of 24
S
1
The Predominance of Hydrogen Evolution on Transition Metal Sulfides
and Phosphides
u
nder CO
2
Reduction Conditions: An Experimental and
Theoretical Study
Alan T. Landers,
a, b, 1
Meredith Fields,
b, c, 1
Daniel A. Torelli,
d, 1
Jianping Xiao,
b, c
Thomas R.
Hellstern,
b, c
Sonja A. Francis,
d
Charlie Tsai,
b, c
Jakob Kibsgaard,
b, c, e
Nathan S. Lewis,
d, 2
Karen
Chan,
b, c, 2
Christopher Hahn,
b, c, 2
Thomas F. Jaramillo
b, c, 2
a. Department of Chemistry, Stanford University, Stanford, CA 94305
b.
SUNCAT Center
for Interface Science and Catalysis, SLAC National Accelerator Laboratory,
Menlo Park, CA 94025
c. Department of Chemical Engineering, Stanford University, Stanford, CA 94305
d.
Division of Chemistry and Chemical Engineering, California Institute of Techno
logy,
Pasadena, CA 91125
e.
Department of Physics,
Technical University of Denmark
, DK
-
2800 Kongens Lyngby,
Denmark
1. A.T.L, M.F., and D.A.T. contributed equally to this work.
2. To whom correspondence should be addressed: N.S.L.
nslewis@caltech.edu
, K.C.
chank@stanford.edu
, C.H.
chahn@slac.stanford.edu
, T.F.J.
jaramillo@stanford.edu
S
2
Supporting
Info
rmation
Experimental
Methods
The molybdenum sulfide and TM phosphide thin films were synthesized using previously
reported methods.
1
-
2
First, ten nanometers of the TM were deposited onto a silicon substrate using
electron beam physical vapor deposition. For the TM
-
doped molybdenum samples, the films were
deposited with a Mo:TM
ratio of 3:1. Subsequently, a vapor
-
assisted process in a tube furnace
converts the metal thin films into the corresponding TM sulfide or phosphide. The molybdenum
thin films convert to the sulfide when heated to 250
o
C under a mixture of H
2
and H
2
S gas
(
Caution:
H
2
S is a highly toxic gas. Both H
2
S and H
2
are flammable gases)
. To convert the samples to
phosphides, the metal thin films and a sample of red phosphorus were heated in a tube furnace
while flowing H
2
gas
(
Caution:
Red phosphorus is a highly flam
mable solid with an auto
-
ignition
temperature as low as 260
o
C
. H
2
is a flammable gas)
.
1
X
-
ray photoelectron spectroscopy
confirmed the formation of the ionic materials based on the presence of a phosphide peak in the P
2p region, a sulfide peak in the S 2p region, and appropriate metal oxidation sta
tes which were
consistent with previous reports.
1
-
2
The materials were evaluated for CO
2
R activity in CO
2
sparged
0.1
0
M KHCO
3
using a previously described flow cell.
3
The synthesis for nanoparticulate catalysts
is reported elsewhere.
4
-
6
Briefly, tri
-
n
-
octyphosphine (TOP) was added as a phosphorus source in
equal volume to a 1:1 mixture of 1
-
o
ctadecene and oleylamine. This mixture was heated to 120
°C
under vacuum f
or 1
hour
in a three
-
necked round
-
bottom flask equipped with a reflux condenser.
The mixture was then heated to 330
°C under Ar and premade metallic nanoparticles suspended
in degassed TOP were added and solution was stirred for 1 hr. The resulting solutio
n was then
cooled to room temperature, centrifuged and washed with a mixture of hexanes and isopropanol.
The resulting nanoparticles were then suspended in hexanes under N
2
. To prepare a working
S
3
electrode, this suspension was drop cast onto a pyrolytic gra
phite plate at a loading of 1 mg/cm
2
.
The electrodes were then annealed under 5% H
2
/N
2
at 400
°C for 1 hr. Thin films of SnS were
synthesized using literature procedures.
7
-
8
Briefly, bulk SnS powder was dissolved in 11:1 vol/vol
mixture of ethylene
diamine and 1,2
-
ethanedithiol at 50 °C for 15 h at a concentration of 60 mg
mL
-
1
. Solutions were then filtered using a 0.45 μm filter and spin coated onto FTO substrates.
Samples were then annealed on a hot plate at 350 °C under flowing N
2
and finally at 5
00 °C in a
tube furnace with flowing N
2
to increase robustness for electrochemical measurements.
7
-
8
Computational Methods
We have employed QUANTUM ESPRESSO code for total energy calculations, with
plane
-
wave and density cutoffs of 500 and 5000 eV, respectively. K
-
point sampling grids of (2 ×
2 × 1) for sulfide surfaces and (4 x 4 x 1) for phosphide surfaces as well as a 0.1 e
V Fermi
-
level
smearing were chosen
based on
convergence tests from previous work. All calculations
implemented the Bayesian error estimation Functional with Van der Waals correction (BEEF
-
vdW) exchange correlation functional. All structures were relaxed un
til all force components were
< 0.05 eV
1
.
In addition, spin
-
polarized calculations were performed for all systems containing
Ni, Fe, and Co.
For sulfide surfaces,
a monolayer of water and explicit H
3
O
+
were
used to
determine electrochemical transition state for CO protonation to CHO. Barriers were determined
using
the
climbing image
nudged elastic band (NEB) method
,
9
and a charge extrapolation
method
10
-
11
was used to determine the potential dependence of the electrochemical barriers.
As
detailed by Chan and N
ørskov
,
11
a Bader analysis
12
was applied to determine the degree of charge
transfer across the electrochemical interface at the transition states, which
provides the
corresponding transfer coefficients.
For further calculation details, lattice constants, and
S
4
optimization parameters for both phosphides and sulfide surfaces, see corresponding references
.
13
-
15
S
5
a.
b.
S
6
c.
d.
Figure S
1
.
X
-
ray photoelectron spectra of MoP prior to
(
a
) and after (
b
)
electrochemical testing.
Peaks
attributed
to Mo
3+
and P
3
-
are identified.
1
X
-
ray photoelectron spectra of MoS
2
prior to
(
c
)
and after (
d
)
electrochemical testing
.
Peaks corresponding to Mo
4+
and S
2
-
are identified.
2
Note
that neither the pre
-
nor the post
-
reaction characterization necessarily reflect the surface under
electrochemical conditions. Oxidized species at the surface could be reduced at the negative
potentials applied during electrolysis. After the electroly
sis finishes, the sample briefly returns to
its open circuit value in the electrolyte before being removed to the atmosphere and transferred to
S
7
the XPS. During this transfer process, the surface of the material could become oxidized relative
to its state u
nder electrochemical conditions. To understand the actual surface during CO
2
R
conditions, in situ characterization would be required.
S
8
Figure S
2
:
CO preferentially adsorbs on the metal sites of the CoP surface. While the *CO
binding energy decreases slightly with increasing *CO coverage, it is only when all Co metal
sites are saturated and the *CO is forced to occupy a P
-
site that the binding energy
weakens
substantially. It is possible that at steady state, the CoP surface is operating at higher coverages
of *CO than addressed in the DFT thermodynamic analysis. Full kinetic analysis would be
required to determine true steady state coverages.
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
0
1
2
3
4
5
6
Binding Energy (eV)
# Occupied Sites
CO Coverage
H Coverage
S
9
a.
b
.
Figure S3.
Panel (a)
shows CO bound to a phosphorus site on the CoP surface.
Panel (b)
shows
CO bound to a metallic Co site on the same CoP surface.
Charge density differences
are defined so that:
=
푠푢푟푓
+
푎푑푠
푎푑푠
푠푢푟푓
Decreased Electron Density
Increased Electron Density
Decreased Electron Density
Increased Electron Density
S
10
w
here
푎푑푠
is the charge density of the gas phase adsorbate,
푠푢푟푓
the density of the pristine slab,
and
푠푢푟푓
+
푎푑푠
the density of the adsorbed system. Therefore,
Δ
represents
the charge transfer
between the adsorbate and slab. In these calculatio
ns, the full system is fully relaxed, and a single
point charge density calculation is then calculated for each density component. Red volumes
represent regions of decreased electron density, while blue volumes represent regions of increased
electron densi
ty. All isosurfaces are visualized in VMD and taken at +/
-
0.001 isovalues
(e Bohr
-
1
). For CO bound to metallic sites, significant stabilization is seen in comparison to CO
bound to P
-
sites. This may be attributed to CO back
-
bonding.
S
11
Table S
1
. Complete Product Distribution for All Tested Catalysts
Material
Morphology
Potential
(V vs.
RHE)
Current
Density
(mA/cm
2
)
H
2
FE
1
(%)
CO
FE
1
(%)
CH
4
FE
1
(%)
HCOO
-
FE
1
(%)
CA or
CP
2
MoP
Thin film
-
0.39
-
0.6
10
4
0
0
0
CA
-
0.47
-
1.9
12
9
0
0
0
CA
-
0.50
-
3.
4
9
6
0
0
0
CA
-
0.54
-
5.1
11
0
0
0
0
CA
-
0.57
-
5.0
103
0
0
0
CP
-
0.
59
-
1
4.4
94
0
0
0
CA
CoP
Thin film
-
0.29
-
0.
4
10
4
0
0
0
CA
-
0.39
-
0.7
11
0
0
0
0
CA
-
0.48
-
1.
2
10
2
0
0
0
CA
-
0.57
-
1.
9
11
0
0
0
0
CA
-
0.61
-
5.0
92
0
0
0
CP
-
0.65
-
4.0
10
3
0
0
0
CA
-
0.6
7
-
6.
2
10
4
0
0
0
CA
Ni
x
P
Thin film
-
0.40
-
0.1
101
0
Trace
0
CA
-
0.50
-
0.2
12
3
Trace
Trace
0
CA
-
0.59
-
0.6
10
7
0
0
0
CA
-
0.68
-
1.
3
9
4
0
0
0
CA
-
0.75
-
3.
6
84
Trace
0
0
CA
-
0.7
8
-
9.0
1
05
Trace
0
0
CA
MoS
2
Thin film
-
0.53
-
0.7
10
0
0
0
0
CA
-
0.62
-
1.9
9
5
0
0
0
CA
-
0.69
-
4.
7
1
12
0
0
0
CA
-
0.7
4
-
5.0
9
0
0
Trace
Trace
CP
-
0.74
-
10.0
9
9
0
0
0
CA
-
0.7
5
-
1
4.6
113
0
0
0
CA
Ni
-
MoS
x
Thin film
-
0.50
-
0.4
98
0
0
0
CA
-
0.59
-
1.3
10
1
0
0
0
CA
-
0.67
-
3.
4
92
0
0
0
CA
-
0.71
-
9.0
9
4
0
0
0
CA
S
12
-
0.7
5
-
1
4.8
95
0
0
0
CA
-
0.80
-
5.0
97
0
Trace
Trace
CP
Co
-
MoS
x
Thin film
-
0.40
-
0.3
109
0
0
0
CA
-
0.49
-
0.8
88
0
0
0
CA
-
0.59
-
1.9
87
0
0
0
CA
-
0.66
-
4.0
114
0
0
0
CA
-
0.73
-
8.0
108
0
0
0
CA
Fe
-
MoS
x
Thin film
-
0.60
-
0.6
9
5
0
0
0
CA
-
0.68
-
1.8
12
2
0
0
0
CA
-
0.75
-
4.
9
11
7
0
0
0
CA
-
0.79
-
11.
9
89
0
0
0
CA
SnS
Thin Film
-
0.8
-
0.6
78
1.4
0
0
CA
-
1.0
-
1.3
71
3.4
0
0
CA
MoP
Nanoparticles
-
0.7
-
2.9
89
Trace
0
0
CA
-
0.9
-
4.0
76
Trace
Trace
Trace
CA
-
1.0
-
15.7
90
Trace
Trace
Trace
CA
Ni
2
P
Nanoparticles
-
0.7
-
9.7
92
0
0
0
CA
-
0.85
-
12.2
85
0
0
0
CA
-
0.90
-
13.2
92
0
0
0
CA
CoP
Nanoparticles
-
0.69
-
12.0
95
0
Trace
0
CA
-
0.89
-
19.6
80
0
0
0
CA
-
0.92
-
19.5
91
0
Trace
0
CA
WP
Nanoparticles
-
0.85
-
6.2
92
Trace
Trace
0
CA
-
1.2
-
15.2
91
Trace
Trace
0
CA
IrP
Nanoparticles
-
0.9
-
11.9
99
0
0
0
CA
RhP
Nanoparticles
-
0.9
-
11.4
1
15
0
Trace
0
CA
1.
FE refers to Faradaic efficiency.
2.
CA refers to
chronoamperometery while CP refers to chronopotentiometry
S
13
Figure Data
Figure 1
Material
Morphology
Potential
(V vs. RHE)
H
2
FE (%)
MoP
Thin film
-
0.39
10
4
-
0.47
12
9
-
0.50
9
6
-
0.54
11
0
-
0.57
103
-
0.
59
94
CoP
Thin film
-
0.29
10
4
-
0.39
11
0
-
0.48
10
2
-
0.57
11
0
-
0.61
92
-
0.6
5
10
3
-
0.6
7
10
4
Ni
x
P
Thin film
-
0.40
101
-
0.50
12
3
-
0.59
10
7
-
0.68
9
4
-
0.75
84
-
0.7
8
1
05
MoS
2
Thin film
-
0.53
10
0
-
0.62
9
5
-
0.69
1
12
-
0.7
4
9
9
-
0.74
90
-
0.7
5
1
13
Ni
-
MoS
x
Thin film
-
0.50
98
-
0.59
10
2
-
0.67
92
-
0.71
94
-
0.7
5
95
S
14
-
0.80
9
7
Co
-
MoS
x
Thin film
-
0.40
109
-
0.49
88
-
0.59
87
-
0.66
114
-
0.73
108
Fe
-
MoS
x
Thin film
-
0.60
9
5
-
0.68
12
2
-
0.75
11
7
-
0.79
89
MoP
Nanoparticles
-
0.7
0
89
-
0.9
0
76
-
1.0
0
90
Ni
2
P
Nanoparticles
-
0.7
0
92
-
0.85
85
-
0.90
92
CoP
Nanoparticles
-
0.69
95
-
0.89
80
-
0.92
91
WP
Nanoparticles
-
0.85
92
-
1.2
0
91
IrP
Nanoparticles
-
0.9
0
99
RhP
Nanoparticles
-
0.9
0
1
15
SnS
Thin Film
-
0.8
0
78
-
1.0
0
71