(Tianyou Fan) Mathematical Theory of Elasticity of (BookFi)

Tianyou Fan
Mathematical Theory of Elasticity of

Quasicrystals and Its Applications
Tianyou Fan
Mathematical Theory of
Elasticity of Quasicrystals
and Its Applications
With 82 figures
Author
Tianyou Fan
Department of Physics, School of Science
Beijing Institue of Technology
Beijing 100081, China
ISBN 978-7-03-025669-0
Science Press Beijing
ISBN 978-3-642-14642-8 e-ISBN 978-3-642-14643-5

Springer Heidelberg Dordrecht London New York
Library of Congress Control Number: 2010931505
© Science Press Beijing and Springer-Verlag Berlin Heidelberg 2011

This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of
this publication or parts thereof is permitted only under the provisions of the German Copyright Law
of September 9, 1965, in its current version, and permission for use must always be obtained from
Springer. Violations are liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, 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.
Cover design: Frido Steinen-Broo, EStudio Calamar, Spain
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

Preface
This monograph is devoted to the development of a mathematical theory of elasticity
of quasicrystals and its applications. Some results on elastodynamics and plasticity
of quasicrystals are also included to document preliminary advances in recent years.
The first quasicrystal was observed in 1982 and reported in November 1984. At
that time several physical and mathematical theories for quasicrystal study already
existed. Soon after the discovery, the theory of elasticity of quasicrystals was put
forward. Based on Landau-Anderson symmetry-breaking, a new elementary exci-
tation - the phason - was introduced in addition to the well known phonon. The
phason concept was suggested in the 1960’s in incommensurate phase theory. Group
theory and discrete geometry e.g. the Penrose tiling provide further explanations
to quasicrystals and their elasticity from the standpoint of algebra and geometry.
The phonon and phason elementary excitations form the basis of the theory of elas-
ticity of this new solid phase. Many theoretical (condensed matter) physicists have
spent a great deal of effort on constructing the fundamental physical framework of
the theory of elasticity of quasicrystals. Applications of group theory and group
representation theory further enhance the physical basis of the description. On the
basis of the physical framework and by extending the methodology of mathematical
physics and classical elasticity, the mathematical theory of elasticity of quasicrystals
has been developed. Recent studies on the elasto-/hydro-dynamics and the plastic-
ity of quasicrystals have made preliminary but significant progress. As regards the
dynamics, there are various viewpoints in the quasicrystal community, which reveal
the unusual characteristics of phason dynamics. Yet the effect of the phason de-
grees of freedom on plastic deformation is not well understood, and the basic plastic
properties of the material are virtually unknown. Because of many unsolved critical
issues, the study of quasicrystals has attracted many researchers. The contex of the
last few chapters in this book is a probe of the fascinating research area.
As this book is focused on the mathematical theory of elasticity of quasicrystals,
it does not include in-depth discussions on the physics of the phason degrees of
freedom and the physical nature of the phason variables. These research subjects
are important to the quasicrystal study.
I sincerely thank the National Natural Science Foundation of China and the
Alexander von Humboldt Foundation of Germany for their support over the years.
I also thank Professors Fanghua Li (Institute of Physics in the Chinese Academy of
vi Preface
Science), Longan Ying (School of Mathematics of Peking University), Dihua Ding

(Department of Physics, Wuhan University), Jiann-Quo Tarn (Cheng Kung Uni-
versity / Zhejiang University), Weiqiu Chen (Zhejiang University), Qingping Sun
(Hong Kong Science and Technology University), U. Messerschmidt (Max Planck
Institut fuer Mikrostruktur Physik, Halle), H.-R. Trebin (Institut fuer Theoretis-
che und Angewandte Physik, Universitaet Stuttgart), and K. Edagawa (University
of Tokyo) for their helpful discussions. Last but not least, I thank my co-workers
and former and present students, especially Professor Xianfang Li (Central South
University), for their help and contributions.
Tianyou Fan
January 2010, Beijing
Contents
Preface
Chapter 1 Crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Periodicity of crystal structure, crystal cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Three-dimensional lattice types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Symmetry and point groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Reciprocal lattice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Appendix of Chapter 1: Some basic concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Chapter 2 Framework of the classical theory of elasticity . . . . . . . . . . . . . 13
2.1 Review on some basic concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Basic assumptions of theory of elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 Displacement and deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Stress analysis and equations of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5 Generalized Hooke’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Elastodynamics, wave motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Chapter 3 Quasicrystal and its properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.1 Discovery of quasicrystal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Structure and symmetry of quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 A brief introduction on physical properties of quasicrystals . . . . . . . . . . . . . 29
3.4 One-, two- and three-dimensional quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5 Two-dimensional quasicrystals and planar quasicrystals . . . . . . . . . . . . . . . . 30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Chapter 4 The physical basis of elasticity of quasicrystals . . . . . . . . . . . . . 35
4.1 Physical basis of elasticity of quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2 Deformation tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Stress tensors and the equations of motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4 Free energy and elastic constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.5 Generalized Hooke’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.6 Boundary conditions and initial conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.7 A brief introduction on relevant material constants of quasicrystals . . . . . 43
4.8 Summary and mathematical solvability of boundary value or initial-
viii Contents
boundary value problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.9 Appendix of Chapter 4: Description on physical basis of elasticity of
quasicrystals based on the Landau density wave theory . . . . . . . . . . . . . . . . 46
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Chapter 5 Elasticity theory of one-dimensional quasicrystals and
simplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.1 Elasticity of hexagonal quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.2 Decomposition of the problem into plane and anti-plane problems . . . . . . 56
5.3 Elasticity of monoclinic quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.4 Elasticity of orthorhombic quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.5 Tetragonal quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.6 The space elasticity of hexagonal quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.7 Other results of elasticity of one-dimensional quasicrystals . . . . . . . . . . . . . 65
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Chapter 6 Elasticity of two-dimensional quasicrystals and
simplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.1 Basic equations of plane elasticity of two-dimensional quasicrystals:
point groups 5m and 10mm in five- and ten-fold symmetries . . . . . . . . . . . 71
6.2 Simplification of the basic equation set: displacement potential
function method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.3 Simplification of the basic equations set: stress potential function
method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.4 Plane elasticity of point group 5, 5 pentagonal and point group 10, 10
decagonal quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.5 Plane elasticity of point group 12mm of dodecagonal quasicrystals . . . . . 85
6.6 Plane elasticity of point group 8mm of octagonal quasicrystals,
displacement potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.7 Stress potential of point group 5, 5 pentagonal and point group 10,
10 decagonal quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
6.8 Stress potential of point group 8mm octagonal quasicrystals . . . . . . . . . . . . 95
6.9 Engineering and mathematical elasticity of quasicrystals . . . . . . . . . . . . . . . 98
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Chapter 7 Application I: Some dislocation and interface problems
and solutions in one- and two-dimensional quasicrystals . . . . 103
7.1 Dislocations in one-dimensional hexagonal quasicrystals . . . . . . . . . . . . . . . 104
7.2 Dislocations in quasicrystals with point groups 5m and 10mm
symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
7.3 Dislocations in quasicrystals with point groups 5, 5̄ five-fold and 10,
Contents ix
10 ten-fold symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

7.4 Dislocations in quasicrystals with eight-fold symmetry . . . . . . . . . . . . . . . . 117
7.5 Dislocations in dodecagonal quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
7.6 Interface between quasicrystal and crystal . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
7.7 Conclusion and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Chapter 8 Application II: Solutions of notch and crack problems
of one-and two-dimensional quasicrystals . . . . . . . . . . . . . . . . . 127
8.1 Crack problem and solution of one-dimensional quasicrystals . . . . . . . . . . 128
8.2 Crack problem in finite-sized one-dimensional quasicrystals . . . . . . . . . . . 134
8.3 Griffith crack problems in point groups 5m and 10mm quasicrystals
based on displacement potential function method . . . . . . . . . . . . . . . . . . . . . 139
8.4 Stress potential function formulation and complex variable function
method for solving notch and crack problems of quasicrystals of point
groups 5, 5̄ and 10, 10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
8.5 Solutions of crack/notch problems of two-dimensional octagonal
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
8.6 Other solutions of crack problems in one-and two-dimensional
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
8.7 Appendix of Chapter 8: Derivation of solution of Section 8.1 . . . . . . . . . . 154
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Chapter 9 Theory of elasticity of three-dimensional quasicrystals
and its applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
9.1 Basic equations of elasticity of icosahedral quasicrystals . . . . . . . . . . . . . . . 160
9.2 Anti-plane elasticity of icosahedral quasicrystals and problem of
interface between quasicrystal and crystal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
9.3 Phonon-phason decoupled plane elasticity of icosahedral
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
9.4 Phonon-phason coupled plane elasticity of icosahedral quasicrystals—
displacement potential formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
9.5 Phonon-phason coupled plane elasticity of icosahedral quasicrystals—
stress potential formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
9.6 A straight dislocation in an icosahedral quasicrystal . . . . . . . . . . . . . . . . . . . 173
9.7 An elliptic notch/Griffith crack in an icosahedral quasicrystal . . . . . . . . . 178
9.8 Elasticity of cubic quasicrystals—the anti-plane and axisymmetric
deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
x Contents
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
Chapter 10 Dynamics of elasticity and defects of quasicrystals . . . . . . 191
10.1 Elastodynamics of quasicrystals followed the Bak’s argument . . . . . . . . 192
10.2 Elastodynamics of anti-plane elasticity for some quasicrystals . . . . . . . . 192
10.3 Moving screw dislocation in anti-plane elasticity . . . . . . . . . . . . . . . . . . . . . 194
10.4 Mode III moving Griffith crack in anti-plane elasticity . . . . . . . . . . . . . . . 197
10.5 Elasto-/hydro-dynamics of quasicrystals and approximate analytic
solution for moving screw dislocation in anti-plane elasticity . . . . . . . . . 199
10.6 Elasto-/hydro-dynamics and solutions of two-dimensional decagonal
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
10.7 Elasto-/hydro-dynamics and applications to fracture dynamics of
icosahedral quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
10.8 Appendix of Chapter 10: The detail of finite difference scheme . . . . . . . 221
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Chapter 11 Complex variable function method for elasticity of
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
11.1 Harmonic and quasi-biharmonic equations in anti-plane elasticity of
one-dimensional quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
11.2 Biharmonic equations in plane elasticity of point group 12mm two-
dimensional quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
11.3 The complex variable function method of quadruple harmonic
equations and applications in two-dimensional quasicrystals . . . . . . . . . . 231
11.4 Complex variable function method for sextuple harmonic equation
and applications to icosahedral quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . 243
11.5 Complex analysis and solution of quadruple quasiharmonic
equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
11.6 Conclusion and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Chapter 12 Variational principle of elasticity of quasicrystals,
numerical analysis and applications . . . . . . . . . . . . . . . . . . . . . . . 257
12.1 Basic relations of plane elasticity of two-dimensional quasicrystals . . . . 258
12.2 Generalized variational principle for static elasticity of quasicrystals . . . . 259
12.3 Finite element method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
12.4 Numerical examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Chapter 13 Some mathematical principles on solutions of elasticity
of quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
13.1 Uniqueness of solution of elasticity of quasicrystals . . . . . . . . . . . . . . . . . . . 273
Contents xi
13.2 Generalized Lax-Milgram theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

13.3 Matrix expression of elasticity of three-dimensional quasicrystals . . . . . 278
13.4 The weak solution of boundary value problem of elasticity of
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
13.5 The uniqueness of weak solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Chapter 14 Nonlinear behaviour of quasicrystals . . . . . . . . . . . . . . . . . . . . . . 289
14.1 Macroscopic behaviour of plastic deformation of quasicrystals . . . . . . . . 290
14.2 Possible scheme of plastic constitutive equations . . . . . . . . . . . . . . . . . . . . . 292
14.3 Nonlinear elasticity and its formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
14.4 Nonlinear solutions based on simple models . . . . . . . . . . . . . . . . . . . . . . . . . . 295
14.5 Nonlinear analysis based on the generalized Eshelby theory . . . . . . . . . . 301
14.6 Nonlinear analysis based on the dislocation model . . . . . . . . . . . . . . . . . . . 305
14.8 Appendix of Chapter 14: Some mathematical details . . . . . . . . . . . . . . . . . 309
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
Chapter 15 Fracture theory of quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
15.1 Linear fracture theory of quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
15.2 Measurement of G IC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
15.3 Nonlinear fracture mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
15.4 Dynamic fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
15.5 Measurement of fracture toughness and relevant mechanical
parameters of quasicrystalline material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Chapter 16 Remarkable conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Major Appendix: On some mathematical materials . . . . . . . . . . . . . . . . . . . . 333
Appendix I Outline of complex variable functions and some additional
calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
A.I.1 Complex functions, analytic functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
A.I.2 Cauchy’s formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
A.I.3 Poles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
A.I.4 Residue theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
A.I.5 Analytic extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
A.I.6 Conformal mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
A.I.7 Additional derivation of solution (8.2-19) . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
A.I.8 Additional derivation of solution (11.3-53) . . . . . . . . . . . . . . . . . . . . . . . . . . 342
xii Contents
A.I.9 Detail of complex analysis of generalized cohesive force model for plane
elasticity of two-dimensional point groups 5m, 10mm and 10, 10
quasicrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
A.I.10 On the calculation of integral (9.2-14) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Appendix II Dual integral equations and some additional calculations . . . . . 348
A.II.1 Dual integral equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
A.II.2 Additional derivation on the solution of dual integral equations
(8.3-8) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
(9.8-8) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
.
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Chapter 1
Crystals
This book discusses mainly elasticity and defects of quasicrystals, however qua-
sicrystals have inherent connection with crystals. This chapter provides the basic
knowledge on crystals which may be beneficial to study quasicrystals and relevant
topics.
1.1 Periodicity of crystal structure, crystal cell

Based on X-ray diffraction patterns, it is known that a crystal consists of particles
(i.e., collections of ions, atoms and molecules) which are arranged regularly in space.
The arrangement is a repetition of the smallest unit, called a unit cell, resulting in the
periodicity of a complete crystal. The frame of the periodic arrangement of centres of
gravity of particles is called a lattice. Thus, the properties of corresponding points
of different cells in a crystal are the same. The positions of these points can be
defined by the radius vectors r and r in the coordinate frame e1 , e2 , e3 , and a, b
and c are three non-mutually co-linear vectors, respectively (the general concept on
vector refering to Chapter 2). Hence we have
r = r + la + mb + nc, (1.1-1)
in which a, b and c are the basic translational vectors describing the particle arrange-
ment in a complete crystal, and l, m and n are arbitrary integers. If the physical
properties are described by function f (r), the above invariance can be expressed
mathematically as
f (r ) = f (r + la + mb + nc) = f (r). (1.1-2)
This is called the translational symmetry or long-range translational order of a crys-
tal, because the symmetry is realized by the operation of translation.
Formula (1.1-1) represents a kind of translational transform, while (1.1-2) shows
that the lattice is invariant under transformation (1.1-1). The collection of all trans-
lational transform remaining lattice invariant constitutes the translational group.
1.2 Three-dimensional lattice types

Cells of lattice may be described by a parallel hexahedron having lengths of its
three sides a, b and c and angles α, β and γ between the sides. According to the
relationship of length of sides and angles there are seven different forms observed for
the cells, which form seven crystal systems given in Table 1.2-1.
T. Fan, Mathematical Theory of Elasticity of Quasicrystals and Its Applications

2 Chapter 1 Crystals
Table 1.2-1 Crystals and the relationship of the length of sides and angles
Crystal system Characters of cell
Triclinic a = b = c, α = β = γ
Monoclinic a = b = c, α = γ = 90◦ = β
Orthorhombic a = b = c, α = β = γ = 90◦
Rhombohedral a = b = c, α = β = γ = 90◦
Tetragonal a = b = c, α = β = γ = 90◦
Hexagonal a = b = c, α = β = 90◦ , γ = 120◦
Cubic a = b = c, α = β = γ = 90◦
Among each crystal system there are certain classes of crystals that are classified
based on the configuration such that whether the face centre or body centre contains
lattice point. For example, the cubic system can be classified as three classes: the
simple cubic, body centre cubic and face centre cubic. According to this classifica-
tion, the seven crystal systems contain 14 different lattice types, called Bravais cells
as shown in Fig. 1.2-1.
Fig. 1.2-1 The 14 Crystal cells of three-dimension

(a) Simple triclinic, (b) Simple monoclinic, (c) Button centre monoclinic,
(d) Simple orthorhombic, (e) Button centre orthorhombic, (f) Body centre orthorhombic,
(g) Face centre orthorhombic, (h) Hexagonal, (i) Rhombohedral,
(j) Simple tetragonal, (k) Body centre tetragonal, (l) Simple cubic,
(m) Body centre cubic, (n) Face centre cubic.
Apart from the above mentioned 14 Bravais cells with three-dimensional lattices,
there are 5 Bravais cells of two-dimensional lattice, we do not go any further.
1.3 Symmetry and point groups 3
1.3 Symmetry and point groups

In Section 1.1 we have discussed the translational symmetry of crystals. Here we
point out that the symmetry reveals invariance of crystals under translational trans-
formation
T = la + mb + nc. (1.3-1)
Formula (1.3-1) is referred to as an operation of symmetry, which is a translational
operation. Apart from this, there are rotation operation and reflection (or mapping)
operation, they belong to so-called point operation. A brief introduction on the
rotation operation and orientational symmetry of crystals is given below.
By rotating about an axis through a lattice, the crystal always returns to the orig-
inal state since the rotational angles are 2π/1, 2π/2, 2π/3, 2π/4 and 2π/6
or integer times of these values. This is the
orientational symmetry or the long-range
orientational order of a crystal. Because
of the constraint of translational symme-
try, the orientational symmetry holds for
n = 1, 2, 3, 4 and 6 only, which is neither
equal to 5 nor more greater than 6 where n
is the denominator of 2π/n (e.g. a molecule
can have five-fold rotation symmetry, but
a crystal cannot have this symmetry be-
cause the cells either overlap or have gaps
between them when n = 5, Fig. 1.3-1 is an Fig.1.3-1 There is no five-fold rotational
example). The fact constitutes the follow- symmetry in crystals
ing fundamental law of crystallography:
Law of symmetry of crystals Under rotation operation, n-fold symmetry
axis is marked by n. Due to the constraint of translational symmetry, there are axes
n = 1, 2, 3, 4 and 6 exist only, neither 5 nor number greater than 6 exists.
In contrast to translational symmetry, rotation is a point symmetry. Other point
symmetries are: plane of symmetry, the corresponding operation is mapping, marked
by m; centre of symmetry, the corresponding operation is inversion, marked by I;
rotation-inversion axis, the corresponding operation is composition of rotation and
inversion, when the inversion after rotation 2π/n, marked by n̄.
For crystals, the point operation consists of eight independent ones only, i.e.,
1, 2, 3, 4, 6, ¯
I = (I), m = 2̄, 4̄, (1.3-2)
which are the basic symmetric elements of point symmetry.

The rotation operation is also denoted by Cn , n = 1, 2, 3, 4, 6.
The mapping operation can also be expressed by σ. The horizontal mapping by
mh and Sh , the vertical one by Sv .
The mapping-rotation is a composite operation, denoted by Sn , which can be

understood as
Sn = Cn σh = σh Cn .
The inversion can be understood as
I = S2 = C2 σh = σh C2 .
Another composite operation—rotation-inversion n̄ is related to Sn , e.g. 1̄ =
S2 = I, 2̄ = S1 = σ, 3̄ = S6 , 4̄ = S4 , 6̄ = S3 .
So that (1.3-2) can be redescribed as
C1 , C2 , C3 , C4 , C6 , I, σ, S4 . (1.3-3)
The collection of each symmetric operation among these eight basic operations con-
stitutes a point group, the collection of their composition forms 32 point groups
listed in Table 1.3-1.
Table 1.3-1 32 Point groups of crystals

Sign Meaning of sign Point group Number
Cn Having n-fold axis C1 , C2 , C3 , C4 , C6 5
I Symmetry central I(i) 1
σ(m) Mapping σ(m) 1
Having n-fold axis and
Cnh C2h , C3h , C4h , C6h 4
horizontal symmetry plane
Having n-fold axis and
Cnν C2ν , C3ν , C4ν , C6ν 4
vertical symmetry plane
Having n-fold axis and n 2-fold axes,
Dn D2 , D3 , D4 , D6 4
they are perpendicular to each other
Meaning of h is the
Dnh D2h , D3h , D4h , D6h 4
same as before
d means in Dn there is a
Dnd symmetry plane dividing D2d , D3d 2
the angle between two 2-fold axes
Having n-fold
Sn S4 , S6 = C3i 2
mapping-rotation axis
Having four 3-fold axes and
T T 1
three 2-fold axes
Meaning of h is the
Th Th 1
same as before
Meaning of d is the
Td Td 1
same as previous
Having three 4-fold axes
O which perpendicular to each other and six O, Oh 2
2-fold axes and four 3-fold axes
Note: T = C3 D2 means the composition between operations C3 and D2 , where suffixes 3
denotes a 3-fold axis.
O = C3 C4 C2 means the composition between operations C3 , C4 and C2 , where 3 repre-
sents a 3-fold axis, 2 represents a 2-fold axis.
1.4 Reciprocal lattice 5
The concept and sign of point groups will be used in the subsequent chapters.
1.4 Reciprocal lattice

The concept of the reciprocal lattice will be concerned in the subsequent chapters,
here is a brief introduction.
Assume that there is relation between the base vectors a1 , a2 and a3 of a lattice
(L) and the base vectors b1 , b2 and b3 for another lattice (LR )

1, i = j,
bi · aj = δij = i, j = 1, 2, 3, (1.4-1)
0, i = j,
such that the lattice with the base vectors {b1 , b2 , b3 } is the reciprocal lattice LR of
crystal lattice L with the base vectors {a1 , a2 , a3 }. Between bi and aj there exist
the relationship
a2 × a3 a3 × a1 a1 × a2
b1 = , b2 = , b3 = , (1.4-2)
Ω Ω Ω
where Ω = a1 · (a2 × a3 ) is the volume of lattice cell.

Denote
Ω ∗ = b1 · (b2 × b3 ),
then
1
Ω∗ = .
Ω
The position of any point in the reciprocal lattice can be expressed by
G = h1 b1 + h2 b2 + h3 b3 (1.4-3)
in which G called the reciprocal vector and
h1 , h2 , h3 = ±1, ±2, · · ·
Points in the lattice can be described by a1 , a2 , a3 as well as by b1 , b2 , b3 .

In a similar fashion the concept of the reciprocal lattice can be extended to
the higher dimensional space, e.g. six-dimensional space, which will be used in the
discussion in Chapter 4.
The brief introduction above provides a background for reading the subsequent
text of the book. Further information on the crystals, the diffraction theory, and
the point groups can be found in books[1, 2]. We will recall the concepts in the
following text.
1.5 Appendix of Chapter 1: Some basic concepts

Some basic concepts will be described in the following chapters, with which most
of physicists are familiar. For the readers of non-physicists, a simple introduction is
provided as follows; the details can be found in the relevant references.
1.5.1 Concept of phonon
In general, the course of crystallography does not contain the contents given in this
section. Because the discussion here is dependent on quasicrystals, especially with
the elasticity of quasicrystals, we have to introduce some of the simplest relevant
arguments.
In 1900, Planck put forward the quantum theory. Soon after Einstein developed
the theory and explained the photo-electric effect, which leads to the photon concept.
Einstein also studied the specific heat cv of crystals by using the Planck quantum
theory. There are some unsatisfactory points in the work of Einstein although his
formula explained the phenomenon of cv = 0 at T = 0, where T denotes the absolute
temperature. To improve Einstein s work, Debye[3] and Born et al. [4,5] applied the
quantum theory to study the specific heat arising from lattice vibration in 1912 and
1913 respectively, and got a great success. The theoretical prediction is in excellent
agreement to the experimental results, at least for the atom crystals.
The propagation of the lattice vibration is called the lattice wave. Under the
long-wavelength approximation, the lattice vibration can be seen as continuum elas-
tic vibration, i.e., the lattice wave can be approximately seen as the continuum
elastic wave. The motion is a mechanical motion, but Debye and Born assumed
that the energy can be quantized based on Planck’s hypothesis. With the elastic
wave approximation and quantization, Debye and Born successfully explained the
specific heat of crystals at low temperature, and the theoretical prediction is con-
sistent with the experimental results in all range of temperature, at least for the
atomic crystals. The quanta of the elastic vibration, or the smallest unit of energy
of the elastic wave is named phonon, because the elastic wave is one of acoustic
waves. Unlike photon, the phonon is not an elementary particle, but in the sense of
quantization, the phonon presents natural similarity to that of photon and other ele-
mentary particles, thus can be named quasi-particle. The concept created by Debye
and Born opened the study on lattice dynamics, an important branch of solid state
physics. Yet according to the view point at present, the Debye and Born theory on
solid belongs to a phenomenological theory, though they used the classical quantum
theory.
Landau[6] further developed the phenomenological theory and put forward the
concept of elementary excitation. According to his concept, photon and phonon etc.
1.5 Appendix of Chapter 1: Some basic concepts 7
belong to elementary excitations. In general one elementary excitation corresponds

to a certain field, e.g. photon corresponds to electro-magnetic wave, phonon cor-
responds to elastic wave, etc. The phonon concept is further extended by Born[5] ;
he pointed out the phonon theory given by Debye corresponds to the vibration as
molecule as a whole body, which is called phonetic frequency vibration modes, or
the phonetic branch of phonon. In this case the physical quantity phonon describ-
ing displacement field deviated from the equilibrium position of particles (atoms,
or ions, or molecules) at lattice is also called as phonon field, or briefly phonon.
Macroscopically it is the displacement vector u of elastic body (if the crystal is re-
garded as an elastic body). But Born emphasized that there is another vibration in
crystals with compound lattice, i.e., the relative vibration between atoms within a
molecule, which is called photonic frequency vibration modes, or photonic branch of
phonon. For this branch the phonon cannot be understood simply as macroscopic
displacement field. But our discussion here is confined to the frame work of contin-
uum medium, with no concern with the photonic branch, so the phonon field is the
macro-displacement field in the consideration.
In many physical systems (classical or quantum systems) the motion presents
the discrete spectrum (the energy spectrum or frequency spectrum, which corre-
sponds to the discrete spectrum of an eigen-value problem of a certain operator
from the mathematical point of view), the lowest energy (frequency) level state is
called ground state, and that beyond the ground state is called excited state. The
so-called elementary excitation induces a transfer from the ground state to the state
with the smallest non-zero energy (or frequency). Strictly speaking, it should be
named lowest energy (or frequency) elementary excitation.
The solid state physics was intensively developed in 1960∼70’s, then evolved into
the condensed matter physics. The condensed matter physics is not only extending
the scope of solid state physics by considering the liquid state and micro-powder
structure, but also developing the basic concepts and principles. Modern condensed
matter physics is established as a result of the construction of its paradigm, in which
the symmetry-breaking is in a central place, which was contributed by Landau[6] and
Anderson[7] and other scientists.
Considering the importance of the concept and principle of symmetry-breaking
in development of elasticity of quasicrystals, we give a brief discussion here.
It is well known that for a system with a constant volume, the equilibrium state
thermodynamically requires the free energy of the system
F = E − TS (1.5-1)
be minimum, in which E is the internal energy, S is the entropy and T is the absolute
temperature.
Landau proposed the so-called second order phase transition theory by introduc-
ing a macroscopic order parameter η to describe (order-disorder) phase transition,
i.e., assuming that the free energy can be expanded as a power series of η,
F (η, T ) = F0 (T ) + A(T )η 2 + B(T )η 4 + · · · , (1.5-2)
in which, according to the requirement of the stability condition of phase transition

(i.e., the variational condition δF = 0 or ∂F/∂η = 0), the coefficients of odd terms
should be taken to zero, and B(T ) > 0. At high temperature the system is in
a disorder state, so A(T ) > 0, too; as the temperature reduces A(T ) changes its
sign; at the critic temperature TC there exists A(TC ) = 0. The simplest choice in
satisfying these conditions is
A(T ) = α(T − TC ), B(T ) = B(TC ), (1.5-3)
in which α is a constant. Without concerning concrete micro-mechanism, the Landau

theory has the merits of simplicity and generality; it can be used to many systems
and has achieved successes, especially for the study of superconductivity, liquid
crystals,high energy physics (To the author’s understanding, the quasicrystals study
is another area that has been achieved following the line of the symmetry-breaking
principle). Applying the above principle to periodic crystals we have[7]
1
F = α(|G|)(T − TC (G))η 2 + higher-order terms, (1.5-4)
2
where the constant α is related to the reciprocal vector G (the concepts on the re-
ciprocal vector and the reciprocal lattice, refer to Section 1.4). Further, Anderson[7]
showed that for crystals if the density of periodic crystals can be expressed by
Fourier series (the expansion exists due to the periodicity of the structure in three-
dimensional lattice or the reciprocal lattice)

ρ(r) = ρG exp{iG · r} = |ρG | exp{−iΦG + iG · r}, (1.5-5)
G∈LR G∈LR
where G is a reciprocal vector, and LR the reciprocal lattice in three-dimensional

space, ρG is a complex number,
ρG = |ρG | e−iΦG (1.5-6)
with the amplitude |ρG | and the phase angle ΦG , due to ρ(r) being real, |ρG | = |ρ−G |
and ΦG = −ΦG , the order parameter is
η = |ρG | . (1.5-7)
1.5 Appendix of Chapter 1: Some basic concepts 9
Anderson pointed out that for crystals the phase angle ΦG contains the phonon
u for crystal, i.e.,
ΦG = G · u, (1.5-8)
in which both G and u are in three-dimensional physical space. If we consider

only the phonetic branch of the phonon, then u can be understood as the displace-
ment field. So the displacement field in periodic crystals can be understood as the
phonon field from the Landau symmetry-breaking hypothesis, though it possesses
an intuitive physical meaning under the approximation of the long-wavelength (refer
to Chapter 2). The description based on the symmetry-breaking, physical quantity
u is connected to the reciprocal vector G and reciprocal lattice LR of crystals, so
it presents more profound insight (a result of symmetry-breaking) than that of the
intuitive description of the displacement field u, though the explanation is still based
on phenomenological theory (because the Landau theory is a phenomenological the-
ory), rather than on the first-principles.
The concept of phonon is originated from Debye[3] and Born[4,5] , which describes
the mechanical vibration of lattice mass points (atoms, or ions , or molecules) de-
viating from their equilibrium position. The propagation of the vibration leads to
the lattice wave, and the motion can be quantized, the quanta is the phonon. This
is an elementary excitation in condensed matter. The symmetry-breaking leads to
appearance of new order phase (e.g. crystals), new order parameter (e.g. wave am-
plitude of mass density wave), the new elementary excitation (e.g. phonon), and
the new conservation law (e.g. the crystal symmetry law given in Section 1.3). The
above description on the phonon from the symmetry-breaking point of view helps
us in understanding in-depth the physical nature of the phonon.
Following this scheme, some elementary excitations (e.g. phason) temporarily
without intuition meaning can also be explained by the Landau theory, and one can
further find out their physical meaning from the point of view of symmetry rather
than from intuition, because for some complex phenomena, the simple intuition
cannot give a complete and correct explanation.
We recall that some of the elementary excitations are related to broken symmetry
or symmetry-breaking, e.g. a liquid behaves arbitrary translational symmetry and
arbitrary orientational symmetry, then the periodicity of crystals (i.e., the lattice)
breaks the translational symmetry and the orientational symmetry of liquid, the
phonon is elementary excitation, resulting from the symmetry-breaking. In this
case we may say the quasicrystal is a result of symmetry-breaking of crystal, which
will be discussed in Chapter 4.
If description on phason concept is unnecessary, this section can be omitted. The
reader can skip this section if not interested in.
1.5.2 Incommensurate crystals
In this book we do not discuss incommensurate phases, but quasicrystals are related
with so-called incommensurate structure, we mention it in brief.
Since 1960’s, incommensurate crystals have been studied by many physicists,
see e.g. [8]. The incommensurate phase means that it is plus an additional in-
commensurate modulate at the basic lattice, in which the modulated ones may be
displacements or compounds of atoms or arrangement of spin etc. As an example,
if a modulated displacement λ is plus to a lattice with period a, and if λ/a is an
rational number, the crystal becomes a super-structure with long period (which is
the integer times of a); and if λ/a is an irrational number, the crystal becomes an
incommensurate structure. In this case along the modulate direction the periodicity
is lost. The modulation can be one-dimensional, e.g. Na2 CO3 , NaNO2 , etc, or
two-dimensional, e.g. TaSe2 , quartz, etc, or three-dimensional, e.g. Fe1−x O, etc.
In the incommensurate phases, the modulation is only a “perturbation” of another
period of the basic lattice, the diffraction pattern of the basic lattice holds, i.e., the
crystallography symmetry holds, so one calls the structure being the incommensu-
rate crystals. It is noticed that there are new degrees of freedom in the phases, be
named the phasons. Here the phason modes present long-wavelength propagation
similar to that of the phonon modes. In Chapter 4 we shall study that the phason
modes in the quasicrystals present quite different nature, i.e., the motion of atoms
exhibits discontinuous jumps rather than the long-wavelength propagation. In addi-
tion, in the quasicrystals there is non-crystallographic orientational symmetry, which
is essentially different from that of incommensurate crystals.
1.5.3 Aphomus(glassy structure)
The crystals are solid with long-range order due to regular atom arrangement. In
contrast, there is a solid without order, but it has short-range order in scale within
the atom size. This material is the aphomus, as a branch of the condensed matter
physics.
References
[1] Kittel C.Introduction to the Solid State Physics. New York: John Wiley & Sons,
Inc, 1976
[2] Wybourne B G. Classical Group Theory for Physicists. New York: John Wiley &
Sons, Inc, 1974
[3] Debye P. Die Eigentuemlichkeit der spezifischen Waermen bei tiefen Temperaturen.
Arch de Genéve, 1912, 33(4): 256–258
References 11
[4] Born M, von Kármán Th. Zur Theorie der spezifischen Waermen, Physikalische
Zeitschrift. 1913, 14(1): 15–19
[5] Born M, Huang K. Dynamic Theory of Crystal Lattices, Oxford: Clarendon Press,
1954
[6] Landau L D, Lifshitz M E. Theoretical Physics V: Statistical Physics. 3rd ed. Oxford:
Pergamon Press, 1980
[7] Anderson P W. Basic Notions of Condensed Matter Physics. Menlo Park: Benjamin-
Cummings, 1984
[8] Blinc B and Lavanyuk AP. Imcommensurate Phases in Dielectrics, I. II. Amsterdan:
North Holand, 1986
Chapter 2
Framework of the classical theory of
elasticity
As the knowledge of crystals is a benefit for understanding quasicrystals, it is worth-

while having a concise review of the classical theory of elasticity before learning the
elasticity of quaicrystals. Here is a brief description to the theory. The detailed
material for this theory can be found in many monographs and textbooks, e.g. Lan-
dau and Lifshitz[1] . Though the discussion here is limited within the framework of
continuum medium mechanics, there are still connections to physical nature of the
elasticity of crystals reflected by phonon concept (discussed in Section 1.5). The
readers are advised to refer to the relevant chapters and sections of monographs of
Born and Huang[2] and Anderson[3] which would help us understanding the phonon
concept so the phason concept and elasticity of quasicrystals, which will be presented
in the following chapters. The practice shows that it would be hard to understand
phason concept and the phason elasticity if we limited our knowledge only within
the classical continuum medium and complete intuition.
For simplicity, the tensor algebra will be used in the text.
2.1 Review on some basic concepts

2.1.1 Vector
A quantity with both magnitude and direction is named vector, denoted by a, and
a = |a| represents its magnitude. The scalar product is defined as a · b = ab cos(a, b)
of two vectors a and b. The vector product is defined as a × b = n ab sin (a, b),
in which n is the unit vector perpendicular to both a and b, so |n| = 1. A more
general definition on vector is given later.
2.1.2 Coordinate frame
To describe the vector and the tensor, it is convenient to introduce the coordinate
frame. We will consider the orthogonal frame. Assume that e1 , e2 and e3 are three
unit vectors and mutually perpendicular, i.e., e1 · e2 = 0, e2 · e3 = 0, e3 · e1 = 0

14 Chapter 2 Framework of the classical theory of elasticity
and e3 = e1 × e2 , e2 = e3 × e1 , e1 = e2 × e3 , then they are the base vectors of an

orthogonal coordinate frame, and often called base vectors briefly.
In the orthogonal coordinate frame e1 , e2 , e3 , any vector a can be expressed by
a = a1 e1 + a2 e2 + a3 e3
or (2.1-1)
a = (a1 , a2 , a3 ).
2.1.3 Coordinate transformation

Considering another orthogonal frame e1 , e2 , e3 which can be expressed in terms of
e1 , e2 , e3 . From (2.1-1), we have
e1 = c11 e1 + c12 e2 + c13 e3 ,

e2 = c21 e1 + c22 e2 + c23 e3 , (2.1-2)
e3 = c31 e1 + c32 e2 + c33 e3 ,
where c11 , c12 , · · · , c33 are some scalar constants. The relation (2.1-2) is the coordi-
nate transformation, which can be expressed in the matrix form
⎡ ⎤ ⎡ ⎤
e1 e1
⎢ ⎥ ⎢ ⎥
⎣ e2 ⎦ = [C] ⎣ e2 ⎦ , (2.1-3)
e3 e3
where ⎡ ⎤
c11 c12 c13
⎢ c c22 c23 ⎥
[C] = ⎣ 21 ⎦,
c31 c32 c33
which is an orthogonal matrix, consequently
[C]T = [C]−1 , (2.1-4)
here notation “T” marks transpose operation, and “−1” the inversion operation. It
follows, ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
e1 e1 e1
⎢ e ⎥ T ⎢ e ⎥ −1 ⎢ e ⎥
⎣ 2 ⎦ = [C] ⎣ 2 ⎦ = [C] ⎣ 2 ⎦ . (2.1-5)
e3 e3 e3
Based on (2.1-1), in the frame e1 , e2 , e3 ,
a = a1 e1 + a2 e2 + a3 e3 (2.1-6)
Substituting (2.1-5) into (2.1-1) yields
a = (c11 a1 + c12 a2 + c13 a3 )e1 + (c21 a1 + c22 a2 + c23 a3 )e2

2.1 Review on some basic concepts 15
+ (c31 a1 + c32 a2 + c33 a3 )e3 . (2.1-7)
It follows that by the comparison between (2.1-6) and (2.1-7)
a1 = c11 a1 + c12 a2 + c13 a3 ,

a2 = c21 a1 + c22 a2 + c23 a3 , (2.1-8)
a3 = c31 a1 + c32 a2 + c33 a3 ,
or in the matrix expression, i.e.,

⎡ ⎤ ⎡ ⎤
a1 a1
⎣ a2 ⎦ = [C] ⎣ a2 ⎦ . (2.1-8 )
a3 a3
Whatever (2.1-8) or (2.1-8 ), there is

3

ai = cij aj = cij aj . (2.1-9)
j=1
The summation sign in the right-handside of (2.1-9) is omitted, when the repeated
indexes in cij aj represent summing. Henceforth, the summation convention will be
used throughout.
A set of number (a1 , a2 , a3 ) satisfying the relation (2.1-9) under the linear trans-
formation (2.1-2) is a vector regardless its physical meaning. This is an algebraic
definition of vector; it is more general than saying that the vector has both magni-
tude and direction.
2.1.4 Tensor
Let us define nine numbers in the orthogonal frame e1 , e2 , e3 as A:
⎡ ⎤
A11 A12 A13
A = ⎣ A21 A22 A23 ⎦ , (2.1-10)
A31 A32 A33
in which the components satisfy the relation

3

Akl = cki clj Aij = cki clj Aij (2.1-11)
i,j=1
under the linear transformation, then A is a tensor of rank 2, where cij are given
by (2.1-3), and the summation sign is omitted in the right-handside of (2.1-11). It
is evident that the concept of tensor is an extension of that of vector. According to
the definition Aij represents a tensor where i=1, 2, 3, j=1, 2, 3. It is understood
that it represents a component with the indexes i and j of the tensor.
2.1.5 Algebraic operation of tensor
1) Unit tensor

0, i = j,
I = δij = (2.1-12)
1, i = j,
which is named the Kronecker delta conventionally.
2) Transpose of tensor
⎡ ⎤
A11 A21 A31
AT = ⎣ A12 A22 A32 ⎦ . (2.1-13)
A13 A23 A33
3) Algebraic sum of a tensors
A ± B = Aij ± Bij . (2.1-14)
4) Product of a tensor by a scalar
mA = mAij . (2.1-15)
5) Product of tensors
AB = Aij Bkl . (2.1-16)
Other operations about tensors will be provided where required.
2.2 Basic assumptions of theory of elasticity

The theory of linear elasticity is a branch of continuum mechanics, it follows the
basic assumptions thereof, whereas there are
1) Continuity
In the theory one assumes that the medium is filled the full space that it occu-
pies, this means the medium is continuous. Connected with this, the field variables
concerning the medium are continuous and differentiable functions of coordinates.
2) Homogeneity
Physical constants describing the medium are independent of coordinates, so the
medium is homogeneous.
3) Small deformation
Assume that displacements ui small and ∂ui /∂xj much less than unit. Due
to small deformation, the boundary conditions are expressed with reference to the
boundaries before deformation though those boundaries have undergone deforma-
tion. This makes the problems to be linearized and simplifies the solution procedure.
2.3 Displacement and deformation 17
2.3 Displacement and deformation

That elastic body exhibits deformation is connected to the relative displacement
between points in it. So we first look for the displacement field.
Consider a region R in an elastic body, refer to Fig. 2.3-1, it turns into another
region R after deformation. The point O with radius vector r, before deformation,
becomes point O with radius vector r after deformation, and u is the displacement
vector of point O during the deformation process (see Fig.2.3-1) i.e.,
Fig. 2.3-1 Displacement of a point in an elastic body
r = r + u (2.3-1)
or
u = r − r = xi − xi . (2.3-1 )
In Fig. 2.3-1, e1 , e2 , e3 depicts any orthogonal coordinate system, especially we
use the rectilinear coordinate system (x1 , x2 , x3 ) or (x, y, z). Assume that O1 in R
is a point near the point O, the radius vector connecting them is dr = dxi . The
point O1 becomes point O1 in R after deformation. The radius vector connecting
points O1 and point O1 is dr = dxi = dxi + dui . The displacement of point O1 is
u , thus
u = u + du, (2.3-2)
i.e.,
dui = ui − ui (2.3-3)
and ∂ui
dui = dxj . (2.3-4)
∂xj
(2.3-4) expresses the Taylor expansion at point O and takes the first order term
only. Under the small deformation assumption, this reaches a very high accuracy.
It denotes
∂ui
= εij + ωij , (2.3-5)
∂xj
in which
1 ∂ui ∂uj
εij = + , (2.3-6)
2 ∂xj ∂xi

1 ∂ui ∂uj
ωij = − , (2.3-7)
2 ∂xj ∂xi
here εij is a symmetric tensor
εij = εji (2.3-8)
and called the strain tensor, while ωij an asymmetric tensor, which has only three
independent components

1 ∂uy ∂uz
Ωx = ωyz = − ,
2 ∂z ∂y

1 ∂uz ∂ux
Ωy = ωzx = − , (2.3-9)
2 ∂x ∂z

1 ∂ux ∂uy
Ωz = ωxy = − .
2 ∂y ∂x
The physical meaning of εij describes the volume and shape change of a cell, and
that of ωij the rigid-body rotation, which is independent of deformation. Henceforth,
it suffices to consider εij .
The components ε11 , ε22 and ε33 (if denote x = x1 , y = x2 , z = x3 , then we have
εxx , εyy and εzz ) represent normal strains, describing the volume change of a cell,
while ε32 = ε23 , ε13 = ε31 and ε12 = ε21 (or εyz = εzy , εzx = εxz and εxy = εyx )
represent shear strains, describing the shape change of a cell.
2.4 Stress analysis and equations of motion

The stress in the internal forces per unit area due to deformation denoted it by σij ,
it is zero if there is no deformation. When the body is in static equilibrium, the
equilibrium equations can be derived from the law of momentum conservation as
follows:
∂σij
+ fj = 0, (2.4-1)
∂xi
which holds for any infinitesimal volume element of the body, σij represents the
components of the stress tensor as mentioned above, and suffix j the acting direction
of the component, i the direction of outward normal vector of the surface element
on which the stress component exerted, fi the body force density vector. Among all,
the components of σij , σxx , σyy and σzz are normal to the surface elements which
they exerted, and σyz , σzy , σzx , σxz , σxy and σyx are along the tangent directions
2.5 Generalized Hooke’s law 19
of the surface elements, the former are called normal stresses, and the latter shear
stresses.
According to the angular momentum conservation, one finds that
σij = σji , (2.4-2)
this means the stress tensor is a symmetric tensor, and (2.4-2) is named the shear
stress mutual equal law.
The external surface forces density (tractions) Ti subjected to the surface of a
body should be balanced with the internal stresses, this leads to
σij nj = Ti , (2.4-3)
where nj is the unit vector along the outward normal to the surface element. People
also call Ti area force density.
Equation (2.4-3) describes the stress boundary conditions which play a very
important role for elasticity.
2.5 Generalized Hooke’s law

Between the stresses σij and the strains εij , there exists a certain relationship de-
pended upon the material behaviour of the body. Hereafter we consider only the
linear elastic behaviour of materials, and the state without initial stresses. In the
case, the classical experimental law——Hooke’s law can be generalized as
∂U
σij = = Cijkl εkl , (2.5-1)
∂εij
in which U denotes the free energy, or the strain energy density, i.e.,
1
U =F = Cijkl εij εkl (2.5-2)
2
and Cijkl is the elastic constant tensor, consisting of 81 components. Due to the
symmetry of σij and εij , each of them has 6 independent components only, such
that the independent components of Cijkl reduce to 36. Formula (2.5-2) shows that
U is a quadratic function of εij , considering the symmetry of εij , then we have
Cijkl = Cklij , (2.5-3)
so the independent components 36 reduce to 21.

The relation (2.5-1) with 21 independent elastic constants is named generalized
Hooke’s law.
The generalized Hooke’s law describes anisotropic elastic bodies including crys-
tals. Stress and strain tensors can also be expressed by corresponding vectors with
6 independent elements, then can be denoted by the corresponding elastic constants

matrix [Bijkl ],
⎡ ⎤ ⎡ ⎤⎡ ⎤
σ11 B1111 B1122 B1133 B1123 B1131 B1112 ε11
⎢ σ22 ⎥ ⎢ B2222 B2233 B2223 B2231 B2212 ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ε22 ⎥
⎢ σ33 ⎥ ⎢ B3333 B3323 B3331 B3312 ⎥ ⎢ ε33 ⎥
⎥ ⎢
⎢ ⎥ ⎢ ⎥.
⎢ σ23 ⎥ = ⎢ B2323 B2331 B2312 ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ε23 ⎥
⎣ σ31 ⎦ ⎣ (symmetry) B3131 B3112 ⎦ ⎣ ε31 ⎦
σ12 B1212 ε12
(2.5-4)
Applying formula (2.5-4) to crystals, between the elements Cijkl (or Bijkl ) there are
some relations by considering certain symmetry of the crystals, so that, the resulting
number of the elastic constants for certain individual crystal systems may be less
than 21. In the following we give a brief discussion on the argument.
1) Triclinic system (classes 1 or C1 and Ci )
The triclinic symmetry does not add any restrictions to the components of ten-
sor Cijkl (or Bijkl in (2.5-4)), however, appropriate choice of the coordinate system
enables us to reduce the number of non-zero independent elastic constants. Because
the orientation of the coordinate system is determined by three rotation angles, this
provides three conditions to restrict some components in Cijkl (or Bijkl in (2.5-4));
for example. one can take three of them to be zero, such that, the triclinic crystal
system has 18 components of elastic moduli.
2) Monoclinic system (classes Cs , C2 and C2h )
In the class Cs , there is plane of symmetry, we take it as x3 = 0(z = 0) in
coordinate system {e1 , e2 , e3 }. Making a coordinate transformation with this plane
of symmetry one can obtain a new coordinate system {e1 , e2 , e3 } . Between these
two coordinate systems, there are relations
e1 = e1 , e2 = e2 , e3 = −e3 . (2.5-5)

This operation is the reflection or mapping. In addition, we know that between σij

in e1 , e2 , e3 and σij in e1 , e2 , e3 , there are (refer to Section 2.1)

σkl = αkj αli σji , (2.5-6)
in which αij are the coefficients of linear transformation, i.e.,
ei = αij ej . (2.5-7)
Under the transformation (2.5-5), there are
α11 = 1, α22 = 1, α33 = −1, others = 0. (2.5-8)

2.5 Generalized Hooke’s law 21
Therefore, under the transformation, for Cijkl in (2.5-1) (or Bijkl in (2.5-4)) whose
suffixes containing 3 with an odd number of times (1 or 3) will change sign, while
the others will remain invariant. Considering the symmetry of the crystal, however,
the physical properties including Cijkl (or Bijkl in (2.5-4)) should remain unchanged
under symmetric operation (including the reflection). So it is obvious that all com-
ponents with an odd number of suffixes 3 must vanish, i.e.,
B1123 = B1131 = B2223 = B2231 = B3323 = B3331 = B2312 = B3112 = 0. (2.5-9)
Consequently, there are only 13 independent elastic constants.

A similar discussion can be done for the classes C2 and C2h .
3) Orthorhombic system (classes C2v , D2 and D2h )
This crystal system has a macroscopic correspondence, i.e., the orthotropic ma-
terials, in which there exist two planes of symmetry perpendicular to each other.
Let us take x3 = 0 and x1 = 0 as the planes. If on reflection in plane x3 = 0, it
is just the case for monoclinic system mentioned above. Subsequently consider the
mapping in plane x1 = 0, between the new and old coordinate systems, there is the
relation such as ⎡ ⎤ ⎡ ⎤⎡ ⎤
e1 −1 0 0 e1
⎣ e2 ⎦ = ⎣ 0 1 0 ⎦ ⎣ e2 ⎦ .
e3 0 0 1 e3
By a similar description to that for monoclinic system, one finds that
B1112 = B2212 = B3312 = B2331 = 0. (2.5-10)
Collecting to (2.5-9), the system contains 9 independent elastic constants.

4) Tetragonal system (classes C4v , D2d , D4 and D4h )
This crystal system has 4 axes of symmetry. Similar to previous discussion, that
independent elastic moduli are
B1111 , B3333 , B1122 , B1212 , B1133 , B1313 .
The total number of them is six.

5) Rhombohedral system (classes C3v , 3 or C3 , D3 , D3d and S6 )
In this system there is a third-order axis of symmetry (or three-fold symmetric
axis). We can take axis of symmetry as the axis e3 , after a lengthy derivation that
six independent elastic constants are as follows
B3333 , Bξηξη , Bξξηη , Bξη33 , Bξ3η3 , Bξξξ3
with
ξ = x1 + ix2 , η = x1 − ix2 .
The moduli can also be written in conventional version as

B3333 , B1212 , B1122 , B1323 , B1113 .
6) Hexagonal system (class C6 )
The crystal system has a macroscopic correspondence—the transversely isotropic
material, whose elasticity presents fundamental importance to elasticity of one- and
two-dimensional quasicrystals.
There is a sixth-order axis of symmetry (or say six-fold symmetric axis) in the sys-
tem. Take this axis as x3 -axis, and use the coordinate substitution ξ = x1 + ix2 , η =
x1 − ix2 . In a rotation with angle 2π/6 about the x3 -axis, the new coordinates ξ and
η are transformed by ξ → ξei2π/6 , η → ηe−i2π/6 . Then one can see that only those
components Cijkl (or Bijkl in (2.5-4)) do not vanish which have the same number
of suffixes ξ and η. These are
B3333 , Bξηξη , Bξξηη , Bξη33 , Bξ3ηξ ,
or in conventional expressions,
C1111 = C2222 , C3333 , C2323 = C3131 , C1122 , C1133 = C2233 , C1212 ,
in which 2C1212 = C1111 −C1122 , so the number of independent elastic constants is 5.
7) Cubic system (clases T, O)
For this system there are 3 four-fold symmetric axes, in which there is tetragonal
symmetry. If taking the four-fold symmetric axis of the tetragonal symmetry in the
x3 -direction, the number of independent components of Cijkl (or Bijkl in (2.5-4)) are
B1111 , B1122 , B1212 .
8) Isotropic body
In this case there are two elastic moduli, e.g. the Young’s modulus and Poisson’s
ratio, E, ν, respectively, or the Lamé constants
νE E
λ= , μ= , (2.5-11)
(1 + ν)(1 − 2ν) 2(1 + ν)
or the bulk modulus of compression and shear modulus
E E
K= ,μ = = G.
3(1 − 2ν) 2(1 + ν)
In this case the generalized Hooke’s law presents very simple form, i.e.,
σij = 2μεij + λεkk δij , (2.5-12)
where εkk = ε11 + ε22 + ε33 = εxx + εyy + εzz , δij is the unit tensor. An equivalent
form of (2.5-12) is
1+ν ν
εij = σij − σkk δij , (2.5-13)
E E
in which σkk = σ11 + σ22 + σ33 = σxx + σyy + σzz .
2.6 Elastodynamics, wave motion 23
2.6 Elastodynamics, wave motion

When the inertia effect is considered in (2.4-1), then it becomes
∂σij ∂ 2 ui
+ fi = ρ 2 , (2.6-1)
∂xj ∂t
where ρ is the mass density of the material.
Considering an isotropic medium and omitting body forces, from (2.6-1), (2.3-6)
and (2.5-12) the equations of wave motion are obtained as
∂ 2 ui ∂ 2 uj ∂ 2 uj
(c21 − c22 ) + c22 = , (2.6-2)
∂xi ∂xj ∂x2i ∂t2
where c1 and c2 defined by
12 12
λ + 2μ μ
c1 = , c2 = , (2.6-3)
ρ ρ
which are speeds of elastic longitudinal and transverse waves respectively. If put
u = ∇φ + ∇ × ψ, (2.6-4)
then (2.6-2) can be reduced to

1 ∂2φ 1 ∂2ψ
∇2 φ = , ∇2 ψ = , (2.6-5)
c21 ∂t2 c22 ∂t2
∂2 ∂2
where φ is the scalar potential, and ψ the vector potential, and ∇2 = 2
+ 2+
∂x ∂y
∂2
, (2.6-5) are typical wave equations of mathematical physics. To solve the prob-
∂z 2
lem apart from the boundary conditions one needs initial conditions, i.e.,
ui (xi , 0) = ui0 (xi ),
xi ∈ Ω .
u̇i (xi , 0) = u̇i0 (xi ),
2.7 Summary
The classical theory of elasticity is concluded to solve the following initial-boundary
value problem
1 ∂ui ∂uj
εij = + ,
2 ∂xj ∂xi
∂σij ∂ 2 uj
= ρ 2 − fj , t > 0, xi ∈ Ω ,
∂xi ∂t
σij = Cijkl εkl ,
ui (xj , 0) = ui0 (xj )
xj ∈ Ω (Initial condition)
u̇i (xj , 0) = u̇i0 (xj )
σij nj = Ti , t > 0, xj ∈ Γt
(Boundary condition)
ui = ūi , t > 0, xj ∈ Γu
where ui0 (xj ), u̇i0 (xj ), Ti and ūi are known functions, Ω denotes the region of ma-
terials we studied, Γt and Γu are parts of boundary Γ on which the tractions and
∂ 2 uj
displacements are prescribed respectively and Γ = Γt + Γu . If = 0, the prob-
∂t2
lem reduces to a static problem as pure boundary value problem, there are no initial
conditions at all.
References
[1] Landau L D, Lifshitz E M. Theoretical Physics V: Theory of Elasticity. Oxford:
Pergamon Press, 1986
[2] Born M, Huang K. Dynamic Theory of Crystal Lattices. Oxford: Clarendon Press,
1954
[3] Anderson P W. Basic Notions of Condensed Matter Physics. Menlo Park: Benjamin-
Cummings, 1984
Chapter 3
Quasicrystal and its properties
The observation of quasicrystal in 1982 was a dramatic event and is an important

discovery in the development history of physics. This has given rise to a great deal of
attention of researchers to the unusual properties of the new structure of solid. Due
to the thermodynamic stability of the most quasicrystals produced to date, which
become a novel material. It is necessary to give a brief discussion on the structural
characters and physical properties of quasicrystals for understanding their elasticity
and defects which are the main focus of the following chapters.
3.1 Discovery of quasicrystal∗

The first observation of quasicrystal was done in April 1982 while D. Shechtman
was working in the Bureau of Standards in USA as a guest scholar. He observed
from the electronic microscopy that a rapid cooled Al-Mn alloy exhibits five-fold
orientational symmetry by the diffraction patterns with bright diffraction spots and
sharp Bragg reflections, as shown in Fig. 3.1-1. Because the five-fold orientational
symmetry is of contradiction with the basic law of symmetry of crystals (refer to
Chapter 1), the result could not be understood within the first two year since the
discovery. I. Blech in Israel a colleague of Shechtman, gave him a powerful support,
explaining that it might be an icosahedral glass. They drafted a paper concerning
the experimental results and submitted to a journal, but was rejected. Then they
submitted it to another journal, which could not be published too. J. W. Cahn, the
hosting scientist in the Bureau of Standards recommended streamlining the paper;
leaving out details of the model and experiment, and limiting it solely the experi-
mental findings. After consulting with D. Gratias, a mathematical crystallographer
at the Centre National de la Recherche Scientifique in France, the group submitted
an abbreviated article to Physical Review Letters (PRL) in October 1984, more than
two years after Shechtman’s initial experiment. The article was published several
weeks later. This is the Ref. [1].
After four weeks, Levine and Steinhardt[2] published their work by PRL to intro-
∗ This is referred to article PRL Top 10: #8 of APS.

26 Chapter 3 Quasicrystal and its properties
duce ordered structure with quasiperiodicity and called the novel alloy as “quasicrys-
tal” formally, in which their theoretical (computed) diffraction pattern in excellent
agreement with that of the experimental observation. Soon after, other groups, for
example, Ye et al[3] , Zhang et al[4] etc. also found similar structure of five-fold
symmetry and icosahedral quasicrystal in Ni-V and Ni-Ti alloys.
The icosahedral quasicrystals are one of three-dimensional quasicrystals, in which
the atomic arrangement is quasiperiodic in three directions. Another three-dimensio-
nal quasicrystal is cubic quasicrystal observed by Fung et al[5] later.
Fig. 3.1-1 The patterns of diffraction of icosahedral quasicrystal

(a) The five-fold symmetry, (b) the stereographic structure of the quasicrystal.
Successively two-dimensional quasicrystals were observed. Here, the atomic ar-

rangement is quasiperiodic along two directions and periodic along the third direc-
tion, which is just the directions of five-, eight-, ten- and twelve-fold symmetrical
axes of the two-dimensional quasicrystals observed to date. Such that one finds
four kinds of two-dimensional quasicrystals with five-, eight-, ten- and twelve-fold
rotation symmetries which are also called pentagonal, octagonal, decagonal and do-
decagonal quasicrystals, respectively (see Bendersky[6] , Chatopadihyay et al[7] , Fung
et al[8] , Urban et al[9] , Wang et al[10] , Li et al[11] ).
There is another class of quasicrystals, the one-dimensional quasicrystals, in
which the atomic arrangement is quasiperiodic along one axis, and periodic along
the plane perpendicular to the axis (see Merdin et al[12] , Hu et al[13] ,Fung et al[14] ,
Teranchdi et al[15] , Chen et al[16] , and Yang et al[17] ).
The discovery of this novel matter with long-range order and non-crystallographic
symmetry changes the traditional concept of classifying solids into two classes: crys-
tal and noncrystal, and gives a strong impact on traditional crystallography, and
brings new physical idea into the matter structure and symmetry.
3.2 Structure and symmetry of quasicrystals 27
The unusual structure of quasicrystals leads to a series of properties different

from those of crystals, which has created a great deal of attention by researchers in
a range of fields, such as physics, crystallography, chemistry etc.
A decade before the discovery, Penrose[18] put forward a mathematical model
of quasicrystals, afterward called Penrose tiling, in which a tiling without an over-
lap or a gap in two different rhombohedra can result in the quasiperiodic struc-
ture. After the discovery of quasicrystals, Penrose tiling has become a geome-
try tool of the new solid phase. As an example, a Penrose tiling for describing
a two-dimensional five-fold symmetry is shown in Fig. 3.1-2, in which the local
structure of the tiling is similar, this is called
local isomorphism (LI). The quasiperiodic
symmetry and local isomorphism of the Pen-
rose tiling present significance for describing
the new solid phase—quasicrystals.The dis-
covery of quasicrystal physically gave rise to
the development of the Penrose geometry and
the relevant discrete geometry, and the de-
velopment of ergodic theory, group theory,
Fourier analysis etc.
Quasicrystals with thermodynamical sta-
bility are becoming a new class of functional
and structural materials, which have many po-
tential engineering applications. They gen- Fig. 3.1-2 Penrose tiling of
erate the study on physical and mechanical two-dimensional quasicrystal with
properties of the material. The elasticity five-fold symmetry
and defects are of important topics among
the mechanical behaviour of quasicrystals. These provide many new challenges as
well as opportunities to the continuum mechanics.
3.2 Structure and symmetry of quasicrystals

The quasicrystals are different from the periodic crystals. With a certain symmetry,
it is one of kind of aperiodic crystals. The unusual characters of the quasicrystals are
originated from their special atomic constitution. The character of this structure
is explored by diffraction patterns. Just through these diffraction patterns peo-
ple discovered differences between quasicrystals and crystals, so did the quasicry-
stals.
Similar to other aperiodic crystals, quasiperiodicity induces new degrees of free-
dom, which can be explained as follows. In crystallography and solid state physics,
the Miller indices (h, k, l)are often used to describe the structure of crystals. These
indices can explain the spectra of diffraction patterns of all crystals. In Chapter 1 we
mentioned that the number of base vectors for crystal N is identical to the number
of the dimensions d of the crystal, i.e., N = d. However, because quasicrystals have
quasiperiodic symmetry (including both or either quasiperiodic translational and ori-
entational symmetries disallowed by the rule of crystallography), the Miller indices
cannot be used and instead we need to employ six indices (n1 , n2 , n3 , n4 , n5 , n6 ).
This feature implies, it is necessary to introduce higher dimensional (four-, or five-
or six-dimensional) spaces to characterize the symmetry of quasicrystals. This idea
is identical to that of group theory, i.e., the quasiperiodic structure is periodic in
higher dimensional space (four-, or five- or six-dimensional space). Quasicrystals in
the real three-dimensional space (physical space) may be seen as a projection of a
periodic lattice in the higher dimensional space (mathematical space). The projec-
tion of the periodic lattice at four-, five-, and six-dimensional space to the physical
space generates one-, two- and three-dimensional quasicrystals, respectively. The
six-dimensional space is denoted by E 6 , which consists of two sub-spaces, one is the
physical space, called the parallel space and denoted by E3 , another is the comple-
3
mentary space also called the vertical space and denoted by E⊥ , so that
E 6 = E3 ⊕ E⊥
3
, (3.2-1)
where ⊕ denotes the direct sum.

For one-, two- and three-dimensional quasicrystals, the number of the base vec-
tors is N = 4, 5, 6; the number of the realistic dimension of the material (in physical
space) is d = 3, so N > d, this is different from that of crystals.
The method of group theory is the most appropriate method to describe the
symmetry of quasicrystals.
The one-dimensional quasicrystals have 31 point groups, consisting of 6 quasicrys-
tal systems and 10 Laue classes, in which the all the point groups are crystallographic
point groups listed in Table 3.2-1 .
Table 3.2-1 The systems, Laue classes and point groups of

one-dimensional quasicrystals
System Laue class Point group
Triclinic 1 1, 1
Monoclinic 2 2, mh , 2/mh
3 2h , m, 2h /m
Orthorhombic 4 2h 2h 2, mm2, 2h mmh , mmmh
Tetragonal 5 4, 4, 4/mh
6 42h 2h , 4mm, 4/mh , 4/mh mm
Rhombohedral 7 3, 3
8 32h , 3m, 3m
Hexagonal 9 6, 6, 6/mh
10 62h 2h , 6mm, 6m2h , 6/mh mm
3.3 A brief introduction on physical properties of quasicrystals 29
The two-dimensional quasicrystals have 57 point groups, in which 31 are crystal-

lographic point groups listed in Table 3.2-1, and other 26 are non-crystallographic
point groups listed in Table 3.2-2.
The three-dimensional quasicrystals have 60 point groups. They are: (1) 32 crys-
tallographic point groups and 28 non-crystallographic point groups, i.e., icosahedral
point groups (235, m3̄5̄) and 26 point groups with 5-,8-,10-and12-fold symmetries
(5, 5,52, 5m, 5m, and N ,N ,N/m, N 22, N mm, N m2 , N/mmm, N = 8, 10, 12), the
latter have been listed in Table 3.2-2.
Table 3.2-2 The systems, Laue classess and point groups of

non-crystallography of two-dimensional quasicrystals
System Laue class Point groups
Pentagonal 11 5, 5
12 5m, 52, 5m
Decagonal 13 10, 10, 10/m
14 10mm, 1022, 10m2, 10/mmm
Octagonal 15 8, 8, 8/m
16 8mm, 822, 8m2, 8/mmm
Dodecagonal 17 12, 12, 12/m
18 12mm, 1222, 12m2, 12/mmm
3.3 A brief introduction on physical properties of quasicrys-

tals
The unusual structure of quasicrystals leads to some new physical properties of the
material.
The mechanical behaviour of quasicrystals, especially, the distinct features of
elasticity to those of crystals in particular, have aroused a great deal of research
interest. These will be discussed in detail starting from Chapter 4. We do not talk
about them any further here.
The thermal properties of quasicrystals are a field that attracts attention[19∼27] .
The thermal conductivity of quasicrystals is lower than that of the conventional
metals.
Among the profound properties of quasicrystals being studied, the first is the
structural and elastic properties, the second is the properties of electricity of the ma-
terial. The conductivity of electricity of quasicrystals is lower. The Hall effect[28∼32]
was well studied. The absolute number of the Hall coefficient RH is two orders of
magnitude greater than that of the conventional metals. In addition, the pressure-
resistance properties of quasicrystals have also been discussed, see, e.g. [33].
The light conductivity rate of quasicrystals is quite different from that of the
conventional metals, the singularity is of interest, see [34∼36]. Recently the study of
quasicrystals photonic crystal study becomes a focus[37∼39] , with a trend of further

development[40∼42] .
The electronic structure of quasicrystals and relevant topics have also been con-
cerned after the discovery[43∼45] . Due to lack of periodicity, the Bloch theorem and
Brillouin zone concept cannot be used. By some simple models and through nu-
merical computation, one can obtain results on the electronic energy spectra. Some
results on wave functions obtained exhibit behaviour neither the in extending state
nor in the localization state. For some quasicrystalline materials, e.g. Al-Cu-Li,
Al-Fe etc, there are pseudogaps when energy is over the Fremi energy.
Because the present book discusses elasticity of quasicrystals only, the other
properties will not be dealt with. The above descrption provides a very simple
introduction only.
3.4 One-, two- and three-dimensional quasicrystals

It is needed to recall the concept on one-, two- and three-dimensional quasicrys-
tals. The one-dimensional quasicrystals are the ones in which the atom arrangement
is quasiperiodic in one direction, and periodic in other two directions. The two-
dimensional quasicrystals belong to ones in which the atom arragement is quasiperi-
odic in two directions and periodic in the other one. The three-dimensional qua-
sicrystals behave in such a manner that the arrangement presents quasiperiodicity
in all three directions.
There exist 200 quasicrystals observed to date, in which about 100 are icosa-
hedral and about 70 decagonal quasicrystals, respectively, so these two classes of
quasicrystal systems present major importance in the material.
3.5 Two-dimensional quasicrystals and planar quasicrystals

The two-dimensional quasicrystals and planar quasicrystlas are different concepts.
The two-dimensional quasicrystals have been introduced in the previous section,
which represent a three-dimensional structure with two-dimensional quasiperiodic
planes stacked along the third direction; in this direction the atom arrangement
is periodic. While the planar quasicrystals belong to a two-dimensional structure
within the plane the atom arrangement is quasiperiodic, and there is no third di-
mension.
References
[1] Shechtman D, Blech I, Gratias D et al. Metallic phase with long-range orientational
order and no translational symmetry. Phys Rev Lett, 1984, 53(20): 1951–1953
References 31
[2] Levine D, Steinhardt P J.Quasicrystals: A new class of ordered structure. Phys Rev
Lett, 1984, 53(26): 2477–2450
[3] Ye H Q, Wang D, Kuo K H. Five-fold symmetry in real and reciprocal space. Ultra-
microssopy, 1985, 16(2): 273–277
[4] Zhang Z, Ye H Q, Kuo K H. A new icosahedral phase with m35 symmetry. Phil Mag
A, 1985, 52(6): L49–L52
[5] Feng Y C, Lu G, Witers R I. An incommensurate structure with cubic point group
symmetry in rapidly solidified V-Vi-Si alloy. J Phys Condens Matter, 1989, 1(23):
3695–3700
[6] Bendersky L. Quasicrystal with one-dimensional translational symmetry and a ten-
fold rotation axis. Phys Rev Lett, 1985, 55(14): 1461–1463
[7] Chattopadhyay K, Lele S, Thangarai N et al. Vacancy ordered phases and one-
dimensional quasiperiodicity. Acta Metall, 1987, 35(3): 727–733
[8] Fung K K, Yang C Y, Zhou Y Q et al. Icosahedrally related decagonal quasicrystal
in rapidly cooled Al-14-at.%-Fe alloy. Phys Rev Lett, 1986, 56(19): 2060–2063
[9] Urban K, Mayer J, Rapp M et al. Studies on aperiodic crystals in Al-Mn and Al-V
alloys by means of transmission electron microscopy. Journal de Physique Colloque,
1986, 47C(3): 465–475
[10] Wang N, Chen H, Kuo K H. Two-dimensional quasicrystal with eightfold rotational
symmetry. Phys Rev Lett, 1987, 59(9): 1010–1013
[11] Li X Z, Guo K H. Decagonal quasicrystals with different periodicities along the
tenfold axis in rapidly solidified Al-Ni alloys. Phil Mag Lett, 1988, 58(3): 167–171
[12] Merdin R, Bajema K, Clarke R et al. Quasiperiodic GaAs-AlAs heterostructures.
Phys Rev Lett, 1985, 55(17): 1768–1770
[13] Hu A, Tien C, Li X et al. X-ray diffraction pattern of quasiperiodic (Fibonacci)
Nb-Cu superlattices. Phys Lett A, 1986, 119(6): 313–314
[14] Feng D, Hu A, Chen K J et al. Research on quasiperiodic superlattice. Mater Sci
Forum, 1987, 22(24): 489–498
[15] Terauchi H, Noda Y, Kamigami K et al. X-Ray diffraction patterns of configurational
Fibonacci lattices. J Phys Jpn, 1988, 57(7): 2416–2424
[16] Chen K J, Mao G M, Fend D et al. Quasiperiodic a-Si: H/a-SiNx: H multilayer
structures. J Non-cryst Solids, 1987, 97(1): 341–344
[17] Yang W G, Wang R H, Gui J. Some new stable one-dimensional quasicrystals in
Al65 Cu20 Fe10 Mn5 alloy. Phil Mag Lett, 1996, 74(5): 357–366
[18] He L X, Li X Z, Zhang Z et al. One-dimensional quasicrystal in rapid solidified
alloys. Phys Rev Lett, 1988, 61(9): 1116–1118
[19] Biham O, Mukamel D, Shtrikman. Symmetry and stability of icosahedral and other
quasicrystalline phases. Phys Rev Lett, 1986, 56(20): 2191–2194
[20] Scheater R J, Bendersky L A. Introduction to Quasicrystals. Jaric M, Boston V.

MA: Academic Press, 1988
[21] Widom M, Deng D P, Henleg C L. Transfer-matrix analysis of a two-dimensional
quasicrystal. Phys Rev Lett, 1989, 63(3): 310–313
[22] Yang W G, Ding D H, Wang R H, et al. Thermodynamics of equilibrium properties
of quasicrystals. Z Phys. B, 100(3), 447–454
[23] Hu C Z, Yang W G, Wang R H et al. Quasicrystal symmetry and physical properties.
Progress in Phys, 1997, 17(4): 345–367
[24] Fan T Y. A study on specific heat of one-dimensional hexagonal quasicrystal. J Phys:
Condens Matter, 1999, 11(45): L513–L517
[25] Fan T Y, Mai Y W. Partition function and state equation of point group 12mm
two-dimensional dodecagonal quasicrystals. Euro Phys J B, 2003, 31(2): 17–21
[26] Li C Y, Liu Y Y. Phason-strain influence on low-temperature specific heat of the
decagonal Al-Ni-Co quasicrystal. Chin Phys Lett, 2001, 18(4): 570–573
[27] Li C Y, Liu Y Y. Low-temperature lattice excitation of icosahedral Al-Mn-Pd qua-
sicrystals. Phys Rev B, 2001, 63(6): 064203
[28] Biggs B D, Li Y, Poon S J. Electronic properties of icosahedral, approximant, and
amorphous phases of an Al-Cu-Fe alloy. Phys Rev B, 1991, 43(10): 8747–8750
[29] Lindqvist P, Berger C, Klein T et al. Role of Fe and sign reversal of the Hall coefficient
in quasicrystalline Al-Cu-Fe. Phys Rev B, 1993, 48(1): 630–633
[30] Klein T, Gozlen A, Berger C et al. Anomalous transport properties in pure Al-Cu-Fe
icosahedral phases of high structural quality. Europhys Lett, 1990, 13(2): 129–134
[31] Pierce F S, Guo Q, Poon S J. Enhanced insulatorlike electron transport behavior
of thermally tuned quasicrystalline states of Al-Pd-Re alloys. Phys Rev Lett, 1994,
73(16): 2220–2223
[32] Wang A J, Zhou X, Hu C Z et al. Properties of nonlinear elasticity of quasicrystals
with five-fold symmetry. Wuhan University Journal Nat Sci, 2005, 51(5): 536–566
[33] Zhou X, Hu C Z, Gong P et al. Piezoresistance properties of quasicrystals. J Phys:
Condens Matter, 2004, 16(30): 5419–5425
[34] Homes C C,Timusk T, Wu X et al. Optical conductivity of the stable icosahedral
quasicrystal Al63.5 Cu24.5 Fe12 . Phy Rev Lett, 1991, 67(19): 2694–2696
[35] Burkov S E. Optical conductivity of icosahedral quasicrystals. J Phys: Condens
Matter, 1992, 4(47): 9447–9458
[36] BasovD N, Timusk T, Barakat F et al. Anisotropic optical conductivity of decagonal
quasicrystals. Phys Rev Lett, 1994, 72(12): 1937–1940
[37] Villa D, Enoch S, Tayeb G et al. Band gap formation and multiple scattering in
photonic quasicrystals with a Penrose-type lattice. Phys Rev Lett, 2005, 94(18):
183903
References 33
[38] Mayou D. Generalized Drude formula for the optical conductivity of quasicrystals.
Phys Rev Lett, 2000, 85(6), 1290–1293
[39] Notomi M, Suzuk H, Tamamura T et al. Document lasing action due to the two-
dimensional quasiperiodicity of photonic quasicrystals with a Penrose lattice. Phys
Rev Lett, 2004, 92(12): 123906
[40] Feng Z F, Zhang X D, Wang Y Q. Negative refraction and imaging using 12-fold-
symmetry quasicrystals. Phys Rev Lett, 2005, 94(24): 247402
[41] Lifshitz R, Arie A, Bahabad A. Photonic quasicrystals for nonlinear optical frequency
conversion. Phys Rev Lett, 2005, 95(13): 13390
[42] Marn W, Megens M. Experimental measurement of the photonic properties of icosa-
hedral quasicrystals. Nature, 2005, 436(7053): 993–996
[43] Kohmoto M, Sutherland B. Electronic states on a Penrose lattice. Phys Rev Lett,
1986, 56(25): 2740–2743
[44] Sutherland B. Simple system with quasiperiodic dynamics: a spin in a magnetic field.
Phys Rev B, 1986, 34(8): 5208–5211
[45] Fujiwara T, Yokokawa T. Universal pseudogap at Fermi energy in quasicrystals. Phys
Rev Lett, 1991, 66(3): 333–336
Chapter 4
The physical basis of elasticity of
quasicrystals
The physical background on elasticity of quasicrystals is quite different from that of

the classical elasticity, the discussion about this provides a basis of the subsequent
contents of the book.
4.1 Physical basis of elasticity of quasicrystals

Quasicrystal has become a type of functional and structural materials, having po-
tential engineering applications. As a material, quasicrystal is deformable under
applied forces, thermal loads and certain internal effects. The deformation of crys-
tals has been discussed in Chapter 2. Questions arise as to what the characteristics
in the deformation process of the quasicrystal are. How to describe mathematically
the behaviour of the quasicrystal deformation? To answer these questions, it is nec-
essary to consider the physical background of elasticity of quasicrystals. The study
in this regard was conducted soon after discovery of the new solid phase.
Because the quasicrystal is a new structure of solid, theoretical physicists have
proposed various descriptions of its elasticity. The majority agrees that the Landau
density wave theory (Refs[1-25]) is the physical basis of elasticity of the quasicrystals.
We shall introduce the theory in Appendix of this chapter (Section 4.9). Essentially,
the description suggested that there are two displacement fields u and w in a qua-
sicrystal: the former is similar to that in crystals, named as the phonon field, its
macro-mechanical behaviour has been discussed in Chapter 2; the latter is a displace-
ment field, named as the phason field. The total displacement field in a quasicrystal
is expressed by
ū = u ⊕ u⊥ = u ⊕ w, (4.1-1)
where ⊕ denotes the direct sum; u is in the physical space, or the parallel space E3 ;
3
w is in the complement space, or the perpendicular space E⊥ , which is an internal
space
Furthermore, the two displacement vectors depend on the coordinate vector r

36 Chapter 4 The physical basis of elasticity of quasicrystals
in the physical space,

u = u(r ), w = w(r ) (4.1-2)
For simplicity, the superscript of r will be removed hereafter. From the angle of
mathematical theory of elasticity of quasicrystals and its technological applications
the formulas (4.1-1) and (4.1-2) are sufficient for comprehending the following con-
tents of the book. If readers are interested in more of the physical background on
the phonon and phason fields in quasicrystals, we suggest that they could read the
Appendix of this chapter (i.e., the Section 4.9).
With basic formulas (4.1-1) and (4.1-2) and some fundamental conservation laws
well known in physics, the macroscopic basis of the continuous medium model of
elasticity of quasicrystals can be set up, in some extent, the discussion is an extension
to that in Chapter 2, which will be done in the following sections.
4.2 Deformation tensors

In Chapter 2 we introduced that the deformation of phonon field lies in the relative
displacement (i.e., the rigid-body translation and rotation do not result in deforma-
tion), which is expressed by
du = u − u.
If we set up an orthogonal coordinate system (x1 , x2 , x3 ) or (x, y, z), then u =

(u1 , u2 , u3 ) = (ux , uy , uz ) and
∂ui
dui = dxj (4.2-1)
∂xj
in which ∂ui /∂xj has the meaning of the gradient of vector u. It may also be denoted
as ⎡ ⎤
∂ux ∂ux ∂ux
⎢ ∂x ∂y ∂z ⎥
⎢ ⎥
⎢ ⎥
∂ui ⎢ ∂uy ∂uy ∂uy ⎥
∇u = =⎢ ⎢ ⎥ (4.2-2)
∂xj ⎢ ∂x ∂y ∂z ⎥ ⎥
⎢ ⎥
⎣ ∂uz ∂uz ∂uz ⎦
∂x ∂y ∂z
and ⎡ ⎤
∂ux ∂ux ∂ux
⎢ ∂x ∂y ∂z ⎥
⎢ ⎥
⎢ ⎥
⎢ ∂uy ∂uy ∂uy ⎥
⎢ ⎥
⎢ ∂x ∂y ∂z ⎥
⎢ ⎥
⎢ ⎥
⎣ ∂uz ∂uz ∂uz ⎦
∂x ∂y ∂z
4.2 Deformation tensors 37
⎡ ⎤
∂ux 1 ∂ux ∂uy 1 ∂ux ∂uz
⎢ + + ⎥
⎢ ∂x 2 ∂y ∂x 2 ∂z ∂x
⎥
⎢ ⎥
⎢ 1 ∂u ∂uy ∂uy 1 ∂uy ∂uz ⎥
⎢ x ⎥
=⎢
⎢ 2 ∂y
+
∂x ∂y 2 ∂z
+
∂y
⎥
⎥
⎢ ⎥
⎢ ⎥
⎢ 1 ∂ux ∂uz 1 ∂uy ∂uz ∂uz ⎥
⎣ + + ⎦
2 ∂z ∂x 2 ∂z ∂y ∂z
⎡ ⎤
1 ∂uy ∂ux 1 ∂uz ∂ux
⎢ 0 − − − − ⎥
⎢ 2 ∂x ∂y 2 ∂x ∂z ⎥
⎢ ⎥
⎢ 1 ∂u ∂uy

1 ∂uz ∂uy
⎥
⎢ − x
− − − ⎥
+⎢
⎢ 2 ∂y ∂x
0
2 ∂y ∂z
⎥
⎥
⎢ ⎥
⎢ ⎥
⎢ 1 ∂ux ∂uz 1 ∂uy ∂uz ⎥
⎣ − − − − 0 ⎦
2 ∂z ∂x 2 ∂z ∂y

1 ∂ui ∂uj 1 ∂uj ∂ui
= + − − = εij + ωij ,
2 ∂xj ∂xi 2 ∂xi ∂xj

1 ∂ui ∂uj
εij = + , (4.2-3)
2 ∂xj ∂xi

1 ∂ui ∂uj
ωij = − , (4.2-4)
2 ∂xj ∂xi
which means the gradient of the phonon vector u can be decomposed into two
parts εij and ωij , in which εij has contribution to the deformation energy, and ωij
represents a kind of rigid-body rotations. We consider only εij , which is the phonon
deformation tensor, or the strain tensor, it is a symmetric tensor: εij = εji .
Similarly for the phason field we have
∂wi
dwi = dxj (4.2-5)
∂xj
and ⎡ ⎤
∂wx ∂wx ∂wx
⎢ ∂x ∂y ∂z ⎥
⎢ ⎥
⎢ ⎥
∂wi ⎢ ∂wy ∂wy ∂wy ⎥
∇w = =⎢
⎢ ∂x
⎥, (4.2-6)
∂xj ⎢ ∂y ∂z ⎥
⎥
⎢ ⎥
⎣ ∂wz ∂wz ∂wz ⎦
∂x ∂y ∂z
though it can be decomposed into symmetric and asymmetric parts, all components
∂wi
of contribute to the deformation of the quasicrystals, the phason deformation
∂xj
tensor or the phason strain tensor is defined by

∂wi
wij = , (4.2-7)
∂xj
which is asymmetric tensor wij = wji , and describes the local rearrangement of
atoms in a cell.
The difference between εij and wij given by (4.2-3) and (4.2-7) are originated
from the physical properties of the phonon modes and the phason modes. This
can also be explained by the group theory, in that, they follow different irreducible
representations for some symmetry transformations for most quasicrystal systems,
except the three-dimensional cubic quasicrystal system. The detail about this is
omitted here.
For the three-dimensional cubic quasicrystals, the phason modes exibit the same
behaviour as that of the phonon modes, which will be particularly discussed in
Chapter 9.
4.3 Stress tensors and the equations of motion

The gradient of the displacement field w figures out the local rearrangement of
atoms in a cell in the quasicrystals. External forces are needed to drive the atoms
through barriers when they make the local rearrangement in a cell. Such that, there
is another kind of body forces and tractions apart from the conventional body forces
f and tractions T for deformed quasicrystals, named as the generalized body forces
(density) g and the generalized tractions ( the generalized area forces density) h.
At first, we consider the static case.
Denoting the stress tensor corresponding to εij by σij , the phonon stress tensor,
and that to wij by Hij , the phason stress tensor, we have the following equilibrium
equations ⎧
⎪ ∂σij + f = 0,
⎪
⎪
⎨ ∂xj i
x, y, z ∈ Ω (4.3-1)
⎪
⎪
⎪ ∂Hij + gi = 0,
⎩
∂xj
based on the momentum conservation law.
Apply the angular momentum conservation law to the phonon field

d
r × ρu̇dΩ = r × f dΩ + r × T dΓ (4.3-2)
dt Ω Ω Ω
by using the Gauss theorem, it follows that
σij = σji , (4.3-3)

4.3 Stress tensors and the equations of motion 39
this indicates that the phonon stress tensor is symmetric,

Since r and w(g, h) transform under different representations of the point
groups, more precisely that, the former transform like a vector, but latter does
not, the product representation: r × w, r × g and r × h do not contain any
vector representations. This implies that for the phason field there is no equation
analogous to (4.3-2), from which it follows that, generally
Hij = Hji (4.3-4)
except the case for three-dimensional cubic quasicrystals.

In the dynamic case, the deformation process is rather complicated; there are
different arguments. Lubensky et al[5] claimed that the phonon modes and the
phason modes are different based on their role in six-dimensional hydrodynamics,
the phonons are wave propagation while the phasons are diffusive with very large
diffusive time. Physically the phason modes represent a relative motion of the con-
stituent density waves. Dolinsek et al[22,23] further developed the idea by bringing in
the concept of atom flip or atom hopping for the phason dynamics. But according
to Bak[1,2] , the phason describes particular structural disorders or structure fluctu-
ations in quasicrystals, and it can be formulated based on a six-dimensional space
description. Since there are six continuous symmetries, there exist six hydrodynamic
vibration modes. In the following we give a brief introduction on elastodynamics
based on the Bak’s argument as well as argument of Lubensky et al.
Ding et al[26] and Hu et al[16] derived that
⎧
⎪ ∂σij ∂ 2 ui
⎪
⎪ + f = ρ ,
⎨ ∂xj i
∂t2
(x, y, z) ∈ Ω , t > 0 (4.3-5)
⎪
⎪ ∂Hij ∂ 2 wi
⎪
⎩ + gi = ρ 2 ,
∂xj ∂t
based on the law of momentum conservation. We believe that the derivation is

carried out by following the Bak’s argument, in which ρ is the mass density of the
quasicrystals.
If following the argument of Lubensky et al, people cannot obtain (4.3-5), instead
we have ⎧
⎪
⎪ ∂σij ∂ 2 ui
⎪
⎨ ∂xj + f i = ρ ,
∂t2
(x, y, z) ∈ Ω , t > 0, (4.3-6)
⎪
⎪ ∂H ∂w
⎪
⎩ ij
+ gi = κ
i
,
∂xj ∂t
in which κ = 1/Γw , Γw is the kinetic coefficient of the phason field. The equations
are given by Fan et al[27] , which are identical to those given by Lubensky et al[5] for
the linear case. Lubensky et al gave their formulation based on the hydrodynamics
principle, so equation (4.3-6) may be seen as elasto-/hydro-dynamic equation of
quasicrystals. In particular, the second equation of (4.3-6) presents the dissipation
feature of motion of the phason degrees in the dynamic process, it is irreversible
thermodynamically.
The elastodynamics of quasicrystals poses a great challenge, see e.g. [32, 33], we
discuss it only in a quite narrow scope (only in Chapter 10), the influence of those
distinguishing arguments in the circle of the quasicrystal study is limited to the
range in Chapter 10. The content therein does not affect the results and conclusions
presented in other chapters.
4.4 Free energy and elastic constants

Consider the free energy or the strain energy density of a quasicrystal F (εij , wij )
whose general expression is difficult to obtain. We take a Taylor expansion in the
neighbourhood of εij = 0 and wij = 0, and remain up to the second order term,
then

1 ∂2F 1 ∂2F
F (εij , wij ) = εij εkl + εij wkl
2 ∂εij ∂wkl 0 2 ∂εij ∂wkl 0

1 ∂2F 1 ∂2F
+ wij wkl + wij εkl
2 ∂wij ∂wkl 0 2 ∂wij ∂εkl 0
1 1 1 1
= Cijkl εij εkl + Rijkl εij wkl + Kijkl wij wkl + Rijkl wij εkl
2 2 2 2
= Fu + Fw + Fuw , (4.4-1)
where Fu , Fw and Fuw denote the parts contributed by the phonon, the phason and
the phonon-phason coupling, respectively, and

∂2F
Cijkl = (4.4-2)
∂εij ∂εkl 0
is the phonon elastic constant tensor, discussed in Chapter 2 already, and
Cijkl = Cklij = Cjikl = Cijlk , (4.4-3)
the tensor can be expressed by the symmetric matrix
[C]9×9 .
In (4.4-1) another elastic constant tensor is

∂2F
Kijkl = , (4.4-4)
∂wij ∂wkl 0
4.4 Free energy and elastic constants 41
in which the suffixes j, l belong to space E3 and i, k to space E⊥

3
, and
Kijkl = Kklij . (4.4-5)
All components of Kijkl can also be expressed by the symmetric matrix
[K]9×9 .
In addition,
∂2F
Rijkl = , (4.4-6)
∂εij ∂wkl 0

∂2F
Rijkl = (4.4-7)
∂wij ∂εkl 0
are the elastic constants of the phonon-phason coupling. To be noted that, the
suffixes i, j, l belong to space E3 and k belongs to space E⊥
3
, and

Rijkl = Rjikl , Rijkl = Rklij , Rklij = Rijkl , (4.4-8)
but

Rijkl = Rklij , Rijkl = Rklij , (4.4-9)
of which all components can be expressed by the symmetric matrices
[R]9×9 , [R ]9×9
and
[R]T = [R ] , (4.4-10)
where T denotes the transpose operator. The composition of four matrices [C] , [K] ,
[R] and [R ] forms a 18 × 18 matrix
⎡ ⎤
[C] [R] [C] [R]
[C, K, R] = =⎣ ⎦. (4.4-11)
T
[R ] [K] [R] [K]
If the strain tensor is expressed by a row vector with 18 elements, i.e.,

ε11 , ε22 , ε33 , ε23 , ε31 , ε12 , ε32 , ε13 , ε21 ,
[εij , wij ] = , (4.4-12)
w11 , w22 , w33 , w23 , w31 , w12 , w32 , w13 , w21
the transpose of which denotes the array vector, then the free energy (or strain
energy density) may be expressed by
⎡ ⎤
1 [C] [R]
F = [εij , wij ] ⎣ ⎦ [εij , wij ]T , (4.4-13)
2 T
[R] [K]
which is identical to that given by (4.4-1)

4.5 Generalized Hooke’s law

To apply theory of elasticity of quasicrytals, one must determine the displacement
field and stress field, this requires that we need to set up relationship between the
strains and the stresses, the relations are the generalized Hooke’s law of quasicrys-
talline material. From the free energy (4.4-1) or (4.4-13), we have
∂F
σij = = Cijkl εkl + Rijkl wkl ,
∂εij
(4.5-1)
∂F
Hij = = Kijkl wkl + Rklij εkl ,
∂wij
or in the matrix form

σij [C] [R] εij
= T , (4.5-2)
Hij [R] [K] wij
where
σij
= [σij , Hij ]T ,
Hij
(4.5-3)
εij T
= [εij , wij ] .
wij
4.6 Boundary conditions and initial conditions

The above general formulas give a description of the basic law of elasticity of qua-
sicrystals, and provide a key to solve those problems in application for theoretical
research and engineering practice, the formulas hold in any interior of the body, i.e.,
(x, y, z) ∈ Ω , where (x, y, z) denote the coordinates of any point of the interior, and
Ω denotes the body. The formulas are concluded as some partial differential equa-
tions. To solve them, it is necessary to know the situation of the field variables at
the boundary Γ of Ω , without appropriate information at the boundary, the solution
has no any physical meaning. According to practical case the boundary Γ consists
of two parts Γt and Γu , i.e., Γ = Γt + Γu , at Γt the tractions are given and at Γu
the displacements are prescribed. For the former case,

σij nj = Ti ,
(x, y, z) ∈ Γt , (4.6-1)
Hij nj = hi ,
where nj represents the unit outward normal vector at any point at Γ , Ti and hi the
traction and generalized traction vectors, which are given functions at the boundary.
Formula (4.6-1) is called the stress boundary conditions. And for the latter case,

ui = ūi ,
(x, y, z) ∈ Γu , (4.6-2)
wi = w̄i ,
4.7 A brief introduction on relevant material constants of quasicrystals 43
where ūi and w̄i are known functions at the boundary. Formula (4.6-2) is named
the displacement boundary conditions.
If Γ = Γt (i.e., Γu = 0), the problem for solving equations (4.2-3), (4.2-7),
(4.3-1) and (4.4-14) under boundary conditions (4.6-1) is a traction boundary value
problem. While Γ = Γu (i.e., Γt = 0) , the problem for solving equations (4.2-3),
(4.2-7), (4.3-1) and (4.4-14) under boundary conditions (4.6.2) is a displacement
boundary value problem.
If Γ = Γu + Γt and both Γt = 0, Γu = 0, the problem for solving equations
(4.2-3), (4.2-7), (4.3-1) and (4.4-14) under boundary conditions (4.6-1) and (4.6.2)
is a mixed boundary value problem.
For the dynamic problem, if taking the wave equations (4.3-5) together with equa-
tion (4.2-3), (4.2-7) and (4.4-14), besides boundary conditions (4.6-1) and
(4.6-2), we must prescribe initial value conditions:

ui (x, y, z, 0) = ui0 (x, y, z), u̇i (x, y, z, 0) = u̇i0 (x, y, z),

(x, y, z) ∈ Ω , (4.6-3)
wi (x, y, z, 0) = wi0 (x, y, z), ẇi (x, y, z, 0) = ẇi0 (x, y, z),
in which ui0 (x, y, z, 0), u̇i0 (x, y, z, 0),wi0 (x, y, z, 0) and ẇi0 (x, y, z, 0) are given func-
∂ui
tions and u̇i = etc. In this case the problem is called initial-boundary value
∂t
problem.
If taking the wave equations coupling diffusion equations (4.3-6) together with
(4.2-3) and (4.4-14), the initial value conditions are

ui (x, y, z, 0) = ui0 (x, y, z), u̇i (x, y, z, 0) = u̇i0 (x, y, z),

(x, y, z) ∈ Ω , (4.6-4)
wi (x, y, z, 0) = wi0 (x, y, z),
this is also an initial-boundary value problem, but different from the previous one.
4.7 A brief introduction on relevant material constants of

quasicrystals
In above discussion we find that the quasicrystals posses different nature with those
of the crystals. Connection with this, the material constants of the solid should be
different from those of the crystals and other conventional structural materials, in
which the constants appearing in the above basic equations are interesting at least.
We here give a brief introduction to help readers in conceptional point of view. The
detailed introduction will be given in Chapters 6, 9 and 10.
The measurement of material constants of quasicrystals is difficult, but the ex-
perimental technique is in progress, especially in recent years. Due to the majority
of icosahedral and decagonal quasicrystals in the material, the measured data are
mainly for these two kinds of the solid phase.
For icosahedral quasicrystals, the independent nonzero components of the phonon

elastic constants Cij are only λ and μ, the phason elastic contants Kij only K1 and
K2 , and the phonon-phason coupling elastic constants Rij only R. For the most
important icosahedral Al-Pd-Mn quasicrystal, the measured data including the mass
density and kinetic coefficient of phason are[28,29] :
3
ρ = 5.1g/cm , λ = 74.9, μ = 72.4GPa, K1 = 72, K2 = −37MPa, R ≈ 0.01μ,
Γw = 4.8 × 10−10 cm3 · μs/g = 4.8 × 10−19 m3 · s/kg

and two-dimensional quasicrystals, the independent nonzero components of the phonon
elastic constants are only C11 , C33 , C44 , C12 , C13 , and C66 = (C11 − C12 )/2, the pha-
son ones only K1 , K2 , K3 , and the phonon-phason coupling ones only R1 , R2 . For
decagonal Al-Ni-Co quasicrystal, the measued data are[28] :
3
ρ = 4.186g/cm , C11 = 234.3, C33 = 232.22, C44 = 70.19, C12 = 57.41,
C13 = 66.63(GPa),
R1 = −1.1, |R2 | < 0.2(GPa)
and there are no measured data for K1 , K2 (but we can use those obtained by the
Monte-Carlo simulation), and Γw can approximately be taken the value of the icosa-
hedral quasicrystal. In addition, the tensile strength σc = 450MPa for decagonal
Al-Cu-Co quasicrystals before annealing, and σc = 550MPa after annealing. The
hardness for decagonal Al-Cu-Co quasicrystals is 4.10GPa[30,31] , the fracture tough-
√
ness is 1.0 ∼ 1.2MPa m [30] .
With these basic data, the computation for stress analysis for statics and dy-
namics can be undertaken.
4.8 Summary and mathematical solvability of boundary value

or initial-boundary value problem
For a static equilibrium problem, the mathematical formulation is
∂σij ∂Hij
+ fi = 0, + hi = 0, xi ∈ Ω , (4.8-1)
∂xj ∂xj

1 ∂ui ∂uj ∂wi
εij = + , wij = , xi ∈ Ω , (4.8-2)
2 ∂xj ∂xi ∂xj

σij = Cijkl εkl + Rijkl wkl ,

xi ∈ Ω , (4.8-3)
Hij = Kijkl wkl + Rklij εkl ,
σij nj = Ti , Hij nj = h, xi ∈ Γt , (4.8-4)
4.8 Summary and mathematical solvability of boundary value or... 45
ui = ūi , wi = w̄i , xi ∈ Γ u . (4.8-5)
For a dynamic problem, based on the Bak’s argument, the mathematical formu-
lation is
∂σij ∂ 2 ui ∂Hij ∂ 2 wi
+ fi = ρ 2 , + gi = ρ 2 , xi ∈ Ω , t > 0, (4.8-6)
∂xj ∂t ∂xj ∂t

1 ∂ui ∂uj ∂wi
εij = + , wij = , xi ∈ Ω , t > 0, (4.8-7)
2 ∂xj ∂xi ∂xj


xi ∈ Ω , t > 0, (4.8-8)
σij nj = Ti , Hij nj = h, xi ∈ Γt , t > 0, (4.8-9)
ui = ūi , wi = w̄i , xi ∈ Γu , t > 0, (4.8-10)
ui |t=0 = ui0 , u̇i |t=0 = u̇i0 , wi |t=0 = wi0 , ẇi |t=0 = ẇi0 , xi ∈ Ω . (4.8-11)
For a dynamic problem, based on the argument of Lubensky et al, the mathe-
matical formulation is
∂σij ∂ 2 ui ∂Hij ∂wi 1

+ fi = ρ 2 , + hi = κ ,κ = , xi ∈ Ω , t > 0, (4.8-12)
∂xj ∂t ∂xj ∂t Γw

1 ∂ui ∂uj ∂wi
εij = + , wij = , xi ∈ Ω , t > 0, (4.8-13)
2 ∂xj ∂xi ∂xj


xi ∈ Ω , t > 0, (4.8-14)
σij nj = Ti , Hij nj = h, xi ∈ Γt , t > 0, (4.8-15)
ui = ūi , wi = w̄i , xi ∈ Γu , t > 0, (4.8-16)
ui |t=0 = ui0 , u̇i |t=0 = u̇i0 , wi |t=0 = wi0 , xi ∈ Ω . (4.8-17)
The solution satisfying all equations and corresponding initial conditions and
boundary conditions is just the realistic solution of elasticity of quasicrystals math-
ematically and has physical meaning.
The existence and uniqueness of solutions of elasticity of quasicrystals will further
be discussed in Chapter 13.
4.9 Appendix of Chapter 4: Description on physical basis of

elasticity of quasicrystals based on the Landau density
wave theory
In Section 4.1 we gave the formula (4.1-1) as the physical basis of elasticity of qua-
sicrystals and did not discuss its profound physical source, because the discussion is
concerned with quite complicated background, which is unnecessary for the begin-
ner before reading Sections 4.2∼4.8. At this point, it is advantageous to study the
physical background. The reader is advised to read Section 1.5 (the Appendix of
Chapter 1) first.
Chapter 3 shows quasicrystals belong to subject of condensed matter physics
rather than traditional solid state physics, though the former is evolved from the
latter. In the development, the symmetry-breaking (or broken symmetry) forms the
core concept and principle of the condensed matter physics. According to the under-
standing of physicists that the Landau density wave description on the elasticity of
quasicrystals is a natural choice, though there are some other descriptions, e.g. the
unit cell description based on the Penrose tiling. Now the difficulty lies in that the
workers in other disciplines are not so familiar with the Landau theory and the rele-
vant topics. For this reason we have introduced the Landau theory in the Appendix
of Chapter 1 (i.e., Section 1.5, in which the concept on incommensurate crystals
is also introduced in brief, which will be concerned though it is not related to the
Landau theory) and the Penrose tiling in Section 3.1. These important physical and
mathematical results can help us to understand the elasticity of quasicrystals.
Immediately after the discovery of quasicrystals, Bak[1] published the theory of
elasticity in which he used the three important results in physics and mathematics
mentioned above, but the core is the Landau theory on elementary excitation and
symmetry-breaking of condensed matter. Bak[1,2] pointed out too, ideally, one would
like to explain the structure from first-principles calculations taking into account the
actual electronic properties of constituent atoms. Such a calculation is hardly possi-
ble to date, so he suggested that the Landau phenomenological theory[3] on structural
transition can be used, i.e., the condensed phase is described by a symmetry-breaking
order parameter which transforms as an irreducible representation of the symmetry
group of a liquid with full translational and rotational symmetry. According to
the Landau theory, the order parameter of quasicrystals is the density wave. For
the density of the ordered, low-temperature d-dimensional quasicrystal can be ex-
pressed as a Fourier series by extended formula (1.5-5) (the expansion exists due to
the periodicity in lattice or reciprocal lattice of higher dimensional space)

ρ(r) = ρG exp{iG · r} = |ρG | exp{−iΦG + iG · r}, (4.9-1)
G∈LR G∈LR
4.9 Appendix of Chapter 4: Description on physical basis of elasticity... 47
where G is a reciprocal vector, and LR the reciprocal lattice (the concepts on the
reciprocal vector and reciprocal lattice, referring to Chapter 1), ρG is a complex
number
ρG = |ρG | e−iΦG (4.9-2)
with an amplitude |ρG | and phase angle ΦG , due to ρ(r) being real, |ρG | = |ρ−G |
and ΦG = −Φ−G .
There exists a set of N base vectors, {Gn }, so that each G ∈ LR can be written

as mn Gn for integers mn . Furthermore N = kd, where k is the number of
the mutually incommensurate vectors in the d-dimensional quasicrystal. In general
k = 2. A convenient parametrization of the phase angle is given by
Φn = Gn · u + G⊥
n · w, (4.9-3)
in which u can be understood similar to the phonon like that in conventional crystals,
while w can be understood the phason degrees of freedom in quasicrystals, which
describe the local rearrangement of unit cell description based on the Penrose tiling.
Both are functions of the position vector in the physical space only, where Gn is the
reciprocal vector in the physical space E3 just mentioned and G⊥ n is the conjugate
3
vector in the perpendicular space E⊥ . We can realize that the above mentioned Bak’s
hypothesis is a natural development of Anderson’s theory introduced in Section 1.5.
Almost in the same time, Levine et al[4] , Lubensky et al [5∼8] , Kalugin et al[9] ,
Torian and Mermin[10] , Jaric[11] , Duneau and Katz[12] , Socolar et al[13] , Gahler
and Phyner[14] carried out the study on elasticity of quasicrystals. Though the
researchers studied the elasticity from different descriptions, e.g. the unit-cell de-
scription based on the Penrose tiling is adopted too, but the density-wave description
based on the Laudau phenomenological theory on symmetry-breaking of condensed
matter has played the central role and been widely acknowledged. This means there
are two elementary excitations of low-energy, phonon u and phason w for quasicrys-
tals, in which vector u is in the parallel space E3 and vector w is in the perpendicular
3
space E⊥ , respectively. So the total displacement field for quasicrystals is
ū = u ⊕ u⊥ = u ⊕ w,
which is the formula (4.1-1), where ⊕ represents the direct sum.

According to the argument of Bak etc, that
u = u(r ), w = w(r ),
i.e., u and w depend upon special radius vector r in parallel space E3 only, this is
formula (4.1-2). For simplicity the superscript of r is removed in the presentation
in Sections 4.2∼4.8.
Even if introducing u and w by such a way the concept of phason is hard to

be accepted by some readers. We would like to give some additional explanation in
terms of projection concept as below.
We previously said that the quasicrystal in three-dimensional space may be seen
as a projection of periodic structure in higher dimensional space. For example, one-
dimensional quasicrystals in physical space may be seen as a projection of “periodic
crystals” in four-dimensional space, while in the one-dimensional quasicrystals, the
atom arrangement is quasiperiodic only in one direction, say z-axis direction, and
periodic in other two directions. The atom arrangement quasiperiodic axis may seen
as a projection of two-dimensional periodic crystal shown in Fig. 4.9-1 (a), in which
dots form the two-dimensional, e.g. right square, crystal, and lines with slop of
irrational numbers can correspond to quasiperiodic structures (by contrast, if the
slope is rational number it corresponds to the periodic structure). For this purpose
we can use the so-called Fibonacci sequence formed by a longer segment L and a
shorter segment S according to the recurrence (whose geometry depiction is shown
in Fig. 4.9-1 (b))
Fn+1 = Fn + Fn−1
and
F0 : S,
F1 : L,
F2 : LS,
F3 : LSL,
F4 : LSLLS.
The Fibonacci sequence is ordered but not periodic. The geometric expression of
the sequence can be shown in the axis E1 i.e., E in Fig. 4.9-1 (a), and E is the so-
called parallel space, and that perpendicular to which is the so-called vertical space
E2 , i.e., E⊥ . The Fibonacci sequence is a useful tool to describe the geometry of
one-dimensional quasiperiodic structure, like that the Penrose tiling to describe the
geometry of two- and three-dimensional quasicrystals. The Fig. 4.9-1 may help us
to understand the internal-space E⊥ . For one-dimensional quasicrystals, the figure
can give an explicit description, while for two- and three-dimensional quasicrystals
there is no such an explicit graph.
Since quasicrystals belong to one of incommensurate phases, and there are pha-
son modes in the incommensurate crystals, denoted by w(r ), which may be un-
derstood as the corresponding new displacement field, if people have knowledge on
incommensurate phases, then they may easily understand the origin of phason modes
in quasicrystal, though conventional incommensurate crystals are not certainly the
actual quasicrystals.
4.9 Appendix of Chapter 4: Description on physical basis of elasticity... 49
Fig. 4.9-1 A geometrical representation for one-dimensional quasiperiodic structure

(a) A projection of two-dimensional crystal can generate a one-dimensional quasiperiodic
structure, (b) The Fibonacci sequence
The phonon variables u(r ) appearing in the physical space E3 , vector u rep-
resents the displacement of lattice point deviated from its equilibrium position due
to the vibration of the lattice. The propagation of this vibration is sound waves in
solids. Though vibration is a mechnical motion, which can be quantized, the quan-
tum of this motion is named phonon. So the u field is called phonon field in physical
terminology. The gradient of the u field characterizes the changes in volume and in
shape of cells—this is identical to that in the classical elasticity (see, e.g. Chapter
2, and previous sections of this chapter).
As mentioned before, the phason variables are substantively related with struc-
tural transitions of alloys, some of them can be observed from the characteristics of
diffraction patterns. Lubensky et al[5,7] and Horn et al[15] discussed the connection
between the phenomena and phason strain. These profound observations could not
be discussed here, reader can refer to the review given by Hu et al[16] . This makes us
know that the phason modes exist. The physical meaning of phason variables can be
explained as a quantity to describe the local rearrangement of atoms in a cell. We
know that the phase transition in crystalline materials is just induced by the atomic
local rearrangement. The unit-cell description on quasicrystals mentioned above
predicts that w describes the local arrangement of Penrose titling. These findings
may help us to understand the meaning of the unusual field variables. Afterward ex-
perimental investigations by neutron scattering[17∼20] , Moessbauer spectroscopy[21] ,
NMR (Nuclear Magnetic Resonance)[22,23] and specific heat measurements[24,25] , the
concept of thermal-induced phason flips has been suggested, this is identical to the
diffusive essentiality of phasons. Note that the so-called diffusion here is quite differ-
ent from that in metallic periodic crystals (which mainly results from the presence
of vacancies in the lattice, and the vacancies are not necessary for atomic motion in
the quasicrystal structures). We will discuss this aspect in Chapter 10.
It should be noted that vector u and vector w are different in nature under
certain symmetry operations. This can be explained by the group theory. The
discussion is omitted here.
References
[1] Bak P. Phenomenological theory of icosahedral incommensurate (quaisiperiodic) order,
in Mn-Al alloys. Phys Rev Lett, 1985, 54(8): 1517–1519
[2] Bak P. Symmetry, stability and elastic properties of icosahedral incommensurate crys-
tals. Phys Rev B, 1985, 32(9): 5764–5772
[3] Landau L D, Lifshitz E M. Theoretical Physics V: Statistical Physics. 3rd ed. New
York: Pregamen Press, 1980
[4] Levine D, Lubensky T C, Ostlund S et al. Elasticity and dislocations in pentagonal
and icosahedral quasicrystals. Phys Rev Lett, 1985, 54(8): 1520–1523
[5] Lubensky T C, Ramaswamy S, Nad Toner J. Hydrodynamics of icosahedral quasicrys-
tals. Phys Rev B, 1985, 32(11): 7444–7452
[6] Lubensky T C, Ramaswamy S, Toner J. Dislocation motion in quasicrystals and im-
plications for macroscopic properties. Phys Rev B, 1986, 33(11): 7715–7719
[7] Lubensky T C, Socolar J E S, Steinhardt P J et al. Distortion and peak broadening
in quasicrystal diffraction patterns. Phys Rev Lett, 1986, 57(12): 1440–1443
[8] Lubensky T C. Introduction to Quasicrystals. Jaric M V. Boston: Academic Press,
1988
[9] Kalugin P A, Kitaev A, Levitov L S. 6-dimensional properties of Al0.86 Mn0.14 alloy.
J Phys Lett, 1985, 46(13): 601–607
[10] Torian S M, Mermin D. Mean-field theory of quasicrystalline order. Phys Rev Lett,
1985, 54(14): 1524–1527
[11] Jaric M V. Long-range icosahedral orientational order and quasicrystals. Phys Rev
Lett, 1985, 55(6): 607–610
[12] Duneau M, Katz A. Quasiperiodic patterns, Phys Rev Lett, 1985, 54(25): 2688–2691
[13] Socolar J E S, Lubensky T C, Steinhardt P J. Phonons, phasons, and dislocations in
quasicrystals. Phys Rev B, 1986, 34(5): 3345–3360
References 51
[14] Gahler F, Rhyner J. Equivalence of the generalised grid and projection methods for
the construction of quasiperiodic tilings. J Phys A: Math Gen, 1986, 19(2): 267–277
[15] Horn P M, Melzfeldt W, Di Vincenzo D P et al. Systematics of disorder in quasiperi-
odic material. Phys Rev Lett, 1986, 57(12): 1444–1447
[16] Hu C Z, Wang R H, Ding D H. Symmetry groups, physical property tensors, elasticity
and dislocations in quasicrystals. Rep Prog Phys, 2000, 63(1): 1–39
[17] Coddens G, Bellissent R, Calvayrac Y et al. Evidence for phason hopping in icosahe-
dral Al-Fe-Cu quasicrystals. Europhys Lett, 1991, 16(3): 271–276
[18] Coddens G, Sturer W. Time-of-flight neutron-scattering study of phason hopping in
decagonal Al-Co-Ni quasicrystals. 1999, Phys Rev B, 1999, 60(1): 270–276
[19] Coddens G, Lyonnard S, Hennion B et al. Triple-axis neutron-scattering study of
phason dynamics in Al-Mn-Pd quasicrystals. Phys Rev B 2000, 62(10): 6268–6295
[20] Coddens G, Lyonnard S, Calvayrac Y et al. Atomic (phason) hopping in perfect icosa-
hedral quasicrystals Al70.3 Pd21.4 Mn8.3 by time-of-flight quasielastic neutron scatter-
ing. Phys Rev B, 1996, 53(6): 3150–3160
[21] Coddens G, Lyonnard S, Sepilo B et al. Evidence for atomic hopping of Fe in perfectly
icosahedral Al-Fe-Cu quasicrystals by57 Fe Moessbauer spectroscopy. J Phys, 1995,
5(7): 771–776
[22] Dolisek J, Ambrosini B, Vonlanthen P et al. Atomic motion in quasicrystalline
Al70 Re8.6 Pd21.4 : A two-dimensional exchange NMR study. Phys Rev Lett, 1998, 81(17):
3671–3674
[23] Dolisek J, Apih T, Simsic M et al. Self-diffusion in icosahedral Al72.4 Pd20.5 Mn7.1 and
27
phason percolation at low temperatures studied by Al NMR. Phys Rev Lett, 1999,
82(3): 572–575
[24] Edagawa K, Kajiyama K. High temperature specific heat of Al-Pd-Mn and Al-Cu-Co
quasicrystals, Mater Sci and Eng A, 2000, 294∼296(5): 646–649
[25] Edagawa K, Kajiyama K, Tamura R et al. High-temperature specific heat of qua-
sicrystals and a crystal approximant. Mater Sci and Eng A, 2001, 312(1∼2): 293–298
[26] Ding D H, Yang W G, Hu C Z et al. Generalized elasticity theory of quasicrystals.
Phys Rev B, 1993, 48(10): 7003–7010
[27] Fan T Y, Wang X F, Li W et al. Elasto-/hydro-dynamics of quasicrystals. Phil Mag,
2009, 89(6), 501–512
[28] Edagawa K, Takeuchi S. Elasticity, dislocations and their motion in quasicrystals,
Dislocation in Solids. 2007, Chapter 76. ed. by Nabarro E R N, Hirth J P, 367–417
[29] Edagawa K, Giso Y. Experimental evaluation of phonon-phason coupling in icosahe-
dral quasicrystals. Phil Mag, 2007, 87(1): 77–95
[30] Meng X M, Tong B Y, Wu Y K. Mechanical properties of quasicrystal Al65 Cu20 Co15 .
Acta Metallurgica Sinica, 1994, 30(2): 61–64 (in Chinese)
[31] Takeuchi S, Iwanhaga H, Shibuya T. Hardness of quasicrystals. Japanese J Appl Phys,

1991, 30(3): 561–562
[32] Francoual S, Levit F, de Boussieu M et al. Dynamics of phason fluctuations in the
i-Al-Pd-Mn quasicrystals. Phys Rev Lett, 2003, 91(22): 225501
[33] Coddens G. On the problem of the relation between phason elasticity and phason
dynamics in quasicrystals. Eur Phys J B, 2006, 54(1): 37–65
Chapter 5
Elasticity theory of one-dimensional
quasicrystals and simplification
As mentioned in Chapter 4, there exist one-, two- and three-dimensional quasicrys-

tals. Each can be further divided into sub-classes by symmetry consideration. In
the class of one-dimensional quasicrystals, the atom arrangement is quasiperiodic
in one direction (say, z-direction) and is periodic in the plane perpendicular to the
direction (xy-plane). Although the quasicrystal is one-dimensional, the structure
is three-dimensional; it is generated in a three-dimensional body. One-dimensional
quasicrystals can be regarded as a projection of periodic crystals in four-dimensional
space to physical space. There are 4 non-zero displacements, ux , uy , uz and wz
(wx = wy = 0). In this sense, the elasticity of one-dimensional quasicrystals is a
four-dimensional problem.
Table 5.0-1 Systems, Laue classes and point groups of one-dimensional

quasicrystals
Systems Laue classes Point groups
Triclinic 1 1 1̄
Monoclinic 2 2 mh 2/mh
3 2h m 2h /m
Orthorhombic 4 2h 2h 2 mm2 2h mmh mmmh
Tetragonal 5 4 4̄ 4/mh
6 42h 2h 4mm 4̄2h m 4/mh mm
Rhombohedral 7 3 3̄
(or trigonal) 8 32h 3m 3̄m
Hexgonal 9 6 6̄ 6/mh
10 62h 2h 6mm 6̄m2h 6/mh mm
We here briefly list the crystal systems and Laue classes of one-dimensional qua-
sicrystals, in which the concept of point group must be concerned, and we do not
concern with the concept of space group.

54 Chapter 5 Elasticity theory of one-dimensional quasicrystals and simplification
In the following we discuss the elasticity of quasicrystals of certain types listed

in Table 5.0-1.
5.1 Elasticity of hexagonal quasicrystals

As pointed out previously that for one-dimensional quasicrystals, there are phonon
displacements ux , uy , uz and phason displacement wz (and wx = wy = 0), the
corresponding strains are
∂ux ∂uy ∂uz

εxx = , εyy = , εzz = , (5.1-1)
∂x ∂y ∂z

1 ∂uz ∂uy 1 ∂uz ∂ux
εyz = εzy = + , εzx = εxz = + ,
2 ∂y ∂z 2 ∂x ∂z

1 ∂ux ∂uy
εxy = εyx = + ,
2 ∂y ∂x
∂wz ∂wz ∂wz
wzx = , wzy = , wzz = (5.1-2)
∂x ∂y ∂z
and other wij = 0. Formulas (5.1-1) and (5.1-2) hold for all one-dimensional qua-
sicrystals. In this section we only discuss one-dimensional hexagonal quasicrytals.
Let the strains given by (5.1-1) and (5.1-2) be expressed by the vector with 9
components,
[ε11 , ε22 , ε33 , 2ε23 , 2ε31 , 2ε12 , w33 , w31 , w32 ] (5.1-3)
or
[εxx , εyy , εzz , 2εyz , 2εzx , 2εxy , wzz , wzx , wzy ] , (5.1-4)
The corresponding stresses are
[σxx , σyy , σzz , σyz , σzx , σxy , Hzz , Hzx , Hzy ] (5.1-5)
The elastic constant matrix is

⎡ ⎤
C11 C12 C13 0 0 0 R1 0 0
⎢ C12 C11 C13 0 0 0 R1 0 0 ⎥
⎢ ⎥
⎢ C13 C13 C33 0 0 0 R2 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 C44 0 0 0 0 R3 ⎥
⎢ ⎥
[CKR] = ⎢ 0 0 0 0 C44 0 0 R3 0 ⎥,
⎢ ⎥
⎢ 0 0 0 0 0 C66 0 0 0 ⎥
⎢ ⎥
⎢ R1 R1 R2 0 0 0 K1 0 0 ⎥
⎢ ⎥
⎣ 0 0 0 0 R3 0 0 K2 0 ⎦
0 0 0 R3 0 0 0 0 K2
5.1 Elasticity of hexagonal quasicrystals 55
where a short notation for the phonon elastic constant tensor is used, i.e., index
11→1, 22→2, 33→3, 23→4, 31→5, 12→6 and Cijkl is denoted as Cpq accordingly,
C11 = C1111 = C2222 , C12 = C1122 , C33 = C3333 , C44 = C2323 = C3131 ,
C11 − C12 C1111 − C1122

C13 = C1133 = C2233 , C66 = = .
2 2
There are five independent phonon elastic constants, in addition K1 = K3333 , K2 =
K3131 = K3232 , i.e., two independent phason elastic constants, and R1 = R1133 =
R2233 , R2 = R3333 , R3 = R2332 = R3131 , i.e., three phonon-phason coupling elastic
constants.
The corresponding stress-strain relations are:
⎧
⎪
⎪ σxx = C11 εxx + C12 εyy + C13 εzz + R1 wzz ,
⎪
⎪
⎪
⎪ σyy = C12 εxx + C11 εyy + C13 εzz + R1 wzz ,
⎪
⎪
⎪
⎪ σzz = C13 εxx + C13 εyy + C33 εzz + R2 wzz ,
⎪
⎪
⎨σyz = σzy = 2C44 εyz + R3 wzy ,
σzx = σxz = 2C44 εzx + R3 wzx , (5.1-6)
⎪
⎪
⎪σxy = σyx = 2C66 εxy ,
⎪
⎪
⎪
⎪Hzz = R1 (εxx + εyy ) + R2 εzz + K1 wzz ,
⎪
⎪
⎪
⎪Hzx = 2R3 εzx + K2 wzx ,
⎪
⎩
Hzy = 2R3 εyz + K2 wzy
and other Hij = 0.

The equilibrium equations are:
⎧ ∂σ ∂σxy ∂σxz
⎪
⎪
xx
+ + = 0,
⎪
⎪ ∂x ∂y ∂z
⎪
⎪
⎪
⎪
⎪
⎪
∂σyx ∂σyy ∂σyz
⎨ + + = 0,
∂x ∂y ∂z
(5.1-7)
⎪
⎪ ∂σzx ∂σzy ∂σzz
⎪
⎪ + + = 0,
⎪
⎪ ∂x ∂y ∂z
⎪
⎪
⎪
⎪
⎩ ∂Hzx + ∂Hzy + ∂Hzz = 0.
∂x ∂y ∂z
The above results are obtained by Wang et al[1] .

The elastic equilibrium problem of one-dimensional hexagonal quasicrystals is
more complicated than that of three-dimensional classical elasticity. Here there are
4 displacements, 9 strains and 9 stresses, which added up to 22 field variables. The
corresponding field equations are also 22, consisting of 4 for equilibrium equations
and 9 equations of deformation geometry and 9 stress-strain relations. We will
present a rigorous treatment of the problem later on. In the following we give a
simplified treatment.
5.2 Decomposition of the problem into plane and anti-plane

problems
If there is a straight dislocation or a Griffith crack along the direction of the atom
quasiporiodic arrangement, the deformation is independent of the z-axis, say
∂
= 0, (5.2-1)
∂z
Then
∂ui ∂wz
= 0, i = 1, 2, 3, = 0. (5.2-2)
∂z ∂z
Hence,
1 ∂uz 1 ∂uz
εzz = wzz = 0, εyz = εzy = , εzx = εxz = , (5.2-3)
2 ∂y 2 ∂x
∂σij ∂Hij
= 0, = 0. (5.2-4)
∂z ∂z
The generalized Hooke’s law is simplified as
⎧
⎪
⎪ σxx = C11 εxx + C12 εyy ,
⎪
⎪
⎪
⎪
⎪
⎪ σyy = C12 εxx + C11 εyy ,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ σxy = σyx = 2C66 εxy ,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ σzz = C13 (εxx + εyy ),
⎪
⎪
⎪
⎨
σyz = σzy = 2C44 εyz + R3 wzy ,
(5.2-5)
⎪
⎪
⎪
⎪
⎪
⎪ σzx = σxz = 2C44 εzx + R3 wzx ,
⎪
⎪
⎪
⎪
⎪
⎪ Hzz = R1 (εxx + εyy ),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ H = 2R3 εzx + K2 wzx ,
⎪ zx
⎪
⎪
⎪
⎪
⎪
⎪
⎩ Hzy = 2R3 εyz + K2 wzy .
In the absence of the body force and generalized body force the equilibrium equations
are
∂σxx ∂σxy ∂σyx ∂σyy ∂σzx ∂σzy
+ = 0, + = 0, + = 0, (5.2-6)
∂x ∂y ∂x ∂y ∂x ∂y
∂Hzx ∂Hzy
+ = 0. (5.2-7)
∂x ∂y
5.2 Decomposition of the problem into plane and anti-plane problems 57
Equations(5.1-2), (5.1-3), (5.2-5)∼(5.2-7)define two uncoupled problems[2] . The first

of them is
⎧
⎪
⎪ σxx = C11 εxx + C12 εyy ,
⎪
⎪
⎪
⎪ σyy = C12 εxx + C11 εyy ,
⎪
⎪
⎪
⎪
⎪
⎪ σxy = (C11 − C12 )εxy ,
⎪
⎪
⎪
⎪
⎪
⎪
⎨ σzz = C13 (εxx + εyy ),
Hzz = R1 (εxx + εyy ), (5.2-8)
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ ∂σxx ∂σxy ∂σyx ∂σyy
⎪
⎪ + = 0, + = 0,
⎪
⎪ ∂x ∂y ∂x ∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂ux ∂uy 1 ∂uy ∂ux
⎩ εxx = , εyy = , εxy = + ,
∂x ∂y 2 ∂x ∂y
this is the classical plane elasticity of conventional hexagonal crystals. The second
one is ⎧
⎪
⎪ σyz = σzy = 2C44 εyz + R3 wzy ,
⎪
⎪
⎪
⎪
⎪
⎪ σzx = σxz = 2C44 εzx + R3 wzx ,
⎪
⎪
⎪
⎪
⎪
⎪ Hzx = 2R3 εzx + K2 wzx ,
⎪
⎪
⎪
⎪
⎪
⎪ Hzy = 2R3 εyz + K2 wzy ,
⎨
∂σzx ∂σzy ∂Hzx ∂σzy (5.2-9)
⎪
⎪ + = 0, + = 0,
⎪
⎪
⎪
⎪
∂x ∂y ∂x ∂y
⎪
⎪
⎪
⎪ 1 ∂uz 1 ∂uz
⎪
⎪ εzx = = εxz , εzy = = εyz ,
⎪
⎪ 2 ∂x 2 ∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂wz ∂wz
⎩ wzx = , wzy = ,
∂x ∂y
which is a phonon-phason coupling elasticity problem, involving only the displace-
ments uz and wz . It is an anti-plane elasticity problem.
The plane elasticity described by (5.2-8) has been studied, extensively using the
stress function approach, e.g. it introduces
∂2U ∂2U ∂2U

σxx = , σyy = , σxy = − ,
∂y2 ∂x2 ∂x∂y
then equations (5.2-8) are reduced to solve
∇2 ∇2 U = 0.
The problem is considerably discussed in crystal (or classical) elasticity, we do not

consider it here.
We are interested in the phonon-phason coupling anti-plane elasticity described

by (5.2-9), which may bring new insight into the scope of elasticity of quasicry-
stals.
Substituting the deformation geometry relations into the stress-strain relations,
then into the equilibrium equations yields the final governing equations such as

C44 ∇2 uz + R3 ∇2 wz = 0,
(5.2-10)
R3 ∇2 uz + K2 ∇2 wz = 0,
because C44 K2 − R32 = 0, we have
∇2 uz = 0, ∇2 wz = 0, (5.2-11)
∂ ∂
where ∇2 = 2
+ 2 , so uz and wz are harmonic functions.
∂x ∂y
It is well known that the two-dimensional harmonic functions uz and wz can be
a real part or an imaginary part of any analytic functions φ(t) and ψ(t) of complex
√
variable t = x + iy, i = −1, respectively, i.e.,

uz (x, y) = Re φ(t),
(5.2-12)
wz (x, y) = Re ψ(t).
In this version, equations (5.2-11) should be identically satisfied. Determination of

φ(t) and ψ(t) depends upon appropriate boundary conditions, which will be dis-
cussed in detail in Chapters 7 and 8. The complex variable function method for
solving elasticity of one-, two- and three-dimensional quasicrystals will be summa-
rized fully in Chapter 11.
5.3 Elasticity of monoclinic quasicrystals

Decomposition procedure suggested by the author e.g. in Ref.[2,3] is applicable not
only for hexagonal quasicrystals but also for other one-dimensional quasicrystals.
Herein we discuss monoclinic quasicrystals.
For this kind of one-dimensional quasicrystals, there are 25 non-zero elastic con-
stants in total, namely, C1111 , C2222 , C3333 , C1122 , C1133 , C1112 , C2233 , C2212 , C3312 ,
C3232 , C3231 , C3131 , C1212 for the phonon field, K3333 , K3131 , K3232 , K3132 for the
phason field and R1133 , R2233 , R3333 , R1233 , R2331 , R2332 , R3132 , R1233 for the
phonon-phason coupling.
The generalized Hooke’s law for the case is as follows[1] :
5.3 Elasticity of monoclinic quasicrystals 59
⎧
⎪ σxx = C11 εxx + C12 εyy + C13 εzz + 2C16 εxy + R1 wzz ,
⎪
⎪
⎪
⎪
⎪
⎪ σyy = C12 εxx + C22 εyy + C23 εzz + 2C26 εxy + R2 wzz ,
⎪
⎪
⎪
⎪
⎪
⎪ σzz = C13 εxx + C23 εyy + C33 εzz + 2C36 εxy + R3 wzz ,
⎪
⎪
⎪
⎪
⎪
⎪ σyz = σzy = 2C44 εyz + 2C45 εzx + R4 wzx + R5 wzy ,
⎨
σzx = σxz = 2C45 εyz + 2C55 εzx + R6 wzx + R7 wzy , (5.3-1)
⎪
⎪
⎪
⎪
⎪
⎪ σxy = σyx = C16 εxx + C26 εyy + C36 εzz + 2C66 εxy + R8 wzz ,
⎪
⎪
⎪
⎪
⎪
⎪ Hzx = 2R4 εyz + 2R6 εzx + K1 wzx + K4 wzy ,
⎪
⎪
⎪
⎪
⎪
⎪ Hzy = 2R5 εyz + 2R7 εzx + K4 wzx + K2 wzy ,
⎪
⎩
Hzz = R1 εxx + R2 εyy + R3 εzz + 2R8 εxy + K3 wzz ,
where the short notation is used for the phonon elastic constant tensor, such that
the index 11→1,22→2,33→3, 23→4,31→5,12→6 and Cijkl is denoted as Cpq ; for the
phason elastic constants, K3131 = K1 , K3232 = K2 , K3333 = K3 ,K3132 = K4 and for
the phonon-phason coupling elastic constants, R1133 = R1 , R2233 = R2 , R3333 = R3 ,
R2331 = R4 , R2332 = R5 , R3131 = R6 , R3132 = R7 , R1233 = R8 .
Under the assumption (5.2-1), the problem can be decomposed into two separate
problems as follows:
⎧
⎪
⎪ σxx = C11 εxx + C12 εyy + 2C16 εxy ,
⎪
⎪
⎨ σyy = C12 εxx + C22 εyy + 2C26 εxy ,
σxy = σyx = C16 εxx + C26 εyy + 2C66 εxy , (5.3-2)
⎪
⎪
⎪
⎪ σzz = C 13 ε xx + C 23 εyy + 2C36 ε xy ,
⎩
Hzz = R1 εxx + R2 εyy + R3 εzz + 2R8 εxy ,
and ⎧
⎪
⎪ σyz = σzy = 2C44 εyz + 2C45 εzx + R4 wzx + R5 wzy ,
⎨
σzx = σxz = 2C45 εyz + 2C55 εzx + R6 wzx + R7 wzy ,
(5.3-3)
⎪
⎪ Hzx = 2R4 εyz + 2R6 εzx + K1 wzx + K4 wzy ,
⎩
Hzy = 2R5 εyz + 2R7 εzx + K4 wzx + K2 wzy
in which the problem described by equations (5.3-2) is plane elasticity of monocline
crystals. By introducing the displacement potential G(x, y),

∂2 ∂2 ∂2
ux = C16 2 + C26 2 + (C12 + C66 ) G,
∂x ∂y ∂x∂y

∂2 ∂2 ∂2
uy = − C11 2 + C66 2 + 2C16 G,
∂x ∂y ∂x∂y
the equations of elasticity reduce to

∂4 ∂4 ∂4 ∂4 ∂4
c1 4 + c2 3 + c3 2 2 + c4 + c5 4 G = 0
∂x ∂x ∂y ∂x ∂y ∂x∂y3 ∂y
with constants
2
c1 = C16 − C11 C66 , c2 = 2(C16 C12 − C11 C26 ),
2
c3 = C12 − 2C16 C26 + 2C12 C66 − C11 C22 ,
2
c4 = 2(C26 C12 − C16 C22 ), c5 = C26 − C22 C66 .
Since this is classical elasticity and does not have a direct connection to phason
elasticity of one-dimensional quasicrystal, we do not consider it further.
We are interested in the problem described by equation (5.3-3), which is a
phonon-phason coupling problem. Substituting the equations of deformation ge-
ometry into the stress-strain relations and then into the equilibrium equations, we
obtain the governing equation:

∂4 ∂4 ∂4 ∂4 ∂4
a1 4 + a2 3 + a3 2 2 + a4 + a5 4 F = 0, (5.3-4)
∂x ∂x ∂y ∂x ∂y ∂x∂y 3 ∂y
where ⎧
⎪ ∂2 ∂2 ∂2
⎪
⎪ u = R 6 + R 5 + (R 4 + R 7 ) F,
⎨ z ∂x2 ∂y 2 ∂x∂y
(5.3-5)
⎪
⎪ ∂2 ∂2 ∂2
⎪
⎩ wz = − C55 + C + 2C F,
44 45
∂x2 ∂y 2 ∂x∂y
and
⎧
⎨ a1 = R62 − K1 C56 , a2 = 2(R6 (R4 + R7 ) − K1 C45 − K4 C55 ),
a3 = 2R5 R6 + (R4 + R7 )2 − K1 C44 − K2 C55 − 4K4 C45 , (5.3-6)
⎩
a4 = 2[R5 (R4 + R7 ) − K2 C45 − K4 C44 ], a5 = R52 − K2 C44 .
In the following, we consider only anti-plane elasticity of monocline quasicrystals,

its solution has the complex representation
2

F (x, y) = 2Re Fk (zk ), zk = x + μk y, (5.3-7)
k=1
where Fk (zk ) are analytic functions of zk and
μk = αk + iβk (5.3-8)
are the distinct complex parameters to be determined by the characteristic equation
a5 μ4 + a4 μ3 + a3 μ2 + a2 μ + a1 = 0 (5.3-9)
and μ1 = μ2 .
If the roots of equation (5.3-9) are double roots, i.e., μ1 = μ2 , then
F (x, y) = 2Re[F1 (z1 ) + z̄1 F2 (z1 )], z1 = x + μ1 y. (5.3-10)

5.4 Elasticity of orthorhombic quasicrystals 61
Substitution of formula (5.3-7) into (5.3-5) and then into (5.3-3) we arrive at the
complex representation of the displacements and stresses as follows:
⎧
⎪ 2
⎪
⎪
⎪
⎪ u = 2Re [R6 + (R4 + R7 )μk + R5 μ2k ]fk (zk ),
⎪
⎪
z
⎪
⎪ k=1
⎪
⎪
⎪
⎪ 2
⎪
⎪
⎪
⎪ w z = −2Re (C55 + 2C45 μk + C44 μ2k )fk (zk ),
⎪
⎪
⎪
⎪ k=1
⎪
⎪
⎪
⎪ σ = σ
⎪
⎪
zy yz
⎪
⎪ 2
⎪
⎪
⎪
⎪ = 2Re [R6 C45 − R4 C55 + (R6 C44 − R4 C45 + R7 C45 − R5 C55 )μk
⎪
⎪
⎪
⎪
⎪
⎪
k=1
⎪
⎪ +(R7 C44 − R5 C45 )μ2k ]fk (zk ),
⎪
⎪
⎪
⎪
⎨ σzx = σxz
⎪
⎪ 2
⎪
⎪
⎪
⎪ = 2Re [R4 C55 − R6 C45 + (R5 C55 + R4 C45 − R6 C44 − R5 C55 )μk
⎪
⎪
⎪
⎪ k=1
⎪
⎪
⎪
⎪ +(R3 C45 − R7 C44 )μ2k ]μk fk (zk ),
⎪
⎪
⎪
⎪ 2
⎪
⎪
⎪
⎪ H = 2Re [(R7 + R5 μk )(R6 + R4 μk + R7 μk + R5 μ2k )
⎪
⎪ zx
⎪
⎪
⎪
⎪
k=1
⎪
⎪ −(K4 + K2 μk )(C55 + 2C45 μk + C44 μ2k )]fk (zk ),
⎪
⎪
⎪
⎪
⎪
⎪ 2
⎪
⎪
⎪
⎪ H = 2Re [(R6 + R4 μk )(R6 + R4 μk + R7 μk + R5 μ2k )
⎪
⎪
zy
⎪
⎪ k=1
⎪
⎩ −(K + K μ )(C + 2C μ + C μ2 )]f (z ),
4 2 k 55 45 k 44 k k k
(5.3-11)
2

where fk (zk ) ≡ ∂ Fk (zk ) ∂zk2
= Fk (zk ).
Determination of analytic functions Fk (zk ) must be supplemented with the bound-
ary conditions of specific problems, which will be given in Chapters 7 and 8.
5.4 Elasticity of orthorhombic quasicrystals

In Table 5.0-1, we know that orthorhombic quasicrystals contain the points 2h 2h 2,
mm2, 2h mmh and mmmh ,which belong to one Laue class 4. Due to the increase of
symmetric elements of this quasicrystal system in comparison with the monocline
quasicrystal system, one has
C16 = C26 = C36 = C45 = K4 = R4 = R7 = R8 = 0. (5.4-1)
Thus, the total number of non-zero elastic constants in the case reduces to 17,i.e.,
C11 , C22 , C33 , C12 , C44 , C55 , C66 for the phonon field, K1 , K2 , K3 for the phason
fields and R1 , R2 , R3 , R5 , R6 for the phonon-phason coupling field.
With results (5.4-1), then (5.3-6) is simplified to
a2 = a4 = 0, a1 = R62 −K1 C55 , a5 = R52 −K2 C66

a3 = 2R5 R6 −K1 C44 −K2 C55 ,
(5.4-2)
and a1 and a5 remain the same as in (5.3-6). Then, equation (5.3-4) reduces to

∂4 ∂4 ∂4
a1 4 + a3 2 2 + a5 4 F = 0, (5.4-3)
∂x ∂x ∂y ∂y
The solution takes the form
⎧
2
⎪
⎪
⎪
⎪ uz = 2Re [R6 + R5 μ2k ]fk (zk ),
⎪
⎪
⎪
⎪
⎪
⎪
k=1
⎪
⎪ 2
⎪
⎪ −2Re (C55 + C44 μ2k )fk (zk ),
⎪
⎪
w z =
⎪
⎪
⎪
⎪
k=1
⎪
⎪ 2
⎪
⎪ (R6 C44 − R5 C55 )μk fk (zk ),
⎪
⎪ σzy = σ yz = 2Re
⎪
⎨ k=1
2 (5.4-4)
⎪
⎪ σzx = σxz = 2Re (R5 C55 − R6 C44 )μ2k fk (zk ),
⎪
⎪
⎪
⎪
⎪
⎪
k=1
⎪
⎪ 2
⎪
⎪ [R5 R6 − K2 C55 + (R52 − K2 C44 )μ2k )
⎪
⎪
Hzy = 2Re
⎪
⎪
⎪
⎪ k=1
⎪
⎪ − (K
4 + K2 μk )]μk fk (zk ),
⎪
⎪
⎪
⎪ 2
⎪
⎪
⎪
⎩ Hzx = 2Re [R62 − K1 C55 + (R5 R6 − K1 C44 )μ2k ]fk (zk ).
k=1
in which fk (zk ) represents analytic function of zk .
5.5 Tetragonal quasicrystals

From Table 5.0-1 we know that one-dimensional tetragonal quasicrystal have 7 point
groups, in which groups 4̄2h m, 4mm, 42h 2h and 4/mh mm belong to Laue class 6.
Besides (5.4-1), we also have
C11 = C22 , C13 = C23 , C44 = C55 , K1 = K2 , R1 = R2 , R5 = R6 . (5.5-1)
The total number of non-zero elastic constants reduces to 11.

With (5.5-1) and (5.4-2), one can find the complex representation of the solution
for the anti-plane elasticity of quasicrystals of Laue class 6.
The point groups 4 ,4̄ and 4/mh belong to Laue class 5, in which the solution of
the anti-plane elasticity can also be derived in a similar manner.
The governing equation of anti-plane elasticity for this kind of quasicrystals is
∇2 ∇2 F = 0. (5.5-2)
5.6 The space elasticity of hexagonal quasicrystals 63
5.6 The space elasticity of hexagonal quasicrystals

The decomposition procedure proposed in Ref[2] has been developed to simplify
the elasticity of various systems of one-dimensional quasicrystals, in the previous
sections. The main feature of the procedure lies in decomposition a space (three-
dimensional) elasticity into plane elasticity and anti-plane elasticity for the qua-
sicrystalline material under study. The merits will be shown in Chapters 7 and 8
and other chapters.
In some cases, when the procedure is not applicable, we have to solve the problem
of space elasticity. As an example, we discuss the solution of space elasticity of
hexagonal quasicrystals here.
Substituting (5.5-1) into (5.1-6), then into (5.1-7) we obtain the equilibrium
equations expressed in terms of the displacements as follows
⎧
⎪
⎪ ∂2 ∂2 ∂2 ∂ 2 uy
⎪
⎪ C 11 + C 66 + C 44 u x + (C 11 − C 66 )
⎪
⎪ ∂x2 ∂y 2 ∂z 2 ∂x∂y
⎪
⎪
⎪
⎪
⎪
⎪
2
∂ uz 2
∂ wz
⎪
⎪ +(C13 + C44 ) + (R1 + R3 ) = 0,
⎪
⎪ ∂x∂z ∂x∂z
⎪
⎪
⎪
⎪
⎪
⎪ ∂ 2 ux ∂2 ∂2 ∂2
⎪
⎪ (C11 − C66 )
⎪
⎪ + C 66 + C 11 + C 44 uy
⎪
⎪ ∂x∂y ∂x2 ∂y2 ∂z 2
⎪
⎪
⎪
⎪
⎪
⎪ ∂ 2 uz ∂ 2 wz
⎪
⎪ +(C13 + C44 ) + (R1 + R3 ) = 0,
⎨ ∂y∂z ∂y∂z
2 (5.6-1)
⎪
⎪ ∂ ux ∂ 2 uy ∂2 ∂2 ∂2
⎪
⎪
⎪
⎪
(C13 + C44 ) + + C44 2 + C44 2 + C33 2 uz
⎪
⎪ ∂x∂z ∂y∂z ∂x ∂y ∂z
⎪
⎪
⎪
⎪
⎪
⎪ ∂2 ∂2 ∂2
⎪
⎪ + R3 + 2 + R2 2 wz = 0,
⎪
⎪ ∂x 2 ∂y ∂z
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂ 2 ux ∂ 2 uy ∂ ∂2 ∂2
⎪
⎪ (R + R ) + + R + + R uz
⎪
⎪ 1 3 3
∂x2 ∂y 2
2
∂z 2
⎪
⎪ ∂x∂z ∂y∂z
⎪
⎪ 2
⎪
⎪
⎪
⎪ ∂ ∂2 ∂2
⎩ + K2 + + K 1 wz = 0.
∂x2 ∂y 2 ∂z 2
If we introduce 4 displacement functions as

⎧
⎪ ∂ ∂F4 ∂ ∂F4
⎨ ux = (F1 + F2 + F3 ) − , uy = (F1 + F2 + F3 ) + ,
∂x ∂y ∂y ∂x (5.6-2)
⎪
⎩ uz = ∂ (m1 F1 + m2 F2 + m3 F3 ), wz = ∂ (l1 F1 + l2 F2 + l3 F3 )
∂z ∂z
and
∇2i Fi = 0, i = 1, 2, 3, 4, (5.6-3)
∂2 ∂2 2 ∂
2
∇2i = + + γ , i = 1, 2, 3, 4 (5.6-4)
∂x2 ∂y 2 i
∂z 2
with mi , li and γi defined by
C44 + (C13 + C44 )mi + (R1 + R3 )li

C11
C33 mi + R2 li
=
C13 + C44 + C44 mi + R3 li
R2 mi + K1 li C44
= = γi2 , i = 1, 2, 3, = γ42 , (5.6-5)
R1 + R3 + R3 mi + K2 li C66
consequently, equations (5.6-1) are identically satisfied.

The final governing equations (5.6-3) are three-dimensional harmonic equations.
By substituting (5.6-2) into (5.1-1), (5.1-2), and then into (5.1-6), we express the
stresses in terms of F1 , F2 , F3 and F4 as

∂2 ∂2 ∂ 2 F4
σxx = C11 2 + (C11 − 2C66 ) 2 (F1 + F2 + F3 ) − 2C66
∂x ∂y ∂x∂y
2 2
∂ ∂
+ C13 2 (m1 F1 + m2 F2 + m3 F3 ) + R1 2 (l1 F1 + l2 F2 + l3 F3 ),
∂z ∂z
∂2 ∂2 ∂ 2 F4
σyy = (C11 − 2C66 ) 2 + C11 2 (F1 + F2 + F3 ) + 2C66
∂x ∂y ∂x∂y
∂2 ∂2
+ C13 2 (m1 F1 + m2 F2 + m3 F3 ) + R1 2 (l1 F1 + l2 F2 + l3 F3 ),
∂z ∂z
∂2 2 2 2 ∂2
σzz = −C13 2 (γ1 F1 + γ2 F2 + γ3 F3 ) + C33 2 (m1 F1 + m2 F2 + m3 F3 )
∂z ∂z
∂2
+ R2 2 (l1 F1 + l2 F2 + l3 F3 ),
∂z 2
∂2 ∂ ∂2
σxy = σyx = 2C66 (F1 + F2 + F3 ) + C66 − F4 ,
∂x∂y ∂x2 ∂y 2
∂2
σyz = σzy = C44 [(m1 + 1)F1 + (m2 + 1)F2 + (m3 + 1)F3 ]
∂y∂z
∂ 2 F4 ∂2
+ C44 + R3 (l1 F1 + l2 F2 + l3 F3 ),
∂x∂z ∂y∂z
∂2
σzx = σxz = C44 [(m1 + 1)F1 + (m2 + 1)F2 + (m3 + 1)F3 ]
∂x∂z
∂ 2 F4 ∂2
− C44 + R3 (l1 F1 + l2 F2 + l3 F3 ),
∂y∂z ∂x∂z
∂2 ∂2
Hzz = −R1 2 (γ12 F1 + γ22 F2 + γ32 F3 ) + R2 2 (m1 F1 + m2 F2 + m3 F3 )
∂z ∂z
∂2
+ K1 2 (l1 F1 + l2 F2 + l3 F3 ),
∂z
References 65
∂2
Hzx = R3 [(m1 + 1)F1 + (m2 + 1)F2 + (m3 + 1)F3 ]
∂x∂z
∂ 2 F4 ∂2
− R3 + K2 (l1 F1 + l2 F2 + l3 F3 ),
∂y∂z ∂x∂z
∂2
Hzy = R3 [(m1 + 1)F1 + (m2 + 1)F2 + (m3 + 1)F3 ]
∂y∂z
∂ 2 F4 ∂2
+ R3 + K2 (l1 F1 + l2 F2 + l3 F3 ).
∂y∂z ∂y∂z
Harmonic equations (5.6-3) can be solved under appropriate boundary condi-
tions, which will be discussed in Chapter 8.
5.7 Other results of elasticity of one-dimensional quasicrys-

tals
There are other results of elasticity of one-dimensional quasicrystals, e.g. Fan et al[8]
on elasticity of one-dimensional quasicrystal-crystal coexisting phase (refer to Chap-
ter 7 for a brief discussion), Chen et al[9] , Wang et al,[10] etc on the three-dimensional
elasticity of hexagonal quasicrystals, they carried out considerable research in the
area, and have got quite a lot of achievements. Due to the space limitation, the
details of their work could not be handled here and reader can refer to the original
literatures.
References
[1] Wang R H, Yang W G, Hu C Z et al. Point and space groups and elastic behaviour
of one-dimensional quasicrystals. J Phys Condens Matter, 1997, 9(11): 2411–2422
[2] Fan T Y. Mathematical theory of elasticity and defects of quasicrystals. Advances
in Mechanics, 2000, 30(2): 161–174 (in Chinese)
[3] Fan T Y, Mai Y W. Elasticity theory, fracture mechanics and some relevant thermal
properties of quasicrystalline materials. Appl Mech Rev, 2004, 57(5): 325–344
[4] Liu G T, Fan T Y, Guo R P. Governing equations and general solutions of plane
elasticity of one-dimensional quasicrystals. Int J Solid and Structures, 2004, 41(14):
3949–3959
[5] Liu G T. The complex variable function method of the elastic theory of quasicrys-
tals and defects and auxiliary equation method for solving some nonlinear evalution
equations, Dissertation. Beijing Institute of Technology, 2004 (in Chinese)
[6] Peng Y Z, Fan T Y. Elastic theory of 1-D quasiperiodic stacking 2-D crystals. J.
Phys: Condens Matter, 2000, 12(45): 9381–9387
[7] Peng Y Z. Study on elastic three-dimensional problems of cracks for quasicrystals,
Dissertation. Beijing Institute of Technology, 2002 (in Chinese)
[8] Fan T Y, Fan L, Xie L Y et al. Study on interface of quasicrystal-crystal. J. Phys.:

Condens. Matter, 2009, submitted
[9] Chen W Q, Ma Y L, Ding H J. On three-dimensional elastic problems of one-
dimensional hexagonal quasicrystal bodies. Mech Res Commun, 2004, 31(5): 633–
641
[10] Wang X. The general solution of one-dimensional hexagonal quasicrystal. Appl Math
Mech, 2006, 33(4): 576–580
Chapter 6
Elasticity of two-dimensional quasicrystals
and simplification
As has been shown in Chapter 5, in one-dimensional quasicrystals, elasticity can be

decomposed into plane elasticity and anti-plane elasticity in case that the configu-
ration is independent of the quasiperiodic axis. In this case, the plane elasticity is
a classical elasticity problem and its solution is well known, whereas the anti-plane
elasticity is a problem concerned with the quasiperiodic structure, which is our main
concern. This decomposition leads to great simplifications for the solution.
We now consider elasticity of two-dimensional quasicrystals, which is mathe-
matically much more complicated than that of one-dimensional quasicrystals. The
decomposition procedure developed in Chapter 5 hints that the elasticity of two-
dimensional quasicrystals may somehow also be made of a decomposition for a wide
range of applications. In this way the problem can be greatly simplified and it is
helpful to solve the boundary value problems by analytic method.
Two-dimensional quasicrystals so far observed cover four systems, i.e., those
involving five-fold, eight-fold, ten-fold and twelve-fold symmetries, named the pen-
tagonal, octagonal, decagonal and dodecagonal respectively, and among them there
are different Laue classes. The importance of two-dimensional quasicrystals is only
less than that of the three-dimensional icosahedral quasicrystals. To date, among
the almost 200 quasicrystals, there are 100 icosahedral quasicrystals, and 70 decago-
nal quasicrystals. These two kinds of quasicrystals constitute the majority of the
material. The elasticity of icosahedral quasicrystals will be discussed in Chapter 9.
We will, at first, give a brief description of the point groups and Laue classes of the
two-dimensional quasicrystals. There are 31 kinds of crystallographic point groups
and 26 kinds of noncrystallographic point groups of the quasicrystals, the former
will not be discussed here and we focus on the latter, which is further divided into
eight Laue classes and four quasicrystal systems observed so far, as listed in Table
6.0-1.
Like that in the one-dimensional quasicrystals, the phonon field of two-dimensional
quasicrystals is transversally isotropic. If we take the xy- plane (or x1 x2 - plane) as

68 Chapter 6 Elasticity of two-dimensional quasicrystals and simplification
the quasiperiodic plane, and axis z (or x3 ) as the periodic axis, then xy- plane is
the elasticity isotropic plane, within which the elastic constants are
C1111 = C2222 = C11 ,
C1122 = C12 ,
C1212 = C1111 − C1122 = C11 − C12 = 2C66 .
Table 6.0-1 Quasicrystal systems, Laue classes and point groups of

two-dimensional quasicrystal
Quasicrystal systems Laue classes Point groups
Pentagonal 11 5, 5̄
12 5m, 52, 5̄m
Decagonal 13 10, 10, 10/m
14 10mm, 1022, 10m2, 10/mmm
Octagonal 15 8, 8̄, 8/m
16 8mm, 822, 8̄m2, 8/mmm
Dodecagonal 17 12, 12, 12/m
18 12mm, 1222, 12m2, 12/mmm
This shows that C66 is not independent. Other independent elastic constants are
out of the plane, i.e.,
C2323 = C3131 = C44 ,
C1133 = C2233 = C13 ,
C3333 = C33 .
which are listed in Table 6.0-2.
Table 6.0-2 Phonon elastic constants in two-dimensional quasicrystals (Cijkl )

11 22 33 23 31 12
11 C11 C12 C13 0 0 0
22 C12 C11 C13 0 0 0
33 C13 C13 C33 0 0 0
23 0 0 0 C44 0 0
31 0 0 0 0 C66 0
12 0 0 0 0 0 C66
The relevant phason elastic constants and phonon-phason coupling elastic con-
stants are listed in Tables 6.0-3∼6.0-6.
Table 6.0-3 Phason elastic constants for Laue class 11 (Kijkl )

11 22 23 12 13 21
11 K1 K2 K7 0 K6 0
22 K2 K1 K7 0 K6 0
23 K7 K7 K4 K6 0 −K6
12 0 0 K6 K1 −K7 −K2
13 K6 K6 0 −K7 K4 K7
21 0 0 −K6 −K2 K7 K1
Chapter 6 Elasticity of two-dimensional quasicrystals and simplification 69
Table 6.0-4 Phonon-phason coupling elastic

constants for Laue class 11 (Rijkl )
XXX
XXXwij 11 22 23 12 13 21
εij X X
X
11 R1 R1 R6 R2 R5 −R2
12 −R1 −R1 −R6 −R2 −R5 R2
33 0 0 0 0 0 0
23 R4 −R4 0 R3 0 R3
31 −R3 R3 0 R4 0 R4
12 R2 R2 −R5 −R1 R6 R1
For Laue class 12:

If 2//x1 , m⊥x1 : K6 = R2 = R3 = R6 = 0.
If 2//x2 , m⊥x2 :K7 = R2 = R4 = R6 = 0.
For Laue class 13:
K6 = K7 = R3 = R4 = R5 = R6 = 0.
For Laue class 14:
K6 = K7 = R2 = R3 = R4 = R5 = R6 = 0.
Table 6.0-5 The phason elastic constants for Laue class 15 (Kijkl )
11 22 23 12 13 21
11 K1 K2 0 K5 0 K5
22 K2 K1 0 −K5 0 −K5
23 0 0 K4 0 0 0
12 K5 −K5 0 K 0 K3
13 0 0 0 0 K4 0
21 K5 −K5 0 K3 0 K
K = K1 + K2 + K3 .
Table 6.0-6 The phonon-phason coupling elastic constants

for Laue class 15 (Rijkl )
```
``` wij
``` 11 22 23 12 13 21
εij `
`
11 R1 R1 0 R2 0 −R2
22 −R1 −R1 0 −R2 0 R2
33 0 0 0 0 0 0
23 0 0 0 0 0 0
31 0 0 0 0 0 0
12 R2 R2 0 −R1 0 R1
For Laue class 16: K6 = R2 = 0.

For Laue class 17: the constants Kijkl are the same as table 6.0-5 and Rijkl = 0.
For Laue class 18: the constants Kijkl are the same as those of Laue 16 and Rijkl = 0.
The experimental measurement of the material data for different quasicrystals is

essential in studying the elasticity. We here list some data for decagonal quasicrystals
in Tables 6.0-7∼6.0-9.
Table 6.0-7 Phonon elastic constants of decagonal quasicrystal (the unit of

Cij , B, G is GPa) by experimental measurement[1]
Alloy C11 C33 C44 C12 C13 B G ν
Al-Ni-Co 234.33 232.22 70.19 57.41 66.63 120.25 79.98 0.228
in which B is the bulk modulus, G the shear modulus, and ν the Poisson’s ratio,
respectively.
The phason elastic constants, for a decagonal Al-Ni-Co quasicrystal, anisotropic
diffuse scattering has been observed in synchrotron X-ray diffraction measurements[2] .
It has been shown that the measurement can attributed the phason elastic constants,
although no quantitative evaluation on K1 and K2 . The Monto-Carlo simulation
was used to evaluate the phason elastic constants, e.g. given by Table 6.0-8[3] .
Table 6.0-8 Phason elastic constants of an Al-Ni-Co decagonal quasicrystal

by Monte-Carlo simulation[3]
Alloy K1 /(1012 dyn/cm2 ) K2 /(1012 dyn/cm2 )
Al-Ni-Co 1.22 0.24
2
where 1GPa = 1010 dyn/cm . It should be noted that the values given by the Monte-
Carlo simulation are subjected to verification.
Recently the experimental measurement for phonon-phason coupling elastic con-
stants for decagonal quasicrystals have been achieved, the results are listed in Table
6.0-9.
Table 6.0-9 Coupling elastic constants for point group 10, 10, Al-Ni-Co
decagonal quasicrystals[4]
Alloy R1 /GPa |R2 |/GPa
Al-Ni-Co −1.1 <0.2
Tables 6.0-2∼6.0-6 indicate that both the phonon elasticity of two-dimensional

quasicrystals, and the phason and phonon-phason coupling elasticity are three-
dimensional, not two-dimensional. In general, they can not be reduced to two-
dimensional problems. In this case there are 29 field variables and 29 field equations,
it is very difficult to obtain their analytic solution.
In practice there is often a case in which the configuration of a system is uniform
along a periodic axis, say the axis z, both physically and geometrically, so that the
field variables are free from the coordinate z, i.e.,
∂
= 0. (6.0-1)
∂z
6.1 Basic equations of plane elasticity of two-dimensional quasicrystals: point . . . 71
Under this condition, the elasticity of two-dimensional quasicrystals can be decom-

posed into plane elasticity of quasicrystal and anti-plane elasticity of crystal, where
the latter is a pure phonon or classical elasticity, whose governing equation and
boundary conditions are decoupled with the plane elasticity of quasiperiodic struc-
ture, which can be treated separately.
Uniform configuration along the z-axis, in physical and geometrical sense, repre-
sents a number of useful physical problems, for example, a straight dislocation line
or a Griffith crack along the direction, and so on. This shows that the decomposition
procedure presents important physical meaning.
In what follows, we derive the final governing equations for the plane and anti-
plane elasticity of the four different systems, by introducing some displacement or
stress potential functions. We will see that the field equations are dramatically
simplified, this is helpful to solve them by analytic method, Furthermore we use
the Fourier analysis or complex variable function method to construct analytic so-
lutions of some boundary value problems of the equations. But the mathematical
methods and numerous exact solutions for different boundary value problems will
be performed in Chapters 7 and 8 in detail.
6.1 Basic equations of plane elasticity of two-dimensional qua-

sicrystals: point groups 5m and 10mm in five- and ten-
fold symmetries
Following the introduction given in Chapter 1, it is understood that the sign 10 in the
symbol 10mm represents a ten-fold rotation symmetry and that m signifies a mirror
symmetry. Thus, the notation 10mm therefore means a combined operation of one
rotation symmetry and two mirror symmetries. For other signs the explanation is
similar.
The point groups 5m and 10mm quasicrystals are quasicrystals with five-fold
and ten-fold symmetries repectively, whose plane elasticity has been studied earlier,
see e.g. in reviews given by Fan and Mai[5] or by Hu et al[6] or by Bohsung et
al[7] . The first solution of dislocation of pentagonal quasicrystals was given by De
and Pelcovis[8] , according to our understanding, it is a solution for point group 5m
two-dimensional quasicrystal. The formulation and the solution method developed
here and herein are different from those given by Ref. [8].
The diffraction pattern and relevant Penrrose tiling are shown in Figs. 6.1-1 and
6.1-2. The Penrose tiling of pentagonal quasirystals has been shown in Chapter 3.
Quasicrystals with these symmetries belong to the class of two-dimensional qua-
sicrystals, i.e., the atomic arrangement is quasiperiodic in a plane, and periodic in
the third direction. For clarity, the quasicrystals can be defined as being generated
Fig. 6.1-1 Diffraction pattern of two-dimensional quasicrystal with ten-fold symmetry
Fig. 6.1-2 Penrose tiling of ten-fold symmetry quasicrystals
by stacking from planar quasiperiodic structures of five-fold or ten-fold symmetry

along the third symmetry axis. Here the quasiperiodic plane is the xy- plane, and the
five-fold or ten-fold rotational axis is the z-axis, which is the axis of symmetry. Due
to the presence of scaling of incommensurate length in the quasiperiodic plane, it
leads to the additional degree of freedom that does not exist in conventional crystals,
named the phason field w.
Although two-dimensional quasicrystal can be understood as an stacked pla-
nar quasiperiodic structure along the periodic symmetrical axis, yet it is three-
dimensional in elasticity, and is different from the plane problem in classical elas-
ticity. And it can be decomposed into plane and anti-plane elasticity in special
cases. Here we consider only the plane elasticity, because the anti-plane elasticity is
a classical one and independent of the phason variables.
Tables 6.0-2∼6.0-6 listed all the elastic constants. Consider a plane in two-
dimensional quasicrystals, and assume that it is perpendicular to the periodic sym-
metrical axis (say, axis z). In this case,
u = (ux , uy , uz ), w = (wx , wy , 0),
so the strains are wzz = wzx = wxz = wzy = wyz = 0. Assumption (6.0-1) leads to
εzz = εxz = εyz = 0, and Table 6.0-2 is simplified to the following Table 6.1-1 for
the phonon elastic constants for the plane elasticity.
Table 6.1-1 Phonon elastic constants for plane

elasticity of two-dimensional quasicrystals
11 22 12 21
11 C11 C12 0 0
22 C12 C11 0 0
12 0 0 C66 C66
21 0 0 C66 C66
The phonon elastic constants listed in Table 6.1-1 can be expressed in the fourth-
order tensor
Cijkl = Lδij δkl + M (δjk δjl + δil δjk ), i, j, k, l = 1, 2, (6.1-1)

C11 − C12
L = C12 , M= = C66 . (6.1-2)
2
There are only two independent phonon elastic constants.
The data given by Table 6.1-1 and relation by equations (6.1-1), (6.1-2) hold for
all two-dimensional quasicrystals of plane elasticity.
In the plane elasticity for two-dimensional quasicrystals with point groups 5m,
52, 5m, 10mm, 1022, 10mm and 10/mmm, Table 6.0-3 is simplified to Table 6.1-2
below.
Table 6.1-2 Phason elastic constants for plane elasticity for point group
10mm quasicrystals
11 22 12 21
11 K1 K2 0 0
22 K2 K1 0 0
12 0 0 K1 −K2
21 0 0 −K2 K1
This means
K1111 = K2222 = K2121 = K1 ,
(6.1-3)
K1122 = K2211 = −K2112 = −K1221 = K2
and other Kijkl = 0, and the expression of them by tensor of four rank is
Kijkl = K1 δik δjl + K2 (δij δkl − δil δjk ), i, j, k, l = 1, 2. (6.1-4)
Table 6.0-4 in the present case is simplified into Table 6.1-3 as follows.
Table 6.1-3 Phonon-phason coupling elastic constants for plane elasticity

hhh
hhh wij
h hhh 11 22 12 21
εij hh
h
11 R R 0 0
22 −R −R 0 0
12 0 0 −R R
21 0 0 −R R
This shows that
R1111 = R1122 = −R2222 = R1221 = R2121 = −R1212 = −R2211 = −R2121 = R

(6.1-5)
or
Rijkl = R(δi1 − δi2 )(δij δkl − δik δjl + δiljk ), i, j, k, l = 1, 2. (6.1-6)
One can find that for plane elasticity point group 5m has the same elastic constants
with point group 10mm, so they can be discussed in the same line.
The definition of strain tensor are as given in Chapter 4, i.e. ,

1 ∂ui ∂uj ∂wi
εij = + , wij = . (6.1-7)
2 ∂xj ∂xi ∂xj
In Chapter 4 it was seen that the stress, strain and elastic constant tensors can be
expressed by matrices. The above mentioned elastic constants may be denoted by
matrix[CKR]. For the present case, the strain vector defined by (4.4-12) is simplified
into
[εij wij ] = [ε11 ε22 ε12 ε21 w11 w22 w12 w21 ] (6.1-8)
and [CKR] is
⎡ ⎤
L + 2M L 0 0 R R 0 0
⎢ L L + 2M 0 0 −R −R 0 0 ⎥
⎢ ⎥
⎢ 0 0 M M 0 0 −R R ⎥
⎢ ⎥
⎢ 0 0 M M 0 0 −R R ⎥
[CKR] = ⎢
⎢
⎥
⎥
⎢ R −R 0 0 K 1 K2 0 0 ⎥
⎢ R −R 0 0 K2 K1 0 0 ⎥
⎢ ⎥
⎣ 0 0 −R −R 0 0 K1 −K2 ⎦
0 0 R R 0 0 −K2 K1
⎡ ⎤
L + 2M L 0 0 R R 0 0
⎢ L + 2M 0 0 −R −R 0 0 ⎥
⎢ ⎥
⎢ M M 0 0 −R R ⎥
⎢ ⎥
⎢ M 0 0 −R R ⎥
=⎢ ⎥ . (6.1-9)
⎢ K1 K2 0 0 ⎥
⎢ ⎥
⎢ (symmetry) K1 0 0 ⎥
⎢ ⎥
⎣ K1 −K2 ⎦
K1
The free energy (or strain energy density) is
1 1
F = Lεii εii + M εij εij + K1 wij wij + K2 (wxx wyy − wxy wxy )
2 2
+ R [(εxx − εyy )(wxx + wyy ) + 2εxy (wxy − wyx )] (6.1-10)
By (6.1-9) and (4.5-3) or by (6.1-10) and (4.5-1), the generalized Hooke’s law for
plane elasticity of point group 10mm quasicrystals of ten-fold symmetry is
⎧
⎪ σxx = L(εxx + εyy ) + 2M εxx + R(wxx + wyy ),
⎪
⎪
⎪
⎪
⎪
⎪ σyy = L(εxx + εyy ) + 2M εyy − R(wxx + wyy ),
⎪
⎪
⎪
⎪
⎨ σxy = σyx = 2M εxy + R(wyx − wxy ),
⎪
Hxx = K1 wxx + K2 wyy + R(εxx − εyy ), (6.1-11)
⎪
⎪
⎪
⎪ Hyy = K1 wyy + K2 wxx + R(εxx − εyy ),
⎪
⎪
⎪
⎪
⎪
⎪ Hxy = K1 wxy − K2 wyx − 2Rεxy ,
⎪
⎪
⎩
Hyx = K1 wyx − K2 wxy + 2Rεxy .
In addition, for anti-plane elasticity there is

σxz = 2M εxz ,
(6.1-12)
σyz = 2M εyz .
Monograph [9] pointed first out that (6.1-9)∼(6.1-11) hold for the plane elasticity
for both point group 5m and point group 10mm, this is identical to the argument of
Ref. [6].
Equations (6.1-11) are the physical basis of elasticity of point groups 5m and
10mm quasicrystals. The geometry (or kinetics) basis of the subject is equation
(6.1-7). Another necessary basis comes from statics, i.e.,
⎧
⎪
⎨ ∂x + ∂y = 0, + = 0,
∂x ∂y
(6.1-13)
⎪
⎪ ∂Hxx ∂Hxy ∂Hyx ∂Hyy
⎩ + = 0, + = 0.
∂x ∂y ∂x ∂y
In addition, for the anti-plane elasticity there is

∂σzx ∂σzy
+ = 0. (6.1-14)
∂x ∂y
Here the body force density is omitted.
From (6.1-7),(6.1-11) and (6.1-13) ,we found that there are 18 field variables,
i.e., four displacements ux , uy , wx , wy , seven strains εxx , εyy , εxy = εyx , wxx ,
wyy , wxy , wyx and seven stresses σxx , σyy , σxy = σyx , Hxx , Hyy , Hxy , Hyx . And
the number of corresponding field equations is also 18, including four equations
of statics, seven stress-strain relations and seven equations of deformation geometry.
The number of field equations is equal to that of field variables. This means that the
mathematical presentation of the problem is consistent and solvable under appropri-
ate boundary conditions. In addition, (6.1-7), (6.1-12) and (6.1-14) give the descrip-
tion of anti-plane pure classical (phonon) elasticity. De and Pelcovits[10] solved the
equations for dislocation problem by using the Green function method and iteration
procedure. Ding et al[11] solved the similar problem in terms of Fourier transforms
and the Green function method. Taking different way from either De and Pelcovits
or Ding et al, we at first reduce the 18 equations into a single partial differential of
higher order by introducing some displacement potentials or stress potentials, which
enable us to solve the problem by the Fourier transform method and complex vari-
able function method. In the following we derive the final governing equations and
fundamental solutions for various quasicrystal systems. And the applications of the
theory and methodology will be given in detail in Chapters 7 and 8 for dislocation
and crack problems of quasicrystals, respectively.
6.2 Simplification of the basic equation set: displacement po-

tential function method
The number of basic equations as given in the previous section is too large and it is
very hard to solve them directly. In mathematical physics, a conventional procedure
is to reduce the number of field equations. In the classical theory of elasticity, there
is a way by introducing the so-called displacement or stress potential functions to
realize the purpose, and the procedure is called the displacement potential method
or stress potential method, respectively. In this section we eliminate the stress and
strain components in the basic equations, and obtain some equilibrium equations
expressed by displacement components, thus the strain compatibility is automati-
cally satisfied. Furthermore, by introducing a displacement potential function, the
final governing equations become a single quadruple harmonic equation concerning
the displacement potential function, thus this is the so-called displacement potential
method. In the next section, we will introduce the stress potential method, there
6.2 Simplification of the basic equation set: displacement potential function method 77
the governing equation becomes a single quadruple harmonic equation concerning

the stress potential function that would be the so-called stress potential method. In
this way the huge number of equations involving elasticity can be simplified very
much.
Substituting (6.1-7) into (6.1-11) and then into (6.1-12), we obtain
⎧ 2
⎪ ∂ ∂ wx ∂ 2 wy ∂wx
⎪
⎪ M ∇ 2
u + (L + M ) ∇ · u + R + 2 − = 0,
⎪
⎪
x
∂x ∂x2 ∂x∂y ∂y 2
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂ ∂ wy ∂ 2 wx ∂ 2 wy
⎪
⎪ ∇ 2
∇ · u − −
⎪
⎨ M u y + (L + M ) + R 2 = 0,
∂y ∂x2 ∂x∂y ∂y 2
2 (6.2-1)
⎪
⎪ ∂ ux ∂ 2 uy ∂ 2 ux
⎪
⎪ ∇ 2
− −
⎪
⎪ K1 w x + R
∂x2
2
∂y 2
= 0,
⎪
⎪ ∂x∂y
⎪
⎪ 2
⎪
⎪
⎪
⎪ ∂ uy ∂ 2 ux ∂ 2 uy
⎩ K1 ∇2 wy + R + 2 − =0
∂x2 ∂x∂y ∂y2
in which
∂2 ∂2 ∂ux ∂uy
∇2 = 2
+ 2, ∇·u= + .
∂x ∂y ∂x ∂y
Equations (6.2-1) are actually the displacement equilibrium equations of the

plane elasticity of point group 5m with five-fold symmetry or point group 10mm
with ten-fold symmetry, as mentioned in the previous section. Here there are only 4
displacement components ux , uy , wx , wy , i.e., the number of field variables are only
4, and the number of order of equations is 8.
Observation for the first two equations of (6.2-1) suggests that by introducing
the unknown functions ϕ(x, y) and ψ(x, y) as follows:
⎧
⎪
⎪ ∂2ϕ ∂2ψ ∂2ψ
⎪
⎪ ux = (L + M ) + M 2 + (L + 2M ) 2 ,
⎪
⎪ ∂x∂y ∂x ∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂2ϕ ∂2ϕ ∂2ψ
⎪
⎪ u = − (L + 2M ) 2 + M 2 + (L + M ) ,
⎪
⎨ y ∂x ∂y ∂x∂y
(6.2-2)
⎪
⎪ 2
⎪
⎪ M (L + 2M ) ∂ ϕ ∂2ψ ∂2ψ
⎪
⎪ wx = − 2 + − ,
⎪
⎪ R ∂x∂y ∂x2 ∂y2
⎪
⎪
⎪
⎪
⎪
⎪ M (L + 2M ) ∂ 2 ϕ ∂ 2 ϕ ∂ 2ψ
⎪
⎩ wy = − − 2 .
R ∂x2 ∂y 2 ∂x∂y
then the first two equations are automatically satisfied already. Substituting (6.2-2)
into the last two equations in (6.2-1), there follows the equations
⎧
⎪
⎪ ∂2ϕ ∂2 ∂2
⎪
⎨ (αΠ 1 + βΠ 2 ) + αΠ1 − βΠ2 ψ = 0,
∂x∂y ∂y2 ∂x2
(6.2-3)
⎪
⎪ ∂2 ∂2 ∂2ψ
⎪
⎩ αΠ2 2 − βΠ1 2 ϕ + (αΠ2 + βΠ1 ) = 0.
∂x ∂y ∂x∂y
where
∂2 ∂2 ∂2 ∂2
Π1 = 3 2
− 2, Π2 = 3 2
− 2, (6.2-4)
∂x ∂y ∂y ∂x
⎧
⎨ α = R(L + 2M ) − ωK1 , β = RM − ωK1 ,
(6.2-5)
⎩ ω = M (L + 2M ) .
R
Equation (6.2-3) is much simpler compared with (6.2-1), still it can be simplified by
letting

∂2 ∂2 ∂2F
ϕ = βΠ2 2 − αΠ1 2 F, ψ = (αΠ1 + βΠ2 ) . (6.2-6)
∂x ∂y ∂x∂y
in which F (x, y) is any function, then the first equation of (6.2-3) is satisfied. Substi-
tuting (6.2-6) into the second one of equations (6.2-3), after manipulation, it reduces
to
∇2 ∇2 ∇2 ∇2 F = 0. (6.2-7)
This is the final governing equation of the plane elasticity of point groups 5m and
10mm quasicrystals. We call F (x, y) as a displacement potential function, or simply
displacement potential. The equation (6.2-7) is a quadruple harmonic equation, the
order of which is much higher than that in the classical elasticity, where that is
biharmonic equation.
All of displacement and stress components can be expressed by potential function
F (x, y) as follows:
⎧
⎪
⎪ ∂2 2
⎨ ux = [M αΠ1 + (L + 2M )βΠ2 ] ∂x∂y ∇ F,
⎪
(6.2-8)
⎪
⎪ ∂2 ∂2
⎪
⎩ uy = M αΠ1 2 − (L + 2M )βΠ2 2 ∇2 F,
∂y ∂x
⎧
⎪
⎪ ∂2
⎨ wx = ω(α − β)Π1 Π2 ∂x∂y F,
⎪
(6.2-9)
⎪
⎪ 2 ∂
2
2 ∂
2
⎪
⎩ wy = −ω αΠ1 2 + βΠ2 2 F,
∂y ∂x
6.3 Simplification of the basic equations set: stress potential function method 79
⎧
⎪ ∂3 2
⎪
⎪ σ = 2M (L + M )αΠ 1 ∇ F,
⎪
⎪
xx
∂y 3
⎪
⎪
⎪
⎪
⎨
∂3
σyy = 2M (L + M )αΠ1 2 ∇2 F, (6.2-10)
⎪
⎪ ∂x ∂y
⎪
⎪
⎪
⎪
⎪
⎪ 3
⎪
⎩ σxy = σyx = −2M (L + M )αΠ1 ∂ ∇2 F,
∂x∂y 2
⎧
⎪ ∂ 2 2 2 ∂ 2 ∂2 2 ∂2
⎪
⎪ Hxx = αβ ∇ ∇ ∇ F + ω(K1 − K2 ) αΠ1 2 + βΠ2 2 F,
⎪
⎪ ∂y ∂y ∂y ∂x
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ ∂ ∂3
⎪
⎪ Hyy = αβ ∇2 ∇2 ∇2 F − ω(K1 − K2 )(α − β)Π1 Π2 2 F,
⎨ ∂y ∂x ∂y
⎪
⎪
⎪ ∂ 2 2 2 ∂ 2 2
⎪ 2 ∂ 2 ∂
⎪
⎪ H xy = −αβ ∇ ∇ ∇ F − ω(K 1 − K 2 ) αΠ1 + βΠ 2 F,
⎪
⎪ ∂x ∂x ∂y 2 ∂x2
⎪
⎪
⎪
⎪
⎪
⎪ 3
⎪
⎩ Hyx = αβ ∂ ∇2 ∇2 ∇2 F + ω(K1 − K2 )(α − β)Π1 Π2 ∂ F.
∂x ∂x∂y 2
(6.2-11)
[12]
Li and Fan have suggested the approach, and the practice shows that it is very
effective. In the next two chapters many applications of the approach will be given.
In addition (6.1-12) and (6.1-14) yield the final governing equation for anti-plane
elasticity
∇2 uz = 0. (6.2-12)
It is obvious that the anti-plane problem is decoupled with the plane one.
6.3 Simplification of the basic equations set: stress potential

function method
The stress potential method has been widely used in classical theory of elasticity.
We have extended it to elasticity of quasicrystals [13], [14].
From (6.1-7) the strain compatibility equations
⎧ 2
⎪ ∂ εxx ∂ 2 εyy ∂ 2 εxy
⎪
⎪
⎨ ∂y 2 + ∂x2 = 2 ∂x∂y ,
(6.3-1)
⎪
⎪
⎪
⎩ ∂wxy = ∂wxx , ∂wyx
=
∂wyy
.
∂x ∂y ∂y ∂x
The strain components εij and wij can be expressed by the stress components σij
and Hij , i.e.,

⎧
⎪ 1 1
⎪
⎪ εxx = (σxx + σyy ) + [(K1 + K2 )(σxx + σyy ) − 2R(Hxx + Hyy )],
⎪
⎪ 4(L + M ) 4C
⎪
⎪
⎪
⎪
⎪
⎪ 1
(σxx + σyy ) −
1
[(K1 + K2 )(σxx + σyy ) − 2R(Hxx + Hyy )],
⎪ εyy =
⎪
⎪
⎪ 4(L + M ) 4C
⎪
⎪
⎪
⎪ 1
⎪
⎪ εxy = εyx = [(K1 + K2 )σxy − R(Hxy + Hyx )],
⎪
⎪
⎪
⎪
C
⎪
⎨ 1 1
wxx = (Hxx − Hyy ) + [M (Hxx + Hyy ) − R(σxx − σyy )],
⎪ 2(K 1 − K 2 ) 2C
⎪
⎪
⎪
⎪ 1 1
⎪
⎪ wyy = − (Hxx − Hyy ) + [M (Hxx + Hyy ) − R(σxx − σyy )],
⎪
⎪ −
⎪
⎪
2(K 1 K 2 ) 2C
⎪
⎪
⎪
⎪ 1 (R2 − M K1 )Hxy + (R2 − M K2 )Hyx
⎪
⎪ w = Rσ − ,
⎪
⎪ xy xy
K1 − K 2
⎪
⎪
C
⎪
⎪
⎪
⎪ 1 (R2 − M K1 )Hxy + (R2 − M K2 )Hyx
⎪
⎩ wyx = −Rσxy − .
C K1 − K 2
(6.3-2)
with
C = M (K1 + K2 ) − 2R2 . (6.3-3)
If substituting (6.3-2) into (6.3-1), one can find the strain compatibility equations
expressed by stress components (because they are too long to be listed, but we will
give them in Section 6.9). In addition, there are equilibrium equations
⎧
⎪
⎨ ∂x + ∂y = 0, + = 0,
∂x ∂y
(6.3-4)
⎪
⎩ + = 0, + = 0.
∂x ∂y ∂x ∂y
So that we have 7 equations in total in which there are 3 compatibility equations ex-
pressed by stress components referring to (6.9-2a)∼(6.9.2c) and 4 equilibrium equa-
tions, and the number of the unknown functions is also 7,i.e., σxx , σyy , σxy = σyx ,
Hxx , Hyy , Hxy , Hyx . The equation set is closed and solvable.
If introducing the stress functions ϕ, ψ1 and ψ2 , such as
⎧
⎪ ∂2ϕ ∂2ϕ ∂2ϕ
⎪
⎪ σ = , σ = , σ = σ = − ,
⎨ xx
∂y 2 yy
∂x2 xy yx
∂x∂y
(6.3-5)
⎪
⎪ ∂ψ1 ∂ψ1 ∂ψ2 ∂ψ2
⎪ xx
⎩ H = , H xy = − , H yx = − , H yy = .
∂y ∂x ∂y ∂x
then the equilibrium equations (6.3-4) are automatically satisfied. Substituting

(6.3-5) into deformation compatibility equations expressed by stress components
6.4 Plane elasticity of point group 5, 5 pentagonal and point group 10, 10 . . . 81
yields
⎧
⎪
⎪ 1 2 2 K1 + K 2 2 2 R ∂ ∂
⎪
⎪ ∇ ∇ ϕ+ ∇ ∇ ϕ+ Π1 ψ1 − Π2 ψ2 = 0,
⎪
⎪ 2C(L + M ) 2C C ∂y ∂x
⎪
⎪
⎨ C

∂
+ M ∇2 ψ1 + R Π1 ϕ = 0,
⎪ K1 − K2
⎪ ∂y
⎪
⎪
⎪
⎪
⎪
⎪ C ∂
⎩ + M ∇2 ψ2 − R Π2 ϕ = 0.
K1 − K2 ∂x
(6.3-6)
in which Π1 and Π2 are defined by (6.2-4), C is given by (6.3-3). By now, the
numbers of equations and unknown functions have been reduced to 3.
Now introducing a new unknown function G(x, y), such as
⎧
⎪
⎪ ϕ = D ∂ Π1 ∇2 G,
⎪
⎪
⎪
⎪ ∂y
⎪
⎪
⎪
⎪ 1
⎪ 2 2 2 2
⎨ ψ1 = − R (M K1 − R2 )[(L + 2M )(K1 − K2 ) − 2R ]∇ ∇ ∇ G
⎪
(6.3-7)
⎪
⎪ ∂2
⎪
⎪ + (L + M )(K1 − K2 )R Π1 Π2 G,
⎪
⎪ ∂x∂y
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂
⎩ ψ2 = (L + M )(K1 − K2 )R Π1 Π2 G.
∂x∂y
If
∇2 ∇2 ∇2 ∇2 G = 0, (6.3-8)
then equations (6.3-6) are satisfied, in which
D = 2(M K1 − R2 )(L + M ). (6.3-9)
At the same time the equation (6.2-12) holds for anti-plane elasticity too.
6.4 Plane elasticity of point group 5, 5 pentagonal and point

group 10, 10 decagonal quasicrystals
For plane elasticity, the five-fold and ten-fold symmetries quasicrystals of point
groups 5, 5 and 10, 10 are different in elasticity with that of point groups 5m and
10mm, the difference lies in only the phonon-phason coupling elastic constants, in
which the former has two coupling elastic constants R1 and R2 rather than one
constant R. Thus we have
Rijkl = R1 (δi1 − δi2 )(δij δkl − δik δjl + δil δjk )

+ R2 [(1 − δij )δkl + δij (δi1 − δi2 )(δk1 δl2 − δk2 δl1 )], i, j, k, l = 1, 2. (6.4-1)
Apart from this, the phonon and phason elastic constants of point group 5,5 and
point group 10, 10 quasicrystals are the same as those of point groups 5m and 10mm.
The corresponding elastic constant matrix is
⎡ ⎤
L + 2M L 0 0 R1 R1 R2 −R2
⎢ ⎥
⎢ L L + 2M 0 0 −R1 −R1 −R2 R2 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 M M R2 R2 −R1 R1 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 M M R2 R2 −R1 R1 ⎥
[CKR] = ⎢ ⎢ ⎥
⎥
⎢ R1 −R1 R2 R2 K1 K2 0 0 ⎥
⎢ ⎥
⎢
⎢ R1 −R1 R2 R2 K2 K1 0 0 ⎥ ⎥
⎢ ⎥
⎢
⎣ R2 −R2 −R1 −R1 0 0 K1 −K2 ⎥ ⎦
−R2 R2 R1 R1 0 0 −K2 K1
⎡ ⎤
L + 2M L 0 0 R1 R1 R2 −R2
⎢ ⎥
⎢ L + 2M 0 0 −R1 −R1 −R2 R2 ⎥
⎢ ⎥
⎢ ⎥
⎢ M M R2 R2 −R1 R1 ⎥
⎢ ⎥
⎢ ⎥
⎢ M R2 R2 −R1 R1 ⎥
=⎢
⎢
⎥.
⎥
⎢ (symmetry) K1 K2 0 0 ⎥
⎢ ⎥
⎢ 0 ⎥
⎢ K1 0 ⎥
⎢ ⎥
⎢ −K2 ⎥
⎣ K1 ⎦
K1
(6.4-2)
By this elastic constant matrix, the stress-strain relation can be written as
⎧
⎪ σxx = L(εxx + εyy ) + 2M εxx + R1 (wxx + wyy ) + R2 (wxy − wyx ),
⎪
⎪
⎪
⎪
⎪
⎪ σyy = L(εxx + εyy ) + 2M εyy − R1 (wxx + wyy ) − R2 (wxy − wyx ),
⎪
⎪
⎪
⎪
⎪
⎪ σ = σyx = 2M εxy + R1 (wyx − wxy ) + R2 (wxx + wyy ),
⎨ xy
Hxx = K1 wxx + K2 wyy + R1 (εxx − εyy ) + 2R2 εxy , (6.4-3)
⎪
⎪
⎪
⎪
⎪
⎪ Hyy = K1 wyy + K2 wxx + R1 (εxx − εyy ) + 2R2 εxy ,
⎪
⎪
⎪
⎪ Hxy = K1 wxy − K2 wyx − 2R1 εxy + R2 (εxx − εyy ),
⎪
⎪
⎪
⎩
Hyx = K1 wyx − K2 wxy + 2R1 εxy − R2 (εxx − εyy ).
In addition, the stresses σij and Hij satisfy the same equilibrium equations as
(6.1-13).
Substituting (6.1-7) into (6.4-3), then into (6.1-12) leads to the equilibrium equa-
6.4 Plane elasticity of point group 5, 5 pentagonal and point group 10, 10 . . . 83
tions expressed by displacements as

⎧ 2
⎪
⎪ 2 ∂ ∂ wx ∂ 2 wy ∂ 2 wx
⎪
⎪ M ∇ u x + (L + M ) ∇ · u + R 1 + 2 −
⎪
⎪ ∂x ∂x2 ∂x∂y ∂y 2
⎪
⎪
⎪
⎪ ∂ 2 wy ∂ 2 wx ∂ 2 wy
⎪
⎪
⎪
⎪ −R2 −2 − = 0,
⎪
⎪ ∂x 2 ∂x∂y ∂y 2
⎪
⎪ 2
⎪
⎪ ∂ 2 wx ∂ 2 wy
⎪
⎪ 2 ∂ ∂ wy
⎨ M ∇ uy + (L + M ) ∂y ∇ · u + R1 ∂x2 − 2 ∂x∂y − ∂y 2
⎪
2
⎪
⎪ ∂ wx ∂ 2 wy ∂ 2 wx
⎪
⎪ +R + 2 − = 0,
⎪
⎪ 2
∂x2 ∂x∂y ∂y2
⎪
⎪
⎪
⎪ 2 2
⎪
⎪ ∂ ux ∂ 2 uy ∂ 2 ux ∂ uy ∂ 2 ux ∂ 2 uy
⎪ 2
⎪ K1 ∇ wx + R1 ∂x2 − 2 ∂x∂y − ∂y 2 + R2 ∂x2 + 2 ∂x∂y − ∂y2 = 0,
⎪
⎪
⎪
⎪
⎪
⎪ 2 2
⎪
⎪ ∂ uy ∂ 2 ux ∂ 2 uy ∂ ux ∂ 2 uy ∂ 2 ux
⎪
⎩ K1 ∇2 wy + R1 + 2 − − R 2 − 2 − = 0.
∂x2 ∂x∂y ∂y 2 ∂x2 ∂x∂y ∂y 2
(6.4-4)
The equations are similar to those of (6.2-1), the definitions on operators ∇2 and
∇· are the same there. It is obvious that equations (6.4-4) are more complex than
those of (6.2-1). We now simplify the equations in terms of the displacement poten-
tial function method.
Introducing displacement potentials ϕ(x, y) and ψ(x, y) as below:
⎧
⎪
⎪ ∂2ϕ ∂2ψ ∂2ψ
⎪
⎪ ux = (L + M ) + M 2 + (L + 2M ) 2 ,
⎪
⎪ ∂x∂y ∂x ∂y
⎪
⎪
⎪
⎪ 2 2
∂2ψ
⎪
⎪ ∂ ϕ ∂ ϕ
⎪
⎪ u y = − (L + 2M ) + M + (L + M ) ,
⎪
⎪ ∂x2 ∂y 2 ∂x∂y
⎪
⎪
2
⎪
⎪ 2 2
⎪
⎪ wx = −ω 2R1 ∂ − R2 ∂ − ∂
⎪
⎨ ϕ
∂x∂y ∂x2 ∂y 2
2 (6.4-5)
⎪
⎪ ∂ ∂2 ∂2
⎪
⎪ + R − + 2R ψ ,
⎪
⎪
1
∂x2 ∂y2
2
∂x∂y
⎪
⎪
⎪
⎪

⎪
⎪ ∂ 2
∂ 2
∂2
⎪
⎪ wy = ω R 1 − + 2R2 ϕ
⎪
⎪ ∂x2 ∂y 2 ∂x∂y
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂2 ∂ ∂2
⎪
⎩ − 2R 1 − R 2 − ψ .
∂x∂y ∂x2 ∂y 2
in which
M (L + 2M )
ω= , R2 = R12 + R22 . (6.4-6)
R2
The functions ϕ(x, y) and ψ(x, y) defined by (6.4-4) automatically satisfy the first
two of equations (6.4-4), and the substitution of formulas (6.4-6) into the second two
of equations (6.4-4) results in

⎧
⎪ ∂ ∂ ∂ ∂
⎪
⎨ (L + 2M )c2 ∂x Λ1 ϕ + M c1 ∂y Λ2 ϕ + (L + 2M )c2 ∂y Λ1 ψ − M c1 ∂x Λ2 ψ = 0,
⎪
⎪ ∂ ∂ ∂ ∂
⎩ (L + 2M )c2 Λ2 ϕ − M c1 Λ1 ϕ + (L + 2M )c2 Λ2 ψ + M c1 Λ1 ψ = 0
∂x ∂y ∂y ∂x
(6.4-7)
with ⎧
⎪ ∂ ∂
⎪
⎨ Λ1 = R1 ∂y Π1 + R2 ∂x Π2 ,
(6.4-8)
⎪
⎪ ∂ ∂
⎩ Λ2 = R1 Π2 − R2 Π1 ,
∂x ∂y
c1 = (L + 2M )K1 − R2 , c2 = M K1 − R2 . (6.4-9)
If we introduce a new function F (x, y) by means of

⎧
⎪ ∂ ∂
⎪
⎨ ϕ = −(L + 2M )c2 R ∂y Λ1 F + M c1 R ∂x Λ2 F,
(6.4-10)
⎪
⎪ ∂ ∂
⎩ ψ = (L + 2M )c2 R Λ1 F + M c1 R Λ2 F.
∂x ∂y
Then the first one of equations (6.4-7) is identically satisfied, and from the second
one we find that
∇2 ∇2 ∇2 ∇2 F = 0. (6.4-11)
The definitions of operators ∇2 and ∇· are the same as before.

All components of displacement vectors and stress tensors can be expressed by
the potential function F (x, y) as

∂ ∂
ux = R c2 Λ1 + c1 Λ2 ∇2 F, (6.4-12a)
∂x ∂y

∂ ∂
uy = R c2 Λ1 − c1 Λ2 ∇2 F, (6.4-12b)
∂y ∂x
wx = −c0 Λ1 Λ2 F, (6.4-12c)
wy = −R−1 [c2 (L + 2M )Λ21 + c1 M Λ22 ]F, (6.4-12d)

∂2
σxx = 2c0 c2 Λ1 ∇2 F, (6.4-12e)
∂y 2
∂2
σyy = 2c0 c2 Λ1 ∇2 F, (6.4-12f)
∂x2
∂2
σxy = σyx = −2c0 c2 Λ1 ∇2 F, (6.4-12g)
∂x∂y
6.5 Plane elasticity of point group 12mm of dodecagonal quasicrystals 85
∂ 2 2 2 ∂
Hxx = −c1 c2 R ∇ ∇ ∇ F + R−1 K0 [c2 (L + 2M )Λ21 + c1 M Λ22 ]F, (6.4-12h)
∂y ∂y
∂ 2 2 2 ∂
Hxy = c1 c2 R ∇ ∇ ∇ F − R−1 K0 [c2 (L + 2M )Λ21 + c1 M Λ22 ]F, (6.4-12i)
∂x ∂x
∂ 2 2 2 ∂
Hyx = −c1 c2 R ∇ ∇ ∇ F − c0 K0 Λ1 Λ2 F, (6.4-12j)
∂x ∂y
∂ 2 2 2 ∂
Hyy = −c1 c2 R ∇ ∇ ∇ F + c0 K0 Λ1 Λ2 F (6.4-12k)
∂y ∂x
with
c0 = R(L + M ), K0 = K1 − K2 . (6.4-13)
The results were given by Li and Fan[15] and Li et al[16] . Recently, Li and Fan[14] have
derived the final governing equation of elasticity of the same point groups through the
stress potential method, the resulting equation is also quadruple harmonic equation,
of course the unknown function is the stress potential. Application will be shown
in Chapter 8 for solving notch problem of point group 5, 5 and point group 10, 10
two-dimensional quasicrystals.
The frequent appearance of the quadruple harmonic equations in (6.2-7), (6.3-8)
and (6.4-11), suggests that this kind of equations is very important in theory and
practice.
6.5 Plane elasticity of point group 12mm of dodecagonal qua-

sicrystals
Point group 12mm dodecagonal quasicrystals are two-dimensional ones, the Penrose
tiling and the diffraction pattern are shown in Figs. 6.5-1 and 6.5-2 respectively. If
taking z-axis as periodic arrangement direction, and supposing that the field vari-
ables independent of the coordinate z, then the elasticity problem can be decomposed
into plane elasticity and anti-plane elasticity.
As mentioned in previous sections, the quasiperiodic plane is taken as xy-plane,
so the z-axis represents the twelve-fold symmetry axis. As in other two-dimensional
quasicrystals, the quasiperiodic plane is an elastic isotropic plane. There are two
non-zero independent elastic constants Cijke in the plane, i.e., L and M ,
C11 − C12
L = C12 , M= = C66 . (6.5-1)
2
Fig.6.5-1 Penrose tiling of dodecagonal quasicrystal
Here we have non-zero Kijkl as

K1111 = K2222 = K1 , K1122 = K2211 = K2 ,
(6.5-2)
K1221 = K2112 = K3 , K2121 = K1212 = K1 + K2 + K3
and others are zero. The results can also be expressed as
Kijkl = (K1 − K2 − K3 )(δik − δil ) + K2 δij δkl
+ K3 δil δjk + 2(K2 + K3 )(δi1 δj2 δk1 δl2 + δi2 δj1 δk2 δl1 ),
i, j, k, l = 1, 2. (6.5-3)
In addition, the phonon and phason are decoupled, i.e.,
Rijkl = 0. (6.5-4)
Fig.6.5-2 Diffraction pattern of quasicrystal of twelve-fold symmetry, in the condition of

the phason strain field being uniformly
6.5 Plane elasticity of point group 12mm of dodecagonal quasicrystals 87
The elastic constant matrix is

⎡ ⎤
L + 2M L 0 0 0 0 0 0
⎢ ⎥
⎢ ⎥
⎢ L L + 2M 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 M M 0 0 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 M M 0 0 0 0 ⎥
⎢ ⎥
[CKR] = ⎢ ⎥
⎢ 0 0 0 0 K1 K 2 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 0 0 K2 K1 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 0 0 0 0 K1 + K2 + K3 K3 ⎥
⎣ ⎦
0 0 0 0 0 0 K3 K1 + K2 + K3
⎡ ⎤
L + 2M L 0 0 0 0 0 0
⎢ L + 2M 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ M M 0 0 0 0 ⎥
⎢ ⎥
⎢ M 0 0 0 0 ⎥
=⎢ ⎥.
⎢ K1 K2 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎣ K1 + K2 + K3 K3 ⎦
K 1 + K2 + K3
(6.5-5)
The relevant free energy (or strain energy density) is

1 1 1
F = L(∇ · u)2 + M εij εij + K1 wij wij + K2 (wyx
2 2
+ wxy + 2wxx wyy )
2 2 2
1
+ K3 (wyx + wxy )2 , (6.5-6)
2
where εij and wij denote the strain tensors

1 ∂ui ∂uj ∂wi
εij = + , wij = . (6.5-7)
2 ∂xj ∂xi ∂xj
From (4.5-1) in Chapter 4 and (6.5-5) or (6.5-6), the generalized Hooke’s law for
plane elasticity of dodecagonal quasicrystals is
⎧
⎪ σxx = L(εxx + εyy ) + 2M εxx ,
⎪
⎪
⎪
⎪
⎪
⎪ σyy = L(εxx + εyy ) + 2M εyy ,
⎪
⎪
⎪
⎪
⎪
⎪ σ = σyx = 2M εxy ,
⎨ xy
Hxx = K1 wxx + K2 wyy , (6.5-8)
⎪
⎪
⎪
⎪
⎪
⎪ Hyy = K1 wyy + K2 wxx ,
⎪
⎪
⎪
⎪
⎪
⎪ Hxy = (K1 + K2 + K3 )wxy + K3 wyx ,
⎪
⎩
Hyx = (K1 + K2 + K3 )wyx + K3 wxy
and there are the equilibrium equations in the absence of the body force
⎧
⎪
⎪
⎨ ∂x + ∂y = 0, + = 0,
∂x ∂y
(6.5-9)
⎪
⎪
⎩ + = 0, + = 0.
∂x ∂y ∂x ∂y
Eliminating the stress and strain components from (6.5-7)∼(6.5-9), we obtain the
equilibrium equations expressed by displacement components as below:
⎧
⎪ ∂
⎪
⎪ M ∇2 ux + (L + M ) ∇ · u = 0,
⎪
⎪ ∂x
⎪
⎪
⎪
⎪
⎪
⎪ ∂
⎪
⎪ M ∇2 uy + (L + M ) ∇ · u = 0,
⎨ ∂y
(6.5-10)
⎪
⎪ ∂ ∂wx ∂wy
⎪
⎪ K 1 ∇ 2
w + (K 2 + K3 ) + = 0,
⎪
⎪
x
∂y ∂y ∂x
⎪
⎪
⎪
⎪
⎪
⎪ ∂ ∂wx ∂wy
⎪
⎩ K1 ∇2 wy + (K2 + K3 ) + = 0.
∂x ∂y ∂x
If defining two displacement potentials F (x, y) and G(x, y), such as

⎧
⎪
⎪ ∂2F
⎪
⎪ u x = (L + M ) ,
⎪
⎪ ∂x∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂2F ∂2F
⎪
⎪ u = −(L + 2M ) − M ,
⎨ y
∂x2 ∂y 2
(6.5-11)
⎪
⎪ ∂2G
⎪
⎪
⎪
⎪ w x = (K2 + K 3 ) ,
⎪
⎪ ∂x∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂2G ∂2G
⎩ wy = −K1 2 − (K1 + K2 + K3 ) 2 ,
∂x ∂y
then equations (6.5-10) will be reduced to
∇2 ∇2 F = 0, ∇2 ∇2 G = 0, (6.5-12)
which is found in Ref.[12].

One can see that the problem is concluded to solve two biharmonic equations,
and the theory and method studying this kind of equations are well developed in the
theory of classical elasticity, which can be used in studying elasticity of quasicrystals.
In this respect, the most systematic method is the complex variable function method.
If assume ϕ1 (z), ψ1 (z), π1 (z) and χ1 (z) are analytic functions of complex variable
6.6 Plane elasticity of point group 8mm of octagonal quasicrystals, displacement . . . 89
√
z = x + iy(i = −1), then
⎧
⎪
⎪
⎪ F (x, y) = Re z̄ϕ1 (z) + ψ1 (z)dz ,
⎨
(6.5-13)
⎪
⎪
⎪
⎩ G(x, y) = Re z̄π1 (z) + χ1 (z)dz .
where z̄ = x − iy and Re denotes the real part of a complex number. Theory

of analytic functions is a powerful tool to solve the boundary value problems of
harmonic, biharmonic and multiple harmonic equations, the work on quadruple and
sextuple harmonic equations is developed by the study of elasticity of quasicrystals,
the details will be given in the subsequent chapters. As a complete description for
the new development of the method some detailed summarization will be displayed
in Chapter 11.
6.6 Plane elasticity of point group 8mm of octagonal qua-

sicrystals, displacement potential
Octagonal quasicrystals belong to two-dimensional quasicrystals, the Penrose tiling
of its quasiperiodic plane is shown in Fig.6.6-1. The stacking of the planes along the
third direction perpendicular them will result in the quasicrystals.
Fig.6.6-1 Penrose tiling of eight-fold symmetry quasicrystal

In the following considering only a simple case of elasticity of the material, i.e., all
of the field variables are independent of the axis along which the atom arrangement
is periodic. In the case, the elasticity can be decomposed into plane elasticity and
anti-plane elasticity. We focus on the solution for the plane elasticity. Within the
quasiperiodic plane, the phonon elasticity is isotropic, so Cijkl are the same of those
discussed in previous sections. The phason elasticity here is anisotropic, but Kijkl
are the same with those of point group 12mm quasicrystals. Between phonon and
phason fields there is coupling, the corresponding elastic constants Rijkl are the
same given by (6.1-6), i.e.,
Rijkl = R(δi1 − δi2 )(δij δkl − δik δjl + δil δjk ), i, j, k, l = 1, 2. (6.6-1)
such that we have the elastic constants matrix

⎡ ⎤
L + 2M L 0 0 R R 0 0
⎢ ⎥
⎢ L L + 2M 0 0 −R −R 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 M M 0 0 −R R ⎥
⎢ ⎥
⎢ 0 0 M M 0 0 −R R ⎥
⎢ ⎥
[CKR] = ⎢ ⎥
⎢ R −R 0 0 K1 K2 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ R −R 0 0 K2 K1 0 0 ⎥
⎢ ⎥
⎢ −R −R ⎥
⎣ 0 0 0 0 K1 +K2 +K3 K3 ⎦
0 0 R R 0 0 K3 K1 +K2 +K3
⎡ ⎤
L+2M L 0 0 R R 0 0
⎢ ⎥
⎢ L+2M 0 0 −R −R 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ M M 0 0 −R R ⎥
⎢ ⎥
⎢ M 0 0 −R R ⎥
⎢ ⎥
=⎢ ⎥.
⎢ K1 K2 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ ⎥
⎢ ⎥
⎣ K1 +K2 +K3 K3 ⎦
K1 +K2 +K3
(6.6-2)
The corresponding free energy is
1 1 1
F = L(∇ · u)2 + M εij εij + K1 wij wij + K2 (wxy
2 2
+ wyx + 2wxx wyy )
2 2 2
1
+ K3 (wxy + wyx )2 + R[(εxx − εyy )(wxx + wyy ) + 2εxy (wyx − wxy )] (6.6-3)
2
where εij and wij are defined as before.
From (6.6-3) and (4.5-1) the generalized Hooke’s law can be expressed as
⎧
⎪
⎪
⎪
⎪
⎪
⎪ σyy = L(εxx + εyy ) + 2M εyy − R(wxx + wyy ),
⎪
⎪
⎪
⎪
⎪ σxy = σyx = 2M εxy + R(wyx − wxy ),
⎪
⎨
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ Hxy = (K1 + K2 + K3 )wxy + K3 wyx − 2Rεxy ,
⎪
⎪
⎪
⎩
Hyx = (K1 + K2 + K3 )wyx + K3 wxy + 2Rεxy .
For brevity, stress equilibrium equations are not listed.

Through a similar procedure, equilibrium equations given by displacement com-
ponents are as follows:
⎧ 2
⎪ ∂ ∂ wx ∂ 2 wy ∂ 2 wx
⎪
⎪
2
M ∇ ux + (L + M ) ∇ · u + R +2 − = 0,
⎪
⎪ ∂x ∂x2 ∂x∂y ∂y 2
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂ ∂ wy ∂ 2 wx ∂ 2 wy
⎪
⎪ 2
⎨ M ∇ uy + (L + M ) ∂y ∇ · u + R ∂x2 − 2 ∂x∂y − ∂y 2
⎪ = 0,
⎪ 2 2
⎪
⎪ ∂ wx ∂ 2 wy ∂ ux ∂ 2 uy ∂ 2 ux
⎪ K1 ∇ wx + (K2 + K3 )
⎪
2
+ +R −2 − = 0,
⎪
⎪ ∂y 2 ∂x∂y ∂x2 ∂x∂y ∂y2
⎪
⎪
⎪
⎪ 2 2
⎪
⎪ ∂ wx ∂ 2 wy ∂ uy ∂ 2 ux ∂ 2 uy
⎪
⎩ K1 ∇ wy + (K2 + K3 )
2
+ +R +2 − = 0.
∂x∂y ∂x2 ∂x2 ∂x∂y ∂y 2
(6.6-5)
It is evident that, if K2 + K3 = 0, the equations will be reduced to (6.2-1). In
fact, K2 + K3 = 0, so the equations are more complex than those given in the
previous sections. But the final governing equation for the present case presents
more interesting in mathematical physics, we can see immediately.
At first we introduce two auxiliary function ϕ(x, y) and ψ(x, y) in such way
⎧
⎪ ∂2ϕ ∂2ψ ∂2ψ
⎪
⎪ u = (L + M ) + M + (L + 2M ) ,
⎪
⎪
x
∂x∂y ∂x2 ∂y 2
⎪
⎪
⎪
⎪
⎪
⎪ ∂2ϕ ∂2ϕ ∂2ψ
⎪
⎪ u = − (L + 2M ) 2 + M 2 + (L + M )
⎪
⎨ y ∂x ∂y ∂x∂y
,
⎪ 2 (6.6-6)
⎪
⎪ ∂ ϕ ∂2ψ ∂2ψ
⎪ wx = −ω 2
⎪ + − ,
⎪
⎪ ∂x∂y ∂x2 ∂y 2
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂ ϕ ∂2ϕ ∂2ψ
⎪
⎩ wy = ω − 2 −2 .
∂x2 ∂y ∂x∂y
where ω = M (L + M )/R, so that (6.6-5) is simplified as
⎧
⎪
⎪ ∂2ϕ ∂2 ∂2
⎨ (γΠ1 + δΠ2 ) ∂x∂y + αΠ1 ∂y 2 − βΠ2 ∂x2 ψ = 0,
⎪
(6.6-7)
⎪
⎪ ∂2ψ ∂2 ∂2
⎪
⎩ (γΠ2 + δΠ1 ) + αΠ2 2 − βΠ1 2 ϕ = 0,
∂x∂y ∂y ∂x
in which
⎧
⎪ ∂2 ∂2 ∂2 ∂2
⎪
⎪ Π 1 = 3 − , Π 2 = 3 − ,
⎪
⎪ ∂x2 ∂y 2 ∂y2 ∂x2
⎪
⎨
α = R(L + 2M ) − ω(K1 + K2 + K3 ), (6.6-8)
⎪
⎪
⎪
⎪ β = RM − ωK1 , δ = RM − ω(K1 + K2 + K3 ),
⎪
⎪
⎩
γ = R(L + 2M ) − ωK1 .
and ω is given above. At last the displacement potential F (x, y) is introduced
⎧
⎪
⎪ ∂2 ∂2
⎪
⎨ ϕ = βΠ 2 − αΠ1 F,
∂x2 ∂y2
(6.6-9)
⎪
⎪ ∂2F
⎪
⎩ ψ = (γΠ1 + δΠ2 )
∂x∂y
and (6.6-7) reduces to a single equation as below:
(∇2 ∇2 ∇2 ∇2 − 4ε∇2 ∇2 Λ2 Λ2 + 4εΛ2 Λ2 Λ2 Λ2 )F = 0, (6.6-10)
where
⎧
⎪
⎪ 2 ∂2 ∂2 2 ∂2 ∂2
⎪
⎨ ∇ = + , Λ = − ,
∂x2 ∂y 2 ∂x2 ∂y 2
(6.6-11)
⎪
⎪ R2 (L + M )(K2 + K3 )
⎪
⎩ ε= .
[M (K1 + K2 + K3 ) − R2 ][(L + 2M )K1 − R2 ]
It is obvious that if K2 + K3 = 0, then ε = 0, equation (6.6-10) will be reduced to

(6.2-7).
If F (x, y) is the solution of (6.6-10), substituting it into (6.6-9) then into (6.6-6),
one can obtain the displacement field
⎧
⎪
⎪ ∂2
⎪
⎪ u = [M αΠ1 + (L + 2M )βΠ2 ] + 4ω(K2 + K3 )
⎪ x ∂x∂y
⎪
⎪
⎪
⎪
⎪
⎪
⎪ ∂2 ∂2
⎪
⎪
⎪
⎪ × M 2 + (L + 2M ) 2 Λ2 F,
⎪
⎪ ∂x ∂y
⎪
⎪
⎪
⎪
⎪
⎨ ∂2 ∂2
uy = M αΠ1 2 − (L + 2M )βΠ2 2 ∇2 F
∂y ∂x (6.6-12)
⎪
⎪
⎪
⎪ ∂4 2
⎪
⎪ −4ω(K2 + K3 )(L + M ) 2 2 Λ F,
⎪
⎪ ∂x ∂y
⎪
⎪
⎪
⎪ 2
⎪
⎪ ∂
⎪
⎪ wx = [ω(α − β)Π1 Π2 − 4ω 2 (K2 + K3 )Λ2 Λ2 ]F,
⎪
⎪ ∂x∂y
⎪
⎪
⎪
⎪ 2 2
∂4
⎪
⎩ wy = −ω αΠ12 ∂ + βΠ22 ∂ F − 8ω 2
(K + K ) Λ2 F.
2 3
∂y 2 ∂x2 ∂x2 ∂y 2
Similarly, the stress field has the following expressions:

⎧
⎪ ∂3 2 ∂5
⎪
⎪ σxx = 2M (L + M )αΠ1 ∇ F + 8M ω(L + M )(K2 + K3 ) 2 3 Λ2 F,
⎪
⎪ ∂y 3 ∂x ∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂3 ∂5
⎪
⎪ σyy = 2M (L + M )αΠ1 ∇2 F + 8M ω(L + M )(K2 + K3 ) 4 Λ2 F,
⎪
⎪ 2
∂x ∂y ∂x ∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂3 2 ∂5
⎪
⎪ σxy = σyx = −2M (L + M )αΠ1 ∇ F −8M ω(L+M )(K 2 +K 3 ) Λ2 F,
⎪
⎪ ∂x∂y 2 ∂x3 ∂y 2
⎪
⎪

⎪
⎪ ∂2 ∂3
⎪
⎪ ∂
⎪
⎪ Hxx = R M αΠ1 Λ + 2(L + 2M )βΠ2 2 ∇2 + 4Rω(K2 + K3 ) 2
2
⎪
⎪ ∂y ∂x ∂x ∂y
⎪
⎪
⎪
⎪ 2 2 3
⎪
⎪ ∂ ∂ ∂
⎪
⎪ M 2 + (2L + 5M ) 2 Λ2 + K1 ω 2 [(α − β)Π1 Π2 − 4ω(K2 + K3 )
⎪
⎪ ∂x ∂y ∂x ∂y
⎪
⎪
⎪
⎪ ∂3 2 ∂3 ∂5
⎪
⎪ 2 2 2 2
⎪
⎪ Λ Λ ] − K2 ω α 3 Π1 + β 2 Π2 + 8ω(K2 + K3 ) 2 3 Λ F,
⎪
⎪ ∂y ∂x ∂y ∂x ∂y
⎪
⎪

⎪
⎪ ∂2
⎪
⎪ ∂
⎪
⎪ Hyy = R M αΠ1 Λ2 + 2(L + 2M )βΠ2 2 ∇2
⎪
⎪ ∂y ∂x
⎪
⎪
⎪
⎪ 3 2
∂2
⎪
⎪ ∂ ∂
⎪
⎪ + 4Rω(K2 + K3 ) 2 M 2 + (2L + 5M ) 2 Λ2
⎪
⎪ ∂x ∂y ∂x ∂y
⎪
⎪
⎪
⎨ ∂3
+ K2 ω [(α − β)Π1 Π2 − 4ω(K2 + K3 )Λ2 Λ2 ]
⎪ ∂x2 ∂y
⎪
⎪
⎪
⎪ ∂3 ∂3 ∂5
⎪
⎪
⎪
⎪ − K1 ω α 3 Π12 + β 2 Π22 + 8ω(K2 + K3 ) 2 3 Λ2 F,
⎪
⎪ ∂y ∂x ∂y ∂x ∂y
⎪
⎪

⎪
⎪ ∂ ∂2
⎪
⎪ Hxy = −R 2M α 2 Π1 − (L + 2M )βΠ2 Λ2 ∇2
⎪
⎪ ∂x ∂y
⎪
⎪
⎪
⎪ ∂3
⎪
⎪
⎪
⎪ + 4Rω(L + 2M )(K2 + K3 ) Λ2 Λ2
⎪
⎪ ∂x∂y 2
⎪
⎪
⎪
⎪ ∂3
⎪
⎪ + (K1 + K2 + K3 )ω [(α − β)Π1 Π2 − 4ω(K2 + K3 )Λ2 Λ2 ]
⎪
⎪ ∂x∂y 2
⎪
⎪
⎪
⎪ ∂3 ∂3 2 ∂5
⎪
⎪ − K3 ω α 2
Π + β 3 Π2 + 8ω(K2 + K3 ) 3 2 Λ 2
⎪
⎪ F,
⎪
⎪ ∂x∂y2 1 ∂x ∂x ∂y
⎪
⎪

⎪
⎪ ∂ ∂2
⎪
⎪ Hyx = R 2M α 2 Π1 − (L + 2M )βΠ2 Λ2 ∇2
⎪
⎪ ∂x ∂y
⎪
⎪
⎪
⎪ ∂3
⎪
⎪
⎪
⎪ − 4Rω(L + 2M )(K2 + K3 ) Λ2 Λ2
⎪
⎪ ∂x∂y 2
⎪
⎪
⎪
⎪ ∂3
⎪
⎪ + K3 ω [(α − β)Π1 Π2 − 4ω(K2 + K3 )Λ2 Λ2 ] − (K1 + K2 + K3 )ω
⎪
⎪ ∂x∂y 2
⎪
⎪
⎪
⎪ ∂3 ∂3 2 ∂5
⎪
⎩ α Π 2
+ β Π + 8ω(K + K ) Λ 2
F.
2 3
∂x∂y 2 1 ∂x3 2 ∂x3 ∂y 2
(6.6-13)
A part of above results was reported by Refs. [12], [15] and [17].
6.7 Stress potential of point group 5, 5 pentagonal and point

group 10, 10 decagonal quasicrystals
In Section 4 we discussed the displacement potential of plane elasticity of point
group 5, 5 pentagonal and point group 10, 10 decagonal quasicrystals. But the stress
potential for the quasicrystals is also beneficial, which will be introduced in this
section.
From the basic formulas listed in Section 6.4, if we exclude the displacements
then there are the deformation compatibility equations
∂ 2 εxx ∂ 2 εyy ∂ 2 εxy ∂wxy ∂wxx ∂wyx ∂wyy
+ = 2 , = , = . (6.7-1)
∂y 2 ∂x2 ∂x∂y ∂x ∂y ∂y ∂x
If introducing stress potential functions φ(x, y), ψ1 (x, y) and ψ2 (x, y), such as
∂2φ ∂2φ ∂2φ
σxx = 2
, σyy = 2
, σxy = σyx = − ,
∂y ∂x ∂x∂y
∂ψ1 ∂ψ1 ∂ψ2 ∂ψ2
Hxx = , Hxy = − , Hyx = − , Hyy = , (6.7-2)
∂y ∂x ∂y ∂x
then equilibrium equations ∂σij /∂xj = 0 and ∂Hij /∂xj = 0 will be automatically
satisfied. Based on the generalized Hooke’s law (6.4-3), all strain components can
be expressed by relevant stress components:
⎧
⎪ 1 1
⎪
⎪ ε = (σ + σyy ) + [(K1 + K2 )(σxx − σyy )
⎪ xx 4(L + M ) xx
⎪ 4c
⎪
⎪
⎪
⎪ − 2R1 (Hxx + Hyy ) − 2R2 (Hxy − Hyx )],
⎪
⎪
⎪
⎪
⎪
⎪
1 1
(σxx + σyy ) − [(K1 + K2 )(σxx − σyy )
⎪
⎪ εyy =
⎪
⎪ 4(L + M ) 4c
⎪
⎪
⎪
⎪ − 2R 1 (H + H ) − 2R 2 (Hxy − H)],
⎪
⎪
xx yy
⎪
⎪ 1
⎪
⎪ ε = εyx = [(K1 + K2 )σxy − R2 (Hxx + Hyy ) + R1 (Hxy − Hyx )],
⎪
⎨ xy 2c
1 1
⎪
⎪ wxx = (Hxx − Hyy ) + [M (Hxx + Hyy ) − R1 (σxx − σyy ) − 2R2 σxy ],
⎪
⎪ 2(K 1 − K 2 ) 2c
⎪
⎪
⎪
⎪ 1 1
⎪
⎪ wyy = − (Hxx − Hyy ) + [M (Hxx + Hyy )
⎪
⎪ 2(K1 − K2 ) 2c
⎪
⎪
⎪
⎪ − R (σ − σ ) − 2R σ ],
⎪
⎪ 1 xx yy 2 xy
⎪
⎪
⎪
⎪ 1
wxy = [−R2 (σxx −σyy )+2R1 σxy ]+
1 M
(Hxy +Hyx )+ (Hxy −Hyx ),
⎪
⎪
⎪
⎪ 2c 2(K1 −K2 ) 2c
⎪
⎪
⎪
⎪ 1 1 M
⎩wyx = [R2 (σxx −σyy)−2R1 σxy]+ (Hxy +Hyx )− (Hxy −Hyx ),
2c 2(K1 −K2 ) 2c
(6.7-3)
6.8 Stress potential of point group 8mm octagonal quasicrystals 95
in which
c = M (K1 + K2 ) − 2(R12 + R22 ). (6.7-4)
So the deformation compatibility equations (6.7-2) can be rewritten by the stresses

σij , Hij , then by employing (6.4-6), one has
⎧
⎪
⎪ 1 K1 + K2 2 2 R1 ∂ ∂
⎪
⎪ + ∇ ∇ φ + Π ψ
1 1 − Π ψ
2 2
⎪
⎪ 2(L + M ) 2c c ∂y ∂x
⎪
⎪
⎪
⎪
⎪
⎪ R2 ∂ ∂
⎪
⎨ + c Π2 ψ1 + Π1 ψ2 = 0,
∂x ∂y
(6.7-5)
⎪
⎪ c ∂ ∂
⎪
⎪ 2
+ M ∇ ψ1 + R1 Π1 φ + R2 Π2 φ = 0,
⎪
⎪ K1 − K2 ∂y ∂x
⎪
⎪
⎪
⎪
⎪
⎪ c ∂ ∂
⎪
⎩ + M ∇2 ψ2 − R1 Π2 φ + R2 Π1 φ = 0,
K1 − K2 ∂x ∂y
where
∂2 ∂2 ∂2 ∂2 ∂2 ∂2
∇2 = 2
+ 2, Π1 = 3 2
− 2, Π2 = 3 2
− 2.
∂x ∂y ∂x ∂y ∂y ∂x
The equations (6.7-5) will be satisfied when we choose a new function G, which
is called the stress function, such that
φ = c1 ∇2 ∇2 G,

∂ ∂
ψ1 = − R1 Π1 + R2 Π2 ∇2 G,
∂y ∂x

∂ ∂
ψ2 = R1 Π2 − R2 Π1 ∇2 G, (6.7-6)
∂x ∂y
in which
c
c1 = +M (6.7-7)
K1 − K 2
and
∇2 ∇2 ∇2 ∇2 G = 0. (6.7-8)
So the final governing equation based on the stress potential is the same as that
based on the displacement potential.
6.8 Stress potential of point group 8mm octagonal quasicrys-

tals
The final governing equation of plane elasticity of point group 8mm octagonal quas-
sicrystals was given in Section 6 by displacement potential, similarly we can also
give a derivation by stress potential.
The strain compatibility equations and the definition on stress potential are the
same as those given by (6.7-1) and (6.7-2), and the strain-stress relations are as
follows:

1 1
εxx = (K1 + K2 )(L + 2M ) − 2R2 σxx
(L + M )c 4

1 1
− (K1 + K2 )L + 2R2 σyy − R(L + M )(Hxx + Hyy ) , (6.8-1a)
4 2

1 1 1
εyy = − (K1 + K2 )L + 2R2 σxx + [(K1 + K2 )(L + 2M )
(L + M )c 4 4

1
−2R2 ]σyy + R(L + M )(Hxx + Hyy ) , (6.8-1b)
2
1
εxy = εyx = [(K1 + K2 )σxy + R(Hxy − Hyx )], (6.8-1c)
2c
1 1
wxx = R(K1 − K2 )(σyy − σxx )
(K1 − K2 )c 2

+ (K1 M − R2 )Hxx − (K2 M − R2 )Hyy , (6.8-1d)
1 1
wyy = R(K1 − K2 )(σyy − σxx )
(K1 − K2 )c 2

− (K2 M − R2 )Hxx + (K1 M − R2 )Hyy , (6.8-1e)
1
wxy = {R(K1 + K2 + 2K3 )σxy
(K1 + K2 + 2K3 )c

+ (K1 + K2 + K3 )M − R2 Hxy − (K3 M + R2 )Hyx , (6.8-1f)
1
wyx = {−R(K1 + K2 + 2K3 )σxy
(K1 + K2 + 2K3 )c

−(K3 M + R2 )Hxy + (K1 + K2 + K3 )M − R2 Hyx , (6.8-1g)
in which
c = M (K1 + K2 ) − 2R2 .
If introducing stress potential functions φ(x, y), ψ1 (x, y) and ψ2 (x, y) such as
∂2φ ∂ 2φ ∂2φ
σxx = , σyy = , σ xy = σ yx = − ,
∂y2 ∂x2 ∂x∂y
∂ψ1 ∂ψ1 ∂ψ2 ∂ψ2
Hxx = , Hxy =− , Hyx = − , Hyy = . (6.8-2)
∂y ∂x ∂y ∂x
then equilibrium equations (6.5-9) (or ∂σij /∂xj = 0, ∂Hij /∂xj = 0) will be auto-
matically satisfied.
Substituting the stress formulas (6.8-2) into the strain-stress relation (6.8-1) then
6.8 Stress potential of point group 8mm octagonal quasicrystals 97
into the deformation compatibility equations (6.7-1), then we have

⎧
⎪ ∂ ∂
⎪
⎪ c1 ∇2 ∇2 φ + 2R(L + M ) Π1 ψ1 − 2R(L + M ) Π2 ψ2 = 0,
⎪
⎪ ∂y ∂x
⎪
⎪
⎨ 1 ∂ 2
∂ ψ2 2
∂ ψ1 2
∂ ψ1
− R Π1 φ + c2 − c3 − c4 = 0, (6.8-3)
⎪
⎪ 2 ∂y ∂x∂y ∂x 2 ∂y 2
⎪
⎪
⎪
⎪ 2 2 2
⎪
⎩ 1 R ∂ Π2 φ + c2 ∂ ψ1 − c4 ∂ ψ2 − c3 ∂ ψ2 = 0,
2 ∂x ∂x∂y ∂x2 ∂y 2
where
⎧
⎪ ∂2 ∂2 ∂2 ∂2 ∂2 ∂2
⎪
⎪ ∇ 2
= + , Π 1 = 3 − , Π 2 = 3 − ,
⎪
⎪ ∂x2 ∂y 2 ∂x2 ∂y2 ∂y2 ∂x2
⎪
⎪
⎪
⎨
K3 M + R2 K2 M + R2
c1 = (K1 + K 2 )(L + 2M ) − 2R2 , c2 = + ,
⎪
⎪ K1 + K 2 + 2K3 K1 − K 2
⎪
⎪
⎪
⎪
⎪ 2 2
⎩ c3 = (K1 + K 2 + K3 )M − R , c4 = K1 M + R .
⎪
K1 + K 2 + 2K3 K1 − K 2
(6.8-4)
∂
The manipulation of the second equation of (6.8-3) by × Π2 + the third equation
∂x
∂
of (6.8-3) by× Π1 , adding them together, leads to
∂y

∂3 ∂3 ∂3
c2 Π 1 − c3 Π 2 − c4 Π 2 ψ1
∂x∂y2 ∂x3 ∂x∂y 2

∂3 ∂3 ∂3
= c4 Π1 + c3 Π 1 − c2 Π2 ψ2 . (6.8-5)
∂x2 ∂y ∂y 3 ∂x2 ∂y
There exists a function A(x, y) such that
⎧
⎪ ∂3 ∂3 ∂3
⎪
⎪ ψ = c Π + c Π − c Π2 A,
⎨ 1 4
∂x2 ∂y
1 3
∂y 3
1 2
∂x2 ∂y
(6.8-6)
⎪
⎪ ∂3 ∂3 ∂3
⎪
⎩ ψ2 = c2 Π − c Π − c Π
1 3 2 4 2 A.
∂x∂y 2 ∂x3 ∂x∂y2
Substituting (6.8-6) into the third equation of (6.8-3), we arrive at the function
G(x, y) satisfying the relations such as
φ = −c3 c4 ∇2 ∇2 G,
1
A= RG. (6.8-7)
2
In the derivation the relation c2 = c4 − c3 has been used. The stress potentials
φ, ψ1 , ψ2 can be expressed by the new function G(x, y), i.e.,
⎧
⎪
⎪ φ = −c3 c4 ∇2 ∇2 G,
⎪
⎪
⎪
⎪ ∂3 ∂3 ∂3
⎨ 1
ψ1 = R c4 2 Π1 + c3 3 Π1 − c2 2 Π2 G,
2 ∂x ∂y ∂y ∂x ∂y (6.8-8)
⎪
⎪
⎪
⎪ 3 3
∂3
⎪
⎪ 1 ∂ ∂
⎩ ψ2 = R c2 Π1 − c3 3 Π2 − c4 Π2 G.
2 ∂x∂y 2 ∂x ∂x∂y 2
Substituting (6.8-8) into the first equation of (6.8-3) yields

∂4 ∂4 2 ∂4
− c1 c3 c4 ∇2 ∇2 ∇2 ∇2 G + R2 (L + M ) c4 Π 2
+ c3 Π − 2c 2 Π1 Π2
∂x2 ∂y 2 1 ∂y 4 1 ∂x2 ∂y 2
∂4 2 ∂4
2
+ c3 Π 2 + c 4 Π2 G = 0. (6.8-9)
∂y 4 ∂x2 ∂y 2
Considering the following relations
∂4 1 !
= ∇2 ∇2 − Λ2 Λ2 , Π1 Π2 = ∇2 ∇2 − 4Λ2 Λ2 , c2 = c4 − c3 ,
∂x2 ∂y 2 4
∂2 ∂2 ∂2 ∂2
∇2 = 2
+ 2 , Λ2 = 2
− 2,
∂x ∂y ∂x ∂y
equation (6.8-9) can be simplified as
∇2 ∇2 ∇2 ∇2 G − 4ε∇2 ∇2 Λ2 Λ2 G + 4εΛ2 Λ2 Λ2 Λ2 G = 0 (6.8-10)
with
R2 (L + M )(c3 − c4 ) R2 (L + M )(K2 + K3 )
ε= = .
−c1 c3 + R2 (L + M )c3 [(K1 + K2 + K3 )M − R2 ][K1 (L + 2M ) − R2 ]
(6.8-11)
The final governing equation in this case is exactly in agreement to that given by
the displacement potential formulation, discussed in Section 6.
6.9 Engineering and mathematical elasticity of quasicrystals

The complicated structure of quasicrystals leads to tremendous complexity of their
elasticity equations. Even though, the solution can also be carried out, and fruitful
results are found. This formulation is effective not only for linear elasticity of qua-
sicrystals but also for nonlinear material response of the solids if using some simple
physical models[18,19] . Those studies belong to mathematical theory of elasticity of
quasicrystals.
In the subsequent chapters we can realize that the difficulty for analytic solutions
does not only lie in the complexity of the equations but also the boundary condi-
tions. In some cases the boundary conditions can be simplified, useful approximate
6.9 Engineering and mathematical elasticity of quasicrystals 99
solutions may be constructed with relative ease. For this purpose we consider an
example.
From Section 6.3, we have deformation compatibility equations
⎧ 2
⎪ ∂ εxx ∂ 2 εyy ∂ 2 εxy
⎪
⎨ ∂y 2 + ∂x2 = 2 ∂x∂y ,
(6.9-1)
⎪
⎪
⎩ ∂wxy = ∂wxx , ∂wyx = ∂wyy .
∂x ∂y ∂y ∂x
Then the substitution of (6.3-2) into (6.9-1) yields the compatibility equations ex-
pressed by σij and Hij

2 L+M ∂2 ∂2
∇ (σxx + σyy ) − −
C ∂x2 ∂y2
× [(K1 + K2 )(σxx − σyy ) − 2R(Hxx + Hyy )]
L + M ∂2
=8 [(K1 + K2 )σxy − R(Hxy + Hyx )] (6.9-2a)
C ∂x∂y

1 ∂ (R2 − M K1 )Hxy + (R2 − M K2 )Hyx
Rσxy −
C ∂x K1 − K 2
1 ∂ 1 ∂
= (Hxx −Hyy )+ [M (Hxx +Hyy )−R(σxx −σyy )] (6.9-2b)
2(K1 −K2 ) ∂y 2C ∂y

1 ∂ (R2 − M K1 )Hxy + (R2 − M K2 )Hyx
−Rσxy −
C ∂y K1 − K 2
−1 ∂ 1 ∂
= (Hxx −Hyy )+ [M (Hxx +Hyy )−R(σxx −σyy )], (6.9-2c)
2(K1 −K2 ) ∂x 2C ∂x
in which C, M, L, K1 , K2 and R are material constants given in Section 6.3. These

three equations are combined with the equilibrium equations
⎧
⎪
⎨ ∂x + ∂y = 0, + = 0,
∂x ∂y
(6.9-3)
⎪
⎪ ∂Hxx + ∂Hxy = 0, ∂Hyx + ∂Hyy = 0
⎩
∂x ∂y ∂x ∂y
and this provides a basis for solving the problem.

A very meaningful example is pure bending of beam, shown in Fig.6.9-1. For this
sample, at the upper and lower surfaces of the beam there are following boundary
conditions
σyy = 0, σyx = 0,
(6.9-4)
Hyy = 0, Hyx = 0,
Fig.6.9-1 Beam of quasicrystal of ten-fold symmetry under pure bending
in addition there are so-called St. Venant boundary conditions at the end sections
⎧ h/2 h/2
⎪
⎪
⎪
⎪ σxx dy = 0, yσxx dy = Mz ,
⎪
⎪
⎪ −h/2
⎪
−h/2
⎪
⎨ h/2 h/2
Hxx dy = 0, yHxx dy = Lz , (6.9-5)
⎪
⎪ −h/2 −h/2
⎪
⎪
⎪
⎪ h/2 h/2
⎪
⎪
⎪
⎩ σxy dy = 0, Hyx dy = 0.
−h/2 −h/2
where Mz and Lz represent the resultant moments of σxx and Hxx . The direction
of vector of the moments is z. Boundary conditions (6.9-5) are relaxation boundary
conditions, this gives some flexibility for solution.
At first we assume that the value of Lz momentarily be undetermined, and
assume further
σxx = A1 y, σyy = σxy = σyx = 0, Hyy = Hxy = Hyx = 0, Hxx = f (y)
and A1 and f (y) are to be determined. Substituting σxx = A1 y into (6.9-5), we find
that
2Mz
A1 = 3 (6.9-6)
h
so that
Mz
σxx = y, (6.9-7)
I
where I = 1 · h3 /12 represents the inertia moment of the transverse section with
height h and width 1.
It can be verified that the above results have satisfied equations (6.7-3) and
(6.7-2).
Substituting (6.9-7) into (6.9-2b) yields
Rσxx 2RMz y
Hxx = = , (6.9-8)
1 M 1 M
+ I +
2(K1 − K2 ) 2C K1 − K2 C
References 101
so the equation is satisfied. So that we have the solution

⎧
⎪ Mz
⎪
⎨ σxx = I y, σyy = σxy = σyx = 0,
(6.9-9)
⎪
⎪ 2RMz y
⎩ Hxx = , Hyy = Hxy = Hyx = 0.
I[1/(K1 − K2 ) + M/C]
At last, combining (6.9-9) and (6.9-5) determines the value of Lz .
Here we utilize the physical character of the sample to solve the problem eas-
ily. This procedure is quite effective for some practical problems, and needs not so
hard mathematics. This treatment is similar to that in the engineering elasticity of
conventional structural materials, namely there is engineering elasticity of quasicrys-
tals. But this monograph mainly discusses mathematical elasticity of quasicrystals
on problems with topological or metric defects such as dislocations, cracks concern-
ing stress singularities, or with local discontinuity of physical quantities such as in
contact, shock wave problems. The complexity of boundary conditions renders the
above engineering approximate treatment ineffective, we must develop systematic
and direct analysis methods. In the classical elasticity, such systematic and direct
methods were developed by Muskhelishvili[20] in 1930’s and Sneddon[21] in 1940’s
and present important effects promoting elasticity and relevant disciplines of science
and engineering. By extending the methods in the classical elasticity to quasicrys-
tal elasticity we will display in detail the application in Chapters 7, 8 for one-and
two-dimensional quasicrystals, in Chapter 9 for three-dimensional quasicrystals, re-
spectively.
References
[1] Chernikov M A, Ott H R, Bianchi A et al. Elastic moduli of a single quasicrystal of
decagonal Al-Ni-Co: evidence for transverse elastic isotropy. Phys Rev Lett, 1998,
80(2): 321–324
[2] Abe H, Taruma N, Le Bolloc’h D et al. Anomalous-X-ray scattering associated with
short-range order in an Al70 Ni15 Co15 decagonal quasicrystal. Mater Sci and Eng A,
2000, 294∼296(12): 299–302
[3] Jeong H C, Steinhardt P J. Finite-temperature elasticity phase transition in decago-
nal quasicrystals. Phys Rev B, 1993, 48(13): 9394–9403
[4] Edagawa K. Phonon-phason coupling in decagonal quasicrystals. Phil Mag, 2007,
87(18∼21): 2789–2798
[5] Fan T Y, Mai Y W. Elasticity theory, fracture mechanics and relevant thermal prop-
erties of quasicrystalline materials. Appl Mech Rev, 2004, 57(5), 325–344
[6] Hu C Z, Yang W G, Wang R H et al. Symmetry and physical properties of qua-
sicrystals. Prog Phys, 1997, 17(4): 345–374
[7] Bohsung J, Trebin H R. Introduction to the Quasicrystal Mathematics. Jaric M V,

New York: Academic Press, 1989
[8] De P, Pelcovis R A. Disclinations in pentagonal quasicrystals. Phys Rev B, 1987, 35
(17): 9604–9607
[9] Yang S H, Ding D H. Foundation of Dislocation Theory of Crystals. Vol. II. Beijing:
Science Press, 1998 (in Chinese)
[10] De P, Pelcovits R A. Linear elasticity of pentagonal quasicrystals. Phys Rev B, 1987,
36(13): 8609–8620
[11] Ding D H, Wang R H, Yang W G et al. General expressions for the elastic displace-
ment fields induced by dislocations in quasicrystals. J Phys: Condens Matter, 1995,
7(28): 5423–5426
[12] Li X F, Fan T Y. New method for solving elasticity problems of some planar qua-
sicrystals and solutions. Chin Phys Letter, 1998, 15(4): 278–280
[13] Guo Y C, Fan T Y. Mode II Griffith crack in decagonal quasicrystals. Appl Math
Mech, English Edition, 2001, 22(10): 1311–1317
[14] Li L H, Fan T Y. Complex function method for solving notch problem of point 10
two-dimensional quasicrystal based on the stress potential function. J Phys: Condens
Matter, 2006, 18(47): 10631–10641
[15] Li X F, Fan T Y, Sun Y F. A decagonal quasicrystal with a Griffith crack. Phil Mag
A, 1999, 79(7): 1943–1952
[16] Li X F, Duan X Y, Fan T Y, et al. Elastic field for a straight dislocation in a
decagonal quasicryatal. J Phys: Condens Matter, 1999, 11(3): 703–711
[17] Zhou W M, Fan T Y. Plane elasticity of octagonal quasicrystals and solutions. Chin
Phys, 2001, 10(8): 743–747
[18] Fan T Y, Trebin A R, Messerschmidt U et al. Plastic flow coupled with a Griffith
crack in some one- and two-dimensional quasicrystals. J Phys: Condens Matter,
2004, 16 (37): 5419–5429
[19] Fan T Y and Fan L. Plastic fracture of quasicrystals, Phil. Mag., 2008, 88(4), 321–
335
[20] Muskhelishvili N I. Some Basic Problems in the Mathematical Theory of Elasticity.
Groringen, P Noordhoff Ltd: 1953
[21] Sneddon I N. Fourier Transforms. New York: McGraw-Hill, 1951
Chapter 7
Application I−−Some dislocation and
interface problems and solutions in one-
and two-dimensional quasicrystals
In Chapters 5 and 6, with the physical basis of quasicrystal elasticity based on the
density wave model, we have performed some mathematical operations, by proper
simplification, to reduce the original problems to the boundary value problems of
high-order partial differential equations, and to establish the standard solving proce-
dure and the fundamental solutions. This work is the development of the boundary
value problems in classical elasticity. Here, we need to pose a question: Do these
mathematical operations contribute to solving the quasicrystal elasticity problems?
This is answered only by practice. The following two chapters will provide the ap-
plications of these theories, including the solutions to some dislocation, crack and
interface problems in one- and two-dimensional quasicrystals. The calculation re-
sults indicate that the mathematical operations discussed in Chapters 5 and 6 are
powerful in solving these problems.
As we know, almost all monographs of classical elasticity do not place their main
focus upon dislocation and crack problems. These problems have been investigated
in monographs relating to dislocation theory and fracture mechanics, respectively.
As a monograph on theories of elasticity of quasicrystals, the current work is not
going to focus entirely on dislocations and cracks in quasicrystals. Therefore, as
an attempt to examine the theories developed in Chapters 5 and 6 and to show
their applications in elasticity of quasicrystals, we present some calculation exam-
ples of certain realistic dislocation and crack problems in elasticity of quasicrystals.
The methods developed in this work can be used for studying other problems in
quasicrystals.
Historically, soon after the discovery of quasicrystals, scientists proposed the
possibility of existing dislocations in quasicrystals. De and Pelcovits[1,2] first studied
elastic field around dislocations and disclinations in quasicrystals. Furthermore,
Ding et al[3,4] conducted systematic investigation of this topic by using the Green

104 Chapter 7 Application I—Some dislocation and interface problems and . . .
function method.
Now people have recognized that quasicrystal is a kind of ordered phase of
quasiperiodic long-range. Similar to crystals with ordered phase of long-range, the
breaking of long-range regularity takes place usually through topological defects,
i.e., dislocations lead to the breaking of the long-range symmetry. As mentioned
in Chapter 3, quasicrystals also possess the orientational symmetries incompatible
with those permitted in crystal theory. In quasicrystals, another type of defect,
disclinations, also exists simultaneously, which lead to breaking of the orientational
symmetry in quasicrystals. In some cases, the crystalline phases often coexist with
the quasicrystalline phases. So there is another kind of defects, the interface be-
tween quasicrystal and crystal. The existence of these defects including the stacking
fault dramatically affects the mechanical properties of quasicrystals. Therefore, it is
important to study the elastic properties of quasicrystals with dislocations, discli-
nations, interfaces and stacking faults.
As aforementioned, during the study of dislocations in quasicrystals, physicists
have developed some mathematical methods, such as the Green function method.
These important methods are available in relevant literature. Hereafter, by using the
elementary methods developed in Chapters 5 and 6, complex variable function and
Fourier transform, we study the displacement and stress fields around dislocations
in one- and two-dimensional quasicrystals quantitatively. The three-dimensional
dislocation and dynamic dislocation problems will be discussed in Chapters 9 and
10 respectively.
7.1 Dislocations in one-dimensional hexagonal quasicrystals

Following the sequence from simplicity to complexity, we first study the dislocations
in one-dimensional quasicrystals. A dislocation in a n-dimensional quasicrystal bears
|| || ||
the Burgers vector b|| ⊕b⊥ , where b|| =(b1 , b2 , · · · , bn ) is its Burgers vector of phonon
field, and b⊥ =(b⊥ ⊥ ⊥
1 ,b2 , · · · , bn ) the Burgers vector of phase field. And b is located
||
in the physical space E|| , while b⊥ in the supplementary space or the vertical space
|| || ||
E⊥ . For one-dimensional quasicrystals, b|| =(b1 , b2 , b3 ) and b⊥ =(0,0, b⊥ 3 ) since
⊥ ⊥ || ⊥ || || || ⊥
b1 = b2 =0. Therefore, b ⊕ b =(b1 , b2 , b3 , b3 ) here, which can be dealt with as
|| || || || ||
the superposition of (b1 , b2 ,0,0) and (0,0, b3 , b⊥ 3 ). The (b1 ,b2 ,0,0) corresponds to the
blade dislocation in regular hexagonal crystals, whose elastic solution is available in
common metal physics or dislocation monographs, (e.g. Refs.[7],[8]). The one corre-
||
spondence to the component (0,0,b3 , b⊥ 3 ) is the screw dislocation in one-dimensional
quasicrystal. We solve the elastic field induced by this dislocation in the following.
In Section 5.2, we have obtained the governing equations for anti-plane strain
7.1 Dislocations in one-dimensional hexagonal quasicrystals 105
problems in one-dimensional hexagonal quasicrystal elasticity such as
∇2 uz = 0, ∇2 wz = 0. (7.1-1)
The boundary conditions corresponding to the screw dislocation with the Burgers
||
vector (0, 0, b3 , b⊥
3 ) are
⎧ " "
⎨ uz ""y=0+ − uz ""y=0− = b ,
3
" " (7.1-2a)
⎩
wz "y=0+ − wz "y=0− = b⊥ 3
or

duz = b3 , dwz = b⊥
3, (7.1-2b)
Γ Γ
where Γ indicates an arbitrary contour surrounding the dislocation core. The solu-
tion to the boundary value problem (7.1-1) and (7.1-2) is

b3 y b⊥
3 y
uz = arctan , wz = arctan . (7.1-3)
2π x 2π x
The stress components can be extracted according to the stress-strain relation (5.2-8)
in Chapter 5 as
⎧
⎪
⎪ C44 b3 y R3 b⊥
3 y
⎪
⎪ σ = σ = − − ,
⎪
⎪
xz zx
2π x2 + y 2 2π x2 + y 2
⎪
⎪
⎪
⎪
⎪
⎪ C44 b3 x R3 b⊥ x
⎪
⎪ σ = σ = − + 3
,
⎨ yz zy
2π x2 + y 2 2π x2 + y 2
(7.1-4)
⎪
⎪
⎪ K 2 b⊥
y R 3 b y
⎪
⎪ Hzx = − 3
− 3
,
⎪
⎪ 2π x2 + y 2 2π x2 + y 2
⎪
⎪
⎪
⎪
⎪
⎪ K b⊥ R3 b3
⎪
⎩ Hzy = − 2 3
x
+
x
.
2π x2 + y 2 2π x2 + y 2
||
The displacement and stress fields corresponding to the Burgers vector b|| ⊕ b⊥ =(b1 ,
|| ||
b2 , b3 , b⊥
3 ) can be obtained by superposing the above elastic field on that of the
regular hexagonal crystals.
The elastic strain energy induced by the screw dislocation is

1 ∂uz ∂wz
W = σzj + Hzj dx1 dx2
2 Ω ∂xj ∂xj

1 R0 2π ∂uz ∂wz
= r σzj + Hzj drdθ
2 r0 0 ∂xj ∂xj
2 1 R0
= (C44 b3 + K2 b⊥2 ⊥
3 + 2R3 b3 b3 ) ln , (7.1-5)
4π r0
where r0 is the size of the dislocation core, R0 is the size of dislocation network or
inclusion, which are available in theory of dislocations in regular crystals.
7.2 Dislocations in quasicrystals with point groups 5m and

10mm symmetries
Consider a dislocation in pentagonal point group 5m or decagonal point group 10mm
|| || ||
quasicrystals with the Burgers vector b|| ⊕ b⊥ =(b1 ,b2 ,b3 ,b⊥ ⊥ ||
1 ,b2 ), where b is its
⊥
Burgers vector of phonon field, and b is its Burgers vector of phase field. Here,
b⊥
3 =0 due to wz =0.
This dislocation problem can be dealt with by superposition of two dislocation
|| || ||
problems of the Burgers vectors (b1 ,b2 ,0,b⊥ ⊥
1 ,b2 ) and (0,0,b3 ), respectively. On the
dislocation line parallel to the direction of periodic arrangement (z-direction), the
|| ||
elastic field induced by the dislocation of the Burgers vector (b1 ,b2 ,0,b⊥ ⊥
1 ,b2 ) can be
simplified as an in-plane elastic field in quasicrystal elasticity, which can be analyzed
by the methods discussed in Section 6.1 and Section 6.2. The elastic field induced
||
by (0,0,b3 ) is governed by equation ∇2 uz =0, and its solution has been obtained in
Section 7.1. Therefore, in the following two sections, we always assume the disloca-
tion line parallel to the direction of atom periodic arrangement in quasicrystals, and
we only study the elastic field induced by the dislocation with the Burgers vector
|| ||
(b1 ,b2 ,0,b⊥ ⊥
1 ,b2 ).
To demonstrate the methodology concisely, we first consider a special case of
||
(b1 ,0,0,b⊥1 ,0). During this procedure, in order to further simplify the mathematical
|| ||
process, we first analyze the case of b1 = 0 and b2 = b⊥ ⊥
1 = b2 = 0. In this case,
we consider the problem in the upper half-space of y 0, and the problem has the
following boundary conditions:
⎧ #
⎪
⎪ σij (x, y) → 0, Hij (x, y) → 0, x2 + y 2 → ∞,
⎪
⎪
⎪
⎪
⎨ σyy (x, 0) = 0,
wx (x, 0) = wy (x, 0) = 0, (7.2-1)
⎪
⎪
⎪
⎪
⎪
⎪
⎩ ux (x, 0+ ) − uy (x, 0− ) = b or
dux = b1 ,
1
Γ
where Γ indicates an arbitrary contour surrounding the dislocation core.

The dislocation problem is reduced to solve equation (6.2-7), i.e., the following
equation
∇2 ∇2 ∇2 ∇2 F = 0
under boundary conditions (7.2-1).

7.2 Dislocations in quasicrystals with point groups 5m and 10mm symmetries 107
By introducing the Fourier transform

∞
F̂ (ξ, y) = F (x, y)eiξx dx, (7.2-2)
−∞
where ξ is the Fourier transform parameter, the above equation can be reduced into
4
d2
− ξ2 F̂ = 0. (7.2-3)
dy 2
This is an ordinary differential equation with constant coefficients, and its general
solution is
F̂ (ξ, y) = [A1 + B1 y + C1 y 2 + D1 y 3 ]e−|ξ|y

+ [A2 + B2 y + C2 y 2 + D2 y 3 ]e|ξ|y , (7.2-4)
where A1 , B1 , · · · , D2 are functions with respect to ξ to be determined by boundary

conditions.
We are going to use the Fourier transform to search for solutions in Chapters
7∼10. For convenience, we only consider the problem in the upper half-space (or
the lower half-space) based on the symmetry or anti-symmetry of the problems.
It should be cautious that here the symmetry or anti-symmetry is in the view of
macroscopic continuum media, namely the even and odd characteristics of the dis-
placement function F (x,y) or stress function G(x,y) with respect to x, but not in the
scale of crystalline structures. Furthermore, during the use of the Fourier transform,
we attempt to make the boundary condition homogeneous at infinity (zero boundary
condition). Therefore, we only consider the case in the upper half-space, and the
formal solution (7.2-4) can be simplified as
F̂ (ξ, y) = [A1 + B1 y + C1 y 2 + D1 y 3 ]e−|ξ|y . (7.2-5)
In the evident reason the suffix of A1 , B1 , C1 , D1 can be removed in the following.

For conciseness in writing, introduce the following indicator:
X = (A, B, C, D), Y = (1, y, y 2 , y 3 )T , (7.2-6)
where the symbol T indicates the matrix transpose. Therefore, (7.2-5) can be ex-
pressed as
F̂ (ξ, y) = XY e−|ξ|y . (7.2-5 )
Since A, B, C, and D are arbitrary functions with respect to ξ, without loss of

generality, (7.2-5) can be rewritten as
F̂ (ξ, y) = (4ξ 4 )−1 XY e−|ξ|y . (7.2-7)
By performing the Fourier transform on the displacement expressions (6.2-8),

(6.2-9) and stress expressions (6.2-10), (6.2-11) and using the notation (7.2-7), we
have
ûx (ξ, y) = iξ −1 X[2nξ 2 Y + (m − 5n) |ξ| Y − (2m − 5n)Y ]e−|ξ|y , (7.2-8a)
ûy (ξ, y) = |ξ|−1 X[2nξ 2 Y − (m + 5n) |ξ| Y + (2m + 5n)Y ]e−|ξ|y , (7.2-8b)
ŵx (ξ, y) = −iω(α − β)ξ −1 X[2n |ξ|3 Y − 12ξ 2 Y
+ 15 |ξ| Y − 10Y ]e−|ξ|y , (7.2-9a)
ŵy (ξ, y) = −ω(α − β) |ξ|−1 X[4 |ξ|3 Y − 12ξ 2 Y
+ 15 |ξ| Y − (10 + e0 − e1 )Y ]e−|ξ|y , (7.2-9b)
σ̂xx (ξ, y) = 2M α(L + M )X(−2ξ 2 Y + 8 |ξ| Y − 13Y )e−|ξ|y , (7.2-10a)
σ̂yy (ξ, y) = 2M α(L + M )X(2ξ 2 Y − 4 |ξ| Y + 3Y )e−|ξ|y , (7.2-10b)
−1 2
σ̂xy (ξ, y) = σ̂yx (ξ, y) = i2M α(L + M )ξ |ξ| X(2ξ Y − 6 |ξ| Y
+ δY )e−|ξ|y , (7.2-10c)
−1 3
Ĥxx (ξ, y) = −ω(α − β)(K1 − K2 ) |ξ| X[4 |ξ| Y
− 16ξ 2 Y + 27 |ξ| Y − (25 + e2 )Y ]e−|ξ|y (7.2-11a)
3
Ĥyy (ξ, y) = −ω(α − β)(K1 − K2 )X[−4 |ξ| Y + 12ξ 2 Y − 15 |ξ| Y
+ (10 − e1 )Y ]e−|ξ|y , (7.2-11b)
3
Ĥxy (ξ, y) = iω(α − β)(K1 − K2 )ξ −1 |ξ| X[−4 |ξ| Y + 12ξ 2 Y
− 15 |ξ| Y + (10 + e2 )Y ]e−|ξ|y , (7.2-11c)
3
Ĥyx (ξ, y) = −iω(α − β)(K1 − K2 )ξ −1 |ξ| X[−4 |ξ| Y + 16ξ 2 Y
− 27 |ξ| Y + (25 − e1 )Y ]e−|ξ|y , (7.2-11d)
where
⎧
⎨ m = M α + (L + 2M )β, n = M α − (L + 2M )β,
2αβ 2αβ α+β (7.2-12)
⎩ e1 = , e2 = +
ω(α − β)(K1 − K2 ) ω(α − β)(K1 − K2 ) α − β
and α, β, and ω are given in (6.2-5).
The inverse of Fourier transform reads
∞
1
F (x, y) = F̂ (ξ, y)e−iξx dξ, (7.2-13)
2π −∞
Similarly, the inverses of displacements and stresses are obtained as follows:

∞
1
uj (x, y) = ûj (ξ, y)e−iξx dξ, (7.2-14)
2π −∞
∞
1
wj (x, y) = ŵj (ξ, y)e−iξx dξ, (7.2-15)
2π −∞
∞
1
σjk (x, y) = σ̂jk (ξ, y)e−iξx dξ, (7.2-16)
2π −∞
∞
1
Hjk (x, y) = Ĥjk (ξ, y)e−iξx dξ. (7.2-17)
2π −∞
In the above dislocation problem, ux (x,y) is an odd function with respect to x,
therefore F (x, y) must be an odd function with respect to x according to the first
expression in (6.2-8). For an odd function, the Fourier transform (7.2-2) is rewritten
as ∞
F̂ (ξ, y) = F (x, y) sin(ξx)dx (7.2-18)
0
and its inversion is ∞
2
F (x, y) = F̂ (ξ, y) sin(ξx)dξ. (7.2-19)
π 0
If F (x,y) is an even function with respect to x, it has
∞
F̂ (ξ, y) = F (x, y) cos(ξx)dx (7.2-20)
0
and its inversion is ∞

2
F (x, y) = F̂ (ξ, y) cos(ξx)dξ. (7.2-21)
π 0
Thereby, the corresponding integrals relating uj , wj , σjk and Hjk can be simpli-
fied.
So far, the displacement and stress components in forms of integrals have been
obtained, which all include the unknown functions A(ξ), B(ξ), C(ξ) and D(ξ) to be
determined by the second expressions in the boundary conditions (7.2-1) such that
9Jsgnξ Jsgnξ
A= , B = 2Jξ 2 , C= , D = 0, (7.2-22)
4ξ 2 2ξ 2
where J is a constant to be determined. By using the last expression in (7.2-1), i.e.,
the dislocation condition, J is found to be

b1
J= . (7.2-23)
8(n − m)
Substitution of the above expressions into (7.2-8)∼(7.2-11) and then into the inver-
sion formulas (7.2-14)∼(7.2-17) leads to the displacement and stress fields induced
||
by dislocation with Burgers vector (b1 ,0,0,0) in quasicrystals of point group 10mm
ten-fold symmetry as below

b1 y (L + M )K1 xy
ux = arctan + , (7.2-24a)
2π x (L + M )K1 + (M K1 − R2 ) r2
r
b1 (M K1 − R2 )
uy = − ln
2π (L + M )K1 + (M K1 − R2 ) a
2
(L + M )K1 y
+ , (7.2-24b)
(L + M )K1 + (M K1 − R2 ) r2
$ % 3

b1 (L + M )K1 2x y
wx = , (7.2-25a)
2π (L + M )K1 + (M K1 − R2 ) r4
$ % 2 2

b1 (L + M )K1 2x y
wy = , (7.2-25b)
2π (L + M )K1 + (M K1 − R2 ) r4
y(3x2 + y 2 )
σxx = −A , (7.2-26a)
r4
2 2
y(x − y )
σyy =A , (7.2-26b)
r4
x(x2 − y 2 )
σxy = σyx = A , (7.2-26c)
r4
R(K1 − K2 ) x2 y(3x2 − y 2 )
Hxx = −A , (7.2-27a)
M K1 − R 2 r6
2
R(K1 − K2 ) x y(3y2 − x2 )
Hyy = −A , (7.2-27b)
M K1 − R 2 r6
2 2
R(K1 − K2 ) xy (3x − y 2 )
Hxy =A , (7.2-27c)
M K1 − R 2 r6
3 2
R(K1 − K2 ) x (3y − x2 )
Hyx = −A , (7.2-27d)
M K1 − R 2 r6
#
where r = x2 + y 2 , a is the size of dislocation core and
$ %

b1 (L + M )(M K1 − R2 )
A= . (7.2-28)
π (L + M )K1 + (M K1 − R2 )
Considering L = C12 , and M =(C11 − C12 )/2=C66 in (6.1-1), and substituting

them into (7.2-24a) and (7.2-25), we find that our solutions exactly consistent with
those given by Ding et al[4,7] . They used the Green function method. This examines
the correction of our derivation.
In the case of R=0, the above solution recovers the solution of dislocation in
hexagonal crystals[8] . If the material is isotropic, therefore L = λ and M = μ.
Substitution of L and M into (7.2-28) yields

μb1
A= , (7.2-29)
2π(1 − ν)
where λ and μ are the Lamé constants, and ν is the Poisson’s ratio of phonon field.
In this case, ui and σij recover those in a regular isotropic crystal with an edge
dislocation while wi =0 and Hij =0.
||
Let us consider another special case of b⊥ ⊥
1 = 0 and b1 = b2 = b2 = 0. In this
case, the boundary conditions can be described as
⎧ #
⎪
⎪ σij (x, y) → 0, Hij (x, y) → 0, x2 + y 2 → ∞,
⎪
⎪
⎪
⎪
⎨ Hyy (x, 0) = 0,
ux (x, 0) = uy (x, 0) = 0, (7.2-30)
⎪
⎪
⎪
⎪
⎪
⎪
⎩ wx (x, 0+ ) − wx (x, 0− ) = b⊥
1 or dwx = b⊥ 1,
Γ
where the meaning of Γ is aforementioned.

By using the similar procedure, the displacement and stress fields can be deter-
mined as

b⊥1 k0 xy c1 − c2 2xy 3
ux = − , (7.2-31a)
2πc2 r2 2c1 r4

b⊥1 k0 xy c1 − c2 y 2 (x2 − y 2 )
uy = − 2 + , (7.2-31b)
2πc2 r 2c1 r4
y c k xy(3x2 − y 2 )(3y 2 − x2 )
b⊥ 0 0
wx = 1 arctan + , (7.2-31c)
2π x 2c1 c2 3r 6

b⊥ L + 2M M r c0 k0 y 2 (3x2 − y 2 )2
wy = 1 1− − ln + , (7.2-31d)
2π 2c1 2c2 a 2c1 c2 3r 6
⊥ 2
c0 b1 k0 x y(3x2 − y 2 )
σxx = − , (7.2-32a)
πc1 R r6
⊥ 3 2
c0 b1 k0 y (3x − y 2 )
σyy = − , (7.2-32b)
πc1 R r6
⊥
c0 b1 k0 2xy 2 (y 2 − x2 )
σxy = σyx = , (7.2-32c)
πc1 R r6

k0 b⊥1 y 2x2 y(3x2 − y 2 )(3y 2 − x2 )
Hxx = (e1 + e2 ) 2 − , (7.2-33a)
2πe1 r r8
⊥
k 0 b1 y 2(x2 − y 2 ) (x2 − y 2 )(3x2 − y 2 )(3y 2 − x2 )
Hyy = − + , (7.2-33b)
2πe1 r4 r8

k0 b⊥1 x 2xy2 (3x2 − y 2 )(3y 2 − x2 )
Hxy = (e1 + e2 ) 2 + , (7.2-33c)
2πe1 r r8

k 0 b⊥
1x 2(x2 − y 2 ) (x2 − y 2 )(3x2 − y 2 )(3y 2 − x2 )
Hyx = − + , (7.2-33d)
2πe1 r4 r8
where
2c1 c2 c1 c2 c1 c
e1 = , e2 = + 2 ,
c0 k0 c0 k0 c1 c2
c1 = (L + 2M )K2 − R2 , c2 = M K2 − R2
and c0 , c1 ,c2 and k0 are given in Chapter 6, i.e.,
c0 = (L + 2M )R, c1 = (L + 2M )K1 − R2 , c2 = M K1 − R2 ,
k0 = R(K1 − K2 ).
||
Superposition of the above two solutions yields the solution of dislocation (b1 , 0,
||
b1 , 0). The solution of (0, b2 , 0, b⊥
⊥
2 ) can be determined similarly. Their superposition
|| ||
yields the solution of (b1 , b2 , b⊥ ⊥
1 b2 ). Part of this work can be found in paper given
,
by Li and Fan[5] . Readers may examine that the present solutions are identical to
those given by Ding et al[4] using the method of Green functions (Note that L = C12
and M = (C11 − C12 )/2=C66 ).
7.3 Dislocations in quasicrystals with point groups 5, 5̄ five-

fold and 10, 10 ten-fold symmetries
Similar to those discussed in the proceeding section, as two-dimensional quasicrys-
|| || ||
tals, the Burgers vector of the dislocation is b|| ⊕ b⊥ =(b1 ,b2 ,b3 ,b⊥ ⊥
1 , b2 ). Due to
w3 =0 and our assumption that dislocation line is parallel to the direction of pe-
riodic arrangement (z-direction), the above Burgers vector can be regarded as the
||
superposition of (b1 ,b2 ,b⊥ ⊥ ⊥
1 ,b2 ) and (0,0,b3 ). Under this condition, the field variables
do not vary with the change of the variable z. Therefore, the elastic field corre-
|| ||
sponding to (b1 ,b2 ,b⊥ ⊥
1 ,b2 ) can be determined by using the theory of plane elasticity
of two-dimensional quasicrystals discussed in Section 6.4. The anti-plane elastic field
induced by (0,0,b⊥ 3 ) is very simple, and can be solved by using the method discussed
in Section 7.1.
From Section 6.4, we have obtained the governing equation of plane elastic field
of quasicrystals with point groups 5, 5̄ five-fold and 10, 1̄0̄ ten-fold symmetries:
∇2 ∇2 ∇2 ∇2 F = 0, (7.3-1)
where F (x,y) is the displacement potential function, defined in (6.4-10). By per-

forming the Fourier transform of (7.2-2) and considering the case of upper half-space
(y 0), its Fourier transform F̂ (ξ, y) has the form of
F̂ (ξ, y) = (4ξ 4 R2 )−1 XY e−|ξ|y , (7.3-2)

7.3 Dislocations in quasicrystals with point groups 5, 5̄ five-fold and 10, 10 . . . 113
where X and Y have the same meanings as those in Section 7.2, ξ is the parameter
of Fourier transform and
R2 = R12 + R22 . (7.3-3)
After some algebraic manipulation with (6.2-8)∼(6.2-11), the displacement com-
ponents in the Fourier transform domain are
ûx (ξ, y) = iξ −1 R̄0 X[2nξ 2 Y + (m − 5n) |ξ| Y − (2m − 5n)Y ]e−|ξ|y , (7.3-4a)
−1 2 −|ξ|y
ûy (ξ, y) = |ξ| R̄0 X[2nξ Y − (m + 5n) |ξ| Y + (2m + 5n)Y ]e , (7.3-4b)
3
ŵx (ξ, y) = ic0 ξ R̄02 X[4 |ξ| Y − 12ξ 2 Y + 15 |ξ| Y − 10Y ]e−|ξ|y ,
−1
(7.3-4c)
−1 2 3 2 2 −|ξ|y
ŵy (ξ, y) = c0 |ξ| R̄0 X[4 |ξ| Y − 12ξ Y + 15 |ξ| Y − (10 + e0 R0 )Y ]e ,
(7.3-4d)
where
⎧
⎪
⎪ −[C66 c1 + C11 c2 ]
⎨ m = c2 + c1 , n = c2 − c1 , e0 = ,
Rc0
(7.3-5)
⎪
⎪ R1 − iR2 sgn ξ
⎩ R0 = R1 + iR2 sgn ξ , R̄0 = .
R R
Similarly, the stress components in the Fourier transform domain are
σ̂xx (ξ, y) = 2c0 c2 R−1 R̄0 X(−2ξ 2 Y + 8 |ξ| Y − 13Y )e−|ξ|y , (7.3-6a)
−1 2 −|ξ|y
σ̂yy (ξ, y) = 2c0 c2 R R̄0 X(2ξ Y − 4 |ξ| Y + 3Y )e , (7.3-6b)
−1
σ̂xy (ξ, y) = σ̂yx (ξ, y) = i2c0 c2 R R̄0 (sgnξ)
· X(2ξ Y − 6 |ξ| Y + 7Y )e−|ξ|y ,
2
(7.3-6c)
3
Ĥxx (ξ, y) = c0 k0 R̄02 X[4 |ξ| Y − 16ξ 2 Y + 27 |ξ| Y
− (25 + e2 R02 )Y ]e−|ξ|y , (7.3-6d)
3
Ĥyy (ξ, y) = c0 k0 R̄02 X[−4 |ξ| Y + 12ξ 2 Y − 15 |ξ| Y
+ (10 − e1 R02 )Y ]e−|ξ|y , (7.3-6e)
Ĥxy (ξ, y) = ic0 k0 R̄02 (sgnξ)X[−4 |ξ|3 Y + 12ξ 2 Y − 15 |ξ| Y
+ (10 + e2 R02 )Y ]e−|ξ|y , (7.3-6f)
3
Ĥyx (ξ, y) = ic0 k0 R̄02 (sgnξ)X[−4 |ξ| Y
+ 16ξ 2 Y − 27 |ξ| Y + (25 − e1 R02 )Y ]e−|ξ|y , (7.3-6g)
⎧
⎪
⎨ e1 = 2c1 c2 , e2 = c1 c2 c1 c2
+ ,
c0 k0 R c0 k0 R c1 c2 (7.3-7)
⎪
⎩
c1 = (L + 2M )K2 − R2 , c2 = M K2 − R2 ,
where c0 and K0 are given in (6.4-13).
As introduced in the proceeding section, X includes four unknowns A(ξ), B(ξ),

C(ξ) and D(ξ), which can be determined by boundary conditions. Once A(ξ), B(ξ),
C(ξ) and D(ξ) are determined, ui , wi , σjk and Hjk can be determined by the Fourier
inverse.
||
In the case of dislocation problem b|| ⊕b⊥ =(b1 ,0,b⊥ 1 ,0), we try to determine these
functions and their Fourier inverses in the following.
||
In order to simplify the calculation, we first consider the case of b1 = 0 and
b⊥
1 = 0. In this case the boundary conditions are
⎧ #
⎪
⎪ σ ij (x, y) → 0, H ij (x, y) → 0, x2 + y 2 → ∞,
⎪
⎪
⎪
⎪ σyy (x, 0) = 0,
⎨
wx (x, 0) = wy (x, 0) = 0, (7.3-8)
⎪
⎪
⎪
⎪
⎪
⎪
⎩
dux = b1 , duy = 0.
Γ Γ
In the Fourier transform domain, solutions (7.3-4), (7.3-6) and the first three
expressions in boundary conditions (7.3-8) yield
9C(ξ)
A(ξ) = , B(ξ) = 2C(ξ), D(ξ) = 0, (7.3-9)
4
and the dislocation condition leads to
⎧
⎪
⎪ dux = −4c1 R−1 (R1 ReC + R2 sgnξImC),
⎨
Γ
(7.3-10)
⎪
⎪
⎩ duy = 4c1 R−1 (R2 ReC − R1 sgnξImC),
Γ
where the indicator Re and Im stand for the real and imaginary parts of a complex,
respectively. Finally, expressions (7.3-10) and (7.3-8) define

(R1 + iR2 sgnξ)b1 R0 b1
C(ξ) = ReC + iImC = − =− , (7.3-11)
4πRc1 4πc1
where R0 is given by (7.3-5), and c1 is given by (6.4-9), i.e.,
c1 = (L + 2M )K1 − R2 , R2 = R12 + R22 . (7.3-12)
Therefore,

b1 y c − c xy
1 2
ux = arctan + , (7.3-13a)
2π x c1 r2

b1 r c1 − c2 r y2
uy = − ln + ln + 2 , (7.3-13b)
2π a c1 a r
7.3 Dislocations in quasicrystals with point groups 5, 5̄ five-fold and 10, 10 . . . 115

c 0 b1 R1 2x3 yy 2 (3x2 + y 2 )
R2
wx = + , (7.3.13c)
2πc1 R r4 r4R
3
c 0 b1 R1 y 2 (3x2 + y 2 ) R2 2x y
wy = + , (7.3-13d)
2πc1 R r4 R r4
where c0 is given by (6.4-13), and a indicates the radius of dislocation core.
The corresponding stress components can be determined by (7.3-13) and gener-
alized Hooke’s law (6.4-3):
$ %

c0 c2 b1 y(3x2 + y 2 )
σxx = − , (7.3-14a)
πc1 R r4
$ %

c0 c2 b1 y(x2 − y 2 )
σyy = , (7.3-14b)
πc1 R r4
$ %

c0 c2 b1 x(x2 − y 2 )
σxy = σyx = , (7.3-14c)
πc1 R r4
3 2
c0 k0 b1 R1 x2 y(3x2 − y 2 ) R2 x (3y − x2 )
Hxx = − + , (7.3-14d)
πc1 R r6 R r6
2 2
c0 k0 b1 R1 x2 y(3x2 − y 2 ) R2 xy (3x − y 2 )
Hyy = − − , (7.3-14e)
πc1 R r6 R r6
2
c0 k0 b1 R1 xy 2 (3x2 − y 2 ) R2 x y(3y 2 − x2 )
Hxy = − + , (7.3-14f)
πc1 R r6 R r6
2
c0 k0 b1 R1 x3 (3y 2 − x2 ) R2 x y(3x2 − y 2 )
Hyx = − − + . (7.3-14g)
πc1 R r6 R r6
||
Now let us consider the case of b1 = 0 and b⊥ 0. In this case, the corresponding
1 =
boundary conditions are
⎧ #
⎪
⎪ σij (x, y) → 0, Hij (x, y) → 0, x2 + y 2 → ∞,
⎪
⎪
⎪
⎪ H (x, 0) = 0,
⎨ yy
ux (x, 0) = uy (x, 0) = 0, (7.3-15)
⎪
⎪
⎪
⎪
⎪
⎪
⎩ dwx = b⊥1, dwy = 0.
Γ Γ
By using the similar analysis, the corresponding displacement and stress fields can
be determined as the following formulas (7.3-16) and (7.3-17),

c1 b⊥
1 R1 xy c1 − c2 2xy 3
ux = −
πc0 e1 R r2 c1 r4

R2 y 2 c1 − c2 y 2 (x2 − y 2 )
+ + , (7.3-16a)
R r2 c1 r4

c1 b⊥
1 R1 y 2 c1 − c2 y 2 (x2 − y 2 )
uy = − −
πc0 e1 R r2 c1 r4

R2 xy c1 − c2 2xy 3
+ + , (7.3-16b)
R r2 c1 r4
y R2 − R2 xy(3x2 − y 2 )(3y 2 − x2 )
b⊥
1 1 2
wx = arctan +
2π x e1 R2 3r 6

2R1 R2 y 2 (3x2 − y 2 )2
+ , (7.3-16c)
e1 R2 3r 6
2
b⊥
1 r R1 − R22 y 2 (3x2 − y 2 )2
wy = e2 ln +
2πe1 a R2 3r 6

2R1 R2 xy(3x2 − y 2 )(3y 2 − x2 )
− , (7.3-16d)
e1 R2 3r 6

2c2 b⊥1 R1 x2 y(3x2 − y 2 ) R2 x3 (3y 2 − x2 )
σxx =− + , (7.3-17a)
πe1 R R r6 R r6

2c2 b⊥1 R1 y 3 (3x2 − y 2 ) R2 xy 2 (3y 2 − x2 )
σyy =− + , (7.3-17b)
πe1 R R r6 R r6

2c2 b⊥
1 R1 xy 2 (3x2 − y 2 ) R2 x2 (3y 2 − x2 )
σxy = σyx = − + , (7.3-17c)
πe1 R R r6 R r6
2
K0 b ⊥
1 y R1 −R22 2R1 R2
Hxx = −(e1 +e2 ) 2 +x h21 (x, y)− h22 (x, y) , (7.3-17d)
2πe1 r R2 R2
2
K0 b ⊥1 y R1 − R2
2
2R1 R2
Hyy =− h22 (x, y) + h 21 (x, y) , (7.3-17e)
2πe1 R2 R2
2
K0 b ⊥
1 x R1 −R22 2R1 R2
Hxy = (e1 +e2 ) 2 +y h21 (x, y)− h22 (x, y) , (7.3-17f)
2πe1 r R2 R2
2
K0 b ⊥1 x R 1 − R2
2
2R1 R2
Hyx =− h 22 (x, y) + h 21 (x, y) , (7.3-17g)
2πe1 R2 R2
in which
2xy(3x2 − y 2 )(3y 2 − x2 )
h21 (x, y) = ,
r8
(7.3-18)
2(x2 − y 2 ) (x2 − y 2 )(3x2 − y 2 )(3y 2 − x2 )
h22 (x, y) = + .
r4 r8
||
The solution to dislocation problem (b1 ,0,b⊥
1 ,0) can be determined by superpos-
7.4 Dislocations in quasicrystals with eight-fold symmetry 117
ing (7.3-13) and (7.3-14) onto (7.3-16) and (7.3-17). The solution to dislocation
||
problem (0,b2 ,0,b⊥
2 ) can be determined in the similar way. As a result, the solu-
|| ||
tion to dislocation problem with Burgers vector (b1 ,b2 ,b⊥ ⊥
1 ,b2 ) can be determined
completely.
The above work can be found in literature [6].
7.4 Dislocations in quasicrystals with eight-fold symmetry

The final governing equation of elasticity of two-dimensional quasicrystals with eight-
fold symmetry can be expressed
(∇2 ∇2 ∇2 ∇2 − 4ε∇2 ∇2 Λ2 Λ2 + 4εΛ2 Λ2 Λ2 Λ2 )F = 0, (7.4-1)
in which ⎧
⎪
⎪ 2 ∂2 ∂2 2 ∂2 ∂2
⎪
⎨ ∇ = ∂x2 + ∂y 2 , Λ = ∂x2 − ∂y 2 ,
(7.4-2)
⎪
⎪ R2 (L + M )(K2 + K3 )
⎪
⎩ ε=
[M (K1 + K2 + K3 ) − R2 ][(L + 2M )K1 − R2 ]
(see (6.6-10),(6.6-11) for detail). Equation (7.4-1) is more complicated than those
of (6.2-7) and (7.3-1), so the solution of which is also more complex than those
discussed in the previous sections. Due to the space limitation we cannot list whole
procedure of the solution, only give some main results of them in the following, in
which the Fourier transform and complex variable function methods are respectively
used.
7.4.1 Fourier transform method[9]

||
Consider dislocation problem b|| ⊕ b⊥ =(b1 ,0,b⊥1 ,0,0), we try to determine the dis-
placement field under action of the boundary conditions
⎧ #
⎪
⎪ σij (x, y) → 0, Hij (x, y) → 0, x2 + y 2 → ∞,
⎪
⎪
⎨
σyy (x, 0) = 0, Hyy (x, 0) = 0, (7.4-3)
⎪
⎪
⎪
⎪
⎩

dux = b1 , dwx = b⊥ 1.
Γ Γ
By performing the Fourier transform to (7.4-1), it can be reduced to

4 2 2 2 4
d2 2 d 2 d 2
−ξ − 4ε −ξ + 4ε +ξ F̂ = 0. (7.4-4)
dy 2 dy 2 dy 2
The eigen roots of equation (7.4-4) depend on the value of parameter ε, Zhou[9]
gave a detailed discussion for the solutions corresponding to case (1): 0 < ε < 1 and
case (2): ε < 0, but calculation is tremendous complex and lengthy which cannot
be included here. For the case (1) the solution is
⎧ 2 2

⎪
⎪ 1 b1 λ1 +λ22 y λ2 λ1 +λ22 y λ2
⎪
⎪ x
u (x, y) = arctan + +arctan −
⎪
⎪ 2π 2 λ x λ1 λ1 x λ1
⎪
⎪ 1
⎪
⎪
⎪
⎪ 2λ 3 xy 2λ1 xy
⎪
⎪ +(F 3 C + F4 D) arctan − arctan
⎪
⎪ x2 − (λ23 + λ24 )y 2 x2 − (λ21 + λ22 )y 2
⎪
⎪
⎪
⎪ x2 +2λ2 xy+(λ21 +λ22 )y 2 x2 +2λ4 xy+(λ23 +λ24 )y 2
⎪
⎪ 1
⎪
⎪ + F 5 ln 2 +F6 ln 2
⎪
⎪ 4π x −2λ2 xy+(λ21 +λ22 )y 2 x −2λ4 xy+(λ23 +λ24 )y 2
⎪
⎪

⎪
⎪ 2λ1 λ2 y 2
⎪
⎪ 1 λ2
⎪
⎪ uy = H1 arctan 2 − 2 arctan
⎪
⎪ 2π x + (λ21 − λ22 )y 2 λ1
⎪
⎪
⎪
⎪ 2
⎪
⎪ 2λ3 λ4 y λ4
⎪
⎪
⎪
+H2 arctan 2
x + (λ 2 − λ2 )y 2 − 2 arctan λ
⎪
⎪ 3 4 3
⎪
⎪

⎪
⎪ 1 4 2
x + 2(λ1 − λ2 )x y 2 2 2
⎪
⎪ + H3 ln 1 +
⎪
⎪ 4π (λ21 + λ22 )2 y 4
⎪
⎪
⎪
⎪
⎪
⎪ x4 + 2(λ23 − λ24 )x2 y 2
⎪
⎪ +H ln 1 + ,
⎪
⎪ 4
(λ23 + λ24 )2 y 4
⎪
⎪
⎪
⎪
2 2 2
⎨ 1 b⊥ 1 λ1 +λ2 y λ2 λ1 +λ22 y λ2
wx (x, y) = arctan + +arctan −
⎪
⎪
2π 2 λ1 x λ1 λ1 x λ1
⎪
⎪
⎪
⎪ 2λ3 xy
⎪
⎪ +(G C + G D) × arctan 2
⎪
⎪ 3 4
x − (λ23 + λ24 )y 2
⎪
⎪
⎪
⎪
⎪
⎪ 2λ1 xy
⎪
⎪ − arctan 2
⎪
⎪ x − (λ 2 + λ2 )y 2
⎪
⎪ 1 2
⎪
⎪
⎪
⎪ 1 x 2
+2λ xy+(λ 2 2 2
1 +λ2 )y x2 + 2λ4 xy+(λ23 +λ24 )y 2
⎪
⎪ + G5 ln 2
2
+G6 ln 2 ,
⎪
⎪ 4π x −2λ2 xy+(λ21 +λ22 )y 2 x −2λ4 xy+(λ23 + λ24 )y 2
⎪
⎪
⎪
⎪

⎪
⎪ 1 2λ1 λ2 y 2 λ2
⎪
⎪ w = I arctan − 2 arctan
⎪
⎪ y
2π
1
x2 + (λ21 − λ22 )y 2 λ1
⎪
⎪
⎪
⎪
⎪
⎪ 2λ3 λ4 y 2
λ4
⎪
⎪
⎪
⎪
+I2 arctan 2
x + (λ 2 − λ2 )y 2 − 2 arctan λ
⎪
⎪ 3 4 3
⎪
⎪

⎪
⎪ 1 4 2
x + 2(λ1 − λ2 )x y 2 2 2
⎪
⎪ + I3 ln 1 +
⎪
⎪ (λ21 + λ22 )2 y 4
⎪
⎪
4π
⎪
⎪
⎪
⎪ x4 + 2(λ23 − λ24 )x2 y 2
⎪
⎪ +I4 ln 1 + ,
⎪
⎩ (λ23 + λ24 )2 y 4
(7.4-5)
in which F1 , · · · , F6 , G1 , · · · , G6 , H1 , · · · , H4 and I1 , · · · , I4 are some functions of
λ1 , λ2 , λ3 and λ4 which are constants constituted from the original material constants
M, L, K1 , K2 , K3 and R, the expressions are very complicated and lengthy, we have
7.4 Dislocations in quasicrystals with eight-fold symmetry 119
to omit them.
By using the similar procedure, the solution for the case (2) can also be obtained.
But the solving procedure is very tedious due to the complexity of the final governing
equation (7.4-1). We omit them for simplicity.
7.4.2 Complex variable function method
The equation (7.4-1) can also be solved by complex variable function method. For
this purpose the equation can be rewritten as
8
∂ ∂8 ∂8
+ 4(1 − 4ε) + 2(3 + 16ε)
∂x8 ∂x6 ∂y 2 ∂x4 ∂y 4

∂8 ∂8
+4(1 − 4ε) 2 6 + 8 F = 0. (7.4-6)
∂x ∂y ∂y
The solution of equation (7.4-6) can be expressed in terms of four analytic functions
Fk (zk ) of complex variable zk (k = 1, 2, 3, 4), i.e.,
4

F (x, y) = 2Re Fk (zk ), zk = x + μk y (7.4-7)
k=1
and μk = αk + iβk (i = 1, 2, 3, 4) are complex parameters and determined by the

roots of the following eigenvalue equation:
μ8 + 4(1 − 4ε)μ6 + 2(3 + 16ε)μ4 + 4(1 − 4ε)μ2 + 1 = 0. (7.4-8)
Based on the displacement expressions (6.6-12) and the dislocation condition

dux = b1 , duy = b2 , dwx = b⊥ 1 , dwy = b⊥ 2, (7.4-9)
Γ Γ Γ Γ
we can obtain the solution as follows:

4
4

ux = 2Re a1k fk (zk ), uy = 2Re a2k fk (zk ),
k=1 k=1
4
4

wx = 2Re a3k fk (zk ), wy = 2Re a4k fk (zk ),
k=1 k=1
where
∂ 6 Fk (zk )
fk (zk ) = . (7.4-10)
∂zk6
Some detail of this work can be found in Ref. [10].
7.5 Dislocations in dodecagonal quasicrystals

In monograph [7] Ding et al give the solution in terms of the Green function method
which has not been introduced in the present book, so the detail is omitted and only
list the results, such as the phonon displacement field
⎧
⎪
⎪ b1 x2 L + M x 1 x2
⎪
⎪ u1 = 2π arctan x + L + 2M r2
⎪
⎪ 1
⎪
⎪
⎪
⎪
⎪
⎪ b2 M r L + M x22
⎪
⎪ + ln + ,
⎪
⎪ 2π L + 2M r0 L + 2M r2
⎪
⎨

⎪ b1 M r L + M x21 (7.5-1)
⎪
⎪ u2 = − ln + 2
⎪
⎪ 2π L + 2M r0 L + 2M r
⎪
⎪
⎪
⎪
⎪
⎪ b2 x2 L + M x1 x2
⎪
⎪ + arctan − ,
⎪
⎪ 2π x1 L + 2M r2
⎪
⎪
⎩
u3 = 0.
and the phason displacement field
⎧
⎪
⎪ b⊥
1 x2 (K1 + K2 )(K2 + K3 ) x1 x2
⎪
⎪ w1 = arctan −
⎪
⎪ 2π x1 2K1 (K1 + K2 + K3 ) r2
⎪
⎪
⎪
⎪ b⊥ K2 (K1 + K2 + K3 ) − K1 K3 (K1 + K2 )(K2 + K3 ) x22
⎪
⎪ 2 r
⎪
⎨ + − ln + ,
4π K1 (K1 + K2 + K3 ) r0 K1 (K1 + K2 + K3 ) r2
⎪
⎪
⎪ b⊥
1 K2 (K1 + K2 + K3 ) − K1 K3 r (K1 + K2 )(K2 + K3 ) x21
⎪
⎪ w2 = 4π ln −
⎪
⎪ K1 (K1 + K2 + K3 ) r0 K1 (K1 + K2 + K3 ) r2
⎪
⎪
⎪
⎪
⎪
⎪ b⊥ x2 (K1 + K2 )(K2 + K3 ) x1 x2
⎩ + 2 arctan + .
2π x1 2K1 (K1 + K2 + K3 ) r2
(7.5-2)
and the sign L, M are phonon elastic constants and others have the same mean-
ing as before. Because the phonons and phasons are decoupled for dodecagonal
quasicrystals, the coupling constant R = 0.
7.6 Interface between quasicrystal and crystal

In the previous sections we discussed dislocation problems in one- and two-dimensional
quasicrystals, a series analytic solutions are obtained. In Chapter 8 we will discuss
crack problems in the material. Apart from dislocations and cracks, interface is
a kind of defects in quasicrystals too, which is of particular significance for some
physical processes.
We know all quasicrystals observed to date are alloys. This type of alloys can
be in crystalline phase, in quasicrystalline phase, or in crystal-quasicrystal coex-
isting phase. Li et al[16,17] observed the crystal-quasicrystal phase transition, and
7.6 Interface between quasicrystal and crystal 121
which is continuous transition. The transition process gives rise to interface be-
tween crystal and quasicrystal. Consequently, analysis of the interface problem of
quasicrystals is of significance. In this section we give a phenomenological study on
elastic behaviour of the interface for one-dimensional quasicrystal-isotropic crystal.
Refs.[15∼17] pointed out that the phase transition is induced by the phason strains.
Such a problem is formidable. Here we focus only on the determination of the
strains, and further studies will be given in Chapter 9 for icosahedral quasicrystal-
cubic crystal interface.
Consider an orthorhombic quasicrystal at the upper half-plane (i.e., y > 0), the
phonon-phason coupling problem is governed by (5.4-3), i.e.,

∂4 ∂4 ∂4
a1 4 + a3 2 2 + a5 4 F = 0, (7.6-1)
∂x ∂x ∂y ∂y
in which F (x, y) denotes the displacement potential, a1 , a3 , a5 the material constants
composed of Cij , Ki and Ri defined by (5.3-6). We assume that the crystal coexisting
with the quasicrystal is located at the lower half-plane with thickness h(i.e., −h <
y < 0). The plane y = 0 is the interface between the quasicrystal and crystal,
as shown in Fig.7.6-1. For simplicity, suppose the crystal be an isotropic material,
(c)
characterizing by elastic constants Cij (E (c) , μ(c) ). At the interface there are the
Fig. 7.6-1 Coexisting phase of quasicrystal-crystal
y = 0, −∞ < x < ∞ : σzy = τ f (x) + ku(x), Hzy = 0, (7.6-2)

in which f (x) is the distribution function of applied stress at the interface, u(x) =
uz (x, 0) the value of displacement component of phonon field at the interface, τ a
constant shear stress and k a material constant to be taken as
μ(c)
k= (7.6-3)
h
and where μ(c) and h the shear modulus and the thickness of the crystal. We further
assume that the outer boundaries are stress free.
Taking the Fourier transform
∞
F̂ (ξ, y) = F (x, y)eiξx dx (7.6-4)
−∞
to the equation (7.6-1) yields

d4 d2
a5 4 + a3 ξ 2 2 + a1 ξ 4 F̂ = 0. (7.6-5)
dy dy
If put solution of equation (7.6-5) being
F̂ (ξ, y) = e−λ|ξ|y ,
where λ is a parameter, and substituting it into (7.6-5) leads to
a5 λ4 − a3 λ2 + a1 = 0 (7.6-6)
which has the roots & &
# #
2a3 ± a23 − 4a1 a5 2a3 ± a23 − 4a1 a5
λ1 , λ2 , λ3 , λ4 = ,− . (7.6-7)
2a5 2a5
There follows
F̂ (ξ, y) = Ae−λ1 |ξ|y + Be−λ2 |ξ|y + Ce−λ3 |ξ|y + De−λ4 |ξ|y .
By considering the condition of stress free at y = ∞, then C = D = 0, i.e.,
F̂ (ξ, y) = Ae−λ1 |ξ|y + Be−λ2 |ξ|y . (7.6-8)
According to Section 5.4, we have, e.g.
⎧
⎪
⎪ ∂3F
⎪
⎪ σ zy = (R 6 C 44 − R 5 C 55 ) ,
⎪
⎪ ∂x2 ∂y
⎪
⎨
∂2 ∂2 ∂ ∂
Hzy = − C55 2 + C44 2 K1 + K2 F, (7.6-9)
⎪
⎪ ∂x ∂y ∂x ∂y
⎪
⎪
⎪
⎪ ∂2 ∂2
⎪
⎩ uz = R6 2 + R5 2 F.
∂x ∂y
The Fourier transforms of the stresses and displacement are
⎧
⎪
⎪ dF̂
⎪
⎪ σ̂zy = −(R6 C44 − R5 C55 )ξ 2 ,
⎪
⎪ dy
⎪
⎨
2 2 dF̂ d2 F̂ d3 F̂
Ĥ zy = −iC55 K1 |ξ| ξ F̂ + C55 K2 ξ + iC44 K1 |ξ| 2 − C44 K2 3 ,
⎪
⎪ dy dy dy
⎪
⎪
⎪
⎪ d 2
⎪ û = −ξ 2 R + R
⎩ F̂ .
z 6 5
dy 2
(7.6-10)
7.6 Interface between quasicrystal and crystal 123
Substituting (7.6-8) into the second one of (7.6-10) then into the second one of (7.6-2)
yields
B = αA, (7.6-11)
where
−λ1 c2 − λ31 c4 + i(c1 + λ21 c3 )
α= ,
λ2 c2 + λ32 c4 + i(−c1 − λ22 c3 )
c1 = C55 K1 , c2 = C55 K2 , c3 = C44 K1 , c4 = C44 K2 . (7.6-12)
The Fourier transform of the phonon stress and displacement component at the
interface are
σ̂zy (ξ, 0) = A(ξ) |ξ| (λ1 e−λ1 |ξ|y + αλ2 e−λ2 |ξ|y ),
ûz (ξ, 0) = A(ξ)ξ 2 [(−R6 + λ21 R5 )e−λ1 |ξ|y + α(R6 − λ22 R5 )e−λ2 |ξ|y ]. (7.6-13)
The Fourier transform of the first equation of (7.6-2) is
σ̂zy (ξ, 0) = τ fˆ(ξ) + kûz (ξ, 0). (7.6-14)
From (7.6-13) and (7.6-14), one determines the unknown function
τ fˆ(ξ)
A(ξ) = . (7.6-15)
|ξ| (λ1 + αλ2 ) − kξ 2 [−R6 + λ21 R5 + α(R6 − λ21 R5 )]
Thus, the stress and displacement components for phonon and phason fields can be
evaluated, e.g.
⎧ ∞ ∞
⎪
⎪ 1 −iξx 1
⎪ σzy =
⎪ σ̂zy e dξ = A(ξ) |ξ|(λ1 e−λ1 |ξ|y + αλ2 e−λ2 |ξ|y )e−iξx dξ,
⎪
⎪ 2π −∞ 2π −∞
⎪
⎪ ∞
⎪
⎪
⎪
⎪ 1
⎪
⎪ u z = ûz e−iξx dξ
⎪
⎪ 2π −∞
⎪
⎪ ∞
⎨ 1
= A(ξ)ξ 2 [(−R6 + λ21 R5 )e−λ1 |ξ|y + α(R6 − λ22 R5 )e−λ2 |ξ|y ]e−iξx dξ,
⎪
⎪ 2π −∞
⎪
⎪ ∞
⎪
⎪
⎪
⎪ 1
ŵz e−iξx dξ
⎪
⎪ wz =
⎪
⎪ 2π −∞
⎪
⎪ ∞
⎪
⎪
⎪
⎪ =−
1
A(ξ)ξ 2 [(−C55 +λ21 C66 )e−λ1 |ξ|y +α(C66 −λ22 C55 )e−λ2 |ξ|y ]e−iξx dξ.
⎩
2π −∞
(7.6-16)
The phason strain field presents important effect in the phase transition of
crystal-quasicrystal, which is determined by the above solution such as
⎧ ∞
⎪
⎪ ∂wz 1
⎪
⎪ w zx = = −i A(ξ)ξ 3 [(−C55 + λ21 C66 )e−λ1 |ξ|y
⎪
⎪ ∂x 2π −∞
⎪
⎪
⎪
⎪ +α(C66 − λ22 C55 )e−λ2 |ξ|y ]e−iξx dξ,
⎪
⎪
⎨ ∂wz
wzy = (7.6-17)
⎪
⎪ ∂y
⎪
⎪ ∞
⎪
⎪ 1
⎪
⎪
⎪ = A(ξ)ξ 3 [λ1 |ξ| (−C55 + λ21 C66 )e−λ1 |ξ|y
⎪
⎪ 2π −∞
⎪
⎩ +λ |ξ| α(C − λ2 C )e−λ2 |ξ|y ]e−iξx dξ.
2 66 2 55
The solution varies from material constants Cij , Ki , Ri of quasicrystals, and the
material constant μ(c) of crystals, applied stress τ and the size h of the crystals, so
the results are interesting.
Further discussion on interface will be given in Section 9.2 of Chapter 9.
7.7 Conclusion and discussion

The main object of this chapter lies in demonstration on effect of the formulation in
Chapters 5 and 6, it does not intend to fully explore the nature of dislocations and in-
terfaces in quasicrystals. There are other work e.g. [11∼15,18,19] can be referenced.
It has been found that singularity around the dislocation core occurs, for instance,
stresses σij , Hij ∼ 1/r(r → 0), which is the direct result of broken symmetry due
to appearance of the topological defect, where r is the distance measured from the
dislocation core. This is similar to that of crystals, i.e., the broken symmetry also
leads to the appearance of singularity in crystals. Physically, the high stress grade
or the stress concentration around dislocation core results in plastic flow in crystals
as well as in quasicrystals. So dislocation and other defects in quasicrystals affect
the mechanical and physical properties of the material. The dislocation solutions
present important application in studying plastic deformation and plastic fracture,
as referring to [14] or Chapter 14 of this book. The discussion on interface here
is in a primary version, but it may be helpful to study crystal-quasicrystal phase
transition. However this is a very complicated problem, the knowledge in this re-
spect is very limited so far. The dislocations and interfaces and the solutions in
three-dimensional quasicrystals will be introduced in Chapter 9, and the dynamic
dislocation problem can be referred in Chapter 10.
References
[1] De P, Pelcovits R A. Linear elasticity theory of pentagonal quasicrystals. Phys Rev
B, 1987, 35(16): 8609–8620
[2] De P, Pelcovits R A. Disclination in pentagonal quasicrystals. Phys Rev B, 1987,
36(17): 9304–9307
References 125
[3] Ding D H, Wang R H, Yang W G et al. General expressions for the elastic displace-
ment fields induced by dislocation in quasicrystals. J Phys Condens Matter, 1995,
7(28): 5423–5436
[4] Ding D H, Wang R H, Yang W G et al. Elasticity theory of straight dislocation in
quasicrystals. Phil Mag Lett, 1995, 72(5): 353–359
sicrystals and solutions. Chin Phys Lett, 1998, 15(4): 278–280
[6] Li X F, Duan X Y, Fan T Y et al. Elastic field for a straight dislocation in a decagonal
quasicrystal. J Phys: Condens Matter, 1999, 11(3): 703–711
[7] Yang S H, Ding D H. Fundamentals to Theory of Crystal Dislocations. Vol II. Beijing:
Science Press, 1998 (in Chinese)
[8] Firth J P, Lothe J. Theory of Dislocations. John Wiley, Sons. New York, 1982
[9] Zhou W M. Dislocation, crack and contact problems in two- and three-dimensional
quasicrystals. Dissertation Beijing Institute of Technology, 2000 (in Chinese)
[10] Li L H. Study on complex variable function method and exact analytic solutions of
elasticity of quasicrystals. Dissertation. Beijing Institute of Technology, 2008 (in
Chinese)
[11] Fan T Y, Li X F, Sun Y F. A moving screw dislocation in a one-dimensional hexagonal
quasicrystals. Acta Physica Sinica (Oversea Edition), 1999, 8(3): 288–295
[12] Li X F, Fan T Y. A straight dislocation in one-dimensional hexagonal quasicrystals.
Phy Stat Sol (b), 1999, 212(1): 19–26
[13] Edagawa K. Dislocations in quasicrystals, Mater Sci Eng A, 2001, 309∼310(2):
528–538
[14] Fan T Y, Trebin H R, Messeschmidt U et al. Plastic flow coupled with a crack
in some one- and two-dimensional quasicrystals. J Phys: Condens Matter, 2004,
16(37): 5229–5240
[16] Li F H. In: Jacaman M J, Torres M. Crystal-Quasicrystal Transitions. Elsevier Sci
Publ, 1988, 13–47
[17] Li F H, Teng C M, Huang Z R et al. In between crystalline and quasicrystalline
states. Phil Mag Lett, 1988, 57 (1): 113–118
[18] Fan T Y, Xie L Y, Fan L et al. Study on interface of quasicrystal-crystal. Chin.
Phys. B, submitted 2009
[19] Kordak M, Fluckider T, Kortan A R et al. Crystal-quasicrystal interface in Al-Pd-
Mn. Prog Surface Sci, 2004, 75(3∼8): 161–175
Chapter 8
Application II: Solutions of notch and
crack problems of one-and
two-dimensional quasicrystals
Quasicrystals are potential materials to be developed for structural use, and their str-
ength and toughness attract attention of researchers. Experimental observations[1,2]
have shown that quasicrystals are brittle. With common experience of conventional
structural materials, we know that failure of brittle materials is mainly related to
the existence and growth of cracks. Chapter 7 indicated that dislocations have been
observed in quasicrystals, and the accumulation of dislocations will eventually lead
to cracking of the material. Now let us study crack problems in quasicrystals that
have both theoretical and practical value in the view of applications in future.
Chapters 5∼7 have discussed some elasticity and dislocation problems in one-
and two-dimensional quasicrystals. It has shown that when the quasicrystal config-
uration is independent of one coordinate, e.g. variable z, its elasticity problem can
be decoupled into a plane problem and an anti-plane problem. In the case of one-
dimensional quasicrystals, if the z-direction accords with the quasiperiodic axis, the
above plane problem belongs to classic elasticity problem, and the anti-plane problem
is a coupling problem of phonon and phason fields. In the case of two-dimensional
quasicrystals, if z-direction represents the periodic axis, the above plane problem
is a coupling problem of phonon and phason fields, and the anti-plane problem be-
longs to a classical elasticity problem. Due to use of decomposition procedure, the
resulting problem can be dramatically simplified. Chapters 5 and 6 have given their
corresponding fundamental solutions, and Chapter 7 conducted the solutions of dis-
locations in detail. The present chapter is going to focus on crack problems, to
continue using the above schemes, such as the fundamental solutions developed in
Chapters 5 and 6, and the Fourier transform and complex variable function meth-
ods used in Chapter 7. But it emphasizes the complex function method, which
will be developed in Sections 8.1, 8.2 and 8.4, this approach is powerful. Problems
displayed in Sections 8.1 and 8.2 are relatively simpler, the detailed introduction

128 Chapter 8 Application II: Solutions of notch and crack problems of one-and...
may help reader understand and further handle the principle and technique of the
complex potential method, though the representation does not beyond the classi-
cal Muskhelishvili[18] method, this is helpful to understand the solutions for more
complicated problems displayed in Section 8.4 and in the next chapter. The further
summary of the method will be introduced in Chapter 11, because the contents in
Section 8.4 and Chapter 9 bring some new insights into the study and go beyond
the Muskhelishvili method, the farther general discussion is necessary and may be
beneficial.
Based on the common nature of exact solutions of different static and dynamic
cracks in different quasicrystal systems for linear and nonlinear deformation (which
are discussed in this chapter and Chapters 9, 10 and 14), the fracture theory of qua-
sicrystalline material is suggested in Chapter 15, which can be seen as a development
of fracture mechanics of conventional structural materials.
8.1 Crack problem and solution of one-dimensional quasicrys-

tals
8.1.1 Griffith crack
As shown in Fig. 8.1-1, assume a Griffith crack along the quasiperiodic axis (z-
direction) of a one-dimensional hexangular quasicrystal. Under the action of ex-
(∞) (∞)
ternal traction σyz = τ1 and/or Hzy = τ2 . The deformation induced is often
termed longitudinal shearing. Obviously, the geometry of the crack is independent
of variable z. In this case, all field variables are independent of variable z. There-
fore, based on the analysis in Chapter 5, this one-dimensional quasicrystal elasticity
Fig. 8.1-1 A Griffith crack subjected to a longitudinal shear where τ = τ1 corresponding

to phonon field, and τ = τ2 corresponding to phason field
8.1 Crack problem and solution of one-dimensional quasicrystals 129
can be decomposed into a plane elasticity problem of regular crystal and an anti-
plane elasticity problem of the phonon-phason coupling field. Plane elasticity prob-
lems of regular crystal have been studied extensively in classical elasticity, and its
crack problems have also been studied in the classical fracture theory, see e.g. Ref.
[3]. Therefore, we skip the discussion of crack problems in regular crystal. The anti-
plane elasticity problem of the phason-phonon coupling field is described by using
the following basic equations:
⎧
⎪
⎪ σyz = σzy = 2C44 εyz + R3 wzy
⎪
⎪
⎪
⎨ σzx = σxz = 2C44 εzx + R3 wzx
(8.1-1)
⎪
⎪ Hzy = K2 wzy + 2R3 εzy
⎪
⎪
⎪
⎩ H = K w + 2R ε
zx 2 zx 3 zx
⎧ 1 ∂uz
⎪
⎪ εyz = εzy = ,
⎪
⎪ 2 ∂y
⎪
⎪
⎪
⎪
⎪
⎪ 1 ∂uz
⎪
⎨ εzx = εxz = 2 ∂x
(8.1-2)
⎪
⎪ ∂wz
⎪
⎪ wzy = ,
⎪
⎪ ∂y
⎪
⎪
⎪
⎪
⎪
⎩ w = ∂wz
zx
∂x
∂σzx ∂σzy ∂Hzx ∂Hzy
+ = 0, + =0 (8.1-3)
∂x ∂y ∂x ∂y
The derivation in Chapter 5 shows that the above equations can be reduced to
∇2 uz = 0, ∇ 2 wz = 0 (8.1-4)
where ∇2 = ∂ 2 /∂x2 + ∂ 2 /∂y 2 .

Figure 8.1-1 shows that the anti-plane deformation of a Griffith crack has the
#
x2 + y 2 → ∞ : σyz = τ1 , Hzy = τ2 , σzx = Hzx = 0
(8.1-5)
y = 0, |x| < a : σyz = 0, Hzy = 0
where a is the half-length of the crack.
Based on the results of linear elasticity analysis, if the quasicrystal is traction
free at infinity instead there are a traction σyz = −τ1 and a generalized traction
Hzy = −τ2 at the crack surface, then the boundary conditions stand for
#
x2 + y 2 → ∞ : σyz = σzx = Hzx = Hzy = 0
(8.1-6)
y = 0, |x| < a : σyz = −τ1 , Hzy = −τ2
Note that the phason stress τ2 at the crack surface is assumed from point of view
physically, though its measurement result has not been reported yet. For simplicity,
we can assume τ2 =0, sometimes.
In the following we are going to solve the boundary value problem of (8.1-4)
and (8.1-5) first. The complex variable function method will be used. To do so, we
introduce the complex variable
√
t = x + iy = reiθ , i = −1 (8.1-7)
From equations (8.1-4), it is known that both uz (x, y) and wz (x, y) are harmonic
functions that can be expressed in terms of the real part or the imaginary part of
two arbitrary analytic functions φ1 (t) and ψ1 (t) of complex variable t in a region
occupied by the quasicrystal. For simplicity we can call φ1 (t) and ψ1 (t) as complex
potentials. Here, assume
uz (x, y) = Reφ1 (t)
(8.1-8)
wz (x, y) = Reψ1 (t)
in which the symbol Re indicates the real part of a complex number.
It is well known that if a function F (t) is analytic, then
∂F dF ∂F idF
= , = . (a)
∂x dt ∂y dt
Furthermore, assume
F (t) = P (x, y) + iQ(x, y)

= ReF (t) + iImF (t), (b)
where symbol Im denotes the imaginary part of a complex number, and P (x, y) and
Q(x, y) represent the real and imaginary parts of F (t), respectively. Therefore, the
Cauchy-Riemann relations of an analytic function is
∂P ∂Q ∂P ∂Q
= , =− . (c)
∂x ∂y ∂y ∂x
With the aid of relation (a), formula (8.1-8) and equations (8.1-1) and (8.1-2)
lead to ⎧
⎪ ∂ ∂
⎪
⎪ σyz = σzy = C44 Reφ1 + R3 Reψ1
⎪
⎪ ∂y ∂y
⎪
⎪
⎪
⎪
⎪
⎪ ∂ ∂
⎪
⎨ σzx = σxz = C44 ∂x Reφ1 + R3 ∂x Reψ1
(8.1-9)
⎪
⎪ ∂ ∂
⎪
⎪ Hzx = K2 Reψ1 + R3 Reφ1
⎪
⎪ ∂x ∂x
⎪
⎪
⎪
⎪
⎪
⎪ ∂ ∂
⎩ Hzy = K2 Reψ1 + R3 Reφ1
∂y ∂y
Based on the Cauchy-Riemann relation (c), the above equation can be rewritten
as
σzx − iσzy = C44 φ1 + R3 ψ1
(8.1-10)
Hzx − iHzy = K2 ψ1 + R3 φ1
where φ1 = dφ1 /dt, ψ1 = dψ1 /dt.
According to formula (8.1-10), we have

σyz = σzy = −Im(C44 φ1 + R3 ψ1 )
(d)
Hzy = −Im(K2 ψ1 + R3 φ1 )
Furthermore, for an arbitrary complex function F (t), its imaginary part is

1
ImF (t) = (F − F )
2i
where F indicates the conjugate of F . Then the above formula (d) can be expressed
as ⎧
⎪ 1
⎪ σyz = σzy = − [C44 (φ1 − φ1 ) + R3 (ψ1 − ψ1 )] + τ1
⎨ 2i
(8.1-11)
⎪
⎪ 1
⎩ Hzy = − [K2 (ψ1 − ψ1 ) + R3 (φ1 − φ1 )] + τ2
2i
The expression given by (8.1-11) ensures the solution constructed in the following
to automatically satisfy the boundary condition at infinity, the solution is
⎧ ⎛ & ⎞
⎪
⎪ 2
⎪
⎪ ia(K 2 τ 1 − R 3 τ 2 ) t t
⎪
⎪ φ1 (t) = ⎝ − − 1⎠
⎪
⎨ C44 K2 − R32 a a
⎛ & ⎞ (8.1-12)
⎪
⎪ 2
⎪
⎪ ia(C τ
44 2 − R τ
3 1 ⎝ ) t t
⎪
⎪ ψ1 (t) = − − 1⎠
⎪
⎩ C44 K2 − R32 a a
whose derivation can be obtained in terms of the strict complex variable function
method given in Section 8.7—Appendix of Chapter 8.
From (8.1-12)
⎧
⎪
⎪ i(K2 τ1 − R3 τ2 ) t
⎪
⎨ φ1 (t) = C K − R2 1− √
44 2 3 t2 − a2
(8.1-13)
⎪
⎪ i(C 44 τ 2 − R 3 τ 1 ) t
⎪
⎩ ψ1 (t) = 1− √
C44 K2 − R32 t2 − a2
and substitution of the above expressions into the first expression in (8.1-11) yields

t
σzx − iσzy = iτ1 − √ (8.1-14)
t2 − a2
Separation of the real and imaginary parts of (8.1-14) leads to

⎧
⎪ τ1 r 1 1
⎪
⎨ σxz = σzx = (r1 r2 )1/2 sin θ − 2 θ1 − 2 θ2
⎪
⎪ (8.1-15)
⎪
⎪ τ1 r 1 1
⎩ σyz = σzy = cos θ − θ1 − θ2
(r1 r2 )1.2 2 2
where
t = reiθ , t − a = r1 eiθ1 , t + a = r2 eiθ2 (8.1-16)
or
⎧ # # #
⎪
⎪ r = x2 + y 2 , r1 = (x − a)2 + y 2 , r2 = (x + a)2 + y 2
⎨
y (8.1-16 )
⎪
⎪ y y
⎩ θ = arctan , θ1 = arctan , θ2 = arctan
x x−a x+a
which can be shown in Fig. 8.1-2.
Fig. 8.1-2 The coordinate system at the crack tip
Similarly
t
Hzx − iHzy = iτ2 − √ (8.1-17)
t2 − a2
For these stress components there are expressions similar to (8.1-15).
As a result, ⎧
⎨ √ τ1 x , |x| > a
σzy (x, 0) = x2 − a2 (8.1-18)
⎩
0, |x| < a
⎧
⎨ √ τ2 x , |x| > a
Hzy (x, 0) = x2 − a2 (8.1-19)
⎩
0, |x| < a
The above two formulas indicate that at y = 0 and |x| < a : σzy = 0, Hzy = 0.
Therefore, the solution given above also satisfies the boundary conditions at the
crack surfaces.
#
Formulas (8.1-14) and (8.1-17) also show that when x2 + y 2 → ∞, σyz = τ1 ,
σxz = 0 and Hzy = τ2 , Hzx = 0, namely, the solution given above satisfies boundary
conditions at infinity.
8.1.2 Brittle fracture theory
The above formulas show that stresses have singular characteristics near crack tips,
for example
⎧ τ1 x +
⎨ σzy (x, 0) = √x2 − a2 → ∞, x → a
⎪
(8.1-20)
⎪
⎩ Hzy (x, 0) = √ τ2 x → ∞, x → a+
x2 − a2
Define the mode III stress intensity factors of the phonon and phason fields such
that:
||
#
KIII = lim
+
2π(x − a)σzy (x, 0),
x→a
⊥
#
KIII = lim 2π(x − a)Hzy (x, 0),
x→a+
then
|| √ ⊥
√
KIII = πaτ1 , KIII = πaτ2 , (8.1-21)
where suffix “III” stands for model III (longitudinal shearing mode) [3].
Now let us calculate the crack strain energy:
a
WIII = 2 (σzy ⊕ Hzy ) (uz ⊕ wz )dx
0 a
=2 [σzy (x, 0)uz (x, 0) + Hzy (x, 0)wz (x, 0)]dx (8.1-22)
0
From (8.1-8) and (8.1-12), we have

⎧ + x 2
⎪
⎪ K2 τ1 − R3 τ2
⎪
⎨ u z (x, 0) = Re(φ 1 (t)) t=x = a 1 − |x| < a
C44 K2 − R32 a
+ (8.1-23)
⎪
⎪
⎪ C τ − R3 τ 1 x 2
⎩ wz (x, 0) = Re(ψ1 (t))t=x = a 44 2 2 1− |x| < a
C44 K2 − R3 a
In addition, considering the equivalency of problems (8.1-5) and (8.1-6), we can

take σyz (x, 0) = −τ1 and Hzy = −τ2 as |x| < a, then substitution of above results
into (8.1-22) yields
K2 τ12 + C44 τ22 − 2R3 τ1 τ2 2
WIII = πa (8.1-24)
C44 K2 − R32
From (8.1-24), we can determine the crack strain energy release rate (crack growth
force) such that
1 ∂WIII K2 τ12 + C44 τ22 − 2R3 τ1 τ2

GIII = = πa
2 ∂a C44 K2 − R32
2 ⊥2
⊥
K2 KIII + C44 KIII − 2R3 KIII KIII
= 2 (8.1-25)
C44 K2 − R3
Clearly, the crack energy and energy release rate are related not only to the phonon
but also to the phason and the phonon-phason coupling fields.
If τ2 = 0, then

K2 (KIII )2
GIII = (8.1-26)
C44 K2 − R32
Furthermore if R3 = 0, we have
||
πaτ12 (KIII )2
GIII = , or GIII = . (8.1-27)
C44 C44
Since GIII comprehensively describes the coupling effect of the phonon and phason
fields and the stress states near crack tip, we recommend
GIII = GIIIC (8.1-28)
as the fracture criterion of quasicrystals under mode III deformation, where GIIIC
is the critical (threshold) value of GIII , the mode III fracture toughness to be deter-
mined by testing.
The stress intensity factor and the strain energy release rate are fundamental
physical parameters, and constitute the basis of brittle fracture theory for both the
conventional crystalline as well as the quasicrystalline materials.
8.2 Crack problem in finite-sized one-dimensional quasicrys-

tals
The preceding section discussed the Griffith crack problems in one-dimensional qua-
sicrystals, and obtained their exact solutions. In this section, we are going to discuss
the solution of another kind of important crack problems, which go beyond the scope
of the Griffith crack. In the preceding section, we assumed that the size of quasicrys-
tal be much more larger than that of the defect, while in this case we consider the
quasicrystal as an infinite body. Therefore, this section is aimed to the study of
crack problems in quasicrystals of finite size.
8.2 Crack problem in finite-sized one-dimensional quasicrystals 135
8.2.1 Cracked quasicrystal strip with finite height

As shown in Fig. 8.2-1, a one-dimensional hexangonal quasicrystal strip of height 2H
has a semi-infinite crack embedded at the mid plane with crack tip corresponding to
coordinate origin. Crack surfaces near the crack tip, i.e., y = ±0 and −a < x < 0,
are assumed under the action of uniform shear traction σzy = −τ1 , Hzy = 0, where
a is a length used to simulate a finite-size crack. The upper and lower strip surfaces
are assumed being traction free, therefore the boundary conditions are
⎧
⎪
⎪ y = ±H, −∞ < x < ∞ : σzy = 0, Hzy = 0,
⎪
⎪
⎪
⎨ x = ±∞, −H < y < H : σzx = 0, Hzx = 0,
(8.2-1)
⎪
⎪ y = ±0, −∞ < x < −a : σzy = 0, Hzy = 0,
⎪
⎪
⎪
⎩ y = ±0, −a < x < 0 : σ = −τ , H = 0.
zy 1 zy
Fig. 8.2-1 Finite strip of hexangonal quasicrystal with a crack
Formulas (8.1-1)∼(8.1-10) in the preceding section still hold in this section, and
other notation and symbol are similar. Therefore, formula (8.1-11) should be modi-
fied as
C44 (φ1 − φ1 ) + R3 (ψ1 − ψ1 ) = 2iτ1 f (x),
(8.2-2)
K2 (ψ1 − ψ1 ) + R3 (φ1 − φ1 ) = 0,
where
0, x < −a,
f (x) = (8.2-3)
1, −a < x < 0.
The conformal mapping
2
H 1+ζ
t = ω(ζ) = ln 1 + , (8.2-4)
π 1−ζ
maps the domain in the t-plane onto the interior of the unit circle γ in the ζ-plane,
ζ = ξ + iη. Therefore, crack tip t = 0 corresponds to ζ = −1, and t = −a accords to
two points on the unit circle in the ζ-plane such that
⎧ √
⎪
⎪ −e−πa/H + 2i 1 − e−πa/H
⎪ σ
⎨ −a = ,
2 − e−πa/H
√ (8.2-5)
⎪
⎪ −πa/H
⎩ σ−a = −e
⎪ − 2i 1 − e−πa/H
,
2 − e−πa/H
where σ = eiϕ = ζ||ζ|=1 denotes the value of ζ on γ.
Equations in Section 8.1 and the Appendix of this chapter (i.e., Section 8.7) are
useful. Nevertheless, the first 2iω (σ)τ in (8.7-5) should be changed to 2if ω (σ)τ ,
and the first term in (8.7-5)

2iτ 1 ω (σ)
dσ
C44 2πi γ σ−ζ
should be modified as
2iτ 1 ω (σ)
f dσ.
C44 2πi γ σ−ζ
Thus, equation (8.7-5) for the present problem becomes
⎧
⎪ R3 2iτ1 1 ω (σ)
⎪ φ (ζ) +
⎨ ψ (ζ) = f dσ,
C44 C44 2πi γ σ − ζ
(8.2-6)
⎪
⎪ R3
⎩ ψ (ζ) + φ (ζ) = 0,
K2
where f is the function given by formula (8.2-3), which takes values between σ−a
and σ−a in the ζ-plane.
Integration of the right-handside of equation (8.2-6) leads to

1 ω (σ) 1 1 1+ζ
f dσ = ln(σ − 1) − ln(σ − ζ)
2πi γ σ − ζ 2πτ 1 − ζ (1 − ζ)(1 + ζ 2 )
σ=σ−a
ζ 2 1 σ−i
− ln(1 + σ ) + ln
2(1 + ζ 2 ) 2(1 − ζ 2 ) σ + i σ=σa
≡F (ζ) (8.2-7)
where σ−a and σ−a are given by formula (8.2-5). Based on the above relation, (8.2-6)
is further reduced to
K2 τ1 R3 τ1
φ (ζ) = 2iF (ζ), ψ = − 2iF (ζ). (8.2-8)
C44 K2 − R32 C44 K2 − R32
Now let us calculate the stresses. During this process, the following relations will
be used:
φ (ζ) 2iK2 τ1 F (ζ)
φ1 (t) = = · ,
ω (ζ) C44 K2 − R32 ω (ζ)
8.2 Crack problem in finite-sized one-dimensional quasicrystals 137
ψ (ζ) 2iR3 τ1 F (ζ)

ψ1 (t) = = − · .
ω (ζ) C44 K2 − R32 ω (ζ)
Substitution of the above expressions into (8.1-10) leads to

⎧
⎪
⎨ σzx − iσzy = iτ1 F (ζ) ,
ω (ζ) (8.2-9)
⎪
⎩ H − iH = 0,
zx zy
which shows that the stress distribution is independent of the material constants.
Substitution of (8.2-7) into (8.2-9) leads to the explicit expressions of stresses.
The forms of F (ζ) and F (ζ) are relatively complex, while the expressions of stress
in terms of variable ζ(= ξ + iη) are concise. In an attempt to invert them to the
t-plane, inverse of transform (8.2-4) must be used,
√
−1 −(eπt/H − 2) ± 2i eπt/H − 2
ζ = ω (t) = . (8.2-10)
eπt/H − 2
Substitution of (8.2-10) into (8.2-9) yields to the final expressions of σzx , σzy , Hzx
and Hzx in the t-plane, which are very complex. Here we skip this procedure.
Now let us calculate the stress intensity factors. According to (8.1-15) and
(8.1-16), it is known that at the region such that r1 /a 1,
|| ||
K θ1 K θ1
σzx = − √ III sin , σzy = √ III cos ,
2πr1 2 2πr1 2
therefore
||
K θ1 θ1 K
σzx − iσzy = − √ III sin − i cos = − √ III , (8.2-11)
2πr1 2 2 2πt1
where t1 = r1 eiθ1 = x1 + iy1 .

By using conformal mapping t = ω(ζ), we have
t1 = ω(ζ1 ), (8.2-12)
where ζ1 is the point in the ζ-plane corresponding to t1 . With relations (8.2-11),

(8.2-12), and (8.2-9), we define
√ F (ζ) ⊥
KIII = lim πτ1 # , KIII =0 (8.2-13)
ζ→−1 ω (ζ)
where ζ = −1 is the point corresponding to the crack tip. Substitution of (8.2-4)

and (8.2-7) into (8.2-13) yields
√ √
2Hτ1 2eπa/H − 1 + 2eπa/H 1 − e−πa/H
KIII = ln √ , (8.2-14)
2π 2eπa/H − 1 − 2eπa/H 1 − e−πa/H
⊥
If we do not assume τ2 = 0 in (8.2-1), then the stress intensity factor KIII can be
evaluated, and the expression is similar to (8.2-14), this is the extension of work
given by Ref. [4] for classical elasticity to quasicrystal elasticity.
8.2.2 Finite strip with two cracks
The configuration is shown in Fig. 8.2-2, there are the following boundary conditions
⎧
⎪ y = ±H, −∞ < x < ∞ : σzy = 0, Hzy = 0
⎪
⎪
⎪
⎪
⎪
⎪ x = ±∞, −H < y < H : σzx = 0, Hzx = 0
⎪
⎪
⎨
y = ±0, −∞ < x < −a : σzy = 0, Hzy = 0; (8.2-15)
⎪
⎪
⎪
⎪
⎪
⎪ −a < x < 0 : σzy = −τ1 , Hzy = −τ2 ;
⎪
⎪
⎪
⎩
L < x < ∞ : σzy = 0, Hzy = 0
The following conformal mapping function:

2
H 1 + α (1 − ζ/1 + ζ)
t = ω(ζ) = ln (8.2-16)
π 1 + βα (1 − ζ/1 + ζ)2
transforms the region at z-plane onto the interior of the unit circle γ at ζ-plane, in
which
1 − e−πa/H
α= , β = e−πL/H . (8.2-17)
1 − e−π(a+L)/H
Fig. 8.2-2 Two cracks in a strip
Then substitute (8.2-17) into (8.2-6), we can find the solution φ (ζ) so the stress
intensity factors such as
8.3 Griffith crack problems in point groups 5m and 10mm quasicrystals based... 139
⎧

⎪ (0,0) # F (ζ)
⎪
⎪ KIII = lim −2 2πω(ζ)τ1
⎪
⎪ ω (ζ)
⎪
⎪
ζ→1+0
⎪
⎪ √ √ √
⎪
⎪ 2Hτ1 1+ α √ 1 + αβ
⎪
⎪ = √ ln √ − β ln √ ,
⎨ π 1−β 1− α 1 − αβ

(8.2-18)
⎪
⎪ (L,0) # F (ζ)
⎪
⎪ KIII = lim −2 2π(L − ω(ζ))τ1
⎪
⎪ ω (ζ)
⎪
⎪
ζ→−1−0
⎪
⎪ √ √ √
⎪
⎪ 2Hτ1 √ 1+ α 1 + αβ
⎪
⎩ = √ β ln √ − ln √ ,
π 1−β 1− α 1 − αβ
in which
K2 τ1 R3 τ1
φ (ζ) = 2iF (ζ), ψ (ζ) = − 2iF (ζ),
C44 K2 − R32 C44 K2 − R32
√ √
2H αA αβM
F (ζ) = 2 −
π (1 + ζ)2 + α(1 − ζ)2 (1 + ζ)2 + αβ(1 − ζ)2
2H iα(1 − β)(1 − ζ 2 ) i−ζ
+ 2 2 2 2 2
ln , (8.2-19)
π [(1 + ζ) + α(1 − ζ) ][(1 + ζ) + αβ(1 − ζ) ] 1 − iζ
√ √
(1 + α) 1 + αβ
A = ln √ , M = ln √ .
1− α 1 − αβ
⊥
If we do not assume τ2 = 0 in (8.2-15), then the stress intensity factor KIII can
be similarly evaluated, and the expression is similar to (8.2-18), this extends the
study for the classical elasticity. The detail can be found in Refs. [5] and [6], some
calculations on the function F (ζ) refer to Major Appendix of this book.
8.3 Griffith crack problems in point groups 5m and 10mm

quasicrystals based on displacement potential function
method
Literature [2] reported that Chinese materials scientists had started to characterize
the fracture toughness of quasicrystals. Due to the lack of theoretical solutions to
crack problems in quasicrystals, their experimental work was performed largely by
indirectly measuring the fracture toughness. If the crack solution had been known
then, the fracture toughness could have been determined by direct measurement, a
much more simpler and accurate method.
This section aims to solve the mode I Griffith crack in quasicrystal with ten-fold
symmetry by using method of displacement functions. Therefore, the fundamental
formulas in Section 6.2 and formulas using Fourier transform in Section 7.2 are the
basis for this section. To save space, we do not plan to list those formulas in detail
here, and interested readers may refer to the previous two chapters. In the next
section, we are going to solve this problem using method of stress functions. On
the one hand, this demonstrates the problem-solving procedure based on method of
stress functions; on the other hand this is targeted to examine the results obtained
by the method of displacement functions. For a correct solution, which can be
examined using any available method.
(∞)
Consider a Griffith crack under the action of external traction, i.e., σyy = p, and
the crack is assumed to penetrate the periodic axis(z-direction) of the quasicrystals,
as shown in Fig. 8.3-1. Similar to the analysis in the preceding section, within
the framework of Griffith’s theory, this problem can be replaced by an equivalent
crack problem shown in Fig. 8.3-2. Furthermore, assume the external traction being
independent of z, therefore the deformation of the quasicrystal is also independent
of z, namely
∂ui ∂wi
= 0, = 0, i = 1, 2, 3. (8.3-1)
∂z ∂z
According to the analysis performed in Chapter 6, under this case, the two-
dimensional quasicrystal elasticity problem can be decoupled into a plane elasticity
problem of phonon-phason coupling and an anti-plane pure elasticity problem. In
this case, the latter only has a trivial solution under mode I external traction, which
can be neglected. The plane elasticity problem of phonon-phason coupling with
point groups 5m and 10mm has been studied in Section 6.2, and its final governing
equation is
∇2 ∇2 ∇2 ∇2 F = 0. (8.3-2)
Here F (x, y) is the displacement potential function introduced in Section 6.2.
Fig. 8.3-1 Griffith crack along the periodic axis of quasicrystal and
subjected to a tension
Fig. 8.3-2 The same Griffith crack as Figure 8.3-1 with external traction
acting on crack surfaces
As shown in Fig. 8.3-2, the Griffith crack is under the action of uniform traction
at crack surfaces and without far-field traction, i.e., σyy (x, 0) = −p, |x| < a. We
modify this problem into the semi-plane problem, i.e., only study the case in the
upper half-plane or the lower half-plane under the following conditions:
⎧ #
⎪
⎪ x2 + y 2 → ∞ : σij = Hij = 0,
⎪
⎪
⎪
⎪
⎪
⎪ y = 0, |x| < a : σyy = −p, σyx = 0,
⎨
Hyy = 0, Hyx = 0, (8.3-3)
⎪
⎪
⎪
⎪
⎪
⎪ y = 0, |x| > a : σyx = 0, Hyx = 0,
⎪
⎪
⎩
uy = 0, wy = 0.
By performing the Fourier transform on equation (8.3-2)
∞
F̂ (ξ, y) = F (x, y)eiξx dx. (8.3-4)
−∞
which is therefore reduced to an ordinary differential equation such that

2 4
d 2
−ξ F̂ (ξ, y) = 0. (8.3-5)
dy 2
If we choose the upper half-plane y > 0 for our study, the solution of the above
equation is
F̂ (ξ, y) = (4ξ 4 )−1 XY e−|ξ|y , (8.3-6)
where X = (A, B, C, D), Y = (1, y, y 2 , y 3 )T , and A, B, C, and D are arbitrary
functions with respect to ξ to be determined according to boundary condition, “T”
stands for the transpose of a matrix. Fourier transforms of the displacement and
stress components can be expressed in term of F̂ (ξ, y), i.e., X and Y as discussed in
Section 7.2.
Solution (8.2-6) has satisfied the boundary condition (8.2-3) at infinity, and the
left boundary condition in (8.2-3) results in
⎧
⎪ 21C(ξ)|ξ| − 3(32 − e2 )D(ξ)
⎪
⎨ A(ξ) = 3 ,
2 |ξ|
(8.3-7)
⎪
⎩ B(ξ) = 6C(ξ)|ξ| − 21D(ξ)
⎪
ξ2
and the following set of dual integral equations:
⎧ ∞
⎪ 2
⎪
⎪ [C(ξ)ξ − 6D(ξ)] cos(ξx)dξ = −p, 0 < x < a,
⎪
⎪
⎪ d11 0
⎪
⎪
⎪ ∞
⎪
⎪
⎪
⎨ ξ −1 [C(ξ)ξ − 6D(ξ)] cos(ξx)dξ = 0, x > a,
0
∞ (8.3-8)
⎪
⎪ 2
⎪
⎪ D(ξ) cos(ξx)dξ = 0, 0 < x < a,
⎪
⎪ d12 0
⎪
⎪
⎪
⎪ ∞
⎪
⎪
⎩ ξ −1 D(ξ) cos(ξx)dξ = 0, x > a.
0
Here, e2 is given by the second relation in (7.2-12), i.e.,
2αβ α−β
e2 = + ,
ω(α − β)(K1 − K2 ) α + β
α and β are given in (6.2-5), i.e.,
M (L + 2M )
α = R(L + 2M ) − ωK1 , β = RM − ωK1 , ω= ,
R
d11 and d22 are given as
⎧ nR
⎪
⎪ d11 = ,
⎪
⎪ (4M/L + M )(L + 2M )(M K1 − R2 )
⎪
⎪
⎪
⎪
⎨ nR2
d12 = , (8.3-9)
⎪ d0 M (L + 2M )
⎪
⎪
⎪
⎪ d0 = −{(M K1 − R2 )[(L + 2M )(K1 + K2 ) − 2R2 ]
⎪
⎪
⎪
⎩
−[(L + 2M )K1 − R2 ][M (K1 + K2 ) − 2R2 ]}
and n is determined by (7.2-12), namely,
n = M α − (L + 2M )β.
The theory of dual integral equations is offered in the Major Appendix of this mono-
graph. Accordingly, solution to the set of equations (8.3-8) is
2C(ξ)ξ = d11 pαJ1 (aξ), D(ξ) = 0, (8.3-10)
where J1 (aξ) is the first order Bessel function of the first kind.
So far, the unknown functions A(ξ), B(ξ), C(ξ), and D(ξ) have been determined
completely. In the view of mathematics, this problem has been solved in the Fourier
transform space. However, in the view of physics, we need to perform the Fourier
inverse ∞
1
F (x, y) = F̂ (ξ, y)e−iξx dξ (8.3-11)
2π −∞
in order to express the field variables in the physical space.
Obviously, once F (x, y) is determined from the integral (8.3-11), uj , σjk , and
Hjk can be determined by substituting F (x, y) into (8.3-8)∼(8.3-11). Alternatively,
ûj (ξ, y), ŵj (ξ, y), σ̂jk (ξ, y), and Ĥjk (ξ, y) can be determined by substituting X(ξ)
and Y (ξ) into (8.3-8)∼(8.3-11), and then their Fourier inverses finally lead to uj , σjk ,
and Hjk . Luckily, the above integrals with Bessel function can be expressed explicitly
using elementary functions. Nevertheless, their final expressions in terms of variables
x and y are extremely complex. However, the final expressions appear more concise
if using (r, θ), (r1 , θ1 ), and (r2 , θ2 ) to stand for the three polar coordinate systems
with the crack center, left crack tip, and right crack tip as origins, respectively, as
shown in Fig. 8.1-2, similar to (8.1-16), i.e.,

x = r cos θ = a + r1 cos θ1 = −a + r2 cos θ2 ,
(8.3-12)
y = r sin θ = r1 sin θ1 = r2 sin θ2 .
The following infinite integrals involving Bessel functions are used for stress cal-
culation: ⎧
⎪
⎪
∞
1 z
⎪
⎪ −ξz
J1 (aξ)e dξ = 1− ,
⎪
⎪ 0 1/2
⎪
⎪
a (a2 − z 2 )
⎪
⎪ ∞ ∞
⎪
⎪
⎪
⎪ ξJ1 (aξ)e−ξz dξ =
a
⎪
⎨ 0 ξJ 1 ,
0 (a2 − z 2 )3/2
(8.3-13)
⎪ ∞
⎪ 3az
⎪
⎪ 2 −ξz
ξ J1 (aξ)e dξ = ,
⎪
⎪
⎪
⎪ 0 (a2 − z 2 )5/2
⎪
⎪ ∞
⎪
⎪
⎪
⎪ 3 −ξz 3a(4z 2 − a2 )
⎪
⎩ ξ J 1 (aξ)e dξ = 7/2
,
0 (a2 − z 2 )
where z = x + iy.
After proper calculation, we have

⎧
⎪
⎪ σxx = −p[1 + r(r1 r2 )−3/2 cos(θ − θ)] − pr(r1 r2 )−3/2 sin θ sin 3θ,
⎪
⎪
⎪
⎪ σyy = −p[1 − r(r1 r2 )−3/2 cos(θ − θ)] + pr(r1 r2 )−3/2 sin θ sin 3θ,
⎪
⎪
⎪
⎪
⎪
⎪ σxy = σyx = pr(r1 r2 )−3/2 sin θ cos 3θ,
⎪
⎨
Hxx = −4d21 pr(r1 r2 )−3/2 sin θ cos 3θ − 6d21 pr3 (r1 r2 )−5/2 sin2 θ cos(θ − 5θ),
⎪
⎪
⎪
⎪ Hyy = −6d21 pr3 (r1 r2 )−5/2 sin2 θ cos(θ − 5θ),
⎪
⎪
⎪
⎪
⎪
⎪ Hxy = 6d21 pr3 (r1 r2 )−5/2 sin2 θ sin(θ − 5θ),
⎪
⎪
⎪
⎩
Hyx = 4d21 pr(r1 r2 )−3/2 sin θ cos 3θ + 6d21 pr3 (r1 r2 )−5/2 sin2 θ sin(θ − 5θ),
(8.3-14a)
in which r = r/a, r1 = r1 /a, r2 = r2 /a, θ = (θ1 + θ2 )/2,
R(K1 − K2 )
d21 = (8.3-14b)
4(M K1 − R2 )
Similarly, displacement components uj and wj can be also expressed in terms of
elementary functions. We list here only the component
⎧
⎨ 0, |x| > a,
uy (x, 0) = p K1 1 √ (8.3-15)
⎩ + 2 2
a − x , |x| < a,
2 M K1 − R2 L + M
From these results, we can find the stress intensity factor

# √
KI = lim+ 2π(x − a)σyy (x, 0) = πap. (8.3-16)
x→a
and from (8.3-15) we have the crack strain energy

a
WI =2 (σyy (x, 0) ⊕ Hyy (x, 0))(uy (x, 0) ⊕ wy (x, 0))dx
0

πa2 p2 1 K1
= + . (8.3-17)
4 L+M M K1 − R2
So there is the crack energy release rate

1 ∂W1 1 1 K1
GI = = + (KI )2 . (8.3-18)
2 ∂a 4 L+M M K1 − R2
It is evident that the crack energy release rate depends on not only phonon elastic
constants L(= C12 ), M (= (C11 − C12 )/2), but also phason elastic constant K1 and
phonon-phason coupling elastic constant R.
The further meaning and applications of these quantities will be discussed in
detail in Chapter 15. This provides the basis of fracture mechanics of quasicrystalline
materials.
The detail on the work can be found in Ref [7].
8.4 Stress potential function formulation and complex variable function method... 145
8.4 Stress potential function formulation and complex vari-

able function method for solving notch and crack prob-
lems of quasicrystals of point groups 5, 5̄ and 10, 10
The Fourier method cannot solve notch problems, while complex variable function
method with conformal mapping can solve them. In this section we develop complex
potential method for notch/crack problems for plane elasticity of point groups 5, 5̄
and 10, 10 quasicrystals. We will use the stress potential formulation[9] , of course,
we can also use the displacement potential formulation given in Section 6.4.
8.4.1 Complex function method
From Section 6.7, we find that, based on stress potential method the final governing
equation of plane elasticity of point group 10, 10 decagonal quasicrystals is
∇2 ∇2 ∇2 ∇2 G = 0. (8.4-1)
The general solution of the equation (8.4-1) is

1 1
G = 2Re g1 (z) + z̄g2 (z) + z̄ 2 g3 (z) + z̄ 3 g4 (z) , (8.4-2)
2 6
where gi (z)(i = 1, 2, 3, 4) are four analytic functions of a single complex variable

z = x + iy = reiθ . The bar over the quantity denotes the complex conjugate
hereinafter, i.e., z̄ = x − iy = re−iθ . We can see that the complex analysis here
will be more complicated than that of Muskhelishvili method for plane elasticity of
classical elasticity.
8.4.2 The complex representation of stresses and displacements
Substituting expression (8.4-2) into equation (6.7-6) then into equations (6.7-2) leads
to
⎧
⎪ σxx = −32c1 Re[Ω (z) − 2g4 (z)],
⎪
⎪
⎪
⎪
⎪
⎪ σyy = 32c1 Re[Ω (z) + 2g4 (z)],
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ σxy = σyx = 32c1 ImΩ (z),
⎪
⎨
Hxx = 32R1 Re[Θ (z) − Ω (z)) − 32R2 Im(Θ (z) − Ω (z)], (8.4-3)
⎪
⎪
⎪
⎪
⎪
⎪ Hxy = −32R1 Im[Θ (z) + Ω (z)) − 32R2 Re(Θ (z) + Ω (z)],
⎪
⎪
⎪
⎪
⎪
⎪ Hyx = −32R1 Im[Θ (z) − Ω (z)) − 32R2 Re(Θ (z) − Ω (z)],
⎪
⎪
⎪
⎪
⎩
Hyy = −32R1 Re[Θ (z) + Ω (z)) + 32R2 Im(Θ (z) + Ω ],
where ⎧
⎨ Θ(z) = g (IV) (z) + z̄g (IV) (z) + 1 z̄ 2 g (IV) (z)
2 3 4
2 (8.4-4)
⎩ (IV) (IV)
Ω (z) = g3 (z) + z̄g4 (z)
in which the prime, two prime, three prime and superscript (IV) denote the first
to fourth order derivatives of gj (z)(j = 1, · · · , 4) to variable z, in addition Θ (z) =
dΘ(z)/dz.
We further derive the complex representations of displacement components of
phonon and phason fields. The first two equations of (6.7-3) can be rewritten as
⎧
⎪ K1 + K2 1
⎪
⎨ εxx = c2 (σxx + σyy ) − σyy − [R1 (Hxx + Hyy ) + R2 (Hxy − Hyx )],
2c 2c
⎪
⎪ K + K 1
⎩ εyy = c2 (σxx + σyy ) − 1 2
σxx + [R1 (Hxx + Hyy ) + R2 (Hxy − Hyx )],
2c 2c
(8.4-5)
where
c + (L + M )(K1 + K2 )
c2 = . (8.4-6)
4(L + M )c
Substituting equations (8.4-3) into (8.4-5) and by integration yield
K1 + K2 ∂
ux =128c1 c2 Reg4 (z) − φ
2c ∂x
32(R12 + R22 )
+ Re[g3 (z) + z̄g4 (z) − g4 (z)] + f1 (y),
c
K1 + K2 ∂
uy =128c1 c2 Img4 (z) − φ
2c ∂y
32(R12 + R22 )
− Im[g3 (z) + z̄g4 (z) + g4 (z)] + f2 (x).
c
With these results and other equations of equations (6.7-3) one finds that
df1 (y) df2 (x)
− = .
dy dx
This means these two functions must be constants which only give rigid-body dis-
placements. Omitting the trial functions f1 (y), f2 (x), one obtains
ux + iuy = 32(4c1 c2 − c3 − c1 c4 )g4 (z) − 32(c1 c4 − c3 )[g3 (z) + zg4 (z)], (8.4-7)
where
R12 + R22 K1 + K 2
c3 = , c4 = . (8.4-8)
c c
Similarly, the complex representations of displacement components of phason
fields can be expressed as follows:
32(R1 − iR2 )
wx + iwy = Θ(z). (8.4-9)
K1 − K2
8.4.3 Elliptic notch problem

To illustrate the effect of the stress potential and complex variable function method
to the complicated stress boundary value problems of eight order partial differential
equations given above, we here calculate the stress and displacement fields induced
x2 y2
by an elliptic notch L : 2 + 2 = 1, see Fig. 8.4-1 (a), the edge of which is
a b
subjected to a uniform pressure p, this problem is equivalent to the case that the
body is subjected to a tension at infinity and the surface of the notch is stress free
shown in Fig. 8.4-1(b) if α = π/2. The equivalency is proved in Subsection 11.3.9 of
Chapter 11.
Fig. 8.4-1 (a) The elliptic notch in a decagonal quasicrystal subjected to a uniform
pressure and traction tree at infinity; (b) An infinite decagonal quasicrystal with an
elliptic notch subjected to a tension
For the problem shown by Fig. 8.4-1 (a) the boundary conditions can be expressed
as follows:
σxx cos(n, x) + σxy cos(n, y) = Tx , σxy cos(n, x) + σyy cos(n, y) = Ty , (x, y) ∈ L,

(8.4-10)
Hxx cos(n, x)+Hxy cos(n, y) = hx , (x, y) ∈ L,
Hyx cos(n, x)+Hyy cos(n, y) = hy ,
(8.4-11)
where Tx = −p cos(n, x), Ty = −p cos(n, y) denote the components of surface trac-
tion, p is the magnitude of the pressure, hx and hy are generalized surface tractions,
n represents the outward unit normal vector of any point of the boundary. But the
measurement of generalized tractions has not been reported so far, for simplicity, we
assume that hx = 0, hy = 0.
From equations (8.4-3), (8.4-4) and (8.4-10), one has

i 1
g4 (z) + g3 (z) + zg4 (z) = (Tx + iTy )ds = − pz, z ∈ L. (8.4-12)
32c1 32c1
Taking conjugate on both sides of equation (8.4-12) yields
1
g4 (z) + g3 (z) + z̄g4 (z) = − pz̄, z ∈ L. (8.4-13)
32c1
From equations (8.4-3), (8.4-4) and (8.4-11), we have
R1 ImΘ(z) + R2 ReΘ(z) = 0,
z ∈ L. (8.4-14)
−R1 ReΘ(z) + R2 ImΘ(z) = 0,
Multiplying the second formula of (8.4-14) by i and adding it to the first, one obtains
Θ(z) = 0, z ∈ L. (8.4-15)
Because the function g1 (z) does not appear in the displacement and stress formu-
las, boundary equations (8.4-12), (8.4-13) and (8.4-15) are enough for determining
the unknown functions g2 (z), g3 (z) and g4 (z). However the calculation cannot be
completed at the z-plane due to the complicity of the evaluation, we must use the
conformal mapping

1
z = ω(ζ) = R0 + mζ (8.4-16)
ζ
to transform the region with ellipse at the z-plane onto the interior of the unit circle γ
a+b a−b
at the ζ-plane, refer to Fig.8.4-2, where ζ = ξ+iη = ρeiϕ and R0 = ,m= .
2 a+b
Fig. 8.4-2 Conformal mapping from the region at z-plane onto the interior of the unit
circle γ at ζ-plane
For simplicity, we introduce the following new symbols

(IV)
g2 (z) = F2 (z), g3 (z) = F3 (z), g4 (z) = F4 (z). (8.4-17)
And we have
Φj (ζ)
Fj (z) = Fj (ω(ζ)) = Φj (ζ), Fj (z) = , (j = 1, · · · , 4). (8.4-18)
ω (ζ)
Substituting (8.4-17) into (8.4-12), (8.4-13) and (8.4-15), then multiplying both sides
1 dσ
of equations by , and integrating along the unit circle, we have
2πi σ − ζ

1 Φ4 (σ)dσ 1 Φ3 (σ)dσ 1 ω(σ) Φ4 (σ)dσ
+ +
2πi γ σ−ζ 2πi γ σ − ζ 2πi γ ω(σ) σ − ζ

p 1 ω(σ)dσ
=− ,
32c1 2πi γ σ − ζ

1 Φ4 (σ)dσ 1 Φ3 (σ)dσ 1 ω(σ) Φ4 (σ)dσ
+ +
2πi γ σ − ζ 2πi γ σ − ζ 2πi γ ω(σ) σ − ζ

p 1 ω(σ)dσ
=− ,
32c1 2πi γ σ − ζ

1 Φ2 (σ)dσ 1 ω(σ) Φ3 (σ)dσ
+
2πi γ σ − ζ 2πi γ ω (σ) σ − ζ
2 2
1 ω(σ) Φ4 (σ)dσ ω(σ) ω (σ) Φ4 (σ)dσ
+ − = 0, (8.4-19)
2πi γ [ω (σ)]2 σ − ζ γ [ω (σ)]3 σ−ζ
where σ = eiϕ (ρ = 1) represents the value of ζ at the unit circle.

According to Cauchy’s integral formula and analytic extension of the complex

variable function theory, similar to the calculations at Sections 8.1 and 8.2, from the
first and the second equation of (8.4-19), one obtains
pR0 (1 + m2 )ζ
Φ3 (ζ) = ,
32c1 mζ 2 − 1
pR0
Φ4 (ζ) = − mζ. (8.4-20)
32c1
Substitution of
2
ω(σ) σ2 + m ω(σ) ω (σ) 2σ(σ 2 + m)2
= σ , = −
ω (σ) mσ 2 − 1 ω (σ)3 (mσ 2 − 1)3
and (8.4-20) into the third equation of (8.4-19) yields

1 Φ2 (σ)dσ 1 σ 2 + m Φ3 (σ)dσ 1 σ(σ 2 + m)2 Φ4 (σ)dσ
+ σ + = 0.
2πi γ σ−ζ 2πi γ mσ 2 − 1 σ − ζ 2πi γ (mσ 2 − 1)3 σ − ζ
By Cauchy’s integral formula, we have

1 Φ2 (σ)dσ
= Φ2 (ζ),
2πi γ σ−ζ

1 σ 2 + m Φ3 (σ)dσ ζ2 + m
σ = ζ Φ (ζ),
2πi γ mσ 2 − 1 σ − ζ mζ 2 − 1 3

1 σ(σ 2 + m)2 Φ4 (σ)dσ ζ(ζ 2 + m)2
= Φ (ζ).
2πi γ (mσ 2 − 1)3 σ − ζ (mζ 2 − 1)3 4
Substituting these equations into (8.4-19), one finds that at last
pR0 ζ(ζ 2 + m)[(1 + m2 )(1 + mζ 2 ) − (ζ 2 + m)]

Φ2 (ζ) = . (8.4-21)
32c1 (mζ 2 − 1)3
Utilizing the above mentioned results, the phonon and phason stresses can be de-
termined at the ζ-plane. We here only give a simple example, i.e., along the edge of
notch (ρ = 1), there are phonon stress components such as
1 − 3m2 + 2m cos 2ϕ
σϕϕ = p , σρρ = −p, σρϕ = σϕρ = 0,
1 + m2 − 2m cos 2ϕ
which are identical to the well-known results of the classical elasticity theory.
8.4.4 Elastic field caused by a Griffith crack

The solution of the Griffith crack subjected to a uniform pressure, has been observed
by Li[10] in terms of the Fourier transform method, which can also be obtained from
a
the notch solution in corresponding to the case m = 1, R0 = of the present work.
2
For explicitness, we express the solution at z-plane. The inversion of transformation
(8.4-16) is as m = 1,
1 #
ζ = (z − z 2 − a2 ). (8.4-22)
a
From equations (8.4-20), (8.4-25) and (8.4-22), we have
⎧
⎪ (IV) pa2 z2
⎪
⎪ g2 (z) = − # ,
⎪
⎪ 128c1 (z 2 − a2 )3
⎪
⎪
⎨
p a2
g3 (z) = − √ , (8.4-23)
⎪
⎪ 64c1 z 2 − a2
⎪
⎪
⎪
⎪ p √ 2
⎪
⎩ g4 (z) = ( z − a2 − z).
64c1
So the stresses and the displacements can be expressed with complex variable z.
Similar to (8.1-31), we introduce three pairs of polar coordinates (r, θ), (r1 , θ1 )
and (r2 , θ2 ) with the origin at crack center, at the right crack tip and at the left crack
tip, i.e., z = reiθ , z − a = r1 eiθ1 , z + a = r2 eiθ2 respectively, the analytic expressions
for the stress and displacement fields can be obtained.
Moreover, the stress intensity factor and free energy of the crack and so on can be
evaluated as the direct results of the solution. We here only list the stress intensity
factor and energy release rate as below

# √
KI = lim+ ( 2π(x − a)σyy (x, 0) = πap,
x→a

∞
1 ∂
GI = 2 [(σyy (x, 0) ⊕ Hyy (x, 0))(ux (x, 0) ⊕ wx (x, 0))]dx
2 ∂a 0
L(K1 + K2 ) + 2(R12 + R22 ) 2

= (KI ) , (8.4-24)
8(L + M )c
where c = M (K1 + K2 ) − 2(R12 + R22 ) and L = C12 , M = (C11 − C12 )/2 = C66 . The
crack energy release rate GI is dependent upon not only the phonon elastic constants
L(= C12 ), M (= (C11 − C12 )/2), but also the phason elastic constants K1 , K2 , and
the phonon-phason coupling elastic constants R1 , R2 , though we have assumed the
generalized tractions hx = hy = 0.
It is evident that the present solution covers the solution for point groups 5m,
10mm quasicrystals.
The detail for some further principle of the complex function method will be
discussed in-depth in Chapter 11. For the further implications and applications of
the results to fracture mechanics of quasicrystals, refer to Chapter 15.
8.5 Solutions of crack/notch problems of two-dimensional oc-

tagonal quasicrystals
Zhou and Fan[11] and Zhou[12] obtained the solution of a Griffith crack in octagonal
quasicrystals in terms of the Fourier transform and dual integral equations, the
calculation is very complex and lengthy, which cannot be listed here.
Li[13] gave solutions for a notch/Griffith crack problem in terms of complex vari-
able function method based on the stress potential formulation, an outline of the
algorithm was listed in Section 6.8.
The final governing equation of plane elasticity of octagonal quasicrystals of point
group 8mm based on the stress potential formulation is the same of that of (6.6-12),
but we here rewrite as
8
∂ ∂8 ∂8 ∂8 ∂8
+ 4(1 − 4ε) + 2(3 + 16ε) + 4(1 − 4ε) + G = 0,
∂x8 ∂x6 ∂y 2 ∂x4 ∂y4 ∂x2 ∂y6 ∂y 8
(8.5-1)
in which G(x, y) is the stress potential function, and the material constant ε is the
same as that given in Chapters 6 and 7. The complex representation of equation
(8.5-1) is
4
G(x, y) = 2Re Gk (zk ), zk = x + μk y, (8.5-2)
k=1
in which unknown functions Gk (zk ) are analytic functions of complex variable zk (k =

1, · · · , 4), to be determined, and μk = αk +iβk (k = 1, · · · , 4) are complex parameters
and determined by the roots of the following eigenvalue equation:
μ8 + 4(1 − 4ε)μ6 + 2(3 + 16ε)μ4 + 4(1 − 4ε)μ2 + 1 = 0. (8.5-3)
The stresses can be expressed by functions Gk (zk ) such as

4

σxx = −2c3 c4 Re (μ2k + 2μ4k + μ6k )gk (zk ), (8.5-4a)
k=1
4

σyy = −2c3 c4 Re (1 + 2μ2k + μ4k )gk (zk ), (8.5-4b)
k=1
4

σxy = σyx = 2c3 c4 Re (μk + 2μ3k + μ5k )gk (zk ), (8.5-4c)
k=1
8.6 Other solutions of crack problems in one- and two-dimensional quasicrystals 153
4

Hxx = RRe [(4c4 − c3 )μ2k + 2(3c3 − 2c4 )μ4k − c3 μ6k )]gk (zk ), (8.5-4d)
k=1
4

Hxy = −RRe [(4c4 − c3 )μk + 2(3c3 − 2c4 )μ3k − c3 μ5k )]gk (zk ), (8.5-4e)
k=1
4

Hyx = −RRe [c3 μk + 2(c4 − 2c3 )μ3k − (4c4 − c3 )μ5k )]gk (zk ), (8.5-4f)
k=1
4

Hyy = RRe [c3 + 2(c4 − 2c3 )μ2k − (4c4 − c3 )μ4k )]gk (zk ), (8.5-4g)
k=1
in which
∂ 6 Gk (zk ) dgk (zk )
gk (zk ) = , gk (zk ) = ,
∂zk6 dzk
(K1 + K2 + K3 )M − R2 K1 M − R 2
c3 = , c4 = .
K1 + K2 + 2K3 K1 − K 2
We now consider an elliptic hole L : x2 /a2 + y 2 /b2 = 1, at which there are the
boundary conditions
σxx cos(n, x)+σxy cos(n, y) = Tx , σxy cos(n, x)+σyy cos(n, y) = Ty , (x, y) ∈ L,
Hxx cos(n, x)+Hxy cos(n, y) = hx , Hyx cos(n, x)+Hyy cos(n, y) = hy , (x, y) ∈ L.
(8.5-5)
The complex variable zk can be rewritten as

zk = xk + iyk ,
(8.5-6)
xk = x + αk y, yk = βk y.
The second formula of equations (8.5-6) represents a coordinate transformation.
Further taking conformal mapping, the complex potentials satisfying the bound-
ary conditions can be determined. The detail is omitted here.
8.6 Other solutions of crack problems in one- and two-dimens-

ional quasicrystals
Fan and Guo[14] constructed the solution of three-dimensional elliptic crack in hexag-
onal quasicrystals by using perturbation procedure and hydrodynamic analogy, this
is the first solution of elliptic crack in quasicrystalline materials, and the solution
on penny-shaped crack of hexagonal quasicrystals given by Peng and Fan[15] is the
special case of the solution of Fan and Guo[14] .
Liu and Fan[16] and Liu[17] found the solution of Griffith crack of point group
10mm quasicrystals again by complex variable function-conformal mapping tech-
nique, Liu offered a series crack solutions and solutions of interaction between cracks
and dislocations for various one-dimensional quasicrystals.
The characteristics of crack solutions mentioned here and those given in the
previous sections provide a basis of fracture theory of quasicrystalline materials
summarized in Chpater 15.
8.7 Appendix of Chapter 8: Derivation of solution of Section

8.1
In Section 8.1 we focused on the physical aspect of the problem, and omitted some
mathematical details. Because the mathematical details are similar to that in Section
8.4, there is no need to repeat. But the strict derivation for complex displacement
potentials (8.1-12) is significant, which also provides a basis for the solutions in
Section 8.2, we outline the derivation as follows.
For simplicity we consider the boundary value problem (8.1-5). From (8.1-11)
we take ⎧
⎪ 1
⎪
⎨ σyz = σzy = − 2i [C44 (φ1 − φ1 ) + R3 (ψ1 − ψ1 )] + τ1
(8.7-1)
⎪
⎪ 1
⎩ Hzy = − [K2 (ψ1 − ψ1 ) + R3 (φ1 − φ1 )] + τ2

2i
In the following we use L to represent the crack surface under consideration, and
then the second expression of the boundary condition at L exhibited by (8.1-5) can
be written as

C44 (φ1 − φ1 ) + R3 (ψ1 − ψ1 ) − 2iτ1 = 0, t ∈ L
(8.7-2)
K2 (ψ1 − ψ1 ) + R3 (φ1 − φ1 ) − 2iτ2 = 0, t∈L
By using the following conformal mapping,

a 1
z = ω(ζ) = ζ+ (8.7-3)
2 ζ
which maps the domain in the t-plane with a Griffith crack onto the interior of the
unit circle γ in the ζ-plane (ζ = ξ + iη = ρeiϕ ), and the elliptic hole L onto a unit
circle γ (similar to Fig. 8.4-2).
On the unit circle γ, ζ = σ ≡ eiϕ , where ρ = 1. Furthermore, with the conformal
mapping (8.7-3), the unknown functions φ1 (t) and ψ1 (t) and their corresponding
derivatives can be expressed as

φ1 (t) = φ1 [ω(ζ)] = φ(ζ), ψ1 (t) = ψ1 [ω(ζ)] = ψ(ζ)
(8.7-4)
φ1 (t) = φ (ζ)/ω (ζ), ψ1 (t) = ψ (ζ)/ω (ζ)
Similar to the manipulation undertaken in Section 8.4 the boundary conditions

8.7 Appendix of Chapter 8: Derivation of solution of Section 8.1 155
(8.7-2) are reduced to the following function equations

⎧
⎪ 1 φ (σ) 1 ω (σ) dσ R3 1 ψ (σ)
⎪
⎪ dσ − φ (σ) + dσ
⎪
⎪ 2πi σ − ζ 2πi σ − ζ C 2πi γ σ−ζ
⎪
⎪ γ γ ω (σ) 44
⎪
⎪
⎪
⎪ R3 1 ω (σ) dσ 2iτ1 1 ω (σ)
⎪
⎪ − ψ (σ) = dσ
⎪
⎨ C44 2πi γ ω (σ) σ−ζ C44 2πi γ σ − ζ
⎪ (8.7-5)
⎪
⎪ 1 ψ (σ) 1 ω (σ) dσ R3 1 φ (σ)
⎪
⎪ dσ − ψ (σ) + dσ
⎪
⎪ 2πi σ − ζ 2πi ω (σ) σ − ζ K 2 2πi σ −ζ
⎪
⎪
γ γ γ
⎪
⎪
⎪
⎪ R3 1 ω (σ) dσ 2iτ2 1 ω (σ)
⎪
⎩ − φ (σ) = dσ
K2 2πi γ ω (σ) σ−ζ K2 2πi γ σ − ζ
With the similar procedure that adopted in Section 8.4, we can find
⎧
⎪
⎪ 1 φ (σ) 1 ω (σ) dσ
⎪
⎪ dσ = φ (ζ), φ (σ) =0
⎪
⎪ 2πi γ σ − ζ
2πi γ ω (σ) σ−ζ
⎪
⎪
⎨ 1 ψ (σ) 1 ω (σ) dσ
dσ = ψ (ζ), − ψ (σ) =0 (a)
⎪
⎪ 2πi γ σ − ζ
2πi γ ω (σ) σ−ζ
⎪
⎪
⎪
⎪ 1 ω (σ) a
⎪
⎪
⎩ 2πi σ−ζ
dσ =
2
γ
so the solution of function equations (8.7-5) is

K 2 τ 1 − R3 τ 2 C44 τ2 − R3 τ1
φ (ζ) = ia , ψ (ζ) = ia (8.7-6)
C44 K2 − R32 C44 K2 − R32
By integration from the solution (8.7-6) we have

K2 τ1 − R3 τ2 C44 τ2 − R3 τ1
φ(ζ) = ia ζ, ψ(ζ) = ia ζ (8.7-7)
C44 K2 − R32 C44 K2 − R32
The single-valued inversion of conformal mapping
&
2
t t
ζ = ω −1 (t) = − −1 (8.7-8)
a a
makes |z| = ∞ correspond to ζ = 0, and substituting it into (8.7-8) yields

K 2 τ 1 − R3 τ 2
φ(ζ) = φ(ω −1 (t)) = φ1 (t) = ia ζ
C44 K2 − R32
⎛ & ⎞
2
K2 τ1 − R3 τ2 ⎝ t t
= ia − − 1⎠
C44 K2 − R32 a a
C44 τ2 − R3 τ1
ψ(ζ) = ψ(ω −1 (t)) = ψ1 (t) = ia ζ
C44 K2 − R32
156 Chapter 8 Application II: Solutions of notch and crack problems and solutions...
⎛ & ⎞
2
C44 τ2 − R3 τ1 ⎝ t t
= ia − − 1⎠
C44 K2 − R32 a a
these are complex displacement potentials (8.1-12).
References
[1] Hu C Z, Yang W Z, Wang R H et al. Symmetry and physical properties of quasicrys-
tals. Adv Phys, 1997, 17(4): 345–376 (in Chinese)
[2] Meng X M, Dong B Y, Wu Y K. Mechanical property of quasi-crystal Al65 Cu20 Co15 .
Acta Metal Sinica, 1994, 30(2): 61–64 (in Chinese)
[3] Kanninen M F, Popelar C H, Advanced Fracture Mechanics. Cambridge: Cambridge
University Press, 1985; Fan T Y. Foundation of Fracture Theory. Beijing: Science
Press, 2003 (in Chinese)
[4] Fan T Y. Exact analytic solutions of stationary and fast propagating cracks in a
strip. Science in China, A, 1991, 34(5): 560–569
[5] Fan T Y. Mathematical Theory of Elasticity of Quasicrystals and Its Application.
Beijing Institute of Technology Press. Beijing, 1999 (in Chinese)
[6] Li L H, Fan T Y. Exact solutions of two semi-infinite collinear cracks in a strip of
one-dimensional hexagonal quasicrystal. Applied Mathematics and Computation,
2008, 196(1): 1–5
[7] Shen D W, Fan T Y. Exact solutions of two semi-infinite collinear cracks in a strip.
Eng Fracture Mech, 2003, 70(8): 813–822
A, 1999, 79(8): 1943–1952
[9] Li L H, Fan T Y. Complex function method for solving notch problem of point group
10, 10 two-dimensional quasicrystals based on the stress potential function. J Phys:
Condens Matter, 2007, 18(47): 10631–10641
[10] Li X F. Defect problems and their analytic solutions of the theory of elasticity of
quasicrystals. Dissertation. Beijing Institute of Technology, 1999 (in Chinese)
[11] Zhou W M, Fan T Y. Plane elasticity problem of two-dimensional octagonal qua-
sicrystal and crack problem. Chin Phys, 2001, 10(8): 743–747
[12] Zhou W M. Mathematical analysis of elasticity and defects of two- and three-dimens-
ional quasicrystals . Dissertation. Beijing: Beijing Institute of Technology, 2000 (in
Chinese)
[13] Li L H. Study on complex variable function method and analytic solutions of elasticity
of quasicrystals. Dissertation. Beijing: Beijing Institute of Technology, 2008 (in
Chinese)
References 157
[14] Fan T Y, Guo R P. Three-dimensional elliptic crack in one-dimensional hexagonal

quasicrystals. Appl Math Mech, submitted, 2009
[15] Peng Y Z, Fan T Y. Elastic theory of 1D quasiperiodic stacking of 2D crystals. J
Phys: Condens Matter, 2000, 12(45): 9381–9387
[16] Liu G T, Fan T Y. The complex method of the plane elasticity in 2D quasicrystals
point group 10mm ten-fold rotation symmetry notch problems. Science in China E,
2003, 46(3): 326–336
tals and defects and auxiliary equation method for solving some nonlinear evolution
equations. Dissertation. Beijing Institute of Technology, 2004 (in Chinese)
[18] Muskhelishrili N I, Some Basic Problems of the Mathematical Theory of Elasticity,
Groningen: Noordhoff ltd, 1953
Chapter 9
Theory of elasticity of three-dimensional
quasicrystals and its applications
In Chapters 5∼8 we discussed the theories of elasticity of one- and two-dimensional

quasicrystals and their applications. In this chapter the theory and applications of
elasticity of three-dimensional quasicrystals will be dealt with. The three-dimensional
quasicrystals include icosahedral quasicrystals and cubic quasicrystals. In all about
200 individual quasicrystals observed to date there are almost 100 icosahedral qua-
sicrystals, so that they play the central role in this kind of solids. This suggests the
major importance of elasticity of icosahedral quasicrystals in the study of mechanical
behaviour of quasicrystalline material.
There are some polyhedrons with the icosa-
hedral symmetry, one among them is shown in
Fig. 9.0-1, which consists of 20 right triangles
and contains 12 five-fold symmetric axes A5, 20
three-fold symmetric axes A3 and 30 two-fold
symmetric axes A2. One of the diffraction pat-
tern is shown in Fig. 3.1-1 and the stereographic
structure of one of icosahedral point groups is
also depicted in Fig. 3.1-1.
The elasticity of icosahedral quasicrystals
was studied immediately after the discovery of
the structure, which is the pioneering work of
the field. The outlook about this was figured out
in the Chapter 4, in which the contribution of pi-
oneers such as P. Bak etc was introduced. After-
ward Ding et al[1] set up the physical framework
of elasticity of icosahedral quasicrystals, they[2] Fig. 9.0-1 Icosahedral quasicrystal
also summarized the basic relationship of elasti-
city of cubic quasicrystals. In terms of the Green function method, Yang et al[3] gave
an approximate solution on dislocation for a special case, i.e., the phonon-phason
decoupled plane elasticity of icosahedarl quasicrystal. In this chapter we mainly
discuss the general theory of elasticity of icosahedral quasicrystals and the applica-
tions, in addition, those for cubic quasicrystals are also concerned. We focus on the

160 Chapter 9 Theory of elasticity of three-dimensional quasicrystals and its...
mathematical theory of the elasticity and the analytic solutions. Because of the large
number of field variables and field equations involving elasticity of these two kinds
of three-dimensional quasicrystals, the solution presents tremendous difficulty. We
continue to develop the decomposition procedure adopted in the previous chapters,
this can reduce the number of the field variables and field equations, and three-
dimensional elasticity can be simplified to two-dimensional elasticity to solve for
some cases with important practical applications. The introducing of displacement
potentials or stress potentials[4,5] can further simplify the problems. In the work
some systematic and direct methods of mathematical physics and function theory
have been developed, and a series of analytic solutions are constructed, which will
be included in the chapter. Because the calculations are very complex, we would like
to introduce them in detail as much as possible in order to facilitate comprehension
of the text.
9.1 Basic equations of elasticity of icosahedral quasicrystals

The equations of deformation geometry are

1 ∂ui ∂uj ∂wi
εij = + , wij = , (9.1-1)
2 ∂xj ∂xi ∂xj
which are similar in form to those given in previous chapters, but here ui and wi
have 6 components, and εij and wij have 15 components in total.
The equilibrium equations are as follows:
∂σij ∂Hij
= 0, = 0, (9.1-2)
∂xj ∂xj
which are also similar in form to those listed in previous chapters, however here
adding σij and Hij gives 15 stress components.
Between the stresses and strains there is the generalized Hooke’s law such as
σij = Cijkl εkl + Rijkl wkl , Hij = Rklij εkl + Kijkl wkl , (9.1-3)
in which the phonon elastic constants are described by
Cijkl = λδij δkl + μ(δik δjl + δil δjk ), (9.1-4)
where λ and μ (= G in some references) the Lamé constants.

If the strain components are arranged as a vector according to the order
[εij , wij ] = [ε11 ε22 ε33 ε23 ε31 ε12 w11 w22 w33 w23 w32 w12 w32 w13 w21 ] (9.1-5 )
9.1 Basic equations of elasticity of icosahedral quasicrystals 161
and the stress components are also arranged according to the same order, i.e.,
[σij , Hij ] = [σ11 σ22 σ33 σ23 σ31 σ12 H11 H22 H33 H23 H12 H32 H13 H21 ] (9.1-5 )
then phason and phonon-phason coupling elastic constants can be expressed by

matrices of [K] and [R],
⎡ ⎤
K1 0 0 0 K2 0 0 K2 0
⎢ 0 K ⎥
⎢ 1 0 0 −K2 0 0 K2 0 ⎥
⎢ ⎥
⎢ 0
⎢ 0 K2 + K1 0 0 0 0 0 0 ⎥⎥
⎢ ⎥
⎢ 0 0 0 K1 − K 2 0 K2 0 0 −K2 ⎥
⎢ ⎥
[K] = ⎢
⎢ K2 −K2 0 0 K 1 − K2 0 0 0 0 ⎥⎥,
⎢ ⎥
⎢ 0 0 0 K 2 0 K1 −K 2 0 0 ⎥
⎢ ⎥
⎢ 0 0 0 0 0 −K K − K 0 −K ⎥
2⎥
⎢ 2 1 2
⎢ ⎥
⎣ K2 K2 0 0 0 0 0 K 1 − K2 0 ⎦
0 0 0 −K2 0 0 −K2 0 K1
⎡ ⎤
1 1 1 0 0 0 0 1 0
⎢
⎢ −1 −1 1 0 0 0 0 −1 0 ⎥
⎥
⎢ ⎥
⎢ 0 0 −2 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢ 0 0 0 0 0 −1 1 0 −1 ⎥
⎢ ⎥
[R] = R ⎢
⎢ 1 −1 0 0 1 0 0 0 0 ⎥
⎥. (9.1-6)
⎢ ⎥
⎢ 0 0 0 −1 0 −1 0 0 1 ⎥
⎢ ⎥
⎢ 0 0 0 0 0 −1 1 0 −1 ⎥
⎢ ⎥
⎢ ⎥
⎣ 1 −1 0 0 1 0 0 0 0 ⎦
0 0 0 −1 0 −1 0 0 1
Equations (9.1-1)∼(9.1-3) are basic equations of elasticity of icosahedral quasicrys-

tals, there are 36 equations in total, and the number of the field variables is also 36.
It is consistent and solvable mathematically.
Due to the huge number of the field variables and field equations, the mathe-
matical solution is highly complex. One way to solve the elasticity problem is to
reduce the number of the field variables and field equations above mentioned. For
this purpose we can utilize the eliminating element method in classical mathematical
physics.
Based on the matrix expression of the generalized Hooke’s law of (4.5-3) in the
Chapter 4, then one has the explicit relationship between stresses and strains as
below:
⎧
⎪
⎪ σxx = λθ + 2μεxx + R(wxx + wyy + wzz + wxz ),
⎪
⎪
⎪
⎪ σyy = λθ + 2μεyy − R(wxx + wyy − wzz + wxz ),
⎪
⎪
⎪
⎪ σzz = λθ + 2μεyy − 2Rwzz ,
⎪
⎪
⎪
⎪ σyz = 2μεyz + R(wzy − wxy − wyx ) = σzy ,
⎪
⎪
⎪
⎪ σzx = 2μεzx + R(wxx − wyy − wzx ) = σxz ,
⎪
⎪
⎪
⎪ σxy = 2μεxy + R(wyx − wyz − wxy ) = σyx ,
⎪
⎪
⎨ Hxx = R(εxx − εyy + 2εzx ) + K1 wxx + K2 (wzx + wxz ),
Hyy = R(εxx − εyy − 2εzx ) + K1 wyy + K2 (wxz − wzx ), (9.1-7)
⎪
⎪
⎪
⎪ Hzz = R(εxx + εyy − 2εzz ) + (K1 + K2 )wzz ,
⎪
⎪
⎪
⎪ Hyz = −2Rεxy + (K1 − K2 )wyz + K2 (wxy − wyx ),
⎪
⎪
⎪
⎪ Hzx = 2Rεzx + (K1 − K2 )wzx + K2 (wxx − wyy ),
⎪
⎪
⎪
⎪ Hxy = −2R(εyz + εxy ) + K1 wxy + K2 (wyz − wzy ),
⎪
⎪
⎪
⎪ Hzy = 2Rεyz + (K1 − K2 )wzy − K2 (wxy + wyx ),
⎪
⎪
⎪
⎪ Hxz = R(εxx − εyy ) + K2 (wxx + wyy ) + (K1 − K2 )wxz ,
⎩
Hyx = 2R(εxy − εyx ) + K1 wyx − K2 (wyz + wzy ),
where θ = εxx + εyy + εzz denotes the volume strain and εij and wij are defined by
(9.1-1). This explicit expression was first given by Ding et al[1] .
Substituting (9.1-7) into (9.1-2) yields one of forms of the final governing equations—
the
⎧ equilibrium equations in terms of displacements as follows:
⎪
⎪ ∂
⎪
⎪ μ∇2 ux + (λ + μ) ∇ · u
⎪
⎪ 2 ∂x 2
⎪
⎪
⎪
⎪ ∂ wx ∂ wx ∂ 2 wx ∂ 2 wy ∂ 2 wy ∂ 2 wz
⎪
⎪ + R + − + 2 − 2 + 2 = 0,
⎪
⎪ ∂x2 ∂x∂z ∂y 2 ∂x∂y ∂y∂z ∂x∂z
⎪
⎪
⎪
⎪ ∂
⎪
⎪ μ∇2 uy + (λ + μ) ∇ · u
⎪
⎪ ∂y
⎪
⎪ 2
⎪
⎪ ∂ wy ∂ 2 wx ∂ 2 wy ∂ 2 wx ∂ 2 wy ∂ 2 wz
⎪
⎪ +R − − −2 −2 +2 = 0,
⎪
⎪ ∂x2 ∂x∂y ∂y 2 ∂z∂y ∂x∂z ∂y∂z
⎪
⎪
⎪
⎪ ∂
⎪
⎪ μ∇2 uz + (λ + μ) ∇ · u
⎪
⎪
⎪
⎪ 2 ∂z
⎪
⎪ ∂ 2 wx ∂ 2 wx ∂ 2 wz ∂ 2 wz ∂ 2 wz
⎪
⎪
∂ wx
− − −
⎪
⎨ + R 2
2 + + 2 = 0,
∂x ∂x∂y ∂y 2 ∂x2 ∂y 2 ∂z2
⎪ ∂ 2 wx ∂ 2 wx ∂ 2 wy ∂ 2 wz ∂ 2 wz (9.1-8)
⎪
⎪ K1 ∇2 wx + K2 2 − +2 + −
⎪
⎪ ∂x∂z ∂z 2 ∂y∂z ∂x 2 ∂y 2
⎪
⎪ 2
⎪
⎪ ∂ u ∂ 2
u ∂ 2
u ∂ 2
u ∂ 2
u ∂ 2
uz ∂ 2 uz
⎪
⎪ +R
x
+2
x
−
x
−2
y
−2
y
+ − = 0,
⎪
⎪ ∂x2 ∂x∂z ∂y 2 ∂x∂y ∂y∂z ∂x2 ∂y 2
⎪
⎪
⎪
⎪ ∂ 2 wx ∂ 2 wy ∂ 2 wz ∂ 2 wy
⎪
⎪ K1 ∇2 wy + K2 2 −2 −2 −
⎪
⎪
⎪
⎪ ∂y∂z ∂x∂z ∂x∂y ∂z 2
⎪
⎪ 2 2 2 2

⎪
⎪ ∂ uy ∂ ux ∂ uy ∂ ux ∂ 2 uy ∂ 2 uz
⎪
⎪ +R −2 + +2 −2 −2 = 0,
⎪
⎪ ∂x2 ∂y∂z ∂y 2 ∂x∂y ∂x∂z ∂x∂y
⎪
⎪
⎪
⎪ ∂ 2 wz ∂ 2 wy ∂ 2 wx ∂ 2 wx
⎪
⎪ (K1 − K2 )∇2 wz + K2 2 −2 + −
⎪
⎪ 2 ∂x2 ∂y 2
⎪
⎪
∂z ∂y∂x

⎪
⎪ 2 2 2 2
⎪
⎪ ∂ uz ∂ ux ∂ uz ∂ uz
⎩ +R +2 + − = 0,
∂x2 ∂x∂z ∂y 2 ∂z 2
9.1 Basic equations of elasticity of icosahedral quasicrystals 163
where
∂2 ∂2 ∂2 ∂ux ∂uy ∂uz
∇2 = + + , ∇·u= + + .
∂x2 ∂y2 ∂z 2 ∂x ∂y ∂z
Equations (9.1-8) are 6 partial differential equations of second order on displace-
ments ui and wi . So the number of the field variables and field equations is reduced
already. But obtaining solution is still very difficult, one of reasons is the boundary
conditions for quasicrystals being much more complicated than those of the classical
theory of elasticity. In the subsequent sections we will make a great effort to solve
some complex boundary value problems through different approaches.
It is obvious that the material constants of λ, μ, K1 , K2 and R are very important
for stress analysis for different icosahedral quasicrystals, which are experimentally
measured through various methods (e.g. X-ray diffraction, neutron scattering etc)
and listed by Tables 9.1-1∼9.1-3 respectively as follows:
Table 9.1-1 Phonon elastic constants of various icosahedral quasicrystals

Alloys λ μ(G) B ν Refs.
Al-Li-Cu 30 35 53 0.23 [6]
Al-Li-Cu 30.4 40.9 57.7 0.213 [7]
Al-Cu-Fe 59.1 68.1 104 0.213 [8]
Al-Cu-Fe-Ru 48.4 57.9 87.0 0.228 [8]
Al-Pd-Mn 74.9 72.4 123 0.254 [8]
Al-Pd-Mn 74.2 70.4 121 0.256 [9]
Ti-Zr-Ni 85.5 38.3 111 0.345 [10]
Cu-Yh 35.28 25.28 52.13 0.2913 [11]
Zn-Mg-Y 33.0 46.5 64.0 0.208 [12]
In this table the measurement unit of λ, μ and B is GPa, and B = (3λ + 2μ)/3
represents the bulk modulus, and ν = λ/2(λ + μ) the Poisson’s ratio, respectively.
Table 9.1-2 Phason elastic constants of various icosahedral quasicrystals

Alloys Source Meas. Temp. K1 /MPa K2 /MPa Refs.
Al-Pd-Mn X-ray R.T. 43 −22 [13]
Al-Pd-Mn Neutron R.T. 72 −37 [13]
Al-Pd-Mn Neutron 1043K 125 −50 [13]
Zn-Mg-Sc X-ray R.T. 300 −45 [14]
Table 9.1-3 Phonon-phason coupling elastic constant of various icosahedral

quasicrystals
Alloys Source R Refs.
Mg-Ga-Al-Zn X-ray −0.04μ [15]
Al-Cu-Fe X-ray 0.004μ [15]
It is needed to point out that equations (9.1-8) are not the only form of final
governing equation of elasticity of icosahedral quasicrystals, there are other forms
which will be discussed in Section 9.5.
9.2 Anti-plane elasticity of icosahedral quasicrystals and prob-

lem of interface between quasicrystal and crystal
People can find equations (9.1-8) very complex, but they can be simplified for some
meaningful cases physically. One of them is the so-called anti-plane case where the
non-zero displacements are only uz and wz , and the other displacements vanish. In
particular these two displacements and relevant strains and stresses are independent
of the coordinate x3 (or z). If there is a Griffith crack along the axis z (see Fig.
9.2-1) or a straight dislocation line along the direction, etc., in addition, the applied
external fields are independent of variable z, so

∂ ∂
= = 0. (9.2-1)
∂x3 ∂z
Fig. 9.2-1 One of configuration of plane or anti-plane elasticity
Because there are only two components uz and wz , and others have vanished, the
corresponding strains are only
1 ∂uz 1 ∂uz ∂wz ∂wz
εyz = εzy = , εxz = εzx = , wzy = , wzx = . (9.2-2)
2 ∂y 2 ∂x ∂y ∂x
From the formulas listed in Section 9.1, the non-zero stress components are
⎧
⎪
⎪ σxz = σzx = 2μεxz + Rwzx ,
⎪
⎪
⎪
⎪ σ yz = σzy = 2μεyz + Rwzy ,
⎪
⎪
⎪ zx = (K1 − K2 )wzx + 2Rεxz ,
⎪ H
⎨
Hzy = (K1 − K2 )wzy + 2Rεyz ,
(9.2-3)
⎪
⎪ H xx = 2Rεxz + K2 wzx ,
⎪
⎪
⎪
⎪ Hyy = −2Rεxz − K2 wzx ,
⎪
⎪
⎪
⎪ Hxy = −2Rεyz − K2 wzy ,
⎩
Hyx = −2Rεyz − K2 wzy
9.2 Anti-plane elasticity of icosahedral quasicrystals and problem of interface... 165
and the equilibrium equations stand for

⎧
⎪ ∂σzx ∂σzy ∂Hzx ∂Hzy
⎪
⎨ + = 0, + = 0,
∂x ∂y ∂x ∂y
(9.2-4)
⎪
⎩ + = 0, + = 0.
∂x ∂y ∂x ∂y
Problem described by equations (9.2-2)∼(9.2-4) is anti-plane elasticity problem, and
we have the final governing equations
∇21 uz = 0, ∇21 wz = 0, (9.2-5)
∂2 ∂2
where ∇21 = + .
∂x2 ∂y 2
One can see that equations (9.2-5) are similar to (5.2-11), which can be solved
using a procedure similar to that adopted in Chapters 5, 7 and 8.
As an example of solution of anti-plane elasticity of icosahedral quasicrystals we
discuss the interface problem between centre-body cubic crystals and icosahedral
quasicrystals.
The physical model is similar to that proposed in Section 7.6, i.e., the icosahedral
quasicrystal is located in upper half-space y > 0, whose governing equations are
listed above, while the centre-body cubic crystal lies in lower space y < 0 with finite
thickness h (refer to Fig. 7.6-1) and governed by the following equation:
∇2 u(c)
z = 0, (9.2-6)
with the following stress-strain relations

(c) (c)
σzy = σyz = 2μ(c) ε(c)
yz ,
(c)
σzx (c)
= σxz = 2μ(c) ε(c)
xz
(c) (c) (c) (c)

for crystals, in which εij = (∂ui /∂xj + ∂uj /∂xi )/2, μ(c) = C44 .
After the Fourier transform, the solution of (9.2-5) is very easy to obtain such as
∞ ∞
1 1
uz (x, y) = A(ξ)e−|ξ|y−iξx dξ, wz (x, y) = B(ξ)e−|ξ|y−iξx dξ
2π −∞ 2π −∞
(9.2-7)
and the relevant stresses, e.g.
∞ ∞
1 −|ξ|y−iξx 1
σzy = −2μ |ξ|A(ξ)e dξ − R |ξ|B(ξ)e−|ξ|y−iξx dξ,
2π −∞ 2π −∞
∞ ∞
1 −|ξ|y−iξx 1
Hzy = −R |ξ|A(ξ)e dξ − (K1 − K2 ) |ξ|B(ξ)e−|ξ|y−iξx dξ,
2π −∞ 2π −∞
(9.2-8)
in which y > 0 and A(ξ) and B(ξ) are arbitrary functions to be determined.
According to the second condition of the boundary conditions at the interface
y = 0, −∞ < x < ∞ : σzy = τ f (x) + ku(x), Hzy = 0, (9.2-9)
where
μ(c)
k= , (9.2-10)
h
then the relation between the two unknown functions is obtained
R
B(ξ) = − A(ξ). (9.2-11)
K1 − K 2
From (9.2-7) and (9.2-8), we have

R2
σ̂zy = − 2μ + A(ξ) |ξ|
K1 − K2
and from the first one of conditions (9.2-9), one determines the unknown function
τ fˆ(ξ)
A(ξ) = − 2
(9.2-12)
R
+ μ |ξ| + k
K1 − K 2
so that
Rτ fˆ(ξ)
B(ξ) = , (9.2-13)
R2
(K1 − K2 ) + μ |ξ| + k
K1 − K 2
in which τ and k is defined by (9.2-9) and (9.2-10) respectively. Thus the problem
is solved. The phason strain field can be determined as
⎧ ∞
⎪
⎪ 1 |ξ| fˆ(ξ)
⎪ zy
⎨ w (x, y) = −Rτ e−|ξ|y−iξx dξ,
2π −∞ (R − μ(K1 − K2 )) |ξ| + k(K1 − K2 )
2
⎪ ∞
⎪
⎪ 1 ξ fˆ(ξ)
⎩ wzx (x, y) = iRτ 2
e−|ξ|y−iξx dξ.
2π −∞ (R − μ(K1 − K2 )) |ξ| + k(K1 − K2 )
(9.2-14)
Note that y > 0. The integrals in (9.2-14) can be evaluated by the residue
theorem introduced in the Major Appendix of this book.
For illustration, consider the first example. Let that f (x) = 1, as −a/2 < x <
2 a
a/2, and f (x) = 0, as x < −a/2 and x > a/2, so fˆ(ξ) = sin ξ , then from
ξ 2
solution (9.2-14), we obtain
9.2 Anti-plane elasticity of icosahedral quasicrystals and problem of interface... 167
⎧
⎪
⎪ a μ(c) R(K1 − K2 )τ 1 μ(c) (K1 − K2 ) a
⎪
⎪ w zy (x, y) = sin
⎪
⎪ h [μ(K1 − K2 ) − R2 ]2 2 μ(K1 − K2 ) − R2 h
⎪
⎪ (c)
⎪
⎪ μ(c) (K1 − K2 ) y μ (K1 − K2 ) x
⎪
⎪
⎪
⎨ × exp − cos
μ(K1 − K2 ) − R2 h μ(K1 − K2 ) − R2 h
⎪
⎪
⎪ a μ(c) R(K1 − K2 )τ 1 μ(c) (K1 − K2 ) a
⎪
⎪ w (x, y) = sin
⎪
⎪
zx
h [μ(K1 − K2 ) − R2 ]2 2 μ(K1 − K2 ) − R2 h
⎪
⎪ (c)
⎪
⎪
⎪
⎪ μ(c) (K1 − K2 ) y μ (K1 − K2 ) x
⎩ × exp − sin
μ(K1 − K2 ) − R2 h μ(K1 − K2 ) − R2 h
(9.2-15)
(c)
in which k = μ /h and the normalized expression has been used, i.e., x/h, y/h.
Then consider the second example f (x) = δ(x), then the integrals (9.2-14) will be
⎧
⎪
⎪ μ(c) R(K1 − K2 )τ
⎪
⎪ w (x, y) =
⎪
⎪
zy
[μ(K1 − K2 ) − R2 ]2
⎪
⎪
⎪
⎪ (c)
⎪
⎪ μ(c) (K1 − K2 ) y μ (K1 − K2 ) x
⎪
⎪ × −
⎪
⎨ exp sin
μ(K1 − K2 ) − R2 h μ(K1 − K2 ) − R2 h
(9.2-16)
⎪
⎪ (c)
R(K1 − K2 )τ
⎪
⎪ μ
⎪
⎪ wzx (x, y) =
⎪
⎪ [μ(K1 − K2 ) − R2 ]2
⎪
⎪
⎪
⎪ (c)
⎪
⎪ μ(c) (K1 − K2 ) y μ (K1 − K2 ) x
⎪
⎩ × exp − cos
μ(K1 − K2 ) − R2 h μ(K1 − K2 ) − R2 h
The detail of the evaluation is given in the Major Appendix of the book.
The results are quite interesting. In the first example, the phason strain field is
dominated by the elastic constants μ, K1 , K2 , R and μ(c) of quasicrystal and crystal,
applied stress τ and geometry parameters a and h, while in the second example
the geometry parameter is only h. For different τ /μ, μ(c) /μ,a/h and given values of
μ, K1 , K2 and R, one can find a rich set of numerical results. The computation used
the measured values of these quantities for Al-Pd-Mn icosahedral quasicrystals are
provided by Tables 9.1-1, 9.1-2 and 9.1-3:
μ = 72.4GPa, K1 = 125MPa, K2 = −50MPa, R = 0.04μ
The numerical results show that the influence of the ratio μ(c) /μ of shear modulus
of crystal and quasicrystal is very evident. In addition the influence of the applied
stress τ /μ is also very important. While the influence of a/h is not evident for the
first example.
Another feature of solution here is quite different from that in Section 7.6 due
to the difference of quasicrytalline systems.
This work is given in the Ref. [24].
9.3 Phonon-phason decoupled plane elasticity of icosahedral

quasicrystals
Yang et al[3] presented an approximate solution of a straight dislocation in icosahe-
dral quasicrystals under the assumptions
∂
=0 (9.3-1)
∂z
and
R=0 (9.3-2)
The conditions (9.3-1) and (9.3-2) result in a phonon-phason decoupled plane elas-
ticity, in which there are
εzz = 0, wzz = wyz = wxz = 0. (9.3-3)
Based on conditions (9.3-1) and (9.3-2), the final governing equations (9.1-8)
reduce to
⎧
⎪ ∂
⎪
⎪ μ∇21 ux + (λ + μ) ∇1 · u1 = 0,
⎪
⎪ ∂x
⎪
⎪
⎪
⎪
⎪
⎪ ∂
⎪
⎪ μ∇21 uy + (λ + μ) ∇1 · u1 = 0,
⎪
⎪ ∂y
⎪
⎪
⎪
⎪ μ∇ 2
u = 0,
⎨ 1 z 2
2 ∂ wz ∂ 2 wz (9.3-4)
⎪ K1 ∇1 wx + K2 ∂x2 − ∂y 2 = 0,
⎪
⎪
⎪
⎪
⎪
⎪
⎪ ∂ 2 wz
⎪
⎪ K1 ∇21 wy − 2K2 = 0,
⎪
⎪
⎪
⎪
∂x∂y
⎪
⎪ 2
⎪
⎪ ∂ wx ∂ 2 wy ∂ 2 wy
⎪ 2
⎩ (K1 − K2 )∇1 wz + K2 −2 − = 0,
∂x2 ∂x∂y ∂y 2
where
∂2 ∂2 ∂ux ∂uy
∇21 = 2
+ 2, u1 = (ux , uy ), ∇1 · u1 = + .
∂x ∂y ∂x ∂y
Because the phonons and phasons are decoupled, the first three equations of (9.3-
4) are pure phonon equilibrium equations, in addition uz is independent of ux and
uy , and the second three equations in (9.3-4) are pure phason equilibrium equations.
Yang et al[3] solved the equations under the dislocation conditions

dui = bi , dwi = b⊥i , (9.3-5)
Γ Γ
9.4 Phonon-phason coupled plane elasticity of icosahedral quasicrystals... 169
where Γ represents a path enclosing the dislocation core. The authors used the
Green function method to calculate. The results are
⎧
⎪
⎪ b1 y λ + μ xy b2 μ r λ + μ x2
⎪
⎪ ux = arctan + + ln + ,
⎪
⎪ 2π x λ + 2μ r2 2π λ + 2μ r0 λ + 2μ r2
⎪
⎪
⎪
⎪
⎪
⎪ b μ r λ + μ y2 b2 y λ + μ xy
⎪ uy = − 1
⎪ ln + + arctan − ,
⎪
⎪ 2π λ + 2μ r0 λ + 2μ r 2 2π x λ + 2μ r2
⎪
⎪
⎪
⎪
⎪ uz = b3 arctan y ,
⎪
⎪
⎪
⎪
⎨
2π x
b⊥
1 y K 2 2xy3 xy
wx = arctan + 2 −
⎪
⎪ 2π x 2K5 r4 r2
⎪
⎪
⎪
⎪ b⊥ K 2 r 2x2 y 2 b⊥ K2 xy
⎪
⎪ − 2 2 ln + 3
⎪
⎪
+ 4 2π K1 r2
,
⎪
⎪ 4π K5 r0 r
⎪
⎪ ⊥ ⊥
⎪
⎪ b⊥ K 2 r 2x2 y 2 b2 y K22 xy 3 − x3 y b3 K2 y 2
⎪ wy = 1 2 ln −
⎪ + arctan + − ,
⎪
⎪ 4π K5 r0 r4 2π x 2K5 r4 2π K1 r2
⎪
⎪
⎪
⎪ b⊥ K xy b⊥ 2
b⊥
⎪
⎩ wz = 1 K1 2 − 2 K1 K2 y
+ 3
arctan ,
y
2π K5 r2 2π K5 r2 2π x
(9.3-6)
in which
K5 = K12 − K1 K2 − K22 . (9.3-7)
The first three of (9.3-6) are well known solution of pure phonon field in the
classical theory of dislocation, and the second three of (9.3-6) are new results for
pure phason field. Because Yang et al[3] ignored the coupling terms, the interaction
between phonons and phasons could not be revealed.
9.4 Phonon-phason coupled plane elasticity of icosahedral qu-

asicrystals—displacement potential formulation
In the previous section, Yang et al[3] introduced the assumption (9.3-2), which is
not valid, and leads to loss a lot of information coming from the phonon-phason
coupling. In studies of Fan and Guo[4] , Zhu and Fan[16] , Zhu, Fan and Guo[17] , and
Li and Fan[5] considered the coupling effects, i.e.,
R = 0
and obtained the complete theory for the plane elasticity of quasicrystals. In the
study the assumption (9.2-1) or (9.3-1) still maintains, i.e.,
∂
= 0. (9.4-1)
∂z
In this case the three-dimensional elasticity can be reduced into a plane elasticity
problem. From condition (9.3-1) directly we have
εzz = wzz = wxz = wyz = 0. (9.4-2)
Thus the number of the field variables and field equations are reduced from 36 to
32. Though the reduction of the total number is not so much, the resulting equation
system has been greatly simplified and with the following form:
⎧ 2
⎪ ∂ ∂ wx ∂ 2 wy ∂ 2 wy
⎪
⎪ μ∇ 2
u + (λ + μ) ∇ 1 · u 1 + R + 2 − = 0,
⎪
⎪
1 x
∂x ∂x2 ∂x∂y ∂y 2
⎪
⎪
⎪
⎪ ∂ ∂ 2 wy ∂ 2 wx ∂ 2 wy
⎪
⎪ μ∇21 uy + (λ + μ) ∇1 · u1 + R − −
⎪
⎪ 2 = 0,
⎪
⎪ ∂y ∂x2 ∂x∂y ∂y 2
⎪
⎪
⎪
⎪ ∂ 2 wx ∂ 2 wy ∂ 2 wx
⎪ 2
⎨ μ∇1 uz + R − 2 − + ∇21 wz = 0,
∂x2 ∂x∂y ∂y 2
2 2
⎪
⎪ ∂ wz ∂ 2 wz ∂ ux ∂ 2 uy ∂ 2 ux ∂ 2 uz ∂ 2 uz
⎪
⎪ K ∇ 2
1 1 x w +K 2 − +R −2 − + − = 0,
⎪
⎪ ∂x2 ∂y 2 ∂x2 ∂x∂y ∂y 2 ∂x2 ∂y 2
⎪
⎪
⎪
⎪ ∂ 2 wz ∂ 2 uy ∂ 2 ux ∂ 2 uy ∂ 2 uz
⎪
⎪ ∇ 2
− − −
⎪
⎪ K 1 1 w y 2K 2 + R + 2 2 = 0,
⎪
⎪ ∂x∂y ∂x2 ∂x∂y ∂y 2 ∂x∂y
⎪
⎪
⎪
⎪ ∂ 2 wx ∂ 2 wy ∂ 2 wy
⎩ (K1 − K2 )∇21 wz + K2 − 2 − + R∇21 uz = 0,
∂x2 ∂x∂y ∂y 2
(9.4-3)
where ∇21 and ∇1 · u1 are the same of those in Section 9.3, but the suffix 1 of the
two-dimensional Laplace operator will be omitted in the following for simplicity.
The equation set is much simpler than that of (9.1-8) but still quite complicated.
If we introduce a displacement potential F (x, y) such as
⎧
⎪ ∂2
⎪
⎪ u = R ∇2 ∇2 [μαΠ1 + β(λ + 2μ)Π2 ]F
⎪
⎪
x
∂x∂y
⎪
⎪
⎪
⎪ ∂2 ∂4 ∂4 ∂4
⎪
⎪ − −
⎪
⎪ +c 0 R Λ (3μ λ)
∂x4
+ 10(λ + μ)
∂x2 ∂y 2
(5λ + 9μ)
∂y 4
F
⎪
⎪ ∂x∂y

⎪
⎪
⎪
⎪ ∂2 ∂2
⎪
⎪ u = R∇ 2 2
∇ μα Π1 − β(λ + 2μ) Π 2 F
⎪
⎪
y
∂y 2 ∂x2
⎪
⎪
⎪
⎪ ∂6 ∂6 ∂6 ∂6
⎪
⎪ 2
+c0 RΛ (λ + 2μ) 6 − 5(2λ + 3μ) 4 2 + 5λ 2 4 + μ 6 F,
⎨ ∂x ∂x ∂y ∂x ∂y ∂y
2
2 2

⎪
⎪ ∂ ∂ ∂
⎪
⎪ uz = c1 (α − β)Λ2 Π1 Π2 + α 2 Π12 + β 2 Π22 F,
⎪
⎪ ∂x∂y ∂y ∂x
⎪
⎪
⎪
⎪ ∂2
⎪
⎪ −ω ∇2 2c0 Λ2 ∇2 − (α − β)Π1 Π2 F
⎪
⎪ w x =
⎪
⎪ ∂x∂y
⎪
⎪
⎪
⎪ ∂2 2 ∂2 2
⎪
⎪
2 2 2 2
wy = −ω∇ c0 Λ Λ ∇ + α 2 Π1 + β 2 Π2 F,
⎪
⎪ ∂y ∂x
⎪
⎪
⎪
⎪ ∂ 2
∂ 2
∂2
⎪
⎩ wz = c2 (α − β)Λ2 Π1 Π2 + α 2 Π12 + β 2 Π22 F,
∂x∂y ∂y ∂x
(9.4-4)
9.4 Phonon-phason coupled plane elasticity of icosahedral quasicrystals... 171
then the field equations mentioned above will be satisfied if
∇2 ∇2 ∇2 ∇2 ∇2 ∇2 F (x, y) + ∇2 LF (x, y) = 0, (9.4-5)
where
⎧
⎪
⎪ α = (λ + 2μ)R2 − ωK1 , β = μR2 − ωK1 , ω = μ(λ + 2μ),
⎪
⎪
⎪
⎪
⎨ μK22 +(K1 −3K2 )R2 (K1 −2K2 )Rω (K2 μ−R2 )ω
c0 = ω , c1 = , c2 = ,
μ(K1 −K2 )−R 2 μ(K1 −K2 )−R 2 μ(K1 −K2 )−R2
⎪
⎪
⎪
⎪
⎪
⎪ ∂2 ∂2 ∂2 ∂2 ∂2 ∂2 ∂2 ∂2
⎩ Π1 = 3 2 − 2 , Π2 = 3 2 − 2 , ∇2 = 2 + 2 , Λ2 = 2 − 2 ,
∂x ∂y ∂y ∂x ∂x ∂y ∂x ∂y
(9.4-6)
in which the suffix 1 of the two-dimensional Laplace operator is omitted and

c0 ∂ 10 α ∂ 10 α ∂ 10
L= − 10 + 5 4 − 5 8 2
− 10 11 − 10
β ∂x β ∂x ∂y β ∂x6 ∂y4

α ∂ 10 α ∂ 10 α ∂ 10
+ 10 10 − 11 − 5 5 − 4 − . (9.4-7)
β ∂x4 ∂y6 β ∂x2 ∂y8 β ∂y 10
Assuming
R2
1 (9.4-8)
μK1
(this is understandable, because the coupling effect is weaker than that of phonon),
then from equations (9.4-6) and (9.4-7),
β c0 2 2 2 2 2 2
→ 1, ∇2 L = ∇ ∇ ∇ ∇ ∇ ∇ , (9.4-9)
α β
substituting (9.4-9) into (9.4-5), we find that
∇2 ∇2 ∇2 ∇2 ∇2 ∇2 F (x, y) = 0. (9.4-10)
This sextuple harmonic equation is the governing equation for plane elasticity
of quasicrystals based on the displacement potential formulation. With the aid of
the generalized Hooke’s law, the phonon and phason stress components can also be
expressed in terms of potential function F (x, y) and these expressions are omitted
here due to the limitation of the space.
In other words, equation set (9.4-4) gives a fundamental solution in terms of
F (x, y) for the plane elasticity problem of an icosahedral quasicrystal. Once the
function F (x, y) satisfying equation (9.4-4) is determined for prescribed boundary
conditions, the entire elastic field of an icosahedral quasicrystal can be found from
(9.4-4). This formulation has been reported briefly by Fan and Guo[4] . In some
extent it is a development of Li and Fan[18] for elasticity of two-dimensional qua-
sicrystals. The application of the formulation and relevant solution will be given in
Section 6.
9.5 Phonon-phason coupled plane elasticity of icosahedral qu-

asicrystals—stress potential formulation
In the previous section the displacement formulation exhibits its effect, which reduces
a very complicated partial differential equation set into a single partial differential
equation with higher order (12th order), the latter will be easily to solve. At the
meantime the stress potential formulation is effective too. In this section we will
introduce the formulation.
To contrast the displacement potential formulation we now retain the stress com-
ponents and eliminate the displacement components. From the deformation geome-
try equations (9.1-1) and considering (9.3-1) and (9.3-2) we obtain the deformation
compatibility equations as follows:
⎧ 2
⎪ ∂ εxx ∂ 2 εyy ∂ 2 εxy ∂εyz ∂εzx
⎪
⎨ ∂y 2 + 2
= 2 , = ,
∂x ∂x∂y ∂x ∂y
(9.5-1)
⎪
⎪ ∂wxy ∂wxx ∂wyy ∂wyx ∂wzy ∂wzx
⎩ = , = , = .
∂x ∂y ∂x ∂y ∂x ∂y
Thus the displacements are eliminated already.

By combining (9.1-7) and (9.5-1) we obtain the deformation compatibility equa-
tions expressed by stress components, so that the strain components have been
eliminated up to now ( those equations are too lengthy we here do not list them).
So far one has the deformation compatibility equations expressed by stresses and
equilibrium equations only.
If we introduce stress potential functions
ϕ1 (x, y), ϕ2 (x, y), ψ1 (x, y), ψ2 (x, y), ψ3 (x, y)
such as
⎧
⎪ ∂ 2 ϕ1 ∂ 2 ϕ1 ∂ 2 ϕ1
⎪
⎪ σxx = , σ = − , σ = ,
⎪
⎪ ∂y 2
xy
∂x∂y
yy
∂x2
⎪
⎪
⎪
⎪
⎪
⎪ ∂ϕ2 ∂ϕ2
⎪
⎨ σzx = ∂y , σzy = − ∂x ,
⎪
(9.5-2)
⎪
⎪ ∂ψ1 ∂ψ1 ∂ψ2
⎪
⎪
⎪ Hxx = ∂y , Hxy = − ∂x , Hyx = ∂y ,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎩ Hyy = − ∂ψ2 , Hzx = ∂ψ3 , Hzy = − ∂ψ3 ,
∂x ∂y ∂x
with
9.6 A straight dislocation in an icosahedral quasicrystal 173
⎧ 2
⎪
⎪ ∂ ∂ 2
⎪
⎪ ϕ 1 = c 2 c3 R 2 Π 2 − Λ Π 1 ∇2 ∇2 G,
⎪
⎪ ∂y ∂x2
⎪
⎪
⎪
⎪ ϕ2 = −c3 c4 ∇2 ∇2 ∇2 ∇2 ∇2 G
⎪
⎪
⎪
⎪ 2
⎪
⎨ ∂2 ∂
ψ1 = c1 c2 R 2 2 2 Π1 Π2 − Λ Π1 ∇2 G + c2 c4 Λ2 ∇2 ∇2 ∇2 ∇2 G,
2 2
⎪ ∂y ∂x
⎪
⎪ 2
⎪
⎪ 2
∂2
⎪
⎪
∂ ∂
Π 2
− 2
Π 2
∇2 ∇2 ∇2 ∇2 G,
⎪ 2
⎪ ψ = c1 2c R 2 2 2 Λ 1 2 ∇ G − 2c2 c4
Π
⎪
⎪ ∂x∂y ∂x ∂x∂y
⎪
⎪
⎪
⎪ 1
⎩ ψ3 = − K2 c3 c4 ∇2 ∇2 ∇2 ∇2 ∇2 G,
R
(9.5-3)
then the equilibrium equations and the deformation compatibility equations will be
identically satisfied if
∇2 ∇2 ∇2 ∇2 ∇2 ∇2 G = 0 (9.5-4)
under the approximation R2 /K1 μ 1, which is the final governing equation of plane
elasticity of icosahedral quasicrystals, function G(x, y) is named the stress potential
in which ⎧
⎪ R(2K2 − K1 )(μK1 + μK2 − 3R2 )
⎪
⎪ c 1 = ,
⎪
⎪ 2(μK1 − 2R2 )
⎪
⎪
⎪
⎪
⎪
⎪ 1
⎪
⎪ c2 = K2 (μK2 − R2 ) − R(2K2 − K1 ),
⎪
⎪ R
⎪
⎪
⎪
⎪ (μK2 − R2 )2
⎪
⎪ 2
⎪
⎨ c 3 = μ(K 1 − K 2 ) − R − ,
μK1 − 2R2
(9.5-5)
⎪
⎪ 1 μK 1 − 2R 2
⎪
⎪ c4 = c1 R + c3 K1 + ,
⎪
⎪ 2 λ+μ
⎪
⎪
⎪
⎪
⎪
⎪ ∂2 ∂2 ∂2 ∂2
⎪
⎪ Π1 = 3 2 − 2 , Π2 = 3 2 − 2 ,
⎪
⎪ ∂x ∂y ∂y ∂x
⎪
⎪
⎪
⎪
⎪
⎪
2 2 2 2
⎩ ∇2 = ∂ + ∂ , Λ2 = ∂ − ∂
∂x2 ∂y2 ∂x2 ∂y2
In derivation of (9.5-4) the approximation (9.4-8) is used at the last step.
This work is given by Ref. [5], which may be seen as a development of the study
for two-dimensional quasicrystals given by Guo and Fan (see e.g. Fan[19] or Guo and
Fan[20] ).
9.6 A straight dislocation in an icosahedral quasicrystal

The formulations exhibited in the Sections 9.4 and 9.5 are meaningful, which have
greatly simplified the complicated equations involving elasticity. Their applications
will be addressed in this and subsequent sections, in which the Fourier analysis and
complex variable function method play important roles.
We introduced the dislocation solution of Yang et al[3] in an icosahedral qua-

sicrystal. The authors of Ref. [3] assumed that the coupling effect between phonons
and phasons is omitted, i.e., R = 0. In the case the phonon solution of the dis-
location is the same to that of the classical isotropic elastic solution of an edge
dislocation. And the phason solution of the problem is independent of the phonon
field. In this section we try to give a complete analysis of the problem in which the
phonon-phason coupling effect is taken into account.
For a dislocation along x3 -axis (or z-axis) in an icosahedral quasicrystal with the

core at the origin: the Burgers vector is denoted as b = b ⊕b⊥ = (b1 , b2 , b3 , b⊥ ⊥ ⊥
1 , b2 , b3 ),
where the dislocation conditions are

duj = bj , dwj = b⊥j , (9.6-1)
Γ Γ
in which x1 = x, x2 = y, x3 = z, and the integrals in (9.6-1) should be taken along

the Burgers circuit surrounding the dislocation core in space E . By using the
superposition principle, we here calculate first the elastic field for a special case, i.e.,

which corresponds to b1 = 0, b⊥ ⊥ ⊥
1 = 0, b2 = b3 = 0 and b2 = b3 = 0.
For simplicity we can solve a half-plane problem by considering symmetry and
anti-symmetry of relevant field variables, so there are the following boundary con-
ditions including the dislocation condition:
σyy (x, 0) = 0, (9.6-2a)
σzy (x, 0) = 0, (9.6-2b)
Hyy (x, 0) = 0, (9.6-2c)
Hzy (x, 0) = 0, (9.6-2d)

dux = b1 , (9.6-2e)
Γ

dwx = b⊥
1. (9.6-2f)
Γ
In addition there are boundary conditions at infinity:

#
σij (x, y) → 0, Hij (x, y) → 0, x2 + y 2 → ∞. (9.6-3)
In the following we use the formulation of Section 9.4 to solve the above boundary
value problem. Performing the Fourier transform to equation (9.4-10) and the above
boundary conditions, we obtain the solution at the transformed domain, then taking
inversion of the Fourier transform, we obtain the solution as follows:

⎧
⎪
⎪ 1 y xy xy 3
⎪
⎪ u = b arctan + c12 2 + c13 4 ,
2π 1
x
⎪
⎪ x r r
⎪
⎪
⎪
⎪
⎪
⎪ 1 r y2 y 2 (y 2 − x2 )
⎪
⎪ uy = −c21 ln + c22 2 + c23 ,
⎪
⎪ 2π r0 r 2r 4
⎪
⎪
⎪
⎪ xy 3
⎪
⎪ 1 y xy
⎪
⎨ uz = −c 31 arctan + c 32 + c 33 ,
2π x r2 r4
(9.6-4)
⎪
⎪ 1 y xy xy 3
⎪
⎪ x
w = b ⊥
arctan + c + c ,
⎪
⎪ 2π 1 x
42 2
r
43 4
r
⎪
⎪
⎪
⎪
⎪
⎪ 1 r y2 y 2 (y 2 − x2 )
⎪
⎪ wy = −c51 ln + c52 2 + c53 ,
⎪
⎪ 2π r0 r 2r 4
⎪
⎪
⎪
⎪
⎪
⎪ 1 y xy xy 3
⎪
⎩ wz = −c61 arctan + c62 2 + c63 4 ,
2π x r r
in which r 2 = x2 + y2 , r0 is the radius of the dislocation core and cij are constants
shown as follows:
⎧
⎪ 2 2 2 2 ⊥
⎪
⎪ c12 = 2c0 (μ(2R +c0 μ)(λ +3λμ+2μ )b1 +R(−e(λ+μ)+2μc0 (λ+2μ) )b1 ) ,
⎪
⎪
⎪
⎪ −e(2e+μc0 (λ+2μ))+μc0 (λ+2μ)(e+2μc0 (λ + 2μ))
⎪
⎪
⎪
⎪
⎪
⎪

2c0 R(λ + μ)(2Rμ(λ + μ)b1 + 2μc0 (λ + 2μ)b⊥
⎪
⎪ 1)
⎪
⎪
c13 =
−e(2e
,
⎪
⎪ + μc 0 (λ + 2μ)) + μc0 (λ + 2μ)(e + 2μc0 (λ + 2μ))
⎪
⎪
⎪
⎪
⎪
⎪ (2c20 μ3 (λ + 2μ) − 2e2 )b1 + 2c0 R(λ + 3μ)eb⊥
⎪
⎪ c = 1
,
⎪
⎪ 21
−e(2e + μc (λ + 2μ)) + μc (λ + 2μ)(e + 2μc (λ + 2μ))
⎪
⎪ 0 0 0
⎪
⎪
⎪
⎪
⎪
⎪ 2c0 (−μ2 (λ + μ)(−2R2 + c0 (λ + 2μ))b1 + R(−(λ + μ)e + 2c0 μ2 )b⊥ 1)
⎪
⎪ c 22 = ,
⎪
⎪ −e(2e + μc0 (λ + 2μ)) + μc0 (λ + 2μ)(e + 2μc0 (λ + 2μ))
⎪
⎨

⎪ 2c0 R(λ + μ)(2Rμ(λ + μ)b1 + 2c0 μ2 b⊥ 1)
⎪
⎪ c23 = ,
⎪
⎪ −e(2e + μc 0 (λ + 2μ)) + μc0 (λ + 2μ)(e + 2μc0 (λ + 2μ))
⎪
⎪
⎪
⎪
⎪
⎪
3c1 e{2(c0 μ + 7e)μc0 (λ + 2μ)b1 + R(54c20 (λ2 + 3λμ + μ2 )
⎪
⎪
⎪
⎪
⎪
⎪ −2(α − β)(e + μc0 (λ + 2μ)))b⊥
⎪
⎪ 1}
⎪
⎪ c 31 = ,
⎪
⎪ 4c0 R(−e(2e + μc 0 (λ + 2μ)) + μc 0 (λ + 2μ)(e + 2μc0 (λ + 2μ))
⎪
⎪
⎪
⎪
⎪
⎪
3c1 e(2μ(−e + μc0 (λ + 2μ))b1 + R(−2e + 2μc0 (λ + 2μ))b⊥
⎪
⎪ 1)
⎪ 32
⎪ c =
−e(2e
,
⎪
⎪ + μc 0 (λ + 2μ)) + μc 0 (λ + 2μ)(e + 2μc 0 (λ + 2μ))
⎪
⎪
⎪
⎪
⎪
⎪ −3ec1 (2Rμ(λ + μ)b1 + 2μc0 (λ + 2μ)b⊥ 1)
⎪
⎩ c33 = ,
−e(2e + μc0 (λ + 2μ)) + μc0 (λ + 2μ)(e + 2μc0 (λ + 2μ))
⎧
⎪
⎪ −2e(2Rμ(λ + μ)b1 + 2μc0 (λ + 2μ)b⊥ 1)
⎪
⎪ c42 = ,
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
c43 = 0,
⎪
⎪
⎪
⎪
⎪
⎪ {−4eμ2 c0 (λ + 2μ)b1 + R(2(λ + 2μ)(e + 0.5μc0 )
⎪
⎪
⎪
⎪
⎪
⎪ +μ(2β 2 μ + 2c20 (λ + 2μ)2 + c0 (λ + 2μ)(−βμ + R2 (λ + μ)))b⊥
1}
⎪
⎪ c51 =− ,
⎪
⎪ R(−e(2e + μc0 (λ + 2μ)) + μc0 (λ + 2μ)(e + 2μc0 (λ + 2μ)))
⎪
⎪
⎪
⎪
⎪
⎪
2e(2Rμ(λ + μ)b1 + 2μc0 (λ + 2μ)b⊥
⎪
⎪ 1)
⎪
⎨ c52 =− ,
⎪
⎪ c53 =0
⎪
⎪
⎪
⎪
⎪
⎪
3c2 e{(2(c0 μ + 7e)μc0 (λ + 2μ)b1 + R(54c20 (λ2 + 3λμ + μ2 )
⎪
⎪
⎪
⎪
⎪
⎪ −2(α − β)(e + μc0 (λ + 2μ)))b⊥
⎪
⎪ 1 )}
⎪
⎪ c61 =− ,
⎪
⎪ 4c0 R(−e(2e + μc0 (λ + 2μ))(e + 2μc0 (λ + 2μ)))
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 3ec2 (2μ(−e + μc0 (λ + 2μ))b1 + R(−2e + 2μc0 (λ + 2μ))b⊥1)
⎪
⎪ c62 = ,
⎪
⎪
⎪
⎪
⎪
⎪
⎪ −3ec2 (2Rμ(λ + μ)b1 + 2μc0 (λ + 2μ)b⊥1)
⎪
⎩ c63 = .
(9.6-5)
2
with e = −(λ + μ)R .
For the other two typical problems, in which the Burgers vector of the dislocation

is denoted by (0, b2 , 0, 0, b⊥ ⊥
2 , 0) and (0, 0, b3 , 0, 0, b3 ) respectively, a complete similar
consideration will yield similar results, which are omitted here. Alternatively, the
(2) (2) (3) (3)
expressions are denoted as uj , wj , uj , wj .

Analytic expressions for elastic field for a dislocation (b1 , b2 , b3 , b⊥ ⊥ ⊥
1 , b2 , b3 ) in the
icosahedral quasicrystal can be obtained by superposition of the corresponding ex-

pressions for the elastic fields for (b1 , 0, 0, b⊥ ⊥
1 , 0, 0), (0, b2 , 0, 0, b2 , 0) and (0, 0, b3 , 0, 0,
b⊥
3 ), namely,
(1) (2) (3) (1) (2) (3)

uj = uj + uj + uj , wj = wj + wj + wj , i, j = 1, 2, 3. (9.6-6)
We can see that the interaction among phonon-phonon, phason-phason and phonon-
phason is very evident, so the solution (9.6-4) is quite different from the solution
given by Yang et al[3] (whose solution for phonon displacement field is given by the
first three formulas of equations (9.3-6), and will be quoted again in the following, see
formula (9.6-7)), where they took R = 0, i.e., they assumed the phonon and phason
are decoupled, so the solution for phonon is the same as the classical solution for
crystals. It is obvious that our solution given by (9.6-4) explores that the realistic
case for quasicrystals is quite different from that of crystal. To illustrate the coupling
effect we give some numerical results in Fig. 9.6-1 and Fig. 9.6-2 for the normalized

Fig. 9.6-1 The displacement u1 /b1 versus x for different coupling elastic constants

Fig. 9.6-2 The displacement u1 /b1 versus y for different coupling constants

displacement u1 /b1 versus x and y respectively, in which the results exhibit the
influence of the coupling constant R; the coupling effect is quite remarkable. In the
calculation we take the data of elastic moduli as
λ = 74.9, μ = 72.9(GPa), K1 = 72, K2 = −73(MPa),
and the phonon-phason coupling elastic constant for three different cases: i.e., R/μ =
0, R/μ = 0.04 and R/μ = 0.06, in which the first one corresponds to decoupled case.
The figures show that the coupling effect is very important, and the displacement
is increasing with the growth of value of R.
For icosahedral quasicrystals with the presence of a dislocation, there are five
independent elastic constants. If R = 0, wi = 0, our solution is exactly reduced to
the solution of dislocation of crystals, i.e.,
⎧
⎪
⎪ b1 y λ + μ xy b2 μ r λ + μ x2
⎪
⎪ u = arctan + + ln + ,
⎪
⎪
x
2π x λ + 2μ r2 2π λ + 2μ r0 λ + 2μ r2
⎪
⎪
⎪
⎨
b1 μ r λ + μ x2 b2 y λ + μ xy
⎪ uy = − ln + + arctan − ,
⎪
⎪ 2π λ + 2μ r0 λ + 2μ r2 2π x λ + 2μ r2
⎪
⎪
⎪
⎪
⎪
⎪
⎩ u = b3 arctan y .
z
2π x
(9.6-7)
The displacement potential function formulation establishes the basis for solving
defects problem in icosahedral quasicrystals. The formulation is simplified the so-
lution process. In the subsequent steps a systematic Fourier analysis is developed,
which provides a constructive procedure to find the analytic solution, it is effective
not only for dislocation problem, but also for more complicated mixed boundary
value problems (e.g. crack problems refer to Ref. [25]).
The present solution can be used as a fundamental solution for a dislocation in
an icosahedral quasicrystal. Therefore, many elasticity problems in an icosahedral
quasicrystal can be directly solved with the aid of this fundamental solution by
superposition.
This work has been published in Ref. [17].
9.7 An elliptic notch/Griffith crack in an icosahedral qua-

sicrystal
The solution for notch problem of icosahedral quasicrystals is not available until
2006. The difficulty lies in that cannot be solved by the Fourier transform method.
We must develop other methods, in which the complex variable function-conformal
mapping method is particularly effective. The results are given by Ref. [21], which
may be seen as a development of the work of the same authors in Ref. [22].
In the present section, we consider an icosahedral quasicrystal with an elliptic
notch along the z-axis. On the basis of the general solution obtained in Section 5,
explicit expressions of stress and displacement components of phonon and phason
fields in the quasicrystals are given. With the help of conformal mapping, analytic
solution for elliptic notch problem of the quasicrystals is presented. The solution of
the Griffith crack problem can be observed as a special case of the results, which
9.7 An elliptic notch/Griffith crack in an icosahedral quasicrystal 179
will be reduced to the well-known results in a conventional material if the phason

field is absent.
The solution of (9.5-4) can be expressed as
G(x, y) = Re[g1 (z) + z̄g2 (z) + z̄ 2 g3 (z) + z̄ 3 g4 (z) + z̄ 4 g5 (z) + z̄ 5 g6 (z)], (9.7-1)
where gi (z) are arbitrary analytic functions of z = x + iy, and the bar over complex
variable or complex function denotes the complex conjugate.
By equations (9.5-2)∼(9.5-4) and (9.7-1), the stresses can be expressed as follows:
⎧
⎪
⎪ σxx + σyy = 48c2 c3 RImΓ (z),
⎪
⎪
⎪
⎪
⎪
⎪ σyy − σxx + 2iσxy = 8ic2 c3 R(12Ψ (z) − Ω (z)),
⎪
⎪
⎪
⎪
⎪
⎪ σzy − iσzx = −960c3 c4 f6 (z),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 24λR
⎪
⎪ σzz = c2 c3 ImΓ (z),
⎪
⎪ (μ + λ)
⎪
⎨
Hxy − Hyx − i(Hxx + Hyy ) = −96c2 c5 Ψ (z) − 8c1 c2 RΩ (z), (9.7-2)
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ Hyx − Hxy + i(Hxx − Hyy ) = −480c2 c5 f6 (z) − 4c1 c2 RΘ (z),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ Hyz + iHxz = 48c2 c6 Γ (z) − 4c2 R2 (2K2 − K1 )Ω (z),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 24R2
⎪
⎩ Hzz = c2 c3 ImΓ (z),
(μ + λ)
where
⎧
⎪
⎪ R(2K2 − K1 )(μK1 + μK2 − 3R2 )
⎪
⎪ c 1 = ,
⎪
⎪ 2(μK1 − 2R2 )
⎪
⎪
⎪
⎪
⎪
⎪ 1
⎪
⎪ c = K (μK2 − R2 ) − R(2K2 − K1 ),
⎨ 2 R 2
⎪
⎪ (μK2 − R2 )2
⎪
⎪ c3 = μ(K1 − K2 ) − R2 − ,
⎪
⎪ μK1 − 2R2
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 1 μK1 − 2R2
⎩ c4 = c1 R + c3 K1 + ,
2 λ+μ
and
⎧
⎪
⎪ Ψ (z) = f5 (z) + 5z̄f6 (z),
⎪
⎪
⎪
⎪ Γ (z) = f4 (z) + 4z̄f5 (z) + 10z̄ 2 f6 (z),
⎪
⎪
⎪
⎪
⎨ Ω (z) = f (z) + 3z̄f (z) + 6z̄ 2 f (z) + 10z̄ 3 f (z),
3 4 5 6
(9.7-3)
⎪
⎪ Θ(z) = f2 (z) + 2z̄f (z) + 3z̄ 2 f (z) + 4z̄ 3 f (z) + 5z̄ 4 f (IV ) (z),
⎪
⎪ 3 4 5 6
⎪
⎪
⎪
⎪ 2
⎪ c5 = 2c4 − c1 R, c6 = (2K2 − K1 )R2 − 4c4 μK2 − R .
⎪
⎩
μK1 − 2R2
In the above expressions, the function g1 (z) does not appear, this implies that for
stress boundary value problem in this formalism only five complex potentials g2 (z),
g3 (z), g4 (z), g5 (z) and g6 (z) are needed, and one can take g1 (z) = 0. For simplicity,
we have introduced the following new symbols:
(9) (8) (7)
g2 (z) = f2 (z), g3 (z) = f3 (z), g4 (z) = f4 (z),
(9.7-4)
(6) (5)
g5 (z) = f5 (z), g6 (z) = f6 (z),
(n)
where gi denotes the n-th derivative with the argument z, accordingly f1 (z) = 0.
Similar to formulation in Chapter 8, the complex representations of displacement
components can be written as follows (here we have omitted the rigid-body displace-
ments)
⎧
⎪
⎪ uy + iux = −6c2 R
2c3
+ c7 − 2c2 c7 RΩ (z),
⎪
⎪
⎪
⎪ μ+λ
⎪
⎪
⎪
⎪ 4
⎪
⎪
⎪ uz = μ(K + K ) − 3R2 (240c10 Imf6 (z))
⎪
⎪
⎪ 1 2
⎪
⎪
⎨ 2
+ c1 c2 R Im(Θ(z) − 2Ω (z) + 6Γ (z) − 24Ψ (z))),
(9.7-5)
⎪
⎪ R
⎪
⎪ w + iw = − (24c 9 Ψ (z) − c8 Θ(z)),
⎪
⎪
y x
c1 (μK1 − 2R2 )
⎪
⎪
⎪
⎪
⎪
⎪ w = 4(μK2 − R2 )
⎪
⎪ (240c10 Imf6 (z))
⎪
⎪
z
(K1 − 2K2 )R(μ(K1 + K2 ) − 3R2 )
⎪
⎪
⎩
+ c1 c2 R2 Im(Θ(z) − 2Ω (z) + 6Γ (z) − 24Ψ (z))),
in which
⎧ c3 K1 + 2c1 R
⎪
⎪ c7 = ,
⎪
⎪ μK1 − 2R2
⎪
⎪
⎪
⎪
⎨ c8 = c1 c2 R(μ(K1 − K2 ) − R2 ),
(9.7-6)
⎪
⎪ (μK2 − R2 )2
⎪
⎪ c = c + 2c c c − ,
⎪
⎪
9 8 2 4 3
μK1 − 2R2
⎪
⎪
⎩
c10 = c1 c2 R2 − c4 (c2 R − c3 K1 ).
9.7.2 Elliptic notch problem

We consider an icosahedral quasicrystal solid with an elliptic notch, which penetrates
through the medium along the z-axis direction,the edge of the elliptic notch is subject
to a uniform pressure p, see Fig. 9.7-1.
The boundary conditions of this problem can be expressed as follows:
σxx cos(n, x) + σxy cos(n, y) = Tx , (x, y) ∈ L,

σxy cos(n, x) + σyy cos(n, y) = Ty ,
(9.7-7)
Hxx cos(n, x)+Hxy cos(n, y) = hx , Hyx cos(n, x)+Hyy cos(n, y) = hy , (x, y) ∈ L,
(9.7-8)
σzx cos(n, x) + σzy cos(n, y) = 0, Hzx cos(n, x) + Hzy cos(n, y) = 0, (x, y) ∈ L,
(9.7-9)
where
dy dx
cos(n, x) = , cos(n, y) = − , Tx = −p cos(n, x), Ty = −p cos(n, y),
ds ds
(9.7-7 )
Fig. 9.7-1 An elliptic notch subject to an inner pressure in an icosahedral quasicrystal
Tx , Ty denote the components of surface traction, p is the magnitude of the

pressure, hx , hy are components of the generalized surface traction, n is the outward
x2 y 2
unit normal vector of any point of the boundary, L : 2 + 2 = 1 is the edge of the
a b
elliptic notch. Since the measurement of generalized traction has not been reported
so far, for simplicity, we assume that
hx = 0, hy = 0. (9.7-8 )
Utilizing equations (9.7-2), (9.7-3), (9.7-7) and (9.7-7 ), one has
− 4c2 c3 R[3(f4 (z) + 4z̄f5 (z) + 10z̄ 2 f6 (z))

− (f3 (z) + 3zf4 (z) + 6z 2 f5 (z) + 10z 3 f6 (z))]

= (Tx + iTy )ds = ipz. (9.7-10)
Taking complex conjugate on both sides of equation (9.7-10) yields
− 4c2 c3 R[3(f4 (z) + 4zf5 (z) + 10z 2 f6 (z))

− (f3 (z) + 3z̄f4 (z) + 6z 2 f5 (z) + 10z̄ 3 f6 (z))] = −ipz̄. (9.7-11)
From equations (9.7-2), (9.7-3), (9.7-8) and (9.7-8 ), we have
48c3 (2c4 − c1 R)ReΨ (z) + 2c1 c3 RReΘ(z) = 0,

(9.7-12)
−48c3 (2c4 − c1 R)ImΨ (z) − 2c1 c3 RImΘ(z) = 0.
Multiplying the second formula of (9.1-12) by -i and adding it to the first, we obtain
48c3 (2c4 − c1 R)Ψ (z) + 2c1 c3 RΘ(z) = 0. (9.7-13)
By equations (9.7-2), (9.7-3) and (9.7-9), the following expression results in
f6 (z) + f6 (z) = 0,
4c11 Re[f5 (z) + 5z̄f6 (z)] + (2K2 − K1 )RRe[f4 (z) + 4z̄f5 (z)
+ 10z̄ 2 f6 (z) + 20f6 (z)] = 0, (9.7-14)
in which
4c4 (μK2 − R2 )
c11 = (2K2 − K1 )R − . (9.7-15)
(μK1 − 2R2 )R
However further calculation will be very difficult on the z-plane owing to the com-
plexity of the manipulation, we must employ conformal mapping

1
z = ω(ζ) = R0 + mζ (9.7-16)
ζ
to transform the region with the ellipse at the z-plane onto the interior of the unit
circle γ at ζ-plane, in which
a+b a−b
R0 = , m= .
2 a+b
Let
fi (z) = fi [ω(ζ)] = Φi (ζ), i = 2, 3, · · · 6. (9.7-17)
Substituting (9.7-16) into (9.7-10), (9.7-11), (9.7-13) and (9.7-14), then multiplying
both sides of equations by dσ/[2πi(σ − ζ)](σ represents the value of ζ at the unit
circle) and integrating around the unit circle γ, by means of Cauchy’s integral for-
mula and analytic extension of the complex variable function theory, we obtain (see
the Major Appendix of this book for details)
⎧
⎪ R0 ipζ(ζ 2 + m)(m3 ζ 2 + 1) (2K2 − K1 )R0
⎪
⎪ Φ2 (ζ) = + ×
⎪
⎪ 2c2 c3 R (mζ 2 − 1)3 2c2 c3 c11
⎪
⎪
⎪
⎪
⎪
⎪ pmζ 3 (ζ 2 + m)[m2 ζ 6 − (m3 + 4m)ζ 4 + (2m4 + 4m2 + 5)ζ 2 + m]
⎪
⎪ ,
⎪
⎪ (mζ 2 − 1)5
⎪
⎪
⎨
R0 ipζ(m2 + 1) (2K2 − K1 )R0 pmζ 3 (ζ 2 + m)(mζ 2 − m2 − 2)
Φ3 (ζ) = − ,
⎪
⎪ 4c2 c3 R (mζ 2 − 1) 12c2 c3 c11 (mζ 2 − 1)3
⎪
⎪
⎪
⎪
⎪
⎪ R0 (2K2 − K1 )R0 pmζ(ζ 2 + m)
⎪
⎪ Φ4 (ζ) = − ipmζ − ,
⎪
⎪ 12c2 c3 R 2c2 c3 c11 (mζ 2 − 1)
⎪
⎪
⎪
⎪
⎩ Φ5 (ζ) = − (2K2 − K1 )R0 pmζ, Φ6 (ζ) = 0.
⎪
48c2 c3 c11
(9.7-18)
The elliptic notch problem has been solved.
The solution of the Griffith crack subjected to a uniform pressure can be obtained
if put m = 1, R0 = a/2 in the above notch solution. The solution of crack can be
expressed explicitly in the z-plane, for example,
$ %
z ia2 y 3(2K2 − K1 )R ipa2 y
σyy =Im ip √ +# −1 + #
z 2 − a2 (z 2 − a2 )3 2c11 (z 2 − a2 )3
(2K2 − K1 )R ipy(2a4 − 3z z̄) (2K2 − K1 )R a2 pz(z z̄ − a2 )
+ # − #
2c11 (z 2 − a2 )5 4c11 (z 2 − a2 )5

(2K2 − K1 )R a2 pz̄
+ # , (9.7-19)
4c11 (z 2 − a2 )3
#
2c3 ip
uy = − 6c2 R + c7 Re (z − z 2 − a2 )
μ+λ 24c2 c3 R
#
2K2 − K1 z z̄ a2
+ p √ −√ − z 2 − a2
24c2 c3 c11 z 2 − a2 z 2 − a2

ip z z̄ a2 2K2 − K1
− 2c2 c7 Re √ −√ − z̄ − ipy
8c2 c3 R z 2 − a2 z 2 − a2 4c2 c3 c11
$ %
2K2 − K1 a2 [(z z̄ − a2 ) + 2iyz̄]
+ p #
16c2 c3 c11 (z 2 − a2 )3
#
2K2 − K1 a2 2z z̄
+ p √ −√ +2 z −a2 2 . (9.7-20)
16c2 c3 c11 z 2 − a2 z 2 − a2
From equations (9.7-19) and (9.7-20), the stress intensity factor and energy release
rate can be evaluated as follows:
√
K1 = πap,
a
1 ∂
GI = 2 (σyy (x, 0) ⊕ H(x, 0))(uy (x, 0) ⊕ wy (x, 0))dx
2 ∂a −a

1 1 c7
= + (KI )2 , (9.7-21)
2 λ + μ c3
in which material constant c3 is given by (9.5-5) and c7 by (9.7-6), it is evident that

the crack energy release depends upon not only phonon elastic constants λ, μ but
also phason elastic constants K1 , K2 and phono-phason coupling elastic constant R,
though we assumed phason tractions hx = hy = 0.
9.7.3 Brief summary

The notch problem can be solved only by complex variable function method, the solu-
tion includes that of Griffith crack problem naturally. Though the Fourier transform
can solve the Griffith crack problem, referring to Zhu and Fan [25], it cannot solve
the notch problem. Whatever solution for notch or crack here reveals the effects
of not only phonon but also phason and phonon-phason coupling. The numerical
examples on crack opening displacement δ(x) = uy (x, +0) − uy (x, −0) and energy
release rate GI for different values of R/μ given by Ref. [25] are shown in Figs. 9.7-2
and 9.7-3 respectively.
Both solutions given by the complex variable function method and the Fourier
transform reduce to that of the classical theory when phason is absent, this is helpful
to examine the present work.
Fig. 9.7-2 Influence of phason and phonon-phason coupling to the crack opening
displacement
9.8 Elasticity of cubic quasicrystals—the anti-plane and axisymmetric deformation 185
Fig. 9.7-3 Influence of phason and phonon-phason coupling to the energy release rate
This work developed the previous work for elasticity of two dimensional qua-
sicrystals of Fan and co-workers. The work is helpful to understand quantitatively
the influence of elliptic notch and crack on the mechanical behaviour of icosahedral
quasicrystals. The stress intensity factor and energy release rate are also obtained
as the direct results of the solution, which are useful for fracture mechanics.
The rigorous theory on the complex potential method will be summarized in
Chapter 11, which can also be referred to Ref [26].
9.8 Elasticity of cubic quasicrystals—the anti-plane and ax-

isymmetric deformation
Cubic quasicrystal is one of important three-dimensional quasicrystals. There are
few analytic solutions availabe in the literature. Since the phasons in this case have
the same irreducible representation with the phonons, the stress and strain tensors
are symmetric. With this feature we consider two cases, one for anti-plane and
another for axisymmetric elastic theory of cubic quasicrystal, and the latter can re-
veal the three-dimensional effect of the elasticity. In addition, a penny-shaped crack
problem under tensile loading in the material is investigated, and the exact ana-
lytic solution is obtained by using the Hankel transform and dual integral equations
theory, and the stress intensity factor and the strain energy release rate are then
determined, which provide some useful information for studying deformation and
fracture of the quasicrystalline material.
From the physical basis provided by Hu et al[2] , we first discuss the anti-plane
elasticity such as, i.e., the stress-strain relation
σ23 = 2C44 ε23 + R44 w23 ,

σ31 = 2C44 ε31 + R44 w31 ,
H23 = 2R44 ε23 + K44 w23 ,
H31 = 2R44 ε31 + K44 w31 ,
the deformation geometry
1 ∂u3 1 ∂u3 ∂w3 ∂w3
ε23 = , ε31 = , w23 = , w31 =
2 ∂x2 2 ∂x1 ∂x2 ∂x1
and the equilibrium equations
∂σ31 ∂σ32 ∂H31 ∂H32
+ = 0, + = 0.
∂x1 ∂x2 ∂x1 ∂x2
These equations are exactly similar to those of anti-plane elasticity of one-dimensional
and icosahedral quasicrystals, so result in the final governing equations
∇2 u3 = 0, ∇2 w3 = 0.
The solution can be derived from the relevant discussion in Chapters 5,7,8 and
Section 9.2, so it need not be mentioned again.
For the axisymmetric case Zhou and Fan[23] developed a displacement potential
theory to reduce basic equations to a single partial differential equation with higher
order in circular cylindrical coordinate system (r, θ, z), i.e., we assume
∂
=0 (9.8-1)
∂θ
and by the generalized Hooke’s law,
⎧
⎪ σrr = C11 εrr + C12 (εθθ + εzz ) + R11 wrr + R12 (wθθ + wzz ),
⎪
⎪
⎪
⎪
⎪
⎪ σθθ = C11 εθθ + C12 (εrr + εzz ) + R11 wθθ + R12 (wrr + wzz ),
⎪
⎪
⎪
⎪
⎪
⎪ σzz = C11 εzz + C12 (εrr + εθθ ) + R11 wzz + R12 (wθθ + wrr ),
⎪
⎪
⎪
⎨ σzr = σrz = 2C44 εrz + 2R44 wrz ,
(9.8-2)
⎪
⎪ Hzz = R11 εzz + R12 (εθθ + εrr ) + K11 wzz + K12 (wθθ + wrr ),
⎪
⎪
⎪
⎪
⎪
⎪ Hrr = R11 εrr + R12 (εθθ + εzz ) + K11 wrr + K12 (wθθ + wzz ),
⎪
⎪
⎪
⎪
⎪
⎪ Hθθ = R11 εθθ + R12 (εrr + εzz ) + K11 wθθ + K12 (wrr + wzz ),
⎪
⎪
⎩
Hzr = Hrz = 2R44 εrz + 2K44 wrz ,
and the equations of deformation geometry

1 ∂ui ∂uj 1 ∂wi ∂wj
εij = + , wij = + ,
2 ∂xj ∂xi 2 ∂xj ∂xi
9.8 Elasticity of cubic quasicrystals—the anti-plane and axisymmetric deformation 187
which are ⎧
⎪
⎪
∂ur ur ∂uz
⎪
⎪ εrr = , εθθ = , εzz = ,
⎪
⎪ ∂r r ∂z
⎪
⎪
⎪
⎪ 1 ∂ur ∂uz
⎪
⎨ εrz = εzr = + ,
2 ∂z ∂r
(9.8-3)
⎪
⎪ ∂wr wr ∂wz
⎪
⎪ wrr = , wθθ = , wzz = ,
⎪
⎪ ∂r r ∂z
⎪
⎪
⎪
⎪ 1 ∂wr ∂wz
⎪
⎩ wrz = wzr = +
2 ∂z ∂r
and the equations of equilibrium
⎧ ∂σ ∂σrz σrr − σθθ

⎪
⎪
rr
+ + = 0,
⎪
⎪ ∂r ∂z r
⎪
⎪
⎪
⎪
⎪
⎪ ∂σzr ∂σzz σzr
⎪
⎨ ∂r + ∂z + r = 0,
(9.8-4)
⎪
⎪ ∂Hrr ∂Hrz Hrr − Hθθ
⎪
⎪ + + = 0,
⎪
⎪ ∂r ∂z r
⎪
⎪
⎪
⎪
⎪ ∂Hzr
⎩ ∂Hzz Hzr
+ + = 0,
∂r ∂z r
If all the displacements and stresses can be expressed by a potential F (r, z) (the
detail on definition of F (r, z) is given by [23]), which satisfies
2 6 2 2 4
∂8 ∂ 1 ∂ ∂ ∂ 1 ∂ ∂
8
−b 2
+ 6
+c 2
+
∂z ∂r r ∂r ∂z ∂r r ∂r ∂z 4
2 3 2 2 4
∂ 1 ∂ ∂ ∂ 1 ∂
−d + + e + F = 0, (9.8-5)
∂r2 r ∂r ∂z 2 ∂r2 r ∂r
then the equations (9.8-2)∼(9.8-4) are identically satisfied.

As an application of above theory and method, the solutions of elastic field of
cubic quasicrystal with a penny-shaped crack are discussed in the following.
Assume a penny-shaped crack with radius a in the centre of the cubic quasicrystal
material, the size of the crack is much smaller than that of the solid, so that the
size of the material can be considered as infinite. At the infinity, the quasicrystal
material is subjected to a tension p in z-direction. The origin of coordinate system
is at the centre of the crack (as shown in Fig. 9.8-1).
From the symmetry of the problem, it is sufficient to study the upper half-space
z > 0 or the lower half-space z < 0. In this case, for studying the upper half-space,
the boundary conditions of the problem are described by
⎧ √
⎪
⎪ r2 + z 2 → ∞ : σzz = p0 Hzz = 0, σrz = 0, Hrz = 0,
⎨
z = 0, 0 r a, σzz = σrz = 0; Hzz = Hrz = 0, (9.8-6)
⎪
⎪
⎩
z = 0, r > a : σrz = 0, uz = 0; Hrz = 0, wz = 0.
These boundary conditions can be replaced by

⎧ √
⎪
⎪ r2 + z 2 → ∞ : σij = 0, Hij = 0,
⎨
z = 0, 0 r a, σzz = −p, σrz = 0; Hzz = Hrz = 0, (9.8-6 )
⎪
⎪
⎩
z = 0, r > a : σrz = 0, uz = 0; Hrz = 0, wz = 0,
which are equivalent to (9.8-6) in the sense of fracture mechanics, if p = p0 .
Fig. 9.8-1 Penny-shaped crack subject to a tension in a cubic quasicrystal
By taking the Hankel transform to equation (9.8-5) and boundary conditions

(9.8-6 ), the solution at the transformed space is such as
F̄ (ξ, z) = A1 e−λ1 ξz + A2 e−λ2 ξz + A3 e−λ3 ξz + A4 e−λ4 ξz , (9.8-7)
where Ai (i = 1, 2, 3, 4) are unknown functions of ξ to be determined and λi (i =

1, 2, 3, 4) are eigen roots obtained from the ordinary differential equation of F̄ (ξ, r).
Accoding to the boundary conditions Ai (ξ) can be determined by solving the fol-
lowing dual integral equations
References 189
⎧ ∞
⎪
⎪ ξAi (ξ)J0 (ξr)dξ = Mi p0 , 0 < r < a,
⎨
0
∞ (9.8-8)
⎪
⎪
⎩ Ai (ξ)J0 (ξr)dξ = 0, r > a
0
and i = 1, 2, 3, 4, in which Mi are some constants consisting of elastic moduli, J0 (ξr)

the first kind Bessel function of zero order.
According to the theory of dual integral equations (refer to Major Appendix),
we obtain the solution of dual integral equations (9.8-8) as follows:
Ai (ξ) = 2a2 Mi p(2πaξ)−1/2 ξ −7 J3/2 (aξ), (9.8-9)
in which J3/2 (aξ) is the first kind Bessel function of 3/2-order (refer to the Major
Appendix for the detailed calculation).
After some calculation, the stress intensity factor KI , strain energy WI and strain
energy release rate GI can be obtained as follows:
2√ 1 ∂WI 3M p2 a
KI = πap, WI = M p2 a3 , GI = = , (9.8-10)
π 2πa ∂a 2π
where M is the constant composed of the elastic constants which is quite lengthy so
has not been included here.
References
[1] Ding D H, Yang W G, Hu C Z et al. Generalized theory of elasticity of quasicrystals.
Phys Rev B, 1993, 48(10): 7003–7010
[2] Hu C Z, Wang R H, Ding D H et al. Point groups and elastic properties of two-
dimensional quasicrystals. Acta Crystallog A, 1996, 52(2): 251–256
[3] Yang W G, Ding D H et al. Atomtic model of dislocation in icosahedral quasicrystals.
Phil Mag A, 1998, 78(6): 1481–1497
[4] Fan T Y, Guo L H, Final governing equation of plane elasticity of icosahedral qua-
sicrystals. Phys Lett A, 2005, 341(5): 235–239
[5] Li L H, Fan T Y. Final governing equation of plane elasticity of icosahedral quasicrystals–
stress potential method. Chin Phys Lett, 2006, 24(9): 2519–2521
[6] Reynolds G A M, Golding B, Kortan A R et al. Isotropic elasticity of the Al-Cu-Li
quasicrystal. Phys Rev B, 1990, 41(2): 1194–1195
[7] Spoor P S, Maynard J D, Kortan A R. Elastic isotropy and anisotropy in quasicrys-
talline and cubic Al-Cu-Li. Phys Rev Lett, 1995, 75(19): 3462–3465
[8] Tanaka K, Mitarai, Koiwa M. Elastic constants of Al-based icosahedral quasicrystals.
Phil Mag A, 1996, 76(10): 1715–1723
[9] Duquesne J-Y, Perrin B. Elastic wave interaction in icosahedral AlPdMn. Physica
B, 2002, 316–317: 317–320
[10] Foster K, Leisure R G, Shaklee A et al. Elastic moduli of a Ti-Zr-Ni icosahedral

quasicrystal and a 1/1 bcc crystal approximant. Phys Rev B, 1999, 59(17): 11132–
11135
[11] Schreuer J, Steurer W, Lograsso T A et al. Elastic properties of icosahedral i-
Cd84 Yb16 and hexagonal h-Cd51 Yb14 . Phil Mag Lett, 2004, 84(10): 643–653
[12] Sterzel R, Hinkel C, Haas A et al. Ultrasonic measurements on FCI Zn-Mg-Y single
crystals. Europhys Lett, 2000, 49(6): 742–747
[13] Letoublon A, de Boissieu M, Boudard M et al. Phason elastic constants of the
icosahedral Al-Pd-Mn phase derived from diffuse scattering measurements. Phil Mag
Lett, 2001, 81(4): 273–283
[14] de Boissieu M, Francoual S, Kaneko Y et al. Diffuse scattering and phason fluctua-
tions in the Zn-Mg-Sc icosahedral quasicrystal and its Zn-Sc periodic approximant.
Phys Rev Lett, 2005, 95(10): 105503
[15] Edagawa K, So GI Y. Experimental evaluation of phonon-phason coupling in icosa-
hedral quasicrystals. Phil Mag, 2007, 87(1): 77–95
[16] Zhu A Y, Fan T Y, Elastic field of a mode II Griffith crack in icosahedral quasicrystals.
Chinese Physics. B. 2007, 16(4): 1111–1118
[17] Zhu A Y, Fan T Y, Guo L H. A straight dislocation in an icosahedral quasicrystal.
J Phys: Condens Matter, 2007, 19(23): 236216
sicrystals. Chin Phys Lett, 1998, 15(4): 278–280
[19] Fan T Y, Mathematical Theory of Elasticity of Quasicrystals and Its Applications.
Beijing: Beijing Institute of Technology Press, 1999 (in Chinese)
[20] Guo Y C, Fan T Y, A mode-II Griffith crack in decagonal quasicrystals. Appl Math
Mech, 2001, 22(11): 1311–1317
[21] Li L H, Fan T Y. Complex variable function method for solving Griffith crack in an
icosahedral quasicrystal. Science in China G, 2008, 51(6): 773–780
[22] Li L H, Fan T Y. Complex function method for solving notch problem of point 10 two-
dimensional quasicrystal based on the stress potential function. J Phys: Condens,
Matter, 2006, 18(47): 10631–10641
[23] Zhou W M, Fan T Y. Axisymmetric elasticity problem of cubic quasicrystal. Chinese
Physics, 2000, 9(4): 294–303
[24] Fan T Y, Xie L Y, Fan L et al. Study on interface of quasicrystal-crystal. Chin.
Phys. B, submitted, 2009
[25] Zhu A Y and Fan T Y. Elastic analysis of a Griffith crack in icosahedral Al-Pd-Mn
quasicrystal, Int. J. Mod. Phys. B, 2009, 23(16): 3429–3444
[26] Fan T Y, Tang Z Y. The strict theory of complex varible function method of sextuple
harmonic equation and applications, J. Math. Phys., 51(5), 053519, 2010
Chapter 10
Dynamics of elasticity and defects of
quasicrystals
The discussion in Chapters 5−9 is limited in the scope of elastostatics of quasicrys-

tals. In this chapter we study the dynamic problems of the solid phase. Elasto-
dynamics of quasicrystals is a topic with different points of view. The focus of
contradictions between different scholar circles lies in the role of phason variables in
the dynamic process.
Lubensky et al[1] , Socolar and Lubensky[2] pointed out that the phonon field u
and the phason field w play very different roles in the hydrodynamics of quasicrys-
tals, because w is insensitive to spatial translations, the phason modes represent the
relative motion of the constituent density waves. They claimed that the phasons are
diffusive, not oscillatory, and with very large diffusion times. On the other hand ac-
cording to Bak[3,4] , the phason describes particular structure disorders or structure
fluctuations in quasicrystals, and it can be formulated based on a six-dimensional
space description. Since there are six continuous symmetries, there exist six hydro-
dynamic vibration modes. Following this point of view u and w play similar roles in
the dynamics. It is evident that the difference between the argument of Lubensky et
al and Bak’s argument lies only in the dynamics. There is no difference between the
arguments in statics (i.e., static equilibrium). Based on this reason in the discussions
of the previous chapters, we need not distinguish the arguments of either Lubensky
et al or Bak.
Probably due to simpler mathematical formulation, many authors followed the
Bak’s argument e.g.[5 ∼ 12] in dynamic study. In this chapter, we will present some
results given in the references. These are Sections 10.1∼10.4 which constitute the
part one of this chapter. In the meantime, we introduce some other results[13,14,27]
which were carried out by following the argument of Lubensky et al. In this line of
thinking it appears that elastodynamics and hydrodynamics are combined in some
extent so it can be called elasto-/hydro-dynamics of quasicrystals. The discussion
given in Sections 10.5∼10.7 constitutes the part two of this chapter.
The results based on different hypothesis are presented to provide readers for

192 Chapter 10 Dynamics of elasticity and defects of quasicrystals
their consideration and comparison. Though researchers believe that the hydrody-
namics based on the argument of Lubensky et al is more fundamentally sound, the
major shortcoming so far is lack of proper experimental data confirmation. Recently
the research interest in the respect is growing up[13∼16] , but the most important ac-
complishment shall still be the quantitative results.
10.1 Elastodynamics of quasicrystals followed the Bak’s ar-

gument
Ding et al[5] first discussed the elastodynamics of quasicrystals. The basic equa-
tions in deformation geometry and generalized Hooke’s law are the same as those of
elastostatics, i.e.,
1 ∂ui ∂uj ∂wi
εij = + , wij = , (10.1-1)
2 ∂xj ∂xi ∂xj
Hij = Kijkl wkl + Rklij εkl . (10.1-2)
They claimed that law of the momentum conservation holds for both phonons and
phasons, namely for linear and small deformation case the equations of motion are
∂σij ∂ 2 ui ∂Hij ∂ 2 wi
=ρ 2 , =ρ 2 , (10.1-3)
∂xj ∂t ∂xj ∂t
where ρ denotes the average mass density of the material.

This implies that they follow the Bak’s argument. After seven years, Hu et al[6]
confirmed the point of view again. In fact, the final elastodynamic equations can
be deduced by substituting (10.1-1) and (10.1-2) into (10.1-3). The mathematical
structure of this theory is relatively simpler, the formulations are the extension
to that of the classical elastodynamics, so many authors take this formulation to
develop the elastodynamics of quasicrystals and give some applications in defect
dynamics and thermodynamics. In the subsequent sections we will present examples
of applications of the theory.
10.2 Elastodynamics of anti-plane elasticity for some qua-

sicrystals
For three-dimensional icosahedral or cubic or one-dimensional hexagonal quasicrys-
tals, in the anti-plane elasticity the basic equations have similar form. First we
consider icosahedral quasicrystals, which have the constitutive relations as
10.2 Elastodynamics of anti-plane elasticity for some quasicrystals 193
⎧
⎪ ∂uz ∂wz
⎪
⎪ σzy = σyz = μ +R ,
⎪
⎪ ∂y ∂y
⎪
⎪
⎪ ∂u ∂w
⎨ σxz = σzx = μ z + R z ,
∂x ∂x (10.2-1)
⎪
⎪ ∂wz ∂uz
⎪
⎪ Hzy = (K1 − K2 ) +R ,
⎪
⎪ ∂y ∂y
⎪
⎪
⎩ H = (K − K ) ∂wz + R ∂uz .
zx 1 2
∂x ∂x
Substituting expressions (10.2-1) into equations of motion of (10.1-3) yields
∂ 2 uz
μ∇2 uz + R∇2 wz = ρ ,
∂t2
∂ 2 wz
R∇2 uz + (K1 − K2 )∇2 wz = ρ . (10.2-2)
∂t2
If define displacement functions φ and ψ as[7]
uz = αφ − Rψ, wz = Rφ + αψ, (10.2-3)
then equations (10.2-2) reduce to the standard wave equations
1 ∂2φ 1 ∂ 2ψ
∇2 φ = , ∇2 ψ = , (10.2-4)
s21 ∂t2 s22 ∂t2
where
1 #
α= [μ − (K1 − K2 ) + (μ − (K1 − K2 ))2 + 4R2 ] (10.2-5)
2
and +
εj
sj = , j = 1, 2, (10.2-6)
ρ
1 #
ε1,2 = [μ + (K1 − K2 ) ± (μ − (K1 − K2 ))2 + 4R2 ],
2
sj can be understood as the speeds of wave propagation in anti-plane deformation
of the material. It is obvious that the wave speeds result from the phonon-phason
coupling. If there is no coupling, i.e., R → 0, then
+ &
μ (K1 − K2 )
s1 → , s2 → , (10.2-7)
ρ ρ
+ +
μ K1 − K2
where represents the speed of transverse wave of the phonon field and
ρ ρ
represents the speed of the pure phason elastic wave, requiring K1 − K2 > 0.
Substituting (10.2-3) into (10.2-1) the stresses can be expressed by φ and ψ
⎧
⎪ ∂φ ∂ψ
⎪
⎪ σyz = σzy = (αμ + R2 ) + R(α − μ) ,
⎪
⎪ ∂y ∂y
⎪
⎪
⎪
⎪ ∂φ ∂ψ
⎨ σxz = σzx = (αμ + R2 ) + R(α − μ) ,
∂x ∂x
(10.2-8)
⎪
⎪ Hzy = R(α + (K1 − K2 )) ∂φ 2 ∂ψ
⎪
⎪ + (α(K 1 − K 2 ) − R ) ,
⎪
⎪ ∂y ∂y
⎪
⎪
⎪
⎩ Hzx = R3 (α + (K1 − K2 )) ∂φ ∂ψ
+ (α(K1 − K2 ) − R2 ) .
∂x ∂x
Formulas (10.2-3) and (10.2-8) give the expressions for displacements and stresses in
terms of displacement functions φ and ψ, which satisfy the standard wave equations
(10.2-4) for elastodynamics of anti-plane elasticity of three-dimensional icosahedral
quasicrystals.
The above discussion is valid for anti-plane elasticity of three-dimensional cu-
bic quasicrystals or one-dimensional quasicrystals too. The difference between these
quasicrystals is only the material constants. If μ, K1 − K2 and R are replaced by
C44 , K44 and R44 (see Section 9.8) for cubic quasicrystals, or by C44 , K2 and R3 for
one-dimensional hexagonal quasicrystals with the Laue classes 6/mh and 6/mh mm
(see Sections 7.1 or 8.1) one can find the similar equations.
The solution of (10.2-4) can be done by using method for solving pure wave
equations in classical mathematical physics.
10.3 Moving screw dislocation in anti-plane elasticity

Assume a straight screw dislocation line parallel to the axis which moves along one
of axes, say, the x-axis in the plane perpendicular to the x-axis. For simplicity,
considering the dislocation moves with constant velocity V .
For the problem, a dislocation condition is assumed

||
duz = b3 , dwz = b⊥3, (10.3-1)
Γ Γ
||
i.e., we assume that the dislocation has the Burgers vector (0, 0, b3 , 0, b⊥
3 ), and Γ
denotes the Burgers circuit surrounding the core of the moving dislocation .
Starting now we denote the fixed coordinates as (x1 , x2 , t) and moving ones as
(x, y).
By introducing the Galilean transformation
x = x1 − V t, y = x2 , (10.3-2)

1 ∂2
wave equations (10.2-4) reduce to the Laplace equations i.e., ∇2 − 2 2 → ∇21 ,
s1 ∂t
2 2 2
1 ∂ ∂ ∂
∇2 − 2 2 → ∇22 , ∇2 = +
s2 ∂t ∂x21 ∂x22
∇21 φ = 0, ∇22 ψ = 0, (10.3-3)
10.3 Moving screw dislocation in anti-plane elasticity 195
where
∂2 ∂2 ∂2 ∂2
∇21 = 2
+ 2 , ∇21 = 2
+ 2, (10.3-4a)
∂x ∂y1 ∂x ∂y2
,
yj = βj y, βj = 1 − V 2 /s2j , j = 1, 2. (10.3-4b)
Let complex variable zj be
√
zj = x + iyj , i= −1, (10.3-5)
the solution of equations (10.3-3) is
φ = ImF1 (z1 ), ψ = ImF2 (z2 ), (10.3-6)
where F1 (z1 ) and F2 (z2 ) are analytic functions of z1 and z2 respectively, and notation
Im marks the imaginary part of a complex function.
The boundary condition (10.3-1) determines the analytic functions as[7]
A1 y1 A2 y2
φ(x, y1 ) = arctan , ψ(x, y1 ) = arctan (10.3-7a)
2π x 2π x
with constants
|| ||
αb3 + Rb⊥ 3 αb3 − Rb⊥ 3
A1 = , A 2 = , (10.3-7b)
α 2 + R2 α2 + R2
the displacement field is determined in the fixed coordinate system as follows

1 2 β1 y 2 β2 y ||
uz (x, y, t) = α arctan + R arctan b
2π(α2 + R2 ) x−Vt x−Vt 3
(10.3-8a)
β1 y β2 y
+ arctan − arctan αRb⊥ 3 ,
x−Vt x−Vt

1 2 β1 y 2 β2 y
wz (x, y, t) = R arctan + α 3 arctan b⊥
2π(α2 + R2 ) x−Vt x−Vt 3

β1 y β2 y ||
+ arctan − arctan αRb3 . (10.3-8b)
x−Vt x−Vt
The expressions for strains and stresses are omitted here due to limitation of space.
We give the evaluation on the energy of the moving dislocation. Denote energy
W per unit length on the moving dislocation which consists of the kinetic energy
Wk and potential energy Wp defined by the integrals
2 2
1 ∂uz ∂wz
Wk = ρ + dx1 dx2 ,
2 ∂t ∂t
Ω
(10.3-9)
1 ∂uz ∂wz
Wp = σij + Hij dx1 dx2 ,
2 ∂t ∂t
Ω
respectively, where the integration should be taken over a ring r0 < r < R0 , r0
denotes the size of the dislocation core, and R0 the size of so-called dislocation net
similar to those in conventional crystals. which are introduced in Section 7.1. In

general r0 ∼ 10−8 cm, and R0 ∼ 104 r0 . Substituting displacement formulas and
corresponding stress formulas into (10.3-9), we obtain
k k R0 k p R0
Wk = ln , Wp = ln (10.3-10)
4π r0 4π r0
with
ρV 2 (α2 + R2 )
A21 A22
kk = + ,
2 β1 β2

A21 2 2 2 1
kp = (μα + (K1 − K2 ))R + 2αR ) β1 +
2 β1
(10.3-11)
A22 2 2 2 1
+ (μR + (K1 − K2 )α − 2αR ) β2 +
2 β2
and A1 , A2 given by (10.3-7). Therefore the total energy is
kk + kp R0
W = ln . (10.3-12)
4π r0
It is concluded that when V → s2 , i.e., β2 → 0, this leads to the infinity of the
energy, and is invalid, and thus s2 is the limit of the velocity of a moving dislocation.
Additionally, if V << s2 , the total energy can be written in the following simple
form:
1 || 2 1 R0 1
W ≈ W0 + ρV 2 [(b3 )2 + (b⊥ 3) ] ln = W0 + m0 V 2 , (10.3-13)
2 4π r0 2
where W0 is the elastic energy per unit length of a rest screw dislocation, i.e.,
|| || 1 R0
W0 = [μ(b3 )2 + R(b⊥ 2 ⊥
3 ) + 2b3 b3 R] ln (10.3-14)
4π r0
and m0 , the so-called “apparent” mass of the dislocation per unit length in the case
considered
|| || ⊥ 1 R0
m0 = [μ(b3 )2 + R(b⊥ 2
3 ) + 2b3 b3 R] ln . (10.3-15)
4π r0
It is evident that if V = 0, the solution reduces to that of static dislocation, which
is given in Section 7.1.
# #
Furthermore, if b⊥3 = 0, R = 0, then ε1 = μ, β1 = 1 − V 2 /c22 , s1 = c2 = μ/ρ
is the speed of transverse wave of the conventional crystal, the above solution reduces
to
b β1 y
uz (x − V t, y) = arctan ,
2π x−Vt
b μβ1 (x − V t)
σyz = σzy = , (10.3-16)
2π (x − V t)2 + β12 y 2
b μβ1 y
σxz = σzx = − ,
2π (x − V t)2 + β12 y 2
10.4 Mode III moving Griffith crack in anti-plane elasticity 197

1 1 R0
W ≈ μb + ρV 2 b2
2
ln ,
2 4π r0
ρb2 R0
m0 = ln ,
4π r0
this is exactly identical to the well known Eshellby solution for crystals[17] .
The above discussion holds for anti-plane elasticity of three-dimensional cubic
or one-dimensional hexagonal quasicrystals, only the material constants μ, K1 − K2
and R should be replaced by C44 , K44 and R44 , or by C44 , K2 and R3 , respectively.
10.4 Mode III moving Griffith crack in anti-plane elasticity

As another application of above elastodynamic theory, we study a moving Griffith
crack of Mode III, which moves with constant speed V along x1 (see Fig. 10.4-1).
Here we also take the fixed coordinates (x1 , x2 , t) and the moving ones (x, y).
Fig.10.4-1 Moving Griffith crack of Mode III
In the moving coordinates the boundary conditions are

#
x2 + y 2 → ∞ : σij = 0, Hij = 0,
y = 0, |x| < a : σyz = −τ, Hyz = 0. (10.4-1)
Assume that the Laplace equations have the solution
φ(x1 , y1 ) = ReF1 (z1 ), ψ(x1 , y2 ) = ReF2 (z2 ), (10.4-2)
where F1 (z1 ) and F2 (z2 ) are any analytic functions of z1 and z2 , and notation Re
denotes the real part of a complex number.
Because the boundary conditions (10.4-1) are more complicated than those given
by (10.3-1), we must use conformal mapping
a
z1 , z2 = ω(ζ) = (ζ + ζ −1 ) (10.4-3)
2
to solve the problem at ζ(= ξ + iη)-plane .

After some calculations, we find the solution
iΔ1 iΔ2
F1 (z1 ) = F1 [ω(ζ)] = G1 (ζ) = ζ, F2 (z2 ) = F2 [ω(ζ)] = G2 (ζ) = ζ,
Δ Δ
(10.4-4)
in which
Δ = β1 β2 [(αμ + R2 )(α(K1 − K2 ) − R2 ) − R2 (α + (K1 − K2 ))(α − μ)],
Δ1 = τ αβ2 (α(K1 − K2 ) − R2 ), Δ2 = τ αβ1 R(α + (K1 − K2 )). (10.4-5)

Because there is the inverse mapping as
+ +
−1 z1 z1 2 −1 z2 z2 2
ζ = ω (z1 ) = − − 1 = ω (z2 ) = − − 1. (10.4-6)
a a a a
The manipulation afterward can also be done in z1 -plane/z2 -plane.
The corresponding stresses are expressed by
⎧
⎪ ∂ ∂
⎪
⎪ σyz = σzy = (αμ + R2 )β1 ReF1 (z1 ) + R(α − μ)β2 ReF2 (z2 ),
⎪
⎪ ∂y1 ∂y2
⎪
⎪
⎪
⎨ σxz = σzx = (αμ + R2 ) ∂ ReF1 (z1 ) + R3 (α − μ)β2 ∂ ReF2 (z2 ),
∂x ∂x
⎪
⎪ Hzy = R3 (α + (K1 − K2 ))β1
∂
ReF1 (z1 ) + (α(K1 − K2 ) − R2 )
∂
⎪
⎪ ReF2 (z2 ),
⎪
⎪ ∂y 1 ∂y 2
⎪
⎪ ∂ ∂
⎩ Hzx = R3 (α + (K1 − K2 ))β1 ReF1 (z1 ) + (α(K1 − K2 ) − R2 ) ReF2 (z2 ).
∂x ∂x
(10.4-7)
Substituting (10.4-6) into (10.4-4) then into (10.4-7), the stresses can be evaluated
in explicit form, e.g.
σyz = σzy
τ 2 2 d 1 1
= − (αμ + R )β1 β2 (α(K1 − K2 ) − R ) 1 − cos θ − θ1 − θ2
Δ (d1 d2 )1/2 2 2

τ D 1 1
+ β1 β2 R2 (α + (K1 − K2 ))(α − μ) 1 − cos Θ − Θ1 − Θ 2
Δ (D1 D2 )1/2 2 2
(10.4-8)
with
⎧ # # #
⎪
⎪ d = x2 + y12 , d1 = (x − a)2 + y12 , d2 = (x + a)2 + y12 ,
⎪
⎪
⎪
⎪ # # #
⎪
⎪ D = x2 + y22 , D1 = (x − a)2 + y22 , D2 = (x + a)2 + y22 ,
⎪
⎨
y
⎪ 1 y1 y1
⎪
⎪ θ = arctan , θ1 = arctan , θ2 = arctan ,
⎪
⎪ x x−a x+a
⎪
⎪ y
⎪
⎪ y2 y2
⎩ Θ = arctan 2 , Θ1 = arctan , Θ2 = arctan .
x x−a x+a
(10.4-9)
10.5 Elasto-/hydro-dynamics of quasicrystals and approximate analytic solution... 199
It is easy to prove that (10.4-9) satisfies the relevant boundary conditions, and
is the exact solution.
Similarly, σxz = σzx , Hzx and Hzy can also be expressed explicitly.
From (10.4-8), as y = 0, it yields
xτ
√ − τ, |x| > a,
σyz (x, 0) = x − a2
2 (10.4-10)
−τ, |x| < a
The stress presents singularity of order (x − a)−1/2 as x → a.
The stress intensity factor for Mode III for phonon field is
||
# √
KI = lim+ π(x − a)σyz (x, 0) = πaτ. (10.4-11)
x→a
This is identical to the classical Yoffe solution[18] , there the stress intensity factor is
also independent of crack moving speed V .
Now we calculate energy of the moving crack, which is defined by
a
W=2 [σzy (x, 0) ⊕ Hzy (x, 0)][uz (x, 0) ⊕ wz (x, 0)]dx
0
1 1
= (Δ1 α−Δ2 R)τ πa = [αβ2 (α(K1 −K2 ) − R2 )−β1 R2 (α + (K1 − K2 ))]πa2 τ.
Δ Δ
(10.4-12)
The crack energy release rate is
1 ∂W 1
G= = [αβ2 (α(K1 −K2 )−R2 ) − β1 R2 (α + (K1 −K2 )](KI )2 . (10.4-13)
2 ∂a 2Δ
The above discussion, results and conclusions hold too for anti-plane elasticity
of three-dimensional cubic or one-dimensional hexagonal quasicrystals, only the ma-
terial constants μ, K1 − K2 and R need to be replaced by C44 , K44 and R44 or by
C44 , K2 and R3 , respectively. The results are given by the author in the Chinese
edition of this book in 1999.
10.5 Elasto-/hydro-dynamics of quasicrystals and approxi-

mate analytic solution for moving screw dislocation in
anti-plane elasticity
In the previous sections the formulation following Bak’s argument has been devel-
oped and certain results are obtained. The character of the work lies in the equations
of motion being taken as
∂ 2 ui ∂σij
ρ 2 = ,
∂t ∂xj
∂ 2 wi ∂Hij
ρ 2 = ,
∂t ∂xj
which are different from those based on the argument of Lubensky et al. Meanwhile
Rochal and Lorman[15] did not completely agree to the above equations even though
they followed the Bak’s argument. They suggested the density ρ in the second equa-
tion listed above should be replaced by ρef , where ρef represents the generalized
effective phason density, but the meaning of the quantity is not very clear and its
measurement is difficult. Fan et al[13] , Rochal and Lorman[16] have respectively sug-
gested a new version on elasto-/hydro-dynamics of quasicrystals, this compromises
different models, e.g. those proposed by Ref. [1] and Ref. [5]. At present the experi-
mental data are not so sufficient to verify each model individually. It is the anthor’s
opinion that it shall be highly desirable that all researchers collaborate the common
features of different models and find a simplest version for elasto-/hydro-dynamics
of quasicrystals, then carry out systematic, theoretical, numerical and experimental
work; that would be beneficial for promoting the study in the field.
Fan et al[13] presented results incorporating the arguments of both Bak and
Lubensky et al. They suggested the equations of motion under linear and small
deformation be written as ⎧
⎪
⎪ ∂ 2 ui ∂σij
⎨ ρ 2 = ,
∂t ∂xj (10.5-1)
⎪
⎪ ∂wi ∂Hij
⎩ κ = ,
∂t ∂xj
It is obvious that the first equation of (10.5-1) is equation of conventional elastody-
namics, in which ρ is the mass density as mentioned in the previous section, while
the second one is diffusion equation, where κ = 1/Γw , in which Γw the kinematic
coefficient of phason field of the material defined by Lubensky et al[1] . It can be seen
that the second equation of (10.5-1) is a linearized result of hydrodynamics of qua-
sicrystals of Lubensky et al. The dynamic equations (10.5-1) should be named as the
elasto-/hydro-dynamic equations for quasicrystals which are identical to equations
(6) of Ref. [16]. This treatment is believed to be physically sound and reconciles the
contradiction between the arguments of Bak and Lubensky et al; it also reconciles
the contradiction between Refs. [5 ∼ 12] and Ref. [15], though they all follow the
framework of Bak’s argument.
As an example we discuss herein an approximate analytic solution of moving
dislocation in anti-plane problem of three-dimensional icosahedral or cubic or one-
dimensional hexagonal quasicrystal, see Fan et al[13] . The stress-strain relations in
this case are the same as (9.2-3), the difference lies in only the equations of motion
which are given by (10.5.1), such that for anti-plane which become
⎧ 2
⎪
⎨ μ∇2 uz + R∇2 wz = ρ ∂ uz ,
∂t2 (10.5-2)
⎪
⎩ R∇ uz + (K1 − K2 )∇2 wz = κ ∂wz .
2
∂t
Equations (10.5-2) are wave-diffusion mixed type equations which are different from
(10.2-2), so the solution will be difficult to obtain.
The following boundary conditions for screw dislocation moving along direction
Ox with speed V are considered as
⎧ #
2 2
⎨ x + y → ∞ : σij (x, y, t) → 0,
⎪ Hij (x, y, t) → 0,
(10.5-3)
⎪
⎩
duz = b3 , dwz = b⊥3.
Γ Γ
Solving boundary value problem (10.5-2), (10.5-3) is more complicated than solv-
ing either a pure wave propagation problem or a pure diffusion problem. We take
the perturbation method along with variational method. Based on the physical
consideration R/μ should be small and it can be taken as the perturbation param-
eter. Firstly, we calculate the zeroth-order perturbation solution. The governing
equations (10.5-2) reduce to
⎧
⎪
⎪ ∂ 2 u0
⎨ ρ 2 − μ∇2 u0 = 0,
∂t
(10.5-4)
⎪
⎪ ∂w 0
⎩ κ − (K1 − K2 )∇2 w 0 = 0,
∂t
where u0 , w0 denote the zeroth-order perturbation solution of the problem.

A coordinate substitution
x = x1 − V t (10.5-5)
is used, here (x1 , y, t) denotes fixed coordinate system and (x, y) the moving coor-
dinate system, and
∂ ∂ ∂2 ∂2
= −V , 2
= V2 2. (10.5-6)
∂t ∂x ∂t ∂x
Putting (10.5-6) into (10.5-4) yields

⎧ 2 0
⎪ ∂ u 1 ∂ 2 u0
⎪
⎨ + 2 = 0,
∂x 2 β1 ∂y 2
(10.5-7)
⎪
⎪ ∂ 2 w0 2 ∂w
0
∂ 2 w0
⎩ + β 2 + = 0,
∂x2 ∂x ∂y 2
where
V2 Vκ V
β12 = 1 − , β22 = = , (10.5-8)
c2 K1 − K2 (K1 − K2 )Γw
#
in which c = μ/ρ for icosahedral quasicrystals is the velocity of transverse wave
of phonon field.
Because the equations in (10.5-7) are decoupled, and considering dislocation
conditions in (10.5-3) we have the following two separate dislocation problems such
as
⎧ 2 0
⎪
⎪ ∂ u 1 ∂ 2 u0
⎨ 2
+ 2 = 0,
∂x β1 ∂y2
(10.5-9a)
⎪
⎪ 0
⎩ du = b3 ,
Γ
⎧ 2 0
⎪
⎪ ∂ w ∂w0 ∂ 2 w0
⎨ 2
+ β22 + = 0,
∂x ∂x ∂y 2
(10.5-9b)
⎪
⎪
⎩ dw0 = b⊥3.
Γ
The solution of problem (10.5-9a) is known, i.e.,

b3 β1 y
u0 = arctan . (10.5-10)
2π x
To solve problem (10.5-9b) we put
w0 = w1 + w2 , (10.5-11)
in which w1 satisfies
⎧ 2
⎪
⎪ ∂ w1 ∂ 2 w1
⎨ + = 0,
∂x2 ∂y2
(10.5-12)
⎪
⎪
⎩ dw1 = b⊥
3,
Γ
so it has the solution

b⊥
3 y
w1 = arctan (10.5-13)
2π x
and w2 satisfies
⎧
⎪
⎪ ∂ 2 w2 2 ∂w2 ∂ 2 w2 β2 b⊥
3y
⎨ 2
+ β 2 + 2
= 2 + y2 )
,
∂x ∂x ∂y 2π(x (10.5-14)
⎪
⎪
⎩ dw2 = 0.
Γ
The equation in problem (10.5-14) is more complex than that in (10.5-9a) and
(10.5-12), so its analytic solution is very difficult to obtain. We can do only ap-
proximate solution by variational principle (refer to Section 13.2 in Chapter 13).
Omitting the detail the approximate solution is
b⊥
3 β22 x2 y
w2 = − . (10.5-15)
2π (x2 + y 2 )(1 + β22 x)
From (10.5-11), (10.5-13) and (10.5-15), the zeroth-order approximate solution

is
b⊥ y β22 x2 y
w = 3
0
arctan − 2 . (10.5-16)
2π x (x + y 2 )(1 + β22 x)
Then we calculate the first-order solution, the governing equations (10.5-2) are
⎧
⎪
⎪ ∂ 2 u1
⎨ ρ 2 − μ∇2 u1 = R∇2 w 0 ,
∂t (10.5-17)
⎪
⎪ ∂w 1
⎩ κ − (K1 − K2 )∇2 w 1 = R∇2 u0 ,
∂t
where u1 , w1 are denoted the first-order perturbation solution of the problem.

According to the boundary condition (10.5-3) and the governing equation
(10.5-17), one finds the first order pertubation solution by the variational method
that
b⊥
3 R 2β22 x2 y(3x2 − y 2 + 6β22 x3 + 2β24 x4 )
u1 = − ,
2π (ρV 2 − μ) (x2 + y 2 )2 (1 + β22 x)3

(10.5-18)
1 b3 R 2(β13 − β1 )x3 y
w =− .
2π(K1 − K2 ) (x2 + β12 y 2 )2 (1 + β22 x)
The approximate solution up to the first pertubation of the moving dislocation

problem is
⎧
⎪ uz (x1 , y, t) ≈ u0z (x1 , y, t)+u1z (x1 , y, t)
⎪
⎪
⎪
⎪
⎪
⎪
β1 y b⊥ 2β22 x2 y(3x2 −y 2 + 6β22 x3 +2β24 x4 )
⎪
⎪ b3
− 3
R
⎨ = arctan ,
2π x 2π (ρV −μ)2 (x2 +y 2 )2 (1+β22 x)3
0 1
⎪
⎪ wz (x1 , y, t) ≈ wz (x1 , y, t)+wz (x1 , y, t)
⎪
⎪
⎪
⎪ b⊥
⎪
⎪ y β22 x2 y

b3 R 2(β13 −β1 )x3 y
⎪
⎩ = 3 arctan − − ,
2π 2
x (1+β2 x)(x +y ) 2 2 2π(K1 − K2 ) (x +β12 y 2 )2 (1+β22 x)
2
(10.5-19)
where we recall x = x1 − V t.
So the corresponding strains are as follows:
⎧
⎪
⎪ 1 b3 β1 x b⊥ β22 R
⎪
⎪ε = ε = + 3
⎪
⎪
yz zy 2 + β 2 y2 π ρV 2 − μ
⎪
⎪
2 2π x 1
⎪
⎪ 6 − 12x4 y 2 + y 6 + 6(x7 − 3x5 y 2 )β 2 + 2(x8 − 3x6 y 2 )β 4
⎪
⎪ 3x
⎪
⎪ × 2 2
,
⎪
⎪ (x2 + y 2 )3 (1 + β22 x)3
⎪
⎪
⎪
⎪
⎪
⎪
b⊥ β22 R
⎪
⎪ 1 b β1 y
⎪
⎪εzx = εxz = − 3 2 + 3
⎪
⎪ 2 2π x + β1 y 2 2 π ρV 2 − μ
⎪
⎪
⎨ 14x y −2xy +(−3x y+38x4 y 3+x2 y 5 )β22 −8(x7 y−3x5 y 3 )β24−2(x8 y−3x6 y3 )β26
3 3 5 6
× ,
⎪
⎪
2 2 3
(x +y ) (1+β2 x) 2 4
⎪
⎪
⎪
⎪
⎪
⎪w b⊥ 2 2
3 x(x + y + 2β2 xy )
2 2 b3 R x3 (x2 − 3β12 y 2 )(β13 − β1 )
⎪
⎪ = − ,
⎪
⎪
zy
2π (x 2 + y 2 )2 (1 + β 2 x) π K − K 2 2
2 (1 + β2 x)(x + β1 y )
2 2 3
⎪
⎪ 2 1
⎪
⎪
⎪
⎪ b⊥ y(2β22 x3 + y 2 + 4β22 xy 2 + x2 (1 + 2β24 y 2 ))
⎪
⎪wzx = − 3
⎪
⎪ (x2 + y 2 )2 (1 + β22 x)2
⎪
⎪
2π
⎪
⎪
⎪
⎪
x2 y(x2 + 2β22 x3 − 2β12 β22 xy 2 − 3β12 y 2 )(β13 − β1 )
⎪
⎪
b R
⎩ + 3
π K1 − K2 (1 + β22 x)2 (x2 + β12 y 2 )3
(10.5-20)
and the stresses can be expressed by
⎧
⎪
⎪ σyz = σzy
⎪
⎪
⎪
⎪
⎪
⎪ b⊥ 2β22 Rμ 3x6 −12x4 y 2 +y 6 +6(x7 −3x5 y 2 )β22 +2(x8 −3x6 y 2 )β24
⎪
⎪ 3
⎪
⎪ =
⎪
⎪ 2π ρV 2 −μ (x2 +y 2 )3 (1+β22 x)3
⎪
⎪
⎪
⎪
⎪
⎪
2 2
Rx(x +y +2β2 xy ) 2 2 b μβ1 x 2R2 x3 (x2 −3β12 y 2 )(β13 −β1 )
⎪
⎪ + + 3 − ,
⎪
⎪ 2 2 2
(x +y ) (1+β2 x) 2 2
2π x +β1 y 2 2 K1−K2 (1+β2 x)(x +β1 6y )
2 2 2 2 3
⎪
⎪
⎪
⎪
⎪
⎪
⊥
b R 2β2 μ 2
⎪
⎪ σzx = σxz = 3
⎪
⎪ 2π ρV 2 −μ
⎪
⎪
⎪
⎪
⎪
⎪ 14x3 y 3−2xy 5+(−3x6 y+38x4 y 3 +x2 y 5 )β22−8(x7 y−3x5 y 3 )β24−2(x8 y−3x6 y 3 )β26 )
⎪
⎪
⎪
⎪ ×
⎪
⎪ (x2 +y 2 )3 (1+β22 x)4
⎪
⎪
⎪
⎪ y(2β2 x +y +4β2 xy +x (1+2β24 y 2 ))
2 3 2 2 2 2
⎪
⎪
⎪
⎪ −
⎪
⎪ (x2 +y 2 )2 (1+β22 x)2
⎪
⎪
⎪
⎨
b 2R2 x2 y(x2 +2β22 x3 −2β12 β22 xy 2 −3β12 y 2 )(β13 −β1 ) μβ1 y
+ 3 − ,
⎪
⎪ 2π K1 −K2 2 2
(1+β2 x)2 (x2 +β1 y 2 )3 2
x2 +β1 y 2
⎪
⎪ ⊥ 2β 2 R2 3x6 −12x4 y 2 +y 6 +6(x7 −3x5 y 2 )β 2 +2(x8 −3x6 y 2 )β 4
⎪
⎪ b
⎪
⎪ 3 2 2 2
⎪Hzy = 2π ρV 2 −μ
⎪ (x2 +y 2 )3 (1+β22 x)3
⎪
⎪
⎪
⎪
⎪
⎪ x(x2 +y 2 +2β22 xy2 ) b3 R 2x3 (x2−3β12 y 2 )(β13 −β1 )
⎪
⎪ β1 x
⎪
⎪ +(K1 −K2 ) 2 + − .
⎪
⎪
2 2
(x +y ) (1+β2 x) 2 2
2π x +β1 y 2 2 2 2
(1+β2 x)(x +β1 y ) 2 2 3
⎪
⎪
⎪
⎪ b⊥ 2β22 R
⎪
⎪
⎪
⎪Hzx = 3
⎪
⎪ 2π ρV 2 −μ
⎪
⎪
⎪
⎪ 14x3 y 3−2xy 5+(−3x6 y+38x4 y 3+x2 y 5 )β22−8(x7 y−3x5 y 3 )β24−2(x8 y−3x6 y 3 )β26
⎪
⎪
⎪
⎪ ×
⎪
⎪ (x2+y 2 )3 (1+β22 x)4
⎪
⎪
⎪
⎪ y(2β2 x +y +4β2 xy +x2 (1+2β24 y 2 ))
2 3 2 2 2
⎪
⎪
⎪
⎪ −(K 1−K 2 )
⎪
⎪ (x2 +y 2 )2 (1+β22 x)2
⎪
⎪
⎪
⎪ 2
⎪
⎪ b R 2x y(x2 +2β22 x3 −2β12 β22 xy2 −3β12 y 2 )(β13 −β1 ) β1 y
⎪
⎩ + 3 − .
2π 2 2
(1+β2 x)2 (x2 +β1 y 2 )3 2
x2 +β1 y2
(10.5-21)
For cubic or hexagonal quasicrystals, the solutions are exactly similar, with
μ, (K1 − K2 ) and R replaced by C44 , K44 and R44 or by C44 , K2 and R3 respec-
tively in the above equations.
It is easy to observe that the solution contains the contributions coming from
three parts of wave propagation, diffusion and their interaction. The solution on
phonon field is dominated by wave propagation, and the solution on phason field is
dominated by diffusion, so the result is quite different from the dynamic solution of
the dislocation based on the Bak’s argument, which was discussed in Section 10.3
already. The present solution can reduce to that in static dislocation as V = 0,
which can also reduce to the solution of crystals, this demonstrates the correctness
of the work. The solution reveals that the phason field resists motion of dislocation,
this is the reason of difficulty of plastic deformation of quasicrystalline materials.
The illustrations of the solution and comparison with other solutions for the case
#
V = 0.1c, c = μ/ρ are depicted in Figs.10.5-1∼10.5-4, in which the solution given
by Ref. [7] has been introduced in Section 10.3. In the numerical computation we
take the data of material constants as

b⊥
3 = 0.8b3

Fig.10.5-1 Variation of uz /b3 at x = 0.01 mm versus y

Fig.10.5-2 Variation of uz /b3 at y = 0.01 mm versus x

Fig.10.5-3 Variation wz /b3 at x = 0.01 mm versus y

Fig.10.5-4 Variation wz /b3 at y = 0.01 mm versus x
R
ρ = 4.1kg/cm3 , μ = 70GPa, K1 = 74.2, K2 = −37MPa, = 0.004,
μ
1
Γw = = 4.8 × 10−19 m3 · s/kg = 4.8 × 10−10 cm3 · μs/g
κ
#
given in Tables 9.1-1∼9.1-3. We recall x = x1 − V t, c = μ/ρ.
10.6 Elasto-/hydro-dynamics and solutions of two-dimensio-

nal decagonal quasicrystals
In this section we would like to give a detailed description on the solution of two-
dimensional quasicrystals based on the elasto-/hydro-dynamics formulation[14] .
The equations of deformation geometry and the generalized Hooke’s law are the
same as before which will not be repeated here.
By using the dynamic equations (10.5-1), i.e., the so-called elasto-/hydro-dynamic
equations for quasicrystals, we can derive the mathematical formalism for the new
10.6 Elasto-/hydro-dynamics and solutions of two-dimensional decagonal... 207
dynamics of two-dimensional quasicrystals. As an application of the formulation,

some dynamic crack solutions are given in this section.
10.6.1 The mathematical formulation of dynamic crack problems of decag-
onal quasicrystals
Among 200 quasicrystals observed to date, there are 70 two-dimensional decagonal
quasicrystals; so this kind of solid phases plays an important role in the material.
For simplicity, herein only point group 10mm two-dimensional decagonal quasicrystal
will be considered. We denote the periodic direction as the z-axis and the quasiperi-
odic plane as the xy-plane. Assume that a crack in the solid along the periodic
direction, i.e., the z-axis. It is obvious that elastic field induced by a uniform tensile
stress at upper and lower surfaces of the specimen is independent of z, so ∂/∂z = 0.
In this case, the stress-strain relations are reduced to
⎧
⎪
⎪
⎪
⎪
⎪ σ = L(εxx + εyy ) + 2M εyy − R(wxx + wyy ),
⎪
⎪ yy
⎪
⎪
⎨ σxy = σyx = 2M εxy + R(wyx − wxy ),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ Hxy = K1 wxy − K2 wyx − 2Rεxy ,
⎪
⎩ H = K w − K w + 2Rε ,
yx 1 yx 2 xy xy
where L = C12 , M = (C11 − C12 )/2 are the phonon elastic constants, K1 and K2
the phason elastic constants, R phonon-phason coupling elastic constant.
Substituting (10.6-1) into (10.5-1), we obtain the equations of motion of decago-
nal quasicrystals as follows:
⎧ 2 2 2 2
2
⎪ ∂ ux 2 ∂ ux 2 ∂ uy 2 ∂ ux ∂ wx ∂ 2 wy ∂ 2 wx
⎪
⎪ = c1 2
+ (c1 − c2 ) + c + c3 2
+2 − ,
⎪
⎪ ∂t2 ∂x2 ∂x∂y 2
∂y 2 ∂x2 ∂x∂y ∂y 2
⎪
⎪ 2
⎪ ∂ 2 uy
⎪ 2 2 2
∂ 2 wx ∂ 2 wy
⎪
⎪ 2 ∂ uy 2 2 ∂ ux 2 ∂ uy 2 ∂ wy
⎨ = c 2 + (c1 − c 2 ) + c 1 + c 3 − 2 − ,
∂t2 ∂x2 ∂x∂y ∂y 2 ∂x2 ∂x∂y ∂y 2
2
⎪
⎪ ∂wx ∂ wx ∂ 2 wx ∂ 2 ux ∂ 2 uy ∂ 2 ux
⎪
⎪ = d 2
+ + d 2
− 2 − ,
⎪
⎪ ∂t 1
∂x2 ∂y2 2
∂x2 ∂x∂y ∂y 2
⎪
⎪
⎪
⎪ ∂wy ∂ 2 wy ∂ 2 wy ∂ 2 uy ∂ 2 ux ∂ 2 uy
⎪
⎩ = d21 + + d2
+ 2 − ,
2
∂t ∂x2 ∂y 2 ∂x2 ∂x∂y ∂y 2
(10.6-2)
in which
& & & + + +
L + 2M M R K1 R K2
c1 = , c2 = , c3 = , d1 = , d2 = , d3 = ,
ρ ρ ρ κ κ κ
(10.6-3)
note that constants c1 , c2 and c3 have the meaning of elastic wave speeds, while
d21 , d22 and d23 do not represent wave speeds, they are diffusive coefficients.
The decagonal quasicrystal with a crack is shown in Fig.10.6-1. It is a rectangu-
lar specimen with a central crack of length 2a(t) subjected to a dynamic or static
tensile stress at its ends ED and F C, in which a(t)represents the crack length as
a function of time, and for dynamic initiation of crack growth, the crack is stable,
so a(t) = a0 = constant, for fast crack propagation, a(t) varies with time. At first
we consider dynamic initiation of crack growth, then study crack fast propagation.
Fig.10.6-1 The specimen with a central crack
Due to the symmetry of the specimen only the upper right quarter is considered.
Referring to the upper right part and considering a fix grips case, the following
boundary conditions should be satisfied:
⎧
⎪
⎪ ux = 0, σyx = 0, wx = 0, Hyx = 0, on x = 0 for 0 y H,
⎪
⎪
⎪
⎨ σxx = 0, σyx = 0, Hxx = 0, Hyx = 0, on x = L for 0 y H,
⎪
σyy = p(t), σxy = 0, Hyy = 0, Hxy = 0, on y = H for 0 x L, (10.6-4)
⎪
⎪
⎪
⎪ σyy = 0, σxy = 0, Hyy = 0, Hxy = 0, on y = 0 for 0 x a(t),
⎪
⎪
⎩
uy = 0, σxy = 0, wy = 0, Hxy = 0, on y = 0 for a(t) x L,
where p(t) = p0 f (t) is a dynamic load if f (t) varies with time, otherwise it is a static
load (i.e., if f (t) = const), and p0 = const with the stress dimension.
The initial conditions
⎧ are
⎪
⎪
ux (x, y, t)|t=0 = 0, uy (x, y, t)|t=0 = 0,
⎨
wx (x, y, t)|t=0 = 0, wy (x, y, t)|t=0 = 0, (10.6-5)
⎪
⎪ ∂ux (x, y, t) ∂uy (x, y, t)
⎩ |t=0 = 0, |t=0 = 0.
∂t ∂t
For implementation of finite difference all field variables in both governing equa-
tions (10.6-2) and boundary-initial conditions (10.6-4), (10.6-5) must be expressed
by displacements and their derivatives. This can be done through the constitutive
equations (10.6-1). The detail of the finite difference scheme is given in Appendix
of this chapter.
For the related parameters in this section, the experimentally determined mass
density for decagonal Al-Ni-Co quasicrystal: ρ = 4.186 × 10−3 g · mm−3 is used and
elastic phonon moduli are taken as C11 = 2.3433, C12 = 0.5741(1012 dyn/cm2 =
102 GPa) which are obtained by resonant ultrasound spectroscopy[19] , we have also
chosen phason elastic constants K1 = 1.22 and K2 = 0.24(1012 dyn/cm2 = 102 GPa)
estimated by Monto-Carlo simulation[20] and Γw = 1/κ = 4.8 × 10−19 m3 · s/kg =
4.8×10−10 cm3 ·μs/g[21] .The coupling constant R has been measured for some special
cases recently, see Chapter 6 and Chapter 9 respectively. In computation we take
R/M = 0.01 for coupling case corresponding to quasicrystals, and R/M = 0 for
decoupled case which corresponds to crystals.
10.6.2 Examination on the physical model

In order to verify the correctness of the suggested model and the numerical simu-
lation, we first explore the specimen without a crack. We know that there are the
fundamental solutions characterizing time variation natures based on wave propa-
gation of phonon field and on diffusion according to the mathematical physics
⎧
⎨ u ∼ eiω(t−x/c) ,
1 2 (10.6-6)
⎩ w∼ √ e−(x−x0 ) /Γw (t−t0 ) ,
t − t0
where ω is a frequency and c is a speed of the wave, t is the time and t0 is a special
value of t, x is a distance, x0 is a special value of x, and Γw is the kinetic coefficient
of phason defined previously.
Comparison results are shown in Fig.10.6-2 (a∼c), in which the solid line rep-
resents the numerical solution of quasicrystals and the dotted line represents fun-
damental solution of equation (10.6-6). From Figs. 10.6-2(a) and (b) we can see
that both displacement components of phonon field are in excellent agreement to
the fundamental solutions. However, there are some differences because the phonon
field is influenced by phason field and the phonon-phason coupling effect. From Fig.
10.6-2(c), in the phason field we find that the phason mode presents diffusive nature
in the overall tendency, but because of influence of the phonon and phonon-phason
coupling, it also has some character of fluctuation. So the model describes the dy-
namic behaviour of phonon field and phason field. This also shows the mathematical
modelling of the present work is valid.
(a) Displacement component of phonon field ux versus time
(b) Displacement component of phonon field uy versus time
(c) Displacement component of phason field wx versus time

Fig.10.6-2 Solutions without crack and comparison to fundamental solutions
10.6.3 Testing the scheme and the computer program

10.6.3.1 Stability of the scheme
The stability of the scheme is the core problem of finite difference method which
depends upon the choice of parameter α = c1 τ /h, the ratio between time step and
space step substantively. The choice is related to the ratio c1 /c2 , i.e., the ratio
between speeds of elastic longitudinal and transverse waves of the phonon field. To
determine the upper bound for the ratio to guarantee the stability, according to our
computational practice and experiences of computations for conventional materials,
α = 0.8 is chosen in all cases. The computational results are stable.
10.6.3.2 Accuracy test
The stability is only a necessary condition for successful computation. We must

check the accuracy of the numerical solution. This can be realized through some
comparison with some well known classical solutions (analytic as well as numerical
solutions). For this purpose the material constants in the computation are chosen
3
as c1 = 7.34, c2 = 3.92 (mm/μs) and ρ = 5 × 103 kg/m , p0 = 1MPa which are the
same as those in Refs. [22∼24] (but are different from those listed in Subsection
10.6.1).At first the comparison to the classical exact analytic solution is carried out,
in this case we put wx = wy = 0 (i.e., K1 = K2 = R = 0) for the numerical
solution. The comparison has been done with the key physical quantity—dynamic
stress intensity factor, which is defined by
#
KI (t) = lim π(x − a0 )σyy (x, 0, t). (10.6-7)
x→a+
0
The normalized dynamic stress intensity factor can be denoted as KI (t)/KIstatic , in

which KIstatic is the corresponding static stress intensity factor, whose value here is
√
taken as πa0 p0 (the reason for this refer to Subsection 10.6.3.3). For the dynamic
initiation of crack growth in classical fracture dynamics there is the only exact
analytic solution—the Maue’s solution[22] , but the configuration of whose specimen
is quite different from that of our specimen. Maue studied a semi-infinite crack in
an infinite body, and subjected to a Heaviside impact loading at the crack surface.
While our specimen is a finite size rectangular plate with a central crack, and the
applied stress is around the external boundary of the specimen. Generally, the
Maue model cannot describe the interaction between wave and external boundary.
However, consider a very short time interval, i.e., during the period between the
stress wave from the external boundary arriving at the crack tip (this time is denoted
by t1 ) and before the reflecting by external boundary stress wave emanating from
the crack tip in the finite size specimen (the time is marked as t2 ). During this
special short time interval our specimen can be seen as an “infinite specimen”.
The comparison given by Fig.10.6-3 shows the numerical results are in excellent
agreement with those of Maue’s solution within the short interval in which the
solution is valid.
Our solution for the case of wx = wy = 0 is also compared with numerical
solutions of conventional crystals, e.g. Murti’s solution[23] and Chen’s solutions[24] ,
which are also shown in Fig.10.6-3, it is evident, our solution is in good agreement
with available solutions.
Fig.10.6-3 Comparison of our solution with analytic solution and other

numerical solution for crystalline materials
10.6.3.3 Influence of mesh size (space step)
The mesh size or the space step of the algorithm can influence the computational
accuracy too. To check the accuracy of the algorithm we take different space steps
shown in Table 10.6-1, which indicates if h = a0 /40 the accuracy is good enough.
The check is carried out through static solution, because the static crack problem
in infinite body of decagonal quasicrystals has exact solution, see Chapter 8, the
normalized static intensity factor is equal to unit. In the static case, there is no
wave propagation effect, L/a0 3, H/a0 3 the effect of boundary to solution is
very weak, and for our present specimen L/a0 4, H/a0 8, which may be seen as
an infinite specimen, so the normalized static stress intensity factor is approximately
but with highly precise equal to unit. The table shows that the algorithm is with a
quite highly accuracy when h = a0 /40.
Table 10.6-1 The normalized static S.I.F. of quasicrystals for different space
steps
h a0 /10 a0 /15 a0 /20 a0 /30 a0 /40
K̄ 0.9259 0.94829 0.96229 0.97723 0.99516
Errors 7.410% 5.171% 3.771% 2.277% 0.484%
10.6.4 Results of dynamic initiation of crack growth

The dynamic crack problem indicates two “phases”in the process: the dynamic
initiation of crack growth and fast crack propagation. In the phase of dynamic
initiation of crack growth, the length of the crack is constant, assuming a(t) = a0 .
The specimen with stationary crack subjected to a rapidly varying applied load
p(t) = p0 f (t), where p0 is a constant with stress dimension and f (t) is taken as
the Heaviside function. It is well known the coupling effect between phonon and
phason is very important, which reveals the distinctive physical properties including
mechanical properties, and makes quasicrystals different from the periodic crystals.
So studying the coupling effect is significant.
The dynamic stress intensity factor KI (t) for quasicrystals has the same defi-
nition given by (10.6-7), whose numerical results are plotted in Fig.10.6-4, where
√
the normalized dynamics stress intensity factor K I (t)/ πa0 p0 is used. There are
two curves in the figure, one represents quasicrystal, i.e.,R/M = 0.01, the other
describes periodic crystals corresponding to R/M = 0, the two curves of the fig-
ure is apparently different, though they are similar to some extent. Because of the
phonon-phason coupling effect, the mechanical properties of the quasicrystals are
obviously different from the crystals. Thus, the coupling effect plays an important
role.
In Fig.10.6-4, t0 represents the time that the wave from the external boundary
propagates to the crack surface, in which t0 = 2.6735μs. So the velocity of the wave
propagation is ν0 = H/t0 = 7.4807km/s, which is just equal to the longitudinal
#
wave speed c1 = (L + 2M )/ρ. This indicates that for the complex system of
coupled motion of wave propagation and diffusion the phonon wave propagation
plays dominating role.
Fig.10.6-4 Normalized dynamics stress intensity factor (DSIF) versus time

There are many oscillations in the figure, especially the stress intensity factor.
These oscillations characterize the reflection and diffraction between waves coming
from the crack surface and the specimen boundary surfaces. The oscillations are
influenced by the material constants and specimen geometry including the shape
and size.
10.6.5 Results of the fast crack propagation
In this section, we focus on the discussion for the “phase” of fast crack propagation.
To explore the inertia effect caused by the fast crack propagation, the specimen
is designed under the action of constant load P (t) = p rather than time-varying
load, but the crack grows with high speed in this case. The problem for fast crack
propagation is a nonlinear problem, because one part of the boundaries—crack is
with unknown length beforehand. For this moving boundary problem, we must give
additional condition for determining solution. That is, we must give a criterion
checking crack propagation or crack arrest at the growing crack tip. This criterion
can be imposed in different ways, e.g., the critical stress criterion or critical energy
criterion. The stress criterion is used in this paper:σyy < σc represents crack arrest,
σyy = σc represents critical state and σyy > σc represents crack propagation. Here
we take σc = 450M P a for decagonal Al-Ni-Co quasicrystals, which was obtained by
referring measured value by Meng et al for decagonal Al-Cu-Co quasicrystals, refer
to Ref.[2] in Chapter 8, the modification by referring the hardness of alloys Al-Ni-Co
and Al-Cu-Co, and the hardness on decagonal Al-Ni-Co can be found in paper given
by Takeuchi et al[30] . The simulation of a fracturing process runs as follows:
Given the specimen geometry and its material constants we first solve the initial
dynamic problem in the way previously described. When the stress σyy reaches a
prescribed critical value σc the crack is extended by one grid interval. The crack
now continues to grow, by one grid interval at a time, as long as the σyy stress level
ahead of the propagating crack tip reaches the value of σc . During the propagation
stage the time that elapses between two sequential extensions is recorded and the
corresponding velocity is evaluated.
The crack velocity for quasicrystals and periodic crystals is constructed in Fig.10.6-
5, from the figure, we observe that the velocity in quasicrystals is lower than that
of the periodic crystals; the phonon-phason coupling makes the quasicrystals being
different from periodic crystals. The reason for this is not so clear. Though Mikulla
et al[25] gave an explanation for crack slowly propagation (i.e., the quasi-static crack
growth) in decagonal quasi crystals, the case is quite different from the present crack
fast propagation.
We find that the fast crack propagating velocity is obviously different in qua-
sicrystals compared to the conventional engineering material. Next we will explore

the velocity under different loads in quasicrystal. The above described procedure was
conducted, keeping the same geometry and material constants. With various loads,
the relation between velocity and crack growth is constructed in Fig.10.6-6. The
crack velocity increases with increasing applied load smoothly. It is understandable
that as the load increases the time to reach the critical stress is less, so the velocity
increases.
Fig.10.6-5 The crack velocity versus normalized crack size with

different coupling constants
Fig.10.6-6 Variation of crack velocity versus crack growth size for different load levels
As shown in Fig.10.6-7, the calculated crack propagation results show some

roughness as the load level increases. Currently there is no experimental obser-
vation for fast crack propagation, though Ebert et al[26] made some observation by
scanning tunneling microscopy for quasi-static crack growth. Because the fast propa-
gation and quasi-static crack growth belong to two different regimes, the comparison
cannot be easily made.
Fig.10.6-7 Normalized crack growth size (a − a0 )/a0 of crack tip versus

time for different load levels
10.7 Elasto-/hydro-dynamics and applications to fracture dy-

namics of icosahedral quasicrystals
10.7.1 Basic equations, boundary and initial conditions
The elasto-/hydro-dynamics of icosahedral Al-Pd-Mn quasicrystals is more inter-
esting topic than that of decagonal Al-Ni-Co quasicrystals, especially a rich set of
experimental data for elastic constants can be used for the computation described
here. From the previous section we know there are lack of measured data for pha-
son elastic constants, which are obtained by Monte Carlo simulation, and results in
uncertainty for computation for decagonal quasicrystals. This shows the discussion
on icosahedral quasicrystals is more necessary, and the formulation and numerical
results are presented in this section.
If considering only the plane problem, especially for the crack problems, there
are much of similarities with those discussed in the previous section. We present
herein only the distinguishing part that are different.
For the plane problem, i.e.,
∂
= 0, (10.7-1)
∂z
the linearized elasto-/hydro-dynamics of icosahedral quasicrystals has non-zero dis-
placements uz , wz apart from ux , uy , wx , wy , so in the strain tensors,
10.7 Elasto-/hydro-dynamics and applications to fracture dynamics of... 217

1 ∂ui ∂uj ∂wi
εij = + , wij = ,
2 ∂xj ∂xi ∂xj
it increases some non-zero components compared with those in two-dimensional

quasicrystals. In connecting with this, in the stress tensors, the non-zero components
increase too. With these reasons, the stress-strain relation presents different nature
with that of decagonal quasicrystals though the generalized Hooke’s law has the
same form with that in one- and two-dimensional quasicrystals, i.e.,
σij = Cijkl εkl + Rijkl wkl , Hij = Rklij εkl + Kijkl wkl .
In particular the elastic constants are quite different from those discussed in the
previous sections, in which the phonon elastic constants can be expressed such as
Cijkl = λδij δkl + μ(δik δjl + δil δjk ) (10.7-2)
and the phason elastic constant matrix [K] and phonon-phason coupling elastic one
[R] are defined by formula (9.1-6) in Chapter 9, which are not listed here again.
Substituting these non-zero stress components into the equations of motion
∂ 2 ui ∂σij ∂wi ∂Hij

ρ 2
= , κ = (10.7-3)
∂t ∂xj ∂t ∂xj
and through the generalized Hooke’s law and strain-displacement relation we obtain
the final dynamic equations as follows:
2
∂ 2 ux ∂ux 2
2 ∂ ux 2
2
2 ∂ uy
2
2 ∂ ux 2 ∂ wx ∂ 2 wy ∂ 2 wx
+θ = c1 +(c1 −c2 ) +c +c3 +2 − ,
∂t2 ∂t ∂x2 ∂x∂y 2 ∂y2 ∂x2 ∂x∂y ∂y 2

∂ 2 uy ∂uy ∂ 2 uy ∂ 2 ux ∂ 2 uy ∂ 2 wy ∂ 2 wx ∂ 2 wy
2
+θ = c22 2
+(c21 −c22 ) +c21 2
+c23 2
−2 − ,
∂t ∂t ∂x ∂x∂y ∂y ∂x ∂x∂y ∂y 2
2 2
∂ 2 uz ∂uz 2 ∂ ∂2 2 ∂ wx ∂ 2 wx ∂ 2 wy ∂ 2 wz ∂ 2 wz
+θ = c2 + uz +c 3 − −2 + + ,
∂t2 ∂t ∂x2 ∂y 2 ∂x2 ∂y2 ∂x∂y ∂x2 ∂y2
2
∂wx ∂ ∂2 ∂2 ∂2
+θwx = d1 + w x +d 2 − wz
∂t ∂x2 ∂y 2 ∂x2 ∂y2
2
∂ ux ∂ 2 uy ∂ 2 ux ∂ 2 uz ∂ 2 uz
+d3 −2 − + −
∂x2 ∂x∂y ∂y 2 ∂x2 ∂y2
2 2
2
2
∂wy ∂ ∂ ∂ wz ∂ uy ∂ 2 ux ∂ 2 uy ∂ 2 uz
+θwy = d1 + w y −d 2 +d 3 +2 − −2 ,
∂t ∂x2 ∂y 2 ∂x∂y ∂x2 ∂x∂y ∂y2 ∂x∂y
2
∂wz ∂ ∂2
+θwz = (d1 −d2 ) + wz
∂t ∂x2 ∂y2
2 2
∂ wx ∂ 2 wx ∂ 2 wy ∂ ∂2
+d2 − −2 +d 3 + uz ,
∂x2 ∂y 2 ∂x∂y ∂x2 ∂y 2
(10.7-4)
in which
& + &
λ + 2μ μ R K1 K2 R
c1 = , c2 = , c3 = , d1 = , d2 = , d3 = ,
ρ ρ ρ κ κ κ
(10.7-5)
note that constants c1 , c2 and c3 have the meaning of elastic wave speeds, while d1 , d2
and d3 do not represent wave speed, but are diffusive coefficients and parameter θ
may be understood as a manmade damping coefficient as in the previous section.
Consider an icosahedral quasicrystal specimen with a central crack shown in
Fig.10.6-1, all parameters of geometry and loading are the same with those given
in the previous section, but in the boundary conditions there are some different
points, which are given as below if only the upper right quarter of the specimen is
considered:
⎧
⎪
⎪ u = 0, σyx = 0, σzx = 0, wx = 0, Hyx = 0, Hzx = 0, on x = 0 for 0 y H,
⎪ x
⎪
⎪
⎪
⎨ σxx = 0, σyx = 0, σzx = 0, Hxx = 0, Hyx = 0, Hzx = 0, on x = L for 0 y H,
⎪
σyy = p(t), σxy = 0, σzy = 0, Hyy = 0, Hxy = 0, Hzy = 0, on y = H for 0 x L,
⎪
⎪
⎪
⎪ σyy = 0, σxy = 0, σzy = 0, Hyy = 0, Hxy = 0, Hzy = 0, on y = 0 for 0 x a(t),
⎪
⎪
⎪
⎩ u = 0, σ = 0, σ = 0, w = 0, H = 0, H = 0, on y = 0 for a(t) < x L.
y xy zy y xy zy
(10.7-6)
The initial conditions are
ux (x, y, t)|t=0 = 0, uy (x, y, t)|t=0 = 0, uz (x, y, t)|t=0 = 0,

wx (x, y, t)|t=0 = 0, wy (x, y, t)|t=0 = 0, wz (x, y, t)|t=0 = 0, (10.7-7)
∂ux (x, y, t) ∂uy (x, y, t) ∂uz (x, y, t)
|t=0 = 0, |t=0 = 0, |t=0 = 0.
∂t ∂t ∂t
10.7.2 Some results
We now concentrate on investigating the phonon and phason fields in the icosahedral
3
Al-Pd-Mn quasicrystal, in which we take ρ = 5.1g/cm and λ = 74.2, μ = 70.4GPa
of the phonon elastic moduli, for phason ones K1 = 72, K2 = −37MPa (refer to
Chapter 9) and the constant relevant to diffusion coefficient of phason is Γw =
[21]
1/κ = 4.8 × 10−10 cm3 · μs/g . On the phonon-phason coupling constant, there
is no measured result for icosahedral quasicrystals so far, we take R/μ = 0.01 for
quasicrystals, and R/μ = 0 for “decoupled quasicrystals” or crystals.
The problem is solved by the finite difference method, the principle, scheme and
algorithm are almost the same as shown in the previous section, and shall not be
repeated here. The testing for the physical model, scheme, algorithm and computer
program are similar to those given in Section 10.6.
10.7 Elasto-/hydro-dynamics and applications to fracture dynamics of... 219
The numerical results for dynamic initiation of crack growth problem, the phonon
and phason displacements are shown in Fig.10.7-1.
Fig.10.7-1 Displacement components of quasicrystals versus time

(a)displacement component ux ; (b)displacement component uy ; (c)displacement component wx ;
(d)displacement component wy
The dynamic stress intensity factor KI (t) is defined by

#
KI (t) = lim π(x − a0 )σyy (x, 0, t)
x→a+
0
Fig.10.7-2 Normalized dynamic stress intensity factor of central crack specimen

under impact loading versus time
√
and the normalized dynamics stress intensity factor (S.I.F.) K̃I (t) = KI (t)/ πa0 p0
is used, the results are illustrated in Fig.10.7-2, in which the comparison with those
of crystals are shown, one can see the effects of phason and phonon-phason coupling
are evident very much.
For the fast crack propagation problem the primary results of the dynamic stress
intensity factor versus time are given in Fig.10.7-3.
Fig. 10.7-3 Normalized stress intensity factor of propagating crack with constant crack
speed versus time
Details of this work are reported in Ref. [27].
10.7.3 Conclusion and discussion
In Sections 10.5∼10.7 a new model on dynamic response of quasicrystals is formu-

lated based on both argument of Bak as well as argument of Lubensky et al. This
model can be regarded as an elasto-/hydro-dynamics model for the material, or as a
collaborating model of wave propagation and diffusion. This model is more complex
than that of pure wave propagation model for conventional crystals, and analytic
solution is very difficult to obtain, except for a few simple examples given in Sec-
tion 10.5. Numerical procedure based on finite difference algorithm are developed,
the computed results confirm the validity of wave propagation behaviour of phonon
field, and behaviour of diffusion of phason field. The interaction between phonons
and phasons are also recorded.
The finite difference formalism is applied to analyze dynamic initiation of crack
growth and crack fast propagation for two-dimensional decagonal Al-Ni-Co and
three-dimensional icosahedral Al-Pd-Mn quasicrystals, the displacement and stress
fields around the tip of stationary and propagating cracks are investigated. The
10.8 Appendix of Chapter 10: The detail of finite difference scheme 221
stress presents singularity with order r −1/2 , in which r denotes the distance measured
from the crack tip. For the fast crack propagation, which is a nonlinear problem—
moving boundary problem, one must provide additional condition for determining
solution. For this purpose we give a criterion for checking crack propagation/crack
arrest based on the critical stress criterion. Applications of this additional condition
for determining solution enable us to conduct numerical simulation of the moving
boundary value problem and examine the crack length-time evolution. However,
much important and difficult problems are left open for further study[31] .
10.8 Appendix of Chapter 10: The detail of finite difference

scheme
Equations (10.6-2) subjected to conditions (10.6-4) and (10.6-5) are very compli-
cated, analytic solution for the boundary-initial value problem is not available at
present, which have to be solved by numerical method. Here we extend the method
of finite difference of Shmuely and Alterman[28] scheme for analyzing crack problem
for conventional engineering materials to quasicrystalline materials.
A grid is imposed on the upper right of the specimen shown in Fig.10.8-1. For
Fig.10.8-1 Scheme of grid used in the text
convenience, the mesh size h is taken to be the same in both x and y directions. The
grid is extended beyond the half step by adding four special grid lines x = −h/2, x =
L + h/2, y = −h/2, y = H + h/2, which form the grid boundaries.
Denoting the time step by τ and using central difference approximations, the
finite difference formulation of equations (10.6-2), valid at the inner part of the grids
is
⎧
⎪
⎪ ux (x, y, t + τ ) = 2ux (x, y, t) − ux (x, y, t − τ )
⎪
⎪ τ 2
⎪
⎪
⎪
⎪ + c1 [ux (x + h, y, t) − 2ux (x, y, t) + ux (x − h, y, t)]
⎪
⎪ h 2
⎪
⎪ τ
⎪
⎪ + (c21 − c22 )[uy (x + h, y + h, t)
⎪
⎪
⎪
⎪ h
⎪
⎪ −uy (x + h, y − h, t) − uy (x − h, y + h, t) + uy (x − h, y − h, t)]
⎪
⎪
⎪
⎪ τ 2
⎪
⎪ c2 [ux (x, y + h, t) − 2ux (x, y, t) + ux (x, y − h, t)]
⎪
⎪ +
⎪
⎪ h 2
⎪
⎪ τ
⎪
⎪ + c3 [wx (x + h, y, t) − 2wx (x, y, t) + wx (x − h, y, t)]
⎪
⎪ h 2
⎪
⎪ τ
⎪
⎪ +2 c23 [wy (x + h, y + h, t) − wy (x + h, y − h, t)
⎪
⎪ h
⎪
⎪
⎪
⎪ −wy (x − h, y + h, t) + wy (x − h, y − h, t)]
⎪
⎪ τ 2
⎪
⎪
⎪
⎪ − c3 [wx (x, y + h, t) − 2wx (x, y, t) + wx (x, y − h, t)],
⎪
⎪ h
⎪
⎪
⎪
⎪ uy (x, y, t + τ ) = 2uy (x, y, t) − uy (x, y, t − τ )
⎪
⎪
⎪
⎪ τ 2
⎪
⎪ c2 [uy (x + h, y, t) − 2uy (x, y, t) + uy (x − h, y, t)]
⎪
⎪ +
⎪
⎪ hτ 2
⎪
⎪
⎪
⎪ + (c21 − c22 )[ux (x + h, y + h, t) − ux (x + h, y − h, t)
⎪
⎪ 2h
⎪
⎪
⎪
⎪ −ux (x − h, y + h, t) + ux (x − h, y − h, t)
⎪
⎪
⎪
⎪ τ 2
⎪
⎪
⎪
⎪ + c1 [uy (x, y + h, t) − 2uy (x, y, t) + uy (x, y − h, t)]
⎪
⎪ h 2
⎪
⎪ τ
⎪
⎪ + c3 [wy (x + h, y, t) − 2wy (x, y, t) + wy (x − h, y, t)]
⎪
⎪
⎪
⎪ h τ 2
⎨
−2 c23 [wx (x + h, y + h, t) − wx (x + h, y − h, t)
⎪ 2h
⎪
⎪ −wx (x − h, y + h, t) + wx (x − h, y − h, t)]
⎪
⎪
⎪
⎪ τ 2
⎪
⎪
⎪
⎪ − c3 [wy (x, y + h, t) − 2wy (x, y, t) + wy (x, y − h, t)],
⎪
⎪ h
⎪
⎪ τ
⎪
⎪ wx (x, y, t + τ ) = wx (x, y, t) + d22 2 [ux (x + h, y, t) − 2ux (x, y, t) + ux (x − h, y, t)]
⎪
⎪ h
⎪
⎪
⎪
⎪ τ
+d21 2 wx (x + h, y, t) + wx (x − h, y, t) − 4wx (x, y, t)
⎪
⎪
⎪
⎪ h
⎪
⎪
⎪
⎪ −
⎪
⎪ +w x (x, y + h, t) + w x (x, y h, t)
⎪
⎪
⎪
⎪
⎪
⎪ −2d 2 τ
uy (x + h, y + h, t) − uy (x + h, y − h, t)
⎪
⎪ 2
⎪
⎪ (2h)2
⎪
⎪
⎪
⎪ −u − − −
⎪
⎪ y (x h, y + h, t) + uy (x h, y h, t)
⎪
⎪
⎪
⎪ τ
⎪
⎪ −d22 2 [ux (x, y + h, t) − 2ux (x, y, t) + ux (x, y − h, t)] ,
⎪
⎪ h
⎪
⎪ 2 τ
⎪
⎪ wy (x, y, t + τ ) = wy (x, y, t) + d2 h2 [uy (x + h, y, t) − 2uy (x, y, t) + uy (x − h, y, t)]
⎪
⎪
⎪
⎪ τ
⎪
⎪ +d21 2 wy (x + h, y, t) + wy (x − h, y, t) − 4wy (x, y, t)
⎪
⎪
⎪
⎪ h
⎪
⎪
⎪
⎪ −
⎪
⎪
+w y (x, y + h, t) + w y (x, y h, t)
⎪
⎪
⎪
⎪ τ
⎪
⎪ 2
ux (x + h, y + h, t) − ux (x + h, y − h, t)
⎪
⎪ +2d 2
⎪
⎪ (2h)2
⎪
⎪
⎪
⎪ −u
⎪ x (x − h, y + h, t) + ux (x − h, y − h, t)
⎪
⎪
⎪
⎪ τ (10.8-1)
⎩ −d22 2 [uy (x, y + h, t) − 2uy (x, y, t) + uy (x, y − h, t)] .
h
10.8 Appendix of Chapter 10: The detail of finite difference scheme 223
The displacements at mesh points located at the special lines are determined
by satisfying the boundary conditions, we obtain respectively for points on the grid
lines x = −h/2 and x = L + h/2.
h h
− −
ux L+2 h , y, t =ux L−2 h , y, t
2 2
1 d21 (c21 − 2c22 ) + c23 d22 h2 h

± uy L− h , y + h, t − uy L−
2
h , y − h, t
2 c21 d21 − c23 d22 2 2
1 c23 (d21 − d23 ) h2 h

± w y L− h , y + h, t − wy 2
, y − h, t ,
2 c21 d21 − c23 d22 2 L− h2
h (10.8-2a)
−h
2 −
wx L+ h
, y, t=wx L−2 h , y, t
2 2
d2 (c2 − 2c2 ) h2 h
±2 22 21 2 22 uy L− h , y + h, t − u y
2
L− h
, y − h, t
c3 d2 − c1 d1 2 2
2 2 2 2
1 c3 d2 − c1 d3 h h
± 2 2 w y
2
, y + h, t − w y
2
, y − h, t ,
2 c3 d2 − c21 d21
h h
L− 2 L− 2
(10.8-2b)
h h 1 h h
−2 −2
uy L+ h , y, t = uy L− h , y, t ± ux L− h , y + h, t − ux L− h , y − h, t
2 2
2 2 2 2 2
1 c23 (d21 − d23 ) h2 h

± 2 2 w x , y + h, t − w x
2
, y − h, t ,
2 c2 d1 − c23 d22
h h
L− 2 L− 2
(10.8-2c)
−h
2
h
2
wy L+ h
, y, t = wy L− h
, y, t
2 2
1 c23 d22 − c22 d23 h2 h

2 2± 2 2 wx L− h , y + h, t − wx L−
2
h , y − h, t ,
2 c2 d1 − c3 d2 2 2
(10.8-2d)
where the equations (10.8-2a) and (10.8-2b) related to x = −h/2 is not valid. From
the first condition of (10.6-5), at x = 0, ux = 0 and wx = 0. To satisfy the condition
the displacements ux and wx at x = −h/2 is approximated by
⎧
⎪
⎪ h h
⎨ ux x, − , t = −ux x, , t ,
2 2 (10.8-3)
⎪
⎪ h h
⎩ wx x, − , t = −wx x, , t .
2 2
On the grid line y = −h/2 and y = H + h/2, we obtain

1
−h −h h h
ux x,H+2
h ,t = ux x,H− 2
h ,t ± 2
uy x + h,H− h ,t − uy x − h,H−
2
h ,t
2 2 2 2 2
2 2 2
1 c (d − d ) h h
± 23 2 1 2 3 2 wy x + h,H− 2
h ,t − wy x − h,H−2
h ,t ,
2 c2 d1 − c3 d2 2 2
(10.8-4a)

−h
2 −h
2
wx x,H+ h ,t = wx x,H− h ,t
2 2
1 c2 d2 − c22 d23 h
h

± 32 22 wy x + h, 2
, t − w y x − h, 2
, t ,
2 c2 d1 − c23 d22
h h
H− 2 H− 2
(10.8-4b)
h
−h
2 2
uy x,H+ h , t = uy x,H− h ,t
2 2
1 + d21 (c21 − 2c22 )

c23 d22 h h
± 2 2 2 2 ux x + h,H−2
h ,t − ux x − h,H−
2
h ,t
2 c1 d1 − c3 d2 2 2
2 2 2
1 c (d − d ) h h
± 23 2 3 2 1 2 wx x + h,H− 2
h ,t − wx x − h,H−2
h ,t ,
2 c1 d1 − c3 d2 2 2
(10.8-4c)
h
−h
2 2
wy x,H+ h , t = wy x,H− h ,t
2 2
d2 (c2 − c2 ) h
h

± 22 2 1 2 2 2 ux x + h,H−2
h ,t − ux x − h,H−
2
h ,t
c1 d1 − c3 d2 2 2
2 2 2 2
1 c d − c3 d2 h
h

± 12 32 2 2
2
wx x + h,H− h ,t − wx x − h,H−
2
h ,t ,
2 c1 d1 − c3 d2 2 2
(10.8-4d)
in which, equations (10.8-4c) and (10.8-4d) related to y = −h/2 is valid only along
the crack surface, namely, only for x a − h/2 at y = 0, in which the crack
terminates. From the last condition of (10.6-5), at y = 0 and the ahead of the crack,
uy = 0, wy = 0. To satisfy this condition the displacements uy and wy at y = −h/2
is approximated by
⎧
⎪ h h
⎪
⎨ uy x, − , t = −uy x, , t ,
2 2
(10.8-5)
⎪
⎪ h h
⎩ wy x, − , t = −wy x, , t .
2 2
In constructing the approximation (10.8-2∼10.8-5) we follow a method proposed

by Shmuely and Peretz[29] which was also successfully employed in Ref. [28] for con-
ventional engineering materials. According to this method, derivatives perpendicular
to the boundary are proposed by uncentered differences and derivatives parallel to
the boundary by centered difference. The real boundary can be considered as located
at a distance of half the mesh size from the grid boundaries.
The four grid corners require a special treatment. Different methods of handling
the discontinuities at such points have been proposed in the past. Here we found
that satisfactory results are obtained when the displacements sought are extrapo-
lated from those given along both sides of the corner in question. Accordingly, the
components ux , uy , wx , wy at (−h/2, −h/2) are given by
References 225

ux h h ux h h ux h h
− ,− ,t = ,− ,t + − , ,t
uy 2 2 uy uy
2 2 2 2
ux 3h h u h 3h
−0.5 ,− ,t + x − , ,t ,
uy 2 2 uy 2 2

wx h h w h h wx h h
− ,− ,t = x ,− ,t + − , ,t
wy 2 2 wy wy
2 2 2 2 (10.8-6)
wx 3h h wx h 3h
−0.5 ,− ,t + − , ,t .
wy 2 2 wy 2 2
Similar expressions are used for deriving the displacement components at (−h/2, H +
h/2), (L + h/2, H + h/2)and(L + h/2, −h/2).
By following relevant stability criterion of the scheme the computation is always
stable and achieves high precision. Discussions on this aspect are omitted here due
to space limitation.
References
[1] Lubensky T C , Ramaswamy S and Joner J. Hydrodynamics of icosahedral quasicrys-
tals. Phys Rev B, 1985, 32(11): 7444–7452
[2] Socolar J E S, Lubensky T C and Steinhardt P J. Phonons, phasons and dislocations
in quasicrystals. Phys Rev B, 1986, 34(5): 3345–3360
[3] Bak P. Phenomenological theory of icosahedral in commensurate (quasiperiodic)order
in Mn-Al alloys. Phys Rev Lett, 1985, 54(14): 1517–1519
[4] Bak P. Symmetry, stability and elastic properties of icosahedral in commensurate
crystals. Phys Rev B, 1985, 32(9): 5764–5772
Phys Rev B, 1993, 48(10): 7003–7010
and dislocations in quasicrystals. Reports on Progress in Physics, 2000, 63(1): 1–39
[7] Fan T Y, Li X F, Sun Y F. A moving screw dislocation in one-dimensional hexagonal
quasicrystal. Acta Physica Sinica(Overseas Edition), 1999, 8(3): 288–295
[8] Fan T Y. A study on special heat of one-dimensional hexagonal quasicrystals. J
Phys: Condens Matter, 1999, 11(45): L 513–L 517
two-dimensional quasicrystals. Euro Phys J B, 2003, 31(1): 25–27
[11] Li C L, Liu Y Y. Phason-strain influences on low-temperature specific heat of the
[12] Li C L, Liu Y Y. Low-temperature lattice excitation of icosahedral Al-Mn-Pd qua-

sicrystals. Phys Rev B, 2001, 63(6): 064203
[13] Fan T Y , Wang X F, Li W et al. Elasto-hydrodynamics of quasicrystals. Phil Mag,
2009, 89(6): 501–512
[14] Zhu A Y, Fan T Y. Dynamic crack propagation in a decagonal Al-Ni-Co quasicrystal.
J Phys: Condens Matter, 2008, 20(29): 295217
[15] Rochal S B, Lorman V L. Anisotropy of acoustic-phonon properties of an icosahedral
quasicrystal at high temperature due to phonon-phason coupling Phys Rev B, 2000,
62(2): 874–879
[16] Rochal S B, Lorman V L. Minimal model of the phonon-phason dynamics on icosa-
hedral quasicrystals and its application for the problem of internal friction in the
i-AlPdMn alloys. Phys Rev B, 2002, 66(14): 144204
[17] Hirth J P, Lorthe J. Theory of Dislocations. 2nd ed On Eshelby’s solution one can
refer to New York: John Wiely & Sons, 1982
[18] Yoffe E H. Moving Griffith crack. Phil Mag,1951, 43(10): 739–750
[19] Chernikov M A, Ott H R, Bianchi A et al. Elastic moduli of a single quasicrystal of
decagonal Al-Ni-Co: evidence for transverse elastic isotropy. Phys Rev Lett, 1998,
80(2): 321–324
[20] Jeong H C, Steinhardt P J. Finite-temperature elasticity phase transition in decago-
nal quasicrystals. Phys Rev B, 1993, 48(13): 9394–9403
[21] Walz C. Zur Hydrodynamik in Quasikristallen. Diplomarbeit. Universitaet Stuttgart,
2003
[22] Maue A W. Die entspannungswelle bei ploetzlischem Einschnitt eines gespannten
elastischen Koepores. Zeitschrift fuer angewandte Mathematik und Mechanik, 1954,
14(1): 1–12
[23] Murti V, Vlliappan S. The use of quarter point element in dynamic crack analysis.
Engineering Fracture Mechanics, 1982, 23(3): 585–614
[24] Chen Y M. Numerical computation of dynamic stress intensity factor s by a La-
grangian finite-difference method (the HEMP code). Engineering Fracture Mechan-
ics, 1975, 7(8): 653–660
[25] Mikulla R, Stadler J, Krul F et al. Crack propagation in quasicrystals. Phys Rev
Lett, 1998, 81(15): 3163–3166
[26] Ebert Ph, Feuerbacher M, Tamura N et al. Evidence for a cluster-based on structure
of Al-Pd-Mn single quasicrystals. Phys Rev Lett, 1996, 77(18): 3827–3830
[27] Wang X F, Fan T Y, Zhu A Y. Dynamic behaviour of the icosahedral Al-Pd-Mn
quasicrystal with a Griffith Crack, Chin Phys B, 2009, 18(2), 709-714. (in Section
10.7 a part data are introduced from Zhu A Y and Fan T Y’s unpublished work:
Fast crack propagation in three-dimensional icosahedral Al-Pd-Mn quasicrystals)
References 227
[28] Shmuely M, Alterman Z S. Crack propagation analysis by finite differences. Journal

of Applied Mechanics, 1973, 40(4): 902–908
[29] Shmuely M, Peretz D. Static and dynamic analysis of the DCB problem in fracture
mechanics. Int J of Solids and Structures, 1976, 12(1): 67–67
[30] Takeuchi S, Iwanaga H, Shibuya T. Hardness of quasicrystals. Japanese J Appl Phys,
1991, 30(3): 561–562
[31] Coddens G. On the problem of the relation between phason elasticity and phason
dynamics in quasicrystals, Eur Phys. J B, 2006, 54(1): 37–65
Chapter 11
Complex variable function method for
elasticity of quasicrystals
In Chapters 7∼9 we frequently used the complex variable function method to solve
problems of elasticity of quasicrystals, and many exact analytic solutions were ob-
tained by the method. In those chapters we only provided the results, and the
underlying principle and details of the method could not be discussed. Considering
the relative new feature and particular effect of the method, it is helpful to attempt a
further discussion in depth. Of course this may lead to a slight repetition of relevant
content of Chapters 7∼9.
It is well-known that the so-called complex potential method in the classical
elasticity is effective, in general, mainly for solving harmonic and biharmonic partial
differential equations, for these equations the solutions can be expressed by analytic
√
functions of single complex variable z = x + iy, i = −1. In addition, the quasi-
biharmonic partial differential equations can be solved by analytic functions of two
complex variables such as z1 = x + α1 y, z2 = x + α2 y, in which α1 , α2 are com-
plex constants. The study of elasticity of quasicrystals has led to discovery of some
multi-harmonic and multi-quasiharmonic equations, which cover a quite wide range
of partial differential equations appearing in the field to date and have been intro-
duced in Chapters 5∼9. The discussion on the complex analysis for these equations
is significant. We know that the Muskhelishvili complex potentials method for the
classical isotropic plane elasticity[1] , which solves mainly the biharmonic equation,
made great contributions for a quite wide range of fields in science and engineer-
ing. And the complex potential method developed by Lekhnitzkii[2] for the classical
anisotropic plane elasticity which solves mainly the quasi-biharmonic equation. The
present formulation and solutions of the complex analysis for, e.g. quadruple and
sextuple harmonic equations and quadruple-quasiharmonic equation, are at least a
new development of the complex potential method of the classical elasticity. Though
the new method is used to solve elasticity problems of quasicrystals at present, it
may be extended into other disciplines of science and engineering in future.
At first we simply review the complex potential method for harmonic and bi-

230 Chapter 11 Complex variable function method for elasticity of quasicrystals
harmonic equations, which does not belong to a new innovation from point of view
of methodology, then focus on those for quadruple and sextuple harmonic equa-
tions and quadruple quasiharmonic equation, and with discussions in detail pre-
senting their new features from the angle of elasticity as well as complex potential
method.
11.1 Harmonic and quasi-biharmonic equations in anti-plane

elasticity of one-dimensional quasicrystals
The final governing equations of elasticity of one-dimensional quasicrystals present
the following two kinds discussed in Chapter 5:

c44 ∇2 uz + R3 ∇2 wz = 0,
(11.1-1)
R3 ∇2 uz + K 2 ∇2 wz = 0,

∂4 ∂4 ∂4 ∂4 ∂4
c1 4 + c2 3 + c3 2 2 + c4 + c5 G = 0, (11.1-2)
∂x ∂x ∂y ∂x ∂y ∂x∂y 3 ∂y 4
in which equations (11.1-1) are actually two decoupled harmonic equations of uz and
wz , whose complex variable function method was introduced in the Sections 8.1, 8.2
and 8.7, here we do not repeat any more.
Equation (11.1-2) is a quasi-biharmonic equation which describes the phonon-
phason coupling elasticity field for some kinds of one-dimensional quasicrystal sys-
tems, Liu, Fan and Guo[3] and Liu[4] studied some solutions of them in terms of the
complex variable function method, whose origin comes from the classical work of
Lekhnitskii[2] , reader can find some beneficial hints in the monograph.
11.2 Biharmonic equations in plane elasticity of point group

12mm two-dimensional quasicrystals
From Chapter 6 we know that in elasticity of dodecagonal quasicrystals, the phonon
and phason fields are decoupled each other. For such a plane elasticity we have the
final governing equations as follows:
∇2 ∇2 F = 0, ∇2 ∇2 G = 0. (11.2-1)
The complex representation of solution of (11.2-1) is

-
F (x, y) = Re[z̄ϕ1 (z) + ψ1 (z)dz],
- (11.2-2)
G(x, y) = Re[z̄π1 (z) + χ1 (z)dz],
where φ1 (z), ψ1 (z), π1 (z) and χ1 (z) are any analytic functions of complex variable
√
z = x + iy, (i = −1). For these kinds of biharmonic equations Muskhelishvili[1]
11.3 The complex variable function method of quadruple harmonic equations and... 231
developed systematical complex variable function method, reader can find some
details in the well known monograph, we need not discuss them any more.
11.3 The complex variable function method of quadruple har-

monic equations and applications in two-dimensional qua-
sicrystals
As it was discussed in Chapters 6∼8, for point groups 5m and 10mm or point groups
5, 5 and 10, 10 quasicrystals, either by the displacement potential formulation or by
the stress potential formulation we obtain the final governing equation is quadruple
harmonic equation, whose complex variable function method is newly created by
Liu and Fan[5] based on the displacement potential formulation and by Li and Fan[6]
based on the stress potential formulation. This complex potential method develops
the methodology was used in the classical elasticity. It is necessary to give some
further discussions in-depth. The representation in the following will be done based
on the stress potential formulation, and discussion is given only for point groups 5,
5 and 10, 10 quasicrystals, because the point groups 5m and 10mm ones can be seen
as a special case of the former.
11.3.1 Complex representation of solution of the governing equation
Because it is relatively simpler for the case of point groups 5m and 10mm, which
belong to the special case of point group 5, 5 and point groups 10, 10, we here discuss
only the final governing equation of plane elasticity of pentagonal of point groups 5,
5 and decagonal quasicrystals of point groups 10, 10
∇2 ∇2 ∇2 ∇2 G = 0, (11.3-1)
where G(x, y) is the stress potential function. The solution of equation (11.3-1) is

1 2 1 3
G = 2Re g1 (z) + z̄g2 (z) + z̄ g3 (z) + z̄ g4 (z) , (11.3-2)
2 6
where gj (z)(j = 1, · · · , 4) are four analytic functions of a single complex variable

z ≡ x + iy = reiθ . The bar denotes the complex conjugate hereinafter, i.e., z̄ =
x − iy = re−iθ . We call these functions the complex stress potentials, or the complex
potentials in brief.
11.3.2 Complex representation of the stresses and displacements
The Section 8.4 shows, from fundamental solution (11.3-2) one can find the complex
representation of the stresses as below:
⎧
⎪
⎪ σxx = −32c1 Re(Ω (z) − 2g4 (z)),
⎪
⎪
⎪
⎪ σyy = 32c1 Re(Ω (z) + 2g4 (z)),
⎪
⎪
⎪
⎪
⎪
⎪
⎨ σxy = σyx = 32c1 ImΩ (z),
Hxx = 32R1 Re(Θ (z) − Ω (z)) − 32R2 Im(Θ (z) − Ω (z)), (11.3-3)
⎪
⎪
⎪ H = −32R Im(Θ (z) + Ω (z)) − 32R Re(Θ (z) + Ω (z)),
⎪
⎪
⎪ xy 1 2
⎪
⎪
⎪
⎪ Hyx = −32R1 Im(Θ (z) − Ω (z)) − 32R2 Re(Θ (z) − Ω (z)),
⎪
⎪
⎩
Hyy = −32R1 Re(Θ (z) + Ω (z)) + 32R2 Im(Θ (z) + Ω (z),
where ⎧
⎨ (IV ) (IV ) 1 (IV )
Θ(z) = g2 (z) + z̄g3 (z) + z̄ 2 g4 (z),
2 (11.3-4)
⎩ Ω (z) = g (IV ) (z) + z̄g (IV ) (z),
3 4
in which the prime, two prime, three prime and superscript (IV ) denote the first to
fourth order derivatives of gi (z) to variable z, in addition Θ (z) = dΘ(z)/dz, it is
evident that Θ(z) and Ω (z) are not analytic functions.
By some derivation from (11.3-3) we have the complex representation of the
displacements such as
ux + iuy = 32(4c1 c2 − c3 − c1 c4 )g4 (z) − 32(c1 c4 − c3 )(g3 (z) + zg4 (z)), (11.3-5)
32(R1 − iR2 )
wx + iwy = Θ(z) (11.3-6)
K1 − K 2
with constants
c
c = M (K1 + K2 ) − 2(R12 + R22 ), c1 = + M,
K1 − K2
c + (L + M )(K1 + K2 ) 2
R + R22
K1 + K 2 (11.3-7)
c2 = , c3 = 1 , c4 = .
4(L + M )c c c
11.3.3 The complex representation of boundary conditions
In the following we consider only the stress boundary value problem, i.e., at the
boundary curve Lt the tractions (Tx , Ty ) and generalized tractions (hx , hy ) are given,
and there are the stress boundary conditions such as
σxx cos(n, x)+σxy cos(n, y) = Tx , σxy cos(n, x)+σyy cos(n, y) = Ty , (x, y) ∈ Lt ,

(11.3-8)
Hxx cos(n, x)+Hxy cos(n, y) = hx , (x, y) ∈ Lt ,
Hxy cos(n, x)+Hyy cos(n, y) = hy ,
(11.3-9)
where T x , T y and hx , hy are tractions and generalized tractions at the boundary
Lt where the stresses are prescribed.
From (11.3-8) and after some derivation the phonon stress boundary condition
can be reduced to the equivalent form

i
g4 (z) + g3 (z) + zg4 (z) = (Tx + iTy )ds, z ∈ Lt . (11.3-10)
32c1
From equations. (11.3-9), (11.3-3) and (11.3-4), we have

(R2 − iR1 )Θ(z) = i (hx + ihy )ds, z ∈ Lt . (11.3-11)
11.3.4 Structure of complex potentials

11.3.4.1 Arbitrariness in the definition of the complex potentials
For simplicity, we introduce the following new symbols

(IV )
g2 (z) = h2 (z), g3 (z) = h3 (z), g4 (z) = h4 (z), (11.3-12)
then equation (11.3-3) can be written as follows:
σxx + σyy = 128c1 Re h4 (z), (11.3-13)
σyy − σxx + 2iσxy = 64c1 Ω (z) = 64c1 [h3 (z) + z̄h4 (z)], (11.3-14)
Hxy − Hyx − i(Hxx + Hyy ) = 64(iR1 − R2 )Ω (z), (11.3.15)
(Hxx − Hyy ) − i(Hxy + Hyx ) = 64(R1 + R2 )Θ (z). (11.3-16)
By inspection, the state of phonon and phason stresses is not altered if replacing
h4 (z) by h4 (z) + Diz + γ, (11.3-17)
h3 (z) by h3 (z) + γ , (11.3-18)
h2 (z) by h2 (z) + γ , (11.3-19)
in equations (11.3-13)–(11.3-16) where D is a real constant and γ, γ , γ are arbitrary

complex constants.
Now consider how these substitutions affect the components of the displacements
which were determined by the formulas (11.3-5) and (11.3-6).Direct substitution
shows that
ux + iuy = 32(4c1 c2 − c3 − c1 c4 )h4 (z) − 32(c1 c4 − c3 )(h3 (z) + zh4 (z))

+32(4c1 c2 − 2c3 )Diz + [32(4c1 c2 − c3 − c1 c4 )γ − 32(c1 c4 − c3 )γ ],
(11.3-20)

32(R1 − iR2 ) 1 2 32(R1 − iR2 )
wx + iwy = h2 (z) + zh3 (z) + z h4 (z) + γ .
K1 − K 2 2 K1 − K 2
(11.3-21)
Formulas (11.3-20) and (11.3-21) show that a substitution of the form (11.3-17),
(11.3-19) will affect the displacements, unless
c1 c4 − c3
D = 0, γ= γ , γ = 0.
4c1 c2 − c3 − c1 c4
11.3.4.2 General formulas for finite multiply connected region
Consider now the case when the region S, occupied by the quasicrystal, is multiply
connected. In general the region is bounded by several simple closed contours s1 ,
s2 , · · · , sm , sm+1 , the last of these contours is to contain all the others, depicted in
Fig. 11.3-1, i.e., a plate with holes. We assume that the contours do not intersect
themselves and have no points in common. Sometimes we call s1 , s2 , · · · , sm as inner
boundaries and sm+1 as outer boundary of the region. It is evident that the points
z1 , z2 , · · · , zm are fixed points in the holes, but located out of the material.
Fig. 11.3-1 Finite multi-connected region
Similar to the discussion of the classical elasticity theory (refer to [1]), we can obtain

m
h4 (z) = Ak ln(z − zk ) + h4∗ (z), (11.3-22)
k=1

m
m
h4 (z) = Ak z ln(z − zk ) + γk ln(z − zk ) + h4∗ (z), (11.3-23)
k=1 k=1

m
h3 (z) = γk ln(z − zk ) + h3∗ (z), (11.3-24)
k=1
recalling zk denote fixed points outside the region S, h3∗ (z), h4∗ (z) are holomorphic
(analytic and single-valued, refer to the Major Appendix of this book) in S, Ak real
constants, γk , γk complex constants.
By substituting (11.3-22)∼(11.3-24) into (11.3-16) one can find that

m
h2 (z) = γk ln(z − zk ) + h2∗ (z), (11.3-25)
k=1
h2∗ (z)is holomorphic in S and γk are complex constants.

Consideration will be given to the condition of single-valuedness of phonon dis-
placements. By equation (11.3-5), one has
ux + iuy = 32(4c1 c2 − c3 − c1 c4 )h4 (z) − 32(c1 c4 − c3 )(h3 (z) + zh4 (z)). (11.3-26)
Substituting (11.3-23)∼(11.3-25) into (11.3-26), it is immediately seen that
[ux+iuy ]sk = 2πi{[32(4c1 c2−c3−c1 c4 )+32(c1 c4−c3 )]Ak z+32(4c1 c2−c3−c1 c4 )γk+γk (z)},
(11.3-27)
in which [ ]Sk denotes the increase undergone by the expression in brackets for one
anti-clockwise circuit of the contour sk . Hence it is necessary and sufficient for the
single-valuedness of phonon displacements that in the formulas (11.3-22)∼(11.3-25)
Ak = 0, 32(4c1 c2 − c3 − c1 c4 )γk + γk = 0. (11.3-28)
Similar to the above mentioned discussion, by equation (11.3-6), one has
32(R1 − iR2 )
[wx + iwy ]sk = (−2πi)γk . (11.3-29)
K1 − K 2
Hence it is necessary and sufficient for the single-valuedness of phason displacements
if
γk = 0. (11.3-30)
It will now be shown shat the quantities γk , γk may be very simply expressed in
terms of Xk , Yk , where (Xk , Yk ) denotes the resultant vector of the external stresses,
exerted on the contour sk . By (11.3-10), applying it to the contour sk , one has
−32c1 i[h4 (z) + h3 (z) + zh4 (z)]sk = Xk + iYk (11.3-31)
with
Xk = Tx ds, Yk = Ty ds.
sk sk
In the present case the normal vector n must be directed outwards with respect
to the region Sk . Consequently, the contour sk must be traversed in the clockwise
direction. Taking this fact into consideration, one obtains
i
−2πi(γk − γk ) = (Xk + iYk ). (11.3-32)
32c1
By equations (11.3-28), (11.3-31) and (11.3-32), one has
Ak = 0,
γk = d1 (Xk + iYk ), γk = d2 (Xk − iYk ), (11.3-33)
where
1 4c1 c2 − c3 − c1 c4
d1 = , d2 = −
64c1 π[32(4c1 c2 − c3 − c1 c4 ) + 1] 2c1 π[32(4c1 c2 − c3 − c1 c4 ) + 1]
(11.3-34)
and which are independent of the suffix k. So that
⎧
⎪
⎪ m
⎪
⎪ h (z) = d (Xk + iYk ) ln(z − zk ) + h4∗ (z),
⎪
⎪
4 1
⎨ k=1
m
(11.3-35)
⎪
⎪ h3 (z) = d 2 (Xk − iYk ) ln(z − zk ) + h3∗ (z),
⎪
⎪
⎪
⎪
⎩ h (z) = h k=1 (z).
2 2∗
We can conclude that the complex functions h2 (z), h3 (z), h4 (z) must be expressed
by formulas (11.3-35) to assure the single-valuedness of stresses and displacements,
where h2∗ (z), h3∗ (z), h4∗ (z) are holomorphic in S.
11.3.4.3 Case of infinite regions
From the point of view of application in Chapter 9, this chapter and Chapters 14,
15 the consideration of infinite regions is likewise of major interest. We assume that
the contour sm+1 goes to infinity.
Since equations (11.3-13) and (11.3-14) are similar to the classical elasticity the-
ory, we have

h4 (z) = d1 (X + iY ) ln z + (B + iC)z + h04 (z),

(11.3-36)
h3 (z) = d2 (X − iY ) ln z + (B + iC )z + h03 (z),
where B, C, B , C are unkonwn real constants to be determined and

m
m
X= Xk , Y = Yk ,
k=1 k=1
h03 (z), h04 (z) are functions, holomorphic in region S, including the point at infinity,i.e,
for sufficiently large |z| they may be expanded into series of the form
a1 a2 a1 a2
h04 (z) = a0 + + 2 + ··· , h03 (z) = a0 + + 2 + ··· . (11.3-37)
z z z z
On the basis of (11.3-2) the state of stresses will not be altered by assuming
a0 = a0 = 0.
By the theorem of Laurent expansion, the function h2∗ (z) may be represented in
region S including point at infinity by the series
+∞

h2∗ (z) = cn z n . (11.3-38)
−∞
Substituting equations (11.3-36) and (11.3-38) into equation (11.3-16), one has
(Hxx − Hyy ) − i(H xy + Hyx )
+∞

d 2 1 2d1
= 2 × 32(R1 + R2 cn nz n−1 + z̄ − 2 + h03 (z) + z̄ 2 + h04 (z) ,
−∞
z 2 z3
(11.3-39)
hence it follows that for the stresses to remain finite as |z| → ∞, one must have
cn = 0, n 2.
It is obvious that the phonon and phason stresses will be bounded, if these conditions
are satisfied. Hence one has finally
⎧
⎨ h4 (z) = d1 (X + iY ) ln z + (B + iC)z + h04 (z),
h3 (z) = d2 (X − iY ) ln z + (B + iC )z + h03 (z), (11.3-40)
⎩
h2 (z) = (B + iC )z + h02 (z),
where B , C are unknown real constants to be determined, h02 (z) is function, holo-
morphic in region S, including the point at infinity, thus it has the form similar to
that of (11.3-37),
a a
h02 (z) = a0 + 1 + 22 + · · · . (11.3-41)
z z
We have assumed that a0 = a0 = 0 already, now further assume a0 = 0, i.e.,
h04 (∞) = h03 (∞) = h02 (∞) = 0,
then from (11.3-40) and (11.3-13)∼(11.3-16) one can determine

⎧ (∞) (∞) (∞) (∞) (∞)
⎪
⎪ σxx + σyy σxx − σyy σxy
⎪
⎪ B = , B
= , C
= ,
⎪
⎪ 128c1 64c1 32c1
⎪
⎪
⎨ (∞) (∞) (∞) (∞)
R2 (Hxy − Hyx ) − R1 (Hxx + Hyy )
B = , (11.3-42)
⎪
⎪ 2 2
64(R1 − R2 )
⎪
⎪
⎪
⎪ (∞) (∞) (∞) (∞)
⎪
⎪ R1 (Hxy − Hyx ) − R2 (Hxx + Hyy )
⎩ C =
64(R12 − R22 )
(∞) (∞)
in which C has no usage, we put it to be zero, and σij and Hij represent the
applied stresses at point of infinity.
11.3.5 Conformal mapping
If we constrain our discussion only for case of stress boundary value problems, then
the problems will be solved under boundary conditions (11.3-10) and (11.3-11). For
some complicated regions solutions of the problems cannot be directly obtained in
the physical plane (i.e., the z-plane). We must use a conformal mapping
z = ω(ζ) (11.3-43)
to transform the region studied in the plane onto interior of the unit circle γ in the
mapping plane (say e.g. ζ-plane).
Substituting (11.3-43) into (11.3-40) we have
⎧
⎨ h4 (z) = Φ4 (ζ) = d1 (X + iY ) ln ω(ζ) + Bω(ζ) + Φ40 (ζ),
h3 (z) = Φ3 (ζ) = d2 (X − iY ) ln ω(ζ) + (B + iC )ω(ζ) + Φ30 (ζ), (11.3-44)
⎩
h2 (z) = Φ2 (ζ) = (B + iC )ω(ζ) + Φ20 (ζ),
where
Φj (ζ) = hj [ω(ζ)], Φj0 (ζ) = h0j [ω(ζ)], j = 1, · · · , 4.
In addition
Φi (ζ)
hj (z) = .
ω (ζ)
At the mapping plane the boundary conditions (11.3-10) and (11.3-11) stand for

Φ (σ) i
Φ4 (σ) + Φ3 (σ) + ω(σ) 4 = (Tx + iTy )ds, (11.3-10 )
ω (σ) 32c1

(R2 − iR1 )Θ(σ) = i (hx + ihy )ds, (11.3-11 )
where σ = eiϕ represents the value of ζ at the unit circle (i.e., ρ = 1). From these
boundary value equations we can determine the unknown functions Φi (ζ)(i = 2, 3, 4).
11.3.6 Reduction of the boundary value problem to function equations
Due to Φ1 (ζ) = 0, we now have three unknown functions Φj (ζ)(j = 2, 3, 4). Taking
conjugate of (11.3-10 ) yields

Φ (σ) i
Φ4 (σ) + Φ3 (σ) + ω(σ) 4 =− (Tx − iTy )ds. (11.3-10 )
ω (σ) 32c1
Substituting the first one of equations (11.3-4) into (11.3-11 ), then multiplying
dσ/2πi(σ − ζ) to both sides of (11.3-10 ) and (11.3-11 ) and integrating along the
unit circle γ leads to
⎧
⎪
⎪ 1 Φ4 (σ)dσ 1 Φ3 (σ)dσ 1 ω(σ) Φ 4 (σ)dσ 1 1 tdσ
⎪
⎪ + + =
⎪
⎪ 2πi σ−ζ 2πi σ−ζ 2πi ω (σ) σ − ζ 32c1 2πi σ−ζ
⎪
⎪
γ γ γ γ
⎪
⎪
⎪
⎪ 1 Φ4 (σ)dσ 1 Φ3 (σ)dσ 1 ω(σ) Φ4 (σ)dσ 1 1 tdσ
⎪
⎪ + + =
⎪
⎪ 2πi σ−ζ 2πi σ−ζ 2πi
ω (σ) σ − ζ 32c1 2πi σ−ζ
⎪
⎪ γ γ γ γ
⎪
⎨
1 Φ2 (σ)dσ 1 ω(σ) Φ 3 (σ)dσ
+
⎪
⎪ 2πi γ σ − ζ 2πi γ ω (σ) σ − ζ
⎪
⎪
⎪
⎪ 2 2
⎪
⎪ 1 ω(σ) Φ4 (σ)dσ ω(σ) ω (σ) Φ4 (σ)dσ
⎪
⎪ + −
⎪
⎪ 2πi γ [ω (σ)]2 σ − ζ [ω (σ)]3 σ−ζ
⎪
⎪ γ
⎪
⎪
⎪
⎪ 1 1 hdσ
⎪
⎩ = ,
R1 − iR2 2πi γ σ − ζ
(11.3-45)
- - -
where t = i (Tx + iTy )ds, t = −i (Tx − iTy )ds, h = i (h1 + ih2 )ds in equations
(11.3-45), which are the function equations to determine complex potentials Φj (ζ)
which are analytic in the interior of the unit circle γ, and satisfy boundary value
conditions (11.3-45) at the unit circle.
11.3.7 Solution of the function equations

According to the Cauchy’s integral formula (refer to Appendix I)

1 Φi (σ) 1 Φi (σ)
dσ = Φi (ζ), dσ = Φi (0), |ζ| < 1.
2πi γ σ−ζ 2πi γ σ−ζ
So that (11.3-45) are reduced to

⎧
⎪
⎪ 1 ω(σ) Φ4 (σ)dσ i 1 tdσ
⎪
⎪ Φ4 (ζ) + Φ 3 (0) + = ,
⎪
⎪ 2πi γ ω(σ) σ − ζ 32c1 2πi σ −ζ
⎪
⎪
γ

⎪
⎪
⎪
⎪ 1 ω(σ) Φ4 (σ)dσ i 1 tdσ
⎪
⎪ Φ (0) + Φ (ζ) + = − ,
⎨ 4 3
2πi γ ω(σ) σ − ζ 32c1 2πi γ σ − ζ
⎪ 2 (11.3-46)
⎪
⎪ 1 ω(σ) Φ3 (σ)dσ 1 ω(σ) Φ4 (σ)dσ
⎪
⎪ Φ2 (ζ) + +
⎪
⎪ 2πi γ ω (σ) σ − ζ 2πi γ [ω (σ)]2 σ − ζ
⎪
⎪
⎪
⎪ 2
⎪
⎪ ω(σ) ω (σ) Φ4 (σ)dσ i 1 hdσ
⎪ −
⎩ = .
γ [ω (σ)]3 σ − ζ R 1 − iR2 2πi γ σ −ζ
The calculation of integrals in (11.3-46) depends upon configuration of the sample,

so the mapping function ω(ζ), and the applied stresses t and h. In the following we
will give some concrete solutions for given configuration and applied traction.
11.3.8 Example 1 Elliptic notch/crack problem and solution

We here calculate the stress and displacement fields induced by an elliptic notch

x2 y 2
L : + = 1 in an infinite plane of icosahedral quasicrystal, see Fig.11.3-2,
a2 b2
the edge of which is subjected to a uniform pressure p.
Fig. 11.3-2 An elliptic notch in a decagonal quasicrystal
The boundary conditions can be expressed by equations (11.3-10) and (11.3-11), for
simplicity, we assume hx = hy = 0. Thus
⎧
⎪
⎨ i (Tx + iTy )ds = i (−p cos(n, x) − ip cos(n, y))ds = −pz = −pω(σ),

⎪
⎩ i (hx + ihy )ds = 0.
(11.3-47)
In addition in this case in formulas (11.3-44)
X = Y = 0,
B = 0, B = C = 0, B = C = 0, (11.3-48)
so Φi (ζ) = Φi0 (ζ), but in the following we omit the superscript of the functions Φi0 (ζ)
for simplicity.
The conformal mapping is

1
z = ω(ζ) = R0 + mζ , (11.3-49)
ζ
to transform the region containing ellipse at the z-plane onto the interior of the
unit circle at the ζ-plane, refer to Fig.11.3-3, where ζ = ξ + iη = ρeiϕ and R0 =
a+b a−b
,m= .
2 a+b
Substituting (11.3-48) and (11.3-49) into the first two equations of function equation
(11.3-46), one obtains
Fig.11.3-3 Conformal mapping from the exterior of the elliptic hole at z-plane onto the
interior of the unit circle at ζ-plane
⎧
⎪
⎪ pR0 (1 + m2 )ζ
⎨ Φ3 (ζ) = , (11.3-50a)
32c1 mζ 2 − 1
⎪
⎪ pR0 (11.3-50b)
⎩ Φ4 (ζ) = − mζ.
32c1
Substitution of
2
ω(σ) σ2 + m ω(σ) ω (σ) 2σ(σ 2 + m)2
= σ , = −
ω (σ) mσ 2 − 1 ω (σ)3 (mσ 2 − 1)3
and (11.3-50a)(11.3-50b) into the third equation of (11.3-46) yields

1 Φ2 (σ)dσ 1 σ 2 + m Φ3 (σ)dσ 1 σ(σ 2 + m)2 Φ4 (σ)dσ
+ σ + = 0.
2πi γ σ − ζ 2πi γ mσ 2 − 1 σ − ζ 2πi γ (mσ 2 − 1)3 σ − ζ
(11.3-50c)
By Cauchy’s integral formula, we have

1 Φ2 (σ)dσ
= Φ2 (ζ),
2πi γ σ − ζ

1 σ2 + m Φ3 (σ)dσ ζ2 + m
σ = ζ Φ (ζ),
2πi γ mσ 2 − 1 σ − ζ mζ 2 − 1 3

1 σ(σ 2 + m)2 Φ4 (σ)dσ ζ(ζ 2 + m)2
= Φ (ζ).
2 3
2πi γ (mσ − 1) σ − ζ (mζ 2 − 1)3 4
Substituting these equations and (11.3-50a), (11.3-50b) into (11.3-50c), one obtains
pR0 ζ(ζ 2 + m)[(1 + m2 )(1 + mζ 2 ) − (ζ 2 + m)]

Φ2 (ζ) = . (11.3-50d)
32c1 (mζ 2 − 1)3
Utilizing the above mentioned results, the phonon and phason stresses can be deter-
mined at the ζ-plane, so the displacements.
If take m = 1, from (11.3-50) we can obtain solution of the Griffith crack, in particu-
lar the explicit solution at z-plane can be explored by taking inversion ζ = ω −1 (z) =
#
z/a − z 2 /a2 − 1 (as m = 1) into the relevant formulas. The concrete results are
given in Section 8.4, which are omitted here.
11.3.9 Example 2 Infinite plane with an elliptic hole subjected to a ten-

sion at infinity
In this case
p
X = Y = 0, Tx = Ty = 0, B= , B = C = 0,
64c1 (11.3-51)
B = C = 0, t = t = h = 0,
so that from (11.3-44),
h4 (z) = Φ4 (ζ) = Bω(ζ) + Φ40 (ζ),

h3 (z) = Φ3 (ζ) = Φ30 (ζ), (11.3-52)
h2 (z) = Φ2 (ζ) = Φ20 (ζ),
Substituting (11.3-52) into (11.3-45), we obtain the similar equations on functions

Φj0 (ζ)(j = 2, 3, 4), so the solution is similar to (11.3-50).
11.3.10 Example 3 Infinite plane with an elliptic hole subjected to a

distributed pressure at a part of surface of the hole
The problem is shown in Fig.11.3-4, but the conformal mapping is different to that
of example 1 and example 2.
Fig. 11.3-4 Infinite plane with an elliptic hole subjected to a distributed pressure at a
part of surface of the hole and its conformal mapping at ζ-plane, in which the region at
z-plane is mapped onto the exterior of the unit circle at ζ-plane.
11.4 Complex variable function method for sextuple harmonic equation and... 243
In terms of the similar procedure we find the solution[9] as follows

⎧
⎪ 1 p mR0 σ2 σ2 − ζ
⎪
⎪ Φ 4 (ζ) = · · − ln + z ln + z1 ln(σ 1 − ζ) − z2 ln(σ 2 − ζ)
⎪
⎪ 32c1 2πi ζ σ1 σ1 − ζ
⎪
⎪
⎪
⎪ + ip(d1 − d2 )(z1 − z2 ) ln ζ
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ 1 p (1 + m2 )R0 ζ σ2 R0 (σ1 − σ2 )(1 + mζ 2 )
⎪
⎪ Φ 3 (ζ) = · · − ln +
⎪
⎪ 32c1 2πi (ζ 2 − m) σ1 (ζ 2 − m)
⎪
⎪
⎪
⎪
⎪
⎪ − z2 ln(σ2 − ζ) + z1 ln(σ1 − ζ)
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ (1 + m2 )
⎪
⎪ − ip(d 1 + d 2 ) · (z1 − z 2 ) ln ζ + (z 1 − z2 ) ,
⎨ ζ2 − m

⎪ 1 pR0 (mζ 2 + 1)(ζ 2 + m) σ2 σ2 − σ1
⎪
⎪ Φ 2 (ζ) = · · ln +
⎪
⎪ 32c1 2πi (ζ 2 − m)3 σ1 (σ2 − ζ)(σ1 − ζ)
⎪
⎪
⎪
⎪ 1 p (mζ 2
+ 1) σ2 − σ1
⎪
⎪ · · 2 × 2Re z2 ·
⎪
⎪
+
− 2 − ζ)(σ1 − ζ)
⎪
⎪ 32c
1
2πi

(ζ m)

(σ 2

⎪
⎪
⎪
⎪ m (σ2 − ζ)(σ1 − ζ) + (σ2 + σ1 − 2ζ)(σ2 − σ1 )
⎪
⎪ + z 2 − R 0 ζ − ·
⎪
⎪ ζ (σ2 − ζ)(σ1 − ζ)
⎪
⎪
⎪
⎪ (mζ 2
+ 1)(ζ 2
+ m) 1
⎪
⎪ ip d1 (z1 − z2 − z1 + z2 )
⎪
⎪ (ζ − m)
2 3 ζ − σ1
⎪
⎪
⎪
⎪
⎪
⎪ 1 1 1
⎩ + (d2 − d1 )(z1 − z2 ) 2 + +
ζ ζ (ζ − σ1 )2
(11.3-53)
where

m m
z1 = R σ1 + , z2 = R σ2 + .
σ1 σ2
The detail is given in Major appendix of this book.
11.4 Complex variable function method for sextuple harmonic

equation and applications to icosahedral quasicrystals
Plane elasticity of icosahedral quasicrystals has been reduced to a sextuple harmonic
equation to solve in Chapter 9, where we have shown the solution procedure of the
equation for a notch/crack problem by complex variable function method, we here
provide further discussion in-depth from view-point of complex function theory. The
aim is to develop the complex potential method for higher order multi-harmonic
equations. Though there are some similar natures in the following description with
that introduced in the previous section, the discussion here is necessary, because the
governing equation and boundary conditions for icosahedral quasicrystals are quite
different from those for decagonal quasicrystals.

In Section 9.5 by the stress potential we obtain the final governing equation under
the approximation R2 /μK1 1
∇2 ∇2 ∇2 ∇2 ∇2 ∇2 G = 0. (11.4-1)
Fundamental solution of equation (11.4-1) can be expressed in six analytic functions
of complex variable z, i.e.,
G(x, y) = Re[g1 (z) + z̄g2 (z) + z̄ 2 g3 (z) + z̄ 3 g4 (z) + z̄ 4 g5 (z) + z̄ 5 g6 (z)], (11.4-2)
where gi (z) are arbitrary analytic functions of z = x + iy, the bar denotes the
complex conjugate.
By equations (11.4.1),(11.4-2) and (9.5-2), (9.5-3), the stresses can be expressed
as follows:
⎧
⎪
⎪ σxx + σyy = 48c2 c3 RImΓ (z), σyy − σxx + 2iσxy = 8ic2 c3 R(12Ψ (z) − Ω (z)),
⎪
⎪
⎪
⎪ 24λR
⎪
⎪ σzy − iσzx = −960c3 c4 f6 (z), σzz = c2 c3 ImΓ (z),
⎪
⎪ (μ + λ)
⎪
⎪
⎪ H − H − i(H + H ) = −96c c Ψ (z) − 8c c RΩ (z),
⎪
⎨ xy yx xx yy 2 5 1 2
⎪
⎪ Hyx + Hxy + i(Hxx − Hyy ) = −480c2 c5 f6 (z) − 4c1 c2 RΘ (z),
⎪
⎪
⎪
⎪
⎪
⎪ Hyz + iHxz = 48c2 c6 Γ (z) − 4c2 R2 (2K2 − K1 )Ω (z),
⎪
⎪
⎪
⎪ 24R2
⎪
⎪
⎩ Hzz = c2 c3 ImΓ (z),
(μ + λ)
(11.4-3)
where
⎧
⎪ Ψ (z) = f5 (z) + 5z̄ f6 (z),
⎪
⎪
⎪
⎪
⎪
⎪ Γ (z) = f4 (z) + 4z̄ f5 (z) + 10z̄ 2 f6 (z),
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ Ω (z) = f3 (z) + 3z̄ f4 (z) + 6z̄ 2 f5 (z) + 10z̄ 3 f6 (z),
⎪
⎪
⎪
⎪
⎪
⎪ (IV)
Θ(z) = f2 (z) + 2z̄ f3 (z) + 3z̄ 2 f4 (z) + 4z̄ 3 f5 (z) + 5z̄ 4 f6
⎪
⎪
(z),
⎪
⎪
⎪
⎪
⎪
⎪ R(2K2 − K1 )(μK1 + μK2 − 3R2 )
⎨ c1 = ,
2(μK1 − 2R2 ) (11.4-4)
⎪
⎪ 1
⎪
⎪ c3 = K2 (μK2 − R2 ) − R(2K2 − K1 ),
⎪
⎪ R
⎪
⎪
⎪
⎪ (μK2 − R2 )2
⎪
⎪
⎪
⎪ c2 = μ(K1 − K2 ) − R2 − ,
⎪
⎪ μK1 − 2R2
⎪
⎪
⎪
⎪ 1 μK1 − 2R2
⎪
⎪ c4 = c1 R + c3 K1 +
⎪
⎪ 2 λ+μ
⎪
⎪
⎪
⎪ μK2 − R2
⎪
⎩ c5 = 2c4 − c1 R, c6 = (2K2 − K1 )R2 − 4c4 .
μK1 − 2R2
In the above expressions,the function g1 (z)is not used and to be assumed g1 (z) = 0
so f1 (z) = 0 for simplicity, we have introduced the following new symbols:
(9) (8) (7)
g2 (z) = f2 (z), g3 (z) = f3 (z), g4 (z) = f4 (z),
(6) (5)
(11.4-5)
g5 (z) = f5 (z), g6 (z) = f6 (z),
(n)
where gi denotes n-th derivative with the argument z. Similar to manipulation in
the previous section, the complex representations of displacement components can
be written as follows (here we have omitted the rigid body displacements):
⎧
⎪
⎪ 2c2
⎪
⎪ u y + iu x = −6c 3 R + c 7 Γ (z) − 2c3 c7 RΩ (z),
⎪
⎪ μ+λ
⎪
⎪
⎪
⎪ 4
⎪
⎪ uz = (240c10 Imf6 (z)
⎪
⎪ μ(K + K 2 ) − 3R
2
⎪
⎪
1
⎪
⎨ +c1 c2 R2 Im(Θ(z) − 2Ω (z) + 6Γ (z) − 24Ψ (z))),
(11.4-6)
⎪
⎪ R
⎪
⎪ wy + iwx = − c1 (μK1 − 2R2 ) (24c9 Ψ (z) − c8 Θ(z)),
⎪
⎪
⎪
⎪
⎪
⎪ 4(μK2 − R2 )
⎪
⎪ wz =
⎪
⎪ (240c10 Imf6 (z)
⎪
⎪ (K1 − 2K2 )R(μ(K1 + K2 ) − 3R2 )
⎪
⎩
+c1 c2 R2 Im(Θ(z) − 2Ω (z) + 6Γ (z) − 24Ψ (z))),
in which
c2 K1 + 2c1 R
c7 = , c8 = c1 c3 R(μ(K1 − K2 ) − R2 ),
μK1 − 2R2
(μK2 − R2 )2
c9 = c8 + 2c3 c4 c2 − , c10 = c1 c3 R2 − c4 (c3 R − c2 K1 ).
μK1 − 2R2
(11.4-7)
11.4.2 The complex representation of boundary conditions

The boundary conditions of plane elasticity of icosahedral quasicrystals can be ex-
pressed as follows:
σxx l + σxy m = Tx , σyx l + σyy m = Ty , σzx l + σzy m = Tz , (11.4-8)
Hxx l + Hxy m = hx , Hyx l + Hyy m = hy , Hzx l + Hzy m = hz (11.4-9)

for (x, y) ∈ L, which represents the boundary of a multi-connected quasicrystalline
material, and
dy dx
l = cos(n, x) = , m = cos(n, y) = − ,
ds ds
T = (Tx , Ty , Tz ) and h = (hx , hy , hz ) denote the surface traction vector and gener-
alized surface traction vector, n represents the outward unit normal vector of any
point of the boundary, respectively.
Utilizing equation (11.4-3) and the first two formulas of equations (11.4-8) ,one has
−4c2 c3 R[3(f4 (z)+4z̄f5 (z)+10z̄ 2 f6 (z))−(f3 (z)+3zf4 (z)+6z 2 f5 (z)+10z 3 f6 (z))]
-
= i (Tx + iTy )ds, z ∈ L.
(11.4-10)
Taking conjugate on both sides of equation (11.4-10) yields
−4c2 c3 R[3(f4 (z)+4zf5 (z)+10z 2 f6 (z))−(f3 (z)+3z̄f4 (z)+6z 2 f5 (z)+10z̄ 3 f6 (z)
-
= −i (Tx − iTy )ds, z ∈ L.
(11.4-11)
Similarly, from equations (11.4-3) and the first two formulas of (11.4-9) ,one obtains

48c2 (2c4 − c1 R)Ψ (z) + 2c1 c2 RΘ(z) = i (hx + ihy )ds, z ∈ L. (11.4-12)
Furthermore we assume
Tz = hz = 0, (11.4-13)
for simplicity and by the third equations in (11.4-8) and (11.4-9) and the formulas
of (11.4-3) and (11.4-13),one has

f6 (z)+f6 (z) = 0,
4c11 Re[f5 (z)+5z̄f6 (z)]+(2K2 −K1 )RRe[f4 (z)+4z̄f5 (z)+10z̄ 2 f6 (z)+20f6 (z)] = 0,
z∈L (11.4-14)
in which
4c4 (μK2 − R2 )
c11 = (2K2 − K1 )R − . (11.4-15)
(μK1 − 2R2 )R
As we have shown in the previous section, complex analytic functions (i.e., the
complex potentials) must be determined by boundary value equations, and the dis-
cussion is as below.
11.4.3 Structure of complex potentials

11.4.3.1 The arbitrariness of the complex potentials
For explicit description, equation (11.4-3) can be written as follows:

⎧
⎪ σzy − iσzx = −960c3 c4 f6 (z),
⎪
⎪
⎪
⎪ c1 (σyy − σxx − 2iσxy ) + ic2 [Hxy − Hyx + i(Hxx + Hyy )] = −192ic2 c3 c4 Ψ (z),
⎪
⎪
⎪
⎨ 2c1 (Hzy + iHzx ) − R(2K2 − K1 )[Hxy − Hyx + i(Hxx + Hyy )]
⎪
⎪ = 96c3 cR(2K2 − K1 )Ψ (z) + 96c1 c3 c6 Γ (z),
⎪
⎪
⎪
⎪ c5 (σyy − σxx + 2iσxy ) + ic2 R[Hxy − Hyx − i(Hxx + Hyy )] = −16ic2 c3 c4 Ω (z),
⎪
⎪
⎩
Hyx + Hxy + i(Hxx − Hyy ) = −480c2 c5 f6 (z) − 4c1 c2 RΘ (z).
(11.4-16)
Similar to the discussion of two-dimensional quasicrystal, from the equations,it is

obvious that a state of phonon and phason stresses is not altered, if one replaces
fi (z) by fi (z) + γi , i = 2, · · · , 6, (11.4-17)
where γi are arbitrary complex constants.

Now consider how these substitutions affect the components of the displacement
vectors which were determined by the formula (11.4-6). Substituting (11.4-13) into
(11.4-8)∼(11.4-12) shows that if the complex constants γi (i = 2, · · ·, 6) satisfy
⎧
⎪
⎪ 2c2
⎪ 3
⎪ + c7 γ4 + c7 γ3 = 0,
⎨ μ+λ
24c9 γ5 − c8 γ2 = 0, (11.4-18)
⎪
⎪
⎪
⎪ c9 2c 2
⎩ 40c10 γ6 − c1 c3 R2 4 1 − γ5 − γ4 = 0,
c8 (μ + λ)c7
then the substitution (11.4-17) will not affect the displacements.
11.4.3.2 General formulas for finite multiply connected region
Consider now the case when the region S, occupied by the body, is multiply con-
nected,see Fig 11.3-1.
Since the stress must be single-valued and equation (11.4-16)
σzy − iσzx = −960c3 c4 f6 (z). (11.4-19)
we know that f6 (z)is holomorphic and hence single-valued in the region with contour
sm+1 , so the complex function can be expressed as
z
f6 (z) = f6 (z)dz + constant, (11.4-20)
z0
where z0 denotes fixed point. By equation(11.4-20), we have
f6 (z) = bk ln(z − zk ) + f6∗ (z), (11.4-21)
f6∗ (z) is holomorphic in the region with contour sm+1 .

Substituting (11.4-21) into the second formula of equations (11.4-16), i.e.,
c1 (σyy − σxx − 2iσxy ) + ic2 [Hxy − Hyx + i(Hxx + Hyy )] = −192ic2 c3 c4 Ψ (z)
shows that f5 (z) is holomorphic in the region enclosed by contour sm+1 , so one has
f5 (z) = ck ln(z − zk ) + f5∗ (z), (11.4-22)
f5∗ (z) is holomorphic in the region of interior of contour sm+1 .

Similar to the above mentioned discussion, by equations (11.4-16)∼(11.4-18), the

complex functions fi (i = 2, 3, 4) can be written as
⎧
⎨ f4 (z) = dk ln(z − zk ) + f4∗ (z),
f3 (z) = ek ln(z − zk ) + f3∗ (z), (11.4-23)
⎩
f2 (z) = tk ln(z − zk ) + f2∗ (z)
where dk , ek and tk are complex constants, fi∗ (z)(i = 2, 3, 4) is holomorphic in the
region with contour sm+1 .
By substituting (11.4-21)∼(11.4-23) into the complex expressions of displace-
ments, the condition of single-valuedness of displacements will be given as follows:
⎧
⎪
⎪ 2c2
⎪
⎨ −3 + c7 dk + c7 ek = 0,
μ+λ
(11.4-24)
⎪ 24c9 ck + c8 tk = 0,
⎪
⎪
⎩
240c10 bk + c1 c3 R2 (tk − 2ek + 6dk − 24ck ) = 0.s
Applying the boundary conditions given above to the contour sk and by equation
(11.4-24), we know that the above complex contants may be very simply expressed
in terms of surface traction and generalized surface traction as
⎧
⎪ c1 c3 R2 12c2 c9
⎪
⎪ bk = dk + 24 1 + ck ,
⎪
⎪ 240c10 (μ + λ)c7 c8
⎪
⎪
⎪
⎪ c8
⎪
⎪ c = (hx − ihy ),
⎪ k
⎪ −96π[c 3 c8 (2c 4 − c1 R) − c1 c3 R]
⎪
⎨ c8
tk = (hx + ihy ), (11.4-25)
⎪ 4π[c3 c8 (2c4 − c1 R) − c1 c3 R]
⎪
⎪
⎪
⎪ (μ + λ)c7
⎪
⎪ dk = (Tx + iTy ),
⎪
⎪ 24πc c R(2c 2 + (μ + λ)c7 )
⎪
⎪ 2 3
⎪
⎪ 2c + (μ + λ)c7
⎪
⎩ ek = − 2 (Tx − iTy ).
16πc22 c3 R
We can easily extend the above results to the case there are m inner boundaries.
11.4.4 Case of infinite regions

From the point of view of application the consideration of infinite regions is likewise
of major interest. We assume that the contour sm+1 has entirely moved to infinity.
Similar to the discussion of two-dimensional quasicrystal, we have
⎧ m m
⎪
⎪
⎪
⎪ f6 (z) = b k ln z + f6∗∗ (z), f 5 (z) = ck ln z + f5∗∗ (z),
⎪
⎪
⎪
⎪ k=1 k=1
⎨ m m
f4 (z) = dk ln z + f4∗∗ (z), f3 (z) = ek ln z + f3∗∗ (z), (11.4-26)
⎪
⎪
⎪
⎪ k=1 k=1
⎪
⎪ m
⎪
⎪ f2 (z) =
⎩ tk ln z + f2∗∗ (z),
k=1
fi∗∗ (z)(i = 2, · · · , 6) are functions, holomorphic outside sm+1 , not including the
point at infinity. By the theorem of Laurent expansion, the function h2∗ (z) may be
represented outside sm+1 by the series
+∞

fi∗∗ (z) = ain z n (i = 2, · · · , 6.) (11.4-27)
n=−∞
Substituting the first equation of (11.4-26) and (11.4-27) into equation (11.4-16),one
has $m ∞
%
1
σzy − iσzx = −960c3 c4 bk + na6n z n−1 , (11.4-28)
z −∞
k=1
hence it follows that for the stress to remain finite as |z| → ∞, one must have
a6n = 0, n 2. (11.4-29)
Similarly, from equations (11.4-15)∼(11.4-18), to make the stress be bounded, the
following conditions are also be satisfied
ain = 0, n 2, (i = 2, · · · , 6.) (11.4-30)
So we can obtain the expressions of the complex function fi (z)(i = 2, · · · , 6) for the
stress to remain finite as |z| → ∞,for example

m
f6 (z) = bk ln z + (B + iC)z + f60 (z), (11.4-31)
k=1
where B, C are unknown real constants to be determined, f60 (z) is function, holomor-
phic outside sm+1 , including the point at infinity. The determination of unknown
contants B, C etc is similar to that given in subsection 11.3.4, but the details are
omitted here due to limitation of the space.
11.4.5 Conformal mapping and function equations at ζ- plane

We now have five equations of boundary value (11.4-10)∼(11.4-12) and (11.4-14),
from which the unknown functions fi (z)(i = 2, · · · , 6)will be determined, in addition
we have assumed that f1 (z) = 0, because it has no usage. For some complicated
regions the function equations cannot be directly solved at the physical plane (i.e.,
the z - plane), and the conformal mapping is particular meaningful in the case.
Assume that a conformal mapping
z = ω(ζ) (11.4-32)
is used to transform the region at z- plane onto the interior of the unit circle γ at
ζ- plane. Under the mapping the unknown functions fi (z) become
fi (z) = fi [ω(ζ)] = Φi (ζ) i = 2, 3, · · · , 6. (11.4-33)

Substituting (11.4-32) and (11.4-33) into the first relation of boundary condition
(11.4-14) and multiplying dσ/2π(σ−ζ) to both sides of the equation, then integrating
along the unit circle γ and |ζ| < 1 yields

1 Φ6 (σ) 1 Φ6 (σ)
dσ + dσ = 0.
2πi σ−ζ 2πi σ−ζ
γ γ
This shows
Φ6 (ζ) = 0 (11.4-34)
according to the Cauchy integral formula.
Substitution of (11.4-32)∼(11.4-34) into boundary conditions (11.4-10)∼
(11.4-12) and the second one of conditions (11.4-14) and taking manipulation sim-
ilarly to above leads to boundary value equations for determining the unknown
functions Φi (ζ)(i = 2, · · · , 5) at ζ - plane, i.e.,

3 Φ4 (σ) 4 ω(σ) Φ5 (σ) 1 Φ3 (σ)
dσ + (σ) σ − ζ
dσ − dσ
2πi σ − ζ 2πi ω 2πi γ σ−ζ
γ

γ

1 ω(σ) Φ4 (σ) 1 [ω(σ)]2 Φ5 (σ) [ω(σ)]2 ω (σ) dσ
−3 dσ − 6 2 − 3 Φ5 (σ)
2πi γ ω (σ) σ − ζ 2πi γ ω (σ) ω (σ) σ −ζ

1 1 t
= dσ,
4c2 c3 2πi γ σ − ζ
(11.4-35)

3 Φ4 (σ) 4 ω(σ) Φ4 (σ) 1 Φ3 (σ)
dσ + dσ − dσ
2πi γ σ − ζ 2πi γ ω (σ) σ − ζ 2πi γ σ − ζ
2 2
1 ω(σ) Φ3 (σ) 1 ω(σ) Φ5 (σ) ω(σ) ω (σ)Φ5 (σ) dσ
−3 dσ − 6 i −
2πi γ ω (σ) σ − ζ 2π γ [ω (σ)]2 [ω (σ)]3 σ−ζ

1 1 t
= dσ,
4c2 c3 R 2πi γ σ − ζ
(11.4-36)

1 Φ2 (σ) 1 ω(σ) Φ3 (σ)
dσ + 2 dσ+
2 2 2
1 ω(σ) Φ4 (σ) ω(σ) ω (σ)Φ4 (σ) dσ 1 ω(σ) Φ 5 (σ)
3 − + 4
2πi γ [ω (σ)]2 [ω (σ)]3 σ−ζ 2πi γ [ω (σ)]3
3 3 3
ω(σ) ω (σ)Φ5 (σ) ω(σ) ω (σ)Φ5 (σ) ω(σ) ω (σ)Φ5 (σ) dσ
−3 4
+ 3 −
[ω (σ)] [ω (σ)]5 [ω (σ)]4 σ−ζ
1 h
= dσ,
2πi σ−ζ
(11.4-37)

4c11 Φ5 (σ) (2K2 − K1 )R Φ4 (σ) ω(σ) Φ5 (σ)
dσ + +4 dσ = 0, (11.4-38)
2πi γ σ − ζ 2πi γ σ−ζ ω (σ) σ − ζ

in which t = i (Tx + iTy )ds, t = −i (Tx − iTy )ds, h = i (h1 + ih2 )ds. For
given configuration and applied stresses we can obtain the solution by solving these
function equations.
11.4.6 Example: Elliptic notch problem and solution

We consider an icosahedral quasicrystal solid with an elliptic notch, which penetrates
through the medium along the z-axis direction,the edge of the elliptic notch subjected
to the uniform pressure p, similar to Fig11.3-2.
Since the measurement of generalized traction has not been reported so far, for
simplicity, we assume that hx = 0, hy = 0.
However the calculation can not be completed at the z-plane owing to the com-
plicity, we have to employ the conformal mapping

1
z = ω(ζ) = R0 + mζ (11.4-39)
ζ
to transform the exterior of the ellipse at the z-plane onto the interior of the unit
circle γ at the ζ− plane, in which
R0 = (a + b)/2, m = (a − b)/(a + b) (11.4-39)
Substituting (11.4-39) and( 11.4-34) into (11.4-35)-(11.4-38), considering
ω(σ) σ2 + m
= σ
ω (σ) mσ 2 − 1
and
ζ2 + m ζ2 + m
ζ Φ 5 (ζ) = ζ (α1 + 2α2 ζ + 3α3 ζ 2 + · · ·)
mζ 2 − 1 mζ 2 − 1

1 Φ4 (σ)
dσ = Φ4 (ζ)
2πi σ−ς
γ
are analytic in |ζ| < 1 and continuous in the unit circle γ, by means of Cauchy
integral formula, from equation(11.4-34), we have

1 σ 2 + m Φ5 (σ) ζ2 + m
σ dσ = ζ Φ (ζ)
2πi mσ 2 − 1 σ − ς mζ 2 − 1 5
γ
Substituting
ω(σ) 1 mσ 2 + 1 ω(σ)2 ω (σ) 2σ(mσ 2 + 1)2

=− , 3 =
ω (σ) σ σ2 − m ω (σ) (σ 2 − m)3
into equation(11.4-34), and note that

1 mζ 2 + 1 1 mζ 2 + 1 β2 β3
− Φ (ζ) = − β1 + 2 + 3 2 + · · ·
ζ ζ2 − m 4 ζ ζ2 − m ζ ζ

2ζ(mζ 2 + 1)2 2ζ(mζ 2 + 1)2 α2 α3
Φ 5 (ζ) = α1 + 2 +3 2 +···
(ζ 2 − m)3 (ζ 2 − m)3 ζ ζ
are analytic in |ζ| > 1 and continuous in the unit circle γ, by means of Cauchy
integral formula and analytic extension of the complex variable function theory,
from equation(11.4-34), we obtain

1 Φ3 (σ) 1 ω(σ) Φ4 (σ)
dσ = 0, dσ = 0
2πi σ−ς 2πi ω (σ) σ − ς
γ γ

1 ω(σ) Φ5 (σ) ω(σ)2 ω (σ)
2
dσ
2 − 3 Φ5 (σ) =0
2πi ω (σ) ω (σ) σ −ς
γ
Substituting the above results into equation(11.4-34), with the help of equation
(11.4-38),one has
R0 (2K2 − K1 )R0 pmζ(ζ 2 + m)

Φ4 (ζ) = pmζ −
12c2 c3 R 2c2 c3 C 11 (mζ 2 − 1) (11.4-36)
Φ5 (ζ) = − (2K 2 −K1 )R0
48c2 c3 C 11 pmζ
Similar to the above discussion, by equations (11.4-36) and (11.4-37), one

has
R0 pζ(ζ 2 + m)(m3 ζ 2 + 1)
Φ2 (ζ) = −
2c2 c3 R (mζ 2 − 1)3
(2K2 −K1 )R0 pmζ 3 (ζ 2 +m)[m2 ζ 6 −(m3 +4m)ζ 4 +(2m4 +4m2 +5)ζ 2 +m]
+
2c2 c3 C 11 (mζ 2 −1)5
R0 pζ(m2 + 1) (2K2 − K1 )R0 pmζ 3 (ζ 2 + m)(mζ 2 − m2 − 2)
Φ3 (ζ) =− −
4c2 c3 R (mζ 2 − 1) 12c2 c3 C 11 (mζ 2 − 1)3
(11.4-41)
The elliptic notch problem is solved. The solution of the Griffith crack subjected
to a uniform pressure can be obtained corresponding to the case m = 1, R0 = a/2 of
the above solution. The solution of crack can be expressed explicitly in the z−plane,
the concrete results refer to Section 9.7 in Chapter 9.
11.5 Complex analysis and solution of quadruple quasiharmonic equation 253
11.5 Complex analysis and solution of quadruple quasihar-

monic equation
In Chapters 6∼8 we have known that the plane elasticity of octagonal quasicrystals
is governed by the final equation
(∇2 ∇2 ∇2 ∇2 − 4ε∇2 ∇2 Λ2 Λ2 + 4εΛ2 Λ2 Λ2 Λ2 )F = 0 (11.5-1)
either by displacement potential or by stress potentail, in which

⎧
⎪ ∂2 ∂2 ∂2 ∂2
⎪
⎨ ∇ =
2
+ , Λ 2
= − ,
∂x2 ∂y2 ∂x2 ∂y2
(11.5-2)
⎪
⎪ R2 (L + M )(K2 + K3 )
⎩ ε=
[M (K1 + K2 + K3 ) − R2 ][(L + 2M )K1 − R2 ]
Due to the appearance of operator Λ2 it seems there is no any connection with

complex variable functions in solving equation (11.5-1). But if we rewrite it as
8
∂ ∂8 ∂8 ∂8 ∂8
+ 4(1 − 4ε) 6 2 + 2(3 + 16ε) 4 4 + 4(1 − 4ε) 2 6 + 8 F = 0,
∂x8 ∂x ∂y ∂x ∂y ∂x ∂y ∂y
(11.5-3)
then find that this is one of typical multi-quasiharmonic partial differential equation
with 8th order, and there is complex representation of solution such as
4

F (x, y) = 2Re Fk (zk ), zk = x + μk y, (11.5-4)
k=1
in which functions Fk (zk ) are analytic functions of complex variable zk (k = 1, 2, 3, 4),

and μk = αk + iβk (k = 1, 2, 3, 4) are complex parameters and determined by the
roots of the following eigenvalue equation
μ8 + 4(1 − 4ε)μ6 + 2(3 + 16ε)μ4 + 4(1 − 4ε)μ2 + 1 = 0. (11.5-5)
We have shown in Chapters 7 and 8 some solutions of dislocations (based on dis-

placement potentail formulation) and notches/cracks (based on stress potential for-
mulation) can be found in terms of this complex analysis. In the procedure it must
carry out some calculations on determinants of 4th order, so the solution expressions
are quite lengthy, but which are analytic substantively.

The discovery of quadruple and sextuple harmonic equations is interesting for the
elasticity. This chapter gives a comprehensive discussion on the complex analysis
for solving the equations, we think the study is preliminary.
In some sense, the complex potential approach above mentioned is a new devel-
opment of Muskhelishvili approach of the classical plane elasticity.
Apart from the development to the complex potential theory and method, the
conformal mapping technique has also been extended. According to the monograph
[1], application of the conformal mapping is limited within the rational function
class. But it can be extended into the transcendental function class, and some exact
analytic solutions for more complicated cracked configurations are achieved, see e.g.
Chapter 8.
The method is effective not only for solving elasticity problems but also for
solving some simple plasticity problems, see e.g. Fan and Fan[10] , Li and Fan[11,12]
and Fan and Tang[13] .
References
[1] Muskhelishvili N I. Some Basic Problems of the Mathematical Theory of Elasticity.
Groningen: P Noordhoff., 1956
[2] Lekhnitskii S G. Theory of Elasticity of an Anisotropic Body. Holden-Day. San
Francisco, 1963
[3] Liu G T, Fan T Y, Guo R P. Governing equations and general solutions of plane
elasticity of one-dimensional quasicrystals. Int J Solid and Structures, 2004, 41(14):
3949−3959
tals and defects and auxiliary equation method for solving some nonlinear evolution
equations. Dissertation. Beijing Institute of Technology, 2004 (in Chinese)
[5] Liu G T, Fan T Y. The complex method of the plane elasticity in 2D quasicrystals
point group 10mm ten-fold rotation symmetry notch problems. Science in China E,
2003, 46(3): 326−336
[6] Li L H, Fan T Y. Final governing equation of plane elasticity of icosahedral quasicrystals–
stress potential method. Chin Phys Lett, 2006, 24(9): 2519−2521
[7] Li L H, Fan T Y. Complex function method for solving notch problem of point group
10 and 10 two-dimensional quasicrystal based on the stress potential function, J.
Phys.: Condens. Matter, 2006, 18(47): 10631−10641
[8] Li L H, Fan T Y. Complex function method for notch problem of plane elasticity of
icosahedral quasicrystals, Science in China, G, 2008, 51(6): 773−780
[9] Li W, Fan T Y. Study on elastic analysis of crack problem of two-dimensional decago-
nal quasicrystals of point group 10,10. Mod Phys Lett, B, 2009, 23(16), 1989–1999
[10] Fan T Y, Fan L. Plastic fracture of quasicrystals. Phil Mag, 2008, 88(4): 321−335
[11] Li W, Fan T Y. Study on plastic analysis of crack in decagonal Al-Ni-Co quasicrystals.
Mod. Phys. Lett, B, in press, 2009
References 255
[12] Li W, Fan T Y. Plastic solution of crack in three-dimensional Al-Pd-Mn quasicrystals,

Phil.Mag., 2009, 89(31), 2823–2831
[13] Fan T Y, Tang Z Y, 2009, The strict theory of complex variable function method of
sextuple harmonic equation and applications, J. Math. Phys., accepted, 2010
Chapter 12
Variational principle of elasticity of
quasicrystals, numerical analysis and
applications
From Chapter 5 to Chapter 11, we developed mainly analytic theories and meth-
ods. The elasticity problems of quasicrystals were reduced to boundary value or
initial-boundary value problems of some partial differential equations to solve, in
which some analytic methods were used. For some boundary value problems, these
methods are extremely powerful, even capable of obtaining exact analytic solutions.
In Chapter 14 and Major Appendix we will further develop the analytic method in
studying some problems such as nonlinear deformation etc.
However there are limitations themselves for these analytic methods. In general,
they can only treat some problems with simple configurations and simple boundary
conditions; while for more complicated problems, the methods cannot display their
power, one must use numerical methods.
In Chapter 10 we have applied the finite difference method to determining the
numerical solutions for some elasto-/hydro-dynamics problems of quasicrystals. In
this chapter we derive a variational principle of elasticity of quasicrystals, which
is the foundation of the subsequent development of the finite element method for
problems of quasicrystals. Discretization is the main feature of finite difference
method and finite element method. It has been shown that solutions obtained by
these two schemes can approach to the analytic solutions as the discreted mesh
(or element) tends to infinitesimal. In the other hand, in contrast to the analytic
solutions (or classical solutions), finite element solutions belong to one kind of so-
called weak solutions (or generalized solutions) according to the modern theory
of partial differential equations. The further mathematical principle on the weak
solutions of elasticity of quasicrystals will be developed in Chapter 13, which will
help us to understand that the finite element method is one of important tools to
implement weak solutions from other angles.

258 Chapter 12 Variational principle of elasticity of quasicrystals, numerical analysis...
12.1 Basic relations of plane elasticity of two-dimensional

quasicrystals
Here we only consider the two-dimensional quasicrystals, and further assume that
along the periodic direction, i.e., along the x3 -axis all field variables are uniform, so
that
∂
= 0. (12.1-1)
∂x3
This assumption and the fact
w3 = 0 (12.1-2)
reduce the elasticity to a plane phonon-phason elasticity and an anti-plane pure

phonon elasticity, respectively, and the solution is greatly simplified. In the following
only the plane elasticity is considered.
In this case, the generalized Hooke’s law stands for

(12.1-3)
where
Cijkl = Lδij δkl + M (δik δjl + δil δjk ), (12.1-4)
C11 − C12
L = C12 , M= , (12.1-5)
2
Kijkl = (K1 − K2 − K3 )δik δjl + K2 δij δkl + K3 δil δ
(12.1-6)
+2(K2 + K3 )(δi1 δj2 δk1 δl2 + δi2 δj1 δk2 δl1 ),
Rijkl = R1 (δi1 − δ)(δij δkl − δik δjl + δil δjk )
+R2 [(1 − δij )δkl + δij (δi1 − δi2 )(δk1 δl2 − δk2 δl1 )] i, j, k, l = 1, 2.
(12.1-7)
and the definitions about C12 , C12 (orL, M ), K1 , K2 , K3 , R1 and R2 have been dis-
cussed in detail in Chapters 6∼8.
For the points group 10mm, R1 > 0, R2 = 0, K1 > 0, K2 > 0, K2 + K3 = 0; for
the point group 8mm, R1 > 0, R2 = 0, K1 > 0, K2 > 0, K3 > 0; for the point group
10,10 R1 > 0, R2 > 0, K1 > 0, K2 > 0, K2 + K3 = 0; for the point group 12mm,
R1 = 0, R2 = 0,K1 > 0,K2 > 0,K3 > 0 (but some measured results show K2 < 0,
this does not influence the extreme value of energy functional, only influences the
minimum value of the functional in the proof for theorem of variational principle of
elasticity of quasicrystals, in Section 12.2).
The strains in above relevant formulas are denoted by
12.2 Generalized variational principle for static elasticity of quasicrystals 259

1 ∂ui ∂uj ∂wi
εij = + , wij = (12.1-8)
2 ∂xj ∂xi ∂xj
and stresses satisfy the equilibrium equations
∂σij ∂Hij
+ fi = 0, + gi = 0. (12.1-9)
∂xj ∂xj
The above formulas hold in any interior point of region Ω , and at boundary Γt , the
stresses satisfy the boundary conditions

σij nj = Ti ,
(x1 , x2 ) ∈ Γt , (12.1-10)
Hij nj = hi
and at boundary Γu the displacements satisfy the boundary conditions

ui = ūi ,
(x1 , x2 ) ∈ Γu , (12.1-11)
wi = w̄i ,
where Ti is the traction vector, hi the generalized traction vector at boundary Γt , ūi
and w̄i the given displacements at boundary Γu , ni the unit outward normal vector
at any point of the boundary and Γ = Γu + Γt .
In the following only the static problems are studied, and the initial value condi-
tions will not be concerned. For the dynamic problems, the initial value conditions
must be used, which have been discussed in Chapter 10.
12.2 Generalized variational principle for static elasticity of

quasicrystals
The variational principle in mathematical physics is one of basic principles, which
reveals that the extreme value (or stationary value) of energy functional of a system
is equivalent to the governing equations and the corresponding boundary value (or
initial-boundary value) conditions of the system. Accordingly solutions of initial-
boundary value problem of the partial differential equations can be converted to
determine the extreme value of the corresponding energy functional. And the latter
will be implemented by a discretization procedure, one among them is the finite
element method.
We here extend the minimum potential energy principle of classical elasticity to
describe elasticity of quasicrystals.
Theorem(Variational principle of elasticity of quasicrystals) For sufficient
smooth boundary, if all ui and wi satisfy equations of deformation geometry (12.1-8)
and displacement boundary conditions (12.1-11), let the energy functional of qua-
sicrystals

Π = F dΩ + (fi ui + gi wi )dΩ + (Ti ui + hi wi )dΓ (12.2-1)
Ω Ω Γt
to take a minimum value, then they will be the solution satisfying the equilibrium
equations (12.1-9) and the stress boundary conditions (12.1-10), in which F is defined
by
⎧ εij wij
⎪
⎪ F =
⎪
⎪ σ ij dε ij + Hij dwij = Fu + Fw + Fuw ,
⎪
⎪ 0 0
⎪
⎪
⎪
⎨ 1
Fu = Cijkl εkl ,
2 (12.2-2)
⎪
⎪
⎪
⎪ 1
⎪
⎪ Fw = 2 Kijkl wkl ,
⎪
⎪
⎪
⎩
Fuw = Rijkl εij wkl
or by (4.4-1), Ω is the region occupied by the quasicrystal and Γ is the boundary of

Ω.
In addition, the conditions in the theorem are sufficient and necessary.
Proof
(1) Necessity
Assume that the functional Π takes its extreme value, i.e.,δ Π = 0. From (12.2-
1),

∂F ∂F
δΠ = δεij + δwij − fi δui − gi δwi dΩ
Ω ∂εij ∂wij

+ (Ti δui + hi δwi )dΓ = 0 (12.2-3)
Γt
where
∂F ∂F
σij = , Hij = , (12.2-4)
∂εij ∂wij
which have been introduced in Chapter 4.

∂F
Noting that the suffixes of quantities are symmetric, therefore we obtain
∂εij

∂F ∂F ∂ui
δεij dΩ = δ dΩ .
Ω ∂εij Ω ∂εij ∂xj
Making use of the Green formula to above formula yields

12.2 Generalized variational principle for static elasticity of quasicrystals 261

∂F ∂ ∂F ∂ ∂F
δεij dΩ = δui dΩ − δui dΩ
Ω ∂εij ∂xj ∂εij ∂xj ∂εij
Ω Ω
∂F ∂ ∂F
= nj δui dΓ − δui dΩ .
Γu +Γt ∂εij Ω ∂x j ∂ε ij
Because the displacements are given at the boundary Γu , at which δui = δ ūi = 0,
the above formula has been reduced to

∂F ∂F ∂ ∂F
δεij dΩ = nj δui dΓ − δui dΩ . (12.2-5)
Ω ∂εij Γt ∂εij Ω ∂xj ∂εij
Due to wij = ∂wi /∂xj , a similar analysis to what adopted just above gives rise to

∂F ∂F ∂ ∂F
δwij dΩ = nj δwi dΓ − δui dΩ . (12.2-6)
Ω ∂wij Γt ∂wij Ω ∂xj ∂wij
Substituting (12.2-5) and (12.2-6) into (12.2-3) leads to

∂ ∂F ∂ ∂F
δ Π =− + fi δui + + gi δwi dΩ
∂x ∂εij ∂x ∂wij

Ω j j (12.2-7)
∂F ∂F
+ nj − Ti δui + nj − hi δwi dΓ = 0.
Γt ∂εij ∂w ij
Since δui and δwi are of arbitrary and independent variation at region Ω and bound-
ary Γ there follows

∂ ∂F ∂ ∂F
+ fi = 0, + gi = 0, (x1 , x2 ) ∈ Ω ,
∂xj ∂εij ∂xj ∂wij

∂F ∂F
nj − hi = 0, nj − Ti = 0, (x1 , x2 ) ∈ Γt .
∂wij ∂εij
Substituting (12.2-4) into the above formulas yields the equilibrium equations and
stress boundary conditions. This shows that ui and wi satisfying equations of defor-
mation geometry, stress-strain relations and displacement boundary conditions and
making energy functional to have minimum value, should be the solution satisfying
the equilibrium equations and stress boundary conditions.
(2) Sufficiency
The sufficiency of the conditions given by the theorem means that, if among
all sets of the displacements ui and wi satisfy relations of deformation geometry,
displacement boundary, the one that satisfies the equilibrium equations and stress
boundary conditions makes in (12.2-1) an extremum. In what follows, we show that
Π is minimum.
Suppose that quantities ui , εij , wi and wij obey the stress-strain relations
(12.1-3) and satisfy the displacement boundary conditions (12.1-13). Let

∗
εij = εij + δεij , u∗i = ui + δui ,
∗ (12.2-8)
wij = wij + δwij , wi∗ = wi + δwi
through the displacement- strain relations, we have

⎧
⎨ δε = 1 (δu + δu ),
ij i,j j,i
2 (12.2-9)
⎩ δw = δw ,
ij i,j
where ui,j = ∂ui /∂xj etc.

The free energy F (ε∗ij , wij
∗
) can be expanded into the Taylor series as follows:
F (ε∗ij , wij
∗
) = F (εij + δεij , wij + δwij )
∂F ∂F
= F (εij , wij ) + δεij + δwij
∂εij ∂wij
1 ∂2F 1 ∂2F ∂2F
+ δεij δεkl + δwij δwkl + δεij δwkl +· · · .
2 ∂εij ∂εkl 2 ∂wij ∂wkl ∂εij ∂wkl
(12.2-10)
If the free energy is a homogeneous quantity of strain components of second order,
then the expansion of energy functional corresponding to (12.2-10) does not contain
terms higher than second order, i.e.,
Π ∗ = Π + δ Π + δ 2 Π + O(δ 3 ), (12.2-11)
in which

∂F ∂F
δΠ = δεij + δwij − fi δui − gi δwi dΩ − (Ti δui + hi δwi )dΓ ,
Ω ∂εij ∂wij Γt

(12.2-12)

2 1 ∂2F 1 ∂2F ∂2F
δ Π = δεij δεkl + δwij δwkl + δεij δwkl dΩ .
Ω 2 ∂εij ∂εkl 2 ∂wij ∂wkl ∂εij ∂wkl
(12.2-13)
Applying the Green formula to (12.2-12) leads to
δ Π = 0.
Because ui and wi (through the corresponding σij and Hij ) satisfy the equilibrium
equations and stress boundary conditions, this means that the energy functional
takes extreme value.
According to the discussion in Ref.[2], we do some extension, i.e., L + M >
0, M > 0, K1 > 0, K2 > 0, K3 > 0, M K1 > R2 ,then elasticity of two-dimensional
quasicrystals presents stability, in the case the stress-strain elastic matrix should be
positive definite, so
12.3 Finite element method 263
δ 2 Π > 0.
This ensures that δ Π = 0 takes not only an extreme value, but also the minimum
value of the energy functional.
But recent experimental results and some simulation show that there may be
K2 < 0, this does not influence the energy taking extreme value, but the extreme
value may not be minimum value.
Collaborating variational principle with theory of functional analysis, we can
prove that existence, uniqueness and stability of solution of the boundary value
problem (12.1-3),(12.1-8)∼(12.1-11).This is concerned not only with the numerical
implementation, but also with other topics, the detailed discussion of which will be
given in Chapter 13 or see Guo and Fan[4] .
The variational principle can be extended to dynamic case, in which it is needed
only to extend the energy functional (12.2-1) to be as follows:

..
Π = F dΩ + [(fi − ρ ui )ui + (gi − κẇi )wi )dΩ + (Ti ui + hi wi )dΓ , (12.2-14)
Ω Ω Γt
where the meaning of ρ and κ can be found in Chapter 10. From (12.2-14) we
can obtain the corresponding variational equation similar to (12.2-3), but which is
equivalent to the elasto-/hydro-dynamic equations and related boundary and initial
conditions of quasicrystals. The further discussion is omitted here.
12.3 Finite element method

A discretization of the variational equation and domain Ω yields the finite element
algorithm.
Dividing the quasicrystalline body studied into N elements, and we here use the
harmonic element of tetragon, i.e., we require that
(1) Within every element, displacements are continuous and single-valued;
(2) Along the inter-element boundaries, the displacements are continuous, i.e.,
(m) (m )
ui = ui
(m) (m )
(x1 , x2 ) ∈ Γ (mm ) , (12.3-1)
wi = ui ,

where Γ (mm ) represents the interface of element m and element m ;
(3) At the displacement boundary Γu ,the displacements should satisfy the con-
ditions such as
ui = ūi ,
(x1 , x2 ) ∈ Γu . (12.3-2)
wi = w̄i ,
By satisfying these conditions, the discrete form of the energy functional Π is given
by
N

(m) (m) (m) (m) (m) (m) (m) (m)
Π∗ = (F (m) −fi ui −gi wi )dΩ − (Ti ui +hi wi )dΓ .
(m)
m=1 Ω (m) Γt
(12.3-3)
Taking variation to (12.3-3) and letting it to be zero, by using the method similar
to that adopted in the previous section, we can show that within an element the
equilibrium equations are satisfied by
⎧ (m)
⎪
⎪ ∂ ∂F (m)
⎪
⎨ + fi = 0,
∂xj ∂εij
(m) (x1 , x2 ) ∈ Ωm , (12.3-4)
⎪
⎪ ∂ ∂F (m)
⎪
⎩ + gi = 0,
∂xj ∂wij
the stress boundary conditions are satisfied by

⎧
⎪
⎪ ∂F (m) (m) (m)
⎪
⎨ nj = T i ,
∂εij (m)
(m) (x1 , x2 ) ∈ Γt (12.3-5)
⎪
⎪ ∂F (m) (m)
⎪
⎩ nj = h i ,
∂wij
and at the interface between elements the displacement continuous conditions

(12.3-1) are satisfied.
The discrete energy functional (12.3-3) can also be written in the following matrix
form
N

∗ 1 (m) T (m) (m) (m) T (m)
Π = {u } {K }{u } − {u } {R } , (12.3-6)
m=1
2
where column vector {u(m) } represents the element displacement (including phonons
and phasons) vector, {u(m) }T is its transpose, the row vector, and {K (m) } the
element stiffness matrix , i.e.,

T
[K (m) ] = [B] [D][B]Δdx1 dx2 , (12.3-7)
Ωm
in which, Δ denotes the thickness of the element and Δ=1.0 for plane strain case
(i.e., the thick body, and we do not consider plane stress case, i.e., the thin body),
[B] = [B1 , B2 , · · · Bn ] = [L][I] (12.3-8)
is the element strain matrix , n is the number of nodes with differential operator
12.3 Finite element method 265
matrix, ⎡ ⎤
∂
⎢ 0 0 0 ⎥
⎢ ∂x ⎥
⎢ ∂
0 ⎥
⎢ 0 0 ⎥
⎢ ∂y ⎥
⎢ ⎥
⎢ ∂ ∂ ⎥
⎢ 0 0 ⎥
⎢ ∂y ∂x ⎥
⎢ ⎥
[L] = ⎢
⎢ 0 0
∂
0 ⎥
⎥ (12.3-9)
⎢ ∂x ⎥
⎢ ∂ ⎥
⎢ 0 0 0 ⎥
⎢ ∂y ⎥
⎢ ⎥
⎢ ∂ ⎥
⎢ 0 0 0 ⎥
⎢ ∂y ⎥
⎣ ∂ ⎦
0 0 0
∂x
[Bi ] = [L][Ii ] (12.3-9 )
and [I] = [I1 , I2 , · · · , In ] is the element interpolation matrix, n is the number of
nodes as pointed out just before, here n = 4, i.e.,
⎡ ⎤
Ii 0 0 0
⎢ 0 Ii 0 0 ⎥
[Ii ] = ⎢ ⎥
⎣ 0 0 Ii 0 ⎦ , i = 1, 2, 3, 4. (12.3-10)
0 0 0 Ii
where Ii is the i-th node interpolation function. In (12.3-7) matrix [D] denotes the
plane elastic constants matrix of two-dimensional quasicrystals, namely
⎡ ⎤
L + 2M L 0 R1 R1 R2 −R2
⎢ L L + 2M 0 −R1 −R1 −R2 R2 ⎥
⎢ ⎥
⎢ 0 0 2M 2R2 2R2 −2R1 2R1 ⎥
⎢ ⎥
⎢
[D] = ⎢ 1R R 2R K K 0 0 ⎥.
1 2 1 2 ⎥
⎢ R1 −R 2R K K 0 0 ⎥
⎢ 1 2 2 1 ⎥
⎣ R2 −R2 2R1 0 0 K1 + K2 + K3 K3 ⎦
−R2 R2 2R1 0 0 K3 K1 + K 2 + K 3
(12.3-11)
The strain vector and stress vector corresponding to (12.3-11) are arranged by
[εij , wij ] = [ε11 , ε22 , ε12 , w11 , w22 , w12 , w21 ],

[σij , Hij ] = [σ11 , σ22 , σ12 , H11 , H22 , H12 , H21 ].
The external force vector {R(m) } in (12.3-6) is

(m) T f T T
{R } = [I] Δdx1 dx2 + [I] ΔdΓ , (12.3-12)
Ωm g Γ
(m) h
t
in which

T
f
= [f1x , f1y , g2x , g2y , · · · , f4x , f4y , g4x , g4y ],
g

T
T
= [T1x , T1y , h1x , h1y , · · · , T4x , T4y , h4x , h4y ]
h

f T
and , are the transposes of the row vectors, x = x1 , y = x2 .
g h
After assemblage of the element matrix, (12.3-6) may be rewritten in terms of
matrix form
1
Π ∗ = {u}T [K]{u} − {u}T {R}. (12.3-13)
2
Carrying out the operation
δΠ ∗ = 0, (12.3-14)
we obtain the finite element equation
[K]{u} = {R}, (12.3-15)
where [K], {u} and {R} are the global stiffness matrix, displacement and external
force vectors generated by assemblage of element stiffness matrixes, element dis-
placement and external force vectors, and {u} is solution vector unknown to be
determined. By solving algebraic equation set (12.3-15) the displacement vector can
be determined, and so do the strains and stresses. The stresses of quasicrystals at the
Gauss integration points can be evaluated through the element node displacements
in accordance with the following version

(m)
(m)
(m)
(m)
σ u u u
= [S] = [D][B] = [D][L][I] ,
H m m m
(12.3-16)
where
[S] = [D][B] = [D][L][I]
is the element stress matrix

(m)T
σ
= [σxx , σyy , σxy , Hxx , Hyy , Hxy , Hyx ]
H
is the phonon and phason stress vector in element m and

(m)T
u
= [u1x , u1y , w1x , w1y , · · · u4x , u4y , w4x , w4y ]
w
is the node displacement (including phonones and phasons) vector at element m.

12.4 Numerical examples 267
12.4 Numerical examples

To illustrate the application of the finite element method to stress analysis of qua-
sicrystals, some numerical results are displayed in this section
Example 1 A hollow cylinder of octagonal quasicrystal material subjected to
an inner pressure is shown in Fig. 12.4-1.
Fig. 12.4-1 Hollow cylinder of octagonal quasicrystal subjected to a pressure
There are boundary conditions
r = r0 : σrr = −p, σrθ = 0, Hrr = 0, Hrθ = 0

(12.4-1)
r = r1 : σrr = 0, σrθ = 0, Hrr = 0, Hrθ = 0.
The phonon and phason stresses, displacements and free energy (strain energy den-
sity) are ploted versus radial distance in Figs.12.4-2∼12.4-6 for different material
constants, respectively.
In the Fig.12.4-3, it seems that the phason stress is one order of magnitude less
than the phonon stress, but this by no means shows that the phason stress are less
than phonon stress in general. One of the reasons resulting in the lower phason
stresses compared with the phonon stresses lies in the boundary conditions (12.4-1).
Instead of (12.4-1), if we take
r = r0 : σrr = 0, σrθ = 0, Hrr = q, Hrθ = 0,

(12.4-2)
r = r1 : σrr = 0, σrθ = 0, Hrr = 0, Hrθ = 0,
then the computation shows that the phason stresses are greater than the phonon
stresses. This suggests that the boundary conditions play an important role in the
analysis. Due to the lack of the knowledge especially lack of the measuring data
on the generalized tractions hi , one often supposes that the tractions to be zero,
this influences the correctness of the computational results, especially to those of
the phason field. Even so, the computaion reveals the effects of the phason and
phonon-phason coupling.
Fig.12.4-2 Phonon stress versus radial distance

1. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
2. K1 /M = 0.7, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
3. K1 /M = 0.8, K2 /M = 0.7, K3 /M = 0.4, R1 /M = 0.1;
4. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.2
Fig.12.4-3 Phason stress versus radial distance

1. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
2. K1 /M = 0.7, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
3. K1 /M = 0.8, K2 /M = 0.7, K3 /M = 0.4, R1 /M = 0.1;
4. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.2
Fig.12.4-4 Phonon displacement versus radial distance

1. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
2. K1 /M = 0.7, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
3. K1 /M = 0.8, K2 /M = 0.7, K3 /M = 0.4, R1 /M = 0.1;
4. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.2
Fig.12.4-5 Phason displacement versus radial distance

1. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
2. K1 /M = 0.7, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1;
3. K1 /M = 0.8, K2 /M = 0.7, K3 /M = 0.4, R1 /M = 0.1;
4. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.2
Fig.12.4-6 shows that the variation of the free energy (or the strain energy den-
sity) in the cylinder as function of coordinates, versus radial distance for different
material constants.
The mesh configuration of the phason field presents some changes before and
after deformation, which are shown in Fig. 12.4-7, this explores the anisotropic
behaviour of the phason field of octagonal quasicrystal. The axisymmetry of the

mesh configuration of phonon field remains after deformation, because the phonon
field in xy-plane is the isotropic plane of elasticity. The results are identical to those
discussed by Refs. [2,3].
Fig.12.4-6 Strain energy density of hollow cylinder versus radial distance

1 1.75E+0.5, 2 2.50E+05, 3 3.61E+0.5, 4 4.55E+05, 5 5.40E+05, 6 6.41E+05, 7 7.34E+05
Fig. 12.4-7 The mesh of phason field before and after deformation:
K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R1 /M = 0.1[9]
Example 2 Numerical solution of Griffith crack in an octagonal quasicrystal

with point group 8mm[9]
The problem has been solved analytically and given by Zhou and Fan[8] but it
is more difficult than that for decagonal[4] . Here the numerical solutions for two
different materials R/M =0.0 and R/M =0.1, and other elastic constants are the
same as given in example 1. The results are depicted by Fig. 12.4-8. The solution
for the former material, is the solution for dodecagonal quasicrystal in fact. All
results are in agreement with those of analytic solutions, see Chapter 8 and the
Refs. [5]∼[7] of this chapter for more detailed discussions.
Fig. 12.4-8 FEM numerical results of normalized phonon stress near the crack tip versus
normalized distance
1. K1 /M = 0.8, K2 /M = 0.6, K3 /M = 0.4, R/M = 0.0
2. K1 /M = 0.7, K2 /M = 0.6, K3 /M = 0.4, R/M = 0.1
References
[1] Hu H C. Variational Principles of Elasticity, Beijing. Science Press, 1981 (in Chinese)
[2] De P, Pelcorits R A. Linear elasticity theory of pentagonal quasicrystals. Phys Rev
B, 1987 35(28): 8609−8620
Phy Rev B, 1993 48(16). 7003−7010
A, 1999, 79(8): 1943−1952
[5] Guo L H, Fan T Y. Solvability of boundary value problems of elasticity of three-
dimensional quasicrystal. Appl Math Mech, English Edition 2007, 28(8): 1061−1070
[6] Guo Y C, Fan T Y. Mode II Griffith crack in decagonal quasicrystals. Appl Math
Mech, English Edition, 2001, 22(10), 1311−1317
[7] Fan T Y and Mai Y W. Elasticity theory, fracture mechanics and some relevant
thermal properties of quasicrystalline materials, Appl. Mech. Rev, 2004, 57(5),
325−344.
[8] Zhou W M, Fan T Y, Plane elasticity and crack problem of octagonal quasicrystals.
Chin. Phys. 2001, 10 (6), 277−284
[9] Wu X F, Numerical analysis of power-law hardening material and quasicrystalline
material with defects, Dissertation, Beijing Institute of Technology, 1998 (in Chinese)
Chapter 13
Some mathematical principles on solutions
of elasticity of quasicrystals
Starting from Chapter 4 we studied the elasticity of quasicrystals and gave many
solutions. A further discussion on some common features of the solutions will be
offered in this chapter.
13.1 Uniqueness of solution of elasticity of quasicrystals

Theorem Assume quasicrystal occupied by region Ω is in equilibrium under
the boundary conditions

σij nj = Ti , Hij nj = hi , xi ∈ Γt ,
(13.1-1)
ui = ūi , wi = w̄i , xi ∈ Γu ,
i.e., if the equations

∂σij ∂Hij
+ fi = 0, + gi = 0, xi ∈ Ω (13.1-2)
∂xj ∂xj
coupled by
(13.1-3)
Hij = Rklij εkl + Kijkl wkl
and
1 ∂ui ∂uj ∂wi
εij = + , wij = (13.1-4)
2 ∂xj ∂xi ∂xj
satisfying boundary conditions (13.1-1), then the boundary value problem
(13.1-1)∼(13.1-4) has a unique solution.
Proof If the conclusion is not true, suppose there are two solutions of equations
(13.1-2)∼(13.1-4) under boundary conditions (13.1-1):
(1) (1) (2) (2)
ui ⊕ w i , ui ⊕ w i ,
(1) (1) (2) (2)
εij ⊕ wij , εij ⊕ wij ,
(1) (1) (2) (2)
σij ⊕ Hij , σij ⊕ Hij ,

274 Chapter 13 Some mathematical principles on solutions of elasticity of quasicrystals
both solutions satisfy equations of deformation geometry, stress-strain relationship,

equilibrium equations and boundary conditions.
According to the linear superposition principle, the difference between these two
solutions
(1) (2) (1) (2)
(ui − ui ) ⊕ (wi − wi ) = Δui ⊕ Δwi ,
(1) (2) (1) (2)

(σij − σij ) ⊕ (Hij − Hij ) = Δσij ⊕ ΔHij
should be the solution of the problem too. The “differences” should satisfy zero
boundary conditions. The work done by the external forces is

0= (Δf ⊕ Δg) · (Δu ⊕ Δw)dΩ
Ω

+ (ΔT ⊕ Δh) · (Δu ⊕ Δw)dΓ = 2 ΔU dΩ , (13.1-5)
Γ Ω
where ΔU represents the strain energy density corresponding to the differences of

two sets of solutions, and the derivation of the last step of (13.1-5) used the Gauss
theorem.
Because ΔU is the positive definite quadratic form of Δεij and Δwij , there is
ΔU 0, (13.1-6)
considering the left-handside of (13.1-5) be zero, it follows that
ΔU = 0. (13.1-7)
Based on the positive definite property of ΔU , Δεij and Δwij must be zero,
(1) (2) (1) (2)
apart from only rigid displacements, so that εij = εij , wij = wij , etc.
At the same time, at boundary Γu , one has
Δu = 0, Δw = 0.
This means the rigid-body displacements cannot exist, it must be
(1) (2) (1) (2)

ui = ui , wi = wi .
The theorem is a powerful tool in the study of elasticity of quasicrystals, e.g.

uniqueness of the solution ensures that different approaches for a problem of elasticity
of quasicrystals will not result in different solutions so long as the equations of
elasticity of quasicrystals and the prescribed boundary conditions are satisfied.
13.2 Generalized Lax-Milgram theorem 275
13.2 Generalized Lax-Milgram theorem

Consider a quasicrystal described by equations (13.1-2)∼(13.1-4) and boundary con-
ditions (13.1-1). In the case the displacements satisfy the following partial differential
equations:

∂ 2 uk ∂ 2 wk
− Cijkl + Rijkl = fi (x, y, z)),
∂xj ∂xl ∂xj ∂xl
(x, y, z) ∈ Ω . (13.2-1)
∂ 2 uk ∂ 2 wk
− Rklij + Kijkl = gi (x, y, z),
∂xj ∂xl ∂xj ∂xl
At boundary Γu , they satisfy homogenous conditions
ui |Γu = 0, wi |Γu = 0. (13.2-2)
And at boundary Γt , the corresponding stresses satisfy non-homogenous conditions

⎧
⎪
⎪ ∂uk ∂wk
⎨ Cijkl ∂x + Rijkl ∂x nj |Γt = Ti (x, y, z),
⎪
l l
(13.2-3)
⎪
⎪ ∂u ∂w
⎪
⎩ Rklij
k
+ Kijkl
k
+ nj |Γt = hi (x, y, z),
∂xl ∂xl
in which x1 = x, x2 = y, x3 = z, Γ = Γu + Γt .
If ui , wi ∈ C 2 (Ω )∩C 1 (∂Ω ) and satisfy equation (13.2-1) and boundary conditions
(13.2-2) and (13.2-3), then they will be called a classical solution of boundary value
problem (13.2-1)∼(13.2-3), where ∂Ω = Γ = Γu + Γt .
The boundary value problem (13.2-1)∼(13.2-3) can be transformed into a varia-
tional problem. To this end, let us introduce the norm
⎧ 2 2 2
⎪
⎪ ∂u ∂u ∂u
⎪ 2
⎪ ui 1,Ω =
i
+
i
+
i
dΩ ,
⎪
⎨ ∂x ∂y ∂z
Ω
⎪ 2 2 2 (13.2-4)
⎪
⎪ ∂w ∂w ∂w
⎪ 2
⎪ wi 1,Ω =
i
+
i
+
i
dΩ .
⎩ ∂x ∂y ∂z
Ω
This norm is suitable only for case of homogeneous boundary conditions (13.2-2),
otherwise the norm will not be in this form.
With the norm in (13.2-4), denote the space defining ui (x, y, z) and wi (x, y, z) is
by H (Ω ). If introduce inner product
⎧ $ (1) (2) %
⎪ (1) (2) (1) (2)
⎪
⎪ (1) (2) ∂ui ∂ui ∂ui ∂ui ∂ui ∂ui
⎪
⎪ (u , ui ) = + + dΩ ,
⎨ i Ω ∂x ∂x ∂y ∂y ∂z ∂z
⎪ $ (1) (2) (1) (2) (1) (2)
% (13.2-5)
⎪
⎪ ∂wi ∂wi ∂wi ∂wi ∂wi ∂wi
⎪
⎪
(1) (2)
⎩ (wi , wi ) = ∂x ∂x
+
∂y ∂y
+
∂z ∂z
dΩ ,
Ω
then H (Ω ) is a Hilbert space, it is also named Sobolev space.

We can define space V such as
V = {(ui , wi ) ∈ H (Ω ), (ui )Γu = 0, (wi )Γu = 0}. (13.2-6)
and the inner product at space V (13.2-5), then V is also a Hilbert space and
V ⊂ H (Ω ). (13.2-7)
Define the space
H = {X = (u1 , u2 , u3 , w1 , w2 , w3 ) ∈ (H (Ω ))61 , (ui )Γu = 0, (wi )Γu = 0} (13.2-8)
and the norm

3 1/2
2 2
X1,Ω = [ui 1,Ω + wi 1,Ω ) . (13.2-9)
i=1
For any X = (u1 , u2 , u3 , w1 , w2 , w3 ) ∈ (H (Ω ))61 , by using the strain-displacement

relations
1 ∂ui ∂uj
εij (X) = εji (X) = + ,
2 ∂xj ∂xi
∂wi
wij (X) =
∂xj
and the stress-strain relations of quasicrystals
σij (X) = σji (X) = Cijkl εkl (X) + Rijkl wkl (X),
Hij = Rklij εkl (X) + Kijkl wkl (X),
Define the bilinear functional

B(X (1) , X (2) ) = [σij (X (1) )εij (X (2) ) + Hij (X (1) )wij (X (2) )]dΩ , (13.2-10)
Ω
which possesses symmetry, continuity and positive definite properties.

Take a linear functional

l(X) = [fi ui + gi wi ]dΩ + [Ti ui + hi wi ]dΓ , (13.2-11)
Ω Γt
where (f1 , f2 , f3 , g1 , g2 , g3 ) ∈ (L2 (Ω ))6 , (T1 , T2 , T3 , h1 , h2 , h3 ) ∈ (L2 (Ω ))6 are given

functions at Ω and Γ respectively. One can prove that l(X) is a continuous functional
of X at region Ω , and L2 represents the Lebesgue square integrable meaning.
In what follows, we state a number of theorems on the solution of the problem
of elasticity of quasicrystals.
13.2 Generalized Lax-Milgram theorem 277
Theorem 1 The variational problem associated with the boundary value prob-
lem (13.2-1)∼(13.2-3) is to obtain X ∈ H so that X
B(X, Y ) = l(Y ), ∀Y ∈ H, (13.2-12)
in which l(Y ) is a linear functional defined by (13.2-11).

(Proof is omitted.)
Theorem 2 If X = (u1 , u2 , u3 , w1 , w2 , w3 ) ∈ (C 2 (Ω ))6 is a classical solu-
tion of boundary value problem (13.2-1)∼(13.2-3), then X must be solution of the
variational problem (13.2-12), i.e., X is also a generalized solution (or weak solu-
tion). Alternatively, if X is a solution of the variational problem (13.2-12), and
X ∈ (C 2 (Ω ))6 , then must be the classical solution of the boundary value problem
(13.2-1)∼(13.2-3).
(Proof is omitted.)
Theorem 3(Generalized Lax-Milgram theorem) Assume that H is the Hilbert
space defined above, for elasticity of quasicrystals, B(X, Y ) is a bilinear functional
at H × H and satisfies
B(X, Y ) = f (X), ∀Y ∈ H, (13.2-13)
admits a unique solution X ∈ H and
1
X f (H) , (13.2-14)
α
where (H) is the dual space of H consisting of all bounded linear functionals and
equipped with the norm
f (Y )
f (H) = sup .
Y ∈H,Y
=0 Y
This theorem gives an extension of Lax-Milgram theorem[3] .

(Proof is omitted.)
Theorem 4 Assume
1
J(X) = B(X, X) − l(X) (13.2-15)
2
and H is the same as mentioned previously, B(X, Y ) is a bilinear functional at H ×H
defined by (13.2-10), then the variational problem for having X ∈ H so that J(X)
being minimum
J(X0 ) = min J(X) (13.2-16)
X∈H
in which
(1) Exists solution, and the number of solutions does not exceed one;
(2) If there is a solution of problem (13.2-16), which must be the solution of
problem (13.2-12), and vise versa;
(3) If X ∗ is their solution, then

1
J(X) − J(X ∗ ) = B(X − X ∗ , X − X ∗ ), ∀X ∈ H. (13.2-17)
2
(Proof is omitted.)
The discussion of this section may provide preparedness for the following text.
13.3 Matrix expression of elasticity of three-dimensional qua-

sicrystals
In the previous section the variational problem on elasticity of quasicrystals was
discussed in which the concept of weak solution (generalized solution) is mentioned.
In the following sections we will deal with the weak solution of elasticity of three-
dimensional icosahedral quasicrystals, for simplicity a matrix expression for the elas-
ticity will be used. From the discussion of previous chapters we know that the basic
equations for elasticity of quasicrystals except cubic quasicrystals
⎧
⎪ ∂σij ∂ 2 ui ∂Hij ∂ 2 wi ∂wi
⎪
⎪ + fi = ρ 2 , + gi = ρ 2 or κ , (13.3-1)
⎪
⎪
⎨ ∂xj
∂ t

∂xj ∂t ∂t
1 ∂ui ∂uj ∂wi
⎪
⎪ εij = + , wij = , (13.3-2)
⎪
⎪ 2 ∂x ∂x ∂xj
⎪
⎩
j i
σij = Cijkl εkl + Rijkl wkl , Hij = Kijkl wkl + Rklij εkl , (13.3-3)
where xi ∈ Ω , i, j = 1, 2, 3, t > 0, u = (ux , uy , uz ) represents the phonon displace-

ment vector, w = (wx , wy , wz ) the phason one, εij , wij the phonon and phason
strains, σij , Hij the corresponding phonon and phason stresses, Cijkl , Kijkl , Rijkl
the elastic constants, fi , gi the body force and generalized body force densities, ρ
the mass density and κ = 1/Γw , respectively.
Denoting ∂Ω = (∂Ω )u + (∂Ω )σ to express the boundary of region occupied by
the quasicrystal there are the boundary conditions
x ∈ (∂Ω )u : ui = u0i , wi = wi0 , (13.3-4)
x ∈ (∂Ω )σ : σij nj = Ti , Hij nj = hi , (13.3-5)

where u0i and wi0 represent the known functions at boundary (∂Ω )u , Ti and hi the
prescribed traction and generalized traction density at boundary (∂Ω )σ , nj the unit
outward normal vector at any point of the boundary.
There are initial conditions for dynamic problem
ui |t=0 = ai (x), wi |t=0 = bi (x), u̇i |t=0 = ci (x), ẇ |t=0 = di (x),

(13.3-6)
(or ui |t=0 = ai (x), wi |t=0 = bi (x), u̇i |t=0 = ci (x))
13.3 Matrix expression of elasticity of three-dimensional quasicrystals 279
where ai , bi , ci , di are known functions, x = (x1 , x2 , x3 ) ∈ Ω .

Putting
T T
Ũ = (u1 , u2 , u3 , w1 , w2 , w3 , )1×6 , F̃ = (fi gi )1×6 = (f1 , f2 , f3 , g1 , g2 , g3 )1×6 ,
σ̃ T = (σ11 , σ22 , σ33 , σ12 , σ23 , σ31 , H11 , H22 , H33 , H12 , H23 , H31 , H13 , H21 ,
H32 )1×15 ,
T
ε̃ = (ε11 , ε22 , ε33 , 2ε12 , 2ε23 , 2ε31 , w11 , w22 , w33 , w12 , w23 , w31 , w13 , w21 ,
w32 )1×15 ,
⎡ ⎤
∂1 0 0
⎡ ⎤ ⎢ 0 ∂2 0 ⎥
∂1 0 0 ⎢ ⎥
⎢ ⎥
⎢ ⎥ ⎢ 0 0 ∂3 ⎥
⎡ ⎤ ⎢ 0 ∂2 0 ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ∂2 0 0 ⎥
˜ (1)
∂ 0 ⎢
⎢ 0 0 ∂3 ⎥
⎥
⎢
⎢
⎥
⎥
˜ =⎣
∂ ⎦, ˜ (1)
∂ =⎢ ⎥, ˜ (2)
∂ =⎢ 0 ∂3 0 ⎥,
˜ (2)
∂ ⎢ ∂2 ∂1 0 ⎥ ⎢ ⎥
0 ⎢ ⎥ ⎢ 0 0 ∂1 ⎥
⎢ 0 ∂3 ∂2 ⎥ ⎢ ⎥
⎣ ⎦ ⎢ ⎥
⎢ ∂3 0 0 ⎥
∂3 0 ∂1 ⎢ ⎥
⎣ 0 ∂1 0 ⎦
0 0 ∂2
∂
∂i = ,
∂xi
T T
Ũ , F̃ , ∂ T , σ̃ T , ε̃T denote the transpose of Ũ , F̃ , ∂, σ̃, ε̃, σ i = (σi1 , σi2 , σi3 ) the
i-th row of matrix (σij )3×3 , H i = (Hi1 , Hi2 , Hi3 ) the i-th row of matrix (Hij )3×3 .
So that, (13.3-2) can be rewritten as the matrix form
ε̃ = ∂ Ũ . (13.3-2 )
T
∂σ1j ∂σ2j ∂σ3j ∂H1j ∂H2j ∂H3j
Note that = ∂ T σ̃, then (13.3.1)
∂xj ∂xj ∂xj ∂xj ∂xj ∂xj
can be rewritten by
. = ρU
∂ T σ̃ + F .̈ . (13.3-1 )
Putting
⎡ ⎤ ⎡ ⎤
··· C11ij ··· C1111 C1122 C1133 C1112 C1123 C1131
⎢ ··· ··· ⎥ ⎢ ⎥
⎢ C22ij ⎥ ⎢ C2211 C2222 C2233 C2212 C2223 C2231 ⎥
⎢ ⎥ ⎢ ⎥
⎢ ··· C33ij ··· ⎥ ⎢ C3311 C3322 C3333 C3312 C3323 C3331 ⎥
C=⎢
⎢
⎥
⎥ =⎢
⎢
⎥
⎥ ,
⎢ ··· C12ij ··· ⎥ ⎢ C1211 C1222 C1233 C1212 C1223 C1231 ⎥
⎢ ⎥ ⎢ ⎥
⎣ ··· C23ij ··· ⎦ ⎣ C2311 C2322 C2333 C2312 C2323 C2331 ⎦
··· C31ij ··· 6×6
C3111 C3122 C3133 C3112 C3123 C3131 6×6
⎡ ⎤
··· K11ij ···
⎢ ··· K22ij ··· ⎥
⎢ ⎥
⎢ ··· K33ij ··· ⎥
⎢ ⎥
⎢ ··· K12ij ··· ⎥
⎢ ⎥
K =⎢
⎢ ··· K23ij ··· ⎥
⎥
⎢ ··· K31ij ··· ⎥
⎢ ⎥
⎢ ··· K13ij ··· ⎥
⎢ ⎥
⎣ ··· K21ij ··· ⎦
··· K32ij ···
⎡ ⎤
K1111 K1122 K1133 K1112 K1123 K1131 K1113 K1121 K1132
⎢ K2211 K2222 K2233 K2212 K2223 K2231 K2213 K2221 K2232 ⎥
⎢ ⎥
⎢ K3311 K3322 K3333 K3312 K3323 K3331 K3313 K3321 K3332 ⎥
⎢ ⎥
⎢ K1211 K1222 K1233 K1212 K1223 K1231 K1213 K1221 K1232 ⎥
⎢ ⎥
=⎢
⎢ K2311 K2322 K2333 K2312 K2323 K2331 K2313 K2321 K2332 ⎥
⎥ ,
⎢ K3111 K3122 K3133 K3112 K3123 K3131 K3113 K3121 K3132 ⎥
⎢ ⎥
⎢ K1311 K1322 K1333 K1312 K1323 K1331 K1313 K1321 K1332 ⎥
⎢ ⎥
⎣ K2111 K2122 K2133 K2112 K2123 K2131 K2113 K2121 K2132 ⎦
K3211 K3222 K3233 K3212 K3223 K3231 K3213 K3221 K3232 9×9
⎡ ⎤
··· R11ij ···
⎢ ··· R22ij ··· ⎥
⎢ ⎥
⎢ ··· R33ij ··· ⎥
R =⎢ ⎥
⎢ ··· R12ij ··· ⎥
⎢ ⎥
⎣ ··· R23ij ··· ⎦
··· R31ij ··· 6×9
⎡ ⎤
R1111 R1122 R1133 R1112 R1123 R1131 R1113 R1121 R1132
⎢ R2211 R2222 R2233 R2212 R2223 R2231 R2213 R2221 R2232 ⎥
⎢ ⎥
⎢ R3311 R3322 R3333 R3312 R3323 R3331 R3313 R3321 R2232 ⎥
=⎢ ⎥ ,
⎢ R1211 R1222 R1233 R1212 R1223 R1231 R1213 R1221 R1232 ⎥
⎢ ⎥
⎣ R2311 R2322 R2333 R2312 R2323 R2331 R2313 R2321 R2332 ⎦
R3111 R3122 R3133 R3112 R3123 R3131 R3113 R3121 R3132 6×9
then
C R
D = (dij )15×15 = .
RT K
Here the order of index i, j of C is the same with those of the phonon strain tensor,
the order of index i, j of K, R is the same with those of phason strain tensor, and
RT is the transpose of R. From the above expressions one can find that due to the
symmetry of C and K (see e.g. (4.4-3) and (4.4-5)), the matrix D is symmetric.
The generalized Hooke’s law (13.3-3) can be rewritten as
σ̃ = Dε̃, (13.3-3 )
13.3 Matrix expression of elasticity of three-dimensional quasicrystals 281
(13.3-1 ) and (13.3-2 ) can be collected as below:

.̈ .
∂ T D∂ Ũ + F̃ = ρU (13.3-7)
Put
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
a1 (x) c1 (x) u01
⎢ a2 (x) ⎥ ⎢ c2 (x) ⎥ ⎢ u02 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ a3 (x) ⎥ ⎢ c3 (x) ⎥ 0 ⎢ u03 ⎥
A(x) = ⎢ ⎥ , B(x) = ⎢ ⎥ , Ũ = ⎢ ⎥ ,
⎢ b1 (x) ⎥ ⎢ d1 (x) ⎥ ⎢ w10 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣ b2 (x) ⎦ ⎣ d2 (x) ⎦ ⎣ w20 ⎦
b3 (x) 6×1
d3 (x) 6×1
w30 6×1
⎡ ⎤
T1
⎢ T2 ⎥ ⎡ ⎤
⎢ ⎥ ˜ (1)
⎢ T3 ⎥ ∂ 0
σ̃ = ⎢
0 ⎥ , ñ = ⎣
∂
n ⎦,
⎢ ⎥
⎢ h1 ⎥ 0 ˜ (2)
∂
⎣ h2 ⎦ n
h3 6×1
⎡ ⎤
cos(n, x1 ) 0 0
⎢ 0 cos(n, x2 ) 0 ⎥
⎢ ⎥
⎢ ⎥
˜ (1)
∂ =⎢
0 0 cos(n, x3 ) ⎥,
n ⎢ cos(n, x2 ) cos(n, x1 ) 0 ⎥
⎢ ⎥
⎣ 0 cos(n, x3 ) cos(n, x2 ) ⎦
cos(n, x3 ) 0 cos(n, x1 )
⎡ ⎤
cos(n, x1 ) 0 0
⎢ 0 cos(n, x2 ) 0 ⎥
⎢ ⎥
⎢ 0 0 cos(n, x3 ) ⎥
⎢ ⎥
⎢ cos(n, x2 ) 0 0 ⎥
⎢ ⎥
˜ (2)
∂ =⎢ 0 cos(n, x3 ) 0 ⎥,
n ⎢ ⎥
⎢ 0 0 cos(n, x1 ) ⎥
⎢ ⎥
⎢ cos(n, x3 ) 0 0 ⎥
⎢ ⎥
⎣ 0 cos(n, x1 ) 0 ⎦
0 0 cos(n, x2 )
˜ (1) , ∂
ñ , ∂
where ∂ ˜ (2) are obtained from the differential operator matrices ∂ ˜ (1) , ∂
˜ ,∂ ˜ (2)
n n
through a replacement ∂ i by cos(n, xi ).
Equation (13.3-4) can be rewritten as
0
Ũ (x, t) = Ũ , x ∈ (∂Ω )u . (13.3-4 )
Considering the similarity of the left-handside of (13.3-5) with the first term of (13.3-
1), then (13.3-5) can be rewritten as ∂ ˜ T σ̃ = σ̃ 0 , x ∈ (∂Ω )σ , In addition by using
n
(13.3-2 ) and (13.3-3 ) there is
˜ T D∂
∂ ˜ Ũ = σ̃ 0 , x ∈ (∂Ω )σ . (13.3-5 )
n
.̈ = 0 (i.e., the inertia forces vanish),

If the quasicrystal is in static equilibrium: ρU
in the case it needs not the initial conditions, and the boundary value problem of
quasicrystals is interpreted as follows:
T
−∂ ˜ D∂ ˜ Ũ = F̃ , x ∈ Ω , t > 0, (13.3-8)
" "
0
Ũ (x, t)"(∂Ω) = Ũ , ∂˜ T D∂˜ Ũ (x, t)" = σ̃0 , (13.3-9)
u
n (∂Ω)σ
where
∂Ω = (∂Ω )u + (∂Ω )σ .
13.4 The weak solution of boundary value problem of elas-

ticity of quasicrystals
For simplicity only the displacement boundary problem (or say the Dirichlet prob-
lem) is dealt with in the following. Suppose that F̃ ∈ (L2 (Ω ))6 , if Ũ (x) ∈ (C 2 (Ω̃ ))6
being the
solution of problem (13.3-8), (13.3-9), then for any vector function η =
1
η3×1 T
2 = η11 η21 η31 η12 η22 η32 1×6 ∈ (C0∞ (Ω ))6 , multiplying the both
η3×1
sides of (13.3-8) by η (by making scalar product), and then integrating along Ω , we
have
T
˜ ˜
(−∂ D ∂ Ũ ) · ηdx = F̃ · ηdx. (13.4-1)
Ω Ω
From (13.3-3) and Chapter 4 known σij = σji , in terms of Gauss formula, (13.3-2 )
and (13.3-3 ), there exists

˜ T D∂ ˜ Ũ · ηdx = − ∂σij 1 ∂Hij 2
−∂ ηi + ηi dx
∂xj ∂xj
Ω Ω

∂ 1 2 ∂ηi1 ∂ηi2
= − (σij ηi + Hij ηi ) − σij + hij dx
∂xj ∂xj ∂xj
Ω

∂η 1 ∂η 2
= − (σij ηi1 +Hij ηi2 )nj dS + σij i +Hij i dx
∂Ω ∂xj ∂xj
Ω
$ %
1 ∂ηi1 1 ∂ηj1 ∂η 2
= σij + σij + Hij i dx
2 ∂xj 2 ∂xi ∂xj
Ω

= [(σij (Ũ )εij (η 1 ) + Hij (Ũ )wij (η 2 )]dx
Ω

T
= σ̃(Ũ ) · ε̃(η)dx = ˜
(∂η) ˜ Ũ dx,
D∂ (13.4-2)
Ω Ω
13.5 The uniqueness of weak solution 283
(13.4-1) and (13.4-2) yield

˜ T D∂
(∂η) ˜ Ũ dx = σ̃(Ũ ) · ε̃(η)dx = F̃ · ηdx. (13.4-3)
Ω Ω Ω
Because C0∞ is dense in H01 (Ω ),(13.4-3) holds for ∀η(x) ∈ (H01 (Ω ))6 .
In contrast, if Ũ (x) ∈ (C (Ω̄ ))6 , and (13.4-3) is valid ∀η(x) ∈ (H01 (Ω ))6 , we can
2
do derivation in counter order of the above procedure, and find (13.4-1) through the
fundamental lemma of variational method[2] . So that we have
Definition Assume F̃ ∈ (L2 (Ω ))6 , if Ũ (x) ∈ (H01 (Ω ))6 , and (13.4-3) holds
∀η(x) ∈ (H01 (Ω ))6 , then say Ũ (x) being the weak solution (or generalized solution)
of the boundary value problem

−∂ ˜ T D∂˜ Ũ (x) = F̃ (x), x ∈ Ω , t > 0 (13.3-8)
Ũ (x) |∂Ω = 0. (13.3-9 )
13.5 The uniqueness of weak solution

Making use of (·, ·) to express inner product in L2 (Ω ), and the corresponding norm
1/2
2
is · : v = v dx , for the scalar function v ∈ L2 (Ω ). And in terms of (·, ·)1
Ω
to denote the inner product in H01 (Ω ), the corresponding norm is ·1 : v1 =
3 2 1/2 3 2 1/2
∂v ∂v
2
v dx+ dx , and semi-norm is |·|1 : |v|1= dx ,
Ω Ω ∂xk
k=1 Ω ∂xk k=1
for scalar function v ∈ L2 (Ω ).
Note 1 Norm ·1 is equivalent to semi-norm |·|1 .
Note 2 The norm and semi-norm of vector function v = (v1 , v2 , · · · , vn ) ∈
(H01 (Ω ))n (it is marked by H01 (Ω ) some times)are denoted as ·1 and |·|1 , such as
n n n 3 2
∂vi
v21 = vi 21 = vi2 dx + dx = |v|2 dx + |vx |2 dx,
i=1 i=1 Ω i=1 Ω ∂xk Ω Ω
k=1
3
n 2
∂vi
|v|21 = dx,
i=1 k=1 Ω ∂xk

n 3
n 2
2 2 ∂vi ∂v ∂v1 ∂v2 ∂vn
where |v| = vi2 , |vx | = = , , ,··· , . Obvi-
i=1 i=1 k=1
∂xk ∂xk ∂xk ∂xk ∂xk
ously, the Note 1 holds for vector function v, too.
Lemma(Korn inequality[4,5] ) Assume Ω is a bounded region with boundary
∂Ω of sufficient smooth in Rn , and ∀v = (v1 , v2 , · · · , vn ) ∈ H01 (Ω ), there is
n 2
∂vi ∂vk 2
+ dx c1 v1 ,
Ω ∂x k ∂x i
i,k=1
in which the positive constant c1 is only dependent on Ω .

Theorem Suppose Ω is a bounded region in R3 and with sufficient smooth
boundary ∂Ω , if real symmetric matrix D = (dij ) satisfies the inequality
15
15
15

λ1 ξi2 ξi dij ξj λ2 ξi2 ,
i=1 i,j=1 i=1
where λ1 , λ2 are positive constants, then for any F̃ ∈ (L2 (Ω ))6 , displacement bound-
ary value problem (13.3-8), (13.3-9 ) exists unique weak solution (or generalized
solution).
T
Proof Put Ũ , η = (∂η) D∂ Ũ dx, then (13.4-3) can be rewritten as
Ω
Ũ , η = (F̃ , η), ∀η ∈ (H01 (Ω ))6 . (13.5-1)
At first we prove ·, · is a new inner product at (H01 (Ω ))6 . For this purpose it
needs to prove: Ũ , Ũ 0, and Ũ , Ũ = 0 ⇔ Ũ = 0, ∀Ũ ∈ (H01 (Ω ))6 .
In the following we give only an outline of the proof, the detail is omitted. In
addition Ũ , η in (13.5-1) is positive definite bilinear functional at H01 (Ω ), the proof
can be done from the Lax-Milgram theorem (see Section 13.2).
Due to the assumption, matrix D = (dij )15×15 being positive definite, matrix
D = (dij )15×15 and unit matrix I are in contract, i.e., there exists a reversible matrix
C such that
D = CTC
(note that here C is not the phonon elastic constant matrix). Then

T
Ũ , Ũ = ˜ ˜
(∂ Ũ ) D ∂ Ũ dx = ˜ Ũ )T (C T C)∂
(∂ ˜ Ũ dx
Ω Ω

T
= ˜ Ũ ) (C ∂
(C ∂ ˜ Ũ )dx 0,
Ω

.,U
U . = 0 ⇔ .U
(C ∂ . )T (C ∂
.U .U
. )dx = 0 ⇔ C ∂ . = 0.
Ω
.U
Because C is reversible, ∂ . = 0, i.e.,
∂ui ∂ui ∂uj ∂wi
= 0, + = 0 (i = j), = 0, i, j = 1, 2, 3.
∂xi ∂xj ∂xi ∂xj
∂wi
It follows = 0 that wi should be constant, besides Ũ |∂Ω = 0, and wi = 0 at
∂xj
the boundary. In similar analysis we find that ui = 0 at boundary. Thus Ũ = 0 at
boundary.
13.5 The uniqueness of weak solution 285
In this way we have proved ·, · is a new inner product at (H01 (Ω ))6 , the corre-
sponding norm is U . (1) = U .,U . 12 .
Secondly, at (H01 (Ω ))6 the new norm ·(1) is equivalent to the initial norm ·1 .
We are going to give the proof about this.
From the Cauchy’s inequality, the assumption of the theorem and Note 1, there
is
/ /2 15
/ / ˜ Ũ )i dij (∂
˜ Ũ )j dx λ2 ˜ Ũ )T (∂
˜ Ũ )dx
/Ũ / = (∂ (∂
(1)
Ω i,j=1 Ω
⎡ ⎤
3 2 3 2 3 2
⎢ ∂ui ∂ui ∂uj ∂wi ⎥
= λ2 ⎣ + + + ⎦ dx
i=1
∂xi i,j=1
∂xj ∂xi i,j=1
∂xj
Ω i<j
⎧ ⎫
⎪ 3 2 2 2 3 2 ⎪
⎨ ∂u
3
∂ui ∂uj ∂wi ⎬
i
λ2 +2 + + dx
⎪
⎩ i=1 ∂xi i,j=1
∂xj ∂xi ∂xj ⎪ ⎭
Ω i,j=1
i<j

3
2 2
∂ui ∂wi
2λ2 + dx
i,j=1
∂xj ∂xj
Ω
" "2 / /2
" " / /
= 2λ2 "Ũ " 2c /Ũ / .
1 1
In other hand, from the assumption of the theorem, the Korn inequality and
Note 1 there is
/ /2 15
/ / ˜ Ũ )T (∂
/Ũ / = (∂˜Ũ )i dij (∂
˜ Ũ )j dx λ1 (∂ ˜ Ũ )dx
(1)
Ω i,j=1 Ω
⎡ ⎤
3 2 3 2 3 2
⎢ ∂ui ∂ui ∂uj ∂wi ⎥
= λ1 ⎣ + + + ⎦ dx
i=1
∂xi i,j=1
∂xj ∂xi i,j=1
∂xj
Ω i<j
⎡ ⎤
3
2 3 2
⎣1 ∂ui ∂uj ∂wi ⎦
= λ1 + + dx
4 i,j=1 ∂xj ∂xi i,j=1
∂xj
Ω
2 3 2
1 ∂ui ∂wi
λ1 c2 dx + λ1 dx
4 ∂xj i,j=1 Ω
∂xj
Ω

" " / /2
1 " "2 / /
min λ1 c2 , λ1 "Ũ " c /Ũ / .
4 1 1
Consequently, we have proved the equivalency between the new norm ·(1) and the
initial norm ·1 .
Finally, for F̃ ∈ (L2 (Ω ))6 , by using Schwarz inequality and the fact that the
embedding H01 (Ω ) ⊂ → L2 (Ω ) is a compact embedding, we have
" "
" " / / / / / /
" " / /
" F̃ · ηdx " /F̃ / · η M /
/ F̃
/
/ · η M
/ /
1 /F̃ / · η1 , ∀η ∈ (H01 (Ω ))6 ,
" " 1
" "
Ω

i.e., η → F̃ · ηdx (∀η ∈ (H01 (Ω ))6 ) is a unique continuous linear functional
Ω
at (H01 (Ω ))6 .
Therefore, from Riesz theorem, there must be a unique y F̃ ∈ (H01 (Ω ))6 , such
that
F̃ · ηdx = y F̃ , η , ∀η ∈ (H01 (Ω ))6 .
Ω
Thus (13.5-1) is rewritten as
3 4
Ũ , η = y F̃ , η , ∀η ∈ (H01 (Ω ))6 .
This shows Ũ = y F̃ is the unique weak solution (generalized solution) of displace-
ment value problem (13.3-8), (13.3-9 ).
In the above proof, using Korn inequality is crucial (the second Korn inequality
will be used for the stress boundary value problem, but this discussion is not included
here).

In some extent the discussion Sections 13.3∼13.5 is a development of Section 13.2.
These discussions provide a basis for the numerical methods developed in Chapters
10 and 12 from point of view of weak solution. Of course for the stress boundary
value problems and dynamic problems the discussion needs to do some extensions
of the previous discussions.
It is evident that the discussion in Sections 13.3∼13.5 is valid for any systems of
one-, two- and three-dimensional quasicrystals, except cubic quasicrystals, for which
wij = ∂wi /∂xj should be replaced by wij = (∂wi /∂xj +∂wj /∂xi )/2 in the equations
(13.3-2) only. The difference of the formulation for different quasicrystal systems
lies in the elastic constant matrix D, one can directly use the above formulation for
any quasicrystal systems by substituting the concrete matrix D into the relevant
formulas.
References
References 287
[2] Courant R, Hilbert D. Methods of Mathematical Physics. New York: Interscience

Publisher Inc. 1955
[3] Oden J J, Reddy J H. An Introduction to the Mathematical Theory of Finite Element.
New York: John Wiley & Sons. 1976
[4] Guo L H, Fan T Y. Solvability on boundary-value problems of elasticity of three-
dimensional quasicrystals. Applied Mathematics and Mechanics, 2007, 28(8): 1061–
1070
[5] Fikera G. Existence Theorems of Elasticity Theory. Moscow: World Press, 1974 (in
Russian)
[6] Kondratjev W A, Oleinik O A. Korn Inequality of Boundary Value Problems for
Systems of Theory of Elasticity in Boundless Region. Moscow: UMN Press, 1988 (in
Russian)
Chapter 14
Nonlinear behaviour of quasicrystals
From Chapter 4 to Chapter 13 we mainly discussed the elasticity and relevant prop-
erties of quasicrystals, which belong to linear regime both physically and mathemat-
ically. Their mathematical treatment is relatively easy though the calculations are
quite complex.
The current chapter is to give a simple description on deformation and fracture
of quasicrystals with nonlinear behaviour, considering the great difficulty in this
topic. For the conventional engineering materials including crystalline material, the
nonlinear behaviour means mainly plasticity. In the study on the classical plasticity
there are two different theories, one is the macroscopic plasticity theory, which is
based on some assumptions concluded from certain experimental data, and the other
is so-called crystal plasticity theory, which is based on the mechanism of motion of
dislocation, and in some extent can be seen as a “microscopic” theory. The difficulty
for quasicrystal plasticity lies in lack of both enough macro- and micro-data. At
present the macroscopic experiments have not, as yet, been properly undertaken.
Though there is some work on the mechanism in microscopy of the plasticity, the
data are very limited. This leads to the constitutive law of quasicrystals being
essentially unknown. Due to the reason the systematic mathematical analysis on
deformation and fracture for the material is not available so far.
In spite of these difficulties, study on plasticity of quasicrystals has aroused a
great deal of attention of researchers[1∼8] . But the analytic quantitative work may
be at an infant stage. Considering the readers’ interest and the development level, it
is beneficial to give a brief discussion on some simple problems of nonlinear behaviour
of the material with some simple models and by extending results in the study of
linear regime. Of course, these discussions are not complete, which may provide
some hints for further development of the area.
This chapter is arranged as follows. First, we discuss some experimental results
on the nonlinear deformation behaviour of quasicrystals, then describe a possible
plastic constitutive equation of the material. In view of the difficulty for setting
up the equations, we turn to introduce nonlinear elastic constitutive equations of
quasicrystals which are available at present though not equivalent to the plastic

290 Chapter 14 Nonlinear behaviour of quasicrystals
constitutive equations. The Sections 14.4 and 14.5 may be seen as applications of
the macro-constitutive law in which some nonlinear solutions on quasicrystals are
presented. In Section 14.6 another version of the study based on the dislocation
model or “microscopic model” is exhibited, which achieves the same results given in
Sections 14.4 and 14.5.
14.1 Macroscopic behaviour of plastic deformation of qua-

sicrystals
At medium and low temperature, quasicrystals exhibit brittleness, but they present
plasticity-ductility at high temperature. In addition, near the high stress concen-
tration zone, e.g. around dislocation core or crack tip, plastic flow appears. It was
observed in experiments that plastic deformation of quasicrystals is induced by mo-
tion of dislocations in the material. This reveals the important connection between
plasticity and the structural defects in quasicrystals. In some extent there are simi-
larities between the phenomena of crystals and quasicrystals. But the latter presents
salient structural features, fundamentally differing from those of conventional crys-
tals. The plasticity of quasicrystals must be studied in a quite different way from
that for studying the classical plasticity.
The fundamental step in studying plasticity of quasicrystals is, of course, exper-
imental observation.
As in previous discussion, this chapter mainly deals with icosahedral and decago-
nal quasicrystals.
Recently there are many experimental data for Al-Mn-Pd icosahedral quasicrys-
tals, which are the most intensively studied and possess a very high brittle-ductile
transition temperature. At strain rate of 10−5 /s the ductile range sets in at about
690 ◦ C corresponding to a homologous temperature (i.e., the absolute temperature
scaled by the absolute melting temperature) of about 0.82. At lower strain rates
of 10−6 /s and below limited ductility can be observed down to temperature about
480 ◦ C. Fig. 14.1-1 shows the stress-strain curves of the quasicrystal in the high-
temperature range 730 − 800 ◦ C at strain rate 10−5 /s.
As another class of icosahedral quasicrystals, icosahedral Zn-Ma-Dy, the stress-
strain curves measured by experiments are similar to those given by Fig. 14.1-1.
Feuerbacher and Urban[9] , Guyot and Canova[10] , Feuerbacher et al[11] , among
others, discussed the constitutive equation based on the experimental data of dislo-
cation density and dislocation velocity, e.g. the plastic strain rate and the applied
stress in a power-law form
ε̇p = B(σ/σ̂)m , (14.1-1)
in which B and m are temperature dependent parameters and σ̂ is the internal

14.1 Macroscopic behaviour of plastic deformation of quasicrystals 291
variable that can be regarded as a reference stress representing the current mi-
crostructural state of the material and is used to accommodate the model to the
description of different materials or hardening mechanism[12] . Combining relevant
information, formula (14.1-1) can be used to well predict experimental curves, e.g.
recorded by Fig. 14.1-1. It should be pointed out, though there are some similar
forms to (14.1-1) in the classical plasticity, they are quite different substantively.
For example, the current parameters B, m and σ̂ are different from those appearing
in relevant formulas in the classical plasticity. Those parameters in the classical
plasticity were measured from pure macroscopic approach rather than dislocation
model, some detail about the latter can be found in Refs. [9-12], which, from the
angle of methodology, are different from those adopted in the classical plasticity. Un-
fortunately there is lack of the comprehensive macroscopic experimental data (e.g.
the data arising from multiaxial loading condition) for quasicrystal plasticity so far.
Fig. 14.1-1 Stress-strain curves of icosahedral Al-Mn-Pd single quasicrystals at strain

rate of 10−5 /s[9]
Decagonal quasicrystals are an interesting topic for plastic deformation studies

since the phase possesses quasiperiodic as well as periodic lattice directions. There-
fore the influences of periodicity and quasiperiodicity on the plastic deformation
can be directly revealed on one material by investigating the plastic properties in
different deformation geometries—an anisotropic behaviour is expected.
In Ref. [9] the experimental data for stress-strain curves for Al-Ni-Co decagonal
quasicrystals are depicted in Fig. 14.1-2 (note that: the results obtained for the ba-
sic cobalt (b-Co) modification of decagonal Al-Ni-Co), it presents plastic properties
different from those of icosahedral quasicrystals. The reason for this is the distin-
guishing nature of quasicrystalline lattice between the two phases. It is different
from icosahedral phase, decagonal phase possesses quasiperiodic as well as periodic

lattice directions. Therefore, the influences of periodicity and quasiperiodicity on
the plastic deformation lead to an anisotropic behaviour. To reveal this feature the
Fig. 14.1-2 gives the data to those orientations as A , A⊥ and A45 respectively. It
is obvious that the effect of anisotropy is important.
Fig. 14.1-2 (a) Stress-strain curves of decagonal b-Co decagonal Al-Ni-Co single
quasicrystals for different specimen orientations at strain rate of 10−5 /s, (b) Definition of
the specimen orientations A⊥ , A and A45 [9]
The experimental data are worthy of a theoretical study.
14.2 Possible scheme of plastic constitutive equations

The formula (14.1-1) results from the fitting of experimental data and assists in
studying plastic constitutive equations of quasicrystals.
From point of view macroscopically, the result is obtained for uniaxial loading
condition as, at present there is lack of data on multiaxial loading condition. We
assume that, if sufficient experimental data are availabe on yield surface/loading
surface for some quasicrystals, then we can obtain an equation for the yield surface
such as
Φ = σeff − Y = 0, (14.2-1)
where σeff = σe + f (Hij ) denotes a generalized effective stress in which σe represents

the effective stress of the phonon stresses σij and f (Hij ) the part coming from
phason stresses Hij . If
Y = σY = const, (14.2-2)
14.2 Possible scheme of plastic constitutive equations 293
in which σY is the uniaxial yield limit of the material, then (14.2-1) represents the
initial yield surface. On the other hand if
Y = Y (h), (14.2-3)
where h is a parameter related to deformation history, then (14.2-1) describes the

evolution law of material deformation. When the yield surface/loading surface like
(14.2-1) is available, one can construct the following plastic constitutive equations
as ⎧ 1 ∂Φ
⎪
⎪ ε̇ij = σ̇eff ,
⎨ H(σeff ) ∂σij
(14.2-4)
⎪
⎪ 1 ∂Φ
⎩ ẇij = σ̇eff
H(σeff ) ∂Hij
provided that the flow rule of the isotropic hardening is taken, where Φ is the above
mentioned yield surface/loading surface function. The dot over physical quantities
denotes the variation rate of the quantities, and H(σeff ) the hardening modulus of
the material, which can be calibrated by a test of a simple stress-strain state, e.g.
given by (14.1-1). Then adding the elastic constitutive relationship to (14.2-4) the
elastic-plastic constitutive equation will be set up. The elastic constitutive law was
fully discussed in previous chapters.
The assumed constitutive law shown in equations (14.2-4) which are incremen-
tal plastic equations, can describe effect of deformation history including load-
ing/unloading state within the deformation process. This may be a complete consti-
tutive equation. There is a possible and relative simpler constitutive law belonging
to total plasticity theory or deformation plasticity theory. That is if defining the ef-
fective stress σeff and effective strain εeff , in which the former was introduced above,
and the latter has a similar definition and consists of phonon strains εij as well as
phason strains wij , then between the strains and stresses there are relations
⎧
⎪
⎪ 1 3εeff 1
⎪
⎨ ε ij − ε kk δij = σ ij − σkk δij ,
3 2σeff 3
(14.2-5)
⎪
⎪
⎪ wij − 1 wkk δij = 3εeff Hij − 1 Hkk δij
⎩
3 2σeff 3
in which
εkk = εxx + εyy + εzz , wkk = wxx + wyy + wzz , σkk = σxx + σyy + σzz ,
Hkk = Hxx + Hyy + Hzz

and we assume that
(e)
εeff , σeff < σ0 ,
εeff = (14.2-6)
A(σeff )n , σeff > σ0 ,
where σ0 , A and n are the material constants of the quasicrystals, which can be
(e)
measured through a uniaxial test, εeff represents the quantity at elastic deformation
stage and σ0 the uniaxial tensile yield stress.
In contrast to (14.2-4), the equations (14.2-5) can not describe deformation his-
tory, as they are substantively nonlinear elastic constitutive equations rather than
plastic constitutive equations related to quasicrystalline materials. However they
can describe plastic deformation in the case of proportional loading and no unload-
ing. It is evident either (14.2-4) or (14.2-5) belong to a supposed incremental plastic
constitutive law or total plastic constitutive law for quasicrystals. One cannot say
whether they are correct or not due to lack of enough experimental data.
If one has the constitutive equations (14.2-4) or (14.2-5) then by coupling the
equations of deformation geometry

1 ∂ui ∂uj ∂wi
εij = + , wij = (14.2-7)
2 ∂xj ∂xi ∂xj
and the equilibrium equations
∂σij ∂Hij
= 0, = 0, (14.2-8)
∂xj ∂xj
the basic framework of the theory of macro-plasticity of quasicrystals, in the sense
of incremental or total deformation, of quasicrystals can be set up.
At present there is lack of such data, so the equations (14.2-1), (14.2-2) and
equations (14.2-4) have not been established yet. With the same reason the equations
(14.2-5) have not been set up either. This is the major difficulty of macro-plasticity
theory currently. It is evident the possible theory is nonlinear, because the material
parameters are not constants any more, and the mechanical behaviour is dependent
with the history of deformation process in general. The solution is of course more
difficult than that for elastic deformation.
Due to relative simplicity of the equations (14.2-5), for some simple configu-
rations, e.g. anti-plane elasticity of one-dimensional hexagonal, three-dimensional
cubic and three-dimensional icosahedral quasicrystals, one can probe into plastic
analysis by using the proposed constitutive equations.
14.3 Nonlinear elasticity and its formulation

As pointed out in the previous section due to lack of constitutive equations of plas-
ticity of quasicrystals up to now, analysis of plastic deformation for the material
is not available. By this reason we can perform some simpler nonlinear elastic
analysis first, though the mechanisms of nonlinear elastic deformation and plastic
deformation are quite different. Furthermore, there are some differences between
14.4 Nonlinear solutions based on simple models 295
the following nonlinear elasticity and the total plasticity introduced by equations
(14.2-5). We here do not constrain the concrete form of relationship between stresses
and strains. The results obtained in the following may provide some useful hints for
further plastic analysis.
Consider the following nonlinear elastic constitutive relations i.e., define the free
energy (or strain energy density)
εij wij
F (εij , wij ) = σij dεij + Hij dwij , (14.3-1)
0 0
then there is
∂F ∂F
σij = , Hij = . (14.3-2)
∂εij ∂wij
So the formulas (14.3-1) and (14.3-2) can be seen as constitutive law for linear as well
as nonlinear elasticity of quasicrystals, which, in general, cannot describe the plastic
deformation. If there is proportional loading and without unloading the relationship
can give appropriate description of the plastic deformation.
In above formulas εij and wij are phonon and phason strain tensors given by

1 ∂ui ∂uj ∂wi
εij = + , wij = (14.3-3)
2 ∂xj ∂xi ∂xj
where ui and wi denote the phonon and phason displacement vectors, σij and Hij
the phonon and phason stress tensors, respectively, which satisfy the equilibrium
equations (if the body forces and generalized body forces are omitted)
∂σij ∂Hij
= 0, = 0. (14.3-4)
∂xj ∂xj
The equations (14.3-1)∼(14.3-4) are the basic equations describing nonlinear

elastic deformation of a quasicrystal. In Section 14.5 we will give some applications
of formulation (14.3-1)∼(14.3-4), and show that it constitutes the basis of those
nonlinear analysis of Section 14.4.
14.4 Nonlinear solutions based on simple models

In this section we give some nonlinear solutions based on some simple models which
substantively are extended from the framework of Eshelby s work[13] , the classical
Eshelby hypothesis was set up for linear elasticity, which can be extended into non-
linear elasticity even total plasticity (or deformation plasticity). Fan and Fan[14]
carried out the investigation.
14.4.1 Generalized Dugdale-Barenblatt model for anti-plane elasticity

for some quasicrystals
The crack problem of nonlinear deformation of quasicrystals is very interesting.
For solving this problem, one of the available models is the generalized Dugdale-
Barenblatt model (or DB model for simplicity)[15,16] originated from the classical
nonlinear fracture theory which was used for conventional structural materials in-
cluding crystalline materials.
The linear solutions for a crack for anti-plane for some quasicrystals have been
discussed in Chapters 8 and 9. We now discuss only the nonlinear solution.
Assume that there is an atomic cohesive force zone at the crack tip with length d
shown in Fig. 14.4-1, the value of which is unknown at moment to be determined. In
continuum theory of quasicrystals the atomic cohesive force zone is the plastic zone
macroscopically, and the distribution of atomic cohesive force must be evaluated
by experimental observation. Due to lack of the data of the new material, if we
assume that the atomic cohesive is a constant τc , the shear yield limit (or shear
yield strength) of the material, within the zone, then the problem is simplified. In
this version the nonlinear boundary value problem is linearized already. So the
previous formulation exhibited e.g. in Chapter 8 can be used to solve the present
problem. The most effective method for solving the problem is the complex variable
function method, but the formulation and calculation are quite complicated and
lengthy, which are much more complex than those given in previous chapters. The
details are omitted, we here list only the final results:

2i 2θ1 ζ2 1 2i e2iθ1 − ζ 2
F1 (ζ) = τ c − τ1 + τc ln −2iθ1 , F2 (ζ) = 0,
C44 πζ 2 ζ − 1 2πi C44
2 e − ζ2
(14.4-1)
Fig. 14.4-1 Atomic cohesive force zone near the crack tip for anti-plane elasticity for
some quasicrystals
in which
R3
F1 (ζ) = φ (ζ) + ψ (ζ),
C44
R3
F2 (ζ) = ψ (ζ) + φ (ζ)
K2
and
φ(ζ) = φ1 (t) = φ1 (ω(ζ)), ψ(ζ) = ψ1 (t) = ψ1 (ω(ζ)),
where
a+d
t = ω(ζ) = (ζ + ζ −1 )
2
represents the conformal mapping from t-plane (xy-plane) onto ζ-plane, under the
mapping the region of xy-plane is transformed onto the interior of the unit circle
at ζ-plane, and φ (ζ) and ψ (ζ) are derivatives of the functions to the new complex
variable ζ, and θ1 is the angle of point at the unit circle corresponding to the crack tip
(i.e., y = 0, x = a, and there is relation cos θ1 = a/(a + d) between the corresponding
points at t-plane and ζ-plane).
From the solution, the size of plastic zone d is determined as

πτ1
d = a sec −1 , (14.4-2)
2τc
πτ1
i.e., θ1 = and the crack tip tearing displacement is
2τc

4K2 τc a πτ1
δIII = ln sec . (14.4-3)
π(C44 K2 − R32 ) 2τc
It is obvious that the results are very simple and explicit, and the effects of phason to
plastic deformation and fracture are explored. If taking the tearing displacement as a
nonlinear fracture parameter, then a fracture criterion for one-dimensional hexagonal
quasicrystal is
δIII = δIIIC (14.4-4)
is suggested in which δIIIC is the critical value the crack tip tearing displacement,
measured by experiment, a material constant.
In the low stress level case, i.e., τ1 /τc 1, the deformation is linear elastic, in
which 2
πτ1 1 πτ1
sec =1+ + ···
2τc 2 2τc
if we remain the first two terms then
2 4
πτ1 1 πτ1 1 πτ1
ln sec = + + ···
2τc 2 2τc 12 2τc
so that in the linear case

GIII
δIII = . (14.4-5)
τc
This shows that parameter δIII is equivalent to the energy release rate GIII in the
linear case.
The above results are obtained for one-dimensional hexagonal quasicrystals,
which hold for three-dimensional icosahedral quasicrystals, if the material constants
C44 , K2 and R3 are replaced by μ, K1 − K2 and R; which hold for three-dimensional
cubic quasicrystals too, if the constants are replaced by C 44 , K44 and R44 , respec-
tively.
14.4.2 Generalized Dugdale-Barenblatt model for plane elasticity of two-

dimensional point groups 5m, 10mm and 5, 5, 10, 10 quasicrystals
The linear analysis for a crack in plane elasticity in point groups 5m and 10mm
as well as in groups 5, 5 and 10, 10 of two-dimensional quasicrystals are given in
Sections 8.3 and 8.4 in terms of the Fourier method and complex variable function
method respectively, where we pointed out that if taking R1 = R, R2 = 0 then the
solution on point groups 5, 5 and 10, 10 will exactly reduces to that on point groups
5m and 10mm. In the following we discuss the nonlinear solutions only based on the
complex variable function method.
For two-dimensional quasicrystals a similar generalized Dugdale-Barenblatt model
to that given in the Subsection 14.4.1 is suggested as below, refer to Fig. 14.4-2. That
is, there is the so-called atomic cohesive force zone, or the plastic zone macroscop-
ically, with length d, temporarily unknown and to be determined. If we suppose
that the distribution of the atomic cohesive force distribution and magnitude of the
atomic cohesive force σc = σc (x) are known, further more σc =constant within the
zone, then the nonlinear problem is linearized already, in which σc represents the
Fig. 14.4-2 The generalized DB model for two-dimensional decagonal quasicrystal

macro-tensile yield strength of the quasicrystal. Therefore the formulation intro-

duced above can be used to solve the present problem. The powerful method for
the problem is the complex variable function-conformal mapping method, which
was displayed briefly in the Subsection 14.4.1. But the calculation is much more
complicated and lengthy than those listed in the previous subsection. The detail of
complex variable function-conformal mapping method is given in Subsection 11.3.10
in Chapter 11 and Subsection A.I.9 in Major Appendix, here we list the final plastic
solution only as follows.
The size of the plastic zone is determined as

πp
d = a sec −1 , (14.4-6)
2σc
which is similar to that given by (14.4-2), but the meaning is different. And the
crack tip opening displacement is for point groups 5m and 10mm quasicrystals

2σc a 1 K1 πp
δI = + ln sec (14.4-71 )
π L+M M K1 − R 2 2σc
and for point groups 5, 5, 10, 10 ones

2σc a 1 K1 πp
δI = + ln sec , (14.4-72 )
π L+M M K1 − (R12 + R22 ) 2σc
which is also similar to that obtained by (14.4-3) in which L = C12 , M = (C11 −
C12 )/2 are the phonon elastic constants, K1 the phason elastic constant and R, R1 , R2
the phonon-phason coupling constants respectively. It is very interesting that the
solution (14.4-7) exactly covers the solutions of crystals and conventional structural
materials, as K1 = R = 0, e.g. for classical conventional materials
⎧
⎪ (1 + κ)σc a π p 4(1 − ν)σc a π p
⎪
⎪ ln sec = ln sec ,
⎪
⎪ πμ 2 σc πμ 2 σc
⎪
⎪
⎨ plane strain state,
δt = CT OD =
⎪
⎪ (1 + κ )σc a π p 4σs a π p
⎪
⎪ ln sec = ln sec ,
⎪
⎪ πμ 2 σc (1 + ν)πμ 2 σc
⎪
⎩
plane stress state,
3−ν
where κ = 3−4ν for plane strain, κ = for plane stress, ν is the Poisson’s ratio
1+ν
of the materials. So the solutions of crystals and conventional structural materials
are the specific cases of our solution of quasicrystals. Fig. 14.4-3 shows the normal-
ized crack tip opening displacement δI /a versus normalized stress p/σc and gives the
comparison between quasicrystals and crystals as well as the comparison between
point groups 10mm and 10, 10 quasicrystals. The detail of solution including point
group 10, 10 is given in Subsection A.I.9 of Major Appendix of this book.
Fig. 14.4-3 Normalized crack tip opening displacement versus normalized stress and
comparison[27]
Based on the parameter δI given by (14.4-7) we can suggest the nonlinear fracture
criterion for pentalgonal and/or decagonal quasicrystals for the Mode I loading as
δI = δIC , (14.4-8)
in which δIC denotes critical value of the crack opening displacement, measured by
experiment, a material constant.
In linear elastic case, i.e., p/σc 1, through a similar analysis like in the previous
subsection the crack tip opening displacement reduces to
GI
δI = , (14.4-9)
σc
where GI is defined by (8.3-18), this gives the simple connection between these two
parameters, so the criterion (14.4-8) reduces to energy release rate criterion for linear
elastic case discussed in Chapter 8.
14.4.3 Generalized Dugdale-Barenblatt model for plane elasticity of three-

dimensional icosahedral quasicrystals
The schematic figure of the model for icosahedral quasicrystals is similar to Fig.
14.4.2, the plastic zone is the same given by (14.4-6), and the crack tip opening
displacement is as follows

1 c4 σc a π p
δI = lim 2uy (x, 0) = lim 2uy (x, 0) = 2 + · · ln sec
x→l ϕ→ϕ2 λ + μ c2 π 2 σc
(14.4-10)
14.5 Nonlinear analysis based on the generalized Eshelby theory 301
in which

(μK2 − R2 )2
2 1 μK1 − 2R2
c2 = μ(K1 − K2 ) − R − , c4 = c1 R + c2 K1 +
μK1 − 2R2 2 λ+μ
(14.4-11)
R(2K2 − K1 )(μK1 + μK2 − 3R2 )
with c1 = .
2(μK1 − 2R2 )
The variation of the normalized crack tip opening displacement versus the normalized
applied stress is shown in Fig. 14.4-4.

Icosahedral quasicrystal Al-Mn-Pd
Icosahedral quasicrystal Al-Cu-Li
Conventional crystal aluminum

δI/a

p/σc
Fig. 14.4-4 Normalized crack tip opening displacement of icosahedral quasicrystals

versus normalized applied stress [27]
14.5 Nonlinear analysis based on the generalized Eshelby the-

ory
In the previous section we obtained some nonlinear solutions for one-, two-and three-
dimensional quasicrystals by using some simple physical models. We will show that
those solutions have some inherent relations to the generalized Eshelby theory orig-
inated from the classical reference [13] for crystals. Fan and Fan[14] did some work
to extend the classical Eshelby model for crystals to that for quasicrystals.
14.5.1 Generalized Eshelby energy-momentum tensor and generalized

Eshelby integral
Fan and Fan[14] (also refer to Ref. [18]) defined the generalized Eshelby energy-
momentum tensor as
∂ui ∂wj
G = F n1 − σij nj − Hij nj , (14.5-1)
∂x1 ∂x1
where F is defined by (14.3-1) and ni the unit vector of outward normal at any point
of an arc in a quasicrystal, and at which there are
σij nj = Ti , Hij nj = hi ,
where Ti denotes the traction vector and hi the generalized traction vector, shown
in Fig. 14.5-1, respectively.
Fig. 14.5-1 The path of the generalized Eshelby integral
Furthermore we define an integral such as

E= GdΓ , (14.5-2)
Γ
in which Γ is an integration path enclosing a crack tip in the material. To memory

Eshelby the integral is named generalized Eshelby integral. The integral exhibits
path independency. In Section 14.8, i.e., Appendix of Chapter 14, the proof on the
path-independency of the integral is given.
Because the integral presents the above important character, it has some appli-
cations in fracture analysis of quasicrystals.
Another important character is the value of the integral which is equal to the
energy release rate GIII (under Mode III loading), or GII (under Mode II loading), or
GI (under Mode I loading) of one-, two- and three-dimensional quasicrystals when
the material is in linear elastic deformation state. The mathematical proof about
this will be provided in Section 14.8, i.e., Appendix of Chapter 14.
The features of the generalized Eshelby integral (14.5-2) exhibited above imply
that it may be significant for the analysis of nonlinear fracture of quasicrystals. In
the following an application for this purpose is discussed.
14.5 Nonlinear analysis based on the generalized Eshelby theory 303
14.5.2 Relation between crack tip opening displavement and the gener-
alized Eshelby integral
By using the path-independency of E-integral (14.5-2), we take the integration path
Γ shown in Fig. 14.5-2 and let the path close to
the surface of the plastic zone as
close as possible. Now the integration path is , and along segment AB and
ACB
BC, i.e.,
dy = 0, T1 = 0, T 2 = σc , h2 = Hc
so that
B
∂u ∂wy
E= GdΓ = GdΓ = dx + Hc
σc dx
Γ ACB A ∂x ∂x
= σc [(uy )B − (uy )A ] + Hc [(wy )B − (wy )A ] ≈ σc [(uy )B − (uy )A ] = σc δI . (14.5-3)
Fig. 14.5-2 The integration path for evaluating crack tip opening displacement
This proves the relation between the generalized Eshelby integral and crack tip
opening displacement, it shows the E-integral presents an equivalency to the crack
tip opening displacement, where σc is the atomic cohesive force (or the plastic yield
strength in macroscopic sense) and the “generalized atomic cohesive force” of the
quasicrystal material. In microscopy the quantity is meaningful, but it has not
been measured macroscopically at present, the effect of which is omitted in formula
(14.5-3) by this reason. For Mode II and Mode III crack the proof is similar.
Since by the generalized BCS model and generalized DB model one can obtain
the same result on crack tip opening displacement, which is also as a direct result
of the generalized Eshelby integral for quasicrystals, we realize that the generalized
Eshelby energy-momentum tensor theory is the uniform physical basis of the two
models.
14.5.3 Some further interpretation on application of E-integral to the

nonlinear fracture analysis of quasicrystals
The above subsections and the appendices exhibit the generalized Eshelby integral
can be as a uniform basis of the generalized BCS model as well as the generalized
DB model for nonlinear fracture analysis of some one-, two- and three-dimensional
quasicrystals. We suggest taking the E-integral as a fracture parameter, and
E = Ec (14.5-4)
as fracture criterion, where Ec is the critical value of the integral, a material constant,
which can be measured through some conventional specimens, the discussion will be
given in Section 14.8 (the Appendix of Chapter 14).
Though the measurement of the critical value of the generalized Eshelby integral
may be easier, but the evaluation of values of E-integral for plastic deformation of
quasicrystals is very difficult. Therefore the implementation of the criterion (14.5-
4) is not so convenient in practice. Instead, people can use fracture criterion of
crack tip opening displacement considering the equivalency between the crack tip
opening displacement δ(δI , δII , δIII ) and E-integral, we suggest taking the crack tip
opening displacement as a fracture parameter. Elasto-plastic crack solutions for
some one- and two-and three-dimensional quasicrystals have been found based on
the generalized Dugdale-Barenblatt model, refer to Refs. [14, 18, 27], in which the
size of the plastic zone and crack tip opening displacement are determined, with
these data an equivalent plastic fracture criterion is suggested as
δ = δc , (14.5-5)
in which δc represents the critical value of the crack tip opening displacement, a
material constant. The evaluation of crack tip opening displacement has been exactly
completed for large plates with central crack and narrow plastic zone and introduced
in Section 14.4 or refer to Refs. [14, 18, 27], and some approximate solution for other
configuration can be obtained by complex variable function method (an outline of
the method may be interpreted in Refs. [14, 18, 27]) and other methods (mainly the
approximate methods and numerical methods). The shortcoming of this criterion
lies in some difficulties of the determination of δc , but which can be obtained through
relation
Ec
δc = . (14.5-6)
σc
and the Section 14.8 shows the measurement of Ec is easier.
Thus, the collaboration of criterions (14.5-4) and (14.5-5) makes the nonlinear
fracture analysis for quasicrystals possible. Some details will be given in the Ap-
pendix of this chapter (i.e., Section 14.8).
14.6 Nonlinear analysis based on the dislocation model 305
14.6 Nonlinear analysis based on the dislocation model

We pointed out at the begining of this chapter that in the study on the classical
plasticity there are two different theories, one is the macroscopic plasticity theory,
and the other is so-called crystal plasticity theory, in some extent the latter can
be seen as a “microscopic” theory. which is based on the mechanism of motion of
dislocation. We discussed quite more dislocation solutions in Chapter 7 for one- and
two-dimensional quasicrystals and in Chapter 9 for three-dimensional quasicrystals,
those results are helpful to explore the plastic deformation and fracture of the ma-
terial. Because this analysis needs not to rely on nonlinear macroscopic constitutive
law and it reduces a great difficulty in mathematical treatment. Here we focus
on the discussion concerning plastic flow around crack tip for some one-, two- and
three-dimensional quasicrystals, respectively.
14.6.1 Screw dislocation pile-up for hexagonal or icosahedral or cubic

quasicrystals
For the topic the basis is the generalized BCS model developed by Fan and co-
workers[8] , which is stated as below.
Assume there is a Griffith crack with length 2l along the direction of z axis in a
one- or three-dimensional quasicrystal, subjected to a longitudinal shear stress τ (∞)
at infinity see Fig. 14.6-1. Around the crack tip there is a dislocation pile-up with size
d which is unknown so far to be determined. Within the zone of dislocation pile-up
the material presents plastic flow, the stress σyz is equal to the critic shear stress τc ,
the atomic cohesive force, or so-called the flow limit of the material macroscopically.
For simplicity the external applied stress at the infinity can be removed, instead it
is applied at the crack surface. The latter is equivalent with the former from point
of view of fracture behaviour for any studied systems. This is the generalized BCS
model
Fig. 14.6-1 Schematic picture of a crew dislocation pile-up coupled with a crack in
anti-plane elasticity of quasicrystals
for icosahedral quasicrystals. The origin of the model was given by Bilby, Cottrell
and Swinden[20] and Bilby, Cottrell, Smith et al[21] for crystals, we developed model
to the study for some one- and two-dimensional quasicrystals[8] . The model can
be formulated by the following boundary conditions for one-dimensional hexagonal
quasicrystals:
(x2 + y 2 )1/2 → ∞ : σij = 0, Hij = 0,
y = 0, |x| < l : σyz = −τ (∞) , Hyz = 0, (14.6-1)
y = 0, l < |x| < l + d : σyz = −τ (∞) + τc = 0, Hyz = 0.
Some detailed statement for the formulation can refer to Ref. [18].
The boundary conditions (14.6-1) has been shown that the nonlinear (plastic
deformation) problem is linearized mathematically. So it has been transformed into
a linear problem (or an “equivalent” elasticity problem) described by the final gov-
erning equations
∇2 uz = 0, ∇2 wz = 0. (14.6-2)
and boundary conditions (14.6-1), detail about this can be found in the text of
Chapters 5, 7 and 8. The nonlinear fracture problem can be solved if we can obtain
the solution of equation (14.6-2) under boundary conditions (14.6-1). It is obvious
the problem is complicated. Though we can solve the problem in the methodology
developed in Chapter 8, but there is simpler procedure to solve it if we are interested
in only some plastic deformation parameters around crack tip and if we have a
dislocation solution, the latter is given in Section 7.1. For this purpose we first
introduce a dislocation density function f (ξ), then problem (14.6-1) and (14.6-2)
can transformed into the following singular integral equation problem such as

f (ξ)dξ τ (x)
= , (14.6-3)
L ξ−x A
in which ξ denotes the dislocation source point coordinate, x denotes the field point
coordinate at the real axis and L represents interval (l, l + d). From the dislocation
solution given in Section 7.1 the constantA is defined by

−τ (∞) , |x| < l,

τ (x) = (∞) (14.6-4)
−τ + τc , l < |x| < l + d,

(C44 K2 − R32 )b3
A=− .
2πK2
In terms of the Muskhelishvili[26] theory, Fan et al[8] solved the integral equation
(14.6-3) under condition (14.6-4), the solution is
& &
1 x + (l + d) ξ − (l + d) dξ
f (x) = − 2 τ (ξ)
π A x − (l + d) L ξ + (l + d) ξ−x
14.6 Nonlinear analysis based on the dislocation model 307
&

1 x + (l + d) l (∞)
=− 2 i 2τc arccos −τ π
π A x − (l + d) l+d
" " " " (14.6-5)
τc " (l + d)2 − lx " " (l + d)2 + lx "
+ 2 "
arc cosh " " "
− arc cosh " " .
π A (l + d)(l − x) " (l + d)(l + x) "
Because the function f (x) should be a real number, the factor multiplying the
imaginary number i must be zero in the first term of right-handside of (14.6-5), this
leads to
l
2τc arccos − τ (∞) π = 0,
l+d
i.e., (∞)
πτ
d = l sec −1 . (14.6-6)
2τ c
This is the same as (14.4-2) if l is replaced by a.
From solution (14.6-5) we evaluate amount of dislocations N (x) such as
x
N (x) = f (ξ)dξ. (14.6-7)
0
Substituting (16.6-5) (coupled with (14.6-6)) into (14.6-7), we can get N (l + d) and
N (l), so the amount of dislocation motion is

2b lτc l+d 4K2 τc l πτ (∞)
δIII = b3 [N (l + d) − N (l)] = 32 ln = ln sec .
π A l π(C44 K2 − R32 ) 2τc
(14.6-8)
This is the same as (14.4-3) apart from difference of notations.
14.6.2 Edge dislocation pile-up for pentagonal or decagonal two-

dimensional quasicrystals
For plane elasticity the corresponding problem is “edge” dislocation shown in Fig.
14.6-2, here we consider the pile-up in the pentagonal or decagonal two-dimensional
quasicrystals. There are the boundary conditions
⎧ 2
⎨ (x + y 2 )1/2 → ∞ : σij = 0, Hij = 0,
y = 0, |x| < l : σyx = −τ (∞) , Hyx = 0, σyz = 0, Hyz = 0,
⎩
y = 0, l < |x| < l + d : σyx = −τ (∞) + τc , Hyx = 0, σyz = 0, Hyz = 0,
(14.6-9)
and the governing equation
∇2 ∇2 ∇2 ∇2 F = 0. (14.6-10)
Instead solving the boundary value problem (14.6-9) and (14.6-10), one can de-
termine the dislocation density which satisfies the integral equation (14.6-3) if the
relevant quantity A is known beforehand.
Fig. 14.6-2 Schematic picture of an “edge” dislocation pile-up coupled with a crack in a
plane elasticity of quasicrystals
In the singular integral equation the shear stress distribution function, τ (x) is
the same as before, but here A is replaced by:

b1 (L + M )(M K1 − R2 )
A= . (14.6-11)
π (L + M )K1 + (M K1 − R2 )
which has been found in Section 7.2 for point groups 5m and 10mm quasicrystals.
According to the condition concerning factor multiplying the imaginary i be zero
in (14.6-5), the plastic zone size is determined by (14.6-6), this is similar to (14.4-6),
but the physical meaning is somewhat different.
By a similar calculation on amount of dislocation motion we can obtain the slip
of the crack tip such as
(∞) (∞)
2b l τc (L+M ) 2τ 2τc l 1 K1 2τ
δII = 21 ln sec = + 2
ln sec .
π A M +K1 πτc π M +K1 M K1 −R πτc
(14.6-12)
It is evident that the result of (14.6-12) is similar to that of (14.4-7), but the phys-
ical meaning is different. According to the macroscopic fracture mechanics, the
crack tip slip corresponding to the Mode II crack tip displacement (i.e., the crack
tip sliding displacement) δII . But for the Mode I crack tip opening displacement
(14.4-7) there is no physical basis of dislocation model, though the similar math-
ematical treatment can be easily given, which is omitted here. Even if this, the
dislocation model provides a powerful support to the nonlinear analysis of previous
sections.
14.6.3 Edge dislocation pile-up for three-dimensional icosahedral qua-

sicrystals
(x2 + y 2 )1/2 → ∞ : σij = 0, Hij = 0,

14.7 Conclusion and discussion 309
y = 0, |x| < l : σyx = −τ (∞) , Hyx = 0, σyz = 0, Hyz = 0, σyy = 0, Hyy = 0,

y = 0, l < |x| < l + d : σyx = −τ (∞) + τc , Hyx = 0, σyz = 0, Hyz = 0, σyy = 0,
Hyy = 0,
∇2 ∇2 ∇2 ∇2 ∇2 ∇2 F (x, y) = 0.
The derivation can be offered similar to above. Because the formulas are too
lengthy (please refer to Section 9.6), the discussion is omitted.
The dislocation model provides a powerful support to the generalized cohesive
force model and generalized Eshelby model presented in previous sections, even if
there is difficulty physically, we are unable to derive δI directly from the model, and
results in a slight faulty aspect of the model.

This chapter discussed the deformation and fracture exhibiting nonlinear behaviour
of quasicrystals, this is a topic which has not been well studied so far. The discus-
sion in the chapter shows that the nonlinear elastic constitutive equation possesses
some meaning for the investigation due to lack of plastic constitutive equation at
present. The generalized Eshelby principle based on the nonlinear elastic constitu-
tive law plays an important role in revealing the fracture behaviour of quasicrystals
under nonlinear deformation which may be a basis of generalized DB model in some
extent. We realize that the generalized Eshelby principle and generalized DB model
are based on the macro-nonlinear continuous mechanics of quasicrystals. The gener-
alized BCS model developed in Section 14.6 emphasizes dislocation mechanism for
the plastic deformation, it is different from those of the generalized DB hypothesis
as well as the generalized Eshelby principle at physical basis and methodology. The
attractiveness is that results derived from these quite different hypothesizes achieve
the exact agreement in the study.
14.8 Appendix of Chapter 14—Some mathematical details

14.8.1 Proof on path-independency of E-integral
Consider a single-connected region D with a closed boundary C. From the Green

formula the first, second and third terms of the left-handside of the integral (14.5-2)
will be
∂F
F n1 dΓ = F dx2 = dx1 dx2 ,
C C ∂x1
D

∂ui ∂wi ∂ ∂ui ∂wi
σij nj + Hij nj dΓ = σij + Hij dx1 dx2 .
C ∂x1 ∂x1 ∂xj ∂x1 ∂x1
D
So that

∂F ∂ ∂ui ∂wi
GdΓ = − σij + Hij dx1 dx2 . (14.8-1)
C ∂x1 ∂xj ∂x1 ∂x1
D
Meantime, (14.3-2) provides that

∂F ∂F ∂εij ∂F ∂wij ∂εij ∂wij
= + = σij + Hij
∂x1 ∂εij ∂x1 ∂wij ∂x1 ∂x1 ∂x1

∂ 1 ∂ui ∂uj ∂ ∂wi
= σij + + Hij
∂x1 2 ∂xj ∂xi ∂x1 ∂xj

∂ ∂ui ∂ ∂wi
= σij + Hij ,
∂x1 ∂xj ∂x1 ∂xj
where the symmetry of tensor εij (given by the first equation of (14.3-3)) is used.
The above result can be rewritten as

∂F ∂ ∂ui ∂σij ∂ui ∂ ∂wi ∂Hij ∂wi
= σij − + Hij −
∂x1 ∂xj ∂x1 ∂xj ∂x1 ∂xj ∂x1 ∂xj ∂x1

∂ ∂ui ∂wi
= σij + Hij , (14.8-2)
∂xj ∂x1 ∂x1
in the last step in driving (14.8-2) the equilibrium equations (14.3-4) are used.
Substituting (14.8-2) into (14.8-1) we have

GdΓ = 0.
C
If taking C = Γ + BB − Γ + A A, in which Γ is a path similar to Γ , these two

curves intersect the upper face of the crack at points B and B , and the lower face
at points A and A , respectively. Due to
dx2 = dy = 0, Ti = 0, hi = 0
at the segments BB and A A, there are

GdΓ = 0, GdΓ = GdΓ .
BB +A A Γ Γ
Because Γ and Γ are arbitrary paths satisfying the requirements of E-integral

definition, the last equality proves its path-independency.
14.8.2 Proof on the equivalency of E-integral to the energy release rate

under the linear elastic case for quasicrystals
In the text of this chapter we argue that the E-integral is equivalent to the en-
ergy release rate when the quasicrystals are in linear elastic deformation state, the
14.8 Appendix of Chapter 14—Some mathematical details 311
mathematical proof about this is given here. The proof can be done for one-, two-
and three-dimensional quasicrystals and for Mode I, Mode II and Mode III cracks.
For simplicity here we only discuss Mode III crack for one-dimensional hexagonal
quasicrystals.
At the linear elastic case, the generalized Hooke’s law
Hij = Kijkl wkl + Rklij εkl
is reduced to
σyz = σyz = 2C44 εyz + R3 wzy ,
σzx = σxz = 2C44 εzx + R3 wzx ,
Hzy = K2 wzy + 2R3 εzy ,
Hzx = K2 wzx + 2R3 εzx
for the phonon-phason coupling anti-plane shearing (or longitudinal shearing) state
in elasticity of one-dimensional hexagonal quasicrystals, one finds the free energy
(or strain energy density) is
1
F = C44 (ε2xz + ε2yz ) + K2 (wzx
2 2
+ wzy ) + R3 (εzx wzx + εyz wzy ). (14.8-3)
2
Substituting the crack solution given in Chapter 8 into (14.8-3) we have
1 ⊥ 2 ⊥
F = {C44 [K2 KIII − R3 KIII ] + 2K2 [C44 KIII − R3 KIII ]2
4(C44 K2 − R32 )
⊥ ⊥
+ 2R3 [K2 KIII − R3 KIII ][C44 KIII − R3 KIII ]}, (14.8-4)
√ ⊥
√
where KIII = πaτ1 and KIII = πaτ2 are the stress intensity factors associated
with phonon and phason fields respectively.
Integrating the quantity given by (14.8-4) around a path enclosing the crack
tip, then one can obtain the first term of the E-integral. Because of the path-
independency of the integral, we can take a half-circle with the crack tip as its origin
and with radius r, such that
−π −π
F dy = F r cos θdθ = 0, (14.8-5)
π π
where (r, θ) denote the crack tip coordinates.

Now we calculate the second and third terms of the integral.
According to the definition of Section 14.3, it is known that between the traction,
generalized traction and the phonon, phason stresses there are
Tz = σzx nx + σzy ny = σzx cos θ + σzy sin θ,
hz = Hzx nx + Hzy ny = Hzx cos θ + Hzy sin θ,

by substituting the crack solution given in Section 8.1 into the above formulas, we
find
1 1 1 ⊥ 1
Tz = √ KIII sin θ, hz = √ KIII sin θ. (14.8-6)
2πr 2 2πr 2
Based on the crack solution given in Section 8.1, there are
⎧ K 2 τ 1 − R3 τ 2 √ 1
⎪
⎨ uz = 2ar sin θ,
C44 K2 − R3 2 2
C τ − R3 τ1 √ (14.8-7)
⎪
⎩ wz = 44 2
1
2ar sin θ.
C44 K2 − R32 2
In addition,
∂ ∂ ∂r ∂ ∂θ ∂ ∂
= + = cos θ − sin θ .
∂x ∂r ∂x ∂θ ∂x ∂r ∂θ
Substituting these results and relations into the second and third terms of the
integral we obtain
−π
∂ui ∂wi K2 (KIII )2 + C44 (KIII
⊥ 2 ⊥
) − 2R3 KIII KIII
− σij nj + Hij nj dΓ = 2 .
π ∂x1 ∂x1 C44 K2 − R3
(14.8-8)
From (14.8-5) and (14.8-8), we find that

K2 (KIII )2 + C44 (KIII
⊥ 2 ⊥
) − 2R3 KIII KIII
EIII = 2 = GIII . (14.8-9)
C44 K2 − R3
which is just (8.1-25) given in Chapter 8.

Similarly one can obtain
EI = GI (14.8-10)
and
EII = GII , (14.8-11)
where the suffixes mean the Mode I, Mode II and Mode III of the cracks. The above
demonstration indicates that the generalized Eshelby integral is equivalent to the
energy release rate of crack in linear elasticity of quasicrystals.
Under plastic deformation of quasicrystals, in general, the generalized Eshelby
integral does not represent energy release rate, the reason of this is the unloading
can appear due to crack extension, and stress-strain relations described by (14.3-1)
14.8 Appendix of Chapter 14—Some mathematical details 313
and (14.3-2) do not remain one to one correspondence, the physical background of
the E-integral does not hold. But if we define the potential energy per unit thickness
for a plane (i.e., a two-dimensional) region Ω occupied by a quasicryslal, and denote
the boundary of the region by Γ

V = F dxdy − (Ti ui + hi wi )dΓ , (14.8-12)
Ω Γ
then the total potential energy is
Π = BV, (14.8-13)
where B denotes the thickness of the specimen, and there is relation

δΠ 1 δΠ
E=− =− , (14.8-14)
δA B δa
where E-integral can be expressed only by the “difference quotient” rather than the
“differential quotient”. The difference quotient is not the energy release rate, only
represents the ratio between individual difference value of the energy over difference
value of crack length for the same configuration specimen but with different initial
crack lengths. Considering the situation the physical meaning of (14.8-14) is left
for further discussion. Even if there is the faulty aspect, (14.8-14) provides useful
application for calibrating the measurement data of the integral in experiments. The
relevant discussion will be done in the next subsection.
14.8.3 On the evaluation of the critic value of E-integral

The E-integral provides not only the theoretical basis for generalized BCS model as
well as generalized DB model but also an effective tool for measuring the material
constants Ec and δc . We mentioned that the measurement of δc is difficult, but the
measurement of Ec is easier. Due to the connection between these two quantities
the value of the former can be obtained from that of the latter.
At low and conventional temperature quasicrystals present brittle behaviour, the
fracture belongs to linear elastic fracture, and the measurement of fracture toughness
can be carried out by indentation technique, see e.g. Meng et al[19] . At high tem-
perature, the materials present large plastic deformation, the measurement can be
done by the bending experiment. The three point bending specimen, see Fig. 14.8-1,
is one of common specimens. There are two procedures either by multi-specimens
testing or by single specimen testing.
For simplicity we here discuss the single specimen testing only. For the three
point bending specimen shown in Fig. 14.8-1, based on formula (14.8-14) the value
of E-integral is approximately by
Fig. 14.8-1 Three point bending specimen
2U
E= , (14.8-15)
B(W − a)
where W represents the width, B the thickness of the specimen and U the area
under P -Δ curve (shown by Fig. 14.8-2), i.e.,

U = P dΔ, (14.8-16)
in which P is the load (force per unit thickness) at loading point, and Δ the dis-
placement of the same point. When the initiation of crack growth is observed, then
the value of E-integral is marked as the fracture toughness of the quasicrystal.
If the phason field is absent, the material is
degenerated to conventional structural material.
In this case, the E-integral is reduced to conven-
tional Eshelby integral or J-integral, the latter
was introduced by Rice[22] and Cherepanov[23] .
Begley and Landes[24,25] put forward the exper-
imental study on J-integral, further promoting
the development of nonlinear fracture theory and
its applications for conventional engineering ma-
terials. These experiences would be helpful for
Fig. 14.8-2 The depiction of the the experimental study of nonlinear fracture of
specimen deformation energy quasicrystalline materials.
References
[1] Calliard D. Dislocation mechanism and plasticity of quasicrystals: TEM observations
in icosahedral Al-Pd-Mn. Materials Sci Forum, 2006, 509(1): 49–56
[2] Geyer B, Bartsch M, Wollgarten M et al. Plastic deformation of icosahedral Al-Pd-
Mn single quasicrystals. Experimental results. Phil Mag A, 2000, 80(7): 1151–1164
References 315
[3] Messerschmidt U, Bartsch M, Wollgarten M et al. Plastic deformation of icosahedral

Al-Pd-Mn single quasicrystals. II. Interpretation of experimental results. Phil Mag
A, 2000, 80(7): 1165–1181
[4] Urban K, Wollgarten M. Dislocation and plasticity of quasicrstals. Materials Sci
Forum, 1994, 150–151(2): 315–322
[5] Wollgarten M, Bartsch M, Messerschmidt U et al. In-situ observation of dislocation
motion in icosahedral Al-Pd-Mn single quasicrystals. Phil Mag Lett, 1995, 71(1):
99–105
[6] Feuerbacher M, Metzmacher C, Wollgarten M et al. 1997, Plastic deformation of
decagonal Al-Ni-Co quasicrystals. Phil Mag Lett, 76(4): 396–375
[7] Messerschmidt U, Bartsch M, Feuerbacher M et al. Friction mechanism of dislocation
motion in icosadedral Al-Pd-Mn single quasicrystals. Phil Mag A, 1999, 79(11):
2123–2135
[8] Fan T Y, Trebin H-R, Messerschmidt U et al. Plastic flow coupled with a Griffith
crack in some one- and two-dimensional quasicrystals. J Phys: Condens Matter,
2004, 16(47): 5229–5240
[9] Feuerbacher M, Urban K. Platic behaviour of quasicrystalline materials. in: Qua-
sicrystals, Trebin H R, Wiely Press, Berlin, 2003
[10] Guyot P, Canova G. The plasticity of icosahedral quasicrystals. Phil Mag A, 1999,
79(11): 2815–2822
[11] Feuerbacher M, Schall P, Estrin Y et al. Aconstitntive model for quasicystal plastic-
ity. Phil Mag Lett, 2001, 81(7): 473–482
[12] Estrin Y. in: Unified Constitutive Laws of Plastic Deformation, Krausz A, Krausz
K, Academic Press. New York, 1996
[13] Eshelby J D. The continuum theory of dislocations in crystals, Solid State Physics
(ed by Seits F et al). Vol. 3. New York: Academic Press, 1956
[14] Fan T Y, Fan L. Plastic fracture of quasicrystals. Phil Mag, 2008, 88(4): 523–535
[15] Dugdale D S. Yielding of steel sheets containing slits. J Mech Phys Solids, 1960,
32(2): 105–108
[16] Barenblatt G I. The mathematical theory of equilibrium of crack in brittle fracture.
Advances in Applied Mechanics. 1962, 7: 55–129
[18] Fan T Y, Fan L. Relation between generalized Eshelby integral and generalized BCS
and generalized D B models for some one- and two-dimensional quasicrystals. Chin.
Phys. B in press, 2010
[19] Meng X M, Tong B Y, Wu Y K. Mechanical properties of quasicrystals Al65 Cu20 Co15 .
Acta Metallurgica Sinica A, 1994, 30(3): 61–64 (in Chinese)
[20] Bilby B A, Cottrell A H, Swinden K H. The spread of plastic yield from a notch.
Proc R Soc A, 1963, 272(2): 304–314
[21] Bilby B A, Cottrell A H, Smith E et al. Plastic yealding from sharp notches, ibid,
1964, 279(1): 1–9
[22] Rice J R. A path independent integral and approximate analysis of strain concentra-
tion by notches and cracks. J Appl Mech, 1968, 35(4): 379–386
[23] Cherepanov G P. On crack propagation in solids. Int J Solids and Structures, 1969,
5(8): 863–871
[24] Begley G T, Landes J D. 1972, The J-integral as a fracture criterion, Fracture Tough-
ness, ASTM STP 514, American Society for Testing and Materials, Philadelphia,
1–20
[25] Landes J D, Begley G T. 1972, The effect of specimen geometry on JIC , Fracture
Toughness, ASTM STP 514, American Society for Testing and Materials, Philadel-
phia, 24–29
[26] Muskhelishvili N I. Singular Integral Equations. Groningen, Noordhoff, 1956
[27] Li W, Fan T Y. Study on plastic analysis of crack problem in icosahedral quasicrys-
tals. Phil. Mag., 2009, 89(31): 2823–2831
Chapter 15
Fracture theory of quasicrystals
Quasicrystals are brittle, study on fracture behaviour of the material is significant. In

the previous chapters many crack problems are investigated, and the exact analytic,
approximate and numerical solutions are constructed. This provides a basis for
discussing the fracture theory. The exact solutions of crack problem of quasicrystals
present particular importance, which reveal the essential nature describing fracture
behaviour of this kind material. The exact analytic solution on a Griffith crack in
decagonal quasicrystals was given by Li and Fan[1] in 1999. Afterward Fan[2] , Fan
and Mai[3] developed the idea of linear elastic fracture mechanics of quasicrystals
based on the common feature of crack solutions obtained in that date. Rudhart
et al[4] discussed the relevant topic from other point of view. Recently Fan and
Guo[5] , Zhu and Fan[6] , Li and Fan[7] observed the analytic methods and solutions
of cracks in three-dimensional icosahedral quasicrystals, Fan and Fan[8,9] , Li and
Fan[15] found the plastic analytic solutions of cracks, Zhu and Fan[10,11] carried out
the numerical analysis on dynamic crack of quasicrystals, respectively. And the
measurement of fracture toughness of the material is also reported[12] . The quite
fruitful information mentioned above suggests the need to give a summary on the
fracture study of quasicrystals. This chapter undertakes such a task and may put
forward some pointers for the fracture mechanics of quasicrystals.
15.1 Linear fracture theory of quasicrystals

The descriptions on crack solutions in different chapters reveal some common nature
of stress field and displacement field around crack tip. For all exact solutions of
cracks of one-, two- and three-dimensional quasicrystals in linear elasticity regime
the resulting expressions of stress and displacement components hold with respect
to arbitrary variables x and y. In addition we found quite a lot of solutions of cracks
of dynamic state and nonlinear behaviour of quasicrystals, some common feature
are also exhibited. At first we analyze the linear elastic solutions. As indicted in
Section 8.1, within the framework of linear elasticity and infinite-sharp crack model,
near crack tip, i.e.,
r1 /a 1, (15.1-1)

318 Chapter 15 Fracture theory of quasicrystals
−1/2
stresses for phonons as well as for phasons appear singular in order of r1 (r1 →0),
and other terms can be ignored when compared to this term. Although stress sin-
gularity is implausible, it is the result of idealized mathematical model. Quite
a few researchers indicated its severe weakness in theory and the paradox of its
methodology[1,2] . However, prior to the actual establishment of more reasonable
fracture theory, we still continue using this theory in expectable near future.
If temporarily accepting this theory, we focus on the field variables near crack
tip. If in (8.3-14) only keeping the term in order of (r1 /a)−1/2 , we have
⎧
⎪
⎪
⎪ KI 1 1 3
⎪ xx
⎪ σ = √ cos θ 1 1 − sin θ 1 sin θ 1 ,
⎪
⎪ 2πr1 2 2 2
⎪
⎪
⎪
⎪
⎪
⎪ KI 1 1 3
⎪
⎪ σ = √ cos θ 1 + sin θ sin θ 1 ,
⎪
⎪
yy
2πr1 2
1
2
1
2
⎪
⎪
⎪
⎪
⎪
⎪
⎪
⎪ √
KI 1 3
⎪
⎪ σ xy = σ yx = cos θ1 cos θ1 ,
⎪
⎪ 2πr 1 2 2
⎪
⎪
⎨
d21 K 3 3 5
Hxx = − √ I sin θ1 2 sin θ1 + sin θ1 cos θ1 , (15.1-2)
⎪
⎪ 2πr1 2 2 2
⎪
⎪
⎪
⎪
⎪
⎪

d21 KI 3 5
⎪ Hyy = √
⎪ sin2 θ1 cos θ1 ,
⎪
⎪
⎪
⎪ 2πr 1 2 2
⎪
⎪
⎪
⎪
⎪
⎪ d21 KI 3 5
⎪ xy
⎪ H = − √ sin2 θ1 sin θ1 ,
⎪
⎪ 2πr 1 2 2
⎪
⎪
⎪
⎪
⎪
⎪ d21 KI 3 3 5
⎪
⎩ yx
H = √ sin θ 1 2 cos θ1 − sin θ 1 sin θ 1 ,
2πr1 2 2 2
where d21 = R(K1 − K2 )/4(M K1 − R2 ) given in Section 8.3 and

||
# √
KI = lim 2π(x − a)σyy (x, 0) = π ap, (15.1-3)
x→a+
in which ⎧
⎨ x
p √ − 1 , |x| > a,
σyy (x, 0) =
⎩ x2 − a2 (15.1-4)
−p, |x| < a,
It is one of results of Sections 8.3 and 8.4, and (15.1-3) represents a physical param-
eter describing fracture behaviour of quasicrystals under Mode I loading condition.
The physical meaning of the generalized surface tractions hi = Hij nj is clear,
but have not been measured so far, we do not considered hi at the physical boundary
||
(simply assume zero). Therefore, we only obtain the KI , but the stress intensity
factor for phason field still exits if we do not assume the generalized tractions hi to be
||
zero. Can we use the stress intensity factor KI in the parallel space (physical space)
15.1 Linear fracture theory of quasicrystals 319
as the physical parameter to control the crack stability (instability) in quasicrystals?

This only depends upon tests.
||
It can be found that KI given by (15.1-3) is not directly related to the material
constants of quasicrystals in the case for any self-equilibrium applied stress state
(this is similar to that for conventional fracture mechanics for structural materials).
Nevertheless, it does not mean that it cannot be used as the physical parameter to
govern the crack stability/instability in quasicrystals. Further study is needed in
this topic.
But the displacement field concerning crack is strongly related with material
constants (this is also similar to that of conventional fracture mechanics for structural
materials), we must distinguish the results for different quasicrystal systems.
For point groups 5m and 10mm it can be found from (8.3-17) that
⎧ ⎧
⎪
⎪ ⎪
⎨ 0, |x| > a,
⎪
⎪
⎨ uy (x, 0) = p K1 1 √
⎪
⎩ + a2 − x2 , |x| < a, (15.1-5)
⎪
⎪ 2 M K1 − R 2 L + M
⎪
⎪
⎩
wy (x, 0) = 0, |x| < ∞.
The strain energy of the system due to the existence of crack is

a
WI = 2 (σyy (x, 0) ⊕ Hyy (x, 0)(uy (x, 0) ⊕ wy (x, 0))dx
0
a
=2 σyy (x, 0)uy (x, 0)dx
0

πa2 p2 1 K1
= + , (15.1-6)
4 L+M M K1 − R2
which is called crack strain energy with the suffix “I” to indicate the Mode I crack.
It can be found from solution in Sections 8.3 and 8.4 that under the assumption
of generalized surface traction hi = Hij nj being free, the crack strain energy is still
relevant to both the elastic constant K1 of the phason field and the phonon-phason
coupling elasticity constant R apart from relevant to the phonon elastic constants
L = C12 , M = (C11 − C12 )/2.
In Section 8.3, we define the strain energy release rate (crack growth force) for
point groups 5m and 10mm,

1 ∂W1 πap2 1 K1
GI = = +
2 ∂a 4 L+M M K1 − R 2

1 1 K1
= + (KI )2 , (15.1-7)
4 L+M M K1 − R
for point groups 5, 5 and 10, 10 quasicrystals

L(K1 + K2 ) + 2(R12 + R22 ) 2
GI = (KI ) , (15.1-8)
8(L + M )c
c = M (K1 + K2 ) − 2(R12 + R22 )

and for icosahedral quasicrystals

1 1 c7
GI = + (KI )2 , (15.1-9)
2 μ + λ c3
R(2K2 − K1 )(μK1 + μK2 − 3R2 )

c1 =
2(μK1 − 2R2 )
(μK2 − R2 )2
c3 = μ(K1 − K2 ) − R2 − ,
μK1 − 2R2
c3 K1 + 2c1 R
c7 = ,
μK1 − 2R2
which also indicate that the strain energy release rate depends on not only the
phonon elastic constants, but also the phason elastic constants and phonon-phason
coupling constants.
In the above relations, for point groups 5m and 10mm, due to L + M > 0,
M K1 − R2 > 0, M + L > 0 and crack energy WI and crack energy release rate
GI are all positive, meaningful in physical sense. For point groups 5, 5 and 10, 10,
icosahedral and other quasicrystals too.
Considering the obvious physical meaning of GI , we recommend
GI = GIC (15.1-10)
as the crack initiation criterion, where GIC is the critical value, a material constant
determined experimentally. With the availability of explicit expression GI , the mea-
surement of GIC is convenient, to be discussed in the next section. The above results
have been documented in Chapters 8, 9 and the relevant references.
With these common features of cracks in quasicrystalline materials, the funda-
mental of fracture theory for the material can be set up.
15.2 Measurement of GIC

Meng et al[12] measured the fracture toughness for Al65 Cu20 Co15 decagonal qua-
sicrystal by using nonstandard specimen, because the expressions on the stress in-
tensity factors and the energy release rate were not available then.
During the characterization of mechanical properties of quasicrystals, similar to
conventional structural materials, standard samples are expected to use, such as
15.2 Measurement of GIC 321
cracked samples. Here we recommend three-point bending specimen shown by Fig.

14.8-1 of Appendix of Chapter 14 and compact tension specimen shown by Fig. 15.2-
1 for determining GIC . The corresponding GI expressions are obtained by extending
the formula (15.1-7) and others.
Fig. 15.2-1 Compact tension specimen for measuring fracture toughness of

quasicrystalline material
15.2.1 Characterization of GI and GIC of three-point bending quasicrys-

tal samples
||
Due to KI independent of material constants, according to fracture mechanics, the
stress intensity factor corresponding to the three-point bending specimen as shown
in Fig.14.8-1 of Appendix of Chapter 14 is
a 3/2 a 5/2
PS a 1/2
KI = 29 − 4.6 + 21.8
BW 3/2 W W W
a 7/2 a 9/2 (15.2-1)
−37.6 + 38.7 .
W W
Therefore extension of (15.1-17) leads to

a 3/2
1 1 K1 PS a 1/2
GI = + 29 − 4.6
4 L+M M K1 − R2 BW 3/2 W W
a 5/2 a 7/2 a 9/2 2 (15.2-2)
+21.8 − 37.6 + 38.7
W W W
for point groups 5m and 10mm, where S is the sample span, B the sample thickness,
W the sample width, a the crack length plus the size of the machined notch, and P
is the external force (per unit length). Finally, the GIC value can be determined by
measuring the critical external force PC . For other quasicrystal systems there are
similar results.
15.2.2 Characterization of GI and GIC of compact tension quasicrystal

sample
It can be found in fracture mechanics that the stress intensity factor of the compact
tensile sample as shown in Fig. 15.2-1 is
a 1/2 a 3/2
PS
KI = 3/2
29.6 − 185.5
BW W W
a 5/2 a 7/2 a 9/2
+ 655.7 − 1017.0 + 638.9 . (15.2-3)
W W W
Therefore for point groups 5m and 10mm quasicrystals.

a 1/2
1 1 K1 P
GI = + 29.6
4 L+M M K1 − R 2 BW 3/2 W
a 3/2 a 5/2 a 7/2 a 9/2 2
−185.5 + 655.7 − 1017.0 + 638.9 ,
W W W W
(15.2-4)
where B, W , a, and P have the same meanings above. The GIC value can be deter-
mined by measuring the critical external force PC . For other quasicrystal systems
there are similar results.
15.3 Nonlinear fracture mechanics

In the regime for nonlinear deformation of quasicrystals, the stress intensity factor
and energy release rate cannot be used as a fracture parameter, and we must carry
out elasto-plastic analysis and the exact solutions of the crack problems have not
been obtained except a few of special cases. Fortunately this difficult topic has been
discussed in Chapter 14 in detail, now it is needed listing some key points only.
Instead stress intensity factor and energy release rate the crack tip opening dis-
placement or Eshelby integral may be a parameter characterizing the mechanical
behaviour of crack tip under nonlinear deformation of quasicrystals. These quanti-
ties are strongly related to material constants, so the discussion must be done for
distinguishing quasicrystal systems.
15.4 Dynamic fracture 323
For one-dimensional hexagonal quasicrystals we have obtained the crack tip slid-
ing displacement for Mode III crack as

4K2 τc a πτ1
δIII = ln sec (15.3-1)
π(C44 K2 − R32 ) 2τc
and for two-dimensional quasicrystals with point groups 5m and 10mm the crack tip
opening displacement for Mode I crack is

2σc a 1 K1 πp
δI = + ln sec . (15.3-2)
π L+M M K1 − R2 2σc
For other quasicrystal systems the results are similar.
The plastic zone size around the crack tip is

πτ1
d = a sec −1
2τc
for one-dimensional hexagonal quasicrystals and

πp
d = a sec −1
2σc
for two-dimensional quasicrystals with point groups 5m and 10mm.
And we have fracture criterion for mode I crack
δI = δIC . (15.3-3)
For Mode II and Mode III crack there are similar criterion, which have been discussed
in Chapter 14.
As pointed out in Chapter 14 the Eshelby integral can also be a fracture param-
eter, and based on which one can set up a fracture criterion, the full discussion can
be found there.
The experimental measurement of nonlinear fracture toughness of quasicrystals
has been introduced in Section 14.8 of Chapter 14, it does not mention any more.
15.4 Dynamic fracture

As we have known from Chapter 10, the study on dynamics for quasicrystals presents
a difficult situation, so the study for dynamic fracture.
Nevertheless the Chapter 10 provides some beneficial data for us.
By taking the so-called elasto-/hydro-dynamic equation system for quasicrystals,
the dynamic crack initiation problem can be solved by finite difference method. In
linear case the crack dynamic initiation can be described by dynamic stress intensity
factor, for some samples the results are listed in Chapter 10, which are sensitive func-
tions of loading type, loading rate and sample geometry including crack geometry.
One of results can be seen from the Fig. 15.4-1.
Fig. 15.4-1 Normalized dynamic stress intensity factor of central stationary crack
specimen under Heaviside impact loading (for icosahedral Al-Pd-Mn quasicrystals)
With the results we can propose the fracture criterion for dynamic crack initiation
KI (t) = KId (σ̇), (15.4-1)
in which KI (t) is the dynamic stress intensity factor evaluated by different ap-
proaches, KId (σ̇) represents the dynamic fracture toughness for initiation of crack
growth of the material and is measured by test, a material constant, but is function
of loading rate σ̇. While for fast crack propagation/crack arrest problems we have
results, e.g. shown in Fig. 15.4-2 for the central crack specimen there is fracture
criterion such as
Fig. 15.4-2 Normalized dynamic stress intensity factor of propagating crack with
constant crack speed of central crack specimen (for icosahedral Al-Pd-Mn quasicrystals)
15.5 Measurement of fracture toughness and relevant mechanical parameters of ... 325
KI (t) KID (V ), (15.4-2)

where KI (t) is also the dynamic stress intensity factor, a computational quantity,
while KID (V ) denotes the fracture toughness for fast propagating crack, which must
be measured by test, a material constant, but is the function of crack speed V =
da/dt. In (15.4-2), the equality sign represents crack propagation condition, and the
inequality sign marks the crack arrest condition.
15.5 Measurement of fracture toughness and relevant me-

chanical parameters of quasicrystalline material
Ref. [12] reported the measurement of fracture toughness of two-dimensional decago-
nal quasicrystal Al65 Cu20 Co15 as well as three-dimensional icosahedral Al-Li-Cu
quasicrystal, the authors used the indentation approach.
15.5.1 Fracture toughness

The material is locally pressed, the crack around the indentation will appear as the
compressible stress reaches a certain value, this describes the ability of fracture of
the material along the direction of the compressible stress. When the crack length
2a is 2.5 times greater than the length 2c of diagonal of the indentation, then the
fracture toughness can be evaluated by
+
√ c
KIC = 0.203HV c 3 , (15.5-1)
a
in which HV denotes hardness of the material.
Their results of measurement for decagonal Al-Ni-Co quasicrystal is
√
KIC = 1.0 ∼ 1.2MPa m (15.5-2)
with HV = 11.0 ∼ 11.5GPa, and for icosahedral Al-Li-Cu quasicrystal is

√
KIC = 0.94MPa m, (15.5-3)
in which HV = 4.10GPa.
The values of fracture toughness for general alloys for black metals measured by
Ma et al[13] are much greater than the above data, those for aluminum alloys and
√
other colour metals are also, e.g., for aluminum 33MPa m refer to Fan[14] . So one
finds that quasicrystals are very brittle.
The author of the book minds the indirect measurement for fracture tough-
ness of quasicrystals through indentation perhaps is not so exact, because formula
(15.5-1) is empirical, the exact measurement should use the stress intensity factor
formula. Due to the high brittleness of the material, maybe it is easy by taking the
indentation approach.
15.5.2 Tension strength

The tensile strength σc is measured through the formula
2
σc = 0.187Pa/a , (15.5-4)
they obtained
σc = 450MPa (15.5-5)
before annealing and
σc = 550MPa (15.5-6)
after annealing for decagonal Al-Ni-Co quasicrystal.
The Fig. 15.5-1 shows the SEM morphology of grain interior containing large
hole, Fig. 15.5-2 shows diagram of indentation crack under applied load 100g before
annealing and after annealing, Fig. 15.5-3 shows the SEM fractograph and fracture
feature for decagonal Al65 Cu20 Co15 quasicrystal.
Fig. 15.5-1 SEM morphology of grain interior hole
Fig. 15.5-2 Morphology of cracks near impression under 100g load

References 327
Fig. 15.5-3 SEM fractograph
References
[1] Li X F, The defect problems and their analytic solutions in the theory of elasticity
of quasicrystals, Dissertation, Beijing Institute of Technology, Beijing, 1999 (in Chi-
nese); Li X F, Fan T Y, Sun Y F. A decagonal quasicrystal with a Griffith crack.
Phil Mag A, 1999, 79(8): 1943–1952
[2] Fan T Y. Mathematical theory of elasticity and defects of quasicrystals. Advances
of Mechanics, 2000, 30(2): 161–174 (in Chinese)
[3] Fan T Y, Mai Y W. Theory of elasticity, fracture mechanics and some relevant
thermal properties of quasicrystalline materials. Appl Mech Rev, 2004, 57(5): 235–
244
[4] Rudhart C, Gumbsch P, Trebin H R. Crack propagation in quasicrystals. Quasicrys-
tals, Trebin H R. Berlin: Wiely Press, 2003
[5] Fan T Y, Guo L H. The final governing equation of plane elasticity of icosahedral
quasicrystals. Phys Lett A, 2005, 341(5): 235–239
[6] Zhu A Y, Fan T Y. Elastic analysis of Mode II Griffith crack in an icosahedral
quasicrystal. Chinese Physics, 2007, 16(4): 1111–1118
[7] Li L H, Fan T Y. Complex variable method for plane elasticity of icosahedral qua-
sicrystals and elliptic notch problem. Science in China, G, 2008, 51(6): 1–8
[9] Fan T Y, Fan L. Relation between generalized Eshelby integral and generalized BCS
model and generalized DB model. Chin. Phys. B, in press, 2010
[10] Zhu A Y, Fan T Y. Dynamic crack propagation of decagonal Al-Ni-Co decagonal
quasicrystals. J Phys: Condens Matter, 2008, 20(29): 295217
[11] Wang X F, Fan T Y and Zhu A Y. Dynamic behaviour or the icosahedral Al-Pd-Mn
quasicrystal with a Griffith crack, Chin. Phys. B, 2009, 18(2): 709–714. or Zhu A
Y and Fan T Y, Fast crack propagctim in three-dimensional icosahedral Al-Pd-Mn

quasicrystals, J. Phys. A, 2008, submitted
[12] Meng X M, Tong B Y, Wu Y K. Mechanical behaviour of Al65 Cu20 Co15 quasicrystal.
Acta Metallurgy Sinica, 1994, 30(1): 61–64 (in Chinese)
[13] Ma D L. Handbook for Parameters of Fracture Mechanics Behaviour of General Black
Metals. Beijing: Industrial Press, 1994 (in Chinese)
[14] Fan T Y. Foundation of Fracture Theory, Beijing: Science Press, 2003 (in Chinese)
[15] Li W, Fan T Y. Study on elastic analysis of crack problem decagonal Al-Ni-Co
quasicrystals of point group 10, 10. J Phys: Int. Mod. Phys. Lett. B, 2009, 23(16),
1989–1999.
Chapter 16
Remarkable Conclusion
The text of discussion on elasticity of quasicrystals and some applications is ended till
the Chapter 15. This work in some extent is a development of crystal (or classical)
elasticity. The crystal (or classical) elasticity is a branch of theoretical physics, see
e.g. Sommerfeld[1] and Landau and Lifshitz[2] . Since the time of Hooke, Bernoulli,
Euler, Navie, Cauchy, Saint-Venant, among others, the classical theory of elasticity
has been mature, representing one of the most elegant branches of continuum me-
chanics, which challenges the applied mathematics and becomes one of sources of
many powerful methods of mathematical physics. The study of elasticity of qua-
sicrystals not only enlarges the scope of the classical elasticity, but also strength-
ens the link between elasticity and other branches of physics, e.g. the elemen-
tary excitation theory of condensed matter[3] , symmetry and symmetry-breaking[4] ,
hydrodynamics[5] etc. It strengthens the link between elasticity and applied math-
ematics as well, for instance, the partial differential equations, complex variable
function method, group theory, discrete geometry, numerical analysis etc, those
have been partly concerned in the text. In particular, the mathematical solutions of
elasticity of quasicrystals promote the development of partial differential equations
of higher order, applied complex analysis, the Fourier method, weak solution the-
ory and some numerical methods in continuum mechanics and applied mathematics,
which have been touched upon in the previous chapters. The study of the elastic-
ity also promotes the development of defect theories (dislocation theory[6] , fracture
mechanics[7] etc), plasticity[8−10] and elasto-/hydro-dynamics[11,12] . It is believed
that extensive research on the new area will bring forth useful results of theoretical
interest and practical importance.
Due to the space limitation, some relevant subjects could not be included in
the volume, but they present important meaning, for examples, the effect of pha-
son strain field to phase transition of crystal-quasicrystal[13−22] , specific heat of
quasicrystals[23−27] based on the classical Debye[28] theory, evaluating indirectly elas-
tic modulus of phason and phonon-phason coupling[29,30] . The readers may refer to
the following references on these interesting subjects.

330 Please number this as Chapter 16
References
[1] Sommerfeld A. Vorlesungen ueber theoretische Physik. Vol. V: Elastische Theorie.
Wiesbaden: Diederich Verlag, 1952
[2] Landau L D, Lifshitz E M. Theoretical Physics. Vol. VII: Theory of Elasticity.
Oxford: Pergamon Press, 1986
[3] Bak P. Symmetry, stability and elastic properties of icosahedral incommensurate
crystals. Phys Rev B, 1985, 32(9): 5764–5772
[4] Horn P M, Malzfeldt W, DiVincenzo D P et al. Systematics of disorder in quasiperi-
odic Material. Phys Rev Lett, 1986, 57(12): 1444–1447
[5] Lubensky T C. Introduction to Quasicrystals. Jaric M V. Boston: Academic Press,
1988.
[7] Fan T Y, Mai Y W. Elasticity theory, fracture mechanic and some thermal properties
of quasicrystalline materials. Appl Mech Rev, 2004, 57(5), 325–344
[8] Feuerbacher M, Urban K. Platic behaviour of quasicrystalline materials. in: Qua-
sicrystals. Trebin H R, Wiely Press, Berlin, 2003
[9] Fan T Y, Trebin H R, Messerschmidt U et al. Plastic flow coupled with a crack
in some one- and two-dimensional quasicrystals. J Phys: Condens Matter, 2004,
16(47): 5229–5240
[11] Fan T Y, Wang X F, Li W et al. Elasto-hydrodynamics of quasicrystals. Phil Mag,
2009, 89(6), 501–512
[12] Wang X F, Fan T Y. Dynamic behaviour of the icosahedral AL-Pd-Mn quasicrystals
with a Griffith crack. Chin Phys, 2009, 18(2), 709–714
[13] Audier M, Guyot P. Al4 Mn quasicrystal atomic structure, diffraction data and Pen-
rose tiling. Phil Mag B, 1986, 53(1): L43–L51
[14] Jaric M V, Gratias D. Extended Icosahedral Structures. Boston: Academic, 1989
[15] Jaric M V, Lundquist S. Proc Adiatico Research Conf Qusicrystals, Singapore: World
Scientific, 1990
[16] Elser V, Henley C L. Crystal and quasicrystal structures in Al-Mn-Si alloys. Phys
Rev Lett, 1985, 55(26): 2883–2886
[17] Henley C L, Elser V. Quasicrystal structure of (Al, Zn)49 Mg32 . Phil Mag B, 1986,
53(3): L59–L61
[18] Lubensky T C, Ramaswamy S, Toner J. Hydrodynamics of icosahedral quasicrystals.
Phys Rev B, 1985, 32(11): 7444–7452
[19] Lubensky T C, Socolar J E S, Steinhardt P J et al. Distortion and peak broadening
in quasicrystal diffraction patterns. Phys Rev Lett, 1986, 57(12): 1440–1443
References 331
[20] Li F H. in Crystal-Quasicrystal Transitions. Jacaman M J, Torres M, Amsterdam:

Elsevier Sci Publ. 1993
[21] Li F H, Teng C M, Huang Z R et al. In between crystalline and quasicrystalline
states. Phil Mag Lett, 1988, 57(1): 113–118
[22] Xie T Y, Fan L, et al. Study on interface of quasicrystal-crystal. Chin. Phys. B,
submitted 2009
[23] Fan T Y. A study of specific heat of one-dimensional hexagonal quasicrystals. J
Phys: Condens, Matter, 1999, 11(45): L513–L517
[24] Li C L, Liu Y Y. Phason-strain influence on low-temperature specific heat of the
[25] Li C L, Liu Y Y. Lower-temperature lattice excitation of icosahedral AL-MN-pd
quasicrystals. Phys Rev B, 2001, 63(6): 064203
dodecagonal quasicrystals. Euro Phys J B, 2003, 31(2): 25–27
[27] Wang J, Fan T Y. An analytic study on specific heat of icosahedral Al-Pd-Mn qua-
sicrystals. Modern Phys Lett B, 2008, 22(17): 1651–1659
[28] Debye P. Eigentuemlichkeit der spezifischen Waermen bei tiefen Temperaturen. Arch
de Genéve, 1912, 33(7): 256–258
[29] Edagawa K, Giso Y. Experimental evaluation of phonon-phason coupling in icosahe-
dral quasicrystals. Phil. Mag, 2007, 87(1): 77–95
[30] Edagawa K. Phonon-phason coupling in decagonal quasicrystals. Phil Mag, 2007,
87(7): 2789–2798
Major Appendix:
On Some mathematical materials
Mathematical solutions for individual boundary value problems were given in the
previous chapters, in which crack problems are more difficult than dislocation prob-
lems. Due to the complexity of boundary conditions for the formers, we must use
some special techniques, in which the complex function method collaborating con-
formal mapping and Fourier analysis collaborating dual integral equations are the
most effective procedures. The use of the procedures were exhibited in the previ-
ous chapters. Though some appendices in relevant chapters were given in the text
for discussing some special problems, we list some necessary additional materials
on complex variable functions and dual integral equations in separate two parts in
the present appendix. It is interesting in particular, between these two parts there
are some close inherent connections. Perhaps these materials are not necessary for
readers majoring applied mathematics, but which may be helpful for young scholars
and post graduate students who are not majoring applied mathematics.
Appendix I Outline of complex variable functions and some

additional calculations
It is enlightened that Muskhelishvili[1] gave extensive description in detail on complex
variable functions in due presentation of elasticity in his classical monograph, this is
very beneficial to readers. However there is no the possibility for the present book.
We list here a few of points only of the function theory which were frequently cited
in the text. These can be referred for readers who are advised to read books of I. I.
Privalov[2] and M. A. Lavrentjev and B. V. Schabat[3] for the further details. Other
knowledge has been provided in due succession of the text of Chapters 7∼9 and 11.
The present contents can also be seen as a supplement in reading the material given
in Chapters 7∼9 and 11 if it is needed. The importance of complex variable functions
is not only in deriving the solutions by the complex potential formulation but also
in dealing with the solutions by integral transforms and dual integral equations to
be discussed in the next section of the appendix.
334 Major Appendix: On Some mathematical materials
A.I.1 Complex functions, analytic functions

√
Usually z = x + iy is denoted as a complex variable in which −1 = i, or z = reiθ ,
# y
and r = x2 + y 2 , called the modulus of the complex number, θ = arctan , the
x
argument of z. Assume f (z) to be a function of one complex variable, or complex
function in abbreviation, denote
f (z) = P (x, y) + iQ(x, y), (A.I-1)
in which both P (x, y) and Q(x, y) are functions with real variables, and called the
real and imaginary parts, respectively, and marked by
P (x, y) = Ref (z), Q(x, y) = Imf (z).
There is a sort of complex functions called analytic functions (or regular func-
tions, and single-valued analytic functions are called holomorphic functions) which
have important applications in many branches of mathematics, physics and engi-
neering. The concepts related with this are discussed as follows.
The complex function f (z) is analytic in a given region, this means that it can
be expanded in the neighbourhood of any point z0 of the region into a non-negative
integer power series (i.e., the Taylor series) of the form
∞

f (z) = an (z − z0 )n , (A.I-2)
n=0
in which an is a constant (in general, a complex number). The concept will be used
frequently in later calculation.
Another definition of an analytic function is that if the complex function f (z)
given in the region, whose real part P (x, y) and imaginary part Q(x, y) are single
valued, have continuous partial derivatives of the first order, and satisfy Cauchy-
Riemann conditions such as
∂P ∂Q ∂P ∂Q
= , =− (A.I-3)
∂x ∂y ∂y ∂x
in the region.
This kind of function, P and Q are named mutually conjugate harmonic ones.
From (A.I-3) it follows that
2 2
∂ ∂2 ∂ ∂2
∇2 P = + P = 0, ∇2
Q = + Q = 0.
∂x2 ∂y2 ∂x2 ∂y 2
This concept will also be often used in the following.

An analytic function can also be defined in integral form. Assuming f (z) is an
analytic function f (z) in a certain region D at z-plane, and Γ is any simple smooth
Appendix I Outline of complex variable functions and some additional calculations 335
closed curve (sometime called simple curve for simplicity), we can obtain that f (z)
is analytic in the region if
f (z)dz = 0. (A.I-4)
Γ
The result is known as the Cauchy’s integral theorem (or simply called the Cauchy’s
theorem) which is frequently used in some chapters in the text and in subsequent
description here.
The theory of complex functions proves that the above definitions are mutually
equivalent.
A.I.2 Cauchy’s formula
An important result of the Cauchy’s theorem is so-called the Cauchy’s formula, i.e.,
if f (z) analytic in a single-connected region D + bounded by a closed curve Γ and
continuous in D+ + Γ (Fig. A.I-1), then

1 f (t)
dt = f (z), (A.I-5)
2πi Γ t−z
in which z is an arbitrary point in D+ .
Fig. A.I-1 A finite region D+
Proof Taking z as the centre, ρ as the radius, make a small circle in D+ .

According to Cauchy’s theorem (A.I-4),

f (t) f (t)
dt = dt. (A.I-6)
Γ t−z γ t−z
As f (z) is analytic in D + and continuous in D+ + Γ , there is a small number ε > 0,

for any point t and γ, if ρ is sufficient small, such as
|f (t) − f (z)| < ε

and note that |t − z| = ρ , hence

f (t) f (z)
lim dt = dt. (A.I-7)
ε→0 Γ t−z γ t−z
Just as mentioned in the beginning that f (z) is analytic in D + , the value of the
integral
f (z)
dt
γ −z
t
will not be changed when ρ is reducing. Therefore the limit mark in the left-handside
of (A.I-7) can be removed. In addition
2π iθ
f (z) dt ρe
dt = f (z) = f (z) dθ = 2πif (z).
γ t − z γ t − z 0 ρeiθ
Based on (A.I-6) and this result, formula (A.I-5) is proved.

In formula (A.I-5), if z is taken its values in region D− consisting of the points
lying outside Γ (see Fig. A.I-1), then

1 f (t)
dt = 0. (A.I-8)
2πi Γ t − z
In fact, this is a direct consequence of the Cauchy’s theorem, because in the case
the integrand f (ζ)/(ζ − z) as function of ζ is analytic in region D + where ζ denotes
the point in the region D+ .
Suppose all conditions are the same as those for (A.I-5), then

1 f (t)
dt = f (0). (A.I-9)
2πi Γ t−z
Proof For simplicity here the proof is given for the case Γ as a circle. Being
analytic in the region D + , f (z) may be expanded as non-negative integer power
series, in which, taking z0 = 0, such that
1
f (z) = a0 + a1 z + a2 z 2 + · · · = f (0) + f (0)z +
f (0)z 2 + · · · .
2!

1
The function f (z) in formula (A.I-9) is the value of f¯ at the circle Γ , here
z

1 1 1 1
f¯ = f (0) + f (0) + f (0) 2 + · · ·
z z 2! z
is an analytic function in D− . From the Cauchy’s formula,

1 dt 1, k = 0,
=
2πi Γ tk (t − z) 0, k > 0,
such that (A.I-9) is proved.

In contrast to above, current function f (z) is analytic in D − (including z = ∞),
then
1 f (z) −f (z) + f (∞), z ∈ D− ,

dt = (A.I-10)
2πi Γ t − z f (∞), z ∈ D+ .
The proof of this formula can be offered in the similar manner to that for (A.I-5),
but the following points must be noted:
(i) The analytic function f (z) in D − (includingz = ∞) may be expanded as the
following series:
1 1
f (z) = c0 + c1 + c2 2 + · · · ;
z z

−
1 c0 0, z ∈ D ,
(ii) dt =
2πi Γ t − z c0 , z ∈ D + .
where c0 = f (∞) = 0.
If all conditions are same as that for formula (A.I-10), there exists

1 f (t)
dt = 0. (A.I-11)
2πi Γ t−z
A.I.3 Poles
Suppose a finite point in z-plane (i.e., z is not a point at infinity), and in the
neighbourhood of the point, the function presents the form as follows:
f (z) = G(z) + f0 (z), (A.I-12)
in which f0 (z) is an analytic function in the neighbourhood of point a and
A0 A1 Am
G(z) = + + ··· + , (A.I-13)
z − a (z − a)2 (z − a)m
where A1 , A2 , · · · , Am are constants, such that f (z) is called having a pole with
order m, z = a is the pole.
If a is a point at infinity, f0 (z) in (A.I-12) is regular at point at infinity (i.e.,
f (t) = c0 + c1 z −1 + c2 z −2 + · · · ), while at z = ∞,
G(z) = A0 + A1 z + · · · + Am z m , (A.I-14)
then we say that f (z) has a pole of order m at z = ∞.
A.I.4 Residue theorem

If the function f (z) has pole a with order m, its integral may be evaluated simply
by computing residue.
What is the meaning of the residue? Suppose f (z) analytic in the neighbourhood
of point z = a, but except z = a, and infinite at z = a. In this case the point z = a
is named isolated singular point. The residue of the function f (z) at point z = a is
the value of the integral

1
f (z)dz,
2πi Γ
in which Γ represents any closed contour enclosing point z = a. For a residue we
will use the resignation as Res f (a).
If z = a is a m-order pole of f (z), its residue may be evaluated from the following
formula and
1 dm−1
Resf (a) = lim m−1 {(z − a)m f (z)} , (A.I-15)
(m − 1)! z→a dz
obviously the integral is

f (z)dz = 2πiResf (a).
Γ
So evaluation of integrals may be reduced to the calculation of derivatives, it is

greatly simplified. Especially if z = a is a first order pole, then
Resf (a) = lim (z − a)f (z), (A.I-16)

z→a
the calculating is much simpler.

What follows the residue theorem is introduced as: let the function f (z) be ana-
lytic in region D and continuous in D+Γ except at finite isolated poles a1 , a2 , · · · , an ,
then

n
f (z)dz = 2πi Resf (ak ), (A.I-17)
Γ k=1
where Γ represents the boundary of region D.

Example Calculate the integral
∞
1 1
e−iωt dω = I (A.I-18)
2π −∞ −mω 2 + k
in terms of the residue theorem, where m and k are positive constants.

Though the integral is a real integral, it is difficult to evaluate because the in-
tegration limit is infinite and there are two singular points at the integration path,
but is easily completed by using the residue theorem. At first we extend the real
variable ω to a complex one, i.e., put ω = ω1 + iω2 , and ω1 , ω2 are real variables. At
the complex plane ω, a half-circle with origin (0, 0) and radius R → ∞ is taken as
an additional integral path referring to Fig. A.I-2. Along the real axis, the integrand
# #
of the integral has two poles (− k/m, 0) and ( k/m,0), the value of the integral
is equal to
∞
1 1 −iωt
e dω = I1 = lim + + + + + ,
2π −∞ −mω 2 + k R→∞,r→0 CR 1 2 3 C1 C2
(A.I-19)
Fig. A.I-2 Integration path at ω-plane
where the first integral in the right-handside of (A.I-19) is carried out on path of
the grand half-circle, the second and fourth ones are on the path along the real axis
# # # #
except intervals (−r − k/m, − k/m + r) and (−r + k/m, k/m + r), the fifth
#
and sixth ones are on two small half-circle arcs C1 and C2 with origins (− k/m, 0)
#
and ( k/m, 0) and radius r respectively. Because the integrand in the interior
enclosing by the integration path in (A.I-19) is analytic, according to the Cauchy’s
theorem (referring to formula (A.I-3))
I1 = 0. (A.I-20)
Based on the bebaviour of the integrand and the Jordan lemma, the first one in the
right-handside of (A.I-19) must be zero. So that

lim + + + + =0
R→∞,r→0 1 2 3 C1 C2
and
lim + + = I = − lim + .
R→∞,r→0 1 2 3 r→0 C1 C2
# #
At arc C1 : ω + k/m = reiθ1 , dω = ireiθ1 dθ1 , and at C2 : ω − k/m = reiθ2 , dω =
ireiθ2 dθ2 . Substituting these into the above integrals and after some simple calcula-
tions, we obtain
+
π k
I= # sin t (A.I-21)
m k/m m
This result will be used in the evaluation of some integrals in Chapter 9, which
can be seen the Subsection AI.10 of this appendix.
A.I.5 Analytic extension

A function f1 (z) analytic at region D1 , if one can construct another function f2 (z)
analytic at region D2 , D1 and D2 have common bounding Γ , furthermore,
f1 (z) = f2 (z), z ∈ Γ,
we can say that f1 (z) and f2 (z) are analytic extension to each other, we can also

f1 (z) as z ∈ D1
say that function F (z) =
f2 (z) as z ∈ D2
analytic at D = D1 + D2 is an analytic extension of f1 (z) as well as f2 (z).
A.I.6 Conformal mapping

In the text of Chapters 7∼9, and 11, by using one or several analyic functions which
are also named complex potentials, we have expressed the solutions of harmonic, bi-
harmonic, quadruple harmonic, sextuple harmonic equations, and quasi-biharmonic
and quasi-quadruple harmonic equations, which is called the complex representa-
tion. We can see that only the complex representation is the first step for solving
some boundary value problems. For some problems with complicated boundaries,
one must utilize the conformal mapping to transform the problem into the mapping
plane, in this plane the function equations concluded from the boundary conditions
are formulated at a unit circle, and the calculation can be put forward, in some
cases, exact analytic solutions are available.
The so-called conformal mapping is that the complex variable z = x + iy and
another one ζ = ξ + iη can be connected by
z = ω(ζ), (A.I-22)
in which ω(ζ) is a single-valued analytic function of ζ = ξ + iη in some region.

Except certain points the inversion of mapping (A.I-22) exists. If for a certain
region the mapping is single-valued, we say it is a single-valued conformal mapping.
In general, the mapping is single-valued, but the inversion ζ = ω −1 (z) is impossible
single-valued. It has the following properties:
(1) A angle at point z = z0 after the mapping becomes a angle at point ζ = ζ0 ,
but the both angles have the same value of the argument, and the rotation is either
in the same direction, this is the first kind of conformal mapping, or in counter
direction, which is the second kind of conformal mapping.
(2) If ω(ζ) is analytic and single-valued in region Ω and transforms the region
into region D, then the inversion ζ = ω −1 (z) is analytic and single-valued in region
D and maps D onto Ω .
(3) If D is a region and c is a simple closed curve in it, and its interior belongs to
D, and if ω −1 (z) is analytic, and maps c onto a closed curve γ at Ω region bilaterally
single-valued, then ω(ζ) is analytic and single-valued in the region and maps D onto
the interior of Ω .
In the text we mainly used the following two kinds of conformal mapping, i.e.,
(1) Rational function conformal mapping, e.g.
c
ω(ζ) = + a0 + a1 ζ + · · · + an ζ n (A.I-23)
ζ
or
1 1
ω(ζ) = Rζ + b0 + b1 + · · · + bn n , (A.I-24)
ζ ζ
in which, c, a0 , a1 , · · · , an , R, b0 , b1 , · · · , bn are constants. These mappings can be
used in studying infinite region with a crack at physical plane onto the interior of
unit circle at mapping plane. In the monograph of Muskhelishvili[3] , he postulated
that his method is only suitable for this kind of mapping functions. Fan[4] extended
it to transcendental mapping functions and achieved exact analytic solutions for
crack problems for complicated configuration.
(2) Transcendental functions, e.g.

H (1 + ζ)2
ω(ζ) = ln 1 + (A.I-25)
π (1 − ζ)2
and 5# πa 6
2W
ω(ζ) = arctan 1 − ζ 2 tan − a, (A.I-26)
π 2W
which can be used to transform a finite specimen with a crack onto the interior of
unit circle at mapping plane, where H, W and a represent sample sizes and crack
size.
A.I.7 Additional derivation of solution (8.2-19)

Formula (8.2-19) is obtained from the integral of (8.2-6) in which the conformal
mapping ω(ζ) is given by (8.2-16). The difficulty of the calculation lies in the
integral path is a part of circle rather than whole circle. Substitution (8.2-16) into
the right-handside of (8.2-6) yields
1 1
p ω(σ) p ω (σ)
F (ζ) = − dσ = − dσ,
2πi −1 (σ − ζ)2 2πi −1 σ−ζ
where
4Hα(1 − β) 1 − σ (1 + σ)2
ω (σ) = − .
π 1 + σ [(1 + σ) + α(1 − σ 2 )2 ][(1 + σ)2 + βα(1 − σ)2 ]
2
Substituting eiϕ by (1 + ix)/(1 − ix) yields that
1−σ 2i 4
= −2x, dσ = dx, (1 + σ)2 = ,
1+σ (1 − ix)2 (1 − ix)2
1 1 − ix (1 − ix)[1 − ζ − ix(1 + ζ)] 1 − ζ − x2 (1 + ζ) − 2ix
= = − .
σ−ζ 1 − ζ + ix(1 + ζ) (1 − ζ)2 + x2 (1 + ζ)2 (1 − ζ)2 + x2 (1 + ζ)2
So that
1
pHα(1 − β) ix[1 − ζ − x2 (1 + ζ) − 2ix]
F (ζ) = − dx
π2 −1 (1 − αx2 )(1− βαx2 )[(1 − ζ)2 + x2 (1 + ζ)2 ]
1
4pHα(1 − β) x2
=− dx
π2 0 (1 − αx2 )(1 − βαx2 )[(1 − ζ)2 + x2 (1 + ζ)2 ]
2pH α(1 − β)(1 − ζ 2 )
=
π [α(1 − ζ) + (1 + ζ)2 ][βα(1 − ζ)2 + (1 + ζ)2 ]
2 2

1+ζ 1+ζ
× arctan − arctan
1−ζ −1 − ζ
√ √ √ √
4pH α arctanh α βα arctanh α
− 2 − .
π α(1 − ζ)2 + (1 + ζ)2 γα(1 − ζ)2 + (1 + ζ)2
(A.I-27)
In the last step the evaluation is used of the Mathematica 3.0[19] .
By considering
√
1+ α √
A = ln √ = 2 arctanh α,
1− α
√
1 + βα √
M = ln √ = 2 arctanh βα,
1 − βα

1+ζ 1+ζ i i−ζ
arctan = − arctan = ln ,
−1 + ζ 1−ζ 2 1 − iζ
then (A.I-27) is just the formula (8.2-19).
A.I.8 Additional derivation of solution (11.3-53)

In example 3 of Section 11.3, the calculation is quite lengthy, here we provide some
details on the evaluation. As an example, we can show the derivation on function
Φ4 (ζ), which is
Φ4 (ζ) = d1 (X + iY ) ln ζ + Bω(ζ) + Φ4∗ (ζ),
in which the first two terms are known referring to the text, and the single-valued
analytic function Φ4∗ (ζ) satisfies the following boundary condition:
ω(σ)
Φ4∗ (σ) + Φ3∗ (σ) +

· Φ4∗ (σ) = f0 ,
ω (σ)
where

i ω(σ)
f0 = (Tx + iTy )ds − (d1 − d2 )(X + iY ) ln σ − · d1 (X − iY ) · σ
32c1 ω (σ)
−2Bω(σ) − (B − iC )ω(σ)
(A.I-28)
and (referring to Fig. 11.3-4)
Tx = −p cos(n, x), Ty = −p cos(n, y) at z1 M z2 ,

⎧
⎪
⎪
⎪
⎨ ipdz, z1 M z2 ,

(Tx + iTy )ds =
⎪
⎪
⎪
⎩ 0, z Nz ,

2 1
(A.I-29)

X + iY = (Tx + iTy )ds = ip(z1 − z2 ).
1 dσ
Multiplying both sides of the above equation by , and integrating around
2πi σ − ζ
the unit circle, we have

Φ4∗ (σ)

1 1 ω(σ) Φ4∗ (σ) 1 Φ3∗ (σ) 1 f0
dσ + dσ + dσ = dσ,
2πi γ σ − ζ 2πi γ σ − ζ
(A.I-30)
in which according to the Cauchy’s integral formula (referring to formula (A.I-5))
there is
1 Φ4∗ (σ)
dσ =Φ4∗ (ζ)
2πi γ σ − ζ
and in terms of analytic extension principle and the Cauchy’s theorem (referring to
Sections A.I-5 and A.I-1)

1 ω(σ) Φ4∗ (σ)
=0
2πi γ ω (σ) σ − ζ
and according to formula (A.I-9)

1 Φ3∗ (σ)
dσ = const.
2πi γ σ−ζ
So that (A.I-30) reduces to

1 f0
Φ4∗ (ζ) = dσ + const
2πi γ σ−ζ
And substituting (A.I-28) and (A.I-29) into the right-handside yields

1 f0 pR0 m dσ
σ2 pz2 σ2 dσ
dσ = +
σ+
2πi γ σ−ζ 2πic1 σ1 σ σ−ζ 2πi σ1 σ − ζ

p(z1 − z2 ) 1 ln σ
+ dσ
2πi 2πi γ σ − ζ

p(z1 − z2 ) 1 σ 2 + m dσ
− ,
2πi 2πi γ 1 − mσ 2 σ − ζ
in which

σ2 m dσ m σ2 m σ2 − ζ
σ+ = σ2 − σ1 − ln + ζ+ ln ,
σ1 σ σ−ζ ζ σ1 ζ σ1 − ζ
σ2
dσ σ1 − ζ
= ln .
σ1 σ−ζ σ2 − ζ
In accordance with the Cauchy’s theorem (referring to formula (A.I-4))

1 σ 2 + m dσ
= 0,
2πi γ 1 − mσ 2 σ − ζ
because the integrand is single-valued analytic function in the region outside the
unit circle γ. The remaining term is

1 ln σ
I(ζ) = dσ.
2πi γ σ − ζ
For calculating it we consider

dI 1 ln σ 1 1
= dσ = − ln σd
dζ 2πi γ (σ − ζ)2 2πi γ σ−ζ

ln σ σ=σ1 1 dσ 1 1 exp i(ϕ1 + 2π) 1
=− + =− ln −
σ − ζ σ=σ 2πi γ σ(σ − ζ) 2πi σ1 − ζ exp(iϕ1 ) ζ
1
2πi 1
= − .
σ1 − ζ ζ
So that
I(ζ) = ln(σ1 − ζ) − ln ζ + const.
Hence function Φ4∗ (ζ) is determined so the function Φ4 (ζ) in which the constant
term is omitted:

1 p mR0 σ2 σ2 − ζ
Φ4 (ζ) = · · − ln + z ln + z1 ln(σ1 − ζ) − z2 ln(σ2 − ζ)
32c1 2πi ζ σ1 σ1 − ζ
+ip(d1 − d2 )(z1 − z2 ) ln ζ,
which is just the first formula of (11.3-53), others can be similarly derived. In the
derivation the classical work of Muskhelishvili[1] is referred.
A.I.9 Detail of complex analysis of generalized cohesive force model for

plane elasticity of two-dimensional point groups 5m, 10mm and
10, 10 quasicrystals
The elasticity solution (11.3-53) based on complex analysis can be used to solve the
present problem.
The generalized Dugdale-Barenblatt model or generalized cohesive force model
for plane elasticity of the quasicrystals can be reduced to solve the equation
∇2 ∇2 ∇2 ∇2 G = 0 (A.I-31)
under boundary conditions

⎧ #
⎪
⎪ σyy = p, σxx = σxy = 0, Hxx = Hyy = Hxy = Hyx = 0 x2 + y 2 → ∞,
⎨
σyy = σxy = 0, Hyy = Hyx = 0 y = 0, |x| < a,
⎪
⎪
⎩
σyy = σc , σxy = 0, Hyy = Hyx = 0 y = 0, a < |x| < a + d,
(A.I-32)
which can be decomposed into two cases, among them one is
⎧ #
⎨ σxx = σxy = σyy = 0, Hxx = Hyy = Hxy = Hyx = 0 x2 + y 2 → ∞,
σ = σxy = 0, Hyy = Hyx = 0 y = 0, |x| < a,
⎩ yy
σyy = σc , σxy = 0, Hyy = Hyx = 0 a < |x| < a + d
(A.I-33)
and another

#
σyy = p, σxx = σxy = 0, Hxx = Hyy = Hxy = Hyx = 0 x2 + y 2 → ∞,
σyy = σxy = 0, Hyy = Hyx = 0 y = 0, |x| < a + d.
(A.I-34)
The solution of problem (A.I-31), (A.I-33) can be obtained from (11.3-53), i.e.,
if put m = 1, R0 = (a + d)/2 then the elliptic hole reduced to a Griffith crack with
half-length (a + d). In Fig. 11.3-3, let z1 = (a + d, +0), z2 = (a, +0), from (11.3-53),
we can obtain a solution, similarly put
z1 = (a + d, −0), z2 = (a, −0), z1 = (−a − d, +0), z2 = (−a, +0),
z1 = (−a − d, −0), z2 = (−a, −0)
respectively from (11.3-53), one can find other corresponding three solutions, by
superposing which one can obtain solution
⎧
⎪ (1) 1 σc (a + d)ϕ2 1 1 σc
⎪
⎪ Φ4 (ζ) = · · − ·
⎪
⎪ 32c 1 π ζ 32c 1 2πi
⎪
⎪
⎨ σ −ζ σ2 + ζ

(ζ − σ2 )(ζ + σ2 )

2
z ln + ln − a ln
⎪
⎪ σ2 − ζ σ2 + ζ (ζ + σ2 )(ζ − σ2 )
⎪
⎪
⎪
⎪
⎪ Φ (1) (ζ) = 1 · σc (a + d)ϕ2 · 2ζ − 1 · σc a ln (ζ − σ2 )(ζ + σ2 ) ,
⎩ 3
32c1 π ζ 2 − 1 32c1 2πi (ζ + σ2 )(ζ − σ2 )
(A.I-35)
iϕ
where σ = e represents the value of ζ at the unit circle in the mapping plane and
σ2 = eiϕ2 , a = (a + d) cos ϕ2 .
And the solution of problem (A.I-31), (A.I-34) is known, i.e.,
⎧ 1 p 1
⎪ (2)
⎨ Φ4 (ζ) = − 32c1 2 (a + d) ζ ,
⎪
(A.I-36)
⎪
⎪ p ζ
⎩ Φ3(2) (ζ) = − (a + d) .
32c1 (ζ 2 − 1)
(1)
The superposition of (A.I-35) and (A.I-36) gives the total solution Φ4 (ζ) = Φ4 (ζ)+
(2) (1) (2)
Φ4 (ζ), Φ3 (ζ) = Φ3 (ζ) + Φ3 (ζ), for example the first term of Φ4 (ζ) is
1 p 1 1 σc (a + d)ϕ2 1
− (a + d) + · · (A.I-37)
32c1 2 ζ 32c1 π ζ
and Φ2 (ζ) has not been listed here because it is too lengthy. So the stresses and
displacements are determined already. In addition, we know that
Φ (ζ)
σij , Hij ∼ (A.I-38)
ω (ζ)
here Φ(ζ) means Φ4 (ζ) or Φ3 (ζ), and ω (ζ) ∼ 1/(1 − ζ 2 ).

From Section 14.4 in the text we know that there is no stress singularity at
the Dugdale-Barenblatt crack tip, this fact and conjunct with (A.I-37) and (A.I-38)
require that the value of formula (A.I-37) must be zero, which leads to (14.4-6).
Considering this the final version of Φ4 (ζ) is

1 σc σ2 − ζ σ2 + ζ (ζ − σ2 )(ζ + σ2 )
Φ4 (ζ) = − · z ln + ln − a ln . (A.I-39)
32c1 2πi σ2 − ζ σ2 + ζ (ζ + σ2 )(ζ − σ2 )
And the displacement at the crack surface presents the form
uy (x, 0) = (128c1 c2 − 64c3 )Im(Φ4 (ζ))ζ=σ . (A.I-40)

After some calculation we find that

(4c1 c2 −2c3 ) σc (a + d) sin(ϕ2 −ϕ) (sin ϕ2 −sin ϕ)
uy (x, 0) = · · cos ϕ ln −cos ϕ2 ln ,
c1 2π sin(ϕ2 +ϕ) (sin ϕ2 +sin ϕ)
(A.I-41)
so the crack tip opening displacement
δI = CT OD = lim 2uy (x, 0) = lim 2uy (x, 0)

x→l ϕ→ϕ2
(A.I-42)
(8c1 c2 − 4c3 )σc a π σ (∞)
= ln sec ,
c1 π 2 σc
in which the constants c1 , c2 , c3 are defined in Section 11.3, so the solution holds for
point group 10, 10 decagonal quasicrystals as well. When we assume R1 = R, R2 = 0
in the equation (A.I-42), δI will be the corresponding solution of point group 10mm
decagonal quasicrystals, i.e.,

2σc a 1 K1 π σ (∞)
δI = CT OD = + ln sec ,
π L+M M K1 − R2 2 σc
which is just the (14.4-7). If let K1 = R = 0, L = λ, M = μ in above formula then

it exactly reduces to the classical Dugdale solution holding for engineering material
(or structural material) including crystalline material (referring to Section 14.4).
The more details can be found in article given by [25].
A.I.10 On the calculation of integral (9.2-14)
In formulas (9.2-14) y > 0 result in the integrals being convergent. We let ξ to

extend to complex number ξ1 + iξ2 , take integration path at complex ξ-plane similar
to Fig. A.I-2, and by physical consideration k(K1 − K2 ) > 0, μ(K1 − K2 ) − R2 > 0,
and k = μ(c) /h, then one can find the integrand of (9.2-14) in the interior of the
region bounded by the integration path is analytic except poles
(1) k(K1 − K2 ) (2) k(K1 − K2 )

ξ1 = < 0, ξ1 = − 2 >0 (A.I-43)
R2 − μ(K1 − K2 ) R − μ(K1 − K2 )
at real axis ξ1 , and by a generalized Jordan lemma the integral along the big half-
circle is zero.
# (1) # (2)
If in Fig. A.I-2 we put ω = ξ, ω1 = ξ1 , ω2 = ξ2 , k/m = ξ1 , − k/m = ξ1 ,
according to the additional integration path at complex ξ-plane, through the similar
manner for evaluating integral (A.I-18) then we can obtain results (9.2-15) and (9.2-
(1) (2)
16) respectively where ξ1 , ξ1 are defined by (A.I-43).
Appendix II Dual integral equations and some additional cal-

culations
Application of integral transforms in elasticity of quasicrystals may be effective and
more widely used than that of complex variable function method, and has got much
analytic solutions, see e.g. Li[20] , Zhou and Fan[21] , Zhou[22] , Zhu and Fan[23,24] etc,
due to the space limitation, the results have not been introduced. The application
of the method to crack problems often leads to some dual integral equations, so it
is beneficial for discussion about this.
A.II.1 Dual integral equations

It is well-known the Fourier transform or Hankel transform are very useful tool in
solving partial differential equations which has been shown in Chapters 7∼9 though
the introduction is very limited. For non-harmonic or non-multiharmonic equations
the complex potential method is not effective, we have to use the Fourier transform,
Hankel transform, Mellin transform or others. After the transform, the boundary
value problems of the dislocations are reduced to some algebraic equations to solve
(this is relatively simpler), while those of the cracks are concluded for solving the
following dual integral equations
⎧ ∞
⎪
⎪ y α f (y)Jν (xy)dy = g(x), 0 < x < 1,
⎨
0
∞ (A.II-1)
⎪
⎪
⎩ f (y)Jν (xy)dy = 0, x>1
0
or
⎧ ∞ n
⎪
⎪
⎪
⎪ y αj
ajk fj (y)Jνj (xy)dy = gj (x), 0 < x < 1,
⎨ 0
k=1
∞ j = 1, 2, · · · , n (A.II-2)
⎪
⎪
n
⎪
⎪
⎩ ajk fj (y)Jνj (xy)dy = 0, x > 1,
0 k=1
or
⎧ ∞ ∞
⎪
⎪ g1 (ξ1 , ξ2 , s, x1 , x2 )f (ξ1 , ξ2 , s)Jα (ξ1 x1 )Jβ (ξ2 x2 )dξ1 dξ2 = h(x1 , x2 , s),
⎪
⎪
⎪
⎪ 0 0
⎪
⎪
⎨ (x1 , x2 ) ∈ Ω1 ,
∞ ∞
⎪
⎪
⎪
⎪ g2 (ξ1 , ξ2 , s, x1 , x2 )f (ξ1 , ξ2 , s)Jα (ξ1 x1 )Jβ (ξ2 x2 )dξ1 dξ2 = 0,
⎪
⎪
⎪
⎪ 0 0
⎩
(x1 , x2 ) ∈ Ω2 .
(A.II-3)
Appendix II Dual integral equations and some additional calculations 349
Among them the equations (A.II-1) are the simplest ones, which will be discussed in
the following only. The equations (A.II-2) are dealt with multi-unknown functions,
and (A.II-3) are the two-dimensional dual integral equations, these two kinds of dual
integral equations are more complicated.
In equations (A.II-1), f (x) is unknown function to be determined, g(x) is known
one, α and ν are constants, Jν (xy) the first kind Bessel function of ν order. Titchm-
arsh[6] and Busbridge[7] gave the analytic solution of the equations. Various au-
thors[8∼16] discussed the solutions with different methods. Here only the procedure
of Refs. [6,7] is introduced. Titchmarsh[6] gave formal solution for the case α >
0. Busbridge extended the discussion to the case α > −2 and gave proof for the
existence of the solution. The solution is given through a complex integral as follows
k+i∞
1 Γ(1/2 + ν/2 + s/2)
f (x) = 2s−α ψ(s)x−s ds, (A.II-4)
2πi k−i∞ Γ(1/2 + ν/2 + α/2 − s/2)
in which s = σ + iτ and
C+i∞
1 Γ(1/2 + ν/2 − α/2 + w/2) g(α + 1 − w)
ψ(s) = · dw, (A.II-5)
2πi C−i∞ Γ(1/2 + ν/2 + w/2) w−s
where w = u + iv (in which v represents the imaginary part of complex variable w,

do not confuse with ν—the suffix of the Bessel function, which represents the order
of the Bessel function), σ < u and
1
g(α + 1 − w) = g(x)xα−w dx
0
in above formulas Γ(x)represent the Euler gamma function. The solution (A.II-4)
holds for both α > 0 and α > −2.
For α > 0, the solution can be expressed by real integral as
1
(2x)1−α/2 1 1+α/2
f (x) = μ Jν+α/2 (μx)dμ g(ρμ)ρν+1 (1 − ρ2 )α/2−1 dρ
Γ(α/2) 0 0
(A.II-4 )
and for α > −2, which is in form
1
2−α/2 x−α
f (x) = μ1+α/2 Jν+α/2 (x) y ν+1 (1 − y2 )α/2 g(y)dy
Γ(1 + α/2) 0
1 1 (A.II-4 )
+ y α+1 (1 − y 2 )α/2 dy (xu)2+α/2 g(yu)Jν+1+α/2 (xu)du .
0 0
1
Theorem If α > −2, − ν − 1 < α − < ν + 1, the Mellin transforms of g(x)
2
and f (x) exist, the latter is analytic in the strip region −ν < Re s = σ < α and
σ−α+ε
has the order O(|t| ) (ε > 0, t → ∞), where s = σ + iτ is the Mellin transform
parameter, then equations (A.II-1) have one and only one solution (A.II-4).
Proof Because the strict proof given by Busbridge [7] is very lengthy, we
cannot quote its all details here, instead, only a rough outline of the proof is figured
out in the following. One can find that in the proof a quite lot of complex variable
function knowledge are used, this seems that the theory on dual integral equations
presents inherent connection with complex analysis. So the Section I of this appendix
is helpful for the present discussion too.
1
At first assume that 0 < α < 2, −ν − 1 < α − < ν + 1 and the Mellin transform
2
of f (x) ∞
f (s) = f (x)xs−1 dx, s = σ + iτ
0
is analytic in region −ν < σ < α, and assume as ε > 0, as t → ∞ has order
−α+ε
O(|t| ).
According to the definition, the Mellin transform of y α Jν (xy) is
∞
2α+s−1 Γ(α/2 + ν/2 + s/2)
J α (s) ≡ [yα Jν (xy)]y s−1 dy = α+s . (A.II-6)
0 x Γ(1 − α/2 + ν/2 − s/2)
Recall that s = σ + iτ . By using the notations of relevant Mellin transform the
left-handside of the first and second equations (A.II-1) become
∞ C+i∞
1
α
y f (y)Jν (xy)dy = f (s)J α (1 − s)ds,
0 2πi C−i∞
∞ C+i∞
1
f (y)Jν (xy)dy = f (s)J 0 (1 − s)ds
0 2πi C−i∞
and substituting (A.II-6) into the above formulas yields
C+i∞ α−s
1 2 Γ(1/2 + α/2 + ν/2 − s/2)
f (s)ds = g(x), 0 < x < 1,
2πi C−i∞ x1−s Γ(1/2 − α/2 + ν/2 + s/2)
C+i∞ α−s
1 2 Γ(1/2 + ν/2 − s/2)
f (s)ds = 0, x > 1.
2πi C−i∞ Γ(1/2 + ν/2 + s/2)
Put
2α−s Γ(1/2 + ν/2 + s/2)
f (s) = ψ(s). (A.II-7)
Γ(1/2 + α/2 + ν/2 − s/2)
Then the above equations reduce to
⎧ C+i∞
⎪
⎪ 1 Γ(1/2 + ν/2 + s/2)
⎪
⎨ ψ(s)xs−1−α ds = g(x), 0 < x < 1,
2πi C−i∞ Γ(1/2 + ν/2 − α/2 + s/2)
⎪ C+i∞
⎪ 1
⎪ Γ(1/2 + ν/2 − s/2)
⎩ ψ(s)xs−1 ds = 0, x > 1.
2πi C−i∞ Γ(1/2 + ν/2 + α/2 − s/2)
(A.II-8)
Multiplying xα−w to the first one of (A.II-8), where w = u + iv and σ − u > 0, then
integrating over (0, 1) to x and
C+i∞
1 Γ(1/2 + ν/2 + s/2) ds
ψ(s) = g(α − w + 1), (A.II-9)
2πi C−i∞ Γ(1/2 + ν/2 − α/2 + s/2) s−w
where u < C and 1
g(α − w + 1) = g(x)xα−w dx.
0
The left-handside of equation (A.II-9) is analytic everywhere in the strip zone
−ν < σ < α
−α+ε
except the simple pole s = w, and behaves order O(|t| ). If we move the inte-
gration path from σ = C to σ = C < u, see Fig. A.II-1, based on the Cauchy’s
integral formula (referring to formula (A.I-5)),
C +i∞
1 Γ(1/2 + ν/2 + s/2) ds
ψ(s)
2πi C −i∞ Γ(1/2 + ν/2 − α/2 + s/2) s−w
Γ(1/2 + ν/2 + w/2)
= g(α − w + 1) − ψ(w).
Γ(1/2 + ν/2 − α/2 + w/2)
This translation of the integration line corresponds to form a closed region, the value
of the integral around the closed region is just equal to the second term of the above
formula including the sign of the term. Because the left-handside is analytic as
u > C , so the right-handside. In addition,
Γ(1/2 + ν/2 − α/2 + w/2)
ψ(w) − g(α − w + 1) (A.II-10)
Γ(1/2 + ν/2 + α/2)
is analytic for the case
1 1 1 1
+ ν − α + w = 0, −1, −2, · · · .
2 2 2 2
Fig. A.II-1 The integration path in s = σ + iτ -plane

Integrating function (A.II-10) along a big rectangle whose corners are
C − iT, C + iT, −T + iT, −T − iT, T > |v|
and one can find the absolute values of the integrals

" " " " " "
" −T +iT " " −T −iT " " C−iT "
" " " " " "
" ", " ", " "
" C+iT " " −T +iT " " −T −iT "
−α/2+ε
have order O(|t| ), and the value of ε can be always taken less than α/2, then
they can tend to zero as T → ∞. According to the Cauchy’s integral theorem
(referring to formula (A.I-4))
C+i∞

1 Γ(1/2 + ν/2 − α/2 + w/2) ds
ψ(s) − g(α − s + 1) = 0, u < C.
2πi C−i∞ Γ(1/2 + ν/2 + s/2) s−w
(A.II-11)
−w
Similarly, multiplying x to the second one of equations (A.II-8), where σ − w < 0,
then integrating to x over (1, ∞)
C +i∞
1 Γ(1/2 + ν/2 − s/2) ds
ψ(s) = 0. (u > C )
2πi C −i∞ Γ(1/2 + ν/2 + α/2 − s/2) s−w
Moving the integration path and find that
C+i∞
1 ds
ψ(w) = ψ(s) . (u < C) (A.II-12)
2πi C−i∞ s−w
Comparing (A.II-11) and (A.II-12), we find (A.II-5), substituing it into (A.II-7),
then taking inversion of the Mellin transform yields (A.II-4), the theorem is proved.
In above 1
g(α − s + 1) = g(ξ)ξ α−s dξ.
0
In addition, one has
1
1
= η s−w−1 dη.
s−w 0
If exchange integral order, then (A.II-5) may be rewritten as
1 1 C+i∞ −s
1 Γ(1/2 + ν/2 − α/2 + s/2) ξ
ψ(ω) = g(ξ)ξ α dξ η −w−1 dη× ds,
0 0 2πi C−i∞ Γ(1/2 + ν/2 + s/2) η
in which the integration[26,27]
C+i∞ −s
1 Γ(1/2 + ν/2 − α/2 + s/2) ξ
ds
2πi C−i∞ Γ(1/2 + ν/2 + s/2) η
⎧
⎨ 2
ξ 1+ν−α (η 2 − ξ 2 )α/2−1 η 1−ν , η ξ,
= Γ(α/2)
⎩ 0, 0 < η < ξ.
So that
1 1
2 1+ν
ψ(w) = g(ξ)ξ dξ η −w−ν (η 2 − ξ 2 )α/2−1 dη.
Γ(α/2) 0 ξ
By exchanging the integral order (this means the integration zone at ξη-plane is
changed), we may find that
1 1
2
ψ(w) = η −w−ν dη g(ξ)ξ 1+ν (η 2 − ξ 2 )α/2−1 dξ
Γ(α/2) 0 0
1 1
2
= η α−ω
dη g(ξ)ξ 1+ν (1 − ξ 2 )α/2−1 dξ.
Γ(α/2) 0 0
Substituting it into (A.II-4) yields

1 1
2
f (x) = η α dη g(ηζ)ζ 1+ν (1 − ζ 2 )α/2−1 dζ
Γ(α/2) 0 0
C+i∞
1 Γ(1/2 + ν/2 + s/2)
× 2s−α (xη)−s ds.
2πi C−i∞ Γ(1/2 + ν/2 + α/2 − s/2)
By utilizing inversion of the Mellin transform[26,27] ,

C+i∞
1 Γ(1/2 + ν/2 + s/2)
2s−α (at)−s ds = 2−α/2 (at)1−α/2 Jν+α/2 (at),
2πi C−i∞ Γ(1/2 + ν/2 + α/2 − s/2)
then one finds (AII.4 ).

For the case α > −2, the derivation is similar. For some details reader can refer
to Busbridge[7] .
In the following some examples are discussed in detail, which are the dual integral
equations appearing in Chapters 8 and 9 respectively, where only the solutions were
listed without derivation detail.

(8.3-8)
⎧ ∞
⎪ 2
⎪
⎪ [C(ξ) |ξ| − 6D(ξ)] cos(ξx)dξ = −p, 0 < x < a,
⎪
⎪ d11 0
⎪
⎪
⎪ ∞
⎪
⎪
⎪
⎪
⎨ ξ −1 [C(ξ) |ξ| − 6D(ξ)] cos(ξx)dξ = 0, x > a,
0
∞ (A.II-13)
⎪
⎪ 2
⎪
⎪ D(ξ) cos(ξx)dξ = 0, 0 < x < a,
⎪
⎪ d12 0
⎪
⎪
⎪ ∞
⎪
⎪
⎪
⎩ ξ −1 D(ξ) cos(ξx)dξ = 0, x > a.
0
It is evident that the second pair of dual integral equations (A.II-6) has the zero
solution, i.e., D(ξ) = 0, and we only consider the first pair in (A.II-13), which is
⎧ ∞
⎪ 2
⎪
⎨ C(ξ) |ξ| cos(ξx)dξ = −p, 0 < x < a,
d11 0
∞ (A.II-14)
⎪
⎪
⎩ −1
ξ C(ξ) |ξ| cos(ξx)dξ = 0, x > a.
0
Because 1/2
πξx
cos(ξx) = J−1/2 (ξx)
2
and denoting
1/2
1/2 x πad11
ξ C(ξ) = f (ξ), η = aξ, ρ= , g(ρ) = a p,
a 2ρ
then (A.II-14) is reduced to
⎧ ∞
⎪ 2
⎪
⎨ d ηf (η)J−1/2 (ηρ)dη = g(ρ), 0 < ρ < 1,
11 0
∞ (A.II-14 )
⎪
⎪
⎩ f (η)J−1/2 (ηρ)dη = 0, ρ > 1,
0
which becomes one of the standard dual integral equations shown by (A.II-1) with
1
α = 1, ν=− , g(ρ) = g0 ρ−1/2 , g0 = const = a(πad11 )1/2 p.
2
In this case it is very easy to calculate the solution of dual integral equations
(A.II-14) (or (A.II-14 )) by formulas (A.II-4) and (A.II-5), but the key step is choos-
ing the integration path. In previous introduction on Titchmarsh-Busbridge solu-
tion, we mentioned that it must require that −ν > k > α, −ν < C < α and k < C.
At present case ν = −1/2, α = 1 such that 1/2 < k < 1 and 1/2 < C < 1. The
concrete calculation is as below. Due to
1 1
g0
g(α + 1 − t) = g(ρ)ρα−t dρ = g0 ρ−1/2 ρ1−t dρ = ,
0 0 3/2 − t
where t = t1 + it2 represents a complex variable, and requires that t1 < 3/2. Sub-
stituting the relevant data and the above result into (A.II-5), we have
C+i∞
1 Γ(−1/4 + t/2) 1 1
ψ(s) = g0 dt,
2πi C−i∞ Γ(1/4 + t/2) t − s 3/2 − t
the integration path is shown in Fig. A.II-2. In this case the integrand has only one
pole at point t = 3/2 of first order. According to the formula for evaluating residue
(A.I-15) the above integral is very easy obtained as
√
Γ(1/2) 1 π
ψ(s) = g0 = g0 , (A.II-15)
Γ(1) 3/2 − s 3/2 − s

1 √
whereas Γ(1) = 1, Γ = π. Substituting the result into formula (AII.4) leads
2
to k+i∞ s−α
√ 1 2 Γ(1/4 + s/2) −s
f (η) = g0 π η ds. (A.II-16)
2πi k−i∞ Γ(1/4 − s/2)
In terms of the inversion of the Mellin transform[26,27] ,
k+i∞
1 Γ(1/2 + μ/2 + s/2)
2s−λ (βη)−s ds
2πi k−i∞ Γ(1/2 + μ/2 + λ/2 − s/2) (A.II-17)
−λ/2 1−λ/2
=2 (βη) Jμ+λ/2 (βη).
In formula (A.II-16), μ = −1/2, λ = 3, β = 1, so that

√
f (η) = g0 π(2η)−1/2 J1 (η)
and
πad11 −1
C(ξ) = ξ −1/2 f (ξ) = ξ J1 (aξ).
2
This is just the result given by (8.3-10).
Fig. A.II-2 The integration path in t = t1 + it2 -plane
The calculation through (A.II-4 ) and (A.II-4 ) yields the same result, so the
correctness of the result is demonstrated.

(9.8-8)
In Section 9.8, the dual integral equations
∞
ξAi (ξ)J0 (ξr)dξ = Mi p0 , 0 < r < a,
0
∞ (A.II-18)
Ai (ξ)J0 (ξr)dξ = 0, r>a
0
is solved and obtained the solution (9.8-8). We here give the detail for the derivation.
According to the standard type of the equations here α = 1, ν = 0, g(ρ) = g0 =
const, so
1
g0
g(α + 1 − t) = g(ρ)ρα−t dρ = , Ret = t1 < 2,
0 2−t
C+i∞
1 Γ(1/2) 1 1 2 1
ψ(s) = g0 dt = g0 √ ,
2πi C−i∞ Γ(1/2 + t/2) t − s 2 − t π2−s
in which s = σ + iτ , and the integral is evaluated through the residue of pole t = 2

and the integration path is shown in Fig. A.II-3.
Substituting the result into (A.II-4) yields
k+i∞
2 1 2s Γ(1/2 + s/2) −s 2g0
Ai (ξ) = f (ξ) = g0 √ ξ ds = √ ξ −1/2 J3/2 (ξ),
π 2πi k−i∞ Γ(2 − s/2) 2π
Fig. A.II-3 Integration path at the t = t1 + it2 -plane
which is just the solution (9.8-8), a little bit difference with that lies in we here used
the normalized coordinate ρ = r/a. In the last step of the calculation the inversion
of the Mellin transform (A.II-17) was used.
The above two subsections demonstrate the effect and simplicity of complex vari-
able function method in evaluating solutions of Titchmarsh-Busbridge dual integral
equations.
References 357
The system of dual integral equations (AII-2) and its applications are discussed
by Fan[17] , the two-dimensional dual integral equations (AII-3) are solved approxi-
mately by Fan and Sun[18] , in which some applications are also given.
References
[1] Muskhelishvili N I. Some Basic Problems of the Mathematical Theory of Elasticity,
Groningen: Noordhoff Ltd, 1953
[2] Privalov I I. Introduction to Complex Variable Functions Theory. Moscow: Science,
1984 (in Russian)
[3] Lavrentjev M A, Schabat B V. Methods of Complex Functions Theory. 6th ed.
Moscow, Science, 1986 (in Russian)
[4] Fan T Y. Semi-infinite crack in a strip. Chin Phys Lett, 1990, 8(9): 401–404
[5] Shen T W, Fan T Y. Two collinear semi-infinite cracks in a strip. Eng Fract Mech,
2003, 70(8): 813–822
[6] Titchmarsh B C. Introduction to the Fourier Integrals. Oxford: Clarenden Press,
1937
[7] Busbridge I W. Dual integral equations. Math Soc Proc, 1938, 44(2): 115–129
[8] Weber H. Ueber die Besselschen Functionen und ihre Anwendung auf die Theorie der
electrischen Stroeme. J fuer reihe und angewandte Mathematik, 1873, 75(1): 5–105
[9] McDonald H M. The electrical distribution induced on a disk plased in any fluid of
force. Phil Mag, 1895, 26(1): 257–260
[10] King L V. On the acoustic radiation pressure on circular disk. Roy Soc London Proc
Ser A, 1935, 153(1): 1–16
[11] Copson E T. On the problem of the electric disk. Edinburg Math Soc Proc, 1947,
8(1): 5–14
[12] Tranter C J. On some dual integral equations. Quar J Math Ser 2, 1951, 2(1): 60–66
[13] Sneddon I N. Fourier Transforms. New York: McGraw-Hill, 1951
[14] Gordon A N. Dual integral equations. London Math Soc J, 1954, 29(5): 360–369
[15] Noble B. On some dual integral equations. Quar J Math Ser 2, 1955, 6(2): 61–67
[16] Noble B. The solutions of Bessel function dual integral equations by a multiplying
factor method. Cambridge Phil Soc Proceeding, 1963, 59(4): 351–362
[17] Fan T Y. Dual integral equations and system of dual integral equations and their
applications in solid mechanics and fluid mechanics. Mathematica Applicata Sinica,
1979, 2(3): 212–230 (in Chinese)
[18] Fan T Y, Sun Z F. A class of two-dimensional dual integral equations and applica-
tions. Appl Math Mech, 2007, 28(2): 247–252
[19] Wolfran St. The Mathematica Book. 3 rd . Wolfran St. Cambridge: Cambridge
University Press, 1996
[20] Li X F. The defect problems and their analytic solutions in the theory of elasticity
of quasicrystals. Dissertation. Beijing Institute of Technology, 1999 (in Chinese)
[21] Zhou W M, Fan T Y. Plane elasticity problem of two-dimensional octagonal qua-
sicrystal and crack problem. Chin Phys, 2001, 10(8): 743–747
[22] Zhou W M. Mathematical analysis of elasticity and defects of two- and three-dimen-
sional quasicrystals. Dissertation. Beijing: Beijing Institute of Technology, 2000 (in
Chinese)
[23] Zhu A Y, Fan T Y, 2007, Elastic analysis of a mode II crack in an icosahedral
quasicrystal, Chin. Phys., 16(5), 1111–1119
[24] Zhu A Y, Fan T Y, 2009, Elastic analysis of a Griffith crack in an icosahedral Al-
Pd-Mn quasicrystal, Int. J. Modern Phys. B, 23(10), 1–16
[25] Li W, Fan T Y, Study on plastic analysis of crack problem in icosahedral quasicrys-
tals. Phil. Mag., 2009, 89(31): 2823–2831.
[26] Watson G N. A Treatise on the Theory of Bessel Functions. New York: Amazon
Com, 1955
[27] Erdelyi A. Higher Transcendental Functions. Vol.2. New York: McGraw-Hill, 1953
Index
A Conjugate vector of reciprocal vector, 45

Absolute temperature, 6 Complex variable function, 127
Acoustic wave, 6 Conformal mapping, 132
Al-Mn alloy, 24 Compact embedding, 280
Amount of dislocation, 300 Compact tension specimen, 314
Analytic function, 56 Complete crystal, 3
Analytic extension, 332 Constitutive law, 282
Angular momentum conservation, 36 Contact problem, 99
Anti-plane elasticity, 54 Coordinate frame, 52
Aphomus (glassy structure), 10 Coordinate transformation, 52
“Apparent” mass of dislocation, 192 Crack, 99
Atomic cohesive force, 289 Crack speed, 193
Axisymmetric case, 182 Crack growth force, 131
B Crack strain energy release rate, 131
Bessel function, 185 Crack tip opening displacement, 292
Biharmonic equation, 86 Crack tip tearing displacement, 290
Bilinear functional, 271 Crystal-quasicrystal coexisting, 124
Bloch theorem, 28 Crystal-quasicrystal phase transition, 128
Body force, 6 Cubic system, 22
Boundary conditions, 40 Cubic quasicrystal, 37
Bragg reflection, 24 D
Brillouin zone, 28 Decagonal quasicrystal, 28
Brittle fracture, 130 Decagonal Al-Ni-Co quasicrystal, 285
Burgers vector, 102 Decomposition, 54
C Deformation, 16
Cauchy’s formula, 327 Deformation energy, 35
Cauchy’s integral, 132 Deformation tensor, 34
Cauchy-Riemann relations, 127 “differenc quotient”, 313
Classical elasticity, 12 Diffraction spot, 24
Classical solution, 271 Diffraction pattern, 70
Column vector, 258 Diffusion, 48
360 Index
Direct sum, 27 Fermi energy, 29

Discrete form of the energy functional, 258 Fibonacci sequence, 46
Discrete spectrum, 7 Finite difference scheme, 205
Dislocation, 99 Finite multi-connected region p229
Dislocation core, 103 Finite element method, 257
Dislocation density function, 299 Five-fold symmetry, 25
Dislocation network, 103 Flow rule of isotropic hardening, 256
Dislocation pile-up, 298 Fourier transform, 105
Dislocation velocity, 190 Fracture criterion, 131
Displacement, 16 Fracture toughness, 318
Displacement boundary condition, 40 Free energy, 8
Displacement potential function, 61
Function equation, 233
Dodecagonal quasicrystal, 28
Fundamental solution, 205
Dual integral equations, 137
G
Dynamic fracture, 316
Galilean transformation, 190
Dynamic initiation of crack growth, 204
Gauss theorem, 36
Dynamic stress intensity factor, 207
Generalized BCS model, 295
E
Generalized body force, 36
Eight-fold symmetry, 25
Generalized Dugdal-Barenblatt model, 289
Elastic constant matrix, 19
Generalized Eshelby theory, 294
Elastic wave, 6
Generalized Eshelby energy-momentum ten-
Elasticity, 15
sor, 294
Elastodynamics, 22
Generalized Hooke’s law, 18
Elasto-/hydro-dynamics, 38
Generalized Eshelby integral, 295
Electronic microscopy, 24
Generalized Lax-Milgram theorem, 269
Element interpolation matrix, 259
Element stiffness matrix, 258 Generalized solution (weak solution), 271
Element strain matrix, 258 Generalized traction, 40
Elementary excitation, 7 Generalized variational principle, 253

Elliptic notch, 144 Global stiffness matrix, 260
Embedding, 280 Griffith crack, 125
Energy functional, 254 H
Engineering elasticity, 96 Hall effect, 28
Entropy, 7 Hankel transform, 184
Equilibrium state, 7 Hardening modulus, 286
Euler gamma function, 341 Hardness, 318
F Hexagonal system, 21
Fast crack propagation, 204 Higher dimensional space, 27
Index 361
Hilbert space, 270 M

Holomorphic, 230 Manmade damping coefficient, 214
Horizontal mapping, 3 Mathematical elasticity, 96
I Matrix representation of elasticity, 272
Icosahedral Al-Pd-Mn quasicrystal, 214 Mesh, 217
Incommensurate phase, 10 Miller index, 27
Incremental plasticity, 286 Mellin transform, 340
Initial condition, 40 Metric defect, 99
Inner product, 270 Minimum value, 257
Inner variable, 284 Mixed boundary condition, 40
Interface, 101 Moessbauer spectroscopy, 48
Inversion, 4 Momentum conservation, 36
Inversion of Fourier transform, 106 Monoclinic system, 19
Inversion of Mellin transform, 345 Monte-Carlo simulation, 68
Isotropic body, 22 Moving boundary problem, 210
J Moving Griffith crack, 193
Jordan lemma, 331 Mutually conjugate harmonic functions, 326
K N
Kinetic coefficient, 37 Neutron scattering, 48
Korn inequality, 277 NMR (nuclear magnetic resonance), 48
L Node, 285
Lattice dynamics, 6 Non-crystallographic point groups, 27
Lattice wave, 6 Nonlinear elasticity, 287
Lattice vibration, 6 Nonlinear behaviour of quasicrystals, 282
Laue class, 27 Nonlinear fracture mechanics, 315
Law of crystal symmetry, 3 Norm, 277
Lebesgue square integrable, 270 Number of nodes, 285
Light conductivity rate, 28 O
Linear fracture theory, 310 One-dimensional quasicrystals, 28
Linear functional, 271 Order parameter, 9
Linear transformation, 14 Orientational symmetry, 3
Loading surface, 285 Orthorhombic system, 20
Local discontinuity, 99 P
Local isomorphism, 26 Paradigm, 7
Longitudinal shear, 125 Parallel space, 27
Long-range order, 11 Path independency, 295
Long-wavelength, 6 Penrose tiling, 26
362 Index
Penny-shaped crack, 184 Row vector, 258

Pentagonal quasicrystal, 28 S
Periodicity, 1 SEM morphology, 319
Perturbation, 197 SEM fractolograph, 320
Phonetic frequency mode, 7 Semi-norm, 277
Phonon, 6 Sextuple harmonic equation, 167
Phonon strain tensor, 35 Singular integral equation, 299
Phason, 10 Sobolev space, 270
Phason strain tensor, 35 Solvability, 281
Photon, 6 Space elasticity, 61
Photonic frequency mode, 6 Space step, 207
Photonic crystal, 28
Specific heat, 6
Planar quasicrystal, 29
Stability (of finite difference scheme), 207
Planck quantum theory, 6
Strain tensor, 16
Plane elasticity, 54
Strain rate, 283
Plastic constitutive equation, 285
Stress potential function, 77
Plasticity, 282
Stress singularity, 99
Plastic zone, 280
Stress tensor, 17
Point operation, 3
St. Venant boundary condition, 98
Point group, 4
Symmetry-breaking, 7
Poles, 329
T
Positive definite, 256
Taylor expansion, 16
Power law form, 283
Ten-fold symmetry, 25
Projection, 46
Tetragonal system, 3
Pseudogap p29
Theorem of Laurent expansion, 232
Q
Quadruple harmonic equation, 76 Three-dimensional quasicrystal, 28
Quanta, 6 Three point bending specimen, 307
Quantum, 6 Time step, 207
Quasicrystal, 12 Topological defect, 99
Quasi-particle, 6 Total plasticity, 286
Quasiperiodic symmetry, 26 Traction, 18

R Transcendental mapping function, 333
Reciprocal lattice, 5 Translational symmetry, 1
Residue theorem, 329 Triclinic system, 2
Rhombohedral system, 3 Twelve-fold symmetry, 25
Rotation-inversion, 4 Two-dimensional quasicrystal, 28
Index 363
U Vertical mapping, 3
Uniqueness of solution, 267 X
V X-ray diffraction patterns, 1
Vertical space, 27 Y
Variational principle, 252 Yield surface, 285
Variational method Yield strength, 285

(Tianyou Fan) Mathematical Theory of Elasticity of (BookFi)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Tianyou Fan) Mathematical Theory of Elasticity of (BookFi)

Uploaded by

Copyright:

Available Formats

Tianyou Fan

Mathematical Theory of Elasticity of

ISBN 978-3-642-14642-8 e-ISBN 978-3-642-14643-5

Library of Congress Control Number: 2010931505

© Science Press Beijing and Springer-Verlag Berlin Heidelberg 2011

Cover design: Frido Steinen-Broo, EStudio Calamar, Spain

Printed on acid-free paper

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

Science), Longan Ying (School of Mathematics of Peking University), Dihua Ding

boundary value problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

10 ten-fold symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

13.2 Generalized Lax-Milgram theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

1.1 Periodicity of crystal structure, crystal cell

1.2 Three-dimensional lattice types

T. Fan, Mathematical Theory of Elasticity of Quasicrystals and Its Applications

Fig. 1.2-1 The 14 Crystal cells of three-dimension

1.3 Symmetry and point groups

which are the basic symmetric elements of point symmetry.

The mapping-rotation is a composite operation, denoted by Sn , which can be

Table 1.3-1 32 Point groups of crystals

1.4 Reciprocal lattice

where Ω = a1 · (a2 × a3 ) is the volume of lattice cell.

in which G called the reciprocal vector and

Points in the lattice can be described by a1 , a2 , a3 as well as by b1 , b2 , b3 .

1.5 Appendix of Chapter 1: Some basic concepts

1.5.1 Concept of phonon

belong to elementary excitations. In general one elementary excitation corresponds

F (η, T ) = F0 (T ) + A(T )η 2 + B(T )η 4 + · · · , (1.5-2)

in which, according to the requirement of the stability condition of phase transition

A(T ) = α(T − TC ), B(T ) = B(TC ), (1.5-3)

in which α is a constant. Without concerning concrete micro-mechanism, the Landau

where G is a reciprocal vector, and LR the reciprocal lattice in three-dimensional

ρG = |ρG | e−iΦG (1.5-6)

in which both G and u are in three-dimensional physical space. If we consider

1.5.2 Incommensurate crystals

1.5.3 Aphomus(glassy structure)

As the knowledge of crystals is a beneﬁt for understanding quasicrystals, it is worth-

2.1 Review on some basic concepts

2.1.2 Coordinate frame

T. Fan, Mathematical Theory of Elasticity of Quasicrystals and Its Applications

and e3 = e1 × e2 , e2 = e3 × e1 , e1 = e2 × e3 , then they are the base vectors of an

2.1.3 Coordinate transformation

e1 = c11 e1 + c12 e2 + c13 e3 ,

[C]T = [C]−1 , (2.1-4)

Based on (2.1-1), in the frame e1 , e2 , e3 ,

a = a1 e1 + a2 e2 + a3 e3 (2.1-6)

Substituting (2.1-5) into (2.1-1) yields

a = (c11 a1 + c12 a2 + c13 a3 )e1 + (c21 a1 + c22 a2 + c23 a3 )e2

+ (c31 a1 + c32 a2 + c33 a3 )e3 . (2.1-7)

It follows that by the comparison between (2.1-6) and (2.1-7)

a1 = c11 a1 + c12 a2 + c13 a3 ,

or in the matrix expression, i.e.,

Whatever (2.1-8) or (2.1-8 ), there is

in which the components satisfy the relation

2.1.5 Algebraic operation of tensor

3) Algebraic sum of a tensors

A ± B = Aij ± Bij . (2.1-14)

4) Product of a tensor by a scalar

Other operations about tensors will be provided where required.

2.2 Basic assumptions of theory of elasticity

2.3 Displacement and deformation

Fig. 2.3-1 Displacement of a point in an elastic body

2.4 Stress analysis and equations of motion

in which the suﬃxes j, l belong to space E3 and i, k to space E⊥