrolysates or lipidic mediums. PHAs are
unbranched polymers predominantly composed of R-3-hydroxyalkanoic acid mono-
mers, ranging from 3 to 14 carbons in length with hydrogen or alkyl up to nonyl
radicals. Copolymers of different length monomers or introduction of other monomers
are often proposed. Isolation and purification costs could be high for these products, as
they are obtained from complex mixtures. Monsanto stopped the commercialization
of the poly(3)-hydroxybutyrate-hydroxyvalerate (PHB/V) product Biopol® in 1999.
Since that date, others have entered the market and have built new facilities for pilot
plant production of microbial PHA, as in the case of Tianan Biologic Material
produced in Ningbo (China), Biocycle® by PHB Industrial in Brazil, or “Telles,” a
Metabolix/Archer Daniels Midland joint venture that plans to produce “Mirel.” DSM
(Nl) announced a $20 million dollar investment in China in partnership with TIANJIN
to produce 10,000 t/year of PHA. Recently, the production of PHAs with GM crop
plants was proposed with the potential for producing large amounts of bio-polymers at
a low cost. A commercially relevant yield of 7–14% dry weight has been achieved
with transgenic sugarcane or switch grass. Metabolix is one of the major industrial
actors in this field.
PHA-based plastic properties look like PE ones but their rigidity is much higher
due to a higher glass transition temperature (+5 /+10 C, compared to 120 C for
PE). Actual production of these microbial polyesters is still very low (between 0.6
and 1.2 kT/year) with cost between 1.8 and 5.0€/kg according to the quality and
production process.
The second group of bio-plastics refers to “chemical synthesis” polymers,
which are also sometimes called “artificial bio-polymers.” They can be: (i) bio-
sourced (renewable origin) and biodegradable or compostable (e.g., PLA), (ii) bio-
sourced and nonbiodegradable (e.g., bio-PE, bio PP), or (iii) petrol-based (fossil
origin) but biodegradable (e.g., PBAT or PBSA). These polymers are synthesized
by conventional chemical polymerization of monomers issued either from fossil oil
cracking or renewable resource “deconstruction” (e.g., products of a plant biore-
finery). These may benefit from the possibility of designed tailor-made polymers
with controlled properties but the environmental impact is sometimes controversial.
Polylactic (and polyglycolic) acids are mainly produced by chemical polymeri-
zation of lactic acid (and glycolic acid) obtained by Lactobacillus fermentation.
Biomass is used for fermentation derivatives of corn flour (corn starch hydroly-
sates), but use of the whole plant (including straw and lignocellulosic parts) is
expected in the near future (second generation). Commercial applications of PLA
materials are growing very rapidly under the trademarks of Ingeo™ from Cargill
(USA) or Purasorb from Purac (NL). PLA-based plastic properties are similar to PE
properties but their rigidity is much higher due to a higher glass transition tempera-
ture (þ5 /þ60 C, instead of 120 C for PE). PLA biodegrades at temperatures
higher than Tg; the standard PLA (Tg ¼ þ50/þ60 C) is not considered as biode-
gradable according to the ASTM standards but is compostable. In 2008, PLA
production was slightly above 70 kT/year with an average price of 1.3–2.6€/kg.
Bio-sourced but not biodegradable/compostable plastics polymers (biopolyole-
fins, biopolyurethan, polycarbonate, etc.), produced with monomers from renew-
able resources and “green chemistry” transformation, have been announced for the
near future. For example, Braskem, Brazil’s largest petrochemical firm, invested
US$150 million in a 200 KT plant to produce PE from sugarcane-based ethanol
in 2009.
Petrol-based biodegradable polyesters are produced mainly by chemical synthe-
sis, by companies such as Basf with Ecoflex (-PBAT-), Showa Highpolymer with
Bionolle (polybutene succinate adipate -PBSA-), and Solvay with Capa (Polyca-
prolactone -PCL-). Actual production of these polyesters is between 10 and 25 kT/
year with price around 2.0–3.0€/kg. Several research projects to replace these last
fossil biodegradable polyesters with renewable ones are underway. Some are based
on plant oil, fatty acid or glycerol chemical modification, including polymerization.
Other projects are based on the production by fermentation of some monomers for
copolymer building. As an example, Roquette (FR) and DSM (NL) announced the
future building of a pilot plant for the production by fermentation of succinic acid
for PBSA fabrication.
26.3 Applications of Bio-plastics
All commercial bio-plastic materials can be processed using conventional shaping

processes (e.g., extrusion, injection, injection molding) or coating processes
(casting of solution in adequate solvents). Their applications can be classified into
four categories: (i) plastics to be composted; (in fields where reuse or fine recovery
is difficult; for short shelf-life products such as disposable items or fast food
packaging; in fields where recycling is not industrially feasible, such as fiber
composite), (ii) plastics used in natural environments (in fields where recovery is
not economically or practically feasible, such as mulching of plastics for agricul-
ture), (iii) specialty plastics (in fields with specific features where bio-plastics
possess preferential properties), and (iv) commodity plastic substitutes with renew-
able origin (supposedly has a better environmental impact than nonbio-sourced
ones).
Due to their relatively high cost and low availability, actual commercial applica-
tions of biodegradable plastics are limited to special niches with environmental
considerations. Loose fill packaging and compost bags are the two major end uses,
constituting nearly 50% of demand in 2008 but applications in the field of agricul-
ture and food packaging are expected to dominate in the near future.
Actual commercial bio-plastic materials have different features that can be
matched with different applications. For instance, considering the two major
products on the market, Mater-bi material of Novamont and PLA of Cargill, the
Mater-bi material is more suitable for films or bags, while PLA is mostly used for
packaging and trays (unless plasticized). Manufacturers presently have a double
function as both buyer and provider. Further, producers using the same applications
are not yet competitive; they blend polymers from one producer to another to obtain
specific or intended material properties.
Commercial applications of biodegradable materials are growing very rapidly,
but because they are complex to develop and sometimes hard to produce, the total
production of “thermoplastic bio-plastics” is actually low, with a total world
consumption (2008) of only 120,000–150,000 t. Thus, large-scale application is
not possible. This value is inferior to the production capacity of bio-plastics, which
is around 300,000 t; a 2-year shift from capacity is observed.
The market for biodegradable plastics is young compared to that of traditional
plastics. It started 20 years ago and has been booming for 2 or 3 years (Advanced
Bioplastics 2003). World production passed from the pilot scale (during 1990s)
to the industrial scale. However, this amount represents 0.2–0.3% of the total
plastic market distributed using biodegradable polymers, ranging from fossil
origin (50,000 t) to renewable resources (250,000 t). The global production
capacity of biodegradable polymers is expected to reach 1,000,000 t in 2015.
Nonbiodegradable, bio-sourced plastics are under development and/or are at the

pilot plant production stage (e.g., bio-sourced PE from sugarcane alcohol in Brazil).
Completion of the first commercial products is expected in 2010 (but prices are not
available). Expected markets are linked to public or private markets requiring a
minimal and mandatory ratio for renewable based materials (e.g., plastic- or fiber-
based material for construction).
Presently, the bio-plastic market is under the authority of a limited number of
producers, such as American Cargill, Italian Novamont, and German BASF, and to
a lesser extent, French Sphere and Biolice. Others entering the market are American
Dupont, Procter & Gamble, Novomer and Metabolix, British UCB, Brazilan Bio-
cycle and Braskem, French Roquette, and Dutch DSM.
Actual manufacturers expect to expand their production capacity, and construc-
tion of numerous plants is planned in Europe, USA, Brazil, and China. However,
the existing economic crisis combined with some uncertainty about the environ-
mental impact of bio-plastics (few global life cycle analyses [LCAs]) might hamper
these ambitions.
26.4 Original Properties and Active Materials

for Food Packaging
There are very few reliable data concerning the environmental footprint of bio-
plastics. Some life cycle analysis (LCA) is available but the scope of such is
generally limited to the factory (from cradle to gate). Without cradle to tomb
analysis (e.g., cradle to composting) or better, cradle to cradle, it is very difficult
to advise on the real environmental impact (compared to conventional plastic
materials). It is also true that due to the absence of an optimized supply chain on
one part, and industrial local composting facilities on another part, it is unfair to
compare an emerging chain to a well-established one. For these reasons, govern-
mental agencies generally support the R&D effort in the domain of bio-plastics
study, to help and promote the emergence of a viable industry.
Beyond environmental consideration, the development of bioplastics is driven
by their unique and original properties when compared to conventional plastics.
This is mainly due to very different molecular architecture and in most cases their
hydrophilic character. These specific properties, mainly observed for extractible
agropolymers, can be of great interest for designing gas selective or active materials
for food packaging. Gas selectivity, obtained by the ratio of gas permeability,
represents the ability of materials to favor transfer of one gas compared with
another. For instance, various CO2/O2 permeation ratio is expected in modified
atmosphere packaging (MAP) to meet physiological requirements of a large range
of fresh produces. Active packaging has deliberately incorporated active agents that
are intended to release or absorb substances into, onto or from the packaged food or
the environment surrounding the food, and is defined in the food contact material
framework regulation 1935/2004 (Danielli et al. 2008). Such new packaging tech-
nologies are expected to play a major role in response to consumer demand for more
convenient, safer, and mildly preserved products with longer storage duration. In
addition, globalization of markets results in longer distribution distances and major
challenges in the food packaging industry, acting as a driving force for development
of packaging concepts that extend shelf-life while maintaining food quality and
safety. This concept is developing throughout Europe and is enhancing the compet-
itiveness of the food and packaging industry, with a huge potential market. Japan
and the USA share more than 50% and 22% of the global Active and Intelligent
(A&I) packaging market, respectively, against only 15% for EU. Simultaneously,
the use of modified atmosphere packaging (MAP), passive packaging with gas-
selective materials, or active packaging with additional gas (or vapor or any other
volatile agent) emitter or absorber for the preservation of fresh or minimally
processed fruits and vegetables constitutes one of the most important challenges.
To illustrate these properties, we chose to develop the case of an extractible
agropolymer: wheat gluten proteins. Wheat gluten films appears to be of great interest
because of their high gas permeability, adapted to the respiration of fresh fruits and
vegetable, and high selectivity values (CO2 permeability/O2 permeability), which
are able to create a unique low O2 and CO2 atmosphere adaptable to the preservation
of CO2-sensitive commodities (Gontard et al. 1996; Barron et al. 2002). However,
despite their low cost and interesting functional properties for food packaging,
these protein-based materials needs to be combined with:
– fiber-based materials, such as paper to overcome its poor mechanical properties
(Han and Krochta 1999; Rhim et al. 2006; Gastaldi et al. 2007),
– or nanofillers such as montmorillonite (MMT) to reduce moisture sensitivity (Wang
et al. 2005; Hedenqvist et al. 2006; Olabarrieta et al. 2006; Tunc et al. 2007).
An experimental support paper was coated with an acid wheat gluten solution
(20% w/v, pH4) containing or not MMT (2.5% w/w of proteins) to produce
composite or nanocomposite materials, respectively. O2 and CO2 permeability of
composite and nanocomposite materials was evaluated at 25 C, and 80% and 90%
relative humidity (RH), respectively, in comparison to control uncoated materials.
The continuous layer formed by wheat gluten proteins on support paper greatly
reduced gas permeability of the paper. Coated paper could thus not be considered as
porous any longer. O2 and CO2 permeability values were not significantly affected
by the presence of MMT in the wheat gluten network regardless of RH. Since
permeability is known to be governed by two mechanisms, diffusion and sorption, it
was assumed that introduction of MMT did not change solubility nor diffusivity of
O2 and CO2, as observed in a previous study (Tunc et al. 2007). It should be pointed
out that O2 and CO2 permeability of both composite and nanocomposite materials
increased when increasing RH. This phenomenon was also observed on pure wheat
gluten films [Gontard et al. 1996], suggesting that the wheat gluten-based coating
layer is the key element of gas-barrier properties in the studied materials (composite
and nanocomposite). The increasing RH effect was attributed to a modification of
the wheat gluten network structure and polymeric chain mobility and is related to
the change from the glassy to the rubbery state (Gontard et al. 1996; Gontard and
Ring 1996). The increase of CO2 permeability was more pronounced than O2
permeability. This phenomenon was explained by a selective sorption of CO2 due
to the specific interactions occurring between CO2 and the water-plasticized protein
matrix, especially the high-content amide groups of wheat gluten protein (Pochat-
Bohatier et al. 2006). At high RH value, adsorption of water should provide better
accessibility to active sites of CO2 sorption located on the mobile polymeric protein
chains. As a consequence, gas selectivity (COE permeability/O2 permeability) was
highly affected by RH (from 80 to 90% RH) and rose from 1.9 to 7.9 and 1.5 to 7.5
for composite and nanocomposite materials, respectively. These results show that
the unique gas selectivity properties are preserved when combined with paper or
MMTs. If we consider the potential use of the nanocomposite material for MAP of
fresh produces at 90% RH, CO2 should go out the packaging headspace at a rate
7.5-fold higher than O2 comes in.
Passive MAP experiments were conducted on parsley with the uncoated support
paper as a control and the nanocomposite material at a controlled RH of 80%. As
expected for a highly porous material, O2 and CO2 partial pressures at the steady
state obtained using control paper were close to air composition (21 and 0 kPa,
respectively). Such an atmosphere was detrimental to the quality of the produce.
After only 4 days of storage, more than 50% of ascorbic acid and chlorophyll were
lost, and parsley leaves were fully yellow. Nanocomposite material generated a
headspace atmosphere containing lower O2 (11 kPa) and higher CO2 (4 kPa)
content. This steady atmosphere clearly improved the quality attributes of parsley
during storage by maintaining a high chlorophyll content (directly linked to pars-
ley’s green color) and ascorbic acid during 8 days of storage. A critical level of 60%
initial vitamin C content was reached after only 3 days of storage with uncoated
paper, against more than 8 days for composite material. The use of nanocomposite
materials for MAP of parsley led to equilibrium atmosphere favorable to maintain-
ing parsley quality, by slowing down oxidation and physiological reactions respon-
sible for product degradation. Similar results were obtained with mushrooms, and
other fruits and vegetables.
Another promising application of extractible agropolymers, and in particular
protein-based materials, is their use as active antimicrobial packaging. This relies
on the ability of bio-based polymers to entrap active compounds and to release
these compounds in a controlled way, with a moisture- and temperature-triggering
effect (Carlin et al. 2001; Gennadios and Weller 1991; Guilbert et al. 1997).
Protein-based materials have also been demonstrated to be efficient for antimicro-
bial packaging because of their ability to control, in a relevant way, the release of
volatile extracts of various essential oils (allylisothyocyanate, carvacrol, cinnamal-
dehyde, and eugenol) (Ben Arfa et al. 2007a, b; Chalier et al. 2007) that could open
new routes for the development of active MAP.
It has been shown that the release of the compound carvacrol from paper coated
with wheat gluten is RH-dependent with or without the addition of nano-particles of
MMTs. Release of this volatile compound was assessed at 25 C on all materials as a
function of time, using a two-step gradient for RH. Composite material lost more
than 70% of carvacrol within 20 days of storage at 60% RH. This meant that only
30% of active agent would be available for release into the food during storage
within packaging. Placed at 100% RH, 30% of active agent was entirely released
within 8 days. In the presence of MMT introduced into the wheat gluten network
(nanocomposite material), only 20% of carvacrol was released during 20 days of
storage at 60% RH. Consequently, 80% of the volatile active agent remained
available for release during the period food is packaged. Placed at 100% RH,
80% of active agent was entirely released within 13 days (from day 22 to 35).
It can be concluded from these results that the release of cavacrol from wheat
gluten-coated paper is RH-dependent with or without MMT. Such behavior is
highly interesting for both limiting volatile active agent losses before using the
material as food packaging, and for triggering the active agent release in the
presence of the food.
26.5 Conclusions
Considering the relative novelty of packaging materials based on renewable

resources, the following conclusions or question can be drawn:
– There is no clear definition of “bio-plastic,” and different concepts sometimes
contradictory are emerging accordingly in reference to polymer origin (natural/
fossil), chemistry (conventional/biosynthesis), and/or end-life (conventional/
biodegradable).
– Properties of biodegradable polymers are still far from conventional plastics
(i.e., glass transition temperature, melting temperature, processability, mechan-
ical properties, and water sensitivity).
– The necessity to use plasticizers for many extractible polymers-based materials
is a major drawback for food contact materials (how to control their migration
into the product).
– Biodegradable packaging is often not suitable for packaging of long shelf-life
products with high water content (water sensitivity of materials and growth of
microorganisms on packaging, e.g., on some made of proteins that need a
combination with nanofillers to overcome this drawback).
– Prices are still high and susceptible to change according to the cost of renewable
raw materials.
– Some of the most promising polymers are obtained through GMO plant or GMO
microorganism origins (e.g., PLA, PHA, etc.), whereas “GMO” origin is not
compatible with the “green” image of biodegradable or bio-sourced products.
– Apart from extractible polymers, which are of limited application, other bio-
plastics are obtained through a deconstruction/reconstruction procedure (e.g.,
natural polymer hydrolysis, followed by fermentation, purification, polymeriza-
tion stages, etc.). What is the rationality for such a complex and costly procedure?
– In actual context, it is difficult to analyze the life cycle of bio-plastics from cradle
to cradle due to poor availability of materials, lack of composting facilities, the
necessity to collect materials separately, and the sorting of bio-plastic material

waste for correct treatment to avoid pollution and interaction with industrial
recycling facilities for conventional materials (e.g., PET recycling).
– What is the economic and environmental rationality for bio-sourced conven-
tional plastics? Orientation on extractible plant and microbial polymers (starch,
protein, cellulose, bio-polyesters, etc.) is probably more rational than producing
monomers from natural origins (i.e., deconstruction of high-value natural
polymers) for reconstruction of bio-sourced plastic. Due to competition in the
food/nonfood usage of agricultural products (even if relatively low land surface
is required for production of bio-plastics compared to other usages such as
energy or food), the rationality is to use limited volumes; one can also recom-
mend that bio-plastics be reserved for high-value specialty plastics (materials
with original/specific properties such as biodegradability, respirability, gas
selectivity, controlled release, etc.), not for substitution of commodity plastics.
These conclusions sound very negative but this is mainly due to the lack of
reliable data on the subject, and to the fact that some materials or concepts (e.g.,
bio-sourced conventional plastics) are very new and to the tendency of scientific
and economic actors to treat the subject in a passion mode. More research is needed
in polymer/materials science, plasturgy, properties enhancement, migration studies,
safety issues of biodegradable food contact materials, eco-conception, LCA, etc.
References
Auvergne R, Morel M-H, Menut P, Guilbert S, Robin JJ (2008) Reactivity of Wheat gluten protein
during mechanical mixing: radical and nucleophilic reactions for the addition of molecules on
sulfur. Biomacromolecules 9:664–671
Barron C, Varoquaux P, Guilbert S, Gontard N, Gouble B (2002) Modified atmosphere packaging
of cultivated mushroom (Agaricus bisporus L.) with hydrophilic films. J Food Sci 67:251–255
Ben Arfa A, Chrakabandhu Y, Presiozi-Belloy L, Chalier P, Gontard N (2007a) Coating paper
with soy protein isolates as inclusion matrix of carvacrol. Food Res Int 40:22–32
Ben Arfa A, Presiozi-Belloy L, Chalier P, Gontard N (2007b) Antimicrobial paper based on soy
protein isolate or modified starch coating including carvacrol and cinnamaldehyde. J Agric
Food Chem 55:2155–2162
Chalier P, Ben Arfa A, Presiozi-Belloy L, Gontard N (2007) Carvacrol losses from antimicrobial
soy protein coated papers as a function of drying paper conditions. J Appl Polym Sci
106:611–620
Carlin F, Gontard N, Reich M, Nguyen-The C (2001) Utilization of zein coating and sorbic acid to
reduce Listeria monocytogenes growth on cooked sweet corn. J Food Sci 66:1385–1389
Fritz HG, Seidenst€ucker T, B€olz U, Juza M, Schroeter J, Endres HJ (1994) Study on production of
thermoplastics and fibers based mainly on biological materials. Science Research Develop-
ment. European Comission, EUR 16102EN
Gastaldi E, Chalier P, Guillemin A, Gontard N (2007) Microstructure of protein-coated paper as
affected by physico-chemical properties of coating solutions. Colloid Surf A 301:301–310
Gennadios A, Weller CL (1991) Edible films and coatings from soymilk and soy protein. Cereal
Food World 36:1004–1009
Gontard N, Guilbert S, Cuq JL (1993) Water and glycerol as plasticizers affect mechanical and
water vapor barrier properties of an edible wheat gluten film. J Food Sci 58(1):206–211
Gontard N, Guilbert S (1994) Bio-packaging: technology and properties of edible and/or

biodegradable material of agricultural origin. In: Mathlouthi M (ed) Food packaging and
preservation. Blackie Academic & Professional, Glasgow, pp 159–181
Gontard N, Thibault R, Cuq B, Guilbert S (1996) Influence of relative humidity and film
composition on oxygen and carbon dioxide permeabilities of edible films. J Agric Food
Chem 44:1064–1069
Gontard N, Ring S (1996) Edible wheat gluten film: influence of water content on glass transition
temperature. J Agric Food Chem 44:3474–3478
Gontard N, Guillaume C (2009) Food packaging and the shelf life of fruits and vegetables. In:
Robertson GL (ed) Food packaging and shelf life: a practical guide. CRC Press, Boca Raton,
FL, chapter 16
Guilbert S, Cuq B, Gontard N (1997) Recent innovations in edible and/or biodegradable packaging
materials. Food Addit Contam 14:741–751
Guilbert S, Cuq B (2002) Protein as raw material for biodegradable products. In: Bastioli C (ed)
Handbook of biodegradable polymers. Rapra tech Ltd. London
Guilbert S, Feuilloley P, Bellon Fontaine V (2005) Biodegradable polymers in agricultural
applications. In: Smith R (ed) Biodegradable polymers for industrial applications. Woodhead
Publishing, Cambridge, chapter 25
Guilbert S, Morel MH, Gontard N, Cuq B (2006) Protein-based plastics and composites as smart
green materials. In: Bozell JJ, Patel MK (eds) Feedstocks for the future: renewables for the
production of chemicals and materials. Oxford University Press, New York, pp 334–350,
(n 921). American Chemical Society
Guillaume C, Chalier P, Gontard N (2008) Modified atmosphere packaging using environmentally
compatible and active food packaging materials. In: Chiellini E (ed) Environmentally compat-
ible food packaging. Woodhead Publishing, Cambridge, pp 396–418
Han JH, Krochta JM (1999) Wetting properties and water vapor permeability of whey-protein-
coated paper. Tranns ASAE 42:1375–1382
Hedenqvist MS, Backman A, Gallstedt M, Boyd RH, Gedde UW (2006) Morphology and diffusion
properties of whey/montmorillonite nanocomposites. Compos Sci Technol 66:2350–2359
Olabarrieta I, Gallstedt M, Ispizua I, Sarasua JR, Hedenqvist MS (2006) Properties of aged
montmorillionite-wheat gluten composite films. J Agric Food Chem 54:1283–1288
Pochat-Bohatier C, Sanchez J, Gontard N (2006) Influence of relative humidity on carbon dioxide
sorption in wheat gluten films. J Food Eng 77:983–991
Rhim JW, Lee JH, Hong SI (2006) Water resistance and mechanical properties of biopolymer
(alginate and soy protein) coated paperboards. LWT Food Sci Technol 39:806–813
Tunc S, Angellier H, Cahyana Y, Chalier P, Gontard N, Gastaldi E (2007) Functional properties of
wheat gluten/montmorillonite films processed by casting. J Membr Sci 289:159–168
Wang SF, Shen L, Tong YJ, Chen L, Phang IY, Lim PQ, Liu TX (2005) Biopolymer chitosan/
montmorillonite nanocomposites: preparation and characterization. Polym Degrad Stab 90:123–131
Chapter 27
Edible Coatings to Improve Food Quality
and Safety
Noemı́ Zaritzky
27.1 Introduction
Edible films and coatings are useful materials mainly produced from edible biopo-
lymers and food-grade additives (GRAS). Films are usually made from polymers
that are able to provide mechanical strength to a stand-alone thin structure
(Han 2005; Han and Gennadios 2005). Coatings are a particular form of films
directly applied to the surface of materials. An edible coating has a close and
continuous association with the food until consumption and it is regarded as a
part of the final product. Edible coatings enhance the quality of food products,
protecting them from physical, chemical, and biological deterioration (Kester and
Fennema 1986; Han and Gennadios 2005).
The application of edible coatings can improve the physical strength of food
products, reduce particle clustering, and improve visual and tactile features on
product surfaces (Cuq et al. 1995; Cisneros-Zevallos et al. 1997). The coatings
can also protect food products from moisture migration, microbial growth on the
surface, light-induced chemical changes, oxidation of nutrients, etc. Edible coatings
can act as barriers against oils, gases, or vapors and as carriers of active substances
(antioxidants, antimicrobials, colors, and flavors) (Kester and Fennema 1986;
Gennadios and Weller 1990; Guilbert and Gontard 1995; Krochta and De
Mulder-Johnston 1997; Miller et al. 1998). These functions enhance the quality
of food products, resulting in shelf-life extension and safety improvement. Further,
edible coatings can be utilized as active films when applied to modify the atmo-
sphere of food surface conditions (Cuq et al. 1995; Guilbert and Gontard 2005).
As an example of edible films, yuba (soy-milk skin) has been traditionally used in
Asian countries since the fifteenth century (Wu and Bates 1972; Park et al. 2002). Wax
coatings were applied to citrus fruits in the twelfth and thirteenth centuries, but only
N. Zaritzky
Centro de Investigación y Desarrollo en Criotecnologı́a de Alimentos (CIDCA), Universidad
Nacional de La Plata (UNLP) – CONICET La Plata and Depto de Ingenierı́a Quı́mica, La Plata,
Argentina
e-mail: zaritzky@ing.unlp.edu.ar
632 N. Zaritzky
commercially utilized on apples and pears since 1930 (Baldwin 1994; Debeaufort
et al. 1998; Park 1999). Lipid coatings (larding) on meats and cheeses have been used
since the Middle Ages for shrinkage prevention (Donhowe and Fennema 1994).
The objectives of this presentation are to: (1) discuss the basic concepts of food
coating applications, (2) review the most significant publications in this area, and
(3) describe the different experiments carried out in our laboratory studying the
performance of edible coatings to improve food quality and safety.
27.2 Edible Coating Materials
Coating materials can be either hydrophilic or hydrophobic; the main film-forming

materials are biopolymers, such as proteins and polysaccharides (carbohydrates
and gums); lipids and resins are also used, however they are not biopolymers
(Gennadios et al. 1997). Coating materials can be used alone or in combinations
(Krochta et al. 1994; Baldwin et al. 1997). The formulation of bioplastics or edible
films implies the use of at least one component able to form a matrix, having
sufficient cohesion and continuity (Nussinovitch 2003).
Polysaccharides: The most common polysaccharides used for edible coatings
are methyl cellulose (MC), carboxymethyl cellulose (CMC), hydroxypropyl cellu-
lose (HPC), hydroxypropyl methyl cellulose (HPMC), starch, modified starches,
amylose, hydroxypropyl amylose (HPA), alginate, carrageenan, pectin, chitosan,
gellan gum, xanthan gum, etc. The occurrence of relatively large numbers of
hydroxyl groups in the structure of polysaccharides indicates that hydrogen bonds
may play significant roles in film formation and characteristics. Starch is the
most common agricultural raw material used for biodegradable and edible films
formulation, and is an appropriate matrix-forming material with lower cost com-
pared to other alternatives. Starch is also widely available and relatively easy to
handle; amylose is responsible for the film-forming capacity of starch (Guilbert and
Gontard 2005). Cellulose derivatives are composed of linear chains of b (1–4)
glucosidic units with methyl, hydroxypropyl, or carboxyl substituents. MC has
excellent film-making properties, as well as high solubility and efficient O2 and
lipid-barrier properties (Donhowe and Fennema 1994; Nisperos-Carriedo 1994). Two
other examples, chitin and its deacetylated product, chitosan, have received much
interest regarding application in the food industry due to their biocompatibility,
biodegradability, and bioactivity.
Proteins: Different proteins such as whey protein, casein, collagen, gelatin, egg
white protein, keratin, fish, myofibrillar protein, corn zein, wheat gluten, soy protein,
pea protein, peanut protein, cottonseed protein, rice bran protein, etc., have been used
in edible coatings. The most distinctive characteristics of proteins, compared to other
film-forming materials, would be the conformational denaturation, electrostatic
charges, and amphiphilic nature of proteins. Many factors can affect the conforma-
tion of proteins, such as charge density and hydrophilic-hydrophobic balance, which
in turn can affect the physical and mechanical properties of films and coatings.
27 Edible Coatings to Improve Food Quality and Safety 633
Protein film-forming materials are derived from animal tissues, milk, eggs, grains,
and oil seeds (Krochta 1997, 2002). Among the proteins considered, milk proteins
(casein, whey proteins), soy proteins, and cereal proteins (wheat gluten, zein) have
been more extensively studied (Gennadios et al. 1994; Guilbert et al. 2001). Since
several edible food coatings are made with proteins that can cause allergic responses
in some consumers and problems in people affected by Celiac disease (gluten
intolerance), the use of a food coating with a known allergen or the presence of
gluten in a food must be appropriately declared and clearly labeled on the product.
Waxes/Lipids/Resins: In this group, beeswax, candelilla wax, carnauba wax,
cocoa butter, milk fat fractions, acetylated monoglycerides, fatty acids, and resins
(shellac, terpene) are used in edible coatings. It should be noted that lipids and
resins are not biopolymers; nevertheless, they are edible, biodegradable, and con-
sidered to be cohesive water barrier biomaterials. Most lipids and edible resins are
soft-solids at room temperature and show characteristic phase transition tempera-
tures. Films or coatings made from lipids have very high water resistance and low
surface energy, due to their hydrophobic nature. Lipids are generally applied in thin
layers or as a composite with a polymeric matrix. Lipids can be combined with
other film-forming materials, such as proteins or polysaccharides, as emulsion
particles or as multilayer coatings in order to increase the resistance to water
penetration (Greener and Fennema 1989; Greener and Fennema 1992; Baldwin
et al. 1997; Gennadios et al. 1997; Perez-Gago and Krochta 2002; Perez-Gago and
Krochta 2005).
Plasticizers: The structures of films containing polysaccharides and proteins are
often brittle and stiff due to extensive interactions between the polymer molecules
(Krochta 2002). The film-forming mechanisms of biopolymers include intermole-
cular forces such as covalent bonds (e.g., disulfide bonds and cross-linking) and/or
electrostatic, hydrophobic, or ionic interactions. The addition of plasticizers over-
comes this film brittleness and improves flexibility and extensibility. Plasticizers are
low molecular weight agents incorporated into the polymeric film-forming material;
the plasticizer positions itself between the polymer molecules and interferes with the
polymer-polymer interaction, increasing the flexibility and processability of
the material (Guilbert and Gontard 1995). Plasticizers must be compatible with the
film-forming polymer; they reduce the intermolecular forces and increase the
mobility of polymer chains. Plasticizers increase the free volume of polymer
structures or the molecular mobility of polymer molecules (Sothornvit and Krochta
2000, 2001), decreasing the ratio between the crystalline and the amorphous regions,
and decreasing the glass transition temperature (Tg) of the polymer (Krochta 2002;
Guilbert and Gontard 2005). Most plasticizers are very hydrophilic and hygroscopic;
they attract water molecules and form a large hydrodynamic plasticizer-water
complex. Hydrophilic compounds, such as polyols (glycerol, sorbitol, and polyethy-
lene glycol) are commonly used as plasticizers in hydrophilic film formulations.
Functional additives: This group includes antioxidants, antimicrobials, nutri-
ents, nutraceuticals, flavors, and colorants, which can be combined with film-
forming biopolymers to modify the functionality of films or coatings. The most
beneficial characteristics of edible films and coatings are their edibility and inherent
634 N. Zaritzky
biodegradability (Guilbert et al. 1996). To maintain edibility, all film components

(i.e., biopolymers, plasticizers, and other additives) should be food-grade ingredients
and all process facilities should be acceptable for food processing; solvents used
should be restricted to water and ethanol.
27.3 Application and Distribution of Edible Coatings

on the Food Surface
The application and distribution of the film-coating material in a liquid form can be
achieved by hand spreading with a paint brush, spraying, falling film enrobing,
dipping and subsequent dripping, and distribution in a revolving pan (pan coating),
etc. (Guilbert 1986). Edible coating formulations must be wet when spread on the
food’s surface; upon drying the formulation must form a film coating with adequate
adhesion, cohesion, and durability to function properly.
Different tests can be performed on the filmogenic suspensions before the
coating is applied on a foodstuff. Rheological analysis and viscoelastic properties
of the suspensions help to determine the suitability of the food coating method
(immersion, spraying, etc.). Rheological behaviors (flow curves) are normally eval-
uated using viscometers; oscillatory rheometers allow measurement of the visco-
elastic parameters of the film-forming suspension (Garcı́a et al. 2004b, 2006;
Lopez et al. 2008).
Some negatively charged gums, such as alginate, pectin, and CMC, show
significantly different rheological properties in acidic conditions as opposed to
neutral or alkaline conditions. Rheological behavior and surface tension of the
film-forming suspension are important factors linked to suspension spreadability
and coating adhesion capacity. A decrease in the surface tension of the filmogenic
suspension leads to a better coating adhesion to foodstuffs. The peel or surface of
many vegetables has low surface tension for protection purposes; however, this
natural advantage is a drawback in the case of aqueous coating applications. The
properties of the edible coating depend on the type of film-forming materials and
especially on their structural characteristics (Han and Gennadios 2005).
Cohesion is the attractive force between the molecules of the same substance; it
depends on the structure of the polymer, the molecular length, geometry, and
molecular weight distribution, and on the type and position of the lateral groups.
Cohesion of film-forming materials influences the mechanical strength of the films
(Guilbert et al. 1996). If the film-forming material contains different components
that are not compatible with the main biopolymers, the cohesion decreases and film
strength weakens. Plasticizers reduce the cohesion of film-forming polymers.
Adhesion of film-forming materials is an important parameter for film casting
and coating processes (Guilbert and Gontard 2005). Adhesion is an attractive force
between the surface molecules of coating materials and food surfaces. A low
adhesion force leads to an incomplete coating on the food surface. A large diffe-
rence between the surface energy of a coating material, and the uncoated product
surface, results in a poor coating performance. Surface active agents, such as
emulsifiers in the film-forming solution, reduce the surface tension of the coating
solution; then the difference between the solid surface energy and the surface
tension of the coating solution decreases, increasing the work of adhesion (Cuq
et al. 1995; Guilbert and Gontard 2005).
27.4 Edible Coating Characteristics as Related to Polymer

Structure and Physico-chemical Properties
The physical and chemical characteristics of the biopolymers greatly influence the
properties of resulting films and coatings. An edible film is essentially a dried and
extensively interacting polymer network of a three-dimensional structure. The film-
forming mechanisms of biopolymers include intermolecular forces such as covalent
bonds (e.g., disulfide bonds and cross-linking) and/or electrostatic, hydrophobic, or
ionic interactions. It is important to determine the relationship between the polymer
structural chemistry and the physico-chemical and mechanical properties of the film
or coating. Several techniques can be used to characterize the structure of the
biopolymers forming the coatings, such as scanning electron microscopy (SEM),
X-ray diffraction, differential scanning calorimetry (DSC), thermal mechanical
analysis (TMA), dynamic mechanical thermal analysis (DMTA), and Fourier
transform infrared spectroscopy (FTIR), etc. The main properties of a coating are
as follows: moisture and gas permeation (barrier properties), water vapor sorption
isotherms, optical attributes (light transmittance, color), thermoplastic characteris-
tics, crystallinity, mechanical properties of the film (strength, elasticity), and film
solubility in water and oil.
The optical properties impact the surface appearance of coated foods. Surface
film color is generally measured using a colorimeter and opacity is determined by
spectrophotometry. Film opacity is a critical property of a food coating; it can be
measured using the method proposed by Gontard et al. (1992), in which the
absorption spectrum is recorded in the visible range. The opacity is estimated as
the area below the absorption curve; this area has a very low value for transparent
films. The color, shininess, and transparency of edible films and coating layers vary
significantly depending on the chemical composition and the structure of the
polymer used. For example, the color stability of whey protein isolate (WPI)
coatings is affected by Maillard reactions (produced by milk proteins during
storage), causing yellowing.
The water sorption isotherms of the coatings can be measured on films formed
under standard conditions. They are useful to estimate stability at different ambient
conditions, mainly because the hydrocolloid film properties are highly dependent
636 N. Zaritzky
on relative humidity (RH). The presence of plasticizers, such as glycerol or sorbitol

increases the hygroscopic characteristics of the films.
Microstructure characterization of the coatings helps to understand the perfor-
mance of the different formulations. Light microscopy observation (stereomicro-
scopy) is used to analyze coating thickness and uniformity; SEM permits evaluation
of film homogeneity, the presence of different layers, pores, and cracks, and the
smoothness of the coated surface. FTIR is a useful technique for composite
film microstructure characterization since it allows evaluation of interactions and
compatibility between the components (Lacroix and Le Tien 2005), and permits
identification of the functional groups in the polymers after modifications.
The spectral regions 3,300–3,000 cm1 are generally used for elucidation of the
free or interacting hydroxyl groups; 2,950–2,850 cm1 for alkyl (-CH2-) chains; at
1,750–1,650 cm1 the bands represent carbonyl groups and amides bands. Pinotti
et al. (2007) applied FTIR spectrometry to determine the possible interactions
between chitosan and MC composite systems.
X-ray diffraction patterns of the films are characterized by sharp peaks asso-
ciated with the crystalline fraction and amorphous zones. The amorphous fraction
of the sample can be estimated by the area between the smooth curve, drawn to
follow the scattering hump, and the baseline joining the background within the low-
and high-angle points; the crystalline fraction can be estimated by the upper region
above the smooth curve. X-rays can also detect the recrystallization process during
storage, such as retrogradation in starch-based films and coatings (Garcı́a et al.
2000b; Mali et al. 2002, 2006)
Glass transition is a reversible change that takes place in the polymer between
the rubbery and glassy states; the temperature at which it occurs is the Tg. The
properties of amorphous or semicrystalline materials are seriously modified when
the temperature of the compounds rises above Tg. Generally, fully amorphous
bioplastic applications are limited by the fact that Tg of a polymer is highly affected
by the RH (especially for hydrophilic polymers). Below Tg the material is rigid, and
above Tg, it becomes rubbery and visco-elastic (Guilbert and Gontard 2005). Below
Tg, weak and non-cooperative local vibrations and rotation movements of the
molecules are possible; above Tg, strong and cooperative movements of whole
molecules and polymer segments can be observed. The knowledge of Tg is impor-
tant since it determines both the mechanical and barrier properties (Mali et al. 2005,
2006). DSC and DMTA techniques can be used to determine Tg. DMTA is
considered the most sensitive method for measuring the Tg of a material. In
DMTA, an oscillating strain is applied to a sample and the resulting stress deve-
loped in the sample is measured. The output signals are analyzed, and, using
established mathematical methods, the rheological parameters are also computed.
The ratio of viscous modulus (E00 ) to elastic modulus (E0 ) is the tangent of the phase
angle shift between the stress and strain vectors (E00 /E0 ¼ tan d); it measures the
damping ability of the material. The glass transition of a film is detected as a sudden
and marked decrease in the elastic modulus E´ and a peak in the tan delta curve as
temperature increases. Plasticization decreases the intermolecular forces between
polymer chains, and reduces Tg.
27.5 Barrier Properties
The quality of most food products deteriorates from moisture absorption, O2

invasion, flavor loss, undesirable odor absorption, and migration of packaging
components into the food (Kester and Fennema 1986; Debeaufort et al. 1998;
Miller et al. 1998; Krochta 2002). These mass transfer phenomena can occur
between the food and the environment, the food and packaging materials, or
among heterogeneous ingredients in the food product itself (Krochta 1997). The
diffusion of atmospheric O2 into foods causes oxidation of food ingredients; dried
foods can absorb moisture from the environment or from other wet zones in
the food, leading to the loss of crispness. Edible films and coatings may be used
to wrap these food products or located between the heterogeneous parts of food
products to prevent migration phenomena and to preserve quality (Guilbert et al.
1997; Krochta 2002).
To characterize the barrier properties of edible films and coatings, the transmis-
sion rates of specific migrants should be determined using stand-alone edible films
(McHugh and Krochta 1994). Most of the research on this subject reports the water
vapor, O2, CO2 and flavor permeabilities, and oil resistance of edible films. How-
ever, barrier properties of films and coatings determine their applicability to
improving the storage life of different products. Permeability depends not only on
the diffusion coefficient but also on the solubility of permeants in the system.
Barrier properties are strongly related to film structure since permeants are moved
through the amorphous zone. Film formulation, especially with the addition of
plasticizers, affects both the barrier and mechanical properties because these prop-
erties modify film structure, chain mobility, and diffusion coefficients of permeants.
27.5.1 Water Vapor Permeability
Water vapor permeability (WVP) determinations can be performed using com-

mercial instruments; in addition, many researchers applied the ASTM (1995)
method E96 with the modifications introduced by Gennadios et al. (1994). In
this method the film sample is sealed over a circular opening of known area in a
permeation cell that is stored at constant temperature in a dissecator. Anhydrous
calcium chloride or silica gel (0% RH) is placed inside the cell and a saturated
solution (constant RH) is used in the dissecator; thus water vapor transport is
determined from the weight gain of the permeation cell (Garcı́a et al. 1999,
2000a, 2004b, 2006; Romero-Bastida et al. 2005).
Synthetic moisture-barrier films such as low-density polyethylene (LDPE) show
low values of WVP ¼ 9.14 1013 g m1 s1 Pa1 in comparison with polysacchar-
ides and proteins, which are polymeric and hydrophilic by nature and poor moisture
barriers. The WVP permeabilities of proteins, such as gluten and zein plastified
with glycerol, range between 7.00 1010 and 8.90 1010 g m1 s1 Pa1.
638 N. Zaritzky
Lower WVP values were reported for several polysaccharides; sorbitol plastified yam
starch (3%) and corn starch (2%) films have WVP values of 1.50 1010 and
1.75 1010 g m1 s1 Pa1, respectively (Mali et al. 2002, 2004, 2006). Films of
MC (1%) and Chitosan (1%) have WVP values similar to that of cellophane ranging
between 7.55 1011 and 9.03 1011 g m1 s1 Pa1 (Pinotti et al. 2007).
It is important to remark that edible films made of proteins and polysaccharides
have low water-vapor barrier properties, with WVP values at one and two orders of
magnitude higher than that of LDPE; therefore, they can only be used as protective
barriers to limit moisture exchange in short-term applications. However, these
edible films can be of considerable value for numerous processes, as in the case
of modified atmosphere packaging of fresh, minimally processed or fermented
foods (fish, meat, fruits, vegetables, and cheeses).
Lipids or waxes for edible applications show better water-vapor barrier properties
than starch or protein-based materials do, but they still have significantly lower
water-vapor barrier properties than most synthetic plastic films; WVP values rang-
ing between 2.75 and 3.1 1012 g m1 s1 Pa1 were reported for beeswax and
carnauba wax (Shellhammer and Krochta 1997; Perez-Gago and Krochta 2005).
Lipid compounds, such as animal and vegetable fats (natural waxes and deriva-
tives, acetoglycerides, surfactants, etc.), could be proposed as components of edible
coatings given their excellent moisture-barrier properties; however, they can cause
textural and sensory problems due to oxidation and a waxy taste; further, their
nonpolymeric nature limits the ability to form cohesive films (Perez-Gago and
Krochta 2005). The addition of lipids to polysaccharide-and protein-based films
decreases WVP due to hydrophobicity; this is an important property in the case of
fresh fruits and vegetables because composite coatings retard moisture loss and
shriveling in fresh products.
27.5.2 Gas Permeabilities
CO2 and O2 barrier properties are important in understanding the quality and
physiological aspects of coated food during storage and determine food coatings
applications. Gas permeability is commonly measured on isolated films. O2 perme-
ability can be measured using available commercial equipment; moreover, CO2 and
O2 permeability of films can be assessed by the accumulation method in a specially
designed cell (Garcı́a et al. 1999, 2000b; Bifani et al. 2007). This quasistatic method
was based on measuring the amount of gas diffusing through a film, quantified by
gas chromatography.
Garcı́a et al. (1999) measured gaseous permeabilities to O2 (PO2) and CO2
(PCO2) for different starch-based films; values of PO2 ¼ 5.69 109 cm3
m1s1 Pa1 and PCO2 ¼ 0.248 109 cm3 m1 s1 Pa1 were reported for
corn starch films plasticized with glycerol. All tested starch-based films, plasticized
with either glycerol or sorbitol, showed higher gas permeability to CO2 than O2,
with a ratio of permeability (PCO2/PO2) ranging between 8 and 10 (selectivity
coefficient). These results can be attributed to a higher solubility of CO2 than O2 in

the coatings (McHugh and Krochta 1994). Synthetic materials such as LDPE have
lower selectivity coefficients between CO2 and O2 (around 4) compared to an
average ratio of 9 in the case of starch-based films (Cuq et al. 1998; Garcı́a et al.
1998ab; Park 1999). Cuq et al. (1995) compared the CO2 to O2 permeability ratio
for several synthetic and edible films and reported that edible films show higher
selectivity than synthetic ones, with ratios in the range of 8–30. Very high gas
selectivity is particularly interesting for modified atmosphere packaging of cheeses
(to control proliferation of microflora), and for fresh fruits and vegetables (to
control respiration rates).
Gas permeability is strongly related to crystallinity of polymeric chains, where
in general, the higher the degree of crystallinity, the lower the permeability. As for
synthetic materials, gas and vapor permeabilities depend on several factors, such as
the ratio between crystalline and amorphous zones, polymeric chain mobility, and
specific interactions between the functional groups in the polymers and gases in the
amorphous zones. Permeability increases with a decreasing crystalline-amorphous
ratio because permeation occurs through the amorphous zones of the film.
Starch and proteins have gas-barrier properties in dry conditions, especially
against O2. For example, the O2 permeability of a wheat gluten film was 800
times lower than that of LDPE. However, O2 permeability is very sensitive to RH
(Guilbert et al. 1997). At higher RH conditions, O2 permeability increases signifi-
cantly; thus, it is important to maintain low RH in the environment to maximize the
effectiveness of edible films as gas barriers. Values of the selectivity coefficient
between CO2 and O2 ranging between 4 and 35 were reported by Gontard et al.
(1996); this selectivity coefficient is very sensitive to moisture and temperature for
gluten films, whereas for synthetic polymers it remains relatively constant ranging
between 4 and 6 (Guilbert and Gontard 2005).
27.6 Composite Film Formation
Composite films and coatings can be formulated to combine the advantages of each
component. Biopolymer composites can modify film properties and create film
structures for specific applications (Wu et al. 2002). A composite film that com-
bines lipids and hydrocolloids can be produced as either a bi-layer or a stable
emulsion. In bi-layer composite films, the lipid forms a second layer over the
polysaccharide or protein layer. In emulsion composite films, the lipid is dispersed
and entrapped in the supporting matrix of protein or polysaccharide (Perez-Gago
and Krochta 2005). The emulsifying character of proteins makes them appropriate
for this technique; however, in the case of polysaccharides, the addition of an
emulsifier is required to improve emulsion stability. Emulsion preparation para-
meters, such as stirring velocity and emulsifier addition are important factors in
obtaining homogenous films and coatings, since they determine emulsion stability.
640 N. Zaritzky
27.7 Examples of Coating Applications
In general, an ideal coating should be safe, unperceived by the consumer, have no

off-flavor and taste, be a desirable moisture and gas barrier, and have a certain
mechanical strength. Many factors determine the success of edible coatings in
improving the quality and extending the shelf-life of foods. Factors such as chemi-
cal composition, structure, methods used to form the films or coatings, storage
conditions, and properties of the food itself (maturity stage in the case of fruits and
vegetables, moisture and lipid content, etc.) are all important. When applying
edible films and coatings, these factors have to be carefully considered (Zhao and
McDaniel 2005).
Quality maintenance and product enhancement are significant functions of
edible films and coatings (Krochta 1997). They can retard surface dehydration,
moisture absorption, oxidation of ingredients, aroma loss, frying oil absorption,
ripening/aging, and microbial deterioration of food products. In addition to the
physical and chemical quality enhancements, edible films and coatings contribute
to visual quality, surface smoothness, flavor carriage, edible color print, and other
marketing-related quality factors. Edible films and coatings may be used to preserve
the quality of several food commodities. The O2-barrier property of films and
coating layers can prevent oxidation of lipid ingredients, colorants, and flavors of
food products such as nuts, confectionary, and fried foods (Baldwin et al. 1997).
This property is also used to retard the respiration rate of fruits and vegetables
(Baldwin et al. 1995). Many climacteric fruits and vegetables can be coated with
edible film-forming materials to slow down their respiration rate (Park 1999;
Amarante and Banks 2001).
High-fat meat and fish products, such as sausages and fillets, can be protected
from oxidation using O2 barrier edible coatings. Moisture loss is the most critical
quality degradation factor in fresh produce (Guilbert et al. 1997). Moisture-barrier
properties of edible films and coatings can protect fresh fruits and vegetables from
dehydration. The moisture-barrier property can also be utilized to prevent moisture
migration between heterogeneous food product ingredients, for example between
raisins and breakfast cereals (Kester and Fennema 1986), pie fillings, and crusts,
etc. Different materials used in edible-coating formulation have been reported in
literature (Kester and Fennema 1986; Gennadios and Weller 1990; Krochta et al.
1994; Krochta and De Moulder Johnston 1997; Debeaufort et al. 1998; Zhao and
McDaniel 2005; Lin and Zhao 2007).
Some selected examples are listed below:
l Wheat gluten, whey proteins, corn zein, waxes (beeswax, carnauba, candelilla),
cellulose derivatives and pectins for fruits, grains, and vegetables as O2 and
moisture barriers
l Waxes or fatty acids on fruits and vegetables to delay spoilage and to reduce
water loss and, on cheeses, to prevent mold growth
l Gelatin, seaweed, and pectinate to eliminate dripping, and as moisture barriers

on cheese, ice cream, and yogurt
l Corn zein, milk and whey proteins, waxes, and methyl cellulose in confections
l Cellulose derivatives on fried foods to reduce oil absorption
l Gelatin, carrageenan and alginate, whey protein, collagen, casein, and cellulose
derivatives on poultry, beef, fish and seafood
l Acetylated monoglyceride and whey protein on frozen salmon to reduce mois-
ture loss and lipid oxidation
l Casein, xanthan gum on peeled carrots to reduce dehydration and white blush
formation
l Chitosan on strawberries to delay spoilage, and on tomatoes, to extend shelf-life
l Cellulose, chitosan, and sodium caseinate on bell peppers to reduce O2 and CO2
permeability
l Starch in coated prunes, strawberries, and nuts to prolong storage life
l Dextrin on apples to delay oxidative browning
l Corn zein on tomatoes to delay color changes and loss of firmness
l Wheat gluten, cellulose, soy protein, and whey protein applied individually on
egg shells to improve shell strength and to reduce microbial contamination
l Starch, corn zein, whey protein, and acetylated monoglyceride on nuts to delay
rancidity
l Soy protein on apples to retard changes in firmness and color
l Chitosan-based coatings to reduce drip loss on frozen-thawed raspberries
l Alginate on fresh meat, poultry, and precooked ground pork to reduce shrinkage,
oxidative rancidity, moisture migration, and oil absorption
l Mixtures of stearic-palmitic acids and hydroxypropyl cellulose, and MC and
palmitic acid coatings, as moisture barriers for heterogeneous foods (puree,
cake, ice cream cones)
27.8 Incorporating Functional Ingredients into Edible

Films and Coatings
A unique feature of edible films and coatings is their capacity to carry many
functional ingredients, including antioxidants, antimicrobial agents, flavorings,
and colorants. Integration of these ingredients can enhance food stability, quality,
functionality, and safety. For instance, food microbial stability can be improved
using edible-active layers, which act as surface-retention agents to limit the diffu-
sion of food additives into the food core; these layers can be used with simultaneous
treatments such as controlled atmosphere and refrigeration. In another example,
high concentration of a chemical preservative incorporated into an edible coating
may allow a decrease in the total amount of the preservative in the food, attaining
similar performance.
642 N. Zaritzky
27.8.1 Antioxidant Edible Coatings
Antioxidants, such as tocopherol and butylated hydroxytoluene (BHT) were

integrated into edible coatings to inhibit lipid oxidation in precooked beef patties
and fish muscle, respectively (Zhao and McDaniel 2005). The antioxidant effects of
ascorbic acid in whey protein film coatings applied on roasted peanuts for the
control of lipid oxidation were also studied by Min and Krochta. (2007). By
measuring the peroxide values and the thiobarbituric acid reactive substances, it
was shown that the addition of ascorbic acid to the coating retarded lipid oxidation
in peanuts.
27.8.2 Antimicrobial Edible Films
The concept of active packaging was introduced with antimicrobial films and
coatings. Edible films can serve as carriers of antimicrobials to extend product
shelf-life and to reduce the risk of pathogen growth on food surfaces. Edible films
have been developed that can reduce, inhibit, or delay the growth of microorgan-
isms on the surface of coated foods. Excellent reviews on antimicrobial edible films
have been published by different authors (Han 2000, 2002, 2003a, b; Franssen and
Krochta 2003; Cagri et al. 2004).
Chitosan: A polysaccharide obtained by deacetylation of chitin, chitosan origi-
nates from crustacean exoskeleton and fungal cell walls. It has been widely used in
antimicrobial films and coatings due to its property of inhibiting the growth of many
pathogenic bacteria and fungi. In some fungi, chitosan can alter membrane func-
tions by interacting with the strongly electronegative microbial surface, leading to
changes in permeability, metabolic disturbances, and eventually death.
Natural preservatives: There has been a tendency in the last several years to use
natural preservatives in edible-coating formulations. Some of the most commonly
used preservatives and antimicrobials are:
l Organic acids (acetic, benzoic, lactic, propionic, sorbic)
l Parabens (methyl paraben)
l Fatty acid esters (glyceryl monolaurate)
l Polypeptides (lysozyme, peroxidase, lactoferrin, nisin)
l Nitrites (potassium nitrite, sodium nitrite)
l Sulfites (potassium sulfite, sodium sulfite)
l Natural preservatives (essential oils, spices, extracts such as cinnamon, sage,
allicin, and liquid smoke)
l Other preservatives (atamycin, EDTA)
These agents can be used in protein-, polysaccharide-, and lipid-based edible
coatings and films.
Benzoic acid and sodium benzoate: These two are among the first chemical
preservatives permitted in foods by the Food and Drug Administration. Benzoic
acid (pKa ¼ 4.20) and sodium benzoate are generally regarded as safe preserva-
tives at levels up to 0.1%. Sodium benzoate and benzoic acid are inhibitory to mold,
yeast, and pathogenic and psychrotrophic spoilage bacteria. Sodium benzoate is one
of the most commonly used antimicrobials in edible films because it is soluble in
most film solutions and remains active after film preparation. Organic acids are
more effective in the undissociated form and, therefore, the antimicrobial activity is
related to pH. At pH ¼ 4.0, 60% of benzoic acid is undissociated (Cagri et al.
2004).This antimicrobial can be incorporated into MC, collagen, and chitosan films,
all of which have a relatively low pH. Edible films containing benzoic acid and its
sodium salt are adequate for acidic foods like cheeses and fermented meat products.
Sorbic acid: This is a straight chain unsaturated monocarboxylic acid; the
carboxyl group reacts to form calcium, sodium, or potassium salts. Potassium
sorbate is highly soluble in water (58.2% at 20 C). Edible films containing potas-
sium sorbate are most effective at pH < 6.0. Sorbic acid salts were widely used as
antimicrobial agents in carbohydrate- and protein-based edible films formulated
with MC, WPI, and chitosan. Films containing sorbates have been tested against
spoilage bacteria, pathogenic bacteria, yeasts, and molds in laboratory media.
Propionic acid: This acid is commonly used as a food preservative because of its
wide spectrum of activity. Propionic acid, which is a monocarboxylic acid, is
produced by Propionibacterium freudenreichii subsp. shermanii. Antimicrobial
activity of propionates is also pH dependent. Although this acid is primarily
active against molds, some yeasts and bacteria are also inhibited. The amount of
propionate used in foods is generally less than 0.4% (Cagri et al. 2004).
Parabens: Produced by esterification of the carboxyl group of benzoic acid,
most parabens are active at pH 3.0–8.0. The methyl, propyl, and heptyl parabens
can be used as food preservatives in most countries. Parabens can also be used
effectively in a wide range of foods; they are generally more active against molds
and yeasts than bacteria. For complete inhibition of bacteria and fungi, concentra-
tions of esters of p-hydroxybenzoic acid ranging between 0.033 and 1.0 mg/L are
necessary (Cagri et al. 2004).
Free fatty acids: Low concentrations of long-chain fatty acids are inhibitory to
microorganisms, especially gram-positive bacteria and yeasts. Saturated fatty acids
with chain lengths of C12 to C16 and C10 to C12 have the most antimicrobial
activity against bacteria and yeasts, respectively (Cagri et al. 2004). Fatty acids are
more active at low pH (5.0), and both fatty acids and monoglycerides are inhibitory
to many bacterial species. Monoesters of glycerols and esters of sucrose are more
antimicrobial than their corresponding free acids.
Acetic acid and sodium diacetate: Both are active against various spoilage and
pathogenic bacteria and have been used in many foods. Like other organic acids,
acetic acid can be used to acidify edible coatings containing chitosan, alginate,
collagen, and WPI.
Lactic acid: This acid is used for improving and controlling the quality and
microbial stability of foods. Lactic acid sprays (1–3% solutions) have been widely
644 N. Zaritzky
used to sanitize meat carcass surfaces, with gram-negative psychrotrophs generally

being more sensitive to treatment than gram-positive organisms. Lactic acid can be
used as an acidulant in chitosan and collagen films.
Nisin: The first bacteriocin used in the food industry, nisin is one of the most
investigated bacteriocins in antimicrobial edible films. It is produced by fermenta-
tion using Lactococcus lactis subsp. Lactis. It was recognized as a safe biological
food preservative by a joint FAO/World Health Organization commission on
food additives in 1968, and accepted by the Food and Drug Administration in
1988. Nisin is a polycyclic peptide antibacterial with 34 amino acids residues.
It is obtained commercially from natural substrates, including milk, and is not
chemically synthesized. It inhibits gram-positive spoilage and pathogenic bacteria
(Listeria monocytogenes, Staphylococcus aureus, Clostridium botulinum). Incor-
poration of nisin into packaging materials inhibits the growth of gram-positive
bacteria but does not inhibit gram-negative bacteria such as E. coli.
Nisin can be incorporated into the film suspension or applied directly to the film
surface after casting. Various nisin-containing protein-based films (e.g., whey
protein, corn zein, wheat protein, soy protein) have been assessed for antimicrobial
activity against gram-positive bacteria such as L. monocytogenes and lactic acid
bacteria (Cagri et al. 2004).
Pediocins: Produced by Pediococcus acidilactici, this group of bacteriocins can
be used in edible films and coatings because of their wide spectrum of antimicrobial
activity and effectiveness over a wide range of pH values and temperatures.
Antimicrobial activity of pediocin is retained at 100 C, decreasing at 121 C, and
is most evident at pH values between 4 and 7, with substantial losses at pH 3 and 9.
Pediocin remains active following treatment with lipase, phospholipase C,
lysozyme, DNase, or RNase, but its activity is destroyed by protease, papain, and
a-chymotrypsin (Cagri et al. 2004).
Lysozyme: This enzyme is found in egg whites, human tears, and other secre-
tions. It is responsible for breaking down the polysaccharide walls of many kinds of
bacteria and thus it provides some protection against infection. The enzyme is
formed by 129 amino acids cross-linked by four disulfide bonds. Dried egg white,
the commercial source for lysozyme, contains about 3.5% lysozyme. It is heat
stable (100 C) at pH 5.3 but inactivated at lower temperatures when pH is
increased. Lysozyme is most active against gram-positive bacteria (Cagri et al.
2004).
Spices, herbs, and essential oils: Essential oils are responsible for the odor,
aroma, and flavor of spices and herbs. These compounds can be added to edible
films to modify flavor, aroma, and odor and to introduce antimicrobial properties.
Films containing these ethanol-soluble compounds show activity against both
gram-negative and gram-positive bacteria. Essential oils (cardamom, cinnamon,
cloves, coriander, garlic, nutmeg, oregano, parsley, rosemary, sage, thyme, etc.) are
inhibitory to various spoilage or pathogenic bacteria, molds, and yeasts. However,
the use of these oils as food additives is limited due to their strong flavor.
Lactoferrin (lactotransferrin): This iron-binding glycoprotein is present in
bovine milk. It inhibits the growth of some bacteria; however, other bacteria with
low iron requirements, such as lactic acid bacteria, are not inhibited by lactoferrin.
Pseudomonas fluorescens, Enterococcus faecalis, and Bifidobacterium bifidum
strains are highly resistant to this peptide. The mode of action of lactoferrin has
not been fully elucidated, but presumably it alters membrane permeability because
of its cationic nature.
Liquid smoke: Used in processed meats, sausages, and cheeses, liquid smoke
contains phenols and acetic acid, which are bactericidal at relatively low concentra-
tions. It can inactivate common food-borne pathogens, including E. coli, Salmonella,
Staphylococcus aureus, and L. monocytogenes. It is generally very acidic and has
antimicrobial, antioxidant, and flavor properties, making it a potentially attractive
edible coating additive. However, incorporation of liquid smoke has only been
studied for edible collagen casings.
Cha and Chinnan (2004) reviewed important aspects of biopolymer-based anti-
microbial packaging. The incorporation of several antimicrobial agents (EDTA)
and grapefruit seed extract into films, made of Na-alginate and K-carrageenan,
produced a strong inhibition against Micrococcus luteus, L. innocua, Salmonella
enteritidis, E. coli, and S. aureus; sodium alginate-based films, with the addition of
nisin, lysozyme, and EDTA, showed the strongest inhibition against the same
bacterial strains. Chitosan films containing nisin showed a high antimicrobial
activity against M. luteus. Whey protein film coating was used as a vehicle for
combining antimicrobials (grape seed extract, nisin, malic acid, and EDTA)
(Cha and Chinnan 2004).
The primary advantage of antimicrobial edible films is that the inhibitory agents
in these films can be specifically targeted to postprocessing contaminants on the
food surface. It is important to predict and control the preservative release (Torres
and Karel 1985; Torres et al. 1985; Redl et al. 1996).The diffusion rate of the
antimicrobial into the product is partially controlled by agents incorporated into the
film. The edible film matrix entraps the antimicrobial and decreases its diffusion
during storage; thus, lower levels of preservative addition would be needed in an
edible coating to achieve a targeted shelf-life, as compared to spraying the antimi-
crobial on the surface.
Controlling the antimicrobial release from edible films is very important; there-
fore, measurement of the diffusion coefficients of the antimicrobials in edible films
should be carried out under controlled conditions (Vojdani and Torres 1990; Brody
et al. 2001; Min et al. 2008). While antimicrobial coatings may extend the shelf-life
and protect food products, it must be emphasized that such advantages are never
substitutes for proper handling, storage, and good manufacturing practices.
27.8.3 Edible Coatings as Carriers of Nutraceutical Ingredients
There is an increasing consumer interest in the health-enhancing roles of specific

foods and physiologically active food components (nutraceuticals or functional
foods). Edible coatings could provide an excellent way to enhance the nutritional
646 N. Zaritzky
value of foods by carrying nutrients or nutraceuticals that are lacking, or present in

low quantities (Mei and Zhao 2003; Park and Zhao 2004; Park 2004). Zhao and
Mc Daniel (2005) integrated high concentrations of minerals and vitamins into
edible films and coatings.
Mei et al. (2002) developed xanthan gum coatings containing a high concentra-
tion of calcium and vitamin E, and applied such coatings to fresh baby carrots. Han
et al. (2004) applied chitosan coatings with high concentrations of calcium or
vitamin E on fresh and frozen strawberries and red raspberries.
27.9 Edible Coatings to Improve Quality and Extend

Shelf Life of Foods - Case Studies
27.9.1 Edible Coatings Acting as Oil Barriers in Fried Products
One of the applications of coatings is to reduce the OU of products during deep-fat

frying (Balasubramaniam et al. 1997). The excess of fat in the diet has been linked
to coronary heart disease; thus, coatings applied to food before frying can help
reduce health problems associated with fat consumption. Frying occurs during the
immersion of the product in oil at temperatures higher than the boiling temperature
of water (150–200 C). Deep-fat frying is a complex process involving simultaneous
heat and mass transfer (Bertolini Suarez et al. 2008)
Various hydrocolloids, especially cellulose derivatives, can be used as edible
coatings to decrease oil content during frying. Williams and Mittal (1999) found
that MC films showed the best barrier properties, reducing fat uptake more than
hydroxypropylcellulose and gellan gum films. Cellulose derivatives, including MC
and HPMC exhibit thermal gelation. When suspensions are heated a gel is formed
that reverts back below the gelation temperature, and the original suspension
viscosity is recovered. MC forms a gel at high temperatures; hydrophobic polymer
chain interactions are involved in the thermal gelation process with a predominance
of intermolecular hydrogen bonding over intramolecular hydrogen bonding within
the cellulose ether. Extensive research on edible coatings from cellulose derivatives
was carried out in our laboratories (Garcı́a et al. 2002, 2004a; Bertolini Suarez et al.
2008). The objectives of the experiments were to analyze the coating performance
in reducing OU in French fries and fried pastry.
Aqueous suspensions of 1% Methocel A4M (based on MC), 2% K100LV, and
2% E15LV (based on MC and HPMC), with and without sorbitol (0.25–1%), were
used for coating applications. The rheological behavior of the suspensions and the
viscoelastic moduli were measured with an oscillatory rheometer. MC suspensions
were selected for coating application because they showed the highest elastic (G0 )
and complex (G*) moduli.
Model dough discs (60 mm dia. and 7 mm thick) were prepared with refined
wheat flour and distilled water. Samples were dipped in coating suspensions for 30 s
and drained. Coated and uncoated samples were fried in sunflower oil under
controlled temperature conditions (160 0.5 C, previously determined by a sen-
sory panel). Surface color (lightness and chromaticity parameters), firmness (instru-
mental texture analysis), and OU of the control and coated fried samples were
measured at different frying times.
During frying of both the coated and uncoated samples a moving boundary was
produced within the product, separating the dehydrated zone from the wet core
(Fig. 27.1). A protective layer (approximately 15 mm thick) was formed on the
surface of the coated samples during the initial stages of frying, due to MC thermal
gelation above 60 C; the coating was becoming dehydrated and remained attached
to the surface of the product. SEM observations showed the integrity of the MC
layer and good adhesion of this coating to the fried product (Fig. 27.1); the addition
of a plasticizer (sorbitol) in the coating formulation was necessary to achieve
coating integrity and uniformity.
160-180oC
Food core OIL

(wet zone)
Heat transfer
Water vapor
transfer
Water vapor
bubbles
COATING
Moving
boundary
FOOD
DEHYDRATED
ZONE (DZ)
Fig. 27.1 Schematic description of frying process in a coated product, and scanning electron
microscopy (SEM) micrograph of the methylcellulose coating applied on a fried dough sample
648 N. Zaritzky
OU was determined by combining a preliminary batch extraction process with

a semicontinuous extraction in a Soxhlet; the extracted lipid phases were analyzed
by HPLC (Garcı́a et al. 2002, 2004a). MC coatings (1% A4M and 0.75% sorbitol)
were effective in decreasing oil content without affecting the water content
of the samples (Bertolini Suarez et al. 2008); the OU in coated samples was 30%
lower than in the uncoated ones at the final frying time (Fig. 27.2).
Concerning the quality attributes of the fried dough, differences between the
instrumental color parameters and firmness values of coated and uncoated samples
were not significant, during the frying process (t < 720 s); besides, the panelists
could not distinguish the differences between the control and coated samples. These
results demonstrate the effectiveness of the MC coating to reduce OU without
affecting the sensory attributes of the samples (Garcı́a et al. 2002).
The frying process was mathematically modeled (Bertolini Suarez et al. 2008)
and experimentally validated; temperature profiles and water-content distributions
were predicted by the model. Nonlinear-coupled energy and mass transfer partial
differential equations, under unsteady state conditions, considering a moving
boundary, were numerically solved considering both thermal and physical proper-
ties change with temperature and moisture content. The model simulated appropri-
ately the experimental data of temperature and water content during the different
frying stages.
Microstructural changes are produced during the frying process; the heat and
mass transfer model predicted the position of the vaporization front and the
thickness of the dehydrated zone (dZ) as a function of frying time (Fig. 27.2).
In addition, water content correlated linearly with the vaporization front position
corresponding to the thickness of the humidity core (Bertolini Suarez et al. 2008).
During frying the vapor leaves voids in the product and allows for the oil to enter
later; OU is mainly produced when the product is removed from the frying medium.
0.1 0.006
0.09
0.005
0.08
OU (dry basis)
0.07
0.004
0.06
dZ (m)
0.05 0.003
0.04
OU uncoated 0.002
0.03
0.02 OU coated 0.001
0.01 Dehydrated zone (dz)
0 0
0 200 400 600 800 1000 1200
t (s)
Fig. 27.2 Effect of methylcellulose coating on oil uptake in fried dough. The presence of coating
decreases oil uptake. Oil uptake curve can be closely related to increase of dehydrated zone
thickness (dZ)
OU is affected by the microstructure of the dehydrated zone and is dependent on the

balance between the oil retained on the surface and the oil drained after retrieval of
the product from the oil bath; the oil retained by the surface penetrates into the pores
left by the water evaporation. The amount of oil retained at the sample surface is
limited, due to the oil surface wetting, the property related to the interfacial tension
that governs these phenomena. Figure 27.2 shows similar shapes for the OU curves
(experimental data) and the thickness of the dehydrated zone (predicted values).
From these results, OU was linearly correlated with the thickness of the dehydrated
zone during the initial frying period (Bertolini-Suarez et al. 2008).
Other results obtained in our laboratory on using edible coatings to reduce OU
during frying (4 min/180 C) showed a reduction of 40.6% in oil content of French
fries (0.7 0.7 5 cm) using MC (1%) and sorbitol, without causing detrimental
effects on the quality attributes (surface color and texture).
27.9.2 Starch-Based Edible Coatings to Prolong Storage

Life of Refrigerated Highly Perishable Fruits
The strategies explored to extend postharvest life of fruits should consider several
challenges, such as extending maturation and senescence periods, reducing dehy-
dration, and reducing onset and rate of microbial growth. Edible films and coatings
could offer solutions to all such challenges simultaneously, the use of which has
been reported by several researchers for different vegetables and fruits (Drake and
Nelson 1990; El Gaouth et al. 1991; Park 1999, 2003, Park et al. 2005). The use of
coatings with selective permeabilities on fruits has also been shown to have an
influence on fruit physiology in retarding ripening and postharvest metabolism,
thus, extending the fruit’s storage life (Baldwin 1994).
Edible coatings can be made from food materials regarded as GRAS, such as
proteins, lipids, cellulose derivatives, starch, and other polysaccharides. Composite
films can be formulated to combine the advantages of both lipid and hydrocolloid
components (Guilbert 1986). In the case of fruit and vegetable conservation, the
lipid component in coating formulation can serve as a good barrier against water
vapor, while the hydrocolloid component can provide a selective barrier against O2
and CO2 and the necessary supporting matrix. The functional, organoleptic, nutri-
tional, and mechanical properties of an edible film can be modified with the
addition of various chemicals in minor amounts, such as plasticizers, antimicrobial
agents, and lipids.
Strawberries (Fragaria ananassa), a typical soft fruit variety, have high phy-
siological postharvest activity that transforms them into a highly perishable fruit
with a low period of commercialization, which is mainly limited by fungal infec-
tion. As a consequence, strawberries have short ripening and senescence periods,
which makes coating this fruit a good alternative for extending storage life. In
this case, the effect of adding plasticizer and lipid and antimicrobial agents to
650 N. Zaritzky
starch-based coatings, on their performance, applied to strawberries is discussed

(Garcı́a et al. 1998a, b, 2001).
Strawberries (Fragaria ananassa, cv Selva) at the commercial ripening stage
(75% red color), grown on a local farm in greenhouses, were harvested and
immediately treated. The strawberries were dipped in chlorinated water (0.25 g
Cl2/L), dried with air, dipped in coating suspensions, and dried again with air
(1.2 m/s, 20 C and 84.8% RH). Uncoated fruit (control) was treated similarly,
replacing immersion in starch suspensions with distilled water. Samples were
stored in a cold chamber at 0 C and 84.8% RH.
Two different starches were used for coating formulations: commercial corn starch
with 25% amylose and a high amylose corn starch; and amylomaize (Amylomaize
VII, Amaizo, USA) with 65% amylose. Starches were cold-gelatinized using
sodium hydroxide; suspensions were neutralized with H3P04. The gelatinization
method was selected to comply with the dipping procedure for coating application,
and the heat-sensitive characteristic of strawberries. Glycerol or sorbitol was added
as plasticizer at a concentration of 20 g/L. Additionally, 2 g/L sunflower oil were
added and emulsions were homogenized at 7,800 rpm. Potassium sorbate (2 g/L)
and citric acid (to reach pH 4) were also included in the formulations as antimi-
crobial agents.
Rheological behavior of coating suspensions was analyzed. The suspensions
showed pseudoplastic behavior (n < 1) and the Power law rheological model
satisfactorily fit experimental data (r2 > 0.96). Plasticizer and lipid addition to
corn and amylomaize suspensions decreased the flow behavior index and increased
the consistency index. Apparent viscosity of corn starch suspensions also decreased
with plasticizer and lipid addition.
Microstructure characterization of the coatings contributed to understanding the
performance of the different formulations. This characterization included light
microscopy observation (stereomicroscopy) and SEM. The stereomicroscope was
used to determine coating thickness (40–50 mm) and to check coating uniformity;
coated samples were stained with iodine solution to improve coating visualization.
Polarized light microscopy observations revealed that starch was totally gelatinized
in the filmogenic suspensions (Garcı́a et al. 1998a, b). SEM observations showed
that plasticizer addition was necessary for film integrity, to avoid pores and cracks.
Plasticized films containing lipid (Fig. 27.3b) exhibited smooth surfaces and a
compact structure, thereby indicating the homogeneous dispersion of lipids in the
film matrix (Garcı́a et al. 1999, 2000a, b).
Modifications of the crystalline structure of films stored at 20 C and 63.8% RH
were evaluated by DSC and x-ray diffraction, using Ka Cu radiation (l ¼ 1.5418 Å).
X-rays detected the starch recrystallization process during storage. In the case of
starch-based films, peak width slightly decreased and peak intensities increased
with storage time, showing a tendency toward increase in crystallite size, which can
be attributed to a slow recrystallization process (Garcı́a et al. 2000b, 2001). In the
case of starch coatings obtained by alkaline treatment, X-ray diffraction patterns
showed that both corn and amylomaize films attained a common crystalline
structure of higher stability with storage time, exhibiting an A-type starch pattern
a b c
500 DSC
Scanning Electron Microscopy
X-ray diffraction SEM
400
Initial
Cuentas / seg
5 days
dQ/dt (mJ/seg)
300
7 days
200 15 days
100 36 days
0 1 mJ/seg
0 10 20 30 40
2Θ
20 40 60 80 100
Temperature (°C)
d e
40 10
35
L (Hunter units)
8
30
Weight loss (%)
25 6
20
4
15
10 2
5
0
0 1 7 13 19
0 10 20 30
Time (days)
Time (days)
Fig. 27.3 Application of starch-based coatings to prolong storage life of refrigerated strawberries.
(a, b, c) Coating characterization; (a) X-Ray diffraction pattern of amylomaize coating containing
20 g/L glycerol and 2 g/L sunflower oil; (b) SEM micrograph of corn starch-based films with
20 g/L sorbitol and 2 g/L sunflower oil. Magnification: 100 mm between marks; (c) DSC thermo-
grams of amylomaize film stored at 20 C and 63.8% relative humidity. (d, e) Effect of coating
formulation on quality parameters of refrigerated strawberries during storage at 0 C and 84% RH;
(d) weight losses, (e) lightness differences. Control samples, uncoated fruits, (○); composition of
coating formulation: (□) amylomaize without plasticizer; (D) amylomaize with 20 g/L sorbitol;
and (e) amylomaize with 20 g/L sorbitol and 2 g/L sunflower oil. Bars indicate standard errors
(Fig. 27.3a). Films without plasticizer showed higher crystallinity (higher peaks)
than films containing plasticizer, which showed a larger amorphous zone and lower
peaks.
Amylomaize films containing neither plasticizer nor lipid showed a higher
crystallinity (higher numbers of peaks) than films containing plasticizer and lipid.
These results agree with the DSC studies that showed endothermic transitions with
lower peak temperatures and lower DH in films with plasticizer and lipid than in
control films. Presence of lipid and plasticizer in film formulations did not alter the
x-ray pattern of the most stable structure developed in stored control films; peaks
maintained their initial positions.
Crystalline evolution of the film matrix during storage was also evaluated by
DSC; this technique also allowed determining the Tg. Starch coatings obtained by
alkaline treatment, showed an endothermic transition with a peak temperature
around 50 C during storage; this peak became narrower and its temperature and
the corresponding enthalpy (DH) increased with storage time (Fig. 27.3c).
652 N. Zaritzky
Table 27.1 Barrier properties of starch-based coatings: water vapor permeability (WVP) and
gaseous permeabilities
Starch-based Additives WVP 1010 CO2 permeabilitya O2 permeabilitya 109
coating (g m s Pa ) 109 (g m1 s1 Pa1) (g m1 s1 Pa1)
1 1 1
Corn Without 3.68 2.24b 29.21 13.89 1.59 0.30

additives
G 2.57 1.04 5.69 0.97 0.46 0.05
G þ SO 1.92 0.47 5.87 0.58 ND
S 1.75 0.14 4.19 0.81 0.24 0.03
S þ SO 1.22 0.11 4.72 0.65 ND
Amylo-maize Without 2.62 1.39 28.05 7.37 2.64 0.25
additives
G 2.14 0.75 3.85 1.28 0.321 0.02
G þ SO 1.76 0.37 4.39 0.90 ND
S 1.21 0.15 2.96 0.46 0.23 0.03
S þ SO 0.97 0.08 3.43 0.21 ND
G glycerol (20 g/L), S sorbitol (20 g/L), SO sunflower oil (2 g/L), ND not determined
a
At standard pressure and temperature conditions
b
Value standard deviation
Barrier properties, which include water vapor (WVP), and O2 and CO2 perme-
abilities of starch-based coating formulations, were determined on films obtained
by casting (Garcı́a et al. 2000b). The addition of lipids to starch-based films
decreased WVP due to its hydrophobicity (Table 27.1). This is an important
property in the case of fresh fruits and vegetables because composite coatings
retard moisture loss and subsequent shriveling of fresh products. Table 27.1
shows that O2 permeabilities were much lower than those of CO2, indicating a
selective action of these films on gas permeabilities. Amylomaize films with higher
amylose content showed higher crystallinity and therefore, lower permeabilities
than corn starch films. The development of edible coatings with selective gas
permeabilities is a good example of active packaging’s role in controlling respira-
tory exchange and improving the conservation of fresh vegetables. Moreover, the
addition of lipid, which is necessary to reduce WVP, maintained the selective gas
permeability property since CO2 and O2 permeabilities did not differ from those of
plasticized starch-based films.
The effect of coatings on the quality attributes of refrigerated strawberries was
analyzed and the following results were obtained:
Weight loss: The same fruits were weighed at the beginning of the experiment
and during storage; weight loss was expressed as percentage loss of initial weight.
All the tested coatings showed a beneficial effect on weight loss (Fig. 27.3d).
Weight loss in fruits coated with starch formulations, without plasticizer, was
similar to those of control fruits due to pores and cracks. Coatings with sorbitol
led to significantly lower fruit weight loss than glycerol, regardless of starch type.
However, weight loss was unacceptable after 3 weeks of storage for coated fruits
without lipid. The addition of sunflower oil was necessary to reduce weight loss
significantly (P < 0.05) and to increase storage life, provided the microbial counts
were below the established limit (106 CFU/g fruit), even at 28 days of storage. The
maximum weight loss reduction at 28 days of storage was 63.2%, obtained with
coating formulations that included sunflower oil (Garcı́a et al. 2001).
Texture changes: Important modifications in texture can occur in fruits and
vegetables during storage, determining the postharvest storage life of the product.
Fruit softening is generally attributed to degradation of the cell wall components,
mainly pectins, due to the activity of specific enzymes such as polygalacturonase.
The rate and extension of firmness loss during ripening of soft fruits, such as
strawberries, is one of the main factors; for both control and coated fruits, the
breaking force decreased as a function of storage time. The formulations that
minimized weight loss maintained better firmness, since this attribute is also highly
influenced by water content.
Color changes: The effect of coatings on surface color modifications in
strawberries was also analyzed, because this quality attribute may determine
consumer acceptability of the fruit. Lightness (L) and chromaticity parameters
(a and b) were recorded at 1st, 8th, 15th and 22nd days of storage. Both
plasticizers significantly delayed surface color development. Formulations con-
taining glycerol gave better surface color results compared to those containing
sorbitol. Oil addition did not modify surface color results significantly, regardless
of plasticizer used in formulation (Fig. 27.3e).
Physiological parameters of fruit, such as titratable acids, pH, anthocyanin,
and sugar content are good indicators of maturation and senescence. Coatings
may alter natural physiological behavior, modifying the organoleptic character-
istics of fruit such as color, taste, or flavor. In coated strawberries, these
physiological parameters were slowed down but reached commercially accept-
able values. The results indicate that starch-based coatings retard the metabolic
reactions, and thus, senescence of coated refrigerated fruits is delayed. This
result is attributed to the differential gaseous permeability of films. The O2 and
CO2 barriers lead to a reduction in respiration rate by limiting the exposure to
ambient O2, increasing internal CO2, delaying ripening, senescence, and extend-
ing storage life of treated fruits (Baldwin 1994; Avena-Bustillos et al. 1997;
Garcı́a et al. 1998a, b, 2001).
In microbiological analysis, surface microbial growth is the main cause of
spoilage for many food products. Microorganisms growing on strawberries were
mainly yeasts, molds, and sugar-fermenting bacteria. Aerobic mesophilic and
psychrotrophic bacteria, molds and yeasts, and coliform microorganisms were
analyzed at different storage times. Viable counts were expressed as log CFU/g
fruit. Microbial counts in unplasticized, coated strawberries did not differ signifi-
cantly from the control. Coatings with potassium sorbate, a well-known effective
antifungal agent, significantly decreased (P < 0.05) yeast and mold counts on
coated strawberries. Since the undissociated form is the active antimicrobial
agent of sorbic acid, citric acid at fruit pH was added to coating formulation to
increase potassium sorbate effectiveness. The addition of 0.2 g/L potassium sorbate
654 N. Zaritzky
and critic acid to the starch-based coating was the most effective formulation for
decreasing microbial counts. This active coating allows using lower amounts of
preservatives, maintaining the same antimicrobial efficacy as traditional techni-
ques, with consequent health and economic benefits.
The storage life of a refrigerated fruit, defined as the time necessary to reach
106 CFU/g, was determined for the different formulations. At 0 C, the storage life
of the uncoated fruit was 14 days; coatings with sorbitol extended fruit storage life
to 21 days. At maximum storage time assayed (28 days), formulations with sorbitol
and potassium sorbate showed microbial counts below 106 CFU/g in fruit. The
addition of potassium sorbate enhanced the effectiveness of starch coatings; more-
over, the addition of citric acid (to reach pH 4) increased the antimicrobial action
of potassium sorbate, leading to a shelf-life of more than 28 days for coated
strawberry.
The presence of antimicrobial agents in the coating provided a local high and
effective concentration of the preservative. Edible films and coatings act as surface-
retention agents and limit preservative diffusion in the food core. This allows
reducing the total amount of preservative added to a food compared to traditional
methods (e.g., dipping in chemical preservative solutions) (Garcı́a et al. 1998a, b,
2001).
In conclusion, starch-based coatings with selective gaseous permeabilities
proved to extend the storage life of refrigerated strawberries. Plasticizer and lipid
addition improved coating performance by decreasing WVP. Weight losses were
reduced, color changes delayed, and firmness of tissue and fruit appearance
improved. Modifications of the physiological parameters of the fruit were slowed
down but reached commercially acceptable values. Starch-based coatings contain-
ing potassium sorbate and citric acid helped decrease microbial growth; as a result,
shelf-life of the fruit was extended to more than 28 days compared with 14 days for
uncoated fruits.
27.10 Final Remarks
Edible films and coatings enhance the quality of food products, protecting them
from physical, chemical, and biological deterioration, which results in extended
shelf-life and improved safety. They can be used on fruits, vegetables, seafood,
meats, and confectionery products. Edible coatings can help retain or improve food
product quality by:
l Forming an efficient barrier to prevent moisture loss.
l Delaying ripening process in vegetables, through selective permeability to gases
that affects postharvest metabolism, extending the storage life.
l Reducing OU in frying process.
l Adding vitamins or other functional ingredients to enhance quality.
l Incorporating active additives such as antimicrobial agents and antioxidants.

Active components are located at the surface of the product, where their action is
required. Coating matrix limits diffusion into the core of the food and thus,
minor additive concentrations are needed compared to bulk addition.
The factors contributing to renewed interest in the development of edible
coatings include:
l Consumer demand for safe and high quality foods, with longer shelf-life.
l Interest of food processors in new storage techniques.
l Opportunities for creating new markets for film-forming ingredients derived
from under-utilized agricultural commodities.
Successful growth in the area of edible films will require strong interaction
between food technologies and polymer science. Collaborative research work will
allow edible coatings to be used in target applications. Moreover, the development
of new technologies (functionalization, cross-linking, etc.) to improve the film
properties (control release, bioactivity protection, resistance to water etc.) of active
packaging and coatings is a major focus for future research.
References
Amarante C, Banks NH (2001) Post harvest physiology and quality of coated fruits and vegetables.
Hort Rev 26:161–238
ASTM (1995) Standard test methods for water vapor transmission of material. In: Annual book of
ASTM. American Society for Testing and Materials, Philadelphia, PA, pp E96–95
Avena-Bustillos RJ, Krochta JM, Saltveit ME (1997) Water vapor resistance of red delicious
apples and celery sticks coated with edible caseinate-acetylated mono-glyceride films. J Food
Sci 62(2):351–354
Balasubramaniam M, Chinnan MS, Mallikarjunan P, Phillips RD (1997) The effect of edible film
on oil uptake and moisture retention of a deep-fat fried poultry product. J Food Process Eng
20:17–29
Baldwin EA (1994) Edible coatings for fresh fruits and vegetables: past, present and future.
In: Krochta JM, Baldwin EA, Nisperos-Carriedo M (eds) Edible coatings and films to improve
food quality. Technomic Publishing, Lancaster, PA, pp 25–64
Baldwin EA, Nispero-Carriedo MO, Baker RA (1995) Edible coatings for lightly processed fruits
and vegetables. Hort Sci 30(1):35–38
Baldwin EA, Nispero-Carriedo MO, Hagenmaier RD, Baker RA (1997) Use of lipids in coatings
for food products. Food Technol 51(6):56–64
Bertolini Suarez R, Campañone LA, Garcı́a MA, Zaritzky NE (2008) Comparison of the deep
frying process in coated and uncoated dough systems. J Food Eng 84:383–393
Bifani V, Ramı́rez C, Ihla M, Rubilara M, Garcı́a MA, Zaritzky N (2007) Effects of murta
(Ugni molinae Turcz) extract on gas and water vapor permeability of carboxymethylcellulose
based edible films. Lebensm Wiss Technol-Food Sci Technol 40:1473–1481
Brody AL, Strupinsky ER, Kline LR (2001) Antimicrobial packaging. In: Active packaging for
food applications. CRC, Boca Raton, FL, pp 131–194
Cagri A, Ustunol Z, Ryser ET (2004) Antimicrobial edible films and coatings. J Food Prot
67(4):833–848
656 N. Zaritzky
Cha DS, Chinnan MS (2004) Biopolymer-based antimicrobial packaging: a review. Crit Rev Food
Sci Nutr 44:223–237
Cisneros-Zevallos L, Saltveit ME, Krochta JM (1997) Hygroscopic coatings control surface white
discoloration of peeled (minimally processes) carrots during storage. J Food Sci
62(2):363–366, 398
Cuq B, Gontard N, Guilbert S (1995) Edible films and coatings as active layers. In: Rooney M (ed)
Active food packaging. Blackie Academic and Professional, Glasgow, UK, pp 111–142
Cuq B, Gontard N, Guilbert S (1998) Proteins as agricultural polymers for packaging production.
Cereal Chem 75(1):1–9
Debeaufort F, Quezada-Gallo JA, Voilley A (1998) Edible films and coatings: tomorrow’s
packaging: a review. Crit Rev Food Sci Nutr 38(4):299–313
Donhowe IG, Fennema O (1994) Edible films and coatings: characteristics, formation, definitions,
and testing methods. In: Krochta JM, Baldwin EA, Nisperos-Carriedo M (eds) Edible coatings
and films to improve food quality. Technomic Publishing, Lancaster, PA, pp 1–24
Drake SR, Nelson JW (1990) Storage quality of waxed and nonwaxed “Delicious” and “Golden
Delicious” apples. J Food Qual 13:331–334
El Gaouth A, Arul J, Ponnampalam R, Boulet M (1991) Chitosan coating effect on storability and
quality of fresh strawberry. J Food Sci 56(6):618–1620
Franssen LR, Krochta JM (2003) Edible coatings containing natural antimicrobials for processed
foods. In: Roller S (ed) Natural antimicrobials for the minimal processing of foods. Woodhead
Publishing Limited, Cambridge, UK
Garcı́a MA, Martino MN, Zaritzky NE (1998a) Plasticizer effect on starch-based coatings applied
to strawberries (Fragaria ananassa). J Agr Food Chem 46:3758–3767
Garcı́a MA, Martino MN, Zaritzky NE (1998b) Starch-based coatings: effect on refrigerated
strawberry (Fragaria ananassa) quality. J Sci Food Agric 76(3):411–420
Garcı́a MA, Martino MN, Zaritzky NE (1999) Edible starch films and coatings characterization:
scanning electron microscopy, water vapor, and gas permeabilities. Scanning 21(5):348–353
Garcı́a MA, Martino MN, Zaritzky NE (2000a) Lipid addition to improve barrier properties of
edible starch-based films and coatings. J Food Sci 65(6):941–947
Garcı́a MA, Martino MN, Zaritzky NE (2000b) Microstructural characterization of plasticized
starch-based films. Starch/Staerke 52(4):118–124
Garcı́a MA, Martino MN, Zaritzky NE (2001) Composite starch-based coatings applied to
strawberries (Fragaria ananassa). Nahrung 45(4):267–272
Garcı́a MA, Ferrero C, Campana A, Bértola N, Martino M, Zaritzky N (2004a) Methylcellulose
coatings applied to reduce oil uptake in fried products. Food Sci Technol Int 10(5):339–346
Garcı́a MA, Pinotti A, Martino MN, Zaritzky NE (2004b) Characterization of composite hydro-
colloid films. Carbohydr Polym 56(3):339–345
Garcı́a MA, Pinotti A, Zaritzky NE (2006) Physicochemical, water vapor barrier and mechanical
properties of corn starch and chitosan composite films. Starch/Staerke 58(9):453–463
Garcı̀a MA, Ferrero C, Bértola N, Martino M, Zaritzky N (2002) Edible coatings from cellulose
derivatives to reduce oil uptake in fried products. Innovative Food Sci Emerg Technol 3(4):391–397
Gennadios A, Weller CL (1990) Edible films and coatings from wheat and corn proteins. Food
Technol 44(10):63–69
Gennadios A, McHugh TH, Weller CL, Krochta JM (1994) Edible coatings and film based
proteins. In: Krochta JM, Baldwin EA, Nisperos-Carriedo M (eds) Edible coatings and films
to improve food quality. Technomic Publishing, Lancaster, PA, pp 201–277
Gennadios A, Hanna MA, Kurth LB (1997) Application of edible coatings on meats, poultry and
seafoods: a review. Lebensm Wiss Technol 30(4):337–350
Gontard N, Guilbert S, Cuq JL (1992) Edible wheat gluten films: influence of the main process
variables on film properties using response surface methodology. J Food Sci 57(190–195):199
Gontard N, Thibault R, Cuq B, Guilbert S (1996) Influence of relative humidity and film
composition on oxygen and carbon dioxide permeabilities of edible films. J Agr Food Chem
44(4):1064–1069
Greener IK, Fennema O (1989) Barrier properties and surface characteristics of edible, bilayer
films. J Food Sci 54:1393–1399
Greener IK, Fennema O (1992) Lipid-based edible films and coatings. Lipid Technol 4:34–38
Guilbert S (1986) Technology and application of edible protective films. In: Mathlouthi M (ed)
Food packaging and preservation. Theory and practice. Elsevier Applied Science Publishing
Co, London, UK, pp 371–394
Guilbert S, Gontard N (1995) Edible and biodegradable food packaging. In: Ackermann P, Jagerstad
M, Ohlsson T (eds) Foods and packaging materials – chemical interactions. The Royal Society
of Chemistry, Cambridge, England, pp 159–168
Guilbert S, Gontard N (2005) Agro-polymers for edible and biodegradable films: review of
agricultural polymeric materials, physical and mechanical characteristics, Chapter 4.16. In:
Han JH (ed) Innovations in food packaging. Elsevier Academic, Oxford, England, pp 263–276
Guilbert S, Gontard N, Gorris LGM (1996) Prolongation of the shelf life of perishable food products
using biodegradable films and coatings. Lebensm Wiss Technol-Food Sci Technol 29:10–17
Guilbert S, Cuq B, Gontard N (1997) Recent innovations in edible and/or biodegradable packaging
materials. Food Addit Contam 14(6–7):741–751
Guilbert S, Gontard N, Morel MH, Chalier P, Micard X, Redl A (2001) Formation and properties
of wheat gluten films and coatings. In: Gennadios A (ed) Protein-based films and coatings.
CRC, Boca Raton, FL, pp 69–122
Han JH (2000) Antimicrobial food packaging. Food Technol 54(3):56–65
Han JH (2002) Protein-based edible films and coatings carrying antimicrobial agents. In:
Gennadios A (ed) Protein-based films and coatings. CRC, Boca Raton, FL, pp 485–499
Han JH (2003a) Antimicrobial food packaging. In: Ahvenainen R (ed) Novel food packaging
techniques. Woodhead Publishing Ltd, Cambridge, UK, pp 50–70
Han JH (2003b) Design of antimicrobial packaging systems. Int Rev Food Sci Technol
11:106–109
Han JH (2005) New technologies in food packaging: overview. In: Han JH (ed) Innovations in
food packaging. Elsevier Academic, Oxford, England, pp 3–11
Han JH, Gennadios A (2005) Edible films and coatings: a review. In: Han JH (ed) Innovations in
food packaging. Elsevier Academic, Oxford, England, pp 239–262
Han C, Zhao Y, Leonard SW, Traber MG (2004) Edible coatings to improve storability and
enhance nutritional value of fresh and frozen strawberries (Fragaria ananassa) and raspberries
(Rubus ideaus). Postharvest Biol Technol 33(1):67–78
Kester JJ, Fennema OR (1986) Edible films and coatings: a review. Food Technol 48(12):47–59
Krochta JM (1997) Edible protein films and coatings. In: Damodaran S, Paraf A (eds) Food
proteins and their applications. Marcel Dekker, New York, pp 529–549
Krochta JM (2002) Proteins as raw materials for films and coatings: definitions, current status, and
opportunities. In: Gennadios A (ed) Protein-based films and coatings. CRC, Boca Raton, FL,
pp 1–41
Krochta JM, De Mulder-Johnston C (1997) Edible and biodegradable polymer films: challenges
and opportunities. Food Technol 51(2):61–74
Krochta JM, Baldwin EA, Nisperos-Carriedo MO (1994) Edible coatings and films to improve
food quality. Technomic, Lancaster, PA
Lacroix M, Le Tien C (2005) Edible films and coatings from nonstarch polysaccharides. In: Han
JH (ed) Innovations in food packaging. Elsevier Academic, Oxford, England, pp 338–359
Lin D, Zhao Y (2007) Innovations in the development and application of edible coatings for fresh
and minimally processed fruits and vegetables. Compr Rev Food Sci Food Safety 6:60–74
Lopez OV, Garcia MA, Zaritzky N (2008) Film forming capacity of chemically modified corn
starches. Carbohydr Polym 73:573–581
Mali S, Grossmann MVE, Garcia MA, Martino MN, Zaritzky NE (2002) Microstructural charac-
terization of yam starch films. Carbohydr Polym 50(4):379–386
Mali S, Grossmann MVE, Garcia MA, Martino MN, Zaritzky NE (2004) Barrier, mechanical and
optical properties of plasticized yam starch films. Carbohydr Polym 56(2):129–135
658 N. Zaritzky
Mali S, Grossmann MVE, Garcı́a MA, Martino MN, Zaritzky NE (2005) Mechanical and thermal
properties of yam starch films. Food Hydrocolloids 19(1):157–164
Mali S, Grossmann MVE, Garcia MA, Martino MN, Zaritzky NE (2006) Effects of controlled
storage on thermal, mechanical and barrier properties of plasticized films from different starch
sources. J Food Eng 75(4):453–460
McHugh TH, Krochta JM (1994) Permeability properties of edible films. In: Krochta JM, Baldwin
EA, Nisperos-Carriedo M (eds) Edible coatings and films to improve food quality. Technomic,
Lancaster, PA, pp 139–187
Mei Y, Zhao Y (2003) Barrier and mechanical properties of milk protein-based edible films
incorporated with nutraceuticals. J Agr Food Chem 51(7):1914–1918
Mei Y, Zhao Y, Farr H (2002) Enhancement of nutritional and sensory qualities of fresh baby
carrots by edible coatings. J Food Sci 67(5):1964–1968
Miller KS, Upadhyaya SK, Krochta JM (1998) Permeability of d-limonene in whey protein films.
J Food Sci 63(2):244–247
Min S, Krochta JM (2007) Ascorbic acid-containing whey protein film coatings for control of
oxidation. J Agr Food Chem 55(8):2964–2969
Min S, Rumsey TR, Krochta JM (2008) Diffusion of the antimicrobial lysozyme from a whey
protein coating on smoked salmon. J Food Eng 84:39–47
Nisperos-Carriedo MO (1994) Edible films and coatings based on polysaccharides. In: Krochta JM,
Baldwin EA, Nisperos-Carriedo MO (eds) Edible coatings and films to improve food quality.
Technomic Publishing, Lancaster, PA, pp 305–335
Nussinovitch A (2003) Water soluble polymer applications in foods. Blackwell Science, Oxford,
UK, pp 29–69
Park HJ (1999) Development of advanced edible coatings for fruits. Trends Food Sci Technol
10:254–260
Park HJ (2005) Edible coatings for fruit. In: Jongen W (ed) Fruit and vegetable processing. CRC,
Boca Raton, FL
Park SI, Zhao Y (2004) Incorporation of a high concentration of mineral or vitamin into chitosan-
based films. J Agr Food Chem 52:1933–1939
Park SK, Hettiarachchy NS, Ju ZY, Gennadios A (2002) Formation and properties of soy protein
films and coatings. In: Gennadios A (ed) Protein-based films and coatings. CRC, Boca Raton,
FL, pp 123–137
Park SI, Daeschel MA, Zhao Y (2004) Functional properties of antimicrobial lysozyme chitosan
composite films. J Food Sci 69:215–221
Park SI, Stan SD, Daeschel MA, Zhao Y (2005) Antifungal coatings on fresh strawberries
(Fragaria ananassa) to control mold growth during cold storage. J Food Sci 70:202–207
Pérez-Gago MB, Krochta JM (2005) Emulsion and bi-layer edible films. In: Han JH (ed) Innova-
tions in food packaging. Elsevier Academic, Oxford, England, pp 384–402
Pinotti A, Garcı́a MA, Martino MN, Zaritzky NE (2007) Study on microstructure and physical
properties of composite films based on chitosan and methylcellulose. Food Hydrocolloids
21(1):66–72
Redl A, Gontard N, Guilbert S (1996) Determination of sorbic acid diffusivity in edible wheat
gluten and lipid based films. J Food Sci 61:116–120
Romero-Bastida CA, Bello-Pérez LA, Garcı́a MA, Martino MN, Solorza-Feria J, Zaritzky NE
(2005) Physicochemical and microstructural characterization of films prepared by thermal and
cold gelatinization from non-conventional sources of starches. Carbohydr Polym
60(2):235–244
Shellhammer TH, Krochta JM (1997) Whey protein emulsion film performance as affected by
lipid type amount. J Food Sci 62(2):390–394
Sothornvit R, Krochta JM (2000) Plasticizer effect on oxygen permeability of beta -lactoglobulin
films. J Agr Food Chem 48:6298–6302
Sothornvit R, Krochta JM (2001) Plasticizer effect on mechanical properties of beta -lactoglobulin
films. J Food Eng 50:149–155
Torres JA, Karel M (1985) Microbial stabilization of intermediate moisture food surfaces III.
Effects of surface preservative concentration and surface pH control on microbial stability of
an intermediate moisture cheese analog. J Food Process Preserv 9:107–119
Torres JA, Bouzas J0, Karel M (1985) Microbial stabilization of intermediate moisture food
surfaces II. Control of surface pH. J Food Process Preserv 9:93–106
Vojdani F, Torres JA (1990) Potassium sorbate permeability of methylcellulose and hydroxypro-
pyl methylcellulose coatings: effect of fatty acid. J Food Sci 55:841–846
Williams R, Mittal GS (1999) Low-fat fried foods with edible coatings: modeling and simulation.
J Food Sci 64:317–322
Wu LC, Bates RF (1972) Soy protein-lipid films. I. Studies on the film formation phenomenon.
J Food Sci 37(1):36–39
Wu Y, Weller CL, Hamouz F, Cuppett SL, Schnepf M (2002) Development and application of
multicomponent edible coatings and films: a review. Adv Food Nutr Res 44:347–394
Zhao Y, McDaniel M (2005) Sensory quality of foods associated with edible film and coating
systems and shelf life extension. In: Han JH (ed) Innovations in food packaging. Elsevier
Academic, Oxford, England, pp 434–453
.
Chapter 28
Physical Properties of Edible Gelatin Films
Colored with Chlorophyllide
Paulo J.A. Sobral, Rosemary A. Carvalho, and Carmen S. Fávaro-Trindade
28.1 Introduction
The question of the environmental impact of synthetic packaging has generally

favored research on edible films and, in particular, biodegradable films, which are
flexible materials elaborated with biological macromolecules capable of forming a
continuous matrix (Gontard and Guilbert 1996; Krochta and De-Mulder-Johnston
1997; Van de Velde and Kiekens 2002; Tharanathan 2003). The main biopolymers
of agricultural origin used in the elaboration of films are polysaccharides (Nisperos-
Carriedo 1994) and proteins (Gennadios et al. 1994; Torres 1994). Synthetic
biopolymers have also been studied in depth in recent decades (Van de Velde and
Kiekens 2002).
Among the proteins, gelatin has stimulated much interest due to its excellent
filmogenic properties (Arvanitoyannis 2002). Gelatin was one of the first macro-
molecules used in the production of films, and has continued to be widely studied
for this purpose due to its functional properties, large scale production, and com-
petitive prices (Arvanitoyannis 2002).
The properties of the macromolecules directly influence the final characteristics
of the edible films. The properties of protein-based films are determined by the
protein-protein and protein-water interactions, which can be controlled by the
preparation conditions of the FFS and by the addition of plasticizers (Gennadios
et al. 1994; Torres 1994; Gontard and Guilbert 1996; Arvanitoyannis 2002).
Plasticizers are low molecular weight molecules and therefore can penetrate the
biopolymeric matrix, increasing the free space between the macromolecular chains,
causing a decrease in the intermolecular forces throughout the matrix (Banker
1966; Krochta and De-Mulder-Johnston 1997).
In practical terms, the final characteristics of edible and/or biodegradable films
are the result of numerous parameters, such as the characteristics and concentra-
tions of the macromolecule and other constituents (solvent, plasticizer, etc.); pH;
P.J.A. Sobral (*), R.A. Carvalho, and C.S. Fávaro-Trindade

Department of Food Engineering, University of São Paulo, Pirassununga (SP), Brazil
e-mail: pjsobral@usp.br
662 P.J.A. Sobral et al.
denaturation conditions (in the case of proteins); type of support used; drying
conditions; and environmental conditions (temperature and humidity).
Protein-based films, including those from gelatin, generally present excellent
gas-barrier properties and good mechanical properties. In addition, they can serve
as a support for active substances such as antioxidants, antimicrobial agents, etc.
(Gómez-Guillén et al. 2007; López-Carballo et al. 2008; Gómez-Estaca et al. 2009).
On another side, two papers on biopolymer-based films containing natural pigments
can be found, in which the pigments in both cases are derived from chlorophyll
(Corat et al. 2007; López-Carballo et al. 2008).
Since chlorophyll is a pigment, it can confer a green color on the material, which
can be an added attraction. However, the properties of the material, notably the
light-barrier properties, may be affected by the presence of this pigment (Corat et al.
2007). Thus, this chapter will present and discuss the effects of adding chlorophyll
to the formulation of gelatin-based edible films, which would also permit the
production of more attractive packaging with improved light-barrier properties.
28.2 Gelatin and Edible Films
Gelatin is a protein of animal origin resulting from the acidic or basic hydrolysis of
the collagen obtained from bones, bovine or swine hides, or connective tissue
(Gennadios et al. 1994). Configuration of the peptide chain in gelatin is normally
controlled by interactions between the solvent and the amino acids in the main
chain, and by certain preferential orientations of the peptide bonds, such that the
amino acid sequence is responsible for the characteristics of the gelatin (Veis 1964).
In general, glycine, proline, and hydroxyproline constitute one-third of the total
amino acids in gelatin (Veis 1964; Gennadios et al. 1994).
The molecular weight of gelatin depends on the type of raw material and process
conditions used, varying from 3,000 to 200,000 Da (Gennadios et al. 1994).
According to the type of pretreatment used to remove impurities and to start the
hydrolysis of the raw material, the gelatin is classified as Type A when an acid
pretreatment is used, resulting in an isoelectric point between 7.0 and 9.4; or Type
B, when a basic pretreatment is used, resulting in an isoelectric point between 4.5
and 5.3 (Veis 1964).
Gelatin is soluble in water at temperatures above 30 C. When submitted to
temperatures above its melting point, the gelatin swells, dissolves, and forms
heat-reversible gels when the temperature returns to room temperature (Slade and
Levine 1987). At the molecular level, the formation of a gelatin gel involves
restructuring of the proteins, implying transformation from a disordered state to
an ordered state, formed by triple-helix structures characteristic of collagen in the
active state, the structure and physical properties of these gels being a result of the
formation of microcrystalline links (Slade and Levine 1987; Achet and He 1995;
Ziegler and Foegeding 1990).
28 Physical Properties of Edible Gelatin Films Colored with Chlorophyllide 663
Various studies involving the formation of edible and/or biodegradable films

have been carried out with mammalian (Arvanitoyannis et al. 1997, 1998a, b; Lim
et al. 1999; Menegalli et al. 1999; Sobral 1999; Sobral et al. 2001; Carvalho and
Grosso 2004; Bertan et al. 2005; Thomazine et al. 2005; Vanin et al. 2005; Carvalho
and Grosso 2006; Carvalho et al. 2008a) and fish (Jongjareonrak et al. 2006a, b;
Gómez-Guillén et al. 2007; Zhang et al. 2007; Carvalho et al. 2008b) gelatin. In
general, gelatin-based films show good mechanical resistance but reduced effi-
ciency as a water vapor barrier. On the other hand, due to the hydrophilic char-
acteristics of gelatin, these films show an elevated susceptibility to environmental
conditions, making the films difficult to use as packaging materials. Various alter-
natives have been studied in an attempt to improve these characteristics, such as
enzymatic modifications (Lim et al. 1999; Carvalho and Grosso 2004), chemical
modifications (Bigi et al. 2001; Carvalho and Grosso 2004, 2006; Cao et al. 2007;
Carvalho et al. 2008a), the incorporation of different types of plasticizers (Vanin
et al. 2005), the use of plasticizer blends (Thomazine et al. 2005), and lipid
incorporation (Bertan et al. 2005), among others. In general, no significant
improvements were observed with respect to the water vapor barrier and mechani-
cal properties.
Another alternative used in an attempt to improve the mechanical resistance
of these materials was the mixture of the biopolymers with synthetic polymers
(Tharanathan 2003). One very interesting synthetic polymer for this type of study is
the polyvinyl alcohol (PVA), which, despite being synthetic, is hydrophilic and
biodegradable (Matsumura et al. 1999). Some films based on blends of PVA with
gelatin have been developed (Chiellini et al. 2001; Bergo et al. 2006; Mendieta-
Taboada et al. 2008; Maria et al. 2008; Silva et al. 2008; Carvalho et al. 2009).
However, all these films were colorless, that is, the inclusion of pigments to make
them more attractive was not studied.
28.3 Chlorophylls
Chlorophylls are the most abundant natural pigments in plants. They occur in the
chloroplasts of leaves and other vegetable tissues, being very common in fruits and
vegetables (Streit et al. 2005; Humphrey 2004). Precursors and derivatives of the
chlorophylls are used in photodynamic medical treatments (Schoefs 2002). In
addition, chlorophyllian pigments are currently of great industrial importance,
since they can be used both as pigments and as antioxidants, although there is
controversy in the research studies with respect to the latter property (Lanfer-
Marquez 2003).
Due to their color and physical properties, the chlorophylls are also used as
colorants for food products (Humphrey 2004; Chattopadhyay et al. 2008). Most
chlorophyll colorants exist in a water-soluble form and are used in dairy products,
soups, oils, sugar confections, drinks, and cosmetics (Francis 2002).
In plants, the chlorophyll molecules are associated with proteins, carotenoids,

and tocopherols by way of noncovalent interactions, which give them certain
stability (Delgado-Vargas and Paredes-López 2002). After their extraction from
the vegetable tissues, these pigments are chemically unstable and can easily be
altered or destroyed, modifying the perception and quality of the products. In
general, the chlorophylls are relatively unstable and sensitive to light, heating,
oxygen, and chemical degradation (Kidmose et al. 2002; Schoefs 2002).
Chlorophylls are molecules formed from complex porphyrin derivatives, con-
stituting the most important subgroup of pigments within the tetrapyrrole group
(Fig. 28.1) (Chattopadhyay et al. 2008). Structurally, they are formed by four
pyrrolic rings and a fifth isocyclic ring located at the side of the third pyrrolic
ring, bonded to each other via methylenic bridges and containing one magnesium
atom on their inside. A molecule of propionic acid esterified to a long chain acyclic
alcohol, usually a phytol, can be found on the fourth pyrrolic ring, conferring a
hydrophobic character to the chlorophyll (Delgado-Vargas and Paredes-López
2002; Francis 2002; Lanfer-Marquez 2003; Humphrey 2004; Chattopadhyay et al.
2008). Porphyrin is a stable ring-shaped molecule around which electrons are free
to migrate. Because the electrons move freely, the ring has the potential to gain or
lose electrons easily, and thus to provide energized electrons to other molecules.
This is the fundamental process by which chlorophyll captures the energy of
sunlight (Wilska-Jeszka 2007).
Two chlorophylls, chlorophyll A and chlorophyll B, are currently important as
food dyes (Chattopadhyay et al. 2008). These pigments are obtained from plants
and only differ with respect to the position of the functional groups –CH3 and
–CHO, respectively, on carbon 3 (Kidmose et al. 2002; Humphrey 2004). These
differences in the structure of the chlorophyll are sufficient to produce different
H Chlorophyll b
H H
O H
H
Chlorophyll a 2 3 CH3
H
H3C 4
1 CH2
H H N N
H CH3
H Mg H
2 3 CH3
H3C 4 H3C 8 N N
1 CH2 5
N N H CH3
H 7 6
H Mg H
CH2 10
H3C 8 N N H2C H 9
5 CO2CH3 O
H CH3 CH3
7 6 O OCH2
H
CH2 10
H2C H 9 CH3 CH3 CH3 CH3
CO2CH3 O
CH3
O OCH2
CH3 CH3 CH3 CH3
Fig. 28.1 Chemical structure of chlorophylls

wavelength absorptions and hence a variety of different green hues. The colors vary
from greenish-yellow to greenish-blue, and the derivatives of these chlorophylls
would be probable producers of orange hues or, under drastic chemical conditions,
a red color (Delgado-Vargas and Paredes-López 2002).
Moreover, chlorophyllide is a green pigment obtained by removal of the phytyl
group of chlorophyll through hydrolysis in dilute alkali or the action of chloro-
phyllase, and is thus water soluble (Kidmose et al. 2002; Lanfer-Marquez 2003).
These chlorophylls are green in color because they absorb strongly in the red and
blue regions of the visible spectrum (Kidmose et al. 2002).
28.4 Experimental Considerations
The films were produced from FFSs containing 2 g gelatin/100 g FFS and 25 g
sorbitol/100 g gelatin. Pigskin gelatin was used, provided by Gelita South America
(São Paulo, Brazil). To prepare these solutions, the gelatin was first hydrated for
30 min. at room temperature, followed by dissolution at 55 C in a water bath (TE
184-Tecnal) (Sobral et al. 2001). After dissolution the plasticizer previously dis-
solved in water and the chlorophyll pigment were added. A water soluble form of
chlorophyll, the chlorophyllide (10% solution of chlorophyll 70008) provided by
Germinal aditivos para alimentos (Cabreúva, Brazil), was used for practical rea-
sons. To test the effect of the pigment concentrations, 0, 2, 4, 6, 8, and 10 g of
chlorophyll solution/100 g gelatin were added to the FFS.
After drying in an oven with air circulation (MA 037-TECNAL) at 30 C for
about 24 h, easily handled films were obtained with thicknesses between 0.075 and
0.081 mm, determined using a Mitutoyo digital micrometer (0.001 mm). The
films were then characterized after preconditioning in desiccators containing NaBr
(relative humidity 58%) at 25 C for at least 7 days. The film characteristics
(mechanical properties, solubility, moisture content, water vapor, and optical bar-
rier properties) were determined in an acclimatized room with a temperature of
about 22 C and relative humidity between 55% and 65%.
The moisture content of the films was determined in an oven at 105 C to
constant weight. The water solubility of the films was determined after 24 h of
immersion, according to Gontard et al. (1993), and expressed in terms of dissolved
dry mass. Water vapor permeability was determined using the method proposed by
Gontard et al. (1993). The films were fixed in cells containing silica gel, which were
then placed in desiccators containing distilled water and maintained at 25 C in an
oven (BOD TE 390 TECNAL) (0.2 C). Weight gain of the system was deter-
mined at 24 h intervals, over a period of 120 h. Water vapor permeability (WVP)
was calculated from (28.1) (Gontard et al. 1993).
w x
WVP ¼ (28.1)
tA DP
where x is the mean film thickness (mm), A is the permeation area (cm2), DP is
the partial vapor pressure difference between the inside of the cell (silica gel,
P1 ¼ 0 Pa) and the distilled water (P2 ¼ 3,166 Pa), and the term w/t corresponds to
the angular coefficient of the linear regression of the graph of mass versus time (g/h).
The mechanical properties (stress at break, elongation at break, and elastic
modulus) of the films were determined by a tensile stress test using the TA.XT2i
texturometer (Stable Micro Systems) with the tensile grip probe, moving at 0.9 mm/s,
according to Thomazine et al. (2005). The color parameters (a*, b*, and L*), the total
color difference (DE*), and the opacity were determined according to Sobral (1999)
using the Miniscan XE (HunterLab) colorimeter. The parameters a*, b*, and L* were
determined by the superimposition of the films on a white standard and the total
difference in color according to (28.2) (Gennadios et al. 1996).
h i0;5
DE ¼ ðDLÞ2 þ ðDaÞ2 þ ðDbÞ2 (28.2)
where DL* ¼ L*standard L*sample, Da* ¼ a*standard a*sample, Db* ¼ b*standard

b*sample.
Opacity was determined according to Sobral (2000) with the same apparatus and
computer program used to measure the color, and was calculated as the ratio
between the opacity of the film superimposed on the black standard (Yb) and that
on the white standard (Yw) (Y ¼ Yb/Yw). The gloss of the films was determined
using the Rhopoint NGL 20/60 glossimeter at an angle of 20 according to
Villalobos et al. (2005). This angle was sufficient because gelatin-based films
have high gloss values (Maria et al. 2008; Silva et al. 2008). The UV and visible
light barrier properties were measured at wavelengths varying from 190 to 800 nm,
using a UV-Visible spectrophotometer (Biochrom, Libra S22), as described by
Fang et al. (2002). The Duncan test was used to compare the means of these results
(a ¼ 5%) using the SAS computer program (version 6.8, SAS Inc., Carry, NC,
USA).
The color parameters were also evaluated in relation to the variation in film
thickness. This study was carried out by superimposing the films from each
formulation and of known thickness, one on top of the other. A colorimeter reading
was made after the addition of each film. The regression analyses were carried out
using the software Statistica 8.0 (StatSoft, Inc.).
28.5 Physical Properties of Gelatin-Based Films Colored

with Chlorophyllide
In general the chlorophyllide concentration in the FFS affected the main physical
properties studied, although with no logical sequence, that is, one could not affirm
that the properties presented in Table 28.1 increased or decreased as a function of
increase in chlorophyllide concentration, despite observing significant differences
Table 28.1 Range of Properties Range

variation in physical
Moisture content 12.0–13.8 g water/100 g film
properties of colored gelatin
Solubility in water 38.6–44.1 g/100 g dry film
films
Water vapor permeability 3.4–6.7 108 g mm/h cm2 Pa
Tensile strength 45.6–53.0 MPa
Elongation at break 11.3–24.1%
Elasticity modulus 13.6–18.1 MPa/%
Gloss at 20 134–198
Opacity 0.3–0.4
(P < 0.05) between the means in some cases. The moisture content of the films
colored with chlorophyll were similar to those determined by Sobral et al. (2001)
and Thomazine et al. (2005) for gelatin-based films plasticized with sorbitol and
conditioned in the same way.
The addition of chlorophyllide did not significantly (P > 0.05) affect the water
solubility of the films, independent of the concentration added. These water solu-
bility result values were slightly higher than those observed by Carvalho and
Grosso (2004, 2006) for gelatin-based (25–30 g/100 g of film) films that were
modified chemically and enzymatically and plasticized with glycerol.
The water vapor permeability oscillated between 3 and 7 108 g mm/h cm2 Pa,
with no clear behavior as a function of increase in chlorophyllide concentration in
the film. Despite some significant differences (P < 0.05) observed between some
of the mean values, such variation did not constitute an important effect, that is, all
the films continued to be highly permeable to water vapor.
The addition of chlorophyllide to the film did affect the mechanical resistance,
although there was no direct relationship between the pigment concentration and
the stress at break. The observed stresses at break values were similar to those
obtained by Thomazine et al. (2005) for gelatin-based films plasticized with sorbitol
(25 g sorbitol/100 g of gelatin). On the other hand, it was observed that an increase
in chlorophyllide concentration caused a greater variation in elongation at break.
Independent of the addition of chlorophyllide, the elongation at break values was
higher than that observed by Thomazine et al. (2005). With respect to the elastic
modulus, the property that indicates the rigidity of the material, an increase in
chlorophyllide concentration did not cause significant variations in this property.
Apparently, the incorporation of chlorophyll can affect the mobility of the poly-
meric matrix due to its structure and possible interaction with protein, a behavior
typical of plasticizing agents. In fact, such behavior can also be observed with
synthetic materials. According to Durston (2006), there are limits to the amount of
colored pigment that can be added to polyethylene, without affecting the mechani-
cal properties of the film.
High gloss samples are better differentiated using measurements at smaller
angles (Villalobos et al. 2005). Thus, only the results obtained at 20 were presented
in Table 28.1. In general, the addition of chlorophyllide caused an increase in gloss
on the film surface, indicating that the incorporation of pigment favored the
morphology of the polymeric matrix, that is, decreased surface roughness as
compared to pure gelatin. The films developed in the present study presented
greater gloss than whey protein isolate-based films dried in a microwave dryer,
with gloss values between 87 and 96 (Kaya and Kaya 2000), or hydroxypropyl
methylcellulose-based films with gloss values below 100 (Villalobos et al. 2005).
The opposite behavior is normally observed with synthetic materials, for which a
specific coloring process should be used to prevent it from adversely affecting the
surface gloss of the films (Durston 2006).
Opacity was also not significantly (P > 0.05) affected by increases in the chlor-
ophyllide concentration in the gelatin films. Thus, it could be suggested that the
inclusion of chlorophyllide in the film formulation, at the concentrations studied, did
not affect the translucence of the films, which continued similar to that of pure
gelatin films, which are extremely transparent (very low opacity) (Sobral 1999;
Vanin et al. 2005). Since many packaging applications require the contents to be
visible, film transparency is of considerable interest (Hanlon et al. 1998).
In addition, this result for opacity suggests that the chlorophyllide was
completely dissolved in the polymeric matrix, since it is known that the presence
of a dispersed, nonmiscible phase promotes opacity as a function of the differences
in refractive indexes of the phases and the concentration and particle size of the
dispersed phase (Villalobos et al. 2005).
28.6 UV and Light Barrier Properties of Gelatin-Based Films

Colored with Chlorophyllide
Although gelatin constitutes an excellent UV light barrier (Fig. 28.2) the presence
of chlorophyllide in the gelatin-based films contributed to an increase in this
property, since the films showed no transmittance at 200 nm, and between 2.6
and 6.0 at 280 nm. The chlorophyllide absorbed UV radiation, having a protective
effect against the degradation of the material (Hanlon et al. 1998) and the packaged
foodstuff (Coltro et al. 2003). Thus, films colored with chlorophyllide show excel-
lent UV barrier properties, suggesting that this type of material could be used to
package foods rich in lipids and/or oxygen-susceptible vitamins (Bekbolet 1990;
Fang et al. 2002; Artharn et al. 2007). According to Saffert et al. (2009), light is
known to have a damaging effect on several foodstuffs, such as milk and other dairy
products, due to the presence of light-sensitive vitamins like vitamins A and B2.
The most pronounced effect of light catalyzed reactions is observed with light in the
lower wavelengths of the visible spectrum and in the UV spectrum (Bekbolet 1990).
Contrarily, Shiku et al. (2004) working on films based on fish sarcoplasmic
proteins, and Artharn et al. (2007), working on films based on fish muscle proteins,
determined no transmittance between 200 and 280 nm. It is highly possible that this
excellent capacity as a UV barrier was caused by the presence of nondissolved
proteins, constituting physical barriers (Villalobos et al. 2005; Monedero et al.
2009). However, Fang et al. (2002), in working with whey-based films, obtained
90
80
70
Transmittance (%)
60
50
40
30
20
10
0
190 290 390 490 590 690 790
Wave length (nm)
Fig. 28.2 Spectra of gelatin films colored with chlorophyllide
values for transmittance between 6 and 7 also between 200 and 280 nm. The great
barrier property of gelatin-based films could be explained by the presence of amino
acids containing aromatic rings in their residues, as can be seen in the aminogram
determined by Sobral et al. (2001).
28.7 Color Characteristics of Gelatin-Based Films Colored

with Chlorophyllide
The addition of chlorophyllide to the gelatin films caused significant alterations in

the parameters a* and b* (Fig. 28.3), but did not affect parameter L*, remaining
practically constant at about 91, which is almost equivalent to the parameter of the
white standard used as the support.
The parameter a*, which varies from green () to red (+) decreased linearly with
the pigment concentration (Cp) (28.3), indicating, effectively, that the material
became greener with increase in concentration of the pigment as expected. On the
other hand, the parameter b*, which varies from blue () to yellow (+), increased
linearly with increase in the pigment concentration (28.4), suggesting that the films
presented a yellowy green coloration.
a ¼ 0:94 0:14 Cp ; R2 ¼ 0:987 (28.3)
b ¼ 2:75 þ 0:19 Cp ; R2 ¼ 0:963 (28.4)
The values of the color parameters determined in the present study were similar
to those obtained by López-Carballo et al. (2008), who characterized edible films
prepared with 10 g of gelatin/100 mL, 2.5 g glycerol/100 mL and 80 mg of
Concentration of chlorophyllide
0.0
0 2 4 6 8 10
–0.5
–1.0
a* –1.5
–2.0
–2.5
–3.0
6.0
5.0
4.0
b* 3.0
2.0
1.0
0.0
0 2 4 6 8 10
Fig. 28.3 Parameters a* (top) and b* (bottom) of gelatin films containing different chlorophyllide
concentrations
chlorophyllin E-140 or E-141/mL: a*¼ 3.2 and 4.4, b* ¼ 7.2 and 4.6, and
L* ¼ 85 and 87, respectively. According to these authors, the films had a vivid
green-yellow color as compared to the gelatin films.
An increase in chlorophyllide concentration also caused a linear increase (28.5)
in the total color difference (Fig. 28.4), signifying that the films containing additive
evidently became more colored with greater pigment concentration.
DE ¼ 3:51 þ 0:19 Cp ; R2 ¼ 0:864 (28.5)
Nevertheless, the values for DE* were not as high as those that would have been
expected from the presence of pigment in the formulation. Although the gelatin films
with added chlorophyll were more colored than films based on various other proteins
without the addition of pigments, such as gelatin (Vanin et al. 2005) and ovoalbumins
4
DE*
3
0
0 2 4 6 8 10
Fig. 28.4 Color difference (DE*) in gelatin films containing different chlorophyllide concentrations
Table 28.2 Examples of color parameters in protein-based edible films

Films DEi ai bi Li References
Egg albumen films, 40 g glycerol/100 g 1.7 0.5 3.5 96 Gennadios et al.
protein. (i ¼ 1) (1996)
Water-soluble fish proteins films, 25 g 7.0 0.8 8 33 Bourtoom et al.
sorbitol/100 g protein. (i ¼ *) (2006)
Laboratory-prepared soy protein isolate 8.5 2.3 9.9 95 Kunte et al. (1997)
films, 30 g glycerol/100 g protein.
(i ¼ 1)
Films based on Amaranth flour, 22.5 g 8.9 1.2 8.1 90 Tápia-Blácido
glycerol/100 g flour. (i ¼ *) et al. (2007)
Round scad mince based film, dark – 2.5 20.7 84 Artharn et al.
muscle, 50 g glycerol/100 g protein. (2007)
(i ¼ *)
Wheat gluten films, 30 g glycerol/100 g – 10.5 45.4 75 Micard et al.
protein, heat treated at 140 C/15 min. (2000)
(i ¼ *)
Fish muscle protein, 40 g glycerol/100 g 8.5 1.6 8.9 97 Paschoalick et al.
protein, heat treated at 40 C/30 min. (2003)
(i ¼ *)
Mammalian gelatin,
53 g glycerin/100 g gelatin (i ¼ *) 3 1 2.8 91 Sobral (1999)
Blends of gelatin and PVA, hydrolysis 3.2 1 2.6 92 Maria et al. (2008)
degree of 95.7%, 25 g glycerol/100 g
macromolecules (i ¼ *)
Fish skin gelatin,
25 g glycerol/100 g proteins. (i ¼ *) – 0.4 1.6 9.5 Jongjareonrak
et al. (2006a)
(i ¼ 1): HunterLab; (i ¼ *): CIELab
(Gennadios et al. 1996), they were apparently less colored than other films, also those
containing no pigments such as films based on Nile Tilapia myofibrillar proteins
(Sobral 2000) and soy proteins (Kunte et al. 1997) (Table 28.2).
However, despite the behavior of increasing coloration with increasing chlo-

rophyll concentration in the film formulation, the coloration was practically
imperceptible to the eye. According to Nassau (1996), the appearance of color
as detected by the eye and interpreted by the brain depends significantly on the
exact viewing circumstances, including sample transparency. In fact, the color
of a pigmented medium, such as a plastic, depends on both transmittance and
reflectance, which in turn are different if the material is opaque or transparent
(Saunderson 1942). Moreover, according to Hanlon et al. (1998), in some
plastics the thickness of translucent material has a direct bearing on the chroma
{C* ¼ [(a*)2 + (b*)2]0.5}.
Thus, considering that, although transparent, the opacity of gelatin films
increases with increase in thickness (Sobral 1999), the perception of the color of
the material would be different depending on the thickness of the films. Thus,
analyses of the color of superimposed films, simulating films with increasing
thickness, allowed for the confirmation that the thickness effectively influenced
film coloration.
Figures 28.5 and 28.6 show that the parameters a*, b*, L* and the total color
difference (DE*) varied linearly with the thickness (x) of the films, with different
slopes, depending on the chlorophyllide concentrations (Cp) in the film formulation.
Thus, these data were submitted to a multilinear regression analysis to observe both
effects of x and Cp (28.6)–(28.9):
a ¼ 0:89 0:57x 0:77 Cp ; R2 ¼ 0:886 (28.6)
b ¼ 0:70 þ 0:73x þ 0:60 Cp ; R2 ¼ 0:927 (28.7)
L ¼ 95:76 0:97x 0:26 Cp ; R2 ¼ 0:985 (28.8)
DE ¼ 2:79 þ 0:91x þ 0:42 Cp ; R2 ¼ 0:969 (28.9)
The quality of fitting can be also observed in Fig. 28.7, where it is possible to
observe the capacity of all models to predict the experimental data. The data are
closely banded around the straight line with a slope of 45 , which indicates the
effectiveness of the (28.6)–(28.9) in describing the effect of pigment concentration
and thickness on color parameters.
The linear behavior between the total color difference and thickness of biopolymer-
based films was verified by Sobral (1999, 2000) working with gelatin and Tilapia
myofibrillar protein films (respectively). However, no paper was found in the
specialized literature on the effect of the concentration of a pigment on the actual
effect of the thickness.
Moreover, it can be noted now that the thicker films (x!0.400 mm) presented
very high color difference values (DE* ffi 23), much higher than those observed for
colorless films (Table 28.2).
Thickness (mm)
0
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450
–1
–2
–3
a* –4
0g chlorophillide
–5
2g chlorophillide
–6 4g chlorophillide
8g chlorophfilide
–9
18
16
14
12
b* 10
8
6
4
2
0
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450
Thickness (mm)
95
90
85
L*
80
75
70
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450
Thickness (mm)
Fig. 28.5 Parameters a* (top), b* (middle), and L* (bottom) of gelatin films containing different
chlorophyllide concentrations, as a function of thickness
25
0 g chlorophillide
20 2 g chlorophillide
4 g chlorophillide
6 g chlorophillide
15 8 g chlorophillide
DE*
10 g chlorophillide
10
0
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450
Thickness (mm)
Fig. 28.6 Total color difference (DE*) in gelatin films containing different chlorophyllide con-
centrations, as a function of thickness
25 95
20
Predicted DE* values
90
Predicted L* values
15
85
10
80
5
0 75
0 5 10 15 20 25 75 80 85 90 95
Experimental DE* data Experimental L* data
Experimental a* data 15
0
–8 –6 –4 –2 0
12
Predicted b* values
–2
Predicted a* values
–4 6
3
–6
0
0 3 6 9 12 15
–8 Experimental b* data
Fig. 28.7 Predicted values using (28.7)–(28.9) against experimental data of color parameters
(slope of straight lines is 45 )
28.8 Conclusion
The use of a water-soluble pigment such as chlorophyll in the form of chlorophyl-

lide is perfectly viable in edible film technology. The addition of this pigment does
not affect the principal physical or functional properties of the material as compared
to pure gelatin. However, perception of the green coloration is difficult, since the
material is transparent. An increase in film thickness could contribute to a better
perception of the color provided by the pigment.
Acknowledgements to the Foundation for Research Support of the State of São Paulo (FAPESP)
and the National Council for Scientific and Technological Development (CNPq) for their support.
References
Achet DÇ, He XW (1995) Determination of the renaturation level in gelatin films. Polymer
36(4):787–791
Artharn A, Benjakul S, Prodpran T, Tanaka M (2007) Properties of a protein-based film from
round scad (Decapterus maruadsi) as affected by muscle types and washing. Food Chem
103:867–874
Arvanitoyannis IS (2002) Formation and properties of collagen and gelatin films and coatings. In:
Gennadios A (ed) Protein-based films and coatings. CRC Press, Boca Raton, pp 275–304
Arvanitoyannis I, Psomiadou E, Nakayama A, Aiba S, Yamamoto N (1997) Edible films made
from gelatin, soluble starch and polyols, Part 3. Food Chem 60(4):593–604
Arvanitoyannis I, Nakayama A, Aiba S (1998a) Edible films made from hydroxypropyl starch and
gelatin and plasticized by polyols and water. Carbohydr Polym 36:105–119
Arvanitoyannis I, Nakayama A, Aiba S (1998b) Chitosan and gelatin-based edible films: state
diagrams, mechanical and permeation properties. Carbohydr Polym 37:371–382
Banker GS (1966) Film coating, theory and practice. J Pharmacol Sci 55:81–89
Bekbolet M (1990) Light effects on food. J Food Prot 53:430–440
Bergo PV, Carvalho RA, Sobral PJA, Bevilacqua FRS, Pinto JKC, Souza JP (2006) Microwave
insertion loss measurements in gelatin-based films. Meas Sci Technol 17:3261–3264
Bertan LC, Tanada-Palmu PS, Siani ACC, Grosso RF (2005) Effect of fatty acids and ‘Brazilian
elemi’ on composite films based on gelatin. Food Hydrocolloids 19(1):73–82
Bigi A, Cojazzi G, Panzavolta S, Rubini K, Roveri N (2001) Mechanical and thermal properties of
gelatin films at different degrees of glutaraldehyde crosslinking. Biomaterials 22(7):763–768
Bourtoom T, Chinnan MS, Jantawat P, Sanguandeeku R (2006) Effect of select parameters on the
properties of edible film from water-soluble fish proteins in surimi wash-water. Lebensmittel
Wiss Technol 39:405–418
Cao N, Fu Y, He J (2007) Mechanical properties of gelatin films cross-linked, respectively, by
ferulic acid and tannin acid. Food Hydrocolloids 21(4):575–584
Carvalho RA, Grosso CRF (2004) Characterization of gelatin based films modified with transglu-
taminase, glyoxal and formaldehyde. Food Hydrocolloids 18:717–726
Carvalho RA, Grosso CRF (2006) Properties of chemically modified gelatin films. Braz J Chem
Eng 23(1):45–53
Carvalho RA, Grosso CRF, Sobral PJA (2008a) Effect of chemical treatment on the mechanical
properties, water vapour permeability and sorption isotherms of gelatin-based films. Packag
Technol Sci 21:165–169
Carvalho RA, Sobral PJA, Thomazine M, Habitante AMQB, Giménez B, Guillen CG, Montero P
(2008b) Development of edible films based on differently processed Atlantic halibut (Hippo-
glossus hippoglossus) skin gelatin. Food Hydrocolloids 22:1117–1123
Carvalho RA, Moraes ICF, Bergo PVA, Kamimura ES, Habitante AMQB, Sobral PJA (2009)
Study of some physical properties of biodegradable films based on blends of gelatin and poly
(vinyl alcohol) using a response-surface methodology. Mater Sci Eng, C 29:485–491
Chattopadhyay P, Chatterjee S, Sen SK (2008) Biotechnological potential of natural food grade
biocolorants. Afr J Biotechnol 7(17):2972–2985
Chiellini E, Cinelli P, Fernandes EG, Kenawy ES, Lazzeri A (2001) Gelatin-based blends and
composites. Morphological and thermal mechanical characterization. Biomacromolecules
2:806–811
Coltro L, Padula M, Saron ES, Borghetti J, Buratin AEP (2003) Evaluation of a UV absorber added
to PET bottles for edible oil packaging. Packag Technol Sci 16(1):15–20
Corat M, Carvalho RA, Trindade CSF, Sobral PJA (2007) Produção e caracterização de filmes a
base de gelatina coloridos com clorofila. Alimentos: Ciencia e Ingenierı́a 16:94–96
Delgado-Vargas F, Paredes-López O (2002) Natural colorants for food and nutraceutical uses.
CRC Press, Boca Raton
Durston J (2006) Flexible package closures and sealing systems. In: Theobald N, Winder B (eds)
Package closures and sealing systems. CRC Press, Boca Raton, pp 204–230
Fang Y, Tung MA, Britt IJ, Yada S, Dalgleish DG (2002) Tensile and barrier properties of edible
films made from whey proteins. J Food Sci 67:188–193
Francis FJ (2002) Food colorings. In: MacDougall DB (ed) Colour in food. CRC Press, New York,
pp 307–340
Gennadios A, McHugh TH, Weller CL, Krochta JM (1994) Edible coating and films based on
proteins. In: Krochta JM, Baldwin EA, Nisperos-Carriedo MO (eds) Edible coatings and to
improve food quality. Technomic Publishing Company, Lancaster, pp 201–277
Gennadios A, Weller CL, Handa MA, Froning GW (1996) Mechanical properties of egg albumen
films. J Food Sci 61(3):585–589
Gómez-Estaca J, Giménez B, Montero P, Gómez-Guillén MC (2009) Incorporation of antioxidant
borage extract into edible films based on sole skin gelatin or a commercial fish gelatin. J Food
Eng 92(1):78–85
Gómez-Guillén MC, Ihl M, Bifani V, Silva A, Montero P (2007) Edible films made from tuna-fish
gelatin with antioxidant extracts of two different murta ecotypes leaves (Ugni molinae Turcz).
Food Hydrocolloids 21:1133–1143
Gontard N, Guilbert S (1996) Bio-packaging: technology and properties of edible and/or biode-
gradable material of agricultural origin. Boletim da SBCTA 30(1):3–15
Gontard N, Guilbert S, Cuq J-L (1993) Water and glycerol as plasticizers effect mechanical and
water vapor barrier properties of an edible wheat gluten film. J Agr Food Chem 58:206–211
Hanlon JF, Kelsey RJ, Forcinio HE (1998) Handbook of package engineering, 3rd edn. Technomic
Publishing Company Inc., Lancaster
Humphrey AM (2004) Chlorophyll as a color and functional ingredient. J Food Sci 69(5):
C422–C425
Jongjareonrak A, Benjakul S, Visessanguan W, Prodpran T, Tanaka M (2006a) Characterization of
edible films from skin gelatin of brownstripe red snapper and bigeye snapper. Food Hydro-
colloids 20(4):492–501
Jongjareonrak A, Benjakul S, Visessanguan W, Tanaka M (2006b) Effects of plasticizers on the
properties of edible films from skin gelatin of bigeye snapper and brownstripe red snapper. Eur
Food Res Technol 222:229–235
Kaya S, Kaya A (2000) Microwave drying effects on properties of whey protein isolate edible.
J Food Eng 43:91–96
Kidmose U, Edelenbos M, Nørbæk R, Christensen LP (2002) Colour stability in vegetables.
In: MacDougall DB (ed) Colour in food. CRC Press, New York, pp 189–242
Krochta JM, De-Mulder-Johnston CD (1997) Edible and biodegradable polymer films: challenges
and opportunities. Food Technol 51:61–74
Kunte LA, Gennadios A, Cuppett SL, Hanna MA, Weller CL (1997) Cast films from soy protein
isolates and fractions. Cereal Chem 74(2):115–118
Lanfer-Marquez UM (2003) O papel da clorofila na alimentação: uma revisão. Braz J Pharm Sci
39(3):227–242
Lim LT, Mine Y, Tung A (1999) Barrier and tensile properties of transglutaminase cross-linked
gelatin films as affect by relative humidity, temperature, and glycerol content. J Food Sci
64(4):616–622
López-Carballo G, Hernández-Muñoz P, Gavara R, Ocio MJ (2008) Photoactivated chlorophyllin-
based gelatin films and coatings to prevent microbial contamination of food products. Int J
Maria TMC, Carvalho RA, Sobral PJA, Habitante AMBQ, Solorza-Feria JS (2008) The effect of
the degree of hydrolysis of the PVA and the plasticizer concentration on the color, opacity, and
thermal and mechanical properties of films based on PVA and gelatin blends. J Food Eng
87:191–199
Matsumura S, Tomizawa N, Toki A, Nishikawa K, Toshima K (1999) Novel poly(vinyl alcohol)-
degrading enzyme and the degradation mechanism. Macromolecules 32:7753–7761
Mendieta-Taboada OWM, Sobral PJA, Carvalho RA, Habitante AMBQ (2008) Thermomechani-
cal properties of biodegradable films based on blends of gelatin and poly(vinyl alcohol). Food
Hydrocolloids 22:1485–1492
Menegalli FC, Sobral PJA, Roques M, Laurent S (1999) Characteristics of gelatin biofilms in
relation to drying process conditions near melting. Drying Technol 17:1697–1706
Micard V, Balamri R, Morel M-H, Guilbert S (2000) Properties of chemically and physically
treated wheat gluten films. J Agric Food Chem 48(7):2948–2953
Monedero FM, Fabra MJ, Talens P, Chiralt A (2009) Effect of oleic acid-beeswax mixtures on
mechanical, optical and water barrier properties of soy protein isolate based films. J Food Eng
91(4):509–515
Nassau K (1996) Fundamentals of color science. In: Nassau K (ed) Color for science, art and
technology. North Holland, Lebanon, pp 1–30
Nisperos-Carriedo MO (1994) Edible coatings and films based on polysaccharides. In: Krochta
JM, Baldwin EA, Nisperos-Carriedo M (eds) Edible coatings and films to improve food
quality. Technomic Publishing Company, Lancaster, pp 305–336
Paschoalick TM, Garcia FT, Sobral PJA, Habitante AMQB (2003) Characterization of some
functional properties of edible films based on muscle proteins of Nile Tilapia. Food Hydro-
colloids 17:419–427
Saffert A, Pieper G, Jetten J (2009) Effect of package light transmittance on the vitamin content of
milk, part 3: fortified UHT low-fat milk. Packag Technol Sci 22:31–37
Saunderson JL (1942) Calculation of the color of pigmented plastics. J Opt Soc Am 32(12):
727–729
Schoefs B (2002) Chlorophyll and carotenoid analysis in food products. Properties of the pigments
and methods of analysis. Trends Food Sci Technol 13:361–371
Shiku Y, Hamaguchi PY, Benjakul S, Visessanguan W, Tanaka M (2004) Effect of surimi quality
on properties of edible films based on Alaska Pollack. Food Chem 86:493–499
Silva GGD, Sobral PJA, Carvalho RA, Bergo PVA, Mendieta-Taboada OW, Habitante AMQB
(2008) Biodegradable films based on blends of gelatin and poly (vinyl alcohol): effect of
PVA type or concentration on some physical properties of films. J Polym Environ
16:276–285
Slade L, Levine H (1987) Polymer-chemical properties of gelatin in foods. In: Pearson AM,
Dutson TR, Bailey AJ (eds) Advances in meat research, collagen as a food. Elsevier Applied
Science, London, pp 251–266
Sobral PJA (1999) Propriedades funcionais de biofilmes de gelatina em função da espessura.
Ciência & Engenharia 8:60–67
Sobral PJA (2000) Influência da espessura sobre certas propriedades de biofilmes à base de
proteı́nas miofibrilares. Pesqui Agropecu Bras 35:1251–1259
Sobral PJA, Menegalli FC, Hubinguer MD, Roques MA (2001) Mechanical, water vapor barrier
and thermal properties of gelatin based edible films. Food Hydrocolloids 15:423–32
Streit MN, Canterle LP, Canto MW, Hecktheuer LHH (2005) As clorofilas. Ciência Rural
35:748–755
Tapia-BLácido DT, Mauri A, Menegalli FC, Sobral PJA, Añon MC (2007) Contribution of the
starch, protein and lipid fractions to the physical, thermal and structural properties of amaranth
(amaranthus caudatus) flour films. J Food Sci 72:E293–E300
Tharanathan RN (2003) Biodegradable films and composite coatings: past, present and future.
Trends Food Sci Technol 14:71–78
Thomazine M, Carvalho RA, Sobral PJA (2005) Physical properties of gelatin films plasticized by
blends of glycerol and sorbitol. J Food Sci 70:172–176
Torres JA (1994) Edible films and coatings from proteins. In: Hettiarachy NS, Ziegler GR (eds)
Protein functionality in food systems. Marcel Dekker, New York, pp 467–507
Van de Velde K, Kiekens P (2002) Biopolymers: overview of several properties and consequences
on their applications. Polym Test 21(4):433–442
Vanin FM, Sobral PJA, Menegalli FC, Carvalho RA, Habitante AMQB (2005) Effects of plasti-
cizers and their concentrations on thermal and functional properties of gelatin based films.
Food Hydrocolloids 70:172–176
Veis A (1964) The molecular characterization of gelatin. In: The macromolecular chemistry
of gelatin. Academic, New York
Villalobos R, Chanona J, Hernández P, Gutiérrez G, Chiralt A (2005) Gloss and transparency
of hydroxypropyl methylcellulose films containing surfactants as affected by their microstruc-
ture. Food Hydrocolloids 19:53–61
Wilska-Jeszka J (2007) Food Colorants. In: Sikorski ZE (ed) Chemical and functional properties
of food components. CRC Press, Boca Raton, pp 245–274
Zhang S, Wang Y, Herring JL, Oh J-H (2007) Characterization of edible film fabricated with
channel catfish (Ictalurus punctatus) gelatin extract using selected pretreatment methods.
J Food Sci 72(9):C498–C503
Ziegler CR, Foegeding EA (1990) The gelation of proteins. Adv Food Nutr Res 34:203–298
Index
A Axial dispersion coefficient

ABRS. See Automated batch retort system dimensionless plot, 422, 423
Active antimicrobial packaging, 627 Peclet number, 414
Adaptation mechanisms, 553 Axially dispersed plug flow (ADPF), 401, 408
Adaptive random search method
description, 267–268 B
logistic curve utilization, 268–270 Bacteriocins, 308–309
pedestal frequency distribution Barrier properties, edible coatings
transformation, 266–267 gas permeability, 638–639
Adhesion mechanisms water vapor permeability (WVP),
caking processes, 493 637–638
liquid and solid bridges formation, 494 Binary diffusion coefficient, 432, 435
sintering mechanisms, 495 Bio-plastics
sinter phases, 496, 497 applications of, 624–625
AGV. See Automated guided vehicles categories of, 620–621
Air evaluation, 519 chemical synthesis, 623–624
Airflow profiles simulation, 522 nature-made polymers, 621–623
Anhydrobiotes, 554–555 Bioreactors
Antifreeze proteins (AFPs), 563–564 gas-lift reactors, 130, 131
Antral contraction waves (ACWs), 99–101 important factors, 127–128
Apple tissue, PSD membrane reactors, 130–131
freezing, 603–604 packedbed bioreactors, 128–129
heterogeneity, 610 Bitter pit symptoms, 590
parameters and standard deviations, 608 Brabender® farinograph, 32–34
procedure, 604–606 Branching effects, alkali-soluble corn
radial orientation, 604, 605 arabinoxylans
reconstructed 3-D X-ray, 608, 609 extensional rheology, 91
spectrum computation, 606–608 fractionation, 89–90
spectrums, 609 HPSEC-MALS, 92–93
X-ray cross-sectional image, 605 melt shear rheology, 90
Arabinoxylans, 74–75 solution shear rheology, 92
Aseptic processing, 60
Atomic doping, 581 C
Automated batch retort system (ABRS), Caking
251, 252 adhesion mechanisms
Automated guided vehicles (AGV), caking processes, 493
252–253 liquid and solid bridges formation, 494
DOI 10.1007/978-1-4419-7475-4, # Springer ScienceþBusiness Media, LLC 2011
680 Index
sintering mechanisms, 495 food packaging

sinter phases, 496, 497 indicator, PUI, 203–211
caked powders, 491, 492 levels of, 199
crystalline powders, 508 utilizing resources, 200
crystallinity, skim milk powder, 513 waste management, 200–203
dextrose syrup greenhouse gases (GHG)
glass transition temperatures, sintering, indicators for, 171–174
504, 505 sources, 170
moisture content and visual appearance, manufacturing steps and processes, 167
503, 504 water intake and discharge
particle package collapse, 506 indicators, 190–199
sinter bridge diameter calculation, pollutants, 189
505, 508 withdrawn amount, 188–189
unconfined yield strength, 507 Canned foods sterilization, CFD methods,
yeast, shear locus, 506, 507 58–59
materials, 501–502 Canning plant
measurement method, 502–503 optimization, 277–281
powder flowability and stress states schedule
Mohr’s circle, 500 mathematical model, 273–274
mono-and bi-axial stresses, 500, 501 problem definition, 273
normal and shear stresses action, simultaneous sterilization method,
498, 499 270–272
prism-shaped powder element, 499 Capillarity, 538–539
yield locus, 501 Capillary rheometry, fibers, 85–86
powder mixes Cell aggregation, 127
dehydrated food products, 508–509 Cereal food, glass transitions, 475–476
dextrose syrup, 511, 512 Chlorophylls
unconfined yield strength, 510, 511 A and B, 664, 665
yield loci measurement, 509 chemical structure, 664
processes, 497–498 colorants, 663
re-crystallization influence, 512–514 molecule structure, 581
ring shear tester principle, 502, 503 natural pigments, 663
supra-molecular structure and material Clostridum botulinum studies, 348–349
properties, 492–493 Clostridum sporogenes, 346–348
time consolidation trials, 502, 503 CO2 extraction, kinetic parameters
Canadian food and beverage industry (FBI) axial dispersion coefficient
classification, NAICS, 166 dimensionless plot, 422, 423
ECI and GHGEI indicator Peclet number, 414
calculation methods, 171–172, 175–176 compositions of, 414–418
national results and interpretation, effective diffusivity, solid matrix
176–182 binary diffusion coefficient, 432, 435
plant grouping parameters, 173–174 correction factor, 432–434
provincial results and interpretation, microstructural factor, 428
182–187 oregano, essential oils, 430, 431
response options, 187 particle diameter effect, 436, 437
employees and productivity, 165–166 plant essential oils, 432
energy consumption size-reduction pretreatment, 438–439
direct and indirect impact, 168 superficial glands, 438
indicators for, 171–174 external mass transfer coefficient
input cost, 169 dimensionless plot, 424–426
statistics of, 168–169 experimental vs. correlated values, 427
sub-sectors, 168–169 extraction rate reduction, 426–427
environmental standards, 167 Grashof number, 427–428
Index 681
extraction conditions, 411–413 modification strategies, 76–77

molecular weights (MW) and critical processing of, 73
volume estimation, 414, 419–420 rheological properties, 84–93
sage essential oils composition, 414, 421 structural characterization, 79–83
Cold plasma Correction factor, 432–434
advantages, 304–305 Cow’s udder, 581–582
effect on microorganisms, 306 CT. See Computed tomography
processing conditions, 305–306 Curriculum design
Compression heating, 345–346 doctoral program, 242
Computational fluid dynamics (CFD) guidelines, 240
advantages, 45–46 master degree program, 241–242
applications systemic or systems theory, 238–239
aseptic processing, 60 systems thinking, 239
cooking, 62–63 undergraduate courses, 240–241
drying, 60–62
pasteurization, 59 D
sterilization, 58–59 Darcy’s law, 57, 541, 548
challenges Definitions, food engineering, 4–5
control food processes, 63–64 Dense phase carbon dioxide (DPCD)
efficiency, solution process, 63 design, 308
sensitivity analysis need, 64 spore inactivation, 307
turbulence, 64 thermodynamic diagram, CO2, 306, 307
modeling properties, fluids Desiccation, 554–555
density, 50–51 Desorption-dissolution-diffusion (DDD)
viscosity, 51 models, 401, 406, 407
multiphase flows Dextrose syrup
categories of, 55 glass transition temperatures, sintering,
definition, 54–55 504, 505
Eulerian–Eulerian concept, 55–56 moisture content and visual appearance,
Lagrangian–Eulerian model, 56 503, 504
volume of fluid model, 55 particle package collapse, 506
numerical methods, 57–58 powder mixes, 511, 512
partial differential equations (PDEs) sinter bridge diameter calculation, 505, 508
applications, 46–48 unconfined yield strength, 507
conservation of energy, 49–50 yeast, shear locus, 506, 507
conservation of mass and momentum, Dietary fiber
48–49 fortification, 71
porous media flows, 56–57 importance, 69–70
simulation, 522, 531–532 processing and chemical evaluation
turbulent flows corn fiber, 73–77
large eddy simulation (LES), 52 extrusion processing, 71–73
prediction performance, 51–52 rheological characterization
Reynolds stress models, 52 capillary rheometer, 77–78
RNG k-e model, 53–54 lubricated squeezing flow technique,
standard k-e model, 53 78–79
viscosity models, 53 solution rheology, 77
Computed tomography (CT) solubility, 70
synchrotron X-ray, 592, 593 sources, 70–71
X-ray, 591 structural characterization
Convenient foods, 13 chemical analysis, 79–81
Cooking oven, CFD methods, 62–63 fermentation profiling, 83
Corn fiber. See also Dietary fiber HPSEC techniques, 82
chemistry, 73–75 light scattering, 82–83
682 Index
spectroscopy techniques, 81 computational fluid dynamics (CFD)

thermal analysis, 81 simulation, 60–62, 522, 531–532
structure and functionality, corn bran drying height, 519–522
chemical treatment, 84 gray-level intensity, 527, 529
extrusion, 93–94 mass flow rate, 519
rheological properties, 84–93 mean particle diameter, 519
Diffusion model moisture content and sampling of
assumptions of, 398–399 material, 518
limitations of, 405–406 nonlinear dynamics, 522–523
mathematical models summary, 400–405 non steady-state air patterns, 524–526
source-and-transfer term, 399 oscillations, FDT, 528, 531
Digestion, human stomach fluid flow phase-space diagrams, FDT, 528, 530
ACW speed effect, 111–113 product temperature, 519
circular tube validation, 105–106 simulated trajectories, 527, 528
contraction depth effect, 113–115 spray dryer, 517–518
deformation of, 103–105 spray drying tests, 515–516
density effect of, 111 stages of, 523, 524
flow field types, 100–102 surface 3-D plots, 527, 529
geometry of, 102–103 temperature profile, 524, 525
modeling procedures, 101–115 testing material, 517
pressure validation, 106–107
viscosity effect of, 109–111 E
3D imaging, fruit microstructure Eco-efficiency indicators, Canadian FBI
apple and pear fruit, 594, 595 ECI and GHGEI indicator
fruit and methods, 592, 594 calculation methods, 171–172, 175–176
multiscale modeling, 594–597 national results and interpretation,
3D numerical simulation, mixing flows 176–182
continuous mixer and extruders plant grouping parameters, 173–174
conveying screws, 34–35 provincial results and interpretation,
dispersive mixing, 38–39 182–187
distributive mixing, 37–38 response options, 187
single vs. twin paddle mixer, 35–37 food packaging
time averaged efficiencies, 37–38 indicator, PUI, 203–211
viscoelastic constitutive, 39–42 levels of, 199
dough mixers and kneaders utilizing resources, 200
blades types, 31–32 waste management, 200–203
Brabender® farinograph, 32–34 water intake (WII) and discharge intensity
deformation quantification of, 32, 33 (WDI)
optimizing research, 31 indicators, 190–199
require actions of, 30 pollutants, 189
viscoelastic fluid model, 33 withdrawn amount, 188–189
stirred tank reactors/batch mixers Edible coatings
anchor effects, 29–31 acetic acid and sodium diacetate, 643
DHR impeller, 29 application of, 631, 634–635
impeller designs, 26–28 barrier properties
laminar mixing, 28 gas permeabilities, 638–639
Do-corder, dough kneaders, 31, 32 water vapor permeability (WVP),
Doctoral program, curriculum design, 242 637–638
DPCD. See Dense phase carbon dioxide benzoic acid and sodium benzoate, 643
Drying zones characteristics, 635–636
air evaluation, 519 chitosan, 642
airflow profiles simulation, 522 coating applications, 640–641
Index 683
composite film formation, 639 confectionery, 183

distribution of, 634–635 distilleries sub-sector, 185
food product quality improvement, fruit and vegetable sector, 184
654–655 grain and oilseed sector, 184
free fatty acids, 643 limitations, 187
incorporating functional ingredients meat sector, 183–184
antimicrobial edible films, 642–645 seafood sector, 184
antioxidant, 642 response options, 187
nutraceutical ingredients carriers, Ester production, 134
645–646 Evolution
lactic acid, 643–644 journals, 8
lactoferrin (lactotransferrin), 644–645 national and international meetings, 7
liquid smoke, 645 research
lysozyme, 644 affordable food supply, 11–13
material convenient foods, 13
functional additives, 633–634 innovative new products, 15–16
plasticizers, 633 kinetic models, 9–10
polysaccharides, 632 process design, 10–11
proteins, 632–633 product quality improvements, 14–15
waxes/lipids/resins, 633 safe and wholesome foods, 11
natural preservatives, 642 significant changes, 8
nisin, 644 textbooks, 6–7
parabens, 643 Extend shelf life. See Quality improvement
pediocins, 644 External mass transfer coefficient
physico-chemical properties, 635–636 dimensionless plot, 424–426
polymer structure, 635–636 experimental vs. correlated values, 427
propionic acid, 643 extraction rate reduction, 426–427
quality improvement and extend shelf life Grashof number, 427–428
oil barriers, in fried products, 646–649 Extracellular ice nucleation rate, 562–563
storage life prolongation, 649–654 Extractible agropolymers, 619–620
sorbic acid, 643 Extraction process
spices, herbs, and essential oils, 644 critical parameters, 366–367
Effective diffusivity, solid matrix efficiency, 365
binary diffusion coefficient, 432, 435 lipid-based nutraceuticals
correction factor, 432–434 b-carotene and lycopene, 368–373
microstructural factor, 428 carotenoids, 368–373
oregano, essential oils, 430, 431 specialty oils, 367–368
particle diameter effect, 436, 437 phytochemicals, 373–374
plant essential oils, 432 plant material pretreatment, 365–366
size-reduction pretreatment, 438–439 Extrusion, 481
superficial glands, 438
ELECTRE III method, microbial risk F
assessment, 159–161 Farinograph, dough kneaders, 31, 32
Encapsulation process, 481–482 Fats
Energy consumption intensity (ECI), Canadian crystal networks, 584
FBI hierarchical arrangements, 583–584
calculation methods, 171–172 Fermentation, immobilized cells
expressed units, 175 alcohol, 119–120
national indicators beer, 120
sectoral features, 177–179 bioreactor, 127–131
sub-sectoral features, 176–177 carrier selection and design, 121–127
plant grouping parameters, 173–174 cider and wine, 120–121
provincial indicators flavor formation in, 132–139
684 Index
Fiber. See Dietary fiber multifactorial risk prioritization

Flexible retortable packages frameworks
design and control, 255, 256 consumer assessment, 153–156
flexible and semi-rigid properties, 255 ELECTRE III method, 159–161
shelf-stable canned foods, 255 FAO/WHO guidelines, 150–151
Fluidized bed drying, CFD methods, 60–61 market-level assessment, 152–154
Food microstructures MCDA techniques, 158–159
atomic doping, 581 PROMETHEE analysis, 160–161
cow’s udder, 581–582 public health impact, 151–152
fat crystal networks, 584 ranking risks, 149–150
health, 584–585 social factor assessment, 156
hierarchical arrangements, fats, 583–584 risk management frameworks, 148–149
molecule structure, chlorophyll, 581 Food technology applications, 566–567
muscle tissue, 580 Fortification, fiber, 71
nature, the ultimate provider, 579–580 Fractionation
novel ingredients and nutraceuticals, alkali-soluble corn arabinoxylans, 89–90
586–587 lipid-based nutraceuticals
pleasure, 586 carotenoids, 380–381
processed foods, 577–578 polyunsaturated fatty acids, 375–378
structure formation kinetics, 582–583 tocopherols and phytosterols, 378–380
structure matters, 578–579 phytochemicals
unique materials, 578 essential oils, 381–382
whipped cream, 583 propolis tincture, 381
Food process economics rosemary and sage herb, 381
capital cost, 220 protocols, 374–375
importance of, 219 Free fatty acids, 643
ingredients plants, 235 Freeze-dried foods, 13
manufacturing plants Freezing and drying
characteristics, 230–232 adaptation mechanisms, 553
yogurt, 232–235 desiccation strategies, 554–555
operating cost equilibrium curves, 556, 558, 559
labor requirements, 221–222 food technology applications, 566–567
raw food materials and packaging frozen environments survival, 555–557
materials, 220–221 glass transitions, 477–479
utilities, 222 ice nucleation and crystal growth rates,
preservation plants, orange juice 560, 561
annual operating cost, 229–231 involved mechanisms
annual operating time, 227 extracellular ice nucleation rate,
equipment cost, 227 562–563
processes, 227, 228 high water content vitrification,
process profitability 564–565
capital cost, 222–225 ice crystal growth inhibition, 563–564
discounted cash flow, 225–226 recalcitrant seeds problem, 565–566
manufacturing cost, 225 solids crystallization avoidance,
plant profitability evaluation, 226 561–562
Food rehydration, 547 sugar crystallization, 568–570
Food safety transient ice-like embryos, 559
challenges and implications for Frozen food, glass transitions, 474–475
research group, 161 Fruit microstructure evaluation
research tools and funding, 162 3-D imaging
specialized knowledge, 162 apple and pear fruit, 594, 595
information cards, 156–158 fruit and methods, 592, 594
Index 685
quality and microstructure Glass formation, anhydrobiotes, 554–555

apple, 589, 590 Glass transitions
biological food products, 589–590 challenges
bitter pit symptoms, 590 amorphous lactose, crystallization
disorders, 590, 591 rate, 484
water core, 590 dielectric thermal analysis (DEA), 485
synchrotron X-ray CT, 592, 593 differential scanning calorimetry
X-ray CT, 591 (DSC), 486, 487
Fusel alcohol production, 132–134 dynamic mechanical analysis (DMA),
485, 486
G time dependent characteristics and
Galerkin method, mixing processes, 24–25 relaxation, 483
Gas-lift reactors, 130, 131 confectionary, 474
Gas permeability, 625–626, 638–639 dextrose syrup, 504, 505
Gastric food digestion, human stomach food
modeling procedures cereal, 475–476
ACW speed effect, 111–113 frozen, 474–475
circular tube validation, 105–106 powder and dehydrated, 476–477
contraction depth effect, 113–115 molecular mobility, 473
deformation of, 103–105 opportunities
density effect of, 111 encapsulation process, 481–482
geometry of, 102–103 extrusion, 481
pressure validation, 106–107 freezing and freeze-drying, 477–479
viscosity effect of, 109–111 spray drying, 479–481
stomach and fluid flow vitrification, 477
antral contraction waves (ACWs), solid-fluid state transformation, 473
99–101 Grashof number, 427–428
eddy structures, 101 Greenhouse gas emission intensity (GHGEI),
meal volume effects, 100 Canadian FBI
parts and role of, 99 calculation methods, 171–172
peristaltic flow analysis, 101–102 expressed units, 175
retropulsive jet flow, 100–101 national indicators
size and shape, 100 bakery and seafood sectors, 180
Gelatin films, edible energy type(s) used, 179
biodegradable films, 661 fruit and vegetable sector, 181
colored with chlorophyllide GHGEI vs. energy consumption, 180
color parameters, 669–670, 673 grain and oilseed sector, 181–182
multilinear regression analysis, 672 provincial indicators, 185–187
physical properties, 666–668 response options, 187
total color difference, 670–671, 674
transmittance and reflectance, 672 H
UV and light barrier properties, Health
668–669 food microstructures, 584–585
color parameters, 666 microbial multifactorial risk assessment,
filmogenic properties, 661 151–152
heat-reversible gels, 662 Higher alcohol production, 132–134
mechanical resistance, 663 High hydrostatic pressure
moisture content, 665 effects on, 290–294
molecular weight, 662 PATS, 295–297
opacity, 666 pulsed electric field device examples,
peptide chain configuration, 662 295, 296
preparation methods, 665 temperature effect, 294
tensile stress test, 666 High-pressure-induced effects
WVP, 665–666 bacterial and fungal spore inactivation, 326
686 Index
High-pressure-induced effects (cont.) higher alcohol production, 132–134

challenges, 336 malolactic fermentation, 138–139
food applications, 325–326 secondary fermentation, 137–138
thermal sterilization Incorporating functional ingredients
advantage, isostatic, 326 antimicrobial edible films, 642–645
dormant spore features, 327 antioxidant, 642
industrial applications, 331–333 nutraceutical ingredients carriers, 645–646
Maillard reactions, 326–327 Innovative food products, 15–16
microbial safety, 327 Intact and broken cells (IBC) hypothesis, 402,
temperature controlled spore 407–408
inactivation, 327–331 Internal mass transfer mechanisms
vegetative microorganisms and enzymes axially dispersed plug flow (ADPF),
b-amylase inactivation, 335, 336 401, 408
flow cytometry, 333 desorption–dissolution–diffusion (DDD)
lethal effects, 333 models, 401, 406, 407
polyphenol oxidase inactivation, 335 intact and broken cells (IBC) hypothesis,
pressure-temperature isorate diagram, 402, 407–408
333, 334 linear driving force (LDF) model, 403, 409
High-pressure size exclusion chromatography perfectly mixed multistages (PMMS),
(HPSEC), 82 403, 408
High pressure sterilization, food Reverchon–Sesti Ossèo (R-SO) model, 410
aseptic processing, 341 shrinking-core (SC) models, 401, 406
compression heating, 345–346 International Congress of Engineering and
pasteurization Food (ICEF), 4
dual phase destruction kinetics, 342
log-linear model, 343 J
nutritional and sensory quality, 342 Journals, 8
pulse effect (PE), 343
spore inactivation K
Clostridum botulinum studies, 348–349 Kelvin equation, 543–544
Clostridum sporogenes, 346–348 Kneaders, dough
High water content vitrification, 564–565 blades types, 31–32
Hydraulic diffusivity, 541–542 Brabender® farinograph, 32–34
deformation quantification of, 32, 33
I optimizing research, 31
Ice crystal growth inhibition, 555–556, require actions of, 30
563–564 viscoelastic fluid model, 33
Ice nucleating agents (INAs), 562–563
Ice nucleation kinetics, 555–556 L
Immobilized cell technology (ICT) Lactic acids, 643–644
alcohol fermentation, 119–120 Laminar mixing, 28–29
beer fermentation, 120 Large eddy simulation (LES), 52
bioreactor design, 127–131 Life cycle analysis (LCA), 625
carrier selection and design Light scattering analysis, polysaccharides,
cell aggregation, 127 82–83
entrapment, porous matrix, 125–127 Linear driving force (LDF) model, 403, 409
membrane barrier, 127 Liquid smoke, 645
prerequisites, 121 Lubricated squeezing flow technique, fibers,
solid carrier surfaces, 121–124 78–79, 86–87
cider and wine production, 120–121
flavor formation M
carbonyl compounds production, Malolactic fermentation (MLF), 138–139
135–137 Manufacturing plant economics
ester production, 134 characteristics, 230–232
Index 687
yogurt, 232–235 Microbial multifactorial risk assessment. See

Mass flow rate, 519 also Food safety
Mass transfer models consumer assessment, 153–156
alternative internal mass transfer ELECTRE III method, 159–161
mechanisms FAO/WHO guidelines, 150–151
axially dispersed plug flow (ADPF), market-level assessment, 152–154
401, 408 MCDA techniques, 158–159
desorption–dissolution–diffusion PROMETHEE analysis, 160–161
(DDD) models, 401, 406, 407 public health impact, 151–152
intact and broken cells (IBC) ranking risks, 149–150
hypothesis, 402, 407–408 social factor assessment, 156
linear driving force (LDF) model, Microbial polyesters, 622–623
403, 409 Microstructural factor, 428
perfectly mixed multistages (PMMS), Milk processing, CFD methods, 60
403, 408 Mixed-integer linear programming
Reverchon–Sesti Ossèo (R-SO) (MILP) model
model, 410 Mixing processes
shrinking-core (SC) models, 401, 406 devices used, 19–20
diffusion model dispersive and non-dispersive, 19
assumptions of, 398–399 3D numerical simulation
limitations of, 405–406 continuous mixer and extruders, 34–42
mathematical models summary, dough mixers and kneaders, 30–34
400–405 stirred tank reactors/batch mixers,
source-and-transfer term, 399 26–31
Master degree program, curriculum design, flow calculation, 22–23
241–242 numerical methods for simulation
Materials, edible coatings CFD method, 25
functional additives, 633–634 FEM technique, 23–24
plasticizers, 633 Galerkin method, 24–25
polysaccharides, 632 single screw mixer, 25
proteins, 632–633 twin screw mixer, 26
waxes/lipids/resins, 633 theoretical measures, 20–22
Mathematical models, 400–405 Mixograph, dough kneaders, 31, 32
Melt shear rheology, arabinoxylans, 90 Model (ternary) systems, essential oil
Membrane reactors, 130–131 fractionation
Mesh superposition technique (MST), 26 phase equilibria, 448–449
MFA. See Multifractal analysis separation factors, terpenes and aroma,
Microbial inactivation model 450, 451
activation shoulder, 316, 317 use of, 449
concavity, 316 vapor–liquid equilibria, 449
dose–response curve, 311 Model (binary) systems, solubility in CO2
mathematical models, 311–315 high-pressure equilibrium, 439–441
shapes, 310, 311 limonene, 441–443
Weibullian model, 316 plant essential oil, 444–446
Microbial kinetics predictive model implementation, 446–448
Arrhenius equation, 248 sampling problem, 443
bacteria inactivation, 248 Modified atmosphere packaging (MAP),
death-time kinetic parameters, 248–249 625–627
D- vs. Z-values, 249–250 Mohr’s circle, 500
error minimization method, 249 Mono-and bi-axial stress, 500, 501
inoculated liquid product, 249 Montmorillonite (MMT), 626–628
PEIE method, 249–250 Multi-criteria decision analysis (MCDA)
temperature dependency, 248 techniques, food safety, 158–159
688 Index
Multifractal analysis (MFA) high hydrostatic pressure

apple tissue, PSD effects on, 290–294
freezing, 603–604 PATS, 295–297
heterogeneity, 610 pulsed electric field device examples,
parameters and standard 295, 296
deviations, 608 temperature effect, 294
procedure, 604–606 microbial inactivation model
radial orientation, 604, 605 activation shoulder, 316, 317
reconstructed 3-D X-ray, 608, 609 concavity, 316
spectrum computation, 606–608 dose–response curve, 311
spectrums, 609 mathematical models, 311–315
X-ray cross-sectional image, 605 shapes, 310, 311
fractal and multifractal concepts, 599–600 Weibullian model, 316
potential parameters, 614 PEF
presliced pork hams, FSD cost, 301
ham qualities evaluation, 612 effects on, 298–299
parameters and standard deviations, electrical arcing, 299
612, 613 electrode erosion, 300
procedure, 611 extraction process, 300–301
spectrums, 612, 613 liquid food processing, 297
theory of, 600–602 treatment chambers, 299
Multiphase fluid system, CFD preservation factors, 285
categories of, 55 ultrasound
definition, 54–55 anthocyanin content, 303
Eulerian–Eulerian concept, 55–56 extraction, 302–303
Lagrangian–Eulerian Model, 56 research equipment, 304
volume of fluid model, 55 Nutraceuticals, bioseparation. See Supercritical
Multiscale modeling, 594–597 carbon dioxide
Multistage fixedbed tower (MFBT) bioreactor,
128–129 O
Oil barriers, in fried products
N frying process, 647
National and international meetings, 7 hydrocolloids, 646
Nisin, 644 methylcellulose coating effect, 648, 649
Nonlinear dynamics, 522–523 Oilseed safety assessment, Canadian FBI
Nonthermal technologies energy consumption intensity, 184
bacteria inactivation, 287, 288 greenhouse gas emission intensity,
cold plasma 181–182
advantages, 304–305 packaging use intensity indicator, 206
effect on microorganisms, 306 Operational solubility and sorption phenomena
processing conditions, 305–306 best-fit values, 453, 454
consumer trends, 286–287 plant essential oils, 455, 456
DPCD solute binding effect, 457
design, 308 solute partition coefficients, 457, 458
spore inactivation, 307 sorption isotherm models, 459, 460
thermodynamic diagram, CO2, Orange juice concentrate (OJC), preservation
306, 307 plants
food component modification, 287, 289 annual operating cost, 229–231
food processing annual operating time, 227
bacteriocins, 308–309 equipment cost, 227
plasma, 310 processes, 227, 228
ultrasound, 309 Origin, food engineering, 4–6
ultraviolet treatment, 308 Oscillations, drying zones, 528, 531
Index 689
P nutritional and sensory quality, 342

Packaging materials, renewable resources pulse effect (PE), 343
bio-plastics PATS. See Pressure assisted thermal
applications of, 624–625 sterilization
categories of, 620–621 Peclet number, 414
chemical synthesis, 623–624 Pediocins, 644
nature-made polymers, 621–623 PEF. See Pulsed electric fields
extractible agropolymers, 619–620 Petrol-based biodegradable polyesters,
microbial polyesters, 622–623 623–624
novelty of, 628–629 Phase equilibrium effects
petrol-based biodegradable polyesters, essential oil fractionation, model (ternary)
623–624 systems
polylactic acids, 623 phase equilibria, 448–449
properties and active materials separation factors, terpenes and aroma,
active antimicrobial packaging, 627 450, 451
gas permeability, 625–626 use of, 449
life cycle analysis (LCA), 625 vapor–liquid equilibria, 449
modified atmosphere packaging (MAP), operational solubility and sorption
625–627 phenomena
montmorillonite (MMT), 626–628 best-fit values, 453, 454
wheat gluten films, 626 plant essential oils, 455, 456
Packaging use intensity indicator (PUI), solute binding effect, 457
Canadian FBI solute partition coefficients, 457, 458
calculation method, 203–204 sorption isotherm models, 459, 460
levels, 199 solubility in CO2, model (binary) systems
limitations of, 210 high-pressure equilibrium, 439–441
national results and interpretation limonene, 441–443
activity sector and sub-sector, plant essential oil, 444–446
204–205 predictive model implementation,
bakeries sector, 206–207 446–448
fruit, vegetable and beverage sector, sampling problem, 443
205–206 thermodynamic and operational solubility
grain and oilseed sector, 206 comparison of, 451–453
sugar and confectionery sector, 207 equilibrium isotherms, 451
provincial results and interpretation Plant essential oils
Atlantic provinces, 209 chemistry and localization of, 394–396
British Columbia, 207–208 kinetic parameters, CO2 extraction
Ontario, 209 axial dispersion coefficient, 414,
Prairies, 208–209 421–424
response options, 210–211 compositions of, 414–418
waste management effective diffusivity, solid matrix,
environmental issues and cost of, 428–439
201–202 external mass transfer coefficient,
government support, 202 424–428
incineration, 200 extraction conditions, 411–413
reduction of, 203 molecular weights (MW) and critical
Packed bed bioreactors, 128–129 volume estimation, 414, 419–420
Particle diameter effect, 436, 437 sage essential oils composition,
Pasteurization 414, 421
CFD methods, 59 mass transfer models
high pressure, 342–343 alternative internal mass transfer
dual phase destruction kinetics, 342 mechanisms, 406–411
log-linear model, 343 diffusion model, 398–406
690 Index
Plant essential oils (cont.) Pressure assisted thermal sterilization (PATS),

phase equilibrium effects, 461 295–297
fractionation, model (ternary) systems, Process design research, 10–11
448–451 PROMETHEE analysis, microbial risk
operational solubility and sorption assessment, 160–161
phenomena, 453–460 Propionic acid, 643
solubility in CO2, model (binary) Pseudo steady state (PSS) technique, 40–42
systems, 439–448 Pulsed electric fields (PEF)
thermodynamic and operational cost, 301
solubility, 451–453 effects on, 298–299
supercritical fluid extraction (SCFE), electrical arcing, 299
393–394, 396–398 electrode erosion, 300
Pleasure, food microstructures, 586 extraction process, 300–301
Polylactic acids, 623 liquid food processing, 297
Polyvinyl alcohol (PVA), cell immobilization, treatment chambers, 299
125–127
Porous matrix entrapment, 125–127 Q
Porous media Quality and microstructure, fruit evaluation
boundary, 541–542 apple, 589, 590
capillarity and tension head, 538–539 biological food products, 589–590
capillary flow, 537–538 bitter pit symptoms, 590
flows, CFD methods, 56–57 disorders, 590, 591
food rehydration, 547 water core, 590
initial conditions, 541–542 Quality improvement
Richards equation, 541–542 oil barriers, in fried products
theory of flow frying process, 647
diffusion and relaxation mechanisms, hydrocolloids, 646
546, 547 methylcellulose coating effect, 648, 649
four-step procedure, 543–546 research, 14–15
two-phase characteristic curves, 542 storage life prolongation
water retention curve (RC), 539–541 barrier properties, 652
Pouched foods sterilization, CFD methods, 59 coating formulations, 650
Powder flowability and stress states coatings effect, 652–653
Mohr’s circle, 500 edible coatings, 649
mono-and bi-axial stress, 500, 501 fruit physiological parameters, 653
normal and shear stresses action, 498, 499 microstructure characterization, 650
prism-shaped powder element, 499 starch-based coatings application,
yield locus, 501 651–653
Powder mixes surface microbial growth, 653–654
dehydrated food products, 508–509
dextrose syrup, 511, 512 R
unconfined yield strength, 510, 511 RC. See Water retention curve
yield loci measurement, 509 Recalcitrant seeds problem, 565–566
Preservation plant economics Rehydration modeling
annual operating cost, 229–231 capillary flow, porous media, 537–538
annual operating time, 227 food trends, 535
equipment cost, 227 mathematical model
processes, 227, 228 diffusion model, 536–537
Presliced pork hams, FSD empirical and semiempirical, 536
ham qualities evaluation, 612 paradigm shift, 537–538
parameters and standard deviations, 612, physical mechanisms, 548
613 transport mechanisms, 548–549
procedure, 611 unsaturated porous media flow
spectrums, 612, 613 boundary, 541–542
Index 691
capillarity and tension head, 538–539 Spore inactivation

food rehydration, 547 Clostridum botulinum studies, 348–349
initial conditions, 541–542 Clostridum sporogenes, 346–348
Richards equation, 541–542 Spray dryer, 517–518
theory of flow, 542–547 Spray drying
water retention curve (RC), 539–541 CFD methods, 61
X-ray MCT, 549 particle stickiness problem, 479–480
Relaxation and diffusion mechanisms, powder characteristics, 480
546, 547 Sterilization
Retort equipment systems adaptive random search method, 266–270
ABRS, 251, 252 description, 267–268
AGV, 252–254 logistic curve utilization, 268–270
Retropulsive jet flow, stomach, 100–101 pedestal frequency distribution
Reynolds stress method, CFD, 52 transformation, 266–267
Rheological properties, alkali-treated bran CFD methods, 58–59
branching effects, 89–94 computer simulation, 265–266
capillary rheometry, 85–86 lethality constraint, 264–265
expansion process, 87–88 penalty functions, 265
extensional rheology, 86–87 retort function, 264
extrusion trials, 84–85 Stickiness, 480
structure and composition effects, 88–89 Stomach fluid flow
Rushton turbine impeller, 29–31 antral contraction waves (ACWs), 99–101
eddy structures, 101
S meal volume effects, 100
Seafood safety assessment, Canadian FBI modeling procedures, 102–115
energy consumption intensity (ECI), 184 parts and role of, 99
greenhouse gas emission intensity, 180 peristaltic flow analysis, 101–102
Sectional expansion index (SEI), fibers, 84–85 retropulsive jet flow, 100–101
Shrinking-core (SC) models, 401, 406 size and shape, 100
Simultaneous sterilization method, 270–272 Storage life prolongation
Sinter bridge, 505, 508 barrier properties, 652
Sintering mechanisms, 495 coating formulations, 650
Solid carrier surfaces, 121–124 coatings effect, 652–653
Solids crystallization, 561–562 edible coatings, 649
Solubility fruit physiological parameters, 653
determination and correlation microstructure characterization, 650
b-carotene, 361–362 starch-based coatings application,
caffeine, logarithmic plot, 364 651–653
Chrastil’s model, 365 surface microbial growth, 653–654
density-based correlation methods, Sugar crystallization, 568–570
363, 364 Supercritical carbon dioxide
equations of state (EoS), 363 commercialization, 382–384
off-line and on-line sampling, 362 extraction process
supercritical fluids lipid-based nutraceuticals, 367–373
caffeine, isotherms and crossover phytochemicals, 373–374
pressure, 359, 360 fractionation
cosolvent effect, 360–361 lipid-based nutraceuticals, 375–381
retrograde condensation, 360 phytochemicals, 381–382
solute properties, 359–360 protocols, 374–375
Solute binding effect, 457 phase diagram, CO2, 354
Sorbic acid, 643 physical and transport properties
Sorption isotherm models, 459, 460 density vs. pressure, 355
Spiral mixer, 31, 32 diffusion coefficients, 357
692 Index
thermal conductivity vs. pressure, 358 microbial kinetics

viscosity vs. pressure, 356 Arrhenius equation, 248
solubility bacteria inactivation, 248
determination and correlation, 361–365 death-time kinetic parameters, 248–249
supercritical fluids, 359–361 D-vs. Z-values, 249–250
Supercritical fluid extraction (SCFE), 393–394, error minimization method, 249
396–398 inoculated liquid product, 249
PEIE method, 249–250
T temperature dependency, 248
Theory of flow preservation, 247
diffusion and relaxation mechanisms, retort equipment systems
546, 547 ABRS, 251, 252
four-step procedure, 543–546 AGV, 252–254
two-phase characteristic curves, 542 Thermodynamic and operational solubility,
Thermal processing, optimization 451–453
calculation Time consolidation, 502, 503
canning plant, 277–281 Transient ice-like embryos, 559
cubic spline approximation, 262 Turbulent flows, CFD methods
global algorithms, 262 large eddy simulation (LES), 52
nonlinear programming (NLP) problem, prediction performance, 51–52
262 Reynolds stress models, 52
process time minimization, 275–277 RNG k-e model, 53–54
quality retention maximization, 275 standard k-e model, 53
VRT, 262 viscosity models, 53
canning plant schedule
mathematical model, 273–274 U
problem definition, 273 Ultrasound
simultaneous sterilization method, anthocyanin content, 303
270–272 extraction, 302–303
features, 261 research equipment, 304
food canneries, schedule, 262 Undergraduate courses, curriculum design,
MILP model, 263 240–241
sterilization Unsaturated porous media flow
adaptive random search method, boundary, 541–542
266–270 capillarity and tension head, 538–539
computer simulation, 265–266 food rehydration, 547
lethality constraint, 264–265 initial conditions, 541–542
penalty functions, 265 Richards equation, 541–542
retort function, 264 theory of flow
Thermal sterilization, high pressure diffusion and relaxation mechanisms,
advantage, isostatic, 326 546, 547
dormant spore features, 327 four-step procedure, 543–546
industrial applications, 331–333 two-phase characteristic curves, 542
Maillard reactions, 326–327 water retention curve (RC), 539–541
microbial safety, 327
temperature controlled spore inactivation, V
327–331 Variable retort temperature (VRT)
Thermal treatment, food processing, 262
flexible retortable packages Vegetative microorganisms and enzymes
design and control, 255, 256 high-pressure-induced effects
flexible and semi-rigid properties, 255 b-amylase inactivation, 335, 336
shelf-stable canned foods, 255 flow cytometry, 333
market implications, 256–258 lethal effects, 333
Index 693
polyphenol oxidase inactivation, 335 particle package collapse, 506

pressure-temperature isorate diagram, sinter bridge diameter calculation, 505, 508
333, 334 unconfined yield strength, 507
VRT. See Variable retort temperature yeast, shear locus, 506, 507
processing Water vapor permeability (WVP), 637–638,
665–666
W Weibullian model, 316
Water consumption, Canadian FBI Wheat gluten films, 626
indicators, intake and discharge Whipped cream, 583
calculation method, 190–191 WVP. See Water vapor permeability
limitations, 191
national results and interpretation, X
193–195 X-ray
plant grouping parameters, 192 apple tissue, PSD, 605, 608, 609
provincial results and interpretation, computed tomography (CT), 591–593
196–198 MCT, 549
response options, 198–199
intake volumes, 188–189 Y
usages, 188 Yoghurt
wastewater flow in, 189 manufacturing plants
Water core, 590 annual operating cost, 232–234
Water retention curve (RC), 539–541 annual sales income, 234
Water-soluble powders, dextrose syrup cumulative cash flow, 234–235
glass transition temperatures, sintering, material and energy balances, 232
504, 505 packaging, 232
moisture content and visual appearance, process, 233
503, 504 processing, CFD methods, 60

AGUILERA y Col - Food Engineering Interfaces PDF

Uploaded by

Copyright:

Available Formats

AGUILERA y Col - Food Engineering Interfaces PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AGUILERA y Col - Food Engineering Interfaces PDF

Uploaded by

Copyright:

Available Formats

What are some of the main topics covered in the book?

What are some of the main topics covered in the book?

What are some of the editing and publishing details of the book?

What are some of the editing and publishing details of the book?

Food Engineering Series

For other titles published in this series, go to

Cánovas Ricardo Simpson Jorge

Welti-Chanes Daniela Bermúdez-Aguirre

Food Engineering Interfaces

Dr. Daniela Bermúdez-Aguirre

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

Part I Selected Topics in Food Engineering

1 The Beginning, Current, and Future of Food Engineering:

2 Advances in 3D Numerical Simulation of Viscous

3 CFD: An Innovative and Effective Design Tool

4 Incorporation of Fibers in Foods: A Food Engineering Challenge . . 69

5 Gastric Digestion of Foods: Mathematical Modeling of Flow

6 State of the Art in Immobilized/Encapsulated Cell Technology

7 Multifactorial Assessment of Microbial Risks in Foods: Merging

8 Development of Eco-efficiency Indicators to Assess

9 Food Process Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

10 Systemic Approach to Curriculum Design and Development . . . . . . . 237

Part II Advances in Food Process Engineering

11 Innovations in Thermal Treatment of Food . . . . . . . . . . . . . . . . . . . . . . . . . . 247

12 Optimization of Food Thermal Processing: Sterilization

13 Recent Advances in Emerging Nonthermal Technologies . . . . . . . . . . . 285

14 High-Pressure-Induced Effects on Bacterial Spores, Vegetative

15 High Pressure Sterilization of Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341

16 Bioseparation of Nutraceuticals Using Supercritical

17 Mass Transfer and Equilibrium Parameters on High-Pressure

Part III Water Management in Food

18 Glass Transitions: Opportunities and Challenges . . . . . . . . . . . . . . . . . . . . 473

19 Caking of Water-Soluble Amorphous and Crystalline

20 Effective Drying Zones and Nonlinear Dynamics

21 Rehydration Modeling of Food Particulates Utilizing

22 Responses of Living Organisms to Freezing and Drying: Potential

Part IV Food Microstructure

23 Food Microstructures for Health, Well-being, and Pleasure . . . . . . . . 577

24 Fruit Microstructure Evaluation Using Synchrotron X-Ray

25 Multifractal Characterization of Apple Pore and Ham

Part V Food Packaging

26 New Packaging Materials Based on Renewable Resources:

27 Edible Coatings to Improve Food Quality and Safety . . . . . . . . . . . . . . . 631

28 Physical Properties of Edible Gelatin Films Colored

Alik Abakarov Departamento de Ingenierı́a Quı́mica y Ambiental, Universidad

José Miguel Aguilera Department of Chemical and Bioprocess Engineering,

Liliana Alamilla-Beltrán Departamento de Graduados e Investigación en

Gustavo V. Barbosa-Cánovas Center for Nonthermal Processing of Food,

Daniela Bermúdez-Aguirre Center for Nonthermal Processing of Food,

Gerd Brunner Thermische Verfahrenstechnik, Technische Universität Hamburg-

Branko Bugarski Department of Chemical Engineering, University of Belgrade,

Osvaldo Campanella Agricultural and Biological Engineering Department, Purdue

Rosemary A. Carvalho Department of Food Engineering, University of São

Jose Jorge Chanona-Pérez Departamento de Graduados e Investigación en

Peter Cloetens European Synchrotron Radiation Facility, 6 Rue Jules Horowitz,

Valerie Davidson School of Engineering, University of Guelph, N1G 2W1

Juan C. de la Fuente Departamento de Procesos Quı́micos, Biotecnológicos y

Adriana Delgado FRCFT Group, Biosystems Engineering, UCD Agriculture &