ARTICLE

Hybrid Metal
Semiconductor
Nanostructure for Ultrahigh Optical
Absorption and Low Electrical
Resistance at Optoelectronic Interfaces
Vijay K. Narasimhan,† Thomas M. Hymel,† Ruby A. Lai,‡ and Yi Cui*,†,§
†
Department of Materials Science and Engineering, Stanford University, 476 Lomita Mall, Stanford, California 94305, United States, ‡Department of Physics,
Stanford University, 382 Via Pueblo, Stanford, California 94305, United States, and §Stanford Institute for Materials and Energy Sciences, SLAC National Accelerator
Laboratory, 2575 Sand Hill Road, Menlo Park, California 94025, United States

ABSTRACT Engineered optoelectronic surfaces must control both the flow of light and the flow of
electrons at an interface; however, nanostructures for photon and electron management have typically
been studied and optimized separately. In this work, we unify these concepts in a new hybrid
metal
semiconductor surface that offers both strong light absorption and high electrical conductivity.
We use metal-assisted chemical etching to nanostructure the surface of a silicon wafer, creating an array of
silicon nanopillars protruding through holes in a gold film. When coated with a silicon nitride anti-
reflection layer, we observe broad-band absorption of up to 97% in this structure, which is remarkable
considering that metal covers 60% of the top surface. We use optical simulations to show that Mie-like
resonances in the nanopillars funnel light around the metal layer and into the substrate, rendering the
metal nearly transparent to the incoming light. Our results show that, across a wide parameter space, hybrid metal
semiconductor surfaces with
absorption above 90% and sheet resistance below 20 Ω/0 are realizable, suggesting a new paradigm wherein transparent electrodes and photon
management textures are designed and fabricated together to create high-performance optoelectronic interfaces.

KEYWORDS: photon management . light trapping . Mie resonators . transparent conductors . metal nanowire networks .
metal-assisted chemical etching

A
variety of nanostructured surfaces,
including nanopillars,1 nanocones,2
nanowires,3
6 inverted pyramids,7
nanospheres,8 nanoshells,9 and nano-
domes,10 have enabled strong, broad-band
absorption in semiconductors for optoelec-
tronic applications. Dielectric nanostruc-
tures are extremely effective at coupling
and directing light across an interface be-
tween the outside world and optically active
materials owing to a rich set of optical
phenomena, including gradual refractive
index matching, scattering, and coupling
to guided modes.11,12 However, in many
applications, a nanostructured optical inter-
face must also serve as a conduit for elec-
trical current between the semiconductor
and an external circuit. In addition to pro-
viding photon management, the interface
must offer high electrical conductivity.
Typically, electrical conductivity at this
interface has been provided by depositing
a separate transparent electrode over the
photon management structure, which re-
duces resistance while letting light pass
through to underlying layers. Sputtered
indium-doped tin oxide (ITO) is the most
common transparent conductor currently
used; however, sputtering ITO evenly over
highly textured surfaces can be challenging,
and the thick layers needed for applications
that require very low resistance result in
long sputtering times and thus lower device
throughput.13 Metal nanowire networks are
a promising alternative.14
16 However, when
high surface conductivity is desired, the
concomitantly high metal surface coverage * Address correspondence to
yicui@stanford.edu.
required can result in significant reflection
loss and thus decreased absorption in the Received for review July 1, 2015
active layer.17 Thus, the surface coverage of and accepted October 8, 2015.
metal networks is limited to about 10% for
Published online
practical applications.18 10.1021/acsnano.5b04034
Despite the need for both strong ab-
sorption and high conductivity at many C XXXX American Chemical Society

NARASIMHAN ET AL. VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX A

www.acsnano.org
Figure 1. Comparison of perforated gold films on silicon substrates with and without silicon nanopillars protruding through
the holes. (a) Schematic, (b) false-color 25° tilted-view SEM (scale bar = 500 nm, with gold in yellow), and (c) digital photograph
(scale bar = 1 mm) for a 16 nm thick gold grid sitting atop a planar silicon substrate. The surface, which is 65% covered by
ARTICLE
metal, is highly reflective. (d) Schematic, (e) false-color 25° tilted-view SEM micrograph (scale bar = 500 nm, with gold in
yellow), and (f) digital photograph (scale bar = 1 mm) for an identical 16 nm thick gold film with silicon nanopillars protruding
through the holes. The patterned area is dark red, indicating a significant decrease in reflection. (g) Angle-integrated
reflection spectra of perforated metal films with (dashed blue trace) and without (dotted red trace) protruding silicon
nanopillars. The perforated gold film on the planar substrate exhibits 50% integrated reflection averaged over the solar
spectrum, while the film with protruding silicon nanopillars has only 19% reflection. The reflection can be further reduced to
10% by coating the metal
silicon nanopillar sample with a 50 nm silicon nitride antireflection layer (solid black trace).

optoelectronic interfaces and the commensurate size
scale of semiconductor photon management struc-
tures and metal nanowire networks, these nanostruc-
tured materials have been studied and developed in
relative isolation. In particular, the transparency of
metal nanowire networks is optimized and measured
considering the network in isolation, that is, suspended
in free space or sometimes on a transparent substrate.
At real interfaces with nanostructured semiconductors,
the profile of the light, and therefore the performance
of the network, could differ significantly. Here, we
demonstrate that unifying a metal nanowire grid
with extremely high surface coverage and a photon
management texture in a single optoelectronic inter-
face can simultaneously improve optical absorption
and electrical conductivity, overcoming the conven-
tional trade-offs encountered when designing these
structures separately.

NARASIMHAN ET AL. VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX B

www.acsnano.org

Our design is based on nanopillars with small aspect
ratios (∼1:1). These structures have been shown to act
as broad-band light “funnels” by strongly coupling and
confining light into Mie-like resonances that leak into
the substrate through the small volume beneath each
pillar.19
21 By aligning nanopillars protruding from
a substrate through the holes in a metal film, we
expected that most of the electric field intensity would
be concentrated in the nanopillars and thus not over
areas of the substrate covered by the metal. Therefore,
we hypothesized that even in metal films with high
surface coverage and small holes, the introduction of
dielectric nanopillars would substantially reduce the
reflection losses from the metal, allowing light to be
easily absorbed in the underlying substrate.
RESULTS
We began by examining the optical properties of a
high-surface-coverage metal film perforated with
nanoscale holes sitting on top of a planar silicon sub-
strate. Figure 1a,b shows a schematic and the scanning
electron micrograph (SEM) of a 16 nm thick gold film
patterned with an array of 295 nm  295 nm square
holes with a center-to-center spacing of 500 nm. The
patterned film on the planar silicon substrate appears
shiny and reflective, only slightly darker than the
unperforated gold areas surrounding it (Figure 1c). This
appearance is expected considering the high surface
coverage of the metal: 65% of the silicon surface is
covered with gold.
Next, we examined the properties of an identical
gold film with silicon nanopillars protruding from the
substrate through the holes. We fabricated this struc-
ture with metal-assisted chemical etching (MACE),22
using the patterned gold film to anisotropically etch
the silicon. This process is described in more detail
in the Methods section and shown schematically in
Figure 2. The gold acts as a catalyst for the etching of
silicon, and thus the gold film sinks into the substrate,
creating an array of nanopillars that protrude through
holes in the film. The silicon nanopillars are thus
automatically aligned to the holes in the metal film in
a single step.
Figure 2. Metal-assisted chemical etching process. (a) Process flow illustrating the key steps for fabricating a gold film with
protruding silicon nanopillars using metal-assisted chemical etching. A silicon substrate with a patterned gold film is
immersed into a solution of hydrofluoric acid and hydrogen peroxide to create the nanopillar array. (b) Detailed schematic of
the etching step. The silicon directly beneath the gold layer is oxidized and attacked, resulting in the metal layer sinking into
the substrate. The silicon nanopillars are thus self-aligned with the holes in the metal film.
NARASIMHAN ET AL.
ARTICLE
Figure 1d,e shows a gold film identical to that shown
in Figure 1a,b through which an array of silicon nano-
pillars protrudes from the substrate 330 nm above the
metal film. The change in the optical properties of this
surface is dramatic; the patterned area in the etched
sample with protruding nanopillars is a deep red color
(Figure 1f).
We measured the combined specular and diffuse
reflection (R) of the gold
silicon surfaces at different
angles of incidence. The reflection spectra of the two
samples confirms the dramatic change conferred by
the nanopillar array (Figure 1g): the perforated film
on the planar substrate reflects, on average, 50% of
the incident light, whereas the etched sample with
protruding silicon nanopillars reflects only 19% of the
light. By adding a 50 nm thick silicon nitride anti-
reflection coating, the reflection of the nanopillar
sample decreases further, to 10%. Considering that
the gold film still occupies 65% of the top surface of the
silicon in these samples, the magnitude of the change
in the reflectance is surprising.
To explain the remarkable optical properties of the
film, we next performed detailed simulations of a
model of our structure using S4, a freely available fre-
quency modal method software package.23 Figure 3a
shows the simulated reflection spectra for a silicon
substrate with a metal film with 300 nm wide holes on a
500 nm spacing with a 15 nm thick gold layer, with and
without protruding silicon nanopillars. The surface
with nanopillars protruding 330 nm above the metal
demonstrates much less reflection, confirming our
experimental observations. Moreover, Figure 3a also
shows the simulated reflection spectra for a structure
coated with a 50 nm thick silicon nitride layer, both
with and without the embedded metal film. The
difference is negligible (2%), suggesting, as hypothe-
sized, that the presence of the metal does not have a
significant impact on the optical properties of this
architecture. We also simulated the electric field pat-
terns in the structures. Field patterns at an incident
wavelength of 550 nm are also shown in Figure 3. While
the patterned metal film sitting on top of a planar silicon
surface strongly reflects incoming light (Figure 3b), the
VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX C
www.acsnano.org

Figure 3. Detailed optical simulations showing leaky Mie-
like resonances in gold films with protruding silicon nano-
pillars. (a) Simulated normal incidence absorption as a
function of wavelength for the perforated metal film sitting
atop a planar silicon surface (dotted red trace), the metal
film with protruding silicon nanopillars (dashed blue trace),
the perforated film with protruding silicon nanopillars
covered with a 50 nm silicon nitride anti-reflection coating
(solid black trace), and a gold-free silicon nanopillar array
with a 50 nm silicon nitride coating (dashed black trace). In
these simulations, the metal film has 300 nm  300 nm
holes on a 500 nm spacing, and the gold is 15 nm thick. The
nanopillars also measure 300 nm  300 nm laterally and
protrude 330 nm through the holes in the metal film. The
surfaces with protruding nanopillars have a simulated
reflection lower than that of the perforated films on planar
substrates, and the effect of the metal layer on the reflec-
tance spectrum is negligible. (b) Cross section (through the
center of a hole) of the simulated electric field intensity with
550 nm incident light for the perforated metal film sitting
atop a planar silicon surface (scale bar = 100 nm). Silicon and
gold outlines are shown for clarity. Light is strongly re-
flected from the film. (c) Cross section (through the center of
a pillar) of the simulated electric field intensity with 550 nm
incident light for the perforated gold film with protruding
silicon nanopillars with a silicon nitride coating (scale bar =
100 nm). Silicon, gold, and nitride outlines are shown for
clarity. A Mie-like resonance confines light inside the pillar,
and the light leaks from the pillar into the substrate,
circumventing the metal film.
nanopillar structure confines light within a Mie-like
resonance in the pillar that leaks into the substrate
and circumvents the metal almost entirely, as pre-
dicted (Figure 3c).
We performed further simulations to explore the
optical properties of structures with different geo-
metries and materials. One of the most interesting
findings from these simulations was the dramatic
effect of the nanopillar height on the optical properties
of the structure. As shown in Figure 4, a pillar array that
protrudes even 50
100 nm above a metal film is
NARASIMHAN ET AL.
ARTICLE
sufficient to greatly suppress reflection as compared
to simply filling the holes in the metal with silicon. We
observed this finding experimentally during the etch-
ing process. Video S1 in the Supporting Information
shows that the surface covered by the patterned metal
film undergoes a significant color change from shiny
gold to red to dark purple as the etching reaction
proceeds and the nanopillars begin to protrude
through the metal surface.
Other results from our simulations are shown in
Figure S1. We found that integrated absorption above
90% is possible even with metal layers 50 nm thick,
suggesting that highly absorptive optoelectronic
surfaces with extremely high conductivity could be
manufactured. Simulations of the width and spacing of
the pillars showed that the minimum reflection values
correspond to nanopillar arrays with a surface cover-
age of just less than 50%. Also, high absorption from
metal
semiconductor hybrid interfaces is expected in
a number of other optical material systems, including
gallium arsenide and indium phosphide. While the
specific spectral features in each absorption spectrum
generally shift with increasing size in any dimension,
our simulations show that high absorption in substrates
covered by metal films with protruding dielectric
nanopillars is possible across a variety of geometries
and materials.
With these insights in hand, we fabricated three
additional samples, varying the width of the gold lines
and the etching time to produce metal
silicon nano-
pillar hybrid surfaces with different geometries, as
shown in Figure S2. Figure 5a
c shows an electron
micrograph, the 5° off-normal incidence absorption
spectrum (the smallest angle measurable in the inte-
grating sphere), and average absorption as a function
of angle of incidence for our best sample (a digital
photograph is provided in Figure S3). This structure has
pillars that are 320 nm wide by 150 nm tall with a
500 nm center-to-center spacing. Remarkably, with a
50 nm thick silicon nitride anti-reflection layer, the
structure has an average angle-integrated absorption
of 97% over the standard reference solar spectrum.24
Further, absorption above 90% is preserved even to
angles of incidence as high as 40°. From simulations of
the electric field distribution in this structure, we
deduce that only 4.2% of the incoming light is ab-
sorbed in the metal. Owing to the modal profile of the
resonances in the nanopillars, almost all of the interface's
absorption takes place in the silicon (see Figure S4).
Despite the fact that 60% of the top surface of the
silicon is covered with a metal film, the structure dem-
onstrates strong, broad-band absorption across the
visible spectrum and at a variety of angles of incidence.
All four hybrid surfaces we fabricated had high
conductivity while also demonstrating extremely high
absorption. Each sample had a sheet resistance of less
than 20 Ω/0 (compared to 10 kΩ/0 for the silicon
VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX D
www.acsnano.org
 ARTICLE
Figure 4. Effect of the nanopillar height on simulated
reflection. The simulated short-wavelength reflectance
from a perforated metal film with silicon filling its holes
(dotted red trace) is still very high, whereas reflection drops
significantly if the silicon protrudes 50 nm (dashed blue
trace) or 100 nm (solid black trace) above the metal surface.
For these simulations, we assumed 200 nm square holes
with a 450 nm center-to-center spacing in a 50 nm thick
metal film. All structures were assumed to be covered with a
50 nm thick silicon nitride coating.

substrates with the native oxide removed). These

values were possible owing to the high metal coverage
of over 40%. Despite the high metal coverage, all the
samples demonstrated more than 89% absorption
averaged over the standard reference solar spectrum24
from 400 to 900 nm (compared to 61% for the bare
silicon substrate). In fact, the performance of the
optoelectronic interfaces we designed is equivalent
to placing the highest performing transparent elec-
trodes from the literature atop semiconductors coated
with an ideal anti-reflection coating (see Figure S5).
To demonstrate the potential of a metal
silicon
hybrid nanostructure in an optoelectronic device, we
compare the front surface reflectance between 400 and
900 nm of the interface in Figure 5 with the reported
front surface reflectance of a PERC-type silicon solar
cell with front metal contacts25 in Figure S6. The Figure 5. Optical characterization of the best performing
reference silicon solar cell shows a front surface reflec- hybrid metal
silicon nanopillar interface. (a) Tilted-view
SEM micrograph (scale bar = 500 nm). In this structure, the
tance of 6.5% averaged over the standard reference nanopillars are 320 nm wide with a 500 nm center-to-center
solar spectrum compared to 2.8% from the gold
spacing and protrude 150 nm from the 16 nm thick gold
silicon hybrid structure. Remarkably, despite its lower film. (b) 5° Off-normal absorption spectrum for bare silicon
(dotted red trace), the as-made gold
silicon nanopillar
reflectance, the metal
silicon hybrid interface has surface shown in (a) (dashed blue trace), and the gold

close to 12 times the metal coverage of the reference silicon nanopillar surface covered with a 50 nm silicon
solar cell (60% versus 5.1%). Even including the calcu- nitride coating (solid black trace). (c) Absorption of the
gold
silicon nanopillar hybrid surface with a 50 nm silicon
lated 4.2% absorption loss in the metal, the optical nitride coating as a function of incident light angle, aver-
losses of the hybrid structure are comparable to those aged over the standard hemispherical reference solar spec-
of the silicon solar cell (the absorption losses of the trum.24 Absorption as high as 97% is observed in this
sample, and absorption above 90% is preserved even at
front metal contacts were not reported directly for high angles of up to 40° from the normal even though 60%
the reference cell). The advantage of enabling higher of the top surface is covered with gold.
metal coverage in a solar cell while maintaining similar
optical performance is that the series resistance could lines, this resistance (0.1 Ω cm2) could essentially be
be substantially reduced. For example, the lateral eliminated. With an appropriate metal contact stack,
resistance in the emitter comprises one-sixth of the the nearly continuous metal coverage of the metal

total series resistance in the reference PERC solar cell; silicon hybrid interface could thus result in a higher fill
by providing a high-coverage grid with nanoscale factor while maintaining the same short circuit current.

NARASIMHAN ET AL. VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX E

www.acsnano.org

However, the contact-emitter resistance and the surface
passivation of the nanostructure would need to be
carefully considered to ensure improved efficiency.
DISCUSSION
There are a variety of applications in which high-
coverage metal films with nanoscale holes would be
desirable to provide dense, highly conductive electrical
pathways to a semiconductor. These include electro-
des for photosensors, solar cells, light-emitting diodes
(LEDs), and displays;26 conductive patches for trans-
parent antennas;27 interconnects for fully transparent
electronics;28 and even current collectors for transpar-
ent batteries.29 However, to date, there has been no
way to cover a semiconductor surface with a high
fraction of metal without incurring significant reflec-
tion loss at the interface. Only plasmonic resonances
have been extensively studied for achieving extra-
ordinary transmission through high-coverage metal
films,30
32 and these resonances are very sensitive
to the shape and size of the holes.33
35 Plasmon-
mediated extraordinary transmission is fairly narrow-
band and polarization-sensitive; even state of the art
structures report only 50
80% transmission36,37 for
certain colors and greatly suppressed transmission in
the rest of the spectrum. By contrast, we have shown
through simulations and experiments that the strong
enhancement of transmission above 90% through
metal films created by resonances in semiconductor
nanopillars is universal across a wide parameter space
of geometries and optically important material sys-
tems. While fabrication complexity, ohmic contact,
junction resistance, and impurities would all have to
be carefully considered, our results provide strong
evidence that metal films with protruding dielectric
nanostructures could make a compelling candidate
system for high-performance interfaces in a variety of
optoelectronic applications.
Moreover, the process we have designed to fabri-
cate metal
semiconductor hybrid nanostructures is
very versatile. The metal grid is automatically self-
registered to the nanopillar array, so no subsequent
patterning or alignment is required after the formation
METHODS
Fabrication of the Metal Film with Protruding Dielectric Nanopillars by
Metal-Assisted Chemical Etching. To fabricate the structure, we
started with silicon substrates (n-type, 1
10 Ω cm, 36

350 μm thick). We removed the native oxide by immersion in
a 2% hydrofluoric acid solution. We then diluted ma-N 2405
negative-tone electron-beam lithography resist (Micro Resist
Technology) 1:1 with mr-T 1090 thinner (Micro Resist
Technology) and spin-coated it onto the silicon pieces at
4000 rpm for 40 s. After spin-coating, we baked the sample
on a hot plate at 90 °C for 1 min and exposed an array of square
areas with a dose of 240 μC/cm2 in a JEOL JBX-6400FS electron-
beam lithography system with a 100 keV beam energy.
NARASIMHAN ET AL.
ARTICLE
of the photon management structure. All processing
occurs at low temperatures, below 90 °C, and the
dramatic optical change associated with etching could
be useful for end point detection or for monitoring
process uniformity. MACE can be used to etch many
metal
semiconductor systems, including single-crystal
and polycrystalline silicon38 and III
V substrates39 using
aluminum, silver, platinum, palladium, rhodium, and
gold catalysts (or combinations thereof).40,41 Etching
with thicker metal films, which would demonstrate
lower resistivity, is also possible.42
Further, our simulations and experiments have de-
monstrated a wide design space for strong absorption
and high conductivity in metal
semiconductor hybrid
nanostructures, which suggests that the structure
could tolerate polydispersity in the nanopillar geome-
try, making scalable random patterning of the metal
layer, for example, by nanosphere lithography43 or
metal dewetting,44 a viable option. It is also possible
that by tuning the pattern of the metal grid and the
photonic properties of the dielectric resonators, the
architecture could harness photonic crystal effects and
be useful not only for high absorption and coupling but
also for extraordinary reflection, for example, as an
intermediate reflector in a tandem solar cell45 or as a
light-extracting structure for LEDs.46
CONCLUSION
Our simulations and experiments have confirmed
that metal
semiconductor nanopillar hybrid structures
can exhibit high conductivity with robust, broad-band
absorption across a very wide parameter space of
materials and geometries. In our best sample, we
observed 97% average absorption with a sheet resis-
tance of only 16 Ω/0. More importantly, we have
shown that in a hybrid optoelectronic interface, the
optical absorption profile of the photon management
structure can be used to greatly enhance transmission
through a high-coverage metal nanowire network.
These results suggest a new paradigm wherein trans-
parent electrodes and photon management textures
should be designed and fabricated together to create
high-performance optoelectronic interfaces.
We developed the pattern by immersion in Microposit
MF-319 (DOW Chemical) for 30 s, followed by a water rinse
and drying with a nitrogen gun.
We next prepared the sample for wet etching by clean-
ing it in an RF plasma system for 10 s (Plasma-Prep, 2% O2
and 98% Ar) to descum the resist and leave the surface
of the silicon oxygen-terminated. We evaporated a nano-
porous 16 nm thick gold film onto the sample at 0.5 Å/s
in a Kurt J. Lesker LAB 18 electron-beam evaporator
system.
We then immersed the patterned, metalized silicon sub-
strate into a solution of hydrofluoric acid (12.04 M) and hydrogen
peroxide (0.82 M) for 6
10 s. Though there is some disagree-
ment as to the exact mechanism of MACE, a proposal that
VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX F
www.acsnano.org

explains most experimental observations is that the reaction
proceeds by a two-step mechanism:22
H2 O2 þ 2Hþ f 2H2 O þ 2hþ (1)
Si þ nhþ þ 6HF f H2 SiF6 þ nHþ þ 0:5(4
n)H2 v (2)
In the first step, the hydrogen peroxide is decomposed to
water at the metal surface and in the process transfers holes to
the silicon via the metal. Next, the hole-rich silicon, now in an
oxidized state, reacts with the hydrofluoric acid to form the
soluble species H2SiF6 and hydrogen gas. Etching proceeds
anisotropically directly beneath the metal, and the metal sinks
into the nanostructure, resulting in a patterned metal film with a
self-aligned protruding array of nanopillars in a single step.
Metal-assisted chemical etching is possible with both thin47 and
thick metal films.42
We rinsed the samples with deionized water after etching.
We then removed the resist using heated N-methyl pyrrolidone
and heated acetone, sometimes with sonication. For some
samples, we used a PlasmaTherm Versaline high-density plas-
ma-enhanced chemical vapor deposition system to perform a
low-temperature (90 °C) deposition of silicon nitride.
Optical Measurements. We used an integrating sphere setup
(painted with Spectraflect paint) with a white light source and
monochromator, as reported elsewhere,10 to measure the total
(specular and diffuse) reflection spectra of the gold
silicon
surfaces at different angles of incidence. The reflectance was
measured over angles of incidence ranging from 5 to 45° such
that specular reflection from the sample would not escape the
integrating sphere through the sphere's aperture. Transmission
through the samples was below the measurement threshold of
the setup. The average reflectance or absorptance was calcu-
lated as a weighted average over the standard reference solar
spectrum24 from 400 to 900 nm.
Optical Simulations. We conducted electromagnetic simula-
tions using S4, a free frequency modal method solver.23 We
used standard refractive index values from two online sources,
namely, http://refractiveindex.info/ for the metals and silicon
nitride and http://www.filmetrics.com/refractive-index-database/
for the semiconductors (Si, GaAs, and InP). We used a plane wave
source with equal parts s- and p-polarized light. We used
24 spatial frequency modes in most simulations to calculate
reflection spectra and the fields in our structure with enough
resolution. For finer resolution in some simulations, we in-
creased the number of modes to 37 or 100. Data points were
taken at wavelengths 10 nm apart from 400 to 900 nm. Fields
were extracted on a 5 nm grid spacing.
Electrical Measurements. Our structure included two large gold
pads electrically connected to each side of the patterned metal
film. We applied a conductive adhesive to each pad and used a
two-point probe measurement to measure the resistance (R) of
the sample. The sample size was measured in an optical
microscope, and the sheet resistance (Rs) was calculated using
the expression Rs = R  W/L, where W and L are the width of the
pads and the distance between the pads, respectively.
Conflict of Interest: The authors declare no competing
financial interest.
Supporting Information Available: The Supporting Informa-
tion is available free of charge on the ACS Publications website
at DOI: 10.1021/acsnano.5b04034.
Video of the MACE process (AVI)
Additional optical simulations and all data collected from
fabricated samples (PDF)
Acknowledgment. This material is based upon work sup-
ported by the Department of Energy through the Bay Area
Photovoltaic Consortium (BAPVC) under Award Number
DE-EE0004946 and was also supported by the Stanford Global
Climate and Energy Project (GCEP). The authors would like to
acknowledge the support of the staff at the Stanford Nano
Shared Facilities and the Stanford Nanofabrication Facility, espe-
cially R. Tiberio and C. Knollenberg. V.K.N. would like to thank
the International Fulbright Science and Technology Fellowship
NARASIMHAN ET AL.
ARTICLE
Program for support, and he would also like to thank P.-C. Hsu
and I. Karakasoglu for helpful discussions. V.K.N. conceived the
structure. V.K.N., T.M.H., R.A.L., and Y.C. designed and reviewed
results from the experiments. V.K.N. conducted the etching, AR
coating, and characterization experiments, analyzed the results,
and performed electromagnetic simulations. T.M.H. and R.A.L.
provided some silicon wafers. V.K.N. and T.M.H. optimized the
gold thickness and surface treatment for the MACE process and
performed sheet resistance measurements. V.K.N. prepared the
figures with layout and design assistance from R.A.L. and T.M.H.
All authors helped prepare the text of the manuscript.
REFERENCES AND NOTES
1. Kapadia, R.; Fan, Z.; Takei, K.; Javey, A. Nanopillar Photo-
voltaics: Materials, Processes, and Devices. Nano Energy
2012, 1, 132–144.
2. Jeong, S.; Garnett, E. C.; Wang, S.; Yu, Z.; Fan, S.; Brongersma,
M. L.; McGehee, M. D.; Cui, Y. Hybrid Silicon Nanocone-
Polymer Solar Cells. Nano Lett. 2012, 12, 2971–2976.
3. Garnett, E.; Yang, P. Light Trapping in Silicon Nanowire
Solar Cells. Nano Lett. 2010, 10, 1082–1087.
4. Tian, B.; Zheng, X.; Kempa, T. J.; Fang, Y.; Yu, N.; Yu, G.;
Huang, J.; Lieber, C. M. Coaxial Silicon Nanowires as Solar
Cells and Nanoelectronic Power Sources. Nature 2007,
449, 885–889.
5. Cao, L.; Fan, P.; Vasudev, A. P.; White, J. S.; Yu, Z.; Cai, W.;
Schuller, J. A.; Fan, S.; Brongersma, M. L. Semiconductor
Nanowire Optical Antenna Solar Absorbers. Nano Lett.
2010, 10, 439–445.
6. Wallentin, J.; Anttu, N.; Asoli, D.; Huffman, M.; Aberg, I.;
Magnusson, M. H.; Siefer, G.; Fuss-Kailuweit, P.; Dimroth, F.;
Witzigmann, B.; et al. InP Nanowire Array Solar Cells
Achieving 13.8% Efficiency by Exceeding the Ray Optics
Limit. Science 2013, 339, 1057–1060.
7. Mavrokefalos, A.; Han, S. E.; Yerci, S.; Branham, M. S.; Chen,
G. Efficient Light Trapping in Inverted Nanopyramid Thin
Crystalline Silicon Membranes for Solar Cell Applications.
Nano Lett. 2012, 12, 2792–2796.
8. Grandidier, J.; Callahan, D. M.; Munday, J. N.; Atwater, H. A.
Light Absorption Enhancement in Thin-Film Solar Cells
Using Whispering Gallery Modes in Dielectric Nano-
spheres. Adv. Mater. 2011, 23, 1272–1276.
9. Yao, Y.; Yao, J.; Narasimhan, V. K.; Ruan, Z.; Xie, C.; Fan, S.;
Cui, Y. Broadband Light Management Using Low-Q
Whispering Gallery Modes in Spherical Nanoshells. Nat.
Commun. 2012, 3, 664.
10. Zhu, J.; Hsu, C.-M.; Yu, Z.; Fan, S.; Cui, Y. Nanodome Solar
Cells With Efficient Light Management and Self-Cleaning.
Nano Lett. 2010, 10, 1979–1984.
11. Narasimhan, V. K.; Cui, Y. Nanostructures for Photon Man-
agement in Solar Cells. Nanophotonics 2013, 2, 187–210.
12. Brongersma, M. L.; Cui, Y.; Fan, S. Light Management
for Photovoltaics Using High-Index Nanostructures. Nat.
Mater. 2014, 13, 451–460.
13. Ye, S.; Rathmell, A. R.; Chen, Z.; Stewart, I. E.; Wiley, B. J.
Metal Nanowire Networks: The Next Generation of Trans-
parent Conductors. Adv. Mater. 2014, 26, 6670–6687.
14. Lee, J.-Y.; Connor, S. T.; Cui, Y.; Peumans, P. Solution-
Processed Metal Nanowire Mesh Transparent Electrodes.
Nano Lett. 2008, 8, 689–692.
15. Hu, L.; Wu, H.; Cui, Y. Metal Nanogrids, Nanowires, and
Nanofibers for Transparent Electrodes. MRS Bull. 2011, 36,
760–765.
16. Rathmell, A. R.; Wiley, B. J. The Synthesis and Coating of
Long, Thin Copper Nanowires to Make Flexible, Transpar-
ent Conducting Films on Plastic Substrates. Adv. Mater.
2011, 23, 4798–4803.
17. Hsu, P.-C.; Wang, S.; Wu, H.; Narasimhan, V. K.; Kong, D.;
Ryoung Lee, H.; Cui, Y. Performance Enhancement of Metal
Nanowire Transparent Conducting Electrodes by Meso-
scale Metal Wires. Nat. Commun. 2013, 4, 2522.
18. Rowell, M. W.; McGehee, M. D. Transparent Electrode
Requirements for Thin Film Solar Cell Modules. Energy
Environ. Sci. 2011, 4, 131.
VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX G
www.acsnano.org
 ARTICLE
19. Spinelli, P.; Verschuuren, M. A.; Polman, A. Broadband 40. Dimova-Malinovska, D.; Sendova-Vassileva, M.; Tzenov,
Omnidirectional Antireflection Coating Based on Sub- N. U.; Kamenova, M. Preparation of Thin Porous Silicon
wavelength Surface Mie Resonators. Nat. Commun. Layers by Stain Etching. Thin Solid Films 1997, 297, 9–12.
2012, 3, 692. 41. Yae, S.; Morii, Y.; Fukumuro, N.; Matsuda, H. Catalytic
20. Wang, K. X.; Yu, Z.; Sandhu, S.; Liu, V.; Fan, S. Condition for Activity of Noble Metals for Metal-Assisted Chemical
Perfect Antireflection by Optical Resonance at Material Etching of Silicon. Nanoscale Res. Lett. 2012, 7, 352.
Interface. Optica 2014, 1, 388. 42. Chang, C.; Sakdinawat, A. Ultra-High Aspect Ratio High-
21. Yu, Y.; Cao, L. Leaky Mode Engineering: A General Design Resolution Nanofabrication for Hard X-Ray Diffractive
Principle for Dielectric Optical Antenna Solar Absorbers. Optics. Nat. Commun. 2014, 5, 4243.
Opt. Commun. 2014, 314, 79–85. 43. Mikhael, B.; Elise, B.; Xavier, M.; Sebastian, S.; Johann, M.;
22. Chartier, C.; Bastide, S.; Lévy-Clément, C. Metal-Assisted Laetitia, P. New Silicon Architectures by Gold-Assisted
Chemical Etching of Silicon in HF
H2O2. Electrochim. Acta Chemical Etching. ACS Appl. Mater. Interfaces 2011, 3,
2008, 53, 5509–5516. 3866–3873.
23. Liu, V.; Fan, S. S4: A Free Electromagnetic Solver for 44. Liu, R.; Zhang, F.; Con, C.; Cui, B.; Sun, B. Lithography-Free
Layered Periodic Structures. Comput. Phys. Commun. Fabrication of Silicon Nanowire and Nanohole Arrays by
2012, 183, 2233–2244. Metal-Assisted Chemical Etching. Nanoscale Res. Lett.
24. International ASTM. ASTM G173
03(2012), Standard 2013, 8, 155.
Tables for Reference Solar Spectral Irradiances: Direct 45. Bielawny, A.; Upping, J.; Miclea, P. T.; Wehrspohn, R. B.;
Normal and Hemispherical on 37° Tilted Surface; West Rockstuhl, C.; Lederer, F.; Peters, M.; Steidl, L.; Zentel, R.;
Conshohocken, PA, 2012. Lee, S.-M.; et al. 3D Photonic Crystal Intermediate
25. Tous, L.; Russell, R.; Debucquoy, M.; Posthuma, N.; Reflector for Micromorph Thin-Film Tandem Solar Cell.
Duerinckx, F.; Mertens, R.; Poortmans, J. Power-Loss Phys. Status Solidi A 2008, 205, 2796–2810.
Analysis of Advanced PERC Cells Reaching 20.5% 46. Zhen, A.; Ma, P.; Zhang, Y.; Guo, E.; Tian, Y.; Liu, B.; Guo, S.;
Energy Conversion Efficiency. Energy Procedia 2013, 38, Shan, L.; Wang, J.; Li, J. Embeded Photonic Crystal at the
467–473. Interface of P-GaN and Ag Reflector to Improve Light
26. Ellmer, K. Past Achievements and Future Challenges in the Extraction of GaN-Based Flip-Chip Light-Emitting Diode.
Development of Optically Transparent Electrodes. Nat. Appl. Phys. Lett. 2014, 105, 251103.
Photonics 2012, 6, 809–817. 47. Li, L.; Liu, Y.; Zhao, X.; Lin, Z.; Wong, C.-P. Uniform Vertical
27. Ryan, C. G. M.; Eleftheriades, G. V. Single- and Dual-Band Trench Etching on Silicon with High Aspect Ratio by Metal-
Transparent Circularly Polarized Patch Antennas With Assisted Chemical Etching Using Nanoporous Catalysts.
Metamaterial Loading. IEEE Antennas Wirel. Propag. Lett. ACS Appl. Mater. Interfaces 2014, 6, 575–584.
2015, 14, 470–473.
28. Transparent Electronics; Facchetti, A., Marks, T. J., Eds.; John
Wiley & Sons, Ltd.: Chichester, UK, 2010.
29. Yang, Y.; Jeong, S.; Hu, L.; Wu, H.; Lee, S. W.; Cui, Y.
Transparent Lithium-Ion Batteries. Proc. Natl. Acad. Sci. U.
S. A. 2011, 108, 13013–13018.
30. Ebbesen, T. W.; Lezec, H. J.; Ghaemi, H. F.; Thio, T.; Wolff,
P. A. Extraordinary Optical Transmission through Sub-
Wavelength Hole Arrays. Nature 1998, 391, 667–669.
31. Martín-Moreno, L.; García-Vidal, F.; Lezec, H.; Pellerin, K.;
Thio, T.; Pendry, J.; Ebbesen, T. Theory of Extraordinary
Optical Transmission through Subwavelength Hole Arrays.
Phys. Rev. Lett. 2001, 86, 1114–1117.
32. Genet, C.; Ebbesen, T. W. Light in Tiny Holes. Nature 2007,
445, 39–46.
33. Koerkamp, K. J. K.; Enoch, S.; Segerink, F. B.; van Hulst, N. F.;
Kuipers, L. Strong Influence of Hole Shape on Extraordi-
nary Transmission through Periodic Arrays of Subwave-
length Holes. Phys. Rev. Lett. 2004, 92, 183901.
34. Van der Molen, K. L.; Segerink, F. B.; van Hulst, N. F.; Kuipers,
L. Influence of Hole Size on the Extraordinary Transmission
through Subwavelength Hole Arrays. Appl. Phys. Lett.
2004, 85, 4316.
35. Van der Molen, K. L.; Klein Koerkamp, K. J.; Enoch, S.;
Segerink, F. B.; van Hulst, N. F.; Kuipers, L. Role of Shape
and Localized Resonances in Extraordinary Transmission
through Periodic Arrays of Subwavelength Holes: Experi-
ment and Theory. Phys. Rev. B: Condens. Matter Mater.
Phys. 2005, 72, 045421.
36. Yokogawa, S.; Burgos, S. P.; Atwater, H. A. Plasmonic Color
Filters for CMOS Image Sensor Applications. Nano Lett.
2012, 12, 4349–4354.
37. Zeng, B.; Gao, Y.; Bartoli, F. J. Ultrathin Nanostructured
Metals for Highly Transmissive Plasmonic Subtractive
Color Filters. Sci. Rep. 2013, 3, 2840.
38. Li, L.; Zhao, X.; Wong, C.-P. Deep Etching of Single- and
Polycrystalline Silicon with High Speed, High Aspect Ratio,
High Uniformity, and 3D Complexity by Electric Bias-
Attenuated Metal-Assisted Chemical Etching (EMaCE).
ACS Appl. Mater. Interfaces 2014, 6, 16782–16791.
39. DeJarld, M.; Shin, J. C.; Chern, W.; Chanda, D.; Balasundaram,
K.; Rogers, J. A.; Li, X. Formation of High Aspect Ratio GaAs
Nanostructures with Metal-Assisted Chemical Etching.
Nano Lett. 2011, 11, 5259–5263.

NARASIMHAN ET AL. VOL. XXX ’ NO. XX ’ 000–000 ’ XXXX H

www.acsnano.org