Behavior of electrodeposited cd and pb schottky junctions on C H3 -Terminated n-Si(111) Surfaces

Article abstract of DOI:10.1149/1.3021450

n-SiCd and n-SiPb Schottky junctions have been prepared by electrodeposition of Cd or Pb from acidic aqueous solutions onto H-terminated and C H3 -terminated n-type Si(111) surfaces. For both nondegenerately (n-) and degenerately (n+ -) doped H-Si(111) el

Full text of DOI:10.1149/1.3021450

Journal of The Electrochemical Society, 156 ͑2͒ H123-H128 ͑2009͒
H123
0013-4651/2008/156͑2͒/H123/6/$23.00 © The Electrochemical Society
Behavior of Electrodeposited Cd and Pb Schottky Junctions
on CH3-Terminated n-Si(111) Surfaces
,z
*
Stephen Maldonado and Nathan S. Lewis
Beckman Institute and Kavli Nanoscience Institute, Division of Chemistry and Chemical Engineering,
California Institute of Technology, Pasadena, California 91125, USA
n-Si/Cd and n-Si/Pb Schottky junctions have been prepared by electrodeposition of Cd or Pb from acidic aqueous solutions onto
H-terminated and CH3-terminated n-type Si͑111͒ surfaces. For both nondegenerately ͑n-͒ and degenerately ͑n+-͒ doped H–Si͑111͒
electrodes, Cd and Pb were readily electroplated and oxidatively stripped, consistent with a small barrier height ͑⌽b͒ at the
Si/solution and the Si/metal junctions. Electrodeposition of Cd or Pb onto degenerately doped CH3-terminated n+-Si͑111͒ elec-
trodes occurred at the same potentials as Cd or Pd electrodeposition onto H-terminated n+-Si͑111͒. However, electrodeposition on
nondegenerately doped CH3-terminated n-Si͑111͒ surfaces was significantly shifted to more negative applied potentials ͑by −130
and −347 mV, respectively͒, and the anodic stripping of the electrodeposited metals was severely attenuated, indicating large
values of ⌽b for contacts on nondegenerately doped n-type CH3–Si͑111͒ surfaces. With either Cd or Pb, current–voltage mea-
surements on the dry, electrodeposited Schottky junctions indicated that much larger values of ⌽b were obtained on
CH3-terminated n-Si͑111͒ surfaces than on H-terminated n-Si͑111͒ surfaces. Chronoamperometric data indicated that
CH3–Si͑111͒ surfaces possessed an order-of-magnitude lower density of nucleation sites for metal electrodeposition than did
H–Si͑111͒ surfaces, attesting to the high degree of structural passivation afforded by the CH3–Si surface modification.
© 2008 The Electrochemical Society. ͓DOI: 10.1149/1.3021450͔ All rights reserved.
Manuscript submitted June 2, 2008; revised manuscript received October 13, 2008. Published December 9, 2008.
The interfacial properties of semiconductor electrodes can be
controlled through deliberate chemical functionalization.1-8 A two-
step chlorination–alkylation reaction sequence has been shown to
modify Si with covalently attached alkyl groups,9 significantly re-
ducing the rate of oxidation of Si in air while simultaneously pro-
ducing a low density of interfacial electrical trap states at the Si
surface.9,10
The presence of covalently bonded alkyl functionalities also in-
troduces a surface dipole that shifts the energetics of the Si valence
͑Evb͒ and conduction ͑Ecb͒ bandedges at the interface toward the
energy of the vacuum level ͑Evac͒.11 UV photoelectron spectro-
scopic data obtained under ultrahigh-vacuum conditions have indi-
cated that the electron affinity ͑␹͒ ͑the energy difference from Ecb to
Evac͒ is ϳ0.5 eV less for CH3–Si͑111͒ surfaces than for H–Si͑111͒
surfaces.12 For an ideal contact between an n-type semiconductor
and another conductive phase ͑e.g., a metal, a conductive polymer,
or a solution with a dissolved redox couple͒, the equilibrium junc-
tion barrier height, ⌽b, is given by
Extension of the Si/Hg results to other metals thus requires soft
methods for deposition of metal contacts onto Si surfaces. Elec-
trodeposition has been used extensively to form microscale and
macroscale metal/semiconductor contacts18-30 as well as to deposit
metal-catalyst islands onto semiconductor photoelectrodes.25,31-35
An attractive feature of electrodeposition relative to other metal-
deposition techniques is that the process can be monitored in situ
while providing qualitative and quantitative information on the en-
ergetics and physical state of the semiconductor surface.18,29,30,36,37
We report herein the electrodeposition of Cd and Pb films onto
H- and CH3-terminated n-Si͑111͒. Neither Cd nor Pb readily forms
silicides at ambient temperature and pressure,14 so these metals pro-
vide interesting systems to evaluate whether electrodeposited metal
contacts can yield the expected variation in barrier height as
the dipole of the Si͑111͒ surface is changed by alkylation. The elec-
trochemical responses, and corresponding solid-state electrical junc-
tion data, of electrodeposited Cd and Pb junctions using degener-
ately doped and nondegenerately doped n-type H–Si͑111͒ and
CH3–Si͑111͒ surfaces have been investigated to determine whether
CH3 monolayers predictably and consistently influence the electro-
chemical and electrical properties of Si interfaces. In addition,
chronoamperometric data have been used to compare the available
nucleation-site density for metal electrodeposition onto CH3–
Si͑111͒ and H–Si͑111͒ surfaces, allowing an investigation of
whether these different surface functionalities can effect different
metal-plating properties on Si surfaces.
− q⌽b = EF + ␹
͓1͔
where q is the unsigned elementary electronic charge, EF is the
Fermi level of the conductive phase, and more negative energies are
further from the vacuum level. Hence, the change in ␹ produced by
alkylation should produce different barrier heights for ideal
semiconductor/metal Schottky barriers on H–Si͑111͒ and
CH3–Si͑111͒ surfaces.
For Si͑111͒/Hg Schottky junctions, the measured differences in
⌽b between H-terminated and CH3-terminated Si͑111͒ are in excel-
lent agreement with the change in ␹ measured upon alkylation.13
The ideal behavior of the Hg junctions likely arises from the “soft”
nature of Hg contacts to Si, because Hg-silicide does not form at
ambient conditions and Hg does not react with the Si–C surface
bonds.14-16 In contrast, Schottky barriers produced by thermal
evaporation of Cu, Au, and Ag onto CH3–Si͑111͒ do not exhibit the
expected increase in barrier height relative to Schottky barriers pro-
duced by evaporation of these metals onto H–Si͑111͒ surfaces.17
This behavior likely reflects the propensity of evaporated transition
metals to form interfacial silicides14 or to disrupt the CH3–Si surface
layer.
Experimental
Materials.— Single-side polished, nondegenerately phosphorus-
doped ͑n-type͒, 525 ␮m thick Si͑111͒ wafers, with miscut angles
Ϯ0.5° and resistivities between 1 and 10 ⍀ cm, were purchased
from ITME ͑Warsaw, Poland͒. Double-side polished, degenerately
phosphorus-doped ͑n+͒, 300 ␮m thick silicon wafers, with miscut
angles Ϯ0.5° and nominal resistivities of ഛ0.004 ⍀ cm, were ob-
tained from Addison Engineering, Inc. ͑San Jose, California͒. Prior
to chemical treatment, the specific resistivity of each wafer was
determined using four-point-probe methods. Replicate measure-
ments were conducted using Si samples cut from the same parent
wafer.
All chemicals used for silicon cleaning, silicon surface function-
alization, and electrochemical analyses were used as received. Water
with a resistivity Ͼ18 M ⍀ cm was obtained from a Barnstead
Nanopure system and was used throughout.
*
Electrochemical Society Active Member.
z E-mail: nslewis@its.caltech.edu

H124
Journal of The Electrochemical Society, 156 ͑2͒ H123-H128 ͑2009͒
Si surface functionalization.— Before surface functionalization,
a)
b)
10
5
10
5
silicon wafers were sectioned into 1 ϫ 5 cm pieces that were de-
greased by sequentially immersing and sonicating in methanol ͑Al-
drich͒, acetone ͑Aldrich͒, 1,1,1-trichloroethane ͑Aldrich͒, dichlo-
romethane ͑Fisher͒, 1,1,1-trichloroethane, acetone, and methanol for
5 min each. The Si pieces were then immersed in a 3:1 ͑by volume͒
solution of 18 M H2SO4 ͑Fisher͒ and 30% H2O2͑aq͒ ͑v/v, Fisher͒ at
ϳ80° for ജ1 h, followed by a rinse with H2O. H-terminated
Si͑111͒ surfaces were prepared by immersion of these Si samples for
10 min in 11 M NH4F͑aq͒ ͑Transene͒,38-42 with periodic agitation
to minimize accumulation of bubbles at the surface. The Si samples
were then rinsed thoroughly with water and were used immediately
for further chemical functionalization or for electrochemical analy-
sis. The samples that were used for electrochemical measurements
without further functionalization were contacted with In–Ga eutectic
on the back side prior to etching the top surface.
CH3-terminated Si͑111͒ surfaces were prepared as described
previously.13,43 First, freshly H-terminated Si͑111͒ surfaces were
chlorinated by immersion for 50 min at 90°C into a saturated PCl5
͑Aldrich͒ solution in chlorobenzene ͑Aldrich͒. The chlorine-
terminated Si͑111͒ samples were then rinsed sequentially with chlo-
robenzene and tetrahydrofuran ͑THF, Aldrich͒, followed by immer-
sion for 8–12 h at 70°C in a 1 M CH3MgBr solution in THF. The
CH3-terminated Si͑111͒ surfaces were then rinsed with THF, rinsed
with methanol, and sonicated sequentially in methanol, acetone, and
methanol for 5 min each. The samples were then rinsed with water
and dried under a stream of N2͑g͒.
Cd2+/0
Pb2+/0
0
0
-5
-1.2 -0.9 -0.6 -0.3 0.0
-1.2 -0.9 -0.6 -0.3 0.0
E /V vs. SCE
E /V vs. SCE
Figure 1. Representative cyclic voltammograms of ͑dashed͒ H-terminated
and ͑solid͒ CH3-terminated nondegenerately doped n-Si͑111͒ electrodes im-
mersed in deaerated aqueous acidic solutions: ͑a͒ J-E data for electrodepo-
sition of Cd from a 0.1 M H2SO4 solution containing 0.005 M CdSO4 and
͑b͒ J-E data for electrodeposition of Pb from a 0.1 M HClO4 solution con-
taining 0.005 M Pb͑ClO4͒2. Scan rate: 0.030 V s−1
potential steps from open circuit ͑−0.1 Ϯ 0.1 V͒ to − 0.95 V vs
SCE were used, with a subsequent 10 min potential step at −0.88 V
vs SCE.
All voltammetry and chronoamperometry data were collected
with the electrodes in the dark, in deaerated solutions, with scans
initiated at the OCP of the working electrode. For electrochemical
current density–potential ͑J-E͒ data, cathodic ͑reducing͒ and anodic
͑oxidizing͒ currents are plotted as negative and positive, respec-
tively. For steady-state electrical current density–voltage ͑J-V͒ data,
forward-bias voltages and currents are plotted as positive. Voltam-
metric data were obtained using a Solartron Instruments potentiostat
͑SI1287͒ interfaced to a computer for data acquisition. All reported
cyclic voltammetric data were recorded at a scan rate of 0.030 V s−1
and were averages of separate measurements on at least five differ-
ent samples. Chronoamperometry was performed with a Princeton
Applied Research ͑PAR͒ model 362 potentiostat coupled to a PAR
model 175 universal programmer. Data were collected with a Tek-
tronix TK200 digital oscilloscope that was interfaced to a data-
acquisition computer. A spring-loaded brass top contact was used to
make electrical connection to the deposited metal film for solid-state
J-V measurements.
Electrochemical metal deposition and solid-state device mea-
surements.— After chemical treatment, the n-Si͑111͒ samples were
diced into 1 ϫ 1 cm pieces. To make ohmic contacts to these Si
samples, the back side was lightly scratched with a diamond scribe,
etched for 10 min with 11 M NH4F͑aq͒, and then rubbed with a
Ga–In eutectic. The Si samples were then placed on a 1 ϫ 2.5
ϫ 0.2 cm stainless steel support and were fitted onto a stainless steel
base plate. A Teflon cell with a 0.2 cm diameter Viton O-ring was
then press-fit on top of the Si sample, exposing a surface area of
0.128 cm2.13
For full metal-film preparation, a two-step nucleation-growth
chronoamperometric protocol was used.44 A standard three-electrode
setup with a reference and a Pt mesh counter electrode was used.
Solutions for Cd electrodeposition consisted of 5 mM CdSO4
͑Fisher͒ dissolved in 0.1 M H2SO4͑aq͒ ͑Fisher͒. Pb electrodeposi-
tion solutions contained 5 mM Pb͑ClO4͒2 ͑Aldrich͒ in 0.1 M
HClO4͑aq͒ ͑Fisher͒.
Metal-film electrodeposition was performed using a combination
of potential steps at high overpotentials, followed by steps to low
overpotentials to generate finely sized and spaced metal nuclei45-47
and to subsequently produce even film growth between metal
clusters.48 For Cd electrodeposition on H-terminated n-Si͑111͒, and
for both types of n+-Si͑111͒ electrode samples, a series of ten
100 ms steps, from open circuit to −0.90 V vs a saturated calomel
electrode ͑SCE͒, were applied, followed by a 10 min step at
−0.75 V vs SCE. The open-circuit potentials ͑OCPs͒ were
−0.10 Ϯ 0.08 and −0.2 Ϯ 0.2 V vs SCE for H-terminated n-Si͑111͒
and n+-Si͑111͒ electrodes, respectively. The OCP for CH3
-terminated n+-Si͑111͒ was 0.17 Ϯ 0.02 V vs SCE. For Cd-film
electrodeposition on CH3-terminated n-Si͑111͒, thirty 100 ms poten-
tial steps, from open circuit ͑−0.05 Ϯ 0.04 V͒ to − 1.0 V vs SCE,
were used, followed by a single 10 min step at −0.90 V vs SCE. For
Pb electrodeposition on H-terminated n-Si͑111͒, and for both
n+-Si͑111͒ surfaces, ten 100 ms potential steps, from open circuit to
−0.60 V vs SCE, were applied, followed by a 10 min step at
−0.55 V vs SCE. The OCPs for H-terminated n-Si͑111͒ and
n+-Si͑111͒ electrodes were −0.09 Ϯ 0.04 and −0.22 Ϯ 0.02 V vs
SCE, respectively, and was 0.1 Ϯ 0.1 V vs SCE for CH3-terminated
n+-Si͑111͒. For CH3-terminated n-Si͑111͒ electrodes, thirty 100 ms
Characterization.— X-ray photoelectron spectroscopic ͑XPS͒
data were obtained using an M-Probe spectrometer equipped with a
1486.6 eV Al K␣ X-ray source.49 The pressure during data collec-
tion was ഛ5 ϫ 10−9 Torr, and data were obtained using an incident
angle of 35° off of the surface normal. For all samples, Si 2p spectra
were obtained in the 106–96 eV energy region. All spectra were
averaged over 30 scans that were obtained with a spot diameter of
300 ␮m and a resolution of 0.6 eV. Following the nucleation step,
metallized Si samples were immediately loaded into the spectro-
meter, and scans were acquired after the base pressure had been
obtained.
Results
Voltammetric responses of n-type Si(111) in aqueous Cd2ϩ and
Pb2ϩ solutions
.— Figure 1 displays representative J-E data for
n-Si͑111͒ working electrodes immersed in aqueous acidic solutions
that contained either dissolved Cd2+ or Pb2+. In accord with previous
reports, reversible Cd or Pb metal reduction and metal oxidation
͑stripping͒ was observed for freshly etched, H-terminated, n-Si͑111͒
electrodes ͑Fig. 1a and b͒.26-29 The ratios of the anodic to cathodic
charge density ͑Qa/Qc͒ passed in the voltammograms in Fig. 1a and
b for H-terminated n-Si͑111͒ were near unity ͑Table I͒, indicating
that these Si/metal junctions were not strongly rectifying, i.e., an-
odic and cathodic current densities ӷ1 mA cm−2 in magnitude were
observed at positive and negative potentials, respectively, under
these experimental conditions.

Journal of The Electrochemical Society, 156 ͑2͒ H123-H128 ͑2009͒
Table I. Voltammetric responses for metal electrodeposition on n-type Si(111).a
H125
Surface
functionalization
b
c
d
e
f
Metal
Cd
Ep,c ͑V vs SCE͒
͉⌬Ep,c͉ ͑V͒
Ep,a ͑V vs SCE͒
͉⌬Ep,a͉ ͑V͒
Qa/Qc
ND ͑cm−3
͒
q⌽JV ͑eV͒
Nondegenerate doping
–H
–CH3
−0.833 Ϯ 0.003
−0.964 Ϯ 0.017
−0.586 Ϯ 0.012
−0.656 Ϯ 0.008
0.90 Ϯ 0.03
͑3 Ϯ 2͒ ϫ 1015
ഛ0.4
0.31 Ϯ 0.04 ͑1.5 Ϯ 0.3͒ ϫ 1015 0.81 Ϯ 0.05
0.131 Ϯ 0.017
0.000 Ϯ 0.008
0.070 Ϯ 0.014
0.004 Ϯ 0.005
Degenerate doping
–H
–CH3
−0.827 Ϯ 0.004
−0.827 Ϯ 0.007
−0.585 Ϯ 0.003
−0.581 Ϯ 0.004
0.962 Ϯ 0.007 ͑3.4 Ϯ 0.5͒ ϫ 1019
0.966 Ϯ 0.006 ͑3.4 Ϯ 0.5͒ ϫ 1019
—
—
Pb
Nondegenerate doping
–H
–CH3
−0.577 Ϯ 0.010
−0.924 Ϯ 0.031
−0.408 Ϯ 0.003
0.84 Ϯ 0.02
͑2 Ϯ 1͒ ϫ 1015
ഛ0.59
—
0.02 Ϯ 0.01 ͑1.3 Ϯ 0.9͒ ϫ 1015 0.82 Ϯ 0.03
0.347 Ϯ 0.024
0.006 Ϯ 0.030
—
Degenerate doping
–H
–CH3
−0.577 Ϯ 0.014
−0.583 Ϯ 0.026
−0.410 Ϯ 0.005
−0.401 Ϯ 0.005
0.91 Ϯ 0.03 ͑3.4 Ϯ 0.5͒ ϫ 1019
0.90 Ϯ 0.02 ͑3.4 Ϯ 0.5͒ ϫ 1019
—
—
0.009 Ϯ 0.007
a T = 295 Ϯ 3 K; 1 scan at 0.030 V s−1; the averages and standard deviations were determined from five separate samples for each condition.
b
͉⌬Ep,c͉ = ͉Ep,c,Si–H − Ep,c,Si–CH ͉.
͉⌬Ep,a͉ = ͉Ep,a,Si–H − Ep,a,Si–CH ͉.
3
c
d Qc = integrated cathodic ͑met3al deposition͒ charge density passed; Qa = integrated anodic ͑stripping͒ charge density passed.
e Measured by four-point probe.
f Measured from dark J-V responses of electrochemically deposited dry device.
n-Si͑111͒ electrodes. For Cd electrodeposition at 0.030 V s−1
,
Essentially, identical behavior was observed when the dopant
density of the H-terminated n-type Si͑111͒ electrodes was increased
from 1015 to 1019 cm−3 ͑Fig. 2͒. The Qa/Qc ratios were only slightly
larger for H-terminated n+-Si͑111͒ electrodes than for H-terminated
n-Si͑111͒ electrodes ͑Table I͒. The cathodic and anodic peak poten-
tials ͑Ep,c and Ep,a, respectively͒ for Cd2+ or Pb2+ electrodeposition
were statistically indistinguishable between nondegenerately doped
and degenerately doped H-terminated Si͑111͒ electrodes ͑Table I͒.
The close similarity between the data for nondegenerately doped
and degenerately doped H-terminated Si͑111͒ indicates that the ob-
served electrodeposition overpotentials are not dominated by the
presence of a significant built-in voltage ͑Vbi͒ in the H-terminated
n-Si͑111͒ electrodes.
changing from H- to CH3-terminated n-Si͑111͒ produced a shift in
Ep,c of −0.131 Ϯ 0.007 V ͑Fig. 1a͒, while for Pb electrodeposition
the value of Ep,c shifted by −0.347 Ϯ 0.024 V ͑Fig. 1b͒. In addition,
a greater extent of “crossover” was observed between the forward
and reverse scans37,50 on CH3-terminated n-Si͑111͒ than on
H-terminated n-Si͑111͒. Also, the anodic currents on the oxidative
scan were suppressed on CH3-terminated n-Si͑111͒ for both Cd2+
and Pb2+, with Qa/Qc values much smaller than unity for Cd2+ and
nearly 0 for Pb2+ ͑Fig. 1, Table I͒. For oxidative stripping of Cd,
changing from H–Si͑111͒ to CH3–Si͑111͒ electrodes produced a
shift in Ep,a of −0.070 Ϯ 0.014 V, whereas no anodic voltammetric
wave for Pb oxidative stripping was observed with CH3-terminated
n-Si͑111͒ electrodes.
In contrast, significant differences were observed between the
voltammetry of Cd2+ or Pb2+ at CH3- and H-terminated n-Si͑111͒
electrodes ͑Fig. 1͒. The voltammetric wave shapes and peak heights
observed with CH3-terminated n-Si͑111͒ electrodes were less re-
versible in both reduction and oxidation than with H-terminated
Unlike the situation for H-terminated Si͑111͒ surfaces, the volta-
mmetric behavior of CH3-terminated n-type Si͑111͒ was strongly
influenced by the dopant density, ND. In fact, the voltammetric re-
sponses of CH3-terminated n+-Si͑111͒ electrodes closely resembled
those of H-terminated Si͑111͒ ͑Fig. 2͒. The values of Qa/Qc, Ep,c
,
15
and Ep,a for Cd and Pb electrodeposition did not differ statistically
between CH3-terminated and H-terminated n+-Si͑111͒ electrodes
͑Fig. 2, Table I͒. The CH3-terminated n+-Si͑111͒ electrodes exhib-
ited clear and narrow stripping peaks on the back scan, with values
of Qa/Qc Ϸ 1.0 within this potential range.
a)
b)
10
5
Cd2+/0
Pb2+/0
10
5
The change in the voltammetry of CH3-terminated Si͑111͒ elec-
trodes in response to a change in the bulk doping density implies
that a significant built-in voltage within the near-surface region of
the Si contributed to the electrodeposition/oxidation overpotentials
observed for CH3-terminated n-Si͑111͒ electrodes. The contribution
of Vbi to the required electrodeposition overpotential is minimized at
0
0
-5
-1.2 -0.9 -0.6 -0.3 0.0
-1.2 -0.9 -0.6 -0.3 0.0
E /V vs. SCE
E /V vs. SCE
n+-Si͑111͒ electrodes, because the depletion-layer width is suffi-
51
Figure 2. Representative cyclic voltammograms of ͑dashed͒ H-terminated
and ͑solid͒ CH3-terminated degenerately doped n+-Si͑111͒ electrodes im-
mersed in deaerated aqueous acidic solutions: ͑a͒ J-E data for electrodepo-
sition of Cd from a 0.1 M H2SO4 solution containing 0.005 M CdSO4 and
͑b͒ J-E data for electrodeposition of Pb from a 0.1 M HClO4 solution con-
taining 0.005 M Pb͑ClO4͒2. Scan rate: 0.030 V s−1
ciently thin ͑ϳ10−9 m for ⌽b = 1 V for Si with ND = 1019 cm−3
͒
that large tunneling currents of electrons from the bulk semiconduc-
tor through the depletion region into solution are possible.37,52 The
similarity of the values of Ep,c and Ep,a for CH3-terminated and
H-terminated n+-Si͑111͒ electrodes further indicates that charge

H126
Journal of The Electrochemical Society, 156 ͑2͒ H123-H128 ͑2009͒
a)
b)
c)
a)
b)
6
4
2
0.4
0.2
0.0
0.4
0.2
0.0
7
6
5
10
n-Si/Cd
n-Si/Pb
5
0
-0.6 -0.3 0.0
0.3
0.6
-0.6 -0.3 0.0
0.3
0.6
0.0 0.05 0.10
0.0
0.3
0.6
0.0
0.5
3/2 x 102/s3/2
3/2 x 103/s3/2
V /V
V /V
t
t /s
t
Figure 4. Representative electrical responses of electrodeposited metal/n-
Si͑111͒ Schottky junctions: ͑a͒ J-V data for Cd Schottky junctions with
͑dashed͒ H-terminated and ͑solid͒ CH3-terminated n-Si͑111͒, where T
= 296 Ϯ 2 K, and ͑b͒ J-V data for Pb Schottky junctions with ͑dashed͒
H-terminated and ͑solid͒ CH3-terminated n-Si͑111͒, where T = 296 Ϯ 2 K.
Film thicknesses were approximately 200 nm.
Figure 3. Potentiostatic pulse deposition of Cd nuclei at −0.9 V vs SCE in
deaerated 0.1 M H2SO4 with 0.005 M CdSO4: ͑a͒ J-t responses for ͑open
circles͒ H-terminated and ͑closed squares͒ CH3-terminated n+-Si͑111͒ elec-
trodes, ͑b͒ J-t3/2 response for H-terminated n+-Si͑111͒, and ͑c͒ J-t3/2 re-
sponse for CH3-terminated n+-Si͑111͒. The y axes for ͑b͒ and ͑c͒ have been
chosen to emphasize the linear region used for analysis, as described in the
text.
Electrical properties of electrochemically deposited n-Si(111)/Cd
and n-Si(111)/Pb Schottky junctions.— Figure 4 depicts the electri-
cal properties of electrodeposited n-Si/Cd and n-Si/Pb Schottky
junctions. At room temperature, the junctions with CH3-terminated
n-Si͑111͒ were considerably more rectifying than the junctions with
H-terminated n-Si͑111͒ ͑Fig. 4a and b͒. The values of ⌽b for the Cd
and Pb Schottky devices with CH3-Si͑111͒ surfaces were 0.81 and
0.82 Ϯ 0.03 V, respectively, as determined from the dark J-V data
͑Table I͒. In contrast, the H-terminated n-Si͑111͒ substrates exhib-
ited much lower barrier heights, with Cd and Pb junctions exhibiting
transfer is not substantially ͑i.e., by Ͼ6 mV overpotential͒ impeded
by through-bond tunneling through the surface passivation layer.53
Chronoamperometric deposition of Cd onto H-terminated and
CH -terminated nϩ-Si͑111͒
.— Continuous metal films were more
3
readily deposited on H-Si͑111͒ than on CH3-Si͑111͒ surfaces. This
effect was investigated quantitatively by measurement of the tran-
sient currents for the first nucleation step on H- and CH3-terminated
n+-Si͑111͒ working electrodes. At an overpotential of ϳ150 mV
and long times, the total charge passed was essentially the same at
both electrode types. In contrast, the initial current density–time
͑J-t͒ responses at this overpotential differed significantly. Figure 3
depicts representative chronoamperometric data at short times for
the initial Cd nucleation step for both H- and CH3-terminated
n+-Si͑111͒ electrodes biased at −0.90 V vs SCE, with the
H-terminated n+-Si͑111͒ samples exhibiting larger initial transient
currents. Following the analysis of Hills et al.,54 the currents re-
corded at early times provide mechanistic information on the metal-
nucleation process. For an electrode held at a constant potential,
progressive nucleation, in which the number of nucleation events
⌽
Ͻ0.4 and Ͻ0.59 V, respectively. For these junctions, the J-V
b
responses were not sufficiently rectifying at room temperature to
allow unambiguous determination of ⌽b from the J-V data.13
A
small fraction of the tested H-terminated n-Si͑111͒/Pb devices ex-
hibited values of ⌽b as large as 0.59 V, but the majority of the
H-terminated n-Si͑111͒/Pb junctions showed poor or no rectifica-
tion, consistent with previous reports.56 Hence, the values for elec-
trochemically prepared H-terminated n-Si͑111͒ devices are reported
as upper estimates of ⌽b. For all junctions that exhibited rectifica-
tion, the overall diode quality factors ranged between 1.8 and 2.
X-ray photoelectron spectra of Si(111) surfaces after elec-
37,54
increases with time ͑t͒, is indicated if J is linear with t3/2
trodeposition of Cd and Pb.— Figure
5 displays representative
1/2
3/2
3/2
high-resolution Si 2p XPS data for H- and CH3-terminated n-Si͑111͒
zNϱ␲͑2DCd2+CCd2+
1/2
͒
M
t
Cd
2+
J͑t͒ =
͓2͔
3␳
Cd
where n is the number of electrons transferred, DCd2+ is the diffusion
coefficient of Cd2+ in solution ͑cm2 s−1͒, CCd2+ is the solution con-
centration of Cd2+ ͑moles cm−3͒, MCd2+ is the atomic mass of Cd
a)
b)
͑g mol−1͒, ␳ is the density of Cd ͑g cm−3͒, and Nϱ is the nucle-
As-prepared
NH4F etch
Cd
ation density ͑cm−2͒ of Cd at infinite time. Figures 3b and c explic-
itly depict the early-time, linear dependence of J with t3/2 for the H-
and CH3-terminated n+-Si͑111͒ electrodes of Fig. 3a. From the
slopes of the linear fits in Fig. 3b and c, the values of Nϱ for
H-terminated Si͑111͒ and CH3-terminated Si͑111͒ surfaces were de-
termined to be ͑1.1 Ϯ 0.2͒ ϫ 109 cm−2 and ͑0.13 Ϯ 0.05͒
ϫ 109 cm−2, respectively.
Cd nucleation
Cd nucleation
For electrodeposition of Au or Ag on freshly etched single-
crystalline Si, values of Nϱ at mild overpotentials ͑Ͻ200 mV͒,
where diffusion of electroactive species is not the dominating pro-
Pb nucleation
Pb nucleation
cess in nucleation, have been reported to be between 109 and
18,19,44,55
1011 cm−2
.
These Nϱ values are in agreement with the data
105
102
99
105
102
99
observed herein for deposition of Cd or Pb on H-terminated
n+-Si͑111͒. In contrast, the total apparent number of existing nucle-
ating sites for a one-pulse electrodeposition experiment of Cd or Pb
on CH3-terminated Si͑111͒ was approximately an order of magni-
tude smaller than that for 11 M NH4F͑aq͒ etched n+-Si͑111͒ elec-
trodes.
Binding Energy /eV
Binding Energy /eV
Figure 5. X-ray photoelectron spectra of the Si 2p region of ͑a͒
H-terminated and ͑b͒ CH3-terminated n-Si͑111͒ electrode surfaces ͑top͒ as-
prepared, ͑middle͒ after initial Cd nucleation step, and ͑bottom͒ after initial
Pb nucleation step.

Journal of The Electrochemical Society, 156 ͑2͒ H123-H128 ͑2009͒
H127
electrodes after the initial metal nucleation step during the deposi-
tion of Cd or Pb films. No oxide signatures were apparent between
101 and 105 eV for any of the analyzed surfaces. For both H- and
CH3-terminated n-Si͑111͒ surfaces, these results suggest that neither
Cd nor Pb electrodeposition resulted in interfacial silicon oxide
thicknesses ജ1 Å.
The data depicted in Fig. 2 argue against the plausibility of
through-bond tunneling across the CH3 groups as a significant factor
in the differences observed between metal electrodeposition at CH3-
and H-terminated n-type Si͑111͒ electrodes. On Au͑111͒ electrodes,
Schneeweiss et al. have shown that butanethiol, pentanethiol, and
hexanethiol monolayers can act as large electron tunneling barriers
that slow the overall rate of Cu electrodeposition at low overpoten-
tials. These insulating monolayers effected Ͼ200 mV shifts in the
peak potential of the cyclic voltammetry of Cu electrodeposition.53
For comparison, with electrodeposition of Cd and Pb, respectively,
onto Si ͑Fig. 2͒, the values of Ep,c and Ep,a were not statistically
different between the two n+-Si͑111͒ electrode types. These data
suggest similar metal electrodeposition charge-transfer rate con-
stants for these H- and CH3-terminated Si surfaces. For nondegen-
erately doped semiconductor electrodes, evaluation of explicit
charge-transfer kinetics from descriptive cylic voltammetric features
such as ⌬Ep is not straightforward, because the J-E responses of
semiconductor electrodes are sensitive to the standard rate constant
of the charge-transfer process and to the potential dependence of the
population of transferable electrons at the interface.52,60-63 For metal
electrodeposition/oxidation, the determination of interfacial charge-
transfer rate constants is further complicated by the spatial disper-
sion of deposited metal nuclei, the magnitude of current flow at the
metal/semiconductor contacts, and the complex, multicharge-
transfer step mechanisms.29,30,48,64 Accordingly, the data presented
herein do not allow for direct comparison of explicit charge-transfer
rate constants at H- and CH3-terminated Si͑111͒ electrodes. Never-
theless, the similarities between the electrodeposition behavior for
both Cd and Pb onto degenerately doped CH3-terminated
n+-Si͑111͒ indicates that through-bond electron tunneling is not a
dominant factor in these electrodeposition processes at
CH3-terminated n-Si͑111͒ electrodes.
The electrical and electrochemical data were also not influenced
significantly by insulating interfacial silicon oxide layers. Although
oxidation of Si interfaces immersed in aqueous electrolytes at pH 1
is thermodynamically favored at potentials ജ − 1.1 V vs
Ag/AgCl,58 the data in Fig. 5 suggest that the formation of a
multilayer oxide on n-Si in deaerated acidic solutions, biased ca-
thodically and in the absence of illumination, is kinetically slow on
the timescale of the deposition experiments. Unlike aqueous acidic
Au3+ or Ag+ solutions, which readily oxidize Si surfaces, solutions
of dissolved Cd2+ or Pb2+ apparently do not have sufficient oxidiz-
ing power to form appreciable SiOx layers on either CH3- or
H-terminated Si.27,58 Hence, under the conditions employed herein,
the electrochemical and electrical properties of the Si/metal inter-
faces reflect the native surface chemistries of CH3- and
H-terminated Si͑111͒.
Discussion
Influence of surficial CH3-groups on the barrier height of Si/Cd
and Si/Pb Schottky junctions.— The electrochemical and electrical
data support the expectation that the presence of a surface dipole
arising from the attachment of CH3- groups to Si͑111͒ surfaces will
substantially ͑i.e., by ϳ0.5 V͒ shift the bandedges closer to the
vacuum level as compared to the H-terminated Si͑111͒
surface.11-13,57 The net change in the value of ␹ at the Si͑111͒ inter-
face is manifest in the increased junction barrier heights for the
investigated CH3-terminated Si͑111͒/metal and Si͑111͒/solution
junctions. The J-V properties of the electrodeposited dry
n-Si͑111͒/Cd and n-Si͑111͒/Pb Schottky junctions were ohmic for
H-terminated n-Si͑111͒, while these same junctions were strongly
rectifying with CH3-terminated Si͑111͒. These Cd and Pb Schottky
junctions prepared by electrodeposition displayed electrical proper-
ties that are in agreement with the previously reported J-V responses
of n-Si͑111͒/Hg junctions13 and support the hypothesis that Schottky
junctions prepared by thermal metal evaporation do not reflect the
innate properties of the CH3-terminated Si interfaces because of
physical damage or chemical reactivity during the deposition pro-
cess.
The voltammetric data of the n-type Si electrodes in contact with
the metal plating solutions further illustrate the differences in sur-
face energetics of CH3-terminated and H-terminated Si͑111͒. The
close correspondence between the charge density passed in the for-
ward and reverse sweeps of the voltammograms in Fig. 2 for
H-terminated Si͑111͒ is consistent with the hypothesis that these Si
electrodes are not strongly depleted48 during Cd and Pb electrodepo-
sition, i.e., the H-terminated n-Si͑111͒ interface is not strongly
rectifying toward reduction/oxidation of either Cd2+/0 or Pb2+/0
.
In acidic ͑non-HF containing͒ aqueous solutions, the potential of
the conduction bandedge, Ecb, of H-terminated Si͑111͒ has been
estimated as −0.50 Ϯ 0.05 V vs SCE.58 The potentials of Cd2+/0
and Pb2+/0 electrodeposition ͑−0.74 and −0.52 V vs SCE,
28,29
respectively͒
are near, or more negative than, the conduction
bandedge of H-terminated n-Si͑111͒, so a depletion region in
n-Si͑111͒ is not expected during the electroreduction of these met-
als. The identical features of the voltammetric responses displayed
for H-terminated n-Si͑111͒ and n+-Si͑111͒ electrodes support this
premise, with the surface energetics and Vbi having minimal influ-
ence on the interfacial charge transfer at degenerately doped semi-
conductor electrodes.52
Chemical passivation of Si by CH3 surface groups.— Chemical
modification of Si͑111͒ surfaces by the chlorination/Grignard alkyl-
ation method results in an essentially complete, protective layer at-
tached to the topmost Si atoms. Scanning tunneling microscopy of
CH3-terminated Si͑111͒ surfaces has revealed a low density of de-
fects in the monolayer coverage of chemically grafted CH3
groups.65 Transient photoconductivity measurements have also dem-
onstrated that CH3-terminated Si͑111͒ surfaces possess an unusually
low density of electrically active surface trap states.10 The lower
measured value of Nϱ for CH3-terminated Si͑111͒ surfaces relative
to NH4-etched Si͑111͒ surfaces is in qualitative agreement with
these prior observations. Direct comparisons between the areal den-
sities of nucleation sites and electrical trap states are confounded by
the well-known difficulties in precisely assessing the nucleation site
density solely from electrochemical data.37,66 Nevertheless, because
electrodeposition of metal films occurs by metal cluster nucleation
at individual high-energy surface sites that then grow larger and fuse
together,67-69 the observation of hindered film growth on
CH3-terminated Si͑111͒ agrees with expectations and with the chro-
noamperometric data. From a practical standpoint, a decrease in the
available nucleation site density on Si͑111͒ as a result of the intro-
In contrast, the voltammetric responses indicate that
CH3-terminated n-Si͑111͒ electrodes are largely depleted of majority
carriers during the electrodeposition of Cd or Pb. The susceptibility
of CH3-terminated Si͑111͒ surfaces to electrochemical oxidation at
potentials Ͼ0.3 V vs SCE59 precluded direct measurement of Ecb
and ⌽b by standard impedance methods. However, examination of
the J-E behavior of CH3-terminated n-Si͑111͒ electrodes in aqueous
solutions containing dissolved methyl viologen has demonstrated
that the bandedges of CH3-terminated Si͑111͒ are invariant to
changes in solution pH and in acidic solutions are situated at sig-
nificantly ͑i.e., 0.3–0.4 V͒ more negative potentials than the band-
edges of freshly etched n-Si͑111͒.11 The electrochemical behavior of
CH3-terminated n+-Si͑111͒ is expected to more closely approximate
that of H-terminated Si͑111͒ electrodes, if the additional elec-
trodeposition overpotentials arise largely from Vbi. Comparison of
Fig. 1 and 2 indeed shows a significant difference in the voltamme-
tric behavior of CH3-terminated Si͑111͒ electrodes as the bulk Si
donor density is varied, consistent with the hypothesis that
CH3-terminated n-Si͑111͒ surfaces are in depletion for the elec-
trodeposition of Cd and Pb under our experimental conditions.

H128
Journal of The Electrochemical Society, 156 ͑2͒ H123-H128 ͑2009͒
16. Y. J. Liu and H. Z. Yu, ChemPhysChem, 4, 335 ͑2003͒.
17. R. Hunger, R. Fritsche, B. Jaeckel, L. J. Webb, W. Jaegermann, and N. S. Lewis,
Surf. Sci., 601, 2896 ͑2007͒.
18. M. L. Munford, F. Maroun, R. Cortes, P. Allongue, and A. A. Pasa, Surf. Sci., 537,
95 ͑2003͒.
19. K. Marquez, G. Staikov, and J. W. Schultze, Electrochim. Acta, 48, 875 ͑2003͒.
20. J. M. DeLucca, S. E. Mohney, F. D. Auret, and S. A. Goodman, J. Appl. Phys., 88,
2593 ͑2000͒.
21. M. E. Kiziroglou, A. A. Zhukov, M. Abdelsalam, X. L. Li, P. A. J. de Groot, P. N.
Bartlett, and C. H. de Groot, IEEE Trans. Magn., 41, 2639 ͑2005͒.
22. P. M. Vereecken, K. Rodbell, C. X. Ji, and P. C. Searson, Appl. Phys. Lett., 86,
121916 ͑2005͒.
23. G. Oskam, D. Vanmaekelbergh, and J. J. Kelly, Electrochim. Acta, 38, 1115
͑1993͒.
24. K. Ghosh and N. K. D. Chowdhury, Int. J. Electron., 54, 615 ͑1983͒.
25. Y. L. Kawamura, T. Sakka, and Y. H. Ogata, J. Electrochem. Soc., 152, C701
͑2005͒.
26. B. Rashkova, B. Guel, R. T. Potzschke, G. Staikov, and W. J. Lorenz, Electrochim.
Acta, 43, 3021 ͑1998͒.
27. J. C. Ziegler, A. Reitzle, O. Bunk, J. Zegenhagen, and D. M. Kolb, Electrochim.
duction of surficial CH3 groups implies possible difficulty with the
uniform and dense decoration of such surfaces with electrodepos-
ited, microscopic, metallic electrocatalysts. Poorly controlled unifor-
mity and spatial dispersion of electrodeposited, metallic electrocata-
lysts on semiconductor surfaces can be detrimental to the use of
such modified semiconductors in photoelectrochemical systems,
where charge-carrier recombination at the semiconductor surface de-
pends critically on the size and spacing of metallic clusters.70,71
Interestingly, Takabayashi et al. reported significant differences in
catalyst dispersion on n-Si͑111͒ photoelectrodes modified with dif-
ferent alkyl surface moieties.34 Irregular catalyst dispersion and low
resultant energy-conversion efficiencies were reported with n-Si
photoelectrodes modified by linear alkyl groups. We are currently
exploring the use of unsaturated organic surface groups to provide
ample sites for catalyst attachment while simultaneously controlling
the underlying electrical surface passivation properties.72
Acta, 45, 4599 ͑2000͒.
Conclusions
28. J. C. Ziegler, R. I. Wielgosz, and D. M. Kolb, Electrochim. Acta, 45, 827 ͑1999͒.
29. R. Krumm, B. Guel, C. Schmitz, and G. Staikov, Electrochim. Acta, 45, 3255
͑2000͒.
Significant differences were observed between the electrochemi-
cal and electrical properties of Schottky barriers formed on H- and
CH3-terminated Si͑111͒ surfaces. Cyclic voltammetric data on the
Pb or Cd electrodeposition processes on such surfaces were consis-
tent with previous observations that the band-edge energetics in
aqueous acidic solution of H-terminated n-Si͑111͒ are offset from
those of CH3-terminated n-Si͑111͒ by several tenths of a volt. The
electrochemical observations are also in accord with prior UV pho-
toelectron spectroscopic measurements that have elucidated the dif-
ferences in interfacial dipole magnitude between the H-Si͑111͒ and
CH3–Si͑111͒ surfaces. The dark J-V responses of electrodeposited
Cd and Pb films clearly exhibited more rectifying behavior on
n-type CH3–Si͑111͒ surfaces than on n-type H-Si͑111͒ surfaces.
Chronoamperometric data on n+-Si͑111͒ substrates has quantita-
tively indicated that nucleation events during electrodeposition are
an order of magnitude lower on CH3–Si͑111͒ surfaces than on
H-Si͑111͒ surfaces. The electrochemical and electrical properties of
the electrodeposited Si͑111͒ Schottky barriers reported herein
complement previous studies on CH3-terminated Si͑111͒ surfaces
and further emphasize the durable, well-defined, and unique electro-
chemical, electrical, chemical, and electronic properties of these in-
terfaces.
30. P. Bindra, H. Gerischer, and D. M. Kolb, J. Electrochem. Soc., 124, 1012 ͑1977͒.
31. E. Aharonshalom and A. Heller, J. Electrochem. Soc., 129, 2865 ͑1982͒.
32. A. Heller and R. G. Vadimsky, Phys. Rev. Lett., 46, 1153 ͑1981͒.
33. S. Yae, M. Kitagaki, T. Hagihara, Y. Miyoshi, H. Matsuda, B. A. Parkinson, and Y.
Nakato, Electrochim. Acta, 47, 345 ͑2001͒.
34. S. Takabayashi, M. Ohashi, K. Mashima, Y. Liu, S. Yamazaki, and Y. Nakato,
Langmuir, 21, 8832 ͑2005͒.
35. S. U. M. Khan and J. O. Bockris, J. Phys. Chem., 88, 2504 ͑1984͒.
36. T. Zambelli, M. L. Munford, F. Pillier, M. C. Bernard, and P. Allongue, J. Elec-
trochem. Soc., 148, C614 ͑2001͒.
37. R. M. Stiger, S. Gorer, B. Craft, and R. M. Penner, Langmuir, 15, 790 ͑1999͒.
38. P. Jakob, Y. J. Chabal, K. Kuhnke, and S. B. Christman, Surf. Sci., 302, 49 ͑1994͒.
39. P. Dumas, Y. J. Chabal, R. Gunther, A. T. Ibrahimi, and Y. Petroff, Prog. Surf. Sci.,
48, 313 ͑1995͒.
40. C. P. Wade and C. E. D. Chidsey, Appl. Phys. Lett., 71, 1679 ͑1997͒.
41. H. Fukidome and M. Matsumura, Appl. Surf. Sci., 130–132, 146 ͑1998͒.
42. P. Allongue, V. Kielingb, and H. Gerischer, Electrochim. Acta, 40, 1353 ͑1995͒.
43. A. Bansal, X. L. Li, I. Lauermann, N. S. Lewis, S. I. Yi, and W. H. Weinberg, J.
Am. Chem. Soc., 118, 7225 ͑1996͒.
44. G. Oskam, D. van Heerden, and P. C. Searson, Appl. Phys. Lett., 73, 3241 ͑1998͒.
45. G. Holmbom and B. E. Jacobson, J. Electrochem. Soc., 135, 2720 ͑1988͒.
46. K. M. S. Youssef, C. C. Koch, and P. S. Fedkiw, J. Electrochem. Soc., 151, C103
͑2004͒.
47. M. S. Chandrasekar and M. Pushpavanam, Electrochim. Acta, 53, 3313 ͑2008͒.
48. G. Oskam, J. G. Long, A. Natarajan, and P. C. Searson, J. Phys. D, 31, 1927
͑1998͒.
49. J. A. Haber and N. S. Lewis, J. Phys. Chem. B, 106, 3639 ͑2002͒.
50. G. Gunawardena, G. Hills, and I. Montenegro, J. Electroanal. Chem. Interfacial
Electrochem., 184, 371 ͑1985͒.
Acknowledgments
We thank the National Science Foundation ͑Grant CHE-
0604894͒ for financial support and the Gordon and Betty Moore
Foundation for postdoctoral fellowship support ͑S.M.͒.
51. S. M. Sze, Physics of Semiconductor Devices, John Wiley & Sons, New York
͑1981͒.
52. M. X. Tan and N. S. Lewis, Inorg. Chim. Acta, 242, 311 ͑1996͒.
53. M. A. Schneeweiss, H. Hagenstrom, M. J. Esplandiu, and D. M. Kolb, Appl. Phys.
A, 69, 537 ͑1999͒.
54. G. J. Hills, D. J. Schiffrin, and J. Thompson, Electrochim. Acta, 19, 671 ͑1974͒.
55. G. Oskam and P. C. Searson, Surf. Sci., 446, 103 ͑2000͒.
56. W. Monch, Phys. Status Solidi A, 159, 25 ͑1997͒.
57. S. Maldonado, D. Knapp, and N. S. Lewis, J. Am. Chem. Soc., 130, 3300 ͑2008͒.
58. X. G. Zhang, Electrochemistry of Silicon and Its Oxide, p. 45, Kluwer Academic/
Plenum Publishers, New York ͑2001͒.
California Institute of Technology assisted in meeting the publication
costs of this article.
References
1. P. W. Loscutoff and S. F. Bent, Annu. Rev. Phys. Chem., 57, 467 ͑2006͒.
2. A. G. Aberle, Prog. Photovoltaics, 8, 473 ͑2000͒.
3. Y. J. Chabal and K. Raghavachari, Surf. Sci., 502, 41 ͑2002͒.
4. R. Cohen, L. Kronik, A. Shanzer, D. Cahen, A. Liu, Y. Rosenwaks, J. K. Lorenz,
and A. B. Ellis, J. Am. Chem. Soc., 121, 10545 ͑1999͒.
5. J. M. Buriak, Chem. Rev. (Washington, D.C.), 102, 1271 ͑2002͒.
6. J. S. Hovis, S. Lee, H. B. Liu, and R. J. Hamers, J. Vac. Sci. Technol. B, 15, 1153
͑1997͒.
7. R. Boukherroub, S. Morin, F. Bensebaa, and D. D. M. Wayner, Langmuir, 15,
3831 ͑1999͒.
8. T. K. Mischki, R. L. Donkers, B. J. Eves, G. P. Lopinski, and D. D. M. Wayner,
Langmuir, 22, 8359 ͑2006͒.
59. Unpublished results.
60. P. G. Santangelo, G. M. Miskelly, and N. S. Lewis, J. Phys. Chem., 93, 6128
͑1989͒.
61. P. G. Santangelo, G. M. Miskelly, and N. S. Lewis, J. Phys. Chem., 92, 6359
͑1988͒.
62. P. G. Santangelo, M. Lieberman, and N. S. Lewis, J. Phys. Chem. B, 102, 4731
͑1998͒.
63. S. R. Morrison, Electrochemistry at Semiconductor and Oxidized Metal Electrodes,
Springer, New York ͑1980͒.
64. P. Allongue, E. Souteyrand, and L. Allemand, J. Electroanal. Chem., 362, 89
͑1993͒.
9. A. Bansal, X. L. Li, S. I. Yi, W. H. Weinberg, and N. S. Lewis, J. Phys. Chem. B,
105, 10266 ͑2001͒.
65. H. B. Yu, L. J. Webb, R. S. Ries, S. D. Solares, W. A. Goddard, J. R. Heath, and N.
S. Lewis, J. Phys. Chem. B, 109, 671 ͑2005͒.
66. F. Gloaguenb, J. M. Légerb, C. Lamyb, A. Marmanna, U. Stimming, and R. Vogel,
10. W. J. Royea, A. Juang, and N. S. Lewis, Appl. Phys. Lett., 77, 1988 ͑2000͒.
11. T. W. Hamann and N. S. Lewis, J. Phys. Chem. B, 110, 22291 ͑2006͒.
12. R. Hunger, R. Fritsche, B. Jaeckel, W. Jaegermann, L. J. Webb, and N. S. Lewis,
Phys. Rev. B, 72, 045317 ͑2005͒.
13. S. Maldonado, K. E. Plass, D. Knapp, and N. S. Lewis, J. Phys. Chem. C, 111,
17690 ͑2007͒.
14. E. H. Rhoderick and R. H. Williams, Metal-Semiconductor Contacts, Clarendon
Press, Oxford ͑1988͒.
15. A. Salomon, T. Boecking, O. Seitz, T. Markus, F. Amy, C. Chan, W. Zhao, D.
Cahen, and A. Kahn, Adv. Mater. (Weinheim, Ger.), 19, 445 ͑2007͒.
Electrochim. Acta, 44, 1805 ͑1999͒.
67. R. M. Penner, J. Phys. Chem. B, 105, 8672 ͑2001͒.
68. H. Liu and R. M. Penner, J. Phys. Chem. B, 104, 9131 ͑2000͒.
69. P. Allongue and F. Maroun, J. Phys. Condens. Matter, 18, S97 ͑2006͒.
70. R. C. Rossi, M. X. Tan, and N. S. Lewis, Appl. Phys. Lett., 77, 2698 ͑2000͒.
71. R. C. Rossi and N. S. Lewis, J. Phys. Chem. B, 105, 12303 ͑2001͒.
72. K. E. Plass, X. Liu, B. S. Brunschwig, and N. S. Lewis, Chem. Mater., 20, 2228
͑2008͒.