Synthesis and characterisation of [(triphos)Fe(CO)H2] and its protonation to a dihydrogen complex via an unconventional hydrogen-bonded intermediate

Article abstract of DOI:10.1039/b801652f

The complex [(triphos)Fe(CO)H₂] 1 has been prepared and characterised by ¹H and ³¹P NMR and IR spectroscopies, and by single-crystal X-ray diffraction. Complex 1 is fluxional at ambient temperatures in solution; NMR spectral simulation shows that both a Bailar twist of the triphos ligand and pairwise hydride exchange contribute to this fluxionality, with the latter process apparently more facile. DFT-calculated vibrational frequencies for 1 indicate that both these processes are thermally accessible at ambient temperatures. The X-ray crystal structure of complex 1 shows that the triphos ligand adopts a fac arrangement in a distorted octahedral structure, with the two hydride ligands mutually cis, and a non-linear Fe-C-O moiety. Protonation of complex 1 with hexafluoroisopropanol (HFIP) initially produces a hydrogen-bonded intermediate; reaction proceeds further to give the cationic complex [(triphos)Fe(CO)H(η²-H₂)] ⁺2. Complex 2 decomposes above 250 K in solution, but has been characterised by NMR spectroscopy. T_{1 min} and J_HD measurements indicate that 2 contains a stretched dihydrogen ligand, with r _HH = 1.03 A, in apparent disagreement with DFT calculations that suggest that the equilibrium H-H distance is 0.85 A, indicative of a conventional dihydrogen moiety. However, the very shallow nature of the potential energy surface for the Ru(η²-H₂) moiety can account for this difference between experimental and calculated values. The experimental and computed structural and spectroscopic features deduced for 1 and 2 are compared with those of their known ruthenium analogues, and are discussed in the context of other hydride and dihydrogen complexes of the Group 8 metals. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique.

Full text of DOI:10.1039/b801652f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Synthesis and characterisation of [(triphos)Fe(CO)H2] and its
protonation to a dihydrogen complex via an unconventional
hydrogen-bonded intermediatewz
Gemma Guilera,a G. Sean McGrady,*b Jonathan W. Steed,*c
Richard P. L. Burchell,b Peter Sirschb and Anthony J. Deemingd
Received (in Montpellier, France) 29th January 2008, Accepted 9th April 2008
First published as an Advance Article on the web 15th May 2008
DOI: 10.1039/b801652f
1
The complex [(triphos)Fe(CO)H2] 1 has been prepared and characterised by H and 31P NMR
and IR spectroscopies, and by single-crystal X-ray diffraction. Complex 1 is fluxional at ambient
temperatures in solution; NMR spectral simulation shows that both a Bailar twist of the triphos
ligand and pairwise hydride exchange contribute to this fluxionality, with the latter process
apparently more facile. DFT-calculated vibrational frequencies for 1 indicate that both these
processes are thermally accessible at ambient temperatures. The X-ray crystal structure of
complex 1 shows that the triphos ligand adopts a fac arrangement in a distorted octahedral
structure, with the two hydride ligands mutually cis, and a non-linear Fe–C–O moiety.
Protonation of complex 1 with hexafluoroisopropanol (HFIP) initially produces a hydrogen-
bonded intermediate; reaction proceeds further to give the cationic complex
[(triphos)Fe(CO)H(Z2-H2)]+ 2. Complex 2 decomposes above 250 K in solution, but has been
characterised by NMR spectroscopy. T1 min and JHD measurements indicate that 2 contains a
stretched dihydrogen ligand, with rHH = 1.03 A, in apparent disagreement with DFT calculations
that suggest that the equilibrium H–H distance is 0.85 A, indicative of a conventional dihydrogen
moiety. However, the very shallow nature of the potential energy surface for the Ru(Z2-H2)
moiety can account for this difference between experimental and calculated values. The
experimental and computed structural and spectroscopic features deduced for 1 and 2 are
compared with those of their known ruthenium analogues, and are discussed in the context of
other hydride and dihydrogen complexes of the Group 8 metals.
play—in many metal-catalysed reactions involving CO and
Introduction
H2. These include such important industrial processes as the
water gas shift reaction, Fischer Tropsch catalysis, hydrofor-
mylation, reductive carbonylation and hydrogenation of
alkenes.2 The involvement in many of these processes of
molecular hydrogen complexes is becoming increasingly
widely realised.3 Furthermore, complexes of the type
[L4MH2] and their protonation products [L4MH3]+ have
played a seminal role in studies of stereochemical non-rigidity
and in the understanding of molecular hydrogen complexes.4
We report here the synthesis and characterisation of the novel
[L4MH2] complex [(triphos)Fe(CO)H2] 1 [triphos = MeC-
(CH2CH2PPh2)3], and a study by NMR spectroscopy of its
protonation via an intermediate proton–hydride hydrogen-
bonded complex to give the cationic dihydrogen–hydride
[L4MH3]+ product [(triphos)Fe(CO)(H)(Z2-H2)] 2.
With the first report by Hieber of [H2Fe(CO)4] in 1931,1
carbonyl hydrides are amongst the longest known of organo–
transition-metal complexes, and predate the birth of
organometallic chemistry as a discipline by at least two
decades. Carbonyl hydride phosphine complexes in particular
have represented an area of intense research, due in part to
their structural and spectroscopic diversity, but also on
account of the crucial role they play—or are suspected to
a Experiments Division, CELLS-ALBA, Ed. C3, Campus UAB,
Bellaterra, Spain
b Department of Chemistry, University of New Brunswick, 30 Dineen
Drive, Fredericton, N.B. 3B 6E2, Canada
c Department of Chemistry, University of Durham, South Road,
Durham, UK DH1 3LE
d Department of Chemistry, University College London, 20 Gordon
Street, London, UK WC1H 0AJ
w Dedicated to the memory of Dr D. A. ‘Fred’ Armitage (King’s
College London 1941–2008).
Results and discussion
z Electronic supplementary information (ESI) available: NMR simu-
lations for exchange processes A and B (Scheme 1) alone; Arrhenius
and Eyring plots for combined exchange involving A and B; DFT
calculated structures for [(triphos0)Fe(CO)H2] 10 and [(tri-
phos0)Fe(CO)H(Z2-H2)]+ 20. CCDC reference number 244046. For
ESI or crystallographic data in CIF or other electronic format see
DOI: 10.1039/b801652f
Synthesis and characterisation of [(triphos)Fe(CO)H2] 1
During a study of the hydride–borohydride complex [(tri-
phos)Fe(H)(Z2-BH4)] and its conversion to the remarkable
species [(triphos)Fe(m,Z4:Z4-BH6)Fe(triphos)]+,5 we observed
the formation of a small amount of a yellow by-product,
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1573–1581 | 1573

View Article Online
which was characterised as [(triphos)Fe(CO)H2] 1. Subsequent
deliberate synthesis of this air-stable complex proceeded by
two different routes. In the first, [Fe(H2O)6](BF4)2 was reacted
with triphos and an excess of NaBH4 in THF solution through
which a stream of CO was passed. The second method drew on
the procedure employed by Bakhmutov et al. in their synthesis
of the ruthenium analogue of 1.6 In this preparation, [(tri-
phos)FeH(Z2-BH4)] was reacted with a slight excess of KOtBu
in anhydrous THF, followed by passage of CO through the
solution. The IR spectrum of a solid film of 1 showed a strong
ligands. The Fe–H(1) distance is 1.57(3) A, and Fe–H(2) is
1.53(4) A. The molecular structure of 1 in the crystalline state
is slightly less distorted than its ruthenium analogue [(triphos)-
Ru(CO)H2] 3, reported by Bakhmutov et al.,6 although the
Fe–C–O moiety in 1 deviates slightly more (21) from linearity.
In contrast to its Ru analogue 3, [(triphos)Fe(CO)H2] 1
1
displays a non-rigid structure as revealed by H and 31P{1H}
variable-temperature (VT) NMR spectroscopic experiments
over the range 186–300 K. The pseudo-quartet observed in the
hydride region of the 1H NMR spectrum at room temperature
transforms into a second-order feature at 260 K. In addition,
inequivalent phosphorus nuclei are evident in the VT 31P{1H}
NMR spectrum, where the room temperature singlet at d =
57.1 ppm evolves at 280 K into a triplet at d = 61.3 ppm and a
doublet at d = 54.9 ppm (JP,P = 39.4 Hz). 1H NMR spectro-
scopy showed 1 to have a T1 min of 208.0 ms at 210 K, in the
regime expected for a classical transition-metal hydride.7
Fig. 2a shows VT 1H NMR spectra of 1 in CD2Cl2,
revealing the transformation of a quartet at 295 K (apparently
equal coupling of the three 31P nuclei to the two 1H nuclei) into
a second-order pattern on cooling to 200 K. Simulation of
these spectra (Fig. 2b) using the program gNMR reveals that
the two processes described in Scheme 1 are responsible for the
transformations observed. A combination of a Bailar twist
(turnstile rotation)8 of the triphos ligand with respect to the
FeH2(CO) moiety (process A), together with a pairwise ex-
change of the hydride ligands (process B), accurately matches
the observed spectra between 280 and 230 K.
band at 1880 cmꢁ1 (nCO), and a broad feature at 1663 cmꢁ1
,
assigned to both symmetric and antisymmetric nFe–H modes,
which could not be resolved. These features correspond to
DFT-calculated harmonic frequencies of 1930 and 1921 cmꢁ1
.
The calculated nCO mode occurs at 2020 cmꢁ1, exhibiting
strong coupling with the symmetric nFe–H feature. The scaling
factors of 0.93 and 0.86, respectively, for the nCO and nFe–H
modes reflects a greater degree of anharmonicity in the FeH2
moiety. The 1H NMR spectrum of 1 in CD2Cl2 at 298 K
revealed a well-resolved pseudo-quartet in the hydride region
at ꢁ10.5 ppm (JP,H = 21 Hz). The 31P{1H} NMR spectrum
displayed a broad peak at 57.1 ppm. The structure of 1
determined by single-crystal X-ray diffraction (Fig. 1) shows
a distorted octahedral environment around the Fe centre on
account of the fac-symmetry imposed by the triphos ligand.
The two hydride ligands are mutually cis, and the Fe–C–O
moiety deviates from linearity by 61. The three Fe–P distances
are very similar (FeꢁPav = 2.196 A), reflecting the compar-
able trans influences exerted by the hydride and carbonyl
It is clear from the simulations that pairwise hydride ex-
change (process B) is the lower energy pathway. Arrhenius and
Eyring plots (see ESIz) provide the following estimates for this
process: Ea = 24.3(4.2) kJ molꢁ1; DHz = 22.1(4.0) kJ molꢁ1
and DSz = ꢁ83.5(3.3) J Kꢁ1 molꢁ1. Presumably this exchange
is facilitated by the existence of a low-lying dihydrogen com-
plex of the form [(triphos)Fe(Z2-H2)(CO)], as is the case for a
wide range of polyhydride complexes.9 For example, ab initio
calculations have shown the [L4MH2] model systems [(CO)4-
FeH2] and [(PH3)4RuH2] to undergo polytopal rearrangement
Fig. 1 X-Ray crystal structure of [(triphos)Fe(CO)H2] 1. Selected
bond lengths (in A): Fe–H(1) 1.57(3), Fe–H(2) 1.53(4). Fe–P(1)
2.1925(12), Fe–P(2) 2.1998(12), Fe–P(3) 2.1954(11). Selected angles
Fig. 2 VT 1H NMR spectrum and simulation of [(triphos)Fe(CO)H2]
1 in CD2Cl2, showing the loss of fluxionality and second-order
behaviour below 260 K. (a) Observed spectrum; (b) simulated
spectrum.
(in
1):
H(1)–Fe–H(2)
88.7(18),
C(42)–Fe–H(1)
65.6(12),
C(42)–Fe–H(2) 80.6(14), Fe–C(42)–O(1) 174.0(4), P(3)–Fe–H(1)
171.0(12), P(2)–Fe–H(2) 175.6(14), P(3)–Fe–C(42) 106.69(15).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1574 | New J. Chem., 2008, 32, 1573–1581

View Article Online
1
Scheme 1 Intramolecular exchange processes (Bailar twist A and pairwise hydride exchange B) used to model the VT H NMR spectra of 1.
via a square-pyramidal dihydrogen complex in which the H2
ligand occupies an axial site.10 Although the ligand environ-
ment in 1 is more severely restrictive than in these simple model
systems, the mutually cis hydride ligands can easily attain a
similar transition-state geometry if the triphos ligand occupies
two equatorial and one axial site. It is noteworthy in this regard
that the more rigid complex [(tetraphos)FeH2] [tetraphos =
Characterisation of the dihydrogen complex
[(triphos)Fe(CO)H(g2-H2)]+
2
First-row transition metals have a greater proclivity to produce
dihydrogen complexes than do their second- and third-row
congeners owing to the less diffuse nature of the
3d orbitals compared to 4d and 5d.13 As [(triphos)Ru-
(CO)H(Z2-H2)]+ 4 was already known, we prepared the ana-
logous Fe complex [(triphos)Fe(CO)H(Z2-H2)]+ 2 by addition
at low temperature of an excess (16 eq.) of the medium strength
1
P(CH2CH2PPh2)3] shows fluxional behaviour in its H NMR
spectrum as low as 183 K.11 However, compounds of the
formula [(tetraphos)MH2]+ (M = Co, Rh, Ir) have been
shown to be classical dihydrides by T1 measurements, with
M–H relaxation process being the major relaxation component
rather than the H–H relaxation process.12 The second-row
congener of 1, viz. [(triphos)Ru(CO)H2] 3 is stereochemically
rigid in solution at ambient temperatures.6 The stronger M–H
bonds generally found in second-row metal complexes are
likely to stabilise the ground state dihydride in this system
with respect to the putative dihydrogen transition state re-
quired for site exchange.13
acid hexafluoroisopropanol (HFIP), (CF3)2CHOH, to
CD2Cl2 solution of [(triphos)Fe(CO)H2] 1 in an NMR tube.
The 1H NMR spectrum of the resulting complex at
a
We carried out DFT calculations on complex 1 to corrobo-
rate the experimental X-ray crystal structure (Fig. 1), and to
provide further insight into the fluxionality of the complex
(Scheme 1), as suggested by simulated NMR spectra (Fig. 2).
In our initial calculations the six C6H5 substituents bound to
the three P atoms of the triphos ligand were replaced with H
atoms (10). Subsequently, we were able to optimize the
geometry of the full complex [(triphos)Fe(CO)H2] (1).
The calculated structures of 1 and 10 show good agreement
with the experimentally determined structure (Fig. 3 and
Fig. S1z).
Our NMR spectral simulations revealed the fluxionality of 1
to be a combination of processes A and B in Scheme 1. From
the DFT-optimised geometry of 1, normal modes of vibration
that displace the geometry towards the reaction coordinate
corresponding to a Bailar twist (process A) appear at 404 and
474 (FeH2 deformation), and at 504 and 514 cmꢁ1 (FeL3
twist). Those corresponding to pairwise hydride exchange
(process B) appear at 594 (coupled FeH2/CO stretching mo-
tion), 726 (MH2 rocking) and 826 cmꢁ1 (MH2 symmetric
wag). The relatively soft vibrational potential characterising
each of these two processes thus indicates that they are both
sufficiently thermally accessible to allow the fluxional beha-
viour observed in solution at ambient temperatures.
Fig. 3 DFT calculated structure of [(triphos)Fe(CO)H2] (1) calcu-
lated at the B3LYP level of theory with a 6-31G(d,p) basis set. Selected
bond lengths (in A): Fe–H(1) 1.52, Fe–H(2) 1.52. Fe–P(1) 2.228,
Fe–P(2) 2.224, Fe–P(3) 2.217. Selected angles (in 1): H(1)–Fe–H(2)
84.5, C(42)–Fe–H(1) 79.5, C(42)–Fe–H(2) 78.8, Fe–C(42)–O(1) 172.9,
P(3)–Fe–H(1) 178.1, P(2)–Fe–H(2) 171.4, P(1)–Fe–C(42) 157.9.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1573–1581 | 1575

View Article Online
Scheme 2 Reaction pathway for the conversion of [(triphos)Fe(CO)H2] 1 into [(triphos)Fe(CO)H(Z2-H2)]+ 2.
190 K displayed a new broad hydride resonance at ꢁ8.6 ppm
(3H). The short T1 min value corresponding to this feature (20.6
ms at 210 K) argues persuasively in favour of a non-classical
dihydrogen moiety. The yellow, air-sensitive complex [(tri-
phos)Fe(CO)(Z2-H2)H]+ 2 is highly fluxional in solution, with
fast hydride–dihydrogen exchange occurring down to 180 K.
Complex 2 is stable below 250 K, but slowly eliminates H2
above this temperature, as evidenced by the growth of a feature
at d = 4.6 ppm in the 1H NMR spectrum assigned to free H2.
In the 31P{1H} NMR spectrum at 190 K, the transformation of
1 [Pa = 61.3 ppm (t), Pb = 54.9 ppm (d)] into 2 [Pa = 41.3 ppm
(t), Pb = 46.0 ppm (d)] is clearly observed, as shown in Fig. 4.
The phosphorus nucleus trans to CO is characterised by a
significant low-frequency shift (Dd = 20.1 ppm). The conver-
sion of 1 to 2 apparently occurs via the intermediate hydrogen-
bonded adduct [(triphos)(CO)(H)Fe–Hꢂ ꢂ ꢂH–OCH(CF3)2] 1a,
as depicted in Scheme 2. This conclusion is based on the
observance of a low frequency temperature-dependent shift
of the hydride resonance (ꢁ10.48 to ꢁ10.57 ppm at 250 K and
to ꢁ10.80 ppm at 190 K) upon addition of HFIP. The results
we observed are consistent with those of Shubina et al. in their
study of the analogous complex [(triphos)Ru(CO)H2] under
protonation.14
to 1) a signal was observed at d = ꢁ8.57 ppm, while a second
feature in the spectrum, at d = ꢁ8.46 ppm is assigned to
[(triphos)Fe(CO)D(Z2-HD)]+ 2-d2. Along with the character-
istic resonance of free HD at d = 4.60 ppm (JHD = 43.2 Hz),
these features attest to the lability of the dihydrogen ligand in
complex 2; a modest low-frequency shift of the resonance is
characteristic of a dihydrogen complex.17–19 Unfortunately
these resonances are broad and we were unable to determine
the value of JHD reliably. The T1
method proposed by
min
Hamilton and Crabtree affords a more precise estimate of the
H–H distance,20a and this approach was employed in a study
of the Ru analogue 4, the H–H distance in which was
accurately estimated using the T1 min slow-spinning model.6
Accordingly, we carried out a similar analysis of complex 2.
Insertion of our experimental values (T1 min for 1 = 208.06 ms
obs
1 min
and T for 2 = 20.61 ms) into eqn (1) (which accounts for
the averaging of T1 min by the exchange process) gives T
obs
1 min
(2Fe(H )) = 14.21 ms, and substitution of this value into eqn (2)
2
to obtain the true dipole–dipole relation time gives T1
(2Fe(H )) = 15.25 ms. The H–H distance of the dihydrogen
-
min
2
ligand in 2 can now be obtained from eqn (3). According to the
fast- (C* = 0.7937) and slow-rotation (C* = 1) models, the
empirical rHH values for complex 2 are 0.816 and 1.03 A,
respectively. The small relaxation due to phosphorus was
neglected in order to make the analysis exactly comparable
with the Ru analogue.20b As analogous dihydrogen complexes
containing different metal centres present the same spinning
behaviour,21 and as the fast-rotation model can be ruled out
for the Ru analogue, we are confident that slow rotation also
applies to 2, giving rHH (2) = 1.03 A. The NMR results thus
imply that [(triphos)Fe(CO)H(Z2-H2)] 2 contains a slightly
stretched dihydrogen ligand, in contrast to the Ru analogue 4
(rHH = 0.92 A by JHD and 0.94 A measured by T1 min), which
contains a normal dihydrogen moiety.
A simple method for calculating the H–H distance in a
dihydrogen complex involves partial deuteriation, and relies
on an inverse correlation between rHH and JHD for the (Z2-
HD) ligand, which was first proposed by Morris et al.15 and
improved by Heinekey and co-workers.16 For [(triphos)Fe
(CO)H(Z2-HD)]+ 2-d1 (produced by addition of CF3CO2D
obs
1/T
1 min
(2Fe,H ) = (1/3)/T1 min (1Fe,H
obs
)
2
3
+ (2/3)/T
(2Fe(H )
)
(1)
1 min
2
obs
1 min
1/T
(2Fe(H )) = 1/T1 min (1Fe,H ) + 1/T1 min (2Fe(H )) (2)
2 2 2
rHH [A] = 5.815 C* (T1 min/n)1/6
(3)
T1 min (1Fe,H ) = minimum spin–lattice relaxation time of
the hydride nuclei in complex 1.
obs
2
T
1 min
(2Fe,H ) = observed minimum spin–lattice relaxation
3
Fig. 4 31P{1H} NMR spectra in CD2Cl2 showing the transformation
of [(triphos)Fe(CO)H2] 1 (lower spectrum) into [(triphos)Fe(CO)(Z2-
H2)H]+ 2 (upper spectrum) via the hydrogen-bonded intermediate
[(triphos)(CO)(H)Fe–Hꢂ ꢂ ꢂH–OCH(CF3)2] 1a (middle spectrum) upon
addition of (HFIP).
time of the hydride resonance in complex 2.
obs
T
1 min
(2Fe(H )) = observed minimum spin–lattice relaxation
2
time of the dihydrogen ligand in complex 2, calculated from
eqn (1).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1576 | New J. Chem., 2008, 32, 1573–1581

View Article Online
T1 min (2Fe(H )) = minimum spin–lattice relaxation time of
2
the dihydrogen ligand in complex 2, calculated from eqn (2).
This conclusion runs counter to the expected change in
behaviour on going from a first-row transition-metal complex
to its second-row analogue. The strength of M-(Z2-H2) back-
bonding largely determines whether [LnMH2] or [LnM(Z2-H2)]
is formed.22 The general trend down a transition-metal triad is
an increase in M–H bond strength and a corresponding
stabilisation of the dihydride oxidative addition product as
the metal changes from 3d through 4d to 5d.23 For example,
Group 9 complexes of the type [(tetraphos)MH2]+ are re-
ported to be non-classical in the solid state and solution only
when M = Co; the Rh analogue is possibly non-classical in
solution whereas the Ir version is classical in both phases.24
However, Group 8 complexes of the form [L4MH(Z2-H2)]+
appear to buck this trend. In an extensive study of [trans-
(P2)2MH(Z2-H2)]+ complexes [P2 = Et2PCH2CH2PEt2 (depe)
or Ph2PCH2CH2PPh2 (dppe)], Bautista et al. concluded that
the Ru dihydrogen complex is less stable than its Fe counter-
part.25 In this series of complexes, nMH modes revealed the s-
bond strength of the M–H moiety to increase in the order Fe
o Ru o Os as expected, whereas the M–(Z2-H2) bond
strength was found to increase in the order Ru o Fe o Os;
this was accompanied by a corresponding decrease in the H–H
bonding interaction in the order Os o Fe o Ru. Hence, Fe
acts as a better-than-expected dp(M) - s*(Z2-H2) back-
donor.25 In a similar vein, experimental and theoretical studies
of the Group 8 complexes [(tetraphos)MH(Z2-H2)]+ by
Bianchini et al. have shown the Ru species to display the
weakest dp(M) - s*(Z2-H2) back-donation.26
Fig. 5 DFT calculated structure of [(triphos)Fe(CO)(Z2-H2)H]+ (2)
calculated at the B3LYP level of theory with a 6-31G(d,p) basis set.
Selected bond lengths (in A): Fe–H(1) 1.50, Fe–H(2) 1.57, Fe–H(3)
1.61, H(1)–H(2) 1.71, H(2)–H(3) 0.85, Fe–P(1) 2.323, Fe–P(2) 2.284,
Fe–P(3) 2.312. Selected angles (in 1): H(1)–Fe–H(2) 67.7,
H(1)–Fe–H(3) 98.5, H(2)–Fe–H(3) 30.8, C(42)–Fe–H(1) 83.0,
C(42)–Fe–H(2) 86.5, C(42)–Fe–H(3) 89.5, Fe–C(42)–O(1) 172.7,
P(3)–Fe–H(1) 175.7, P(2)–Fe–H(2) 150.1, P(2)–Fe–H(3) 174.8,
P(1)–Fe–C(42) 169.6.
Hence, our NMR studies agree with those of other workers,
in that the Fe complex 2 contains a more elongated dihydrogen
ligand than its Ru analogue 4. The implicitly stronger M–(Z2-
H2) interaction in the first-row dihydrogen complex may
possibly arise from an attractive ‘cis-effect’ between the hydride
and the neighbouring dihydrogen ligand.27 In the present case
the H–H distance may well also be influenced by interactions
H–H distance of 0.85 A being obtained for 2, identical to that
deduced for 20.
This case is apparently similar to that of
[OsCl2(H2)(NHQCPh2)(PiPr3)2], which was originally re-
ported by Barea et al.,28 and analysed in detail by Heinekey
and co-workers.29 Analogous NMR experiments to those
presented above yielded a value of 1.24 A for the H–H
distance, virtually identical to that we obtained for 2. Optimi-
sation of the complex at the B3LYP level of theory yielded an
H–H distance of 1.294 A, with only a single minimum.
However, the potential energy curve for the H–H stretching
motion (calculated at several levels of theory) showed that
varying the H–H distance between 1.00 and 1.60 A, requires
less than 4 kJ molꢁ1; in other words, the H–H separation in
this complex is virtually independent of energy. In this regime,
oxidative addition and reductive elimination of the H2 ligand
can occur with essentially no activation barrier. In a similar
vein, the complex [Cp*Ru(dppm)(H2)]+ exhibits an H–H
distance of 1.10 A as determined by neutron diffraction, which
agrees well with the value calculated from T1 measurements
and JHD measurements in solution.30 Optimisation of the
complex at the B3LYP level of theory correctly reproduced
the experimental geometry, with the exception of the H–H and
Ru–H distances.31 The H–H distance converged at 0.888 A (a
normal dihydrogen complex). This anomaly persisted when
calculations were performed with extended basis sets, and the
experimental geometry clearly did not conform to the ground
state geometry derived from the calculations. Analysis of the
ꢁ
with the OC(CF3)2 anion since 2 is in equilibrium with 1a
(Scheme 1).
We carried out DFT calculations on [(triphos)Fe(CO)(Z2-
H2)H]+ 2, to corroborate the structural parameters of the
dihydrogen moiety obtained from the T1 NMR measurements.
Initially, we employed the same approach as in our calcula-
tions on 1, with the six C6H5 substituents bound to the three P
atoms of the triphos ligand being replaced with H atoms (20).
This resulted in a calculated H–H distance of 0.85 A, implying
a normal dihydrogen ligand rather than the elongated ligand
of 1.03 A suggested by the T1 NMR measurements. It
appeared at this stage that the simplified triphos ligand,
calculated at the B3LYP level of theory with a 6-31G(d,p)
basis set, had produced an inaccurate result through signifi-
cant changes to the electronic nature of the ligand, and hence
to the Fe centre, even though we have used this approach
successfully for a closely related system.5 Accordingly, we
proceeded to a full structural optimisation of 2. The calculated
structures and selected structural parameters of 2 and 20 are
presented in Fig. 5 and as Fig. S2,z respectively. In the event,
the simplification of the triphos ligand appears to have had a
negligible effect on the overall equilibrium structure, with an
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1573–1581 | 1577

View Article Online
potential energy surface for the RuH2 moiety revealed that the
H2 ligand is highly delocalised in [Cp*Ru(dppm)(H2)]+: the
potential energy surface is highly anharmonic and the H–H
and Ru–H2 stretching modes are inseparable; there is extensive
coupling between both degrees of freedom. The flat, anhar-
monic potential for the vibrationally coupled RuH2 moiety in
this system thus permits significant thermal population of
higher vibrational states with large amplitudes of motion,
leading to an average H–H distance in solution that is
significantly longer than that derived for the electronic ground
state by DFT methods. We are led to conclude that similar
effects are at play in the case of [(triphos)Fe(CO)(Z2-H2)H]+
2, although an investigation of the potential energy surface of
the MH(Z2-H2) moiety in this system is beyond the scope of
our study. It is worth noting here that the discrepancy we
observe for 2 between the average H–H distance in solution
and the corresponding distance in the ground state may also
exist for some of the other dihydrogen complexes of Group 8
discussed above, challenging the experimental conclusion that
the M–(Z2-H2) bond strength is actually weaker for Ru than
for Fe.25,26
ephemeral proton–hydride interactions exist mainly in solu-
tion. There exist two NMR criteria for identifying
M–Hꢂ ꢂ ꢂH–X bonding. The first is a rather modest shift to
low frequency of the hydride signal, which is temperature- and
concentration-dependent. The second is a decrease in the
spin–lattice relaxation time (T1 min) of the coordinated hydride
ligand relative to the free hydride. Generally, T1 min decreases
by a factor of 1.5 to 3. Unfortunately, we were unable to
determine the Fe–Hꢂ ꢂ ꢂH–O distance in 1a, owing to a lack of
reproducibility in the T1 values we measured.
NMR spectroscopy can also be useful for extracting ther-
modynamic data relating to the formation of proton–hydride
complexes. Few attempts to determine thermodynamic para-
meters of proton transfer from an alcohol to a hydride have
been reported in the literature, presumably on account of the
complexity of the process.34 Proton–hydride bond enthalpies
can be obtained from the temperature-dependence of the
formation constants of the corresponding intermediate (van’t
Hoff method). We studied the proton-transfer leading to the
formation of [(triphos)Fe(CO)H(Z2-H2)]+ 2, with the results
shown in Fig. 6a. In this experiment, a two-fold excess of
HFIP was introduced into a CD2Cl2 solution of 1 at 190 K in a
5 mm NMR tube. This produced a low-frequency shift of the
hydride resonance, from d = ꢁ10.80 ppm to d = ꢁ10.57 ppm
at 250 K (lower curve). In the absence of the proton donor, the
chemical shift of the hydride resonance in 1 is practically
independent of temperature (upper curve). This shift (Dd =
ꢁ0.23 ppm) is reproducible and larger than the limits of
experimental error, and demonstrates clearly the existence of
an Fe–Hꢂ ꢂ ꢂHOCH(CF3)2 interaction. Hence, the resonance at
d = ꢁ10.80 ppm can be assigned to a hydrogen-bonded
adduct of the type shown as 1a.
Proton transfer to 1 via an Fe–Hꢂ ꢂ ꢂH–O intermediate
Formation of 2 from 1 occurs through proton transfer from
HFIP to an Fe–H moiety of 1, as described in Scheme 2. The
details of the reaction coordinate corresponding to this type of
process have been the subject of vigorous investigations for the
last decade.32 Although formation of (Z2-H2) species occurs
via M–Hꢂ ꢂ ꢂH–X intermediates, there is a dearth of charac-
terised examples of this kind, on account of the low tempera-
tures necessary to observe the labile M–Hꢂ ꢂ ꢂH–X intermediate
spectroscopically. The subtle chemical and physical changes in
this multi-step process make it very complicated to study, and
it is far from fully understood.33
Fig. 6b shows the strikingly similar results obtained from an
analogous study of the protonation of [(triphos)Ru(CO)H2] 3 by
Bakhmutov and co-workers,6 who calculated DH1f and DS1f for
a proton–hydride intermediate species 3a. However, no experi-
mental error was quoted for this study, nor was any possible
T1 measurements through NMR spectroscopy provide a
unique method for determination of M–Hꢂ ꢂ ꢂH–X distances in
such cases, because complexes containing these weak and
contribution considered from
a
1 : 2 hydrogen-bonded
Fig. 6 (a) Temperature-dependence of the hydride chemical shift in CD2Cl2 solution of [(triphos)Fe(CO)H2] 1 (upper curve); and of 1 in the
presence of a two-fold excess of HFIP (lower curve). (b) Temperature-dependence of the hydride chemical shift in CD2Cl2 solution of
[(triphos)Ru(CO)H2] 3 (upper curve); and of 3 in the presence of a two-fold excess of HFIP (lower curve). Data for 3 are taken from ref. 6.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1578 | New J. Chem., 2008, 32, 1573–1581

View Article Online
intermediate of the form [(triphos)M(CO)(Hꢂ ꢂ ꢂHOCH(CF3)2)2],
whose existence is impossible to discount in both the Fe and Ru
systems, and whose involvement will vitiate any quantitative
conclusions. Accordingly, we restrict our discussion to a quali-
tative comparison of the Fe and Ru intermediates 1a and 3a.
For [(triphos)Ru(CO)(H)Hꢂ ꢂ ꢂHOCH(CF3)2] 3a, DH1f was
calculated as ꢁ29.7 kJ molꢁ1, placing the adduct in the range
of medium strength H-bonds, as expected.32 The chemical
shift difference, Dd, between the free hydride complex [(tri-
phos)M(CO)H2] and the proton–hydride intermediate [(tri-
phos)M(CO)(H)Hꢂ ꢂ ꢂHOCH(CF3)2] is slightly greater for M
= Ru (0.42 ppm) than for M = Fe (0.32 ppm); possibly
implying a stronger M–Hꢂ ꢂ ꢂH–X interaction in 3a than in 1a,
and consistent with the higher basicity of 3 than of 1.
The DSo value reported for formation of 3a is
ꢁ79.5 J Kꢁ1 molꢁ1. In view of the higher fluxionality of the
Fe system 1 as evidenced by its variable-temperature NMR
spectra, we anticipate a larger loss of entropy in this system
upon protonation to form the more rigid species 1a. However,
differences in M–H bond strengths, zero-point energy contri-
butions, and the effects of solvation and electrostriction are
impossible to assess; these can easily tip the balance in terms of
entropy considerations.
rate constants derived from these simulations carry an esti-
mated uncertainty of 10%.
Preparation of 1
[(triphos)FeH2(CO)] was synthesised via two different meth-
ods.
Method A. A solution of NaBH4 (0.3 g, 8 mmol) in ethanol
(25 cm3) was added to a solution of [Fe(H2O)6][BF4]2 (0.337,
1 mmol) and triphos (0.625, 1 mmol) in THF (40 mL). The
resulting solution was heated under reflux for 20 min to give a
dark red solution. After filtration, a stream of CO (g) was
passed through the solution for 10 min. The resulting dark
orange solution was layered with hexane and stored at ꢁ40 1C,
producing a yellow powder after 1 d. The solution was reduced
in vacuo to 20% and the remaining solution decanted. The
precipitated solid was dried in vacuo and recrystallised from a
mixture of CH2Cl2–EtOH. Yield: 23% (0.16 g, 0.23 mmol).
Method B. A vigorously stirred solution of [(triphos)-
FeH(Z2-BH4)] (0.17, 0.24 mmol) in THF (7 mL) was treated
with solid KOtBu (0.08 g, 0.73 mmol), added in small portions.
After 1 d stirring, the pink solution became orange-brown.
This was filtered and a stream of CO (g) was passed through
for 5 min. EtOH (6 mL) saturated with CO (g) was layered on
top of the solution and left at ꢁ40 1C to produce a yellow
powder. Yield: 11% (0.08 g, 0.11 mmol). Anal. calc. for
C42H41FeOP3: C, 67.09; H, 5.49. Found: C, 68.34; H,
5.26%. IR (solid, cmꢁ1): nCO 1880 (br), nFeH 1663 (s). 1H
NMR (CD2Cl2, 295 K): d 7.40–6.90 (m, 30H, phenyl H), 2.17
(s, 6H, CH3(CH2PPh2)3), 1.53 (s, 3H, CH3(CH2PPh2)3),
ꢁ10.51 (q, 2H, FeH2, 2JHP = 21 Hz). 31P{1H} NMR (CD2Cl2,
295 K): d 57.12 (s). 1H NMR (CD2Cl2, 200 K): d 7.40–6.90 (m,
30H, phenyl H), 2.08 (m, 6H, CH3(CH2PPh2)3), 1.43 (s, 3H,
Experimental
General
All operations were performed under an inert atmosphere with
the use of Schlenk techniques and a glove box (MBraun,
Labmaster 130), and using freshly distilled solvents. Routine
1H and 31P NMR spectra were recorded using a Bruker AV360
spectrometer operating at 360 MHz, and 145 MHz, respec-
tively, or a Bruker AV400 operating at 400 and 162. FT-IR
spectra were recorded on a Perkin Elmer Spectrum One
spectrometer with a beam-condensing accessory (BCA), using
a diamond compression cell. Microanalyses were performed at
the University of North London. T1 measurements and VT
NMR experiments were carried out in standard 5 mm NMR
tubes in CD2Cl2 and the data were recorded using a Bruker
DRX500 spectrometer operating at 500 or 202.5 MHz. If
HFIP was added to the sample, this was carried out at ꢁ78
1C by cooling the NMR tube in a dry ice–acetone bath, and
the acid was transferred via a syringe. The mixture was
degassed by passage of argon, and the Teflon valve was closed
immediately. The NMR tube was then rapidly inserted into
the probe of the spectrometer which had been pre-cooled to
190 K. All samples were allowed at least 10 min to equilibrate
at each temperature. The conventional inversion–recovery
method (180–t–90) was used to determine the T1 relaxation
time as a function of temperature. Relaxation times were
calculated using the non-linear three-parameter fitting routine
of the spectrometers. The pulses were controlled at each
temperature. In each experiment, the waiting period was much
longer than the expected relaxation time, and 16 variable
delays were employed. VT studies were performed in the range
190–300 K. Spectral simulations were carried out using gNMR
(Cherwell Scientific Publishing Ltd, Oxford, version 4.1); the
CH3(CH2PPh2)3), ꢁ10.46 (2nd order, 2H, FeH2, Tchange
=
260 K). 31P{1H} NMR (CD2Cl2, 200 K): d 61.31 (t, 1P, Pa,
2JP P = 39.4), 54.9 (d, 2P, Pb). T1 min = 208.06 ms (210 K).
a
b
Preparation of 2
A 5 mm NMR tube was charged in the glove-box with 1
(3.7 mg, 5.2 mmol) and CD2Cl2 (0.7 mL) was added. The NMR
tube was cooled to ꢁ78 1C, an excess of [(CF3)2CHOH]
(10.4 mL, 8.3 mmol) was added via a syringe and the tube
immediately sealed. The tube was rapidly inserted into the
1
spectrometer probe maintained at 180 K, and the H NMR
spectrum immediately recorded at this temperature, revealing
quantitative transformation of 1 into 2. 1H NMR (CD2Cl2,
200 K): d 7.80–6.60 (m, 30H, phenyl H), 2.50–2.30 (br, 6H,
CH3(CH2PPh2)3), 1.84 (br, 3H, CH3(CH2PPh2)3), ꢁ8.57 (br,
3H, FeH). 31P{1H} NMR (CD2Cl2, 220 K): d 45.96 (d, 2P, Pb,
2JP P = 53.7 Hz), 41.25 (t, 1P, Pa). T1 min = 20.61 ms (210 K).
a
b
Identification of the proton–hydride bonded adduct 1a
Under the experimental conditions described above, two
equivalents of HFIP (1.3 mL, 1.0 ꢃ 10ꢁ2 mmol) were added
via syringe to a solution of 1 prepared as described above.
Low-temperature 1H NMR spectra revealed the complete
conversion of 1 into the proton–hydride bonded adduct 1a.
Addition of a subsequent 9.1 mL of HFIP resulted, as
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1573–1581 | 1579

View Article Online
1
described, in the formation of 2. H NMR (CD2Cl2, 195 K):
a thorough understanding of both 1 and 2, and of the proton-
transfer process which links them. The NMR experiments
imply that complex 2 contains a stretched dihydrogen ligand
(rHH = 1.23 A), with an H–H distance about 0.3 A longer than
in its Ru analogue 4. This elongation apparently arises from a
greater degree of M-(Z2-H2) back-donation in 2 compared to
in 4, reversing the expected trend in H–H distances for
dihydrogen complexes, which normally increases down a tran-
sition-metal triad. However, DFT calculations on 2 indicated a
short H–H distance of 0.85 A, characteristic of a normal
dihydrogen ligand. A similar discrepancy between experimen-
tal and theoretical H–H distances has been noted for several
other dihydrogen complexes of Group 8 metals. The proton-
transfer process from HFIP to 1 proceeds through the
proton–hydride hydrogen-bonded intermediate [(triphos)Fe-
(CO)(H)Hꢂ ꢂ ꢂHOCH(CF3)2] 1a. A qualitative study of the
M–Hꢂ ꢂ ꢂH–X bond in 1a reveals a slightly lower basicity
for the Fe complex 1 than for [(triphos)Ru(CO)H2] 3.
d 7.80–6.60 (m, 30H, phenyl H), 2.50–2.30 (br, 6H, CH3-
(CH2PPh2)3), 1.82 (br, 3H, CH3(CH2PPh2)3), ꢁ10.80 (br, 3H,
FeH). 31P{1H} NMR (CD2Cl2, 220 K): d 60.44 (t, 1P, Pa,
2JP P = 41.2 Hz), 54.7 (d, 2P, Pb). T1 measurements were not
a
b
reproducible to the desired accuracy.
X-Ray crystallography studies
Suitable crystals of 1 were mounted on a thin glass fibre using
silicon grease and cooled on the diffractometer to 120 K using
Oxford Cryostream Liquid N2 device. Approximate unit cell
dimensions were determined by the Nonius Collect program
using a Nonius KappaCCD diffractometer (graphite monochro-
matic Mo-Ka radiation, l = 0.71073 A), with a detector-to-
crystal distance of 30 mm. Crystals were indexed using the
DENZO-SMN package, and positional data were refined
along with diffractometer constants to give the final unit cell
parameters. Integration and scaling (DENZO-SMN, Scale-
pack) resulted in unique data sets corrected for Lorentz and
polarisation effects and for the effects of crystal decay and
absorption, by a combination of averaging of equivalent
reflections and an overall volume and scaling correction.
Structures were solved using SHELXS-97 and developed via
alternating least-squares cycles and difference Fourier synth-
esis (SHELXL-97). All non-hydrogen atoms were modelled
anisotropically.
Acknowledgements
We thank King’s College London for a studentship (to G.G.)
and NSERC of Canada for financial support of this work. We
also thank ACEnet for access to computational facilities.
References
Crystal data for 1. C42H41FeOP3, M = 710.51, yellow
block, 0.30 ꢃ 0.25 ꢃ 0.25 mm, monoclinic, space group
P21/n (No. 14), a = 10.1997(3), b = 18.1058(6), c =
18.9884(7) A, b = 96.489(2)1, V = 3484.2(2) A3, Z = 4, Dc
= 1.354 g cmꢁ3, F000 = 1488, Nonius KappaCCD, Mo-Ka
radiation, l = 0.71073 A, T = 120(2) K, 2ymax = 50.01, 17650
reflections collected, 5907 unique (Rint = 0.1198). Final GooF
= 1.038, R1 = 0.0588, wR2 = 0.0869, R indices based on
3612 reflections with I 4 2s(I) (refinement on F2), 433
parameters, 0 restraints. Lp and absorption corrections ap-
plied, m = 0.604 mmꢁ1. The hydride atoms were located on
the difference Fourier map, the thermal parameter of H(1) was
fixed at 0.04, and both hydrides were refined freely.
1 W. Hieber, Ber. Dtsch. Chem. Ges., 1931, 64, 2342.
2 M. Torrent, M. Sola and G. Frenking, Chem. Rev., 2000, 100, 439
and references therein.
3 M. A. Esteruelas and L. A. Oro, Chem. Rev., 1998, 98, 577.
4 D. G. Gusev, R. Hubener, P. Burger, O. Orama and H. Berke,
J. Am. Chem. Soc., 1997, 119, 3716 and references therein.
5 G. Guilera, G. S. McGrady, J. W. Steed and N. Kaltsoyannis, New
J. Chem., 2004, 28, 444.
6 V. I. Bakhmutov, E. V. Bakhmutova, N. V. Belkova, C. Bianchini,
L. M. Epstein, D. Masi, M. Peruzzini, E. S. Shubina, E. V.
Vorontsov and F. Zanobini, Can. J. Chem., 2001, 79, 479.
7 D. M. Heinekey and W. J. Oldham, Jr, Chem. Rev., 1993, 93, 913.
8 See, for example, Inorganic Chemistry. Principles of Structure and
Reactivity, ed. J. E. Huheey, E. A. Keiter and R. L. Keiter, Harper
Collins, New York, 4th edn, 1993.
9 F. Maseras, A. Lledo
2000, 100, 601.
10 C. Soubra, Y. Oishi, T. A. Albright and H. Fujimoto, Inorg.
Chem., 2001, 40, 620.
´
s, E. Clot and O. Eisenstein, Chem. Rev.,
DFT calculations on 1, 10, 2 and 20
Gas phase structures and vibrational frequencies were calcu-
lated for each complex at the B3LYP level of theory with a
6-31G(d,p) basis set. Calculations were performed using
Gaussian 03 on the Placentia2 cluster housed at the Memorial
University of Newfoundland. Placentia2 is part of the ACEnet
(Atlantic Computational Excellence Network) group of clus-
ters in Maritime Canada.
11 C. Bianchini, F. Laschi, M. Peruzzini, F. M. Ottaviani, A. Vacca
and P. Zanello, Inorg. Chem., 1990, 29, 3394.
12 D. M. Heinekey and M. van Roon, J. Am. Chem. Soc., 1996, 118,
12134.
13 M. T. Haward, M. W. George, P. Hamley and M. Poliakoff,
J. Chem. Soc., Chem. Commun., 1991, 1101.
14 E. S. Shubina, N. V. Belkova, A. N. Krylov, E. V. Vorontsov, L.
M. Epstein, D. G. Gusev, M. Niedermann and H. Berke, J. Am.
Chem. Soc., 1996, 118, 1105.
15 P. A. Maltby, M. Schlaf, M. Steinbeck, A. J. Lough, R. H. Morris,
W. T. Klooster, T. F. Koetzle and R. C. Srivastava, J. Am. Chem.
Soc., 1996, 118, 5396.
Conclusions
16 K. L. James, H. Mellows and D. M. Heinekey, J. Am. Chem. Soc.,
2001, 129, 2085.
17 X. L. Luo and R. H. Crabtree, J. Am. Chem. Soc., 1990, 112, 4813.
18 N. Bampos and L. D. Field, Inorg. Chem., 1990, 29, 587.
19 W. J. Oldham, Jr, A. S. Hinkle and D. M. Heinekey, J. Am. Chem.
Soc., 1997, 119, 11028.
20 (a) D. G. Hamilton and R. H. Crabtree, J. Am. Chem. Soc., 1988,
110, 4126; (b) P. J. Desrosiers, L. H. Cai, Z. R. Lin, R. Richards
and J. Halpern, J. Am. Chem. Soc., 1991, 113, 4173.
The novel complex [(triphos)Fe(CO)H2] 1 has been synthesised
and characterised by 1H and 31P NMR and IR spectroscopies,
and its structure has been determined by X-ray diffraction.
Hence, it is shown to contain classical hydride ligands. Proto-
nation of 1 by the medium strength acid HFIP gives the novel
dihydrogen complex [(triphos)Fe(CO)H(Z2-H2)]+ 2. Conven-
tional NMR experiments and T1 measurements have permitted
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008
1580 | New J. Chem., 2008, 32, 1573–1581

View Article Online
21 M. T. Bautista, K. A. Earl, P. A. Maltby, R. H. Morris, C. T.
Schweitzer and A. Sella, J. Am. Chem. Soc., 1988, 110
7031.
22 R. H. Crabtree, Angew. Chem., Int. Ed. Engl., 1993, 32, 789.
23 See, for example, Ch. Elschenbroich and A. Salzer, Organo-
metallics, VCH, Weinheim, 2nd edn, 1991.
24 C. Bianchini, M. Peruzzini and F. Zanobini, J. Organomet. Chem.,
1998, 354, C19.
25 M. T. Bautista, E. P. Cappellani, S. D. Drouin, R. H. Morris, C. T.
Schweitzer, A. Sella and J. Zubkowski, J. Am. Chem. Soc., 1991,
113, 4876 and references therein.
27 L. S. van der Sluys, J. Eckert, O. Eisenstein, J. H. Hall, J. C.
Huffmann, S. A. Jackson, T. F. Koetzle, G. J. Kubas, P. J.
Vergamini and K. G. Caulton, J. Am. Chem. Soc., 1990, 112, 4831.
28 G. Barea, M. A. Esteruelas, A. Lledos, A. M. Lopez and J. I.
Tolosa, Inorg. Chem., 1998, 37, 5033.
29 D. M. Heinekey, A. Lledos and J. M. Lluch, Chem. Soc. Rev.,
2004, 33, 175.
30 M. S. Cinn and D. M. Heinekey, J. Am. Chem. Soc., 1987, 109, 5865.
31 R. Gelabert, M. Moreno, J. M. Lluch and A. Lledos, J. Am. Chem.
Soc., 1997, 119, 9840.
32 R. H. Crabtree, P. E. M. Siegbahn, O. Eisenstein, A. L. Rheingold
and T. F. Koetzle, Acc. Chem. Res., 1996, 29, 348.
33 L. M. Epstein and E. S. Shubina, Coord. Chem. Rev., 2002, 231, 165.
34 J. A. Ayllon, C. Gervaux, S. Sabo-Etienne and B. Chaudret,
Organometallics, 1997, 16, 2000.
26 (a) C. Bianchini, K. Linn, D. Masi, M. Peruzzini, A. Polo, A.
Vacca and F. Zanobini, Inorg. Chem., 1993, 32, 2366; (b) C.
Bianchini, D. Masi, M. Peruzzini, M. Casarin, C. Maccato and
G. A. Rizzi, Inorg. Chem., 1997, 36, 1061.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2008 New J. Chem., 2008, 32, 1573–1581 | 1581