Mo/W-N2 and -N2H2 complexes with trans nitrile ligands: Electronic structure, spectroscopic properties and relevance to nitrogen fixation

Article abstract of DOI:10.1016/S0020-1693(02)00988-X

Molybdenum and tungsten dinitrogen complexes with trans nitrile groups are of significant interest to synthetic nitrogen fixation as they can be protonated to the corresponding N₂H₂ complexes with retention of the trans ligand. This is in contrast to their bis(dinitrogen) counterparts where one N₂ group is exchanged by a ligand X deriving from the acid used for protonation (X=Hal^-, HSO₄^- etc.). Here the first crystal structure of a tungsten(0) dinitrogen nitrile complex having a P₄ ligand set is presented. The electronic and vibrational properties of this and analogous N₂ and N₂H₂ systems with trans nitrile ligands are investigated using IR and Raman spectroscopies coupled to DFT calculations. Force constants are evaluated from experimental frequencies and isotope shifts by the QCA-NCA procedure developed earlier [N. Lehnert, F. Tuczek, Inorg. Chem. 38 (1999) 1659]. The resulting electronic structure descriptions are compared with those obtained earlier for the bis(dinitrogen) and trans fluoro N₂H₂ complexes, respectively [N. Lehnert, F. Tuczek, Inorg. Chem. 38 (1999) 1671]. Importantly, the N₂ ligand is found to be activated to a higher degree in the trans nitrilo as compared with the corresponding bis(dinitrogen) system. On the other hand, the N₂H₂ ligand is less activated in the trans nitrilo than in the analogous trans fluoro complexes. Further, bonding of the nitrile group becomes labile at the N₂H₂ stage of N₂ reduction. These results are interpreted based on the electronic donor/acceptor properties of the respective trans ligands, and the consequences regarding the reactivity of these systems towards further protonation and their potential use for synthetic nitrogen fixation are discussed.

Full text of DOI:10.1016/S0020-1693(02)00988-X

Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
www.elsevier.com/locate/ica
Mo/Wꢀ
/
N2 and ꢀ
/
N2H2 complexes with trans nitrile ligands: electronic
structure, spectroscopic properties and relevance to nitrogen fixation
Carsten M. Habeck, Nicolai Lehnert, Christian Nather, Felix Tuczek *
¨
Institut fur Anorganische Chemie, Christian-Albrechts-Universitat Kiel, Olshausenstraße 40, D-24098 Kiel, Germany
¨ ¨
Received 16 January 2002; accepted 8 April 2002
Dedicated to Professor Dr. Karl Wieghardt on the occasion of his 60th birthday
Abstract
Molybdenum and tungsten dinitrogen complexes with trans nitrile groups are of significant interest to synthetic nitrogen fixation
as they can be protonated to the corresponding N2H2 complexes with retention of the trans ligand. This is in contrast to their
bis(dinitrogen) counterparts where one N2 group is exchanged by a ligand X deriving from the acid used for protonation (Xꢀ ,
Halꢁ
/
ꢁ
HSO4 etc.). Here the first crystal structure of a tungsten(0) dinitrogen nitrile complex having a P4 ligand set is presented. The
electronic and vibrational properties of this and analogous N2 and N2H2 systems with trans nitrile ligands are investigated using IR
and Raman spectroscopies coupled to DFT calculations. Force constants are evaluated from experimental frequencies and isotope
shifts by the QCAꢂNCA procedure developed earlier [N. Lehnert, F. Tuczek, Inorg. Chem. 38 (1999) 1659]. The resulting electronic
/
structure descriptions are compared with those obtained earlier for the bis(dinitrogen) and trans fluoro N2H2 complexes,
respectively [N. Lehnert, F. Tuczek, Inorg. Chem. 38 (1999) 1671]. Importantly, the N2 ligand is found to be activated to a higher
degree in the trans nitrilo as compared with the corresponding bis(dinitrogen) system. On the other hand, the N2H2 ligand is less
activated in the trans nitrilo than in the analogous trans fluoro complexes. Further, bonding of the nitrile group becomes labile at
the N2H2 stage of N2 reduction. These results are interpreted based on the electronic donor/acceptor properties of the respective
trans ligands, and the consequences regarding the reactivity of these systems towards further protonation and their potential use for
synthetic nitrogen fixation are discussed.
# 2002 Published by Elsevier Science B.V.
Keywords: Nitrogen fixation; Dinitrogen complexes; Hydrazido(2-) complexes; Nitrile complexes
1. Introduction
protein. It is believed that N2 is bound and reduced to
NH3 at the ironꢂmolybdenum cofactor (FeMoco), a
/
A synthetic route to nitrogen fixation under ambient
conditions is one of the classic goals of bioinorganic,
metal-organic and coordination chemistry. This requires
a detailed knowledge of the specific conditions for the
activation and reduction of dinitrogen. It has been
demonstrated that information relevant to this subject
can be obtained from the investigation of well-defined
model complexes that are able to bind and reduce N2 [1].
In nature nitrogen fixation is mediated by the enzyme
nitrogenase [2] which is composed of two components,
unique Fe7MoS9 cluster present in the MoFe protein.
One possible coordination geometry of N2 to the
FeMoco involves the end-on terminal coordination [3]
which also exists in simple mononuclear Mo/Wꢀ
/
N2
complexes of the type [M(N2)2(P4)], where MꢀW or
/
Mo and P is one phosphorus donor of mono- or
bidentate phosphine ligands. Remarkably, these systems
can be protonated in a stepwise manner leading to well-
defined intermediates [4]. The hypothesis has been put
forward that protonation of N2 in nitrogenase proceeds
in a similar manner as in these low molecular-weight
complexes [5]. In this sense they can be considered as
functional models for the end-on terminal reduction and
protonation pathway of nitrogenase.
the iron- (Fe-) and the molybdenumꢂiron (MoFe-)
/
* Corresponding author. Tel.: ꢃ
/
49-431-880 2406; fax: ꢃ49-431-880
/
1520
0020-1693/02/$ - see front matter # 2002 Published by Elsevier Science B.V.
PII: S 0 0 2 0 - 1 6 9 3 ( 0 2 ) 0 0 9 8 8 - X

12
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
The bis(dinitrogen) complexes of Mo and W react
with mineral acids to give ‘diazenido(-)’ (NNHꢁ),
detailed insight into the electronic structure of these
intermediates. Since the DFT calculations require struc-
tural input, a crystal structure determination of the
trans nitrilo dinitrogen complex [W(N2)(NCEt)(dppe)2]
has been carried out. This represents the first crystal-
lographic information available for this class of com-
pounds. The consequences of the bonding descriptions
of the N2 and N2H2 intermediates with trans nitrile
ligands with respect to the design of a catalyst for N2
reduction and protonation are discussed.
2ꢁ
‘hydrazido(2-)’
(NNH2
)
and
‘hydrazidium’
(NNH3ꢁ) compounds which in the case of bidentate
phosphine coligands have been structurally and spectro-
scopically characterized [4]. After the splitting of the Nꢀ
/
N bond, nitrido-, imido- and amido species have been
identified [6]. Our research involves the detailed spectro-
scopic and quantum-chemical characterization of these
complex-bound intermediates as well as the study of the
proton and electron transfer steps involved in their
interconversion in order to further explore the physico-
chemical conditions under which a catalytic cycle based
on such systems could actually work [1a,7].
2. Experimental
An ideal metal complex for the homogeneous cata-
lysis of nitrogen fixation should provide a framework
for the binding and reduction of N2 that stays more or
less intact during the entire catalytic cycle. From this
perspective, one of the serious drawbacks of the men-
tioned Mo/W bis(dinitrogen) systems is the exchange of
one N2 group in the very first protonation step against a
2.1. Synthesis, isotopic substitution and sample
preparation
The dinitrogen complexes of tungsten (I, III, V) as
well as the protonated species (II, IV) with the dppe
(dppeꢀbis(diphenylphosphino)ethane) ligand were pre-
/
ꢁ
ligand X (Halꢁ, HSO4 etc.). This strongly affects the
pared following literature procedures [8,12] with slight
variations. The nitrile group was introduced by irradia-
tion of a solution of the bis(dinitrogen) complex with a
100 W tungsten lamp in the presence of an excess of the
respective nitrile. For protonation, degassed sulfuric
acid (96%) was used. Protonation with HBF4 was found
electronic structure of the metal complex mediating the
transformation of N2 to NH3. From the viewpoint of
catalysis, it also corresponds to a loss of 50% of bound
substrate. In the cases reported so far, X is either
identical with the conjugated base of the mineral acid
employed in the protonation (e.g. HSO4ꢁ), or X is a
Lewis base being contained in the conjugated base of
that acid (e.g. Fꢁ in BF4ꢁ). The ligand exchange
obviously represents an unwanted kinetic barrier to
the metal-centered reduction and protonation of N2,
and therefore systems are desirable which avoid this
process.
to lead to decomposition of the nitrileꢂN2 complexes.
/
All protonated species could be deprotonated to the
corresponding dinitrogen complexes with triethylamine.
The 15N isotopomer of W(15N2)2(dppe)2, 0b, was
synthesized using 15N2. The 15N labeled compounds
W(15N2)(NCEt)(dppe)2, Ib, [W(15N2H2)(NCEt)(dppe)2]-
(HSO4)2, IIb, W(15N2)(NCPrn) (dppe)2, IIIb, [W(15N2-
H2)(NCPrn)(dppe)2](HSO4)2, IVb, and W(15N2)-
(NCPh)(dppe)2, Vb, were prepared from the 15N sub-
stituted complex 0b. Because the compounds are air
sensitive, syntheses were performed using the Schlenk-
In this respect, it is noteworthy that Mo/Wꢀ
/
N2
complexes with trans-nitrile groups have been shown
to be protonable on the N2 ligand without ligand
exchange. Studies on their reactivity, [8] their electro-
chemical [9] and NMR spectroscopic [10] properties
have been performed earlier. Electrochemical oxidation
technique. The dinitrogenꢂnitrile systems I and III
/
exchange nitrile against dinitrogen within weeks in the
solid state. In solution, the exchange is much faster.
Sample preparation for vibrational spectroscopy was
carried out in a glovebox. All liquid reagents and
solvents were dried under argon or dinitrogen gas,
with the exception of sulfuric acid which was degassed
by treatment in an ultrasonic bath.
W(N2)2(dppe)2 (0), obtained as orange solid in 50%
yield. Anal. Calc. for C52H48N4P4W: C, 60.2; H, 4.7; N,
5.4. Found: C, 59.1; H, 4.6; N, 5.3%.
potentials have been correlated with trends from Nꢀ
and CꢀN stretching frequencies [9]. In addition, nitriles
have been used to attach Mo/WꢀN2 complexes to
electrodes for the electrosynthesis of ammonia [11]. In
this study, we compare Mo/Wꢂdinitrogen and -hydra-
/
N
/
/
/
zido(2-) systems with trans nitrile ligands to their
bis(dinitrogen) and trans fluoro counterparts, respec-
tively, which we have investigated earlier [7]. In parti-
cular, spectroscopy coupled to quantum-chemical
calculations (DFT) is employed in order to quantita-
tively describe the bonding and activation of the N2 and
N2H2 groups in these systems.
W(N2)(NCEt)(dppe)2 (I), obtained as dark red solid in
60% yield. Anal. Calc. for C55H53N3P4W: C, 62.1; H,
5.0; N, 3.9. Found: C, 61.9; H, 4.9; N, 3.8%.
The results of IR and Raman spectroscopic investiga-
tions are evaluated with a quantum-chemistry assisted
W(N2H2)(NCEt)(dppe)2(HSO4)2 (II), obtained as
white solid in 60% yield. Anal. Calc. for
C55H57N3P4S2O8W: C, 52.4; H, 4.6; N, 3.3. Found: C,
49.7; H, 4.5; N, 2.9%.
normal coordinate analysis (QCAꢂ/NCA) which in
conjunction with the DFT calculations allows to obtain

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
13
W(N2)(NCPrn)(dppe)2(HSO4)2 (III), obtained as dark
red solid in 50% yield. Anal. Calc. for C56H55N3P4W: C,
62.9; H, 5.6; N, 3.8. Found: C, 58.7; H, 5.3; N, 3.0%.
W(N2H2)(NCPrn)(dppe)2(HSO4)2 (IV), obtained as
light green solid in quantitative yield. Anal. Calc. for
C56H59N3P4S2O8W: C, 52.3; H, 4.6; N, 3.3. Found: C,
52.3; H, 4.9; N, 3.0%.
Table 1
Crystallographic data for [W(N2)(NCEt)(dppe)2]
Chemical formula
Formula weight (g molꢁ1
Crystal colour
[W(N2)(C3H5N)(C26H24P2)2]
)
1063.73
red
triclinic
Crystal system
Space group
¯
P
/1/
˚
a (A)
10.2526 (7)
10.7823 (9)
23.438 (2)
87.73 (1)
88.90 (1)
64.96 (1)
2345.6 (3)
2
W(N2)(NCPh)(dppe)2 (V), obtained as dark green
solid in 60% yield. Anal. Calc. for C59H53N3P4W: C,
63.7; H, 4.8; N, 3.8. Found: C, 63.4; H, 4.5; N, 3.2%.
˚
b (A)
˚
c (A)
a (8)
b (8)
g (8)
2.1.1. IR spectroscopy
3
V (A )
˚
MIR spectra were obtained from KBr pellets using a
Mattson Genesis Typ I spectrometer. FIR spectra were
obtained from RbI pellets using a Bruker IFS 66 FTIR
spectrometer. Both instruments were equipped with a
Cryogenics helium cryostat. The spectra were recorded
Z
Dcalc. (g cmꢁ3
F(000)
)
1.506
1076
2u-Range (8)
h/k/l Ranges
4ꢂ52
/
ꢁ11/12, ꢁ13/13, ꢁ28/28
2.640
M (mmꢁ1
T (K)
)
at 10 K and the resolution was set to 2 cmꢁ1
.
150
16 915
8477
Measured reflections
Independant reflections
Rint
2.1.2. Raman spectroscopy
0.0668
569
6862
Raman spectra were obtained on a Bruker IFS 66/CS
NIR FT-Raman spectrometer at 270 K. The setup
involves a 350 mW NdYAG-Laser with an excitation
wavelength of 1064 nm. The samples were pressed into
the groove of a holder which can be sealed with a glass
plate to ensure inert gas conditions.
Refined parameters
Reflections with Foꢀ4s(Fo)
R1 [Foꢀ4s(Fo)]
WR2 (all data)
0.0417
0.0942
1.002
Goodness-of-fit
Residual electron density (e A
ꢁ3
˚
)
1.564/ꢁ2.368
2.1.3. X-ray structure analysis
Intensity data were collected using a STOE Imaging
˜
˜
constants (matrix f ) of model M have been calculated.
In order to remove artificial interactions and to allow an
easier handling, the H atoms of the PH3 ligands were
removed and the models [W(N2)(NCCH3)(P)4] (I?) and
[W(NNH2)(NCCH3)(P)4]2ꢃ (II?) were obtained. Trun-
Plate Diffraction System with MoꢂKa radiation. The
/
structure was solved with direct methods using SHELXS-
97 [13]. Refinement was done against F2 using SHELXL-
97. All non-H atoms were refined anisotropically. In the
subsequent refinement, the H atoms were positioned
with idealized geometry and refined with isotropic
displacement parameters using a riding model. Further
information is contained in Table 1 and in the Supple-
mentary Material.
˜
cation of the matrix f leads to matrix f? which can be
divided into two parts: the force constants of the CH3ꢀ
/
Cꢀ
/
Nꢀ
/
Wꢀ
/
Nꢀ
/
N and the CH3ꢀ
/
Cꢀ
/
Nꢀ
/
Wꢀ
/
Nꢀ/NH2 unit
(core), respectively, and the force constants of the WP4
unit (frame). Small non-diagonal elements were ne-
glected. The force constants of the Cꢀ
/
Nꢀ
/
Wꢀ
/
Nꢀ
/
N
and the CꢀNꢀWꢀNꢀNH2 core were fit to the experi-
/
/
/
/
2.1.4. Normal coordinate analysis
mental frequencies. The force constants of the frame
and those between core and frame were fixed at their
theoretical values. Thus only selected force constants
Normal coordinate calculations were performed using
the QCPE computer program 576 by M.R. Peterson and
D.F. McIntosh. It involves solution of the secular
were refined in the QCAꢂ
/
NCA procedure.
equation GFLꢀLL by the diagonalization procedure
/
of Miyazawa. The calculations are based on a general
valence force field and the force constants are refined
using the non-linear optimization routine of the simplex
algorithm according to Nelder and Mead [14].
2.1.5. DFT calculations
˜
˜
For the models I and II spin-restricted DFT calcula-
tions were performed using Becke’s three parameter
hybrid functional with the LYP correlation functional of
Lee, Yang and Parr. The LanL2DZ basis set was used
for the calculations. It applies Dunning/Huzinaga full
double zeta (D95) [15] basis functions on the first row
and Los Alamos effective core potentials plus DZ
functions on all other atoms [16]. Charges were analyzed
using the natural bond orbital (NBO) formalism [17].
All computational procedures were used as they are
The QCAꢂNCA procedure [7d] is useful for the
/
treatment of large molecules. If the real molecule M is
too large to be handled completely with ab initio theory,
a simplification is necessary. In the case of the com-
plexes I and II, the substitution of the dppe ligands by
PH3 and the usage of molybdenum instead of tungsten
˜
leads to model complexes I and II (model M) which are
˜
˜
suitable for DFT calculations. Frequencies and force

14
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
implemented in the GAUSSIAN-98 package [18]. Wave-
functions were plotted with the visualization program
MOLDEN [19]. The force constants in internal coordi-
nates were extracted from the GAUSSIAN output using
the program REDONG [20]. The structures of the models
Table 2
Comparison of selected experimental and calculated bond lengths and
angles for [W(N2)(NCEt)(dppe)2]
Experimental
Calculated
˚
Bond lengths (A)
˜
˜
I and II are fully optimized. The MO scheme of the
NNH2 ligand has been calculated as well using a fully
optimized structure [7e]. For all model complexes, W
has been substituted by Mo in order to simplify the
calculations.
N(1)ꢀN(2)
N(1)ꢀW(1)
W(1)ꢀN(3)
W(1)ꢀP(1-4)
N(3)ꢀC(5)
C(5)ꢀC(6)
1.120 (7)
1.982 (5)
2.110 (4)
1.170
1.993
2.098
Â2.49
1.189
1.467
2.419 (2) ꢂ2.450 (2)
1.146 (7)
1.477 (8)
/
Bond angles (8)
N(1)ꢀW(1)ꢀN(3)
N(2)ꢀN(1)ꢀW(1)
N(3)ꢀC(5)ꢀC(6)
C(5)ꢀN(3)ꢀW(1)
P(2)ꢀW(1)ꢀP(1)
P(3)ꢀW(1)ꢀP(4)
172.7 (2)
176.5 (5)
175.6 (6)
171.9 (4)
80.97 (5)
79.72 (5)
178.6
177.7
179.2
179.1
89.4
3. Crystal structure determination
The crystal structure of W(N2)(NCEt)(dppe)2 (I, Fig.
1) represents the first structural characterization of a
tungsten(0) dinitrogen nitrile complex having a P4 ligand
set. The central tungsten atom is surrounded by one
dinitrogen, one propionitrile and two bidentate bis(pho-
sphine) ligands in a distorted octahedral geometry. The
nitrile ligand occupies an axial position trans to the
90.6
Table 2. Selected Crystallographic data are presented in
Table 1.
The Nꢀ
magnitude compared with that in the structures of
N distance of I of
˚
N distance of 1.120 (7) A is of similar
/
dinitrogen ligand. The Nꢀ
/
Wꢀ
/
N angle of 172.7 (2)8
N distance of
1.120 (7) A is somewhat longer than in free dinitrogen
deviates slightly from linearity. The Nꢀ
/
[W(N2)2(dppe)2] (0) [22]. The Wꢀ
/
˚
˚
1.982 (5) A is significantly shorter than in the first
polymorph of 0, [22a] but comparable with those in the
˚
(1.0975 A) [1c]. The elongation of the Nꢀ
/
N distance is
explained by back donation of electrons from the metal
non-bonding d orbitals into antibonding p* orbitals of
dinitrogen. The two bis(phosphine) ligands are bound in
second one [22b]. The average Wꢀ
comparable in all systems. All PꢀWꢀ
significantly from 908 present in an ideal octahedral
environment.
/
P distances are
P angles deviate
/
/
equatorial position. The Pꢀ
/
Wꢀ
/
P bite angle is 80.97 (2)
and 79.72 (5)8, respectively. The average Pꢀ
/
W bond
C bond distance
˚
distance is 2.430 (2) A. The average Pꢀ
/
˚
of 1.861 (6) A in the phosphine ligand of I is slightly
longer than in the uncomplexed ligand (1.821 and 1.818
˚
˚
for dinitrogen and 2.110 (4) A for the nitrile nitrogen
4. Electronic structure
˚
N distances are 1.982 (5) A
A, respectively) [21]. The Wꢀ
/
atom. The Cꢀ
/
N distance in the nitrile ligand is 1.146 (7)
˚
A. Relevant bond distances and angles are listed in
4.1. Input geometries and geometry optimizations
˜
The structure of the model system I used for the
calculation is derived from the crystal structure of
complex I and is fully optimized. Experimentally
determined bond lengths and angles are compared
with calculated values (after optimization) in Table 2.
In order to simplify the computational problem, the
coordinating phosphorus donors were substituted by
ideal PH3 ligands in a geometry retrieved from literature
data. The alkyl group of EtCN was substituted by
˜
NCCH3. Input coordinates of I are collected in Table 3.
˜
The structure of model system II was derived from that
˜
of I by adding two hydrogen atoms to the N2 group at a
distance known from the NNH2 complex
[MoF(NNH2)(dppe)2](BF4) [23]. Input coordinates are
˜
˜
given in Table 4. Structure V was obtained from I by
replacing the methyl group on the nitrile ligand by
phenyl. Input coordinates are given in Table 5. All
Fig. 1. Crystal structure of [W(N2)(NCEt)(dppe)2] (I) with labeling
and displacement ellipsoids drawn at the 50% probability level (the
hydrogen atoms are omitted for clarity).

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
15
Table 3
Coordinates of [Mo(N2)(NCCH3)(PH3)4] (
Table 4
Coordinates of [Mo(N2H2)(NCCH3)(PH3)4] (
˜
˜
II)/
/
I)
/
/
a
a
Atom
Position
Atom
Position
x
y
z
x
y
z
C
0.049
0.038
ꢁ4.473
ꢁ3.006
ꢁ1.817
0.281
0.000
0.000
C
4.549
3.088
1.913
0.0046
0.003
ꢁ0.0347
ꢁ0.0249
ꢁ0.0203
ꢁ0.001
1.925
C
C
N
Mo
P
0.012
0.000
0.000
0.000
N
Mo
P
ꢁ0.001
ꢁ0.006
ꢁ1.747
1.957
ꢁ0.362
ꢁ0.191
ꢁ0.108
ꢁ0.244
ꢁ2.133
ꢁ3.447
ꢁ0.136
ꢁ1.082
ꢁ0.582
1.011
ꢁ2.507
0.025
0.314
0.255
0.000
2.495
P
P
1.714
N
N
P
ꢁ0.059
ꢁ0.139
2.487
2.273
3.440
0.000
0.000
0.000
P
1.769
ꢁ1.907
0.006
0.011
N
N
P
ꢁ0.011
ꢁ0.012
ꢁ1.963
2.917
0.308
0.255
P
0.025
ꢁ0.934
1.185
ꢁ2.495
3.286
3.245
ꢁ1.712
ꢁ1.759
ꢁ3.246
ꢁ2.144
ꢁ1.247
ꢁ2.689
ꢁ2.563
1.614
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
1.000
0.698
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
1.400
2.409
ꢁ0.139
3.236
3.236
ꢁ0.995
0.933
0.933
3.213
ꢁ1.074
1.074
0.000
ꢁ0.103
ꢁ1.172
1.011
ꢁ3.316
ꢁ2.061
ꢁ1.966
ꢁ3.136
ꢁ1.607
ꢁ1.815
2.166
3.244
1.185
ꢁ0.930
0.698
ꢁ3.246
ꢁ3.213
ꢁ3.286
0.000
ꢁ0.042
0.857
ꢁ0.139
ꢁ0.934
ꢁ3.152
ꢁ3.313
ꢁ3.313
ꢁ0.461
ꢁ0.461
1.077
ꢁ0.995
1.000
1.608
2.888
2.798
ꢁ1.319
ꢁ1.233
0.931
2.567
2.690
ꢁ0.249
ꢁ0.249
ꢁ4.869
ꢁ4.869
ꢁ4.858
ꢁ1.067
1.067
0.888
1.865
3.297
0.996
0.105
4.929
1.260
ꢁ0.308
0.957
ꢁ0.888
0.000
4.938
4.926
ꢁ0.258
ꢁ0.727
ꢁ0.298
0.269
ꢁ0.760
ꢁ0.817
0.845
ꢁ3.964
ꢁ3.957
a
˚
Coordinates of the atoms in A.
a
˚
Coordinates of the atoms in A.
structures were optimized. Relevant bond lengths of the
optimized structures are indicated in Fig. 2.
Notably, backbonding acts to put electron density
primarily on the b atom of both ligands, and dinitrogen
participates more than nitrile in the bonding combina-
4.2. Electronic structure of the N2 complex with a trans
alkyl nitrile ligand
tions (about 15 vs. 9% orbital contribution for Nꢀ
CꢀN, respectively).
Metalꢂligand coordination is complemented by s
/
N and
The molecular orbital (MO) scheme of the Molybde-
num N2 model complex with a trans acetonitrile ligand,
˜
I; is very similar to that of the Mo bis(dinitrogen)
complex which we have studied earlier (Fig. 3) [7d,7e].
Charge decompositions and shapes of relevant orbitals
are given in Table 6 and Fig. 4, respectively. The
/
/
bonding. For the N2 group this is mediated by the
ps(NN)_dz2 orbital that is heavily admixed by phos-
phine functions and has metal dz2 and N2 contributions
of about the same magnitude (15 and 24%, respectively,
Fig. 4). The participation of the CN group in this orbital
is weak. Nitrile s bonding is primarily mediated by the
ps(CN) orbital that has a very high contribution of the
CN group, especially at NC (NC: 56%, CN: 17%).
Furthermore, it has a 10% contribution of N2 and a
metal contribution of only 8% (Table 6). The
ps(NN)_ps(CN)_dz2 orbital at low energy is a super-
position of N2 and nitrile s orbitals (Fig. 4). As in the
bis(dinitrogen) complex, p donation from the ligand p
orbitals is found to be negligible.
˜
HOMO of I is contained within the twofold degenerate
set of bonding combinations between the Mo dxz and
ꢄ
dyz orbitals and the ligand (i.e. dinitrogen and nitrile) px
ꢄ
and py orbitals, respectively. These combinations are
ꢄ
ꢄ
designated dxz_px and dyz_py . For both ligands, the
contribution of the coordinating atom in these combi-
nations (Na for dinitrogen, NC for nitrile) is very small
relative to that of the b atom (Nb and CN, respectively).
˜
The LUMO of I is part of the twofold degenerate set of
non-bonding combinations of p*(NN) and p*(CN)
orbitals in x- and y-direction, respectively. At much
higher energy, finally, the antibonding combinations of
ligand p* orbitals with the metal dxz and dyz orbitals,
Comparing the NPA charges on the two axial ligands
˜
of I (Table 7), it becomes apparent that the CN group in
˜
I carries a smaller negative charge (ꢁ
/
0.05) than the
ꢄ
ꢄ
px_dxz and py _dyz, are found. This orbital pattern
corresponds to dp(metal) 0 p* (ligand) back donation.
dinitrogen ligand (ꢁ
charge on Nb is twice as large in I (ꢁ0.12) as in the
/
0.22). In particular, the negative
˜
/
/

16
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
Table 5
Coordinates of [Mo(N2)(NCPh)(PH3)4] (/
˜
V)/
a
Atom
Position
x
y
z
C
0.055
0.033
3.305
1.880
0.685
0.000
0.000
0.000
0.000
2.507
0.000
ꢁ2.507
0.000
0.000
0.000
ꢁ3.240
ꢁ3.256
ꢁ3.262
1.075
ꢁ1.075
0.000
3.256
3.262
3.240
0.000
1.067
ꢁ1.067
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
C
N
Mo
P
0.000
ꢁ0.016
0.007
ꢁ1.378
ꢁ1.401
ꢁ1.353
ꢁ1.401
ꢁ3.376
ꢁ4.541
ꢁ1.351
ꢁ2.408
ꢁ1.555
ꢁ0.267
ꢁ1.937
ꢁ1.937
ꢁ0.075
ꢁ1.555
ꢁ0.267
ꢁ2.408
ꢁ2.621
ꢁ0.756
ꢁ0.756
4.010
P
ꢁ2.526
0.007
P
N
N
P
ꢁ0.095
ꢁ0.195
2.474
H
H
H
H
H
H
H
H
H
H
H
H
C
ꢁ0.727
1.239
ꢁ0.481
3.248
3.248
3.162
1.239
ꢁ0.481
ꢁ0.727
ꢁ3.217
ꢁ3.305
ꢁ3.305
1.292
C
1.306
0.098
5.413
6.141
C
C
ꢁ1.132
ꢁ1.161
2.224
5.450
4.048
C
H
H
H
H
H
3.451
5.940
2.257
0.115
7.228
6.006
3.518
ꢁ2.066
ꢁ2.109
a
˚
Coordinates of the atoms in A.
corresponding bis(dinitrogen) complex (ꢁ0.06) which is
/
important to protonation. Consequently a higher
amount of negative charge is transferred from the metal
˜
to the single N2 ligand in I than to each of the two N2
˜
groups in the (N2)2 complex; i.e. the N2 group in I is
Fig. 2. Optimized structures for the model complexes.
activated to a higher degree than each of the two groups
of the bis(dinitrogen) system. Hence, nitrile must be a
weaker p acceptor and/or a stronger s or p donor than
N2. From the above observations, nitrile is in fact a
weaker p acceptor than N2. As in case of N2, p donation
from nitrile to the metal is very weak (note the very
small metal contribution in the px, py orbitals being
largely localized on NC).
not donated to the metal to a significant degree. The
overlap between the CꢀN s orbital ps(CN) having lone-
/
pair character on NC and the metal is even weaker than
in case of the N2 ps orbital. This probably derives from
the fact that the coordinating nitrogen atom of the
nitrile group is bound to a positively charged C atom
and hence, cannot donate electron density to a signifi-
cant extent (see below). To conclude, the primary
bonding interaction for the nitrile group with the Mo/
W(0) center is backbonding, but is weaker than in the
case of the N2 ligand. For this reason, the dinitrogen
systems with trans nitrile ligands should be thermo-
dynamically less stable than their bis(dinitrogen) coun-
terparts. This is also evident from the back
transformation of the trans alkyl nitrile complexes I
In addition, the nitrile ligand is found to be a weaker
s donor than dinitrogen. Upon going from N2 to nitrile,
one N atom is replaced by the weaker electronegative
Cꢀ
origin for the large polarization of the CN group leading
to a charge of ꢁ
0.40 at NC and of ꢃ0.35 at CN.
However, this large amount of negative charge on NC is
/
CH3 group. This difference in electronegativity is the
/
/

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
17
˜
Fig. 3. Orbital schemes of I and II:
˜
/
and III into the bis(dinitrogen) complex 0 at room
temperature under N2 (vide supra).
transformed into a p donor orbital at Na, and the
bonding combination of this orbital with dxz primarily
has ligand character. This corresponds to an effective
transfer of about two electrons from the metal to the N2
ligand, initiated by the double protonation reaction. In
contrast, the bonding combination of dyz with the out-
˜
4.3. Electronic structure of the N2H2 complex II
/
˜
Addition of two protons to the dinitrogen complex I
˜
leads to the ‘hydrazido(2-)’ model compound II with a
ꢄ
of plane p* orbital of the N2H2 group, pv , still has
predominantly metal character. The occupation of this
orbital corresponds to back donation from dyz. The
LUMO contains the antibonding combination of dxz
trans acetonitrile ligand which is related to the analo-
gous intermediate with a trans fluoro ligand that has
been characterized earlier [7d,7e]. Charge decomposi-
tions and shapes of relevant orbitals are given in Table 8
and Fig. 5, respectively. As in the case of the corre-
sponding trans fluoro system, the non-bonding dxy
ꢄ
with ph. Importantly, this orbital still carries a bonding
ꢄ
contribution of the nitrile px orbital, like in the parent
˜
complex I (Fig. 5). To higher energy, the antibonding
ꢄ
combination of dyz with the pv orbital is found. The
˜
orbital is the HOMO of complex II (Figs. 3 and 5).
ꢄ
ꢄ
antibonding combinations of nitrile px and py functions
with dxz and dyz, respectively, are located at still higher
energy as shown in Fig. 3.
Below the HOMO are the bonding combinations of dxz
ꢄ
ꢄ
and dyz with the N2H2 ph and pv orbitals, respectively.
ꢄ
The p* orbital lying within the N2H2 plane, ph , has

18
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
Table 6
Charge contributions of [Mo(N2)(NCCH3)(PH3)4] (/
˜
I)/
Orbital
Label
Energy (Hartree)
Charge decomposition
% Mo
% P % Na
% Nb
% NC
% CN
% C % HC
% HP
ꢄ
ꢄ
ꢄ
ꢄ
ꢄ
ꢄ
px (NN)_px (CN)_dxz
ꢁ54ꢂ
ꢁ53ꢂ
ꢁ52ꢂ
ꢁ43ꢂ
ꢁ42ꢂ
ꢁ41ꢂ
ꢁ40ꢂ
ꢁ39ꢂ
ꢁ38ꢂ
ꢁ37ꢂ
ꢁ35ꢂ
ꢁ34ꢂ
ꢁ33ꢂ
ꢁ21ꢂ
ꢁ17ꢂ
0.0615
0.0540
5
40
18
22
24
24
2
2
1
1
0
1
2
3
1
1
0
0
0
0
0
6
6
0
5
4
1
8
6
7
7
2
2
0
0
0
5
5
0
1
1
49
23
21
9
py (NN)_py (CN)_dyz
9
15
8
11
7
8
5
ꢄ
px (NN)_px (CN)
0.0507
0.0011
0.0011
18
17
18
61
60
68
7
9
7
ꢄ
py (NN)_py (CN)
7
11
12
15
14
0
9
16
15
9
ꢄ
ꢄ
px (NN)_px (CN)
8
9
8
ꢄ
ꢄ
Dyz _py (NN)_py (CN)
ꢁ0.1372
ꢁ0.1373
ꢁ0.1456
ꢁ0.2812
ꢁ0.2814
ꢁ0.3460
ꢁ0.3470
ꢁ0.3485
ꢁ0.4315
ꢁ0.5392
1
0
11
11
28
10
10
6
ꢄ
ꢄ
D
xz _px (NN)_px (CN)
2
1
0
9
Dxy
3
0
0
0
px (CN)_px(NN)
py (CN)_py(NN)
py (CN)
77
77
4
1
2
2
2
7
1
2
2
2
2
0
0
49
48
2
27
27
1
px (CN)
2
4
1
1
5
ps (NN)_dz2
ps(CN)
ps (NN)_ps (CN)_dz2
15
8
51
2
9
15
6
7
/
/
4
51
56
8
17
7
1
7
1
20
0
P and H are the sums over all corresponding atoms.
˜
To summarize, the characteristics of metalꢂ
/
N2H2
binding in II are very similar to those in the correspond-
increase of metalꢂnitrile s bonding upon going from I
/
˜
˜
to II: The metal contribution to the orbital
ps(CN)_ps(NN)_dz2 remains low (Fig. 5 and Table 8).
As p donation is not increased either, the bonding
between the nitrile group and the Mo center now
ing trans fluoro N2H2 complex. In both cases, the metal
center can be assigned a formal charge of ꢃ
/
II. Nitrile
˜
bonding in the N2H2 complex II; however, is very
˜
different from that present in the precursor I : while the
˜
nitrile group participates in backbonding in I; it has
negligible contributions to the occupied orbitals with
becomes very weak. This is also reflected by a very
˚
N(nitrile) bond length (calc. 2.275 A after
˚
optimization, vs. 2.098 A in I).
large Moꢀ
/
˜
metal d character in II and, therefore, is practically non-
Comparing the charges on the N2H2 ligand between
the trans fluoro and trans nitrilo complexes shows that
ꢄ
ꢄ
bonding with its px and py orbitals. Considering the
˜
NPA charges on the nitrile group of II; it becomes
apparent that the polarization of the CN group, which
˜
the nitrilo complex II has a weaker activated N2H2
ligand than the corresponding fluoro compound (ꢃ0.20
vs. ꢁ0.03 total charge on N2H2, cf. Table 7). This
/
already is large in the uncoordinated ligand and in
C
/
˜
˜
complex I; has further increased in II (N : ꢁ
/
0.50, CN:
difference is obviously due to the different charges of the
trans ligand. In [MoF(NNH2)(PH3)4]ꢃ the fluoro
ꢃ0.56). Surprisingly, however, this does not lead to an
/
˜
Fig. 4. Important orbitals of I:
/

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
19
Table 7
NPA Charges of different Mo complexes
Complex
Atom
Mo
Na
Nb
H
Nc
CN
˜
/P/
Free CH3CN ligand
[Mo(N2)2(PH3)4][7e]
˜
ꢁ0.37
0.33
ꢁ1.01
ꢁ0.89
ꢁ0.84
ꢁ0.02
0.04
ꢁ0.07
ꢁ0.10
ꢁ0.09
ꢁ0.10
ꢁ0.21
ꢁ0.36
ꢁ0.06
ꢁ0.12
ꢁ0.10
ꢁ0.58
ꢁ0.68
ꢁ0.60
0.31
0.29
0.29
0.20
0.24
0.20
[Mo(N2)(NCCH3)(PH3)4] (
˜
/
I)
/
-0.40
ꢁ0.39
ꢁ0.50
0.35
0.34
0.56
[Mo(N2)(NCPh)(PH3)4] (
/
V)
/
[Mo(NNH2)(NCCH3)(PH3)4]2ꢃ
[MoF(NNH2)(PH3)4]ꢃ[7e]
[MoF(NNH3)(PH3)4]2-[29]
(/
II)
0.44
0.43
0.48
˜
/
0.19
c
Na is the coordinating and Nb the terminal dinitrogen atom of N2. N is the dinitrogen atom from the nitrile. P is the average of all P atoms.
˜
[7d,7e]. These are missing in the case of the trans nitrile
complex (the CN group carries a positive charge of ꢃ
/
0.06). We have pointed out earlier that the N2H2 ligand
in the MoF(NNH2) complex corresponds to neutral
isodiazene and that the hydrazido(2-) formulation is
misleading [7d,7e]. The same applies to the weaker
activated N2H2 intermediate of the trans nitrile system.
4.4. Electronic structure of the N2 complex with a trans
˜
benzonitrile ligand V
/
It has been found earlier that the N2 complex can also
be prepared with a trans benzonitrile ligand [8]. How-
ever, a corresponding N2H2 intermediate has not been
described. It therefore appeared of interest to investigate
the electronic-structural differences between dinitrogen
complexes with alkyl nitrile and benzonitrile groups.
Table 7 shows that in fact the negative charge on the N2
ligand is slightly lower in the system coordinated by an
aromatic nitrile than in a system with a trans aliphatic
nitrile ligand. Fig. 6 shows the electronic structural
origin of this difference: there is a delocalization of
˜
Fig. 5. Important orbitals of II:
/
ligand carries a negative charge of ꢁ0.55, i.e. has
transferred 0.45 charge units to the MoNNH2 fragment
/
Table 8
Charge contributions of [Mo(N2H2)(NCCH3)(PH3)4]2ꢃ
˜
II)
(/
/
Orbital
Label Energy (Hartree)
Charge decomposition
% Mo % P % Na % Nb % NC % CN % C % HC % HP % HN
/
/
/
/
pvꢄ(CN)_pvꢄ(NN)_dyz
phꢄ(CN)_dxz
dyz_pvꢄ(CN)_pꢄv (NN)
dxz_phꢄ(CN)
dxy
pvꢄ(NN)_dyz
/
ꢁ47ꢂ
ꢁ46ꢂ
ꢁ43ꢂ
ꢁ42ꢂ
ꢁ41ꢂ
ꢁ40ꢂ
ꢁ37ꢂ
ꢁ34ꢂ
ꢁ33ꢂ
ꢁ31ꢂ
ꢁ25ꢂ
ꢁ24ꢂ
ꢁ23ꢂ
ꢁ20ꢂ
ꢁ0.2421
ꢁ0.2469
ꢁ0.3077
ꢁ0.3348
ꢁ0.4588
ꢁ0.4890
ꢁ0.5732
ꢁ0.6216
ꢁ0.6227
ꢁ0.6527
ꢁ0.6764
ꢁ0.6772
ꢁ0.7066
ꢁ0.7620
14
11
25
47
88
43
23
2
1
8
2
29
29
3
34
36
11
8
1
11
11
2
1
0
3
0
6
0
0
7
0
0
0
0
0
0
1
/
4
2
1
0
2
/
10
9
31
23
0
13
0
0
4
/
1
0
1
4
0
0
0
0
0
0
12
2
/
/
/
0
10
53
0
42
4
0
2
0
0
dxz_phꢄ(NN)
/
10
3
0
0
0
0
4
pv (CN)
ph (CN)
pv (NN)
0
45
45
1
19
18
0
18
18
1
10
10
0
4
2
4
0
0
4
2
23
46
18
2
21
4
24
3
28
43
16
2
pv (NN)
3
0
0
0
0
/pv(NN)_dyz
/ps(CN)_ps(NN)_dz2
/ps(CC)_ps(NN)_dz2
/
7
31
6
23
4
0
0
2
1
/
9
45
6
17
22
8
27
1
/
8
0
7
9
6
0
P and H are the sums over all corresponding atoms.

20
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
stretching vibration between tungsten and N2, n(WN).
In the 14N compound, the n(WN) peak overlaps with
another signal which is the reason for the appearance of
two peaks in the 15N spectra. The unshifted peak cannot
be assigned. It is neither the n(WN) stretch of the nitrile
(based on the frequency predicted from the calculations,
vide infra), nor does it originate from a splitting of the
former vibration. Other intense features in the Raman
spectra which remain unshifted in the 15N compound
are found at 2185, 1584, 1028, 1093, 999, 525 and 178
˜
Fig. 6. Important orbital of V:
/
cmꢁ1. The signal at 2185 cmꢁ1 is assigned to the Cꢀ
stretching mode n(CN) of the nitrile and the peak at 178
cmꢁ1 to the symmetric Wꢀ
P stretch n(WP). Compar-
/N
charge from the backbonding dyz orbital into the p*
system of the aromatic ring such that the charge transfer
/
ꢄ
into the N2 pv orbital is lowered. This reduction of
negative charge is obviously sufficient to prevent proto-
nation of the coordinated N2 ligand.
ison with data from [W(N2)2(dppe)2] suggests that the
remaining vibrations are due to the phenyl residues of
the dppe ligand: thus, the peak at 1586 cmꢁ1 belongs to
a Cꢃ
/
C stretch; 1093 and 1028 cmꢁ1 are the in-plane
bending vibrations d(CCC) and d(CCH), and the
feature at 525 cmꢁ1 is either an out-of-plane bend
5. Vibrational spectra and QCA
/
NCA
ꢂ
g(CCC) or the Pꢀ
The infrared spectra of complex I are given in Fig. 8.
Besides the NꢀN stretching vibration n(NN) at 1897
cmꢁ1 which shifts to 1836 cmꢁ1 on 15N substitution,
another isotope sensitive band is found below 600 cmꢁ1
/phenyl stretch [24].
5.1. Vibrational spectra of [W(N2)(NCEt)(dppe)2] (I)
/
The Raman spectra of I (Fig. 7) exhibit two peaks at
1897 and 462 cmꢁ1 that shift in the 15N2-substituted
compound. The signal at 1897 cmꢁ1 shifts by about 60
.
The band at 555 cmꢁ1 is assigned to the d(WNN) bend
and shifts by about 17 cmꢁ1 into a broad feature. The
other isotope sensitive signals at 2010 and 1947 cmꢁ1 in
the 14N spectra (15N: 1958 and 1881 cmꢁ1) are due to
cmꢁ1 down to 1834 cmꢁ1 and is assigned to the Nꢀ
/
N
stretching vibration. The peak at 462 cmꢁ1 shifts by 18
cmꢁ1 on 15N substitution and is assigned to the Wꢀ
/
N
Fig. 7. Raman spectra of [W(N2)(NCEt)(dppe)2] (I).

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
21
Fig. 8. IR spectra of [W(N2)(NCEt)(dppe)2] (I).
the symmetric and antisymmetric NN vibrations of the
precursor [W(N2)2(dppe)2] (0). The n(CN) stretching
mode is found at 2190 cmꢁ1. The observed signals at
1585, 1093, 1026 and 1000 cmꢁ1 are assigned as
presented above. Frequencies and assignments are
collected in Table 9.
experimental data. Table 9 shows the comparison
between the experimentally determined peak positions,
the calculated data retrieved from DFT (B3LYP) and
the QCAꢂ/NCA results. The f-matrix for the relevant
CꢀNꢀWꢀNꢀ
/
/
/
/N core is shown in Table 10.
The agreement between the DFT results and the
experiment is good. The n(CN) stretch is found between
2185 and 2190 cmꢁ1 and is calculated to be at 2199
cmꢁ1. The n(NN) stretch is calculated by DFT to 1957
cmꢁ1, but is found at 1897 cmꢁ1. For the bending mode
d(WNN) only one signal at 555 cmꢁ1 is observed (see
Table 9) although calculations indicate a splitting into
two signals at 555 cmꢁ1 and 393 cmꢁ1. This could not
be confirmed by experimental data. The n(WN) stretch
5.2. QCAꢂNCA of [W(N2)(NCEt)(dppe)2] (I)
/
The structure of model system [W(N2)(NCCH3)(P)4]
(I?) used for the normal coordinate analysis is shown in
Fig. 9. The QCAꢂNCA procedure is performed as
/
described in the experimental section. The n(NN),
n(CN), d(WNN), and n(WN) vibrations are fit to the
Table 9
Comparison of the observed and calculated frequencies of [W(N2)(NCEt)(dppe)2]
Mode
Experimental
14N (cmꢁ1
QCAꢂ
/
NCA
B3LYP
14
)
15N (cmꢁ1
)
14N (cmꢁ1
)
15N (cmꢁ1
)
N
(cmꢁ1
)
Calc
a
a
a
a
n(CN)
n(NN)
2190/2185
1895/1897
555
2190/2185
1836/1834
538
2190
1897
555, 376
462
2190
1837
537, 364
448
2199
1957
d(WNN)
n(WNN)
n(WNC)
555, 393
438
285
a
a
462
n.o.
444
n.o.
270
269
n.o., not observed.
a
Raman-data, unlabeled data are IR-data.

22
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
ꢁ1
(bis(dinitrogen) complex: 2.66 mdyn Aꢁ1), and
˚
˚
A
that of the corresponding bending vibration d(WNN)
1.01 mdyn Aꢁ1. The force constant of the n(WN)
˚
stretch between tungsten and the nitrile is 1.53 mdyn
˚
A
ꢁ1, which is much lower than that of the correspond-
ing Wꢀ
/
N2 stretch. This again reflects the weak interac-
tion between the metal and the nitrile ligand.
5.3. Vibrational spectra of
[W(N2H2)(NCEt)(dppe)2][HSO4]2 (II)
Due to fluorescence, the Raman spectra of II could
not be obtained. Overview IR spectra are shown in Fig.
10; selected features are presented in Fig. 11. The IR
spectra exhibit a very broad feature between 2700 and
3300 cmꢁ1 which is caused by the counterion and
overlaps with the n(NH) vibrations. These are observed
at 3329 and 3213 cmꢁ1, shifting to about 3279 (broad
band) and 3181 cmꢁ1 on 15N isotope substitution.
Other isotope sensitive signals are found at 1617 cmꢁ1
and at about 440 cmꢁ1. Both show very small shifts and
are assigned to d(NNH) and d(WNN), respectively. The
n(WN) stretch is masked by the broad band from the
Fig. 9. Structure of model system I?.
is found at 462 cmꢁ1 while the B3LYP calculations
predict this vibration at 438 cmꢁ1, showing a deviation
of 24 cmꢁ1. The n(WN) stretch for the nitrile is not
observed in the spectra, but the DFT calculations
predict the signal at 285 cmꢁ1
.
Comparison of the experimental data and the QCAꢂ
/
NCA result shows excellent agreement as given in Table
9. For the bending mode d(WNN) only one signal at
555 cmꢁ1 could be observed experimentally. After
fitting to the experimental data, the observed isotope
shift as well as the 14N frequency are reproduced (see
ꢁ
HSO4 counterion at about 590 cmꢁ1. The n(CN)
stretch is found between 2262 and 2267 cmꢁ1 (see Table
11), and shows a shift to higher wavenumbers as
compared with the free nitrile.
Table 9). The second signal is then predicted by QCAꢂ
/
NCA to be at 376 cmꢁ1 in the 14N complex and 364
cmꢁ1 in the isotope substituted compound. The n(WN)
stretch for the nitrile is not observed in the spectra but
5.4. QCAꢂNCA of
/
[W(N2H2)(NCEt)(dppe)2][HSO4]2 (II)
predicted after QCAꢂ
/
NCA to be located at 270 cmꢁ1
.
The calculated force constants for I? are collected in
The structure of model system [W(N2H2)(NCCH3)-
(P)4] (II?) used for normal coordinate analysis is shown
Table 12. The force constant of the Nꢀ
/
N stretch is 14.95
mdyn Aꢁ1, which reflects the activation of the dinitro-
˚
in Fig. 12; the QCAꢂNCA procedure is performed as
/
gen in the complex compared with the free N2 molecule
(22.42 mdyn Aꢁ1) [7d]. Importantly, this value is lower
described in the Experimental section. The n(NN),
n(CN), d(NNH), n(NH) and d(WNN) vibrations are
fit to the experimental data and comparison of the
˚
than in the corresponding bis(dinitrogen) complex
which has an fNN value of 16.43 mdyn Aꢁ1. The CN
˚
experimental data, the QCAꢂNCA fit and the predicted
/
force constant of the coordinated nitrile (16.70 mdyn
˚
vibrational energies (from DFT) are shown in Table 11.
The DFT calculations predict all vibrations of this
system at higher frequencies than measured experimen-
tally. The two NH stretching modes (symmetric and
antisymmetric) are calculated at 3671 and 3523 cmꢁ1 in
A
mdyn A
ꢁ1) is somewhat lower than in the free nitrile (17.01
ꢁ1
˚
calculated by DFT), showing a weak effect
of backbonding. The force constant for the n(WN)
stretch between tungsten and dinitrogen is 3.15 mdyn
Table 10
f Matrix of the central CꢀNꢀWꢀNꢀN unit in [W(NCCH3)(N2)(P4)]
n(NN)
n(WNN)
n(WNC)
n(CN)
dx (WNN)
dx (CNW)
dy (WNN)
dy (CNW)
f1
f9
0
f2
f10
0
f3
f11
0
0
f4
0
0
0
f5
0
0
0
0
0
f6
0
0
0
0
0
0
f7
0
0
0
0
0
0
0
F8

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
23
Fig. 10. IR spectra of [W(NNH2)(NCEt)(dppe)2][HSO4]2 (II).
Fig. 11. Selected IR spectral features of II along with 15N substitution data (left: n(NN), middle: n(WN) (masked), right: d(WNN)). Arrows indicate
positions of 15N shifted peaks.

24
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
Table 11
Comparison of the observed and calculated frequencies of [W(N2H2)(NCEt)(depe)2][HSO4]2
Mode
Experimental
QCAꢂ
/
NCA
B3LYP
14
14N (cmꢁ1
)
15N (cmꢁ1
)
14N (cmꢁ1
)
15N (cmꢁ1
)
N
(cmꢁ1
)
Calc
nas(NH)
ns(NH)
n(CN)
3329
3215
2263
1617
n.o.
3309
3194
2263
1614
1384
Â590
Â440
n.o.
3313
3230
2261
1617
1428
592
3302
3225
2261
1614
1382
576
3671
3523
2299
1673
1500
634
d(NNH)
n(NN)
n(WNN)
d(WNN)
n(WNC)
Â590
Â440
n.o.
446
221
442
220
448
204
n.o., not observed.
broad feature was determined based on the position of
the 15N peak. In the 15N spectra the n(NN) signal is
observed at 1382 cmꢁ1. Based on QCAꢂNCA the 14N
/
n(NN) stretch should be found at 1428 cmꢁ1 (cf. Table
11). The tungsten dinitrogen stretch is predicted at 592
cmꢁ1 for 14N by QCA-NCA, but is masked experimen-
tally by a band of the counterion (vide supra). The
tungsten nitrile stretch, finally, is calculated at 220 cmꢁ1
but is not observed (Table 11). The d(WNN) stretch was
identified at about 440 cmꢁ1, based on QCAꢂ
(446 cmꢁ1 (14N)/442 cmꢁ1 (15N)), respectively.
NH2 core is shown
/
NCA
The f-Matrix for the Cꢀ
/
Nꢀ
/
Wꢀ
/
Nꢀ
/
in Table 13. Important calculated force constants for I?
are collected in Table 12. The NN stretch has a diagonal
force constant of 7.60 mdyn Aꢁ1. This is reduced
˚
Fig. 12. Structure of model system II?.
compared with free NNH2 (calculated 11.66 mdyn
˚
A
ꢁ1), but higher than in [MoF(NNH2)(dppe)2](BF4)
˚
the 14N compound but observed at 3329 and 3215 cmꢁ1
.
(7.197 mdyn Aꢁ1) [7d]. The force constant for the CN
ꢁ1
n(CN) is found experimentally at 2263 and calculated at
2299 cmꢁ1. For the d(NNH) bending mode, a differ-
ence of 56 cmꢁ1 between 1673 cmꢁ1 (calculated by
DFT) and 1617 cmꢁ1 (observed) is obtained (Table 11).
The position of the 14N n(NN) stretch which is masked
by a broad feature was determined based on the position
of the 15N peak. In the 15N spectra the n(NN) signal is
observed at 1382 cmꢁ1 and calculated to be at 1500
cmꢁ1. Based on the DFT results, the tungsten dinitro-
gen stretch should be found at 634 cmꢁ1 but is masked
in experiment by a band of the counterion (vide supra).
The tungsten nitrile stretch is calculated at 204 cmꢁ1 but
is not observed experimentally (Table 11). The d(WNN)
stretch was identified at about 440 cmꢁ1, based on the
DFT prediction (448 cmꢁ1).
˚
stretch is with 17.60 mdyn A
larger than in the free
ligand (17.01 mdyn A ) or in I (16.70 mdyn Aꢁ1, vide
supra). The force constants for the NH stretches are 5.90
and for the n(WN) stretches 6.44 (between tungsten and
ꢁ1
˚
˚
ꢁ1
˚
dinitrogen) and 0.88 mdyn A (between tungsten and
nitrile), respectively. The value of the force constant for
˚
the d(WNN) bend is 0.46 mdyn A
ꢁ1
.
5.5. Vibrational spectra of [W(N2)(NCPrn)(dppe)2]
(III)
The vibrational properties of III are similar to those
of I. The Raman spectrum of III exhibits three isotope
sensitive peaks at 1893 shifting to 1830 on 15N-substitu-
tion, at 449 cmꢁ1 shifting by 22 to 427 cmꢁ1 and a weak
signal at 2017 shifting to 1949 cmꢁ1 (Fig. 13). The peak
at 1893 cmꢁ1 is assigned to n(NN) and the peak at 449
cmꢁ1 to n(WN). In the 14N compound, the n(WN) peak
overlaps with another signal which is the reason for the
appearance of two peaks in the 15N spectra (vide supra).
The features at 2017 cmꢁ1 in the 14N spectra and at
1949 cmꢁ1 in the 15N spectra are assigned to the
After QCAꢂNCA good agreement between experi-
/
mental and theoretical values is achieved as shown in
Table 11 The two NH stretching modes (symmetric and
antisymmetric) are fit by QCAꢂ
cmꢁ1 in the 14N compound and the 15N shifts are ꢁ
and ꢁ
5 cmꢁ1, respectively; somewhat smaller than
obtained experimentally (ꢁ20 and ꢁ
19 cmꢁ1). The
position of the 14N n(NN) stretch which is masked by a
/NCA to 3313 and 3230
/
11
/
/
/

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
25
Table 12
Calculated force constants for [W(N2)(NCEt)(PH3)4] (I?) and [W(N2H2)(NCEt)(PH3)4]2ꢃ (II?)
Force constant
Free N2 [7d]
22.42
Free CH3CN
[W(N2)(NCC2H5)(dppe)2]
[W(N2H2)(NCC2H5)(dppe)2]2ꢃ
Free NNH2 [30]
11.66
n(NN)
n(CN)
14.95
16.70
1.01
7.60
17.60
0.46
6.44
0.88
5.90
5.90
17.01
d(WNN)
n(WNN)
n(WNC)
nas(NH)
ns(NH)
3.15
1.53
5.89/4.61
ꢁ1
ꢁ1
˚
Units are mdyn/A
˚
for stretching and mdyn×A
for bending.
precursor, showing that the exchange reaction between
nitrile and N2 is not totally complete.
The IR spectrum of III (Fig. 14) is very similar to that
5.6. Vibrational spectra of
[W(N2H2)(NCPrn)(dppe)2][HSO4]2 (IV)
of I. Isotope sensitive signals are found at 1895 cmꢁ1
,
,
The Raman spectra of IV are obscured by fluores-
cence such that no spectroscopic information could be
shifting to 1826 on 15N substitution, and at 553 cmꢁ1
548 cmꢁ1. They are assigned to n(NN)
obtained. The IR spectra exhibit
a very broad
shifting by 5ꢂ
/
feature between 2700 and 3300 cmꢁ1 which is caused
by the counterion and overlaps with the n(NH) vibra-
and d(WNN), respectively. The features at 1583, 1093,
1026, 1000, 530 and 179 cmꢁ1 are assigned as in I. The
tions. These are observed at 3321 and 3215 cmꢁ1
,
nitrile Cꢀ
/
N stretch is found at 2171 cmꢁ1. The bands at
shifting to about 3309 (broad band) and 3194 cmꢁ1 on
15N isotope substitution, respectively. Other isotope
sensitive signals are found at 1604 cmꢁ1 shifting to
1601 cmꢁ1 and at about 440 cmꢁ1 showing a very small
shift, which are assigned to d(NNH) and d(WNN),
respectively. The n(WN) stretch is most probably
1944 and 2015 cmꢁ1 (14N) as well as the bands at 1955
and 1877 cmꢁ1 in the 15N spectra are caused by the
precursor. Frequencies and assignments are collected in
Table 14.
Fig. 13. Raman spectra of [W(N2)(NCPrn)(dppe)2] (III).

Table 13
f Matrix of the central CꢀNꢀWꢀNꢀNH2 unit in [W(NCCH3)(NNH2)(P4)]
n(NN)
n(WNC)
n(CN)
n(WNN)
n(NH)
n(NH)
dx(NNH)
dy(NNH)
dx(WNN)
g(NNHH)
dy(CNW)
dy(WNN)
t(MoN)
t(MoC)
f1
0
f2
f20
f21
0
0
f3
0
0
0
0
0
0
0
0
0
0
0
f15
f16
f17
f18
f19
0
f4
0
f5
0
0
f22
0
f6
f24
0
0
0
f7
f25
0
0
0
f23
0
f8
0
0
0
0
0
0
0
0
0
f9
0
0
0
0
0
0
0
f10
0
0
0
0
0
0
0
0
f11
0
0
0
0
0
0
0
0
0
f12
0
0
0
0
0
0
0
0
0
0
f13
0
0
0
0
0
0
0
0
0
0
0
f14

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
27
Fig. 14. IR spectra of [W(N2)(NCPrn)(dppe)2] (III).
Table 14
Observed
Table 15
Observed frequencies of [W(N2H2)(NCPrn)(dppe)2][HSO4]2
frequencies
[W(N2)(NCPh)(dppe)2]
of
[W(N2)(NCPrn)(dppe)2]
[W(N2)(NCPh)(dppe)2]
and
Mode
[W(N2N2)(NCC3H8)(dppe)2]
Mode
[W(N2)(NCPrn)(dppe)2]
14N (cmꢁ1 15N (cmꢁ1
14N (cmꢁ1
)
15N (cmꢁ1
)
)
)
14N (cmꢁ1
)
15N (cmꢁ1
)
nas(NH)
ns(NH)
n(CN)
3321
3215
3309
3194
a
a
a
a
n(CN)
n(NN)
2171/2171
1895/1893
2181/2171
1826/1830
548
2094
1919
540
2090
1855
a
a
2263/2263
1604
2265/2263
1601
d(WNN) 553
n(WNN) 449
Schulter 530
n.o.
ds(NNH)
n(WNN)
n(NN)
a
a
427
n.o.
Â590
n.o.
Â590
1384
n.o., not observed.
a
Raman-data, unlabeled data are IR-data.
d(WNN)
Â440
Â440
n.o., not observed.
a
Data from Raman, unlabeled data are IR-data.
ꢁ
masked by the broad band from the HSO4 counterion
at about 590 cmꢁ1. The n(CN) stretch is found
between 2263 and 2265 cmꢁ1 (see Table 15), showing
a shift to higher wavenumbers as compared with the free
nitrile.
assigned as in the case of compound I (see above).
Isotope sensitive signals were found at 1919, assigned as
n(NN), and at 540 cmꢁ1 (broad shoulder), assigned to
d(WNN). They shift by 64 cmꢁ1 (n(NN)) and 11 cmꢁ1
(d(WNN)) to lower wavenumbers on 15N substitution,
respectively (cf. Table 15). The band at 3049 cmꢁ1 is
assigned to n(CH), that at 2090 cmꢁ1 to n(CN). Bands
at 1582, 1088 and 1026 cmꢁ1 are assigned as in
compound I. The features at 1951 (14N) and 1886
cmꢁ1 (15N) correspond again to the NN- stretch of
the precursor (vide supra).
5.7. Vibrational spectra of [W(N2)(NCPh)(dppe)2]
(V)
This compound loses dinitrogen on irradiation with
laser light and thus only IR data could be obtained (Fig.
15). Intense features are found in the 14N spectra at
3049, 2094, 1951, 1582, 1088 and 1026 cmꢁ1 which are

28
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
Fig. 15. IR spectra of [W(N2)(NCPh)(dppe)2] (V).
6. Discussion and conclusions
is more favorable in the trans nitrilo complex. Likewise,
N stretching frequency in the trans nitrilo system
[W(N2)(NCEt)(dppe)2] (I) of 1895 cmꢁ1 has a lower
value than that in the bis(dinitrogen) complex
[W(N2)2(dppe)2] (0) (2007 cmꢁ1 for the symmetric and
1954 cmꢁ1 for the antisymmetric combination of the
the Nꢀ
/
In the preceding sections the electronic and vibra-
tional structure descriptions of dinitrogen and diaze-
nido(2-) complexes with trans nitrile ligands have been
developed. Due to the fact that protonation of the
dinitrogen complex with a trans nitrile group to the
corresponding N2H2 complex proceeds under retention
of this axial ligand, these systems appear to be particu-
larly attractive candidates for the metal-centered cata-
lysis of N2 fixation. In contrast, one of the two
dinitrogen ligands in the usual bis(dinitrogen) Mo and
W systems is exchanged by a Lewis base deriving from
the acid used for protonation, leading to a trans halide,
Nꢀ
/
N stretches). This is also reflected by the obtained
force constants: for the nitrilo complex, the QCAꢂNCA
/
procedure gives an fNN value of 14.95 mdyn Aꢁ1 which
˚
is lower than 16.43 mdyn A
˚
ꢁ1
calculated for the
corresponding bis(dinitrogen) complex. Both, the trans-
fer of electronic charge to Nb and the lowering of the
Nꢀ/N force constant with respect to the free ligand are
due to backbonding into the p* orbitals of dinitrogen. In
agreement with the above findings, this interaction is
found to be stronger in the DFT calculations for the
single N2 group in the trans nitrilo system than for each
of the N2 ligands in the bis(dinitrogen) complex. The
stronger activation of the single N2 group in I is also
hydrogen sulfate or sulfonic acid Mo/Wꢀ
/N2H2 complex
[5]. This exchange of the axial ligand is unfavorable with
respect to the realization of a catalytic cycle for nitrogen
fixation (cf. Section 1).
Most importantly, the activation of the single N2
group bound in trans position to nitrile is found to be
higher than that of each of the two N2 ligands in the
corresponding bis(dinitrogen) system. This is evidenced
by an increased electronic charge on the Nb atom of the
N2 group in the trans nitrilo as compared with the trans
evident from the shorter Wꢀ
/
N distance found in the
˚
˚
crystal structure (1.983 A in I vs. 2.004 A in 0).
To a certain extent, the nitrile group competes with
the trans dinitrogen ligand for back donation from the
Mo/W(0) center: for all trans nitrilo dinitrogen systems,
bis(dinitrogen) complex (ꢁ
sequence, the protonation leading to the N2H2 complex
/
0.12 vs. ꢁ
/
0.06). As a con-
the Cꢀ
/
N stretching frequency is found to be lower than
in the corresponding free nitrile (EtCN: 2190 vs. 2249

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
29
cmꢁ1; iPrCN: 2171 vs. 2245 cmꢁ1, PhCN: 2094 vs. 2231
cmꢁ1) [8]. Nevertheless, the nitriles are weaker p
acceptors than N2. This is evidenced by a smaller
lowering of the stretching frequency of the coordinated
nitrile with respect to the free ligand (lowering of n(CN)
corresponding 14Nꢀ
/
14N stretch is obscured in the nitrilo
complex), indicating a lower degree of backbonding.
The origin for the lower activation of NNH2 trans to
nitrile as compared with trans halide is the fact that
nitrile is a neutral ligand while a trans halide ligand is
able to donate additional electron density into the Mo/
60ꢂ
n(NN) upon coordination over 400 cmꢁ1). Likewise,
there is less negative charge on the CꢀN as compared
with the NꢀN group (ꢁ0.05 vs. ꢁ0.22). Remarkably,
/
75 cmꢁ1) than in the case of dinitrogen (lowering of
Wꢀ
ligand is assigned an NPA charge of ꢁ
transferred 0.45 charge units to the Mo/W complex. In
contrast, the CꢀN group of the nitrile ligand acquires a
total charge of ꢃ0.06 at the NNH2 stage, practically the
/
NNH2 moiety. In the trans fluoro complex, the F
/
/0.55, i.e. has
/
/
/
the lower extent of backdonation is not compensated by
stronger s donation from nitrile; in contrast, this ligand
is also found to be a weaker s donor than N2. This
probably derives from the fact that s donation from the
nitrile ps orbital is limited due to the bonding of a
positively charged C atom (see below). Taken together,
these observations indicate that the bonding of nitriles
to the Mo/W(0) center is much weaker than that of the
/
/
same value as in the N2 complex (vide supra). Hence, in
the case of nitrile there is no synergetic effect to the
protonation process due to donation from the trans
ligand. Nevertheless the example of the nitrile systems
demonstrates that protonation of an N2 complex is
possible without exchange of the trans axial ligand.
It has been found by 15N NMR spectroscopy that the
shielding of Na and Nb in the NNH2 ligand as a function
N2 group. The lability of the metalꢂnitrile bond is also
/
evidenced by the thermally allowed transformation of
the trans nitrile systems into their bis(dinitrogen)
counterparts. On the other hand, the lower extent of
backdonation into the nitrile p* orbitals, is the origin for
the higher activation found for the single dinitrogen
group in the trans nitrilo complex as compared with that
of each of the N2 ligands in the bis(dinitrogen) system.
The dinitrogen complex with a trans PhCN ligand has
a weaker activated N2 group than the corresponding
trans alkyl nitrile systems. This is evident from the
of the trans ligand decreases in the sequence halideꢀ
/
HSO4ꢁ ꢀNCPrn [10]. The weaker shielding (or stronger
/
ꢁ
deshielding) of NNH2 in the trans HSO4 complex as
compared with trans halide systems has been suggested
to be associated with a movement of electronic charge
towards the trans NNH2 group which then should be
more susceptible to protonation than in the case of the
trans halide systems [25]. From the above trend of the
Na and Nb chemical shifts, this should be even more
pronounced for the trans NCPrn complex, i.e. this
complex should have the highest activated NNH2
ligand. Our results indicate that the opposite is the
case. Therefore, the shielding of Na and Nb and
activation of the NNH2 ligand have to go in parallel.
This would also be in agreement with the stronger
shielding of Na and Nb observed for W than for Mo/N2
higher Nꢀ
a slightly lower electronic charge calculated for the N2
group. Nevertheless, the lowering of the CꢀN stretching
/
N stretching frequency (1919 cmꢁ1) and from
/
frequency upon coordination of the free ligand to the
metal center is largest for this group (137 vs. about 70
cmꢁ1 for the alkyl nitriles; vide supra). The electronic
reason for these observations is the delocalization of
electronic charge from the d orbitals of the metal into
the p* orbitals of the aromatic nitrile ligand which acts
to diminish the charge transfer to the N2 group. The
lower activation of the N2 ligand trans to the PhCN
group as compared with systems with trans alkyl nitrile
ligands is in agreement with trends obtained from
electrochemical oxidation potentials [9].
The second important result of the spectroscopic
investigations and DFT calculations is the fact that the
activation of the NNH2 group with a trans nitrilo ligand
is lower than in corresponding NNH2 systems having
trans halide ligands. This is evidenced by the positive
and
ꢀ
/
N2H2 complexes corresponding to a higher
activation of NNH2 and N2, respectively [10,26]. As a
consequence, however, the activation of NNH2 in the
trans HSO4 complex would be lower than in trans
ꢁ
halide systems. This would correspond to the weaker
Lewis basicity of HSO4ꢁ as compared with halide, again
reflecting the influence of the donor properties of the
trans ligand on the activation of the N2H2 group.
Remarkably, the bonding of the nitrile group which
has been found to be weak in the case of the N2 complex
(vide supra), becomes even weaker in the NNH2
complex. The calculations show that nitrile is in fact a
very weak ligand at this stage of N2 reduction. There is
no backbonding into nitrile p* orbitals any more.
Neither is nitrile bound by a significant s donor
charge on this group in the NNH2ꢂ
(ꢃ0.20), while the corresponding trans fluoroꢂ
system has an almost neutral NNH2 ligand (charge
0.03). Clearly, the hydrazido(2-) formulation is even
/nitrile complex
/
/NNH2
ꢁ
/
interaction to the metal center. This is illustrated by
˚
N(nitrile) bond distance (2.275 A) and the
less appropriate for the description of the NNH2 group
in the trans nitrile than in the trans fluoro complex. In
the large Moꢀ
/
˚
short Cꢀ
free nitrile). In the N2H2 complex there is a further
increase of the polarization of the CꢀN group which
already has been strong in the trans dinitrogen complex.
/
N bond distance (calc. 1.175 vs. 1.189 A in the
addition, the Nꢀ
nitriloꢀ
NNH2 (n(15Nꢀ
the fluoro NNH2 system (n(15Nꢀ
/
N stretching frequency is higher for the
15N): 1384 cmꢁ1) compared with
15N): 1343 cmꢁ1; the
/
/
/
/

30
C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ31
/
However, the negative charge at the nitrile N atom is not
transferred to the metal to a significant degree. A
possible explanation for this observation is provided
by the fact that the nitrogen donor of the nitrile group is
bound to a positively charged C atom and hence, cannot
act as a strong donor.
ligands). Nevertheless, due to the weak covalency at
intermediate stages of N2 reduction the electron transfer
properties of this group may also become poor during
the catalytic cycle. As a consequence, it appears more
promising to use a connection via the phosphine groups
to the electrode to mediate electron transfer to com-
plexes like I.
This bonding description is also supported by the
results of vibrational spectroscopy. In contrast to the
If Mo/W phosphine complexes with axial nitrile
ligands are to be used for the homogeneous catalysis
of N2 reduction, it also has to be taken into account that
the nitrile ligand is potentially labile during most of the
catalytic cycle. In spite of this drawback, it may be
meaningful to use this ligand in trans position to the N2
group as it both stabilizes the Mo/W(0) oxidation state
and activates the N2 group to the first (and most
difficult) protonation step; this is not possible with
ligands having p donor interactions [28]. As a conse-
quence, the nitrile ligand has to be further attached to
the metal phosphine complex in order to prevent (i) loss
of this ligand during the catalytic cycle and/or (ii)
exchange against a Lewis base deriving from the acid
used for protonation. If this can be achieved syntheti-
cally, the trans position of the N2 group would
effectively be blocked, i.e. protected from attack by an
external ligand. This way, the unfavorable ligand
exchange reactions at the trans position of N2 could in
fact be avoided.
trans dinitrogen complexes where the Cꢀ
/
N stretching
frequencies have been found to be lowered with respect
to the free ligands (vide supra), they are shifted to higher
values in the NNH2 system. This increase of the Cꢀ
stretching frequency which is also observed for other
nitrile complexes [8] is obviously correlated to a short-
/
N
ening of the Cꢀ
/
N bond occurring as a result of the
increase of polarization of the Cꢀ
/N unit upon bonding
to a metal center. Only in the case of coordination to an
electron-rich metal, this effect is overcompensated by
backbonding leading to a lowering of the Cꢀ
ing frequency with respect to the free ligand, as observed
in the case of the Mo/W(0) nitrile systems.
This finding bears an interesting relationship to the
theoretical results of Frenking et al. on the properties of
CO coordinated to metal centers [27]. Here, the Cꢀ
triple bond also becomes shorter when the metal cation
approaches the carbon atom; only at the onset of
/N stretch-
/O
metal0
/
CO back donation this trend is reversed, and
the CꢀO bond becomes longer again. With respect to
/
the polarization of the coordinating group, however,
there is a remarkable difference between CO and RCN:
upon coordination of a metal or Hꢃ to the C atom of
CO, the polarization between C and O is found to
decrease whereas in the case of a metal coordinating to
the N atom of the nitrile ligand, the polarization
between N and C is found to increase (vide supra).
This difference is certainly due to the fact that in the case
of CO the less electronegative atom acts as a s donor to
the metal while in case of RCN this function is located
at the more electronegative atom.
7. Supplementary material
Tables of details of the structure determination,
atomic coordinates, isotropic and anisotropic displace-
ment parameters as well as bond lengths and angles for I
are available upon request.
Acknowledgements
What are the consequences of the above results with
respect to the rational design of metal complexes able to
catalyze N2 fixation? As a first application, nitrile
groups have been used to attach Mo/W dinitrogen
complexes with equatorial phosphine ligands to electro-
des [11]. In this geometry, the nitrile ligand has two
functions: it first acts to fix the complex mechanically to
the surface of the electrode and secondly acts as a
C.H. thanks U. Cornelissen for help with the spectro-
scopic measurements and N. Bores for help with
¨
preparative work. F.T. thanks the Fonds der Che-
mischen Industrie (FCI) for financial support.
References
molecular wire, i.e. funnels electrons into the Mo/Wꢀ
/
N2
[1] (a) F. Tuczek, N. Lehnert, Angew. Chem. 110 (1998) 2780;
(b) R.L. Rawls, Chem. Eng. News 76 (1998) 29;
system. However, since the bonding of nitrile has been
found to become weak at an intermediate stage of N2
reduction, the connection to the electrode established
this way may become labile during the catalytic cycle
and the complex may dissociate away from the surface.
As a consequence, it appears to be necessary to
additionally fix the Mo/W phosphine system to the
surface (e.g. by linkers deriving from the phosphine
(c) M. Fryzuk, S.A. Johnson, Coord. Chem. Rev. 200 (2000) 379;
(d) D. Sellmann, J. Utz, N. Blum, F.W. Heinemann, Coord.
Chem. Rev. 190ꢂ192 (0) (1999) 607;
/
(e) R.L. Richards, Coord. Chem. Rev. 154 (1996) 83;
(f) M. Hidai, Y. Mizobe, Chem. Rev. 95 (1995) 1115;
(g) G.J. Leigh, Accts. Chem. Res. 25 (1992) 177.
[2] (a) E.W. Triplett (Ed.), Prokaryotic Nitrogen Fixation: A Model
System for the Analysis of a Biological Process, Horizon Scientific

C.M. Habeck et al. / Inorganica Chimica Acta 337 (2002) 11ꢂ
/31
31
Press, Wymondham, UK, 2000;
[9] J. Chatt, G.J. Leigh, H. Neukomm, C.J. Pickett, D.R. Stanley, J.
Chem. Soc., Dalton Trans. (1980) 121.
(b) B.K. Burgess, D.J. Lowe, Chem. Rev. 96 (1996) 2983;
(c) J. Kim, D.C. Rees, Science 257 (1992) 1667;
(d) J. Kim, D.C. Rees, Nature 360 (1992) 553;
(e) M.K. Chan, J. Kim, D.C. Rees, Science 260 (1993) 792;
(f) J. Chen, J. Christiansen, N. Campobasso, J.T. Bolin, R.C.
Tittsworth, B.J. Hales, J.J. Rehr, S.P. Cramer, Angew. Chem. 105
(1993) 1661;
[10] S. Donovan-Mtunzi, R.L. Richards, J. Chem. Soc., Dalton Trans.
(1984) 1329.
[11] G.J. Leigh, C.J. Pickett, J. Chem. Soc., Dalton Trans. (1977)
1797.
[12] J.R. Dilworth, R.L. Richards, Inorg. Synth. 20 (1980) 119.
[13] G.M. Sheldrick, SHELXS-97 and SHELXS-97, in: Program for the
Solution and Refinement of Crystal Structures, University of
(g) J.B. Howard, D.C. Rees, Chem. Rev. 96 (1996) 2965;
(h) J.W. Peters, M.H.B. Stowell, S.M. Soltis, M.G. Finnegan,
M.K. Johnson, D.C. Rees, Biochemistry 36 (1997) 1181;
(i) J.T. Bolin, A.E. Ronco, T.V. Morgan, L.E. Mortenson, N.H.
Xuong, Proc. Natl. Acad. Sci. USA 90 (1993) 1078.
[3] (a) H. Deng, R. Hoffmann, Angew Chem. 105 (1993) 1125;
(b) W. Plass, J. Mol. Struct., Theochem. 315 (1994) 53.
[4] (a) M. Jimenez-Tenorio, M.C. Puerta, P. Valerga, D.L. Hughes, J.
Chem. Soc., Dalton Trans. (1994) 2431;
Gottingen, Gottingen, Germany, 1997.
¨ ¨
[14] J.A. Nelder, R. Mead, Comput. J. 7 (1965) 308.
[15] T.H. Dunning, Jr., P.J. Hay, in: H.F. Schaefer, III (Ed.), Modern
Theoretical Chemistry, Plenum Press, New York, 1976.
[16] (a) P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 270 and 299;
(b) W.R. Wadt, P.J. Hay, J. Chem. Phys. 82 (1985) 284.
[17] (a) J.P. Foster, F. Weinhold, J. Am. Chem. Soc. 102 (1980) 7211;
(b) A.B. Rives, F. Weinhold, Int. J. Quantum Chem. Symp. 14
(1980) 201;
(b) T.A. George, B.B. Kaul, Q. Chen, J. Zubieta, Inorg. Chem. 32
(1993) 1706;
(c) A.E. Reed, R.B. Weinstock, F. Weinhold, J. Chem. Phys. 83
(1985) 735;
(c) T.A. George, L. Ma, S.N. Shailh, R.C. Tisdale, J. Zubieta,
Inorg. Chem. 29 (1990) 4789;
(d) A.E. Reed, L.A. Curtiss, F. Weinhold, Chem. Rev. 88 (1988)
899.
(d) A. Galindo, A. Hills, D.L. Hughes, D.L. Richards, M.
Hughes, J. Mason, J. Chem. Soc., Dalton Trans. (1990) 283;
(e) J.E. Barclay, A. Hills, D.L. Hughes, G.J. Leigh, C.J.
Macdonald, M. Abu Bakar, H. Mohd-Ali, J. Chem. Soc., Dalton
Trans. (1990) 2503;
[18] GAUSSIAN 98, Revision A.7, M.J. Frisch, G.W. Trucks, H.B.
Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G.
Zakrzewski, J.A. Montgomery, Jr., R.E. Stratmann, J.C. Burant,
S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C.
Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B.
Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A.
Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D.
Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V.
Ortiz, A.G. Baboul, B.B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T.
Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonza-
lez, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W.
Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S.
Replogle, J.A. Pople, Gaussian, Inc., Pittsburgh, PA, 1998.
[19] G. Schaftenaar, MOLDEN version 3.2; CAOS/CAMM Center,
University of Nijmegen, Nijmegen, The Netherlands.
[20] A. Allouche, J. Pourcin, Spectrochim. Acta 49A (1993) 571.
[21] C. Pelizzi, G. Pelizzi, Acta Crystallogr., Sect. B 35 (1979) 1785.
[22] (a) T. Wagner, U. Englert, Struct. Chem. 8 (1997) 357;
(b) C. Hue, W.C. Hodgeman, D.W. Benett, Inorg. Chem. 35
(1996) 1621.
(f) M. Abu Bakar, D.L. Hughes, W. Hussain, G.J. Leigh, C.J.
Macdonald, H. Mohd-Ali, J. Chem. Soc., Dalton Trans. (1988)
2545;
(g) T.A. George, R.C. Tisdale, J. Am. Chem. Soc. 107 (1985)
5157;
(h) J. Chatt, M.E. Fakley, P.B. Hitchcock, R.L. Richards, N.T.
Luong-Thi, J. Chem. Soc., Dalton Trans. (1982) 345;
(i) T. Takahasi, Y. Mizobe, M. Sato, Y. Uchida, M. Hidai, J. Am.
Chem. Soc. 102 (1980) 7461;
(j) J. Chatt, A.J. Pearman, R.L. Richards, J. Chem. Soc., Dalton
Trans. (1978) 1766;
(k) J. Chatt, A.J. Pearman, R.L. Richards, J. Chem. Soc., Dalton
Trans. (1977) 2139;
(l) J. Chatt, A.J. Pearman, R.L. Richards, J. Chem. Soc., Dalton
Trans. (1977) 1852;
(m) J. Chatt, A.J. Pearman, R.L. Richards, J. Chem. Soc., Dalton
Trans. (1976) 1520;
[23] M. Hidai, T. Kondama, M. Sato, M. Harakawa, Inorg. Chem. 15
(1976) 2694.
(n) G.A. Heath, R. Mason, K.M. Thomas, J. Am. Chem. Soc. 96
(1974) 260;
[24] N.B. Colthup, L.H. Daly, S.E. Wiberley, Introduction to Infrared
and Raman Spectroscopy, third ed., Academic Press, Orlando,
FL, 1990.
(o) J. Chatt, G.A. Heath, R.L. Richards, J. Chem. Soc., Dalton
Trans. (1974) 2074;
(p) J. Chatt, G.A. Heath, R.L. Richards, J. Chem. Soc., Chem.
Commun. (1972) 1010.
[25] (a) J.R. Dilworth, R.L. Richards, in: G. Wilkinson, F.G.A. Stone,
E.W. Abel (Eds.), Comprehensive Organometallic Chemistry, vol.
60 (and references therein), Pergamon Press, Oxford, 1982, p.
1073;
[5] R.N.F. Thorneley, D.J. Lowe, in: T.G. Spiro (Ed.), Molybdenum
Enzymes, Wiley, New York, 1985.
[6] R.A. Henderson, G.J. Leigh, C. Pickett, Adv. Inorg. Chem.
Radiochem. 27 (1983) 197.
(b) J. Chatt, R.L. Richards, J. Organomet. Chem. 239 (1982) 65
(and references therein).
[7] (a) N. Lehnert, B. Wiesler, F. Tuczek, A. Hennige, D. Sellmann,
J. Am. Chem. Soc. 119 (1997) 8869;
[26] S. Donovan-Mtunzi, R.L. Richards, J. Chem. Soc., Dalton
Trans. (1984) 469.
(b) N. Lehnert, B. Wiesler, F. Tuczek, A. Hennige, D. Sellmann,
J. Am. Chem. Soc. 119 (1997) 8879;
[27] A.J. Lupinetti, S. Fau, G. Frenking, S.H. Strauss, J. Phys. Chem.
A 101 (1997) 9558.
[28] O. Franke, B.E. Wiesler, N. Lehnert, C. Nather, V. Ksenofontov,
(c) E.B. Wiesler, N. Lehnert, F. Tuczek, J. Neuhausen, W.
Tremel, Angew. Chem. 110 (1998) 856(Int. Ed. 37 815);
(d) N. Lehnert, F. Tuczek, Inorg. Chem. 38 (1999) 1659;
(e) N. Lehnert, F. Tuczek, Inorg. Chem. 38 (1999) 1671.
[8] T. Tatsumi, M. Hidai, Y. Uchida, Inorg. Chem. 14 (1975) 2530.
¨
J. Neuhausen, F. Tuczek, Inorg. Chem. (2002), in press.
[29] K.H. Horn, F. Tuczek, in preparation.
[30] H.J.A. Jensen, P. Jørgensen, T. Helgaker, J. Am. Chem. Soc. 109
(1987) 2895.