Contribution of Nitrogen Vacancies to Ammonia Synthesis over Metal Nitride Catalysts

Article abstract of DOI:10.1021/jacs.0c06624

Ammonia is one of the most important feedstocks for the production of fertilizer and as a potential energy carrier. Nitride compounds such as LaN have recently attracted considerable attention due to their nitrogen vacancy sites that can activate N2 for ammonia synthesis. Here, we propose a general rule for the design of nitride-based catalysts for ammonia synthesis, in which the nitrogen vacancy formation energy (ENV) dominates the catalytic performance. The relatively low ENV (ca. 1.3 eV) of CeN means it can serve as an efficient and stable catalyst upon Ni loading. The catalytic activity of Ni/CeN reached 6.5 mmol·g-1·h-1 with an effluent NH3 concentration (ENH3) of 0.45 vol %, reaching the thermodynamic equilibrium (ENH3 = 0.45 vol %) at 400 °C and 0.1 MPa, thereby circumventing the bottleneck for N2 activation on Ni metal with an extremely weak nitrogen binding energy. The activity far exceeds those for other Co- and Ni-based catalysts, and is even comparable to those for Ru-based catalysts. It was determined that CeN itself can produce ammonia without Ni-loading at almost the same activation energy. Kinetic analysis and isotope experiments combined with density functional theory (DFT) calculations indicate that the nitrogen vacancies in CeN can activate both N2 and H2 during the reaction, which accounts for the much higher catalytic performance than other reported nonloaded catalysts for ammonia synthesis.

Full text of DOI:10.1021/jacs.0c06624

pubs.acs.org/JACS
Article
Contribution of Nitrogen Vacancies to Ammonia Synthesis over
Metal Nitride Catalysts
Tian-Nan Ye,# Sang-Won Park,# Yangfan Lu,# Jiang Li, Masato Sasase, Masaaki Kitano,*
and Hideo Hosono*
Cite This: https://dx.doi.org/10.1021/jacs.0c06624
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Ammonia is one of the most important feedstocks for
the production of fertilizer and as a potential energy carrier. Nitride
compounds such as LaN have recently attracted considerable
attention due to their nitrogen vacancy sites that can activate N2
for ammonia synthesis. Here, we propose a general rule for the design
of nitride-based catalysts for ammonia synthesis, in which the
nitrogen vacancy formation energy (ENV) dominates the catalytic
performance. The relatively low ENV (ca. 1.3 eV) of CeN means it can
serve as an efficient and stable catalyst upon Ni loading. The catalytic
activity of Ni/CeN reached 6.5 mmol·g−1·h−1 with an effluent NH3
concentration (ENH3) of 0.45 vol %, reaching the thermodynamic
equilibrium (ENH3 = 0.45 vol %) at 400 °C and 0.1 MPa, thereby
circumventing the bottleneck for N2 activation on Ni metal with an extremely weak nitrogen binding energy. The activity far exceeds
those for other Co- and Ni-based catalysts, and is even comparable to those for Ru-based catalysts. It was determined that CeN itself
can produce ammonia without Ni-loading at almost the same activation energy. Kinetic analysis and isotope experiments combined
with density functional theory (DFT) calculations indicate that the nitrogen vacancies in CeN can activate both N2 and H2 during
the reaction, which accounts for the much higher catalytic performance than other reported nonloaded catalysts for ammonia
synthesis.
CO over metal oxides such as V2O5 and CeO2, which are
interpreted within the Mars−van Krevelen mechanism.14−17
This mechanism was further extended to ammonia synthesis
recently using oxyhydride and oxynitride based materials, in
which incorporated N3− and/or H− ions play a vital role for
ammonia production.18−20 For example, the transition metal
(TM)-loaded 12CaO·7Al2O3 electride and Ca(NH2)2 exhibit
excellent catalytic activity for ammonia synthesis by the
presence of anionic electrons and incorporated H− ions.21−26
TM-loaded metal hydrides (AHx, A = Li, Na, K, Ba, Ca) can
extract dissociated N from TMs to form metal-imide species,
which then undergo hydrogenation to the final NH3
product.27−29 Although high catalytic performance has been
achieved, these efficient support materials are still dependent
on the loading of active metals. Some perovskite materials,
such as BaCeO3−xNyHz and BaTiO2.5H0.5, also generate
ammonia, even without TM loading.30−33 For each case, the
INTRODUCTION
■
Ammonia is important for the production of chemicals and
fertilizer worldwide.1 N2 activation to form ammonia (NH3) is
the general route for ammonia production, which requires the
large NN bond energy (945 kJ·mol−1) to be overcome.2,3
The major milestone for industrial ammonia production was
the development of the Haber−Bosch process. It was
demonstrated that Fe-based catalysts could generate NH3
from N2 and H2 molecules under high temperature (400−
600 °C) and high pressure (20−40 MPa) conditions.4,5
Although electrochemical and photocatalytic approaches have
been attempted under milder conditions,6−10 the efficiency and
production scale have been far from industrial requirements,
which has led to the preference of the traditional thermal
catalytic process for large scale NH3 production. There is thus
a high demand to develop more efficient substitutes as
industrial catalysts, and as such, further fundamental scientific
research has been undertaken.
Received: June 19, 2020
The surface properties of heterogeneous catalysts are critical
for catalytic performance, and the introduction of surface
defect sites has emerged as an effective strategy to achieve high
activity for various chemical reactions.11−13 For instance,
oxygen vacancies are considered to participate in chemical
reactions such as the oxidation of aromatic hydrocarbons and
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. Catalytic activity of Ni-loaded different nitrides. (a) Catalytic activity (blue, bulk; red, nanoparticles (NPs)) for ammonia synthesis over
various nitride-supported Ni catalysts. (b) Arrhenius plots for ammonia synthesis over Ni/CeN NPs, Ni/LaN NPs, Ni/CeN bulk, and Ni/LaN
bulk catalysts at 0.1 MPa. (c) Temperature dependence of ammonia synthesis activity over Ni/CeN NPs and Ni/CeN bulk at 0.1 and 0.9 MPa.
The inset of (c) shows an Arrhenius plot for ammonia synthesis over Ni/CeN NPs at 0.9 MPa. (d) Time course of ammonia synthesis over the Ni/
CeN NP catalyst at 0.1 and 0.9 MPa. (Reaction conditions: 0.1 g catalyst, N2:H2 = 15:45 mL min−1, 150−400 °C, 0.1 and 0.9 MPa).
formation of nitride-hydride surface species significantly
promotes the catalytic performance of TMs. Accordingly, a
support material with high nitrogen affinity is considered to
enhance the catalytic performance of the TMs.
ammonia synthesis over nitride compounds, with and without
Ni loading.
RESULTS AND DISCUSSION
■
In this context, we recently demonstrated that a (LaN-Ni)
catalyst with dual active sites could separately activate H2 and
N2 molecules, and successfully overcome the scaling relation-
ship that has restricted the use of Ni-based catalysts for
ammonia synthesis due to the extremely weak nitrogen binding
energy of Ni.34 One of the highlighted points of LaN is its low
nitrogen vacancy formation energy (ENV). It was demonstrated
that the lattice N of LaN directly reacts with dissociated
hydrogen (H*) from the Ni surface to generate NH3 and N
vacancies. N2 molecules are subsequently adsorbed and
activated at the N vacancy sites, and continuously react with
H* to realize a stable catalytic cycle. The role of nitride
supports and their N vacancies has therefore inspired us to
consider ENV as an important factor, and a general rule for the
design of Ni-based catalysts for ammonia synthesis.
In the present study, the correlation of ENV over various
nitride supported Ni catalysts for ammonia synthesis was
clarified first. Among these tested nitrides, CeN showed the
highest activity with Ni loading, which is characterized by a
lower ENV and a suppressed activation barrier toward ammonia
synthesis. CeN itself also serves as a stable catalyst for
ammonia synthesis, even without TMs-loading. Kinetic
analysis and isotope experiments indicate that the N vacancies
in CeN contribute to the activation of both N2 and H2
molecules. The experimental data obtained are consistent
with the theoretical calculations, which indicate that the ease of
N vacancy formation determines the catalytic activity for
To test the effect of the nitride support on the catalytic
performance of Ni, rare-earth metals (Re) based nitrides ReN
(Re = Ce, La, Y, and Sc) and early transition metals (Et) based
nitrides EtN (Et = Zr, Ti, Ta, and Nb) were employed as
supports with the same amount of Ni loading (ca. 5 wt % for
bulk, ca. 10 wt % for nanoparticles (NPs)) (Figures S1−S3).
Figure 1a compares the activities of Ni/ReN and Ni/EtN
catalysts at 400 °C and at a reactant gas (H2:N2 = 3:1)
pressure of 0.1 MPa. Ni/ReN effectively produced ammonia,
while the activities of Ni/EtN were below the detection limit.
In particular, Ni/CeN exhibited the highest activities, which
were as high as 3100 and 6500 μmol·g−1·h−1 for the bulk and
NPs, respectively. In the turnover frequency (TOF) calcu-
lation, the concentration of surface nitrogen vacancies (VN)
used was 3.9 × 1014 cm−2 from the amount of top layer lattice
N of CeN, which are the active sites of the Ni/CeN catalyst
(details for the active sites are discussed later). The TOF of
Ni/CeN was estimated to be 0.064 s−1. The catalytic activities
and TOF value far exceed those of the other reported Ni- and
Co-based catalysts, and are even comparable to those of noble
metal catalysts, such as the Ru catalysts (Tables S1,
S2).24,28,35−40 The activities of the Ni/ReN catalysts are
ranked in the order of Ni/CeN > Ni/LaN > Ni/YN > Ni/ScN
for both the bulk and NP samples (as shown by the red and
blue bars in Figure 1a). Considering the similar specific surface
areas and Ni particle sizes of these Ni/ReN samples (Figure
S4), this order for the capability to promote the reaction is
B
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 2. Characterization of Ni/CeN. (a) HAADF-STEM image of CeN regions in N-deficient Ni/CeNv along the [111] direction. The inset of
(a) shows the corresponding crystal structure of CeN along the [111] direction. Ce and N atoms are represented as yellow and blue balls,
respectively. (b−d) AES spectra of fresh and N-deficient Ni/CeNv. (e) HAADF-STEM image and (f−h) EDX mapping results for N, Ce, and Ni in
the used Ni/CeN catalyst.
considered to originate from the role of the nitride support
rather than that of the Ni sites. The apparent activation
energies (Ea) of the Ni/CeN and Ni/LaN catalysts are close to
ca. 60 kJ·mol−1 in the temperature range of 280−380 °C at 0.1
MPa (Figure 1b). The Ni/YN bulk catalyst shows a higher Ea
(ca. 90 kJ·mol−1) in the relative high temperature range of
400−500 °C at 0.1 MPa (Figure S5), which suggests that the
reaction mechanism over the Ni/CeN and Ni/LaN catalysts is
similar, i.e., H2 and N2 molecules are separately activated on
the loaded Ni and surface VN sites of the ReN support, which
was confirmed through an isotope experiment using 15N2 and
H2 and the isotopic exchange reaction between 14N2 and 15N2
(Figure S6, S7).34
As the most active sample, the generation of surface VN on
the CeN support was further confirmed by high-angle annular
dark-field scanning transmission electron microscopy
(HAADF-STEM) along the [111] direction. Local distortion
could be identified after the treatment of Ni/CeN in a H2
atmosphere at 400 °C, which suggests the dislocation of Ce
atoms due to the formation of a substantial amount of N
defects (Figure 2a). To further confirm the generation of N
defects, Auger electron spectroscopy (AES) was performed to
investigate the surface composition. The depth dependence of
the N peak (ca. 387 eV) from the AES spectra show that N
defects are more likely to be generated on the CeN surface
(Figures 2b and S8). Figure 2b shows that fresh CeN has a
stronger N peak than that after H2 treatment, while the Ce and
Ni peaks remains unchanged (Figure 2c,d). Consideration of
both the results of local distortion and reduced N AES peaks
suggests that a substantial amount of N defects were formed on
the Ni/CeN surface by H2 treatment. Energy-dispersive X-ray
spectroscopy (EDX) mapping of the Ni/CeN sample collected
after reaction (Figure 2e−h) revealed that Ce and N are
uniformly dispersed on CeN. The mean Ni particle size was
estimated to be ca. 30 nm, which is consistent with scanning
electron microscopy (SEM) observations (Figure S4a).
To further understand the degree of the metal−support
interaction, we performed the PDOS calculations of Ni8 cluster
on nitrogen vacancy-free CeN surface (Ni8/Ce48N48), Ni8
cluster on VN existed CeN surface (Ni8/Ce48N47) and pure
VN existed CeN surface (Ce48N47). As shown in Figure S9, in
both Ni loaded cases, the density of states near Fermi level was
mainly contributed by the Ni 3d orbitals. Compared with the
PDOS of Ni8/Ce48N48, Ni8/Ce48N47 gave a significant shift due
to the electron doping, associated with the generation of N
vacancy. Therefore, the electronic energy region of the Fermi
level shifting is dominated by Ni 3d orbitals, and thus doped
electrons should be transferred from the VN to the Ni sites.
While CeN support and Ni metal do not form ternary
compounds, Ni8 cluster can be strongly interacted with CeN
through such electron transfer during the catalytic cycles since
each of them was electronically connected. The orbital
hybridization between Ni8 cluster and N 2p orbitals is also
an important interaction between Ni and CeN support. As
shown in Figure S10, Ni[1] 3d orbitals hybridize with the
nearest N[2] 2p orbitals, and exhibit a substantial overlapping
of PDOS at around −4 < E − EF < −2 eV. Such contribution
cannot be seen for the second nearest N[1] at the
corresponding energy range. This effect stabilizes the loaded
Ni8 cluster as well, and the stabilization energy was calculated
to be −8.8 eV per Ni8 cluster.
The reaction orders of the Ni/CeN catalyst were estimated
to be N2 (α) = 1.2, H2 (β) = 1.6, and NH3 (γ) = −1.4, which
would result in a desirable high-pressure effect (Table S3). The
reaction rates increase monotonically with the reactant gas
pressure and no hydrogen poisoning was observed (Figure 1c
and S11). It should be noted that Ea for Ni/CeN NPs was
measured to be 54.6 kJ·mol−1 under 0.9 MPa (Figure 1c inset),
which is close to that under 0.1 MPa (Figure 1b) and indicates
the reaction mechanism remains unchanged under high-
pressure conditions. The catalyst produced ammonia con-
tinuously without degradation for 80 h at 340 °C and at both
C
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. Comparison of ammonia production, H2-TPR results and DFT studies over various nitride-supported Ni catalysts. (a) Ammonia
synthesis activity and N vacancy formation energies (ENV) over various nitride-supported Ni catalysts. (b) Cumulative amounts of NH3 generated
over various nitride bulk-supported Ni catalysts as a function of time using pure H2 gas supply at 400 °C. (c) H2-TPR profiles for various nitride-
supported Ni catalysts. (d) Energy profiles from DFT studies of the reaction path for ammonia synthesis with various nitride-supported Ni
catalysts.
0.1 and 0.9 MPa, which confirmed excellent stability (Figure
1d).
reaction.34 Here, the reaction barrier for Nlattice hydrogenation
over Ni/ReN (Re = Ce, La, and Y) catalysts was also
calculated. The energy profiles of the intermediate structures
and transition states (TSs) as the key elementary steps are
shown in Figure 3d and Table S4−S6. The TS energy of the
Nlattice hydrogenation step (TS1) is the highest among all the
reaction steps for the Ni/LaN and Ni/CeN catalysts.
Therefore, the overall activation barrier for NH3 formation
on Ni/CeN is ca. 0.37 eV, which is slightly lower than that on
Ni/LaN (ca. 0.46 eV), and indicates that Nlattice in CeN is
more easily hydrogenated by H* than that in LaN, which is in
good agreement with the distinction of the experimentally
obtained Ea (Figure 1b). However, in the case of Ni/YN, the
overall activation barrier comes from the hydrogenation of
NH2 species (TS3) with an exceptionally large barrier of ca.
1.05 eV. This could be ascribed to the relatively large ENV of
Ni/YN (Figure 3a), which makes it difficult to remove the
Nlattice from the YN lattice and is consistent with the
experimentally obtained Ea of ca. 0.95 eV (ca. 91.5 kJ·mol−1)
with Ni/YN (Figure S5). Moreover, from the final the energy
state in step VII (Figure 3d), VN on Ni/CeN are much more
stable than those on Ni/LaN and Ni/YN, which is also in good
agreement with the distinction in the Ea (Figure 1b) and ENV
(Figure 3a). We also calculated the N2 adsorption energy
(EN2) into the VN of these Ni-loaded nitrides to be Ni/CeN
To determine the origin of the different catalytic perform-
ance among the Ni/ReN catalysts with similar surface area and
particle size (Figure S4), DFT calculations were conducted
and nitrogen vacancy formation energies (ENV) were calculated
to be Ni/CeN (1.29 eV) < Ni/LaN (1.52 eV) < Ni/YN (2.43
eV) < Ni/ScN (2.69 eV), which indicates that Nlattice in Ni/
CeN are more easily removed than Nlattice in Ni/LaN, Ni/YN,
and Ni/ScN (Figure 3a). Accordingly, the measurements of
Nlattice removal from these Ni-loaded nitride catalysts reveal a
higher NH3 production rate over Ni/CeN, with the lowest ENV
among these Ni/ReN catalysts (Figure 3b). Furthermore, H2-
temperature-programmed reduction (H2-TPR) data showed
finite NH3 desorption from Ni/CeN and Ni/LaN (Figure 3c).
The NH3 desorption temperature of CeN (220 °C) is slightly
lower than that of LaN (250 °C), which is consistent with the
results of the pure H2 supply experiment (Figure 3b). In
contrast, negligible NH3 was detected over Ni/YN and Ni/
ScN (Figure 3c). These experimental results indicate that the
removal of Nlattice from Ni/CeN by H2 proceeds more easily
than from Ni/LaN, Ni/YN, and Ni/ScN, which is consistent
with the different ENV values (Figure 3a). As we reported for
Ni/LaN, the combined process of H* and Nlattice in LaN is the
rate-determining step (RDS) for the ammonia synthesis
D
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. Catalytic performance for ammonia synthesis over various nonloaded nitride catalysts. (a) Catalytic activity (blue, 0.1 MPa; red, 0.9
MPa) for ammonia synthesis over pure CeN, LaN, and YN NPs. (b) Arrhenius plots for ammonia synthesis over pure CeN NPs at 0.1 and 0.9
MPa. (c) Pressure dependence of ammonia synthesis activity over pure CeN NPs. (d) Time course for ammonia synthesis over pure CeN NPs at
0.1 and 0.9 MPa. (Reaction conditions: 0.1 g catalyst, N2:H2 = 15:45 mL min−1, 300−400 °C, 0.1−0.9 MPa).
(EN2 = −1.21 eV), Ni/LaN (EN2 = −1.03 eV), and Ni/YN
(EN2 = −1.56 eV). Accordingly, the NN bond lengths were
increased to 1.25 Å for Ni/CeN, 1.26 Å, for Ni/LaN and 1.32
Å for Ni/YN, which suggests that the removal of the top N
(end-on adsorbed N2) proceeds more easily from Ni/YN than
from Ni/CeN and Ni/LaN. However, the reaction rate over
Ni/YN is much lower than that over Ni/CeN and Ni/LaN,
which indicates that down N (equal to Nlattice) hydrogenation
controls the reaction rate; this is strongly related to the ENV we
have proposed (Figure 3a). Therefore, the calculation results
presented here indicate that the RDS for ammonia synthesis
over the Ni/ReN catalysts is not N2 dissociation, but one of the
Nlattice hydrogenation steps. The reaction barrier for Nlattice
hydrogenation (RDS) over Ni/CeN is obviously lower than
those over Ni/LaN and Ni/YN because of the relatively low
ENV (ca. 1.3 eV), which results in the significantly higher
catalytic activity of Ni/CeN (Figure 1a). Figure S12 shows the
comparison of NH3 desorption from Ni/CeN bulk and NPs
during H2-TPR. The NH3 desorption reflects the reaction of
hydrogen with lattice nitrogen on Ni/CeN catalysts and
simultaneous VN formation. The integrated peak area of the
Ni/CeN NPs for NH3 desorption is almost twice that of Ni/
CeN bulk sample. Thus, the number of active sites on Ni/CeN
NPs is almost double that of Ni/CeN bulk catalyst, which is
consistent with the difference in the catalytic activities of these
two catalysts (Figure 1a). This result further supports our
proposed mechanism.
formation energy compared with that for cation vacancies,41
which makes it almost impossible to remove Nlattice from these
compounds, and this was also confirmed by the absence of
NH3 in the H2-TPR results for the Ni/EtN catalysts with
similar specific surface areas and Ni particle sizes (Figure S13,
S14). All of these computational and experimental results
indicate that the catalytic activities of Ni/ReN (Re = Ce, La,
and Y) are closely related to ENV, which demonstrates the
importance of VN in various nitrides for N2 activation and
hydrogenation toward Ni loaded catalysts for ammonia
synthesis.
Overall, the Ni/CeN catalyst was demonstrated to be the
most efficient Ni-based catalyst for ammonia synthesis
reported to date. During the reaction, dissociated H* from
the Ni surface can directly react with Nlattice of CeN to generate
NH3 and nitrogen vacancies. N2 molecules are subsequently
activated at the defect sites and again react with H* to realize a
stable catalytic cycle. Here, we consider whether nitrogen
defects can also activate H2 molecules, i.e., whether CeN itself
can also act as a stable catalyst for ammonia synthesis. Given
the much higher surface areas of the NP samples among pure
CeN, LaN, and YN, the catalytic performance for ammonia
synthesis over pure NP catalysts is shown in Figure 4a. Pure
CeN NPs exhibited catalytic activities of 250 and 1450 μmol·
g−1·h−1 at 0.1 and 0.9 MPa (400 °C, WHSV: 36 000 mL·g−1·
h−1), respectively, which are 2 times higher than those with
pure LaN and YN NPs. To the best of our knowledge, the
catalytic activity of pure CeN NPs is superior to those of other
reported non-TM-loaded catalysts, such as BaCeO3−xNyHz
(430 μmol·g−1·h−1, 400 °C, 0.9 MPa, WHSV: 36 000 mL·
g−1·h−1) and BaTiO2.5H0.5 (700 μmol·g−1·h−1, 400 °C, 5 MPa,
In the cases of Ni/EtN, the ENV values for Ni/ZrN and Ni/
NbN were calculated to be 3.36 and 1.84 eV, respectively.
Here, we consider that the negligible catalytic performance
could be ascribed to the reported larger nitrogen vacancy
E
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 5. Characterization and DFT calculations of various nitrides. (a) Reaction time profiles for NH3 synthesis from 15N2 and H2 over fresh CeN.
(b) Reaction time profiles for the ratio changes of 17/18 and 16/18 according to (a), where in mass spectrometry, NH2/NH3 = 0.8 and NH/NH3
= 0.075. (c) AES spectra N signals for fresh and used Ni/CeN, and pure CeN catalysts. (d) N vacancy formation energies (ENV) and nitride
formation energies (ENF) for pure CeN, LaN and YN catalysts. (e) TPD profiles for ammonia synthesis over used CeN, LaN, and YN NPs catalysts.
(f) Cumulative amounts of NH3 generated over pure CeN, LaN, and YN NP catalysts as a function of time using pure H2 gas supply at 400 °C.
WHSV: 54 000 mL·g−1·h−1), as shown in Figure S15.30,32 The
Ea for the pure CeN NPs were estimated to be 56.5 and 57.2
kJ·mol−1 at 0.1 and 0.9 MPa (Figure 4b), which were
comparable to those for the Ni/CeN catalyst (Figure 1b),
but smaller than those for other reported non-TM-loaded
catalysts (70−80 kJ·mol−1) (Figure S15). The reaction orders
with respect to N2, H2, and NH3 were measured to be 1.0, 1.3,
and −1.6 (Figure S16 and Table S3), similar to those of Ni/
CeN catalyst, indicating a similar reaction mechanism which
will be discussed later. A positive pressure effect, almost 6-fold
enhancement of the reaction rate, was achieved when the
pressure was increased from 0.1 to 0.9 MPa (Figure 4c). The
crystal structure and the initial ammonia synthesis rates were
maintained after 120 h under both ambient and high pressure
conditions (Figure 4d and S17), and the total amount of
generated ammonia reached 17.4 mmol, which was more than
26 times the total nitrogen content in the CeN catalyst (0.65
mmol). This result indicates that the produced ammonia is not
derived from the decomposition of CeN, but from the catalytic
cycle.
To further clarify the mechanism, the isotopic effect was
measured using 15N2 and H2 over the pure CeN NP catalyst.
Figure 5a shows the changes in intensity for mass numbers as a
function of time. Ammonia isotopomer-related species with the
mass signals of m/z = 16 (15NH or 14NH2), 17 (15NH2 or
14NH3), 18 (15NH3), and 30 (15N2) were all detected. After the
introduction of isotope gas 15N2 and H2, the signals measured
for the intensity ratios of m/z 17/18 and 16/18 were larger
than the theoretical values of 0.8 (m/z 17/18) and 0.075 (m/z
16/18), which indicated the generation of both 14NH3 and
15NH3 (Figure 5b). These results clearly indicate that the
lattice 14N of CeN participate in the formation of ammonia,
which is consistent with the case for Ni/CeN (Figure S6).
These results raise the following question. How can H2
molecules be activated and dissociated to form reactive
hydrogen species? To address this point, the change of the
surface composition was estimated from AES measurements
(Figures 5c and S18). The variation in the surface composition
of Ni/CeN shows that the N peak (ca. 387 eV) remains largely
unchanged after reaction, which indicates that the in situ
generated surface N vacancies are readily replenished by gas-
phase N2 molecules during ammonia synthesis. Figure S19
shows the calculated adsorption energies of H2 on the Ni
surface and N vacancies are −1.30 and −0.62 eV, respectively,
which indicates that H2 is more likely to be adsorbed on the Ni
surface than on the N vacancies of CeN. Given the extreme
weak nitrogen adsorption energy of Ni, the N2 molecules tend
to occupy the surface VN sites with an adsorption energy of
−1.21 eV (Figure S19c), which is consistent with the
unchanged surface N content indicated by the AES spectra
(Figure 5c). In contrast, the surface N content of pure CeN
decreased significantly after reaction (Figure 5c), which is
ascribed to the competition between H2 and N2 molecules to
occupy the surface VN sites. Accordingly, ca. 0.14 hydrogen
atoms per formula unit are desorbed from the used CeN
catalyst above ca. 250 °C (Figure S20). Also, no amide
(NH2 ) and imide (NH2−) species could be confirmed in
−
Raman spectra the used CeN catalyst (Figure S21). In
combination with the AES surface composition analysis and
the large amount of H desorption from the TPD measure-
ments, it is reasonable to deduce that the surface VN sites of
pure CeN are mainly occupied by hydrogen species. However,
unlike other Ce-based oxide catalysts, H2 molecules dissociate
homolytically and form H− ions at the VN sites of CeN, which
is evidenced by Bader charge analysis (−0.7 for each
dissociated H atom).30 This large amount of H− ions that
occupy VN sites then participate in the formation of ammonia.
F
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
for 2 h. The corresponding CeN and LaN bulk samples were then
synthesized by heating CeH3 and LaH3 powders at 600 °C under a N2
gas flow for 6 h. YH3 and ScH2 bulk powders were prepared by
reacting Y and Sc chips (99.99%, Kojundo Chemical Lab. Co.) in a
1.5 MPa H2 atmosphere at 400 °C for 4 h. YN and ScN bulk samples
were synthesized by heating bulk YH3 and ScH2 at 800 and 1000 °C,
respectively, under a N2 flow gas for 24 h. CeH3, LaH3, and YH3 NPs
were obtained by an Ar/H2 arc evaporation method.42 CeN and LaN
NPs were prepared by heating CeH3 and LaH3 NPs at 400 °C under a
N2 gas flow for 48 h. YN NPs were synthesized by the heat treatment
of YH3 NPs at 700 °C under a N2 gas flow for 24 h. Other nitride
materials, such as ZrN, TiN, TaN, and NbN, were obtained as
commercially available products. For Ni loading, nickelocene
(Ni(C5H5)2) was used as a Ni precursor, which was mixed with
various supports in an agate mortar within an Ar-filled glovebox. The
mixture was then heated at 250 °C under a H2 gas flow for 1.5 h. The
nitrides are moisture sensitive; therefore, all of the catalyst preparation
procedures were performed in an Ar-filled glovebox.
In the Ni/ReN catalyst, ENV determines the catalytic
performance for ammonia synthesis upon Ni. The comparable
Ea (ca. 60 kJ·mol−1) of pure CeN (Figure 4b) indicates a
similar reaction mechanism is valid for Ni/CeN. Here, the
hydrogenation process of Nlattice in pure CeN was also
investigated using DFT calculations (Figure S22 and Table
S7). The H2 molecules can be activated and dissociated on the
VN sites of CeN because of the negligible energy barriers (I →
II, IV → V). The surface Nlattice reacts step-by-step with H*
formed from the nearby VN sites. After hydrogenation of the
surface Nlattice by two H*, NH2 is ejected from the lattice by a
dissociated H* by an energy-favored state (step VI). The final
hydrogenation step of NH2 with the dissociated H* creates an
overall energy barrier of 0.64 eV for the entire reaction (VI →
VII). The calculated reaction paths over pure CeN are slightly
different from those over the Ni/CeN catalyst because the
overall energy barrier originates from the different hydro-
genation steps (Figure 3d). The relatively lower catalytic
activity of pure CeN than the Ni-loaded catalyst may be
ascribed to the large number of H2 molecules that occupy the
surface VN sites, which inhibits the adsorption and activation of
N2 molecules (Figure 5c). Here, ENV was also calculated for
pure CeN, LaN, and YN in the following order: CeN (1.39
eV) < LaN (1.72 eV) < YN (2.65 eV), which indicates that
Nlattice is more easily removed from CeN than from LaN and
YN (Figure 5d).
The order determined for the nitride formation energy
(ENF) was YN (−3.48 eV) < LaN (−2.65 eV) < CeN (−2.44
eV), which indicates the thermodynamic stability of YN is
much higher than that for LaN and CeN, and that the most
active Nlattice is that in CeN, which is consistent with the
calculated ENV. Accordingly, N2 desorption from the used CeN
catalyst occurs in a lower temperature range than that from
LaN and YN (Figure 5e). To further investigate this
relationship with ammonia generation, NH3 production was
investigated using each catalyst under a pure H2 gas supply
(Figure 5f).The NH3 production over CeN was higher than
that over LaN and YN in the following order: YN < LaN <
CeN. Therefore, the ease of N vacancy formation also
determines the ammonia synthesis reaction over pure nitride
compounds without Ni.
Catalytic Reaction. Catalytic reactions were conducted in a fixed-
bed flow system with a flow of H2 (>99.99999%) and N2
(>99.99995%) with a H2/N2 flow ratio of 3. The concentration of
ammonia in the effluent stream from the catalyst bed (0.1 g) was
tested under steady-state conditions at a gas flow rate of 60 mL·min−1
in the temperature range of 150−400 °C and under pressure in the
range of 0.1−0.9 MPa. The reaction temperature errors were
minimized by directly placing a thermocouple into the catalyst in
the fixed-bed flow reactor. The ammonia produced was dissolved in 5
+
mM sulfuric acid solution and the amount of NH4 ions in the
solution was identified using ion chromatography (Prominence,
Shimadzu) with an electrical conductivity detector. All kinetic results
were performed far from equilibrium limitations. The reaction orders
with respect to N2 and H2 were determined with Ar gas as a diluent to
ensure a total flow of 60 mL·min−1 when changing the flow rate of N2
and H2 at 340 °C and 0.1 MPa.
The ammonia synthesis reaction can be described as shown in eq 1:
1/2N2 + 3/2H2 → NH3
(1)
Then, the effluent NH3 concentration can be calculated from the
reaction rate r0 (eqs 2−4):
QNH3 = r0·Vm·mcat
(2)
Q total = QN2 − 1/2QNH3 + QH2 − 3/2QNH3 + QNH3
=QN2 + QH2 − QNH3
(3)
ENH3 = QNH3/Q total
(4)
CONCLUSION
■
where r0 is the reaction rate, Vm is the molar volume, mcat is the weight
of catalyst, QN2, QH2, and QNH3 are the volume flow rates of the feed
and product gas, respectively, and ENH3 is the effluent NH3
concentration. Here, thermodynamic limits of effluent NH3
concentration at 400 °C and under 0.1 MPa is 0.45 vol %.
For calculation of the TOFs, surface VN sites of Ni/CeN were
regarded as the active sites for the activation and adsorption of N2
molecules, and then for further hydrogenation to produce NH3, which
could be determined from the concentration of the top layer Nlattice of
CeN. The amount of top layer Nlattice was estimated from the unit cell
of pure CeN to be 3.9 × 1014 cm−2. The TOFs were then calculated
from the reaction rate r0, derived from the number of surface N atoms
(eq 5):
The present work demonstrates that rare earth metal nitrides
such as CeN, LaN, and YN can act as efficient supports and/or
catalysts for ammonia synthesis, in which nitrogen vacancy
formation has a strong effect on the catalytic performance. In
the case of Ni-loaded nitrides, H2 and N2 are separately
activated at Ni metal and VN sites, respectively. Without Ni
loading, VN sites can activate both H2 and N2 simultaneously
for stable ammonia production. Due to the relative low ENV
(ca. 1.3 eV), CeN itself and with Ni loading were
demonstrated to be the most efficient and stable catalyst for
ammonia synthesis among the different nitride catalysts
investigated. Understanding the role of nitrides may shed
light on the design and development of efficient early transition
metal based heterogeneous catalysts for other chemical
processes.
TOFs = r0·N /Nsurface N·SBET
(5)
A
where r0 is the reaction rate, NA is Avogadro’s constant, Nsurface N is the
amount of top layer Nlattice of CeN, and SBET is the specific surface
area of the catalyst. The surface of CeN was loaded with Ni NPs,
which partially hindered the exposure of surface N atoms. Therefore,
the TOFs obtained from this calculation method were under-
estimated.
EXPERIMENTAL SECTION
■
Sample Preparation. CeH3 and LaH3 bulk samples were
prepared by the treatment of Ce and La chips (99.99%, Kojundo
Chemical Lab. Co.) in a 1.5 MPa H2 atmosphere at room temperature
Sample Characterization. The crystal structure was analyzed
using X-ray diffraction (XRD; D8 Advance, Bruker) with Cu Kα
G
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
radiation (λ = 0.15418 nm). The samples were placed in Ar-filled
capsules to protect them from air exposure. The morphology of the
samples was evaluated using field-emission scanning electron
microscopy (FE-SEM; JSM-7600F, JEOL) and the component
elements were analyzed using energy dispersive X-ray spectroscopy
(EDX; JED-2300, JEOL). Transmission electron microscopy (TEM)
and high-angle annular dark-field scanning transmission electron
microscopy (HAADF−STEM) images were obtained using an atomic
resolution analytical electron microscope (JEM-ARM200F, JEOL)
operated at an accelerating voltage of 200 kV. Raman spectra of the
samples were measured with a spectrometer (HR-800, Horiba Jobin
Yvon), using a laser with a wavelength of 457.4 nm. Auger electron
spectroscopy (AES) was performed with 10 keV primary electrons
using a scanning Auger nanoprobe system (PHI 710, Ulvac-Phi).
Nitrogen sorption measurements (BELSORP-mini II, BEL) were
applied to evaluate the Brunauer−Emmett−Teller (BET) surface
areas of the catalysts. Temperature-programmed reduction with H2
(H2-TPR; BELCAT-A, BEL) was also conducted. Prior to measure-
ments, the catalysts (0.1 g) were introduced into a quartz glass cell in
an Ar-filled glovebox. The glass cell was then heated (10 °C·min−1) in
a stream of 4.8% H2/Ar, and the H2 consumption and gas products
were monitored with a mass spectrometer (Bell Mass, BEL).
Temperature-programmed desorption (TPD) of hydrogen (BEL-
CAT-A, BEL) was performed using the same instrument for H2-TPR.
Prior to measurements, the catalysts (0.1 g) were introduced into a
quartz glass cell in an Ar-filled glovebox and the glass cell was heated
(10 °C·min−1) in an Ar stream (50 mL·min−1), and the concentration
of H2 was monitored with a thermal conductivity detector (TCD) and
mass spectrometer (Bell Mass, BEL).
they were most favored for exposure. The ReN(001) surfaces were
modeled by an 11 layer slab with a 4 × 4 lateral unit cell. A 20 Å thick
vacuum region was set to prevent interaction between the slabs. The
central 3 layers of atoms of the surface models were kept fixed to hold
the characteristics of realistic surfaces, while the remainder of the unit
cell was allowed to be fully relaxed during the geometry optimizations.
A cutoff energy of 500 eV and a Monkhorst−Pack K-mesh setting of 3
× 3 × 1 were employed in the calculation for the ReN(001) surfaces.
The vacuum level was confirmed as the energy level at which the
electrostatic potential became constant. The atoms in the 4 bottom
layers were then removed, and the atoms in the 4 top layers were
relaxed for further VN formation and TS calculations. The model of
the Ni-loaded ReN surface was constructed by binding a Ni8 cluster
onto a ReN(001) surface. Here, the VN site is located at the second
nearest neighbor site (N[1]) with respect to Ni8 cluster (Figure S23).
The nitrogen vacancy formation energy (ENV) of the ReN(001) and
Ni/ReN surfaces are defined as (eq 6):
ENV = Etot(VN/surface) − Etot(surface) + 1/2Etot(N2)
(6)
where Etot(VN/surface) is the total energy for the optimized surface
nitrogen atom desorption configuration at different surfaces,
Etot(surface) is the total energy of the surface model, and Etot(N2)
is the total energy of an N2 molecule.
The adsorption energies of X [Ead(X)] (X represents H2 or N2)
species on the surfaces are defined as (eq 7):
Ead(X) = Etot(X/surface) − Etot(surface) − Etot(X)
(7)
where Etot(X/surface) is the total energy of the optimized X
molecule/atom adsorption configuration, and Etot(X) is the total
energy of an X molecule or atom. The TSs were searched using the
climbing image-nudged elastic band (CI-NEB) method.48 The energy
and force convergence criteria were set to 10−6 eV and 0.05 eV·Å−1,
respectively. All molecules in the gas phase were relaxed in a box with
dimensions of 20 × 20 × 20 Å3.
Isotope Experiment. Ammonia synthesis from 15N2 isotope
(98%) and H2 was performed using a U-shaped glass reactor
connected to a closed gas circulation system. The mixture of 15N2 and
H2 gases (total pressure = 60 kPa, 15N2:H2 = 1:3) was introduced into
the glass system, and the catalyst (0.2 g) was heated at different
temperatures (Ni/CeN: 340 °C, CeN: 400 °C). The circulating gas
was monitored using gas chromatography (GC; GC-8A, Shimadzu)
with a Chromosorb 103 column to separate the NH3 gas from the
mixture, and the outlet gases from the chromatograph were also
detected with a quadrupole mass spectrometer (M-101QA-TDM,
Canon Anelva Corp.) and He was used as the carrier gas. The m/z =
2, 16, 17, 18, 28, 29, and 30 masses were monitored as a function of
time to follow the reaction. The N2 isotope exchange experiment was
performed using a U-shaped glass reactor in connection with a closed
gas circulation system. A mixture of 15N2 and 14N2, with total pressure
of 20 kPa (15N2:14N2 = 1:4), was adsorbed on the catalyst at the
reaction temperature until an adsorption equilibrium was reached.
The circulating pump placed in the system removes diffusional and
adsorption/desorption limitations. The circulating gas was monitored
with a quadrupole mass spectrometer (M-101QA-TDM, Canon
Anelva Corp.), and the mass-to-charge ratios (m/z) of 28, 29, and 30
were measured as a function of time.
Computational Calculations. All of the structural relaxation and
electronic structure calculations were performed using density
functional theory (DFT) as implemented in the Vienna Ab Initio
Simulation Package (VASP).43,44 The generalized gradient approx-
imation (GGA) with the Perdew−Burke−Ernzerhof (PBE) func-
tional45 was adopted in the DFT calculations, and the core electrons
were described using the projector augmented wave (PAW)
method.46,47 The PAW potentials used in this work were N, H, Ni,
La, Ce_3, and Y_sv from the VASP distribution, where the Ce_3
potential is that where one 4f electron is kept frozen in the core and
the valency of Ce is fixed to three. The Ce_3 potential was selected
because nitrogen vacancy generation leads to inordinate electron
doping into the Ce 4f bands with the DFT+U method using the Ce
potential. Spin polarization in Ni was considered. First, the lattice
parameters for bulk Ni and ReN were relaxed using Monkhorst−Pack
grids of 20 × 20 × 20 and 12 × 12 × 12, respectively. An energy
cutoff of 500 eV and total energy convergence of 10−6 eV were used
to create the plane wave basis set. The ReN(001) surfaces, based on
the optimized bulk lattice parameters, were selected for study because
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.0c06624.
■
sı
Structural characterizations, catalytic activity data, and
DFT calculation details including Figures S1−S23 and
Tables S1−S7 (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
Masaaki Kitano − Materials Research Center for Element
Strategy, Tokyo Institute of Technology, Midori-ku, Yokohama
226-8503, Japan; Precursory Research for Embryonic Science
and Technology (PRESTO), Japan Science and Technology
Agency (JST), Kawaguchi, Saitama 332-0012, Japan;
orcid.org/0000-0003-4466-7387; Email: kitano.m.aa@
m.titech.ac.jp
Hideo Hosono − Materials Research Center for Element
Strategy, Tokyo Institute of Technology, Midori-ku, Yokohama
226-8503, Japan; International Center for Materials
Nanoarchitectonics, National Institute for Materials Science,
Tsukuba, Ibaraki 305-0044, Japan; orcid.org/0000-0001-
9260-6728; Email: hosono@mces.titech.ac.jp
Authors
Tian-Nan Ye − Materials Research Center for Element Strategy,
Tokyo Institute of Technology, Midori-ku, Yokohama 226-
8503, Japan; orcid.org/0000-0001-9039-8232
Sang-Won Park − Materials Research Center for Element
Strategy, Tokyo Institute of Technology, Midori-ku, Yokohama
H
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(9) Zhu, D.; Zhang, L. H.; Ruther, R. E.; Hamers, R. J. Photo-
illuminated diamond as a solid-state source of solvated electrons in
water for nitrogen reduction. Nat. Mater. 2013, 12, 836−841.
(10) Li, H.; Shang, J.; Ai, Z. H.; Zhang, L. Z. Efficient visible light
nitrogen fixation with BiOBr nanosheets of oxygen vacancies on the
exposed {001} facets. J. Am. Chem. Soc. 2015, 137, 6393−6399.
(11) Tompsett, D. A.; Parker, S. C.; Islam, M. S. Rutile (β-)MnO2
surfaces and vacancy formation for high electrochemical and catalytic
performance. J. Am. Chem. Soc. 2014, 136, 1418−1426.
226-8503, Japan; International Center for Materials
Nanoarchitectonics, National Institute for Materials Science,
Tsukuba, Ibaraki 305-0044, Japan; orcid.org/0000-0002-
2843-9803
Yangfan Lu − Materials Research Center for Element Strategy,
Tokyo Institute of Technology, Midori-ku, Yokohama 226-
8503, Japan; orcid.org/0000-0003-2084-9549
Jiang Li − Materials Research Center for Element Strategy, Tokyo
Institute of Technology, Midori-ku, Yokohama 226-8503, Japan
Masato Sasase − Materials Research Center for Element
Strategy, Tokyo Institute of Technology, Midori-ku, Yokohama
226-8503, Japan
(12) Jeen, H.; Bi, Z.; Choi, W.; Chisholm, M. F.; Bridges, C. A.;
Paranthaman, M. P.; Lee, H. Orienting oxygen vacancies for fast
catalytic reaction. Adv. Mater. 2013, 25, 6459−6463.
(13) Zhao, D.; Chen, Z.; Yang, W.; Liu, S.; Zhang, X.; Yu, Y.;
Cheong, W.; Zheng, L.; Ren, F.; Ying, G.; Cao, X.; Wang, D.; Peng,
Q.; Wang, G.; Chen, C. MXene (Ti3C2) vacancy-confined single-atom
catalyst for efficient functionalization of CO2. J. Am. Chem. Soc. 2019,
141, 4086−4093.
(14) Mars, P.; van Krevelen, D. W. Oxidations carried out by means
of vanadium oxide catalysts. Chem. Eng. Sci. 1954, 3, 41−59.
(15) Campbell, C. T.; Peden, C. H. F. Oxygen vacancies and
catalysis on ceria surfaces. Science 2005, 309, 713−714.
(16) Esch, F.; Fabris, S.; Zhou, L.; Montini, T.; Africh, C.;
Fornasiero, P.; Comelli, G.; Rosei, R. Electron localization determines
defect formation on ceria substrates. Science 2005, 309, 752−755.
(17) Lawrence, N. J.; Brewer, J. R.; Wang, L.; Wu, T.; Wells-
Kingsbury, J.; Ihrig, M. M.; Wang, G.; Soo, Y.; Mei, W.; Cheung, C.
Defect engineering in cubic cerium oxide nanostructures for catalytic
oxidation. Nano Lett. 2011, 11, 2666−2671.
(18) Hargreaves, J. S. J. Nitrides as ammonia synthesis catalysts and
as potential nitrogen transfer reagents. Appl. Petrochem. Res. 2014, 4,
3−10.
(19) Zeinalipour-Yazdi, C. D.; Hargreaves, J. S. J.; Catlow, C. R. A.
Nitrogen activation in a Mars−van Krevelen mechanism for ammonia
synthesis on Co3Mo3N. J. Phys. Chem. C 2015, 119, 28368−28376.
(20) Laassiri, S.; Zeinalipour-Yazdi, C. D.; Catlow, C. R. A.;
Hargreaves, J. S. J. The potential of manganese nitride based materials
as nitrogen transfer reagents for nitrogen chemical looping. Appl.
Catal., B 2018, 223, 60−66.
(21) Kitano, M.; Inoue, Y.; Yamazaki, Y.; Hayashi, F.; Kanbara, S.;
Matsuishi, S.; Yokoyama, T.; Kim, S. W.; Hara, M.; Hosono, H.
Ammonia synthesis using a stable electride as an electron donor and
reversible hydrogen store. Nat. Chem. 2012, 4, 934−940.
(22) Kitano, M.; Kanbara, S.; Inoue, Y.; Kuganathan, N.; Sushko, P.
V.; Yokoyama, T.; Hara, M.; Hosono, H. Electride support boosts
nitrogen dissociation over ruthenium catalyst and shifts the bottleneck
in ammonia synthesis. Nat. Commun. 2015, 6, 6731.
(23) Kitano, M.; Inoue, Y.; Ishikawa, H.; Yamagata, K.; Nakao, T.;
Tada, T.; Matsuishi, S.; Yokoyama, T.; Hara, M.; Hosono, H.
Essential role of hydride ion in ruthenium-based ammonia synthesis
catalysts. Chem. Sci. 2016, 7, 4036−4043.
(24) Kitano, M.; Inoue, Y.; Sasase, M.; Kishida, K.; Kobayashi, Y.;
Nishiyama, K.; Tada, T.; Kawamura, S.; Yokoyama, T.; Hara, M.;
Hosono, H. Self-organized ruthenium-barium core-shell nanoparticles
on a mesoporous calcium amide matrix for efficient low temperature
ammonia synthesis. Angew. Chem., Int. Ed. 2018, 57, 2648−2652.
(25) Inoue, Y.; Kitano, M.; Kishida, K.; Abe, H.; Niwa, Y.; Sasase,
M.; Fujita, Y.; Ishikawa, H.; Yokoyama, T.; Hara, M.; Hosono, H.
Efficient and stable ammonia synthesis by self-organized flat Ru
nanoparticles on calcium amide. ACS Catal. 2016, 6, 7577−7584.
(26) Kanbara, S.; Kitano, M.; Inoue, Y.; Yokoyama, T.; Hara, M.;
Hosono, H. Mechanism switching of ammonia synthesis over Ru-
loaded electride catalyst at metal-insulator transition. J. Am. Chem.
Soc. 2015, 137, 14517−14524.
(27) Wang, P. K.; Chang, F.; Gao, W. B.; Guo, J. P.; Wu, G. T.; He,
T.; Chen, P. Breaking scaling relations to achieve low-temperature
ammonia synthesis through LiH-mediated nitrogen transfer and
hydrogenation. Nat. Chem. 2017, 9, 64−70.
(28) Chang, F.; Guan, Y. Q.; Chang, X. H.; Guo, J. P.; Wang, P. K.;
Gao, W. B.; Wu, G. T.; Zheng, J.; Li, X. G.; Chen, P. Alkali and
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.0c06624
Author Contributions
#T.N.Y., S.W.P., and Y.L. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by a grant from the Element Strategy
Initiative to Form Core Research Center (No.
JPMXP0112101001) of the Ministry of Education, Culture,
Sports, Science and Technology of Japan (MEXT). A part of
this work was supported by a PRESTO Grant (No.
JPMJPR18T6) from the Japan Science and Technology
Agency (JST) and Kakenhi Grants-in-Aid (Nos. 17H06153,
JP19H05051, and JP19H02512) from the Japan Society for the
Promotion of Science (JSPS). T. N. Ye is supported by a JSPS
fellowship for International Research Fellows (No. P18361). Y.
F. Lu is supported by a JSPS fellowship for young scientists
(No. 18J00745). The authors thank Dr. Y. Sato (Tokyo
Institute of Technology) for technical support with the Auger
electron spectroscopy measurements.
REFERENCES
■
(1) Smil, V. Detonator of the population explosion. Nature 1999,
400, 415.
(2) Pool, A. J.; Lobkovsky, E.; Chirik, P. J. Hydrogenation and
cleavage of dinitrogen to ammonia with a zirconium complex. Nature
2004, 427, 527−530.
(3) Gambarotta, S.; Scott, J. Multimetallic cooperative activation of
N2. Angew. Chem., Int. Ed. 2004, 43, 5298−5308.
(4) Honkala, K.; Hellman, A.; Remediakis, I. N.; Logadottir, A.;
Carlsson, A.; Dahl, S.; Christensen, C. H.; Nørskov, J. K. Ammonia
synthesis from first-principles calculations. Science 2005, 307, 555−
558.
(5) Erisman, J. W.; Sutton, M. A.; Galloway, J.; Klimont, Z.;
Winiwarter, W. How a century of ammonia synthesis changed the
world. Nat. Geosci. 2008, 1, 636−639.
(6) 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. Beyond fossil fuel-
driven nitrogen transformations. Science 2018, 360, eaar6611.
(7) Licht, S.; Cui, B.; Wang, B.; Li, F.; Lau, J.; Liu, S. Ammonia
synthesis by N2 and steam electrolysis in molten hydroxide
suspensions of nanoscale Fe2O3. Science 2014, 345, 637−640.
(8) Hao, Y.; Guo, Y.; Chen, L.; Shu, M.; Wang, X.; Bu, T.; Gao, W.;
Zhang, N.; Su, X.; Feng, X.; Zhou, J.; Wang, B.; Hu, C.; Yin, A.; Si, R.;
Zhang, Y.; Yan, C. Promoting nitrogen electro-reduction to ammonia
with bismuth nanocrystals and potassium cations in water. Nat. Catal.
2019, 2, 448−456.
I
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
̈
alkaline earth hydrides-driven N2 activation and transformation over
Mn nitride catalyst. J. Am. Chem. Soc. 2018, 140, 14799−14806.
(29) Gao, W. B.; Wang, P. K.; Guo, J. P.; Chang, F.; He, T.; Wang,
Q. R.; Wu, G. T.; Chen, P. Barium hydride-mediated nitrogen transfer
and hydrogenation for ammonia synthesis: A case study of cobalt.
ACS Catal. 2017, 7, 3654−3661.
(30) Kitano, M.; Kujirai, J.; Ogasawara, K.; Matsuishi, S.; Tada, T.;
Abe, H.; Niwa, Y.; Hosono, H. Low-temperature synthesis of
perovskite oxynitride-hydrides as ammonia synthesis catalysts. J. Am.
Chem. Soc. 2019, 141, 20344−20353.
(46) Blochl, P. E. Projector augmented-wave method. Phys. Rev. B:
Condens. Matter Mater. Phys. 1994, 50, 17953−17979.
(47) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient
approximation made simple. [Phys. Rev. Lett. 1996, 77, 3865]. Phys.
Rev. Lett. 1997, 78, 1396−1396.
́
(48) Henkelman, G.; Jonsson, H. Improved tangent estimate in the
nudged elastic band method for finding minimum energy paths and
saddle points. J. Chem. Phys. 2000, 113, 9978−9985.
(31) Kobayashi, Y.; Hernandez, O. J.; Sakaguchi, T.; Yajima, T.;
Roisnel, T.; Tsujimoto, Y.; Morita, M.; Noda, Y.; Mogami, Y.; Kitada,
A.; Ohkura, M.; Hosokawa, S.; Li, Z. F.; Hayashi, K.; Kusano, Y.; Kim,
J. E.; Tsuji, N.; Fujiwara, A.; Matsushita, Y.; Yoshimura, K.;
Takegoshi, K.; Inoue, M.; Takano, M.; Kageyama, H. An oxyhydride
of BaTiO3 exhibiting hydride exchange and electronic conductivity.
Nat. Mater. 2012, 11, 507−511.
(32) Kobayashi, Y.; Tang, Y.; Kageyama, T.; Yamashita, H.; Masuda,
N.; Hosokawa, S.; Kageyama, H. Titanium-based hydrides as
heterogeneous catalysts for ammonia synthesis. J. Am. Chem. Soc.
2017, 139, 18240−18246.
(33) Tang, Y.; Kobayashi, Y.; Masuda, N.; Uchida, Y.; Okamoto, H.;
Kageyama, T.; Hosokawa, S.; Loyer, F.; Mitsuhara, K.; Yamanaka, K.;
Tamenori, Y.; Tassel, C.; Yamamoto, T.; Tanaka, T.; Kageyama, H.
Metal-dependent support effects of oxyhydride-supported Ru, Fe, Co
Catalysts for ammonia synthesis. Adv. Energy Mater. 2018, 8,
1801772.
(34) Ye, T. N.; Park, S. W.; Lu, Y.; Li, J.; Sasase, M.; Kitano, M.;
Tada, T.; Hosono, H. Vacancy-enabled N2 activation for ammonia
synthesis on an Ni-loaded catalyst. Nature 2020, 583, 391−395.
(35) Gao, W.; Guo, J.; Wang, P.; Wang, Q.; Chang, F.; Pei, Q.;
Zhang, W.; Liu, L.; Chen, P. Production of ammonia via a chemical
looping process based on metal imides as nitrogen carriers. Nat.
Energy 2018, 3, 1067−1075.
(36) Kojima, R.; Aika, K. Cobalt molybdenum bimetallic nitride
catalysts for ammonia synthesis: Part 2. Kinetic study. Appl. Catal., A
2001, 218, 121−128.
(37) Bion, N.; Can, F.; Cook, J.; Hargreaves, J. S. J.; Hector, A. L.;
Levason, W.; McFarlane, A. R.; Richard, M.; Sardar, K. The role of
preparation route upon the ambient pressure ammonia synthesis
activity of Ni2Mo3N. Appl. Catal., A 2015, 504, 44−50.
(38) Inoue, Y.; Kitano, M.; Tokunari, M.; Taniguchi, T.; Ooya, K.;
Abe, H.; Niwa, Y.; Sasase, M.; Hara, M.; Hosono, H. Direct activation
of cobalt catalyst by 12CaO·7Al2O3 electride for ammonia synthesis.
ACS Catal. 2019, 9, 1670−1679.
(39) Gong, Y. T.; Wu, J. Z.; Kitano, M.; Wang, J. J.; Ye, T. N.; Li, J.;
Kobayashi, Y.; Kishida, K.; Abe, H.; Niwa, Y.; Yang, H. S.; Tada, T.;
Hosono, H. Ternary intermetallic LaCoSi as a catalyst for N2
activation. Nat. Catal. 2018, 1, 178−185.
(40) Wu, J.; Li, J.; Gong, Y.; Kitano, M.; Inoshita, T.; Hosono, H.
Intermetallic electride catalyst as a platform for ammonia synthesis.
Angew. Chem., Int. Ed. 2019, 58, 825−829.
(41) Balasubramanian, K.; Khare, S. V.; Gall, D. Energetics of point
defects in rocksalt structure transition metal nitrides: Thermodynamic
reasons for deviations from stoichiometry. Acta Mater. 2018, 159, 77−
88.
(42) Lu, Y.; Li, J.; Ye, T. N.; Kobayashi, Y.; Sasase, M.; Kitano, M.;
Hosono, H. Synthesis of rare-earth-based metallic electride nano-
particles and their catalytic applications to selective hydrogenation
and ammonia synthesis. ACS Catal. 2018, 8, 11054−11058.
(43) Kresse, G.; Furthmuller, J. Efficient iterative schemes for ab
̈
initio total-energy calculations using a plane-wave basis set. Phys. Rev.
B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186.
(44) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the
projector augmented-wave method. Phys. Rev. B: Condens. Matter
Mater. Phys. 1999, 59, 1758−1775.
(45) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient
approximation made simple. Phys. Rev. Lett. 1996, 77, 3865−3868.
J
https://dx.doi.org/10.1021/jacs.0c06624
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX