LETTERS

Mechanical properties of ultrahigh-strength

gold nanowires
BIN WU1*, ANDREAS HEIDELBERG1,2* AND JOHN J. BOLAND1†
1
Centre for Research on Adaptive Nanostructures and Nanodevices (CRANN) and the Department of Chemistry, Trinity College Dublin, Dublin 2, Ireland
2
AGEF e.V.-Institut an der Heinrich-Heine-Universität Düsseldorf, D-40225 Düsseldorf, Germany
*These authors contributed equally to this work
†
e-mail: [email protected]

Published online: 5 June 2005; doi:10.1038/nmat1403

anowires have attracted considerable interest as nanoscale

N
a
interconnects and as the active components of both Lateral force signal
electronic and electromechanical devices. Nanomechanical
measurements are a challenge, but remain key to the development
Laser
and processing of novel nanowire-based devices. Here, we report a
general method to measure the spectrum of nanowire mechanical
properties based on nanowire bending under the lateral load from an
Photodiode
atomic force microscope tip. We find that for Au nanowires, Young’s
modulus is essentially independent of diameter, whereas the yield
strength is largest for the smallest diameter wires, with strengths
up to 100 times that of bulk materials, and substantially larger than
that reported for bulk nanocrystalline metals (BNMs)1–5. In contrast
to BNMs, nanowire plasticity is characterized by strain-hardening, b c
demonstrating that dislocation motion and pile-up is still operative
down to diameters of 40 nm. Possible origins for the different
mechanical properties of nanowires and BNMs are discussed.
Although the size dependence of mechanical properties is well
documented, the ability to structure materials at the nanometre
scale has generated renewed interest in this area. Early studies
showed that micrometre-sized whiskers have yield strengths
that are over ten times that of the corresponding bulk materials6.
Other measurements on BNMs revealed remarkably different
properties when compared with their coarse-grain counterparts1–5.
Nanocrystalline Cu exhibits high yield strengths (approaching
0.4 GPa) and near-perfect elastoplasticity without work hardening
or necking1, and room-temperature superplasticity2, whereas Figure 1 The bending test for nanowire mechanical measurements.
nanocrystalline nickel shows reversible X-ray-diffraction peak a, Schematic of fixed wire in a lateral bending test with an AFM tip. b, SEM image
broadening during plastic deformation3. In contrast, much less is of a 200-nm Au nanowire suspended on a trench. c, SEM image of a Au nanowire
known about mechanical properties of free-standing nanoscale mechanically fixed by electron-beam-induced deposition of Pt lines. The scale bars
objects such as nanowires due to difficulties associated with standard on both images are 500 nm.
tensile or bending tests. Several approaches have been reported and
include the use of a nanostressing stage within scanning electron
microscope (SEM)7, transmission electron microscope (TEM)8,9
and atomic force microscope (AFM)10–19 measurements. The latter positioning the wires across pores on substrates11–14. In general, these
has attracted particular attention due to its high spatial resolution methods suffer from complications due to wire–substrate friction.
and force-sensing capabilities. To date, elastic force–displacement In the case of pinned wires, the flexure lengths are uncertain due
(F–d) data have been reported for a variety of nanowire systems (SiC to leakage of pinning materials in the shadow-mask process10.
nanowires, carbon nanotubes, polypyrrole and WS2 nanotubes), and Measurements of yield and ultimate strengths are even more difficult
involve either pinning one end of these wires to the substrate10 or and involve pulling7,20,21 or bending experiments10, and it is often

nature materials | VOL 4 | JULY 2005 | www.nature.com/naturematerials 525

©2005 Nature Publishing Group

LETTERS
a b from Young’s modulus E, yield strength, plastic deformation and
failure. We show that whereas E is essentially diameter-independent
and comparable to that of bulk gold, yield strengths are up to 100
Loading Unloading times larger than bulk, with ultimate strengths approaching the
theoretical value of E/10 for smaller-diameter nanowires22. Gold
1 nanowires exhibit well-defined yield points and undergo strain-
Baseline
hardening characterized by a universal F–d curve that is traced
out by successive yield points during repeated loading–unloading
Force

cycles. Nanowire yielding and plastic deformation properties are

2
dramatically different from those of BNMs with comparable grain
size, suggesting that the constrained material volume and limited
Forward number of grains across the nanowire diameter preclude the elasto-
Reverse plasticity behaviour reported in BNMs1.
3 4,000 nN This method involves AFM lateral manipulation of nanowires
4,000 nN that are mechanically fixed to and positioned over trenches on a SiO2
substrate (Fig. 1a). Trenches were fabricated by focused-ion-beam
0 200 400 0 200 400 600
milling and were 0.3–4 µm wide and 250–500 nm deep. Nanowires
Manipulation path (nm) Manipulation path (nm)
ranging in diameter from 40 nm to 250 nm were synthesized
electrochemically in alumina templates, released and dispersed
onto the substrate. TEM investigations of 40-nm Au nanowires
c d reveal that individual grains span the diameter and are separated
from adjacent grains by low-angle boundaries (See Supplementary
Information, Fig. S1). The grain structure of 200-nm nanowires
could not be investigated due to electron-beam attenuation. Wires
that were found to span trenches in an orthogonal configuration
were subsequently pinned down at the trench edges by electron-
beam-induced deposition of Pt lines (Fig. 1b and 1c)23. In this
manner, wire–substrate friction and difficulties associated with the
precise control of the pinning points were eliminated. The resulting
structures were then subjected to lateral loading by an AFM
cantilever under closed-loop X–Y control. The AFM tip was swept in
e f a programmed manipulation path in the XY plane with the axis of the
cantilever aligned parallel to the nanowire (see Fig. 1). With the tip
positioned 100–300 nm below the plane of the substrate, the lateral
deflection of the cantilever was monitored during the manipulation
of the tip using a four-quadrant photodiode and the corresponding
F–d trace was recorded. The normal force component was recorded
simultaneously but was less than 3% of the lateral force and does not
contribute significantly to the mechanical deformation of the wire.
Before each manipulation, a tapping-mode image of the
suspended nanowire was recorded, the tip oscillation was then
switched off and a trajectory was defined in which the tip approached,
contacted and reversed back from the wire. Typical tip velocities
Figure 2 Mechanical deformation of a 200-nm Au nanowire. a, F–d curves
were 50 nm s–1 and the forward and reverse paths were always
recorded during manipulation without and with contact between the AFM tip and
coincident, perpendicular to the wire axis and intersected the mid-
the bottom of the trench. The presence of contact is evident from the baseline
point of the wire. Figure 2 shows a typical set of F–d curves for a
curves and the shape of the F–d curve. b, F–d curves recorded during consecutive
200-nm Au nanowire and the corresponding AFM images recorded
manipulation by AFM tip-induced lateral bending of a 200-nm Au nanowire. The
during the bending test (Supplementary Information Fig. S2 shows
cantilever’s turning points, that is, reversal of manipulation direction, are identified
manipulations and F–d data for a 58-nm Au nanowire). Figure 2a
by the vertical dashed lines. The scale of the force is given by the vertical bar in
shows that during the approach the lateral force is essentially zero
the two plots. c, AFM image taken after elastic manipulation (F–d curve labelled
until contact is established with the wire, at which point the force
1) of a 200-nm Au nanowire. d, e, AFM images obtained after successive plastic
increases linearly as the wire is loaded, reflecting the elastic response
manipulations (F–d curves labelled 2 and 3). f, SEM image of a Au nanowire
of the wire–cantilever system. When the tip reverses direction, the
following the bending test agrees in detail with the AFM image shown in e. All scale
wire unloads yielding a symmetric F–d curve. The baseline curve
bars are 1 µm. See text for additional details.
measured away from the wire location demonstrates that the F–d
data is entirely due to mechanical deformation of the wire. However,
as we do not have closed-loop control in the z direction whenever
the tip is positioned below the wire, its actual position is not well
difficult to establish whether the measured failure is due to the wire controlled and contact with the bottom of the trench is possible. In
itself or the contact between the wire and AFM tip or substrate. The the case of tip contact with trench bottom, friction forces induce a
methods available at present are not capable of providing reliable lateral deflection of the cantilever. When the manipulation direction
measures of nanowire strength or their behaviour during plastic is switched from forward to reverse the cantilever bends in the
deformation and failure. opposite direction resulting in a jump in the force trace (Fig. 2a),
Here we report an AFM-based method that unambiguously and this shift in the baseline results in a asymmetric F–d curve. Note
measures the full spectrum of mechanical properties; ranging that inadvertent contact with the bottom of the trench does not
526 nature materials | VOL 4 | JULY 2005 | www.nature.com/naturematerials

©2005 Nature Publishing Group

 LETTERS
a elongation of the wire. Increased loading eventually resulted in
140
extended plastic deformation and ultimately failure of the wire
120
(images in Fig. 2e and f). Note that the wire in AFM image Fig. 2e
and SEM image Fig. 2f is actually broken, but has sprung back on
100 itself after the tip was removed. To confirm that failure is due to the
Young's modulus (GPa)

Bulk gold mechanical properties of the wire and not the interaction with the
80 AFM probe tips, experiments were also performed in which tips
were ion-beam-milled into a smooth conical shape but these failed
60
to result in any significant change in the measured mechanical
40
properties (see Supplementary Information, Fig. S3).
The linear–elastic F–d curves in Fig. 2a and b are consistent
20 with elastic beam-bending theory, and were analysed to
determine the mechanical properties of the nanowire system.
0 The relationship between Young’s modulus E and the elastic
0 50 100 150 200 250
deformation of a cylindrical beam or wire of radius r and suspended
Diameter of Au wire (nm)
length L subjected to a load F at its mid-point is given by24

10 3
b E = FL
9 192 dI (1)
Au nanowires (circle)
8
Annealed Au nanowires (star) where the moment of inertia I = (πr4)/4. The relationship between
7 the lateral force F exerted by a rectangular Si cantilever of length Lc,
Yield strength (GPa)

6 width w and thickness t and the measured photodiode signal ∆Ulat

(volts) is well established25:
5
( (
4
3
(
F = Gwt2
4Lc l
3
( 1 ∆U
Sver lat
(2)
2
Bulk gold where G is the shear modulus of Si and l the tip length (typically
1 15–17 µm). The vertical sensitivity Sver was determined by
0 vertical F–d curves. The AFM tip and cantilever dimensions
0 50 100 150 200 250 were determined from high-resolution SEM imaging. Note
Diameter of Au wire (nm) that the quantity F/d in (1) is actually the spring constant of the
combined wire–cantilever system kobs, because the observed
displacement d has contributions from both. kobs is related to
Figure 3 Measured Young’s modulus and yield-strength values for Au the spring constant of the wire kw and that of cantilever kc by
nanowires. a, Young’s modulus values measured for Au nanowires ranging from
40 nm to 250 nm in diameter. The average measured value is 70 ± 11 GPa. The
kwkc .
kobs =
line drawn shows the average values of the Young’s modulus for bulk Au. b, Yield- kw + kc (3)
strength values for untreated (circle) and annealed (star) Au nanowires as a function
of Au nanowire diameters. The average measured value is 3.5 ± 1.1 GPa and The quantity kc was independently determined by measuring the
5.6 ± 1.4 GPa for nanowires with diameters of 200 nm and 40 nm, respectively. cantilever’s response when laterally loading an immoveable hard
The thick black line indicates the typical range for the yield strength of bulk Au. object such as the trench. Failure to account for cantilever loading
Note that annealed nanowires have lower average yield strength, 2.37 ± 0.4 GPa. can result in significant errors (see Supplementary Information,
All errors shown represent one standard deviation obtained for all data points (see Fig. S4). Young’s modulus was then calculated using E = kwL3/(192I)
Methods for discussion of errors). and the yield strength was estimated from the yield force Fy in F–d
curves before the onset of plastic deformation using the equation24:

Fy L
σy = .
influence the F–d curve slope, or complicate subsequent analysis of 2πr 3 (4)
the mechanical properties. The forward trace of the F–d curve before
contacting the nanowire is taken as the reference line of zero force. Figure 3a shows values of E determined for Au nanowires
Figure 2b shows F–d curves recorded during a sequence of with diameters ranging from 40 to 250 nm. The average modulus
repeated loading and unloading cycles. The top two curves show E is 70 ± 11 GPa, essentially independent of diameter and close
that the wire was elastically loaded and unloaded during the to the values reported for bulk gold (78 GPa)26. The measured
forward and reverse trajectory, respectively. Note that the slope is yield strengths in Fig. 3b are diameter-dependent, with average
identical in both the forward and reverse directions, and the F–d values of 3.5 ± 1.1 GPa and 5.6 ± 1.4 GPa for 200-nm and 40-nm
curves for each loading/unloading cycle are highly reproducible. Au nanowires, respectively. These yield strengths are substantially
The AFM image in Fig. 2c recorded afterwards reveals no greater than the 55–200 MPa (actual value dependent on history
permanent deformation of the wire. However, increased loading and heat treatment) reported for bulk Au27, but close to the
eventually results in plastic deformation: the third and fourth F–d maximum theoretical shear stress τmax ≈ G/2π = 4.8 GPa for gold22.
curves show the same initial linear region (with the same slope) Our measurements are also in good agreement with other studies.
followed by a clear break corresponding to the yield point, after For instance, a value of 4 GPa was reported for 1–8-nm gold-atom
which deformation occurs more readily. This is confirmed by chains measured by pulling single atomic wires from a gold contact20.
the subsequent AFM image in Fig. 2d, which shows a permanent Mean yield strengths in the range of 5–10 GPa were also obtained by
nature materials | VOL 4 | JULY 2005 | www.nature.com/naturematerials 527

©2005 Nature Publishing Group

LETTERS
a failure, which then represents the ultimate strength of the nanowire.
In this limit, equation (4) provides an estimate of the ultimate
strength of this strain-hardened brittle material, yielding values
4,000 nN
of 5.8 GPa and 7.9 GPa for 200-nm and 40-nm Au nanowires, the
σmax(F max)
latter being close to the theoretical value of E/10. Second, plastic
σy3 deformation in Au nanowires is accompanied by strain-hardening
Force

with estimated strain-hardening indices that range from 0.29 to 0.4

σy2
(see Supplementary Fig. S5). Although these values are somewhat
Failure lower than expected for face-centred-cubic metals26, they indicate
σy1
that dislocation motion and pile-up is still operative in nanowires
down to diameters of at least 40 nm. Lastly, individual F–d curves
all fall within the envelope of the single-shot experiment, and the
mechanical properties of these wires exactly follow the single-shot
curve throughout the strain-hardening process, indicating that
0 250 500 750 1,000 the material has a pre-determined maximum strain-hardened
strength. This classical strain-hardening behaviour indicates that
Manipulation path (nm)
the materials properties of nanowires are not grossly affected
by moderate cold-working. The sensitivity to the detailed grain
b structure is demonstrated in Fig. 4b, which shows F–d curves for
two identical 200-nm nanowires, one of which was annealed in N2
at 300 °C for 21 h. Under these conditions, recrystallization and
Au nanowire Annealed Au nanowire grain growth occur, with a reduction in dislocation density26. The
curves in Fig. 4b have near-identical elastic slopes, demonstrating
that E is a materials property and unaffected by grain size.
However, the plastic deformation range is 30–50% larger for
Force

annealed nanowires and is characterized by an initial ‘easy glide’

region followed by a strain-hardening region with an index similar
to that for the unannealed nanowire.
These results highlight clear differences between the mechanical
4,000 nN properties of BNMs and metallic nanowires. The former comprised
20–80-nm grains and in the case of Cu exhibited yield strengths
of less than 0.4 GPa without substantial evidence of strain-
hardening even after extensive cold-working1. Molecular dynamics
0 250 500 750 1,000 simulations suggest that grain boundaries in BNMs may source
Manipulation Path (nm) and sink dislocations without build up of a residual dislocation
network30. In contrast, the 40-nm Au nanowires reported here
have yield strengths of 5.6 GPa, whereas corresponding studies on
Figure 4 Plastic deformation of Au nanowires. a, F–d curves recorded during a similar-sized Cu nanowires have an average yield strength of 7 GPa
series of consecutive plastic deformations of a 200-nm Au wire. The curves shown (A. H., B. W. & J. J. B., in preparation). Both nanowire systems
demonstrate that the maximum plastic deformation on a given cycle determines the exhibit strain-hardening, suggesting that the conventional strain-
yield point for the next cycle. Consecutive yield points are labelled σy1, σy2 and so hardening mechanism involving dislocation motion and pile-up is
on. The red curve corresponds to a single-shot experiment in which an identical Au still operative. Although the detailed microscopic mechanism is not
nanowire was plastically deformed and failed in a single manipulation. For clarity, known, it is likely that the finite material volume and grain size in
the unloading trace is plotted by a dashed line. b, Comparison of an untreated and nanowires have important consequences. First, there is a dramatic
annealed 200-nm Au nanowire during which both wires were deformed and failed reduction in the number of defects included during growth, and
in a single-shot experiment. hence a strengthening of the material. Second, the limited number
of grains across the diameter precludes perfect emission and re-
absorption of dislocations during plastic deformation. The latter
is consistent with the non-zero yet reduced work-hardening index
nano-indentation measurements on Au single crystals28,29. The yield- reported here, and suggests that although nanowires are more
strength data in Fig. 3 are therefore consistent with an increased elastoplastic-like than bulk materials, they do not exhibit the near-
strength of the Au nanowires compared with the bulk, presumably perfect elasto-plasticity reported in BNMs1.
due to a corresponding reduction in the number of defects in the Our manipulation and force-sensing method that measures
nanowire material. the full spectrum of nanowire mechanical properties is expected to
Nanowire plasticity was studied by repeatedly cycling wires have applications in a wide range of nanoscale systems, and may be
beyond their yield point. Figure 4a shows the results obtained easily extended to allow electromechanical measurements16.
from a series of consecutive plastic deformations on a single
200-nm Au wire. The F–d curves shown demonstrate that the
maximum plastic deformation on a given cycle determines the METHODS
yield point for the next cycle. Figure 4a also shows a single-shot
SAMPLE PREPARATION AND CHARACTERIZATION
experiment in which an identical Au nanowire was strained Porous alumina anodic oxide (AAO) templates with pore diameters from 40 nm to 100 nm were
beyond the failure point in a single manipulation. There are several prepared in-house by a two-step anodization process as described elsewhere31. Commercially available
remarkable aspects of these data. First, a well-defined yield point Whatman AAO membranes were used to produce 200–300-nm nanowires. Metallic nanowires were
fabricated by electrodepositing metals (gold and copper) into AAO pores. Before electrodeposition, a
exists during successive plastic deformations and the yield force thin silver layer (200–400 nm) was sputtered onto one side of an AAO template as a conductive working
Fy asymptotically approaches the maximum force Fmax just before electrode. In the case of Au-nanowire formation, silver was deposited first to fill the pores, followed

528 nature materials | VOL 4 | JULY 2005 | www.nature.com/naturematerials

©2005 Nature Publishing Group

 LETTERS
by Au filling32. Au and Cu nanowires were electroplated by potentiostatic control from the following 5. Uchic, M. D., Dimiduk, D. M., Florando, J. N. & Nix, W. D. Sample dimensions influence strength
bath: 40 g l–1 KAu(CN)2, 100 g l–1 KH2PO4 and 256 g l–1 CuSO4 aqueous solutions. Silver was removed and crystal plasticity. Science 305, 986–989 (2004).
by using 6 M HNO3 from the Au/AAO template. Templates were dissolved in aqueous NaOH (3 M) for 6. Levitt, A. P. Whisker Technology (A. P. Levitt) Ch. 6 (Wiley-Interscience, New York, 1970).
1 h and nanowires were washed several times with ultrapure Millipore water followed by re-dispersing 7. Yu, M. F. et al. Strength and breaking mechanism of multiwalled carbon nanotubes under tensile
them in deionized water. A drop of nanowire suspension was dispersed on a carbon-coated, 400-mesh, load. Science 287, 637–640 (2000).
copper grid for TEM analysis. Trenches were fabricated by focused-ion-beam milling (10 kV, 1,000 pA) 8. Poncharal, P., Wang, Z. L., Ugarte, D. & De Heer, W. A. Electrostatic deflections and
and were 0.3–4 µm wide and 250–500 nm deep. Nanowires bridging trenches were pinned down at the electromechanical resonances of carbon nanotubes. Science 283, 1513–1516 (1999).
trench edge by the deposition of Pt line using 10 keV electrons at an electron-beam current of 2 pA in 9. Gao, R. P. et al. Nanomechanics of individual carbon nanotubes from pyrolytically grown arrays.
scanning mode. The chamber pressure of the SEM (FEI Strata D8 225) used was 4.6 × 10–6 mbar. On Phys. Rev. Lett. 85, 622–625 (2000).
Pt deposition, the pressure increased by one decade but recovered within 1 min after the deposition 10. Wong, E. W., Sheehan, P. E. & Lieber, C. M. Nanobeam mechanics: elasticity, strength, and toughness
finished. Depending on the line length and height, deposition times were typically 3 to 10 s. The of nanorods and nanotubes. Science 277, 1971–1975 (1997).
deposited Pt lines have a typical height ranging from 40–70 nm and 140–170 nm for 50 nm and 200 nm 11. Salvetat, J. P. et al. Elastic and shear moduli of single-walled carbon nanotube ropes. Phys. Rev. Lett.
nanowires, respectively. SEM and AFM (DI Dimension 3100) were used to determine all critical 82, 944–947 (1999).
dimensions, such as diameter and suspended length of the pinned nanowires. SEM imaging and Pt 12. Salvetat, J. P. et al. Elastic modulus of ordered and disordered multiwalled carbon nanotubes. Adv.
deposition do not influence the mechanical properties, as shown by oxygen-plasma cleaning studies and Mater. 11, 161–165 (1999).
TEM analysis of the pinning contacts, respectively. 13. Kis, A. et al. Shear and Young’s moduli of MoS2 nanotube ropes. Adv. Mater. 15, 733–736 (2003).
14. Cuenot, S., Champagne, S. D. & Nysten, B. Elastic modulus of polypyrrole nanotubes. Phys. Rev.
AFM CANTILEVER CALIBRATION
Lett. 85, 1690–1693 (2000).
Rectangular cantilevers (Budgetsensors) with average normal force constants of 20 to 40 N m–1
15. Walters, D. A. et al. Elastic strain of freely suspended single-wall carbon nanotube ropes. Appl. Phys.
(300 KHz) and 1 to 3 N m–1 (75 KHz) were used to perform bending tests on 200-nm and 50-nm
Lett. 74, 3803–3805 (1999).
Au nanowires, respectively. The cantilever dimensions were accurately determined by SEM. For the
16. Tombler, T. W. et al. Reversible electromechanical characteristics of carbon nanotubes under local-
lateral force calculation, it is assumed that the laser beam spot on the photodiode is circular, and that
probe manipulation. Nature 405, 769–772 (2000).
the angular changes of the reflected laser beam induced by vertical bending and lateral torsion of the
17. Minot, E. D. et al. Tuning carbon nanotube band gaps with strain. Phys. Rev. Lett. 90, 1564011 (2003).
rectangular cantilever results in equal changes of the photodiode signals. After laser positioning no
18. Falvo, M. R. et al. Bending and buckling of carbon nanotubes under large strain. Nature 389,
adjustments were performed during the entire calibration and wire-manipulation process. The lateral
582–584 (1997).
spring constant of the cantilever was calibrated by pushing tips against the wall of a trench in a SiO2
19. Li, X. D., Gao, H. S., Murphy, C. J. & Caswell, K. K. Nanoindentation of silver nanowires. Nano Lett.
calibration grid with 200-nm-deep square holes. Typical cantilever lateral force constants kc are of
3, 1495–1498 (2003).
the order of 200 to 600 N m–1 for 300 kHz cantilevers and 50 to 100 N m–1 for 75 kHz cantilevers. Tip
20. Agrait, N., Rubio, G. & Vieira, S. Plastic deformation of nanometer-scale gold connective necks.
velocities in the x–y direction from 10 nm s–1 to 1 µm s–1 were investigated to check the influence of
Phys. Rev. Lett. 74, 3995–3998 (1995).
tip speed on both F–d curves of calibration of cantilevers and the wire bending. The tip velocity has no
21. Pan, Z. W. et al. Tensile tests of ropes of very long aligned multiwall carbon nanotubes. Appl. Phys.
influence on the observed data.
Lett. 74, 3152–3154 (1999).
MECHANICAL MEASUREMENTS OF NANOWIRES 22. Barrett, C. R., Nix, W. D. & Tetelman, A. S. in The Principles of Engineering Materials Ch. 6–8
A Dimension 3100 AFM (DI instruments) equipped with the Nanoman software package and closed (Prentice-Hall, Englewood Cliffs, New Jersey, 1973).
x–y loop was used to conduct bending tests on wires in air at room temperature. With the help of built- 23. Koops, H. W. P., Kaya, A. & Weber, M. Fabrication and characterization of platinum nanocrystalline
in high-magnification optical microscopy and several SEM images of wire configuration, the clamped material grown by electron-beam induced deposition. J. Vac. Sci. Technol. B 13, 2400–2403 (1995).
nanowires were positioned in a direction where the axis line of wires is parallel to the cantilever axis, 24. Gere, J. M. & Timoshenko, S. P. in Mechanics of Materials Ch. 5.1.8 (PWS-KENT, Boston,
which minimizes the possibility of the slippage between wires and tip during the loading process. Massachusettes, 1990).
This also ensures that the cantilever twists during manipulation, and eliminates any possible vertical 25. Schwarz, U. D., Koster, P. & Wiesendanger, R. Quantitative analysis of lateral force microscopy
deflection of the cantilever. experiments. Rev. Sci. Instrum. 67, 2560–2567 (1996).
26. Callister, W. D. Jr in Materials Science and Engineering 3rd edn, Ch. 6–8 (Wiley, New York, 1994).
ERROR ANALYSIS
27. Espinosa, H. D., Prorok, B. C. & Peng, B. Plasticity size effects in free-standing submicron
The mechanical measurements shown here rely on estimates of the physical dimensions of the wire
polycrystalline FCC films subjected to pure tension. J. Mech. Phys. Solid 52, 667–689 (2004).
and cantilever, which are the dominant source of errors. The suspended length and the diameter for
28. Kiely, J. D. & Houston, J. E. Nanomechanical properties of Au (111), (001), and (110) surfaces. Phys.
smallest nanowires are rather precisely measured by SEM and AFM with an accuracy of the order of 5%.
Rev. B 57, 12588–12594 (1998).
Although the error propagation for Young’s modulus and yield strength are similar, the relative error
29. Corcoran, S. G. & Colton, R. J. Anomalous plastic deformation at surfaces: Nanoindentation of gold
for yield strength is smaller. The maximum relative error in the nanowire modulus and yield strength,
single crystals. Phys. Rev. B 55, R16057–R16060 (1997).
accounting for uncertainties in the cantilever and wire dimensions, together with the photodiode
30. Schiotz, J. & Jacobsen, K. W. A maximum in the strength of nanocrystalline copper. Science 301,
sensitivity, is estimated to be no more than 40%. In addition, the maximum variation in the vertical
1357–1359 (2003).
positioning of the tip due to the lack of z closed-loop control is 200 nm, which, when compared with the
31. Masuda, H. & Fukuda, K. Ordered metal nanohole arrays made by a two-step replication of
overall tip length of 15–17 µm, results in a 1% error.
honeycomb structures of anodic alumina. Science 268, 1466–1468 (1995).
Received 28 January 2005; accepted 18 April 2005; published 5 June 2005. 32. Kovtyukhova, N. I. & Mallouk, T. E. Nanowires as building blocks for self-assembling logic and
memory circuits. Chem. Eur. J. 8, 4354–4363 (2002).
References
1 Champion, Y. et al. Near-perfect elastoplasticity in pure nanocrystalline copper. Science 300, Acknowledgements
310–311 (2003). We thank John Monaghan and Patrick Prendergast for the valuable discussions. This work was
2. Lu, L., Sui, M. L. & Lu, K. Superplastic extensibility of nanocrystalline copper at room temperature. supported by Science Foundation Ireland under grant 00/PI.1/C077A.2.
Science 287, 1463–1466 (2000). Correspondence and requests for materials should be addressed to J.J.B.
3. Budrovic, Z., Swygenhoven, H. V., Derlet, P. M., Petegem, S. V. & Schmitt, B. Plastic deformation Supplementary Information accompanies the paper on www.nature.com/naturematerials.
with reversible peak broadening in nanocrystalline nickel. Science 304, 273–276 (2004).
4. Lu, L., Shen, Y. F., Chen, X. H., Qian, L. H. & Lu, K. Ultrahigh strength and high electrical Competing Financial Interests
conductivity in copper. Science 304, 422–426 (2004). The authors declare that they have no competing financial interests.

nature materials | VOL 4 | JULY 2005 | www.nature.com/naturematerials 529

©2005 Nature Publishing Group