Stability of Residual Oxides in Oxide-Derived Copper Catalysts for Electrochemical CO2 Reduction Investigated with 18O Labeling

Article abstract of DOI:10.1002/anie.201710590

Oxide-derived (OD) Cu catalysts have high selectivity towards the formation of multi-carbon products (C₂/C₃) for aqueous electrochemical CO₂ reduction (CO₂R). It has been proposed that a large fraction of the initial oxide can be surprisingly resistant to reduction, and these residual oxides play a crucial catalytic role. The stability of residual oxides was investigated by synthesizing ¹⁸O-enriched OD Cu catalysts and testing them for CO₂R. These catalysts maintain a high selectivity towards C₂/C₃ products (ca. 60 %) for up to 5 h in 0.1 m KHCO₃ at ?1.0 V vs. RHE. However, secondary-ion mass spectrometry measurements show that only a small fraction (<1 %) of the original ¹⁸O content remains, showing that residual oxides are not present in significant amounts during CO₂R. Furthermore, we show that OD Cu can reoxidize rapidly, which could compromise the accuracy of ex situ methods for determining the true oxygen content.

Full text of DOI:10.1002/anie.201710590

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: Stability of residual oxides in oxide-derived Cu catalysts for
electrochemical CO2 reduction investigated with 18O labeling
Authors: Yanwei Lum and Joel W. Ager
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: Angew. Chem. Int. Ed. 10.1002/anie.201710590
Angew. Chem. 10.1002/ange.201710590
Link to VoR: http://dx.doi.org/10.1002/anie.201710590
http://dx.doi.org/10.1002/ange.201710590

Angewandte Chemie International Edition
10.1002/anie.201710590
COMMUNICATION
Stability of residual oxides in oxide-derived Cu catalysts for
1
8
electrochemical CO reduction investigated with O labeling
2
Yanwei Lum and Joel W. Ager*
Abstract: Oxide-derived (OD) Cu catalysts have high selectivity
towards the formation of multi-carbon products (C /C ) for aqueous
electrochemical CO reduction (CO R). It has been proposed that a
amount of oxygen residing in the subsurface, possibly modifying
the electronic structure of the catalyst and creating active sites
with higher CO binding energy.
2
3
[
15]
2
2
large fraction of the initial oxide can be surprisingly resistant to
reduction and that these residual oxides play a crucial role in
First principles calculations performed by Goddard and co-
workers found that the presence of oxygen in the subsurface
+
0
promoting the formation of C
stability of residual oxides by synthesizing O enriched OD Cu
catalysts and testing them for CO R. These catalysts maintain a high
selectivity towards C /C products (~60%) for up to 5h in 0.1 M
KHCO at -1.0 vs RHE. However, secondary ion mass
spectrometry measurements show that only a small fraction (< 1%)
/C
2 3
products. We investigate the
would generate a mix of Cu and Cu on the surface. These
could then work with adsorbed H O and aid in CO activation by
forming chemisorbed CO , which is the first step of CO
separate report, they investigated a partially reduced copper
18
2
2
[
16]
2
2
2
R. In a
2
3
0
+
[17]
3
V
oxide matrix consisting of a mix of Cu and Cu regions. They
+
0
showed that Cu and Cu regions could act synergistically to
18
of the original O content remains, showing that residual oxides are
not present in significant amounts during CO R. Furthermore, we
promote CO dimerization and suppress C
boosting catalyst selectivity.
1
pathways, thereby
2
show that OD Cu reoxidizes rapidly in the absence of a reducing
potential, which could compromise the accuracy of ex-situ methods
for determining the true oxygen content.
However, a few aspects of this emerging picture of the role
of residual oxides in controlling the activity of OD Cu are
puzzling. (1) Cu nanomaterials have been shown to readily
oxidize, which could create difficulties in performing accurate ex-
[
18,19]
situ quantification of the oxygen content.
(2) Moisture and
2 2
Electrochemical reduction of CO (CO R) into chemical fuels and
oxygen are known to cause corrosion (oxidation) of Cu and
exposure to both is difficult to avoid. Corrosion might also be
exacerbated in a porous, high surface area material such as OD
Cu. (3) OD Cu possesses a high density of grain boundaries,
which are known to accelerate the oxidation process by serving
feedstock, powered by renewable electrical energy has been
proposed as a strategy to mitigate rising greenhouse gas
[1–3]
emissions.
Two main challenges in this area of research are
improving the product selectivity and reducing the overpotentials
[
2,4,5]
required to drive CO
2
R.
“Oxide-derived” Cu (OD Cu)
as nucleation sites and as pathways where diffusion can occur
catalysts have attracted much attention because they exhibit
higher selectivity towards potentially valuable multi-carbon
products (e.g. ethanol and ethylene) with lower overpotential
[20–24]
at a faster rate.
It is therefore possible that upon removal of
a reducing potential, Cu rapidly reoxidizes and as a result, the O
content characterized ex-situ would not be representative of the
[
6–13]
requirements.
2
actual case during CO R.
OD Cu catalysts are formed by oxidizing Cu and
subsequently reducing it. Recently, it has been proposed that
residual oxides are responsible for its remarkable catalytic
18
To address these concerns, we employed O isotope
labeling in order to confirm the presence/absence of residual
18
2
oxides in OD Cu during CO R. O enriched OD Cu catalysts
[
11,12,14]
properties.
Using ex-situ energy dispersive X-ray
18
were synthesized by oxidation/reduction cycling in
2
H O
spectroscopy (EDS), Cuenya and co-workers reported that a
large fraction of the initial oxide can be resistant to reduction
even under strongly reducing potentials typically used for
following the procedure of Nilsson and co-workers (see SI for full
[25]
details). These OD Cu catalysts (OD18 Cu) were then tested
18
2
for CO R for various times and the residual O content was
[
11,12]
+
CO
2
R.
They proposed that the presence of Cu on the
analyzed ex-situ using secondary ion-mass spectrometry (SIMS).
This ex-situ method allows us to determine what the "in-situ"
[12]
surface was important for the formation of C
2
3
/C products.
Nilsson and co-workers studied an OD Cu catalyst with ambient
pressure X-ray photoelectron spectroscopy (XPS) and electron
energy loss spectroscopy (EELS) and they found a small
oxygen content of the catalyst is during operation as any
18
subsequent reoxidation of the catalyst would not result in
O
enrichment above the natural isotopic abundance of 0.2
atomic %.
Figure 1a shows the electrochemical oxidation and reduction
process used to generate OD18 Cu by sweeping to an anodic
potential and then to a cathodic potential in an electrolyte
[*]
Y. Lum and Prof. J.W. Ager
Joint Center for Artificial Photosynthesis and Materials Sciences
Division, Lawrence Berkeley National Laboratory, California 94720,
United States
18
consisting of 0.1 M KHCO
3 2
and 32 mM of KCl in H O. Working
electrodes were 1 µm thick Cu films deposited on Si, and
1
8
oxidation results in formation of Cu
2
O nanocubes on the
Department of Materials Science and Engineering, University of
California, Berkeley, California 94720, United States
surface. Reduction results in loss of the cubical morphology and
formation of a porous oxide-derived layer (roughness factor ~22),
which is approximately 100 nm thick. Additional characterization,
including electroactive surface area, additional SEM images,
Raman spectroscopy, x-ray photoelectron spectroscopy and x-
E-mail: jwager@lbl.gov
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201710590
COMMUNICATION
ray diffraction are available in the SI (Figures S1 to S8).
agreement with what has been reported for OD Cu in the
literature (Table S6). During long-term testing (5 h) the gas
product distribution was relatively stable, with an ethylene
selectivity of ~34% throughout (Figures 1c and S9).
a
30
20
10
0
Next, CO
mins, 30 mins, 1 hr and 5 hrs with fresh samples of OD18 Cu.
After CO R, the samples were rinsed with copious amounts of DI
water, dried and then quickly stored in a N glove box. The
2
R was performed under identical conditions for 10
Cu
Cu O
2
2
2
1
8
16
residual O and O contents of these samples, as well as a
freshly prepared sample (0 mins) were determined ex-situ using
SIMS (Figure 2). Also, a control sample was prepared by
-10
-20
-30
16
oxidation and reduction in H
for 1 hr (OD16 Cu 1 hr).
2
2
O electrolyte and tested for CO R
Cu
Cu O
2
-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2
Potential vs Ag/AgCl (V)
a
b
70
60
50
40
30
20
10
0
C2/C3
C1
Hydrogen
23
1
1
1
0
0
0
OD18 Cu 0 mins
OD18 Cu 10 mins
OD18 Cu 5 hrs
18O
2
2
2
2
1
0
"
Control" OD16 Cu 1 hr
After
CO2R
10
19
1
0
0
1
8
7
1
OD Cu
Cu film
Cu foil
1
10
-2
-2
-2
(
10.9 mA cm ) (6.7 mA cm ) (6.9 mA cm )
0
100
200
300
400
500
600
c
Depth (nm)
4
0
5
b
-
-
6
8
1
.2
OD18 Cu 10 mins
OD18 Cu 30 mins
OD18 Cu 1 hr
1
8
Residual O
3
1.0
0.8
OD18 Cu 5 hrs
-
-
-
10
12
14
30
0.6
0.4
0.2
0.0
2
5
0
2
0
.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Time (hours)
0
20
40
60
80
100
120
Figure 1. (a) Electrochemical oxidation and reduction procedure for
generating OD18 Cu. The blue box highlights the oxidation stage and the
green box highlights the reduction stage. SEM images are shown as insets
Depth (nm)
c
OD18 Cu 0 mins
OD18 Cu 10 mins
OD18 Cu 5 hrs
(
scale bars: 0.5 µm) (b) Faradaic efficiency to C
Cu, Cu film and Cu foil (70 min tests). Average current densities are given in
parenthesis (full data in Tables S2 to S5). (c) 5 h long CO R with OD18 Cu.
Graph shows current density (black line) and faradaic efficiency to ethylene
red squares) vs time. See Figure S9 for faradaic efficiency to other gas
2
/C
3
, C
1
and hydrogen for OD
16O
2.5
2.0
2
"Control" OD16 Cu 1 hr
(
products. Error bars are standard deviations from 3 repeat experiments.
1.5
The product distributions obtained by operating OD18 Cu
After
CO2R
catalysts at -1.0 V vs RHE with CO
2
saturated 0.1 M KHCO
O electrolyte for 70 mins are shown in Figure 1b along with
results from the unmodified Cu film and a Cu foil. As expected,
OD18 Cu has a higher selectivity towards C /C products
60.1%) and lower selectivity towards C products (6.5%)
relative to planar Cu controls. Values obtained here are in good
3
in
1.0
1
6
H
2
0.5
2
3
0
20
40
60
80
100
120
(
1
Depth (nm)
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201710590
COMMUNICATION
1
8
16
18
18
2
Figure 2. (a) O and (c) O content of OD Cu catalysts with different CO R
the O content is not simply due to trapped H
2
O in these
durations measured by SIMS. Note the different scales in (a) and (c). For ease
of viewing, the results for 30 mins and 1 hour are omitted in (a) and (c) but are
available in Figures S17 and S18. (b) shows the percentage of O that
remains in OD18 Cu after CO R (see SI for calculation details)
porous catalysts. This is evidenced by Raman peak shifts of
Cu O consistent with isotopic substitution (Figure S11a). Taken
together, all these results highlight the difficulty of ex-situ
18
2
2
measurements of the oxygen content, since OD Cu can be
rapidly reoxidized.
1
8
As seen from Figure 2a, the OD18 Cu 0 mins has
O
2
2
content on the order of 10 atoms/cc, which is on the same
order of magnitude as the atomic density of Cu, showing
2
3
2
10
"Control" OD16 Cu 1 hr
18O
Oxidation in
electrolyte
18
1
8
OD16 Cu CO R in H O
successful incorporation of
However, after CO R has been carried out, the O content
drops drastically. Figure 2b shows the percentage of the initial
O into the catalyst structure.
2 2
2
18
OD16 Cu soaked in H2
18
10
10
10
10
10
O
2
1
8
21
20
O that remains after CO
2
R (details of calculation in SI) in the
top ~100 nm layer (estimated thickness of the oxide-derived
1
8
layer). Surprisingly, only <1% of the original O remains after
1
8
CO
2
R and, in fact, the O content drops to the same order of
1
1
9
8
magnitude as the control sample, OD16 Cu 1 hr, which was not
1
8
18
enriched with O. (We note that the O in the control sample is
due to its natural abundance of 0.2 atomic %.) This means that
the oxide layer is completely reduced, as would be expected
since the applied potential typical of CO
the standard reduction potential of Cu O. There therefore cannot
be a large concentration of oxygen remaining in the catalyst
under CO R conditions. Another possibility is that during CO R,
O atoms that leave as a result of reduction are replaced
2
R is >1 V negative of
17
1
0
0
100
200
300
400
500
600
2
Depth (nm)
2
2
1
8
1
8
Figure 3. O content in OD16 Cu 1 hr, OD16 Cu with CO R for 3 mins in
2
1
6
18
2
18
2
16
instantly via oxidation, with O taking its place. However this is
unlikely as the oxidation step is thermodynamically unfavorable
because of the highly reducing potentials applied.
H
O electrolyte and OD16 Cu that was soaked in H
content was also analyzed (Figure S11b).
O for 3 mins. The
O
1
6
Intriguingly even after CO
2
R, the O content in OD18 Cu
16
Finally, we consider the reasons for the rapid reoxidation of
OD Cu. Firstly, Cu nanomaterials are known to be easily
oxidized due to numerous highly reactive undercoordinated
atoms on the surface. Secondly, previous studies have shown
that grain boundaries can act as nucleation sites for oxide
growth as well as provide channels diffusion can take place at a
does not change and actually increases (Figure 2c). In fact,
O
2
2
contents after CO
2
R are on the order of 10
atoms/cc,
suggesting that samples have been heavily oxidized. However,
in our experimental process, care was taken to ensure that
exposure to ambient air was as minimal as possible. We
therefore considered the possibility that reoxidation of OD Cu
occurs rapidly, as soon as the reducing potential is turned off. In-
situ Raman spectroscopy measurements by Yeo and co-workers
[
19]
[
20–24]
faster rate (compared to diffusion through the bulk lattice).
OD Cu has a high density of grain boundaries,
[
8,27,28]
which could
explain why OD Cu can reoxidize so quickly via exposure to
ambient air and moisture (see Schematic 1). Comparatively, Cu
with fewer grain boundaries (Cu film) does not oxidize as quickly,
as illustrated by our Cu Auger LMM measurements (Figure S8).
on Cu
vs RHE resulted in loss of the characteristic Cu
and they concluded that complete reduction to metallic Cu had
2
O films revealed that application of a potential of -0.99 V
2
O Raman peaks,
0
[
6]
taken place. Surprisingly, once the potential was removed,
Cu O peaks began to appear again after 60 s and by 120 s had
2
become very distinct; meaning that rapid reoxidation of OD Cu
had taken place. In this work, in-situ Raman spectroscopy was
also carried out (Figure S10) and similar observations were
made.
To further emphasize the rapid oxidation of OD Cu, a freshly
1
8
prepared OD16 Cu sample was soaked in H
2
O (with 0.1 M
KHCO ) for 3 mins and then rinsed immediately with DI water.
3
1
8
SIMS measurements revealed that the O content in this
sample (Figure 3, blue trace) indeed rises by an order of
2
magnitude. To further investigate the reoxidation process, CO R
18
was carried out on an OD16 Cu sample in H
2
O electrolyte (0.1
M KHCO ) at -1.0 V vs RHE for 3 mins. This sample was then
3
1
8
washed with DI water and the O content in this sample was
Scheme 1. Rapid reoxidation of OD Cu occurs because oxygen diffusion and
oxide nucleation can be facilitated by grain boundaries.
similarly analyzed ex-situ using SIMS (Figure 3, red trace).
Startlingly, this sample showed a very high incorporation of O,
with concentrations on the order of ~10 atoms/cc, meaning that
a large degree of reoxidation had occurred. We emphasize that
1
8
2
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201710590
COMMUNICATION
In conclusion, the stability of residual oxides was
B. Roldan Cuenya, ACS Nano 2017, 11, 4825–4831.
1
8
18
investigated by O isotope labeling. O enriched oxide-derived
Cu catalysts were synthesized and CO R was performed with
them. The residual O content was then analyzed ex-situ with
[12]
H. Mistry, A. S. A. S. Varela, C. S. C. S. C. S. Bonifacio, I.
Zegkinoglou, I. Sinev, Y.-W. Y.-W. Choi, K. Kisslinger, E. A. E. A.
Stach, J. C. J. C. Yang, P. Strasser, et al., Nat. Commun. 2016, 7,
2
18
1
8
SIMS and it was found that only <1% of the original O content
remained in the samples. We therefore conclude that residual
oxides are unstable in our oxide-derived Cu catalysts during
12123.
[
[
[
13]
14]
15]
S. Lee, D. Kim, J. Lee, Angew. Chemie 2015, 127, 14914–14918.
C. H. Lee, M. W. Kanan, ACS Catal. 2015, 5, 465–469.
A. Eilert, F. Cavalca, F. S. Roberts, J. Osterwalder, C. Liu, M.
Favaro, E. J. Crumlin, H. Ogasawara, D. Friebel, L. G. M.
Pettersson, et al., J. Phys. Chem. Lett. 2017, 8, 285–290.
M. Favaro, H. Xiao, T. Cheng, W. A. Goddard, J. Yano, E. J.
Crumlin, Proc. Natl. Acad. Sci. 2017, 114, 201701405.
H. Xiao, W. A. Goddard, T. Cheng, Y. Liu, Proc. Natl. Acad. Sci.
2017, 201702405.
2 2 3
CO R. We believe that the high C /C product selectivity
exhibited by these catalysts is more likely due to a high density
of grain boundaries as previously proposed by Kanan and co-
[28,29]
workers
. We also show that these catalysts can rapidly
[
[
[
16]
17]
18]
reoxidize due, possibly, to the numerous grain boundaries
present in the material. The rapid reoxidation process could
compromise the accuracy of ex-situ methods for determining the
2
true oxygen content of the catalyst during CO R conditions.
I. Platzman, R. Brener, H. Haick, R. Tannenbaum, J. Phys. Chem. C
2
008, 112, 1101–1108.
L. Xu, Y. Yang, Z.-W. Hu, S.-H. Yu, ACS Nano 2016, 10, 3823–
834.
Acknowledgements
[19]
3
This material is based upon work performed by the Joint Center
for Artificial Photosynthesis, a DOE Energy Innovation Hub,
supported through the Office of Science of the U.S. Department
of Energy under Award Number DE-SC0004993, which
supported synthesis and electrochemical measurements. SIMS
analysis was supported by the Singapore Berkeley Initiative for
[20]
[21]
[22]
S. Matsnnaga, T. Homma, Oxid. Met. 1976, 10, 361–376.
C. Gattinoni, A. Michaelides, Surf. Sci. Rep. 2015, 70, 424–447.
K. Fujita, D. Ando, M. Uchikoshi, K. Mimura, M. Isshiki, Appl. Surf.
Sci. 2013, 276, 347–358.
[23]
Q. Zhu, L. Zou, G. Zhou, W. A. Saidi, J. C. Yang, Surf. Sci. 2016,
652, 98–113.
Sustainable
Energy
(SinBeRISE),
National
Research
[
24]
A. Paul, T. Laurila, V. Vuorinen, S. V Divinski, in (Eds.: A. Paul, T.
Laurila, V. Vuorinen, S. V Divinski), Springer International
Publishing, Cham, 2014, pp. 429–491.
Foundation (NRF) of Singapore. Y.L. acknowledges the support
of an A*STAR National Science Scholarship. The authors thank
Alan Wan from EAG Laboratories for SIMS analysis, Peter Zhao
for SEM analysis, and Prof. Thomas Devine for helpful
discussions.
[
[
25]
26]
F. S. Roberts, K. P. Kuhl, A. Nilsson, Angew. Chemie 2015, 127,
5268–5271.
M. Favaro, H. Xiao, T. Cheng, W. A. Goddard, J. Yano, E. J.
Crumlin, Proc. Natl. Acad. Sci. U. S. A. 2017, 114, 6706–6711.
C. W. Li, J. Ciston, M. W. Kanan, Nature 2014, 508, 504–507.
A. Verdaguer-Casadevall, C. W. Li, T. P. Johansson, S. B. Scott, J.
T. McKeown, M. Kumar, I. E. L. Stephens, M. W. Kanan, I.
Chorkendorff, J. Am. Chem. Soc. 2015, 137, 9808–9811.
X. Feng, K. Jiang, S. Fan, M. W. Kanan, ACS Cent. Sci. 2016, 2,
Keywords: CO
2
reduction • copper oxides • electrocatalysis •
[
[
27]
28]
oxide stability • isotope labeling
[
[
1]
2]
C. Graves, S. D. Ebbesen, M. Mogensen, K. S. Lackner, Renew.
Sustain. Energy Rev. 2011, 15, 1–23.
[
29]
Y. Hori, in Mod. Asp. Electrochem. (Eds.: C.G. Vayenas, R.E. White,
M.E. Gamboa-Aldeco), Springer New York, New York, NY, 2008, pp.
169–174.
8
9–189.
J.-P. Jones, G. K. S. Prakash, G. A. Olah, Isr. J. Chem. 2014, 54,
451–1466.
[
[
[
[
[
3]
4]
5]
6]
7]
1
K. P. Kuhl, E. R. Cave, D. N. Abram, T. F. Jaramillo, Energy Environ.
Sci. 2012, 5, 7050–7059.
K. P. Kuhl, T. Hatsukade, E. R. Cave, D. N. Abram, J. Kibsgaard, T.
F. Jaramillo, J. Am. Chem. Soc. 2014, 136, 14107–14113.
D. Ren, Y. Deng, A. D. Handoko, C. S. Chen, S. Malkhandi, B. S.
Yeo, ACS Catal. 2015, 5, 2814–2821.
M. Ma, K. Djanashvili, W. A. Smith, Angew. Chemie Int. Ed. 2016,
55, 6680–6684.
[
[
8]
9]
C. W. Li, M. W. Kanan, J. Am. Chem. Soc. 2012, 134, 7231–7234.
R. Kas, R. Kortlever, A. Milbrat, M. T. M. Koper, G. Mul, J.
Baltrusaitis, Phys. Chem. Chem. Phys. 2014, 16, 12194.
Y. Lum, B. Yue, P. Lobaccaro, A. T. Bell, J. W. Ager, J. Phys. Chem.
C 2017, 121, 14191–14203.
[
[
10]
11]
D. Gao, I. Zegkinoglou, N. J. Divins, F. Scholten, I. Sinev, P. Grosse,
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.201710590
COMMUNICATION
Layout 2:
COMMUNICATION
Yanwei Lum, Joel W. Ager
Page No. – Page No.
Stability of residual oxides in oxide-
derived Cu catalysts for
2
electrochemical CO reduction
1
8
investigated with O labeling
In this work, we show that residual oxides in oxide-derived Cu catalysts are not
2
stable under strongly reducing potentials during electrochemical CO reduction.
1
8
This is demonstrated with the use of O labelled oxide-derived Cu catalysts.
Instead we find that the surface is rapidly reoxidized in the absence of a reducing
potential.
This article is protected by copyright. All rights reserved.