1
Supporting Information
.
A. Materials
Methanol
(BakerDRY, Mallinckrodt
Baker) and lithium perchlorate
(battery grade,
Sigma
-Aldrich) were used as received,
without further purifica
tion. Me
2
Fc
(95%
, Sigma
-
Aldrich) was sublimated at room temperature and stored
under an inert atmosphere
until use.
Me
2
FcBF
4
was synthesized as described previously.
1
B. VLS
–Catalyzed Wire Growth
Si microwire arrays were grown using th
e vapor
–liquid
–solid (VLS) growth method,
using thermally evaporated Au or Cu
(ESPI, 99.9999%)
as the VLS growth catalyst.
Degenerately doped
(111)
–oriented
n-Si
wafers with a resistivity of
ρ
< 0.007
Ω
-cm
and with
300 nm of
thermal oxide (Addison Enginee
ring, Inc.) were
used as the growth substrates. A
positive photoresist (Microchem S1813)
was used to pattern t
he wafers with 3 μm diameter
circular holes
, with
a 7
μm center
-to-center spacing
, in a square (for Au) or hexagonal
(for Cu)
array. The exposed
thermal oxide was etched in buffered HF(aq) (BHF, Transene Inc.) for 4
min. Immediately following the HF etch, 400 nm of Au or Cu was thermally evaporated onto
the patterned growth substrate. Lift
-off was performed in acetone,
and the patterned wafers w
ere
then
cleaved into 1.3 x 2.0 cm pieces. To perform VLS growth, the samples were annealed in a
tube furnace at 1000º
C for 20 min with 500 sccm of
H
2
at atmospheric pressure
. W
ire growth
was
induced
by introduction of SiCl
4
(6N, Strem
) in 50 sccm of
He
into the reactor for 20 min.
After VLS growth, the Au VLS catalyst was subsequently
removed by a 10 s BHF etch followed
by etching in a Au etchant solution (gold etch TFA, Transene Inc.) for 45 min. T
he
Cu
growth
catalyst was likewise
removed by a
10 s BHF etch
, immediately followed by an etch in
6:1:1
(by
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
2
volume
) of
H
2
O:HCl:H
2
O
2
at 70º
C (RCA 2)
for 15 min. Prior to photoelectrochemical
measurements, the Si wire arrays were etched for 5 s in 5% HF(aq), rinsed with > 18
M
Ω
-cm
resistivity
H
2
O, and dried thoroughly under a stream of N
2
(g).
C. Four
-point Resistance
and Gate
-Dependent
Measurements
Four
-point resistance measurements were performed
as
described
previously.
2
After
removal of the VLS catalyst,
an area of 3 x 3 mm of
Si microwires was
mechanically removed
from the growth substrate with a razor bl
ade,
and
the microwires were suspended in isopropanol.
The wires were then spin
-coated onto a silicon wafer that had been coated with 300 nm of Si
3
N
4
(University Wafer). Contacts were patterned on individual wires using a lift
-off resist (LOR10A,
Microch
em) and a positive photoresist (S1813, Microchem). Immediately following a 5 s BHF
etch, 800 nm of Al (5N, Kurt J. Lesker) and 200 nm of Ag (4N, Kurt J. Lesker) were deposited
by electron
-beam evaporation onto the patterned wafer, to form ohmic contacts t
o the wires.
The
conductivity of the wires was measured with varying gate bias potentials, between -
10 V and +10
V, to determine the carrier type in the wires.
For both Au and Cu catalyzed wires, three wire
arrays were sampled to measure the variability i
n wire dopant density among growths, and, for
each growth, at least ten wires were measured
. For Au catalyzed wire arrays,
the measured
resistivities
of wires from three different arrays were 1400
±
900
Ω
-cm,
800
±
700
Ω
-cm,
and
600
±
300
Ω
-cm
. For Cu catalyzed wire arrays, the measured resistivities of wires from three
different arrays were 310
±
70
Ω
-cm,
1000
±
600
Ω
-cm,
and
600
±
400
Ω
-cm.
The resistivities
of wires from the arrays measured for their
J
-
E
performance are
provided in the main article.
D. Electrode Fabrication
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
3
To fabricate electrodes for photoelectrochemical
measurements, wire arrays
were cleaved
into 4 x 4 mm samples. The backs of the samples were scratched with a SiC scribe that was
coated in Ga:In eutectic
, to make ohmic contact to the Si substrate. The samples were then
mounted with Ag print
( GC Electronics
) on a coiled wire
that was passed through a
glass tube
, so
that the electrode was positioned
in a face
-down configuration. The
active area of the electrodes
was defined using Loctite 9460 F epoxy, and the back contact and wire coil
were
insulated using
Hysol 1C epoxy. Prior to electrochemical measurements, the electrodes were placed
for 2 h in
an oven heated to 70º
C, to further
cure the epoxy to obtain
enhanced chemical stability.
Electrode areas were measured with a high
-resolution scanner
, and
were
calculated using Adobe
Photoshop software. Electrode areas were ~
0.03 cm
2
.
E. Photoelectrochemical Measurements
Current density
vs. potential (
J
-
E
) measurements were performed with bottom
illumination in an air
-tight, flat
-bottomed
glass cell. The electrolyte
solution consisted of
200
mM of dimethylferrocene (Me
2
Fc),
0.4
mM of Me
2
FcBF
4
, and 1.0
M LiClO
4
in 30 mL of
methanol. The
cell was assembled and sealed under an
inert atmosphere (< 10 ppm O
2
) before
being placed under a
positive Ar pressure outside of the dry
box. A methanol bubbler was used
to prevent evaporation of the solution during an Ar purge. The three
-electrode cel
l consisted of a
high
-area Pt mesh as the counter electrode, a Pt wire in a Luggin capillary filled with the cell’s
solution as the reference
electrode, and a Si working electrode. The solution potential versus the
reference
was continuously monitored usi
ng a 4
-digit voltmeter (Keithley)
, and deviated from
the reference by <10 mV
.
J
-
E
measurements were obtained
at a scan rate of 5 mV s
-1
using a
Princeton Applied Research (PAR) Model 273 potentios
tat in conjunction with CoreWare
software
.
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
4
The cell was i
lluminated using a 150 W Xe lamp with an
AM 1.5 G filter
(Newport/Oriel) coupled to a UV
-enhanced Al mirror
, to direct the light through the bottom of
the cell. The incident light intensity was calibrated using a Si photodiode that was placed in
the
solut
ion at the position of the working electrode. The light intensity was adjusted until the short
-
circuit photocurrent density on the Si diode was the same as the value produced by
100 mW cm
-2
of AM 1.5G illumination. The cell was vigorously stirred during
J
-
E
measurements.
Data were
collected and averaged for s
ix wire array samples
, for both Au-
and Cu
-catalyzed Si microwire
arrays.
To demonstrate the correction for concentration overpotential losses, 25
mM Me
2
FcBF
4
was added to the cell. A 1 W 808 nm diode laser (Thor
labs)
was used as the illumination source,
and
J
-
E
data
were collected by matching the
J
sc
value to the value of
J
sc
that was obtained under
simulated AM 1.5 G illumination. This process required ~60 mW cm
-2
of 808 nm illumination.
Correcti
ons for the c
oncentration overpotential (
η
conc
) and series
resistance
(
R
s
) losses
were
performed
according to eqs
. 1 and 2.
(1)
(2)
where
k
B
is Boltzmann’s constant;
T
is the absolute temperature;
q
is the (u
nsigned) charge on an
electron;
n
is stoichiometric number of electrons transferred in the electrode reaction (
n
= 1 for
Me
2
Fc
+/0
); and
J
l,a
and
J
l,c
are the anodic and cathodic mass
-transport
-limited current densities,
respectively. A Pt foil working ele
ctrode of comparable area to the Si working electrodes was
used to measure
J
l,a
and
J
l,c
, and
R
s
of the cell
. The limiting anodic current density was 80 mA
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
5
cm
-2
and the limiting
cathodic
current densities were 0.15 and 9.8 mA cm
-2
, for 0.4 mM and 25
mM Me
2
FcBF
4
, respectively. The measured value of
R
s
was dependent on the placement of the
working electrode with respect to the Luggin capillary, and typically varied from 50–150
Ω
. A
value of
R
s
= 50
Ω
was used in the calculations to avoid overcorrection of the data, resulting in
conservative, potentially underestimated, values for the intrinsic fill factor and photoelectrode
efficiency of the Si/CH
3
OH-
Me
2
Fc
+/0
contact.
F. Angle
-resolved Spectral Response and Optical Measurements
For angle
-resolved spectral
response measurements, side
-facing electrodes
of
high
-fidelity Si microwire arrays with dimensions of ~
8 x 8 mm
were fabricated as
described
previously
for photoelectrochemical measurements performed under 1 Sun. Care was taken to
ensure that all electr
odes had the same orientation in all three dimensions, with the long 14 μm
axis of the hexagona
l pattern oriented vertically, as the axis of rotation (
θ
y
). Angle
-resolved
spectral response measurements were performed using an apparatus that has been descr
ibed
previously, which consisted of a chopped (
f
= 30 Hz) Fianium supercontinuum laser coupled to a
monochromator, with two rotational stages to allow for rotation around both the
θ
x
and
θ
y
axes.
3
A custom, air
-tight, round-
bottom flask with a side window was utilized for angle
-resolved
spectral response measurements, allowing free rotation about the
θ
y
axis
. The configuration of
the electrochemical cell was identical to that of the cell that was used for
J
-
E
measurements,
except that 10 mM Me
2
Fc was used, to reduce light absorption from this reagent. The working
electrode was poised at the solution potential of the cell, referenced by a Pt wire in solution. The
photoelectrode was aligned in the cell by utilizing the reflected optical diffraction pattern, and
normal incidence (
θ
x
,
y
= 0º)
was determined by minimizing the photocurrent of each electrode.
A calibrated Si photodiode (Thorlabs) that was positioned inside the cell was used to calculate
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
6
the
Γ
ext
of the Si microwire array photoelectrodes. The cell was constantly purged with Ar, and
no degradation of the response of the Si working
electrode
was observed over the duration of the
experiment.
After electrochemical measurements, the electrodes were thoroughly rinsed
, and
polydimethylsiloxane (PDMS; Sylgard 184, Dow Corning)
was
drop-
cast into the e
lectrodes. To
ensure a uniform
thin film, a transparency was placed on top of the curing PDMS, to create a thin
film (< 1 mm) that exhibited
little optical distortion. The PDMS was allowed to cure at room
temperature
for 48 h, and the tra
nsparency was then
removed from the top of the f
ilm. Si w
ires
embedded in PDMS were subsequently peeled
off
of
the electrode using a razor blade, and the
films were mounted onto a quartz slide.
Optical transmission and reflection measurements
as a
functi
on of wavelength (
λ
) and incident angle of illumination (
θ
y
) were
performed on the peeled–
off films using an integrating sphere
.
3
The optical diffraction patterns of the arrays were used to
orient the films relative to the rotational axes (
θ
x
,
θ
y
). The maximization of transmission in the
films was taken to be normal incidence to the wire array, and was 4º off specular. This angle
corresponded to the off
-cut of the (111) growth substrate, which caused the microwires to grow
at 4º from the surface normal.
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
7
References.
1.
J. R. Maiolo, H. A. Atwater and N. S. Lewis,
J Phys Chem C
, 2008,
112
, 6194-
6201.
2.
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.
3.
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.
4.
J. R. Maiolo, B. M. Kayes, M. A. Filler, M. C. Putnam, M. D. Kelzenberg, H. A. Atwater
and N. S. Lew
is,
J Am Chem Soc
, 2007,
129
, 12346-
12347.
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
8
SI Figure Captions.
Figure S1.
Two
-point and four
-point resistivity measurements of an undoped Si microwire. The
inset is a scanning electron microscope image of a contacted single wire for the measur
ement
(scale bar 40 μm).
Figure S2.
Scanning electron microscope
image
of the top of a Si microwire after chemical
etching had been performed to remove the metallic Cu VLS catalyst (scale bar 1 μm).
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
9
Figure S1.
Table S1. Figures of Merit of Si Mic
rowire Array Cells
V
oc
(mV)
J
sc
(mA cm
-
2
)
ff
Efficiency (%)
Au Catalyzed (AM 1.5 G)
334 ± 21
10.0 ± 1.3
0.34 ± 0.05
1.1 ± 0.3
Au Catalyzed (808 nm)
332 ± 18
10.2 ± 1.2
0.47 ± 0.04
2.7 ± 0.7
Corrected Au
334 ± 21
10.4 ± 1.4
0.57
± 0.05
2.0
± 0.5
Au Wi
res Removed
223 ± 38
1.0 ± 0.2
0.20 ± 0.04
0.04 ± 0.01
Cu Catalyzed (AM 1.5 G)
437 ± 8
7.9 ± 0.6
0.40 ± 0.02
1.4 ± 0.1
Cu Catalyzed (808 nm)
435
± 10
7.8 ± 0.4
0.60
± 0.02
3.4
± 0.2
Corrected Cu
437 ± 8
8.0 ± 0.7
0.61
± 0.04
2.1
± 0.1
Wires Removed
274
± 1
1.4 ± 0.1
0.22
± 0.002
0.08
± 0.006
Previous Result
4
389
±
18
1.43 ± 0.14
0.16
± 0.02
0.09
± 0.01
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012
10
Figure S2.
Electronic Supplementary Material (ESI) for Energy & Environmental Science
This journal is © The Royal Society of Chemistry 2012