Catalytic nitrogen fixation via direct cleavage of nitrogen-nitrogen triple bond of molecular dinitrogen under ambient reaction conditions

Article abstract of DOI:10.1246/bcsj.20170197

We have now found that molybdenumiodide complexes bearing a PNP-pincer ligand have a higher catalytic activity than the so far reported molybdenumdinitrogen complexes for ammonia formation from nitrogen gas under ambient reaction conditions, up to 830 equiv being produced based on a dinitrogen-bridged dimolybdenum complex (415 equiv of ammonia based on the molybdenum atom). This remarkable catalytic activity is induced by a novel reaction pathway, where the generation of a dinitrogen-bridged dimolybdenumiodide complex is a key point to promote direct cleavage of the nitrogennitrogen triple bond of the bridging dinitrogen ligand in the MoN τNMo core.

Full text of DOI:10.1246/bcsj.20170197

Advance Publication Cover Page
Catalytic Nitrogen Fixation via Direct Cleavage of Nitrogen–Nitrogen Triple Bond
of Molecular Dinitrogen under Ambient Reaction Conditions
Kazuya Arashiba, Aya Eizawa, Hiromasa Tanaka, Kazunari Nakajima,
Kazunari Yoshizawa,* Yoshiaki Nishibayashi*
Advance Publication on the web July 27, 2017
doi:10.1246/bcsj.20170197
©
2017 The Chemical Society of Japan

Catalytic Nitrogen Fixation via Direct Cleavage of Nitrogen–Nitrogen Triple Bond of Molecular
Dinitrogen under Ambient Reaction Conditions
1
1
2
1
2
Kazuya Arashiba, Aya Eizawa, Hiromasa Tanaka, Kazunari Nakajima, Kazunari Yoshizawa, * Yoshiaki
Nishibayashi *
1
1
Department of Systems Innovation, School of Engineering, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan
2
Institute for Materials Chemistry and Engineering, Kyushu University, Nishi-ku, Fukuoka 819-0395, Japan
E-mail: kazunari@ms.ifoc.kyushu-u.ac.jp, ynishiba@sys.t.u-tokyo.ac.jp
Received: June 17, 2017; Accepted: July 00, 2017; Web Released: July 00, 2017
Abstract
We have now found that molybdenum–iodide complexes
bearing a PNP-pincer ligand have a higher catalytic activity
than the so far reported molybdenum–dinitrogen complexes for
ammonia formation from nitrogen gas under ambient reaction
conditions, up to 830 equiv being produced based on a
dinitrogen-bridged dimolybdenum complex (415 equiv of
ammonia based on the molybdenum atom).
This remarkable
catalytic activity is induced by a novel reaction pathway, where
the generation of a dinitrogen-bridged dimolybdenum–iodide
complex is a key point to promote direct cleavage of
nitrogen–nitrogen triple bond of the bridging dinitrogen ligand
in the Mo–N≡N–Mo core.
On the basis of these previous findings of the
catalytic nitrogen fixation, we have envisaged that the
generation of the nitride complex via direct cleavage of the
bridging dinitrogen ligand in the catalytic cycle may provide
a novel pathway for the catalytic nitrogen fixation
under ambient reaction conditions (Path B in Figure 1
(b)). In this new reaction pathway, the dinitrogen complex
is directly converted into the corresponding nitride complex,
where the diazenide, hydrazide, and hydrazidium complexes
are not involved as reactive intermediates.
1
. Introduction
Industrially, the Haber–Bosch process is well known
to produce ammonia from nitrogen and hydrogen gases
under drastic reaction conditions such as high
temperature and pressure.
annual primary energy supply in the world are consumed in
this ammonia production. In sharp contrast, nitrogenase
enzymes promote the conversion of nitrogen gas into
1
At least 1-2% of the
ammonia under ambient reaction conditions.
Recent
Since the breakthrough discovered by Cummins and a co-
worker, many groups have already found direct cleavage of
1
4
studies have revealed the molecular structure of the
active site of nitrogenase in FeMo cofactor although the
detailed mechanism of the function of nitrogenase still
nitrogen–nitrogen triple bond of the bridging dinitrogen ligand
1
5,16
in the M–N≡N–M core.
Recently, several groups have
2
,3
remains unclear.
reported further transformations of the nitride complexes,
which are generated via cleavage of nitrogen–nitrogen triple
bond of the bridging dinitrogen, into nitrogen-containing
The development of catalytic nitrogen fixation under mild
reaction conditions using transition metal–dinitrogen
1
6a,h-j
complexes as catalysts has been one of the most challenging
organic compounds such as organic nitriles (RC≡N).
In
4
topics in chemistry.
Despite extensive studies on the
some cases, pseudocatalytic cycles for the stepwise conversion
of molecular dinitrogen into nitrogen-containing organic
compounds have been achieved by using a stoichiometric
preparation of various transition metal–dinitrogen complexes
and their reactivity, successful examples of direct reduction of
molecular dinitrogen into ammonia and/or hydrazine
1
6a,h-j
amount of transition metal complexes.
On the other hand,
(
NH NH ) under mild reaction conditions using transition
only a few nitride complexes generated via cleavage of
nitrogen–nitrogen triple bond of the bridging dinitrogen have
been reported to form nitrogen–hydrogen bonds leading to the
2
2
metal–dinitrogen complexes as catalysts have been quite
limited until now.
nitride complexes (M≡N) are considered to be key reactive
5
-10
In the molybdenum-catalyzed reactions,
1
6e,17
stoichiometric formation of ammonia.
However, there is
intermediates via diazenide (M-NNH), hydrazide (M≡NNH ),
no successful example of catalytic nitrogen fixation via direct
cleavage of molecular dinitrogen as a key step under ambient
reaction conditions until now.
2
+
and hydrazidium (M-NNH ) complexes after the formation of
3
1
1,12
ammonia.
In the iron- and cobalt-catalyzed reactions, a
similar reaction pathway is proposed involving diazenide,
hydrazide, and hydrazine (M-NH NH ) complexes as key
Quite recently, we have found that some
molybdenum–nitride
pincer ligands worked as effective catalysts toward
ammonia formation from nitrogen gas under
ambient reaction conditions. After an extensive
complexes
bearing
PNP-type
2
2
reactive intermediates although a detailed reaction pathway
1
3
has not yet been clarified.
In both catalytic systems, the first
1
2,18,19
step is considered to be the protonation to the terminal
dinitrogen ligand on the metal center to form a nitrogen–
hydrogen bond (Path A in Figure 1 (a)).
study, we have now found that similar molybdenum–
nitride complexes can be

Table 1. Catalytic nitrogen fixation via a novel reaction pathway.[a]
generated from dinitrogen-bridged dimolybdenum–iodide
complexes via cleavage of nitrogen–nitrogen triple bond of the
bridging dinitrogen in the Mo–N≡N–Mo core and they have a
high catalytic activity toward nitrogen fixation using molecular
catalysts, where up to 830 equiv are formed based on the
dinitrogen-bridged dimolybdenum complex (415 equiv of
R1
M
OTf
R2
R1
R1
R1
cat.
1
N2 + 6
1 atm)
R
+ 6
2 NH3 ( + H2 )
1
1
R
R
toluene
rt, 20 h
(
N
H
R1
R1
R1
1
2
CoCp*2: M = Co, R = Me [ColH]OTf: R = Me
CrCp*2: M = Cr, R = Me [LutH]OTf: R = H
CoCp2: M = Co, R = H
1
2
ammonia based on the molybdenum atom).
We are sure that
1
this high catalytic activity is realized by a novel reaction
pathway via direct cleavage of nitrogen–nitrogen triple bond of
nitrogen gas under ambient reaction conditions.
Entry Cat. Reductant Proton source NH3 (equiv/Mo) NH3 (%)[b] H2 (equiv/Mo) H2 (%)[b]
10.9±0.2[c]
91±2[c]
0.2±0.1[c]
1±1[c]
1
2
3
4
5
6
1a
2a
3
CoCp*2
CoCp*2
CoCp*2
CrCp*2
CoCp2
[ColH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
[LutH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
[ColH]OTf
12.2±0.1[c]
7.1±0.4[c]
10.3±0.2[c]
4.2
3.1±0.2[c]
50.7±0.4[c]
40.5
96±1[c]
59±4[c]
87±1[c]
35
26±2[c]
84±1[c]
69
0.2±0.1[c]
4.0±0.2[c]
1.3±0.2[c]
5.7
1.1±0.1[c]
0.5±0.1[c]
0.9
1±1[c]
22±1[c]
7±1[c]
32
7±1[c]
3±1[c]
5
1a
1a
1a
2
. Results and Discussion
CoCp*2
CoCp*2
CoCp*2
CoCp*2
CoCp*2
CoCp*2
CoCp*2
CoCp*2
7[d] 1a
8[d]
9[d]
Treatment
of
[MoI (PNP)]
(1a;
PNP
=
3
1c
1b
4a
4b
2c
2b
2
,6-bis(di-tert-butylphosphinomethyl)pyridine) with 3.3 equiv
24.4
41
2.4
13
5
of decamethylcobaltocene CoCp*2 (Cp* =  -C Me ) as a
5
5
10
11
6.3
53
1.4
8
reductant in toluene at room temperature for 2 h under an
atmospheric pressure of nitrogen gas gave the corresponding
nitride complex [Mo(≡N)I(PNP)] (2a) in 75% NMR yield
together with a dinitrogen-bridged dimolybdenum(0) complex
6.4
53
1.7
9
1
1
2
3
11.1
88
0.3
2
8.4
66
0.7
4
[
a] To a mixture of molybdenum complex (0.010 mmol) and [ColH]OTf (0.48 mmol) in toluene (1 mL) was
added a solution of CoCp*2 (0.36 mmol) in toluene (4 mL) at room temperature over a period of 1 h, followed
by stirring at room temperature for another 19 h under N2. [b] Yield based on a reductant. [c] Average of
some individual experiments (at least 2). [d] 1 (0.002 mmol) was used as a catalyst.
[
Mo(N ) (PNP)] (-N ) (3) in 4% yield (Scheme 1).
2 2 2 2
N
t
N
Bu2
P
N
X
Mo
X
N
t
t
N
X
P Bu2
P Bu2
t
t
P Bu2
P Bu2
Ph
N
X
N
Mo
N
Mo
N
N
N
Mo
N
P
Mo
X
X
P
PtBu2
t
P
P
t
X
Bu2
N t
P
t
Bu2
Bu2
Bu2
1
1
1
a: X = I
b: X =Cl
c: X = Br
2a: X = I
2b: X =Cl
2c: X = Br
N
3
N
4a: X = I
4b: X =Cl
3
from reduction of 1a with CoCp*2 under an
atmospheric pressure of nitrogen gas. This new finding
prompted us to investigate catalytic nitrogen fixation
using CoCp*2 as a reductant in the presence of a catalytic
amount of 1a under ambient reaction conditions.
To
collidinium trifluoromethanesulfonate (48 equiv to Mo atom;
ColH]OTf) as a proton source in toluene was added a
a
mixture
of
1a
and
2,4,6-
The nitride complex 2a was characterized spectroscopically.
Detailed molecular structure of 2a was determined by X-ray
[
solution of CoCp* (36 equiv to 1a) as a reductant in toluene
2
2
0
crystallography (Figure 2).
The result of the stoichiometric
via a syringe pump at room temperature over a period of 1 h
under an atmospheric pressure of nitrogen gas, followed
by stirring at room temperature for another 19 h under an
reaction indicates that complex 2a may be formed from a
dinitrogen-bridged dimolybdenum–iodide complex via
cleavage of nitrogen–nitrogen triple bond of the bridging
dinitrogen in the Mo–N≡N–Mo core, which was generated in
atmospheric pressure of nitrogen gas.
After the reaction,
ammonia (10.9 equiv based on Mo atom) was formed
together with a small amount of hydrogen gas (0.2 equiv
based on Mo atom), where ammonia and molecular
dihydrogen were obtained in 91% and 1% yields based on
situ from reduction of 1a with CoCp* .
This proposal is
2
1
5
supported by the
N-labelling experiment, where
1
5
15
[
Mo(≡ N)I(PNP)] (2a’) was obtained when N gas was used
2
1
4
in place of normal N gas under the same reaction conditions.
2
CoCp* , respectively (Table 1, Entry 1).
The amounts
2
Many research groups have already found direct cleavage of
nitrogen–nitrogen triple bond of a bridging dinitrogen ligand in
the M–N≡N–M core (vide infra).
of
ammonia
and
molecular
dihydrogen were
21
determined by indophenol method and GC, respectively.
No other product such as hydrazine was observed at all.
When 2a was used as a catalyst in place of 1a, 12.2 equiv
of ammonia based on Mo atom (96% yield) and 0.2 equiv
of hydrogen gas were obtained based on Mo atom (1%
yield), respectively (Table 1, Entry 2).
Separately, we
confirmed the direct conversion of molecular dinitrogen
into ammonia by using 1a as a catalyst under an atmospheric
1
5
14
22
pressure of N gas in place of normal N gas.
2
2
For comparison, the use of 3 as a catalyst in place of
a under the same reaction conditions gave only a lower
1
amount of ammonia (7.1 equiv based on Mo atom; 59% yield
based on CoCp*2) together with larger amount of
hydrogen gas (4.0 equiv based on Mo atom; 22% yield based
on CoCp* ) (Table 1, Entry 3). This result indicated that the
catalytic activity of 3 toward the ammonia formation (59%
yield) is substantially lower than that of 1a and 2a
The experimental result shown in Scheme 1 is in sharp
a
contrast to the previous report that reduction of [MoCl (PNP)]
3
(
1b) with an excess amount of Na-Hg (6 equiv) as reductant in
THF at room temperature for 12 under an
atmospheric pressure of nitrogen gas gave only 3 in 63%
2
h
6
yield.
the halide ligand in molybdenum complexes as well as
reductants. These results suggest that the formation of 2a is
preferred than that of
Thus, the reactivity much depended on the nature of
(91% and 96% yields, respectively).
In the reaction
using 1a and 2a as catalysts, we did not observe any free
PNP ligand after the catalytic reaction.
result was in sharp contrast to
This experimental

our previous reaction system using 3 as a catalyst, where we
observed only a free PNP ligand from the reaction mixture after
used as catalysts in place of [MoI
yield), showing that the catalytic activity of [MoX
3
(PNP)] 1a (50.7 equiv; 84%
(PNP)]
3
the catalytic reaction.
It is noteworthy that the use of
toward ammonia formation depended on the nature of the
halide ligand in molybdenum complexes (Table 1, Entries
7-9).
molybdenum–iodide complexes such as 1a and 2a as catalysts
dramatically improved the catalytic activity toward the
formation of ammonia and completely inhibited the formation
of hydrogen gas.
Almost the same result was observed when CrCp* was
used as a slightly weaker reducing reagent in place of CoCp*2,
2
6,27
Quite recently, Mézailles and co-workers reported the
Cy
formation of a similar nitride complex [Mo(≡N)I( PPP)]
Cy
2
( PPP = bis(cyclohexylphosphinoethyl)phenylphosphine) via
2
3
direct cleavage of nitrogen–nitrogen triple bond of the bridging
5
Cy
but the use of CoCp (Cp =  -C H ) as a weaker reducing
dinitrogen ligand produced by reduction of [MoCl ( PPP)] in
2
5
5
3
2
8
reagent in place of CoCp* and CrCp* dramatically decreased
the amount of ammonia (4.2 equiv based on Mo atom, 35%
yield) (Table 1, Entries 4 and 5). On the other hand, slightly
more acidic [LutH]OTf than [ColH]OTf was not applicable as
a suitable proton source although we have previously found
the presence of NaI.
We have newly confirmed that similar
2
2
tBu
tBu
molybdenum complexes [MoX3( PPP)] ( PPP = bis(di-tert-
butylphosphinoethyl)phenylphosphine; 4a: X = I; 4b: X = Cl )
worked as less effective catalysts than [MoX (PNP)] 1 under
2
9
2
4
3
the same reaction conditions (Table 1, Entries 10 and 11). This
result indicates that the application of not PPP ligand but PNP-
pincer ligand to the molybdenum atom is necessary to achieve
the higher catalytic performance on catalytic nitrogen fixation.
For comparison of the catalytic activity, we monitored the
that [LutH]OTf was the most efficient proton source in our
6
catalytic reactions (Table 1, Entry 6).
These results
indicated that the combination of CoCp*2 having a stronger
reducing ability and [ColH]OTf having a weaker acidity is
suitable to achieve the higher efficiency than the previous
combination of CoCp2 and [LutH]OTf.
observed no formation of ammonia in the absence of
molybdenum complexes as catalysts.
formation of ammonia by using [MoX
Typical results are shown in Figure 3, together with its time
profile using 3 as a catalyst. The catalytic reactions using 1a
and 1c as catalysts proceeded more rapidly than that using 1b
as a catalyst. The turnover frequencies (TOFs) for ammonia
(PNP)] as catalysts.
3
Separately, we
Next, we carried out the catalytic reaction in the presence
of 1a by using larger amounts of CoCp* as a reductant and
(determined as moles of ammonia per Mo atom produced in the
2
-1
-1
-1
[ColH]OTf as a proton source under similar reaction conditions.
initial 1 h) using 1a, 1c, and 1b were 28 h , 29 h , and 23 h
-1
To prevent the formation of hydrogen gas from the direct
respectively, compared with 12
h
using 3.
These
reaction of CoCp* with [ColH]OTf in the reaction mixture, we
added CoCp*2 and [ColH]OTf to the reaction mixture in a
experimental results shown in Table 1, Scheme 2, and Figure 3
indicate that 1a has the highest performance on not only the
catalytic activity but also on the rate of ammonia formation.
Previously, we reported the formation of hydrogen gas as a
byproduct for the catalytic formation of ammonia using 3 as a
2
stepwise fashion.
The detailed experimental procedure and
typical results are shown in the Supporting Information.
With larger amounts of CoCp* and [ColH]OTf were employed
2
3
0
under the same reaction conditions, larger amounts of ammonia
were produced based on 1a as a catalyst although the yield of
catalyst.
formation of ammonia and hydrogen gas is complementary in
the catalytic reaction. In sharp contrast to the result using 3
as a catalyst, the use of 1 as catalysts inhibited almost
completely the formation of hydrogen gas. The time profiles
for the formation of hydrogen gas using 1 as catalysts were
quite different from that using 3 as a catalyst (Figure 3). In
all cases using 1 as catalysts, less than 2 equiv of hydrogen gas
were formed based on the molybdenum atom. The
suppression of the formation of hydrogen gas may contribute to
The previous information suggested that the
ammonia decreased gradually.
Finally, the use of 2160 equiv
of CoCp* and 2880 equiv of [ColH]OTf to the Mo atom of 1a
2
in the presence of a catalytic amount of 1a gave the maximum
amount of ammonia (415 equiv based on Mo atom; 59% yield
based on CoCp* ) together with 241 equiv of hydrogen gas
2
(
23% yield based on CoCp* ) (Scheme 2).
This is so far the
2
OTf
1a
(
0.001 mmol)
N2 + 6
1 atm)
Co
+ 6
2 NH3
toluene
rt, 20 h
(
N
H
4
15 equiv based on Mo atom
830 equiv based on the catalyst)
59% yield)
2
160 equiv 2880 equiv
(
(
Scheme 2. Catalytic reaction with larger amounts of CoCp*2 as a reductant and [ColH]OTf
as a proton source.
highest TON in catalytic nitrogen fixation using homogeneous
transition metal complexes as catalysts under ambient
2
5
reaction conditions.
As described in our previous report,
only 12 equiv of ammonia based on the Mo atom (23 equiv
based on the dinitrogen-bridged dimolybdenum complex) were
produced even by using larger amounts of reductants and
proton sources under similar reaction conditions when 3
6
was used as a catalyst.
Thus, the catalytic activity of 1a
as a pre-catalyst is 35 times higher than that of 3 as a catalyst
415 vs 12 equiv of ammonia based on the Mo atom; 830 vs
3 equiv based on the dinitrogen-bridged dimolybdenum
(
2
6
complex).
We then investigated the effect of halide ligands in
MoX (PNP)] 1 toward the present catalytic reaction. In
[
3
order to clarify the differences among the results, the catalytic
reactions were carried out using a smaller amount of 1
(0.002 mmol) as catalysts.
A lower amount of ammonia
(40.5 and 24.4 equiv respectively; 69% and 41% yields,
respectively) was obtained when [MoBr (PNP)] (1c) and
3
[
MoCl (PNP)] 1b were
3

the efficient formation of ammonia.
Thus, we believe that 1
[Mo(≡N)I(PNP)]
2a
>
[Mo(≡N)Br(PNP)]
2c
>>
has a unique catalytic behaviour not only toward ammonia
formation but also dihydrogen suppression in the present
reaction system.
[Mo(≡N)Cl(PNP)] 2b (Figure 5 (b)). This is due to the weaker
electron-donating ability of the iodide and bromide ligands than
2
6
that of chloride ligand.
The result revealed that reduction
As described in the previous section, the catalytic
behaviour of 1 was revealed to be quite different from that of 3.
These experimental results promoted us to investigate the
catalytic behaviour of both 1a and 3 from a viewpoint of
steps of catalytic reactions by using molybdenum–iodide and –
bromide complexes as catalysts proceed more smoothly than
those by using molybdenum–chloride complexes as catalysts.
These results suggest that not protonation steps but reduction
step may be involved in the rate-determining step of the present
kinetic study.
different initial concentrations of the catalyst at the early stage
of catalytic reactions. The plot of the v(NH ) against the
concentration of the molybdenum ([Mo]) is shown in Figure 4.
This plot shows that v(NH ) was independent on [Mo] when 1a
We estimated the kinetic data of 1a and 3 for
catalytic reaction.
molybdenum complexes under the same reaction conditions
was as follows: [Mo(≡N)I(PNP)] 2a (96%)
In fact, the catalytic activity of the
3
>
[Mo(≡N)Br(PNP)] 2c (88%) >> [Mo(≡N)Cl(PNP)] 2b (66%)
(Table 1, Entries 2, 12, and 13).
On the basis of new findings described in the present
manuscript, we have proposed a reaction pathway shown in
3
was used as a catalyst, while the v(NH ) apparently depended
3
on [Mo] when 3 was used as a catalyst.
According to the
plot of the log(v(NH )) against log [Mo] shown in Figure 4, the
3
reaction orders with respect to the molybdenum were estimated
Figure 6 (a). First, [MoI (PNP)] 1a is reduced with 2 equiv of
3
to be 0 for 1a and ca. 1 for 3, respectively.
reaction order to the molybdenum depends on the nature of
catalysts. The result of kinetic study suggests that the
reaction mechanism differs between the reaction systems using
Thus, the
CoCp*2 under an atmospheric pressure of nitrogen gas to
generate [MoI(PNP)–N≡N–MoI(PNP)] (B) from mononuclear
dinitrogen complex [Mo(N )I(PNP)] (A).
Then,
2
[Mo(≡N)I(PNP)] 2a is formed from B via cleavage of
nitrogen–nitrogen triple bond of the bridging dinitrogen in B.
Previously, Schrock and co-workers reported a similar
1
a and 3 as catalysts.
To obtain more information on reactive intermediates of the
catalytic reaction, we attempted to observe reactive species
from the reaction mixture using 1a as a catalyst. From the
reaction mixture, we detected 2a by TOF-MS. Separately, we
confirmed that a stoichiometric reaction of 2a with 3 equiv of
CoCp*2 and 4 equiv of [ColH]OTf under an atmospheric
pressure of argon gas gave ammonia in 60% yield based on the
molybdenum atom (Figure 5 (a)). These results clearly show
that the nitride complex 2a worked as a key reactive
intermediate in the present catalytic reaction. We estimated
the electronic property of the nitride complexes
[Mo(≡N)X(PNP)] (2a: X = I; 2b: X = Cl; 2c: X = Br) by
measuring the cyclic voltammetry of 2. The cyclic
voltammetry of 2 revealed one oxidative wave in the anodic
scan and one reductive wave in the reverse cathodic scan,
respectively, where the tendency to be reduced was as follows:

stoichiometric reaction by using [MoI (POCOP)] (POCOP =
is likely to occur only in the Mo(I)–N≡N–Mo(I) core.
2
1
,3-bis((di-tert-butylphosphanyl)oxy)benzenide) to generate the
As described in the present manuscript, molybdenum–
iodide complexes bearing a PNP-pincer ligand such as 1a and
2a have a higher catalytic activity than the so far reported
corresponding nitride complex [Mo(≡N)I(POCOP)] via direct
cleavage of nitrogen–nitrogen triple bond of the bridging
dinitrogen, although this molybdenum complex was found to
have no catalytic activity toward the formation of ammonia
2
9,35
molybdenum–dinitrogen complexes
from nitrogen gas under ambient reaction conditions, up to 830
equiv being produced based on dinitrogen-bridged
for ammonia formation
1
6e
from nitrogen gas under the same reaction conditions.
a
Stepwise protonation and reduction of 2a by [ColH]OTf and
dimolybdenum complex (415 equiv of ammonia based on the
molybdenum atom). If the formation of ammonia by using
molybdenum–iodide complexes as catalysts proceeds via the
same reaction pathway, where 3 worked as a catalyst (Path A
in Figure 1 (a)), we cannot explain this remarkable catalytic
activity of molybdenum–iodide complexes, the higher TONs
and TOFs of 1a than those of 3 and the higher selectivity of 1a
than that of 3 toward the formation of ammonia prior to
CoCp*2 give the corresponding ammonia complex
3
1
[
Mo(NH )I(PNP)] (C).
Previously, we confirmed that the
protonation of 2b by pyridinium trifluoromethanesulfonate
C H NH]OTf as a proton source under ambient reaction
3
[
5
5
conditions gave the corresponding imide complex
1
2
[
Mo(≡NH)Cl(PNP)(C H N)]OTf.
Finally, the ligand
5
5
exchange of C with another dinitrogen molecule affords the
starting complex A together with ammonia.
hydrogen gas (Table 1 and Figure 3).
The result of the
In order to support our plausible reaction pathway shown in
Figure 6 (a), we carried out DFT calculations at the B3LYP-D3
level of theory on direct cleavage of nitrogen–nitrogen triple
stoichiometric reaction shown in Scheme 1 indicates that the
formation of mononuclear dinitrogen complex such as A is
preferred than that of 3 from the reduction of 1a with an excess
bond of the bridging dinitrogen in B.
cleavage of nitrogen–nitrogen triple bond is shown in Figure 6
b) as the free energy changes at 298.15 K (G298) in toluene.
The electronic ground state of the Mo(I)–Mo(I) complex B
reactant complex, RC) is triplet, where both Mo atoms possess
an unpaired electron. The open- and closed-shell singlet
states of B lie 3.7 and 9.0 kcal/mol above the triplet state,
respectively. On the other hand, the ground spin state of the
product complex (PC) consisting of two molecule of 2a is the
closed-shell singlet. Two unpaired electrons in the triplet
The energetics of
amount of CoCp* under an atmospheric pressure of nitrogen
gas. After the formation of A, 2a is promptly produced via
2
(
cleavage of nitrogen–nitrogen triple bond of a bridging
3
6
dinitrogen in the Mo–N≡N–Mo core such as B.
(
We should consider another possibility that mononuclear
dinitrogen complexes such as A play key roles to promote the
catalytic reaction via the protonation of dinitrogen complex as
the first step (Path A in Figure 1 (a)).
However, the
protonation of A by [ColH]OTf is endergonic by 6.8 kcal/mol,
while the formation of B from A is exergonic by 14.7 kcal/
3
3
state of PC are localized at one of two 2a, because the formal
charge of +4 in the Mo atom of 2a forces the localization of the
mol.
This result indicates that the formation of the stable
Mo-N≡N-Mo core is more favored than the protonation of A by
[ColH]OTf. As a result, Path B in Figure 1 (b) proceeds more
smoothly than Path A in Figure 1 (a) when molybdenum–
iodide complexes such as 1a and 2a were used as catalysts.
On the basis of the findings described in the present
manuscript, we can point out that the catalytic formation of
ammonia using molybdenum–iodide complexes as catalysts
mainly proceeds not via the previous reaction pathway (Path A
in Figure 1 (a)) but via the novel reaction pathway (Path B in
Figure 1 (b)).
unpaired electrons at one Mo atom.
Thus, the energy gap
between the singlet and triplet states of PC (19.4 kcal/mol)
corresponds to that between the singlet and triplet states of 2a.
The closed-shell singlet of the transition state (TS) is 21.8
kcal/mol higher in energy than the triplet state of B, and the
triplet state of TS is much higher (34.5 kcal/mol).
These
results suggest that the spin inversion should occur to induce
cleavage of nitrogen–nitrogen triple bond at room temperature.
If so, the cleavage should proceed in a highly exergonic way
(
G298 = –27.6 kcal/mol) with an activation energy of 21.8
kcal/mol.
As described in the previous section, the use of
The novel reaction system has two advantages over the
previous reaction system.
One is diminution of
reaction steps of protonation and reduction in the catalytic
molybdenum–iodide complexes bearing a PNP-pincer ligand
such as 1a and 2a as catalysts has realized novel nitrogen
fixation via direct cleavage of nitrogen–nitrogen triple bond of
a bridging dinitrogen ligand in the Mo–N≡N–Mo core under
ambient reaction conditions. At present, we consider that the
generation of molybdenum(I)–iodide complex is a key point to
promote such direct cleavage. This hypothesis is supported by
our previous experimental result that no cleavage of nitrogen–
nitrogen triple bond of the bridging dinitrogen occurred at
cycle.
pathway,
In sharp contrast to the previous reaction
where six pairs of stepwise protonation and
6
,12
reduction are necessary, only three pairs to the molybdenum–
nitride complex promote the formation of ammonia from
dinitrogen molecule.
dihydrogen formation.
pathway, where hydrogen gas was produced as a side-product
when 3 was used as a catalyst, the formation of only a small
amount of hydrogen gas was observed in the novel reaction
The other is the inhibition of
In contrast to the previous reaction
dinitrogen-bridged
dimolybdenum(II)
tetrachloride
DFT calculations
system.
The formation of nitride complex via direct
3
2
[
MoCl (PNP)–N≡N–MoCl (PNP)].
cleavage of nitrogen–nitrogen triple bond may inhibit the
production of hydrogen gas because the Mo center of the
nitride complex does not have enough reactivity toward the
protonation by proton source to form molybdenum–hydrogen
2
2
demonstrated that the bridging dinitrogen ligands in the Mo(0)–
Mo(0) complex 3 and Mo(II)–Mo(II) complexes such
as [MoI (PNP)–N≡N–MoI (PNP)] (D) would not split.
If
2
2
intermediate
D
with
the
geometry
same
as
bond for the production of hydrogen gas.
On the basis of the
3
2
[
MoCl (PNP)–N≡N–MoCl (PNP)] is formed, the N≡N bond
findings described in the present manuscript, we can state that
the catalytic activity of molybdenum–iodide complexes such as
1a and 2a meets the requirement of the novel reaction pathway.
2
2
cleavage
of
D
yielding
two
molecules
of
trans-[Mo(V)I (≡N)(PNP)] is endergonic by 31.2 kcal/mol with
2
a very high activation energy of 62.6 kcal/mol.
bond cleavage of yielding two molecules of
trans-[Mo(III)(N ) (≡N)(PNP)] is also extremely endergonic by
The N≡N
3
3. Conclusion
2
2
3
3
7
1.6 kcal/mol.
These results indicate that direct cleavage of
We have now found that molybdenum–iodide complexes
bearing a PNP-pincer ligand have a higher catalytic activity
than the so far reported molybdenum–dinitrogen complexes
nitrogen–nitrogen triple bond of a bridging dinitrogen ligand
2
9,35

for ammonia formation from nitrogen gas under ambient
reaction conditions, up to 830 equiv being produced based on
the dinitrogen-bridged dimolybdenum complex (415 equiv of
ammonia based on the molybdenum atom). This remarkable
catalytic activity is realized by a novel reaction pathway, where
the generation of a dinitrogen-bridged dimolybdenum–iodide
complex is a key point to promote direct cleavage of nitrogen–
nitrogen triple bond of the bridging dinitrogen ligand in the
Mo–N≡N–Mo core. Further investigations are necessary to
Peters, Angew. Chem. Int. Ed. 2017, 56, 6921–6926.
4. Laplaza, C. C. Cummins, Science 1995, 268, 861–863.
5. a) C. Rebreyend, B. de Bruin, Angew. Chem. Int. Ed.
1
1
2
015, 54, 42-44. b) M. J. Bendek, P. J. Chirik, Angew.
Chem. Int. Ed. 2016, 55, 7892-7896.
1
6. For recent selected examples, see: a) J. S. Figueroa, N. A.
Piro, C. R. Clough, C. C. Cummins, J. Am. Chem. Soc.
2
006, 128, 940–950. b) H. Kunkely, A. Vogler, Angew.
Chem. Int. Ed. 2010, 49, 1591-1593. c) M. M. Rodriguez,
E. Bill, W. W. Brennessel, P. L. Holland, Science 2011,
334, 780-783. d) B. Askevold, J. T. Nieto, S. Tussupbayev,
M. Diefenbach, M. C. Holthausen, S. Schneider, Nat.
Chem. 2011, 3, 532–537. e) T. J. Hebden, R. R. Schrock,
M. K. Takase, P. Müller, Chem. Commun. 2012, 48,
elucidate the reaction mechanism.
We believe that the result
described in this paper provides a striking advance in the
research area of the development of catalytic nitrogen
3
7,38
fixation.
Acknowledgement
The present project is supported by CREST, JST
JPMJCR1541). We thank Grants-in-Aid for Scientific
1
851-1853. f) T. Shima, S. Hu, G. Luo, X. Kang, Y. Luo,
(
Z. Hou, Science, 2013, 340, 1549–1552. g) I. Klopsch, M.
Finger, C. Würtele, B. Milde, D. B. Werz, S. Schneider, J.
Am. Chem. Soc. 2014, 136, 6881–6883. h) A. J. Keane, W.
S. Farrell, B. L. Youke, P. Y. Zavalij, L. R. Sita, Angew.
Chem. Int. Ed. 2015, 54, 10220-10224. i) I. Klopsch, M.
Kinauer, M. Finger, C. Würtele, S. Schneider, Angew.
Chem. Int. Ed. 2016, 55, 4786-4789. j) M. M. Guru, T.
Shima, Z. Hou, Angew. Chem. Int. Ed. 2016, 55,
Research (Nos. JP17H01201, JP15H05798 to Y.N., Nos.
JP24109014, JP15K13710 to K.Y., and No. JP26888008 to
H.T.) from JSPS and MEXT.
A.E. is a recipient of the JSPS
Predoctoral Fellowships for Young Scientists.
Supporting Information
Experimental
data
are
available
on
http://dx.doi.org/10.000/bcsj000000.
1
2316-12320. k) F. Schendzielorz, M. Finger, C.
References
Volkmann, C. Würtele, S. Schneider, Angew. Chem. Int.
Ed. 2016, 55, 11417–11420. l) G. A. Silantyev, M. Fӧrster,
B. Schluschab, J. Abbenseth, C. Würtele, C. Volkmann, M.
C. Holthausen, S. Schneider, Angew. Chem. Int. Ed. 2017,
1
.
Ammonia synthesis catalysis: Innovation and Practice,
Ed: H. Liu, World Scientific, Beijing, 2013.
2
.
T. Spatzal, M. Aksoyoglu, L. Zhang, S. L. A. Andrade, E.
Schleicher, S. Weber, D. C. Rees, O. Einsle, Science 2011,
5
6, 5872–5876. m) Y. Nakanishi, Y. Ishida, H. Kawaguchi,
Angew. Chem. Int. Ed. 2017, 56, early view (DOI:
0.1002/anie.201704286).
3
34, 940.
3
4
.
.
K.M. Lancaster, M. Roemelt, P. Ettenhuber, Y. Hu, M. W.
Ribbe, F. Neese, U. Bergmann, S. DeBeer, Science 2011,
1
1
7. T. Miyazaki, H. Tanaka, Y. Tanabe, M. Yuki, K. Nakajima,
K. Yoshizawa, Y. Nishibayashi, Angew. Chem. Int. Ed.
3
34, 974–977.
a) Y. Tanabe, Y. Nishibayashi, Chem. Rec. 2016, 16,
549–1577. b) R. J. Burford, M. D. Fryzuk, Nat. Rev.
2
014, 53, 11488–11492.
1
1
1
8. Y. Nishibayashi, Inorg. Chem. 2015, 54, 9234–9247.
9. H. Tanaka, Y. Nishibayashi, K. Yoshizawa, Acc. Chem.
Res. 2016, 49, 987–995.
0. CCDC 1524460 (1a), 1524461 (2a), and 1524462 (2c)
contain the supplementary crystallographic data for this
Chem. 2017, 1, 0026. c) Nitrogen Fixation: Topics in
Organometallic Chemistry 60, Ed: Y. Nishibayashi,
Springer, 2017.
2
5
6
.
.
D. V. Yandulov, R. R. Schrock, Science 2003, 301, 76–78.
K. Arashiba, Y. Miyake, Y. Nishibayashi, Nat. Chem.
paper.
These data can be obtained free of charge from
2
011, 3, 120–125.
J. S. Anderson, J. Rittle, J. C. Peters, Nature 2013, 501,
4–87.
The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
7
8
.
.
8
2
2
1. M. W. Weatherburn, Anal. Chem. 1967, 39, 971-974.
S. Kuriyama, K. Arashiba, K. Nakajima, Y. Matsuo, H.
Tanaka, K. Ishii, K. Yoshizawa, Y. Nishibayashi, Nat.
Commun. 2016, 7, 12181.
15
1
2. The formation of NH3 was confirmed by H and
15
1
N{ H} NMR (see Supporting Information in details).
+
2
3. E1/2 Values of reductants in MeCN vs Ag/Ag are as
9
.
S. Kuriyama, K. Arashiba, H. Tanaka, Y. Matsuo, K.
Nakajima, K. Yoshizawa, Y. Nishibayashi, Angew. Chem.
Int. Ed. 2016, 55, 14291–14295.
follows; CoCp* = –1.78 V, CrCp* = –1.35 V, CoCp =
2
2
2
–
1.15 V.
M. P. Castellani, S. J. Geib, A. L. Rheingold,
W. C. Trogler, Organometallcs 1987, 6, 1703–1712.
1
1
1
0. P. J. Hill, L. R. Doyle, A. D. Crawford, W. K. Myers, A. E.
Ashley, J. Am. Chem. Soc. 2016, 138, 13521-13524.
1. R. R. Schrock, Angew. Chem. Int. Ed. 2008, 47,
2
2
4. pKa Value of proton sources in H O are as follows;
2
+
+
[ColH] = 7.48, [LutH] = 6.77.
K. Clarke, K.
Rothwell, J. Chem. Soc. 1960, 1885–1895.
5
512–5522.
5. Quite recently, Peters and co-workers have found the
efficient iron-catalyzed conversion of molecular
dinitrogen into ammonia, where the catalytic reaction was
2. H. Tanaka, K. Arashiba, S. Kuriyama, A. Sasada, K.
Nakajima, K. Yoshizawa, Y. Nishibayashi, Nat. Commun.
2
014, 5, 3737.
carried out at very low reaction temperature such as
1
3. a) T. J. D. Castillo, N. B. Thompson, J. C. Peters, J. Am.
Chem. Soc. 2016, 138, 5341–5350. b) M. J. Chalkey, T. J.
D. Castillo, B. D. Matson, J. C. Peters, ACS Cent. Sci.
13
–
78 °C.
However, simple comparison of between the
Iron-system and Mo-system is quite difficult from a
viewpoint of TON and TOF of the catalytic reaction
because of the different reaction conditions (room
2
017, 3, 217-223. c) T. M. Buscagan, P. H. Oyala, J. C.

temperature for Mo-system vs at -78 °C for Fe-system).
2017, early view (DOI: 10.1002/ejic.201700569).
2
2
6. K. Fagnou, M. Lautens, Angew. Chem. Int. Ed. 2012, 41,
2
6–47.
7. Unique effect of NaX (X = Cl, Br, and I) as additive on
the catalytic reaction using 3 as catalyst was observed in
the present reaction system (see Supporting Information
in details).
2
2
8. Q. Liao, A. Cavaillé, N. Saffon-Merceron, N. Mézailles,
Angew. Chem. Int. Ed. 2016, 55, 11212–11216.
9. K. Arashiba, E. Kinoshita, S. Kuriyama, A. Eizawa, K.
Nakajima, H. Tanaka, K. Yoshizawa, Y. Nishibayashi, J.
Am. Chem. Soc. 2015, 137, 5666–5669.
3
3
0. S. Kuriyama, K. Arashiba, K. Nakajima, H. Tanaka, N.
Kamaru, K. Yoshizawa, Y. Nishibayashi, J. Am. Chem.
Soc. 2014, 136, 9719–9731.
1. Quite recently, Peters and co-workers pointed out the
possibility of proton-coupled electron transfer (PCET)
process instead of proton transfer (PT)/electron transfer
1
3b
(ET) process in the catalytic reaction.
Based on the
proposal, we consider the possibility of PCET process in
our catalytic reaction using CoCp* as a reductant and
2
[
ColH]OTf as a proton source.
The formation of
4
+
endo-[Cp*Co( -C Me H)] is exergonic by 4.3 kcal/mol
in the protonation of CoCp*2 by [ColH] with an
activation energy of 22.3 kcal/mol. On the other hand,
the protonation of 2a by [ColH] is almost isoenergetic
G298 = +0.7 kcal/mol) with virtually no activation
5
5
+
+
barrier.
Therefore, in our catalytic system, the DFT
results suggest that the transformation of 2a to C is likely
to proceed via a stepwise process of PT and ET.
3
3
2. K. Arashiba, S. Kuriyama, K. Nakajima, Y. Nishibayashi,
Chem. Commun. 2013, 49, 11215–11217.
3. We have not completely excluded the possibility that
cleavage of nitrogen–nitrogen triple bond of a bridging
dinitrogen occurs at other oxidation states of the
3
4
molybdenum atom in the Mo–N≡N–Mo core. Further
work is necessary to elucidate detailed reaction
mechanism.
See detailed results of DFT calculations in
the Supporting Information.
3
3
3
4. We cannot consider the deprotonation of the CH groups
2
at the benzylic position of PNP ligand because the
catalytic reaction was carried out under acidic conditions.
5. A. Eizawa, K. Arashiba, H. Tanaka, S. Kuriyama, Y.
Matsuo, K. Nakajima, K. Yoshizawa, Y. Nishibayashi,
Nat. Commun. 2017, 8, 14874.
6. Separately, we confirmed the formation of 2a from the
reaction of 3 with 1 equiv of I2.
suggest that 2a is formed via B after the reaction of 3
with I2. See detailed experimental result in the
This result may
Supporting Information.
3
3
7. a) R. Araake, K. Sakadani, M. Tada, Y. Sakai, Y. Ohki, J.
Am. Chem. Soc. 2017, 139, 5596–5606. b) L. A.
Wickramasinghe, T. Ogawa, R. R. Schrock, P. Müller, J.
Am. Chem. Soc. 2017, 139, 9132-9135. c) D. E.
Prokopchuk, E. S. Wiedner, E. D. Walter, C. V. Popescu,
N. A. Piro, W. S. Kassel, R. M. Bullock, M. T. Mock, J.
Am. Chem. Soc. 2017, 139, 9291-9301.
8. a) R. Imayoshi, K. Nakajima, Y. Nishibayashi, Chem. Lett.
2
017, 45, 466–468. b) R. Imayoshi, K. Nakajima, J.
Takaya, N. Iwasawa, Y. Nishibayashi, Eur. J. Inorg. Chem.