Porous Metals With Directional Pores

Hideo Nakajima
Porous
Metals with
Directional
Pores
Porous Metals with Directional Pores
Hideo Nakajima
Porous Metals
with Directional Pores
Hideo Nakajima
Director, The Wakasa Wan Energy Research Center
Tsuruga, Japan
Emeritus Professor, Osaka University
Suita, Japan
ISBN 978-4-431-54016-8 ISBN 978-4-431-54017-5 (eBook)

DOI 10.1007/978-4-431-54017-5
Springer Tokyo Heidelberg New York Dordrecht London
Library of Congress Control Number: 2013942871
© Springer Japan 2013

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed. Exempted from this legal reservation are brief excerpts
in connection with reviews or scholarly analysis or material supplied specifically for the purpose of being
entered and executed on a computer system, for exclusive use by the purchaser of the work. Duplication
of this publication or parts thereof is permitted only under the provisions of the Copyright Law of the
Publisher’s location, in its current version, and permission for use must always be obtained from
Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance Center.
Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Preface
Porous metallic materials such as foamed metals, sponge-like metals, structural cellular
metals, metals with directional pores, and sintered metals are increasingly looked upon
as potential light-weight structural and functional materials with, for example, superior
sound absorption, damping, and filtering properties. The porous metals are a new
category of promising engineering materials from the point of view of both fundamen-
tal science and industrial applications. To realize such applications, various problems
in fabrication such as uniformity in pore size and porosity controlled with variable
factors must be solved; furthermore, their properties must be sufficiently understood
and elucidated. Toward that end, investigations into the science and technology of
porous and foamed metallic materials definitely have been expanded recently.
It is well understood that porous and foamed metals should be useful in solving some
major issues of the twenty-first century such as environmental preservation, aging
societies, and energy problems. For example, foamed aluminum is expected to be
used for crash absorbers in automobiles and for sound absorption, while porous metals
with elongated directional pores may be useful for medical devices, machine tools, heat
sinks, and in other ways. There are various fabrication methods, classified by power
sintering, foaming, and casting techniques. Porous metals are further grouped into
porous and cellular metals depending upon the magnitude of their porosity.
Among these porous and foamed metals, the porous metals with directional pores,
the so-called lotus and gasar metals, have been attracting attention owing to their long
cylindrical pores aligned in one direction. These are considered new types of porous
metals. Recently, methods for their fabrication have almost been established at the
mass production level. Various unique physical, chemical, and mechanical
properties have been discovered and already have been fairly well understood. In
addition, several applications are proceeding at the industrial manufacturing level.
Thus, I consider that the present is a good time to organize and present the science and
relevant technology of porous metals with directional pores. I hope that readers of
this book can understand the present status of research and development of porous
metals with directional pores for the benefit of progress in their research.
Tsuruga, Japan Hideo Nakajima

Suita, Japan
v
Acknowledgements
The author would like to acknowledge the collaborations with Prof. S.K. Hyun
of Inha University in Korea, Prof. T. Ikeda of Ibaraki University, Prof. S. Suzuki
of Waseda University, Dr. M. Tane, Dr. T. Ide, Prof. K. Nakata, Prof. M. Hirao,
Prof. H. Utsunomiya, Prof. T. Nakano, Prof. S. Fujimoto of Osaka University,
Prof. O. Yoshinari of Nagoya Institute of Technology, Dr. S. Ueno of Nippon
University, Dr. T. Murakami of Tohoku University, Prof. B.-Y. Hur of Gyeongsang
National University in Korea, Dr. V. Shapovalov of MER Corporation, Ltd. of
USA, Dr. Y. Higuchi of Osaka Dental University, Prof. T. Ogushi of Hiroshima
International University and Dr. H. Chiba of Mitsubishi Electric Corporation, Ltd.
The author also expresses his appreciation to Prof. G. Stephani of Fraunhofer
Institute and Prof. J. Banhart of Helmholtz Center Berlin of Germany, and
Prof. D. Dunand of Northwestern University of USA.
vii
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Porous Materials Widespread in Natural World . . . . . . . . . . . . . 1
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Various Fabrication Methods of Cellular
Metals and Foamed Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1 Materials Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Various Fabrication Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.1 Melt Gas Injection (Air Bubbling) . . . . . . . . . . . . . . . . . 8
2.2.2 Gas-Releasing Particle Decomposition
in the Melt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
in Semisolids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.4 Casting Using a Polymer or Wax Precursor
as Template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.5 Metal Deposition on Cellular Preforms . . . . . . . . . . . . . . 10
2.2.6 Entrapped Gas Expansion . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.7 Hollow Sphere Structures . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.8 Co-compaction or Casting of Two Materials,
One Leachable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3 Fabrication Methods of Porous Metals
with Directional Pores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1 Historical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.1 Ice Wormholes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.2 Porous Metals with Directional Pores . . . . . . . . . . . . . . . 16
3.2 High-Pressure Gas Method (PGM) . . . . . . . . . . . . . . . . . . . . . . 16
3.2.1 Mold Casting Technique . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.2 Continuous Zone Melting Technique . . . . . . . . . . . . . . . 20
3.2.3 Continuous Casting Technique . . . . . . . . . . . . . . . . . . . . 24
ix
x Contents
3.3 Thermal Decomposition Method (TDM) . . . . . . . . . . . . . . . . . . 36

3.3.1 Mold Casting Technique through TDM . . . . . . . . . . . . . . 37
3.3.2 Continuous Zone Melting Technique
through TDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3.3 Continuous Casting Technique through TDM . . . . . . . . . 52
3.4 Moisture Decomposition Method . . . . . . . . . . . . . . . . . . . . . . . . 54
3.4.1 Fabrication of Lotus Nickel by Moisture . . . . . . . . . . . . . 54
3.4.2 Fabrication of Lotus Cobalt and Silicon
by Moisture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4 Nucleation and Growth Mechanism of Pores in Metals . . . . . . . . . 65
4.1 Solubility of Gases in Metals: Sieverts’ Law . . . . . . . . . . . . . . . 65
4.2 Evolution Process of Directional Gas Pores . . . . . . . . . . . . . . . . 66
4.2.1 Pore Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.2.2 Pore Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3 Model Experiment on Unidirectional Solidification
of Water Containing Carbon Dioxide . . . . . . . . . . . . . . . . . . . . . 74
4.4 Influence of Ultrasonic Agitation on Pore
Morphology during Unidirectional Solidification
of Water Containing Carbon Dioxide . . . . . . . . . . . . . . . . . . . . . 76
4.5 Evolution of Spherical Pores during
Foaming Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5 Control of Pore Size and Porosity in Lotus-Type
Porous Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.1 Control of Pore Size by Solidification Velocity . . . . . . . . . . . . . 83
5.2 Control of Pore Size and Porosity by Ambient
Gas Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.3 Control of Pore Size by Addition of Oxide Particles . . . . . . . . . . 89
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
6 Details of Fabrication Techniques of Various
Lotus Metals and Alloys, Lotus Intermetallic Compounds,
Lotus Semiconductors, and Lotus Ceramics . . . . . . . . . . . . . . . . . . 93
6.1 Fabrication of Lotus Iron by Nitrogen . . . . . . . . . . . . . . . . . . . 93
6.2 Fabrication of Lotus Silver by Oxygen . . . . . . . . . . . . . . . . . . . 98
6.3 Fabrication of Lotus Intermetallic Compounds . . . . . . . . . . . . . 102
6.4 Fabrication of Lotus Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.5 Fabrication of Lotus Alumina by Unidirectional
Solidification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
6.6 Fabrication of Lotus Brass by Solid-State Diffusion . . . . . . . . . 109
6.7 Effect of Microstructure on Pore Morphology
in Lotus Magnesium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Contents xi
6.8 Fabrication of Lotus Carbon Steel by Continuous

Casting Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
6.9 Fabrication of Lotus Aluminum by Continuous
Casting Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
7 Mechanical Properties of Lotus Metals and Alloys . . . . . . . . . . . . . 127
7.1 Elastic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.2 Internal Friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.3 Tensile Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
7.3.1 Ultimate Tensile Strength . . . . . . . . . . . . . . . . . . . . . . . . 138
7.3.2 Investigation of Tensile Deformation
by Acoustic Emission Method . . . . . . . . . . . . . . . . . . . . 142
7.4 Compressive Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.4.1 Compressive Yield Strength . . . . . . . . . . . . . . . . . . . . . . 148
7.4.2 Energy Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
7.4.3 Effect of Pore Orientation on Compressive Properties . . . 153
7.4.4 Strain Rate Dependence of Compressive Behavior . . . . . 160
7.4.5 Compression Deformation Behavior of Lotus (γ-TiAl) . . . 164
7.5 Bending Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
7.6 Fatigue Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
8 Various Physical and Chemical Properties of Lotus Metals . . . . . . 183
8.1 Sound Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
8.2 Thermal Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
8.2.1 Measurement of Effective Thermal
Conductivity of Lotus Copper . . . . . . . . . . . . . . . . . . . . 190
8.2.2 Analysis of Effective Thermal Conductivity
of Lotus Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
8.3 Electrical Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
8.3.1 Measurement of Electrical Conductivity
of Lotus Nickel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
8.3.2 Analysis of Electrical Conductivity of Lotus Nickel . . . . 194
8.4 Magnetization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
8.5 Thermal Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
8.6 Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
9 Processing of Lotus Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
9.1 Weldability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
9.1.1 Weld Fusion Property of Lotus Copper . . . . . . . . . . . . . . 211
9.1.2 Weld Fusion Property and Joint
of Lotus Magnesium . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
9.2 Equal-Channel Angular Extrusion Process . . . . . . . . . . . . . . . . . 221
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
xii Contents
10 Various Applications of Lotus Metals . . . . . . . . . . . . . . . . . . . . . . . 231

10.1 Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
10.2 Heat Sink . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
10.2.1 Heat Sink for Air Cooling . . . . . . . . . . . . . . . . . . . . . 234
10.2.2 Heat Sink for Water Cooling . . . . . . . . . . . . . . . . . . . 241
10.3 Vibration–Damping Materials . . . . . . . . . . . . . . . . . . . . . . . . . 244
10.4 Golf Putter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
10.5 Medical Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
10.5.1 In Vitro Cytocompatibility of Lotus
Ni-Free Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . 253
10.5.2 In Vivo Osteocompatibility of Lotus
Ni-Free Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . 256
10.5.3 Biocompatibility of Lotus Stainless
Steel and Titanium in Alveolar Bone . . . . . . . . . . . . . 263
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
About the Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Chapter 1
Introduction
Abstract In natural world, a number of porous materials are observed, for

example, woods, animal bones, leaves, and stalks. These holes or pores carry out
the functions of supply, lightweight property, fluid permeability, keeping warm
ability, masticability, etc. Wood, trabecular bone, and bamboo are all anisotropic;
their mechanical properties depend on the direction of loading. Natural cellular
materials exploit anisotropy to increase their mechanical efficiency, placing mate-
rial where it is most needed to resist the applied loads. Throughout this book, we
shall see the way in which porous materials exploit anisotropy to give unique
functional performance.
Keywords Bamboo • Cancellous bone • Cellular materials • Natural biomaterials

• Solidification defects
1.1 Porous Materials Widespread in Natural World
When we look around the natural world, we notice that a number of porous
materials are existent such as woods, animal bones, leaves, and stalks. Besides,
artificial materials such as foods, clothes, and buildings are not immaculate, but
porous. Figure 1.1 shows the microstructure of several cellular materials [1]. Wood
and cork are honeycomb-like cellular materials with prismatic cells like the hexag-
onal cells in a bee’s honeycomb. Sponge and cancellous tissue consist of connected
ligaments, while a coral and a bone of cuttlefish consist of stacked rectangular cells.
On the other hand, leaves are network of round cells and stems of plants have also
similar shape, but these are cross section of bundles of many tubes. As shown in
Fig. 1.2, the bones are much familiar to our human body [1]. Most bones are an
elaborate construction, made up of an outer shell of dense, compact bone, enclosing
a core of porous cellular, cancellous, or trabecular bone. Most of the configuration
minimizes the weight of bone while still providing a large bearing area, a design
which reduces the bearing stresses at the joint [1]. In a sense these bones are
H. Nakajima, Porous Metals with Directional Pores, 1

DOI 10.1007/978-4-431-54017-5_1, © Springer Japan 2013
2 1 Introduction
Fig. 1.1 Natural cellular materials: (a) cork, (b) balsa, (c) sponge, (d) cancellous bone, (e) coral,
(f) cuttlefish bone, (g) iris leaf, and (h) stalk of a plant (Reprinted with permission from [1] © 1997
Cambridge University Press)
considered as functionally gradient materials; the outer immaculate skin layer holds
strength and the porosity of the bone increases as the depth from the top surface is
inward. It is more interesting that the bones of birds that fly in sky are much lighter
weight, just like a tube, and more porous than those of animals that crawled on
1.1 Porous Materials Widespread in Natural World 3
Fig. 1.2 Cross-sectional views of (a) the head of a femur, (b) the tibia, and (c) a lumbar vertebra.
In each case, there is an outer shell of almost fully dense compact bone surrounding a core of
porous, low-density, cancellous bone (Reprinted with permission from [1] © 1997 Cambridge
University Press)
the ground as shown in Fig. 1.3 [http://www3.famille.ne.jp/~ochi/kaisetsu-01/05-

te-ashi.html]. Thus, it may be considered that such natural biomaterials are one
example to be designed as advanced materials. We, researchers who are investi-
gating metallic and inorganic advanced materials, should learn various ideas from
natural environment, in particular, the biomaterials in nature. These holes or
pores carry out the functions of channels of supply, lightweight property, fluid
permeability, keeping warm ability, masticability etc. Superficially, bones look
fairly solid, but deceptive.
It is no exaggeration to say that most of various parts of industrial products are
manufactured by casting or powder sintering techniques. In these processes, casting
defects and sintering defects such as gas pores are usually considered as harmful
defects to impede an efficiency or functional properties of the manufactured
products. Therefore, it is said to be indispensable for high performance of the
products to manufacture high-density materials with a porosity as small as possible.
Figure 1.4 shows cross-sectional view of rimmed steel ingot [2]. The top part of the
ingot has large shrinkage cavity, which is caused by volume reduction during
solidification (not appear in this photo). Inside the ingot, there are a number of
cylindrical pores which are called casting (solidification) defects. Such defects are
known to be formed by evolution of hydrogen bubbles, carbon monoxide gas
bubbles, etc. In particular, in the bottom part, some longer elongated pores are
observed. Such long pores are grown and elongated in the direction of solidifica-
tion. This book focuses attention to formation of the elongated solidification
defects, properties of metals with the elongated defects, and utilization of such
elongated defects in metals.
4 1 Introduction
Fig. 1.3 Cross sections of a bone of a bird that flies in sky (Reprinted with permission from [http://
www3.famille.ne.jp/~ochi/kaisetsu-01/05-te-ashi.html] © Nature Photo Gallery, Shinji Ochi)
Fig. 1.4 Cross-sectional view of rimmed steel ingot. Two typical solidification defects are
observed; top part is shrinkage cavity and surroundings are gas pores [2]
Up to the present, for example, penetrated pores can be used for filtering
materials and large surface area resulted from high porosity can be utilized as
electrode materials. However, if the porous materials whose mechanical strength
does not become inferior significantly can be produced, wide and various
applications to lightweight structural, functional materials, transportation materials,
etc., could be possible prospectively. The porous materials with directional pores
may meet these demands.
Natural structures such as bones have a gradient in density, rather than two
distinct solid and cellular components. For example, bamboo epitomizes this
(Fig. 1.5); the volume fraction of dense fibers increases radially toward the periph-
ery of the stem. Bamboo is also tubular, again increasing the bending stiffness of its
cross section. Wood, trabecular bone and bamboo are all anisotropic; their mechan-
ical properties depend on the direction of loading. Natural cellular materials exploit
anisotropy to increase their mechanical efficiency, placing material where it is most
needed to resist the applied loads. In tree, for instance, the highest stresses, resulting
from bending in the wind, act along the length of the trunk and branches. Wood is
much stiffer and stronger in this direction, along the grain, than across it, as a result
of its honeycomb-like cellular microstructure as well as the composite nature of the
1.1 Porous Materials Widespread in Natural World 5
Fig. 1.5 (a) A cross-section of bamboo, showing the tubular (b) structure. (b) A longitudinal
section of bamboo, showing the more or less evenly spaced diaphragms. (c) Scanning electron
micrograph of a cross-section of bamboo, showing the radial density gradient. (Reprinted with
permission from [3] © 2010 Cambridge University Press
solid cell wall material. Bone grows in response to applied loads; the trabeculae in
human vertebrae, for instance, which are subjected primarily to compressive
loading from the weight of the body, align in the vertical direction, increasing the
stiffness and strength in that direction. Throughout this book, we shall see the way
in which porous materials exploit anisotropy to give exceptional mechanical
performance.
This book is divided into three parts. The first part summarized the various
fabrication methods of cellular and foamed metals (Chap. 2), fabrication methods
of porous metals with directional pores, in particular explaining various casting
techniques and various gas-supplying techniques including the historical back-
ground (Chap. 3). Furthermore, nucleation and growth mechanism of pores in
metals in Chap. 4 and control methods of pore size and porosity in metals in
6 1 Introduction
Chap. 5 were described. Then we described the details of fabrication techniques of

various materials through various fabrication methods in Chap. 6. In the second
part, mechanical properties (Chap. 7) and various physical and chemical properties
(Chap. 8) and processing techniques (Chap. 9) are explained. Finally, in the third
part, various applications to heat sinks, vibration-damping materials, golf putter,
and medical devices are described somehow in details in Chap. 10. In Chap. 11, the
conclusions are summarized.
References
1. Gibson LJ, Ashby MF (1997) Cellular solids. Cambridge University Press, Cambridge
2. Hultgren A, Phragmen G (1939) Solidification of rimming-steel ingots. J Iron Steel Inst
139:133–244
3. Gibson LJ, Ashby MF, Harley BA (2010) Cellular materials in nature and medicine. Cambridge
University Press, Cambridge
Chapter 2
Various Fabrication Methods of Cellular
Metals and Foamed Metals
Abstract This book concerns about porous metals with directional pores. In
general, porous materials are defined as holey materials, which are mainly
characterized by porosity. The porous materials with high porosity and low density
are called as foamed materials, which resemble bubbling materials such as soap and
beer. Cellular materials look like honeycomb, whose porosity is also high and the
density is low. The holey materials, whose porosity is less than 70 %, are usually
called porous materials. Depending upon the shape of pores, isotropic and aniso-
tropic porous materials are classified. The former has rather isotropic spherical
pores, while the latter has cylindrically elongated pores aligned in one direction.
In this chapter, various fabrication techniques of foamed metals and cellular metals
are presented.
Keywords Closed pores • Hollow spheres • Metal foams • Open pores • Porosity
2.1 Materials Definition
There are several nomenclatures to call metallic materials which possess high
porosity. Usually, it is accustomed to classify these materials by their relative
density, ρ*/ρs, the density of the high-porosity metal ρ* divided by that of the solid
metal (nonporous metal) ρs. Special ultra-low-density foams can be made with a
relative density as low as 0.001. Polymeric foams used for cushioning, packing, and
insulation have relative densities which are usually between 0.05 and 0.2. As the
relative density of the foamed metals increases, the cell walls thicken and the pore
space shrinks; above about 0.3 there is a transition from a cellular structure to porous
structure which is better thought of as a solid containing isolated pores. Thus, the
porous materials are defined as the materials whose porosity ranges less than 70 %.
The shape of pores in most of foamed and sintered materials is nearly spherical
and isotropic, while the shape in lotus and gasar-type porous materials is

8 2 Various Fabrication Methods of Cellular Metals and Foamed Metals
anisotropic and long cylindrical, which is evolved unidirectionally during the

solidification. Thus, the pore morphology is defined with the following parameters:
Pore diameter (or pore size)
Pore length
Pore aspect ratio: the ratio of pore length against the diameter
Pore orientation
Porosity: pore volume fraction
Regardless of porosity, pore size, and pore shape, two kinds of pores are existent
depending upon the fabrication methods: closed pores and open pores. “Closed
pores” are surrounded by material wall and each pore is isolated, which are usually
filled with gas, while “open pores” are pores connected to each other in space; the
pores are held by struts. The closed pores are not permeable to gas or fluid, but
the open pores are usually permeable.
2.2 Various Fabrication Methods
A metallic foam was first tried to be produced by Sosnik in 1948 [1]. In order to
create pores, he put mercury into molten aluminum. In 1956 Elliot replaced
mercury with foaming agents generating gas by thermal decomposition [2], and
until now, many researchers have developed metallic foam using this technique
without the toxicity of mercury. Then in 1963, Allen invented a powder-compact
foaming technique to manufacture metallic foams, and the basic processing
techniques for such metallic foams were almost completed [3].
A variety of cellular and foamed metals have been fabricated by researchers and
industries. Metal foams and porous metals are made by the following processes,
which are summarized from the book [4].
2.2.1 Melt Gas Injection (Air Bubbling) [4]
Pure liquid metals cannot easily be caused to foam by bubbling a gas into them.
Drainage of liquid down the walls of the bubbles usually occurs too quickly to
create a foam that remains stable long enough to solidify. However, 10–30 % of
small, insoluble, or slowly dissolving particles, such as aluminum oxide or silicon
carbide, raise the viscosity of the aluminum melt and impede drainage in the bubble
membrane, stabilizing the foam. Gas-injection processes are easiest to implement
with aluminum alloys because they have a low density and do not excessively
oxidize when the melt is exposed to air or other gases containing oxygen. A variety
of gases can be used to create bubbles within liquid aluminum. Bubbles formed
by this process float to the melt surface, drain, and then begin to solidify. Low
relative density, closed-cell foams can be produced by carefully controlling the
2.2 Various Fabrication Methods 9
gas-injection process and the cooling rate of the foam. Various techniques can be
used to draw off the foam and create large (up to 1 m wide and 0.2 m thick) slabs
containing closed cell pores with diameters between 5 and 20 mm. This technique is
the least costly to implement and results in a foam with relative densities in the
range 0.03–0.1.
2.2.2 Gas-Releasing Particle Decomposition in the Melt [4]
Metal alloys can be foamed by mixing into them a foaming agent that releases gas
when heated. The widely used foaming agent titanium hydride (TiH2) begins to
decompose into Ti and gaseous H2 when heated above about 738 K. By adding
titanium hydride particles to aluminum melt, large volumes of hydrogen gas are
rapidly produced, creating bubbles that can lead to a closed-cell foam, provided
foam drainage is sufficiently slow, which requires a high-melt viscosity. The
Shinko Wire Company has developed an aluminum foam trade named Alporas
using this technique. The process begins by melting aluminum and stabilizing the
melt temperature between 943 and 963 K. Its viscosity is then raised by adding
1–2 % of calcium which rapidly oxidized and forms finely dispersed CaO and
CaAl2O4 particles. The melt is then aggressively stirred and 1–2 % of TiH2 is added
in the form of 5–20 μm diameter particles. As soon as these are dispersed in the
melt, the stirring system is withdrawn, and a foam is allowed to form above the
melt. Control of the process is achieved by adjusting the overpressure, temperature,
and time. When foaming is complete, the melt is cooled to solidify the foam before
the hydrogen escapes and the bubbles coalesce or collapse. The cell size can be
varied from 0.5 to 5 mm by changing the TiH2 content, and the foaming and cooling
conditions. Relative densities from 0.2 to as low as 0.07 can be obtained. As
produced, the Alporas foam has predominantly closed cells, though a subsequent
rolling treatment can be used to fracture many of the cell walls in order to increase
their acoustic damping.

in Semisolids [4]
Foaming agents can be introduced into metals in the solid state by mixing and
consolidating powders. Titanium hydride, a widely used foaming agent, begins to
decompose at about 738 K, which is well below the melting point of pure aluminum
and of its alloys. This raises the possibility of creating a foam by dispersing the
foaming agent in solid aluminum using powder metallurgy processes and then
raising the temperature sufficiently to cause gas release and partial or full melting
of the metal, allowing bubble growth. Cooling then stabilizes the foam. It begins by
combining particles of a foaming agent with an aluminum alloy powder. After the
ingredients are thoroughly mixed, the powder is cold compacted and then extruded
into a bar or plate of near theoretical density. This precursor material is chopped into
small pieces, placed inside a sealed split mold, and heated to a little above the solidus
temperature of the alloy. The titanium hydride then decomposes, creating voids with
a high internal pressure. These expand by semisolid flow and the aluminum swells,
creating foam that fills the mold. The process results in components with the same
shape as the container and relative densities as low as 0.08. The foam has closed cells
with diameters that range from 1 to 5 mm in diameter.
2.2.4 Casting Using a Polymer or Wax Precursor

as Template [4]
An open-cell polymer foam mold template with the desired cell size and relative
density is first selected. This can be coated with a mold casting slurry which is then
dried and embedded in casting sand. The mold is then baked both to harden the
casting material and to decompose the polymer template, leaving behind a negative
image of the foam. This mold is subsequently filled with a metal alloy and allowed
to cool. The use of a moderate pressure during melt infiltration can overcome the
resistance to flow of some liquid alloys. After directional solidification and cooling,
the mold materials are removed, leaving behind the metal equivalent of the original
polymer foam. Metal powder slurries can also be used instead of liquid metals.
These are subsequently sintered. The method gives open-cell foams with pore sizes
of 1–5 mm and relative densities as low as 0.05. The process can be used to
manufacture foams from almost any metal that can be investment cast.
2.2.5 Metal Deposition on Cellular Preforms [4]
Open-cell polymer foams can serve as templates upon which metals are deposited by
chemical vapor decomposition (CVD), by evaporation, or by electrodeposition.
Nickel is deposited by the decomposition of nickel carbonyl, Ni(CO)4. An open-
cell polymer is placed in a CVD reactor and nickel carbonyl is introduced. This gas
decomposes to nickel and carbon monoxide at a temperature of about 373 K and
coats all the exposed heated surfaces within the reactor. Infrared or RF heating can
be used to heat only the polymer foam. After several tens of micrometers of the metal
have been deposited, the metal-coated polymer foam is removed from the CVD
reactor and the polymer is burnt out by heating in air. This results in a cellular metal
structure with hollow ligaments. A subsequent sintering step is used to densify
the ligaments. The foam with open sizes in the 100–300 μm diameter range is
available. It gives the lowest relative density (0.02–0.05) foams available today.
2.2 Various Fabrication Methods 11
2.2.6 Entrapped Gas Expansion [4]
In the process Ti-6Al-4V powder is sealed in a canister of the same alloy. The
canister is evacuated to remove any oxygen and then backfilled with between 0.3
and 0.5 MPa of argon. The canister is then sealed and consolidated to a high
relative density (0.9–0.98) by HIPing causing an eightfold increase in void
pressure. This is too low to cause expansion of Ti-6Al-4V at room temperature.
The number of pores present in the consolidated sample is relatively low, so a
rolling step is introduced to refine the structure and create a more uniform
distribution of small pores. In titanium alloys, rolling at 1,173–1,213 K results
in void flattening and elongation in the rolling direction. As the voids flatten,
void faces come into contact and diffusion bond, creating strings of smaller
gas-filled pores. Cross rolling improves the uniformity of their distribution.
Various cold sheet forming processes can then be used to shape the as-rolled
plates. The final step in the process sequence is expansion by heating at 1,173 K
for 20–30 h. The high temperature raises the internal pore pressure by the ratio of
the absolute temperature of the furnace to that of the ambient, i.e., to between
10 and 16 MPa, causing creep dilation and a reduction in the overall density of
the sample. This process results in shaped Ti alloy sandwich construction
components with a core containing a closed-cell void fraction of up to 0.5 and
a void size of 10–300 μm.
2.2.7 Hollow Sphere Structures [4]
Several approaches have recently emerged for synthesizing hollow metal

spheres. For one example, hollow spheres are formed from a slurry composed
of a decomposable precursor such as TiH2, together with organic binders and
solvents. The spheres are hardened by evaporation during their flight in a tall drop
tower, heated to drive off the solvents and to volatilize the binder. A final heat
treatment decomposes the metal hydride, leaving hollow metal spheres. The
approach can be applied to many materials and is not limited to hydrides. As
an example, an oxide mixture such as Fe2O3 plus Cr2O3 can be reduced to create
stainless steel.
In another method developed at IFAM, Bremen, polystyrene spheres are
coated with a metal slurry and sintered, giving hollow metal spheres of high
uniformity. The consolidation of hollow spheres gives a structure with a mixture
of open and closed porosity. The ratio of the two types of porosity and overall
relative density can be tailored by varying the starting relative density of the
hollow spheres and the extent of densification during consolidation. Overall
relative densities as low as 0.05 are feasible with a pore size in the range
100 μm to several millimeters.
2.2.8 Co-compaction or Casting of Two Materials,

One Leachable [4]
Two powders, neither with a volume fraction below 25 %, are mixed and
compacted, forming double-connected structures of both phases. After consolida-
tion one powder (e.g., salt) is leached out in a suitable solvent. Foams based on
powder mixes of aluminum alloys with sodium chloride have successfully been
made in large sections with uniform structures. The resulting cell shapes differ
markedly from those of foams made by other methods. In practice the method is
limited to producing materials with relative densities between 0.3 and 0.5. The cell
size is determined by the powder particle size and lies in the range 10 μm–10 mm.
References
1. Sosnik B (1948) US Patent No. 2,434,775

2. Elliot JC (1956) US Patent No. 2,751,289
3. Allen BC (1963) US Patent No. 3,087,807
4. Ashby MF, Evans A, Fleck NA, Gibson LJ, Hutchinson JW, Wadley HNG (2000) Metal foams:
a design guide. Butterworth-Heinemann, Boston
Chapter 3
Fabrication Methods of Porous Metals
with Directional Pores
Abstract Porous metals with directional pores were investigated from long time ago
from the viewpoint to elucidate solidification or casting defects. In 1980s Shapovalov
et al. indicated some applicability porous metals with directional pores fabricated
through gasar process using high-pressure hydrogen. Then, Nakajima et al. carried
out systematic investigations using various casting techniques such as mold casting
technique, continuous zone melting technique, and continuous casting technique.
The latter is the most superior to control pore size and porosity and is the most
suitable for fabricating long-sized casting slabs. To overcome technical difficulty of
use of high-pressure hydrogen, very simple fabrication method was invented recently
by adding only gas-forming compounds into melt.
Keywords Continuous casting technique • Continuous zone melting technique

• Gasar process • High-pressure gas method • Lotus metals • Mold casting technique
• Thermal decomposition method
3.1 Historical Background
3.1.1 Ice Wormholes
Chalmers [1] observed porous ice with directional pores called as “ice worms” and
discussed the mechanism of the pore formation to the following. Air is a solute that
is rejected by water ahead of the freezing process during the solidification of water.
It accumulates in the advancing interface until its concentration is high enough for
bubbles to nucleate. Once a bubble has formed, it grows because air diffuses into
it. If the interface continues to move forward, the bubble cannot grow laterally, and
so, it grows forward to form the cylindrical bubble sometimes known as an ice
worm as shown in Fig. 3.1. Close inspection shows that ice worms never start at the
surface of the ice, but always a little way in; this is because some freezing must

14 3 Fabrication Methods of Porous Metals with Directional Pores
Fig. 3.1 Ice worms form during the growth of an ice cube because the ice crystal (grew from the
bottom) rejects air that was dissolved in the water (Reprinted with permission from [1] © 1959
Scientific American, Inc.)
occur before the air accumulating at the interface reaches sufficient concentration to
cause the nucleation of a bubble. Here again growth occurs more easily than the
nucleation that starts it.
The ice worms that grow in ice cubes frequently look like strings of pearls. This
reflects the fluctuation in the freezing rate due to the intermittent operation of the
refrigerator. When freezing is slow, more air diffuses into the bubbles and they
grow larger; during periods of fast growth, there is less times for diffusion and the
bubble decreases in cross section. Fast freezing suppresses the formation of ice
worms. Because insufficient air diffuses into the bubbles to permit them to grow,
the ice contains a large number of very small, round bubbles. On the other hand,
very slow freezing permits the rejected air to diffuse away from the interface;
because the concentration of air never reaches the nucleation point, neither bubbles
nor ice worms appear. Ice grown in flowing water is also usually free of bubbles and
ice worms; the continuous removal of the water prevents the buildup of a high
concentration of dissolved air, hence the absence of bubbles in the ice cubes
produced by some commercial freezers.
For freezing to take place continuously, the heat of fusion must somehow be
carried away from the region where ice is forming. Normally, as in the growth of ice
cubes or of ice on the surface of a lake, the heat is extracted by conduction through
the ice that has already formed. The water–ice interface tends to remain flat,
because any region of convex curvature is cooled less efficiently than the remain-
der, and its rate of freezing is retarded until a smooth interface is restored. However,
if the ice crystal is growing in supercooled water, its growth takes an entirely
3.1 Historical Background 15
Fig. 3.2 Growth of ice

worm. Diagonal hatching
shows accrual of ice; the
arrows, dissolved air
diffusing in to worm. (a) A
pore nucleates in water in
vicinity of ice. (b) As the
interface between ice and
water proceeds upward,
enriched air which is rejected
in the ice diffuses into the
pore. (c) An elongated-
cylindrical pore continues to
grow in the direction of
movement of the interface.
(Reprinted with permission
from [1] © 1959 Scientific
American, Inc.)
different form. Because the latent heat of fusion flows outward into the supercooled
liquid, cooling is more efficient at a convex region; convex regions therefore
become even more convex. The latent heat produced by the growth of a convexity
suppresses growth in the vicinity, and so a projection becomes an isolated spike as
illustrated in Fig. 3.2. The direction in which such branching or “dendritic” growth
may go in various substances depends on the characteristic molecular arrangement
of the crystal. In the case of ice, the spike may grow in any one of six directions.
Once a spike has formed, the supercooled environment may permit the growth of
lateral secondary, ternary, and even quaternary branching spikes, always in the
proper crystallographic directions. This explanation is highly suggestive to investi-
gate the following porous metal science.
3.1.2 Porous Metals with Directional Pores
Recently, a new type of porous metals whose long-cylindrical pores are aligned in
one direction has been fabricated by unidirectional solidification under a
pressurized hydrogen, nitrogen, or oxygen gas. Many gas pores are evolved from
insoluble hydrogen (nitrogen or oxygen) in solids while hydrogen (nitrogen or
oxygen) dissolves significantly in liquids. Formation of elongated gas pores during
solidification has been investigated by Imabayashi et al. [2], by Svensson and
Fredriksson [3], and by Knacke et al. [4]. Furthermore, Shapovalov et al. [5]
fabricated longer cylindrical pores by adopting a unidirectional solidification tech-
nique at a high pressure of hydrogen. Nakajima et al. [6, 7] also produced porous
iron, copper, magnesium, nickel, and those alloys in high-pressure hydrogen or
nitrogen atmosphere and porous silver in high-pressure oxygen by means of the
Czochralski method and a unidirectional solidification method. The processing
technique is different from that of cellular and foamed metals, since it allows a
control of pore size, pore direction, and overall porosity. High-pressure hydrogen or
other gas is filled in the pores of the porous metals during an invariant reaction of
the so-called gas-evolution crystallization reaction, where the melt is solidified to
transform into a solid solution and a gas phase [8]. During solidification, the gas is
rejected from the solid metal at the solid–liquid interface and forms long pores that
are aligned parallel to the solidification direction. Shapovalov et al. named this
method as gasar, which means a Ukrainian acronym for gas-reinforced composite
metals [9]. However, it turns out by the systematic investigations by Nakajima
et al. that gas pores do not affect to reinforce the porous metals. Therefore, a term
from their shape is considered to be more suitable for the name of the porous metals.
Nakajima group called their materials as “lotus-type porous metals” (hereafter we
call lotus metals) because the morphology of the material resembles that of a lotus
root. The strength of lotus metals is superior to that of conventional porous metals
such as cellular and foamed metals [7, 10].
Thus, such lotus and gasar metals are expected as a new category of engineering
materials.
3.2 High-Pressure Gas Method (PGM)
3.2.1 Mold Casting Technique
As mentioned above, the gas pores are evolved by insoluble gas atoms in the solid
when the melt dissolving gas is solidified because of the gas solubility difference
between liquid and solid. Figure 3.3 shows the temperature dependence of hydro-
gen solubility in both states of various metals [11]. The solubility of hydrogen
increases with increasing temperature in both solid and liquid. For the metals in
which a discontinuous abrupt decrease is observed in the solubility at the melting
3.2 High-Pressure Gas Method (PGM) 17
100
10-1 Mg
10-2 Fe
Solubility / mol%
10-3
Cu Si
10-4
Al
10-5
10-6
400 600 800 1000 1200 1400 1600 1800
Temperature /K
Fig. 3.3 Temperature dependence of hydrogen solubility in solids and liquids of various metals
under the hydrogen pressure of 0.1 MPa [11]
point when the temperature is lowered, significant pore evolution occurs in the
solidified metals so that the lotus and gasar structure can be easily fabricated during
the unidirectional solidification.
So far, casting method is usually adopted as the fabrication technique of the lotus
and gasar metals. The peculiar difference of lotus and gasar metals from the cellular
and foamed metals is the pore morphology; while the former has elongated cylin-
drical pores aligned in the solidification direction, the latter has always almost
spherical pores distributed randomly in the matrix. Figure 3.4a shows schematic
drawing of the mold casting technique to fabricate lotus and gasar metals; a metal
inside a crucible is melted by an induction heating in a high-pressure gas atmo-
sphere. The gas is dissolved up to the equilibrium gas concentration into the molten
metal under a given gas pressure according to the Sieverts law [12]. The melt
saturated with gas is poured into the mold. When some part of the mold is cooled
down by a chiller or circulated water, the melt can be solidified unidirectionally
from the vicinity of the cooling part. Figure 3.4b shows the overviews of the
fabrication apparatus of mold casting technique installed at Nakajima Laboratory
of Osaka University. The elongated pores can evolve and grow by the influence of
the unidirectional solidification. The pore growth direction can be controlled by
changing the location of the cooling part (chiller). Figure 3.5 illustrates three
different types of the molds. When the bottom of the mold is water-cooled, the
molten metal is unidirectionally solidified upwards from the bottom part so that the
directional pores grow in the upward direction. When the lateral side of the mold is
cooled down, the solidification takes place inwardly from the surrounding, and thus,
the pore distribution becomes radial. If a particular cooling part is not set up in the
Fig. 3.4 (a) Schematic drawing of principle for fabrication of lotus metals through mold casting
technique (Reprinted with permission from literature of Prog Mater Sci 52(2007) 1091–1173,
©2006 Elsevier Ltd). (b) Overview of the apparatus for fabrication of lotus metals through mold
casting technique installed at Nakajima Laboratory of Osaka University. The right-hand upper
figure shows that the metal is melted by the crucible by heating radio-frequency induction coil. The
right-hand lower figure shows that after the stopper rod is lifted to open the bottom hole of the
crucible, the melt is dropped down through a funnel into the hearth whose bottom is cooled by a
chiller. Unidirectional solidification takes place to produce the lotus metals
a b c
Mold Mold
Mold
Chiller
Chiller
Fig. 3.5 Three different types of molds for casting and resulting pore configurations in the porous
metals: (a) unidirectional pores, (b) radial pores, and (c) isolated spherical pores. (a) and (b) are
anisotropic pores, while (c) is isotropic pores (Reprinted with permission from literature of Prog
Mater Sci 52(2007) 1091–1173, © 2006 Elsevier Ltd)
mold, the spherical pores start to grow everywhere and distributed randomly whose
morphology resembles pumice stones.
In general, the porous, cellular, and foamed metals are characterized by pore
growth direction, pore size, and porosity. As mentioned earlier, while the pore
direction, pore size, and porosity of the cellular and foamed metals are not easy to
control by the fabrication methods, those of the lotus and gasar metals are relatively
easier by those fabrication methods because the direction of the solidification, gas
pressure, gas content, and the solidification velocity can be controlled. The
parameters to control the pore morphology can be listed up as
the melt temperature
the solidification rate
the temperature gradient of the interfacial region of liquid–solid phase
the dissolving gas pressure during melting and solidification
the inert gas pressure during melting and solidification
Table 3.1 Lotus metals to be fabricated and available gases

Available gas Lotus metals to be fabricated
Hydrogen Iron, carbon steel, stainless steel, aluminum, magnesium, nickel,
copper, cobalt, tungsten, manganese, chromium, beryllium, and
those alloys
Nitrogen Iron, carbon steel, stainless steel
Oxygen Silver
The lotus metals to be fabricated and available gases are compiled in Table 3.1.
Hydrogen gas is used to fabricate various lotus metals and alloys: iron, nickel,
aluminum, copper, magnesium, cobalt, tungsten, manganese, chromium, beryllium,
and those alloys. Fortunately, most of the base metals for commercially available
practical alloys can be made porous. However, there are some problems; as the
hydrogen concentration dissolving in molten aluminum is small, the porosity of
lotus aluminum is less than 40 % by the present technique. On the other hand,
although fabrication of porous titanium is possible, the pore shape is not cylindrical,
but spherical, because of high dissolving concentration of hydrogen.
Since hydrogen is inflammable and explosive when oxygen is present, its use is
not convenient from the industrial point of view. Use of other gases than hydrogen
is desirable. It is well known that nitrogen is an important alloying element widely
used to improve corrosion resistance and mechanical properties of steels. The
temperature dependence of nitrogen solubility in solid and liquid of iron is similar
to that of hydrogen, which exhibits a large nitrogen solubility difference between
solid and liquid of iron at the melting temperature [13–15]. Moreover, it is known
that an invariant reaction [8] of “gas-evolution crystallization reaction” takes place
in the Fe-N system, in which the iron melt dissolving nitrogen is solidified to
transform into a primary solid solution and nitrogen gas. Utilizing the nitrogen
solubility difference between liquid and solid, lotus iron was fabricated by the
abovementioned technique [16]. Using oxygen gas, porous silver can be produced,
because metallic oxide of silver cannot be formed in oxygen atmosphere when the
molten silver is solidified [17].
Figure 3.6 shows typical examples of optical micrographs on the cross section
(above) and the longitudinal sections (below) of lotus copper.
3.2.2 Continuous Zone Melting Technique
The mold casting technique was usually used to produce lotus-type porous metals.
For a few porous metals such as copper [6, 7] and magnesium [18] as shown in
Fig. 3.6, long-cylindrical pores were grown in the same direction as the unidirec-
tional solidification. The pore size in the section perpendicular to the solidification
direction depends on the solidification velocity; the higher the solidification veloc-
ity, the smaller the pore size [19]. Since these metals exhibit high-thermal
Fig. 3.6 Typical examples of optical micrographs on the cross sections of lotus-type porous
copper fabricated through mold casting technique at different hydrogen pressure. The above is the
cross sections perpendicular to the solidification and the below is the cross sections parallel to the
solidification. (a) 0.4 MPa hydrogen pressure, porosity 44.9 %, and (b) 0.8 MPa hydrogen
pressure, porosity 36.6 % (Reprinted with permission from literature of Prog Mater Sci 52
(2007) 1091–1173, © 2006 Elsevier Ltd)
conductivity, the solidification proceeds with almost constant solidification rate

through the whole of ingot. Thus, porous metals with uniform pore size and porosity
can be produced by this technique as shown in Fig. 3.6. However, the same
technique cannot be applied to fabricate porous metals and alloys with
low-thermal conductivity such as stainless steel [20]. For the metals and alloys
with low-thermal conductivity, although the heat from the melt is easily dissipated
to the water-cooled plate during the solidification process, the cooling becomes
slower at the upper part of the solidified ingot where it is far from the cooling part
and thus the pores become coarse. As a result only porous metals and alloys with
nonuniform pore size and porosity can be produced as illustrated in Fig. 3.7. In order
a b
Copper, Magnesium Stainless steel
(High thermal conductivity) (Low thermal conductivity)
gradually decreases
Solidification rate is
Solidification rate
constant
Chiller Chiller
2 mm 2 mm
Fig. 3.7 Comparison of the evolution of pores in porous metals fabricated by mold casting
technique in gas atmosphere. The above is schematic drawings of pore evolution during the
unidirectional solidification. The below is optical micrographs of the sectional views perpendicular
to the solidification. The left is lotus copper, and the right is stainless steel. (a) The uniform pore
size and porosity are observed in copper and magnesium with high thermal conductivity, and (b)
various pore size and porosity are found in stainless steel with low-thermal conductivity. The
magnitude of the thermal conductivity affects the solidification velocity of the melt (Reprinted
with permission from literature of Prog Mater Sci 52(2007) 1091–1173, © 2006 Elsevier Ltd)
to overcome the shortcoming, a novel technique was invented by the present author
to fabricate pore-elongated lotus metals and alloys even with low-thermal conduc-
tivity [21, 22]. Figure 3.8 illustrates the schematic setup of the continuous zone
melting technique, which consists of radio-frequency induction coil, blowers,
specimen rod, and movable specimen holders; the induction coil is used for the
zone (restricted area) melting of the rod-shaped specimen, while the blower is
helpful for further cooling of the melt metal. These components are placed into a
high-pressure chamber filled with gases such as hydrogen (or nitrogen) and argon.
While a part of the specimen rod is melted by induction heating, the hydrogen
(or nitrogen) gas is absorbed into the melt up to the gas equilibrium solubility in the
pressurized gas atmosphere according to Sieverts’ law. Concurrently, the specimen
Continuous Zone Melting Technique
Under pressurized
Metal rod
hydrogen gas
Nonporous metal
Induction coil
Blower Melting zone
Moving
Solidifying
Porous metal direction
Vector sum of
inward cooling
and upward cooling
Moving velocity
160, 330, 500 µm/s
Fig. 3.8 Schematic drawings and a photograph of the overview for the melting part of continuous
zone melting technique. After passing the melting zone, the rod in the lower part is expanded to
form lotus metals (Reprinted with permission from literature of Prog Mater Sci 52(2007)
1091–1173, © 2006 Elsevier Ltd)
rod is moved downward at a given transfer velocity. In the lower part of the melt
zone, the solidification takes place simultaneously. Then, directional elongated
pores are evolved by precipitation of insoluble gas of hydrogen (or nitrogen) in
the solidified specimen rod. If the transfer velocity is kept constant, the solidifica-
tion velocity becomes constant so that the pore size should be constant. The
direction of the growing pores is determined by the vector sum of two kinds of
solidification directions: an inner cooling vector and an upward cooling vector as
shown in the same figure. It is noticed after observation of the sectional view of the
specimen that the upward cooling is usually dominant in the interior part of the rod,
while the inner cooling effect cannot be ignored near the surface region of the
specimen rod. The stainless steel (e.g., SUS304L) exhibits a low-thermal conduc-
tivity so that porous stainless steel with homogeneous pore size and porosity is
impossible by the conventional mold casting technique as shown in Fig. 3.7.
However, lotus stainless steel with homogeneous pore size and porosity is success-
fully fabricated by so-called continuous zone melting technique. This technique has
an advantage that we can control the solidification velocity by changing the
Fig. 3.9 Overview of the fabrication apparatus of lotus metals through continuous zone melting
technique installed at Nakajima Laboratory of Osaka University
lowering speed of the specimen rod regardless of the magnitude of thermal conduc-
tivity [23]. On the other hand, the solidification velocity is not controllable to a wide
range by the conventional mold casting technique and is almost uniquely deter-
mined by its own inherent thermal conductivity. Figure 3.9 shows the photographs
of the overview of the fabrication apparatus for continuous zone melting technique.
Figure 3.10 shows a sectional view of a lotus stainless steel rod (SUS304L) and the
pore size and the porosity are almost identical everywhere through the solidified
specimen rod more than 300 mm in length. The length of the lotus metals and alloys
to be fabricated by the continuous zone melting technique is essentially endless, but
in the present chamber, the movable height is limited to less than 300 mm. This
continuous zone melting technique is a promising technique in order to produce
long-sized lotus rods for commercial application.
3.2.3 Continuous Casting Technique
The continuous casting technique is extensively used as a mass-production method

for ferrous and nonferrous metals and alloys. In this conventional continuous
casting process, the solidified ingot can pass through the mold smoothly due to
the solidification shrinkage from the melt. However, for fabrication process of lotus
Fig. 3.10 Sectional views of lotus stainless steel fabricated by continuous zone melting technique
in the 2.0 MPa hydrogen atmosphere. The transfer velocity of the rod is 330 μm s1. Resulting
porosity and average pore size are 40 % and 320 μm, respectively, both of which are almost
uniform in the whole part of the ingot (Reprinted with permission from literature of Prog Mater Sci
52(2007) 1091–1173, © 2006 Elsevier Ltd)
metals, a large volume expansion due to large amount of pore formation occurs
when the solidified ingot is passed through the mold. It was initially thought that
such an expansion would present problems with the stacking of the ingot in the
mold and that consequently this technique could not be applied to the fabrication of
lotus metals. However, it was later realized that such a large expansion inherent
from the pore evolution was released to push the volume toward the copper part of
the molten metal, so that the melt can accommodate the large strain of the solidified
ingot. That is why the continuous casting technique is applicable to the fabrication
of lotus metals.
Through this technique, the solidification velocity can be controlled by the
transfer velocity in hydrogen gas atmosphere. Since the pore morphology of lotus
metal is related to the solidification velocity, it is suggested that the pore morphology
of lotus metals can be easily controlled by the technique. Park et al. successfully
fabricated long-sized lotus copper by controlling its pore size and porosity [24].
Slabs of lotus copper were fabricated by a vacuum-assisted and pressurized
continuous casting apparatus, as illustrated in Fig. 3.11. The apparatus consists of
a crucible with a rectangular hole at the bottom, a dummy bar for preventing the
melt from flowing through the hole, and induction heating coil, and a mold which is
surrounded by a water-cooled chill block and pinch rollers to control the transfer
velocity of the dummy bar. The solidified ingot bar (slab) can be produced to
700 mm long. Pure copper was melted in the crucible by radio-frequency induction
Fig. 3.11 Schematic drawing of continuous casting apparatus (Reprinted with permission from
[24]. © 2007 Acta Materialia Inc.)
heating under a hydrogen gas pressure of 2.0 MPa. The temperature of the molten
copper in the crucible was monitored by a W-5Re/W-26Re thermocouple, which
was available in the hydrogen atmosphere and was set to be 1,573 K. The melt was
pulled down by the dummy bar of nonporous copper through the cooled mold at a
given transfer velocity. By the cooling of the melt through both the mold and the
connected dummy copper bar, the melt was simultaneously and continuously
solidified. Then, the hydrogen in the melt was rejected at the solid–liquid interface
due to the solubility gap of hydrogen between liquid and solid, and cylindrical pores
aligned parallel to the solidification direction were formed. Figure 3.12 shows the
photographs of the overview of fabrication apparatus for continuous casting tech-
nique installed at Nakajima Laboratory of Osaka University.
Figure 3.13a, b shows outer view and cross-sectional view of lotus copper rod
fabricated in the mixture gas of hydrogen 0.25 MPa and argon 0.15 MPa under
transfer velocity of 100 mm min1, respectively. The pore size and porosity are
almost identical everywhere through the solidified specimen plate more than
700 mm. The length of the fabricated lotus metals should be essentially endless,
but in this chamber, there is some size limitation; the movable height was limited to
less than 1,000 mm. Figure 3.14 shows the cross-sectional views of the lotus copper
fabricated at various transfer velocities under various hydrogen gas pressure; these
views are parallel and perpendicular to the transfer direction. In the observed cross
Fig. 3.12 Photographs of the overview of fabrication apparatus for the continuous casting
technique installed at Nakajima Laboratory of Osaka University. The central upper photo shows
the upper chamber set up by heating part and solidification part, while the central lower photo
shows lower chamber set up by mechanical pinch roller to pull down the ingot slab
section of the slabs fabricated at 1 mm min1 under 1.0 or 2.0 MPa hydrogen. Three
large (diameter 5 mm) and many small pores are distributed inhomogeneously;
many small pores are also found in the interior of the large pores. The shapes of the
pores are irregular. Thus, it is considered that the large pores are found by the
cohesion of a few small pores; the volume of the small pores increases during
solidification when the amount of hydrogen diffused from the solid to the pores
increases with decreasing transfer velocity. However, the slabs fabricated at a high-
transfer velocity possess long-cylindrical pores distributed homogeneously. The
pore size decreases and the number density of the pores increases with an increase
in not only the transfer velocity but also the hydrogen gas pressure.
Figure 3.15 shows the relationship between the porosity and the transfer velocity
under hydrogen gas pressures of 1.0 and 2.0 MPa. The porosity decreases with
increasing hydrogen gas pressure, while the porosity is almost constant and inde-
pendent of the transfer velocity. Such a tendency is consistent with the results
reported by Hyun and Nakajima [19] and Ikeda et al. [23]; they respectively
fabricated lotus copper and stainless steel using the mold casting and continuous
zone melting techniques at various solidification velocities under various gas
pressure. Figure 3.16 shows the effect of transfer velocity on the average pore
Fig. 3.13 (a) A photograph of the overview of lotus copper fabricated using the continuous
casting technique and (b) cross sections parallel and perpendicular to the transfer direction of the
lotus copper fabricated in mixture gases of hydrogen 0.25 MPa and argon 0.15 MPa by continuous
casting technique (Reprinted with permission from literature of Prog Mater Sci 52(2007) 1091-
1173, © 2006 Elsevier Ltd)
Fig. 3.14 Cross sections perpendicular and parallel to the transfer direction of lotus copper
fabricated at various transfer velocities under hydrogen gas pressure of 1.0 and 2.0 MPa (Reprinted
with permission from [24]. © 2007 Acta Materialia Inc.)
Fig. 3.15 Porosity against the transfer velocity of lotus copper fabricated under hydrogen gas
pressure of 1.0 and 2.0 MPa (Reprinted with permission from [24]. © 2007 Acta Materialia Inc.)
Fig. 3.16 Average pore diameter against the transfer velocity of the lotus copper fabricated under
hydrogen gas pressure of 1.0 and 2.0 MPa (Reprinted with permission from [24]. © 2007 Acta
Materialia Inc.)
Fig. 3.17 Pore number density calculated using the average pore diameter and the porosity in a
unit area (1.0 1.0 mm2) on a cross section perpendicular to the transfer direction at various
transfer velocities under hydrogen gas pressure of 1.0 and 2.0 MPa (Reprinted with permission
from [24]. © 2007 Acta Materialia Inc.)
diameter d under each hydrogen gas pressure. The average pore diameter decreases
with an increase in not only hydrogen gas pressure but also the transfer velocity.
In addition, the density of the pore number increases with increasing transfer
velocity as shown in Fig. 3.14. Figure 3.17 shows the change of the pore number
density Np calculated using the average pore diameter and the porosity in a cross-
sectional area of 1.0 1.0 mm2, with the transfer velocity. The pore number
density increases with increasing transfer velocity, while an increase in hydrogen
gas pressure brings about an increase in the pore number density for a given
transfer velocity. These indicate that the pore number density is affected not
only by the transfer velocity but also by the total gas pressure. The pore length l
was also affected by the transfer velocity as shown in Fig. 3.18, which decreases
with increasing transfer velocity.
The pore growth direction is affected by the transfer velocity. Figure 3.19 shows
the angle θ between the pore growth direction and the transfer direction in various
positions of the slab in a direction perpendicular to transfer direction. In the center
of the slab, the angle is almost zero and independent of the transfer velocity. On the
other hand, the pore growth angle increases when the position of the slab moves
from center to near the surface, and the pore growth angle increases more with
increasing transfer velocity. Thus, the pore growth angle is affected by the transfer
velocity, and the effect increases with increasing transfer velocity. Figure 3.20
shows the magnifications and the schematics of the pore formation position near the
surface. The thickness t decreases with increasing transfer velocity as shown in
Fig. 3.21.
Fig. 3.18 Pore length against the transfer velocity of lotus copper fabricated under hydrogen gas
pressure of 1.0 and 2.0 MPa (Reprinted with permission from [24]. © 2007 Acta Materialia Inc.)
Fig. 3.19 Angle between the pore growth direction and the transfer direction in various positions
of the slab in a direction perpendicular to the transfer direction; the slab was fabricated at various
transfer velocities under a hydrogen gas pressure of 1.0 MPa (Reprinted with permission from
[24]. © 2007 Acta Materialia Inc.)
Fig. 3.20 Photographs (upper row) and schematic (lower row) of the pore formation position near
the surface fabricated by the continuous casting technique at the transfer velocity of (a)
20 mm min1, (b) 50 mm min1, and (c) 100 mm min1 under a hydrogen pressure of 1.0 MPa
(Reprinted with permission from [24]. © 2007 Acta Materialia Inc.)
Fig. 3.21 Thickness of the skin layer plotted against the transfer velocity of lotus copper
fabricated under a hydrogen gas pressure of 1.0 MPa (Reprinted with permission from [24].
© 2007 Acta Materialia Inc.)
In order to discuss the pore diameter and the pore length, the pore nucleation
mechanism must be taken into consideration. Pore nucleation in the liquid has been
investigated by many researchers [25, 26], and it has been reported that inhomoge-
neous pore nucleation occurs because the surface Gibbs free energy is lowered by
the existence of impurities and inclusions. Fisher [26] suggested the following
relationship between the pore nucleation rate I and the critical Gibbs free energy
ðΔGhetero Þ for heterogeneous pore nucleation:

NkT ΔGa þ ΔGhetero
I¼ exp ; (3.1)
h kT
where N is the number of atoms in the liquid, k is Boltzmann’s constant, T is the

temperature, h is Planck’s constant, and ΔGa is the activation energy for motion of
an individual atom. Here, ΔGa is negligible compared to ΔGhetero , and one can
obtain the expression

NkT 1 16π γ 3
I¼ exp f ðθc Þ ; (3.2)
h kT 3 ΔP2
where γ is the surface energy of the pore, ΔP is the difference between the ambient
and the internal pressure of the pore, and f ðθc Þ a function of the surface energy that
depends on the contact angle θc between the solid and the pore. The pore nucleation
rate is closely related to ΔP, and ΔP is proportional to the undercooling ΔT through
the Clausius–Clapeyron equation [27]:
ΔP / ΔT: (3.3)
In addition, it is generally reported that the solidification velocity v affects the

undercooling ΔT during the solidification process as follows:
v / ΔT n ð1 n 2Þ; (3.4)
where n is an empirical constant. Thus, the following expression can be finally

obtained from Eqs. (3.3) and (3.4):
v / ΔPn : (3.5)
Thus, the pore nucleation rate in Eq. (3.2) increases with increasing transfer
velocity because of the relationship between v and ΔP in Eq. (3.5); the solidification
velocity v is assumed to be equal to the transfer velocity.
On the other hand, since the hydrogen content in the melts is almost constant at
the melting point under a constant hydrogen gas pressure, the hydrogen content
diffused in each pore during the solidification decreases with an increase in the pore
nucleation rate. Thus, it is considered that the pore diameter decreases with a
decrease in each pore volume. This is in good agreement with the results of pore
length and pore aspect ratio; the changing in the pore length with the transfer
velocity is similar to the change in the pore diameter, and therefore, the pore aspect
ratio is not changed very much. This indicates that the pore dimensions such as
diameter and length are determined by the relationship between the pore nucleation
rate I and hydrogen solubility in the melt.
The pore grows in the direction perpendicular to the solid–liquid interface. Thus,
the pore growth direction depends on the shape of the interface during solidifica-
tion. The shape is determined by the flow of the heat emitted from the liquid during
solidification. Assuming that the heat flow rate is constant in a unit area and the
amount of extracted heat during solidification increases with increasing transfer
velocity during unit time, it is considered that the heat extracted from the liquid at
lower velocities is sufficient even at a flat interface. However, if the amount of the
heat increases by an increase in transfer velocity, then a large interface area may be
required at higher velocities in order to emit the additional heat. Therefore, it is
thought that the interface shape changes from flat at lower velocities to concave at
higher velocities. Because the depth at the center in the concave shape is propor-
tional to the transfer velocity, the depth increases with increasing velocity. Thus,
the pore growth angle shows a change such as that displayed in Fig. 3.19.
It is considered that the thickness of the skin layer is related not only to the pore
growth angle but also to the distance between the pore and the surface of the slab;
the thickness of the skin layer changes with a relationship of sine function between
the angle and the distance as shown Fig. 3.20. The distance decreases with increas-
ing transfer velocity as shown in Figs. 3.20 and 3.22. Assuming that the hydrogen
rejected from the solid was only transported away by diffusion when solidification
began, it is considered that the concentration profile of the hydrogen just before the
pore formation can be expressed as shown in Fig. 3.23. There will be a buildup of
the hydrogen ahead of the solid, which is accelerated by increasing solidification
velocity since it is more difficult to diffuse the hydrogen from the solid to the liquid
by increasing solidification velocity. If a critical hydrogen concentration Cpore
H for
the pore formation exists, then the pore forms and grows when the hydrogen
concentration ahead of the solid reaches Cpore H by hydrogen buildup. When the
pore is formed, the solidified distance xc is expressed as follows [28, 29]:
D
xc ¼ ; (3.6)
k0 v
where D is the hydrogen diffusivity in the liquid copper at the melting point and k0
is the equilibrium distribution coefficient; k0 is 0.31, which is evaluated as the
proportion of the hydrogen solubility in the solid and liquid copper at 1,357 K under
PH2 ¼ 0:1 MPa by using an equation suggested by Fromm and Gebhardt
[30]. Assuming that the solidification velocity is equal to the transfer velocity, the
Fig. 3.22 Distance between the pore and the surface plotted against the transfer velocity of lotus
copper fabricated under a hydrogen pressure of 1.0 MPa (Reprinted with permission from [24].
Fig. 3.23 Schematic of the hydrogen concentration profile from the beginning of solidification to
pore formation in (a) lower solidification velocity v1 and (b) higher solidification velocity v2 (x:
solidified distance, ClH : hydrogen concentration in liquid, CsH : hydrogen concentration in solid,
Cpore
H : critical hydrogen concentration for pore nucleation, and xc : the solidified distance until the
pore is formed) (Reprinted with permission from [24]. © 2007 Acta Materialia Inc.)
distance described in Fig. 3.22 can be plotted as a function of v1, as shown in

Fig. 3.24, where the distance is in inverse proportion to the transfer velocity as
shown in Eq. (3.6). The D value calculated from the slope of the dashed line is
2.28 107 m2 s1, which is close to the value of 0.99 107 m2 s1 at 1,374 K
reported by Wright and Hocking [31].
Fig. 3.24 Distance from the surface to the pore plotted as a function of reciprocal transfer velocity
v1 in lotus copper fabricated under a hydrogen gas pressure of 1.0 MPa (Error bars are the
standard deviation, and the dotted line is the fitted line.) (Reprinted with permission from [24].
Consequently, the thickness of the skin layer decreases with increasing transfer
velocity, which is affected by a change not only in the pore growth but also in the
distance between the pore and the surface of the slab.
3.3 Thermal Decomposition Method (TDM)
As already mentioned above, the high-pressure gas method is a convenient method

to supply the gas source in the molten metals for evolving gas pores in the solidified
metals. However, the PGM possesses two disadvantages: (1) necessity of strongly
built high-pressure chamber and application of high-pressure regulation and
(2) necessity of safety procedure to protect against hydrogen as an inflammable
and explosive gas. Mass production and commercialization in industries are
hesitated from such disadvantages. Therefore, a technique that does not require
high-pressure hydrogen to fabricate lotus metals is highly desirable. In order to
overcome these difficulties, Nakajima group [32] proposed an alternative, but
simple, method to fabricate such lotus metals by using a thermal decomposition
method (TDM) of compounds containing gas elements in a non-hydrogen atmo-
sphere under nearly atmospheric pressure.
The compounds suitable for TDM can be listed up to the following [33]:
TiH2, MgH2, ZrH2, Fe4N, TiN, Mn4N, CrN, Mo2N, Ca(OH)2, Cu2O, B2O3, CaCO3,
SrCO3, MgCO3, BaCO3, and NaHCO3.
3.3 Thermal Decomposition Method (TDM) 37
The compounds should be decomposed into gas element and another type of
metallic compound at the temperature (dissolving temperature) just below the
melting point of the solvent metal to be solidified. If the dissolving temperature
of the compounds is higher than the melting point of the metal, sufficient dissolu-
tion of the gas element into the melt cannot be ensured. The principle of TDM and
the first versatile method to control the pore morphology, including pore size and
porosity of lotus metals are described.
3.3.1 Mold Casting Technique through TDM
3.3.1.1 Fabrication of Lotus Copper through TDM
Copper (99.99 % pure) was melted by radio-frequency induction heating in a

graphite crucible under an argon atmosphere from 0.1 to 0.5 MPa. The melt was
poured into the mold, which had a copper bottom plate cooled by water and lateral
side walls made of 0.1-mm-thick cylindrical stainless steel. A few pellets of
titanium hydride, which ranged between 0.075 and 0.25 g in mass for melting
copper of 200 g, were set on the bottom plate of the mold. Unidirectional solidifi-
cation occurred in the mold so that a lotus copper ingot was obtained, as illustrated
in Fig. 3.25a. The ingot size was 28 mm in diameter and had a height of maximum
90 mm. The ingot was cut by a spark-erosion wire-cutting machine parallel and
perpendicular to the solidification direction.
Another method was also used to supply the hydrogen source in the melt. As
shown in Fig. 3.25b, a few pellets of titanium hydride were set in the crucible along
with copper. During the copper melt, hydrogen, which decomposed from titanium
hydride, was dissolved into the melt. The molten copper was poured from the
crucible into the mold without pellets of titanium hydride for unidirectional solidi-
fication. The ingots were investigated using the same method described previously.
Figure 3.26 shows the optical micrographs of the cross-sectional views of lotus
copper parallel and perpendicular to the solidification direction [32]. The pore
growth direction is coincident to the direction of the unidirectional solidification,
which is in agreement with that of the conventional high-pressure gas method
(PGM). Therefore, the principal mechanism of this pore evolution is suggested to
be similar to that of PGM; when the molten metal dissolving a gas is solidified,
insoluble gas precipitates and evolves the gas pores in the solid metals.
Figure 3.27 shows dependence of the porosity and the average pore diameter on
the mass of titanium hydride for the lotus copper fabricated by TDM in a 0.1 MPa
argon atmosphere. The porosity abruptly increases by the addition of 0.10 g of
titanium hydride and becomes constant to be about 55 % in the porosity on further
addition. When 0.10 g of titanium hydride was added to 200 g of molten copper, the
concentration of hydrogen in the melt was evaluated as 0.128 at. pct. Directly
comparing the hydrogen concentration from the titanium hydride to the available
data on the solubility in molten copper just above the melting temperature (Tm),
a b
Graphite crucible Induction
coil
H2 gas
Molten Cu
TiH2
pellets
Mold
Pore
TiH2 pellets
Chiller
Fig. 3.25 Schematic drawings of the principle to fabricate lotus metals by a mold casting
technique: (a) pellets of titanium hydride are set in the mold and (b) pellets of titanium hydride
are set in the crucible (Reprinted with permission from [32]. © 2008 The Minerals, Metals &
Materials Society and ASM International)
Fig. 3.26 Optical micrographs of cross-sectional views of lotus copper (upper views) perpendic-
ular and (lower views) parallel to the solidification direction. Mass of titanium hydride added to the
mold is (a) 0.075 g, (b) 0.10 g, (c) 0.125 g, and (d) 0.25 g. Melting and subsequent solidification
were carried out in 0.1 MPa argon atmosphere (Reprinted with permission from [32]. © 2008
The Minerals, Metals & Materials Society and ASM International)
1000
Average pore diameter /µm

800
600
400
200
100 0
80
porosity (%)
60
40
20
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Mass of TiH2 /g
Fig. 3.27 Dependence of the porosity and the average pore diameter on the mass of titanium
hydride. Melting and subsequent solidification were carried out in 0.1 MPa argon atmosphere
(Reprinted with permission from [32]. © 2008 The Minerals, Metals & Materials Society and
ASM International)
1,083 K, is impossible, because the atmospheric hydrogen pressure of the TDM

process differs from that of the solubility measurement (0.1 MPa). However, the
concentration of dissolving hydrogen in the TDM process is comparable to the
solubility of hydrogen in the liquid phase of copper near Tm [30]. If more than 0.1 g
of titanium hydride is added to the melt, the supersaturated hydrogen may generate
gas bubbles, which are then liberated from the melt to the atmosphere. The
hydrogen dissolved in the melt evolves pores at the solid–liquid interface during
solidification. If less than 0.1 g of titanium hydride is added, then all hydrogen can
dissolve in the melt without bubbling so that some of the hydride forms pores in the
solid–liquid interface, but the porosity may be smaller than that when more than
0.1 g of the hydride is added. Such redistribution of hydrogen in the liquid phase to
the solid phase and the atmosphere is exhibited in Fig. 3.28a, b, depending on the
mass of titanium hydride. For comparison, Fig. 3.28c shows the PGM case. The
TDM process is based on a nonequilibrium state, while the PGM process is based
on an equilibrium state in the chamber where a constant high-pressure gas is
maintained. On the other hand as shown in Fig. 3.27, the average pore diameter is
constant and nearly independent of the mass of titanium hydride. It has been
previously confirmed that the pore size is affected by the supercooling rate from
the liquid to the solid, which can be controlled by the solidification velocity in the
PGM [19, 23, 24]. The same reasoning can be used to interpret the present data.
a b c
TDM TDM PGM
(supersaturated) (undersaturated)
H2 H2
H H Melt
TiH2 H
pellets
Mold Porous
H H metal
Pore H2 H2 H2 H
Chiller
Fig. 3.28 Schematic drawings of the redistribution of hydrogen dissolved in the melt to the solid
phase and atmosphere. (a) Thermal decomposition method: the melt contains supersaturated
hydrogen. (b) Thermal decomposition method: the melt contains undersaturated hydrogen. (c)
High-pressure gas method. In all cases, hydrogen dissolved in the melt near the solid–liquid
interface moves directly into the pores, but some of the hydrogen moves in the solidified metal,
while insoluble hydrogen diffuses into the pores via the solid or diffuses back to the melt
ASM International)
Next, the atmospheric pressure effect was investigated. For this study, the mass
of titanium hydride was constant at 0.25 g, and argon was selected as the atmo-
spheric gas, which served as the external pressure. The argon pressure was varied
from 0.1 to 0.5 MPa. Figure 3.29 shows sectional views of lotus copper parallel and
perpendicular to the solidification direction as a function of argon pressure. The
effect of external pressure is obvious, and the pore growth is suppressed at higher
pressure. Figure 3.30 shows the dependence of the porosity and the average pore
diameter as a function of argon pressure. Both the porosity and the average pore
diameter decrease with increasing argon pressure. The pore volume v, which is
equal to the porosity, is inversely proportional to the external argon pressure P,
which can be described by the Boyle–Charles law, v ¼ nRT/P, where n, R, and T
are the hydrogen molar number, the gas constant, and the temperature, respectively.
Therefore, the pore diameter can be written as d/ P1/3. The tendency of the
pressure dependence of the porosity and the pore diameter are explained by the law.
However, it seems that such a pressure effect may be more significant than that
predicted by the Boyle–Charles law. This difference may be attributed to the
possibility that the molar number of hydrogen is not constant under changing
pressure. The decomposition rate of titanium hydride should be retarded as the
external pressure of argon increases. Thus, the porosity and the pore diameter may
decrease remarkably.
Finally, we examined the role of the other metallic element, Ti, through the
thermal decomposition of titanium hydride. Titanium hydride is decomposed into
hydrogen and titanium. The latter is a very reactive element, which easily reacts
with residual oxygen in the molten copper. Consequently, titanium oxide particles
are formed and dispersed, which may serve as the nucleation sites for the hydrogen
Fig. 3.29 Cross-sectional views of lotus copper fabricated under different argon pressure.
(a) 0.1 MPa, (b) 0.25 MPa, and (c) 0.5 MPa. Upper and lower views are the cross section
perpendicular and parallel to the solidification direction, respectively. Titanium hydride (0.25 g)
was added to the mold during solidification (Reprinted with permission from [32]. © 2008
1000
Average pore diameter /µm
800
600
400
200
60 0
Porosity (%)
40
20
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Argon pressure / MPa
Fig. 3.30 Dependence of the porosity and the average pore diameter on argon pressure. Titanium
hydride (0.25 g) was added to the mold during solidification (Reprinted with permission from
[32]. © 2008 The Minerals, Metals & Materials Society and ASM International)
pores in the solid–liquid interface during unidirectional solidification. It is well

known that the pores evolve by heterogeneous nucleation in metal melts in the
presence of small amounts of foreign particles [34]. Therefore, the same reasoning
can be applied to the present case, and it is surmised that the pore size and porosity
become homogeneous by uniformly distributed nucleation sites. Thus, TDM
exhibits another advantage to produce lotus metals with more homogeneous
pore size and porosity than those by PGM, which does not provide intentional
nucleation sites.
3.3.1.2 Fabrication of Lotus Aluminum through TDM
The investigation was undertaken to fabricate lotus aluminum with the porosity
more than 10 % through TDM [35]. For this purpose, three types of the compounds
containing hydrogen were used: calcium hydroxide, sodium bicarbonate, and
titanium hydride. The pore morphology is usually characterized with the pore
size and porosity, which were controlled by the temperature of melt, amount of
compounds, and external argon pressure in this work.
Pure aluminum (99.99 % pure) of about 100 g was melted in a graphite crucible
by an induction heating coil in vacuum. The bottom of the mold was copper plate
cooled by a circulated water chiller, while the side was made of stainless steel. 0.2 g
of the compounds such as calcium hydroxide, sodium bicarbonate, or titanium
hydride was usually wrapped with aluminum foil and was set on the bottom plate
of the mold. The temperature of the liquid in the graphite crucible was monitored by
an infrared pyrometer. After pouring the molten aluminum at 1,023 K in the
crucible into the mold, hydrogen decomposed from the compounds dissolves in
the melt, which is then solidified so that insoluble hydrogen precipitates to evolve
the hydrogen pores. In order to investigate the dependence of the pore size and
porosity on the mass of calcium hydroxide, the amount of compounds was changed
from 0.1 to 1.0 g. In order to investigate the external pressure dependence of the
pore size and porosity, the argon pressure was changed from vacuum to 0.04 MPa.
When the solidification took place unidirectionally, lotus aluminum with direc-
tional cylindrical pores was fabricated. The size of the obtained ingot was about
25 mm in diameter and about 60 mm in height. The specimens were cut using a
spark-erosion wire-cutting machine in both directions parallel and perpendicular to
the solidification direction. Each cross section was polished with a series of emery
papers and was observed using an optical microscope.
The constituent elements of pores were investigated using gas analyzers. After
the specimen kept into salt-saturated distilled water was cut, the gas bubbles were
collected into a vial in the distilled water. The collected gas was analyzed by a gas
chromatograph (GC-14A, Shimadzu Co., Kyoto, Japan) in order to identify hydro-
gen, carbon monoxide, and carbon dioxide. Oxygen was analyzed by a gas analyzer
(TC-300, LECO Co., St. Joseph, MI).
Fig. 3.31 Optical micrographs of lotus aluminum fabricated by mold casting technique with
different compounds in vacuum at 1,023 K. The upper and lower micrographs are the cross
sections perpendicular and parallel to the solidification direction, respectively (Reprinted with
permission from [35]. © 2009 The Minerals, Metals & Materials Society and ASM International)
Figure 3.31 shows the microstructure of lotus aluminum on cross sections

parallel and perpendicular to the solidification direction. The pore morphologies
perpendicular to the solidification direction are shown on the upper photos, while
those parallel to the solidification direction are shown in the lower photos. In all
specimens, the aligned pores parallel to the solidification direction were observed.
It is considered that the cylindrical gas pores are evolved by TDM from the
compounds during solidification of aluminum containing hydrogen gas.
Table 3.2 compiled the decomposition reactions of three compounds in the
aluminum melt. Calcium hydroxide and sodium bicarbonate decompose into
vapor and compounds and then, the vapor decomposes to hydrogen and metallic
oxide. On the other hand, titanium hydride decomposes into hydrogen and titanium.
These decomposed gas elements can dissolve in aluminum melt. When the melt is
solidified, lotus aluminum can be produced.
Figure 3.32 shows the pore size and porosity of lotus aluminum with different
compounds using the mold casting technique in vacuum. It is seen that the pores
with average size 400, 268, and 1,086 μm are evolved by using calcium hydroxide,
sodium bicarbonate, and titanium hydride, respectively. The porosity of lotus
aluminum was similar, being as much as 20 % regardless of the kinds of the
compounds. The pore size was small, as much as 400 μm, and its distribution was
homogeneous in porous aluminum through TDM of calcium hydroxide and sodium
bicarbonate, while the pore size was large as much as 1,000 μm and its distribution
is not uniform through the TDM of titanium hydride.
Table 3.2 Decomposition reaction and temperature of gas-forming compounds

Decomposition Gas atoms or molecules
Reactions temperature (K) to be dissolved
Ca(OH)2 ! CaO + H2O 853 H
H2O ! metallic oxide + 2H
2NaHCO3 ! Na2CO3 + H2O + CO2 473 H, CO, CO2, O
H2O ! metallic oxide + 2H
TiH2 ! Ti + 2H 723 H
Ti ! dissolved in liquid
Fig. 3.32 Pore size and porosity of lotus aluminum fabricated by mold casting technique using
compounds in vacuum (Reprinted with permission from [35]. © 2009 The Minerals, Metals &
Materials Society and ASM International)
Calcium hydroxide or sodium bicarbonate decomposes into compounds (CaO,

Na2CO3), carbon dioxide, and vapor, the latter of which furthermore decomposes
into metallic oxide and hydrogen. Small pore size with homogeneous distribution
was observed through these reactions. On the other hand, titanium hydride
decomposes directly into titanium and hydrogen. Large pores with inhomogeneous
distribution were observed. It is not clear at present, but evolution of small pores
with homogeneous distribution may be attributed to the existence of the oxide
particles, which may serve as the nucleation sites of the pores. In order to investi-
gate the content of pore gas, the gas analysis was carried out. Figure 3.33 shows the
results of gas analysis of lotus aluminum using different compounds. It can be seen
that hydrogen was the pore-formation source in the case of titanium hydride,
calcium hydroxide, and sodium bicarbonate. Although the absolute values of
concentration of each gas element cannot be compared in the analysis, it is apparent
that the major gas is hydrogen in three kinds of the specimens. The contents of
Fig. 3.33 Results of gas analysis in the pore in lotus aluminum. (a) Hydrogen, (b) carbon dioxide,
and (c) oxygen (Reprinted with permission from [35]. © 2009 The Minerals, Metals & Materials
Society and ASM International)
Fig. 3.34 Lotus aluminum fabricated using calcium hydroxide in vacuum or under argon pressure
(0.01–0.04 MPa). The amount of calcium hydroxide was kept to be 0.2 g. (a) Pore morphology.
The upper and lower micrographs are the cross sections perpendicular and parallel to the solidifi-
cation direction, respectively. (b) Variation of the porosity and pore size as a function of argon
pressure (Reprinted with permission from [35]. © 2009 The Minerals, Metals & Materials Society
and ASM International)
oxygen and carbon dioxide are very small; some amount of them may be contained
in the collection handling procedure of the pores.
Figure 3.34a shows the pore structures on cross sections of lotus aluminum
perpendicular (upper) and parallel (lower) to the solidification direction. The lotus
aluminum was fabricated using calcium hydroxide in vacuum and under argon
atmosphere. The aligned pores formed under the pressure less than 0.03 MPa Ar;
however, spherical pores formed over 0.04 MPa argon. This may indicate that the
applied pressure in the chamber reduces the driving force for pore nucleation and
growth. Figure 3.34b shows the dependence of the porosity and the average pore size
as a function of argon pressure. The effect of external pressure is obvious, and
the pore growth is suppressed under higher pressure. Both the porosity and the
average pore size decrease with increasing argon pressure. The pore volume v,
which is equal to the porosity, is inversely proportional to the external argon

pressure P, which can be described by the Boyle law, v ¼ nRT/P, where n, R, and
T are the hydrogen molar number, the gas constant, and the temperature, respec-
tively. Therefore, the pore diameter can be written as d / P1/3. The tendency of the
pressure dependence of the porosity and pore size are explained by the law.
3.3.2 Continuous Zone Melting Technique through TDM
Recently, Makaya and Fredriksson produced porous Fe-based materials by

dissolving CrN compound into a metallic melt of Fe-based alloy in an argon
atmosphere [36]. The decomposition of the compound leads to dissolution of a
gas into the melt. They solidified the melt in the crucible to produce porous metals
with an isotropic pore structure. However, no investigations were carried out to
produce porous metals with an anisotropic cylindrical pore structure until Nakajima
and Ide performed theirs. Nakajima and Ide studied fabrication of lotus copper
using titanium hydride in an argon atmosphere instead of a pressurized hydrogen
atmosphere, which is called as TDM [32]. In order to fabricate the lotus metals,
unidirectional solidification was usually carried out using the mold casting tech-
nique, as mentioned already. However, this technique cannot be applied to the
metals and alloys with low- thermal conductivity. The solidification rate changes
from the bottom to the top: the cooling speed near the bottom is faster, while it
becomes slower near the top where the pore is coarsened because the heat cannot be
dissipated. Thus, the pore size significantly changes with the height; the pore size
increases with increasing height of the mold. In order to solve this problem, Ikeda
et al. developed a continuous zone melting technique [23]. The partial melting zone
is moved to retain the constant solidification velocity so that uniform distribution of
porosity and pore size are obtained. This technique was adopted to fabricate lotus
iron by Wada et al. [37].
Iron rods of 10.0 mm in diameter and 80 mm in length were used as specimens.
In order to supply a constant source of chromium nitride, a hole of 2.0 mm in
diameter in a central part of the cross section was made by an electric drill as shown
in Fig. 3.35. The compound was filled into the hole. The chromium nitride was
selected as a gas dissolving compound, since the temperature of gas releasing
chromium nitride is the closest to the melting temperature of iron. The Cr1.18N
and Cr2N powders were used. The former was a mixture of 81.4 mass pct CrN and
18.6 mass pct Cr2N. 1.0 mass pct of Cr1.18N powders was filled into the permeable
straight hole of the iron rods. The continuous zone melting was carried out by
controlling the transfer velocity from 80 to 580 μm s1.
Figure 3.36 shows the cross-sectional views of lotus iron perpendicular and
parallel to the solidification direction as a function of the transfer velocity. No pores
were observed in the ingot fabricated with the velocity of 80 μm s1. Figure 3.37
exhibits the transfer velocity dependence of the porosity and average pore diameter.
The porosity increases with increasing transfer velocity, while the variation of the
Fig. 3.35 Outer view of iron

rod: (a) longitudinal
direction and (b) cross
section perpendicular to the
longitudinal direction
from [37]. © 2009 The
Minerals, Metals & Materials
Society and ASM
International)
Fig. 3.36 Cross-sectional views of lotus iron perpendicular (upper) and parallel (lower) to the
solidification direction. Transfer velocity: (a) 80 μm s1, (b) 250 μm s1, (c) 410 μm s1, and (d)
580 μm s1. Atmosphere is 0.5 MPa helium (Reprinted with permission from [37]. © 2009 The
Minerals, Metals & Materials Society and ASM International)
a b
50 1000
Average pore diameter / µm

40 800
porosity (%)
30 600
20 400
10 200
0 0
0 200 400 600 0 200 400 600
Transfer velocity / µm s-1 Transfer velocity / µm s-1
Fig. 3.37 Transfer velocity dependence of (a) the porosity and (b) average pore diameter of lotus
iron. Atmosphere is 0.5 MPa helium (Reprinted with permission from [37]. © 2009 The Minerals,
Metals & Materials Society and ASM International)
pore diameter with the velocity is relatively limited compared with previous results
[23, 24]. This suggests that the mechanism of pore formation obtained by this study
by TDM is different from that obtained by PGM.
In the mold casting, the gas compound reacts and simultaneously dissolves a gas
into the melt during the casting without preheating of the gas compound. However,
in the continuous zone melting, when the rod is solidified unidirectionally, it is
preheated by heat conduction before melting so that it has a temperature distribu-
tion. In general, the behavior of thermal decomposition of compounds is strongly
affected by the heating rate [38]. The possibility of forming pores by continuous
zone melting will be discussed, taking into consideration the transfer velocity
dependence of the temperature distribution in the iron rod and the thermal decom-
position of nitride. The relation between the temperature T of preheated rod and the
distance x from the solid–liquid interface under the transfer velocity v can be
expressed by applying an equation of redistribution of solute in material solidified
in the zone melting at constant velocity to a heat conduction [39].

d dT dT
λ þ ρcv ¼ 0; (3.7)
dx dx dx
where ρ, c, and λ are the density of iron, the specific heat of iron, and the thermal
conductivity of iron, respectively. Figure 3.38 shows the temperature and the
heating rate as a function of the distance from the interface between liquid and
solid at different transfer velocities. Both the temperature gradient and heating rate
increase with increasing transfer velocity. Therefore, it is considered that the
position in the rod where the gas is released from chromium nitride varies with
the transfer velocity.
a b
1800 Transfer velocity / µm s-1 6000
80 250
410 580
Heating rate / K min-1

Temperature / K
1600
4000
1400
2000
1200
0
0 10 20 30 40 50 0 10 20 30 40 50
Distance from melt / mm Distance from melt / mm
Fig. 3.38 Changes in (a) temperature and (b) heating rate as a function of the distance from the
interface between liquid and solid at different transfer velocities (Reprinted with permission from
[37]. © 2009 The Minerals, Metals & Materials Society and ASM International)
Derivative thermogravimetry (DTG) of chromium nitrides of Cr1.18N and Cr2N

powders was carried out using a thermal analyzer in an argon flow of 0.1 MPa in the
temperature range from room temperature to 1,773 K. The heating rate was changed
from 10, 20, 30, and 40 K min1. Figure 3.39 shows DTG curves with the heating
rate of 10 K min1 for Cr1.18N and Cr2N powders. Two peaks for gas release from
Cr1.18N were observed at 1,240 K(Tp1) and 1,570 K(Tp2), while one peak from
Cr2N was found at 1,680 K. Thus, the 1,570 K peak observed in higher temperature
is attributed to the gas release from Cr2N. Since Cr1.18N is composed of CrN and
Cr2N, the 1,240 K peak of lower temperature is due to the gas release from CrN.
According to the DTG measurement of Cr1.18N at different heating rate, the starting
temperature and peak temperature of the gas release are shifted to higher tempera-
ture with increasing heating rate.
According to Kissinger [40], the heating rate β is related to the peak temperature
Tp for gas release by the following equation:

d ln Tβ2 E
p ¼ a ; (3.8)
d Tp1 R
where Ea is an activation energy of reaction and R is the gas constant. The Kissinger
plots of the peak temperature were obtained from the measured DTG curves. It is
predicted from the plots that the heating rates with which the peak temperature for
the gas release is equal to the melting temperature (1809 K) of iron are 3.58104
K min1 (H.R1) for CrN and 5.70 102 K min1 (H.R2) for Cr2N. In order to have
decomposition of the nitrides into the melt, the transfer velocity has to be such that
the heating rate in the rod exceeds the predicted value, while there is an optimal
heating rate. Figure 3.40 shows the relation between the temperature of the rod and
Cr1.18N
Ts1 Ts2
DTG Tp2
Cr2N Tp1
0.5 % min-1
400 800 1200 1600

Temperature / K
Fig. 3.39 DTG curves measured by the DTG with the heating rate of 10 K min1 for Cr1.18 N and
Cr2N powders. The Ts and Tp are the starting temperature and peak temperature of the gas release,
respectively (Reprinted with permission from [37]. © 2009 The Minerals, Metals & Materials
Society and ASM International)
Transfer velocity / µm s-1 Melting

80 250 Ts2 zone
410 580
104
Heating rate / K min-1
103
102
1000 1200 1400 1600 1800

Temperature / K
Fig. 3.40 Relation between heating rate and temperature of the rod at different transfer velocities
ASM International)
heating rate, which is obtained from the results of Fig. 3.38a, b. Since the predicted
H.R1 is higher than the heating rates at all transfer velocity, gas from CrN contained
in Cr1.18N is released before melting and does not contribute to pore evolution. On
the other hand, H.R2 in higher temperature than the starting temperature (about
1,450 K slightly depending on the heating rates) for gas release is slower than the
heating rates in the transfer velocity more than 250 μm s1. Therefore, it is surmised
that no pore formation at lower transfer velocity of 80 μm s1 is attributed to
insufficient nitrogen release into the molten iron, because most of nitrogen gas
released from the nitride escapes to the atmosphere; the rod is heated at high
temperature by heat conduction from the melt part of a longer time. It is considered
that gas from Cr2N contained in Cr1.18N cannot be released until part of the rod
melts and is used effectively to evolve the pores. Thus, for transfer velocity such
that the decomposition peak temperature lies just above the melting point, decom-
position of the nitrides in the melt is optimal.
Different from PGM, a gas equilibrium between the atmosphere (nitrogen partial
pressure ~ 0) and the melt is not maintained in TDM [32]. During melting, nitrogen
escapes from the melt to the atmosphere. Since the holding time of the melting
condition decreases with increasing transfer velocity, the amount of escaped nitro-
gen may decrease; thus, the porosity increases with increasing transfer velocity, as
shown in Fig. 3.37a.
3.3.3 Continuous Casting Technique through TDM
As mentioned above, TDM is desirable to avoid inherent risk of employing high-

pressure hydrogen gas, and the continuous casting technique is suitable to produce
large-sized lotus metal slabs with controllable pore size and porosity. Combining
both techniques is the most suitable to fabricate lotus metals. Nakajima and Ide
[41] fabricated long-sized lotus copper slabs through TDM by continuous casting
technique. Figure 3.41a illustrates the continuous casting apparatus for TDM; for
comparison, the apparatus of continuous casting technique through PGM is shown
in Fig. 3.41b. The apparatus consists of three parts. (1) The first part is the melting
part; the copper were melted in the graphite crucible by the induction heating. The
upper cylindrical graphite bar was pushed down. (2) The overflowing copper melt
dropped down into a tundish, in which pellets of titanium hydride were supplied
through a nozzle at a constant time interval in order to maintain the dissolving
hydrogen constant. (3) The third part is the mold cooled by water and pinch rollers
which pulled the dummy bar and the solidified slab ingot horizontally. Figure 3.42
shows one example of lotus copper fabricated with addition of a pellet, 0.2 g, per
minute at the melt temperature of 1,573 K in argon atmosphere of 0.1 MPa with
different transfer velocity. The concentration of titanium hydride in the melt
added in the tundish was evaluated as 0.001mass%. Almost uniform pore distri-
bution in size and position was observed in the cross section parallel to the
Fig. 3.41 Apparatus of continuous casting technique to fabricate lotus metals through (a) TDM
and (b) PGM (Reprinted with permission from [41]. © 2012 Deutsche Gesellschaft fur
Materialskunde e.V.)
Transfer
velocity 20 40 60 80 100
mm/min
section
// section
10 mm
Fig. 3.42 Lotus copper fabricated by continuous casting technique through TDM. Upper and
lower are the cross sections perpendicular and parallel to the solidification direction, respectively
a Flat b Curved
interface interface
Solidification Liquid
Interface
Solid
Pores grow Tilted pores
upward collapse and
coarsen
Fig. 3.43 Relation between alignment of evolving pores and the shape of liquid–solid interface.
(a) The alignment of the evolving pores is unidirectional. (b) The alignment is tilted. The tilted
pores collapse to coarse pores in the central part
solidification direction, but not uniform in the section perpendicular. This is

because tilted pores collapsed to cause coarse pores in the central part, as shown
in Fig. 3.43b. The alignment of the evolving pores should be unidirectional as
shown in Fig. 3.43a. If a flat liquid–solid interface can be obtained by controlling
the chiller temperature in the hearth, this problem will be solved.
3.4 Moisture Decomposition Method
3.4.1 Fabrication of Lotus Nickel by Moisture
Lotus nickel was successfully fabricated by unidirectional solidification using

molybdenum-sheet mold coated by Al2O3–Na2SiO3 in an argon atmosphere. Dur-
ing solidification, the moisture absorbed in Al2O3–Na2SiO3 at atmospheric temper-
ature is dissolved into the melt and then the gas pores are formed at the solid–liquid
interface. Nickel in a magnesia crucible was melted by induction heating at a given
pressure of argon gas in the high-pressure chamber. As shown in Fig. 3.44, the
molten nickel was poured into the mold whose bottom was water-cooled by a
chiller from the crucible. The mixture paste of alumina, sodium silicate, and
water was coated on the inside lateral surface on the molybdenum cylindrical
sheet of the mold in order to suppress a reaction between molybdenum sheet
and the molten nickel as insulated ceramics and also to maintain the moisture.
Figure 3.45 shows the optical micrographs of lotus nickel in parallel section to the
solidification direction. The moisture contents in the mold weigh (a) 0.0596 g,
(b) 0.0876 g, (c) 0.1070 g, and (d) 0.1201 g. In the ingot with the smallest amount of
3.4 Moisture Decomposition Method 55
Fig. 3.44 Schematic

Ni(melt)
drawing of the process of
Ar atmosphere
mold casting technique for
fabrication of lotus nickel
with moisture in argon
atmosphere (Reprinted with
permission from [42].
© 2004 Japan Institute Ni(melt)
of Metals)
Mo sheet
Al2O3 powder
+ Na2SiO3
+ H2O
Mold
Chiller Pore
Solidified Ni
moisture, almost all pores are spherical and long-cylindrical pores were not
observed. The number of elongated pores increases with increasing moisture.
Such longer pore evolution may be attributed to increase in hydrogen content
which serves as a source for pore growth.
Figure 3.46 shows the magnified optical micrographs of cross-sectional planes
(a) parallel and (b) perpendicular to the solidification direction at 4.5 mm in height
from the bottom of the ingot (c) in Fig. 3.45, whose porosity is 44.7 % and the
average pore size is 105 μm, ranging from 5 to 200 μm in diameter. On the other
hand, for example, the porosity and pore size are 49.8 % and 499 μm, respectively,
in lotus nickel fabricated using the dried mold without moisture under the pressure
of 0.6 MPa argon and 0.2 MPa hydrogen. Thus, the porosity in lotus nickel
fabricated in the mold with moisture in argon is as high as that in lotus nickel
without moisture in hydrogen, and moreover, the pores can be more minute than
that using hydrogen.
The pore evolution model by moisture was proposed by Suematsu
et al. [42]. Figure 3.47 illustrates schematic of pore evolution process before
solidification, after pouring the melt and during solidification. The moisture is
absorbed in the ceramic coat on the lateral side of the mold, and then, when the
molten nickel is poured into the mold, the moisture is dissolved into hydrogen and
oxygen as a result of chemical reaction between the molten metal and moisture:
n H2 OðgasÞ þ m M $ Mm On þ 2n H; (3.9)
where M is a metallic element to be oxidized easily. Then, the hydrogen is

absorbed in the molten nickel and insoluble hydrogen when nickel is solidified is
Fig. 3.45 Optical micrographs of lotus nickel in parallel section to the solidification
direction (argon pressure, 0.3 MPa); moisture contents in mold are (a) 0.0596 g, (b) 0.0876 g,
(c) 0.1070 g, and (d) 0.1201 g (Reprinted with permission from [42]. © 2004 Japan Institute of
Metals)
able to form pores, whose principle is the same as that in fabrication of lotus
metals. On the other hand, oxygen forms oxides with, for example, nickel which
may become to be heterogeneous nucleation sites. It is well known that the pore
nucleation takes place by heterogeneous nucleation mechanism. Thus, dissolution
of oxygen in the melt increases the number of the pore nucleation sites. Even if the
same amount of hydrogen dissolves in the molten nickel, a number of more minute
elongated pores can be evolved using the mold with sufficient moisture in compar-
ison with that using the dry mold in hydrogen atmosphere. Therefore, lotus-type
porous nickel with minute elongated gas pores can be successfully produced by
using the mold with moisture, which is far easy and simple technique to fabricate
lotus metals.
Fig. 3.46 Optical micrographs of cross sections of lotus nickel with 44.7 % porosity (argon
pressure, 0.3 MPa). The cross-sectional plane (a) parallel and (b) perpendicular to solidification
direction at 4.5 mm from the bottom plane in the ingot (Reprinted with permission from [42].
© 2004 Japan Institute of Metals)
Fig. 3.47 Schematic drawings of pore-forming process. (a) Mold assembly before solidification,
(b) molten nickel in the mold just after pouring of the melt, and (c) process of unidirectional
solidification (Reprinted with permission from [42]. © 2004 Japan Institute of Metals)
3.4.2 Fabrication of Lotus Cobalt and Silicon by Moisture
Previous study as described in Sect. 3.4.1 suggests that formation of pores strongly
depends on the dissociation constant of water at the casting temperature
[42]. Because the free energy changes of Eq. (3.9) are known to be positive values,
the dissociation constant PH2 =PH2 O is very small [43, 44]. It is worthwhile to
investigate in detail whether the feasibility of pore formation on other materials is
also closely related with the magnitude of the dissociation constant of water in the
moist atmosphere. The feasibility to fabricate lotus silicon, cobalt, and copper was
investigated by unidirectional solidification using moisture contained in the
mold [45].
The metal in the crucible was melted by an induction coil and the melted metal
was poured into the mold. The bottom of the mold was cooled by a water chiller,
while the lateral wall was made of molybdenum whose surface was coated with a
mixture of alumina, water glass (54.5 % H2O, 31.2 % SiO2, and 14.3 % Na2O), and
water with a weight ratio of 1:1.5:1. The mold was dried in an oven at 423 K for
7.2 ks. Then, the mold was kept in a closed chamber with constant humidity (90 %),
which was measured by a hydrometer attached to the chamber for a given time, and
0.15 g of water, which corresponds to 0.083 mol of H2O, was reabsorbed into the
coating material. This value is sufficient to produce pores. About 2 mol of silicon,
cobalt, and copper ingots with 99.9 % purity were used for casting. The size of the
solidified ingots was 28 mm in diameter and 50 mm in length.
Figure 3.48 shows the cross-sectional views of solidified ingots of Si, Co, and Cu
perpendicular and parallel to the solidification direction under a given pressure of
the argon atmosphere. The pores elongated along the solidification direction were
observed for Si and Co, while no pores were observed for Cu. Moreover, lotus Si
and Co were also fabricated in the hydrogen atmosphere in the mold without
moisture. The pore size in the lotus Si and Co fabricated in hydrogen gas is much
larger than that by moisture dissociation, as shown in Fig. 3.49. When the moisture
contained in the oxides mixture coated on the molybdenum mold was completely
removed by drying, only nonporous cobalt ingot was obtained so that the pore
evolution is attributed to hydrogen decomposed from the moisture in the mold.
According to the previous investigation [10], lotus copper was easily produced by
unidirectional solidification in pressurized hydrogen atmosphere. However, lotus
copper was not produced by the moisture in the mold. In order to elucidate the
reason why the moisture can produce lotus cobalt and silicon, not lotus copper,
the ratio of pressure of hydrogen decomposed to that of the moisture was evaluated
to the following reactions:
Co þ H2 O ¼ CoO þ H2 ; (3.10)
Si þ 2H2 O ¼ SiO2 þ 2H2 : (3.11)
They assume that the moisture on the mold decomposes to hydrogen and oxygen
atoms at the initial stage of casting. According to the Ellingham diagram, the ratio
of hydrogen to moisture, PH2 =PH2 O , can be derived from the Gibbs free energy
changes [43, 44]. Figure 3.50 shows the formation free energy changes of water and
nickel, silicon, cobalt, and copper oxides. Porous nickel was fabricated by the mold
casting method ratio of PH2 =PH2 O , at the casting temperature was approximately
1.18 102, as calculated using the following equation:

PH2 ΔG
¼ exp ; (3.12)
PH 2 O 2RT
where ΔG, R, and T denote the Gibbs free energy change, the gas constant, and the
casting temperature, respectively. The ratio of PH2 =PH2 O , for the formation of
Fig. 3.48 (a) Transverse cross-sectional and (b) longitudinal cross-sectional views of lotus
silicon, cobalt, and copper fabricated under Ar pressure. Argon pressure for each fabrication is
0.4 MPa for Si, 0.8 MPa for Co, and 0.4 MPa for Cu (Reprinted with permission from [45]. © 2008
silicon, cobalt, and copper oxides, is estimated as 2.15 104 at 1,773 K,

6.93 102 at 1,873 K, and 4.91 104 at 1,673 K, respectively [43, 44], as
shown in Table 3.3. Assuming that the initial stage of casting is in the equilibrium
state, the amount of moisture decomposed into hydrogen and oxygen atoms and the
amount of hydrogen gas dissolved into the molten metal can be evaluated by the
Fig. 3.49 (a) Longitudinal cross-sectional and (b) transverse cross-sectional views of lotus silicon
fabricated under hydrogen pressure of 0.4 MPa and cobalt fabricated under hydrogen pressure
of 0.15 MPa and argon pressure of 0.65 MPa (Reprinted with permission from [45]. © 2008
ratio of PH2 =PH2 O and the Sieverts law. The formation mechanism of the moisture
decomposition into the molten metal is essentially the same as that of the method
using pressurized hydrogen gas. These results indicate that the pores can be
evolved, even if the value of PH2 =PH2 O is very small. Since no hydrogen gas was
introduced in the chamber, the hydrogen partial pressure in the atmosphere is
considered as 0 MPa. Hence, the hydrogen gas used in the formation of pores is
formed by decomposition of the moisture on the mold during solidification.
The pore diameter for the lotus silicon and cobalt fabricated in the moisture
method is much smaller than that of those fabricated under pressurized hydrogen
atmosphere. It is well known that heterogeneous nucleation takes place in metal
melts in the presence of small amounts of foreign elements [34]. Then, it is
reasonable to consider that oxygen decomposed from the moisture can result with
a metallic element to form oxide clusters, which are nuclei for pore evolution as
illustrated in Fig. 3.51. Such dispersive oxide clusters may promote uniform small
Fig. 3.50 Temperature dependence of the change of the free energy for formation of water, silica,
cobalt oxide, nickel oxide, and copper oxide [43, 44] (Reprinted with permission from [45].
© 2008 The Minerals, Metals & Materials Society and ASM International)
Table 3.3 Free energy changes of dissociation of water for Si, Co, Cu, and Ni
R reaction ΔG (kJ mol1) Casting temperature (K) PH2 =PH2 O
Si + 2H2O ¼ SiO2 +2H2 298.5 1,773 2.15 104
Co + H2O ¼ CoO + H2 41.56 1,873 6.93 102
2Cu + H2O ¼ Cu2O + H2 53.41 1,673 4.91 104
Ni + H2O ¼ NiO + H2 69.13 – 1.18 102
Fig. 3.51 Schematic growing of the movement of oxide during fabrication and pore generation
ASM International)
pore distribution. Although quantitative comparison of the moisture decomposition

method with the pressurized hydrogen method is not possible, there is a tendency
that the moisture method can produce smaller-sized pores and a large number of
density of pores. This tendency can be explained in terms of the oxide clusters for
possible nucleation sites.
References
1. Chalmers B (1959) Sci Am 200:114–122

2. Imabayashi M, Ichimura M, Kanno Y (1983) Trans JIM 24:93–100
3. Svensson I, Fredriksson HS (1980). In: Proceedings of international conference organized by
the applied metallurgy and metals tech group of TMS, University of Warwick, pp. 376–380
4. Knacke O, Probst H, Wernekinck J (1979) Z Metallkde 70:1–6
5. Bioko LV, Shapovalov VL, Chernykh EA (1991) Metallurgiya 346:78–81
6. Hyun SK, Shiota Y, Murakami K, Nakajima H (1999) In: Koiwa M, Otsuka K, Miyazaki T (eds)
Proceedings of international conference on solid-solid phase transformations ’99 (JIMIC-3),
Japan Inst Metals, Kyoto, pp. 341–344
7. Nakajima H, Hyun SK, Ohashi K, Ota K, Murakami K (2001) Colloids Surf A:Physicochem
Eng Aspects 197:209–214
8. Nakajima H (2001) Mater Trans 42:1827–1829
9. Shapovalov VI (1994) MRS Bull XIX:24–28
10. Hyun SK, Murakami K, Nakajima H (2001) Mater Sci Eng A 299:241–248
11. Smith DP (1947) Hydrogen in metals. The University of Chicago Press, Chicago, p 34
12. Reed-Hill RE (1964) Physical metallurgy principles. D.Van Nostrand Company Inc,
Princeton, p 393
13. Massalski TB (1986) Binary alloy phase diagram. American Society for Metals, Metals Park,
p 1079
14. Levinsky Y (1997) Pressure dependent phase diagrams of binary alloys. ASM International,
Materials Park, p.693
15. Satir-Kolorz AH, Feichtinger HK (1991) Z Metallkde 82:689–697
16. Hyun SK, Nakajima H (2002) Mater Trans 43:526–531
17. Nakahata T, Nakajima H (2005) Mater Trans 46:587–592
18. Ikeda T, Nakajima H (2002) Japan Foundry Eng Soc 74:812–816
19. Hyun SK, Nakajima H (2003) Mater Lett 57:3149–3154
20. Ikeda T, Tsukamoto M, Nakajima H (2002) Mater Trans 43:2678–2684
21. Nakajima H, Ikeda T, Hyun SK (2003) In: Banhart J, Fleck A (eds) Cellular metals: manufac-
ture, properties, applications. MIT, Berlin, pp 191–202
22. Nakajima H, Ikeda T, Hyun CK (2004) Adv Eng Mater 6:377–384
23. Ikeda T, Aoki T, Nakajima H (2005) Metall Mater Trans A 36A:77–86
24. Park JS, Hyun SK, Suzuki S, Nakajima H (2007) Acta Mater 55:5646–5654
25. Campbell J (1991) Castings. Butterworth-Heinemann, Oxford
26. Fisher JC (1948) J Appl Phys 19:1062–1067
27. Flemings MC (1974) Solidification processing. McGraw-Hill, New York
28. Chalmers B (1964) Principles of solidification. Wiley, New York
29. Porter DA, Easterling KE (1992) Phase transformations in metals and alloys. Chapman & Hall,
London
30. Fromm E, Gebhardt E (1976) Gases and carbon in metals. Springer, Berlin
31. Wright JH, Hocking MG (1972) Metall Trans 3:1749–1753
32. Nakajima H, Ide T (2008) Metall Mater Trans A 39A:390–394
References 63
33. Nakajima H, Ide T (2007) Method for manufacturing porous body. PCT/JP2007/062769
(patent pending)
34. Fredriksson H, Akerlind U (2006) Materials processing during casting. Wiley, Chichester,
pp 141–142
35. Kim SY, Park JS, Nakajima H (2009) Metall Mater Trans A 40A:937–942
36. Makaya A, Fredriksson H (2005) Mater Sci Eng A 413A–414A:533–537
37. Wada T, Ide T, Nakajima H (2009) Metall Mater Trans A 40A:3204–3209
38. Murray P, White J (1955) Trans Br Ceram Soc 54:204–237
39. Tiller WA, Jackson KA, Rutter JW, Chalmers B (1953) Acta Metall 1:428–437
40. Kissinger HE (1957) Anal Chem 29:1702–1706
41. Nakajima H, Ide T (2012) In: Proceedings of cellular materials (Cellmat2012), Deutsche
Gesellschaft fur Materialskunde e.V, pp. 1–4
42. Suematsu T, Hyun SK, Nakajima H (2004) J Japan Inst Metals 68:257–261
43. Kubaschewski O, Alcock CB (1979) Metallurgical thermochemistry, 5th edn. Pergamon,
Oxford
44. Elliott JF, Gleiser M, Ramakrishna V (1963) Thermochemistry for steelmaking, vol 1.
Addition-Wesley, New York, pp 161–215
45. Onishi H, Ueno S, Hyun SK, Nakajima H (2009) Metall Mater Trans A 40A:438–443
Chapter 4
Nucleation and Growth Mechanism
of Pores in Metals
Abstract In this chapter, the nucleation and growth mechanism of pores in metals
are represented in order to explain why the directional pores are grown during the
unidirectional solidification. The presented models are consistent with the model
experiment of ice-freezing from water containing carbon dioxide. It is shown that
the mechanism of pore formation in lotus materials is much different from that of
foaming process.
Keywords Carbon dioxide • Ice • Pore growth • Pore nucleation • Sieverts’ law
4.1 Solubility of Gases in Metals: Sieverts’ Law
The difference between the solubility of gases in a metal melt and in the solid phase
after solidification causes gas precipitation in the metal at casting. The solubility of
a gas in a metal melt depends on its partial pressure on the surrounding atmosphere.
Most common gases are diatomic at room temperature. At high temperature
(temperature of the metal melt), a dissociation of the gas G2 occurs at the metal
surface in the presence of metal atoms M:
M þ G2 $ M þ 2G: (4.1)
A chemical equilibrium between the dissolved gas atoms G and the surrounding
gas is established. Guldberg–Waage’s law controls the equilibrium:
½G2
¼ constant; (4.2)
pG 2
where the value of the constant depends on the temperature. The gas atoms then
diffuse into the melt and/or the solid phase and stay there in the state of a solution.
We denote the concentration in the solution with ½G, where the underlined letter

66 4 Nucleation and Growth Mechanism of Pores in Metals
means that the gas is dissolved in the metal as G atoms and not G2 molecules. At
equilibrium, the concentration of the dissolved gas atoms in the melt or in the solid
phase is proportional to the square root of the partial pressure of the gas in the
surrounding atmosphere, which can be expressed as Sieverts’ law [1]:
pffiffiffiffiffiffiffi
½G ¼ constant pG 2 : (4.3)
4.2 Evolution Process of Directional Gas Pores
Figure 4.1 shows two different modes of gas pore evolution processes: (a) gas
nucleation and growth of pores in the case of water boiling or supersaturation in
soda in a glass and (b) pore nucleation and growth in unidirectional solidification
[2]. In the former case, for example, when the soda is poured into a glass cup,
heterogeneous nucleation occurs to nucleate the gas bubbles on the glass wall.
Fig. 4.1 Two different modes of gas pore evolution processes. (a) Gas nucleation and growth of
pores during water boiling and supersaturation in soda in a glass and (b) pore nucleation and
growth in unidirectional solidification in gas atmosphere (Reprinted with permission from [2].
© 1998 Materials Research Society)
4.2 Evolution Process of Directional Gas Pores 67
If the pore continues to absorb the carbon dioxide to coarsen, the interfacial surface
area, that is, the interfacial energy, between liquid and gas phases above the critical
size of the bubble becomes too large to maintain the stable gas pores’ condition, and
finally, the pores are detached from the glass wall and floated upward in order
to minimize the interfacial energy. Such phenomena are repeated in the soda
everywhere. We can see many bubbles evolution in the soda. On the other hand,
in the case of (b) during unidirectional solidification, the gas pores nucleate in the
interface between solidified metal and liquid metal by precipitation of insoluble gas
in the solid. If the pore growth rate is just identical to the solidification velocity, the
interfacial area (the interfacial energy) between liquid and gas phases can be kept
constant so that the coarsening so as to increase the pore diameter does not occur.
As a result, cylindrical pores can grow upward in the direction of the unidirectional
solidification. However, in practice, split and adhesion of the isolated pores are
often observed which are attributed to imperfection of flatness of the interface
between solid and liquid, the existence of impurity inclusion, nonuniformity of the
temperature gradient, convection of liquid, etc. In order to suppress such irregular
configuration of pores, precise control of temperature gradient during solidification,
suppression of liquid convection, and impurity control is necessary.
4.2.1 Pore Nucleation
According to the nucleation theory by Fisher [3], when the pores of hydrogen are
formed in the liquid phase during solidification by homogeneous nucleation, the gas
pressure required for the pore nucleation is evaluated to be a few GPa. This
magnitude of the pressure is not realistic [4]. Therefore, the pores are considered
to nucleate heterogeneously. In most cases, a few ppm of oxygen will be dissolved
in the metal melt, because the metal is melt in a vacuum of 100–101 Pa. As the
solidification proceeds, oxygen may concentrate at the solid–liquid interface to
form an oxide layer. When the hydrogen amount in the metal melt is large in the
case of high hydrogen pressure, the oxide layer at the solid–liquid interface is
resolved by hydrogen. However, in the case of low hydrogen pressure, the oxide
layer may remain at the solid–liquid interface. This oxide layer becomes a nucle-
ation site and causes formation of large size pores.
4.2.2 Pore Growth
The pore growth model is considered in the copper–hydrogen system on the basis of
the idea that the gas pores are formed by the hydrogen released at the solidification
front owing to the solubility difference between liquid and solid copper and by the
hydrogen diffused out from the solid copper due to the decrease in hydrogen
Fig. 4.2 Schematic drawing Gas

for unidirectional x
solidification process of PH2 PAr
liquid under the pressure of
Liquid
hydrogen PH2 and argon PAr.
The long pore grows upward
into which the gas atoms
dissolving in the solid and
liquid diffuse (Reprinted l+dl
with permission from dl
[5]. © 2001 Elsevier l r
Science B.V.) l+dl
P
l e
1-e
0
Solid
Pore
solubility of solid copper with deceasing temperature [5]. The following

assumptions are made to develop the model:
1. The solidification front advances at a constant velocity of the interface between
liquid and solid.
2. The solubilities of argon in liquid and solid copper are negligible.
3. The hydrogen concentration in solid copper is constant at a given distance from
the solidification front and is equal to that at the inner surface of the pore.
4. The temperature gradient in the ingot is constant.
Figure 4.2 shows a schematic drawing for the unidirectional solidification
process of the liquid copper. While the liquid copper in the crucible is saturated
with hydrogen during melting, the hydrogen partially escapes from the liquid to the
atmosphere during pouring and/or solidification. Let a be the ratio of the amount of
the escaped hydrogen to that of the hydrogen remaining in the liquid copper. As the
length of the columnar pore increases from l to l + dl, the mass balance of hydrogen
is written as follows:
(initial amount of hydrogen contained in the liquid copper in the volume element)
þ (amount of hydrogen contained in the liquid copper flowing into the volume
element to compensate for solidification shrinkage)
(amount of hydrogen contained in the liquid copper out of the volume element
due to the formation of pores)
þ (amount of hydrogen in the pore of a length of l )
þ (amount of hydrogen dissolved in the solid copper around the pore length of l )
¼ (amount of hydrogen in the pore of a length of l + dl)
þ (amount of hydrogen dissolved in the solid copper around the pore of length of
l + dl)
Cm Cm Cm
ρL ðl dlÞ þ ρL fβð1 εÞðl dlÞg ρL fεðl dlÞg þ W l þ Sl
aþ1 aþ1 aþ1
¼ Wlþdl þ Slþdl ; (4.4)
where ρL, liquid density, 8.00 103 kg m3; ρs, solid density, 8.40 103 kg m3;
β, solidification shrinkage of copper, 0.0476; Cm, hydrogen concentration of copper
melt in equilibrium with the mixture gas of argon and hydrogen at the holding
temperature; Tm, holding temperature of copper melt in the mixture gas of argon
and hydrogen (1,523 K); Wl, mass of hydrogen in the pore of length l; Wl+dl, mass of
hydrogen in the pore of length l + d; Sl, mass of hydrogen dissolved in solid copper
around the pore of length l; and Sl+dl, mass of hydrogen dissolved in solid copper
around the pore length l + dl.
First, the amount of hydrogen in the pore was calculated. Imagine a vertical
copper cylinder that has a cylindrical bore along its center and has a length of l as
shown in Fig. 4.2. Cross-sectional area is unity for the copper cylinder, whereas it is
ε for the cylindrical bore. This means that the porosity of the cylinder is equal to ε.
The bore is filled with hydrogen at a pressure of P. The temperature at the top end of
the cylinder corresponds to the nonvariant temperature Tn at which the reaction of
liquid copper!solid copper + gaseous hydrogen occurs. When the vertical tem-
perature profile in the cylinder is described as T ¼ G x þ ðTm G lÞ, the mass of
hydrogen dw in a small volume element of the bore with a length of dx is given by
M PðεdxÞ
dw ¼ ; (4.5)
RfGx þ ðTn GlÞg
where M, molecular weight of hydrogen, 2.016 kg mol1; R, gas constant; and G,

temperature gradient in the bore. The total mass of hydrogen W in the bore with a
length of l is obtained by integrating Eq. (4.5) from x ¼ 0 to x ¼ l; thus,
Z l
MPε Tn
W¼ dw ¼ ln : (4.6)
0 RG Tn Gl
The increase in the mass of hydrogen dW in the bore resulting from its growth by
dl is given by
dW MPε
dW ¼ dl ¼ dl: (4.7)
dl RðTn GlÞ
Next, the amount of hydrogen in the solid phase of copper around the pores was
calculated. Since the solubilities of argon in the solid and liquid phases of copper
are negligibly small, the principal gas species in the pore can be assumed to be
hydrogen. The hydrogen concentration in the solid phase of copper Cs is written as
pffiffiffi
Cs ðTÞ ¼ ηðTÞ P: (4.8)
Fig. 4.3 Distributions of

hydrogen concentration in
solid copper with pore length
of l and l + dl (Reprinted
with permission from
[5]. © 2001 Elsevier
Science B.V.)
ηðTÞ ¼ 4:34 107 expð5:888 103 =TÞ wt. fraction Pa1/2. The amounts of
hydrogen Sl and Sl+dl, in the solid phase of copper around the pores of length of l
and l + dl can be given as
Z l
Sl ¼ ρs ð1 εÞ Cs ðTÞdx (4.9)
0
and
Z lþdl
Slþdl ¼ ρs ð1 εÞ Cs ðTÞdx; (4.10)
0
respectively. Figure 4.3 shows the variation of the hydrogen concentration in the
solid phase with the distance in the solidification direction for pore lengths l and
l + dl. Since the area ABCD equals the area HIFG, the difference in the mass of
hydrogen Sl+sl–Sl corresponds to the area EBIH. For a small length of dl, the area
EBIH is approximately equal to the area of rectangle ABIH. Therefore,
Slþdl Sl ¼ ρs dl C0 : (4.11)
Substituting Eqs. (4.7) and (4.11) in Eq. (4.4), one finally obtains
" rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h ffi # 2
pffiffiffiffiffiffiffiffii
ρs ð1 εÞηðTm Þ þ ρ2s ð1 εÞ2 ηðTn GlÞ2 ½4 Mε=RðTn GlÞ ρL ð1 þ βÞð1 εÞξðTm Þ PH2 =ða þ 1Þ
P¼ :
2 Mε=RðTn GlÞ
(4.12)
Assuming that the hydrostatic pressure of the copper melt is negligible, the
pressure P in the pore is given by the sum of the pressure in the chamber and
the capillary pressure arising from the surface tension of the copper melt; thus,
P ¼ PH2 þ PAr þ Pr ; (4.13)
where PAr is argon pressure in the chamber, and Pr ¼ 2σ=r, where r is the radius of
the pore “cap” at the solidification front (assumed to be equal to the pore radius),
and σ is the surface tension of copper melt.
The relationship between the pressure of hydrogen and argon is expressed as
" rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h ffi # 2
pffiffiffiffiffiffiffiffii
ρs ð1 εÞηðTm Þ þ ρ2s ð1 εÞ2 ηðTn GlÞ2 ½4 Mε=RðTn GlÞ ρL ð1 þ βÞð1 εÞξðTm Þ PH2 =ða þ 1Þ
PAr ¼
2 Mε=RðTn GlÞ

2σ
: PH2
r
(4.14)
The solid lines and the dotted lines in Fig. 4.4 show the results calculated from
equation for the combination of the hydrogen pressure and the argon pressure to
give various porosity ε. In the calculation, the value a is assumed to be 0.52. This
means that 34:3 % ð¼ 0:52=ð1 þ 0:52Þ 100Þ of the hydrogen dissolved in the
copper melt at the melting temperature of 1,523 K escapes from the melt to the
atmosphere during pouring and/or solidification and, therefore, that the copper melt
to form the porous copper has the composition at which the nonvariant reaction
occurs. Hence, in the calculation, only the hydrogen dissolved in the melt of the
nonvariant composition is assumed to be incorporated in the ingot. Under such
condition, the escape factor a is not a function of the hydrogen pressure, but is
dependent only on the holding temperature. When the holding temperature is given,
the amount of hydrogen incorporated in the ingot is determined by the applied
hydrogen pressure. Since the pore is filled with hydrogen whose pressure is equal to
the sum of the total pressure in the chamber and the capillary pressure, the porosity
decreases with increasing argon pressure at a given hydrogen pressure. This
prediction is in agreement with the experimental results indicated in Fig. 4.4. The
amount of hydrogen dissolved in the melt is proportional to the square root of
the applied hydrogen pressure, whereas the pressure in the pore is proportional to
the total pressure in the chamber. At a given argon pressure, therefore, the porosity
Fig. 4.4 Porosity map of lotus copper. The numerals in the figure are the measured porosity. The
lines are the calculated results for the combination of the hydrogen pressure and argon pressure to
give various porosity values ε. In the calculation, the holding temperature of the copper melt and
the escape factor are 1,523 K and 0.52, respectively (Reprinted with permission from [5]. © 2001
Elsevier Science B.V.)
first increases with increasingly applied hydrogen pressure and then decreases at
higher hydrogen pressure due to densification of hydrogen in the pore as shown
in Fig. 4.4.
Figure 4.5 shows the calculation for the dependence of the porosity on the
holding temperature Tm. In the calculation, the escape factor a, the porosity ε,
and the mean pore radius r are assumed to be 0.52, 20 %, and 50 μm, respectively.
The hydrogen solubility in the copper melt increases with increasing holding
temperature, resulting in increase in the porosity. It is also known from Fig. 4.5
that the porosity decreases as the melting temperature of the copper decreases for a
given set of partial pressure of hydrogen and argon and for the factor of a ¼ 0.52.
For the purpose of avoiding this decrease in the porosity, it is necessary to
reduce the partial pressure of argon. When Tm is kept constant at 1,523 K, but the
escape factor a is varied, the relation between the argon pressure and the hydrogen
pressure to give specific porosity values is calculated from Eq. (4.14). The result is
shown in Fig. 4.6. As the escape factor a decreases, the porosity increases because
of increasing hydrogen amount that generates pores. The value a is associated with
various parameters and also depends on Tm. The factor a smaller than 0.52
Fig. 4.5 Calculated contribution of argon partial pressure and hydrogen partial pressure to give a
porosity of 20 % for various holding temperatures of the copper melt. Escape factor and mean pore
radius are 0.52 and 50 μm, respectively (Reprinted with permission from [5]. © 2001 Elsevier
Science B.V.)
Fig. 4.6 Calculated combination of argon partial pressure and hydrogen partial pressure to give
a porosity of 20 % for various escape factors. Holding temperature of the melt and mean pore radius are
1,523 K and 50 μm, respectively (Reprinted with permission from [5]. © 2001 Elsevier Science B.V.)
corresponds to the situation where the gas phase appeared in the liquid phase just
ahead of the solidification front incorporated in the solid phase.
For derivation of Eq. (4.14), the solidification is assumed to be under a steady
state, but the pore nucleation is not considered. As the undercooling of the melt in
the first stage of solidification will vary depending on the holding temperature Tm,
the cooling rate, etc., the nucleation rate may change. Further investigations are
necessary to elucidate the effect of the undercooling on the porosity.
Drenchev et al. [6] and Liu et al. [7] investigated a model of porosity formation
and spatial distribution of pores in lotus metals.
4.3 Model Experiment on Unidirectional Solidification

of Water Containing Carbon Dioxide
The formation of the gas pores (or blowholes) in castings and ingots has been
extensively studied [4, 8–10], since the gas pores have been regarded as defects that
degrade the mechanical properties of the metal products. For the purpose of
studying the formation of gas pores, model experiments were widely carried
out using water–gas system, and the nucleation and the growth of gas bubbles
were directly observed in the course of solidification. Chalmers described the
dependence of bubble formation on the freezing rate in water–air solution
[11]. When the freezing is slow, more air diffuses into the bubbles from the
surrounding water and they grow larger, while in the case where the growth rate
is high, there is less time for diffusion and the cross section of the bubble decreases.
Very slow freezing permits the air rejected at the ice/water interface to diffuse away
from the interface and neither bubbles nor long columnar pores appear. It was
reported [12, 13] that in the solidification of the water–air system and the
water–carbon dioxide system in nonsteady conditions, the shape of the gas pores,
most of which are short pores rather than long columnar pores, depends on the
growth rate of ice growing from the chill. In a unidirectional solidification of water
dissolving air at controlled growth rates, Geguzin and Dzuba [14] showed that the
air bubbles nucleated at the advancing solidification front grow into long columnar
pores when the displacement rate of the pore cap is equal to the growth rate of the
solid. They observed the periodic formation of elliptical pores along the solidifica-
tion direction, which they attributed to the alternative accumulation and “drop
down” of the liquid concentration near the solidification front.
For this purpose of simulating the formation of gas pores in lotus metals, it is
preferable to unidirectionally solidify water–gas solutions at a constant rate and to
use a gas having large solubility in water since the size and the pore density can be
changed over a wide range by changing the concentration of the gas. The formation
of gas pores in water–carbon dioxide system solidifying at constant rates is directly
observed to clarify the factors governing the morphology and the distribution of the
gas pores by Murakami and Nakajima [15, 16].
4.3 Model Experiment on Unidirectional Solidification of Water Containing. . . 75
Fig. 4.7 Schematic drawing of setup of model experiment using water–carbon dioxide solution
saturated with carbon dioxide. The glass cell filled with water–carbon dioxide solution is lowered
immersing into alcohol at 253 K. The gas pores are formed at the solidified ice (Reprinted with
permission from [15]. © 2002 Japan Institute of Metals)
Water–carbon dioxide solution saturated with carbon dioxide was prepared by

dissolving carbon dioxide in deaerated distilled water at 293 K and at atmospheric
pressure [17]. The solution in the glass cell was solidified unidirectionally upwards
by moving the glass cell down into an alcohol bath held at 253 K as shown in
Fig. 4.7. The moving velocity of the glass cell was in the range from 3 to 45 μm s1.
As the solidification proceeded, the pores of carbon dioxide grew in the solidifica-
tion direction with ice. The experimental setup was placed in a plastic desiccator
filled with flowing nitrogen to keep the low-temperature surface of the glass cell
from being misted by humidity of the air. The nucleation process and growth
process of the pores were observed directly by using a digital microscope through
the glass window of the desiccator. The growth rate of ice was the same as that of
the moving rate of the glass cell as long as the solidification front was observed
under the microscope.
The morphology of the carbon dioxide pores depends on the growth rate of ice as
shown in Fig. 4.8 in the case of the solution with the degree of saturation of 1/2.
At the growth rate of 3 μm s1, spherical pores that nucleated at the solidification
front grow into long columnar pores elongated in the solidification direction as
shown in Fig. 4.8a. The solidification front is “planar” except that it is concave
toward the solidification direction and that grain boundary grooves exist on the
solidification front. As the growth rate increases, the length and the diameter of
the columnar pores decrease and, at the growth rate of 12 and 18 μm s1, only very
Fig. 4.8 Effect of the growth rate on the pore morphology in ice. The degree of saturation of the
solution is 1/2. The transfer velocity is (a) 3 μm s1, (b) 6 μm s1, (c) 9 μm s1, (d) 12 μm s1, and
(e) 18 μm s1 (Reprinted with permission from [15]. © 2002 Japan Institute of Metals)
short columnar pores or chestnut-shaped pores are formed as shown in Fig. 4.8d, e.
The morphology of the solidification front changes from planar one to cellular one
with increasing growth rate due to the constitutional supercooling that occurs in the
liquid adjacent to the solidification front [18].
4.4 Influence of Ultrasonic Agitation on Pore Morphology

during Unidirectional Solidification of Water
Containing Carbon Dioxide
Tane and Nakajima studied the influence of the agitation of liquid on pore
formation and growth during unidirectional solidification [19]. In situ observation
during solidification is useful to study the pore formation and growth. For metals,
however, such observation is difficult, because pores are formed inside metals.
Thus, model experiments were widely carried out using water–gas system. They
also adopted water–carbon dioxide model system. The piezoelectric transducer,
attached to the glass cell with an adhesive, excites ultrasonic vibration of 44 kHz in
water–carbon dioxide solution; the magnitude (amplitude) of the ultrasonic
4.4 Influence of Ultrasonic Agitation on Pore Morphology during Unidirectional. . . 77
a b
Water
under US
vibration
Ice
under no US
vibration
CO2 pore
c
1 mm
under US
vibration
Fig. 4.9 Morphologies of pores formed during unidirectional solidification of water–carbon

dioxide solution when the moving velocity of glass cell is 3.0 μm s1: (a) No ultrasonic vibration
is applied (the electric voltage applied to a piezoelectric transducer is 0 Vpp), and (b) and (c)
ultrasonic (US) vibration is applied to the solution (the voltage applied to the transducer is
20 Vpp). The application of ultrasonic vibration to the solution was started just after the
photography of (a). After 960 and 1,500 s, photographs (a) and (b) were taken, respectively
(Reprinted with permission from [19]. © 2006 Japan Institute of Metals)
vibration is proportional to the electric voltage of alternate current signal applied to

the transducer (the energy of vibration is proportional to the square of the amplitude
of vibration). It is estimated that longitudinal vibration mainly is transmitted to
water because transverse vibration cannot transmit in water. The solution in the
glass cell solidifies unidirectionally upwards by moving the glass cell down at
3.0 μm s1 into an alcohol bath cooled at 253 K with a chiller. Pore formation and
growth under the agitation by ultrasonic vibration were observed with a digital
microscope.
Figure 4.9a shows a morphology of pores formed during unidirectional solidifi-
cation when no ultrasonic vibration is applied. Cylindrical pores grow in the
moving direction of the glass cell. Figure 4.9b, c shows morphologies of pore
formed under ultrasonic vibration by piezoelectric transducer applied by electric
voltage of 20 Vpp. The lower pores in Fig. 4.9b are formed when no ultrasonic
vibration is applied, while the upper pores are formed under ultrasonic vibration.
The application of ultrasonic vibration results in the halt of pore growth. After this
Fig. 4.10 Schematic

illustration showing the mean
solute concentration in liquid
near the solid–liquid
interface during
unidirectional solidification
of a binary solution with
planar solidification front:
(a) agitation of liquid is
not present and (b) liquid
is agitated by disturbance
from [19]. © 2006 Japan
Institute of Metals)
halt, new pores nucleate and grow. All the pores in Fig. 4.9c are formed under
ultrasonic vibration. Flask-shaped pores are formed under the agitation by ultra-
sonic vibration.
The supply quantity of carbon dioxide depends on the concentration profile of
carbon dioxide in the water near the solid–liquid interface. When the convection
of water exists during solidification of binary solution, the concentration profile of
solute in the front of the interface can be modeled by using the stagnant layer model
[9, 18], which was applied to discuss the influence of agitation. When the amount of
carbon dioxide absorbed by pores is balanced with the amount of carbon dioxide
supplied with advance in solidification, the steady-state growth of pores is achieved
and cylindrical pores are formed. On the other hand, insufficient supply of carbon
dioxide prevents the steady-state growth of pores and results in the halt of pore
growth. Thus, the sufficient supply of carbon dioxide is important for the steady-
state growth of cylindrical pores. Figure 4.10 shows schematic illustrations
exhibiting mean solute concentration in the liquid near the solid–liquid interface
with and without agitation. The mean concentration with agitation drastically
decreases with increasing distance in growth direction x. For x > δ, liquid is
agitated and the solute concentration is constant. For x < δ, liquid is not agitated
owing to the viscosity of liquid, and in this layer, the mass transport is due to only
diffusion. The δ decreases with increase in the magnitude of agitation.
4.5 Evolution of Spherical Pores during Foaming Process 79
When ultrasonic vibration is applied to water–carbon dioxide solution, its

concentration near solid–liquid interface decreases. Thus, the supply of carbon
dioxide becomes insufficient for the steady-state growth of cylindrical pores and
the growth of pores stops as shown in Fig. 4.9b. During the halt of pore growth,
the solute concentration increases with solidification because carbon dioxide
rejected with advance in solidification is not used for the growth of pores. Then,
the concentration of carbon dioxide reaches a critical value and pores nucleate. Just
after the nucleation, the pores grow by absorbing the carbon dioxide accumulated
during the halt so that the diameter of pores at the start of growth is large. However,
the carbon dioxide supplied with advance in solidification is not sufficient to
maintain the initial diameter of pores owing to the decrease of carbon dioxide
near the solid–liquid interface by ultrasonic agitation. Therefore, the diameter of
pores decreases with advance in solidification and the flask-shaped pores are
formed as shown in Fig. 4.9b.
4.5 Evolution of Spherical Pores during Foaming Process
In lotus metals fabricated gasar process, the directional pores are evolved by
nucleation and growth of insoluble gas during solidification. Here, it is worthwhile
to show growth mechanism of foamed metals during foaming process when the
molten metal dissolving gas is solidified. This mechanism is due to the swelling of
the isotropic spherical bubbles, which is clearly different from that of pores in
lotus metals.
The evolution of the cell structure of foams produced by gas bubble expansion in
a molten metal is numerically simulated by the Lattice Boltzmann Method by
Korner et al. [20]. The decomposition of the blowing agent is modeled by a constant
homogeneous volume source of hydrogen. The hydrogen is dissolved in the liquid
aluminum and diffuses to stochastically distributed bubble nuclei already present at
the beginning of foaming. The gas flow into bubbles and the surrounding atmo-
sphere is determined by the diffusion equation and Sieverts law. Movement of the
melt is given by the Navier–Stokes equations. Figure 4.11 shows the temporal
evolution and subsequent collapse of the cellular structure. From the very begin-
ning, foam expansion is accompanied by bubble coalescence if the density of the
initial bubble nuclei is high enough. Coalescence is depicted in more detail in
Fig. 4.12 and compared with microcomputer tomography (μCT) pictures of real
foams. The number of bubbles decreases via bubble coalescence during foam
expansion. If the initial density of nuclei is high, then substantial coalescence
takes place during expansion from the very beginning. The maximum expansion
factor is determined by two mechanism of gas loss. The first one is gas loss by
diffusion to the foam surface. The contribution of this effect depends on the foam
surface to volume ratio, the ambient pressure and atmosphere, and the foaming
velocity. For samples smaller than 5 mm and foaming times larger than some
minutes, this kind of gas loss can be substantial and suppresses expansion.
Fig. 4.11 Temporal evolution of the foam structure. Growth of bubble nuclei, bubble coales-
cence, gas loss to the surrounding by bursting bubbles on the foam surface, and eventually, foam
collapse (Reprinted with permission from [20]. © 2001 MIT-Verlag)
Fig. 4.12 Coalescence. From the very beginning, foam growth is intimately correlated with
bubble coalescence. (Left) Simulated structures for small expansion ratios (600 initial nuclei,
parameter as in Fig.4.11). (Right) μCT pictures of a Mepura wrought alloy for two different
expansion states. The samples were heated up under an external pressure of 100 bar and were
subsequently foamed by decreasing the pressure (Reprinted with permission from [20]. © 2001
MIT-Verlag)
The second origin of gas loss is bursting cell walls located at the foam surface
(see Fig. 4.11). In this case, the cell disappears and its gas constant is completely
lost. As long as the cell size is small, this effect is insignificant. It starts to become
important when the cell size reaches the order of magnitude of the foam dimension.
In this case, the amount of gas lost to the environment is higher than the expansion
volume itself. As a result, the foam collapses.
The in situ observation of the foaming process of aluminum was carried out by
Banhart group by means of real-time radioscopy using 150 kV microfocus X-ray
4.5 Evolution of Spherical Pores during Foaming Process 81
Fig. 4.13 X-ray radiograms of AlSi11 foams with different alloying elements (Fe, Sn, Sb, and In)
at 200 s after start of heating, foaming without mold (Reprinted with permission from [21]. © 2005
Japan Institute of Metals)
source with 5 μm spot size [21]. Although the micrographs are not a temporal
evolution of the cellular structure, Fig. 4.13 shows examples of X-ray radiograms of
AlSi11 foams with different additional alloying elements. 200 s after the heating,
it foamed without mold. Obviously, there is drastic collapse of the sample with Fe,
followed by the one with Sn. On the other hand, the samples with Sb and In are
more homogenous than the sample without. After 200 s, no significant differences
in drainage between the samples can be found.
References
1. Fredriksson H, Akerlind U (2006) Materials processing during casting. Wiley, Chichester,

pp 258–259
2. Shapovalov VI (1998). In: Scharttz DS, Shih DS, Evans AG, Wadley HNG (ed) Porous and
cellular materials for structural applications, Materials Research Society, Warrendale,
pp 281–290
3. Fisher JC (1948) J Appl Phys 19:1062–1067
4. Campbell J (1993) Casting. Butterworth-Heinemann, Oxford, p 163
5. Yamamura S, Shiota H, Murakami K, Nakajima H (2001) Mater Sci Eng A 318:137–143
6. Drenchev L, Sobczak J, Sobczak N, Sha W, Malinov S (2007) Acta Mater 55:6459–6471
7. Wan J, Li Y, Liu Y (2007) J Mater Sci 42:6446–6452
8. Lampman S, Moosbrugger C, DeGuire E (eds) (1998) Metals handbook, 9th edn, vol 15:
Casting. ASM International, Metals Park, p 82
9. Flemings MC (1974) Solidification processing. McGraw-Hill, New York, p 208
10. Fredriksson H, Svensson I (1976) Metall Trans 7B:599–606
11. Chalmers B (1959) Sci Amer 200:114–122
12. Wei PS, Kuo YK, Chiu SH, Ho CY (2000) Int J Heat Mass Transfer 43:263–280
13. Vasconcellos KF, Beech J (1975) J Cryst Growth 28:85–92
14. Geguzin TE, Dzuba AS (1981) J Cryst Growth 52:337–344
15. Murakami K, Nakajima H (2002) Mater Trans 43:2582–2588
16. Murakami K, Nakai Y, Nakajima H (2002) Int J Cast Metals Res 15:459–463
17. Handbook of chemistry (1984). In: Chem Soc Japan (ed) Pure chemistry II, 3rd edn, Maruzen,
Tokyo, p 158
18. Chalmers B (1964) Principles of solidification. Wiley, New York, p 150
19. Tane M, Nakajima H (2006) Mater Trans 47:2183–2187
20. Korner C, Thies M, Arnold M, Singer RF (2001) In: Banhart J, Ashby MF, Fleck NA (eds)
Cellular metals and metal foaming technology. MIT, Bremen, pp 93–98
21. Weise J, Marchetto O, Haesche M, Garcia-Moreno F, Banhart J (2005) In: Nakajima H,
Kanetake N (eds) Porous metals and metal foaming technology. Japan Inst Metals,
Sendai, pp 123–128
Chapter 5
Control of Pore Size and Porosity
in Lotus-Type Porous Metals
Abstract The porous metals with directional elongated pores are characterized by
three important parameters: the pore direction, pore size, and porosity. The pore
direction can be adjusted by changing the direction of unidirectional solidification
as shown before. Here we describe the way how to control the pore size and the
porosity by adjusting the solidification velocity and ambient gas pressure and by
addition of oxide particles.
Keywords Boyle’s law • Gas pressure • Moisture • Pore size • Solidification

velocity
5.1 Control of Pore Size by Solidification Velocity
The lotus copper was fabricated with the mold casting technique by the vacuum-
assisted and pressurized casting apparatus consisting of a graphite crucible with a
hole on the bottom of the crucible, a stopper stick for preventing the melt flow
through the hole, an induction heating coil, and a mold with water-cooled copper
plate. High-purity copper was melted in the crucible by radio-frequency heating
under high-pressure mixture gas of hydrogen and argon. The molten copper was
poured into the mold whose bottom plate was cooled down with water circulated
through a chiller. The lateral side of the mold was made of alumina-coated stainless
steel tube, which was suitable for insulating in order to be solidified in one direction
from the bottom to the top. During the solidification, hydrogen in the melt was
rejected at the solid–liquid interface due to the solubility gap of hydrogen between
liquid and solid and forms cylindrical pores that were aligned parallel to the
solidification direction. The samples were solidified at different solidification
velocities through the condition whether a ceramic sheet was inserted between a
carbon plate and a copper plate of chiller as shown in Fig. 5.1. R-type
thermocouples were inserted in the mold and coupled with a computer-controlled
data acquisition system for recording the cooling curves.

84 5 Control of Pore Size and Porosity in Lotus-Type Porous Metals
Fig. 5.1 Schematic drawing for measurement of solidification velocity in mold casting technique.
The molten metal is dropped through a funnel from the crucible into the bottom mold. Several
thermocouples are set up in the vertical direction on the lateral side of the mold. In order to control
the solidification velocity, ceramic sheets are inserted between the copper chiller and the mold
(Reprinted with permission from [1]. © 2003 Elsevier Ltd)
It is considered that each solidification velocity is constant since the beginning

time of solidification linearly increases with increasing distance from bottom of
mold, in which the thermal conductivity of copper can be sufficiently high. The
samples were solidified with solidification velocities of 1.185 and 0.697 mm s1
through different conditions of the ceramic sheet or not. The optical micrographs of
the cross section of the lotus copper are shown in Fig. 5.2. It is apparent that the pore
morphology is significantly affected by the solidification velocity. The average
5.1 Control of Pore Size by Solidification Velocity 85
Distance from bottom of mold

15 mm 25 mm 40 mm
0.697 mms-1
Growth velocity during solidification
1.185 mms-1
500 µm
Fig. 5.2 Microstructure of lotus copper in perpendicular section to the solidification direction.
Lotus copper ingots were fabricated with two different solidification velocities by the mold casting
technique. The micrographs were taken in the sections at the position from different distances from
the bottom of the mold (Reprinted with permission from [1]. © 2003 Elsevier Ltd)
150
Average pore diameter / µm
0.697 mms-1
100
50
1.185 mms-1
0
15 20 25 30 35 40
Distance from bottom of mold / mm
Fig. 5.3 Average pore diameter plotted against distance from the bottom of the mold in lotus
copper fabricated by mold casting technique with two different solidification velocities (Reprinted
with permission from [1]. © 2003 Elsevier Ltd)
pore diameter is plotted against the distance from the bottom of the mold as shown
in Fig. 5.3. The pore diameter in the lotus copper fabricated with the solidification
velocity of 0.697 mm s1 is twice larger than that with 1.185 mm s1. It can be
understood that the amount of hydrogen diffusing from liquid to the pores increases
with decreasing solidification velocity and then the pores formed with the velocity
Fig. 5.4 Cross sections perpendicular to the solidification direction of lotus stainless steel
fabricated under hydrogen gas of 1.0 MPa by continuous zone melting technique. The upper
figures (a), (b), and (c) are of the rods fabricated without gas-blow cooling and the lower figures
(d), (e), and (f) are of those fabricated with gas-blow cooling. The transfer velocities are
160 μm s1 for (a) and (d), 330 μm s1 for (b) and (e), and 500 μm s1 for (c) and (f) (Reprinted
with permission from [2]. © 2005 The Minerals, Metals & Materials Society and ASM
International)
of 0.697 mm s1 become larger than those with 1.185 mm s1. As the solidification
velocity increases, the pore size decreases but the number density of pores
increases. Increasing the solidification velocity causes an increase in the hydrogen
supersaturation of the solid–liquid interface and then the driving force to nucleate
the pores increases. If the entire amount of hydrogen dissolved in the melt is
diffused into the pores, the pore size can decrease with increasing number density
of pores. Consequently it can be understood that the solidification velocity can
affect the average pore diameter of lotus copper [1].
In Fig. 5.4, the cross-sectional planes perpendicular to the solidification direction
in lotus stainless steel rods fabricated by the continuous zone melting technique
under the hydrogen atmosphere of 1.0 MPa without or with the gas-blow cooling gas
are shown. The average pore diameter and porosity determined by image analyses
are shown in Table 5.1. The average pore diameter decreases with increasing
transfer velocity, while the porosities are not much different. It is found that the
cooling rate is almost proportional to the transfer velocity in fabricating porous
metal rods with gas-blow cooling. The cause of the smaller pore diameters in the
rods fabricated with the larger velocities is considered to be due to the higher cooling
rates [2]. Such tendency is consistent with the experimental results in the fabrication
5.2 Control of Pore Size and Porosity by Ambient Gas Pressure 87
Table 5.1 Porosity and average pore diameter of lotus stainless steel fabricated under hydrogen
atmosphere of 1.0 MPa by continuous zone melting technique with/without gas-blow cooling
Average pore
Transfer velocity (μm s1) Gas-blow cooling Porosity (%) diameter (μm)
160 52 890
330 49 710
500 64 1,070
160 ○ 59 1,020
330 ○ 48 550
550 ○ 46 480
of porous copper by the casting technique [1]. Furthermore, since the temperature
gradient increases by using gas-blower, it is considered that the solidification
velocity increases by the gas-blower and thus the pore size becomes smaller.
5.2 Control of Pore Size and Porosity by Ambient

Gas Pressure
The average pore diameter and porosity of the lotus stainless steel rods that were
fabricated under hydrogen and argon or helium of various pressures with the
transfer velocity of 330 μm s1 are summarized in Fig. 5.5a, b. The average pore
diameter and porosity decrease with increasing pressure under the atmosphere of
only hydrogen. The concentration of hydrogen dissolving in liquid metal is propor-
tional to the square root of hydrogen pressure according to Sieverts’ law. On the
other hand, the volume of gas in the pores is reciprocally proportional to the
pressure of the ambient gas according to Boyle’s law since the pressure of pores
should be almost in balance with the pressure of ambient gas. Therefore, it is
considered that the decrease in the average pore diameter and porosity with increase
in pressure of hydrogen is caused by more significant influence of pore volumes
than that of the solubility of hydrogen.
Under mixed gas of hydrogen and argon of a constant total pressure, the porosity
increases and average pore diameter decreases with increasing partial pressure of
hydrogen. Hydrogen solubility in the melt of stainless steel increases with increase
of the hydrogen pressure according to Sieverts’ law. The pores are composed of
only hydrogen since argon is not dissolved in the melts. Since the pressure in the
pores should be balanced with the total pressure, the pressure in the pores is
constant regardless of the variation of the hydrogen partial pressure under a
constant total pressure. Therefore, it is considered that the increase in the hydrogen
solubility due to the increase of the partial pressure of hydrogen leads to the
increase in porosity under a constant total pressure.
Such a pressure dependence of porosity of lotus stainless steel fabricated by the
continuous zone melting technique is qualitatively consistent with that of lotus
Fig. 5.5 Average pore diameter (upper number) and porosity (lower number) of the lotus stainless
steel fabricated by continuous zone melting technique with gas-blowing cooling under various
pressures of mixed gases composed of (a) hydrogen and argon or (b) hydrogen and helium. These
diagrams are for the cases where the transfer velocity is 330 μm s1. (c) The porosity of lotus
stainless steel fabricated by mold casting technique (Reprinted with permission from [2]. © 2005
stainless steel fabricated by mold casting technique [3] shown in Fig. 5.5c. How-
ever, the porosity by the continuous zone melting technique is a little lower than
that by the mold casting technique. It is understood that the hydrogen concentration
in the melt zone may not be saturated since the hydrogen in the ambient atmosphere
can be dissolved into the melt only while the metal is transferred across the melt
zone of at most 10 mm. Therefore, the hydrogen concentration is a little smaller for
the continuous zone melting technique. This leads to result in the smaller porosity
of the lotus stainless steel fabricated by the continuous zone melting technique than
that by the mold casting technique. In Table 5.1, the porosity decreases little by
little with increase of the transfer velocity in the case with gas-blow cooling,
5.3 Control of Pore Size by Addition of Oxide Particles 89
comparing with the case without the cooling. This is also possibly due to the
decrease of hydrogen dissolution into melt with increase of the transfer velocity.
On the other hand, the average pore diameter increases with increasing argon
pressure under a constant total pressure. This is possibly related with the heat
conductivity of ambient gas. The heat conductivity of hydrogen is
21.18 102 Wm1 K1 (373 K) [4] and that of argon is 2.12 102 Wm1 K1
(373 K) [4], the latter of which is an order of magnitude lower than that of
hydrogen. Therefore, the heat conductivity of ambient gas significantly decreases
with increasing partial pressure of argon and hence the efficiency of heat transmis-
sion from the metal rod to the ambient gas by gas-blow cooling also decreases.
Consequently, this leads to the lowering of the cooling rate and hence to the
increase in average pore diameter.
Figure 5.5b shows the pressure dependence of average pore diameter and porosity
of the lotus stainless steel fabricated under mixed gas of hydrogen and helium. The
porosity decreases with increasing helium pressure under a constant total pressure.
This trend is almost consistent with the case where mixed gas of hydrogen and argon
is used. The average pore diameter also has a trend of increase with increasing
helium pressure. The heat conductivity of helium is 17.77 102 Wm1 K1
(373 K) [4], which is a little smaller than that of hydrogen but in the same order of
magnitude as hydrogen. Therefore, it is expected that the heat conductivity of
ambient gas and hence the cooling rate does not much vary by increasing helium
partial pressure under a constant total pressure. Nevertheless, the average pore
diameter depends on the helium pressure under a constant total pressure similar to
the case of hydrogen-argon gas. The reason for this is not very clear at present.
Further investigations are required.
5.3 Control of Pore Size by Addition of Oxide Particles
Suematsu et al. proposed a fabrication method of lotus nickel using moisture during
the solidification in argon atmosphere [5]. The pore size of the sample fabricated by
this method is smaller than that of the sample fabricated in pressurized hydrogen
atmosphere (PGM). Such smaller pore formation is attributed to formation of
nucleation sites of the oxide. According to the following reaction,
n H2 OðgasÞ þ m M $ Mm On þ 2n H; (5.1)
the moisture decomposes into hydrogen and metal oxide. The former produces
hydrogen pores, while the latter may serve as the nucleation sites for the pores.
Although the moisture could be produced by the reverse reaction (1), the moisture
(H2O) itself cannot be dissolved into the molten nickel. Therefore, it is considered
that the moisture does not contribute to evolve any pores in the solidified nickel.
The experimental results by Suematsu et al. suggests that the pore size in lotus
nickel can be controlled by the amount and/or particle size of NiO powder [5].
Fig. 5.6 The longitudinal cross-sectional (a) and transversal cross-sectional (b) views of the
sample fabricated in mixture gas of 0.65 MPa Ar and 0.15 MPa H2 using NiO powder whose
weight is 0.5 g and particle size is 7 μm. (c) and (d) show the longitudinal cross-sectional and
transversal cross-sectional views of the sample fabricated in the same atmosphere without NiO
powder (Reprinted with permission from [6]. © 2008 Japan Institute of Metals)
Here the effect of NiO powder on the formation of pores was examined by
Onishi et al. in order to elucidate the relation among pore size in lotus nickel, the
amount, and particle size of NiO powder [6]. The solidification was conducted by
mold casting under pressurized gas of argon and hydrogen. 120 g of nickel ingot
with 99.9 % purity was heated in alumina crucible by an induction coil and the
molten nickel was poured into the mold whose bottom was made of copper and the
side was made by molybdenum. The bottom of the mold was cooled by water
chiller. An appropriate amount of NiO powder was put on the copper mold before
casting.
Figure 5.6a, b shows the longitudinal and transversal cross-sectional views of
the sample fabricated in the mixture of gas 0.65 MPa Ar and 0.15 MPa H2 using
NiO powder whose weight is 0.5 g and particle size is 7 μm. On the other hand,
Fig. 5.6c, d shows longitudinal and transversal cross-sectional views of the sample
fabricated under the same atmosphere without NiO powder. The porosity and
average pore diameter calculated from Fig. 5.6 are 29 % and 31 μm for (b) and
62 % and 427 μm for (d), respectively. Apparently, the average pore diameter for
(b) is smaller than that of (d). On the other hand, the aspect ratio of pores for (a), 2.2,
is smaller than that of (c), 4.7. When the melt is solidified, supersaturated hydrogen
in the solid diffuses into pores. If the number density of pores increases, the amount
5.3 Control of Pore Size by Addition of Oxide Particles 91
Fig. 5.7 Dependence of

porosity on amount of NiO
powder with different sizes
of the particles. Lotus nickel
was fabricated in mixture gas
of 0.85 MPa Ar and
0.15 MPa H2 (Reprinted with
permission from [6]. © 2008
Fig. 5.8 The average pore

diameter of the samples as a
function of mass of NiO
powder with different sizes
of the particles. Lotus nickel
was fabricated in mixture gas
of 0.85 MPa Ar and
0.15 MPa H2 (Reprinted with
permission from [6]. © 2008
of hydrogen which diffuses into each pore from the surrounding region decreases so
that the pores cannot grow continuously in the direction of the solidification. Thus,
the aspect ratio of the pores in the case of addition of NiO is shorter than that
without NiO. Since the pore size decreases and the number of density of pores
increases by addition of NiO, it is considered that the oxide powders act as
nucleation sites for pores.
Figure 5.7 shows the porosity change with the amount of NiO powder. All
samples in Fig. 5.7 were fabricated in the mixture of 0.85 MPa Ar and 0.15 MPa
H2. The porosity increases with increasing amount of NiO powder. It was suggested
that about half of the hydrogen amount with the solubility gap between solid and
liquid is released to the atmosphere not to contribute to the pore formation
[7]. However, if the number of the pore nucleation sites increases by addition of
excess NiO powder, more insoluble hydrogen may be trapped by the nucleation
sites. Thus, higher porosity was observed with increasing mass of NiO powder, as
shown in Fig. 5.7. However, no significant effect of the particles size of NiO
powder on the porosity was found.
Figure 5.8 shows the average pore diameter of the samples as a function of the
mass of NiO powder. The pore diameter decreases with decreasing particle size of
NiO powder. Since the number density of particles significantly increases with
decreasing particle size when the same mass of NiO powder is added, the number of
pores increases and the resulting pore size decreases as shown in Fig. 5.8. This may
be attributed to the increase in the number of hydrogen atoms for formation of the
pores by NiO powder addition. On the other hand, the average pore diameter
monotonously increases with increasing amount of NiO powder. It is concluded
that NiO powder can serve as nucleation sites for the pore formation in the process
of the solidification.
References
1. Hyun SK, Nakajima H (2003) Mater Lett 57:3149–3154

2. Ikeda T, Aoki T, Nakajima H (2005) Metall Mater Trans A 36A:77–86
3. Ikeda T, Tsukamoto M, Nakajima H (2002) Mater Trans 43:2678–2684
4. National Astronomical Observatory (ed) (2002) Chronological scientific tables. Maruzen Co.
Ltd, Tokyo, p 404
5. Suematsu Y, Hyun SK, Nakajima H (2004) J Jpn Inst Metals 68:257–261
6. Onishi H, Ueno S, Nakajima H (2008) Mater Trans 49:2670–2672
7. Yamamura S, Shiota H, Murakami K, Nakajima H (2001) Mater Sci Eng A 318:137–143
Chapter 6
Details of Fabrication Techniques of Various
Lotus Metals and Alloys, Lotus Intermetallic
Compounds, Lotus Semiconductors,
and Lotus Ceramics
Abstract Case studies are shown by different atmospheric gases and by different
materials such as intermetallic compounds, semiconductors, ceramics, and magne-
sium and iron alloys.
Keywords Alumina • Aluminum • Ceramics • Czochralski technique • Intermetal-

lic compounds • Iron • Silicon • Solid-state diffusion
6.1 Fabrication of Lotus Iron by Nitrogen
So far hydrogen has widely been used as a dissolving gas to fabricate several kinds
of porous metals such as lotus metals. However, since the hydrogen gas is inflam-
mable and explosive, its use is not convenient from the industrial point of view. Use
of other gases except hydrogen would be desirable. It is well known that nitrogen is
an important alloying element widely used to improve corrosion resistance and
mechanical properties of steels [1]. Hyun and Nakajima noticed that the tempera-
ture dependence of nitrogen solubility in solid and liquid of iron is similar to that of
hydrogen [2], which exhibits large nitrogen solubility difference between solid and
liquid iron at the melting temperature [3]. Moreover, it is noticed that an invariant
reaction of “gas-evolution crystallization reaction” [4] takes place in the Fe–N
system, in which the iron melt dissolving nitrogen is solidified to transform into the
primary solid solution and nitrogen gas phase. Utilizing the nitrogen solubility
difference between liquid and solid, it is possible to fabricate porous iron using safe
nitrogen gas instead of hydrogen gas. This new type of porous iron is promising for
commercial application.
Figure 6.1 shows the typical optical micrographs of the cross section at the
bottom part of the ingot. The overall porosity was measured as functions of the
partial pressures of nitrogen and argon. The porosity was found to be affected by
the nitrogen and argon pressures, which is similar to the system of hydrogen–argon
gas. At a given nitrogen pressure, the porosity decreases with increasing argon

94 6 Details of Fabrication Techniques of Various Lotus Metals and Alloys. . .
Fig. 6.1 Optical micrographs of cross sections of lotus iron with 37.7 % porosity. The total
pressure 1.5 MPa with the partial pressure of nitrogen of 1.0 MPa and the partial pressure of argon
of 0.5 MPa. (a) Cross section perpendicular to the pore axis and (b) cross section parallel to the
pore axis (Reprinted with permission from [2]. © 2002 Japan Institute of Metals)
pressure as shown in Fig. 6.2a. At a given argon pressure, the porosity increases
with increasing nitrogen pressure as shown in Fig. 6.2b.
The solubility of nitrogen in iron is proportional to the square root of the nitrogen
pffiffiffiffiffiffiffiffi
gas pressure, PN2 , according to Sieverts’ law:
pffiffiffiffiffiffiffiffi )
CLN ¼ K L PN 2
δ δ
pffiffiffiffiffiffiffiffi ; (6.1)
C N ¼ K PN 2
6.1 Fabrication of Lotus Iron by Nitrogen 95
Fig. 6.2 Dependence of a 50

porosity in lotus iron on (a)
the partial pressure of argon Partial Pressure of Nitrogen = 0.5 MPa
at nitrogen pressure of 40
0.5 MPa and (b) the partial
pressure of nitrogen at argon
Porosity (%)
pressure of 1.0 MPa 30
from [2]. © 2002 Japan
Institute of Metals) 20
10
0
0.0 0.5 1.0 1.5
Partial Pressure of Argon, PAr / MPa
b 50
Partial Pressure of Argon = 1.0 MPa
40
Porosity (%)
30
20
10
0
0.0 0.5 1.0 1.5
Partial Pressure of Nitrogen, PN2 / MPa
where CLN and CδN are the nitrogen concentration in the liquid and the δ-iron,
respectively, and KL and Kδ are equilibrium constants given by

ΔG ΔS ΔH
K ¼ exp ¼ exp ; (6.2)
RT R RT
where ΔG is the Gibbs energy change for the solution of nitrogen, T is an absolute
temperature, R is the gas constant, and ΔH and ΔS are the standard enthalpy and
entropy of solution of nitrogen, respectively. The solubility of nitrogen in the
different iron phases is usually evaluated from the slope of the decomposition
pressure on logarithmic scale against 1/T [5]. The solubility of nitrogen in the
molten iron is much higher than that of δ-iron as shown in Fig. 6.3. Therefore,
nitrogen in the molten iron is rejected at the solid–liquid interface during solidifi-
cation and the pores nucleate and grow unidirectionally. Considering the mixture
a
α+Fe4N
C mp
Nitrogen Solubility
L
C α↔γ
γ
mp
γ↔δ Cδ
C
δ
T(α↔γ) T(γ↔δ) T(melting point)

b Pore Pore Pore
N N
N N
N N
α-phase+Fe4N γ-phase δ-phase
Cooling
Fig. 6.3 (a) Temperature dependence of nitrogen solubility in iron and (b) schematic of nitrogen
movement during cooling after solidification in lotus iron fabricated with nitrogen gas (Reprinted
with permission from [2]. © 2002 Japan Institute of Metals)
gas of nitrogen and argon, although the solubility of nitrogen in liquid is determined
by the partial pressure of nitrogen, the nitrogen solubility near the pores is related
with the internal gas pressure Pi for satisfying with the equilibrium between solid
phase around the pores and gas phase in the pores. Therefore, Eq. (6.1) can be
written as
pffiffiffiffiffi
CδN ¼ K δ Pi : (6.3)
From this equation, it can be seen that the solubility of nitrogen in solid is related
with total pressure of nitrogen and argon. With increasing partial pressure of argon
at a given total pressure, the nitrogen solubility in liquid decreases but the nitrogen
solubility in solid is almost constant.
Figure 6.4 shows the nitrogen concentration in the matrix of lotus iron as a
function of the partial pressure of nitrogen when the total gas pressure of nitrogen
and argon is kept constant at 2.0 MPa. It was found that the measured nitrogen
concentration is in the range from 0.06 to 0.135 mass%. According to the solubility
curve of nitrogen in iron [5], the nitrogen solubility in γ-iron decreases with
increasing temperature and the maximum solubility is 0.135 mass% at the lowest
6.1 Fabrication of Lotus Iron by Nitrogen 97
Fig. 6.4 Nitrogen concentration in the matrix of lotus iron versus the partial pressure of nitrogen
gas under constant total pressure of 2.0 MPa (Reprinted with permission from [2]. © 2002 Japan
Institute of Metals)
temperature of the γ-phase. On the other hand, in δ-iron, the nitrogen solubility
increases with increasing temperature and the maximum solubility is calculated as
0.06 mass% at the temperature just below the melting point of iron. Thus, it is
reasonable that the measured nitrogen concentration range as shown in Fig. 6.4 is in
between these two limits of the solubility. The reason why the measured nitrogen
concentration is much higher than the nitrogen solubility in δ-iron can be explained
as follows. As shown in Fig. 6.3 when the molten iron is solidified in the pressurized
mixture gas of nitrogen and argon, the pores are formed by the evolution of nitrogen
insoluble in the solid iron. In the subsequent cooling process in the solid, the
nitrogen solubility in δ-iron gradually decreases with decreasing temperature.
γ$δ
Then, the insoluble nitrogen atoms ðCmp δ Cδ Þ diffuse into the pores from
δ-iron matrix. However, in further cooling process in γ-iron, the nitrogen solubility
abruptly increases at the δ ! γ phase transformation temperature and continues to
increase with decreasing temperature until the γ ! α phase transformation tem-
perature is reached. In this process the nitrogen gas in the pores is dissolved into the
γ-iron. Finally, two phases of α(ferrite) phase and Fe4N compound are formed on
the phase transformation from γ-phase to α-phase.
Figure 6.5 shows the microstructure of the specimen containing 0.137 mass%
nitrogen in the matrix of quenched porous iron which was fabricated in the mixture
gas of PN2 ¼ 1:25 MPa and PAr ¼ 0:75 MPa. The proeutectoid ferrite was observed
on the grain boundaries and also inside the grains in the as-cast porous iron.
Fig. 6.5 Optical micrograph of as-quenched lotus iron containing 0.137 mass% nitrogen in the
matrix fabricated in the mixture gases of nitrogen 1.25 MPa and argon 0.75 MPa (Reprinted with
permission from [2]. © 2002 Japan Institute of Metals)
6.2 Fabrication of Lotus Silver by Oxygen
In most combinations of the metals and gas such as hydrogen or nitrogen, the gas
solubility in both of liquid and solid phases tends to decrease with decreasing
temperature. However, a unique tendency is observed in the silver–oxygen system;
the oxygen solubility in liquid silver increases with decreasing temperature [6]. It is
interesting to know the effect of such an inverse temperature dependence of
solubility on the pore formation and growth during the unidirectional solidification.
The lotus-type porous silver was produced by unidirectional solidification through
the mold casting technique and the Czochralski technique under pressurized
oxygen [7]. Figure 6.6 shows a schematic picture of the apparatus for the
Czochralski technique consisting of a melting part and a transfer mechanism of
the rod in a high-pressure chamber. First, silver in an alumina crucible was melted
by an induction heating in a mixture gas of oxygen and argon under a given
pressure. A silver rod set upward was moved down to the surface of the melt, and
then the rod was moved upward at the rate of 1.6 mm min1. The dimensions of the
resulting silver ingot were about 15 mm in diameter and 100 mm in height.
The shape of pores in porous silver is different from those in other lotus copper
and iron. Figure 6.7 shows an example of optical microscopic observation of the
structure of lotus-type porous silver fabricated by the mold casting technique,
indicating the cross section parallel and perpendicular to the solidification direction.
In the cross section parallel to the solidification direction, although characteristic
6.2 Fabrication of Lotus Silver by Oxygen 99
Fig. 6.6 Schematic drawing

of the apparatus of
Czochralski method for 1 2
fabricating lotus silver
(Reprinted with permission 3
from [7]. © 2005 Japan
Institute of Metals) 4
5
10
6
7
9
1.Transference mechanism
2.Silver rod
3.High pressure chamber
4.Alumina crucible
5.Inlet and outlet of oxygen and argon
6.Platinum crucible
7.Refractory
8. Tube for evacuation
9.Molten silver
10.Induction heating coil
cylindrical pores were observed to extend to the solidification direction, their

shapes are not straight and their diameters are not constant in the direction.
Similarly the pore shape is distorted, the distribution density of pores is not
homogeneous, and furthermore, the pore size distribution is broad in the cross
section of the perpendicular direction compared with other lotus metals. Such
particular features to porous silver are discussed on the basis of the binary equilib-
rium phase diagram. A schematic drawing of the binary equilibrium phase diagram
for silver–oxygen system is illustrated in Fig. 6.8, where Tm and Tn mean the
holding temperatures of silver melt in oxygen (1,573 K) and the gas-evolution
crystallization reaction temperature, respectively. On the assumption that the solid-
ification progresses in equilibrium state, oxygen concentration of the liquid silver
increases from A to B with decreasing temperature. Subsequently at Tn the reaction
of L ! (S) + G occurs. In this case the solid phase appears below Tn, expecting
that the pore shape would be similar to that of other lotus metals. However, the
experimental results are not the case. Thus, it is reasonable to consider that
the solidification progresses in nonequilibrium state by rapid cooling without
Fig. 6.7 Cross sections of lotus silver fabricated by the mold casting technique under 0.05 MPa of
oxygen and 0.15 MPa of argon. (a) Parallel to the solidification direction and (b, c) perpendicular
to the solidification direction at the distance (b) 10 mm and (c) 5 mm from the copper chiller (the
bottom). (d) Cross section perpendicular to the solidification direction of lotus copper (Reprinted
oxygen supply from the gas phase to the liquid phase and the oxygen concentration
of the liquid phase is constant from A to C. In this case unlike other lotus metals, the
solidification progresses through L + (S) phase: the solid phase coexists with the
liquid phase. This can prevent the growth of cylindrical pores similar to the case of
lotus magnesium alloys [8]. Thus, the solidification in the L + (S) phase could be a
part of reason for the distorted pore shape.
Figure 6.9 shows the cross section parallel and perpendicular to the solidification
direction of the specimen fabricated by the Czochralski technique. The distribution
density of pores fabricated by the Czochralski technique is more homogeneous than
that by the mold casting technique. In the Czochralski technique, the liquid phase
continuously contacts the gas phase during the solidification. Thus, it is expected
that oxygen can easily be supplied from gas phase to the liquid–solid interface and
A
Tm
Temperature
L+G
L
C
C’
(S) + L B
Tn (S)
(S) + G
Silver Oxygen
Oxygen Concentration
Fig. 6.8 Silver-rich side of binary equilibrium phase diagram for silver–oxygen system
(Reprinted with permission from [7]. © 2005 Japan Institute of Metals)
a b
Porosity 44%
5mm 1mm
c d
Porosity 41%
5mm 1mm
Fig. 6.9 Cross sections of lotus silver by the Czochralski method. (a) Parallel to the solidification
direction and (b) perpendicular to the solidification direction. Cross sections of the lotus silver
fabricated by the mold casting technique: (c) parallel to the solidification direction and (d)
perpendicular to the solidification direction. The atmosphere was a mixture of 0.55 MPa of oxygen
and 0.55 MPa of argon (Reprinted with permission from [7]. © 2005 Japan Institute of Metals)
then the porosity also must increase. The distribution of the pore size σ and the
circularity R at the cross section perpendicular to the solidification direction are
calculated as
σ ¼ σ s =da
R ¼ 4π A=L2 ; (6.4)
where σ s, da, A, and L are the standard deviation of pore sizes, the average pore
diameter, the area of pores, and the circumference length of the pores, respectively.
It turned out that the distribution of pore size σ decreases and the circularity R
increases in the lotus silver made by the Czochralski technique compared with the
ingot obtained by the mold casting method. These results support the assumption
that the oxygen gas in the pores formed at the liquid–solid interface redissolves to
the molten silver with insufficient oxygen concentration in the fabrication by the
unidirectional solidification method.
6.3 Fabrication of Lotus Intermetallic Compounds
Intermetallic compounds exhibit superior properties of high-temperature strength,

oxidation resistance, and anti-corrosion and are expected to be used as high-
temperature structural materials and functional materials. If elongated pores
aligned in one direction can be evolved in the intermetallic compounds, various
applications such as lightweight high-temperature structural materials, high-
temperature catalyst, and high-temperature filters can be expected. In particular,
aluminide intermetallic compounds have attracted much attention for lightweight
high-temperature materials [9–11]. Recently Hyun et al. [12] fabricated lotus-type
porous intermetallic compounds of Ni3Al and TiNi for the first time.
Figure 6.10 shows optical micrographs of the cross-sectional views of lotus-
type porous NiAl compounds perpendicular to the solidification direction fabricated
by the continuous zone melting technique. The fabrication was carried out at
2.5 MPa hydrogen and with the transfer velocity of 330 μm s1. The pores in
lotus Ni-15%Al alloy are round, while those in lotus Ni-28%Al and Ni-31%Al
alloys are in distorted morphology. The solidification structures in longitudinal
cross section parallel to the solidification direction are shown in Fig. 6.11. The
dendritic structure is observed in the Ni-28%Al alloy. When the pores nucleated
in apertures of the dendrite grow, the pores may touch to the dendritic arms
and cannot grow straight, which result in distorted pore morphology. On the other
hand, in Ni-15%Al alloy no dendrite structure is observed so that round-shaped
pores observed in the section perpendicular to the solidification direction grow
straight in the solidification direction. This is attributed to the narrow two
phases’ region of mixture of liquid and solid and to the suppressive growth of the
dendritic arms.
6.3 Fabrication of Lotus Intermetallic Compounds 103
Fig. 6.10 Optical micrographs of lotus (a, d) Ni-15%Al, (b, e) Ni-28%Al, and (c, f) Ni-31%Al
intermetallic compounds in the cross sections perpendicular to the solidification direction
fabricated by continuous zone melting technique in 2.5 MPa hydrogen atmosphere (Reprinted
Fig. 6.11 Optical micrographs of lotus (a) Ni-15%Al and (b) Ni-28%Al intermetallic compounds
in section parallel to the solidification direction (Reprinted with permission from [12]. © 2004
The porosity decreases from 40 to 18 % with increasing Al content from 15 to

31 %. The porosity is closely related with and proportional to the hydrogen
concentration rejected at the liquid–solid interface. According to Lange and
Schenck [13], the addition of aluminum to nickel suppresses the hydrogen solubil-
ity in the liquid. For example, the hydrogen solubility in Ni-25%Al alloy at 1,773 K
under 0.1 MPa hydrogen pressure is about 50 % smaller than that in pure nickel.
Therefore, the decrease in porosity may be due to the decrease in hydrogen content
available to the pore evolution. The average pore diameter decreases with increas-
ing Al content, which is also attributed to the same reasoning.
6.4 Fabrication of Lotus Silicon
Porous silicon with nano-order sized pores was fabricated by electrolyte reaction
[14–16]. However, porous semiconductors with unidirectional elongated micro-
order sized pores have never synthesized so far. Nakahata and Nakajima were
succeeded in fabricating a lotus silicon by unidirectional solidification in a
pressurized hydrogen atmosphere [17]. Different from the conventional nano-
sized porous silicon, the pore size of the lotus silicon ranges from 10 μm to 1 mm
in diameter. This type of porous silicon is considered to be novel porous semicon-
ducting material. The lotus silicon with directional cylindrical gas pores was
fabricated by unidirectional solidification through mold casting technique of the
melt dissolving hydrogen in a pressurized hydrogen atmosphere.
Figure 6.12a shows an overview of lotus silicon fabricated in the hydrogen
atmosphere of 0.21 MPa. Figure 6.12b, c shows the optical micrographs of the
cross section perpendicular and parallel to the solidification direction, respectively.
The pores in the cross section perpendicular to the solidification direction are
round-shaped, while those parallel to the solidification direction are long and
Fig. 6.12 (a) Overview and optical micrographs of cross section of lotus silicon overview, (b)
cross section perpendicular to the solidification direction, and (c) cross section parallel to the
solidification direction. The solidification occurred upward as shown in figure (c) (Reprinted with
permission from [17]. © 2004 Elsevier B.V.)
6.4 Fabrication of Lotus Silicon 105
a 40 b 350
Average pore diameter/μm

300
30 250
Porosity (%)
200
20
150
100
10
50
0 0
0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2
Hydrogen Pressure, PH2 / MPa Hydrogen Pressure, PH2 / MPa
Fig. 6.13 (a) Porosity of lotus silicon as a function of hydrogen pressure: ( filled circle) evaluated
from apparent density of lotus silicon and (open circle) determined by the microstructural analysis
of the cross sections at the distance 5 mm from the copper chiller. (b) Average pore diameter of the
lotus silicon as a function of hydrogen pressure. The closed circles represent the average pore
diameter determined by microstructural analysis of the cross sections at the distance 5 mm from
the copper chiller and the broken line shows the fitted one by linear regression (Reprinted with
permission from [17]. © 2004 Elsevier B.V.)
cylindrical. Figure 6.13a shows hydrogen pressure dependence of the porosity of

the lotus silicon. The porosity decreases with increasing hydrogen pressure.
Figure 6.13b shows hydrogen pressure dependence of the average pore diameter
in the cross section on the sample plane at 5 mm in height from the bottom of the
ingot. The pore diameter decreases with increasing hydrogen pressure; the pore size
is about 100 μm in diameter at hydrogen pressure of 0.99 MPa. From Sieverts’ law
and Boyle’s law, the following expression is derived:
1=2
ðkl ks Þ PH2 R Tn
V¼ ; (6.5)
2ðPH2 þ 2σ=rÞ
where kl and ks are the equilibrium constants of hydrogen dissolving reaction in

liquid and solid, respectively. σ and r are the surface tension of the melt and pore
radius. Since the surface tension of silicon at Tn (1,687 K) is about 0.74 Nm1 [18]
and the pore diameter in the range from 103 to 286 μm, the value of 2σ/r is
estimated at most to be 0.03 MPa, which is in fact negligible small in comparison
with the corresponding PH2 of 0.99 MPa. Thus, the pore volume is inversely
1=2
proportional to PH2 . The porosity increases by a factor of about 2.7 when the
hydrogen pressure decreases from 0.99 to 0.12 MPa, which is in good agreement
with the estimation by the above equation; the porosity increases by a factor of
(0.99/0.12)1/2 2.9. Although the present evaluation is not exact because of
neglecting some effect such as temperature gradient, the pore evolution behavior
can be explained by supersaturation of hydrogen in the solidified silicon.
6.5 Fabrication of Lotus Alumina by Unidirectional

Solidification
Porous oxide ceramics have been used as hot gas or molten metal filters and catalyst
because of their excellent mechanical properties, high-temperature resistance, and
chemical stabilities. Porous ceramics with highly oriented cylindrical pores are
convenient to use as a gas filter and as a catalyst carrier. Ishizaki et al[19] suggested
that the appropriate pore size for gas filter application is between 10 and several
hundred micrometer. Hence, in the development of a gas filter, control of the pore
size is an important issue. Several research groups have proposed to the fabrication
processes of porous alumina with oriented cylindrical pores. Zhang et al. [20]
fabricated porous alumina with unidirectionally aligned continuous pores via the
slurry coating of fugitive cotton filter. Porous alumina with approximately 150-μm-
diameter pores and 35 % porosity was fabricated using their method. Ding
et al. [21] prepared a porous alumina with oriented pores by combining a foaming
method with sol–gel technology. A porous alumina with 1-mm-diameter pores and
35 % porosity was fabricated from their method. Isobe et al. [22] prepared porous
alumina with oriented cylindrical pores by an extrusion method. In their method,
a porous alumina with 14-μm-diameter pores and 35 % porosity was fabricated.
Fukasawa et al. [23] have developed a fabrication method for porous alumina with
oriented pores using the freeze-drying technique. However, it is difficult to control
the porosity and/or pore size of the porous alumina by these methods.
On the other hand, in the field of metals, porous metals with cylindrical pores
called as lotus metals were proposed. The porous metals are fabricated by unidirec-
tional solidification under pressurized hydrogen gas, utilizing the hydrogen solubil-
ity gap in between liquid and solid. For such metals, the phase diagrams for
metal–hydrogen system have been established. However, data of hydrogen gas
solubility in oxides are hardly available. Ueno et al. [24] investigated lotus alumina
fabricated by unidirectional solidification under pressurized hydrogen gas to eluci-
date the formation mechanism of pores.
Feed rods were prepared by high-purity alumina powder (99.99 % purity). The
powder was mixed with a binder in water and a green rod was prepared by the slip-
casting method. After drying in air, calcination was performed at 1,473 K for 7.2 ks
in air, and then, a feed rod with 8 mm in diameter and 150 mm in length was
obtained. The solidification was carried out using an optical floating zone apparatus
under pressurized hydrogen gas where 100%H2 gas, 50%H2-Ar mixture gas, or
10%H2-Ar mixture gas was used as the environmental gas, as illustrated in
Fig. 6.14. A xenon lamp was used as an optical source. The light from the xenon
lamp was put on the focus by the elliptical mirror to set up the melting zone of the
feed rod. The top of the feed rod was hooked on the upper shaft and the bottom was
fixed to the lower shaft. The melting system was set up in a quartz tube. During the
unidirectional zone melting, the solidification rate and the rod rotation were fixed to
200 mm h1 and 20 rpm, respectively.
6.5 Fabrication of Lotus Alumina by Unidirectional Solidification 107
Pressure
valve Gas outlet
Elliptical mirror
Quartz tube
Seed rod
CCD Camera
Melting zone
Xenon lamp
Hydrogen
Pressure valve cylinder
Moving direction
Fig. 6.14 Schematic drawing of fabrication apparatus for lotus ceramics
Lotus alumina was successfully fabricated by unidirectional solidification under

a pressurized hydrogen atmosphere. Figure 6.15a, b shows transversal and longitu-
dinal cross-sectional views of the solidified sample prepared in 100%H2 gas under
the pressure of 0.8 MPa. Many cylindrical pores were aligned along the solidification
direction. Some large pores were seen in the inner periphery and in the center of the
sample. In the longitudinal cross-sectional view, the large pores possessed a spheri-
cal shape, which is quite different from the morphology of cylindrical pores. Hence,
it is considered that the formation mechanism of these two types of pores is different.
For the cylindrical pores, which are the dominant pores found in the figure, the shape
of the pores is very similar to that of lotus metals. Thus, this type of pore is formed at
the solid–liquid interface during the solidification due to a hydrogen solubility gap
between liquid and solid phases. One possible model for the formation of sphere-
shaped pores is trapping gas bubbles that are released from the solid–liquid interface
to the liquid phase by the solid phase during unidirectional solidification.
Figure 6.16 shows the total pressure dependence of the porosity in the samples
prepared in 100%H2, 50%H2-Ar, and 10%H2-Ar gases. For the samples prepared
in 100%H2, the porosity once increases with increasing the pressure up to 0.4 MPa
and then gradually decreases. A similar tendency was observed for lotus copper.
Yamamura et al. suggested that there are two factors controlling the porosity in the
formation of lotus metals: Sieverts’ law and Boyle’s law [25]. It is known
that the porosity of lotus metals is proportional to hydrogen partial pressure.
However, the pressure inside the pore is driven by environmental pressure
according to Boyle’s law. Hence, the porosity decreases with increasing total
pressure according to Boyles law. Thus, the porosity of lotus materials is deter-
mined by two contrary laws related to hydrogen partial pressure and total pressure.
Because Ar gas cannot be dissolved into liquid and solid metal phases, Ar gas in the
Solidification direction
1 mm
1 mm
Fig. 6.15 Transversal cross-sectional and longitudinal cross-sectional views of the solidified
sample prepared in mixed gas of 50%H2-50%Ar at 0.8 MPa (Reprinted with permission from
[24]. © 2007 The American Ceramic Society)
Fig. 6.16 Porosity changes

in the samples prepared in
100%H2, 50%H2-Ar, and
10%H2-Ar gases (Reprinted
with permission from [24].
© 2007 The American
Ceramic Society)
H2-Ar mixture gas affects only to increase the total pressure so that the effect of
Boyle’s law became dominant. The same reasoning is also applied to lotus alumina
as shown in Fig. 6.16.
Figure 6.17 shows the total pressure dependence of the average pore diameter of
the samples. The pore size decreases with increasing total pressure of the gas, which
is explained by Boyle’s law. On the other hand, since the absolute amount of
hydrogen gas increases with increasing hydrogen partial pressure, the pore volume
increases with increasing hydrogen partial pressure.
6.7 Effect of Microstructure on Pore Morphology in Lotus Magnesium Alloys 109
Fig. 6.17 Longitudinal

cross-sectional views of the
solidified sample prepared in
50%H2-Ar gas at 0.8 MPa
from [24]. © 2007 The
American Ceramic Society)
6.6 Fabrication of Lotus Brass by Solid-State Diffusion
It is not easy to fabricate lotus brass by unidirectional solidification method in gas

atmosphere. This is probably because of its very small hydrogen solubility difference
between solid and liquid of brass. Instead of solidification, lotus brass is fabricated
by utilizing solid-state diffusion of zinc into porous copper which is made by
unidirectional solidification in high pressure of hydrogen atmosphere [26]. The
vapor deposition method is adopted; the lotus copper sample was wrapped in a
tantalum foil and was annealed together with the zinc piece in a quartz tube with
vacuum of 1 104 Pa. During annealing, zinc was evaporated, deposited onto the
lotus copper, and diffused into the copper. After annealing, the samples were
observed by an optical microscope and their compositions were analyzed by an
electron probe microanalysis (EPMA).
Figure 6.18 is the result of the analysis by EPMA on the sample annealed at
953 K for 4.32 104 s. No zinc pieces remained in the quartz tubes after annealing.
All of zinc atoms evaporated and diffused into the samples. It was found that the
composition was almost constant between two neighboring pores on the cross
section parallel to pore direction; the zinc concentration is around 40 at.% and
almost uniform composition. Thus, the lotus brass was fabricated by zinc diffusion
into lotus copper.
6.7 Effect of Microstructure on Pore Morphology

in Lotus Magnesium Alloys
Magnesium is considered to be a rather new material as a lightweight structural metal.

Usually the magnesium alloys are used for high-mechanical strength and good
anti-corrosion resistance. As mentioned before, lotus magnesium can be produced
by unidirectional solidification through the mold casting technique [27]. However,
Fig. 6.18 Results of electron probe microanalysis in the samples annealed at 935 K for
4.32 104 s in a quartz tube with a zinc piece: (a) map of Zn Kα intensity, (b) map of Cu Kα
intensity, and (c) composition–distance profiles between two pores (Reprinted with permission
from [26]. © 2003 Japan Institute of Metals)
it was found by Hoshiyama et al. [28] that addition of alloying elements significantly
changes the pore shape and the porosity. Three kinds of alloys, Mg-9%Al-0.75%Zn
(AZ91D), Mg-3Al-1Zn (AZ31B), and Mg-3Al-0.25Zn (AZ91D), the latter of which is
diluted with magnesium and is abbreviated as (1/3)AZ91D, were used. Figure 6.19
shows photographs on the longitudinal sections of porous (1/3)AZ91D, AZ31B, and
AZ91D alloys. In the lower region near the bottom chiller of (1/3)AZ91D and AZ31B
alloys, elongated pore growth is observed, while almost all pores in AZ91D alloy are
spherical, not cylindrical. Comparing (1/3)AZ91D with AZ31B the pores in the former
alloy are longer than that in the latter, because total concentration of the alloying
elements is lower in the former. The number of pores and porosity decrease with
increasing concentration of the alloying elements Al, Zn, and Mn. This result is
explained in terms of the decrease in hydrogen solubilities in the magnesium alloys
due to addition of alloying elements [29]. Figure 6.20 shows Mg-Al equilibrium phase
diagram in magnesium-rich side. The temperature interval ΔT between liquidus and
solidus lines at each composition increases by 43.5 K, 52.5 K, and 150 K in order
6.7 Effect of Microstructure on Pore Morphology in Lotus Magnesium Alloys 111
Fig. 6.19 Photographs on longitudinal sections of lotus magnesium alloys. (a) (1/3)AZ91D, (b)
AZ31B, and (c)AZ91D alloys. Copper chiller was used for the bottom plate of the mold (Reprinted
with permission from [28]. © 2008 Freund Publishing House Ltd)
Fig. 6.20 Magnesium-rich side of Mg-Al equilibrium phase diagram (Reprinted with permission
from [28]. © 2008 Freund Publishing House Ltd)
of (1/3)AZ91D, AZ31, and AZ91D, respectively; ΔT corresponds to increase in

coexistence region of solid and liquid phases.
Let us think about pore growth mechanism in such mushy zone. When the alloy
is solidified by way of the mushy zone, the ratio of solid phase portion in the mushy
zone increases with decreasing temperature and the hydrogen rejected in the solid
phase forms the pores. The schematic drawing of wide mushy zone is illustrated in
Fig. 6.21a. Even if the pores are nucleated, the solid phase which is precipitated in
front of evolved pores prevents the pores from growing toward the solidification
direction. On the other hand, in the short mushy zone in Fig. 6.21b, the top of the
cylindrical pores can grow without the precipitation of solid phase. Thus, it is
concluded that the shorter the mushy zone is, the longer the pores grow.
Fig. 6.21 Effect of the mushy zone on pore growth. Width of the mushy zone is (a) wide and (b)
narrow. The mushy zone affects the pore evolution in unidirectional solidification (Reprinted with
permission from [28]. © 2008 Freund Publishing House Ltd)
6.8 Fabrication of Lotus Carbon Steel by Continuous

Casting Technique
Lotus carbon steel is suitable for lightweight high-strength structural material. For
practical purpose, lotus carbon steel plates are expected to be used for movable
structural bodies of machine tools because of the effects of reduction of weight and
vibration damping, which improve the precision and productivity of machining, the
tool life, etc. Although the continuous casting apparatus was used to fabricate lotus
copper, magnesium, and aluminum, some modifications of the apparatus were
required for fabrication of lotus carbon steel, because there are several problems as
follows. Since iron reacts easily with carbon, nitrides, and so on, endurance of the
refractory, a ceramic material which is not reactive with iron, was selected as the
refractory. In order to improve the endurance of the refractory, a technique for a rigid
connection among the crucible, the break ring, and the mold was developed [30]. The
second problem was the friction between the mold and the carbon steel, which causes
the blockage of melt in the mold during continuous casting. To solve this problem,
intermittent motion was adopted, where a cycle of transfer and stop is repeated
periodically. This technique can give a vibration to the carbon steel sample in order
to deduce the friction. In the following sections, transfer velocity means the average
value, which is the transfer distance per unit time over several cycles as shown in
Fig. 6.22. By developing this technique, it became possible to fabricate long lotus
carbon steel up to 600 mm in length.
Figure 6.23 shows the fabricated lotus carbon steel at 100 mm min1. On the
surface, ripple marks are observed periodically. These ripple marks were generated
at each cycle of the intermittent motion and correspond to the solid–liquid interface.
The ripple marks show that the solid–liquid interface changed from horizontal to
concave shape with a depth h during the continuous casting. The change can be
explained by the position of interface, which is considered to have dropped during
continuous casting, because cooling capacity of the mold was not enough to solidify
the total amount of the melt, which entered during one cycle. The length of the
6.8 Fabrication of Lotus Carbon Steel by Continuous Casting Technique 113
Fig. 6.22 Intermittent

motion of transfer (Reprinted
with permission from [30].
© 2007 MIT-Verlag)
Ripple Solid/liquid
mark interface
h Molten
steel
0
-x
Number of ripple marks N
Length of solidified part L
h vN
L
Ripple
mark
Lotus
carbon
steel
10mm
Fig. 6.23 Fabricated lotus carbon steel. (left) Ripple mark on the surface of solidified slab of lotus
carbon steel. (right) Schematic drawing around the solid/liquid interface during continuous casting
(Reprinted with permission from [30]. © 2007 MIT-Verlag)
solidified part L until the Nth cycle (see Fig. 6.22) was measured by the ripple mark.
Then, the position of the solid/liquid interface can be evaluated by the difference
between L and vN, which is the moving distance of the dummy bar, where v is a
moving distance per a cycle.
Figure 6.24 shows the measured position and depth of the solid–liquid interface
at each cycle in the experiment of 100 mm min1. At the beginning of the
continuous casting, the position is close to the crucible. The small h-value
shows that the interface is nearly horizontal. In this case small pores are distributed.
Number of cycle, N
0 20 40 60 80 100 120
0
-10
Position of S/L interface, -20
L
(a)
Edge
-30
x / mm
-40 S
L
-50 S
-60 Center
(b)
h
-70
-80
0 200 400 600
Moving distance of dummy bar, vN/mm
a b
T.D.
5mm
Fig. 6.24 Position of the solid–liquid interface at each cycle number. The pictures show the cross
sections parallel to the transfer direction at the initial transference (a) and at about 80 cycles (b).
The dotted lines show the solid–liquid interface (Reprinted with permission from [30]. © 2007
MIT-Verlag)
With increase of the cycle number, the position dropped and the interface became a
deep concave shape with a large value of h. In this case coarsened irregular pores
are observed.
The mechanism of these results is illustrated in Fig. 6.25. In the initial state (a),
the melt is cooled mainly by the heat flow in the longitudinal direction, because
this is a heating zone. Then the solid–liquid interface is horizontal. But if the
solid–liquid interface is located in the cooling zone (b), the heat flow is mainly in
the width direction and the interface becomes concave. As the pores grow perpen-
dicular to the solid–liquid interface, the pores merge with each other in the center.
Coarsened irregular pores are generated.
From these results the mechanism of pore formation depending on the transfer
velocity became clear as shown in Fig. 6.26. Then one can find the condition for
small pores distributed homogeneously at around 20 mm min1.
6.9 Fabrication of Lotus Aluminum by Continuous Casting Technique 115
a b
Heating
zone
Cooling
zone
Fig. 6.25 Mechanism of the descent position of the solid–liquid interface and he change into a
concave shape. (a) The initial transference and (b) after several ten cycles (Reprinted with
permission from [30]. © 2007 MIT-Verlag)
low Transference Velocity high
Molten
steel
S/L
interface
Pore
small Super cooling large
small Depth of curvature large
Fig. 6.26 Schematic illustration of the dependence of pore formation on transfer velocity
(Reprinted with permission from [30]. © 2007 MIT-Verlag)
6.9 Fabrication of Lotus Aluminum by Continuous

Casting Technique
Lotus aluminum possessing slender directional pores aligned in one direction is one
of the most promising candidates for lightweight structural materials [31]. Numer-
ous investigations have focused on the fabrication of porous aluminum and its
alloys with slender directional pores formed by unidirectional solidification in a
hydrogen atmosphere. Particularly, solidification defects have been emphasized for
Table 6.1 Previously reported porosity and pore size results in lotus aluminum fabricated in
hydrogen atmosphere
Materials Atmosphere Pore diameter First author
(mass%) (H2/MPa) (μm) Porosity (%) (year) [ref]
Al (99.8%pure) 0.1 <1.6 Shinada (1980) [32]
Al 0.05 40-60 0.2 Shahani (1985) [33]
0.1 50-60 0.9
0.3 125-150 0.6
0.5 150-200 1.1
Al 0.11 difficult to produce regular Shapovalov (1993) [34]
high-porosity Al
Al-(2-8)%Fe 0.11 Al3Fe precipitates promote
pore formation
Al 0.1 60 5.1 Zhang (2007) [35]
0.18H2+0.22Ar 75 <0.1
air 60 0.1
air 60 0.3
basic and applied research of a lightweight material. Table 6.1 is a compilation of

these works [32–35]. None of the previous works produced porous aluminum with
directional pores and a high porosity; the porosity is limited to 5 % at most.
Lotus aluminum can be fabricated utilizing the hydrogen solubility gap between
a liquid and a solid at the melting point in the unidirectional solidification process in
a hydrogen atmosphere. When the melt dissolving hydrogen is solidified, the
insoluble hydrogen is precipitated to evolve directional pores in the solidified
region at the solid–liquid interface. Therefore, the following conditions are critical
to fabricate highly porous lotus metals.
1. The hydrogen concentration dissolved in the molten metal should exceed several
atomic percent.
2. The metal should hold a smaller solid solubility of hydrogen so that the solubility
gap between liquid and solid metals becomes larger.
However, the hydrogen solubility in liquid and solid aluminum [36] is only
about 2 % of those for magnesium, copper, and transition metals [37, 38]. Conse-
quently, previous researchers considered the fabrication of highly porous lotus
aluminum difficult; none of the previous reports successfully fabricated lotus
aluminum with a porosity above 5 %.
The influence of solidification conditions (solidification velocity, hydrogen
partial pressure, temperature gradient, and melt temperature) on pore formation
of lotus aluminum was experimentally and theoretically investigated by Ide et al.
[39] to elucidate the pore formation mechanism as well as to control porosity. They
fabricated lotus aluminum with slender unidirectional pores using a continuous
casting technique under controlled solidification conditions. In particular, it was
clarified that the porosity and the pore size are affected significantly by the
solidification velocity as the velocity decreases the porosity and pore size by one
to two orders of magnitude compared to those for copper, magnesium, and transi-
tion metals. Based on this knowledge, lotus aluminum with a porosity as high as
40 % was obtained for the first time.
A continuous casting technique was utilized for unidirectional solidification in a
pressurized hydrogen atmosphere under controlled solidification conditions. The
temperature of the melt in the crucible was measured using two WRe 5/26-type
thermocouples located 5 mm from the bottom of the crucible. Then to solidify the
melt in a continuous downward direction, a graphite dummy bar pulled molten
aluminum through a cooling mold at a constant solidification velocity. The solidifi-
cation temperature using two K-type thermocouples inserted in the graphite dummy
bar. To ensure the temperature gradient was at a steady state, the temperature was
measured 20 mm above the top of the dummy bar. The measured temperature was
recorded every second using a temperature recorder (Keyence Co., Ltd.; GR-3500).
The temperature gradient was calculated using the obtained cooling curve in the
temperature range from 933 to 873 K (temperaturetime curve). The temperature
gradient was determined using the following equation:
V
G¼ , (6.6)
R
where G, V, and R are the temperature gradient (K mm1), cooling rate during
unidirectional solidification (K s1), and solidification velocity (mm s1), respec-
tively. The temperature gradient was determined from the average of two tempera-
ture gradients, whose difference was less than 0.3 K mm1).
The effects of different characteristics on pore morphology were examined. The
influence of the solidification rate was probed by melting pure aluminum and
unidirectionally solidifying in a mixture of hydrogen 0.25 MPa and argon
0.25 MPa with a constant temperature gradient and melt temperature of
9.7 K mm1 and 1,223 K, respectively, while changing the solidification velocity
from 0.5 to 0.9 mm min1. The impact of hydrogen partial pressure was
investigated by melting aluminum and unidirectionally solidifying in three different
atmospheres: hydrogen 0.5 MPa, argon 0.5 MPa, or mixed gas (0.5 MPa) consisting
of hydrogen (0.25 MPa) and argon (0.25 MPa) when the solidification velocity,
temperature gradient, and melt temperature were set to 0.9 mm min1, 9.5 K mm1,
and 1,223 K, respectively.
Figure 6.27 shows the typical pore morphology perpendicular (upper row) and
parallel (lower row) to the solidification direction of lotus aluminum as a function
of solidification velocity from 0.5 mm min1 to 0.9 mm min1. Lotus aluminum
was fabricated by unidirectional solidification in a mixture of hydrogen (0.25 MPa)
and argon (0.25 MPa). Figure 6.27f shows an outer view of lotus aluminum with the
highest porosity of nearly 40 %. Unidirectional pores aligned parallel to the
solidification direction are observed in the solidified lotus aluminum. The porosity
and average pore diameter decrease as the solidification velocity increases.
Figure 6.28 shows the pore morphology in lotus aluminum fabricated by unidi-
rectional solidification at a solidification velocity of 0.9 mm min1 as a function of
Fig. 6.27 Cross sections perpendicular (upper row) and parallel (lower row) to the solidification
direction of lotus aluminum. Solidification velocity: (a) 0.5 mm min1, (b) 0.6 mm min1, (c)
0.7 mm min1, (d) 0.8 mm min1, and (e) 0.9 mm min1. (f) Outer view of lotus aluminum with
the highest porosity of nearly 40 %
direction of lotus aluminum fabricated in different atmospheres: (a) argon (0.5 MPa), (b) mixed
gas of hydrogen (0.25 MPa) and argon (0.25 MPa), and (c) hydrogen (0.5 MPa). Temperature
gradient and melt temperature are 9.5 K mm1 and 1,223 K, respectively
direction of lotus aluminum fabricated with different temperature gradients at the solid–liquid
interface. Lotus aluminum is fabricated in a mixed gas of hydrogen (0.5 MPa) and argon (0.5 MPa)
at a melt temperature of 1,373 K and a fixed solidification velocity of 0.5 mm min1
the hydrogen partial pressure; photos in the upper and lower rows are the cross-
sectional views perpendicular and parallel to the solidification direction, respec-
tively. The atmospheric total pressure was kept constant at 0.5 MPa, while the
partial pressure of hydrogen was either zero, 0.25 MPa, or 0.5 MPa and the partial
pressure of argon was 0.5, 0.25 MPa, or zero, respectively. The ingot solidified in an
argon atmosphere does not exhibit pores (Fig. 6.28a). In contrast, ingots prepared in
a hydrogen atmosphere contain pores (Fig. 6.28b, c). It is apparent that dissolving
gas in the melt is responsible for the evolution of pores, because hydrogen but not
argon can dissolve in molten aluminum. Porosity and pore diameter increase as the
hydrogen partial pressure increases.
Figure 6.29 shows cross-sectional views of the pore morphology perpendicular and
parallel to the solidification direction of lotus aluminum fabricated at different temper-
ature gradients. The porosity and average pore diameter decrease as the temperature
gradient increases. Figure 6.30 shows the change in pore morphology due to the
difference in the melt temperature of lotus aluminum fabricated with a solidification
velocity of 0.9 mm min1 in a 0.5 MPa hydrogen atmosphere. The porosity and
average pore diameter increase as the melt temperature increases.
As mentioned above, the porosity of lotus aluminum depends on not only hydrogen
partial pressure but also the solidification velocity, temperature gradient, and melt
Fig. 6.30 Dependence of porosity and pore diameter on melt temperature of lotus aluminum
fabricated with solidification velocity of 0.9 mm min1 in 0.5 MPa hydrogen. Temperature
gradient is 9.5 0.5 K mm1
temperature. These dependencies are similar to those of dual-phase formation in

eutectic alloys; the flux of solute atoms due to solidification conditions affects the
morphology of the phases. This suggests that a change in the hydrogen flux by different
solidification conditions significantly affects pore formation in lotus aluminum.
The solubility difference of hydrogen in aluminum is more than one order of
magnitude smaller than that in copper, whereas the diffusion coefficients are similar
in both aluminum and copper. To promote pore growth in lotus aluminum, hydro-
gen in aluminum should diffuse over a longer distance than in copper because the
solubility of hydrogen is lower in aluminum. Figure 6.31 schematically depicts the
effect of a long diffusion distance of hydrogen rejected in a solid in the vicinity of
the solid–liquid interface on pore evolution in lotus aluminum. Assuming the
diffusion distance of hydrogen increases, the porosity and pore diameter in lotus
aluminum increase as the number of hydrogen atoms increases, resulting in the
formation and growth of pores. Thus, the mechanism for pore evolution in lotus
aluminum may be the opposite of the mechanism for lotus copper and stainless steel
where the formation and growth of pores easily occur even for short distant
diffusion of hydrogen due to the high content of hydrogen rejected in the solid.
Therefore, the condition of a lower solidification velocity is certainly crucial to
realize highly porous lotus aluminum.
The porosity and pore diameter decrease as the solidification velocity and
temperature gradient increase, but increase as the hydrogen partial pressure and
Fig. 6.31 Two-dimensional model for the growth of unidirectional pores near the liquid–solid
interface. Hydrogen rejected in solidified aluminum accumulates in the liquid near the liquid–solid
interface. (a) When the solidification velocity is fast, hydrogen diffuses only a short distance.
Consequently, the impact of hydrogen on growth is small, leading to small pores and a low
porosity. (b) When the solidification velocity is slow, hydrogen diffuses over a longer distance.
Because a large amount of hydrogen contributes to growth, large pores and a high porosity
are obtained
melt temperature increase. Previous works have reported the effects of solidifica-
tion velocity and hydrogen partial pressure on pore formation of lotus metals.
However, the influence of solidification velocity on pore formation in lotus alumi-
num differs from previous reports using other lotus metals such as copper [25]
and stainless steel [40]. Similar to the fabrication of lotus copper or stainless
steel, the pore diameter decreases as the solidification velocity increases. However,
unlike for lotus aluminum, the porosity for lotus copper or stainless steel is
independent of solidification velocity. Such a difference is attributed to the hydro-
gen solubility between the liquid and solid phases. For example, the hydrogen
solubility gap between the liquid and solid phases in aluminum (4.93 104 mol%
under H2 0.1 MPa [36]) is about 40 times smaller than that of copper
(2.09 102 mol% at H2 0.1 MPa [37]), but there is not a significant difference
in the diffusion coefficients [41, 42]. Therefore, the supersaturated hydrogen atoms
in aluminum have to migrate a relatively long distance toward the pores to contrib-
ute to pore formation and growth. For a fast solidification velocity, the hydrogen
atoms in aluminum cannot migrate a sufficient distance to grow large pores. On the
other hand, for a slow solidification velocity, the hydrogen atoms in aluminum can
migrate a longer distance, allowing more supersaturated hydrogen atoms to con-
tribute to pore formation and growth, producing a larger porosity.
Next, we discuss effects of a hydrogen flux, which depends on the solidification
conditions, on pore formation in lotus aluminum. In eutectic alloys, the morphology
of a lamellar or rod structure is affected by solidification conditions. Such a change
is due to the change in the flux of the solute. For a binary system composed of metal
and hydrogen, we consider a dual phase, which consists of a combination of pores
and aluminum. Figure 6.32 shows a two-dimensional model for the competitive
Fig. 6.32 Schematic Z

drawing of the diffusion flux
of hydrogen and aluminum
during pore evolution and
growth in the solidification L
front Flux of hydrogen
Flux of aluminum
λ/2
Pore Aluminum
Y
0 fλ/2 λ/2
growth between directional pores and solid aluminum under steady-state

conditions. Ide et al. [39] treated the dual phase in two dimensions, and due to
symmetry, only half of a lamellae for each phase needs to be considered [43]. As
shown in Fig. 6.32, aluminum rejects hydrogen atoms due to the solubility gap in
the melt, whereas the pore rejects aluminum atoms during unidirectional solidifica-
tion. As well as the solute distribution in the eutectic structure, the mass of diffusing
hydrogen (aluminum) atoms into pore (aluminum) per unit of time can be obtained
by solving the diffusion equation for solute in the diffusion layer at the solidifica-
tion front. Therefore, when the z-axis is parallel to the solidification direction and
the y-axis is perpendicular to the solidification direction in the yz coordinate
system (Fig. 6.32), the distribution of the solute concentration at the solidification
front can be expressed by [43]

Vz 2πz 2πy
C ¼ C0 þ A exp þ B exp cos , (6.7)
D λ λ
where C0 and D are concentration (mol) and diffusion coefficient (m2 s1) of solute
in molten metal, respectively, V is the growth velocity (m s1), and λ is the distance
between periodic phase structure (m). Here A and B are, respectively,

A ¼ f CβS CαS þ CE CβS , (6.8)

f ð1 f ÞVλ CβS CαS
B¼ , (6.9)
2Dsin ðπf Þ
where f is the volume fraction of theαphase, CSα and CSβ are the solid solubility limit
of B atoms in the α and β phases (mol m3), respectively. When B atoms are
rejected by the α phase and diffuse into the β phase through y = fλ/2, the flux of B
atoms, J (mol s1 m2) is obtained by
2πz
∂C β α
J ¼ D ¼ πf ð1 f ÞV CS CS exp : (6.10)
∂y y¼ f λ λ
2
Here, the width of the inspissated solute layer (diffusion layer) is considered to
be about λ/2. When a small distance dl is solidified, the amount of diffusing
hydrogen, Cdl (mol m1), into the nearby phase through the small Δl is given by
ð λ=2
dl
Jdl ¼ Jdz ¼ f ð1 f Þ CβS CαS λdl, (6.11)
0 V
where α and β are the aluminum and pore phases, respectively, and A and B atoms
denote aluminum and hydrogen atoms, respectively. Cα and Cβ are the hydrogen
solubility in solid aluminum and the gas phase (mol m3), respectively. Because
solid aluminum is not dissolved in the gas phase, the molar ratio of hydrogen is
equal to 1. Therefore, Cα and Cβ are expressed as
ηTn pffiffiffiffiffiffiffiffi
Cα ¼ PH2 , (6.12)
VAl
2RTn pffiffiffiffiffiffiffiffi
Cβ ¼ PH 2 : (6.13)
P
Here the volume fraction in the equilibrium phase diagram is used as the volume
fraction f. However, because the porosity of lotus aluminum depends on the
solidification conditions, the volume fraction of aluminum phase f is obtained
from the experimental results. In the present work, Yamamura’s model for mass
balance of lotus copper [25] is modified by considering hydrogen diffusion, i.e., the
terms indicating the mass of hydrogen are substituted using Eqs. (6.9)(6.11).
[Initial amount of hydrogen contained in the liquid in the volume element]
+ [Amount of hydrogen contained in the liquid flowing into the volume element to
compensate for solidification shrinkage]
– [Amount of hydrogen contained in the liquid flowing out of the volume element
due to pore formation]
+ [Amount of hydrogen dissolved in solid copper around the pore with a length
of l + dl].
In Yamamura’s model, the hydrogen mass balance for a unit length as the
solidification front advances dl is considered. Because the number of pores per
unit length and supplied hydrogen from both sides of the pores have to be considered
as the solidification front advances dl, the mass of diffusing hydrogen is written as
2 M
WH ¼ f ð1 f Þ Cβ Cα λ dl (6.14)
λ 2
a d
30
40
25
30 20
Porosity / %
Porosity / %
15
20
10
10 5
0
0
0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.00 0.25 0.50
Transfer velocity / mm min-1 Hydrogen partial pressure / MPa
c b
60 35
50 30
25
40
Porosity / %
Porosity / %
20
30
15
20
10
10 5
0 0
1160 1200 1240 1280 6 8 10 12 14 16
Melt Temperature / K Temperature gradient / K mm-1
Fig. 6.33 Dependence of porosity on (a) solidification velocity, (b) hydrogen partial pressure,
(c) melt temperature, and (d) temperature gradient. Closed and open circles indicate experimental
data and the calculated results using Eq. (6.16), respectively. Dotted lines are the predicted results
using Yamamura’s model
In the present work, λ is given by following equation using porosity ε (%) where
pore diameter dp (m) given by experimental results:
dp
λ¼ : (6.15)
ε
Equation (6.14) is substituted in Yamamura’s model and the porosity of lotus
metal is expressed as
pffiffiffi
fMð1 f Þ Cβ Cα þ ηTn ηTn Gl P 2VMAl
ε¼ MP
pffiffiffi (6.16)
RðTn GlÞ þ ηTn ηTn Gl P 2VMAl
Figure 6.33 shows the dependence of porosity on (a) solidification velocity,

(b) hydrogen partial pressure, (c) melt temperature, and (d) temperature gradient.
References 125
To compare the experimental results and the calculated results predicted using Eq.
(6.16) with Yamamura’s model [25], closed circles, open circles, and the dotted line
denote the experimental results, calculated results using Eq. (6.16), and the results
predicted by Yamamura’s model, respectively. The calculated porosities using Eq.
(6.16) agree well with the experimental results. In Yamamura’s model, the porosity
was calculated by assuming that all of the rejected hydrogen atoms due to the
solubility gap form pores; parameter a = 1. Although the dependence of the
porosity calculated from Yamamura’s model on the hydrogen partial pressure is
consistent with the experimental results, the dependence of the porosity for the
other factors using Yamamura’s model is inconsistent with the experimental results.
These observations are reasonable because Yamamura’s model does not consider
the effect of the solidification condition.
Moreover, even if all of hydrogen atoms rejected in the solidified metal evolve
into pores, the maximum porosity calculated by Yamamura’s model is less than
15 %, which is far below the maximum experimental porosity of 40 %. This
discrepancy suggests that the contribution of hydrogen diffusion in the melt
rejected in the solid phase near the solid–liquid interface is critical. Thus, control
of solidification conditions such as the solidification velocity, hydrogen partial
pressure, temperature gradient, and melt temperature is crucial to increase the
porosity of lotus metal with a low hydrogen solubility.
References
1. Simmons JW (1996) Mater Sci Eng A 207:159–169

2. Hyun SK, Nakajima H (2002) Mater Trans 43:526–531
3. Satir-Kolorz AH, Feichtinger HK (1991) Z Metallkde 82:689–697
4. Levinsky Y (ed) (1997) Pressure dependent phase diagrams of binary alloys. ASM Interna-
tional, Materials Park, p 693
5. Raghaban V (ed) (1987) Phase diagram of ternary iron alloys. Indian Inst Tech, Delhi, p 143
6. Shah ID, Parlee NAD (1967) Trans AIME 239:763–764
7. Nakahata T, Nakajima H (2005) Mater Trans 46:587–592
8. Hoshiyama H, Ikeda T, Murakami K, Nakajima H (2003) J Jpn Inst Metals 67:714–720
9. Hirano T (1990) Acta Metall 38:2667–2671
10. Yamaguchi M, Inui H, Ito K (2000) Acta Mater 48:307–322
11. Deevi SC, Sikka VK (1996) Intermetallics 4:357–375
12. Hyun SK, Ikeda T, Nakajima H (2004) J Jpn Inst Metals 68:39–42
13. Lange KW, Schenck H (1969) Z Metallkd 60:638
14. Turner DR (1958) J Electrochem Soc 105:402–408
15. Unagami T, Seki M (1978) J Electrochem Soc 125:1339–1344
16. Cullis AG, Canham LT (1991) Nature 353:335–338
17. Nakahata T, Nakajima H (2004) Mater Sci Eng A 384:373–376
18. Przyborowski M, Hibiya T, Eguchi M, Egry I (1995) J Cryst Growth 151:60–65
19. Ishizaki K, Komarneni S, Nanko M (1998) Porous materials: process technology and
applications, Materials technology series. Kluwer, Dordrecht
20. Zhang GJ, Yang JF, Ohji T (2001) J Am Ceram Soc 84:1395–1397
21. Ding XJ, Zhang JZ, Wang RD, Feng CD (2002) J Eur Ceram Soc 22:411–414
22. Isobe T, Tomita T, Kameshima Y, Nakajima A, Okada K (2006) J Eur Ceram Soc 26:957–960
23. Fukasawa T, Deng ZY, Ando M, Ohji T (2001) J Ceram Soc Jpn 109:1035–1038
24. Ueno S, Lin LM, Nakajima H (2008) J Am Ceram Soc 91:223–226
25. Yamamura S, Shiota H, Murakami K, Nakajima H (2001) Mater Sci Eng A318:137–143
26. Aoki T, Ikeda T, Nakajima H (2003) Mater Trans 44:89–93
27. Ikeda T, Nakajima H (2002) J Jpn Foundry Eng Soc 74:812–816
28. Hoshiyama H, Ikeda T, Nakajima H (2007) High Temp Mater Process 26:303–316
29. Watanabe T (1976) Light Metals 25:167
30. Kashihara M, Yonetani H, Suzuki S, Hyun SK, Kim SY, Kawamura Y, Nakajima H (2007)
Porous metals and metallic foams. MIT, Boston, pp 201–204
31. Nakajima H (2010) Proc Jpn Acad Ser B 86:884–899
32. Shinada H, Nishi S (1980) J Jpn Inst Light Metals 30:317–323
33. Shahani H, Fredriksson H (1985) Scand J Metall 14:316–320
34. Shapovalov VI, Timchenko AG (1993) Phys Met Metall 76:335–337
35. Zhang H, Li Y, Liu Y (2007) Acta Metall Sinica 43:11–16
36. Qiu C, Olson GB, Opalka SM, Anton DL (2004) J Phase Equilib Diff 25:520–527
37. Fromm E, Gebhardt E (1976) Gases and carbon in metals. Springer, Berlin
38. San-Martin A, Manchester FD (1987) Bull Alloy Phase Diagrams 8:431–437
39. Ide T, Iio Y, Nakajima H (2012) Metall Mater Trans A 43A:5140–5152
40. Ikeda T, Aoki T, Nakajima H (2005) Metall Mater Trans A36:77–86
41. Papp K, Csetenyi EK (1981) Scr Metall 15:161–164
42. Wright JH, Hocking MG (1972) Metall Trans 3:1749–1753
43. Kurz W, Fisher DJ (1998) Fundamental of solidification. Trans Tech Publications, Switzerland
Chapter 7
Mechanical Properties of Lotus
Metals and Alloys
Abstract Many conventional porous materials possess almost nearly spherical

pores and exhibit that those mechanical properties are usually isotropic. Moreover,
the shape of pores is not always round and distorted so that the stress concentration
takes place easily, which causes to weaken the strength of those materials. In
addition, the overall uniformity of the pore size and the porosity is very crucial
which may affect the stress concentration; some local strain under the stress is
accumulated into the nonuniform region to degrade the materials strength. Thus, the
foamed materials, cellular-structured materials, or sintered materials exhibit infe-
rior mechanical properties.
Different from such conventional porous materials, lotus materials have elon-
gated cylindrical pores aligned unidirectionally, and various mechanical properties
are remarkably different from those of isotropic porous materials and show signifi-
cant anisotropic behavior. Although the data on the mechanical properties of such
anisotropic porous materials are not so much accumulated compared with the
isotropic porous materials, it is at present good enough to review some systematic
view of the mechanical properties of the lotus metals.
Keywords Compressive strength • Elastic constants • Fatigue strength • Internal

friction • Tensile strength
7.1 Elastic Properties
Tane et al. measured the anisotropic elastic constants of lotus iron fabricated using
the continuous zone melting technique [1, 2], extended the effective-mean-field
(EMF) theory so as to take account of the pore-orientation effect on the effective
elastic constants, and applied the extended EMF theory to lotus iron to validate the
theory as a prediction method of the elastic properties of lotus metals. They
prepared two kinds of lotus iron, containing hydrogen pores or nitrogen pores,
with various porosities. The specimens were cut out of the ingots and machined into

128 7 Mechanical Properties of Lotus Metals and Alloys
the rectangular parallelepipeds with the surfaces normal to the pore (solidification)
direction.
When the material possesses orthorhombic symmetry, there are eight indepen-
dent groups of free vibrations: OD(dilatation), EV(torsion), OX, OY, OZ (shear),
and EX, EY, EZ (flexure) [3, 4]. The resonance frequencies of the vibration modes
depend on the mass density, dimensions, and elastic constants of the specimen.
With the known elastic constants, the resonance frequencies can be calculated using
the measured density ρ and dimensions:
ð
1 @ui @uj @uk @ul
ρω2 ¼ Cijkl þ þ dV; (7.1)
V4 @xj @xi @xl @xk
where Cijkl is an elastic constant tensor in the four-index notation and ω ¼ 2πf is the
angular frequency, with f the resonance frequency. Equation (7.1) is computed by
the Rayleigh–Ritz method, in which the displacement vector u is expressed by the
linear combination of the basis functions {ϕ}:
u ¼ ðap ϕp ; aq ϕq ; ar ϕr Þ: (7.2)
Products of the Legendre functions are often chosen for {ϕ} [3]. It is not
straightforward to find the elastic constants from the resonance spectrum. They
are determined inversely through iterative calculations. First, the resonance
frequencies are measured using two piezoelectric transducers; one sends continuous
waves into the specimens and the other detects the signals to pick up resonance
peaks. Next, the resonance frequencies are calculated using the assumed elastic
constants and are compared with the measured resonance frequencies. If a good
agreement between the two spectra is achieved, the assumed values are regarded as
the true values. During this procedure, one should know an exact correspondence
between the resonance peaks and the vibration modes. For this, one has employed
the mode-selective electromagnetic acoustic resonance method (EMAR), which
can excite and detect only one vibration group among OX, OY, OZ, and OD
vibration groups. Figure 7.1 shows the excitation, as an example, of OX group by
EMAR. The resonance frequencies have been measured at room temperature within
the frequency range of 50–300 kHz at 0.01-kHz steps.
Figure 7.2 shows the porosity dependence of the Young’s moduli E⊥ and Ek and
the elastic stiffness c11, c33, c13, c44, and c66 of lotus iron with hydrogen or nitrogen
pores, where Ek and E⊥ indicate Young’s moduli in the directions parallel and
perpendicular to the pore direction, respectively. As well as the porosity depen-
dence of other lotus metals [5, 6], Ek decreases linearly, while E⊥ drops steeply in
the small porosity region. On the other hand, Hyun et al. have found that the
mechanical strength of lotus iron prepared in nitrogen atmosphere is much higher
than that in hydrogen atmosphere owing to the solid solution hardening by the
solute nitrogen [7]. Regarding its elastic properties, however, any significant
modifications was not observed [8].
7.1 Elastic Properties 129
Fig. 7.1 Excitation of the OX vibration group by the EMAR method. (a) A specimen is inserted
into the coil and the magnetic field is applied. (b) Eddy currents on the specimen surface are
induced by RF current in the coil, and the shearing Lorentz force is generated. (c) The Lorentz
forces deform the specimen as shown in thick arrows. The displacement u2 is an even function, for
example, against the x2 mirror plane. (d) The parity of the displacements for the OX group. The
displacement u2 in (c) only satisfies the OX parity (Reprinted with permission from [5] © 2002
Elsevier Science Ltd)
a b c
Young's modulus, E (GPa)
300 100
200 E⊥ c11 c44
c11 , c33, c13 (GPa)
250 80
E// c33 c33
c44 , c66 (GPa)
150 E// c66

200 c13 60 c44
100 E⊥ 150 c11
40 c66
100 c13
50 20
50
0 0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Porosity , p Porosity , p Porosity , p
Fig. 7.2 Porosity dependence of (a) two Young’s moduli, Ek in the direction parallel to the x3-axis
and E⊥ in the direction perpendicular to X3, and the elastic stiffness coefficients, (b) c11, c33, c13,
(c) c44 and c66 of lotus iron. The transverse isotropy condition c66 ¼ ðc11 c12 Þ=2 holds. Each line
is obtained by fitting the equation (7.3) to the measurements (Reprinted with permission from [8]
© 2004 Elsevier Ltd)
By extensive researches on the effective physical properties (e.g., the electrical

conductivity, yield stress, and elastic modulus) of the porous materials, their poros-
ity dependencies are experimentally found to follow the power-law formula [9–13]:
M ¼ M0 ð1 pÞm ; (7.3)
a E// E⊥ b
2
Young's modulus, E (GPa)
200 p=0 0°
Young's modulus, E
30°
60°
150 90°
100
9
100 p = 0.240 8
7
p = 0.397 6
50 5
p = 0.499
4
0 4 5 6 7 8 9
0 20 40 60 80 1
Angle, w (degree) Relative density, 1-p
Fig. 7.3 (a) Young’s modulus as a function of the angle ω and (b) log–log plot of Young’s
modulus versus porosity in the direction of ω from the x3-axis for lotus iron (Reprinted with
permission from [8] © 2004 Elsevier Ltd)
where M and M0 are the physical properties of porous and nonporous material,
respectively, and m is the coefficient empirically determined. The solid and broken
lines in Fig. 7.2 indicate Eq. (7.3) fitted to the measurement data. Thus, one finds
that the measurements lie on them, indicating that Eq. (7.3) holds for the effective
elastic constants of the anisotropic porous metals. The values of m for lotus iron are
similar to those for lotus magnesium, i.e., they seems to be independent of Poisson’s
ratio of the isotropic matrix and mainly depend on the shape and alignment of the
pores. In contrast, the values of m for lotus copper are slightly different from the
others. They appear to depend also on the elastic anisotropy of the matrix.
Since lotus metals are fabricated by using the unidirectional solidification
method, preferential orientations in the crystal growth usually appear. This is
the case for lotus copper and lotus magnesium [5, 6]. In the case of lotus iron,
however, X-ray diffraction proved that the matrix has no texture (not presented
here). Hence, despite the fact that the single-crystal iron shows a relatively strong
elastic anisotropy ð2c44 =ðc11 c12 Þ 2:4Þ [12, 13], the nonporous iron has
isotropic elastic constants and the two directional Young’s moduli of nonporous
iron are substantially equal to each other. Although the matrix of lotus magnesium
has a texture, it virtually exhibits elastic isotropy because the elastic property of
single-crystal magnesium is nearly isotropic [6]. In contrast, the matrix of lotus
copper shows elastic anisotropy that originates from a texture consisting of
elastically anisotropic crystals [5]. Thus, the texture effects on the matrix metal
are observed case by case.
Since all the independent elastic constants were determined, the elastic modulus
in an arbitrary direction is available through the coordinate conversion. Figure 7.3a
shows the Young’s modulus E of lotus iron as a function of the angle ω from the x3
axis. The Young’s modulus monotonically decreases with increase in ω, which
simply arises from the increase of the stress concentration around pores. Figure 7.3b
7.2 Internal Friction 131
shows the log–log plot of Young’s modulus of lotus magnesium versus porosity in
the direction of ω ¼ 0, 30, 60, and 90 . The solid lines in Fig. 7.3b show fits of
Eq. (7.3) to the measurement data. Thus, Young’s moduli in arbitrary directions can
also be expressed using Eq. (7.3).
7.2 Internal Friction
In lotus materials hydrogen exits in a solid solution state in metals and in gaseous state
in pores. For the practical use of these materials, it is important to know the behavior
of hydrogen, especially transportation paths and desorption temperature. Internal
friction (IF) method has been known to be sensitive to impurity atoms and lattice
defects. Ota et al. measured IF of lotus copper and found a large IF peak around 700 K
[14]. They suggested that hydrogen is responsible for the IF peak, although
the detailed mechanism was not known yet. Recently Yoshinari et al. investigated
the IF and hydrogen desorption behavior of lotus copper with various pore sizes in
order to discuss the mechanism of the influence of hydrogen on internal friction [15].
The specimens for the IF measurement which have a cylindrical shape (2 mm in
diameter and 20–30 mm in length) were cut from the lotus ingot by a spark-erosion
wire-cutting machine. Two types of specimens were prepared; the cylindrical axis
of specimens are selected so as to be parallel or perpendicular to the direction of
the lotus pores whose average pore size and porosity are compiled in Table 7.1. The
crystal grains in specimens were grown in the transfer direction. The grain size of
lotus copper was almost the same as the pore size, while that of nonporous copper
was about 0.5 mm. Hereafter lotus specimens with directional pores which are
parallel and perpendicular to the specimen axis will be referred to as, e.g., lotus15k
and lotus15⊥, respectively. An apparatus of a forced excitation torsion pendulum
type was used for IF measurements as illustrated in Fig. 7.4. The measurements
were done under the condition s of measuring frequencies of 0.1–10 Hz, tempera-
ture range of 300–1,100 K, and a heating rate of 3 K min1. The torsional vibration
was excited by supplying a torque to a magnet attached to torsion axis through the
external AC magnetic field. The torsion strain was detected by an optical lever
system. Both the excitation system and the detection system were connected to a
personal computer via DA and AD converters, respectively. The IF, Q1, is
detected as tanφ, where φ is a phase lag between stress and strain (or between
Table 7.1 Fabrication conditions for prepared specimens

Ar pressure H2 pressure Transfer velocity Average pore
Specimen (MPa) (MPa) (mm min1) Porosity (%) size (mm)
lotus15 0.6 0.4 15 34.2 0.30
lotus20 0.6 0.4 20 35.8 0.27
lotus50 0.6 0.4 50 35.1 0.06
Nonporous 1.0 0.0 20 – –
Internal Friction meas.

thin foil
light spot magnet

position detector Drive
coil
Detect
AD DA
Converter Converter vacuum
PC
Forced Driving Method

Drive Detect
phase internal
j ª Q -1
lag friction
specimen
The IF spectra for various frequencies (0.1 – 10 Hz)

can be obtained on a single heating run.
Fig. 7.4 An apparatus for measurement of internal friction and the internal friction spectra for
various frequencies (0.1–10 Hz)
driving force and detected strain). The phase lag was obtained by curve fitting of
stress and strain signals. The typical accuracy for φ is about 104 (0.005 ). The IF
spectra for various frequencies were measured on a single heating run. Thermal
desorption spectrum (TDS) of hydrogen was measured as follows: hydrogen gas
desorbed from specimen was detected by a quadrupole mass spectrometer during
constant heating of 12 K min1 and dynamic evacuating. Spectrum of hydrogen
pressure versus temperature was obtained in a range between 300 and 1,000 K.
Figure 7.5 shows the results of IF measurements for lotus15k, lotus20k, and
lotus50k specimens. On the first heating runs for as-prepared specimens, a broad IF
peak with a peak temperature of 610–710 K was observed. After heating up to
1,100 K, the specimens were subjected to the second heating measurements
(Fig. 7.5b). On the second run, an IF peak appears about 100 K higher than that
on the first run. In order to confirm whether this peak shifts from the lower
temperatures or it newly appears, measurements were repeated for another lotus50k
specimen with raising the upper limit of the heating as shown in Fig. 7.6. The peak,
which appears around 620 K on the first run, decreases and shifts toward lower
7.2 Internal Friction 133
a 0.12 b 0.12
0.10 1st heat 2nd heat
0.10
f = 1Hz f = 1Hz lotus50||
0.08 lotus50|| 0.08
lotus20||
-1
-1
0.06 0.06
Q
Q
lotus15||
0.04 0.04
0.02 0.02 lotus15||
lotus20||
0.00 0.00
200 300 400 500 600 700 800 900 1000 1100 200 300 400 500 600 700 800 900 1000 1100
T/K T/K
Fig. 7.5 IF spectra of lotus copper for frequency of 1 Hz. (a) 1st and (b) 2nd heating runs
(Reprinted with permission from [15] © 2012 GSIntervision)
a 0.10 b
800
P2 lotus50||
0.08 P1 700
TP/K
lotus50||
TP(P1) TP(P2) 600
0.06 f=1Hz
Q-1
500
Q-1 max
0.05 Q-1 max(P2)

0.04 0.04
0.03 Q-1 max(P1)
0.02 0.02
0.01
0 0
200 300 400 500 600 700 800 900 1000 1100 600 700 800 900 1000
T /K TA/K
Fig. 7.6 (a) IF spectra of lotus 50k measured repeatedly by increasing the upper heating limit.
(b) Variation of peak temperature (Tp) for f ¼ 1 Hz and peak height (Q1max) for P1 and P2 with
annealing temperature (TA) (Reprinted with permission from [15] © 2012 GSIntervision)
temperature with the annealing. A new peak appears around 750 K after annealings
at temperatures higher than 850 K. It should be noted that the two peaks coexist
after the annealings at 850–950 K. Therefore, it is clarified that the two peaks have
different origins. In the following, 620 and 750 K peaks are referred as P1 and P2,
respectively. In Fig. 7.6b the heights, Q1max, and peak temperatures, Tp, of P1 and
P2 are summarized as a function of the annealing temperature, TA, which is a
maximum temperature of the previous heating run. Although Q1max of P2 also
decreases with TA, it should be noted that the relation between Q1max and Tp is
opposite for P1 and P2. The IF was also measured for specimens in which
directional pores lie in the direction perpendicular to that of specimen axis. The
results for lotus15⊥ specimen are shown in Fig. 7.7.
The IF of nonporous specimen was also measured for comparison. Figure 7.8
shows the first and second heating measurements for nonporous specimen. It is
shown that there is almost no IF peak except for a continuously rising background
on the first heating run, while a peak is observed around 670 K on the second run.
Since the properties of the peak are similar to P2 observed for lotus specimens, it is
considered that it has the same origin as P2.
0.12
lotus15|| 1st heat
0.10 f= 1Hz
lotus15|| 2nd heat
0.08 lotus15⊥ 1st heat
Q-1 lotus15⊥ 2nd heat
0.06
0.04
0.02
0
200 300 400 500 600 700 800 900 1000 1100
T/K
Fig. 7.7 IF spectra with pore direction perpendicular to specimen axis (lotus15⊥) comparing with
parallel pore direction (lotus 15k), f ¼ 1 Hz (Reprinted with permission from [15] © 2012
GSIntervision)
0.06 nonporous Cu f= 1Hz

1st heat
2nd heat
0.04
Q-1
0.02
0
200 300 400 500 600 700 800 900 1000 1100
T/K
Fig. 7.8 IF spectra of nonporous copper specimens on first and second heating runs. f ¼ 1 Hz
Figure 7.9 shows hydrogen spectrum for lotus50 specimen. Two peaks are
observed; a clear peak is observed around 690 K and a shoulder peak is also
observed around 850 K. On the second heating run, the 690 K peak completely
disappears, while the 850 K peak still remains although its magnitude somewhat
decreases. Therefore, it is considered the 690 K peak corresponds the hydrogen
desorption from the specimen and the 850 K peak contains hydrogen release from
places other than the specimen. The amount of hydrogen estimated from the 690 K
peak was about 40 and 10 at.ppm for lotus50 and lotus20 specimens, respectively.
The amount of hydrogen is much smaller than the hydrogen solubility of 470 at.
ppm under the specimen fabrication condition (melt temperature 1,523 K and
hydrogen pressure 0.4 MPa) [16]. It is considered that a large amount of hydrogen
was released during the cooling process of specimen fabrication.
7.3 Tensile Strength 135
lotus50
2
PH2/ 10-6 Pa
0
600 700 800 900 1000
T/K
Fig. 7.9 Hydrogen TDS for lotus 50 specimens. Dashed line shows a suitable background
The origin of P2 may not be related with hydrogen because P2 appears after the
hydrogen release from specimens and P2 is also observed for nonporous specimen
which is fabricated under atmosphere without hydrogen. It is well known for alumi-
num and copper that a large IF peak is observed for a temperature range of 0.4–0.6
Tm, where Tm is the melting temperature. The mechanism of the peak was believed to
be the grain boundary (GB) sliding, because no IF peak is observed for single-
crystalline specimens and the peak has been referred to as the GB peak or Kê peak
[17]. Although it is not clear whether P2 is caused by GB sliding or dislocation
movements, anelastic strain which is responsible for P1 may be the same as that of P2
since the peak height of P1 is compatible to P2 and the magnitudes of them are
complementary as shown in Fig. 7.6b. In the case of CW peak observed in bcc metals,
the existence of hydrogen retards the dislocation motion and the CW peak appears at
higher temperatures than the original peak. It is considered for the case of P1 that
the existence of hydrogen helps the motion of the defects which is responsible for the
P2 relaxation. As a result, P1 appears at lower temperatures than P2. It is a new idea
that the hydrogen affects the so-called GB relaxation and it may be helpful for the
elucidation of its mechanism. Furthermore, it is interesting to use lotus specimens
for such a purpose since anisotropic GBs and hydrogen atoms simultaneously
exist in them.
7.3 Tensile Strength
Little data have been available on the mechanical properties of this type of lotus
metals. Wolla and Trovenzano [18] and Simone and Gibson [19] measured the
tensile strength of porous copper with pore orientation parallel to the tensile
direction. However, their data exhibited large scatter, which may be attributed to
Fig. 7.10 Stress–strain a

curves for the specimens
with the pore orientation 140
(a) parallel and (b) 120
perpendicular to the tensile
direction of lotus copper 100 1
2
Stress (MPa)
3
from [20] © 2001 Elsevier
Science Ltd) 60
1. porosity 20%
40 2. porosity 25%
20 3. porosity 29.8%
0
0 20 40 60 80 100
Strain(%)
b
140
120
100
Stress (MPa)
80 1 1. porosity 19%
60 2 2. porosity 21.3%
3
40 4 3. porosity 24.2%
4. porosity 34%
20
0
0 20 40 60 80 100
Strain(%)
the microstructural variations both within each specimen and among specimens.
Furthermore, no investigations were carried out on the anisotropy in the mechanical
properties of the porous metals with elongated cylindrical pores. Hyun et al. [20]
undertook to measure the ultimate tensile strength and the yield strength of the lotus
copper with elongated cylindrical pores and to elucidate the uniaxial tensile behav-
ior of lotus copper with the pore orientation parallel to and perpendicular to the
tensile direction. Moreover, the fracture surfaces of the selected tensile test
specimens were investigated to obtain information concerning the fracture mode
of these materials. Typical stress–strain curves for the lotus copper specimens are
shown in Fig. 7.10. The curves show linear elastic behavior at small strains,
followed by yield and strain hardening up to the peak stress. The ductility of the
specimens decreases with increasing porosity. The ultimate tensile strength
of specimens with cylindrical pore orientation parallel to the tensile direction
is plotted against the porosity in Fig. 7.11. The data points for the ultimate tensile
strengths lie on a straight line which passes through the point of 0 MPa at the
porosity of 100 %; the specific ultimate tensile strength does not change by
the pore existence. This fact indicates that the pores whose axes are aligned parallel
Fig. 7.11 Ultimate tensile 160

strength and yield strength of st (Pore//Tension)
lotus copper in the direction 140 sy (Pore//Tension)
parallel and perpendicular to
Ultimate Tensile Strength, σt (MPa)

st (Pore^Tension)
pore axis versus porosity sy (Pore^Tension)
120
Yield Strength σy (MPa)

Science Ltd) 100
80
60
K
K==11
40
20 K =3
K =3
0
0 20 40 60 80 100
Porosity (%)
Fig. 7.12 Comparison of Hyun et. al. (2001)

data obtained from two Simone and Gibson (1996)
different groups for ultimate 160
tensile strength of lotus
copper in the direction 140
parallel to the pore axis
Ultimate Tensile Strength / MPa

from [20] © 2001 Elsevier 120
Science Ltd)
100
80
60
40
20
0
0 20 40 60 80 100
Porosity(%)
to the tensile direction cause little stress concentration in the tensile specimens.
Simple rule of mixture of the empty pores and the solid body can be applied to
these specimens.
As shown in Fig. 7.12, the ultimate tensile strength obtained by Hyun et al. [20]
agrees well with the data by Simone and Gibson [19] at a low porosity levels, but
the strength is considerably higher in the higher porosity range. This difference may
be due to the production method of tensile specimens.
On the other hand, the ultimate strengths of specimens with cylindrical pores
perpendicular to the tensile direction are also plotted against the porosity in Fig. 7.11
The ultimate tensile strength of the specimen with pores perpendicular to the
tensile direction is much lower than that of specimen with the pores parallel to the
tensile direction at a given porosity.
Correlation between the ultimate tensile strength and the porosity of sintered
porous materials was studied extensively [21–23]. The measurements were limited
in the low porosity levels from 0 to 40 %, because the strength for porosity higher
than about 60 % becomes almost 0 MPa. It is considered that both load-bearing
areas model [21] and stress concentration model [23] are reasonable to determine
the strength of porous materials, since load-bearing area is reduced with increasing
porosity and causes stress concentration around the pores.
7.3.1 Ultimate Tensile Strength
Balshin [22] suggested a relation of the form
σ ¼ σ 0 ð1 pÞK ; (7.4)
to describe the variation of strength with the porosity, where K is the empirical
constant which depends on the materials and the fabrication method. Recently
Baccaccini et al. [23] suggested that the empirical constant K is related with the
stress concentration around pores in the porous materials. The stress concentration
factor of the pores depends on the pore geometry and orientation with respect to the
direction of applied stress and can be expressed by
σ max
K¼ ; (7.5)
σ
where σ max is the maximum value of the stress, which is given by the following
equation using the radius parameters, a and b, of an ellipse-shaped pore in the
tensile specimen as shown in Fig. 7.13 [24].
a
σ max ¼ σ 1 þ 2 : (7.6)
b
For the specimen with the cylindrical pores aligned parallel to the tensile
direction, since b becomes infinite, the value of K approaches unity, where no
stress concentration takes place. Thus, Eq. (7.6) can be simply rewritten as
σ ¼ σ 0 ð1 pÞ: (7.7)
Fig. 7.13 Stress

concentration in the vicinity
of elliptical pore in the
infinitely large panel. The
pore is located in the center.
The stress is concentrated in
both edges of the pore
smax
Science Ltd)
s
b
a
For the specimen with cylindrical pores oriented perpendicular to the tensile
direction, the equality a ¼ b can be held, so that the value of K approaches 3. Then,
the stress can be expressed as
σ ¼ σ 0 ð1 pÞ3 : (7.8)
The ultimate tensile strength evaluated from Eqs. (7.7) and (7.8) is shown in
Fig. 7.11, where it is found that the experimental results are in good agreement with
the dotted lines estimated from Eqs. (7.7) and (7.8).
Such agreement suggests that (1) the ultimate tensile strength for porous copper
with cylindrical pores parallel to the tensile direction exhibits no stress concentra-
tion and (2) the ultimate tensile strength with the pore orientation perpendicular
to the tensile direction can be quantitatively explained in terms of stress concen-
tration. Such anisotropic behavior for the tensile strength is applicable not only to
lotus copper but also to other lotus metals and alloys. Thus, the anisotropy is a
general trend.
Although the mechanical properties of porous materials have been an object of
study for a long time, there is little attention to the correlation between the yield
strength and the porosity. Some authors [25, 26] reported that two types of yielding
occurred. At low porosity level the yield strength linearly decreased with increasing
porosity, and the yield strength over the critical porosity level drastically decreased
with increasing porosity. Experimental results shown in Fig. 7.11 are in agreement
with their results. Although it is difficult to explain perfectly the reason for this, they
suggested the following mechanism. As seen in Fig. 7.14, if a pore is in the tensile
plate, the stress concentrates at the edge of the pore. Therefore, the stress level of
this region is higher than that of matrix. Before the stress level reaches the yield
strength, at first plastic deformation locally occurs in the region of stress concen-
tration. Plastic deformation is accompanied by plastic hardening; therefore, this
Fig. 7.14 Formation of

barrier to dislocation
movement in the region of
stress concentration
Science Ltd)
region becomes barrier such as Cottrell–Lomer locks [27, 28], against moving
dislocations in the matrix until the cross slip makes progress. Consequently, the
yield strength of this sample will not rapidly decrease with increasing porosity at
low porosity levels.
Here, we have to remark on abnormal tensile properties of lotus iron fabricated
by nitrogen. Hyun et al. [7] fabricated the lotus iron, not only by hydrogen gas but
also by nitrogen gas. They found extraordinary merit of the superior mechanical
strength of lotus iron. Typical stress–strain curves of the lotus iron specimens
fabricated by hydrogen or nitrogen gas are shown in Fig. 7.15. The tensile direction
is parallel to the direction of the pore axes. The curves show a linear elastic
behavior at small strains, followed by yield and strain hardening up to the peak
stress. Drastic difference in the strength of the specimens fabricated with hydrogen
and nitrogen was observed even at similar porosity; the ultimate tensile strength of
the specimen fabricated by nitrogen is about twice higher than that by hydrogen.
According to the chemical analysis, hydrogen of 27.7 mass ppm is contained in the
specimens fabricated under hydrogen atmosphere, which is not considered to affect
the strength of iron, since the tensile data extrapolated to the zero porosity as shown
in Fig. 7.11 are good fit to the result of nonporous pure iron which hardly contains
hydrogen. On the other hand, nitrogen as much as 0.0873 mass% is contained in the
specimens fabricated with nitrogen atmosphere. Thus, such remarkable enhance-
ment of the tensile strength is attributed to solid solution strengthening due to solute
nitrogen atoms.
The ultimate tensile strength of specimens with cylindrical pore orientation
parallel to the tensile direction is much higher than that perpendicular as shown
in Fig. 7.16. Such anisotropic strength was interpreted by the previous work on
lotus copper [2] mentioned above. The pores whose axes are aligned parallel to the
250
a
200
Stress / MPa
150
b
100
50
0.00 0.05 0.10 0.15 0.20

Strain
Fig. 7.15 Stress–strain curves for lotus iron with 50.1 2.6 % porosity fabricated by (a) nitrogen
or (b) hydrogen gas. The tensile direction is parallel to the direction of the pore axis (Reprinted
with permission from [7] © 2003 Elsevier Ltd)
a 400 b 400
Ultimate Tensile Strength / MPa
Fe-N (Pore // Tension)

Fe-N (Pore // Tension)
350 350 Fe-N (Pore ^ Tension)
Fe-N (Pore ^ Tension)
Fe-H (Pore // Tension)
Yield Strenght / MPa
Fe-H (Pore // Tension)

300 Fe-H (Pore ^ Tension)
300 Fe-H (Pore ^ Tension)
250 250
200 200
150 150
100 100
50 50
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Porosity (%) Porosity (%)
Fig. 7.16 Porosity dependence of (a) yield strength and (b) ultimate tensile strength of lotus iron
fabricated with hydrogen and nitrogen in the direction parallel and perpendicular to pore growth
orientation (Reprinted with permission from [7] © 2003 Elsevier Ltd)
tensile direction cause little stress concentration; the specific tensile strength is
constant in spite of the pore existence in the lotus iron. On the other hand, the
ultimate tensile strength with the pore orientation perpendicular to the tensile
direction can be quantitatively explained in terms of the stress concentration; the
stress is accumulated in the vicinity of the pores.
7.3.2 Investigation of Tensile Deformation

by Acoustic Emission Method
Anisotropic tensile deformation of lotus copper was investigated with a particular

focus on crack formation during deformation. The anisotropy of microscopic
deformation was analyzed by an acoustic emission (AE) method that can detect
crack formation. The deformed microstructures were observed using a scanning
electron microscope (SEM). By correlating the observations pertaining to macro-
scopic deformation with those pertaining to microscopic deformation, the aniso-
tropic tensile deformation mechanism was examined [29].
A broadband AE sensor was placed on the specimen during the tensile tests, and
the AE signals generated by the deformation were used to analyze crack formation.
The detected signals were recorded on a personal computer through a discriminator
and oscilloscope. Since understanding the crack formation mechanism was central
to this study, only the burst AE signals relating to crack formation were analyzed,
although some continuous AE signals were also detected. Figure 7.17 shows the
amplitudes of burst AE signals detected during the deformation of lotus copper and
the corresponding stress–strain curves for parallel loading (Fig. 7.17a) and perpen-
dicular loading (Fig. 7.17b). It is apparent that the elongation and peak stress
(ultimate tensile strength) for perpendicular loading are much smaller than those
for parallel loading; this result is consistent with that of the previous report
mentioned in 7.3.1. Many burst AE signals are detected after yielding to fracture
regardless of the loading direction.
In contrast to lotus copper, burst AE signals are not detected when tensile
loadings are applied in directions parallel and perpendicular to the solidification
direction in nonporous copper as shown in Fig. 7.17c. This implies that the burst AE
signals detected in lotus copper are correlated with the porosity. Furthermore, there
is no other source of burst AE signals except for crack formation when the normal
plastic deformation of ductile copper takes place. Thus, the burst AE signals
associated with lotus copper can be attributed to pore-related crack formation.
The cumulative AE counts for the parallel and perpendicular tensile loadings
illustrate the anisotropy in crack formation (Fig. 7.18). The slopes of cumulative
AE counts for perpendicular loading are larger than those for parallel loading,
which means that cracks are formed more easily under perpendicular loading.
The change in the cross section, ΔS, during the parallel and perpendicular tensile
loadings for lotus copper is shown in Fig. 7.19. ΔS for nonporous copper is also
shown for comparison and is calculated assuming a plastic Poisson’s ration of 0.5.
The absolute value of ΔS for perpendicular loading is smaller than that for parallel
loading, while the absolute values of ΔS in lotus copper are much smaller than those
in nonporous copper. It is noted that the cross section for perpendicular loading
hardly changes despite the occurrence of a large tensile deformation exceeding
20 % strain. These results indicate that the plastic Poisson’s ratios of lotus copper
are much smaller than the value of 0.5 corresponding to nonporous copper; the
porosity strongly decreases the plastic Poisson’s ratio.
Fig. 7.17 Amplitude of burst AE signals detected during deformation and the corresponding
stress–strain curves for porous copper under (a) parallel and (b) perpendicular loadings.
(c) Amplitude of burst AE signals detected during deformation and the corresponding
stress–strain curves when tensile loadings parallel and perpendicular to the solidification direction
are applied to nonporous copper (Reprinted with permission from [29] © 2010 Materials
Research Society)
In addition to other tensile properties, the apparent work-hardening coefficients

of lotus copper exhibit anisotropy. The apparent work-hardening coefficients, n,
were evaluated by fitting the equation σ ¼ Kεn ; where K is the strength coefficient,
to the true-stress–true-strain curves calculated on the basis of changes in cross
section. The apparent work-hardening coefficient for perpendicular loading is
Fig. 7.18 Cumulative AE

counts as a function of
nominal strain in the case of
parallel and perpendicular
loadings (Reprinted with
permission from [29] © 2010
Materials Research Society)
Fig. 7.19 Change in cross

section as a function of
nominal strain in the case of
parallel and perpendicular
loadings. The change in cross
section of nonporous copper
is shown for comparison.
The plotted values
correspond to the average
of five measurements, and
the error bars are standard
deviation values (Reprinted
with permission from [29]
© 2010 Materials Research
Society)
0.13 0.01, which is smaller than the value of 0.34 0.04 for parallel loading.
Because the apparent work-hardening coefficients of lotus metals increase with
the degree of plastic deformation with regard to applied strain, it is deduced
that perpendicular loading produces comparatively less plastic deformation. The
reduced plastic deformation is probably the result of the localization of deformation
around the pores.
a b
200 μm
Fig. 7.20 SEM images of deformed porous structure in the case of parallel loading. The applied
nominal strains are (a) 31.5 % and (b) 41.0 %, which correspond to conditions just before the peak
stress and just after the peak stress in the stress–strain curve, respectively (Reprinted with
permission from [29] © 2010 Materials Research Society)
Figure 7.20a shows the SEM image of a deformed lotus structure following
parallel loading, where the deformation is applied up to just before the point of peak
stress in the stress–strain curve. The cracks whose propagation direction is perpen-
dicular to the pore direction are formed as indicated by the arrows. The cracks
become remarkably large just after the peak stress, as shown in Fig. 7.20b. Not one
crack but many cracks are formed in the vicinity of the peak stress, and they are
distributed in the specimen as indicated by the arrows in Fig. 7.21a, b. Figure 7.22a
shows the SEM images of a deformed lotus structure just before the peak stress for
perpendicular loading. A small crack is formed around the pores as indicated by the
arrow. Note that it is more difficult to find cracks on the specimen surface under
perpendicular loading than under parallel loading. Just after the peak stress, cracks
grow and connect with others as shown in Fig. 7.22b. Figure 7.23a denotes the
deformed microstructures on the fracture surfaces under parallel loading. Serpen-
tine glides are formed on the entire observed surface as indicated by the arrows,
where the serpentine glide is a sharp and well-defined interwoven pattern [30].
This indicates that slip deformation occurs homogeneously and that the final
fracture is caused by glide plane decohesion. In contrast to parallel loading, the
deformed microstructures depend on the local positioning in the case of perpendic-
ular loading as shown in Fig. 7.23b. The semibrittle surface is formed near the final
fracture region, while slip lines are formed in regions away from the final fracture
region. In general, a metal tends to deform in a brittle fashion under high-stress
triaxiality [31]. This finding indicates that deformation occurs locally under high-
stress triaxiality in the case of perpendicular loading.
The tensile elongation and absolute value of change in the cross section of the
specimen are much smaller for perpendicular loading than the corresponding values
in the case of tensile deformation of nonporous copper. Under perpendicular
a b
Loading direction 50 m
1 mm
Fig. 7.21 (a) SEM images of deformed porous structure in the case of parallel loading and
(b) magnified version of (a). The applied strain is 51.0 %, which corresponds to a strain in the
vicinity of peak stress (Reprinted with permission from [29] © 2010 Materials Research Society)
a b
200 µm
Loading direction
Fig. 7.22 SEM images of deformed porous structure in the case of perpendicular loading. The
applied nominal strains are (a) 7.0 % and (b) 9.0 %, which correspond to conditions just before
the peak stress and just after the peak stress, respectively (Reprinted with permission from [29]
© 2010 Materials Research Society)
loading, lotus copper macroscopically behaves a semibrittle material, although

nonporous copper is ductile as shown in Fig. 7.24. Under parallel loading, lotus
copper behaves in an intermediate manner between deformation under perpendicu-
lar loading and that of nonporous copper as shown in Fig. 7.24. These deformation
behaviors of lotus copper are qualitatively consistent with that of a ductile metal,
which tends to deform as a brittle material under high-stress triaxiality [31].
a b
Pore
Pore
Matrix
Matrix 50 µm 20 µm
Fig. 7.23 Deformed microstructure on the fracture surfaces for (a) parallel and (b) perpendicular
loadings (Reprinted with permission from [29] © 2010 Materials Research Society)
Fig. 7.24 Macroscopic

deformations of porous
copper and nonporous copper
from [29] © 2010 Materials
Research Society)
Crack formation and growth affect tensile deformation in lotus copper, closed-cell
metal foams [32, 33], and open-cell metal foams [34] in the same way, being
independent of porous structure. Furthermore, in all three materials, the cracks
are preferentially formed in weak regions of the matrix. When lotus copper is
subjected to parallel loading, reduced stress concentration and stress triaxiality
occur around pores as compared with the cases of other porous metals. Thus, crack
formation is more difficult, which results in the superior tensile strength and
elongation. Therefore, it is important to make the porous structure so homogeneous
that the crack formation is suppressed.
7.4 Compressive Strength
7.4.1 Compressive Yield Strength
Compressive properties of foamed metals have been extensively investigated.

In the strain–stress curves, there is a plateau region, where the stress remains nearly
constant with increasing strain [35, 36]. Energy absorption is closely related with
this region. On the other hand, Simone and Gibson [37] investigated compressive
properties of porous copper fabricated by unidirectional solidification, but no
investigation on anisotropy for uniaxial compressive properties was carried out.
From the point of view, Hyun and Nakajima [38] investigated the uniaxial com-
pressive properties of lotus copper with elongated cylindrical pores parallel and
perpendicular to the compressive direction.
Compressive stress–strain curves for the lotus copper specimens with different
porosities and compressive direction are shown in Fig. 7.25. The slope of the
stress–strain curves decreases with increasing porosity and depends on compressive
direction with respect to the pore growth direction at a given porosity. At low strain
level, the stress of the specimen with cylindrical pores parallel to the compressive
direction is higher than that of specimen with pores perpendicular. However, this
1200 Porosity
pore // compression 0.0 %
1000 pore ⊥ compression
30.4 %
800
Stress, σ / MPa
48.4 %
600
400 59.5 %
200
0 20 40 60 80 100
Strain, ε (%)
Fig. 7.25 Compressive stress–strain curves of lotus copper and nonporous copper with different
porosities. The compression directions are parallel and perpendicular to the pore growth direction
(Reprinted with permission from [38] © 2002 Elsevier Science Ltd)
7.4 Compressive Strength 149
Fig. 7.26 Optical micrographs of cross section of lotus copper with 59.5 % porosity in the
direction parallel to pore axis after compression test: strain (a) 0 %, (b) 30 %, (c) 50 % and
(d) 80 % (Reprinted with permission from [38] © 2002 Elsevier Science Ltd)
tendency is reversed with increasing strain. For understanding this behavior, two
reasonable possibilities were taken into consideration [38].
First, while the stress concentration occurs around the pores in the specimen with
pores perpendicular to the compressive direction, there is little stress concentration
around the pores in the specimen with the cylindrical pore parallel to the compres-
sive direction. The stress concentration is important factor for determination on the
strength of porous metals [20, 23]. The pores perpendicular to the compressive
direction are easily deformed at lower stress.
Next, different types of deformation such as buckling occur in the specimens
depending on compressive direction with respect to the pore axis. The buckling in
a specimen may result from loading misalignment and the specimen can be
deformed at lower stresses. The macrographs of the lotus copper with 59.5 %
porosity compressed by 0, 30, 50, and 80 % in the direction parallel and
perpendicular to the pore axis are shown in Figs. 7.26 and 7.27, respectively. In
Fig. 7.27 Optical micrographs of cross sections of lotus copper with 59.5 % porosity in the
direction perpendicular to pore axis after compression test: strain (a) 0 %, (b) 30 %, (c) 50 %, and
(d) 80 % (Reprinted with permission from [38] © 2002 Elsevier Science Ltd)
the parallel case, two types of buckling are observed in the specimen during
compression. One is the macro-buckling observed outside of the specimens as
shown in Fig. 7.26c. Another is the micro-buckling observed around the pores.
Figure 7.28 shows the microstructure around the pores before and after 50 %
height reduction, in which the copper ligaments between pores after compression
is buckled like zigzag.
However, in the perpendicular case, macro-buckling is not observed outside of
specimens during compression, while the micro-buckling is found inside of
the specimens. Furthermore, the volume fraction of pores after 50 % compression
in the perpendicular case is lower than that in the parallel case. Moreover, the slope
of stress–strain curve in the perpendicular case is rapidly increased at 50 %
compressive strain as shown in Fig. 7.25. It can be explained that the pores in the
perpendicular case are easily crushed during compression due to the stress concen-
tration around the pores. Consequently, such anisotropy of the slope of the
Fig. 7.28 Optical micrographs of cross section of lotus copper with 59.5 % porosity in the
direction parallel to pore axis before and after compression test: (a) ε ¼ 0 and (b) ε ¼ 50 %,
respectively (Reprinted with permission from [38] © 2002 Elsevier Science Ltd)
stress–strain curves may be attributed to the stress concentration, the buckling, and
the volume fraction of pores.
7.4.2 Energy Absorption
Schematic stress–strain curve of the lotus-type porous copper is shown in Fig. 7.29.
The curve exhibits three regions: (I) elastic region characterized by elastic modulus;
(II) stress-plateau region characterized by a shallow slope corresponding to the
plastic yielding, buckling, and bending of cell ligaments; and (III) densification
region characterized by a relative steep slope. From the compression test data, the
energy absorption during the compression can be calculated by integrating the area
under the stress–strain curve [38]:
ðε
W ¼ σðεÞdε; (7.9)
0
where W is the absorbed energy per unit volume and ε is the strain.
Stress, s I II III
eD
Strain, e
I. Elastic region
II. Stress plateau region (caused by the plastic yielding, buckling and bending of cell ligament)
III. Densitification region
Fig. 7.29 Schematic compressive stress–strain curve of lotus metal. (I ) elastic region, (II) stress-
plateau region (caused by the plastic yielding, buckling and bending of cell ligament), and (III)
densification region (Reprinted with permission from [38] © 2002 Elsevier Science Ltd)
Table 7.2 Densification strain and absorption energy of lotus copper

Absorbed
Apparent Compressive energy per
Porosity density direction with Densification Stress volume
(%) (Mg m3) pore axis strain (%) (MPa) (MJ m3)
30.4 6.25 Parallel 48.4 234 70.8
Perpendicular 29.8 144 25.7
48.4 4.64 Parallel 51.2 159 47.1
Perpendicular 33.8 70.5 14.4
59.5 3.57 Parallel 46.3 91.5 27.8
Perpendicular 33.0 36.1 7.2
Table 7.2 compiles the absorbed energy up to densification strain, which is

defined as the strain corresponding to the end of the stress plateau of lotus copper.
The stress and the absorbed energy at densification strain decrease with increasing
porosity. At a given porosity the absorbed energy of the specimen in the parallel
case is higher than that of the specimen in the perpendicular case, because the
lotus copper with pores parallel to the compressive direction has higher strength
and ductility than that with pores perpendicular to the compressive direction.
It can be seen that the energy absorption capacity of lotus copper with the pores
parallel to the compressive direction is much higher than that perpendicular to the
compressive direction at a given porosity.
Fig. 7.30 Schematic

illustration showing (a)
specimen coordinate system
and (b) three Euler angles, ψ,
θ, and φ defined between
specimen and pore
coordinate system. The
longitudinal axis of the pore
is parallel to the x;3 axis
Research Society)
7.4.3 Effect of Pore Orientation on Compressive Properties
Ide et al. investigated the compressive properties of lotus stainless steel not only in
the direction parallel and perpendicular to the pore direction but also in other
directions. The results of compressive tests on lotus stainless steel were described
and the yield behavior by using micromechanical mean-field theory was
discussed [39].
Nonporous and porous stainless steels (SUS304L) were fabricated by a
continuous zone melting technique. Cubic specimens (5 5 5 mm3 or 8 8
8 mm3) with various values of θ were cut out from the nonporous and porous
ingots, where θ is the angle between the pore direction (x0 3 axis) and compressive
direction (x3 axis) as shown in Fig. 7.30. The pores inside the specimens are
invisible so that they assume that the angle of inside pores is the same as that of
pores on the surface and determine θ by analyzing the pores only on the surface.
The angle of the ith pore inside a porous specimen, θ(i), is given by
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
θðiÞ ¼ tan1 tan2 θ1 ðiÞ þ tan2 θ2 ðiÞ ; (7.10)
where θ1 ðiÞ is the projection angle between the pore direction and the x3 axis on
the x2–x3 plane and θ2 ðiÞ is the projection angle between the pore direction and the
x3 axis on the x1–x3 plane. However, since the angles cannot be measured simul-
taneously for a pore on the same surface, they determined the two projection
*
angles by averaging θ1 ðiÞ and θ2 ðjÞ on the respective surfaces: θ1 ¼ n11 θ1 ðiÞ and
*
θ2 ¼ n12 θ2 ðjÞ , where n1 and n2 denote the number of pores on the x2–x3 and x1–x3
planes, respectively. Note that the absolute values are used to take account of the
misorientation of pores. The angle θ is thus given by
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
θ ¼ tan1 tan θ1 þ tan θ2 : (7.11)
Figure 7.31a shows the compressive stress–strain curves of nonporous stainless

steel fabricated through unidirectional solidification, where σ k and σ ⊥, respectively,
denote the compression stress in the directions parallel and perpendicular to the
solidification direction. The compression behavior does not exhibit anisotropy
despite the presence of the texture. Figure 7.31b shows the stress–strain curves of
lotus stainless steel for compression in the direction of θ. The compressive stress in
the direction of 3 < θ < 15 and 58 < θ <87 is insensitive to θ. In all θ, the
compressive stress steeply increases in the high-strain range, and its slope tends to
increase with increased θ. Figure 7.31c shows the stress–strain curves of stainless
steel subjected to compression in the direction parallel and perpendicular to the
pore direction. The compressive stress in both directions decreases with increased
porosity. In the low-strain range, σ k is higher than σ ⊥. However, this magnitude
relation is reversed in the high-strain range. The crossover points of σ k and σ ⊥ shift
to higher strain with increase in porosity.
Figure 7.32a shows the yield stress (0.2 % proof stress) of lotus stainless steel in
the direction of θ, σ yθ. The yield stress σ yθ monotonically decreases with increasing
of θ. Figure 7.32b shows the porosity dependence of the yield stress of lotus
stainless steel. The yield stress σ yk decreases almost linearly with increased poros-
ity, while σ y⊥ rapidly decreases. Thus, σ y⊥ is lower than σ yk over the whole range of
porosity. The yield stress of lotus stainless steel is expressed by the power-law
equation [20] σ y ¼ σ y0 ð1 pÞm . The values of m were found to be 1.0 and 2.4 for
the direction parallel and perpendicular to the pore direction, respectively.
The compressive behavior of nonporous stainless steel does not show anistropy;
the macroscopic deformation modes is thought to originate from the anisotropic
pore structure. Although average compressive strain inside pores, εp, and that inside
the matrix, εM, cannot be easily measured, the magnitude relationship between
Fig. 7.31 (a) Stress–strain curves of nonporous stainless steel fabricated through unidirectional
solidification for compression in the direction parallel and perpendicular to the solidification
direction. (b) Stress–strain curves of lotus stainless steel (p ¼ 39 %) for compression in the
direction of θ. (c) Stress–strain curves of lotus stainless steel with different porosities for
compression in the direction parallel and perpendicular to the pore direction (Reprinted with
Fig. 7.32 Yield stress of

lotus stainless steel measured
in the compression tests:
(a) Yield stress in the
direction of θ. (b) Porosity
dependence of yield stress
in the directions parallel and
perpendicular to the pore
direction. The plots represent
the measurement result, and
the lines represent fitting of
the power-law equation
mentioned above to the
measurements (Reprinted
© 2006 Materials Research
Society)
them can be calculated according to Qiu and Weng’s theory on the elastoplastic
behavior of composite materials [40], in which the ductility of the matrix is taken
into account by its secant moduli. For compression in the direction perpendicular
to the pore direction, it is concluded from the theory that εp is larger than εM; the
deformation rate of pores with regard to the macroscopic compressive strain is
larger than that of the matrix. Thus, in the perpendicular case, porous specimens
are densified with increased macroscopic comressive strain, and specimens
macroscopocally deforn only in the compressive direction. On the other hand, for
compression in the pore direction, the Q–W theory concludes εp almost equals εM;
the deformation rate of the pore with regard to the macroscopic compressive strain
is almost the same as that of the matrix. Therefore, it is apparent that the pores may
Fig. 7.33 Normalized

stress–plastic strain curves of
lotus stainless steel, where
the compressive stress of
lotus stainless steel is
normalized with its yield
stress: (a) compression in the
direction parallel to the pore
direction and (b)
compression in the direction
direction (Reprinted with
remain, even in the high-strain range, and barreling may occur as well as in the case
of nonporous materials.
A plateau-stress region exists in conventional porous metals; in this region, first,
local weak regions deform and harden, and thus deformation in these regions
does not proceed further. Subsequently, other weak regions deform and harden.
Densification proceeds by iteration of this phenomenon; pores collapse not simul-
taneously but sequentially. Figure 7.33 shows the normalized
stress–plastic strain
εp and that inside the matrix, εM, curves. σ jj σ yjj and σ ? σ y? denote the normalized
stress in the directions parellel and perpendicular to the pore direction. εpl denotes
the plastic strain.
For p ¼ 49 %, σ ? σ y? is lower than that for p ¼ 0 % when εpl ¼ 0.15–0.5,

while for p ¼ 26 % and 36 %, σ ? σ y? agrees with that for p ¼ 0 %. It is surmised
that for p ¼ 49 %, the specimen is densified by sequential pore collapse in the same
way as for conventional porous materials, which results in the stress-increase rate of
p ¼ 49 % being smaller than that of the nonporous materials. On the other hand, for
p ¼ 26 and 36 %, the rates of stress increase in porous materials, similar to those of

nonporous materials. This is because the collapse of the pores appears to occur
simultaneously, this densification mode differs from that of conventional porous
materials.
For p ¼ 26 %, σ jj σ yjj agrees quite closely with that for p ¼ 0 % over the whole
strain range. This suggests that the specimen is not densified even in the high-strain
range,since stress concentration does not occur around the pores; εp almost quals
εM. σ jj σ yjj is roughly equal to that for p ¼ 36 and 49 %, σ jj σ yjj almost agreew with

that for p ¼ 0 % when εpl < 0.3, while σ jj σ yjj is smaller than that for p ¼ 0 % when
εpl 0.3. This is because the pores bend with barreling when the porosity is high,
which results in stress concentration around the pores. The pores then begin to be
densified, after the plateau-stress region appears. The deformation modes therefore
depend not only on the compressive direction but also on porosity. The rapid stress
increase in the latter range of compression also depends on the compressive
deformation mode and yield stress.
The yield behavior of lotus stainless steel was dicussed by using
micromechanical mean-field theory based on Qiu and Weng’s concept [40]. On
the assumption that a macroscopic yield is experimentally observed when local
region R yields, the yield function for x 2 R is expressed as
!2
X
ðRÞ
fPR ¼ Y2; (7.12)
eq
!2
X
ðRÞ 0 2 X 2
2 2 2
¼ σeq þ R σ 0 eq ¼ σeq þR eq
σeq ; (7.13)
eq
where Y is the yield stress of the matrix, ∑eq is the macroscopic effective stress in
the matrix, and ∑eq(R) is the effective stress for x 2 R. The term R is a parameter
introduced by Tane et al. and can take into account the microyield around pores; R
depends on the local-yield region for macroscopic yield to appear, and R > 1
means that local yield is dominant for macroscopic yield [8]. The average and the
σ eq Þ2 and ðσ 0eq Þ2 , are defined as
deviation of local effective stress, ð
X 2 1n o
¼ σ 2 σ3 Þ2 þ ð
ð σ 3 σ1 Þ2 þ þ 6ð
σ 6 Þ2
eq 2
1
ð h 0 0
i2 h 0 0
i2 h 0 i2
þ σ 2 ðxÞ σ 3 ðxÞ þ σ 3 ðxÞ σ 1 ðxÞ þ þ6 σ 6 ðxÞ dV
2VM VM
0 2
σ eq Þ2 þ ðσ eq Þ ;
¼ ð
(7.14)
σ eq Þ2 and ðσ 0eq Þ2 depend on the average stress and the stress variance of the
where ð
matrix, respectively; VM is the volume of the matrix. The numerical-
Fig. 7.34 Yield stress of

lotus stainless steel
calculated by
micromechanical mean-field-
theory: (a) yield stress in the
direction of θ from the
longitudinal axis of the pore,
(b) porosity dependence of
yield stress in the direction
parallel and perpendicular to
the pore direction. The plots
and lines represent the
measurement and calculation
results, respectively
Research Society)
differentiation method based on Eshelby’s inclusion theory and Mori–Tanaka’s

mean-field theory was used to calculate the yield stress by using Eqs. (7.12) and
(7.13). Figure 7.34a shows the yield stress of lotus stainless steel in the direction of
θ, where the plots denote the measurements and lines denote the calculations for
R ¼ 1,3, and 4, and a3/a1 ¼ 1. For R ¼ 1, the calculations are higher than the
measurements, since the local yield around the pores is not adequately taken into
account. On the other hand, the calculations agree with the measurements for R ¼ 3
and 4, because the local yield is adequately taken into account. Figure 7.34b shows
the porosity dependence of σ yk and σ y⊥ for R ¼ 1, 3, and 4. The yield stress σ yk is
independent of R, because stress deviation does not appear for compression in the
pore direction. The yield stress σ y⊥ depends on R, and σ y⊥ decreases with increased
R. This is because high-stress deviation appears on the loading in the direction
perpendicular to the pore direction. The calculations for R ¼ 4 provide a better fit to
the measurements on the whole.
Incident wave Transmission wave

Stress wave
High-pressure
air compressor
Specimen
Striker
Incident pressure Transmitter

bar pressure bar
Stopper
Fig. 7.35 Apparatus for high strain rate compression test by split Hopkinson pressure bar method
7.4.4 Strain Rate Dependence of Compressive Behavior
Lotus metals are promising not only as functional materials such as sound absorbers
but also as lightweight structural materials. There are also high expectations of lotus
metals as impact energy absorbers. Because lotus metals have considerably better
mechanical strength than metal foams, it is expected that those also exhibit high-
energy absorption. However, the high strain rate compression behavior of lotus
metals has yet to be investigated. Here high strain rate compression behavior and
energy absorption characteristics of lotus iron are described. High strain rate
(~103 s1) compression tests in the directions parallel and perpendicular to the
pore direction were done with the split Hopkinson pressure bar (SHPB) method
[41, 42]. Low and middle strain rate compression tests were also carried out with a
universal Instron testing machine. The apparatus of SHPB mainly consists of
striker, incident pressure bar, transmitter bar, and high-pressure air guns shown
in Fig. 7.35. The striker bar was made of a SUS304 stainless steel for 20 mm in
diameter and 700 mm in length. The incident and transmitter pressure bars are made
of 30 mm diameter S45C steel bars with the lengths of 1,600 and 1,100 mm,
respectively. The absorbed energy was calculated by Eq. (7.15), while the effi-
ciency of the absorbed energy, η, is calculated by
W
η¼ ; (7.15)
σ d εd
where σ d is the stress at the densification strain εd. Figure 7.36a shows the low,
middle, and high strain rate nominal stress–nominal plastic strain curves of porous
iron when the compressive direction is parallel to the pore direction [43]. In the
low and middle strain rate compression, the shapes of stress–strain curves are
almost the same between the low and middle strain rates: (1) The flow stress
monotonically increases with increasing strain up to εp 20 %, (2) the deforma-
tion proceeds with a low work-hardening rate in the range of εp 20–40 %,
Fig. 7.36 Low, middle, and high strain rate nominal stress–nominal plastic strain curves of porous
iron in directions (a) parallel and (b) perpendicular to the pore direction. All specimens were
prepared in the same hydrogen atmosphere; averages of porosity are in range of 47.9–49.7 %.
Specimen heights H are 6, 6, and 3 mm for low, middle, and high strain rates, respectively
(Reprinted with permission from [43] © 2010 Materials Research Society)
where the work-hardening rate of porous metals depends on the intrinsic work
hardening of matrix iron and the change in the porous structure as the deformation
progresses, and (3) the densification region, where stress steeply increases with
increasing strain, appears above εp 47.5 %. On the other hand, the shape of
stress–strain curve in the high strain rate compression is different from those in the
other two strain rate compressions. The flow stress slightly increases with increas-
ing strain up to εp 7.5 %. Then, the flow stress slightly decreases with increas-
ing strain in the range of εp 7.5–20 %. As a result, the flow stress shows hardly
any increases up to εp 30 %; a clear plateau-stress region appears. After the
plateau-stress region, the densification region appears. In three strain rates,
the flow stress increases with increasing strain rate up to εp 20 % as in the
nonporous iron. However, the difference among three strain rates becomes small
for εp 20–30 % as a result of the appearance of the clear plateau-stress region in
the high rate compression.
Figure 7.36b shows low, middle, and high strain rate stress–strain curves of lotus
iron in the compressive direction perpendicular to the pore direction. Comparison
with Fig. 7.36a indicates that the flow stress in the perpendicular direction is lower
than that in the parallel direction in the initial stage of the stress–strain curves.
However, the magnitude of the relationship is reversed in the latter stage of the
stress–strain curves in low and middle strain rates. In contrast to the parallel
direction, the shapes of stress–strain curves in the perpendicular direction are
similar for all three strain rates: (1) the flow stress monotonically increases with
increasing strain and (2) the densification region appears in the latter stage of
stress–strain curves. The flow stress increases with increasing strain rate in the
entire strain range as in the case of nonporous iron.
Fig. 7.37 Efficiency of absorbed energy and absorbed energy per mass in the high strain rate
compression in various porous metals: porous iron in directions parallel and perpendicular to pore
direction, closed-cell aluminum alloy foams (Alporas), closed-cell aluminum alloy foams
(Alulight), open-cell aluminum alloy foam (M-Pore), open-cell aluminum alloy foam (Duocell),
and open-cell aluminum alloy foam (Spacer) prepared by the spacer method [43] (Reprinted with
Figure 7.37 compares the efficiency of absorbed energy η and absorbed energy
per mass W in high strain rate compression among various porous metals. The
absorbed energy W and its efficiency η of lotus iron exhibit significant anisotropy;
W and η in the parallel direction are higher than those in the perpendicular direction.
W of lotus iron in both directions are higher than those of metal foams. It is of
particular interest that the absorbed energy in the parallel direction is much higher
than it is in metal foams. On the other hand, η of lotus iron in both directions is
comparable with that of metal foams; in fact, the efficiency in the parallel direction
is comparable with that of closed-cell aluminum alloy foams (Alporas), which has
the highest efficiency among all of the metal foams.
Figure 7.38a shows SEM images of microstructure on the inside wall of a pore
before deformation and those after (b) low, (c) middle, and (d) high strain rate
compression of approximately 5 % plastic strain is applied along the pore direction.
In the deformed microstructure after the high strain rate compression, the deforma-
tion bands are observed as indicated by the arrows. However, the deformed
microstructure after low and middle strain rate compression is almost the same as
the microstructure before compression, because the applied plastic strain is so small
that visible slip lines are not formed. Figure 7.39 shows SEM images of the
deformation on the inside walls of pores after (a) low, (b) middle, and (c) high
Fig. 7.38 (a) Microstructure on inside wall of pore before deformation. Deformation on inside
walls of pores after (b) low, (c) middle, and (d) high strain is applied along the pore direction
strain rate compression. The microscopic deformation behaviors are clearly differ-
ent among three strain rates. In the low strain rate compression, fine slip lines are
formed as indicated by the arrows. In the middle rate compression, the slip lines are
thicker than those in the low strain compression as indicated by the arrows. In the
high strain rate compression, large deformation bands are formed, which are further
apart than the slip lines in other two levels of strain rate compression. Comparison
between Figs. 7.38d and 7.39c indicates that deformation bands grow with an
increase in applied strain from 5 to 35 %.
The strain rate-dependent microscopic deformation behavior can be explained
by the strain rate-dependent assistance of thermal activation for dislocations to
move. In the low strain rate compression of εp 35 %, conventional fine slip lines
are formed. In the middle strain rate, dislocation becomes increasingly difficult to
move because the assistance from the thermal activation decreases. As a result,
slip deformation becomes concentrated on particular slip planes, forming thick slip
lines as shown in Fig. 7.39b. In the case of high strain rate compression,
dislocations become increasingly difficult to move. Thus, deformation occurs
only along limited planes, resulting in the formation of large deformation bands
Fig. 7.39 Deformation on inside walls of pores after (a) low, (b) middle, and (c) high strain rate
compression. Compression of approximately 35 % plastic strain is applied along the pore direction
as shown in Fig. 7.39c. Similar deformation behavior dependent on a strain rate

probably occurs not only on the observed inside wall of pores but also all around the
matrix iron. In the early stage of compression, the microscopic deformation behav-
ior is probably similar to those in the middle stage of compression of εp 35 %,
although slip lines cannot be observed in the low and middle strain rate compres-
sion because of the small applied strain. Thus, the effects of strain rate on the both
microscopic deformation behavior and initial yield stress can be explained by the
thermal activation of dislocations.
7.4.5 Compression Deformation Behavior of Lotus (γ-TiAl)
Gamma TiAl(γ-TiAl) has been known to be one of the most promising candidates
for high-temperature materials owing to its high strength at elevated temperature
and lower density compared with those of NiAl and Ni3Al. Intermetallic compounds
Fig. 7.40 Compressive stress–strain curves of lotus TiAl with various porosities (Reprinted with
permission from [44] © 2009 Elsevier B.V.)
including TiAl are expected to lightweight structural materials at elevated temp-

erature. However, no investigation of TiAl was carried out. Ide et al. investigated
fabrication and compressive behavior of γ-TiAl [44].
Raw ingot whose chemical composition was Ti48Al52 was prepared by skull
melting. Rods of 10 mm in diameter and 100 mm in length were cut from the raw
ingot with a spark-erosion wire- cutting machine. The TiAl rod was unidirectionally
solidified utilizing a continuous zone melting technique in a chamber pressurized
by mixtures of hydrogen and helium gas of 2.5 MPa.
Figure 7.40 shows the stress–strain curves of lotus TiAl with various porosities.
In both (parallel and perpendicular) directions, the compressive stress and the strain
at the peak strength decrease with increasing porosity. For lotus metals with ductile
matrix, the plateau-stress region appears owing to the gradual densification of the
specimen [38]. However, for lotus TiAl, the plateau-stress region does not appear in
both directions. For lotus TiAl with the similar porosity, the compressive stress in
the loading direction parallel to the elongated pore direction (parallel direction) is
higher than that in the loading direction perpendicular to the elongated pore
direction (perpendicular direction), and the strain at the peak strength in parallel
direction is larger than that in perpendicular direction. It is reported that the
polycrystalline Ti48Al52 shows 4 % plastic strain at room temperature [45]. There-
fore, plastic strain at room temperature in lotus TiAl significantly increases com-
pared with plastic strain in nonporous TiAl.
Figure 7.41 shows the porosity dependence of the 0.2 % proof stress of lotus TiAl,
where 0.2 % proof stress of nonporous TiAl was referred to as that of polycrystalline
Ti48Al52 with the same grain size (about 300 μm):332 MPa [46]. In both direc-
tions, 0.2 % proof stress decreases with increasing porosity. While 0.2 % proof stress
rapidly decreases with increasing porosity in the perpendicular direction, it
Fig. 7.41 Porosity dependence of the 0.2 % proof stress of lotus TiAl. Symbols and lines represent
the measurements and fittings of the power-law equation to the measurements, respectively
(Reprinted with permission from [44] © 2009 Elsevier B.V.)
gradually decreases with increasing porosity in the parallel direction. The porosity
dependence of the yield stress (0.2 % proof stress) of lotus metals is expressed using
the following equation:
p K
σ pγ ¼ σ 0γ 1 ; (7.16)
100
where σ pγ and σ 0γ are yield stress of lotus and nonporous material, respectively, and K
is a constant determined experimentally. With the experimentally obtained K ¼ 1.4
in the parallel direction and K ¼ 3.3 in the perpendicular direction, the porosity
dependence of the 0.2 % proof stress of lotus TiAl is expressed by the curve using
Eq. (7.16).
The anisotropic compressive properties of lotus TiAl are due to anisotropic
porous structure. However, the porosity dependence of the 0.2 % proof stress of
lotus TiAl differs from that of lotus metals with ductile matrix. Particularly, in
parallel direction, the K value obtained from Eq. (7.16) is higher than that of lotus
metals with ductile matrix (approximately K ¼ 1), i.e., the 0.2 % proof stress of
lotus TiAl decreases more with increasing porosity compared with that of lotus
metals with ductile matrix. Figure 7.42 shows periodic unloading–reloading
Fig. 7.42 Cyclic stress–strain curves of lotus TiAl. Porosities of specimens are, respectively,
(a) 9 16 and (b) 17 %. Loading directions are (a) parallel and (b) perpendicular to the elongated
pore direction (Reprinted with permission from [44] © 2009 Elsevier B.V.)
compressive stress–strain curves of lotus TiAl. The loading directions are

(a) parallel and (b) perpendicular to the elongated pore direction. When the stress
is unloaded, the strain remains in the specimen even in the very low stress level,
which is corresponding to the origins that compressive stress approximately line-
arly increases with increasing compressive strain in parallel and perpendicular
directions. Figure 7.43 shows the optical micrographs of cross section of lotus
TiAl after periodic unloading–reloading compression tests. The loading direction is
perpendicular to the elongated pore direction.
Deformation and cracks locally exist around the pores and between the pores as
shown in Fig. 7.43b. Since the deformation and cracking are caused by structural
defect in porous metal, it is thought that those can also occur in lotus metal with
ductile matrix. However, the specific strength in parallel direction in lotus metal
with ductile matrix holds constant despite the presence of pores, and pore collapse
does not affect the stress–strain response at early stage of compressive deforma-
tion. Therefore, 0.2 % proof stress of lotus metals with ductile matrix is not
sensitive to local large stress concentration. On the other hand, for lotus TiAl, it
is considered that the lack of ductility of the matrix is the cause of a stress–strain
response which is affected by local large stress concentration. The strain locally
increases with increasing stress and the macroscopic strain is rapidly increased by
local strain and crack caused by local large stress concentration, depending on
ductility of matrix. The region causing local deformation and fracture increases
with increasing porosity. Therefore, the 0.2 % proof strength of lotus TiAl
decreases more significantly with increasing porosity, compared with lotus metal
with ductile matrix.
The effect of the brittleness of matrix on compressive properties significantly
appears in high-strain region. Figure 7.44 shows optical micrographs of cross
Fig. 7.43 Optical micrographs of cross section of lotus TiAl, which is fabricated in mixture gas of
1.0 MPa hydrogen and 1.5 MPa helium, after cyclic compression test at (a) low and (b) high
magnifications. Porosity and average pore diameter of specimen are 17 % and 635 μm, respec-
tively. Loading direction is perpendicular to the elongated pore direction (Reprinted with permis-
sion from [44] © 2009 Elsevier B.V.)
section of lotus TiAl during compression; the loading direction is perpendicular to

elongated pore direction. At the strain 1.5 % as shown in Fig. 7.44a, initial
deformation occurs around the pores. Then, deformation and initial cracks are
observed around the pores (Fig. 7.44b), which progress between locally neighbor
pores (Fig. 7.44c, d). When the loading direction is parallel to elongated pore
direction, the local deformation occurs and progresses around the pores
(Fig. 7.45a, b) as in perpendicular compression. The compressive stress of the
specimen which apparently yielded increases with increasing strain without frac-
ture of specimen. Therefore, the increase of plastic strain in lotus TiAl compared
with plastic strain in nonporous TiAl is caused not by the improvement in ductility
of matrix but by successive deformation and cracking. For lotus TiAl in both
compressive directions, the deformation occurs around the pores and progresses
only in local region between pores. The fracture of specimen is caused by buckling
of cell wall without the deformation of overall specimen owing to barreling or
uniaxial deformation, which is observed in compressive deformation in lotus metals
with ductile matrix. However, the densification of the specimen does not occur so
that obvious plateau-stress region does not appear. Thus, the deformation behavior
in lotus TiAl with brittle matrix significantly differs from that in lotus metals with
ductile matrix.
7.5 Bending Strength 169
Fig. 7.44 Optical micrographs of cross section of lotus TiAl fabricated in mixture gas of 1.0 MPa
hydrogen and 1.5 MPa helium (perpendicular compression). Strain: (a) 1.5 %, (b) 3.0 %, (c) 5.0 %,
and (d) 10.0 %. Porosity and average pore diameter of specimens are 27 % and 619 μm,
respectively (Reprinted with permission from [44] © 2009 Elsevier B.V.)
7.5 Bending Strength
Bending is one of the most representative process to deform the porous structure
parts and it is necessary to understand bending properties. The bending properties of
lotus copper were investigated by Hyun et al. [47]. Figure 7.46 shows schematic
drawings of specimens for three-point bending tests; x, y, and z are defined as the
orientation of pore axis in the test specimens: (a) type X specimen whose pore axis
is parallel to the bending direction, (b) type Y specimen, and (c) type Z specimen
Fig. 7.45 Optical micrographs of cross section of lotus TiAl fabricated in mixture gas of 1.0 MPa
hydrogen and 1.5 MPa helium (parallel compression). Strain: (a) 5.0 % and (b) 7.0 %. Porosity and
average pore diameter of specimens are 27 % and 642 μm, respectively (Reprinted with permission
from [44] © 2009 Elsevier B.V.)
whose pore axes are perpendicular to the bending direction. Figure 7.47 shows the
typical load–displacement curves of type X, type Y, and type Z specimens with 31.4 %
porosity. The yield point and plastic deformation are observed in all specimens. The
maximum load and the fracture are observed in type Y and type Z specimens, but it is
not observed in the type X specimen; the load is increased continuously. Type Y and
type Z specimens are fractured preferentially because the strength, which involves
both tensile and compressive stresses in bending, is lower than that of type X
specimen. Such result can be explained by the behavior of the tensile and compressive
strength on the lotus copper; the specimen with cylindrical pores parallel to the tensile
and compressive direction has higher strength than that with pores perpendicular to the
direction [20, 38]. Furthermore, it is found that the absorbed energy, which is defined
by the value calculated from the sectional area of the load–displacement curve of type
X specimen, is higher than that of Y and Z specimens.
After the bending tests, the optical micrographs are shown in Fig. 7.48. In type X
specimen no crack is observed after bending, while the crack is observed in other
type of specimens: type Y and Z. Especially the crack is shown on the upper surface
where the tensile stress is loaded during bending. Since it is known that the strength
and elongation of the specimen with pores parallel to the tensile direction are much
higher than those perpendicular to the tensile direction, it is reasonable that type X
specimen can be deformed without crack. Thus, it can be explained that the
bendability of type X specimen is much better than that of type Y and Z specimens.
In order to understand the bending properties of type X specimens, the bending
properties of this type of specimens were measured as functions of porosity and
7.5 Bending Strength 171
Fig. 7.46 Schematic a Load

drawings of specimens for
bending test; x, y, and z are
defined as the orientation of
pore axis in the test
specimens. (a) Type X
specimen, (b) type Y Pore
5 mm (D)
specimen, and (c) type Z
specimen (Reprinted with 12 mm
Elsevier B.V.) 0.7-5.3 mm (t)
D+3t
b
Pore
Pore
z
y
thickness of the specimens. The bending load–displacement curves for type X

specimens with different porosities are shown in Fig. 7.49. The slope of the curves
decreases with increasing porosity. This tendency agrees with the results of tensile
and compressive properties shown before [20, 38]. Some interesting features are
observed; the crack propagation is ceased by the elongated pores as shown in
Fig. 7.50b, while the cracks occurred locally from several parts on the surface as
shown in Fig. 7.50c. Thus, it is suggested that the elongated pores prevent cracks
from propagating and then the specimen after bending is not broken down perfectly
by the cracks.
600
a b
400
200
Load / N
0
0 1 2 3 4 5
600 Displacement / mm
c
400
200
0
0 1 2 3 4 5
Displacement / mm
Fig. 7.47 Bending load–displacement curves of lotus copper with 31.4 % porosity at 3.3-mm
thickness. (a) Type X specimen, (b) type Y specimen, and (c) type Z specimen (Reprinted with
7.6 Fatigue Strength
For practical use of lotus metal as structural materials, the knowledge of fatigue
behavior is important because the effect of pore morphology on the fatigue behavior
probably differs from that on tensile (or compressive) strength. So far the fatigue
behavior of conventional porous metals with irregular pores has been investigated
[32, 48–51], and the effects of porosity, the local strength of cell walls, etc., were
clarified. For lotus metals, however, detailed studies have not been carried out,
e.g., the effects of pore morphology on fatigue strength at finite life and fracture
behavior have not been investigated. From a standpoint of practical use, the fatigue
behavior of lotus metals has to be investigated in details.
Lotus copper was employed as a model of lotus metals to study the effects of
pore morphology and applied-stress direction on the fatigue strength at finite life
and fracture surface characteristics [52]. Dogbone-type specimens with a gauge
diameter of 4 mm and gauge length of 6 mm were turned from the square rods.
The angles θ between the longitudinal direction of the dogbone-type specimens and
the longitudinal direction of cylindrical pores were 0, 16, 19, 21, 40, and 90 .
The directions of θ ¼ 0 and 90 are the directions parallel and perpendicular to the
longitudinal direction of pores, respectively. Constant stress amplitude fatigue tests
7.6 Fatigue Strength 173
Fig. 7.48 Optical

micrographs of lotus copper
with 31.4 % porosity after
bending test. (a) Type X
specimen, (b) type Y
specimen, and (c) type Z
specimen (Reprinted with
Elsevier B.V.)
were carried out with a servovalve-controlled electrohydraulic testing machine in

air at room temperature. Cyclic tension–compression stress of R ¼ 1 was applied
to nonporous and lotus copper specimens, where R is the ratio of the minimum
stress to the maximum stress. The frequency of cyclic stress was 5 Hz. Stress
amplitude σ a was calculated from an applied load and cross-sectional area including
pores. Constant stress amplitude fatigue tests with various stress amplitudes σ a were
carried out, and S–N curves which give the relation between stress amplitude
(S) and the number of cycle to fatigue (N) were obtained.
Fig. 7.49 Bending 800

load–displacement curves for 0 % porosity
nonporous copper and X 700 22 % porosity
31 % porosity
specimens (lotus copper) of 600 53 % porosity
thickness of 3.3 mm with
different porosities 500
Load / N
B.V.) 300
200
100
0
0 1 2 3 4 5 6 7
Displacement / mm
Figure 7.51 shows the log plots of σ a against the number of cycles to failure Nf
for nonporous and lotus copper, where cyclic stress was applied in the direction
(a) parallel and (b) perpendicular to the longitudinal axis of pores. The lines denote
fittings of the following function to the experimental data:
log αa ¼ C log Nf þ D; (7.17)
where C and D are fitting coefficients. For both directions the numbers of cycles to
failure of lotus copper and nonporous copper decrease with increasing stress
amplitude. The fatigue strength at finite life decreases in both directions with
increasing porosity. Nonporous copper does not show anisotropy in the fatigue
strength at finite life. On the other hand, the fatigue strength at finite life of lotus
copper shows significant anisotropy; the fatigue strength in the perpendicular
direction is lower than that in the parallel direction.
Figure 7.52 shows the number of cycles to failure of lotus copper with porosity
of 40 %, when cyclic stress of σ a ¼ 32 MPa was applied in the direction of θ ¼ 0,
16, 19, 21, 40, and 90 . The number of cycles to failure decreases with increasing
angle θ. Thus, the fatigue life is the longest, when stress is applied in the direction
parallel to the longitudinal axis of the pores (θ ¼ 0 ).
Figure 7.53 shows the fracture surface of lotus copper with porosity 30 % when
cyclic stress of σ a ¼ 47 MPa was applied in the direction parallel to the longitudinal
axis of the pores, where the number of cycles to failure Nf was 541,336. As shown
in Fig. 7.53b, a stage I-type facet is formed in the lower right region on the entire
fracture surface. This indicates that a crack initiated near this region. As shown in
Fig. 7.53c, a striation of stage II is formed in the center of the entire fracture surface.
The solid arrow denotes the direction in which a primary crack propagates. How-
ever, the direction of crack propagation changes when a crack reaches a pore as
denoted by the broken arrow in Fig. 7.53c. A dimple pattern is not formed. Instead,
the elongation of copper matrix that is analogous to a mountain ridge is formed as
shown in Fig. 7.53d. The direction of the primary crack propagation indicates that
Fig. 7.50 Optical

micrographs of the X
specimens appeared on the
cracks on the surface after
bending of lotus copper
with 5.3-mm thickness.
(a) and (b): the porosity
is 31.4 % and (c) the
porosity is 53.3 %
B.V.)
this surface is the final fracture surface of stage III. The fracture surface
characteristics for the parallel loadings were independent of the number of cycles
to failure Nf.
Figure 7.54 shows the fracture surface of lotus copper with porosity of 30 %
when cyclic stress of σ a ¼ 32 MPa was applied in the direction perpendicular to the
longitudinal axis off pores, where the number of cycles to failure Nf was 24,127.
Fig. 7.51 Log plots of stress

amplitude against the number
of cycles to failure for lotus
copper, where cyclic stress
was applied in the direction
(a) parallel and
(b) perpendicular to the
longitudinal axis of
cylindrical pores
(Reprinted with
permission from [52]
© 2007 Materials
Research Society)
Figure 7.54b shows a stage I-type facet. The regions showing this type of facet are
randomly distributed on the entire fracture surface, which indicates cracks initiated
in various sites. Figure 7.54c shows a striation pattern of stage II, and the regions
showing this type of surface are also randomly distributed on the fracture surface.
A crack does not propagate in the shortest path between pores, as shown by the
arrow in Fig. 7.54c. Furthermore, a primary direction in which a crack propagates
Fig. 7.52 The number of cycles to failure of lotus copper with porosity of 40 % as a function of
angle θ. Θ is an angle from longitudinal axis of pore, and cyclic stress of 32 MPa was applied in the
direction of θ (Reprinted with permission from [52] © 2007 Materials Research Society)
Fig. 7.53 The fracture surface of lotus copper with porosity of 30 % when cyclic stress of 47 MPa
was applied in the direction parallel to the longitudinal axis of cylindrical pores where the number
of cycle to failure was 541,336. (a) Entire fracture surface and (b–d) higher magnifications of (a).
The magnified points are denoted in (a) (Reprinted with permission from [52] © 2007 Materials
Research Society)
Fig. 7.54 The fracture surface of lotus copper with porosity of 30 % when cyclic stress of 32 MPa
was applied in the direction perpendicular to the longitudinal axis of cylindrical pores where the
number of cycle to failure was 24,127. (a) Entire fracture surface and (b–d) higher magnifications
of (a). The magnified points are denoted in (a) (Reprinted with permission from [52] © 2007
does not exist. Figure 7.54d shows the final fracture surface of stage III, and this
type of surface is formed near only this region. The ridge of a mountain-ridge-like
pattern is aligned in the direction parallel to the longitudinal axis of pores. The
fracture surface characteristics for the perpendicular loadings were independent of
the number of cycles to failure Nf as well as that for the parallel loadings.
It is well known that the fatigue strength of some metals is related to the ultimate
tensile strength [53]. The ultimate tensile strength of lotus copper in the parallel and
perpendicular directions can be expressed by the following simple formula [20]:
σ ¼ σ 0 ð1 pÞm ;
where σ a is the ultimate tensile strength of nonporous copper, m ¼ 1 for the parallel
direction, and m ¼ 3 for the perpendicular direction. The values of m were deter-
mined in the previous study [20]. Figure 7.55 shows the fatigue strength at Nf ¼ 105
of lotus copper with various porosities in the parallel and perpendicular directions
as a function of the ultimate tensile strength of lotus copper with the same
References 179
Fig. 7.55 The fatigue

strength at Nf ¼ 105 of
nonporous and lotus copper
with porosity of 23, 30, and
40 % as a function of the
ultimate tensile strength of
lotus copper (Reprinted with
porosities. The fatigue strength at Nf ¼ 105 was estimated from the fitting curves in
Fig. 7.51. The ultimate tensile strength was calculated by using Eq. (7.3) and
σ 0 ¼ 143 MPa [54]. The fatigue strength is proportional to the ultimate tensile
strength. This implies that the fatigue strength follows a simple formula of
σ ¼ σ 0 (1 – p)m, where σ 0 is the fatigue strength at finite life of nonporous copper.
The proportional relationship between the fatigue strength and ultimate tensile
strength indicates that the value of m for the fatigue strength is almost the same
as those of the ultimate tensile strength.
Most of the pores on the surface of specimens were closed by thin-film walls
owing to machining, and some surface pores were open. In general, fatigue strength
depends on defects on the surface of the specimens, because the defects become a
crack-initiation site. There are many possible crack-initiation sites in lotus copper
specimen, because there are many pores corresponding to the defects. Therefore,
the effect of the surface open pores on the fatigue strength is probably negligible
compared with that of inner pores.
References
1. Nakajima H, Ikeda T, Hyun SK (2003) In: Banhart J, Fleck A (eds) Cellular metals: manufac-
ture, properties, applications. MIT, Berlin, pp 191–202
2. Nakajima H, Ikeda T, Hyun SK (2004) Adv Eng Mater 6:377–384
3. Demarest HH Jr (1971) J Acoust Soc Am 49:768–775
4. Ohno I (1976) J Phys Earth 24:355–379

5. Ichitsubo T, Tane M, Ogi H, Hirao M, Ikeda T, Nakajima H (2002) Acta Mater 50:4105–4115
6. Tane M, Ichitsubo T, Hirao M, Ikeda T, Nakajima H (2004) J Appl Phys 96:3696–3701
7. Hyun SK, Ikeda T, Nakajima H (2004) Sci Tech Adv Mater 5:201–205
8. Tane M, Ichitsubo T, Nakajima H, Hyun SK, Hirao M (2004) Acta Mater 52:5195–5201
9. Phani HH (1986) Am Ceram Soc Bull 65:1584–1586
10. Kovacik J (1998) Acta Mater 46:5413–5422
11. Kovacik J, Simancik F (1998) Scr Mater 39:239–246
12. Rotter CA, Smith CS (1966) J Phys Chem Solids 27:267–276
13. Simmons G, Wang H (eds) (1971) Single crystal elastic constants and calculated aggregate
properties: a handbook, 2nd edn. MIT, Cambridge
14. Ota K, Ohashi K, Nakajima H (2003) Mater Sci Eng A341:139–143
15. Yoshinari O, Kobayashi T, Nakajima H, Ide T (2012) In: Hur BY, Kim BK, Kim SE, Hyun SK
(eds) Porous metals and metallic foams. GSIntervision, Seoul, pp 479–485
16. Smithells Metals Reference Book (1992) 7th edn. edited by Brandes EA, Brook GB,
Butterworth- Heinemann, Oxford, pp 12–13
17. Kê TS (1947) Phys Rev 71:533–546
18. Wolla JM, Provenzano V (1995) Mater Res Soc Symp Proc 371:377–382
19. Simone AE, Gibson LJ (1996) Acta Metall 44:1437–1447
21. Eudier M (1962) Powder Metall 5:278–290
22. Balshin MY (1949) Doklady Akad Sci USSR 67:831–996
23. Boccaccini AR, Ondracek G, Mombello E (1995) J Mater Sci Lett 14:534–536
24. Peterson AR (1953) Stress concentration design factors. Wiley, New York, p 1
25. Dehoff RT, Gillard JP (1971) Mod Dev. In: Hausner HH (ed) Powder Metall, vol 5. Plenum,
New York, p 281
26. Lund JA (1984) Int J Powder Metall Powder Tech 20:141–148
27. Cottrell AH (1952) Philos Mag 43:645–647
28. Lomer WM (1951) Philos Mag 42:1327–1331
29. Tane T, Okamoto R, Nakajima H (2010) J Mater Res 25:1975–1982
30. ASM Handbook (1987), vol 12. American Society for Metals, Materials Park
31. Agogino AM (1978) J Eng Mater Technol 100:348–355
32. Sugimura Y, Meyer J, He MY, Bart-Smith H, Grenstedt J, Evans AG (1997) Acta Mater
45:5245–5259
33. Motz C, Pippan R (2001) Acta Mater 49:2463–2470
34. Amsterdam E, de Vries JHB, De Hosson JTM, Onck PR (2008) Acta Mater 56:609–618
35. Gibson LJ, Ashby MF (1997) Cellular solids. Cambridge University Press, Cambridge
36. Ashby MF, Evans AG, Fleck NA, Gibson LJ, Hutchinson JW, Wadley HNG (2000) Metal
foams: a design guide. Butterworth-Heinemann, Woburn
37. Simone AE, Gibson LJ (1997) J Mater Sci 32:451–457
38. Hyun SK, Nakajima H (2003) Mater Sci Eng A 340:258–264
39. Ide T, Tane M, Ikeda T, Hyun SK, Nakajima H (2006) J Mater Res 21:185–193
40. Qiu YP, Weng GJ (1992) J Appl Mech-Trans ASME 59:261–268
41. Deshpande VS, Fleck NA (2000) Int J Impact Eng 24:277–298
42. Kolsky H (1949) Proc Phys Soc London, Sect B 62:676–700
43. Tane M, Kawashima T, Yamada H, Horikawa K, Kobayashi H, Nakajima H (2010) J Mater
Res 25:1179–1190
44. Ide T, Tane M, Nakajima H (2009) Mater Sci Eng A 508:220–225
45. Rao PP, Tangri K (1991) Mater Sci Eng A 132:49–59
46. Simone AE, Gibson LJ (1998) Acta Mater 46:3109–3123
47. Hyun SK, Nakajima H, Boyko LV, Shapovalov VI (2004) Mater Lett 58:1082–1086
48. Zhou J, Soboyejo WO (2004) Mater Sci Eng A369:23–35
49. Olurin OB, McCullough KYG, Fleck NA, Ashby MF (2001) Int J Fatigue 23:375–382
References 181
50. Harte A-M, Fleck NA, Ashby MF (1999) Acta Mater 47:2511–2524
51. Sugimura Y, Rabiei A, Evans AG, Harte AM, Fleck NA (1999) Mater Sci Eng A269:38–48
52. Seki H, Tane M, Otsuka M, Nakajima H (2007) J Mater Res 22:1331–1338
53. Suresh S (1998) Fatigue of materials, 2nd edn. Cambridge University Press, Cambridge
54. Gerber TL, Fuchs HO (1968) J Mater 3:359–374
Chapter 8
Various Physical and Chemical
Properties of Lotus Metals
Abstract Lotus metals have anisotropic pore configuration. Such anisotropic pores
yield anisotropic behavior of sound absorption, electrical and thermal conductivity,
magnetization, and corrosion behavior, because the pore itself affects those
materials characteristics. However, it does not affect thermal expansion, because
only nonporous body controls thermal expansion. Thus, lotus metals exhibit unique
characteristics.
Keywords Corrosion • Magnetization • Sound absorption • Thermal conductivity •

Thermal expansion
8.1 Sound Absorption
Sound absorption materials with an advanced performance to noises are required

for mufflers of cars, air-conditioner parts, pump chambers, elevated roads, etc. The
glass wool and foamed aluminum with closed pores are used most frequently as
marketed sound-absorbing materials at present. In most cases, these materials have
low strength, though they have good sound absorption capacity. The development
of the sound-absorbing material with comprehensive characteristics such as
high strength, lightweight, and good sound absorption capacity is desirable.
In order to investigate whether lotus metals exhibit significant sound absorption
characteristics, Xie et al. [1–3] measured the sound absorption coefficient of lotus
magnesium and copper plates by standing-wave method [4]. It is well known that
the absorption coefficient depends on the angle between the material surface and
the incident sound wave. In the experiments, the specimen surface was set perpen-
dicular to the incident sound wave. Figure 8.1 shows the schematic drawing for
measurements of the sound absorption coefficient. The specimen was set on the
rigid wall in the sound tube. A pure sound with a single frequency was generated
from the speaker installed at the other end. The standing-wave is caused by
the interference between an incidence wave and a reflection wave in the tube.

184 8 Various Physical and Chemical Properties of Lotus Metals
Rigid wall Acoustic tube Microphone Speaker

Specimen
Amplitude
Standing wave
Pmax
Pmin
Fig. 8.1 Schematic drawings for measurement of sound absorption coefficient by standing-wave
method
The sound pressure becomes the maximum at each quarter of the wavelength. The
maximum value of the sound pressure is written as j pjmax ¼ jA þ Bj, where A and B
are the amplitude of incidence wave and reflection wave, respectively. The mini-
mum value of the sound pressure is written as j pjmin ¼ jA Bj. The ratio between
the maximum and minimum of the sound pressure, n, is given by
j pjmax jA þ Bj
¼ ¼ n: (8.1)
j pjmin jA Bj
The sound reflectivity of the specimen is written as

B n 1
rp ¼ ¼
A n þ 1 : (8.2)
The absorption coefficient, α0, is given by
2 4
α 0 ¼ 1 rp ¼ : (8.3)
n þ ð1=nÞ þ 2
j pjmax and j pjmin are measured by moving the microphone in the tube to determine
the value of n. Then the absorption coefficient can be calculated using Eq. (8.3).
This measuring method is called a standing-wave method, which is one of the tube
methods, and the details are provided in JIS A 1405–1963 standards [4].
8.1 Sound Absorption 185
Fig. 8.2 Schematic drawing for measurement of flow resistance (Reprinted with permission from
[1] © 2004 Elsevier B.V.)
Both the flow resistance and the absorption coefficient show amount of the
performance of absorbing sound in the porous material. Therefore, it is necessary
to measure the flow resistance of the lotus copper. The flow resistance of the sound-
absorbing material is basically the same as the ventilation resistance, used to show
the ventilations such as cloth and paper. Unit area flow resistance of the porous
sound-absorbing material is defined by
Δp
Rf ¼ ; (8.4)
u
where u is the flow ratio when the constant air is passed through the vertical
direction on the surface of the material and Δp is the difference in pressure at
both sides of material. A pump is operated as shown in Fig. 8.2, and the differential
pressure between both sides of the specimen is measured with a U manometer. The
flow ratio u is written as
Q
u¼ ; (8.5)
S
where S and Q are area of specimen and flow volume of air, respectively.
Zie et al. measured the absorption coefficients α0 for the lotus copper as a
function of pore diameter under the condition that the specimen thickness and the
porosity were constant [1]. As shown in Fig. 8.3, α0 increases with decrease in pore
diameter from 660 to 460 μm in the whole frequency range up to 4 kHz. Figure 8.4
shows the porosity dependence of the absorption coefficient for constant pore
diameter 380 μm and constant specimen thickness 10 mm. The absorption coeffi-
cient increases with increasing porosity from 43 to 62 %. There are some data
scattering, which is attributed to coexistence of non-permeable and permeable
pores. Figure 8.5 shows the dependence of the absorption coefficient on specimen
Fig. 8.3 Effect of pore

diameter on sound absorption
coefficient of lotus copper
B.V.)
Fig. 8.4 Effect of porosity

on sound absorption
B.V.)
Fig. 8.5 Effect of specimen

thickness with 10 and 20 mm
on sound absorption
B.V.)
8.1 Sound Absorption 187
Fig. 8.6 Comparison of

absorption coefficient of
various materials (Reprinted
© 2004 Elsevier B.V.)
thickness when the pore diameter and porosity are constant. The absorption coeffi-
cient increases with increasing thickness. Especially, the absorption coefficient
increases significantly in high-frequency range. The maximum of α0 was observed
at 3.1 kHz in the specimen of 20 mm thick, while such maximum value was not
found until 4 kHz in the specimen of 10 mm thick. A similar tendency was observed
for lotus magnesium [5].
The glass wool has a peculiar mechanism of absorbing sound and is used widely
as the sound-absorbing material. The absorption coefficients of lotus copper, the
foam aluminum, and the glass wool with the same thickness, in the same frequency
region, were compared as shown in Fig. 8.6. All of them were measured by a
standard-wave method. The glass wool [6] and foam aluminum [7] exhibit superior
absorption capacity. The foam aluminum is composed of many independent closed
pores. Continuous pores are necessary to have high-sound absorption capacity [6]
so that minute cracks are introduced by rolling to connect the pores of the foam
aluminum.
For mechanism of sound absorption, it is thought that the viscosity resistance of
air in pores plays an important role in absorbing the sound for the porous material.
The sound is absorbed by the resistance in the fiber and this space of pores, when the
sound enters into the open pores in porous materials [8]. The sound is also absorbed
by disturbance of the movement of air. The absorption of the sound in porous
material is considered to be mainly due to the consumption of the sound energy by
the viscosity and the thermal conduction when the sound is propagating into this
tube. It is difficult to analyze this strictly because the pores in the porous material
are arranged to have complex shape.
The lotus metal can be considered as an assembly of many parallel thin tubes.
In order to simplify the analysis, first consider how a sound propagates in only one
tube. When a sound propagates in a thin tube, the attenuation of a sound depends on
the material of the tube. The attenuation in a smooth metal tube has been reported to
be larger than the attenuation in air [9]. According to Igarasi [10], the attenuation
constant β is expressed as
0:0102 1=2
β¼ f ; (8.6)
cr
where c and r are speed of sound and radius of tube, respectively. The attenuation of
sound can often be disregarded when the inner diameter exceeds several
centimeters because attenuation is reversely proportional to the inner diameter in
Eq. (8.6). On the other hand, since the radius of pores of the lotus copper is from
200 μm to 1 mm, the attenuation increases when the sound enters lotus copper. The
attenuation constants of N pores are given by
β 0:0102 1=2
βN ¼ ¼ f : (8.7)
N crN
The relation between the pore number and the porosity is written as
r12 ε
N¼ ; (8.8)
r2
where r1 is a radius of the specimen. From Eqs. (8.7) and (8.8), the attenuation
constant in the lotus copper is given by
0:0102 1=2
βN ¼ f : (8.9)
cεr12
It is thought that the reason of absorbing sound is mainly due to viscous friction
of air in the lotus copper. The absorption effect by only permeable pores is taken
into consideration. Therefore, it is necessary to measure the porosity of only
permeable pores. The thicker the specimen is, the more pores become difficult
to be permeable because the length of elongated pores is limited in the lotus copper.
It has been known that the pores are hardly permeable in the specimen of 20 mm or
more in thickness. Therefore, some differences were seen in the absorption
coefficient–frequency curve when the porosity was increased from 43 to 62 % in
Fig. 8.4.
The attenuation mechanism of the lotus copper is considered to be the change of
absorption sound energy into thermal energy by the viscous friction in the pores.
From this viewpoint, the attenuation constant in the lotus copper is related to the
radius, the porosity, the pore diameter, the thickness of the specimen, and the
frequency. Figure 8.7 shows the relation of the absorption coefficient and attenua-
tion constant by calculating from Eq. (8.9). Thus, it is found that the absorption
coefficient is related to the attenuation constant of the specimen.
8.2 Thermal Conductivity 189
Fig. 8.7 A relation between

absorption coefficient and
attenuation constant of lotus
copper with 10-mm thickness
B.V.)
8.2 Thermal Conductivity
To use lotus copper effectively as heat sinks mentioned later, it is very important to
know its effective thermal conductivity and consider the pore effect on the heat
flow. There is much in the literatures devoted to the effective thermal conductivity
of composite materials with cylindrical inclusions. Behrens [11] analytically
investigated the effective thermal conductivities of composite materials under the
assumption of orthorhombic symmetry and proposed a simple equation for
predicting the effective thermal conductivity. Perrin et al. [12] proposed a method
for predicting transport properties that included the thermal conductivity of circular
cylinders in square and hexagonal arrays. Han and Cosner [13] concluded a
numerical investigation of the effective thermal conductivities of composites with
uniform fibers in unidirectional orientation and layered composites with fibers laid
alternately along two mutually perpendicular directions. A numerical method was
devoted by Sangani and Yao [14] to determine the effective thermal conductivity of
a composite medium consisting of parallel circular cylinders in random arrays.
They cited that the conductivity appeared to be a relatively weak function of the
detailed arrangement of the cylinders. Mityushev [15] extended analysis of the
resolution of the Laplace equation in composite material with a collection of
non-overlapping, identical, circular disks. Moctezuma-Berthier et al. [16] showed
the predominant influence of the total porosity on the thermal properties of vulgar
porous media. Ogushi et al. [17] investigated the effective thermal conductivities
parallel and perpendicular to the pore axis of lotus copper both experimentally and
analytically. Since the thermal conductivity of the fluid in the pores is negligible in
comparison to lotus copper material in the application of lotus copper to heat sinks,
a very simple equation can describe the thermal conductivities of lotus copper.
8.2.1 Measurement of Effective Thermal

Conductivity of Lotus Copper
The effective thermal conductivity of lotus copper keff is defined by
Q
q¼ ¼ keff rT; (8.10)
A
where q is the heat flux from heat flow Q divided by heat flowing through the cross-
sectional area A in lotus copper including the pores, and T is the temperature in lotus
copper. The tensor keff is orthorhombic and is expressed as
0 1
keffjj
keff ¼ @ keff A: (8.11)
keff
The effective thermal conductivity of lotus copper is anisotropic. The parallel

and perpendicular effective thermal conductivities, keff|| and keff⊥, of lotus copper
are defined as the thermal conductivities for heat flow parallel and perpendicular to
the pore axis, respectively.
Figure 8.8 presents the experimental apparatus for measuring the effective
thermal conductivity. A cylindrical specimen with a diameter of 30 mm and a
length of 30 mm was located between upper and lower copper rods of known
thermal conductivity. The upper rod was heated by electrical heaters from the top
surface, while the bottom surface of the lower rod was cooled by cooling water in
order to transmit a certain amount of heat through the specimen. K-type
thermocouples were located each at 5-mm spacing in the specimen and the upper
and the lower rods to measure the temperature. An example of the temperature
distribution of the experimental setup is shown in Fig. 8.9. As the heat flowed in one
direction through the rods and specimen from top to bottom, heat flux q through the
specimen was obtained by the following one-dimensional equations:
q1 þ q2
q¼ ; (8.12)
2

@T
q1 ¼ kCu ; (8.13)
@x 1

@T
q2 ¼ kCu ; (8.14)
@x 2
where q1 is a heat flux from the upper rod to the specimen and q2 is the heat flux
from the specimen to the lower rod, kCu is the thermal conductivity of the rod, and x
is the direction of heat flow from the upper to the lower rod. From
8.2 Thermal Conductivity 191
Fig. 8.8 Experimental

apparatus for measuring
effective thermal
conductivity (Reprinted with
American Institute of
Physics)
Fig. 8.9 Temperature

distribution in specimen and
rods (Reprinted with
American Institute of
Physics)
Eqs. (8.10)–(8.14), effective thermal conductivity keff|| and keff⊥ are obtained by the
following equation:
q þq
keffjj ; keff? ¼ 1@T 2 ; (8.15)
2 @x lotus

where ð@T @xÞlotus is the temperature gradient in the specimen with pores parallel
or perpendicular to heat flow direction x.
8.2.2 Analysis of Effective Thermal Conductivity

of Lotus Copper
Since the heat flow cross-sectional areas parallel to the pore axis in lotus copper is
proportional to (1ε), the effective thermal conductivity keff|| is expressed by the
following equation:
keffjj
¼ 1 ε; (8.16)
ks
where ks is the thermal conductivity of nonporous copper and ε is the porosity

expressed by the volume ratio of pores versus the total volume of lotus copper.
Behrens [11] derived the effective thermal conductivity of composite materials
with orthorhombic symmetry. By applying his equation to the thermal conductivity
of lotus copper, the effective thermal conductivity perpendicular to the pores can be
expressed by the following equation:
keff? ðβ þ 1Þ þ εðβ 1Þ
¼ ; (8.17)
ks ðβ þ 1Þ εðβ 1Þ
where β(¼kp/ks) is the conductivity ratio, that is, pore conductivity kp divided by
material conductivity ks of lotus copper. Because the thermal conductivity of the
hydrogen gas or air in the pores of lotus copper is negligible compared with that of
lotus material, the effective thermal conductivity of lotus copper is derived as the
following equation by setting β ¼ 0 in the above equation:
keff? 1 ε
¼ : (8.18)
ks 1þε
Han and Cosner [13] performed a numerical study on the effective thermal
conductivities of fibrous composites using a unit-cell approach under a uniform
fiber diameter condition. Since the diameter of lotus copper is distributed around a
certain range, numerical simulation for the thermal conductivity perpendicular to
the pores under a nonuniform pore diameter condition was conducted to verify the
applicability of Eq. (8.18) to lotus copper.
Figure 8.10a shows a comparison between the experimental data and the results
evaluated by the analytical equation, Eq. (8.16), for thermal conductivity parallel to
the pores. Experimental data for keff|| showed good agreement with the analytical
results derived from the assumption that heat flow through the cross-sectional area
parallel to the pore axis is proportional to (1ε). A value of 335 W/(mK) for
thermal conductivity ks of the lotus copper material was used for comparison.
Figure 8.10b gives a comparison between the experimental data and Eq. (8.18),
where effective thermal conductivity keff⊥ perpendicular to the pores was lower
than that of the parallel ones(keff||) and was 40 % of lotus copper material ks with a
8.3 Electrical Conductivity 193
Fig. 8.10 (a) Comparison of experimental data of the effective thermal conductivity of lotus
copper parallel to pores with analysis data. (b) Comparison of experimental results of the effective
thermal conductivity of lotus copper perpendicular to pores with analysis data (Reprinted with
permission from [17] © 2004 American Institute of Physics)
porosity of 0.4. The analytical values evaluated by Eq. (8.18) showed good
agreement with the experimental data, indicating that Eq. (8.18) can be used to
predict the effective thermal conductivity perpendicular to pores of lotus copper
within experimental accuracy of 10 %. The results show that lotus copper
displayed anisotropy of the effective thermal conductivity. The effective thermal
conductivity keff⊥ perpendicular to pores was lower than that of parallel to keff|| ones.
8.3 Electrical Conductivity
For the permeability and large surface area, porous metals are indispensable as
electrode materials of a battery. The conventional porous electrode materials are
fabricated with powder metallurgy, and therefore the porous structure is complicated,
i.e., the pores are nonuniform and dispersed irregularly. This complicated structure
results in a degradation of the strength and permeability of the electrolyte solution.
Because of this unique porous structure of the directional porosity metals, lotus
metals show more superior mechanical properties than the conventional porous
metals [18]. Suematsu et al. [19] succeeded to fabricate lotus nickel; this porous
nickel is promising as a new type of electrode material that possesses the superior
strength and permeability. For this application, the effect of pore structure and
porosity on the electrical conductivity needs to be investigated. Thus, Tane
et al. [20] studied the electrical conductivity of lotus-type porous nickel. The first
part describes the experimental result on the conductivity parallel and perpendicular to
the longitudinal pore directions measured by four-probe method. Next, a calculation
method is shown for the electrical conductivity of such an anisotropic porous structure
on the basis of the effective-mean-field (EMF) theory [21]. Finally the validity of
the calculation by the method is discussed, comparing with the measurement data.
Fig. 8.11 Porosity 14x106

dependence of the electrical
Effective electrical conductivity,

conductivity of lotus nickel. 12
The plots indicate measured σ||
result, and the lines indicate 10 σ⊥
σ||, σ⊥ [Ω-1m-1]
fitting the equation to the
measurements (Reprinted 8
© 2005 American Institute of 6
Physics)
4
0
0.0 0.2 0.4 0.6 0.8 1.0
Porosity , p
8.3.1 Measurement of Electrical Conductivity

of Lotus Nickel
Lotus nickel ingots with long straight pores aligned in one direction were fabricated
through unidirectional solidification in a pressurized hydrogen or argon gas. Con-
tinuous current electrical conductivity along the longitudinal direction of the
specimen was measured with the four-probe method at room temperature.
Figure 8.11 shows the porosity dependence of the normalized effective electrical
conductivities σ k =σ 0 and σ ? =σ 0 , where σ k =σ 0 and σ ? =σ 0 denote the normalized
effective electrical conductivity parallel and perpendicular to the longitudinal pore
direction, respectively. The electrical conductivity of nonporous material, σ 0, is
1.41 107 Ω1 m1. The electrical conductivity shows the anisotropy that reflects
the porous structure. For the parallel to the longitudinal pore direction, the specific
conductivity is almost retained. This is because the flow direction of the electrical
current in the nickel matrix is almost parallel to the applied electric field. For the
perpendicular direction, the electrical current needs to flow detouring around pores.
Therefore, a distance that the electrical current flows increases, which results in the
increase of the effective electrical resistivity (decrease of the effective electrical
conductivity).
8.3.2 Analysis of Electrical Conductivity of Lotus Nickel
The previous extensive researches about porous rocks elucidated that porosity
dependence of the effective conductivity follows an empirical formula (called
Archie’s law [22]):
σ ¼ σ 0 ð1 pÞm ; (8.19)
8.3 Electrical Conductivity 195
Fig. 8.12 Porosity 14x106

dependence of the effective
a3/a1
Effective electrical conductivity,

electrical conductivity of 12
lotus nickel parallel and ∞
perpendicular to the pore 10 || 10
σ||, σ⊥ [Ω-1m-1]
growth direction in the case 5
of a3/a1 ¼ 5, 10, and 1, 8
which are calculated with ⊥
EMF theory. Plots indicate 6
the measurement data, and
lines indicate the calculation 4 σ||
results with EMF theory
2 σ⊥
from [20] © 2005 American
0
Institute of Physics)
0.0 0.2 0.4 0.6 0.8 1.0
Porosity , p
where σ and σ 0 are the effective electrical conductivity of porous and nonporous
material, respectively, and m is the coefficient determined empirically. Eq. (8.19)
was fitted to the measurement data, and lines in Fig. 8.11 show the fitting curves. It
is found that power-law relation holds in the anisotropic porous metals. The
coefficient m for parallel and perpendicular conductivity is estimated at 1.1 and
1.8, respectively.
Tane et al. [20] applied the concept of EMF theory to derive the effective
electrical conductivity of composites. Composites consist of the matrix and one
type of inclusions, whose volume fractions are denoted by fM and fI ð¼ 1 fM Þ,
respectively. Spatial averages of an electrical current density J (31 vector) and
electric field E (31 vector) of composites are expressed as J ¼ fM JM þfI JI and
E ¼ fM EM þfI EI , where JM ¼ σ M EM and JI ¼ σ I EI and σ M and σ I are electrical
conductivity of matrix and inclusion, respectively (σ is 33 matrix, and the
component is 0 when i 6¼ j). The electric field E is defined as E ¼ gradϕ ,
where ϕ is the electric potential. Then, the effective electrical conductivity σ of
the composites are defined as J ¼ σE. When expressing EI ¼ A EM , the effective
electrical conductivity σ of the composite can be written as
σ ¼ ð fM σ M þ fI σ I AÞ½ fM I þ fI A1 ; (8.20)
where I is the unit matrix (33 matrix). Using Eshelby’s equivalent inclusion
theory [23] and mean-field theory [24], A (33 matrix) can be expressed as A
1
¼ ½Sσ 1
M ðσ I σ M Þ þ I in analogy with the elasticity problem [25]. S is the
2nd-rank Eshelby tensor.
When the present EMF theory is applied to lotus nickel, we assumed that the
pore shape is ellipsoidal of a1 ¼ a2, and the electric conductivity for the pore σI is
0, where a1, a2, and a3 denote the radii of the ellipsoidal inclusion. Figure 8.12
shows the porosity dependence of σ k =σ 0 and σ ? =σ 0 in the case of a3 =a1 ¼ 5; 10
and 1. Measurement data are shown for comparison. Practical a3/a1 of lotus nickel
is 5–10 [26], and in this range the calculations agree well with the measurements.
Furthermore, the calculation can fully simulate the power-law formula. When
a3/a1 ! 1, i.e., in the case of typical lotus metal, the effective electrical
conductivities parallel and perpendicular to the longitudinal pore direction are
expressed with the power-law formula:
)
σ k ¼ σ 0 ð1 pÞ1 ;
(8.21)
σ ? ¼ σ 0 ð1 pÞ2 :
The electrical conductivity of lotus nickel with cylindrical pores aligned unidi-
rectionally was measured. It is found that lotus nickel shows the anisotropy in the
electrical conductivity because of the anisotropic porous structure; the effective
conductivity parallel to longitudinal pore direction decreases linearly, while that
perpendicular to the pore direction decreases steeply. These porosity dependencies
follow Archie’s power-law formula. The electrical conductivity of lotus nickel
calculated with this theory is consistent with measurement data, and the EMF
theory can fully simulate Archie’s power-law formula.
8.4 Magnetization
In the application of lotus materials to magnetic materials, lotus materials are

expected to have a useful merit that the geometrical magnetic anisotropy, called
shape magnetic anisotropy, can be controlled by the growth direction of pores,
without any change in the outer shape of material. For example, although the
perpendicular direction of platelike samples is a hard axis of magnetization, it is
possible to make the perpendicular direction be an easy axis of magnetization by
introducing elongated pores aligned perpendicular to the plate. With regard to
magnetic properties of porous materials, the effect of pore size has been widely
studied for materials possessing isotropic pores [27, 28], but lotus materials have
never been investigated. Onishi et al. [29] measured magnetization curves of lotus
nickel and cobalt to characterize the anisotropy of the magnetization process, and
the effect of porosity on the anisotropy was analyzed in terms of the empirical
Archie’s law.
The saturation magnetization of the porous and nonporous specimens was
measured in the magnetic fields applied parallel and perpendicular to the pore
growth direction. Due to the fact that the magnitudes of saturation magnetization
of normal ferromagnetic materials such as nickel are given by the atomistic origin,
it does not depend on the shapes of specimens and the directions of applied
magnetic field, and it was confirmed that the saturation magnetization observed
for porous nickel decreased linearly with increasing porosity. The same behavior
was observed for cobalt. The relationship between the saturation magnetization of
porous nickel and cobalt and their porosities is summarized in Fig. 8.13. The ratios
8.4 Magnetization 197
Fig. 8.13 Porosity

dependence of the
normalized saturation
magnetization ðMsat =M0sat Þ of
lotus nickel ( filled circles)
and cobalt (triangles)
from [29] © 2008 American
Institute of Physics)
of the saturation magnetization of the porous materials to that of the nonporous

material naturally correspond with the volume ratios of the net metal to the whole
porous material (1p) [30]. As a result, the saturation magnetization of porous
materials is expressed by the following formula:
Msat ¼ M0sat ð1 pÞ; (8.22)
where Msat and M0sat denote the saturation magnetization of the porous material and
that of the nonporous material, respectively. This result shows that the porosity of a
porous material can be estimated easily by comparing its saturation magnetization
with that of nonporous material.
Next, the magnetic field dependence of magnetization was in detail measured for
the lotus nickel and cobalt. Although there is no remarkable difference in the
obtained results between nickel and cobalt, effects of the pore direction are clearer
in nickel. This is due to the soft magnetic properties of nickel. Figure 8.14a, b shows
the normalized magnetization curves if lotus nickel, i.e., the magnetic field depen-
dence of the ratio of the magnetization of porous specimens (M ) to their saturation
magnetization (Msat), in the magnetic field parallel and perpendicular to the pore
growth direction. It was observed that for all the cases, the M/Msat values were close
to unity above 3 kOe, showing the nearly saturated state. When the magnetic field
was applied parallel to the pore growth direction, the normalized magnetization
curve of the porous specimen corresponded with that of the nonporous specimen, as
shown in Fig. 8.14a. In the case of perpendicular direction, on the other hand,
different slopes of the normalized magnetization curves appeared at low-magnetic
fields and the M/Msat value decreased with the increase in the porosity as shown in
Fig. 8.14b. The different slopes between the parallel and perpendicular directions
were caused by geometric magnetic anisotropy of specimens containing anisotropic
pores. In Fig. 8.15, M was normalized against the magnetization of the nonporous
specimen (M0) in order to discuss the porosity dependence of the magnitude of
anisotropy in terms of Archie’s formula. Only in the case of the magnetization
perpendicular to the pore growth direction, significant lowering of the M/M0 was
Fig. 8.14 Magnetization

curves for specimens of 0 %
porosity (solid line), 11 %
porosity (broken line), 19 %
porosity, and 35 % porosity
(chain line) in the
low-magnetic field (a)
parallel and (b)
growth direction (Reprinted
© 2008 American Institute of
Physics)
observed, at the low-magnetic field of ~200 Oe and was enhanced with increasing
porosity. The relationship between the porosity and M/M0 at a low- magnetic field
(~200 Oe), an intermediate magnetic field (~1 kOe), and a high-magnetic field
(~10 kOe) were shown in Fig. 8.16a–c, respectively. The relationship for the cases
could be fitted by the following Archie’s law:
M ¼ M0 ð1 pÞn ; (8.23)
where the exponent n is a fitting parameter. In this equation, it is considered that

difference between parallel and perpendicular directions represents the relationship
8.4 Magnetization 199
Fig. 8.15 Normalized

magnetization for specimens
of (I) 0 % porosity, (II) 11 %
porosity, (III) 19 % porosity,
and (IV) 35 % porosity in the
low-magnetic field (a)
parallel and (b)
growth direction (Reprinted
© 2008 American Institute
of Physics)
between anisotropy and porosity. The n value decreased toward unity with
increasing applied magnetic field. The relationship between the applied magnetic
field and n is shown in Fig. 8.17. The n values take a maximum, approximately 1.8
(perpendicular) and approximately 1.1 (parallel) at around 200 Oe. The anisotropy
of magnetization clearly appears when n > 1.1, i.e., magnetic field less than
2.3 kOe. The n values of 1.8 and 1.1 are coincident with those reported by Tane
et al. [20] for electrical conductivity of lotus nickel in the parallel and perpendicular
direction, implying the existence of remarkable similarity between electrical con-
ductivity and magnetization.
A possible mechanism of the coincidence in n may come from similar spatial
distribution of the electrical current and magnetic flux in the lotus nickel. When a
low-magnetic field is applied, the magnetic anisotropy originates from the area of
the pore walls perpendicular to the direction of applied magnetic field because
magnetic poles created at the pore wall increase the total energy of the system, and
therefore, this area acts as a resistance to magnetization. The anisotropic shape of
pore is considered to cause the observed magnetic anisotropy, and at a low-magnetic
field, the magnetic flux in lotus materials flows without crossing pore wall. Since
electrical current also flows without crossing pore walls, the spatial distribution of
electrical current and magnetic flux should be similar at a low-magnetic field, and
similar values of n are expected to appear in electrical conductivity and
magnetization.
a b
1.0 1.0
: M// /M0// : M// /M0//
0.8 : M⊥ /M0⊥ 0.8 : M⊥ /M0⊥
0.6 0.6
M/M0
M/M0
0.4 0.4
0.2 0.2
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Porosity, p Porosity, p
c 1.0
: M// /M0//
0.8 : M⊥ /M0⊥
0.6
M/M0
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Porosity, p
Fig. 8.16 Porosity dependence of the normalized magnetization at the magnetic field of
(a) 10 kOe, (b) 1 kOe, and (c) 200 Oe. The plots indicate the experimental data, and the lines
indicate the fitting results of formula (8.23) to the experimental data (Reprinted with permission
from [29] © 2008 American Institute of Physics)
Fig. 8.17 Dependence of the fitting parameter n on magnetic field applied to parallel ( filled
triangles) and perpendicular ( filled circles) to the pore growth direction. The broken lines indicate
n ¼ 1 (the case in which all the pores are penetrated) and n ¼ 1.1 (10 % anisotropy at the point)
(Reprinted with permission from [29] © 2008 American Institute of Physics)
8.5 Thermal Expansion 201
a 2.0 b 20
Nonporous Nonporous
Parallel (porosity 9.1 %) Parallel (Porosity 9.1 %)
Parallel (porosity 35.1%) 18 Parallel (Porosity 35.1%)
Elongation, ΔL/L0 (%)
1.5
mean CTE / 10-6 K-1

Perpendicular (porosity 9.1 %) Perpendicular (Porosity 9.1 %)
Perpendicular (porosity 38.3%) Perpendicular (Porosity 38.3%)
16
1.0
14
0.5
12
0.0 10
200 400 600 800 1000 1200 1400 200 400 600 800 1000 1200 1400
Temperature / K Temperature / K
Fig. 8.18 (a) Elongation of lotus nickel in the direction parallel and perpendicular to the pore axis
with different porosities as a function of temperature. (b) Coefficient of linear thermal expansion
of lotus nickel in the direction parallel and perpendicular to the pore axis with different porosities
as a function of temperature (Reprinted with permission from [31] © 2006 Elsevier Ltd)
8.5 Thermal Expansion
The understanding of the properties of thermal expansion of lotus-type porous

metals is important for the practical application such as the electrode, heat sink,
and high-temperature parts. Measurements of thermal expansion were carried out
by Tane et al. [31] in a thermal dilatometer. The equipment uses an alumina
pushrod in contact with the specimen to record the change in linear dimension
during a heating. The samples were in the shape of rods of nonporous and porous
nickel of 6 mm in diameter and 15 mm in length. The temperature was increased
from 294 to 1,273 K at a rate of 1 K min1. Measurements were performed under
mechanical-pump vacuum which was sufficient to prevent from oxidation, as
verified by measuring negligible mass changes after the experiments.
The thermal expansion data for nickel samples with different porosities are
shown in Fig. 8.18a, where △L ¼ LL0 and L0 is the original length of samples
at room temperature. In general, it is convenient to express the thermal expansion
behavior by the coefficient of linear thermal expansion (CTE) as follows:
1 ΔL
α¼ ; (8.24)
L0 ΔT
where α is the coefficient of thermal expansion (K1), L0 is the original sample

length (m), and △T ¼ TT0 is the temperature interval (T0 ¼ 293 K). The values
of CTE α for specific temperature ranges are shown in Fig. 8.18b. These results
show no significant effect of the porosity and the pore direction in the lotus-type
porous nickel on the CTE.
For prediction of the CTE of metal-matrix composite, the models are derived by
Kerner [32] and Schapery [33]. Kerner’s model gives the CTE of composite as
Kp Km
þ Vp ð1 Vp Þðαp αm Þ
αc ¼ α ; (8.25)
ð1 Vp ÞKm þ Vp Kp þ ð3Kp Km =4Gm Þ
is the rule-of-mixtures value given by Vp αp þ ð1 Vp Þαm. K and G are the

where α
bulk and shear moduli, V is the volume fraction, and α is the coefficient of thermal
expansion. The subscripts c, p, and m denote composite, particle, and matrix,
respectively.
The Schapery’s model is written as
ð1=Kc Þ ð1=Kp Þ
αc ¼ αp þ ðαp αm Þ : (8.26)
ð1=Km Þ ð1=Kp Þ
In the case of porous metals, the Kp in Eqs. (8.25) and (8.26) is zero and then
αc ¼ αm. This indicates the CTE of lotus-type porous metals depends on that of the
matrix, which the effect of porosity on CTE can be neglected.
8.6 Corrosion
It is well known that the properties of passive films on stainless steels depend on the
type of solution used for the corrosion tests. Stainless steels exhibit different corro-
sion resistance in alkaline, neutral, and acidic environments, since the mechanisms of
formation and breakdown of the passive film are different [34]. AISI 316L stainless
steel has been used as a surgical implant material for many years exhibiting sufficient
corrosion resistance under body fluids which contain Cl ion as well as many other
components such as water, dissolved oxygen, bacteria, cells, enzymes, and proteins.
The improved corrosion resistance of type AISI 316L stainless steel in a chloride
environment is attributed to the addition of molybdenum which is effective in
stabilizing the passive film in the presence of chlorides [35]. Recently it has been
found that nitrogen addition to austenitic stainless steel improves the pitting corrosion
resistance [36–38]. In addition, nitrogen as an alloy component allows a nickel
content reduction in the austenitic stainless steel, since the allergic reactions caused
by this element is suppressed.
Potential applications of this new family of stainless steels include components
subjected to corrosion and wear mechanisms, structural components, as well as
medical and biocompatible parts, because nitrogen confers high-yield strength,
ultimate tensile strength and ductility, as well as fracture toughness, low-magnetic
permeability, very favorable localized corrosion resistance, and higher wear and
abrasion resistance. Due to the combination of these excellent properties, currently
there is also much interest in the fabrication of porous parts made of high-nitrogen
Ni-free stainless steels.
8.6 Corrosion 203
Reference
Purging N2 Working
Potentiostat
ASTM G-5 type corrosion cell Ag/AgCl Counter
Reference
Electrolyte: NaCl 3.5% +HCl 1M electrode:
Room temperature
Electrodes: Auxiliary electrode: Pt,
Working electrode: Ni-free SUS, X-Y Recorder
Reference electrode: Ag/AgCl Temperature
Scan Rate: 1 mV/s controlled bath
Open circuit potential: 60 min Platinum wire
Counter electrode
Reaction cell
Salt bridge 1mol KCl sol.
Specimen
Electrolyte
Fig. 8.19 Experimental setup for measurement of electrochemical polarization curves
The corrosion behavior of porous materials in general differs from that of the
bulk materials of the same composition. Moreover, the pore morphology, porosity,
pore size, and pore surface condition also interfere in the degradation process, in
view of the fact that a porous material is attacked not only on its surface but also
from inside. Alvarez et al. measured the corrosion rates, corrosion potentials, and
breakdown potentials in dehydrogenized porous and nonporous specimens in order
to investigate the corrosion resistance of lotus high-nitrogen Ni-free stainless steels
and to compare it with the nonporous material [39].
Electrochemical polarization experiments were conducted using a potentiostat
unit connected to a potential and current x–y recorder. The results were finally
processed in a personal computer using a data acquisition program. The
potentiodynamic polarizations were performed at a scan rate of 1 mV s1, and
the scans were started at +100 mV with respect to the open-circuit potential. The
details of the experimental setup are illustrated in Fig. 8.19.
The nitrogen effect in the polarization curves was examined firstly in nonporous
samples. Potentiodynamic polarization curves for the nonporous nitrided and
non-nitrided alloys are shown in Fig. 8.20, where the current density is a measure
of the corrosion rate. It was confirmed that the nitrogen alloying suppressed the
pitting corrosion even in such a low pH and high Cl1 concentration solution.
A more stable passivity region was established due to the nitrogen alloying.
Moreover, the solution-nitrided samples exhibited lower passivation current Ipass.
The breakdown potentials for all the solution-nitrided samples were found to be
more noble than the non-nitrided samples, thus exhibiting improved corrosion
resistance in 3.5 wt% NaCl + 1 M HCL. The high-pitting potential of high-nitrogen
Fig. 8.20 Nitrogen alloying effect in the potentiodynamic polarization curves of the investigated
samples (Reprinted with permission from [39] © 2007 Elsevier Ltd)
steels is attributed to the amount of N dissolved within the austenite. Together with
a sufficient Cr and Mo content, which bring about a passive layer (Cr) and a
pre-passive film (M), respectively, nitrogen leads to an increase of the pH value
within the surrounding medium by a decomposition of NH3 to NH4 + + OH
[40]. Indeed, Fe-23wt%Cr-2wt%Mo-1wt%N alloy resulted to have the higher
corrosion potential. Inspection of the non-nitrided samples after polarization
showed pitting in the surfaces of the AISI 316L and also in AISI 446 stainless
steel and Fe-25wt%Cr, Fe-23wt%Cr-2tMo alloys, while in the solution-nitrided
samples, the surface after the polarization looked only etched by the electrolyte.
When comparing results for the porous and nonporous alloys, in the former,
a clear susceptibility to the attack of chloride ions was found. Figure 8.21 shows the
polarization curves of the porous and nonporous solution nitride alloys. A clear
defined passive region was not observed in the porous alloys; in contrast, the
nonporous alloys exhibited a very wide passive region. The passive current was
2 orders of magnitude higher for the porous alloys than for nonporous alloys.
Regarding the breakdown potential, the nonporous alloys exhibited a slightly
more noble breakdown potential. The porous configuration could have introduced
many small crevices which are conductive to localized corrosion, especially in the
internal surface of the pores that cannot be mechanically polished. In these sites a
marked increase in the acidity within a confined space could be able to accelerate
localized corrosion.
8.6 Corrosion 205
Fig. 8.21 Potentiodynamic polarization curves of lotus Fe-25wt%Cr-1wt%N, Fe-23wt%Cr-2wt%

Mo-1wt%N, and AISI446-1wt%N steels and nonporous steels in 3.5wt%NaCl + 1 M HCl solution
(Reprinted with permission from [39] © 2007 Elsevier Ltd)
Figure 8.22 shows the electropolishing effect in the porous specimens. This
study revealed that solution nitriding + electropolishing produces a polarization
curve with a very wide passive region. The current density in this passive region is
within 14–300 μA cm2, almost as low as the nonporous samples’ current density.
Therefore, a way of increasing the corrosion resistance of lotus alloys is to
electropolish the porous surface in order to remove deformed layer, improve
surface roughness, and also form a durable passive film on the surface. This more
uniform oxide layer resulting from electropolishing will increase corrosion resis-
tance in stainless steels. It is worth noting that the electropolishing treatment
applied in the porous alloys did not alter the breakdown potential significantly.
However, the passive current densities in the active–passive transition were greatly
reduced and the passivating ability, given by the wide passivation range of
the polarization curves, also increased by this final surface finish. These results
indicate that a higher corrosion resistance is obtained after the electropolishing
surface treatment.
The results of pitting corrosion susceptibility can also be interpreted using the
images of the surface microstructure after corrosion obtained by scanning electron
microscopy. Selected SEM micrographs of lotus nitrided alloys after the polariza-
tion test in 3.5wt%NaCl + 1 M HCl are shown in Fig. 8.23a–d. SEM observation
revealed that lotus AISI 316L stainless steel experienced severe localized
Fig. 8.22 Potentiodynamic polarization curves of lotus Fe-25wt%Cr-1wt%N, Fe-23wt%Cr-2wt%
Mo-1wt%N, and AISI446-1wt%N steels with and without electropolishing using an electrolyte
composed in phosphoric acid 56 %, sulphuric acid 27 %, glycerine 7 %, and DI water at 333 K
(Reprinted with permission from [39] © 2007 Elsevier Ltd)
Fig. 8.23 SEM micrographs of the investigated lotus alloys after the polarization in 3.5wt%
NaCl + HCl 1 N. (a) AISI316L, (b) Fe-25wt%Cr-1wt%N, (c) AISI446-1wt%N, and (d) Fe-23wt
%Cr-2wt%Mo-1wt%N (Reprinted with permission from [39] © 2007 Elsevier Ltd)
8.6 Corrosion 207
Fig. 8.24 Weight loss in 10wt%FeCl3・6H2O after 48 h immersion for the porous and nonporous
high-nitrogen Ni-free stainless steels (Reprinted with permission from [39] © 2007 Elsevier Ltd)
corrosion. The metal base (zone A in Fig. 8.23a) as well as the internal surroundings
of the pores (zone B in Fig. 8.23a) were attacked. The pits found in the metal base of
AISI 316L consist of an open center, perforated middle region, and an outer ring.
The center and the middle regions are typical perforated or “lacy” pit morphology
in which subsurface undercutting expands the pit from the central initiation point.
Contrastingly, no obvious corrosion pits were found in any of the high-nitrogen
Ni-free alloys (Fig. 8.23b–d). In the case of Fe-25wt%Cr-1wt%N and AISI 446-1wt
%N (Fig. 8.23b, c, respectively), it was observed that the internal wall of the pores,
marked with arrows in Fig. 8.23b, c, seems to be degraded and not smooth like in
the case of Fe-23wt%Cr-2wt%Mo -1wt%N Fig. 8.23d). This difference in the
microstructure of the inner part of the pores in the high-nitrogen Ni-free stainless
steels is probably because Fe-23wt%Cr-2wt%Mo-1wt%N alloy exhibited the lower
corrosion rate under the experimental conditions. These results confirm the higher
corrosion susceptibility of AISI 316L compared to high-nitrogen Ni-free stainless
steels. In every high-nitrogen Ni-free alloy investigated, severe open perforated pit
structure was not observed after the polarization test.
The results after immersion test in 10wt%FeCl36H2O solution for 172.8 ks are
shown in Fig. 8.24. It was observed that lotus AISI 316L exhibited the highest
weight losses. Metallographic observations revealed that AISI 316L stainless steel
coupons displayed much more pits than the rest of the solution nitride samples.
Moreover, the pits developed on AISI 316L were wider and deeper with average
size of 100 μm for the nonporous specimen and 150 μm for the lotus specimen. On
the contrary the high-nitrogen Ni-free stainless steel specimens after 172.8 ks
immersion showed a very smooth and bright surface with a small number of
shallow pits with a diameter not exceeding 20 μm. The corrosion rate determined
by weight loss method of the high-nitrogen Ni-free stainless steel specimens in this
highly oxidizing chloride environment was less than 0.04 mm y1. In contrast, AISI
316L stainless steel-corroded samples showed a very high average corrosion rate.
Corrosion rate of duplicate runs determined by weight loss method showed average
values of 0.86 and 1.52 mm y1 for nonporous and lotus AISI 316L stainless steel,
respectively. Therefore, it can be concluded that lotus high-nitrogen Ni-free stain-
less steels possess the general corrosion resistance much better than conventional
austenitic stainless steels in acidic chloride environment.
The results of this investigation showed that the general corrosion resistance as
well as the localized corrosion resistance of lotus high-nitrogen Ni-free stainless
steels is higher than AISI 316L. From the microscopic observation and the electro-
chemical characterization, the alloy with best performance was lotus Fe-23wt%Cr-
2wt%Mo-1wt%N since it resulted to be immune to pitting attack under these testing
conditions. From the weight loss measurements, lotus Fe-25wt%Cr-1wt%N alloy
tolerated very good the immersion in the severe aggressive media much better than
any other alloy.
References
1. Xie ZK, Ikeda T, Okuda Y, Nakajima H (2004) Mater Sci Eng A 386:390–395
2. Xie ZK, Ikeda T, Okuda Y, Nakajima H (2004) Jpn J Appl Phys 43:7315–7319
3. Xie ZK, Ikeda T, Okuda Y, Nakajima H (2004) Mater Sci Forum 449–452:661–664
4. Kimura S (1977) Architectural sound and anti-noise plan. Shokokusha Publishing, Tokyo,
p 142 (in Japanese)
5. Xie ZK, Ikeda T, Okuda Y, Nakajima H (2003) J Jpn Inst Metals 67:708–713
6. Okuda Y (1990) Proc J Acoustic Soc Jpn 581
7. Akiyama S, Itoh M, Ishii E (1991) National Industrial Research Institute. Report of Kyushu,
p 2928
8. Nisimaki S (1971) Electro-acoustics and vibration. Corona Publishing, Tokyo, Japan, p 79
9. Beranek LL (1950) Acoustic measurement. Wiley, New York
10. Igarasi J (1970) Sound and vibration. Kyoritsu Shuppan Co., Tokyo, p 37
11. Behrens E (1968) J Compos Mater 2:2–17
12. Perrins WT, McKenzie DR, McPhedran RC (1979) Proc Roy Soc London Ser 369:207
13. Han LS, Cosner AA (1981) J Heat Transfer 103:387–392
14. Sangani AS, Yao C (1988) Phys Fluids 31:2435–2444
15. Mityushev V (1999) Proc Roy Soc London A 455:2513–2528
16. Moctezuma-Berthier A, Vizika O, Adler PM (2002) Trans Porous Media 49:331–332
17. Ogushi T, Chiba H, Nakajima H, Ikeda T (2004) J Appl Phys 95:5843–5847
19. Suematsu Y, Hyun SK, Nakajima H (2004) J Jpn Inst Metals 68:257–261
20. Tane M, Hyun SK, Nakajima H (2005) J Appl Phys 97:103701-1–103701-4
21. Tane M, Ichitsubo T (2004) Appl Phys Lett 85:197–199
22. Archie GE (1942) Trans AIME 146:54–63
References 209
23. Eshelby JD (1957) Proc Roy Soc London A 241:376–396

24. Mori T, Tanaka K (1973) Acta Metall 21:571–574
25. Hatta H, Taya M (1986) J Eng Mater 24:1159–1172
26. Ohnishi Y, Hyun SK, Nakajima H (2006) In: Nakajima H, Kanetake N (eds) Porous metals and
metal foaming technology. Jpn Inst Metals, Sendai, pp 423–426
27. Noudem JG, Reddy ES, Schmitz GJ (2003) Physica C 390:286–290
28. Castano FJ, Nielsch K, Ross CA, Robinson JWA, Krishnan R (2004) Appl Phys Lett
85:2873–2875
29. Onishi H, Hyun SK, Nakajima H, Mitani S, Takanashi K, Yakushiji K (2008) J Appl Phys
103:093539
30. Igarashi H, Okazaki K (1976) J Am Ceram Soc 59:371–372
31. Tane M, Hyun SK, Nakajima H (2006) Scr Mater 54:545–552
32. Kerner EH (1956) Proc Phys Soc London B 69:808–813
33. Schapery RA (1968) J Comp Mater 2:380–404
34. Dillon CP (1995) Corrosion resistance of stainless steels. Marcel Dekker, New York, p 258
35. Seah KHW (1993) Corros Sci 34:1841–1851
36. Fossati A, Borgioli F, Galvaneto E, Bacci T (2006) Corros Sci 48:1513–1527
37. Di Schino A, Barteri M, Kenny JM (2003) J Mater Sci 38:4725–4733
38. Rechsteiner A, Speidel M (1997) Proceedings of 1st European stainless steel conference, vol
2, Florence, p107
39. Alvarez K, Hyun SK, Tsuchiya H, Fujimoto S, Nakajima H (2008) Corros Sci 50:183–193
40. Ernest P, Laycock NJ, Moayed MH, Newman RC (1997) Corros Sci 39:1133–1136
Chapter 9
Processing of Lotus Metals
Abstract In order to use lotus metals for various industrial products, processing is
indispensable. Here, two important topics are chosen: joining technique and severe
deformation without shape change of metals.
Keywords Equal-channel angular extrusion process • Laser beam • Vickers

hardness • Welding
9.1 Weldability
9.1.1 Weld Fusion Property of Lotus Copper
For industrial use of the porous metals as various parts, reliable joining techniques
such as welding is required as well as processing techniques. There have been some
reports on laser and arc welding of foamed aluminum manufactured by using foaming
agent, TiH2, or CaCO3 [1–3]. However, weldability of porous metals with controlled
pore direction such as lotus metals has not been investigated, although the welding is
necessary for fabricating engineering materials. Murakami et al. [4] investigated the
weldability of lotus copper, and in their study the laser beam with narrow weld bead
width and heat-affected zone was applied to the welding of lotus metal. Figure 9.1
shows schematic views of specimens using laser welding. The laser beam was
delivered using a 600-μm-diameter optical fiber. The focal length for Nd:yttrium–
aluminum–garnet (Nd:YAG) laser with a 1,064-nm wavelength was 100 mm. The
diameter of the laser beam at the focus point was 300 μm. The laser beam was
irradiated at an 80 forward angle relative to a workpiece surface damage to the optics
by a reflection of laser beam. Argon was used as a shielding gas with a flow rate of
5.0 104 m3 s1. Figure 9.2 shows a relation between the welding speed and the
penetration depth of the weld bead in depth at the thickness of 5 mm from the work
piece surface when the laser beam power of 3.2 kW was irradiated on both types of
samples (a) and (b) shown in Fig. 9.1. The penetration depth in both directions

212 9 Processing of Lotus Metals
Fig. 9.1 Schematic views of specimens during laser welding in the lotus copper with the pore
growth (a) perpendicular and (b) parallel to the work piece surface (Reprinted with permission
from [4] © 2003 Elsevier Science B.V.)
Fig. 9.2 Relation between welding speed and penetration depth of the weld bead in depth at the
thickness of 5 mm from the work piece surface when the laser beam power of 3.2 kW was
irradiated on both types of sample (a) and (b) shown in Fig. 9.1 (Reprinted with permission from
[4] © 2003 Elsevier Science B.V.)
decreases with increasing welding speed. The penetration depth in the parallel direc-
tion (filled square) is always shallower than that in the perpendicular direction (filled
circle). Figure 9.3 shows the top view of the specimens after laser welding in the
thickness of 5 mm at the constant laser beam power of 3.2 kW in the welding speed
range of 1.67–16.7 mm s1 in (a)–(c) perpendicular and (d)–(g) parallel pore
directions. In the perpendicular direction, only the groove was observed in the
condition of 16.7 mm s1 as shown in Fig. 9.3a. The weld beads were found below
the welding speed of 8.33 mm s1 as shown in Fig. 9.3b, c, while a smooth weld bead
was observed at the welding speed of 3.33 mm s1. In contrast, in the parallel
direction, no weld bead was made in the welding speed range from 3.33 to
16.7 mm s1, and a smooth weld bead on all specimens was not observed. At the
9.1 Weldability 213
Fig. 9.3 Top views of laser welded specimen at different welding speeds with 3.2-kW constant
laser power in 5-mm thickness in (a–c) perpendicular and (d–g) parallel to the pore growth
directions. Welding speed: (a) and (b) 16.7 mm s1, (b) and (e) 8.33 mm s1, (c) and
(f) 3.33 mm s1, and (g) 1.67 mm s1 (Reprinted with permission from [4] © 2003 Elsevier
Science B.V.)
welding speed of 1.67 mm s1, the weld bead was formed only at the end of weld,
though nonuniform. These indicate that the formation of the weld bead has strong
anisotropy of pore direction in relation to the irradiation direction of laser beam.
The effect of the laser power on the penetration depth in the thickness of 4 mm
for two pore directions at the constant welding speed of 3.33 mm s1 was shown in
Fig. 9.4. The penetration depth of the weld bead perpendicular to the pore growth
direction shows a sharp rise at the laser beam power of 2.4 kW. On the other hand,
the penetration depth in the parallel to the pore growth direction increases gradually
in comparison with that in the perpendicular. These facts indicate that the keyhole
formed at the laser-irradiated point perpendicular to the pore growth direction may
be attributed to the metal vaporization [5] more than the laser beam power of
2.4 kW, but not in the case parallel to the pore direction. Figure 9.5 shows the cross
sections of laser welding direction in the thickness of 4 mm in the welding speed of
3.33 mm s1 at the laser beam power of 3.2, 2.5, and 2.0 kW (a)–(c) perpendicular
and (d)–(f) parallel to the pore growth directions. As shown in Fig. 9.5a, the weld
part at the laser beam power of 3.2 kW in the perpendicular case was completely
melted through the bottom; shortly, the fully penetrated weld bead was formed. The
shape of the fusion boundary is nearly straight. The straight line of the fusion
boundary is nearly parallel to the direction of the original pores. Some large and
many small pores in the weld metal were observed in the weld bead. The weld metal
at the laser beam power of 2.5 kW in the perpendicular shown in Fig. 9.5b did not
Fig. 9.4 Effect of laser power on penetration depth of weld bead for perpendicular and parallel to
the pore growth direction in depth of 4 mm thick from the work piece surface at 3.33 mm s1
welding speed (Reprinted with permission from [4] © 2003 Elsevier Science B.V.)
Fig. 9.5 Cross sections of laser weld perpendicular to the welding direction in 4-mm thick-
ness in the welding speed of 3.33 mm s1 at the laser beam power of 3.2, 2.5, and 2.0 kW
(a–c) perpendicular and (d–f) parallel to the pore growth directions, which are slightly
inclined to the work piece surface with about 10 (Reprinted with permission from [4]
© 2003 Elsevier Science B.V.)
9.1 Weldability 215
Fig. 9.6 (a) X-ray radiograph of the workpiece after welding with the welding speed of
3.33 mm s1 at the laser beam power of 3.2-kW depth of 4 mm thick. The drilling points
determined were in the open circle including the shadow of weld pore as A-1 and A-2. (b)
Dependence of mass spectral intensity of gas detected from A-2 on drilling time and the distance
from the weld bead surface (Reprinted with permission from [4] ©2003 Elsevier Science B.V.)
melt through the opposite side. Some larger pores evolved by the weld were
observed in the weld metal. At the laser power of 2.0 kW shown in Fig. 9.5c, the
weld metal was not observed, and the diameters of the grooves formed instead of
weld bead were about 0.6 mm. In the parallel direction case, the shape of groove
was similar to Fig. 9.5c, though the groove depth increased with increasing laser
power. Obviously, the shape of the weld bead and the groove exhibits anisotropy
with respect to the pore growth direction. Thus, the weldability of lotus copper in
the perpendicular direction is more significant than that in the parallel direction.
X-ray radiograph of the work piece after welding with the welding speed of
3.33 mm s1 at the laser beam power of 3.2 kW taken in depth of 4 mm thick is
shown in Fig. 9.6a. The weld bead after welding was drilled in 3-mm diameter in
the vacuum chamber, and the residual gas in the pores was analyzed using a mass
spectrometer. The drilling points were determined in the open circle including the
shadow of the weld pores such as A-1 and A-2. Figure 9.6b shows the drilling time
and the distance from the weld bead surface dependence of mass spectral intensity
of residual gas detected from the pores in weld metal (Fig. 9.5a) with the welding
speed of 3.33 mm s1 at the laser beam power of 3.2 kW in depth of 4 mm thick.
When the drilling depth increases to more than 0.5 mm, various kinds of gases were
clearly detected at the same time with high intensity, which were evolved from the
large weld pore at this point. These gases are identified as H2, H2O, N2, Ar, and CO2
by the mass spectrometer. No obvious change in the intensity was detected during
drilling for other gases up to 100 mass numbers. The gas concentration was
determined using a net intensity, which eliminated the difference between the
peak intensity and the background intensity in depth of 0.5 mm from the surface.
Some small hydrogen peaks with comparably high intensity were detected in depth
of 0.4 and 0.48 mm. These may correspond to small weld pores which were formed
during welding. The major component of A-1 was argon comprised beyond 73 vol.
%, and other components were 14 vol.% nitrogen and water. The major component
of A-2 was water which comprised about 60–70 vol.%, and other components were
about 20 vol.% argon, several to 10 vol.% nitrogen, and several volume percent
hydrogen.
It is generally considered that the pores in the laser weld bead of dense materials
are formed by the shielding gas trapped in the weld pool due to keyhole instability
by vaporized molten metal [6]. Therefore, it is likely that argon of the shielding gas
was detected as a main component. The detected nitrogen may result in the gas from
the air which is trapped together with the shielding gas. In their study, however, the
presence of a large amount of water and small amount of hydrogen is not explain-
able in terms of only keyhole instability. The reason why those components are
present in the pores is that high-pressure hydrogen gas remains in the closed
original pores in the lotus copper formed during the unidirectional solidification
under high pressure. It is considered that at remelting of workpiece by laser
welding, the remaining hydrogen gas in the original pore was released into the
molten weld metal and then coalesced into the large pore which is generated by
keyhole instability. Subsequently the hydrogen gas reacted with oxygen containing
in air trapped with the shielding gas, and then water vapor was evolved. Besides,
there is a small amount of hydrogen gas that does not react with oxygen in some
pores. Nitrogen gas in air was also trapped but remained in the pores because of no
reaction with molten copper. Similar reason was supposed in the case of CO2 gas
formation. Therefore, the large pores in the weld metal are comprised of the
shielding gas, the vaporized copper, the air trapped with the shielding gas, and
the hydrogen gas. On the other hand, the small gas pores are composed of the
supersaturated hydrogen supplied from the original closed pores in lotus copper.
As mentioned above, significant anisotropy in the morphology of laser weld
metal was observed, depending on the pore growth direction. The laser beam
mostly reflects on the copper surface due to the high reflectivity of copper
[7, 8]. A part of the laser beam penetrates deeply into the pores in the lotus copper
with cylindrical pores perpendicular to the welding direction, since the direction of
pore growth is nearly identical to the incident laser beam direction; the multiple
reflections of the laser beam at the pore inner wall increases the amount of heat
input absorbed to the specimen. In addition, two-dimensional thermal conductivity
on the parallel plane to the specimen surface is low, which enhances the melting of
the specimen. The amount of heat input for the lotus copper with cylindrical pores
9.1 Weldability 217
parallel to the welding direction is smaller than that of the perpendicular, because of
no multiple reflection and high-thermal conductivity. Moreover, the molten copper
in the weld pool is blown out more significantly by high-pressure hydrogen gas
released from the original closed pores.
In the laser beam welding of lotus copper plate, the pore growth direction should
be perpendicular to the work piece surface. When the lotus copper plate with the
original pores parallel to the work piece is used, the surface treatment should be
required in order to increase the laser beam absorption.
9.1.2 Weld Fusion Property and Joint of Lotus Magnesium
Figure 9.7 shows the bead appearance and macrostructures in the cross section of
the welded specimen with pores perpendicular to the specimen surface at nominal
laser power of 0.8, 1.0, and 1.2 kW and spot diameters of 0.3, 0.45, and 0.6 mm
[9]. The fusion area of the specimen at the laser power of 0.8 kW and the spot
diameter of 0.6 mm did not penetrate to the bottom although complete penetration
of the fusion area was achieved for the spot diameter of 0.45 mm as well as for
increased laser powers. This is because energy density in the former case is lower
than that in the latter case. In the weld metal, some blowholes, which form due to
the remnant hydrogen gas in the original closed pores, were observed at the laser
power of 1.2 kW and the spot diameter of 0.45 mm, as shown in Fig. 9.7.
Macrostructures of the welded joint with pores parallel to the specimen surface
at different nominal laser powers are shown in Fig. 9.8. The weld at the laser powers
of 1.5 and 2.0 kW completely melted the bottom, while the weld at the laser power
of 1.0 kW did not melt the bottom completely. The blowholes, which are indicated
by black arrows in Fig. 9.8, were observed for all the conditions. There were more
blowholes in the weld metal of the specimen with pores parallel to the specimen
surface (parallel case) than in the weld metal of the specimen with pores perpen-
dicular to the specimen surface (perpendicular case). This was because the
blowholes were caused by the remnant hydrogen gas in the original closed pores,
and the number of original closed pores for the specimen in the parallel case was
much more numerous than that for the specimen in the perpendicular case.
By comparing Fig. 9.7, which represents the laser power of 1.0 kW and the spot
diameter of 0.6 mm, with Fig. 9.8a, it was observed that the profiles of the weld
metal in the perpendicular and parallel cases were different, although the welding
condition was identical. The depth of penetration in the former case is greater than
that in the latter case. In the previous report [4], the melting property of lotus copper
differed greatly according to the pregrowth direction. The weld bead in the parallel
case was not formed in most welding conditions, and only a groove was observed.
The difference in the melting property was attributed to the high reflectivity of
copper for a laser beam. On the other hand, the reflectivity of magnesium is much
lower than that of copper. Therefore, it is expected that the melting property of lotus
magnesium is mainly controlled by the thermal conductivity.
Fig. 9.7 Bead appearance and macrostructure of laser welded specimen with the pores perpen-
dicular to the specimen surface at the laser powers of 0.8, 1.0, and 1.2 kW and the spot diameters of
0.3, 0.45, and 0.6 mm (Reprinted with permission from [9] ©2006 Elsevier B.V.)
The effect of pore growth direction on the joint strength of lotus magnesium was
investigated. Figure 9.9 shows the geometry of tensile specimens of the base metal
and the welded joint. A tensile test specimen with a gage section of 1.8 mm 5 mm
was cut out of the specimen before the welding for the base metal and after the
welding for the welded joint using the spark erosion wire-cutting machine. The
tensile test was performed on the specimen in a tensile test machine at a rate of
1.67 102 mm s1 at room temperature. The tensile strengths of the weld beads
and the base metals in the parallel and perpendicular cases are shown in Fig. 9.10
9.1 Weldability 219
Fig. 9.8 Macrostructure

of welded specimen with
the pores parallel to the
specimen surface at the
spot diameter of 0.6 mm
and the laser powers of
(a) 1.0 kW, (b) 1.5 kW, and
(c) 2.0 kW (Reprinted with
Elsevier B.V.)
together with the cross-sectional view of the fractured joints. The open and closed
circles in Fig. 9.10 denote the samples fractured in the base metal or in the weld
bead, respectively. The average tensile strengths of the base metal in the parallel
and perpendicular cases were approximately 55 and 30 MPa, respectively.
The average tensile strength of the weld bead with the pores perpendicular to the
specimen surface was 29 MPa, which was similar to that of the base metal because
the joint was fractured in the base metal as shown in Fig. 9.10b. On the other hand,
the weld bead in the parallel case was fractured at the joint interface between
the weld metal and the base metal as shown in Fig. 9.10a. The average joint strength
in the parallel case was 41 MP, which is lower than that of the base metal for
the same case.
Fig. 9.9 Geometry in the tensile test specimens of base metal and welded joint with the pores
(a) perpendicular and (b) parallel to the specimen surface (Reprinted with permission from [9]
Fig. 9.10 Tensile strength of the base metal and welded joint with the pores perpendicular
and parallel to the specimen surface together with cross section of fractured joint. (a) Fractured
in the weld bead and (b) fractured in the base metal (Reprinted with permission from [9] © 2006
Elsevier B.V.)
9.2 Equal-Channel Angular Extrusion Process 221
Ikeda et al. [10, 11] reported the tensile strength of nonporous magnesium
produced by unidirectional solidification. The tensile strength parallel to the
solidification direction was larger than that perpendicular to the solidification
direction. The highly preferred orientations of the nonporous and lotus
magnesium fabricated by unidirectional solidification were h0002i and h1120i
[10, 11]. The slip system of magnesium at room temperature is only (0002) h1120i
[12]. When a load perpendicular to the pore growth direction is applied to porous
magnesium with a h11 20i orientation, the (0002) h1120i slip occurs because the
normal direction of the (0002) basal plane is perpendicular to the solidification
direction. On the other hand, when a load parallel to the growth direction is
applied to porous magnesium with h0002i and h1120i orientations, the Schmid
factor is close to 0. Therefore, the anisotropic tensile strength of the lotus
magnesium is controlled not only by the pore growth direction but also by the
crystallographic direction.
9.2 Equal-Channel Angular Extrusion Process
Suzuki et al. investigated the possibility to improve the mechanical properties of

the metallic part of porous metals by plastic deformation [13]. As a kind of
porous metals, lotus metals were selected, which are expected to be suitable for
improvement of metallic parts by plastic deformation. Since the lotus metals have
more metallic parts than metallic foams with up to 90 % of porosity, it is expected
that the pores do not disappear and the effect of plastic deformation is obvious.
Even if the pores are compressed, metallic part work hardens so that specific
strength may increase. The densified materials are expected still to have micro- or
nano-pores, which induce novel functions of materials. As a plastic deformation
process the equal-channel angular extrusion (ECAE) [14] was adopted as shown
in Fig. 9.11. In this process a sample rod is inserted in the entry channel and
extruded to the exit channel through a die with a channel angle Φ. At the corner
between the two channels, the sample is sheared with severe plastic deformation.
Because of several advantages, nowadays this process has been investigated
actively [15]. First, as the shape and the diameter of the sample rod after the
extrusion are the same as that before the extrusion, it is thus possible to extrude
the same sample many times through the same die. Secondly, as various processing
routes can be selected arbitrarily, it is possible to vary the deformation pattern.
Thirdly, lower hydrostatic stress than forging or extrusion does not cause
significant densification.
Until now there is no report about ECAE of porous metals. Severe plastic
deformation of porous materials has several special characteristics compared with
that of nonporous materials because of the densification process and thus requires
newly established plastic deformation models. Here the fundamental plastic defor-
mation behavior of lotus copper during ECAE process is described.
Fig. 9.11 Schematic load punch

illustration of ECAE displacement
process (Reprinted with
permission from [13] © 2008 punch
die
Elsevier B.V.)
entry Φ
channel specimen
exit
Ψ channel
extruded length
Figure 9.12 shows the cross section of lotus copper which was extruded halfway
of the first pass through the die with a channel angle of 90 . In the entry channel part
(Fig. 9.12b), some in polygonal shapes still remain, although a lot of pores are seen
to be compressed. The pore morphologies are similar to those of specimens
deformed uniaxially by compression tests [16]. In the exit channel close to the
corner, almost no pores in the original size are observed but many streaks inclined
toward the exit channel axis (Fig. 9.12c, d). Many pores remain in the leading end
of the specimen, which is the initial unsteady part. Figure 9.13 shows the cross
section of the completely extruded sample.
The volume change ΔV during the process can be calculated by
V0 V 1 p0 p1
ΔVð%Þ ¼ 100 ¼ 100; (9.1)
V0 100 p1
where V and p are the volume and the porosity, respectively. The suffixes 0 and
1 mean the states before and after the deformation, respectively. The porosity of the
sample cut from the part in the entry channel (Fig. 9.12b) was 17 %, and that in the
exit channel, shown within the dotted line in Fig. 9.13 was 6 %. The volume change
in the entry channel was calculated to be 40 % with p0 ¼ 50 % and p1 ¼ 17 %, and
that in the exit channel was calculated to be 12 % with p0 ¼ 17 % and p1 ¼ 6 %. As
shown in Fig. 9.13, the pores after the shear deformation are inclined from the
central axis. This inclination can be explained by using a geometrical model shown
in Fig. 9.14. One assumes a square region “klmn” on a cross section of a sample rod
(Fig. 9.14a); the square turns and deforms into a parallelogram (Fig. 9.14b) shape
by the shear stress at the corner. In the case of nonporous materials, the line “lm”
inclines theoretically 26.5 to the exit channel axis [14].
In the case of lotus metals which are densified during the shear deformation, the
area “lmn” should be decreased. As the volume reduction by shearing at the corner
was 12 %, the area S0 of the square “klmn” should be reduced. Therefore, the side
“nm” reduces from 1 to 0.88 as shown in Fig. 9.14b, and the inclination θ of “lm”
can be estimated as 28 . The observed inclination of pores agrees well with the
estimation. From these results, the deformation behavior during the extrusion
through ECAE process can be schematically illustrated in Fig. 9.15.
Lotus-type
Pore porous copper
specimen
200µm
Pore
c d
aligned micro-pores
200µm 10µm
Fig. 9.12 Cross sections of a lotus copper specimen before and after extrusion approximately
36 mm in a single pass through an ECAE die with a channel angle of 90 . (a) Before extrusion. (b)
Compressed part in the entry channel. (c) Shear-deformed part in the exit channel close to the
corner. The arrows indicate micropores. (d) Micropore magnified in the window in (c) (Reprinted
with permission from [13] © 2008 Elsevier B.V.)
The extruded samples after the second pass are shown in Fig. 9.16a, b. On the
cross section of the extruded sample via route A, these pores aligned in one
direction are observed. Even after the second pass, the walls of the pores seem
not to be bonded, since hydrogen captured in the pores prevented complete com-
pression pores. On the other hand, the pores were opened again in the case of route
C where the sample was rotated axially 180 from the first pass. This deformation
behavior can also be explained by using the deformation model in Fig. 9.15. It may
be possible to change the pore morphology by ECAE process.
28°
porosity : 6% 2mm
Fig. 9.13 Cross section of a lotus copper specimen after completely extruded through an ECAE
die with a channel angle of 90 in a single pass. The porous metal within the square drawn with a
dotted line has a porosity of 6 % (Reprinted with permission from [13] © 2008 Elsevier B.V.)
a b
k n
k n
l m m
n m
k n non-porous
SNP
1 1
S0
θ=26.5°
l o
m k l
2
1
porous n m
SP
θ=28°
k l o
1.88
Fig. 9.14 Geometrical model for the deformation behavior of sample rod through ECAE process
with a channel angle of 90 . A square “klmn” on the cross section shown in (a) turns and deforms
into a parallelogram shown in (b). The area S0 of the square “klmn” in the entry channel is the
same as the area SNP of “klmn” in the exit channel. Sp is 12 % smaller than S0 (Reprinted with
densification by uni-
axial compression
pore
porosity
50%
densification by shearing
Porosity 6%
micro-pores
porosity
17%
Fig. 9.15 Schematic illustration of deformation of lotus metals through ECAE process with a
channel angle of 90 . It demonstrates the case of porosity of 50 % before the extrusion (Reprinted
with permission from [13] © 2008 Elsevier B.V.)
Figure 9.17 shows the diagrams of the punch load–displacement and extruded
length. In the beginning of the loading, although the load increases with increasing
punch displacement, the extruded length does not increase. In this stage only
densification takes place in the entry channel by uniaxial compression. Then at
about 20 mm of the punch displacement, the extrusion began and the increment of
the extruded length is close to that of the punch displacement. As the initial length
of the sample rod is 50 mm, the volume reduction is 40 % when the punch
displacement is 20 mm. The value of the volume reduction estimated from the
punch displacement agrees well with that estimated from the porosity change. After
this stage densification in the entry channel does not take place, but the extrusion is
done by shear deformation.
The mechanism of the onset of the extrusion is shown in Fig. 9.18, assuming no
friction between the rod and the die. Although the sample is loaded by the punch,
the normal force F3 of the shear flow stress reacts against the punch load F1. At the
beginning of the pressing, the porosity is as high as 50 % that the compressive flow
stress σ of the entry channel by uniaxial compression is smaller than the stress
caused by the punch load F1, and successively the sample is compressed and
densified. During this process, the compressive flow stress σ increases with a
a b
route A
route C
turned180°
2mm 2mm
aligned micro-pores pores opened again
200µm
200µm 200µm
200µm
Fig. 9.16 Cross sections of lotus copper specimens after extrusion approximately 30 mm in a
second pass through an ECAE die with a channel angle of 90 . (a) Route A where the direction of
the sample of the second pass is the same as that of the first pass. (b) Route C where the direction of
the sample of the second pass is 180 turned axially from the first pass (Reprinted with permission
40 50
Extruded length /mm
Onset of extrusion
40
30
Load/ kN
Load Extruded 30
20
length
20
10
10
0 0
0 10 20 30 40 50
Punch displacement/ mm
Fig. 9.17 Change of the punch load and the extrusion length during the first pass through an
ECAE die with a channel angle of 90 . The onset of the extrusion was the place where the increase
of the extruded length began (Reprinted with permission from [13] © 2008 Elsevier B.V.)
σ
F1 die
F2
σ
F2’
F3 τ
Fshear
/2 τ
sample rod
Fig. 9.18 Mechanical balance model for ECAE process (Reprinted with permission from [13]
decrease of the porosity. It is considered that the force F2 caused by the compressive
flow stress σ is balanced with the force F3 caused by the shear flow stress τ at 20 mm
of the punch displacement.
At the beginning of pressing of lotus copper, the load increases slower than that
of nonporous metals. At the punch displacement of about 30 mm, the punch load
achieved a plateau value. From these results, the criteria for the onset of the
extrusion are expressed by the balance of the forces:
F 1 ¼ F2 > F3 :
Thus,
F1 ¼ Aσ > Aτ cotðΦ=2Þ; (9.2)
where A is the area of the cross section perpendicular to the axis of the entrance
channel. Thus, the criteria can be expressed as

σ Φ
> cot : (9.3)
τ 2
Or Φ ¼ 90 , the criteria is expressed as σ=τ > 1. Although the shear stress τ of
lotus copper has not been measured exactly, the expression indicates the tendency
that the criteria can be satisfied easily when the channel angle Φ is large.
2 pass
1 pass
Route A Route C
1mm
Fig. 9.19 Cross sections of lotus copper specimens after extrusion in the first and the second
pass through an ECAE die with a channel angle of 150 (Reprinted with permission from [13]
If the channel angle of 150 is chosen, an extrusion without densification can be

expected. Figure 9.19 shows the cross sections parallel to the sample axis of the
extruded sample obtained via 1 pass, route A (2 passes) and route C (2 passes).
These photos show that the pore morphology is not changed after 1 pass, and even
after 2 passes. Thus, it is concluded that it is possible to extrude without densifica-
tion under certain suitable conditions.
Figure 9.20a shows the change of the hardness in the case of the channel angle
of 90 . During the first pass, the hardness increases in the entry channel because of
the densification caused by uniaxial compression. Then, the shear deformation at
the corner increases the value of HV. Furthermore, the value of HV increases
through both the entry and exit channel of the second pass. There is not so
significant difference between route A and C of the second pass. Figure 9.20b
shows the HV in the case of the channel angle of 150 . In the entry channel of the
first pass, the value of HV is not so different from that before the extrusion. Then in
the end channel, HV increases. During the second pass, HV also does not increase
in the entry channel and then increases in the exit channel. Compared with the
sample extruded through the die with the channel angle of 90 , the increment of the
HV is not so large. It is apparent that ECAE process can give work hardening to the
sample in the case of channel angle of 150 maintaining the pores. The HV
increases with increase of the pass number. Therefore, it is also expected that
harder metal with pores after multi-pass ECAE can be obtained through the die
with the channel angle of 150 .
References 229
Fig. 9.20 Change of

Vickers hardness during the
first pass through an ECAE.
(a) Channel angle of 90
and (b) channel angle of
150 (Reprinted with
Elsevier B.V.)
References
1. Pogibenko AG, Konkevich VY, Arkuzova LA, Ryazantsev VI (2001) Weld Int 156:312–316
2. Bollinghaus T, Bleck W (2001) Cellular metals and metal foaming technology. MIT, Bremen,
pp 495–500
3. Haferkamp H, Ostendorf A, Goede M, Bunte J (2001) Cellular metals and metal foaming
technology. MIT, Bremen, pp 479–484
4. Murakami T, Nakata K, Ikeda T, Nakajima H, Ushio M (2003) Mater Sci Eng A357:134–140
5. Yamaoka H, Yuki M, Tsuchiya K (2000) Q J Jpn Weld Soc 18:422–430
6. Seto N, Katayama S, Matsunawa A (2000) Q J Jpn Weld Soc 18:243–255
7. Brandes EA (ed) (1983) Metal reference book, 6th edn. Butterworths, London
8. Watanabe H, Susa M, Nagata K (1997) Metall Mater Trans A 28:2507–2513
9. Murakami T, Tsumura T, Ikeda T, Nakajima H, Nakata K (2007) Mater Sci Eng A
456:278–285
10. Ikeda T, Nakajima H (2002) J Jpn Foundry Eng Soc 74:812–816
11. Ikeda T, Hoshiyama H, Nakajima H (2004) J Jpn Light Metals 54:388–393
12. Asada H, Yoshinaga H (1959) J Jpn Inst Metals 23:67–71
13. Suzuki S, Utsunomiya H, Nakajima H (2008) Mater Sci Eng A 490:465–470
14. Segal VM, Rezinikov VI, Drobyshevkiy AE, Kopylov VI (1981) Russ Metall 1:99–105
15. Valiev RZ, Langdon TG (2006) Prog Mater Sci 51:881–981
16. Hyun SK, Nakajima H (2003) Mater Sci Eng A 340:258–264
Chapter 10
Various Applications of Lotus Metals
Abstract In this chapter, four examples of application of lotus metals are

described. For heat sinks, penetrable pores are utilized to cool electronic devices.
For artificial teeth, open pores are effective to let bone tissue intrude for fixation of
the teeth into alveolar bones. For use of vibration–damping materials and golf
putter, closed pores serve to damping effect. Thus, anisotropic pores in lotus metals
contribute to such functions of products.
Keywords Biocompatibility • Golf putter • Heat sink • Implant

• Vibration–damping
10.1 Nomenclature
Ab: baseplate area of heat sink [m2]

Aw : cross-sectional area [m2]
D e: hydraulic diameter [m]
dpmean: average diameter of pores [m]
fg: fin gap of groove fins [m]
ft : thickness of groove fins [m]
hb: heat transfer coefficient based on Ab [W/(m2∙K)]
hdp: heat transfer coefficient on surface of pores [W/(m2∙K)]
H f: fin height [m]
hfin: heat transfer coefficient on surface of groove fins [W/(m2∙K)]
hm : average heat transfer coefficient [W/(m2∙K)]
k: thermal conductivity [W/(mK)]
L: thickness of lotus type porous copper fins [m]
Lg: length of groove fins [m]
Nt: total number of pores
p: porosity
Pr: Prandtl number of fluid

232 10 Various Applications of Lotus Metals
Q: heat flow [W]

Redp: Reynolds number defined by dp
ReDe : Reynolds number defined by De
Rfi : thermal resistance of heat sink [K/W]
Ra : thermal resistance of fluid [K/W]
Sf: total surface of pores [m2]
Tb: temperature of baseplate [K]
Ti: inlet temperature of fluid [K]
To: outlet temperature of fluid [K]
Ut: total fluid rate [m3/s]
udpmean: velocity of fluid through pore[m/s]
W: width of heat sink [m]
△P: pressure drop [Pa]
η: fin efficiency
a: air
f: cooling water
10.2 Heat Sink
In recent years, heat dissipation rates in power devices and laser diodes have been
increasing to more than 100 W/cm2 and in high-frequency electronic device, it is
increasing to more than 1,000 W/cm2 under the trend of miniaturization and
growing capacity. Figure 10.1 shows typical cooling loads and cooling technologies
compiled by Nishio [1]. When the heat flux increases, the temperature increases,
and a heat flux of 100 W/cm2 is as high as an equivalent heat flux from nuclear blast
at 2,000 K. However, a power device or a laser diode should be kept at normal
temperature level in spite of a large-heat flux of more than 100 W/cm2. Therefore,
novel heat sinks with high-heat transfer performance are required to cool these
devices. Among various types of heat sinks, heat sinks utilizing microchannels
with channel diameters of several tens of microns are expected to have excellent
cooling performance because a higher heat transfer capacity is obtained with
smaller channel diameters. Wei and Joshi investigated three-dimensional stacked
microchannel heat sinks to enhance the cooling performance of heat sinks with
microchannels [2]. Porous material with open pores is preferable for three-
dimensional microchannels because of the higher surface area per unit volume
and lower product cost. Bastawros and Evans investigated cellular metals as a heat
transfer medium [3]. Various porous materials such as sintered porous metals,
cellular metals, and fibrous composites have been investigated for heat sink
applications [4, 5]. However, heat sinks using porous materials were clarified to
have a high-pressure drop because the cooling fluid flow through the pores of the
porous materials is complex [6].
Among the described porous materials, a lotus metal with straight pores is
preferable for heat sinks due to the small pressure drop of the cooling fluid flowing
10.2 Heat Sink 233
Fig. 10.1 Typical cooling loads and cooling technology [1]
through the pores. The following are the main features of lotus metals: (1) the pores
are straight, (2) the pore size and porosity are controllable, and (3) porous metals
can be produced with pores as small as a hundred microns in diameter. To
effectively use lotus copper as a heat sink, its effective thermal conductivity must
be known, by which the effect of directional pores on the heat flow must be
considered. One defined perpendicular effective thermal conductivity keff⊥ of
lotus copper as the thermal conductivity for heat flow perpendicular to the pore
axis is shown in Fig. 10.2.
To predict the heat transfer capacity of a lotus copper heat sink, the correlation
between heat conduction in lotus metal and heat transfer to the fluid in the pores
must be considered. This correlation is generally expressed by fin efficiency. As a
result of numerical analysis, the fin efficiency of the lotus copper fin was verified to
be predictable by a simple straight fin model using effective thermal conductivity
keff⊥ and the surface area ratio between the surface area of lotus copper fin and that
of the straight fin [7]. Chiba et al. investigated the model for predicting the heat
transfer capacity of heat sink using the lotus copper. This model takes into consid-
eration the heat conduction in lotus copper and the heat transfer to the fluid in the
pores. Finally, cooling performance of lotus copper heat sink designed by using the
model for air cooling and water cooling is investigated [8].
Fig. 10.2 Lotus copper with anisotropic pores: (a) outer view of lotus copper and (b) definition of
thermal conductivity (Reprinted with permission from [8] © 2011 American Society of Mechani-
cal Engineering)
10.2.1 Heat Sink for Air Cooling
Generally, a groove fin is used for cooling the power devices for air cooling or water
cooling. The groove fin composes of a lot of boards connected vertically on the
base. The heat transfer capacities of two types of heat sinks: (1) conventional
groove fins and (2) lotus copper heat sinks, were examined. The configuration
and specifications of the conventional grove fins are shown in Fig. 10.3. The
conventional groove fins made of copper have a 3-mm fin gap size fg, a 1-mm fin
thickness ft, and a 20-mm fin height Hf. The heat transfer capacity of the conven-
tional groove fins is only derived from calculations because many experimental
formulas of the fins have been established so far [9, 10]. Figure 10.4 shows the
configuration of the heat sink using three lotus copper fins. W is a width. Lg is a
length. Because the heat sink has 3 lotus copper fins, base area Ab of the heat sink is
expressed by 3WL.
Figure 10.5 illustrates the experimental apparatus for measuring the heat transfer
capacity of heat sinks. Cooling air was blown by a blower into the test duct in which
the heat sink is located. The heat sink consists of fins brazed on one side of a copper
10.2 Heat Sink 235
Fig. 10.3 Configuration of conventional groove fins (Reprinted with permission from [8] © 2011
American Society of Mechanical Engineering)
Fig. 10.4 Configuration of lotus copper heat sink (Reprinted with permission from [8] © 2011
Fig. 10.5 Experimental apparatus for measuring heat transfer capacity (Reprinted with permis-
sion from [8] © 2011 American Society of Mechanical Engineering)
baseplate and the heating block with a heater soldered to the other side of the
baseplate. The inlet temperature of cooling air Ti, the temperature of copper
baseplates Tb1, Tb2, and Tb3, and the outlet temperature of the cooling air T0 are
measured by K-type thermocouples.
The heat transfer capacity by heat transfer coefficient hb based on baseplate area
Ab of the heat sink is as follows:
Q
hb ¼ ; (10.1)
Ab ðTb Ti Þ
Q ¼ ρ Cp Ut ðT0 Ti Þ; (10.2)
where Q is the heat input to the heat sink, ρ and Cp denote the density and the
specific heat of the fluid, respectively. Ut is the total flow rate of the air through the
test duct. Tb( ¼(Tb1 + Tb2 + Tb3)/3) is the mean temperature of the copper base-
plate. The pressure drop between the inlet side and the outlet side of the heat sinks
was measured within an experimental accuracy of 5 % by a pressure sensor.
Heat transfer capacity of heat sink was evaluated by using the heat transfer
coefficient hb based on baseplate surface area Ab. hb is expressed by Eq. (10.3):
1
hb ¼ ; (10.3)
Rfi Ab
where Rfi is the thermal resistance of the heat sink between the temperature of the
baseplate and the fluid’s inlet temperature. Rfi is expressed by
Tb Ti Ra
Rfi ¼ ¼ ; (10.4)
Q 1 exp RRaf
where Ra is the thermal resistance between the inlet and outlet of the fluid and Rf is
the thermal resistance based on the logarithmic mean temperature difference
between the base of the heat sink and the fluid. These thermal resistances are
expressed by
ΔTm 1
Rf ¼ ¼ ; (10.5)
Q hdpmean Sf η
T0 Ti
ΔTm ¼ ; (10.6)
ln TTbbT
Ti
0
T 0 Ti 1
Ra ¼ ¼ ; (10.7)
Q ρ Cp Ut
10.2 Heat Sink 237
where hdpmean is the heat transfer coefficient on surface of pores Sf in the lotus copper
fin. As the flow characteristics through the pores in the lotus copper fins is
considered to be similar in a circle pipe, heat transfer coefficient hdpmean is expressed
by the following correlations for the circle pipe under laminar flow regime
(Rep < 3,000) [9]:
hdpmean dpmean
10=9 3=10
Nup ¼ ¼ 5:364 1 þ f ð220=π Þ Xþ g
ka
8 0 15=3 93=10
>
< þ
>
=
B π =ð115:2X Þ C
1 þ @h i 1=2 h i 3=5 A 1;
>
: >
;
1 þ ðPra =0:0207Þ2=3 1 þ f ð220=π Þ Xþ g 10=9
(10.8)
L=dpmean
Xþ ¼ ; (10.9)
Rep Pra
udpmean dpmean
Rep ¼ ; (10.10)
υa
Ut
udpmean ¼ ; (10.11)
p Aw
where Nup and Rep are Nusselt number and Reynolds number which are defined by
average diameter for pore dpmean, and νa and ka are dynamic viscosity and thermal
conductivity of the air, respectively. udpmean is the velocity of air through the pore. η
is the fin efficiency of the lotus copper fin, which is calculated from the model by
which lotus copper fin was assumed to be a straight fin model as follows:
tanhðm Hf Þ
η¼ ; (10.12)
m Hf
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
hdpmean ξ 2ðW þ LÞ
m¼ : (10.13)
keff? ðW LÞ
Pressure drop of lotus copper fin, ΔP, is expressed by the following correlation
for the circle pipe under laminar flow regime:

64 L 1
ΔPL ¼ þ ζ ρ udpmean 2 ; (10.14)
Rep dpmean 2
where ζ is the total of the pressure loss coefficient generated when the fluid flows
into and out of the pores, where ζ ¼ 1.4 was assumed.
Because the conventional groove fins have often been researched so far, empiri-
cal formula for predicting their heat transfer capacity and pressure drop are
established [10]. Based on the baseplate surface area, the heat transfer coefficient
hb is predicted by Eq. (10.15)
hf A f η
hb ¼ ; (10.15)
Ab
where hf is the heat transfer coefficient on fin surface Af of the groove fins and is
predicted by Eq. (10.16)
Nu1 Nu2
Nufg ¼ ðPra 0:71Þ þ Nu2 ; (10.16)
10 Pra
Nu1 ¼ 2:80136 2:10514X þ 0:411783X2 þ 4:11; (10.17)
Nu2 ¼ 4:1880 3:14709X þ 0:611075X2 þ 4:11; (10.18)

L
X ¼ ln Refg Pra ; (10.19)
fg
where Refg and Nufg are the Reynolds number and the Nusselt number defined by the
fin gap. u is the air velocity through the fin gap.
Pressure drop ΔP of the conventional groove fins under a laminar flow regime is
expressed as follows:
Lg
X¼ ; (10.20)
ReDe De

4Lg 1
ΔP ¼ f þ ζ ρ u2 ; (10.21)
De 2
3:44 24 þ 0:674
4X X0:5
3:44
f ¼ þ ; (10.22)
X0:5 ReDe ð1 þ 0:000029 X2 Þ ReDe
where ReDe is the Reynolds number defined by hydraulic diameter De (¼ 4Hf fg/
(2Hf + 2fg)).
Chiba et al. measured the heat transfer coefficients and pressure drop using
experimental apparatus of air-cooling heat sink [11]. The heat transfer coefficient
hb, based on the baseplate surface area defined by Eq. (10.3) for the experiment and
Eq. (10.1) for the calculated results, respectively, is plotted as a function of the inlet
velocity u0 (¼Ut/Aw) in Fig. 10.6. The prediction for the lotus copper heat sink
showed a good agreement with the experimental data within accuracy of 5 %.
10.2 Heat Sink 239
Fig. 10.6 Comparison of heat transfer coefficient hb between experimental and predicted data
(Reprinted with permission from [8] © 2008 Elsevier B.V.) (Reprinted with permission from [8] ©
2011 American Society of Mechanical Engineering)
The experimental data for lotus copper heat sink with thickness L ¼ 1 mm showed
a very large-heat transfer coefficient of 5,000 W/(m2 K) under an inlet velocity of
1.0 m/s, which is 13.2 times higher than that for the conventional groove fins.
The pressure drop ΔP in lotus copper heat sinks is plotted as a function of the
inlet velocity to heat sink u0 in Fig. 10.7. The predicted pressure drop of lotus heat
sink, denoted by 3ΔPL, shows a good agreement with the experimental data within
accuracy of 10 %. Comparing the pressure drop among all of heat sinks at the
inlet velocity u0 of 1.0 m/s, the experimental data show that the pressure drop of
lotus copper heat sink with thickness L ¼ 3 mm is 13.5 times higher than that of the
conventional groove fins.
The heat transfer coefficient hb is shown as a function of the pumping power
ΔPUt that is defined by the product of the pressure drop ΔP and the total flow rate
Ut (Fig. 10.8). A comparison of the heat transfer coefficient among all lotus copper
heat sinks under a pumping power of 0.02 W revealed that the experimental data for
lotus copper heat sink with thickness of 1 mm are 11.3 times higher than those for
the conventional groove fin.
In summary, the thinner the lotus copper fin is, the higher is the heat transfer
capacity of a lotus copper heat sink under identical pressure drop and identical
pumping power. It is preferable that the thickness of the lotus copper is small when
a lotus copper fin is used for heat sink.
Fig. 10.7 Comparison of pressure drop ΔP between experimental and predicted data (Reprinted
with permission from [8] © 2011 American Society of Mechanical Engineering)
Fig. 10.8 Comparison of heat transfer coefficient hb as a function of pumping power ΔPUt
(Reprinted with permission from [8] © 2011 American Society of Mechanical Engineering)
10.2 Heat Sink 241
Fig. 10.9 Configuration groove fins and microchannel copper heat sink (Reprinted with permis-
sion from [8] © 2011 American Society of Mechanical Engineering)
Fig. 10.10 Lotus copper heat sink for water cooling (Reprinted with permission from [8] © 2011
10.2.2 Heat Sink for Water Cooling
The heat transfer capacity of three types of heat sinks for water cooling with
conventional groove fins, groove fins with smaller fin gap (microchannel), and
lotus copper were investigated. The configuration and the specifications of the
conventional groove fins and microchannels are shown in Fig. 10.9. The conven-
tional groove fins have fin gap of 3 mm and fin thickness of 1 mm. The
microchannels have fin gap of 0.5 mm and fin thickness of 0.5 mm. The heat
transfer capacity of the conventional groove fins is only for calculation. On the
other hand, the configuration of lotus copper heat sink for water cooling is shown in
Fig. 10.10. The heat sink has three pieces of lotus copper fins with the thickness of
3 mm along the flow direction. The lotus copper fins have pores with the mean
diameter of 0.3 mm and the porosity of 39 %. Experimental apparatus for measur-
ing the heat transfer capacity of heat sink is shown in Fig. 10.11. A cooling water
was circulated through a filter and the test duct in which the heat sink is located.
The circulator has a pump and water-cooling equipment. The heat sink consists of
fins that are brazed on one side of a copper baseplate. The heat sink block with a
heater was soldered on the other side of the baseplate. Inlet temperature of cooling
Fig. 10.11 Experimental apparatus for measuring heat transfer capacity for water cooling
water Ti, a temperature of copper baseplate Tb’, and outlet temperature of

cooling water T0 were measured by K-type thermocouples. From the measurement
result, heat transfer coefficient was calculated by Eq. (10.1). The pressure drop
between the inlet and the outlet of the fins of every heat sink was measured.
The heat transfer coefficients based on baseplate surface area Ab defined by
Eq. (10.1) for experiment and Eq. (10.3) for prediction, respectively, are plotted for
all of heat sinks as a function of the inlet velocity to the heat sinks u0 in Fig. 10.12.
The prediction for lotus copper heat sink showed good agreement with the experi-
mental data within an accuracy of 15 %. The experimental data of lotus copper
heat sink showed very large-heat transfer coefficient of 80,000 W/(m2 K) under the
velocity u0 of 0.2 m/s that is 1.7 times higher than that of the microchannels and 6.5
times higher than that of the conventional groove fins.
The value of pressure drop of all the heat sinks is compared in Fig. 10.13 as a
function of u0. The predicted pressure drop of lotus copper heat sink showed good
agreement with the experimental data within an accuracy of 5 %. Comparing the
pressure drop among all of the heat sinks under the velocity u0 of 0.2 m/s, the
experimental data of the lotus copper heat sink is 2.5 times higher than that of
the microchannels and 38 times higher than that of the conventional groove fins.
The heat transfer coefficient hb is shown as a function of pumping power that is
defined by a product of a flow rate Ut and ΔP is shown in Fig. 10.14. Comparing the
heat transfer coefficients among all of the heat sinks under the pumping power of
0.01 W, the experimental data of lotus copper heat sink is 1.3 times higher than that
of the microchannels and 4 times higher than that of the conventional groove fins.
Thus, the use of lotus metals is the most superior for the heat sinks of air cooling and
water cooling.
10.2 Heat Sink 243
Fig. 10.12 Comparison of heat transfer coefficient hb between experimental and predicted data
Fig. 10.13 Comparison of pressure drop ΔP between experimental and predicted data (Reprinted
with permission from [8] © 2011 American Society of Mechanical Engineering)
Fig. 10.14 Comparison of heat transfer coefficient hb as a function of pumping power ΔPUt
10.3 Vibration–Damping Materials
Nowadays, the demand for high-damping materials is increasing in various fields

such as aerospace, transformation, and manufacturing industries. Also in an envi-
ronmental problem, high-damping materials play an important role on the reduction
of noise. Therefore, various damping materials have been produced and developed
so far [12]. It is known that porous materials such as cellular materials and foamed
metals show a high-damping capacity [13]. The combination of the two
characteristics of lightness in weight and high-damping capacity makes porous
metals more attractive. However, the mechanical strength of conventional porous
materials is quite low due to their complicated porous structure, and this weak
mechanical strength limits their applications. The strength of lotus metals is
superior to that of conventional porous materials. If lotus metals show a high-
damping capacity as well as the conventional porous materials, lotus metal can be a
new high-damping material combining two characteristics of lightness in weight
and high strength. Xie et al. investigated the damping capacity of lotus magnesium,
which was measured by hammering–vibration–damping tests [14].
Figure 10.15 shows the schematic drawing of the apparatus for the hammering–
vibration–damping method. The disc-shaped specimen is hanged with two strings,
and the stainless steel sphere of 10 mm in diameter and 4 g in weight is also hanged
with a string. They lift the sphere up to 20 mm in height and release it. The swung
10.3 Vibration–Damping Materials 245
Amplifier
Amplifier
Support body
String AD converter
Stainless
steel ball
Amplitude
Amplitude
time
time
Microphone
Specimen
85mm
(thickness: 2mm) Personal computer
Fig. 10.15 Schematic illustration showing the measurement system for the damping capacity
(attenuation coefficient) by hammering–vibration–damping tests (Reprinted with permission
sphere hits the center of the disc-shaped specimen and excites the free vibration.
The acoustic microphone detected the amplitude of the specimen vibration through
the acoustic waves. The AD converter transforms the analog signals obtained with
the microphone into the digital signals, where the sampling frequency of the AD
converter is 42 kHz. Figure 10.16 shows the vibration–damping curves of the
nonporous and lotus magnesium with the porosity of 44.4 % after excitation of
the vibration. While for nonporous magnesium, the amplitude of a free vibration
gradually decreases with increase in time, the amplitude decreases steeply for lotus
magnesium. Figure 10.17 shows the Fourier transforms of the vibration–damping
curves of nonporous and lotus magnesium. The Fourier transforms of nonporous
magnesium possess one large peak and two small peaks, which indicate that a
hammering mainly excites the fundamental mode of resonance vibrations. On the
other hand, many peaks observed for lotus magnesium indicate that a hammering
excites various resonant-vibration modes.
Here, one considers a material vibration at a resonant frequency by cyclic
external force. When the external force is removed, the resonant-vibration damps
gradually. Then, the amplitude of a damping vibration, A(t), can be expressed as
AðtÞ ¼ A0 expðαt þ i2πfr tÞ; (10.23)
where t denotes the time after removal of an external force, A0 denotes the
amplitude at t ¼ 0, and fr denotes a resonant frequency. α denotes the attenuation
(damping) coefficient, which depends on the damping capacity of materials.
Fig. 10.16 Vibration–damping curves of (a) nonporous magnesium (porosity 0 %) and

(b) lotus magnesium with the porosity of 44.4 % (Reprinted with permission from [14] © 2005
Elsevier B.V.)
For lotus magnesium, various vibration modes are excited, and therefore they
cannot obtain the attenuation coefficient by fitting Eq. (10.23) to the
vibration–damping curve. To determine α in such a case, they picked up the
maximum value of the positive amplitude of the damping curve and fitted the
equation
Fig. 10.17 The Fourier transforms of the vibration–damping curves of (a) nonporous magnesium
(porosity 0 %) and (b) lotus magnesium with the porosity of 44.4 % (Reprinted with permission
AðtÞ ¼ A0 expðαtÞ þ C (10.24)
to the amplitude and obtained the apparent attenuation coefficient of lotus

magnesium, where C denotes a fitting coefficient. Figure 10.18 shows the
vibration–damping curve of lotus magnesium of 44.4 % in porosity and the curves
obtained by fitting Eq. (10.24) to the vibration–damping curves. Figure 10.19
exhibits the porosity dependence of the apparent attenuation coefficient of lotus
magnesium. The apparent attenuation coefficient increases with increase in porosity.
When a resonant vibration of a material at a resonant frequency damps, the
attenuation coefficient depends on the resonant frequency and intrinsic internal
friction, Q1:
Fig. 10.18 The vibration–damping curves of (a) nonporous magnesium and (b) lotus magnesium
with the porosity of 44.4 %. The curves were obtained by fitting Eq. (10.24) to the
vibration–damping curves (Reprinted with permission from [14] © 2005 Elsevier B.V.)
α ¼ πfr Q1 : (10.25)
The measurement of the internal friction of lotus copper [15] revealed that the
internal friction hardly depends on porosity at room temperature. Therefore,
one can assume that Q1 of lotus magnesium is independent of its porosity and
Fig. 10.19 Porosity dependence of the attenuation coefficients of lotus magnesium, which was
measured by hammering–vibration–damping tests (Reprinted with permission from [14] © 2005
Elsevier B.V.)
resonant-vibration mode, i.e., the attenuation coefficient depends only on the

resonant frequency. The resonance frequency depends on the elastic constants,
density, and dimension of the specimen. The presence of the pores decreases the
frequency of the resonant-vibration modes, and therefore, the resonant frequency of
the fundamental mode of the lotus magnesium of 44.4 % in porosity is lower than
that of nonporous magnesium. If a hammering excites only fundamental mode, the
attenuation coefficient of lotus magnesium is smaller than that of nonporous mag-
nesium. However, in fact, various vibration modes are excited for lotus magnesium,
which affects the apparent attenuation coefficient. In order to discuss the effect of
such a mode excitation, they picked up the vibration–damping curves consisting of
specific frequency using Fourier transforms technique. Figure 10.20 shows the
vibration–damping curve of a specific vibration mode, where the curves containing
the overall frequencies are shown for comparison. The vibration of nonporous
magnesium mainly consists of one resonant-vibration mode, and the apparent
attenuation coefficient mainly depends on the resonant frequency of 3.3 kHz. On
the other hand, the vibration of the lotus magnesium of 44.4 % in porosity mainly
consists of various resonance modes, and therefore, the amplitude of the overall
vibration disperses to the various resonance modes. At t ¼ 0, the amplitude of
resonance mode of 10.2 kHz is larger than that of 1.6 kHz. However, the resonant
vibration of 10.2 kHz damps faster than that of 1.6 kHz. For lotus magnesium, the
frequencies of most of excited vibrations are higher than that of fundamental mode
of nonporous magnesium, and the resonance modes of high frequencies are domi-
nant in the apparent attenuation coefficient. Therefore, the attenuation coefficients
of lotus magnesium are larger than that of nonporous magnesium.
Fig. 10.20 Vibration–damping curves of (a) nonporous magnesium and (b) lotus magnesium with
the porosity of 44.4 %. (a) Black line denotes the vibration–damping curve consisting of the overall
frequencies. Dark-gray and light-gray lines denote the damping curves of the resonant modes of 3.3
and 14.1 kHz, respectively. (b) Black line denotes the vibration–damping curve that consists of the
overall frequencies. Dark-gray and light-gray lines denote the damping curves of the resonant modes
of 10.2 and 1.6 kHz, respectively (Reprinted with permission from [14] © 2005 Elsevier B.V.)
In order to compare the vibration–damping capacity of lotus magnesium with

those of various metals and alloys, Xie et al. measured the attenuation coefficients
of various metals and alloys using the hammering–vibration–damping method.
10.4 Golf Putter 251
120
Porosity
44.4%
Attenuation coefficient (%) 100
Porosity
80 40.4%
60
Porosity
40 28.1%
20
0
Sta
Co
No
Alu
Po
Po
Po
rou
rou
rou
pp
np
mi
inle
er
oro
niu
sM
sM
sM
ss
us
g
ste
Mg
el
Fig. 10.21 Attenuation coefficients of nonporous aluminum, copper, stainless steel magnesium,
and lotus magnesium with different porosity (Reprinted with permission from [16] © 2006
Elsevier Ltd)
Figure 10.21 shows the attenuation coefficients of nonporous aluminum, copper,

and stainless steel. The attenuation coefficients of lotus magnesium are larger than
those of nonporous aluminum, copper, and stainless steel which indicates that lotus
magnesium is newly considered as desirable candidate of high-damping material.
10.4 Golf Putter
Usually, new materials have been first adopted to produce sporting goods, weapons,
space materials, etc., whose costs can be almost disregarded. Since the lotus metals
are rather new materials, commercialization of lotus metals to sporting goods is
reasonable. In 2002, lotus copper was for the first time commercialized as golf
putter manufactured by Ryobi Corporation in Japan.
Nonporous copper, lotus copper, and resin were put into the putter flame as
inserted materials. The vibration–damping capacity of the inserted materials such
as nonporous copper, lotus copper, and resin was investigated [16]. Three
damping curves were measured by hammering–vibration–damping method.
Figure 10.22 shows the configuration of putter head, measurement apparatus of
damping, the damping results, and the attenuation coefficients. The results
Fig. 10.22 (a) Configuration of putter head, (b) damping measurement apparatus, (c) the damping
property; the vertical axis shows the amplitude and the transversal axis shows the time (upper
left, nonporous copper; upper right, lotus copper; lower left, lotus copper inserted by the rubber;
lower right, resin)
exhibited that lotus copper had superior damping capacity, and thus, the company
decided to commercialize the putter with lotus copper. It is said that feeling of a
golf putter hitting a ball is mild. Figure 10.23 shows the golf putter whose central
10.5 Medical Devices 253
Fig. 10.23 Golf putter whose central part is made of lotus copper. This putter exhibits superior
vibration–damping capacity. Players feel soft to hit a ball (Reprinted with permission from [16]
© 2006 Elsevier Ltd)
part is made of lotus copper: the porosity is 40 % and the pore size is about 100 μm
in diameter.
10.5 Medical Devices
Porous metallic biomaterials have been developed for the reconstruction of hard
tissues since those provide obvious advantages over typical monolithic implants.
One of these is the possibility to match the mechanical properties to those of bone
to prevent stress shielding and loosening of the implant. Metals and alloys
processed to create a porous structure by means of powder metallurgy processes
[17–19], solid free-form fabrication [20], or vapor deposition of the metal,
i.e., tantalum on vitreous polymer network, have been shown to enhance the
bone ingrowth and osseointegration [21–24]. In addition, a porous surface
improves the mechanical interlocking between the host bone and the implant
biomaterial [25], reinforcing the stability of the implant by biological fixation.
Thus, application of the porous metallic biomaterials to medical devices has been
increasingly expected.
10.5.1 In Vitro Cytocompatibility of Lotus Ni-Free

Stainless Steel
Three kinds of stainless steels containing different contents of chromium were

employed in the investigation of in vitro cytocompatibility by Alvarez et al. [26]:
Fe-25Cr-1N, Fe-23Cr-2Mo-1N, and AISI 446-1N [composition (wt.%)]. Beside
AISI 316L stainless steel was also tested for comparison. Lotus stainless steels were
fabricated using continuous zone melting technique. In order to alloy the samples
with nitrogen, a high temperature solution nitriding treatment was applied. The high
temperature solution nitriding treatment was carried out by sealing the plate
samples into a quartz tube of 20 mmϕ 170 mm in a nitrogen atmosphere of
0.07 MPa. The encapsulated samples were heat-treated at 1,100 C for 604.8 ks and
then quenched into a water bath. After this heat treatment the stainless steel samples
of originally ferritic structure become austenitic.
Cytoxicity of the solution nitride samples was evaluated using MC3T3-E1
murine calvaria osteoblasts-like cells. The cell line was acquired from an
authorized cell and gene bank (Riken Cell Bank, Ibaraki, Japan). The cells
were grown and maintained in α minimum essential medium (α-MEM)
supplemented with 1 % kanamycin and 10 vol.% fetal bovine serum in a
humidified incubator with 95 % air and 5 % CO2 at 37 C. Passaging and
preparation of single ell suspensions for seeding on the metallic samples was
achieved by enzymatic digestion using 0.25 vol.% trypsin, 0.02 % (w/v) EDTA
solution in PBS at pH 7.4. Cells counts were performed using a Tatai-type
hemacytometer under an inverted microscope. In vitro cytotoxicity was assessed
by direct contact method. Disc-shaped specimens with approximate thickness
of 1 mm and diameter of 10 mm were obtained from the nonporous alloys
rods using electric discharge machining. Prior to cytotoxicity testing, each
specimen was ground with 2000 grit SiC paper and ultrasonically cleaned with
alcohol, and then rinsed in distilled water. Subsequently, the discs were
autoclaved at 121 C for 0.9 ks.
Cells were incubated at 37 C in a humidified 5 % CO2 atmosphere for 3- and
7-day intervals without renewal of the culture medium. The cells were examined
under optical microscope and photographed with a Nikon DS-L2 color digital
camera. The total cell number was analyzed after staining using digital image
analysis. The measurements of the cell spreading area were analyzed at the different
culture intervals and normalized to the maximum area available in the metallic
surfaces. Cell proliferation and the assessment of extent of spreading and cell
morphology assays were repeated four times. After 3 days of proliferation assays,
media from the cells cultured with metallic materials were saved in quadruplicate
for inductively coupled plasma (ICP) analysis.
Cytotoxicity testing of surgical implant materials is an important way to verify
their biocompatibility. The quantitative results of the proliferation test obtained
with the image analysis software are presented in Fig. 10.24 [26].
The cells cultured for 7 days multiplied actively on the four stainless steel
surfaces compared to those cultured for 3 days. After each culture period, cells
grown on stainless steel surfaces displayed a significantly lower proliferation rate
than the cells cultured on type IV collagen-coated cell culture dish. One-way
ANOVA showed no statistically significant difference for the cell numbers of
four stainless steel samples after 3 and 7 days of culturing. In the case of pure Ni
samples after 7 days in culture, the cell numbers were significantly smaller than that
of four stainless steel samples and negative controls. While in the stainless steel
Fig. 10.24 Proliferation rates after 3 and 7 days in direct contact with the evaluated metallic
materials with respect to the control cultures grown on type IV collagen-coated cell culture dish
(100 %). (Reprinted with permission from [26] © 2008 Springer Science + Business Media, LLC)
surface cells proliferated forming multilayers after 7 days of culture, in the pure Ni
sample, the formation of randomly located multilayer cell nodules was observed,
indicating that the spreading was scarce and not uniform.
Morphology of adherent cells after 3-day incubation period is shown in
Fig. 10.25. Microscopic analysis of the cell monolayers revealed that there were
no notable differences in cell morphology between the type IV collagen cell culture
dish (Fig. 10.25c) and the stainless steels (Fig. 10.25a–h). The body of osteoblasts
has a mean size of 15 μm and many of them developed armlike cytoplasmic
extensions with lengths up to 25 μm. MC3T3-E1 cells seemed to adhere tightly to
surface under high magnification (750magnification) (Fig. 10.25e–h) in accor-
dance with the flattened morphology surrounded by cytoplasmic prolongations.
MC3T3-E1 osteoblasts cells proliferated uniformly on the surface of Ni-free
stainless steels under optical microscope (100magnification) (Fig. 10.25a–d).
Microscopic analysis of the cell cultures on pure Ni samples revealed that the cell
morphology was altered. Cells grown on pure Ni exhibited reduced cytoplasmic
area and also abnormally large cells (giant cells) were observed in both cultivation
periods (Fig. 10.26).
After 3 days of culture, the MC3T3-E1 cellular spreading was good in all the
evaluated stainless steel surfaces but not as extensive as in the cell culture dish
treated with type IV collagen, and after 7 days of culture, osteoblasts were evenly
Cytotoxicity Assay
a e
C Fe-23Cr-2Mo-1N
b f
250 µm
Fe-25Cr-1N
c g
Cell line:
MC3T3-E1
(Osteoblasts)
AISI 446-1N
d h
AISI 316L
100 µm 25 µm
Fig. 10.25 Representative optical microscopy images of spread MC3T3-E1 osteoblasts cells after
3 days of direct cultivation on (C) type IV collagen-coated cell culture dish (a) and (e) Fe-23Cr-
2Mo-1N, (b) and (f) Fe-25Cr-1N, (c) and (g) AISI 446-1N, and (d) and (h) AISI 316L. MC3T3-E1
cells are approaching confluence and display normal morphology. Original magnification:
(C)15, (a–d) 100, (e–h) 750 (Reprinted with permission from [26] © 2008 Springer
Science + Business Media, LLC)
spread covering the entire available stainless steel surfaces. Statistically significant
difference between the average area of cells cultivated on the Ni-free stainless steel
samples and AISI 316L samples was not observed.
10.5.2 In Vivo Osteocompatibility of Lotus Ni-Free

Stainless Steel
A new generation of stainless steels, called nickel-free austenitic stainless steels,

does not contain any nickel or the amount of nickel is very reduced (<0.5 wt%)
[27], and those have been developed to be used in biological systems. The high
nitrogen content is the reason of their superior corrosion resistance and elevated
mechanical properties, and these advantages have been widely reported in the
literatures [28]. As for the mechanical properties are concerned, the stiffness of
100 µm 100%Ni
100 µm
100%Ni
Fig. 10.26 Optical microscopic images of MC3T3-E1 osteoblast cells grown on pure Ni samples.
A higher frequency of reduced cytoplasmic area cells and giant cells was observed (Reprinted with
permission from [26] © 2008 Springer Science + Business Media, LLC)
stainless steels is far higher than that of compact bone, and this can lead to stress
shielding to the host bone, which may result in detrimental resorptive bone
remodeling. Moreover, high elastic moduli of stainless steel implants can cause
delayed callus formation and bone healing. However, in order to solve this problem,
a number of fabrication processes to produce porous bulk metals and alloys,
including stainless steels, have been developed.
The introduction of pores into a structural material will drastically change its
mechanical properties of stiffness, strength, and fatigue resistance. Moreover, the
geometric microstructure of the porosity is able to change the physical properties
such as specific surface area, apparent density, and degree of anisotropy. Based on
clinical outcomes and histological evidence, the advantages of porous biomaterials
for bone–implant applications are multifold. Bone ingrowth is possible, which

could greatly improve the bone–implant interface and may allow for efficient soft
tissue attachment supplementing the stability of the implant by biological fixation.
A strong bone–implant interface reduces harmful tissue strain and micromotion and
therefore promotes the osseointegration of the implant, which is a basic requirement
for its successful functioning as host organism. Additionally, by increasing the
porosity of the implant material, its stiffness decreases correspondingly, reducing
the stress-shielding effect, which can lead to resorption of the bone and the
loosening of the implant.
For many clinical situations, a metallic porous graft offers structural integrity.
These porous structures must provide temporary support and bone ingrowth. The
greatest application fields are in the osseous reconstructive surgery, such as treat-
ment of small bone defects, cranio-maxillofacial reconstruction, tooth root
replacements, and vertebral body replacement. In applications in which a porous
structure bridges gaps in tissues, the success of bone ingrowth is affected by the
pore structure characteristics, such as pore density; pore orientation along length,
width, or radial direction; pore size; and pore spacing. New porous metallic
structures have to be developed to help advance the current generation of metallic
grafts and address unresolved clinical applications. The evaluation of the bone
apposition rates within differently oriented pores is also necessary in order to
accomplish an efficient implant design.
Alvarez et al. [29] produced a lotus nickel-free austenitic stainless steel with
longitudinally oriented pores by continuous zone-melting technique. In the former
section, lotus structure was tested in vivo in a canine model achieving excellent
bone ingrowth. This section describes further and detailed evaluation of lotus
nickel-free austenitic stainless steel as a biomaterial candidate for the treatment
of osseous defects and osseous reconstructive repairs in persons with strong nickel
allergy predisposition.
The lotus rods of AISI 316L and AISI 446 were fabricated using the continuous
zone-melting technique. AISI 316L is commonly used in biomedical applications
and was selected as a reference material. AISI 446 stainless steel is a ferritic-type
stainless steel that contains 0.3 wt% Ni as impurity. Sixty male, 8 weeks of age
Sprague–Dawley rats, received bilateral femoral and trial transcortical implants.
In each femur and tibia, one sterile implants of the same alloy was carefully placed
into the defect with the pores parallel or perpendicular to the long axis of the rat bone
using a sterile forceps (Fig. 10.27). Animals were placed in pairs in cages and were
given free access to food and water. The rats were sacrificed by ether suffocation.
The left and right tibia bones were retrieved at 4, 8, and 12 weeks after implantation
for histology analysis. The femurs also retrieved at 4, 8, and 12 weeks were reserved
for mechanical testing. Twenty rats were used for each implantation period.
Push-out testing was performed after 4, 8, and 12 weeks postoperatively by
applying a longitudinal compressive force to the cylinder implants. Schematic
diagram for testing shear bonding strength between bone tissues and the implant
is shown in Fig. 10.28. Push-out testing of the bone–implant samples was
a b
S.D. Rat
2.2 mm
2.2 mm
8.0 mm
Left tibia
8.0 mm
Right tibia
// ^
Fig. 10.27 Radiographs showing the implant placement configuration. (a) The elongated direc-
tion of the pores of the implant was parallel to the long axis of the rat bone. (b) The elongated
direction of the pores of the implant was perpendicular to the long axis of the rat bone
Fig. 10.28 Schematic diagram of the arrangement used for the mechanical push-out tests. F force
applied on implant, t thickness of the cortical bone surrounding the implant, ϕ implant diameter
performed by using an INSTRON 4482 universal testing machine. The crosshead

speed was set at 0.5 mm min1 and the cylinder implants were tested to failure.
Signs of almost complete repair were observed with both types of implants.
Histological evaluations revealed bone ingrowth even in the most central part of the
lotus implants. Photomicrographs of representative histologic sections after
12 weeks implantation are shown in Fig. 10.29. At 12 weeks the histological
Fig. 10.29 (a) Lotus-type porous AISI 316L implant inserted with the elongated direction of the
pores parallel (//) to the tibia bone long axis direction. (b) Lotus-type porous AISI 316L implant
inserted with the elongated direction of the pores perpendicular (?) to the tibia bone long axis
direction. (c) Lotus-type porous AISI 446-IN implant inserted with the elongated direction of the
pores parallel (//) to the tibia bone long axis direction. (d) Lotus-type porous AISI 446-IN implant
inserted with the elongated direction of the pores perpendicular (?) to the tibia bone long axis
direction. Light microscopy histological sections after 12 weeks of implantation stained with
Villanueva-Goldner’s-Trichrome. (Reprinted with permission from [29] © 2008 Elsevier B.V.)
sections displayed an increase of bone contact and bone ingrowth. Indeed, it was
found in many histological sections that the area of bone ingrowth within the
implant surface was almost complete at 12 weeks. At higher magnifications,
the fabricated implants demonstrated healthy cortical bone directly interfacing
with the metallic materials (Fig. 10.30). The concept of bone–implant contact
(osseointegration) is based on the light microscopic appearance of the biological
side of the interface. Osseointegration implies an interface with remodeled, viable
bone in direct contact with the implant surface, with no interposed fibrous tissue
membrane. This study showed that osseointegration can be attained regularly after
12 weeks implantation time. In the lotus nickel-free implants, the mean bone
ingrowth distance through the porous graft was 1,000 μm at 12 weeks, and the
appearance of the host cortical bone located far from the implants was similar to the
newly ingrown bone inside the pores.
Differences in bone ingrowth between AISI 316L and AISI 446-1N implants are
presented in Fig. 10.31. In general, the amount of newly mineralized bone and the
bone ingrowth distance into the implants progressively increased with time and
maximum bone ingrowth was noted by 12 weeks. The results for both implant types
showed that the alignment of the pores parallel (k) to the bone long axis permitted a
higher percentage of bone ingrowth. Histological analysis and computer-assisted
Fig. 10.30 Higher magnification light micrograph showing bone formation in apparent close
contact with lotus AISI 446-1N Ni-free stainless steel at the internal part of the implant. Bone is
stained pink, bone marrow purple, and Ni-free stainless steel is observed in black. Bar, 50 μm
Fig. 10.31 Percentage of the newly formed bone in the total available pore space in the region of
interest for lotus AISI 446-1N Ni-free stainless steel and lotus AISI 316L implants. // and ⊥
indicate parallel and perpendicular pore orientation with respect to the tibia long axis, respectively
Fig. 10.32 Changes in the bone ingrowth as a function of the depth for each implantation period
for lotus AISI-1N Ni-free stainless steel implants inserted with the elongated direction of the
pores parallel with respect to the tibia long axis (Reprinted with permission from [29] © 2008
Elsevier B.V.)
quantification of the bone growth formed revealed that the bone formation within
lotus AISI 446-1N had already reached in average 27.9 and 20.9 % after 4 weeks in
the implants inserted parallel (k) and perpendicular (⊥) to the bone axis direction,
respectively. On the other hand, for lotus AISI 316L by 4 weeks, newly formed
bone was 22.5 and 20.6 % in the implants inserted parallel and perpendicular to the
bone long axis direction, respectively. Between the two types of materials, no
statistically significant differences were noted with respect bone ingrowth.
To determine how deep the bone grew inside the lotus implants, bone ingrowth
was determined from two consecutive histological sections. The bone ingrowth
parameter was then examined as a function of depth into the implant. For lotus
AIAI 446-1N implants, these results are presented in Fig. 10.32. The bore ingrowth
decreases with increasing depth into lotus implants. After 12 weeks in the middle of
the implants, 82 and 79 % of bone ingrowth were obtained when the implants were
inserted with the elongated direction of the pores oriented parallel or perpendicular
to the tibia bone long axis direction, respectively.
Mechanical testing was performed since higher bone apposition does not neces-
sarily imply greater bone-bonding strength. The applications of compressive shear
forces at the bone–implant interface are a useful means of distinguishing bone-
bonding quality and bone contact at the interface. Changes of bonding strength
between bone tissues and the implants with the AISI 316L stainless steels are shown
in Fig. 10.33. During the push-out tests, all the implants failed through the interface
of implant and bone, and two specimens were broken on the bone side. For both
alloys, the shear strength of the interface between bone and implants increased with
an increase of the implantation time during the first 8 weeks, and this could be
Fig. 10.33 Changes in the interface shear strength between the rat femur bone and the lotus
implants with the implantation time. // and ⊥ indicate parallel and perpendicular pore orientation
with respect to the tibia long axis, respectively (Reprinted with permission from [29] © 2008
Elsevier B.V.)
attributed to the increasing with time of the extent of penetration of new bone across
the full diameter of the implants.
Although there were no significant differences among the 12 weeks strength
values, maximum interfaces shear strength of approximately 22 3 and
24 1 MPa for lotus AISI 446-1N and lotus AISI 316L, respectively, were
obtained at 12 weeks, and these values differed significantly from those of the
4 weeks group. When comparing in the literatures the failure loads of nonporous
(smooth or rough surface) and porous-coated metallic implants with the failure
loads of lotus implants, it turned out that the strength of bone–implant bonding of
the fabricated lotus implants is about ten times higher than in porous-coated or
nonporous metallic implants. According to the high shear forces needed to loose the
implants at the interface, it is concluded that fabricated lotus implants provide tight
fixation offering adequate pore volume for bone ingrowth.
10.5.3 Biocompatibility of Lotus Stainless Steel

and Titanium in Alveolar Bone
Metallic materials for artificial hip joints, femoral heads, and metal plates to fix
fragments are too strong and heavy for elderly patients, causing wear of the
patients’ bones, damaging and fracturing metals. In dental implantology,
Fig. 10.34 Insertion of lotus metals to alveolar bone (Reprinted with permission from [32]
© 2006 WILEY-Verlag GmbH & Co.)
insufficient trabecular bones contribute to poorer clinical results. It is due to

insufficient bone–metal contact. One of the solutions to this phenomenon is the
use of porous metallic materials. In natural teeth, Sharpey’s fibers make bundles to
go into cementum. Minute pores of the cementum work as scaffold to another fiber
bundles. The fiber bundles then resist vertical, lateral, and torsional stresses applied
on the teeth [30]. The biomaterials with 150–200 μm pores are confirmed to
facilitate bone ingrowth [31] and there exists a possibility to further improve
functionality of the biomaterials. The conventional manufacturing technology,
however, did not control shape and size of the pores so as to maintain certain
mechanical strength. Higuchi et al. paid special attention to lotus metals whose pore
size and orientation are controlled by utilizing gas pressure to investigate the
biocompatibility of lotus stainless steel and titanium in alveolar bone in order to
investigate the possibility of an implant by animal experiments [32].
Ten 5-year-old beagle dogs were used in this animal experiment. The dogs were
under general anesthesia with intravenously injected pentobarbital sodium.
The mandible premolars were extracted. When bone healing was obtained after
three months, epithelial mucosa was stripped at the mandible premolars to expose
Fig. 10.35 Osteogenesis in the lotus stainless steel. Observed with actual condition by
optical microscope after 4 weeks (Reprinted with permission from [32] © 2006 WILEY-Verlag
GmbH & Co.)
mandibular bone. The cortical bone was drilled with a round bur under saline
irrigation to create the implantation holes as shown in Fig. 10.34. The lotus metal
samples were then implanted and the tissue was sutured. The samples were placed
so that the pore direction is mesio-lateral to the mandibular bone. If the pore
direction is buccolingual, the openings of the pores will contact the cortical bone,
making it difficult to have active tissue ingrowth into the pores. The dogs were fed
with water and pellet feeds for 2, 4, and 8 weeks before being sacrificed. At the
end of the study period, the dogs were sacrificed with overdose of anesthetics.
The specimens were observed with actual condition microscope, optical micro-
scope, and SEM.
As a result, newly formed bones were not observed in lotus stainless steel SUS
304L after 2 weeks. In 4 weeks the active osteogenesis was observed on the surface
of the pores as shown in Fig. 10.35, and it advanced into the depth of the pores in
8 weeks. Angiogenesis was rapid in cancellous bones and slower in cortical bones
as shown in Fig. 10.36. After 4 weeks the sinusoidal vessels in porous
dehydrogenated titanium formed dense vascular plexus advancing into depth of
the pores (Fig. 10.37). Osteogenesis was identified in as early as 2 weeks. Osteo-
genesis was confirmed in the pores both after 4 and 8 weeks as shown in Fig. 10.38.
There is a technical problem to fabricate lotus titanium with uniform porosity
and pore size like stainless steel. Living tissues, however, actively grew into the
pores. In 2 weeks, active formation of vessels was observed around the metal with
invasion into the pores, which was not observed in lotus stainless steel. In 4 weeks
the sinusoidal vessels were developed into dense plexus, entering into the deep area
of the pores. It is the same with osteogenesis. Newly formed bones were observable
in 2 weeks. Pores were filled with new bones in 4 weeks. High biocompatibility of
titanium seemed to include new bone formation without induction materials.
Biocompatibility of titanium includes excellent cell adhesion property and
Fig. 10.36 Osteogenesis in the lotus stainless steel. Staining with truiduin blue after 8 weeks
(Reprinted with permission from [32] © 2006 WILEY-Verlag GmbH & Co.)
Fig. 10.37 Osteogenesis in porous titanium. Observed with actual condition by optical micro-
scope after 4 weeks (Reprinted with permission from [32] © 2006 WILEY-Verlag GmbH & Co.)
biomineralization or the ability to deposit hydroxyapatite-like calcium phosphate

on the surface under the presence of body fluid. The cells thus strongly adhered to
the titanium surface release extracellular matrix to form scaffold for tissue forma-
tion. The vessels go into the pores along the scaffold, forming network earlier than
lotus stainless steel. Early formation of capillary vessels allows early migration of
osteoblasts in the porous titanium. Combination of osteoblasts and biomineraliza-
tion [33] provides titanium better osteogenic environment than stainless steel.
Though titanium is a metal, it has tissue compatibility equal to that of ceramics.
One of the reasons for this is that titanium is easily oxidized and covered with an
oxide layer. Exposure of titanium to body fluid through this oxide layer was
Fig. 10.38 Osteogenesis in porous titanium. Staining with truiduin blue after 8 weeks (Reprinted
with permission from [32] © 2006 WILEY-Verlag GmbH & Co.)
considered to be causing ceramic-like behavior. Recently, however, it was

found that immersion of titanium in body fluid-like solution produced the layer of
hydroxyapatite. This phenomenon was not observed in gold alloy, gold–silver–
palladium alloy, cobalt–chromium–nickel alloy, nickel–chromium alloy, or in
stainless steel alloys. Furthermore, deposit on the Ti-6Al-4V or on Ti-50Ni
alloy is not apatite. Higuchi et al. [33] conducted in vitro study that extracellular
matrix became more absorbable to titanium through Ca ions. The favorable envi-
ronment for extracellular matrix adsorption via Ca ion is thus created around the
implant surface. Extracellular matrix formed in the tissue around titanium makes
barrier to through extracellular matrix increases physical resistance, contributing to
the initial fixation of titanium implants.
Regenerative medicine using cell proliferation has been the topic of research
recently. It takes long time, however, before this technique becomes a mainstream
due to such problems as carcinogenicity, ethical issues, complicated culture opera-
tion, and high cost. Scaffold for cell proliferation is indispensable for living tissue
architecture in the regenerative medicine. The materials currently used are conven-
tional biomaterials. Such hard tissue as bones and teeth needs to have stress
resistance. It is highly expected to develop better combination of stiff biomaterials
such as lotus metals and regenerative medicine for tissue repair. Development of
artificial bones (Fig. 10.39) and joints of metals and ceramics has been active
reconstruct structure supporting body and interlocking with joints.
Fig. 10.39 Parenchymatous defect of the mandible and biomaterial (Reprinted with permission
from [32] © 2006 WILEY-Verlag GmbH & Co.)
Fig. 10.40 Application of lotus metal for parenchymatous defect of the mandible (Reprinted with
permission from [32] © 2006 WILEY-Verlag GmbH & Co.)
Metals have been used as biomaterials since it has sufficient stiffness to resist
stress concentration fracture. Bones bear body weight, the stress of which acts as
stimulus for repetitive remodeling. Remodeling is important in maintaining
trabecular structures. Long implantation of stiff metals in the bone reduces
stimulus to the adjunct bone, causing disuse atrophy. The challenge for the future
is to make metal stiffness close to that of the bone (elastic modulus: 7–30 GPa).
The lotus materials allow porosity control, giving metals elasticity. It is one of the
ways to provide artificial materials with elastic modulus similar to bone modulus
(Fig. 10.40). Porous metals with pore of 150–200 μm are said to be ideal for
Fig. 10.41 In the implant case, the bone resorption in maxilla mandible posterior is remarkable
(Reprinted with permission from [32] © 2006 WILEY-Verlag GmbH & Co.)
artificial bones. The pores allow ingrowth of vessels and bones in the pores.
They have a potential to develop into artificial bones with bone marrow, which
is the final goal.
The implants allow contact between the metal and the gingiva. Absence of
fibrous connection makes it more probable that bacteria and food residues enter
from peri-implant space. Without more careful cleaning after implantology treat-
ment, patients will experience peri-implant inflammation, which may be the cause
of implant removal. Implants are composed of three components: super structures,
fixtures, and abutment going through mucosa. Greatest difference between dental
implants and other biomaterials is that implants communicate with ex vivo envi-
ronment. Biomaterials other than dental implants are implanted inside the body,
giving little opportunity for external bacteria coming into the body. Dental
implants, on the other hand, are easily exposed by external bacteria. It is the reason
why we have to make thorough oral care.
It has been impossible or less successful to use conventional dental implants for
bones with compromised volume or osteoconduction (Fig. 10.41). Lack of fibrous
connection gives no buffer against occlusive impacts, reducing abutment–mucous
membrane connection. Porous implants should be developed to solve issues we
currently face. There are several ways to manufacture porous implants: metal
deposition onto fixture surface under high temperature and deposition through
melting metal beads and fixture surface. The manufacturing process of porous
implants with high functionality will allow to create pores in fixtures without
depositing metal beads.
Lotus stainless steel and titanium can be utilized to create pores of 150–200 μm.
It will allow growth of bony tissues alone in the pores to enhance holding force.
Pore size can be controlled to 10 to 20 μm to allow fibers alone to come into the
pores to add buffer against stress. Tailor-made many different types of implant can
Fig. 10.42 Serve as basis for bone and fibrous tissue regeneration (Reprinted with permission
from [32] © 2006 WILEY-Verlag GmbH & Co.)
be produced. Some can have strongly physical connection by fibers at the site where
body communicates with external environment. The lotus metals can be better
controlled to improve implant functions (Fig. 10.42). Fixture pore size can be
controlled to 150–200 μm to allow only bony tissue to come into the pores and to
enhance holding force. The size can be decreased to 10–20 μm to allow only fibrous
tissue to come into pores, thus adding buffer to the stress. The fibrous connection is
applied at the site of communication between implants and the external environ-
ment, enhancing physical binding strength. In the future the porous metals with
porosity of 80 % can be used to tailor-made artificial bones.
References
1. Nishio S (2001) J Thermal Eng. Jpn Soc Mech Eng 617–622

2. Wei X, Joshi Y (2004) J Electron Packaging 126:60–66
3. Bastawros AF, Evans AG (1999) ASME EEP (Am Soc Mech Eng Electronic Packaging Div)
1:733–736
4. Hunt ML, Tien CL (1988) Int J Heat Mass Trans 31:301–309
5. Kamiuto K, Yee SS (2005) Int Comm Heat Mass Trans 32:947–953
6. Zhang HY, Pinjala D, Joshi YK, Wong TN, Toh KC, Ilyer MK (2004) Proceedings of IEEE
ITherm 2004, pp 640–647
7. Chiba H, Ogushi T, Nakajima H, Ikeda T (2003) JSME J Series B—Fluids Thermal Eng
47:5843–5847
8. Chiba H, Ogushi T, Nakajima H (2011) Proceedings of ASME/JME 2011 8th thermal
engineering joint conference, 13–17 March, Honolulu, ATTEC2011-44108
References 271
9. Shah RK, London AL (1978) Advances in heat transfer series, Suppl 1. Academic, New York,
pp 192–197
10. Shah RK, London AL (1978) J Fluids Eng 100:177–179
11. Chiba H, Ogushi T, Nakajima H (2010) J Thermal Sci Tech 5:222–237
12. Yoshida I, Monma D, Iino K, Otsuka K, Asai M, Tsuzuki H (2003) J Alloy Comp 355:79–84
13. Golovin IS, Sinning HR (2003) J Alloy Comp 355:2–9
14. Xie ZK, Tane M, Hyun SK, Okuda Y, Nakajima H (2006) Mater Sci Eng A 417:129–133
15. Ota K, Ohashi K, Nakajima H (2003) Mater Sci Eng A341:139–143
16. Nakajima H (2007) Prog Mater Sci 52:1091–1173
17. Greiner C, Oppenheimer SM, Dunand DC (2005) Acta Biomater 1:705–716
18. Jee CSY, Ozguven N, Guo ZX, Evans JRG (2000) Metall Mater Trans B 31:1345–1352
19. Oh IH, Nomura N, Masahi N, Hanada S (2003) Scr Mater 49:1197–1202
20. Bram M (2000) Adv Eng Mater 2:196–199
21. Cameron HU, Lilliar RM, Macnab I (1976) J Biomed Mater Res 10:295–302
22. Bobyn JD, Toh KK, Hacking SA, Tanzer M, Krigier JJ (1999) J Arthroplasty 14:347–354
23. Galante J, Rostoker W, Lueck R, Ray RD (1971) J Bone Jt Surg 53A:101–114
24. Zhang X, Ayers RA, Thorne K, Moore JJ, Schowengerdt F (2001) Biomed Sci Instrum
37:463–468
25. Story BJ, Wagner WR, Gaisser DM, Cook SD, Rust-Dawicki AM (1998) Int J Oral Maxillofac
Implants 13:749–756
26. Alvarez K, Hyun SK, Fujimoto S, Nakajima H (2008) J Mater Sci Mater Med 19:3385–3397
27. Speidel MO (2004). US Patent 6,682,582
28. Gavriljuk VG, Berns H (1999) High nitrogen steels. Springer, Berlin, pp 267–268
29. Alvarez K, Hyun SK, Nakano T, Umakoshi Y, Nakajima H (2009) Mater Sci Eng C
29:1182–1190
30. Sims MR (1984) Connec Tissue Res 13:59–67
31. Hulbert SF, Cooke FW, Klawitter JJ, Leonard RB, Sauer BW, Moyle DD (1973) J Biomed
Mater Res 4:1–23
32. Higuchi Y, Ohashi Y, Nakajima H (2006) Adv Eng Mater 8:907–912
33. Higuchi Y, Takahashi K, Asai T, Murata Y (2001) J Osaka Dent Univ 35:83–86
Chapter 11
Summary
This overview summarized the present status of the research of porous materials
with directionally elongated pores. In comparison with the conventional porous and
foamed materials with isotropically almost spherical pores, the lotus materials
exhibit peculiar features derived from alignment of long pores such as anisotropy
of various properties. Besides, these materials possess lightweight and integrated
mechanical properties. Thus, these materials should be considered as one of the
advanced engineering materials from the point of view of not only materials science
but also applications for the practical use. In this review, the author mentioned only
a few kinds of applications, but there are a variety of possible prospective
applications such as air bearings, filters, mechanical parts of airplanes, and
automobiles. We hope the further research and development of this kind of
materials will surely be expanded to widen a path to porous materials science and
technology.

About the Author
Hideo Nakajima received his Bachelor’s degree in 1971 and Ph.D. from Tohoku
University in 1977. Then, he was a postdoctoral associate at Rensselaer Polytechnic
Institute, USA, until 1980. From 1980 to 1992, he was an assistant and associate
professor at Institute for Materials Research, Tohoku University. In 1992 he moved
to Iwate University as a professor. Since 1996, he had been a professor of Institute
of Scientific and Industrial Research, Osaka University. He has been the director of
the Wakasa Wan Energy Research Center at Tsuruga in Fukui Prefecture since
2012 and is an emeritus professor of Osaka University.
His research interests focus on fabrication, properties, and application of porous
metals. He was awarded the Medal with Purple Ribbon in 2009 by the Emperor of
Japan.

DOI 10.1007/978-4-431-54017-5, © Springer Japan 2013
Index
A Apparent attenuation coefficient, 246–247, 249

Absorption coefficient-frequency curve, 188 Arc welding, 211
Absorption sound energy, 188 Argon, 11, 22, 26, 28, 37–42, 46, 47, 50,
Acidic chloride environment, 208 52, 54–60, 68, 69, 71–73, 87–90,
Acoustic emission method, 142–148 93–101, 117–119, 211, 216
Activation energy, 33, 50 Artificial bones, 267–270
Active-passive transition, 205 Artificial hip joint, 263
Agitation, 76–79 As-cast porous iron, 97
Air bearings, 273 Atmosphere, 16, 20, 22, 25, 26, 37–40, 46–49,
Air bubbles, 74 52, 54–58, 60, 65, 66, 68, 71, 79,
Air cooling, 233–242 86–91, 101, 103, 104, 107, 109,
Airplanes, 273 115–119, 128, 135, 140, 161, 254
AISI 446, 204–207, 253, 256, 258, 260–263 Attenuation(damping) coefficient, 245–247,
AISI 316L, 202, 204–208, 253–254, 256, 249–251
258, 260–263 Attenuation constant, 187–189
Alporas, 9, 162 Attenuation of a sound, 187, 188
Aluminide, 102 Austenitic stainless steels, 202, 208, 256, 258
Aluminum, v, 8–10, 12, 20, 42–47, 79, 80, Automobiles, v, 273
103, 112, 115–125, 162, 183,
187, 211, 251
Aluminum oxide, 8 B
Alveolar bone, 263–270 Bamboo, 4, 5
Ambient gas pressure, 87–89 Base-plate surface area, 236, 238, 242
Anelastic strain, 135 Beagle dogs, 264
Animal bones, 1 Bending, 4, 151, 152, 158, 169–175
Animal experiments, 264 Bending strength, 169–172
Anisotropic behavior, 139, 140 Beryllium, 20
Anisotropic cylindrical pore structure, 47 Binary equilibrium phase diagram, 99, 101
Anisotropic tensile deformation, 142 Biocompatibility, 254, 263–270
Anisotropy, 4, 5, 8, 19, 47, 127, 130, 131, 135, Biological fixation, 253, 258
136, 139, 140, 142, 143, 148, 150, Biomineralization, 265–266
154, 162, 166, 174, 190, 192–200, Blowholes, 74, 217
213, 215, 221, 234, 257, 273 Bone, 1–4, 261, 266–270
Annealing, 109, 132–133 healing, 257, 264–265
Anti-corrosion, 102, 109 ingrowth, 253, 258–260, 262–264

DOI 10.1007/978-4-431-54017-5, © Springer Japan 2013
278 Index
Boyle–Charles law, 40 Control of pore size, 16, 83–92

Boyle’s law, 47, 87, 105. 107, 108 Conventional groove fins, 234, 235, 238,
Breakdown potentials, 203–205 239, 241, 242
Bubble coalescence, 79, 80 Conventional mold casting technique, 23, 24
Buckling, 149–152, 168 Cooling performance, 232, 233
Burst AE signals, 142, 144 Copper, 16, 67, 83, 98, 130, 183, 211, 231
Corrosion
potentials, 203, 204
C rates, 203, 207, 208
Calcium hydroxide, 36, 42–44, 46 resistance, 20, 93, 202, 203,
Carbon dioxide (CO2), 42, 44–47, 67, 74–79, 205, 208, 256
216, 254 Cortical bone, 259, 260, 265
Carbon monoxide (CO), 3, 10, 42, 44 Cottrell–Lomer locks, 139–140
Casting, v, 3, 5, 10, 12, 16–28, 32, 37–47, 49, Crack formation, 142, 147, 148
52–54, 57–60, 65, 74, 83–88, 90, Crack-initiation sites, 179
97, 98, 100–102, 104, 106, 109, Cross-sectional area, 30, 69, 173, 190, 192
112–117 CTE. See Coefficient of linear thermal
Casting defects, 3 expansion (CTE)
Casting technique, v, 5, 16–47, 87 Cumulative AE counts, 142, 144
Cellular materials, 1, 2, 4, 244 CVD. See Chemical vapor decomposition
Ceramics, 54, 55, 83, 84, 93–125, 266–267 (CVD)
Channel angle, 221–229 Cylindrical pores, 77
Chemical equilibrium, 65 Cytoplasmic prolongations., 255
Chemical vapor decomposition Cytotoxicity testing, 254
(CVD), 10 Czochralski method, 16, 99, 101
Chesnut-shaped pores, 75–76
Chiller, 17–19, 42, 54, 58, 77, 83, 84, 90,
100, 105, 110, 111 D
Chromium, 20, 253 Damping capacity, 244, 245, 251–252
Chromium nitride, 47, 49, 50 Damping materials, 244, 251
Clausius–Clapeyron equation, 33 Decomposition peak temperature, 52
Closed pores, 8, 183, 187, 216, 217 Decomposition pressure, 95
Cobalt, 20, 57–61, 196, 197, 267 Deformation rate, 156
Coefficient of linear thermal expansion Dendrite, 102
(CTE), 201 Dendritic arms, 102
Collagen cell culture dish, 255 Dendritic growth, 15
Collapse, 9, 54, 79–81, 157, 158, 167 Dendritic structure, 102
Columnar pores, 74–76 Densification, 225
Composite materials, 156, 189, 192 Densification region, 151, 152, 161
Compressive direction, 148, 149, 152, 153, Dental implants, 269
156, 158, 160, 161, 168, 170 Derivative thermogravimetry (DTG), 50
Compressive properties, 148, 153–159, 167, 171 Diatomic, 65
Compressive strength, 148–169 Diffusion, 11, 14, 34, 74, 78, 79, 109,
Compressive stress, 148, 152, 154, 157, 120–123, 125
165–168, 170 Diffusion coefficients, 120–122
Compressive yield strength, 148–151 Dimple pattern, 174
Concave shape, 34, 112, 114, 115 Directional pores, v, 1, 4, 5, 13–62, 79,
Constitutional supercooling, 76 115–117, 121, 131, 133, 193, 233
Continuous casting technique, 24–36, Direction of solidification, 3
52–54, 112–125 Dislocation motion, 135
Continuous zone melting technique, 20–25, Dissociation constant, 57
27, 47–52, 86–88, 102, 103, 127, Dissolving gas, 16, 19, 47, 79, 93, 113
153, 165, 254, 258 Dissolving temperature, 37
Index 279
Distilled water, 42, 75, 254 Excited vibrations, 245, 249

Dog-bone type specimens, 172 Exit channel, 221–224, 228
Drainage, 8, 9, 91 Explosive gas, 36
Dual-phase formation, 120 External pressure, 40, 42, 46, 80
Ductile copper, 142 Extruded length, 225, 226
Ductility, 136, 152, 156, 167, 168, 202 Extrusion, 106, 221–223, 225–228
Dummy bar, 25, 26, 52, 113, 117
Dynamic viscosity, 237
F
Fabrication techniques, 6, 8–62, 89,
E 93–125, 138, 257
ECAE. See Equal-channel angular extrusion Fatigue strength, 172–179
(ECAE) Femoral heads, 263
Effective electrical conductivity, 194, 195 Fe4N, 36, 97
Effective electrical resistivity, 194 α(Ferrite)-iron, 97
Effective-mean-field (EMF), 127, 193, Ferromagnetic materials, 196
195, 196 Filtering materials, 4
Effective thermal conductivities, Filters, 102, 106, 241, 273
189–193, 233 Fin gap, 234, 238, 241
Elastic anisotropy, 130 Fluid flows, 123, 237
Elastic constants, 127, 128, 130, 249 Fluid permeability, 3
Elastic modulus, 129, 130, 151, 268 Foam, v, 5, 7–12, 16–19, 79–81, 162, 187
Elastic properties, 127–131 Foamed aluminum, v, 183, 211
Elastic region, 151, 152 Foamed materials, 273
Elastic stiffness, 128, 129 Foaming agent, 8–10, 211
Electrical conductivity, 129, 193–196, 199 Fourier transforms, 245, 247, 249
Electrical current, 194, 199 Four-probe method, 193, 194
Electric current density, 195 Fracture mode, 136
Electric field, 194, 195 Fracture toughness, 202
Electrochemical polarization Functional materials, v, 5, 102, 160
experiments, 203
Electrode materials, 4, 193
Electrodeposition, 10 G
Electromagnetic acoustic resonance Gas analyzer, 42
method, 128 Gasar process, 79
Electron probe microanalysis (EPMA), Gas-blow cooling, 86–89
109, 110 Gas-blower, 87
Ellingham diagram, 58 Gas bubbles, 3, 39, 42, 66, 74, 79, 107
Elliptical mirror, 106 Gas chromatograph, 42
Elliptical pores, 74, 139 Gas concentration, 17, 216
Elongated pores, 3, 17, 23, 55, 56, 102, 110, Gas element, 36, 37, 43, 44
165, 167–168, 171, 188, 196, 273 Gas-evolution crystallization reaction,
EMF. See Effective-mean-field (EMF) 16, 20, 93, 99
Empirical Archie’s law, 194, 196 Gas filter, 106
Energy absorption, 148, 151–152, 160 Gas-forming compounds, 44
Entry channel, 221–225, 228 Gas-injection, 8–9
Equal-channel angular extrusion (ECAE), Gas precipitation, 65
221–229 Gibbs free energy, 33, 58
Equilibrium distribution coefficient, 34 Glass wool, 183, 187
Eshelby’s equivalent inclusion theory, 195 Glide plane decohesion, 145
Eutectic alloys, 120, 121 Golf putter, 6, 251–253
Evaporation, 10, 11 Grain boundary (GB) sliding, 135
Excitation, 128, 129, 131, 249 Graphite crucible, 37, 42, 52, 83
280 Index
Groove fins, 234, 235, 238, 239, 241, 242 Intermetallic compounds, 93–125, 164–165
Growth, 5, 9, 14, 15, 19, 65–81, 98, 100, Intermittent motion, 14, 112, 113
121–122, 147, 221 Internal friction (IF) method, 131–135, 247, 248
Guldberg–Waage’s law, 65 In vitro cytocompatibility, 253–256
In vivo osteocompatibility, 256–263
Iron, 16, 20, 47–50, 93–98, 112,
H 127–130, 140, 141, 160–162, 164
Hammering-vibration-damping tests, δ-Iron, 94–95, 97
244, 245, 249–251 γ-Iron, 96–97
Heat dissipation rates, 232 Isotropic, 7–8, 19, 79, 130, 196
Heat flow rate, 34, 236, 242
Heat flux, 190, 232
Heat sink block, 241 J
Heat sinks, v, 6, 189, 232–244 Joining techniques, 211
Heat transfer capacity, 232–236, 238, 239,
241, 242
Heat transfer coefficient, 236–240, 242–244 K
Heat transmission, 89 Kissinger plots, 50
Helium, 48, 49, 87–89, 165, 168–170
Heterogeneous pore nucleation, 33
High frequency electronic device, 232 L
High-pressure gas method (PGM), 16–38, 40 Lamellae, 122
High-pressure hydrogen, 16, 36, 52, 67, Laminar flow, 237, 238
216, 217 Laser beam, 216
High temperature strength, 102 power, 211–215
Hollow sphere, 11 welding, 217
Homogeneous nucleation, 67 Laser diodes, 232
Honeycomb, 1, 4 Laser welding, 211–214, 216, 218
Hydraulic diameter, 238 Lattice Boltzmann method, 33, 79
Hydrogen diffusivity, 27, 34, 40, 67–68, Lattice defects, 131
85–86, 121, 123, 125 Leaves, 1, 2
Hydrogen solubility, 17, 34, 72, 87, 103, Legendre functions, 128
106, 107, 109, 110, 116, 121, Light-weight property, 3, 273
123, 125, 134 Lightweight structural materials, v, 4,
Hydroxyapatite, 267 102, 112, 115, 160, 164–165
Hydroxyapatite-like calcium phosphate, Liquid-solid phase, 19, 39, 98, 107, 121
265–266 Load-bearing areas model, 138
Loading misalignment, 149
Local effective stress, 158
I Local strain, 167
Ice-freezing, 13, 14, 74 Longitudinal direction, 48, 114, 172, 194
Ice wormholes, 13–15 Lotus alumina, 106–109
Impact energy absorbers, 160 Lotus aluminum, 20, 42–47, 115–125
Implant, 202, 253, 254, 257–264, 267, Lotus brass, 109
269, 270 Lotus carbon steel, 112–115
Impurity atoms, 131 Lotus copper, 20, 72, 83, 98, 130, 185,
Induction heating coil, 25, 42, 70, 83 211, 233
Infrared pyrometer, 42 fin, 233, 234, 237, 239, 241
Inhomogeneous distribution, 44 heat sinks, 233, 234, 238, 239, 241, 242
Inlet temperature, 236, 241–242 Lotus magnesium, 100, 109–112, 130–131,
In situ observation, 76, 80 183, 187, 217–221, 244–251
Insoluble gas, 16, 23, 37, 67, 79 Lotus materials, 107, 131, 196, 199, 244,
Instron 4482 universal testing machine, 258–259 268, 273
Index 281
Lotus metals, 16, 18, 20, 22–26, 36–38, 42, Mold casting technique, 16–24, 37–47,
47, 52, 74, 93–125, 127–179, 83–85, 88, 98, 100, 101, 104, 109
183–208, 211–229, 231–270 Molten metal, 17, 25, 36, 37, 55, 59, 60, 79,
Lotus nickel-free austenitic stainless steel, 84, 106, 116, 122, 216
256, 258 Molybdenum-sheet mold, 54
Lotus nickel-free implants, 260 Monolithic implants, 253
Lotus stainless steel, 23–25, 86–89, 153–158, Mori–Tanaka’s mean-field theory, 159
254, 263–270 Mushy zone, 111, 112
Low strain rate, 163
N
M Nano-sized porous silicon, 104
Macro-buckling, 150 Navier–Stokes equation, 79
Magnesium (Mg), 16, 20, 22, 54, 100, NiAl, 102, 164
109–112, 116, 117, 130, 131, Ni3Al, 102, 164
152, 187, 217–221, 244–251 Nickel (Ni), 10, 16, 20, 54–58, 89–91, 103,
Magnetic fields, 129, 131, 196–200, 202 193–197, 199, 201, 202, 256, 258
Magnetic flux, 199 Ni-free stainless steels, 202, 203, 207, 208,
Magnetic materials, 196 253–263
Magnetic permeability, 202 NiO powder, 89–92
Magnetic poles, 199 Nitrogen (N), 16, 20, 52, 93–98, 127, 128,
Magnetization, 196–201 140, 202–204, 216, 254, 256
Mandibular bone, 265 Non-permeable pores, 185, 188
Manganese (Mn), 20, 110 Nucleation, 5, 14, 62, 65–81
Mass spectrometer, 132, 215, 216 sites, 40–42, 44, 56, 89, 91, 92
Masticability, 3 theory, 67
Materials strength, 112, 138 Number of cycle to fatigue, 173
μCT. See Micro computer tomography (μCT) Numerical-differentiation method, 158–159
Mean-field theory, 153, 158, 159, 195
Mechanical interlocking, 253
Mechanical property, v, 6, 20, 74, 93, 106, O
127–179, 193, 221, 253, 256, Open pores, 8, 179, 187, 232
257, 273 Optical micrographs, 20–22, 37, 38, 43,
Medical devices, v, 6, 253–270 54–57, 84, 93, 94, 98, 102–104,
Melting point, 9, 33, 34, 37, 52, 96, 97, 116 149–151, 167–170, 173, 175
Melting temperature, 9, 19, 20, 37, 47, 50, Orientation of pore axis, 169, 171
52, 71, 72, 93, 116–121, 124, 125, Orthorhombic symmetry, 128, 189, 192
134, 135 Osseointegration, 253, 258, 260
Mercury, 8 Osteoblast, 254–257, 266
Metallic compound, 37 Osteoblast cells, 257
Micro-buckling, 150 Outlet temperature, 236, 241–242
Microchannels, 232, 241, 242 Overall porosity, 16, 93
Micro computer tomography (μCT), 79 Oxidation resistance, 102
Microfocus X-ray source, 80–81 Oxide particles, 40, 42, 44, 89–92
Micromechanical mean-field theory, Oxygen, 9, 11, 16, 20, 40, 42, 45–46, 55,
153, 158, 159 56, 58, 59, 98–102, 202, 216
Microscopic deformation, 142, 163, 164
Middle strain rate, 160–164
Misorientation of pores, 154 P
Moisture, 54–62, 89 Parallel loading, 142–148, 175, 178, 221
Moisture dissociation, 58 Partial pressure, 52, 60, 65, 66, 72, 73, 87,
Mold, 10, 17, 19, 24–26, 37–43, 52, 54–60, 81, 89, 93–97, 107–108, 116–121,
83–85, 90, 98, 100, 101, 111, 112, 117 124–125
282 Index
Passive current, 204, 205 Porous materials, 1–8, 129–130, 138, 139,
Passive film, 202, 204, 205 157–158, 185, 187, 196–197, 203,
Peak temperature, 50–52, 132, 133 221, 232–233, 244, 273
Penetration depth, 211–214, 217 Porous metallic biomaterials, 253, 264
Permeability, 3, 8, 47, 185, 188, 193, 202 Porous metals, 5, 8, 13–62, 83–93, 106, 130,
Permeable pores, 8, 185, 188 136, 147–149, 157, 160–162, 167,
Perpendicular loading, 142–147, 178 172, 193, 195, 201, 202, 211, 221,
Piezoelectric transducer, 76–77, 128 224, 232, 233, 244, 253, 258, 264,
Pinch roller, 25, 27, 52 268–270
Pitting corrosion, 202, 203, 205 Porous metals with directional pores,
Pitting corrosion susceptibility, 205, 207 5, 13–62
Plastic deformation, 139–140, 142, 144, Porous oxide ceramics, 106
170, 221 Porous semiconductors, 104
Plastic hardening, 139–140 Potentiodynamic polarizations, 203–206
Plastic Poisson’s ratio, 142 Powder metallurgy, 9, 193, 253
Plastic yielding, 151, 152 Power devices, 232, 234
Plateau-stress region, 157, 158, 161, 165, 168 Power-law relation, 195
Poisson’s ratio, 130 Precipitation, 23, 37, 42, 65, 67, 111, 116
Pores, 3, 7, 13, 66, 83, 94, 127, 183, 211, Precursor, 10, 11
232, 273 Preheating, 49
aspect ratio, 8, 34, 90, 91 Pressure drop (Pa), 232–233, 236–240,
diameter, 8, 29–31, 34, 37, 39–41, 242, 243
47–49, 60, 67, 84–92, 102, 103, Pressurized hydrogen, 16, 47, 58, 60, 62,
105, 106, 108, 116, 117, 119–121, 89, 104, 106, 107, 117, 194
124, 168–170, 185–188, 192 Proeutectoid ferrite, 97
evolution, 16–17, 22, 25, 37, 52, 58, 60, 66, Projection angle, 154
103, 105, 112, 120, 122 Proliferation test, 254, 255, 267
evolution model, 55 0.2% Proof stress, 154, 165–167
growth direction, 17, 19, 30, 31, 34, 37, Pumping power, 239, 240, 242, 244
148, 195–200, 213–218, 221 Punch displacement, 225, 227
length, 8, 30, 31, 33, 34, 68–70 Push-out testing, 258–259, 262
morphology, 8, 17, 19, 25, 37, 42, 46,
76–79, 84, 102, 109–112, 117, 119,
172, 203, 223, 228 Q
nucleation, 15, 35, 46, 56, 66, 67, 74, 91 Quadrupole mass spectrometer, 132
mechanism, 33 Quenched porous iron, 97
rate, 33–34
number density, 27, 30, 86, 90–91
orientation, 8, 127, 135, 136, 139–141, R
153–160, 258, 261, 263 Rat bone, 258, 259
shape, 7–8, 20, 74, 76, 78, 79, 98–100, 102, Real-time radioscopy, 80–81
107, 109–110, 138, 195, 199, 264 Relative density, 7–11
size, 5–6, 8, 10, 11, 16, 19–27, 37, 39, Remodeling, 257, 268
42–44, 46, 47, 52, 55, 58, 67, 83–92, Resonance frequencies, 128, 249
99, 102, 104–106, 108, 116–117, Resonant-vibration, 245, 247–249
131, 196, 203, 233, 252–253, 258, Reynolds number, 237, 238
264, 265, 269, 270 Ripple marks, 112, 113
volume fraction, 8, 150–151 Round-shape pores, 102, 104–105
wall, 162–164, 199, 207, 223
Porosity, 2, 7, 16, 69, 86, 93, 128, 185,
221, 233 S
Porous biomaterials, 257–258 Saturation magnetization, 196–197
Porous electrode materials, 193 Scaffold, 264, 266, 267
Index 283
Scanning electron microscope (SEM), 5, 142, Split Hopkinson pressure bar (SHPB)
145, 146, 162, 205–207, 265 method, 160
Schmid factor, 221 Sponge, 1, 2
SEM. See Scanning electron microscope Sprague–Dawley rats, 258
(SEM) Stainless steel, 11, 20–25, 27, 37, 42, 83,
Semibrittle surface, 145 86–89, 120, 121, 153–160,
Semiconductors, 93–125 202–205, 207, 208, 244, 251,
Severe plastic deformation, 139–140, 221 253–270
Shear flow stress, 225, 227 Stalks, 1, 2
Shielding gas, 211, 216 Standing-wave method, 183–184
SHPB method. See Split Hopkinson pressure Stiffness, 4, 5, 128, 129, 256–258, 268
bar (SHPB) method Straight pores, 194, 232, 233
Sieverts’ law, 17, 22, 59–60, 65–66, 79, 87, Stress amplitude, 172–174, 176
94–95, 105, 107 Stress concentration, 130, 137, 140, 141,
Silicon, 57–62, 104–105 147–151, 158, 167, 268
Silicon carbide, 8 Stress concentration model, 138
Silver, 16, 20, 98–102, 267 Stress plateau, 151, 152, 157, 158, 161,
Silver-oxygen system, 98, 99, 101 165, 168
Simple straight fin model, 233 Stress-shielding effect, 258
Sintered materials, 7–8, 10, 11, 138, 232 Stress-strain curves, 136, 140–143, 145, 148,
Skin layer, 1–2, 32, 34, 36 150–152, 154, 155, 160, 161,
S-N curves, 173 165–167
Sodium bicarbonate, 42–44 Stress triaxiality, 145–148
Solidification defects, 3, 4, 115–116 Striation, 174, 176
Solidification direction, 16, 17, 20, 23, 26, Supersaturated hydrogen, 39, 40, 86, 90, 105,
37, 38, 40–43, 46–48, 53–58, 70, 121, 216
75, 83, 85, 86, 98–105, 107, 111, Surgical implant material, 202, 254
117–119, 122, 127–128, 142, 143, Swelling, 10, 79
154, 155, 221
Solidification rate, 19–21, 47, 106, 117
Solidification shrinkage, 24, 68, 69, 123 T
Solidification velocity, 19, 20, 22–25, TDM. See Thermal decomposition method
33–35, 39, 47, 67, 83–87, (TDM)
116–122, 124, 125 TDS. See Thermal desorption
Solidified ingots, 3, 4, 21, 24–25, 37, 42, spectrum (TDS)
47, 55, 58, 68, 71, 85, 90, 102, Temperature distribution, 49, 190, 191
119, 194 Temperature gradient, 19, 49, 67–69, 87,
Solid-state diffusion, 109 105, 116–120, 124, 125, 191
Sound absorbers, 160 Template, 10
Sound absorbing material, 183, 185, Tensile elongation, 145
187, 188 Tensile strength, 135–148
Sound absorption, 183–189 Thermal conduction, 187
Sound absorption coefficient, 183, 184, 186 Thermal conductivity, 21, 22, 47, 49, 84,
Sound energy, 187, 188 189–193, 216, 217, 233, 237
Sound pressure, 184 Thermal decomposition, 8
Sound reflectivity, 184 Thermal decomposition method
Sound tube, 183, 187 (TDM), 36–54
Spark-erosion wire cutting machine, 37, 42, Thermal desorption spectrum
131, 165, 218 (TDS), 132, 135
Spatial distribution, 74, 199 Thermal dilatometer, 201
Specific strength, 167, 221 Thermal energy, 188
Sphere-shaped pores, 107 Thermal expansion, 201–202
Spherical pores, 17, 19, 46, 54–55, 75, 79–81 Thermal resistance, 236
284 Index
TiAl, 164–170 Viscosity, 8, 9, 78, 187, 237

TiNi, 102 Viscous friction, 188
Tissue, 1, 253, 258, 260, 262, 265–267, 270
Titanium, 11, 20, 40, 43, 44, 263–270
Titanium hydride (TiH2), 9–11, 20, 37–44, W
47, 52 Water-carbon dioxide solution, 75–77, 79
Trabecular bone, 1, 4, 263–264 Water cooling, 17, 21, 25, 54, 83, 233, 234,
Transfer direction, 26, 28, 30, 31, 48, 241–244
114, 131 Weldability, 211–221
Transfer velocity, 22–23, 25–36, 47–52, 76, Welded joint, 217, 218, 220
87–89, 102, 112, 114, 115, 131 Welding, 211–218
True-stress-true-strain curves, 143 Welding speed, 211–215
Tungsten, 20 Wood, 1, 4
Work-hardening, 143, 144, 160, 161, 221, 228
Work-hardening coefficients, 143, 144
U
Ultimate tensile strength, 136–142, 178,
179, 202 X
Ultrasonic vibration, 76–79 Xenon lamp, 106, 107
Uniaxial tensile behaviour, 136 X-ray
Unidirectional solidification, 16–18, 22, 37, diffraction, 130
42, 47, 54, 57, 58, 66–68, 74–79, radiograms, 81
98, 102, 104, 106–109, 112, radiograph, 215
115–117, 122, 130, 148, 154,
155, 194, 216, 221
Universal Instron testing machine, 160 Y
Yield strength, 129, 136, 137, 139–141, 148–151,
154, 156–159, 164, 166, 202
V Young’s moduli, 128–131
Vapor deposition method, 109, 253
Vibration damping, 112, 244–251, 253
Vibration-damping curve, 245–250 Z
Vibration modes, 128, 245, 246, 249 Zinc (Zn), 109, 110

Porous Metals With Directional Pores

Uploaded by

Copyright:

Available Formats

Porous Metals With Directional Pores

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

Porous Metals With Directional Pores

Uploaded by

Copyright:

Available Formats

Hideo Nakajima

ISBN 978-4-431-54016-8 ISBN 978-4-431-54017-5 (eBook)

© Springer Japan 2013

Printed on acid-free paper

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

Tsuruga, Japan Hideo Nakajima

3.3 Thermal Decomposition Method (TDM) . . . . . . . . . . . . . . . . . . 36

6.8 Fabrication of Lotus Carbon Steel by Continuous

10 Various Applications of Lotus Metals . . . . . . . . . . . . . . . . . . . . . . . 231

About the Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

Abstract In natural world, a number of porous materials are observed, for

Keywords Bamboo • Cancellous bone • Cellular materials • Natural biomaterials

1.1 Porous Materials Widespread in Natural World

H. Nakajima, Porous Metals with Directional Pores, 1

the ground as shown in Fig. 1.3 [http://www3.famille.ne.jp/~ochi/kaisetsu-01/05-

Chap. 5 were described. Then we described the details of fabrication techniques of

2.1 Materials Definition

H. Nakajima, Porous Metals with Directional Pores, 7

anisotropic and long cylindrical, which is evolved unidirectionally during the

2.2 Various Fabrication Methods

2.2.1 Melt Gas Injection (Air Bubbling) [4]

2.2.2 Gas-Releasing Particle Decomposition in the Melt [4]

2.2.3 Gas-Releasing Particle Decomposition

2.2.4 Casting Using a Polymer or Wax Precursor

2.2.5 Metal Deposition on Cellular Preforms [4]

2.2.6 Entrapped Gas Expansion [4]

2.2.7 Hollow Sphere Structures [4]

Several approaches have recently emerged for synthesizing hollow metal

2.2.8 Co-compaction or Casting of Two Materials,

1. Sosnik B (1948) US Patent No. 2,434,775

Keywords Continuous casting technique • Continuous zone melting technique

3.1 Historical Background

3.1.1 Ice Wormholes

H. Nakajima, Porous Metals with Directional Pores, 13

Fig. 3.2 Growth of ice

3.1.2 Porous Metals with Directional Pores

3.2 High-Pressure Gas Method (PGM)

3.2.1 Mold Casting Technique

Table 3.1 Lotus metals to be fabricated and available gases

3.2.2 Continuous Zone Melting Technique

conductivity, the solidification proceeds with almost constant solidification rate

Continuous Zone Melting Technique

Blower Melting zone

3.2.3 Continuous Casting Technique

The continuous casting technique is extensively used as a mass-production method

where N is the number of atoms in the liquid, k is Boltzmann’s constant, T is the

In addition, it is generally reported that the solidification velocity v affects the

where n is an empirical constant. Thus, the following expression can be finally

distance described in Fig. 3.22 can be plotted as a function of v1, as shown in

3.3 Thermal Decomposition Method (TDM)

As already mentioned above, the high-pressure gas method is a convenient method

3.3.1 Mold Casting Technique through TDM

3.3.1.1 Fabrication of Lotus Copper through TDM

Copper (99.99 % pure) was melted by radio-frequency induction heating in a

Average pore diameter /µm

1,083 K, is impossible, because the atmospheric hydrogen pressure of the TDM

pores in the solid–liquid interface during unidirectional solidification. It is well

3.3.1.2 Fabrication of Lotus Aluminum through TDM

Figure 3.31 shows the microstructure of lotus aluminum on cross sections

Table 3.2 Decomposition reaction and temperature of gas-forming compounds