The reaction of M(CO)3(Ph2PCH2CH 2PPh2) (M = Fe, Ru) with parahydrogen: Probing the electronic structure of reaction intermediates and the internal rearrangement mechanism for the dihydride products

Article abstract of DOI:10.1039/b407457b

The photochemical reaction of Ru(CO)₃(dppe) and Fe(CO) ₃(dppe) (dppe = Ph₂PCH₂CH₂PPh ₂) with parahydrogen has been studied by in situ-photochemistry resulting in NMR spectra of Ru(CO)₂(dppe)(H)₂ that show significant enhancement of the hydride resonances while normal signals are seen in Fe(CO)₂(dppe)(H)₂. This effect is associated with a singlet electronic state for the key intermediate Ru(CO)₂(dppe) while Fe(CO)₂(dppe) is a triplet. DFT calculations reveal electronic ground states consistent with this picture. The fluxionality of Ru(CO) ₂(dppe)(H)₂ and Fe(CO)₂(dppe)(H)₂ has been examined by NMR spectroscopy and rationalised by theoretical methods which show that two pathways for ligand exchange exist. In the first, the phosphorus and carbonyl centres interchange positions while the two hydride ligands are unaffected. A second pathway involving interchange of all three ligand sets was found at slightly higher energy. The H-H distances in the transition states are consistent with metal-bonded dihydrogen ligands. However, no local minimum (intermediate) was found along the rearrangement pathways.

Full text of DOI:10.1039/b407457b

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
The reaction of M(CO) (Ph PCH CH PPh ) (M = Fe, Ru)
3
2
2
2
2
with parahydrogen: probing the electronic structure of reaction
intermediates and the internal rearrangement mechanism for the
dihydride products†
a
a
a
a
a
Danièle Schott, Philip Callaghan, John Dunne, Simon B. Duckett,* Cyril Godard,
d
b
a
a
a
José M. Goicoechea, Jeremy N. Harvey, John. P. Lowe, Roger J. Mawby, Georg Müller,
Robin N. Perutz, Rinaldo Poli and Michael K. Whittlesey
Department of Chemistry, University of York, UK YO10 5DD. E-mail: sbd3@york.ac.uk
Department of Chemistry, University of Bristol, UK BS8 1TS
Laboratoire de Chimie de Coordination, 205 Route de Narbonne, 31077, Toulouse Cedex,
France
a
c
d
a
b
c
d
Department of Chemistry, University of Bath, Claverton Down, Bath, UK BA2 7AY
Received 19thaMꢀyy20ꢁcceA9ed29tꢁꢂ4ꢂꢃ9ꢀyy2
Firꢃ9Aꢂbꢄiꢃtedaꢃasꢁdvasceꢁr9icꢄeꢅs9teꢆebꢀꢇ9tꢁꢂ4ꢂꢃ9ꢀyy2
The photochemical reaction of Ru(CO)
has been studied by in situ-photochemistry resulting in NMR spectra of Ru(CO)
enhancement of the hydride resonances while normal signals are seen in Fe(CO)
with a singlet electronic state for the key intermediate Ru(CO) (dppe) while Fe(CO)
calculations reveal electronic ground states consistent with this picture. The fluxionality of Ru(CO)
and Fe(CO) (dppe)(H) has been examined by NMR spectroscopy and rationalised by theoretical methods which
3
(dppe) and Fe(CO)
3
(dppe) (dppe = Ph
2
PCH
(dppe)(H)
(dppe)(H)
(dppe) is a triplet. DFT
(dppe)(H)
2
CH
2
PPh
2
) with parahydrogen
2
2
that show significant
2
. This effect is associated
2
2
2
2
2
2
2
show that two pathways for ligand exchange exist. In the first, the phosphorus and carbonyl centres interchange
positions while the two hydride ligands are unaffected. A second pathway involving interchange of all three ligand
sets was found at slightly higher energy. The H–H distances in the transition states are consistent with metal-bonded
dihydrogen ligands. However, no local minimum (intermediate) was found along the rearrangement pathways.
support triplet ground states for these species.8,9 Understanding
how such electronic changes influence a metal centre’s potential
to cleave the C–H bond of a alkane remains an important goal
in C–H bond functionalisation.10
Introduction
Reaction mechanisms implicating intermediates with two or
more electronic states close in energy are increasingly common.1
They include reactions involving carbenes and nitrenes as well as
their isolobal transition metal analogues. In view of the differing
rates of singlet–singlet and singlet–triplet combinations, the
existence of the two spin states can affect product distributions
significantly. It is also possible that while both reagents and
products have the same electronic spin state, spin crossover
occurs during the transformation and a highly complex reaction
pathway is followed. Although breakthroughs have been made
in the spectroscopic examination of reaction intermediates,
their short lifetime makes detailed study experimentally
challenging. One reaction that has been studied in great detail
In many ways, NMR spectroscopy is an ideal technique for
studying solution based reactions, since most elements exhibit
at least one NMR-active isotope, and isotopic labelling can
be undertaken if greater sensitivity is required. Nevertheless,
NMR spectroscopy remains a technique with low sensitivity
that is hard to adapt to the study of short-lived species. One
route that increases the sensitivity of the NMR experiment
1
1
involves the use of parahydrogen (p-H
observation of substantially enhanced signals for nuclei that
originate in the p-H molecule or couple to these nuclei. The
effect has been called parahydrogen induced polarisation
PHIP) or PASADENA (parahydrogen and synthesis allow
2
). This leads to the
2
2
–4
is the oxidative addition of H
has been shown that rate of H
2
to Fe(CO)
2
4
.
Experimentally it
is
(
addition to triplet Fe(CO)
4
dramatically enhanced nuclear alignment) and has been
three orders of magnitude slower than that to related singlet
extensively reviewed.1
1–15
5
states. This phenomenon has been attributed to the spin
In this paper we employ the parahydrogen effect in conjunction
with in-situ photolysis16 to study reaction intermediates.
Several groups have demonstrated how UV photolysis of an
forbidden character of the singlet–triplet combination. The
chemistry of the group 6 metallocences, (g -C
ansa-bridged analogues that can be generated photochemically
from the corresponding dihydrides or methyl hydrides is
6
5
5 5 2
H ) W and their
NMR sample within the NMR probe can be used to generate
6
a
6b
new compounds.1
6,17
Reports also describe how this method
7
also dominated by singlet/triplet effects. A further example
enables species to be characterised that might normally be
considered reaction intermediates by conducting experiments at
temperatures low enough to extend their lifetime substantially.16
Here, we use a 325 nm He–Cd continuous wave laser to generate
intermediates that yield metal dihydride complexes upon
5
is provided by (g -C
5
H
5
)Co(CO), formed by photolysis of
5
8
(
g -C
5 5 2
H )Co(CO) . This species is a spin triplet whereas its
rhodium and iridium analogues have singlet ground states.
Remarkably, triplet (g -C
5
5 5
H )Co(CO), proves to be the most
reactive species towards Si–H activation. Theoretical analyses
1
8
2 2
reaction with H . When p-H is used the electronic structure
5
5
5 5 5 5 3 2 4
of (g -C H )Co(CO) and (g -C H )CoL (L = PH and C H )
of the reaction intermediate is shown to be indicated by key
features of the resultant NMR spectra of the reaction product,
even though the reaction intermediate itself is not directly
observed. Density functional theory (DFT) is used to confirm
the electronic structure of these intermediates and map the
reaction pathways.
†
Electronic supplementary information (ESI) available: Cartesian
coordinates for all optimised geometries, synthesis and NMR
parameters for Fe(CO) (dppe)(H) and rate data for ligand interchange.
2
2
See http://www.rsc.org/suppdata/dt/b4/b407457b/
3
2 1 8
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

View Article Online
Results and discussion
In order to carry out a definitive experiment, the analogous
iron complex Fe(CO)
uously photolysed at 203 K with p-H
spectra were recorded. The immediate generation of the known
dihydride Fe(CO) (dppe)(H) 2 was indicated by the observation
of two hydride resonances as doublets of doublets of doublets
3
(dppe) was prepared and a sample contin-
Reaction of p-H
2
with Ru(CO)
2
(dppe) and Fe(CO)
2
(dppe)
2
whilst appropriate NMR
We have previously reported that Ru(CO)
3
(dppe) (dppe = Ph -
2
PCH CH PPh ) reacts with p-H under photolytic conditions at
95 K (Scheme 1) to yield enhanced signals for the corresponding
2
2
2
2
2
2
2
1
8
2
2
dihydride Ru(CO)
2
(dppe)(H)
2
1 (Fig. 1(a)). This observation
centred at d −8.8 (| JHH| = 14.5 Hz, | JPH| = 28.5 Hz and 56.4 Hz)
2
2
20
can be reproduced at 203 K. Since p-H
2
corresponds to H in
2
and d −10.24 (| JHH| = 14.5 Hz, | JPH| = 35.4 Hz and 57.4 Hz).
These signals had normal phase profiles (Fig. 1(b)), and no evi-
dence for p-H amplification was evident. A solution containing
a mixture of Ru(CO) (dppe) and Fe(CO) (dppe) under 3-atm of
p-H was also prepared and irradiated in CD CN at 243 K. The
the antisymmetric nuclear spin state (ab–ba), any reaction that
leads to a product in which this spin encoding is retained, will
yield NMR signals that are derived from a non-Boltzmann spin
population. When this process is described using the product
2
3
3
2
3
operator formalism, the resultant I
z
S
z
product state leads to
hydride signals for the ruthenium dihydride 1 were still enhanced,
while those for 2 again showed normal shape. The most logical
explanation for this effect requires different electronic spin states
at the metal. We changed to the more coordinating solvent
because we expected the Fe reaction to proceed associatively
I
z
S
x
and I terms upon interrogation. Each of these terms
x z
S
corresponds to an anti-phase signal, one for the I spin and one
for the S spin, which are separated by JIS which in this case
corresponds to JHH. However, when the related CO labelled
1
3
1
3
iron complex Fe( CO)
5
was photolysed with p-H
2
no enhanced
(H) . This
result is surprising since the C label forces the hydride ligands
to belong to a second-order spin system and p-H activity is
expected. Such an enhancement has already been reported by
via the solvent complex Fe(CO)
2 3
(dppe)(CD CN) which was
1
3
hydride resonances were observed for Fe( CO)
4
2
expected to be a singlet in CD CN. We therefore expected that
3
1
3
2 would be formed through a singlet-based pathway and hence
polarisation would be visible in the hydride resonances of 2.
Although parahydrogen induced polarisation was observed for
1, the signals for 2 were unaffected, and fast depletion of the
parahydrogen reservoir was observed. This suggests that these
2
1
3
ourselves when cis-trans-cis-Ru( CO)
thermally driven H exchange with p-H
accepted that the unsolvated 16-electron intermediate Fe(CO)
2
(PMe
3
)
2
(H)
2
undergoes
1
8
2
2
.
It is generally
2 2
reactions still involve triplet Fe(CO) (dppe) reacting with H .
4
adopts a triplet ground state while related Ru(0) species are
expected to have singlet ground states. We therefore postulated
2
Characterisation of Fe(CO)
2
2
(dppe)(H) (2)
that quenching of the p-H
addition of H to Fe(CO) and that this accounts for the failure
to detect any signal enhancement. This simple observation is,
however, complicated by the known fluxionality of FeL (H)
L = two-electron donor) systems which could also quench the
2
spin state would occur during the
In order to characterise Fe(CO)
2
(dppe)(H)
2
(2) a series of NMR
spectra were recorded at 193 K. The P{ H} NMR spectrum of
2
4
31
1
2
9
contains a pair of mutually coupled resonances at d 92.4 and
7.8 (JP–P = 16 Hz), due to two inequivalent phosphorus centres
4
2
(
in the product. An assignment of the relative positions of the
phosphorus ligands was made on the basis of NOE data and
decoupling experiments as shown in Fig. 2. It is interesting to
note that the phosphorus atom that is cis to both hydride ligands
yields the larger hydride phosphorus couplings. This differs from
the situation at ruthenium where the trans hydride–phosphorus
coupling is larger than a cis coupling, reinforcing the need
for care to be taken when ligand arrangements are based on
coupling data. A full description of the methods used in the
characterisation can be found in the ESI.†
1
9
antiphase signal by excessive line broadening.
Scheme 1
2 2
Fig. 2 Representation of the structure for Fe(CO) (dppe)(H) 2
illustrating key NMR features.
Theoretical studies on the fragments Fe(CO)
Ru(CO) (dpe)
2
(dpe) and
2
Recently, some of us have shown that 16-electron mixed
phosphine and carbonyl Fe(0) species of the type Fe(CO)
2
(PH
3
)
4
2
3
do indeed have a triplet ground state like those of Fe(CO)
2
1
and Fe(PR
3
)
4
.
However, the use of chelating phosphines is
known to change the reactivity of metal fragments dramatically
because of the P–M–P bond angles that result. An appropriate
theoretical study was therefore undertaken. The importance of
this study was heightened by the fact that the electronic structure
of the related complex, cis-cis-Fe(CO)
in the singlet even though the lowest energy trans-cis isomer
Fe(CO) (PH corresponded to a triplet.
Calculations were carried out on model complexes contain-
ing the PH CH CH PH (dpe) ligand in place of dppe at the
same level of theory previously employed for Fe(CO)
2 3 2
(PH ) , only optimised
2
3 2
)
2
2
2
2
Fig. 1 Hydride signals observed by in situ UV photolysis of
M(CO) (dppe) in the presence of 3 atm of p-H : (a) M = Ru,
b) M = Fe.
2
2
2 3 2
(PH ) ,
3
2
(
since a benchmark investigation had shown this to give the
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4
3 2 1 9

View Article Online
Table 1 Calculated bond lengths (Å) and angles (°) for all species using the B3PW91* functional
Compound
M–Lax
M–Leq
HH
Lax–M–Lax
Leq–M–Leq
H–M–H
H
Fe(CO)
2
—
2.213 (P)
—
0.748
—
—
176.9
—
133.7
—
—
1
2
2
(dpe)
(dpe)
2.160 (P)
1.733 (C)
2.240 (P)
1.786 (C)
2.272 (P)
1.782 (C)
2.186 (P)
1.769 (C)
1
.755 (C)
2.271 (P)
.805 (C)
2.307 (P)
.816 (C)
2.171 (P)
.751 (C)
3Fe(CO)
—
156.0
157.1
157.0
98.4
97.4
—
1
1/3Fe(CO)
(MECP)
0.775
2.000
0.868
22.2
82.4
31.1
2
(dpe) + H
2
1
Fe(CO)
2
(dpe)(H)
2
103.4
1
Fe(CO)
TS1
2
(dpe)(H)
2
2.198 (P)
1.749 (C)
84.6 (PFeP)
88.7 (CFeC)
50.2, 152.0 (PFeC)
84.2 (PFeP)
87.0 (CFeC)
1
1
Fe(CO)
TS2
2
(dpe)(H)
2
2.192 (P)
1.753 (C)
0.859
30.7
51.0, 151.6 (PFeC)
1Ru(CO)
3Ru(CO)
(dpe)
(dpe)
2.342 (P)
2.319 (P)
1.868 (C)
2.298 (P)
1.867 (C)
2.346 (P)
1.919 (C)
—
174.3
159.6
162.7
149.9
—
2
1
.878 (C)
2.359 (P)
.914 (C)
2.323 (P)
.887 (C)
2
—
97.6
—
1
Ru(CO)
2
(dpe)(H)
2
2
2.136
0.863
104.1
81.5
27.9
1
Ru(CO)
TS1
2
(dpe)(H)
2.342 (P)
1.879 (C)
82.3 (PRuP)
89.3 (CRuC)
50.0, 151.7 (PRuC)
82.1 (PRuP)
87.7 (CRuC)
1
1
Ru(CO)
TS2
2
(dpe)(H)
2
2.330 (P)
1.887 (C)
0.853
27.5
50.9, 151.7 (PRuC)
most reliable results for Fe(CO)
B3PW91*, where the coefficient c
exact”, Hartree–Fock, exchange is reduced to 0.15 (vs. 0.20
in B3PW91). The 16-electron fragment Fe(CO) (dpe) (I) was
found to adopt a triplet ground state and possess a distorted
trigonal-bipyramid geometry with one vacant equatorial site
where the Pax–Fe–COax angle was 156.0° and the Peq–Fe–COeq
angle is 98.4° (Fig. 3). Both the Fe–Pax distance of 2.271 Å and
the Fe–COax distance of 1.805 Å in I proved to be longer than
the corresponding equatorial distances of 2.240 and 1.786 Å
respectively (Table 1). This trend is common in four coordinate
systems with this geometry.
4
.22 This corresponds to using
describing the mixing of the
geometry that is not constrained by the chelate effect are 155.4
and 129.9°, respectively.21 The major difference between the
chelate and non-chelate structures therefore manifests itself in a
change in the equatorial ligand orientation. The comparison of
the bond lengths in the singlet and triplet forms reveals minimal
differences.
3
“
2
The predicted energy difference between the singlet and
triplet isomers of Fe(CO) (dpe) (I) is of great significance
2
to the photochemical driven p-H addition studies, since the
2
NMR experiments suggest that p-H enhanced hydride reso-
2
nances for Fe(CO) (dppe)(H) should be observed if singlet
2
2
Fe(CO) (dppe) is formed. Hydride signals for the same spe-
2
2 2 2
cies, Fe(CO) (dppe)(H) , formed from triplet-Fe(CO) (dppe),
overlap exactly with those derived from the singlet route since
they would arise from chemically identical species but have
normal signal strength. In the case of 1, we have demonstrated
that the enhancement factor approaches 15000. Assuming that
the results are thermodynamically derived, the lack of enhance-
ment observed in the chemistry of 2 implies that the equilibrium
concentration of the singlet would have to be at most 10−5 that of
the triplet, which in turn suggests an energy difference between
−1
spin states exceeding 15 kJ mol . Taking account that the singlet
is more reactive than the triplet, and assuming singlet and triplet
states are able to interconvert faster than addition to the triplet,
the minimum energy difference between states would have to be
still larger to account for the absence of enhancement. This fits
in with the theoretical values, which predict an energy difference
−
1
between the spin isomers of almost 40 kJ mol .
The Ru(CO) (dpe) system II was previously investigated by
Eisenstein et al. at a theory level quite similar to ours,
investigation was limited to the singlet state. We find the triplet
2
23,24
but the
Fig. 3 Labelling scheme for all calculated molecules.
−
1
state to lie 58 kJ mol higher in energy implying that it will not
be populated at room temperature, mirroring the experimental
study. Structurally, the same trends are observed for the singlet
isomer which has wider bond angles and shorter bond lengths
than the triplet species, as outlined in Table 1. In the computed
structure of singlet (II), both Ru–P and Ru–C bond lengths are
shorter for the equatorially bonded ligands, in agreement with
greater back-bonding in those positions. The Ru–C distances
are slightly shorter than those previously calculated for the
same system23 and in close agreement with those determined
experimentally for the related unsaturated Ru(0) complex
The singlet state of Fe(CO)
2
(dpe) (I) optimised with a similar
distorted trigonal bipyramid geometry, and proved to lie
−
1
4
0 kJ mol above the triplet ground state (Fig. 4). Comparison
of the singlet Pax–Fe–COax and Peq–Fe–COeq angles (176.9 and
33.7°, respectively) with those of the triplet (156.0 and 98.4°,
1
respectively) reveals that the structure distorts towards the
trigonalbipyramiduponpairingtheelectrons.Thecorresponding
ligand–metal bond lengths are shorter in the singlet (Table 1).
The corresponding bond angles for the singlet cis-cis isomer
2 3 2
of Fe(CO) (PH ) where the phosphines are free to adopt a
3
2 2 0
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4

View Article Online
2 2
Fig. 4 Calculated reaction coordinate, thermodynamic parameters and illustrations of the relevant geometries for the H addition to M(CO) (dpe)
and for the ligand scrambling processes (M = Fe, Ru).
t
Ru(CO)
1
2
(P Bu
2
5,26
Me)
2
with trans phosphines (1.886(10) and
Our calculation yields a wider Lax–Ru–Lax and
a narrower Leq–Ru–Leq angle than those reported previously by
the H–Fe–H angle shrinking to 82.4° to accommodate this,
although the HH distance is 2.00 Å. The distorted geometry
is seen for Ru(CO) (dpe)(H) (IV), but in this case the bond
2 2
2
.854(6) Å).
2
4
Eisenstein et al. The relatively important variations in these
optimized parameters that result from minor changes in the
theory level provide a sign of a relatively flat potential energy
surface along this coordinate, as also suggested by a study
of the related Ru(CO)
Ru(CO) (dpe) has a similar Lax–Ru–Lax angle to Fe(CO)
but a wider Leq–Ru–Leq angle. The triplet states of Ru(CO)
and Fe(CO)
Ru complex are also rather different from those determined
distances are longer than their iron counterparts (Table 1).
This allows the Pax–Ru–COax angle to increase to 162.7° so that
the ruthenium structure distorts less from the ideal octahedral
arrangement than the iron system. The Peq–Ru–COeq angle of
104° is essentially the same as that seen at iron although the
H–Ru–H angle closes slightly to 81.5°. The greater Ru–H bond
lengths result in a longer HH contact of 2.136 Å. It should be
noted that no X-ray data are available for a monomeric Ru(II)
complex of this type for comparison.
2
4
2 3 2
(PH ) system. The singlet state of
2
2
(dpe)
(dpe)
2
2
(dpe) have similar angles. The angles in the singlet
t
experimentally for complex Ru(CO)
2
(P Bu
2
Me)
2
(axial:
1
65.56(8)°; equatorial: 133.3(4)°), suggesting that the ground
Minimum energy crossing point
state geometry is highly affected by the phosphine (steric effect,
constraint imposed by the backbone).
The explicit determination of the MECP yields a structure
whose geometry and energy are illustrated in Table 1 and Fig. 4,
an expanded figure is available in the ESI.† This is closely related
Theoretical treatment of Fe(CO)
Ru(CO) (dpe)(H) (IV)
2 2
(dpe)(H) (III) and
to the MECP for the H
2
addition to other 16-electron Fe(0) frag-
2
2
ments.21 The H
ligand is found to approach the metal centre
2
The energy gains associated with the oxidative addition of H
to M(CO) (dpe) [M = Fe (I), Ru (II)] (Fig. 4) show that, as ex-
pected, the reaction is more exothermic for Ru (124 kJ mol )
2
with a parallel orientation relative to the equatorial plane of the
trigonal bipyramid and sideways relative to the empty coordina-
tion site. This preferred mode of attack was attributed to the
2
−
1
−
1
than Fe (97.9 kJ mol ). Since the two 16-electron fragments
have different spin states, it is interesting to consider also the
energy gain for the two species relating to the same spin state.
The energy gain relative to the triplet intermediate is much
smaller for Fe than for Ru, whereas the energy gain relative to
the singlet intermediate is greater for Fe. The large decrease in
the energy gain from the triplet state on going from Ru to Fe
may be related to the much greater cost of pairing the electrons
for the smaller Fe ion.
greater repulsion by the a
ideal C2v FeL fragment.
For the previously reported 1/3FeL
1
orbital relative to the b
1
orbital of the
4
4
+ H MECPs, it was
2
found that the FeH distance decreases and the HH dis-
tance increases with an increase of the ligand’s donor strength.21
However, the calculated FeH distances of 1.997 and 2.031 Å
and the HH distance of 0.775 Å for the crossing point of the
Fe(CO)
cally similar Fe(CO)
2
(dpe) system are quite close to those of the electroni-
(PH system, for which the comparative
2
3 2
)
The Fe(CO)
2
(dpe)(H)
2
III product optimised with a distorted
FeH distances are 1.981 and 2.052 Å, and the HH distance
is 0.778 Å.21 Energetically, the MECP is located 28.3 kJ mol
−1
octahedral structure. In the calculated structure of III, the bond
lengths for the axial phosphine and CO ligands are now shorter
than the corresponding equatorial ligands (Table 1). This is con-
sistent with the known trans influence of the hydride ligand. The
basic structure is however similar to that reported for the related
higher than the triplet Fe(CO)
2
(dpe) fragment, a value simi-
lar to those of the FeL systems investigated previously. In a
4
recent contribution, we computed a singlet/triplet energy split-
−
1
−1
ting of 50.2 kJ mol for Fe(dmpe)
2
and of 62.1 kJ mol for
2
7,28
21
−1
Fe(II) complex, tcc-Fe(CO)
2
[P(OPh)
3
]
2
H
2
.
2
Fe(dpe) . In the latter case, the MECP lies 25.6 kJ mol above
−
1
In the calculated structure of (III), the Pax–Fe–COax angle
was found to be 157.0°, rather than the ideal value of 180° and
the two groups are bent towards the hydride ligands. In the
equatorial plane the Peq–Fe–COeq angle widens to 103.4° with
the reactants which is similar to the 28.3 kJ mol found here.
Similar rate coefficients for H addition would therefore be
expected for the two systems. The Fe(dmpe) system was found
to react with H with a rate constant of 8 × 10 dm mol s at
2
2
5
3
−1 −1
2
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4
3 2 2 1

View Article Online
2
9
2
room temperature. Based on the failure to detect any enhanced
hydride resonance we would predict the lifetime of the triplet at
an g -H ligand which, in view of the small normal kinetic
2
isotope effect and apparently restricted rotation had only a weak
1
95 K to be less than 1 ms.
H–H interaction. This deduction was based on the observation
that at 306.2 K, the rates of hydride, phosphorus and carbonyl
interchange were the same within experimental error suggesting
synchronised exchange. The activation parameters were as
Fluxionality with Ru(CO)
2
(dppe)(H)
2
and Fe(CO)
2
(dppe)(H)
2
While the 1H NMR spectrum of ccc-Fe(CO)
contained two inequivalent hydride signals at 245 K upon
warming to 295 K a single broad peak is observed at d −9.9.
The corresponding P{ H) spectrum also consists of a very
broad signal centred at d 96 at this temperature showing that
Fe(CO)
used to determine the rate of exchange of the hydride ligands in
ccc-Fe(CO)
recorded at intervals of around 5 K between 189 and 296 K in
toluene-d
solution. Spectra were then analysed using g-NMR30
and rates obtained for data between 216 and 270 K. In order
to improve the number of reliable temperatures, a coalescence
measurement was also employed which allowed the hydride ligand
(dppe)(H)
2
‡
−1
‡
−1
−1
2
2
follows: DH = 85.5 ± 2 kJ mol and DS = 34 ± 7 J mol K
‡
−1 18
and a DG value of 75 kJ mol .
We have now carried out calculations on IV that predict two
independent ligand exchange pathways exist for 1 in direct
analogy with 2. The corresponding calculated enthalpies of
31
1
−
1
−1
2
(dppe)(H)
2
is highly fluxional. Line-shape analysis was
activation for TS1 and TS2 are 78.6 kJ mol and 82.0 kJ mol ,
very close to the experimental values for hydride ligand
1
‡
2 2
(dppe)(H) . For this purpose, H NMR spectra were
exchange in IV. The experimentally determined DS value of
−
1
−1
34 ± 7 J mol
K
is, however, significantly different from the
8
zero value predicted theoretically. Clearly, the entropy effects
are not adequately modelled using this level of theory. The H–H
bond distances in TS1 and TS2 for III and IV are in the range
2
2
seen for g -H ligands, those of TS1 (0.868 and 0.863 Å for Fe
−1
exchange rate at 271.3 K to be determined as 1300 s . Rate data
areprovidedintheESI.†Theactivationparametersforthehydride
and Ru, respectively) being only slightly longer than those of
TS2 (0.859 and 0.853 Å). These distances suggest that both
exchange processes place the hydride ligands sufficiently close
together for the formation of a five-coordinate geometry with
a metal-bonded “dihydrogen” ligand. However, no local mini-
mum corresponding to such complex was found; thus no distinct
dihydrogen intermediates exist along the exchange pathways.
‡
ligand exchange process were determined to be DH = 48 ± 6 kJ
mol− and DS = −10 ± 20 J mol
1
‡
−1
K for this process.
−1
3
1
1
Line shape analysis of the corresponding P{ H} spectra
between 200 and 290 K was used to examine the phosphorus
interchange pathway. The activation parameters for this
process were calculated to be DH = 33 ± 5 kJ mol and
DS = −48 ± 20 J mol
from those for hydride interchange. The observation that the
phosphorus interchange process has the lower activation barrier
is consistent with rates of 90 and 24 s , respectively, that were
determined for the two pathways at 232 K. This suggests that
phosphorus and hydride interchange can occur independently
of one another.
‡
−1
‡
−1
−1
K
and are substantially different
Reductive elimination of H
Fe(CO) (dppe)(H)
2 2 2
from Ru(CO) (dppe)(H) and
2
2
We have previously described18 how exchange between metal
dihydride and free dihydrogen involving 1 can be examined
quantitatively over the temperature range 343–373 K via the
mapping of hydride/dihydrogen connections obtained as a
−
1
3
2
function of mixing time using the 1D-NOE sequence of Keeler.
Theory indicated that there are two possible ways of
‡
−1
Subsequent analysis provided values of DH = 97 ± 10 kJ mol
2 2
exchanging the hydride ligands in M(CO) (dpe)(H) (M = Ru,
‡
−1
−1
and DS = 2 ± 2 J K mol for this reaction. At 369 K the rate
Fe) (Fig. 4). The two pathways can be described as pseudo-
rotations involving the two hydride ligands, each one occurring
in a different direction (clockwise and anticlockwise). One route
−
1
of reductive elimination of H
value of DS is not unexpected.
2
from 1 was 0.253 s . The small
‡
18,33
As can be seen from Fig. 4,
this experimental enthalpy is higher than that calculated for
TS1/TS2, but much lower than that calculated to reach singlet
Ru(CO) (dpe). This suggests that the reductive elimination step
2
is not purely dissociative.
(
through the transition state TS1) involves only P and CO site
interchange, with the hydride ligands remaining unaffected. In
the second pathway (through TS2), all three sets of ligands
exchange sites simultaneously. In the iron system, the calculated
enthalpy change to reach TS1 (DH ) from the ground state was
found to be +55.6 kJ mol , while for TS2 the enthalpy is higher
at +58.9 kJ mol . This is consistent with the experimental
observation that P–P mutual exchange has the lower activation
barrier. However, the experimental activation enthalpies are
substantially lower than those calculated for both processes
which suggests that a steric effect arising from the phenyl
substituents is not fully represented in the calculations where
they are replaced by hydrogen at the small iron centre.
When the reaction of Fe(CO)
similar way, no exchange peaks to free H
2
(dppe)(H)
2
was monitored in a
were observed up to
63 K, the highest readily accessible temperature. The failure to
‡
−
1
2
3
−
1
2
detect magnetisation transfer to H could be attributed either
to paramagnetic quenching of the associated spin encoding,
or to a process which was too slow to observe on the timescale
of the EXSY experiment. We therefore examined the reaction
of 2 with CO to form the precursor Fe(CO) (dppe) by NMR
3
spectroscopy at 333 K. A slow exponential decay of the signal
due to the hydride resonances of 2 was observed which yielded a
‡
The calculated entropy change (DS ) for these processes were
−
5
−1
‡
−1
first-order rate constant of 6 × 10 s , or DG = 109 kJ mol at
33 K. This observation requires slow H elimination from 2. In
the case of Fe(CO) {(C PCH CH P(C }(H) , Brookhart
et al. reported that trapping of the H loss product with P(OMe)
proceeded at the substantially faster rate of 7 × 10−3
52 K. This confirms that the use of the fluorinated phosphine
found to be slightly negative as shown in Fig. 4 and indicates that
there is little or no change in order on moving from the ground
state to the transition state. Negative entropy terms were also
obtained by experiment, although the errors are substantial.
The related complex Fe(CO) {(C F ) P(CH CH P(C F ) }(H)
2 2 5 2 2 2 2 5 2 2
has been prepared by Brookhart et al. in a similar manner. This
complex also proved to be stereochemically nonrigid with the
hydride exchange rate being estimated at 3.2 × 10 s at 235 K.
Slower phosphorus site exchange was also observed (4 × 10 s
3
2
2
2
F
5
)
2
2
2
F
2 5
)
2
2
2
3
,
at
−1
s
2
31
destabilises the Fe(II) centre and aids in H
2
elimination.31
3
−1
Conclusions
3
−1
at 257 K).
Literature studies on the laser-initiated reaction of
Ru(CO) (dmpe) (dmpe = Me PCH CH PMe ) with H have
described how CO dissociation leads to the rapid forma-
tion of Ru(CO) (dmpe)(solvent). On a longer timescale,
reaction with H leads to the generation of the stable complex
Ru(CO) (dmpe)(H)
used to prepare the analogous complex, Ru(CO)
from the readily accessible precursor Ru(CO) (dppe) and then
Someof ushavepreviouslyshownthatccc-Ru(CO)
is fluxional, undergoing intramolecular exchange of the two
2
(dppe)(H)
2
3
2
2
2
2
2
1
8
hydride ligands on the NMR timescale. Quantitative analysis
of this behaviour led us to suggest that hydride interchange was
accompanied by synchronised exchange of the two carbonyl
ligands and the dppe phosphorus atoms. We proposed that this
process involves accessing a five-coordinate complex containing
2
2
34
2
2
.
We have described how this route can be
(dppe)(H) (1)
2
2
3
3
2 2 2
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4

View Article Online
extended this approach to Fe(CO)
Fe(CO) (dppe)(H)
The proposed role of the zerovalent d intermediates,
3
(dppe) and hence prepared
interchange pathway. This is consistent with increased H–H
bonding character, although no local minimum (intermediate)
was found along the rearrangement pathway.
2
2
.
8
‡
Ru(CO)
keeping with related Ru(CO)
feature Ru(CO) and Ru(PR
2
(dppe) and Fe(CO)
2
(dppe), in these reactions is in
and Ru(PR (H) systems which
intermediates. The reactivity
In the case of 2, the measured DH barrier to hydride inter-
−
1
5
3
)
4
2
–4
change proved to be 48 ± 6 kJ mol while a second pathway
1
‡
4
3
)
4
that exchanges P and CO but not H has a DH barrier of
−
1
−1
of such species is, however, influenced dramatically by the elec-
tron configuration, be it singlet or triplet, which in turn depends
on the strength of interaction between the metal and ligand. In
33 kJ mol . The calculated values of 55.6 and 58.9 kJ mol
for the two analogous processes in Fe(CO) (dpe)(H) seem to
2
2
overestimate the barriers. The calculations, however, agree with
the observation that the iron system is more fluxional. For the
iron system III, the HH distance changes from 2.00 to 0.859 Å
in the transition state, which again suggests an increase in H–H
bonding character.
the case of the related complexes Fe(CO)
4 3 4
, Fe(PH ) and more
recently Fe(CO) (PH , it has been established by appropriate
2
3 2
)
2
1
theory that triplet ground states are adopted. The energy
difference between the triplet and the singlet reduces as the
r-donor properties of the phosphine increase, and the p-accept-
ing properties of the ligands increase. The resultant need for spin
crossover during reactions producing diamagnetic 18-electron
products necessitates an increased activation barrier and hence
Studies on the reductive elimination of H from 1 yielded
2
‡
−1
activation parameters of DH = 97 ± 10 kJ mol
and
‡
−1
−1
DS = 2 ± 2 J K mol for this reaction. In contrast, the
calculations indicated a barrier to H loss of 124.8 kJ mol .
2
−
1
a slower reaction rate. The analogous Ru(CO)
complexes on the other hand have been shown to adopt singlet
ground states.
4
and Ru(PH
3
)
4
The low experimental barrier can be explained by a role for
the solvent or the incoming ligand in stabilising the transition
state. In other words, H
in character.
2
substitution is not purely dissociative
Here we have studied the reactions of the intermediates
Ru(CO)
scopy. Evidence has been presented that when Fe(CO)
reacts to form Fe(CO) (dppe)(H) , no signal enhancements
are observed for the corresponding hydride ligand signals.
In contrast, when Ru(CO) (dppe) is used, signal enhance-
ments are observed for the analogous hydride resonances of
Ru(CO) (dppe)(H) . This result is consistent with the fast
reaction of a diamagnetic complex leading to retention of the
nuclear spin encoding and generation of 1 which retains p-H
activity. When Fe(CO) (dppe) is irradiated however, the triplet
or paramagnet, Fe(CO) (dppe), adds H with concomitant
2
(dppe) and Fe(CO)
2
(dppe) with p-H
2
by NMR spectro-
2
(dppe)
Experimental
Computational details
2
2
2
All calculations were performed using the Gaussian 98
36
program together with the modified form of the B3PW91
2
2
functional in which the proportion of exact exchange has been
modified from 0.20 to 0.15. We also used flexible polarised
22
2
basis sets on all atoms, as in our previous work on related
3
21–22
systems.
Thus, for Fe, C and O and P, and the incoming H
2
2
2
37
ligand, the triple-zeta basis sets of Schäfer et al. were used,
3
5
loss of the spin-encoding due to enhanced relaxation. This
simple observation demonstrates that the electronic spin state
of a reaction centre can be probed readily in conjunction with
with additional polarization functions. For other hydrogen
atoms, the smaller split-valence basis of the same authors was
used. Finally, for ruthenium, the 28 core electrons were treated
a suitable nuclear spin reservoir such as p-H
suggests that when substantial p-H enhancements have been
observed the reaction pathways involve diamagnetic centres.
Since the studies in CD CN produced similar results, we suggest
that these reactions proceed via direct H addition to the naked
6-electron fragment, rather than associative displacement from
the 18-electron solvent complex.
2
. Furthermore, it
38
using the ECP with the valence electrons treated using the
2
associated SDD basis set, augmented by one f polarisation
function (a = 1.0). All minima and transition states were
fully optimised and characterised by computing vibrational
frequencies at the same level of theory. The MECP was located
using Gaussian 98, together with the code developed by one of
3
2
1
39
the authors.
Given the propensity for chelate structures to influ-
ence reactivity, we completed a B3PW91* study of the
Experimental details
structures of Ru(CO)
2 2
(dpe) and Fe(CO) (dpe) (where
dpe = PH CH CH PH ). As expected, the iron system proved
2
2
2
2
The syntheses of Ru(CO)
performed according to literature methods. The complex
Ru(CO) (dppe)(H) (1) was obtained by in-situ irradiation
of a toluene-d solution of Ru(CO) (dppe) using a Kimmon
40 mW He–Cd CW laser in the presence of 3 atm of p-H
The preparation of the Fe(CO) (dppe)(H) used here was
achieved by first converting Fe(CO) by reaction with HSiPh
4 3
to Fe(CO) (SiPh )H. This white compound was then dissolved
(dppe)40 and Fe(CO) (dppe)41 were
3 3
to be a triplet while the ruthenium complex was a singlet. The
energy difference between the iron triplet, and the excited singlet
2
2
−
1
was determined to be 40 kJ mol . This value is sufficiently large
to account for the failure to see enhancement from a small equi-
librium amount of the singlet.
8
3
1
6
2
.
2
2
TheNMRdataforboth1and2indicatedthattheyarefluxional
5
3
1
31
42
on the NMR timescale. In the case of 1, a series of H, P and
1
3
C, studies revealed an apparently synchronised interchange of
the hydride, phosphine and carbonyl pairs. Measured activation
in benzene and 1 equivalent of dppe added. Argon was bubbled
through the solution over 2 h to drive off the CO that is released.
Upon concentration, the solution yielded an orange solid which
‡
−1
parameters for these processes were DH = 85.5 ± 2 kJ mol and
‡
−1
−1
DS = 34 ± 7 J mol K . When this process was modelled for
the dpe analogue, two pathways were revealed, one where the
phosphorus and carbonyl centres alone interchange positions,
and one where all three groups interchange. The theoretical
activation enthalpies for these processes were DH = 78.6 and
8
was recrystallised from warm Et
2
O, to yield yellow crystals
of Fe(CO) (dppe)(H) . This procedure has been reported
2 2
previously.20
8
The toluene-d (Apollo Scientific) was dried over potassium
‡
and degassed prior to use. The NMR measurements were made
using NMR tubes fitted with J. Young Teflon valves and sol-
vents were added by vacuum transfer on a high vacuum line.
For the parahydrogen induced polarization (PHIP) experi-
ments, hydrogen enriched in the para spin state was prepared
2 kJ mol− respectively. Our NMR observations failed to
1
distinguish these two pathways and suggest that the calculations
on the model complex underestimate the barrier to reaction in
−
1
the case of 1 by about 7 kJ mol and that the two processes have
2
similar activation barriers. The role of an g -H
2
interaction in
by cooling H
2 2 3
to 77 K over a paramagnetic catalyst (Fe O ) as
14
many hydride exchange processes has been suggested with H–H
bond lengths between 0.8 and 1.0 Å in the transition states.
Our calculations reveal that the HH distance changes from
previously described. Spectra were recorded on a Bruker DSX
1
31
400 spectrometer with H at 400.13 MHz and P at 161.9 MHz.
1
H NMR chemical shifts are reported in ppm relative to residual
H signals in the deuterated solvents (toluene-d
1
31
2
.136 Å in IV to 0.853 Å in the transition state for the hydride
7
, d 2.13),
P
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4
3 2 2 3

View Article Online
NMR chemical shifts are relative to an external 85% solution
of phosphoric acid.
19 (a) G. E. Ball and B. E. Mann, J. Chem. Soc., Chem. Commun., 1992,
5
61; (b) P. Meakin, E. L. Muetterties and J. P. Jesson, J. Am. Chem.
Soc., 1973, 95, 75.
2
2
2
2
0 U. Schubert and M. Knorr, Inorg. Chem., 1989, 28, 1765.
1 J. N. Harvey and R. Poli, Dalton Trans., 2003, 4100.
2 J. Harvey and M. Aschi, Faraday Discuss., 2003, 124, 129.
3 T. Gottschalk-Gaudig, J. C. Huffman, K. G. Caulton, H. Gerard
and O. Eisenstein, J. Am. Chem. Soc., 1999, 121, 3242.
4 T. Gottschalk-Gaudig, J. C. Huffman, H. Gérard, O. Eisenstein and
K. G. Caulton, Inorg. Chem., 2000, 39, 3957.
Acknowledgements
We are grateful to the European Union for partial funding under
the HYDROCHEM network (contract HPRN-CT-2002-00176)
and to CINES for a grant of free computational time. Support
from the University of York, the ERASMUS programme and
the EPSRC is also acknowledged.
2
25 M. Ogasawara, S. A. MacGregor, W. E. Streib, K. Folting,
O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1995, 117,
8
869.
References
2
6 M. Ogasawara, S. A. MacGregor, W. E. Streib, K. Folting,
O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1996, 118,
10189.
1
(a) R. Poli and J. N. Harvey, Chem. Soc. Rev., 2003, 32, 1;
b) J. N. Harvey, R. Poli and K. M. Smith, Coord. Chem. Rev., 2003,
38–239, 347–361; (c) R. Poli, J. Organomet. Chem. in press.
(a) M. Poliakoff and E. Weitz, Acc. Chem. Res., 1987, 20, 408;
b) W. Wang, A. A. Narducci, P. G. House and E. Weitz, J. Am.
(
2
27 V. Arion, J. J. Brunet and D. Neibecker, Inorg. Chem., 2001, 40,
2
2628.
(
28 Nan Nhat Ho, R. Bau and S. A. Mason, J. Organomet. Chem., 2003,
676, 85.
Chem. Soc., 1996, 118, 8654.
3
4
P. T. Snee, C. K. Payne, S. D. Mebane, K. T. Kotz and C. B. Harris,
J. Am. Chem. Soc., 2001, 123, 6909.
29 M. K. Whittlesey,R. J. Mawby,R. Osman,R. N. Perutz,L. D. Field,
M. P. Wilkinson and M. W. George, J. Am. Chem. Soc., 1993, 115,
8627.
30 G-NMR V5. Adept Scientific. Plc 2003.
31 M. Brookhart, W. A. Chandler, A. C. Pfister, C. C. Santini and
P. S. White, Organometallics, 1992, 11, 1263.
H. Ihee, J. M. Cao and A. H. Zewail, Angew. Chem., Int. Ed., 2001,
4
0, 1532.
5
6
R. J. Ryther and E. Weitz, J. Phys. Chem., 1991, 95, 9841.
(a) J. Chetwynd-Talbot, P. Grebenik and R. N. Perutz, Inorg. Chem.,
1
982, 21, 3647; (b) P. A. Cox, P. Grebenik, R. N. Perutz, M. D.
Robinson, R. Grinter and D. R. Stern, Inorg. Chem., 1983, 22, 3614;
c) R. G. Graham, R. Grinter and R. N. Perutz, J. Am. Chem. Soc.,
998, 110, 7036; (d) F. G. N. Cloke, J. C. Green, M. L. H. Green
32 K. Stott, J. Stonehouse, J. Keeler, T. L. Hwang and A. J. Shaka,
J. Am. Chem. Soc., 1995, 117, 4199.
33 L. A. Watson and O. Eisenstein, J. Chem. Educ., 2002, 79, 1269.
34 M. K. Whittlesey, R. N. Perutz, I. P. Virrels and M. W. George,
Organometallics, 1997, 16, 268.
(
1
and C. P. Morley, Chem. Commun., 1985, 945; (e) see: L. Labella,
A. Chernega and M. L. H. Green, J. Chem. Soc., Dalton Trans.,
35 T. Nilsson and J. Kowalewski, Mol. Phys., 1617, 98, 2000.
36 Gaussian98,RevisionA.11.1,M. J. Frisch,G. W. Trucks,H. Schlegel,
G. E. Scuseruia, 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. Peterson,
P. Y. Ayala, Q. Cui, K. Morokuma, P. Salvador, J. J. Dannerberg,
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. Liaskenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,
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, 2001.
1
995, 395, and references therein.
7
8
(a) J. C. Green, J. N. Harvey and R. Poli, J. Chem. Soc., Dalton
Trans., 2002, 1861; (b) J. C. Green, Chem. Soc. Rev., 1998, 27, 263.
(a) A. A. Bengali, R. G. Bergman and C. B. Moore, J. Am. Chem.
Soc., 1995, 117, 3879; (b) P. T. Snee, C. K. Payne, C. T. Kotz,
H. Yang and C. B. Harris, J. Am. Chem. Soc., 2001, 123, 2255.
R. Poli and K. M. Smith, Eur. J. Inorg. Chem., 1999, 12, 877.
9
1
1
0 C. Hall and R. N. Perutz, Chem. Rev., 1996, 96, 3125.
1 C. R. Bowers, D. H. Jones, N. D. Kurur and J. A. Labinger, Adv.
Magn. Reson., 1990, 14, 269.
1
1
1
2 R. Eisenberg, Acc. Chem. Res., 1991, 24, 100.
3 J. Natherer and J. Bargon, Prog. Nucl. Magn. Res., 1997, 31, 293.
4 S. B. Duckett and C. J. Sleigh, Prog. Nucl. Magn. Res., 1999, 34,
7
1.
1
1
5 S. B. Duckett and D. Blazina, Eur. J. Inorg. Chem., 2003, 16, 2901.
6 C. Godard, P. Callaghan, J. L. Cunningham, S. B. Duckett,
J. A. B. Lohman and R. N. Perutz, Chem. Commun., 2002, 2836.
7 (a) S. Geftakis and G. E. Ball, J. Am. Chem. Soc., 1999, 121,
37 A. Schaefer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97,
2571.
38 D. Andrae, U. Haussermann, M. Dolg, H. Stoll and H. Preuss,
Theor. Chim. Acta, 1990, 77, 123.
39 J. N. Harvey, M. Aschi, H. Schwarz and W. Koch, Theor. Chem.
Acc., 1998, 99, 95.
40 R. A. Sanchez-Delgado, J. S. Bradley and G. Wilkinson, J. Chem.
Soc., Dalton Trans., 1976, 5, 399.
1
1
6
336; (b) N. A. Knatochnil, J. A. Pachinson, P. J. Bedmaishi and
P. J. Sadler, Angew. Chem., Int. Ed., 1999, 38, 1460; (c) C. E. Lyon,
J. L. Lopez, B.-M. Cho and P. J. Hore, Mol. Phys., 2002, 100, 1261;
(
d) T. Kühn and H. Schwalbe, J. Am. Chem. Soc., 2002, 122, 6169.
8 D. Schott, C. J. Sleigh, J. P. Lowe, S. B. Duckett, R. J. Mawby and
41 T. A. Manuel, Inorg. Chem., 1963, 2, 854.
42 W. Jetz and W. A. G. Graham, Inorg. Chem., 1971, 10, 4.
M. G. Partridge, Inorg. Chem., 2002, 41, 2960.
3
2 2 4
D a l t o n T r a n s . , 2 0 0 4 , 3 2 1 8 – 3 2 2 4