SO2-Induced Selectivity Change in CO2 Electroreduction

Article abstract of DOI:10.1021/jacs.9b03215

Electrochemical conversion of carbon dioxide (CO₂) to value-added chemicals has attracted much attention in recent years as a potential alternative to fossil resources. Although significant works have studied the influence of impurities in the electrolyte (e.g., metal ions), few studies have been performed to understand the influence of gaseous impurities in CO₂ electroreduction. Herein, we study the effects of sulfur dioxide (SO₂) on Ag-, Sn-, and Cu-catalyzed CO₂ electrolysis in a flow-cell electrolyzer in near-neutral electrolyte, representing a broad range of CO₂ reduction catalysts. We show that the presence of SO₂ impurity reduces the efficiency of converting CO₂ due to the preferential reduction of SO₂. In the cases of Ag and Sn, the effect of SO₂ impurity was reversible and the catalytic activities of both catalysts were recovered. On the contrary, a shift in selectivity toward formate accompanied by a suppression of multicarbon (C₂₊) products was observed on Cu catalyst, demonstrating that Cu is highly sensitive to SO₂ impurity. Our results suggest that CO₂ obtained from direct air capture technologies or biorefineries could be more suitable for Cu-catalyzed CO₂ electrolysis as these CO₂ sources would be relatively cleaner (SO₂-free) than fossil-derived sources such as power plants and can be directly coupled with distributed renewable energy sources such as wind and solar.

Full text of DOI:10.1021/jacs.9b03215

Subscriber access provided by UNIV OF SOUTHERN INDIANA
Article
SO2-induced Selectivity Change in CO2 Electroreduction
Wesley W Luc, Byung Hee Ko, Shyam Kattel, Shuang Li, Dong Su, Jingguang G Chen, and Feng Jiao
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b03215 • Publication Date (Web): 03 Jun 2019
Downloaded from http://pubs.acs.org on June 3, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 17
Journal of the American Chemical Society
1
2
3
4
SO2-induced Selectivity Change in CO2 Electroreduction
5
6
7
8
Wesley Luc,1† Byung Hee Ko, 1† Shyam Kattel,2 Shuang Li,3 Dong Su,3 Jingguang G. Chen,4* and
Feng Jiao1*
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[1] Center for Catalytic Science and Technology, Department of Chemical and Biomolecular
Engineering, University of Delaware, Newark, DE, 19716 (USA)
[2] Department of Physics, Florida A&M University, Tallahassee, Florida, 32307 (USA)
[3] Center for Functional Nanomaterials, Brookhaven National Laboratory, Upton, New York,
11973 (USA)
[4] Department of Chemical Engineering, Columbia University, New York, New York, 10027
(USA)
†These authors contributed equally to this work.
*Corresponding author: jgchen@columbia.edu, jiao@udel.edu
Abstract
Electrochemical conversion of carbon dioxide (CO2) to value-added chemicals has
attracted much attention in recent years as a potential alternative to fossil resources. Although
significant works have studied the influence of impurities in the electrolyte (e.g. metal ions), few
studies have been performed to understand the influence of gaseous impurities in CO2
electroreduction. Herein, we study the effects of sulfur dioxide (SO2) on Ag-, Sn-, and Cu-
catalyzed CO2 electrolysis in a flow-cell electrolyzer in near-neutral electrolyte, representing a
broad range of CO2 reduction catalysts. We show that the presence of SO2 impurity reduces the
efficiency of converting CO2 due to the preferential reduction of SO2. In the case of Ag and Sn,
the effect of SO2 impurity was reversible and the catalytic activities of both catalysts were
recovered. On the contrary, a shift in selectivity toward formate accompanied with a suppression
of multi-carbon (C2+) products was observed on Cu catalyst, demonstrating that Cu is highly
sensitive to SO2 impurity. Our results suggest that CO2 obtained from direct air capture
technologies or biorefineries could be more suitable for Cu-catalyzed CO2 electrolysis as these
CO2 sources would be relatively cleaner (SO2-free) than fossil-derived sources such as power
plants and can be directly coupled with distributed renewable energy sources such as wind and
solar.
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 17
1
2
3
4
1. Introduction
5
6
7
8
The electrochemical production of value-added chemicals using carbon dioxide (CO2) as
the carbon source has attracted much attention in recent years and has been motivated by ever
increasing CO2 emission and anticipated cheap electricity prices derived from renewable energy
sources.1 These efforts include fundamental investigation combining experimental and
computational studies to establish design principles for the rational design of catalysts,2 novel
synthesis approaches for constructing high surface area materials with specified active sites,3,4 and
reactor engineering to design practical devices.5-8 In the latter case, the recent development of
flow-cell electrolyzers,8-10 in which catalysts are positioned on a well-engineered electrode-
electrolyte interface allowing gaseous reactant to be fed from one side while electrolyte is fed on
the other, has enabled the electrochemical conversion of CO2 at practical rates of reaction that
cannot be achieved with typical batch reactors, further pushing this technology toward
commercialization.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Critical to the development of CO2 electrolysis technologies is also the key understanding
of potential sources of impurities in the incoming feed (e.g. electrolyte and CO2), as well as the
identification of potential mechanisms for performance degradation. It has been well studied that
trace impurities in the electrolyte such as metal ions (e.g. Fe2+ and Zn2+)11 can dramatically degrade
CO2 reduction performance where the catalytic surface can be irreversibly altered through the
deposition of these metal ions under reducing potentials, shifting selectivities toward undesirable
products.12 Strategies such as pre-electrolysis using sacrificial electrodes,12 pretreatment of the
electrolyte with metal-chelexing resin,13 and designing catalysts where the catalytic surface can
reconstruct under operating conditions14 have been proposed to overcome these challenges.
Overall, the electrolyte most likely needs to be relatively pure and free of residual metal ions for
commercial electrochemical conversion of CO2. On the contrary, the effect of gaseous impurities
in the CO2 feed has barely been studied15, and further work is necessary to elucidate the impact of
such impurities on the performance in CO2 electrolysis.
The CO2 feedstock for commercial CO2 electrolysis can be obtained from direct air
capture16,17 or point sources such as power plants or chemical facilities.18,19 The former is currently
an area of intensive research due to the advantage of not being geographically limited; however,
such technology has not yet been fully commercialized and still under pilot plant development.
The latter accounts for a large fraction of anthropogenic CO2 emissions with the current state-of-
the-art CO2 capture technology being amine-based absorption technologies. However, the exhaust
streams from point sources such as power plants contain impurities such as SOx, NOx, or volatile
organic compounds.20 Although current carbon capture processes can purify these exhaust streams
using energy intensive processes involving series of scrubbing and separating units, it is critical to
understand how these impurities affect CO2 electrolysis to establish engineering criteria. To the
best of our knowledge, majority of CO2 research to date uses highly pure CO2 gas as reactant and
little is known on the effect of gaseous impurities.
Herein, we study the influence of sulfur dioxide (SO2) in the electrochemical conversion
of CO2 in a flow-cell electrolyzer. Three different catalysts, silver (Ag), tin (Sn), and copper (Cu)
as selective catalysts for producing carbon monoxide (CO), formate, and multi-carbon (C2+)
products, respectively, are investigated as these materials are the most commonly studied in CO2
2
ACS Paragon Plus Environment

Page 3 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
electrolysis (Figure 1).11 The catalysts are examined for CO2 reduction free of SO2 impurity, during
the presence of SO2, and finally after SO2 exposure to determine if the gaseous impurity has an
irreversible effect on the catalytic performance. The CO2 reduction products are tracked over the
course of electrolysis and ex-situ surface characterization is conducted at various time points
during operation to elucidate the underpinning of catalytic change. The results show that the
presence of SO2 impurity in the incoming CO2 feed reduces the efficiency of converting CO2 due
to the preferential reduction of SO2, as it is thermodynamically more favorable to reduce SO2 than
CO2. In the case of Ag and Sn catalysts, SO2 impurity does not change the CO2 reduction product
selectivities and full recovery of the catalytic activities can be observed after the catalysts were
subjected to SO2. However, in the case of the Cu catalyst, SO2 impurity has an irreversible,
detrimental effect as the overall selectivity is shifted toward formate while suppressing the
formation of C2+ products. Characterization and computational efforts indicate that the formation
of copper(I) sulfide (Cu2S) is likely responsible for the change in selectivity for Cu-catalyzed CO2
electrolysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Schematic of CO2 electrolyzer and potential feed sources.
2. Results and Discussion
Initial Characterization and Experimental Procedures
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 17
1
2
3
4
5
6
7
8
9
Commercial Ag, Sn, and Cu nanoparticles, representing a broad range of CO2 reduction
catalysts, were loaded on a gas-diffusion layer (GDL) with a catalyst loading of 0.3 mg cm-2 using
a drop-casting method as previously described.21 Scanning electron microscopy (SEM) images
(Figure S1) show the well-dispersed nature of the nanoparticles on the GDL and the porosity of
the catalyst layer, which is critical for maintaining the electrode-electrolyte interface by allowing
facile transport of gaseous reactants and products to and away from the catalytic surface. Powder
X-ray diffraction (PXRD) (Figure S2) of the as-purchased nanoparticles reveals the phase purity
of the nanoparticles. In the case of Sn and Cu nanoparticles, the presence of oxidized Sn and Cu
in the form of SnO and Cu2O, respectively, is apparent. The surface composition of the as-prepared
electrodes was analyzed using ex-situ X-ray photoelectron spectroscopy (XPS) (Figure S3). In the
case of Sn and Cu, the surfaces are partially-oxidized, most likely from air exposure during sample
handling. The as-deposited nanoparticles were analyzed for the influence of SO2 on CO2
electroreduction in a three compartment flow-cell electrolyzer (Figure S4) with two compartments
for the anolyte and catholyte and one compartment for the gaseous reactants and products.
Although previous studies have shown that conducting CO2 electrolysis in alkaline electrolytes
(e.g. KOH) has improvement in CO2 reduction performances,8,21,22 CO2 strongly reacts with
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
-
2-
hydroxide (OH-) to form a bicarbonate/carbonate (KHCO3 /CO3 ) mixture, which degrades the
electrolyte and reduces the efficiency of CO2 utilization.9,23 Therefore, CO2 electrolysis
experiments were conducted in near-neutral conditions in 1.0 M potassium bicarbonate (KHCO3)
solution. A syringe pump was used to accurately control the feed rate of SO2 gas into the CO2 feed
before entering the flow-cell electrolyzer, and electrolysis was conducted in the constant current
mode. In short, pure CO2 was initially fed for 30 minutes, then SO2 was introduced into the CO2
feed for a duration of 30 minutes, and then the SO2 feed was stopped for the remainder of the
electrolysis. The gaseous products were quantified by an in-line gas-chromatography (GC) during
operation while liquid products were collected and analyzed using nuclear magnetic resonance
(NMR). Depending on the source, the concentration of SO2 in flue gas is ~100 ppm (~0.01%);
however, SO2 concentration can be as high as ~10000 ppm (~1%) in concentrated CO2 streams
that have not been desulfurized.20 As a conservative limit, a stream of 1% SO2 was chosen for
initial studies.
Influence of SO2 on Ag and Sn-catalyzed CO2 electroreduction
The influence of SO2 on CO2 (CO2 + SO2) electroreduction was first studied on Ag and Sn
catalysts, as these materials are widely studied and commercial devices are currently being
developed for the selective electrochemical conversion of CO2 to CO and formic acid.24,25 As
shown in Figure 2, when pure CO2 was initially fed, indeed high selectivities (>85%) toward CO
and formate, respectively, were achieved on Ag and Sn catalysts in near-neutral electrolyte. The
fluctuation in voltage is due to the accumulation of gas products as bubbles and eventual flush out
of the catholyte chamber. However, when both catalysts were subjected to a stream of 1% SO2
(Figure 2, red region), the total Faradaic efficiencies toward CO2 reduction products decreased,
and the remaining charge balance is likely attributed to the reduction of SO2 since it is
thermodynamically more favorable to reduce SO2 than CO2 (e.g. SO2 + 4H+ + 4e-  S + 2H2O Eo
= 0.50 V, CO2 + 2H+ + 2e-  CO + H2O Eo = -0.11 V, CO2 + 2H+ + 2e-  HCOOH Eo = -0.25 V).
There is also a large overpotential to drive SO2 reduction under CO2 electrolysis conditions.
Previous ultra-high vacuum studies have also indicated the facile adsorption of SO2 on transition
metal surface and in most cases, spontaneous catalyzed-activation of SO2 near room
4
ACS Paragon Plus Environment

Page 5 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
temperature.26,27 To further elucidate the adsorption of SO2 compared to CO2, Gibbs free energy
of adsorption as well as binding energies were calculated (Table S3), and calculations indeed show
that SO2 adsorption is stronger on all surfaces (Ag, Sn, and Cu) compared to that of CO2. This
further suggests the preferential reduction of SO2 over CO2. XPS measurements confirmed the
reduction of SO2 toward the formation of metal sulfide species (Figures S5&6, Table S4), and XPS
measurements at each time point were obtained from separate electrolysis experiments with new
electrodes operating at different electrolysis durations. Based on the peak positions at 161.7 and
161.6 eV, the metal sulfides were identified as Ag2S and SnS2 with surface concentrations of ~35.6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
at.% and ~3.3 at.% on Ag and Sn catalysts, respectively.28-30 Sulfite (SO3 ) and sulfate (SO4 )
were also detected. The former is likely formed during sample handling from residual surface
metal oxides reacting with SO2 in the presence of water when no potential is applied, and the latter
is likely formed from the natural oxidation of metal sulfides.31 The observed shifts in Ag peak
position to lower binding energies are likely due to the presence of S species, since Ag has a lower
electronegativity than S (S: 2.59, and Ag: 1.87).32,33 On the contrary, the shifts in Sn peak position
are due to different concentration of tin oxides (SnO/SnO2) in each sample as Sn rapidly oxidizes
in air during sample handling. Due to the resolution of the spectrometer, accurately differentiating
these two oxides is difficult. The formation of surface metal sulfides (Ag2S and SnS2) was also
confirmed with high angle annular dark field scanning transmission electron microscopy
(HAADF-STEM) and electron energy loss spectrum (EELS) imaging (Figures S7&8), which
shows S species positioned near the edges of the particles. Bright field (BF)-transmission electron
microscopy (TEM) imaging and the corresponding selected area electron diffraction (SAED)
indeed show the formation of SnS2 (Figure S8f). In the case of Ag (Figure S7f), diffraction patterns
associated with Ag2S was not observed, indicating that structure of Ag2S is likely amorphous. This
observation was further confirmed with high resolution (HR)-TEM (Figure S9), showing an
amorphous layer on the surface of the Ag catalyst. For all TEM work, FAA ionomer binder was
used instead of Nafion, as Nafion has sulfonic functional groups that would convolute the detection
of S species associated to the metal sulfides. SEM images (Figures S10&11), after 1 hour of
electrolysis with both catalysts subjected to a feed of 1% SO2 for 30 minutes, show the formation
of bulky particles with smooth surfaces, a common characteristic of metal sulfides. Future work
using density functional theory (DFT) can provide additional insight on the mechanism of metal
sulfide formation during CO2 electrolysis. It must be noted, that the exact surface characteristic
under reaction condition remains unclear due to the lack of in-situ surface sensitive methods and
the limitation of ex-situ measurements. Gas chromatography-mass spectrometry (GC-MS) was
also employed to detect other possible sulfur-containing species; however, the only sulfur-
containing species detected was unreacted SO2. It is likely that other residual species were
produced below the detection limit and cannot be easily detected by GC-MS or NMR.
2-
2-
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Performance of CO2 + SO2 electrolysis over (a) Ag and (b) Sn catalysts at constant
current density of 100 mA cm-2 in 1M KHCO3 over the span of 3 hours. Corresponding Faradaic
efficiencies can be found in Supporting Information (Tables S1&2).
Interestingly, when the feed of SO2 was stopped, the CO2 reduction performances
recovered on both catalysts, and in the case of Sn, a slight improvement in formate Faradaic
efficiency was observed. A recent study has suggested that doping Sn surfaces with sulfur (S)
atoms, and then selectively removing S atoms can create highly active undercoordinated Sn sites
for improved CO2 reduction to formate.30 Thermodynamically, Ag2S and SnS2 are not stable under
CO2 electrolysis conditions, and this was confirmed with post-reaction XPS that shows a decrease
in metal sulfides (Table S4). Particle aggregation can be seen in post-reaction SEM images, which
is likely due to surface rearrangement from the reduction of metal sulfides (Figures S10&S11).
The SO2 + CO2 experiments suggest that the presence of small amounts of metal sulfides (Ag2S
and SnS2) on Ag and Sn catalysts do not shift the CO2 reduction product selectivities, as the
dominant CO2 reduction products remained as CO and formate, respectively. However, the
reduction of the total CO2 Faradaic efficiency indicates that SO2 impurity can reduce the efficiency
of converting CO2, since the reduction of SO2 is more favorable than CO2. Sn catalyst was also
examined under longer SO2 exposure (Figure S12), and a plateauing of total Faradaic efficiency
was observed (~50%), suggesting a steady-state competitive reaction between SO2 and CO2
reduction. To further understand the effect of SO2 near flue gas concentrations, both Ag and Sn
catalysts were subjected to 0.01% SO2 (Figure S13), and the results show that SO2 at lower
concentration does not significantly affect CO2 electroreduction performances for both catalysts.
Influence of SO2 on Cu-catalyzed CO2 electroreduction
Next, CO2 + SO2 electroreduction experiments were performed on Cu catalyst to study the
influence of SO2 on Cu-catalyzed CO2 reduction, as Cu is the only monometallic material that can
reduce CO2 to high-value C2+ products with appreciable selectivities. As shown in Figure 3a,
indeed Cu can convert CO2 to an array of products including CO, formate, ethylene (C2H4), ethanol
(EtOH), n-propanol (PrOH), and a small amount of acetate and methane (CH4). Long-term CO2
electrolysis over Cu catalyst demonstrates a stable performance of converting CO2 to C2+ products
(Figure S14). When the Cu catalyst was subjected to 1% SO2 (Figure 3a, red region), as similar to
6
ACS Paragon Plus Environment

Page 7 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
both Ag and Sn catalysts, the total CO2 reduction Faradaic efficiency decreased, likely due to
preferential reduction of SO2. Interestingly, the CO2 reduction product selectivity shifted, and
formate became the dominant CO2 reduction product, while the other CO2 reduction products were
suppressed. XPS analysis (Figures 3b&c) shows the formation of metal sulfide (Table S4) after
one hour of electrolysis and the peak position at ~162.0 eV was assigned to Cu2S with a surface
concentration of ~12 at.%.34 This peak assignment was also verified with commercial Cu2S and
copper(II) sulfide (CuS) particles (Figure S15). Again, sulfite and sulfate were also detected likely
due to sample handling. Unreacted SO2 was the only sulfur-containing compound that was
detected with GC-MS; however, there is likely residual compounds produced under the detection
limit. The formation of aggregated bulky-like particles after 1 hour of electrolysis, subjected to a
feed of 1% SO2 for 30 minutes, can be seen (Figure S16), and this observation was further
confirmed with BF-TEM imaging (Figure 3d). The corresponding SAED pattern shows the
presence of residual Cu2S, metallic Cu, and CuO that was likely formed from air exposure during
sample handling. Similar to the Ag and Sn catalysts, HAADF-STEM and the corresponding EELS
imaging (Figure 3e-j) reveal highly disperse S species positioned near the edges of the particles,
suggesting that formation of Cu2S likely resides near the surface. This observation was also
confirmed with HR-TEM imaging and the corresponding fast Fourier transform (FTT) pattern
(Figure S17), also showing the formation of Cu2S near the particle surface. Additional HAADF-
STEM and EELS mapping can be found in the Supporting Information (Figure S18). Again, FAA
ionomer binder was used instead of Nafion for TEM work. On the as-prepared Cu catalyst, there
were no S species detected (Figure S19). The utilization of FAA ionomer showed a similar trend
in CO2 + SO2 electroreduction as Nafion, and thus, the ionomer is likely not the cause of the
observed shift in selectivity (Figure S20). Again, the exact surface characteristic under reaction
condition remains unclear due to the lack of in-situ surface sensitive methods to examine Cu
particles under CO2 electrolysis conditions while operating at high current densities. However,
Chorkendorff, Jaramillo, and co-workers very recently demonstrated that the surface structure of
polycrystalline Cu thin film is fully converted to the metallic phase under CO electroreduction
conditions by using in-situ grazing incident X-ray diffraction with a synchrotron radiation source.35
Future work employing similar techniques could further elucidate the surface structure of Cu in
the presence of SO2 under CO2 electrolysis conditions, and this will gain additional insight on the
formation of surface Cu2S.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. (a) Performance of Cu catalyst at constant current density of 100 mA cm-2 in 1M KHCO3
with a 1% SO2 feed over the span of 5 hours. Corresponding Faradaic efficiencies can be found in
Supporting Information (Table S5). XPS spectrum of (b) Cu and (c) S region at various time points.
(d) BF-TEM image. Inset: SAED pattern, (e-f) HAADF-STEM images, and (g-j) EELS mapping
of Cu catalyst after 1 hour of electrolysis in CO2 + SO2 experiment, subjected to a 1% SO2 feed
for 30 minutes.
Interestingly, when the SO2 feed was stopped, formate remained as the dominant CO2
reduction product, and an incremental recovery was observed as the selectivity toward CO
production slowly increased, followed by C2H4 production. Mechanistically, this observation is
consistent with literature as it is well known that the pathway toward C2+ products goes through a
CO intermediate on Cu-catalyzed CO2 electroreduction, while the pathway toward formate is a
dead-end.36 Therefore, the observed trend in recovery is CO followed by C2+ products. In addition,
an enhancement in hydrogen evolution was also observed during and after SO2 injection. Similar
to Ag2S and SnS2, Cu2S is also thermodynamically unstable under CO2 electrolysis conditions.
However, even after 5 hours of electrolysis, Cu2S was still detected on the Cu surface (Figures 3c
& S21), indicating that the reduction of Cu2S is kinetically slow. The presence of Cu2S may be the
reason for the inability of the Cu catalyst to recover its initial performance, and the observation in
the current study is consistent with previous works that have shown that S-modified nanostructured
Cu catalysts can selectively convert CO2 to formate.37,38 In addition, commercial Cu2S particles
were also examined and the dominant CO2 reduction product was formate (Figure S22), further
supporting that surface Cu2S is likely responsible for the change in selectivity.
8
ACS Paragon Plus Environment

Page 9 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
To further probe the sensitivity of the Cu catalyst toward SO2 impurity, similar experiments
were conducted with lower SO2 concentrations (0.1% and 0.01%). In both cases, a shift in
selectivity toward formate was observed (Figures S23&24), and XPS measurements confirmed the
presence of Cu2S (Table S6). With lower SO2 concentrations, there was a relatively faster rate of
recovery; however full recovery was not achieved even after 5 hours of electrolysis. Again, an
incremental recovery of CO followed by C2+ products can be seen. A recovery up to ~10% and
~5% for CO and C2H4 Faradaic efficiencies, respectively, was achieved in 0.1% SO2 after 5 hours.
At lower concentration of 0.01% SO2, the recovery of CO and C2H4 quickly plateaued to those
same Faradaic efficiencies, and no further improvement was observed. Post-reaction XPS analysis
after 5 hours electrolysis shows similar Cu2S composition on both catalytic surfaces after 0.1%
and 0.01% SO2 treatments (Table S6), suggesting that there may be residual surface Cu2S that are
stable under CO2 electrolysis conditions. In an attempt to recover the initial performance after 1%
SO2, the flow-cell electrolyzer was operated at 1 A cm-2 for 30 minute under Ar to locally generate
a high-flux of hydrogen to potentially reduce the metal sulfides (Figure S25). Although slight
improvements were observed, a full recovery was not achieved, and post-reaction XPS analysis
also indicated the presence of Cu2S (Table S6). These observations suggest that Cu catalysts are
sensitive to SO2 and that residual Cu2S on the catalytic surface may dramatically impact the CO2
electroreduction performance. Therefore, it is concluded that the CO2 gas feed must be relatively
pure (SO2-free) for Cu-catalyzed CO2 electrolysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Density Functional Theory Calculations
To further gain insight on the shift in selectivity, DFT calculations were performed on S
modified Ag, Sn, and Cu catalysts, and the surfaces were modeled as S-doped Ag(111), Sn(111),
and Cu(111), respectively. The binding energies of the hydrogen evolution reaction intermediate
(e.g. *H) and CO2 reduction intermediates for CO production (e.g. *COOH) and formate
production (e.g. *OCHO) were calculated on clean and S-doped surfaces (Table S7). The DFT
optimized geometries on Cu(111) and S-doped Cu(111) (Figure 4) show that *H prefers to bind at
a hollow site; *COOH binds at a top site via C atom and *OCHO binds at top sites via two O
atoms. The binding geometries of these intermediates on undoped and S-doped (111) surfaces of
Ag and Sn were also found to be similar to those on Cu(111). It is noted that *OCHO is relatively
more stable than the *COOH intermediate on all surfaces and the stability of *OCHO increases in
the order: Sn(111) > Ag(111) > Cu(111).
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. DFT optimized binding configurations of (a) *H, (b) *COOH, and (c) *OCHO on
Cu(111), (d) *H, (e) *COOH and (f) *OCHO on S-doped Cu(111), and (g) *H, (h) *COOH, and
(i) *OCHO on Cu2S(100). Blue: Cu, yellow: S, red: O, brown: C and white: H.
The binding energies (Table S7) show that *OCHO is more stable than *COOH by 1.25
eV, 1.35 eV and 1.13 eV on Ag(111), Sn(111) and Cu(111), respectively. In contrast, the *OCHO
stability over *COOH (0.62 eV) decreases on S-doped Ag(111) compared to Ag(111). However,
the *OCHO stability over *COOH increases on S-doped Sn(111) (1.54 eV) and S-doped Cu(111)
(1.23 eV) compared to Sn(111) and Cu(111), respectively. To account for the formation of Cu2S
as experimentally observed, further DFT calculations were performed to calculate binding energies
of *H, *COOH and *OCHO (Figure 4) on the Cu2S(100) surface that has mixed Cu and S surface
termination. Consistent with the prediction on S-doped Cu(111), the results show that the *OCHO
stability over *COOH (1.24 eV; Table S7) is enhanced on Cu2S(100) compared to Cu(111). In
summary, the DFT calculations predict that surface S atoms on Cu surfaces selectively stabilize
10
ACS Paragon Plus Environment

Page 11 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
*OCHO, and thus the formation of formate is promoted as in agreement with the experimental
observations. A similar promotional effect on the formate selectivity is predicted on S-doped Sn
(111). However, such promotion is not predicted on S-doped Ag(111) since the *OCHO stability
over *COOH is less on S-doped Ag(111) than on Ag(111). It must be noted that there is a
discrepancy between experimental and computational predication on the Ag catalyst as DFT
predicts that formate is more favorable than CO formation; while experimentally, CO was
observed as the major CO2 reduction product. A similar discrepancy is also highlighted in a recent
work by Bohra et al.39 However, by incorporating solvation effects on transition states and surface
coverage of *H, *COOH and *OCHO, they successfully demonstrated the promotion of CO over
formate formation on the Ag surface. Future work should also focus on exploring these effects on
Sn- and Cu- catalyzed CO2 reduction.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. Conclusions and Outlook
In summary, the influence of SO2 impurity on Ag, Sn, and Cu-catalyzed CO2 electrolysis
was studied in a flow-cell electrolyzer. The results show that the presence of SO2 impurity reduces
the efficiency of converting CO2 as the reduction of SO2 is thermodynamically more favorable. In
the case of Ag and Sn, full recovery was observed after both catalysts were subjected to SO2 and
the selectivities of converting CO2 to CO and formate were preserved, respectively. However, in
the case of Cu, the catalytic performance was highly sensitive to SO2 and a shift in selectivity
toward formate, accompanied by a suppression of C2+ products, was observed. Computational
efforts suggest that residual Cu2S formed on the surface, as confirmed with XPS characterization,
is likely responsible for the shift in selectivity.
Our experimental observations further motivate the development of direct air capture
(DAC) technologies as the CO2 feedstock from these sources would be relatively cleaner than that
obtained from fossil-burning point sources such as power plants and chemical facilities. The main
challenge is the capture and concentration of CO2 from a dilute source, the air. Although DAC
technologies has yet been fully commercialized and previous cost estimation has indicated that
these technologies are too expensive (>$600 per ton of CO2) to be practical,40 small pilot-plant
facilities, such as the Climeworks AG facility in Switzerland with the ability to capture 900 tons
of CO2 annually,41 are currently being developed. A recent work by Keith and co-workers provided
the first comprehensive cost breakdown of a DAC facility, estimating the cost to be between $94
and $232 per ton of CO2.17 Another advantage of DAC is that it is not geographically limited, and
thus, this technology can be coupled with distributed renewable energy sources that is expected to
be inexpensive in the near future.1
Lastly, additional work should also focus on studying the influence of NOx as well as O2
impurities in CO2/CO electrolysis. Although the concentration of NOx and O2 may be trace amount
in the CO2 feed, these impurities may also have detrimental effect on CO2/CO electrolysis as
demonstrated by the current SO2 study. Similarly, both NOx and O2 could co-adsorb on the
catalytic sites under reducing potentials or interact with the electrolyte, potentially altering the
local environment or surface structure and consequently influencing catalytic performance. In
particular, it has been hypothesized that subsurface oxygen atoms in Cu-catalyzed CO2/CO
electroreduction may potentially enhance the formation of C2+ products,42 and it would be
interesting to determine if adsorbed oxygen species can influence the catalytic performance. Thus,
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 17
1
2
3
4
5
future work should focus on further understanding the influence of these gaseous impurities on
CO2/CO electroreduction activity.
6
7
8
9
4. Experimental Setup
Electrode preparation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Commercial Ag (<100 nm, 99.5%) and Cu (25nm) powders were purchased from Sigma
Aldrich. Commercial Sn (0.1 micron), IrO2 (99.99%), Cu2S (99.5%), and CuS (99.8%) powders
were purchased from Alfa Aesar. The catalyst inks were prepared by mixing 25 mg of metal
particles, 20 l of Nafion solution (5 wt.% in 50/50 water and isopropanol) and 3 ml of isopropanol.
휇
After the catalyst inks were sonicated for at least 30 minutes prior to drop casting, 0.3 mg cm-2 of
the metal particles were drop casted onto a Sigracet 29 BC GDL (Fuel Cell Store). Ag, Sn, and Cu
electrodes were used as cathodes while IrO2 electrodes were used as anodes.
CO2 electrolysis in flow-cell electrolyzer
The electrochemical measurements were conducted in a three chamber flow-cell
electrolyzer with channel dimensions of 2 cm
0.5 cm
0.15 cm. 1M KHCO3 (99.95%, Sigma
×
×
Aldrich) was fed to the catholyte and anolyte chamber at ~1 ml min-1 with peristaltic pumps. CO2
was fed to the flow-cell at 17 sccm via a Brooks GF40 mass flow controller while SO2 was fed
from a syringe via a syringe pump (Cole-Parmer) at 0.17, 0.017, and 0.0017 sccm. To obtain SO2
gas, a H-cell was continuously purged with SO2 (Matheson Gas) for 30 minutes and then a 10 mL
gas-tight syringe (#1010, Hamilton Company) was used to extract the gas. The cathode and anode
chambers were separated by a FAA-3 hydroxide exchange membrane (Fumatech). The
backpressure of the gas in the gas chamber was controlled via a backpressure controller (Cole-
Parmer).
Chronopotentiometry experiments were conducted via an Autolab PG128N. The half-cell
potentials were measured using an external Ag/AgCl reference electrode (Pine Research). The
resistance between the cathode and reference electrode was determined using the current-interrupt
technique prior to electrolysis, and the measured applied potential was iR-corrected after
electrolysis. The gas products were sampled directly to a Multiple Gas Analyzer #5 gas
chromatography system (SRI Instruments) equipped with a Molseive 5A and a HayeSep D
columns connected to a thermal conductivity detector (TCD) and a flame ionization detector (FID).
The electrolytes containing liquid products were collected at the exit of the flow cell for 200
seconds and were analyzed using 1H NMR with water suppression using a pre-saturation method
(Bruker AVIII 600 MHz NMR spectrometer). 500 l of collected sample was mixed with 100 l
휇
휇
of D2O containing 25 ppm (m/m) dimethyl sulphoxide (99.9%, Alfa Aesar) as the internal
standard. To prevent SO2 from entering the GC during SO2 injection, a base trap was positioned
between the outlet of the flow-cell electrolyzer gas chamber and the inlet of the GC. Ar was fed at
15 sccm through a tee-connector to push the gas products along with unreacted CO2 through the
base trap before entering the GC for quantification. In order to detect potential sulfur-containing
products, the outlet gas stream from the electrolyzer was connected to a gas tight batch cell and
was continuously purged. A gas-tight syringe (Hamilton Company) was used to draw and inject
12
ACS Paragon Plus Environment

Page 13 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
the gas sample into an integrated GC-MS (Aglient 59771A) system equipped with a DB-FFAP
column and a mass spectrometry system. The mass fragmentation patterns were compared to those
of the National Institute of Standards and Technology (NIST) database. As for the catalyst
recovery, a National Instrument power source (RMX-4121) was used to achieve a current density
of 1 A cm-2. After 30 minutes of electrolysis under Ar, the flow-cell electrolyzer was disassembled,
the cathode was quickly dried due to severe flooding, and new anode was replaced prior to CO2
electrolysis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Material characterization
PXRD was conducted using Cu K radiation source (Bruker D8 Discover diffractometer).
훼
SEM images were obtained from Auriga 60 CrossBeam. XPS was performed with a K-Alpha X-
ray Photoelectron Spectrometer System (Thermo Fisher Scientific). All peaks were fitted with
Thermo Avantage software with the C 1s signal calibrated to 284.8 eV.
For TEM characterization, the samples were thoroughly washed with DI water
immediately after reaction to remove the residual electrolyte and then gently abraded from the
GDL support to acquire particles for the TEM experiment. In order to distinguish the sulfur atoms
that correspond to Cu2S, FAA ionomer binder was used instead of Nafion binder, since Nafion has
sulfonic functional groups. Bright field (BF), high-resolution transmission electron microscopy
(HRTEM) images and selected area electron diffraction pattern (SAED) were performed with a
transmission electron microscope (JEM-2100F, JEOL) at an acceleration voltage of 200 kV. High
angle annular dark field scanning transmission electron microscopy (HAADF-STEM) images and
electron energy loss spectrum (EELS) were obtained on a Hitachi HD2700C STEM with a probe
aberration corrector.
Computational methods
Spin polarized periodic density functional theory (DFT)43,44 calculations were performed at
the GGA level within the PAW-PW91 formalism45,46 using the Vienna Ab-Initio Simulation
Package (VASP) code.47,48 The total energy calculations were performed using a 3 × 3 × 1
Monkhorst-Pack grid49 and a plane wave cut-off energy of 400 eV.
Cu(111), Ag(111) and Sn(111) surfaces were modeled using a four layer 3 × 3 surface slabs.
One surface Cu, Ag, and Sn atom was replaced with a S atom to model S-doped Cu(111), Ag(111)
and Sn(111) surfaces, respectively. A vacuum layer of ~15 Å thick was added in the slab cell along
the direction perpendicular to the surface to minimize the artificial interactions between the surface
and its periodic images. During geometry optimization, atoms in the bottom two layers were fixed
and all other atoms including adsorbates were allowed to relax until the force on each ion was
smaller than 0.02 eV/Å. The binding energy (BE) of an adsorbate was calculated as:
BE(adsorbate) = E(slab + adsorbate) ‒ E(slab) ‒ E(adsorbate)
where E(slab + adsorbate), E(slab) and E(adsorbate) are the total energy of slab with adsorbate, the energy of
clean slab and the energy of adsorbate in the gas phase, respectively. E(H) is taken as one-half the
total energy of the H2 molecule during the calculation of hydrogen binding energy.
The free energy of adsorption of SO2 and CO2 on Cu(111), Ag(111), and Sn(111) at T =
298.15 K was calculated as:
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 17
1
2
3
4
5
∆Gad = G(slab + adsorbate) – G(slab) – G(adsorbate)
6
7
8
9
where G = E(DFT) + ZPE - T∆S. E(DFT) is the total energy obtained from DFT calculations, ZPE
and T∆S are zero point energy and entropic contributions, respectively. The DFT calculations were
performed to calculate ZPE of adsorbed species. ZPE and T∆S values of gas phase species were
taken from NIST database.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Acknowledgments
This material is based upon work supported by the Department of Energy under Award Number
DE-FE0029868. The authors at the University of Delaware also thank the National Science
Foundation for financial support (Award No. CBET-1803200). DFT calculations were supported
by the US Department of Energy, Catalysis Program (Grant No. DE-FG02-13ER16381).
This research used resources of the Center for Functional Nanomaterials, which is a U.S. DOE
Office of Science Facility, at Brookhaven National Laboratory under contract No. DE-
SC0012704.
Supporting Information
Schematic of flow-cell electrolyzer, SEM and TEM images, XPS analysis, tables of Faradaic
efficiencies, and theoretical calculations
References
(1) Haegel, N. M.; Margolis, R.; Buonassisi, T.; Feldman, D.; Froitzheim, A.; Garabedian, R.;
Green, M.; Glunz, S.; Henning, H. M.; Holder, B.; Kaizuka, I.; Kroposki, B.; Matsubara,
K.; Niki, S.; Sakurai, K.; Schindler, R. A.; Tumas, W.; Weber, E. R.; Wilson, G.;
Woodhouse, M.; Kurtz, S. Science 2017, 356, 141.
(2) Seh, Z. W.; Kibsgaard, J.; Dickens, C. F.; Chorkendorff, I. B.; Norskov, J. K.; Jaramillo,
T. F. Science 2017, 355 eaad4998.
(3) Lu, Q.; Rosen, J.; Zhou, Y.; Hutchings, G. S.; Kimmel, Y. C.; Chen, J. G. G.; Jiao, F.
Nature Communications 2014, 5 3242.
(4) Rosen, J.; Hutchings, G. S.; Lu, Q.; Forest, R. V.; Moore, A.; Jiao, F. ACS Catalysis 2015,
5, 4586.
(5) Ripatti, D. S.; Veltman, T. R.; Kanan, M. W. Joule 2019, 3, 240.
(6) Delacourt, C.; Ridgway, P. L.; Kerr, J. B.; Newman, J. Journal of the Electrochemical
Society 2008, 155, B42.
(7) Kim, B.; Hillman, F.; Ariyoshi, M.; Fujikawa, S.; Kenis, P. J. A. Journal of Power
Sources 2016, 312, 192.
(8) Verma, S.; Lu, X.; Ma, S. C.; Masel, R. I.; Kenis, P. J. A. Physical Chemistry Chemical
Physics 2016, 18, 7075.
(9) Verma, S.; Hamasaki, Y.; Kim, C.; Huang, W. X.; Lu, S.; Jhong, H. R. M.; Gewirth, A.
A.; Fujigaya, T.; Nakashima, N.; Kenis, P. J. A. ACS Energy Letters 2018, 3, 193.
(10) Ma, S. C.; Sadakiyo, M.; Luo, R.; Heima, M.; Yamauchi, M.; Kenis, P. J. A. Journal of
Power Sources 2016, 301, 219.
(11) Hori, Y.; Wakebe, H.; Tsukamoto, T.; Koga, O. Electrochimica Acta 1994, 39, 1833.
14
ACS Paragon Plus Environment

Page 15 of 17
Journal of the American Chemical Society
1
2
3
4
5
(12) Hori, Y.; Konishi, H.; Futamura, T.; Murata, A.; Koga, O.; Sakurai, H.; Oguma, K.
Electrochimica Acta 2005, 50, 5354.
(13) Wuttig, A.; Surendranath, Y. ACS Catalysis 2015, 5, 4479.
(14) Weng, Z.; Zhang, X.; Wu, Y. S.; Huo, S. J.; Jiang, J. B.; Liu, W.; He, G. J.; Liang, Y. Y.;
Wang, H. L. Angewandte Chemie-International Edition 2017, 56, 13135.
(15) Zhai, Y. M.; Chiachiarelli, L.; Sridhar, N. ECS Transactions 2009, 19, 1.
(16) Sanz-Perez, E. S.; Murdock, C. R.; Didas, S. A.; Jones, C. W. Chemical Reviews 2016,
116, 11840.
(17) Keith, D. W.; Holmes, G.; Angelo, D. S.; Heidel, K. Joule 2018, 2, 1573.
(18) Koytsoumpa, E. I.; Bergins, C.; Kakaras, E. Journal of Supercritical Fluids 2018, 132, 3.
(19) Wu, X. M.; Yu, Y. S.; Qin, Z.; Zhang, Z. X. In 12th International Conference on
Greenhouse Gas Control Technologies, Ghgt-12; Dixon, T., Herzog, H., Twinning, S.,
Eds. 2014; Vol. 63, p 1339.
(20) Last, G. V.; Schmick, M. T. Identification and Selection of Major Carbon Dioxide Stream
Compositions, 2011.
(21) Lv, J. J.; Jouny, M.; Luc, W.; Zhu, W. L.; Zhu, J. J.; Jiao, F. Advanced Materials 2018, 30
1803111.
(22) Dinh, C. T.; Burdyny, T.; Kibria, M. G.; Seifitokaldani, A.; Gabardo, C. M.; de Arquer, F.
P. G.; Kiani, A.; Edwards, J. P.; De Luna, P.; Bushuyev, O. S.; Zou, C. Q.; Quintero-
Bermudez, R.; Pang, Y. J.; Sinton, D.; Sargent, E. H. Science 2018, 360, 783.
(23) Dinh, C. T.; de Arquer, F. P. G.; Sinton, D.; Sargent, E. H. ACS Energy Letters 2018, 3,
2835.
(24) Liu, Z. C.; Yang, H. Z.; Kutz, R.; Masel, R. I. Journal of the Electrochemical Society
2018, 165, J3371.
(25) Yang, H. Z.; Kaczur, J. J.; Sajjad, S. D.; Masel, R. I. ECS Transactions 2017, 77, 1425.
(26) Haase, J. Journal of Physics-Condensed Matter 1997, 9, 3647.
(27) Rodriguez, J. A.; Jirsak, T.; Chaturvedi, S.; Hrbek, J. Journal of the American Chemical
Society 1998, 120, 11149.
(28) Ghafoor, S.; Ata, S.; Mahmood, N.; Arshad, S. N. Scientific Reports 2017, 7 255.
(29) Hota, G.; Idage, S. B.; Khilar, K. C. Colloids and Surfaces a-Physicochemical and
Engineering Aspects 2007, 293, 5.
(30) Zheng, X. L.; De Luna, P.; de Arquer, F. P. G.; Zhang, B.; Becknell, N.; Ross, M. B.; Li,
Y. F.; Banis, M. N.; Li, Y. Z.; Liu, M.; Voznyy, O.; Dinh, C. T.; Zhuang, T. T.; Stadler,
P.; Cui, Y.; Du, X. W.; Yang, P. D.; Sargent, E. H. Joule 2017, 1, 794.
(31) Aastrup, T.; Wadsak, M.; Leygraf, C.; Schreiner, M. Journal of the Electrochemical
Society 2000, 147, 2543.
(32) Mann, J. B.; Meek, T. L.; Knight, E. T.; Capitani, J. F.; Allen, L. C. Journal of the
American Chemical Society 2000, 122, 5132.
(33) Allen, L. C. Journal of the American Chemical Society 1989, 111, 9003.
(34) Xie, Y.; Riedinger, A.; Prato, M.; Casu, A.; Genovese, A.; Guardia, P.; Sottini, S.;
Sangregorio, C.; Miszta, K.; Ghosh, S.; Pellegrino, T.; Manna, L. Journal of the American
Chemical Society 2013, 135, 17630.
(35) Scott, S. B.; Hogg, T. V.; Landers, A. T.; Maagaard, T.; Bertheussen, E.; Lin, J. C.; Davis,
R. C.; Beeman, J. W.; Higgins, D.; Drisdell, W. S.; Hahn, C.; Mehta, A.; Seger, B.;
Jaramillo, T. F.; Chorkendorff, I. Acs Energy Letters 2019, 4, 803.
(36) Ren, D.; Fong, J. H.; Yeo, B. S. Nature Communications 2018, 9 925.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 17
1
2
3
4
5
6
(37) Shinagawa, T.; Larrazabal, G. O.; Martin, A. J.; Krumeich, F.; Perez-Ramirez, J. ACS
Catalysis 2018, 8, 837.
(38) Deng, Y. L.; Huang, Y.; Ren, D.; Handoko, A. D.; Seh, Z. W.; Hirunsit, P.; Yeo, B. S.
ACS Applied Materials & Interfaces 2018, 10, 28572.
7
(39) Bohra, D.; Ledezma-Yanez, I.; Li, G. N.; de Jong, W.; Pidko, E. A.; Smith, W. A.
Angewandte Chemie-International Edition 2019, 58, 1345.
(40) Lackner, K. S.; Brennan, S.; Matter, J. M.; Park, A. H. A.; Wright, A.; van der Zwaan, B.
Proceedings of the National Academy of Sciences of the United States of America 2012,
109, 13156.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(41) www.climeworks.com
(42) Eilert, A.; Cavalca, F.; Roberts, F. S.; Osterwalder, J.; Liu, C.; Favaro, M.; Crumlin, E. J.;
Ogasawara, H.; Friebel, D.; Pettersson, L. G. M.; Nilsson, A. Journal of Physical
Chemistry Letters 2017, 8, 285.
(43) Hohenberg, P.; Kohn, W. Physical Review B 1964, 136, B864.
(44) Kohn, W.; Sham, L. J. Physical Review 1965, 140, A1133.
(45) Blochl, P. E. Physical Review B 1994, 50, 17953.
(46) Perdew, J. P.; Wang, Y. Physical Review B 1992, 45, 13244.
(47) Kresse, G.; Furthmuller, J. Computational Materials Science 1996, 6, 15.
(48) Kresse, G.; Hafner, J. Physical Review B 1993, 48, 13115.
(49) Monkhorst, H. J.; Pack, J. D. Physical Review B 1976, 13, 5188.
16
ACS Paragon Plus Environment

Page 17 of 17
Journal of the American Chemical Society
1
2
3
4
TOC Figure:
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
ACS Paragon Plus Environment