A combined parahydrogen and theoretical study of H2 activation by 16-electron d8 ruthenium(0) complexes and their subsequent catalytic behaviour

Article abstract of DOI:10.1039/b410912k

The photochemical reaction of Ru(CO)₃(L)₂, where L = PPh₃, PMe₃, PCy₃ and P(p-tolyl)₃ with parahydrogen (p-H₂) has been studied by in-situ NMR spectroscopy and shown to result in two competing processes. The first of these involves loss of CO and results in the formation of the cis-cis-trans-L isomer of Ru(CO)₂(L)₂(H)₂, while in the second, a single photon induces loss of both CO and L and leads to the formation of cis-cis-cis Ru(CO)₂(L)₂(H)₂ and Ru(CO)₂(L) (solvent)(H)₂ where solvent = toluene, THF and pyridine (py). In the case of L = PPh₃, cis-cis-trans-L Ru(CO)₂(L) ₂(H)₂ is shown to be an effective hydrogenation catalyst with rate limiting phosphine dissociation proceeding at a rate of 2.2 s ^-1 in pyridine at 355 K. Theoretical calculations and experimental observations show that H₂ addition to the Ru(CO)₂(L) ₂ proceeds to form cis-cis-trans-L Ru(CO)₂(L) ₂(H)₂ as the major product via addition over the π-accepting OC-Ru-CO axis.

Full text of DOI:10.1039/b410912k

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
A combined parahydrogen and theoretical study of H2 activation by
16-electron d8 ruthenium(0) complexes and their subsequent catalytic
behaviour†
John P. Dunne,a Damir Blazina,a Stuart Aiken,a Hilary A. Carteret,b Simon B. Duckett,*a
Jonathan. A. Jones,c Rinaldo Polid and Adrian C. Whitwooda
a
Department of Chemistry, University of York, UK YO10 5DD
LITQ, Departement d’Informatique et Recherche Operationelle, Universite de Montreal,
b
Pavillon Andre-Aisenstadt, Montreal, Quebec H3C 3J7, Canada
Oxford Centre for Quantum Computation, Clarendon Laboratory, Parks Road,
c
Oxford, UK OX1 3PU
Laboratoire de Chimie de Coordination, 205 Route de Narbonne, 31077 Toulouse Cedex,
d
France
Received 16th July 2004, Accepted 25th August 2004
First published as an Advance Article on the web 16th September 2004
The photochemical reaction of Ru(CO)3(L)2, where L = PPh3, PMe3, PCy3 and P(p-tolyl)3 with parahydrogen (p-
H2) has been studied by in-situ NMR spectroscopy and shown to result in two competing processes. The first of
these involves loss of CO and results in the formation of the cis-cis-trans-L isomer of Ru(CO)2(L)2(H)2, while in the
second, a single photon induces loss of both CO and L and leads to the formation of cis-cis-cis Ru(CO)2(L)2(H)2
and Ru(CO)2(L)(solvent)(H)2 where solvent = toluene, THF and pyridine (py). In the case of L = PPh3, cis-cis-
trans-L Ru(CO)2(L)2(H)2 is shown to be an effective hydrogenation catalyst with rate limiting phosphine dissociation
proceeding at a rate of 2.2 s−1 in pyridine at 355 K. Theoretical calculations and experimental observations show that
H2 addition to the Ru(CO)2(L)2 proceeds to form cis-cis-trans-L Ru(CO)2(L)2(H)2 as the major product via addition
over the p-accepting OC–Ru–CO axis.
In 1965, Wilkinson and co-workers showed that Ru(CO)3(PPh3)2
Introduction
can catalyse the hydroformylation of alkenes to aldehydes,4,5
The examination of the oxidative addition of H2 to d8 square-
and that this complex was converted to Ru(CO)2(PPh3)2(H)2,
planar transition metal complexes has been significant in
enabling the understanding of how transition metal catalysts
operate. For example, the accepted mechanism of H2 addition
with cis carbonyls and hydrides, and mutually trans phosphine
ligands (cct-L), before hydroformylation commenced.6 Gordon
and Eisenberg reported that this reaction can be initiated and
to the square-planar Ir(CO)Cl(PPh3)2 is concerted and occurs
accelerated photochemically.7 Related ruthenium complexes
across the Cl–Ir–CO axis.1 Recent work in our group has shown
such as Ru(CO)(PPh3)3(H)2 have found use in catalytic
that a minor product is also formed by H2 addition over the
C–C bond-forming reactions involving ketones and suitable
P–Ir–P axis (Scheme 1).2,3
alkenes.8,9 Other ruthenium complexes have also been described
that are successful in asymmetric hydrogenation.10
It is well known that a step in many catalytic transforma-
tions involves the binding of the transforming substrate
to the catalyst. Hence, it is important to understand the
selectivity shown by the addition of H2 to potential catalysts,
such as the ruthenium complexes that feature here. Recently,
Caulton and co-workers isolated the 16-electron ruthenium(0)
complex Ru(CO)2(PMetBu2)2, which has a trigonal-bipyramidal
structure with a vacant equatorial site.11 This suggests that H2
addition could take place either parallel or orthogonal to the
OC–Ru–CO axis. Previous NMR studies on Ru(CO)2(L)2(H)2
(L = PMe3, PMe2Ph, and AsMe2Ph) have revealed that these
compounds can exist in three geometries, cis, cis, cis (ccc),
cct-L and cct-CO, with equilibrium ratios that are highly
dependent on the electronic properties of L (see Scheme 2).12–14
If these reactions involve the same Ru(CO)2(L)2 intermediate
Scheme 1 Addition of H2 to IrCl(CO)(L)2 species.
as characterised with L = PMetBu2, then H2 addition
across both the ligand–metal axes could account for two of
these products. When L = PMe3, the ccc form proved to be
visible only when parahydrogen (p-H2)13,14 was used to amplify
its spectral features and hence studying this problem poses
many challenges. In contrast, when L = AsMe2Ph, the ccc- and
cct-L forms were found to be present in similar quantities and
a cct-CO isomer was detectable. At elevated temperatures, the
AsMe2Ph complexes proved to be in equilibrium, while for
the other systems dynamic behaviour was observed. Fluxional
behaviour within Ru(CO)2(L)2(H)2 could therefore also account
We therefore set out to examine H2 addition to the iso-
electronic ruthenium species Ru(CO)2(L)2, where L = PPh3,
PMe3, PCy3 and P(p-tolyl)3, and L2 = dppe (Ph2PCH2CH2PPh2),
which are formed by photolysis of Ru(CO)3(L)2 complexes.
Ruthenium complexes of this type have attracted attention
because of their ability to catalyse organic transformations.4–10
† Electronic supplementary information (ESI) available: NMR
characterisation, catalytic studies, and XYZ coordinates determined for
optimised structures. See http://www.rsc.org/suppdata/dt/b4/b410912k/
3 6 1 6
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
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
for the observed product distribution. In the case of L2 =
dppe, hydride site interchange within the ccc form has been
observed and shown to be accompanied by synchronised CO
and phosphorus centre interchange. This process was suggested
to involve the formation of a trigonal-bipyramidal transition
state that contained an g2-H2 ligand with little H–H bond-
ing character. Subsequent theoretical studies on the related
Ru(CO)2(PH2CH2CH2PH2)(H)2 supported this view.15 The
potential fluxionality of these ruthenium dihydride species
therefore complicates the original aim of exploring the addition
of H2 to Ru(CO)2(L)2.13
the NMR spectrometer, no reaction was observed according
to 31P NMR spectroscopy. However, when the photolysis was
repeated on a fresh sample in the presence of 3 atm of p-H2,
1
the initial 32-scan H NMR spectrum revealed the selective
formation of the cis-cis-trans-L isomer of the known complex
Ru(CO)2(PPh3)2(H)2, cct-2-PPh3.36 This was evident from the
hydride signal that was seen for the two chemically equivalent
hydride ligands of cct-2-PPh3 at d −6.35 which showed an
unexpected p-H2 enhancement (Fig. 1). Since p-H2 corresponds
to H2 in 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. In chemical reactions that
produce a new molecule where the two hydrogen atoms become
distinct (I and S), the atoms become separately addressable and
under these conditions are described in the product operator
formalism as providing IzSz magnetisation. This state leads to
observable IzSx and IxSz terms which correspond to anti-phase
signals, one for the I spin and one for the S spin, which are
separated by JIS (with hydrogen, JHH). However, in the case of
cct-2-PPh3 a more complicated situation results since the two
hydrides should form part of an A2 spin system; this will be
commented on later in the text.
Scheme 2 Observed isomers of Ru(AsMe2Ph)2(CO)2(H)2.
There have been a number of reports on the use of UV
photolysis of an NMR sample within the NMR probe to
study in-situ reactions.16–20 Here we have employed NMR
spectroscopy in conjunction with in-situ photolysis and
p-H2 to study H2 addition to a series of ruthenium complexes.
Utilisation of the p-H2 effect was necessary to allow low
concentration photoproducts to be detected via the observa-
tion of enhanced NMR signals for nuclei that originate in the
p-H2 molecule. Recent photochemical studies in our group on
Ru(CO)3(dppe) have demonstrated that the hydride ligands
of the product Ru(CO)2(dppe)(H)2 are enhanced by a factor
of 28400.21 That study indicated the feasibility of using p-H2
to initialise an NMR quantum computer, and the molecules
described in the current paper were initially prepared to test
their suitability in such an application.
The p-H2 effect has been called PHIP (parahydrogen
induced polarisation)22 and PASADENA (parahydrogen and
synthesis allow dramatically enhanced nuclear alignment)23
and has been extensively reviewed.24–27 A notable achievement
in this area that is relevant to this study is the demonstration by
Aime et al. that Os3(l-H)2(CO)10, a species with magnetically
equivalent hydrides, can be enhanced,28 and that the enhanced
hydride signal arises via the involvement of an intermediate
with inequivalent hydrides. Similar studies involving
Ru3(CO)11(NCMe) yielded an enhanced emission signal for
molecular hydrogen that indicated a reversible interaction of
p-H2 with the Ru3 cluster containing inequivalent hydrides.29
More recently, PHIP has been employed in the sensitisation
of a hydroformylation product containing a single p-H2 atom30
and the transfer of polarisation via a 13C nucleus to deuterium
after the hydrogenation of a perdeuterated substrate.31 It has
also been successfully exploited in the study of catalytic trans-
formations by mono-,32,33 di-34 and tri-nuclear35 species.
Fig. 1 Selected regions of NMR spectra obtained at 255 K during
the reaction between Ru(CO)3(PPh3)2 and p-H2 with concurrent laser
irradiation: (a) 32-transient 1H spectrum of Ru(CO)2(PPh3)2(H)2,
cct-2, in toluene-d8; (b) 1H spectrum illustrating the hydride resonances
of Ru(CO)2(PPh3)(THF-d8)(H)2 4b; (c) 1H spectrum in pyridine-d5
showing Ru(CO)2(PPh3)(py-d5)(H)2 4c.
When a further 128 transients were recorded with con-
current photolysis, the new 1H{31P} NMR spectrum contained
additional enhanced hydride resonances for the known ccc
isomer of Ru(CO)2(PPh3)2(H)2, ccc-2-PPh3, and weaker signals
derived from the two equivalent hydride ligands of the fac
isomer of the known complex Ru(CO)3(PPh3)(H)2 and the
two inequivalent hydrides of the corresponding mer isomer.36
Appropriate NMR data for these complexes are provided in
Table 1 and their structures are indicated in Fig. 2.
The formation of a further photoproduct became evident
upon longer photolysis in spectra recorded with a large number
of transients, where the greater signal intensities revealed
two further hydride signals at d −2.95 and −4.91, with 5% of
the intensity of the signal for cct-2-PPh3. These signals were
assigned to 4a-PPh3 and appeared as doublets of antiphase
doublets, due to mutual JHH couplings of −5 Hz and single
31P couplings of 21 and 114 Hz, respectively. On stopping the
laser irradiation, the hydride signals for all the products except
cct-2-PPh3 decayed in intensity until after 1 min they were no
longer observable. At this point the hydride signal for cct-2-PPh3
This paper illustrates: (i) investigations into the mechanism
of H2 addition to 16-electron d8 ruthenium(0) complexes of the
type Ru(CO)2(L)2; (ii) the linking of experimental observations
and high-level DFT calculations and (iii) the catalytic proper-
ties of Ru(CO)3(L)2 systems towards hydrogenation and hydro-
formylation.
Results and discussion
Photochemical reactions of Ru(CO)3(PPh3)2
When a toluene-d8 solution of Ru(CO)3(PPh3)2, 1-PPh3, was
irradiated with a 325-nm He/Cd CW laser at 255 K inside
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
3 6 1 7

View Article Online
3 6 1 8
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8

View Article Online
and the pyridine complex 4c-PPh3 now being observed. These
data suggest that any 4a-PPh3 that is formed under these condi-
tions reacts with pyridine to yield 4c-PPh3, or with phosphine
to re-form cct-2-PPh3 more rapidly than the NMR detection
time scale. The reduction in the observed signal strengths of
the hydride resonances for ccc-2-PPh3 also suggests that one
route to its formation involves the displacement of toluene in
4a-PPh3 by CO. We note that no evidence for H–D exchange was
observed in these experiments, so reversible hydride exchange
with the solvent via C–H bond activation is not occurring.
Effect of phosphine on the product distribution
In order to probe the effect of the phosphine on this reaction,
analogous complexes containing PMe3, PCy3, P(p-tolyl)3 and
the chelating diphosphine 1,2-bis(diphenylphosphino)ethane
(dppe) were prepared and examined under identical conditions.
In the case of Ru(CO)3[P(p-tolyl)3]2 (viz. 1-P(p-tolyl)3),
photolysis in toluene-d8 at 255 K led to the initial observation
of cct-Ru(CO)2[P(p-tolyl)3]2(H)2, cct-2-P(p-tolyl)3. The hydride
signal for this species appears at d −6.04 in the 1H NMR spec-
trum, with a similar signal profile under these conditions to that
described earlier for cct-2-PPh3. On longer irradiation, hydride
signals for mer-Ru(CO)3[P(p-tolyl)3](H)2, mer-3-P(p-tolyl)3, and
the toluene solvent complex 4a-P(p-tolyl)3 were observed. No
evidence was obtained in these spectra for the formation of
the isomers ccc-2-P(p-tolyl)3 and fac-3-P(p-tolyl)3. When the
solvent was changed to pyridine-d5, the major product proved
to be cct-2-P(p-tolyl)3 and signals for the pyridine solvent
complex 4c-P(p-tolyl)3 were detected. Appropriate resonances
for these species are listed in Table 1.
When a sample containing Ru(CO)3(PMe3)2 (viz. 1-PMe3)
was photolysed under p-H2 in toluene-d8 at 255 K only cct-2-
PMe3 was observed.13 Upon repeating this experiment at 295 K,
hydride resonances for both the cct and ccc isomers of 2-PMe3
were detected, although once irradiation was stopped the signals
from ccc-2-PMe3 were no longer visible. The failure to observe
the solvent complex 4c-PMe3 or mer- and fac-3-PMe3 in this
reaction suggests that the stronger donating ability of PMe3
reduces the propensity for photochemically induced phosphine
loss. However, when 1-PMe3 was photolysed in pyridine-d5 in
the presence of three atm of p-H2 at 255 K, very weak signals
for a pair of enhanced hydride resonances at d −4.04 and −4.25
due to 4c-PMe3 were observed. The formation of ccc-2-PMe3
was again quenched.
Fig. 2 Proposed structures for compounds 2–4 and 6 identified in
this study, where R = Ph, Me, Cy or p-tolyl.
had a normal signal profile, which confirmed its hydride
ligands do not exchange with free H2 at this temperature. This
information indicates that the hydride signals for ccc-2-PPh3,
fac-3-PPh3 and 4a-PPh3 are only enhanced at 255 K because
they are formed photochemically, and that they either convert
to cct-2-PPh3 or are present in such low amounts as to be
undetectable under normal conditions.
The identity of the previously unknown species 4a-PPh3 was
deduced by changing the solvent. When the same experiment
was repeated in THF-d8 at 255 K, a pair of analogous hydride
signals were observed at d −3.66 and −5.26 due to 4b-PPh3
(Fig. 1). It should be noted that no enhanced hydride reso-
nances corresponding to ccc-2-PPh3 were observed in spectra
recorded in THF-d8, although on longer photolysis hydride
signals for mer- and fac-3-PPh3 were again detected. On moving
to pyridine-d5, a series of similar observations were made,
with the related complex 4c-PPh3 now exhibiting enhanced
hydride resonances at d −3.52 and −4.14 (Fig. 1). However,
in pyridine-d5, the hydride signals for cct-2-PPh3 and 4c-PPh3
were equally intense at the onset of irradiation. Once again, the
signals for ccc-2-PPh3 could not be observed. It should be noted
that enhanced hydride signals for cct-2-PPh3 were immediately
apparent in all these experiments.
These observations indicate that products of the type
4-PPh3 correspond to the highly reactive solvent complex
Ru(CO)2(PPh3)(solvent)(H)2. The NMR signatures of these
complexes require structures where phosphine and CO ligands
are trans to the two hydrides. For 4b-PPh3 and 4c-PPh3 solvent
coordination via a heteroatom lone pair is expected while
in 4a-PPh3 the toluene ligand is predicted to coordinate in
an g2-fashion.16 Unfortunately, this reaction could not be
examined in a non-coordinating solvent such as cyclohexane
or methylcyclohexane due to the insolubility of 1-PPh3 in
these solvents. Under such conditions, the formation of the
unsaturated 16-electron complex Ru(CO)2(PPh3)(H)2 complex
with a square-based pyramidal geometry and inequivalent
equatorial hydride ligands would have been expected.37 Since
this complex would contain a hydride ligand that is trans to a
vacant site and hence yield a hydride signal at very high field
(d −20 to −40),38 any suggestion that 4a-PPh3 contains a vacant
coordination site can be discounted.
In order to investigate the effect of introducing a more
sterically demanding phosphine, the complex Ru(CO)3(PCy3)2
was prepared. Crystals suitable for X-ray analysis were grown
from a 1:1 mixture of THF–hexane at room temperature.
This complex adopts a trigonal-bipyramid geometry (Fig. 3
and Tables 2 and 3) with equatorial CO groups and mutually
trans axial tricyclohexylphosphine ligands. The cyclohexyl
groups adopt a staggered orientation relative to the Ru(CO)3
The effect of the solvent on this reaction is clearly sub-
stantial, since although cct-2-PPh3, mer and fac-3-PPh3 and
4-PPh3 were seen in THF and pyridine, signals for ccc-2-PPh3
were absent. In order to probe the effect of the coordinating
strength of the solvent more directly, we examined a toluene-
d8 solution of 1-PPh3 containing a 20-fold excess of PPh3 and
p-H2 at 273 K. Signals for mer and fac-3-PPh3 and ccc-2-PPh3
were observed but with dramatically reduced signal intensities
relative to the situation without phosphine, while signals for
4a-PPh3 were absent. This study was then repeated in toluene-d8
using 20 lL of pyridine instead of PPh3. Under these condi-
tions the formation of the toluene solvent complex 4a-PPh3
was again suppressed, with ccc-2-PPh3, mer- and fac-3-PPh3
Fig. 3 Crystal structure of Ru(CO)3(PCy3)2.
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
3 6 1 9

View Article Online
to a 31P centre and the other trans to CO. However, when
the photolysis is performed in pyridine-d5 two new minor
species can be observed (see Fig. 4). The first of these, 5a,
shows 1H signals at d −3.89 (JHH = −5 Hz, JHP = 21 Hz)
and −4.36 (JHH = −5 Hz, JHP = 104 Hz), which are split by a
single phosphorus nucleus. The 31P centre giving rise to these
couplings was located at d 39.17 by HMQC spectroscopy.
The presence of the solitary 1H–31P coupling indicates that
5a corresponds to a species in which the dppe ligand is un-
chelated, i.e. Ru(CO)2(g1-dppe)(py)(H)2. From the hydride
chemical shifts, the pyridine moiety can be deduced to be cis
to both hydrides, with one hydride trans to phosphine and
the other trans to CO, as shown in Fig. 5. It should be noted
that due to the unstable nature of this species and the need for
p-H2 amplification, the uncoordinated 31P centre could not be
detected at this point.
Table 2 Crystal data and structure refinement for Ru(CO)3(PCy3)2
1-PCy3
Empirical formula
Formula weight
Temperature/K
Wavelength/Å
Crystal system
Space group
C39H66O3P2Ru
745.93
115(2)
0.71073
Monoclinic
P21/n
Unit cell dimensions
a/Å
b/Å
15.587(2)
12.4953(18)
c/Å
19.860(3)
95.843(4)
3848.0(9)
b/°
V/Å3
Z
4
Dc/Mg m−3
1.288
0.525
l/mm−1
F(000)
1592
Crystal size/mm
Theta range for data collection/°
Index ranges, hkl
Reflections collected
Independent reflections (Rint)
Completeness to theta
Absorption correction
Refinement method
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I > 2r(I)]
R indices (all data)
Largest diff. peak, hole/e Å−3
0.43 × 0.26 × 0.15
1.58 to 27.53
−20 to 20, −16 to 13, −18 to 25
25489
8828 (0.0470)
27.53° → 99.6%
None
Full-matrix least-squares on F2
8828/0/450
0.964
R1 = 0.0320, wR2 = 0.0668
R1 = 0.0533, wR2 = 0.0725
0.541, −0.485
Table 3 Relevant bond lengths and angles for Ru(CO)3(PCy)3 1-PCy3
C(37)–Ru(1)
C(38)–Ru(1)
C(39)–Ru(1)
P(2)–Ru(1)
P(3)–Ru(1)
C(1)–P(2)
1.910(2)
1.903(2)
1.915(2)
2.3783(6)
2.3788(6)
1.864(2)
1.856(2)
1.885(5)
C(13B)–P(2)
C(19)–P(3)
C(25)–P(3)
C(31)–P(3)
C(37)–O(2)
C(38)–O(1)
C(39)–O(3)
1.874(5)
1.852(2)
1.8679(19)
1.8792(19)
1.162(2)
1.161(2)
1.154(2)
Fig. 4 1H{31P} PHIP-enhanced spectrum at 275 K with concurrent
photolysis, showing species 5, 5a and 5b.
C(7)–P(2)
C(13A)–P(2)
O(2)–C(37)–Ru(1)
O(1)–C(38)–Ru(1)
O(3)–C(39)–Ru(1)
P(2)–Ru(1)–P(3)
178.0(2)
C(38)–Ru(1)–C(37)
C(38)–Ru(1)–C(39)
C(37)–Ru(1)–C(39)
119.74(9)
116.77(9)
123.49(9)
176.60(18)
175.26(19)
176.942(19)
core. This structure is directly analogous to that reported for
39
related ruthenium complexes such as Ru(CO)3(PPh3)2 and
Ru(CO)3(PMe3)2.40 The Ru–P distances were found to be iden-
tical at 2.378 Å while the Ru–CO bond lengths differ slightly
(1.910(2), 1.903(2), 1.915(2) Å). Notably, the Ru–P bond length
is longer than that reported for the related PMe3 complex where
it is 2.34 Å. The Pax–Ru–Pax angle in 1-PCy3 is slightly bent at
176.942(19)° with the COeq–Ru–COeq angles being inequiva-
lent at 116.77(9), 119.74(9) and 123.49(9)°; all the O–C–Ru
angles are also slightly bent away from linearity (e.g. 176.60°).
It should also be noted that one of the cyclohexyl rings of a
phosphine ligand is disordered due to conformational effects.
Photolysis of 1-PCy3 with p-H2 in toluene-d8 at 253 K
exclusively yielded the cct-2-PCy3 isomer with no evidence
for a solvent dihydride analogous to 4a or for the ccc isomer
being obtained. In addition, no evidence for the formation of
the ccc-2 isomer was found even when these experiments were
repeated at 295 K. This selectivity can be attributed to the
steric bulk of tricyclohexylphosphine, which should disfavour
the required cis arrangement in ccc-2-PCy3. When the photolysis
was performed in pyridine-d5, NMR signals for two isomers of
Ru(CO)2(PCy3)(py)(H)2 were evident (Table 1).
Fig. 5 Products formed from reactions of Ru(CO)3(dppe) with 3 atm
of p-H2 in pyridine-d5.
1
The second species, 5b, exhibits H hydride resonances at
d −4.46 and d −4.65 (JHH = −5 Hz), both of which couple to
two 31P nuclei. The second of these resonances exhibits a large
31P coupling of 122 Hz to a 31P nucleus resonating at d 89.05
which is indicative of a trans arrangement between the asso-
ciated ligands. A second coupling of 23 Hz was present due to
a further 31P centre that was detected at d 65.41 in the corres-
ponding 31P NMR spectrum; the size of the JHP is now indica-
tive of a cis arrangement of the respective nuclei. The hydride
resonance at d −4.46, meanwhile, showed two cis couplings of
24 and 30 Hz to 31P nuclei as detailed in Table 4.
Utilisation of a bidentate phosphine
Photolysis of Ru(CO)3(dppe) in the presence of H2 exclusively
yielded the ccc isomer of the dihydride Ru(CO)2(dppe)(H)2
5 in toluene-d8.13 In this species, one hydride ligand is trans
On the basis of these data, 5b can be concluded to be
a second type of solvent complex, Ru(CO)(dppe)(py)(H)2,
3 6 2 0
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8

View Article Online
Table 4 NMR data for the products formed in the reaction of Ru(CO)3(dppe) with 3 atm of p-H2 in pyridine-d5 at 335 K
Compound
d(1H)
d(31P{1H})
Ru(H)2(CO)2(dppe) (5)
−6.46, ddd, JHP = 22 Hz (cis-Pa), JHP = 72 Hz (trans-Pb), JHH = −5 Hz
−7.55, ddd, JHP = 19 Hz (cis-Pa), JHP = 27 Hz (cis-Pb), JHH = −5 Hz
−3.89, dd, JHP = 21 Hz (cis), JHH = −5 Hz
−4.36, dd, JHP = 104 Hz (trans), JHH = −5 Hz
−4.77, dd, JHP = 26 Hz (cis), JHH = −7 Hz
65.2, Pb, d, JPP = 13 Hz
71.7, Pa, d, JPP = 13 Hz
39.17, d, JPP = 50 Hz
Ru(H)2(CO)2(g1-dppe)(py) OC-6–13 (5a)
Ru(H)2(CO)2(g1-dppe)(py) OC-6–31 (5a)
Ru(H)2(CO)(dppe)(py) OC-6–13 (5b)
Ru(H)2(CO)(dppe)(py) OC-6–14 (5b)
49.5
−13.97, dd, JHP = 27 Hz (cis), JHH = −7 Hz
−4.46, ddd, JHP = 24 Hz (cis-Pa), JHP = 30 Hz (cis-Pb), JHH = −5 Hz
−4.65, ddd, JHP = 122 Hz (trans-Pb), JHP = 23 Hz (cis-Pa), JHH = −5 Hz
−4.61, unknown multiplicity
65.41, Pa, d, JPP = 16 Hz
89.05, Pb, d, JPP = 16 Hz
45.5, Pb
−16.38, ddd, JHP = 16 Hz (cis), JHP = 29 Hz (cis) JHH = −7 Hz
76.0, Pa
where the dppe ligand is coordinated in a bidentate fashion
and the pyridine moiety is cis to both hydrides (Fig. 5).
Comparable amounts of 5a to 5b are formed on the basis of
the corresponding hydride signal intensities (5:4). However, if
it is assumed that identical hydride enhancements are seen on
a per mole basis for each of the three products, 5a and 5b are
formed at 1% of the level of 5.
We have previously commented in this paper that the
appearance of the hydride resonance of cct-2-PPh3 at d −6.35
was unusual because the antiphase components of the p-H2
enhanced signal are separated by 2 × JHP rather than the more
usual JHH value. Since the two hydride ligands of cct-2-PPh3
are in chemically identical environments, an A2 spin system
would be expected and no signal enhancement should be seen.
The observation of an enhanced hydride resonance, however,
implies that these two nuclei actually belong to a complex
second-order spin system. Several examples of p-H2 based
signal enhancements under such circumstances have been
reported.28,41,42 The weak spin-spin coupling between the hydride
ligands and the twelve ortho-phenyl protons of cct-2-PPh3 results
in a complex spin system where the two chemically equivalent,
but strongly coupled, phosphine ligands add to the complexity.
This effect accounts for the hydride signal enhancement seen
in complexes of the type cct-Ru(CO)2(L)2(H)2. For analogous
cct-2 species containing different phosphines, the enhancement
is suggested to arise via similar interactions between the hydrides
and the corresponding protons on the phosphine ligands.
When a toluene-d8 solution of the complex 1-PMe3 was
heated in the presence of p-H2, no reaction was observed until
355 K; at this point hydride signals for both cct-2-PMe3 and
ccc-2-PMe3 were observed. This corresponds to the point where
Ru–CO bond breakage in 1-PMe3 occurs. In contrast to the PPh3
system, no monophosphine species such as Ru(CO)3(PMe3)(H)2
were detected, which suggests that the Ru–PMe3 bonds remain
intact under these conditions. Furthermore, the addition of a
small amount of PMe3 to the system resulted in substantial
hydride signal enhancements being observed for the known
complex Ru(H)2(CO)(PMe3)3.43 However, when a sample of
1-PMe3 was heated in the presence of p-H2 in pyridine-d5, the
enhanced hydride resonances of cct and ccc-2-PMe3 were seen
at 315 K. It can therefore be concluded that pyridine facilitates
the CO substitution process. The product distribution in this
reaction is significant and will be commented on later. However,
it should be noted at this point that the largest set of hydride
signals corresponded to a pair of doublets of antiphase doublets
at d −4.04 and −4.25 that arose from 4c-PMe3. This observation
suggests that pyridine also facilitates the loss of CO and PMe3
from 1-PMe3. Heating this sample further to 355 K in the
presence of p-H2 resulted in the observation of the second
isomer of the solvent complex, 4c-PMe3, as well as the double
solvent substitution product 6.
Thermal reactions of Ru(CO)3(L)2 complexes
Previous reports indicate that heating a toluene-d8 solution
of 1-PPh3 to 335 K in the presence of p-H2 yields a mixture
of cct-2-PPh3 (major product), ccc-2-PPh3, mer-3-PPh3 and
fac-3-PPh3.36 We have repeated this reaction in pyridine-d5
and found the same core product distribution as seen in
the photochemical studies described above, although two
additional hydride containing products were evident. The
first of these showed enhanced hydride resonances at d −4.44
and −13.73. Both these resonances appeared as doublets of
antiphase doublets due to cis 1H–31P couplings of 29 and
26 Hz, respectively; the 31P nucleus was located at d 27.0 by
HMQC methodology. The chemical shifts of these resonances
indicate that while the former is trans to CO, the latter is trans
to pyridine. This identifies this species as further isomer of the
pyridine solvent complex, 4c (see Fig. 2 for structure). The
identity of 4c was confirmed by repeating the experiment with
15N-labelled pyridine. Under these conditions, the resonance at
d −13.73 exhibited a 13 Hz trans coupling to 15N but in the case
of 4c-PPh3, the low intensity of the hydride resonances pre-
cluded the determination of the bound pyridine’s 15N chemical
shift. The coordinated pyridine ligand of 4c-PPh3 was, however,
located at d 248 by HMQC methods.
The second new product detected in these spectra, 6,
yielded hydride resonances at d −2.0 and −11.26. Both of
these appeared as simple antiphase doublets (JHH = −8 Hz),
suggesting that 6 does not contain a phosphine ligand and that
it therefore corresponds to the double substitution product
ccc-Ru(CO)2(py)2(H)2. It is worth noting that species of type
4c and 4c are also observed under thermal conditions when
20 lL of pyridine is added to toluene-d8 solution of 1-PPh3
under H2. In contrast, species 6 is only observed under these
conditions in neat pyridine. This concentration dependence
supports the assignment of 6 as the double substitution pro-
duct. We further note that both 4c and 6 were absent from the
photochemical studies and were only detectable under thermal
conditions.
Upon heating a sample of 1-P(p-tolyl)3 in pyridine-d5 to
355 K under p-H2, the complexes cct-2-P(p-tolyl)3, ccc-2-P(p-
tolyl)3, mer-3-P(p-tolyl)3 were seen, as were weaker signals for
the solvent complexes 4c-P(p-tolyl)3, 4c-P(p-tolyl)3 and 6. In
contrast, heating 1-PCy3 under identical conditions yielded
only cct-2-PCy3 and the solvent complexes 4c-PCy3, 4c-PCy3
and 6. The failure to observe mer-3 and fac-3 with the strongly
basic phosphines PMe3 and PCy3 will be commented on further
in the section on catalytic behaviour.
On examining Ru(CO)3(dppe) under p-H2 in pyridine-d5 at
315 K in the absence of photolysis, enhanced hydride signals
corresponding to species 5, 5a and 5b were again detected. In
contrast to photochemical investigations, where 5 was clearly
1
Another surprising feature of these p-H2 based H NMR
spectra was noted when they were recorded at or above 335 K,
in both pyridine-d5 and toluene-d8. Under these conditions,
an enhanced peak was visible at d +7.89. This corresponds to
a signal for the ortho-phenyl protons of the PPh3 ligands in
2
cct-2-PPh3. EXSY investigations and H labelling experiments
demonstrated that there was no exchange between hydride
and ortho-phenyl proton sites in this species. However, high-
resolution COSY spectra employing p-H2 enabled the detection
of a small (0.05 Hz) spin–spin coupling between the protons in
these two locations.
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
3 6 2 1

View Article Online
But2PCH2CH2PBut2 ligand. For the simpler PH3 system, isomer
C was found to be 5.5 kJ mol−1 less stable than isomer A.47 No
information was reported on isomer B. In a separate study, we
also carried out calculations on Ru(CO)2(H2PCH2CH2PH2)
which focused on studying the size of the singlet–triplet energy
gap in comparison with the analogous Fe system.15 We have
now carried out DFT calculations to examine the structural
arrangements and the relative energies of the three possible
isomers of Ru(CO)2(L)2 when L = PH3, PMe3, P(OMe)3,
AsMe3 and PF3. Since the chemistry involved here concerns
only the singlet state, and since the triplet was shown to
dominant, the ratio of 5:5a:5b under thermal conditions was
1:1:1, assuming identical extents of enhancement on a per-
mole basis. On increasing the temperature to 335 K, 5a became
the major species, exhibiting twice the signal intensities of the
other two. This indicates that heating Ru(CO)3(dppe) in pyridine
facilitates the de-chelation of the dppe ligand. In addition,
species 6 was observed, as were four new signals due to products
present at approximately 3% of the level of 5a. The first of these
species yielded signals, at d −4.77 and −13.97, which were shown
to couple to each other in the corresponding COSY spectrum.
Both these hydride signals appeared as doublets of antiphase
doublets, which indicates that they originate in a species that
contains an unchelated dppe ligand. The chemical shift of the
latter hydride resonance suggests it arises from a hydride ligand
that is trans to pyridine. This species is therefore assigned to
5a, an isomeric form of 5a (see Table 4 and Fig. 5). The final
two resonances, which also coupled to one another, appeared at
d −4.61 and −16.38. The multiplicity of the lower field resonance
could not be determined because of signal overlap; however,
the higher field hydride appeared as a doublet of doublets of
antiphase doublets. This splitting pattern is indicative of two cis
31P couplings and, given the chemical shift, arises from a hydride
that is trans to pyridine. This species therefore corresponds to
5b, a further isomer of 5b with the structure shown in Fig. 5.
It should be noted that the low signal intensity for 5a and 5b
precluded the determination of JPP values from the associated
HMQC spectra.
47
have a significantly higher energy for both the PH3 and the
H2PCH2CH2PH215 systems at ruthenium, we have confined our
calculations to the singlet state.
The calculations reported here were carried out using the
same level of theory previously employed for Fe(CO)2(PH3)2,48
where a benchmark investigation demonstrated that it gave the
most reliable results for Fe(CO)4.49
With PH3 as the phosphine, the most stable isomer was
found to correspond to A-PH3, which is 10 kJ mol−1 lower
in energy than C-PH3. This result parallels that already
reported by Eisenstein and coworkers.47 The corresponding
isomer with equatorial phosphines B-PH3 failed to optimise,
converting instead to A-PH3. Changing the phosphine to
trimethylphosphine gave a similar result, with isomer A-PMe3
now being 27.9 kJ mol−1 more stable than C-PMe3. The same
relative energy profile is also found with trimethyl phosphite
where A-P(OMe)3 is 16.0 kJ mol−1 lower in energy than
C-P(OMe)3. These data suggest that the introduction of an
increased p-accepting ability in the phosphine lowers the
energy difference between the two isomeric forms. To test this
hypothesis the strongly p-acidic phosphine PF3 was examined.
In this case the lowest energy isomer proved to be B-PF3, which
was now 11.6 kJ mol−1 lower in energy than C-PF3, which was
in turn 1.2 kJ mol−1 lower in energy than A-PF3. These data
confirm that the strongest p-acceptor prefers to locate itself
in an equatorial site as previously suggested.11 In the case of
L = AsMe3, the relative energies of A-AsMe3 and C-AsMe3
differ by 21.1 kJ mol−1, a difference which although lower
than that found for PMe3 is very similar to the P(OMe)3 value.
The similar relative proportions of the ccc and cct-L isomers
of Ru(CO)2(AsMe2Ph)2(H)2 observed experimentally cannot
therefore be attributed simply to the relative energies of these
intermediates.
Theoretical examination of the 16-electron Ru(CO)2(L)2
intermediates
We have already commented on the experimental determination
of the structure of Ru(CO)2(PMetBu2)2 as a trigonal bipyramid
with axial phosphines.11 The X-ray structure of this species
indicates that the associated vacant equatorial position is not
stabilised by an agostic interaction with C–H bonds of the
phosphine ligand. However, with less sterically demanding
phosphines, other isomers of such 16-electron species might
exist in solution. Three isomers A, B and C are possible as shown
in Fig. 6. Upon H2 addition to these 16-electron fragments,
three isomers of Ru(CO)2(L)2(H)2 can be formed, with (i) all
cis ligand arrangements, (ii) cis hydrides and CO’s and trans
phosphines and (iii) cis hydrides and cis phosphines and trans
CO’s. In the case of AsMe2Ph, as noted previously, all three of
these geometries can be detected in solution.13,42
Relevant optimised distances and angles for these species are
shown in Table 5. All optimised A isomers have very similar
geometries, with the axial P–Ru–P angle being close to 170° and
the equatorial OC–Ru–CO angle close to 135°. The axial Ru–P
bonds were found to bend slightly towards the empty equatorial
site. These parameters are in relatively good agreement with
those experimentally determined for Ru(CO)2(PMetBu2)2 where
the corresponding angles are 166.5° and 133.3° respectively.11
As shown previously,47 the potential energy surface along
the coordinate corresponding to the opening and closing of
these angles is rather flat, and wide structural changes may
be expected as a result of minor changes in the ligands. The
Ru–L distances are, however, quite similar for PH3 (2.304 Å)
and PMe3 (2.304 Å), whereas they significantly shorten on going
to the stronger p-acids P(OMe)3 (2.295 Å) and PF3 (2.229 Å)
but lengthen for AsMe3. These distances do, however, compare
well with the experimentally determined value of 2.357 Å for
the PMetBu2 complex.11 The equatorial Ru–CO distances show
a much smaller variation across the series.
Fig. 6 Relative energies calculated for the 16-electron intermediates.
Theoretical work has previously been devoted to the analysis
of RuL4 species.11,15,44–47 For the specific case of Ru(CO)2(L)2,
MP2 calculations have been carried out on A-PH3 to confirm
the preference for the experimentally observed nonplanar C2v
structure, whereas the isoelectronic [Rh(CO)2(PH3)2]+ species
prefers a planar geometry.11,46 Subsequent DFT studies at a
theory level closely related to ours have calculated the structure
of Ru(CO)2(H2PCH2CH2PH2), whose geometry is constrained
as C, to model the experimentally observed compound with a
The failure to detect a cct-CO H2 addition isomer in all the
phosphine systems examined here with p-H2 suggests that the
barrier to H2 addition over the P–Ru–P axis of isomer A of
Ru(CO)2(L)2 exceeds 20 kJ mol−1. This value is deduced on the
basis of the potential 15000-fold enhancement that could be
observed in this reaction with p-H2 at 253 K.
The optimised structures corresponding to isomer C show a
greater variability of angular parameters. However, the overall
3 6 2 2
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8

View Article Online
Table 5 Selected bond lengths (Å) and bond angles (°) for all calculated Ru(CO)2(L)2 fragments using B3PW91* level of theory
Compound
M–Lax/Å
M–Leq/Å
Lax–X/Å
Leq–X/Å
Lax–M–Lax/°
Leq–M–Leq/°
Ru(CO)2(PH3)2 (A-PH3)
Ru(CO)2(PH3)2 (C-PH3)
2.304
1.902
1.154
1.434
171.2
95.9
138.7
92.1
1.864 (C)
2.330 (P)
2.304
2.370 (P)
1.855 (CO)
2.309
1.859 (C)
2.358 (P)
1.875
2.364 (P)
1.855 (CO)
1.862
1.154 (CO)
1.425 (PH3)
1.850 (P)
1.841 (P)
1.159 (CO)
1.827
1.556 (CO)
1.427 (PH3)
1.159 (CO)
1.841 (P)
1.160 (CO)
1.165
Ru(CO)2(PMe3)2 (A-PMe3)
Ru(CO)2(PMe3)2 (C-PMe3)
170.0
151.5
135.3
151.5
Ru(CO)2(PPh3)2 (A-PPh3)
Ru(CO)2(PPh3)2 (C-PPh3)
Ru(CO)2(PF3)2 (A-PF3)
Ru(CO)2(PF3)2 (B-PF3)
Ru(CO)2(PF3)2 (C-PF3)
162.4
127.9
2.2291
1.938
2.250 (P)
1.918 (CO)
2.295
1.89864
2.178
1.886 (P)
1.886 (CO)
1.876
1.594
1.140
1.594 (P)
1.141 (CO)
1.622
1.1466
1.602
1.611 (P)
1.147 (CO)
1.160
166.89
167.22
164.3
135.93
123.5
128.4
Ru(CO)2{P(OMe)3}2
(A-P(OMe)3)
Ru(CO)2{P(OMe)3}2
(C-P(OMe)3)
Ru(CO)2(AsMe3)2 (A-AsMe3)
Ru(CO)2(AsMe3)2 (C-AsMe3)
169.3
143.9
133.5
150.4
2.300 (P)
1.871 (CO)
2.433
2.486 (As)
1.852 (CO)
2.323 (P)
1.871 (CO)
1.869
2.486 (As)
1.852 (CO)
1.626 (P)
1.155 (CO)
1.973
1.991 (As)
1.171 (CO)
1.638 (P)
1.155 (CO)
1.1179
1.978 (As)
1.172 (CO)
170.4
151.1
131.5
151.1
a All species were optimised in C1 symmetry.
difference between the axial and equatorial P–Ru–C bond
angles is always smaller than that found in isomer A and in the
case of PMe3 they are the same. Eisenstein and coworkers have
proposedthatsuchaligandarrangementfeaturesinthetransition
stateforsiteexchangewhenL = PH3;thiswasfoundtobeonly1.3
kJ mol−1 higher in energy than isomer C. Thus, depending on the
phosphine, this geometry may become a local energy minimum,
which again highlights the flatness of the potential energy surface
along this reaction coordinate. All the Ru–P distances were found
to be longer in isomer C than in the corresponding isomer A.
This is most likely a direct consequence of the trans influence of
CO vs. phosphine but may reflect the steric cost of bringing the
phosphine ligands closer together.
The geometry for isomer B yielded a stable minimum only
for the PF3 analogue. Here, the Ru–PF3 distance is shorter than
that in isomer A, a situation that is reversed when the Ru–CO
distance is considered. This observation confirms the greater p-
basicity of the metal with respect to the equatorial interactions,
a phenomenon that has been previously highlighted.44,46
systems under thermal conditions and confirm that it is the
relative energies of the 18-electron products, not the 16-electron
intermediate Ru(CO)2(L)2 that is important in controlling the
product distribution. This deduction is consistent with the fact
that at elevated temperatures, the corresponding AsMe2Ph
complexes proved to be in equilibrium.13,42
Mechanistic considerations at low temperature
In the low-temperature photochemical studies described here,
the quenching of ccc-2-PPh3 and ccc-2-PMe3 in coordinating
solvents such as pyridine suggests that isomer interchange via
the cct-L form at low temperatures is not responsible for the
observation of the ccc isomer.
In order to investigate this process further, authentic samples
of cct-2-PPh3 and cct-2-PCy3 were prepared and irradiated.
Under these conditions, concurrent 325-nm irradiation did
not lead to the observation of any new species in the corres-
1
ponding H NMR spectra. The photochemical formation of
Ru(CO)2L(H)2 from cct-2 does not therefore account for the
observation of ccc-2 or 4.
Relative ground-state energies of the cct-L and ccc isomers of
These observations suggest that the starting Ru(CO)3(L)2
complexes must be able to undergo photochemical ejection
of both CO and PPh3 to form the 14-electron intermediate
Ru(CO)2(L). This intermediate subsequently yields the solvent
dihydride complexes 4 and both ccc-2 and 3, a deduction
consistent with the fact that pyridine quenches the formation of
ccc-2 and 3. Further support for this proposal comes from the
observation that under photochemical conditions, the addition
of CO to a sample of 1 under p-H2 suppresses the formation of
all the dihydride species, while the addition of free phosphine
allows the formation of cct-2 but suppresses the formation
of the solvent complexes, as well as yielding the known
trisphosphine dihydride species Ru(CO)(PPh3)3(H)2.43
The formation of the 14-electron intermediate Ru(CO)2(L)
could occur in a one- or two-photon process. The use of a series
of neutral density filters (25, 50, 75%) enabled us to show the
ratio of cct-2-PPh3 to 4c-PPh3 was independent of the photon
flux and hence deduce that both compounds are formed in a
one-photon process.52 The existence of such processes has been
invoked previously in the case of related iron analogues.53,54
In this manuscript we have not described the photochemical
reactions of Ru(AsMe2Ph)2(CO)3 or Ru(AsPh3)2(CO)3 with p-H2
because the corresponding dihydride products only contain one
easily accessible NMR handle, and hence their unambiguous
characterisation is almost impossible. We note, however,
that when toluene-d8 solutions of Ru(AsMe2Ph)2(CO)3 is
irradiated at 255 K, signals for the corresponding ccc isomers
Ru(CO)2(L)2(H)2, L = PH3, PMe3 and AsMe3
In the case of reactions with monodentate-phosphine
containing precursors Ru(CO)3(L)2, the first species observed
upon photolysis proved to be cct-L Ru(CO)2(L)2(H)2, cct-2-L.
This product geometry can be formed by loss of an equatorial
CO ligand from 1, followed by H2 addition over the OC–Ru–CO
axis of intermediate A (Scheme 3). However, accounting for the
formation of ccc-2, as well as species 3 and 4, at low temperature
is more complex. For instance, ccc-2 could be formed by re-
arrangement of intermediate A to C followed by H2 addition,
or from cct-2 by ligand re-arrangement (Scheme 3). The relative
energies of the dihydrides therefore also need to be assessed.
The calculated structures for the dihydride species
Ru(CO)2(L)2(H)2 (where L = PH3, PMe3 and AsMe3) optimise
with distorted octahedral geometries (Table 6). Their basic
structure is, however, similar to that reported for the related Fe(II)
complex, tcc-Fe(CO)2[P(OPh)3]2H2.50,51 In the case of L = PH3,
calculations show that the ccc isomer of Ru(CO)2(PH3)2(H)2
is favoured over the cct-PH3 isomer by 12.7 kJ mol−1. In
contrast, when the L = PMe3 the cct-L isomer is favoured by
19.6 kJ mol−1 For AsMe3, the calculated relative energies of
.
the isomers at 295 K are much closer with the cct isomer being
9.42 kJ mol−1 more stable than the ccc isomer, and 17.71 kJ mol−1
more stable than the ccc-CO isomer. These relative energies are
consistent with experimentally observed isomer populations of
both the Ru(CO)2(PMe3)2(H)2 and Ru(CO)2(AsMe2Ph)2(H)2
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
3 6 2 3

View Article Online
Table 6 Selected calculated bond lengths (Å) and bond angles (°) for Ru(CO)2(L)2H2 species using B3PW91* level of theory
Compound
Ru–Lax/Å
2.296 (P)
Ru–Leq/Å
Lax–M–Lax/°
Leq–M–Leq/°
H–M–H/°
Ru(CO)2(PH3)2(H)2 cct-PH3
Ru(CO)2(PH3)2(H)2 ccc-PH3
Ru(CO)2(PMe3)2(H)2 cct-PMe3
Ru(CO)2(PMe3)2(H)2 ccc-PMe3
Ru(CO)2(PMe3)2(H)2 cct-CO
Ru(CO)2(AsMe3)2(H)2 ccc-AsMe3
Ru(CO)2(AsMe3)2(H)2 cct-CO
Ru(CO)2(AsMe3)2(H)2 cct-AsMe3
1.922 (CO)
1.659 (H)
2.358 (P)
1.923 (CO)
1.908 (CO)
1.665 (H)
2.376 (P)
156.7
158.2
157.7
158.1
157.6
157.7
158.2
157.4
100.2
100.7
99.6
82.0
82.7
82.7
82.9
84.8
82.9
84.4
81.9
2.329 (P)
1.882 (CO)
2.324 (P)
2.353 (P)
1.879 (CO)
1.906 (CO)
97.2
1.907 (CO)
2.384 (P)
101.1
98.5
1.635 (H)
2.477 (As)
1.910 (CO)
2.488 (As)
1.636 (H)
1.915 (As)
1.678 (H)
2.452 (As)
1.868 (CO)
1.907 (CO)
101.3
99.6
2.428 (As)
Scheme 3 Summary of mode of addition of H2 to Ru(CO)2(PPh3)2 fragments and subsequent reactions.
rate quantified by EXSY methods. This was achieved via the
monitoring of the transfer of magnetisation from the hydride
resonance of cct-2-PPh3 into the hydrogenation product and free
hydrogen (Fig. 7) as a function of reaction time. Analysis of these
spectra was achieved by simulation.55–58 The rate of H2 transfer
from 2-PPh3 into cis-stilbene proved to be 2.2 s−1 while the rate
of H2 elimination was 3.1 s−1. The observed rate constants
obtained for these two processes proved to be independent
of the alkyne excess relative to 1-PPh3. In comparison, the
related trinuclear species Ru3(H)(l-H)(CO)9(PPh3)2 catalyses
hydrogenation under much milder conditions (an apparent
rate of 0.31 s−1 was observed at 300 K)57,58 and at 355 K,
hydrogenation by the cluster is fast on the NMR time scale,
indicating (qualitatively) that the cluster is a better catalyst.
of the dihydrides are visible at d −6.70 and −8.44 and the tcc
isomer at d −7.29. Upon changing the solvent to pyridine-d5,
the size of the signals due to the ccc isomers is substantially
reduced, although on running these reactions at 295 K they
again become strongly enhanced. These additional data fully
support the previous deductions (Scheme 3). However, these
are not the only signals that are seen in these spectra, and the
situation with AsPh3 is even more complex. These reactions are
currently being explored in more detail.
Catalytic hydrogenation studies
A sample of 1-PPh3 was then examined by NMR spectroscopy
in pyridine-d5 in the presence of p-H2 and a 100-fold excess of
diphenylacetylene (PhCCPh). When a sample was introduced
into the NMR spectrometer at 335 K in the absence of photolysis,
the hydride-containing species cct-2-PPh3 and 4c-PPh3 were
1
immediately observed in the corresponding H NMR spectra.
As the sample temperature increases to the set point, the signals
for 4c-PPh3 vanish and enhanced resonances due to cis-stilbene
become evident, indicative of a thermally initiated hydrogenation
reaction. It should be noted that when the experiment was
repeated with concurrent UV irradiation, the intensity of
both the hydride resonance of 2-PPh3 and the enhanced cis-
stilbene peak increased, which suggests that photochemical
promotion of this reaction, although not required, is possible.7
At 355 K, hydrogenation could be monitored directly and the
Fig. 7 EXSY spectrum depicting magnetisation transfer from cct-2-
PPh3 to cis-stilbene and free H2.
3 6 2 4
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8

View Article Online
When toluene-d8 was used as the solvent, no transfer of
magnetisation from cct-2-PPh3 to cis-stilbene was observed at
355 K, but a p-H2 enhanced signal for the hydrogenation pro-
duct cis-stilbene was evident. It can therefore be concluded that
catalytic hydrogenation does occur in toluene, but the associated
rate is too slow to allow direct magnetisation transfer to be
observed.
the situation with the trinuclear precursor Ru3(CO)10(PPh3)2,
where only cis-stilbene was observed, even at prolonged reaction
times.57,58 Another interesting feature of this investigation
was the observation of 31P resonances corresponding to free
phosphine ligands in all cases, which confirms that phosphine
loss takes place during catalysis.
When Ru(CO)3(dppe) was examined as the catalytic pre-
cursor, the resonances for 5, 5a and 5b were again observed,
as were polarised resonances for the hydrogenation product
cis-stilbene. Interestingly, the resonances for 5b were 5 times
more intense than without substrate, indicating that this
species now cycles hydrogen more rapidly and hence is an active
hydrogenation catalyst. However, the rate of H2 transfer to the
hydrogenation product was again too slow to monitor directly
by EXSY methods.
To examine this situation further, samples of 1-PPh3, 1-PCy3,
1-P(p-tolyl)3, 1-PMe3 and Ru(CO)3(dppe) were prepared with
identical metal complex and substrate concentrations and
were examined for catalytic hydrogenation under identical
temperatures and H2 pressures. Based on the build-up of
cis-stilbene resonances over a period of 5 min, the catalytic
activity order is PPh3 > P(p-tolyl)3 > PMe3 > PCy3 > dppe.
The trend clearly shows that the electronic properties of the
phosphine play a role in controlling catalysis with the more
strongly donating phosphines being less suitable. Since catalysis
is known to involve loss of a phosphine, it is clear that Ru–P
bond breakage becomes more difficult as the basicity of the
phosphine is increased, and hence catalysis is retarded. This is in
agreement with the observation that 1-PCy3 and 1-PMe3, with
strongly basic phosphines, do not form the monophosphine
species mer-3 and fac-3 under thermal conditions in toluene. In
the case of dppe, the chelating nature of the phosphine makes
phosphine loss less favourable and thus accounts for the low
catalytic activity observed with this ligand.
Since cct-2-PPh3 is coordinatively saturated, ligand loss must
take place to facilitate catalysis. While the addition of 1 atm
of CO to the system decreased the intensities of the hydride
resonance for cct-2-PPh3 and the cis-stilbene peak, the measured
rate of hydrogen transfer from 2-PPh3 to cis-stilbene was
unchanged. In contrast, the addition of a 50-fold excess of PPh3
to the system essentially quenches the hydrogenation process.
It can therefore be concluded that the first step in catalytic
hydrogenation by 2-PPh3 involves phosphine loss and the
formation of the 16-electron intermediate Ru(CO)2(PPh3)(H)2.
Although the observation of the solvent complex 4c-PPh3
is suppressed by the addition of PhCCPh, Ru(CO)2(PPh3)-
(PhCCPh)(H)2 is not detected at 355 K. This species has
however been detected when Ru3(CO)10(PPh3)2 was examined
with p-H2 and PhCCPh in C6D6 at 300 K.57,58 A complete
catalytic cycle can therefore be assembled for this reaction, as
shown in Scheme 4. In order to account for the dramatic solvent
participation in the hydrogenation rate, we suggest that solvent
participation assists in Ru–PPh3 bond breakage.
When a sample of 1-PPh3 was examined in the presence
of diphenylacetylene, p-H2 and CO, very slow build-up
of a resonance at d 9.84 due to an aldehydic proton was
observed. This resonance is indicative of the presence of slow
hydroformylation.
Scheme 4 Proposed catalytic cycle for hydrogenation of diphenyl-
acetylene by cct-2-PPh3.
Conclusions
In this paper we have shown that photolysis of the compounds
Ru(CO)3(L)2, where L = PPh3, PMe3, PCy3 and P(p-tolyl)3, in
the presence of hydrogen, yields a product distribution that
is dependent on the temperature and solvent. However, one
species, cis-cis-trans-L Ru(CO)2(L)2(H)2, is always produced,
and is characterised by an unusual parahydrogen enhanced
hydride resonance. This species proved unsuitable for quantum
computation experiments, however, the study of this system has
yielded valuable insights into the mechanisms of H2 addition to
d8 Ru(0) complexes and their subsequent catalytic behaviour.
Photochemical addition of H2 to Ru(CO)3(L)2 has
been shown to proceed via initial loss of CO and the
subsequent reaction of the Ru(CO)2(L)2 intermediate with
hydrogen. Previous work indicates that the related species
Ru(CO)2(PMetBu2)2 adopts a trigonal-bipyramidal geometry
with axial phosphines and a vacant equatorial site.11 Hence, H2
addition across the more p-accepting OC–Ru–CO axis readily
accounts for the observation of cct-L Ru(CO)2(L)2(H)2, while
addition across the L–Ru–L axis is not observed, even with
p-H2 amplification. Literature studies on the laser-initiated
reaction of Ru(CO)3(dmpe) (dmpe = Me2PCH2CH2PMe2)
with H2 have described how CO dissociation leads to the rapid
formation of Ru(CO)2(dmpe)(solvent). On a longer timescale,
reaction with H2 leads to the generation of the stable complex
Ru(CO)2(dmpe)(H)2.59
The rate of hydrogenation by cct-2-PPh3 proved to be
independent of the dihydrogen concentration. This indicates
that dihydrogen transfer to the organic substrate and subsequent
product elimination must occur before another H2 molecule
binds to the active species. This contrasts with the situation
observed for Ru3(H)(l-H)(CO)9(PPh3)2, where the rate of
hydrogenation was found to depend on the H2 concentration.57,58
When samples of 1-PMe3, 1-P(p-tolyl)3 and 1-PCy3 were
investigated as catalytic precursors, enhanced resonances for
cct-2 and cis-stilbene protons were again observed at 355 K,
which are indicative of catalytic hydrogenation. However, the
intensity of the cis-stilbene resonance was, in all three cases,
lower than that seen with 1-PPh3, and no direct magnetisation
transfer could be observed from cct-2 species into cis-stilbene.
This indicates that systems based on 1-PCy3, 1-P(p-tolyl)3 and
1-PMe3 are not as catalytically active as 1-PPh3. However,
reductive elimination of H2 from cct-2-P(p-(tolyl)3 and cct-
2-PMe3 was still rapid enough to follow by EXSY methods.
The associated rate constants at 355 K were 2.7 and 1.3 s−1,
respectively. Both these rates are slower than those found for
cct-2-PPh3 under identical conditions and parallel the electron
donating ability of the phosphine, which is expected to stabilise
the dihydrides.
At long reaction times (16 h under catalytic conditions), stable
resonances corresponding to cct-2 and significant amounts of
cis-stilbene were observed with all catalysts. In addition, slow
isomerisation to trans-stilbene was evident in all cases, with
approximately 5% of cis-stilbene being converted. This contrasts
When the reaction is carried out in toluene, the photochemical
formation of a second isomer of this species, cis-cis-cis
Ru(CO)2(L)2(H)2, does occur. Previous work has shown that
this product is stable in its own right when L = AsMe2Ph,12,13
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
3 6 2 5

View Article Online
but with phosphine ligands this minor isomer is only
observed when Ru(CO)2(L)2(H)2 is warmed with p-H2, at
which point the signal amplification of the associated hydride
resonances allows its detection. The formation of cis-cis-cis
Ru(CO)2(L)2(H)2 cannot be achieved by the direct addition of
H2 to an Ru(CO)2(L)2 intermediate that is isostructural with
Ru(CO)2(PMetBu2)2. However, it may be possible to form
this product by H2 addition to a minor intermediate with
cis phosphines. Theoretical calculations have revealed that
when L = PMe3, the isomer A-PMe3 with trans phosphines
is 27.9 kJ mol−1 more stable than isomer C-PMe3 with cis
phosphines (Fig. 6) necessary to form ccc-PMe3 directly. If it
is assumed that the p-H2 based signal enhancement is at the
theoretical maximum (a factor of 15000 is available with a
simple 45° read pulse),21 an isomer with an energy difference
of 20 kJ mol−1 could be detected. The calculations therefore
indicate that cis-cis-cis Ru(CO)2(PMe3)2(H)2 does not arise
because of H2 addition to C-PMe3.
We have also described how the in-situ irradiation of
Ru(CO)3(L)2 led to the detection of the solvent complexes
Ru(CO)2(L)(sol)(H)2 (where sol is toluene, THF and pyridine).
These species were shown to originate from the minor
photoproduct Ru(CO)2(L), which is formed via a single photon
pathway. This intermediate was shown to be responsible for
the generation of ccc-Ru(CO)2(L)2(H)2 as well as fac and mer
isomers of Ru(CO)3(L)(H)2 at 253 K. This example serves to
illustrate how the enhanced signal intensities provided by para-
hydrogen in conjunction with high-level DFT calculations can
be used to elucidate the mechanism of H2 addition to a metal
complex, and differentiate the mechanisms by which the various
product isomers are formed.
Compounds of the type Ru(CO)3(L)2 also undergo thermal
addition of H2, and are active hydrogenation catalysts at
elevated temperatures. The activity of the system is dependent
on the phosphine, such that the most active system is produced
when L = PPh3. The catalytic activity also depends on the
solvent, such that strongly coordinating solvents facilitate
the catalytic process. In the most favourable case, the thermal
generation of cct-PPh3 Ru(CO)2(PPh3)2(H)2 in pyridine enabled
direct magnetisation transfer into the hydrogenation product cis-
stilbene to be observed. The mechanism of this transformation
has been elucidated based on the direct detection of all the
intermediates in the catalytic cycle and a rate of hydrogenation
has been determined. At 355 K, catalytic hydrogenation
proceeds via rate-limiting phosphine dissociation with a rate
constant of 2.2 s−1. The mechanistic information provided by
this study represents the first step towards designing improved
hydrogenation catalysts based on d8 Ru(0) systems.
1H NMR chemical shifts are reported in ppm relative to
residual 1H signals in the deuterated solvents (toluene-d7, d 2.13,
THF-d7, d 1.73 and pyridine-d4, d 8.74), 31P NMR in ppm
downfield of an external 85% solution of phosphoric acid, 13C
NMR relative to toluene-d8, d 21.3 and pyridine-d5, d 150.35
and 15N relative to an external sample of CH3NO2. Modified
COSY, HMQC and EXSY pulse sequences were used as
previously described.61–63
In the hydrogenation studies, 1D EXSY spectra were
acquired immediately after exposing the sample to hydrogen
and introducing it into the probe. The results were modelled
allowing for hydride transfer into the hydrogenation product,
as described previously.55–58 The rate constants obtained in this
way were multiplied by two to take into account the analysis
method.64
Synthesis of Ru(CO)3(L)2
1-PPh3, 1-PMe3, Ru(CO)3(dppe), Ru(AsMe2Ph)2(CO)3 and
Ru(AsPh3)2(CO)3 were prepared by literature procedures6,65–67
and purified by recrystallisation from a hot solution of
1:1 THF–hexane. The synthesis of Ru(CO)3(PCy3)2,
1-PCy3, followed a literature procedure66 with purification
by recrystallisation from a hot solution of 1:1 THF–hexane.
NMR (toluene-d8, 295 K): dH 1.34 (m), 1.68 (m), 1.85 (m), 2.27
(m), 2.41 (m); dP 49.0, (s); dC 202.0 (t, JC–P = 9 Hz, CO), 38.60 (t,
JC–P = 11 Hz, Cipso), 30.7 (s, Cortho), 28.2 (t, JC–P = 5 Hz, Cpara), 27.0
(s, Cmeta). Crystals of 1-PCy3 suitable for X-ray diffraction were
grown at room temperature from a 1:1 THF–hexane mixture.
ThecomplexRu(CO)3[P(p-tolyl)3]2, 1-P(p-tolyl)3 wasprepared
by a modified literature procedure.68 2.7 g (9 mmol) of P(p-
tolyl)3 was dissolved in 100 ml of degassed 2-methoxyethanol
and heated to reflux. To this solution was added 0.39 g
(1.5 mmol) of RuCl3·3H2O dissolved in 30 ml of degassed
2-methoxyethanol, followed rapidly by 40 ml of degassed
aqueous formaldehyde (37% wt) and 0.6 g potassium hydroxide
dissolved in 30 ml of 2-methoxyethanol. This yellow solution
was then reflux under nitrogen for 4 h by which time the product
Ru(CO)3[P(p-tolyl)3]2 precipitated as yellow microcrystals. The
solution was allowed to cool and the product filtered off. The
yellow solid was washed with 20 ml of cold ethanol, then 10 ml
of ice-cold water followed by 10 ml of ethanol and 30 ml of
hexane. Ru(CO)3[P(p-tolyl)3]2 was recrystallised from a boiling
solution of THF–hexane (2:1) to yield 845 mg (75% yield) of
a pale yellow solid.
Preparation of Ru(CO)2(L)2(H)2
To prepare the dihydride complexes Ru(CO)2(L)2(H)2, where
L = PCy3 or PPh3, a 50 mg sample of the corresponding
complex Ru(CO)3(L)2 was dissolved in a minimum amount of
dry toluene and degassed by three freeze–pump–thaw cycles.
The sample was then filled with 1 atm of H2 and photolysed
under UV light (Xe-arc lamp) for 2 h, during which period the
solution was degassed and refilled with fresh H2 every 30 min.
A gradual colour change from yellow to colourless was observed.
The solvent was then removed in vacuo to yield a white powder
corresponding to Ru(CO)2(L)2(H)2, which was characterised by
NMR spectroscopy and used without further purification.
Experimental
General conditions and reagents
All reactions and purifications were carried out under nitrogen
using glove-box, high-vacuum or Schlenk-line techniques. All
chemicals were purchased from commercial suppliers and used
without further purification.
NMR experiments
All NMR solvents (Apollo Scientific) were dried using
appropriate methods and degassed prior to use. The NMR
measurements were made using NMR tubes fitted with
J. Young Teflon valves and solvents were added by vacuum
transfer on a high vacuum line. For the parahydrogen induced
polarization (PHIP) experiments, hydrogen enriched in the
para spin state was prepared by cooling H2 to 20 K over a para-
magnetic catalyst (activated charcoal) as described previously.60
All NMR studies were carried out with sample concentrations
of approximately 1 mM and spectra were recorded on a
In-situ photolysis
This was achieved using a modified NMR probe that was
equipped for in-situ photolysis, as described previously.16
A Kimmon IK3202R-D 325-nm He–Cd 27 mW continuous
wave (CW) laser was used as the radiation source.
Computational details
All calculations were performed using the Gaussian 03 program69
together with the modified form of the B3PW91 functional in
conjunction with a flexible polarisable basis sets.70 Specifically,
1
Bruker DRX-400 spectrometer with H at 400.13 MHz, 31P at
161.9 MHz, 13C at 100.0 MHz and 15N at 40.5 MHz, respectively.
3 6 2 6
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8

View Article Online
24 R. Eisenberg, Acc. Chem. Res., 1991, 24, 100.
25 J. Natherer and J. Bargon, Prog. Nucl. Magn. Reson, Spectrosc.,
1997, 31, 293.
26 S. B. Duckett and C. J. Sleigh, Prog. Nucl. Magn. Res. Spectrosc.,
1999, 34, 71.
27 S. B. Duckett and D. Blazina, Eur. J. Inorg. Chem., 2003, 2901.
28 S. Aime, R. Gobetto and D. Canet, J. Am. Chem. Soc., 1998, 120,
6770.
29 S. Aime, W. Dastru, R. Gobetto, A. Russo, A. Viale and D. Canet,
J. Phys. Chem. A., 1999, 103, 9702.
30 A. B. Permin and R. Eisenberg, J. Am. Chem. Soc., 2002, 124,
12406.
31 S. Aime, R. Gobetto, F. Raineri and D. Canet, J. Chem. Phys.,
2003, 119, 8890.
the c3 coefficient in Becke’s original three-parameter fit to
thermochemical data71 was changed to 0.15, to give the B3PW91*
functional. Atoms C, O, P and As were described by the
triple-zeta basis sets of Schäfer et al.72 augmented by one d
polarisation function on C, O, P, As (a = 0.8, 1.2, 0.55 and
0.34, respectively). The Ru atom was described with the SDD
basis set,70 which uses the Stuttgart/Dresden ECP and double-
zeta functions for all valence electrons, augmented with an
f polarisation function (a = 1). All minima were fully optimised
and characterised by computing vibrational frequencies at the
same level of theory.
32 A. Koch and J. Bargon, Magn. Reson. Chem., 2000, 38, 216.
33 M. Stephan, O. Kohlman, H. G. Niessen, A. Eichhorn and
J. Bargon, Magn. Reson. Chem., 2002, 40, 157.
34 S. A. Colebrooke, S. B. Duckett and J. A. B. Lohman, Chem.
Commun., 2000, 685.
35 R. Gobetto, L. Milone, F. Reineri, L. Salassa, A. Viale and
E. Rosenberg, Organometallics, 2002, 21, 1919.
36 C. J. Sleigh, S. B. Duckett, R. J. Mawby and J. P. Lowe, Chem.
Commun., 1999, 1223.
Crystallography
Experimental details are collected in Table 2 and reported in the
CIF file.
CCDC reference number 244908.
See http://www.rsc.org/suppdata/dt/b4/b410912k/ for crystal-
lographic data in CIF or other electronic format.
37 R. H. Heyn, S. A. Macgregor, T. T. Nadasdi, M. Ogasawara,
O. Eisenstein and K. G. Caulton, Inorg. Chem. Acta., 1997, 259, 5.
38 A. C. Cooper, W. E. Streib, O. Eisenstein and K. G. Caulton, J.
Am. Chem. Soc., 1997, 119, 9069.
39 F. Dahan, S. Sabo and B. Chaudret, Acta. Crystallogr. Sect. C:
Cryst. Struct. Commun., 1984, 40, 786.
40 R. A. Jones, G. Wilkinson, A. M. R. Galas, M. B. Hursthouse and
K. M. A. Malik, J. Chem. Soc., Dalton Trans., 1980, 1771.
41 M. Haake, J. Barkemeyer and J. Bargon, J. Phys. Chem., 1995, 99,
17539.
Acknowledgements
We are grateful to the EPSRC, and the European Union for
funding under the HYDROCHEM network (contract HPRN-
CT-2002-00176) and to CINES for a grant of free computa-
tional time. We are also grateful for discussions with Prof. R. N.
Perutz, Dr R. J. Mawby, Dr J. N. Moore and Dr J. Lynam and
experimental help from Dr P. Callaghan.
42 S. B. Duckett, R. J. Mawby and M. G. Partridge, Chem. Commun.,
1996, 383.
References
43 C. J. Sleigh, DPhil thesis, University of York, 1999.
44 M. Elian and R. Hoffmann, Inorg. Chem., 1975, 14, 1058.
45 J. Li, G. Schreckenbach and T. Ziegler, J. Am. Chem. Soc., 1995,
117, 486.
1 L. Vaska and J. DiLuzio, J. Am. Chem. Soc., 1961, 83, 2784.
2 S. K. Hasnip, S. B. Duckett, C. J. Sleigh, D. R. Taylor, G. K. Barlow
and M. J. Taylor, Chem. Commun., 1999, 1717.
3 S. K. Hasnip, S. A. Colebrooke, C. J. Sleigh, S. B. Duckett, D. R.
Taylor, G. K. Barlow and M. J. Taylor, J. Chem. Soc., Dalton
Trans., 2002, 743.
4 D. Evans, J. A. Osborn, F. H. Jardine and G. Wilkinson, Nature,
1965, 208, 1203.
5 D. Evans, J. A. Osborn and J. Wilkinson, J. Chem. Soc. A, 1968,
3133.
6 R. A. Sanchez-Delgado, J. S. Bradley and G. Wilkinson, J. Chem.
Soc., Dalton Trans, 1976, 399.
7 E. M. Gordon and R. Eisenberg, J. Organomet. Chem., 1986, 306,
C53.
46 M. Ogasawara,
S. A. Macgregor,
W. E. Streib,
K. Folting,
O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1996, 118,
10189.
47 T. Gottschalk Gaudig, J. C. Huffman, H. Gérard, O. Eisenstein
and K. G. Caulton, Inorg. Chem., 2000, 39, 3957.
48 J. N. Harvey and R. Poli, Dalton Trans., 2003, 4100.
49 J. Harvey and M. Aschi, Faraday Discuss., 2003, 124, 129.
50 V. Arion, J. J. Brunet and D. Neibecker, Inorg. Chem., 2001, 40,
2628.
51 Nan Nhat Ho, R. Bau and S. A. Mason, J. Organomet. Chem.,
2003, 676, 85.
52 I. K. Lednev, T. Q. Ye, L. C. Abbott, R. E. Hester and J. N. Moore,
J. Phys. Chem. A, 1998, 102, 9161.
53 V. Balcher, F. W. Grevels, K. Kerpen, G. Olbrich and K. Schaffner,
Organometallics, 2003, 22, 1696.
54 M. K. Kuimova, W. Z. Alsindi, J. Dyer, D. C. Grills, O. S. Jina,
P. Matousek, A. W. Parker, P. Portius, X. Z. Sun, M. Towrie,
C. Wilson, J. Yang and M. W. George, Dalton Trans., 2003, 3996.
55 W. D. Jones, G. P. Rosini and J. A. Maguire, Organometallics, 1999,
18, 1754.
8 S. Murai, F. Kakiuchi, S. Sekini, Y. Tanaka, A. Kamatani,
M. Sonoda and N. Chatani, Nature, 1993, 366, 529.
9 R. F. R. Jazzar, M. F. Mahon and M. K. Whittlesey, Organo-
metallics, 2001, 20, 3745.
10 R. Noyori and T. Ohkuma, Angew. Chem., Int. Ed., 2000, 40, 40.
11 M. Ogasawara,
S. A. MacGregor,
W. E. Streib,
K. Polting,
O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1995, 117,
8869.
12 J. M. Bray and R. J. Mawby, J. Chem. Soc., Dalton Trans., 1987,
2989.
56 D. Blazina, S. B. Duckett, P. J. Dyson, B. F. G. Johnson, J. A. B.
Lohman and C. J. Sleigh, J. Am. Chem. Soc., 2001, 123, 9760.
57 D. Blazina, S. B. Duckett, P. J. Dyson and J. A. B. Lohman, Angew.
Chem., Int. Ed., 2001, 40, 3874.
58 D. Blazina, S. B. Duckett, P. J. Dyson and J. A. B. Lohman, Chem.
Eur. J., 2003, 9, 1046.
59 M. K. Whittlesey, R. N. Perutz, I. P. Virrels and M. W. George,
Organometallics, 1997, 16, 268.
60 M. S. Anwar, D. Blazina, H. A. Carteret, S. B. Duckett, T. K.
Halstead, J. A. Jones and R. J. K. Taylor, Magn. Reson. Chem.,
submitted.
61 B. A. Messerle, C. J. Sleigh, M. G. Partridge and S. B. Duckett,
J. Chem. Soc., Dalton Trans., 1999, 1429.
62 S. A. Colebrooke, S. B. Duckett and J. A. B. Lohman, Chem.
Commun., 2000, 685.
63 P. Hubler and J. Bargon, Angew. Chem., Int. Ed., 2000, 39, 3701.
64 M. L. H. Green, L. L. Wong and A. Sella, Organometallics, 1992,
11, 2660.
13 S. B. Duckett, R. J. Mawby and M. G. Partridge, Chem. Commun.,
1996, 383.
14 D. Schott, C. J. Sleigh, J. P. Lowe, S. B. Duckett and R. J. Mawby,
Inorg. Chem., 2002, 41, 2960.
15 D. Schott, J. P. Dunne, S. B. Duckett, C. Godard, J. N. Harvey,
R. J. Mawby, G. Muller, R. N. Perutz, R. Poli and M. K. Whittlesey,
Dalton Trans., 2004, 3218.
16 C. Godard, P. L. Callaghan, J. L. Cunningham, S. B. Duckett,
J. A. B. Lohman and R. N. Perutz, Chem. Commun., 2002, 2836.
17 S. Geftakis and G. E. Ball, J. Am. Chem. Soc., 1999, 121, 6336.
18 N. A. Knatochnil, J. A. Pachinson, P. J. Bedmaishi and P. J. Sadler,
Angew. Chem., Int. Ed., 1999, 38, 1460.
19 C. E. Lyon, J. L. Lopez, B. M. Cho and P. J. Hore, Mol. Phys., 2002,
100, 1261.
20 T. Kühn and H. Schwalbe, J. Am. Chem. Soc., 2002, 122, 6169.
21 M. S. Anwar, D. Blazina, H. A. Carteret, S. B. Duckett, T. K.
Halstead, J. A. Jones, C. M. Kozak and R. J. K. Taylor, Phys. Rev.
Lett., 2004, 93, 40501.
65 K. Stott, J. Stonehouse, J. Keeler, T. L. Hwang and A. J. Shaka,
J. Am. Chem. Soc., 1995, 117, 4199.
66 L. Song and W. C. Trogler, J. Am. Chem. Soc., 1992, 114, 3355.
67 G. Bellachioma, G. Cardaci, A. Macchioni and A. Madami, Inorg.
Chem., 1993, 32, 554.
22 C. R. Bowers and D. P. Weitekamp, J. Am. Chem. Soc., 1987, 109,
5541.
23 T. C. Eisenschmid, R. U. Kirss, P. A. Deutsch, S. I. Hommeltoft,
R. Eisenberg, J. Bargon, R. G. Lawler and A. L. Balch, J. Am.
Chem. Soc., 1987, 109, 8089.
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8
3 6 2 7

View Article Online
68 N. Ahmed, S. D. Robinson and M. F. Uttley, J. Chem. Soc., Dalton
Trans., 1972, 843.
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, 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, C. Gonzalez, and J. A.
Pople, Gaussian, Inc., Pittsburgh, PA, 2003.
69 Gaussian 03, Revision B.05. M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A.
Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M.
Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi,
G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida,
T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li,
J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,
70 J. Harvey and M. Aschi, Faraday Discuss., 2003, 124, 129.
71 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
72 A. Schaefer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97,
2571–2577.
3 6 2 8
D a l t o n T r a n s . , 2 0 0 4 , 3 6 1 6 – 3 6 2 8