Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

INSTITUTE OF PHYSICS PUBLISHING MODELLING AND SIMULATION IN MATERIALS SCIENCE AND ENGINEERING

Modelling Simul. Mater. Sci. Eng. 15 (2007) S181–S192 doi:10.1088/0965-0393/15/1/S15

Plastic deformation in nanoscale gold single crystals


and open-celled nanoporous gold∗

Dongyun Lee1,2 , Xiaoding Wei1,2 , Manhong Zhao1,3 , Xi Chen1,3 ,


Seong C Jun2 , James Hone2 and Jeffrey W Kysar1,2,4
1 Columbia Nanomechanics Research Centre, Columbia University, 220 S. W. Mudd,

500 W 120th Street, New York, NY 10027, USA


2 Department of Mechanical Engineering, Columbia University, 220 S. W. Mudd,

500 W 120th Street, New York, NY 10027, USA


3 Department of Civil Engineering and Engineering Mechanics, Columbia University,

610 S. W. Mudd, 500 W 120th Street, New York, NY 10027, USA

E-mail: jk2079@columbia.edu

Received 23 June 2006, in final form 12 September 2006


Published 7 December 2006
Online at stacks.iop.org/MSMSE/15/S181

Abstract
The results of two sets of experiments to measure the elastic–plastic behaviour
of gold at the nanometre length scale are reported. One set of experiments was
on free-standing nanoscale single crystals of gold, and the other was on free-
standing nanoscale specimens of open-celled nanoporous gold. Both types of
specimens were fabricated from commercially available leaf which was either
pure Au or a Au/Ag alloy following by dealloying of the Ag. Mechanical testing
specimens of a ‘dog-bone’ shape were fabricated from the leaf using standard
lithographic procedures after the leaf had been glued onto a silicon wafer.
The thickness of the gauge portion of the specimens was about 100 nm, the
width between 250 nm and 300 nm and the length 7 µm. The specimens were
mechanically loaded with a nanoindenter (MTS) at the approximate midpoint of
the gauge length. The resulting force–displacement curve of the single crystal
gold was serrated and it was evident that slip localization occurred on individual
slip systems; however, the early stages of the plastic deformation occurred in
a non-localized manner. The results of detailed finite element analyses of the
specimen suggest that the critical resolved shear stress of the gold single crystal
was as high as 135 MPa which would lead to a maximum uniaxial stress of about
500 MPa after several per cent strain. The behaviour of the nanoporous gold
was substantially different. It exhibited an apparent elastic behaviour until the
point where it failed in an apparently brittle manner, although it is assumed
that plastic deformation occurred in the ligaments prior to failure. The average

∗ Presented at the IUTAM Symposium on Plasticity at the Micron Scale, Technical University of Denmark,
Copenhagen, Denmark.
4 Author to whom any correspondence should be addressed.

0965-0393/07/010181+12$30.00 © 2007 IOP Publishing Ltd Printed in the UK S181


S182 D Lee et al

elastic stiffness of three specimens was measured to be Enp = 8.8 GPa and
the stress at ultimate failure averaged 190 MPa for the three specimens tested.
Scaling arguments suggest that the stress in the individual ligaments could
approach the theoretical shear strength.
(Some figures in this article are in colour only in the electronic version)

1. Introduction

It is well-documented that the mechanical properties of metals at length scales below about
10 µm deviate from those at larger length scales. This is due to the fact that the characteristic
length scales of the various physical mechanisms which govern plastic deformation—be they
dislocations, twinning, grain boundary sliding, among others—begin to be comparable to
length scales associated with the specimen size, gradients in applied loading or gradients in
deformation state. Length scales and phenomena which can affect plastic deformation include
grain size [1–3], specimen size [4], the presence of large strain gradients [5], the presence
of a passivation layer [6], and the absence of dislocation sources, among many others [7–9].
Experimental methods to elucidate the mechanical properties of metals at small length scales
include whisker tension and torsion [5], nanoindentation [10, 11], a beam bend test [12],
micro-column compression [4], microtensile or MEMS testing [13–20], a bulge test [21], a
membrane deflection technique [22], wafer curvature [23], deflective tensile test [24], electron
back scatter diffraction (EBSD) [25–27], x-ray diffraction methods [28], flexural-based testing
methods [29], as well as in situ transmission electron microscope (TEM) [30] and SEM [31]
measurements.
In this study, we focus on the mechanical properties of free-standing nanoscale specimens
which are subjected to a combination of tensile and bending loads. As such, the results are of
fundamental interest for use in elucidating parameters in multiscale simulations. In addition,
nanoscale metal components have been employed in NEMS devices [32] so the mechanical
properties of such components have potential industrial interest. The results of two sets of
experiments are reported in this paper. The specimens are made of commercially available gold
leaf and subsequently fabricated into free-standing dog-bone specimens with thicknesses of
about 100 nm and widths between 250 and 300 nm which are anchored to a Si substrate at their
ends. One set of specimens is in single crystal form and the other set consists of an open-cellular
nanoporous gold. The specimens are mechanically loaded by deflecting the free-standing speci-
mens toward the substrate using a nanoindenter. The resulting force versus displacement curves
are then analysed to determine the mechanical properties of interest. Based on detailed finite
element analysis, the results suggest that the tensile stress at which the pure single crystals began
to yield plastically is approximately 375 MPa, which corresponds to an initial critical resolved
shear stress of about 135 MPa. In addition, the results suggest that the maximum stress in the
ligaments of the nanoporous gold approaches that of the theoretical shear stress of the material.
The details of specimen fabrication and mechanical loading are discussed in the following
section. The results of the experiments on the single crystal gold specimens are given in
section 3, and the results of the experiments on the nanoporous gold specimens are given in
section 4. Conclusions are drawn in section 5.

2. Specimen preparation and mechanical testing

Two types of commercially available leaf were used in this study. Pure gold leaf (Monarch™
24 carat, Sepp Leaf Products, Inc.) was used to fabricate the single crystal specimens and a
Plastic deformation in nanoscale gold single crystals S183

Figure 1. Fabrication processes.

gold–silver alloy leaf (Monarch™ 12 carat white gold-Au37.4 Ag62.6 at.%, Sepp Leaf Products,
Inc.) was used as the base material for the nanoporous gold. Each type of leaf had an as-received
thickness of about 150 nm as measured by atomic force microscopy (AFM). The leaf was glued
onto a (001) Si wafer with epoxy in tetrahydrofuran (C4 H8 O) solution distributed uniformly
by spin coating, followed by flattening with a smooth glass surface to obtain a final thickness
of about 100 nm. An array of dog-bone shape specimens was fabricated from the leaf using the
process illustrated in figure 1. Positive polymethylmethacrylate (PMMA) electron beam resist
was deposited on the film to allow electron beam masking. After developing the resist to expose
the portion of the film to become the specimens, a 40 nm Ni film was deposited by thermal
vapour deposition (Auto 306, BOC Edwards Inc.). The remaining resist under the Ni film
was lifted off to leave the Ni deposited directly onto the leaf to act as a mask for subsequent
processing. The leaf exception was removed by inductively coupled plasma (ICP) etching
(anisotropic Ar sputtering mode, Oxford PlasmaLab 80 Plus ICP 65) and the specimens were
suspended by isotropic CF4 /O2 reactive ion etching (RIE) (gas pressure 250 m Torr, 200 W, RF
power, TECHNICS Series 800 RIE) of the underlying silicon substrate followed by removal
of the Ni film by wet etching. At this point, the specimens were anchored to the underlying
substrate at the two ends, but were otherwise free-standing. One additional and final step in
the fabrication of the nanoporous gold was to dealloy the suspended specimens in ACS grade
70% concentrated nitric acid (approximately 16 mols) for 45 min at ambient temperature in
order to remove the silver from the alloy. In the dealloying process, the silver reacts with the
nitric acid and is thus removed from the alloy. The surface diffusion of the remaining gold is
enhanced which allows the formation of open cellular nanoporous gold [33–35].
The grain structure of the pure gold leaf was measured by EBSD prior to the fabrication.
The results indicated an in-plane grain size of several tens of micrometres and a strong [001]
texture, as in figures 2(a) and (b). Since the overall length of each specimen was approximately
18 µm, it proved possible to selectively fabricate the pure gold specimens within individual
grains of the leaf so that the specimens consisted of single crystals. An example of an array
of specimens is shown in figures 2(c) and (d). The gauge length of each specimen was 7 µm.
The width of the single crystal gold specimens was 250 nm and the width of the nanoporous
gold specimens was 300 nm. The final thickness of the specimens was determined as the
average value of at least ten measurements via the SEM. For the single crystal gold specimens
the average thickness was 105.5 ± 9.9 nm and for the nanoporous gold specimens the average
thickness was 99.8±7.1 nm. It should be noted that since the leaf was adhered to the substrate at
S184 D Lee et al

Figure 2. Preparation of specimens: (a) EBSD of film; (b) pole figure showing texture; (c) array
of specimens; (d) specimens.

ambient temperature and the deformation experiments were performed at ambient temperature
there was expected to be very little residual stress in the specimens. The initial dislocation
structure of the nanoscale gold single crystals will be measured via TEM in future studies.
The deformation tests were carried out by deflecting the centre of the suspended bridge-like
specimens using a nanoindenter (Nanovision™ , MTS Systems Corporation, Oak Ridge, TN)
with a dynamic contact module (DCM). The force used to actuate the system was generated
via a magnet and coil assembly. The general features for measuring displacement and load
are described in detail elsewhere [10]. A Berkovich indenter was used for this study. To
perform the test, two adjacent specimens were first located using the nanoindenter in a contact
scanning mode (Nanovision™ ). In the process the two specimens were destroyed but the
positions of registry marks on the specimens were used to determine the positions of the
remaining specimens. Once the tip came into contact with an adjacent undeformed specimen,
the stiffness of the gauge section was sampled at several locations across its width in 17 nm
increments after which the tip was returned to the position of highest stiffness which was
presumed to be the specimen centre. To load the specimens, the nanoindenter was operated
under position control (30 nm s−1 ) to push the specimen toward the substrate until the specimen
was either broken or touched the substrate, and the resulting force versus displacement data
were obtained. Three specimens each of the single crystal gold specimens and the nanoporous
gold specimens were tested. In addition to the deflection tests, nanoindentation experiments
were also performed on nanoporous Au while it still adhered to the substrate. It is estimated
that the tip of the nanoindenter had a radius of curvature of about 300 nm.

3. Results of experiments on gold single crystals

Figure 3(a) shows single crystal specimen 2 after the loading. The indenter tip lay along the
longitudinal centreline of the specimen as a result of the stiffness measurements; however
the point of contact of the indenter was approximately 0.5 µm from the midpoint between
the two anchors. It is apparent that plastic deformation occurred heterogeneously, but in
Plastic deformation in nanoscale gold single crystals S185

Figure 3. Deformed singe crystal specimen 2: (a) global view; (b) close-up view of slip localization;
(c) plan view of slip localization; (d) force versus displacment data for three specimens.

an initially non-localized manner, along the gauge length of the specimen. Views at higher
magnification in figures 3(b) and (c) show the presence of slip localization on a single slip
system characteristic of plastic deformation in single crystals. The force–displacement curves
for the three free-standing single crystal gold specimens are shown in figure 3(d). In the elastic
range, the magnitude of the force is expected to scale with the cube of displacement once the
displacement of the bridge becomes comparable to its thickness [36]. It can be seen that the
initial response of all the tested specimens was similar until a load of about 1 µN had been
achieved after which sample 1 suffered a slight load drop; the corresponding displacement
at this load was equal to about two specimen thicknesses. It is assumed that the load drop
corresponded to the onset of plastic deformation in that specimen, especially since the force
increased at a slower rate than the other two specimens upon further loading. The force–
displacement curves for the other two specimens closely followed each other until a load of
about 1.7 µN when the slope of the curve for specimen 2 began to decrease which presumably
signaled the initiation of plastic deformation in specimen 2. Specimen 3 experienced a sudden
load drop at about 3.5 µN. In addition to a monotonic loading, sample 2 was unloaded after
achieving a load of 2.5 µN. It is assumed that the plastic deformation up to this point occurred
in a non-localized manner as has been seen in submicron diameter compression specimens [4].
Upon reloading, the specimen experienced a sudden load drop at about 3.5 µN which was
likely precipitated by the formation of the localized slip shown in figure 3(b). In addition it is
likely that the slip localization did not occur until sufficiently large deflection had occurred so
that the bending character of the deformation was significantly less important than the tensile
character of the deformation.
S186 D Lee et al

Figure 4. Finite element simulations of experiments: (a) overall mesh with rigid indenter; (b) close-
up of deformed mesh showing total slip contours; (c) comparison of simulations to experiment; (d)
predicted uniaxial behaviour.

Since the plastic slip occurred in a non-localized manner in the early stages of the
deformation a continuum analysis of the deformation is warranted, with the goal of determining
the overall stress and strain response of the material in order to ascertain the approximate
mechanical properties of the specimens. Numerical simulations were performed using the
commercial finite element code ABAQUS [37] with a user-material subroutine incorporating
single crystal plasticity constitutive relations which account for the kinematics of finite
deformation and lattice rotation associated with single crystal plasiticity [38]. The dimensions
of the single crystal—here called a beam—in the analysis were 7000 nm in length, 250 nm
in width, and 100 nm in thickness. An indenter, modelled as a rigid contact surface, was
positioned on the beam at a point 3000 nm away from one end of the gauge section of the
beam. The finite element mesh is shown in figure 4(a) and consisted of 28320 8-node linear
hybrid elements with reduced integration and hourglass control [37]. The indenter had a conical
shape with a half-apex angle α = 70.3◦ , so that the ratio of cross-sectional area to depth was
the same as the Berkovich indenter. Additionally, the indenter tip was rounded to a radius
of 60 nm; however, the results were insensitive to the precise radius chosen. The Coulomb
friction law was used between contacting surfaces, and the friction coefficient was taken to be
0.15. The elastic moduli of the face-centred cubic gold were C11 = 186 GPa, C12 = 157 GPa
and C44 = 42.0 GPa and the plastic slip systems were {1 1 1}1 1 0 [39]. The crystallographic
orientation of the surface normal (parallel to the direction of motion of the indenter) was taken
to be [0 0 1] based upon the EBSD analysis. The crystallographic orientation along the gauge
length of the specimen was taken to be 1̄ 5.67 0 which is rotated approximately 10◦ from
[0 1 0], consistent with the EBSD results and also with the orientation of the slip localization
{1 1 1} slip plane in figure 3(b). The constitutive relation was considered to be rate-independent
Plastic deformation in nanoscale gold single crystals S187

in the limit of rate-dependent viscoplasticity [40]. Specifically, the rate of slip, γ̇ (α) , of slip
system α was determined as
 
 (α)   τ (α) n
γ̇ = ȧ sgn τ
(α) (α)   (1)
 g (α)  ,

where τ (α) is the resolved shear stress, ȧ (α) is the reference strain rate, g (α) is the current strength
of slip system α, and n isthe rate sensitivity exponent. The rate of increase in strength of the
slip systems was ġ (α) = β hαβ γ̇ (β) , where the hαβ contains the so-called hardening moduli.
The initial values of g (α) were the same for all the slip systems. A simple phenomenological
functional form for the hardening moduli was given by [41]
 

2  h0 γ 

hαα = h(γ ) = h0 sec h  (no sum on α), (2)
τ −τ  s 0
where h0 is the initial hardening modulus, τ0 the yield stress which equals the initial value of
current strength g (α)
, τsthe stage 1 stress (or the break-through stress where large plastic flow
t
initiates), and γ = α 0 |γ̇ (α) | dt, where t is the elapsed time of the simulation. In the present
simulation, ȧ (α) = 10−3 s−1 and n = 50. It was assumed that no latent hardening occurred.
The contours of total slip in the specimen near the indented region are shown in figure 4(b).
A set of parameters given by h0 = 2 GPa, τ0 = 134.2 MPa and τs = 189.0 MPa was found to
closely approximate the mechanical response of specimen 2, as seen in figure 4(c). Specimens
1 and 3 were also simulated in an attempt to determine appropriate constitutive parameters;
however since these specimens were not unloaded during the experiments a very broad range
of parameters could be used to simulate the experiments with a high degree of accuracy.
Therefore, these parameters are not reported.
In addition to the simulation of the deflected single crystal, another simulation was
performed to predict the response of the crystal loaded in uniaxial tension in the 1̄ 5.67 0
direction using the parameters above. The mesh used in the simulation was ten times longer
than its width to minimize end effects, but otherwise all procedures and elements were the
same. The results are shown in figure 4(d). Based upon the parameters chosen (which are
non-unique) the results suggest that the tensile yield stress was approximately 375 MPa, and
due to the physical interpretation of the variable τ0 , the initial critical resolved shear stress of
the pure gold single crystal was about 135 MPa. Another simulation with the uniaxial tension
in the 0 1 0 direction was very similar except that the magnitudes of the stresses were about
10% lower.

4. Results of experiments on nanoporous gold

The results of the experiments on the nanoporous gold are now discussed. The morphology of
the nanoporous gold after dealloying is shown in figure 5(a). The ligaments had a characteristic
cross-section size of 20–40 nm, similar to that of other studies [33–35,42,43]. Energy
dispersive x-ray (EDX) analysis confirmed that negligible silver was present. Therefore, the
nanoporous Au fabricated from white gold leaf had a relative density of 35.2% compared with
that of monolithic gold.
Three specimens were loaded following the same procedure as that for the single crystal
gold specimens, except that the indenter was in the middle of the gauge length. Specimen 2 is
shown in its deformed configuration in figure 5(b). The force–displacement curves of the three
free-standing nanoporous gold specimens are shown in figure 5(c). A sudden drop in load for
each of the three specimens signaled the onset of deformation localization which ultimately
led to failure across the width of the specimen. The elastic modulus of the nanoporous Au
S188 D Lee et al

Figure 5. Experiments on nanoporous gold: (a) open-celled structure after dealloying; (b) sample
2 after deformation; (c) force versus displacement data for three specimens; (d) curve fit relative
to experiment for sample 2.

film was determined by comparing the force–displacement data to an analytical model of a


centre-loaded double cantilever beam for which sliding at the anchor-points is prohibited. The
model which assumes a linear elastic stress–strain constitutive behaviour, yet accounts for
finite displacement and residual stress, is an approximate solution based upon the principle
of minimum potential energy [44, 45]. Accordingly, the force, F , necessary to cause an
elastic vertical displacement, δ, at the centre of the gauge section of the suspended film is
approximately
     4 3    4 3  3
F π2 σ0 h δ π h δ π h δ
= + + , (3)
Ewh 2 E L h 6 L h 8 L h

where L is the gauge length of specimen, w is the width of specimen, h is thickness of the film,
σ0 is the residual stress and E is elastic modulus. The parameters E and σ0 in equation (3)
are fitted to the experimental data to determine values for elastic modulus and residual stress.
Figure 5(d) shows the data and the fitted curve for specimen 2. The resulting values are:
specimen 1, Enp = 8.99 GPA, σ0 = 65.0 MPa; specimen 2, Enp = 8.33 GPa, σ0 = 52.9 MPa;
and specimen 3, Enp = 9.08 GPa, σ0 = 77.7 MPa, which gives an average of Enp = 8.8 GPa
and σ0 = 65.2 MPa, where Enp refers to elastic modulus of the nanoporous gold film.
The results of the deflection experiments can be evaluated by comparing them to scaling
laws. We adopt Gibson and Ashby’s analysis [46] of the physical and mechanical properties
of open-cell forms to the nanoporous gold for which the mass density of the nanoporous gold
Plastic deformation in nanoscale gold single crystals S189

Figure 6. Indentation on nanoporous film: (a) experimental results with bounds from simulation;
(b) example of deformed mesh showing void volume fraction.

is expected to scale as
 2
ρnp t
= C1 , (4)
ρs l
where ρnp and ρs are the mass density of the nanoporous material and the mass density of the
solid material, respectively, t is ligament thickness and l is length of the ligaments. C1 is a
geometric constant that is approximately unity. Average values of t and l for the nanoporous
gold were determined using a standard metallographic method [47] via SEM micrographs of
nanoporous gold to obtain t ∼ = 32 nm and l ∼ = 53.4 nm, which yields a relative density of
35.9%; this value agrees well with the relative density of gold in the alloy prior to dealloying,
which is 35.2%.
According to Gibson and Ashby [46], the elastic modulus for a porous material scales as
 
Enp ρnp 2
= C2 , (5)
Es ρs
where Es is the elastic modulus of the corresponding monolithic material and C2 is a constant
which is approximately unity for most materials. The elastic modulus of bulk gold [47], Es
is about 80 GPa, so the predicted elastic modulus of the nanoporous gold film, Enp , is in the
range of 9.91 to 10.31 GPa depending on relative densities. This value is in good agreement
with the value measured by the deflection test.
The stress in the specimen which corresponds to the sudden load drop was also determined
by converting the force–displacement data obtained from the deflection test to stress–strain
data along the neutral axis of the specimen using the approximate solution based upon the
principle of minimum potential energy; the average value obtained for these specimens was
about 190 MPa. This should be considered the ‘ultimate’ stress at which the weakest set of
ligaments across the section of the specimen was strained to its point of failure, while the
deformation outside the region of strain localization remained elastic.
Nanoindentation experiments were also performed on a nanoporous gold film while it
adhered to the substrate; the process steps to obtain the film were identical to that described
above except that the RIE etching to suspend the films was not performed. The results are
shown as thin solid lines in figure 6(a). The elastic modulus was estimated using the data to a
depth of 20 nm by continuous stiffness measurement and standard analysis methods [48]. The
average of more than 20 indentations gives an elastic modulus of Enp = 13.2 ± 3.4 GPa, which
is slightly higher than the values measured by the deflection test and other studies [49, 50],
S190 D Lee et al

which could be due to the presence of the underlying substrate as well as the large ratio of the
indenter tip radius to film thickness.
The stress at which irreversible deformation initiates, here called the yield stress σy , was
estimated by nanoindentation on nanoporous gold films bonded to a silicon substrate. The
data to the depth of 100 nm were analysed using recently proposed finite element techniques
[51–53], which incorporate both the substrate effect and the densification effect for porous
materials. The effect of the substrate on the data was removed based upon a substrate effect
map [51] by assuming that the substrate (Si) was much harder and stiffer than the film. Detailed
finite element analyses of nanoindentation on a porous media were then carried out at fixed
porosity, employing the film elastic modulus, residual stress and film thickness values obtained
from the experiment. The plastic properties of the nanoporous gold films were varied in
the simulations, including the yield stress, σy , and work hardening coefficient, n, which is
typically about 0.2 to 0.25 for gold [54]. Finally, by fitting to the upper and lower bounds of
the experimental data, the yield stress of the nanoporous gold films was estimated from reverse
analyses of the finite element simulations to obtain the range σy = 76–111 MPa for n = 0.25
and σy = 110–146 MPa for n = 0.2 which bound the experimental results in figure 6(a).
The average yield stress of these values is σy = 111 ± 35 MPa. Figure 6(b) shows contours
of the void volume fraction during the simulation. Biener and co-workers [49, 50] reported
similar values for mechanical properties of nanoporous gold based on the nanoindentation test,
although they focused on a bulk specimen with ligaments significantly larger than those in the
present study.
The yield stress for a porous material is expected to scale as [46]
   
σy ρnp 3/2 ρnp 1/2
= 0.23 1+ , (6)
σs ρs ρs
where σy and σs are the yield stress of nanoporous gold and solid gold in the ligaments,
respectively. Using the value of ρnp /ρs = 0.352 and σy = 111 MPa leads to an estimated value
of the yield stress of the individual ligaments of the nanoporous gold to be σs = 1.45 GPa.
This implies that the yield stress of the ligaments of nanoporous gold approaches the intrinsic
yield stress of gold which is on the order of 1.5–8 GPa [55–59].
The deflection tests of the free-standing film and the nanoindentation tests of the film
bonded to the substrate lead to different stress values. The nanoindentation experiments
estimate the stress at which plastic, or irreversible, deformation first occurs to be about 111 MPa.
This stress is defined as the force per unit area of the nanoporous medium, where the area
includes both ligaments and cells. This value was then used to estimate the stress on the
ligaments (defined as force per unit area of the ligament) at initial yield which can be as high
as 1.45 GPa. The deflection test, on the other hand, measures the stress at the final stages
of localized deformation at the point where the plastically deforming ligaments begin to fail.
This ultimate stress is defined as the force per unit area of the nanoporous medium and is on
average 190 MPa for the three specimens tested. This suggests that the initial yield occurred
at a stress of 111 MPa and that the ligaments hardened with plastic deformation up to an
ultimate stress level of 190 MPa at the point of failure. According to the values of the assumed
strain hardening coefficients, this would indicate that on average a 10% to 15% plastic strain
occurred in the ligaments prior to ultimate failure. However, this interpretation assumes that
the localized deformation occurs uniformly across the width of the specimen, which cannot
be verified with the current experimental set-up.
It is of interest to understand how gold can attain such high yield stress in its pure state.
Recent experiments and analyses by Nix and co-workers [56,60], suggest that a concept termed
dislocation starvation may lead to a yield stress of up to 1 GPa. However, the ligaments of the
Plastic deformation in nanoscale gold single crystals S191

nanoporous gold studied herein are much smaller than the specimens used in those studies,
so the mechanisms which lead to the high yield stress in the nanoporous metal are as yet
undetermined.

5. Conclusions

In conclusion, free-standing nanoscale specimens of pure single crystal gold and also
nanoporous gold were fabricated and the mechanical properties were probed both with a
bridge deflection technique and by nanoindentation.
The data and subsequent detailed finite element analysis for the single crystal specimens
suggest that the uniaxial tensile behaviour in the 1̄ 5.67 0 direction of a crystal with cross-
sectional dimensions of 100 nm by 250 nm reached a value of almost 500 MPa after several
per cent strain. Previous uniaxial compression experiments [56, 60] on gold single crystals
in the 0 1 0 direction (which is oriented 10◦ from the specimens in the present experiments)
with a 400 nm diameter found similar values. Also, similar studies on Ni3 Al exhibit flow
stresses of approximately 250 MPa [4]. In addition, it is interesting to note that the force–
displacement data of the other two specimens tested in this experiment have different values
of stress at which plastic deformation began, which suggests that the plastic behaviour at these
small length scales can be highly sensitive to slight variations in specimen geometry or loading
conditions, as seen in other studies [4].
The data and subsequent analysis for the nanoporous specimens suggest that the elastic
modulus of the nanoporous gold is about 8.8 GPa, the initial yield stress of the nanoporous gold
is about 111 MPa and the ultimate stress at failure due to deformation localization is as high
as 190 MPa. According to scaling laws, this would lead to a yield stress within an individual
ligament as high as 1.45 GPa, which approaches the theoretical shear strength of the material
at which adjacent atomic planes slide past one another without the mediating influence of
dislocations.

Acknowledgments

Support from AFOSR FA5550-06-1-0214, NSF CMS-0134226 and NSF CMS-0407743 and
the Academic Quality Fund from Columbia University is gratefully acknowledged. This work
has used the shared experimental facilities that are supported primarily by the MRSEC Program
of the National Science Foundation under Award Number DMR-0213574 and by the New York
State Office of Science, Technology and Academic Research (NYSTAR). Shared experimental
facilities of NSEC Program of the National Science Foundation under NSF CHE-0117752 are
also gratefully acknowledged.

References

[1] Hall E O 1951 Proc. Phys. Soc. B 64 747–53


[2] Petch N J 1953 J. Iron Steel Inst. 174 25–8
[3] Schiotz J and Jacobsen K W 2003 Science 301 1357–9
[4] Uchic M D, Dimiduk D M, Florando J N and Nix W D 2004 Science 305 986–9
[5] Fleck N A, Muller G M, Ashby M F and Hutchinson J W 1994 Acta Metall. Mater. 42 475–87
[6] Dehm G, Balk T J, Edongue H and Arzt E 2003 Microelectron. Eng. 70 412–24
[7] Kraft O, Freund L B, Phillips R and Arzt E 2002 MRS Bull. 27 30–7
[8] Arzt E, Dehm G, Gumbsch P, Kraft O and Weiss D 2001 Prog. Mater. Sci. 46 283–307
[9] Arzt E 1998 Acta Mater. 46 5611–26
[10] Oliver W C and Pharr G M 2004 J. Mater. Res. 19 3–20
S192 D Lee et al

[11] Pharr G M and Oliver W C 1992 Mrs Bull. 17 28–33


[12] Stölken J S and Evans A G 1998 Acta Mater. 46 5109–15
[13] Haque M A and Saif M T A 2003 Exp. Mech. 43 248–55
[14] Kraft O and Volkert C A 2001 Adv. Eng. Mater. 3 99–110
[15] Read D T and Dally J W 1993 J. Mater. Res. 8 1542–9
[16] Ruud J A, Josell D, Spaepen F and Greer A L 1993 J. Mater. Res. 8 112–17
[17] Kahn H, Chen L, Ballarini R and Heuer A H 2006 Acta Mater. 54 667–78
[18] Gianola D S and Sharpe W 2004 Exp. Tech. 28 23–7
[19] Chasiotis I and Knauss W G 2003 J. Mech. Phys. Solids 51 1533–50
[20] Wang Y M, Wang K, Pan D, Lu K, Hemker K J and Ma E 2003 Scr. Mater. 48 1581–6
[21] Vlassak J J and Nix W D 1992 J. Mater. Res. 7 3242–9
[22] Maner K C, Begley M R and Oliver W C 2004 Acta Mater. 52 5451–60
[23] Nix W D 1989 Metall. Trans. A (Phys. Metall. Mater. Sci.) 20 2217–45
[24] Espinosa H D, Prorok B C and Fischer M 2003 J. Mech. Phys. Solids 51 47–67
[25] Schwartz A J, King W E, Campbell G H, Stolken J S, Lassila D H, Sun S and Adams B L 1999 J. Eng. Mater.
Technol.—Trans. ASME 121 178–81
[26] Chen H Q, Kysar J W and Yao Y L 2004 J. Appl. Mech.—Trans. ASME 71 713–23
[27] Sun S, Adams B L and King W E 2000 Phil. Mag. A (Phys. Condens. Matter Struct. Defects Mech. Prop.) 80
9–25
[28] Larson B C, Yang W, Tischler J Z, Ice G E, Budai J D, Liu W and Weiland H 2004 Int. J. Plast. 20 543–60
[29] Gludlavalleti S, Gearing B R and Anand L 2005 Exp. Mech. 45 412–9
[30] Chen M W, Ma E, Hemker K J, Sheng H W, Wang Y M and Cheng X M 2003 Science 300 1275–7
[31] Chen D J and Chiang F P 1993 Appl. Opt. 32 225–36
[32] Huang X M H, Manolidis M, Jun S C and Hone J 2005 Appl. Phys. Lett. 86 143104
[33] Erlebacher J 2004 J. Electrochem. Soc. 151 C614–26
[34] Erlebacher J, Aziz M J, Karma A, Dimitrov N and Sieradzki K 2001 Nature 410 450–53
[35] Erlebacher J and Sieradzki K 2003 Scr. Mater. 49 991–6
[36] Lee D, Wei X, Chen X, Zhao M, Jun S C, Hone J, Herbert E K, Oliver W C and Kysar J W 2006 Scr. Mater. at
press
[37] ABAQUS 2004 ABAQUS 6.4 User’s Manual (Pawtucket, RI: ABAQUS Inc.)
[38] Huang Y 1991 Division of Applied Science (Cambridge, MA: Harvard University) pp 1–46
[39] Hirth J P and Lothe J 1992 Theory of Dislocations 2nd edn (Malabar, Fl: Krieger Publishing Company)
[40] Hutchinson J W 1976 Proc. R. Soc. Lond. Ser. A (Math. Phys. Eng. Sci.) 348 101–27
[41] Peirce D, Asaro R J and Needleman A 1982 Acta Metall. 30 1087–119
[42] Ding Y and Erlebacher J 2003 J. Am. Chem. Soc. 125 7772–3
[43] Ding Y, Kim Y J and Erlebacher J 2004 Adv. Mater. 16 1897–1900
[44] Senturia S D 2001 Microsys. Des. (Boston: Kluwer Academic Publishers)
[45] Timoshenko S P and Goodier J N 1970 Theory of Elasticity 3rd edn (NewYork: McGraw-Hill)
[46] Gibson L J and Ashby M F 1997 Cellular Solids: Structure and Properties 2nd edn (Cambridge: Cambridge
University press)
[47] 1998 Metals Handbook 2nd edn (Cleveland: ASM International)
[48] Oliver W C and Pharr G M 1992 J. Mater. Res. 7 1564–83
[49] Biener J, Hodge A M and Hamza A V 2005 Appl. Phys. Lett. 87 121908
[50] Biener J, Hodge A M, Hamza A V, Hsiung L M and Satcher J H 2005 J. Appl. Phys. 97
[51] Chen X and Vlassak J J 2001 J. Mater. Res. 16 2974–82
[52] Chen X, Xiang Y and Vlassak J J 2006 J. Mater. Res. 21 715–24
[53] Xiang Y, Chen X, Tsui T Y, Jang J I and Vlassak J J 2006 J. Mater. Res. 21 386–95
[54] Ashby M F and Jones D R H 1996 Engineering. Materials. I 2nd edn (Oxford: Butterworth Heinemann)
[55] Agrait N, Rubio G and Vieira S 1995 Phys. Rev. Lett. 74 3995–8
[56] Greer J R, Oliver W C and Nix W D 2005 Acta Mater. 53 1821–30
[57] Kiely E D and Houston J E 1998 Phys. Rev. B 57 12588–94
[58] Stalder A and Durig U 1996 Appl. Phys. Lett. 68 637–9
[59] Wu B, Heidelberg A and Boland J J 2005 Nature Mater. 4 525–9
[60] Greer J R, Oliver W C and Nix W D 2006 Acta Mater. 54 1705

You might also like