1
Supplementary Materials
F
acet
-
D
ependent Oxygen Evolution Reaction Activity of IrO
2
from Quantum Mechanics
and Experiments
Soonho Kwon,
1
† Kelsey A. Stoerzinger,
2
† Reshma Rao,
3
Liang Qiao,
4
William A. Goddard
III,
1
* Yang Shao
-
Horn
5,6,7
*
1
Materials and Process Simulation Center (MSC), California Insti
tute of Technology,
Pasadena, California
91125
, United States
2
Department of Chemical Engineering and Materials Science, University of Minnesota,
Minneapolis, M
innesota
55455,
United States
3
Department of Materials, Imperial College London, South Kensington Campus, London
SW7
2AZ, United Kingdom
4
Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, 1 Bethel Valley
Road, Oak Ridge, Tennessee 37831, United States
5
Department of Materials Science and Engineering, Massachusetts Institute of Technology,
Cambridge,
M
essachusetts
02139,
United States
6
Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge,
M
essachusetts
02139,
United States
7
Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge,
M
essachusetts
02139,
United States
†These authors contributed equally to this work.
*Corresponding authors:
wag@caltech.
edu
;
shaohorn@mit.edu
ORCID: SK: 0000
-
0002
-
9225
-
3018
ORCID: KAS: 0000
-
0002
-
3431
-
8290
ORCID: RR: 0000
-
0002
-
6655
-
3105
ORCID: LQ: 0000
-
0003
-
2400
-
2986
ORCID: WAG:0000
-
0003
-
0097
-
5716
ORCID: YSH: 0000
-
0001
-
8714
-
2121
2
1.
Materials and Methods
1.1
Film Fabrication and Characterization
Epitaxial films were fabricated by pulsed laser deposition (PLD) on single crystal (001)
-
oriented LaAlO
3
or MgO substrates or (101), (111) or (001) TiO
2
from an Ir metal target. The
MgO substrate was first covered by a 2 nm layer of BaTiO
3
(BTO) by PLD to prevent formation
of IrMg intermetallics. PLD was performed using a KrF excimer laser (λ = 248 nm) at ~750
°C to obtain films at least 25 nm thick, which were deposited at an oxygen pressure of 200
mTorr and cooled to room temperature unde
r 200 Torr oxygen
. Thin film X
-
ray diffraction
(XRD) was performed using a four
-
circle diffractometer (X
’
Pert PRO, PANalytical
). Film
surface morphologies examined by atomic force microscopy (AFM; see
Fig.
S1 in the
Supporting Information) (Bruker Dimension Icon). Current densities are normalized to the
geometric area of the electrode, yielding similar charge (q*) upon cycling in Argon (Table
S2
)
after OER measurements. This was previously reported to scale with active area.
(
1
,
2
)
1.2
Electrochemical Measurements
Electrical contacts were applied to the front of the conductive IrO
2
film. A Gallium
−
indium
eutectic (Sigma
-
Aldrich, 99.99%) was scratched into a small corner, and a Ti
wire (Sigma
-
Aldrich, 99.99%) was affixed with silver paint (Ted Pella, Leitsilber 200). The back and sides
of the electrode as well as the wire were covered with a nonconductive, chemically resistant
epoxy (Omegabond 101), so only the catalyst surface was
exposed to the electrolyte.
Electrochemical measurements were conducted with a Biologic SP
-
300 potentiostat in a
∼
120 mL solution of 0.1 M HClO
4
(70% Veritas® double distilled) prepared from deionized
water (Millipore, >18.2 M
Ω
cm) and saturated with O
2
to fix the reversible potential. Potentials
were referenced to a saturated Ag/AgCl electrode (Pine), calibrated to the RHE scale in 0.1 M
HClO
4
. Electrical impedance spectroscopy (EIS) was performed at the open circuit voltage
with an amplitude of 10 mV. Potentials were corrected for t
he electrolyte/cell resistance from
the high frequency intercept of the real impedance (
∼
45
-
60
Ω
). Discrete points on the Tafel
plot were obtained from potentiostatic measurements. Lines on the Tafel plot were obtained
from the capacitance
-
corrected CV, where the forward and reverse scans were averaged.
1.3
Density Functional Theory
Our model system
s
employed (100), (110), (001), (101), (111) surface slab
s
of IrO
2
with at
least 15 Å vacuum between them.
For these facets, the slab thickness of the fully oxidized
surface corresponds to 9.00, 15.86, 12.31, 12.75, and 13.40 Å, respectively. Only the top two
layers are allowed to relax for (100), one for (110), two for (001), one for (101), and three for
(111)
IrO2 layers.
We
used
2×2, 2×1, and 1×1 surface supercells for (100), (110)
,
and the rest
of IrO
2
facets
, respectively
. The Vienna ab initio simulatio
n package (VASP ver. 5.4.5)
(
3
)
with the VASPsol solvation model
(
4
)
was used for geometry optimization. Electron exchange
and correlation were treated within the generalized gradient approximation (GGA)
(
5
)
in the
form of the PBE functional, including the D3 correction for London Dispersion (van der
3
Waals attraction)
(
6
)
. The interaction between the ionic core and the valence electrons was
described by the projector
-
augmented wave (PAW) method.
(
7
)
We used a plane
-
wave basis
set with an energy cutoff of 500 eV. The Brillouin zone was sampled using the 4×3×1,
3×3×1, 4×4×1, 3×4×1, 4×4×1 Monkhorst
-
Pack grid
(
8
)
for the electronic structure for (100),
(110), (001), (101), and (111) IrO
2
surface slabs, respectively. The convergence criteria for
the electronic structure and the atomic geometry were 10
–
5
eV and 0.03 eV/Å
.
For transition
state searching, we used the climbing image nudged elastic band (CI
-
NEB) method.
(
9
)
After the geometry optimization,
we car
r
ied out
single point calculations using the CANDLE
solvation model
(
10
)
as incorporated in the joint density
-
functional theory (JDFTx)
(
11
)
to
calculate the grand
canonical
free energy as a function of applied potential
(GCP
-
K)
. We
used the GBRV ultrasoft pseudopotential (USPP)
(
12
)
with a plane wave cutoff of 544 eV (20
Hartree). The ionic screening of different net charges was achieved with 0.1 M K
+
and 0.1 M
F
-
in the fluid model. All other settings were similar to those in VASP calculations.
The Gibbs free energy (G) of all surface states at 298 K and 1 atm was calculated as:
퐺
=
퐻
−
푇
∆
푆
=
퐸
퐷퐹푇
+
퐸
푍푃퐸
+
퐸
푠표푙푣
−
푇
∆
푆
+
∫
퐶
푣
푑푇
298
0
where
퐸
퐷퐹푇
is the electronic total energy,
퐸
푍푃퐸
is the zero
-
point vibrational energy, and
퐸
푠표푙푣
is the solvation energy. The enthalpy (
∫
퐶
푣
푑푇
298
0
) and entropy (
∆
푆
) contributions at room
temperature were calculated from the vibrational modes of the system.
1.4
Surface energy diagram
For a given facet, the surface free energy changes (
Δ
G)
for
various surface states
relative to
the fully reduced state were calculated as a function of
applied
potential. The energy of
different surface states depends on the surface oxidation state and the configuration of surface
species.
The equilibrium surface populations of oxyl, hydroxyl, and water on a facet at a potential (U)
are estimated by
θ
푖
(
푈
)
=
1
푍
(
푈
)
∑
푓
푖푗
∙
e
−
∆
퐺
푗
(
푈
)
푁
휇
∙
푘
퐵
푇
푗
where
푓
푖푗
is the fraction of the surface oxygen species
i
on a surface state
j
,
∆
퐺
푗
(
푈
)
is
relative Gibbs free energy of the surface state (
j
) compared to the fully reduced state,
푁
휇
is
the total number of surface
μ
1
-
and
μ
2
-
sites in a given orientation,
푘
퐵
is Boltzmann
constant, and T is room temperature (298.15 K). The partition function,
푍
(
푈
)
, is obtained by
considering all the surface states for a given surface orientation as:
4
Z
(
U
)
=
∑
e
−
∆
퐺
푗
(
푈
)
푁
휇
∙
푘
퐵
푇
푗
The charge Q flowed from surface redox reaction at U is calculated:
Q
(
U
)
=
1
푍
(
푈
)
∑
e
−
∆
퐺
푗
(
푈
)
푁
휇
∙
푘
퐵
푇
푗
∙
푁
푗
푑푒푝푟표
∙
푞
푒
퐴
푠푢푟푓
where
푁
푗
푑푒푝푟표
is number of redox reactions to a surface state
j
,
푞
푒
is the elementary charge
and
퐴
푠푢푟푓
is the surface area.
The theoretical current density is estimated by
푑푄
(
푈
)
푑푡
=
푑푄
(
푈
)
푑푈
∙
푑푈
푑푡
The
푑푄
(
푈
)
푑푈
is
obtained by taking the gradient of Q with respect to U, us
ing
50 mV/s for the
scan rate (
푑푈
푑푡
) to match with our experiment condition.
The accuracy of calculations for estimating
Δ
G can be further improved by increasing the
size of supercell with a larger number of intermediate combinations, exploiting the grand
canonical approach for fixing electrochemical potential
,
or including explicit solvent in
Helmholtz layer.
1.5
Energetics for redox reactions
For CV simulation, a
ll the redox reaction energ
ies
at a
given
potential
are
described using
the
computational hydrogen electrode (CHE) method
(
13
)
. The CHE method
assumes the
solvated proton and electron are in equilibrium with the
g
as phase
hydrogen
at standard
condition,
휇
퐻
+
0
+
휇
푒
−
0
=
1
2
⁄
휇
퐻
2
(
푔
)
0
The proton chemical potential
and electrochemical potential
for proton
-
coupled electron
transfer (PCET) step
in specific working condition
is expressed as
휇
퐻
+
+
휇
푒
−
=
휇
퐻
+
0
+
휇
푒
−
0
+
푘
퐵
푇
ln
푎
퐻
+
−
푒푈
where
푎
퐻
+
is the activity of proton and e is unit charge and U is the applied bias.
1.6
Grand canonical potential kinetics (GCP
-
K)
The reaction free energy (
Δ
G) and kinetic barrier (
Δ
G
‡
) of
all elementary reaction step along
OER pathway including
non
-
electrochemical O
-
O coupling steps were
all
described using the
GCP
-
K scheme.
(
14
)
All GCP for the initial, transition, and final states were obtained to
calculate the
Δ
G
(for all elementary steps)
and
Δ
G
‡
(for O
-
O coupling steps)
for all facets.
The grand canonical free energy of an electrochemical system is defined as
G
(
n
;
U
)
=
퐹
(
푛
)
−
푛푒
(
푈
푆퐻퐸
−
푈
)
5
where
퐹
is the total free energy of a system with number of electron (
푛
),
푒
is unit charge,
푈
푆퐻퐸
is absolute potential of the standard hydrogen electrode (SHE)
,
and
푈
is an applied
potential in absolute scale.
Assuming
a
quadratic dependence of
퐹
(
푛
)
with respect to
푛
, we obtaine the Grand
Canonical Potential (GCP) by minimizing
G
(
n
;
U
)
at a fixed potential (U),
(
14
)
GCP
(
U
)
=
푒
2
퐶
푑푖푓푓
2
(
푈
−
푈
푃푍퐶
)
2
+
푛
0
푒푈
+
퐹
0
−
푛
0
휇
푒
,
푆퐻퐸
where
퐶
푑푖푓푓
is the differential capacitance,
푈
푃푍퐶
is the potential of zero charge,
휇
푒
,
푆퐻퐸
is
the electrochemical potential of SHE,
푛
0
is the total number of electron
s
at zero net charge
and
퐹
0
푖푠
the total free energy a
t n
0
. The
퐶
푑푖푓푓
,
푈
푃푍퐶
, and
퐹
0
values for all surface states
are listed in Table
S4
.
The
Δ
G
‡
of water nucleophilic attack (WNA) is calculated as
∆
퐺
‡
=
퐺퐶푃
푇푆
−
퐺퐶푃
∗
−
휇
퐻
2
푂
−
∆
퐻
2
푂
where the GCP
TS
represent
s
the GCP of
the
transition state
(
Fig. S8 and S9
)
and
GCP
*
the
resting
state of
the
WNA step
without reactant (H
2
O)
. We refer
ence
the chemical potential
to
bulk water,
휇
퐻
2
푂
which is applied
equivalently
to all orientations. The constant correction
term (0.
15
eV),
훿
퐻
2
푂
, is introduced for all facet in order to reflect the
stability of
interfacial
water network
prior to WNA
which is determined by fitting theoretical current densities from
microkinetic analysis to the experimental current densities (See Parameter fitting section).
1.7
Microkinetic analysis
We built microkinetic
models in terms of the surface state population rather than the
intermediate coverage to distinguish the reduction of the same species for different surface
states having different intermediate
-
intermediate interaction. For example, we can separate
the OH
reduction of a
μ
-
OH/
μ
-
O state from a
μ
-
OH/
μ
-
H
2
O
or 2
μ
-
OH
state using this
approach.
Due to the O
2
overbinding of
semi
-
local functional (PBE)
more than > 0.4 eV compared to
hybrid functionals
(
15
,
16
)
,
we assumed a fast
equilibrium O
2
/H
2
O populations based on the
exchange thermodynamics.
For a given reaction, we considered both forward and backward reaction to estimate the net
reaction rate without exception.
The full list of elementary reaction and partial differential
equations for each facet are listed in the
Appendix
. The steady
-
state coverage as non
-
trivial
solutions for the microkinetic equations are obtained using
the
open source SUNDIAL
differential equation solver
(
17
)
through
the
Assimulo package.
(
18
)
The
CVODE solver is
used for the Backward Differentiation Formulas (BDFs)
,
which
is
suitable for stiff problems.
6
T
he turnover frequency (TOF)
for O
2
gas generation
is calculated
by summing up all possible
rate from
O
2
/H
2
O exchange steps, then the
steady
-
state
OER current density
(
퐽
)
for an IrO
2
facet
is calculated as
퐽
=
TOF
∙
푛
푃퐶퐸푇
∙
F
푁
퐴
∙
푛
푆푖푡푒
/
퐴
푠푢푟푓
where
푛
푃퐶퐸푇
is the number of proton
-
coupled electron steps for a single OER cycle,
F
is
the Faradaic constant,
푁
퐴
is the Avogadro constant,
푛
푆푖푡푒
is the number of active sites in
the supercell, and
퐴
푠푢푟푓
is the
surface area of the supercell.
1.8
Parameter fitting
We use
DFT
solely
to
describes the surface chemistry.
In order to account for
specific
working condition
s
, we employed a parameter
(
훿
퐻
2
푂
)
for
휇
퐻
2
푂
and two
additional
parameters
(
α and β
)
for the
kinetic
barrier of PCET step. The former
parameter
is employed
to
assess
the free energy
distinction
between
our
single water model as a reactant (H
2
O) for
WNA step and the
interfacial water
with hydrogen bond network
. The kinetic barrier of
PCET step
(deprotonation step) is described by Bell
–
Evans
–
Polanyi (BEP) principle
(
19
,
20
)
assuming the PCET energy linearly depends on the
reaction free energy
,
∆
퐺
푃퐶퐸푇
‡
=
훼
+
훽
×
∆
퐺
푃퐶퐸푇
To
calibrate
the theoretical current density (
푗
푡
ℎ
푒표푟푦
) to
align with
the
experimental current
density (
푗
푒푥푝
),
we
employed
the
second
Linear Sweep Voltammetry (
LSV
)
curve for each
facet from experiment to minimize the
impact of
corrosion while activated.
The
error
is calculated as
error
=
∑
∑
[
푗
푡
ℎ
푒표푟푦
(
푖
,
푈
)
−
푗
푒푥푝
(
푖
,
푈
)
]
푈
푖
where
푖
represents
a facet and
푈
denotes the
applied
potential
being
considered
(
1.60, 1.65,
1.70 V vs RHE
which are within Tafel region in experiment).
The optim
al parameter set obtained
i
s
as:
훿
퐻
2
푂
=0.1
5
eV, α
=
0.
79
eV and β
=
0.
46
.
7
Supplemental Figures and Tables
(100)
(110)
(001)
(101)
(111)A
(111)B
ρ(
퐼푟
퐶푈푆
)
[#/cm
2
]
6.91E+14
4.92E+14
4.91E+14
8.02E+14
3.68E+14
3.68E+14
γ
[J/m
2
]
2.263
1.765
2.747
1.928
2.276
2.555
ρ(
휇
1
−
푂
) [#/cm
2
]
6.91E+14
4.92E+14
9.82E+14
8.02E+14
3.68E+14
1.47E+15
Q [
μ
C/cm
2
]
1.04E+02
8.88E+01
1.18E+02
9.73E+01
1.73E+02
1.05E+02
d(O
-
O) [Å]
3.19
3.18
3.21
(2.80*)
2.77
5.52
2.81
(2.69*)
Table S1 Physical properties on five different IrO
2
facets.
The density of Ir
CUS
site
(
ρ(
퐼푟
퐶푈푆
)
), surface energy (
γ
) of stoichiometric surface, and the density of terminal oxygen
(
ρ(
휇
1
−
푂
)
) on fully oxidized surface.
T
he theoretical integrated charge (
Q
) transferred during
anodic scan (0.3 V
RHE
→
1.2 V
RHE
), and the atomic distance between two
μ
1
-
O. * denotes the
distance which in both oxygen are bonded to the same surface Ir atom.
Orientation
q*
cath
acid
(μC/cm
2
)
(110)
178
(100)
140
(101)
129
(001)
257
(111)
182
Table
S2
.
Integrated charge density from the region below the OER onset, referred to as q*
Facet
V vs RHE
Lowest energy surface
IrO
2
(100)
0.00~0.05
4μ
2
-
OH/μ
1
-
OH/
3
μ
1
-
H
2
O
0.06~0.22
4μ
2
-
OH/2μ
1
-
OH/2μ
1
-
H
2
O
0.23~0.55
4μ
2
-
OH/3μ
1
-
OH/μ
1
-
H
2
O
0.56~0.96
μ
1
-
H
2
O/3μ
1
-
OH/3μ
2
-
OH
0.97~1.08
2μ
1
-
H
2
O/2μ
1
-
OH/
μ
2
-
OH
1.09~1.22
μ
1
-
H
2
O/3μ
1
-
OH
8
1.23~1.50
4μ
1
-
OH
1.51~2.00
4μ
1
-
O/4μ
2
-
O
IrO
2
(110)
0.00~0.59
2μ
2
-
OH/μ
1
-
OH/μ
1
-
H
2
O
0.6~1.03
2μ
2
-
OH/2μ
1
-
OH
1.04~1.17
μ
2
-
OH/2μ
1
-
OH
1.18~1.50
2μ
1
-
OH
1.51~2.00
2μ
1
-
O/2μ
2
-
O
IrO
2
(001)
0.00~0.23
2μ
2
-
OH/μ
1
-
OH/μ
1
-
H
2
O
0.24~0.77
μ
2
-
OH/μ
1
-
OH/
2
μ
1
-
H
2
O
0.78~1.24
μ
1
-
OH/
2
μ
1
-
H
2
O
1.25~1.50
2μ
1
-
OH
1.51~2.00
2μ
2
-
O/2μ
1
-
O
IrO
2
(101)
0.00~0.
18
2μ
2
-
OH/μ
1
-
OH/2μ
1
-
H
2
O
0.00~0.84
2μ
1
-
H
2
O
0.85~1.37
μ
1
-
OH/μ
1
-
H
2
O
1.38~1.72
2μ
1
-
OH
1.73~2.00
2μ
2
-
O/2μ
1
-
O
IrO
2
(111)B
0.00~0.
01
2μ
1L
-
H
2
O/μ
1L
-
OH/μ
1H
-
H
2
O
0.02~0.91
μ
1
L
-
H
2
O/2μ
1L
-
OH/μ
1H
-
H
2
O
0.92~1.11
3μ
1
L
-
OH/μ
1
H
-
H
2
O
1.12~1.42
3μ
1
L
-
OH/μ
1
H
-
OH
1.43~1.63
μ
1
L
-
OH
1.64~2.00
3μ
1
L
-
O/μ
1
H
-
O/μ
2
-
O
Table S3.
The most stable surface state
for each
applied potential for five IrO
2
facets.
IrO
2
facet
Final state
state
a
b
c
(100)
μ
1
-
OOH
/
μ
1
-
OH
IS
0.468681646
-
6.451141234
-
49021.13115
TS
0.557310425
-
6.058714685
-
49490.54355
FS
0.482804408
-
6.04283103
-
49490.91063
μ
1
-
OOH
/
μ
2
-
OH
IS
0.468681646
-
6.451141234
-
49021.13115
TS
0.464021117
-
6.015318388
-
49490.59655
FS
0.53390765
-
5.946603436
-
49490.7984
μ
1
-
OOH
/
μ
1
-
H
2
O
IS
0.327887599
-
6.295492632
-
49038.50311
TS
0.516728238
-
6.021518673
-
49507.97617
FS
0.574569427
-
5.907697111
-
49508.2223
(110)
μ
1
-
OOH
/μ
1
-
OH
IS
0.729933858
-
6.292781568
-
59979.87605
TS
0.668442909
-
5.849800517
-
60449.2684
FS
0.853870394
-
5.675155349
-
60449.76368
μ
1
-
OOH/
μ
2
-
OH
IS
0.729933858
-
6.292781568
-
59979.87605
TS
0.771446818
-
5.801444833
-
60449.25371
FS
0.955362624
-
5.244093564
-
60449.31189
μ
1
-
OOH
/
μ
1
-
H
2
O
IS
0.611473092
-
6.126148779
-
59997.21745
TS
0.875729239
-
5.474874385
-
60466.61146
FS
1.010700869
-
5.143694485
-
60466.70784
(001)
μ
1
-
OOH
/
μ
1
-
OH
IS
1.779872355
-
5.859181137
-
24512.25086
TS
1.619376166
-
5.376354231
-
24981.3527
FS
2.173408268
-
5.331910903
-
24981.99679
μ
1
-
OOH
/
μ
1
-
H
2
O
IS
1.916995998
-
5.550949935
-
24529.4671
TS
1.72957967
-
4.854963326
-
24998.57556
FS
2.191929136
-
4.153147576
-
24998.73211
(101)
μ
1
-
OOH
/
μ
1
-
OH
IS
1.363600096
-
6.276799196
-
30423.79651
TS
1.385048891
-
5.526295529
-
30893.13699
9
FS
1.289202223
-
5.59731014
-
30893.76223
μ
1
-
OOH
/
μ
2
-
H
2
O
IS
1.06064302
-
6.241692136
-
30441.18809
TS
1.551559021
-
5.60503226
-
30910.53309
FS
1.811408818
-
5.212155155
-
30910.73042
μ
2
-
O
2
IS
1.363600096
-
6.276799196
-
30423.79651
TS
1.394816526
-
6.244771964
-
30423.3349
FS
1.608231915
-
6.073571089
-
30423.564
(111)
B
μ
1H
-
OOH
/
μ
1L
-
OH
IS
1.27151856
-
4.997679554
-
33374.74688
TS
1.057712548
-
4.969714819
-
33843.87785
FS
1.377643768
-
4.952860724
-
33844.56199
μ
1L
-
OOH
/
μ
1
H
-
OH
IS
1.27151856
-
4.997679554
-
33374.74688
TS
1.17507214
-
4.887254356
-
33843.7863
FS
1.33497349
-
4.220876169
-
33844.40317
μ
1
H
-
OOH
/
μ
1
L
-
H
2
O
IS
1.370023498
-
4.987064361
-
33392.15832
TS
1.265768222
-
4.773066986
-
33860.89775
FS
2.228613357
-
2.779372205
-
33860.95188
μ
1
L
-
OOH
/
μ
1
H
-
H
2
O
IS
1.291167656
-
4.860462635
-
33391.94736
TS
1.297841725
-
4.838712159
-
33861.1375
FS
1.055471456
-
4.614027346
-
33861.43354
Table
S4
Grand canonical potential
–
kinetics (
GCP
-
K)
Parameters for
O
-
O
coupling
step
.
The a, b, c parameters for the quadratic dependence of free energy F(n) = a(n
-
n
0
)
2
+
b(n
-
n
0
) + c to get grand canonical potential.
(
14
)
Orientation
Tafel slope in 0.1 M
HClO
4
(mV/decade)
(001)
78.7
(110)
66.9
(100)
112
(111)
219
(101)
170
Table
S5
.
Tafel slope from chronoamperometry (measured after cycling).
10
Figure S1.
(A)
XRD and representative
(B
-
F)
AFM
.
(
21
)
The sharpest XRD peaks arise from
the substrate, with broader, lower intensity features arising from the film. Film peaks include
(111) and (222), (001), (200), (101) and (202), (110) (220) and (330). The peak marked with
* arises from <0.3% (100)
-
orientat
ion. The (100) film has additional substrate peaks from
growth on
(001)
-
SrTiO
3
.
RMS roughness (nm) from the shown images are (B) 5.97, (C) 6.20,
(D) 3.22, (E) 1.87, and (F) 3.15 nm.
11
Figure S2.
Tafel plot from chronoamperometry (measured after cycling)
in 0.1 M HClO
4
.
Figure
S3
.
CV (corrected for capacitance and iR
drop) normalized by q*
(
Table
S5
)
in 0.1 M
HClO
4
, cycle
2
.
Trends between orientations are comparable to normalization by surface
area.
12
Figure
S4
.
Surface geometries
of five IrO
2
facets
(
A
to
F
) stoichiometric surfaces (
G
to
L
)
fully reduced
state (
M
to
R
) fully oxidized state.
The location of the 1
-
fold (
μ
1
-
O) and 2
-
fold
oxygen sites (
μ
2
-
O) are indicated in each fully oxidized surfaces.
All
μ
1
-
species are bonded
to the Ir
CUS sites regardless of facet orientation.
On IrO
2
(111)
B
surface,
μ
1l
-
denotes the 1
-
fold site bonded to low
-
coordinated surface Ir
4c
and
μ
1h
-
denotes the 1
-
fold site bonded to
high
-
coordinated surface Ir
5c
.
* The pseudo
-
stoichiometric IrO
2
(111)A surface has
O:Ir=24:11 to have 5 coordination of surface Ir.
13
Figure
S5
.
(
A
to
F
)
Surface free energy diagrams,
(
G
to
L
)
surface intermediate coverages,
(
M
to
R
)
Charge flow and
corresponding
current density from surface redox reaction and
(
S
to
X
)
current density comparison
s
between theory
and experiment
for (
A
,
G
,
M
,
S
)
IrO
2
(100), (
B
,
H
,
N
,
T
) IrO
2
(110), (
C
,
I
,
O
,
U
) IrO
2
(001), (
D
,
J
,
P
,
V
) IrO
2
(101), (
E
,
K
,
Q
,
W
) IrO
2
(111)
A, and (
F
,
L
,
R
,
X
) IrO
2
(111)B
.