S
1
Supporting Information
Nickel
-
Gallium
-
Catalyzed Electrochemical Reduction of CO
2
to Highly Reduced Products
at Low Overpotentials
Daniel A. Torelli
¶†‡
, Sonja A. Francis
¶†‡
£
, J. Chance Crompton
†‡
, Alnald Javier
‡
, Jonathan R.
Thompson
ǁ
, Bruce S. Brunschwig
*
§
, Manuel P. Soriaga
*
‡
, and Nathan S. Lewis
*
†‡§
†
Division of Chemistry and Chemical Engineering,
‡
Joint Center for Artificial Photosynthesis,
£
Resnick Sustainability Institute,
ǁ
Division of
Engineering and Applied Sciences,
and
§
Beckman
Institute Molecular Materials Research Center,
California Institute of Technology
, Pasadena,
California 91125,
United States
¶These authors contributed equally
to this work
*E
-
mail:
bsb@caltech.edu
,
msoriaga@caltech.ed
u
,
nslewis@caltech.edu
S
2
Experimental.
Sample Preparation.
Quartz slides (VWR),
gallium (III) nitrate hydrate 99.999% [69% gallium by mass] (Sigma
Aldrich) and nickel (II) nitrate hexahydrate 99.999% (Sigma Aldrich) were used as received.
Pyrolytic graphite plates (GraphiteStore) were cleaned in aqua regia and polished to a mirror
fi
nish before use. Water with a resistivity > 18 M
Ω
cm obtained from a Barnsted Nanopure
system was used throughout.
Aqueous stock solutions of 0.052 M nickel nitrate and 0.036 M gallium nitrate were prepared
and were combined in appropriate ratios to prepar
e precursor solutions for NiGa, Ni
5
Ga
3
, and
Ni
3
Ga. For an individual sample, 0.5 mL of the precursor solution was drop
-
cast onto
a graphite
plate heated to 100 °
C. After drying on a hotplate for five min, the samples were placed in
porcelain boats and then
loaded into a quartz tube in a Carbolite tube furnace. The tube furnace
was placed under a constant 4 L min
-
1
flow of forming gas (5% H
2
, 95% N
2
) and then heated to
700 °C for 3 h, before cooling to room temperature. The back of the graphite plate was the
n
covered in 3M 470 Electroplating Tape (Uline.com) to mask from electrical contact with the
solution, and an alligator clip was used to make contact to the front part of the plate with the
Ni
x
Ga
y
film.
Control Ni
-
only samples and Ga
-
only samples were made
by the same method but
only drop
-
casting one of the precursor solutions.
Powder X
-
ray diffraction (XRD) patterns were collected at room temperature on a Bruker D2
Phaser Diffractometer with a copper K
α
source
(
λ
= 1.54184
Å
) and a LynxEye
-
1D detector.
Simulated XRD patterns
were generated by the Crystal Maker and Crystal Diffract software
package.
Electrochemical Testing.
A BioLogic SP
-
200 potentiostat (Biologic, Grenoble, France) was used for all electrochemical
testing. T
he uncompensated cell resistance was determined from a single
-
point high
-
frequency
impedance measurement and was compensated (85%) by the built
-
in positive
-
feedback software.
A modified two
-
compartment cell was used for all electrochemical measurements. Th
e cell
consisted of a Pyrex weigh bottle that had been modified with ground
-
glass joints in its lid
(
Figure S3
). The joints were used to insert electrodes and make seals during electrolysis, while
the bottle was used so that the entire lid of the cell coul
d be removed to accommodate larger (2
-
4
cm
2
) working electrodes. A polyether ether ketone (PEEK) tube was modified with a detachable
ring at the bottom
that
was fit with a fresh Selemion anion
-
exchange membrane before each
electrolysis or
set of
cyclic vol
tammograms (CV). The Pt mesh counter electrode was behind the
membrane. A Ag/AgCl (3M NaCl) fritted reference electrode (CH Instruments) was used, and
potentials were converted to
values relative to a
reversible hydrogen electrode (RHE) using the
equation
:
E
(vs. RHE) =
E
(vs. Ag/AgCl) + 0.197 V + 0.0591 V x pH. The working electrode
was the Ni
-
Ga film supported on the graphite plate described above. Electrical contact was made
with a stainless steel alligator clip.
The electrolyte used in all cycling test
s was potassium
phosphate buffered to pH 7. No anodic cycling was performed before electrolyses. All cycling
was performed on fresh electrodes to show their electrochemical behavior. During the first
minute of electrolysis, the current decreased by ~5
-
10x
as the native oxide was reduced. At these
S
3
potentials, a Pourbaix diagram shows that only the metallic state of Ni is stable.
Bulk electroly
sis
was performed in a cell that was sealed under 1 atm CO
2
and stirred at ~ 1,000 rpm.
Electrolyses
were performed at varied potentials ranging from
-
0.9 V to
-
1.8 V versus a Ag/AgCl (3 M NaCl)
reference electrode, in 0.1 M Na
2
CO
3
(Sigma
-
Aldrich
≥
99.999% metal basis) that had been
acidified to pH 6.8 with 1 atm CO
2
creating a HCO
3
-
/CO
2
buff
ered system similar to the standard
protocol.
1
Na
2
CO
3
was used instead of
K
HCO
3
due to the much higher purity available for
Na
2
CO
3
.
Although the faradaic efficiency can be affected by the cation in the electrolyte,
for a
Cu electrode,
the production of C
2
products is lower for Na
+
than K
+
.
3
It is consequently
reasonable
to assume sim
ilar, if not higher
,
yields of C
2
products with Ni
x
Ga
y
when
a KHCO
3
electrolyte
is used
instead of NaHCO
3
.
Electrolyses were performed until a set amount of charge,
generally 40 C
cm
-
2
was passed. However, for some of the lower potentials
, a lower amount
of
charge was passed in the experiments
.
Electrolyses at
-
0.5 V were run for
>12 h to obtain
a
significant
concentration
of
the
products. The calculated yields suggest the electrodes are stable
on this timescale.
In contrast, po
lycrystalline
Cu has
been
re
ported
to deactivate over the same
period of time, at the same potential.
6
Additionally, the
13
CO
2
experiment was run for > 10 h
with
a data
point approximately every 2 h
. No decrease
s
in the rate
s
of
13
CH
4
,
13
C
2
H
4
, or
13
C
2
H
6
w
ere
observed over this period of time.
CO Reduction
.
For experiments where CO or N
2
was required
,
K
2
HPO
4
was used as the electrolyte and
buffered to pH 7 to create similar conditions to
those
under CO
2
.
The use of the CO
2
-
free buffer
eliminates the possibili
ty that CO
2
produced from the equilibration of HCO
3
-
with CO
2
was
responsible for the measured response rather than being attributable to CO or N
2
.
The CV data
suggest
that
Ga slows the binding of CO rather than weakens the
Ni
-
CO
interaction. The
potential at which CO oxidation is observed does not shift significantly (<50 mV) between the
Ni film and the Ni
5
Ga
3
film, suggesting
that
the strength of the metal
-
CO bond is approximately
equivalent between the two films. However, the t
ime it takes for CO to rebind is significantly
longer on the Ni
5
Ga
3
f
ilm compared to the Ni film, suggesting
the presence of
kinetic differences
between the two films.
Direct infrared spectroscopic detection of surface
-
bound CO on the Ni
-
Ga systems has not
been observed to date.
Product Analysis.
An Agilent 7890A gas chromatograph (GC), with two thermal conductivity detectors, was used
to separate and quantify the gases in the headspace of the electrochemical cell. The oven was set
to 50 °C for 9 min foll
owed by ramping at a rate of 8 °C min
-
1
to 80 °C for a total run time of 14
min. For isotopic labeling experiments that involved smaller amounts of charge passed, an
Agilent 7820A GC coupled with a 5977E MS with a heated cold quadrupole detector and a
cap
illary CarbonPLOT column was used for identification and quantification of the products.
The oven was set to 35 °C for 6.6 min and was then ra
mped to 150 °C at a rate of 20
°
C
min
-
1
and held for 2 min to allow heavier molecules such as ethylene and ethane
to elute. Both the GC
and GCMS instrumentation was calibrated using tanks of 15% CH
4
, 10% C
2
H
4
, and 5% C
2
H
6
each mixed with N
2
. Dilutions were performed by filling a sealed
1
or
3
L round
-
bottom flask
with N
2
or CO
2
,
and injecting known amounts of the calibration gas. The gaseous mixture was
allowed to stir for ~1 min at which time aliquots were removed for GC or GCMS calibration.
1
H
NMR spectroscopy was performed on a Bruker 400 MHz Spectrometer. Standards of 10 to 10
0
μ
M solutions of the analytes (sodium formate, methanol, etc.) were prepared by serial dilution,
S
4
and were used to calibrate the instrument. In general, 0.1
μ
L of the internal standard (dimethyl
formamide, DMF) was added to a 2 mL aliquot of the standard s
olution. 0.5 mL of this solution
was then transferred to a NMR tube that contained 200
μ
L of deuterated water. A water
suppression method was used to suppress the signal of the water in the electrolyte and to allow
visualization of the analyte peaks. The s
ame procedure was used to quantify the liquid CO
2
reduction products, with 0.1
μ
L of DMF added to 2 mL of the electrolyte after electrolysis, and
0.5 mL of this solution added in an NMR tube to 200
μ
L of D
2
O.
Surface Analysis.
Scanning
-
electron micrograp
hs were obtained using a Nova NanoSEM 450 microscope (FEI,
Hillsboro, OR, USA) with an accelerating voltage of 20 kV and a working distance of 5.0 mm.
X
-
ray photoelectron spectroscopy (XPS) data were obtained using an AXIS Ultra DLD
instrument (Kratos Anal
ytical) at a background pressure of 1
×
10
−
9
Torr. High
-
intensity
excitation was provided by monochromatic Al K
α
X
-
rays having an energy of 1486.6 eV with an
instrumental resolution of 0.2 eV full width at half
-
maximum. Photoelectrons were collected at
0°
from the surface normal at a retarding (pass) energy of 80 eV for the survey scans, whereas a
pass energy of 20 eV was used for the high
-
resolution scans. The peak energies were calibrated
against the binding energy,
BE
, of the adventitious C 1s peak. For
quantitative analysis, the XPS
signals were fitted using CasaXPS software (CASA Ltd., Teignmouth, United Kingdom) to
symmetric Voigt line shapes that were composed of Gaussian (70%) and Lorentzian (30%)
functions that employed a Shirley background.
Electrochemical surface area measureme
nts were
preformed and electrodes were found to be only slightly roughened.
S
5
Tables and Figures.
Figure S
1
.
SEM of Ni
x
Ga
y
films after annealing sho
wing the aggregation of microparticles into a
cracked film, (right) zoomed out, (left) zoomed in.
Table S
1
.
Tabulated XRD reflections seen for each nickel
-
gallium phase.
NiGa
Ni
5
Ga
3
Ni
3
Ga
2 Theta
Reflection
2 Theta
Reflection
2
Theta
Reflection
44.4
110
43.2
221
43.5
111
64.6
200
48.4
131
50.9
200
76.7
quartz
50.9
220 (Ni
3
Ga)
74.8
220
81.7
211
54.5
440
76.7
quartz
94.1
quartz
75.2
440
90.8
311
98
220
76.7
quartz
94.1
quartz
86.5
223
96.1
222
94.1
quartz
50
μ
m
1
μ
m
S
6
Figure S
2
. XP spectra showing (a) the Ni 2
p
and (b) Ga 3
p
regions. In the Ni spectra the peaks
at ~852 eV and ~870 eV correspond to Ni in the nickel
-
gallium films, while the peaks at ~856
eV and ~874 eV correspond to oxidized Ni.
Other peaks in the Ni spectra are satellite peaks. In
the Ga 3p region the peaks at ~
106 eV and ~109 eV correspond
to oxidized Ga while the small
peak at ~104 eV corresponds to metallic Ga.
S
7
Table S
2
. Comparison of the stoichiom
etric ratios of Ga
2
p
and Ni 2p from XPS data to the
target values.
Phase
Ga
2
p
Ni 2
p
Observed
Ni:Ga
Target Ni:Ga
NiGa
51
49
0.95
1
Ni
3
Ga
28
72
2.53
3
Ni
5
Ga
3
43
57
1.35
1.667
Figure S
3
.
Diagram of the cell design employed in this study.
S
8
Figure S
4
.
Electrochemical surface area measurements showing a slightly roughened film
by
lower (a) capacitative current on carbon substrate compared to (b) capacitative cu
rrent on Ni
3
Ga
.
S
9
Figure S
5
.
Potential
-
dependent Faradaic efficiencies (solid lines) and current densities (dotted
lines) for CO
2
reduction in 0.1M Na
2
CO
3
(aq)
acidified to pH 6.8 with 1 atm CO
2
(g)
to
methane
(triangles),
ethylene (squares) and ethane (exes). (a) NiGa and (b) Ni
3
Ga show similar behavior
to Ni
5
Ga
3
(See
Figure 2a
) with respect to product distributions at different potentials.
S
10
Figure S
6
. Representative example of a gas chromatog
ram for the products of CO
2
reduction at a
nickel
-
gallium film. The example shows specifically results from electrolysis with Ni
5
Ga
3
at
−
0.68
V vs. RHE in
0.1M Na
2
CO
3
(aq) acidified to pH 6.8 with 1 atm CO
2
(g)
where 150
C was
passed. Given the poor separation between
the
C
2
H
4
and CO
2
peaks at these concentrations, GC
-
MS rather than GC was used to quantify the F.E. for C
2
H
4
. Inset shows the full
-
scale
chromatogram. Relative peak heights for O
2
, N
2
and CO
2
are represen
tative for both control
experiments and bulk electrolyses.
S
11
Figure S
7
.
Histograms showing (a) Faradaic efficiency and (b) partial current density towards
hydrocarbon products for the nickel
-
gallium films under CO
2
or under CO.