1
Supplementary
Information
Investigations
of the Stability of
Etched or Platinized p-InP(100) Photocathodes for Solar-
driven Hydrogen Evolution
in Acidic or Alkaline
Aqueous
Electrolytes
Weilai Yu
1
, Matthias
H. Richter
1,2
, Pakpoom Buabthong
2
, Ivan A.
Moreno-Hernandez
1
, Carlos
G.
Read
1
, Ethan Simonoff
1
, Bruce S. Brunschwig
3
and Nathan S.
Lewis
1,3*
1
Division of
Chemistry and Chemical Engineering,
2
Division of Engineering and Applied Science,
and
3
Beckman Institute Molecular Materials Resource
Center
(MMRC), California Institute of
Technology,
Pasadena CA, 91125
Email:
nslewis@caltech.edu
Electronic
Supplementary
Material
(ESI)
for
Energy
&
Environmental
Science.
This
journal
is
©
The
Royal
Society
of
Chemistry
2021
2
Experimental
Materials
All
chemicals
and
materials
were
used
as
received,
including
sulfuric
acid
(H
2
SO
4
,
Fisher
Scientific,
TraceMetal
Grade,
93-98%),
1.0
M sulfuric
acid
solutions
(H
2
SO
4
,
VWR
Chemicals),
potassium
hydroxide
(KOH,
Sigma-Aldrich,
semiconductor
grade,
99.99
%
trace-metal
basis),
bromine
(Br
2
,
reagent
grade,
Sigma-Aldrich),
ammonia/methanol
solution
(4.0
M,
Sigma-Aldrich),
and
methanol
(CH
3
OH,
VWR
Analytical,
ACS,
99.8
%).
Deionized
water
with
a resistivity
of
18.2
MΩ
cm
was
obtained
from
a Barnstead
Millipore
system.
Single-side
polished,
(100)-oriented,
Sn-doped
(N
d
= 1-
5×10
17
cm
-3
),
n-type
InP
wafers
and
Zn-doped
(N
d
= 1-5×10
17
cm
-3
),
p-type
InP
wafers
were
obtained
from
AXT
Inc.
Indium
foil
(0.25
mm
thick,
99.99
%)
was
purchased
from
VWR
International.
Nafion
(proton-exchange
membrane)
and
Fumasep
(anion-exchange
membrane)
were
purchased from the Fuel Cell Store.
Purification
of 1.0 M
H
2
SO
4
(aq) and
1.0 M KOH(aq) by pre-electrolysis
Prior
to
use,
the
prepared
1.0
M
H
2
SO
4
(aq)
and
1.0
M
KOH(aq)
electrolytes
were
pre-electrolyzed
in
a
two-compartment
electrochemical
cell,
with
the
two
compartments
separated
by
either
a Nafion
(acid)
or
Fumasep
(base)
membrane.
The
pre-electrolysis
was
performed
for
> 24
h under
either
a
constant
> 3 V
bias
or
under
a constant
current
of
6 mA,
using
two
separate
carbon
cloth
or
carbon
rod
electrodes.
In
some
pre-electrolysis
experiments
of
1.0
M
KOH(aq),
a Ni
foil
electrode
was
also
used
as the
anode
for
pre-electrolysis
of
1.0
M
KOH(aq).
Only
the
pre-electrolyzed
electrolyte
in the
cathode
compartment
was
used
in
subsequent
electrochemical
measurements.
After
the
electrolysis,
H
2
O
2
was
not
detectable
in
the
catholyte,
as
determined
by
spectrophotometric
analysis
using
titanium oxalate.
1
Back contacts to InP electrodes
Ohmic
back
contacts
to
n-type
InP
were
obtained
by
sputtering
20
nm
of
Ni
onto
the
back
(unpolished)
side
of
the
n-InP
wafer
and
then
annealing
the
sample
under
forming
gas
at
400
C for
10
min.
Sputter
deposition
of
Ni
was
performed
in
an
AJA
Orion
sputtering
system.
Ohmic
back
contacts
to
p-type
InP
were
obtained
by
sputtering
10
nm
of
Zn
and
then
90
nm
of
Au
onto
the
back
side of the wafer and then
annealing the sample
under forming gas at 400
C for 10 min.
2
3
Epoxy-encapsulated electrodes
For
electrochemical
measurements
made
outside
the
glovebox,
the
back
contact
to the
InP
sample
was
attached
to
a coiled,
tin-plated
Cu
wire
(McMaster-Carr)
using
high-purity
Ag
paint
(SPI
supplies).
The
Cu
wire
was
then
threaded
through
a piece
of
glass
tubing
(Corning
Incorporated,
Pyrex
tubing,
7740
glass).
The
sample
was
encapsulated
in,
and
sealed
to,
the
glass
tube
using
Hysol
9460
epoxy,
which
was
allowed
to
dry
overnight
under
ambient
conditions.
The
exposed
areas
of
each
electrode
were
imaged
using
a high-resolution
optical
scanner
(Epson
Perfection
V370
with
a
resolution
of
1200
psi)
and
the
geometric
areas
of
the
electrodes
(typically
0.03-0.10
cm
2
)
were
analyzed using ImageJ software.
InP etching
Prior to electrochemical measurements, n-InP and p-InP
electrodes
were etched in
0.04%
(by volume)
Br
2
/CH
3
OH
for
30
s, then
in
4.0
M
NH
3
/CH
3
OH
for
30
s, and
then
rinsed
in
pure
CH
3
OH
for
10
s.
3
The
etching
and
rinsing
cycle
was
then
repeated
two
more
times.
The
electrodes
were
blown
dry
for
>10 s under
a stream of flowing
N
2
(g). Etching was performed outside the glove box.
Electrochemical measurements
in
an O
2
-free
environment
Electrochemical
measurements
on
samples
that
were
analyzed
by
XPS
were
performed
in a nitrogen-
filled
glovebox
(VAC,
OMNI-LAB)
using
electrolytes
that
had
been
degassed
in
a Schlenk
line
(Figure
S2a).
The
concentration
of
O
2
in
the
glove
box
was
< 1.0
ppm
as monitored
by
an
O
2
-
sensitive electrode (Fuel Cell Sensor, AMI Acetic, Type T2).
To
facilitate
XPS
analysis
of
the
samples,
electrochemical
measurements
were
performed
using
a
custom
compression
cell
fabricated
from
PEEK
(Figure
S1).
4
The
cell
had
two
compartments
separated
by
an ion-exchange
membrane
(Nafion
for
measurements
in
H
2
SO
4
,
and
Fumasep
for
measurements
in
KOH).
Electrochemical
measurements
were
performed
using
a SP-200
potentiostat
(BioLogic
Science
Instruments)
and
a three-electrode
set-up
with
either
a Pt
foil
(for
H
2
SO
4
)
or
a Ni
foil
(for
KOH)
as
the
counter
electrode,
and
either
a leakless
miniature
AgCl/Ag
electrode
(eDAQ,
ET072-1)
or a hydrogen
electrode
HydroFlex
(Gaskatel)
as
the
reference
electrode.
Electrochemical
data
were
acquired
without
compensation
for
the
series
resistance
of
the
solution.
During
electrochemical
experiments,
H
2
(g)
was
fed
into
the
glovebox,
passed
through
a gas
bubbler,
and
fed
4
into
the
catholyte
to
constantly
purge
the
solution
(Figure
S2b).
Separate
outlets
were
provided
for
H
2
from
the
cathode
chamber
and
for
O
2
from
the
anode
chamber.
Prior
to each
experiment,
the
electrochemical
cell
was
assembled
immediately
after
InP
etching,
and
the
assembled
cell
was
then
promptly transferred into
the glovebox.
Electrochemical
data
were
acquired
on
an
SP-200
potentiostat
(BioLogic
Science
Instruments)
without
compensation
for
the
series
resistance
of
the
solution.
To
periodically
measure
the
J
-
E
behavior
of
illuminated
p-InP/Pt
electrodes,
chronoamperometry
(CA)
was
first
interrupted
by
a
short
period
(15
s) of
open
circuit
to
measure
E
oc
.
In
each
cycle,
cyclic
voltammetry
(CV)
was
started
from
E
start
=
E
oc
-30
mV
to
< 0 V
vs
RHE
and
then
ended
at
E
end
=
E
oc
-80
mV,
to
minimize
passing
anodic current
through the electrodes. Three CVs were measured during
each cycle.
The
main
goal
of this
study
is
to
elucidate
the
underlying
corrosion
pathways
of
p-InP
photocathodes,
by
correlating
the
electrode
dissolution
and
surface
changes
with
the
evolution
of
the
J-E
behaviors
as a function
of
time.
The
total
periods
of
the
CA
testing
were
determined
based
on
when
the
J-E
behavior
of
the
electrodes
stopped
changing
over
time.
Thus,
the
elapsed
times
used
for
the stability experiments do not reflect the
ultimate limits in stability
for p-InP/Pt electrodes.
In
this
work,
evaluation
of
the
stability
of
etched
p-InP
and
p-InP/Pt
photocathodes
was
primarily
performed
at E=0
V
vs.
RHE,
to
allow
comparison
with
analogous
studies
in
other
systems.
5–7
Due
to
the
presence
of
a photovoltage,
applying
this
E
produces
substantial
cathodic
current
density
at start
of
the
CA,
as
expected
based
on
the
position
of
the
maximum
power
point
in
an
efficient
PEC
system.
The
cathodic
current
density
also
determines
the
surface
quasi-Fermi
level
position
of minority
carriers,
as
well
as
the
corrosion
pathways
of
InP
in
the
Pourbaix
diagram.
Thus,
our
further
evaluation
of
surface
chemistry
was
performed
at
this
specific
E. Similarly
for
n-InP
electrodes,
E=-0.1
V
vs.
RHE
was
chosen
to
match
the
quasi-Fermi
level
position
of
majority
carriers with the
overpotential of Pt HER catalyst.
A
miniature
fiber-optic
adjustable-arm
light
equipped
with
a 150
W halogen
bulb
was
used
as
the
illumination
source.
Illumination
was
introduced
from
the
source
into
the
glove
box
via
the
fiber
optic.
The
illumination
intensity
at the
location
of
the
sample
in
the
cell
was
calibrated
to 1 Sun
(100
mW
cm
-2
)
using
a Si
photodiode
(FDS100,
Thorlabs).
The
volume
of
electrolyte
in the
cathode
5
chamber
was
3-4
mL.
For
each
ICP-MS
analysis,
0.2
mL
of
electrolyte
was
removed
from
the
cell
at
different
time
intervals
and
0.2
mL
of fresh
electrolyte
was
added
into
the
cell.
After
each
experiment
was
completed,
the
cell
was
disassembled
inside
the
glove
box,
and
the
electrode
sample
was
thoroughly
cleaned
with
deionized
water,
dried
under
stream
of
flowing
N
2
(g),
and
stored
inside
the glove box until the
surface was analyzed
by XPS.
Figure
S32
compares
the
corrosion
thickness
of
n-InP
electrodes
(dark)
held
at
E=-0.1
V vs.
RHE
in
H
2
-saturated
and
O
2
-saturated
1.0
KOH
electrolytes.
However,
the
underlying
corrosion
mechanism
of
InP
induced
by
oxygen
exposure
is the
topic
of a separate
study
but
is not
relevant
to
an
operating
photocathode
in
which
the
catholyte
should
be maintained
under
1 atm
of
H
2
to obtain
intrinsically
safe
operation
as
well
as to
minimize
efficiency
losses
due
to
reduction
of
O
2
in
competition with the
HER.
Electrodeposition of
Pt
particles
A solution
of 5 mM
K
4
PtCl
6
and
0.5
M
KCl
was
used
to electrodeposit
Pt
particles
onto
InP
samples.
A
constant
current
density
of
-0.2
mA
cm
-2
was
applied
using
a two-electrode
configuration
until
20
mC
cm
-2
had
passed.
Before
each
deposition,
bubbles
accumulating
in the
electrode
area
were
carefully
removed
to
ensure
the
uniformity
of
electrodeposition.
The
deposition
was
performed
under
~1
Sun
illumination
for
p-InP
and
in
the
dark
for
n-InP.
Either
a Pt
wire
or
a carbon
rod
was
used
as
the
counter
electrode
for
the
deposition.
The
cell
was
then
thoroughly
cleaned
with
deionized
water
at
least
3 times
before
the
1.0
M
H
2
SO
4
or
1.0
M KOH
electrolyte
used
for
electrochemical
experiments
was
added
to
the
cell.
For
the
stability
test
of
p-InP/Pt
in 1.0
M KOH(aq),
the
Pt
deposition
and
the
subsequent
hydrogen-evolution
reaction
(HER)
process
were
performed
in
two
separate cells to minimize cross-contamination.
Electrochemistry outside
glovebox
A series
of
chronoamperometry
(CA)
and
cyclic
voltammetry
(CV)
experiments
for
n-InP
(dark)
and
p-InP
(light)
were
performed
in
a two-compartment
cell
outside
the
glove
box
under
a continuous
H
2
purge
(Figure
S2).
These
experiments
did
not
involve
XPS
analyses
after
electrochemical
operation.
The
CA
experiments
were
performed
by
potentiostatically
holding
freshly
etched
InP
electrodes
at
6
specified
potentials
for
30
min.
At
least
three
voltammetric
cycles
were
then
scanned
in
the
positive
direction
from
the
original
polarization
potential
to
potentials
sufficient
to oxidize
any
deposited
In
metal.
Electrochemical
measurements
were
performed
using
a
custom
two-compartment
electrochemical
cell
with
a Nafion
(1.0
M
H
2
SO
4
)
or
Fumasep
(1.0
M
KOH)
membrane
separating
the
two
compartments.
The
electrochemical
cell
was
cleaned
with
aqua
regia
before
use.
The
volume
of
the
electrolyte
used
in
the
cathode
compartment
was
20
or
25
mL.
A
mercury/mercurous
sulfate
(Hg/HgSO
4
in saturated
K
2
SO
4
(aq),
CH
Instruments,
CH151)
reference
electrode
was
used
for
measurements
in
1.0
M H
2
SO
4
(aq).
A
mercury/mercury
oxide
(Hg/HgO
in
1.0
M KOH
(aq),
CH
Instruments,
CH152)
reference
electrode
was
used
for
measurements
in 1.0
M
KOH
(aq).
A
carbon
rod
placed
within
a fritted
glass
tube
(gas
dispersion
tube
Pro-D,
Aceglass,
Inc.)
was
used
as
the
counter
electrode.
Both
the
Hg/HgSO
4
and
the
Hg/HgO
reference
electrodes
were
calibrated
versus
a
reversible
hydrogen
electrode
(RHE).
The
Hg/HgO
electrode
potential
was
0.910
V
versus
RHE
in
1.0 M KOH (aq). The Hg/HgSO
4
electrode potential
was 0.683 V
versus RHE in
1.0 M H
2
SO
4
(aq).
Electrochemical
data
were
acquired
on
an
MPG-2
multichannel
potentiostat
(BioLogic
Science
Instruments)
without
compensation
for
the
series
resistance
of
the
solution.
During
measurements,
the
electrolyte
was
continually
bubbled
with
1 atm
of H
2
(g)
and
vigorously
agitated
with
a magnetic
stir
bar
driven
by
a model-train
motor
(Pittman)
with
a Railpower
1370
speed
controller
(Model
Rectifier
Corporation)
or
a magnetic
stirrer
(IKA
Topolino).
50
W
ENH-type
(Philips
MR16)
W-
halogen
lamps
with
dichroic
rear
reflectors
and
custom
housings
and
transformers
(Staco
Energy
Products
Co.)
were
used
for
photoelectrochemical
measurements.
The
illumination
intensity
at
the
position
of the
working
electrode
in
the
electrochemical
cell
was
determined
by
placing
a calibrated
Si
photodiode
(FDS100-Cal,
Thor
Labs)
into
the
cell
at
the
same
position
occupied
by
the
photoelectrode.
To
illuminate
the
bottom-facing
photoelectrodes,
a broadband
reflection
mirror
(Newport dielectric mirror)
was used to direct the uniform light beam in
the vertical direction.
Mott-Schottky Analyses
Impedance
measurements
of freshly
etched
n-InP
and
p-InP
electrodes,
in contact
with
1.0
M
H
2
SO
4
(aq)
or
1.0
M
KOH(aq)
under
H
2
(g)
bubbling,
were
collected
in the
dark
over
a frequency
7
range
of
20
Hz
to
20
kHz
with
a sinusoidal
wave
amplitude
of
25
mV.
The
measurements
were
performed
outside
the
glove
box.
The
impedance
measurements
were
fit
with
a circuit
consisting
of
a
resistor
in
series
with
an
additional
component
consisting
of
a resistor
and
a capacitor
in
parallel.
8,9
The
potential
dependence
of
the
differential
capacitance
was
analyzed
using
the
Mott-Schottky
relationship:
퐶
푑
‒
2
=
2
푞
퐴
2
휀
0
휀
푟
푁
푑
(
푉
푎푝푝
+
푉
푏푖
‒
푘
퐵
푇
푞
)
where
C
d
is the
differential
capacitance,
q
is
the
unsigned
charge
of
an
electron,
A
is
the
electrode
area,
ε
0
is
the
vacuum
permittivity,
ε
r
is
the
relative
permittivity
of
the
semiconductor,
N
d
is
the
dopant
concentration
of
the
semiconductor,
V
app
is the
applied
electrode
potential,
V
bi
is the
built-in
voltage
in
the
semiconductor,
and
T
is the
temperature
of
the
electrode
while
the
impedance
data
were collected. A relative
permittivity of
ε
r
= 8.8 was used to analyze data for InP electrodes.
Analytical Methods
Inductively coupled plasma mass spectrometry (ICP-MS)
Inductively
coupled
plasma
mass
spectrometry
(ICP-MS)
data
were
collected
using
an Agilent
8800
Triple
Quadrupole
ICP-MS
system.
Calibration
solutions
were
prepared
by
diluting
the
multi-
element
standard
solutions
for
ICP
with
water
having
a resistivity
of
18.2
MΩ
cm.
All
electrolyte
samples
were
acidified
to
pH
≤2
before
the
ICP-MS
measurements.
The
total
amounts
of
In
dissolved
from
the
electrodes
were
calculated
and
normalized
to
the
geometric
electrode
area
to yield
the
equivalent
depth
of
material
removed
from
the
InP
electrode.
The
error
bars
of
each
data
point
represent
the
standard
deviations
of five
consecutive
measurements
using
the
instrument.
Due
to the
high
detection
limit
of phosphorus
(P)
by
ICP-MS,
only
the
concentrations
of
In
ions
were
used
to
calculate
the
corrosion
thickness
of InP.
The
conversion
equation
relating
the
In
concentration
(ug/L)
as measured by ICP-MS to the dissolution thickness of InP
(nm)
is shown below:
8
퐼푛푃 푇ℎ푖푐푘푛푒푠푠
(
푛푚
)
=
푥 휇푔
1 퐿
(
퐼푛 푐표푛푐푒푛푡푟푎푡푖표푛
)
∗
3.2 푚퐿
0.2 푚퐿
(
푑푖푙푢푡푖표푛 푓푎푐푡표푟
)
∗
4 푚퐿
(
푡표푡푎푙 푣표푙푢푚푒푟 표푓 푐푎푡ℎ표푙푦푡푒
)
∗
1 퐿
1000 푚퐿
∗
1
∗
10
‒
6
푔
1 휇푔
∗
1 푚표푙
114.8 푔
(
퐼푛 푚표푙푎푟 푚푎푠푠
)
∗
(
114.8
+
31
)
푔
1 푚표푙
(
퐼푛푃 푚표푙푎푟 푚푎푠푠
)
∗
1 푐푚
3
4.81 푔
(
퐼푛푃 푑푒푛푠푖푡푦
)
∗
1
0.2
푐푚
2
(
푒푙푒푐푡푟표푑푒 푎푟푒푎
)
∗
1
∗
10
7
푛푚
1 푐푚
The
dilution
factor,
the
total
volume
of
catholyte
and
the
electrode
area
shown
above
are
typical
values which
are
specific to each experiment.
X-ray photoelectron spectroscopy
X-ray
photoelectron
spectroscopy
(XPS)
was
performed
using
a Kratos
Axis
Ultra
system
with
a
base
pressure
of
< 1×10
-9
Torr
equipped
with
a monochromatic
Al
Kα
X-ray
source
with
a photon
energy
of 1486.6
eV.
Photoelectrons
were
collected
at
0°
from
the
surface
normal
with
a retarding
pass
energy
of
160
eV
for
survey
XPS
scans
(step
sizes
of
1.0
eV
and
10
eV)
and
for
high-resolution
core-level scans (step
size
0.025 eV).
Prior
to
XPS
measurements,
samples
were
mounted
on
a sample
holder
and
loaded
in
a transfer
box
inside
the
glove
box
under
nitrogen.
The
transfer
box
was
attached
to
the
load
lock
of
the
Kratos
Axis
Ultra
system
with
the
transfer
box
gate
valve
(VAT)
still
closed
(Figure
S2c).
The
load
lock
was
first
pumped
down
and
purged
again
with
N
2
.
After
pumping
down
the
load
lock
again
to 1×10
-
2
Torr,
the
gate
valve
to
the
transfer
box
was
opened
and
the
turbo
molecular
pump
was
switched
on.
After
achieving
a pressure
of
<1
x 10
-6
Torr,
the
sample
was
transferred
to the
sample-transfer
chamber (base pressure <1×10
-9
Torr)
before
further
transfer to
the analysis chamber.
All
XPS
peak
fitting
was
performed
using
CasaXPS
software
version
2.3.18.
All
binding
energies
are
referenced
to
the
adventitious
carbon
peak
at 284.8
eV.
Before
fitting
the
data,
a Shirley
background
was
calculated
and
subtracted
from
the
original
spectra.
The
major
peak
of
the
In
3+
cations
of InP
in
the
In
3d
spectra
was
fit using
an asymmetric
Lorentzian
function
LF(1,1,15,80).
All
other
peaks
were
fit
using
the
70%
Gaussian/30%
Lorentzian
Voigt-function.
The
surface
In/P
atomic
ratios
were
calculated
using
the
relative
sensitive
factors
(RSF)
in
the
database
of
the
Kratos
instrument
and
the
peak
areas
(In=7.265,
P=0.486).
In addition,
the
criteria
of
XPS
peak
assignments
9
to
various
In-containing
species
including
InP,
InO
x
,
In(OH)
x
and
InPO
x
is shown
in Table
S4.
10–22
We
also
collected
XPS
data
of
various
control
samples
(InP
with
native
oxide,
ITO
and
In
foils)
on
the same instrument to further
support the
XPS
peak assignment,
which are shown in
Figure
S33-35.
In this work,
we used single-crystal p-InP(100)
electrodes with thicknesses of 625
μm.
However,
electrochemical
experiments
using
either
etched
p-InP
or
p-InP/Pt
electrodes
only
revealed
surface
changes
over
the
range
of a few
nm.
Due
to
its
relatively
large
penetration
depth,
XRD
analyses
of
both
etched
p-InP
and
p-InP/Pt
electrodes
before
and
after
CA
will
likely
yield
similar
patterns
with
major
features
assigned
to
the
lattice
of
crystalline
InP.
In
contrast,
the
high
surface
sensitivity
and
relatively
low
photoelectron
escape
depth
of XPS
provides
a more
direct
probe
of
changes
in
the
surface
and
consequently
facilitates
determination
of the
corrosion
pathways.
In
addition,
the
dipole
energy
formed
at the
InP/electrolyte
interface
shown
in Scheme
1 is
not
measurable
by
ultraviolet
photoelectron spectroscopy (UPS) under ultrahigh-vacuum conditions.
Microscopy
Scanning-electron
microscopy
(SEM)
images
were
obtained
using
a calibrated
Nova
NanoSEM
450
(FEI) with an accelerating voltage of 5 kV.
Transmission-electron
microscopy
(TEM)
cross-sections
of
the
samples
were
prepared
using
a
focused
Ga
ion
beam
(FIB),
on
a FEI
Nova-600
Nanolab
FIB/FESEM
or
a FEI
Helios
NanoLab
G4
Dual
Beam.
Pt
and
C protection
layers
were
applied
prior
to exposure
to
the
FIB.
TEM
images
of the
prepared
lamella
samples
were
obtained
using
a Tecnai
Polara
(F30)
TEM
at
an
accelerating
voltage
of
300
kV,
or
a FEI
Osiris
at an
accelerating
voltage
of
200
kV
equipped
with
a Gatan
2K
TEM
camera and Bruker EDS.
Atomic-force
microscopy
(AFM)
images
were
obtained
on
a Bruker
Dimension
Icon
using
Bruker
ScanAsyst-Air
probes
(silicon
tip,
silicon
nitride
cantilever,
spring
constant:
0.4
N
m
-1
,
frequency:
50-90
kHz),
operating
in
the
ScanAsyst
mode.
Images
were
analyzed
using
the
Nanoscope Analysis
software (version
1.9).
10
Figure
S1.
(a,b)
Two
views
of
the
design
of the
photoelectrochemical
compression
cell
used
for
electrochemical
measurements
in
an O
2
-free
environment.
The
current
collector
(Al
foil)
at
the
back
of
both
working
and
counter
electrode
are
omitted
for
simplicity.
Before
each
experiment,
electrolyte
was
added
through
the
ports
of H
2
and
O
2
outlets
before
the
respective
tubes
were
connected.
During
each
experiment,
the
electrolyte
in
the
catholyte
compartment
was
sampled
(0.2
mL)
through
the
H
2
-
outlet port
after disconnecting the tube,
while the working electrode was still under potential control.
11
Figure
S2.
Experimental
setup.
(a)
The
custom
cell
illustrated
in
Figure
S1
being
tested
inside
a
N
2
-filled
glovebox.
(b)
Modifications
to
the
glove
box
to allow
experiments
to
be
performed
with
H
2
purging,
light
illumination
and
post-test
N
2
purging.
(c)
The
air-free
transfer
arm
attached
to
the
XPS
instrument.
After
the
EC
cell
was
disassembled,
the
sample
was
rinsed
with
deionized
water,
purged
with
N
2
,
and
loaded
onto
the
transfer
arm
inside
the
glove
box.
The
transfer
arm
was
transferred
from
the
glove
box
to
the
XPS.
(d)
A two-compartment
cell
used
for
testing outside the
glove
box while purging with H
2
.