Molybdenum-catalyzed transformation of molecular dinitrogen into silylamine: Experimental and DFT study on the remarkable role of ferrocenyldiphosphine ligands

Article abstract of DOI:10.1021/ja109181n

A molybdenum-dinitrogen complex bearing two ancillary ferrocenyldiphosphine ligands, trans-[Mo(N₂)₂(depf)₂] (depf = 1,1′-bis(diethylphosphino)ferrocene), catalyzes the conversion of molecular dinitrogen (N₂) into silylamine (N(SiMe₃) ₃), which can be readily converted into NH₃ by acid treatment. The conversion has been achieved in the presence of Me ₃SiCl and Na at room temperature with a turnover number (TON) of 226 for the N(SiMe₃)₃ generation for 200 h. This TON is significantly improved relative to those ever reported by Hidai's group for mononuclear molybdenum complexes having monophosphine coligands [J. Am. Chem. Soc.1989, 111, 1939]. Density functional theory (DFT) calculations have been performed to figure out the mechanism of the catalytic N₂ conversion. On the basis of some pieces of experimental information, SiMe₃ radical is assumed to serve as an active species in the catalytic cycle. Calculated results also support that SiMe₃ radical is capable of working as an active species. The formation of five-coordinate intermediates, in which one of the N₂ ligands or one of the Mo-P bonds is dissociated, is essential in an early stage of the N₂ conversion. The SiMe ₃ addition to a "hydrazido(2-)" intermediate having the NN(SiMe₃)₂ group will give a "hydrazido(1-)" intermediate having the (Me₃Si)NN(SiMe₃)₂ group rather than a pair of a nitrido (-N) intermediate and N(SiMe₃) ₃. The N(SiMe₃)₃ generation would not occur at the Mo center but proceed after the (Me₃Si)NN(SiMe₃) ₂ group is released from the Mo center. The flexibility of the Mo-P bond between Mo and depf would play a vital role in the high catalysis of the Mo-Fe complex.

Full text of DOI:10.1021/ja109181n

ARTICLE
pubs.acs.org/JACS
Molybdenum-Catalyzed Transformation of Molecular Dinitrogen into
Silylamine: Experimental and DFT Study on the Remarkable Role of
Ferrocenyldiphosphine Ligands
†
†
†
‡
‡
Hiromasa Tanaka, Akira Sasada, Tomohisa Kouno, Masahiro Yuki, Yoshihiro Miyake,
§
,‡
,†
Haruyuki Nakanishi, Yoshiaki Nishibayashi,* and Kazunari Yoshizawa*
†
Institute for Materials Chemistry and Engineering, Kyushu University, Nishi-ku, Fukuoka, Fukuoka 819-0395, Japan
‡
Institute of Engineering Innovation, School of Engineering, The University of Tokyo, Yayoi, Bunkyo-ku, Tokyo 113-8656, Japan
§
Fuel Cell System Development Center, Toyota Motor Corporation, Mishuku, Susono, Shizuoka 410-1193, Japan
S Supporting Information
b
ABSTRACT: A molybdenum-dinitrogen complex bearing two
ancillary ferrocenyldiphosphine ligands, trans-[Mo(N ) (depf) ]
2
2
2
0
(
depf = 1,1 -bis(diethylphosphino)ferrocene), catalyzes the con-
version of molecular dinitrogen (N ) into silylamine (N(SiMe ) ),
2
3 3
which can be readily converted into NH by acid treatment. The
3
conversion has been achieved in the presence of Me SiCl and Na at
3
room temperature with a turnover number (TON) of 226 for the
N(SiMe ) generation for 200 h. This TON is significantly improved
3
3
relative to those ever reported by Hidai's group for mononuclear
molybdenum complexes having monophosphine coligands [J. Am. Chem. Soc. 1989, 111, 1939]. Density functional theory (DFT)
calculations have been performed to figure out the mechanism of the catalytic N conversion. On the basis of some pieces of
2
experimentalinformation, SiMe radical isassumedtoserve asanactive species inthecatalytic cycle. Calculatedresultsalso supportthat
3
SiMe radical is capable of working as an active species. The formation of five-coordinate intermediates, in which one of the N ligands
3
2
or one of the Mo-P bonds is dissociated, is essential in an early stage of the N conversion. The SiMe addition to a “hydrazido(2-)”
2
3
intermediate having the NN(SiMe ) group will give a “hydrazido(1-)” intermediate having the (Me Si)NN(SiMe ) group rather
3
2
3
3 2
than a pair of a nitrido (tN) intermediate and N(SiMe ) . The N(SiMe ) generation would not occur at the Mo center but proceed
3
3
3 3
after the (Me Si)NN(SiMe ) group is released from the Mo center. The flexibility of the Mo-P bond between Mo and depf would
3
3 2
play a vital role in the high catalysis of the Mo-Fe complex.
’
INTRODUCTION
transformation of N into NH and/or its NH equivalent using
2 3 3
16
transition metal complexes under mild reaction conditions. In
1972, Shiina discovered the metal-chloride-catalyzed reductive
transformation of N into tris(trimethylsilyl)amine (N(SiMe ) ),
Nitrogen fixation under ambient reaction conditions is one of
1
the most important and challenging topics in chemistry. Mo-
lecular dinitrogen (N ) is chemically inert due to its extremely
strong, nonpolar triple bond (225 kcal/mol) as well as the large
HOMO-LUMO gap. Industrial nitrogen fixation typified by the
Haber-Bosch process requires drastic reaction conditions of
high pressures and high temperatures. Since the discovery of the
2
3 3
2
which can be readily converted into NH by acid treatment, in
3
1
7
the presence of Me SiCl and lithium. The turnover number
3
(
TON) of N(SiMe ) generation was up to 5.4 (CrCl ). Hidai,
3 3 3
Mizobe, and co-workers reported in 1989 a more effective
2
þ
method for N(SiMe ) generation using Me SiCl and sodium
3
3
3
first transition metal-N complex [Ru(N )(NH ) ] by Allen
2
2
3 5
2
together with molybdenum-N complexes such as [Mo(N ) -
2
2 2
and Senoff in 1965, a great deal of effort has been devoted to the
development of artificial nitrogen fixation systems that are
capable of working under mild reaction conditions. Nowadays,
18
(
PMe Ph) ] as a catalyst. In their reaction system, the TON
2 4
reached 24 based on the Mo atom. The mechanism of this
catalytic reaction still remains to be elucidated although they
proposed a mechanism involving a silyldiazenido intermediate
formed by the attack of SiMe radical on the N ligand. Recently,
a great number of well-defined N complexes are known for almost
2
all d-block transition metals as well as some f-block metals, and a lot
of studies have been reported so far on stoichiometric transforma-
tion of their coordinated N into ammonia (NH ) and hydrazine
3
2
Yandulov and Schrock reported the direct conversion of N into
2
2
3
3-14
NH
3
using a molybdenum-N complex bearing an ancillary tri-
2
(NH NH ).
In 1975, for example, Chatt and co-workers found
2
2
19
amidoamine ligand. They adopted a pair of proton and electron
the formation of NH by protonolysis of tungsten-N and
3
2
15
molybdenum-N complexes [M(N ) (PR ) ] (M = Mo, W).
2
2 2
3 4
In sharp contrast to such stoichiometric reactivity of transition-
Received: October 19, 2010
Published: February 22, 2011
metal-N complexes, there are only a few examples of the catalytic
2
r 2011 American Chemical Society
3498
dx.doi.org/10.1021/ja109181n
|J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
donors, lutidinium and decamethylchromocene, for the nitrogen
fixation and proposed a catalytic mechanism containing successive
experimental information that can be directly associated with the
determination of the catalytic mechanism. Computational quantum
chemistry will provide a valuable insight into our understanding.
Actually, some of us have computationally revealed the mechanisms
hydrogenation of N through alternating steps of protonation and
2
reduction. The hydrogenation of a Mo-NNH intermediate gives a
2
nitrido (tN) intermediate and the first molecule of NH , and then
on the activation and triple-bond cleavage of N by a cubane-type
3
2
28
29
the nitrido ligand on Mo is converted into the second molecule of
RuIr S cluster as well as a hydride-bridged diniobium complex.
3 4
NH . At present, the validity of this mechanism is strongly supported
In this Article, we describe typical results of the catalytic conversion of
3
by the isolation and observation of a large part of reactive inter-
mediates as well as intensive theoretical studies on the catalytic
N into N(SiMe ) using 1a as a catalyst and propose a possible
2
3 3
reaction pathway by quantum chemical calculations.
20-24
cycle.
very high (up to 8 based on the Mo atom).
Unfortunately, the TON for the NH generation was not
3
25
’ RESULTS AND DISCUSSION
Quite recently, some of us have reported the synthesis and the
Catalytic Conversion of N into N(SiMe ) by 1a. Typical
stoichiometric reactivity of molybdenum-N and tungsten-N
2
3 3
2
2
experimental results are summarized in Table 1 for the conversion
of N into N(SiMe ) by using 1a as a catalyst. Details are described
complexes bearing ferrocenyldiphosphines as auxiliary ligands,
0
trans-[M(N ) (depf) ] (M = Mo (1a), W (2a); depf = 1,1 -
2
3 3
2
2
2
in the Supporting Information. Treatment of Na (60 mmol) and
bis(diethylphosphino)ferrocene), where the electron transfer
process from the ferrocene moiety to the tungsten or molybde-
num center may be expected to assist the reduction of the
Me SiCl (60 mmol) in the presence of a catalytic amount of 1a
3
(
0.015 mmol) in THF (tetrahydrofuran; 40 mL) under an atmo-
26
spheric pressure of N at room temperature for 20 h gave 1.35
coordinated N into NH . During the continuous study on
2
2
3
mmol of N(SiMe ) together with some side products such as
the development of novel nitrogen fixation system under mild
3 3
n
27
Me SiSiMe , Me Si(CH ) OSiMe , and BuOSiMe (eq 1; Ta-
reaction conditions, a more efficient catalytic nitrogen fixation
system was found by using 1a as a catalyst. Dinitrogen under an
atmospheric pressure was catalytically converted into N(SiMe3)3
3
3
3
2 4
3
3
ble 1, run 1). The generated N(SiMe ) was isolated from the
3 3
1
13
reaction mixture and confirmed by H and C NMR and MS.
in the presence of Me SiCl and sodium in quite a high TON.
3
catalyst
ð0:015 mmolÞ
N2 þ reductant þ Me3SiCl
f
NðSiMe3Þ3 ð1Þ
ð1 atmÞ
ð60 mmolÞ
ð60 mmolÞ
THF; rt
The amounts of N(SiMe ) and other side-products were deter-
3
3
mined by GLC analysis. The TON for the formation of N(SiMe3)3
was 90 based on the Mo atom of 1a. A longer reaction time (100 h)
increased the amount of N(SiMe ) up to 1.77 mmol, corresponding
3
3
to a TON of 118 (Table 1, run 2). In the reaction with Li as a
reductant in place of Na (Table 1, run 3) for 100 h, a smaller amount
of N(SiMe ) (0.56 mmol, TON = 37) was formed together with a
3
3
larger amount of Me SiSiMe (16.53 mmol). The reaction in a more
3
3
dilute solution (60 mL of THF) yielded a larger amount of N-
(
SiMe ) (2.25 mmol, TON = 150; Table 1, run 4) than the result of
3
3
An intriguing subject is to elucidate how the coordinated N is
run 2 of Table 1. The TON reached 184 after 150 h, and Me SiCl was
2
3
converted in this multimetallic system and how the depf ligands
confirmed to be depleted in the reaction solution (Table 1, run 5).
enhanced the catalysis for the conversion of N into N(SiMe ) . For
Here, we considered that the presence of more Na and Me SiCl may
2
3 3
3
the present reaction system, unfortunately, we have not obtained
produce an extra amount of N(SiMe ) from the reaction mixture
3
3
Table 1. Reactions Using Reductant (60 mmol) and Me SiCl (60 mmol) Were Carried Out in the Presence of a Catalyst (0.015
3
mmol) under Atmospheric Pressure of N at Room Temperature in THF
2
a
b
run
catalyst
reductant
THF (mL)
time (h)
amount of N(SiMe
3
)
3
(mmol)
TON
1
2
3
4
1a
1a
1a
1a
1a
3
Na
Na
Li
40
40
40
60
60
40
40
40
40
40
40
40
40
20
100
100
100
200
20
1.35
1.77
0.55
2.25
3.39
0.48
0.81
0.01
0.90
0.24
0.18
0.04
0.66
90
118
37
Na
Na
Na
Na
Na
Na
Na
Na
Na
Na
150
226
32
c
5
6
7
8
9
1b
4
20
54
20
<1
60
2a
5
20
1
1
1
1
0
1
2
3
20
16
6
20
12
7
20
3
8
20
44
a
b
c
Determined by GLC. TON = N(SiMe
3
)
3
/catalyst. Another THF solution (60 mL) of Me
3
SiCl (60 mmol) and Na (60 mmol) was added after 100 h.
3
499
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
when the catalytic ability of 1a is still alive at this stage. In fact, addition
N(SiMe ) catalyzed 1a. We briefly discuss the active silyl species
in the reaction solution and then propose a plausible mechanism
3
3
of Na (60 mmol) and Me SiCl (60 mmol) to the reaction solution
3
after 100 h produced an extra amount of N(SiMe ) from the
solution used in reaction mixture. The total amount of N(SiMe3)3
was finally measured to be 3.39 mmol with a TON of 226 (Table 1,
of the catalytic N conversion. All calculations were carried out by
3
3
2
3
0
using the Gaussian09 program package. All intermediates
proposed in this study were searched on potential energy surfaces
3
1
run 5). Both N gas and 1a were confirmed to be indispensable for
by using the hybrid density functional B3LYP method. For
optimization, the LANL2DZ and 6-31G(d) basis sets were chosen
for the metal atoms (Mo, Fe) and the other atoms, respectively
2
the formation of N(SiMe ) (see the Supporting Information).
3
3
Under the reaction conditions similar to run 1 in Table 1,
32
replacement of 1a with trans-[Mo(N ) (depf)(PMePh )] (1b)
(BS1). Optimized structures were confirmed to have no ima-
ginary frequencies by vibrational analyses. Ancillary depf ligands
were included without any simplification. The total charge of 1a is
equal to 0, and thus, the formal charges of Mo and Fe in 1a are 0
and þ2, respectively. To determine the energy profile of the
proposed catalytic cycle, we performed single-point calculations at
the optimized geometries using the 6-311þG(d) basis set instead
of the 6-31G(d) basis set (BS2). Solvation effects (THF) were
taken into account by using the polarizable continuum model
2
2
2
and cis-[Mo(N ) (PMe Ph) ] (3) decreased the TONs to 54
2
2
2
4
and 32, respectively (Table 1, runs 6 and 7). Previously, Hidai’s
group reported that cis-[W(N ) (PMe Ph) ] (4) did not work as
2
2
2
4
1
8
a catalyst for the formation of N(SiMe ) . We also confirmed
3
3
that 4 did not catalyze this reaction under the same reaction
conditions (Table 1, run 8). Interestingly, the reaction in the
presence of a catalytic amount of trans-[W(N ) (depf) ] (2a)
2
2
2
gave N(SiMe ) , where the TON reached 60 (Table 1, run 9). In
3
3
33
sharp contrast to the efficient catalysis of 1a and 2a, low catalysis
was observed for molybdenum-N and tungsten-N com-
(PCM). Zero-point energy corrections were applied for enthal-
py changes (ΔH ) and activation energies (E ) calculated for each
2
2
0
a
plexes having other diphosphines as ancillary ligands such as
reaction step.
trans-[Mo(N )(dppe) ] (dppe = 1,2-bis(diphenylphosphino)
1. Active Silyl Species. As mentioned above, SiMe radical is a
2
2
3
ethane; 5), trans-[Mo(N ) (depe) ] (depe = 1,2-bis(diethylphos-
possible active silyl species in the present reaction system. Here
we would like to deduce the active silyl species based on
theoretical calculations. If the active silyl species is directly
2
2
2
0
phino)ethane; 6), and trans-[W(N ) (depr) ] (depr = 1,1 -bis-
2
2
2
(diethylphosphino)ruthenocene; 7). The TONs for the N(SiMe3)3
formation were measured to be 16 for 5, 12for6, and3for7(Table 1,
runs 10-12). All the experimental results clearly show that the depf
ligands in 1a and 2a are responsible for their efficient catalysis.
Unfortunately, we were not able to detect any structurally defined
complexes by NMR after the catalytic reaction. As a piece of experi-
mental information leading to elucidation of the catalytic mechanism,
it should be noted that N(SiMe ) was not observed at all when dry
generated by the Si-Cl bond cleavage of Me SiCl, the hetero-
3
þ
3
-
lysis of Me SiCl gives SiMe and Cl (eq 4), while the
3
homolysis gives a pair of SiMe and Cl radicals (eq 5):
3
þ
-
Me3SiCl f Me3Si þ Cl
ð4Þ
ð5Þ
Me3SiCl f Me3Si• þ Cl•
3
3
air was used in place of N . This would imply that O in the air
2
2
inhibited the generation of radical species such as SiMe radical.
The Si-Cl bond energy in THF is calculated to be 70.2 kcal/mol
for the heterolysis and 107.3 kcal/mol for the homolysis at the
B3LYP/6-311þG(d) level of theory. The large Si-Cl bond
energies for both the homolysis and heterolysis indicate that a
3
Previously, Hidai's group reported the synthesis of a series of silyl-
diazenido complexes by the reaction of Mo and W dinitrogen com-
plexes such as 3, 4, and 5 with Me SiI, Me SiCl/NaI, or Me SiOTf. It
3
3
3
was further revealed that reduction of trans-[MI(NNSiMe )(PMe -
neutral Me
SiMe radical at room temperature. We can therefore rule out a
3
SiCl molecule would give neither SiMe cation nor
3
3
2
Ph) ] (M = Mo, W) with Na in the presence of Me SiCl under Ar
3
4
3
1
8,34
affords N(SiMe ) as a principal nitrogenous product.
We investi-
reaction mechanism containing “alternative protonation/reduc-
tion steps” from consideration, which is commonly accepted for
3
3
gated the stoichiometric reaction of 1a with 1 equiv of Me SiOTf
3
1
9-24
(
OTf = OSO CF ) at room temperature for 1 h, which gave the
nitrogen fixation catalyzed by Schrock’s Mo complex.
Here we consider the role of sodium in the reaction mixture. If
sodium in the reaction solution reduces Me SiCl in advance, the
generated Me SiCl anion is readily cleaved into SiMe
2
3
corresponding silyldiazenido complex (8) in 69% isolated yield (eq 2).
When 8 was used as a catalyst in place of 1a, the TON for N(SiMe3)3
was 44 (Table 1, run 13). In addition, the stoichiometric reaction of 8
3
radical
3
3
with Na (20 equiv) and Me SiCl (20 equiv) under an atmospheric
and Cl anion (eqs 6 and 7):
3
pressure of argon afforded N(SiMe ) in 88% yield based on 8 (eq 3).
These results suggest that the diazenido complex 8can be regarded as a
3
3
reduction
Me SiCl
f
Me SiCl-
ð6Þ
ð7Þ
3
3
reactive intermediate in the catalytic formation of N(SiMe ) .
3
3
Me3SiCl- f Me3Si• þ Cl-
From the difference in energy between the optimized structures
-
of Me SiCl and Me SiCl , the adiabatic electron affinity of
3
3
Me SiCl is calculated to be -13.2 kcal/mol. The Si-Cl bond
3
-
cleavage of Me SiCl in THF is exergonic by 7.3 kcal/mol. These
3
values are in good agreement with the experimental fact that
Me SiCl molecules slowly generate disilane Me SiSiMe in the
3
3
3
presence of Na in THF. From both the experimental and
theoretical results, the following discussion focuses on a radical
8
þ
Na þ Me3SiCl
f
NðSiMe3Þ3
ð3Þ
ð20 equivÞ
THF; rt; 20 h
Ar ð1 atmÞ
mechanism in which SiMe radical plays a role in the catalysis.
ð20 equivÞ
88% yield
3
. Catalytic Mechanism for the Dinitrogen-to-Silylamine
2
Theoretical Calculations. In this section, we describe a
Conversion. We present in Figure 1 a plausible mechanism on
theoretical study on a mechanism of the conversion of N into
the conversion of N into N(SiMe ) via successive addition of
2
2
3 3
3
500
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
Figure 1. Possible mechanism of the catalytic conversion of N
2
into N(SiMe
3
)
3
by a molybdenum complex bearing ferrocenyldiphosphines. (a) Three
pathways to yield the four-coordinate intermediate IVb. (b) The formation of (Me Si) NN(SiMe ) anion and the start of the next catalytic cycle. SiMe
3
2
3
3
represents trimethylsilyl radical.
SiMe radicals, which is partially relevant to Hidai and Mizobe’s
bond cleavage of VI result in the formation of [Mo(depf) ] VII
3
2
34
proposal for cis-[Mo(N ) (PMe Ph) ] 3. In their catalytic
and (Me Si) NN(SiMe ) anion. This anion liberated into the
2
2
2
4
3
2
3
mechanism, a five-coordinate intermediate is first formed by
reaction solution would be readily converted into two mole-
the dissociation of one N ligand in 3. Figure 2 shows optimized
cules of N(SiMe ) by Me SiCl and SiMe radicals. The next
2
3 3
3
3
structures of 1a and intermediates II-VIII proposed in the present
calculation. Figure 3 shows the overall energy profile containing the
enthalpy change and activation energy (in parentheses) for each
reaction step. Detailed information on the reaction steps, such as
optimized structures of reactant complexes, transition states, and
product complexes, are described in the Supporting Information.
The conversion starts with the addition of SiMe to one of the N
cycle of the conversion of N starts with the binding of a
2
new N molecule by VII. The addition of SiMe radical to
2
3
[Mo(N )(depf) ] VIII gives IIIa and the following reactions
2
2
proceed along path A or B.
3. Addition of first SiMe Radical and Formation of Five-
Coordinate Intermediates. As shown in Figures 1 and 3, the
3
catalytic cycle starts with the addition of SiMe radical to the
3
2
3
ligands in 1a to yield [Mo(N )(NNSiMe )(depf) ] IIa. There are
initial complex 1a, which is exergonic by 2.6 kcal/mol in THF.
The calculated activation barrier (8.3 kcal/mol) is low enough to
overcome at room temperature. As mentioned above, there are
three possible reaction pathways for the formation of five-
coordinate intermediates from the generated intermediate IIa:
2
3
2
three possible pathways (paths A-C) leading to the doubly
silylated intermediate IVb having the NN(SiMe ) group. In paths
3
2
A and B, IIa releases the N ligand to form a five-coordinate
2
intermediate [Mo(NNSiMe )(depf) ] IIIa. In path C, on the
3
2
other hand, one of the four Mo-P bonds in IIa is cleaved to form
the Mo-N bond dissociation leading to IIIa (paths A and B)
2
another five-coordinate intermediate IIb. In path A, the second
and the Mo-P bond dissociation leading to IIb (path C). As
SiMe radical attacks IIIa to yield a five-coordinate intermediate
shown in Figure 3, the enthalpy change for the Mo-N bond
3
2
IVa having the NN(SiMe ) group, and then one of the Mo-P
dissociation of IIa (IIa f IIIa) is calculated to be þ2.4 kcal/mol
3
2
bonds in IVa is cleaved to give a four-coordinate intermediate IVb.
InpathB, one ofthe Mo-P bonds in IIIa is cleaved in advance, and
then the generated four-coordinate intermediate IIIb is attacked by
(E = 13.7 kcal/mol) in THF, which is much smaller than that for
a
the Mo-N bond dissociation of 1a (þ18.4 kcal/mol; E =
2
a
20.0 kcal/mol). These values suggest that the silylation of one
N ligand prompts the dissociation of the other N ligand. In
the second SiMe radical for the formation of IVb. In path C, the
3
2
2
five-coordinate IIb accepts the second SiMe radical to form
the Mo-depf system, a five-coordinate intermediate will be
formed after the silylation of the initial complex, while the
catalytic mechanism proposed for cis-[Mo(N ) (PMe Ph) ] 3
3
[
Mo(N )(NN(SiMe ) )(depf) ] V. The N ligand in V is then
2 3 2 2 2
dissociated to give IVb. The third SiMe radical attacks the N
3
2 2
2
4
atom adjacent to the Mo atom of IVb, leading to VI having the
Me SiNN(SiMe ) group. One-electron reduction and the Mo-N
supposed that one N ligand dissociates from 3 before the
2
20
addition of SiMe3.
3
3 2
3
501
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
Figure 2. Optimized structures of 1a and possible intermediates II-VIII. Hydrogenatoms are omittedfor clarity. Interatomic distances are presented inÅ.
Figure 4 presents computed energy changes of 1a and IIa
plotted as a function of one of the Mo-P bond distances r.
Optimization at each fixed r was carried out at the B3LYP/BS1
level of theory, and the total energy in THF was obtained with a
single-point calculation at the B3LYP/BS2 level of theory. For
comparison, we show in this figure energy changes calculated for
trans-[Mo(N ) (dppe) ] 5 having bidentate phosphine coli-
IIa, the five-coordinate structure of 9 (Mo-P = 5.885 Å) is
1.8 kcal/mol less stable than the six-coordinate one and E (8.4
a
kcal/mol) is much higher than that of IIa. If the catalytic
formation of N(SiMe ) by 5 proceeds in a similar manner to
3
3
1a, high catalytic ability of 1a relative to 5 may be explained by
facile Mo-P bond cleavage in IIa as well as the higher thermo-
dynamic stability of the five-coordinate IIb than the six-coordi-
2
2
2
gands and its silylated complex trans-[Mo(N )(NNSiMe )-
nate IIa. This is because one of the Mo-P bonds or the Mo-N
2
3
2
(
dppe) ] 9. The Mo-P bond cleavage of 1a requires E of 9.8
bond must be dissociated before accepting the second SiMe
2
a
3
kcal/mol, and the corresponding five-coordinate structure
radical, as mentioned later in subsections 4 and 5.
(
Mo-P = 6.371 Å) is 6.0 kcal/mol less stable than 1a. The
For the following reaction steps leading to IVb having the
activation energy for the Mo-P cleavage of 5 is calculated to be
much higher (18.7 kcal/mol), and the five-coordinate structure
NN(SiMe ) group, we consider both pathways that start from
3
2
IIIa and IIb. The addition of SiMe to IIa to optimize a “diazene”
3
(
Mo-P = 5.205 Å) is 16.8 kcal/mol less stable than the six-
intermediate having the Me SiNdNSiMe3 group gave the
3
coordinate structure. On the other hand, the silylation of 1a
makes a five-coordinate structure energetically more favorable,
corresponding intermediate having the NN(SiMe ) group.
3
4. Formation of Four-Coordinate IVb via IIIa (Paths A
2
probably due to the steric repulsion between the SiMe group
and the bulky depf ligand; the five-coordinate intermediate IIb
and B). An intermediate having the NN(SiMe ) group must
3
3 2
dissociate one of the Mo-P bonds to accept the third SiMe radical,
3
(
Mo-P = 6.876 Å) is 8.5 kcal/mol more stable than IIa, and Ea
and therefore, intermediate IVb has a four-coordinate structure.
for the Mo-P bond cleavage is only 2.5 kcal/mol. Contrary to
As shown in Figure 1, two reaction pathways are possible to lead
3
502
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
Figure 3. Energy profile of the catalytic conversion of N
2
into N(SiMe
3
)
3
by 1a. [Mo] and SiMe
3
represent [Mo(depf)
2
] and trimethylsilyl radical,
respectively. The enthalpy change ΔH and activation energy (E , in parentheses) for each reaction step were calculated at the B3LYP/BS2 level of
0
a
theory, and solvation effects (THF as a solvent) were taken into account. The energies are presented in kcal/mol.
Figure 4. Energy changes for a Mo-P bond cleavage in (a) 1a and 5 and (b) their silylated complexes IIa and 9. Units are presented in kcal/mol.
to IVb from IIIa: The addition of SiMe to IIIa is followed by
NN(SiMe ) group readily dissociates one Mo-P bond to give
3
3 2
the Mo-P bond cleavage (path A; IIIa f IVa f IVb). In the
the four-coordinate IVb (ΔH = -14.5 kcal/mol, E = 0.1 kcal/mol).
0
a
other pathway, one of the Mo-P bonds in IIIa is dissociated
In path B, the Mo-P bond cleavage of IIIa proceeds in an
endergonic way by 3.0 kcal/mol and is hampered by a low
activation barrier (4.5 kcal/mol). Instead, the addition of the
in advance, and then SiMe radical attacks to form IVb (path B;
3
IIIa f IIIb f IVb). As shown in Figure 3, the addition
of SiMe to IIIa yielding IVa is calculated to be an exergonic
second SiMe radical to IIIb yielding IVb is a highly exergonic
3
3
reaction (ΔH0 = -10.6 kcal/mol) with a moderate E
reaction(ΔH = -27.1kcal/mol) with E of9.3 kcal/mol, whichis
a
0
a
(15.5 kcal/mol). The five-coordinate IVa having the bulky
lower than that for IIIa f IVa. The calculated ΔH and E values
0 a
3
503
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
Figure 5. Energy diagrams for the addition of SiMe
3
radical to IVb and V. Interatomic distances and relative energies are presented in Å and kcal/mol,
respectively.
suggest that the formation of IVb undergoes via both
the pathways.
the Mo center, a “hydrazido(1-)” intermediate having the
Me SiNN(SiMe ) group is formed. On the other hand, the
3
3 2
The experimental result that trans-[Mo(NNSiMe )(OTf)-
silylation of the doubly silylated N atom will give a “hydrazidium”
3
(
depf) ] 8 served as a catalyst for the formation of N(SiMe )
intermediate having the NN(SiMe ) group or cause the N-N
2
3 3
3 3
implies that a Mo-Cl species such as trans-[Mo(NNSiMe )-
bond dissociation to generate a N(SiMe ) molecule and a
3
3 3
(
Cl)(depf) ] could be considered as an intermediate, although
nitrido (tN) intermediate. Figure 5 displays energy diagrams for
2
such a silyldiazenido intermediate was not observed. If this
intermediate is reduced by Na, IIIa will be formed via the elimi-
nation of Cl anion. The enthalpy change for reaction [Mo(NN-
the SiMe addition to IVb and V. The silyl addition to the N atom
3
adjacent to the Mo center in IVb yields VI via TSIVb-VI. It is
noteworthy that TSIVb-VI and VI adopt unique bent Mo-N-N
linkages whose angles are 127.0° and 106.2°, respectively. The
Mo-N and N-N bond distances in VI are calculated to be 2.081
and 1.512 Å, respectively, implying that the Me SiNN(SiMe )
-
-
SiMe )(Cl)(dppe) ] f IIIa þ Cl is calculated to be -16.6
3
2
kcal/mol in THF at the B3LYP/BS2//B3LYP/BS1 level of
theory. The calculated result also suggests a possibility that
trans-[Mo(NNSiMe )(Cl)(depf) ] was temporally formed in
3
3 2
group is still bound to the Mo center; the formation of VI is
3
2
the reaction solution.
exergonic by 10.9 kcal/mol with E of 21.6 kcal/mol.
a
5
. Formation of Four-Coordinate Intermediate IVb via IIb
The silyl addition to the doubly silylated N atom in IVb causes
a spontaneous cleavage of the N-N bond due to the bulkiness of
the generated N(SiMe ) moiety and then gives the first mole-
(
Path C). Another pathway leading to IVb starts from the five-
coordinate intermediate IIb, in which the Mo center has one
3
3
Mo-N bond with the N ligand and three Mo-P bonds with
cule of N(SiMe ) and the nitrido intermediate IX. Although this
2
3 3
two depf ligands (Figure 1). As shown in Figure 3, the addition of
reaction is highly exergonic (ΔH = -89.6 kcal/mol) relative to
0
the second SiMe radical to IIb gives intermediate V having the
the formation of VI, the activation barrier of 28.7 kcal/mol seems
too high for a reaction occurring at room temperature. Inter-
mediate V also undergoes a spontaneous N-N bond dissociation
by the addition of SiMe . The formation of X and N(SiMe ) is
3
NN(SiMe ) group, which is exergonic by 25.8 kcal/mol (E =
3
2
a
9
.0 kcal/mol). All attempts to optimize a six-coordinate structure
of V having four Mo-P bonds resulted in failure. This is a reason
why we considered the formation of the five-coordinate inter-
3
3 3
highly exergonic by 81.9 kcal/mol, but the activation barrier of
42.2 kcal/mol is too high to overcome at room temperature,
suggesting that a five-coordinate structure is not suitable for the
mediates IIIa and IIb. The dissociation of N in V to give IVb
2
proceeds in an endergonic way (ΔH = þ16.4 kcal/mol) and has
0
a relatively high activation barrier (21.6 kcal/mol). Although this
reaction step seems undesirable from both thermodynamic and
third SiMe addition. We did not obtain any intermediate or
3
transition state having the Me SiNN(SiMe ) group for the
3
3 2
kinetic points of view, addition of the third SiMe radical to the
Mo-Fe system does not occur at room temperature without
considering this process, as described later in subsection 6.
addition of SiMe to V.
Intermediate VI is not able to accept any more SiMe radical,
3
3
3
and thus, the Mo-N bond is cleaved to release the Me SiNN-
3
6
. Addition of Third SiMe Radical. As shown in Figure 1,
(SiMe ) group. Elongation of the Mo-N bond by partial
3
3 2
-
intermediates IVb and V have two N atoms that can be attacked
optimizations invoked heterolysis to form Me SiNN(SiMe )
.
3
3 2
by SiMe radical. If SiMe radical attacks the N atom adjacent to
In fact, the enthalpy change for the homolysis of the Mo-N
3
3
3
504
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
ARTICLE
bond in VI (VI f VII þ Me SiNN(SiMe ) ) is þ3.9 kcal/mol
group is released from the Mo complex as Me SiNN(SiMe )
anion and the anion in the reaction solution is readily converted
3
3 2
3
3 2
(
endergonic) in THF at the B3LYP/BS2 level of theory, while
þ
that for the heterolysis of VI (VI f VII þ Me SiNN-
into two molecules of N(SiMe ) by Me SiCl and SiMe radicals.
3
3 2
3
3
(
SiMe3)2-) is -52.9 kcal/mol (highly exergonic). The cationic
To afford the Me SiNN(SiMe ) anion, the Mo-Fe system
3
3 2
VII would be readily neutralized by sodium in the reaction
solution because the neutral VII is 8.3 kcal/mol energetically
more favorable than the cation. There is another reaction path-
way that VI is reduced by sodium in advance and then the anionic
should be reduced by sodium before or after the Mo-N bond
cleavage. Intermediate VII traps a single N molecule for starting
2
the next cycle of the N(SiMe ) generation. The catalytic
3
3
mechanism proposed in this study is different from the Yandulov-
Schrock mechanism, in which ammonia is generated step-by-
step on their Mo complex. The catalysis of 1a would be sterically
controlled by the depf ligands through not only its bulkiness but
also the flexible Mo-P bond that allows the Mo center to adopt
variouscoordination numbers. Althoughweexpected that the depf
ligands contribute to the high catalysis electronically through an
electron transfer from the ferrocene moiety to the Mo center, such
theoretical suggestions have not been obtained in the framework
of theradical mechanism. For all the intermediates, theFeatomsin
the depf ligands have the formal charge of þ2 in their electronic
ground states, and the spin state of the ferrocene moieties is always
closed-shell singlet. Nevertheless, both the experimental and
-
VI releases Me SiNN(SiMe ) . The anionic VI is 45.5 kcal/mol
3
3 2
lower in energy than the neutral VI, and the Mo-N bond cleavage
of the anionic VI is exergonic by 15.7 kcal/mol. In total, the one-
electron reduction and Mo-N bond cleavage of VI to yield VII
-
and Me SiNN(SiMe ) proceed in an exergonic way by 61.2
3
3 2
-
kcal/mol. Me SiNN(SiMe ) liberated into the reaction solution
3
3 2
will be converted into two molecules of N(SiMe ) by Me SiCl
3
3
3
and SiMe radicals; the enthalpy change for reaction Me SiNN-
3
3
-
-
(
SiMe3)2 þ Me SiCl þ 2SiMe f 2N(SiMe ) þ Cl is calcu-
3
3
3 3
lated to be -163.8 kcal/mol in THF. The calculated results
suggest that 1a mediates the conversion of N into N(SiMe )
2
3 3
in a different manner from the Yandulov-Schrock mecha-
nism containing a stepwise generation of NH via a nitrido
theoretical results provide convincing evidence that SiMe radical
3
3
1
9
intermediate.
. Starting the Next Cycle of N Conversion. As shown in
plays a key role in the present reaction system. SiMe cation would
3
7
not work as an active silyl species because of the presence of a
2
Figures 1 and 3, Intermediate VII traps a new N molecule at its
strong reductant (Na) as well as the highly endergonic nature of
2
Mo center to initiate the next cycle of the N conversion. The
the heterolysis of Me SiCl. We therefore conclude that the radical
2
3
binding of N proceeds in an exergonic way (ΔH = -39.3 kcal/
mechanism presented in this report is highly possible for the
2
0
mol) as virtually a barrierless reaction (E = 0.6 kcal/mol). The
catalytic N conversion by1a. This conclusion doesnot necessarily
a
2
generated intermediate VIII is silylated to form intermediate IIIa
exclude different reaction mechanisms. To reconsider the electro-
nic contribution of the depf ligand to the catalytic ability of 1a, we
are currently investigating a mechanism in which Mo complexes
(
ΔH = -34.9 kcal/mol, E = 0.3 kcal/mol), and the following
0
a
reaction steps proceed along path A or B. In conclusion, the rate-
determining step in all the reaction steps examined here is the
addition of the third Si radical to IVb yielding intermediate VI
are reduced by Na in place of Me SiCl and the anionic Mo species
3
react with the neutral Me SiCl .
3
(
Ea = 21.6 kcal/mol).
’
CONCLUSIONS
’ ASSOCIATED CONTENT
A molybdenum-N complex bearing ancillary ferrocenyldi-
2
S
Supporting Information. Details on experiments, com-
b
phosphine ligands, trans-[Mo(N ) (depf) ] 1a, has catalyzed the
2
2
2
puted energetics, and structural changes in the proposed reaction
steps, Cartesian coordinates for all optimized intermediates, and
the complete author list of ref 30. This material is available free of
charge via the Internet at http://pubs.acs.org.
conversion of N into N(SiMe ) in the presence of Me SiCl
2
3 3
3
and Na at room temperature. The turnover number for the
N(SiMe ) generation finally reached 226 for 200 h, which is
3
3
greatly improved relative to the precedent molybdenum and
tungsten complexes having mono- and diphosphine ligands, such
18
’ AUTHOR INFORMATION
as [M(N ) (PMe Ph) ] and [M(N ) (dppe) ] (M = Mo, W).
2
2
2
4
2 2
2
This is so far the most efficient conversion of N into NH3
equivalent under ambient conditions. With an assumption that
Corresponding Author
kazunari@ms.ifoc.kyushu-u.ac.jp; ynishiba@sogo.t.u-tokyo.ac.jp
2
SiMe radical works as an active silyl species in the catalytic cycle,
3
we have shown a plausible mechanism of the N conversion
2
’
ACKNOWLEDGMENT
by DFT calculations and determined the energy profile of the
catalytic cycle from both thermodynamic and kinetic points of
view. Almost all of the reaction steps proceed in an exergonic way
with reasonably low activation barriers, even the first silylation of
K.Y. acknowledges Grants-in-Aid (Nos. 18066013, 18GS0-
207, and 22245028) for Scientific Research from Japan
Society for the Promotion of Science (JSPS) and the Ministry
of Education, Culture, Sports, Science and Technology of
Japan (MEXT), the Nanotechnology Support Project of
MEXT, the MEXT Project of Integrated Research on Chem-
ical Synthesis, and the Kyushu University Global COE
Project for their support of this work. Y.N. acknowledges
Grants-in-Aid for Scientific Research for Young Scientists (S)
(No. 19675002) and for Scientific Research on Priority Areas
(No. 18066003) from MEXT. This paper is dedicated to the
late Professor Yasushi Mizobe, Institute of Industrial Science,
The University of Tokyo.
N . It is notable that the formation of five-coordinate intermedi-
2
ates is essential in an early stage of the conversion. The first
silylation of one N ligand in 1a prompts bond dissociation
2
between Mo and N as well as Mo and P, while these bonds in 1a
2
are strong enough to adopt a six-coordinate structure. For the
silylation of intermediate IVb having the NN(SiMe ) group,
3
2
SiMe radical will not attack the N atom having two SiMe group
3
3
to form a pair of a nitrido intermediate and N(SiMe ) but attack
3
3
the N atom adjacent to the Mo center to give intermediate VI
having the Me SiNN(SiMe ) group. The Me SiNN(SiMe )
3
3 2
3
3 2
3
505
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506

Journal of the American Chemical Society
REFERENCES
ARTICLE
’
(25) (a) Schrock, R. R. Acc. Chem. Res. 2005, 38, 955. (b) Schrock,
R. R. Angew. Chem., Int. Ed. 2008, 47, 5512.
(1) (a) Ertl, G. In Catalytic Ammonia Synthesis; Jennings, J. R., Ed.;
(
26) (a) Yuki, M.; Miyake, Y.; Nishibayashi, Y.; Wakiji, I.; Hidai, M.
Plenum Press: New York, 1991 and references therein. (b) Appl, M.
Ammonia; Wiley-VCH: Weinheim, 1999 and references therein.
Organometallics 2008, 27, 3947. (b) Yuki, M.; Midorikawa, T.; Miyake,
Y.; Nishibayashi, Y. Organometallics 2009, 28, 4741. (c) Yuki, M.;
Miyake, Y.; Nishibayashi, Y. Organometallics 2009, 28, 5821.
27) (a) Nishibayashi, Y.; Iwai, S.; Hidai, M. Science 1998, 279, 506.
b) Nishibayashi, Y.; Takemoto, S.; Iwai, S.; Hidai, M. Inorg. Chem. 2000,
9, 5946. (c) Nishibayashi, Y; Saito, M.; Uemura, S.; Takekuma, S.;
Takekuma, H.; Yoshida, Z. Nature 2004, 428, 279.
28) (a) Tanaka, H.; Mori, H.; Seino, H.; Hidai, M.; Mizobe, Y.;
(2) (a) Allen, A. D.; Senoff, C. V. Chem. Commun. 1965, 621. (b)
Bottomley, F.; Nyburg, S. C. Chem. Commun. 1966, 897. (c) Senoff,
C. V. J. Chem. Educ. 1990, 67, 368.
(
(
3
(
3) (a) Hidai, M.; Mizobe, Y. Chem. Rev. 1995, 95, 1115. (b) Hidai,
M.; Mizobe, Y. Can. J. Chem. 2005, 83, 358.
4) (a) Gambarotta, S. J. Organomet. Chem. 1995, 500, 117. (b)
Gambarotta, S.; Scott, J. Angew. Chem., Int. Ed. 2004, 43, 5298.
5) (a) Fryzuk, M. D.; Johnson, S. A. Coord. Chem. Rev. 2000, 200-
02, 379. (b) MacKay, B. A.; Fryzuk, M. D. Chem. Rev. 2004, 104, 385.
c) Fryzuk, M. D. Acc. Chem. Res. 2009, 42, 127. (d) Ballmann, J.;
Munh ꢀa m, R. F.; Fryzuk, N. D. Chem. Commun. 2010, 46, 1013.
(
(
Yoshizawa, K. J. Am. Chem. Soc. 2008, 130, 9037. (b) Tanaka, H.;
Ohsako, F.; Seino, H.; Mizobe, Y.; Yoshizawa, K. Inorg. Chem. 2010, 49,
(
2
(
2
464.
29) Tanaka, H.; Shiota, Y.; Matsuo, T.; Kawaguchi, H.; Yoshizawa,
K. Inorg. Chem. 2009, 48, 3875.
30) Frisch, M. J.; et al. Gaussian 09, revision A.01; Gaussian, Inc.:
Wallingford CT, 2009.
31) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (b) Becke, A. D. J.
Chem. Phys. 1993, 98, 5648. (c) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B
988, 37, 785. (d) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.;
Frisch, M. J. J. Phys. Chem. 1994, 98, 11623.
32) (a) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys.
971, 54, 724. (b) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem.
Phys. 1972, 56, 2257. (c) Hariharan, P. C.; Pople, J. A. Theor. Chim.
Acta 1973, 28, 213. (d) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.;
Schleyer, P. v. R. J. Comput. Chem. 1983, 4, 294. (e) Francl, M. M.;
Petro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S.; DeFrees, D. J.;
Pople, J. A. J. Chem. Phys. 1982, 77, 3654. (f) Hay, P. J.; Wadt, W. R. J.
Chem. Phys. 1985, 82, 299.
(
(
(
(
6) Himmel, H.-J.; Reiher, M. Angew. Chem., Int. Ed. 2006, 45, 6264.
7) Chirik, P. J. J. Chem. Soc., Dalton Trans. 2007, 16.
8) (a) Pool, J. A.; Lobkovsky, E.; Chirik, P. J. Nature 2004, 427, 527.
(
(
(
b) Bart, S. C.; Chzopek, K.; Bill, E.; Bouwkamp, M. W.; Lobkovsky, E.;
Neese, F.; Wieghardt, K.; Chirik, P. J. J. Am. Chem. Soc. 2006, 128, 13901.
c) Bart, S. C.; Bowman, A. C.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem.
1
(
Soc. 2007, 129, 7212. (d) Knobloch, D. J.; Toomey, H. E.; Chirik, P. J.
J. Am. Chem. Soc. 2008, 130, 4248. (e) Pun, D.; Bradley, C. A.; Lobkovsky,
E.; Keresztes, I.; Chirik, P. J. J. Am. Chem. Soc. 2008, 130, 14046. (f) Pun,
D.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2008, 130, 6047.
(
1
(
g) Knobloch, D. J.; Lobkovsky, E.; Chirik, P. J. Nat. Chem. 2010, 2, 30.
9) Gilbertson, J. D.; Szymczak, N. K.; Tyler, D. R. J. Am. Chem. Soc.
005, 127, 10184.
10) (a) Curley, J. J.; Sceats, E. L.; Cummins, C. C. J. Am. Chem. Soc.
006, 128, 14036. (b) Curley, J. J.; Cook, T. R.; Reece, S. Y.; M €u ller, P.;
(
2
(
2
(
33) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
999.
34) Hidai, M.; Mizobe, Y. Chem. Rev. 1995, 95, 1115.
Cummins, C. C. J. Am. Chem. Soc. 2008, 130, 9394.
2
(11) (a) Evans, W. J.; Lee, D. S.; Ziller, J. W.; Kaltsoyannis, N. J. Am.
(
Chem. Soc. 2006, 128, 14176. (b) Evans, W. J.; Fang, M.; Zucchi, G.;
Furche, F.; Ziller, J. W.; Hoekstra, R. M.; Zink, J. I. J. Am. Chem. Soc.
2
009, 131, 11195.
12) Mori, H.; Seino, H.; Hidai, M.; Mizobe, Y. Angew. Chem., Int. Ed.
007, 46, 5431.
13) (a) Kawaguchi, H.; Matsuo, T. Angew. Chem., Int. Ed. 2002, 41,
(
2
(
2792. (b) Akagi, F.; Matsuo, T.; Kawaguchi, H. Angew. Chem., Int. Ed.
2
007, 46, 8778.
(
14) Hirotsu, M.; Fontaine, P. P.; Zavalij, P. Y.; Sita, L. R. J. Am.
Chem. Soc. 2007, 129, 12690.
(
(
15) Chatt, J.; Pearman, A. J.; Richards, R. L. Nature 1975, 253, 39.
16) Shilov and co-workers reported the catalytic reduction of
molecular dinitrogen into hydrazine. For recent reviews, see:(a) Shilov,
A. E. Coord. Chem. Rev. 1995, 144, 69. (b) Shilov, A. E. Russ. Chem. Bull.
2
003, 52, 2555.
(
17) Shiina, K. J. Am. Chem. Soc. 1972, 94, 9266.
(18) (a) Komori, K.; Oshita, H.; Mizobe, Y.; Hidai, M. J. Am. Chem.
Soc. 1989, 111, 1939. (b) Komori, K.; Sugiura, S.; Mizobe, Y.; Yamada,
M.; Hidai, M. Bull. Chem. Soc. Jpn. 1989, 62, 2953. (c) Oshita, H.;
Mizobe, Y.; Hidai, M. J. Organomet. Chem. 1993, 456, 213. (d) Mori, M.
J. Organomet. Chem. 2004, 689, 4210 and references therein.
(19) (a) Yandulov, D. V.; Schrock, R. R. Science 2003, 301, 79. (b)
Ritleng, V.; Yandulov, D. V.; Weare, W. W.; Schrock, R. R.; Hock, A. S.;
Davis, W. M. J. Am. Chem. Soc. 2004, 126, 6150.
(20) Cao, Z.; Zhou, Z.; Wan, H.; Zhang, Q. Int. J. Quantum Chem.
2
005, 103, 344.
(
(
21) Studt, F.; Tuczek, F. Angew. Chem., Int. Ed. 2005, 44, 5639.
22) (a) Le Guennic, B.; Kirchner, B.; Reiher, M. Chem.—Eur. J.
2
2
005, 11, 7448. (b) Reiher, M.; Le Guennic, B.; Kirchner, B. Inorg. Chem.
005, 44, 9640. (c) Schenk, S.; Le Guennic, B.; Kirchner, B.; Reiher, M.
Inorg. Chem. 2008, 47, 3634. (d) Schenk, S.; Reiher, M. Inorg. Chem.
2
009, 48, 1638.
(
(
23) H €o lscher, M.; Leitner, W. Eur. J. Inorg. Chem. 2006, 4407.
24) Magistrato, A.; Robertazzi, A.; Carloni, P. J. Chem. Theory
Comput. 2007, 3, 1708.
3
506
dx.doi.org/10.1021/ja109181n |J. Am. Chem. Soc. 2011, 133, 3498–3506