Advances in Applied Mechanics

Visit the link below to download the full version of this book:

https://medipdf.com/product/advances-in-applied-mechanics/

Click Download Now

Academic Press is an imprint of Elsevier
525 B Street, Suite 1900, San Diego, CA 92101-4495, USA
30 Corporate Drive, Suite 400, Burlington, MA 01803, USA
84 Theobald’s Road, London WC1X 8RR, UK
Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands

First edition 2009

Copyright # 2009 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system

or transmitted in any form or by any means electronic, mechanical, photocopying,
recording or otherwise without the prior written permission of the publisher.

Permissions may be sought directly from Elsevier’s Science & Technology Rights
Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333;
email: [email protected]. Alternatively you can submit your request online by
visiting the Elsevier website at http://elsevier.com/locate/permissions, and selecting
Obtaining permission to use Elsevier material.

Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons or
property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions or ideas contained in the material herein.

ISBN: 978-0-12-374813-3
ISSN: 0065-2156

For information on all Academic Press publications

visit our website at www.elsevierdirect.com

Printed and bound in USA

09 10 11 12 10 9 8 7 6 5 4 3 2 1
 Contents

PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

A Shell Theory for Carbon Nanotubes Based on the Interatomic

Potential and Atomic Structure

J. Wu, K. C. Hwang and Y. Huang

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Interatomic Potentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3. A Plate Theory for Graphene Based on the Interatomic Potential . . . . . 10
4. A Linear Shell Theory for Carbon Nanotubes Based on the
Interatomic Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5. Can a Single-Wall Carbon Nanotube Be Modeled as a Thin Shell?. . . . 25
6. A Nonlinear, Finite-Deformation Shell Theory for Carbon Nanotubes
Based on the Interatomic Potential . . . . . . . . . . . . . . . . . . . . . . . . . 34
7. Instability of Carbon Nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . 49
8. Concluding Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
A. The Incremental Form of Equilibrium Equation . . . . . . . . . . . . . . . . 63
B. The Symmetric Membrane Stress and Moment, and Covariant
Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Tensegrity: 60 Years of Art, Science, and Engineering

C. Sultan

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2. Tensegrity Origins: The Pioneers. . . . . . . . . . . . . . . . . . . . . . . . . . 71

v
vi Contents

3. From Abstract Art to Abstract Science . . . . . . . . . . . . . . . . . . . . . 76

4. The Blossoming 1990s and Beyond . . . . . . . . . . . . . . . . . . . . . . . 84
5. Advances in Statics Research . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6. Advances in Dynamics Research . . . . . . . . . . . . . . . . . . . . . . . . . 97
7. Deploying Tensegrity Structures . . . . . . . . . . . . . . . . . . . . . . . . 112
8. Controllable Tensegrity Structures . . . . . . . . . . . . . . . . . . . . . . . 117
9. Tensegrity Structures in Biology . . . . . . . . . . . . . . . . . . . . . . . . 125
10. The Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
11. Challenges for Controllable Tensegrity Structures . . . . . . . . . . . . . 134
12. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Skin Biothermomechanics: Modeling and

Experimental Characterization

F. Xu and T. J. Lu

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
2. Review of Related Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3. Modeling of Skin Biothermomechanics and Thermal Pain . . . . . . . . 162
4. Experimental Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
5. Biothermomechanical Behavior of Skin Tissue. . . . . . . . . . . . . . . . 196
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

SUBJECT INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

 Preface

There are few areas in science and engineering where applied mechanics does not
continue to make significant contributions. This volume of Advances in Applied
Mechanics is proof of this by highlighting the role of solid mechanics in topics
ranging from carbon nanotubes to the response of skin tissue.
Carbon nanotubes (CNTs) hold many promises, including improved mechan-
ical behavior of materials based on or reinforced with CNTs. Predictive modeling
of such properties requires that the CNT behavior can be cast in a continuum
model. One of the challenges there is that the wall of a CNT is only one atom
thick. The chapter by Wu, Hwang, and Huang shows how to systematically
derive a continuum shell description of a CNT on the basis of its atomic structure
and a state-of-the-art interatomic potential for carbon. The resulting CNT model
is different from all existing shell theories, but can be reduced to it for specific
CNTs when allowing for (known) relative errors. The finite-deformation version
of this atomistic-based shell theory is used to study the instability of single-wall
CNTs subject to structural loading states, such as compression and torsion.
The second chapter, by C. Sultan from Virginia Tech, also addresses structures
but of a special type: tensegrity structures. While originating in the avant garde art of
the twentieth century, the concept of a tensegrity structure has made its way into a
variety of applications in engineering and is even recognized nowadays in applied
sciences. Beside a broad historical overview of the evolution of tensegrity struc-
tures, this chapter provides a detailed yet concise summary of the fundamental
mechanics that gives these frameworks their peculiar static and dynamical proper-
ties. Modern applications, for instance in the form of deployable structures for
aeronautics, are discussed as well as the possible role of tensegrity in cell biology.
Biology, or perhaps medicine, is also the area of application for the third and
final chapter of this volume, co-authored by the Cambridge–China team of Xu
and Lu. This chapter provides an extensive outline of the interaction between heat
conduction, thermal damage, and the mechanical response of skin in determining
thermal pain sensations. An experimental methodology encompassing a variety
of techniques is presented against the background of recent models for skin
biothermomechanics and thermal pain. The results supplement knowledge of
the mechanical properties of skin tissue under normal physiological conditions
with detailed characterization of the changes in mechanical properties of skin
tissue due to temperature changes and the corresponding thermal damage. The
new insights and model parameters are expected to be crucial for thermal therapy
in skin disease treatments.
Hassan Aref
Erik van der Giessen

vii
 A Shell Theory for Carbon Nanotubes Based on the
Interatomic Potential and Atomic Structure

J. WU,a K. C. HWANG,a and Y. HUANGb

a
FML, Department of Engineering Mechanics, Tsinghua University, Beijing 100084, China
b
Departments of Civil and Environmental Engineering, and Mechanical Engineering,
Northwestern University, Evanston, Illinois 60208, USA

Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Interatomic Potentials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3. A Plate Theory for Graphene Based on the Interatomic Potential. . . . . 10
4. A Linear Shell Theory for Carbon Nanotubes Based
on the Interatomic Potential. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1. Membrane Strains and Curvatures. . . . . . . . . . . . . . . . . . . . . . . . . 18
4.2. Constitutive Relations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.3. Equilibrium Equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5. Can a Single-Wall Carbon Nanotube Be Modeled as a Thin Shell?. . . 25
5.1. Order of Error (D/R)3: Not a Classical Linear Elastic Shell. . . . . 26
5.2. Order of Error (D/R)2: A Linear Elastic Orthotropic Shell. . . . . . 26
5.3. Order of Error (D/R): A Classical Linear Elastic Shell with a
Universal Shell Thickness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6. A Nonlinear, Finite-Deformation Shell Theory for Carbon Nanotubes
Based on the Interatomic Potential. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.1. General Description of a Curved Surface. . . . . . . . . . . . . . . . . . . . 34
6.2. Membrane Strains and Curvatures of a Deformed Surface. . . . . . 35
6.3. Deformation of a Single-Wall Carbon Nanotube. . . . . . . . . . . . . . 36
6.4. Membrane Stresses and Moments in a Carbon Nanotube. . . . . . . 38
6.5. Equilibrium Equations for Membrane Stresses and Moments. . . 39
6.6. Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7. Instability of Carbon Nanotubes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
7.1. Instability of Carbon Nanotubes in Tension. . . . . . . . . . . . . . . . . . 50
7.2. Instability of Carbon Nanotubes in Compression. . . . . . . . . . . . . 53
7.3. Instability of Carbon Nanotubes Subject to Internal Pressure. . . 56
7.4. Instability of Carbon Nanotubes Subject to External Pressure. . . 58
7.5. Instability of Carbon Nanotubes in Torsion. . . . . . . . . . . . . . . . . . 60
8. Concluding Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

ADVANCES IN APPLIED MECHANICS, VOL. 43 1 Copyright # 2009 by Elsevier Inc.

ISSN: 0065-2156 DOI: 10.1016/S0065-2156(09)43001-1 All rights reserved.
2 J. Wu et al.

Appendix A The Incremental Form of Equilibrium Equation. . . . . . . . . . . . 63

Appendix B The Symmetric Membrane Stress and Moment, and
Covariant Derivatives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Abstract

A finite-deformation shell theory for single-wall carbon nanotubes (CNTs) is established

directly from the atomic structure of CNT and the interatomic potential by accounting for the
important effect of moment and curvature (of CNTs) in the Cauchy–Born rule. The theory
incorporates the effect of bending moment and curvature for a curved surface, and accurately
accounts for the nonlinear, multibody atomistic interactions, as well as the carbon nanotube
chirality. It avoids the ambiguous definition of nanotube thickness, and provides the consti-
tutive relations among stress, moment, strain, and curvature via the interatomic potential.
The constitutive behavior of a CNT is different from that of a classical shell, but its overall
structural response at infinitesimal deformation may still be approximately represented by a
linear elastic thin shell for some representative loadings such as tension, compression,
bending, torsion, internal, and external pressure. The ratio of atomic spacing (D
0.14
nm) to CNT radius, D/R, is used to identify the order of error, where D/R ranges from zero (for
graphene) to about 40% [for the (5,5) armchair CNT (R ¼ 0.35 nm)]. For the order of error O
[(D/R)3] (as compared to unity), which is about 6% for the (5,5) armchair CNT, the structural
response of a CNTs cannot be represented by any classical shell. For the order of error O
[(D/R)2], which is about 16% for the (5,5) armchair CNT, a CNT can be approximated by a
linear elastic orthotropic thin shell. Only for the order of error O(D/R), which is about 40% for
the (5,5) armchair CNT, a universal constant shell thickness and Young’s modulus can be
defined, and CNTs can be represented by an elastic isotropic thin shell.
The instability of single-wall CNTs subject to tension, compression, internal and
external pressure, and torsion is studied.

1. Introduction

A single-wall carbon nanotube (CNT), as illustrated in Fig. 1.1, has a single layer
of carbon atoms. It has a perfect hexagonal lattice structure, a diameter on the order
of 1 nm and a length ranging from 1 to 100 mm (or even longer). The structure of a
single-wall CNT can be conceptualized by wrapping a graphene (one-atom-thick
layer of graphite) into a seamless cylinder. As shown in Fig. 1.2, the graphene sheet is
wrapped in the circumferential direction Ch ¼ na1 þ ma2 of CNT represented by a
pair of indices (n, m). The integers n and m denote the number of unit vectors a1
and a2 along two directions in the honeycomb crystal lattice of graphene. If m ¼ 0,
the nanotubes are called ‘‘zigzag.’’ If n ¼ m, the nanotubes are called ‘‘armchair.’’
Otherwise, they are called ‘‘chiral.’’ The carbon atoms inside a graphene sheet are
 A Shell Theory for Carbon Nanotubes 3

A C

B
D

A C

FIG. 1.1 A carbon nanotube (CNT) prior to deformation: (A) a CNT and (B) a representative
atom (A) and its three nearest-neighbor atoms (B, C, and D).

covalently bonded through overlapping s-orbitals, making it the stiffest material

known today in the absence of defects (Lee, Wei, Kysar, & Hone, 2008). The
p-orbitals of graphene out of the plane start to interact for CNTs with small radius.
As compared to metals, carbon nanotubes have superior properties, such as
low mass density (1/6 of steel mass density), high Young’s modulus on the
order of 1 TPa (6 times steel modulus) (e.g., Lee et al., 2008; Qian, Wagner,
Liu, Yu, & Ruoff, 2002; Srivastava, Menon, & Cho, 2001; Yakobson & Avouris,
2001), and therefore have many potential applications.
There are extensive atomistic studies of CNT mechanical behavior
(e.g., Belytschko, Xiao, Schatz, & Ruoff, 2002; Samsonidze, Samsonidze, &
Yakobson, 2002; Yakobson, Brabec, & Bernholc, 1996; Yakobson, Campbell,
Brabec, & Bernholc, 1997). Atomistic simulations, in general, are limited in time
scale (up to 1 ns) and size scale (up to 1 mm). The atomic-scale finite element
method has been developed recently to extend the continuum finite element
method to atomic scale by explicitly accounting for the nonlinear and multibody
atomistic interactions (Liu, Huang, et al., 2004; Liu, Jiang, Huang, Qu, Yu, et al.,
2005). It provides a seamless linkage between the molecular mechanics and
continuum methods (at least at 0 K).
4 J. Wu et al.

a1
T
a2

(n,0) zigzag

Ch = na1 + ma2

(n,n) armchair

FIG. 1.2 A schematic diagram of graphene sheet wrapped seamlessly to a single carbon
nanotube.

There also exist two types of continuum models for CNTs. The first is to
model CNTs by continuum structural elements, such as spring, truss, beam,
membrane, plate, and shell, and is called the type-I continuum model in this
chapter. For a few representative loading cases (e.g., tension, bending), the
material and geometric properties of continuum elements (e.g., elastic modulus,
thickness) are determined by fitting the atomistic simulation results. For example,
Yakobson et al. (1996) modeled CNTs as classical linear elastic isotropic shells,
which have a strain energy
ð ð ( " 2   
#
1
E¼ D kx þ ky  2 1  v kx ky  kxy
2
2
" 2   #) (1.1)
C 
þ ex þ ey  2 1  v ex ey  exy
2
dS;
1  v2

where D and C are the bending and tension rigidities, respectively; v is the
Poisson’s ratio; kx, ky, and kxy are the curvatures; ex, ey, and exy are the membrane
strains; and the integration is over the CNT surface. By fitting the energy obtained
from molecular dynamics simulations (Robertson, Brenner, & Mintmire, 1992)
 A Shell Theory for Carbon Nanotubes 5

for CNTs in uniaxial tension and in bending (from a graphene sheet to a CNT),
Yakobson et al. (1996) determined D ¼ 0.85 eV, C ¼ 59 eV/atom ¼ 360 J/m2, and
v ¼ 0.19. Carbon nanotubes have also been modeled as Euler beams (Govindjee &
Sackman, 1999), trusses and plates (Odegard, Gates, Nicholson, & Wise, 2002),
trusses and torsion springs (Chang & Gao, 2003; Nasdala & Ernst, 2005; Natsuki,
Tantrakarn, & Endo, 2004), cylinders (Gao & Li, 2003), beams with circular cross
sections (Li & Chou, 2003, 2004), cylindrical shells (Sears & Batra, 2004), trusses
between nearest and second nearest atoms (Leung, Guo, He, & Kitipornchai,
2005), trusses and torsion springs or membranes and torsion springs (Wang,
Zhang, & Lu, 2005), and spatial trusses and shells (Chen & Cao, 2006). These
types of continuum models rely on fitting the energy from atomistic simulations,
and is restricted to linear elastic analysis (and often elastic isotropy).
Another type of continuum models is based on the interatomic potential and
atomic structure of carbon nanotubes, and is called the type-II continuum model in
this chapter. They incorporate the nonlinear, multibody interatomic potential for
carbon directly into the continuum analysis (e.g., Arroyo & Belytschko, 2002,
2004; Jiang et al., 2003; Zhang, Huang, Gao, & Hwang, 2002; Zhang, Huang,
Geubelle, & Hwang, 2002; Zhang, Huang, Geubelle, Klein, & Hwang, 2002;
Zhang, Jiang, Huang, Geubelle, & Hwang, 2004) via the Born rule (Born &
Huang, 1959), that is, to equate the continuum strain energy to energy in atomic
bonds. Arroyo & Belytschko (2002, 2004) used the exponential Born rule and
geodesic to study CNTs and graphene. The atomistic-based continuum theories of
Jiang et al. (2003), Zhang, Huang, Geubelle, Klein, et al. (2002), and Zhang et al.
(2004) account for the CNT chirality, and agree well with molecular dynamics
simulations without any parameter fitting. Their theories have been extended to
account for the effect of finite temperature (e.g., Jiang, Huang, & Hwang, 2005;
Jiang, Liu, et al., 2004) and coupled electromechanical behavior (e.g., Johnson,
Liu, & Huang, 2004; Liu, Jiang, et al., 2004), and to study defect nucleation in
CNTs (e.g., Jiang, Feng, et al., 2004; Song et al., 2006). However, they are
membrane theories that cannot account for the effect of graphene bending,
which is important in the experimental determination of the CNT elastic modulus
from their buckling (Khang et al., 2008), and in many applications such as
CNT-based sensors (e.g., Tombler et al., 2000), oscillators (e.g., Jiang, Yu, et al.,
2004), bio-inspired adhesion (e.g., Zhou, Huang, Liu, Wu, Hwang, et al., 2007),
and CNT-reinforced composite materials (e.g., Shi, Feng, Huang, & Hwang, 2004;
Shi, Feng, Huang, Hwang, & Gao, 2004; Tan, Jiang, Huang, Liu, & Hwang, 2007).
Recently Wu, Hwang, and Huang (2008) developed a finite-deformation shell
theory directly from the interatomic potential for carbon via the modified Born
rule. It belongs to the second type of continuum models discussed above but does
account for the effect of moment and curvature. Its equilibrium equations and
6 J. Wu et al.

boundary conditions are the same as the classical shell theories of Koiter (1966),
Niordson (1985), and Sanders (1963) (see Wu, Hwang, Huang, & Song, 2008a).
However, its constitutive model, which relates increments of second Piola–
Kirchhoff membrane stress T_ and moment M _ to increments of Green strain E_
_
and curvature K, is established from the interatomic potential and is very differ-
ent from the linear elastic isotropic constitutive model in the following three
aspects (see Wu, Peng, Hwang, Song, & Huang, 2008):
(i) The constitutive model becomes nonlinear (with 2% tolerance) for strains
beyond 1%.
(ii) The constitutive model is anisotropic for CNTs with a diameter smaller
than 1 nm. For CNTs with a diameter larger than 1 nm, the constitutive
model also becomes anisotropic as deformation increases.
(iii) The constitutive model has couplings between the membrane stress and
curvature, and between the moment and membrane strain.
Even though the constitutive behavior of CNTs is not strictly linear elastic isotropic,
is it possible that the overall structural response of CNTs be approximately repre-
sented by a linear elastic thin shell for some representative loadings such as tension,
compression, bending, torsion, internal, and external pressure? In other words, can
the structural response (global behavior) of CNTs be approximately represented by
a type-I continuum model of linear elastic thin shell, even if the constitutive
response (local behavior) cannot? If yes, how large is the error? What are the elastic
modulus and shell thickness of CNTs? Is the thin shell elastically isotropic? Peng,
Wu, Hwang, Song, and Huang (2008) answered these questions. Similar to Huang,
Wu, and Hwang (2006) for graphene, they obtained the tension and bending
rigidities (as well as the shear and torsion rigidities) of CNTs directly from the
interatomic potential. The ratio of atomic spacing (D
0.14 nm) to CNT radius (R),
D/R, was used to identify the error, which ranges from zero (for graphene) to about
40% [for a small (5,5) armchair CNT (R ¼ 0.35 nm)]. The order of error in Wu,
Hwang, and Huang’s (2008) atomistic-based shell theory is O[(D/R)3] (as compared
to unity), which is about 6% for the (5,5) armchair CNT. The CNTs can be
approximated by an elastic orthotropic thin shell at small strains, but the order of
error then decreases to O[(D/R)2] (i.e., larger error), which is about 16% for the (5,5)
armchair CNT. Only when the order of the error decreases to O(D/R) (largest error
in the present study), which is about 40% for the (5,5) armchair CNT, a universal
constant shell thickness and Young’s modulus can be defined, and CNTs can be
represented by an isotropic elastic thin shell.
The purpose of this chapter is to review the continuum shell modeling of
CNTs based on the interatomic potential and the atomic structure. It begins with a
review of the Brenner (1990) interatomic potential and its second-generation
 A Shell Theory for Carbon Nanotubes 7

potential (Brenner, Shenderova, Harrison, Stuart, Ni, et al., 2002) in Section 2,

though the proposed continuum shell model can be based on other interatomic
potentials (see, e.g., Pettifor & Oleinik, 1999). The proposed approach to estab-
lish a continuum shell theory from the interatomic potential is illustrated in
Section 3 for graphene, which is a single, planar layer of carbon atoms,
and therefore avoids the complexity of the curved surface and the finite CNT
radius. For such a single layer of atoms, the bending rigidity results from
the multibody atomistic interactions, rather than from tension and compression
on the opposite sides of the neutral plane as in the classical shell theory. Section 4
provides a linear shell theory for CNTs (subject to infinitesimal deformation).
Its constitutive relation is derived directly from the interatomic potential
without introducing any classical structure models (e.g., shell), nor fitting para-
meters (e.g., thickness, elastic modulus). The resulting constitutive relation is
different from the classical shell theories since it involves couplings between
the membrane stress and curvature, and between the moment and membrane
strain. Furthermore, this atomistic-based shell theory does not need any
shell thickness, nor Young’s modulus. Based on this linear theory it is discussed
in Section 5 whether the structural response of a CNT can be modeled by a classical
linear elastic shell and the corresponding order of error. A nonlinear, finite-defor-
mation shell theory is derived from the interatomic potential in Section 6. Once
again it is different from the classical shell theories since it does not involve shell
thickness, nor Young’s modulus. The instability of CNTs subject to various
loadings is studied in Section 7 (Wu, Hwang, Huang, & Song, 2008b).

2. Interatomic Potentials

Brenner (1990) established an interatomic potential, bond order for carbon

from the Tersoff (1988) formalism as

Vði; jÞ ¼ VR ðrij Þ  Bij VA ðrij Þ; (2.1)

where rij is the distance between atoms i and j (no summation for Latin
subscripts), VR and VA are the repulsive and attractive pair terms depending
only on rij
DðeÞ n pffiffiffiffiffi o
VR ðrÞ ¼ exp  2Sbðr  RðeÞ Þ fc ðrÞ; (2.2)
S1

DðeÞ S n pffiffiffiffiffiffiffiffi o
VA ðrÞ ¼ exp  2=Sbðr  RðeÞ Þ fc ðrÞ; (2.3)
S1
8 J. Wu et al.

R(e) ¼ 0.1390 nm is the equilibrium spacing between a pair of carbon atoms,

D(e) ¼ 6.00 eV is the corresponding energy stored in the atomic bond, S ¼ 1.22 is
the ratio of attractive to repulsive bond energies, and b ¼ 21 nm1 characterizes
how fast the energy decreases away from the equilibrium point,
8
> ð1Þ
>
>
18 2 39 r < R
>
<1 < =
pðr  Rð1Þ Þ5
fc ðrÞ ¼ 1 þ cos4 ð2Þ Rð1Þ  r  Rð2Þ (2.4)
>
> 2 : R  R ð1Þ ;
>
>
:
0 r > Rð2Þ

is a smooth cutoff function with R(1) ¼ 0.17 nm and R(2) ¼ 0.2 nm. For bond
lengths beyond 0.2 nm, the above cutoff function suggests vanishing bond
energy.
The multibody coupling term Bij in Eq. (2.1) depends on atoms other than i and j:
2 31=2
X
Bij ¼ 41 þ Gð’ijk Þfc ðrik Þ5 ; (2.5)
kð6¼i; jÞ

where rik is the distance between atoms i and k (k 6¼ i, j), ’ijk is angle between
bonds i–j and i–k, and
" #
c20 c20
Gð’ijk Þ ¼ a0 1 þ 2  2 (2.6)
d0 d0 þ ð1 þ cos ’ijk Þ2

with a0 ¼ 0.00020813, c0 ¼ 330, and d0 ¼ 3.5.

Figure 2.1 shows (by the solid line) the interatomic potential V versus bond
length rij for the fixed bond angle ’ijk ¼ 120 . It reaches the minimum at the bond
length r0 ¼ 0.145 nm, which is the equilibrium bond length for graphene to be
discussed in Section 3.
Recently, Brenner et al. (2002) developed the second-generation interatomic
potential for carbon. It includes both modified analytic functions for the intramole-
cular interactions and an expanded fitting database. It yields a much improved
description of bond energies, lengths, and especially force constants for carbon–
carbon bonds as compared to the earlier effort, and therefore has produced an
improved fit to the elastic properties of diamond and graphite, which in turn yield
better predictions for the energies of several surface reconstructions and intersti-
tial defects. Forces associated with rotation about dihedral angles for carbon–
carbon double bonds as well as angular interactions associated with hydrogen
centers have also been modeled which were not included in the Brenner (1990)
 A Shell Theory for Carbon Nanotubes 9

10
Brenner (1990)
Brenner et al. (2002)
Interatomic potential V (eV)

−5

−10
0.10 0.12 0.14 0.16 0.18 0.20
Bond length rij (nm)

FIG. 2.1 The interatomic potential V according to Eq. (2.1) versus the bond length rij for fixed
bond angle ’ijk ¼ 120 .

potential. The pair terms in Eqs. (2.2) and (2.3) and multibody coupling term in
Eq. (2.5) become
 
Q
VR ðrÞ ¼ 1þ A ear fc ðrÞ; (2.7)
r

X
3
VA ðrÞ ¼ Bn ebn r fc ðrÞ; (2.8)
n¼1

2 31=2
X
Bij ¼ 41 þ Gð’ijk Þfc ðrik Þ5 ; (2.9)
kð6¼ i; jÞ

where the cutoff function is the same as Eq. (2.4); A ¼ 10; 954 eV,
a ¼ 47:465 nm1 , Q ¼ 0:031346 nm, B1 ¼ 12; 389 eV, b1 ¼ 47:205 nm1 ,
B2 ¼ 17:567 eV, b2 ¼ 14:332 nm1 , B3 ¼ 30:715 eV, b3 ¼ 13:827 nm1 ; and
the function Gð’ijk Þ becomes