VOLUME 76, NUMBER 14 PHYSICAL REVIEW LETTERS 1APRIL 1996
Nanomechanics of Carbon Tubes: Instabilities beyond Linear Response
B. I. Yakobson,* C. J. Brabec, and J. Bernholc
Department of Physics, North Carolina State University, Raleigh, North Carolina 27695
(Received 22 September 1995)
Carbon nanotubes subject to large deformations reversibly switch into different morphological
patterns. Each shape change corresponds to an abrupt release of energy and a singularity in the stress-
strain curve. These transformations, simulated using a realistic many-body potential, are explained by
a continuum shell model. With properly chosen parameters, the model provides a remarkably accurate
“roadmap” of nanotube behavior beyond Hooke’s law.
PACS numbers: 61.46.+w, 02.70.Ns, 46.30.Lx, 85.40.Ux
It took the Daedalus’ “hollow molecules” [1] two
decades to become incarnated in a family of fullerenes
[2,3], remaining since then an object of keen interest:
Their growth, atomic makeup, electronic and mechanical
properties, doping, and intercalation are all being inves-
tigated. The variety of shapes is usually associated with
embedded pentagons or heptagons, which accommodate
a positive or a negative Gaussian curvature of otherwise
hexagonal graphene scrolls. Small elastic deviations from
cylindrical shape have also been observed [4]. Latest
findings in transmission electron microscopy of nanotubes
[5–7] suggest that even largely distorted configurations
can be due to elastic deformations with no atomic defects
involved. Such a suggestion is provocative for a graphite-
like material, commonly perceived as rather brittle [8],
and poses important questions: What range of strain can
nanotubes sustain elastically, and what kind of patterns
can they exhibit under different mechanical load? It is
closely related with the ultimate strength of nanotubes
and their lateral stiffness when used as quantum wires
or proximal probes. While direct experimental answers
are pending, a better control of the forces involved, a
sentient atomistic modeling can already provide useful
guidelines.
Recent development of many-body interatomic poten-
tials [9,10] made possible realistic molecular dynamics
(MD) simulations of carbon-made systems. We carried
out such studies for nanotubes under generic modes of
mechanical load: axial compression, bending, and torsion.
A remarkable synergism between the methods of MD and
those of macroscopic structural mechanics was found. A
singular behavior of the nanotube energy at certain levels
of strain corresponds to abrupt changes in morphology,
when a switch between distinct patterns occurs. These
transformations can be explained with a continuous tubule
model. With properly chosen parameters, this analytical
model predicts nanotube behavior not only at small defor-
mations but also beyond the linear response.
One of the outstanding features of fullerenes is their
hollow structure; they consist of atoms densely packed
along a closed surface that defines the overall shape.
This should also manifest itself in dynamic properties of
molecules, resembling so much the macroscopic objects
known in continuum elasticity as shells [11,12]. Macro-
scopic shells and rods have long been of interest: The
first study dates back to Euler, who discovered the elas-
tic instability. A rod subject to longitudinal compression
remains straight but shortens by some fraction ´, propor-
tional to the force, until a critical value (Euler force) is
reached. It then becomes unstable and buckles sideways
at ´.´
cr
, while the force almost does not vary [13]. For
hollow tubules there is also a possibility of local buckling,
in addition to buckling as a whole. Therefore, more than
one bifurcation can be expected.
In application to fullerenes, the theory of shells may
serve as a useful guide, but its relevance for a covalent-
bonded system of only a few atoms in diameter is far
from obvious. MD simulations are much better suited
for objects that small [14]. The discussion below pro-
ceeds therefore in two, almost parallel, paths. First, the
results of MD simulations are presented for a nanotube
under axial compression. We then briefly introduce con-
cepts of elasticity of shells and show how to adapt them
to nanotubes. MD results for other modes of load fol-
low and are compared with those suggested by the con-
tinuum model.
Figure 1 shows a simulated nanotube exposed to ax-
ial compression. The atomic interaction was modeled
by the Tersoff-Brenner potential, which reproduces the
lattice constants, binding energies, and the elastic con-
stants of graphite and diamond [10]. The end atoms
were shifted along the axis by small steps, and the whole
tube was relaxed by a conjugate-gradient method while
keeping the ends constrained. At small strains the to-
tal energy [Fig. 1(a)] grows as Es´d
1
2
E
00
´
2
, where
E
00
59 eVyatom. The presence of four singularities
at higher strains is quite a striking feature, and the pat-
terns 1(b)–1(e) illustrate the corresponding morpholog-
ical changes. The shading indicates strain energy per
atom, equally spaced from below 0.5 eV (brightest) to
above 1.5 eV (darkest). The sequence of singularities in
Es´d corresponds to a loss of molecular symmetry from
D
`h
to S
4
, D
2h
, C
2h
, and C
1
. Can this evolution be put
in the framework of continuum elasticity?
The intrinsic symmetry of a graphite sheet is hexag-
onal. Since the elastic properties of a two-dimensional
0031-9007y96y76(14)y2511(4)$10.00 © 1996 The American Physical Society 2511
VOLUME 76, NUMBER 14 PHYSICAL REVIEW LETTERS 1APRIL 1996
FIG. 1. MD-simulated nanotube of length L 6 nm, dia-
meter d 1 nm, and armchair helicity (7,7) under axial
compression. The strain energy (a) displays four singularities
corresponding to shape changes. At ´
1
0.05 the cylin-
der buckles into the pattern (b), displaying two identical
flattenings—”fins” perpendicular to each other. Further
increase of ´ enhances this pattern gradually until at
´
2
0.076 the tube switches to a three-fin pattern (c),
which still possesses a straight axis. In a buckling side-
ways at ´
3
0.09 the flattenings serve as hinges, and
only a plane of symmetry is preserved (d). At ´
4
0.13
an entirely squashed asymmetric configuration forms (e)
(see text).
hexagonal structure are isotropic [11,15], it can be approx-
imated by a uniform shell with only two elastic parame-
ters: flexural rigidity D, and its resistance to an in-plane
stretching, the in-plane stiffness C. The energy of a shell
is given by a surface integral of the quadratic form of lo-
cal deformation [11,12],
E
1
2
ZZ
Ω
Dfsk
x
1k
y
d
2
22s12ndsk
x
k
y
2k
2
xy
dg
1
C
s1 2n
2
d
fs´
x
1´
y
d
2
22s12nd
3 s´
x
´
y
2´
2
xy
dg
æ
dS , (1)
where k is the curvature variation, ´ is the in-plane
strain, and x and y are local coordinates (see Fig. 2,
inset). In order to adapt this formalism to a graphitic
tubule, the values of D and C can be identified by
comparison with the detailed ab initio and semiempirical
studies of nanotube energetics at small strains [16,17].
Indeed, the second derivative of total energy with respect
to axial strain corresponds to the in-plane rigidity C.
Similarly, the strain energy as a function of tube diameter
FIG. 2. The critical strain levels for a continuous, 1 nm wide
tube as a function of its scaled length LyN. A buckling pattern
sM, Nd is defined by the number of half waves 2M and N in the
y and x directions, respectively, e.g., a (4,4) pattern is shown
in the inset.
d corresponds to 2Dyd
2
in Eq. (1). Using the data
of Ref. [17], we obtain C 59 eVyatom 360 Jym
2
and D 0.85 eV. The Poisson ratio n 0.19 was
extracted from a reduction of the diameter of a tube
stretched in simulations. The same value is obtained from
experimental elastic constants of single crystal graphite
[8]. One can make a further step towards a more
tangible picture of a tube as having wall thickness h
and Young modulus Y . Using the standard relations
D Yh
3
y12s1 2n
2
dand C Yh [11], one finds Y
5.5 TPa and h 0.066 nm [18]. With these parameters,
linear stability analysis [11,12] allows us to assess the
nanotube behavior under strain.
Consider briefly the case of imposed axial strain. A
trial perturbation of a cylinder has a form of Fourier
harmonics, with M azimuthal lobes and N half waves
along the tube (Fig. 2, inset), i.e., sines and cosines of
arguments 2Myyd and NpxyL. At a critical level of
the imposed strain, ´
c
sM, Nd, the energy variation (1)
vanishes for this shape disturbance. The cylinder becomes
unstable and lowers its energy by assuming an sM, Nd
pattern. For tubes of d 1 nm with the shell parameters
identified above, the critical strain is shown in Fig. 2.
According to these plots, for a tube with L . 10 nm the
bifurcation is first attained for M 1, N 1. The tube
preserves its circular cross section and buckles sideways
as a whole; the critical strain is close to that for a simple
rod,
´
c
1
2
spdyLd
2
, (2)
or 4 times less for a tube with hinged edges [11]. For a
shorter tube the situation is different. The lowest critical
strain occurs for M 2 (and N $ 1, see Fig. 2), with
a few separated flattenings in directions perpendicular
to each other, while the axis remains straight. For
such a local buckling, in contrast to (2), the critical
strain depends little on length and estimates to ´
c
4
p
DyCd
21
s2y
p
3ds1 2n
2
d
21y2
hd
21
in the Lorenz
limit [19]. For a nanotube one finds
´
c
s0.077 nmdd
21
. (3)
2512
VOLUME 76, NUMBER 14 PHYSICAL REVIEW LETTERS 1APRIL 1996
Specifically, for the 1 nm wide tube of length L 6 nm,
the lowest critical strains occur for the M 2 and N
2 or 3 (Fig. 2), and are close to the value obtained in MD
simulations [Fig. 1(a)]. This is in accord with the two-
and three-fin patterns seen in Figs. 1(b) and 1(c). Higher
singularities cannot be quantified by the linear analysis,
but they look like a sideways beam buckling, which at
this stage becomes a nonuniform object.
Axially compressed tubes of greater length and/or tubes
simulated with hinged ends (equivalent to a doubled
length [11]) first buckle sideways as a whole at a strain
consistent with (2). After that, the compression at the
ends results in bending and a local buckling inward.
This illustrates the importance of the “beam-bending”
mode, the softest for a long molecule and most likely
to attain significant amplitudes due to either thermal
vibrations [20] or environmental forces. In simulations
of bending, a torque rather than force is applied at the
ends, and the bending angle u increases stepwise. While
a notch in the energy plot [Fig. 3(a)] can be mistaken
for numerical noise, its derivative dEydu unambiguously
shows an increase in tube compliance—a signature of
a buckling event [12]. In bending, only one side of a
tube is compressed and thus can buckle. Assuming that
it buckles when its local strain ´ Kdy2, where K is
the local curvature, is close to that in axial compression
[Eq. (3)], we estimate the critical curvature as
K
c
s0.155 nmdd
22
. (4)
This is in excellent agreement (within 4%) with extensive
simulations of single wall tubes of various diameters,
helicities, and lengths [6]. Due to end effects, the average
curvature is less than the local one, and the simulated
segment buckles somewhat earlier than at u
c
K
c
L,
which is accurate for longer tubes.
FIG. 3. Bending of a tubule of (13,0) helicity, 8 nm long
and 1 nm wide. (a) The strain energy curve (normalized to
its second derivative) switches from harmonic to linear at the
buckling point, while the force (dashed line) drops and remains
almost constant afterwards. (b) Beyond the buckling point, a
distinct kink shape develops.
In simulation of torsion, the increase of azimuthal angle
f between the tube ends results in the changes of energy
and morphology shown in Fig. 4. In the continuum
model, the analysis based on Eq. (1) is similar to that
outlined above, except that it involves skew harmonics
of arguments like NpxyL 6 2Myyd [12]. For overall
beam buckling sM 1d,
f
c
2s1 1ndp, (5)
and for the cylinder-helix flattening sM 2d,
f
c
s0.055 nm
3y2
dLd
25y2
. (6)
The latter should occur first for L # 136d
5y2
nm, which
is true for all the tubes we simulated. However, in
simulations it occurs later than predicted by Eq. (6). The
ends, kept circular in simulation, which is physically
justifiable by a presence of rigid caps on normally closed
ends of a molecule, deter the through flattening necessary
for the helix to form (unlike the local flattening in the case
of an axial load).
Simulations of nanotubes under mechanical duress lead
to shapes very different in appearance. At the same
time there are robust traits in common: A deformation,
proportional to the force within Hooke’s law, eventually
leads to a collapse of the cylinder and an abrupt change
in pattern, or a sequence of such events. The presence
of a snap-through buckling of nanotubes allows for a
possibility of “shape memory” when, in an unloading
cycle, the switch between patterns occurs at a somewhat
lower level of strain. A small hysteresis observed in
simulations is practically eliminated by thermal motion and
can barely be caught in experiment. More interestingly,
video imaging of the simulations at even a low temperature
(50 K) shows strongly enhanced thermal vibrations in
the vicinity of every pattern switch, while before and
after the transition only barely noticeable ripples are seen.
FIG. 4. The energy of a sawtooth (13,0) tube, 23 nm long and
1 nm wide, as a function of torsion angle f (a). At f
1
2.7
the cylinder flattens into a straight axis spiral (b). At f
2
7.2
the whole helix buckles sideways with a loss of the straight axis
(c). It coils up in a “forced” tertiary structure at f
3
8.6.
2513
VOLUME 76, NUMBER 14 PHYSICAL REVIEW LETTERS 1APRIL 1996
Physically, this indicates softening of the system, when
one of the eigenvalues tends to zero in the vicinity of the
bifurcation.
Despite the very large local strain in the nanotubes,
we did not see any bond breaking or switching. Overall,
the simulations show that graphitic layers remain highly
elastic. Conversely, this conclusion, consistent with the
recent detection of the patterns [5–7], serves as another
verification of the potential [9,10] used, which thus proves
to be realistic at large strain. This abets the study of axial
tension, where no shape transformations occur up to an
extreme dilation. Preliminary MD results (at T 50 K)
show a tube stretched by almost 40% with no damage to
its graphitic arrangement. Only at this point an atomic
disorder is seen to nucleate suddenly, the tube weakens,
and, after a pulled out long chain of atoms breaks,
separates into two fragments [21]. This stretchy, rather
than brittle, behavior for such a hard “material” indicates
extreme strength; it deserves further study, thorough in
atomic order analysis and possibly in refinements of the
potential.
An atomistic modeling of multilayer tubes remains
expensive. It makes extrapolation of the continuum
model tempting, but involves an interlayer van der Waals
interaction. The flexural rigidity scales as ,h
3
in the case
of a coherent, and as ,h for an incoherent stack of layers
[22], sliding with respect to each other when the tube is
deformed; this affects the mechanical properties and still
has to be investigated.
We did not simulate tubes under hydrostatic pressure.
At this scale, anisotropic lateral forces in a molecular
crystal packing are more plausible than a uniform pressure
[23]. Instead, one can resort to continuum elasticity and
estimate a pressure leading to an inward buckling as
p
c
ø 2Y shydd
3
[12], that is, thousands of atmospheres
for a nanometer tube. However, it drops fast with the
diameter and is assisted by a flattening effect of twisting
or bending [24] and by van der Waals attraction between
the opposite walls. Elastic collapse can thus impede the
growth of single shell tubes with d . 3 nm.
In summary, our simulations show that carbon nan-
otubes are remarkably resilient, sustaining extreme strain
with no signs of brittleness, plasticity, or atomic rear-
rangements. Their peculiar behavior beyond Hooke’s law
can be well described by a continuum model: At large
deformations an abrupt release of energy is accompanied
by a reversible switch into a different morphological pat-
tern. With properly chosen tubule parameters (n ø 0.19,
Y ø 5.5 TPa, h ø 0.66 Å), the critical deformations at
which these buckling events occur can be estimated by
Eqs. (2)–(6). Corroborated by recent experimental ob-
servations [5–7], this model provides a useful tool for
fullerene studies.
*Electronic address: yakobson@ncsu.edu
[1] D. Jones, New Sci. 32, 245 (1966); D. Jones, The
Inventions of Daedalus (Freeman, Oxford, 1982).
[2] H. Kroto, J. Heath, S. O’Brien, R. Curl, and R. Smal-
ley, Nature (London) 318, 162 (1985); W. Krätschmer,
L. Lamb, K. Fostiropoulos, and D. Huffman, Nature (Lon-
don) 347, 354 (1990); S. Iijima, Nature (London) 354,56
(1991) .
[3] Special issue edited by H. Kroto, C.N.R. Rao, and
E. Osawa, Mater. Res. Soc. Bull. 19, No. 11 (1994).
[4] R. Ruoff et al., Nature (London) 364, 514 (1993).
[5] J. Despres, E. Daguerre, and K. Lafdi, Carbon 33,87
(1995).
[6] S. Iijima, C. J. Brabec, A. Maiti, and J. Bernholc, J. Chem.
Phys. 104, (1996).
[7] N. Chopra, L. Benedict, V. Crespi, M. Cohen, S. Louie,
and A. Zettl, Nature (London) 377, 135 (1995); R. Ruoff
and D. Lorents, Bull. Am. Phys. Soc. 40, 173 (1995).
[8] B. Kelly, Physics of Graphite (Applied Science, London,
1981); M. Dresselhaus, G. Dresselhaus, K. Sugihara,
I. Spain, and H. Goldberg, Graphite Fibers and Filaments
(Springer, Berlin, 1988).
[9] J. Tersoff, Phys. Rev. B 37, 6991 (1988).
[10] D. W. Brenner, Phys. Rev. B 42, 9458 (1990).
[11] L. D. Landau and E. M. Lifshitz, Elasticity Theory (Perga-
mon, Oxford, 1986).
[12] H. Allen and P. Bulson, Background to Buckling
(McGraw-Hill, London, 1980); S. Timoshenko and
J. Gere, Theory of Elastic Stability (McGraw-Hill, New
York, 1988).
[13] R. Feynman, R. Leyton, and M. Sands, The Feynman
Lectures in Physics (Addison-Wesley, Reading, 1964),
Vol. 2.
[14] U. Landman et al., Science 248, 454 (1990).
[15] This explains why the detailed calculations [16,17] did not
reveal significant effects of helicity on elastic properties;
it also excludes the “mechanical chirality,” a different
resistance to left and right twists due to intrinsic helical
structure.
[16] G. Adams et al., Science 256, 1792 (1992); A. Lucas,
P. Lambin, and R. Smalley, J. Phys. Chem. Solids 54, 587
(1993).
[17] D. Robertson, D. Brenner, and J. Mintmire, Phys. Rev. B
45, 12 592 (1992).
[18] h appears to be less than the atomic radius, due to the
“open-work” structure of graphite manifesting itself on
this scale.
[19] R. Lorenz, Phys. Z 12, 241 (1911).
[20] G. Overney, W. Zhong, and D. Tomanek, Z. Phys. D 27,
93 (1993).
[21] Recent experiments indicate that a monoatomic chain can
also be pulled out of a nanotube by a strong electric field;
see A. Rinzler, J. Hafner, P. Nikolaev, L. Lou, S. Kim,
D. Tomanek, p. Norlander, D. Colbert, and R. Smalley,
Science 269, 1550 (1995).
[22] S. Safran, Statistical Thermodynamics of Surface, Inter-
faces and Membranes (Addison-Wesley, Reading, 1994);
D. Srolovitz et al., Phys. Rev. Lett. 74, 1779 (1995).
[23] J. Tersoff and R. Ruoff, Phys. Rev. Lett. 73, 676
(1994).
[24] A. Palmer and J. Martin, Nature (London) 254,46
(1975).
2514
