of 5
Photoanodic behavior of vapor-liquid-solid–grown, lightly doped, crystalline
Si microwire arrays
Elizabeth A. Santori,
d
James R. Maiolo III,
d
Matthew J. Bierman,
d
Nicholas C. Strandwitz,
d
Michael D. Kelzenberg,
b
Bruce S. Brunschwig,
a
Harry A. Atwater
*
bc
and Nathan S. Lewis
*
acd
Received 15th December 2011, Accepted 14th February 2012
DOI: 10.1039/c2ee03468a
Arrays of n-Si microwires have to date exhibited low efficiencies
when measured as photoanodes in contact with a 1-1
0
-dime-
thylferrocene (Me
2
Fc
+/0
)–CH
3
OH solution. Using high-purity Au
or Cu catalysts, arrays of crystalline Si microwires were grown by
a vapor-liquid-solid process without dopants, which produced wires
with electronically active dopant concentrations of 1

10
13
cm

3
.
When measured as photoanodes in contact with a Me
2
Fc
+/0
CH
3
OH solution, the lightly doped Si microwire arrays exhibited
greatly increased fill factors and efficiencies as compared to n-Si
microwires grown previously with a lower purity Au catalyst. In
particular, the Cu-catalyzed Si microwire array photoanodes
exhibited open-circuit voltages of

0.44 V, carrier-collection
efficiencies exceeding

0.75, and an energy-conversion efficiency of
1.4% under simulated air mass 1.5 G illumination. Lightly doped
Cu-catalyzed Si microwire array photoanodes have thus demon-
strated performance that is comparable to that of optimally doped p-
type Si microwire array photocathodes in photoelectrochemical
cells.
The structuring of semiconductors on the nano- and micro-scale is
a promising approach for the fabrication of scalable and efficient
devices for the production of electricity and fuels from sunlight.
1–3
In
contrast to a traditional geometry that is characterized by planar light
absorbers and planar electrical junc
tions, wire-based architectures
orthogonalize the directions of light absorption and carrier collec-
tion.
4
Such a structure provides both a long optical path length for
efficient light absorption and a short distance for minority-carrier
collection, therefore allowing the incorporation of defective materials
with short minority-carrier diffus
ion lengths into devices that can
produce high energy-conversion efficiencies. Specifically, device-
physics modeling of radial junction
p-n
Si microwires has predicted
that crystalline Si microwire arrays can yield efficiencies comparable
to those achieved by wafer-based planar Si solar cells, even for Si
microwires that have a 1
m
m minority-carrier diffusion length.
4
Arrays of crystalline p-Si microw
ires grown by the vapor–liquid–
solid (VLS) process have demonstrated promising performance in
regenerative
5,6
and fuel–forming
7
photoelectrochemical cells, as well
as in photovoltaic devices.
8,9
For example, p-type Si microwire array
a
Beckman Institute, California Institute of Technology, 1200 E. California
Blvd., Pasadena, CA 91125, USA
b
Thomas J. Watson Laboratories of Applied Physics, California Institute of
Technology, 1200 E. California Blvd, Pasadena, CA, 91125, USA. E-mail:
haa@caltech.edu; Fax: +1 626 844-9320; Tel: +1 626-395-2197
c
Member, Kavli Nanoscience Institute, California Institute of Technology,
USA
d
Division of Chemistry and Chemical Engineering, California Institute of
Technology, 1200 E. California Blvd, Pasadena, CA, 91125, USA.
E-mail: nslewis@caltech.edu; Fax: +1 626 395-8867; Tel: +1 626 395-6335
† Electronic supplementary information (ESI) available: Additional
information regarding the VLS growth of Si microwire arrays,
four-point
resistance
measurements,
photoelectrochemical
measurements, concentration overpotential and resistance corrections,
and optical measurements of Si wire array films. See DOI:
10.1039/c2ee03468a
Broadercontext
To make solar energy cost-competitive with fossil fuels and thus drive the large-scale deployment of photovoltaics, significant
technological advances must be made to reduce the installed cost of photovoltaics to < $1/Watt
peak
. Recently, Si wire arrays grown
by the vapor–liquid–solid technique have emerged as a promising technology for the fabrication of efficient and inexpensive
photovoltaics, as well as for artificial photosynthetic devices. The thin-film Si wire array photovoltaic technology utilizes an
atmospheric pressure, chemical vapor deposition growth process, inexpensive Si precursors, and earth-abundant catalysts, and
produces flexible, crystalline Si solar cells that have the potential to approach the efficiencies of wafer-based crystalline Si solar cells.
Herein we demonstrate improved performance of Si microwire array photoanodes in contact with a regenerative non-aqueous
electrolyte that provides a conformal, high barrier-height contact to structured semiconductor materials. Hence, lightly doped Cu–
catalyzed Si microwire array photoanodes can achieve performance comparable to that of optimally doped p-type Si microwire
array photocathodes.
This journal is
ª
The Royal Society of Chemistry 2012
Energy Environ. Sci.
, 2012,
5
, 6867–6871 | 6867
Dynamic Article Links
C
<
Energy &
Environmental Science
Cite this:
Energy Environ. Sci.
, 2012,
5
, 6867
www.rsc.org/ees
COMMUNICATION
Downloaded by California Institute of Technology on 18 May 2012
Published on 20 February 2012 on http://pubs.rsc.org | doi:10.1039/C2EE03468A
View Online
/ Journal Homepage
/ Table of Contents for this issue
photoelectrodes in contact with an aqueous methyl viologen (MV
2+/+
)
redox system have yielded open-circuit voltages (
V
oc
) approaching
0.45 V under 100 mW cm

2
of simulated air mass (AM) 1.5 G illu-
mination, with near-unity internal
quantum yields, demonstrating the
efficient radial collection of carriers in the wire-array geometry.
Arrays of radial junction n
+
p-Si microwires have demonstrated
thermodynamically based photoelectrode energy-conversion effi-
ciencies of >5% for the production of H
2
from H
2
O. Analogous
arrays in solid-state photovoltaic d
evices have achieved an efficiency
of 7.8%, with
V
oc
exceeding 0.5 V, under simulated AM 1.5 G
illumination.
In contrast, n-Si microwire arra
y photoanodes in contact with
a1-1
0
-dimethylferrocene (Me
2
Fc
+/0
)–CH
3
OH solution under simu-
lated AM 1.5 conditions have only exhibited
V
oc
values of 0.39 V, in
conjunction with low fill factors and low short-circuit photocurrent
densities (
J
sc
), resulting in photoelectrode efficiencies,
h
,of

0.1%.
10
Given that the Me
2
Fc
+/0
–CH
3
OH electrolyte in contact with planar,
crystalline n-Si photoanodes produces
V
oc
values that are only
limited by bulk recombination/generation,
11
the comparatively low
performance of n-Si microwire array photoanodes is presumably
indicative of the inferior materia
l quality of the n-Si microwires.
Two factors may have contributed to the poor electronic quality of
the VLS–grown n-Si microwires: th
e purity of the catalyst used and
the choice of metal catalyst. The n-Si microwire arrays were grown
with a 99.999% (5N) Au VLS catalyst,
10,12–14
as compared to the
higher purity 99.9999% (6N) Cu catalyst that has been used to grow
p-Si microwires.
5,7,9,15–18
Uncontrolled impurities in the 5N Au
catalyst produced microwires that had high variability in the
observed electronically active dopant concentrations,
N
d
,withvalues
ranging between 1

10
14
and 1

10
20
cm

3
.
16
Additionally, the use
of Au as the VLS catalyst, as op
posed to Cu, may have limited the
efficiency of the Si microwire arrays. Although both Au and Cu
form mid-gap traps in Si, in planar Si solar cells Cu has a less
detrimental effect than Au, with a minority-carrier lifetime degra-
dation threshold concentration of 4

10
17
cm

3
for Cu as
compared to 3

10
13
cm

3
for Au.
19,20
The focus of this work was
to determine whether the low photoelectrode efficiencies observed
for n-Si microwire arrays in contact with Me
2
Fc
+/0
–CH
3
OH are an
inherent, fundamental property of the system or whether improved
performance could be obtained through control over the electronic
properties of the bulk and surface chemistry of Si wire array
photoelectrodes.
We report herein the photoelectrochemical behavior of Si micro-
wire arrays that have been fabric
ated using a 6N VLS catalyst, for
both Au and Cu. The device performance of the Si microwire arrays
was probed using the Me
2
Fc
+/0
–CH
3
OH junction, because this
system provides a conformal contact to the microwires, and obviates
the need to fabricate a diffused metallurgical junction. Additionally,
the Me
2
Fc
+/0
–CH
3
OH system generates an inversion layer in contact
with n-Si, and the resulting semico
nductor/liquid interface has a low
effective surface recombination velocity.
21
To produce the desired photoanodes, arrays of square–packed Si
microwires were grown on a planar n
+
-Si(111) substrate using the
VLS process, without dopants, but with 6N Au as the growth catalyst
(see ESI†). The resulting Si wires were oriented in the (111) direction,
with diameters of 2.25–3.0
m
m and heights of 65–75
m
m, with an
average areal packing fraction (
h
f
) of 11.0%. Four-point resistance
and gate-dependent conductance measurements indicated that the
undoped silicon microwires were nominally p-type, with resistivities
of 800

700
U
-cm, corresponding to an acceptor concentration,
N
a
,
of

1

10
13
cm

3
(Figure S1).
Current density
vs.
potential (
J
E
) measurements of Au-catalyzed
Si microwire array photoanodes, and of control photoanodes in
which the wires had been physically removed after growth on the n
+
-
Si substrate, were measured in contact with 200 mM Me
2
Fc-0.4 mM
Me
2
FcBF
4
in CH
3
OH under 100 mW cm

2
of simulated AM 1.5 G
illumination (Fig. 1A). The Si wire array photoanodes exhibited
V
oc
¼
334

21 mV,
J
sc
¼
10.0

1.3 mA cm

2
, and fill-factors,
ff
¼
0.34

0.05, with a photoelectrode energy-conversion efficiency
h
¼
1.1

0.3%. The n
+
-Si(111) control substrates for which the wires had
been physically removed after growth produced
V
oc
¼
233

38 mV,
J
sc
¼
1.0

0.2 mA cm

2
,and
ff
¼
0.20

0.04. The observed
properties of the Si microwire array photoelectrodes can therefore
predominately be attributed to the behavior of the VLS–grown
crystalline Si wires in contact with the Me
2
Fc
+/0
–CH
3
OH electrolyte.
The behavior of the same Au-catalyzed Si microwire photoanodes
was also measured in contact with a higher concentration of the
oxidized form of the redox couple, 25 mM Me
2
FcBF
4
, to reduce the
concentration overpotential losses at the photoelectrode. The pho-
toelectrodes were illuminated using an 808 nm laser diode, such that
the
J
sc
value matched the value of
J
sc
that was obtained at low
Me
2
Fc
+
concentrations under 100 mW cm

2
of simulated AM 1.5 G
illumination. Fig. 1B shows the performance of the arrays in the
presence of either 0.4 or 25 mM Me
2
FcBF
4
, with the latter cell
exhibiting a fill factor of
ff
808
¼
0.47

0.04 and an efficiency
h
808
¼
2.7

0.7%. After correcting both the 0.4 and 25 mM Me
2
FcBF
4
J
E
data for concentration overpotenti
al and uncompensated resistance
losses, the corrected fill factor and photoelectrode efficiency values
were
ff
corr
¼
0.57

0.05 and
h
corr
¼
2.0

0.5%, respectively
(calculations in ESI†). The corrected
J
E
data are indicative of the
inherent performance of the Si microwire arrays, without experi-
mental artifacts arising from measurement in an unoptimized elec-
trochemical cell configuration.
The photoelectrode efficiency of Si microwire arrays that were
grown using the 6N-purity Au VLS c
atalyst represents a significant
improvement relative to initial measurements of the photoanodic
performance of n-Si microwire
arrays in contact with the Me
2
Fc
+/0
CH
3
OH system. However, the
V
oc
of the Si microwire arrays grown
with Au was still substantially less than the
V
oc
values produced by
either p-type or n
+
p-Si microwire array photocathodes that were
grownwithCuandtestedincontactwiththeMV
2+/+
(aq) redox
system. For Si microwires grown with a Au VLS catalyst, bulk Au
concentrations up to 1.7

10
16
cm

3
have been previously measured,
corresponding to the thermodynamic equilibrium concentration of
Au in Si at the growth temperature.
14
The concentration of Au within
the wires thus greatly exceeded the degradation threshold concen-
tration of Au in planar Si solar cells, and could have contributed to
the lower
V
oc
values that were measured for wires that were grown
with a Au catalyst. Indeed, Si mic
rowire arrays that were grown by
a Au-catalyzed VLS process have previously shown
V
oc
values up to
500 mV in photovoltaic device structures, but only after repeated
thermal oxidation and etching steps that should getter Au at the
surfaces of the Si wires.
8
Given the low tolerance for Au in Si solar cells, 6N Cu was
subsequently used for VLS-catalyzed Si wire growth. Cu catalyzed,
hexagonally packed Si microwire arrays were fabricated without
dopants, with the resulting wires being 2.0–2.5
m
mindiameterand
70–80
m
m in height, and providing an ave
rage areal packing fraction
6868 |
Energy Environ. Sci.
, 2012,
5
, 6867–6871
This journal is
ª
The Royal Society of Chemistry 2012
Downloaded by California Institute of Technology on 18 May 2012
Published on 20 February 2012 on http://pubs.rsc.org | doi:10.1039/C2EE03468A
View Online
(
h
f
) of 9.4% (Fig. 2). Similar to the Au-catalyzed wires, the
Cu catalyzed wires were slightly p-type, with resistivities of 1000

600
U
-cm, as measured by four-point resistance measurements, cor-
responding to
N
a
of

1

10
13
cm

3
.
J
E
measurements of the Cu-catalyzed Si microwire arrays in
contact with the Me
2
Fc
+/0
–CH
3
OH electrolyte under 100 mW cm

2
of
simulated AM 1.5 G illumination (Fig. 3A) showed
V
oc
¼
437

8mV,
J
sc
¼
7.9

0.5 mA cm

2
,and
ff
¼
0.40

0.02, with a photoelectrode
efficiency of
h
¼
1.4

0.1%. The Cu-catalyzed Si microwire array
photoanodes measured herein exhibited a slightly smaller
J
sc
than the
Au-catalyzed Si microwire array photoanodes, consistent with the Cu-
catalyzed wire arrays possessing a smaller areal packing fraction than
the Au-catalyzed wire arrays. The n
+
-Si(111) control substrates with
the wires physically removed produced
V
oc
¼
262

17 mV,
J
sc
¼
1.3

0.3 mA cm

2
,and
ff
¼
0.21

0.01, again demonstrating that the wafer
substrate did not contribute substan
tially to the photoresponse of Si
microwire photoelectrodes. The Cu-catalyzed Si microwire array
photoanodes measured under 808 nm
illumination in contact with 25
mM Me
2
FcBF
4
, to reduce concentration overpotential losses,
exhibited a fill factor of
ff
808
¼
0.60

0.02 and an efficiency of
h
808
¼
3.4

0.2% (Fig. 3B). After correcting for concentration overpotential
and uncompensated resistance lo
sses, the corrected fill factor and
efficiency were
ff
corr
¼
0.61

0.04 and
h
corr
¼
2.1

0.1%, respectively,
for Si microwire photoanodes measured under AM 1.5 G illumina-
tion. Thus, the undoped Cu-catalyzed Si wire array photoanodes in
contact with Me
2
Fc
+/0
–CH
3
OH not only exhibited improved perfor-
mance relative to the initial n-Si wire array measurements, but
also yielded efficiencies that were very similar to those observed
for optimally doped p-Si wire array photocathodes in contact with
MV
2+/+
(aq).
5,18
To further investigate the carrier-collection efficiency of the Cu-
catalyzed Si microwire arrays, the external quantum yield,
G
ext
,ofSi
microwire photoanodes in contact with Me
2
Fc
+/0
–CH
3
OH was
recorded as a function of the incident angle of illumination (
q
y
). The
photoelectrodes were rotated about the
q
y
axis of the wire array, as
showninFig.2B,from
q
y
¼
0–60

. At normal incidence of illumi-
nation (
q
y
¼
0

), the arrays showed the lowest external quantum
yield, with
G
ext

0.28 under visible illumination (Fig. 4A), corre-
sponding to high transmission through an array that had an average
h
f
¼
9.4%. The prominent resonant peaks in the external quantum
yield can be attributed to whispering-gallery modes in the hexagonal
wires, in which light can circularly propagate at the periphery due to
multiple total internal reflections.
22,23
Despite the low packing fraction
of wires, the arrays effectively col
lected 28% of the incident photons,
Fig. 1
Current density
vs.
potential (
J
E
) data for Au–catalyzed Si microwire array photoelectrodes in contact with the Me
2
Fc
+/0
–CH
3
OH redox system
(a) under 100 mW cm

2
of AM 1.5 G illumination and (b) with varying amounts of varying amounts of Me
2
FcBF
4
, to demonstrate and correct for the
electrochemical concentration overpotential resistance and series resistance losses within the cell.
Fig. 2
(a) Side view scanning electron microscope image of a cleaved
array of Si microwires, scale bar
¼
40
m
m. (b) Top view of a Si microwire
hexagonal array, and the noted axis of rotation
q
y
, scale bar
¼
20
m
m.
Fig. 3
J
E
data for Cu–catalyzed Si microwire array photoelectrodes in contact with the Me
2
Fc
+/0
–CH
3
OH redox system (a) under 100 mW cm

2
of
AM 1.5 G illumination and (b) with varying amounts of Me
2
FcBF
4
.
This journal is
ª
The Royal Society of Chemistry 2012
Energy Environ. Sci.
, 2012,
5
, 6867–6871 | 6869
Downloaded by California Institute of Technology on 18 May 2012
Published on 20 February 2012 on http://pubs.rsc.org | doi:10.1039/C2EE03468A
View Online
demonstrating optical concentration within the array. The spectral
response of the wire arrays strongly depended on the angle of incident
illumination, with a peak of
G
ext
¼
0.86 at
q
y
>52

.Incontrast,
photoanodes for which the Si wires had been physically removed
from the substrate exhibited negligible photocurrent, with
G
ext
< .03.
The current of such electrodes also did not vary with angle, further
indicating that the degeneratel
y doped Si substrate did not substan-
tially contribute to the response of the Si wire array photoanodes.
The significant increase in
G
ext
as
q
y
increased indicates that the
J
sc
of
the wire arrays under AM 1.5 G illumination was primarily limited
by light absorption in the array, and not by carrier collection.
Convolution of the spectral response at
q
y
>52

with the AM 1.5 G
spectraresultedinapredicted
J
sc
value of 26 mA cm

2
.Ashasbeen
demonstrated recently, this calculated
J
sc
can be attained by incor-
porating light-trapping elements such as a back-reflector, anti-
reflective coatings, and/or scattering particles into the device
structure.
9,24
To calculate the internal quantum yield,
G
int
, of the Cu-catalyzed Si
microwire array photoanodes, optical absorption measurements as
a function of wavelength and angle were performed on the same wire
arrays that were used for collection of the external quantum yield
data. An integrating sphere was used to perform optical transmission
and reflection measurements on peeled films of wires that had been
embedded in polydimethylsiloxane (PDMS), as described previ-
ously.
24
As expected, the measured absorption was lowest at normal
incidence, corresponding to large transmission through the sparsely
packed, highly oriented array (Fig. 4B). The absorption rapidly
increased with increasing angle of incident illumination, reaching
a plateau value of 0.86. The value of
G
int
for the Si microwire array
photoelectrodes was subsequently calculated by dividing the
G
ext
at
normal incidence by the absorption of the array at normal incidence.
Fig. 4C compares the values of
G
ext
and absorption at normal inci-
dence, resulting in a peak
G
int
value of 0.73

0.05.
From the values of
G
int
and the optical absorption coefficient
a
(
l
),
the minority-carrier diffusion length
L
n
, for a planar device can be
approximated by use of eqn (1).
11,25
G
int
¼
1
1
þ
1
a
ð
l
Þ
L
n
(1)
Analysis of
G
int
for the Cu-catalyzed Si microwire array photo-
anodes in the near-infrared region (800 nm
#
l
#
950 nm), in which
the optical penetration depth
a

1
did not exceed the length of the wires,
yielded an effective hole diffusion length,
L
n,eff
, of 75–85
m
m. This
value is not a true diffusion length, given that the assumptions of eqn
(1) do not apply in a radial geometry, but rather a comparison to the
diffusion length that would be needed to produce similar near-IR
carrier collection efficiencies in a planar Si device structure (Fig. 5). The
calculated
L
n,eff
value is significantly larger than the 30
m
m minority-
carrier diffusion length that has been measured previously for
moderately doped, Cu-catalyzed, VLS–grown Si microwires.
9,17
This
observation further demonstrates
the advantages of using a radial
junction, which produces a longer effective diffusion length than the
actual minority-carrier diffusio
n length, by extending the device
response further into the near-IR region of the spectrum as compared
to planar Si-based devices with comparable bulk electronic properties.
With
V
oc
values of 0.44 V, a photoelectrode energy-conversion
efficiency of
h
¼
1.4%, and high carrier-collection efficiencies, the
undoped Cu–catalyzed Si microwire array photoanodes have
demonstrated performance that is comparable to that of optimally
doped p-type Si microwire array photocathodes. However, the
measured
G
int
deviated from the value of unity previously measured
for p-type Si microwire photocathodes, and from the unity value that
is predicted by radial junction th
eory for a wire having a radius
smaller than the minority-carrier diffusion length.
4
The
G
int
of the
wire array photoanodes at normal incidence was also lower than the
Fig. 4
Angle-resolved (a) external quantum yield and (b) absorption,
and (c) calculated internal quantum yield at normal incidence, for Si
microwire arrays.
6870 |
Energy Environ. Sci.
, 2012,
5
, 6867–6871
This journal is
ª
The Royal Society of Chemistry 2012
Downloaded by California Institute of Technology on 18 May 2012
Published on 20 February 2012 on http://pubs.rsc.org | doi:10.1039/C2EE03468A
View Online
G
ext
value that was measured at high angles, implying a change in
G
int
with a change in the angle of incident illumination. The Si microwires,
however, are not expected to have a minority-carrier diffusion length
smaller than their 1.25
m
m radius, given that 30
m
m minority carrier
diffusion lengths have been reported previously for Cu–catalyzed
VLS–grown Si microwires. These apparent discrepancies can be
explained by the presence of Cu silicide located at the tops of the
wires, as visualized by a high-cont
rast region in scanning electron
microscopy images (Figure S2). This region persisted after chemical
etching to remove the metallic Cu VLS catalyst, and varied in
thickness from 200–900 nm, dependi
ng on the cooling conditions
after VLS growth. This silicide most likely acted as a region of low
lifetime for carriers that were generated at the tops of the wires, where
a significant fraction of the light is absorbed,
26
thus decreasing the
internal quantum yield, particu
larly at normal incidence. The
quantum yield would therefore not only deviate from unity due to
this electronically defective region, but would also vary as the exci-
tation profile changed with incident
angle of illumination, consistent
with the observations reported herein.
Due to their very low dopant concentration, the Si wires are
expected to be operating under high–level injection conditions, in
which the concentration of photogenerated carriers exceeds the
equilibrium concentration of majority carriers in the wire.
27
The
initial device physics model only encompassed wires under low–level
injection that were not fully depleted, making the model potentially
not applicable to undoped microwires operating under field–free
conditions. Modeling of radial jun
ction nanowire devices under high–
level injection conditions predicts that these devices should have poor
carrier collection efficiencies, due
to the full depletion within the
nanowire resulting in large majority-carrier recombination losses.
16,28
Work is currently underway to understand in more detail the prop-
erties of both nano- and microscale radial geometry devices with low
dopant densities that are operating under high–level injection
conditions.
Acknowledgements
We acknowledge BP, the Gordon and Betty Moore Foundation,
Toyota, and the U.S. Department of Energy for financial support.
NCS acknowledges the NSF for an American Competitiveness in
Chemistry postdoctoral fellowship (CHE-1042006). The angle-
resolved optical characterization work was supported by the US
Department of Energy ‘Light–Material Interactions in Energy
Conversion’ Energy Frontier Res
earch Center Award (grant DE-
SC0001293). We acknowledge critica
l support and infrastructure
provided for this work by the Kavli Nanoscience Institute at Caltech.
References
1 B. Tian, T. J. Kempa and C. M. Lieber,
Chem. Soc. Rev.
, 2009,
38
,
16–24.
2 A. I. Hochbaum and P. D. Yang,
Chem. Rev.
, 2010,
110
, 527–546.
3 M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher,
Q. X. Mi, E. A. Santori and N. S. Lewis,
Chem. Rev.
, 2010,
110
,
6446–6473.
4 B. M. Kayes, H. A. Atwater and N. S. Lewis,
J. Appl. Phys.
, 2005,
97
,
114302.
5 S. W. Boettcher, J. M. Spurgeon, M. C. Putnam, E. L. Warren,
D. B. Turner-Evans, M. D. Kelzenberg, J. R. Maiolo,
H. A. Atwater and N. S. Lewis,
Science
, 2010,
327
, 185–187.
6 J.M.Spurgeon,S.W.Boettcher,M.D.Kelzenberg,B.S.Brunschwig,
H. A. Atwater and N. S. Lewis,
Adv. Mater.
, 2010,
22
, 3277–3281.
7 S. W. Boettcher, E. L. Warren, M. C. Putnam, E. A. Santori,
D. Turner-Evans, M. D. Kelzenberg, M. G. Walter, J. R. McKone,
B. S. Brunschwig, H. A. Atwater and N. S. Lewis,
J. Am. Chem.
Soc.
, 2011,
133
, 1216–1219.
8 C. E. Kendrick, H. P. Yoon, Y. A. Yuwen, G. D. Barber, H. T. Shen,
T. E. Mallouk, E. C. Dickey, T. S. Mayer and J. M. Redwing,
Appl.
Phys. Lett.
, 2010,
97
, 143108.
9 M. C. Putnam, S. W. Boettcher, M. D. Kelzenberg, D. B. Turner-
Evans, J. M. Spurgeon, E. L. Warren, R. M. Briggs, N. S. Lewis
and H. A. Atwater,
Energy Environ. Sci.
, 2010,
3
, 1037–1041.
10 J. R. Maiolo, B. M. Kayes, M. A. Filler, M. C. Putnam,
M. D. Kelzenberg, H. A. Atwater and N. S. Lewis,
J. Am. Chem.
Soc.
, 2007,
129
, 12346–12347.
11 M. L. Rosenbluth and N. S. Lewis,
J. Am. Chem. Soc.
, 1986,
108
,
4689–4695.
12 M. D. Kelzenberg, D. B. Turner-Evans, B. M. Kayes, M. A. Filler,
M. C. Putnam, N. S. Lewis and H. A. Atwater,
Nano Lett.
, 2008,
8
,
710–714.
13 B. M. Kayes, M. A. Filler, M. C. Putnam, M. D. Kelzenberg,
N. S. Lewis and H. A. Atwater,
Appl. Phys. Lett.
, 2007,
91
, 103110.
14 M. C. Putnam, M. A. Filler, B. M. Kayes, M. D. Kelzenberg,
Y. B. Guan, N. S. Lewis, J. M. Eiler and H. A. Atwater,
Nano
Lett.
, 2008,
8
, 3109–3113.
15 M. C. Putnam, D. B. Turner-Evans, M. D. Kelzenberg,
S. W. Boettcher, N. S. Lewis and H. A. Atwater,
Appl. Phys. Lett.
,
2009,
95
, 163116.
16 M. D. Kelzenberg, Ph.D., California Institute of Technology, 2010.
17 M. D. Kelzenberg, D. B. Turner-Evans, M. C. Putnam,
S. W. Boettcher, R. M. Briggs, J. Y. Baek, N. S. Lewis and
H. A. Atwater,
Energy Environ. Sci.
, 2011,
4
, 866–871.
18 E. L. Warren, S. W. Boettcher, M. G. Walter, H. A. Atwater and
N. S. Lewis,
J. Phys. Chem. C
, 2011,
115
, 594–598.
19 R. H. Hopkins and A. Rohatgi,
J. Cryst. Growth
, 1986,
75
, 67–79.
20 J. R. Davis, A. Rohatgi, R. H. Hopkins, P. D. Blais,
P. Raichoudhury, J. R. Mccormick and H. C. Mollenkopf,
IEEE
Trans. Electron Devices
, 1980,
27
, 677–687.
21 F. Gstrein, D. J. Michalak, W. J. Royea and N. S. Lewis,
J. Phys.
Chem. B
, 2002,
106
, 2950–2961.
22 T. Nobis, E. M. Kaidashev, A. Rahm, M. Lorenz and
M. Grundmann,
Phys. Rev. Lett.
, 2004,
93
.
23 L. Y. Cao, J. S. White, J. S. Park, J. A. Schuller, B. M. Clemens and
M. L. Brongersma,
Nat. Mater.
, 2009,
8
, 643–647.
24 M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-
Evans, M. C. Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs,
N. S. Lewis and H. A. Atwater,
Nat. Mater.
, 2010,
9
, 239–244.
25 W. W. Gartner,
Phys. Rev.
, 1959,
116
, 84–87.
26 M. D. Kelzenberg, M. C. Putnam, D. B. Turner-Evans, N. S. Lewis
and H. A. Atwater,
2009 34th IEEE Photovoltaic Specialists
Conference
, Vols. 1–3, 2009, 001948–001953.
27 A. Luque and S. Hegedus, ed.,
Handbook of Photovoltaic Science and
Engineering
, Wiley, Chichester, England; Hoboken, NJ 2003.
28 J. M. Foley, M. J. Price, J. I. Feldblyum and S. Maldonado,
Energy
Environ. Sci.
, 2012,
5
, 5203–5220.
Fig. 5
A comparison of the measured
G
int
of Cu-catalyzed Si microwire
array photoanodes and several calculated
G
int
responses for planar Si
photoelectrodes, with
L
n,
ranging from 5–200
m
m.
This journal is
ª
The Royal Society of Chemistry 2012
Energy Environ. Sci.
, 2012,
5
, 6867–6871 | 6871
Downloaded by California Institute of Technology on 18 May 2012
Published on 20 February 2012 on http://pubs.rsc.org | doi:10.1039/C2EE03468A
View Online