Molecular Scaffolding Strategy with Synergistic Active Centers to Facilitate Electrocatalytic CO2 Reduction to Hydrocarbon/Alcohol

Article abstract of DOI:10.1021/jacs.7b10817

A major impediment to the electrocatalytic CO₂ reduction reaction (CRR) is the lack of electrocatalysts with both high efficiency and good selectivity toward liquid fuels or other valuable chemicals. Effective strategies for the design of electrocatalysts are yet to be discovered to substitute the conventional trial-and-error approach. This work shows that a combination of density functional theory (DFT) computation and experimental validation of molecular scaffolding to coordinate the metal active centers presents a new molecular-level strategy for the development of electrocatalysts with high CRR selectivity toward hydrocarbon/alcohol. Taking the most widely investigated Cu as a probe, our study reveals that the use of graphitic carbon nitride (g-C₃N₄) as a molecular scaffold allows for an appropriate modification of the electronic structure of Cu in the resultant Cu-C₃N₄ complex. As a result, the adsorption behavior of some key reaction intermediates can be optimized on the Cu-C₃N₄ surface, which greatly benefits the activation of CO₂ and leads to a more facile CO₂ reduction to desired products as compared with those on the Cu(111) surface and other kinds of Cu complexes formed on nitrogen-doped carbons. Remarkably, different from the most studied elementary metal surfaces, an intramolecular synergistic catalysis with dual active centers was for the first time observed on the Cu-C₃N₄ complex model, which possesses a unique capability to generate C₂ products. A good agreement between electrochemical measurements and the DFT analysis of the CRR has been achieved on the basis of the newly designed and synthesized Cu-C₃N₄ electrocatalyst.

Full text of DOI:10.1021/jacs.7b10817

Subscriber access provided by READING UNIV
Article
Molecular Scaffolding Strategy with Synergistic Active Centers to
Facilitate Electrocatalytic CO2 Reduction to Hydrocarbon/Alcohol
Yan Jiao, Yao Zheng, Ping Chen, Mietek Jaroniec, and Shi-Zhang Qiao
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b10817 • Publication Date (Web): 19 Nov 2017
Downloaded from http://pubs.acs.org on November 19, 2017
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 free 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 accessible to all
readers and 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.
Journal of the American Chemical Society 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 8
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Molecular Scaffolding Strategy with Synergistic Active Cen-
ters to Facilitate Electrocatalytic CO2 Reduction to Hydrocar-
bon/Alcohol
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
Yan Jiao1,‡, Yao Zheng1,‡, Ping Chen1,2,‡, Mietek Jaroniec,3 ShiꢀZhang Qiao1,4*
1 School of Chemical Engineering, The University of Adelaide, Adelaide, South Australia 5005, Australia
2 School of Chemistry and Chemical Engineering, Anhui University, Hefei, P. R. China
3 Department of Chemistry and Biochemistry, Kent State University, Kent, Ohio 44242, USA
4 School of Materials Science and Engineering, Tianjin University, Tianjin 300072, China
ABSTRACT: A major impediment of the electrocatalytic CO2 reduction reaction (CRR) is the lack of electrocatalysts with both
high efficiency and good selectivity toward liquid fuels or other valuable chemicals. Effective strategies for the design of electroꢀ
catalysts are yet to be discovered to substitute the conventional trialꢀandꢀerror approach. This work shows that a combination of the
density functional theory (DFT) computation and experimental validation on molecular scaffolding to coordinate the metal active
centers presents a new molecularꢀlevel strategy for the development of electrocatalysts with high CRR selectivity toward hydrocarꢀ
bon/alcohol. Taking the most widely investigated Cu as a probe, our study reveals that the use of graphitic carbon nitride (gꢀC3N4)
as a molecular scaffold allows for an appropriate modification of the electronic structure of Cu in the resultant CuꢀC3N4 complex.
As a result, adsorption behavior of some key reaction intermediates can be optimized on the CuꢀC3N4 surface, which greatly beneꢀ
fits the activation of CO2 and leads to a more facile CO2 reduction to desired products as compared with those on the Cu(111) surꢀ
face and other kinds of Cu complexes formed on nitrogenꢀdoped carbons. Remarkably, different from the most studied elementary
metal surfaces, an intramolecular synergistic catalysis with dual active centers was for the first time observed on the CuꢀC3N4 comꢀ
plex model, which possesses unique capability for the generation of C2 products. A good agreement between electrochemical
measurements and the DFT analysis of CRR has been achieved based on the newly designed and synthesized CuꢀC3N4 electrocataꢀ
lyst.
INTRODUCTION
level understanding of the overall CRR mechanism is highly
desired.
Carbon dioxide reduction reaction (CRR) by electrocatalysis
holds a great promise to produce sustainable carbonꢀbased
fuels from renewable resources.1 With the input from sustainꢀ
able energy such as solar and wind, this energy conversion
process utilize carbon as an energy carrier to produce fuels,
addressing the storage issues of the often intermittent reꢀ
sources.2 Because of the inert nature of CO2 molecule, this
normally exothermic process is naturally difficult to proceed
and is greatly dependent on the careful selection of highly
active electrocatalysts.3 Additionally, due to the complexity of
the charge transfer process occurred in CRR, a wide variety of
products is accessible including carbon monoxide (CO) and
formic acid (HCOOH) via a twoꢀelectron reduction pathway,
methanol (CH3OH) via sixꢀelectron reduction, and methane
(CH4) via a complete eightꢀelectron reduction.4 Practically, the
best solution remains the conversion of CO2 to CH3OH or,
further to C2 products like ethylene (C2H4) and ethanol
(C2H5OH) , as thus the established fuelꢀtransporting infrastrucꢀ
ture could be well utilized.5 However, the availability of suitaꢀ
ble electrocatalysts is still very limited.6 Generally, the selecꢀ
tivity of different reduction products strongly depends on both
extrinsic physical properties and intrinsic electronic structure
of the heterogeneous electrocatalyst used, which always act
together in determining the adsorption of intermediates and
activation energies for each reaction.7 This complexity makes
the rational design of suitable electrocatalysts for desired CRR
products a great challenge. Therefore, exploration of atomic
Until now, Cuꢀbased materials remain to be the most effiꢀ
cient and selective CRR electrocatalysts providing a wide vaꢀ
riety of products depending on their chemical and/or physical
properties.8 The density functional theory (DFT) computationꢀ
al studies of these catalysts revealed that their great perforꢀ
mance can be related to the appropriate adsorption strength of
key reaction intermediates.9 Therefore, the design and engiꢀ
neering of Cuꢀbased catalysts with suitable electronic structure
to achieve wider product distribution and/or higher selectivity
is still the mainstream research direction for CRR. Current
strategies to engineer Cu catalysts are mainly relying on the
modification of their physical structure and/or chemical comꢀ
position by straightforward methods. However, due to the
complex nature of CRR and the aforementioned complicated
relationship between the apparent activity/selectivity of a givꢀ
en electrocatalyst and its inherent adsorption energetics toward
key intermediates, almost all of the current approaches toward
development of CRR electrocatalysts are still via a trialꢀandꢀ
error method; an atomic scale design strategy is largely missꢀ
ing.
Due to the ability of transition metals to form coordination
complexes with ligands, coordinating Cu into molecular
frameworks could provide a platform for engineering novel
Cuꢀbased CRR electrocatalysts with unprecedented
properties.10 Previously, based on both theoretical and experiꢀ
mental studies we showed that coordinating a range of transiꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 8
tion metals (e.g., Fe, Co, Ni) into the framework of graphitic TkatchenkoꢀScheffler method was applied during all calculaꢀ
1
2
3
4
5
6
7
8
carbon nitride (gꢀC3N4) can introduce excellent electrocatalytic
activities toward two key oxygen electrode reactions (oxygen
reduction/evolution reactions) for fuel cells and water splitting
applications.11 A strong coordinating capability and particular
functional species of gꢀC3N4 framework make it a suitable new
platform for efficient metal coordination instead of traditional
nitrogenꢀdoped carbons. As regards CRR, in principle, gꢀC3N4
possesses at least two particular properties that differs it from
other nitrogenꢀcontaining carbon substrates. First, the abunꢀ
dant pyridinic nitrogen species in gꢀC3N4 framework show
strong affinity to CO2 under electrocatalytic conditions, which
is considered critical for CO2 activation.12 Second, carbon
species in gꢀC3N4 show high oxophilicity for adsorption of
oxygenꢀcontaining reaction intermediates (*OCHx, *OH, and
*O) during CRR, which appear in the second half of the reacꢀ
tion pathway, benefiting the generation of deep reduction
products.13 Therefore, gꢀC3N4 could serve as an ideal molecuꢀ
lar scaffold for Cu to form CuꢀC3N4 structure for efficient and
effective CRR.
tions to properly address the van der Waals interactions beꢀ
tween atoms.17
Three surface models were studied in this work: CuꢀC3N4,
CuꢀNC, and Cu(111) (for the purpose of comparison). For Cuꢀ
C3N4, the lattice parameters used were 7.089Å × 7.089Å ×
20Å, with the angle between the first two vectors being 120o.
The CuꢀNC model was based on a 5×5 graphene supercell,
and the lattice was optimized to be 12.329Å × 12.329Å × 20Å.
For copper single crystal, the lattice constant was optimized to
be 3.551Å, while the DOS computation was carried out for a
four layer slab model in the [111] direction, with 15Å space
between two slabs. The atomic configuration and electron
charge transfer due to the incorporation of copper to CuꢀC3N4
and CuꢀNC are shown in Figure S1 (supporting information).
Bader charge analysis shows that the electron transfers from
the embedded copper atom to gꢀC3N4 (0.751 e–) or nitrogenꢀ
doped graphene (0.929 e–) frameworks.
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
Reaction Intermediates. First, the CO2 adsorption on Cuꢀ
C3N4 was investigated because this is the first step that occurs
during CRR. The adsorption location and pattern often give
useful information about reaction steps, for example the potenꢀ
tial active sites. As shown in Figure S2, various positions for
CO2 adsorption were investigated and the possible adsorption
sites are listed. CO2 molecule prefers to adsorb on copper and
nitrogen atoms rather than on carbon atoms. Specifically, a
‘mock chemisorption’ exists on nitrogen atom, where CO2 is
adsorbed on CuꢀC3N4 via a carbonꢀnitrogen bond (Figure S2f),
indicating the ability of CuꢀC3N4 toward activating CO2 if
proper electrode potential is applied.12a The formation of this
configuration could be attributed to the electron accumulation
on nitrogen due to its higher electronegativity than that of surꢀ
rounding carbon atoms. The search for all other possible reacꢀ
tion intermediates (*COOH, *CO, *COH, *CH etc., see Table
S1 and S2 for the complete list of investigated states and corꢀ
responding relative stability) on CuꢀC3N4 and CuꢀNC also
begins with various different initial configurations. The free
energy corrections to each state are presented in Table S3 (for
adsorbed states) and Table S4 (for gas phase properties). The
free energy diagram of CO2 reduction on Cu(111) is taken
from reference 18. Equilibrium potentials of several key CRR
products are summarized in Table S5.
In this work, we explored different CO2 reduction pathꢀ
ways on CuꢀC3N4 structures (for comparison, we also studied
Cu supported on nitrogenꢀdoped graphene named as Cuꢀ
NC)14, by investigating a full range of possible reaction interꢀ
mediates. According to the thermodynamic analysis, CuꢀC3N4
shows two unique and critical features for CRR as compared
with Cu(111) surface and other kinds of nitrogenꢀcontaining
carbons with coordinated Cu complexes. First, gꢀC3N4 can
influence copper’s electronic structure by shifting its dꢀband
position toward Fermi level, therefore leading to a stronger
adsorption of intermediates and smaller energy loss toward
hydrogenation of CO2. Second and more importantly, different
from the widely studied bifunctional catalysts composed of a
metallic site and a support for CO/CO2 conversion,15 CuꢀC3N4
shows an intramolecular synergistic effect toward CRR, where
Cu and C act in synergy for different stages of the reaction. As
a result, an enhanced CO2 reduction to C2 species (e.g.,
C2H5OH and C2H6) was achieved in CuꢀC3N4 system due to
the unique dual active center synergistic catalysis. This therꢀ
modynamicꢀbased theoretical prediction was further validated
by electrochemical measurements performed on experimentalꢀ
ly synthesized CuꢀC3N4 catalysts, which showed a wider varieꢀ
ty of C2 products (C2H5OH, C2H6, and C2H4) and higher selecꢀ
tivity of deep reduction products (CH4 and CH3OH) in comꢀ
parison to conventional CuꢀNC counterparts.
Sample Synthesis and Characterization. To match the
models proposed in the computation, the synthesis of CuꢀC3N4
nanomaterials was conducted by mixing different amounts of
CuCl2 aqueous solutions with dicyandiamide (DCDA) in room
temperature. For instance, 1.3, 2.5, 4.8, and 10 mL of 0.1 M
CuCl2 solution with 0.66, 0.64, 0.60, 0.54 g of DCDA and 10
mL of water were selected to prepare the sample with theoretiꢀ
cal ratio of Cu: C3N4 = 1:20, 1:10, 1:5, 1:2, respectively. Difꢀ
ferent ratios were chosen to assure Cu ion coordination with gꢀ
C3N4 as efficient as possible. 50% mass concentration of acid
pretreated carbon black was applied at the same time as a supꢀ
port to enhance the electrode conductivity and to expose the
electrocatalytically active sites of CuꢀC3N4. After stirring
overnight, the mixture was lyophilized and annealed at 600 oC
in N2. The powder was subsequently washed twice with 1M
HNO3 and then several times with water to remove uncoordiꢀ
nated copper/copper oxide particle. The synthesis of CuꢀNC
was carried out by annealing the afterꢀwashed CuꢀC3N4 sample
at 900 oC in N2, during which the molecular scaffold of gꢀC3N4
METHODS
Computational Methods and Models. Electronic strucꢀ
ture calculations were carried out by DFT with PerdewꢀBurkeꢀ
Ernzerhof (PBE) exchangeꢀcorrelation functional using a proꢀ
jector augmentedꢀwave method by VASP code.16 For the
planeꢀwave expansion, a 500 eV kinetic energy cutꢀoff was
used after testing a series of different cutꢀoff energies. The
convergence criterion for electronic structure iteration was set
to be 10ꢀ5 eV and that for geometry optimization was set to be
0.01 eV/Å. A Gaussian smearing of 0.20 eV was applied durꢀ
ing the geometry optimization and for the total energy compuꢀ
tations, whilst for the accurate density of states (DOS) compuꢀ
tation a tetrahedron method with Blöchl correction was emꢀ
ployed. The Kꢀpoints were set to be 5×5×1 for the CuꢀC3N4
unit cell, 3×3×1 for grapheneꢀbased CuꢀNC model. Denser Kꢀ
points were used for the high quality DOS computations
(15×15×1 for CuꢀC3N4, and 9×9×1 for CuꢀNC). The
ACS Paragon Plus Environment

Page 3 of 8
Journal of the American Chemical Society
was decomposed and converted to nitrogenꢀdoped carbon with
bonded Cu.
The near edge Xꢀray absorption fine structure (NEXAFS)
and Xꢀray photoelectron spectroscopy (XPS) measurements
were carried out on soft Xꢀray spectroscopy beamline at the
Australian Synchrotron. In all NEXAFS scans, 50 meV energy
steps were used. In the XPS scans, the excitation energy was
700 eV for better signalꢀtoꢀnoise ratio and the Eꢀpass was set
to 5 eV for optimum energy resolution. The transmission elecꢀ
tron microscopy (TEM) was carried on FEI Tecnai G2 Spirit
TEM operated at 120 kV.
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
Electrochemical Measurements and Product Analysis.
The evaluation of CRR performance on two synthesized samꢀ
ples were done using a homemade Hꢀtype cell with two comꢀ
partments separated by the Nafion membrane. Each compartꢀ
ment contains 50 mL of electrolyte (0.1 M KHCO3) and about
50 mL headspace. The measurements were carried out using
CHI 750E electrochemical workstation at room temperature.
The electrocatalysts were loaded on the glassy carbon serving
as working electrode. A Pt plate and an Ag/AgCl in saturated
KCl solution were used as counter and reference electrodes,
respectively. Before each measurement, the cathodic comꢀ
partment of the cell was degassed and saturated with CO2 at 40
mL/min for at least 20 min. During the reduction, CO2 gas was
continuously bubbled into the cathodic compartment. The
gaseous products were delivered directly to the sampling loop
of an onꢀline gas chromatograph equipped with a thermal conꢀ
ductivity detector using Ar as a carrier gas for H2 analysis and
a flame ionization detector for CO, CH4, C2H4, and C2H6 analꢀ
ysis. The liquid products (e.g. HCOOH, CH3OH, and
C2H5OH) were detected by NMR (Bruker 400 MHz).
Figure 1 (a) Free energy diagram of CO2 reduction pathway to
CH4 on CuꢀC3N4, CuꢀNC, and Cu(111). The values for the
Cu(111) surface are taken from reference 18. Electron density
difference for CO adsorbed on (b) CuꢀC3N4 and (c) CuꢀNC. The
unit of isosurface values is shown in the figure. Green, blue, gold,
and red atoms are carbon, nitrogen, copper and oxygen, correꢀ
spondingly. Yellow is electron accumulation and cyan represents
electron depletion.
Though the first fourꢀelectron transfer steps proceed with
identical reaction intermediates on three copperꢀembedded
surfaces, the reaction pathway starts to diverge when the fifth
proton and electron pair is transferred to the surface. In the
case of CuꢀC3N4 and Cu(111), an *OCH3 group is formed with
oxygen as an anchoring atom on the surface of the aforemenꢀ
tioned catalysts, followed by the formation of an isolated CH4
molecule and adsorbed *O after the sixth proton/electron pair
transfer. As for the reduction pathway on the CuꢀNC surface,
the fifth step includes the generation of *CH2OH, followed by
OH desorption to form *CH2 and therefore two steps of hyꢀ
drogenation are needed to produce CH4 after the last proꢀ
ton/electron pair transfer, as shown in Figure S3b, blue line.
On this pathway, all free energy changes from 2nd to 8th charge
transfer process on CuꢀNC are smaller than that for the first
charge transfer, therefore the rateꢀdetermining step is the proꢀ
tonation of CO2 to *COOH. For the purpose of comparison,
Figure 1a depicts the same CRR pathway to CH4 on all three
surfaces. Despite of the different reduction pathway on the
three electrocatalysts, CuꢀC3N4 generally shows the strongest
binding to reaction intermediates and therefore the lowest free
energy pathway.
RESULTS AND DISCUSSION
Reaction Pathway. We obtained the configurations of all
possible reaction intermediates and their corresponding free
energies by DFT computations. Afterwards the lowest energy
reaction pathway – defined to be the pathway with lowest posꢀ
itive elementary free energy change between any two steps –
toward CH4 production is summarized in Figure 1a. In this
figure, the xꢀaxis represents the transferred numbers of proꢀ
ton/electron pairs to the electrode surfaces including CuꢀC3N4,
CuꢀNC, and traditional Cu(111) slab. Generally, the first few
reduction steps of CO2 to CH4 on these three surfaces follow
the same pathway, which starts with hydrogenation of an adꢀ
sorbed *CO2 to form *COOH. Following the OH desorption, a
*CO is left on the surface and the second protonation process
occurs to form *CHO. These two protonation processes are
traditionally considered as the rateꢀlimitingꢀsteps of CO2 reꢀ
duction; for example, on the Cu(111) slab surface, 0.9 and 0.8
eV free energy change could be observed for CO2 and *CO
hydrogenation, respectively.18 Meanwhile, on the CuꢀC3N4
surface, due to the lower free energy level of the state after the
first hydrogenation step (*COOH), the second hydrogenation
process (*CHO) becomes the only rateꢀlimitingꢀstep. The free
energy change of this step is 0.75 eV uphill, less endothermic
than on Cu(111) slab surface. As for the CuꢀNC surface, the
free energy change of the first hydrogenation step is higher
than that of the second one, with values of 1.34 eV and 0.44
eV, respectively, and therefore the first protonation is considꢀ
ered as the major reaction rateꢀdetermining step.
Following reference 9f we have also examined an alternaꢀ
tive reaction pathway for CH4 production on CuꢀC3N4 and Cuꢀ
NC. For the two evaluated pathways on CuꢀC3N4, as shown in
Figure S3a, the first half of these pathways (until the forth
proton/electron pair transfer) is the same as discussed in the
previous paragraph. The main pathway (black line) involves
the protonation of *CHO to form *OCH2 with the upcoming
hydrogen connecting the carbon atom, with a free energy
change of 0.43 eV for this step. An alternative reaction pathꢀ
way (red line) involves a different reaction intermediate folꢀ
lowing *CHO, where the upcoming hydrogen is connected to
O and forms *CHOH. The free energy change for this interꢀ
mediate state is 0.74 eV, higher than the previously mentioned
pathway. The remaining part of this alternative pathway is OH
desorption to form adsorbed *CH, followed by three steps of
hydrogenation and the final CH4 production. For the CuꢀNC
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 8
surface, as shown in Figure S3b, the alternative reaction pathꢀ higher than Cu(111), which suggests worse activity possessed
1
2
3
4
way considered in reference 9f (red line) is also higher in enꢀ
ergy than the lowest possible pathway (blue line). Therefore,
the aboveꢀmentioned pathways are higher in the free energy
and are thermodynamically unfavorable.
by CuꢀC3N4. Meanwhile, CuꢀNC shows significantly high
limiting potential toward all products, requesting higher overꢀ
potential to kickꢀstart the reactions and therefore poorer perꢀ
formance from the computational point of view.
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 (a) Summary of the onset potential of the products of
CO2 reduction as predicted by the reaction free energy diagrams
on three different surfaces. Detailed free energy diagram of (b)
hydrogen evolution and (c) CO2 reduction to CH3OH on three
copperꢀbased surfaces.
Figure 3 (a) Free energy diagram of CO2 reduction to CH4 on
different sites of CuꢀC3N4. Inset shows the migration of active
center from Cu to carbon. (b) Summary of the onset potential of
the products of CO2 reduction as predicted by the reaction free
energy diagrams on different sites of CuꢀC3N4. Smaller value
indicates the lower onset potential required, hence better perforꢀ
mance of this catalyst.
To explain the lower free energy possessed by reaction inꢀ
termediates on CuꢀC3N4 in comparison to that on CuꢀNC, the
electron density difference analysis was carried out using *CO
adsorption as an example. As regards the CuꢀC3N4 surface, at
an isosurface cutꢀoff of 2×10ꢀ3 e/Bohr3, electron transfers from
the copper present in CuꢀC3N4 and the carbon atom (from CO
molecule) to the CuꢀC bonding area between these two moieꢀ
ties, suggesting a strong interaction between these two parts
(Figure 1b). As a result, the CO triple bond weakens as indiꢀ
cated by the electron depletion alongside the bond, with its
length stretched from 112.8 pm to 115.6 pm (comparable to a
CO double bond). Additionally, COꢀCu bond length within
this configuration is 1.77 Å. While in striking contrast to the
CuꢀNC surface, the distance between carbon on CO and nitroꢀ
gen as the adsorption center is 3.42 Å, much longer than that
on CuꢀC3N4. A significant charge transfer between CuꢀNC
framework and CO is only visible at an isosurface value of
about 1.5×10ꢀ4 e/Bohr3, about one magnitude lower than that
on CuꢀC3N4 (Figure 1c). At the meantime the CO bond length
is 114.4 pm. These considerations show the strong adsorption
CO on CuꢀC3N4 and very weak one on CuꢀNC, which leads to
the much lower free energy level for the COꢀcontaining reacꢀ
tion intermediates as well as other reaction intermediates.
The detailed free energy diagrams that were used to obtain
the aforementioned limiting potentials toward several products
are shown in Figure 2 (b: H2 – via hydrogen evolution reaction
and c: methanol) and Figure S4 (HCOOH, CO, and C2H4). As
indicated by these plots, the free energy levels of all reaction
intermediates on CuꢀC3N4 are generally lower than those on
other two substrates, except for hydrogen evolution. Since the
hydrogen adsorption energy on CuꢀC3N4 is higher than that on
Cu(111), a weaker hydrogen evolution ability of this material
should be expected. In contrary to the gꢀC3N4 framework itꢀ
self, which shows a strong binding to H*, the hydrogen adꢀ
sorption free energy level is elevated and could be attributed to
the copper atom that is already occupying the triangular vaꢀ
cancy formed by surrounding sꢀtriazine units. This feature
would be particularly beneficial for the production of hydroꢀ
carbons, since hydrogen evolution reaction is the major comꢀ
peting reaction with CRR.19 Additionally by comparing the
onset potential required for the generation of formic acid and
carbon monoxide, the CO production requires higher overpoꢀ
tential than that of HCOOH, indicating lower selectivity toꢀ
ward CO. This difference is originated by a strong *CO adꢀ
sorption on CuꢀC3N4, which makes the CO desorption difficult
to proceed.
Onset Potential. When considering the full methane
pathway, the smallest negative potential at which the pathway
becomes exergonic (i.e., the limiting potential) for CuꢀC3N4,
CuꢀNC and Cu(111) is 0.75 V, 2.18 V, and 0.90 V respectiveꢀ
ly. This phenomenon, therefore suggests a better catalytic acꢀ
tivity of CuꢀC3N4 toward CO2 reduction to methane when usꢀ
ing limiting potential as an activity descriptor. The limiting
potential toward other products including H2, CH3OH,
HCOOH, CO, and C2H4, are summarized in Figure 2 (a). Genꢀ
erally, CuꢀC3N4 shows lowest required potential among the
three surfaces to open pathways toward deep reduced prodꢀ
ucts, which indicates smaller electrical energy loss and thereꢀ
fore higher efficiency in generating these products. In comparꢀ
ison, the overpotential required to produce CO on CuꢀC3N4 is
Dual Active Center Mechanism. During the exploration
of all possible reaction pathways on CuꢀC3N4 toward CH4
production, various sites were investigated. Generally, the
binding of reaction intermediates during the first half of the
reaction shows stronger adsorption on the copper atom than on
carbon and nitrogen sites of the CuꢀC3N4 substrate. Meanwhile
during the latter half of the reaction, the reaction intermediates
show stronger binding to the carbon site than on the copper
site. Therefore, the lowest free energy pathway toward CH4
production is given when these two sites are both considered
ACS Paragon Plus Environment

Page 5 of 8
Journal of the American Chemical Society
as a dualꢀactive site. As shown in Figure 3a, the dual active Origin of the Activity. In accordance to the above discusꢀ
1
site pathway leads to an overall limiting potential of 0.75 V. It
is worth noting that the limiting potential toward CH4 producꢀ
tion when considering Cu as the only active center for the
whole reaction is 1.14 V, while that for the carbon as the sole
active center is 1.48 V, both higher than the dualꢀactive site
pathway. The inset shows transfer of active center, which ilꢀ
lustrate the synergistic effect of Cu and gꢀC3N4 framework.
The origin of this behavior is that copper shows strong binding
to reaction intermediates having carbon as an anchoring atom
(e.g. *COOH, *CO, *CHO), while the carbon on gꢀC3N4
shows stronger binding to reaction intermediates having oxyꢀ
gen as the connecting atom (e.g. *OCH2, *OCH3, *O, and
*OH). Therefore, during the first half of the reaction, reaction
intermediates favor copperꢀcarbon bond; while during the
latter half of the reaction, the reaction intermediates are conꢀ
nected to the surface via a carbonꢀoxygen bond. We have also
quantitatively analyzed the limiting potential toward other
products, as shown in Figure 3b, and observed the same beꢀ
havior. Therefore, the unique structure of metal coordinated by
gꢀC3N4 framework leads to a synergistic reaction scheme that
leads to a more feasible reaction pathway induced by the miꢀ
gration of active center.
sion, a linear relationship can be observed between the
*COOH and *CO adsorption strengths, as shown in Figure
S6a, suggesting that the CO2 activation ability is proportional
to CO adsorption on these three surfaces. As indicated in the
aforementioned figure, the CO2 activation ability possessed by
Cu on CuꢀC3N4 is the highest among the three catalysts, indiꢀ
cating a strong interaction between carbon and the copper
atom. This observation also suggests that similar to the scaling
relationship between different metals, a counterpart relation
exists for the same metal species (for example, copper as
shown in this work) when embedded within different frameꢀ
works. The underlying reason for this scaling relationship, as
well as why CuꢀC3N4 possesses the strongest CO binding
strength, could be unveiled by density of states (DOS) analysis
performed on the three surfaces. Figure S6b shows the copper
dꢀorbital density of states on three surfaces, which is considꢀ
ered to play a determining role in interactions with surface
adsorbates. The distribution of states shifts toward the higher
energy region following the order of CuꢀNC → Cu(111) →
CuꢀC3N4, which is consistent with the observed order of the
binding energy strength, as shown in Figure S6c. The general
trend is that with a higher dꢀband position such as CuꢀC3N4, a
stronger adsorption is observed due to upꢀshifting the antiꢀ
bonding state, which leads to lower occupation (of the antiꢀ
bonding state), and results in weaker repulsion between these
two parts.
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
Inspired by this observation that Cu atom serves as the acꢀ
tive center for carbon (and Cꢀcontaining groups) adsorption,
while the adjacent carbon atom on gꢀC3N4 serves as the active
center for reaction intermediates having oxygen as the conꢀ
necting atom, we further explored the possibility of C2 pathꢀ
way on CuꢀC3N4. Here we present the free energy pathway of
CO2 reduction to CH3CH2OH as an example, as shown in Figꢀ
ure S5. The first five proton/electron transfers follow the same
pathway as that shown in Figure 1a, forming an adsorbed
*OCH3 group on the carbon atom. Where starting from the
sixth proton/electron pair transfer, the second CO2 adsorbs on
the copper atom along with a protonation process. The configꢀ
uration of consequently occurring reaction steps is shown in
Figure 4, where the second *COOH undergoes a series of
changes to form *CH3O on the copper atom. From that point, a
bond forms between the two carbonꢀ containing moieties to
form CH3CH2OH, leaving the surface adsorbed oxygen atom,
which is then protonated to form H2O. The existence of this
reaction pathway is due to the molecular scaffold coordinating
copper; the unsymmetrical feature of CuꢀC3N4 could lead to
the C2 pathway and therefore, deep reduced products. Correꢀ
spondingly, the following experimental electrochemical measꢀ
urements also validate that CuꢀC3N4 can produce more C2
species like C2H6, CH3CH2OH than the CuꢀNC sample with
single active centers under the same condition.
Figure 5 (a) TEM image of typical CuꢀC3N4 electrocatalyst supꢀ
ported on carbon black. Arrows indicated atomic Cu clusters. (b)
Comparison of Cu concentrations in different CuꢀC3N4 samples
before and after acid wash. (c) High resolution Cu2p XPS of Cuꢀ
C3N4 before and after acid wash. (d) Nitrogen Kꢀedge NEXAFS
of CuꢀC3N4 and pure gꢀC3N4. Arrows show the weak shoulders in
N Kꢀedge assigning to the CuꢀN interaction. Dotted lines show the
channels of photon energy in two samples.
Experimental Validation. To further provide experiꢀ
mental evidence of gꢀC3N4 scaffold’s unique role in the CRR
process revealed by theoretical calculations, we carried out the
synthesis of CuꢀC3N4 and CuꢀNC electrocatalysts with similar
molecular structures as investigated theoretically (see details
Figure 4 Key reaction intermediates of CH3CH2OH generation on
CuꢀC3N4.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 8
in experimental section). During the thermal treatment, DCDA More strikingly, CuꢀC3N4 samples demonstrate the ability of
1
2
3
4
5
6
7
8
was polymerized into gꢀC3N4 scaffold and simultaneously
coordinated with Cu ions (Figure 5a). After acid washing, all
four samples with different amounts of Cu precursors showed
similar Cu atomic concentration in the CuꢀC3N4 hybrid (Figure
5b), indicating that there is a saturated coordinating ability of
gꢀC3N4 scaffold to metal ions. Taking the 1:5 sample as an
example, after double nitric acid washing, the Cu concentraꢀ
tion in CuꢀC3N4 decreased from 3.1 to 0.2 at.% with almost
unchanged N concentration based on the XPS analysis. (Figꢀ
ure S7). This can be explained by the fact that most Cu in
freshly synthesized product is in the form of Cu(0) metal naꢀ
noparticles, which can be eliminated by washing with acid
solutions. Meanwhile, the remaining Cu species are in the
form of single atoms and clusters bonded with nitrogen atoms
in the gꢀC3N4 scaffold (Figure 5c,d). As can be seen from the
highꢀresolution XPS data, these Cu species possess higher
binding energies than the Cu nanoparticles. The higher valꢀ
ance states of Cu may be attributed to its interaction with gꢀ
C3N4 in the form of electron transfer from copper to nitrogen
atoms (Figure 5c). At the same time, by comparing with the
pure gꢀC3N4, the extra weak shoulders in nitrogen Kꢀedge
NEXAFS spectrum of CuꢀC3N4 also indicate nitrogen atoms,
more specifically, the pyridinic nitrogen in the triazine hetꢀ
erorings, able to accept extra charges from Cu atoms, which
results in a slightly negative shift in its photon energy profile
(Figure 5d). Therefore, analysis of Cu 2p XPS and N Kꢀ edge
NEXAFS spectra confirms that there is a strong chemical inꢀ
teraction (or bonds) between Cu and N atoms in the syntheꢀ
sized CuꢀC3N4 sample, which match the molecular configuraꢀ
tion used in the theoretical computation (Figure S1 in supportꢀ
ing information). It is worth noting that the atomic ratio of Cu
and N in all four samples is around 0.0073~0.0091, which is
equal to the ratio of triꢀsꢀtriazine units to Cu of 8.6 to 11.4
(theoretical value should be 1). This means that not all vacanꢀ
cies in the gꢀC3N4 framework are saturated by Cu species;
most of them are still in the pristine form of gꢀC3N4.
generating C2 species like C2H5OH, C2H4, and C2H6, which is
attributed to the unique dual active center mechanism by the
special configuration of Cu atom embedded in gꢀC3N4 scaffold
(Figure 4). These trends serve as a good agreement between
experimental measurement and the thermodynamicꢀbased
DFT study of these materials.
10
b
a
0
-5
9
8
6
4
2
0
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
-10
-15
-20
Cu-C3N4
Cu-N/C
@-1.6 V vs Ag/AgCl
-2.0 -1.8 -1.6 -1.4 -1.2 -1.0 -0.8
0
200
400
600
800
1000
Potential (V vs Ag/AgCl)
Time (min)
100
c
100
10
1
d
10
1
CO
CH4
HCOOH
CO
CH4
C2H4
0.1
0.01
0.1
C2H6
CH3OH
C2H5OH
Cu-N/C
-2.0
Cu-C3N4
0.01
-2.0
-1.8
-1.6
-1.4
-1.2
-1.8
-1.6
-1.4
-1.2
Potential (V vs Ag/AgCl)
Potential (V vs Ag/AgCl)
Figure 6 (a) Polarization curves of CO2 reduction on two electroꢀ
catalysts, recorded in CO2ꢀsatuated 0.1 M KHCO3. (b) Stability
test of CO2 reduction on CuꢀC3N4 electrocatalyst. (c,d) The measꢀ
ured Faradic efficiencies of various products on CuꢀC3N4 (c) and
CuꢀNC (d) electrodes under different overpotentials.
CONCLUSION
The CRR activity/selectivity differences in the synthesized
CuꢀC3N4 and CuꢀNC electrocatalysts were analyzed to validate
the theoretical considerations. As shown in Figure 6a, the Cuꢀ
C3N4 sample features a more positive onsite potential and
higher current density under a certain overpotential than Cuꢀ
NC. Additionally, the CuꢀC3N4 electrocatalyst was stable durꢀ
ing at least 18 hours operation (Figure 6b). The gaseous and
liquid products analysis showed that the major product of both
CuꢀC3N4 and CuꢀNC electrocatalysts is still hydrogen gas (>50
%) via the competitive cathodic hydrogen evolution reaction.
This can be explained by the fact that there is a large amount
of unꢀcoordinated gꢀC3N4 in the hybrid, which is relatively
active to the hydrogen evolution but inert for CRR, according
to the theoretical analysis (Figure 3a). Therefore, the real
number of active sites on the CuꢀC3N4 material studied is
much lower than that on polycrystalline Cu electrode, which
means that the newly developed CuꢀC3N4 needs further imꢀ
provement (e.g., advanced synthetic methods) to enhance its
potential performance as theoretically predicted.
In summary, we have showed, from both theoretical and exꢀ
perimental perspectives, that for multistep electrochemical
reactions like CRR, a particular molecular scaffold facilitating
coordination of the metallic centers can play a significant role
in affecting the activity and selectivity toward desired reacꢀ
tions. We conducted extensive DFT computations for a Cuꢀ
C3N4 model catalyst to evaluate its potential for CO2 reducꢀ
tion, and compared its performance with that of the Cu (111)
surface and conventional CuꢀNC complex. Based on the comꢀ
putation and experimental results, CuꢀC3N4 shows significantꢀ
ly better CO2 reduction activity with lower onꢀset potential, as
well as a significantly higher C2 production rate than that obꢀ
tained for CuꢀNC. Our study reveals that the molecular scafꢀ
fold can serve as an additional active center for CRR, leading
to the synergistic effect for pathways that leads to deep reꢀ
duced products. This new dual active center model, which has
been validated by our experimental measurements, has never
been observed in the reported CRR electrocatalyst systems.
Additionally, the gꢀC3N4 framework could efficiently uplift
copper’s dꢀband position toward Fermi level, which is critical
in strengthening the adsorption strength of reaction intermediꢀ
ates and CO2 activation. Our investigation suggests that selectꢀ
ing appropriate molecular scaffold can be critical to increase
the electrocatalytic CO2 reduction activity. This atomic level
finding could be helpful to find more efficient and effective
CRR electrocatalysts, and could also be extended to other mulꢀ
As a probing model, indeed, CuꢀC3N4 showed a wider vaꢀ
riety of deep reduction products and higher Faradic efficiency
for each product under the same potential than the CuꢀNC
control sample. For example, CH3OH and CH4 after 6 and 8
electron transfers were only generated on the CuꢀC3N4 elecꢀ
trode (Figure 6c, d), due to the lower overall free energy difꢀ
ference as indicated by the free energy diagram (Figure 1,2).
ACS Paragon Plus Environment

Page 7 of 8
Journal of the American Chemical Society
(6) (a) Li, C. W.; Ciston, J.; Kanan, M. W., Nature 2014, 508, 504;
tiꢀelectron electrocatalytic reactions like nitrogen reduction
(b) Roberts, F. S.; Kuhl, K. P.; Nilsson, A., Angew. Chem. Int. Ed.
2015, 54, 5179; (c) Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo,
T. F., Energy Environ. Sci. 2012, 5, 7050; (d) Ren, D.; Deng, Y.;
Handoko, A. D.; Chen, C. S.; Malkhandi, S.; Yeo, B. S., ACS. Catal.
2015, 5, 2814.
(7) (a) Zheng, Y.; Jiao, Y.; Qiao, S. Z., Adv. Mater. 2015, 27, 5372;
(b) Jiao, Y.; Zheng, Y.; Jaroniec, M. T.; Qiao, S. Z., Chem. Soc. Rev.
2015, 44, 2060.
1
2
3
4
5
6
7
8
reaction for electrochemical ammonia synthesis, oxygen reꢀ
duction/evolution reactions, and alcohol oxidation reactions.
ASSOCIATED CONTENT
Supporting Information. Detailed computational models, conꢀ
figurations of reaction intermediates, more reaction pathways;
more information on materials synthesis, characterization methꢀ
ods, and electrochemical characterization. This material is availaꢀ
ble free of charge via the Internet at http://pubs.acs.org.
(8) (a) Li, C. W.; Kanan, M. W., J. Am. Chem. Soc. 2012, 134, 7231;
(b) Li, M.; Wang, J. J.; Li, P.; Chang, K.; Li, C. L.; Wang, T.; Jiang,
B.; Zhang, H. B.; Liu, H. M.; Yamauchi, Y.; Umezawa, N.; Ye, J. H.,
J. Mater. Chem. A 2016, 4, 4776; (c) Huang, Y.; Handoko, A. D.;
Hirunsit, P.; Yeo, B. S., ACS. Catal. 2017, 7, 1749; (d) Kim, D.; Xie,
C. L.; Becknell, N.; Yu, Y.; Karamad, M.; Chan, K.; Crumlin, E. J.;
Norskov, J. K.; Yang, P. D., J. Am. Chem. Soc. 2017, 139, 8329; (e)
Li, Q.; Fu, J.; Zhu, W.; Chen, Z.; Shen, B.; Wu, L.; Xi, Z.; Wang, T.;
Lu, G.; Zhu, J.ꢀj.; Sun, S., J. Am. Chem. Soc. 2017, 139, 4290; (f) Bai,
S.; Shao, Q.; Wang, P.; Dai, Q.; Wang, X.; Huang, X., J. Am. Chem.
Soc. 2017, 139, 6827; (g) Hirunsit, P.; Soodsawang, W.; Limtrakul, J.,
J. Phys. Chem. C 2015, 119, 8238; (h) Ma, M.; Djanashvili, K.;
Smith, W. A., Angew. Chem. Int. Ed. 2016, 55, 6680.
(9) (a) Nie, X. W.; Luo, W. J.; Janik, M. J.; Asthagiri, A., J. Catal.
2014, 312, 108; (b) Li, Y. W.; Chan, S. H.; Sun, Q., Nanoscale 2015,
7, 8663; (c) Xiao, H.; Cheng, T.; Goddard, W. A., J. Am. Chem. Soc.
2017, 139, 130; (d) Cheng, T.; Xiao, H.; Goddard, W. A., J. Am.
Chem. Soc. 2016, 138, 13802; (e) Peterson, A. A.; AbildꢀPedersen, F.;
Studt, F.; Rossmeisl, J.; Norskov, J. K., Energy Environ. Sci. 2010, 3,
1311; (f) Liu, X. Y.; Xiao, J. P.; Peng, H. J.; Hong, X.; Chan, K.;
Norskov, J. K., Nat. Commun. 2017, 8, 15438.
(10) (a) Rakowski Dubois, M.; Dubois, D. L., Acc. Chem. Res. 2009,
42, 1974; (b) Cheng, M. J.; Kwon, Y.; HeadꢀGordon, M.; Bell, A. T.,
J. Phys. Chem. C 2015, 119, 21345; (c) Back, S.; Lim, J.; Kim, N. Y.;
Kim, Y. H.; Jung, Y., Chem. Sci. 2017, 8, 1090; (d) Shen, H. M.; Li,
Y. W.; Sun, Q., J. Phys. Chem. C 2017, 121, 3963; (e) Li, Y. W.; Su,
H. B.; Chan, S. H.; Sun, Q., ACS. Catal. 2015, 5, 6658; (f) Oh, Y.;
Hu, X. L., Chem. Soc. Rev. 2013, 42, 2253; (g) Back, S.; Lim, J.;
Kim, N.ꢀY.; Kim, Y. H.; Jung, Y., Chem. Sci. 2017, 8, 1090.
(11) Zheng, Y.; Jiao, Y.; Zhu, Y.; Cai, Q.; Vasileff, A.; Li, L. H.;
Han, Y.; Chen, Y.; Qiao, S. Z., J. Am. Chem. Soc. 2017, 139, 3336.
(12) (a) Jiao, Y.; Zheng, Y.; Smith, S. C.; Du, A.; Zhu, Z.,
ChemSusChem 2014, 7, 435; (b) Tan, X.; Tahini, H. A.; Smith, S. C.,
Curr. Opin. Electrochem. 2017, 10.1016/j.coelec.2017.08.006.
(13) Zheng, Y.; Jiao, Y.; Chen, J.; Liu, J.; Liang, J.; Du, A.; Zhang,
W.; Zhu, Z.; Smith, S. C.; Jaroniec, M.; Lu, G. Q.; Qiao, S. Z., J. Am.
Chem. Soc. 2011, 133, 20116.
(14) (a) Gowthaman, N. S. K.; Raj, M. A.; John, S. A., ACS
Sustainable Chem. Eng. 2017, 5, 1648; (b) Liu, L.ꢀL.; Chen, C.ꢀP.;
Zhao, L.ꢀS.; Wang, Y.; Wang, X.ꢀC., Carbon 2017, 115, 773.
(15) (a) Gao, P.; Li, S.; Bu, X.; Dang, S.; Liu, Z.; Wang, H.; Zhong,
L.; Qiu, M.; Yang, C.; Cai, J.; Wei, W.; Sun, Y., Nat. Chem. 2017, 9,
1019; (b) Jiao, F.; Li, J.; Pan, X.; Xiao, J.; Li, H.; Ma, H.; Wei, M.;
Pan, Y.; Zhou, Z.; Li, M.; Miao, S.; Li, J.; Zhu, Y.; Xiao, D.; He, T.;
Yang, J.; Qi, F.; Fu, Q.; Bao, X., Science 2016, 351, 1065; (c) Lin, L.;
Zhou, W.; Gao, R.; Yao, S.; Zhang, X.; Xu, W.; Zheng, S.; Jiang, Z.;
Yu, Q.; Li, Y.ꢀW.; Shi, C.; Wen, X.ꢀD.; Ma, D., Nature 2017, 544, 80.
(16) (a) Kresse, G.; Hafner, J., Phys. Rev. B 1993, 47, 558; (b) Kresse,
G.; Furthmüller, J., Phys. Rev. B 1996, 54, 11169; (c) Kresse, G.;
Joubert, D., Phys. Rev. B 1999, 59, 1758; (d) Perdew, J. P.; Burke, K.;
Ernzerhof, M., Phys. Rev. Lett. 1996, 77, 3865; (e) Perdew, J. P.;
Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1997, 78, 1396.
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
AUTHOR INFORMATION
Corresponding Author
*s.qiao@adelaide.edu.au
Author Contributions
‡These authors contributed equally.
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
We acknowledge financial support by the Australian Research
Council
(DP170104464,
DP160104866,
DP140104062,
DE160101163 and FL170100154). DFT computations within this
research was undertaken with the assistance of resources and serꢀ
vices from the National Computational Infrastructure (NCI),
which is supported by the Australian Government. NEXAFS and
XPS were performed on soft Xꢀray beamlines at Australian Synꢀ
chrotron.
REFERENCES
(1) (a) Gattrell, M.; Gupta, N.; Co, A., J. Electroanal. Chem. 2006,
594, 1; (b) Zhu, D. D.; Liu, J. L.; Qiao, S. Z., Adv. Mater. 2016, 28,
3423; (c) Chen, Y.; Jia, G.; Hu, Y.; Fan, G.; Tsang, Y. H.; Li, Z.; Zou,
Z., Sustainable Energy Fuels 2017, 1, 1875.
(2) (a) Lewis, N. S., Science 2016, 351, aad1920; (b) Jones, J.ꢀP.;
Prakash, G. K. S.; Olah, G. A., Isr. J. Chem. 2014, 54, 1451; (c) Zhan,
Z.; Kobsiriphat, W.; Wilson, J. R.; Pillai, M.; Kim, I.; Barnett, S. A.,
Energy Fuels 2009, 23, 3089; (d) Peng, C.; Reid, G.; Wang, H.; Hu,
P., J. Chem. Phys. 2017, 147, 030901; (e) Centi, G.; Perathoner, S.,
ChemSusChem 2010, 3, 195.
(3) (a) Costentin, C.; Robert, M.; Saveant, J.ꢀM., Chem. Soc. Rev.
2013, 42, 2423; (b) Vasileff, A.; Zheng, Y.; Qiao, S. Z., Adv. Energy.
Mater. 2017, 1700759; (c) Hori, Y., Electrochemical CO2 Reduction
on Metal Electrodes. In Modern Aspects of Electrochemistry,
Vayenas, C. G.; White, R. E.; GamboaꢀAldeco, M. E., Eds. Springer
New York: New York, NY, 2008; pp 89; (d) Malik, K.; Singh, S.;
Basu, S.; Verma, A., WIREs Energy. Environ. 2017, 6, e244; (e)
Simakov, D. S. A., Renewable Synthetic Fuels and Chemicals from
Carbon Dioxide: Fundamentals, Catalysis, Design Considerations
and Technological Challenges. Springer International Publishing:
Switzerland, 2017.
(4) (a) Zhang, L.; Zhao, Z.ꢀJ.; Gong, J., Angew. Chem. Int. Ed. 2017,
56, 11326; (b) Kortlever, R.; Shen, J.; Schouten, K. J. P.; Calleꢀ
Vallejo, F.; Koper, M. T. M., J. Phys. Chem. Lett. 2015, 6, 4073; (c)
Lu, Q.; Jiao, F., Nano Energy 2016, 29, 439.
(5) (a) Whipple, D. T.; Kenis, P. J. A., J. Chem. Phys. Lett. 2010, 1,
3451; (b) Weng, Z.; Jiang, J.; Wu, Y.; Wu, Z.; Guo, X.; Materna, K.
L.; Liu, W.; Batista, V. S.; Brudvig, G. W.; Wang, H., J. Am. Chem.
Soc. 2016, 138, 8076; (c) Gattrell, M.; Gupta, N.; Co, A., Energy
Convers. Manage. 2007, 48, 1255; (d) Kas, R.; Kortlever, R.; Milbrat,
A.; Koper, M. T. M.; Mul, G.; Baltrusaitis, J., Phys. Chem. Chem.
Phys. 2014, 16, 12194.
(17) Tkatchenko, A.; Scheffler, M., Phys. Rev. Lett. 2009, 102,
073005.
(18)Durand, W. J.; Peterson, A. A.; Studt, F.; AbildꢀPedersen, F.;
Norskov, J. K., Surf. Sci. 2011, 605, 1354.
(19) Zhang, Y. J.; Sethuraman, V.; Michalsky, R.; Peterson, A. A.,
ACS. Catal. 2014, 4, 3742.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 8
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
8
ACS Paragon Plus Environment