Bentonite Clay

BENTONITE
CL AY
Environmental Properties
and Applications
Roland Pusch
BENTONITE
CL AY
and Applications
Deposition hole with
Backfilled drift canister surrounded
by “buffer” of dense
smectite-rich clay
Fracture to be
considered
Canister
Function of bentonite “buffer” clay for isolating highly radioactive spent reactor fuel in cop-
per canisters according to a 35 year old concept of geological disposal (KBS-3V).
BENTONITE
CL AY
and Applications
Roland Pusch
Boca Raton London New York
CRC Press is an imprint of the

Taylor & Francis Group, an informa business
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2015 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works

Version Date: 20150427
International Standard Book Number-13: 978-1-4822-4344-4 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information stor-
age or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copy-
right.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222
Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that pro-
vides licenses and registration for a variety of users. For organizations that have been granted a photo-
copy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Dedication
I would like to dedicate this book to my beloved

wife, Greta, my supporter and caretaker
Contents
Preface...................................................................................................................... xv
Acknowledgments....................................................................................................xix
Author......................................................................................................................xxi
Chapter 1 Introduction........................................................................................... 1
1.1 Main Issues.................................................................................1
1.2 Smectite Clay—The Muddiest Type of Soft Clay...................... 1
1.3 Stress–Strain Problems.............................................................. 2
1.4 Preparation of Smectite Clay for Sealing Purposes...................4
1.5 Quality Issues.............................................................................5
1.6 Performance Tests...................................................................... 7
References............................................................................................. 8
Chapter 2 Origin and Occurrence of Smectite Clays: Bentonites......................... 9

2.1 Origin of Smectite Clays............................................................9
2.2 Where Are All These Famous and Magic Clays?......................9
2.2.1 North America...............................................................9
2.2.1.1 The United States........................................ 10
2.2.1.2 Canada......................................................... 11
2.2.2 South America............................................................. 11
2.2.3 Central America.......................................................... 11
2.2.3.1 Mexico......................................................... 11
2.2.4 Africa.......................................................................... 12
2.2.5 Middle East................................................................. 12
2.2.6 Asia.............................................................................. 12
2.2.6.1 China............................................................ 12
2.2.6.2 India............................................................. 13
2.2.6.3 Japan............................................................ 13
2.2.7 Europe......................................................................... 14
2.2.7.1 Germany...................................................... 14
2.2.7.2 Denmark...................................................... 15
2.2.7.3 Italy.............................................................. 15
2.2.7.4 Spain............................................................ 16
2.2.7.5 Czech Republic............................................ 18
2.2.7.6 Greece.......................................................... 18
2.3 Potential Smectite Resources................................................... 18
2.4 Are New Smectites Being Formed Today?.............................. 19
2.5 Quality of Natural Smectite Clays for Practical Use................ 19
2.6 Conclusion................................................................................20
References...........................................................................................20
vii
viii Contents
Chapter 3 Nature of Smectites............................................................................. 23

3.1 Basics........................................................................................ 23
3.2 Smectite Family........................................................................ 23
3.3 Crystal Constitution of Smectites.............................................25
3.4 Chemical Composition of Natural Smectite Clays...................26
3.4.1 Smectite Component...................................................26
3.5 Mineral Composition of Natural Smectite Clays.....................28
3.5.1 Smectite-Rich Clays....................................................28
3.5.2 Clays with Moderate and Low Smectite Contents......28
3.6 Role of Clay Particle Charge.................................................... 29
3.6.1 Basics and Practicalities.............................................. 29
3.6.2 CEC, Anion Exchange Capacity, and Specific
Surface Area................................................................ 31
3.6.3 Mechanisms in Cation Exchange................................ 32
3.6.4 Role of Anions............................................................. 32
3.6.5 Phosphorus and Nitrogen............................................ 33
3.6.6 Sulfur........................................................................... 33
3.6.7 Organic Content..........................................................34
3.6.8 Special Role of Organic Elements...............................34
3.6.8.1 Bacteria........................................................34
3.6.8.2 Identification of Smectite Minerals.............34
3.7 Conclusion................................................................................ 36
References........................................................................................... 36
Chapter 4 Clay–Fluid–Gas Systems.................................................................... 37

4.1 Overview.................................................................................. 37
4.2 Physicochemical Background................................................... 38
4.2.1 Energy Considerations, Soil–Water Potentials............ 38
4.2.2 Hydration of Smectite Clay.........................................40
4.3 Hydraulic Conductivity............................................................ 42
4.3.1 Definition of Hydraulic Conductivity.......................... 42
4.3.2 Determination of the Hydraulic Conductivity
in the Laboratory......................................................... 43
4.3.3 Microstructural Implications.......................................44
4.3.4 Role of Hydraulic Gradients........................................46
4.3.4.1 Test Principles and Theory..........................46
4.3.4.2 Experimental Proof of the Impact of
the Hydraulic Gradient................................ 47
4.3.4.3 Piping and Erosion....................................... 51
4.3.4.4 Experimental Evidence................................ 52
4.3.4.5 Impact of Hydraulic Gradients on
the Permeation of Water-Saturated
Clay Seals.................................................... 55
4.3.5 Impact of Saltwater on the Hydraulic Conductivity......59
Contents ix
4.3.6 Criteria.........................................................................60
4.3.7 Impact of Smectite Content on the Hydraulic
Conductivity of Mixed Soils.......................................64
4.3.7.1 Mixed Clay and Coarser Material...............64
4.3.7.2 Natural Soils................................................66
4.3.7.3 Hydraulic Conductivity of
Unsaturated Clay.........................................66
4.4 Gas Conductivity...................................................................... 67
4.4.1 Basics........................................................................... 67
4.4.2 Determination of Gas Transport Capacity.................. 68
4.4.2.1 Background.................................................. 68
4.4.2.2 Laboratory Technique.................................. 71
4.4.3 Modeling of Gas Penetration Using Capillary
Analogs........................................................................ 71
4.5 Ion Diffusion............................................................................ 71
4.5.1 Definitions................................................................... 71
4.5.2 Mechanisms and Basic Relationships......................... 72
4.5.3 Role of the Microstructure.......................................... 75
4.5.4 Impact of the Microstructural Constitution
on Ion Diffusion.......................................................... 76
4.5.5 Test Principles and Theory.......................................... 78
4.5.5.1 Technique..................................................... 78
4.5.5.2 Role of Electrical Double Layers for
Ion Diffusion................................................80
4.6 Colloid Transport......................................................................80
4.7 Microbiological Filtering..........................................................80
4.8 Heat Transport—Thermal Properties....................................... 81
4.8.1 Overview..................................................................... 81
4.8.2 Influence of Water Content.......................................... 81
4.8.3 Influence of Smectite Content..................................... 81
4.8.4 Influence of Stress and Temperature........................... 82
4.9 Couplings, Dependencies, and Interdependencies................... 82
References...........................................................................................84
Chapter 5 Physical Performance of Smectite Clay Seals: Prediction

and Reality.......................................................................................... 87
5.1 Introduction.............................................................................. 87
5.2 Application of Concepts of Soil Mechanics to Smectite Clay...... 87
5.2.1 Effective Stress Concept............................................. 87
5.2.2 Role of the Effective Stress for the Physical
Stability of Clay Seals................................................. 89
5.2.3 Stress–Strain Definitions and Parameters...................90
5.2.3.1 Compression Modulus K............................. 91
5.2.3.2 Shear Stress Modulus G.............................. 91
5.2.3.3 Oedometer Modulus M................................ 91
x Contents
5.2.3.4 Compression Properties...............................92

5.2.3.5 Time Dependence........................................92
5.2.3.6 Shear Strain................................................. 93
5.3 Role and Mechanisms of Consolidation and Creep.................94
5.3.1 Cases............................................................................94
5.3.2 Consolidation.............................................................. 95
5.3.3 Shear Strain by Creep................................................. 95
5.4 Fundamentals of Thermal Conditions and Performance....... 101
5.4.1 Definitions and Parameters....................................... 101
5.4.2 Influence of Temperature.......................................... 102
5.5 Evolution of Clay Seals.......................................................... 103
5.5.1 Hydration................................................................... 103
5.5.2 Forslind–Jacobsson Model........................................ 105
5.6 Clay Microstructure and Its Role for the Stress–Strain
Behavior of Smectite Clays.................................................... 107
5.6.1 Scale Dependence..................................................... 107
5.6.2 Impact of Heating...................................................... 109
5.6.3 Microstructural Modeling of Smectite-Rich Clay...... 112
5.7 Effect of Combined Wetting and Heating of Clay
Seals—The Buffer Case......................................................... 114
5.7.1 Practical Cases.......................................................... 114
5.7.2 Hydration of Smectite-Rich Buffer Clay under
Temperature Gradients.............................................. 114
5.7.3 Physical Processes Taking Place in Buffer Clay....... 118
5.7.4 Evolution of the Buffer—Temperature...................... 118
5.7.5 Evolution of the Buffer—Expansion and
Consolidation under Hot Conditions.............................119
5.7.6 Modeling of the Hydration of Buffer Clay................ 123
5.7.7 Thermo–Hydro–Mechanical–Chemical
Processes in Buffer Clay—Salt Accumulation......... 124
5.8 Concepts and Techniques for Isolating Moderately
Hazardous Waste.................................................................... 128
5.8.1 Landfills.................................................................... 128
5.8.2 Underground Disposal in Shallow Repositories....... 133
5.8.3 Underground Disposal in Abandoned Mines............ 137
5.9 Concepts for Isolating Highly Radioactive Waste.................. 142
5.9.1 Medium-Deep Repositories....................................... 142
5.9.1.1 The KBS-3V Concept................................ 142
5.9.2 Steep Holes with Two or Several Canisters
(Case A)..................................................................... 147
5.9.3 Big Cavern with Numerous Canisters (Case B-1)..... 147
5.9.4 Tunnels or Drifts with Large Clay-Isolated
Concrete Containers (Case B-2)................................ 147
5.9.5 Inclined Holes (Case C)............................................ 148
5.9.6 Very Long Holes (Case D)........................................ 150
5.10 Very Deep Holes..................................................................... 151
Contents xi
5.11 Correlation of Hydraulic and Mechanical Performances

of Clay Seals........................................................................... 153
5.11.1 Piping and Erosion of Clay Seals............................ 153
5.11.2 VDH Concept.......................................................... 155
5.11.2.1 Clay Mud............................................... 155
5.11.2.2 Clay Block Seals.................................... 155
5.11.2.3 Interaction of Mud and Blocks.............. 156
5.11.3 Usefulness of Rock for Hosting Repositories...........156
5.12 Concluding Remarks.............................................................. 157
References......................................................................................... 160
Chapter 6 Chemical Processes Involved in and Longevity of Smectite Buffer......165

6.1 Chemical Stability of Smectite Clay for Waste Isolation....... 165
6.1.1 Our Starting Point..................................................... 165
6.1.2 Natural Analogs........................................................ 166
6.1.2.1 Conversion of Smectite to
Nonexpanding Minerals............................ 166
6.1.2.2 Kinnekulle—A Comforting Case?............ 168
6.2 Experimental Evidence.......................................................... 168
6.2.1 Overview................................................................... 168
6.2.2 Stripa Project Laboratory Study................................ 169
6.2.2.1 Test Program and Techniques.................... 170
6.2.2.2 Summary of Results.................................. 171
6.2.3 SKB-ANDRA Study................................................. 174
6.2.4 RMN Study............................................................... 176
6.2.5 Swedish–Czech–Chinese University Study.............. 181
6.2.5.1 Montmorillonite-Dominated MX-80......... 184
6.2.5.2 Saponite-Dominated Clay......................... 184
6.2.5.3 Mixed-Layer I/S Clay................................ 186
6.2.6 SKB Field Tests......................................................... 186
6.2.7 Matter of Stiffening................................................... 190
6.2.8 Interaction of Smectite Clay and Other
Components............................................................... 195
6.2.8.1 Smectite Clay Contacting Copper Metal....196
6.2.8.2 Smectite Clay Contacting Iron and Steel....197
6.2.8.3 I/S Mixed-Layer Smectite Clay
Contacting Concrete.................................. 198
6.2.8.4 Montmorillonite-Rich Clay Contacting
Low-pH Concrete...................................... 201
6.3 Summary Respecting Smectite Chemistry and
Mineralogy..........................................................................206
6.3.1 Overview...................................................................206
6.3.2 Geochemical Modeling.............................................207
6.4 Concluding Remarks..............................................................208
References.........................................................................................208
xii Contents
Chapter 7 Processing of Clays and Preparation of Seals................................... 211

7.1 Overview................................................................................ 211
7.2 Block Preparation................................................................... 212
7.2.1 Raw Material............................................................. 212
7.2.2 Achievable Block Density......................................... 212
7.2.3 Preparation of Dense Blocks of Smectite Clay......... 214
7.2.4 Blocks Prepared by Uniaxial Compression............... 216
7.2.5 Blocks Prepared by Isotropic Compression.............. 218
7.2.6 Microstructural Constitution of Compacted
Smectite Clay............................................................ 218
7.3 Clay Liners, Materials, and Principles of Construction.........224
7.3.1 Principles of Design and Construction......................224
7.3.2 Criteria.......................................................................224
7.3.3 Principle of Placement and Compaction................... 226
7.3.4 Microstructural Constitution of Compacted
Smectite Clay Liners................................................. 233
7.3.5 Microstructural Modeling of Smectite Clay............. 233
7.4 Clays for Rock Grouting......................................................... 235
7.4.1 Use of Grouts with respect to Their Physical
Stability..................................................................... 235
7.4.1.1 Argillaceous Cement Grouts..................... 235
7.4.1.2 Role of Palygorskite................................... 236
7.4.2 Penetrability and Sealing Function
of Clay-Based Grouts................................................ 238
7.5 General Aspects on Selection and Use of Smectite Clays
for Waste Isolation.................................................................. 239
References......................................................................................... 239
Chapter 8 Environmental Behavior................................................................... 241

8.1 Waste Isolation by Use of Clay............................................... 241
8.2 VDH—Ostrich Philosophy or a Serious Alternative
for HLW Disposal?................................................................. 242
8.2.1 Background............................................................... 242
8.2.2 Rock Conditions........................................................244
8.2.3 Temperature Conditions............................................246
8.2.4 Conditions Respecting the Chemical
Composition of the Groundwater..............................246
8.3 Sealing Components...............................................................246
8.3.1 Overview...................................................................246
8.3.2 Waste Canisters......................................................... 247
8.3.2.1 Steel Canisters...........................................248
8.3.2.2 Mud Performance...................................... 251
8.3.2.3 Casing Performance................................... 251
8.3.2.4 Supercontainer Performance..................... 253
Contents xiii
8.3.2.5 Canister Performance................................ 253

8.3.2.6 Buffer Clay Performance........................... 253
8.3.3 Concrete.................................................................... 254
8.3.4 Practical Aspects—Placeability................................ 255
8.3.5 Long-Term Function of Clay Components
in the Sealed and Deployment Zones........................ 258
8.3.6 Long-Term Function of Concrete Seals..................... 259
8.3.7 Impact of Gamma Radiation..................................... 259
8.3.8 Unexpected Events.................................................... 259
8.3.9 Environmental Impact...............................................260
8.4 SARC—The Poor Man’s Solution.......................................... 261
8.4.1 Background............................................................... 261
8.4.2 Steps in Siting of SARC............................................ 263
8.4.3 Constitution of a SARC Repository for HLW........... 267
8.4.3.1 General Conditions.................................... 267
8.4.3.2 Bottom Bed................................................ 269
8.4.3.3 Containers and Canisters........................... 272
8.4.3.4 Clay Block Liner........................................ 272
8.4.4 General Scenario of a SARC Repository.................. 272
8.4.4.1 Function of SARC...................................... 272
8.4.4.2 Bottom Bed................................................ 273
8.4.4.3 Clay Block Liner........................................ 273
8.4.4.4 Containers and Canisters........................... 273
8.4.5 Environmental Impact............................................... 274
8.5 Borehole Sealing.................................................................... 274
References......................................................................................... 278
Chapter 9 Pharmacology and Cosmetics........................................................... 281

9.1 Origin of Life.......................................................................... 281
9.2 Interaction of Smectite Clay and Organic Molecules............. 282
9.3 Interaction of Clays and Organics in Medical Treatment...... 288
9.3.1 Issues Considered...................................................... 288
9.3.2 Clays in Natural Medicine........................................ 288
9.3.3 Clay for Healing Wounds.......................................... 289
9.3.4 Extraction of Hazardous Elements Poisoning
the Human Body....................................................... 291
9.3.4.1 Principle..................................................... 291
9.3.4.2 Preference of Clay Minerals...................... 292
9.3.4.3 Interaction of Clay Minerals and Cells
in the Gastric System................................. 293
9.3.4.4 Radionuclides............................................ 295
9.3.5 Potential to Cure Cancer........................................... 296
9.3.6 Summing Up on Clays in Modern Medicine............ 297
9.4 Sun Protection........................................................................ 298
xiv Contents
9.4.1 Pilot Study of the Performance of Organic-

and Clay-Based Sun Creams..................................... 298
9.4.1.1 Tested Creams............................................ 298
9.4.1.2 Testing........................................................ 298
9.4.1.3 Results........................................................ 299
9.4.1.4 Conclusions................................................ 301
9.4.2 Comprehensive Studies.............................................302
9.4.2.1 Tested Creams............................................302
9.4.2.2 Testing........................................................302
9.4.2.3 Results........................................................ 303
9.5 Clay in Cosmetology..............................................................304
9.5.1 Background...............................................................304
9.5.2 Interaction of Clay Minerals and Epidermis.............304
9.5.3 Rheology................................................................... 305
9.5.4 Clay Candidates........................................................ 305
9.5.5 Density and Consistency of Smectite Creams
with and without Organic Liquid.............................. 305
9.6 Summary of Fundamental Properties of Smectitic Creams
on Skin......................................................................................307
References.........................................................................................308
Appendix: Symbols and Definitions.................................................................... 311

Preface
Have you tried to walk across a farmer’s field in Montmorillon in France on a rainy
October day? Marcel and Bertrand, his father, tried to do so in the fall of 1881 with
fatal consequences. We will return to this terrible true story in the first chapter of this
book for explaining how the soil at this place made it happen, continuing throughout
the rest of the book by explaining how the dreadful soil, in fact, also has extremely
beneficial properties in other contexts, like isolation of very hazardous waste. The
place gave name to the mineral montmorillonite, which belongs to a group of clay
minerals called smectites, which have tremendous ability to adsorb water; this prop-
erty transforms the clay from being stiff in dry weather to becoming soft sticking
mud after long periods of rain, from which you cannot pull your feet and legs out.
The harder you try to get them up, the deeper you work yourself down because
intense movement reduces the strength of the clay and makes it fluid. But a minute’s
rest for breathing makes the thixotropic* material stiffen and lock you up, so that you
cannot start moving again (Figure I.1). French farmers in this region and in numer-
ous other areas with this type of clay cannot complete cultivation or even inspect
their land in the rainy season. Much the same kind of clay is found on the island of
Sardinia, Italy, where such clay is in abundance. People in the churchyard visiting the
grave of their deceased relatives are never sure of placing flowers on the right grave
since the graves move about in all directions with the seasonal variation in water
content of the ground. This preface concludes with a comment relating this property
of montmorillonite to a perspective on life and death.
American geologists, such as Hewett in 1917, concluded that the highly colloidal,
plastic clay found near Fort Benton, Wyoming, had been formed by in situ altera-
tion of volcanic ash, and they named the soil bentonite. It turned out to be a close
FIGURE I.1 Clay mud of Wyoming bentonite with a density of 1100 kg/m3 prepared by
mixing smectite clay powder with distilled water and left to rest for 2 minutes. The shear
strength is just enough to avoid slope failure of the gel.
* Thixotropy is defined as “unlimited reversible, isothermal sol/gel transformation,” sols representing

liquid colloidal particle systems, and gels being colloidal materials with some shear resistance.
xv
xvi Preface
relative of the sticky mud in France, for which one could use this term as well.
The term, however, is misused by some in the mineral production branch and utiliz-
ers, for which reason we will instead use the term smectite, which is the scientific
term for clays containing micaceous expandable clay minerals regardless of whether
they originate from weathering of minerals or from volcanic ash. The sticky French
clay gave birth to the name of the most common smectite mineral, namely, montmo-
rillonite, but there are several others with associated designations, such as saponite,
hectorite, and nontronite. Here, we will reserve the term bentonite for smectite clays
that really emanate from volcanic ash.
Bentonite is a highly colloidal ductile clay formed by alteration of volcanic ash in
situ. It is widely distributed in the Cretaceous and Tertiary sediments, as in the Gulf
area and the states of the High Plains in North America, in central and southern Europe,
North Africa, the Middle East, and East Asia. Following the nestor of clay mineralogy
(Grim 1953), there are reasons to believe that alteration of volcanic glass to the smectite
species montmorillonite took place soon after or in conjunction with the fall of ash in
the sea or estuaries. The clays formed sediments that got denser with time upon being
loaded with additional sediments. In fact, enormous amounts of smectite-rich clay are
still being created when large volcanic eruptions take place; the successively densified
clay sediments get the typical physical properties of smectites, that is, tightness, ductil-
ity, and expandability. The first-mentioned character makes such clay layers effective
sealants of underlying oil- and gas-bearing porous rock and is, hence, of extraordinary
importance to the gas and oil industry. The last mentioned is not appreciated by anyone
since it makes excavated slopes with even a small inclination unstable and makes roads
and buildings on them undergo never-ending seasonal upheaval and settlement.
For our purpose, it is the tightness of smectite clays that is of importance since it
makes them very useful for sealing off hazardous waste from groundwater, as I will
discuss throughout the book. Smectite clays have, in fact, additional valuable proper-
ties such as a formidable capacity of sorbing positively charged hazardous ions and
molecules and an ability to retard diffusive transport of such elements with positive
charge. This means that sufficiently thick clay layers can serve as barriers to migra-
tion of many hazardous ion species and totally eliminate the risk of contamination of
groundwater in rock hosting repositories with highly radioactive waste. Quite differ-
ent but equally important applications are in the fields of dermatology, internal medi-
cine, and cosmetics. The effect of clay-based creams smeared on the skin depends on
the transmission of inorganic ions and organic molecules in and out through them,
which means that their microstructure and the types of transport mechanisms, diffu-
sion, and flow are determinants of their efficiency. These matters will become pro-
gressively more important in a world population that gets more and more wizenfaced.
There is every reason to look back in history and consider the enormous impor-
tance of smectite in life on earth. According to various researchers, life was born
on a smectite crystal formed from volcanic glass in a huge magma eruption some
thousand million years ago (Anderson 1989). High up in the atmosphere, where
ultraviolet radiation and heat provided energy, the little electrically charged
crystals sorbed carbons and hydrogens and acted as catalysts in forming amino
acid molecules regularly ordered on the basal surfaces of hydroxyls and oxygens.
The numerous crystals with their organic coatings interacted and formed more
Preface xvii
complex organic molecules, the driving force—the life spark—being provided

by the radiation. Very simple organic forms with just one cell were created but
competition between them led to successively larger organic objects with raised
strive for hegemony. The human being, presently considered as the ultimate living
species, is returned to the complex of clay crystallites while the funeral band plays
“W’re marching on” at the end of the personal life cycle. A good example of the
role of smectites in this context, or rather after it, is the behavior of the blessed soil
in Sardinian churchyards causing anger and upset among mourning Italians who
cannot find the grave they were to put flowers on. I have seen it myself.
In my mind, the mineralogy of clays in general and of smectites in particular is
the most fascinating subject that I can think of both because of their roles for the life
cycle and because they represent the most challenging conditions for geotechnical
engineers, who are responsible for earth dam construction, foundation of roads and
buildings, and long-term isolation of chemical and radioactive waste. We will deal
with a number of these issues in this book.
A final comment: the reader will find that all the chapters are loaded with references
to articles and books that bear my name, some stemming from late-twentieth century
and some from a more recent year. Reference to rather many early reports is deliber-
ate: much of the fundamental knowledge concerning clay properties stem from the
epoch when laboratory work dominated over mathematical modeling and made up the
foundation of our present understanding on the subject, although many recent studies
neglect reference to classical applied geology, which, at least occasionally, represent
“reinventions of the wheel.” These were the times when the basis of soil mechanics
was established as a science and when use of advanced techniques, such as high-volt-
age electron microscopy and nuclear magnetic resonance for working out conceptual
microstructural models of clays, became available. At that time, the development of
smectite-oriented geotechnical science was paralleled by practical engineering that
made it possible to utilize the physicochemical knowledge for design and construction
of advanced building foundations and for constructing deep channels in clays with
very low slope stability factors. Some of this will be mentioned and referred to in the
book that we will go through together. More recent literature has been cited with some
continence, though important contributions are certainly included.
REFERENCES
Anderson, D.M. 1989. Role of Interfacial Water and Water Films in the Origin of Life. Texas
A&M University, College Station, Texas.
Grim, R.E. 1953. Clay Mineralogy. McGraw-Hill, London, UK.
Hewett, D.F. 1917. The origin of bentonite. Journal of the Washington Academy of Sciences,
7:196–198.
Roland Pusch
Luleå University of Technology
and
Drawrite Co. Lund, Sweden
Acknowledgments
I express my deep-felt gratitude for Professors Raymond Yong, emeritus from McGill
University, Canada; Masashi Nakano, emeritus from Tokyo University, Japan; and
Sven Knutsson, present head of the Geotechnical Division at Luleå University of
Technology, Luleå, Sweden, for the inspiring and fruitful scientific cooperation and
interaction over the years, nearly 40, in fact.
Furthermore, I am greatly indebted to my Swedish friends and colleagues
Dr. Lennart Börgesson, Clay Technology AB, Lund, Sweden, and Christer Svemar,
SKB, Stockholm, Sweden, for helpful discussions and support.
xix
Author
Roland Pusch was born on December 4, 1928,
in Stockholm, Sweden. His academic career led
to a PhD degree in soil mechanics in 1962 (Royal
Technical University, KTH) and a PhD degree in
geology in 1970 (Stockholm University) and to
an appointment as chief engineer and head of the
research division of the Swedish Geotechnical Institute
from 1963 to 1967. Dr. Pusch was an associate profes-
sor in soil mechanics and foundation engineering at
Chalmers University of Technology (1967–1974) and
later a professor at the Technical University at Luleå
(1974–1982).
Dr. Pusch’s entire academic career was p aralleled
by work in major consulting companies (Jacobson & Widmark, Blomgren & Co,
Allmänna Ingenjörsbyrån). His major tasks as a consultant were to serve as a respon-
sible chief engineer for the design of Sweden’s biggest channel construction project
(Lindökanalen, Norrköping), from 1957 to 1959, and to be responsible for field and
lab investigations and design of the foundation for most of the bridges along the main
road E22 in Blekinge county between 1970 and 1974.
Dr. Pusch’s engagement in research and development of technology for disposal
of radioactive waste started in 1976 and continues to this day. His major tasks
implied responsibility for design, testing, and construction and follow-up of the per-
formance of the engineering barriers of the SFR repository at Forsmark (1980–2000)
and responsibility for the selection and testing of the soil-based engineered barriers
of SKB’s concepts such as KBS-1, -2, and -3. From 1982 until 1992, his dominant
work was as the principal investigator of engineered barriers in the international
Stripa Project’s three phases, which comprised full-scale testing of clay materials
for embedding HLW radioactive canisters and comprehensive investigation of the
interaction of rock and clay materials.
From 2000 until 2006, he was a designer in a near-surface repository of
Lithuanian LLW, comprising geological/geotechnical surveying for site selection,
calculation of slope stability, and evolution of water saturation and percolation, end-
ing up in a concept that was approved by International Atomic Energy Agency in
2007. He was also a responsible consultant for the Swedish waste-handling company
Ragn-Sell AB in designing landfills consisting of clay-isolated incinerated organic
waste.
Dr. Pusch was also a coordinator and took active part in three projects of the
European Union: Microstructure and chemical parameters of bentonite as deter-
minants of waste isolation efficiency (1995–1998), Low risk disposal technology
(2000–2003), and Cluster repository project (CROP) (2000–2003), concerning
disposal of HLW and long-lived species in LLW and MLW.
xxi
xxii Author
Between 1993 and 2010, Dr. Pusch authored five books: Waste Disposal in Rock
and Rock Mechanics on a Geological Base, both for Elsevier Publishing Company;
Microstructure of Smectite Clays and Engineering Performance (together with
R. N. Yong), for Taylor & Francis; Geological Storage of Radioactive Waste, for
Springer-Verlag; and High-Level Radioactive Waste Disposal, for WIT Press. He
has also published about 200 technical and scientific papers.
1 Introduction
This chapter describes typical geotechnical problems that need to be solved in

conjunction with design and construction of disposals of hazardous waste in

landfills and underground, using smectite clay as the main engineered barrier.
1.1 MAIN ISSUES

We have all realized that the way in which the rate of garbage production increases
will drown us in various rest products unless more intensive recycling and utiliza-
tion of waste takes place globally. Still, there are and will be significant quantities of
hazardous waste emanating from ordinary industrial work, energy production, and
agricultural and medical activities that turn out to be difficult and too expensive to
reprocess. Radioactive waste belongs to this category and challenges the scientific
society’s capacity to realize ways of safe disposal and storage. It is not the worst rest
product, however, since it will be successively less hazardous with time in contrast
to certain chemical wastes such as arsenic, cadmium, and mercury, which all remain
poisonous forever.
In this book, we will examine some often-discussed methods for isolating such
wastes using clay as the major engineered barrier, considering possibilities and limi-
tations. The literature provides numerous works on this issue, most of them deal-
ing with specific matters of largely academic interest, whereas we will be more
pragmatic here and pay particular attention to the construction and performance of
clay seals based on the experience gained in the last few decades. Starting with an
elementary overview of the mechanisms that control their physical behavior, we will
consider construction of disposals and storages, which requires some basic techni-
cal knowledge of soil and rock sciences. Geotechnical issues are in focus since civil
engineering determines what is possible and what is not in planning and constructing
repositories and disposal sites of various types. Swelling clays, primarily smectites,
and their potential to behave differently under different conditions, are in focus since
they are famous for their sealing ability, but also for causing great practical difficul-
ties in farming, road building, and underground construction.
1.2 SMECTITE CLAY—THE MUDDIEST TYPE OF SOFT CLAY

Have you tried to walk across a farmer’s field in Montmorillon in France on a rainy
October day? As stated in the Preface, father and son, Marcel and Bertrand tried to
do so in the rainy season of 1881 and they suffered a terrible death, being stuck in
the clay. What makes soft smectite clays, especially those dominated by the m ineral
species montmorillonite, behave like this is primarily the very low m echanical
strength, the rapid loss/regain of the strength at loading/unloading—a matter of
1
2 Bentonite Clay
thixotropy*—and the very low hydraulic conductivity, that is, the rate of water
permeation under a hydraulic gradient. The latter property means that stepping on
such clay increases the porewater pressure so much that the friction between the
clay particles is lost, causing your feet to sink, while an attempt to pull them up is
hindered by the momentaneous strength regain at unloading and by the difficulty of
pressurized water to be dissipated. In soil mechanics these matters are fundamental
and they determine the physical behavior of smectitic seals, like borehole plugs, and
of clay liners covering landfills of hazardous waste, as well as of clay “buffers” in
repositories where highly radioactive waste is disposed.
1.3 STRESS–STRAIN PROBLEMS

It has already been said that smectite-rich clay, such as bentonites, can cause prob-
lems for agriculture and road traffic because of softening on uptake of water, and it
can be troublesome in a number of situations related to the foundation of buildings
and isolation of hazardous waste, as illustrated by Figure 1.1. The three examples
are repeatedly considered in the book because they represent major geotechnical
problems that are particularly important in environmental protection. The left pic-
ture in Figure 1.1 shows a heavy container (canister) with spent nuclear reactor fuel
surrounded by smectite clay that may fail by insufficient bearing capacity or fracture
at desiccation caused by insufficient access to water from the rock. The pictures to
the right indicate how a smectite-based cover of a landfill of toxic chemical waste
Canister in buffer clay Slope in clay
Backfill
Cover for drainage
and balancing
swelling pressure
Clay liner
Groundwater
level
High shear stresses
FIGURE 1.1 Three examples of geotechnical stability problems that can lead to contami-
nation of the ground and groundwater in a waste disposal area. The clay in the rectangles is
exposed to mechanical stresses that can cause failure (Pusch and Yong 2006; Pusch et al. 2013).
* Thixotropy is defined as “unlimited reversible, isothermal sol/gel transformation,” sols representing

liquid colloidal particle systems, and gels being colloidal materials with some shear resistance.
Introduction 3
can derange by a large strain, causing slope failure and exposure of the underlying
waste to flowing groundwater, erosion, and total loss of isolation potential.
The problem common to these cases is that the shear resistance can be exceeded,
which implies that the safety factor is too low, and that chemically induced changes
to its stress–strain properties can cause brittleness and loss of ductility.
The parameter of greatest importance for the three cases in the figure is shear
strength, which has to be sufficiently higher than the shear stresses generated by the
own weight and extra load acting on the potential slip planes. The strength of any
clay can be expressed in terms of drained or undrained shear strength, depending on
whether water can migrate in or out with time and change the stress conditions, or
if the time for such migration is so short that constant “short-term” conditions with
respect to the water content can be assumed. For smectite clay, the same applies
but there is also another phenomenon called creep, that is, long-term strain, which
plays a role that is a little different from that of nonexpandable clays with lower self-
sealing capacity than smectites, partly because of the thixotropy of the latter. For
smectite clays, creep strain with preserved ability to self-seal can be very large but
sustainable. The different performances depend on the microstructural constitution
and interparticle coupling, which also determine compressibility and expandability.
Stress and strain of smectite clays used as seals are of particular importance for
deep geological disposal of hazardous waste and for sealing deep holes that need to
be isolated from gas- and oil-bearing strata in sedimentary rock. Figure 1.2 shows
the main features of the Swedish concept KBS-3V (Svemar and Pusch 2005) for
5.8 m
6m
Fitted blocks of clay with 2000 kg/m3
density and a preconsolidation (swelling)
pressure according to the table
1.5 m Upward swelling pressure of initially 5 MPa
Dry density, Swelling pressure, MPa

kg/m3 3.5% CaCl2 solution
1m
5.5 m 1430 3.5
Canister
1270 1.0
1070 0.05
0.5 m 2m Blocks of clay with 2000 kg/m3 density

surrounding the canister
FIGURE 1.2 Schematic illustration of the Swedish concept KBS-3V for disposal of highly
radioactive waste from nuclear reactors. The table gives approximate values of the swelling pres-
sure of the block material for different dry densities. The upward pressure of the dense clay in
the hole compresses the tunnel backfill causing drag forces on the canister and surrounding rock.
4 Bentonite Clay
Fracture filled with

grout
Water-bearing
fractures
Concrete plugs
in fracture zone
Clay plugs
Concrete plugs
FIGURE 1.3 Schematic picture of very deep borehole for disposal of highly radioactive
waste, sealed off for preventing axial flow.
disposal of highly radioactive waste, indicating the need for solving practical prob-
lems related to the construction and placement of clay seals, which consist of highly
compacted clay blocks embedding the waste containers (canisters), and of tightly
fitting clay blocks in the tunnel from which the deposition holes will be bored. We
will return to this point several times in the book.
Figure 1.3 illustrates the role of stress and strain in sealing deep boreholes
reaching down to large depth. The problem here is to achieve balance between the
downward-acting vertical stress caused by the weight of the seals and canisters and
the upward- and downward-oriented swelling pressure of the clay seals. Unbalanced,
they will lead to movements and changes in density and thereby in tightness, which
can lead to leakage of gas from corroding canisters and migration of contaminants
from canisters with highly radioactive waste placed deeper down.
1.4 PREPARATION OF SMECTITE CLAY FOR SEALING PURPOSES

It is obvious from the large number of examples in this book that the density of
smectite clay seals has to be high for utilizing their low hydraulic conductivity. For
ordinary landfills of hazardous waste, top and bottom liners of smectite clay are
required, the construction of which includes preparation, placement, and compaction.
Construction of repositories for highly radioactive waste requires very dense
smectite-rich clay, which cannot be prepared and placed like smectite-poor clays.
Together with the requirement that their homogeneity must be high, this calls for
industrial preparation of blocks and precision installation. Blocks are made by uni-
axial or isostatic compression of clay granules or powder to the required dry density,
which can range from 1500 to 1900 kg/m3, depending on the required tightness and
Introduction 5
FIGURE 1.4 Clay block with 1.85 m diameter about to be placed in a 1.95 m bored hole in
SKB’s underground laboratory at Äspö, Sweden. The block is carried by vacuum technique
using stripes as backup. (Courtesy of G. Ramqvist.)
swelling pressure after the expansion that they will undergo in the space they are
placed in, that is, a hole or a tunnel. The need for rational handling and installation
of clay blocks in canister deposition holes speaks in favor of big blocks, as illustrated
by Figure 1.4. It shows a block of smectite-rich clay with a bulk dry density of about
1800 kg/m3, ready for installation in a deposition hole of KBS-3V type. The block
was prepared by uniaxial compaction of granular MX-80 with at least 70% smectite
(montmorillonite) under 100 MPa pressure.
Compaction of dense granules of smectite clay for preparing blocks with high den-
sity can be effected by mixing them with water for reaching a high degree of water satu-
ration after compaction, which is desired for obtaining high heat conductivity so that the
temperature of the canisters and buffer clay will not be too high. The newest technique
for achieving water saturation is to mix air-dry granules with dry water, which is com-
mercially available in the form of microscopic water droplets contained in extremely
thin silica shells (Forsberg et al., in press). The mixture can be so composed that the
material gets the desired dry density with all voids filled with water when compacting it
under the required pressure. Compaction breaks the shells into very small fragments by
which water migrates into the surrounding clay matrix driven by suction.
1.5 QUALITY ISSUES

The design of a clay seal would enable the latter to fulfill the required function
according to theoretical models of its performance. Here, we have a classical
problem in the soil mechanics discipline and in the entire field of construction
of buildings over and under the ground level: design is based on the material
properties as they are in the construction phase. For ordinary buildings, with a
planned lifetime of 50–100 years, no major changes of the performance of the
individual components are expected, but for waste repositories on ground and at
depth, considerable alteration can and will take place and cause contamination
6 Bentonite Clay
of the environment. Low-level radioactive waste can be safely stored in landfills

for a few hundred years after which they may not represent any risk if the halva-
tion time of the radioactivity is low enough. For chemical waste such as mercury,
arsenic, and zinc, the challenge is different: chemical waste retain its risk potential
forever and requires long-lasting confinement and hence a design that implies that
all possible degradation scenarios have been considered, including not only that of
the individual components but also of the system of interacting units. The roles of
several unknown factors such as climatic changes, earthquakes, and tectonics have
to be guessed and only a backup of barriers that have not been initially utilized can
provide the required safety.
Here is an example: Assume that you are asked to work out a design of an under-
ground repository for highly radioactive waste and can choose between a geological
region with very tight rock that is intersected by few fracture zones and one with
rather many zones. The apparently simple choice implying the selection of the for-
mer is in fact very complex and would bring you to suggest the latter. The reason is,
first, that the tightness is associated with very high rock stresses (that is why the rock
is tight!) that cause spalling and rock failure in the construction phase at a depth of
only 400–500 m, and that the clay embedment of the waste canisters will dry out and
permanently lose most of its planned function of tightly confining the canisters. The
alternative, the rock with rather many fracture zones, is “relaxed” and has lower rock
stresses and therefore provides a stable environment of the repository. Naturally, the
many fracture zones provide water for keeping the canister-embedding clay tight
and ductile; admittedly, construction of tunnels, shafts, and rooms is a little more
difficult because of inflowing water.
What sort of clay are we talking about? Well, it is the type of clay that the French
farmers got stuck in, that is, one that is rich in smectite. Such clay is available in
many areas of the planet because it was formed by the numerous volcanic eruptions
and comprehensive weathering of the continents throughout the geological eras. For
fulfilling the demands of long-term performance, repository planners and design-
ers are required to consider all possible changes in mineralogical composition and
dissolution and precipitation processes that one can imagine. Such work has been
going on since the 1950s, and many laboratory and field experiments on different
scales have been made that form the basis of the development of physical and physi-
cochemical models that have been used for predicting the performance of the clays
(discussed later in this book). Nature itself provides a number of examples of the
evolution of clays that simulate the performance and fate of repository clays, but
there has been too little focus on natural evidences, the examination of which is of
fundamental importance in this book.
The clay materials to be used in a repository are as follows: the buffer, being
the clay embedding the waste canisters, and the backfill of tunnel shafts, having
a lower content of smectite, which is prepared by processing of the raw material.
It includes drying, grinding, mixing with sodium carbonate, compaction on site,
and compression of dense blocks. As for all materials in nature, the composition
of raw clay materials always varies, and deviations from what is required has to be
compensated by changing the thickness and bulk density of the seals as specified
in the design. For preparation of liners in landfills, such adaption can be made in
Introduction 7
the course of the construction work by increasing the number of runs of vibratory
rollers or by increasing the intensity of the compaction energy. The achieved density
and uniformity of placed clay seals and of blocks manufactured for placement in a
repository must be checked by testing a sufficient number of samples. Specification
of the physical properties of the clay to be used according to the design worked out
is based on multiple laboratory testing and selection of representative data following
generally accepted geotechnical principles.
1.6 PERFORMANCE TESTS

Organizations responsible for safe performance of hazardous waste, such as
International Atomic Energy Agency and national inspectorates, usually require
instrumentation of repositories for recording the most important processes, such as
the hydration of initially not-water-saturated clay, the temperature, and the swell-
ing pressure. What they forget is that the connection of sensors in the clay seal
via cables to remote recording units can cause short circuiting by water m igrating
from the rock along the cables (Pusch et al. 2014). This phenomenon can give
false results as exemplified by experiments in SKB, the Swedish Nuclear Fuel
and Waste Management Co. underground laboratory at Äspö, Sweden (Figure 1.5).
Such pathways also make it possible for the released toxic species to migrate to the
surroundings. Therefore, installing instrumentation in clay seals that are placed for
isolating radioactive waste is not recommended (Svemar and Pusch 2005).
In a long-time perspective, clay seals will undergo degradation by dissolution or
chemical interaction with the surroundings, implying that their function can only be
adequately predicted by using theoretical models that include the chemical impact of
contacting media, or by comparison with natural analogs.
Total pressure after 270 days

5000
4000
Total pressure (kPa)
3000
2000
1000
0
0 10 20 30
Distance from heater (cm)
FIGURE 1.5 An example of false indication of the wetting rate of clay surrounding a canister
with heater for simulating the conditions in a KBS-3V hole in SKB’s Äspö laboratory. Water
coming from the surrounding rock followed cables and reached the sensor deep in the clay much
quicker than in clay without instrumentation. (Data from Pusch, R. et al., Predicted and Actual
Wetting Rate of Buffer in Repositories for High-Level Radioactive Waste, WIT Press, 2014.)
8 Bentonite Clay
REFERENCES
Forsberg, T., Pusch, R., Knutsson, S. (in press). Rational method for preparing dense clay
blocks with desired degree of water saturation. Engineering Geology, Vol. 182.
Pusch, R., Hatem, M., Kasbohm, J., Knutsson, S. 2013. Roles of clay and concrete in isolating
high-level radioactive waste in very long holes. International Journal of Research and
Reviews in Applied Sciences, 16:263–273.
Pusch, R., Ramqvist, G., Knutsson, S. 2014. Predicted and Actual Wetting Rate of Buffer
in Repositories for High-Level Radioactive Waste. WIT Press, Ashurst Lodge,
Southampton, UK.
Pusch, R., Yong, R.N. 2006. Microstructure of Smectite Clays and Engineering Performance.
Taylor & Francis, New York.
Svemar, C.H., Pusch, R., eds. 2005. Cluster Repository Project (CROP). Final Report of
European Commission Contract FIR1-CT-2000-20023, Brussels, Belgium.
2 Origin and Occurrence
of Smectite Clays
Bentonites
This chapter deals with the origin and occurrence of smectite clays. The purpose is
to give an elementary introduction to the fascinating science of smectite clays and
to provide to readers only what is necessary for digesting the rest of the book. For
anyone interested in the subject, there are kilometers of bookshelves in university
libraries, and articles are very frequently published in a number of geoscientific
journals.
2.1 ORIGIN OF SMECTITE CLAYS

Smectite stems from spontaneous nucleation and growth of crystals in saline water
with the glassy components of volcanic ash as source material, as well as from
weathering of feldspars and heavy minerals in hydrothermally affected rock. The
first-mentioned material is true bentonite, which forms beds of different thickness in
many parts of the world, whereas weathering products are confined to rock masses
with a structure that has been sufficiently pervious to let solutions through, such as
fracture-rich zones in crystalline rock and porous rock of sedimentary or volcanic
origin. This chapter presents an overview to show the widespread occurrence of
smectitic material of potential use. Some North American, Japanese, Middle Eastern,
and European bentonites are given somewhat more space than those from other areas
since they will be referred to which we will return in Chapters 5 through 8.
Commercial data on clay manufacturers and export and import of bentonite mate-
rials is found in the literature (i.e., Roskill Information Services 2001), which illus-
trates that clays rich in smectite are found in almost all parts of the world (Table 2.1).
2.2 WHERE ARE ALL THESE FAMOUS AND MAGIC CLAYS?

2.2.1 North America
The most widely used and mined North American deposits of sodium bentonite
were formed during the Cretaceous age through a combination of geological activity
and environmental conditions that favored formation of bentonites. Central North
America had been inundated by a shallow brackish sea that varied with time, espe-
cially in the northwestern areas (present Alberta and Saskatchewan). Contemporary
with this marine episode was extensive volcanic activity in the Western Cordillera
by which large quantities of ash were formed and blown east by the prevailing winds.
9
10 Bentonite Clay
TABLE 2.1
Production and Exports of Bentonite in 1997 according to Available Statistics,
China Not Included
Production Percentage Export Percentage
Country (kilotons) of World Production (kilotons) of World Export
The United States 6,390 49.4 1007 35.9
Germany 1,111 8.6 73 2.6
Greece 950 7.3 12 0.4
Spaina 846 6.5 603 21.5
Italy 543 4.2 182 6.4
Turkey 531 4.1 62 2.2
Japan 496 3.8 2 0.1
India 367 2.8 174 6.2
Korea 167 1.3 6 0.2
Mexico 163 1.3 13 0.5
The United Kingdom 135 1.0 158 5.6
Argentina 132 1.0 35 1.2
Brazil 110 0.8 – –
Czech Republic 110 0.8 21 0.7
Cyprus 100 0.8 52 1.9
Senegal 100 0.8 – –
Others 691 5.3 404 14.4
Total 12,942 100.0 2804 100.0
a Including sepiolite.
The result was a series of ash deposits on the inland sea. They gradually settled and
became buried by marine or land-derived materials. With time, these deposits were
deeply buried and consolidated by more recent sediments and many of these ancient
ash layers later altered to bentonite (Crawford et al. 1984).
Bentonite beds have a character that depends on the initial ash composition and
granularity, as well as on the pore-fluid chemistry of the deposits. Those in the north
and west of the continent tend to be more ashy with lower smectite content due to
their proximity to the volcanic sources of the ash or to the inclusion of terrestrial
material carried into the deposits during sea-level fluctuations. The deposits in the
central basin tend to be more homogeneous, perhaps as the result of calmer central
waters and finer textured material deposited farther from the volcanic ash sources.
2.2.1.1 The United States
The United States has large bentonite resources of the Upper Cretaceous and Tertiary
ages, the principal producing area being the Black Hills region of South Dakota,
Wyoming and Montana, and the Gulf region, especially in Texas and Mississippi
(Grim and Güven 1978). In the Black Hills area, the bentonite beds have a thickness
of a few centimeters to a few meters and appear in a series of marine shales, marls,
and sandstones. The bentonites in this region are assumed to have been formed
Origin and Occurrence of Smectite Clays 11
by in situ alteration of volcanic ash, which is believed to have been rhyolitic and
stemming from a western source. The best producing horizon has been the Clay Spur
bed forming the top of the Mowry Shale. The Wyoming bentonite has a particularly
good reputation because it has smectite content commonly as high as 75%–85%, with
sodium being the major adsorbed cation. This gives extraordinarily good colloidal,
plastic, and bonding properties.
The beds commonly cover more or less silicified (silicitated) strata, indicating
release of silica from the smectite in conjunction with its formation or early stages of
conversion to illite. At the outcrop, bentonite is often light yellow or green, whereas
it is bluish deeper down, the color shift being due to oxidation of iron or leaching
replacement of exchangeable sodium by calcium. In the Gulf area, most bentonites
have calcium as the major adsorbed cation.
The important question as to what the controlling factors really are for the type of
exchangeable cation in natural bentonites has no definite answer. Marine and lacus-
trine formations may have either sodium or calcium as the dominating cation, the
composition of the volcanic ash being the determinant.
2.2.1.2 Canada
Canada is rich in bentonite, particularly in the Prairie Provinces (Manitoba,
Saskatchewan, and Alberta). The bentonites, which are similar to those in Wyoming
and South Dakota, are of Cretaceous and Lower Tertiary ages. They occur in marine
shales and limestones and their thickness ranges between a centimeter and a few
meters. Alteration of volcanic ash is the generally accepted mode of formation. Most
of the Prairie bentonites have calcium as major exchangeable cation but sodium-
dominant clays occur as well. Cristobalite is a common accessory mineral that may
explain the slight cementation that is characteristic of many of the Prairie bentonites.
In fact, they are often shaly and somewhat brittle and fractured.
2.2.2 South America

South America has bentonites in Argentina, Peru, Uruguay, and Brazil. Clays origi-
nating from altered tuff are found in older, that is, Triassic, formations (Mendoza) in
Argentina (Grim and Güven 1978). Cretaceous smectite-rich bentonites formed by
ash fall in lacustrine environment occur in Brazil in the Ponta Alfa area of Minas
Gerais, and Permian bentonite beds, up to 3.3 m thick, have been found in the Estrada
Nova formation at Acegua in Rio Grande do Sul. Both sorts are rich in montmoril-
lonite, whereas in other Brazilian bentonites, the smectite is diluted with substantial
amounts of other minerals.
2.2.3 Central America
2.2.3.1 Mexico
Mexico is rich in bentonites in the volcanic area (Grim and Güven 1978). They origi-
nate from volcanic ash and have calcium as the major adsorbed cation, although
natural sodium smectites of high quality have been reported as well (American
Colloid Company and various other sources). Chalcedony overlying and underlying
12 Bentonite Clay
certain beds may be related to heating, in conjunction with hydrothermal alteration

of the ash. The extension and quality of the bentonite sources are poorly known.
2.2.4 Africa
Bentonites are common in Algeria and Morocco (Grim and Güven 1978). Some of
them stem from volcanic ash, such as those in the Chelif River area in Algeria,
whereas others have been formed by hydrothermal alteration of rhyolite rock.
The bentonites formed by ash fall vary in thickness from about 10 cm to around
5 m. They represent Cretaceous and Tertiary sediments with interbedded pyroclas-
tics and lava flow, the complex stratification indicating variation in homogeneity and
smectite content. Some of them have high smectite content with sodium as the major
adsorbed cation and with no enrichment of silica below or above. A number of these
bentonites have been extensively utilized commercially.
In Morocco, bentonites of Miocene age are common, often in the form of lay-
ers with a thickness of a few centimeters. They stem from ash fall. An area known
to have smectite-rich bentonites of high quality is Camp-Berteaux in the Taourint
region. The smectites from this area serve as one of the French reference montmoril-
lonite minerals.
In South Africa, bentonite beds of Liassic age occur with a thickness of more
than 2 m in the Karoo system. These beds—which have calcium as the major
exchangeable cation and do not originate from ash fall—are composed of very pure
montmorillonite. In Mozambique, thick bentonite masses, formed by hydrothermal
alteration of perlite, occur but they contain rather much unaltered rock fragments
and cristobalite.
2.2.5 Middle East
Clays are known for their sealing potential in many environmental and remedial
contexts, and the best ones belong to the group of expanding clay minerals known
as smectites. Various types of wastes ranging from municipal to highly radioactive
wastes are being isolated from the biosphere in repositories relying on smectite-
based barriers. Iraq has considerable resources of such clay minerals, especially in
the northern and desertic areas. The Western desert holds more than 22 million tons
of calcium montmorillonite deposits and the Al-Jezira and Southern deserts contain
commercially useful claystone deposits (1–10 m thickness) with montmorillonite and
palygorskite as dominant clay minerals (Altaie 2014). They have been exploited for
isolation of chemical and radioactive wastes produced during and after the wars in
1991 and 2003.
2.2.6 Asia
2.2.6.1 China
Clay has been used in China for pottery since 6000 BC, the art being escalated at
around 3500 BC in the Shandong province, where kaolin clay was found and used
near the confluence of the Yellow, Fen, and Wei Rivers, and later along the upper,
middle, and lower reaches of the Yellow and Yangtze Rivers. Smectites were used
as well, the raw material coming from bentonite beds in volcanic areas. China
has considerable smectite clay resources, mostly in the form of bentonite, and at
present, more than 80 areas of potential use have been discovered, 12 of which
represent reserves of more than 50 million tons and the total reserve is at least
twice as high.
In Xinghe County, Inner Mongolia Autonomous Region (113°58′ to 114°06′E;
40°27′ to 40°44′N), the Gaomiaozi bentonite deposit has been selected as the prefer-
ential candidate for preparing swelling clays for high-level radioactive waste reposi-
tories in China (GMZ). This is the third largest bentonite deposit in China, with
a total reserve of about 120 million tons of sodium bentonite. In the Gaomiaozi
deposit, five bentonite layers have been identified, interbedded in Jurassic rhyolite
and tuff sequences. The dominant clay mineral is montmorillonite (average concen-
tration of more than 60%), with some illite, feldspar, quartz, calcite, and zeolites.
The very large reserves of sodium bentonite in Gaomiaozi are enough for the
present estimated domestic demand for high-level radioactive waste isolation and
serve also to meet the overseas demand for this purpose.
2.2.6.2 India
India has quite pure smectite clays of the Lower Tertiary age (Grim and Güven 1978).
They form large deposits more than 3 m thick and are believed to have stemmed
from in situ weathering of igneous and metamorphic rocks followed by erosion,
transportation, and deposition in the Barmer Embayment.
One-meter-thick bentonite beds have been found in the Jammu and Kashmir
areas. They seem to have been formed by alteration of volcanic glass in ash.
2.2.6.3 Japan
Bentonites formed from volcanic ash and by in situ alteration through hydrother-
mal processes in Tertiary liparites and liparitic tuffs are present on several sites in
Honshu and Hokkaido. At Yamagata and in the Tsukinuno area on the main island,
the deposits were formed in Early Miocene mostly by hydrothermal processes that
gave large irregularly shaped clay bodies. Depending on the composition of the par-
ent rock, zeolites have also been formed in large amounts parallel to smectites. In
the Kwanto district of the Gumwa Prefecture, there are several beds with a total
thickness of up to 5 m, interbedded with mudstones and hence of heterogeneous and
relatively poor quality.
A general conclusion is that most of the clays originated from volcanic ash irre-
spective of the formation process. Since the volcanic activity of Japan has been very
significant from the Neogene epoch, the amounts of clay are very large. Particularly,
the submarine sedimentation of the Miocene dacite pumice was very comprehen-
sive, and most of the deposits gave huge masses of montmorillonite, zeolite, and
opal. Still, most of the smectite clays that are being exploited commercially are
altered tuffs.
Clay formation has been extensively studied in Japan and led to considerable
deepening of the understanding of the mechanisms involved in the formation of
smectites. The conclusions are that weathering tends to produce a suite of clay
14 Bentonite Clay
TABLE 2.2
Geological Profile in the Tsukinuno Area
Smectitic
Age Formation Unit Thickness (m) Formations
Tertiary (Miocene) Black mudstone 170
Nukumi F. Gray tuffaceous sandstone 80
Brown pumiceous tuff 30 +
Dark brown hard mudstone 20
Gray pumiceous tuff 20 +
Tsukinono F. Dark brown hard mudstone 290 +
Brown hard shale 100 +++
Pale green sandy tuff 100 +
+, low to moderate richness; +++, rich.
minerals rather than a single one, and that other clay minerals rather than smectite
are likely to be abundant. In most of the smectites originating from hydrothermal
influence on rhyolitic rock, there are frequent quartz phenocrysts and a considerable
quantity of plagioclase and biotite. The smectite content is hence not very high as
indicated by the list in Table 2.2 of smectite-bearing strata in the Tsukinuno area,
Shiga prefecture.
2.2.7 Europe
European bentonites are well known and are easily available in Spain, Germany,
Denmark, Hungary, the Czech Republic, Italy, Greece, and Bulgaria. We will confine
ourselves here to mention those in Germany, Italy, Spain, the Czech Republic,
Greece, and Denmark. Those from Germany, Spain, and the Czech Republic have
all been examined in the search for suitable clays for isolation of highly radioactive
waste (Pusch et al. 1995; Svemar and Pusch 2005).
2.2.7.1 Germany
Germany is an important source of commercial bentonites, generally emanating from
in situ alteration of acid vitreous tuff. At Moosburg in Bavaria, smectitic clays with
calcium as the major exchangeable cation forms up to 3 m thick beds in a generally
marine section of marls and tuffaceous sands. The quality varies from relatively pure
montmorillonite to material with rather much kaolinite and illite. Typically, there is a
distinct boundary between an upper oxidized (yellow) and a lower blue-colored part
in one of the pit areas (Karnland and Pusch 1990). In the Neubrandenburg region of
northern Germany, there are huge quantities of smectitic mixed-layer clay that are
exploited for isolation of shallow waste piles. This material is being considered as a
possible candidate for backfilling of drifts and shafts in repositories for radioactive
and hazardous chemical waste (Schomburg 1997); see Figure 2.1. The clay, which
is of Tertiary age, has an average content of expandable minerals of 50%–60%, half
FIGURE 2.1 (See color insert.) Open pit mining of water-saturated Friedland clay: the pit
is about 100 m deep and dry.
of which is montmorillonite and the rest being mixed-layer minerals. It is termed

Friedland clay in Mecklenburg-Vorpommern, and we will use this name throughout
this book.
2.2.7.2 Denmark
Denmark is rich in commercially available relatively smectite-poor clays that are
potentially useful for waste isolation. A smectite-rich clay marketed by the Danish
company Dantonite AS is termed Holmehus, which belongs to a Paleocene forma-
tion that also comprises the Æbelö and Rödbyhavn clays, which are all akin to the
German Friedland clay (Pusch et al. 2014). Holmehus clay is an illite/smectite mixed-
layer clay containing also dioctahedral vermiculite/smectite mixed-layer intercala-
tions. The content of expandables is at least 60%. Holmehus clay is composed mainly
of weathered material from the Baltic Shield and has undergone diagenetic processes
after sedimentation (Figure 2.2). We will come back to this clay type several times
in this book.
2.2.7.3 Italy
Italy is rich in bentonites, particularly in the islands Ponza, Sardinia, and Sicily.
Most of them originate from alteration of basic rocks, some due to the effect of
gases left in intrusions of magma, whereas others appear to be of sedimentary ori-
gin. One case of the latter type is concluded to be the presently mined bentonite at
Busachi in central Sardinia (Pusch 1983). Here, molten rhyolite is assumed to have
16 Bentonite Clay
FIGURE 2.2 Mining of smectite-rich Holmehus clay at Taasinge (Fyn), Denmark. (Courtesy
of Dantonite AS.)
moved in, over a more than 7 m thick reddish, rather homogeneous bentonite mass,
yielding sintering at the contact but causing very little changes at larger distance
from the hot contact than about 2 m. This and similar sites represent rich potential
bentonite sources. The Uri, Oristano, and Cagliari regions represent bentonite-rich
areas and have processing plants. At Cagliari, the bentonite reserves represent tens
of millions of cubic meters, a generalized section of the valley with up to 10 m thick
bentonite beds. In the Pedra de Fogu area, there are up to 25 m thick bentonite beds.
At Cagliari, the smectite series is a Miocenic alteration product of a more than 13 m
thick fine pyroclastic stratum, whereas at Pedra de Fogu, it stems from in situ altera-
tion of rhyolitic rock (Lombardi and Mattia 1981).
2.2.7.4 Spain
In recent years, systematic prospecting has given a good overview of the Spanish
resources of smectite clay. Three areas have turned out to be of particular practical
importance. They have all been formed by alteration of Tertiary rhyolite-dacites
through hydrothermal processes or percolation of meteoric water involving removal
of some silica and iron, and the incorporation of aluminum and magnesium.
Concentration of chromium, nickel, and cobalt has caused coloring of some of the
bentonites (Linares 1989).
The Madrid Basin, part of the Tajo Basin, lies in the southern edge of the
Sistema Central. The Tajo Basin is a Tertiary depression, located in the mid-
dle of the Iberian Peninsula, which became filled with continental sediments,
ranging from fluviatile to lacustrine–palustrine environment-type deposits.

A major characteristic of these deposits is a lateral change of facies from detrital
to chemical. Clay materials of saponite type are commonly found in the transi-
tion zones and they represent about 500,000 m 3. Three different units have been
established: (1) a lower unit composed mainly of evaporites; (2) an intermediate
unit made up of clays, sandstones, limestones, and chert; and (3) an upper unit
of essentially detrital materials, associated with tectonic phases that occurred
during the Miocene. The intermediate unit is of particular interest and is made
up of materials from lacustrine and fluviatile environments, with some transition
zones. Predominating in the transition zones are clay materials associated with
silty sands. Biotitic sepiolite deposits, exploited in the Madrid Basin, are found
within these clay formations.
Tests performed in the program of ENRESA* have shown that the Cerro del Aguila
bentonites with a smectite content of more than 85% have the best characteristics of
three deposits identified in the Madrid Basin. Available field data (Linares, 1989)
from borings indicate that the potential reserves and an easy to mine configuration
of the deposits meet the requirements set.
The main Spanish bentonite deposits of hydrothermal origin are located in Cabo
de Gata (Almeria, southeastern Spain). The volcanic activity in this area gave rise to
calc-alkaline rocks, from basaltic andesites to rhyolites in composition. The volcanic
emissions varied from massive lava flows to different types of pyroclastic prod-
ucts. The rocks are heavily fractured and have undergone comprehensive physical
and chemical alterations. Most of the pyroclastic materials have experienced hydro-
thermal alteration. Acid solutions gave rise to the formation of alunite, jarosite, and
kaolinite but the most extensive alteration resulted in the formation of the thick depos-
its of smectite clay spread all over the Cabo de Gata region. The bentonites in Serrata
de Nijar occupy an area of about 11 × 1.5 km2 and extend to at least 20 m depth. The
smectite content is in the interval 85%–95%, the major species being montmorillonite.
Other sites are the montmorillonite-rich zones of Los Trancos and Escullos.
The process of alteration of the volcanic materials to bentonite took place with a
large loss of matter, in particular silica and alkaline elements, accounting for over 50%
of the fresh material. The accompanying decrease in volume gave rise to some mor-
phologic readjustments in the altered pyroclastic bed cover. As a rule, the hydrothermal
solution supplied Mg to the initial fresh rock. The meteoric waters, heated during their
underground southward movement from the metamorphic northern ranges, altered the
pyroclastic beds into bentonite. Since the hydrolysis reaction of these materials is exo-
thermic, the solutions did not need to reach great depths or high temperature bodies to
be heated up. These smectites belong to the montmorillonite–beidellite series and their
compositions vary gradually with the degree of alteration and the chemical composi-
tion of the parent materials. Thus, beidellites correspond to very early alteration stages,
especially in acid rocks. The color of the bentonite depends on the content in trace ele-
ments of the first transition series; all shades of green, blue, red, and white are present.
The bentonite deposits are numerous and important all over the region and reserves are
estimated to be tens of millions of cubic meters.
* ENRESA, a Spanish organization responsible for disposal of highly radioactive waste in Spain.
18 Bentonite Clay
2.2.7.5 Czech Republic

The Bohemian Massif in the Czech Republic is estimated to contain at least
280 million tons of high-grade bentonites representing 16 sites, of which three are
being mined. One of them is the Zelena clay quarry in the upper Miocene Cheb
basin, which holds up to 20 m thick smectitic layers. Both Ca-dominated clay and
Na-dominated clays have been identified and exploited (Prikryl et al. 2001).
2.2.7.6 Greece
The volcanic islands of Greece were formed by transgressive sedimentation on a
crystalline base, a dominant part of the island being covered with sediments of
volcanic origin deposited in late-Tertiary and early-Quaternary times (Wetzenstein
1964; Decher 1997). Young alluvial sediments are found in tectonically gener-
ated depressions. Volcanic activity was initiated in Pliocene forming tuffs and
tuffites and proceeded in five phases, which has yielded considerable complexity.
The geothermal conditions are still very obvious as demonstrated by the fact that
water vapor with 300°C temperature is contained in pores and fractures at less than
1 km depth. The island of Milos represents a major bentonite resource and is briefly
commented here.
The geology of Milos has been described in detail with respect to the presence
and extension of smectitic clays (Decher 1997). There are several mineralogically
and chemically different clay beds, the ones of major practical interest being formed
by in situ alteration of andesitic pyroclastic rock characterized by high contents of
volcanic glass and plagioclase that were transformed to montmorillonite in many
places. The main clay beds in one area, Agrilies, consist of iron-poor smectites with
high quartz and mica contents. Xenolites in the original rock are largely preserved
here. In the Ano Komia area, bentonites with rather little smectite and significant
amounts of cristobalite and tridymite were formed from pyroclastic rock. In certain
areas, particularly in shallow positions and in shear zones, hydrothermal solutions
have percolated and degraded the smectites.
In the area Aspro Chorio, where smectite-rich bentonites are available in very
large quantities, the processes involved in the transformation of pyroclastic rock and
the evolution of smectites have been studied in detail. It is claimed that percolation of
meteoric water of relatively porous rock was a major factor in the smectite formation
in this and other areas. The smectite content is commonly in the range of 50%–70%,
but prospection has shown that there are very large resources of clay material with
more than 70% smectite of montmorillonite type.
2.3 POTENTIAL SMECTITE RESOURCES

The formations described in this chapter represent very large amounts of clay that
can be used as raw material for enrichment of smectite minerals or for use as it is.
Complete data of the quantity of smectite clay are usually not available but approxi-
mate figures can be obtained from certain companies or derived from stratigraphic
data. The total known resources on land of bentonitic materials in the countries
specified earlier in the chapter is estimated at E11–E12 m3, but the actual figure may
be much higher.
The outcropping Black Hill bentonites cover an area of at least 500 km2
corresponding to at least E10 m3 of bentonitic material. The bentonite deposits in the
Gulf States probably represent about the same quantity (Allen 1977).
Japanese bentonites also represent very large volumes. Thus, in the Prefecture of
Gumma, the bentonite zone is about 25 km long and 2 km wide, corresponding to at
least 108 m3 bentonite-holding material. The total amount of exploitable bentonitic
material in Japan can be estimated at E10 m3 (Iwa 1969).
The total European resources, excepting Russia, can be estimated at more than
E10 m3 corresponding to about 2E10 metric tons. In Italy, the Sardinian resources of
bentonitic clay represent no less than E9 m3, and the entire amount of such clays in
the whole of Italy probably also exceeds E9 m3. The Greek and Spanish resources
are estimated to be on the same order of magnitude.
2.4 ARE NEW SMECTITES BEING FORMED TODAY?

In principle, erosion followed by fine particle transport and sedimentation take place
continuously on the planet and create layers of clay on the sea bottom at a rate of
parts of a millimeter per year off the coasts and to decimeters or more per year near
the outflow ends of big rivers. Sedimentary geology tells us that eroded clay particles
that have been transported by ice, wind, or cold water have their crystal constitutions
largely intact, including also reactive cations such as Fe2+. Mineralogical analyses of
deep sea sediments since the early twentieth century show that while the most com-
mon clay mineral is chlorite, where weathering processes are predominantly physi-
cal as in Scandinavia, this mineral makes up only 13% of the clay minerals in the
ocean areas. Here, illite is the most common clay mineral often representing about
50% of the whole series of such minerals. This mineral is believed to result from both
mechanical and chemical weathering in temperate climates or in high altitudes in the
tropics, and brought to the oceans via rivers and wind transport.
It is well known that kaolinite is formed by recrystallization of reaction products
from chemical weathering, and is hence mostly found in low latitudes such as in
deep-sea sediments off equatorial West Africa. Kaolinites like chlorites make up
only about 13% of the clay minerals in the deep sea.
Summing up the concentrations of other clay minerals, we find that smectites are
not very abundant. Except for chemical weathering of certain rocks such as feldspar-
rich granites and other crystalline rock rich in minerals such as hornblende and
amphibole, smectites are chemical alteration products of volcanic glass. They are
hence most common where eruptions occur from volcanoes located close to salt sea
on Iceland. One notices that smectites are most common in the South Pacific, where
they make up about 50% of the clay mineral assembly.
2.5 QUALITY OF NATURAL SMECTITE CLAYS FOR PRACTICAL USE

The quality of smectite clay deposits is determined by various criteria, such as a
minimum smectite content or a maximum content of certain accessory minerals or
organic material. Most important factors are the type of smectite and the mineralogi-
cal and granulometrical homogeneities.
20 Bentonite Clay
Autochtonic bentonites originating from ash fall are commonly relatively

h omogeneous, although thin layers of quartz sand and silt, gypsum, palygorskite, and
cristobalite often occur in the beds. Allochthonous bentonites such as the Triassic
bentonites in Argentina usually show much larger variations in composition because
nonbentonitic material is carried to the area of accumulation together with ash and
smectite.
In situ alteration of rock by hydrothermal effects or percolation of meteoric water
has often yielded homogeneous reaction products but since percolation has usually
taken place in a nonuniform pattern, the altered mass often has irregular shape and
unaltered lenses are often noticed. Furthermore, the quartz and plagioclase com-
ponents, as well as biotite, are commonly intact, meaning that the smectite content
seldom exceeds about 50%.
2.6 CONCLUSION
In this chapter, we have demonstrated that the following are the most important
properties for exploiting a clay source for obtaining waste-isolating materials:
• Available amount of exploitable smectite-rich material

• Smectite type
• Degree of heterogeneity with respect to the smectite content and accessory
minerals
REFERENCES
Allen, B.L. 1977. Mineralogy and soil taxonomy. In Minerals in Soil Environments, Soil
Science Society of America, Madison, WI.
Altaie, L. 2014. Performance of near-surface repository for radioactive waste in the Iraqi
Western Desert. Applied Clay Science, 87:73–80.
Crawford, A.M., Radhakrishna, A.S., Goutama, M., Lau, K.C. 1984. Engineering Materials
for Waste Isolation. Canadian Society for Civil Engineering, Montreal, Quebec, Canada.
Decher, A. 1997. Bentonite der Insel Milos/Griechenland. Aachener Geowissenschaftliche
Beiträge. Verl. der Augustinus-Buchh, Aachen, Germany.
Grim, R.E., Güven, N. 1978. Bentonites. Developments in Sedimentology, vol. 24. Elsevier
Publishing, New York.
Iwa, O.S. 1969. The Clays of Japan. Proceedings of the 1969 International Clay Conference
in Tokyo, Part 1. Geological Survey of Japan, Tokyo, Japan.
Karnland, O., Pusch, R. 1990. Development of Clay Characterization Methods for Use in
Repository Design with Application to a Natural Ca Bentonite Clay Containing a Redox
Front. SKB Technical Report TR 90-42. Swedish Nuclear Fuel and Waste Management,
Stockholm, Sweden.
Linares, J. 1989. Investigacion de bentonitas como materiales de sellado. U.E.I. Fisicoquimica
y Geoquimica Mineral Estacion Experimental del Zaridin (CSIC), Granada, Spain.
Lombardi, G., Mattia, P. 1981. Guidebook for the excursions in Sardinia and central Italy.
Proceedings of the 7th International Clay Conference. AIPEA, Rome, Italy.
Prikryl, R., Ryndova, T., Vejsada, J., Jelinek, E., Bohac, J., Sebec, O. 2001. Tertiary montmo-
rillonite-rich clays from the Cheb basin (Czech Republik): Alternative backfill material
in the nuclear waste disposal. Proceedings of the Joint 6th Biennial SGA-SEG Meeting,
Kraków, Poland, August 26–29.
Pusch, R. 1983. Stability of Deep-Sited Smectite Minerals in Crystalline Rock—Chemical

Aspects. SKBF/KBS Technical Report TR 83-16. Swedish Nuclear Fuel and Waste
Management, Stockholm, Sweden.
Pusch, R., Kasbohm, J., Thao, H.M., Knutsson, S., Ngyuen Thanh, L. 2014. Holmehus clay—
A tertiary smectitic clay of potential use for isolation of hazardous waste. Engineering
Geology, 188:38–47.
Pusch, R. et al. 1995. The Buffer and Backfill Handbook, Part I. SKB Technical Report 95-45.
Swedish Nuclear Fuel and Waste Management, Stockholm, Sweden.
Roskill Information Services. 2001. The Economics of Bentonite. 9th edn. Roskill Information
Services, London.
Schomburg, J. 1997. Data and Literature Collection “Friedland Clay.” Int. Rep. DURTEC,
Neubrandenburg, Germany.
Svemar, C., Pusch, R. 2005. Prototype Repository Project, European Commission Contract
FIKW-2000-00055, Final Report, Brussels, Belgium.
Wetzenstein, W. 1964. Minerogeologisch bedingte Unterschiede der rheologischen
Eigenschaften bei Bentoniten von der Insel Milos/Griechenland. 117 Jahrgang, Heft 8,
Berg- und Hüttenmännische Monathefte.
3 Nature of Smectites
This chapter describes, based on classical literature and current research and
development, the constitution of smectite minerals, of which montmorillonite is the
most common and well-known representative, as a basis for understanding the
behavior of smectite clays in general and their performance in isolation of hazard-
ous waste in particular.
3.1 BASICS
Ordinary clays are more or less fine grained; those containing smectite clay minerals
have extraordinary fineness. Smectite clays are known to be physically unstable by
sorbing or giving off water in conjunction with swelling or contraction much more
than illitic or kaolinitic clays, which makes them particularly sensitive to changes in
water content. In road construction, they are hence feared, but kept confined in bore-
holes or for isolation of hazardous waste down in the rock and in the form of landfills
with sufficient overburden. They are very popular because of their tightness, but
they are chemically unstable and change their behavior if the chemical conditions in
the environment are altered. Smectites are favorites in cosmetics and dermatology
because of their sorption potential and ability to interact well with the human skin.
We will focus on these and other properties of clays that contain smectite, which is
the dominant component of bentonite.
3.2 SMECTITE FAMILY

The minerals that give smectite clay its valuable waste-isolation properties of being
low permeable and expandable consist of stacked lamellae, each of which consists of
two layers of SiO4 tetrahedrons confining a central octahedral layer of hydroxyls, Fe,
Mg, or Li ions (Figure 3.1). These minerals belong to the following groups:
• Montmorillonite: Only Si in the tetrahedrons and Al in the octahedrons

• Beidellite: Si and Al in the tetrahedrons and Al in the octahedrons
• Nontronite: Si and Al in the tetrahedrons and Fe in the octahedrons
• Hectorite: Si in the tetrahedrons and Mg and Li in the octahedrons
• Saponite: Si and Al in the tetrahedrons and Mg in the octahedrons
The lamellae usually do not exist as single units but combine to form stacks with
three or more aligned lamellae depending on the cohesive, or bond, strength, which
is different for different adsorbed cations. Such stacks represent clay particles, which
are more or less interwoven in natural undisturbed smectite-rich clays and in artifi-
cially manufactured ones for serving as waste isolation (Figure 3.2). This interwoven
23
24
Silica tetrahedron Fe, Al, or Mg octahedral unit
Simplified Simplified
view view
= Si = Oxygen = Oxygen or hydroxyl = Al, Fe, Mg
Basic unit layer for smectite consists of an octahedral sheet sandwiched

between two tetrahedral sheets
Basic unit cell Basic unit layer; thickness = 1.0 nm
2:1 layer-lattice mineral particle—smectite
Repeating unit
layers
Sorbed (hydration) water
layers
Repeating unit
layers
Repeating unit
Sorbed (hydration) water
Repeating unit
layers
FIGURE 3.1 Silica tetrahedral and octahedral units with metal cations. The basic unit cell and the basic unit layer have a thickness of about 1 nm
(10Å). The stacking of unit layers with the interlayer or interlamellar space occupied with hydration water. Note that the terms layer (layers) and lamella
(lamellae) mean the same and are often used interchangeably. (Yong et al. 2010)
Bentonite Clay
Nature of Smectites 25
FIGURE 3.2 Typical appearance of stacks of smectite lamellae. Transmission electron

micrograph of ultrathin section (200Å) of acrylate-embedded MX-80 clay with a density at
water saturation of 1670 kg/m3 (dry density 1060 kg/m3). Magnification 5000×.
feature makes it difficult to define and determine the particle size of such clays,
which depends on the manner of sample procurement and testing.
3.3 CRYSTAL CONSTITUTION OF SMECTITES

A description and discussion of how smectite-rich clays perform needs consideration
of the atomic constitution of the crystallites and of the mechanisms that make them
sorb water, ions, and molecules. The development of theoretical models of the crystal
lattice of the most common smectite species, montmorillonite, is in fact a fascinat-
ing story that may not yet have come to an end. The first to suggest a crystal model
was a German team of mineralogists (Hofmann et al. 1933) who in 1933 proposed
a pyrophyllite-type structure composed of two silica tetrahedral sheets with an alu-
mina octahedral sheet in between. The sheets were assumed to be connected, so
that the tips of the tetrahedrons fit the octahedrons (Figure 3.3) that combine to give
lamellae that extend in the crystallographic a and b directions for up to a microm-
eter. In the normal c direction, they build up stacks that can consist of 3–20 lamellae
depending on what cations that dwell between them. Later, Edelman and Favejee
suggested an alternative model that is shown in Figure 3.3 together with Hofmann’s
model. The special and practically very important feature of smectite minerals is
that water and other polar fluids can enter between the lamellae as illustrated in the
figure (Grim 1968).
26 Bentonite Clay
n(H2O)
~10–20
10–20
n(H2O)
OH
O
Mg, Al
Si, Al
FIGURE 3.3 The montmorillonite crystal lattice. Left: Hofmann/Endell/Wilm version.

Right: Edelman/Favejee model. (Data from Edelmann, C.H. and Favejee, J.C.L., Zeitung fuer
Kristallographie, 102, 340–348, 1940.)
Neither Hofmann’s nor Edelman’s models are in complete agreement with cation
exchange capacity (CEC) or X-ray diffraction (XRD) data, which has caused various
modelers to complete certain vital structural changes, such as reducing the fraction
of inverted apical tetrahedrons of the Edelman model and replacing some of the Si
tetrahedrons in Hofmann’s model by (OH)4 tetrahedrons. A third model, based on
Edelman’s model for explaining hydration/dehydration of smectites, has been pro-
posed by Forslind and Jacobsson (1972) and is worth consideration, as we will see.
3.4 CHEMICAL COMPOSITION OF NATURAL SMECTITE CLAYS

3.4.1 Smectite Component
Table 3.1 gives the chemical composition of a number of smectite-rich clays from
various parts of the world in oxide form as well as the layer charge and the water
contained in the interlamellar space (H2O −) after drying at 105°C, following the
ordinary geotechnical procedure for determining the water content. Wyoming-type
montmorillonites are frequently claimed to be superior to others because Na is the
major exchangeable cation providing the best expandability and tightness. MX-80
is a commercial product of Wyoming bentonite marketed by the American Colloid
TABLE 3.1
Composition of Commercial Wyoming-Type Montmorillonites Expressed
as Weight Percentages of Oxides
Wyoming-Type
Oxide 1 2 3 4 5 6 7 8 9
SiO2 64.80 62.00 62.30 62.70 59.73 60.22 60.76 59.91 58.67
Al2O2 24.54 23.42 23.50 22.20 24.30 23.67 23.08 21.97 27.34
Fe2O3 1.27 3.74 3.35 4.62 5.54 6.28 6.10 6.72 3.64
MgO 1.60 0.93 1.95 2.00 2.10 1.46 1.44 2.15 2.00
CaO 0.00 0.68 0.31 0.58 0.00 0.13 0.17 0.34 0.00
Na2O 0.40 0.72 0.40 0.01 0.80 0.09 0.13 0.09 0.62
K2O 0.60 2.63 0.03 0.12 0.22 0.19 0.21 0.11 0.18
Al tetrahedral 0.04 0.08 0.10 0.08 0.24 0.20 0.16 0.17 0.34
Al octahedral 1.72 1.66 1.64 1.55 1.56 1.56 1.56 1.48 1.67
Si tetrahedral 3.96 3.92 3.90 3.92 3.76 3.80 3.84 3.83 3.66
Layer charge 0.34 0.34 0.31 0.33 0.34 0.34 0.30 0.35 0.34
per formula
unit
Source: Pusch, R. The Buffer and Backfill Handbook, Part 1, Swedish Nuclear Fuel and Waste
Management AB, Technical report TR-02-12, SKB, Stockholm, Sweden, 2002.
1—Hojun Mine, Gumma, Japan; 2—Tala, Mendoza, Argentina; 3—Crook County, Wyoming; 4—Rokkaku,
Yamagata, Japan; 5—Amory, Mississippi; 6—Santa Elena, Mendoza, Argentina; 7—San Gabriel, Mendoza,
Argentina; 8—Emilia Calingasta, Argentina; 9—Sin Procedencia, Argentina.
Company. It has been taken as representative of commercial Na smectites in the

research programs of a number of organizations that are responsible for the manage-
ment and disposal of highly radioactive waste, such as the Swedish Nuclear Fuel and
Waste Management Company (SKB). We see that the data show a surprisingly small
variation in the chemical composition of these bentonites, which consequently have
very similar physical properties. The controlling parameter for the latter is the ratio
of SiO2/Al2O3, which ranges between about 2.2 and 2.8.
The smectite clays typically have a high aluminum content and a layer charge of
0.30–0.35 per formula unit. The aluminum content in octahedral positions is approxi-
mately the same for all the clays, and this is also the case for the tetrahedral positions
of Si in almost all of them. The silica content in tetrahedral positions is high for all the
clays, which hence represent relatively pure montmorillonites. The iron content is low
for Japanese montmorillonite from the Hojun mine, whereas it is appreciable for three
of the clays from Argentina. High concentration of iron may cause changes in the crys-
tal constitution of smectite minerals under hydrothermal conditions or lead to formation
of cementing precipitates that reduce their expandability and gel-forming capacity. The
risk is lowest for montmorillonites from Crook County and the Hojun mine. Potassium
is most abundant in the Tala and Hojun bentonites, possibly indicating some illitization.
28 Bentonite Clay
3.5 MINERAL COMPOSITION OF NATURAL SMECTITE CLAYS

3.5.1 Smectite-Rich Clays
Table 3.2 lists the main mineral constituents of commercially available smectite-
dominated clays that have been studied and used for most of the purposes mentioned.
3.5.2 Clays with Moderate and Low Smectite Contents

Natural smectite clays of sedimentary origin, such as most true bentonites, contain
various other minerals deposited together with the smectite minerals or are remain-
ders of particles that were later converted to smectite. Usually, quartz, feldspars,
micas, chlorite, sulfides, sulfates, oxides, and carbonates are therefore present in
various amounts. Significant concentration of certain minerals, such as sulfides,
gypsum, and chlorides, in smectite clay used as embedment of canisters with highly
radioactive waste in repositories is unsuitable since they will attack those made of
metallic copper and iron.
The content of amorphous inorganic material and organic material may be
considerable. The amorphous material commonly consists of SiO2 and Al- and
Fe-hydroxides, usually amounting to a few percent by weight, whereas the organic
content is commonly 0.2 wt%–0.7 wt%. Nonsmectite constituents do not contribute
to the sealing power of the clay except for gypsum and chlorides that will be precipi-
tated in the hot part of clay seals exposed to a thermal gradient. Once the thermal
gradient disappears, they go into solution and diffuse out of the seals. Some of the
rock-forming minerals in natural smectite-rich clays, primarily the feldspars and
TABLE 3.2
Examples of the Mineral Composition of Smectite-Rich Clays in Weight Percent
Central
South- European
Belle European (RMN),
Hojun, Amory, Upton, Fourche, (IBECO), Czech
Mineral Japan MS WY SD Spur, WY Germany Republic
Montmorillonite 83.9 89.3 79.5 91.9 87.7 88 90
Quartz 14.9 5.9 16.3 4.0 10.0 2 2
Muscovite – 0.1 0.1 0.1 – – –
Chlorite – – 0.1 – – –
Feldspars – 4.7 2.3 4.0 2.3 2 <1
Others 1.0 – 1.8 – 0.3 8 7
Total 99.8 100.0 100.1 100.0 100.0 100 99.5
Trace minerals Bio Bio Bio Bio Bio Bio Mica
Lim Lim Lim Lim
org. org.
Bio, Biotite; Lim, Limonite; org., organics.

TABLE 3.3
Mineral Composition of Commercial Smectite Clays Tested by Swedish Nuclear
Fuel and Waste Management Company (SKB)
Clay Smectite Quartz Feldspar Mica Carbonates Pyrite Others
MX-80 (American Colloid 75 *** ** * * * *
Company)
Tixoton (Na) (Sued-Chemie) 90 * * – ** * –
Moosburg (Ca) 65 ** ** *** ** – **
(Sued-Chemie)
RMN (Obrnice) 90 * * * ** – –
Saponite (Geohellas) 70 ** ** ** – – –
Kunigel (Kunimine) 50 *** * – * * –
Friedland (Trade names) 45a *** ** ** * * –
Holmehus (Dantonite) 70a ** * * * * –
*Traces (< 3%); **Moderately abundant (3%–10%); ***Abundant (> 10%).

a Montmorillonite/illite mixed layer.
micas, give off potassium that is consumed in the conversion of smectite to illite,
and by producing cementing agents, as will be discussed in Chapters 4 through 6.
Table 3.3 illustrates the mineral composition of a number of bentonites that have
been investigated with respect to their physical properties in current research work.
MX-80 is the brand that has been particularly well investigated in laboratory and
field experiments (cf. Chapters 5 and 8).
Quartz has been suggested as a suitable component of smectite buffers since it
significantly increases the thermal conductivity. However, at pH > 8, which may well
be reached in mixtures of Na bentonite and quartz, the solubility of silica increases
significantly (i.e., from 3 mol/L to more than 6 mol/L); silica released from the buffer
close to a hot canister in a KBS-3-type repository will migrate toward the cold rock
where it precipitates and causes cementation.
3.6 ROLE OF CLAY PARTICLE CHARGE

3.6.1 Basics and Practicalities
Simple tests show that smectite particles, commonly called quasicrystals, carry a net
negative charge, mirroring the electric conditions in the particles but is a macroscopic
phenomenon. In contrast, the crystal lattice charge is an intrinsic property of smectites
that is conventionally explained by substitutions in the tetrahedral sheets and vacan-
cies in the sheet of individual lamellae. This charge is net negative and assumed to be
balanced by the interlamellar cations that are exchangeable. Smectites with low lattice
charge have smaller cation exchange capacities than those with high charge. Uptake
of water in interlamellar positions causes more swelling for l ow-charge smectites than
for those with high charge. Existing data also s uggest that low-charge smectites form
stronger gels and more viscous suspensions than high-charge smectites.
30 Bentonite Clay
The matter of particle charge is not at all of just scientific interest. It has played a
most important role in practical geotechnical engineering in World War II coupled
to the name of Alfred Neubauer, colonel and appointed head of an organization of
fortification that had the responsibility to construct, among many other things, sites
for harboring German submarines in just-occupied Norway. Some of the sites were
located in clay-rich areas, requiring dredging for providing sufficient water depth, and
this caused problems typical of engineering geology. The major difficulty was insuf-
ficient stability, causing collapse of quays, and slope failure. Sheet piles hammered
down to considerable depth were moved by sliding clay and comprehensive piling did
not help. Here is where Casagrande’s invention provided assistance (Peck et al. 1960).
It is fundamental to soil mechanics that the strength of clay drops with increased
water content and increases if the water content is reduced. Arthur Casagrande found,
using earlier research results of colloid chemists, that direct electrical currents cause
migration of salt ions and thereby of water dragged by them, are excellent for drain-
ing clay soils and that their strength is especially strongly increased if the electrodes
are of iron or aluminum. His brother Leo developed the technique and applied it very
successfully in Norway in the 1940s, as shown in Figure 3.4. The brilliant technique
used on a large scale after the war is termed electro-osmosis. The detailed stabilizing
mechanisms are a bit more complex and require the introduction of chemical processes.
Here, we will confine ourselves to the strengthening effect of the reduced water content.
When the German navyn left Norway in 1944, the electrical current was turned
off and very shortly afterward, the quays and shelters collapsed.
A colleague officer of Alfred Neubauer, Major Peter Winkler, was a very well-
trained and qualified military engineer, who got the responsibility to stabilize clay
slopes adjacent to railway tunnel adits. He was ordered by the German Wehrmacht
commander to investigate if electro-osmosis could be used for preserving potentially
Cathode, Anode, − −
negative positive + +
Perforated Perforated
tubes tubes
Submarine
Stabilized mass
Clay Clay
Positively charged ion migrating

Clay
toward the cathode bringing
water with it
Sheet piles
FIGURE 3.4 Electro-osmosis: principle and application of the technique to Norwegian

submarine harbor.
Migrated water pumped out from

cathodes of perforated pipes
Sloping ground
Potential surface
slip
surface
+
Bars
Clay mass
stabilized by Railway tunnel
dewatering,
water migrates
as shown by
the arrows
FIGURE 3.5 Railway tunnel in clay kept stable by reducing the water content using electro-
osmosis. Two rows of electrodes were installed, the lower consisting of positively charged
iron bars and the upper consisting of negatively charged, perforated cathode pipes from
which water was pumped out.
unstable clay slopes, and he certainly managed. The installation is illustrated in

Figure 3.5, which shows how slip would take place along a curved plane if the shear
strength could not be increased. As for the submarine basins lowering the water con-
tent by directing groundwater to the row of perforated cathode pipes and pumping
it out by them, the slopes were kept stable. When the current was turned off at the
withdrawal of the German troops, the slopes failed.
In the Norwegian cases the current was DC, with a potential of 0.1–2 volts/1 cm dis-
tance between the electrodes requiring an electric potential of some 100–1000 volts.
3.6.2 CEC, Anion Exchange Capacity, and Specific Surface Area

The type of adsorbed cations of smectite determines the degree of hydration and
affects the rheological properties (Sposito 1984). Maximum expandability is obtained
for Li and Na as adsorbed cations, whereas it is at minimum for Ca and polyvalent cat-
ions. Buffer clay for isolating canisters with highly radioactive waste in repositories is
therefore suitably clay of natural Na type or Ca smectite treated with Na carbonate on
an industrial scale for bringing it to Na form. The type of cations in natural smectite
32 Bentonite Clay
TABLE 3.4
Ion Exchange Capacity and Surface Area of Common Soil Minerals
CEC and AEC Surface Area Total Surface Area, Including
Mineral (meq/100 g) (external) (m2/g) Interlamellar Surface Area (m2/g)
Smectite 80–150 (CEC) 80–150 400–800
Finely milled mica 5–50 (CEC) 3–20 5–30
Chlorite 10–40 (CEC) 5–10 10–20
Kaolinite 3–15 (AEC) 5–10 10–20 (200a)
Illite 10–40 10–40 20–50
Finely milled.
a
AEC, anion exchange capacity.
clays reflects the conditions under which they were formed and what changes they
have undergone. One understands that the electro-osmotic stabilization is associated
with migration of iron cations into the clay to be stabilized and that this ion exchanged
the initial sodium cations in the old marine sediments, thereby providing additional
strengthening. Table 3.4 gives typical CEC and surface area values.
3.6.3 Mechanisms in Cation Exchange

Adsorbed cations can be exchanged according to the generally valid rule that the
higher the valence of the cation, the stronger is the replacing potential, except for
hydrogen that has a strong ability to replace Ca and several other bi- or polyvalent
cations. Potassium plays a special role in this respect because of its small size: 2.66Å,
which makes it fit in the cavities of the oxygen layer according to the Hofmann,
Edelman, and Forslind crystal models and become fixed there. From a practical
point of view, it is important to know that Ca readily replaces Na but that the concen-
tration of Na in the porewater must be much higher than that of Ca for replacing Ca
by Na. Coming back for a moment to the Norwegian electro-osmotic projects, it is
obvious that the polyvalent iron cations easily replaced the initially present sodium
ions and created stronger interparticle bonds.
Table 3.5 gives typical examples of the dominant adsorbed cations and of the CEC
of a number of bentonites (Pusch 2002). The table shows that Wyoming bentonite in
its natural form can have a significant fraction of the exchange sites occupied by Mg
and Ca, although Na dominates. The table shows that the CEC is strongly related to the
content of montmorillonite and is, therefore, an indication of the content of this mineral.
3.6.4 Role of Anions
It is commonly believed that anions cannot enter the interlamellar space for charge
reasons, nor be located at the basal surfaces of stacks of lamellae. They can, how-
ever, be coupled to the edges of stacks, which can have positive or negative electrical
charge depending on the pH of the porewater. Under acidic conditions, the edges
can attract anions but since the surface area of the edges is not more than about
TABLE 3.5
Adsorbed Cations and Cation Exchange Capacity in meq/100 g of Commercial
Bentonites
Material Na+ Ca2+ K+ Mg2+ Fe Al3+ CEC (meq/100 g)
Bel Fourche 18.99 36.96 4.92 20.70 0.38 Traces 81.97
Wyoming (blue)* 41.64 29.94 8.09 30.00 0.29 Traces 109.67
Wyoming (yellow)** 32.71 22.30 8.65 24.54 0.21 Traces 88.20
Milos – 52.05 3.90 27.47 – – 81.35
Erbslöh – 105.76 3.55 66.41 – – 71.07
Friedland 11.00 4.00 34.00 2.00 0.50 – 40.00
Czech Republic (RMN) – 46.45 11.53 47.26 0.21 – 62.34
*MX-80 type; **MX-80 type, oxidized.
10%–20% of the total area of the stacks, its role in anion exchange is small. For very
finely milled kaolinite, it can be practically important, however, and be utilized for
catching negatively charged ions such as sulfates, phosphates, nitrates, and chlorides.
The possibility of sorbing radioactive iodine from leaking waste containers can be of
practical importance and can be carried out by letting the most canister-near part of
the buffer consisting of such powder in compacted form.
The most obvious role of the anion exchange capacity is for supplying certain
nutrients to plants. This is of no interest to us since we are not dealing with macro-
biology; however, we are certainly estimating what role nutrients can play in sup-
porting existing and invading corrosion-promoting microbes in clay-surrounding
canisters with highly radioactive waste (Pedersen and Ekendahl 1990).
3.6.5 Phosphorus and Nitrogen

Phosphorus and nitrogen are nutrients to bacteria, which can form organic col-
loids with ability to transport radionuclides, and can, therefore, be of importance.
Colorimetric determination has shown that Wyoming bentonite contains about
100 ppm phosphorus and 200 ppm nitrogen (Pusch 2002). Commercial benton-
ites stockpiled for drying on the ground surface may be contaminated by various
sources, such as cow feces.
3.6.6 Sulfur
Sulfur-bearing minerals are always present in bentonites mostly in the form of sul-
fides, primarily FeS, and are aggressive corrodants of canisters of copper and Navy
Bronze. Table 3.6 shows the content of S (total) and of sulfides in some bentonites.
Experience shows that heating to 425°C of air-dry material for 15 hours significantly
reduces the sulfide content. According to SKB’s original quality indexing, the con-
tent of (total) S was set at 2000 ppm. We see from Table 3.6 that some of the buffer
candidate materials would have to be processed to bring their S content down.
34 Bentonite Clay
TABLE 3.6
Total Sulfur Content and Content of Sulfides in Some Commercial Bentonites
Sulfides Material Heated to a Few
Material Total S (ppm) (Untreated Mtrl) (ppm) Hundred Centigrades (ppm)
MX-80 2700–3200 1,000 150
Wyoming bentonite 300–1700 700 0
Erbslöh (Na-treated) 80–130 – –
Friedland (mixed layer 650 10,000 –
Milos (Ca-form) 5600–6600 – –
Source: Pusch, R. The Buffer and Backfill Handbook, Part 1, Swedish Nuclear Fuel and Waste
Management AB, Technical report TR-02-12, SKB, Stockholm, Sweden, 2002.
3.6.7 Organic Content
Like phosphorus and nitrogen, the organic content of buffers may serve as a nutrient to
bacteria and produce organic colloids. For common commercial bentonites, it has been
determined using a number of techniques, and it has been concluded that the CO2-gas
method gives results that are similar to those of more sophisticated techniques such as
infrared and spectrophotometric analyses and hence sufficiently accurate. For MX-80,
it has been found that the organic content of untreated material ranges between 475
and 750 ppm, whereas heating to 425°C for 15 hours reduces this concentration to
150 ppm (Pusch 2002). Analyses of organic substance extracted from MX-80 have
shown that some form of kerogene dominates but that fulvic and humic acids are also
present. This distinction may have an impact on potential bacterial activity.
3.6.8 Special Role of Organic Elements

3.6.8.1 Bacteria
It has been demonstrated that bacteria and virus can migrate through dense smec-
tite clay despite the assumed limited degree of continuity of voids in matured
canister-embedding buffer clay. The more recent microstructural models that have
been proposed and that will be described in Chapters 4 and 9 imply that artificially
prepared buffer clay contains numerous channels that has clay gels of low density.
The high porosity and interconnectivity of the channels provide possibilities for cor-
rosion-promoting bacteria to migrate in the clay on a large scale and to quickly reach
the canisters (Pedersen and Ekendahl 1990). The effect of strong gamma radiation
and a temperature of about 100°C has not been fully examined yet but would imply
a smaller risk of bacterial invasion and attack.
3.6.8.2 Identification of Smectite Minerals
We cannot conclude this chapter without considering the role of certain organic sub-
stances for identifying smectites and other clay minerals and for further consider-
ation of the crystal constitution of smectites. The basic findings were made early in
the twentieth century in the United States and in conjunction with the development
TABLE 3.7
Characteristic Staining of Clay Minerals
Acid-Treated Acid-Treated Complementary
Strong Hydrochloric Strong Hydrochloric Technique for
Untreated Acid Acid Certification
Clay Mineral Benzidine Safranine Y Malachite Green

Kaolinite No reaction No reaction No reaction SEM or high-
resolution optical
microscopy
Montmorillonite Purple-blue Purple-blue Yellow-red yellow CEC and XRD
Nontronite Blue-green No reaction No reaction XRD
Illite No reaction No reaction No reaction XRD
Palygorskite No reaction No reaction No reaction XRD, SEM, or
high-resolution
optical microscopy
Source: Mielenz, R.C. et al., Staining tests, Report 7, American Petroleum Institute Project 49, Columbia
University, New York, 1950; Grim 1968.
of the branch of organic chemistry in Germany after World War I. The work on
staining tests for clay mineral identification, in fact, started in the late 1880s and
led to the characteristic staining agents and techniques summarized in Table 3.7
(Furukawa and Brindley 1973). Methylene blue has been commonly used in applied
geology and geotechnique for estimating the content of smectite, but it is not of
interest in the present context (Hofmann and Dammler 1969; Higgs 1986). XRD and
CEC analyses are unbeaten technologies in this respect.
The color reactions listed in Table 3.7 have been explained as acid–base reactions
in which the clay serves as an acid, and as oxidation/reduction in which ions such as
ferric iron in the octahedral crystal layer oxidizes the reagent. The color reactions
can be masked by other minerals, and the applicability of staining tests is, therefore,
restricted (Grim 1968; Dodd and Hambleton 2006).
A color test of special interest was proposed by Dodd and Hambleton (2006).
It was based on the adsorption and oxidation of certain amino phenols on clay
mineral surfaces for routine identification of the clay montmorillonoid group, illite
(hydrous mica), and kaolins. The test procedure involves use of alcohol solutions of
the phenols of varying concentrations, with subsequent treatment with hydrochlo-
ric acid. Colors are claimed to be observed of both wet and dry clays and tests
performed were reported to be more accurate than those referred to in Table 3.7 for
members of the montmorillonoid family, as well as for the illite and kaolin groups.
Other literature references claim that aniline dye staining tests are most useful to
confirm the presence of kaolin minerals.
The various observations and attempts to explain the process that gives color
reactions has a bearing on the validity of the montmorillonite crystal models. Thus,
if coloring is caused by oxidation involving surface oxygens, it would be in favor of
36 Bentonite Clay
the Hofmann model, whereas an acid–base reaction in which the clay serves as an
acid would be in principle agreement with both Edelmann’s and Forslind’s models.
The matter is utterly complex and worth additional investigations. High-resolution
optical microscopy would be an ideal technique for identifying the adsorption of
organic tracers on smectites.
3.7 CONCLUSION
The understanding of the constitution of smectite minerals in general and of mont-
morillonite and saponite in particular is incomplete, especially with respect to sorp-
tion and exchange of organic cations and cationic molecules. No wonder so little
research and development has been made on the function of smectite clays used as
engineered barriers to the migration of hazardous cations and anions released from
chemical and particularly from radioactive waste! We will outline some models of
such interaction in the book.
REFERENCES
Dodd, C.G., Hambleton, W.W. 2006. A qualitative color test for rapid identification of the clay
mineral groups, Economic Geology, U.S. Geological Survey, Reston, VA.
Edelmann, C.H., Favejee, J.C.L. 1940. On the crystal structure of montmorillonite and hal-
loysite. Zeitung fuer Kristallographie, 102:340–348.
Forslind, E., Jacobsson, A. 1972. Clay-water systems, in Water a Comprehensive Treatise,
F. Frank (Ed.), Plenum Press, New York.
Furukawa, T., Brindley, G.W. 1973. Adsorption and oxidation of benzidine and aniline by
montmorillonite and hectorite. Clays and Clay Minerals, 21:279–288.
Grim, R.E. 1968. Clay Mineralogy. McGraw-Hill, New York.
Higgs, N.B. 1986. Studies of methylene blue adsorption as a method of evaluating degradable
smectite-bearing sands. Cement and Concrete Research, 16:524–534.
Hofmann, U., Dammler, J. 1969. Die Methylene-Blue sorption on montmorillonite. Chimia
23:476–480.
Hofmann, U., Endell, K., Wilm, D. 1933. Crystal structure and swelling of montmorillonite.
Zeitung fuer Kristallographie, 86:340–348.
Mielenz, R.C., King, M.E., Schieltz, M.C. 1950. Staining tests. Report 7. American Petroleum
Institute Project 49, Columbia University, New York.
Peck, R.B., Casagrande, A., Bjerrum, L., Skempton, A.W. 1960. From Theory to Practice in
Soil Mechanics. Wiley, New York.
Pedersen, K., Ekendahl, S. 1990. Distribution and activity of bacteria in deep granitic ground-
waters of southeastern Sweden. Microbial Ecology, 20:37–52.
Pusch, R. 2002. The Buffer and Backfill Handbook, Part 1. Swedish Nuclear Fuel and Waste
Management AB, Technical report TR-02-12. SKB, Stockholm, Sweden.
Sposito, G. 1984. The Surface Chemistry of Soils. Oxford University Press, Oxford.
Yong, R., Pusch, R., Nakano, M. 2010. Containment of high-level radioactive and hazardous
solid wastes with clay barriers, Spon Press, London; New York.
4 Clay–Fluid–Gas Systems
This chapter deals with migration of dissolved matter and flow and diffusive transport
of porewater solutions as well as with gas movement in buffers and backfills. They are
basic to THMC performance (Trepresenting temperature, H representing hydraulic
performance, M representing mechanical, that is, rheological behavior, and C repre-
senting chemical function). Focus is on the meaning and determination of hydraulic
conductivity, expandability, compressive strength, and ion diffusivity. These properties
determine the rate of water, gas, and ion transport through clay seals.
4.1 OVERVIEW
From design and performance assessment of clay-based barriers to leakage of contami-
nated water from repositories for hazardous waste the following properties are important:
• Fluid penetration and hydraulic conductivity

• Gas penetration and conductivity
• Diffusion of ions and fluids
In designing clay seals, the problem of predicting flow through them is of funda-
mental importance. Unfortunately, the understanding of the mechanisms related to
penetration of water and gas is incomplete, and significant mistakes in determining
the penetrability in laboratory and field tests have caused false conclusions through
the years. A major reason for this is that geotechnicians pay too little attention to the
role of the hydraulic gradient: use of very high gradients for saving time in laboratory
testing causes microstructural changes that can lead to substantial underestimation of
the hydraulic conductivity. On the other hand, overestimation can also occur and lead
to unnecessarily expensive technical solutions because the gradients are in practice
often lower than those one can use in testing. The dilemma is hence that the very low
gradients that prevail in practice cannot be applied in laboratory testing for time and
cost reasons and that the very high gradients used in laboratory testing give irrelevant
and nonconservative results. These issues are dealt with in this chapter, in addition to
a number of practical ones, such as the role of temperature.
As to gas permeation, the problem is similar to that of comprehending the hydrau-
lic performance of smectite clays but different in the sense that gas is held back by a
clay seal until a critical pressure is reached whence it quickly makes its way through
the clay provided that the rate of gas production is maintained. The conditions for
breakthrough depend on the microstructural constitution in a similar way as water
penetration takes place, but the very low viscosity of gas makes it flow much quicker.
Once gas has made its way through the clay, it continues to do so at an increasing rate
because of erosion. All this implies that the role of clay microstructure is understood,
a matter that is in focus throughout this book.
37
38 Bentonite Clay
Regarding ion diffusivity, the conditions are different. Here, the concentration
gradient is the driving force and causes migration of dissolved matter that follows
different paths depending on the microstructural constitution and on the resistance
to migration caused by the different microstructure-affecting properties of dissolved
components, such as Na and Ca.
We will consider here the three transport properties of porewater: flow, pore dif-
fusion, and surface diffusion with respect to the microstructural constitution of the
clay, especially concerning the difference in viscosity of different fractions of the
porewater. They all depend on pressure or concentration gradients, which are often
quite different in laboratory testing and in practice.
4.2 PHYSICOCHEMICAL BACKGROUND

4.2.1 Energy Considerations, Soil–Water Potentials
Before going to practical issues, we should have a quick look on what thermodynam-
ics tells us about water potential, which is the energy of water per unit volume relative
to pure water at rest, that is, under no external pressure or suction. It is a measure of
the tendency of water to move from one spot to another due to chemical forces, osmo-
sis, pressure, or matrix effects such as capillary action. Water potential is typically
expressed in potential energy per unit volume and commonly represented by the Greek
letter ψ. For porewater in smectite clay, the water potential integrates various potential
drivers of water movement, which can operate in the same or different directions.
The electrical charge and colloidal size of clay particles make them hydrate and
interact such that the hydraulic conductivity and stress–strain properties differ from
those of other soils (Yong and Mohamed 1992; Nagy et al. 2010). This is particularly
important for smectites because their net negative charge is higher than that of any
of the other clay mineral type and because the particles are pseudocrystalline and
absorb water within the crystallites. The very fine-tortuous pore systems offer con-
siderable resistance to water and gas flow, especially at higher densities, and cause
capillary suction of water with lower potential such as silty sand, which is much less
hydrophilic. In addition to pressure gradients, gradients in salt concentration and
electric potentials in smectite clay control its physical performance. The introductory
part of this chapter summarizes some basic facts that are required for understanding
these matters and for utilizing smectite clay for preparing seals.
In summary, the energy status of the clay–water system governs the develop-
ment of matric and osmotic forces that are responsible for hydration and dehydration
(Yong et al. 2010). In general, one can explain porewater movement as the net effect
of the thermodynamic terms (Yong et al. 2010):
ψ = ψπ + ψ p + ψm (4.1)
where:
ψ is the total potential
ψπ is the solute potential (ψπ = nRTc, where n is the number of molecules per mole
of salt, R is the universal gas constant, T is the absolute temperature, and c
is the concentration of solutes)
Clay–Fluid–Gas Systems 39
ψp is the pressure potential (porewater pressure or underpressure)

ψm is the matric (matrix) potential (surface tension)
Considering first smectite clay as a soil–water continuum, one can calculate the spac-
ing between the particles and between interparticle lamellae, and the effective pressure
in consolidation tests. One example of this is shown in Figure 4.1, which refers to Na
montmorillonite saturated with 0.003 molar NaCl solution, and which demonstrates that
the effective stress has an enormous impact on the numbers of hydrate layers in the
stacks of lamellae. On increasing the effective consolidation pressure to about 100 MPa,
there is just one monolayer left in the interlamellar space; for higher pressures as well,
this one is squeezed out, causing direct contact between the individual lamellae
(cf. Chapter 3). As to the various potentials, it is generally accepted that the compres-
sion/swelling performance is governed by the ψp potential and that, at complete fluid
saturation, the matric potential ψm = 0. These potentials are averages only and particle
spacing and density variations on the microstructural scale need to be considered for
defining the true thermodynamic conditions in the clay matrix (Yong et al. 2010). We
will return to the meaning and determination of the swelling pressure in Chapter 5.
The coupling of activities and energies in the microscopic and macroscopic voids
determines the hydration and dehydration of clays. Water uptake into an air-dried
smectite particle is initially via hydration forces and subsequently via diffuse double-
layer forces (Pusch and Yong 2006). The energy characteristics defined by the soil–
water potentials indicate how strongly water is held to the soil. On wetting, the soil
potentials change but at final equilibrium, the osmotic or solute potential ψπ of the
microstructural unit, that is, an aggregate of associated stacks of lamellae, will be
equal to the matric potential ψm, which is a characteristic of the macropores and
developed by the microcapillaries, van der Waals forces, osmotic phenomena, and
other surface active forces existent in the macropores.
The potentials are intrinsic and become obvious by any change in volume, tem-
perature, and chemical conditions. Thus, an element of fully water-saturated smectite
clay confined in a cell for becoming hydrated under constant volume conditions is
ultimately in equilibrium with the confinement and the contacting external water.
It exerts a swelling pressure but neither takes up nor gives off porewater. However,
on expanding the space occupied by the clay, it immediately reacts by a change in
70
Separation distance, Å
60
50
40
30
20
10
0
10 1,000 100,000
Pressure in kPa
FIGURE 4.1 Pressure versus interparticle separation in Å for 0.003 molar NaCl porewater
in montmorillonite. (After Yong.)
40 Bentonite Clay
potentials: the clay swells in conjunction with water uptake by which the expanded
clay exerts a lower swelling pressure on the confinement. On compression, the oppo-
site reaction is noted, with hysteresis.
4.2.2 Hydration of Smectite Clay

A number of measurements indicate that no more than three interlamellar hydrate
layers can be established in Na montmorillonite, and we see from the example in
Table 4.1, representing a relatively soft clay, that this number is logically related to
the water content for a given dry density (Pusch and Yong 2006; Nagy et al. 2010).
For water-saturated montmorillonite with higher bulk density than about 1200 kg/m3,
the number of interlamellar hydrates is lower. Thus, that is, for a bulk dry density of
1590 kg/m3 (density at water saturation 2000 kg/m3), the average number is 2, and for
a dry density of 1920 kg/m3, corresponding to 2210 kg/m3 at saturation, the average
number is 1.
It is a remarkable fact that so much water is located in the interlamellar space.
Thus, even for the soft Na montmorillonite clay with only 1200 kg/m3 dry density,
the fraction of interlamellar water is about 80% of the total porewater content at
100% water saturation and even higher for lower degrees of saturation because of
the dominating hydration potential of the interlamellar space. This is of particular
importance in wetting of initially dry clay. Water entering by capillary suction in
the channels formed by interconnected voids migrates quickly into the dry clay and
becomes redistributed from them to finer systems of microvoids, from where it is
sucked by the tremendously hydrophilic stacks of lamellae. On drying, these are the
last to give off water, a process that mobilizes capillary suction for retaining as much
water as possible. Transport of porewater out from them takes place by evaporation
via wider voids and channels. Figure 4.2 illustrates the distribution of water in drying
soil, the solid phase representing dense hydrated stacks of lamellae in smectite clay.
Capillary forces between the particles and particle aggregates cause tension that
greatly affects the macroscopic rheological behavior, especially the shear strength
and strain: nearly dry clay can be as stiff as hardened concrete.
TABLE 4.1
Number of Hydrate Layers in Na Montmorillonite with a
Dry Density of 1200 kg/m3 (1760 kg/m3 at Saturation) for
Saturation with Distilled Water
Water Content Degree of Water Number of
(weight %) Saturation (%) Interlamellar Hydrates
<7 <20 0–1
7–10 20–30 1
10–20 30–60 1–2
20–25 60–80 2
25–35 80–100 2–3
>35 100 3
Liquid phase:
water and
dissolved air
Gas phase: dry air Solid

and water vapor phase
FIGURE 4.2 Schematic representation of an unsaturated smectite clay element. The solid
phase represents mineral particles that are made up of stacks of lamellae with strongly
bound hydrate layers in between. (Data from Pusch, R., Yong, R.N., Microstructure of
Smectite Clays and Engineering Performance, 1st edn., Taylor & Francis, London, 2006.)
Initial water uptake by dry smectite clay is strongly exothermic because the
water is firmly held in the coordination sphere of the cations in the interlamellar
space and in contact with the surface oxygens (Pusch and Yong 2006). Interlamellar
(interlayer) expansion due to sorption of water is determined by the crystal layer
charge, interlayer cations, properties of the adsorbed water or solution, and par-
ticle size. Volume changes by as much as 100% of the original volume of dry
smectite clay can occur when all the interlamellar hydrate layers are established
under the prevailing geometrical restraint represented by confining test cells such
as oedometers.
The sequence of water uptake depends on the nature of the exchangeable cations
in the interlamellar space and on the initial water content of only partly saturated
soil. For swelling clays containing Li or Na as exchangeable cation, continued
uptake beyond the maximum hydration of the interlamellar space takes place on
the basal surfaces of the stacks of lamellae (Figure 4.3) and by hydration of exfoli-
ated single particles consisting of one to three lamellae (Pusch and Yong 2006).
Interaction of such units and of larger stacks of lamellae takes place according to
electrical double-layer theories, implying that the solvation shell of cations plays
a major role. For small monovalent cations, the shells consist of about six water
molecules if the solution is dilute (Sposito 1984), and about three for concentrated
solutions. In the case of divalent cations, the solvation shells move together as a
complex, explaining the different hydration potentials of Na and Ca smectites:
the first mentioned can have three interlamellar hydrates but the other only two
(cf. Chapter 3).
42 Bentonite Clay
Negatively charged pore wall
+ + + + + + + +
+ + + + +
+ + + +
+ +
− −
+
+ +
+
Sorbed
− −
phase + + −
−
−
+ + + + + +
+ +
+ + + + +
+ + + + + + + +
Negatively charged pore wall
FIGURE 4.3 Electrical double layers on basal surfaces of smectite particles exposed in
voids. At small interparticle distance, the particles repel each other and limit the transport
of cations in flowing along water and thereby also of the water itself. (After Neretnieks, U.,
Moreno, L., J. Cont. Hydrol., 14, 163–182, 1993.) (Compare Pusch, R., Yong, R.N.,
Microstructure of Smectite Clays and Engineering Performance, 1st edn., Taylor & Francis,
London, 2006.)
4.3 HYDRAULIC CONDUCTIVITY

4.3.1 Definition of Hydraulic Conductivity
Darcy’s law is applied for the evaluation of ordinary percolation tests:
q
v= = K ⋅ i (4.2)
A
where:
v is the average flow rate (m/s)
q is the amount of percolated fluid per time unit (m3/s)
A is the cross section of percolated volume (m2)
K is the hydraulic conductivity (m/s)
i is the hydraulic gradient, that is, the pressure drop over the length of the perco-
lated sample (m/m)
The hydraulic conductivity K is related to the (intrinsic) permeability k according to

Equation 4.3:
Kμ
k=
ρg
(4.3)
k ⋅ρ⋅ g
K=
μ
where:
μ is the dynamic viscosity (kg/m·s)
ρ is the fluid density (kg/m3)
g is the acceleration by gravity (m/s2)
Another common term in groundwater hydraulics is the transmissivity, which means

the rate at which water is transmitted through unit width of geological material under
unit hydraulic gradient. It is defined as in Equation 4.3 and expressed in meters
squared per second:
T = Kb (4.4)
where:
b is the thickness of the sample or layer
4.3.2 Determination of the Hydraulic Conductivity in the Laboratory

The bulk hydraulic conductivity is determined in the laboratory using permeameters
or oedometer cells, as indicated in the schematic picture in Figure 4.4. Evaluation
of the tests is simple but the values obtained represent the average conductivity for
the particular pressure and pressure gradient, fluid (solution), and temperature. The
microstructural channeling means that there is in fact a large variation in flow rate
and hydraulic conductivity on the microscale, meaning that tracer tests can indicate
concentrated flow through certain parts of a sample and no flow in others. The test
Cross-sectional area
of sample = A
Permeating fluid applied
under constant head
Hydraulic gradient i defined by

constant head of permeating fluid
and height of test sample
Q = kiA for specified time

interval
Permeability coefficient k is
computed using measurements of Q
Calibrated and A under known condition of i
collection and specified time interval.
flask
Quantity Q of permeating fluid
collected in known unit time
FIGURE 4.4 Constant head hydraulic conductivity testing for obtaining test values of q
under specified constant head and time interval. Different constant water pressures are often
applied at the ends of the sample, which establishes a hydraulic gradient. (Data from Pusch, R.,
Yong, R.N., Microstructure of Smectite Clays and Engineering Performance, 1st edn., Taylor &
Francis, London, 2006.)
44 Bentonite Clay
arrangement with water inflow from above in the figure is for pedagogic reasons.
In practice, inflow is usually from below, which is necessary for saturation of i nitially
not fully saturated clay samples in order to get air out.
From a practical viewpoint, one needs to be aware of the risk of leakage along
the oedometer walls and to consider the risk of precipitation of dissolved salt that
can cause clogging of microstructural channels. Thus, flow experiments with very
salty solutions should be followed by several percolation cycles with distilled water
if the purpose is to find out what impact that the salt percolates have had on the
microstructure.
4.3.3 Microstructural Implications

A major factor that influences the rate of water flow through soils is the size of the
voids, as we saw in Figure 4.3. The smaller the particles, the smaller the voids and
the stronger the flow resistance for a given density. This is valid for any soil but plays
an especially important role for smectite-rich clays. Before going deeper in consider-
ing the impact of pressure gradients and other difficulties in the proper conduction
of hydraulic tests, we will return to the microstructural constitution of smectite clays
in Chapter 3 and pay particular attention to the practically important flow paths, that
is, the microchannels (cf. Figure 4.5).
A large number of microstructural analyses indicate that the diameter of the wid-
est channel is in the order of 50 μm in smectite-rich clay with a bulk density at satu-
ration of 1570 kg/m3, 20 μm in clay with 1850 kg/m3 density, and 5 μm in clay with
2130 kg/m3 density.* They can be open or contain soft clay gels. Figure 4.6 shows
a small part of the channel network and illustrates that the channels have different
diameters (Pusch et al. 2001c).
FIGURE 4.5 Typical channel with about 5 μm aperture in dense smectite clay, gel particles
rinsed by percolation of water. (Courtesy of Kasbohm.)
* These densities are used as references in this book.

Channel
Capillaries
Node
FIGURE 4.6 Channel network mapped as a cubic grid with gel-filled channels intersecting
at a node in the grid. (After Neretnieks, U., Moreno, L., J. Cont. Hydrol., 14, 163–182, 1993;
Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering Performance, 1st
edn., Taylor & Francis, London, 2006.)
The Darcy flow model does not consider the detailed mechanisms involved in
permeation, whereas modern flow models such as the 3Dchan code assume networks
of channels do so and make it possible to calculate a relevant bulk hydraulic conduc-
tivity K (Neretnieks and Moreno 1993). The model is based on the assumption that
a characteristic number of interacting channels, each containing a bundle of capil-
laries with soft gel fillings, are responsible for the through-flow. Using the Hagen–
Poiseuille law and a factor representing constrictions of the channels, one gets the
geometrical and conductivity data in Tables 4.2 and 4.3 for room temperature (Pusch
and Yong 2006).
The agreement between the theoretical and experimental K-values of course
depends on the relevance of the theoretical model and on the accuracy of the experi-
mental data, which are backed up by several series of tests of the same and similar
materials. We see that for electrolyte-poor Na- and Ca-rich water there is good agree-
ment between experimental and predicted values; the ones for water with the oceans’
salinity are roughly 10 times higher than for percolation with distilled water because
of the earlier described microstructural differences (cf. Chapter 3). The experiments
were made with hydraulic gradients ranging from 30 to 100 m/m, whereas earlier
TABLE 4.2
Number of Differently Sized Channels per
250 × 250 μm2 Cross-Sectional Area of Montmorillonite-
Rich Clay (MX-80)
Number Number Number
Bulk Density of 20–50 μm of 5–20 μm of 1–5 μm
(kg/m3) Channels Channels Channels
2130 0 0 135
1850 0 10 385
1570 2 85 950
46 Bentonite Clay
TABLE 4.3
Hydraulic Conductivity (K ) of Smectite-Rich Clay
Prepared by Compacting MX-80 Powder and
Saturating It with Electrolyte-Poor Water under
Hydraulic Gradients of 30–100 m Water Pressure
Difference per Meter Flow Length (m/m)
Density at Water Experimental
Saturation (kg/m3) Theoretical K (m/s) K (m/s)
2130 E–13 E–13
1850 5E–12 3E–12
1570 6–11 7E–11
investigations using gradients ranging between 3,000 and 10,000 m/m gave 5–10
times lower conductivities. This matter is of course of considerable practical impor-
tance and deserves closer examination.
4.3.4 Role of Hydraulic Gradients

4.3.4.1 Test Principles and Theory
Very careful measurement of the hydraulic conductivity of fine-grained clays has
shown deviations from Darcy’s law. Thus, for very low hydraulic gradients, there
may be practically no flow because of the quasicrystalline structured water molecule
networks in the interlamellar space and within a few nanometers distance from the
basal planes of smectite particles in the clay matrix and gels in the permeable chan-
nels. When the gradient is increased, breakdown of the water films and new flow
paths are caused by drag forces generated by the water flow, hence increasing the
evaluated conductivity. By raising the gradient further, it is reduced because trans-
port of eroded particle aggregates causes clogging (Pusch et al. 2011), and because
the void space is reduced by consolidation under the compressive force exerted by
water.
The described effects were known decades ago and the American Society for
Testing and Materials (ASTM) therefore advises use of hydraulic gradients up
to only 30 m/m (termed here as icritical) for clayey soils for determination of the
hydraulic conductivity, (ASTM 1994). In geotechnical engineering, the common
practice is to apply high hydraulic gradients for quick determination of the conduc-
tivity, which can cause underestimation of this property. This has been evidenced
by performing numerical calculations using the 3Dchan model and applying fluid
dynamics and Monte Carlo simulations (Pusch and Weston 2003; Hellström 2009).
A further flow-controlling factor has been identified, that is, the migration of
eroded and transported particles collected in the filter at the outflow end of the per-
meameter causing clogging. We will see from the subsequently described experi-
ments that it can play a role for accurate evaluation of the hydraulic conductivity.
4.3.4.2 Experimental Proof of the Impact of the Hydraulic Gradient

The matter has been considered by various investigators (Cheung and Yu 2001;
Pusch 2001a,b) and has been investigated in detail in a study of smectite-rich
clay from Iraq that will be summarized here. It comprised several series of per-
colation tests of these Ca montmorillonites, which provide evidence of the impact
of hydraulic gradients in the range of 5–10,000 m/m (Al-Taie et al. 2013). The
investigation included tests of clay with dry densities of 1000–1500 kg/m3 using
distilled water and 3.5% CaCl2 solution as permeants. Two types of filters were
installed at the outflow end of the permeameters in order to check the impact of
the filter constitution on migrating particles and aggregates. Smectite-rich green
clay sampled from the Fatha formation of the Lower Miocene age was crushed and
thoroughly mixed and sieved giving the granular size distribution: 100%, <1 mm;
92%, <0.8 mm; 76%, <0.5 mm; 47%, <0.2 mm; and 27%, <0.1 mm. Wet sieving
showed that the clay fraction was 65%. X-ray diffraction analysis of randomly
oriented powder in air-dry and ethylene glycolated states showed that montmoril-
lonite in Ca form makes up to 60%–65% of the clay minerals (Al-Taie et al. 2013),
with illite and chlorite representing minor constituents. The liquid limit was 148%.
Skempton’s clay activity (ac) number was found to be 1.4 and the cation exchange
capacity 23.1 meq/100 g. These data points to an actual montmorillonite content
of 40%–60%.
A GDS apparatus of common type was used for the determination of the hydrau-
lic conductivity. Air-dry granules sized 10–1000 μm were compacted in 30 and
50 mm stainless steel oedometer cells after smearing smectite-rich clay (MX-80)
on the cell walls for minimizing leakage along the sample-wall interface. Two types
of filters were used: sintered stainless steel filters with void diameters of 10–75 µm,
and sand/silt filters prepared by static compaction of mixed sand and silt to 20 mm
thickness with a dry density 1600–1630 kg/m3. The porosity was 0.38 and the void
diameter 10–800 μm, as evaluated from applying theoretical packing theory. The
high density made these filters resist compression during saturation and percolation
of the compacted clay.
Two sets of about 20-mm-high samples were prepared: one saturated and per-
meated with distilled water and the other using 3.5% CaCl2 solution supplied from
the lower end of the cells and using very low backpressure. The hydraulic conduc-
tivity tests were conducted under different hydraulic gradients ranging from 5 to
10,000 m/m. Low hydraulic gradients (5–100 m/m) were applied by connecting
a water-filled vessel to the inflow end of the cells (Pusch and Yong 2006). High
gradients (>100 m/m) were achieved by use of the GDS digital controller with an
accuracy of discharge flow of ±1 mm3 and a capacity to provide pressures up to
2500 kPa. The respective gradients were maintained until stable flow was reached.
The relation between the gradient (i) and the conductivity (K) as functions of the
dry density of the clay samples is shown in Figure 4.7.
The left graph (Figure 4.7a) verifies that the hydraulic conductivity is at maximum
for hydraulic gradients lower than 1000 m/m and higher than about 30 m/m. The
right graph (Figure 4.7b) shows an increased evaluated conductivity for increased
gradients. One finds that both techniques gave approximately the same conductivity
48 Bentonite Clay
DW_SIN_1000 DW_SIN_1250 DW_SIN_1375 DW_SIN_1500

1E−06
1E−07
Hydraulic conductivity (K ), m/s
1E−08
1E−09
1E−10
1E−11
1E−12
1 10 100 1,000 10,000
(a)
DW_SS_1000 DW_SS_1250 DW_SS_1375 DW_SS_1500
1E−01
1E−02
1E−03
1E−04
Hydraulic conductivity (K ), m/s
1E−05
1E−06
1E−07
1E−08
1E−09
1E−10
1E−11
1E−12
1 10 100 1,000 10,000
(b)
FIGURE 4.7 The impact of the hydraulic gradient (i, m/m) on the hydraulic conductivity of
the green clay permeated with distilled water for different dry densities (the number following
DW_SIN or SS in kg/m3). Filter type at outflow end: (a) stainless steel filter and (b) sand/silt
filter. (Data from Al-Taie, L. et al., J. Earth Sci. Geotech. Eng., 3, 109–125, 2013.)
for the highest density and gradients of 100–1000 m/m. The evaluated high conduc-
tivities for lower densities indicate that erosion of the clay caused widening of the
channels and significant transport of fines into and through the sand/silt filter. For
dry densities 1000 and 1250 kg/m3, the highest gradient that could be applied was
100 and 700 m/m, respectively (cf. ASTM 1994; Nagy et al. 2010).
Taking the total cross section of the permeable channels in clay with a dry density
of 1500 kg/m3 to be 2% (cf. Chapter 3), percolation under a hydraulic gradient of
10,000 m/m corresponds to a flow rate of about E−4 m/s, which causes dislodgement
of 1 μm particle aggregates (Pusch and Yong 2006). For a dry density of 1000 kg/m3,
the channel cross-sectional area is about 4% and the flow rate about E−2 m/s, which
can erode even smaller particles (<0.5 μm). It is therefore not surprising that small
particles and aggregates can be torn off from the walls of the channels and trans-
ported to constrictions or to the fine-porous filter at the outflow end causing clogging.
For the experiments with sand/silt filter, the erosion of the softest clay led to micro-
structural collapse for higher gradients than 100 m/m.
A most important fact is the obvious drop in evaluated hydraulic conductivity for
very low hydraulic gradients. In many practical cases, such as smectite top liners
of landfills, the hydraulic gradient is commonly below 1 m/m, which likely gives
hydraulic conductivities that are lower by one order than for gradients used in ordi-
nary permeation tests, which, of course, provides extra safety. Such impact of the
gradient on the conductivity was first described by Hansbo (1973) and is illustrated
in Figure 4.8, which shows the relationship between outflow Q and the applied gradi-
ent i, for tests on low-swelling and moderately swelling soils. For smectites, the criti-
cal gradient is higher than for nonswelling clays. The direct proportionality between
flow rate and hydraulic gradient, specified by the Darcy model, is not obtained for
these clays when the hydraulic gradient falls below the critical gradient ic as shown
in Figure 4.8.
High electrolyte concentrations in permeating water are of particular importance
for the performance of bottom liners of landfills. We know from Chapter 3 and from
numerous published studies (cf. Pusch and Yong 2006) that saltwater increases the
Darcy hydraulic
conductivity model
>io
Flow rate, q
,i
i o)
–
k (i Low-to-
q= moderate
swelling soils
ki
q=
ic = critical gradient
io = apparent threshold gradient
0
io ic
Hydraulic gradient, i
FIGURE 4.8 Critical and apparent threshold gradients obtained in hydraulic conductivity
tests on low-to-moderate swelling soils. (Data from Hansbo 1973.)
50 Bentonite Clay
hydraulic conductivity, and this was also found to be the case for the Iraqi clays.
For samples with the dry densities 1000, 1250, and 1500 kg/m3 percolated with dis-
tilled water, the evaluated hydraulic conductivity rose by up to three to seven times
from the value for the gradient 35 m/m when it was increased to 300 to 600 m/m
(Figure 4.9) (Al-Taie et al. 2013). The impact of the hydraulic gradient using salt
water is somewhat different since the big salt-coagulated clay aggregates in the chan-
nels have fewer mutual bonds and are, therefore, more easily released and trans-
ported by flowing saltwater to positions where they cause clogging. However, this
flow-retarding effect is counteracted by the increased number and size of channels
caused by the salt-generated contraction of the stacks of lamellae in the entire clay
1E−04
CaCl2_SIN_1000 CaCl2_SIN_1250 CaCl2_SIN_1500
1E−05
Hydraulic conductivity (K), m/s
1E−06
1E−07
1E−08
1E−09
1E−10
1E−11
10 100 1,000 10,000
(a)
1E−04
CaCl2_SS_1000 CaCl2_SS_1250 CaCl2_SS_1500
1E−05
Hydraulic conductivity (K), m/s
1E−06
1E−07
1E−08
1E−09
1E−10
1E−11
10 100 1,000 10,000
(b)
FIGURE 4.9 The impact of the hydraulic gradient on the hydraulic conductivity of the green
clay permeated with 3.5% CaCl2 solution for different dry densities (the number following
CaCl2 _SIN or SS in kg/m3). Filter type at outflow end: (a) stainless steel filter and (b) sand/silt
filter. (Data from Al-Taie, L. et al., J. Earth Sci. Geotech. Eng., 3, 109–125, 2013.)
i = 10 i=5 i = 10
2.0⋅10−8
1.5⋅10−8
K (m/s)
1.0⋅10−8
0.5⋅10−8
0
0 20 40 60 80 100 120
Time (h)
FIGURE 4.10 Change in evaluated conductivity of clay-poor backfill (10% montmorillonite-

rich clay material resulting primarily from particle transport. Dry density 1200 kg/m3; distilled
water). The hydraulic gradient is termed i. (Data from Pusch, R., Yong, R.N., Microstructure
of Smectite Clays and Engineering Performance, 1st edn., Taylor & Francis, London, 2006.)
matrix. The reducing impact on the evaluated hydraulic conductivity is therefore less
obvious than for percolation with distilled water (Pusch 2001b).
It is concluded that the deviation from linearity in log–log diagrams of the hydrau-
lic conductivity versus the hydraulic gradient begins at a certain low critical gradient
(icritical) that depends on the clay content, clay mineral type, pore size distribution, and
type of permeant, as well as on the test arrangement. The study by Al-Taie indicates
that the critical gradient value 30 m/m recommended by ASTM (1994) can be raised
to 100 m/m in standard permeameter tests for providing conservative values. For top
liners of landfills, this gradient gives a margin of one order of magnitude on the con-
servative side.
The impact of changing the hydraulic gradient on the evaluated hydraulic conduc-
tivity of relatively soft smectite clay is illustrated by Figure 4.10, which also shows
the common tendency of successive tightening after each change because of rear-
rangement of clay gels and buildup of structured water films in the interlamellar
space and on the basal planes of the stacks of lamellae.
4.3.4.3 Piping and Erosion

Dispersion of expansive clays is known to be related to the so-called sodium adsorp-
tion ratio (SAR), which is defined as
Na
SAR = 1/ 2 (4.5)
⎡⎣( Ca + Mg ) / 2 ⎤⎦

where the chemical symbols represent the respective amounts of adsorbed cations.
The dispersibility of such clays is also related to erosion-related parameter (ESP):
52 Bentonite Clay
⎛ Na ⎞
ESP = ⎜ ⎟ ×100 (4.6)
⎝ Total exchange capacity ⎠
where the exchange capacity is expressed in milliequivalents per 100 g dry clay.
Very low SAR and ESP values (<5) are typical of nondispersive clays, whereas
higher values than 15 characteristically point to strong slaking. Experience from
pinhole tests shows that clays with sodium ions occupying more than about 10% of
the exchange complex exhibit dispersion provided that the electrolyte concentration
of the fluid is lower than E−3 N. Both these parameters imply that lesser the salt in
the groundwater, the higher the dispersibility.
4.3.4.4 Experimental Evidence

For understanding the practical importance of piping and erosion, we will consider
the case of water uptake from discrete fractures in the walls of tunnels filled with
air-dry smectite backfills, which is planned for underground repositories for highly
radioactive waste. Mock-up tests with relatively slow inflow of water from local
spots showed the evolution of water penetration in Figure 4.11, indicating piping.
The discharged water can be radioactively contaminated and imply loss of the seal-
ing potential of the fill by causing channeling, which makes the process of funda-
mental importance for the assessment of the risk of radioactive contamination of the
biosphere.
AB AB AB Test V1
1 1 1
CD CD CD
3 3 3 80%
EF EF EF
5 GH 5 GH 5 GH
7 IJ 7 IJ 7 IJ
9 KL 9 KL 9 KL
11 M 11 11 M
12 13 13
13 15 15
15 17 17
17 19 19
19 21 21
21 23 23
25 25 25
30% 30%
30%
50%
50% 50%
70%
70% 70%
80%
80%
(a) (b) (c)
FIGURE 4.11 Distribution of the water content at three vertical planes with 50 mm spacing in
unsaturated montmorillonite-rich clay caused by water inflow from one spot of 0.01 l/s showing
piping (fat broken line with arrows): (a) closest, (b) central, and (c) farthest. (Pusch et al. 2010)
The experiments gave important information on the wetting process of fillings

of smectite granules, which were rod-shaped with 10–15 mm length and 6 mm
diameter. The water content was 8%–15% and the dry density was about 900 kg/m3.
Water inflow took place from single spots simulating a fracture crossing in contact-
ing water-saturated rock. The major finding was that inflow rates up to 0.1 l/min per
inflow spot caused successive wetting of the fill close to the inflow spots by diffusion
and, ultimately, formation of channels oriented perpendicularly to the rock surface
causing erosion. Higher inflow rates gave quicker piping. The channeling process
takes place independently of the inflow rate but is slower and leads to higher critical
pressure if it is low.
Designers of backfills would learn from this that mapping of inflow spots in a
completed tunnel could provide the basis for predicting the distribution of water
inflow if the water pressure in the shallow rock is known. This works for tunnels
made by tunnel boring machines, which leave the apparent rock structure relatively
unchanged, but not for blasted tunnels with considerable excavation disturbance.
For them, it is instead the undisturbed rock structure at 1–3 m distant from the
tunnel walls that counts. In both cases, one can derive patterns of inflow spots,
as in Figure 4.12, in which the spots represent fracture crossings. The practically
most important conclusion concerning hydrology is that the inflow rate per spot
in a blasted tunnel is significantly lower than in a tunnel made by tunnel boring
machines.
Figure 4.13 illustrates the microstructural response to low inflow rates. The
water produces a small hydraulic wedge that stagnates by being blocked by the
expanding granules around it. Many more small wedges are then formed in dif-
ferent directions and they all soon stop growing because of the tightness of the
expanding granules surrounding them. The water pressure is thereby raised, which
Fourth-order Fifth-order
Secondary
discontinuities discontinuities Primary inflow
inflow spots
spots
FIGURE 4.12 Schematic picture of a tunnel wall with visible inflow (primary spots) in
fourth-order discontinuities. Fifth-order discontinuities representing secondary inflow spots.
54 Bentonite Clay
Capillary
meniscs Pellets
Discharge
spot in rock
FIGURE 4.13 First phase of wetting of a fill of smectite granules at low water inflow rate
and still low pressure.
Discharge Pellets
spot in rock
FIGURE 4.14 First stage of wetting of a fill of smectite granules at high water inflow caus-
ing high pressure and piping.
makes the wedge with least resistance grow and propagate as shown in Figure 4.14.
As said, one can therefore expect that the high-pressure case evolves independently
of the inflow rate.
Figure 4.14 indicates the impact on the microstructure by high inflow from a frac-
ture in contacting rock. The high water pressure generates a propagating hydraulic
wedge that displaces granules and wets them by which they expand. All this leads
to the formation of a tube-like channel that continues to propagate if the inflow is
high enough to maintain a sufficiently high pressure at its tip. Gravity will tend to
direct the channel downward directing penetrating water to the lowest part of the fill.
Wetting of the pellets that confine the channel takes place by diffusive migration of
water, which makes the channel walls successively tighter. This means that all water
entering from an inflow spot will be forced through one or a few main channels that
become eroded and widened. They will not self-heal until the water pressure at the
tips has dropped below the critical value for propagation.
While the major part of the inflow into a tunnel backfilled with smectite granules
takes place via primary and secondary inflow spots, that is, fracture crossings or discrete
fractures with varying aperture, water will successively enter from the numerous finer
fractures in the rock matrix when the larger part of the fill starts to become saturated.
4.3.4.5 Impact of Hydraulic Gradients on the Permeation

of Water-Saturated Clay Seals
Natural fine-grained soils with a smectite content of 10%–20% often have a grada-
tion that nearly fulfill ordinary filter criteria, implying that small particles occupy
voids between larger ones. For achieving this, the grain distribution curve shall be
of Fuller- or till-type with no discontinuities of the part representing D < 0.1 mm
(Pusch and Yong 2006).
For artificially prepared clay/aggregate mixtures, the situation can be more dif-
ficult because the clay component is not uniformly distributed and forms more or less
heterogeneous and locally soft coatings of aggregate grains. Large numbers of small
stacks can be washed out and transported in the direction of percolating water as in
the described case of tunnel fills of smectite granules. One must also be aware of
the risk of varying macroscopic density of compacted fills; it can be more than 10%
lower near the walls than in the center of backfilled blasted tunnels (Pusch 2008).
Experimental investigations using microscopy and special equipment for the
determination of the resistance to piping and erosion of water-saturated clay have
demonstrated that there is a critical gradient for hydraulic fracturing of soft clay
gels, such as those located in the voids between coarser grains in mixed soils (Pusch
1995). Typical data are given in Table 4.4. The tests demonstrate that the critical
gradient is in the range of 30–55 m/m for very soft Na bentonite. Other tests have
shown that the critical hydraulic gradient increases very significantly with increasing
density. For Ca clay of low density, the critical gradient is even lower and for reach-
ing values on the order of 35–40, it is required that the density is at least 1400 kg/m3.
The detailed microstructural constitution needs to be considered in the design of
clay-based seals that can be exposed to high hydraulic gradients. A classical example
in civil engineering is the clay core of earth dams, which can be exposed to gradi-
ents well over 3 m/m. Other examples are clayey top liners of landfills for which the
hydraulic gradient can be very high if their thickness is small and the overburden
thick, which makes the water pressure high (Figure 4.15). One understands from this
figure that geotextiles containing a few millimeter or centimeter of smectite clay
with low dry density are not suitable as sealing component of top liners of landfills.
For very effectively compacted mixtures of smectite clay and coarser soil one
can, for a high clay content, bring the density of the clay component up sufficiently
to reduce the risk of piping and erosion, but for low bulk densities and clay contents,
and low density of the clay component, the risk of piping and erosion can be critical.
The fact is that the clay component can hardly get a higher density at saturation than
56 Bentonite Clay
TABLE 4.4
Experimentally Determined Critical Hydraulic
Gradients for Piping of Water-Saturated
Montmorillonite-Rich Clay (MX-80)
ρs (kg/m3)
(saturated) Fluid Critical Gradient
Na-dominated
1080 Dist. 30–40
1080 FF 50–55
1080 Sea 50
1140 Dist. 40
1140 FF 35
1140 Sea 40
Ca-dominated
1380 Sea 35
1380 FF 40
Source: Pusch, R. et al., Handbook of Buffers and Backfills, Part I., SKB
Arbetsrapport AR 95-45, SKB, Stockholm, Sweden, 1995.
ρs, density at fluid saturation; Dist., distilled water; FF, strongly brackish
(11,000 ppm), Ca-dominated solution; Sea, seawater (36,000 ppm).
Top of landfill
Local groundwater level
Erosion-protecting Thickness of Hydraulic

overburden H clay (h), m gradient, m/m
0.01 100 for H = 1 m
0.1 100 for H = 10 m
1.0 10 for H = 10 m
Clay-based sealing 1.0 1 for H = 1 m
component of the top h
liner
FIGURE 4.15 Example of the hydraulic gradient for different thicknesses of the overburden
and of clay sealings. For geotextiles with one or a few centimeter thickness and low density
the hydraulic gradient can be critically high.
about 1500–1600 kg/m3 even for effective compaction. Using microstructural mod-

eling, which we will discuss here, it can be shown that part of the clay component can
have much lower densities than implied by this interval (Figure 4.16).
The homogenization process of clay/aggregate mixtures with 10% clay, sorb-
ing water, has been modeled by the finite element method technique based on the
Sand
particles
Bentonite
granules
Bentonite Ballast grain

grain
T = 11 s
T = 57 s
T=∞
FIGURE 4.16 Schematic microstructure of mixture of 10% MX-80 granules and 90% sand
(upper) and density distribution after complete saturation and expansion of the clay, while the
three lower graphs describe the evolution of density. The clay in the narrowest wrinkles is
soft, whereas the core of the original clay granules is not changed. (After Börgesson.)
assumption that the clay particles and coarser grains are rod-shaped and regularly
arranged in the fashion shown in Figure 4.17. It exhibits the initial grain arrangement
consisting of sand grains with 2.5 mm diameter and clay granules of MX-80 clay
with 1 mm diameter, and the distribution of the density of the fully saturated and
expanded soil as a function of time (Pusch and Yong 2006). The dry density of the
air-dry mixture was 1560 kg/m3.
The calculation, which gave the successive expansion of the clay grains and of
the evolution of voids based on a theoretical ABAQUS model gives the density and
volume of the clay elements, using a relationship between the hydraulic conductivity
and time. The assumed internal friction prevented complete ultimate homogenization.
The calculated density of the water saturated central part of the expanded and
matured clay grains was the original, that is, 1930 kg/m3, whereas the clay in the
58 Bentonite Clay
(a) (b)
FIGURE 4.17 Schematic picture of coagulation of clay particles: (a) small aggregates and
voids and (b) large aggregates separated by large voids for the same dry density.
most narrow portion of the voids between the big aggregate grains got a density of
1420 kg/m3. The study also showed that the narrowest portion of the voids between
the coarse grains will not be filled at all by clay. The resulting open space, which
has the form of isolated unfilled ring-shaped zones around the contact points of
the coarse grains, affects the overall hydraulic conductivity, primarily by providing
pathways for water flowing under high hydraulic gradients. The mixed soil is hence
sensitive to piping and erosion.
For accurate determination and selection of relevant values of the hydraulic
conductivity of the clay component of top liners of landfills, it is essential to con-
sider the impact of temperature, especially in hot climate. For thin overburden, the
temperature can be higher than 50°C of the clay layer and the viscosity lower and
the fluidity of the porewater hence higher than at testing in a geotechnical lab. For
landfills in cold climate, the clay may have a temperature of only a few degrees
on the Celsius scale and consequently be less fluid. Table 4.5 gives an example of
the change in evaluated conductivity by temperature cycling. The most important
impact of a rise or drop in temperature is evaporation and migration under tem-
perature gradients.
TABLE 4.5
Example of the Influence of Temperature Cycling of Clay Prepared
from Smectite-Rich Clay (MX-80) Powder Saturated and Percolated
with Distilled Water
Density at Saturation Dry Density Hydraulic
(kg/m3) (kg/m3) Temperature (°C) Conductivity (m/s)
1660 1050 22 3.4E−12
75 1.2E−11
22 2E−12
The tendency of reduced hydraulic conductivity by heating of confined sam-

ples by a few tens of degrees on the Celsius scale is believed to be a successively
improved microstructural homogeneity in contrast to the coagulation effect of
high temperatures, which we will consider in Chapter 5 with respect to mineral
conversion.
4.3.5 Impact of Saltwater on the Hydraulic Conductivity

We know from Chapter 3 and various investigations (Pusch 2001b) that the micro-
structural constitution is strongly affected by the porewater chemistry. The coagula-
tion of smectite clay formed in freshwater or brackish water caused by permeation
with saltwater is well known and is illustrated in Figure 4.17. The indicated increase
in void size and interconnectivity is the reason for the higher hydraulic conductivity
of saltwater clay.
The difference in conductivity of smectite-rich clays of different origin and
composition is very obvious for low densities and significant even when the den-
sity at water saturation is as high as 1800 kg/m3 as exemplified by Table 4.6. For
densities lower than about 1600 kg/m3 Ca smectite does not form a stable micro-
structural network when percolated by solutions with high electrolyte content.
For comparison with less montmorillonitic clay, values for the illite/smectite (I/S)
mixed-layer Friedland clay with only about 20% montmorillonite are included in
the table.
TABLE 4.6
Hydraulic Conductivity in m/s of MX-80 and Friedland
Clay for Salt Percolate Percolate, (Hydraulic Gradient
100–300 m/m)
Density at Saturation (kg/m3)
Percolate 1600 1800 2000 2050
MX-80
Distilled water 5E−11 7E−12 5E−13 3E−13
0.5 NaCl solution 7E−11 2E−11 2E−12 6E−13
3.5 NaCl solution – E−10 E−11 E−12
0.5 CaCl2 solution E−9 2E−11 2E−12 9E−13
3.5 CaCl2 solution E−8 3E−11 3E−12 E−12
Friedland clay
Distilled water E−8 E−10 5E−12 3E−12
2 NaCl solution 2E−8 – – –
10 NaCl solution 5E−8 E−9 5E−11 2E−11
3.5 CaCl2 solution E−9 6E−10 2E−11 E−11
10 CaCl2 solution – 5E−9 3E−11 2E−11
60 Bentonite Clay
90
20% CaCl2
80
70
K × E10 (m/s)
60
50
40
30
20
10
0
0 2 4 6 8
Time (days)
FIGURE 4.18 Evaluated hydraulic conductivity of dense compacted Friedland clay.

Oedometer test using a hydraulic gradient of 30 m/m. The percolate was a solution with 20%
CaCl2 and the clay density at fluid saturation 2000 kg/m3 (dry density 1590 kg/m3).
One would expect that the mixed-layer Friedland clay would be less affected by
high porewater salinity than the more smectite-rich clay MX-80 since microstruc-
tural collapse by coagulation of elements with lower density than of the average clay
matrix, like the softer gels in the microstructural channels, is limited in comparison
with montmorillonite-rich clays. This is true for low densities, whereas the impact of
the porewater salinity for high densities is small. It is obvious that the conductivity of
the mixed-layer clay is remarkably low even for extreme concentrations of sodium and
calcium in the porewater when the density at saturation is about 2000 kg/m3 or higher.
Figure 4.18 illustrates that the conductivity of smectite clay percolated by 20%
CaCl2 solution dropped successively, implying a reduction of the initially high flow
rate. The experiment is of particular interest because it demonstrates that reaching
steady flow took more than one week and that the evaluated conductivity tended
to approach that for 10% CaCl2 solution (Table 4.6). This suggests that the rather
extreme salinity of 20% did not cause further microstructural changes than those
caused by percolation with the less saline Ca solutions with 3.5% and 10% salinity.
Percolation with any of the salt solutions is believed to have caused clogging of the
microchannels by flow-transported large particle aggregates that were eroded from
the clay matrix (Pusch and Yong 2006).
4.3.6 Criteria
In practice, it is required to design and construct clay-based seals, such as embedment
of containers with hazardous waste, backfills in tunnels, or top liners of landfills, so
that the maximum hydraulic conductivity specified in national or local environmental
regulations is not exceeded. They vary internationally and depend, as we have seen
here, on how the tests are performed and on what statistical variations that are allowed.
In particular, the impact of scale is important: are the properties of small lab samples
representative of a volume that is ten million times larger? No general rules are valid
but national regulations specify allowable bulk conductivities for bottom liners of
landfills of municipal solid waste, such as incinerated organic material (Table 4.7).
TABLE 4.7
Examples of National Specifications for Bottom Liners to
Contain Municipal Solid Waste
Minimum Clay Maximum Allowed
Layer Thickness Hydraulic
Country (m) Conductivity (m/s)
Austria 0.6 E−9
Germany 0.75 5E−10
United Kingdom 1.0 E−9
US Environmental Protection 0.6 E−9
Agency (EPA)
Sweden – E−10
Sources: Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering
Performance, 1st edn., Taylor & Francis, London, 2006; ASTM Standard
D 5084, Standard test method for measurement of hydraulic conductivity
of saturated porous materials using a flexible wall permeameter, American
Society for Testing and Material, West Conshohocken, PA, 1994.
For disposed radioactive and chemical wastes, the demand is stronger and in Sweden
and Finland, the criterion is that the average hydraulic conductivity of clay buffer, that
is, the embedment of canisters with high-level radioactive waste (HLW), must not be
higher than E−12 m/s. For backfills in tunnels with canisters in deposition holes in
underground repositories it is stated by Swedish Nuclear Fuel and Waste Management
Company and others that the conductivity must not exceed that of the surrounding
rock,* which is commonly found to be in the interval E−11 to E−9 m/s (Pusch 2008).
For recently constructed bottom liners of landfills of chloride-rich ashes from incin-
eration of organic waste, the stipulated maximum allowed hydraulic conductivity was
2E−11 m/s for a typical case. One realizes that such demands require measurements
of very high precision and specification of how they shall be conducted with respect
to the type of percolate, hydraulic gradients, scale effects, and temperature conditions.
The frequent question of recommended clay material for different purposes is
easily answered by consulting test protocols from geotechnical labs but the real
problem is what the long-term function is under the conditions that prevail at the
waste disposal site, concerning the groundwater chemistry, hydraulic gradients, tem-
perature, and so on both today and in future. We will look closer at these things in
Chapter 5 and the rest of the book, and confine ourselves here to say that smectite
clays with appropriate density and a minimum of corrosive mineral and organic
components fulfill ordinary criteria. For certain applications, such as disposal of
highly radioactive waste, the rather small impact of even high porewater salinity
on the physical properties of I/S mixed-layer clays makes them attractive as buffer
* This should, in fact, refer to the excavation-disturbed rock with the commonly found bulk conductivity
E−10 to E−7 m/s (Pusch 2008).
62 Bentonite Clay
Quartz
Holmehus clay (bulk)
powder
(randomly oriented)
Intensity (arbitrary units)
Gypsum
Chlorite
Chlorite
Anatase
Quartz
Quartz
Quartz
Quartz
Quartz
Quartz
Quartz
Quartz
Albite
K-Fsp
Pyrite
IS-ml
IS-ml
IS-ml
IS-ml
10 20 30 40 50 60
Position (°2θ CuKα)
FIGURE 4.19 X-ray diffraction spectrum of Holmehus clay (bulk sample).
material, whereas for sealing very deep boreholes with HLW—the deep hole concept
(VDH)—even more stable clay minerals such as saponite can be a candidate, we
will discuss in Chapters 5, 6 and 8. As concerns the behavior of mixed-layer clays,
the Danish Holmehus clay (Dantonite A/S) of Tertiary age, belonging to the series
of I/S mixed-layer clays in northern Europe, is a promising candidate clay. Its major
components are I/S mixed-layer and dioctahedral vermiculite/smectite mixed-layer
minerals (Figure 4.19). The minus 2 μm fraction is about 90% of the total mineral
mass, with the smectite fraction of IS-ml making up 60% and the dioctahedral ver-
miculite/smectite mixed-layer fraction ranging between 20% and 25%. Figure 4.20
shows the appearance of clay material dispersed in distilled water.
For comparison of the two described montmorillonite-rich clays, MX-80 and
Holmehus clay, Tables 4.8 and 4.9 give the respective mineral and chemical compo-
sitions. The cation exchange capacity of Holmehus clay is 29.7–31.6 meq/100 g using
the Cu-triethylenetetramine-method. MX-80 typically has 90 meq/100 g. Table 4.10
shows the hydraulic conductivity of the two clay types as function of density.
The lower content of expandables of the Holmehus clay than of the more mont-
morillonite-rich MX-80 clay makes the I/S clay somewhat more permeable and
somewhat less expandable but both are deemed suitable as buffer clay for isolating
canisters with HLW in deep repositories for densities at water saturation of at least
1900 kg/m3. The buffering capacity of the Holmehus clay by providing dissolved Si
makes Holmehus clay a qualified candidate for use in both deep and near-surface
repositories.
It is worth here to underline the distinction between the microstructural constitution
of undisturbed natural smectite clays, and of those prepared by mixing clay granules
FIGURE 4.20 Transmission electron micrograph of dispersed Holmehus clay showing lath-
like illite–smectite mixed-layer phases. (After Kasbohm.)
TABLE 4.8
Minerals in Two Smectite-Rich Clays Identified by X-Ray Diffractometry
(Weight%; Montmorillonite in MX-80: 77%–85%)
Clay M, 2M1 Chl Q/C P K G J P
Holmehus 10–20 <5 10–15 <5 5 2 2 0.3
MX-80 1–2 – 6–15 5–8 – – – 0.3
Chl, chlorite; G, gypsum; J, jarosite; K, K-feldspar; M, muscovite; P, plagioclase; P, pyrite; Q/C, quartz/
cristobalite.
TABLE 4.9
Element Data of Bulk Samples of Holmehus and MX-80 Clays
Element SiO2 Al2O3 Fe2O3 MgO CaO Na2O K2O
Holmehus 58.6 15.3 6.5 2.2 0.7 1.4 2.8
MX-80 63.6 19.8 5.0 3.2 3.1 2.8 1.0
Note: The SiO2/Al2O3 ratio is 3.83 for Holmehus sample and 3.21 for the MX-80 material.
and aggregate grains. As we have seen in this chapter, the latter have a lower degree of
microstructural homogeneity since the clay grains do not expand enough to fill the voids
between coarser grains with dense clay. The softer parts of the clay matrix, that is, the
gels present in the voids and microchannels, are sensitive to the porewater chemistry.
They make up a very small fraction of dense clays but a considerable part of soft ones.
64 Bentonite Clay
TABLE 4.10
Hydraulic Conductivity K of Holmehus and MX-80 Clays (Hydraulic
Gradient 100–300 m/m)
Density K, Holmehus (m/s) K, MX-80 (m/s)
Density at
Saturation Dry Density 3.5% CaCl2 3.5% CaCl2
(kg/m3) (kg/m3) Distilled Water Solution Distilled Water Solution
1900 1430 2.5E−12 7E−12 2E−12 1E−12
1800 1270 1.5E−11 3E−11 7E−12 3E−11
1675 1070 1.0E−10 1E−09 2E−11 1E−10
4.3.7 Impact of Smectite Content on the Hydraulic

Conductivity of Mixed Soils
For permanent closure of drifts, adits, and ramps of underground disposal of nonhaz-
ardous waste, backfilling of muck from blasting or from tunnel boring machines may
be sufficient. However, where tightness is needed, as for backfills in deposition tun-
nels in repositories for high-level radioactive waste, soil with some smectite content is
required, a common criterion being that backfills must be less conductive than the sur-
rounding rock. They also have to exert a certain minimum effective pressure (usually
set at 100 kPa) on the rock for providing support to the rock and for hindering water
leakage along the contact of rock and backfill. Summarizing available data from test-
ing of mixtures of clay material and coarser soil material, and of natural smectite clays,
gives a basis for selecting either of them (cf. Pusch 1994; Al-Thaie 2013).
4.3.7.1 Mixed Clay and Coarser Material

The composition and density should be such that the bulk hydraulic conductivity is
low enough to fulfill the criteria set, and that the placed material is not damaged by
erosion caused by in- or ad-flowing water. If the tightness criterion for underground
waste disposal implies that the hydraulic conductivity must not exceed a certain value,
for example, the conductivity of the surrounding rock, the clay content and density
must be accordingly adapted, for which one can use available test data such as the ones
summarized in Table 4.11. They stem from systematic laboratory tests of mixtures of
montmorillonite-rich clay (MX-80) and two types of silt/sand materials with the grain
size compositions shown in Table 4.12 and with water added to reach the optimum water
content of 13%. Aggregate 1 had well-rounded grains with a high degree of sphericity
(0.3–0.9), whereas the grains in Aggregate 2 were oblong (sphericity of 0.3–0.7) and had
the surfaces coated with fine debris from the crushed rock. The surface smoothness, of
which the roundness is a measure, is of importance for the possibility of clay to move
into the most narrow space between coarse grains. The grains of Aggregate 1 had a very
profound roundness and can be termed rounded or well-rounded without sharp edges
and corners, using common terminology. Microscopy analyses showed that quartz
formed at least 95% of all the glacial silt/sand material (Aggregate 1).
TABLE 4.11
Typical Hydraulic Conductivity Data (K) for Mixtures of Smectite-Rich Clay
with Na as Dominant Sorbed Cation (MX-80) and Silty/Sandy Soil
(Permeation with Distilled Water)
Smectite Density at Estimated Degree
Aggregate (Montmorillonite) Dry Density Saturation of Proctor
Soil Type Content (%) (kg/m3) (kg/m3) K, (m/s) Compaction (%)
A 1 10 1750 2100 E−9 90
B 1 10 1900 2200 4E−11 100
C 1 20 1770 2100 E−10 90
D 1 20 2010 2270 E−11 100
E 2 10 1970 2240 9E−10 90
F 2 10 2180 2370 2E−11 100
G 2 20 1760 2110 E−11 95
H 2 20 1880 2180 E−11 100
TABLE 4.12
Examples of Grain Size Composition of Sand/Silt Fraction
and Clay Granules for Preparation of Mixed Soils
Content of Material Finer than d mm, %
Grain Size, d mm Aggregate 1 Aggregate 2 MX-80 clay
<10.00 80–90 80 100
4.00 53–63 55 100
2.00 43–55 43 100
1.00 32–46 33 98
0.50 27–36 28 95
0.20 25–27 20 50
0.10 17–21 33 18
0.05 15–19 10 10
Table 4.11 shows that all materials compacted to 100% Proctor density had
approximately the same conductivity. The densities were highest for most of the
mixtures with crushed rock (Aggregate 2), which were, therefore, supposed to have
lower conductivities, but the fine debris enriched along the surfaces of the grains
raised the evaluated conductivity to more than that of mixtures with glacial sand/silt.
The common opinion is that soils with very carefully selected size distribution of
glacial, silty sand can be compacted to higher densities than poorly rounded soil, but
experience shows that finely crushed rock mixed with smectite granules (Table 4.12)
can be even denser (Soil F in Table 4.11). The reason is that the edged grains get
crushed in the compaction phase and automatically generate fragments that fit in
66 Bentonite Clay
voids between larger grains. Such soils should be moderately agitated in the mixing
process in order to avoid coating of larger grains by fine debris. See Table 4.12.
4.3.7.2 Natural Soils

The option of using natural smectite clays with suitable composition is getting popu-
lar both because of the simpler preparation, just drying and milling, and because of
lower cost. The types of soils of potential interest are those with high smectite content
for use as buffers and the ones with intermediate smectite content for top and bot-
tom liners of landfills and for backfilling shafts, tunnels, and drifts in underground
repositories for hazardous waste. The material cost is high for any of the clayey
soils that are of potential interest for liner construction and backfilling purposes and
even higher for smectite-rich clay of the uniform quality that is required for buf-
fers. Considering the two natural montmorillonite-rich clays, MX-80 and Holmehus,
none of them would be useful for the construction of landfill liners since the density
to give the required low conductivity would give a swelling pressure high enough
to lift the overburden. Depending on the criteria set, less expandable clays, such as
the mixed-layer Friedland clay, would do as well as the aforementioned Lithuanian
clay exploited for cement production and being a candidate for isolating low- and
medium-level radioactive waste in an on-ground repository. It will be taken here as
an example of clay useful for liner construction with a minimum of processing.
The clay is of Triassic age and contains Fe-rich chlorite, I/S mixed-layer minerals
(illite/montmorillonite), well- to poorly crystalized illite, kaolinite, feldspars, quartz,
and traces of calcite. The percentage of illitic layers in the I/S varies between 40%
and 85%, indicating that the content of expanding components is relatively small,
that is, 15%–25%. The clay fraction makes up about 30% of the total mass. The
Atterberg plastic limit is 22% and the liquid limit is 55%. For assisting the readers in
characterizing the clay, a typical X-ray diffraction diagram of air-dry and ethylene-
glycol-treated material is shown in Figure 4.21.
Minimum processing of the clay material implies air-drying of stockpiled material
consisting of unsorted chunks and distributing it on the construction site by tractors by
which it disintegrates and provides a size distribution curve of approximately parabolic
Fuller type (Pusch and Yong 2006). Corresponding material for oedometer testing would
imply crushing to crumb size of 0.01–2 mm. In the case described, compaction in the
oedometers was made to the dry densities of 1190, 1660, and 1760 kg/m3. Percolation
of the samples was made with distilled water under hydraulic gradients ranging between
20 and 60 m/m, giving the hydraulic conductivities in Table 4.13.
One finds that the Lithuanian Triassic clay with a dry density of 1750 kg/m3, rep-
resenting heavy compaction, is as tight as mixtures of 20% clay and suitably graded
glacial sand and silt, or crushed rock (Table 4.11, Soils C,D,G, and H).
4.3.7.3 Hydraulic Conductivity of Unsaturated Clay

Flow through unsaturated soil is very complex and the hydraulic conductivity
cannot even be defined if percolation is associated with a successive change in the
degree of water saturation. A general observation is that water migration under a
hydraulic gradient in smectite clays with a lower degree of saturation than about
70%–80% takes places by diffusion (Knudsen flow), leading to a considerably
AD
Sample C
9.97 Å WOI
Quartz
14.3 Å
Chlorite
EG
IS-ml
kaolinite and chlorite
EG
14.2 Å chlorite
7.15 Å
Feldspar
4.41 5.41 6.41 7.41 8.41
IS-ml (∼85%)
IS-ml (∼40%)
IS-ml (∼40%)
IS-ml (∼85%)
16.7 Å IS-ml (∼40%)
4 8 12 16 20 24 28 32
°2θ CoKα
FIGURE 4.21 (See color insert.) Lithuanian smectite clay intended for lining c onstruction.
Legend: AD, air dried; EG, ethylene-glycol saturated; and IS-ml (~85%), illite-smectite
mixed-layer with 85% illitic layers. (After Kasbohm.)
TABLE 4.13
Evaluated Hydraulic Conductivity after Reaching Steady-State Flow
Dry Density Density at Water Water Content Hydraulic
(kg/m3) Saturation (kg/m3) (weight %) Conductivity (m/s)
1190 1750 39.2 1.2E−10
1660 2045 29.6 6.0E−11
1760 2110 21.4 1.4E−11
lower recorded hydraulic conductivity than for completely saturated mixtures. At

least for soils with relatively high clay contents, this can be explained by the fact
that wide microchannels are partly or largely filled with air, causing capillary
retention and lower conductivity at laboratory testing than of completely water-
saturated material.
4.4 GAS CONDUCTIVITY

4.4.1 Basics
Gas transport through clay is an issue of fundamental importance for oil and gas
geologists since it determines the potential of clayey caprock to retain pressur-
ized petroleum in liquid or gaseous form for exploitation. The self-sealing ability
68 Bentonite Clay
of ductile clay explains why they are tight despite seismic events and tectonically
generated displacements since Cretaceous and Tertiary epochs. Since the conductiv-
ity of gas is about a thousand times higher than that of water, its flow is controlled
only by pressure, in contrast with water that moves as a consequence of pressure
and conductivity. The most important issue is the gas pressure that succeeds in gas
penetration through the buffer clay, that is, the critical gas pressure. For temporarily
stagnant gas in clay microchannels, the solubility of the gas is an important matter
since, by dissolution, the bubbles will shrink causing gas to dissolve and diffuse out
of the system. If the gas production rate ceases, microstructural self-healing can take
place by convergence of widened channels and subsequent growth in them of clay
gels released from the surrounding, denser matrix. This is the case for the ductile
North Sea clay beds down to a couple of kilometers below sea bottom but is not valid
for clay buffers surrounding hot canisters with HLW since they can lose their ductil-
ity and self-sealing potential by cementation and stiffening.
The processes and mechanics of gas permeation in and through clay is of greatest
importance in practice as demonstrated by the experience from oil and gas prospec-
tion and production and by the possibility of storing pressurized carbon dioxide in
sedimentary rock that has previously contained hydrocarbons. Data from Russian
gas and oil fields show that natural beds that are rich in montmorillonite clays with
lower porosity than about 30%, corresponding to dry densities of more than 1950 kg/m3,
and with a Na/Ca ratio of 4–12, are to all intents and purposes practically tight
(Osipov et al. 2004).
4.4.2 Determination of Gas Transport Capacity

4.4.2.1 Background
Microstructural heterogeneity has a decisive influence on the critical gas pressure
and transport and for artificially prepared clay seals one realizes the role of micro-
channels: they provide least resistance for gas migration. For microstructurally
homogeneous clay beds in nature, the bond strength of adjacent particles is the key
factor for overcoming gas flow (Horseman and Harrington 1997). For such beds, the
resistance to gas-induced displacement of the clay matrix is related to its swelling
pressure: gas penetration requires separation of dense matrix components causing
fracking. For very dense, artificially prepared smectite-rich clays, the critical gas
pressure is also a function of the bulk swelling pressure but merely for high densities.
This is affirmed by the results shown in Table 4.14 that refers to tests using nitrogen
gas. For lower bulk densities, the critical gas pressure is substantially lower than the
swelling pressure, which is explained by the low capillary retention of the micro-
channels that are easily rinsed from soft clay gels and become locally widened by
penetrating gas in dense clay. Here, the capillary resistance is high.
The impact of scale has been investigated using large permeameters (780 mm
inner diameter and 300 mm height) filled with layer-wise placed montmorillonite-
rich clay* in Na form and wetted to become water saturated and compacted to a dry
* GEKO/QI (Sued-Chemie), Germany.

TABLE 4.14
Experimentally Determined Critical Gas Pressure in MPa for
Smectite-Rich Clay (MX-80) in Na Form
Swelling Pressure of
Smectite-Rich Clay (MX-80)
Density at Saturation Experimentally Determined Saturated with Distilled
(kg/m3) Critical Gas Pressure (MPa) Water
2130 20 20
1850 2 3
1570 0.1 0.4
density of 1080 kg/m3 (1680 kg/m3 at saturation). The hydraulic conductivity and
swelling pressure were 4E−11 m/s and 200 kPa, respectively. In a second series, a
mixture of 10% clay of the same type and 90% ground granitic rock aggregate was
placed and compacted in thin layers to a dry density of 1920 kg/m3 (2200 kg/m3 at
water saturation). The hydraulic conductivity was 2E−9 m/s and the swelling pres-
sure about 50 kPa. For the unmixed clay, gas breakthrough occurred at a gas pres-
sure of 190 kPa about five days after raising the gas pressure to the critical level. For
the 10/90 mixture, the corresponding pressure was only 15 kPa, and breakthrough
occurred after about 10 minutes. This low pressure is consistent with the fact that the
clay component had a low density, that is, about 1600 kg/m3 by saturation with arti-
ficial seawater with a salt content corresponding to 50% of ocean water. Figures 4.22
and 4.23 show the gas flow rate recorded in the tests.
In the experiment with unmixed clay (Figure 4.22), the increased flow rate after
about 80 hours was due to the successively increased pressure gradient when the
gas front approached the discharge points. Steady flow was reached after 160 hours,
4
Gas flow in ml/min
0
0 100 200
Time in hours after application of 190 kPa
gas pressure
FIGURE 4.22 Gas penetration and flow at the critical pressure of Na montmorillonite clay
with 1680 kg/m3 density at saturation (dry density 1080 kg/m3) using saltwater with 50%
concentration of ocean water. Megapermeameter test. (Data from Pusch, R., Yong, R.N.,
Microstructure of Smectite Clays and Engineering Performance, 1st edn., Taylor & Francis,
London, 2006.)
70 Bentonite Clay
12
10
Gas flow in ml/min

8
0
0 1 2 3
Time in hours after application of 15
kPa gas pressure
FIGURE 4.23 Gas penetration process at the critical pressure of a mixture of 10% Na mont-
morillonite and well-graded ballast. Density at saturation with saltwater with 50% concentra-
tion of ocean water was 2200 kg/m3 (dry density 1900 kg/m3). Megapermeameter test. (Data
from Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering Performance,
1st edn., Taylor & Francis, London, 2006.)
1 mm
FIGURE 4.24 Optical micrograph of a thin section (30 μm) of large gas path through water-
saturated montmorillonite-rich clay with 1630 kg/m3 density (dry density 1000 kg/m3). G is the
gas channel. The arrows point at aligned clay coatings consolidated by the gas pressure.
representing a critical pressure of about 150 kPa. In the test with mixed material, gas
quickly made its way by forming some major flow paths already when the pressure
had been raised to 15 kPa.
A conclusion from the tests is that slow buildup of the gas pressure causes
consolidation of the clay in the microchannels in the unmixed clay and of the
clay components in mixed clay (Figure 4.24). When the critical pressure was
reached gas moved faster and caused piping. A general observation was that about
0.02%–0.2% of the water content was discharged before there was any evidence of
gas penetration.
4.4.2.2 Laboratory Technique

Measurement of the gas conductivity can be made in the same fashion as determina-
tion of the hydraulic conductivity. The much more complex transport mechanisms,
such as dissolution and diffusion and transient return to gaseous form at the low
pressure end, make the evaluation uncertain, however. Most gas permeation tests
are focused on determining the critical gas pressure, which is sometimes called the
capillary pressure. It can be determined by use of the equipment employed for deter-
mining the hydraulic conductivity. A major problem is to avoid leakage along the
sample periphery.
The type of gas used for the tests must not react with the soil component, which
means that air containing CO2 and O2 is not suitable. Nitrogen and helium are suit-
able gases whereas hydrogen gives difficulties by leakage and explosion risks.
4.4.3 Modeling of Gas Penetration Using Capillary Analogs

Gas under pressure makes its way through clays in two ways: by penetration and flow
in gaseous form as described, and by molecular diffusion in porewater of dissolved
gas. Pressurized gas diffuses in dissolved form through buffer clay if the gas pres-
sure is low enough compared to the fluid pressure. In gas percolation experiments,
the fluid pressure must therefore be sufficiently high to prevent gas from forming
bubbles, which may block pores. Bubbles move in a different way than gas mol-
ecules, primarily because of capillary effects. The diffusion coefficient of nitrogen
(N2) gas in water is about 1.9E−9 m2/s at room temperature, whereas that of hydrogen
(H2) is about 4.5E−9 m2/s.
A simple model of finger-like penetration of a system of quartz capillaries with
diameters corresponding to those evaluated from microstructural analyses appears to
give values that are in fair agreement with actually recorded pc-values for montmo-
rillonite-rich clay, hence indicating the relevance of the concept of microchannels
(Table 4.15).
One finds that the predicted critical pressure overrates the measured pressures
for lower densities, which is explained by the different dominant mechanisms in gas
transport: the channels act as capillaries for high densities, whereas gas movement in
less dense smectite clay takes place by displacement of soft gels in them.
4.5 ION DIFFUSION

4.5.1 Definitions
The ion diffusivity in smectite clay has different forms and definitions: pore diffu-
sion, surface diffusion, and effective diffusion. Cations take part in ion exchange
processes that alter the microstructure and cause transient changes in diffusivity.
Anions are turned off from the interlamellar space for charge reasons.
72 Bentonite Clay
TABLE 4.15
Theoretical and Actually Measured Critical Gas Pressures (pc) for Very
Dense, Medium Dense, and Soft Clays with High Smectite Content
Critical Pressure (MPa)
Bulk Density Bulk Dry Density
(kg/m3) (kg/m3) Predicted Measured
2113 1800 14–30 20
1850 1350 4–6 2
1567 900 3–5 0.1
Source: Pusch 1995. Handbook of Buffers and Backfills, Part I. SKB Arbetsrapport AR 95-45.
SKB, Stockholm, Sweden.
4.5.2 Mechanisms and Basic Relationships

Diffusion of ions through clayey soils is affected by several physical and chemical
processes, such as ion exchange, physical and chemical sorption, filtration, and precipi-
tation. A common way to mathematically describe the diffusion process for ions migrat-
ing through compacted clay (Brandberg and Skagius 1991; Kato et al. 1995; Muurinen
and Lehikoinen 1995; Wold and Eriksen 2000), is to use the a pparent diffusivity Da
expressed as a function of the distribution coefficient (Kd) as in Equation 4.7:
Dp ⋅ n De
Da = = (4.7)
n + Kd ρd n + Kd ρd

where:
Da is the apparent diffusivity (m2/s)
De is the effective diffusivity (m2/s)
Dp is the diffusivity in pores (pore diffusivity) (m2/s)
n is the porosity
ρd is the dry density (kg/m3)
Kd is the distribution coefficient sorption parameter (m3/kg)
The fact that the apparent diffusivity can be higher for certain cations is explained
by the different sorption ability of different species. It is assumed that diffusion takes
place not only in the void system of the soil but also in the form of surface diffusion
(Figure 4.25). Fick’s second law, Equation 4.8, gives the general expression for the
concentration profile at position x and time t:
∂C ∂ 2C
= D 2 (4.8)
∂t ∂x
where:
D is the diffusion coefficient
C is the concentration of dissolved element
A—Comprehensive
B—Large
C—Minor
FIGURE 4.25 Smectite clay microstructure with dense particle aggregates shown as
hatched areas. Surface diffusion takes place along the exposed free surface of aggregates of
stacks of smectite lamellae. Matrix diffusion occurs within the aggregates, primarily by cat-
ion diffusion in the interlamellar space, whereas pore diffusion takes place in voids of various
types. Legend: A—surface diffusion: quick migration of cations; B—matrix diffusion: slow
migration of cations, quick migration of Li and Na; and C—pore diffusion (free water in
voids): migration of cations and anions. (Pusch et al. 2011.)
The adsorption characteristics of smectite clay control the diffusive transport of ions;
cations can migrate through both large pores and very narrow voids, including the
interlamellar space, whereas anions are assumed to be excluded from the micropores
and interlamellar space by the Donnan potential, and migrate only by pore diffusion.
Diffusion of cations in water is more or less hindered in very small voids because of
the strong adsorption capacity of the exposed mineral surfaces. Thus, surface diffu-
sion is assumed to dominate in these voids (Wold and Eriksen 2000).
Focusing in this chapter on the role of the microstructure for the performance
of clays, we will confine ourselves to consider, in a general manner, how diffusive
migration depends on the microstructure of clay seals. A first attempt to consider
this is to interpret the meaning of the terms pore diffusion and surface diffusion by
considering typical microstructural constitutions. Since many cations move by both
transport mechanisms the retarding effect of increased clay density on the diffu-
sion capacity of such ions is limited, especially for monovalent ions. This is shown
schematically by Figure 4.25, which illustrates three major transport mechanisms
disregarding the impact of possible chemical reactions:
1. Surface diffusion—migration of cations in the electrical double layer

2. Matrix diffusion—diffusive migration of cations in the interlamellar space
3. Pore diffusion—diffusive migration of cations and anions in free porewater
74 Bentonite Clay
In practice, measures of diffusive transport of water and solutes represent averaged

parameter values but the true transport is very complex because of the density varia-
tions on the microstructural scale. This is particularly obvious when considering the
microchannels that can contain just water, or more or less soft clay gels in which pore
diffusion dominates. The small fraction of a cross-sectional area of the clay matrix
that is represented by the channels—a few percent—would mean that they do not
play a significant role but the rate of diffusion in them may be much higher than in
the dense parts of the matrix and still makes them important.
Assuming that the concentration of ions in solution is so low that the linear
adsorption isotherm can be applied (Cs = KdρdC), one can write n = porosity and
De = effective diffusion etc.
∂C Den + Kdρd Ds ∂ 2C
= (4.9)
∂t n + Kdρd ∂x 2
Den + Kdρd Ds
Da = (4.10)
n + Kd ρd
where:
Cs is the concentration of species in the film adsorbed on the mineral surfaces
(kg/m3)
Ds is the surface diffusivity (m2/s)
De is the effective diffusivity (m2/s)
Da is the apparent diffusivity (m2/s)
The true fashion in which cations migrate through smectite clay is not fully under-
stood, especially concerning the role of microchannels, which is probably significant.
For anion diffusion, the distribution coefficient is negligible (Kdρd ≈ 0) because of
the cation adsorption characteristics of the clay:
Da = Dρ (4.11)

The pore diffusivity (Dp) is affected by various soil properties and microstructural
features such as the tortuosity τ (Pusch 2008; Wold and Eriksen 2000), so one has
to be content with an approximate relationship between the pore diffusivity and the
diffusivity in water:
Dp = (τ)2 αDw = fDw (4.12)

f = (τ)2 α (4.13)

where:
τ is the tortuosity factor
α is the viscosity factor
Dw is the intrinsic diffusivity (m2/s)
f is the impedence factor
Since, for cations, the term Kdρd in Equation 4.8 is commonly much bigger than the
porosity, one can, for the sake of simplicity, put the diffusion equation for cations in
the following form:
Den fD n
Da = + Ds = w + Ds (4.14)
Kd ρd Kd ρd
4.5.3 Role of the Microstructure

The diffusive transport capacity expressed as the effective diffusion coefficient refers
to the actual effective porosity and describes ion transport on the microstructural
level. This is in contrast to the apparent diffusion coefficient, which is a general
measure of diffusion derived directly from recorded concentration profiles. Cation
diffusion takes place in several ways, that is, in continuous water-filled voids, along
particle surfaces with electrical double layers, and through the interlamellar space in
smectites. The latter two mechanisms involve ion-exchange mechanisms for which
the sorption parameter, Kd, is relevant. In practice, the ion transport capacity can be
predicted by applying Fick’s law giving reasonably relevant values of the coefficient
of the density-related effective diffusion, De. The density of the clay plays an impor-
tant role as illustrated by Figure 4.26.
Since anions migrate in the pore space, the diffusive anion transport capacity is,
in principle, proportional to the ratio of the space of the voids between the stacks of
smectite lamellae and the total water-filled part of the clay. With increasing density,
there is a strong reduction of the available space for migration, and the diffusion
coefficient of anions, therefore, drops significantly. Figure 4.26 shows that anion
migration would be practically none if the density of the buffer could be raised to
1200
Effective diffusion coefficient
1000
multiplied by E12 (m2/s)
800
600
400
200
0
0 1000 2000 3000
Density at saturation (kg/m3)
FIGURE 4.26 Measured effective diffusivities for smectite clay Upper curve: monovalent
cations. Lower curve: monovalent anions such as chlorine. (Data from Kato, H. et al., Estimation
of effective diffusivity in compacted bentonite. In Scientific Basis for Nuclear Waste Management
XVIII: Proceedings of the Materials Research Society Symposium, vol. 353, Murakami, T.,
Ewing, R.C., Eds., Materials Research Society, Pittsburgh, PA, pp. 277–284, 1995.)
76 Bentonite Clay
TABLE 4.16
Averaged Literature Data of Effective Diffusion
Coefficients for Elements Migrating in Dense
Smectite-Rich Clay with 2000 kg/m3 at Fluid
Saturation
Species De (m2/s)
I-129 E−13 to E−12
Sr-90 E−9 to E−8
Na-22 E−10 to E−9
Cs-137 E−13 to E−11
Pu-238 E−11 to E−10
Am-243 E−11 to E−10
Sources: Brandberg, F., Skagius, K., Porosity, sorption and diffusivity

data compiled for the SKB 91 study, SKB Technical Report
TR 91-16, SKB, Stockholm, Sweden,1991; Kato, H. et al.,
Estimation of effective diffusivity in compacted bentonite.
In Scientific Basis for Nuclear Waste Management XVIII:
Proceedings of the Materials Research Society Symposium,
vol. 353, Murakami, T., Ewing, R.C. (Eds.), Materials
Research Society, Pittsburgh, PA, pp. 277–284, 1995;
Muurinen, A., Lehikoinen, U., Evaluation of phenomena
affecting diffusion of cations in compacted bentonite,
Report YJT-95-05, Voimayhtiöiden Ydinjätetoimikunta,
Helsinki, Finland, 1995; Wold, S., Eriksen, T., Diffusion
of organic colloids in compacted bentonite, SKB Techni
cal Report TR-00-19, SKB, Stockholm, Sweden, 2000;
Kim, H.-T. et al., Waste Management, 13, 303–308, 1991.
more than 2100 kg/m3 at water saturation (dry density 1750 kg/m3). They primarily
move in the microchannels.
The diffusion coefficient of ions is affected by the density, void size distribution,
cation exchange capacity, specific surface area of the soil solids, the chemistry of the
porewater, the ion charge and size, the concentrations of ions, the redox potential,
and pH (Pusch and Yong 2006). Table 4.16 indicates that there is a large span for the
ion species according to published data.
4.5.4 Impact of the Microstructural Constitution on Ion Diffusion

The apparent diffusivity of cations drops with increasing density, which can
be explained by the increase in the fraction of immobile cations on the clay sur-
faces, that is, in the electrical double layers mentioned earlier in this chapter, when
the water content drops (Muurinen 1994; Pusch 1995). It has been proposed that
improved linking of the double layers of adjacent particles would yield an increase in
diffusivity, but there is no proof of this hypothesis. Instead, the microchannels, that
is, the system of interconnected zones of low clay density, may totally control diffuse
transport of cations and anions.
The effect of density is most obvious when the cation has low valency and large
radius, which means that Cs, which is monovalent and larger than the divalent Sr,
responds more strongly to the increase in density. A general opinion is that surface
diffusion of cations dominates over pore diffusion, whereas anion diffusion only takes
place as pore diffusion. This hypothesis can be assessed by comparing the surface dif-
fusion capacity with the clay density as shown in Table 4.17. One sees no differences.
We see from Figure 4.26 that the diffusion coefficient for monovalent anions
moving in montmorillonite-rich buffer clay is up to 100 times lower for a bulk den-
sity of 2100 kg/m3 than for 1100 kg/m3, whereas the fraction of interlamellar water
for these densities are 95% and 10%, respectively, that is, about 10 times lower for the
denser clay. Assuming a constant inverse ratio of diffusion coefficient and content
of interlamellar water, this would suggest that migrating anions can move through
the interlamellar space but rather than rejecting Donnan’s principle, one would favor
the concept of channeling. The microchannels have a cross-sectional area that varies
from more than 50% of an arbitrarily taken section of smectite clay with a density
at water saturation of 1100 kg/m3 to 1% for a density of 2100 kg/m3. This differ-
ence indicates that the role of channels containing soft clay gels or being empty is
very important both for the bulk hydraulic conductivity and for ion diffusion. Anion
transport by diffusion is hence believed to take place very heterogeneously in the
clay matrix of artificially prepared clay seals.
Considering cations, the effective diffusion coefficient for monovalent cations is
nearly three times higher for a bulk density of 2100 kg/m3 than for a density of
1100 kg/m3, whereas the fraction of interlamellar water for these densities are 95%
and 10%, respectively. In contrast to anions, this indicates that diffusive transport
capacity of such cations is as high in free porewater as in the interlamellar space.
TABLE 4.17
Diffusion Parameters for Cs-137a
ρ (kg/m3) Porosity (%) Da (m2/s) Ds (m2/s)
1060 0.94 1.00 × 10 −11 9.80 × 10−12
1270 0.85 7.15 × 10−12 7.05 × 10−12
1580 0.67 3.24 × 10−12 3.21 × 10−11
1780 0.64 4.70 × 10−13 4.67 × 10−13
1960 0.56 1.13 × 10−13 1.12 × 10−13
Source: Brandberg, F., Skagius, K., Porosity, sorption and diffusivity

data compiled for the SKB 91 study, SKB Technical Report
TR 91-16, SKB, Stockholm, Sweden, 1991.
Da is the apparent diffusion coefficient; Ds is the surface diffusion
coefficient.
a K = 0.66 m3/kg.
d
78 Bentonite Clay
4.5.5 Test Principles and Theory

4.5.5.1 Technique
Experimental determination of the diffusivity can be made in several ways but in
practice there are two procedures that can be recommended: the nonsteady (pro-
file) method and the steady-state (through-diffusion) method. Figure 4.27 shows the
1.0
0.9
0.8
0.7
Relative concentration
0.6
0.5 5.0E−12
0.4
1.0E−12 1.0E−11 m2/s
0.3
0.2
0.1
0.0
0 2 4 6 8 10 12 14 16 18 20
x (mm)
Theoretical curves Measured values Cu
High-concentration Low-concentration
side side
Water in Water in
Water out Water out
Sinter
Bentonite
Tube
20
FIGURE 4.27 Principle and results of measuring the diffusion rate of dissolved copper in
dense smectite clay. The concentration gradient is maintained by circulating the respective
solution through the sintered filters. The evaluated apparent coefficient was about E−11 m2/s.
(Data from Pusch, R., Waste Disposal in Rock, Developments in Geotechnical Engineering,
76, Elsevier, New York, 1994; After Eriksen.)
principle of measurement and an example of recorded diffusion of copper ions in

dense smectite-rich clay (MX-80) using the steady-state method.
The measurements give the apparent diffusivity, which can be evaluated from
Equation 4.15:
C x
= erfc (4.15)

C0 2 Da t
where:
C is the ion concentration in the sample (mol/m3)
C0 is the ion concentration in the sample at x = 0 (mol/m3)
x is the distance from the surface (m)
t is the experimental time (s)
At termination of a steady-state diffusion experiment, the sample is sliced and the con-
centration profile determined. This gives the effective and apparent diffusivities, which
can be evaluated in various ways as reported in the literature (Pusch 1995):
Ql
De = (4.16)
A ( Cw1 − Cw 0 )

Ql
Da = (4.17)
A ( C1 − C0 )
l2
Da = (4.18)
6T
where:
Q is the flux through the sample (mol/s)
l is the thickness of the sample (m)
A is the surface area perpendicular to the diffusion flux (m2)
Cw0 is the concentration of the solution on the low-concentration side (mol/m3)
Cw1 is the concentration of the solution on the high-concentration side (mol/m3)
C0 is the concentration in the porous medium on the low-concentration side (mol/m3)
C1 is the concentration in the porous medium on the high-concentration side (mol/m3)
T is the time lag, the point where the asymptote of the breakthrough curve inter-
cepts the time axis (s)
The basic expressions for ion diffusion can be applied only if there is no chemical
reaction between the diffusing elements and the soil components. The risk of such
interaction is particularly obvious if the experiments are made in the common form
of contacting the sample with solutions that contain dissolved carbon dioxide, that is,
without de-airing them.
The indication that surface diffusion can be a major process in clays with high
electrostatic potential, such as smectite-rich clay, would suggest that diffusion of cat-
ions is significant even at low water contents but not in air-dry clay in which sorbed
water does not form a continuum.
80 Bentonite Clay
4.5.5.2 Role of Electrical Double Layers for Ion Diffusion

Double-layer interaction is a key factor for the swelling pressure exerted on the con-
finement of expandable clays and for the diffusive transport of dissolved elements.
We have seen in Chapter 3 that no more than three hydrates are established in mont-
morillonite stacks of lamellae and that two hydrated layers exist in montmorillonite
clay when Ca and most other cations are sorbed. This naturally affects the paths
taken by diffusing cations.
4.6 COLLOID TRANSPORT

The issue of colloid transport through buffers is considered to be an important factor
for the transport of radionuclides from repositories to the biosphere but has not yet
been examined in sufficient detail. In principle, some general conditions have been
defined, as for minute smectite particles, which, through their ability to sorb posi-
tively charged radionuclides and other cations, can play a role in bringing contami-
nants to the biosphere. Such particles, like many other constituents of colloidal size,
carry different electrical charges on different parts of their surfaces and hence form
aggregates, gels, and networks that easily couple to the dense buffer clay (Pusch
1999a). Release of smectite aggregates can take place if the porewater flow rate is
sufficiently high, but this is estimated to be of negligible importance under the very
low regional hydraulic gradients that prevail in the rock after closing the repository.
Buffer clay will tend to enter fractures in the rock and the front of fracture-penetrating
clay may be very soft, so that groundwater flowing in the fractures can tear off small
aggregates and transport them if the flow rate exceeds a few millimeters per second.
Experiments as well as experience demonstrate that their size may be up to 2–50 μm.
Their surface charge makes them interact with rock fracture minerals and become fixed
to them, particularly to chlorite and micas, and it is, therefore, believed that the trans-
port of smectite aggregates will be insignificant over long distances (Pusch 1994).
4.7 MICROBIOLOGICAL FILTERING

Bacteria, of which the sulfate-reducing ones are of particular importance to copper
canisters, may stay alive and multiply in soft smectite clay. They usually carry a
negative surface charge or no net surface charge at all and can move through clay
like large organic molecules such as humus colloids. There are limitations because
they interact by establishing chemical and hydrogen bonds with the basal surfaces
and edges of the smectite stacks, and because of the very small void space in dense
buffer clay (Pusch 1994). It is generally assumed that the small size of interconnected
voids and the limited access to free water prevent bacteria from moving and staying
alive in smectite-rich clay with higher density at water saturation than 1800 kg/m3 of
montmorillonite-rich clay of MX-80 type. Only spores are, therefore, believed to be
operative, but the working hypothesis is that there will ultimately be eradication of
all life forms in the buffer. However, applying the concept of microchanneling on the
microstructural scale, one can still imagine that both bacteria and spores can migrate
through even dense buffer clay and contribute to the corrosion of metal canisters.
4.8 HEAT TRANSPORT—THERMAL PROPERTIES

4.8.1 Overview
The heat conductivity of clay seals can be of great practical importance (Börgesson
et al. 1994). Thus, the buffer clay surrounding HLW canisters shall transfer heat to the
surroundings very effectively for avoiding overheating of the canisters and for avoiding
temperatures that can cause damage to the clay itself. We will confine ourselves here
to transfer some basic information on the matter, assuming that the clay stays chemi-
cally intact. The matter of longevity of clay seals will be the focus of Chapters 6 and 8.
4.8.2 Influence of Water Content

The heat conductivity depends primarily on the amount of water in the clay.
Commercially available bentonites such as MX-80 is delivered with a water content
of 10%–15%, which corresponds to one to two interlamellar hydrate layers, and com-
paction of such material to a dry density of 1500 kg/m3 gives it a heat conductivity
of about 0.75 W/m K (Pusch et al. 1995). Various tests have shown that an increase
in the degree of water saturation from 50% to 100% raises the heat conductivity by
around 100%.
Table 4.18 illustrates the influence of the degree of water saturation on the heat
conductivity of MX-80 clay and Czech RMN clay as reported by the Czech Technical
University in Prague (CEG).
The reported data show that the RMN clay has a somewhat higher heat conductiv-
ity than MX-80 at incomplete fluid saturation while the difference is insignificant at
a high degree of saturation. Differences may be due to different particle organiza-
tion, thickness, and interaction of stacks of lamellae, as well as to the distribution
of porewater, which is not the same for equally smectite-rich clays in sodium and
calcium forms.
4.8.3 Influence of Smectite Content

A number of smectite-rich clays have been investigated with respect to their thermal
properties, and the investigations have led to some understanding of the importance
of the mineral composition. Thus, addition of quartz powder and graphite have been
TABLE 4.18
Thermal Data of the Montmorillonite-Rich RMN and MX-80 Clays
Dry Density Degree of Water Heat Conductivity,
Clay Type (kg/m3) Saturation, Sr (%) λ (W/m, K)
RMN 1475 65 1.07
RMN 1597 99 1.22
MX-80 1570 70 0.90
MX-80 1700 98 1.14
82 Bentonite Clay
shown to significantly increase the heat conductivity, but these additives may cause
difficulties with respect to the distribution in the mixture and—for graphite—raise
the question if this crystalline material may be a nutrient to bacteria. We, therefore,
leave them out of our discussion of smectite performance.
As to clays other than MX-80 for use as buffer, there is information from com-
prehensive testing of a French buffer candidate clay FoCa7 and from ongoing tests
of the Czech candidate clay RMN as exemplified by Table 4.18. The FoCa7 clay
contains about 90% mixed-layer kaolinite/smectite, 5% kaolinite, and 5% smectite
in beidellite form, and is reported to have a thermal conductivity of about 1.6 W/m,
K at 50%–70% degree of water saturation, which made it a primary buffer candidate.
Its microstructural constitution is similar to that of Ca smectite.
4.8.4 Influence of Stress and Temperature

The effective stress and temperature are known to have a certain but rather small
effect on the heat conductivity. A high effective stress implies a high density, which
raises the thermal conductivity. Increased temperature causes a reduction in swelling
pressure, which would reduce the thermal conductivity except if there is an external
pressure that causes consolidation, preserved density, and improved continuity of
thermally conductive elements including porewater. Under a thermal gradient, heat
transfer will be aided by convection, which is most important at low degrees of fluid
saturation.
4.9 COUPLINGS, DEPENDENCIES, AND INTERDEPENDENCIES

Given a heat source and water availability, tracking of the movement of heat and
water in a smectite buffer mass requires one to be aware not only of the various
coupled relationships between heat and fluid fluxes but also on the interdependent
relationships between these and the various internal sets of chemical and mechanical
forces. Water movement in response to a thermal gradient is well appreciated, and a
simple Fickian-type relationship can be written to express this phenomenon—in the
absence of all other factors, forces, processes, and events. However, as indicated in
Chapter 3, changes in water content of a smectite clay buffer in the repository system
are due not only to thermal gradients but also to the various internal gradients. The
effects and outcome of the actions generated for many of these are interdependent.
A brief summary of some of the more prominent factors or events that deserve atten-
tion tells us the following:
• Water content changes at any one point in the clay buffer system arise from
the thermal gradient, evaporation, and condensation, and from internal
energy gradients.
• Swelling of the smectite buffer wetted from one end is primarily a function of
the ambient T, T gradient, concentration c, and of the compressibility of the
unwetted portion.
• Changes in internal energy characterized by the total potential ψ, at any one
point i in the smectite clay barrier are functions of the ambient θ, T, and c.
An obvious example of the interdependencies concerns the concentration of solutes

in pore water. It is considered important since it affects (1) the osmotic (swelling)
pressure, (2) the vapor pressure of the pore water, and (3) because it participates
strongly in reactions with reactive surfaces such as basal planes of smectite particles.
The contributions from vapor and thermocapillary transfer of liquid water under a
temperature gradient for a closed soil-column system, represented by the canister-
embedded clay in a HLW repository, have shown that while transfer of liquid water
takes place toward the cold end, dissolved solutes move in the opposite direction,
creating precipitation of gypsum or chlorides. Vapor movement has been found to be
four to five times faster than that predicted by Fickian diffusion.
For the theoretically minded reader, it should be obvious that a useful technique
to account for coupled effects is to apply the principles of irreversible thermody-
namics to cast the governing relationships. Taking the simple case of heat and mass
transfer in the initially unsaturated buffer mass, the second postulate of irreversible
thermodynamics allows one to write the governing relations. Following Yong, one
can define the major relation as follows:
⎧ ∂θ ⎫
⎧Qw ⎫ ⎡ Lww LwT ⎤ ⎪⎪ ∂x ⎪⎪
⎨ ⎬=⎢ ⎨ ⎬ (4.19)
⎩ QT ⎭ ⎣ LTw LTT ⎥⎦ ⎪ ∂T ⎪
⎪⎩ ∂x ⎪⎭
where:
Qw = (∂θ ∂t ) is the fluid flux
QT = (∂T ∂t ) is the heat flux
∂θ ∂x and ∂T ∂x are the thermodynamic forces due to water content gradient and
temperature gradient, respectively
The phenomenological coefficients are described as follows:
Lww is the diffusion coefficient for fluid flow due to gradient of θ
LTT is the thermal conductivity coefficient for heat transfer due to gradient of T
LwT and LTw are the coupling coefficients
Mass and energy conservation relationships can be specified as follows:
∂θ ∂Q ∂T ∂Q
= − w and Cv = − T (4.20)
∂t ∂x ∂t ∂x
where:
Cv is the specific heat of the buffer material
Accordingly, the governing relationships will be obtained as follows:
⎧ ∂θ ⎫ ⎡1 ⎧ ⎧ ∂θ ⎫⎫
0 ⎤ ⎪
⎪⎪ ∂t ⎪⎪ ⎢ ∂ ⎪ ⎡ Lww LwT ⎤ ⎪⎪ ∂x ⎪⎪⎪⎪
⎨ ⎬=⎢ 1 ⎥⎥ ⎨ ⎢ ⎨ ⎬⎬ (4.21)
⎪ ∂T ⎪ ⎢0 ∂x L Tw LTT ⎥⎦ ⎪ ∂T ⎪⎪
Cv ⎥⎦ ⎪ ⎣
⎪⎩ ∂t ⎪⎭ ⎣ ⎪⎩ ⎪⎩ ∂x ⎪⎭⎪⎭
84 Bentonite Clay
These governing relationships have been used to determine the phenomenological

coefficients. The procedure requires us to assume that the phenomenological coef-
ficients are of certain types of functions of θ and T. They can be derived from tests
under different temperatures using porewater with different chemical compositions.
The reader realizes that we have already made use of them in this chapter and there
is more to come.
REFERENCES
Al-Taie, L., Pusch, R., Al-Ansari, N., Knutsson, S. 2013. Hydraulic properties of smectite
clays from Iraq with special respect to landfills of du-contaminated waste. Journal of
Earth Sciences and Geotechnical Engineering, 3:109–125.
ASTM Standard D 5084, 1994. Standard test method for measurement of hydraulic conductiv-
ity of saturated porous materials using a flexible wall permeameter. American Society
for Testing and Material, West Conshohocken, PA.
Börgesson, L., Fredrikson, A., Johannesson, L.-E. 1994. Heat conductivity of buffer materials.
SKB Technical Report TR 94-29, SKB, Stockholm, Sweden.
Brandberg, F., Skagius, K. 1991. Porosity, sorption and diffusivity data compiled for the SKB
91 study. SKB Technical Report TR 91-16, SKB, Stockholm, Sweden.
Cheung, S.C.H., Yu, R. 2001. The effects of density and hydraulic gradient on clay perme-
ability. In Clay Science for Engineering: Proceedings of the International Symposium
on Suction, Swelling, Permeability and Structure of Clays, Adachi, K., Fukue, M. (Eds.),
A.A. Balkema Publishers, Brookfield, VT.
Hansbo, S. 1973. Influence of mobile particles in soft clay on permeability. Proceedings of the
International Symposium of Soil Structure, Swedish Geotechnical Society, Gothenburg,
Sweden.
Hellström, J.G.I. 2009. Internal erosion in embankment dams, fluid flow through and deforma-
tion of porous media. PhD Thesis, Luleå University of Technology, Sweden.
Horseman, S.T., Harrington, J.F. 1997. Study of Gas Migration in MX-80 Buffer Bentonite: British
Geological Survey. Report WE/97/7. National Environmental Research Council, Swindon.
Kato, H., Muroi, M., Yamada, N., Ishida, H., Sato, H. 1995. Estimation of effective diffusiv-
ity in compacted bentonite. In Scientific Basis for Nuclear Waste Management XVIII:
Proceedings of the Materials Research Society Symposium, vol. 353, Murakami, T.,
Ewing, R.C. (Eds.), Materials Research Society, Pittsburgh, PA, pp. 277–284.
Kim, H.-T., Suk, T.-W., Park, S.-H., and Lee, C.-S. 1991. Diffusivities for ions through com-
pacted Na-bentonite with varying dry density. Waste Management, 13:303–308.
Muurinen, A. 1994. Diffusion of anions and cations in compacted sodium bentonite. PhD
Thesis, University of Helsinki, Helsinki, Finland.
Muurinen, A., Lehikoinen, U. 1995. Evaluation of phenomena affecting diffusion of cations in
compacted bentonite. Report YJT-95-05. Voimayhtiöiden Ydinjätetoimikunta, Helsinki,
Finland.
Nagy, N., Moémi, M., Kónya, J. 2010. Interfacial Chemistry of Rocks and Soils. CRC Press,
Boca Raton, FL.
Neretnieks, U., Moreno, L. 1993. Fluid flow and solute transport in a network of channels.
Journal of Contaminant Hydrology, 14:163–182.
Osipov, V.I., Sokolov, V.N., Eremeev, V.V. 2004. Clay Seals of Oil and Gas Deposits. A.A.
Balkema Publishers, Atington.
Pusch, R. 1994. Waste Disposal in Rock. Developments in Geotechnical Engineering, 76.
Elsevier, New York.
Pusch, R. 1995. Handbook of Buffers and Backfills, Part I. SKB Arbetsrapport AR 95–45.
Pusch, R. 1999a. Clay colloid formation and release from MX-80 buffer. Technical Report
TR-99-31. SKB, Stockholm, Sweden.
Pusch, R. 1999b. Experience from preparation and investigation of clay for microstructural
analysis. Engineering Geology, 54(1/2):187–195.
Pusch, R. 2001a. Can the water content of highly compacted bentonite be increased by apply-
ing a high water pressure? SKB Technical Report TR-01-33. SKB, Stockholm, Sweden.
Pusch, R. 2001b. Experimental study of the effect of high porewater salinity on the physi-
cal properties of a natural smectitic clay. SKB Technical Report TR-01-07. SKB,
Stockholm, Sweden.
Pusch, R. 2001c. The microstructure of MX-80 clay with respect to its bulk physical proper-
ties under different environmental conditions. SKB Technical Report TR-01-08. SKB,
Stockholm, Sweden.
Pusch, R. 2008. Geological Storage of Radioactive Waste. Springer-Verlag, Berlin, Heidelberg.
Pusch, R., Muurinen, A., Lehikoinen, J., Bors, J., Eriksen, T. 1999. Microstructural and chemi-
cal parameters of bentonite as determinants of waste isolation efficiency. Final Report of
EC-contract No.:F14W-CT95-0012. The EC Commission, Brussels, Belgium.
Pusch, R., Ramqvist, G., and Hedin, M. 2011. Final report of experiments with rock blocks
interacting hydraulically with smectitic pellet fills. SKB Research Report R-11-26.
Pusch, R., Weston, R. 2003. Microstructural stability controls the hydraulic conductivity of
smectitic buffer clay. Journal of Applied Clay Science, 23:35–41.
Pusch, R., Yong, R.N. 2006. Microstructure of Smectite Clays and Engineering Performance,
1st edn. Taylor & Francis, London.
Pusch, R., Yong, R.N., Nakano, M. 2011. High-Level Radioactive Waste (HLW) Disposal.
WIT Press, Southampton.
Sposito, G. 1984. The surface chemistry of soils, Oxford University Press, New York.
Wold, S., Eriksen, T. 2000. Diffusion of organic colloids in compacted bentonite. SKB
Technical Report TR-00-19. SKB, Stockholm, Sweden.
Yong, R.N., Mohamed, A.M.O. 1992. A study of particle interaction energies in wetting of
unsaturated expansive clays. Canadian Geotechnical Journal, 29:1060–1070.
Yong, R.N., Nakano, M., Pusch, R. 2010. Containment of High-Level Radioactive and
Hazardous Solid Wastes with Clay Barriers. Spon Press, New York.
5 Physical Performance
of Smectite Clay Seals
Prediction and Reality
This chapter deals with design principles for clay seals and prediction of their function
for isolation of waste and redirecting groundwater flow. The basis for this is modeling
and calculation of the performance using laboratory-determined material properties
and data from field testing, mainly concerning Temperature/hydraulic/Mechanical/
chemical (THMCB). The long-term performance of smectite seals with focus on chem-
ical issues is discussed in Chapters 6 and 8.
5.1 INTRODUCTION
We have seen that the fundamental properties of smectite clays for use as barriers
to flow of water and migration of contaminants are the hydraulic conductivity and
ion diffusivity and that they are strongly dependent on the density in bulk and on
the microstructural scale. The microstructure and the relative amounts of free and
interlamellar porewater are determinants of rheological properties that are in focus
in this chapter. The physical data are valid for clay in equilibrium with the confine-
ment with respect to volume and temperature, but in practice, the physical properties
are transient and develop slowly in the course of hydration and expansion that depend
on the hydraulic and mechanical interactions with the surroundings. We will start by
going through definitions and see how clay seals evolve under defined conditions, and
conclude by showing how they perform in practice. A necessary basis of all this is soil
physics, clay mineralogy, and practical geotechnical engineering.
5.2 APPLICATION OF CONCEPTS OF SOIL

MECHANICS TO SMECTITE CLAY
5.2.1 Effective Stress Concept
The effective stress concept is the basis of all sorts of stress-related expressions used
in soil mechanics. It states that the effective stress, defined according to Equation 5.1,
controls the rheological behavior (Terzaghi et al. 1996).
σʹ = σ − u (5.1)
where:
σ′ is the effective stress
σ is the total stress
u is the pore water pressure
87
88 Bentonite Clay
Equation 5.1 has proven to be valid for a variety of inorganic soils, from sand to
illitic clay as well as for organic soils such as peat, ooze, and slush. However, the
complex microstructure of very fine-grained clayey soils and poorly known consis-
tency and physical properties of the interlamellar water makes the validity of this
concept somewhat uncertain as concerns smectite clays (Sposito and Prost 1982;
Sposito 1984).
According to the classical view, the total stress is defined as follows:
N A A − As
σ= = p s +u (5.2)
A A A
where:
N is the total force on the respective sides of a cubical soil element
A is the total area of the sides of the cube
As is the area of the particle contacts in a cross section of the cube
p is the effective contact pressure, also called grain pressure
u is the pore water pressure
The contact pressure p is very high due to the very small contact area As. The product
pAs is the total force acting over the area, and p ( As A ), the so-called effective stress
σ′c acting on this area (index c stands for contact). Introducing ( As A ) = b the equa-
tion is written as follows:
σ = σʹc + u(1 − b) (5.3)
The effective stress theory is valid if As is very small compared to the total area,
that is, b ≈ 0, which approximately applies to ordinary soils. For smectite clay, the
question is how As shall be interpreted since it is believed that there is commonly
no direct crystal-to-crystal contact. The question is rather what the state is of the
water within and between the lamellar stacks and how the stresses are transferred
within and between them. The water in the microvoids between the stacks of lamel-
lae is Newtonian and reacts on pressure changes as ordinary free water, for which
Poisson’s ratio ν = 0.5. This would imply that the effective stress theory is valid
and b = 0 but as we have seen in preceding chapters of this book, the interlamellar
water dominates especially at high clay densities and that b > 0. This fraction of
the porewater has a high viscosity and exerts a pressure—manifested by the macro-
scopic swelling pressure—on the confinement. The pressure is due to electrochemi-
cal forces in the electrical double layers of adjacent stacks of lamellae and in the
interlamellar space.
For such a system, one can write, for the total stress in one direction (Pusch et al.
1995)
Fr − Fa
σ= + u(1 − b) (5.4)
A
where:
Fr is the sum of the repulsive forces in a cross section
Fa is the sum of the attractive forces in a cross section
Physical Performance of Smectite Clay Seals 89
For this system ( Fr − Fa ) A controls the behavior. Thus, the stress in the structure
over the cross-sectional area A can be considered as an effective stress σʹe (where
index e stands for effective) by which Equation 5.4 turns into Equation 5.5:
σ = σʹe + u (5.5)
A commercial bentonite commonly has a content of nonsmectite minerals of 20%–50%
depending on its origin (Grim and Gueven 1978) (cf. Chapter 3). The content of non-
expandable minerals, called contaminants by purists, means that the interaction of
many particles is different from that of smectites, implying that there are lots of direct
mineral-to-mineral contacts especially at high densities, and the resulting stress is
therefore a combination of Equations 5.2 and 5.4, yielding Equations 5.6 and 5.7:
N A F − Fa A − As
σ= =p s+ r +u (5.6)
A A A A
σ = σʹc + σʹe + u(1 − b) (5.7)
The average of the effective stress is thus the sum of the contact and the repulsive
electrical double-layer pressures:
σʹ = σʹc + σʹe (5.8)

The parameter b would theoretically be up to 0.9 and even higher for very dense smec-
tite clay, hence implying that the total pressure would be nearly the same as the effective
stress. The reason for this can be that smectite particles with their interlamellar water
perform as solid clay particles. The interested reader is encouraged to further investigate
this complex issue, requiring consideration of the clay microstructure; the chemical
potential of water; and the stress–strain properties of smectite minerals, interlamellar,
and free (extralamellar) porewater (cf. Graham et al. 1992). One of course realizes that
while the concept effective stress has a real meaning for sandy soils for which one can
imagine the stress transfer between grains that have mineral-mineral contacts, it is ficti-
tious for smectites. Still, it turns out to work in practical design and construction also for
these clays. The difference between illitic and chloritic clays and smectite ones is that
the time for any imposed strain or change in external pressure on water-saturated ele-
ments to be fully developed is very long for the last mentioned. For example, complete
consolidation of a 20-mm-thick soft montmorillonite sample in an ordinary oedometer
by doubling the pressure can take many months or even years. This has led impatient
laboratory people to give up experiments with commonly sized sample sizes and to test
only tiny samples. These may, however, not be representative of larger volumes.
5.2.2 Role of the Effective Stress for the Physical Stability of Clay Seals
The effective pressure determines the mechanical strength as well as compression
and expansion of soils. At the bottom of the deepest oceans, the porewater pressure
in sediments is very high, about 100 MPa* but the mechanical strength is zero since
* This corresponds to 1000 kg/cm2, which is 10 times the pressure on the floor exerted by a very heavy
lady standing on one of her high-heeled shoes.
90 Bentonite Clay
Groundwater H1
Ground surface level
H1
Groundwater level
H1
Groundwater
level
H2
H2
H2
u = H2 u = H2 u = H2
σ = (H1 + H2)ρ σ = (H1 + H2)ρ σ = (H2 − H1)ρ
σ′ = σ − u = H1ρ + H2(ρ − 1) σ′ = σ − u = H1ρ + H2(ρ − 1) σ′ = σ − u = (H2 − H1)ρ − H2
= (H2ρ − 1) − H1
FIGURE 5.1 Total stress (σ), porewater pressure (u—piezometric head), and effective
stress (σ′) at the lower end of a standpipe for different groundwater levels.
the effective pressure, being the product of the depth below the sea bottom and the
density of the soil, is zero. On land, the effective pressure depends on the depth
below the ground surface and on the groundwater level (Figure 5.1).
The left case in the figure represents common conditions in sediments whereas the
one in the middle shows the stage reached at equilibrium after lowering the ground-
water level, implying a rise in effective stress at the base of the imaginary standpipe.
The right case represents artesian (overpressure) conditions. For H1 = [H2(ρ−1)],
where ρ is the density of the soil with its content of water, the effective stress at this
level is zero meaning that the soil grains are not pressed together and that the soil is
physically unstable. The density of the water is set at unity here.
5.2.3 Stress–Strain Definitions* and Parameters

The rheological behavior of clays can be described as being partly viscoelastic, that
is, with delayed recoverable strain, and partly plastic with unrecoverable strain.
Since, as we just described, any change in stress of a water-saturated smectite clay
element requires a long time to develop the corresponding strain, its rheological
behavior is in practice best described as being visco-elastoplastic.
Two basic types of stress–strain moduli are commonly used: the compression
modulus K and the shear modulus G. The first represents changes in volume and the
second changes in shape (Pusch et al. 1995).
* Described in many textbooks such as Taylor 1948; Pusch 1995; and Terzaghi et al. 1996.
5.2.3.1 Compression Modulus K

Compression is expressed by use of K as defined by Equation 5.9:
d σʹ d σʹ
K= = (5.9)
d ε v d ε1 + d ε2 + d ε3
where:
σ′ is the effective stress
ε v is the volumetric strain
ε1, ε2, and ε3 are linear strains
K can be determined by isotropic, drained triaxial tests*
5.2.3.2 Shear Stress Modulus G

Use of shear boxes imply simple shear:
dτ
G= (5.10)
dγ
For linearly elastic materials, the following relationships between these basic moduli
and Poisson’s ratio v have been derived:
3G
E= (5.11)
1 + G 3K
1 − 2G 3 K
v= (5.12)
2 + 2G 3 K
The modulus of elasticity E can be determined by uniaxial compression tests and the
G modulus can be evaluated from E and K. E can also be derived from undrained
triaxial tests and expressed in terms of the tangent modulus:
d ( σʹ1 − σʹ3 ) d ( σ1 − σ3 )
E= = (5.13)
dε dε
where:
σ1 is the major principal stress
σ3 is the minor principal stress (cell pressure)
5.2.3.3 Oedometer Modulus M

The oedometer modulus (confined compression) is given by
d σʹ
M= (5.14)
dε
* This and other standard tests are described in many textbooks on soil mechanics (cf. Pusch 1995,
Terzaghi et al. 1996).
92 Bentonite Clay
M can be derived from K and G:

4G
M=K+ (5.15)
3
Plotted oedometer and shear strain curves are hardly ever straight lines on any scale
and M and G often obey the following general expression, where D is the deforma-
tion modulus:
B
⎛ σʹ ⎞
D = Aσ j ⎜ ⎟ (5.16)
⎝ σj ⎠
5.2.3.4 Compression Properties

In its most simple form, the compressibility of soils is expressed in terms of the
compression index Cc or the index ε2, which can be evaluated from oedometer tests
(cf. Chapter 3). One finds for linearly scaled void ratio and log-scaled pressure that
Δe
Cc = (5.17)
Δ ( log σʹ )
and if the load is doubled
ε2 (1 + e0 )
Cc = (5.18)
log 2
where:
e0 is the void ratio before the load increase
5.2.3.5 Time Dependence

Compression of water-saturated soil always proceeds as a function of time because
the generated porewater overpressure has to be expelled in order to give com-
pression, which is called primary consolidation, and, after the dissipation of this
pressure, creep effects dominate, called secondary consolidation (Terzaghi et al.
1996). The diffusion-like process of primary consolidation is expressed in ordinary
soil mechanics by use of the consolidation coefficient, cv, which is evaluated from
recording of the compression as a function of time. Conventionally, stepwise load-
ing is made but in modern soil mechanics the compression is often performed with
constant rate of strain and recording of the porewater pressure (Daniel et al. 1994).
The difference is small but it is questionable whether the constant-rate testing really
allows the creep mechanisms to develop without hindrance. We therefore focus here
on the classical technique of step-wise loading.
The time-dependence of the compression (consolidation) is obtained from the two
parameters cv and Tv, which are related according to Equation 5.19. For each load
increment at stepwise loading, Tv is evaluated as
Compression after time t

Tv =
Total compression after completed primary consolidation
Tvd 2
cv = (5.19)
t
where:
cv is the consolidation coefficient
t is the time after onset of consolidation
d is the half thickness of the soil sample or the soil layer (for two-sided drainage)
Secondary consolidation, developed when the porewater overpressure caused by the

primary has dissipated, can be expressed by use of the coefficients αs and Cα in the
following way, using logarithmic time scales:
dε
αs = (5.20)
d (log t )
de
Cα = (5.21)
d (log t )
where:
ε is the compressive strain
e is the void ratio
5.2.3.6 Shear Strain

Shear strain can be evaluated from direct shear tests and from triaxial tests. We will
confine ourselves to describe the latter.
The stress conditions in the sample are defined by the effective average normal
stress pʹ = ( σ1ʹ + 2σʹ3 ) 2 and the deviator stress, q = σ1ʹ − σ3ʹ , which are changed by
altering the major principal total stress σ1—keeping the cell pressure σ3 (minor prin-
cipal total stress) constant.
The angular strain and the volumetric strain are measures of the shear strain and
related to the shear stress as defined in Figure 5.1.
Fs
Fn
δh
Area A δz
zi γ
Fs
Fn
Normal stress σ = Fn/A Normal strain ε = δz/zi

Shear stress τ = Fs/A Shear strain γ = δh/zi
94 Bentonite Clay
10,000
5,000
1,000
500 p/q relationship
from triaxial tests
of smectite-rich clay:
q (kPa)
upper boundary—
100 salt clay;
50 lower boundary—
freshwater clay
10
5
1
1 5 10 50 100 5001,000 5,00010,000
p (kPa)
FIGURE 5.2 Relationship of p/q for smectite-rich clay. For a given effective pressure p the
maximum shear stress is found in the band that indicates that the shear stress at failure can
be twice as high for salt clay than for clay with low-electrolyte porewater. The average inter-
nal friction angle is roughly 30°. (Data from Börgesson, L. et al., Laboratory Investigation
of Highly Compacted Bentonite Blocks for Buffer Material. Compaction Technique and
Material Composition, SKB Djupförvar Projekt Rapport PR 44-93-009, Swedish Nuclear
Fuel and Waste Management [SKB], Stockholm, Sweden, 1993.)
Clay seals like those embedding canisters with hazardous waste and the ones
used as liners in landfills will be exposed to deviator stresses, and the prediction of
the physical stability requires use of methods for calculating the stress conditions
for determining the degree of safety. The fundamental parameters for this are p′ and
q′, which are related as exemplified by the graph in Figure 5.2 that refers to triaxial
shear tests of smectite-rich clay (MX-80).
5.3 ROLE AND MECHANISMS OF CONSOLIDATION

AND CREEP
5.3.1 Cases
Applying simple soil mechanical rules, we find that the mechanical behavior of
smectite clays is controlled by effective stress (Terzagi et al. 1996). In this chapter,
we will consider a number of practical cases where stress-generated strain is of prac-
tical importance, such as the settlement of heavy silos and canisters containing haz-
ardous waste, and the time-dependent shear strain of top liners of landfills. We will
be acquainted with concepts for disposal above and below ground and at very large
depths.
5.3.2 Consolidation
Let us assume that the initial average density of water-saturated buffer clay in a
KBS-3V hole surrounding a canister with highly radioactive waste is 2000 kg/m3;
we know from Chapter 3 that this corresponds to an average amount of interlamel-
lar, immobile, water of about 90% and to an average effective (swelling) pressure of
about 5 MPa. A 20,000 kg canister with a basal surface of 1.13 m2 adds 0.177 MPa
to the existing 5 MPa, that is, an increase by 3.5%. This can cause some expulsion
of free porewater in the voids between the stacks of lamellae from the buffer, but
the time for this will be extremely long as realized by applying common classical
consolidation theory (Taylor 1948). This theory, which is basically an application of
Fick’s second law, expresses the time for evening out the generated porewater over-
pressure as being inversely proportional to the square of the average length of pore-
water drain paths (Equations 5.19 through 5.21). Since this length is many meters in
the deposition hole, the time for consolidation is hundreds to thousands of years and
hence of no practical significance at all.
5.3.3 Shear Strain by Creep

Even very dense smectite-rich buffers and backfills are apt to deform rather much
with time when exposed to a deviator stress because of the weak contacts of adja-
cent stacks of lamellae and because interlamellar water, despite its high viscosity,
undergoes shear strain (Pusch and Yong 2006). This is claimed by modern creep
theory that we will refer to after having a look at some attempts of various inves-
tigators to work out empirically derived expressions of shear strain versus shear
stress.
Such a model has been proposed by Singh and Mitchell (1968):
dγ
γ = = Ae ατ ⋅ t − n (5.22)
dt
where:
γ is the rate of shear strain
γ is the shear strain
τ is the shear stress
t is the time
A, α, and n are constants
Our main interest is to find out how the creep rate is related to the microstructural
response of a smectite clay element to a constant shear stress, and we will do this
by using a thermodynamically based creep theory following principles outlined by
Pusch (1979), Pusch and Feldham (1980), and Pusch and Adey (1999).
Let us start with the assumption that the microstructural constitution is in equi-
librium when the load is applied, and that it undergoes changes by shear-induced slip
within and at the contacts of aggregates of smaller units, that is, the stacks of lamel-
lae. Such a slip takes place under constant volume conditions or in conjunction with
96 Bentonite Clay
Slip units of aligned τ

stacks of lamellae
Dense parts of expanded

granules (aggregates)
τ
FIGURE 5.3 Late stage of evolution of shear strain of networks of clay particles under
a macroscopic shear stress τ. Slip domains are formed by breakdown of weak microstruc-
tural units. (Data from Pusch, R., Feltham, P., A stochastic model of the creep of soils,
Geotechnique, 30, 497–506, 1980.)
consolidation. For smectite clay, lacking true mechanical friction between particles,
creep is stronger than for nonexpandable clays but has the same character of stochas-
tically distributed slip occurring where and when energy barriers are overcome. This
brings us to the use of stochastical mechanics.
The microstructural response to compression or shearing is that the overall defor-
mation of the network of particles changes by disintegration, translation, and rotation
of weaker aggregates, whereas denser and larger retain their strength better (Pusch
and Yong 2006). The breakdown of weak parts transforms them to get a laminated
structure, which is believed to be the reason for the Newtonian rheology of smectite
clay that has undergone large strain (Figure 5.3).
At very late stages of strain and for low bulk stresses, the number of micro-
structural units is more or less constant and the bulk strain rate, caused by rela-
tive displacement of interacting units, termed dislocation glide, is more or less
constant.
Rate process theory, introduced by the American Nobel prize winner Eyring
in the early 1930s (Eyring 1936), provides a basic expression for the creep rate dγ
caused by shearing:
dγ
= Ae( −U / RT ) f (τ) (5.23)
dt
where:
A is the temperature-dependent constant
U is the creep activation energy
R is the universal gas constant
T is the temperature
f(τ) is the stress function
Equation 5.23 implies that the strain rate is constant for constant stress, hence
representing what is termed secondary creep in rheology; however, the introduction
of the concept of an energy barrier spectrum in more modern times improved the
theory and made it possible to employ statistical mechanics and develop models for
transient creep including primary and secondary creep. At present, one can assume
the following major strain-generating mechanisms in a homogeneous soil of visco-
elastic type (Feltham 1979):
• Application of a sufficiently high macroscopic shear stress initiates local

plastization where the interparticle bonds are overstressed.
• Redistribution of microstresses by slip implying higher local stresses where
they were previously low.
• The induced strain makes stronger units interact and successively strength-
ens the structure.
The net effect of the mechanisms depends on the magnitude of the macroscopic
shear stress, the total strain, and the boundary conditions. A sufficiently high stress
produces accumulated structural defects, as depicted in Figure 5.3, by which the
damaging processes outweigh the healing ones and lead to large-scale failure of buf-
fer clay loaded by heavy canisters or to substantial settlement, and of clay slopes if
there is no external restrain. At lower stresses, the healing processes dominate and
creep is retarded.
The detailed microstructural process generating shear strain involves early trig-
gering of low barriers associated with activation of new slip units at the lower end of
the barrier spectrum, which represents the generating barrier end, whereas the upper
end is sorbing. The energy spectrum hence changes over the course of the creep, as
illustrated by Figure 5.4, which represents the case of an initial box-shaped spectrum
with uniform distribution of barriers in the assumed interval 0.1–0.5 eV. The graph
shows the calculated successive change in the number of barriers of different activa-
tion energies as a function of time. At the end of the creep period (stage 5), the number
of very low barriers has dropped from the initially assumed number 100 to 1, whereas
those representing 0.5 eV has increased from 100 to about 400 (Pusch and Feltham
1980). If the shear stress is kept constant, the energy spectrum moves from lower to
higher levels in the course of the mobilization of stronger barriers and the system
successively becomes dominated by strong interparticle bonds. The clay hence stays
macroscopically stable but becomes brittle and sensitive to even a small increase in
macroscopic stress. It can trigger catastrophic instant failure (Pusch et al. 2014a).
The whole particle system remains stable provided that the system of interacting
microstructural units is coherent. Failure can, however, occur ultimately if the bulk
shear stress exceeds a certain critical value, which is believed to be on the order
of 2/3 of the conventionally determined bulk strength (Pusch and Yong 2006). For
lower stresses, the bulk strain corresponds to the accumulated and integrated very
small slips leading to primary creep characterized by attenuating strain rate. For
critically high stresses, the rate of mobilization of new particle bonds at the low-
energy tail of the spectrum is too low to maintain microstructural continuity and
secondary creep, ultimately leading to bulk failure, is reached.
98 Bentonite Clay
103
1
102
2
n(u,t)
4
101
5 0.l–0.6 eV
100
1 2 3 4 5 6 7 8 9 10
ui
FIGURE 5.4 Successive change in distribution of energy barriers in the course of transient
creep. 1, 2, 3, and so on denote time stages and n(u,t) the fraction of barriers of energy level ui.
The initial number of barriers is taken as 100. (Data from Pusch, R., Feltham, P., A stochastic
model of the creep of soils, Geotechnique, 30, 497–506, 1980.)
Applying Eyring’s principle that the frequency of slip is given by Arrhenius’s rate
equation and introducing Feltham’s transition probability parameter to describe the
time-dependent energy shifts, and also the condition that each transition of a slip
unit between consecutive barriers gives the same contribution to the bulk strain, one
gets the expression for the bulk shear strain rate relationship for stresses that are low
enough to leave the microstructure largely intact:
dγ ⎛ 1− t ⎞
= B⎜ ⎟ (5.24)
dt ⎝ to ⎠
The appropriate stress-related function B and the value of to depend on the shear
stress, temperature, and structural details of the slip process. This gives us the
expression in Equation 5.25 for creep with t < to as boundary condition:
⎛ a ⎞
γ = αt − β t 2 , ⎜ t < ⎟ (5.25)
⎝ 2β ⎠
This means that the creep starts off linearly with time and then dies out, which
corresponds to primary creep.
For higher bulk loads, irreversible changes take place in the form of local micro-
structural breakdown and reorganization. This mobilizes the inflow of new low-
energy barriers, which increases the number of slip that is successively being stopped
by meeting higher energy barriers. This type of creep can go on forever without
approaching bulk failure.
According to Feltham (1979), the strain rate is appertaining to logarithmic creep
for thermodynamically appropriately defined limits of the u-spectrum, taking the
form of Equation 5.26:
dγ BT τ (5.26)
=
dt ( t + to )
where:
B is a function of the shear stress τ
to a constant of integration
It leads to a creep relation closely representing the commonly observed logarithmic

type according to which one expresses the creep strain as being proportional to
log(t + to). The significance of to is understood by considering that, in the course
of applying a shear stress, it rises successively from zero to its nominal, final value.
This model implies that the creep strain is as illustrated in Figure 5.5 for moder-
ate deviator stresses that allow for microstructural recovery. Both curves show suc-
cessive adaption to logarithmic strain, the one that is upward convex being typical
for cemented clays with rapidly dropping strain rate associated with quick bond
to (−)
Strain (log scale)
to (+)
Time after onset of creep (log scale)
FIGURE 5.5 Generalization of creep curves of log time type. (Data from Pusch, R.,
Feltham, P., A stochastic model of the creep of soils, Geotechnique, 30, 497–506, 1980.)
100 Bentonite Clay
1.00E−07
1.00E+00
1.00E−01
1.00E−02
Strain rate (strains)

1.00E−08 1.00E−03
1.00E−04
Strain
1.00E−05
1.00E−06
1.00E−09 1.00E−07
1.00E−08
1.00E−09
1.00E−10
1.00E−10 1 10 100
1 10 100 Time after onset of creep (days)
Time after onset of
creep (days)
FIGURE 5.6 Creep of smectite-rich clay (MX-80) with a density at saturation with distilled
water of 1500 kg/m3. Left: steeply declining curve representing primary creep for the shear
stress 6 kPa. Right: Creep rate recorded for the shear stress 23 kPa representing second-
ary creep. (Data from Pusch, R., Geological Storage of Radioactive Waste, Springer-Verlag,
Berlin, Germany, 2008.)
breakage, whereas the downward concave curve represents viscous behavior with
initially constant rate of strain and possible initiation of failure.
Further increase in deviator stress leads to creep with the strain rate tending to be
proportional to time. Depending on the physical boundary conditions, this can lead
to fully developed failure. Figure 5.6 shows recorded creep strain of smectite-rich
clay (MX-80) with a density at saturation with distilled water of 1500 kg/m3 under a
shear stress of 6 kPa and 23 kPa, respectively. For low shear stress, the creep attenu-
ated rapidly, representing the relationship in Equation 5.25, whereas for the high
stress, the strain rate became constant, indicating secondary creep by comprehensive
microstructural distortion; the clay behaved as a fluid.
Figure 5.7 shows the behavior of the same type of clay with a density at saturation
with distilled water of 1940 kg/m3 under the shear stresses 11, 23, and 39 kPa. For
this higher density, all the curves became downward concave, indicating attenuating
creep after about 4 to 110 hours.
For t = 10,000 s, one finds that the strain rate (strain per second) is about 3E−8
for the lowest stress, about 7E−8 for the intermediate stress, and E−7 for the highest
stress, which suggests that the strain rate is roughly proportional to the shear stress,
indicating, in principle, Newtonian behavior.
The nearly constant strain rate under the high stress shown in the right diagram
in Figure 5.6 was recorded after increasing the average shear stress to a level that
was expected to yield instant failure. However, strain-hardening is believed to have
occurred by mobilization of numerous reserve stacks of lamellae successively dis-
persed from the remainders of the granules, causing massive inflow of low-energy
barriers. It can be visualized as formation of an anisotropic matrix of aligned and
partly disintegrated stacks of lamellae oriented in the direction of the maximum
shear stress as indicated in Figure 5.3. The specific surface area of the clay is thereby
successively raised and the shear resistance mainly provided by a steadily increasing
number of formed and broken hydrogen bonds like in viscous gels. This suggests that
1.00E−05
Shear strain (per second) 1.00E−06
1.00E−07
1.00E−08
1.00E−09
100 1,000 10,000 100,000 1,000,000
Time after loading (seconds)
FIGURE 5.7 Creep curves for different shear stresses. Density at saturation with distilled
water 1940 kg/m3. The shear stresses were, from top: 39, 23, and 11 kPa. (Data from Pusch,
R., Geological Storage of Radioactive Waste, Springer-Verlag, Berlin, Germany, 2008.)
smectite clays may not undergo failure like illitic clays, that is, by slip along discrete
failure planes, but deform to yield zone failure, using geotechnical terminology.
5.4 FUNDAMENTALS OF THERMAL CONDITIONS
AND PERFORMANCE
One understands from the discussion of thermodynamic issues in the Chapters 3
and 4 that temperature is an important factor for many functions, which has led to
the concept THMC for the thermal (T), hydraulical (H), mechanical (M) properties,
and behavior of clay seals (C), (Svemar 2005), and we will see some of this in the
book. For understanding all this, some basic definitions of T-related definitions and
parameters need to be given.
5.4.1 Definitions and Parameters

The basic thermal properties of soils used for predicting the temperature evolution in
buffer clay surrounding heat-producing objects such as high-level radioactive waste
(HLW) canisters and clay liners exposed to the temperature of contacting air, are as
follows*:
• Volumetric heat capacity, J m−3 K−1

• Thermal conductivity, W m−1 K−1
• Thermal diffusivity, m2 s−1
* Described by Kahr and Mueller-Von Moos 1982; Knutsson 1983; and Farouki 1986.
102 Bentonite Clay
The contents of air and water are determinants of the thermal conductivity of clay
seals. For buffer clay, it is essential that the conductivity is high and that it has a high
degree of water saturation. For top liners of landfills, it is the opposite: they serve to
protect the underlying waste mass from heat or rather from temperature variation in
order to minimize the risk of condensation.
5.4.2 Influence of Temperature
It was found early that a temperature rise makes the swelling pressure of Na-smectite
drop except for low densities (Börgesson et al. 1991, 1993, 1994), whereas there is
an opposite trend for Ca smectite, which has been explained by different impact of
the osmotic pressure with temperature and by the smaller number of interlamel-
lar hydrates in Ca smectite. This difference has cast doubt of the validity of the
Na-smectite measurements. However, more recent investigations, in which great
care was taken to allow heat-generated porewater overpressure to dissipate, have
demonstrated that an increase in temperature causes a drop in swelling pressure of
montmorillonite-rich smectite clay in Na form. This is demonstrated by the diagrams
in Figures 5.8 and 5.9, showing recorded swelling pressures of clays prepared from
strongly compacted granules of montmorillonite-rich clay (MX-80). The study com-
prised temperature cycling in the interval of 22°C–75°C.
This study showed that the temperature cycling had only a very small impact
on the swelling pressure. The slight drop from initially 4700 to 4500 kPa can be
ascribed to microstructural rearrangement. Tests at room temperature are known
to give largely constant pressure after a few weeks but for clays prepared from very
strongly compacted granules (pellets), it may take several weeks or months before
Heating from 22°C to 75°C

5000
4500
4000
Swelling pressure (kPa)
3500
3000
2500
2000
1500
1000
500
0
0 5 10 15 20
Time (days)
FIGURE 5.8 Recorded swelling pressure of MX-80 clay saturated with distilled water during
heating. Upper curve: 1930 kg/m3 density at saturation. Lower curve: 1660 kg/m3 density. (Data
from Börgesson, L. et al., Final Report of the Rock Sealing Project—Sealing of the Nearfield
Rock around Deposition Holes by Use of Bentonite Grouts, Stripa Project Technical Report
TR 91-34, Swedish Nuclear Fuel and Waste Management [SKB], Stockholm, Sweden, 1991.)
Quick cooling from 75°C to 22°C

5000
4500
Swelling pressure (kPa) 4000
3500
3000
2500
2000
1500
1000
500
0
0 1 2 3 4
Time (days)
FIGURE 5.9 Recorded swelling pressure of MX-80 clay saturated with distilled water at
cooling. Upper curve: 1930 kg/m3 density at saturation. Lower curve: 1660 kg/m3 density. (Data
from Börgesson, L. et al., Final Report of the Rock Sealing Project—Sealing of the Nearfield
Rock around Deposition Holes by Use of Bentonite Grouts, Stripa Project Technical Report
TR 91-34, Swedish Nuclear Fuel and Waste Management [SKB], Stockholm, Sweden, 1991.)
microstructural equilibrium is reached (Pusch et al. 2003). If hydration takes place

at some tens of degrees Celsius above room temperature, the thermally induced
stresses tend to disintegrate the granules, which exposes their interior to water and
breaks cementation bonds, hence speeding up the fluid saturation.
From a practical point of view, the heat-induced expansion of the free, extralamel-
lar porewater in confined dense smectite clay is much more important for the behav-
ior of the buffer clay in a KBS-3 repository although this water makes up a rather
small part of the total porewater content. Being under highest pressure near the hot
canisters, it will be squeezed through the microchannels in the buffer clay toward
the colder rock and migrate upward along the buffer–rock contact to the backfilled
tunnels. The microchannels in the buffer clay, neoformed or widened, can become
permanent by precipitation of salt and dissolved buffer minerals, as we will see in
the subsequent chapter.
5.5 EVOLUTION OF CLAY SEALS

5.5.1 Hydration
We touched on the matter of hydration of smectite clays in Chapters 3 and 4 and come
back to it here for seeing how the microstructural model is related to the perfor-
mance of these clays. The fundamental issue is that hydration of the interlamellar
space depends on the charge and coordination of adsorbed cations and water mol-
ecules. It has fascinated a whole generation of clay mineralogists and chemists and
led to a number of theoretical models of how water molecules are organized and
perform in smectites, particularly in montmorillonite (Hofmann et al. 1933; Forslind
and Jacobsson 1972; Kehres 1983; Pusch 1987a; Nagy et al. 2010).
104 Bentonite Clay
There seems to be general agreement that there can be up to three interlamellar

hydrate layers in montmorillonite, and that there is a monolayer film of water con-
taining cations sorbed on the basal surfaces of these minerals. This film, which
in fact covers practically all silicate minerals except talc* (Sposito 1984), plays a
significant role in diffusive ion transport and osmotic swelling of smectites. The
interlamellar hydrates largely determine the swelling potential, whereas the surface
properties of the stacks of lamellae determine their plasticity and the overall bulk
rheological behavior.
The coordination of interlamellar cations, crystal lattices, and water molecules
depends strongly on the size and charge of the cations and of the charge distribu-
tion in the lattice. As said, three hydrate layers can only be formed in some of the
smectite species, and normally only when Na, Li, or Mg are in interlamellar posi-
tions (Table 5.1). One finds from this table that establishment of three interlamellar
hydrates in montmorillonite is possible only for Na and Mg as sorbed cations.
Submerged in water, hydration of montmorillonite crystallites proceeds until the
maximum number of interlamellar hydrates is formed, provided that there is no geo-
metrical restraint. In humid air, the number of hydrates obtained depends on the
relative humidity. One interlamellar hydrate is formed in Na montmorillonite when
relative humidity (RH) is approximately in the interval 20%–40%, whereas two
hydrates require 40% < RH < 70%. For RH exceeding about 70%, three interlamellar
hydrates can be formed (Mooney et al. 1987).
TABLE 5.1
Number and Thickness of Interlamellar Hydrates in Å
Smectite Clay Ma First Hydrate Second Hydrate Third Hydrate
Montmorillonite Mg 3.00 3.03 3.05
Ca 3.89 2.75 –
Na 3.03 3.23 3.48
K 2.42 3.73 –
Beidellite Mg 2.69 2.69 –
Ca 2.30 2.30 –
Na 2.15 2.15 –
K 2.54 – –
Nontronite Mg 2.92 3.00 –
Ca 3.05 3.37 –
Na 2.70 2.79 –
K 2.60 – –
Source: Kehres, A., Isotherms de deshydratation des argiles. Energies d’hydratation—Diagrammes de

pores surfaces internes et externs, PhD Thesis, Université Paul Sabatier, Tolouse, France, 1983.
a Interlamellar adsorbed cation.
* This hydrophobic mineral is successfully used as fluidizer in concrete.

5.5.2 Forslind –Jacobsson Model

In the 1960s, the Swedish investigators Forslind and Jacobsson introduced a struc-
tural model for interlamellar water that is based on Edelmann–Favejee’s crystal lat-
tice constitution of montmorillonite (Forslind and Jacobsson 1972). It partly explains
the discrepancy in concerning the cation exchange capacity mentioned, that is, that the
Edelmann–Favejee’s model gives a cation exchange capacity value six times higher
than experiments by assuming the apices carrying acidic hydroxyls to exchange their
protons. According to these investigators, there is a coupling of the first-associated
water lattice and the basal planes of the crystallites assuming that the water is ice-
like. Figure 5.10 shows the protruding hydroxyls of the (001) plane with three unit cells
exhibited, and the basal molecules of a superimposed ice lattice. One finds that there are
only four hydrogen bonds formed between the three unit cells and the water lattice, and
that the charge should therefore attain a mean value of two-thirds of a charge unit per
crystal cell. This is also the average cation exchange capacity value recorded in practice.
An obvious consequence of this is that the water molecules are not as closely
packed as in bulk water, and that the water density is hence lower than 1000 kg/m3,
excepting the cations present. A special feature of the Forslind–Jacobsson crystal-
water model is that expulsion of protons from the lattice apices causes a negative
15.62Å
5.16Å
8.94Å
4.51Å
9.02Å
Hydroxyl group of Hydroxyl group of

the silica layer the silica layer
Water molecule
belonging to a basal
ice lattice layer
FIGURE 5.10 Left: protruding hydroxyls of the (001) plane with three unit cells shown.
Right: the basal molecules of an ice lattice superimposed on these hydroxyls. (Data from
Forslind, E., Jacobsson, A., Clay-water systems, In Water, A Comprehensive Treatise, Franks,
G. Ed., Plenum Press, New York, 1972.)
106 Bentonite Clay
particle charge. This can explain the negative charge of montmorillonite particles
without assuming isomorphous substitutions and vacancies in the crystal lattices.
The small Na and Mg cations in interlamellar positions do not disturb the water
lattice, while disturbance is caused by large cations and by those having strong affin-
ity to water like Ca (Sposito and Prost 1982). By exchange of Na for Ca, the solvation
shell of the latter ion binds water so effectively that the water lattice is disturbed and
is no longer strongly associated with the lattice oxygens as implied by the Forslind–
Jacobsson’s model. The water molecules in the interlamellar space are organized
such that the first hydrate is thicker than the second, and that no more hydrates can
be formed, as showed by Table 5.1. The change in reorganization of water mole-
cules in the Forslind–Jacobsson’s model on replacing Na-montmorillonite by Ca in
the interlamellar space can involve conversion of the structural constitution from
Edelmann–Favejee’s form to that of Hofmann–Endell–Wilm’s model. This would
imply a change in coordination of tetrahedral Si and octahedral Al, which should be
detectable by using spin–spin NMR technique, as has in fact been done (Pusch 1993).
For the Na state, the hydrate layers according to the Forslind–Jacobsson model
have been proposed to be as indicated in Figure 5.11, implying that only certain
water lattice configurations are stable. The stable monolayer yields a d-spacing of
17.81Å and a water content w = 11.9%. For two layers d = 22.41Å and w = 23.8% and
for three layers d = 25.17Å and w = 35.7%. These values fit with published (002) and
(003) data, hence making the hydration model trustworthy. However, careful analysis
5.33 water molecules/unit cell
FIGURE 5.11 Model of hydrate configuration in the interlamellar space of montmorillonite

with the Forslind–Jacobsson crystal constitution. The stable monolayer yields a d-spacing of
17.81Å and a water content w = 11.9%, while it is d = 22.41Å for two layers and d = 25.17Å
for three layers. (Data from Forslind, E., Jacobsson, A., Clay-water systems, In Water, A
Comprehensive Treatise, Franks, G. Ed., Plenum Press, New York, 1972.)
100
Na form
80
60
W (%)
40
20
Ca form
0
1000 1200 1400 1600 1800 2000 2100
Bulk density at water saturation (kg/m3)
FIGURE 5.12 Fraction of interlamellar water in montmorillonite. (Data from Pusch, R.,
Yong, R.N., Microstructure of Smectite Clays and Engineering Properties, Taylor & Francis,
New York, 2006; Pusch, R. et al., High-Level Radioactive Waste Disposal, WIT Press,
Southampton, 2012b.)
of recent X-ray diffraction data shows that hydration of smectite-rich samples do not
give perfectly regular water-layer sequences, and the geometry and density varia-
tions of hydrated smectites are still not accurately known.
The organization of clay particles; coarser grains and organic matter, which
together make up the solid phase; and the degree of water saturation determine the
physical properties of clays. The interlamellar, often called internal water, domi-
nates over the free or external water at high densities, especially in Na montmoril-
lonite (Figure 5.12).
5.6 CLAY MICROSTRUCTURE AND ITS ROLE FOR THE

STRESS–STRAIN BEHAVIOR OF SMECTITE CLAYS
5.6.1 Scale Dependence
Starting with the fundamental smectite lattice unit shown in Figure 3.1 and consid-
ering the typical grouping of the stacks of lamellae how they interact, these micro-
structural units can be generalized as in Figure 5.13. It shows the typical difference
between the size and interaction of the stacks forming aggregates in Na clay and Ca
clay, and indicates the difference in their swelling pressure at complete hydration.
They refer primarily to montmorillonite but represent in principle also other smec-
tites. For any bulk dry density, the aggregates of Ca montmorillonite are separated
by larger voids than in those of Na montmorillonite and make the clay more silt-like.
The wider voids and the higher content of external water make Ca montmorillonite
more permeable than Na montmorillonite.
It is obvious from Figure 3.2 and numerous microstructural studies that the inter-
lamellar space varies and that the voids between the aggregates of stacked lamellae
108 Bentonite Clay
a a
Na a Na Ca
Soft Dense Soft to

b dense
b
b
c c c
Low salt Low and

high salt
10,000 10,000 10,000
Ps (kPa)
Ps (kPa)
Ps (kPa)
1,000 Low salt 1,000 High 1,000
salt
100 100 100
High salt
00
20 0
00
00
20 0
00
00
20 0
00
0
0
14
18
14
18
14
18
ρ (kg/m3) ρ (kg/m3) ρ (kg/m3)
FIGURE 5.13 Schematic pictures of stacks of lamellae and influence of density and
porewater salinity on the particle spacing and swelling pressure (Ps). (a) lamella, (b) interla-
mellar space, (c) contact region of interacting electrical double layers, exposed hydroxyls and
polarized sorbed cations. (Data from Pusch, R., Yong, R.N., Microstructure of Smectite Clays
and Engineering Properties, Taylor & Francis, New York, 2006.)
have a varying size, meaning that the hydraulically interacting microchannels have
a varying cross-sectional size. An attempt to mathematically describe this variation
based on transmission electron micrographs of ultrathin sections of water-saturated*
montmorillonite-rich clay samples led to the size distribution of intersected channels
shown in Table 5.2 (Pusch and Yong 2006).
The basis for this specification was an extensive study of transmission electron
microscopy (TEM) and use of the following definitions:
• The microchannels have a circular cross section.

• The diameter of the widest channel is 50 μm in clay with a bulk density of
1570 kg/m3, 20 μm in clay with 1850 kg/m3 density, and 5 μm in clay with
2130 kg/m3 density at water saturation; as estimated from TEM.
The micrograph in Figure 5.14 is an example of the variation in void size and density
of montmorillonite-rich clay (MX-80 Wyoming bentonite) with the rather low dry
bulk density 1270 kg/m3 (1800 kg/m3 at water saturation). Digitalization for estimat-
ing the variation in density and quantification of the related variation in hydrau-
lic conductivity on different scales and of the average bulk conductivity was made
* Ultrathin sections cut from clay in which the water had been replaced by acrylate.
TABLE 5.2
Number of Differently Sized Channels per 250 × 250 μm2 Cross-Sectional
Area of a Representative Elementary Volume (REV) of Na-Montmorillonite-
Rich Clay (MX-80)
Clay Bulk Density Number of 20–50 μm Number of 5–20 μm Number of 1–5 μm
Type (kg/m3) Channels Channels Channels
A 2130 0 0 135
B 1850 0 10 385
C 1570 2 85 950
FIGURE 5.14 (See color insert.) Example of digitalized micrograph of 300Å ultramicro-
tomed section of Wyoming bentonite (MX-80) with a bulk density at saturation of 1800 kg/m3
(dry density 1270 kg/m3). Black represents densest parts of clay matrix and accessory miner-
als; red represents relatively dense parts of the same component; green represents soft, porous
clay gel; white represents open voids. Height of picture is 3 μm.
by use of such pictures taken by TEM of ultramicrotomed sections (300Å–500Å)

(Pusch 1967, 1970, 1999a). They have consistently shown that aggregates of Ca
montmorillonite are separated by large voids making them more permeable than Na
montmorillonite with the same bulk density.
5.6.2 Impact of Heating
The impact on the microstructure of heating implies changes in the structural sta-
bility of the interlamellar water especially when Na or Mg are present. Considering
a water-saturated clay element with constant volume, an increase in temperature
raises the porewater pressure and reduces the grain pressure (effective stress).
Looking at the smallest microstructural entities, the stacks of lamellae, the stability
110 Bentonite Clay
2 k
x
1 θ
r
i
FIGURE 5.15 Mass forces between two bodies with plane interface.
of interlamellar water films can be expressed in terms of attractive mass forces and
entropy effects applying thermodynamical views (Frenkel 1946). Thus, the poten-
tial energy of a body (Figure 5.15) with molecules 1 with respect to a molecule k
in a film of water 2, located at a distance x from the interface, can be expressed by
Equation 5.27. Ordinary cohesive forces c act between molecules i and k represented
by uik = −cik⋅r 6.
uik ( x ) =

∫∫ ni.uik ( r dr sin θdθ) (5.27)
2
where:
ni is the number of molecules i per unit volume
uik is the potential energy of interacting molecules i and k at separation by r
This expression can be used for deriving the mutual potential energy per unit area
of the semi-infinite i-body and the k-body with thickness h in Figure 5.15, as well as
for the case of two 1-bodies confining a 2 film. Frenkel discussed the case of a thin
water film confined between two thin mineral lamellae and concluded that, from
an energy point of view, the water film is not stable under the generated compres-
sive pressure, and that the system collapses forming thicker and bigger particles
(cf. Prikryl, personal comm). The reason why interlamellar water in the stacks of
montmorillonite still exists would be that this water is structured and has suffi-
cient strength to resist compression, and that hydrated interlamellar cations prevent
total collapse. The consequence of heating can be that the induced vibrations break
down the structure to form thinner water films in conjunction with emigration of
interlamellar cations, eventually leading to direct mineral–mineral contact (Pusch
1987a,b). This can explain why no external pressure is required to cause the heat-
induced contraction of stacks of montmorillonite that we see in the central and right
m icrographs in Figure 5.16.
The impact of high temperature on smectite clay is concluded to be as illustrated
in Figure 5.17, which is in agreement with the earlier described drop in swelling
pressure: (1) contraction of stacks of lamellae and (2) precipitation of silicious matter
causing cementation. The two processes combine to increase the hydraulic conduc-
tivity and the shear strength.
In this context, it is challenging and tempting to consider the Forslind–Jacobsson
model of montmorillonite crystal buildup, since it may provide an explanation
of the contraction of the stacks of lamellae at high temperature. Their hydration
model, based on the Edelman–Favejee montmorillonite structure, would not be
(a) (b) (c)
FIGURE 5.16 Comparison of montmorillonite-rich clay at room temperature (a), after

hydrothermal treatment at 150°C for one year (b), and at 200°C (c). Notice the collapse of
stacks of lamellae and formation of dense silicious particles in the heated clays that had a
dry density of 1590 kg/m3. Bar is 1 μm. (Data from Pusch, R., Yong, R.N., Microstructure of
Smectite Clays and Engineering Properties, Taylor & Francis, New York, 2006.)
Nonheated
Si
Heated
FIGURE 5.17 Microstructural changes by heating of smectite-rich buffer clay under con-
stant volume conditions. Upper: virgin state at room temperature. Lower: heating to more
than 60°C–100°C.
stable at higher temperature and structural breakdown of the interlamellar water

structure by heating to more than around 100°C can have the form of inversion of
the protruding apical SiO2 tetrahedrons, transforming the crystal structure to the
traditional Hofmann–Endell–Wilm form (cf. Chapter 3). The conversion would
bring tetrahedral Si in an unstable state, facilitating its exchange by Al, which
transforms the mineral to beidellite that is particularly sensitive to uptake and
fixation of K ions in interlamellar positions, yielding nonexpandable illite. We will
come back to this in Chapter 6, where we will put the chemical stability of smectite
clays in focus.
In summarizing all this and referring to Figure 5.17, it seems that the contraction
effect is small or moderate for temperatures lower than about 100°C (Forslind and
Jacobsson 1972), but that raising the temperature beyond this level can cause break-
down of the interlamellar water structure and promote this effect (Pusch 1987b).
112 Bentonite Clay
5.6.3 Microstructural Modeling of Smectite-Rich Clay

Bentonite being formed by sedimenting particles (Figure 5.18) has a haphazard
structure but consolidation under the weight of overlying sediment gives the soft
clay higher density and a homogeneous constitution of well-fitting aggregates of the
smallest microstructural units, the stacks of lamellae. On a larger scale, the system is
more heterogeneous because of occasional interruption of the sedimentation of fine
particles by coarser material from volcanic eruption or brought on site by bottom
currents. The microstructure of consolidated bentonite clay, although appearing to
be homogeneous, can have considerable variations with respect to orientation and
spacing of aggregated stacks because of the mutual displacements of the aggregates
when true bentonites undergo consolidation in nature.
Since we are concerned in particular with the role of sealing clay in the form of
artificially prepared tight smectite liners, plugs, and buffer clay in this book, we need
to consider the microstructure of such seals. Blocks can be prepared by compress-
ing dense smectite granules from processing plants that manufacture air-dry clay
material by crushing and grinding of raw materials (Figure 5.19) or by compacting
granules or mixtures of granules and coarser aggregate on the construction site to
FIGURE 5.18 Micrograph of 20 μm fluffy aggregate of smectite particles in low-electrolyte

river water (Elbe) photographed by optical microscopy. (Courtesy of Kasbohm.)
FIGURE 5.19 Sector-shaped block of 85% smectite-rich clay mixed with 10% quartz and
5% graphite for enhancing the heat conductivity. Compaction of the air-dry mixture under
100 MPa pressure gave a dry density of 1850 kg/m3 (2170 kg/m3 at water saturation under
confined conditions). The actual bluish color was caused by adding 5% graphite for enhancing
the thermal conductivity.
make liners (layers). Samples from such blocks can be examined by transmission
or scanning electron microscopy for identifying and quantifying the microstructural
constitution (Pusch and Yong 2006). Typical micrographs of artificially prepared clay
blocks are those shown in Figure 5.14 for matured, moderately compacted granules;
and to the left in Figure 5.16 for densely compacted granules after maturation at
room temperature.
Various microstructural models representing compacted smectite-rich granules in
dry and water-saturated form have been proposed, such as the general microstructural
model illustrated in Figure 5.20, which has been used for estimation of the void size
distribution as a basis of modeling percolation of water under a hydraulic gradient
and calculating the bulk hydraulic conductivity (Pusch et al. 1987b, 2001, 1990). As
we saw earlier in this chapter, such conceptual microstructural models can be con-
verted into a generalized form by finite element method (FEM) calculation of the den-
sity distribution in hydrated mixed clay–sand liners using ABAQUS (Computational
Mechanics International Ltd, Southampton, UK), or using boundary element codes
such as the CMI (Computational Mechanics International Ltd, Southampton, UK),
I 1
8
II
II
100
Weight % passing
>200 MPa 3.79520E+000

80 3.45556E+000
3.11591E+000
60 2.77526E+000
2.43662E+000
40 100–200 MPa 2.89697E+000
1.75732E+000
20 1.41768E+000
1.07883E+000
0 <100 MPa 7.38383E−001
10 100 1000 3.98736E−001

5.90890E−002
Grain size (μm) Van Mises effective stress:

Max = 3.8333
Min = 2.09654E−002
FIGURE 5.20 General microstructural model of compacted smectite clay with very dense
granules (I, lower left figure) and voids to be filled with soft exfoliated clay in large voids (II),
and with dense clay in small voids during hydration. (Data from Pusch 1970, 1987a; Pusch,
R. et al., GMM—A general microstructural model for qualitative and quantitative studies of
smectite clays, SKB Technical Report TR-90-43, SKB Swedish Nuclear Fuel and Waste Co
[SKB], Stockholm, Sweden, 1990.) The right picture shows BEM-calculated compression of
a model consisting of 0.35 mm and 0.1 mm MX-80 grains in regular packing after 30% uni-
axial compression. (Data from Pusch, R., Yong, R.N., Microstructure of Smectite Clays and
Engineering Properties, Taylor & Francis, New York, 2006.) The lower right picture shows the
Mises stresses in the big grain in this stage. The stresses are so high that the grain is fragmented.
114 Bentonite Clay
worked out by International Computational Methods, Ashurst, United Kingdom, for

modeling in 3D of the hydrated model clay (Figure 5.20). Here, we will confine our-
selves to consider the geometrical features of smectite microstructure.
5.7 EFFECT OF COMBINED WETTING AND HEATING

OF CLAY SEALS—THE BUFFER CASE
5.7.1 Practical Cases
In many situations, transfer of moisture and heat occurs simultaneously and insepa-
rably. Theoretical studies of the heat-associated mechanisms in soils are based on the
assumption that the soil is homogeneous and that all processes of heat transfer take
place uniformly throughout the porous medium, whereas in reality, vapor transfer or
air convection takes place only in the air-filled pore space, and flow in liquid form
only in saturated pore space forming microchannels. In both cases, slow diffusional
water migration, driven by gradients in water content, can take place (Knudsen flow;
cf. Svemar 2005). Evaporation and condensation are associated with phenomena at
the water–air interfaces (Knutsson 1983; Farouki 1986).
One can identify several cases in which clay seals for isolation of hazardous waste
become heated, typical examples being top liners of landfills hosting waste that gen-
erates heat or is exposed to intense sun radiation, and clay buffer isolating canisters
with highly radioactive waste. The temperature of clay liners in on-ground repositories
seldom exceeds 50°C–75°C and is not allowed to be lower than a few degrees Celsius
plus. It will not cause damage to clay liners but causes seasonal, weekly, or even daily
changes in the degree of water saturation. In contrast, HLW cases need to be consid-
ered in more detail because the temperature and temperature gradients will be much
higher. First of all, one must define, for these cases, what the hydraulic interaction is
between the clay seal and the surrounding rock with special respect to the piezometric
conditions in the rock. Experience tells that the groundwater pressure in the vicinity
of mined-out areas drops temporarily and rather much in the construction and waste
installation phases. This is the case with normally fractured rock making placement of
clay seals simple because the hydraulic gradients across the seals are moderate. In the
course of pregrouting the rock where clay seals will be constructed, like bulwarks with
O-ring type clay seals, the piezometric pressure rises and increases the hydraulic gra-
dients, which can cause piping in them. In very tight rock, as in the presently proposed
very dry Forsmark rock for HLW disposal in Sweden, the problem is the opposite.
Here, no grouting is required, and it can take hundreds of years before the rock has
given off sufficiently much water to make clay seals completely saturated. In this case,
very slow wetting of the buffer in KBS-3-type canister holes can cause comprehensive
precipitation of corroding salts, such as sulfides, as will see in the subsequent chapter.
5.7.2 Hydration of Smectite-Rich Buffer Clay

under Temperature Gradients
Looking first at the KBS-3 concept, launched by the Swedish Nuclear Fuel and Waste
Management Company (SKB), which is responsible for safe isolation of Swedish
HLW from the groundwater for at least 100,000 years, designers need to take the
excavation-damaged near-field rock into consideration for defining the rock condi-
tions with respect to the hydraulic and mechanical performances (Figure 5.21). The
excavation disturbance of the tunnels extends by a few decimeters from the walls
and roof and by 1–1.5 m from the floor of blasted tunnels.* The upper part of the
about 8 m deep holes with clay-embedded hot canisters is hence located in fractured
rock with an average hydraulic conductivity that is up to 1000 times higher than that
of the virgin rock (Börgesson et al. 1995). The rock around the rest of the hole has
a thin boring-disturbed zone with a hydraulic conductivity that is about 10 times
higher than the conductivity of virgin rock (Figure 5.21). The stress-induced damage
observed on site and calculated by FEM technique causes fissuring and makes the
bottom of the holes particularly pervious, which agrees with the experience from
field tests in the Stripa mine (cf. Section 5.7.5). The pervious zones surrounding the
bored holes is in direct contact with the EDZ of the tunnels and are hence hydrauli-
cally interacting with water-bearing fracture zones intersected by the tunnels.
In rock with high stresses, that is, with a major horizontal stress of more than
50 MPa and a minor vertical stress of 10 MPa, critically high hoop stresses can arise
5.0 5.0
4.5 4.5
4.0 4.0
3.5 3.5
3.0 3.0
2.5 2.5
2.0 2.0
1.5 1.5
Backfill
1.0 1.0
0.5 0.5
Blast-damaged floor
0.0 0.0
0 1 2 3 4 0 1 2 3 4
Failed (fractured) rock by 5.0
thermal impact acc. to SKB
4.5
4.0
Canister
3.5
Buffer 3.0
2.5
2.0
1.5
1.0
0.5
0.0
0 1 2 3 4
FIGURE 5.21 Pervious zone around the deposition holes caused by disintegration (left), and
overstressing of the rock (right). The three cases to the right represent the Mohr–Coulomb
failure criterion with c = 0.5 MPa and φ = 30° (upper left); c = 1 MPa and φ = 30° (upper
right), and c = 5 MPa and φ = 40°. (Data from Pusch, R., Rock Mechanics on a Geological
Base, Developments in Geotechnical Engineering 77, Elsevier, New York, 1995.)
* Blasting gives lower hoop stresses than TBM techniques and reduces the risk of spalling, which is
expected to be frequent in the area where the Swedish HLW repository will be built.
116 Bentonite Clay
and cause comprehensive fracturing when the temperature pulse from the radioactive
decay of the HLW hits the rock (Pusch and Weston 2008). This will generate or
activate fractures that expand or become closed depending on the stress state of the
rock, causing an increase or a decrease in hydraulic conductivity. In summary, one
concludes that the conditions for early hydration of the buffer are controlled by the
ability of the surrounding rock mass to provide water. It depends on the rock stress
and structure, on the piezometric conditions, and on time.
The fractures that intersect the deposition holes provide the buffer clay with water
in the hydration phase but can also bring possibly released radionuclides up to the
richly fractured and pervious tunnel floor, which is in turn hydraulically connected
to steep fracture zones in the rock mass. The arrows from the canister to this zone in
Figure 5.22 represent transport paths for released radionuclides.
The so-called very deep hole (VDH) concept, which was proposed many
years ago and is recently gaining new interest (Brady et al. 2009; Pusch et al.
2012a), implies, for the Swedish version, waste disposal in steep holes between
2000 and 4000 m depth and sealing with dense clay of the upper 2000 m.
Moderate to low temperatures (60°C–70°C) prevail down to 2000 m, whereas the
temperature can be up to 150°C in the lower, waste-bearing part. Wetting of the
Thick lines
indicate intersection of
ellipsoids and
deposition hole
FIGURE 5.22 KBS-3V deposition hole with disturbed rock zone connected to fractures that
reach up to the fractured floor of the blasted tunnel. (Data from Pusch, R., Geological Storage
of Radioactive Waste, Springer-Verlag, Berlin, Germany, 2008.)
Upper seal of concrete

under metal plug
0.5 km Sets of blocks of

highly compacted
smectite-rich clay
1.5 km
Sets of HLW canisters
separated by blocks of
highly compacted
smectite-rich clay
2.0 km
Bottom at 4 km depth
0.8 m
FIGURE 5.23 Schematic view of a VDH with constant diameter below the uppermost seal
of concrete extending down to 0.5 km depth, below which the hole is tightly sealed with clay
to 2 km depth. The deployment zone with 0.5 m HLW canisters separated by clay blocks is
shown here in a hole with 0.8 m diameter. (Data from Pusch, R. et al., J. Earth Sci. Geotech.
Eng., 2, 1–24, 2012a.)
buffer will take place under high water pressure (20–40 MPa) and is completed
in a few years.
The stability conditions for VDH depend on the rock stresses, which are higher
than for the KBS-3 concept, especially in the deployment zone (Figure 5.23). One
can estimate the primary maximum and minimum horizontal rock pressures of the
water-filled holes with a diameter of 0.8 m to be 50–60 MPa, and 35–45 MPa at
2 km depth, giving a hoop stress at this depth of up to 100 MPa. At 4 km depth,
the hoop stress has been estimated to be up to 190 MPa (Pusch et al. 2014a).
In the VDH case, relatively dense clay mud is proposed to be used in the con-
struction phase and thereafter, for stabilizing the holes. The difference between
the KBS-3 and VDH concepts with respect to the hydraulic conditions is that the
conductivity of the rock in the deployment zone is much lower for VDH than for
KBS-3V and that the density of the groundwater is sufficiently high to prevent
heat-induced convection of possibly contaminated water to reach higher up than
to the sealed part of the holes. This property is a strong argument for serious
consideration of the candidature of this concept, which we will examine in more
detail later in this chapter.
118 Bentonite Clay
5.7.3 Physical Processes Taking Place in Buffer Clay

For the VDH and the KBS-3 concepts, the latter being illustrated in Figure 5.24,
the evolution of the buffer clay involves a number of processes:
• Redistribution of initially contained porewater in the buffer generated by

the thermal gradient
• Homogenization and subsequent consolidation of the buffer blocks
• Uptake of water of the initially unsaturated buffer from the rock
• Deformation of the buffer caused by the canister load
• Transport of chemical elements into and out of the buffer due to the tem-
perature gradient
• Dissolution and alteration of minerals and precipitation of chemical
compounds
5.7.4 Evolution of the Buffer—Temperature

The temperature distribution in the early phase of hydration of KBS-3 buffer
(10–100 years) is illustrated in Figure 5.25, from which one finds that, for a typi-
cal case, the temperature at the canister surface is about 100°C, whereas it is about
70°C at the buffer–rock contact. The average radial temperature gradient in the buf-
fer is initially several degrees Celsius per centimeter but drops to about 1°C/cm after
10 years. It promotes mineral dissolution, and transport and precipitation of chemi-
cal compounds, and it also causes desiccation and fracturing of the blocks near the
Chamber with backfill material
Corner condition
Fracture/fissure pathway
and portal for water
Canister–buffer
interface
Canister
Buffer-rock interface
Buffer
Host rock
Some factors and processes contributing to long-term system

performance are as follows:
Heat transfer Geochemical processes
Liquid moisture transfer Reaction kinetics (chemical)
Vapor transfer Reaction kinetics (biologically mediated)
Constitutive behavior Illitization and transformations
FIGURE 5.24 Components of the near-field of a deposition hole with hot canister sur-
rounded by buffer clay. (Data from Pusch, R., Yong, R.N., Microstructure of Smectite Clays
and Engineering Properties, Taylor & Francis, New York, 2006.)
Temperature (K)
376.37
366.44
356.51
346.58
336.65
326.72
316.79
306.86
296.93
287
FIGURE 5.25 (See color insert.) Predicted temperature distribution in the near-field of
KBS-3 canisters shortly after the disposal. (Data from Svemar, C., Cluster Repository Project
[CROP], Final Report of European Commission Contract FIR1-CT-2000-20023, Brussels,
Belgium, 2005.)
canisters, that is, in the inner third of the buffer. The heat impact on the surrounding
rock causes a considerable increase in the hoop stresses, which can cause spalling
and fissuring. The swelling pressure exerted by the buffer is not sufficient to prevent
such degradation of the rock. In rock with very low hydraulic conductivity, it can take
a century or more before the buffer is completely saturated (Börgesson et al. 1991)
and very significant changes of the buffer mineralogy and physical properties are
expected in the slow maturation process (Yong et al. 2010).
5.7.5 Evolution of the Buffer—Expansion and Consolidation

under Hot Conditions
An example of combined swelling and consolidation of clay components in holes

containing HLW is shown in Figure 5.26, which shows an early VDH concept with
canisters and dense clay blocks confined in perforated containers forced down into
dense deployment clay mud in the deepest part of the holes. The clay block in each
set sucks water from the deployment mud, expands, and consolidates the soft mud.
The temperature can be up to 100°C in the installation phase, which reduces the
viscosity of the deployment mud.
120 Bentonite Clay
C C
W W
DM DM
B B
600 mm 600 mm
800 mm 800 mm
FIGURE 5.26 Early description of canister/clay set in the deployment part of VDH. Left:
initial state of set of waste canister (W) and clay seal of highly compacted clay (B) in perfo-
rated container. Right: matured state. DM is smectite deployment mud that can be the same
as the bore mud.
Large-scale field experiments have shown the importance of the ability of the
surrounding rock to provide water for saturating the buffer. Such a test was con-
ducted as part of the international Stripa Project in Sweden (1978–1992) (Svemar
2005). It was termed buffer mass test (BMT) and was run in granitic rock at 360 m
depth some 200 km northwest of Stockholm. The rock structure was characterized
by a fracture-poor rock matrix intersected by steep, minor fracture zones with a spac-
ing of 30–70 m, one of them intersecting the drift (Hole 2) in which the experiments
were made. Six holes with 0.76 m diameter and about 3 m depth were core-bored
to about 3 m depth. Blocks of highly compacted smectite-rich clay (MX-80) with a
dry density of about 1750 kg/m3 were placed around electrical heaters with 0.38 m
diameter for simulating canisters. The heaters were powered with 600 to 1200 W.
Figure 5.27 is a perspective view of the Stripe BMT site.
Figure 5.28 shows schematic rock structure models of the holes indicating the
location and orientation of major discrete fractures of importance for the hydration
of the clay, which represents buffer (clay) in the various concepts for disposal of
HLW.
B
A
D
C 6
5
4
3
2 1–6 Heater holes
A Tunnel backfill
1 B Bulwark
C Concrete slab
D Boxing-out
FIGURE 5.27 Persective view of the Stripa underground test site for testing the KBS-3V
concept in general and buffer technology in particular. (Data from Svemar, C., Cluster
Repository Project [CROP], Final Report of European Commission Contract FIR1-CT-
2000-20023, Brussels, Belgium, 2005.)
Figures 5.29 and 5.30 illustrate the recorded swelling pressures. Each of the
graphs have two curves, one representing mid-height of the heater that simulates a
HLW canister and the other representing the lower end of the heater, that is, about
0.3 m over the base of the holes. Hole 2 was located in a minor fracture zone that
made the buffer saturate quicker than in the other holes both at mid-height heater
and at 0.3 m height over the base. Hole 1 was intersected by a number of discrete
1
2
5 6
3 4
FIGURE 5.28 (See color insert.) Schematic rock structure models for BMT holes 1 to 6.
Blue fractures are strongly water-bearing, red fractures are moderately water bearing, green
fractures are poorly water bearing. The arrows indicate north direction.
122 Bentonite Clay
Hole 1 (600 W) Hole 2 (600 W)

6000 12,000
5000 10,000

4000 8,000
3000 6,000
2000 4,000
1000 2,000
0 0
0 20 40 0 20 40
Time (months) Time (months)
Hole 3 (600 W)
450
Hole 3 (1200 W)
400 2500
350
300 2000
250 1500
200
150 1000
100
500
50
0 0
0 5 10 15 0 5 10
FIGURE 5.29 Evolution of the swelling pressure in Holes 1, 2, and 3. The two first-
mentioned holes were located in richly water-bearing rock of the fourth order and the third
in relatively dry rock.
fractures extending from the fracture zone, the buffer being saturated a little slower
than in Hole 2. Hole 3 was located in fracture-poor rock. Here, two canister powers
were used and one notices that the swelling pressure grew quicker for 1200 W than
for 600 W. The lower part (upper curves) of the buffer hydrated more rapidly than
the mid part in both tests, especially in the high-power test. The reason is probably
thermally induced fracturing of the surrounding rock, which raised its hydraulic
conductivity (Pusch et al. 1985; Börgesson et al. 1995).
Evolution of the swelling pressure in Hole 6 was the slowest, which is explained
by the few water-bearing fractures that intersected the hole. As in the high-power
test in Hole 3 and in the test in Hole 4, the buffer at the lower end of the hole (upper
curve in the graphs) hydrated more quickly than at the hotter mid-height of the heater
(Pusch et al. 1985; Börgesson et al. 1995).
The major conclusions from the Stripa BMTs were that the hydration rate of the
buffer was strongly dependent on the ability of the rock to give off water. The buffer
in Hole 2 that was located in a third order fracture zone became saturated at the pre-
dicted rate assuming unlimited access to water, whereas the buffer in the driest hole
took up water so slowly that it would not have become fully saturated until after a
couple of decades. A later series of experiments in SKB’s underground laboratory at
Äspö, some 200 km south of Stockholm and conducted in crystalline rock, also gave
the conclusion that the rate of hydration varies by almost one order of magnitude
Hole 4 (600 W) Hole 5 (600 W)

450 5000
400 4500

350 4000
300 3500
250 3000
2500
200
2000
150 1500
100 1000
50 500
0 0
0 5 10 0 20 40
Hole 6 (600 W)
600
500
400
300
200
100
0
0 10 20 30
Time (months)
FIGURE 5.30 Evolution of swelling pressure in Holes 4, 5, and 6. Hole 5 was intersected by a
major fracture of the fourth order. One finds that the swelling pressure in Hole 5 grew approxi-
mately in the same way as in Hole 1. In Hole 4, the buffer at the lower end of the heater matured
more quickly than at mid-height of the heater. In Hole 5, the two parts behaved similarly. We
have used here a categorization scheme with fracture zones termed first, second, and third
order and discrete fractures called fourth-order discontinuities. (Data from Pusch, R., 1994.
Waste Disposal in Rock, Developments in Geotechnical Engineering 76. Elsevier, New York.)
between individual deposition holes. In contrast, the buffer in the lower part of the
sealed zone and in the deployment zone in VDH will be largely saturated within a
few years because of the high water pressure.
5.7.6 Modeling of the Hydration of Buffer Clay

The desire to work out computer models for predicting the macroscopic rate of satura-
tion and maturation, which includes creep-controlled internal movements and evening-
out of major density differences, has had some success. Application of such codes to
the KBS-3 case shows the same as the field experiments, that is, that the inner half of
the buffer will initially dry even if the rock is assumed to give off unlimited amounts of
water (Figure 5.31). The models predict that the water content will increase with time
and ultimately make the buffer completely water saturated. In practice, especially in
very tight rock, the desiccated state of the canister near the buffer clay will, however,
remain for a long time because the slowly migrating water from the rock can evaporate
and move up to the tunnel in gaseous form and be lost to the tunnel fill, escaping the
buffer clay. For a canister temperature of 100°C, the vapor pressure is about 100 kPa
and is sufficient to keep up the process. For 150°C, the vapor pressure is about 360 kPa.
124 Bentonite Clay
30
25
Water content (%)
20
15
10 X = 54.95 cm Output line

X = 69.45 cm
5 X = 81.05 cm
0
0 20 40 60 80 100
Time (day)
FIGURE 5.31 Example of predicted early evolution of the water content as a function of
time (0–100 days). Rings represent the canister surface, squares the buffer–rock contact, and
triangles half distance between canister and rock. (Data from Svemar, C., Cluster Repository
Project [CROP], Final Report of European Commission Contract FIR1-CT-2000-20023,
Brussels, Belgium, 2005.)
Figure 5.32, obtained using another model (Pusch and Weston 2003), shows the
impact of the average hydraulic conductivity of the surrounding rock for buffer
hydration: the expected successive increase in saturation by diffusive migration of
water under suction of the unsaturated clay takes place in the case of normally con-
ductive rock (K = E−10 m/s), whereas for very tight rock (K = E−12 m/s), there will
initially be underpressure in the rock. In practice, it is not probable that the buffer
would give off water to crystalline rock but well to argillaceous, hydrophilic homo-
geneous rock with high density, which can have such a low conductivity.
The conclusion is that where the rock is very tight, it will give off only little
water to the buffer, which may hence be heated in unsaturated form for more than
100 years, whereas normally fractured rock with common average hydraulic con-
ductivities—E−11 to E−9 m/s at 400 m depth—will lead to a high degree of water
saturation in a few decades up to 100 years. However, complete maturation of KBS-3
buffer clay, implying redistribution of solid matter and microstructural equilibrium,
will not be reached in the first 100 years in normally fractured rock, but well in VDH
for which the piezometric pressure is very high.
5.7.7 Thermo –Hydro –Mechanical–Chemical Processes

in Buffer Clay—Salt Accumulation
One can take the schematic description of the thermohydromechanical process of

buffer evolution in Figure 5.33 as a starting point for considering the conditions
for salt precipitation in the buffer. Assuming that all processes take place radially
in the buffer column, hence disregarding from the risk of loss of water by evapora-
tion upward in the KBS-3 deposition holes, porewater migrates to the desiccation
fractures in the hot part of the buffer and is transported in vapor form toward the
Water saturation ratio, D = 3E−10, K = 1E−10, S = 0.02

1.00
0.95
0.90
Water saturation ratio (%)
0.85
0.80
0.75
0.70
0.65
t = 10 years
0.60 t = 25 years
t = 75 years
0.55
0.50
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Distance from outer boundary (m)

1.0
t = 10 years
t = 25 years
0.9 t = 75 years
0.8
0.7
0.6
0.5
0.4
0.3
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
FIGURE 5.32 (See color insert.) Predicted change in saturation from initially 50% of
KBS-3V buffer. Upper: successive increase in saturation of the buffer in rock with E−10 m/s
hydraulic conductivity. Lower: initial reduction of the buffer saturation followed by delayed
saturation of the buffer clay in rock with E−12 m/s hydraulic conductivity. (Data from Pusch,
R., Weston, R., Appl. Clay Sci., 23, 35–41, 2003.)
126 Bentonite Clay
Isothermal
Largely water-saturated clay matrix
Cold Hot
end end
Water-saturated clay matrix
FIGURE 5.33 Schematic picture of microstructural changes in the initial process of pore-
water redistribution under a thermal gradient. (Data from Pusch, R., Geological Storage of
Radioactive Waste, Springer-Verlag, Berlin, Germany, 2008.)
colder part where it condensates and is sorbed. The colder part of the buffer thereby
expands and compresses the drier, hotter part, the density of which increases on the
expense of the density in the colder part. Migration of water takes place by diffusion
driven by differences in water content.
Water saturation of the buffer under thermal gradients is associated with a
number of chemically induced changes, the extent of which depends on the time
until complete saturation is achieved. In the early phase of maturation of the
hydrated buffer, the absolute temperature is high enough to initiate significant
dissolution of the smectite minerals and accessory minerals, primarily feldspars,
carbonates, and sulfur-bearing minerals (Pusch et al. 1999). The direction of
the temperature gradient implies that minerals with reversed solubility, that is,
carbonates, chlorides, and some sulfur-bearing minerals, will precipitate in the
hottest part of the buffer. This will generate a sink that causes additional ionic
migration and precipitation of new minerals, eventually leading to formation of
a salt crust at the canister surface. For very pure smectite, only Na, Ca, Mg, Cl,
and SO 4 ions entering from the groundwater of the surrounding rock can lead to
precipitation of this type. The rate of their transport to and through the buffer
depends on the diffusivity.
The process is documented by accelerated laboratory experiments in which
Na smectite–rich clay with a dry density of 1270 kg/m3 and 24% degree of water
saturation was exposed to a temperature gradient of 14 C/cm, the cold end being
contacted with 3.5% NaCl solution (Pusch 2008). Figure 5.34 illustrates the redistri-
bution of salt and porewater in the sample after one month. Figure 5.35 exemplifies
40 50
NaCl
NaCl content change (mg/g clay)
W
30 40
Water ratio (%)

20 30
10 20
0 10
−10 0
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
100°C Sample section (n:o) 30°C
FIGURE 5.34 Change in water and NaCl contents from the initial concentrations at
uniform temperature. The successive wetting of the clay and the moving NaCl front are
obvious.
4 μm
FIGURE 5.35 (See color insert.) Example of precipitated NaCl at the cold end of the test
cell in the experiment referred to. Left: scanning electron micrograph of NaCl crystals are
seen embedded in smectite matrix. (Data from Pusch, R., Geological Storage of Radioactive
Waste, Springer-Verlag, Berlin, Germany, 2008.) Right: salt crust at the hot end of the test cell.
the formation of NaCl crystals, which can clog microchannels. When isothermal
conditions eventually occur, the salt can dissolve and diffuse out of the clay.
Under real conditions, the amount of precipitates depends on the diffusion rate in
the surrounding rock of the cations and anions mentioned. Since the process is slow
because of the slowness of Cl and SO4 migration—their diffusion coefficients are
lower than E−11 m2/s in dense clay—the net effect of salt enrichment in the buffer
is believed to be moderate. Most of it occurs early but if the water saturation of the
buffer is very slow, the process may be important and have a practically important
effect on canister corrosion.
128 Bentonite Clay
5.8 CONCEPTS AND TECHNIQUES FOR ISOLATING

MODERATELY HAZARDOUS WASTE
5.8.1 Landfills
We will deal here with the principle of disposal in the form of landfills and consider
then underground disposal in mined, shallow repositories of moderately hazardous
waste, such as low-level radioactive waste and chemical waste. Landfills provide a
relatively cheap and safe form of storing waste for a limited time, a few hundred to
a couple of thousand years because of greater risk for impact of seismic and other
exogenic events, whereas underground disposal can offer very effective isolation for
tens to hundreds of thousands of years. The underground option is primarily con-
sidered in countries with suitable geological conditions, such as salt and very tight
argillaceous rock, and where abandoned mines are available.
We will consider landfills first and begin by taking a look at the ancestors’ clever
way of applying landfill techniques. The capacity of smectite clays to be tight has been
utilized for centuries and millennia. In northern Europe with humid climate, fine-
grained soil containing some smectite has effectively protected constructions such as
moulds from being wetted by rain and meltwater. Many of them consist of rock block
chambers with joints sealed with clayey soil, and they often contain well-preserved
wooden coffins in the dry interior (Figure 5.36). The graves are covered with clay
placed and compacted on top of them and are known to have been successfully kept
dry for thousands of years. They were located on hills where the hydrogeological
settings were favorable and the underground stable as for the 3000-year-old King
Björn’s grave not far from Stockholm. In Germany, the diameter of such mounds is
up to 90 m and the height is up to 15 m, yielding a slope angle of about 18°, that is, on
the same order of magnitude that is required according to stability calculations made
today. Our ancestors were obviously very proficient in soil mechanics.
The old buriers used local clay cut in chunks and placed and compacted them
layerwise using horses and slaves. The same principle, except for slavery, is used
today for deposition of low- and medium-level radioactive waste as exemplified by
Figure 5.37, which shows a concept accepted by the International Atomic Energy
Agency in Vienna and is planned to be built. The bearing principle is to place the
repository consisting of a system of concrete vaults on stiff ground from which it is
separated by dense clay over compacted sand/gravel, and to equip it with an erosion-
resistant draining top cover. This cover can be crushed rock placed over a sand/silt
filter that rests on the top liner of sealing clay. The idea is, for the sake of minimizing
0 20 m
FIGURE 5.36 A 3000-year-old mound at Hersby, Sollentuna, Uppland in Sweden,

according to several classical investigators in the nineteenth and twentieth centuries.
Clay (A) Symmetry

plane
Sand (B)
Slope
angle
3m
17° 4.5–6 m
Rigid concrete cell, inner

height 6 m, thickness 0.5 m
9m
Rigid base
1–1.5 m Clay (D)
FIGURE 5.37 Schematic section of an on-ground repository. Clay (A) minimizes leakage
into the concrete cells where low-level radioactive waste is stored. The layers covering the
clay serve to protect it from erosion and freezing. Clay (D) is proposed to be denser and less
smectitic than Clay (A). (Data from Pusch et al., 2014b.)
cost, to use excavated and stock-piled natural clay with 15%–25% illite/smectite (I/S)
mixed-layer minerals, illite, kaolinite, feldspar, quartz, and traces of calcite, without
processing by screening and milling. Pilot compaction tests of layers with 11–17 cm
thickness of this material gave a dry density of 1610 kg/m3 using 400 kg vibratory
plates and a 7 ton smooth vibratory roller. The hydraulic conductivity was less than
E−10 m/s for percolation with 3.5% CaCl2 solution. A hydrological analysis by use
of the HELP code (Schroeder et al. 1984), assuming a northern European humid cli-
mate, showed that almost no water will penetrate the clay liner and enter the vaults
in the required 300-year period of waste isolation (Pusch et al. 2014b).
Prediction of the performance of this repository concept requires consideration
of several subfunctions: the tightness of the concrete vaults, the stress–strain condi-
tions in the clay surrounding the vaults, and the slope stability of the sand supporting
the clay. A number of typical topics on soil mechanics are hence appearing in the
performance analysis.
The concrete vaults can be designed, so that they remain tight even if earth
shocks of magnitude 5 occur repeatedly (Pusch 2000). The joints are critical fea-
tures, but they can maintain tightness using seals of titanium or copper metal foils.
The performance of the clay surrounding the vaults (Clay A) is controlled by the
shear stresses, which are functions of the vertical load and the lateral earth pres-
sure exerted by the contacting Sand B. The latter will range between the active and
passive Rankine pressures* and presumably average as the earth pressure at rest.
The stress–strain conditions in the whole system require numerical calculation using
finite or boundary element techniques for quantification. This is illustrated in the
graph in Figure 5.38, based on the experimentally determined material parameters:
shear modulus = 4 MPa, Poisson’s ratio = 0.3, and modulus of elasticity = 300 MPa.
* These pressures are defined in classical soil mechanics literature; cf. Terzaghi et al. 1996.
130 Bentonite Clay
1
NODAL SOLUTION
STEP=1 MAY 11 2006
SUB =1 19:28:00
TIME=1
EPTOX (AVG)
RSYS=0
DMX =.073988
SMN =-.003846
SMX =.010391
MX
MN
Y
Z X
-.003846 -.682E-03 .002481 .005645 .008809

rolepsxl -.002264 .899E-03 .004063 .007227 .010391
FIGURE 5.38 (See color insert.) Horizontal strain in the soil components. Horizontal
compression is up to about 0.3% (blue) and tension up to 0.8% (red).
The conservative assumption of complete water saturation of the soil components

already at the end of the construction was made.
The main conclusion from the numerical calculations was that the large sand
mass will be under compressive pressure and stays coherent, and that Clay A is com-
pressed at the contact with the vaults but undergoes an average expansion by about
0.5%, and up to 0.8% at the upper edge of the vaults. This demonstrates that some
expandability of Clay A is required for maintaining coherence and this criterion is
fulfilled by its 25% content of smectite. The lateral earth pressure of the sand on the
clay ranges from 20 kPa at the upper end to 80 kPa at its lower end, which is on the
same order of magnitude as the swelling pressure of the clay. The displacements in
the clay by shear strain are hence not large enough to cause fractures or weakening
that can threaten its tightness.
A similar concept implying use of nonprocessed Iraqi clay compacted to a dry
density of 1700 kg/m3 for preparing a 0.5 m thick clay liner containing 25% Ca
montmorillonite mixed with silty sand has recently been described and analyzed
with respect to its performance under desert conditions (Al-Taie 2013a, 2013b)
(cf. Figure 5.39). According to calculations using the HELP and VADOSE/W codes
(Al-Taie 2013b), the clay will not become water saturated or let water through for
300 years. Calculation of the wetting process of such a repository showed that inter-
mittent periods of heavy rain will not change the condition of long-term and p robably
permanent dryness of the rather thin clay liner (Clay A).
Clay (A) Symmetry

plane
Sand (B)
Slope
angle 17°
0.5 m 4m
Rigid concrete
6m
cell
9m
Rigid base
Sand (C)
Distance between fence and foot of slope = 5 m Building for

hosting data coll.
Slightly convex systems
platform
Vaults with passage

to building
Fence
Natural slope
Moderately
steep slope
(25°–35°) of
reference
case
Road
Steep
slope (32°–35°)
Flat slope
(17°–20°) of alternative case
FIGURE 5.39 Alternative concept of an on-ground repository for low-level radioactive

waste. Upper: design principle with inclined clay layer covering vaults and sand fill (C).
Clay (A) minimizes leakage into the concrete cells where waste is stored. The coarse-grained
layer, Sand (B) covering the clay serves to protect it from erosion and freezing. Lower: appear-
ance of an LLW repository with one concrete vault for hosting the waste.
The main difference between the two design principles indicated in Figures 5.37
and 5.39 is that the latter has a sloping clay liner that can fail by shearing of the clay
if the shear stress is too high, shortly after construction, or much later by creep.
Calculation of the slope stability based on c′/φ′ analyses gave a safety factor of at
least 1.5 both for dry and completely water-saturated clay but the problem with stress
concentrations at the upper edge of the vaults are the same as for the first design as
demonstrated by Figure 5.40. The relatively high safety factor means that failure
132 Bentonite Clay
1
NODAL SOLUTION
STEP=1 OCT 5 2005
SUB=1 18:40:05
TIME=1
SY (AVG)
RSYS=0
DMX =.023363
SMN =–169431
SMX =603806
MN
–169431 2399 174230 346060 517891

–83516 88314 260145 431975 603806
Sigmay
FIGURE 5.40 (See color insert.) Close-up of vertical stresses in Pa in the sand and clay
near the upper edge of the vaults. Compressive stresses are marked dark (max value 170 kPa)
and light blue (minus sign) and tensile stresses are green, yellow, and red (max value 518 kPa).
by creep is improbable. Primary creep that fades out early is believed to occur but
coherence and homogeneity of the clay are maintained by the ductility and self-
healing potential of the smectite component.
The designers of these concepts had to select clays with a smectite content and
density that are sufficiently low to give a swelling pressure that does not exceed the
overburden pressure but sufficiently high to give the required low hydraulic con-
ductivity, E−11 to E−10 m/s. If we compare these concepts with storing waste in
an underground facility, we find them advantageous to the latter in many respects,
primarily because the construction time and cost are much lower and also because
low- and intermediate-level radioactive waste in underground rooms will ultimately
be water saturated and release radionuclides. They are therefore not a priori safer
than those on the ground surface, which can stay totally dry for thousands of years
like King Björn’s mould.
The constructability of on-ground repositories is of course of fundamental impor-
tance, in particular the ways in which compaction on site can be made. For both of
the described concepts, rather light vibratory tools are needed, and for the one shown
in Figure 5.41, the principle of placing and compaction is shown in Figure 5.42.
For the concept shown in Figure 5.39, effective compaction of the sloping clay
liner requires testing of the techniques for placement and compaction (paral-
lel or perpendicular to the slope) and comprehensive quality control. Compaction
14 pav. Kumštelinis (padfoot)

BOMAG firmos vibrotankintuvas
FIGURE 5.41 Example of small padfoot vibrator for Lithuanian NSR construction.
(After Jonynas.)
Lateral support of coarse

soil material, placed and
compacted parallel to the
construction of the clay
layers
Concrete vault
hosting waste
in containers
Clay
Layers of clay compacted on site
FIGURE 5.42 Parallel placement and compaction of clay material and supporting coarser
fill in the construction of a repository.
of flat-lying clay beds, such as bottom liners, is a standard procedure using heavy
vibratory padfoot or sheepfoot rollers (Figure 5.43).
5.8.2 Underground Disposal in Shallow Repositories

For disposal of intermediate-level radioactive waste with long-lived radionuclides, and
of very poisonous chemical waste, national and international legislation and various
134 Bentonite Clay
FIGURE 5.43 Compaction of 30 cm layer of clay with 25% smectite and 20% water content
using 19 t vibratory padfoot roller.
authorities may require underground disposal, of which SKB’s SFR repository some
50 km north of Stockholm serves as an example (Figure 5.44), (Pusch 1994). It con-
sists of a 50 m high on-site cast concrete silo surrounded by about 2 m fill of smectite
granules of Na-montmorillonite-rich clay converted from Ca bentonite (Sud-Chemie,
Germany). The uncompacted fill has a dry density of about 1000 kg/m3 corresponding
Sea level, depth about 10 m
Tunnels for transport of

waste, backfilled at the Radioactive waste 50 m
closure of the repository in silo cells
Cylindrical rock cavern

with 30 m diameter and
60 m height
60 m
Concrete silo surrounded

by 2 m clay fill
Drainage tunnel
backfilled at the closure
of the repository
FIGURE 5.44 The LLW/ILW repository (SFR) at Forsmark, 150 km northeast of Stockholm,
Sweden. (Data from Pusch R, 1994. Waste Disposal in Rock, Developments in Geotechnical
Engineering 76. Elsevier, New York.)
to about 1650 kg/m3 at saturation with the region’s weakly brackish groundwater.
The rock will be kept drained until the space is finally closed at which time the swell-
ing pressure will have increased to about 100 kPa and the water pressure to 0.5–1 MPa.
When the clay has ultimately become water saturated by sorbing water from the rock, it
will slowly transfer water to the silo. The time for penetrating water to reach the waste
is longer than for the concepts for on-ground disposal. Unlike the landfills, there is no
cyclic dehydration–hydration of the clay liner, the whole story is a one-way process.
The mechanisms involved in the water saturation of the clay surrounding the silo
are described Chapters 3 and 4 dealing with hydration by inflow of pressurized water
and/or by diffusion of liquid water or water vapor. For the SFR case, it is a matter of
pure diffusive migration of water. As for any element of smectite clay, hydration is
associated and followed by microstructural homogenization leading to what we call
maturation. It will be reached long after the wetting has ended because of creep-
induced delay in particle reorganization, which will affect the physical properties of
the clay by variations in groundwater chemistry and chemical interaction with the
concrete in the silo. Still, most of the practically important physical properties are
developed in conjunction with the water saturation that is partly controlled by the ris-
ing groundwater pressure after closure of the repository (Pusch 2001c).
The 2-m-thick fill of smectite granules surrounding the concrete silo is in contact
with shotcreted rock that is kept drained until the site will be closed, which means
that the water pressure at the clay–rock contact is very low for several decades, and
that water does not enter into the clay fill by flow. The relative humidity is very high,
however, and causes water to migrate into the fill by diffusion. The matter is of great
importance for defining the pressure conditions at different times, and modeling of
the hydration process was therefore made. The water content predicted was com-
pared with the actually recorded content five years after completing the construc-
tion. Sampling down to 3 m depth in the fill was made both where the rock surface
had appeared wet and where it appeared dry before shotcreting. Pressure sensors of
Gloetzl type had been installed at the rock/shotcrete for measuring the swelling pres-
sure and the groundwater pressure.
The expected increase in water content of the granulate fill, having an initial water
content of 12%–17%, at different distances from the rock contact, was calculated by
assuming diffusive water migration with the diffusion coefficient 3E−10 m2/s, which
had been evaluated from earlier laboratory tests by Börgesson, Clay Technology
AB, Sweden, some 30 years ago. The prediction was based on the assumption that
the large number of interconnected drains at the contact between rock and shotcrete
would even out differences in access to water for the clay. It gave the water content
distributions showed in Table 5.3, implying that the required time for reaching more
than 50% degree of water saturation closer to the rock than 0.3 m would be about five
years. Complete water saturation of the entire clay mass, corresponding to a water
content of 60%, would take several hundred years under drained conditions.
The table shows that there were considerable variations in the degree of wetting
and that the clay had taken up water zone-wise all the way from the rock to the silo,
probably by water vapor moving in channels in the wetted part of the fill. Local
zones with higher water content are assumed to have given off water to neighboring,
drier clay, by diffusion, which will make the wetting more homogeneous with time.
136 Bentonite Clay
TABLE 5.3
Meaured and Predicted Water Contents in 1994, Five Years after Completing
the Filling
Distance from Actual Degree
Wet (W) Shotcreted Measured Water Predicted Water of Water Saturation
or Dry (D) Rock Rock (m) Content (%) Content (%) in the Fill (%)
D 0.1–0.2 51–59 60 85–98
D 0.6–0.8 24–32 25 40–53
D 2.0–2.5 18–29 12–17 30–48
W 0.1–0.2 24–39 60 40–65
W 0.6–0.8 19–20 25 31–33
W 2.0–2.5 19–41 12–17 31–68
The practically most important question was whether a significant swelling pressure
would be built up early and results from published experiments for finding the relation
between the degree of water saturation and swelling pressure of smectite clay were
helpful in this respect. Thus, the diagram published by Shirazia et al. (2010), shown in
slightly different form in Figure 5.45, suggests that significant swelling pressure would
in fact not be expected until the degree of water saturation is about 50% and the record-
ings logically show no pressure in the first five years. The slow rate of diffusive wetting
7
ρd = 1900
kg/m3
6
5
Ps (MPa)
4
ρd = 1670
kg/m3
3
2
ρd = 1480
kg/m3
1
0
0 20 40 60 80 100
Sr
FIGURE 5.45 Relation between dry density and swelling pressure of smectite-rich clay at
45°C and different degrees of water saturation. (After Shirazia, S.M. et al., Environ. Asia, 3,
60–64, 2010.)
means that the initially developed silo pressure will remain largely constant until water
begins to fill the drain system and give unlimited access to water.
In a long-time perspective the pressure on the silo will increase by the fully
developed swelling pressure and of the water pressure, the first mentioned being
300–600 kPa depending on the salinity of the groundwater. In a few tens of years
after closing the repository the water pressure will increase to about 1 MPa at the
base of the silo and to 0.5 MPa at its top.
The question now is what the reaction between the huge concrete mass of the
silo and the clay will be. Cation exchange of the smectite clay from dominantly Na
to Ca will begin when the degree of water saturation has risen to about 50% at the
clay–concrete contact, that is, to a level when the content of free porewater becomes
important. The Portland cement component of the concrete starts to dissolve at this
stage and pH can rise to more than 11–12, which will initiate dissolution of the clay.
We will look closer at the mutual degradation of the materials in Chapter 6 for assess-
ing the risk of mechanical instability of the silo and the ability of the clay to retard
migration of possibly released radionuclides. One of course realises that the risk of
degradation of concrete and clay is much lower for landfills of the described types
than for the underground repository, which will be constantly water saturated and
exposed to adflow of groundwater after some decades.
5.8.3 Underground Disposal in Abandoned Mines

Figure 5.46 shows a section of an abandoned iron mine at Stripa, Sweden, where a
comprehensive research program was carried out in the late 1980s for testing engi-
neered clay barriers. It gave a good overview of the rock structural features and
200
Scale
0 40 m
310
360
410
FIGURE 5.46 Cross section of the iron ore bodies (black) at Stripa. Dotted lines and rec
tangles denote shaft, tunnels, and caverns. The larger part of the ore had been mined out in
the 1970s, leaving a big open cavern and several smaller ones, as well as numerous drifts. The
Buffer Mass Test took place at 360 m level in the granitic rock. (Data from Pusch, R., 1994.
Waste Disposal in Rock, Developments in Geotechnical Engineering 76. Elsevier, New York.)
138 Bentonite Clay
Host rock
Layer-wise
applied and
compacted
granular
waste in big
room
FIGURE 5.47 Proposed use of the big cavern for disposal of hazardous waste such as
batteries or solidified pecticides. The layers of waste are separated by thin layers of smectite
clay compacted on site. The lower pointer indicates a 0.5–1 m thick, layer of on-site com-
pacted clay granules or placed blocks of highly compacted smectite clay. The entrance to the
cavern is via a drift leading to its upper part. (Data from Pusch, R., 1994. Waste Disposal in
Rock, Developments in Geotechnical Engineering 76. Elsevier, New York, 2014.)
served as an example of possible use of mines for disposal of hazardous solid waste
(Figure 5.47).
The hydrological and rock mechanical conditions in the Stripa mine were investi-
gated in a EU Project Low-Risk Deposition Technology (Popov and Pusch 2006) for
predicting the risk of contamination of groundwater if it would be filled with chemical
waste. The study included generalization of the geometry of mined-out major parts to
consist of an actually existing big cavern (100 m length, 50 m height and width) con-
nected to a 100 m long tunnel with 5 m height and width, both taken to be backfilled
with alternating layers of smectite clay (Figure 5.47). The assumed waste consisted of
spent batteries with 25% zinc as the most hazardous component. Dissolution of the
waste was estimated to have an average zinc concentration of 10,000 ppm at the inter-
face at the clay–EDZ* contact, assuming conservatively that the water here had reached
complete saturation with dissolved zinc at the start of the groundwater percolation of the
cavern. The total porewater mass in the cavern, 50,000 m3, would contain 5000 metric
tons of dissolved zinc corresponding to a total battery mass of about 20,000 metric tons.
The problem to be solved in the EU Project was to calculate the contamination of
the water in a vertical well located 110 m from the cavern. The geometry of the model
was as shown in Figure 5.48, while Figure 5.49 indicates the generalized rock structure
with discontinuities in the form of minor fracture zones with a spacing of 50 m, and
orthogonally grouped, persisting discrete water-bearing fractures with 5–10 m spacing.
For the various hydrological and mechanical modeling attempts in the project, the
physical properties in Table 5.4 were used.
* EDZ = excavation-disturbed zone, assumed to be 1 m around the tunnel and 3 m around the cavern.
N (x-direction)
x = −400, y = 900
Z
Big room 1000 m
z = −360 m
Tunnel for disposal
length 300 m
FIGURE 5.48 The Stripa cavern and tunnel.
The contamination of the water in the 110 m distant well would indicate the role
of the system of discontinuities in distributing the groundwater flow. Water flowing
through the cavern and drift would be retarded by the low conductivity of the clay in
the mixed clay and waste and particularly by the ion-absorbing capacity of the clay lin-
ing. The clay material was assumed to be the aforementioned FIM clay consisting of
smectite/illite mixed-layer clay with an ion exchange capacity of 30 mg per 100 g dry
material. For determining the degree of contamination, the source term was defined
as the average concentration of zinc at the clay–EDZ contact, making the over-conser-
vative assumption that the water here had reached complete saturation with dissolved
zinc at the start of the groundwater percolation of the cavern (10,000 ppm).
The outcome of the study was summarized as follows:
• The maximum gradient of hydraulic head over the whole domain is

∇h = 0.05.
• The role of the clay component to delay groundwater ingress into the waste
and to provide support to the rock by its expandability is very obvious.
• The EDZ takes the larger part of the flow in the vicinity of the room.
• Assuming that the selected clay of illite–montmorillonite–chlorite in ratio
50:50 by volume is constantly fully water saturated, and that Kd can be taken as
0.02 m3/kg, application of the relationship R = 1 + (ρ/m) Kd, R = 121, the leak-
age of zinc to the well will be as in Figure 5.50. It implies that in the period
of 120,000 years, the concentration of zinc in the observation well will not
exceed 2 ppm, which is much lower than the critical limit according to com-
mon legislation criteria. The results in fact represent a conservative case and
in reality, it would be very unlikely that the concentrations could reach this
level, because the flow of contaminated water in the rock matrix means that the
cation-sorbing capacity of the fracture minerals will also provide retardation.
140 Bentonite Clay
Second-order discontinuity (boundary) Second-order discontinuity (boundary)
Third-order discontinuity
500–1000 m
Third-order discontinuity
Second-order discontinuity (boundary)
Third-order discontinuities (boundaries)

Fourth-order discontinuities

50
–1
00
m
FIGURE 5.49 Simplified rock structure model of the Stripa host rock. Spacing of second-
order discontinuities (major fracture zones) 500–1000 m. Spacing of third-order disconti-
nuities (minor fracture zones) 50 to 100 m. Spacing of fourth-order discontinuities (discrete
fractures) 5 to 10 m.
TABLE 5.4
Hydraulic and Transport Properties of the Different Entities Used in the Model
Clay Fill (Room Rock (Porous Fracture EDZ Close EDZ Close
Parameter and Tunnel) Matrix) Zones to the Room to the Tunnel
K (m/s) 1E−10 1E−9 1E−7 1E−7 1E−7
D (m2/s) 1E−10 1E−9 1E−7 1E−7 1E−7
K—hydraulic conductivity; D—diffusion coefficient of zinc in clay.

2.0
Time
1.5 48,400
Concentration (ppm)
72,600
1.0
0.5
0.0
−0.5
−800 −600 −400 −200 0
Depth (m)
FIGURE 5.50 Zinc concentration in the well as a function of time (Popov and Pusch 2006).
Occasionally, the idea of constructing a new repository of KBS-3V type for disposal
of hazardous chemical waste comes up and a Swedish mining and mineral process-
ing company recently announced that it plans to construct an underground repository
for the 300,000 tons of arsenic that is presently stored in silos in its backyard. The
problem is that site investigations needed as a basis of the design of such a repository
with rooms, tunnels, and shafts would have to start from scratch. Topographic and
remote geophysical investigation methods provide information that, by experience,
will be largely incomplete and difficult to interpret until systematic deep borings
have been made and construction has started and proceeded. In fact, the information
on the rock constitution will still not be as good as for an abandoned mine because
the experience respecting water flow, rock structure, and stability specified in daily
loggings from the mining operations will not be reached until the construction of a
new one has been largely completed. The possibility of using abandoned mines is
clearly demonstrated by the comprehensive and successful research on geohydrol-
ogy, rock mechanics, and waste isolation conducted in the abandoned Stripa mine
that was converted to an underground laboratory by SKB in the 1980s and 1990s
(Popov and Pusch 2006).
Access to a suitable abandoned mine, where ore containing the same type of haz-
ardous chemical component, such as As, Pb, and Hg, as the waste to be deposited,
would of course be ideal. Still, there can be problems also with utilizing abandoned
mines, primarily with the pH-conditions since they determine the solubility of the
waste. In sulfide-bearing rock found in deep mines, pH can be very low and can lead
to extensive dissolution of placed chemical waste requiring use of smectite clay lin-
ings for retarding wetting of the waste and controlling the migration of hazardous
element to the groundwater.
142 Bentonite Clay
5.9 CONCEPTS FOR ISOLATING HIGHLY RADIOACTIVE WASTE

5.9.1 Medium-Deep Repositories
The concepts depicted in Figure 5.51 are alternatives to the earlier mentioned KBS-3
concept, and we will briefly examine them here for seeing what the major differences are.
5.9.1.1 The KBS-3V Concept

We have discussed the KBS-3 concept, in modern times called KBS-3V, used here as
reference case; we now recall the main features of this concept in Figure 5.52. The
principle to be followed in the site selection process is to locate and orient the reposi-
tory components, deposition tunnels, and main tunnels and shafts, so that they adapt to
the rock structure with discontinuities of different orders, and to the horizontal princi-
pal rock stresses, which often have varying directions. The idealized plan view in this
figure shows the system of deposition tunnels extending from major tunnels for trans-
port, without applying the assumed criterion of a respect distance from the fracture
zones of 50 m. This significantly reduces the factor of usability, which can be down to
20%, being less than half of that of VDH. The initially proposed diameter of 1.5 m for
the KBS-3V deposition holes has been raised to about 1.8 and later to 1.9 m.
We have seen the characteristic section of blasted tunnels with deposition holes bored
from the tunnel floors already in the preface of book and we see it again in Figure 5.53,
which indicates one of the problems related to the performance of the dense smectite-
rich buffer interacting with the overlying backfill of more deformable and less smec-
tite-rich material. Comprehensive numerical calculations for predicting the mechanical
performance and temperature conditions for KBS-3V have been made and they show
A B C
FIGURE 5.51 Four repository concepts (A through D) for crystalline rock. Small cylinders in
bigger ones represent HLW canisters surrounded by buffer clay. Down to the extreme right (C),
we see a concept called (KBS-3i) with the holes turned to the side and axially by 45°. (Data from
Pusch, R., Geological Storage of Radioactive Waste, Springer-Verlag, Berlin, Germany, 2008.)
0 100 200 300 400 500 m

3 3
2
2
3
1
1.5 m
kBG
b a
d
FIGURE 5.52 The KBS-3V concept. Upper: schematic plan of repository adapted to the
presence of a fault (1), major fracture zones (2), and minor fracture zones (3). Thick lines
oriented N/S are main tunnels, the sets of parallel lines are deposition tunnels. Lower left:
completed deposition tunnel with row of deposition holes with 6–8 m spacing and canisters
embedded in dense buffer clay. Lower right: deposition hole intersected by a major fracture into
which clay can migrate and along which shearing can take place. (Data from Pusch, R., 1994.
Waste Disposal in Rock, Developments in Geotechnical Engineering 76. Elsevier, New York;
Pusch, R., Geological Storage of Radioactive Waste, Springer-Verlag, Berlin, Germany, 2008.)
144 Bentonite Clay
ABAQUS
ABAQUS
Orginal mesh
3,2 y 12,7 y
Displaced mesh
Enlarged
1 Mag. factor = + 1.0E+00

area
2
3
FIGURE 5.53 FEM-calculated movement of canister and buffer clay showing displace-
ment of KBS-3V tunnel backfill by upward expanding buffer in deposition hole. Unfilled
space around the deposition hole and on the top of the canister are marked black. Left: after
3.2 years. Right: after 12.7 years. (Data from Pusch, R., Börgesson, L., Performance assess-
ment of bentonite clay barrier in three repository concepts: VDH, KBS3 and VLH. PASS—
Project on Alternative Systems Study, SKB Technical Report TR 92-40, Swedish Nuclear
Fuel and Waste Management [SKB], Stockholm, Sweden, 1992.)
that the upward expansion of the buffer caused by the swelling pressure has a signifi-
cant effect on the overlying backfill that was assumed to consist of on-site compacted
mixture of milled smectite granules and coarser soil in the calculation referred to here.
The graphs in Figure 5.53 show that there will be a gap between the top of the
canister and the overlying backfill. According to the calculations, it stays for more
than 10 years and will be filled with water vapor emanating from the buffer clay that
surrounds the hot canister in this period. This will cause corrosion of the top of the
canister and expose the buffer clay to hot vapor, which is known to cause reduction of
its expandability (Couture 1985). The gap at the upper end of the deposition hole in the
figure will in fact grow and facilitate flow of groundwater and erosion of the backfill
that will soften and promote further upward movement and softening of the buffer.
The gaps in Figure 5.53 and the fracture intersecting a KBS-3 deposition hole in
the lower left picture in Figure 5.52 represent additional problems since water flowing
through them can erode the buffer and reduce its density. The risk of such events is
similar for the other KBS-3 concepts, that is, KBS-3H and KBS-3i, and depends on the
rate of water flow and on the density of the buffer. It has not yet been fully assessed.
Prediction of the temperature evolution in a KBS-3V repository assuming unlim-
ited access to water from the rock, indicates that the temperature of the buffer clay
using SKB’s canisters will increase by about 72°C, and hence reach about 90°C at
100
90
80
Innermost
Temperature (°C)
70
60
Outermost
50
40
30
20
10
0 100 200 300 400 500 600 700 800 900
Time (days)
FIGURE 5.54 Rise of temperature at mid-height of the canister in the wettest hole (No. 1).
Upper curve is for the canister surface and the lowest for the rock contact. (Data from Svemar,
C., Cluster Repository Project [CROP], Final Report of European Commission Contract
FIR1-CT-2000-20023, Brussels, Belgium, 2005.)
mid-height canisters (Figure 5.54). Full-scale experiments at SKB’s underground

laboratory at Äspö, some 300 km south of Stockholm, confirmed this and showed that
the maximum swelling pressure became 4.7 MPa after about two years (Figure 5.55).
One concludes that the recorded temperature still rose after more than four years
and reached about 72°C adjacent to the canister surface and around 60°C at the rock
after about two years. The average temperature gradient was thus about 0.34°C/cm
radial distance.
8.0
Outermost
7.0
6.0
Pressure (MPa)
5.0
4.0
Innermost
3.0
2.0
1.0
0.0
0 100 200 300 400 500 600 700 800 900
Time (days)
FIGURE 5.55 Evolution of total pressure at mid-height of the buffer in the wettest hole at
Äspö. The upper curve indicates a high degree of water saturation already after one year and
the lower curve indicate nearly complete saturation of the whole buffer after two years. (Data
from Svemar, C., Cluster Repository Project [CROP], Final Report of European Commission
Contract FIR1-CT-2000-20023, Brussels, Belgium, 2005.)
146 Bentonite Clay
The evolution of the swelling pressure gave the highest pressure, about 6.7 MPa,
after about two years close to the rock, whereas the lowest (4 MPa) represented the
immediate vicinity of the canister.
Four national numerical calculation codes were used for predicting the evolution
of the swelling pressure, employing the same parameters for the calculations. The
three most accurate codes gave the predicted data in Tables 5.5 and 5.6, which also
give the actually recorded data (Svemar 2005).
Comparing the actually recorded swelling pressure with the various predictions,
one finds (Svemar 2005) that:
• The predicted temperature agreed reasonably well with the actually

recorded ones for all the codes except for THAMES.
• The predicted swelling pressure varied strongly among the codes and only one
of them (COMPASS) gave values in reasonable agreement with the actually
recorded pressure for the canister-contacting buffer, which is most important.
However, it failed to give pressure values on the right level for the rock contact.
The models were concluded to be misleading concerning the evolution of the
swelling pressure and hence also of the rate of hydration of the buffer.
TABLE 5.5
Recorded and Predicted Temperature in Degrees Celsius at the Canister and
Rock Surfaces at Mid-Height of the Wettest Hole after One and Two Years
from Start of Heating
COMPASS CODE_BRIGHT THAMES (JNC),
Location Recorded (UWC), GB (CIMNE, Enresa), Spain Japan
Canister 1 y = 69 1 y = 70 1 y = 70 1 y = 87
2 y = 72 2 y = 72 2 y = 72 2 y = 92
Rock 1 y = 56 1 y = 56 1 y = 57 1 y = 71
2 y = 60 2 y = 59 2 y = 60 2 y = 76
TABLE 5.6
Recorded and Predicted Pressure in MPa at the Canister and Rock Surfaces
at Mid-Height in the Wettest Hole after One and Two Years from Start of
Heating
COMPASS CODE_BRIGHT THAMES (JNC),
Location Recorded (UWC), GB (CIMNE, Enresa), Spain Japan
Canister 1 y = 1.0 1 y = 0.8 1 y = 3.0 1 y = 4.7
2 y = 4.0 2 y = 3.2 2 y = 5.1 2 y = 6.2
Rock 1 y = 6.0 1 y = 2.8 1 y = 5.0 1 y = 6.4
2 y = 6.7 2 y = 3.9 2 y = 7.2 2 y = 7.2
While the wet conditions in the reference hole were well defined with respect to the
hydraulic boundary conditions, the question is naturally how accurately one can pre-
dict the hydration and maturation of buffer in deposition holes with only little access
to water. The experience so far is that it is very uncertain and that it is not known
whether water saturation will take a few tens or hundreds of years for buffer in dry
deposition holes. A valid question is also whether the recorded data are correct and
it in fact turned out that the measurements overrated the pressures obtained in the
first three years, making the COMPASS predictions more correct for the pressure on
the canister whereas those for the rock were still only 50% of the correctly measured
ones. The matter is further treated in Chapter 7.
5.9.2 Steep Holes with Two or Several Canisters (Case A)

The concept is a hybrid of the KBS-3V and VDH concepts and requires the place-
ment of the clay-embedded canisters down in drained holes with a diameter of up
to 2 m. The problems with too much water flowing into the hole that can make the
KBS-3V impractical will be even more troublesome for this concept. We will not
consider it here.
5.9.3 Big Cavern with Numerous Canisters (Case B-1)

This concept was proposed in the late 1980s as a way of minimizing construction cost
and simplifying canister installation, which requires use of fully automatic remote-
handling techniques. The prerequisite is that the cavern is lined with tightly fitting
smectite-rich blocks that effectively hinder percolation of groundwater. Very high
buffer temperature values, that is, more than 150°C, have to be accepted. A valuable
feature is that the time for hydration of the buffer can be many thousands of years.
A less good property is that all the canisters need to be placed in one, continuous
campaign. The problem with the concept is the difficulty in providing radiation pro-
tection in the waste installation phase. Retrievability would be possible.
5.9.4 Tunnels or Drifts with Large Clay-Isolated

Concrete Containers (Case B-2)
The concept has very much in common with the method of using large caverns
with thick liners of compacted clay blocks (Popov and Pusch 2006). For minimiz-
ing the impact of heat, canisters with less HLW, for example, with 2–4 elements
of spent fuel instead of the 12 elements implied by SKB are proposed. Copper
(or Navy Bronze), titanium, or steel can be considered as canister material and
concrete for the large and stiff containers. They can be prepared elsewhere in
the repository and filled with concrete for embedding the canisters, all activities
being made by robot techniques. Transport to the disposal room or drift requires
shielding of these heavy units and remote handling for placing the containers and
filling the space between them with concrete. Figure 5.56 shows the assembly of
5–10 m 3 containers as squares surrounded by pumped in dense cement (congrete)
148 Bentonite Clay
Shotcrete
Dense smectite-rich
clay blocks
Dense low-pH
concrete
Waste packages
FIGURE 5.56 Principle for deposition of spent reactor fuel contained in metal canisters
placed in containers and embedded in low-pH concrete. The upper right pointer indicates
low-pH shotcrete. The lower right pointer indicates a liner of fitted blocks of smectite-rich
clay. The left pointer denotes dense low-pH concrete cast successively in campaigns.
grout (cf. Pusch et al. 2014a) and describes the principle of sarcophagus-type
burial of HLW. The bottom containers rest on cast low-pH concrete and additional
ones stacked on them are placed with great precision. The space between and
around the placed containers is filled with dense silica-rich concrete with a small
content of low-pH cement (Pusch et al. 2014a).
The internal movements in the system are expected to be small and minimized
using somewhat ductile concrete and, most important, with self-sealing ability pro-
vided by the smectite-rich clay liner that has to be 0.5–1 m thick. Designers of can-
didate concepts such as the one in Figure 5.57, according to which the HLW canister
rests on an altar of dense smectite blocks in a BMT tunnel with smectite granules
(pellets) blown to fill the space between the rock and the canister, must have realized
the risk of large uncontrolled movements and potential instability.
5.9.5 Inclined Holes (Case C)

The concept, KBS-3i, was proposed some years ago when SKB had realized that all
deposition holes in a KBS-3V repository will fail because of the high hoop stresses
generated by thermal effects, and when it had become obvious that placement of
individual clay blocks for subsequent insertion of the 20-ton canisters is impractical
and unsafe.
The concept, shown in Figure 5.58, is similar to the KBS-3V concept but has
each hole turned sideways by 45° in two directions with respect to the axial direc-
tion of the tunnels and to the vertical and horizontal axes. The idea behind this
is that the critically high hoop stresses are significantly reduced from those in
KBS-3V deposition holes, and also that the placement of containers would be
much simpler.
Liner
(stainless steel)
y
cla
Buffer
m
o
Bo
(OPC-based concrete)
Overpack
(carbon steel)
Space between overpack

and canisters, filled with
borosilicate glass frit
Disposal gallery lining
(concrete wedge blocks)
Waste canister
Void space,
Mechanical support to be backfilled after
waste emplacement
Gallery floor
FIGURE 5.57 Swiss concept implying placement of the heavy HLW canister on stacked
blocks of highly compacted smectite-rich clay with subsequent filling the open space with
blown smectite granules. (Courtesy of NAGRA, Switzerland.)
Mechanical packer removed in the

backfilling phase, allowing slight
axial displacement
Supercontainer of
perforated tube of navy
bronze with clay blocks and
Tunnel floor canister
Temporary fill
Blocks of highly
compacted blocks of
smectite-rich clay in the
supercontainer
Mud is pressed up
along the
supercontainer Mud filled before
when it is placed submerging the
in the hole supercontainers
Canister
FIGURE 5.58 The inclined deposition hole design, KBS-3i.

150 Bentonite Clay
Inclined deposition holes would reduce the risk of rock spalling and make placement
of the waste safer and simpler using containers (Pusch and Weston 2003). The contain-
ers, being similar to those of SKB’s version of long, horizontally oriented holes, KBS-3H,
can slip down in clay mud in the deposition holes, as in the VDH case, for becoming
embedded in clay mud from which the buffer uniformly sucks water for hydration. As
for KBS-3V, the ability of the rock to provide water for complete water saturation has to
be sufficient but the mud in KBS-3i at least provides early uniform access to a signifi-
cant amount of water. Retrievability will be possible as for the KBS-3V case.
The concept has, in summary, a number of advantages compared to the KBS-3V
concept:
• The overlap of the local temperature fields reduces the hoop stress of the
holes in comparison with those of the KBS-3V case.
• Use of the container principle means that individual handling of canisters
and clay blocks in the KBS-3V case is replaced by working with coherent
and strong supercontainers as for KBS-3H (Case D in Figure 5.51) and
VDH.
• Comparison of the stress levels for the two repository concepts shows that
KBS-3i hole is significantly lower and nowhere reaches up to 60 MPa for
normal rock stress conditions. The latter concept implies significantly safer
conditions since the risk of spalling is more than 20% lower than for KBS-3V.
5.9.6 Very Long Holes (Case D)

This solution, illustrated in Figure 5.59, is SKB’s concept KBS-3H, which has much
in common with VDH in the sense that containers* with clay-embedded canisters
and separated by clay blocks are inserted in holes with about the same diameter as
KBS-3V and KBS-3i but with a few hundred meters length. Where water-bearing
fracture zones are intersected these have to be grouted and confined by twin bulk-
heads between which frictional material shall be filled.
The very long holes will be bored by TBM technique, or subhorizontal
raiseboring, to a diameter of about 2 m. The advantage of this concept compared
with KBS-3V is, in addition to some cost saving, that the volume of excavated rock
Super- Filling Drift end

Copper canister Distance block container block plug Deposition niche
Compartment plug ∼15–25 m
Grouted fracture Qinflow Qinflow Qinflow

1 < Q < 10 l/min Q < 0.1 l/min 0.1 < Q < 1 l/min
Approximately max 300 m
FIGURE 5.59 Schematic view of KBS-3H. The plugs, termed compartment plugs serve to
isolate parts with high water inflow from those hosting supercontainers and distance blocks.
(SKB, 2012.)
* Termed supercontainers by SKB and POSIVA.

Perforated Canister with Permanent disposal

Deposition casing waste cell liner
hole
Engineered
buffer
Permanent
Dense, smectitic sleeve
clay (buffer)
FIGURE 5.60 (See color insert.) Main components of installed supercontainer for spent
reactor fuel. Left: canister surrounded with buffer clay and fitted in perforated container.
Right: French design with sleeve of carbon steel for hosting canisters. (Courtesy of ANDRA,
France.)
is much smaller. A drawback is that the inflow of water from intersected fracture
zones, which statistically are met with every 30–70 m length of the holes, can cause
quick dispersion of the clay blocks that separate the HLW canisters. Another dif-
ficulty is to bring the very heavy containers with clay-embedded canisters on site.
It is proposed to be made by pneumatic technique, which works in an underground
laboratory, but will meet difficulties in practice because the narrow space between
rock and container can make the latter stuck, especially if rock fragments fall from
the roof. In contrast with VDH, in which adjustments of the much smaller con-
tainers can be made remotely by use of a drill rig, the heavy containers may have
to be pulled out for repair and cleaning of the hole and then brought in again.
Anyone familiar with the function of smectite clay in underground construction
would realize the problems with much water flowing into a tunnel, causing disper-
sion of the clay. The least cumbersome of the many practical risks of mishap is that
the floor will be very slippery, a fact that has been underestimated by the designers.
Furthermore, retrievability of the long supercanisters from the inner part of the
holes will not be possible.
Figure 5.60 shows a French version proposed* for horizontal holes. It may well
be used in very tight argillaceous rock if it is combined with injection of water or
clay mud from temporary or permanent bulkheads located at the outer ends of the
holes, for filling all open space in the system. The yellow clay blocks are fitted into
the perforated container that has a steel sleeve for letting the canisters in. The prob-
lem with this concept is the deformation or displacement of the sleeve. Canisters
will not be retrievable because sleeves may become elliptic by nonuniform wetting.
5.10 VERY DEEP HOLES

We have already seen some details of this concept and come back to it here for
describing the principle of locating the deep holes and the main stages in prepar-
ing them as well as for predicting the hydration and maturation of the clay seals.
* Later abandoned.
152 Bentonite Clay
2 A 3
1 2
3 4
3 B
4
A
2 4
2 50–
150
M
3
500
–15
00 M 2 5 4
B
5–
50
M
FIGURE 5.61 Structural hierarchy of large rock block with two major fracture zones
(1 and 2) and smaller blocks with minor fracture zone (3), in turn consisting of even smaller
blocks with discrete fractures of fourth and fifth orders.
The principle followed for finding acceptable canister positions is to identify

fracture-poor parts of pilot holes between major fracture zones, that is, those of
first and second orders (Figure 5.61). Minor zones are of no importance since, at
large depth, the rock stress situation will only cause displacement along major ones
(Pusch et al. 2014a). A possible problem can be that the clay seals and clay/waste
packages expand too much in the placement phase for bringing them down but this
problem can be solved by coating the containers with a paste of mixed clay and talc
or using prewetted blocks as described earlier. Another problem is that axial move-
ments of clay block series and occasional concrete seals can take place in the deep
holes. Minor movements are expected in the process of hydration of the clay and will
be of acceptable magnitude. Large movements can be avoided by letting the clay
seals mature sufficiently long enough before additional load is placed (Pusch et al.
2014a). Retrieval of the containers would be very difficult but is deemed possible.
The site investigation and design work required for construction of a VDH repos-
itory will take at least a decade for 5 to 10 deposition holes in each of 5 to 10 sites
Fracture filled with

grout
Water-bearing
fractures
Concrete plugs
in fracture zone
Clay plugs
Concrete plugs
FIGURE 5.62 Schematic picture of VDH with clay and concrete seals. (Pusch et al. 2014a.)
required for hosting the same amount of waste as KBS-3V. The degree of utilization
of the rock is expected to be 50%–60%. Boring of the deposition holes will be made
using oil-drilling techniques with smectite boremuds. The VDH discussed here has
800 mm diameter and contains HLW from 2000 to 4000 m depth and clay seals
above the 2000 m level, the uppermost 500 m being filled with other soil material
and concrete (Figure 5.62). Moderate to low temperatures prevail in the sealed parts
while the temperature can be up to 150°C in the lower, waste-bearing part. Wetting
of the buffer will take place under high water pressure (20–40 MPa) in this zone
and is completed in a few years. The waste is confined in canisters of copper, Navy
Bronze, or titanium, surrounded by dense montmorillonite or saponite clay blocks,
all held together in perforated supercontainers of the same material. These units are
forced down in the deployment mud, which can be the same as the drilling mud, and
which gives off water to the unsaturated dense clay blocks in the containers. The
blocks hydrate quickly under the high piezometric pressure (Pusch et al. 2014a).
5.11 CORRELATION OF HYDRAULIC AND MECHANICAL

PERFORMANCES OF CLAY SEALS
5.11.1 Piping and Erosion of Clay Seals
We have seen in the discussion of the hydraulic conductivity of smectite clays in
Chapter 4 that evaluation of this property strongly depends on the hydraulic gradi-
ent: low gradients leave the microstructural constitution intact, whereas high ones
can cause rapid flow and transport of eroded fines, but also very slow flow depending
on the test conditions.
154 Bentonite Clay
Piping followed by erosion is a risk for any clay seal that is exposed to a hydraulic
gradient. Physical models and experiments show that a flow rate of E−3 m/s is suf-
ficient to produce dislodgement of particles with a size of 0.5 μm, E−4 m/s for 1 μm
particles, E−5 m/s for 10 μm particle aggregates, and E−7 m/s for aggregates, the
lower sensitivity for small particles being caused by higher numbers of interparticle
bonds (Pusch and Yong 2006).
Comprehensive laboratory testing has shown that the critical gradient is
30–40 m/m for causing quick channel-wise water penetration into soft smec-
tite clay, initiating erosion in the form of particle dislodgement and migration
(Figure 5.63).
The consequence of internal erosion is obvious for clay seals in mined under-
ground facilities such as KBS-3V and KBS-3H repositories and is well known to
occur in clay-rich cores of earthen dams (Figure 5.64), mentioned en passant a few
Grout holes
Highly compacted
bentonite
Concrete
FIGURE 5.63 Upper : piping in the form of a hydraulic wedge penetrating into soft smec-
tite clay causing meandering. One sees 20–50 μm aggregates moved by the flow (E−4 m/s).
(Data from Pusch, R. et al., Piping and Erosion Phenomena in Soft Clay Gels, SKB Technical
Report TR 87-09, Swedish Nuclear Fuel and Waste Management [SKB], Stockholm, Sweden,
1987a.) Lower : case with clay seal of O-ring type around concrete bulkhead in tunnel being
sensitive to piping and erosion by a high hydraulic gradient acting across the bulkhead.
9
V,Y
8
1 2
8
3
4
5 7 5
10
FIGURE 5.64 Embankment dam with low-permeable core, 1. 2 to 4 are filters and 5 side
fills. 6 denotes grouting before construction of the dam, 7 is a concrete slab cast upon the
grouted area. 8 is a layer of coarse-grained material for protection of the slope.
pages ago, in which successive increase in the cross section of channels can lead
to piping and breakthrough followed by complete failure (Hellström 2009; Pusch
et al. 2012b).
5.11.2 VDH Concept
One realizes that, for this, and all other HLW disposal concepts, the microstruc-
tural constitution and its reaction to compressive and shear stresses of smectite
clay seals are determinants of their physical performance, and that it also controls
the transport of water and solutes through them. The VDH is an extreme in this
respect and we will take a closer look at the evolution of the clay seals in the deep
holes and start by considering each of them. They are (1) clay mud for stabilizing
the holes and for providing clay for raising the average ultimate density of the clay
seals, (2) blocks of compacted clay granules for providing effective sealing of the
upper 2 km of the holes, and (3) blocks separating and isolating HLW canisters in
their lower 2 km parts.
5.11.2.1 Clay Mud

A mud of potential value is Na-montmorillonite-rich clay with 1100–1200 kg/m3
density. It can be the same as the bore mud or pumped from below through the drill-
ing rods to replace this mud. Salt groundwater in the holes has to be replaced by
electrolyte-poor water in the installation phase since soft muds coagulate and settle as
sediments. We assume here that the clay mud filled in the holes has montmorillonite
as major smectite component with a density of 1150 kg/m3 (dry density 238 kg/m3).
5.11.2.2 Clay Block Seals

The dense clay consists of precompacted blocks of smectite clay fitted in super-
containers moved down into the clay mud. The containers are of the same type
in the upper part of the holes and in the deployment part in which they contain
both HLW canisters surrounded by clay and dense clay blocks. The blocks are
156 Bentonite Clay
prepared by compaction of granules of smectite-rich clay with a dry density of

1700–1900 kg/m3.
The type of clay for placement down to 2000 m depth, where the hydrothermal
conditions are less severe than deeper down (T > 60°C), is preferably highly expan-
sive montmorillonite. Below this depth, the very high saltwater and high temperature
(150°C) may require use of chemically more stable clays such as saponite or very
dense smectite/illite mixed layer clay minerals.
The about 8–12 m long supercontainers filled with dense clay weigh at least
10 tons in air and sink in the mud because its bearing capacity is much lower than
the pressure exerted by the supercontainers. The surface temperature of the contain-
ers is estimated to be near 100°C in this place, which lowers the viscosity of the mud
and facilitates their insertion.
5.11.2.3 Interaction of Mud and Blocks

The maturation rate of the dense clay in the containers and the consolidation of
the mud generated by its expansion can be too rapid to get them down. The dense
clay has an initial dry density of 1700–1900 kg/m3 with a proposed degree of water
saturation of 50%–60%, implying a very high suction potential (Johannesson et al.
1995). The dense clay sorbs water from the mud, which, at the same time, becomes
consolidated by dense clay moving out through the perforation of the containers.
The process is relatively quick and stiffening of the mud can be significant already
in 4–8 hours, which means that the deepest supercontainer must be down in no
more than 8 hours. Any delay means that the resistance to bringing it down after
this time will require loading, which can be done by use of the drill rig by which the
supercontainer is installed. The required force can be hundreds of kilonewtons after
a couple of days.
Since some unforeseen delay in the operation must be expected, a process or
technique for retarding the maturation of the clay seals will be asked for. It is in
fact provided by the mud since the flow rate of mud porewater to the expanding
blocks is low and successively drops in the course of the consolidation that it
undergoes.
A method for further retardation of the wetting of the dense clay blocks is to coat
the supercontainer with a thin paste of mixed smectite clay and talc (Pusch 2011).
The coating retards migration of water into the dense clay blocks and delays expan-
sion. Another possibility of delaying consolidation of the mud is to use clay blocks
with an initially high degree of water saturation since the suction potential drops
significantly with the degree of water saturation (Figure 5.65).
5.11.3 Usefulness of Rock for Hosting Repositories

We have seen that rock with very low hydraulic conductivity can give such slow wet-
ting of the buffer clay in KBS-3V/i-type repositories that it becomes desiccated for
many decades with concomitant difficulties to become fully hydrated when water
is ultimately available. The problem is to select positions of deposition tunnels and
holes that provide neither too much nor too little water and the results of such optimi-
zation is illustrated in Figure 5.66. Considerable part of the selected host rock have
FIGURE 5.65 Growth of soft clay through the perforation of a copper tube confining a
dense clay core in an 80 mm diameter oedometer with 4 mm space between the oedometer
ring and the copper tube. Appearance at removal of the covering lid 8 hours after start. The
larger part of the clay core is still unaffected by water and will be so even after weeks except
if the water pressure is very high (Pusch, 2008).
to be left out depending on the criteria adopted concerning the respect distances and
necessary distance from the transport tunnels to the deposition tunnels for construct-
ing concrete bulkheads in the latter. We have seen that the fraction of useful rock
volume is lower per deposition tunnel than for the VDH and KBS-3H concepts, and
the need for reserving space for the connections of the two types of tunnels in KBS-
3V/i-type repositories reduces this fraction further.
5.12 CONCLUDING REMARKS

Chapter 5 is comprehensive and readers may require some assistance in finding the
essence of it. Here comes the help:
• Smectite clay serves as a very effective sealing component in all the consid-
ered concepts by retarding groundwater flow around and along the canisters
and by providing a homogeneous ductile embedment of the canisters. For
KBS-3V, the clay surrounding the waste containers is proposed by SKB
to be a smectite-rich clay consisting primarily of montmorillonite. For
the VDH concept, there are two clay components of the same type as for
KBS-3V but placed with different densities, that is, the mud and the com-
pacted blocks of buffer clay.
• The physical and chemical evolution of the buffer clay in KBS-3V is dif-
ferent from that of the mud and dense clay in VDH. For the first-mentioned
repository concept, the water pressure will be relatively low and the ground-
water flow to the deposition holes slow, which can delay water saturation
by tens to hundreds of years. During this period, the desiccated part of the
buffer clay is exposed to at least 100°C and salt enrichment, dissolution,
and precipitation processes can reduce the expandability and self-healing
158
D3 Fracture zone
D3 Fracture zone
N
Deposition hole
30 m
Permplug
with clay in key
Permplug
D3 Fracture
without clay
zone
Templug
Quartz sand
Tunnfill
Transport tunnel
30–40 m
50–100 m Wide EDZ
Normal EDZ
D3 Fracture zone
Deposition tunnels
FIGURE 5.66 Schematic plan view of a KBS-3V repository with a transport tunnel connected to deposition tunnels in rock with minor fracture zones
that have to be accepted in the repository but not allowed to intersect deposition holes (small circles). Tight plugs are keyed into the rock for isolating
the waste-containing parts of the deposition tunnels from fracture zones and from transport tunnels. (Data from Pusch, R., Yong, R.N., Microstructure
Bentonite Clay
of Smectite Clays and Engineering Properties, Taylor & Francis, New York, 2006.)
ability of the ultimately water-saturated clay, and also make it stiffer and
less able to self-seal.
• In all parts of VDH, the clay blocks are quickly water saturated under the
prevailing high water pressure. The ultimate density of the clay blocks in
both KBS-3V and VDH repositories will be 1900 to 2000 kg/m3, which
gives the clay a hydraulic conductivity of no more than E−11 m/s at satura-
tion with Ca-dominated water, and a swelling pressure of at least 1 MPa.
Over the larger part of the length of the supercontainers in a VDH contain-
ing HLW canisters, the density of the mud will be low but the clay will still
not be very permeable.
• In the upper, sealed part of the VDH, the temperature will range from about
15°C to 60°C, causing no heat-induced changes and providing total tight-
ness directly after installation. In the deployment part, which will be heated
up to 150°C, mineralogical changes and less good isolation are expected.
The VDH concept discussed here implies that the boring mud is the
same as the deployment mud. For serving in the boring phase, it must
have the required ability to support the hole being bored and to bring up
the debris from the boring head. The supercontainers with waste canis-
ters and dense blocks of expansive clay shall sink down or be pushed
through the mud to the predetermined levels, which requires that its shear
resistance is sufficiently low. This can be done using equipment such
as PG-Macs* for softening the thixotropic mud. After installation of the
supercontainers, the mud will be consolidated by the expanding dense
blocks, which intermittently seal off the holes in the deployment zone
and completely tighten the holes in the upper sealed part. The performance
of the mud in the two parts will be largely different because of the dif-
ference in temperature and salt conditions. For the deployment part, the
main criterion is that the consolidated mud should stay in contact with the
rock and canisters for minimizing convective water flow. For the upper
sealed part without waste, it is required that the expanded clay stays less
permeable than the surrounding rock, which is conservatively taken here
as E−11 m/s. This criterion is valid and fulfilled for both the VDH and
KBS-3V concepts.
• Prediction of the hydration rate and the evolution of the swelling pressure
in the buffer is very uncertain, the reason being that the processes are tran-
sient and associated with chemical alteration.
Finally, Table 5.7 is a guide to the reader who likes to see typical and available geo-
technical properties of buffers and backfills! M is montmorillonite-dominated clay,
S is saponite-dominated clay, Mx is mixed-layer clay (FIM).
* Hydraulically driven agitators with mud flow capacities of more than 10,000 m3/h.
160 Bentonite Clay
TABLE 5.7
Typical Physical Data for Clays Useful for Sealing Purposes.
Dry Density at
Content of Density Water Saturation
Clay Type Expandables (kg/m3) (kg/m3) Kw/Ks psw/pss (kPa)
M 15–25 950 1600 5E−9/E−8 10/0
M 15–25 1200 1750 E−10/E−9 20/0
M 15–25 1650 2050 5E−11/5E−10 170/10
M 15–25 1750 2100 E−11/5E−11 300/30
M Iraq 40–60 950 1600 E−8/7E−8 0/0
M Iraq 40–60 1200 1750 5E−10/E−9 250/50
M 40–60 1650 2050 5E−12/E−11 1,200/500
M 40–60 1750 2100 E−12/5E−12 1,500/1,000
M 70–90 950 1600 5E−11/5E−10 400/50
M 70–90 1200 1750 E−11/5E−10 1,000/100
M 70–90 1650 2050 E−13/5E−13 5,000/3,000
M 70–90 1750 2100 E−13/5E−13 10,000/10,000
S 50–60 1275 1800 E−12/5E−12 2,500/2,000
S 50–60 1600 2000 5E−13/E−12 8,800/5,000
Mx 30–40 950 1600 5E−9/5E−8 50/0
Mx 30–40 1200 1750 3E−10/3E−9 50/0
Mx 30–40 1650 2050 3E−12/E−11 1,200/800
Mx 30–40 1750 2100 E−12/5E−12 1,500/1,200
Kaolinite 0 1750 2100 2E−10/2E−10 100/100
Palygorskite 50–60 1250 1650 E−12/4E−12 6,500/5,700
In Kw/Ks, Kw—distilled water; Ks—3.5% CaCl2 and in psw/pss, psw—distilled water; pss—3.5% CaCl2.
M—montmorillonite-dominated clay;S—saponite-dominated clay; Mx— mixed-layer clay (FIM).
REFERENCES
Al-Taie, L., Al-Ansari, N., Knutsson, S., Pusch, R. 2013a. Hazardous wastes problems in Iraq:
A suggestion for an environmental solution. Journal of Earth Sciences and Geotechnical
Engineering, 3:81–91.
Al-Taie, L., Pusch, R., Al-Ansari, N., Knutsson, S. 2013b. Hydraulic properties of smectite
clays from Iraq with special respect to landfills of DU-contaminated waste. Journal of
Earth Sciences and Geotechnical Engineering, 3:109–125.
Brady, P.V., Arnold, B.W., Freeze, G.A., Swift, P.N., Bauer, S.J., Kenney, J.L., Rechard, R.P.,
Stein, J.S.; Sandia National Laboratories 2009. Deep Borehole Disposal of High-Level
Radioactive Waste. US Department of Energy, Washington, DC.
Börgesson, L., Fredrikson, A., Johannesson, L.-E. 1994. Heat Conductivity of Buffer Materials.
SKB Teknisk Rapport TR 94-29, Swedish Nuclear Fuel and Waste Management (SKB),
Stockholm, Sweden.
Börgesson, L., Johannesson, L.-E., Fredrikson, A. 1993. Laboratory Investigation of Highly
Compacted Bentonite Blocks for Buffer Material. Compaction Technique and Material
Composition. SKB Djupförvar Projekt Rapport PR 44-93-009. Swedish Nuclear Fuel
and Waste Management (SKB), Stockholm, Sweden.
Börgesson, L., Johannesson, L.-E., Sandén, T., Hernelind, J. 1995. Modelling of the Physical
Behaviour of Water Saturated Clay Barriers. Laboratory Tests, Material Models and
Finite Element Application. SKB Technical Report 95-20, Swedish Nuclear Fuel and
Waste Management (SKB), Stockholm, Sweden.
Börgesson, L., Pusch, R., Fredriksson, A., Hokmark, H., Karnland, O., and Sandén, T.
1991. Final Report of the Rock Sealing Project—Sealing of the Nearfield Rock around
Deposition Holes by Use of Bentonite Grouts. Stripa Project technical Report TR 91-34,
Swedish Nuclear Fuel and Waste Management (SKB), Stockholm, Sweden.
Couture, R.A. 1985. Steam rapidly reduces the swelling capacity of bentonite. Nature,
318(6041):50–52.
Daniel, D.E. 1994. State-of-the-art: Laboratory hydraulic conductivity tests for saturated soils. In
Hydraulic Conductivity and Waste Contaminant Transport in Soil, Daniel, D.E., Trautwein,
S.J. (Eds.), American Society for Testing and Materials, Philadelphia, PA, pp. 30–78.
Eyring, H. 1936. Viscosity, plasticity and diffusion as examples of absolute reaction rates.
Journal of Chemical Physics, 4:283.
Farouki, O.T. 1986. Thermal Properties of Soils. Trans Tech Publications, Clausthal-Zellerfeld,
Germany.
Feltham, P. 1979. The Inter-Relationship: Earth Science-Material Sciences. Mechanisms of
deformation and fracture. Pergamon Press, London, pp. 29,42.
Forslind, E., Jacobsson, A. 1972. Clay-water systems. In Water: A Comprehensive Treatise,
Franks, G. (Ed.), Plenum Press, New York.
Frenkel, J. 1946. Kinetic Theory of Liquids. Clarendon Press, Oxford.
Graham, J., Oswell, J.M., Gray, M.N. 1992. The effective stress concept in saturated sand-clay
buffer. Canadian Geotechnical Journal, 29:1033–1043.
Grim, R.E., Güven, N. 1978. Developments in Sedimentology, Vol. 24: Bentonites. Elsevier,
New York.
Hellström, J.G.I. 2009. Internal erosion in embankment dams, fluid flow through and defor-
mation of porous media. PhD Thesis, Luleå University of Technology, Luleå, Sweden.
Hofmann, U., Endell, K., Wilm, D. 1933. Kristallstruktur und Quellung von Montmorillonit.
Zeitung fuer Kristallographie, 86:340–348.
Johannesson, L.-E., Börgesson, L., Sandén, T. 1995. Compaction of Bentonite Blocks.
SKB Technical Report 95-19, Swedish Nuclear Fuel and Waste Management (SKB),
Stockholm, Sweden.
Kahr, G., Müller-Von Moos, M. 1982. Wärmeleitfähigkeit von Bentonit MX80 und von Montigel
nach der Heizdrahtmethode. Nagra Technischer Bericht 82-06, Wettingen, Switzerland.
Kehres, A. 1983. Dehydration energy isotherms of clays. Diagrams for surfaces of interlamel-
lar and extralamellar (microvoids) space. PhD Thesis, Université Paul Sabatier, Tolouse,
France.
Knutsson, S. 1983. On the Thermal Conductivity and Thermal Diffusivity of Highly Compacted
Bentonite. SKB Technical Report 83-72, Swedish Nuclear Fuel and Waste Management
(SKB), Stockholm, Sweden.
Mooney, R.W., Keenan, A.G., Wood, L.A. 1987. Adsorption of water vapor by montmorillon-
ite. II. Effect of exchangeable ions and lattice swelling as measured by X-ray diffraction.
Journal of the American Chemical Society, 74:1371–1374.
Nagy Noémi M., Kónya, J. 2010. Interfacial Chemistry of Rocks and Soils. CRC Press, Boca
Raton, FL.
Popov, V., Pusch, R., (Eds.), 2006 Disposal of Hazardous Waste in Underground Mines. WIT
Press, Southampton, UK.
Pusch, R. 1967. A technique for investigation of clay microstructure. Journal of Microscopy,
6:963–986.
Pusch, R. 1970. Clay Microstructure. Swedish National Building Research Council, Doc. D:8,
Stockholm, Sweden.
162 Bentonite Clay
Pusch, R. 1979. Creep of Soils. Ruhr-Universität Bochum, Bochum, Germany.

Pusch, R. 1987a. Identification of Na-smectite hydration by use of “humid cell” high voltage
microscopy. Applied Clay Science, 2:343–352.
Pusch, R. 1987b. Permanent crystal lattice contraction—A primary mechanism in thermally induced
alteration of Na bentonite. In Scientific Basis for Nuclear Waste Management X: Proceedings
of the Materials Research Society Symposium, Vol. 84, Bates, J.K., Seefeldt, W.B. (Eds.).
Pusch, R. 1993. Evolution of models for conversion of smectite to non-expandable minerals.
SKB Technical Report TR-93-33, SKB, Stockholm, Sweden.
Pusch, R. 1994. Waste Disposal in Rock, Developments in Geotechnical Engineering 76.
Elsevier, New York.
Pusch, R. 1995. Rock Mechanics on a Geological Base. Developments in Geotechnical
Engineering 77. Elsevier, New York.
Pusch, R. 1999a. Clay Colloid Formation and Release from MX-80 Buffer. SKB Technical Report
TR-99-31. Swedish Nuclear Fuel and Waste Management (SKB), Stockholm, Sweden.
Pusch, R. 1999b. Experience from preparation and investigation of clay for microstructural
Pusch, R. 1999c. Mobility and Survival of Sulphate-Reducing Bacteria in Compacted and
Fully Water Saturated Bentonite—Microstructural Aspects. SKB Technical Report
Pusch, R. 2000. On the Risk of Liquefaction of Buffer and Backfill. SKB Technical Report
Pusch, R. 2001a. Can the Water Content of Highly Compacted Bentonite Be Increased by
Applying a High Water Pressure? SKB Technical Report TR-01-33. Swedish Nuclear
Fuel and Waste Management (SKB), Stockholm, Sweden.
Pusch, R. 2001b. Experimental Study of the Effect of High Porewater Salinity on the Physical
Properties of a Natural Smectitic Clay. SKB Technical Report TR-01-07. Swedish
Nuclear Fuel and Waste Management (SKB), Stockholm, Sweden.
Pusch, R. 2001c. The Microstructure of MX-80 Clay with respect to Its Bulk Physical Properties
under Different Environmental Conditions. SKB Technical Report TR-01-08, Swedish
Pusch, R. 2008. Geological Storage of Radioactive Waste. Springer-Verlag, Berlin, Germany.
Pusch, R. 2011. A technique to delay hydration and maturation of borehole seals of expansive
clay. Engineering Geology, 121:1–6.
Pusch, R., Adey, R. 1999. Creep in Buffer Clay. SKB Technical Report TR-99-32. Swedish
Pusch, R., Bluemling, P., Johnson, L. 2003. Performance of strongly compressed MX-80 pel-
lets under repository-like conditions. Applied Clay Science 23:239–44.
Pusch, R., Börgesson, L. 1992. Performance assessment of bentonite clay barrier in three
repository concepts: VDH, KBS3 and VLH. PASS—Project on Alternative Systems
Study, SKB Technical Report TR 92-40. Swedish Nuclear Fuel and Waste Management
(SKB), Stockholm, Sweden.
Pusch, R., Börgesson, L., Ramqvist, G. 1985. Final Report of the Buffer Mass Test—Volume
II: Test Results. Stripa Project, Technical Report TR-85-12, Swedish Nuclear Fuel and
Waste Management (SKB), Stockholm, Sweden.
Pusch, R., Erlström, M., Börgesson, L. 1987a. Piping and erosion phenomena in soft clay gels.
SKB Technical Report TR 87-09, Swedish Nuclear Fuel and Waste Management (SKB),
Stockholm, Sweden.
Pusch, R., Feltham, P. 1980. A stochastic model of the creep of soils. Geotechnique, 30(4):497–506.
Pusch, R., Hatem, M., Knutsson, S.; Luleå University of Technology 2014a. VDH—A case of
ostrich philosophy or a serious alternative way of disposal of highly radioactive waste?
In Waste Management and the Environment VII, Brebbia, C.A., Passerini, G., Itoh, H.,
(Eds.), WIT Press, Southampton, pp. 377–390.
Pusch, R., Hedin, G., Grahn, P.H. 2014b. Principles of location and design of landfills of
low-level radioactive waste as exemplified by the Lituanian NSR. In Proceedings of the
Waste Management Conference, Pusch, R., Knutsson, S. (Eds.), Luleå, Sweden.
Pusch, R., Hökmark, H., and Börgesson, L. 1987b. Outline of Models and Water Gas Flow
through Smectite Clay Buffers. SKB Technical Report TR 87-10, Swedish Nuclear Fuel
and Waste Management (SKB), Stockholm, Sweden.
Pusch, R., Karnland, O., Hökmark, H. 1990. GMM—A general microstructural model for
qualitative and quantitative studies of smectite clays. SKB Technical Report TR-90-43,
SKB Swedish Nuclear Fuel and Waste Co (SKB), Stockholm, Sweden.
Pusch, R., Moreno, L., Neretnieks, I. 2001. Microstructural modelling of transport in
smectite clay buffer. In Clay Science for Engineering, Adachi, K., Fukue, M. (Eds.),
A.A. Balkema Publishers, Rotterdam, the Netherlands, pp. 47–51.
Pusch, R., Muurinen, A., Lehikoinen, J., Bors, J., Eriksen, T. 1999. Microstructural and
Chemical Parameters of Bentonite as Determinants of Waste Isolation Efficiency. Final
Report Contract No F14W-CT95-0012, European Commission, Luxembourg.
Pusch, R., Popov, V., Adey, R., Kasbohm, J., Knutsson, S. 2014c. Rational and economic
disposal of hazardous waste—use of abandoned mines. Journal of Earth Science and
Geotechnical Engineering, 4(2): 33–55.
Pusch, R., Ramqvist, G., Kasbohm, J., Knutsson, S., Mohammed M.H. 2012a. The concept of
highly radioactive waste (HLW) disposal in very deep boreholes in a new perspective.
Journal of Earth Sciences and Geotechnical Engineering, 2(3):1–24.
Pusch, R., Weston, R. 2003. Microstructural stability controls the hydraulic conductivity of
smectite buffer clay. Applied Clay Science, 23:35–41.
Pusch, R., Yong, R.N. 2006. Microstructure of Smectite Clays and Engineering Properties.
Pusch, R., Yong, R.N., Nakano, M. 2012b. High-Level Radioactive Waste Disposal. WIT
Press, Southampton.
Pusch, R. et al. 1995. The Buffer and Backfill Handbook, Part I. SKB Technical Report 95-45,
Schroeder, P.R., Gibson, A.C., Smolen, M.D. 1984. The hydrologic evaluation of landfill perfor-
mance (HELP) model: Volume II. Documentation for Version 1. EPA/530-SW-84–010,
Municipal Environmental Research Laboratory, Cincinnati, OH.
Shirazia, S.M., Kazama, H., Kuwano, J., Rashid, M.M. 2010. The influence of temperature on
swelling characteristics of compacted bentonite for waste disposal. Environment Asia,
3:60–64.
Singh, A., Mitchell, J.K. 1968. General Stress-Strain-Time Functions for Soils. University of
California, Berkeley, CA
SKB, 2012. KBS-3H Complementary Studies, 2008–2010. SKB Technical Report TR-12-01,
Sposito, G. 1984. The Surface Chemistry of Soils, Oxford University Press, New York,
Clarendon, Oxford.
Sposito, G., Prost, R. 1982. Structure of water adsorbed on smectites. Chemical Reviews,
82(6):553–573.
Svemar, C. 2005. Cluster Repository Project (CROP). Final Report of European Commission
Contract FIR1-CT-2000-20023, Brussels, Belgium.
Taylor, D.W. 1948. Fundamentals of Soil Mechanics. Wiley, New York.
Terzaghi, K., Peck, R.B., Mesri, G. 1996. Soil Mechanics in Engineering Practice, 3rd edn.
Wiley, New York.
Yong, R.N., Pusch, R., Nakano, M. 2010. Containment of High-level Radioactive and
Hazardous Solid Wastes with Clay Barriers. Spon Press, London.
6 Chemical Processes
Involved in and Longevity
of Smectite Buffer
This chapter deals with changes in constitution on the microscale and macroscales
of smectite clays by processes in nature or by anthropological impact. It deals in
particular with the longevity of smectite clays in bentonite beds and in the form of
canister-embedding buffers in repositories for deep geological high-level radioac-
tive waste (HLW) disposal. The basis is the concept of fundamental clay crystal
units described in Chapter 3, the clay–fluid interaction touched upon in Chapter 4,
and the microstructural models discussed in Chapter 5. Most of the message is
inherited from decades of research by my own staff but a number of new findings
of international experts from Denmark, Germany, and Vietnam have added to the
present knowledge.
6.1 CHEMICAL STABILITY OF SMECTITE

CLAY FOR WASTE ISOLATION
6.1.1 Our Starting Point
How can one rely on smectite clay seals over time? What are the threats? These are
questions that need to be answered in order to find practical and safe ways of isolat-
ing hazardous waste. The answer to the last question is temperature, temperature
gradients, softness, porewater chemistry, and time. We will consider in this chapter
the basis for the long-term performance and survival of smectite clay seals; we start
by defining the initial physicochemical condition of man-made clay seals as a matrix
of stochastically distributed stacks of smectite lamellae with voids forming through-
going microchannels that are partly open and partly filled with soft clay gels. This
constitution remains for any period of time and only alters with respect to the void size
distribution if the porewater electrolytes and density change significantly: the fraction
of interlamellar water increases on the expense of extralamellar water if percolation
of freshwater takes place and if the clay undergoes consolidation. Microstructural
changes are largely controlled by the balance of swelling pressure and external effec-
tive pressure. The physical stability is substantial and the clay retains its properties
over long periods of time if the temperature is below about 50°C–60°C. Since we will
be concerned with hot conditions when considering the function of buffer clay, we
need to consider the effect of temperatures higher than this and of temperature gradi-
ents that can be several degrees Celsius per centimeter length of the buffers.
165
166 Bentonite Clay
VDH—vertical hole
SARC—horizontal drift
at 400–500 m depth
Clay/concrete
seals, 2 km
>15 m Clay
Containers with
concrete-embedded
Deployment waste
zone with HLW,
2 km
FIGURE 6.1 Schematic drawing of the concepts VDH and SARC.
Cases of particular interest to the planners and designers of underground repositories for
highly radioactive waste are represented by what we will call here the sarcophagus (SARC),
and very deep hole (VDH) concepts (cf. Brady et al. 2009), which will be in focus in
this chapter. The reason for selecting these two is that very deep HLW disposal is pres-
ently being reexamined internationally and estimated to be relatively cheap, and that
SARC provides effective isolation at even lower cost by use of iron in the canisters and
concrete in confining containers (the poor man’s solution). Before dealing in detail
with them, we should briefly consider some natural analogs that will help us to identify
the major degrading mechanisms in smectite clay exposed to hydrothermal conditions.
Figure 6.1 shows the two concepts reappearing from Chapters 4 and 5. The major dif-
ference between the two is that VDH components are placed under water, whereas the
SARC is constructed at room atmosphere in mined tunnels.
6.1.2 Natural Analogs

6.1.2.1 Conversion of Smectite to Nonexpanding Minerals
The long-term chemical stability of smectite clay, including montmorillonite, has
been investigated by mineralogists since the 1950s and has been the focus of inves-
tigators and designers of repositories for disposal of HLW in Sweden and various
other countries over the last 40–50 years (Pusch et al. 1987; Svemar 2005).
Comprehensive studies of large numbers of samples of the Tertiary age from the
North Sea basin buried at large depths reached the conclusion that smectite in layers
from 1500 to 1700 m below the present sea bottom had undergone only insignificant
transformation to less expanding illite/smectite (I/S) since sedimentation (Nadeau
and Bain 1986; Roaldset 1995). This is in contrast with clay from more than 2000 m
depth, which has undergone significant conversion to I/S mixed-layer minerals.
Chemical Processes Involved in and Longevity of Smectite Buffer 167
The critical temperature for alteration seems to be 65°C–75°C, which has prevailed

for several tens of millions of years at depth. These investigators found that even for
the layers with no illitization, cementation by precipitated matter, mostly of silicious
type, had turned the originally ductile clay into claystone.
Referring to true bentonites, there are numerous examples in the classical lit-
erature of altered clay deposits and attempts to explain the processes leading to
alteration by investigators such as Grim and Guven (1978), Eberl and Hower (1976),
Hoffman and Hower (1979), Pytte and Reynolds (1989), and Grindrod et al. (1999).
All the studies support the general formulation:
S + ( Fk + Mi ) = I + Q + Chl (6.1)

where:
S represents smectite
Fk represents K-feldspar
Mi represents K-mica
I represents illite
Q represents quartz
Chl represents chlorite
The required potassium can emanate from degraded feldspars and micas, released
from various high-temperature reactions, or from the groundwater.
Quartz has been assumed to originate from decomposed feldspars and from sili-
cons lost from the SiO4 tetrahedrons in montmorillonite in conjunction with conver-
sion of this mineral to beidellite as a first step in conversion to illite.
Hoffman and Hower introduced a metamorphic grade index and suggested a scheme
of seven grades with temperature dependence, implying complete conversion from
smectite to illite at 60°C over several hundreds of millions of years, at 100°C in about
100 million years, and at 140°C in 1 million years. Similar e stimates led to the Pytte–
Reynold’s conversion model that is presently favored by some Temperature/Hydraulic/
Mechanical/Chemical (THMC) modelers; it is here formulated by Equation 6.2.
dS ⎡⎛ K + ⎞ m ⎤
− = ⎡⎣ Ae −U / RT ( t ) ⎤⎦ ⎢⎜ + ⎟ S n ⎥ (6.2)
dt ⎢⎝ Na ⎠ ⎥
⎣ ⎦
where:
S is the mole fraction of smectite in I/S assemblages
U is the activation energy
R is the universal gas constant
T is the absolute temperature
t is the time
m and n are coefficients
A is a constant
Use of this equation gives the time of illitization for the activation energy of 27 kcal/
mole for rich access to potassium as a function of temperature: only insignificant
168 Bentonite Clay
loss of smectite will take place in E6 years at 50°C, whereas about 100% of the
original smectite will turn into illite in this period of time at 100°C. According to
this model, a typical buffer temperature history for canisters with HLW in the form
of spent fuel is as follows: 50°C–150°C in the first hundred years followed by a
500-year period with an average temperature of 25°C–100°C, and less than 50°C in
the subsequent 1000 years, which would convert 15% of the initial smectite content
to illite in 1500 years.
The problem is that the activation energy is not known with certainty and that
access to potassium is difficult to define and derive. For an activation energy of
around 20 kcal/mole, which may well be applicable, the model would imply very
significant degradation of the buffer by conversion to illite in 100 years.
6.1.2.2 Kinnekulle—A Comforting Case?

The so-called meta-bentonites at Kinnekulle in southwestern Sweden are exam-
ples of more stable smectites than implied by the models described. They are of
Ordovician age and make up a series of up to 2 m thick beds (Figure 6.2) that have
been exposed to at least 30 MPa pressure in the latest glaciation period. The present
content of montmorillonite in expandable form in the bentonite bed is about 25%,
the rest being illite, quartz, feldspars, chlorite, and carbonates. Analyses of samples
taken from the bed gave the water content values (w) in this figure, indicating that the
clay has expanded from a denser state reached during the maximum ice load some
tens of thousands of years ago (Pusch and Madsen 1995).
The temperature, derived from conodont analysis and calculation of the tempera-
ture evolution after intrusion of Permian basalt, rose from 15°C–20°C to 120°C in
100 years, rose further to 140°C after another 100 years, then dropped to 90°C after
1000 years, and rose to the initial temperature 3000 years after the basalt intrusion.
The average temperature gradient was 0.02°C/cm in the first few hundred years.
Application of the models would give around 25% remaining smectite fraction taking
the activation energy to be 25 kcal/mole—a common assumption—provided that the
potassium content was constantly at least 70 ppm. The present amount of smectite
fits with this estimate but the assumption of constant K-concentration is question-
able considering the fact that access to potassium had to be supplied by diffusive
migration through the huge mass of overlying marine sediments. One has in fact to
assume that comprehensive heat-aided groundwater convection took place bringing
new K-bearing water down to the bentonite.
6.2 EXPERIMENTAL EVIDENCE

6.2.1 Overview
The evolution of smectite clays in saltwater at medium and high temperatures
for use as waste isolation has been frequently discussed in the literature (Weaver
1979; Grindrod and Takase 1993; Kasbohm et al. 2000; Pusch and Kasbohm 2002;
Herbert et al. 2004), indicating that considerable degradation can take place. We
will go through some of the investigations for making ourselves acquainted with the
employed lab techniques and common conclusions.
• Clay formed from volcanish ash (Na form)

Permian Diabase, 30 m • Shallow location (1–7 m below ground surface) at
deglaciation implying strong water flow in the series
• Dry density of clay 1600 kg/m3 (2000 kg/m3 at sat.)
Silurian Shales and • Homogeneous nature

sandstones, 90 m
Present state:
Silicified limestone, 5–10 m 0 ground surface
Bentonite, 2 m
Silicified limestone, 5–10 m
Ordovician Limestone
Glauconite 5m w% wL% S%
23 70 20–25
Dark shales 2 m bentonite bed 28 75 25–30
35 69 20–25
Sandstone 27
32
10 m 27 75
Precambrian Crystalline
160
140
Temperature increase over
ambient (centigrades)
120
100
80
60
40
20
0
0 500 1000 1500
Time after diabase outflow (years)
FIGURE 6.2 The Kinnekulle stratigraphy and temperature history. The big graph describes the
geotechnical/mineralogical data evaluated from analysis of 10 m cores taken from the present
ground surface. wL denotes the Atterberg liquid limit, which is a measure of the smectite content
(S) that is 20%–30% of the mineral mass in the thick bentonite bed. The right graph shows the
evaluated temperature history. (Data from Pusch, R., Stability of Deep-Sited Smectite Minerals in
Crystalline Rock—Chemical Aspects, SKBF/KBS Technical Report TR 83-16, SKB, Stockholm,
Sweden, 1983; Pusch, R., Madsen, F., Clays Clay Miner., 43, 261–270, 1995.)
6.2.2 Stripa Project Laboratory Study

The international Stripa Project described in Chapter 5, comprised, in addition to
the extensive and pioneering field experiments, a systematic study of the chemical
evolution and associated change in rheological behavior of hydrothermally treated
montmorillonite-rich clay (MX-80) (Pusch 1985; Pusch et al. 1985). The density
170 Bentonite Clay
S4-11 MX-80 1300 kgm−3

2.9% IV Freeze-dried + bm-methacr. + catal.
38.7% III 40.2% I Picture width 7 μm
Empty void diameter <3 μm
P/Tw = 2.9%, P/Tw + g = 41.6%
18.2% II
FIGURE 6.3 (See color insert.) Transmission electron micrograph of 300Å–500Å ultrami-
crotomed section of MX-80 clay with 1300 kg/m3density (dry density 480 kg/m3). The digitally
treated version of the left figure gave the picture to the right showing the density variations:
white and green represent open space, red is very dilute clay gel, and black is dense aggregates
of smectite stacks. P/Tw is the ratio of the sectioned area of wholly open voids (white) and the
total sectioned area. P/Tw+g is the ratio of the area of voids with very dilute clay gels and the
total sectioned area. (Data from Pusch, R., Eng. Geol., 54, 187–195, 1999.)
was low, that is, of the kind used for grouting of rock fractures, since the tests
would have to be running for years to reach equilibrium had higher densities been
used. Figure 6.3 shows a typical transmission electron image of MX-80 clay with
1300 kg/m3, showing the continuous network of smectite aggregates and the very
large pore space. Coloring represents different estimated densities. P/T is the two-
dimensional porosity, which can be related to the hydraulic conductivity (Pusch
and Yong 2006).
6.2.2.1 Test Program and Techniques

The study included autoclave experiments with MX-80 clay confined in perme-
able reaction chambers (Pusch 1985; Gray 1993). The aim was to find out how
Piston Solution
Force
Pressure Connection
Autoclave cell chambers of for sampling
placed in oven teflon, filled and refilling
with clay
FIGURE 6.4 Autoclave cell used for the hydrothermal investigations.
montmorillonite-rich clay, saturated with distilled water, 3.5% CaCl2 solution, and
ocean water diluted to a concentration of 1.75 g dissolved salt/l, would react on expo-
sure to temperatures up to 200°C under constant volume conditions. The interesting
issues were (1) dissolution and loss of major crystal components, (2) pH of the pore-
water, and (3) undrained shear strength of the heated clays. The fluid in the autoclave
vessel was distilled water, as just described (Pusch et al. 1985; Gray 1993).
Hydrothermal experiments were made with clay having a dry density of 475 kg/m3
(1300 kg/m3 at fluid saturation) and contained in 6 ml permeable reaction chambers
placed in autoclaves filled with the respective solutions, which were regularly sam-
pled (Figure 6.4). The solutions were under 2 MPa pressure for avoiding boiling and
the cells left for 10, 90, and 270 days heated up to 200°C. The hydraulic conductivity,
pH, and shear strength determined by a laboratory vane bore, were measured after
each of the 10-, 90-, and 270-day periods. The concentration of Si, Al, and Mg in the
sampled solutions was determined as well.
6.2.2.2 Summary of Results

• Hydraulic conductivity: The 10-day samples of salt clays had a hydrau-
lic conductivity that rose from about E−7 to E−6 m/s (hydraulic gradient
33 m/m) when the temperature was increased from 20°C–200°C but was
constantly about 2E−9 m/s for the clay prepared with distilled water in the
same temperature interval. The 270-day samples of salt clays had a slightly
increased conductivity and for the clay prepared with distilled water, it had
increased to about 7E−9 m/s.
• Shear strength: The 10-day samples of salt clays had a shear strength that
increased from 1 kPa at 20°C to 5 kPa at 200°C. For the clay prepared with
distilled water, it increased from 1 to 2.5 kPa. For the 270-day samples
of the salt clays, the shear strength was 2 kPa at 20°C and about 5 kPa at
172 Bentonite Clay
130°C but dropped back to 4 kPa at 200°C. The 270-day sample of the
clay prepared with distilled water had a shear strength of about 2 kPa at
20°C reaching up to 4 kPa for 150°C dropping back to 2 kPa at 200°C.
The strength increased up to 90°C, which was attributed to aggregation
(lensing) and, for higher temperatures, to silicification (Müller-Vonmoos
et al. 1990).
• Solution chemistry—silica (cf. Figure 6.5): The 10-day samples of salt
clays gave off silica to the solutions. The Si content went up from 0 to
60–120 ppm when the temperature rose from 20°C to 200°C. For the 270-
day samples, the Si concentration rose from 0 to 50 ppm at 150°C and fur-
ther to about 120 ppm at 200°C. The data showed that, congruent with silica
release for 120°C–130°C, magnesium and potassium were taken up by the
clay from the salt solutions. No release from the sample prepared with dis-
tilled water had occurred.
The strong dissolution and diffusion of silica out from the clay to the sur-
rounding external water when the temperature was raised indicates that the
90 days
120
100
[Si] change (mg/l)
80
Distilled
60 FF
Sea/2
40
20
0
0 50 100 150 200 250
(a) Temperature (°C)
270 days
120
100
[Si] change (mg/l)
80
Distilled
60 FF
Sea/2
40
20
0
0 50 100 150 200 250
(b) Temperature (°C)
FIGURE 6.5 (a,b) Change in Si concentration by hydrothermal treatment of MX-80 clay.

(Data from Pusch, R. et al., Final Report of the Buffer Mass Test—Volume II: Test Results,
Stripa Project Technical Report 85-12, SKB, Stockholm, Sweden, 1985; Gray, M., OECD/
NEA International Stripa Project 1980–1992: Overview. Vol III, Engineered barriers, SKB,
Stockholm, Sweden, 1993.)
porewater and surrounding external water became saturated with Si well

below 100°C.
• Solution chemistry—aluminum (cf. Figure 6.6): The concentration of Al
was very low except for the 200°C test with distilled water in which the
clay underwent significant dissolution. Figure 6.6 shows that aluminum
was largely maintained in the clay, presumably forming various hydrated
complexes that served as cement (Pusch 2008). The difference between the
tests with three solutions was small. The retainment of Al agrees with the
conclusion from the mineralogical analyses that only little neoformation of
illite took place. The observed formation of illite supports the hypothesis of
Hoffman and Hower (1979) that, given access to potassium at temperatures
above about 120°C–130°C, illite can quickly be formed.
• pH (cf. Figure 6.7): pH of the solutions decreased with temperature in the
first 10 days and was then constant. The clay prepared with distilled water
was alkaline for temperatures lower than 130°C but became strongly acidic
at 200°C.
90 days
15
[Al] change (mg/l)
10
Distilled
FF sol.
Sea sol.
5
0
0 50 100 150 200 250
270 days
15
[Al] change (mg/l)
10
Distilled
FF sol.
Sea sol.
5
0
0 50 100 150 200 250
FIGURE 6.6 (a,b) Change in Al concentration at hydrothermal testing of MX-80. (Data

from Pusch, R. et al., Final Report of the Buffer Mass Test—Volume II: Test Results, Stripa
Project Technical Report 85-12, SKB, Stockholm, Sweden, 1985; Gray, M., OECD/NEA
International Stripa Project 1980–1992: Overview. Vol III, Engineered barriers, SKB,
Stockholm, Sweden, 1993.)
174 Bentonite Clay
90 days
12
10
8 Distilled
FF
pH
6 Sea/2
2
0 50 100 150 200 250
270 days
12
10
8 Distilled
FF
pH
6 Sea/2
2
0 50 100 150 200 250
FIGURE 6.7 (a,b) Change in pH at hydrothermal testing of MX-80. The difference between the
tests with different solutions became smaller with temperature and time. (Data from Pusch, R.
et al., Final Report of the Buffer Mass Test—Volume II: Test Results, Stripa Project Technical
Report 85-12, SKB, Stockholm, Sweden, 1985; Gray, M., OECD/NEA International Stripa
Project 1980–1992: Overview. Vol III, Engineered barriers, SKB, Stockholm, Sweden, 1993.)
The results from the Stripa laboratory study verify that the chemical composition of
the porewater in smectite-rich buffer clay is determined by the interaction of ground-
water and minerals in the host rock. Ion exchange processes play a major role and
also dissolution of smectites and accessory minerals such as carbonates, particularly
calcite, and sulfur-bearing minerals. A typical reaction is when Na montmorillonite
is dispersed in distilled water: Na is given off to the porewater and protons replace
them in the exchange positions—which raises pH of the porewater to as much as 10
(Figure 6.7). If the clay is dispersed in Na-rich water, Na is absorbed and protons are
released, which means that pH of the porewater drops to values that can be as low
as 3. Dispersion in water rich in Ca and Mg causes absorption of these cations and
release of initially absorbed Na; pH will thereby also be changed.
6.2.3 SKB-ANDRA Study

A joint SKB-ANDRA experiment in the early 1990s was made for investigating the
effect of gamma radiation on montmorillonite-rich clay (MX-80) with a dry density
of 1650 kg/m3 (Pusch et al. 1993). The sample was exposed to a radiation dose of
about 3E7 Gy at one end that was sealed by an iron plate, whereas at the opposite
end, the sample was confined by a filter through which it was in contact with weakly
brackish water with Na as dominant cation, contained in a large vessel. The solution
had a very low potassium content (<10 ppm). The radiated end was heated to 130°C
during the one-year experiment and the opposite end to 90°C (thermal gradient about
6°C/cm). The solution was pressurized to 1.5 MPa. The adsorbed radiation dose
was 3972 Gy/hour at the hottest contact, around 700 Gy/hour at half length of the
sample, and 456 Gy/hour at the coldest end. A parallel test of the same duration was
conducted without radiation.
The analyses comprised X-ray diffraction (XRD), electron microscopy with
energy dispersive X-ray (EDX), chemical analysis by atomic absorption spectros-
copy, infrared spectrometry, and determination of cation exchange capacity; the
main data are given in Table 6.1.
The clay samples were saturated beforehand with brackish Na-dominated water
(Total dissolved salt [TDS] = 10 ppm) and transferred to hydrothermal cells with an
iron plate at one end, which was kept at 130°C. The water was pressurized to 1.5 MPa
and circulated through a filter at the end held at 90°C temperature for one year.
Figure 6.8 indicates the changes in mineral composition and content that took place
in this experiment, which was followed by the same type of test with no irradiation.
Chemical analysis of the clay samples showed only minor deviations from the virgin
material. However, mineralogical changes were as shown by the diffractogram in
Figure 6.8 and Table 6.2. An important observation was that iron released from the
hot plate migrated quicker into the irradiated sample than in the ones that were not
exposed to gamma radiation. The impact of the hydrothermal treatment on the rheo-
logical properties of the clay was determined by conducting shear tests with stepwise
loading and recording of the compression. This gave clear evidence of stiffening
of the montmorillonite buffer clay: the shear strain of the 130°C sample was only
TABLE 6.1
Changes in One-Year Hydrothermal Tests of MX-80
Treatment 125°C–130°C 115°C–120°C 105°C–110°C 90°C–95°C
Hydrothermal M- M– M0 M0
without radiation F— - F— F- F-
Chl + Chl + Chl 0 Chl 0
G ++ G +++ G+ G+
K ++ K+ K0 K-
Q+ Q+ Q0 Q0
I+ I0 I0 I0
Source: Pusch, R. et al., Hydrothermal Field Test with French Candidate Clay Embedding Steel Heater
in the Stripa Mine, SKB Technical Report TR-93-02, SKB, Stockholm, Sweden, 1993.
M is montmorillonite; F is feldspars; G is gypsum; Q is quartz; K is kaolinite; Chl is chlorite; and I is illite.
+++ means strong increase; ++ means significant increase; + means slight increase; — - means strong
loss; — means significant loss; - means slight loss; 0 means no change.
176 Bentonite Clay
Gypsum monohydrocalcite
Hexahydrite kaolinite
Quartz mica
Hexahydrite
Amphiboles
Mica illite
Kaolinite
Smectite
Smectite
Feldspar
Feldspar
Feldspar
Chlorite
Quartz
20°C
130°C
0 30 25 20 15 10 5
2θ
FIGURE 6.8 Schematic diffractograms of the reference MX-80 sample (20°C) and the most
heated part of the hydrothermally tested sample (130°C). Feldspars, amphibole, some of the
quartz, and smectite had disappeared in the latter. (Data from Pusch, R. et al., Hydrothermal
Field Test with French Candidate Clay Embedding Steel Heater in the Stripa Mine, SKB
Technical Report TR-93-02, SKB, Stockholm, Sweden, 1993.)
TABLE 6.2
Properties of Oedometer-Tested Samples M1–M4
Sample Distance from Heater (cm) T (°C) ρsat (kg/m3) ps (kPa) K (m/s)
M1 16–18 45–47 1800 430 1.6E−11
M2 12–14 54–56 1945 650 1.9E−11
M3 6–8 67–69 1910 355 2.3E−10
M4 0–2 85–90 1925 310 2.1E−10
Source: Pusch, R., Geological Storage of Radioactive Waste, Springer-Verlag, Berlin, Germany, 2008.
ρsat is the density at water saturation; K is the hydraulic conductivity; and ps is the swelling pressure.
one-third to one-tenth of the strain of the 90°C sample for different shear stresses
(Figure 6.9). The reason was cementation by precipitated silicious compounds.
6.2.4 RMN Study
The appointment of a national buffer clay by the Czech organization RMN, cor-
responding to the Swedish Nuclear Fuel and Waste Management Company (SKB),
was paralleled with the conduction of a mock-up test at the Geotechnical Division of
the Czech Technical University at Prague around the latest century shift (Pacovsky
Shear strain in o/o after 27 hours
Shear stress
520 kPa
2
Shear stress
260 kPa
0
50 60 70 80 90 100 110 120 130 140
Temperature (°C)
FIGURE 6.9 Shear strain of MX-80 hydrothermally treated for one year. Shear box testing
under 6 MPa normal effective pressure. (Data from Pusch, R. et al., High-Level Radioactive
Waste Disposal, WIT Press, Southampton, 2011.)
et al. 2005). This study was made in association with the Geological Department
of Charles University in the same city. The mock-up test simulated SKB’s concept
KBS-3V on half scale, and aimed at providing detailed information on the perfor-
mance of the Czech buffer candidate RMN under repository-like conditions. The test
is referred to here because it indicates how smectite clay, rich in montmorillonite and
iron, performs under such conditions. The clay deposit belongs to a series of argil-
lized basaltic volcanoclastic accumulations of the Tertiary age, formed in lacustrine
basins some 100 km WNW of Prague.
The rock in the test was represented by a steel tube with 80 cm diameter to which
a filter was attached for uniform wetting of the buffer, which consisted of highly
compacted blocks of clay with 75% Ca montmorillonite, 10% finely ground quartz,
and 5% graphite powder (Figure 6.10). The clay blocks had a dry density of 1800 kg/
m3, a radial thickness of 18 cm, and a water content of 7%. The 50 mm gap between
the clay blocks and the filter was filled with RMN granules that were left uncom-
pacted. Samples were taken through 35 mm pipes that passed through the tube and
were replaced by new clay with the same density as the ones removed. The tem-
perature at the heater surface was kept at about 90°C, which gave a temperature
of 45°C–48°C at the outer boundary of the clay buffer at mid-height (temperature
gradient 2°C/cm). The filter was saturated with medium brackish granitic water
under 60 kPa pressure. Nearly complete saturation of the blocks and fill of granules
was reached after two years. Samples taken at the termination were used for miner-
alogical analyses and ordinary geotechnical investigations that are summarized in
Table 6.2 (Pusch 2008).
It is obvious that a large part of the clay buffer had undergone significant changes.
The ratio of the swelling pressures of the M2 and M4 samples, which had practi-
cally the same density, was 650/310 = 2.09. The coldest 50 mm zone had the same
178 Bentonite Clay
Cables for heating and transfer of

signals from sensors to recording
units
Water tank
Heater
Clay blocks
(buffer)
Lead-
through for Granules of clay
sampling block material,
during test separated from the
steel tank by filter
for supplying water
Instrumentation of the clay by thermal elements and gauges for

measuring the swelling pressure in vertical and horizontal directions
FIGURE 6.10 The Czech mock-up. Steel tube with central heater surrounded by buffer
blocks and fill by pouring of clay granules.
hydraulic conductivity as untreated clay with the same density, whereas for the clay
located between 10 and 18 cm from the heater, it was about 10 times higher than of
equally dense, untreated clay. The clay closer than 10 cm from the heater was up to
100 times more permeable than unheated clay with the same density. The swelling
pressure showed a corresponding pattern, that is, a marked drop for the samples
taken close to the heater.
Mineralogical characterization of the original clay was made by use of TEM with
energy dispersive X-ray and coherent scattering domain, that is, coherent scattering
domain data for determining the thickness of particles or stacks of lamellae. The
Koester diagram in Figure 6.11 shows the charge distribution in untreated RMN clay
compared to that of sample M4 after termination of the experiment.
Figure 6.12 shows that very obvious changes in charge distribution took place
in M4. It was concluded that significant dissolution of the Fe montmorillonite and
complete disappearance of the intergrowth of illite and kaolinite had taken place,
suggesting migration and precipitation of released elements at different distances
from the heater. TEM-EDX photos of M4 showed that desiccation fissures, assumed
to have been formed before resaturation took place, remained. This demonstrates
that the clay was not sufficiently expandable to self-heal. The XRD analyses showed
an obvious reduction in montmorillonite content when comparing M3 and M4, and
an increase in illite particle thickness in samples M3 and M4 as compared with M2
(cf. Table 6.3).
diVerm
Charge distribution
RMN diSV-ML
bent.
1.0 Fe-diSV-ML
5.0 0 Fe-Montm.
0.8 1 ML10
5.2
2 ML20
0.6 3 ML30
5.4
4 ML40
er
0.4
lay
0
5.6 0 6 ML60
ral
Ch
4 33
hed
6 0 Montm.
arg
0.2
5.8
cta
eo
f in
fo
0 0
0.0
eo
ter
6.0
arg
l
aye
Ch
r
6.2
6.4
6.6
6.8
14.0 14.2 14.4 14.6 14.8 15.0 15.2 15.4 15.6 15.8 16.0
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
Charge of tetrahedral layerSiIV
FIGURE 6.11 (See color insert.) Distribution of charge in tetrahedral, octahedral, and
interlayers (interlamellar space) in untreated RMN clay. ML represents mixed-layer m inerals;
the figures indicate percentages. (Data from Pusch, R. et al., Appl. Clay Sci., 47, 113–119,
2010; Analyses by Kasbohm.)
The significant changes in mineralogical and physical properties of the larger part
of the RMN buffer caused by mineral conversion and cementation is manifested by
Figure 6.13. It shows the change in Atterberg liquid limit, wL, which is an excellent
measure of the hydration potential of clays. For unheated RMN clay, wL is about
170%, and this figure dropped by about 30% in the hottest part of the buffer and by
about 10% in its peripheral parts.
The main results from the described study of samples from the mock-up test were
as follows (Prikryl et al. 2001; Kasbohm et al. 2000):
• The swelling pressure had dropped by more than 50% in the larger part of
the buffer. The hydraulic conductivity had increased by 10–100 times.
• Intergrowth of illite and kaolinite was obvious in the dioctahedral ver-
miculite in untreated RMN clay. In these aggregates, illite dominated over
kaolinite. In sample M4, most of the illite/kaolinite intergrowths were
dissolved.
180 Bentonite Clay
Charge distribution diVerm

MOCK M4 diSV-ML
1.0 Fe-diSV-ML
5.0 0 Fe-Montm.
0.8 1 ML10
5.2
2 ML20
0.6 3 ML30
5.4 0
er
lay
0 4 ML40
0.4
l
Ch
5.6
dra
5 ML50
arg
5
ahe
77 0 0.2 7 ML70
eo
00
oct
5.8 8
f in
8 ML80
of
ter
0.0 0 Montm.
ge
6.0
la
ar
yer
Ch
6.2
6.4
6.6
6.8
14.0 14.2 14.4 14.6 14.8 15.0 15.2 15.4 15.6 15.8 16.0
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
FIGURE 6.12 (See color insert.) Distribution of charge in tetrahedral, octahedral, and
interlayers (interlamellar space) of the M4 sample. (Data from Pusch, R. et al., Appl. Clay
Sci., 47, 113–119, 2010; Analyses by Kasbohm.)
TABLE 6.3
Mineralogical Changes in the Mock-Up Test
Mineral Sample M2 Sample M3 Sample M4
Montmorillonite Slight dissolution Moderate dissolution
Significant dissolution, low
crystallinity
Illite High disorder, minor illite Significant illite formed, reduced disorder
Talc and kaolinite Low disorder, particle Increased disorder or neoformation
growth
Source: Pusch, R. et al., Phys. Chem. Earth, Parts A/B/C, 32(1–7), 116–122, 2007; Analyses by Kasbohm.
• Fe set free by the dissolution of the Fe montmorillonite could have formed

iron complexes causing cementation in the entire buffer mass, especially in
the most heated part (M4).
• Replacement of octahedral Al by Fe could have caused reduced coherence
of the montmorillonite crystals, promoting easier dissolution.
• Formation of illite in the hottest part of the buffer (sample M4) may have
been caused by uptake of K from dissolved vermiculite.
wL = 170
wL = 160
wL = 140
Confining
tube
wL = 120
RMN clay
Heater
wL = 160
wL = 160
wL = 170
FIGURE 6.13 Evaluated liquid limit of samples extracted from the buffer surrounding the
heater, represented by the inner rectangle. (Data from Pacovsky, J. et al., Phys. Chem. Earth,
32, 767–779, 2005.) The outer rectangle is the confining steel tube.
6.2.5 Swedish–Czech–Chinese University Study

Hydrothermal tests under a higher temperature gradient and higher salinity have
shown the same tendency as the Czech experiment (Xiaodong et al. 2011). They
included investigation of a montmorillonite-rich type of clay (MX-80) and of
Na montmorillonite-rich Chinese bentonite (GMZ), as well as saponite-rich clay
(GeoHellas, DA0464, Na-converted, Greece), and German mixed-layer illite/smectite
(Friedland clay, FIM, Friedland mineral GmbH). Fine-granulated clay was com-
pacted in air-dry form under the same pressure (1.30 MPa) in cells giving the MX-80
and GMZ clays a dry density of 1350 kg/m3 and 1240 kg/m3, respectively, and a dry
density of 1530 kg/m3 of the FIM clay. Heating of one end of the hydrothermal cell
to 95°C and cooling the opposite to about 35°C by circulating 3.5% CaCl2 solution
through a filter, was made for three weeks, implying a thermal gradient of 15°C/cm
(Figure 6.14).
After hydrothermal treatment the samples were saturated with distilled water at
room temperature for dissolving possibly precipitated salt crystals, and sectioned
for determining the hydraulic conductivity, swelling pressure, and compressibil-
ity of the least and most heated parts, and for microstructural and mineralogical
examination. The saponite clay sample was not noticeably affected by the hydro-
thermal treatment,* whereas the MX-80 and GMZ clays had lost a significant part
* The high swelling pressure was caused by the low charge 0.21 per (OH)2O10.
182 Bentonite Clay
Pump
Sintered filter
Cooling
35 mm chamber
30 mm
Cell with bottom part of copper metal
and the other parts of acid-proof
stainless steel. The copper was kept
at 95°C throughout the test.
FIGURE 6.14 Hydrothermal test arrangement with the cold end kept at 35°C by circulating
cooled CaCl2 solution through the upper filter and maintaining the hot end of copper at 95°C.
(Data from Pusch, R., Geological Storage of Radioactive Waste, Springer-Verlag, Berlin,
Germany, 2008.)
of the expandability and showed an obvious increase in hydraulic conductivity

and stiffness. The mixed-layer FIM clay underwent the treatment with some but
not very significant signs of degradation. Figures 6.15 and 6.16 show the changes
in hydraulic conductivity and expandability, in terms of swelling pressure, of the
tested clays. The diagrams verify that the density of the hottest part of all the test
samples was highest, which is explained by compression in the early saturation
phase under the high swelling pressure of the earliest wetted, colder part. The dif-
ferent densities must be kept in mind when interpreting the diagrams from which
one can draw the following general conclusions concerning the impact of hydro-
thermal treatment:
• MX-80, representing montmorillonite-rich clays, was most strongly

affected, the most obvious effect being a significant drop in swelling pres-
sure and hydraulic conductivity of the most heated parts of the sample.
• Saponite-rich clay was affected only by complete exchange of initially
sorbed Na by Ca. The only impact of the treatment was the difference in
density, which was solely caused by the transient saturation process.
• Mixed-layer I/S (FIM) was most strongly affected by the compression in
the saturation phase, causing a very high density of the hot part.
• Temperatures higher than 65°C–75°C gave the most obvious impact on the
hydraulic conductivity and swelling pressure of all the clays.
Saponite 1880 kg/m3 1790 kg/m3 1800 kg/m3

MX-80 2000 kg/m3 1850 kg/m3 1800 kg/m3
FIM 2240 kg/m3 2030 kg/m3 1970 kg/m3
E−14
E−13
MX-80 Saponite FIM
E−12
K (m/s)
E−11
E−10
E−9
E−8
0 5 10 15 20 25 mm
T = 95°C T = 85°C T = 75°C T = 65°C T = 55°C T = 45°C
Distance from 95°C end of hydrothermal cell (mm)
FIGURE 6.15 (See color insert.) Diagram of the hydraulic conductivity (km/s) of the three
clays that had been compacted in air-dry form under 1 MPa pressure giving nearly the same
initial density of MX-80 and saponite but higher initial density of the smectite-poor FIM clay.

MX-80 2000 kg/m3 1850 kg/m3 1800 kg/m3
FIM 2240 kg/m3 2030 kg/m3 1970 kg/m3
2000
1750
Saponite
1500
MX-80
ps (kPa)
1250
500
250
FIM
0
0 5 10 15 20 25 mm
T = 95°C T = 85°C T = 75°C T = 65°C T = 55°C T = 45°C
FIGURE 6.16 (See color insert.) Diagram of the swelling pressure ps of the three clays.
184 Bentonite Clay
Mineralogical investigations gave further information of the impact of hydrothermal

treatment of the montmorillonite-rich, saponite-dominated, and I/S mixed-layer
clays (Kolarikova 2004; Kasbohm et al. 2005). The main findings were as follows:
6.2.5.1 Montmorillonite-Dominated MX-80

TEM-EDX results showed that most of the montmorillonitic particles had converted
to two series of mixed-layer phases: illite-smectite mixed-layer series (IS-ml), and
dioctahedral vermiculite–smectite mixed-layer series (diVS-ml).* The chemical com-
position of the diVS-ml phases was characterized by enrichment of Al in the octa-
hedral layer (Figure 6.17).
According to S'rodon and others, illitization can suitably be expressed in terms
of the relation between the amount of tetrahedral aluminum and of smectite layers
(S%), which, for the finest fraction of virgin MX-80 clay, gives S = 95%. For the
hottest part of the hydrothermally treated clay, S = 80% and for the coldest S = 75%.
This effect was also recognized by an increased distance between the positions of
(001)/(002) and (002/003) reflections in ethylene glycol-saturated samples. The octa-
hedral substitution of Mg by Al obviously reduced the deficit of charge in the octa-
hedral sheet.
Neoformation of illite in the hot part of the MX-80 sample is verified by
Figure 6.18 and by the small 10Å peak in the XRD diagrams for oriented mounts of
material (Figure 6.19).
6.2.5.2 Saponite-Dominated Clay

The mineralogical changes in saponite were very small, which is in agreement with
the conclusions from studies by Gueven and Huang (1990). The only change was that
Ca had replaced initially sorbed Na and Mg. The sum of the Na and Ca concentra-
tions was almost 50% higher in the coldest part than in the hot and central ones.
100 30 100 30
IS-ml diVS-ml IS-ml diVS-ml
Frequency (%)
Frequency (%)
Frequency (%)
Frequency (%)
50 15 50 15
0 0 0 0
tm
ite
tm
tm
ite
tm
rm
l3
l6
l3
l6
l3
l6
l3
l6
er
ill
ill
on
on
on
on
Ve
M
M
V
M
M
di
di
IS-ml diversity diVS-ml diversity IS-ml diversity diVS-ml diversity
FIGURE 6.17 TEM-EDX: mixed-layer phases in hot (left) and cold MX80 cold (right).
IS-ml, illite–smectite mixed-layer phases; diVS-ml, dioctahedral vermiculite–smectite mixed-
layer phases; and diVerm, dioctahedral vermiculite (interlayer charge-deficient illite). (Data
from Kasbohm, J. et al., Environ. Earth Sci., 69, 2569–2579, 2013; Analyses and interpreta-
tion by Kasbohm.)
* Illite with an interlayer charge and K-deficit (<0.89 per half unit cell) is called here dioctahedral
vermiculite.
0.40 μm
FIGURE 6.18 Neoformed pseudohexagonal particles, assumed to be illite and not, as one
may think, kaolinite, in the hottest MX-80 sample. (Data from Pusch, R. et al., Appl. Clay
Sci., 47, 113–119, 2010; SEM analyses by Kasbohm.)
Montmorillonite
Illite/mont-
morillonite
Quartz
Quartz
Gypsum
Chlorite
30 25 20 15 10 5
2Θ
FIGURE 6.19 X-ray diffraction spectrum of oriented sample with <32 μm particles from
the most heated part of the MX-80 sample. Ca uptake altered the typical lamellar spacing
of Na montmorillonite from 12.8Å to 15Å. Quartz and cristobalite, being natural constitu-
ents, remained apparently unaltered, whereas the chloride mineral sinjarite (2.83Å) appears
to have been neoformed in the hottest part, that is, up to 7 to 10 mm distance from the hot
boundary. (Analysis by Kasbohm.)
186 Bentonite Clay
The concentrations of Fe, Al, Mg, and Si dropped slightly in the direction from the
hot to the cold end. The dissolution of minerals was much less significant than in the
MX-80 clay.
6.2.5.3 Mixed-Layer I/S Clay

In the mixed-layer I/S (FIM) clay, Ca replaced the initially sorbed Na nearly com-
pletely in the hottest part and also to a high degree in the central and coldest parts.
The content of Mg sorbed by the clay was partly replaced by Ca or Fe set free in
conjunction with slight dissolution of iron complexes. The largely unchanged Fe, Al,
and Si contents showed that the clay mineral structure remained nearly unchanged
in all parts.
6.2.6 SKB Field Tests

Further data originate from a full-scale five-year field experiment at SKB’s under-
ground laboratory at Äspö with very dense MX-80 clay buffer surrounding a full-
size copper-lined KBS-3 canister in an 8 m deep, 1.75 m diameter deposition hole in
granite. It was evaluated in 2007 with respect to the performance of the clay (Svemar
and Pusch 2005). A filter had been placed in the hole for providing the clay with
groundwater, which had a total salt content of about 6000 ppm with a Na/Ca ratio of
about unity. The canister contained electrical elements for simulating the heat pro-
duced by the highly radioactive content of such canisters and its surface temperature
was maintained at 85°C for a couple of years and at successively lower temperatures
later in the experiment. The average radial thermal gradient was 1°C/cm. The evolu-
tion of the swelling pressure at the buffer–canister contact at mid-height canister is
shown in Figure 6.20.
4.5
4.0
3.5
Swelling pressure (MPa)
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0 500 1000 1500 2000
Time (days)
FIGURE 6.20 Development of swelling pressure at mid-height canister. The first part of
the curve represents the initial phase of hydration and compression of the pellet filling and
closure of joints between the clay blocks.
4500
4000
3500
3000
2500
2000
1500
1000
500
0
0 5 10 15 20 25 30
Distance from heater (cm)
FIGURE 6.21 False interpretation of the evolution of the swelling pressure by migration of
pressurized water along cables to the sensor. (Data from Pusch, R. et al., Eng. Geol., 2014.)
It is obvious that the hydration was much quicker than predicted. Thus, saturation
of the clay buffer in the Äspö hole, which was artificially fed with water, took
place in about the same time as of the clay in the wettest hole in the Stripa
experiment with 50% smaller radial thickness of the clay. The reason for the
discrepancy was that most of the pressure sensors at Stripa were placed at the
rock–buffer contact, whereas the majority of those in the Äspö experiment were
located within the buffer and at the buffer–heater contacts. In the Äspö case,
water under pressure migrated along the tubings from the rock–buffer contact to
the sensors in the interior of the buffer making the sensors react much quicker
than the clay surrounding them, as demonstrated by the graph in Figure 6.21
(Pusch et al. 2015).
A representative sample numbered R8:225: Canister was taken from the nearness
of the heater in the Äspö experiment and investigated by oedometer testing for find-
ing out if the conductivity and expandability deviated from those of equally dense
untreated MX-80 clay, and if chemical and mineralogical changes had taken place.
The comparison of untreated, MX-80 clay, saturated and percolated with distilled
water and having the same density as the R8:225 sample taken from the vicinity of
the heater, showed that it had an average hydraulic conductivity that was about 1/30
of the conductivity of this sample. In contrast, there was no discrepancy between the
swelling pressure of this sample and that of untreated MX-80. One hence concludes
that while the swelling pressure of the clay adjacent to the canister did not indicate
any physicochemical changes, the very obvious increase in hydraulic conductivity
and the reduced dispersibility of this clay caused by the exposure to hydrothermal
conditions, certainly demonstrate such changes. Table 6.4 summarizes the data from
testing of samples from the field experiment.
The main results from the mineralogical analyses are summarized in Table 6.5.
The data were derived from comprehensive (>100 specimens) particle-wise TEM-
EDX measurements (Pusch 2008).
188 Bentonite Clay
TABLE 6.4
Summary of Physical Data of the Sample Adjacent to the Canister at
Mid-Height, and of Untreated MX-80 Clay with the Same Density, That Is,
1970 kg/m3 (1540 kg/m3 Dry Density) Saturated with Distilled Water or
Artificially Prepared Ocean Water (3.5% TDS)
R8:225: Canister Untreated MX-80
Parallel to Perpendicular Distilled Ocean
Sample Canister to Canister Watera Watera
Hydraulic conductivity 2E−11 2E−11 7E−13 2E−12
(m/s)
Swelling pressure (MPa) 4.2 4.3 4.0 3.5–4.0
a Hydraulic gradient 30–300 m/m.
TABLE 6.5
Typical Mineral Changes of MX-80 Clay by Hydrothermal Treatment in Äspö
Buffert Test Project
Parameter Untreated MX-80 R8:225: Canister
Phases Montmorillonite (low-charge Montmorillonite (low-charge
montmorillonite), diVS-ml montmorillonite), diVS-ml
IS-ml IS-ml
Traces: PSV-ml KSV-ml
“albite”
Smectite Layers (S%)
S% in IS-ml phases 85% 95%
S% in diVS-ml phases 90% 80%
Interlayer charge (average of all) 0.222 0.305
Source: Analysis and interpretation by Kasbohm.

Note: Parameter data from TEM-EDX-analyses.
IS-ml, illite–smectite mixed-layer phases; diVS-ml, dioctahedral vermiculite–smectite mixed-layer
phases; PSV-ml, pyrophyllite–smectite–dioctahedral vermiculite mixed-layer phases; and KSV-ml,
kaolinite–smectite–dioctahedral vermiculite mixed-layer phases.
Major findings were as follows:
• MX-80 clay contains two general types of montmorillonite: (1) montmoril-

lonite with a normal charge as end member of IS-ml series and (2) low-charge
montmorillonite as end member of diVS-ml series. The ratio IS:diVS varies
from 1/20 to 2/3.
• The hydrothermally treated clay has undergone substitution of Al3+ by Fe3+
in the octahedral layer causing higher lattice stresses because of the larger
ion radius of Fe3+ than of Al3+ and reduction of the resistance of the mont-
morillonite part to dissolution. The change in the chemical composition of
the montmorillonite is illustrated by the derived formulae:
• Montmorillonite as end member of IS-ml series in virgin MX-80:
Ca0.04 Mg0.09 K<0.01Al1.64 Fe3+0.13 Mg0.23 Ti<0.01 (OH)2Si3.96 Al0.04 O10
• Montmorillonite as end member of IS-ml series in hydrothermally

treated clay:
Ca0.07 Mg0.02 Na0.04 K0.07Al1.46 Fe3+0.23 Mg0.25 Ti0.03 (OH)2Si3.98 Al0.02 O10
• The hydrothermally treated clay has undergone stepwise alteration from

normally charged montmorillonite to a low-charge montmorillonite by
replacement of original Mg by Al. The composition of the montmorillonite
as end member of diVS-ml series in the hydrothermally treated clay has the
following form:
Ca0.02 Mg0.04 Na0.01 K0.03Al1.62 Fe3+0.16 Mg0.17 Ti0.04 (OH)2Si3.99 Al0.01 O10
• Montmorillonite was the dominating mineral phase in both untreated and

hydrothermally treated clay. However, mineralogical changes induced by
the hydrothermal conditions are obvious as indicated by an increase in
smectite layer frequency in the IS-ml phases of the heated clay (95%) com-
pared to the typical number for untreated MX-80 (85%), and by an associ-
ated drop in smectite layer frequency in the diVS-ml phases from 90% of
untreated MX-80 to 80% in the heated clay.
• The lower interlayer charge of R8:225 than of untreated MX-80, means,
according to classical double-layer theory, that it should have higher swell-
ing pressure, which is in agreement with the diagrams in Figure 6.22 and
with the recorded swelling pressure of R8:225 (Table 6.6).
An effect of exposure to hydrothermal conditions that is of importance in conceptual

modeling of the microstructural evolution is the altered particle morphology shown
in Figure 6.18. It demonstrates the appearance of a cardhouse-structured mass of
crystallites of kaolinite-like, discrete particles that are in fact believed to be illite.
The particle assemblies were probably neoformed in the softest parts of the hetero-
geneous microstructure, that is, the microchannels described earlier and illustrated
schematically in Figure 4.11.
In freshly prepared buffer clay of MX-80 type saturated with low-electrolyte
water, the microchannels contain homogeneous clay gels with a density that is low
in large channels and high in narrow ones. Salt porewater causes coagulation of the
gels, which increases their hydraulic conductivity and thereby the bulk conductivity.
If the smectite particles become converted to illite, the conductivity will increase
further and, considering the hottest part of R8:225, it can contribute to the more than
thirtyfold rise in bulk conductivity (cf. Table 6.6).
190 Bentonite Clay
Charge of montmorillonite
4.50
Pusch, THMC
Swelling pressure (in MPa)

4.35
y = 3.115x + 3.549
4.20
4.05
3.90
y = 3.128x + 3.132
2
3.75 R = 0.839
Herbert, NF-PRO
3.60
0.15 0.20 0.25 0.30
(a) Interlayer charge of montmorillonite
Charge of all particles

4.50
Pusch, THMC
Swelling pressure (in MPa)
4.35
y = 2.590x + 3.444
4.20
4.05
3.90
y = 3.452x + 2.813
2
3.75 R = 0.932
Herbert, NF-PRO
3.60
0.20 0.25 0.30 0.35 0.40
(b) Interlayer charge
FIGURE 6.22 (a,b) Comparison of swelling pressure and charge for MX-80 clay showing
the same trend of increased pressure for increased charge in different experiments using
different techniques. (Prepared by Kasbohm.)
TABLE 6.6
Chemical Composition in Weight% of the Si-Rich MERIT 5000 Cement
Type SiO2 Al2O3 MgO Na2O CaO MgO K2O Fe2O3, FeO
Ma 34 13 17 0.5 31 17 0.5 0.3
a M—Merit 5000.
6.2.7 Matter of Stiffening
Precipitation of silica in the form of cristobalite, quartz, and amorphous silica, as
well as iron compounds, and also neoformed particles in the illitization process, can
lead to cementation and stiffening of the clay, and thereby to loss of expandability
and self-sealing potential after mechanical disturbance (Figure 6.23). The combined
action of heat-induced contraction of the stacks of smectite lamellae is hypotheti-
cally imagined in Figure 6.24. The Kinnekulle bentonites give example of silica
precipitation causing cementation that extended to at least 5 m distance from the
main bed (cf. Figure 6.2).
FIGURE 6.23 Silica precipitated on a silt particle in Kinnekulle bentonite. (Data from
Müller-Vonmoos, M. et al., Eng. Geol., 28, 269–280, 1990.)
Virgin state
expanded Cementing
stacks Grains with precipitations Hydrothermally
interlamellar voids treated clay
Clay gels in external voids
Grain
Collapsed
stacks, welded
together
leaving wider
voids
FIGURE 6.24 Illustration of the impact of heating on smectite clay buffer: contraction of
stacks of smectite lamellae causing widening of voids, and increased stiffness by precipita-
tion of cementing agents, which prevent subsequent spontaneous expansion on wetting. The
small picture shows the clay granules before hydration and expansion, which is associated by
growth of clay gels in the voids.
Stiffening by the combined effects of contraction of stacks of lamellae, called

lensing here, and precipitation of cementing agents was investigated in conjunc-
tion with several of the laboratory studies referred to in this chapter (Pusch 1993;
Pusch et al. 1999). The impact is illustrated in Figure 6.25, which compares the
compressive strain of samples of untreated and hydrothermally treated MX-80 clay
192 Bentonite Clay
MX-80 a: Hot end; b: Virgin

0.0
142 kPa
284 kPa
1.0
142 kPa
428 kPa
2.0
500 kPa
284 kPa 575 kPa
3.0
428 kPa
ε (%)
500 kPa a
4.0
575 kPa
5.0
680 kPa
6.0
b
7.0
0 30 60 90 120 150 180
Time (minute)
FIGURE 6.25 Strain versus time of hydrothermally treated MX-80 clay with ρd = 1375 kg/m3
(cf. Section 6.2.5) and untreated virgin MX-80 clay (ρd = 1380 kg/m3). (Data from Xiaodong, L.
et al., Appl. Clay Sci., 52, 419–427, 2011.)
at stepwise increased pressure. The experiment is described in Section 6.2.5. The

average strain rate for the stress range 142–500 kPa was 2.8%/hour for the hydrother-
mally treated clay and 4.7% for the untreated, virgin clay. The toughness of this more
ductile clay is illustrated by the large strain that it could sustain without failing. In
contrast, the hydrothermally treated clay showed brittle failure at a lower stress. The
role of density is illustrated by comparing these data with those from testing of the
earlier described hydrothermally treated MX-80 clay with 1650 kg/m3 dry density
described in Section 6.2.3, in which the corresponding strain was about 0.3%/hour.
Chemical modeling of Si release and precipitation in the near-field of the KBS-3V
repository by Grindrod and Takase (1993) has indicated that cementing silica will
be precipitated in the buffer in contact with the rock after some 500 years, thereby
stiffening the buffer. Impregnation of the microchannels with silicious precipitates
would logically reduce the hydraulic conductivity as well.
This and several other experiments demonstrate that the ductility of freshly
matured montmorillonite-rich clay will be changed to elasto-viscoplastic behavior,
as illustrated by the rheological model in Figure 6.26. The model is in principle of
Eyring’s type, that is, characterized by a time-dependent viscosity component and
an elastic component, yielding creep curves that we recognize from Chapter 4, and
that described the different behavior of untreated and cemented clays in terms of the
to parameter. The upper part of the rheological model in Figure 6.26 represents the
framework of cemented microstructural units, whereas the lower part is the virgin
system of noncemented units. V1 is activated after exceeding a critical stress level
when the cementation bonds have been broken down; the basic behavior, represented
by the lower part then takes over. We see the same behavior of hydrothermally
treated, cemented clays illustrated by Figures 6.27 and 6.28.
V1
E1
V2 E2
FIGURE 6.26 Generalized stress/strain model with dashpots and elastic elements of
Eyring/Kelvin-type forsmectite clay. V1 is viscous with a threshold function representing
cementation, V2 is purely viscous.
10−5
10−6
Strain rate (1/s)
10−7
10−8
10−9
102 103 104 105 106
Time (s)
FIGURE 6.27 Examples of theoretical creep curves derived from the rheological model.
The three curves are characterized by the following data sets (E—modulus of elasticity).
(After Hökmark.) Upper: E = 10 MPa, ηo = E11 Pas, and to = E3 s; Central: E = 10 MPa,
ηo = E11 Pas, and to = E2 s; Lower: E = 10 MPa, ηo = E12 Pas, and to = E3 s.
We will come back to the matter of stiffening in Chapter 8, where the HLW
disposal concepts VDH and SARC will be examined with respect to their per-
formance in reality, and end this part of the present chapter by showing a micro-
graph of illite-converted smectite clay, originating from Silurian volcanism, in
Figure 6.29. The transmission electron micrograph taken by the 1.5 MV in CNRS’
194 Bentonite Clay
10−5
10−6
Strain rate (1/s)
10−7
10−8
10−9
102 103 104 105 106
Time (s)
FIGURE 6.28 Negative to (100 s) resulting from the breakdown of cementation bonds in clay
with E = 10 MPa, ηo = E11 Pas, and to = E2 s. The upper curve is obtained by adding the upper
part of the rheological model with the properties E = E7 Pa, ηo = E8 Pas. (After Hökmark.)
FIGURE 6.29 Transmission electron micrograph of Silurian bentonite with all smectite
being converted to nonexpandable minerals, primarily illite. Micrograph taken of sample
with original porewater by use of the CNRS 1.5 MV microscope in Toulouse, France.
3.0
2.5
Compressive stress (MPa)
2.0
1.5
1.0
0.5
0.0
0 1 2 3 4 5
Strain (%)
FIGURE 6.30 Uniaxial compressive test of Silurian clay.
laboratories at Toulouse, France, was unique by representing undisturbed clay

with water-filled voids. The marine clay, exposed on the southeastern coast of the
Swedish island Gotland, was consolidated under several kilometers of sediments
in Permian time and by 3 km glacial load causing a bulk dry density of heating of
about 1750 kg/m3, and heating to at least 100°C for millions of years. Close exami-
nation of micrographs of this kind shows precipitations in the form of well-rounded
silicious nodules with an approximate size of 0.1 to 0.5 μm. Uniaxial compression
of a sample from the Silurian Burgsvik clay showed a peak at a strain of about
3% compression, indicating brittle failure caused by cementation but quite large
deformation before reaching failure (Figure 6.30). It is a case of Eyring–Kelvin
rheology with the clay matrix undergoing ductile strain until a critical stress level
has been reached where the strongest barriers to deformation fail (cf. the text on
stochastic creep in Chapter 3).
6.2.8 Interaction of Smectite Clay and Other Components

Most concepts for isolation of waste by use of smectite clay imply that it is in contact
with concrete in temporary or permanent supporting constructions and with metal
casings and canisters containing the waste. As to the metals, be it steel- or copper-
based containers, the question of chemical interaction with clay is very complex
and has physical implications by production of hydrogen gas that can form enclosed
bubbles with very high pressure or be able to pass through the clay with or without
creating flow paths that can be permanent (Pusch and Yong 2006). Since concrete
will be used for constructing bulwarks and vaults, we will include a discussion of the
interaction of this material and smectite seals.
196 Bentonite Clay
6.2.8.1 Smectite Clay Contacting Copper Metal

SKB has decided to use a canister design with a central cylindrical core of cast iron
surrounded by a 50 mm copper lining separated from the iron core that contains
space for bundles of spent fuel (Figure 6.31). The total weight is 24,600 kg, of which
copper makes up slightly less than one-third.
The expected chemical degradation is primarily by attack of sulfide minerals
in the embedding clay, whereas oxidation is not believed to play a significant role
(Svemar 2005). Through corrosion of the copper liner is not expected to occur until
after 100,000 years. The chemical interaction with clay will lead to the dissolution
of copper and very slow migration of copper ions into the clay as illustrated by
Figure 6.32, which refers to the experiment described in Section 6.2.5. This study
led to the conclusion that Na-montmorillonite, represented by MX-80, gave much
faster migration of copper ions and hence the corrosion of the copper liner than the
magnesium-smectite saponite and the I/S clay FIM.
Linear extrapolation of the recorded migration, which was initially 1.5 mm in
eight weeks, would give penetration of 10 mm/year and through-corrosion of the
50 mm liner in five years. In practice, however, the process is controlled by diffusive
1050 mm
4835 mm
5 cm Copper liner
Calculated weight (kg):
Copper canister 7,400
Iron insert 13,600
Fuel rods (BWR) 3,600
Total 24,600 BWR PWR
FIGURE 6.31 The SKB canister for HLW. (Data from Svemar, C., Cluster Repository
Project [CROP], Final Report of European Commission Contract FIR1-CT-2000-20023,
Brussels, Belgium, 2005.)
500 μm Carbon coating

(SEM-preparation)
PU101I01 MX-80 hot

6
Magnif.: 25×
EDX-results
Cu not detected
Cu = » 0.1%
Excited volume:
∼ 10 × 10 × 2 μm3
4 5
2 Contact surface
3
a b with Cu-plate
FIGURE 6.32 SEM-EDX analysis showing migration of Cu from the hot end (95°C) of the
hydrothermal cells into MX-80 clay. (Data form Pusch, R., Yong, R.N., Microstructure of
Smectite Clays and Engineering Properties, Taylor & Francis, New York, 2006; Micrograph
and analysis by Kasbohm.)
transport of the corrosion-driving agents, which makes it decline rapidly. The impact
of copper ions on the microstructural constitution of clay buffers and their physical
performance is by ion exchange to Cu from the original sorbed cation, Na or Ca,
depending on the time after deployment. It is generally stated that the physical per-
formance of smectite with Cu as sorbed cation will be practically the same as when
Ca is in the exchange positions.
6.2.8.2 Smectite Clay Contacting Iron and Steel

The role of oxygen is of fundamental importance for the corrosion of iron and steel
canisters including the iron insert of SKB canisters casing and alike (Figure 6.33).
Much has been said about the conditions concerning the oxidation of clay-embedded
steel or iron canisters but it is nowadays commonly stated that anaerobic conditions
prevail soon after placement in medium-deep and deep repositories, implying that
the reaction rate is very low.
Oxidation through the oxygen content that is enclosed in the initially partly air-
filled voids of the clay embedding iron and steel canisters will generate magnetite
early and stay as dominant corrosion product. More magnetite will be formed anaer-
obically following the reaction equation (Equation. 6.3):
3Fe + 4H 2O yielding Fe3O 4 + 4H 2 (6.3)

The Fe2+ concentration in the magnetite is determined by the noncongruent dissolu-
tion of this mineral, which reacts with the hydrogen and yields hematite (Fe2O3) and
goethite (FeOOH) (Svemar 2005). Other corrosion products formed in the absence
198 Bentonite Clay
FIGURE 6.33 Strong corrosion of acid-proof stainless steel oedometer to several hundred
micrometer depth at saturation of montmorillonite-rich clay with 3.5% NaCl solution under a
temperature gradient of 14°C/cm. The initial degree of fluid saturation was 24%.
of oxygen are ferrous hydroxide, Fe(OH)2, and ferric hydroxide, Fe(OH)3, of which
the latter has a strong tendency to form colloids of particles that normally carry a
positive charge and may be attached to smectite minerals. Iron ions will under all
circumstances be free to react with other elements and materials and be sorbed by
the canister-embedding clay through cation exchange, which will make the buffer
stiffer and reduce its self-sealing ability in case of void formation by large strain.
A phenomenon of great practical importance is that magnetite has a specific den-
sity that is only about 50% of that of iron, which means that it would occupy twice
the space that it fills up when placed. This will compress the buffer and raise the
pressure on the confining rock. Hydrogen gas formed in the oxidation process will
have a similar effect if the gas production rate is high since the bubble formed will
compress the buffer and make it so dense and tight that the raised swelling pressure
can break the surrounding rock and damage the canister.
The most important questions are therefore what the gas production rate and the
critical gas pressure of the buffer are. If the production rate is low, gas can diffuse
out of the system causing no harm; if not, we know from Chapter 4, the pressure
for release of gas by channel-wise penetration of the buffer is of the same order of
magnitude as the swelling pressure, that is, several MPa. Consolidation under the gas
bubble pressure will dramatically increase the density of the buffer clay implying a
dramatically increased critical gas pressure. The problem can be solved by selecting
a clay with high gas permeability also at very high densities, which speaks in favor
of I/S mixed-layer clay of FIM, Holmehus, or similar type.
6.2.8.3 I/S Mixed-Layer Smectite Clay Contacting Concrete

Geological disposal of waste will utilize much concrete since the rooms for deploy-
ment need to be supported where the rock is unstable. Ordinary Portland cement
gives a very high pH, which degrades contacting smectite clay as demonstrated inter
alia by batch tests with KOH/NaOH/Ca(OH)2 at 90°C, giving zeolites such as phil-
lipsite and analcime, in a few months (Pusch and Yong 2006).
The most important conclusions from these earlier studies using Portland cement
and montmorillonite-rich clay were as follows:
• High-alkali cement degrades quicker than low-alkali cement. The latter

deteriorates by destruction of the CAH gel.
• Dissolved elements and water migrate from the fresh cement paste to the
smectite in the first few hours. Ca leaving the cement and entering the clay
causes ion exchange and change in the microstructure of the clay by coagu-
lating softer parts.
• The cement paste is dehydrated and its voids become wider; water moves
from the smectite to the dense cement matrix.
• The dense cement paste fissures.
Since pH of Portland cement porewater can be higher than 12 and the critical value
for dissolution of smectite minerals is 10 or 11, there is a risk that Portland cement
in a concrete bulkhead can produce a large high-pH plume that can degrade smec-
tite backfills at a distance of several meters from it. Various attempts have therefore
been made to test low-pH cements, and an early study in Sweden showed positive
results respecting the reaction of the Swedish low-pH cement Merit 5000 with smec-
tite clay (Pusch et al. 1998, 2003). The cement had the chemical composition as
given in Table 6.6. The experiments were conducted using the type of cells shown
in Figure 6.34. Clay powder was compacted in them to a dry density of 1430 kg/m3
(1900 kg/m3 at water saturation). The perforated tube in the center had a thin filter
for preventing clay particles from migrating into the cement water solution that was
contained in the tube. The solution was regularly replaced and analyzed with respect
to the concentration of pH and important elements. The clay was of I/S mixed-layer
type (FIM), selected because of its advantage with respect to physical performance
(high gas permeability) and longevity in HLW repositories.
After five months, parts of the clay samples were extracted and examined using
X-ray diffraction and electron microscopy. The rest of the clay samples were used for
determination of the hydraulic conductivity. Replacement of the cement water once
Central perforated
tube with cement
porewater
Clay sample in
30 mm diameter cell
FIGURE 6.34 Cell for clay–cement experiments.

200 Bentonite Clay
per test period maintained the condition of exposure of the clay to high-concentrate
cement water throughout the experiment. The major conclusions were as follows:
• Slight dissolution of the clay was manifested by increased Si, Al, Mg, and
Fe contents in the solutions.
• Increase in iron in the solutions indicated that dissolution of the chlorite
component of the clay appears to have dominated the degradation of the clay.
• Increase in Ca and drop in Na concentration indicated cation exchange by
uptake of Na from the solutions by the clay and release of Ca to them.
• The insignificant change in K-content in the solutions suggested that illiti-
zation was absent or insignificant.
• XRD analyses (Figure 6.35) verified that no significant changes in the min-
eralogical composition had taken place and that zeolites had not been formed
although X-ray-amorphous particles of such minerals may have been present.
• Electron microscopy showed the typical structure of Friedland clay with
no zeolite minerals present. The element distribution spectra showed a Ca
peak from the interaction with Merit 5000 cement water.
The general conclusion from determination of the physical properties of the clay
saturated with and contacted by Portland cement water in the five-month long test
period did not markedly change its hydraulic conductivity. In fact, the treated clay
had its conductivity reduced to about one-fourth of the value of untreated clay. This
Quartz
Mixed-layer
Montmorillonite/
Mixed-layer
muscovite
Montmorillonite/
muscovite
Chlorite
Chlorite
Chlorite
Illite
Illite
30 25 20 15 10 5
2Θ
FIGURE 6.35 X-ray diffraction diagram of Friedland clay powders from outer (external)
and inner (internal) part of clay reacted with porewater of Merit 5000 cement. (Data from
Pusch, R. et al., Appl. Clay Sci., 23, 203–210, 2003; Pusch, R., Yong, R.N., Microstructure of
Smectite Clays and Engineering Properties, Taylor & Francis, New York, 2006.)
could have been caused by clogging of microchannels by amorphous hydrosilicate

compounds formed by cement-water attack on the clay minerals, or by flow-generated
transport of clay aggregates to clogging positions in the channels.
The satisfactory interaction of Merit 5000 low-pH cement and smectite clay initi-
ated a second test series focussing on the interaction of more montmorillonite-rich
clay of mixed-layer I/S type, and Merit 5000 cement.
6.2.8.4 Montmorillonite-Rich Clay Contacting Low-pH Concrete

A Danish Tertiary mixed-layer I/S clay Holmehus clay formed in nature by sedi-
mentation under marine-to-brackish conditions and composed of smectite, glauco-
nite, quartz, feldspars, calcite, glass, kaolinite, Ti oxide, and pyrite was used in the
study. The smectite makes up 60% of the I/S phase. The clay is similar to but richer
in smectite than the German Friedland (FIM) clay that is widely used for isolating
hazardous waste in landfills (Pusch et al. 2015). The raw material was dried at 60°C,
crushed, and sieved to a maximum grain size of 2 mm.
The samples used for a simple hydrothermal study consisted of two clay discs
separated by one disc of concrete. The cells were placed in ovens for heating to room
temperature, 100°C and 150°C, for two months (Figures 6.36 and 6.37) (Pusch et al.
2012).
The testing comprised determination of the compressive strength of the triple clay–
concrete–clay units and chemical and mineralogical analyses of the respective units.
Figure 6.38 shows the successive breakdown of the sample set stored at room
temperature. The upper clay disc prepared with distilled water was ductile and failed
when the compression was about 3% and the pressure around 500 kPa, correspond-
ing to the undrained shear strength 250 kPa, assuming Mohr–Coulomb failure mode.
The lower, saltwater-saturated clay disc was somewhat stronger but more brittle.
The concrete remained intact until the pressure had reached about 700 kPa and the
Distilled water
mm
15
A ceramic filter
mm
50 mm
30
Holmehus clay
Talc–concrete and very fine quartz
mm
15
3.5% CaCl2 solution

Hydrothermal test of clay/concrete/clay
FIGURE 6.36 Sets of clay/concrete for hydrothermal treatment under isothermal conditions.
202 Bentonite Clay
FIGURE 6.37 Preparation of sample set. Left: compressed, water-saturated clay. Right:
freshly cast concrete over the lower clay sample.
FIGURE 6.38 Compression stages of the sample set stored at room temperature. Upper
left: Initial failure of the upper clay disc for about 500 kPa pressure. Upper right and lower
left: Breakage of the central concrete disc and lower (salt) clay disc at 700 kPa. Lower right:
Total collapse of the set at more than 8% vertical strain.
total compressive strain had become about 8%. Further compression led to complete
brittle failure of all the components.
The behavior of the 100°C sample set was similar to that of the room tempera-
ture set but the compressive strength was higher. Thus, the clay that had been pre-
pared by saturation with distilled water was the first to fail and had a compressive
strength of around 900 kPa, corresponding to a shear strength of about 450 kPa. The
strength was hence nearly doubled compared to that of the clay saturated with dis-
tilled water and stored at room temperature. The lower saltwater-saturated clay was
slightly stronger. The average compression modulus of the clays was similar to that
of the room temperature clays. The concrete remained intact until the pressure had
reached about 1500 kPa when the total compressive strain had become about 40%.
Further compression led to complete failure of the entire set.
Figure 6.39 shows the successive breakdown of the sample set heated at 150°C for two
months. The ductility of the clay saturated with distilled water (upper disc) had been lost
and changed to brittleness. The clay saturated with distilled water failed when the com-
pression was about 3% and the pressure around 1300 kPa, implying that the u ndrained
shear strength was about 650 kPa. The lower disc prepared with saltwater did not fail
until the total compressive strain was about 1.5% and the pressure about 1500 kPa, indi-
cating significant stiffening and strengthening. The concrete remained intact until the
pressure was about 3300 kPa and the total compressive strain had become about 5%.
Further compression led to complete brittle failure of all the components.
The compression modulus of the clays was estimated at 25 MPa, that is, about
three times higher than for the room temperature set, hence showing that the 150°C
set was much stiffer than the less heated ones. The concrete remained intact until
the pressure was 3300 kPa and the total compressive strain had become about 40%.
The major chemical and mineralogical changes are illustrated by Figure 6.40 that
shows spectra from the X-ray diffraction analysis of clay samples from the 20°C,
100°C, and 150°C experiments.
The dominant reflection was that of illite (10Å), being most intense in the room
temperature sample and decreasing at higher temperatures, which can be ascribed
to overlapping of the I/S reflections that show the impact of ion exchange from Na to
Ca and the incorporation of two interlamellar hydrates.
FIGURE 6.39 Compression stages of the samples exposed to 150°C for two months. Initial
failure took place in the upper clay sample at 1300 kPa pressure.
204 Bentonite Clay
2400
.0
10
2300
ite
2200
Ill
2100
2000
1900
1800
1700
1600 6
7.
1500 m
1400 su
Lin (counts)
1300 yp
3
G
1.
.1
1200
ca
1100
15
I/S
1000
7.
ite
4
900
4.
in
5
.1
.0
800
ol
e7
ca
Ka
700
rit
I/S
lo
600
Ch
500
400
300
200
100
0
5 6 7 8 9 10 11 12 13 14 15
2-theta-scale
FIGURE 6.40 (See color insert.) X-ray diffraction analysis of the clay samples showing
the impact of heat and contacting cement: red—21°C, blue—100°C, and green—150°C. The
d-values given are in Å units. The background trace has been subtracted. Lin, Linear counts.
(Original diagram provided by L. Warr, Greifswald University, Germany.)
Scanning electron microscopy combined with EDX analyses of the clay and
cement portions across the contact between the two materials revealed a variety
of microstructural features in both the clay and cement phases at the respective
experimental temperatures (Figure 6.41). The clay subjected to 100°C was simi-
lar to that of the 20°C sample, with similar crenulated edges of the I/S particles
(Figure 6.41C). A notable difference was the occurrence of irregular patches of
cement that coated many of the clay aggregates, and appears to have precipitated
from the interaction of the fluids derived from the cement and the saline CaCl2
solution. These neocrystallized cement precipitations, which include gypsum,
are at least partly responsible for the hardening of the clay beneath the concrete
layer and indicate that cement-derived fluids penetrated significant parts of the
underlying clay layer during the hydrothermal treatment. The 150°C clay samples
showed I/S particles with notably smoother edges that developed in the lower
temperature samples, which can be attributed to enhanced dissolution of the clay.
Localized patches of Ca cement precipitates were also found within the clay, as
in the 100°C sample.
EDX analyses showed some changes in composition of the I/S particles with
increasing temperature. Mg, Fe, and K decreased slightly, indicating possible dis-
solution of illitic and or smectite layers, whereas the Ca content increased by the
exchange of interlayer cations. However, as the clay particles were commonly coated
by nanometer-sized neocrystallized cement grains (e.g., gypsum), which were
increasingly evident at higher temperatures in the cement, these chemical trends
also reflect the increasing degree of cement contamination (Ca phases). The main
constituents of Si and Al showed no recognition changes.
(a) (b)
C
T
I/S
2 μm 2 μm
(c) (d)
T
C
I/S
C
2 μm 2 μm
(e) (f ) T
I/S
2 μm 2 μm
C
FIGURE 6.41 Scanning electron micrographs of hydrothermally treated sets of clay and
cement. Left column shows clay stored at room temperature (a), clay heated at 100°C (c),
and clay heated at 150°C (e). Right column (b, d, and f) shows contacting concrete
correspondingly heated (T is talc).
Summarizing all presently available information on the chemical interaction of

low-pH cement and I/S mixed-layer clay, one concludes the following:
• At room temperature, there is slow cation exchange in the clay of initially

sorbed Na leading to Ca saturation.
• Exposure to hydrothermal conditions at 100°C will give an increase in the
compressive strength and stiffness by approximately 25% compared to the
samples stored at room temperature.
206 Bentonite Clay
• Exposure to hydrothermal conditions at 150°C increases the strength of

the clay (and cement) to become 50% to 100% higher than at room tem-
perature and makes the clay brittle with loss of expandability. Clay satu-
rated with distilled water will be somewhat weaker than for saturation with
3.5% CaCl2 solution.
• Cement porewater can migrate into contacting clay both at room tempera-
ture and higher temperature, bringing gypsum into the clay.
• The chemical stability of I/S clay in contact with low-pH cement at tem-
peratures up to 100°C is obvious in short-term perspectives. Exposure to
150°C is expected to lead to dissolution of both materials to an extent that
is determined by the concentration of Si in the water.
6.3 SUMMARY RESPECTING SMECTITE

CHEMISTRY AND MINERALOGY
Putting together the results of the described laboratory and field tests, and keeping
in mind the experience from the Kinnekulle case, one can make the following brief
comments.
6.3.1 Overview
• In all the described tests, smectite remained the major clay mineral in the
buffer even at temperatures of up to 150°C. However, the fact that natu-
ral analogs with temperature histories similar to that of repositories have
undergone significant loss of montmorillonite demonstrates that buffer clay
will also have its content of this mineral reduced. The rate of conversion of
the montmorillonite content to nonexpandable minerals or complexes is in
fact not yet known; for the Kinnekulle case, the reduction by 50%–75% may
have taken place in about 1000 years. In this context the large difference in
thermal gradient—less than 0.02°C/cm for the Kinnekulle case and more
than 1°C/cm for the experiments—should be considered. It would imply
much quicker changes in a repository.
• Mineralogical changes were similar in all the experiments. The most obvi-
ous phenomenon was dissolution of the accessory minerals calcite and feld-
spars but also, but to a smaller extent, of quartz and montmorillonite. The
released elements Si and Fe were precipitated in the course of the experi-
ments or when cooling took place, forming cementing agents that signifi-
cantly reduced the expandability.
• In the early stage of evolution, desiccation will take place in a large part of
the buffer and frequent fissures and growth of initial voids will appear in
the matrix of contracted stacks of lamellae. Precipitation of Si and Fe can
weld the lamellae, stacks of lamellae, and aggregates of stacks together
and prevent expansion when water later enters this part. This will cause
permanent increase in hydraulic conductivity and drop in swelling pressure.
• The observed silicification in the limestones confining the bentonite beds
at Kinnekulle agrees with the basic smectite-to-illite conversion model that
specifies that silica is released from silicious minerals, primarily feldspars,

and precipitated where lower temperature prevails, yielding cementation.
• Montmorillonite-rich clay in contact with concrete based on Portland
cement degrades significantly in a few years (Pusch et al. 2011). Use of con-
crete with low-pH cement and montmorillonite-rich I/S mixed clay reduces
the negative impact on the clay significantly.
It is important to realize that the conditions under which the tests were conducted
can have saved the clay from stronger degradation by offering unlimited access to
water and by limiting the testing time. In practice, the majority of the buffer in
deposition holes will be supplied with very little water and this water may have a
much higher salt content than in the experiments. The most obvious threat to accept-
able long-term performance of montmorillonite buffers is cementation, which hence
requires particular attention.
6.3.2 Geochemical Modeling
The need for safer prediction of the fate of smectite clays exposed to hydrothermal
conditions has caused intense work, primarily by the oil and gas industry for devel-
oping lasting drill muds and this work together with the R&D of SKB, ANDRA,
and other organizations responsible for HLW disposal, directly or indirectly, has
given further insight in this problem area. Comprehensive work has been made at the
Universities of Strasbourg and Greifswald and we will summarize some of the expe-
rience gained, which has led to development of conceptual models and geochemical
codes. We will start by referring to a few major evolution processes identified in
natural sedimentary smectite clays.
• Neoformation of illite is expected to take place at certain concentrations

of silica (H4SiO4), aluminum, and potassium, yielding crystal nuclei in the
form of laths. When precipitation takes place the potassium concentration
drops locally and the resulting concentration gradient brings in more potas-
sium by which the process continues. Geochemical codes tend to indicate
that illite should be formed from smectites in a certain window of phase
diagrams of silica, aluminum, and potassium, but they do not seem to be
able to indicate whether the conversion takes place via mixed-layer mineral
stages or by dissolution/neoformation. Formation of illite in this way can
take place parallel to conversion of smectite to illite via mixed-layer I/S
minerals.
• As to the role of iron, we saw from the experiments with the Fe-rich Czech
smectite clay with Fe located in octahedral positions in the crystal lattice
that Fe may promote solubility. It can also serve as cement by forming
hydroxides or other compounds.
• A fundamental basis for modeling the chemical evolution of smectite buffer
is that it will undergo two major degrading processes under the hydrother-
mal conditions that will prevail for shorter or longer periods of time in most
HLW repositories.
208 Bentonite Clay
• Practically important dissolution of montmorillonite starts at temperatures

exceeding about 60°C but is assumed to be very slow at temperatures lower
than 80°C–90°C. It depends on the concentration of silicons in the pore-
water, which in turn depends on the dissolution of accessory minerals and
precipitation of silicious compounds, such as illite. For formation of illite,
potassium must be available and it therefore controls the formation of this
mineral. Buffers with very little feldspars containing potassium and calcite
are preferable for long-term stability. The lack of long-lasting tests under
controlled conditions makes predictions uncertain.
• Precipitation of silicious compounds yields illite to an extent that is con-
trolled by the access to potassium, and other crystalline or amorphous mat-
ter. They are formed in the most porous parts of the clay matrix, that is, the
microchannels, which thereby contain minerals with less hydration poten-
tial than montmorillonite and thereby represent permanent, pervious paths.
6.4 CONCLUDING REMARKS

This chapter contains a number of examples of chemically induced changes of clay
materials used or intended for sealing purposes and indicates what the threats can
be to them in landfills and underground repositories. It is obvious that the excel-
lent sealing properties of smectite soils described in Chapter 5 may not—or in
most cases will not—be long-lasting, which forces decision makers to assess what
thicknesses and designs of engineered barriers one should consider for reasonably
safe waste disposal. Laboratory experiments under well-defined testing conditions
make up the first step in selecting the best clay seals; however, for isolation of
highly radioactive waste or of the even more dangerous wastes of concentrated
arsenic, mercury, and pesticides, the conditions on site have to be properly defined
with respect to hydraulic boundary conditions and underground stability. They will
change with time making modeling of long-time performance of engineered barri-
ers uncertain or even a gamble.
REFERENCES
Stein, J.S.; Sandia National Laboratories. 2009. Deep Borehole Disposal of High-Level
Eberl, D., Hower, J. 1976. Kinetics of illite formation. Geological Society of America Bulletin,
87:1326–1330.
Gray, M. 1993. OECD/NEA International Stripa Project 1980–1992: Overview. Vol III,
Engineered barriers. SKB, Stockholm, Sweden.
Grim, R.E., Guven, N. 1978. Bentonites. Developments in Sedimentology, 24. Elsevier, New
York.
Grindrod, P., Takase, H., 1993. Reactive chemical transport within engineered barriers.
Proceedings of the 4th International Conference on the Chemistry and Migration
Behaviour of Actinides and Fission Products in the Geosphere, Charleston, SC,
December 12–17, pp. 773–779. Oldenburg Verlag 1994.
Grindrod, P., Takase, H., Wood, G.P. 1999. Modelling silicitation of a clay buffer subject to
opposing temperature and hydration gradients. Engineering Geology, 54:215–222.
Gueven, N., Huang, W.L. 1990. Effects of Mg2+ and Fe3+ substitutions on the crystallization of
discrete illite and illite/smectite mixed layers. Int. Rep., Dept. Geosciences Texas Tech.
University, Exxon Prod. Res. Co, Houston, TX.
Herbert, H.-J., Kasbohm, J., Henning, K.-H. 2004. Long-term behaviour of the Wyoming ben-
tonite MX-80 in high saline solutions. Applied Clay Sciences, 26:275–291.
Hoffman, J., Hower, J. 1979. Clay mineral assemblages as low grade metamorphic geother-
mometers. Application to the thrust-faulted disturbed belt of Montana, USA. SEPM
Special Bulletin, 26:55–79.
Kasbohm, J., Henning, K.-H., Herbert, H.-J. 2000. Zu Aspekten einer Langzeitstabilität von
Bentonit in hochsalinaren Lösungen. Berichte der DTTG, Jahrestagung Greifswald,
August 30–September 1, 158–170.
Kasbohm, J., Herbert, H.-J., Henning, K.-H. 2005. Short and long term stability of selected
bentonites in high saline solutions. International Symposium on Large Scale Field Tests
in Granite, Sitges, Barcelona, Spain, November 12–14, 2003. Code 88404:231–240.
Kasbohm, J., Pusch, R., Nguyen, T.L., Hoang, M.T. 2013. Labscale performance of selected
expandable clays under HLW repository near-field view. Environmental Earth Sciences,
69:2569–2579.
Kolarikova, I. 2004. Thermal properties of clays in relation to their use in the nuclear reposi-
tory. PhD thesis. Institute of Geochemistry, Mineralogy and Mineral Resources, Charles
University, Prague, Czech Republic.
Müller-Vonmoos, M., Kahr, G., Bucher F., Madsen F. 1990. Investigation of Kinnekulle
K-bentonite aimed at assessing the long-term stability of bentonites under repository
conditions. Engineering Geology, 28:269–280.
Nadeau, P.H., Bain, D.C. 1986. Composition of some smectites and diagenetic illitic clays and
implications for their origin. Clays and Clay Minerals, 34:455–464.
Pacovsky, J., Svoboda. J., Zapletal, L. 2005. Saturation development in the bentonite barrier
of the mock-up CZ geotechnical experiment. Physics and Chemistry of the Earth,
32(8–14):767–779.
Prikryl, R., Ryndova, T., Vejsada, J., Jelinek, E., Bohac, J., Sebec, O. 2001. Tertiary montmo-
rillonite-rich clays from the Cheb basin (Czech Republik): Alternative backfill material
in the nuclear waste disposal. Proceedings of the Joint 6th Biennial SGA-SEG Meeting,
Krakow, Poland, August 26–29. A.A. Balkema Publishers, Abingdon.
Pusch, R. 1983. Stability of Deep-Sited Smectite Minerals in Crystalline Rock—Chemical
Aspects. SKBF/KBS Technical Report TR 83-16. SKB, Stockholm, Sweden.
Pusch, R. 1985. Final Report of the Buffer Mass Test—Volume III: Chemical and Physical
Stability of the Buffer Materials. Stripa Project Technical Report 85-14. SKB, Swedish
Nuclear Fuel and Waste Management AB, Stockholm, Sweden.
Pusch, R. 1993. Evolution of Models for Conversion of Smectite to Non-Expandable Minerals.
SKB Technical Report TR 93-33, SKB, Stockholm, Sweden.
Pusch, R. 1999. Experience from preparation and investigation of clay for microstructural
Pusch, R., Börgesson, L., Erlström, M. 1987. Alteration of Isolating Properties of Dense
Smectite Clay in Repository Environment as Exemplified by Seven Pre-Quaternary
Clays. SKB Technical Report TR 87-29, SKB, Stockholm, Sweden.
Pusch, R., Börgesson, L., Ramqvist, G. 1985. Final Report of the Buffer Mass Test—Volume
II: Test Results. Stripa Project Technical Report 85-12. SKB, Stockholm, Sweden.
Pusch, R., Karnland, O., Lajudie, A., Lechelle, J., Bouchet, A. 1993. Hydrothermal Field Test
with French Candidate Clay Embedding Steel Heater in the Stripa Mine. SKB Technical
Report TR-93-02, SKB, Stockholm, Sweden.
210 Bentonite Clay
Pusch, R., Kasbohm, J., Hoang, T.M.T. 2010. Chemical stability of montmorillonite buffer
clay under repository-like conditions—A synthesis of relevant experimental data.
Applied Clay Science, 47(1/2):113–119.
Pusch, R., Kasbohm, J., Pacovsky, J., Cechova, Z. 2007. Are all smectite clays suitable as
“buffers?” Physics and Chemistry of the Earth, Parts A/B/C, 32(1–7):116–122.
Pusch, R., Kasbohm, J., Thao, H.M., Knutsson, S., Lan, N.T. 2015. Holmehus clay—A
Tertiary smectitic clay of potential use for isolation of hazardous waste. Engineering
Geology, 188:38–47.
Pusch, R., Madsen, F. 1995. Aspects on the illitization of the Kinnekulle bentonites. Clays and
Clay Minerals, 43(3):261–270.
Pusch, R., Muurinen, A., Lehikoinen, J., Bors, J., Erikse, T. 1999. Microstructural and
Report Contract No F14W-CT95-0012, European Commission, Luxembourg.
Pusch, R., Ramqvist, G., Bockgård, N., Ekman, L., 2011. Sealing of investigation boreholes,
Phase 4, Final Report. SKB Research Report R-11-20. SKB, Stockholm, Sweden.
Pusch, R., Ramqvist, G., Kasbohm, J., Knutsson, S., Mohammed M.H., 2012. The concept of
Journal of Earth Sciences and Geotechnical Engineering 2(3):1–24.
Pusch, R., Takase, H., Benbow, S. 1998. Chemical Processes Causing Cementation in Heat-
Affected Smectite—The Kinnekulle Bentonite. SKB Technical Report TR-98-25, SKB,
Stockholm, Sweden.
Pusch, R., Yong, R.N. 2006. Microstructure of Smectite Clays and Engineering Properties.
Pusch, R., Yong, R.N., Nakano, M. 2011. High-Level Radioactive Waste Disposal. WIT Press,
Southampton.
Pusch, R., Zwahr, H., Gerber, R., Schomburg, J. 2003. Interaction of cement and smectite
clay—Theory and practice. Applied Clay Science, 23:203–210.
Pytte, A.M., Reynolds, R.C. 1989. The thermal transformation of smectite to illite. In Thermal
History of Sedimentary Basins, Naeser, N.D., McCulloh, T.H. (Eds.), Springer-Verlag,
Berlin, Germany, pp. 133–140.
Roaldset, E. 2005. Lecture at International Workshop on clay materials for isolation of highly
radioactive waste, Lund, Sweden.
Weaver, C.E. 1979. Geothermal Alteration of Clay Minerals and Shales: Diagenesis. Technical
Report ONWI-21, ET-76-C-06-1830 Contr. Battelle–Office on Nuclear Waste Isolation,
Georgia Institute of technology, Atlanta, GA.
Xiaodong, L., Prikryl, R., Pusch, R. 2011. THMC-testing of three expandable clays of poten-
tial use in HLW repositories, Applied Clay Science, 52:419–427.
7 Processing of Clays and
Preparation of Seals
This chapter deals with preparation of smectite clays for manufacturing dense
blocks for sealing of boreholes and isolation of containers (canisters) with radioac-
tive waste, and for preparing top and bottom liners of landfills of hazardous waste.
7.1 OVERVIEW
Smectite-rich clay has excellent waste-isolating properties by virtue of its low per-
meability and swelling potential, which enables it to establish tight contact with the
confinement. Its cohesion even at low densities is valuable for using such clay for
grouting rock fractures and, as we will see in Chapter 9, as a base for preparing
creams and pastes in cosmetics—ask your betterhalves if they knows that they puts
smectite clay on their lips every day! And ask yourself if you know that the paint you
put on the summerhouse contains smectite for its thixotropy—it is easily rolled on,
does not drip, and stays there without flowing!
We have been focusing on clay for isolation of waste and have seen that for certain
purposes, such as isolation of canisters with high-level radioactive waste (HLW)
confined in metal canisters, very dense clay buffer, of smectite-rich clay is desirable.
But we have also realized that clay layers in the top and bottom liners of landfills of
hazardous waste are preferably moderately smectite-rich that are not too dense.
There are natural clays with very high amounts of smectite minerals, but most
exploitable beds of bentonite or clay-weathered rock do not contain more than
50%–80% of them. Although it is deemed possible to cut out and trim blocks from
such beds for direct placement in repositories, the density of shallow beds would be
too low for HLW isolation, as exemplified by the Kinnekulle bentonites described
in Chapter 6. The best way to prepare dense blocks is to excavate clay from natural
smectite-rich beds, and to dry and mill it for compaction of blocks with required den-
sity (Pusch 1994, 2008). For construction of tight top and bottom liners of landfills,
moderately smectite-rich clays—natural or prepared by mixing smectite-rich clay
with silt or sand—placed and compacted on site by vibratory rollers or plates are
often recommended (Al-Thaie 2012).
We will focus here on the selection and preparation of clay materials for three
purposes: (1) manufacturing of dense blocks; (2) construction of clay liners with
special respect to criteria and use of different techniques; and (3) grouting of rock.
Research for investigating the function of smectite clay as engineered barriers was
initiated some 40 years ago and is accelerating as this book shows.
211
212 Bentonite Clay
7.2 BLOCK PREPARATION

7.2.1 Raw Material
Most natural smectite-rich clays intended for use in repositories or as landfill liners
have Ca as exchangeable cation and are less expandable and tight than Na forms.
This latter state is reached by mixing the raw clay material with 3%–5% sodium
carbonate in processing plants, so that the major sorbed cation will be Na, but the
exchange is not complete and some reverse reaction by Ca released from the Ca car-
bonate formed in the process will take place with time because Ca readily replaces
Na (Sposito 1984). Natural Na-saturated smectites are rare (Grim 1967).
Crushed, air-dried clay can be ground and sieved for reaching a suitable gradation
of clay granules with a dry density of 2000–2200 kg/m3 (Jacobsson and Pusch 1978)
(cf. Figure 7.1). It is suitable for manufacturing blocks of nearly any size. For various
commercial purposes, pellets can be prepared using tablet machines and their dry
density can be up to about 2200 kg/m3. For maximum density of a fill of granules or
compacted pellets, two or more sizes are suitable but not necessary.
7.2.2 Achievable Block Density

A detailed, systematic investigation of the influence of a number of factors on the
compactability of granular smectite-rich clay in the laboratory has been performed
in Sweden by Johannesson et al. (1995). The materials were sodium-converted Ca
bentonite (IBECO-Na) from Germany and natural bentonite with Na as the major
adsorbed cation from Wyoming, United States (MX-80), with similar mineralogical
and chemical properties (liquid limit wL = 400%–520%). The study gave the follow-
ing major conclusions (cf. Figure 7.2):
• Relative humidity higher than 70% makes the granules expand and stick
together.
• The size of individual smectite particles ranges from 30Å (3 lamellae) to
1 μm.
• The compactability of smectite clays drops with increasing smectite content.
• The expansion of samples compacted to reach water saturation is about 4%
on unloading without access to moisture.
• The dry density of a compacted mass of granules under a given pressure
decreases with increasing water content. The maximum dry density is
the specific gravity of the minerals reached asymptotically at increased
compaction.
• The void ratio rises with increasing smectite content for any compaction
pressure.
Various tests have shown that use of very high compaction pressures (150–200 MPa)
of clay granules in vacuum can give dry densities of compacted blocks of slightly
more than 2100 kg/m3. An alternative way of preparing dense clay blocks is to saw
them from natural, heavily overconsolidated sedimented clays below the dry crust
Processing of Clays and Preparation of Seals 213
100
90
80
70
Weight % (passing)
60
50
40
30
20
10
0
1 10 100 1000
Grain size (μm)
45
40
35
30
Percentage
25
20
15
10
0
0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28
Diameter a (μm)
FIGURE 7.1 Size distribution of smectite-rich granules of MX-80 bentonite particles deliv-
ered by the American Colloid Company. Upper: sieved material in air-dry form. Lower:
material dispersed in distilled water and examined by transmission electron microscopy (a is
the longest diameter).
but the density is usually too low for using them as buffer in HLW repositories. This
is because natural pre-Quaternary smectite-rich clays with a present overburden of a
few to tens of meters have expanded through the years and are presently not denser
than 1800–1900 kg/m3.
There are no criteria postulated by authorities and regulators or specified in national
laws concerning the required density or composition of seals of clay blocks in any
type of waste repository, since it is the physical properties in terms of hydraulic
and thermal conductivities and expandability rather than the density that counts.
214 Bentonite Clay
2400
IBECO 100
2200 MX-80 100
SPV200 100
2000 Complete water
saturation, Sr = 100%
ρd (kg/m3)
1800
1600
1400
1200
5 10 15 20 25 30 35
Water content (%)
FIGURE 7.2 The dry density as a function of the water content for smectite-rich clays com-
pacted under 100 MPa uniaxial pressure. The fat curve represents the theoretical relation-
ship between water content and dry density at 100% degree of water saturation (Sr = 100%).
(Courtesy of Clay Technology AB, Lund, Sweden; After Johannesson.)
However, there are recommendations respecting the density and composition of clay
blocks for isolation of highly radioactive waste, with special respect to the perfor-
mance of the confining medium, which is usually the host rock of the repository. The
general design principle is that the hydraulic conductivity of the clay isolation shall
be lower than of the surrounding rock since groundwater will otherwise flow toward
and through the clay and waste rather than around them as illustrated by Figure 7.3.
It shows the redirection of groundwater flux that is achieved for the earlier described
silo repository for radioactive waste by the clay (SKB’s plant SFR at 50–100 m depth
in crystalline rock for disposal of low- and medium level radioactive waste). The
hydraulic conductivity of the rock and the surrounding clay isolation with 1100 kg/
m3 average dry density was estimated at E−8 and E−10 m/s, respectively. The con-
ductivity of the clay represented the case of complete cation exchange from Na to Ca
caused by dissolution of Portland cement in the concrete.
7.2.3 Preparation of Dense Blocks of Smectite Clay

Techniques for block compaction on an industrial scale were first tried in the late
1970s utilizing a method worked out by the Swedish enterprise ASEA Atom AB (now
ABB) with very high cold isostatic compression of granulates (Jacobsson and Pusch
1978). It was used for preparing large cylindrical blocks for the Stripa Project (Gray
1993; Pusch et al. 1982, 1985, 1994; Johannesson et al. 1995). Comprehensive pilot
studies of the pressure required to reach a certain density of granulated smectite clay
were made in the late 1970s in Sweden (Jacobsson and Pusch 1978), and by systematic
R&D by Bucher et al. (1982) in Switzerland. These investigations confirmed that a
pressure of about 100 MPa is required to achieve a dry density of 1900–2000 kg/m3
A
B B
C
D
E
F
50 m
H
G
30 m
FIGURE 7.3 Example of clay isolation (E in right figure) of silo repository for radioactive
waste of low and intermediate levels (SKB’s SFR storage). The left figure shows the differ-
ence calculated by finite element method in horizontal groundwater flux for the case of no
clay (upper) and the case when the space between silo and rock is filled with smectite-rich
clay (lower). (Data from Pusch, R., Waste Disposal in Rock, Developments in Geotechnical
Engineering, 76, Elsevier, New York, 1994.)
of a block of granules with a water content of up to 17%–19%, and that the density is
only insignificantly increased beyond ρd = 2100 kg/m3 even if pressures of 200 MPa
are used. Variation of the water content is known to have a significant effect on the
compaction curve and the work by Bucher et al. demonstrated that a water content of
about 5% gives the highest density of SKB’s reference material, MX-80, which is a
standard product of the American Colloid Company (cf. Figure 7.2). Very dry mate-
rial is known to require high pressures for reaching even moderate densities because
of the high strength of the granules and the high intergranular friction.
A completely new method for preparing blocks with desired degree of saturation
is to mix the granules with dry water (DW), that is, droplets of water confined in very
thin shells of a commercially available hydrophobic silicious substance (Forsberg
et al. 2014). On compaction, the shells break into submicroscopic fragments and
water becomes homogeneously distributed in the mass. The silica coatings prevent
the water droplets from combining during preparation of the mixture, even below the
216 Bentonite Clay
FIGURE 7.4 (See color insert.) A dry water particle in air-dry clay matrix (left) gives off its
water content (light blue) to the smectite particles (upper right), which expand and ultimately
yield a more or less homogeneous clay matrix with very finely fragmented silicious shells.
(Data from Forsberg, T. et al., Eng. Geol., 2014.)
freezing point. The substance can be produced on an industrial scale by exposing

volatile chlorosilanes to high temperature through flame hydrolysis and reaction
with methyl chlorosilanes after cooling. The so prepared powder is hydrophobic.
The specific surface area is 170–230 m2/gram, that is, about one-fourth of that of
montmorillonite clay. The fraction of the residual silanol content of the hydrophilic
silica is 25% and the carbon content is about 2.8% of the solid part of the DW.
On sufficiently large compression, the shells break and release water to the air-dry
clay particles that have been homogeneously mixed with DW particles (Figure 7.4).
The strongly hydrophilic clay matrix sucks up the water that will ultimately be uni-
formly distributed in the clay. The degree of saturation can be predicted to fit the
desired dry density.
Geotechnical data from testing of a smectite-rich clay saturated by the DW tech-
nique, which is patented, are not significantly different from those wetted by dif-
fusive uptake of water from the boundaries. Thus, for a dry density of 1320 kg/m3
(1850 kg/m3 at saturation) both DW clay and conventionally saturated clay had a
hydraulic conductivity of about E−12 m/s, and the same swelling pressure of about
1 MPa. The amount of remaining silica fragments, which are much smaller than
1 μm and hardly discernible in scanning electron micrographs, is less than 10% of
the solid clay constituents.
7.2.4 Blocks Prepared by Uniaxial Compression

This method is simple; it merely involves filling a form with granular clay and com-
pressing it to reach the desired dry density of the resulting blocks, which are coherent and
relatively strong (Figure 7.5). Commonly used compression pressures are 50–200 MPa.
FIGURE 7.5 (See color insert.) Blocks with 1.85 m diameter and about 0.4 m height
prepared by uniaxial compression. Left: solid block with black surface coating of carbon
disulfide for lubrication. Right: hollow block with space for a HLW canister. Weight about
2000 kg, water content about 10%. (Photos by Ramqvist.)
Cubical or rectangular blocks can be made by compression of clay granules in

forms for placement in niches cut in the rock around concrete plugs or in deep bore-
holes (cf. Figure 7.6). They expand on wetting and seal off the space very effectively
by forming an O-ring isolation around the plug. Chemical interaction with concrete
causes slow mutual degradation that needs to be taken into consideration, as dis-
cussed in Chapter 6.
Backfilling of tunnels can be made by stacking blocks as tightly as possible (70%–
90% by volume). This can be made very rationally in Tunnel-boring machine (TBM)-
bored tunnels and shafts but requires detailed planning of design and installation in
blasted tunnels, and trimming of the blocks. In both cases, the space between the
block matrix and the rock can be filled with blown smectite clay in the form of com-
pacted granules or pellets, or with pumped-in dense clay mud (Figure 7.7). Where
water flows into the tunnels, grouting of the rock before installation of the blocks is
required for avoiding piping by flow in the joints of the blocks.
Uniaxial compression of granules filled in a form for preparation of blocks gener-
ates friction along the form wall, which can cause some slight variation in void ratio
and density of the clay. One way of minimizing such heterogeneities is to lubricate
the form with form oil, or by smearing it with carbon disulfide (cf. Figure 7.5). Most
of the oil, which is sprayed on the form walls, evaporates after the compaction and
detailed microscopic studies and chemical analyses have not shown any signs of oil
in the blocks.
Uniaxial compression of big blocks generates shear stresses that can destabilize
the blocks and cause problems by fracturing during transport or when they are being
placed in deposition holes or elsewhere (Figure 7.8). According to SKB’s KBS-3V
concept, the space between the canister and the inside of the hollow blocks is only
10 mm, which can cause contact between the two, and fall of scraped-off clay frag-
ments that can prevent the canister to stand vertically on the plane top surface of the
lowest block in the hole (Figure 7.9). An alternative, safer design would be to make
the gap between the canister and the clay blocks wider for avoiding this problem and
instead reduce the gap between clay blocks and the rock by the same amount, provid-
ing about the same net density.
218 Bentonite Clay
Tunnel backfill
Retaining wall
Concrete/step 1
Concrete/step 2
Compacted bentonite
EDZ
Reinforcement
FIGURE 7.6 Blocks of smectite-rich clay for tunnel sealing. (EDZ is the blast-disturbed
zone of the tunnel wall.) Upper: concrete plug with clay seals constructed in SKB’s under-
ground laboratory, designed to resist 7 MPa axial pressure caused by the piezometric head
and swelling of the smectite backfill. Lower: a few of the about 600 dense blocks in the niche.
(Courtesy of Clay Technology, London; Pusch, R., Yong, R.N., Microstructure of Smectite
Clays and Engineering Performance, Taylor & Francis, London, 2006.)
7.2.5 Blocks Prepared by Isotropic Compression

Isotropic (isostatic) compression of clay granules using tools such as ASEA’s Quintus
press shown in Figure 7.10 causes isotropically to the clay body under compression.
Such blocks are therefore more homogeneous and have fewer defects than those pre-
pared by uniaxial compression in forms.
7.2.6 Microstructural Constitution of Compacted Smectite Clay

The effect of compaction of smectite granules for preparing dense clay blocks can
be imagined by considering their interaction, which has been modeled by numeri-
cal calculation techniques (Pusch and Yong 2006). The matter was in focus in
Fill of dense
clay pellets
Highly compacted
clay blocks
FIGURE 7.7 Regular block pattern in blasted tunnel complemented by blown clay pellets in
the open space between rock and blocks.
Blue color: compressive stress

Yellow color: moderate shear stresses
Red color: very high shear stresses
FIGURE 7.8 (See color insert.) FEM-calculated stress distribution of large block prepared
by uniaxial compression. The left side represents the vertical symmetry axis of the block. The
overstressing of the upper edge causes fissuring of this part that can fracture and fall off at
transporting and installation of the block, or when the heat pulse comes from the hot canister.
(After Börgesson.)
220 Bentonite Clay
Canister being placed

in deposition hole
EDZ
Gap between buffer blocks

and rock filled with smectite Blocks of highly compacted
pellets according to SKB’s smectite clay
concept KBS-3V
FIGURE 7.9 Placement of blocks of highly compacted clay and canister in deposition holes
(KBS-3V concept). EDZ is the blast-disturbed zone that is water-bearing and supplies the
rock surrounding the deposition hole with water. (Data from Pusch, R. et al., High-Level
Radioactive Waste Disposal, WIT Press, Southampton, 2011.)
Chapter 5 and we recognize in Figure 7.11 the unit cell of a mixture of granules

with two diameters (0.10 and 0.35 mm) compressed uniaxially under 100 MPa
pressure. The compressive strain changed the density of the granules and the shear
stresses in them became so high—more than 100 MPa—that granules in contact
were welded together. The ultimate dry density of the unit cell after hydration of
the granules was about 1850 kg/m3 corresponding to 2050 kg/m3 at water satura-
tion. The dry density of the granules to the compressed granules varied from about
1850 kg/m3 in the center of the big granules to a few hundred kg/m3 in their periph-
eral parts. Fragments from the sheared granules accumulate in the voids and form
clay gels of varying density after water saturation.
The net effect of the hydration and expansion of the compacted granules is a
clay matrix with largely varying density and with microchannels of varying cross-
sectional area as illustrated in Figure 7.12. The tortuous, more or less continuous
microchannels are filled with clay with lower density than the granules. This model,
being based on microscopy analyses of sectioned samples of montmorillonite-rich
clays percolated by methylene blue solutions, has given the relationship between
bulk density at water saturation and percentage of channel cross-sectional area in
Figure 7.13 (Pusch 2008).
Use of very dense granules would be expected to give a more homogeneous
microstructure and higher expandability of compacted clay after saturation, but
this is not always the case, as demonstrated by a study in which such granules
FIGURE 7.10 ASEA’s Quintus press for high isostatic (HIPOW) compaction. The granular
clay mass to be compressed is enclosed in a flexible form that is located in oil or water for
cold compression. Hot compression is commonly used for preparing ceramic isolators but is
of potential use for manufacturing copper canisters for isolating HLW. This requires use of
an inert, highly pressurized gas such as argon. (Courtesy of ABB.)
were used for preparing blocks. Smectite-rich clay (MX-80) granules dried to
1.9%–2.3% water content, giving them a dry density of more than 2100 kg/m3,
were compacted under 200 MPa pressure for preparing blocks with a bulk dry
density of 1420–1555 kg/m3 (Pusch et al. 2003). In the course of hydration, the
swelling pressure quickly rose to 8 MPa for the densest sample and to 4 MPa for
a softer one, increasing then to a maximum value of 12 and 6 MPa, respectively,
after about 10 weeks. For clays manufactured using ordinary MX-80 granules with
about 10% water content and compacted to these densities, the swelling pressures
would be 3.5 and 1.5 MPa, respectively, for the two densities. The abnormal pres-
sures for the high-density clay granules remained after 10 weeks, indicating that
microstructural equilibrium persisted and that no further homogenization could
be expected. This was confirmed by the 50%–100% higher hydraulic conductivity
222 Bentonite Clay
FIGURE 7.11 (See color insert.) Unit cell with one-eighth of a big grain (0.35 mm) contacting
one small grain (0.10 mm). The edge length of the cubical unit cell is 0.175 mm. (Data from
Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering Performance,
Taylor & Francis, London, 2006.)
0.1 mm
FIGURE 7.12 (See color insert.) Schematic cross section of the microstructure of fully
matured smectite-rich clay with a bulk density at water saturation of 1800–2000 kg/m3. For
the lower density the bar is 0.1 mm and for the higher 0.01 mm.
of the clays with very dense granules than of ordinarily prepared MX-80 clay. The
lesson learned was that preparation of blocks from very dense smectite pellets or
granules does not make the blocks very homogeneous and low permeable. This is
because highly compressed stacks of smectite lamellae simply do not expand read-
ily and uniformly on hydration (Figure 7.14).
1.75% 16
14
12
Percentage
10
8
6
4
2
0
1500 1600 1700 1800 1900 2000 2100 2200
Density at water saturation (kg/m3)
FIGURE 7.13 Permeable microchannels in compacted smectite-rich clay (MX-80). Left:

photo of a 30 mm diameter specimen prepared and saturated with distilled water to a den-
sity of 2100 kg/m3 and then percolated with methylene blue solution. The dark spots show
the presence and distribution of the microchannels after drying. The white dendritic pattern
represents desiccation fissures developed at the subsequent drying. Right: percentage of chan-
nel cross-sectional area of total sectioned area. (Data from Pusch, R., Geological Storage of
Radioactive Waste, Springer, Berlin, Germany, 2008.)
FIGURE 7.14 Transmission electron micrograph of ultrathin section of clay with a dry den-
sity of 1250 kg/m3 impregnated with acrylate. The very strongly compressed MX-80 pellets
formed dense, largely spherical aggregates and not the wavy network of stacks seen in MX-80
clay blocks prepared from ordinary granules (cf. Figure 3.2). Edge length of the photo 10 μm.
(Courtesy of Kasbohm.)
224 Bentonite Clay
7.3 CLAY LINERS, MATERIALS, AND

PRINCIPLES OF CONSTRUCTION
7.3.1 Principles of Design and Construction
The primary aim of liner systems in landfills of hazardous waste is to confine, iso-
late, and secure the wastes contained by such barriers. The garbage bag principle
adopted in the design and construction of waste landfills satisfies the primary aim.
Figure 7.15 shows such a system of disposed waste implying construction of two
clay-based liners, the upper one serving to delay infiltration of rain and water from
thawed snow into the waste mass, and the lower designed to delay migration of
contaminated water from the waste to the surroundings. The entire system in this
example implies underground placement but still above the groundwater level. One
realizes that this criterion is difficult to fulfill in the light of the ongoing climatic
changes; already today, daily precipitation of more than 100 mm/day can occur and
there can be much more in the future.*
7.3.2 Criteria
Since the required time for isolation of hazardous waste is limited, presently 300
years according to EU directives, the clay liners do not have to be permanently tight
but be composed and constructed, so that the amount of hazardous ions does not
exceed certain limits in the prescribed period of time. The essential parameters for
the clay liners are as follows: (1) water transmissivity, (2) expandability, (3) physical
stability (4) constructability, and (5) cost.
Waste pile Water impermeable cover

Clay-based engineered
barrier
Natural soil base—may be prepared as

compacted and permeability-restricted zone
Filter, membrane and leachate collection, and drainage system
Clay-based engineered barrier
Engineered barrier system
FIGURE 7.15 Garbage bag system in shallow disposal of moderately dangerous waste.
(Data from Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering
Performance, Taylor & Francis, London, 2006.)
* 110 mm fell in Malmö in southern Sweden in one day, August 30/31, 2014.
The rate of percolation of the top liner, which may require tens of years or more
to become water saturated, ultimately determines the downward transport of ions
released from the underlying waste to the underground. The liner should be as tight
as possible for minimizing transmissivity, which would call for a high density. This
will, however, give a high swelling pressure and the risk of expansion and softening,
and the density and composition of the liner must hence be selected such that this
pressure does not exceed the effective overburden pressure.
The problems with the concept in Figure 7.15 are manifold. First, heavy, lasting
rain and snow cover will raise the general groundwater level in the terrain with risk
of flooding, and create a groundwater level in the cover of the top liner that will cause
wetting and percolation of the liner and initiate saturation of the waste mass, which
will start to dissolve and release contaminants to the groundwater. Second, leachate
collection may not work for longer periods than a few decades depending on the per-
manency of the organization responsible for this service and for handling the leachate.
Third, the waste mass will anyhow ultimately be saturated with contaminated water
requiring the excavation and disposal of it elsewhere, leaving the area contaminated
and unsuitable for housing, gardening, or construction of schools and swim stadiums.
Fourth, wetting of the waste can generate gas that can create breakage of the top liner
and reduce its tightness by which penetration of surface water can strongly increase.
Other concepts for isolating hazardous waste on the ground surface are therefore
being considered in recent times (Figure 7.16). The principle followed is to locate the
Hill-type repository on moraine ridge or similar: concrete

cells enclosed in clay covered by erosion-resisting cap
Groundwater level
Overburden of coarse material,

rock blocks on top over
successively finer gravel, sand,
and silt
Top liner of smectite clay
Waste
Permeable
foundation Drainage
Bottom liner of
smectite clay
FIGURE 7.16 Principle of hill-type repository for low-level radioactive waste.

226 Bentonite Clay
landfill on elevated ground, hill-type landfills, with guaranteed run-off conditions of

surface water, and to equip it with a top cover that lets almost no water through in
the specified period of time, which is 300 years for low-level radioactive waste. Such
waste requires confinement in concrete vaults, whereas less dangerous inorganic rest
products like ash from incinerated organic waste can be placed on a suitable foun-
dation bed and covered with a stable and tight overburden, still keeping the basic
requirement of erosion protection in mind.
Also, this kind of disposal gives difficulties. First, effective drainage of the dis-
posal area requires that the topography is suitable and that drainage tubes have suf-
ficient drainage capacity and are made of ceramics (Motiejunas et al. 2005). Second,
clay top liners need to be covered by erosion-resistant fillings for draining off inflow-
ing water, and be sufficiently thick to prevent freezing of the underlying clay and silt-
sand filters. Third, the weight of the overburden must be sufficient to resist the upward
acting swelling pressure of the clay in the top liner. Fourth, steps have to be taken for
preventing ions released from the waste to migrate to the underlying soil or rock via
the bottom liner. This involves exchange of the original sorbed cation, commonly Na,
by cations that can raise the hydraulic conductivity and degrade the clay by chemical
attack, which is the case if pH of the percolate is lower than 7 and higher than 10.
An example of a relatively simple disposal design is represented by a site at
Högbytorp, north of Stockholm, owned and operated by the Swedish waste-handling
company Ragn-Sells AB. Organic waste to be placed is incinerated and the ash is
mixed with cement slurry and filled and compacted on site, yielding concrete of
considerable strength and negligible compressibility under own weight plus several
meters of overburden. The waste mass of such landfills can be several tens of meters
high, placed in campaigns with temporary cover of geotextile and tight plastic sheets
with welded joints.* We will take such a landfill as an example for describing its
function in both short- and long-term perspectives.
7.3.3 Principle of Placement and Compaction

A fundamental question in road and dam building is whether the water content
should be at optimum for reaching the highest possible density or whether a lower
water content should be used. The matter is naturally very important for getting a
clay liner that is as tight as possible but also for the construction cost. Traditionally,
one prefers to choose the water content, so that it becomes equal to the optimum
water content, which is on the order of 10%–12% for moderately smectite soils, but
significantly lower water content can give an even higher density (Figure 7.17). The
controlling factor is the compaction energy.
The reason for the two density maxima in the graph is as follows. When water
is added to an air-dry mixture of nonexpandable frictional soil and smectite clay,
the clay hydrates forms coatings of the coarse grains. These deform easily when the
amount of water is high enough to make the clay a lubricant, hence representing
the conventional optimum water content. For higher water contents, the porewater
* Such organic shields have a limited lifetime and can serve as slip surfaces when degraded. They should
therefore be removed when additional layers of clay are about to be placed.
80% crushed TBM muck + 20% bentonite

2.4
2.3 Proctor
Vibro
2.2 MCV
Dry density (g/cm3)
2.1 Sr = 100%
2.0
1.9
1.8
1.7
1.6
0 5 10 15 20
Water ratio (%)
FIGURE 7.17 Compaction curves of a mixture of 20% MX-80 clay and crushed rock using
different compaction techniques. The optimum water content is 8%–10% but the highest den-
sities were obtained for mixtures with very low water contents. (Data from Pusch, R., Yong,
R.N., Microstructure of Smectite Clays and Engineering Performance, Taylor & Francis,
London, 2006.)
overpressure built up by the compaction will prevent effective approach of the aggre-
gate grains. For very low water contents, clay granules behave like quartz grains and
combine with these to move into the closest possible layering. This will give the most
homogeneous microstructural constitution of the clay component after wetting of
the compacted mixture and a lower hydraulic conductivity compared with mixtures
compacted at higher water contents. Practical application of the technique, which
is termed dry mixing and compaction, has yielded dry densities of up to 2300 kg/m3
at field compaction of mixtures of 10% MX-80 clay and suitably graded silt/sand
material. In terms of microstructure, this corresponds to a dry density of the clay
component of about 1100 kg/m3 or slightly higher.
Compaction of raw material excavated from a thick clay bed of Triassic age in
Lithuania (Motiejunas et al. 2005; Pusch et al. 2011) and stockpiled for drying in
open air for one year reduced the water content to 19%–20% (Figure 7.18), and gave a
dry density of about 1650 kg/m3, corresponding to about 2000 kg/m3, of a compacted
fill at water saturation. This was achieved by moderate compaction effort provided
by a vibrating plate of 420 kg for 11 cm layers, and a 7 t vibratory roller for layers
with 17 cm thickness.* The mineral composition was in the following order start-
ing with the major constituent: (1) Fe-rich chlorite, (2) illite/smectite (I/S)-mixed
layer minerals, (3) well- to poorly ordered illite, and (4) kaolinite (two types),
feldspar, quartz, and traces of calcite. The percentage of illitic layers in the I/S var-
ied between 40% and 85%, indicating that the fraction of expanding components of
the total mineal mass was relatively small, that is, 15%–25%. At water saturation,
* Chunks bigger than about 50 mm were removed.

228 Bentonite Clay
FIGURE 7.18 (See color insert.) Stockpiled Lithuanian clay for drying under open condi-
tions. At rainfall, only the uppermost few centimeters become wetted and hinder further
downward penetration. Removing this very shallow layer gives access to unaffected material.
representing a density of 2000 kg/m3 and a water content of 29%, the clay has a
hydraulic conductivity of 6.0E−11 m/s for percolation with low electrolyte water.
The principle of using slightly dried but otherwise unprocessed clay material, was
further investigated and applied in the construction of the bottom liner of Ragn-Sells
landfill in Högbytorp. For this work, a I/S mixed-layer clay was used, stemming
from the earlier mentioned 100 m thick homogeneous sediment mass of Tertiary
age in northeastern Germany (Friedland clay). It was selected because of the com-
prehensive experience from German landfill construction projects and systematic
laboratory testing of this clay in Sweden. The content of clay-sized particles is about
90%. They consist of irregularly stacked muscovite-montmorillonite with about
30%–40% expandables. Its physical behavior under different conditions, including
temperature gradients, was described in Chapters 5 and 6. Open pit mining of the
other I/S mixed-layer Holmehus clay, discussed in Chapters 2, 4 and 6, is illustrated
in Figure 7.19, which also shows the material stockpiled at the mineral processing
plant of Dantonite A/S in Denmark.
The maximum allowed hydraulic conductivity, specified by the owners of the
Högbytorp disposal site, was E−10 m/s. The basis for selecting the correspond-
ing density was a series of oedometer tests with various porewater compositions
(Figure 7.20). One finds from the diagram that the hydraulic conductivity for low
electrolyte porewater is about 2E−11 m/s for a density at saturation of 1950 kg/m3.
For a density of slightly more than 2000 kg/m3, the conductivity at percolation of
such water was 5E−12 m/s and for saltwater (3.5% CaCl2 solution) about 5E−11 m/s.
The tests were made at room temperature, which slightly overrates the conductivity
in the field. Considering also that the hydraulic gradient will be lower in practice
than in the oedometer tests (roughly 50% lower), the decision was taken to compact
FIGURE 7.19 (See color insert.) Holmehus clay. Upper: view of open pit mining in
Denmark. The precious clay has an overburden of nonsmectite clay of about 10 m. Lower:
stockpiled Holmehus clay stored outdoor for drying. (Courtesy of Dantonite A/S.)
1.00E−07
20% CaCl2
10% CaCl2
1.00E−08 3.5% CaCl2
20% NaCl
Hydraulic conductivity (m/s)
10% NaCl
Dist.
1.00E−09
1.00E−10
1.00E−11
1.00E−12
1700 1750 1800 1850 1900 1950 2000 2050 2100
Density (kg/m3)
FIGURE 7.20 Hydraulic conductivity of fluid-saturated Friedland clay as a function of the

electrolyte content. (Courtesy of Clay Technology AB, Lund, Sweden; Pusch, R., Yong, R.N.,
Microstructure of Smectite Clays and Engineering Performance, Taylor & Francis, London,
2006.)
230 Bentonite Clay
2600
2500 Theoretical maximum
2400 Experimental
2300
Dry density (kg/m3)
2200
2100
2000
1900
1800
1700
1600
0 5 10 15 20 25
Water content (%)
Fractions (mm) Percentage of fractions
2–8 20.0
1–2 20.4
0.1–1 42.4
<0.1 17.2
Total 100
FIGURE 7.21 Compaction tests of clay with the grain size distribution summarized
in the table. The fat curve represents the theoretical maximum density. (Courtesy of Clay
Technology AB, Lund, Sweden.)
the liner of Friedland clay to a dry density of at least 1610 kg/m3, corresponding to a

density at water saturation of 2000 kg/m3.
A particularly valuable property of mixed-layer clays of this type, that is, with the
relatively small amount of smectite being uniformly distributed in the clay, the dry
density at compaction is not significantly affected by the water content (Figure 7.21).
This implies that the material does not require extensive processing such as drying
and wetting for reaching maximum density. The optimal water content is hence not
well defined and untreated raw material can be used as long as the water content is
below the figure representing complete water saturation, that is, about 22% for this
clay at a dry density of about 1600 kg/m3. For reaching a high degree of homogene-
ity of the compacted clay, large and dense chunks should be removed in conjunction
with layer-wise placement and distribution of the soil on the construction site.
At the Högbytorp site the clay was placed in 25 cm layers each compacted by
six runs in perpendicular directions using a 19 t vibratory padfoot roller. The upper
surface was then evened by a few additional runs of a smooth roller. Samples of dif-
ferent sizes were taken systematically for getting representative results, which are
TABLE 7.1
Data from Analysis of Samples
Calculated Sr, Degree
Volume Water Dry Density Density at Water n, Porosity of Water
Layer (dm3) Content (%) (kg/m3) Saturation (kg/m3) (%) Saturation
1 (first) 1.178 22.55 1612 2016 58 70
1 (first) 1.178 22.19 1615 2010 64 71
2 (second) 1.178 22.79 1615 2022 58 87
3 (third) 1.178 19.16 1655 2003 60 71
3 (third) 1.178 23.03 1612 2001 59 80
collected in Table 7.1. The placement and compaction were made under dry weather
conditions and the completed surfaces were sealed by high-density polyethylene
geomembrane for preventing wetting in the construction phase.
The main outcome of the compaction campaign was that the three layers were
uniformly and effectively compacted, and that the average density at water saturation
was 2010 kg/m3, that is, above the design value.
The pictures in Figures 7.22 through 7.24 illustrate the effect of heavy compac-
tion of Friedland clay having a natural water content of about 26%–29% in a German
project. This gave complete saturation much like the condition at compaction of the
same clay at the Högbytorp site.
Another mixed-layer clay of similar composition and potential use is the Tertiary
Roesnaes clay from eastern Denmark extensively used for liner construction and
considered as a candidate material for a landfill deposit for which the hydraulic crite-
rion was set at K < 2E−11 m/s and the maximum swelling pressure at 500 kPa when
saturated and percolated with tap water and 3.5% CaCl2 solution. It was found, as illus-
trated by Table 7.2, that these demands were fulfilled for a dry density of 1270 kg/m3
FIGURE 7.22 Clay noodles of the first layer before compaction. The water content was 29%
and the hydraulic conductivity at saturation and percolation with distilled water 1.9E−10 m/s.
232 Bentonite Clay
FIGURE 7.23 Compaction by vibrating padfoot BOMAG-roller. The water content was 26%
and the hydraulic conductivity at saturation and percolation with distilled water 2.1E−10 m/s.
(Courtesy of Durtec GmbH, Neubrandenburg, Germany.)
FIGURE 7.24 Homogeneous nature of compacted Friedland Ton. The water content of
the uppermost of the three layers was 29% and the hydraulic conductivity at saturation and
percolation with distilled water 1.1E−10 m/s. (Courtesy of Durtec GmbH, Neubrandenburg,
Germany.)
TABLE 7.2
Laboratory Data from Tests on Fully Water Saturated Roesnaes Clay
Sampled in Undisturbed Form in the Field
Material/Water/ Hydraulic Swelling Density (ρs/ρd)
Test No Solution Conductivity (m/s) Pressure (kPa) (kg/m3)
1 Undisturbed/DV 2E−11 145 1850/1350
2 Undisturbed/SV 3E−11 145 1890/1410
DV—distilled water; SV—3.5% CaCl2 solution; ρs—density at water saturation; ρd—dry density.
(1800 kg/m3 at water saturation). The data are similar to those of Friedland clay
with 1590 kg/m3 dry density (2000 kg/m3 at water saturation). The differences are
explained by the higher content of smectite of the Danish clay.
7.3.4 Microstructural Constitution of Compacted Smectite Clay Liners

As we saw from the various tables describing the composition of moderately smec-
tite clays, they all contain grains of rock-forming minerals, primarily quartz and
feldspars, so the difference in microstructural constitution of dried, milled, and com-
pacted natural clays such as the Friedland and Roesnaes brands, as well as the earlier
described Lithuanian clay, is not very different from that of mixed smectite-rich clay
and coarser soil (Figure 7.25). The detailed distribution of smectite is quite differ-
ent, however: practically every little soil particle contains smectite in the unmixed
clays, which gives them a more homogeneous clay matrix with constant density than
mixtures of smectite clay and sand/silt as described in Chapter 4.
Figure 7.26 illustrates the microstructure of mixtures of nonexpandable grains
and smectite granules in air-dry form and hydrated to form a clay matrix with vary-
ing density as indicated in Figure 7.12.
7.3.5 Microstructural Modeling of Smectite Clay

The practical use of microstructural modeling is for explaining the ways in which water
and gas percolate through artificially prepared smectite clay with different bulk densi-
ties. Microstructurally relevant models tell how water, gas, and dissolved ions move
through clays, as demonstrated by comparing predicted and actually recorded bulk
hydraulic conductivities that we saw in Chapter 5 and also how externally applied
stresses are transferred and distributed, which was exemplified in the section of creep
D
G
G
FIGURE 7.25 Schematic picture of the microstructure of clay/aggregate mixture with

low compressibility and low hydraulic conductivity. G—coarse grains; D—clay powder
grains. (Data from Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering
Performance, Taylor & Francis, London, 2006.)
234 Bentonite Clay
Ballast grains Ballast grains

Clay granules Clay gel
FIGURE 7.26 Generalized section of the microstructure of mixed smectite clay and coarse non-
expandable mineral (ballast) grains. The left drawing shows the constitution of mixed dry constit-
uents and the right the system after hydration, expansion, and homogenization. The density of the
clay component is, in fact, strongly varying. (Data from Pusch, R., Yong, R.N., Microstructure of
Smectite Clays and Engineering Performance, Taylor & Francis, London, 2006.)
FIGURE 7.27 Schematic view of the Neretnieks/Moreno 3D model concept of hydrauli-

cally interacting microchannels.
strain. In the future, more advanced modeling of these performances requires defini-
tion of the three-dimensional microstructural constitution, which can be generalized, as
shown in Figure 7.27, respecting the hydraulic conductivity. The microchannels men-
tioned earlier in this chapter and in Chapters 4 and 5 represent pervious paths in the clay
and are characterized by their lengths, widths, apertures, and transmissivities. The rest
of the clay matrix is porous but practically impermeable. The basis of the development
of the present model is the code 3Dchan (Neretnieks and Moreno 1993) (cf. Chapter 4).
Calculation of the bulk hydraulic conductivity can be made by assuming that
a certain number (commonly 6) of channels intersect at each node of the orthogo-
nal network and that each of them consists of a bundle of capillaries with circular
cross-section. We saw in Chapter 5 that use of this model gives fair agreement with
experimental data for clay seals of expandable clay percolated by electrolyte-poor
and saltwater. Further use of the model is for estimating the distribution of gas pene-
tration of water-saturated clay: gas follows the paths of least resistance, implying that
a few major channels are penetrated by gas that consolidates the clay gels in them
and widens them. Propagation of the gas at its tip takes place when the gas pressure
exceeds the counter-pressure, which can be interpreted as the swelling pressure and
has to be overcome in order to displace the clay matrix.
7.4 CLAYS FOR ROCK GROUTING

7.4.1 Use of Grouts with respect to Their Physical Stability
7.4.1.1 Argillaceous Cement Grouts
Backfilling of drifts and shafts in mined repositories at a depth of 300–400 m can be
very difficult because of high local water inflow where fracture zones are intersected.
Grouting is therefore counted for which purpose cement-based materials are consid-
ered in the first place. The very high hydraulic gradients in the rock that are built up
in the course of boring or blasting drifts and tunnels makes grouting of the completed
rooms ineffective; pregrouting is the only practical and safe way of tightening the
rock. The problem of piping is not met in preparation of very deep holes (VDHs) since
no significant hydraulic gradients prevail in them at any time. For mined repositories
the grouts in sealed fractures will be exposed to hydraulic gradients ranging from 20
to 200 m/m (meter water head per meter flow length), which can cause piping and ero-
sion of clay grouts. Refreshing our minds, we recall from Chapters 4 and 5 that pure
montmorillonite grout with a density of 1100–1200 kg/m3 fails by piping at a gradient
of 30–50 m/m (Pusch and Yong 2006). Cement grouts in rock hosting shallow and
medium-deep repositories can withstand such gradients if the time for curing is suffi-
ciently long but they will be successively eroded and soon lose their tightening effect.
In recent times, new grouts consisting of cement mixed with expansive clay, such
as montmorillonite-rich clay or mixed-layer clay, have been proposed and tested.
The mutual degradation of contacting Portland cement and expansive clay miner-
als, caused by the very high pH of the cement porewater, requires use of low-pH
cement. Similarly, degradation of conventional organic fluidizers in the cement calls
for stable inorganic materials for which talc has been found suitable. Grouts based on
Friedland clay mixed with 5% low-pH cement to a water content of about 120% is suf-
ficiently low-viscous for grouting. They can be pumped over considerable distances
for injection and stiffen fairly rapidly. The shear strength rises from 1 to 10 kPa
in one day and to 100 kPa in one week. Recipes for denser argillaceous cement
grouts, that is, 1500–1700 kg/m3, with good fluidity (viscosity 150–200 mPas) can
be prepared using finely milled quartzite/silica powder for aggregate (60%–65%)
236 Bentonite Clay
of material finer than the 100

Cumulative percentage
80
respective size
60
40
20
0
10 100 1000
Particle size (μm)
FIGURE 7.28 Grain size distribution of crushed and ground quartzite as the basic compo-
nent of dense grouts. Photo of grout ready for injection or casting. (Data from Mohammed
et al. 2014.)
(cf. Figure 7.28), low-pH cement (8%–10%), and talc (27%), with the water/cement
ratio being in the interval 7–8 (Mohammed et al. 2014).
The shear strength of cementitious grout with a density of 1314 kg/m3 and palygor-
skite added for quick thixotropic hardening was 0.75 kPa after 6 hours and 3.2 kPa
after 30 hours. The optimal content of palygorskite is estimated to be 15% for reach-
ing a reasonably high packing degree and significant particle-filtering capacity. The
thixotropic character of palygorskite-charged grouts has been documented by tests
that include rapid oscillatory shearing of the grout. They have shown that the fluid-
ity is high for both low and high shear rates, which can make it suitable for dynamic
injection (Pusch 1994; Mohammed et al. 2014).
7.4.1.2 Role of Palygorskite

The need for quick stiffening has led to recipes that include the clay mineral palygor-
skite (attapulgite). It is strongly thixotropic and commonly used in drilling muds.
The chemical formula is (Mg,Al) 2Si4O10(OH)4(H2O), and it is chemically stable
in saltwater and in contact with montmorillonite. Once forced into fractures, such
grouts stiffen and have a filtering function, so that it blocks grouted fractures from
intrusion or loss of fine particles by arching of the needle-shaped particles. The min-
eral particles have a length of up to 2 µm and appear as very thin rods in the electron
TEM micrograph in Figure 7.29.
The problem with grouts of pure clay is their softness. This makes them sensitive
to piping and erosion once they have been injected into rock fractures, where they are
exposed to a high hydraulic gradient that can extrude them or cause piping. Use of
packers for supporting a grout of smectite clay with 1140 kg/m3 density for one day
means, by referring to laboratory experiments, that the shear strength of the grout will
increase to 2.4 kPa in 20 seconds and to 2.7 kPa in one minute. Systematic testing
of grout materials consisting of smectite-rich clay (MX-80) mixed with very finely
milled quartz powder has shown, as expected, that the hydraulic gradient is the con-
trolling parameter for piping that takes place at a certain critical water pressure that
is a function of the length of a grout-filled channel in a rock fracture, while there is
no pressure at its opposite end. One finds, for instance, that a channel with 3 m length
filled with clay with a density of 1100–1200 kg/m3 will be unstable if the water pres-
sure on the clay is about 1 MPa and the outer end of the channel has no pressure at all.
In practice, channels in rock often have the nature shown in Figure 7.30.
0.50 μm
×5000 120 kV 22.02.2005

Microscope: J1210
Operator: Manfred Zander
Preparation: Suspension
Sample: JK41 Sapin
FIGURE 7.29 Palygorskite morphology. Some montmorillonite is also present. (Courtesy

of Kasbohm, Greifswald University, Greifswald, Germany.)
Fracture II
Fracture I
Rock matrix
Contact points with
permeable surrounding
of fracture mineral
Fracture I Fracture II
FIGURE 7.30 Channel formed by the intersection of two fractures in rock.

238 Bentonite Clay
7.4.2 Penetrability and Sealing Function of Clay-Based Grouts

Laboratory and field experiments with cementitious grouts have given a number of
results that are valid in principle also for injecting clayey grouts:
• Effective penetration of grouts into fractures with smaller aperture than

100 µm requires that the viscosity is lower than 50 mPas. For injection
under static, constant pressure, the maximum density for penetration of
grouts based on low-pH cement, talc, and silica-rich aggregate is estimated
to be 1200–1300 kg/m3.
• Effective penetration of grouts into fractures with an aperture of 100–500 µm
can be achieved for a viscosity of 50–1000 mPas. For injection under dynamic
(pulsed) pressure, the maximum density of grouts containing low-pH cement,
talc, and silica-rich aggregate is estimated to be 1300–1400 kg/m3.
• Grouts with a viscosity of 1,000–50,000 mPas can enter fractures with
apertures larger than a few millimeters. The maximum density of grouts is
estimated to be 1400–1600 kg/m3.
Full-scale application of clay-grouting of granitic rock was made in the international

Stripa Project in the 1990s for investigating methods for effective isolation of highly
radioactive waste (Gray 1993; Pusch 1994). Boreholes with 760 mm diameter were
grouted using megapackers for injecting montmorillonite-rich grout with a density
of 1080 kg/m3 and grout with 50/50 montmorillonite/quartz powder with 1200 kg/m3
density. The average hydraulic conductivity of the grouted rock was found to be
6E−10 m/s, whereas ungrouted rock had 5E−9 m/s (Pusch 1994). This campaign
was followed by inserting a heater in the holes and exposing the grouted rock to a
temperature of 100°C for three months after which the hydraulic conductivity was
measured again. This showed that a significant part, that is, about 50% of the gained
reduction in average conductivity, had been lost. The reason for this was believed to
be mechanical degradation of the grout by heat-induced movements of contacting
rock blocks.
Experiments were also made by injecting Li-montmorillonite grout with a den-
sity of 1175 kg/m3 under 4 MPa pressure into fractures with an estimated aperture
of 100 μm. These tests led to the conclusion that the average transmissivity of the
grouted rock had dropped by 40% (Pusch 1994). This very moderate improvement
is believed to be due to the redirection of the water flux from the grouted fracture to
several minor ones that can have been activated by rock displacements caused by the
injection but had not become sealed.
A practically important conclusion from the various experiments and most other
published results from rock sealing projects is that tightening by grouting is not
very effective, especially not where the grouted rock will be heated, like in deposi-
tion holes for heat-producing highly radioactive waste. However, strategic location of
grout injection holes and grouting in campaigns, starting with the inner parts of the
injection holes and using relatively high injection pressures, can be rather successful.
Dynamic grouting is preferred.
7.5 GENERAL ASPECTS ON SELECTION AND USE

OF SMECTITE CLAYS FOR WASTE ISOLATION
Some investigators state that only smectites with a very low content of accessory
minerals and organics should be considered for use as seals. The opinion of expe-
rienced engineers and designers is more liberal: only clays with lower content of
expandables than 50% and a mineral composition that can damage canisters by cor-
rosion, or with more organics than 2000 ppm should be avoided. This opinion has to
be reconsidered, however, when assessing the degradation of the clay that can take
place under the hydrothermal conditions on the respective site. We considered this
item in Chapter 6 and conclude here, on the basis of all matters considered in almost
all of the chapters in the book, that
• For underground disposal of particularly hazardous chemical waste and

radioactive matter, the essential thing is to isolate them from the groundwa-
ter. This can be made by use of dense smectite clay: montmorillonite-rich
material when the temperature is lower than about 100°C and more stable
material like saponite clay for higher temperature. Where the disposal holes
or rooms are intersected by water-bearing fracture zones, grouting prior
to completing the construction is suitable, for which purpose cement-poor
material is recommended.
• For both underground and on-ground disposal of hazardous waste, concrete
has to be used: for filling VDH disposal holes and tunnels where they are
intersected by water-bearing fracture zones, and for storage of the waste in
vaults on-ground. Chemical interaction of ordinary concrete and smectite
causes mutual degradation but concrete with a minimum content of low-pH
cement and talc as fluidizer strongly reduces the risk for comprehensive loss
of the tightness of the contacting seals.
• For on-ground disposal of moderately hazardous chemical waste and
radioactive rest products, top liners of effectively compacted, moderately
smectite-rich clay is best. Sufficiently thick top liners with appropriate
overburden will hinder downward penetrating rain- or meltwater to saturate
the waste and make tight bottom liners needless.
REFERENCES
Al-Thaie, L., Al-Ansari. N., Pusch, R., Knutsson, S. 2012. Proposed selection criteria for
hazardous waste disposal in Iraq. WIT Transaction on Ecology and The Environment,
Vol. 163, WIT Press, Southampton.
Bucher, F., Jeger, P., Kahr, G., Lehner, J. 1982. Herstellung und Homogenität hochverdichteter
Bentonitproben. NAGRA Technischer Bericht 82-05. Wettingen, Switzerland.
Forsberg, T., Pusch, R., Knutsson, S. 2014. Rational method for preparing dense clay blocks
with desired degree of water saturation. Engineering Geology.
Gray, M.N. 1993. OECD/NEA International Stripa Project Overview, Vol III: Engineered
Barriers. Swedish Nuclear Fuel and Waste Management Co (SKB), Stockholm,
Sweden.
Grim, R.E. 1967. Clay mineralogy. McGraw-Hill, New York.
240 Bentonite Clay
Jacobsson, A., Pusch, R. 1978. Egenskaper hos bentonitbaserat buffertmaterial. KBS Teknisk
Rapport 32, SKB, Stockholm, Sweden.
Johannesson, L.-E., Börgesson, L., Sandén, T. 1995. Compaction of Bentonite Blocks. SKB
Technical Report 95-19. SKB, Stockholm, Sweden.
Mohammed, H.M., Pusch, R., Knutsson, S. 2014. A talc-based cement-poor concrete for seal-
ing boreholes in rock.
Motiejunas, S., Pusch, R., Jonynas, J., Cyziene, J., Satkunas, J. 2005. Investigation of Triassic
clay for use as radioactive waste confinement. Proceedings of the 2nd International
Meeting on Natural and Engineered Barriers for Radioactive Confinement, Tours,
France, pp. 305–396.
Neretnieks, I., Moreno, L. 1993. Fluid flow and solute transport in a network of channels.
Journal of Contaminant Hydrology, 14:163–192.
Pusch, R. 1994. Waste Disposal in Rock. Developments in Geotechnical Engineering, 76.
Elsevier, New York.
Pusch, R. 2008. Geological Storage of Radioactive Waste. Springer, Berlin, Germany.
Pusch, R., Bluemling, P., Johnsson, L. 2003. Performance of Strongly Compressed MX-80
Pellets under Repository-Like Conditions. Applied Clay Science, 23:239–244.
Pusch, R., Börgesson, L., Nilsson, J. 1982. Buffer Mass Test—Buffer Materials. Stripa Project
Internal Report 82-06, SKB, Stockholm, Sweden.
Pusch, R., Nilsson, J., Ramqvist, G. 1985. Final Report of the Buffer Mass Test—Volume I:
Scope, Preparative Field Work, and Test Arrangement. Stripa Project Technical Report
TR 85-11, SKB, Stockholm, Sweden.
Pusch, R., Yong, R.N. 2006. Microstructure of Smectite Clays and Engineering Performance.
Taylor & Francis, London.
Pusch, R., Yong, R., Nakano, M. 2011. High-Level Radioactive Waste Disposal. WIT Press,
Southampton.
Sposito, G. 1984. The Surface Chemistry of Soils. Oxford University Press, New York.
8 Environmental Behavior
This chapter deals with the function of smectitic clay seals for isolation of waste
and redirecting groundwater flow. The basis for this is modeling and calculation of
the performance using laboratory-determined material properties and data from
field testing, mainly concerning THMCB. For indicating the procedure in designing
a repository and selecting functional parameters for comparing different concepts,
we will focus on two of considerable potential use. One of them is very deep hole
(VDH), known from Chapters 5 and 6 but looked at here with particular respect to
site selection, constructability, and function. The other, representing the poor man’s
choice, sarcophagus (SARC), is new and described and assessed. It represents a
potentially effective procedure in the planning of a highly radioactive waste (HLW)
repository from scratch.
8.1 WASTE ISOLATION BY USE OF CLAY

Quite obviously, more emphasis is nowadays being put on the need for effective
isolation of hazardous waste than earlier, in most parts of the world. This has led
to regulations, such as those stipulated by International Atomic Energy Agency
(IAEA), for radioactive waste, and numerous rules have been dictated by national
licensing administrations. A number of organizations have been established by
people striving for clean environment, stimulated and prompted by politicians
who search for arguments in the next election. Many of them do not realize that
alternative energy production methods and techniques can have even more serious
consequences than use of nuclear energy, such as the exploitation of gas and oil
in shales by fracking in the United States. Tens of thousands of deep boreholes
have been drilled in this country for this purpose and many of them leak gas and
chemically contaminated water to shallow aquifers because of malfunctioning
concrete liners in the holes. This can be solved by modern borehole sealing meth-
ods based on clay in combination with long-lasting concrete that we will examine
in this chapter.
We will go directly to key questions that appear in examining and assessing the
two concepts VDH and SARC mentioned in Chapter 7. They arose from the uncer-
tainty among technical staffs, geologists, and experts on engineered barriers of how
241
242 Bentonite Clay
radionuclides released from radioactive waste can be retained in the near field. Here
are the two symbols again:
VDH—vertical hole
SARC—horizontal drift
Clay/concrete at 400–500 m depth
seals, 2 km
Clay
>15 m
Containers with
concrete-
Deployment embedded waste
zone with HLW,
2 km
8.2 VDH—OSTRICH PHILOSOPHY OR A SERIOUS

ALTERNATIVE FOR HLW DISPOSAL?
8.2.1 Background
A concept for disposal of highly radioactive waste that was first proposed and
discussed some 50 years ago in the United States implies that the waste is placed in
the lower part of deep boreholes, the upper part of which is filled with effectively
sealing material (Brady et al. 2009). The concept was further developed in the last
few decades and we will examine here a version shown in Figure 8.1, with special
respect to the role of three major components; that is, grout injected in fracture zones
that are intersected by the holes, clay that makes up the major sealing material, and
concrete that is cast where major fracture zones are intersected.
The VDH design type considered here has a drilled diameter of about 800 mm*
and contains waste canisters with spent reactor fuel rods from 2000 to 4000 m depth,
whereas the upper 2000 m is sealed off from the biosphere by dense expandable clay
blocks where the rock is poor in fractures and by concrete cast on site where the
hole intersects fracture zones (Pusch et al. 2012, 2014a, 2014b). Pregrouting of these
zones is recommended for preventing erosion and loss of clay and cement particles
to the fractured rock.
* Recent judgment by deep drilling experts, like John Beswick, consider this to be possible but on the
margin.
Environmental Behavior 243
Ground surface
500 m
Fractures filled with
grout Concrete
Clay plugs
Clay blocks and

cast concrete
1500 m
Water-bearing
fracture zone
Concrete plug in
fracture zone
Clay block
Supercontainer Supercontainers
with HLW
canisters 2000 m
Casing for supporting separated by clay
borehole blocks
Borehole 0.8 m
diameter
Concrete plug
FIGURE 8.1 (See color insert.) Schematic view of proposed VDH design: seals of clay
and concrete down to 2000 m depth and supercontainers with HLW canisters separated by
clay and concrete in the interval 2000–4000 m depth. (Data from Beswick, J., Status of
technology for deep borehole disposal. NDA deep disposal report, Rev.7. Contract No NP
01185. Oxon, 2008; Grundfelt, B., Jämförelse mellan KBS-3-metoden och deponering i djupa
borrhål för slutlig tomhändertagande av använt kärnbränsle, Swedish Nuclear Fuel and
Waste Management Co [SKB], R-10-13, Stockholm, Sweden, 2010; Pusch, R. et al., J. Earth
Sci. Geotech. Eng., 2, 1–24, 2012.)
Several of the holes will be bored from a chamber near the ground surface in
slightly different directions for avoiding overlap of temperature fields. While it may
well be possible to make stable large-diameter boreholes to 4 km depth, extra support
in the form of casings in the larger part of the steeply oriented holes is recommended.
The location of the holes is decided on the basis of the structural constitution
of the rock, especially respecting the presence and orientation of fracture zones.
244 Bentonite Clay
The preparative work includes determination of the compressive strength of

the rock matrix and hydraulic characterization of the rock mass by cross-hole
measurements.
The site investigation and design work will take at least a decade for 5 to 10
deposition holes in each of 5 to 10 sites for hosting the same amount of waste as com-
monly considered in Sweden (5000–7000 metric tons of spent fuel). The degree of
utilization of the rock is expected to be 50%–60%. Boring of deposition holes will be
made using oil-drilling techniques requiring smectitic bore muds. The mechanical
boring disturbance (excavation-disturbed zone [EDZ]) of fracture-poor rock extends
to a few centimeters from the borehole walls and somewhat more where the holes
intersect fracture zones. Its hydraulic conductivity can be taken as E−10 m/s in good
rock based on field measurements and comprehensive measurements on small cores
bored perpendicularly from large holes bored by tunnel boring machines. In fracture
zones, it can be two orders of magnitude higher.
Planners of HLW disposal apply the multibarrier principle, taking the reposi-
tory host rock as the first and the most important barrier to contamination of the
groundwater by possibly escaped radionuclides, and appoint engineered barri-
ers as redundant seals of secondary importance. In a 100,000-year perspective
accumulated tectonic impact, earthquakes and glaciations will change the rock
structure causing great uncertainties in ascribing hydrological or geochemical
waste-isolating functions to the rock. Logically, this requires that engineered bar-
riers consisting of spent fuel canisters and clay and concrete seals take over the
responsibility of providing long-term isolation; this is the trend as well (Pusch
2012; Pusch et al. 2012).
For VDH, being considered as an alternative to shallow disposal in the United
States and in Sweden at some stages, the major HLW-isolating capacity is pro-
vided by the stagnancy of the heavy groundwater in rock deeper than about 2000 m
(Beswick 2008; Brady et al. 2009), thereby reducing the importance and roles of the
rock structure and engineered barriers. Both are, however, still important since they
determine where acceptable positions can be found for individual waste containers
with respect to the risk of damage by rock shear strain, and for placing borehole
seals, so that the deep holes become hydrologically separated from major conductors.
8.2.2 Rock Conditions
The interaction of a VDH with rock is simplified here by assuming the macrostructure
and mega-structure to be of orthogonal rock type* as in Figure 8.2. The number of
major fracture zones (first- and second-order zones) that will be intersected by a 4 km
deep VDH can be estimated at three to five. The spacing of minor (third order) frac-
ture zones can be 30–100 m in crystalline rock that has not undergone strong tectonic
strain, implying that 4 km deep VDHs will be intersected by about 40–100 zones of
this type (cf. Kozlovsky 1984; NEDRA 1992; Pusch 1995). The frequency of zones
that have a potential to undergo seismically or tectonically induced strain is lower at
* Categorization scheme developed for the stone industry (Pusch, 1995; Pusch et al. 2012).
Third-order zones
form network of
hydraulically
active components A 3
First-order zone
4
B
4
4
50–
150
M
5 4
B
Second-order zone
5
VDH
5–
50
M
FIGURE 8.2 (See color insert.) Rock block with optimal location of a VDH. Details to the
right are element assemblies moved out from one of the four 1 km3 blocks. The set of four
blocks is in contact with one large fracture zone (first-order zone) and contains some major
fracture zones (second-order) in a matrix containing a number of interacting minor fracture
zones (third-order). Red spots show intersection of the VDH and major zones. About 100–300
minor zones are intersected.
depth because of the higher average stresses prevailing there. This is indicated by the
fact that their hydraulic conductivity, which is a measure of porosity and strength,
drops at increasing depth as illustrated by typically recorded average values E−8 to
E−6 m/s for 50 m depth, E−10 to E−8 m/s for 200 m depth, and E−11 to E−10 m/s
for 500 m depth (Pusch 1994). At depth, shearing is therefore expected to take place
only in major fracture zones, which are the weakest structural features. Structural
modeling is anyhow necessary for finding acceptable positions of the waste packages.
Rock mechanics gives the hoop stress, which, for common optimal stress con-
ditions, implies that a 4 km VDH will be stable if the hole is filled with water or
soft clay mud. However, a few percent of the rock annulus around the hole will be
overstressed because of the statistical variation in compressive strength and stability
especially where the hole intersects fracture zones. A tightly fitting casing is recom-
mended by some investigators (Beswick 2008) as an extra support and is included in
the presently described VDH concept.
246 Bentonite Clay
8.2.3 Temperature Conditions
In many countries, the rise in bedrock temperature with depth is about 1.6°C/100 m
depth, meaning that the initial temperature at the bottom of a bored but unfilled 4 km
deep VDH is expected to be 65°C–75°C (Brady et al. 2009; Grundfelt 2010). After
installation of canisters of assumed type, the heat production associated with radio-
active decay will give a net temperature at the bottom of the hole of about 150°C,
assuming generally accepted data of heat generation and thermal properties of the
rock. The temperature will be back to ambient in a few hundred years.
8.2.4 Conditions Respecting the Chemical

Composition of the Groundwater
The concentration of total dissolved salt can be as high as 100,000 ppm at 2–4 km
depth (Brady et al. 2009). The density is high enough to maintain the convective
flow of groundwater, generated by the heat production, in the waste-containing part
(Pusch 1994; Beswick 2008; Grundfelt 2010). The Ca/Na ratio is expected to be at
least 2.0 deeper than 2 km, whereas it is estimated to be 1.5 higher up, which affects
the physical properties of the clay seals and leads, sooner or later, to more or less
complete Ca saturation of the smectite buffers. Complete conversion from initially
sorbed Na to Ca takes a few decades in the sealed, upper part of the VDH and a cou-
ple of years in the buffer surrounding the canisters in the deeper deployment zone.
8.3 SEALING COMPONENTS

8.3.1 Overview
The basic function of the VDH concept is that the density of the deep groundwater
is high enough to maintain possibly contaminated water below about 2000 m depth
and would imply that the waste isolating capacity of the engineered barriers does not
need to be high. They are, however, still of importance both in the waste installa-
tion phase and for ensuring that the deep hole with about 800 mm diameter will not
serve as conductor of possibly released and migrating radionuclides. As indicated
in Figure 8.1, the proposed artificial barriers to migration of possibly contaminated
water are as follows:
• Supercontainers with clay blocks or with clay-embedded waste canisters

• Grout injected in fracture zones
• Smectite-rich mud filling the hole in all phases of installing supercontainers
• Dense clay in supercontainers for sealing the upper 2000 m part and for
separating and embedding HLW canisters in the supercontainers from
2000 to 4000 m depth
• Concrete* in the parts of the hole that intersect grouted fracture zones
* Concrete is the term used here for cast, dense cementitious compounds, whereas grout is reserved for
injected, soft compounds.
8.3.2 Waste Canisters
The major engineered barrier for preventing radionuclides from the waste to leave
the deployment zone is the canister. For HLW, including spent reactor fuel, the
material proposed here is copper or copper alloy in the form of Navy Bronze with
90% Cu (Svemar 2005). The assumption adopted is that any seismically or tectoni-
cally induced shearing of the holes will take place in major fracture zones, mean-
ing that the canisters will not be exposed to shear stresses that can cause breakage.
Ductile, mechanically soft canister materials can therefore be considered, and pure
copper with its very significant chemical stability in an oxygen-free environment is
the number one candidate, an alternative with documented low corrosiveness being
the stiffer titanium metal.* We will consider two canister design versions; that is,
the type intended to be used in the forthcoming Swedish repository for spent fuel
(KBS-3) with an iron core with the waste bundles, surrounded by a copper liner
(cf. Figure 7.31), and the High-power technique (HIPOW) type prepared by hot iso-
static compression of copper powder in which the waste has been placed (Figure 8.3)
(Lönnerberg et at 1983; Pusch 2008). The excellence of this canister type is that it
provides complete isolation of the waste from the groundwater for at least 100,000
years. There is of course no meaning in performing any functional analysis of this
rich-man’s canister concept but it is required to repeat the earlier mentioned need
FIGURE 8.3 (See color insert.) Disc sawn from a 1.6 m long column of HIPOW-
compressed copper powder in a copper tube. The black spots in the copper mass represent
the spent fuel, simulated by steel pellets in the experiment. (Data from Lönnerberg, B. et al.,
Encapsulation and Handling of Spent Nuclear Fuel for Final Disposal, SKB/KBS Technical
Report 83-20, Swedish Nuclear Fuel and Waste Management Co [SKB], Stockholm, Sweden,
1983; Courtesy of SKB, Stockholm, Sweden.)
* Steel is a third alternative that is discussed in conjunction with examination of the SARC concept.
248 Bentonite Clay
Sealed upper part Deployment part
Fracture
Fracture
zone
Concrete zone
Highly compacted
clay
Clay mud
Canister
with HLW
Perforated
supercontainer
800 mm 800 mm
hole hole
(a) (b)
FIGURE 8.4 (a,b) Definition of VDH element for examination of the evolution.
of further R&D for making it feasible from production and cost points of view.
Figure 8.4 defines the important components of the VDH.
The clay seals in the upper 2000 m parts of the holes will be exposed to a tem-
perature of up to 60°C for 100,000 years, whereas the buffer clay in the deployment
part will be located where the temperature is up to 150°C for 3000 years followed
by 70°C for 100,000 years (Grundfelt 2010; Pusch et al. 2012). The material compo-
nents and their placement are defined in Table 8.1 together with the estimated time
for them to reach a state at which they no longer provide any significant barrier to
escaping radionuclides. Our aim is here to outline the evolution of each of the two
units, considered as separate entities, and to estimate if and at what time after place-
ment of the heat-generating waste that a critical state with release of radionuclides
may appear. The arguments for the proposed critical times are found after the table.
The description of coupled hydrological (H), thermal (T), chemical (C), and bio-
logical (B) processes is still incomplete but available models (Svemar 2005; Pusch
2008) were taken here as a basis for working out Table 8.1. The critical time is in fact
very uncertain especially since the longevity of buffers and canisters are not expressed
in parametric form and since the physicochemical environment in which specific ele-
ments and reaction products migrate has a transient character (Grindrod et al. 1999;
Kasbohm et al. 2000; Herbert et al. 2004; Lantenois et al. 2005). Two of the canister/
clay interactions require mentioning: (1) the impact of the buffer clay on the corrosion
of steel and copper and (2) the impact of gas embolism and migration from the casing,
supercontainer, and canister through the clay and concrete seals.
8.3.2.1 Steel Canisters

Taking the physical effect of metal corrosion first (Grauer 1986), conversion of iron
and steel to magnetite, which has a specific density of about 5000 kg/m3, causes
TABLE 8.1
VDH Elements, and Prediction of the Time after Installation When They Can Be Expected to Be Out of Order
Part Components Placement Evolution Stage Time to Critical State (years)
Upper Part
Upper, 0–2000 m Rock Natural Boring disturbance (EDZ) >100,000
segment with clay Mud of Na montmorillonite 1150 kg/m3 density, pumped 1. Penetration into EDZ >100,000
Environmental Behavior
seals in from below via drill rods 2. Consolidation under swelling pressure of
dense clay
Casing of steel Inserted in mud, cut out 1. Corrosion with H2 release <10,000
where fracture zones are 2. Fe ion release
intersected
Casing of Navy Bronze or Inserted in mud, and cut off Corrosion with Cu ion release >100,000
copper where fracture zones are
intersected
Supercontainer of steel Moved down in the mud 1. Corrosion with H2 release <10,000
2. Fe ion release
Supercontainer of Navy Moved down in the mud Corrosion with Cu ion release >100,000
Bronze
Canister of steel Goes with supercontainer 1. Corrosion with H2 release <10,000
2. Fe ion release
Canister of copper/iron Goes with supercontainer Corrosion with Cu ion release 100,000
HIPOW copper canister Goes with supercontainer Corrosion with Cu ion release >100,000
Dense smectitic buffer clay Goes with supercontainer 1. Expansion and consolidation of the mud <10,000 for the steel case
(Na montmorillonite) 2. Stiffening by heating >10,000 for the Cu cases
3. Stiffening by Fe or Cu uptake
(Continued)
249
250
TABLE 8.1 (Continued)

VDH Elements, and Prediction of the Time after Installation When They Can Be Expected to Be Out of Order
Part Components Placement Evolution Stage Time to Critical State (years)
Upper Part
Lower, 2000–4000 m Rock Natural Boring disturbance (EDZ) >100,000
segment with HLW Mud of Na >1150 kg/m3 density, 1. Penetration into EDZ <10,000
in canisters montmorillonite/ pumped in from below via 2. Consolidation under swelling pressure
embedded in clay Saponite drill rods of dense clay
Casing of steel Inserted in mud, cut out 1. Corrosion with H2 release <1000
where fracture zones are 2. Fe ion release
intersected
Casing of Navy Bronze or Inserted in mud, cut out Corrosion with Cu ion release <10,000
copper where fracture zones are
intersected
Supercontainer of steel Moved down in the mud 1. Corrosion with H2 release <5000 for the steel case
2. Fe ion release
Supercontainer of Navy Moved down in the mud Corrosion with Cu ion release >10,000
Bronze
Canister of steel Goes with supercontainer 1. Corrosion with H2 release <1000
2. Fe ion release
Canister of copper-lined/ Goes with supercontainer Corrosion with Cu ion release <10,000
iron
HIPOW copper canister Goes with supercontainer Corrosion with Cu ion release >100,000
Dense smectitic buffer clay Goes with supercontainer 1. Expansion and consolidation of the mud <10,000 for the steel case
(saponite) 2. Stiffening by heating >10,000 for the Cu cases
3. Stiffening by Fe or Cu uptake
Bentonite Clay
expansion by nearly 50% and hence compression of the mud and dense clay buffer.
For the sealed upper part of a VDH, the unit element consists, in radial direction,
of 50 mm of mud with a density of 1150 kg/m3, and buffer clay with an initial dry
density of 2000 kg/m3 confined in a supercontainer. The ultimate average density
at saturation of the clay separating the container from the rock will theoretically be
about 2015 kg/m3. Complete corrosion of a supercontainer with outer diameter of
700 mm and 10 mm thickness will make it expand by about 5 mm and cause minor
compression of the clay annulus and the clay remaining in the container. For the
deployment zone, corrosion is more important, of course. The steel canister, with an
outer diameter of 500 mm and an assumed wall thickness of 25 mm will, at complete
corrosion, expand outward by 12.5 mm and give the clay surrounding it an average
density at water saturation of about 2300 kg/m3, which makes it totally impermeable.
The swelling pressure would be more than 200 MPa, which can break the surround-
ing rock, as illustrated in Figure 8.5, by generating tension fractures for rock stress
conditions that may appear in the deployment zone already for a swelling pressure of
just one-tenth of this pressure.
In this context, we also have to consider the role of gas resulting from corro-
sion, the most critical case being the formation of a highly pressurized gas bubble
at the surface of the supercontainer. Its size will depend on the gas pressure, which,
according to what is commonly assumed (Horseman and Harrington 1997), and
stated in Chapter 4 of this book, cannot exceed the sum of the water and the swelling
pressures without causing gas to make its way out through the clay by channeling
(fingering). The sum of the pressures can be more than 40 MPa at the base of the
deployment zone. Escape of gas through pathways generated in the clay will take
place when the gas pressure reaches the level of the swelling pressure.
Having touched on these matters, we are prepared to look closer at the arguments
for selecting the critical times in Table 8.1.
8.3.2.2 Mud Performance

When the supercontainers are being placed, the thereby generated pressure pulse on
the mud makes it penetrate into the boring-generated EDZ by which the fine fissures
in the most shallow millimeters of the rock become sealed. The radial expansion
of the dense clay in the supercontainers consolidates the mud that becomes inte-
grated in the annular clay around them. Calculation using stepwise compression/
water expulsion has shown that consolidation and increase in density is relatively
quick. The density including water exceeds 1500–1600 kg/m3 in a few days and will
be at least 1800 kg/m3 in a week (Pusch and Ramqvist 2006). This is high enough to
carry the container by wall friction. For a temperature of up to 60°C, one concludes,
referring to the evolution of the integrated mud and dense clay described in Chapter 6,
that the clay surrounding the supercontainers will perform acceptably for at least
100,000 years.
8.3.2.3 Casing Performance

Casings of steel will corrode anaerobically, give off hydrogen gas and be totally
converted to magnetite in a very long time perspective. The gas production rate is
expected to be very low and gas may be released by diffusion without leaving any
252 Bentonite Clay
−2.652E+1 to −2.364E+1
−2.364E+1 to −2.075E+1
−2.075E+1 to −1.786E+1
−1.786E+1 to −1.498E+1
−1.498E+1 to −1.209E+1
−1.209E+1 to −9.203E+0
−9.203E+0 to −6.316E+0
−6.316E+0 to −3.430E+0
−3.430E+0 to −5.430E−1
−5.430E−1 to 2.344E+0
2.344E+0 to 5.230E+0
5.230E+0 to 8.117E+0
8.117E+0 to 1.100E+1
1.100E+1 to 1.389E+1
X 15 MPa
Y
E—50,000 MPa
v—0.3
p—internal pressure
(20 MPa)
35 MPa
p MPa
FIGURE 8.5 (See color insert.) Hoop stress in rock around a VDH exposed to an internal
pressure of 20 MPa calculated by the FEM technique. The primary principal stresses in the
rock are 35 and 15 MPa, respectively. Tension stresses appear in the red parts reaching a
maximum value of about 11 MPa. (Data from Pusch, R., Geological Storage of Radioactive
Waste. Springer-Verlag, Berlin, Germany, 2008.)
continuous open channels in the clay around the casings, which provide support
to the rock in the waste installation phase but play no role thereafter. Iron in ionic
form will replace Na and Ca in exchange positions and in the smectite lattices caus-
ing stiffening and loss of the self-healing potential. Corrosion will contribute to the
consolidation of the clay around the supercontainers by the expansion caused by
the transformation to magnetite, the probable net effect of the processes being an
increase in density of the stiffened clay that also becomes tighter. Casings of Navy
Bronze perform as those of pure copper and are expected to stay intact for many tens
of thousands of years without affecting the clay significantly.
8.3.2.4 Supercontainer Performance

Supercontainers of steel will undergo the same evolution as steel casings and hence
corrode anaerobically, giving off hydrogen gas and ultimately being totally con-
verted to magnetite. The gas production rate is expected to be very low and gas may
be released by diffusion without leaving any continuous open channels in the clay
embedding the supercontainers. Dissolved iron will be sorbed by the clay buffer con-
tained in and surrounding the supercontainers. This will cause cementation that starts
from their surfaces and proceeds inward and outward at a rate that is controlled by
diffusion. Applying the diffusion coefficient E−11 to E−10 m2/s, the ion exchange pro-
cess with Fe replacing initially sorbed Na in the thin clay next to the supercontainer
will be over very quickly but the corrosion process that transforms the metal to more
voluminous magnetite will be slow. The delay is naturally caused by the successively
increased density of the adjacent clay that retards precipitation of cementing iron
complexes. One would dare to assume that total degradation of the supercontainers
causes loss of their mechanical strength in a few thousand years. This is of no practi-
cal importance since they have in fact no barrier function as concerns the release of
radionuclides and only serve to make installation of the dense clay seals possible.
Supercontainers of copper or Navy Bronze will stay largely intact for at least
100,000 years and have no negative impact on the clay contained in them.
8.3.2.5 Canister Performance

Here, we talk about considerably higher temperatures, up to 150°C. The role of steel
canisters for the performance of the buffer clay is the same as for the supercontainers
except that they produce heat and expose the buffer to a temperature gradient that
can initially be more than 5°C/cm. The higher temperature in the deployment zone
of a VDH speeds up the corrosion of steel canisters and the ion exchange to Fe in
the buffer clay. Since the environmental conditions are almost the same as for the
supercontainers, it is reasonable to assume that the heat-controlled degradation rate
is the same. Since the thickness of the canister walls is estimated to be larger, that is,
a few tens of millimeters, the critical time will be at least 5000 years and probably
more because of the densification of the surrounding clay by consolidation caused by
the conversion of steel to magnetite.
8.3.2.6 Buffer Clay Performance

Here, we have to distinguish between the well-known clay material candidates: Na
montmorillonite clay, and chemically more stable clays, primarily saponite and I/S
mixed-layer clay. Turning back for a moment to Chapter 6, where we concluded from
the tests referred to that the so-called trioctahedral smectite species saponite (Gueven
and Huang 1990), sustained hydrothermal treatment in salt Ca-dominated water at
95°C under a significant thermal gradient better than the two montmorillonitic com-
petitors. This is in agreement with published chemical/mineralogical arguments, and
254 Bentonite Clay
also with the successful use of saponitic bore muds in deep drillings, speaking in
favor of using saponite as buffer clay in hot, saline environment (Grim and Guven
1978). This type of clay mineral, characterized by having Mg in the octahedral layer
and being less expandable than ordinary dioctahedral montmorillonite, is therefore
a strong buffer clay candidate for use in the deployment zone. In contrast, Na mont-
morillonite, in the form of MX-80 clay,* suffered significantly from such treatment as
demonstrated by the prototype project and the Czech mock-up experiment with Ca
montmorillonite. However, for use in the upper part of a VDH where there will be no
nuclear waste, Na montmorillonite, with its excellent gel-forming capacity compared
to both saponite and I/S mixed-layer clay, will remain competitor, although cementa-
tion effects in the lowermost part of the upper zone will somewhat reduce its expand-
ability and ability to maintain tight contact with the rock in the hole.
The problem is as to what impact on the rheological properties of the clay seal
that can be caused by the casing and supercontainer. As concerns the upper part
of a VDH, the conclusion from assessing the various buffer clay alternatives with
respect to this issue is that steel components will stiffen the clay and make all or
part of it successively less effective as sealant (Grauer 1986; Pusch et al. 1993; Thao
et al. 2014). Taking the often cited corrosion rate as 1–10 µm/year for steel cas-
ings and supercontainers, they will be totally degraded in 1 to 20,000 years, hence
motivating selection of the critical time to be averaged to 10,000 years. For casings
and supercontainers of copper or Navy Bronze, the corresponding critical time is
believed to be at least 100,000 years. The clay will anyhow suffer from cementation
by precipitated silica and iron compounds (Pusch et al. 1998), irrespective of which
of the two metals that will be used, which leads us to consider <10,000 years as the
critical time for the steel case and >10,000 years for the Cu case. In this context,
we need to keep in mind the finding from North Sea explorations that smectite in
clay beds that have been exposed to 65°C–75°C have undergone only insignificant
transformation since Tertiary time, in contrast to clay from more than 2000 m depth,
exposed to 70°C–80°C, which have been largely converted to I/S mixed-layer miner-
als (Lindgreen 1991; Roaldset et al. 1998). These investigators found, however, that
even for the beds with no illitization, cementation by precipitated matter, mostly of
silicious type, had turned the originally ductile clay into rigid claystone.
In the deployment zone, the impact of iron is much stronger, especially in the
hottest, lower-most part. Here, the corrosion rate can be more than 10 times higher,
applying Arrhenius-type impact of the temperature, hence implying that the steel com-
ponents have degraded totally in a thousand years, whereas copper or Navy Bronze
components can remain largely uncorroded for at least 10,000 years. Naturally, one
finds the HIPOW copper canister to be the best choice disregarding from cost.
8.3.3 Concrete
The role of the concrete seals in VDH is to support adjacent clay seals and serve as fil-
ters for minimizing loss of clay particles from them. The chemical impact of concrete
on the clay seals and on the rock shall be as small as possible, which is achieved using
* This clay has, in fact, one-third of the cation exchange positions occupied by Ca.
only about 6% cement of low-pH type and very silica-rich aggregate. About 10%–15%
of very finely ground talc powder serves as fluidizer and contributes to give the con-
crete high strength by reacting with the cement (Pusch et al. 2013, 2014b).
8.3.4 Practical Aspects—Placeability
One realizes that problems can appear in the placement phase: (1) the insertion of
supercontainers in the mud must not take too long time since the expansion of the
dense clay in them, causing the intended consolidation of the mud, can be too fast
and make the containers get stuck, and (2) the consolidation and stiffening of the
integrated mud and clay migrated from the supercontainers after placement may be
too slow for preventing concrete cast over it to penetrate into the clay along the con-
tact between rock and clay. This phenomenon has been observed in field experiments
(Warr and Grathoff 2011).
For realizing the implications and consequences of the first-mentioned problem,
one has to consider the stress conditions in the clay, caused by the own weight of the
containers, and the shear resistance of the mud. Sets are prepared with one or two
canisters fitted in perforated supercontainers that also contain a clay block at each
end and one between the canisters (Figure 8.6). The casing, supercontainer, and
canister are of the same material.
Two canister configurations were initially considered in Sweden, one with four
boiling water reactor (BWR) elements, and another with two BWR and one pres-
sure water reactor) elements and they are both taken here as candidates (Svemar
2005). The 3050 mm long canisters with about 500 mm diameter have room for
Fracture zone
Concrete Fracture-
Clay poor rock
Casing for block
supporting
borehole
Upper end of supercontainer
Supercontainer with canisters and clay block
Canisters
Borehole
0.8 m
diameter
FIGURE 8.6 Supercontainer in the deployment zone. The VDH has 800 mm diameter.
The same type of container filled only with dense clay blocks is used in the 500–2000 m part.
256 Bentonite Clay
claddings with fuel rods, and are equipped with lids welded to them for tight closure.
According to the concept proposed here, they are surrounded by a 100 mm annulus
of highly compacted blocks of dense, expandable clay and fitted in the supercontain-
ers that have 6 m length and 750 mm outer diameter.
The space between casing and supercontainer is 20–30 mm. At the installation,
there is about 50 mm of mud between the rock and the containers.
The performance of the mud is controlled by its density, mineral composition,
temperature, and salinity of the water used for preparing it. They combine, together
with the stirring effect of moving the supercontainers down into the mud, to give a
reduced shear resistance that we can express as viscosity. Various laboratory inves-
tigations have given information on this parameter, which is of fundamental impor-
tance for grouting and kept as a secret by deep-drilling companies. In the present
context, we are interested in determining the bearing capacity of the mud at rest and
the shear resistance at forcing down the supercontainers (Figure 8.7).
Once the shear strength of the mud is known, one can roughly estimate its bear-
ing capacity as the product of a coefficient Nc and the (undrained) shear strength
τfu. Several analytical solutions were proposed some 50 years ago by pioneering
geotechnicians, like Meyerhof (1963) and Terzaghi et al. (1996), who expressed the
shear strength as a function of the internal friction angle, which is on the order
of 10°–20° for smectite-rich clay. For this interval, the dimensionless parameter Nc
ranges between 5 and 9, hence giving a bearing capacity of about 5τfu to 9τfu. Failure
starts at the edges of the supercontainer as indicated in the left picture in Figure 8.7,
and develops in the form of a system of slip planes of which some extend radially
from the edge and some have the form of logarithmic spirals.
Table 8.2 shows the results from the evaluation of the shear resistance as a function
of the density of water-saturated montmorillonite-rich mud by uniaxial compression.
Neglecting the shear resistance generated along the supercontainer and taking only
the own weight of the supercontainer—estimated at 10 tons—into consideration, the
contact pressure on the mud in the first moment will be about 500 kPa. This means
Starting stage Moving stage
Supercontainer
Clay
mud
Local Flow
plastization
FIGURE 8.7 Two stages at insertion of the supercontainers to be considered. Left:

supercontainer exerting pressure on the mud at rest. Right: supercontainer being forced
down into the mud causing shear failure of the mud.
TABLE 8.2
Geotechnical Data of MX-80 Saturated with Distilled Water as Evaluated
from Uniaxial Compression Tests
Density of MX-80 Clay Estimated Bearing
Saturated Distilled Water Hydraulic Shear Strength Capacity for Concentrated
(kg/m3) Conductivity (m/s) (kPa) Loading (kPa)
1100 E−9 50 250–450
1150 5E−10 200 1000–1800
1325 5E−11 400 2000–3600
1400 2E−11 550 2750–4950
1700 E−12 700 3500–6300
that the supercontainer may not sink down under its own weight for a mud density
of 1150 kg/m3, and that a force of a few tons has to be applied to bring it down.
Alternatively, a heavy vibrator placed on the supercanister set can be used. One can
also add more supercontainer units to the one with two units assumed here, but we
will see that longer sets cause difficulties by the ongoing stiffening of the mud.
Once the supercontainer has initiated local plastization of the mud by overstress-
ing, it continues to move by the momentum and the moving stage is entered (right
picture in Figure 8.7), in which the viscosity of the mud determines the rate of move-
ment. There moulded sol state reached by the thixotropic mud makes the super-
container continue to move down assisted by the successively increased thermally
generated drop in mud viscosity when it comes deeper down. The surface tempera-
ture of the containers is estimated to be near 100°C in the placement phase.
There is a problem, however, caused by the successive thickening of the mud by
migration of dense clay through the perforation of supercontainers. The dry density
of the clay contained in them is initially about 2000 kg/m3 with a degree of water
saturation of 50%–60%, which implies a very high suction potential. The dense clay
sorbs water from the mud, which at the same time, becomes consolidated by dense
clay moving out through the perforation of the supercontainers. The process is rela-
tively quick and stiffening of the mud can be significant already in—four to eight
hours, which means that the lowermost supercontainer must be down in no more
than eight hours. Any delay means that the resistance to bring them down after this
time will require loading, which can be done by use of the drill rig required for the
installation. The force needed can be hundreds of kilonewtons after a couple of days
(Pusch 2011).
Since some unforeseen delay in the installation of the supercontainers must be
expected, a process or technique for retarding the maturation of the clay seals is
needed. It is primarily provided by the mud itself since the flow rate of mud pore-
water to the expanding blocks is low and successively drops in the course of the
consolidation that the mud undergoes.
A method for further retardation of the wetting of the dense clay is to coat the
supercontainer with a thin paste of mixed smectite clay and talc (Pusch 2011, 2014a).
The coating retards migration of water into the dense clay and delays expansion of
258 Bentonite Clay
FIGURE 8.8 Borehole seals of very dense montmorillonite-rich clay presaturated to nearly
100% and then stored in distilled water for 24 hours. The clay in A had expanded by about
6 mm and would have largely occupied the initial 6 mm wide space between the perforated
tube and the rock in a 100 mm diameter borehole. B had been coated with silica solution for
delaying hydration of the clay, a technique that was not successful and that made the clay lose
coherence.
them (Figure 8.8). Another possibility of delaying the consolidation of the mud is to

use clay in the supercontainers that has an initially high degree of water saturation
since the suction potential is then significantly lower. Such clay blocks can be pre-
pared by mixing dry clay powder with the aforementioned dry water. The degree of
saturation can be anything from the natural water content to 100% using this method.
8.3.5 Long-Term Function of Clay Components

in the Sealed and Deployment Zones
We saw in Chapters 6 and 7 that temperature is the major controlling parameter

for the mineralogical stability of smectite clay, and that even a very brief period of
simultaneous hydration and exposure of matured buffer to high temperatures gener-
ates mineral changes that can be permanent (cf. the Kinnekulle case in Chapter 6*).
For the upper, sealed zone of a VDH, where the temperature stays below 60°C–70°C,
one can assume, as is supported by the mineralogists’ findings from the North Sea
investigations (Grim and Guven 1978; Gueven 1990; Lindgreen 1991; Pusch 1993;
Roaldset et al. 1998; Herbert et al. 2008), that montmorillonite will remain the domi-
nant clay mineral even beyond 100,000 years. Still, the impact of very high salt con-
tents in combination with high temperature and temperature gradients may change
this and cause cementation, brittleness, and loss of self-sealing potential (Kasbohm
et al. 2000; Pusch and Kasbohm 2002).
The longevity of the individual engineering components depends on the physico-
chemical processes involved in their mutual interaction as indicated by the critical times
in Table 8.1. Such interaction can totally ruin the integrated sealing function of the sys-
tem formed by them, inter alia by creation of H2 gas channels from quickly corroding
iron and steel objects. The impact described as stiffening by Fe or Cu uptake is related
* Pusch and Madsen 1995.

to precipitation of cementing agents but for iron and steel there is also a change in the
structure of montmorillonite crystals and thereby a change in physical performance,
such as expandability. Such impact has been found by Vietnamese researchers who
have observed that octahedral Fe causes destabilization of smectite crystal lattices
(Herbert et al. 2012). In the presence of corroding iron, protons are released from the
lattices and Fe2+ get sorbed on the edges of smectite stacks and migrate from there
into the interlamellar space and further into octahedral positions because of the very
high charge deficit. This destroys the lattices because of the large size of the Fe ions.
Naturally, the corrosion and effect of destabilization of smectite clay goes on faster
in the deployment zone where the temperature is up to 150°C and the salt content
very high (Kaufhold and Dohrmann 2009).
The impact of corrosion by expansion of casings, supercontainers, and canisters
and crystal lattice degradation by Fe uptake is much less important if copper or Navy
Bronze are used as indicated in Table 8.1.
8.3.6 Long-Term Function of Concrete Seals

The concrete seals, commonly called plugs, that are made of low-pH cement and talc as
fluidizer, will be in contact with clay seals and cause mutual degradation (Huertas 2000;
Gaucher and Blanc 2006) that will be very limited, however, because of the relatively
low pH of the porewater in concrete with low-pH cement, as described and discussed in
Chapter 6. The conclusion is that the type of concrete developed for use in boreholes is
very stable and is able to fulfill the criteria of having a high bearing capacity and serving
as filter for preventing through migration of clay particles even if the cement component
would dissolve and be lost (Mohammed et al. 2013; Pusch et al. 2013).
8.3.7 Impact of Gamma Radiation

This matter has been in focus of IAEA and the various national organizations that
are responsible for safe isolation of HLW: SKB in Sweden, ANDRA in France,
NAGRA in Switzerland, ENRESA in Spain, and so on (Svemar 2005). As reported
in Chapter 6, experiments with smectite-rich clay exposed to heavy gamma radiation
doses retain their physical properties (Pusch et al. 1993). The only verified process
in one-year long tests conducted at Saclay, France, was that migration of Fe from
irradiated, metallic iron into contacting water-saturated montmorillonite-rich clay
was speeded up somewhat. Preliminary tests made in the 1970s had indicated that
heavy gamma radiation can cause disintegration of the minerals that leads to particle
breakage and smaller particle size (Jacobsson and Pusch 1974).
8.3.8 Unexpected Events
The VDH concept discussed here implies that the boring mud is the same as the
deployment mud (cf. Beswick 2008; Brady 2009; Grundfelt 2010). For serving in the
boring phase, it must have the required ability to support the hole being bored and
to bring up the debris from the bore head. If this does not work or rock blocks fre-
quently fall in the hole, it has to be abandoned. An unexpected problem can appear
260 Bentonite Clay
if the supercontainers with waste canisters and dense blocks of expansive clay do
not sink down or cannot be forced down into the mud to the predetermined levels
because of too high shear strength of the mud. This can be helped using equipment
such as PG-Macs* for activating the mud between each placement event and using a
heavy vibrator at each placement. After installation of the supercontainers, the mud
will be consolidated by the expanding dense blocks in them, which seal the VDH
holes in the deployment zone at the end of each container, and which completely
occupy the supercontainers in the upper, sealed part of the VDHs. The performance
of the mud in the two parts will be largely different because of the difference in
temperature and salt conditions. For the deployment part, the only criterion to be
fulfilled is that the consolidated mud stays in contact with the rock and canisters for
minimizing convective water flow in the axial direction. For the upper, sealed part
without waste, it is required that the expanded clay stays less permeable than the sur-
rounding rock, which can be conservatively taken as E−11 m/s (Pusch 1994; Pusch
et al. 2014a). We see from the geotechnical data in Chapters 5 and 6 that this crite-
rion can be fulfilled equally well for the VDH and KBS-3V concepts (Svemar 2005).
8.3.9 Environmental Impact
As said, the philosophy behind the VDH concept is that deep groundwater with very
high salt content is confined at depth and kept below the upper sealed zone despite the
impact of heat-producing waste. Once the supercontainers with clay-embedded canisters
have been installed, release of radionuclides from possibly failed canisters will hence
not contaminate groundwater above the deployment zone (Beswick 2008). According
to this concept, the engineered barriers can therefore be regarded as a backup that will
not be required for preventing radioactivity to reach surface-near groundwater under
conditions like those prevailing in present time. They are, however, needed for keeping
the waste isolated in the installation phase, which meets some practical difficulties with
densification of the clay mud as we have seen. The often underrated risk of seismic and
tectonic impact respecting shallow and medium-level repositories is of minor impor-
tance to VDH because the rock strain that can take place is largely confined to the major
low-order discontinuities where no waste shall be placed (Pusch 1995, 2008).
There is considerable uncertainty concerning the survival potential of the engi-
neered barriers in the deployment zone in the relatively short period of heating related
to radioactive decay, the only totally safe canister design being represented by the
HIPOW copper canister. Simpler and cheaper versions require well-performing clay
buffers, which is offered by the described VDH concept.
The main difficulty, as one can see today, is in boring 4 km deep holes with
800 mm diameter and sealing and stabilizing those parts where rock fall can take
place (cf. Figure 8.9). Provided that this can be made, the VDH concept is believed to
be a very good candidate concept (cf. Beswick 2008). The way to handle significant
locally widened VDH is to cast concrete of the talc/low-pH cement type recom-
mended for sealing those parts where fracture zones are intersected, which is also
where widening has occurred.
* Hydraulically driven agitators with mud flow capacities of more than 10,000 m3/h.
214 1,000
214
2,000
2,500
225
3,000
Proterozoic
290
4,000
5,000
250
390
Depth (m)
6,000
7,000
8,000
8,450
225
9,000
Archean
570
10,000
11,000
11,600
240
12,000
200 300 400 500 600 700
370
Cross section of the well (mm)
FIGURE 8.9 Measured variation in diameter of the 12 km deep Murmansk borehole. (After
Kozlovsky, Y.E.A., The world’s deepest well, Scientific American, 251, 106–112, 1984.) The
depth interval 1000–3000 m and >5000 m caused anistropic shape of the hole section.
8.4 SARC—THE POOR MAN’S SOLUTION

8.4.1 Background
Multibarrier isolation of hazardous waste is recommended by organizations and
companies that are responsible for underground disposal of such waste. They tra-
ditionally take the host rock of repositories to be of first and major importance
and consider engineered barriers to serve only as a backup by being less reliable
because of their artificial origin and nature. The isolating role of the rock is exag-
gerated because hydrogeologists have been attracted by the challenge in working out
262 Bentonite Clay
theoretical models for detailed prediction of the intensity and direction of ground-
water flow. Considerable effort and amounts of money have been spent on develop-
ing models that are not valid when tested by field investigations such as cross-hole
hydraulic flow testing. Thus, even the definition of the distribution and magnitude
of hydraulic gradients is hypothetic and the flow paths unpredictable except for
megascopic structural features such as those termed first-order discontinuities,
representing major fracture zones. At best, such models can describe the p resent
large-scale hydrological function of the system of interacting major fracture zones
like the CMI model used in the EU Project Low Risk Deposition Technology
(LRDT). This model was based on a categorization principle that is practical and
useful for groundwater and rock mechanical modeling and in situ characterization
on any scale (Table 8.3).
The discontinuities are termed first-order fractures, second-order fractures, and
so on and correspond to SKB’s categorization scheme using D1, D2, and so on. The
D-system has its equivalents in the structure schemes worked out by other organiza-
tions. Figure 8.10 shows major discontinuities of different size in regular patterns
proposed by SKB’s Spanish equivalent ENRESA.
Computational models for the hydraulic and rock mechanical performances of
crystalline rock were developed in the LRDT Project and were used for calculation
of the risk of contamination of groundwater in nearby wells for drinking water in
this project. A reference case as regards disposal of used batteries in a large mine
cavern showed that isolation of hazardous waste by expandable clay gives acceptable
water quality for any period of time (Popov and Pusch 2006; Pusch et al. 2014a).
One knows from extensive rock testing in the field that excavation damage—the
TABLE 8.3
Categorization Scheme for Rock Discontinuities
Hydraulic
Order Persistence Conductivity Gouge Content Strength D-system (SKB)
Zones
First >Kilometers VH VH VVL D1
Second Kilometers H H VL D2
Third Hundreds of meters M to H M L D3
Rock mass with no discontinuities of lower order

Fourth Tens of meters M to H L L to MH D4
Fifth Meters L VL MH to H –
Sixth Decimeters VL VL H –
Seventh <Decimeters VVL VVL VH –
Note: Orders first to third are fracture zones and fourth to seventh are discrete fractures. A feature is
categorized according to what can be observed or measured but is often part of a larger and more
complex discontinuity.
VH—very high; H—high; M—medium; L—low; VL—very low; VVL—absent.
Regional fractures of
primary order
Block
0 1 2 3 km
Site for waste

placement
Local fractures of
second order
FIGURE 8.10 ENRESA’s general rock structure model with major fracture zones. (Data
from Svemar, C., Cluster Repository Project [CROP], Final Report of European Commission
Contract FIR1-CT-2000-20023, Brussels, Belgium, 2005.)
EDZ issue—by blasting creates continuous major flow paths and introducing them
in the models showed that they played a significant role in the contamination of the
wells by interaction with fracture zones. The isolating and flow-diverting potential of
lining the caverns and drifts with dense blocks of expandable clay was very obvious
as shown by Figure 8.11. The case was a big mined cavern in rock with an average
regional hydraulic gradient of 0.02 m/m. The design principle adopted was to place
and compact layerwise solid battery waste or solidified pesticides mixed with air-dry
smectitic clay granules in the room. The arrows representing velocity vectors indi-
cate that the large majority of the flow takes place around the cavern and not through
it. Figure 8.12 illustrates how drifts can be used for storing big containers with haz-
ardous chemical waste, a concept that makes use of smectite-rich clay of high density
in liners and dense pumped mud for embedment of the containers.
8.4.2 Steps in Siting of SARC

For SARC, intended for disposal of heat-producing HLW, the selection of suitable
tunnels, drifts, or rooms in a mine that qualifies is naturally of primary interest both
with respect to mechanical stability and to the presence and frequency of intersect-
ing water-bearing fracture zones of third and lower orders, where waste should not
be placed. The rock structure is hence the first issue to consider and the detailed
knowledge in this respect of the mining engineers and mine geologists is invalu-
able. The difficulty is to define a rock structure model for hydrological analysis.
264 Bentonite Clay
x
z
FIGURE 8.11 Vector plot of Darcy velocity in vertical section of crystalline rock with
network of third-order discontinuities hosting a big cavern and a drift connected to it.
(Calculation in 3D by CMI, Ashurst, Southampton, UK.)
Thus, the regularity of the various discontinuities and the rock stress conditions vary
considerably in real rock as illustrated by Figure 8.13, which represents granitic rock
in the Forsmark area north of Stockholm selected for locating a HLW repository.
Deeper down, the stress conditions are more homogeneous (cf. Brady 2009) but still
considerable as indicated by Figure 8.14, which shows trends in orientation and spac-
ing of hydraulically and mechanically active fourth-order discontinuities and the rock
pressure variations in the horizontal plane at 360 m depth in the granitic dome in the
Stripa mine area a few hundred kilometers westward from Forsmark (Gray 2003).
The variations in orientation and interconnectivity of the structural features
explain why the average gross hydraulic conductivity can vary by orders of mag-
nitude. These examples also demonstrate that a planned underground repository of
hazardous waste can have some of the tunnels oriented favorably with respect to the
principal stresses in the horizontal plane whereas others, differently oriented, can
have high hoop stresses and be unstable.
A further difficulty in characterization of rock structure for calculating ground-
water flow and quantification of stability conditions is the excavation-damaged zone
Shotcrete
Blocks of compacted
smectite clay
HLW
Smectitic backfill
FIGURE 8.12 Proposed design principle for deposition of hazardous waste in abandoned
mines in rock according to the LRDT concept. The upper left pointer indicates pumped-in
dense smectite-rich mud. The upper right pointer represents shotcreted or cast concrete with
a low content of low-pH concrete (coarsely dotted). The lower right pointer shows a 0.3–0.5 m
lining of blocks of highly compacted expandable clay (finely dotted). (Data from Popov, V.,
Pusch, R., Eds., Disposal of Hazardous Waste in Underground Mines. Series: The sustain-
able World, Vol. 11, WIT Press, Ashurst, 2006; Pusch, R. et al., J. Earth Sci. Geotech. Eng.,
4, 33–54, 2014c.)
around shafts, tunnels, and rooms. The role of the EDZ for groundwater flow in rock
with blasted tunnels and rooms is very important, its average hydraulic conductivity
in the direction of blasted tunnels being 10–100 times higher than that of undis-
turbed rock and up to 1000 times higher in the floor (Pusch 1994, 2008). In modern
tunnels, it extends into the rock by a couple of decimeters to 1.5 m in the floor and
to one to several meters deep thin mines, where the comprehensive use of blasting
certainly did not have the purpose of preserving the rock.
A matter of particular importance but not taken seriously by several of the organi-
zations that are responsible for the disposal of HLW is the change that the rock struc-
ture and hydrological performance will undergo with time. Seismic and tectonic
impact will change the performance of repository rock in a few thousand years and
very much in 100,000 years, in which at least one major glaciation cycle will happen,
causing very significant alteration of the rock stress field. It will have stronger impact
than on a repository with a bored and sealed VDH.
Putting together all the difficulties in characterizing repository rock with respect
to its structural constitution, it is obvious that reasonably reliable calculation of
groundwater flow can only be made by simplifying the structural model to con-
sist of only first-, second-, and third-order discontinuities located in a rock matrix
performing as a porous medium with averaged hydraulic conductivity. Since the
spacing of the third-order discontinuities, representing persisting minor fracture
zones, is commonly 30–70 m (Pusch 1994, 1995); the hydraulic and rock mechanical
266 Bentonite Clay
50
N
Theoretical vertical pressure
100
150
200
0 −10 −20 −30 MPa

250
Scale of stresses
−10 0 +10
Vertical pressure
MPa
FIGURE 8.13 Variation in the stress field in shallow granitic rock in the Forsmark area
north of Stockholm, Sweden. (After Carlsson and Olsson.)
Measured stress on
A principal planes
A—25 MPa
B B—28 MPa
C—24 MPa
D—20 MPa
FIGURE 8.14 (See color insert.) Sets of steeply oriented fractures in an approximately
orthogonal pattern in granite. Blue sets are oriented in the direction of the highest rock
stress and carry more water than the red ones that are compressed by this stress component.
Distance between the lines is 30 m. (Data from Pusch, R., Geological Storage of Radioactive
Waste. Springer-Verlag, Berlin, Germany, 2008.)
performances of large parts of a selected repository host rock remain unpredictable.

The implication of this is that the repository host rock can only be relied on as a
mechanical protection of the chemical apparatus and that engineered barriers are
needed for preventing radionuclides from escaping to the groundwater.
8.4.3 Constitution of a SARC Repository for HLW

8.4.3.1 General Conditions
The fact that strong gamma radiation necessitates remote handling and use of
robot techniques for placement of canister units requires adaption of design and
construction to this condition when following the concept like the one indicated
in Figure 8.12. While this solution admits successive installation of clay blocks
of the liner parallel with the placement of waste containers and other engineering
barriers in adapting tunnels and drifts in a mine to repository, standards have to
be made remotely. The basic principle is to keep the waste dry for very long peri-
ods of time, which requires use of smectitic clay and, for providing mechanical
stability, concrete that is chemically compatible with the clay. Effective ventila-
tion of the drift in the construction phase is required. Preparative steps to be taken
are as follows:
• Selection of tunnels or drifts with low inflow of groundwater.

• Smoothening of the rock surface by trimming.
• Boring of 76 mm circumferential axially oriented holes from niches for
draining the rock in the construction phase. The holes are sealed after com-
pleting installation of the waste.
• Deepening of the floor by sawing for removal of the blast-disturbed rock to
2 m depth and replacing it by concrete with low-pH cement for providing a
flat working platform.
While chemical waste in containers can be handled without use of remote tech-
niques, they are required for HLW, which requires a design for minimum pres-
ence of workers and controllers and great precision in creating space for the
waste containers—such as storages of frozen corpses—and a high degree of
utilization for cost reasons. While retaining the principle of placing waste con-
tainers in drifts or rooms and filling the space around them with dense mud,
one would instead think of creation of a monolithic sarcophagus with minimum
space between the waste containers. This would imply preparation of big rooms
with stable, high walls and flat bottom and roof. The steps taken for this would
comprise the following:
• Constructing a bottom bed for the waste containers.

• Precision concrete casting of walls and roof of the sarcophagus using low-
pH cement and talc as fluidizer.
• Mounting dense smectite clay blocks to form a lining of the sarcophagus
after which the RH is kept below 50% of the room by effective ventilation.
268 Bentonite Clay
• Preparing containers of fiber-reinforced concrete and installation of canis-

ters in them followed by casting nonreinforced concrete in the containers
for embedding the canisters. All parts of this work are made elsewhere on
the same level in the mine.
• Transporting the 25,000 kg radiation-shielded containers to the deposition
site.
• Placing the containers layerwise with intermittent pouring of clay mud for
filling joints and evening the top of the layers of containers before the next
layer is placed.
• Ending the waste placement campaign by injecting dense clay mud in the
space between the roof liner and the top of the stacked containers.
• Stopping water drainage and sealing of the boreholes.
The completed construction can be as indicated in Figure 8.15.

The second step is to prepare for closing the individual drifts and rooms when the
placement of containers has been completed. This requires construction of a concrete
bulwark for which a niche is cut in the rock and part of the bulwark constructed with
an O-ring type seal of blocks of highly compacted smectite clay (cf. Figure 8.16)
(Pusch 1994, 2008). The concrete is preferably of the previously described type with
a small content of low-pH cement and talc as fluidizer for the reasons given earlier in
the book (cf. Pusch et al. 2003).
10–25 m
Drainage holes
Containers
with HLW Blocks of highly
canisters compacted smectitic clay
Concrete cast in
form
Bottom bed of soil

with low clay
content
Concrete cast also in

deepend part
Removed EDZ in
the mine drift
FIGURE 8.15 Schematic section of SARC. The height of the drift or room can be 15–25 m
with access via a transport tunnel. De-airing of the uppermost part of the room is required for
reaching a very high degree of concrete filling.
Slot cut by water jetting and filled

with blocks of highly compacted
Blocks of compacted
smectite-rich clay (1 m deep exc.
smectite clay
in the floor where it extends
to 1.5 m depth)
Blast-generated EDZ 1 m
Shotcreted plug
Cement-poor silica
concrete
Tunnfill
Blast-generated EDZ 1.5 m
Fracture zone
FIGURE 8.16 Plug prepared by multilayer shotcreting in tunnel or drift for separating a
backfilled part from a water-bearing fracture zone. Slots containing highly compacted clay
blocks cut off the EDZ. (Data from Pusch, R., Geological Storage of Radioactive Waste.
Springer-Verlag, Berlin, Germany, 2008.) The “wet” part to the right is filled with pumped-in
cement-poor concrete rich in silica.
8.4.3.2 Bottom Bed

The box-shaped containers, each with six to nine steel canisters with 300–400 mm
outer diameter and 4.76 m length and weighing about 5000 kg, can be stacked on a
bed of well-compacted expandable clay mixed with suitably graded aggregate, or of
dried and crushed natural clay with a smectite content of 10%–15%. The bed should
have low hydraulic conductivity, low expandability and compressibility, and high
bearing capacity. This design principle was followed in constructing the founda-
tion of a 16,000 ton concrete silo for disposal of low- and medium-level radioac-
tive waste at Forsmark some 20 years ago (Pusch 1994), referred to earlier in the
book, and recommended for SARC as well. Field compaction tests using different
mixtures of Na-montmorillonite-rich clay (MX-80) and suitably graded aggregate
(Figure 8.17) gave the achievable dry densities of compositions that would fulfill the
criterion of almost no swelling or compression at complete water saturation. One
finds from Figure 8.18 that for all three aggregate types, the dry density achieved
by ordinary laboratory compaction was higher than 2100 kg/m3 for 10% clay content
and nearly 2000 kg/m3 for 30%. The swelling pressure of the 10% mixture with this
density was about 100 kPa at saturation with freshwater and 20 kPa for ocean water.
The corresponding values for the 30% mixture were 500 and 200 kPa, respectively.
270 Bentonite Clay
100
90
80
70
0–3 0–8 8–16 mm
60
a 0 60% 40%
50 10% 60% 30%
%
b
40 c 5% 60% 35%
b c
30 a
20
10
0
0.002 0.006 002 006 02 06 2 6 20 mm
FIGURE 8.17 Three suitable aggregate types for mixing with air-dry finely ground smectite-
rich clay. The vertical axis shows the percentage of particles smaller than the respective grain
size on the horizontal axis. (Data from Pusch, R., Waste Disposal in Rock, Developments in
Geotechnical Engineering 76. Elsevier, New York, 1994.)
2.2
b
c
a
P 2.1
d
t/m3 c
2.0 b
a
1.9
Bentonite/Ballast 10/90 20/80 30/70
FIGURE 8.18 Dry density of three air-dry mixtures of MX-80 granules and aggregate.
(Data from Pusch, R., Waste Disposal in Rock, Developments in Geotechnical Engineering
76. Elsevier, New York, 1994.)
The density at water saturation of the 10% mixture would be 2260 kg/m3, and the
swelling pressure a little more than 200 kPa when saturated with ocean water.
The hydraulic conductivity would be less than E−11 m/s, which is significantly lower
than the conductivity of the surrounding, blasting-disturbed rock (Pusch 2008).
For the construction of the big concrete silo at Forsmark, the 10% mixture of
MX-80 and aggregate type b were selected and they are deemed suitable also for
SARC.
The compressibility of the bottom bed is of course important only in the construc-
tion stage when the settlement of the container assembly determines its top level and
hence the amount of concrete to be injected between the piles of stacked contain-
ers and the roof (Figure 8.15). Figure 8.19 illustrates the expected compression of
air-dry 10% smectite clay rich in montmorillonite (MX-80) and 90% Fuller-graded
Vertical stress (kPa)

10 50 100 500 1000 3000
0
1
Deformation (%)
Pd = 1.65 g/cm3
Pd = 1.79 g/cm3
3 Pd = 1.91 g/cm3
FIGURE 8.19 Uniaxial compression of air-dry mixtures of MX-80 clay and well-graded
aggregate. (After Börgesson.)
TABLE 8.4
Compressibility of Unsaturated Backfill Materials
Smectite Clay
Aggregate Type Content (%) Dry Density (kg/m3) Water Content (%) m (MPa) B
1 (irregular grain 10 2150 4 200 0.75
shape)
1 10 1650 10 105 0.73
1 10 1790 10 167 0.73
1 10 1910 10 199 0.67
2 (well rounded) 15 1300 14.8 20 0.73
2 15 1900 13–15 160 0.74
2 15 1600 14.5 60 –
2 15 1300 14.8 20 0.73
Source: (After Börgesson.)
aggregate for different dry densities. For any of the mixtures in Figure 8.18, the total
compression of the bottom bed would be less than about 2% or less than 1 cm for
50 cm thickness of the bed. The ultimate water saturation of the bed would cause
some very small expansion, balancing some equally small compression by creep.
For the interested reader, Table 8.4 gives values of the compression parameters
m and β (cf. Chapter 5) evaluated from oedometer testing of mixtures with different
clay contents and types of aggregate. They can be used for predicting the compres-
sion of air-dry filling materials consisting of montmorillonite-rich clay and suitably
composed aggregate.
272 Bentonite Clay
For SARC, the pressure from the assembly of containers, that is, 500–1000 kPa,
is negligible compared to that of the Forsmark silo for which the compression of the
air-dry mixture of 10% montmorillonite-rich clay and aggregate b in Figure 8.18,
compressed under the 80 kPa pressure exerted by the silo, would give a compression
of 2–3 cm, which was also the recorded settlement of the silo in the first five years.
8.4.3.3 Containers and Canisters

The canisters made of iron or steel are assumed to contain four BWR elements with a
length of 4.67 m and similar to the titanium canister surrounded by concrete that SKB
has proposed for VDH. The containers considered for SARC need to be about 7 m
long and to have a square cross section large enough for hosting four canister tubes
with about 0.3 m diameter. The weight will be about 25,000 kg including canisters
and concrete for embedding them. The containers are preferably made of concrete
of the same type as for VDH but with carbon fiber reinforcement.* Each container
exerts a vertical pressure on its base of about 60 kPa and assuming that 15 layers
of containers are stacked, the pressure on the foundation will be about 900 kPa,
which will cause some minor compression of the dry bottom bed as described. After
complete water saturation of the repository rock, buoyancy will reduce the pressure
that will ultimately be in equilibrium with the swelling pressure. The containers are
manufactured, filled with canisters, and injected concrete elsewhere in the repository
and placed by remote robot handling.
8.4.3.4 Clay Block Liner

The liner is the most important engineered barrier to migration of radionuclides and
very low hydraulic conductivity is a fundamental property. Smectite-rich clay, such
as the Danish Holmehus clay or similar, can be utilized for VDH. The weight of
the cubical blocks can be 20 kg (handable) and they shall be mounted and fixed to
the concrete vault before bringing in the containers. The blocks will stay physically
stable by ventilation at RH < 60 until all containers are on site.
8.4.4 General Scenario of a SARC Repository

8.4.4.1 Function of SARC
We will examine the function of the most important SARC constituents in order to
estimate what advantages and drawbacks the concept has with respect to the perfor-
mance and longevity of the clays compared to VDH and a representative version of a
KBS-3V type repository constructed in virgin rock at the same depth as SARC, that
is, about 500 m below the ground surface.
When the waste containers are on site in all the rooms and the confining bulk-
heads have just been completed, the draining boreholes in the surrounding rock will
be sealed, which causes successive rise in water pressure. At this stage, backfilling
of all transport tunnels and shafts in the mine repository starts, which causes further
rise in water pressure to about 5 MPa, depending on the depth. The water pressure
begins to act on the concrete vaults, which in turn transfer the pressure to the liner
* For providing strength and ductility: no testing has yet been made.
of highly compacted clay blocks and the assemblies of containers, as well as to the
bottom bed. The clay liners start to take up water and swell, exerting swelling pres-
sure and transferring water pressures to the well-fitted container assemblies. Since
these are completely filled with concrete embedding the canisters, their compress-
ibility is very low and the rigidity very high, implying that the displacements and
movements in each of the rooms will be very small. Similarly, the piezometric pres-
sure will increase uniformly and since no open space has been left after backfilling
of all transport tunnels and shafts, movements and deformations in the entire reposi-
tory will be very small and of limited practical importance.
Saturation of the liners of dense clay blocks takes place by diffusion with a
diffusion coefficient of about E−10 m2/s (Pusch 2008), meaning that they will be
fully saturated after a few hundred years to a thousand years for a liner with 0.4 m
thickness. Not until then water will start to flow through the liners.
8.4.4.2 Bottom Bed

The very dense bottom bed in the respective room will be confined between the
concrete foundation and the heavy container assembly and when the drainage of the
repository ceases and the bed starts to become water saturated, it begins to exert a
swelling pressure on the container system. Since the swelling pressure will not be
higher than about 200 kPa and the effective pressure caused by the containers is
about 900 kPa, the bed will not undergo measurable expansion. When the entire
repository is finally water saturated, the effective pressure will drop but the bed will
still not become significantly expanded.
8.4.4.3 Clay Block Liner

The clay blocks expand and exert a swelling pressure of a few MPa on the containers.
The role of the blocks is of fundamental importance: water migrating from the rock
via the concrete, which is much tighter than the blasting-disturbed EDZ of the sur-
rounding rock, will proceed into the smectite clay causing self-sealing of the joints
between them. Before the clay is almost fully water saturated, which will take many
hundreds or some thousand years depending on the thickness of the liner, no water
will enter the containers.
8.4.4.4 Containers and Canisters

Since the wetting of the concrete containers and canister-embedding concrete will
be largely uniform, no hydration-generated failure is expected and the intact contain-
ers will be saturated at a rate that is determined by the rate of groundwater percola-
tion through the liner. For the proposed number of BWR elements per volume unit,
the initial surface temperature of the canisters can be kept below 100°C, which can
be sustained by the air-dry concrete in the containers. Realizing that not only the
clay liner but also the concrete of the vaults have very low hydraulic conductivities,
the time for water to saturate the containers is expected to be thousands to tens of
thousands of years. The temperature in the whole repository will be back to ambi-
ent long before that and more than 99% of the radionuclide inventory are no longer
dangerous. For canister temperatures of 100°C, the clay liner will get a maximum
temperature of 50°C–60°C and hence perform as in the upper part of a VDH.
274 Bentonite Clay
8.4.5 Environmental Impact
The philosophy behind the SARC concept proposed here is to provide relatively
cheap but safe disposal of HLW. In contrast to VDH and the KBS-3 disposal concepts,
isolation from groundwater is the key issue. This can be achieved using mines located
at 350–450 m depth and constructing concrete vaults with low content of low-pH
cement, hence avoiding organic fluidizers in the hardening concrete. The vaults with
their low-permeable clay liners become very tight and not until they have become
saturated, water penetration of the containers will begin. The safest conditions are
offered by abandoned mines with alkaline low-salinity groundwater, which are
provided by certain abandoned mines.* According to this concept, the engineered
barriers are of primary importance and the moderate temperature and ambient geo-
chemical conditions suggest considerable longevity of these barriers.
Like all concepts implying location at low and moderate depth, there is risk of
seismic and tectonic impacts, which would be at minimum for using mines in argil-
laceous rock with significant potential to self-seal after vibratory and rock-shearing
events. However, crystalline rock would also do, although the thickness of the smec-
tite clay barrier may have to be increased. No particular practical difficulties in the
construction phase are foreseen while the preparation of containers requires consid-
erable development and refinement.
The moderate temperature in the vault system and engineered barriers lasting
in the brief thermal period, that is, only a few hundred years, minimizes the uncer-
tainty concerning the survival potential of the canisters. Naturally, use of canis-
ters represented by the SKB’s copper/iron or the HIPOW copper versions, would be
excellent but this would turn the poor-man’s concept to one of high cost.
8.5 BOREHOLE SEALING

One realizes that selection of a suitable VDH site or, for SARC, a suitable mine, of
which there are several in Sweden and central Europe, Canada, and the United States
(Popov and Pusch 2006), must be based on proper characterization of the rock struc-
ture and determination of the rock mechanical conditions by stress measurements
and stability calculations. Much of the rock characterization will naturally have been
made by mining engineers and geologists during the mining epoch, but additional
borings for measurements and instrumentation within the mine have to be made for
converting it to a repository and these holes can cause unwanted short-circuiting of
different hydraulic regimes (Figure 8.20). Different techniques for borehole sealing
are used in practice depending on the length and orientation of the holes.
For short boreholes, the clay seals types shown in Figure 8.21 have been tested and
found useful (Pusch and Ramqvist 2006). They are primarily intended to be used in
boreholes with diameters ranging from 56 to 100 mm but at least the Basic and Pellet
plugs can be used in much wider holes. This latter type and the Couronne plug cannot
be used in very deep holes whereas the other two may be used in kilometer-deep holes.
* Backfilling of shafts and transport tunnels with carbonate rock processed for reaching low porosity is
a way of controlling pH in the host rock. Since the density of the clay liners is very high, the increase
in hydraulic conductivity and drop in swelling pressure caused by exchange of Na by Ca are negligible.
Ground surface
Zone I
Zone II
Thick black = clay plugged
Thin black = quartz/cement
Early structural/hydraulic Repository

modeling after flow meas.
in A and B while open
C
Zone III If poorly plugged between the
zones, boreholes A, B, and C form
paths ( ) for radionuclides via all
A B three zones and directly to the
ground surface via hole B.
FIGURE 8.20 Short-circuiting of major water-bearing fracture zones by poorly sealed

(plugged) deep boreholes. (Data from Pusch, R., Ramqvist, G., Cleaning and Sealing of
Boreholes, Final Report of Sub-project 1 on design and modelling of the performance of
borehole plugs, SKB IPR-06-28, SKB, Stockholm, Sweden, 2006.)
The density of the matured clay should be on the order of 2000 kg/m3, corresponding

to a dry density of 1590 kg/m3, for fulfilling the criterion that the hydraulic conduc-
tivity must be lower than that of the surrounding rock.
One recognizes the similarity of the VDH concept and the borehole sealing con-
cept Basic, which was developed in the early 1980s and has been used successfully
for sealing short and long holes as demonstrated by field experiments performed in
conjunction with the international Stripa Project (Gray 1993; Pusch and Ramqvist
2006). It has been used for sealing boreholes from a drill rig at sea for the SFR repos-
itory at Forsmark, which demonstrated the feasibility of the method. The seals con-
sist of perforated copper tubes that contain well-fitting blocks of highly compacted
smectite-rich clay that migrates through the perforation and ultimately embeds the
tube that is left in the hole, as for VDH. As described in conjunction with the pre-
sentation of VDH, the density of the clay formed between the tube and the rock is
very low in the earliest maturation phase but increases by solidification and encloses
the tube with a density that is determined by the geometry and time of maturation.
The plug segments can be a few meters long and connected by a screw mechanism,
bayonet clutch, or by clinching.
In principle, the method is very safe since one has complete control of the posi-
tion of each individual segment and good opportunities to pull up those that may
not be placeable because of unforeseen roughness of the borehole walls or of too
small radius of curvature of the hole. The tightening effect of the four concepts is
276 Bentonite Clay
(a) (b) (c) (d)
FIGURE 8.21 Four alternative borehole sealing concepts based on smectite clay: (a) the
Basic plug with highly compacted MX-80 clay blocks confined in perforated copper tubes;
(b) the Container plug with highly compacted MX-80 clay blocks contained in a cylinder
attached to drilling rods and released when the tip of the cylinder is in the desired position;
(c) Pellet plug with MX-80 pellets poured into the hole without compaction; (d) the Couronne
plug with annuli of highly compacted MX-80 clay stacked around jointed copper rods that
are pushed into the hole.
different. For Basic seals with proper selection of diameter and perforation ratio of
the tube and of the dry density of the clay blocks, the axial hydraulic conductivity of
the matured seal in ocean water is about E−12 m/s and about E−13 m/s in freshwa-
ter. This is also the case for Couronne seals that can hardly be used in holes deeper
than 50–100 m, or in horizontal holes. For Container seals, which can be used in
boreholes with a length of several kilometers, the axial hydraulic conductivity can be
of similar magnitude, whereas for the Pellet case, it will be about E−12 m/s for fresh-
water and E−10 m/s at percolation of salt water. Only the Basic and Container seals
can be installed in upward directed boreholes. A special advantage of the former is
that it can be used for electrical securing of power line towers and alike.
The principle of filling the part of boreholes with concrete where they are inter-
sected by fracture zones applies both to investigation boreholes and to VDH. For
boreholes, the most practical way of constructing the concrete seals is to widen the
hole and cast concrete followed by re-boring for installing clay seals. Figure 8.22
shows the appearance of a Basic clay seal after three weeks in a steel tube with
access to water only from the ends. Figure 8.23 shows an equally old Pellet seal. Talc
concrete with low-pH cement is recommended but the curing time is long.
The wall friction is of great importance in cases where the seal is affected by
high water pressure at one end that can extrude it from the sealed hole. The fric-
tion is determined by the adhesion to the rock, which is determined by the swelling
pressure and the wall friction (Table 8.5). Experiments verify the risk for both VDH
(a) (b)
FIGURE 8.22 Basic clay seal: (a) appearance of the seal extruded after three weeks of
maturation under 500 kPa water pressure and (b) installation of short Basic seal in 76 mm
borehole.
Can with
known
volume
filled with
pellets and
weighed
Borehole drained before
filling of pellets
(a) (b)
FIGURE 8.23 (a) Principle of filling of pellets in short holes and (b) appearance of an
extruded Pellet plug with a density of 1800 kg/m3.
TABLE 8.5
Force Required for Extruding a 2.5 m Long Basic Plug from a
76 mm Tube Simulating a Borehole, and the Corresponding
Mobilized Shear Resistance (Strength) of the Clay Skin
Load Required for Average Shear Stress in the
Time after Placement (h) Displacement (kN) Clay Skin (kPa)
24 11.6 15
48 26.6 35
96 99.6 132
720 350.0 >133
278 Bentonite Clay
supercontainer and borehole seals to be stuck before they have been moved down to
the intended position.
The force required to displace a plug of this length is appreciable already after
one day and delay by more than 12 hours in placing long seals in deep holes can
cause problems. Hence, after four days, a 10 m long plug would require a force of
about 300 kN to move it. We recognize this problem from the earlier description of
possible problems with construction of VDH seals.
REFERENCES
Beswick, J. 2008. Status of technology for deep borehole disposal. NDA deep disposal report,
Rev.7. Contract No NP 01185. Oxon.
Stein, J.S.; Sandia National Laboratories. 2009. Deep Borehole Disposal of High-Level
Gaucher, E.C., Blanc, P. 2006. Cement/clay interactions—A review: Experiments, natural
analogues, and modeling. Waste Management, 26:776–788.
Grauer, R. 1986. Uebermoegliche Wechselwirkungenzwischen Verfuellmaterial und Stahlbehälter
in Endlager. InternerBericht 86-01. NAGRA, Baden, Switzerland.
Gray, M. 1993. OECD/NEA International Stripa Project 1980–1992, Overview Volume III,
Engineered barriers. SKB, Stockholm, Sweden.
Grim, R.E., Guven, N. 1978. Bentonites. Developments in Sedimentology, 24. Elsevier, New
York.
Grundfelt, B. 2010. Jämförelse mellan KBS-3-metoden och deponering i djupa borrhål för
slutlig tomhändertagande av använt kärnbränsle. Swedish Nuclear Fuel and Waste
Management Co (SKB), R-10-13, Stockholm, Sweden.
Gueven, N. 1990. Longevity of bentonite as buffer material in a nuclear-waste repository.
Engineering Geology, 28:233–247.
Gueven, N., Huang, W.-L. 1990. Effects of Mg2+ and Fe3+ substitutions on the crystallization
of discrete illite and illite/smectite mixed layers. Int. report, Dept. Geosciences Texas
Tech University, Exxon Production Research Co, Houston, TX.
Herbert, H.-J., Kasbohm, J., Henning, K.-H. 2004. Long-term behaviour of the Wyoming ben-
tonite MX-80 in high saline solutions. Applied Clay Science, 26:275–291.
Herbert, H.-J., Kasbohm, J., Nguyen T.L., Meyer, L., Hoang, T.M.T., Mingliang, X. 2012.
Fe-Bentonite—Experiments and Modeling of the Interactions of Bentonites with Iron.
GRS–295, Munich, Germany.
Herbert, H.-J., Kasbohm, J., Sprenger, H., Fernández, A.M., Reichelt, C. 2008. Swelling pres-
sures of MX-80 bentonite in solutions of different ionic strength. Physics and Chemistry
of the Earth, 33:327–342.
Hoang-Minh, T., Nguyen-Thanh, L., Nguyen, T.-D., Nguyen, D.-T., Lai, L.T., Thuyet, N.T.H.
2014. Mineralogical characterization of Di Linhbentonite, Vietnam: A methodological
approach using X-ray diffraction and transmission electron microscopy.
Horseman, S.T., Harrington, J.F. 1997. Study of Gas Migration in MX-80 Buffer Bentonite:
British Geol. Survey. Report WE/97/7. Nat. Envir. Research Council, Swindon.
Huertas, F. 2000. Effects of cement on clay barrier performance. ECOCLAY Project. Final
Report Contract No F14W-CT96-0032, European Commission, Brussels, Belgium.
Jacobsson, A., Pusch, R. 1974. SKB Technical Report, SKB, Stockholm, Sweden.
Kasbohm, J., Henning, K.-H., Herbert, H.-J. 2000. Zu Aspekten einer Langzeitstabilität von
Bentonit in hochsalinaren Lösungen. Beiträge zur Jahrestagung, Zuerich, August 30–
September 1, berichte der Deutschen Ton- und Tonmineralgruppe e.V.-DTTG.
Kaufhold, S., Dohrmann, R. 2009. Stability of bentonites in salt solutions sodium chloride.
Applied Clay Science, 45:171–177.
Kozlovsky, Y.E.A. 1984. The world’s deepest well. Scientific American, 251(6):106–112.
Lantenois, S., Lanson, B., Muller, F., Bauer, A., Jullien, M., Plançon, A. 2005. Experimental
study of smectite interaction with metal Fe at low temperature, 1. Smectite destabiliza-
tion, Clays and Clay Minerals, 53:597–612.
Lindgreen, H. 1991. Elemental and structural changes in illite/smectite mixed-layer clay min-
erals during diagenesis in Kimmeridgian- Volgian (-Ryazanian) clays in the Central
Trough, North Sea and the Norwegian-Danish Basin. Bulletin of the Geological Society
of Denmark, 39:1–82.
Lönnerberg, B., Larker, H., Ageskog, I. 1983. Encapsulation and Handling of Spent Nuclear
Fuel for Final Disposal. SKB/KBS Technical Report 83-20, Swedish Nuclear Fuel and
Waste Management Co (SKB), Stockholm, Sweden.
Meyerhof, G.G. 1963. Some recent research on the bearing capacity of foundations. Canadian
Geotechnical Journal, 1(1):16–26.
Mohammed, M.H., Pusch, R., Al-Ansari, N., Knutsson, S., Emborg, M., Nilsson, M.,
Pourbakhtiar, A. 2013. Talc-based concrete for sealing borehole optimized by using
particle packing theory. Journal of Civil Engineering and Architecture, 7(4): 440–455.
NEDRA. 1992. Characterization of crystalline rocks. SKB Technical Report 92-39, Swedish
Nuclear Fuel and Waste Management Co (SKB), Stockholm, Sweden.
Popov, V., Pusch, R. (Eds.) 2006. Disposal of Hazardous Waste in Underground Mines. Series:
The sustainable World, Vol. 11. WIT Press, Ashurst.
Pusch, R. 1993. Evolution of Models for Conversion of Smectite to Non-Expandable Minerals.
SKB Technical Report TR-93-33.SKB, Stockholm, Sweden..
Pusch, R. 1994. Waste Disposal in Rock, Developments in Geotechnical Engineering 76.
Elsevier, New York.
Pusch, R. 1995. Rock Mechanics on a Geological Base. Developments in Geotechnical
Engineering, 77. Elsevier, Amsterdam, the Netherlands.
Pusch, R. 2011. A technique to delay hydration and maturation of borehole seals of expansive
clay. Engineering Geology, 121:1–6.
Pusch, R. 2012. The geological basis for developing concepts for disposal of highly radio-
active waste (HLW) in crystalline rock—A state of art compilation. Communicacoes
Geologicás, 99(1):61–68.
Pusch, R., Karnland, O., Lajudie, A., Lechelle, J., Bouchet, A. 1993. Hydrothermal Field Test
with French Candidate Clay Embedding Steel Heater in the Stripa Mine. SKB Technical
Report TR-93-02, Stockholm, Sweden.
Pusch, R., Kasbohm, J. 2002. Alteration of MX-80 by hydrothermal treatment under high salt
content conditions. SKB Technical Report TR-02-06, SKB, Stockholm, Sweden.
Pusch, R., Madsen, F. 1995. Aspects on the illitization of the Kinnekulle bentonites. Clays and
Clay Minerals, 43(3):261–270.
Pusch, R., Mohammed, M.H., Knutsson, S. 2014a. VDH—A case of ostrich philosophy or a
serious alternative for the disposal of highly radioactive waste. In Waste Management
and the Environment VII, Brebbia, C.A., Passerini, G., Itoh, H. (Eds.), WIT Press,
Southampton, pp. 377–390.
Pusch, R., Popov, V., Adey, R., Kasbohm, J., Knutsson, S. 2014b. Rational and economic
disposal of hazardous waste—use of abandoned mines. Journal of Earth Science and
Geotechnical Engineering, 4(2):33–54.
Pusch, R., Ramqvist, G. 2006. Cleaning and Sealing of Boreholes. Final Report of Sub-project
1 on design and modelling of the performance of borehole plugs. SKB IPR-06-28, SKB,
Stockholm, Sweden.
280 Bentonite Clay
Pusch, R., Ramqvist, G., Kasbohm, J., Knutsson, S., Mohammed H.M. 2012. The concept of
Journal of Earth Science and Geotechnical Engineering, 2(3):1–24.
Pusch, R., Ramqvist, G., Knutsson, S. 2014c. The predicted and actual wetting rate of the
buffer in repositories for high-level radioactive waste. In Waste Management and the
Environment VII, Brebbia, C.A., Passerini, G., Itoh, H. (Eds.), WIT Press, Southampton,
pp. 391–402.
Pusch, R., Takase, H., Benbow, S. 1998. Chemical processes causing cementation in heat-
affected smectite—The Kinnekulle bentonite. SKB Technical Report TR-98-25, SKB,
Stockholm, Sweden.
Pusch, R., Warr, L.N., Grathoff, G., Pourbakhtiar, A., Knutsson, S., Mohammed, M.H.
2013. A talc-based cement-poor concrete for sealing boreholes in rock. Engineering,
5(3):251–267.
Pusch, R., Zwahr, H., Gerber, R., Schomburg, J. 2003. Interaction of cement and smectite
clay—Theory and practice. Applied Clay Science, 23:203–210.
Roaldset, E., Bjaanes, E., Mjöen, K. 1998. Stability of smectite—Environmental aspects
of petroleum geological observations. Contribution to the International Workshop
on Microstructural Modelling with Special Emphasis on the Use of Clays for Waste
Isolation, Lund, Sweden, October 1998. Norwegian Technical University, Trondheim,
Norway.
Terzaghi, K., Peck, R.B., Mesri, G. 1996. Soil Mechanics in Engineering Practice. Wiley,
New York.
Warr, L.N., Grathoff, G.H. 2011. Sealing of investigation boreholes: Mineralogical and geo-
chemical borehole plug analyses. SKB Res. Rep., Ernst-Moritz–Arndt-Universität,
Greifswald, Germany.
9 Pharmacology
and Cosmetics
Through the years, clays, especially those belonging to the smectite family, have
been used in medicine and as cosmetics; some of the experience gained fits in the
book. We will take it as a starting point and then get a glimpse of the scientific
background on which modern knowhow and use of clays for body treatment rests,
internally and externally, and also discuss some other practical cases of interaction
of smectitic clay and organics.
Of course, a discussion of the interaction of organic matter and smectite clay
should have a proper ingress, most logically something on how the evolution of life
on this planet began.
9.1 ORIGIN OF LIFE

Back in 1967, the American scientist Duwayne M. Anderson (1967) discussed the
role of interfaces between silicate surfaces and water in the context of origin of life,
and several others (Bernal, Banin, DeVries, and Drost-Hansen) followed up the rea-
sonings and performed experiments to determine the role of the dynamic character
of clay surface geometry in its interaction with water and ice. He and his colleagues
concluded that low temperatures should have been advantageous in the formation of
RNA and DNA in the chemical reactions in abiotic synthesis of chemical precursors
to organic life. They showed, by use of nuclear magnetic resonance technique, that
the interface of montmorillonite and water has properties that are compatible with
some living organisms down to temperatures far below 0°C. When temperature drops
below about −25°C, the first low-temperature exotherms appear, caused by a phase of
change of the interfacial water. The properties of the unfrozen interfacial water films
in frozen clay are such that a number of organic reactions in the origin of life can have
been facilitated, such as various dynamic geometrical characteristics of clay lattices,
that can have generated ordering and coding in primitive polynucleotides.
Anderson and his colleagues believed that the mineral substrate had the func-
tions of concentration of chemical reactants, and preferential orientation of precur-
sor organic molecules. It may also have initiated and speeded up reaction rates by
catalysis, and caused suppression of randomness in recombination, polymerization,
and condensation reactions, which are all necessary prerequisites for the origin of
life. Considering the water requirements of living cells, the thermodynamic activity
of water is the probable lower limit to growth, which turns out to be the same for the
interfacial zone of one to two monomolecular layers of water and montmorillonite at
a temperature of −25°C. These investigators concluded that the temperature interval
from 0°C to −35°C is compatible with the requirements for the natural appearance of
281
282 Bentonite Clay
life on primitive Earth. Since water and ice exist on Mars and other planets and abi-
otic synthesis can have taken place on them, life can have been born on them as well.
Once initiated, the evolution started as Darwin claimed, and led to the succes-
sively developed enormous spectrum of living species that we think we know a lot
about. In this book, we are forced, by limited space and knowledge, to confine our-
selves to the matter of interaction of smectite clay and organic molecules with some
side-tracking of treatment of human tissues in practical cosmetics and dermatology
with special respect to the role of smectites.
9.2 INTERACTION OF SMECTITE CLAY

AND ORGANIC MOLECULES
An initiative taken for correlating the THMC (Temperature/hydraulic/Mechanical/
chemical) components of smectite barriers to migration of radionuclides was followed
up through an EU Project Microstructural and Chemical Parameters of Bentonite
as Determinants of Waste Isolation Efficiency (Pusch et al. 1999). It focused on the
composition of the porewater of montmorillonite-rich clay (MX-80 clay) with a dry
density of 600 to 1500 kg/m3 in equilibrium with external solutions. It followed two
main paths:
1. Use of microstructural models for defining the distribution of chemical spe-

cies in smectites and for outlining models for ion diffusion and evolution of
the porewater chemistry of cations and organic ions represented by the spe-
cies hexadecylpyridinium (HDP7y+). This substance is interesting since it
is used in drilling muds for reaching desired fluidity by altering the particle
surface charge and the mobility of porewater.
2. Performance of laboratory work, using oedometer cells for determination
of the hydraulic conductivity and swelling pressure, as well as for expulsion
of porewater for chemical analysis. The project included also sorption and
diffusion experiments.
It was verified that determination of the porewater composition through chemical

analysis only gives average values since the ions are not uniformly distributed in
the clay matrix (Figure 9.1). Thus, the interlamellar porewater, which is free from
anions, cannot be distinguished from the free porewater by such analyses but chemi-
cal modeling can be made based on microstructural models of the distribution of
porewater. Expulsion of porewater under successively increased pressure shows
a change in composition since low pressures expel water contained in large voids
whereas high pressures cause migration of water from small voids and ultimately
from the interlamellar space.
Analysis of expelled porewater gave data that agreed very well with predictions
using the code HYDRAQL/CE. The tests were made at room temperature on clay
samples confined in teflon cells that were connected to a vessel with initially pure
water separated from the cells by filters. Ions from the porewater of the initially
fluid-saturated clay diffused into the external solution by which the electrolyte con-
tent increased on expense of that of the porewater. Figure 9.2 shows the change
Pharmacology and Cosmetics 283
FIGURE 9.1 (See color insert.) Relative monovalent anion concentration (%) in a REV for
equilibrating 0.1 M monovalent electrolyte in the microstructural network of MX-80 clay
with a dry density of 1790 kg/m3 (2130 kg/m3 density at saturation with water). The picture
shows that anions (red and yellow) are confined in the center of the wide part of the widest
channels (50 μm) whereas cations (blue) are close to the mineral surfaces and in the channels
(horizontal extension). (After Muurinen, A., Lehikoinen, J., Pusch, R. POSIVA publications
Report YJT-94-22, POSIVA, Helsinki, Finland, 1994.)
16 40
External water
12 30
External water
TDS (g/l)
TDS (g/l)
8 20
Porewater
4 10
Porewater
0 0
0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5
(a) B/W (g/cm3) (b) B/W (g/cm3)
FIGURE 9.2 (a,b) Total dissolved salt (TDS) in porewater for different densities. B/W is
the inverse of the water content and represents the ratio of mass of solids to that of water.
(B/W = 1 means that the dry density was about 1400 kg/m3 whereas B/W = 1.5 corre-
sponds to 1600 kg/m3). Left: MX-80 clay saturated with distilled water. Right: MX-80 clay
saturated with ocean-type water. (Data from Pusch, R. et al., Microstructural and Chemical
Parameters of Bentonite as Determinants of Waste Isolation Efficiency, Final Report
European Commission, Contract No F14W-C95-0012, Stockholm, Sweden, 1999.)
284 Bentonite Clay
in total dissolved salt (TDS) content as function of the successively increased den-
sity of the clay from which porewater was expelled. The total dissolved salt content
increased with increasing density of the clay both for the interlamellar water and for
the free (external) porewater. This unexpected finding led to the conclusion that the
porewater chemistry was largely determined by the dissolution of accessory miner-
als such as calcite.
The uptake of organic species such as HDP7y+ in the interlamellar space has been
proposed to be as illustrated in Figure 9.3.
In the study referred to here, the clay and the organics had been freeze-dried and
ground to granules of the same size. The two air-dry substances were mixed and
compacted in oedometer cells for water saturation in conjunction with measuring
the swelling pressure and determination of the hydraulic conductivity. The unex-
pected observation that the organic substance was sorbed to about 150% of the cation
exchange capacity was explained by assuming uptake in cationic form to 100% cat-
ion exchange capacity and the rest as sorption in molecular form according to some
other bonding mechanism. Alternatively, one can imagine that the clay/water system
appeared in pillared form, that is, in widened interlamellar space. The preparation
turned out to require very high compaction pressures and densities higher than the
achieved ones, 1630 and 1760 kg/m3—at water saturation—could not be reached.
This indicates strong bonds between the organic cations and the clay lattices or very
strong interlamellar water lattices. The digitalized micrograph in Figure 9.4, repre-
senting clay with 1760 kg/m3 density at water saturation, shows a landscape of large
dense aggregates separated by open channels.
Monomolecular
+
Paraffin type
(a)
+ + +
+ +
Bimolecular (b)
+
+ +
(c)
FIGURE 9.3 (a–c) Schematic arrangement of organic molecules in interlamellar space or

between stacks of smectite lamellae. The bonds are of Coulomb type and/or hydrogen bonds.
(Jasmund, K., Lagaly, G., 1993. Tonminerale und Tone—Struktur, Eigenschaften, Anwendung
und Einsatz in Industrie und Umwelt. Steinkopff Verlag, Darmstadt. [Translation: Clay
minerals and clays—Structure, properties, use and application in industry and society].)
FIGURE 9.4 (See color insert.) Digitalized transmission electron micrograph of HDPY+-
treated MX-80 clay with 1760 kg/m3 density in water saturated form (thin-section by ultra-
microtomy). The horizontal edge length of the micrograph is 1 mm. Channels with up to
100 μm width separate large dense and homogeneous aggregates. The white and green areas
represent open space, the red area represents dense clay matrix, and the black area shows very
dense particles representing accessory minerals. (Data from Pusch, R. et al., Microstructural
and Chemical Parameters of Bentonite as Determinants of Waste Isolation Efficiency, Final
Report European Commission, Contract No F14W-C95-0012, Stockholm, Sweden, 1999.)
TABLE 9.1
Swelling Pressure ( ps) of Untreated and HDPy+-Treated MX-80 Clay
Saturated and Permeated with Distilled Water
Density at saturation (kg/m3) ps, kPa of untreated MX-80 ps kPa of HDPy+-treated MX-80
1630 350 600
1760 600 700
The influence of and impact on the swelling pressure due to the presence of this
type of organic species can be seen in Table 9.1. For a density of 1630 kg/m3, the swell-
ing pressure was almost twice as high as that for untreated MX-80. For 1760 kg/m3,
it was, however, about the same as that of untreated MX-80. The diagram in Figure 9.5
shows that development of the pressure was associated with invasion of the interlamel-
lar space by the organic cations, moving in a diffusion-like way. It probably involved
expulsion of the original water molecules in this space.
A study of smectite clay mixed with fat from the refinement of margaric com-
ponents in the food industry for investigating the possible use of such material gave
similar results. Three identical samples of dried and compacted mixtures were water
saturated in oedometers for determining the hydraulic conductivity and swelling
pressure, two of them being tested for 14 days, and the third for 22 days. The material
data were as follows: Sample 1—density at water saturation 930 kg/m3 (dry density
625 kg/m3); Sample 2—saturated density 1040 kg/m3 and dry density 720 kg/m3; and
Sample 3—saturated density 1270 kg/m3 and dry density 981 kg/m3. Incineration
and drying at different temperatures showed that the content of fat was about 30%
286 Bentonite Clay
1000
1.76 g/cm3
1.63 g/cm3
800
Swelling pressure (kPa)
600
400
200
0
1 10 100 1,000 10,000 100,000 1,000,000 1E+07
Time (s)
FIGURE 9.5 Development of swelling pressure in HDPy+-treated MX-80 clay. (Data from
Pusch, R., Muurinen, A., Lehikoinen, J., Bors, J., Eriksen, T., Microstructural and Chemical
Parameters of Bentonite as Determinants of Waste Isolation Efficiency. Final Report EU
Contract No F14-CT95-0012, Stockholm, Sweden, 1999; Pusch, R., Yong, R.N., Microstructure
of Smectite Clays and Engineering Performance, Taylor & Francis, London, 2006.)
of the solid mass and that the clay/fat material to be tested had a water content of
5.6%. As for the HDPy material, the fat hindered effective compaction beyond a dry
density of about 1000 kg/m3.
The conclusions from the tests were, in summary, as follows:
• The hydraulic conductivity, determined using a hydraulic gradient of

20–30 m/m, dropped with time and approached the values for untreated
montmorillonite-rich clay (Figure 9.6). The densest clay had a hydrau-
lic conductivity for percolation with distilled water of 4E-11 m/s. For the
second densest material, the graph shows that the hydraulic conductivity
dropped strongly to begin with, approaching then a conductivity of about
2E-10 m/s.
• Percolation with distilled water gave strong green-colored discharged
water indicating that some of the fat sorbed on and in the aggregates was
washed out. This can have accelerated hydration and formation of clay gels
in the pore space and thereby reduction of the conductivity. The fact that
discharged water still had a weak green color after 14 days indicates that
expulsion of fat molecules in interlamellar positions was not complete even
after this time.
The swelling pressure was very low for all samples. Only Sample 3, with the high-
est density 1270 kg/m3, gave a significant pressure, 120 kPa, which indicates that
25
Hydraulic conductivity multiplied

20
15
by E10 m/s
10
0
0 5 10 15
Time after start (days)
FIGURE 9.6 Evolution of the hydraulic conductivity of Sample 2.
Intercalated
FIGURE 9.7 Possible intercalation of smectite lamellae and protein molecules. (After
Lambert, J.-F., Poncelet, G., Top. Catal., 4, 43–56. 1997.)
the interlamellar space of the mixes with lower densities was fully expanded. The
swelling pressure recorded for Sample 3 was three times higher than that of untreated
MX-80 clay with the same density and porewater chemistry, which suggests that the
pressure was built up by successive, slow migration of organic molecules into the
interlamellar space, possibly in conjunction with intercalation (Figure 9.7).
The establishment of organic molecules expanded the interlamellar space and
caused the recorded swelling pressure as in the case of HDPy uptake. This is compat-
ible with the successive reduction of the hydraulic conductivity. Figure 9.8 illustrates
the typical microstructural constitution of the fat-saturated MX-80 clay character-
ized by embedment of smectite stacks and aggregates.
288 Bentonite Clay
40 μm
FIGURE 9.8 Montmorillonite-rich clay (MX-80) mixed with margaric fat as seen in a
scanning electron microscope. The clay stacks and aggregates are embedded in fat.
9.3 INTERACTION OF CLAYS AND ORGANICS

IN MEDICAL TREATMENT
9.3.1 Issues Considered
The literature contains a very large number of reports on use of clay for various
medical treatments and the interested reader is referred to documents such as 1954,
Annals of the New York Academy of Science; 1961, Medical Annals of the District of
Columbia; and 1996, Industrial Toxicology for further information. We will confine
ourselves here to consider the scientific basis for using clays in natural medicine and
dermatology such as healing of wounds and removal of radionuclides from the skin
and the gastric system. Something will also be said about use of clays for cancer
treatment and ways for counteracting and preventing some forms of this disease
using creams for UV protection.
9.3.2 Clays in Natural Medicine

As pointed out by many investigators of natural medicine, the use of clay has been
transferred in folk tradition from generation to generation in a number of different
cultures all over the world. Since ancient times, people have been using clays for
healing diseases and ailments, as wounded animals do. As is well known, they cover
themselves with clay by diving into mud and instinctively eat it. Since smectites are
common weathering products in Africa, it is no wonder that elephants and hippos
take their curing baths in muddy rivers on this continent.
Eaten clay migrates to the digestive tract causing trapping and neutralization of
toxic matter, or sorption of hazardous molecules and cations for discharge in the
natural way. Particularly important techniques for curing digestive diseases were
developed in ancient Egypt by priests, and even today, the knowledge from this
epoch is used by modern natural healers. They have enough experience of the puri-
fication potential of clay for oral intake and for applying facial clay masks for cos-
metic treatment. The same is noticed from the ancient Aztec Indians’ culture and
a number of similar cases, all associated with the occurrence of smectitic clays.
An interesting example is that Greek women in the countryside until recently and
possibly even today use smectitic clay as shampoo and for cleaning the skin on the
scull. This shows that montmorillonite, palygorskite, saponite, and other members
of the smectite family are particularly effective for use in dermatology, one reason
being their ability to sorb organic and inorganic debris. We realize, of course, that
eaten smectite clay is more effective for removing unwanted polar molecular and
cationic species from the gastric system than kaolinite and illite because of its high
cation exchange capacity and crystal lattice charge.
In modern natural medicine, it is often claimed that clay is anti-inflammatory,
antiseptic, and disinfecting, internally and externally, even for antibiotic resistant
bacteria such as Streptococcus and it is known to heal normally fatal infectious dis-
eases such as the painful Buruli’s ulcer, which eats the skin, muscles, and bones.
9.3.3 Clay for Healing Wounds

This issue has been in focus of natural medicine since stone-agers hurt themselves or
got bitten by insects and wild animals and historians claim that even Neanderthalers
used clay for wound healing. Naturally, a filtering effect of clay pasted on wounds is
that it prevents bacteria and virus to enter them but it is not obvious how healing can
take place in the absence of oxygen caused by the tightness of the clay coating. The
reason can be that oxygen used up in the healing process causes a sink that drives
these molecules through the less than one millimeter thick paste by diffusion. Thick
clay pastes would retard microbial migration very much, which is in fact indirectly
proven by the need of frequent removal of very thin paste and application of new
paste after some delay. The detailed role of molecular diffusion in the healing pro-
cess is almost totally neglected in dermatology.
Use of very surface-active, finely dispersed silicates such as smectite clay for sore
treatment is known to be effective to arrest bleeding. In the olden days, engine crews
on steam vessels used to coat frequent burns and wounds on their hands and arms
with sticky clay paste that stopped bleeding and speeded up healing. The efficiency
of such treatment is illustrated in Figure 9.9, which shows the arrest of intense bleed-
ing from a deep cut in a mouse by spreading fine granulate of zeolites on the wound.
Powdered air-dry smectite has the same impact. This ability is due to activation of
the blood-clotting protein VII by interaction with smectite minerals in conjunction
with the formation of a scab of secretory components. The tight clay/secretory scab
prevents infection and causes quick healing.
An example of the constitution and role of smectite clay pasted on a wound
in a man’s hand is shown in the transmission electron micrograph in Figure 9.10.
The gray wavy matrix is smectite clay covering and being integrated in the wound,
which is thereby sealed off and protected from microbial attack. Healing took place
in two days.
290 Bentonite Clay
FIGURE 9.9 (See color insert.) Blood arrest by treating a deep cut in a mouse with zeolitic
granules (clinoptilolite/chabasite). One sees the abdomen region cut up and granulate applied
causing immediate stop of bleeding by formation of a tight scab. (Patented by Durtec GmbH,
Neubrandenburg, Germany.)
3.00 μm
×1000 120 kV 25.08.2000
Microscope: J1210
Preparation: Section
Sample: White resin
FIGURE 9.10 (See color insert.) The transmission electron micrograph of ultrathin section
of wound present in the skin. The yellow arrows show clay in direct contact with exposed
epidermis. The blue arrow points to the remnants of the shallow epidermis embedded in
the paste of montmorillonite-rich clay that is seen as a wavy mass. Magnification: 100×.
(Microscopy by Greifswald University, Greifswald, Germany.)
These and similar investigations and cases described in historical and modern
medical documents demonstrate that smectitic clay has a capacity to stop bleed-
ing and help curing wounds. The most important functions of smectitic clay pastes
in this respect are (1) the cleaning effect by repeated treatment, (2) the capacity to
remove or bring in cations and organic molecules in shallow and deep parts of the
skin, (3) the effect of a tight barrier to bacteria and viruses, and (4) the ability of
activating blood-clotting proteins of exposed tissues. They are expected, as indicated
in numerous literature references, to be especially effective in treatment of acne by
the strong potential to sorb secretory components, to remove bacteria and virus from
the skin, and to provide protection against invasion of microbes. It is essential in all
these applications to frequently remove and replace the paste, which is preferably
montmorillonite of high purity. A very valuable property of this and other members
of the smectite family is their thixotropic behavior: they are easily smeared on and
washed off but stiffen and stay on the skin when left to mature.
9.3.4 Extraction of Hazardous Elements Poisoning the Human Body

9.3.4.1 Principle
Toxic cations and anions, including radionuclides that have been inhaled or having
entered the gastric system of human beings and animals via eaten food such as milk
and vegetables, can be removed by use of clay preparates. There are examples of
such treatment from reported cases of arsenic and copper having poisoned drinking
water, and of events of radioactive contamination caused by failing nuclear reactors.
Radionuclides in ionic form (Am and Tc) can be amphoteric and difficult to remove
from the body by ion exchange processes. However, the three most dangerous species
iodine, cesium, and strontium have defined charges and may be removed from the body
by use of orally taken tablets of clay minerals through the action of ion exchange. Certain
minerals of this sort, that is, the smectite family, especially montmorillonite, have a high
ion exchange capacity for cations. Others have the same ability for anions, such as very
finely ground kaolinite (Grim 1953). Effective treatment involves release of the toxic
elements and uptake in the body of cations or positively charged organic molecules
earlier sorbed on the clay minerals, and given off from them. Natural excretion of the
mineral particles, which should be smaller than 1 μm, from the gastric system causes the
desired removal of the toxicants. Pretreatment is not meaningful while taking the tablets
immediately after eating or drinking contaminated water or food will be effective.
Basically, the movement of cations that are not fixed to any substrate is controlled by
the concentration gradient. The main mechanism that will remove positively charged
radionuclides is the creation of a sink in the gastric system formed by the sorbing
smectite crystallites. This requires that toxic ion species are preferentially taken up
from the system by the clay. The actual ranking of cations at the prevailing low pH
is Sr > Cs > Ca > K > Na (Grim and Guven 1978), which means that cesium and
strontium, the latter being deposited in bones, are more effectively sorbed by the clay
than the normally appearing cations Na, K, and Ca. The most harmful anionic radio-
nuclide is iodine, which will accumulate in the thyroid gland, a process that can be
counteracted by taking non radioactive isotopes of iodine in tablet form but intake of
tablets of strongly anion-adsorbing clay may in fact be an alternative treatment. In both
292 Bentonite Clay
cases discharge of unwanted radionuclides is through normal excretion. The greatest

potential of smectite to adsorb ions is through cation exchange and chemisorption,
the amount of sorbed ions being determined by the exchange capacity expressed in
meq/100 g, as we know from Chapter 6. The uptake of ions is associated with release
of initially sorbed ions, the loss of which must be accepted by the body for considering
the treatment as successful. The introduction of clay in the gastric system may have
negative effects per se and any treatment by use of phyllosilicates must hence be pre-
ceded by an investigation of what acceptable doses of this substance really are.
Smectitic clay in medical care has been used extensively for a very long time, ear-
lier in the form of poorly defined bentonites and more recently as refined sepiolite,
palygorskite, and montmorillonite. The smectite mineral montmorillonite has the high-
est cation exchange capacity and is therefore preferred in pharmacology. It is on the
order of 100 meq/100 g. Kaolinite is less cation-exchanging but extremely fine-grained
kaolinite has both cation and anion exchange capacities of up to about 100 meq/100 g
(Grim 1953). These two minerals are of primary interest for dermatological treatment.
9.3.4.2 Preference of Clay Minerals

We will focus here on pure montmorillonite and kaolinite and disregard from acces-
sory minerals such as crystalline and amorphous quartz, which are present in natural
bentonites and other smectite-rich clays.
A quick repetition of the basis for using montmorillonite and kaolinite in derma-
tology and cosmetics implies the following:
• Montmorillonite crystallites charged with Na form stacks with 3–5 lamel-

lae between which cations are located that neutralize the charge deficit of
the crystal lattices (van Olphen and Fripiat 1979; Chipera and Bish. 2001;
Vogt et al. 2001; Kasbohm et al. 2005; Pusch and Yong 2006). The particles
consisting of stacked lamellae are oblate with a thickness of up to 100Å
(10 nm), and a length and width of 0.1–1 μm. Water molecules attached to
the cations or forming separate networks are established in the interlamel-
lar space on wetting, causing expansion of the stacks to a certain maxi-
mum thickness. The edges of the stacks expose hydroxyls under normal
or weakly acidic pH conditions. The basal surfaces can be regarded as
electrical double-layer systems with the mineral surface consisting of hexa-
gons of Si atoms. They form negatively charged substrates to which water
molecules are adsorbed, making up external hydrates with considerably
enhanced viscosity within about 5Å–10Å distance from the surface. Here,
cations are distributed in a cloud with maximum concentration close to the
mineral surface. These cations can be exchanged but the major part of the
exchangeable cations are those in the interlamellar space.
• Kaolinite crystallites form stacks with about 20–100 lamellae with very
little deficit of the crystal lattices (Grim 1953; Kranz et al. 1990). The stacks
are oblate particles with a thickness of about 0.1–0.5 μm, and a length and
width of 0.1–2 μm. The edges of the stacks expose hydroxyls whereas the
basal surfaces consist of hexagons of Si atoms forming a negatively charged
substrate on one side of the stacks and of hydroxyls on the opposite. Water
molecules are adsorbed on these surfaces, making up external hydrates.

The individual lamellae are strongly held together, which makes the
crystallites relatively big and the specific surface small. By very effec-
tive grinding, the amount of exposed hydroxyls increases, which strongly
enhances the cation and anion exchange capacities.
9.3.4.3 Interaction of Clay Minerals and Cells in the Gastric System

Experience is available from the literature with respect to the toxicity or harmfulness
of orally administrated clay minerals as summarized here. For estimating the pos-
sible harm to the body that the crystallites may have, their interaction with organic
tissues exposed in the gastric system needs to be considered. This involves assess-
ment of the chemical and physical force-fields established at the contact between the
clay crystallites and the cells, which are of different type and constitution in different
parts of the system.
Two stages need to be considered: first, reactions between contacting crystallites
and organic tissues in the gastric system, and second, the role of elements released
from dissolving clay crystallites. In general, organic cells viewed in cross section
appear to be organized into four parts, externally a slime layer surrounding a mor-
phologically differentiated rigid cell wall, which is separated from the internal
cytoplasm by a plasma membrane. The materials giving rigidity to the cell wall are
carbohydrates in the form of polysaccharides with several functional groups such
as OH and also protons and oxygens exposed on the walls surfaces. The slime coat-
ing, which is considered as nonliving secretion or excretion with no active role in
the metabolism, has a complex composition. In principle, it is made up of polysac-
charides but in contrast to the rigid cell wall, the slime layer has an amorphous and
jelly-like consistency. Staining experiments resulting in coloration have indicated
the presence of polysaccharides with a characteristic ─CHOH─CHOH─ group. The
cell wall can be considered as a rigid, mosaic-like structure (Lamanna and Mallette
1959). The major constituent of ordinary cells is water, which makes up 75%–90%
of the mass. The organic matter is composed of protein (40%–80%), carbohydrates
(1.5%–36%), and lipids (0.4%–39%).
Coupling of the cell walls to montmorillonite or kaolinite stacks occurs by establish-
ment of van der Waals forces and hydrogen bonds between surface-active functional
groups such as carboxyls (COOH), hydroxyls (OH), carbonyls (C─O), and quinone
(C=O). Coulomb bonding can also be assumed to take place between positively charged
edges of smectite stacks and negatively charged cells, and, like in aggregation of clay
particles, by polarization of edge-adsorbed polyvalent cations (Pusch 1970, 1973).
The strength of the bonds between clay stacks and cells is estimated to be on the
same order of magnitude as between aggregated clay particles, which means that
stronger coupling is expected to montmorillonite than to kaolinite crystallites. Very
fine-grained kaolinite may behave like montmorillonite in this respect.
The following medical effects can be imagined:
• Competition between minerals and cells for water

• Osmotic effects
• Agitation
294 Bentonite Clay
As to the first effect, both the cells and the stacks of minerals are, or become,
completely water saturated and there is no transfer of water between them. Osmotic
processes will take place by ions contained in the cell phytoplasma being liberated
and given off to the minerals through the cell wall and replaced by ions released
from the minerals. This process, controlled by the sorption capacity of the organic,
as well as mineral components, is in fact the desired mechanism in removing unde-
sired ions from the body, provided that the minerals can and will be excreted.
The literature provides relevant information on the effect of oral intake of
bentonite (smectite, presumably montmorillonite) through public investigations
(cf. Medical Annals of the District of Columbia 1961). Oral administration dose
rates of 1–10 g/day of smectite or kaolinite appear to be far below any critical level.
At intraveneous injection, which cannot be recommended at present without com-
prehensive research, the corresponding dose should be lower than about 0.1 g/day.
Positive in vivo effects of treatment with smectite were found from clinical
work on treating diarrhea caused by virus infections, food allergy, spastic colitis
(inflammation of the mucous membrane of the large intestine), and food poisoning.
Substantial relief was recorded in 97% of the cases; 80% of abdominal cramps, 71%
of headaches, 85% of nausea, and 100% of weaknesses. These effects have been
ascribed to detoxification of the human alimentary canal, the smectite acting as an
intestinal evacuant when used in gel form. No mutagenicity has been reported.
Negative effects have not been reported except when the intake of smectite in the
form of bentonite exceeded 50% of the diet of experimental animals. At 25% no ill
effects were recorded (Annals of the New York Academy of Sciences 1954). Animal
studies reported publically (Industrial Toxicology 1996) indicate that acute toxicity
of bentonite for laboratory animals is low by all routes of administration. Prolonged
feeding of almost pure bentonite caused benign tumors and toxic injury in the liver
of mice and rats and intravenous bentonite injection is critical (pulmonary edema)
for 10–35 mg bentonite per kilogram body weight. These organizations have found
that the content of crystalline silica in bentonite but not the bentonite itself may
cause carcinogenicity (Harley et al. 1988).
In vitro experiments have been made using commercial bentonite brands (Annals
of the New York Academy of Sciences 1954). The tests, which were performed at
Sias Memorial Laboratory, Brooks Hospital, Massachusetts, had the form of mixing
150 ml clay gel and 1 ml bacteria mixture containing 5E+6 bacteria Serratia marc-
escens, Escherichia coli, and Proteus mirabilis (all gram negative) and recording the
bacterial count after 4–24 hours. The reduction in bacteria of these sorts by sorption
was 85%–100%. In corresponding experiments, with Staphylococcus aureus (gram
positive), the reduction was only 21%–40%. The difference between gram-negative
and gram-positive bacteria, which show different reactions on staining, is primar-
ily that positive ones, in contrast to gram-negative bacteria, are mechanically more
stable and susceptible to penicillin. These differences appear to be related to the
nature of the cell walls.
Exposure of clay crystallites to the low pH conditions in the gastric system
will cause dissolution by which Si and Al are released. From a medical point of
view, it is essential to realize that the silica content of brain, liver, spleen, lung, and
lymph nodes increases with age, and high Si levels are found in the senile plaques
of Alzheimer’s dementia in conjunction with amyloid (Bendz and Lindqvist 1978).

In contrast, the silica content of aorta, skin, thymus, and hair decreases with age.
It is an attractive idea to increase the latter content without enhancing the silica
content in the brain and liver, for which short-term exposure to dissolving clay min-
erals may be a possible procedure. Still, with the present poor understanding of the
migration and reaction with organic compounds of silicons, it is proposed that means
should be taken to minimize uptake of silica by the gastric system, which does not
encourage oral intake of clay minerals.
While silicon biointegration has no role in energy production as in unicellular
algae-like diatoms and DNA synthesis, it is intimately involved with macromol-
ecules displaying endless variations of complex interactions. However, the uptake
of silicons from the gastric system into the cells of blood vessels and further to the
blood requires that there is no finite limit of adaptive mechanisms by which normal
cells and tissues can dispose of excess silicon. The major question is how can the
H4SiO4 molecule be stripped of the water molecules and enter cell walls and pen-
etrate them to form organosilicons. Again, the uncertainty in this matter suggests
that oral intake of clay minerals is left out.
9.3.4.4 Radionuclides
Cesium, strontium, and iodine are main contaminants of shallow soil at nuclear
weapon tests and at nuclear reactor failure as at Chernobyl and Fukushima.137Cs is
particularly hazardous to public health because of its strong gastrointestinal absorp-
tion and distribution in soft tissues such as muscle (Harley et al. 1988). The bio-
logical half-life in adults is from 50 to 100 days causing ulcerations and necrosis of
tissues, as well as susceptibility to infection. Baths in smectitic clay muds and oral
intake have been recommended for removal of radionuclides held by and contami-
nating skin must be made instantly after exposure to air and plants carrying such
species. This is the SPA approach that works but has an after-effect that is neglected:
used clay is charged with radionuclides and has to be safely deposited. The involved
mechanism is simple and self-evident: the smectite clay takes up strontium and
cesium and replaces these ion species by exchangeable cations, usually Na or Ca,
with which the clay has been doped.
There is comprehensive experience from removal and excretion of radioactively
contaminated nutrients by mixing the food with smectite clay powder, which can
be accepted to make up more than 1% of the body weight per day without caus-
ing problems. The process leading to release of radionuclides from tissues and the
gastronomic system is basically cation exchange; using Na or Li montmorillonite
the cations initially sorbed on and within the smectite particles are replaced by posi-
tively charged radionuclides. Iodine, carrying a negative charge, cannot be removed
by this technique but other treatments using inter alia organic material can be used
for the same purpose. Very finely ground kaolinite has an appreciable anion sorption
capacity and may be used for removing iodine.
Cation exchange experiments have shown that the antibacterial component of
the clay can be removed, implicating exchangeable cations in the antibacterial pro-
cess. Furthermore, aqueous leachates of the antibacterial clays effectively kill bac-
teria. Progressive heating of smectite clay leads first to dehydration (200°C), then
296 Bentonite Clay
d ehydroxylation (550°C or more), and finally to destruction of the clay mineral

structure by (~900°C), according to Grim and Gueven (1978). By identifying the
elements lost after each heating step, and testing the bactericidal effect of the heated
product, one has eliminated many toxins from consideration (e.g., microbes, organic
compounds, and volatile elements) and identified several redox-sensitive refractory
metals that are common among antibacterial clays. It has been concluded that the
pH and oxidation state buffered by the clay mineral surfaces are key to controlling
the solution chemistry and redox-related reactions in and at the bacterial cell wall.
9.3.5 Potential to Cure Cancer

People who have cancer, autoimmune diseases, or heavy metal and chemical poison-
ing, have an overall body pH below 7.0, which means that their cells are not being
properly oxygenated and that this can lead to accumulation of toxins. The perfect body
pH is 7.4 but people who are actively using naturopathic principles in detoxification
are advised to keep their pH somewhat more alkaline since cancer cells cannot sur-
vive in a body with pH equal to or higher than 8. The body’s oxygen saturation level
is determined by its overall pH. Montmorillonite clay in calcium form is described
as living clay since it principally consists of minerals that enhance the production of
enzymes in all living organisms (Cancer & Natural Therapy Foundation of Australia
2014). It swells only very little for the densities that are relevant in clay therapy, as we
saw in Chapters 3 and 4. It is interesting that preventive action is taken by American
farmers by feeding such clay to their livestock since they believe, by experience, that
it treats various veterinary diseases. They also gain more weight on less feed!
There are a number of witnesses claiming that clays and various organic sub-
stances have cured skin and internal cancer but no stringent scientific examination
and evaluation of such cases has yet been made to certify their validity. It seems,
however, that such proof has been certified concerning the capacity of smectites to
be an antibacterial agent (Williams and Haydel 2010). These investigators focused
on the ancient practice of geophagy, which is the practice of eating earth materials
containing clay minerals, and on pelotherapy implying adsorption of toxins from
skin and heating to stimulate circulation for rheumatism treatment. Their study dis-
tinguished between healing clays and antibacterial clays and they found that smec-
tites serve to clean and refresh the skin surface and to aid in the healing of topical
blemishes, and are thence the major selling point for many cosmetics. Clays lacking
high cation-exchange capacities, such as I/S mixed-layer clay minerals and kaolinite,
can serve as a nutritional supplement for essential elements. Kaolinite obviously has
many medically beneficial attributes primarily related to its ability to adsorb lipids,
proteins, and some bacteria and viruses.
The high cation exchange capacity of certain smectites has been targeted in the
creation of inorganic antibacterial materials. Synthetic antibacterial clay minerals
have been prepared by exchanging their native ions with known antibacterial ions
such as Ag and Cu.
In vitro testing by various investigators, involving separation of the different size
fractions of the clay and multiple washings and centrifugation cycles in deionized
water, did not eliminate the bactericidal effects of the minerals, hence indicating
that the antibacterial agent is not highly soluble. After cation exchange, it was found
that smectite-rich clay samples (<0.2 µm fraction) no longer killed E. coli, indicating
that the antibacterial agent is linked to ions that are presumably in the exchangeable
sites of the clay. Heating to 550°C does not change the ability of smectite clay to kill
E. coli, hence verifying that the antibacterial agent is not an organic compound as
these are certainly eliminated at this high temperature.
Initial investigations indicate that particular natural clay minerals can have strik-
ing and very specific effects on microbial populations. These effects can range from
enhanced microbial growth to complete growth inhibition, these opposite effects
taking place with clay minerals of similar structure and bulk crystal chemistries
(Williams and Haydel, 2010). One concludes from work and modeling made by these
and other investigators that trace elements or unstable metal complexes stabilized
by the organizing potential of clay minerals like smectites can serve as antibacterial
agents.
9.3.6 Summing Up on Clays in Modern Medicine

As indicated by the investigations referred to the function of smectitic clay, one can
identify three properties that are effective:
• The members of this family of clay minerals have a very large specific sur-
face area (about 800 m2/g for montmorillonite) and therefore a very strong
potential to react with organic and inorganic species in solutions.
• The cation exchange capacity of smectite is high giving it a strong abil-
ity to sorb polar molecules and cations as described earlier in the book.
Different cations and molecules have different exchange potentials; bi- and
polyvalent ones can usually replace monovalent species depending on the
concentration of the solution.
• Smectites can be antibacterial if they are charged with metal ions that are so
located in the clay crystals that they can interact with contacting microbial
populations, prohibiting their growth. Such metal ions can be Cu or Ag.
• Taken internally or used externally, smectite clay draws toxins, cellular
wastes, and pollutants from tissues and organs and traps them inside its
network-like microstructure. It also stimulates the healthy and regenerative
processes in tissues and organs.
• Clay, especially artificially charged montmorillonite, brings minerals and
trace elements that are indispensable for health and beauty. Necessary trace
elements include calcium, aluminum (vital to lots of enzymatic reactions in
the body), magnesium, potassium, sodium, sulfur, iron, phosphorus, zinc,
and manganese. Silica is valuable for the skin, aorta, and liver but not for the
brain. Clay-based creams for dermatological use are hence recommended,
whereas oral intake of clay for providing the gastric system with silicons is
not. Still, short-term exposure may be acceptable and even valuable.
• For easy access to clay-born trace elements for organic tissues and dis-
charge of exchanged ions and molecules, the clay should be placed as a
mixture with water.
298 Bentonite Clay
9.4 SUN PROTECTION

In contrast with clay used for medical treatment, UV-protective clay is c ommonly
mixed with oil or glycerol because the color can make the cream more a ttractive and
desiccation retarded (Vicente et al. 1989; Longstreth et al. 1994; WHO 2002).
UV radiation is a major reason for skin cancer that is expected to be increas-
ingly common due to the ongoing degradation of the earth’s ozone layer (Setlow
1974; WHO 2002). It has recently been reported that the most popular sun protective
creams on the market reduce penetration UV radiation only by a few tens of percent
and Swedish medical authorities have recommended manufacturers to develop more
effective products.
Comparison of a well-known commercial cream, called Tone here, and two
creams MD1 (014/99) and MD2 (015/99) prepared for a scientific pilot study indi-
cates that I/S mixed-layer clay as a base can be competitive. The study was focused
on UV penetration of thin smeared layers of the respective substances.
9.4.1 Pilot Study of the Performance of Organic-

and Clay-Based Sun Creams
9.4.1.1 Tested Creams

The producer of Tone, taken as a representative of typical organic sun creams, is a U.S.
company claiming that it has, like many other commercial sun protective creams, the
following major active ingredients: homosalate, ethylhexyl, p-methoxycinnamate,
2-ethylhexyl salicylate, and oxybenzene. Inactive fillers can be water, glycerine/
glycerol, acrylates, sodium hydroxide, and urea.
The clay-based cream MD1 (014/99) is light-grayish and is composed of mont-
morillonite-rich smectite (Keramont) and glycerine (glycerol, 1,2,3-propantriol) in
proportions 100 g clay with 9% water content/125 ml glycerine. The specific den-
sity of dehydrated smectite is 2780 kg/m3 whereas the water-free glycerine has
a density of 1270 kg/m3. The density of the tested mixture was 1500 kg/m3. The
second cream MD2 (015/99) is dark-grayish and is composed of I/S mixed-layer
(illite/chlorite/montmorillonite) Friedland clay (Figure 9.11) and glycerine (glycerol,
1,2,3-propantriol) in proportions 100 g clay with 1.5% water content/105 ml glycer-
ine. The density of the mixture was 1640 kg/m3.
9.4.1.2 Testing
A 32 W UV-radiation unit of type isel-UV-Belichtungsgerät 1 and a sensor of type
Digital Luxmeter LX-101 were used for the experiments. The samples were applied
in 2 μm foils with very low UV absorption by pressing the cream to form a film of
uniform thickness: 10–15, 20–30, and 40–60 μm. The samples were exposed to radi-
ation for 50 s while recording the penetrated radiation. The average wavelength was
3650Å (365 nm), which represents the upper part of the UV range (140Å–3900Å).
UV radiation with lower wavelength than 2900Å is almost totally absorbed by the
atmosphere and by the most shallow, dense skin. The thicknesses correspond to the
amount of cream normally applied to the skin. 10 μm is estimated to represent a
minimum while 60 μm is a rather thick coating.
[Al Si3 O10]
OH
(OH)
Al
OH
Muscovite
OH (OH)
10.02Å
Layer B [Al Si3 O10]

K
[Al Si3 O10]

OH
(OH)
Al
OH
OH (OH)
Montmorillonite
[(Si,Al)4 O10]
14.0Å
Layer A
Exchangeable cations +nH2O
[(Si,Al)4 O10]
OH
OH (OH)
Al, Fe, Mg
OH
(OH)
[(Si,Al)4 O10]
O OH (OH) Al, Fe, Mg Si, party Al
FIGURE 9.11 Proposed crystal structure of Friedland Ton muscovite-montmorillonite.

It can be described as a system of alternating layers of irregular sequences (e.g.,
…AABAAABBABAAABAA…). The proportion between A and B is 70% A (montmorillon-
ite) to 30% B (muscovite). The cation exchange capacity is about 50 meq/100 g. (Data from
Pusch, R., Yong, R.N., Microstructure of Smectite Clays and Engineering Performance, Taylor
& Francis, London, 2006.)
While Tone can form extremely thin films, the smectite-rich MD1 cream was too
stiff to form thinner films than 40–50 μm, hence requiring softening for serving as
a sun protective cream. The difficulty in preparing very thin MD1 films is believed
to have been caused by thixotropic strength regain after remolding. The MD2 cream
based on Friedland clay has a lower surface potential than the smectite clay and
hence higher fluidity and less thixotropy for the same glycerin content. It could
therefore form 10–15 μm films, which appeared brownish in the foil.
9.4.1.3 Results
Figure 9.12 shows the results from testing of the commercial sun protective cream
and the MD2 cream both with 10–15 μm film thickness, whereas Figure 9.13 shows
300 Bentonite Clay
600
500
UV flux in lux
400
300
200
100
0
0 20 40 60
Time after onset (s)
FIGURE 9.12 (See color insert.) UV penetration of 10–15 μm thick films of tested sun pro-
tection creams. Upper: reference test of foil without clay. Middle: Tone. Lower: Serie MD2.
600
500
UV flux in lux
400
300
200
100
0
0 20 40 60
FIGURE 9.13 (See color insert.) UV penetration of 40–60 μm thick films of tested sun
protection creams. Upper: reference test of foil without clay. Middle: Tone. Lower: MD2.
the corresponding results of the testing of 40–60 μm films. The diagrams give data
from the reference test as well. One finds that the penetrating UV radiation increases
somewhat with time after onset of radiation, the normalized value being taken as the
50 s reading. For 10–15 μm films the flux was 292 lux for MD2 and 433 lux for Tone,
whereas it was 490 lux for the cream-free reference test. A major result of the thin
film testing is hence that Tone reduces UV penetration by only about 12%—which
supports the critics by Swedish medical expertise—whereas MD2 reduces the pen-
etration by about 40%. A thin film of the latter is thus about three times as effective
as one prepared by a typical organic cream such as Tone.
Using 40–60 μm films the 50 s flux reading was 290 lux for Tone and 185 lux for
MD2. The thick Tone film thus reduced the flux by about 40% whereas MD2 reduced
it by about 60%. For these films, which were about four times thicker than the thin
ones, the reduction in penetrated UV radiation of the MD2 cream was hence about
1.5 times stronger than of the Tone cream. The 50 s reading of MD1 gave the value
410 lux, which represents a reduction of UV radiation by 80 lux, or 16%, which is too
little to make this cream a competitive candidate.
TABLE 9.2
Viscosity of the Creams
Cream Viscosity (mPas)
Tone 510
MD1 520
MD2 2980
TABLE 9.3
pH of the Creams
Cream pH
Tone 6.2
MD1 8.5
MD2 7.9
For 20–30 μm films, the UV flux was 260 lux for MD2 and 320 lux for Tone. The
medium-thick Tone film thus reduced the flux by about 35% whereas MD2 reduced it
by about 50%. For these films, the reduction in penetrated UV radiation of the MD2
cream was hence about 1.5 times stronger than of the Tone cream.
The recorded viscosities and pH conditions are given by Tables 9.2 and 9.3.
9.4.1.4 Conclusions
The major conclusions of the study were as follows:
• The MD1 cream based on montmorillonite had a very low UV reduction

potential, whereas MD2 representing I/S clay gave significantly better UV
reduction than the montmorillonite-rich MD1 and the organic Tone cream.
For 10–15 μm film thickness, which is a minimum value in practice, the
I/S-based cream gave a reduction in UV penetration by about 40%, which is
about three times as much as provided by the organic cream. For 20–30 μm
films, the corresponding reduction of the clay cream was about 50%, and for
40–60 μm films, it was about 60%.
• The reason for the efficiency of the I/S clay MD2 is believed to be its
content of the heavy, dark mineral chlorite (about 11%) and the high con-
tent of sheet-structured phyllosilicates such as mica and illite, which pro-
vide it with a strong reflexion ability when smeared on a plane surface.
The montmorillonite-rich MD1 clay should be nearly as effective or even
better because of the alignment of the very large number of flaky particles,
but its poor performance is believed to be caused by the conversion of the
physical state of aligned particles to the flocculated form that is caused by
the thixotropic hardening that it undergoes at rest.
302 Bentonite Clay
• The organic cream has the desired viscosity at smearing and the montmo-
rillonite-rich clay MD1 behaves similarly whereas the I/S clay MD2, being
less thixotropic, is stiffer but still easily applicable to the skin.
• The relatively low pH of the organic Tone cream may explain the commonly
noted irritation of the eyes on contacting them with this type of creams.
The montmorillonite-rich MD1 cream is significantly alkaline as all smec-
tite clays and may be even less well accepted by the skin. The I/S clay MD2
is weakly alkaline and is not expected to cause pH-related skin reactions.
9.4.2 Comprehensive Studies
In recent years, more extensive investigations have been made on the UV-protection
capacity of clays by investigators such as Young et al. (1998); Del Hoyo et al. (1998);
Del Hoyo et al. (2001); Carretero (2002); Diffey (2002); and Hoang-Minh (2006).
A modern study by Hoang-Minh et al. (2010) has clarified the role of UV-protecting
clays in several respects and we will refer to it here. They based their selection of
cream candidates on the experience that certain mineral components of sun creams
such as micronized TiO2 are known to cause unwanted photocatalytic effects, and
focused on the function of clay minerals. Kaolinite, smectite, mixed-layer-dominated
clay and mica-dominated clay were tested for protection against UV radiation in the
range 250–400 nm with special respect to the bulk Fe2O3 content.
9.4.2.1 Tested Creams

The investigated creams contained different clay materials including various types
of kaolinites, smectite-rich clays, and I/S clays. Creams of organic type were also
represented. The air-dry materials were milled to <63 μm and mixed with glycerol
to a mass ratio of 1:2. The material was then mixed with wool-wax-alcohol to a mass
ratio of 10% and 20%, and stirred to become homogeneous. The UV-investigated
thin samples had a thickness of only 3 μm.
9.4.2.2 Testing
UV measurement was made by use of an Analytik Jena AG SPECORD 50 photom-
eter, equipped with a UV lamp as light source giving wavelengths of 250–400 nm.
The light transmission was evaluated by applying the following equation:
I ×100
T= (%) (9.1)
I0
where:
I0 is the intensity of the incident UV ray
I the intensity of the UV ray after passing through the sample
The listed and plotted transmission values were those caused by the clay content
by subtracting those represented by the organic constituents. The results represent
the wavelength interval 280 to 400 nm, that is, covering the entire UV-B and UV-A
spectral ranges.
9.4.2.3 Results
The investigators measured the UV-protection ability as UV transmission and plot-
ted it against wavelength from 280 to 400 nm. Figure 9.14 shows the transmission
of four creams with 20% content of clay of different types and of a popular organic
cream (fine-dotted curve). One finds that the transmission of the very thin clay
creams was similar for the kaolinites and the I/S clay and for all bentonites except for
the relatively iron-rich Garfield smectite clay, which is in fact nontronite. The graphs
also tell that the commercial organic clay let nearly no UV radiation through for
higher wavelengths than about 250 nm. The poor performance of smectite-rich
creams hence underlines the conclusions from the pilot study and provides addi-
tional support by showing that the content of heavy minerals such as Fe2O3 reduces
UV transmission especially for mica-dominated clays (Figure 9.15). One concludes
from the pilot study and the comprehensive rigorous testing that nonthixotropic clay
creams with laminar microstructure after smearing on skin provide the best UV pro-
tection, in particular those containing heavy minerals such as iron oxides. Naturally,
an increased content of clay in creams reduces the UV penetration and so does an
increased thickness of the pasted cream.
50 50
40 40
Transmission (%)
Transmission (%)
30 30
20 Caminau (20%)
20 Wyoming (20%)
SHCa-1 (20%)
Wolfka (20%) STx-1 (20%)
10 Spergau (20%) 10 SAz-1 (20%)
Chambers (20%)
Seilitz (20%) SWy-2 (20%)
Garfield (20%)
Ladival Allerg 20 Ladival Allerg 20
0 0
280 300 320 340 360 380 400 280 300 320 340 360 380 400
(a) Wavelength (nm) (b) Wavelength (nm)
50 50
Plessa (20%)
40 40
Transmission (%)
Transmission (%)
Gorrenberg (20%)
Teistungen (20%)
Thierfeld (20%)
30 30 Ladival Allerg 20
20 20
10 10
Friedland (20%)
Ladival Allerg 20
0 0
280 300 320 340 360 380 400 280 300 320 340 360 380 400
(c) Wavelength (nm) (d) Wavelength (nm)
FIGURE 9.14 UV transmission of creams with 20% kaolin (a), bentonites (b), I/S mixed
clay (c), and mica-dominated clay (d). (Data from Hoang-Minh, T. et al., UV-protection char-
acteristics of some clays, 2010.)
304 Bentonite Clay
35 10
Y = a∗exp (X/t) + Y0 Y = a∗X + b
30 a = 1380.21; t = −4.71; Y0 = 0.42 a = −0.305; b = 9.74
2
R = 0.99 8 2
R = 0.962
25 Bentonites Mica-dominated clays
Fe2O3 (%)
Fe2O3 (%)
20 Mixed-layer series-dominated clay 6 Kaolins
Exponential growth fit of Linear fit of the clays and
15 the bentonites and the clay the kaolins
4
10
5 2
0
0
15 20 25 30 35 40 45 5 10 15 20 25 30 35
(a) UV transmission (%) (b) UV transmission (%)
FIGURE 9.15 Impact of bulk Fe2O3 on the average transmission value of investigated clay
creams: (a) Expandable (bentonites and I/S clays), (b) nonexpandable (kaolins and micaceous
clays). (Data from Hoang-Minh, T. et al., UV-protection characteristics of some clays, 2010.)
9.5 CLAY IN COSMETOLOGY

9.5.1 Background
While the role of clay in curing wounds and solving dermatological problems is
related to the biochemical interaction of clay minerals and organic tissues (Lim et al.
2005), its use in the cosmetic branch is for cleaning exposed skin and for providing
it with a coating that stimulates the most shallow parts of epidermis and provides it
with ions and molecules that affect its rheological performance. The cosmetic branch
deals with this by running numerous in vivo and in vitro experiments trying to find
organic and clay-based creams that can remove wrinkles and rejuvenate the skin.
9.5.2 Interaction of Clay Minerals and Epidermis

Interaction of smectite or kaolinite clay particle surfaces and skin takes place by hydro-
gen bonding and direct coupling via carboxyl groups or bond-sharing of adsorbed
cations and positively charged molecules. The aforementioned coupling of smectite
particle surfaces to organic molecules directly or via interfacial water films explains
why smectite pastes and creams effectively adhere to skin that they are smeared on.
Such interaction is weaker for kaolinite and even weaker for muscovite and other phyl-
losilicates. Once placed, the major time-related processes in the smeared-on smectite
clay are uptake of cations and positively charged organic molecules from the skin and
migration of fluids from or into it depending on the nature of the fluid and the relative
degree of wetting. These processes take place by diffusion in epidermis and by flow
and diffusion in the most shallow skin. A properly composed smectite clay paste sorbs
organic matter from the latter and hence has a cleaning effect. Repeated coating of skin
with smectite gel followed by washing helps to clean and freshen the skin.
The matter is complicated by the function of the liquid that must be mixed with the
clay to make the cream. A common liquid is glycerol that was used for preparing two
of the UV creams described earlier and is believed to enter the interlamellar space and
to embed the stacks of clay platelets. The clay/glycerol mass ratio naturally determines
the interaction of clay and organic tissues and the rheological performance of the cream.
This and related issues are secrets of the companies producing cosmetic products.
9.5.3 Rheology
It is important that pastes and creams for skin treatment have suitable rheological
properties and ability to establish good contact with the tissues and retain contact
with them and also to have their tightness preserved. Thixotropy is desired for pro-
viding fluidity when being pasted and to give a suitable degree of stiffening thereaf-
ter, but, as we have seen, too strong thixotropic action makes UV-protection creams
useless. Thixotropy implies low shear resistance in the coating phase and micro-
structural reorganization when leaving the paste to rest. Naturally the liquid, exem-
plified by glycerol, strongly affects the rheological behavior.
9.5.4 Clay Candidates
As for UV creams, possible candidates for skin treatment are smectitic clays espe-
cially artificially charged montmorillonite for providing the skin with necessary trace
elements such as calcium, silica, aluminum, magnesium, potassium, and sodium.
Following the principle of using only mineral components for skin treatment, the
candidature of expandable hydrophilic clay minerals is self-evident. They sorb cat-
ions and positively charged organic molecules and are impermeable to fl uids and gas
under low pressure. They can balance pH and are excellent agents for cleaning skin.
Considering first pure clay/water mixtures, exposure of water-saturated montmo-
rillonite to air with RH lower than about 80% causes desiccation, leading to intense
fissuring and loss of tightness compared to the state at complete fluid saturation.
Chemically compatible agents can, however, provide microstructural reinforcement
and stability and for montmorillonite and other smectite minerals, lath-shaped min-
erals such as palygorskite (attapulgite) or the finely tubular kaolinite species hal-
loysite can be considered as additives (Figure 9.16). Such components can minimize
shrinkage and provide coherence and tightness of the paste.
Figure 9.17 illustrates the rich fissuring on desiccation of a water saturated mont-
morillonite sample with 1140 kg/m3 dry density dried at 105°C and, in contrast,
the homogeneity of montmorillonite admixed with palygorskite after drying at the
same temperature. The number of fissures in air-dried montmorillonite clay has
been found to be 10 times higher than for equally dense clay consisting of equal
amounts of montmorillonite and palygorskite (Jones and Galan 1988; Arnold 2005).
The aperture of the cracks was similar for both cases meaning that the palygorskite-
mixed montmorillonite retained its homogeneity and tightness much better than the
sample of pure montmorillonite.
9.5.5 Density and Consistency of Smectite Creams

with and without Organic Liquid
Adding organic liquids to clay/water mixtures, exemplified using glycerol in the UV

creams MD1 and MD2, changes the physical properties. Greasy organic bases in
306 Bentonite Clay
0.50 μm
2 μm
1 μm
Transmission electron micrograph of 1:1 by Generalized framework of palygorskite

weight of mixed montmorillonite and particles. Circles are interconnected
palygorskite (needles) montmorillonite aggregates
FIGURE 9.16 Mixture of smectite and palygorskite, weight ratio 1:1. The laths are palygor-
skite with similar crystal lattice constitution as montmorillonite but with different shape. Left:
Transmission electron micrograph with framework of palygorskite particles. Right: Generalized
pattern of montmorillonite matrix stabilized by a palygorskite framework. This model is the basis
of forthcoming theoretical modeling of the rheological performance in bulk.
3.31%
FIGURE 9.17 (See color insert.) Fissuring in montmorillonite-rich clay (left) and in
a m ixture of 50% palygorskite and montmorillonite clay (right) after drying at 105°C.
The diameter of the circular specimen is 30 mm; the scale being the same for both. (Data
from Pusch, R., J. Cosmetics Dermatol. Sci. Appl., 4, 67–72, 2014.)
creams may be irritant to skin and hinder sweat excretion, which can also be irritant.
This is in contrast with pure mineral-based creams as just described but they are less
good in other respects, particularly by undergoing comprehensive microfissuring on
desiccation. Considering first organic-free clay mineral creams, the dry bulk density
for use in dermatology ranges typically between about 160 and 790 kg/m3. A typical
void ratio (ratio of pore volume and total volume) is 16.5 and water content about
590% for the lower density and 2.5 and 89%, respectively, for the higher. A most use-
ful measure of the consistency of water-saturated clay is the Atterberg liquid limit
(wL), which is the water content (mass of water divided by the total mass) that makes
the clay semi-fluid. It is about 300% for smectites saturated with sodium and about
100% for saturation with calcium or magnesium. This parameter is a measure of the
water-sorbing potential of clay.
For easy placement on skin, the density of pure smectite clay pastes saturated
with distilled water can be in the interval 1050 kg/m3 (dry density 80 kg/m3) to
1300 kg/m3 (dry density 480 kg/m3). Ongoing R&D shows that the amount of the
thixotropic stabilizer palygorskite required for providing cohesion and absence of
fissures and voids ranges between 25% and 75%, depending on the humidity of the
environment and the skin.
Continuing now with clay creams based on organic liquids one can take fat oil
with 75% saturated hydrocarbons (primarily paraffins including n-, iso-, and cyclo-
paraffins), and 25% aromatic hydrocarbons (including naphthalenes and alkylben-
zenes) as representative of such liquids. The gel volume of 2 g of such clay in water
is about seven times larger than in glycerol and about 10 times larger than in SAE
40 diesel oil (Mehras 2012) because it preferentially forms films on the basal sur-
faces of smectite stacks of lamellae. The peak shear strength of dense smectite clay
is 50 times higher for oil-soaked smectite clay than for water-saturated clay with the
same density, which means that water-based clay creams are much more fluid than
oil-based ones with the same density. The water-based creams would be most easily
applied on skin and interact with it without interference of oil but they would also
desiccate very quickly, which is the reason for using glycerol and similar fluids for
preparing creams. Effective skin treatment using creams and muds without oil—the
elephant bath type—is of course the most effective for cleaning purposes. This has
been applied by SPAs since the Middle Ages.
9.6 SUMMARY OF FUNDAMENTAL PROPERTIES

OF SMECTITIC CREAMS ON SKIN
The most important functions of smectitic creams and pastes for use as cosmetics
are in summary:
• Cleaning effect: The huge specific surface area of smectite, about 800 m2/g
solid clay, provides a very large number of hydrogen bonds between organic
debris and clay. Washing after treatment clears the skin effectively.
• Very effective barriers against microbial migration into open wounds: This
is obtained by coating them with smectite clay that sorbs secretory compo-
nents and forms a tight scab. This may, for example, retard and cure acne.
308 Bentonite Clay
The pasted clay must not become totally desiccated and fissured, which is
avoided by mixing the smectite with the fibrous clay mineral palygorskite
or the tubular clay mineral halloysite, providing reinforcement and preserv-
ing microstructural cohesion.
• pH effects: A typical reaction when Na montmorillonite is dispersed in dis-
tilled water is that Na is given off to the porewater and protons replacing the
Nas in exchange positions, which raises pH of the water to as much as 10.
If the clay is dispersed in Na-rich water, Na is sorbed and protons released
causing pH to drop even to 3. This is the case when there is sweat on the
skin, which hence means that skin treatment with smectite gels or pastes
can bring pH down. Aged skin has pH 5–6, which implies low epidermal
barrier integrity, whereas treatment of aged skin with substances reducing
pH to no less than 4 is valuable (Blaak et al. 2011).
REFERENCES
1954. Annals of the New York Academy of Sciences, 57:678.
1961. Medical Annals of the District of Columbia, 20:6.
1996. Industrial Toxicology. TLV Documentation FPMD 667, Spring Quarter.
Anderson, D.M. 1967. The interface between ice and silicate surfaces. Journal of Colloid
Science, 25:174–191.
Arnold, D.E. 2005. Maya Blue and Palygorskite: A second possible pre-Columbian source.
Ancient Mesoamerica, 16:51–62.
Bendz, G., Lindqvist, I. (Eds.) 1978. Biochemistry of Silicon Related Problems. Plenum Press,
New York.
Blaak, J., Wohlfart, R., Schuerer, N.Y. 2011. Treatment of aged skin with a pH 4 skin care
product normalizes increased skin surface pH and improves barrier function: Results
of a pilot study. Journal of Cosmectics, Dermatological Sciences and Applications,
1:50–58.
Cancer & Natural Therapy Foundation of Australia. 2014. Calcium Bentonite Living Clay,
Austin, TX.
Carretero, I. 2002. Clay minerals and their beneficial effects upon human health—A review.
Applied Clay Science, 21(3/4):155–163.
Chipera, S.J., Bish, D.L. 2001. Baseline studies of the clay minerals society source clays:
Powder X-ray diffraction analyses. Clays and Clay Minerals, 49(5):398–409.
Del Hoyo, C., Vicente, M.A., Rives, V. 1998. Application of phenyl salycilate-sepiolite sys-
tems as ultraviolet radiation filters. Clay Mineralogy, 33:467–474.
Del Hoyo, C., Vicente, M.A., Rives, V. 2001. Preparation of drug-montmorillonite UV-radiation
protection compounds by gas-solid adsorption. Clay and Clay Minerals, 36:541–546.
Diffey, B.L. 2002. Sources and measurement of ultraviolet radiation. Methods, 28:4–13.
Grim, R.E. 1953. Clay Mineralogy. McGraw-Hill, New York.
Grim, R.E., Guven, N. 1978. Bentonites: Geology, Mineralogy, Properties and Uses.
Developments in Sedimentology, Vol. 24, Elsevier, New York.
Harley, J.H., Schmidt, G.D., Silini, G. (Eds.) 1988. Radionuclides in the Food Chain. Springer-
Verlag, London.
Hoang-Minh, T. 2006. Characterization of clays and clay minerals for industrial applications:
substitution non-natural additives by clays in UV protection. Dissertation, Ernst-Moritz-
Arndt-University Greifswald, Greifswald, Germany.
Hoang-Minh, T., Le, T.L., Kasbohm, J., Gieré, R. 2010. UV-protection characteristics of some
clays. Applied Clay Science, 48:349–357.
Jones, B.F., Galan, E. 1988. Sepiolite and palygorskite. Reviews in Minerals, 19:631–674.
Kasbohm, J., Herbert, H.-J., Henning, K.-H. 2005. Short and long term stability of selected
bentonites in high saline solutions. International Symposium on Large Scale Field Tests
in Granite, Sitges, Barcelona, Spain, November 12–14, 2003. Code 88404:231–240.
Kranz, G., Ruchholz, M., Marx, E. 1990. Neue Referenzmaterialien—Kaoline und Tone.
Silikattechnik, 41(10):330–333.
Lamanna, C., Mallette, M.F. 1959. Basic Bacteriology. Lippincott Williams & Wilkins,
Baltimore, MD.
Lambert, J.-F., Poncelet, G. 1997. Acidity in pillared clays: origin and catalytic manifestations.
Topics in Catalysis, 4:43–56.
Lim, H.W., Gilchrest, B.A., Cooper, K.D. et al. 2005. Sunlight, tanning booths, and vitamin D.
Journal of the American Academy of Dermatology, 52:868–876.
Longstreth, J.D., de Gruijl, F.R., Kripke, M.L., Takizawa, Y., van der Leun, J.C. 1994. Effects
of increased solar ultraviolet radiation on human health. In Environmental Effects of
Ozone Depletion, van der Leun, J.C. (Ed.), United Nations Environment Programme,
Nairobi, Kenya, pp. 23–48.
Mehras, S., 2012. Investigation of method for preventing oil spilled on the ground surface to
migrate downwards and contaminate groundwater. Theoretical modelling, experimen-
tal study and practical technique. MSc thesis, Lulea University of Technology, Dept.
Civ., Environm., Nat. Res. Engineering.
Muurinen, A., Lehikoinen, J., Pusch, R. 1994. POSIVA publications Report YJT-94-22,
POSIVA, Helsinki, Finland.
Pusch, R. 1970. Clay Microstructure. Nat. Swed. Build. Res. Council. Document D8,
Stockholm, Sweden.
Pusch, R. 1973. Influence of Organic Matter on the Geotechnical Properties of Clays. Nat.
Swed. Build. Res. Council. Document D11, Stockholm, Sweden.
Pusch, R. 2014. Flow and ductility of smectite clay for skin treatment. Journal of Cosmetics,
Dermatological Sciences and Applications, 4:67–72.
Pusch, R., Muurinen, A., Lehikoinen, J., Bors, J., Eriksen, T. 1999. Microstructural and
Report European Commission, Contract No F14W-C95-0012, Stockholm, Sweden.
Pusch, R., Yong, R.N. 2006 Microstructure of Smectite Clays and Engineering Performance.
Taylor & Francis, London.
Setlow, R.B. 1974. The wavelengths in sunlight effective in producing cancer: A theoreti-
cal analysis. Proceedings of the National Academy of Sciences of the United States of
America, 71:3363–3366.
van Olphen, H., Fripiat, J.J. 1979. Data Handbook for Clay Minerals and Other Non-Metallic
Materials. Pergamon Press, Oxford.
Vicente, M.A., Sanchez-Camazano, M., Sánchez-Martin, M.J., Del Arco, M., Martin, C., Rives, V.,
Vicente-Hernández, J. 1989. Adsorption and desorption of N-methyl 8-hydroxy quino-
line methyl sulfate on smectite and the potential use of the clay-organic product as an
ultraviolet radiation collector. Clay and Clay Minerals, 37(2):157–163.
Vogt, C., Lauterjung, J., Fischer, R.X. 2002. Investigation of the clay fraction (<2 µm) of the
Clay Minerals Society reference clays. Clays and Clay Minerals, 50(3):388–400.
WHO. 2002. Global Solar UV Index: A Practical Guide 2002. World Health Organization,
Geneva, Switzerland.
Williams, L.-B., Haydel, S.E. 2010. Evaluation of the medicinal use of clays minerals as anti-
bacterial agents. International Geology Review, 52(7/8):745–770.
Young, A.R., Chadwick, C.A., Harrison, G.I., Nikaido, O., Ramsden, J., Potten, C.S. 1998.
The similarity of action spectra for thymine dimmers in human epidermis and ery-
thema suggests that DNA is the chromophore for erythema. Journal of Investigative
Dermatology 111:982–988.
Appendix
Symbols and Definitions*
A.1 SYMBOLS
ac activity
az swelling index
A area, creep parameter
B creep parameter
c, c′ cohesion, cohesion intercept
d diameter
d10 , d60 grain diameter
Da diffusion coefficient (apparent)
De diffusion coefficient (effective)
Dp pore diffusion coefficient
Ds surface diffusion coefficient
e void ratio
E modulus of elasticity
F force
g gravity
go organic content
g1, gc ignition loss
G shear modulus
h height
i hydraulic gradient
IC consistency index
I L liquidity index
IP plasticity index
k (hydraulic) permeability
K hydraulic conductivity, compression modulus
Kd sorption factor
l length
lc content of clay particles
m mass, modulus number
mj modulus number (reference)
ms mass of solid matter
mw mass of water
M oedometer modulus
MI,MII,M III metal cations
n porosity
* Only major ones are listed.
311
312 Appendix
p normal stress
pS swelling pressure
P vapor pressure
q deviator stress
R molar gas constant, roundness
R S shrinkage ratio
s salt content of pore fluid
S sphericity
Sr degree of water saturation, volume percentage
St sensitivity
t time
T temperature
u porewater pressure
v flow rate
V bulk volume
Vd volume of dry soil sample
Vg volume of gas
Vp pore volume
Vs volume of solid matter
VS volumetric shrinkage
Vw volume of water
w water content (ratio) by weight
wL liquid limit
w P plastic limit
wS shrinkage limit
α heat expansion coefficient
αs swelling index
β stress exponent, creep parameter
ε compressive strain
γ shear strain
η viscosity
λ heat conductivity
ν Poisson’s ratio
ρ bulk density
ρd dry density
ρs density of solid particles
ρsat density of fluid-saturated material
ρw density of water
σ total pressure
σ′ effective pressure (stress)
τ shear stress
τfu undrained shear strength
τr remoulded shear strength
φ electrical potential
φ, ϕ′ friction angle
Δl shortening of sample with length (height) l
Appendix 313
A.2 DEFINITIONS
A.2.1 Overview
The basic definitions used here concern mass relationships, consistency, strength
parameters, and transport properties (Pusch 1990). They are identical with those
proposed for the international ISO/DIS standards Geotechnics in Civil Engineering
and Classification of Soil.
A.2.2 Mass Relationships
A.2.2.1 Mass, m (kg)
Mass is the term for the content of material of a body, and it is usually determined by
weighing. The word weight is therefore a common synonym of mass, but it is avoided
here if the meaning becomes unclear.
A.2.2.2 Bulk Volume, V (m3)

Bulk volume refers to the soil volume including isolated as well as continuous voids.
A.2.2.3 Volume of Solid Matter, Vs (m3)

The volume of the solid mass or specific volume refers to the difference between
total volume and the void volume.
A.2.2.4 Bulk Density, ρ (kg/m3)

The bulk density is the ratio of the mass to the bulk volume of a given amount of soil:
ms + mw
ρ= (A.1)
V
A.2.2.5 Dry Density, ρd (kg/m3)
The dry density is the ratio of the solid mass to the bulk volume of a given amount
of soil:
ms
ρd =
V (A.2)

A.2.2.6 Density of Solid Particles, ρs (kg/m3)
The density of solid particles is defined as the ratio of the mass to the true volume of
the solid matter in a given amount of soil:
ms
ρs = (A.3)
Vs

A.2.2.7 Density of Pore Fluid, ρw (kg/m3)
The density of the fluid in the voids is the ratio of the mass of fluid with its content
of dissolved matter to its volume. The pore fluid is conventionally termed porewater,
which we will use throughout the document.
314 Appendix
Vp
Void ratio, e = (A.4)
Vs
The void ratio is the ratio between pore volume and the volume of solid matter.
Vp
Porosity, n = (A.5)
V
Porosity is the ratio between the pore volume and the bulk volume of the soil. It is
usually expressed in percent units.
A.2.2.8 Water Content (Water Ratio) w (%)

The water content gives the ratio between the mass of the porewater and the mass of
the solid substance. It is expressed in percent units:
mw
w= ⋅ 100 (A.6)
ms

A.2.2.9 Degree of Water Saturation, Sr(%)
The degree of water saturation is the ratio between the volume of the porewater and
the pore volume. It is expressed in percent units:
Vw
Sr = ⋅100 (A.7)
Vp

A.2.2.10 Clay Content, lc (%)
The clay content is the ratio between the mass of particles finer than 2 μm and the
mass of all soil particles finer than 0.06 mm. It is expressed in percent units (Pusch
1990; Skempton 1953).
A.2.2.11 Ignition Loss, g and gc (%)

The ignition loss is the ratio between the loss of solid mass by heating dry soil and
the mass of the dry soil before the heating. gc denotes the ignition loss of the mineral
constituents (primarily clay minerals), whereas g means the total ignition loss. Both
are usually expressed in percent units.
A.2.2.12 Organic Content, go (%)

Organic content is the ratio between the mass of the organic substance and the mass
of the dry soil. For carbonate-free soil, it is the difference between the total ignition
loss and a term that contains the clay content.
A.2.2.13 Salt Content of Pore fluid, s (% or ppm)

The salt content is the ratio between the mass of dissolved matter (salt) and the mass
of water. It is often expressed in percent, but for individual elements, such as Na, K,
and Cl, the concentration is usually given in parts per million (ppm).
Appendix 315
A.2.2.14 Gas Content, Vg (%)

The content of gas in gaseous form is defined in terms of the fraction in percent of
the total volume that the gas occupies at normal temperature and pressure.
A.2.2.15 pH (acidity)
pH refers to the negative logarithm for the concentration of free hydrogen ions in the
porewater (cf. Lambe and Whitman 1969; Mitchell 1973).
A.2.3 Consistency Parameters
The term consistency refers in this context primarily to the degree of stiffness and
plasticity of a soil in a remoulded state. Physicochemically, the consistency of the soil
depends on the internal bonds between the soil particles, the cohesion. A characteristic
of the cohesive soils is that, within certain water content limits, they have a plastic con-
sistency in the remoulded state. At lower water content, they have a solid consistency
and at higher water content a liquid consistency. As an intermediate state between the
solid and the plastic consistency, a semi-solid consistency is distinguished:
A coherent, ductile soil sample having a solid consistency can be elastic, vis-
coelastic, or plastic; a nonductile soil sample is brittle with abrupt rupture
occurring at a small deformation. It can be elastic, viscoelastic, or plastic.
A soil sample having a plastic consistency can be remolded and maintains its
shape after deformation.
A soil sample having a liquid consistency is viscous and flows under its own weight.
A.2.3.1 Plastic Limit, wP (%)

The plastic limit is the water content at which the soil begins to crumble when rolled
out to a 3 mm thread.
A.2.3.2 Liquid Limit, wL (%)

The liquid limit wL can be determined according to the percussion method (percus-
sion liquid limit) or the fall-cone method (fall-cone liquid limit).
The percussion liquid limit, determined by use of Casagrande’s liquid limit
device, is defined as the water content at which a remoulded soil sample, placed
in the cupo of the device and divided into halves by a V-shaped groove, has such a
consistency that the two sample halves flow together 13 mm (1/2″) along the bottom
of the groove when the cup is dropped 10 mm 25 times.
The fall-cone liquid limit (earlier called the fineness number), determined by the
fall-cone method, is defined as the water content at which a remoulded soil sample
has such a consistency that the penetration of a 60 g/60° fall-cone is 10 mm (shear
strength 1.7 kPa).* The fall-cone liquid limit generally agrees with the percussion
liquid limit.
* According to British Standard, the (cone) liquid limit is defined as the water content at 20 mm penetra-
tion with an 80 g/30° cone, which corresponds to 10 mm penetration with the Swedish 60 g/60° cone.
316 Appendix
The sticky limit was defined by Atterberg as the lowest water content at which
the soil sticks to the blade of a metal spatula when the spatula is drawn against the
surface of the remoulded soil sample (Pusch 1990).
For high-plastic clays, the sticky limit is lower than the liquid limit. For low-plas-
tic clays, as well as for organic soils, the sticky limit is higher than the liquid limit.
A.2.3.3 Shrinkage Limit, ws (%)

The shrinkage limit ws is defined as the water content at which a saturated soil sam-
ple ceases to shrink when dried. At further desiccation, air starts to penetrate the
voids. At the same time, the color of the sample becomes lighter.
The shrinkage limit can be determined for remoulded samples, as well as for
undisturbed samples.
A.2.3.4 Plasticity Index, Ip [%]

The plasticity index Ip is defined as the difference between the liquid limit and the
plastic limit:
I P = wL − wP (A.8)
The plasticity index can be regarded as a measure of the plasticity of the soil. For
nonplastic soils IP = 0.
With regard to plasticity, clays are classified either on the basis of the liquid limit
wL or the plasticity index IP, in four groups (Table A.1).
For high-plastic soils, the undrained shear strength τfu, determined by use of the
fall cone is reduced with regard to the value of the liquid limit or the plasticity index.
A.2.3.5 Liquidity Index, IL

The liquidity index IL is defined as follows:
w − wP
IL = (A.9)
IP
where:
w is the natural water content of the soil
TABLE A.1
Classification of Clays with respect to
Their Consistency
Plasticity Liquid Limit,
Designation Index, IP (%) wL (%)
Low plasticity <10 <30
Medium plasticity 10–25 30–50
High plasticity >25 >50–80
Very high plasticity >50 >80
Appendix 317
TABLE A.2
Liquidity Index in Relation to Consistency Limits
Water Content in Relation Consistency of
to wP and wL Liquidity Index, IL Remoulded Sample
w < wP IL < 0 Solid and semi-solid
w = wP IL = 0
wP < w < wL 0 < IL < 1 Plastic
w = wL IL ≡ 1
w > wL IL > 1 Semi-liquid–liquid
The liquidity index is a measure of the consistency of a soil in the remoulded state at
the natural water content (Table A.2) and is also used as an indication of the sensitiv-
ity to mechanical disturbance of the soil.
A.2.3.6 Consistency Index, IC

The consistency index is defined as follows:
wL − w
IC = (A.10)
IP
The consistency index is, like the liquidity index, a measure of the consistency of the
soil in the remoulded state at the natural water content (Table A.3). The following
relation exists between the consistency index and the liquidity index (Pusch 1990):
I C = 1 − I L (A.11)
A.2.3.7 Activity, ac
The activity ac is defined as the ratio between the plasticity index and the clay con-
tent of the soil, in percentage of material <0.4 mm:
IP
ac = (A.12)
% clay fraction
TABLE A.3
Consistency Index in Relation to the Consistency of Remoulded Soil
Water Content in Relation to wP Consistency of Remoulded
and wL Consistency Index, IL Sample
w < wP IC > 1 Solid–semi-solid
w = wP IC = 1
wP < w < wL 1 < IC < 0 Plastic
w = wL IC = 0
w > wL IC < 0 Semi-liquid–liquid
318 Appendix
The activity is an indication of the colloidal properties of a clay, and is principally

dependent on the content and the kind of clay minerals and organic colloids, as well
as on the electrolyte content of the pore water (Skempton 1953).
A.2.3.8 Shrinkage Ratio, RS

The shrinkage ratio RS is defined as the ratio between a given volume change,
expressed as a percentage of the dry volume, and the corresponding change in water
content above the shrinkage limit of a soil mass.
The shrinkage ratio is calculated from data obtained in the shrinkage-limit deter-
mination by means of the following formula:
ms
RS =
VD ⋅ρw (A.13)

where:
VD is the volume of the dry sample
Thus, the shrinkage ratio is numerically equal to the bulk density of the dry sample.
Knowing the shrinkage ratio RS, shrinkage limit wS, and original water content w,
the shrinkage of a soil mass when the water content is reduced to, or below, the
shrinkage limit can be calculated.
A.2.3.9 Volumetric Shrinkage, VS (%)

The volumetric shrinkage VS, expressed in percent of the volume of the soil mass at
the shrinkage limit, is obtained by means of the following formula:
VS = ( w − wS ) ⋅ RS ⋅100 (A.14)
A.2.3.10 Linear Shrinkage, LS (%)

The linear shrinkage L S in percent of the original dimension of the soil mass is
evaluated as
⎡ 100 ⎤
LS = ⎢1 − 3 ⎥ 100 (A.15)
⎢⎣ ( w − ws ) RS + 100 ⎥⎦

A.2.4 Strength Parameters

A.2.4.1 Undrained Shear Strength of Undisturbed
Cohesive Soils, τfu (Pa, kPa, MPa)
The undrained shear strength of undisturbed cohesive soils is defined as the maxi-
mum shear stress that a soil sample can sustain under defined conditions with respect
to loading rate and external (total) stresses under undrained conditions (Janbu 1979;
Lambe and Whitman 1969).
Appendix 319
A.2.4.2 Undrained Shear Strength of Remoulded

Cohesive Soils, τr (Pa, kPa)
The undrained shear strength is defined as the maximum shear stress that a soil
sample can sustain in undisturbed or remoulded form.
A.2.4.3 Sensitivity, St
The sensitivity St to mechanical disturbance is defined as the ratio of the shear
strength (τfu) of the undisturbed sample and the shear strength (τr) after remoulding:
τfu
St = (A.16)
τr

With regard to sensitivity, clays are classified in three groups: low-, medium-, and
high-sensitive clays (Table A.4). Quick clay, commonly having St > 50 is distin-
guished as a special group of high-sensitive clays.
A.2.4.4 Drained Shear Strength of Soils, τfu (Pa, kPa, MPa)

The shear strength of soils under drained conditions is conventionally expressed in
terms of effective cohesion c′ and internal friction (angle) ϕ′ applying the effective
stress concept. The parameters are evaluated from a set of recorded maximum shear
stresses that the soil can sustain at different normal stresses under defined conditions
with respect to the loading rate. The general relationship is, in its simplest form,
expressed by Mohr/Coulomb’s failure criterion.
A.2.5 Rheological Parameters
A.2.5.1 Soil Compressibility, Parameters M (kPa), m and β
The compressibility of soils is often expressed by the tangent modulus M of the com-
pression curve obtained from oedometer tests (Larsson 1981):
1−β
⎛ σʹ ⎞
M = mjσʹj ⎜ ⎟ (A.17)
⎝ σʹj ⎠
TABLE A.4
Classification of Clays with respect to
Sensitivity
Sensitivity St (According
Designation to Fall-Cone Tests)
Low-sensitive <10
Medium-sensitive 10–30
High-sensitive >30
320 Appendix
where
mj is the modulus number
β is the stress exponent
σ′ is the effective compression stress, kPa
σj is the reference stress, kPa
A.2.5.2 Compression Index, Cc
Δe
Cc = / kPa (A.18)
Δ ( log σʹ )
where
Δe is the change in void ratio
Δσ′ is the change in effective compression stress (log scale)
A.2.5.3 Swelling Index, Parameters as, Ms, and S
σʹ
Ms = (A.19)
as

S
as = (A.20)
ln ( 0.8σʹc σʹred )

where:
as is the swelling index
S is the rebound on unloading
Ms is the swelling modulus
σʹc is the preconsolidation pressure
σʹred is the effective pressure after load reduction
A.2.5.4 Shear Strain, γ

Shear strain is expressed as the ratio between the shear stress and the angular strain γ at
simple shearing. The shear modulus G is obtained from shear tests in which the shear
strain is plotted as a function of the shear stress. It is expressed as the tangent modulus:
dτ
G= (A.21)
dγ

or secant modulus
Δτ
G= (A.22)
Δγ

A.2.5.5 Compressive Strain, E, G, ν
Uniaxial compression of an elastic body of length l gives the strain ε = Δl l where Δl
is the shortening. The parameters are the modulus of elasticity E and Poisson’s ratio ν:
Appendix 321
E = 2G(1 + n) = 3K (1 − 2n) (A.23)
(3K − 2G )
ν= (A.24)
2(G + 3K )

E
G= (A.25)
2(1 + ν)

1
ε= ⎡σʹx − ν ( σʹy + σʹz ) ⎤⎦ (A.26)
E⎣
A.2.6 Transport Properties
A.2.6.1 Hydraulic Conductivity, K (m/s)
The hydraulic conductivity is the average flow rate at percolation of a confined soil
sample. It is evaluated by use of Darcy’s law:
Q
ν= = K ⋅ i (A.27)
A
where:
ν is the flow rate (m/s)
Q is the flux (m3/s)
A is the total cross section (m2)
I is the hydraulic gradient, difference in meter water hear divided by meter flow
length (m/m)
A.2.6.2 Diffusion Coefficient, Da and De (m2/s)

Diffusive ion transport is described by use of the diffusion coefficient and usually
evaluated from experiments by use of Fick’s law (Muurinen 1994). For pore diffu-
sion, one has
∂C ∂ 2C (A.28)
= Da
∂t ∂ x2
De = Da ( ε + K dρ ) (A.29)

where:
C is the concentration of the diffusing substance (kg/m3)
Da is the apparent diffusion coefficient (m2/s)
De is the effective diffusion coefficient (m2/s)
ε is the porosity available for diffusion
Kd is the sorption factor (m3/kg)
ρ is the bulk density (kg/m3)
322 Appendix
For the case of both pore and surface diffusion, the effective diffusivity is given as
follows:
De = Da ( ε + K dρ ) + K dρDs (A.30)

where:
Ds is the surface diffusion coefficient
REFERENCES
Janbu, N. 1970. Grunnlag i Geoteknikk. Tapir Forlag, Trondheim, Norway.
Lambe, T., Whitman, R.V. 1969. Soil Mechanics. Wiley, New York.
Larsson, R. 1981. Drained Behaviour of Swedish Clays. Report No. 12, Swedish Geotechnical
Institute, Linköping, Sweden.
Mitchell, J.K. 1973. Fundamentals of Soil Behavior, Wiley, New York.
Muurinen, A. 1994. Diffusion of Anions and Cations in Compacted Sodium Bentonite. VTT
Publication No. 168, Technical Research Centre of Finland, Espoo, Finland.
Pusch, R. 1990. Soil Constituents and Structure. Swedish Council for Building Research,
Stockholm, Sweden.
Skempton, A.W. 1953. The colloidal “activity” of clays. Proceedings of the 3rd International
Conference on Soil Mechanics and Foundation Engineering, Zürich, Switzerland,
August.
FIGURE 2.1 Open pit mining of water-saturated Friedland clay: the pit is about 100 m deep
and dry.
AD
Sample C
9.97 Å WOI
Quartz
14.3 Å
Chlorite
EG
IS-ml
kaolinite and chlorite
EG
14.2 Å chlorite
7.15 Å
Feldspar
4.41 5.41 6.41 7.41 8.41

IS-ml (∼85%)
IS-ml (∼40%)
IS-ml (∼40%)
IS-ml (∼85%)
16.7 Å IS-ml (∼40%)
4 8 12 16 20 24 28 32
°2θ CoKα
FIGURE 4.21 Lithuanian smectite clay intended for lining construction. Legend: AD, air
dried; EG, ethylene-glycol saturated; and IS-ml (~85%), illite-smectite mixed-layer with 85%
illitic layers. (After Kasbohm.)
FIGURE 5.14 Example of digitalized micrograph of 300Å ultramicrotomed section of
Wyoming bentonite (MX-80) with a bulk density at saturation of 1800 kg/m3 (dry density
1270 kg/m3). Black represents densest parts of clay matrix and accessory minerals; red rep-
resents relatively dense parts of the same component; green represents soft, porous clay gel;
white represents open voids. Height of picture is 3 μm.
Temperature (K)
376.37
366.44
356.51
346.58
336.65
326.72
316.79
306.86
296.93
287
FIGURE 5.25 Predicted temperature distribution in the near-field of KBS-3 canisters

shortly after the disposal. (Data from Svemar, C., Cluster Repository Project [CROP], Final
Report of European Commission Contract FIR1-CT-2000-20023, Brussels, Belgium, 2005.)
1
2
5 6
3 4
FIGURE 5.28 Schematic rock structure models for BMT holes 1 to 6. Blue fractures are
strongly water-bearing, red fractures are moderately water bearing, green fractures are
poorly water bearing. The arrows indicate north direction.
1.00
0.95
0.90
0.85
0.80
0.75
0.70
0.65
t = 10 years
0.60 t = 25 years
t = 75 years
0.55
0.50
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1.0
t = 10 years
t = 25 years
0.9 t = 75 years
0.8
0.7
0.6
0.5
0.4
0.3
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
FIGURE 5.32 Predicted change in saturation from initially 50% of KBS-3V buffer. Upper:
successive increase in saturation of the buffer in rock with E−10 m/s hydraulic conductivity.
Lower: initial reduction of the buffer saturation followed by delayed saturation of the buffer
clay in rock with E−12 m/s hydraulic conductivity. (Data from Pusch, R., Weston, R., Appl.
Clay Sci., 23, 35–41, 2003.)
4 μm
FIGURE 5.35 Example of precipitated NaCl at the cold end of the test cell in the experi-
ment referred to. Left: scanning electron micrograph of NaCl crystals are seen embedded in
smectite matrix. (Data from Pusch, R., Geological Storage of Radioactive Waste, Springer-
Verlag, Berlin, Germany, 2008.) Right: salt crust at the hot end of the test cell.
1
NODAL SOLUTION
STEP = 1 MAY 11 2006
SUB = 1 19:28:00
TIME = 1
EPTOX (AVG)
RSYS = 0
DMX = .073988
SMN = -.003846
SMX =.010391
MX
MN
Y
Z X
-.003846 -.682E-03 .002481 .005645 .008809

rolepsxl -.002264 .899E-03 .004063 .007227 .010391
FIGURE 5.38 Horizontal strain in the soil components. Horizontal compression is up to

about 0.3% (blue) and tension up to 0.8% (red).
1
NODAL SOLUTION
STEP = 1 OCT 5 2005
SUB = 1 18:40:05
TIME = 1
SY (AVG)
RSYS = 0
DMX = .023363
SMN = –169431
SMX = 603806
MN
–169431 2399 174230 346060 517891

–83516 88314 260145 431975 603806
Sigmay
FIGURE 5.40 Close-up of vertical stresses in Pa in the sand and clay near the upper edge of
the vaults. Compressive stresses are marked dark (max value 170 kPa) and light blue (minus
sign) and tensile stresses are green, yellow, and red (max value 518 kPa).
Perforated Canister with Permanent disposal

Deposition casing waste cell liner
hole
Engineered
buffer
Permanent
Dense, smectitic sleeve
clay (buffer)
FIGURE 5.60 Main components of installed supercontainer for spent reactor fuel. Left:
canister surrounded with buffer clay and fitted in perforated container. Right: French design
with sleeve of carbon steel for hosting canisters. (Courtesy of ANDRA, France.)
S4-11 MX-80 1300 kgm−3
2.9% IV Freeze-dried + bm-methacr. + catal.
38.7% III 40.2% I Picture width 7 μm
Empty void diameter <3 μm
P/Tw = 2.9%, P/Tw + g = 41.6%
18.2% II
FIGURE 6.3 Transmission electron micrograph of 300Å–500Å ultramicrotomed sec-

tion of MX-80 clay with 1300 kg/m3density (dry density 480 kg/m3). The digitally treated
version of the left figure gave the picture to the right showing the density variations: white
and green represent open space, red is very dilute clay gel, and black is dense aggregates of
smectite stacks. P/Tw is the ratio of the sectioned area of wholly open voids (white) and the
total sectioned area. P/Tw+g is the ratio of the area of voids with very dilute clay gels and
the total sectioned area. (Data from Pusch, R., Eng. Geol., 54, 187–195, 1999.)
diVerm
Charge distribution
RMN diSV-ML
bent.
1.0 Fe-diSV-ML
5.0 0 Fe-Montm.
0.8 1 ML10
5.2
2 ML20
0.6 3 ML30
5.4
4 ML40
r
aye
0 0.4
5.6 6 ML60
al l
0
Ch
4 33
r
hed
6 0 Montm.
arg
0.2
5.8
a
eo
oct
f in
0 0
0.0
of
ter
6.0
ge
ayel
ar
Ch
6.2
6.4
6.6
6.8
14.0 14.2 14.4 14.6 14.8 15.0 15.2 15.4 15.6 15.8 16.0
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
IV
Charge of tetrahedral layerSi
FIGURE 6.11 Distribution of charge in tetrahedral, octahedral, and interlayers (interlamellar

space) in untreated RMN clay. ML represents mixed-layer minerals; the figures indicate
percentages. (Data from Pusch, R. et al., Appl. Clay Sci., 47, 113–119, 2010; Analyses by
Kasbohm.)
Charge distribution diVerm
MOCK M4 diSV-ML
1.0 Fe-diSV-ML
5.0 0 Fe-Montm.
0.8 1 ML10
5.2
2 ML20
0.6
5.4 0 3 ML30
r
aye
0 4 ML40
al l 0.4
Ch
5.6 5 ML50
r
hed
a
5
rge
77 0 0.2 7 ML70
cta
8 00
5.8
of
fo
8 ML80
int
eo
0.0
er
0 Montm.
arg
6.0
lay
er
Ch
6.2
6.4
6.6
6.8
14.0 14.2 14.4 14.6 14.8 15.0 15.2 15.4 15.6 15.8 16.0
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
FIGURE 6.12 Distribution of charge in tetrahedral, octahedral, and interlayers (interlamellar

space) of the M4 sample. (Data from Pusch, R. et al., Appl. Clay Sci., 47, 113–119, 2010;
Analyses by Kasbohm.)

3 3
MX-80 2000 kg/m 1850 kg/m 1800 kg/m3
3 3
FIM 2240 kg/m 2030 kg/m 1970 kg/m3
E−14
E−13
MX-80 Saponite FIM
E−12
K (m/s)
E−11
E−10
E−9
E−8
0 5 10 15 20 25 mm
T = 95°C T = 85°C T = 75°C T = 65°C T = 55°C T = 45°C
FIGURE 6.15 Diagram of the hydraulic conductivity (km/s) of the three clays that had been
compacted in air-dry form under 1 MPa pressure giving nearly the same initial density of
MX-80 and saponite but higher initial density of the smectite-poor FIM clay.
MX-80 2000 kg/m3 1850 kg/m3 1800 kg/m3
FIM 2240 kg/m3 2030 kg/m3 1970 kg/m3
2000
1750
Saponite
1500
MX-80
ps (kPa)
1250
500
250
FIM
0
0 5 10 15 20 25 mm
T = 95°C T = 85°C T = 75°C T = 65°C T = 55°C T = 45°C
FIGURE 6.16 Diagram of the swelling pressure ps of the three clays.
2400
.0
10
2300
ite
2200
Ill
2100
2000
1900
1800
1700
1600 6
7.
1500 m
u
1400 ps
Lin (counts)
1300 y
3
G
1.
.1
1200
ca
1100
15
I/S
1000
7.
ite
4
900
4.
lin
05
.1
800
o
7.
ca
Ka
ite
700
I/S
r
lo
600
Ch
500
400
300
200
100
0
5 6 7 8 9 10 11 12 13 14 15
2-theta-scale
FIGURE 6.40 X-ray diffraction analysis of the clay samples showing the impact of heat
and contacting cement: red—21°C, blue—100°C, and green—150°C. The d-values given are
in Å units. The background trace has been subtracted. Lin, Linear counts. (Original diagram
provided by L. Warr, Greifswald University, Germany.)
FIGURE 7.4 A dry water particle in air-dry clay matrix (left) gives off its water content
(light blue) to the smectite particles (upper right), which expand and ultimately yield a more
or less homogeneous clay matrix with very finely fragmented silicious shells. (Data from
Forsberg, T. et al., Eng. Geol., 2014.)
FIGURE 7.5 Blocks with 1.85 m diameter and about 0.4 m height prepared by uniaxial
compression. Left: solid block with black surface coating of carbon disulfide for lubrication.
Right: hollow block with space for a HLW canister. Weight about 2000 kg, water content
about 10%. (Photos by Ramqvist.)
Blue color: compressive stress
Yellow color: moderate shear stresses
Red color: very high shear stresses
FIGURE 7.8 FEM-calculated stress distribution of large block prepared by uniaxial com-
pression. The left side represents the vertical symmetry axis of the block. The overstressing
of the upper edge causes fissuring of this part that can fracture and fall off at transporting
and installation of the block, or when the heat pulse comes from the hot canister. (After
Börgesson.)
FIGURE 7.11 Unit cell with one-eighth of a big grain (0.35 mm) contacting one small grain
(0.10 mm). The edge length of the cubical unit cell is 0.175 mm. (Data from Pusch, R., Yong,
R.N., Microstructure of Smectite Clays and Engineering Performance, Taylor & Francis,
London, 2006.)
0.1 mm
FIGURE 7.12 Schematic cross section of the microstructure of fully matured smectite-rich
clay with a bulk density at water saturation of 1800–2000 kg/m3. For the lower density the bar
is 0.1 mm and for the higher 0.01 mm.
FIGURE 7.18 Stockpiled Lithuanian clay for drying under open conditions. At rainfall,
only the uppermost few centimeters become wetted and hinder further downward penetra-
tion. Removing this very shallow layer gives access to unaffected material.
FIGURE 7.19 Holmehus clay. Upper: view of open pit mining in Denmark. The precious clay has an overburden of nonsmectite clay of about 10 m.
Lower: stockpiled Holmehus clay stored outdoor for drying. (Courtesy of Dantonite A/S.)
Ground surface
500 m
Fractures filled with
grout Concrete
Clay plugs
Clay blocks and

cast concrete
1500 m
Water-bearing
fracture zone
Concrete plug in
fracture zone
Clay block
Supercontainer Supercontainers
with HLW
canisters 2000 m
Casing for supporting separated by clay
borehole
blocks
Borehole 0.8 m
diameter
Concrete plug
FIGURE 8.1 Schematic view of proposed VDH design: seals of clay and concrete down to
2000 m depth and supercontainers with HLW canisters separated by clay and concrete in the
interval 2000–4000 m depth. (Data from Beswick, J., Status of technology for deep borehole
disposal. NDA deep disposal report, Rev.7. Contract No NP 01185. Oxon, 2008; Grundfelt, B.,
Jämförelse mellan KBS-3-metoden och deponering i djupa borrhål för slutlig tomhänderta-
gande av använt kärnbränsle, Swedish Nuclear Fuel and Waste Management Co [SKB], R-10-13,
Stockholm, Sweden, 2010; Pusch, R. et al., J. Earth Sci. Geotech. Eng., 2, 1–24, 2012.)
Third-order zones
form network of
hydraulically
active components A 3
First-order zone
4
B
4
4
50–
150
M
5 4
B
Second-order zone
5
VDH
5–
50
M
FIGURE 8.2 Rock block with optimal location of a VDH. Details to the right are element
assemblies moved out from one of the four 1 km3 blocks. The set of four blocks is in con-
tact with one large fracture zone (first-order zone) and contains some major fracture zones
(second-order) in a matrix containing a number of interacting minor fracture zones (third-
order). Red spots show intersection of the VDH and major zones. About 100–300 minor zones
are intersected.
FIGURE 8.3 Disc sawn from a 1.6 m long column of HIPOW-compressed copper powder in a
copper tube. The black spots in the copper mass represent the spent fuel, simulated by steel pel-
lets in the experiment. (Data from Lönnerberg, B. et al., Encapsulation and Handling of Spent
Nuclear Fuel for Final Disposal, SKB/KBS Technical Report 83-20, Swedish Nuclear Fuel and
Waste Management Co [SKB], Stockholm, Sweden, 1983; Courtesy of SKB, Stockholm, Sweden.)
−2.652E+1 to −2.364E+1
−2.364E+1 to −2.075E+1
−2.075E+1 to −1.786E+1
−1.786E+1 to −1.498E+1
−1.498E+1 to −1.209E+1
−1.209E+1 to −9.203E+0
−9.203E+0 to −6.316E+0
−6.316E+0 to −3.430E+0
−3.430E+0 to −5.430E−1
−5.430E−1 to 2.344E+0
2.344E+0 to 5.230E+0
5.230E+0 to 8.117E+0
8.117E+0 to 1.100E+1
1.100E+1 to 1.389E+1
X 15 MPa
Y
E—50,000 MPa
v—0.3
p—internal pressure
(20 MPa)
35 MPa
p MPa
FIGURE 8.5 Hoop stress in rock around a VDH exposed to an internal pressure of 20 MPa calcu-
lated by the FEM technique. The primary principal stresses in the rock are 35 and 15 MPa, respec-
tively. Tension stresses appear in the red parts reaching a maximum value of about 11 MPa. (Data
from Pusch, R., Geological Storage of Radioactive Waste. Springer-Verlag, Berlin, Germany, 2008.)
Measured stress on
A principal planes
A—25 MPa
B B—28 MPa
C—24 MPa
D—20 MPa
FIGURE 8.14 Sets of steeply oriented fractures in an approximately orthogonal pattern in

granite. Blue sets are oriented in the direction of the highest rock stress and carry more water
than the red ones that are compressed by this stress component. Distance between the lines
is 30 m. (Data from Pusch, R., Geological Storage of Radioactive Waste. Springer-Verlag,
Berlin, Germany, 2008.)
FIGURE 9.1 Relative monovalent anion concentration (%) in a REV for equilibrating 0.1 M
monovalent electrolyte in the microstructural network of MX-80 clay with a dry density of
1790 kg/m3 (2130 kg/m3 density at saturation with water). The picture shows that anions (red
and yellow) are confined in the center of the wide part of the widest channels (50 μm) whereas
cations (blue) are close to the mineral surfaces and in the channels (horizontal extension).
(After Muurinen, A., Lehikoinen, J., Pusch, R. POSIVA publications Report YJT-94-22,
POSIVA, Helsinki, Finland, 1994.)
FIGURE 9.4 Digitalized transmission electron micrograph of HDPY+-treated MX-80 clay
with 1760 kg/m3 density in water saturated form (thin-section by ultramicrotomy). The hori-
zontal edge length of the micrograph is 1 mm. Channels with up to 100 μm width separate
large dense and homogeneous aggregates. The white and green areas represent open space,
the red area represents dense clay matrix, and the black area shows very dense particles
representing accessory minerals. (Data from Pusch, R. et al., Microstructural and Chemical
Parameters of Bentonite as Determinants of Waste Isolation Efficiency, Final Report
European Commission, Contract No F14W-C95-0012, Stockholm, Sweden, 1999.)
FIGURE 9.9 Blood arrest by treating a deep cut in a mouse with zeolitic granules
(clinoptilolite/chabasite). One sees the abdomen region cut up and granulate applied c ausing
immediate stop of bleeding by formation of a tight scab. (Patented by Durtec GmbH,
Neubrandenburg, Germany.)
3.00 μm
×1000 120 kV 25.08.2000

Microscope: J1210
Preparation: Section
Sample: White resin
FIGURE 9.10 The transmission electron micrograph of ultrathin section of wound present
in the skin. The yellow arrows show clay in direct contact with exposed epidermis. The blue
arrow points to the remnants of the shallow epidermis embedded in the paste of montmoril-
lonite-rich clay that is seen as a wavy mass. Magnification: 100×. (Microscopy by Greifswald
University, Greifswald, Germany.)
600
500
UV flux in lux
400
300
200
100
0
0 20 40 60
FIGURE 9.12 UV penetration of 10–15 μm thick films of tested sun protection creams.
Upper: reference test of foil without clay. Middle: Tone. Lower: Serie MD2.
600
500
400
UV flux in lux
300
200
100
0
0 20 40 60
FIGURE 9.13 UV penetration of 40–60 μm thick films of tested sun protection creams.
Upper: reference test of foil without clay. Middle: Tone. Lower: MD2.
3.31%
FIGURE 9.17 Fissuring in montmorillonite-rich clay (left) and in a mixture of 50% palygor-
skite and montmorillonite clay (right) after drying at 105°C. The diameter of the circular
specimen is 30 mm; the scale being the same for both. (Data from Pusch, R., J. Cosmetics
Dermatol. Sci. Appl., 4, 67–72, 2014.)
ENVIRONMENTAL ENGINEERING
BENTONITE CL AY
Environmental Properties and Applications
Studies the Environmental, Cosmetic, and Pharmaceutical Applications

of Bentonite Clay
Bentonite clay, of which members of the smectite family of clay minerals are particularly
important, has proven to be effective in sealing off wastes from groundwater. Bentonite
Clay: Environmental Properties and Applications explores the mineralogy of
clays in general and of smectites in particular that represent challenging conditions for
geotechnical professionals responsible for earth dam construction, the foundations of
roads and buildings, and the long-term isolation of chemical and radioactive wastes. The
author, a world-renowned expert on the subject, places special emphasis on the
environmental behavior of bentonite clay when focused on the isolation of hazardous
wastes and also considers its use in cosmetics and pharmaceuticals.
Based on classical literature and current research and development, this text provides an
in-depth introduction to bentonite soil, explains the origin of smectite-rich clays, and
pinpoints where they can be found. The book describes the interaction of expandable
clay minerals, gas, and fluids, followed by a description of the physical and chemical
properties of smectite clay saturated with water or chemical solutions. It also provides
relevant findings and conclusions concerning the function of bentonite-based sealing
repositories for dangerous waste.
This text:
• Describes the constitution of smectite minerals as a basis for understanding the

behavior of smectite clays and their performance in the isolation of hazardous waste
• Factors in the longevity of smectite clays in bentonite beds and in the form of canister-
embedding buffers in repositories for deep geological disposal of highly radioactive
waste (HLW)
• Covers the design principles for clay seals and considers their function in the isolation
of waste and redirecting groundwater flow
Bentonite Clay: Environmental Properties and Applications documents the

origin, properties, and characteristics of bentonite and its uses. A resource for researchers,
practitioners, regulators, and policy makers, the text examines the use of clay in hazardous
waste and nuclear waste management and provides readers with detailed descriptions of
related technical solutions.
K23433
6000 Broken Sound Parkway, NW

Suite 300, Boca Raton, FL 33487
711 Third Avenue
New York, NY 10017
an informa business
2 Park Square, Milton Park
w w w. c r c p r e s s . c o m Abingdon, Oxon OX14 4RN, UK w w w. c rc p r e s s . c o m

Bentonite Clay

Uploaded by

Copyright:

Available Formats

Bentonite Clay

Uploaded by

Document Information

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

Bentonite Clay

Uploaded by

Copyright:

Available Formats

BENTONITE

Boca Raton London New York

CRC Press is an imprint of the

No claim to original U.S. Government works

International Standard Book Number-13: 978-1-4822-4344-4 (eBook - PDF)

I would like to dedicate this book to my beloved

Chapter 2 Origin and Occurrence of Smectite Clays: Bentonites......................... 9

Chapter 3 Nature of Smectites............................................................................. 23

Chapter 4 Clay–Fluid–Gas Systems.................................................................... 37

Chapter 5 Physical Performance of Smectite Clay Seals: Prediction

5.2.3.4 Compression Properties...............................92

5.11 Correlation of Hydraulic and Mechanical Performances

Chapter 6 Chemical Processes Involved in and Longevity of Smectite Buffer......165

Chapter 7 Processing of Clays and Preparation of Seals................................... 211

Chapter 8 Environmental Behavior................................................................... 241

8.3.2.5 Canister Performance................................ 253

Chapter 9 Pharmacology and Cosmetics........................................................... 281

9.4.1 Pilot Study of the Performance of Organic-

Appendix: Symbols and Definitions.................................................................... 311

* Thixotropy is defined as “unlimited reversible, isothermal sol/gel transformation,” sols representing

complex organic molecules, the driving force—the life spark—being provided

This chapter describes typical geotechnical problems that need to be solved in

1.1 MAIN ISSUES

1.2 SMECTITE CLAY—THE MUDDIEST TYPE OF SOFT CLAY

1.3 STRESS–STRAIN PROBLEMS

Canister in buffer clay Slope in clay

High shear stresses

* Thixotropy is defined as “unlimited reversible, isothermal sol/gel transformation,” sols representing

1.5 m Upward swelling pressure of initially 5 MPa

Dry density, Swelling pressure, MPa

0.5 m 2m Blocks of clay with 2000 kg/m3 density

Fracture filled with

1.4 PREPARATION OF SMECTITE CLAY FOR SEALING PURPOSES

1.5 QUALITY ISSUES

of the environment. Low-level radioactive waste can be safely stored in landfills

1.6 PERFORMANCE TESTS

Total pressure after 270 days

2.1 ORIGIN OF SMECTITE CLAYS

2.2 WHERE ARE ALL THESE FAMOUS AND MAGIC CLAYS?

2.2.2 South America

certain beds may be related to heating, in conjunction with hydrothermal alteration

+, low to moderate richness; +++, rich.

of which is montmorillonite and the rest being mixed-layer minerals. It is termed

ranging from fluviatile to lacustrine–palustrine environment-type deposits.

2.2.7.5 Czech Republic

2.3 POTENTIAL SMECTITE RESOURCES

2.4 ARE NEW SMECTITES BEING FORMED TODAY?

2.5 QUALITY OF NATURAL SMECTITE CLAYS FOR PRACTICAL USE

Autochtonic bentonites originating from ash fall are commonly relatively

• Available amount of exploitable smectite-rich material

Pusch, R. 1983. Stability of Deep-Sited Smectite Minerals in Crystalline Rock—Chemical

3.2 SMECTITE FAMILY

• Montmorillonite: Only Si in the tetrahedrons and Al in the octahedrons

Silica tetrahedron Fe, Al, or Mg octahedral unit

= Si = Oxygen = Oxygen or hydroxyl = Al, Fe, Mg

Basic unit layer for smectite consists of an octahedral sheet sandwiched

2:1 layer-lattice mineral particle—smectite

FIGURE 3.2 Typical appearance of stacks of smectite lamellae. Transmission electron

3.3 CRYSTAL CONSTITUTION OF SMECTITES

FIGURE 3.3 The montmorillonite crystal lattice. Left: Hofmann/Endell/Wilm version.

Traces (< 3%); Moderately abundant (3%–10%); Abundant (> 10%).

4.3.4.5 Impact of Hydraulic Gradients on the Permeation

4.3.5 Impact of Saltwater on the Hydraulic Conductivity

4.3.7 Impact of Smectite Content on the Hydraulic

4.3.7.1 Mixed Clay and Coarser Material