Boosting Formate Production in Electrocatalytic CO2 Reduction over Wide Potential Window on Pd Surfaces

Article abstract of DOI:10.1021/jacs.7b12506

Facile interconversion between CO₂ and formate/formic acid (FA) is of broad interest in energy storage and conversion and neutral carbon emission. Historically, electrochemical CO₂ reduction reaction to formate on Pd surfaces was limited to a narrow potential range positive of -0.25 V (vs RHE). Herein, a boron-doped Pd catalyst (Pd-B/C), with a high CO tolerance to facilitate dehydrogenation of FA/formate to CO₂, is initially explored for electrochemical CO₂ reduction over the potential range of -0.2 V to -1.0 V (vs RHE), with reference to Pd/C. The experimental results demonstrate that the faradaic efficiency for formate (IHCOO^- reaches ca. 70% over 2 h of electrolysis in CO₂-saturated 0.1 M KHCO₃ at -0.5 V (vs RHE) on Pd-B/C, that is ca. 12 times as high as that on homemade or commercial Pd/C, leading to a formate concentration of ca. 234 mM mg^-1 Pd, or ca. 18 times as high as that on Pd/C, without optimization of the catalyst layer and the electrolyte. Furthermore, the competitive selectivity IHCOO^-/I?_CO on Pd-B/C is always significantly higher than that on Pd/C despite a decreases of IHCOO^- and an increases of the CO faradaic efficiency (??_CO) at potentials negative of -0.5 V. The density functional theory (DFT) calculations on energetic aspects of CO₂ reduction reaction on modeled Pd(111) surfaces with and without H-adsorbate reveal that the B-doping in the Pd subsurface favors the formation of the adsorbed HCOO?, an intermediate for the FA pathway, more than that of?COOH, an intermediate for the CO pathway. The present study confers Pd-B/C a unique dual functional catalyst for the HCOOH → CO₂ interconversion.

Full text of DOI:10.1021/jacs.7b12506

Subscriber access provided by MT ROYAL COLLEGE
Article
2
Boosting Formate Production in Electrocatalytic CO
Reduction over Wide Potential Window on Pd Surfaces
Bei Jiang, Xia-Guang Zhang, Kun Jiang, De-Yin Wu, and Wen-Bin Cai
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b12506 • Publication Date (Web): 07 Feb 2018
Downloaded from http://pubs.acs.org on February 7, 2018
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.
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 24
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Boosting Formate Production in Electrocatalytic
CO2 Reduction over Wide Potential Window on
Pd Surfaces
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
Bei Jiang†, XiaꢀGuang Zhang‡, Kun Jiang§, DeꢀYin Wu‡,* and WenꢀBin Cai†,*
† Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Collaborative
Innovation Center of Chemistry for Energy Materials, Department of Chemistry, Fudan
University, Shanghai 200433, China
‡ State Key Laboratory of Physical Chemistry of Solid Surfaces, Collaborative Innovation
Center of Chemistry for Energy Materials, Department of Chemistry, and College of Chemistry
and Chemical Engineering, Xiamen University, Xiamen, 361005, China
§ Rowland Institute, Harvard University, Cambridge, MA 02142, United States.
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 24
1
2
3
4
ABSTRACT
5
6
7
8
Facile interconversion between CO2 and formate/formic acid (FA) is of broad interest in the
energy storage and conversion and the neutral carbon emission. Historically, electrochemical
CO2 reduction reaction to formate on Pd surfaces was limited to a narrow potential range
positive of ꢀ0.25 V (vs. RHE). Herein, a boronꢀdoped Pd catalyst (PdꢀB/C), with a high CO
tolerance to facilitate dehydrogenation of FA/formate to CO2, is initially explored for
electrochemical CO2 reduction over the potential range of ꢀ0.2 V to ꢀ1.0 V (vs. RHE), with
reference to Pd/C. The experimental results demonstrate that the faradaic efficiency for formate
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
(ηHCOOꢀ) reaches ca. 70% over 2 h of electrolysis in CO2 saturated 0.1 M KHCO3 at ꢀ0.5 V (vs.
RHE) on PdꢀB/C, that is ca. 12 times as high as that on homemade or commercial Pd/C, leading
to a formate concentration of ca. 234 mM mgꢀ1Pd, or ca. 18 times as high as that on Pd/C,
without optimization of the catalyst layer and the electrolyte. Furthermore, the competitive
selectivity ηHCOOꢀ/ηCO on PdꢀB/C is always significantly higher than that on Pd/C despite a
decreases of ηHCOOꢀ and an increases of the CO faradaic efficiency (ηCO) at potentials negative
of ꢀ0.5 V. The density functional theory (DFT) calculations on energetic aspects of CO2
reduction reaction on modelled Pd(111) surfaces with and without Hꢀadsorbate reveal that the
Bꢀdoping in the Pd subsurface favors the formation of the adsorbed HCOO*, an intermediate
for the FA pathway, more than that of *COOH, an intermediate for the CO pathway. The
present study confers PdꢀB/C a unique dual functional catalyst for the HCOOH ↔ CO2
interconversion.
KEYWORDS: Carbon dioxide reduction, Pd catalysts, Boronꢀdoping, Electrocatalysis,
Competitive selectivity, Density functional theory calculations.
2
ACS Paragon Plus Environment

Page 3 of 24
Journal of the American Chemical Society
1
2
3
4
1. INTRODUCTION
Controlled facile interconversion between CO2 and formic acid (FA)/formate is of great
significance in energy conversion and storage as well as carbon emission control.1ꢀ6 The
relevant applications include but are not limited to direct FA/formate fuel cell via
electrochemical dehydrogenation of FA/formate to CO2, and fuel regeneration based on
electrochemical hydrogenation of CO2 to formate/FA with a projected neutral CO2 emission.7ꢀ10
Despite the fact that some new catalysts are recently developed with high performances in
converting CO2 to formate, including reduced nanoꢀSnO2,2 Cu@SnO2,11 oxideꢀderied Pb12 and
partially oxidized atomic cobalt layers,13 Pd is still the most promising catalytic metal that
enables the facile interconversion between CO2 and FA/formate. In fact, by using a Pdꢀbased
anode catalyst, a high power density comparable to that of a H2ꢀO2 fuel cell could be achieved
in direct formic acid fuel cells,5,14ꢀ17 and by using a Pd cathode catalyst, smallest overpotentials
are needed to enable the CO2RR to formate due to the unique property of H
adsorption/absorption on/in Pd.18,19 Nevertheless, a major challenge commonly facing such an
interconversion on Pd surfaces is that the parallel COꢀpathway competes strongly with or even
deactivates the desired pathway.18,20ꢀ25
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
So far, the study of CO2RR on Pdꢀbased catalysts was chiefly targeted to achieve high
faradaic efficiencies (FE,
catalyst particle size, shape and composition.5,18,21,26,27 Notably, the formate faradaic efficiency
HCOOꢀ) and the CO faradaic efficiency ( CO) are highly dependent on applied potential, pH
and concentration of electrolyte, and electrolysis time. Despite a large discrepancy in the
reported HCOOꢀ and iHCOOꢀ values owing to different measurement conditions for CO2RR on Pd
η) and partial currents (i) on either formate or CO by tuning the
(η
η
η
electrodes, it seems to reach a consensus that a higher selectivity toward HCOOꢀ against CO is
limited to a very narrow range of potentials negative of ca. 0 V (vs. RHE) in KHCO3 solution.
Specifically, Wang’s group mentioned that
close to 100% in the most commonly used CO2ꢀsaturated 0.1 M KHCO3.21,26,28,29 Kanan’s
group reported HCOOꢀ values were higher than 90% for CO2RR on Pd/C dispersed on Ti in CO2
ηCO on Pd/C between ꢀ0.6 and ꢀ1.2 V (vs. RHE) was
η
–saturated 0.5 M and 2.8 M KHCO3 at potentials positive of ꢀ0.25 V (vs. RHE) but decayed
due to surface CO accumulation.18 A recent detailed study by Wang’s group further revealed
that either ηHCOOꢀ or
η
CO higher than 90% could be attained in 1 M KHCO3 at potentials
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 24
1
2
3
4
positive of ꢀ0.2 V or negative of ꢀ0.5 V (vs. RHE).19 Notably, the high ηHCOOꢀ at a small
5
6
7
overpotential in principle can be counterꢀbalanced by a low total current, and the resulting
i
HCOOꢀ is not necessarily high.3,18,30,31 On the other hand, using Pd/C as a benchmark to screen a
8
9
series of PdxPt1ꢀx/C for CO2RR in CO2ꢀsaturated 0.1 M K2HPO4/KH2PO4 (PBS), Koper’s group
detected a ηHCOOꢀ of ca. 10 ~ 17% for Pd/C at ꢀ0.5 V (vs. RHE) where partial formate current is
highest, depending on the electrolysis time.5
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
Despite these contributions, unfortunately, least effort has been devoted to promote the
formate production in the electrochemical CO2RR on Pd catalysts over a wide potential range,
especially when they are targeted for the interconversion of FA/formate to and from CO2.
PdꢀB/C, a boronꢀdoped Pdꢀbased electrocatalyst, was first developed in our group.32 B insertion
in the Pd lattice leads to the formation of interstitial PdꢀB alloy. Our previous works indicate
that as compared to Pd/C, PdꢀB/C enhances the (electro)catalytic dehydrogenation of FA with
significantly inhibited CO accumulation, promising for the anode catalyst of direct formic acid
fuel cells as well as the heterogeneous catalyst of hydrogen production in FA/formate
solution.22,23 With this in mind, it is enticing to explore PdꢀB/C as a new catalyst for the reverse
electroꢀhydrogenation process, with an expectation to tune up the ratio ηHCOOꢀ/ηCO and boost
the formate production at extended potentials, in an effort to ultimately close the carbon cycle
between CO2 and formate/FA.
In the present work, the CO2RR on PdꢀB/C and Pd/C catalysts in 0.1 M KHCO3 over the
potential range of ꢀ0.2 to ꢀ1.0 V (vs. RHE) is comparatively investigated by electrochemical
measurement, onꢀline gas chromatography and NMR analysis, revealing significantly higher
ηHCOOꢀ/ηCO ratio and formate production rate on PdꢀB/C than on Pd/C under otherwise the
same conditions. To understand the mechanistic aspect at a molecular level, density functional
theory (DFT) calculations are applied to elucidate the competition for the FA and CO pathways
during CO2RR on Hꢀadsorbed and unadsorbed Pd(111) surfaces without and with Bꢀdoping in
the subsurface. The present work signifies that PdꢀB/C is a unique bifunctional catalyst for
promoting the HCOOH ↔ CO2 interconversion.
2. EXPERIMENTAL & CALCULATION METHODS
2.1 Catalyst Synthesis and Characterization. Carbon supported boronꢀdoped palladium
4
ACS Paragon Plus Environment

Page 5 of 24
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
catalyst (PdꢀB/C with 18 wt.% Pd and 6.3 at.% B doping in Pd as determined by ICPꢀAES) was
prepared through an aqueous chemical synthesis with dimethylamine borane (DMAB) as the
reducing agent, based on our previous reports.22,32 The B doping content is confined to ca. 6ꢀ7
at. % by such a synthesis. Further increasing the Bꢀdoping content may be possible via an
organic phase synthesis but may cause a complication in surface cleaning and a possible
structural transition,33 and thus was not attempted in this work. Pd/C (ca. 40 wt. %) from BASF
Fuel Cell Inc. (Lot No. F0301009), noted as Pd/C (com) and homemade Pd/C (ca. 19 wt.%),
noted as Pd/C (home) with sodium borohydride (NaBH4) as the reducing agent in an aqeuous
solution synthesis32 were used for comparison. The morphology and dispersion of PdꢀB/C and
Pd/C catalysts were characterized by A JEOL JEMꢀ2100F tunneling electron microscope
(TEM). An ESCALAB 250 XI Xꢀray photoelectron spectroscopy (XPS) of Thermo Scientific
was used to analyze the surface electronic structures, using a monochromatic Al Kα radiation
(1486.6 eV). The binding energies were calibrated with reference to the C 1s peak at 284.6 eV.
The Xꢀray diffractometer (XRD) of a Bruker D8 advance was used to examine the crystalline
structures. The chemical composition of PdꢀB/C was analyzed with the aid of inductively
coupled plasmaꢀatomic emission spectroscopy (ICPꢀAES) (Thermo Elemental IRIS Intrepid).
2.2. Electrochemical Measurement. The electrochemical measurement on CO2RR was
carried out at room temperature in an Hꢀtype cell with a threeꢀelectrode system of which the
cathode and anode compartments respectively containing 45 mL of the electrolyte were
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
separated by a Nafion 117 membrane. An Lꢀtype glassy carbon (GC,
φ = 5 mm) electrode
covered with a Pd/C or PdꢀB/C catalyst layer served as the working electrode. A catalyst ink
was prepared by ultrasonically dispersing 5 mg of Pd/C (home), 5 mg of PdꢀB/C and 2.5 mg of
Pd/C (com) respectively in 1 mL of isopropanol solvent mixed with 120 ꢀL of 5 wt. % Nafion
isopropanol solution (Alfa Aesar). Around 22 ꢁL of this ink was transferred and dropꢀcoated
onto the GC electrode with a Pd loading of 100 ꢁg cmꢀ2 via a pipette, and dried at room
temperature. Before electrochemical measurement, the catalyst layerꢀcovered GC electrode was
thoroughly rinsed with a copious amount of ultrapure MilliꢀQ water.
A CHI 660B electrochemistry workstation (Shanghai Chenhua Instruments) was used for
potential control and current measurement. A saturated calomel electrode (SCE) and a Pt foil
served as the reference electrode and the counter electrode, respectively. Nevertheless, all
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 24
1
2
3
4
5
6
7
8
9
potentials reported were converted with reference to a reversible hydrogen electrode (RHE).
2.3. Detection and Quantification of Products. The products dissolved in solution were
analyzed with NMR spectroscopy by a 500 MHz nuclear magnetic resonance (NMR) from
Bruker Company.34 Phenol was used as the internal standard for quantification of the dissolved
products. The gas products were analyzed by an onꢀline gas chromatography (GC) with a Gas
Chromatograph 2060 from Ramiin Instruments equipped with a flame ionization detector (FID)
coupled with a methanizer and a thermal conductivity detector (TCD). The detection limits of
the FID and TCD detectors are 0.5 ppm and 5 ppm, respectively. The FID was used to quantify
hydrocarbons, CO and CO2 and the TCD to quantify H2. The data were read out every 15 min
after the gas products were brought in the GC system at desired times of 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
2.4. Computational methods. Density functional theory (DFT) calculations were
performed by using the PerdewꢀBurkeꢀErnzerhof (PBE) functional of generalized gradient
approximation (GGA)35 in Vienna Ab Initio Simulation Package (VASP).36 The
projectorꢀaugmented wave (PAW) method was applied to describe the electronꢀion
interactions.37 Energy cutoff of 500 eV and MethfesselꢀPaxton method with a broadening factor
of 0.1 eV were used. The calculated lattice constant of Pd bulk was 3.96 Å, which agrees the
experimental value of 3.89 Å (error is ca. 1.6 %).38 In this work, a vacuum space of 15 Å was
used for all the calculations. The Γꢀcentered kꢀpoint sampling grid of 4×4×1 was adopted for all
single crystal facets. For calculations of density of states (DOS), the kꢀpoint sampling grid was
increased to 12×12×1. The Pd(111) surface was simulated with a 2×2 fiveꢀlayer slab with the
top three layers relaxed and the bottom two layers fixed in all calculations. On the top of the
optimized structures, vibrational frequencies of the free and adsorbed molecules were
calculated on the basis of the finite difference method where all metal atoms and boron atoms
were frozen.The Bader charge analysis was performed using the code from Henkelman group.39
Adsorption energies (Eads) of reactants, intermediates, and products were estimated by
E =E −Emetal −Emol
(1)
ads
tot
where Etot, Emetal and Emol are total electronic energies of molecules adsorbed on a slab, clean
metal slab and free molecules, respectively. In this work, all thermodynamic properties were
further calculated using the Atomic Simulation Environment suite of programs.40 The Gibbs
6
ACS Paragon Plus Environment

Page 7 of 24
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
free energies were calculated at 298.15 K and 1 atm as outlined below,
G=H −TS = EDFT +EZPE
+
298K CvdT −TS
(2)
∫
0
where EDFT is the total electronic energy obtained from DFT geometry optimization, EZPE is the
9
zeroꢀpoint vibrational energy (ZPE) obtained from calculated vibrational frequencies, the
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
298KC dT
thermal energy
is calculated from the heat capacity, T is the temperature, and S is
v
∫
0
the entropy of studied systems. For the molecules of CO2, H2, CO, HCOOH and H2O, the ideal
gas approximation was used, and for the adsorbates the harmonic approximation was used. The
electrode potential related to the Gibbs free energy of a half electrochemical reaction could be
calculated by the computational hydrogen electrode (CHE) model for CO2RR to HCOOH41.
The CHE model was also used to refer the chemical potential of protonꢀelectron pairs and
hydrogen gas at the standard equilibrium condition,
1
H + ) + e− ↔ H2( g )
(3)
aq
(
2
3. RESULTS AND DISCUSSION
3.1. Catalyst Characterizations. Typical TEM images for the PdꢀB/C, Pd/C (home) and
Pd/C (com) samples are shown in Figure 1AꢀC. The statistics of 300 nanoparticles from
different TEM images of each sample reveals that the sizes of PdꢀB, Pd (home) and Pd (com)
nanoparticles are ca. 4.1 ± 0.5, 4.0 ± 0.3 and 3.8 ± 0.5 nm, respectively.22,23 The ICPꢀAES
analysis indicates an insignificant amount of Bꢀdopant (0~2 at. % B) in Pd nanoparticles for the
Pd/C samples if any, and 6.3 at. % B content in PdꢀB nanoparticles for the PdꢀB/C sample.
According to the ball model proposed by Benfield42, the coverage of B doped in the Pd
subsurface of a 4ꢀnm PdꢀB nanoparticle can be calculated to be higher than 75%. For
simplification, the full coverage of B in Pd(111) subsurface (denoted as Pdꢀ4B, representing 4
B atoms in a 2×2 fiveꢀlayer slab model) will be used in the following DFT calculations to
explain the experimental results.
The XRD patterns in Figure 1D show the diffraction peaks featuring Pd(111), Pd(200) and
Pd(220) planes for PdꢀB/C shift negatively, as compared to those for Pd/C (home) and Pd/C
(com), suggestive of an enlarged PdꢀPd lattice by B doping in the interstices. The coreꢀlevel
XPS spectra in Figure 1E after deconvolution show that the Pd(0) core level binding energy is
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 24
1
2
3
4
5
6
7
8
9
slightly blueꢀshifted for PdꢀB/C with respect to that for Pd/C (home) or Pd/C (com), indicative
of a partial electron transfer between Pd and B as well as a downshifted dꢀband center of
surface Pd atoms. The partial charge transfer is also consistent with the Bader charge analysis
in the following DFT calculation (see Table S2 in the SI). A higher surface Pd(ΙΙ) component
for Pd/C (com) may arise from an aging exposure to the air, and can be readily reduced to Pd(0)
at the CO2RR potentials. The TEM, ICPꢀAES, XRD and XPS results are largely consistent with
those in our previous works for electrocatalytic oxidation of FA,15,22,43 ensuring the reliable
comparison of the CO2RR properties on PdꢀB/C and Pd/C in the following.
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. (AꢀC) TEM images and the size distribution histograms (inserts) for PdꢀB/C, Pd/C
(home) and Pd/C (com), respectively. (D) XRD patterns for PdꢀB/C, Pd/C (home) and Pd/C
8
ACS Paragon Plus Environment

Page 9 of 24
Journal of the American Chemical Society
1
2
3
4
5
(com), standard XRD peaks are indicated with vertical bars, according to JCPDS file 65ꢀ6174.
(E) XPS spectra for Pd 3d regions.
6
7
8
9
The XRD, TEM and XPS characterizations of the catalysts subjected to the CO2RR
measurement in CO2ꢀsaturated 0.1 M KHCO3 at –0.5 V for 2 h were also carried out, and the
results are shown in Figure S1 in the Supporting Information (SI). Bꢀdoping in the Pd lattice
interstices largely remained for PdꢀB/C even after the electrolysis, as indicated from the
negative shift of the feature peak in the XRD pattern and the positive shift of Pd core level in
the XPS spectrum, with respect to the reference values for Pd/C. Furthermore, PdꢀB/C
exhibited a higher resistance to the nanoparticle size growth at a hydrogen evolution potential,
as compared to Pd/C. See the SI for more details.
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
3.2 CO2RR properties on Pd-B/C and Pd/C
Figure 2. Negative going linear sweep voltammograms of (A) unsupported carbon black, (B)
PdꢀB/C, (C) Pd/C (home) and (D) Pd/C (com) recorded at a scan rate of 50 mV sꢀ1 in
N2ꢀsaturated (black) or CO2ꢀsaturated (red) 0.1 M KHCO3 electrolyte.
Figure 2 shows the linear sweep voltammograms (LSVs) recorded at 50 mV sꢀ1 for Pd/C
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 24
1
2
3
4
5
6
7
8
9
(home made and commercial) and PdꢀB/C catalyst layers respectively casted on GC electrode
in either N2ꢀ or CO2ꢀsaturated 0.1 M KHCO3. The baseline LSVs for Pdꢀfree carbon black layer
casted on GC eldctrode in the same electrolytes are also presented, showing a very small
cathodic current mainly attributable to the hydrogen evolution reaction. The significant current
responses on Pd/C and PdꢀB/C in CO2 saturated 0.1 M KHCO3 with pH6.8 may be divided into
three potential regions. At potentials positive of ca. 0.00 V, the apparent cathodic currents are
very small and nearly overlapped in N2ꢀ and CO2ꢀsaturated 0.1 M KHCO3, suggestive of a tiny
CO2RR current if any, consistent with an expected standard equilibrium potential of ca. ꢀ0.02 V
for the CO2/formate redox in CO2ꢀsaturated 0.1 M KHCO3.8 With the potential negatively
shifted, the apparent cathodic current increases with a hump peak ranging from ꢀ0.05 V to ꢀ0.65
V in CO2ꢀsaturated 0.1 M KHCO3, indicating that appreciable CO2RR does occur on each
catalyst at sufficiently negative potentials. At potentials negative of ca. ꢀ0.70 V, the apparent
reduction current in CO2ꢀsaturated 0.1 M KHCO3 turns even lower than that in N2ꢀsaturated 0.1
M KHCO3 on both Pd/C and PdꢀB/C. One possible explanation is that the H2
evolutionꢀdominant process in the N2ꢀsaturated 0.1 M KHCO3 is inhibited by the competitive
CO2RR process in the CO2ꢀsaturated 0.1 M KHCO3 with the inhibition effect being more
pronounced with further decreasing potential.25,44 Potential dependent CO and formate FEs and
relative selectivity will be further discussed next following the analytic determination of CO
and formate in the gas and liquid products by using GC and NMR methods respectively. The
GC chromatograms, potential dependent H2 FEs plots and the NMR spectrum of a liquid
product sample are shown in Figures S2, S3 and S4 in the SI, respectively.
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
3.3 Analysis of Gas and Liquid CO2RR Products. Figure 3 shows the FEs for the
gaseous product CO (Figure 3A) and liquid product HCOOꢀ (Figure 3B) during the
electrocatalytic CO2RR process. The FEs were calculated according to the following equation.
N×c×V ×F
FE =
(4)
Q
where
gas (mL),
period of 2 h,
N
is the number of transferred electrons (here
is the Faradaic constant (96485 C molꢀ1),
is the product concentration as measured by GC or NMR (mol mLꢀ1).
N
= 2),
V
is the total volume of liquid or
F
Q is the overall charge (C) over the
c
To avoid relatively large errors in measuring the gas products by GC at lower
10
ACS Paragon Plus Environment

Page 11 of 24
Journal of the American Chemical Society
1
2
3
4
overpotentials where the total current is quite small in CO2ꢀsaturated 0.1 M KHCO3, only the
5
6
η
CO values at potentials from ꢀ0.4 to ꢀ1.0 V are presented and compared in Figure 3A. It is clear
that the CO yield increases with negatively going potential on all the Pdꢀbased catalysts. The
larger ηCO on Pd/C at higher overpotentials is in agreement with previous reports.21,27,29
Nevertheless, ηCO on PdꢀB/C is much lower than that on both Pd/C (home) and Pd/C (com) at
all given potentials of interest. On the other hand, since the concentration of the liquid product
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
formate can be determined more accurately by NMR at potentials up to ꢀ0.05 V, and
ηHCOOꢀ and
the formate concentration per mg of Pd ( HCOOꢀ) over 2ꢀh electrolysis are shown as a function
C
of potential from ꢀ0.2 to ꢀ1.0 V in Figures 3B and 3C, respectively. The variation of ηHCOOꢀ
with potential on the two Pd/C catalysts demonstrates a profile similar to that reported by
Koper’s group.5 More importantly, at any given potential, PdꢀB/C always outperforms Pd/C in
term of ηHCOOꢀ. The highest ηHCOOꢀ can reach ca. 70% at ꢀ0.5 V on PdꢀB/C over a period of 2
h’s electrolysis, ca. 12 times of that on Pd/C, with the resulting formate concentration of ca. 234
mM mgꢀ1Pd on PdꢀB/C, ca.18 times as high as that on Pd/C. Besides, to our best knowledge the
ηHCOOꢀ value exceeds all reported ones for the CO2RR on Pdꢀbased catalysts in CO2ꢀsaturated
0.1 M KHCO3,26,28,29 and the Pd mass activity in terms of formate concentration is even a bit
higher than that obtained on optimally screened Pd7Pt3/C at ꢀ0.5 V in CO2ꢀsaturated 0.1 M PBS
solution with pH6.7 and smaller volume.5 Notably, in the traditional COꢀformation potential
region (<ꢀ0.5 V), the promoting effect of the Bꢀdoping on formate production on Pd is
particularly appreciated. In fact, at potentials negative of ꢀ0.7 V, virtually no formate could be
detected on the two Pd/C catalysts with
to 15ꢀ30% on the PdꢀB/C catalyst.
ηHCOOꢀ being close to zero. In contrast, ηHCOOꢀ rises up
Since our focus here is to compare the formate production during the CO2RR on Pd/C and
PdꢀB/C under otherwise same conditions, optimization of the electrolyte and the catalyst layer
has not been attempted on purpose. In fact, KHCO3 plays an important role in the CO2RR on
metal surfaces45ꢀ47 and the mass activity for formate production may be further increased by
simply increasing the concentration of KHCO3 (so its pH) at a CO2 saturation concentration of
33 mM. As demonstrated by Kanan et al.18, the average formate partial current can be increased
approximately by a factor of 2ꢀ3 per pH unit in CO2ꢀsaturated electrolyte. In practice, the use of
a gas diffusion electrode and a flow cell device shall greatly increase the output currents for the
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 24
1
2
3
4
CO2RR with the same catalyst.8,19,48
5
6
7
8
Figure 3D shows the logarithm plots of formate partial current versus potential on three
catalysts over the range from ꢀ0.05 to ꢀ0.20 V, yielding an apparent Tafel slope of 137 mV decꢀ1
on PdꢀB/C, smaller than that averaged on the two Pd/C catalysts (i.e., 186 mV decꢀ1). The
smaller Tafel slope on PdꢀB/C is in agreement with the higher electrocatalytic activity, and the
larger Tafel slope on Pd/C may result from its higher tendency to H adsorption and CO
poisoning during CO2RR.18,22,23 Furthermore, these values are more or less close to 150 mV
decꢀ1 reported previously for the CO2RR to formate on Pd electrodes,18,31 largely consistent
with an electrohydrogenation mechanism, in which the rateꢀlimiting step is either a
protonꢀcoupled oneꢀelectron transfer or a hydrogen addition reaction involving electroadsorbed
hydrogen, with the transfer coefficient deviating negatively from 0.5.
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. Potential dependent FEs of CO (A) and formate (B), asꢀformed formate concentration
(C) normalized to Pd masses of the PdꢀB/C (red, square), Pd/C (home) (black, hollow circle)
and Pd/C (com) (blue, solid circle) catalyst layers in CO2ꢀsaturared 0.1 M KHCO3 over 2ꢀh
electrolysis. (D) log(jHCOOꢀ) vs. potential plots for CO2RR on Pdꢀbased catalysts.
Given that
η
HCOOꢀ and ηCO vary with potential, the ηHCOOꢀ and ηCO values obtained on
12
ACS Paragon Plus Environment

Page 13 of 24
Journal of the American Chemical Society
1
2
3
4
PdꢀB/C are first ratioed to their counterparts on either Pd/C (home) or Pd/C (com) at different
5
6
7
8
potentials, respectively. Figures 4A and 4C show the ηHCOOꢀ(PdꢀB)
Figure 3B is far greater than 1 and increases with the potential moving negatively whilst the
ηCO(PdꢀB) CO(Pd) value adapted from Figure 3A keeps less than 1 in the whole potential range,
/ηHCOOꢀ(Pd) value adapted from
/η
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
suggesting that the Bꢀdoping tends to promote the production of formate and inhibit the CO
formation on Pdꢀbased catalysts during the CO2RR at given potentials. The extent to which the
Bꢀdoping affects the resulting pathways on Pd/C can be better seen by introducing the concept
of relative selectivity (RS) defined as follows,
ꢀꢁ = ꢂηꢃꢄꢅꢅꢆꢇ
/ ꢂηꢃꢄꢅꢅꢆꢇ
(5)
η
η
ꢄꢅ
ꢄꢅ
ꢈꢉꢊ
ꢈꢉ
The plot of RS values as a function of potential is presented in Figure 4B for Pd/C (home)
and Figure 4D for Pd/C (com). Despite some variation of the RS values for Pd/C (com) and
Pd/C (home), the RS values are in general one to two orders of magnitude higher with
Bꢀdoping. Furthermore, with decreasing potential from ꢀ0.4 to ꢀ0.7 V, the RS values first
increase and then decrease with the maxima being around ꢀ0.5 V, or the potential where the
PdꢀB/C catalyst yields the highest formate FE and concentration. At potentials negative of ꢀ0.7
V, the RS values tend to increase again.
All the above data strongly suggest that both the absolute formate yield (FE and mass
activity) and relative selectivity toward formate against CO (RS) in the CO2RR can be
effectively enhanced by doping B into the interstices of Pd lattice over a wide potential
window.
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 24
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. (A) The ratios of FEs of formate and CO on PdꢀB/C versus Pd/C (home) catalysts, (B)
RS of PdꢀB vs. Pd (home) catalysts, (C) the ratios of FEs of formate and CO on PdꢀB/C vs.
Pd/C (com) catalysts, (D) RS of PdꢀB vs. Pd (com) catalysts.
For the electrocatalytic CO2RR stability test on PdꢀB/C and Pd/C (com) at ꢀ0.5 V in
CO2ꢀsaturated 0.1 M KHCO3, the chronoamperometric curves as well as the average formate
FEs over different elapsing times up to 5 h were comparatively shown in Figure S5 in the SI.
Tables S1 also lists the mass activities of formate production on these two catalysts over
different times. It can be seen that PdꢀB/C exhibits a higher stabilized total current as well as
higher formate FEs and mass activities than Pd/C does over the 5ꢀh period. The higher stability
of the PdꢀB/C for CO2RR could be attributed to its better resistance to nanoparticle size growth
as indicated from TEM characterization (Figure S1) as well as to CO posioning.22,23
3.4 DFT Calculations on Relative Stability of HCOO* and *COOH Intermediates. To
elucidate the experimentally observed enhancement of formate production on PdꢀB/C, we
theoretically estimated relative stability of reaction intermediates for the HCOOH and CO
pathways of CO2RR on Pd(111) surfaces with and without Bꢀdoping, where HCOOH was
considered as the final product to describe this reaction instead of formate, in line with the DFT
14
ACS Paragon Plus Environment

Page 15 of 24
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
calculation practice by many researchers.3,49ꢀ51 We first determined the configuration for B
doping in Pd(111) by carrying out DFT calculations based on a 2×2 supercell and fiveꢀlayer
slab model, revealing that a full coverage of B atoms in the subsurface of Pd(111) is the
optimized configuration, in agreement with that in a previous report,52 see Figure S6 in the SI
for more details. The Bꢀdoping configuration is referred as Pd(111)ꢀ4B, and the Pd(111) surface
may slightly deform with such a configuration.
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
The conversion from CO2 to FA on a Pd surface proceeds with overall two electrons and
two protons,3,51,53 so does the competitive reaction from CO2 to CO.54 Accordingly, the two
parallel pathways typically involving HCOO* and *COOH intermediates were proposed,11,51
leading to FA and CO, respectively, as illustrated from the following equations where steps 6
and 9 were assumed to be the rateꢀlimiting steps, respectively,55 for the two pathways:
In the FA pathway, 2,3,56
CO2
＋
e－
＋
H+
＋
* → HCOO*
(6)
(7)
(8)
HCOO*
＋
e－
＋
H+ → HCOOH*
HCOOH* → HCOOH(aq)
In the CO pathway, 1,29,48,56,57
CO2
*COOH
＋ *
＋
e－
e－
*CO → CO(g)
＋
H+
＋
* → *COOH
(9)
(10)
(11)
＋
＋
H+ → *CO + H2O(l)
＋
*
Our DFT calculations were first focused on finding relative stable adsorption
configurations for these two intermediates on the two parallel pathways, as a base for the
discussion of the Bꢀdoping effect on the relative selectivity of the formation of FA versus CO in
the CO2RR process. Figure 5 presents the most probable adsorption configurations of HCOO*
(Figure 5A and 5B) and *COOH (Figure 5C and 5D) species on Pd(111) (upper row) and
Pd(111)ꢀ4B (lower row) surfaces, and Table 1 lists the calculated adsorption energies.
In agreement to previous reports,29,51 our calculated results suggest that the HCOO*
intermediate is shortꢀbridge bounded on Pd atoms via two oxygen atoms (Figure 5A) is
energetically more favorable than that singly bonded with one oxygen atom on the top of one
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 24
1
2
3
4
5
6
7
8
9
Pd atom (Figure 5B), about ꢀ17.58 and ꢀ17.48 kcal molꢀ1 on Pd(111) and Pd(111)ꢀ4B,
respectively. For the *COOH intermediate with the C atom bounded on one Pd atom, our
calculated results indicate that the configurations on Pd(111) (upper row) and Pd(111)ꢀ4B
(lower row) surfaces as shown in Figure 5C with the OꢀH bond pointing downward are ꢀ4.38
and ꢀ1.07 kcal molꢀ1 energetically more favorable than the ones with the OꢀH bond pointing
upward, respectively (shown in Figure 5D). Therefore, the two intermediates with surface
configurations illustrated in Figures 5A and 5C are mainly responsible for the CO2RR on
Pd(111) surfaces to yield the products FA and CO, respectively, and will be used for the
following discussion.
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 5. Optimized adsorption configurations of *COOH and HCOO* intermediates on (Ι)
Pd(111) and (ΙΙ) Pd(111)ꢀ4B surfaces. (A) HCOO* adsorbs on two shortꢀbridge Pd atoms via its
two O atoms. (B) HCOO* adsorbed on one Pd atom via single O atom. (C) *COOH adsorbs on
one Pd atom via its C atom with a downward H. (D) *COOH adsorbs on one Pd atom via its C
atom with an upward H. The blue, pink, red, grey, and white balls represent Pd, B, O, C, and H,
respectively.
Next we will focus on the influence of the Bꢀdoping on the relative stability of the two
intermediates with the configurations indicated in Figures 5A and 5C. As shown in Table 1, the
16
ACS Paragon Plus Environment

Page 17 of 24
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
calculated adsorption energies of HCOO* and *COOH are ꢀ2.77 eV and ꢀ2.86 eV on regularly
bare Pd(111) surface, respectively, suggesting that the *COOH is energetically more favorable
than the HCOO*. In contrast, the adsorption energies significantly change to ꢀ2.74 eV and
ꢀ2.25 eV for HCOO* and *COOH species on Pd(111)ꢀ4B surface, respectively. The adsorption
energy changes for the two intermediates clearly indicate that the Bꢀdoping thermodynamically
directs the CO2RR in favor of the FA pathway with an inhibition of the CO pathway. On Pd(111)
surface, the Gibbs free energy of *COOH is 2.33 kcal molꢀ1 more stable than that of HCOO*.
In contrast, on Pd(111)ꢀ4B surface, the Gibbs free energy gap of the two intermediates changes
to ꢀ10.12 kcal molꢀ1. These data show that Bꢀdoping in the subsurface of Pd(111) reverses the
sequence of thermodynamic stability of the two intermediates, in a trend more favorable for the
HCOO* than for the *COOH, which is consistent with the above experimental observation.
H adsorption and absorption were recognized to play an important role in CO2RR on Pd
surfaces, which distinguishes this metal from others. Along this line, we further calculated the
adsorption energies of the two intermediates in the CO2RR on Hꢀadsorbed Pd(111) surfaces.
Here, for the Bꢀundoped Pd(111), fullꢀcoverage H absorption in the subsurface was considered
in addition to its surface adsorption; for the Bꢀdoped Pd(111), only surface H adsorption and no
H absorption in the subsurface was assumed in the calculation due to the existing fullꢀcoverage
B atoms. These optimized structures are shown in Figure S7 in SI, where hydrogen atoms
occupied at the surface hollow site and the octahedral subsurface site, respectively.24 And
Figure S8 indicates that the CO2 adopts a flat configuration parallel to the Hꢀadsorbed metal
surface. It turns out that the adsorption energies of HCOO* and *COOH are ꢀ1.55 eV and ꢀ1.99
eV on Hꢀadsorbed Pd(111) while on Hꢀadsorbed Pd(111)ꢀ4B, they become ꢀ2.51 eV and ꢀ2.26
eV, respectively. In addition, the Gibbs free energy of *COOH relative to that of HCOO* are
12.73 kcal molꢀ1 on the Hꢀadsorbed Pd(111) surface, and ꢀ0.93 kcal molꢀ1 on the Hꢀadsorbed
Pd(111)ꢀ4B surface (see Table S4 for more details). Again, these calculated results strongly
indicate that the B doping makes HCOO* thermodynamically more stable than *COOH on
Hꢀadsorbed Pd(111) surfaces, in further support of the experimental observation of the
enhanced formate selectivity on PdꢀB/C over a wide potential window.
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
Table 1. Calculated adsorption energies (eV) of HCOO* and *COOH on Pd(111) and
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 24
1
2
3
4
Pd(111)ꢀ4B surfaces without and with H adsorption.
5
6
7
8
9
Pd(111)
ꢀ2.77
Pd(111)ꢀ4B
ꢀ2.74
HꢀPd(111)
ꢀ1.55
HꢀPd(111)ꢀ4B
ꢀ2.51
HCOO*
*COOH
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
ꢀ2.86
ꢀ2.25
ꢀ1.99
ꢀ2.26
Figure 6 presents the Gibbs free energy diagrams for the intermediates and products along
the reaction coordinates from an initial state of * + CO2 + 2H+ + 2e− to a final state of either
HCOOH + * or CO + H2O + * on Pd(111) and Pd(111)ꢀ4B surfaces without (Figure 6A and
Figure 6C) and with (Figure 6B and Figure 6D) Hꢀadsorbate. Detailed Gibbs free energies and
related correction parameters for gas phase molecules and adsorption species involved in the
relevant calculations are listed in Tables S3, S4 and S5 in the SI. The calculated results show
that the Gibbs free energy changes from the common initial state to the two intermediate states
(i.e., the proposed rateꢀlimiting step55) are significantly influenced by adsorbed/absorbed H and
doped B.
For the FA pathway (upper panel in Figure 6), regardless of H adsorption, the B doping
decreases the Gibbs free energy for the intermediate state HCOO* + H+ + e− on Pd(111) and
Pd(111)ꢀ4B surfaces in the rate determining step, as shown in Figure 6A (without Hꢀadsorbate)
and Figure 6B (with Hꢀadsorbate), in favor of the FA pathway. Alternatively, for the CO
pathway (lower panel in Figure 6), Figure 6C suggests that without H adsorbate the Gibbs free
energy for the intermediate state *COOH + H+ + e− increases with the Bꢀdoping, resulting in an
inhibited production of CO on the Pd(111)ꢀ4B surfaces. The Hꢀadsorption changes the order of
Gibbs free energies for the *COOH + H+ + e− on Pd(111) and Pd(111)ꢀ4B as shown in Figure
6D, in fact the Gibbs free energy of the intermediate state in the CO pathway decreases on
HꢀPd(111)ꢀ4B as compared to that on HꢀPd(111). Nevertheless, a closer look at Figure 6B, it
suggests in the same time a more significant decrease of Gibbs free energy for HCOO* + H+ +
e− on HꢀPd(111)ꢀ4B. Therefore, the Bꢀdoping promotes the FA pathway more than the CO
pathway on Pd(111) surfaces, regardless of what surface H coverage is used for DFT
calculations. Indeed, this is consistent with the above experimental results on Pd/C and PdꢀB/C.
18
ACS Paragon Plus Environment

Page 19 of 24
Journal of the American Chemical Society
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 6. Gibbs free energy diagrams for the CO2RR on Pd(111) (black) and Pd(111)ꢀ4B (red)
surfaces along the reaction coordinate from the initial state * + CO2 + 2H+ + 2e− to the final
state HCOOH + * (A and B) and CO + H2O + * (C and D) on Pd(111) and Pd(111)ꢀ4B surfaces
without (A and C) and with (B and D) H adsorbate, respectively. All Gibbs free energies are
shown relative to the initial state.
An additional evidence to support the enhanced FA pathway comes from the experimental
and theoretical results on different COꢀadsorption strengths on Pd and PdꢀB surfaces. Our XPS
result suggested a lower Pd dꢀband center with the Bꢀdoping, and the anodic CO stripping
result confirmed a weakened CO adsorption on Pd with the Bꢀdoping.23 The present DFT
calculation also points to a weakened CO adsorption on the Pd surface with the Bꢀdoping, as
can be seen from Figure 6C and Tables S6, S7 and S8 in the SI. All these data are directed to an
unfavorable effect on CO formation on PdꢀB surfaces as compared to that on a pure Pd surfaces.
The lessened CO formation on PdꢀB surfaces is accompanied with the enhanced FA formation
since the CO pathway and the FA pathway are parallel. This may be clearly demonstrated in
Figure 3A and B especially at higher overpotentials.
Figure 7 presents the projected density of states (PDOS) for the d band of the surface Pd
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 24
1
2
3
4
5
6
7
8
9
atoms. The d band centers of surface Pd atoms on Pd(111), Pd(111)ꢀ4B, HꢀPd(111) and
HꢀPd(111)ꢀ4B surfaces are estimated about ꢀ1.48, ꢀ2.04, ꢀ1.83, and ꢀ2.30 eV with respect to the
Fermi level, respectively. The downꢀshift of the d band centers increases the donation bonding
states and decreases the backꢀdonation antiꢀbonding states, weakening the substrateꢀadsorbate
interaction, especially the adsorption of *COOH or *CO on Pd surfaces. The PDOS with
HCOO* and *COOH intermediates adsorbed on Pd(111) are shown in Figure S9. When B
atoms are doped in the Pd(111) subsurface, the PDOS of the Pd d band, which is closer to the
Fermi level, matches more with the molecular orbitals of the HCOO* than with those of the
*COOH. In other words, the adsorption energy of HCOO* is larger than that of *COOH on
Bꢀdoped Pd surfaces. The comparison of Pd electronic structure changes again confirms that
the B doping enhances the FA pathway.
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 7. The PDOS plots of d bands of surface Pd layer. (A) Pd(111), (B) Pd(111)ꢀ4B, (C)
HꢀPd(111) and (D) HꢀPd(111)ꢀ4B.
20
ACS Paragon Plus Environment

Page 21 of 24
Journal of the American Chemical Society
1
2
3
4
4. CONCLUSIONS
5
6
7
8
We have demonstrated that the B doping in the Pd lattice interstices significantly impacts
the pathways of the CO2RR. Remarkably higher selectivity toward formate relative to CO in
the product is found experimentally on PdꢀB/C than on Pd/C. Specifically, a Faradaic efficiency
of ca. 70% for formate after 2 h of electrolysis in 0.1 M KHCO3 at ꢀ0.5 V has been obtained on
PdꢀB/C, which is ca. 12 times as high as that on Pd/C under otherwise same conditions, leading
to ca. 234 mMHCOOꢀ mgꢀ1Pd, ca.18 times as high as that on Pd/C. Comparative DFT calculations
on adsorption energies of the key intermediates HCOO* and *COOH for the two parallel
CO2RR reaction pathways confirm the observed Bꢀdoping effect: on both Hꢀadsorbed and bare
Pd(111) surfaces, the adsorption energy of HCOO* relative to that of *COOH becomes more
negative with B doping in the subsurface of Pd(111). In other words, the relative adsorption
energy shift is energetically more favorable for the FA pathway. Furthermore, the Gibbs free
energy along the reaction coordinate plots indicate that the Gibbs free energy change from the
initial state to the intermediate state, decreases with the B doping, irrespective of Hꢀcoverage
on Pd(111)ꢀ4B surfaces, in favor of the FA pathway. The DFT calculations indeed corroborate
well with the experimental results. In considering that PdꢀB/C is also more efficient for
dehydrogenation of FA both chemically and electrochemically, the present study confers
PdꢀB/C a unique bifunctional catalyst for the HCOOH ↔ CO2 cycle.
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
ASSOCIATED CONTENT
Supporting Information
Details for catalyst characterizations after electrolysis, CO2RR products analysis, longꢀterm
chronoamperometric curves and average FEs, computational model of boron modification in Pd(111), the
Hꢀadsorbed effect on the CO2RR on Pd surfaces, summary of simulation parameters as calculated from DFT,
adsorption geometries and energies of *CO, projected DOS in the presence of molecular adsorption. The
supporting information is available free of charge on the ACS Publications website at http://pubs.acs.org
.
AUTHOR INFORMATION
Corresponding Authors
*Wen-Bin Cai: wbcai@fudan.edu.cn
*De-Yin Wu: dywu@xmu.edu.cn
Notes
The authors declare no competing financial interest.
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 24
1
2
3
4
ACKNOWLEDGMENTS
B.J. and W.B.C. acknowledge the Natural Science Foundation of China (No. 21733004 and 21473039).
X.G.Z. and D.Y.W. acknowledge the Natural Science Foundation of China (No. 21533006, 21621091 and
21373172) and Funds of State Key Laboratory of Physical Chemistry of Solid Surfaces. We thank the
financial supports from the 973 Program of the Chinese Ministry of Science and Technology (No.
2015CB932303).
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
REFERENCES
(1) Chaplin, R. P. S.; Wragg, A. A. J. Appl. Electrochem. 2003, 33, 1107.
(2) Zhang, S.; Kang, P.; Meyer, T. J. J. Am. Chem. Soc. 2014, 136, 1734.
(3) Klinkova, A.; De Luna, P.; Dinh, C. T.; Voznyy, O.; Larin, E. M.; Kumacheva, E.; Sargent, E. H. ACS
Catal. 2016, 6, 8115.
(4) Zhou, X. H.; Liu, R.; Sun, K.; Chen, Y. K.; Verlage, E.; Francis, S. A.; Lewis, N. S.; Xiang, C. X. ACS
Energy Lett. 2016, 1, 764.
(5) Kortlever, R.; Peters, I.; Koper, S.; Koper, M. T. M. ACS Catal. 2015, 5, 3916.
(6) Kapusta, S.; Hackerman, N. J. Electrochem. Soc. 1983, 130, 607.
(7) Whipple, D. T.; Kenis, P. J. A. J. Phys. Chem. Lett. 2010, 1, 3451.
(8) Hori, Y. Electrochemical CO2 Reduction on Metal Electrodes; Modern Aspects of Electrochemistry;
Springer, New York, 2008, 42, 89ꢀ189.
(9) Lu, Q.; Jiao, F. Nano Energy 2016, 29, 439.
(10) Li, W.; In Advances in CO2 Conversion and Utilization: Electrocatalytic Reduction of CO2 to Small
Organic Molecule Fuels on Metal Catalysts; American Chemical Society: Washington, DC, 2010, 1056, 55ꢀ76;
Chapter 5.
(11) Li, Q.; Fu, J. J.; Zhu, W. L.; Chen, Z. Z.; Shen, B.; Wu, L. H.; Xi, Z.; Wang, T. Y.; Lu, G.; Zhu, J. J.; Sun,
S. H. J. Am. Chem. Soc. 2017, 139, 4290.
(12) Lee, C. H.; Kanan, M. W. ACS Catal. 2015, 5, 465.
(13) Gao, S.; Lin, Y.; Jiao, X.; Sun, Y.; Luo, Q.; Zhang, W.; Li, D.; Yang, J.; Xie, Y. Nature 2016, 529, 68.
(14) Mazumder, V.; Sun, S. H. J. Am. Chem. Soc. 2009, 131, 4588.
(15) Wang, J. Y.; Zhang, H. X.; Jiang, K.; Cai, W. B. J. Am. Chem. Soc. 2011, 133, 14876.
(16) Adams, B. D.; Asmussen, R. M.; Ostrom, C. K.; Chen, A. J. Phys. Chem. C 2014, 118, 29903.
(17) Chen, A. C.; Ostrom, C. Chem. Rev. 2015, 115, 11999.
(18) Min, X.; Kanan, M. W. J. Am. Chem. Soc. 2015, 137, 4701.
(19) Gao, D.; Zhou, H.; Cai, F.; Wang, D.; Hu, Y.; Jiang, B.; Cai, W.ꢀB.; Chen, X.; Si, R.; Yang, F.; Miao, S.;
Wang, J.; Wang, G.; Bao, X. Nano Res. 2017, 10, 2181.
(20) Zhou, X.; Liu, R.; Sun, K.; Chen, Y.; Verlage, E.; Francis, S. A.; Lewis, N. S.; Xiang, C. ACS Energy
Lett. 2016, 1, 764.
(21) Gao, D.; Zhou, H.; Wang, J.; Miao, S.; Yang, F.; Wang, G.; Wang, J.; Bao, X. J. Am. Chem. Soc. 2015,
137, 4288.
(22) Jiang, K.; Xu, K.; Zou, S.; Cai, W. B. J. Am. Chem. Soc. 2014, 136, 4861.
(23) Jiang, K.; Chang, J.; Wang, H.; Brimaud, S.; Xing, W.; Behm, R. J.; Cai, W. B. ACS Appl. Mater. Inter.
2016, 8, 7133.
(24) Jiang, K.; Zhang, H. X.; Zou, S.; Cai, W. B. Phys. Chem. Chem. Phys. 2014, 16, 20360.
22
ACS Paragon Plus Environment

Page 23 of 24
Journal of the American Chemical Society
1
2
3
(25) Ohkawa, K.; Hashimoto, K.; Fujishima, A.; Noguchi, Y.; Nakayama, S. J. Electroanal. Chem. 1993, 345,
4
445.
5
6
7
8
(26) Gao, D.; Wang, J.; Wu, H.; Jiang, X.; Miao, S.; Wang, G.; Bao, X. Electrochem. Commun. 2015, 55, 1.
(27) Sheng, W. C.; Kattel, S.; Yao, S. Y.; Yan, B. H.; Liang, Z. X.; Hawxhurst, C. J.; Wu, Q. Y.; Chen, J. G. G.
Energy Environ. Sci. 2017, 10, 1180.
9
(28) Yin, Z.; Gao, D. F.; Yao, S. Y.; Zhao, B.; Cai, F.; Lin, L. L.; Tang, P.; Zhai, P.; Wang, G. X.; Ma, D.; Bao,
X. H. Nano Energy 2016, 27, 35.
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
(29) Huang, H.; Jia, H.; Liu, Z.; Gao, P.; Zhao, J.; Luo, Z.; Yang, J.; Zeng, J. Angew. Chem. 2017, 56, 3594.
(30) Stalder, C. J.; Chao, S.; Wrighton, M. S. J. Am. Chem. Soc. 1984, 106, 3673.
(31) Podlovchenko, B. I.; Kolyadko, E. A.; Lu, S. G. J. Electroanal. Chem. 1994, 373, 185.
(32) Wang, J.ꢀY.; Kang, Y.ꢀY.; Yang, H.; Cai, W.ꢀB. J. Phys. Chem. C 2009, 113, 8366.
(33) Li, J.; Chen, J.; Wang, Q.; Cai, W.ꢀB.; Chen, S. Chem. Mater. 2017, 29, 10060.
(34) Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. Energy Environ. Sci. 2012, 5, 7050.
(35) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865.
(36) Kresse, G.; Furthmüller, J. Phys. Rev. B 1996, 54, 11169.
(37) Blöchl, P. E. Phys. Rev. B 1994, 50, 17953.
(38) Hong, S.; Rahman, T. S. Phys. Rev. B 2007, 75, 155405.
(39) Tang, W.; Sanville, E.; Henkelman, G. J. Phys-Condens. Mat. 2009, 21.
(40) Bahn, S. R.; Jacobsen, K. W. Comput. Sci. Eng. 2002, 4, 56.
(41) Peterson, A. A.; AbildꢀPedersen, F.; Studt, F.; Rossmeisl, J.; Norskov, J. K. Energy Environ. Sci. 2010, 3,
1311.
(42) Benfield, R. E. J. Chem. Soc. Faraday Trans. 1992, 88, 1107.
(43) Wang, M. Z.; Qin, X.; Jiang, K.; Dong, Y.; Shao, M.; Cai, W.ꢀB. J. Phys. Chem. C 2017, 121, 3416.
(44) Ohkawa, K.; Noguchi, Y.; Nakayama, S.; Hashimoto, K.; Fujishima, A. J. Electroanal. Chem. 1994, 367,
165.
(45) Wuttig, A.; Yoon, Y.; Ryu, J.; Surendranath, Y. J. Am. Chem. Soc. 2017, 139, 17109.
(46) Dunwell, M.; Lu, Q.; Heyes, J. M.; Rosen, J.; Chen, J. G. G.; Yan, Y. S.; Jiao, F.; Xu, B. J. J. Am. Chem.
Soc. 2017, 139, 3774.
(47) Zhu, S.; Jiang, B.; Cai, W. B.; Shao, M. J. Am. Chem. Soc. 2017, 139, 15664.
(48) Jiang, K.; Siahrostami, S.; Akey, A. J.; Li, Y.; Lu, Z.; Lattimer, J.; Hu, Y.; Stokes, C.; Gangishetty, M.;
Chen, G.; Zhou, Y.; Hill, W.; Cai, W.ꢀB.; Bell, D.; Chan, K.; Nørskov, J. K.; Cui, Y.; Wang, H. Chem. 2017, 3, 950.
(49) Peng, G. W.; Sibener, S. J.; Schatz, G. C.; Mavrikakis, M. Surf. Sci. 2012, 606, 1050.
(50) Feaster, J. T.; Shi, C.; Cave, E. R.; Hatsukade, T. T.; Abram, D. N.; Kuhl, K. P.; Hahn, C.; Norskov, J. K.;
Jaramillo, T. F. ACS Catal. 2017, 7, 4822.
(51) Yoo, J. S.; Christensen, R.; Vegge, T.; Norskov, J. K.; Studt, F. ChemSusChem 2016, 9, 358.
(52) Yoo, J. S.; Zhao, Z. J.; Nørskov, J. K.; Studt, F. ACS Catal. 2015, 5, 6579.
(53) Peng, G. W.; Sibener, S. J.; Schatz, G. C.; Ceyer, S. T.; Mavrikakis, M. J. Phys. Chem. C 2012, 116,
3001.
(55) Wuttig, A.; Yaguchi, M.; Motobayashi, K.; Osawa, M.; Surendranath, Y. P. Natl. Acad. Sci. USA 2016,
113, 4585.
(54) Dunwell, M.; Yan, Y.; Xu, B. ACS Catal. 2017, 7, 5410.
(56) Hori, Y.; Wakebe, H.; Tsukamoto, T.; Koga, O. Electrochim. Acta 1994, 39, 1833.
(57) Sen, S.; Liu, D.; Palmore, G. T. R. ACS Catal. 2014, 4, 3091.
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 24
1
2
3
4
5
6
Table of Contents (TOC) Graphic
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
24
ACS Paragon Plus Environment