Single nanoporous gold nanowire sensors

Article abstract of DOI:10.1021/jp056940t

Chemisorption from the gas or liquid phase can result in a measurable resistance change in a metallic material when at least one dimension is smaller than the mean free path for electrons. Here we report on the fabrication of single nanoporous gold nanowires and demonstrate that adsorption of an alkanethiol can be monitored in real time. Single nanowire devices were fabricated by in situ etching of Au_0.18Ag_0.82 alloy nanowires in dilute nitric acid. The evolution of the porous structure was characterized by monitoring the resistance change and comparing to cross-sectional images. The feature size of about 10 nm is less than the mean free path for electrons in bulk gold, and hence the resistance is dominated by surface scattering. Adsorption of a monolayer of octadecanethiol onto the nanoporous gold nanowire results in a resistance change of about 3%. The sensitivity factor of 1.0 × 10^-16 cm² is comparable to values reported for adsorption at ultrathin films.

Full text of DOI:10.1021/jp056940t

4318
J. Phys. Chem. B 2006, 110, 4318-4322
Single Nanoporous Gold Nanowire Sensors
Zhu Liu and Peter C. Searson*
Department of Materials Science and Engineering, Johns Hopkins UniVersity, Baltimore, Maryland 21218
ReceiVed: NoVember 29, 2005; In Final Form: January 5, 2006
Chemisorption from the gas or liquid phase can result in a measurable resistance change in a metallic material
when at least one dimension is smaller than the mean free path for electrons. Here we report on the fabrication
of single nanoporous gold nanowires and demonstrate that adsorption of an alkanethiol can be monitored in
real time. Single nanowire devices were fabricated by in situ etching of Au0.18Ag0.82 alloy nanowires in dilute
nitric acid. The evolution of the porous structure was characterized by monitoring the resistance change and
comparing to cross-sectional images. The feature size of about 10 nm is less than the mean free path for
electrons in bulk gold, and hence the resistance is dominated by surface scattering. Adsorption of a monolayer
of octadecanethiol onto the nanoporous gold nanowire results in a resistance change of about 3%. The sensitivity
-
16
2
factor of 1.0 × 10 cm is comparable to values reported for adsorption at ultrathin films.
Introduction
in coarsening of the gold to form solid ligaments with feature
sizes ranging from nanometers to microns depending on the
initial alloy composition and the etching conditions.
The quasi-one-dimensional geometry of nanowires makes
them ideal for device applications based on the transport of
charge, such as field effect transistors (FETs). Single nanowire
FETs have been fabricated using semiconductor nanowires or
carbon nanotubes as the channel with the contacts to each end
of the wire serving as source and drain electrodes. Adsorption
of charged species on the surface of the nanowire gives rise to
In previous work we demonstrated that NPG nanowires can
be fabricated by a two-step technique involving electrochemical
template synthesis of AuxAg1-x alloy nanowires followed by
removal from the template and subsequent etching in nitric
2
2
acid. NPG nanowires exhibit high surface-to-volume ratios
and feature sizes that can be in the nanometer range. Further-
more, NPG nanowires exhibit a small footprint and hence are
attractive for sensor applications. In this paper, we use in situ
etching of AuxAg1-x alloy nanowires to fabricate single NPG
devices and report on the resistance change associated with the
adsorption of an alkanethiol.
1
a change in conductivity that can been exploited for sensing.
In the fabrication of chemical sensors in metallic systems,
finite size effects can be exploited for transduction. In bulk
metals, the resistance is dominated by lattice scattering. How-
ever, when the dimension d of a metallic conductor is smaller
than the mean free path of electrons lm then the resistance
2
becomes dominated by scattering of electrons at the surface.
Experimental Section
In this case, the electronic coupling of adsorbates to the
conduction electrons can give rise to a change in resistance.
The mean free path for most metals is on the order of a few
tens of nanometers, and hence surface scattering is significant
in films where the thickness is less than lm or in wires where
the diameter is less than lm. Ultrathin metal films have been
exploited as sensors for detection of adsorbates from both the
Figure 1 shows a schematic illustration of NPG nanowire
synthesis and device fabrication. Briefly, AuxAg1-x alloy
nanowires were fabricated by electrochemical template synthe-
sis. After removal from the template and suspension in a suitable
solvent, the nanowires were dispersed on a substrate and
electrical contacts were formed using conventional photolithog-
raphy. Finally, the silver was etched from the alloy to form
nanoporous gold nanowires.
An anodic alumina film (Whatman) with a nominal pore size
of 200 nm was used as the template for electrodeposition of
the alloy nanowires. A 600-nm-thick Cu film evaporated onto
one side of the template served as the working electrode in a
three-electrode cell with a Pt mesh counter electrode and a Ag/
AgCl (3 M NaCl) reference electrode (Ueq ) 0.200 V vs SHE).
Au0.18Ag0.82 nanowires approximately15-20 µm long were
deposited from a solution containing 100 mM AgKCN (Alfa
Aesar, 99%), 20 mM AuKCN (Alfa Aesar, 99.99%), and 0.25
M Na2CO3 (Alfa Aesar, 99%), pH 13, at -1.2 V (Ag/AgCl).22
The average diameter was 360 ( 20 nm.
3
-8
9-16
gas phase
and the liquid phase.
Resistance changes in
ultrathin metal films have been used to monitor underpotential
deposition9
alkanethiols.
,12
1
and the adsorption of halide ions
9-11
and
3-16
In the fabrication of compact sensors based on surface
scattering, the challenge is to provide a high surface area with
a relatively small device footprint. In this paper we report on a
single nanoporous gold nanowire sensor that exploits surface
scattering in the solid ligaments of the porous structure for signal
transduction. The nanoporous structure is formed by chemically
etching the less noble component from a single-phase binary
alloy containing a noble metal. For example, nanoporous gold
17-21
(NPG) is formed when Ag is etched from a AuxAg1-x alloy.
Surface diffusion of the gold during the etching process results
After deposition, the Cu film was stripped by etching in a
solution of 0.3 M CuCl2 (Alfa Aefar, 99%) and 0.1 M HCl
(Fisher Scientific, ACS grade). The alumina template was then
*
Author to whom correspondence should be addressed. E-mail: searson@
jhu.edu.
dissolved in 2 M NaOH for more than 2 h at 50-70 °C. After
1
0.1021/jp056940t CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/14/2006

Single Nanoporous Au Nanowire Sensors
J. Phys. Chem. B, Vol. 110, No. 9, 2006 4319
The alloy nanowires were etched in situ to form the
nanoporous structure. After assembling a single Au0.18Ag0.82
nanowire into a chip carrier, a drop of dilute (25 vol %) HNO3
(EMD, ACS grade) was placed on the device. Etching was
stopped by adding MilliQ water. The nanowire resistance was
monitored during the etching process.
Resistance measurements were performed by scanning the
current from -100 nA to 100 nA using a current source
(
(
Keithley 236) and measuring the corresponding voltage change
Keithley 2182 nanovoltmeter).The sensitivity of the resistance
measurement was about 0.01 Ω.
Cross sections of the NPG nanowires were prepared and
imaged using a dual-beam focused ion beam (FIB) system (FEI
NOVA dual-beam FIB). Cross sections were cut using a Ga
ion beam current of 10 pA with a constant milling time of 30
s.
Experiments on thiol adsorption were performed in the
following way. Approximately 5 mL of ethanol (Pharmca, ACS
grade) was placed on the nanowire in the chip carrier. After it
was ensured that there was no change in the resistance,
octadecanethiol (ODT, CH3(CH2)17SH, Sigma-Aldrich) was
injected into the ethanol.
Results and Discussion
Fabrication of NPG Nanowires by In Situ Etching.
Nanoporous gold nanowires are relatively brittle, and processing
steps such as washing and centrifugation can lead to fracture
and hence short wire segments. In this work the NPG nanowires
were formed by in situ etching of the silver from the Au0.18-
Ag0.82 alloy nanowires in dilute nitric acid, resulting in a device
yield that was much higher than that obtained from first forming
the NPG nanowires by etching in suspension.
Figure 1. Schematic illustration of the fabrication of a single NPG
nanowire device.
The measured resistance of single 5-8 µm long Au0.18Ag0.82
nanowire devices was 10-50 Ω. The measured resistance is
larger than the expected resistance of 1-2 Ω based on the
physical dimensions and the resistivity for Au and Ag, indicating
a contact resistance of about 5-25 Ω.
When the Au0.18Ag0.82 alloy nanowires were exposed to dilute
nitric acid, the resistance increased with the etching time, as
shown in Figure 3a. All resistance-time curves exhibited two
characteristic regimes. On introduction of the dilute nitric acid,
the resistance increased by a few ohms within the first minute.
The resistance then remained relatively constant for 4-5 min,
after which the resistance increased dramatically. Etching was
stopped when the resistance was in the range of 100-500 Ω
by rinsing with water. If etching was allowed to continue beyond
this point, the resistance continued to increase, and eventually
the wires broke, resulting in an open circuit. Figure 3b shows
a control experiment where a pure Au nanowire was exposed
to dilute nitric acid, illustrating that there is no change in
resistance.
Figure 2. (a) Optical microscope image of a chip carrier with a single
nanowire device, (b) plan view SEM image of a single nanowire
showing the gold contacts patterned by photolithography, and (c) plan
view SEM image of a NPG nanowire.
being washed in MilliQ water and centrifuged several times,
the Au0.18Ag0.82 alloy nanowires were suspended in 2-propanol
(Sigma Aldrich, ACS grade).
Single nanowire devices were fabricated on Si wafers with
5
00 nm oxide layers (Siliconquest) using photolithography and
In Figure 3a, the resistance rapidly increases by 2.2 Ω in the
first 20 s. The Au0.18Ag0.82 alloy nanowire was 6.5 µm in length
and 360 nm in diameter. Taking an average resistivity of 2 µΩ
cm (FAg ) 1.6 µΩ cm and FAu ) 2.2 µΩ cm), the resistance of
the solid alloy nanowires is determined to be about 1.3 Ω. After
20 s etching, the resistance increases by 2.2 Ω to 3.5 Ω,
equivalent to the resistance of a solid nanowire of the same
length with a diameter of 217 nm (i.e. a cross-section of 0.36
of the original nanowire). In the absence of coarsening, complete
etching of the silver would result in a cross-sectional area 0.18
of the original nanowire. Thus we can infer that about 78% of
the silver is removed in the first 20 s. The extent of the plateau
region where the resistance remains essentially constant is
then assembled into a chip carrier by wire bonding. A drop of
the nanowire suspension was dispersed on the wafer followed
by drying in N2. Photoresist (S1827, Microchem) was spin-
coated on the substrate at 3000 rpm and baked at 115 °C. After
the wafers were exposed using a mask with the lead pattern
and developed, a 200-300 nm gold layer was evaporated onto
the wafer. The photoresist was then removed, leaving the ends
of the nanowires covered by the gold leads. The gold leads were
then connected to a chip carrier (PLCC 28, Spectrum) by 25
µm in diameter gold wire. The length of the nanowires between
the contacts was typically 5-8 µm. Figure 2 shows the chip
carrier, the lead pattern, and a single nanowire.

4320 J. Phys. Chem. B, Vol. 110, No. 9, 2006
Liu and Searson
Figure 3. (a) Resistance change during in situ etching of a single Au0.18
-
Figure 5. (a) Nanowire diameter versus etching time during in situ
Ag0.82 nanowire in dilute HNO with an initial resistance of 13 Ω. The
3
etching in dilute HNO
3
and (b) the porosity versus etching time
inset shows an expanded plot of the resistance during the initial stages
of etching. (b) Resistance versus time plot for a pure Au nanowire in
determined from analysis of the cross-section SEM images.
3
dilute HNO .
2
0 min, respectively. After 1 min of etching, Figure 4c, the
cross-section image shows that the nitric acid has penetrated to
the center of the nanowire and that the morphology is relatively
uniform. From Figures 4d and 4e, it is evident that the porosity
continues to increase with etching time, indicating that Ag is
still being removed from the nanowire.
Figure 5a shows the wire diameter plotted versus etching time,
illustrating that etching and coarsening continue for the first 10
min of etching. After 20 min of etching, energy dispersive
spectroscopy in the transmission electron microscope revealed
that almost no Ag remained in the nanowire, indicating that
the etching process is complete. The continued coarsening
eventually results in fracture since both ends of the nanowire
are constrained by the contact electrodes.
Figure 4. (a) SEM cross-section image of (a) a pure gold nanowire
and (b) a NPG nanowire. Cross-section SEM images of Au0.18Ag0.82
Figure 5b shows the porosity of the nanowires estimated from
analysis of the cross-section images. The boundaries of the solid
regions were first identified by thresholding the images and the
porosity and then determined from the fractional area. This
figure shows that the porosity increases dramatically during the
first minute as a large amount of silver is etched from the alloy.
Both etching and coarsening occur during the first 10 min. This
is also evident from the fact that coarsening alone would result
in a decrease in porosity. After 10 min of etching, both the
diameter and the porosity exhibit a relatively small change,
indicating that most of the Ag has been etched and that the rate
of coarsening is significantly reduced.
3
nanowires after in situ etching in HNO for (c) 1 min, (d) 5 min, and
(e) 20 min.
dependent on the nitric acid concentration. In concentrated nitric
acid, the resistance increases very rapidly to the break point,
and hence the etching process is difficult to control. Thus by
performance of the etching process in situ it is possible to control
the morphology and resistance of the NPG nanowires
Characterization of the Porous Structure. Figure 4a shows
a cross-section image of a Au0.18Ag0.82 alloy nanowire, illustrat-
ing that the nanowires are solid with a diameter of about 360
nm. Figure 4b shows a cross-section image of an NPG nanowire
after etching, illustrating the characteristic structure of nanopo-
rous gold. During the etching process the silver atoms are
removed, leaving gold-rich regions. Diffusion of the gold atoms
at the surface is relatively fast, resulting in coarsening of the
structure. The coarsening of the remaining gold results in a
significant decrease in the wire diameter, as can be seen from
the scanning electron microscope (SEM) images.
We can compare the porosity at long times with the value
estimated from the wire diameter in Figure 5a. After 20 min of
etching the wire diameter is 230 nm. Assuming that all of the
silver is etched from the alloy, then we estimate an average
porosity of 0.56, which is good agreement with the value, 0.65
( 0.08, obtained from analysis of the image in Figure 4e.
We can now compare the resistance change during etching
with the wire diameter and porosity. The SEM images show
that during the first minute the nitric acid has penetrated to the
center of the wires, producing a uniform morphology with a
porosity of 0.53 ( 0.04. As described above the resistance
Figures 4c-e show cross-section images of the porous
structure after in situ etching in dilute nitric acid for 1, 5, and

Single Nanoporous Au Nanowire Sensors
J. Phys. Chem. B, Vol. 110, No. 9, 2006 4321
Figure 7. Normalized resistance change versus time for a single NPG
nanowire on exposure to 10 µL of 1 mM ODT. Note that the ODT
concentration was in excess of the amount needed to form a monolayer
at the surface.
Figure 6. (a) Intensity versus q plot obtained from transformation of
the cross-section SEM images. The etching time is indicated in the
figure. (b) Characteristic feature size associated with the solid ligaments
in the NPG nanowires plotted versus etching time.
change that occurs in the first 20 s corresponds to etching of
about 78% of the silver. Taking into account that the diameter
has decreased to 350 nm, the reduction in area corresponds to
a porosity of 0.61, in reasonable agreement with the value
obtained from the SEM image. After the initial increase, the
resistance remains essentially constant for the next 4-5 min.
During this time coarsening is relatively fast, resulting in a
significant decrease in the diameter of the nanowires. The
remaining silver is also etched during this time so that there is
little change in the porosity. At longer times coarsening
continues, ultimately resulting in fracture of the solid ligaments
in the porous structure and hence a rapid increase in the
resistance.
Analysis of the cross-section images can be used to obtain
the average ligament size in the porous structure. Fast fourier
transforms (FFTs) of the cross-section images can be used to
obtain the ligament size of the porous structure. Figure 6 shows
intensity versus q plots obtained from two-dimensional FFTs
of the cross-section images after thresholding. The peaks in the
plots indicate the characteristic feature size in the nanoporous
structure as a function of etching time. The mean free path for
electrons in gold is about 50 nm,23 significantly larger than the
average ligament size of around 10 nm, as can be seen in Figure
Figure 8. Resistance versus time for a single NPG nanowire device.
The arrows indicate the injection of aliquots of 5 × 10 mol ODT.
-11
than the mean free path of electrons. We note that the resistance
of the NPG nanowires is typically 100-500 Ω, significantly
larger than the contact resistance, and hence the resistance
change is dominated by the change in surface scattering.
The change in resistance is related to the increase in diffusive
scattering at the surface due to the presence of adsorbates. At
an ideal surface, all electrons are specularly reflected (elastically
scattered with no change in momentum in the direction of the
electric field). The presence of adsorbates or other defects can
lead to diffusive (inelastic) scattering of electrons that is
associated with a change in momentum. Thus at a real surface,
as long as the fraction of specularly reflected electrons is
sufficiently large, then adsorption will introduce scattering
centers that will increase the fraction of diffusively scattered
electrons. If the dimensions of the solid are smaller than the
electron mean free path such that the resistance is dominated
by surface scattering, then adsorption can lead to a measurable
change in resistance.
Figure 8 shows an experiment where small aliquots (5 ×
-11
1
0
mol) were sequentially injected into the ethanol on the
6
b. These dimensions are in good agreement with feature sizes
chip carrier. The initial resistance of the nanowire is 236 Ω,
and each aliquot results in a 1 Ω increase in the resistance.
Figure 9 shows the resistance change plotted versus the ODT
surface concentration. The surface concentration was determined
from the amount of ODT at the onset of saturation of the
seen in transmission electron microscope images (not shown).
Single NPG Nanowire Devices. Figure 7 shows the normal-
ized resistance change for a single NPG nanowire on introducing
1
0 µL of 1 mM ODT. This concentration of ODT is in excess
of the amount needed to form a self-assembled monolayer at
the surface. The resistance increases by 3% over a period of
about 1 min, corresponding to the time for the formation of a
monolayer at a planar surface.24 The response time and
14
-2 24
resistance and taking monolayer coverage as 4.6 × 10 cm .
In experiments in the gas phase with ultrathin films, the
resistance increases linearly with coverage and then either
saturates or reaches a maximum that is dependent on the film
thickness, the electron mean free path, and the binding energy
resistance change are comparable to values reported for thiol
adsorption on ultrathin gold films.1
3-16
Although not demon-
25
of the adsorbate on the metal surface. At the low coverage,
strated here, experiments on ultrathin gold films have shown
that the initial resistance can be recovered by exposure to ozone
and ethanol.15
This result demonstrates that the NPG nanowire consists of
an interconnected network of ligaments with dimensions less
the linear relationship can be expressed as
d(∆R/R)
[ d(dn)
Al
m
)
(1)
]
d
nf0

4322 J. Phys. Chem. B, Vol. 110, No. 9, 2006
Liu and Searson
sensors for adsorption of octadecanethiol resulted in a sensitivity
-16
2
of 1.0 × 10 cm , comparable to values reported for adsorption
at ultrathin films.
Acknowledgment. This work was supported by the JHU
MRSEC (National Science Foundation Grant No. DMR05-
2
0491).
References and Notes
(
1) Cui, Y.; Wei, Q. Q.; Park, H. K.; Lieber, C. M. Science 2001, 293,
1
289-1292.
(
2) Sondheimer, E. H. AdV. Phys. 1952, 1, 1-42.
(3) Chopra, K. L.; Bobb, L. C.; Francombe, M. H. J. Appl. Phys. 1963,
Figure 9. Normalized resistance change versus the surface concentra-
34, 1699-1702.
(4) Persson, B. N. J. Phys. ReV. B 1991, 44, 3277-3296.
(5) Persson, B. N. J. J. Chem. Phys. 1993, 98, 1659-1672.
6) Hein, M.; Schumacher, D. J. Phys. D: Appl. Phys. 1995, 28, 1937-
941.
7) Krastev, E. T.; Kuhl, D. E.; Tobin, R. G. Surf. Sci. 1997, 387,
tion of ODT derived from Figure 8. Monolayer coverage corresponds
1
4
-2
to a surface concentration of 4.6 × 10 cm
.
(
1
where the ∆R/R is the normalized resistance change, n is the
number of adsorbate molecules per unit area, d is the film
thickness, and A is a constant. The slope of the linear region
represents a characteristic parameter for a given system.
(
L1051-L1056.
(8) Cabrera, A. L.; Garrido Molina, W.; Colino, J.; Lederman, D.;
Schuller, I. K. Phys. ReV. B 1997, 55, 13999-14004.
(
9) Dickinson, T.; Sutton, P. R. Electrochim. Acta 1974, 19, 427-
-
16
2
From Figure 9 we determine a slope of 1.0 × 10
comparison, a slope of 0.2 × 10 cm has been reported for
ODT adsorption on ultrathin gold films. A sensitivity factor
cm . In
435.
-
16
2
(10) Hansen, W. N. Surf. Sci. 1980, 101, 109-122.
14
(11) Rath, D. L.; Hansen, W. N. Surf. Sci. 1984, 136, 195-216.
(12) Hanewinkel, C.; Winkes, H.; Schumacher, D.; Otto, A. Electrochim.
-
16
2
of 2 × 10
cm has been reported for adsorption of CO on
Acta 1997, 42, 3345-3349.
(13) Zhang, Y.; Terrill, R. H.; Bohn, P. W. J. Am. Chem. Soc. 1998,
120, 9969-9970.
-
16
2
1
0 nm Cu films, and a slope of 0.5 × 10
cm has been
25
reported for adsorption of CO on 10 nm Ni films. These results
show that the single NPG nanowire devices have sensitivities
comparable to relatively large area thin films.
(
14) Zhang, Y.; Terrill, R. H.; Bohn, P. W. Chem. Mater. 1999, 11,
2
1
4
7
191-2198.
(15) Zhang, Y. M.; Terrill, R. H.; Bohn, P. W. Anal. Chem. 1999, 71,
19-125.
(16) Fried, G. A.; Zhang, Y. M.; Bohn, P. W. Thin Solid Films 2001,
Conclusions
01, 171-178.
17) Pickering, H. W.; Wagner, C. J. Electrochem. Soc. 1967, 114, 698-
06.
(
Single NPG nanowire sensors were fabricated by in situ
etching of Au0.18Ag0.82 alloy nanowires using simultaneous
resistance measurements to monitor the etching process. We
show that most of the silver is etched during the first 20 s and
that subsequent coarsening and etching result in a significant
decrease in the nanowire diameter. The final porosity of the
devices is about 0.65. The characteristic feature size of the
porous structure is about 10 nm, smaller than the mean free
path of electrons in bulk gold so that the resistance is dominated
by surface scattering. The performance of the nanowires as
(
18) Forty, A. J. Nature 1979, 282, 597-598.
(19) Li, R.; Sieradzki, K. Phys. ReV. Lett. 1992, 68, 1168-1171.
(
20) Yoon, J.; Chan, M. H. W. Phys. ReV. Lett. 1997, 78, 4801-4804.
(21) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K.
Nature 2001, 410, 450-453.
(22) Ji, C. X.; Searson, P. C. J. Phys. Chem. B 2003, 107, 4494-4499.
(23) Tanner, B. K. Introduction to the Physics of Electrons in Solids;
Cambridge University Press: Cambridge, U. K., 1995.
(
24) Schlenoff, J. B.; Li, M.; Ly, H. J. Am. Chem. Soc. 1995, 117,
1
2528-12536.
(25) Wissman, P. Tracts Mod. Phys. 1975, 77, 1-96.