Mechanistic Insights into Electrochemical Nitrogen Reduction Reaction on Vanadium Nitride Nanoparticles

Article abstract of DOI:10.1021/jacs.8b08379

Renewable production of ammonia, a building block for most fertilizers, via the electrochemical nitrogen reduction reaction (ENRR) is desirable; however, a selective electrocatalyst is lacking. Here we show that vanadium nitride (VN) nanoparticles are active, selective, and stable ENRR catalysts with an ENRR rate and a Faradaic efficiency (FE) of 3.3 × 10^-10 mol s^-1 cm^-2 and 6.0% at -0.1 V within 1 h, respectively. ENRR with ¹⁵N₂ as the feed produces both ¹⁴NH₃ and ¹⁵NH₃, which indicates that the reaction follows a Mars-van Krevelen mechanism. Ex situ X-ray photoelectron spectroscopy characterization of fresh and spent catalysts reveals that multiple vanadium oxide, oxynitride, and nitride species are present on the surface and identified VN_0.7O_0.45 as the active phase in the ENRR. Operando X-ray absorption spectroscopy and catalyst durability test results corroborate this hypothesis and indicate that the conversion of VN_0.7O_0.45 to the VN phase leads to catalyst deactivation. We hypothesize that only the surface N sites adjacent to a surface O are active in the ENRR. An ammonia production rate of 1.1 × 10^-10 mol s^-1 cm^-2 can be maintained for 116 h, with a steady-state turnover number of 431.

Full text of DOI:10.1021/jacs.8b08379

Subscriber access provided by Duquesne University - Gumberg Library
Article
Mechanistic Insights into Electrochemical Nitrogen
Reduction Reaction on Vanadium Nitride Nanoparticles
Xuan Yang, Jared Nash, Jacob Anibal, Marco Dunwell, Shyam Kattel, Eli
Stavitski, Klaus Attenkofer, Jingguang G Chen, Yushan Yan, and Bingjun Xu
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b08379 • Publication Date (Web): 23 Sep 2018
Downloaded from http://pubs.acs.org on September 23, 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.
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 6
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Mechanistic Insights into Electrochemical Nitrogen Reduction Reaction
on Vanadium Nitride Nanoparticles
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
Xuan Yang,†,‡ Jared Nash,†,‡ Jacob Anibal,† Marco Dunwell,† Shyam Kattel,§,# Eli Stavitski,¶ Klaus Atꢀ
tenkofer,¶ Jingguang G. Chen,§,#,* Yushan Yan,†,* and Bingjun Xu†,*
†Center for Catalytic Science and Technology, Department of Chemical and Biomolecular Engineering, University of Delaꢀ
ware, 150 Academy Street, Newark DE, 19716
§Chemistry Division, Brookhaven National Laboratory, Upton, NY 11973, USA
#Department of Chemical Engineering, Columbia University, New York, NY 10027, USA.
¶National Synchrotron Light Source II, Brookhaven National Laboratory, Upton, New York 11973, USA
‡These authors contributed equally to this work.
*Address correspondence to jgchen@columbia.edu, yanys@udel.edu, and bxu@udel.edu
KEYWORDS: Electrochemical nitrogen reduction, vanadium nitride, Mar-van Krevelen mechanism, operando XAS, 15N2
ABSTRACT: Renewable production of ammonia, a building block for most fertilizers, via the electrochemical nitrogen reduction
reaction (ENRR) is desirable; however, selective electrocatalyst is lacking. Here we show that vanadium nitride (VN) nanoparticles
are active, selective, and stable ENRR catalysts with an ENRR rate and Faradaic efficiency (FE) of 3.3 × 10−10 mol s−1 cm−2 and
6.0% at −0.1 V within 1 h, respectively. ENRR with 15N2 as the feed produce both 14NH3 and 15NH3, which indicates that the reacꢀ
tion follows a Marsꢀvan Krevelen mechanism. Ex-situ Xꢀray photoelectron spectroscopy characterization of fresh and spent cataꢀ
lysts reveals that multiple vanadium oxide, oxynitride and nitride species are present on the surface and identified VN0.7O0.45 as the
active phase in ENRR. Operando Xꢀray absorption spectroscopy and catalyst durability test results corroborate with this hypotheꢀ
sis, and indicate that the conversion of VN0.7O0.45 to the VN phase leads to catalyst deactivation. We hypothesize that only the surꢀ
face N sites adjacent to a surface O are active in the ENRR. An ammonia production rate of 1.1 × 10−10 mol s−1 cm−2 can be mainꢀ
tained for 116 h, with a steady state turnover number of 431.
INTRODUCTION
Bosch process, to make nitric acid via electric discharge.8 Furꢀ
ther studies of ENRR have generally focused on high temperaꢀ
ture proton conductors at 773 K; however, the instability of
ammonia at high operating temperature makes it unsuitable for
distributed deployment.9 More recently, there has been a surge
of interest in developing low temperature ENRR processes.
Proton exchange membranes (PEMs) can achieve high proton
conductivity (~100 mS⋅cm−1 at 25 °C for Nafion) at ambient or
modestly elevated temperatures, which opens the possibility
for electrochemical ammonia synthesis at those mild condiꢀ
tions.10 Due to the acidic nature of PEMs, platinum group
metal (PGM) catalysts are the most commonly used catalysts
because of their excellent stability in acidic environments.11−15
Our recent work showed that PGM catalysts generally have
low Faradaic efficiencies (FE <0.1%) for the ENRR with a
membrane electrode assembly (MEA) configuration,13 which
is consistent with computational predictions that metals and
metal alloys are unlikely to be selective for the ENRR.16−18
The low Faradaic efficiency (FE) of PGM catalysts is attributꢀ
ed to the correlated binding energies of various surface interꢀ
mediates in the ENRR and the competing hydrogen evolution
reaction (HER), known as the linear scaling relations.18
Meanwhile, higher FEs on metal catalysts have been reported
in the recent literature mostly in aqueous electrolytes.14,15,19−22
Recent theoretical calculations have suggested that transition
metal nitrides are promising candidates for the ENRR via a
The ability to fixate nitrogen on an industrial scale marked a
turning point in the history of the human society, which supꢀ
ports the livelihood of close to half of world’s population.1
The HaberꢀBosch process operates at high temperatures (~700
K) and pressures (~100 atm) on optimized catalysts,2−7 which
accounts for ~1% of the global energy consumption.2 Hydroꢀ
gen consumed in the HaberꢀBosch process derives exclusively
from the steam reforming of fossil hydrocarbons, making it
also a carbon intensive process.5 Further, the centralized amꢀ
monia production is a poor fit for the distributed nature of
agriculture. Distributed and modular ammonia synthesis via
the electrochemical nitrogen reduction reaction (ENRR) at or
close to ambient conditions, powered by renewable electricity,
emerges as an attractive alternative because it allows for the
onꢀdemand, onꢀsite production of ammonia, and in turn Nꢀ
fertilizers, from ubiquitously available resources, i.e., N2 and
water.3 Widespread adoption of ENRR for ammonia producꢀ
tion could drastically reduce the carbon footprint of agriculturꢀ
al activities. The ENRR is also compatible with the intermitꢀ
tency of renewable energy sources, e.g., solar and wind, as
ammonia and Nꢀfertilizers can be produced and stored when
renewable electricity is abundant or even in surplus.
Electrochemical fixation of atmospheric nitrogen was first
attempted in 1908, before the establishment of the Haberꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
of VN@C is determined to be 177 m2 g−1 with the BET methꢀ
Page 2 of 6
Marsꢀvan Krevelen mechanism under ambient conditions,
however, experimental confirmation has been lacking.23−27
1
A
od (Figure S1ꢀ1).
2
3
4
5
6
7
8
9
key analytical challenge for ENRR studies is the small amount
of ammonia produced, which could be particularly misleading
for Nꢀcontaining catalyst such as metal nitrides. This chalꢀ
lenge can be overcome by increasing the amount of ammonia
produced using the MEA configuration, enabling accurate rate
and FE determination by establishing a quantitative N mass
balance.
Xꢀray photoelectron spectroscopy (XPS) results reveal that
multiple nitride, oxynitride, and oxide species exist on the
surface of VN@C. The survey spectrum confirms the presence
of V and N, along with C and O (Figure S1ꢀ2). The V 2p3/2
XPS region can be deconvoluted into four peaks at 513.6,
515.3, 516.0, and 517.1 eV, which are assigned to V in VN,
VNxOy, VOx (x ≤ 2), and V2O5, respectively (Table S4ꢀ1).29−31
The presence of oxide and oxynitride species on the surface of
metal nitrides is consistent with previous reports.29−31 The N 1s
XPS region (Figure S1ꢀ2C) can also be deconvoluted into four
bands, with the band at 397.3 eV being assigned to N in
VN.29,31,32 The other three N 1s peaks centered at 396.4, 398.4,
and 401.0 eV are assigned to the VNxOy, a satellite peak, and
ammonia/ammonium species, respectively, based on the literaꢀ
ture (Table S4ꢀ2).29,31−33 The small amount of ammoꢀ
nia/ammonium species adsorbed on the surface are likely reꢀ
siduals from urea in the synthesis procedure.
In this work, we demonstrate that vanadium nitride (VN)
nanoparticles are active and selective catalysts for ENRR.
Both the ammonia production rate and the FE are approxiꢀ
mately two orders of magnitude higher than those of noble
metal catalysts. Through a combination of ex-situ and operan-
do characterizations, vanadium oxynitride (VN0.7O0.45) is idenꢀ
tified as the active phase, and the catalyst deactivation is
caused by its conversion to VN. The ENRR is proposed to
proceed via a Marsꢀvan Krevelen mechanism based on results
obtained using 15N2 as the feed in the ENRR. We show that an
ammonia production rate and FE of 1.1 × 10−10 mol s−1 cm−2,
with an FE of 1.6%, can be maintained for 116 h on VN at
−0.1 V.
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
RESULTS AND DISCUSSION
Figure 2. (A) Production rate and Faradaic efficiency of VN cataꢀ
lysts at different potentials for 1 h tests. (B) Timeꢀdependent proꢀ
duction rate and Faradaic efficiency at −0.1 V and −0.2 V for 8 h
tests, respectively. (C) Timeꢀdependent production rate and Faraꢀ
daic efficiency at −0.1 V for 120 h tests, respectively. (D) 1H
nuclear magnetic resonance (NMR) spectra for the postꢀ
electrolysis 0.05 M H2SO4 at −0.1 V in 15N2 and Ar, and at OCP
in 14N2 for 48 h.
Figure 1. (A and B) typical SEM images of VN nanoparticles at
different magnifications. TEM image (C) and XRD pattern (D) of
VN nanoparticles.
Ammonia production rates on VN nanoparticles are evaluꢀ
ated using an MEA configuration to obtain reactivity data
representative of catalytic performance at the device level
(Figure S1ꢀ6). VN nanoparticles (0.5 mg cm−2) and Pt/C (0.4
mgPt cm−2) loaded carbon paper are used as the cathode and
anode, respectively, during ENRR. It is worth noting that the
anode is operating in the H2 atmosphere (1 atm) with a high Pt
loading, and the small overpotential of the hydrogen oxidation
reaction on Pt makes the anode function as a RHE. Thus, the
cell potential is numerically identical to the cathodic potential
versus RHE.13 All reported potentials in this work are referꢀ
enced to the RHE unless noted otherwise. The amount of amꢀ
monia produced in ENRR is quantified by establishing a rigꢀ
orous nitrogen balance encompassing all Nꢀcontaining species
Monodisperse VN nanoparticles with an average size of 10
nm have been synthesized via a ureaꢀassisted pathway.28
Scanning electron microscopy (SEM) images of asꢀ
synthesized VN sample show that the size of nanoparticles is
around 10−20 nm (Figure 1, A and B), while the TEM images
suggest a particle size range of 6−15 nm (Figure 1C). The
discrepancy between SEM and TEM results suggests that VN
nanoparticles are covered by a carbon shell of 2−4 nm, which
has been attributed to the residual carbon as a side product of
the synthesis procedure.28 Xꢀray diffraction (XRD) pattern
(Figure 1D) of the asꢀsynthesized VN@C matches well with
that of cubic VN (JCPDS No. 25ꢀ1252).28,29 The surface area
ACS Paragon Plus Environment

Page 3 of 6
Journal of the American Chemical Society
in the system via a combination of the Nessler method and
by either the V 2p or the N 1s features, decreases from ~0.91
before ENRR to 0.54 after ENRR at potentials ≤ −0.2 V (Figꢀ
ure 3C), while it is 0.77 after 1 h of ENRR at −0.1 V. Control
experiments show that V2O3 is not active for ENRR, and V2O5
is unstable at any negative potentials (Table S6ꢀ1).36 Thereꢀ
fore, it is reasonable to conclude that the active phase in
ENRR is either VN or VNxOy. Since the fraction of V in the
VN phase increases after ENRR, enhanced ENRR rates would
be expected if VN is the active phase for the reaction. The
opposite is observed in the reactivity studies. Thus, we proꢀ
pose that VNxOy is the most likely active phase for ENRR.
The constant VNxOy/VN ratio in the spent catalysts after 1 h of
ENRR at ≤ −0.2 V correlate well with the reactivity results
that they are nearly or completely inactive for ENRR. It is
likely that the exposed VNxOy is entirely converted to VN
during ENRR, causing the catalyst deactivation. Our hypotheꢀ
sis is also consistent with the observation that only about half
of exposed VNxOy was converted during 1 h of ENRR at −0.1
V, leading to its slower deactivation. Based on the ratio of V
2p and N 1s bands (corrected with atomic sensitivity factors)
assigned to the VNxOy phase, an x value of 0.7 could be obꢀ
tained. If a +3 oxidation state of V in VNxOy is assumed, its
1
2
3
4
5
6
7
8
Elementar analysis, as detailed in Section 2 of the Supplemenꢀ
tary Information.34,35
The VN nanoparticles are approximately two orders of
magnitude more active and selective for ENRR than noble
metal catalysts.13 An ammonia production rate of 3.3 × 10−10
mol s−1 cm−2 (based on geometric area, or 6.6 × 10−10 mol s−1
mg−1) and an FE of 6.0% are measured on the VN catalyst at
−0.1 V within the first hour (Figure 2A). The VN catalysts
show the highest ammonia production rate (5.0 × 10−10 mol s−1
cm−2) and FE (6.5%) at −0.2 V within 1 h. At more negative
potentials, both the ammonia production rate and the FE deꢀ
crease significantly, which is attributed to the facile competing
HER. The stability of the VN catalyst in ENRR is highly deꢀ
pendent on the applied potential. The ammonia production on
VN at −0.1 V remains constant in the first two hours with a
decrease of 28% and 62% in the next two (3−4 h) and four
(5−8 h) hours, respectively, which is accompanied by a similar
reduction in the FE (Figure 2B). The ammonia production rate
at −0.1 V becomes stable after 4 h, at 1.1 × 10−10 mol s−1 cm−2
for as long as 116 h, while that at −0.2 V plummets more than
95% to 1.1 × 10−11 mol s−1 cm−2 after 2 h (Figure 2, B and C).
No detectable level of ammonia is detected after 1 h on VN at
potentials below −0.2 V. The total amount of ammonia proꢀ
duced in ENRR in the exhausts of cathode and anode, as well
as the MEA soaking solution, within 120 h at −0.1 V deterꢀ
mined by the Nessler method is consistent with that via the
temperature programmed decomposition with mass specꢀ
trometry (TPDꢀMS, Figure S2ꢀ4). Importantly, the ability of
VN to activate dinitrogen is confirmed by conducting ENRR
with 15N2 at −0.1 V in the batch mode in 0.05 M H2SO4, with
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
composition is calculated to be VN0.7O0.45
.
clear signal of 15NH4 , along with 14NH4 , in the H NMR
spectrum after 48 h (Figures 2D, S3ꢀ1, and S3ꢀ2). Control
+
+
1
Figure 3. XPS spectra of VN catalysts after nitrogen reduction at
−0.1 V for 1 h. (A) XPS deconvolutions of the V 2p region. The
five peaks in the V 2p region are attributed to VN, VNxOy, VOx,
V2O5, and oxygen related species from lower to higher binding
energy, respectively. (B) XPS deconvolutions of the N 1s region.
The four peaks in the N 1s region are attributed to VNxOy, VN,
satellite feature, and ammonia from lower to higher binding enerꢀ
gy, respectively. (C) The ratio of VNxOy to VN before and after
nitrogen reduction at different potentials for 1 h.
experiments show that a minor fraction of the 14NH4 detected
+
after ENRR could be attributed to leaching of N from VN,
which is detected when the VN is soaked in the same electroꢀ
lyte at the open circuit potential (OCP) for 48 h. The applied
potential in ENRR does not appear to impact the rate of N
leaching from VN, as the amount of N leached from VN after
it is held at −0.1 V in Ar is comparable to that at the OCP.
To identify the active phase and understand the reaction and
deactivation mechanisms, fresh and spent VN catalysts are
characterized. SEM images of the VN nanoparticles sprayed
on the carbon paper show an average particle size of ~30 nm,
which could be due to the agglomeration of nanoparticles with
the ionomer (Figure S1ꢀ7, A and B). No significant morphoꢀ
logical change is observed between the fresh and spent VN
catalysts. XRD patterns show that the bulk crystal structure of
VN catalysts remain after being sprayed onto carbon paper
and after ENRR (Figure S1ꢀ8). In contrast, XPS results reveal
significant differences in surface species before and after
ENRR. The V 2p and N 1s spectra of the VN catalyst loaded
on the carbon paper are similar to those of the asꢀsynthesized
catalyst, with VN, VNxOy, VOx, and V2O5 present on the surꢀ
face. The O 1s peak in the range of 530−540 eV is included in
all the V 2p spectra in order to obtain a good baseline.30−32
After 1 h ENRR at all potentials tested in this work, the intenꢀ
sities of the peaks corresponding to V2O5 and VOx decrease by
> 50% (Figure S4ꢀ1−S4ꢀ5), which are likely reduced by either
the negative electrode potential or the produced H2 via HER.
The intensity of the peak corresponding to VNxOy also deꢀ
creases after ENRR, indicating that it is also partially reduced.
Interestingly, the ratio between VNxOy and VN, as quantified
Operando Xꢀray absorption spectroscopic (XAS) studies
further support our hypothesis that VN0.7O0.45 is the active
phase for ENRR. Xꢀray absorption near edge spectroscopic
(XANES) results of the vanadium Kꢀedge are collected at
different potentials (Figures 4A−C and S5ꢀ1). The position
and intensity of the white line at all potentials investigated
remain constant throughout the experiment, indicating that no
detectable level of V leaching occurs during ENRR. The Fouꢀ
rier transform of the extended Xꢀray absorption fine structure
(FTꢀEXAFS) spectra of the catalysts (Figure S5ꢀ2) show two
peaks at 1.6 and 2.4 Å, which correspond to the V−N and
V−V bonds in cubic VN, respectively. The pronounced ampliꢀ
tude of the V−V shells at higher distances (2−7 Å) confirms
the wellꢀdefined longꢀrange order of the VN catalysts.37 The
XANES and FTꢀEXAFS results are consistent with the lack of
change in the XRD patterns of the fresh and spent catalysts
(Figure S1ꢀ7). The characteristic preꢀedge peak at 5468.4 eV
indicates the presence of oxynitride species,37,38 in good
agreement with the XPS results. Importantly, the intensity of
the preꢀedge peak decreases with time during ENRR at all
potentials studied, suggesting that the VN0.7O0.45 is gradually
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 6
show the preꢀedge and white line peaks at −0.1 V as a function of
converted during ENRR, and that the rate of conversion inꢀ
creases as the potential becomes more negative (Figure 4D). In
order to analyze and interpret the preꢀedge features, the backꢀ
ground of the edge under the preꢀedge features needs to be
removed (Figure 4B, left panel), which is achieved by fitting a
polynomial that smoothly reproduces the slope of the data
prior to the preꢀedge and immediately following the preꢀedge
peaks.38 35.7% of VN0.7O0.45 is converted at −0.1 V within 2 h,
while 57.8% conversion is achieved in 1 h at −0.2 V (Figure
4D). The fractions of VN0.7O0.45 that survives after 1 h of
ENRR at ≤ −0.2 V are similar, suggesting that all accessible
VN0.7O0.45 (~58%) has been converted to VN. This is conꢀ
sistent with our DFT calculations showing that oxygen vacanꢀ
cies on the VN0.75O0.25 (111) surface bind with N much more
strongly (by more than 2 eV) than nitrogen vacancies (Table
S7ꢀ1). The XAS results corroborate with the XPS and the reꢀ
activity results, e.g., the slow conversion of VN0.7O0.45 and
slow deactivation at −0.1 V.
1
2
3
4
5
6
7
8
time, respectively. The color scheme used in (A) applies to (B).
Inset in (C) is the corresponding preꢀedge peak. (D) Timeꢀ
dependent preꢀedge area at different potentials.
Our results suggest that ENRR on VN proceeds via a Marsꢀ
van Krevelen mechanism (Figure 5, right circle): an ammonia
molecule is formed by extracting a surface N atom with adꢀ
sorbed hydrogen atoms and leaves behind a N vacancy, and
the catalyst is regenerated by activating N2 and thus healing
the vacancy. This mechanistic hypothesis is supported by the
observation of a mixture of 14NH4+ and 15NH4+ with using 15N2
as the feed in ENRR (Figure 2D), which indicates that surface
N is involved in the activation of N2, as suggested by a few
computational studies.23−26 We hypothesize that only surface N
that are adjacent to a surface oxygen in the VN0.7O0.45 phase
are active in facilitating the catalytic turnover, as catalyst deꢀ
activation is accompanied by the accumulation of the VN
phase. The surface oxygen in the VN0.7O0.45 phase is unstable
at the reducing environment for ENRR, with more negative
potentials leading to faster removal of surface oxygen atoms
and generation of surface vacancies. Since the calculated bindꢀ
ing energy of N on a vacancy is much higher than that of O,
vacancies will be preferentially filled with N, leading to the
inactive VN phase (Figure 5, left circle). Thus, increasing the
density and stability of the surface oxygen of the VN0.7O0.45
phase under ENRR conditions, e.g., via introducing dopants,
could be effective strategies to improve the activity and duraꢀ
bility of the VN catalyst.
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
Based on the reactivity and characterization results, the
turnover number (TON) per active N site, i.e., N sites on the
accessible VN0.7O0.45 (57.8% of total), could be estimated. The
amount of VN0.7O0.45 in the catalyst is determined to be 3.5
µmol per MEA (with a total catalyst loading of 2.5 mg) based
on the elemental analysis. Together with the total amount of
produced ammonia via ENRR (253.1 µmol −0.1 V), an averꢀ
age TON of ~179 could be obtained at −0.1 V within 120 h.
Procedures for the TON calculation are detailed in the Supꢀ
plementary Information (Section 8). Because there is substanꢀ
tial catalyst deactivation within the first 4 h, the TON for acꢀ
tive sites at steady state (5−120 h) is estimated to be 431. It is
worth noting the total amount of nitrogen in the ammonia proꢀ
duced within 120 h is 9.1 times that in the VN catalyst loaded
on the MEA. Furthermore, the total amount of produced amꢀ
monia in this work is more than one order of magnitude higher
than those in previous reports (Table S6ꢀ2). Combined with
the result from the 15N2 experiment (Figure 2D), we concluꢀ
sively show that the produced ammonia originates from N2
activation, rather than N leaching from VN.
Figure 5. Proposed reaction pathway for nitrogen reduction on
the surface of VN0.7O0.45 via a Marsꢀvan Krevelen mechanism and
the catalyst deactivation mechanism.
CONCLUSIONS
To summarize, we demonstrate that VN nanoparticles are
active and selective ENRR catalysts with an ammonia producꢀ
tion rate and FE of 3.3 × 10−10 mol s−1 cm−2 and 6.0% at −0.1
V within 1 h, respectively. A steady state ammonia production
rate of 1.1 × 10−10 mol s−1 cm−2, with an FE of 1.6%, can be
maintained for 116 h. The ENRR with 15N2 as the feed proꢀ
duces a mixture of 14NH3 and 15NH3, which proves the inꢀ
volvement of surface N in the catalytic turnovers. This proꢀ
vides concrete evidence that the reaction proceeds via the
Marsꢀvan Krevelen mechanism. Combined ex-situ and oper-
ando characterization results indicate that the surface
VN0.7O0.45 species is the active phase in ENRR, and the conꢀ
version of this phase to VN is proposed as the deactivation
pathway. We hypothesize that only the surface N sites adjaꢀ
cent to a surface O are active in the ENRR, which leads to a
steady state turnover number of 431.
Figure 4. Operando XAS results of VN catalysts: Kꢀedge
XANES spectra of VN catalysts at different potentials (A) −0.1 V
and (C) −0.2 V as a function of time. (B) Left and Right panels
ACS Paragon Plus Environment

Page 5 of 6
Journal of the American Chemical Society
13. Nash, J.; Yang, X.; Anibal, J.; Wang, J.; Yan, Y.; Xu, B. J. Elec-
ASSOCIATED CONTENT
trochem. Soc. 2017, 164, F1712−F1716.
1
2
3
4
5
6
7
8
Supporting Information. Experiment methods, quantification of
ammonia, Nitrogen reduction in batch mode under 15N2, deconvoꢀ
lution of XPS spectra, operando XAS characterizations of the
catalysts, production rate and FE of other ENRR catalysts, comꢀ
putational methods, and procedure for turnover number calculaꢀ
tions are included in the supporting information. This material is
available free of charge via the Internet at http://pubs.acs.org.
14. Bao, D.; Zhang, Q.; Meng, F.ꢀL.; Zhong, H.ꢀX.; Shi, M.ꢀM.;
Zhang, Y.; Yan, J.ꢀM.; Jiang, Q.; Zhang, X.ꢀB. Adv. Mater.
2017, 29, 1604799.
15. Shi, M.ꢀM.; Bao, D.; Wulan, B.ꢀR.; Li, Y.ꢀH.; Zhang, Y.ꢀF.;
Yan, J.ꢀM.; Jiang, Q. Adv. Mater. 2017, 29, 1606550.
16. Rod, T. H.; Logadottir, A.; Nørskov, J. K. J. Chem. Phys. 2000,
112, 5343−5347.
17. Skúlason, E.; Bligaard, T.; Gudmundsdóttir, S.; Studt, F.;
Rossmeisl, J.; AbildꢀPedersen, F.; Vegge, T.; Jónsson, H.; Nørꢀ
skov, J. K. Phys. Chem. Chem. Phys. 2012, 14, 1235−1245.
18. Montoya, J. H.; Tsai, C.; Vojvodic, A.; Nørskov, J. K.
ChemSusChem 2015, 8, 2180−2186.
19. Wang, J.; Yu, L.; Hu, L.; Chen, G.; Xin, H.; Feng, X. Nat.
Commun. 2018, 9, 1795.
20. Lee, H. K.; Koh, C. S. L.; Lee, Y. H.; Liu, C.; Phang, I. Y.; Han,
X.; Tsung, C.ꢀK.; Ling, X. Y. Sci. Adv. 2018, 4, eaar3208.
21. Nazemi, M.; Panikkanvalappil, S. R.; ElꢀSayed, M. A. Nano
Energy 2018, 49, 316−323.
22. Yang, D.; Chen, T.; Wang, Z. J. Mater. Chem. A 2017, 5,
18967−18971.
23. Abghoui, Y.; Skúlasson, E. Procedia Comput. Sci. 2015, 51,
1897−1906.
24. Abghoui, Y.; Garden, A. L.; Howalt, J. G.; Vegge, T.; Skúlason,
E. ACS Catal. 2016, 6, 635−646.
25. Abghoui, Y.; Skúlason, E. Catal. Today 2017, 286, 78−84.
26. Abghoui, Y.; Skúlason, E. J. Phys. Chem. C 2017, 121,
6141−6151.
27. Zhang, X.; Kong, R.ꢀM.; Du, H.; Xia, L.; Qu, F. Chem. Com-
mun. 2018, 54, 5323−5325.
28. Giordano, C.; Erpen, C.; Yao, W.; Milke. B.; Antonietti, M.
Chem. Mater. 2009, 21, 5136−5144.
29. Zhao, D.; Cui, Z.; Wang, S.; Qin, J.; Cao, M. J. Mater. Chem. A
2016, 4, 7914−7923.
30. Silversmit, G.; Depla, D.; Poelman, H.; Marin, G. B.; De Gryse,
R. J. Electron. Spectrosc. Relat. Phenom. 2004, 135, 167−175.
31. Glaser, A.; Surnev, S.; Netzer, F. P.; Fateh, N.; Fontalvo, G. A.;
Mitterer, C. Surf. Sci. 2007, 601, 1153−1159.
32. Saha, N. C.; Tompkins, H. G. M. J. Appl. Phys. 1992, 72,
3072−3079
33. Wagner C. D.; Naumkin, A. V.; KrautꢀVass, A.; Allison, J. W.;
Powell, C. J.; Rumble, J. R. Jr. NIST Standard Reference Dataꢀ
base 20, Version 3.4 (web version) 2003.
AUTHOR INFORMATION
Corresponding Authors
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
* Address correspondence to jgchen@columbia.edu,
yanys@udel.edu, and bxu@udel.edu
Author Contributions
‡These authors contributed equally.
ACKNOWLEDGMENT
This work is supported by the US Department of Energy under the
grant number of DEꢀSC0016537. The DFT work was supported
by the US Department of Energy under the grant number of DEꢀ
FG02ꢀ13ER16381.
REFERENCES
1. Erisman, J. W.; Sutton, M. A.; Galloway, J.; Klimont, Z.; Winiꢀ
warter, W. Nat. Geosci. 2008, 1, 636−639.
2. Appl, M. Ullmann’s Encyclopedia of Industiral Chemistry,
WileyꢀVCH Verlag GmbH & Co. KGaA, 2000.
3. McEnaney, J. M.; Singh, A. R.; Schwalbe, J. A.; Kibsgaard, J.;
Lin, J. C.; Cargnello, M.; Jaramillo, T. F.; Nøskov, J. K. Energy
Environ. Sci. 2017, 10, 1621−1630.
4. Industrial
Efficiency
Technology
Database,
http://ietd.iipnetwork.org/content/ammonia, accessed September
2017.
5. Nøskov, J.; Chen, J. Sustainable Ammonia Synthesis, DOE
Roundtable Report, U.S. Department of Energy, 2016.
6. Schlögl, R. Ammonia synthesis, Handbook of Heterogeneous
Catalysis, WileyꢀVCH Verlag GmbH & Co. KGaA, 2008.
7. Chen, J. G.; Crooks, R. M.; Seefeldt, L. C.; Bren, K. L.; Bullock,
R. M.; Darensbourg, M. Y.; Holland, P. L.; Hoffman, B.; Janik,
M .J.; Jones, A. K.; Kanatzidis, M. G.; King, P.; Lancaster, K.
M.; Lymar, S. V.; Pfromm, P.; Schneider, W. F.; Schrock, R. R.
Science 2018, 360, eaar6611.
8. Cramp, W.; Hoyle, B. Journal of the Institution of Electrical
Engineers 1908, 42, 297−338.
9. Marnellos, G.; Stoukides, M. Science 1998, 282, 98−100.
10. Amar, I. A.; Lan, R.; Petit, C. T. G.; Tao, S. J. Solid State Elec-
trochem. 2011, 15, 1845−1860.
34. Gondek, K.; MierzwaꢀHersztek, M. Soil Use Manage. 2016, 32,
357−367.
35. Gondek, K.; MierzwaꢀHersztek, M.; Kopeć, M. J. Environ.
Manage. 2018, 210, 87−95.
36. Vanysek, P. Electrochemical Series, CRC Handbook of Chemisꢀ
try and Physics, CRC Press, Boca Raton, 2002.
37. Nakhal, S.; Hermes, W.; Ressler, T.; Pöttgen, R.; Lerch, M. Z.
Anorg. Allg. Chem. 2009, 635, 2016−2020.
11. Kordali, V.; Kyriacou, G.; Lambrou, C. Chem. Commun. 2000,
1673−1674.
12. Lan, R.; Irvine, J. T. S.; Tao, S. Sci. Rep. 2013, 3, 1145.
38. Henderson, G. S.; de Groot, F. M. F.; Moulton, B. J. A. Rev.
Mineral. Geochem. 2014, 78, 75−138.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 6
TOC
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
6
ACS Paragon Plus Environment