ARTICLES

PUBLISHED ONLINE: 14 FEBRUARY 2010 | DOI: 10.1038/NNANO.2010.4

Cold welding of ultrathin gold nanowires

Yang Lu1, Jian Yu Huang2, Chao Wang3, Shouheng Sun3 and Jun Lou1 *
The welding of metals at the nanoscale is likely to have an important role in the bottom-up fabrication of electrical and
mechanical nanodevices. Existing welding techniques use local heating, requiring precise control of the heating mechanism
and introducing the possibility of damage. The welding of metals without heating (or cold welding) has been
demonstrated, but only at macroscopic length scales and under large applied pressures. Here, we demonstrate that singlecrystalline gold nanowires with diameters between 3 and 10 nm can be cold-welded together within seconds by mechanical
contact alone, and under relatively low applied pressures. High-resolution transmission electron microscopy and in situ
measurements reveal that the welds are nearly perfect, with the same crystal orientation, strength and electrical
conductivity as the rest of the nanowire. The high quality of the welds is attributed to the nanoscale sample dimensions,
oriented-attachment mechanisms and mechanically assisted fast surface-atom diffusion. Welds are also demonstrated
between gold and silver, and silver and silver, indicating that the technique may be generally applicable.

elding, with its historic development tracing back to the

Bronze Age, serves modern industry in many areas
where metals are used1. When, in the 1940s, people
started to recognize cold welding as a general phenomenon, it
had already been practised for more than 700 years (refs 2,3).
Unlike normal welding, in which liquids or molten phases need
to be present, cold welding is a solid-state welding process in
which joining takes place without fusion (the process of causing a
material to melt with intense heat) at the interface. However, to
realize cold welding in bulk metals, either a high applied normal/
frictional load or an atomically clean at ductile surface in an ultrahigh-vacuum environment are generally required. About two
decades ago, Whitesides and colleagues discovered that a metallic
thin lm such as gold, supported on compliant elastomers, could
weld together at remarkably low loads under ambient laboratory
conditions3. Although its underlying mechanism was not fully
understood, this nding extended the cold-welding process into
the fabrication of a broad range of modern organic microelectronic/optoelectronic devices, including organic light-emitting
devices (OLED) and photovoltaic cells4. Now, with extensive
research being conducted into nanoelectronic devices and
nanoelectromechanical systems (NEMS), whether or not cold
welding can also be practised at the nanoscale has become an
interesting topic. In recent years, scientists have successfully
realized the joining of individual low-dimensional nanostructures
such as carbon nanotubes58 and metal/semiconductorlled carbon nanotubes912, metal/semiconductor/ceramic nanowires and nanoparticles1321, by either applying voltage/
current5,7,10,14,17,20 or heating the sample stage9,15,21, or by focusing
high-intensity electron or laser beams onto the joining
section6,8,1113,16,18,19. Although these methods certainly have their
advantages, such nanoscale welding techniques have always
involved local heating processes of some kind, which can be
difcult to control precisely at the relevant length scales, and may
change the underlying substructures and related properties of
the original building blocks. Owing to these limitations, the idea
of cold welding, which joins nanostructures without heating, has
become an attractive solution for bottom-up assembly at
the nanoscale.

In this work, cold welding of individual gold nanowires was

performed and monitored inside a high-resolution transmission
electron microscope (HRTEM) equipped with NanofactoryTM
TEM-scanning tunnelling microscopy (STM) and TEM-atomic
force microscopy (AFM) sample holders (see Methods). Ultrathin
gold nanowires (diameters , 10 nm) were chosen, because they
are widely considered to be ideal candidates for achieving extremely
dense logic and memory circuits in future molecular-scale interconnects22. The good resistance to oxidization of gold is another useful
property. Samples were selected from two sources. The rst type
comprises ultrathin gold nanowires with relatively small aspect
ratios, that is, which are 510 nm in diameter and 1050 nm
in length (referred to as nanorods in this paper). They were originally
formed as ligaments that were cut off in situ from home-made porous
gold nanostructures (Fig. 1a). The porous gold nanostructures were
obtained by means of a de-alloying process applied to a goldsilver
alloy23. The second type of sample comprises the recently developed
micrometre-long ultrathin gold nanowires24, which are 39 nm in
diameter (Fig. 1b). HRTEM imaging and selected area diffraction
(SAD) showed that both samples had single-crystalline face-centred
cubic (fcc) structures. The average interfringe distances for both
types of sample were measured to be 0.230.24 nm, corresponding
to (111) lattice spacing (0.23 nm) of the fcc gold crystal.

Nanoscale cold welding

Using the TEMSTM holder, both head-to-head and side-to-side
joining procedures were performed for both sample types (Fig. 2a
and b, respectively). Other welding geometries could also be realized, such as head-to-side joining (see Supplementary
Information). Each individual nanowire was manipulated using a
tungsten or gold STM probe (Fig. 2c) driven by the movable
piezoelectric head of the holder. The joining of two gold nanorods
was rst attempted in a head-to-head orientation (Fig. 3; see also
Supplementary Movie S1). One nanorod was manipulated
towards another, continuously adjusting the alignment until the
two samples approached one another, head to head (Fig. 3a,b).
When the front surfaces of the two nanorods came into contact,
they welded together instantly (1.5 s, Fig. 3c,d). After welding,
the image contrast of the welded nanorod became increasingly

1
Department of Mechanical Engineering and Materials Science, Rice University, Houston, Texas 77005, USA, 2 Center for Integrated Nanotechnologies
(CINT), Sandia National Laboratories, Albuquerque, New Mexico 87185, USA, 3 Department of Chemistry and Division of Engineering, Brown University,
Providence, Rhode Island 02912, USA. * e-mail: jlou@rice.edu

218

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.

NATURE NANOTECHNOLOGY

ARTICLES

DOI: 10.1038/NNANO.2010.4

Figure 1 | Two types of ultrathin gold nanowire samples used for cold-welding experiments. a,b, TEM images of an ultrathin gold nanorod (a, scale bar
5 nm) on a porous gold nanostructure, and micrometre-long ultrathin gold nanowires (b, scale bar 100 nm). Insets: corresponding HRTEM images showing
the crystalline structures. Scale bars, 5 nm. (Chemically fabricated micrometre-long nanowires are usually covered with a layer of surfactant (oleylamine).
Mechanical rubbing between two nanowires can effectively remove the residual surfactant on their surfaces before welding experiments.).

c
a

STM probe

STM probe
STM probe

Figure 2 | Head-to-head and side-to-side cold-welding geometries. a,b, Schematics of two welding geometries for ultrathin gold nanowires: head-to-head (a)
and side-to-side (b) (d represents the virtual bending deection of the top nanowire when contacting the bottom nanowire). c, TEM image showing the
manipulation of a longer nanowire towards a short nanowire. Scale bar, 10 nm.

uniform (Fig. 3df ), indicating continuous substructure evolutions

(Fig. 3d,e) to smooth the surface. Once this process was complete,
the STM probe was retracted (Fig. 3fh) along the direction indicated by the arrow, until it was fully separated from the sample. It
was shown that the as-welded nanorod maintained its morphology
and structure in the free-standing state (Fig. 3i). Similar joining was
also successfully achieved for micrometre-long ultrathin gold nanowires; however, head-to-head welding for the gold nanorods was
easier to perform because there was no problem with buckling, as
was the case when manipulating longer nanowires.
For side-to-side welding, precise alignment is not required, and
successful joining of both types of sample was easily carried out
many times. By simply manipulating the nanowires so that they partially overlapped one another, side-by-side contact could be made,
and cold welding would always occur quickly (Fig. 4a,b). In Fig. 4,
one particular welding process is shown to nish within 34 s
(Fig. 4b,c), and HRTEM imaging in Fig. 4d shows the structure of

the as-welded structure following the relaxation process. To check

its welding quality, in situ pulling of the as-welded nanowire
(Fig. 4ej) was performed immediately after the rst joining.
Surprisingly, the welded nanowire formed a neck and broke at
another location rather than at the joining section of the rst
welding. A comparison of the length of the nanowire remaining
at the bottom during the pulling process (original, 21.5 nm in
Fig. 4a; after breaking, 25.7 nm in Fig. 4j) clearly demonstrates
that the as-welded nanowire broke at a location 4.2 nm above
the original welding spot. This qualitative pulling result implies
that the as-welded structure is as strong as the original nanowire.
For the same sample shown in Fig. 4aj, we also performed a
second welding in the head-to-head mode, and again pulled the
as-welded nanowire until breaking occurred (Supplementary
Movie S2). During this pulling process, the sample formed a neck
near the second welding zone (Fig. 4 k). Fast Fourier transformations (FFT) from images taken from both the welded segment

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.

219

ARTICLES
a

NATURE NANOTECHNOLOGY
0s

22.5 s

1 min 14 s

20 s

37.5 s

1 min 34.5 s

DOI: 10.1038/NNANO.2010.4

21 s

54 s

1 min 57 s

Figure 3 | Head-to-head welding of two gold nanorods. a,b, One nanorod (right) is caused to approach another (left) until their front surfaces come into
contact. ce, The welding process is completed within 1.5 s (c,d) followed by structure relaxation (d,e). fi, After withdrawal of the STM probe (fi), the
as-welded nanowire is left in the free-standing state (i). Triangles indicate the front edges of the two nanorods. Arrows indicate the withdrawing direction
of the STM probe. Scale bars, 5 nm.

and the remaining segment of the nanowire (which also contained

the rst welding zone) conrmed that the second welding zone and
the remaining part of the nanowire (Fig. 4l) were both single crystalline, in the same ,111. orientation.

In situ mechanical and electrical measurements

To quantitatively determine their strength, an in situ TEMAFM
holder was used to perform tensile tests on as-welded nanowires,
as illustrated in Fig. 5a. It should be noted that the two
nanowire samples to be welded were obtained by breaking one
original gold nanowire with a measured tensile strength of
600+50 MPa (engineering stress). A side-to-side cold-welding
experiment was then carried out. Following the completion of
the joining and relaxation process (Fig. 5b,c), the as-welded nanowire was pulled away from the AFM tip. Once again, the breaking
point was not in the welding zone (Supplementary Movie S3).
By measuring the deection of the AFM cantilever (DD,
see Methods), a tensile strength of 580+40 MPa (engineering
stress) was obtained. This compares very well with the strength
of the original nanowire. In contrast, the tensile strength for
bulk gold is normally 100 MPa (ref. 25). This drastically
increased mechanical strength of gold nanowires has previously
been demonstrated both experimentally and computationally26,27
(see discussions in Supplementary Information). The present
measurements clearly conrm that as-welded nanowires retain
the superior mechanical properties of the original singlecrystalline nanowires.
Finally, in situ electrical measurements were conducted for the
original and as-welded nanowires using the TEMSTM holder,
and their currentvoltage (I2V) responses were compared. In
Fig. 6a, a gold nanowire (length, 130 nm; diameter, 7 nm) was
bridged between gold probes. While applying a bias of 21 to
1 mV, I2V measurements were carried out nine times, resulting
220

in an average electrical resistivity of 292.6+5.8 V nm. As-welded

nanowires were then obtained by breaking the original nanowire
and re-welding it, as shown in Fig. 6b,c. Eleven cycles of breaking
and re-welding were performed in side-to-side or head-to-head
modes at different welding locations (because the as-welded nanowires often broke at locations other than the previous welding
spot). Electrical measurements were carried out at least twice for
each as-welded nanowire. The corresponding averaged I2V
curves are plotted in Fig. 6d, together with that of the original nanowire. An average electrical resistivity of 298.1+14.5 V nm was calculated from all eleven successfully welded nanowires, clearly
demonstrating that the electrical resistivity changed very little with
each successful welding, and that the average resistivity of the
as-welded samples was indeed very close to that of the original
nanowire. These results also compare very well with the
resistivity results from ref. 24 (260 V nm) for the same micrometre-long ultrathin gold nanowires. Even without making full
correction for the contact resistance, the electrical resistivities of
these original and as-welded nanowires were already lower than
many other types of gold nanowires with diameters ranging
from 4 to 90 nm (for example, refs 2830), clearly suggesting
the great potential of using ultrathin gold nanowires as future
interconnects and cold welding as an efcient nanoscale
assembly technique.

Comparison with fusion and macroscopic cold welding

In contrast to other existing techniques for joining individual
nanostructures, the demonstrated technique distinguished itself by
requiring no local heating process; that is, there was no need to
apply bias or to use a dedicated heating stage. Additional heating
effects from the electron beam can be ruled out31,32, because only
low-intensity-spread electron beams were used for imaging
(for further discussion, see Supplementary Information). The

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.

NATURE NANOTECHNOLOGY
a

0s

5 min
55 s

ARTICLES

DOI: 10.1038/NNANO.2010.4

21 s

5 min
57 s

55 s

6 min
2.5 s

4 min
36.5 s

6 min
8s

5 min
49.5 s

6 min
10 s

Figure 4 | Side-to-side welding of two gold nanowires. aj, Welding of two ultrathin gold nanowires was completed within 34 s by making side-to-side
contacts (ac), followed by structure relaxation (cd) and in situ pulling of the as-welded nanowire (ej). The thin double-headed arrows in a and j indicate
the bottom nanowire length before and after the rst welding and pulling. The two broken nanowires in j were re-welded by making a second contact,
followed by a second pulling and breaking process. k,l, HRTEM images of the necking area during the second pulling (k) and the remaining nanowire at the
bottom after the second breaking step (l), respectively. Insets: diffraction patterns from the regions marked by squares in both images, calculated by fast
Fourier transformation (FFT). Again, the triangles indicate the two edges of the two nanowires before welding, and the thicker single-headed arrows indicate
the STM probe pulling direction. Scale bars, 5 nm.

cold-welding processes were faster than most other welding processes involving heating11,1315, and were completed close to room
temperature, with no observable fusion occurring at the welding
interface. As a result, the single-crystalline structures of the original
and as-welded nanowires were well maintained during the welding
process, with almost no defects or impurities introduced. The aswelded nanowires was at least as strong as the original nanowires,
due to the fact that the welding zone had the same lattice structure
and connected to the original wires with no observable grain boundaries. It also appears that cold welding has very little effect on
electron conductionthis could again be attributed to the nearperfect welding zone formed during the process. More importantly,
we have successfully extended this technique to other metal systems
such as silversilver nanowires and goldsilver nanowires
(Supplementary Figs S2,S3).
Unlike the traditional cold welding of bulk materials, which
normally requires high load, the cold welding of the ultrathin
nanowires described in this paper can occur easily in head-tohead welding experiments where there are matching crystalline
orientations, and little external force is needed. In the side-to-side

welding mode, joining could occur by simply making mechanical

contact between the nanowires to be welded, and no signicant
deformation due to contact was found in any of the experiments.
Although the exact contact geometry and applied load were difcult to quantify at the joining interface, an estimate of the applied
stress in joining the nanowires in the side-to-side geometry
(Fig. 2b) was attempted using simple beam theory. Assuming
that a small lateral deection of d 1 nm for a 100-nm-long nanowire sample (Fig. 2c) resulted from the side-to-side contact, where
the contact surface area was 10 nm2 and Youngs modulus for
common gold nanowires (70 GPa) is used27, the calculated
applied stress is only 4.7 MPa. This value is considered to be
the upper limit of the actual applied stress, because very little
lateral deection was observed in the actual experiments involving
side-to-side welding (Supplementary Movies S2,S3). This value,
not only much smaller than the requirement for traditional
cold welding of bulk metals, is also smaller than the required
pressure for cold welding of metallic thin lm (.100 MPa in
ref. 4). It should be noted that the reported pressure for cold
welding of gold thin lm (0.1 g cm22; that is, 9.8 Pa in ref. 3)

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.

221

ARTICLES

NATURE NANOTECHNOLOGY
a

DOI: 10.1038/NNANO.2010.4

STM probe

Loading direction
F

Reference bar

Nanowire sample

D
AFM cantilever

Figure 5 | In situ tensile strength measurements of the nanowelds. a, Schematic showing how the AFM cantilever acts as a force sensor (by measuring the
deection of the cantilever, DD) and the STM probe acts as an actuator while the attached nanowire sample is under tensile loading. b,c, Two gold nanowires
before and after cold welding on the TEMAFM holder. d, As-welded nanowire under the maximum load state. e, Broken nanowires at the steady state (note
that the breaking point is no longer the same as the initial contact point). The thin double-headed arrows in d and e indicate the relative displacements
of the AFM cantilever tip with respect to the reference bar. Triangles indicate the edges of the two nanowires before and during welding, and the thicker
single-headed arrows indicate the tensile loading direction. Scale bars, 10 nm.

represents the average stress over a large 1 cm  1 cm area, which

could have many local contact points due to surface asperities.
The actual pressure for individual asperities would probably be
much higher. Therefore, this work may offer a nanoscopic view
of the initial stages of macroscopic cold welding for either bulk
metals or metallic thin lm.

Mechanisms of nanoscale cold welding

Atomic diffusion and surface relaxation, considered important
factors in macroscopic cold welding, were obviously at play in the
aforementioned nanoscale process2,3,33. It is well recognized that
the diffusion barrier for a single metal atom on a metal surface is
quite low (typically less than 1 eV)34. Thermal activation, even at
room temperature, is enough to overcome such low barriers, so isolated metal atoms can diffuse rapidly by means of surface diffusion.
However, to create such isolated atoms demands a much higher
energy cost. It is the combination of the formation and diffusion
energy barriers that determines the cold welding observed in this
work. The mechanical manipulation clearly provided the necessary
extra driving force to facilitate the cold welding and unication of
the two nanowires.
We also believe that the oriented-attachment mechanism, as
reported for PdSe nanocrystal15,35, was playing an important part
in the welding occurring at close to room temperature. Evidence
for this is provided by the fact that cold welding always occurred
easily and instantaneously between two nanowires with the same
growth orientation, in particular for those that were obtained by
breaking one original nanowire into two segments. Matching the
orientation is therefore key to realizing successful welding. It will
certainly be very interesting to see if future work can verify
whether the aforementioned mechanically assisted surface-atom
222

diffusion alone could facilitate the cold welding of nanowires with

different crystal orientations. On the other hand, although
PbSe nanocrystal particles still require low-temperature heating
(100150 8C; ref. 15) for unication, the welding process reported
here had no such requirement. This may be understood as follows.
First, the mechanical manipulation of the nanowire, instead of
local heating, helped to match the orientation of the samples in
what was the beginning of the oriented-attachment process15.
Second, as in traditional cold welding, the use of a clean surface
under conditions of high vacuum is an important factor2,3, and
the gold nanowire samples (particularly the freshly broken nanowires) in the TEM chamber clearly satised this requirement.

Conclusions
We have demonstrated that the cold welding technique has
the capability to join ultrathin gold nanowires without introducing defects. The welding occurs at close to room temperature,
and its exceptional quality is attributed to the nanoscale
sample dimensions, oriented-attachment mechanisms, as well
as mechanically assisted surface atom diffusions. This process
requires no heating or high load, and can be carried out
relatively quickly. More importantly, neither the mechanical
nor electrical properties of the nanowires were affected. These
results provide the rst atomic-scale visualization of the coldwelding process, revealing, for the rst time, the physical mechanisms of the cold welding of nanowires. Combined with other
nano- and microfabrication technologies3638, nanoscale cold
welding is anticipated to have potential applications in the
future bottom-up assembly of metallic one-dimensional nanostructures and next-generation interconnects for extremely
dense logic circuits.

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.

NATURE NANOTECHNOLOGY

DOI: 10.1038/NNANO.2010.4

ARTICLES

accurate (less than +10% error) by measuring cantilever deections in high

magnication TEM images.
All welding experiments were carried out using an FEITM Tecnai G2 F30 TEM,
operating predominantly at 300 kV working voltage (lower working voltages were
also used to rule out the electron-beam heating effect; see discussions in
Supplementary Information). During the welding experiments, no current was
passed through the sample, and a low-intensity electron beam was always used for
imaging and video capturing.

Received 6 November 2009; accepted 13 January 2010;

published online 14 February 2010

References

d
1,000

Current (nA)

500
Original
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
11th

0
500
1,000
1.0

0.5

0.0

0.5

1.0

Bias (mV)

Figure 6 | In situ electrical measurements of the nanowelds. ac, Gold

nanowire (length, 130 nm; diameter, 7 nm) bridged between gold probes
in initial (a), broken (b) and welding (c) states. Scale bars, 10 nm.
d, Comparison of the averaged IV curves of the original and as-welded gold
nanowires. Cold welding was successfully performed 11 times for the same
sample by repeated breaking and re-welding of the nanowire, as shown
in ac. The y-axis (current) error was 2% for the original nanowire
measurement and 5% for each measurement of the as-welded nanowire.

Methods
In situ TEM samples were prepared by adhering nanowires or nanorods onto
tungsten or gold STM probes and then loading the probes into the TEMSTM or
TEMAFM holders (NanoFactoryTM Instruments). Three-dimensional movement
of the STM probes was driven by the piezo-electric heads of the holders. In addition
to the normal manipulation and joining of the nanowires, the TEMSTM holder
could also apply a specic bias and measure the current response of the nanowire
sample bridged across two gold STM probes.
The TEMAFM holder was used primarily for quantitative measurement of
nanowire strength, for which a silicon AFM cantilever beam with known spring
constant (4.8 N m21) was deected by the clamped nanowire sample under
tensile loading. Attachment of the sample to the AFM cantilever was carried out
by pushing the nanowire against the AFM tip surface, which was coated with an
adhesive layer, until strong bonding was formed. Because the deection of the
cantilever was much smaller than its length, a linear relationship between DD
(displacement of the AFM tip, equal to the cantilever deection) and F (force
applied on the nanowire sample) was assumed. During the experiment, a
selected area diffraction (SAD) centre-spot blocking bar, which was free of
movement throughout the process of taking TEM images and videos, was
inserted as the reference for displacement measurements. The tensile strength
was calculated as engineering stress. The stress calculation was reasonably

1. Opening remarks. J. Am. Welding Soc. 1, 3 (1919).

2. Freitas, R. A. & Gilbreath, W. P. (eds) Advanced Automation for Space Missions:
Proceedings of the 1980 NASA/ASEE Summer Study, Appendix 4C.1.
(NASA, 1980).
3. Ferguson, G. S., Chaudhury, M. K., Sigal, G. B. & Whitesides, G. M. Contact
adhesion of thin gold lms on elastomeric supports: cold welding under ambient
conditions. Science 253, 776778 (1991).
4. Kim, C., Burrows, P. E. & Forrest, S. R. Micropatterning of organic electronic
devices by cold-welding. Science 288, 831833 (2000).
5. Jin, C., Suenaga, K. & Iijima, S. Plumbing carbon nanotubes. Nature Nanotech.
3, 1721 (2008).
6. Wang, M., Wang, J., Chen, Q. & Peng, L. M. Fabrication and electrical and
mechanical properties of carbon nanotube interconnections. Adv. Funct. Mater.
15, 18251831 (2005).
7. Hirayama, H., Kawamoto, Y., Hayashi, H. & Takayanagi, K. Nanospot welding of
carbon nanotubes. Appl. Phys. Lett. 79, 11691171 (2001).
8. Madsen, D. N. et al. Soldering of nanotubes onto microelectrodes. Nano Lett. 3,
4749 (2003).
9. Wu, Y. & Yang, P. Melting and welding semiconductor nanowires in nanotubes.
Adv. Mater. 13, 520523 (2001).
10. Dong, L., Tao, X., Zhang, L., Zhang, X. & Nelson, B. J. Nanorobotic spot welding:
controlled metal deposition with attogram precision from copper-lled carbon
nanotubes. Nano Lett. 7, 5863 (2007).
11. Misra, A. & Daraio, C. Sharp carbon-nanotube tips and carbon-nanotube
soldering irons. Adv. Mater. 20, 14 (2008).
12. Rodrguez-Manzo, J. A. et al. Heterojunctions between metals and carbon
nanotubes as ultimate nanocontacts. Proc. Natl Acad. Sci. USA 106,
45914595 (2009).
13. Xu, S. et al. Nanometer-scale modication and welding of silicon and metallic
nanowires with a high-intensity electron beam. Small 1, 12211229 (2005).
14. Tohmyoh, H., Imaizumi, T., Hayashi, H. & Saka, M. Welding of Pt nanowires by
Joule heating. Scripta Mater. 57, 953956 (2007).
15. van Huis, M. A. et al. Low-temperature nanocrystal unication through rotations
and relaxations probed by in situ transmission electron microscopy. Nano Lett.
8, 39593963 (2008).
16. Kizuka, T., Yamada, K., Deguchi, S., Naruse, M. & Tanaka, N. Time-resolved
high-resolution electron microscopy of atomic scale solid-state direct bonding of
gold tips. J. Electron Microsc. 46, 151160 (1997).
17. Tohmyoh, H. A governing parameter for the melting phenomenon at
nanocontacts by Joule heating and its application to joining together two thin
metallic wires. J. Appl. Phys. 105, 014907 (2009).
18. Kim, S. J. & Jang, D. J. Laser-induced nanowelding of gold nanoparticles. Appl.
Phys. Lett. 86, 033112 (2005).
19. Moskalenko, A. V., Burbridge, D. J., Viau, G. & Gordeev, S. N. Electron-beaminduced welding of 3D nano-objects from beneath. Nanotechnology 18,
025304 (2007).
20. Peng, Y., Cullis, T. & Inkson, B. Bottom-up nanoconstruction by the welding of
individual metallic nanoobjects using nanoscale solder. Nano Lett. 9,
9196 (2009).
21. Gu, Z., Ye, H., Smirnova, D., Small, D. & Gracias, D. H. Reow and electrical
characteristics of nanoscale solder. Small 2, 225229 (2006).
22. Lu, W. & Lieber, C. M. Nanoelectronics from the bottom up. Nature Mater. 6,
841850 (2007).
23. Ji, C. & Searson, P. C. Synthesis and characterization of nanoporous gold
nanowires. J. Phys. Chem. B 107, 44944499 (2003).
24. Wang, C., Hu, Y., Lieber, C. M. & Sun, S. Ultrathin Au nanowires and their
transport properties. J. Am. Chem. Soc. 130, 89028903 (2008).
25. Howatson, A. M., Lund, P. G. & Todd, J. D. Engineering Tables and Data 41, 2nd
edn (Chapman and Hall, 1991).
26. Gall, K., Diao, J., Agrait, N. & Dunn, M. L. The strength of gold nanowires.
Nano. Lett. 4, 24312436 (2004).
27. Wu, B., Heidelberg, A. & Boland, J. J. Mechanical properties of ultrahighstrength gold nanowires. Nature Mater. 4, 525529 (2005).
28. Ramsperger, U., Uchihashi, T. & Nejoh, H. Fabrication and lateral
electronic transport measurements of gold nanowires. Appl. Phys. Lett. 78,
8587 (2001).

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.

223

ARTICLES

NATURE NANOTECHNOLOGY

29. Calleja, M., Tello, M., Anguita, J., Garcia, F. & Garcia, R. Fabrication of gold
nanowires on insulating substrates by eld-induced mass transport. Appl. Phys.
Lett. 79, 24712473 (2001).
30. Song, J. H., Wu, Y., Messer, B., Kind, H. & Yang, P. Metal nanowire formation
using Mo3Se32 as reducing and sacricing templates. J. Am. Chem. Soc. 123,
1039710398 (2001).
31. Jose-Yacaman, M. et al. Surface diffusion and coalescence of mobile metal
nanoparticles. J. Phys. Chem. B 109, 97039711 (2005).
32. Rez, P. & Glaisher, R. W. Measurement of energy deposition in transmission
electron microscopy. Ultramicroscopy 35, 6569 (1991).
33. Kizuka, T. Atomic process of point contact in gold studied by time-resolved
high-resolution transmission electron microscopy. Phys. Rev. Lett. 81,
44484451 (1998).
34. Sanders, D. E. & DePristo, A. E. Predicted diffusion rates on fcc (001) metal
surfaces for adsorbate/substrate combinations of Ni, Cu, Rh, Pd, Ag, Pt, Au.
Surf. Sci. 260, 116128 (1992).
35. Cho, K. S., Talapin, D. V., Gaschler, W. & Murray, C. B. Designing PbSe
nanowires and nanorings through oriented attachment of nanoparticles. J. Am.
Chem. Soc. 127, 71407147 (2005).
36. Zhong, Z., Wang, D., Cui, Y., Bockrath, M. W. & Lieber, C. M. Nanowire
crossbar arrays as address decoders for integrated nanosystems. Science 302,
13771379 (2003).

224

DOI: 10.1038/NNANO.2010.4

37. Whang, D., Jin, S., Wu, Y. & Lieber, C. M. Large-scale hierarchical organization
of nanowire arrays for integrated nanosystems. Nano Lett. 3, 12551259 (2003).
38. Huo, F. et al. Polymer pen lithography. Science 321, 16581660 (2008).

Acknowledgements
Y.L. and J.L. acknowledge the nancial support provided by the Air Force Ofce of
Sponsored Research (AFOSR) YIP award FA9550-09-1-0084 and by National Science
Foundation (NSF) grant ECCS-0702766. This work was performed, in part, at the Center
for Integrated Nanotechnologies, a US Department of Energy, Ofce of Basic Energy
Sciences user facility. Sandia National Laboratories is a multiprogram laboratory operated
by Sandia Corporation, a Lockheed-Martin Company, for the US Department of Energy
under contract no. DE-AC04-94AL85000.

Author contributions
Y.L., J.H. and J.L. conceived and designed the experiments. Y.L. performed the experiments.
Y.L., J.H. and J.L. analysed the data. C.W. and S.S. supplied materials. Y.L. and J.L.
composed the manuscript. All authors discussed the results and edited the manuscript.

Additional information
The authors declare no competing nancial interests. Supplementary information
accompanies this paper at www.nature.com/naturenanotechnology. Reprints and
permission information is available online at http://npg.nature.com/reprintsandpermissions/.
Correspondence and requests for materials should be addressed to J.L.

NATURE NANOTECHNOLOGY | VOL 5 | MARCH 2010 | www.nature.com/naturenanotechnology

2010 Macmillan Publishers Limited. All rights reserved.