Pulsed Electrochemical Carbon Monoxide Reduction on Oxide-Derived Copper Catalyst

Article abstract of DOI:10.1002/cssc.202000464

Efficient electroreduction of carbon dioxide has been a widely pursued goal as a sustainable method to produce value-added chemicals while mitigating greenhouse gas emissions. Processes have been demonstrated for the electroreduction of CO₂ to CO at nearly 100 % faradaic efficiency, and as a consequence, there has been growing interest in the further electroreduction of carbon monoxide. Oxide-derived copper catalysts have promising performance for the reduction of CO to hydrocarbons but have still been unable to achieve high selectivity to individual products. A pulsed-bias technique is one strategy for tuning electrochemical selectivity without changing the catalyst. Herein a pulsed-bias electroreduction of CO was investigated on oxide-derived copper catalyst. Increased selectivity for single-carbon products (i.e., formate and methane) was achieved for higher pulse frequencies (<1 s pulse times), as well as an increase in the fraction of charge directed to CO reduction rather than hydrogen evolution.

Full text of DOI:10.1002/cssc.202000464

Chemistry–Sustainability–Energy–Materials
Accepted Article
Title: Pulsed electrochemical carbon monoxide reduction on oxide-
derived copper catalyst
Authors: Joshua M. Spurgeon, Jacob M. Strain, Saumya Gulati, and
Sahar Pishgar
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.202000464
Link to VoR: https://doi.org/10.1002/cssc.202000464
0
1/2020

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
Pulsed electrochemical carbon monoxide reduction on oxide-
derived copper catalyst
Jacob M. Strain, Saumya Gulati, Sahar Pishgar,[a] and Joshua M. Spurgeon*[a]
[
a]
[a]
[a]
J. M. Strain, S. Gulati, S. Pishgar, Dr. J. M. Spurgeon
Conn Center for Renewable Energy Research
University of Louisville
216 Eastern Parkway, Louisville, KY, 40292 (USA)
*
E-mail: joshua.spurgeon@louisville.edu
Supporting information for this article is given via a link at the end of the document
Abstract: Efficient electroreduction of carbon dioxide has
been a widely pursued goal as a sustainable method to
produce value-added chemicals while mitigating greenhouse
gas emissions. Processes have been demonstrated for the
able to achieve near 100% faradaic efficiency (FE) from
CO
at high current density.8
Thus, a promising approach to mitigate the limitations of
the CO RR scaling relations is the further direct
2
2
electroreduction of CO
2
to CO at nearly 100% faradaic
electrochemical reduction of CO in a second electrocatalytic
system. Cascade electrolysis systems have been
efficiency, and as a consequence, there has been growing
interest in the further electroreduction of carbon monoxide.
Oxide-derived copper catalysts have promising performance
for the reduction of CO to hydrocarbons but have still been
unable to achieve high selectivity to individual products. A
pulsed-bias technique is one strategy for tuning
electrochemical selectivity without changing the catalyst.
Herein a pulsed-bias electroreduction of CO was investigated
on oxide-derived copper catalyst. Increased selectivity for
single-carbon products (i.e., formate and methane) was
achieved for higher pulse frequencies (< 1 s pulse times), as
well as an increase in the fraction of charge directed to CO
reduction rather than hydrogen evolution.
2
demonstrated by first reducing CO to CO with Au or Ag,
then subsequently flowing the CO to a second electrolyzer
or catalyst for a more selective secondary reduction.9, 10
With OD-Cu, for instance, the grain boundary active sites
promote adjacent CO binding and subsequent C-C bond
formation, leading to higher FE for C2 products like ethanol
and acetate compared to conventional Cu for the
electrochemical CO reduction reaction (CORR).11 This
approach has motivated significant recent study of CO
reduction,12, 13 which has yielded advancements like > 90%
FE for CORR at a partial current density of over 600 mA
-
2
cm , enabled by CO gas flow-through electrolyzer
designs.14
Beyond modifications to the catalyst, the reaction
branching pathway can also be tuned by affecting the
dynamics at the electrochemical interface by pulsing the
applied bias to the electrolyzer. This approach manipulates
a complex interplay between the pulsing frequency and
vertex potentials with the balance between reaction kinetics
and mass transfer as well as the charging and discharging
of the electrode double-layer capacitance.15 For potential
Introduction
2
Developing viable routes to convert waste CO to value-
added chemicals is receiving increasing interest as a
strategy to decrease greenhouse gas emissions.1 Several
technoeconomic analyses have indicated that with sufficient
performance improvements, several products could be
commercially produced through electrochemical CO
2
pulses in the multi-second time range, improved CO mass
2
reduction.2, 3 Copper is one of the few catalysts that has
shown high activity for producing hydrocarbons and
oxygenates through the carbon dioxide reduction reaction
transport to the electrode during the rest pulse enabled
1
6-18
modestly enhanced FE for CH and C products.
4
2
In the
sub-200 ms pulse time range, the charging/discharging of
(
CO
2
RR).4 Oxide-derived copper (OD-Cu), wherein an
the electrode capacitance favored CO RR intermediate
2
oxidized Cu surface is subsequently electrochemically
reduced, possesses a high electrochemically active surface
area with a high density of grain boundaries, which has led
to increased activity compared to planar Cu.5, 6 However,
desorption and resulted in tunable and selective syngas
1
5
(H +CO) formation. Pulsed-bias CO RR activity was also
2
2
shown to depend strongly on the electrode geometric
surface area, with the differences attributed to induction
time and the rate of double-layer charging.19 Pulsed-bias
electrolysis has further been demonstrated to have the
added benefit of mitigating catalyst poisoning from
contaminants or byproducts by repeated reduction/oxidation
of the electrode interface to release or convert the
undesired adsorbed species.20, 21 Despite the increasing
2
designing catalysts for very high CO RR selectivity to a
single product is hindered by the scaling relations which
limit the degree to which the binding energy can be
simultaneously optimized for the various intermediate
reactions in a multi-step reaction.7 For a simple single-
carbon product like CO, several other catalysts have been
1
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
interest in electrochemical CO reduction, however, few if
any reports exist on the ability of pulsed-bias conditions to
because other groups have reported that it may arise via a non-
electrochemical route in CORR, perhaps as
a hydration
product.11 Some formate is clearly generated through a non-
electrochemical route, as it has been demonstrated to form
under CORR conditions at open-circuit voltage (OCV) with a Cu
electrode.12 We performed additional control experiments to
direct the selectivity of CORR.
Herein, we have
investigated the effects of variable electrolysis parameters
on the product distribution for reducing CO with a highly
nanostructured OD-Cu catalyst surface.
-
further elucidate the behavior of HCOO production on OD-Cu
for CORR. Figure S2 shows the total formate production rate on
OD-Cu at relevant potentials, including where no faradaic
current passed, as well as the formate produced in the absence
Results and Discussion
-
of an electrode. Only 0.09 µmol of HCOO was detected without
First the behavior of OD-Cu for CORR under potentiostatic
conditions without pulsed bias was established (see
Experimental). After in situ reduction of the OD-Cu, current
density vs. potential (J-E) behavior was measured to confirm the
near absence of the peak corresponding to copper oxide
reduction (Fig. S1). Fig. 1 shows the CORR product FEs and
overall current density as determined at the end of a 60 min
electrolysis. The results are the average of three redundant
measurements at each potential. The product FE values (Table
S1) are consistent with other reports for potentiostatic CORR on
the OD-Cu catalyst, which increased to 0.95 µmol for OD-Cu at
0
V vs. RHE and reached as high as 3.99 µmol at -0.35 V vs.
RHE.
While some formate is likely produced through a non-
electrochemical CO hydration, this route alone cannot
explain the observed potential-dependence of HCOO-
concentration. Instead, it is proposed that electrochemically
-
generated OH from alkaline hydrogen evolution at higher
localized concentration in the vicinity of surface-adsorbed
OD-Cu.12, 22 Minimum H
production was observed at -0.30 V vs. CO on the electrode accelerates the rate of CO hydration to
2
RHE, with a FE of 44.1%. Peak C2+ product formation occurred
at -0.35 V vs. RHE, with FEs of 19.7% and 4.6% for ethanol and
acetate, respectively, as well as trace propanol with a FE of
formate (see SI). Faradaic efficiency for formate by an
indirect electrochemical route was therefore determined by
-
subtracting the baseline molar production of HCOO in the
2
0.35%. At -0.40 V vs. RHE, the gaseous C products ethylene
0
.1 M KOH electrolyte with an OD-Cu electrode at open-
and ethane were detected at a combined FE of 8.0%. At -0.25
and -0.30 V vs. RHE, the total FE to measured products was
significantly below 100%, and we attribute the balance to
continued restructuring and reduction of the electrode as well as
unquantified acetaldehyde. Acetaldehyde, which is known to be
an intermediate product on the way to ethanol formation,22 is
highly volatile. Although the low concentration makes it difficult
to quantify by GC with the dynamic flow mode used herein, a
circuit from the value measured at a given applied bias.
Variation in the HCOO- production under pulsed-bias
conditions was even more pronounced, further indicating
that a purely non-electrochemical route was not sufficient to
explain the generation of formate. The formate production
route, along with the proposed reaction pathways and
cathodic half-reactions for the other products for alkaline
CORR on OD-Cu are shown schematically in Figure 2.
For the pulsed-bias CORR experiments, the main
parameter investigated for its effect on the product
distribution was the pulse frequency. A square-wave pulse
static-headspace method has permitted quantification by
others.22 It was recently reported that acetaldehyde is a major
product (~60% FE) at low overpotentials on nanoflower Cu,12
making this a likely source for much of the missing charge in Fig.
1
.
Furthermore, a moderate drop in current density occurred
between a cathodic potential, E
rest potential, E , at 0 V vs. RHE was used consistently
throughout the study (Fig. 3a). The E value was selected
c
, at -0.35 V vs. RHE and a
during the 60 min electrolysis (Fig. S1), most likely due to a
partial loss of the Cu wire mesh-like nanostructures over time as
has been seen in other OD-Cu studies.5, 11
r
c
as the potential for maximum CORR FE, including to C
2
-
A small amount of formate, HCOO , (< 1.7% FE) is
products (Fig. 1). At the E potential, no steady-state
r
reported in Fig. 1 as an electrochemical CORR product on OD-
Cu. The presence of formate as a product is noteworthy
faradaic current passes, which allows time for CO diffusion
without promoting the local oxidation of the CORR products
that might occur at more anodic potentials. Pulsed-bias CO
measurements applied the potential square wave for 60
Hydrogen
Formate
Ethylene
Ethanol
Methane
Ethane
Acetate
1
00
8
6
4
2
0
min, of which only half that time was spent at the active E
c
8
6
4
2
0
0
0
0
0
reductive condition. After switching potentials, a transient
current decayed over ~1 s before reaching a pseudo-
steady-state current within the pulse. An example pulsed-
bias current density vs. time profile is shown in Fig. 3b.
Within each pulse, the current is a combination of non-
faradaic capacitive charging current and the faradaic
currents directing charge to CORR, the hydrogen evolution
reaction (HER), and Cu reduction/oxidation. It is difficult to
determine an unequivocal and quantifiable deconvolution of
the faradaic and non-faradaic components of the charge
passed. Although double-layer capacitive charging can be
fit with an RC circuit exponential fit, the underlying baseline
faradaic current is likely not constant and thus difficult to
-
0.40
-0.35
-0.30
-0.25
Potential, E (V vs. RHE)
Fig. 1. Product distribution for CORR on OD-Cu in CO-saturated aqueous 0.1
M KOH. Overall current density (diamonds) is the value after 60 min. Values
are the average of three measurements at each potential.
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
Fig. 2. Proposed reaction pathways for CORR on OD-Cu in 0.1 M KOH to the observed products. Cu grain boundaries promote adjacent CO binding and
22
subsequent C-C bond formation. Acetaldehyde is a readily desorbed intermediate species on the reduction pathway to ethanol. Acetate formation is believed to
11
proceed through hydroxide attack of a surface-bound ketene.
times deviated appreciably from the values measured for
continuous electrolysis (Fig. 4a). In particular, a lower
fraction of ethanol and acetate was observed for long pulse
times relative to the no-pulse condition.
experiment was therefore conducted with
A
control
a
30 min
continuous electrolysis followed by 30 min at OCV with
continued CO bubbling (i.e., a 1800 s pulse). The 1800 s
pulse product distribution was consistent with the values
measured for 50 s pulse times and also had notably lower
fractions of ethanol and acetate compared to the no-pulse
condition. We therefore attribute the reduced measured
Fig. 3. (a) Pulsed-bias waveform alternating the OD-Cu potential between the
cathodic potential, E , at -0.35 V vs. RHE and a rest potential, E , at 0 V vs.
c
r
RHE. (b) Example pulsed-bias current density response for 1 s pulse times.
accurately quantify. Thus, rather than calculate FEs in
pulsed-bias conditions, the selectivity in these experiments
is reported as the product distribution by charge. Moreover,
OD-Cu morphology after pulsing was investigated by SEM
and capacitance measurements (Figs. S3-6, Table S2).
With no clear trend in surface roughness observed with
pulsing times, we attribute the variations in product
distribution to the dynamic interfacial electrochemistry
rather than changes in the catalyst.
fraction of these liquid C
during the rest periods of the pulsed electrolysis. Gaseous
CH and H products, however, were observed in similar
2
products to changes occurring
4
2
amounts. The discrepancy between the observed liquid
products for long pulses and non-pulsed electrolysis is likely
attributable to the volatility of ethanol and acetate, leading
to the gradual vaporization of these products into the output
CO stream at concentrations below the GC detection limit in
the continuous dynamic flow mode used herein. Because
the NMR liquid product quantification measures the
accumulated product at the end of the experiment, the time
spent under active CO bubbling at the rest potential for
pulsed-bias conditions reduces the observed fraction of
volatile liquid products relative to a continuous electrolysis.
The reported product distribution by charge for ethanol and
acetate under pulsed-bias is therefore a lower limit, and the
Pulsing times varied from 10 ms to 50 s, covering the
wide range of pulse frequencies reported in previous
pulsed-bias studies on Cu for CO
2
RR.15-21 Figure 4 shows
the resulting product distributions measured at each pulse
time along with the total measured charge required to
produce all the detected products (Table S3). Each of the
values are the average of three individual measurements.
The total measured charge was determined by calculating
the required charge to produce all the measured products
2
actual fraction of C species produced is assumed to be
higher. However, precisely accounting for the volatile
product with a static-headspace method is challenging for
CORR due to the low solubility of CO, which makes steady-
state reduction difficult in the absence of active CO
bubbling.
2
from CO and H O reactants.
With increasingly longer pulses, the product distribution
would be predicted to asymptote towards the values
observed for non-pulsed continuous electrolysis as steady-
state conditions increasingly dominate each pulse.15
However, the observed CORR products for 50 s pulse
For non-pulsed CORR electrolysis at -0.35 V vs. RHE,
2
the faradaic efficiency for H was 68.2%, and the total FE
3
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
Fig. 4. (a) Product distribution vs. pulsing time along with the corresponding total measured charge (white diamonds, right axis) attributed to the detected
products, and (b) the fraction of CORR products which are either one- or two-carbon species. (c) The total charge attributable to each detected product. The “PS”
condition refers to a potentiostatic measurement at -0.35 V vs. RHE for 60 min.
for CORR was 28.8% (Fig. 1, Table S1). Accounting for
faradaic current to undetected products, this is equivalent to
a product distribution by charge of 70.3% for HER and
Besides changing the relative rates of HER to CORR,
pulsing frequency also had an effect on the charge
distribution of carbonaceous products. For pulsed-bias
electrolysis, a general trend was observed in which shorter
pulses corresponded to a lower charge fraction of the
2
9.7% for CORR (Fig. 4a, Table S2). Vaporization of
volatile liquid products during the rest period accounted for
a decrease in the charge fraction of CORR, with a product
distribution by charge for CORR of 13% for the 1800 s
pulse time, and corresponding increase in HER to 87%.
The HER charge fraction increased to as high as 92.4% at
2
CORR products being directed towards C species (Fig.
4b). Only minor variation was observed in the CORR
product distribution for pulse times > 5 s, which can be
c
attributed to the majority of the active E period being in
2
2
5 s pulses, then decreased with decreasing pulse times to
9.3% for 10 ms pulses (Fig. 4a, Table S2). CORR became
pseudo-steady-state for pulses of this duration. For pulses
≤ 1 s, however, more significant variation in the CORR
product distribution occurred (Fig. 4a). The charge fraction
the dominant reaction over HER at 100 ms and shorter
pulses, accounting for ~70% of the charge, although the
total charge was lower in this pulse frequency range. While
the partial charge for CORR products increased from 4.1 C
at 25 s pulses to as high as 8.8 C at 100 ms pulses, the
partial charge for HER decreased from 50.1 C to 3.6 C
between the same conditions, respectively (Fig. 4c). Thus,
the decrease in the total measured charge at shorter pulses
can primarily be attributed to a decrease in the rate of HER.
Interestingly, this finding is the reverse trend of the results
2
of C products was notably lower in the sub-second pulse
range, reaching as low as 3.0% of the CORR products for
10 ms pulse times (Fig. 4b). While some pulsed-bias
CO
selectivity toward C
range,16-18 pulsing at 100 ms and below decreased C
2
RR studies have reported notable enhancements in
2
products at pulses in the 1 – 5 s
2
product formation.15 The short pulse time CORR results
herein are consistent with previous reports in which the
non-faradaic component is dominant at sub-second pulse
times and the continuously changing interfacial energetics
affect species adsorption and prevent the pseudo-steady-
state needed to promote adjacent adsorbed CO* species
observed in our previous pulsed-bias work on CO
Cu in neutral aqueous KHCO , in which the rate of H
formation increased with shorter millisecond pulses.15 In
that work, an increase in HER relative to CO RR at short
2
RR on
3
2
2
capable of C-C bond formation to C
Furthermore, with the decreased HER and
production at sub-second pulse times, the production of C
2
products.
pulse times was attributed to the constantly changing
electric field due to double layer charging/discharging
C
2
1
leading to desorption of the CO
2
reduction intermediates
methane and formate species increased as a result. The
partial charge for C products varied from 2.6 C (0.4 C to
and subsequent promotion of HER. In the present case, CO
1
binding and reduction is instead competing with alkaline
formate, 2.2 C to methane) at 25 s pulses to 8.2 C (3.9 C to
formate, 4.3 C to methane) at 0.1 s pulses (Fig. 4c, Fig.
S7). If the strong component of non-faradaic current
throughout the short pulse times inhibits HER and C-C
bond formation, reduction of the single adsorbed CO
becomes the preferred route. Full reduction and protonation
HER, in which water-splitting occurs through
a
OH-
+
intermediate rather than a H . We thus speculate that the
non-faradaic current from charging/discharging which
dominates short pulses in the ms regime could have a
different effect on the binding energy and species
adsorption which acts to inhibit HER in this situation. In
addition, a less reductive potential (-0.35 V vs. RHE) was
used for CORR in the present work compared to the
of the CO leads to CH
4
.23 As mentioned above, some
formate is produced through
a
non-electrochemical
hydration step, but the potential dependence of the formate
concentration indicates an electrochemical pathway exists
as well (Fig. S2). Increased formate production at short
pulses is consistent with the favored reduction of a single
2
potential (-1.0 V vs. RHE) applied for CO RR in the
previous study, leaving less driving force for HER to
overcome competing effects in the present study.15
-
adsorbed CO, as near-surface OH reacts with the CO to
4
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
make formate (see SI). At longer active cathodic pulse
times, these reaction rates compete with diffusion of CO
from the bulk to the electrode, permitting greater HER
relative to CORR.
500 °C, and then a 10 h ramp down to room temperature.
x
The CuO foil was then placed in the electrochemical cell
with ~ 3 cm2 exposed to the 0.1 M KOH electrolyte and
reduced at -0.5 V vs. RHE for 45 - 60 min until the current
density reached a steady-state value for 5 min at ≤ 5 mA
-
2
cm . Previous XPS and XRD characterization has
Conclusions
demonstrated a surface of Cu(II) present as Cu O after
2
thermal oxidation, converting to primarily Cu(0) after the
electrochemical reduction step, with the return of a thin
oxide layer and Cu(II) and Cu(I) due to residual anodic
surface oxidation during pulsed-bias operation.5, 15
Herein, a pulsed-bias technique on a nanostructured Cu
surface was investigated using the pulse time as a parameter
for affecting the CORR selectivity. The 50% duty cycle used
for pulsed-bias conditions resulted in appreciable loss of the
volatile C
2
liquid products to the gas phase relative to a
Electrochemical Measurements
continuous process. Using pulses from 0.01 – 50 s, the
selectivity to CORR was significantly enhanced at sub-second
pulse times mainly due to a decrease in the rate of the
competing HER, which is a different behavior than previously
Electrochemical CO reduction experiments were performed
with a BioLogic SP-200 potentiostat. The electrochemical
cell was made of polycarbonate plates and set up for a
three-electrode experiment with the OD-Cu foil as the
working electrode, Pt mesh as the counter electrode
separated from the cathode compartment by the
membrane, and a Ag/AgCl (CH Instruments, Inc.) in 3.0 M
KCl as the reference electrode.15 Potentials were calculated
from the equation VRHE = VAg/AgCl + 0.210 + 0.059*pH. The
pH of the 0.1 M KOH electrolyte was 13. Following
electrochemical reduction of the OD-Cu, the catholyte was
flushed and replaced with fresh 0.1 M KOH with a catholyte
volume of 7.5 mL. Prior to CORR measurements, the
electrolyte in both the cathode and anode chambers was
bubbled with CO (99.99%, Specialty Gases) at 10 sccm for
20 min and potentiostatic electrochemical impedance
spectroscopy (PEIS) was then used to measure the
uncompensated cell resistance. Typical resistances for the
cell ranged from 15 – 28 Ω. The potentiostat was then set to
compensate for 85% of the uncompensated resistance
during the electrochemical CORR experiment, which was
conducted with a steady CO flow rate of 10 sccm through a
bubbler with 4 - 5 µm pores in the catholyte at the base of
observed for CO
Additionally, the shorter pulses greatly enhanced the
direction of charge to C relative to C products, reaching a
2
reduction in neutral electrolyte.
1
2
ratio as high as 97:3 at 10 ms pulse times. Consequently, sub-
second pulse conditions had significantly higher fractions of
the total charge resulting in the production of methane and
formate. Further study with variable duty cycle and applied
c
potential E during the cathodic pulse may yield additional
conditions for controlling the selectivity of CORR on Cu.
Pulsed-bias studies such as these are a promising way to
achieve an additional systems-level control over the
electroreduction reaction as a way to complement the design
of catalyst materials for high selectivity.
Experimental Section
Membranes and Chemicals
All reagents were used as received, except for KOH
(
reagent grade, Amresco) which was purified using the K
form of Chelex 100 (received as the Na form, Sigma-
Aldrich). All electrolyte solutions were prepared in 18 MΩ-
cm water. The anion exchange membrane used in the
electrochemical cell was 100 µm thick Selemion (AGC
Engineering Co., Ltd.), pretreated by soaking in a bath of 1
M KOH for over 24 hours to ensure there were negligible
membrane contaminants in the electrolyte.24
the OD-Cu electrode.
Electrochemical surface area
(ECSA) measurements were performed to obtain the
double-layer capacitance values of the OD-Cu electrodes
before and after pulsed-bias electrolysis.25
A cyclic
voltammogram was produced at potentials without faradaic
current (0.02 to 0.26 V vs. RHE) and the non-faradaic
current was plotted versus the scan rate to calculate the
double-layer capacitance. The scan rates were 10, 20, 40,
-
1
Electrode Fabrication
60, and 80 mV s . The roughness factor was estimated by
dividing the calculated OD-Cu electrode capacitance by the
approximate capacitance of atomically smooth Cu (29 µF).6
Oxide-derived copper (OD-Cu) electrodes were prepared
from Cu foil (0.127 mm thick, 99.9%, Alfa Aesar) following a
published method.5, 11 The Cu foil was sonicated for 30 min
in acetone and isopropanol for cleaning and then
electropolished in 85% phosphoric acid (Macron Fine
Chemicals) at 2.0 V vs. a secondary Cu foil for 5 min under
Product Analysis
CO reduction products were measured by gas
chromatography (GC, SRI 8610) and nuclear magnetic
resonance (NMR) spectroscopy for the gas and liquid
products, respectively. Both instruments were calibrated
with standard gases or liquid solutions. The concentration
of gaseous products was measured at 15, 30, 45, and 60
min. The liquid products were collected at 0 min and 60
min for each electrolysis experiment. The moles of each
vigorous stirring.
electropolished. The foil was then rinsed in 18 MΩ-cm
water and dried under nitrogen. The Cu foil was
Both sides of the foil were
subsequently placed in a muffle furnace under ambient air
for thermal oxidation with a 1.5 h ramp up to 500 °C, 12 h at
5
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
[
9]
[10]
11]
N. Theaker, J. Strain, B. Kumar, J. P. Brian, S. Kumari and J. M.
Spurgeon, Electrochim. Acta, 2018, 274, 1-8.
liquid product were determined by taking the difference
between the final and initial concentrations. A 400 µL
Gurudayal, D. Perone, S. Malani, Y. Lum, S. Haussener and J. W.
Ager, ACS Appl. Energy Mater., 2019, 2, 4551-4559.
C. W. Li, J. Ciston and M. W. Kanan, Nature, 2014, 508, 504-507.
L. Wang, S. Nitopi, A. B. Wong, J. L. Snider, A. C. Nielander, C. G.
Morales-Guio, M. Orazov, D. C. Higgins, C. Hahn and T. F.
Jaramillo, Nature Catalysis, 2019, 2, 702-708.
2
aliquot of the catholyte was mixed with 100 µL of D O that
[
contained ~100 ppm dimethyl sulfoxide (DMSO, ACS
grade, Amresco). The peak area of the DMSO internal
[12]
standard was compared to the peak area of each liquid
product in 1H-NMR using a Varian 700 MHz cold-probe
[13]
H. Zhang, J. Li, M.-J. Cheng and Q. Lu, ACS Catal., 2019, 9, 49-
65.
[14]
M. Jouny, W. Luc and F. Jiao, Nature Catalysis, 2018, 1, 748-755.
B. Kumar, J. P. Brian, V. Atla, S. Kumari, K. A. Bertram, R. T.
White and J. M. Spurgeon, ACS Catal., 2016, 6, 4739-4745.
J. Yano, T. Morita, K. Shimano, Y. Nagami and S. Yamasaki, J.
Solid State Electrochem., 2007, 11, 554-557.
while using a water peak suppression technique. For non-
pulsed experiments, the faradaic efficiency was determined
directly by comparing the charge required to produce the
measured products to the total charge passed as measured
by the potentiostat. Because of the constantly shifting
convolution of faradaic and transient non-faradaic (i.e.,
double-layer capacitive charging) currents during pulsed-
bias conditions, it is not straightforward to calculate faradaic
efficiency for pulsed-bias experiments. Instead, for pulsed-
bias experiments the selectivity was based on the product
distribution by charge, which was calculated by comparing
the charge required to produce a given product relative to
the charge required to produce all the measured products.
Thus, the product distribution by charge is the same as
faradaic efficiency except that it does not account for
charge passed which did not result in a detected product.
[15]
[16]
[17]
[18]
[19]
[20]
[21]
J. Yano and S. Yamasaki, J. Appl. Electrochem., 2008, 38, 1721-
1
726.
S. Ishimaru, R. Shiratsuchi and G. Nogami, J. Electrochem. Soc.,
000, 147, 1864-1867.
2
K. W. Kimura, K. E. Fritz, J. Kim, J. Suntivich, H. D. Abruna and T.
Hanrath, Chemsuschem, 2018, 11, 1781-1786.
G. Nogami, H. Itagaki and R. Shiratsuchi, J. Electrochem. Soc.,
1994, 141, 1138-1142.
A. Engelbrecht, C. Uhlig, O. Stark, M. Hammerle, G. Schmid, E.
Magori, K. Wiesner-Fleischer, M. Fleischer and R. Moos, J.
Electrochem. Soc., 2018, 165, J3059-J3068.
E. Bertheussen, A. Verdaguer-Casadevall, D. Ravasio, J. H.
Montoya, D. B. Trimarco, C. Roy, S. Meier, J. Wendland, J. K.
Norskov, I. E. L. Stephens and I. Chorkendorff, Angew. Chem.-Int.
Edit., 2016, 55, 1450-1454.
[22]
[23]
A. A. Peterson and J. K. Norskov, J. Phys. Chem. Lett. 2012, 3, 2,
251-258.
[
24]
J. M. Strain and J. M. Spurgeon, Journal of Co2 Utilization, 2020,
35, 298-302.
C. C. L. McCrory, S. Jung, I. M. Ferrer, S. M. Chatman, J. C.
Peters and T. F. Jaramillo, J. Am. Chem. Soc., 2015, 137, 4347-
4357.
[25]
Catalyst Characterization
A scanning electron microscope (FEI NOVA nano-SEM
6
00) was used to image the surface morphology of the Cu
foil. A Bruker D8 powder X-ray diffraction (XRD) system
was used for crystal structure and phase analysis using
non-monochromated Cu-Kα radiation produced by an X-ray
tube operated at 40 kV and 40 mA. The sample XRD
patterns were scanned between 30-80° at a scan speed of
4
s per step with a step size of 0.02°.
Acknowledgements
The authors acknowledge support from the Conn Center for
Renewable Energy Research at the University of Louisville.
Keywords: carbon monoxide • reduction • pulsed bias •
catalysis • electrochemistry
Notes and references
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
O. Bushuyev, P. De Luna, C. T. Dinh, L. Tao, G. Saur, J. van de
Lagemaat, S. O. Kelly and E. H. Sargent, Joule, 2018, 2, 825-832.
J. M. Spurgeon and B. Kumar, Energy Environ. Sci., 2018, 11,
1
536-1551.
M. Jouny, W. Luc and F. Jiao, Ind. Eng. Chem. Res., 2018, 57,
165-2177.
2
K. P. Kuhl, E. R. Cave, D. N. Abram and T. F. Jaramillo, Energy
Environ. Sci., 2012, 5, 7050-7059.
C. W. Li and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 7231-
7234.
Y. W. Lum, B. B. Yue, P. Lobaccaro, A. T. Bell and J. W. Ager, J.
Phys. Chem. C, 2017, 121, 14191-14203.
J. Perez-Ramirez and N. Lopez, Nature Catalysis, 2019, 2, 971-
976.
B. A. Rosen, A. Salehi-Khojin, M. R. Thorson, W. Zhu, D. T.
Whipple, P. J. A. Kenis and R. I. Masel, Science, 2011, 334, 643-
644.
6
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.202000464
FULL PAPER
Entry for the Table of Contents
Variable pulsing frequency in a pulsed-bias electrochemical CO reduction affected selectivity and favored methane and formate at
short pulse times. Pulsed-bias electrosynthesis enables some control of product selectivity that can be complementary to catalyst
2
design. This type of CO reduction can be coupled to high efficiency reduction of CO to CO for a two-step electrosynthesis of useful
products from CO
2
.
7
This article is protected by copyright. All rights reserved.