Selective Electrochemical Reduction of Carbon Dioxide to Ethylene and Ethanol on Copper(I) oxide catalysts

Article abstract of DOI:10.1021/cs502128q

The selective electroreduction of carbon dioxide to C₂ compounds (ethylene and ethanol) on copper(I) oxide films has been investigated at various electrochemical potentials. Aqueous 0.1 M KHCO₃ was used as electrolyte. A remarkable finding is that the faradic yields of ethylene and ethanol can be systematically tuned by changing the thickness of the deposited overlayers. Films 1.7-3.6 μm thick exhibited the best selectivity for these C₂ compounds at -0.99 V vs RHE, with faradic efficiencies (FE) of 34-39% for ethylene and 9-16% for ethanol. Less than 1% methane was formed. A high C₂H₄/CH₄ products' ratio of up to ～100 could be achieved. Scanning electron microscopy, X-ray diffraction, and in situ Raman spectroscopy revealed that the Cu₂O films reduced rapidly and remained as metallic Cu⁰ particles during the CO₂ reduction. The selectivity trends exhibited by the catalysts during CO₂ reduction in phosphate buffer, and KHCO₃ electrolytes suggest that an increase in local pH at the surface of the electrode is not the only factor in enhancing the formation of C₂ products. An optimized surface population of edges and steps on the catalyst is also necessary to facilitate the dissociation of CO₂ and the dimerization of the pertinent CH_xO intermediates to ethylene and ethanol.

Full text of DOI:10.1021/cs502128q

Research Article
pubs.acs.org/acscatalysis
Selective Electrochemical Reduction of Carbon Dioxide to Ethylene
and Ethanol on Copper(I) Oxide Catalysts
Dan Ren, Yilin Deng, Albertus Denny Handoko, Chung Shou Chen, Souradip Malkhandi,
and Boon Siang Yeo*
Department of Chemistry, Faculty of Science, National University of Singapore, 3 Science Drive 3, Singapore 117543
S
* Supporting Information
ABSTRACT: The selective electroreduction of carbon dioxide to
C2 compounds (ethylene and ethanol) on copper(I) oxide films has
been investigated at various electrochemical potentials. Aqueous 0.1
M KHCO3 was used as electrolyte. A remarkable finding is that the
faradic yields of ethylene and ethanol can be systematically tuned by
changing the thickness of the deposited overlayers. Films 1.7−3.6
μm thick exhibited the best selectivity for these C2 compounds at
−0.99 V vs RHE, with faradic efficiencies (FE) of 34−39% for
ethylene and 9−16% for ethanol. Less than 1% methane was
formed. A high C2H4/CH4 products’ ratio of up to ∼100 could be
achieved. Scanning electron microscopy, X-ray diffraction, and in
situ Raman spectroscopy revealed that the Cu2O films reduced rapidly and remained as metallic Cu0 particles during the CO2
reduction. The selectivity trends exhibited by the catalysts during CO2 reduction in phosphate buffer, and KHCO3 electrolytes
suggest that an increase in local pH at the surface of the electrode is not the only factor in enhancing the formation of C2
products. An optimized surface population of edges and steps on the catalyst is also necessary to facilitate the dissociation of CO2
and the dimerization of the pertinent CHxO intermediates to ethylene and ethanol.
KEYWORDS: ethylene, ethanol, selectivity, CO2 reduction, Cu2O, electrocatalysis
applied, and the morphology and surface geometry of the
copper surfaces (Table 1). On polycrystalline Cu surfaces
poised at −5 mA/cm2 in 0.1 M KHCO3, the faradic efficiencies
(FE) for the production of C2 compounds (ethylene and
ethanol) and methane are ∼37 and ∼29%, respectively.17 The
dismal selectivity can be attributed to the heterogeneity of
catalytic sites present on the polycrystalline Cu plane. This can
be ameliorated by tuning the potentials of the working
electrode, although enhancements have not been signifi-
cant.18,26 More improvements were found when CO2 reduction
was performed on single crystal Cu(100) surfaces and even
more so on cleaved Cu(100) substrates with high-indexed
planes.27 In particular, ethylene and ethanol could be produced
from the electroreduction of CO2 with a total FE of ∼57% on
the Cu(S)-[4(100) × (111)] surface. The square arrangement
of the Cu atoms in the (100) terraces and presence of atomic
steps have been proposed to favor C−C bond coupling
between the CHxO intermediates to give C2 compounds;2
however, the use of single crystal surfaces is not practical for
scaling up catalytic processes for the chemical industries.
Recently, our group discovered that copper mesocrystals
formed by the in situ reduction of a CuCl film were active
toward the reduction of CO2 to C2H4 with a C2H4/CH4
1. INTRODUCTION
The reduction of carbon dioxide (CO2) to hydrocarbons and
alcohols has the potential to generate a sustainable supply of
valuable feedstocks for our chemical industries and fuels to
meet our energy needs.1−9 Alongside with carbon sequestra-
tion, it is also an effective way to mitigate anthropogenic CO2
emissions. Among the CO2 reduction products, ethylene
(C2H4) and ethanol (C2H5OH) have higher energy densities
and commercial value than their C1 counterparts such as
methane (CH4).10−13 Ethylene and ethanol can be formed on
the surface of a catalyst via the following half-reactions.2,14,15
2CO2 + 8H2O + 12e− → C2H4 + 12OH−
2CO2 + 9H2O + 12e− → C2H5OH + 12OH−
It is believed that CO2 is first reduced through multiple
proton−electron transfers to surface-bound *CHxO (x = 0, 1,
2). These moieties then undergo intermolecular C−C bond
formation to yield *C2HxO2 (x = 0−4), which are further
reduced to ethylene and ethanol.16 To date, copper-based
materials are the most promising electrocatalysts for this
reaction, albeit still rather unselective.17−21 Considerable and
urgent efforts have therefore been devoted to tuning the
structure and composition of copper catalysts with the aim of
optimizing their CO2-to-C2 selectivity.19−25
Received: December 31, 2014
Revised: March 16, 2015
The type of products formed during CO2 electroreduction is
significantly impacted by the electrolytes used, potentials
© XXXX American Chemical Society
2814
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
a
Table 1. An Overview of Faradic Efficiencies of Products (%) Obtained from CO2 Reduction on Various Copper Catalysts
faradic efficiencies (%)
experimental
conditions
−5 mA/cm2
C2H4/
CH4
catalyst
polycrystalline Cu17
electrolyte
C2H4
30.1
C2H5OH
6.9
C2H6
N.R.
CO
2.0
CH4
29.4
formic acid
9.7
other C2+
3.0
0.1 M
KHCO3
1.0
1.1
1.3
10
polycrystalline Cu18
Cu(100)27
0.1 M
KHCO3
−1.05 V vs
RHE
−5 mA/cm2
26.0
9.8
N.R.
1.1
24.4
2.1
4.8
0.1 M
KHCO3
40.4
9.7
N.R.
0.9
30.4
3.0
7.7
Cu(S)-[4(100) × (111)]27 0.1 M
−5 mA/cm2
50.0
7.4
N.R.
1.1
5.0
4.6
14.1
N.R.
0.1−0.6
KHCO3
0.1 M
KHCO3
Cu-halide confined mesh23 3 M KX
(X = Br,
I, or Cl)
0.1 M
KHCO3
0.5 M
KHCO3
0.1 M
KClO4
Cu mesocrystals26
−0.99 V vs
RHE
27.2
N.R.
1.6−1.9
N.R.
0.55
1.8−2.8
1.47
4.3
18
−2.4 V vs
Ag/AgCl
60.5−79.5
0.8−2.8
4.3−6.6
0.1−0.7
9−17
electrodeposited Cu2O21
electrodeposited Cu2O28
Cu nanoparticles19
−1.1 V vs
RHE
12−33
26
N.R.
N.R.
N.R.
0−9
N.R.
∼1
1−3
6
0−4
1
22
N.R.
N.R.
N.R.
8−12
26
−1.82 V vs
Ag/AgCl
8
−1.1 V vs
RHE
36
34
1
N.R.
36
a
C2+ contains all other compounds with ≥2 carbon atoms. N.R.: not reported.
product ratio of 18.26 High-resolution transmission electron
microscopy revealed the presence of numerous (100) facets
and atomic steps on the Cu mesocrystals, which we assigned as
the catalytically active sites.
Struers). Cu2O layers were then galvanostatically deposited
onto these Cu discs (exposed geometric surface area: 0.385
cm2) from a copper lactate solution (Supporting Information
S1).33 Cu2O films of different thicknesses were obtained by
varying the deposition time between 1 and 60 min. Seven Cu2O
films with varying thicknesses were prepared. Electropolished
Cu surfaces were prepared by electropolishing the copper discs
at +260 mA/cm2 for 60 s in 85% phosphoric acid (Sigma-
Aldrich), followed by rinsing with ultrapure water (Type 1,
Barnstead, Thermo Scientific).17
2.2. Characterization of Copper Catalysts. The thick-
ness and surface morphology of the Cu2O films were imaged
using scanning electron microscopy (JEOL JSM-6710F, 5 keV)
(Supporting Information S1). X-ray diffraction (Bruker D8
Discover with GADDS, 40 keV, 40 mA) was used to analyze
the chemical composition of the films. The films were removed
from the Cu substrates for XRD analysis by using a sharp blade.
The XRD patterns of Cu2O and Cu0 were identified by
comparisons with their respective standards, PDF 00-003-0892
and PDF 00-001-1242.
CuI-halide-coated electrodes, CuO, electrodeposited Cu
nanoparticles and reduced Cu2O films are also catalytically
selective in reducing CO2 to C2 compounds.19,21,23,26,28,29 The
Cu2O films were formed by electrodepositing Cu ions on metal
substrates or thermal heating of Cu metal films.20,21,30 These
films were found to have nanoparticulate morphologies. Kanan
and co-workers demonstrated that reduced Cu2O films could
reduce CO to C2H5OH with a FE of 43%.24 Grain boundaries
on the surfaces of these films have been postulated to be the
driving forces for the C2 product selectivity.19,24 Their role
could be to stabilize and facilitate the dimerization of the
pertinent C1 reaction intermediates. Recently, Kas et al. showed
that C2H4 could be produced with a FE of up to 33% on thick
Cu2O films, but no C2H5OH could be detected.21,31 When the
electrolyte concentration was decreased, the C2H4/CH4 ratio
increased; thus, the selectivity is believed to be caused by a
higher local pH at the surface of the electrode.32
In situ Raman spectroscopy was performed using a confocal
Raman microscope (modular system, Horiba Jobin Yvon) in an
epi-illumination mode (top-down). A schematic diagram of the
setup is provided in Supporting Information S2. The excitation
source was a He−Ne laser with 633 nm wavelength (CVI
Melles Griot). A water immersion objective lens (LUMFL,
Olympus, 60×, numerical aperture: 1.10) covered with a 0.013
mm thin Teflon film (American Durafilm) was used for
focusing and collecting the incident and scattered laser light.34
The backscattered light was filtered by a 633 nm edge filter and
directed into a spectrograph (iHR320)/charge-coupled device
detector (Synapse CCD). The acquisition time for each
spectrum was 5 s. The electrochemical cell was based on a
round Teflon dish, with the working electrode firmly mounted
at its base.34 Counter and reference electrode were inserted into
the cell via port holes.
Inspired by these studies, we have investigated the
electrochemical reduction of CO2 on Cu2O films in aqueous
0.1 M KHCO3 electrolytes. Cu2O films of different thicknesses
were formed by galvanostatic deposition. In addition to ex situ
scanning electron microscopy and X-ray diffraction, in situ
Raman spectroscopy was used for the first time, to probe
changes in the chemical and structural composition of the
Cu2O films during the CO2 reduction. Ethylene and ethanol
were substantially formed on these films. Methane production
was practically suppressed on the thicker films. The factors
underlying the changes in selectivity are discussed in terms of
morphology of the catalysts and effects of local pH at the
surface of the electrodes.
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation. Flat Cu discs (⌀=10 mm,
99.99%, Goodfellow) served as substrates for all the catalysts.
They were first polished to a mirror-like finish using SiC paper
(grit 1200, Struers) and diamond slurries (Diapro 9 and 3 μm,
The electrochemically active surface area (ECSA) of the
catalysts was measured by the double-layer capacitance
measurement method (Supporting Information S1).30
2815
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
2.3. Electrochemical Setup, Identification and Quan-
tification of CO2 Reduction Products. A custom-built,
gastight Teflon electrochemical cell based on the design of Kuhl
et al. was used (Figure 1).18,26 The cathodic and anodic
system under equilibrium condition, only the average of the
third to the sixth GC measurements was used in the data
analysis (Supporting Information S5). The liquid products in
the catholyte were analyzed at the end of the chronoamper-
ometry using 1H nuclear magnetic resonance spectroscopy
(NMR, Avance 300, Bruker) (Supporting Information S4).18,26
The product yields were expressed as faradic efficiencies
(Supporting Information S5). Each data point is an average of
the measurements collected from at least three separate NMR
or GC experiments. Each prepared catalyst was used only once
for CO2 reduction at a chosen potential. Potentials between
−0.59 and −1.19 V were applied.
3. RESULTS AND DISCUSSIONS
3.1. Characterization of Copper Catalysts. The as-
prepared catalysts, which consist of an electropolished Cu
electrode and seven Cu2O films, were characterized by scanning
electron microscopy, X-ray diffraction, and in situ Raman
spectroscopy (Figure 2). The thicknesses of the Cu2O films
were estimated by SEM imaging of their cross sections, and
were found to be 0.2, 0.4, 0.9, 1.7, 3.6, 6.4, and 8.8 μm thick
(Supporting Information S1). Representative SEM images of
these electrodes are shown in Figure 2A−D. As the deposition
time of Cu2O lengthened, the appearance of the pristine Cu
disc changed from flat and featureless to particulate. The size of
the particles increased systematically with thickness of the
deposited Cu2O layer, with ∼100-nm-sized nanoparticles for
the thinnest 0.2 μm film and ∼2−3 μm-sized polyhedron
particles with many well-defined edges for the 8.8 μm film. X-
ray diffraction of the bulk films demonstrated that they are
Cu2O (Figure 2I).
Reduction of CO2 (4200 s) was performed using these
catalysts. At the end of the reaction, their surfaces were again
examined by SEM. Images of electrodes reduced at a
representative potential of −0.99 V are shown in Figures
2E−H. The surface of the electropolished Cu remained smooth
(Figure 2E). The reduced 0.2 μm film consisted of nano-
particles ∼100 nm in size (Figure 2F), and large polyhedron
particles were observed on the thicker films (Figures 2G,H,
Supporting Information S6). These thick films have a rough
surface morphology, as indicated by the numerous small
nanoparticles on their surface and their high surface roughness
factors (Supporting Information S1). Time-resolved ex situ
SEM images show that cracks observed on the thicker films
appear early on during the CO2 reduction process (Figure
2G,H, Supporting Information S6). These could be due to
relief of strains caused by volume changes in the films during
their reduction.36 Indeed, only Cu0 reflexions were observed in
the X-ray diffractograms of the reduced films, which
demonstrates that the bulk of the Cu2O films had reduced to
metallic Cu during the course of the CO2 reduction (Figure 2J).
We also employed in situ Raman spectroscopy to probe the
surface of the Cu2O films in real time during CO2 reduction.
Raman spectra of a representative 1.7 μm Cu2O film held at
−0.99 V are presented in Figure 2K. Cu2O, as evidenced by its
vibrational fingerprints at 147, 218, 526, and 624 cm−1, was
detected at the start of the CO2 reduction (at 0 s).37 These
peaks were attenuated after 30 s, which demonstrate the rapid
reduction of the top layers of the Cu2O film. Two bands
centered at 365 and 502 cm−1 concurrently appeared. These
can be attributed to the Cu−O vibrations of intermediately
reduced Cu oxides.37 From 200 s onward, no peaks could be
observed in the Raman spectrum. This signifies that the surface
Figure 1. Schematic diagram of the electrochemical cell used for CO2
electroreduction. WE: working electrode.
compartments, separated by an anion-exchange membrane
(Selemion AMV, AGC Asahi Glass), were filled with 10 and 8
cm3 of electrolyte separately. The volume of the headspace in
the cathodic compartment was 2 cm3. A coiled platinum wire
and Ag/AgCl (Saturated KCl, Pine) served as counter and
reference electrodes, respectively. The calibration of the
reference electrode was checked against a reversible hydrogen
electrode (RHE, HydroFlex, Gaskatel).
Aqueous 0.1 M KHCO3 (Merck, 99.7%) and phosphate
buffer (0.1 M K2HPO4, Sigma-Aldrich, >99.0% and 0.1 M
KH2PO4, Sigma-Aldrich, >99.0%) were used as electrolytes.
The latter was purified by pre-electrolysis under N2 gas for 20
h.17,35 Prior to CO2 reduction, the electrolytes were saturated
with CO2 gas (99.999%, Linde Gas) for 10 min. During the
experiment, CO2 was bubbled into the electrolyte at a flow rate
of 20 sccm. The electrolyte was stirred using a Teflon-coated
magnetic stirrer bar at 1500 rotations per minute (rpm) to
enhance mass flow of CO2 to the working electrode.
A potentiostat (Gamry Reference 600) was used for
controlling and measuring the potentials/currents. Compensa-
tion for iR drop was made using the current interrupt mode. All
the measured potentials in this work are cited with respect to
the RHE using the following conversion: ERHE (V) = EAg/AgCl
(V) + 0.197 V + (0.059 V × pH). The pH values of the
electrolytes are listed in Supporting Information S3. Unless
otherwise stated, the current density values reported in this
work were normalized to the geometric surface area.
The experimental protocol for identifying and quantifying
the products formed during CO2 electroreduction has been
reported in our previous publication.26 The headspace of the
cathodic compartment was continuously purged by the
incoming CO2 (with reduction products) into the sampling
loops of the gas chromatograph (GC, Agilent, 7890A). Gas
aliquots were analyzed after 3, 14, 26, 37, 48, and 59 min of
chronoamperometry. To ensure that the reported data is from a
2816
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
Figure 2. SEM images of Cu catalysts before CO2 reduction: (A) electropolished Cu; (B) 0.2 μm, (C) 1.7 μm, and (D) 8.8 μm Cu2O films
deposited on Cu disc. SEM images of Cu catalysts after CO2 reduction at −0.99 V: (E) electropolished Cu; (F) 0.2 μm, (G) 1.7 μm, and (H) 8.8 μm
Cu2O films deposited on Cu disc. XRD of Cu catalysts (I) before and (J) after CO2 reduction. The standard patterns (vertical line) are also included
for comparison (PDF 00-003-0892 for Cu2O and PDF 00-001-1242 for Cu). (K) In situ Raman spectra and corresponding chronoamperogram
(inset) of 1.7 μm film at −0.99 V in 0.1 M KHCO3.
of the Cu2O films has been reduced to metallic copper during
CO2 reduction. After the cathodic potential was removed, the
surface reoxidized in tens of seconds to Cu2O, as shown by the
appearance of its Raman bands at 147, 520, and 624 cm−1
(Figure 2K). Similar observations were made (Cu2O reduced to
Cu0 rapidly) when other cathodic potentials were applied
(Supporting Information S2). Because of the insufficient limits
of detection afforded by our spectrometer, no Raman signals
belonging to adsorbed CO2 or its reduced species (whose
surface population is likely to be ≤1 monolayer) could be
discerned.
On the basis of evidence from ex situ XRD and elemental
depth profiling using Auger electron spectroscopy, the oxidized
state of a Cu2O film has been proposed to be partially
conserved during CO2 reduction.28 Cu+ ions were thus
suggested to be catalytic active for reducing CO2 to C2
compounds. We have shown, however, that the surface of a
Cu2O film reduces and remains as metallic Cu particles during
electrochemical CO2 reduction. This finding is consistent with
predictions from the Pourbaix diagram of the copper-water
system.38 We thus believe that Cu0 particles are the catalytic
active species for reducing CO2. To the best of our knowledge,
this is the first in situ spectroscopic study of copper oxide films
during CO2 reduction.
Figure 3. CO2 reduction current as a function of time for four
representative Cu catalysts. The inset is a zoomed-in view of the curves
in the first 5 min. Potential applied: −0.99 V. Electrolyte: 0.1 M
KHCO3.
representative chronoamperometry curves taken at −0.99 V for
electropolished Cu as well as 0.2-, 1.7-, and 8.8-μm-thick Cu2O
catalysts. Reduction peaks were observed in all the
chronoamperometry curves of the Cu2O catalysts during the
initial phase of the CO2 reduction process. This feature, in
agreement with the XRD and Raman spectroscopy results
presented in Section 3.1, can be attributed to the reduction of
3.2. Electrochemical Reduction of CO2 to C2 Products.
CO2 reduction was performed using the Cu catalysts at fixed
potentials between −0.59 and −1.19 V. Figure 3 shows
2817
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
Figure 4. Faradic efficiencies for CO2 electroreduction products as a function of potential. (A) Electropolished Cu; (B) 0.2, (C) 1.7, and (D) 8.8 μm
Cu2O film deposited on Cu disc. Electrolyte: 0.1 M KHCO3.
a
Table 2. Faradic Efficiencies of Products (%) Obtained from CO2 Reduction on Electropolished Cu and Cu2O Catalysts
faradic efficiencies (%)
catalyst
j (mA/cm2)
CO
CH4
C2H4
C2H5OH
C2H6
HCOO−
H2
total
polished Cu
0.2 μm film
0.4 μm film
0.9 μm film
1.7 μm film
3.6 μm film
6.4 μm film
8.8 μm film
−3
−14
−20
−25
−30
−35
−30
−29
8.82
2.25
1.54
0.66
0.49
0.43
0.37
0.52
7.14
9.85
6.93
2.48
0.73
0.32
0.18
0.15
13.79
32.92
37.40
40.25
38.79
34.26
22.76
21.55
N.D.
6.00
9.50
8.66
9.01
16.37
3.71
5.10
N.D.
N.D.
N.D.
0.04
0.08
0.15
0.10
0.15
12.76
12.67
2.48
8.34
4.45
3.94
1.25
N.D.
49.78
26.21
31.01
28.60
38.98
38.87
63.80
67.67
92.29
89.90
88.86
89.03
92.53
94.34
92.17
95.14
a
Reduction carried out in 0.1 M KHCO3 at −0.99 V. N.D.: not detectable.
Cu2O to Cu0. When the steady state currents were compared,
the electrodes deposited with Cu2O films exhibited higher
current densities compared with the electropolished Cu. This
can be attributed to the formers’ larger surface roughness and,
hence, electrochemically active surface areas (Supporting
Information S1). A modest decrease in the currents of the
catalysts could be observed during the reduction reaction. This
could be due to the buildup of gas bubbles at the interface of
the electrode, which blocks available catalytic sites. An increase
in the temperature of the electrolyte during the reaction could
also result in poorer solubility of CO2 in the electrolyte and,
hence, cause a decrease in the current density.18
The faradic efficiency for each CO2 reduction product as a
function of potential is presented in Figure 4. Carbon
monoxide, methane, formate, ethylene, ethanol, and trace
amounts of ethane (0.1%) were found (Supporting Information
S5). H2 was a product of the competitive hydrogen evolution
reaction (HER).14 The faradic efficiencies of the products in all
experiments amount to 89−114%, which showed that all the
major products have been accounted for.17 The product
distribution trend exhibited by the electropolished Cu electrode
2818
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
a
Table 3. Faradic Efficiencies of Products (%) Obtained from CO2 Reduction on Electropolished Cu and Cu2O Catalysts
faradic efficiencies (%)
catalyst
j (mA/cm2)
CO
CH4
C2H4
C2H5OH
C2H6
HCOO−
H2
total
polished Cu
0.2 μm film
1.7 μm film
8.8 μm film
−14
−26
−48
−52
0.02
0.40
0.20
0.28
0.27
9.49
1.11
0.28
0.03
8.09
N.D.
7.90
5.11
2.33
0.01
0.29
0.58
0.09
5.60
8.53
2.15
1.16
95.25
61.20
69.27
97.21
101.18
95.90
16.12
7.00
94.54
108.35
a
Reduction carried out in phosphate buffer saturated with CO2 (pH = 6.73) at −0.99 V. N.D.: not detectable.
toward CO2 reduction is consistent with previous studies
(Figure 4A);18,26 however, the FE of CH4 at −1.19 V is 59%,
which is significantly higher than what was achieved in our
earlier report.26 This could be attributed to the smaller volume
(10 cm3) of catholyte used in the present electrochemical cell,
which enhances the concentration of dissolved CO2. When
CO2 reduction is performed using the Cu2O films, CH4
production dramatically decreases, especially on the thicker
films. For films thicker than 1.7 μm, the FE of CH4 is <2%,
regardless of the applied potential. The Cu2O films exhibited
generally better selectivity for CO2 reduction to C2 compounds.
We observed that the selectivity of Cu2O catalysts toward C2
compounds was optimized at −0.99 V. A detailed analysis of
the compounds formed at this potential is thus presented in
Table 2. A remarkable finding is that the FEs of ethylene and
ethanol change systematically (a parabolic trend) as a function
of thickness of the Cu2O films. The FE of ethylene and ethanol
could be tuned between 22−40%, and 4−16%, respectively. In
contrast, the FE of CH4 decreased rapidly to <1% as the
thickness of the films increased.
The 1.7−3.6-μm-thick Cu2O films exhibited optimum
selectivity toward the formation of C2 products. FEs of 34−
39% and 9−16% for ethylene and ethanol were respectively
formed using these films; the FE of methane was 0.3−0.7%.
The FE of ethanol in this work is notably high (Table 1).19,21,27
The selectivity of our Cu catalysts can also be assessed by its
C2H4/CH4 ratio, which is a useful figure of merit for assessing
the intrinsic C2 selectivity of a material.21,27,39 Here, the 3.6-
μm-thick films exhibit a C2H4/CH4 of up to ∼100, which
compares very favorably with other C2-selective catalysts,
including Cu2O films (Table 1).19,21,26,27
It is significant that C2H4 and C2H5OH were formed at
similar potential regimes, and that the FE of C2H4 is always
higher than that of C2H5OH (Figure 4 and Table 2). These
observations demonstrate that these two molecules originate
from a common reaction intermediate.18 On the basis of
density functional theory (DFT) calculations, *CH2CHO has
been proposed to be the key intermediate.40 The higher FE of
C2H4 compared with C2H5OH has been attributed to a more
energetically favorable hydrogenation of *CH2CHO to C2H4.
Another notable observation is that our reduced Cu2O films
exhibited lower overpotentials by 0.1−0.2 V for the
optimization of C2 selectivity compared with other similarly
synthesized Cu2O films.28
To ascertain if the selectivity differences result from changes
in CO2 reduction activity or H+ reduction activity, we
normalized the partial current density of the products with
the ECSA of the catalysts (Supporting Information S7). The
partial current densities of C2H4 and H2 exhibited by the 1.7−
3.6 μm films were ∼4 and ∼1.3 times larger, respectively, than
those on the electropolished Cu. The partial current density for
CH4 was substantially decreased. We thus attribute the
selectivity differences to mainly changes in the CO2 reduction
activity of the catalysts.
3.3. Effects of Local pH on C2 Products Selectivity. We
assess the effects of local pH on the selectivity of the reduced
Cu2O electrocatalysts by examining the data in Table 2. The
geometric current densities and faradic efficiency of H2
increased continuously with the thickness of the Cu2O films,
which agrees with the work by Kas et al.21 This would result in
a higher local pH at the surface of the electrode, which may
explain the observed suppression of CH4 production and
increased selectivity toward C2H4; however, the FE for C2H4
decreased on the 6.4 and 8.8 μm films. This suggests that the
effects of local pH may not be the sole factor in dictating the
selectivity; morphological factors could also contribute to the
phenomenon. Mass transport limitations of CO2 to these
electrodes that have high electrochemically active surface areas
may also result in a lower C2H4 production. It is difficult,
though, to estimate the extent of this effect in relation to CO2
reduction as the electrolyte is vigorously stirred during the
experiment.
CO2 reduction was also performed on the 0.2-, 1.7-, and 8.8-
μm-thick films in phosphate buffer electrolytes (Table 3).
Buffers mitigate increases in local pH at the surface of the
electrode during CO2 reduction and are expected to favor the
formation of CH4.17 Although the FEs for C2H4 decreases to
<16%, which agrees with previous studies that C2H4 formation
is disfavored in phosphate buffers, CH4 formation does not
increase.14,17 This trend argues against changes in local pH as
the sole factor for the enhanced selectivity of the Cu2O films
toward C2 compounds. The local structure of our catalysts is
likely to be of significant importance.
No gaseous or liquid products could be detected from the
cathodic compartment in the absence of applied potentials on
the working electrode (Supporting Information S4). Thus, all
the products observed in this work must originate from the
electrocatalytic reduction of CO2 on the Cu catalysts.
3.4. Enhanced Intermolecular C−C Coupling of C1
Moieties on Copper Surfaces To Give C2 Products. A
mechanistic model for the electrochemical reduction of CO2 to
ethylene and ethanol is shown in Figure 5. First steps involve
proton and electron transfer to give a *COOH surface moiety,
which can then hydrogenate to give H2O and adsorbed
*CO.41,42 Moderately adsorbed *CO species can further
hydrogenate to CH4, or dimerize/hydrogenate to
*C2HxO2.18,32 This intermediate can then be subsequently
reduced to either ethylene or ethanol. Quantum chemical
simulations suggest that the intermolecular C−C coupling
process is more kinetically favorable when the immediate
reactants are *CHO or *CH2O, rather than *CO.16 We have
also demonstrated in this work using SEM, XRD, and in situ
Raman spectroscopy that the catalytically active sites
responsible for the selective formation of C2 compounds are
Cu0 particles. Both local pH effects and morphological factors
2819
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
more adequately explained by the effects of local pH and
morphological factors.
4. CONCLUSION
In summary, we have investigated the electroreduction of CO2
on electrodeposited Cu2O films at different potentials. The
faradic yields for C2 products can be systematically tuned by
varying the thicknesses of the Cu2O overlayers. The 1.7−3.6-
μm-thick films exhibited the most efficient C−C bond
formation to give ethylene and ethanol with faradic efficiencies
34−39 and 9−16%, respectively. The formation of methane
was significantly suppressed (FE <1%). Thus, an unprecedent-
edly high C2H4/CH4 products’ ratio of up to ∼100 was
achieved.
Figure 5. Proposed mechanism for the electroreduction of CO2 to
ethylene and ethanol on copper surfaces. (H2O molecules are not
drawn).
Materials analysis revealed that the highly selective catalyst
film is metallic Cu in the form of 0.5−1-μm-sized polyhedron
particles. Control experiments with phosphate buffer electro-
lytes suggest that variation in local pH at the surface of the
electrode is not the only factor in influencing the product
selectivity. The higher selectivity of the films toward C2
compounds could also be attributed to their optimized density
of steps and edges, a morphology widely believed to be
essential for the formation of C2 compounds from CO2
reduction. More experimental and theoretical work is needed
to understand the influence of steps and edges and how they
impact the production of ethylene, ethanol, methane, and
hydrogen. This work demonstrates that the careful design of
the morphology of reduced Cu2O films can significantly
enhance its selectivity toward C2 compounds and edges us
closer to the engineering of an efficient, environmentally
friendly and earth-abundant CO2 reduction electrocatalyst.
are likely to play significant roles in driving the selectivity of the
CO2 reduction reaction toward C2 products.
C2 selectivity is optimized with the 1.7−3.6-μm-thick films,
which consisted of 0.5−1-μm-sized Cu polyhedrons. During
CO2 reduction, stepped surfaces with edges and terraces are
likely to be formed as the Cu2O film reduces to Cu.22,24,26
Although it is difficult to quantify their surface population in
our reduced Cu2O films, we propose that an optimum
combination of these features must be necessary to dissociate
CO2 and to optimize the chemisorption energies of the CHxO
intermediates. A critical role of the edges (with under-
coordinated Cu atoms) is to promote the buildup of a large
coverage of CHxO reactive intermediates, to facilitate their
dimerization. This is in agreement with predictions by density
functional theory calculations and with a temperature-
programmed desorption study showing that CO does adsorb
more strongly on Cu steps and edges, as compared with Cu
terraces.16,19,43 We highlight here that Hori et al. had previously
demonstrated that an optimization of step density on Cu(100)
surfaces is necessary for the enhanced formation of C2
compounds.27,39 When either an excessive number of steps or
too few steps are incorporated into a Cu(100) surface, C2
selectivity decreases.
As the thickness of the Cu2O film increases to 6.4−8.8 μm,
production of both C1 and C2 compounds decreases while
hydrogen evolution becomes more efficient (Table 2). This can
be attributed to a larger surface population of Cu atoms with
low coordination numbers (as suggested by the large roughness
factors of these films). These undercoordinated atoms are
expected to bind to atomic H more strongly as compared with
Cu atoms on a planar surface. Because Cu lies on the right-
hand side of the volcano curve for HER, an increase in its
chemisorption strength toward H should result in the
enhancement of HER compared with CO2 reduction.44,45
Similar observations have also been made on SnOx films, where
H2 evolution increased continuously as the SnOx films became
thicker.46 Mass transport limitations of CO2 to the electrode
may also lead to a more favorable catalysis of hydrogen
evolution.
ASSOCIATED CONTENT
■
S
* Supporting Information
The following file is available free of charge on the ACS
Publications website at DOI: 10.1021/cs502128q.
Experimental procedures, characterization of catalysts,
Raman spectroscopy system and data, pH of electrolytes,
NMR data, calculation of faradic efficiencies of products
and SEM figures
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +65 6516 2836. Fax: +65 6779 1691. E-mail:
chmyeos@nus.edu.sg.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work is supported by a start-up grant (R143-000-515-133)
from the National University of Singapore.
The faradic efficiencies for both methane and CO production
are remarkably suppressed to <1% on the thicker (1.7−8.8 μm)
films at −0.99 V. This could not be achieved on Cu single
crystal electrodes, Cu mesocrystals, Cu halides, etc.23,26,27 It is
also significant that CH4 suppression on a bulk Cu electrode
could not be replicated by increasing the local pH at its surface
by using nonbuffer electrolytes or lowering the concentrations
of the KHCO3 electrolyte.17,32 This suggests that the selectivity
of the films (shown in this work) toward C2 products can be
REFERENCES
■
(1) Whipple, D. T.; Kenis, P. J. A. J. Phys. Chem. Lett. 2010, 1, 3451−
3458.
(2) Gattrell, M.; Gupta, N.; Co, A. J. Electroanal. Chem. 2006, 594,
1−19.
(3) Centi, G.; Perathoner, S. Catal. Today 2009, 148, 191−205.
(4) Kuhl, K. P.; Hatsukade, T.; Cave, E. R.; Abram, D. N.; Kibsgaard,
J.; Jaramillo, T. F. J. Am. Chem. Soc. 2014, 136, 14107−14113.
(5) Lu, Q.; Rosen, J.; Jiao, F. ChemCatChem 2014, 7, 38−47.
2820
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821

ACS Catalysis
Research Article
(6) Reske, R.; Mistry, H.; Behafarid, F.; Roldan Cuenya, B.; Strasser,
P. J. Am. Chem. Soc. 2014, 136, 6978−6986.
(7) Baturina, O. A.; Lu, Q.; Padilla, M. A.; Xin, L.; Li, W.; Serov, A.;
Artyushkova, K.; Atanassov, P.; Xu, F.; Epshteyn, A.; Brintlinger, T.;
Schuette, M.; Collins, G. E. ACS Catal. 2014, 4, 3682−3695.
(8) Sen, S.; Liu, D.; Palmore, G. T. R. ACS Catal. 2014, 4, 3091−
3095.
(41) Hatsukade, T.; Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo,
T. F. Phys. Chem. Chem. Phys. 2014, 16, 13814−13819.
(42) Schouten, K.; Kwon, Y.; Van der Ham, C.; Qin, Z.; Koper, M.
Chem. Sci. 2011, 2, 1902−1909.
(43) Vollmer, S.; Witte, G.; Woll, C. Catal. Lett. 2001, 77, 97−101.
̈
(44) Greeley, J.; Jaramillo, T. F.; Bonde, J.; Chorkendorff, I.;
Norskov, J. K. Nat. Mater. 2006, 5, 909−913.
(45) Zhang, Y.-J.; Sethuraman, V.; Michalsky, R.; Peterson, A. A. ACS
Catal. 2014, 4, 3742−3748.
(46) Wu, J.; Risalvato, F. G.; Ma, S.; Zhou, X.-D. J. Mater. Chem. A
2014, 2, 1647−1651.
(9) Ogura, K. J. CO2 Util. 2013, 1, 43−49.
(10) Farrell, A. E.; Plevin, R. J.; Turner, B. T.; Jones, A. D.; O’Hare,
M.; Kammen, D. M. Science 2006, 311, 506−508.
(11) Hill, J.; Nelson, E.; Tilman, D.; Polasky, S.; Tiffany, D. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103, 11206−11210.
(12) Kniel, L.; Winter, O.; Stork, K. Ethylene: Keystone to the
Petrochemical Industry; Marcel Dekker: New York, 1980.
(13) Mills, G. A.; Ecklund, E. E. Annu. Rev. Energy 1987, 12, 47−80.
(14) Hori, Y. In Modern Aspects of Electrochemistry; Vayenas, C.,
White, R., Gamboa-Aldeco, M., Eds.; Springer: New York, 2008; pp
89−189.
(15) Qiao, J.; Liu, Y.; Hong, F.; Zhang, J. Chem. Soc. Rev. 2014, 43,
631−675.
(16) Montoya, J. H.; Peterson, A. A.; Nørskov, J. K. ChemCatChem
2013, 5, 737−742.
(17) Hori, Y.; Murata, A.; Takahashi, R. J. Chem. Soc., Faraday Trans.
1 1989, 85, 2309−2326.
(18) Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. Energy
Environ. Sci. 2012, 5, 7050−7059.
(19) Tang, W.; Peterson, A. A.; Varela, A. S.; Jovanov, Z. P.; Bech, L.;
Durand, W. J.; Dahl, S.; Nørskov, J. K.; Chorkendorff, I. Phys. Chem.
Chem. Phys. 2012, 14, 76−81.
(20) Bugayong, J.; Griffin, G. L. Mater. Res. Soc. Symp. Proc. 2013,
1542, 05−11.
(21) Kas, R.; Kortlever, R.; Milbrat, A.; Koper, M. T. M.; Mul, G.;
Baltrusaitis, J. Phys. Chem. Chem. Phys. 2014, 16, 12194−12201.
(22) Hori, Y.; Takahashi, I.; Koga, O.; Hoshi, N. J. Phys. Chem. B
2002, 106, 15−17.
(23) Yano, H.; Tanaka, T.; Nakayama, M.; Ogura, K. J. Electroanal.
Chem. 2004, 565, 287−293.
(24) Li, C. W.; Ciston, J.; Kanan, M. W. Nature 2014, 508, 504−507.
(25) Reske, R.; Duca, M.; Oezaslan, M.; Schouten, K. J. P.; Koper, M.
T. M.; Strasser, P. J. Phys. Chem. Lett. 2013, 4, 2410−2413.
(26) Chen, C. S.; Handoko, A. D.; Wan, J. H.; Ma, L.; Ren, D.; Yeo,
B. S. Catal. Sci. Technol. 2015, 5, 161−168.
(27) Hori, Y.; Takahashi, I.; Koga, O.; Hoshi, N. J. Mol. Catal. A:
Chem. 2003, 199, 39−47.
(28) Kim, D.; Lee, S.; Ocon, J. D.; Jeong, B.; Lee, J. K.; Lee, J. Phys.
Chem. Chem. Phys. 2015, 17, 824−830.
(29) Yano, J.; Yamasaki, S. J. Appl. Electrochem. 2008, 38, 1721−1726.
(30) Li, C. W.; Kanan, M. W. J. Am. Chem. Soc. 2012, 134, 7231−
7234.
(31) Kas, R.; Kortlever, R.; Yılmaz, H.; Koper, M. T. M.; Mul, G.
ChemElectroChem 2015, 2, 354−358.
(32) Hori, Y.; Takahashi, R.; Yoshinami, Y.; Murata, A. J. Phys. Chem.
B 1997, 101, 7075−7081.
(33) Mukhopadhyay, A. K.; Chakraborty, A. K.; Chatterjee, A. P.;
Lahiri, S. K. Thin Solid Films 1992, 209, 92−96.
(34) Yeo, B. S.; Klaus, S. L.; Ross, P. N.; Mathies, R. A.; Bell, A. T.
ChemPhysChem 2010, 11, 1854−1857.
(35) Le, M.; Ren, M.; Zhang, Z.; Sprunger, P. T.; Kurtz, R. L.; Flake,
J. C. J. Electrochem. Soc. 2011, 158, E45−E49.
(36) Bonvalot-Dubois, B.; Dhalenne, G.; Berthon, J.; Revcolevschi,
A.; Rapp, R. A. J. Am. Ceram. Soc. 1988, 71, 296−301.
(37) Niaura, G. Electrochim. Acta 2000, 45, 3507−3519.
(38) Beverskog, B.; Puigdomenech, I. J. Electrochem. Soc. 1997, 144,
3476−3483.
(39) Takahashi, I.; Koga, O.; Hoshi, N.; Hori, Y. J. Electroanal. Chem.
2002, 533, 135−143.
(40) Calle-Vallejo, F.; Koper, M. T. M. Angew. Chem., Int. Ed. 2013,
52, 7282−7285.
2821
DOI: 10.1021/cs502128q
ACS Catal. 2015, 5, 2814−2821