Fluxionality and Isomerism of the Bis(dihydrogen) Complex RuH2(H2)2(PCy3)2: INS, NMR, and Theoretical Studies

Article abstract of DOI:10.1021/ic970697y

To study the fluxionality of the bis(dihydrogen) complex RuH₂(H₂)₂(PCy₃)₂ (1), NMR spectra were recorded in Freons (mixture of CDCl₃, CDFCl₂, and CDF₂Cl). 1 was found to remain fluxional at all temperatures, but the presence of CDCl₃ necessary for its solubilization induces its transformation into, first, RuHCl(H₂)₂(PCy₃)₂ (3) and the new ruthenium(IV) dihydride RuH₂Cl₂(PCy₃)₂ (4). 4 is produced selectively in pure CDCl₃ but reacts further to give a mixture of chloro complexes. 4 was isolated from the reaction of 1 with aqueous HCl in Et₂O and shows a fluxional process attributed to the interconversion between two symmetrical isomers. The activation parameters of this process were obtained by ₁H NMR line shape analysis, as well as those corresponding to the exchange between 3 and free dihydrogen. The fluxionality of the dihydrogen - hydride system is also evident at a much faster time scale than that of NMR studies in the inelastic neutron scattering observations of the rotation of the dihydrogen ligands. The geometries and relative energies of several isomers of complexes 1, 3, and 4 were studied using density functional theory (DFT) and MP2 methods, together with a few coupled-cluster (CCSD-(T)) calculations. In contrast to what might have been expected, the two hydrides and the two H₂ units of 1 lie in the same plane, due to the attractive cis effect created by the hydrides. The two H₂ ligands adopt cis positions in the lowest-energy isomer. Rotation of the two dihydrogen ligands has been analyzed using DFT calculations. A slight preference for a C₂ conrotatory pathway has been found with a calculated barrier in good agreement with the experimental INS value. Two low-energy isomers of 4 have been characterized computationally, both of which have C_2u symmetry, consistent with the solution NMR spectra.

Full text of DOI:10.1021/ic970697y

Inorg. Chem. 1998, 37, 3475-3485
3475
Fluxionality and Isomerism of the Bis(dihydrogen) Complex RuH2(H2)2(PCy3)2: INS, NMR,
and Theoretical Studies
Venancio Rodriguez, Sylviane Sabo-Etienne, and Bruno Chaudret*
Laboratoire de Chimie de Coordination du CNRS, 205 Route de Narbonne,
31077 Toulouse Cedex 4, France
John Thoburn, Stefan Ulrich, and Hans-Heinrich Limbach
Institut fu¨r Organische Chemie, Fachbereich Chemie, Freie Universita¨t Berlin,
Takustrasse 3, D-14195 Berlin 33, Germany
Juergen Eckert
Manuel Lujan, Jr., Neutron Scattering Center (LANSCE), Los Alamos National Laboratory,
Los Alamos, New Mexico 87545
Jean-Claude Barthelat, Khansaa Hussein,† and Colin J. Marsden
Laboratoire de Physique Quantique, UMR IRSAMC-CNRS UMR 5626, Universite´ Paul Sabatier,
118 Route de Narbonne, 31062 Toulouse Cedex 4, France
ReceiVed June 5, 1997
To study the fluxionality of the bis(dihydrogen) complex RuH2(H2)2(PCy3)2 (1), NMR spectra were recorded in
Freons (mixture of CDCl3, CDFCl2, and CDF2Cl). 1 was found to remain fluxional at all temperatures, but the
presence of CDCl3 necessary for its solubilization induces its transformation into, first, RuHCl(H2)2(PCy3)2 (3)
and the new ruthenium(IV) dihydride RuH2Cl2(PCy3)2 (4). 4 is produced selectively in pure CDCl3 but reacts
further to give a mixture of chloro complexes. 4 was isolated from the reaction of 1 with aqueous HCl in Et2O
and shows a fluxional process attributed to the interconversion between two symmetrical isomers. The activation
1
parameters of this process were obtained by H NMR line shape analysis, as well as those corresponding to the
exchange between 3 and free dihydrogen. The fluxionality of the dihydrogen-hydride system is also evident at
a much faster time scale than that of NMR studies in the inelastic neutron scattering observations of the rotation
of the dihydrogen ligands. The geometries and relative energies of several isomers of complexes 1, 3, and 4
were studied using density functional theory (DFT) and MP2 methods, together with a few coupled-cluster (CCSD-
(T)) calculations. In contrast to what might have been expected, the two hydrides and the two H2 units of 1 lie
in the same plane, due to the attractive “cis effect” created by the hydrides. The two H2 ligands adopt cis positions
in the lowest-energy isomer. Rotation of the two dihydrogen ligands has been analyzed using DFT calculations.
A slight preference for a C2 conrotatory pathway has been found with a calculated barrier in good agreement
with the experimental INS value. Two low-energy isomers of 4 have been characterized computationally, both
of which have C2V symmetry, consistent with the solution NMR spectra.
1. Introduction
chemistry. The complex RuH2(H2)2(PCy3)2 (1)6 has long been
the only reported thermally stable bis(dihydrogen) complex, until
the recent characterization of LRuH(H2)2 (L ) HB(3,5-Me2-
pz), HB(3-iPr-4-Br-pz)7 and of [OsH3(H2)2(PiPr3)2]+,8 respec-
tively, by our group and Caulton and Tilset. Other bis-
(dihydrogen) complexes have either been prepared in a matrix9
or observed in solution.10
The discovery by Kubas of dihydrogen coordination without
dissociation has opened a whole new field in inorganic
chemistry1-5 and stimulated a renewed interest in polyhydride
† Permanent address: Department of Chemistry, Faculty of Sciences,
University Al-Baath, Homs, Syria.
(1) Kubas, G. J. Acc. Chem. Res. 1988, 21, 120.
(2) (a) Crabtree, R. H.; Hamilton, D. G. AdV. Organomet. Chem. 1988,
28, 295. (b) Crabtree, R. H. Acc. Chem. Res. 1990, 23, 95. (c) Crabtree,
R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789.
(3) Jessop, P. G.; Morris, R. H. Coord. Chem. ReV. 1992, 121, 155.
(4) Heinekey, D. M.; Oldham, W. J., Jr., Chem. ReV. 1993, 93, 913.
(5) Burdett, J. K.; Eisenstein, O.; Jackson, S. A. In Transition Metal
Hydrides: Recent AdVances in Theory and Experiment; Dedieu, A.,
Ed.; VCH: New York, 1991; p 149.
A striking feature of most hydrido dihydrogen complexes is
(6) (a) Chaudret, B.; Poilblanc, R. Organometallics 1985, 4, 1722. (b)
Arliguie, T.; Chaudret, B.; Morris, R. H.; Sella, A. Inorg. Chem. 1988,
27, 598.
(7) Moreno, B.; Sabo-Etienne, S.; Chaudret, B.; Rodriguez, A.; Jalon, F.;
Trofimenko, S. J. Am. Chem. Soc. 1995, 117, 7441.
(8) Smith, K.-T.; Tilset, M.; Kuhlman, R.; Caulton, K. G. J. Am. Chem.
Soc. 1995, 117, 9473.
S0020-1669(97)00697-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/19/1998

3476 Inorganic Chemistry, Vol. 37, No. 14, 1998
Rodriguez et al.
Chart 1
their fluxionality leading to the observation of a single peak in
1H NMR at all usual temperatures whenever a hydride is located
cis to a dihydrogen ligand.2-4 This is in particular the case for
1. Only a few static spectra have been reported for such
complexes.11 However, Freon solvents can give access to a
much larger temperature range down to ca. 130 K. Since infrared
data were clearly in agreement with a cis-dihydride structure,
we have investigated the NMR properties of 1 at very low
temperature.
We have studied for several years the reactivity of 1, and for
example, we have previously reported the reactions of 1 with
halocarbons yielding novel 16-electron dihydrogen derivatives.12
However, during the course of this work, we found that 1 reacts
with chloroform to yield a new unstable ruthenium(IV) dihy-
dride. A very similar complex, namely RuH2Cl2(PiPr3)2, was
recently reported by Werner and co-workers.13
Figure 1. Inelastic neutron scattering spectrum of RuH2(H2)2(PCy3)2
at T ) 5 K collected on the IN5 spectrometer at ILL with an incident
wavelength of 6 Å.
In addition, one of us has developed the use of INS for the
determination of the rotation barrier of dihydrogen ligands.14
This rotation barrier can provide information on the mode of
coordination of dihydrogen and possible interactions with nearby
hydrides.
We describe in this paper an INS study in the solid state, an
NMR study of 1 in Freons, and a study of the reactivity of 1
with chloroform leading to a new ruthenium(IV) dihydride
complex 4. Theoretical studies of several possible isomers of
1 are reported, and the concerted rotation of the two dihydrogen
ligands is analyzed. Isomers of 4 are also studied computa-
tionally.
Figure 2. Temperature dependence of the INS spectrum of RuH2(H2)2-
(PCy3)2: 5 K (+); 50 K (*); 75 K (4); 100 K (]).
2. Inelastic Neutron Scattering Studies on RuH2(H2)2-
(PCy3)2
At temperatures below 50 K, the INS spectrum of 1, on the
Data collected on RuH(H2)I(PCy3)2 at various incident
neutron wavelengths failed to reveal any low-frequency rota-
tional tunneling transitions for the dihydrogen ligand. On the
basis of the known energy resolution of the spectrometer for
the longest incident wavelength used, we can place a lower limit
on the barrier to rotation for H2 in this compound of ap-
proximately 2.5 kcal/mol, given that the d(HH) is 1.03 Å from
X-ray diffraction.12a
other hand, consists of the usual pair of bands from the rotational
transitions within the librational ground state at (4.7 cm-1
.
These can be interpreted in terms of planar rotation in a double-
minimum potential well14a to yield a barrier to rotation of 1.1
kcal/mol. The structure which is evident in these bands (Figure
1) is noteworthy when compared with rotational tunneling lines
of other dihydrogen complexes (Figure 4 in ref 14a). It may
indicate that there are at least two inequivalent dihydrogen
ligands in the solid or be the result of interactions between the
two dihydrogen ligands. In view of fluxionality and the
computational results described below, we are inclined to
attribute this observation to the latter possibility.
(9) (a) Sweany, R. L. J. Am. Chem. Soc. 1985, 107, 2374. (b) Upmacis,
R. K.; Poliakoff, M.; Turner, J. J. J. Am. Chem. Soc. 1986, 108, 3645.
(10) (a) Crabtree, R. H.; Lavin, M. J. Chem. Soc., Chem. Commun. 1985,
1661. (b) Michos, D.; Luo, X.-L.; Faller, J. W.; Crabtree, R. H. Inorg.
Chem. 1993, 32, 2, 1370. (c) Fontaine, X. L R.; Fowles, E. H.; Shaw,
B. L. J. Chem. Soc., Chem. Commun. 1988, 482.
(11) (a) Crabtree, R. H.; Lavin, M.; Bonneviot, L. J. Am. Chem. Soc. 1986,
108, 4032. (b) Bianchini, C.; Perez, P. J.; Peruzzini, M.; Zanobini, F.;
Vacca, A. Inorg. Chem. 1991, 30, 279.
(12) (a) Chaudret, B.; Chung, G.; Eisenstein, O.; Jackson, S. A.; Lahoz,
F.; Lopez, J. A. J. Am. Chem. Soc. 1991, 113, 2314. (b) Christ, M.
L.; Sabo-Etienne, S.; Chaudret, B. Organometallics 1994, 13, 3800.
(c) Sabo-Etienne, S.; Chaudret, B. Coord. Chem. ReV. In press.
(13) Gru¨nwald, C.; Gevert, O.; Wolf, J.; Gonzalez-Herrero, P.; Werner,
H. Organometallics 1996, 15, 1960.
(14) (a) Eckert, J.; Kubas, G. J. J. Phys. Chem. 1993, 97, 2378. (b) Limbach,
H.-H.; Ulrich, S.; Buntkowsky, G.; Sabo-Etienne, S.; Chaudret, B.;
Kubas, G. J.; Eckert, J. J. Am. Chem. Soc. In press.
As the temperature is increased (Figure 2), the tunneling peaks
shift to lower energy and broaden in the usual manner where
the width shows an Arrhenius type temperature dependence with
an apparent activation energy of about 0.4 kcal/mol. In addition,
a broad quasielastic feature appears below the inelastic peaks
at temperatures above 75 K. The rotational tunneling peaks
coalesce into a broad quasielastic line below the elastic peak
above 150 K (Figure 3). The width of this line continues to
increase with temperature up to 250 K, the highest temperature
reached in this experiment. The activation energy derived from
the temperature dependence of the width of this quasielastic

Fluxionality and Isomerism of RuH2(H2)2(PCy3)2
Inorganic Chemistry, Vol. 37, No. 14, 1998 3477
Figure 3. Quasielastic scattering from RuH2(H2)2(PCy3)2 at T ) 150
K (s), 200 K (- ‚ -), and 250 K (- - -).
line is about 0.25 kcal/mol, which suggests that the motion
giving rise to this line may be described15 as weakly hindered
rotational diffusion of dihydrogen. In addition, a second, more
narrow quasielastic component is evident at 250 K with a fwhm
1
of approximately 4 cm-1
.
Figure 4. 500 MHz H NMR spectra of 1 dissolved in a mixture of
deuterated Freons (CDCl3:CDFCl2:CDF2Cl ) 1:5:5) as a function of
temperature. For assignment see text.
The increase in width of the rotational tunneling transition
lines as a function of temperature is the result of inhomogeneous
broadening from coupling to the phonon bath and of the
incoherent dihydrogen exchange processes as we recently
described.14b The value of the activation energy for the
inhomogeneous broadening of the tunneling line (0.4 kcal/mol,
or 140 cm-1) is expected16 to be similar to the librational
transition (torsion) which is responsible for the coupling to the
phonon bath in this model. The torsion can be calculated to be
180 cm-1 for the present case with a simple 2-fold rotational
potential. This is in fair agreement with the prediction of this
model16 whereby the remaining discrepancy may indicate the
neglect of other processes such as the incoherent dihydrogen-
hydride exchange.
peaks may be attributed to dihydrogen affected by the rapid
intramolecular exchange of hydride and dihydrogen ligands.
3. NMR Studies of 1 in a Freon Mixture
To gain additional information on the reactivity of 1 with
halogenated hydrocarbons and to try to freeze out the intra-
molecular hydrogen exchange in 1, we decided to prepare cold
solutions of 1 in a mixture of deuterated Freons. 1 was not
soluble in Freons in the absence of CDCl3. The reactions were
therefore carried out in the solvent mixture (CDCl3:CDFCl2:
CDF2Cl ) 1:5:5) described in the Experimental Section. This
mixture allows liquid-state NMR measurements down to 140
K.17 Figure 4 shows the 500 MHz NMR spectra of a sample
measured after storage at 240 K for several days. At 263 K,
the highest accessible temperature in this medium, three main
signals a/b, f, and e are observed in the high-field spectrum
characterized by different spectral changes upon lowering the
temperature. The triplet e appearing at -8.9 ppm corresponds
to the starting complex 1. The signal broadens near 143 K but
does not split into separate signals for dihydrogen and hydride
sites, which demonstrates the fast intramolecular hydride-
dihydrogen exchange within this species. The line broadening
at low temperatures could arise either from the circumstance
that this exchange becomes slower or from a short T2 arising
from the increase in viscosity of the solution or rapid dipole-
dipole relaxation. Signal f appears at -15.1 ppm and consists
The observed non-Arrhenius behavior (i.e., the two different
activation energies for different temperature regimes reported
above) is also consistent with the superposition of inhomoge-
neous broadening from coupling to the phonon bath and of the
incoherent dihydrogen exchange processes. Moreover, it is
interesting to note that, at the temperature at which the rotational
tunneling lines coalesce into a quasielastic line, i.e around 150
K, the NMR data demonstrate the existence of rapid intramo-
lecular hydrogen exchange even though the activation energy
for the latter process is much higher than those observed by
INS for the rotational motions. This may suggest that the
rotational motions of dihydrogen observed by INS are apparently
affected by the presence of the dihydrogen-hydride exchange.
One may then assume that, in the course of this interconversion
of hydride and dihydrogen ligands, the latter can occur in a
variety of environments in addition to that of the octahedral
coordination sites. These considerations would account for the
INS observations at 75 and 100 K, i.e., that the relatively sharp
rotational tunneling bands arise from rotational transitions of
dihydrogen located in their well-defined potential wells while
the broad quasielastic component below the rotational tunneling
of a triplet with Jav ) 11 Hz. It broadens slightly when the
PH
temperature is lowered, and its intensity increases with the
sample storage time. It was assigned to a species (2′) similar
to RuHCl(H2)(PCy3)2 (2), a complex previously obtained from
the reaction of 1 with CH2Cl2, but not to 2 itself, which resonates
at -16.8 ppm and exchanges in the presence of excess
dihydrogen with RuHCl(H2)2(PCy3)2 (3), also present as dis-
cussed below.12b Peak g, appearing at 5.3 ppm, was assigned
to CDHCl2 formed during the reaction together with 2′ and 3.
(15) Be´e, M. Quasielastic Neutron Scattering; Adam Hilger: Bristol, U.K.,
1988.
(16) Hewson, A. C. J. Phys. C: Solid State Phys. 1982, 15, 3855.
(17) Siegel, J. S.; Anet, F. A. L. J. Org. Chem. 1988, 53, 2629.

3478 Inorganic Chemistry, Vol. 37, No. 14, 1998
Rodriguez et al.
Scheme 1
Scheme 2
temperature is lowered. Exchange between free and coordinated
H2 explains the line broadening of both signals when the
temperature is increased. Unfortunately, it was not possible to
heat the samples in order to reach the fast exchange regime
exhibiting one coalesced line. As in the case of 1, only one
signal is observed for the metal-bound hydrogen atoms of 3,
indicating a fast intramolecular hydrogen-dihydrogen conver-
sion. Line broadening of signal c occurs at very low temper-
atures for the same reasons as discussed in the case of 1. In
this Freon mixture, in the presence of excess H2 (released from
reaction), we observe not the 16-electron complex 2 but the
18-electron species 3. In preceding experiments carried out in
toluene-d8, we could see both the 16-electron RuH(H2)I(PCy3)2
and the 18-electron RuH(H2)2I(PCy3)2, even under 6 bar of
dihydrogen.12b In the present case, the exclusive observation
of 3 could result from a better solubility of dihydrogen in Freons.
Figure 5. Superposed experimental and calculated NMR signals a and
b of the spectra shown in Figure 4.
A line shape analysis of signals a-d was carried out in order
to obtain quantitative kinetic parameters. The superposed
experimental and calculated spectra are shown in Figures 5 and
6. The line shape simulations of signals a and b were based
on usual two-state exchange between two unequally populated
environments a and b, corresponding to signals a and b. The
mole fraction xa ) 1 - xb ) 0.47 was obtained by line shape
analysis in the slow exchange region and found to be indepen-
dent of temperature within the margin of error. The chemical
shifts δa and δb were slightly temperature dependent and were
extrapolated to high temperature. The agreement between the
experimental and calculated spectra is excellent, as indicated
in Figure 5. The rate constants obtained can be expressed as
Figure 6. Superposed experimental and calculated NMR signals c and
d of the spectra shown in Figure 4.
The major signal is a triplet labeled a/b, appearing at -12.6
ppm (JPH ) 31 Hz), and was assigned to the new complex RuH2-
Cl2(PCy3)2 (4), the synthesis and characterization of which will
be described in section 5.
The spectra exhibited interesting changes when the sample
was cooled. First, the signal due to 4 broadened and then split
into two sharp triplets labeled as a and b in Figure 5, at -12.7
ppm (JPH ) 29 Hz) and -13.2 ppm (JPH ) 32 Hz). The
coalescence temperature is 220 K. The intensity ratio Ia/Ib of
these peaks was found to be 1.13. Therefore, the line broaden-
ing of signal b is more pronounced than that of signal a when
temperature is increased, and the line shape is strongly asym-
metric at the coalescence point (∼220 K). The longitudinal
relaxation times T1 were measured at 183 K and 500 MHz. We
obtained a common value of 430 ms for each triplet in
agreement with a classical dihydride formulation.
In addition to signals a and b, two new broad peaks appeared
at -7.9 (signal c) and 4.6 ppm (signal d; see Figures 4 and 6).
Signal c was assigned by comparison to a preceding experiment
with RuH(H2)2I(PCy3)212b to RuH(H2)2Cl(PCy3)2 (3), for which
a T1 value of 22 ms was measured at 183 K (500 MHz). Signal
d corresponds to free dihydrogen and sharpens when the
kab ) 1014 exp(-47 kJ mol-1/RT), 193 < T < 243 K,
kab(223 K) ) 600 s-1 (1)
In principle, the line shape of signals c and d in Figure 6
could be described in terms of Scheme 1, which corresponds
to exchange between free and coordinated H2. However, in
the case of the spectra of Figure 6, the concentration of 2 is too
small to be detected and does not influence the actual line shape
any more. According to arguments described previously,18 the
line shape equation can then no longer be based on Scheme 1
but must be based on Scheme 2 where 2 is a nonobserved
intermediate. Scheme 2 can be abbreviated by the equation
(18) Limbach H. H. J. Magn. Reson. 1979, 36, 287.

Fluxionality and Isomerism of RuH2(H2)2(PCy3)2
Inorganic Chemistry, Vol. 37, No. 14, 1998 3479
kex
H2(free) + H* (3) y
\z H*2(free) + H2(3)
(2)
2
where H2(free) (signal d) corresponds to free dissolved dihy-
drogen and H2(3) to dihydrogen bound to Ru in 3 (signal c).
Some uncertainty is introduced because of the short T2 of 3 at
low temperatures. The quantities obtained by line shape
simulations are kdc and kcd, where
kcd ) V/c{H2(3)} ) kex c{H2(free)}
kdc ) kcd xd/xc ) V/c{H2(free)} ) kex c{H2(3)}
(3)
k
ex is the pseudo-second-order rate constant of the reaction and
V its velocity. c{i} represent the concentrations and xi the mole
fractions. A formal kinetic analysis of the reaction network of
Scheme 2 using the usual steady-state condition for the
intermediate 2 shows that kex is related to k1 and k-1 as follows:
kex ) (k1k-1)/(k-1 + k-1) ) k1/2
(4)
The c{i} were not exactly known but were estimated to be
less than 0.1 mol L-1. By contrast, the mole fractions xc ) 1
- xd ) c{H2(3)}/(c{H2(free)} + c{H2(3)}) ) 1 - xd were
obtained by line shape analysis and found to be independent of
temperature; i.e., xc ) 1 - xd ) 0.6. Finally, we obtained
kex c{H2(free)} ) 1.5kex c{H2(3)} )
1014 exp(-47 kJ mol-1/RT), 203 < T < 223 K
Figure 7. The four isomeric structures of 1, RuH2(H2)2(PH3)2,
considered in this work.
kex c{H2(free)}(223 K) ) 1500 s-1
Table 1. Selected Geometrical Parameters (Å and deg) of the
DFT/B3LYP Optimized Stationary Points in the RuH2(H2)2(PH3)2
Model Complex (Basis Set B)
Since c{H2(free)} < 0.1 mol L-1, it follows that the
preexponential factor characterizing the exchange is larger than
1014 s-1. However, because of the limited range where rate
constants could be obtained, all kinetic parameters are only
rough estimates.
1a (C2V)
1b (C2V)
1c (Cs)
1d (C2h)
Ru-H1
1.784
1.818
1.818
1.784
1.621
1.621
2.311
2.311
0.853
0.853
163.8
27.4
1.809
1.809
1.809
1.809
1.621
1.621
2.306
2.306
0.846
0.846
151.0
25.1
1.805
1.805
1.808
1.776
1.619
1.624
2.310
2.310
0.846
0.858
158.6
27.1
1.709
1.709
1.709
1.709
1.682
1.682
2.323
2.323
0.913
0.913
180.0
31.0
149.0
31.0
105.5
180.0
74.5
Ru-H2
Ru-H3
Ru-H4
4. Theoretical Study of a Model for 1
Ru-H5
(i) Stability of Different Isomers. Several different isomers
may be envisaged for 1, depending on the relative orientations
of the various ligands. As the PCy3 ligands involved in 1 itself
contain too many atoms for a reliable ab initio study with current
technology, we replaced them by PH3. This is a standard
approximation, which should not significantly affect the results
obtained. The most important isomers of 1 are sketched in
Figure 7, which shows the atom-labeling scheme used. All may
be described as essentially octahedral with two trans phosphines;
the steric bulk of the PCy3 ligands in the real complex 1 ensures
that they cannot occupy cis positions.
Ru-H6
Ru-P1
Ru-P2
H1-H2
H3-H4
P1-Ru-P2
H1-Ru-H2
H2-Ru-H3
H3-Ru-H4
H4-Ru-H5
H5-Ru-H6
H6-Ru-H1
72.9
27.4
75.8
80.7
91.2
25.1
89.7
89.4
84.0
27.7
77.2
83.6
75.8
89.7
87.8
Isomer 1a has C2V symmetry. All six hydrogen atoms
coordinated to Ru lie in a plane which contains the C2 axis,
and the two hydride ligands are cis. In the case of isomer 1b,
which also has C2V symmetry, the two H2 ligands are perpen-
dicular to the plane which contains the C2 axis. A lower
symmetry (Cs) defines 1c, in which the mirror plane contains
the two hydrides and just one of the dihydrogen ligands, the
other H2 unit being perpendicular to that plane. All the six
hydrogen atoms coordinated to Ru again lie in a plane for isomer
1d (C2h), but the two hydrides are now trans.
(MP2), and density functional methods. A hybrid form of
density functional theory (DFT) usually known as B3LYP was
adopted.20 This method has already been shown to be appropri-
ate for the study of the electronic structures and properties of
dihydrogen complexes.21-23 It is now widely agreed that the
MP2 method is less reliable for transition-metal complexes than
(20) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 1372 and 5648. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(21) (a) Gelabert, R.; Moreno, M.; Lluch, J. M.; Lledos, A. Organometallics
1997, 16, 3805. (b) Ricca, A.; Bauschlicher, C. W. Chem. Phys. Lett.
1995, 245, 150. (c) Siegbahn, P. E. M. J. Am. Chem. Soc. 1996, 118,
1487.
These four isomers, together with several others of lesser
importance,19 were investigated theoretically using Hartree-
Fock (HF), second-order Mo¨ller-Plesset perturbation theory
(22) Bernardi, F.; Bottoni, A.; Calcinari, M.; Rossi, I.; Robb, M. A. J. Phys.
Chem. A 1997, 101, 6310.
(23) Bytheway, I.; Backsay, G. B.; Hush, N. S. J. Phys. Chem. 1996, 100,
6023.
(19) Barthelat, J.-C.; Hussein, K.; Marsden, C. J. Manuscript in preparation.

3480 Inorganic Chemistry, Vol. 37, No. 14, 1998
Rodriguez et al.
Table 2. Relative Energies (kcal/mol) of the Optimized Stationary Points in the RuH2(H2)2(PH3)2 Model Complex
HF//HF
MP2//MP2
DFT//DFT
MP2//DFT/B
C
MP4SDQ//DFT/B
C
CCSD(T)//DFT/B
C
basis set
A
B
A
B
A
B
C
1a
1b
1c
1d
0.0
0.0
0.45
1.77
0.
0.0
1.15
0.0
0.0
0.0
1.97
0.95
0.02
1.22
0.
0.0
1.79
0.13
0.0
1.43
0.26
0.69
0.73
6.52
1.29
1.07
5.32
1.63
0.82
11.64
1.99
0.99
10.51
15.33
12.59
is DFT; our present results, discussed below, provide further
support for this assertion.
Geometrical parameters optimized for the four isomers 1a-d
at the B3LYP/basis B level of theory are presented in Table 1.
The most interesting structural aspect of these complexes is
probably the length of the H-H bond in the coordinated
dihydrogen ligand, relative to its value in free H2. We find an
increase in bond length of 12.4% for isomer 1a at the B3LYP
level of theory, with very similar changes for 1b and for 1c.
MP2 theory predicts a greater lengthening of the H-H bond,
by about 20%. It has already been noted that MP2 theory
overemphasizes donation into the σ* H-H antibonding orbital,
compared to DFT methods;23 the comparison of DFT, MP2,
and CCSD(T) relative energies presented below for isomers of
1 leads us to consider the DFT results as more reliable for these
nonclassical complexes.
Relative energies for the different isomers obtained at various
levels of theory with bases A, B, and C are reported in Table
2. It is clear that the three isomers 1a-c, which differ only in
the orientations of the dihydrogen ligands relative to the plane
containing Ru and the two hydrides, are very close in energy
(separations of the order of 1-2 kcal/mol), whereas isomer 1d,
in which the two hydride ligands are trans, is substantially less
stable (by some 10-15 kcal/mol) than the others. Isomer 1a
is consistently the most stable using HF, DFT, or CCSD(T)
methods, irrespective of the basis used. However, the results
obtained with MP2 theory are inconsistent; basis A gives a small
preference for 1c over 1a, whereas isomer 1c is no longer a
stationary point on the potential surface with the more complete
basis B, as the H2 unit in the plane splits to give two hydride
ligands, increasing the oxidation state of Ru to IV. If there is
an inconsistency between CCSD(T) and MP2 results, we clearly
prefer the far more rigorous CCSD(T) values, and since the DFT
and CCSD(T) results for the energy separations between isomers
are fairly similar, we believe that the DFT results should be
preferred over MP2 data. Other authors previously noted that
DFT theory gives results in good agreement with more elaborate
(CASPT2) methods for organometallic complexes. Even though
the energetic preference for a coplanar arrangement of all six
H atoms bound to Ru in 1 is slight, we believe it to be real; the
consistency of the HF, DFT, and very high-level CCSD(T)
results is striking.
Figure 8. Two isomeric structures of RuH2(H2)(PH3)2.
Table 3. Relative Energies (kcal/mol) and Selected Geometrical
Parameters (Å and deg) of DFT/B3LYP Optimized Stationary
Points in the RuH2(H2)2(PH3)2 Model Complex (Basis Set B)
5a (Cs)
5b (Cs)
Ru-H1
1.767
1.800
1.635
1.565
2.303
2.303
0.860
164.9
27.9
1.804
1.804
1.563
1.630
2.300
2.300
0.850
157.9
27.2
Ru-H2
Ru-H5
Ru-H6
Ru-P1
Ru-P2
H1-H2
P1-Ru-P2
H1-Ru-H2
H2-Ru-H5
H5-Ru-H6
H6-Ru-H1
170.3
87.0
74.8
166.4
89.2
90.2
∆E
0.0
0.79
effect has been discussed by Eisenstein and co-workers in
connection with molecule 2, RuH(H2)I(PCy3)2.12,24 The Ru-H
σ bonding orbital can be partly delocalized into the σ* H-H
antibonding orbital of a neighboring H2 ligand, providing that
all four atoms are coplanar. The effect will clearly be
maximized in 1a, in which all six H atoms are coplanar, absent
in 1b where symmetry prevents its operation, and partly present
in 1c. Wishing to discover whether this stabilizing cis effect
may be used to rationalize the structural preferences of dihy-
drogen complexes in general, rather than being limited just to
complex 1, we have also investigated two related systems, RuH2-
(H2)(PH3)2 (5), sketched in Figure 8, and RuHCl(H2)2(PH3)2
(3).
We have undertaken B3LYP geometry optimizations (basis
B) of isomers 5a and 5b, which differ only by the orientation
of the dihydrogen ligand with respect to the plane containing
Ru and the two hydrides. Structural parameters are reported
in Table 3. 5a, in which the cis effect can operate because the
H2 unit is coplanar with the hydrides, is slightly more stable
than 5b, by just 0.8 kcal/mol. The geometries of the four
isomers 3a-d have been optimized at the same level of theory.
The nature of each of these four stationary points on the
potential energy surface was determined by calculation of the
vibrational frequencies, using basis A. The two isomers 1a and
1b are found to be true minima, whatever level of theory is
used. Isomer 1c has one imaginary vibrational frequency at
the HF and DFT levels. However, it is a minimum at the MP2
level. Isomer 1d is a true minimum only at the HF level. The
relatively high energy calculated for isomer 1d is consistent
with the IR spectrum of 1, which was interpreted to indicate a
cis-dihydride geometry. Before analyzing the cis-trans issue,
we consider the preference for coplanarity of the six H atoms
bound to Ru.
(24) Van der Sluys, L. S.; Eckert, J.; Eisenstein, O.; Hall, J. H.; Huffman,
J. C.; Jackson, S. A.; Koetzle, T. F.; Kubas, G. J.; Vergamini, P. J.;
Caulton, K. G. J. Am. Chem. Soc. 1990, 112, 4831.
The electronic origins of this preference may be traced to
the interaction which has been named the “cis effect”. The

Fluxionality and Isomerism of RuH2(H2)2(PCy3)2
Inorganic Chemistry, Vol. 37, No. 14, 1998 3481
Table 4. Relative Energies (kcal/mol) and Selected Geometrical
Parameters (Å and deg) of DFT/B3LYP Optimized Stationary
Points in the RuHCl(H2)2(PH3)2 Model Complex (Basis Set B)
3a (Cs)
3b (Cs)
3c (Cs)
3d (Cs)
Ru-H1
1.658
1.681
1.868
1.868
1.597
2.463
2.339
2.339
0.945
0.820
162.9
32.9
1.633
1.691
1.923
1.913
1.591
2.465
2.337
2.337
0.938
0.802
166.5
32.5
1.699
1.699
1.865
1.865
1.600
2.461
2.340
2.340
0.899
0.821
156.7
30.7
1.695
1.695
1.892
1.881
1.596
2.468
2.341
2.341
0.904
0.810
161.4
30.9
Ru-H2
Ru-H3
Ru-H4
Ru-H5
Ru-Cl
Ru-P1
Ru-P2
H1-H2
H3-H4
P1-Ru-P2
H1-Ru-H2
H2-Ru-H3
H3-Ru-H4
H4-Ru-Cl
Cl-Ru-H5
H5-Ru-H1
81.9
25.4
86.3
87.3
70.5
24.1
71.7
87.
95.6
24.5
88.5
91.1
83.2
24.8
76.7
88.9
71.9
71.1
88.1
86.7
∆E
0.0
1.33
2.25
3.13
energies of 3b and 3c, suggesting a weak coupling between the
rotational motions of the two H2 ligands. This comparison of
three different systems reveals a consistent picture; the most
stable isomer in each case is that which would be predicted by
the cis effect, an effect already analyzed in the literature.
The low stability of 1d compared to 1a could be anticipated,
given the generally accepted high “trans influence” of the
hydride ligand. From an orbital point of view, we find that the
preference for a cis orientation of the two hydrides in 1 may be
traced to the nature of the two MOs which are primarily
responsible for Ru-H bonding. Bonding between Ru and
dihydrogen does not give rise to any particular preferred
orientation of these ligands. Both Ru-H bonding MOs involve
Ru d orbitals for 1a, but only one MO with good Ru-H bonding
overlap can be constructed using Ru d orbitals if the two
hydrides are trans, the other MO involving a 5p orbital on Ru
which is at substantially higher energy than the 4d set. As a
consequence of these orbital differences, we note that the Ru-
hydride bond lengths in 1a are substantially shorter than the
distances between Ru and the H atoms of the dihydrogen
ligands, by 0.18 Å on average, whereas in 1d, with weak Ru-H
bonds, the difference is only 0.027 Å.
Figure 9. The four isomeric structures of RuHCl(H2)2(PH3)2 (3).
They are sketched in Figure 9. If the cis effect operates for the
Ru-H bond, the dihydrogen close to the hydride will lie in the
equatorial plane. Since Cl is more electronegative than H, one
does not anticipate that there will be a significant tendency of
the Ru-Cl bonding orbital to delocalize. In the absence of any
cis effect involving the Ru-Cl bond, the H2 unit adjacent to Cl
may be expected to lie perpendicular to the other H2, to allow
maximal interaction with the Ru d orbitals; if the two dihydro-
gens are coplanar, they are effectively competing for the same
+
d orbitals on Ru. We note that the two H2 units in Co(H2)2
are indeed perpendicular,25 as are the two ethylene ligands in
Ni(C2H4)2.22
The lowest-energy isomer of 3 is 3a, which follows exactly
the predictions based on the cis effect. It has Cs symmetry and
is a minimum on the potential energy surface. Isomers 3b and
3c both contradict the predictions of the cis effect rule, in one
respect, because the H2 unit cis to Cl lies in the plane for 3b,
whereas the H2 cis to H is perpendicular to the plane for 3c.
One imaginary vibrational frequency is found for both isomers,
and in each case, that motion corresponds to the H2 rotation
which leads to isomer 3a. The relative energies of these isomers
are given in Table 4; the differences are small, as are those of
the isomers 1a-c. We note that the orientational preference
(rotation barrier) is nearly twice as great for the H2 adjacent to
H (isomer 3c) as for that adjacent to Cl (isomer 3b), showing
the importance of the cis effect. The least stable isomer is 3d,
for which both H2 orientations contradict the cis effect rule;
two imaginary vibrational frequencies are found for this isomer,
corresponding to the two H2 rotations needed to return to isomer
3a. While we do not wish to overinterpret rather small energy
differences, we note that the energy separation between 3a and
3d is close to that obtained simply by adding the relative
The IR spectrum of 1 provides valuable information on the
nature and strength of the Ru-H and Ru-(H2) interactions; as
the spectrum has been recorded and partially assigned,6a
a
comparison with the calculated values provides a stringent test
of the quality of the computed data. Two very intense bands
are seen at 1927 and 1890 cm-1, assigned to Ru-H stretching
motions; however, no absorption was observed which could be
attributed to stretching either of the H2 units or of the Ru-(H2)
bonds. The calculated harmonic wavenumbers (B3LYP/basis
B) for the two Ru-H stretching vibrations in 1a are 2031 (a1)
and 1995 (b2) cm-1; both modes are indeed intense, with
absolute intensities of 112 and 134 km/mol, respectively. While
the calculated harmonic frequencies are therefore some 5.5%
higher than the observed fundamentals, a typical margin of
error,26 we note that the difference in frequency between the
two modes is very well reproduced. The stretching motions of
the two H2 units are predicted at 3038 (a1) and 3017 (b2) cm-1
,
with high absolute intensities of 112 and 140 km/mol, respec-
tively. This result is surprising, given the known low intensity
of H-H stretches which have been detected,1,7 and one might
(25) Bauschlicher, C. W.; Maitre, P. J. Phys. Chem. 1995, 99, 3444.
(26) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391.

3482 Inorganic Chemistry, Vol. 37, No. 14, 1998
Rodriguez et al.
Figure 11. RHF and DFT energy profiles for rotation of the two H2
ligands in RuH2(H2)2(PH3)2. Θ indicates the dihedral angle between
the Ru(H2) and Ru(H)2 planes.
at the RHF level of theory using basis A. While the results
obtained at this level may not be quantitatively reliable, the
qualitative features should be informative. Fixed angles of
rotation of the H2 units were adopted, with respect to the plane
containing the Ru and the two hydride ligands, and the
remaining independent geometrical parameters (19 for the C2
pathway, 21 for the Cs) were optimized. It may be seen from
Figure 11 that the C2 pathway is consistently preferred over
the Cs and that the maximum along the pathway is closer to
1b, consistent with the greater stability of 1a than of 1b. In
view of the substantial computational effort needed to scan the
entire C2 and Cs surfaces at higher levels of theory, we
concentrated our attention on those parts of the surfaces near
the higher-energy isomer 1b, i.e. Θ angles near 90°. Extension
of the basis from A to B increases the RHF barrier to H2 rotation
somewhat, from 0.9 to 1.3 kcal/mol, and moves the transition
state closer to 1b; that isomer remains a true minimum, though
the magnitude of its lowest vibrational frequency drops sharply.
When electron correlation effects are included at the B3LYP
level of theory with basis B, the barrier increases again, from
1.3 to 2.0 kcal/mol; more importantly, 1b now becomes a
transition state, as it is a local maximum on the C2 surface.
Thus as the level of theory adopted is improved, the derived
potential governing H2 rotation becomes progressively closer
to the simple 2-fold cosine function used in analysis of the INS
data. In particular, we may take the energy difference between
1a and 1b to represent the potential barrier. Our B3LYP/basis
B value of 2.0 kcal/mol for this barrier agrees remarkably well
with the result inferred from the INS data (2.2 kcal/mol,
assuming uncoupled rotation of the two H2 molecules). How-
ever, the presumably more reliable CCSD(T) value for the
barrier obtained with basis C is a little smaller, at 1.4 kcal/mol,
so the particularly good agreement at the DFT/basis B level is
probably fortuitous. The qualitatively important point is that
the calculated barrier height agrees rather well with that derived
from experiment with the use of a simplified model. We may
also note that the barrier to rotation along the Cs pathway is
only slightly higher, by some 0.1 kcal/mol, than that for coupled
C2 rotation, so at most temperatures both pathways will in fact
be followed. The effects of basis extension on the calculated
barrier are not large; we note in particular that the DFT result
with basis C differs from that with basis B by an utterly trivial
amount (0.02 kcal/mol). This agreement for the barrier therefore
suggests that the other calculated results are also reliable.
Figure 10. The two possible pathways for rotation of the two
dihydrogen ligands in 1. Hydrogen atoms of the PH3 groups have been
omitted for clarity.
suppose that such intense bands would be readily detectable;
unfortunately, they are likely to be masked in the real system
by the many C-H stretching modes within the PCy3 ligands.
Two twisting motions of the H2 units in the RuH2(H2)2 plane
are predicted at 1640 and 1635 cm-1, with moderate to low
intensity (44 and 4 km/mol, respectively), while the vibrations
best described as Ru-(H2) stretches are predicted at 997 (a1)
and 925 (b2) cm-1, but these are both weak with intensities of
only 2 and 7 km/mol. The vibrational wavenumbers were also
obtained for 1a at the MP2 level of theory using basis A. For
the two Ru-H stretching motions, the values are 2123 (a1) and
2116 (b2) cm-1; these are substantially poorer than those
obtained using DFT methods, both for the absolute values of
the wavenumbers, in error here by some 10-11%, and for the
separation between the two modes (only 7 cm-1, compared to
37 cm-1 experimentally or 34 cm-1 with B3LYP theory). These
results provide another indication that the B3LYP method is
more successful than low-order perturbation theory for the study
of organometallic complexes. It is encouraging to note that
extension of the basis from A to B changes the B3LYP
wavenumbers for these two modes by no more than 4 cm-1
.
(ii) Rotation of Dihydrogen Ligands in 1. Since the INS
spectra described above (section 2) have shown that rotation of
the dihydrogen ligands in 1 is facile, there is an obvious interest
in a theoretical study of the potential function for H2 rotation
in 1. As there are two equivalent dihydrogen ligands, both must
participate in the process detected by INS, but it is not clear a
priori whether the two ligands rotate in the same or in opposite
directions or, indeed, whether the coupling between the motion
of the two H2 units is sufficiently strong for a perceptible
difference to exist between the two possibilities. The two
possible pathways from 1a to 1b are sketched in Figure 10; if
the two ligands rotate in opposite senses, a Cs surface is
followed, whereas rotation of the two H2 units in the same senses
generates a C2 pathway. Our most important findings are
displayed in Figure 11, which shows that the calculated rotation
barriers are very small.
We initially scanned the whole of both the C2 and Cs surfaces,

Fluxionality and Isomerism of RuH2(H2)2(PCy3)2
Inorganic Chemistry, Vol. 37, No. 14, 1998 3483
5. NMR Studies of 1 in CDCl3: Synthesis and Theoretical
Studies of RuH2Cl2(PCy3)2 (4)
(i) Synthesis and NMR Studies. Since we suspected that 4
resulted from the reaction of 1 with the chloroform present in
Freon, the direct reaction of 1 with CDCl3 was carried out in
an NMR tube. Vigorous H2 evolution was immediately
observed, and only one triplet signal at -12.38 ppm (JP-H
)
31.5 Hz) was detected in the high-field region which cor-
responded to 4. When the sample was cooled to 233 K, the
signal broadened and split into two broad triplets, well resolved
in Freon (vide supra) or in a mixture of CDCl3/CD2Cl2 at 183
K. The 31P{1H} NMR spectrum in CDCl3 at 297 K presents a
singlet at 91.3 ppm split into a triplet with JP-H ) 31.7 Hz
after decoupling of the protons of the cyclohexyl groups. This
is demonstrative of the presence of two hydrides bound to the
ruthenium. At 208 K, the 31P{1H} NMR spectrum in CDCl3
consists of two broad singlets at 97.1 and 83.1 ppm (integration
1
ratio 1.1:1) in agreement with the observations recorded in H
NMR. Interestingly, the appearance of 4 is not accompanied
by the formation of CDHCl2 as monitored by 1H NMR.
However, as the signal characteristic for 4 disappears, a new
signal appears at +5.28 ppm (1:1:1 triplet, JH-D ) 0.9 Hz)
attributed to CDHCl2 formed during the decomposition. The
chloride ruthenium compounds then formed were not character-
ized.
Attempts at isolating 4 by reacting 1 with CHCl3 using
different experimental procedures (2-fold or excess CHCl3 in
pentane or THF, pure CHCl3) failed because of the lack of
stability of 4 in the presence of CHCl3. For example, the beige
solid obtained from pure chloroform after evaporation and
washing in pentane is a mixture of unidentified chloride
ruthenium complexes and the dihydride complex 4. Similar
results were obtained when 1 was reacted with CCl4, Cl2, or
HCl (either gas or aqueous solution) in THF. Nevertheless,
addition of aqueous HCl to a suspension of 1 in Et2O yielded
a pink suspension which after filtration and washing with
Figure 12. Isomeric structures of 4, RuH2Cl2(PH3)2.
of signals corresponding to two interconverting isomers is
apparent at low temperature. That the different signals observed
for 4 below 203 K correspond to two isomers rather than to
one asymmetric complex results from the integration ratio in
31P and 1H NMR which give reproducibly a 1.1:1 ratio and from
asymmetry in the coalescence spectrum as evidenced by line
shape analysis (vide supra). However, although 4-Os is shown
to consist of a mixture of a symmetric (minor) and an
asymmetric (major) complex, the structures of which have been
proposed after ab initio calculations, complex 4 exists as two
symmetric interconverting isomers. It is not possible to rule
out the possibility for each triplet to result from the presence
of fast interconverting asymmetric isomers even at very low
temperature in Freons. This would however contradict the
results of the theoretical calculations (vide infra).
4 does not react with H2 in benzene but does react with CO
to give RuHCl(CO)2(PCy3)2, identified by comparison with an
authentic sample.12b The reaction proceeds through selective
reductive elimination of HCl rather than that of H2 or Cl2 which
is rather surprising.
(ii) Theoretical Study of a Model for 4. To make feasible
a computational study of the structures and energies of possible
isomers of 4, we adopted the “standard” approximation of
replacing the PCy3 phosphines by PH3, to give 4, whose formula
is thus RuH2Cl2(PH3)2. Geometries were optimized using
B3LYP theory and basis B. Several isomers were studied, of
which the four most important are sketched in Figure 12, which
also indicates the atomic numbering scheme. Optimized
structural parameters and relative energies are reported in Table
5.
1
pentane was identified by H and 31P NMR as pure 4 (yield
60%). All data were then consistent with the formulation RuH2-
Cl2(PCy3)2. 4 is therefore a 16 electron complex and a rare
example of a ruthenium(IV) polyhydride. During the prepara-
tion of this paper, Werner et al. reported the preparation of the
very similar complex RuH2Cl2(PiPr3)2 (4-iPr) identified by an
X-ray crystal structure.13 Several ruthenium(IV) hydrido de-
rivatives have been described,27-29 and we reported a few years
ago the synthesis of a similar but 18-electron complex RuH2-
(OCOCF3)2(PCy3)2. Interestingly, the latter compound, although
electronically saturated, was shown to be thermally unstable
and to lose H2 slowly in solution. A similar pathway is probably
responsible for the degradation of 4 in solution. 4 is also
analogous to the osmium derivative OsH2Cl2(PCy3)2 (4-Os),
prepared and studied by Berke and Caulton.31 The spectroscopic
properties of both compounds are very similar, but as expected,
the osmium complex is more stable. A fluxional process is
present in both compounds, and a decoalescence into two sets
(27) (a) Baird, G. J.; Davies, S. G.; Moon, S. D.; Simpson, S. J.; Jones, R.
H. J. Chem. Soc., Dalton Trans. 1985, 1479. (b) Arliguie, T.; Border,
C.; Chaudret, B.; Devillers, J.; Poilblanc, R. Organometallics 1989,
8, 1308. (c) Paciello, R. R.; Manriquez, J. M.; Bercaw, J. E.
Organometallics 1990, 9, 260.
(28) Borowski, A. F.; Sabo-Etienne, S.; Christ, M. L.; Donnadieu, B.;
Chaudret, B. Organometallics 1996, 15, 1427.
(29) Kono, H.; Wakao, N.; Ito, K.; Nagai, Y. J. Organomet. Chem. 1977,
132, 53.
(30) Chung, G.; Arliguie, T.; Chaudret, B. New J. Chem. 1992, 16, 369.
(31) Gusev, D. G.; Kuhlman, R.; Rambo, J. R.; Berke, H.; Eisenstein, O.;
Caulton, K. G. J. Am. Chem. Soc. 1995, 117, 281.
Although the coordination about Ru in 4 might be regarded
as octahedral, there are substantial deviations from the idealized
values of 90 or 180° for some bond angles in each of 4a-d, so
much so that the octahedral paradigm is misleading. Isomer

3484 Inorganic Chemistry, Vol. 37, No. 14, 1998
Rodriguez et al.
Table 5. Relative Energies (kcal/mol) and Selected Geometrical
Parameters (Å and deg) for the DFT/B3LYP Optimized Stationary
Points in the RuH2Cl2(PH3)2 Model Complex (Basis Set B)
amounts in solution, consistent with the NMR behavior of 4
described above. We were not able to find computationally
any other low-energy isomers of 4. Isomer 4c is sufficiently
high in energy (19.3 kcal/mol above 4a) that we do not expect
it to be detectable by NMR spectroscopy in solution at room
temperature. A nonsymmetric isomer 4d was located (C1 point
group), which may be thought of as having essentially trans
orientations of the hydride, chloride, and phosphine ligands, but
it is substantially higher in energy than either 4a or 4b (15.2
kcal/mol above 4a) and therefore unlikely to be present in
detectable amounts in solution.
4a (C2V)
4b (C2V)
4c (C2V)
4d (C1)
Ru-H1
1.560
1.560
2.379
2.379
2.357
2.357
1.414
179.6
53.9
1.567
1.567
2.387
2.387
2.303
2.303
2.655
112.3
115.8
90.2
1.602
1.602
2.393
2.393
2.236
2.236
2.811
103.3
122.6
87.9
1.554
1.604
2.410
2.396
2.469
2.227
2.567
140.2
108.8
158.1
79.5
Ru-H2
Ru-Cl1
Ru-Cl2
Ru-P1
Ru-P2
H1-H2
P1-Ru-P2
H1-Ru-H2
Cl1-Ru-Cl2
H1-Ru-Cl2
H2-Ru-Cl1
It is intriguing that the relative DFT energies of the isomers
of 4 described here are rather different from those of the
analogous Os system studied by Eisenstein, Caulton, and co-
workers at the MP2 level of theory.31 However, the differences
may be more apparent than real; when MP2 energies are
obtained for 4a and 4b, 4b is slightly more stable, by 2.5 kcal/
mol, i.e., just the reverse of the DFT result, even though the
differences between DFT and MP2 geometrical parameters are
insignificant. Since we have argued above that DFT results
are more reliable than are the MP2 values for 1, we presume
that the same preference will still apply for 4.
149.0
78.5
78.5
112.0
112.0
74.7
74.7
106.6
∆E
0.0
2.32
19.28
15.18
4a has C2V symmetry, with the two chloride and hydride ligands
adopting cis positions in a common plane. The Ru-H distances
are calculated to be a little shorter (0.061 Å) than those in 1a,
implying a reduced covalent radius for Ru(IV) compared to Ru-
(II), but the Ru-P distances change in the reverse order, being
0.046 Å greater in 4a than in 1a. Although the P-Ru-P angle
in 4a is almost exactly 180°, the H-Ru-H angle is notably
acute (only 53.9°) and the Cl-Ru-Cl angle of 149.0° is closer
to linearity than to 90°. One cannot argue convincingly that
the large angle involving the chloride ligands in 4a is due to
steric repulsion between them, since in 4b, which also has C2V
symmetry, the Cl-Ru-Cl angle is only 90.2° and the Ru-Cl
distances in 4a and 4b are almost identical. The most striking
feature of 4b, for which the two chloride and phosphine ligands
occupy the same plane, is the position of the two hydride
ligands; they are decidedly on the phosphine side of the
coordination sphere, making an angle H-Ru-H of only 115.8°.
A related isomer in which the two hydride ligands are in
analogous positions but on the chloride side was also character-
ized computationally, but it is much less stable than 4b, by 31
kcal/mol, and so it is not discussed further here. The structure
of RuH2Cl2(PiPr3)2, a complex which is closely related to 4,
was recently determined by X-ray diffraction;13 the coordination
about Ru in the solid state was described as a distorted square
antiprism with two vacant sites in alternate positions in one
square base. It is similar to that of 4b, with P-Ru-P and Cl-
Ru-Cl angles of 111.7 (calculated 112.3) and 84.3° (calculated
90.2°), respectively. However, there is a significant structural
difference between 4b and RuH2Cl2(PiPr3)2; while the dihedral
angle between the RuPP and RuClCl planes is 48° in RuH2-
Cl2(PiPr3)2, those groups are coplanar in 4b. As the steric bulk
of the phosphine ligands in the experimental system is far greater
than that of the simple model employed computationally, some
structural differences might well be anticipated; Eisenstein and
co-workers have reported that steric effects in the analogous
Os complex favor a twisting of the OsPP plane relative to the
OsClCl unit.31 Isomer 4c is related to 4a in that the two chloride
and two hydride ligands are coplanar but differs in that the
P-Ru-P angle is only 103.3°.
Conclusion
We describe in this paper the high fluxionality of the bis-
(dihydrogen) complex RuH2(H2)2(PCy3)2 (1). The origin of this
fluxionality is both a rapid hydride-dihydrogen interconversion
which cannot be blocked in Freons down to 143 K and a low
barrier to rotation of coordinated dihydrogen (1.1 kcal/mol) as
demonstrated by INS and studied computationally. Theoretical
studies have shown that the bis(dihydrogen) complex 1 has three
isomeric structures within an energy range of only 2 kcal/mol-1
in agreement with the high fluxionality of this molecule. The
geometry of the lowest energy isomer for 1 is unusual, since
the two H2 units and the hydride ligands are located in the same
plane. Analysis of several dihydrogen complexes shows that
the attractive cis effect23 controls the geometrical preference.
The barrier to rotation of dihydrogen is higher in the 16 electron
iodo complex RuHI(H2)(PCy3)2 (>ca. 3.5 kcal/mol-1), but the
interconversion on the NMR time scale remains also rapid at
all accessible temperatures. For the corresponding chloro
complex, we could demonstrate both the rapid equilibrium
between the mono- and bis(dihydrogen) complexes RuHCl(H2)-
(PCy3)2 and RuHCl(H2)2(PCy3)2 and the exchange between free
and coordinated dihydrogen in RuHCl(H2)2(PCy3)2 which could
be blocked at low temperature.
Experimental Section
General Considerations. Microanalyses were performed at our
laboratory’s microanalyses service. Infrared spectra were obtained as
Nujol mulls on a Perkin-Elmer 1725 FT-IR spectrometer. NMR spectra
were recorded on a Bruker AC200 (at 200.13 MHz for 1H and at 81.015
MHz for 31P), while variable-temperature proton spectra were obtained
by using Bruker AM250 (at 250 MHz for 1H and at 101.202 MHz for
31P) and AMX 500 (500 MHz; Freon experiments) spectrometers, all
of these spectrometers operating on the Fourier transform mode. All
manipulations were carried out in argon atmosphere by use of Schlenk
techniques. Solvents were dried and distilled under dinitrogen and
thoroughly degassed under argon before use. For the low-temperature
NMR experiments, a mixture of deuterated Freons was synthesized by
We calculate the energy difference between 4a and 4b to be
small, with 4a more stable by 2.3 kcal/mol. We see no real
contradiction in the detection in the solid state of an isomer
which we have calculated to be slightly less stable than the
global minimum, as packing effects could easily outweigh
energy differences of only 2 kcal/mol. Because the energy
difference between 4a and 4b is so small and because both were
shown to be true minima by calculation of their vibrational
frequencies, we would expect both to be present in detectable
1
the literature methods.17 According to the H and 13C NMR spectra,
the mixture contained 9% CDCl3, 45% CDFCl2, and 45% CDF2Cl.
The solvent was stored in a stainless steel lecture bottle over basic
alumina in order to remove water and acid impurities. The NMR

Fluxionality and Isomerism of RuH2(H2)2(PCy3)2
Inorganic Chemistry, Vol. 37, No. 14, 1998 3485
samples were prepared using vacuum methods as described previously.32
In particular, pressure/vacuum valve NMR tubes (Wilmad, Buena, NY)
were employed, which could easily be attached or removed from the
vacuum line where they were filled with the solvent by vacuum transfer.
The tubes were stored below 240 K. NMR line shape analyses were
carried out using known methods.18
(T)), adopting geometries already obtained at a lower level of theory.
All calculations were performed with the Gaussian 94 series of
programs.34
Materials. RuH2(H2)2(PCy3)2 (1) was prepared according to pub-
lished methods.6 RuCl3‚3H2O was purchased from Johnson Matthey
Ltd.; PCy3 was purchased from Aldrich.
Preparation of RuH2Cl2(PCy3)2 (4). To a suspension of RuH2-
(H2)2(PCy3)2 (250 mg; 0.38 mmol) in 10 mL of diethyl ether was added
HCl(aq) (67 µL, 0.76 mmol). The reaction mixture was stirred for 1
h at room temperature, during which a pink-red solid precipitated. This
was filtered off, washed with ether (8 × 4 mL), and dried in vacuo.
Yield: 60%. Anal. Calcd for RuC36H68Cl2P2. C, 58.83; H, 9.34.
Found: C, 58.51; H, 9.37.
Inelastic Neutron Scattering. Inelastic neutron scattering studies
were carried out on 1 and on RuH(H2)I(PCy3)2 using the cold neutron
time-of-flight spectrometers MIBEMOL and IN5 of the Laboratoire
Le´on Brillouin (CE Saclay, France) and Institut Laue-Langevin
(Grenoble, France), respectively. Data were collected from 1.5 to 250
K using ∼0.8 g of sample in which the PCy3 ligands were deuterated
for the ILL experiment.
Theoretical Calculations. The core electrons for both Ru and P
were represented by pseudopotentials developed in Toulouse.33 Sixteen
electrons were treated explicitly for Ru (those corresponding to the
atomic levels 4s, 4p, 5s, and 4d) and the five valence electrons for P.
The basis used for Ru is of approximately triple-ú quality, being
specified as (8s,6p,6d)/[5s,5p,3d]. A double-ú plus polarization basis
was employed for P and Cl (d-type exponents 0.45 and 0.65,
respectively). The spherical-harmonic representation of d-type functions
was adopted. In our initial studies, a standard double-ú basis was
adopted for all hydrogen atoms; the resulting basis is indicated as “basis
A”. Most of the results we present were obtained with a larger basis
“B”, in which we added a p-type polarization function (exponent 0.9)
to the six hydrogen atoms directly bound to ruthenium. Our best
estimates of some relative energies were obtained with a still larger
basis “C”, in which a set of f-type functions (exponent 1.2) was added
to the Ru basis. These single-point calculations were performed at
higher levels of theory, such as Mo¨ller-Plesset perturbation theory to
fourth order, treating single, double, and quadruple excitations (MP4SDQ),
or coupled-cluster theory, treating explicitly single and double excita-
tions together with a perturbative estimate of triple excitations (CCSD-
Acknowledgment. The authors thank the Institut Laue-
Langevin and the Laboratoire Le´on Brillouin for the use of their
facilities and Dr. Gerrit Coddens for assistance with data
collection at the Laboratoire Le´on Brillouin. Work at Los
Alamos was supported by the office of Basic Energy Sciences,
US Department of Energy. This work was supported by a
PROCOPE grant. V.R. acknowledges a grant from the Minis-
terio de Educacion y Investigacion, J.T. acknowledges the
Humboldt Association (Bonn, Germany), B.C. and S.S.-E.
acknowledge the CNRS, and H.-H.L. acknowledges the Fonds
der Chemischen Industrie (Frankfurt, Germany) for support. The
theoretical part of this work was supported by the Centre
National Universitaire Sud de Calcul, Montpellier, France
(Project irs1013).
IC970697Y
(34) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson,
B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson, G. A.;
Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen,
W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.;
Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.;
Stewart, J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian
94; Gaussian, Inc.: Pittsburgh, PA, 1995.
(32) Limbach, H. H. Dynamic NMR Spectroscopy in the presence of kinetic
Hydrogen/Deuterium Isotope Effects; NMRsBasic Principles and
Progress, Vol. 23; Springer: Heidelberg, Germany, 1991; Chapter 2.
(33) Bouteiller, Y.; Mijoule, C.; Nizam, M.; Barthelat, J.-C.; Daudey, J.
P.; Pe´lissier, M.; Silvi, B. Mol. Phys. 1988, 65, 2664.