Ruthenium dihydride complexes: NMR studies of intramolecular isomerization and fluxionality including the detection of minor isomers by parahydrogen-induced polarization

Article abstract of DOI:10.1021/ic011255w

NMR studies reveal that complexes Ru(CO)₂(H)₂L₂ (L = PMe₃, PMe₂Ph, and AsMe₂Ph) can have three geometries, ccc, cct-L, and cct-CO, with equilibrium ratios that are highly dependent on the electronic properties of L; the cct-L form is favored, because the σ-only hydride donor is located trans to CO rather than L. When L = PMe₃, the ccc form is only visible when p-H₂ is used to amplify its spectral features. In contrast, when L = AsMe₂Ph, the ccc and cct-L forms are present in similar quantities and, hence, must have similar free energies; for this complex, however, the cct-CO isomer is also detectable. These complexes undergo a number of dynamic processes. For L₂ = dppe, an interchange of the hydride positions within the ccc form is shown to be accompanied by synchronized CO exchange and interchange of the two phosphorus atoms. This process is believed to involve the formation of a trigonal bipyramidal transition state containing an η/²-H₂ ligand; in view of the fact that k_HH/k_DD is 1.04 and the synchronized rotation when L₂ = dppe, this transition state must contain little H-H bonding character. Pathways leading to isomer interconversion are suggested to involve related structures containing η²-H₂ ligands. The inverse kinetic isotope effect, k_HH/k_DD = 0.5, observed for the reductive elimination of dihydrogen from Ru(CO)₂(H)₂dppe suggests that substantial H-H bond formation occurs before the H2 is actually released from the complex. Evidence for a substantial steric influence on the entropy of activation explains why Ru(CO)₂(H)₂dppe undergoes the most rapid hydride exchange. Our studies also indicate that the species [Ru(CO)₂L₂], involved in the addition of H2 to form Ru(CO)₂(H)₂L₂, must have singlet electron configurations.

Full text of DOI:10.1021/ic011255w

Inorg. Chem. 2002, 41, 2960−2970
Ruthenium Dihydride Complexes: NMR Studies of Intramolecular
Isomerization and Fluxionality Including the Detection of Minor Isomers
by Parahydrogen-Induced Polarization
Daniele Schott, Christopher J. Sleigh, John P. Lowe, Simon B. Duckett,* and Roger J. Mawby
Department of Chemistry, UniVersity of York, Heslington, York YO10 5DD, U.K.
Martin G. Partridge
Synetix PCEO, P.O. Box 1, Billingham, CleVeland TS23 1LB, U.K.
Received December 6, 2001
NMR studies reveal that complexes Ru(CO)2(H)2L2 (L ) PMe3, PMe2Ph, and AsMe2Ph) can have three geometries,
ccc, cct-L, and cct-CO, with equilibrium ratios that are highly dependent on the electronic properties of L; the cct-L
form is favored, because the σ-only hydride donor is located trans to CO rather than L. When L ) PMe3, the ccc
form is only visible when p-H is used to amplify its spectral features. In contrast, when L ) AsMe2Ph, the ccc and
2
cct-L forms are present in similar quantities and, hence, must have similar free energies; for this complex, however,
the cct-CO isomer is also detectable. These complexes undergo a number of dynamic processes. For L2 ) dppe,
an interchange of the hydride positions within the ccc form is shown to be accompanied by synchronized CO
exchange and interchange of the two phosphorus atoms. This process is believed to involve the formation of a
trigonal bipyramidal transition state containing an η2-H ligand; in view of the fact that kHH/kDD is 1.04 and the
2
synchronized rotation when L2 ) dppe, this transition state must contain little H−H bonding character. Pathways
leading to isomer interconversion are suggested to involve related structures containing η2-H ligands. The inverse
2
kinetic isotope effect, kHH/kDD ) 0.5, observed for the reductive elimination of dihydrogen from Ru(CO)2(H)2dppe
suggests that substantial H−H bond formation occurs before the H is actually released from the complex. Evidence
2
for a substantial steric influence on the entropy of activation explains why Ru(CO)2(H)2dppe undergoes the most
rapid hydride exchange. Our studies also indicate that the species [Ru(CO)2L2], involved in the addition of H to
2
form Ru(CO)2(H)2L2, must have singlet electron configurations.
Introduction
and also that the addition of PPh3 decreased the rate of
hydroformylation. This could imply that the dissociation of
PPh3 is required at some point in the catalytic cycle, perhaps
to provide a vacant site for alkene coordination. However,
it is possible that PPh3 could compete for a site made
available by the dissociation of CO, and in complexes
containing mutually trans hydride and carbonyl ligands, the
trans labilizing influence of the hydride ligand could well
lead to CO loss.3 Eisenberg and co-workers have also
examined the role of Ru(CO)3(PPh3)2 in hydroformylation,
and have reported that the catalysis can be initiated and
accelerated photochemically.4 Related ruthenium complexes
Over the past forty years there have been many reports of
the ability of ruthenium complexes to catalyze reactions. An
early example of ruthenium-based catalysis was the use of
Ru(CO)3(PPh3)2, originally synthesized by Collman and
Roper,1 to catalyze the conversion of alkenes to aldehydes
under high pressures of H2 and CO. In this case, Wilkinson
and co-workers found that Ru(CO)3(PPh3)2 was converted
to Ru(CO)2(H)2(PPh3)2 before hydroformylation commenced2
* Author to whom correspondence should be addressed. E-mail:
sb3@york.ac.uk.
(1) Collman, J. P.; Roper, W. R. J. Am. Chem. Soc. 1965, 87, 4008.
(2) Sanchez-Delgado, R. A.; Bradley, J. S.; Wilkinson, G. J. Chem. Soc.,
Dalton Trans. 1976, 399.
(3) Kaesz, H. D.; Saillant, R. B. Chem. ReV. 1972, 72, 231.
(4) Gordon, E. M.; Eisenberg, R. J. Organomet. Chem. 1986, 306, C53.
2960 Inorganic Chemistry, Vol. 41, No. 11, 2002
10.1021/ic011255w CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/07/2002

Ruthenium Dihydride Complexes
Scheme 1. Labeling Scheme for Ru(H)2(CO)2L2
alkyne binding, combination of alkyne and hydride ligands
to form a vinyl ligand, and reoccupation of the now vacant
site by CO. Elimination of alkene, leaving the 16-electron
species Ru(CO)2(PMe2Ph)2, could be followed by the oxida-
tive addition of the alkyne, and this is consistent with
deuterium labeling studies which indicate that the hydride
ligand in the product Ru(CO)2(CtCR)H(PMe2Ph)2 is derived
from the alkyne.11 An analogue of Ru(CO)2(PMe2Ph)2, the
complex Ru(CO)2(PBut2Me)2, has been isolated and shown
to possess a singlet ground state and a C2V structure.13
Interestingly, the singlet state of the related model complex
Ru(PH3)4 has been calculated to lie 48.9 kJ mol-1 below
the triplet state, while calculations for Fe(PH3)4 indicate that
the triplet state is the more stable by 33.4 kJ mol-1.14 The
complexes described in this paper are Ru(CO)2(H)2(dppe),
1 (dppe ) PPh2CH2CH2PPh2); Ru(CO)2(H)2(PMe2Ph)2, 2;
Ru(CO)2(H)2(PMe3)2, 3; Ru(CO)2(H)2(AsMe2Ph)2, 4; and Ru-
(CO)2(H)2(PMe3)(PMe2Ph), 5. We characterize the fluxional
behavior of the complexes and describe a study of their
reactions with parahydrogen (p-H2). We pay particular
attention to the detection and characterization of minor
isomers of these dihydrides and discuss the shape and
electronic configuration of the associated intermediates. On
the basis of these data, steric and electronic effects are
separated and linked to isomer stability. In addition, structural
interconversion is linked to the formation of an intermediate
containing an η2-H2 ligand. Some of this work has already
been published as a preliminary communication.15
such as Ru(CO)(H2)(PPh3)3 have found use in catalytic C-C
bond-forming reactions involving aromatic ketones and
suitable alkenes that yield ortho-alkylated adducts.5 Ruthe-
nium complexes also feature as successful asymmetric
hydrogenation catalysts6 and have found an ever-widening
role as olefin metathesis catalysts.7
One of the principal problems associated with the exami-
nation of reaction mechanisms relates to the need to
characterize reaction intermediates and probe their dynamic
behavior. This problem is further exacerbated by the fact
that each intermediate can exist in several distinct structural
forms with different activities. Halpern and Brown most
elegantly illustrated the importance of detecting minor species
during studies involving the hydrogenation of prochiral
enamides by cationic rhodium complexes. Careful NMR
work revealed that, contrary to initial suggestions, the less-
stable enamide isomer reacted substantially faster with H2
and was responsible for the observed enantiomeric selectivity
in the hydrogenation product.8 In this paper we report on
the characterization and the reactivity of a series of ruthenium
complexes with the general formula Ru(CO)2(H)2L2. Previ-
ously, complexes of this type containing two monodentate
ligands L had been shown to adopt a preferred geometry in
which the ligands L are mutually trans and the other two
ligand pairs are mutually cis.9 This structure, unambiguously
labeled cct-P, is illustrated in Scheme 1. For Ru(CO)2(H)2-
(AsMe2Ph)2, a second isomer in which all the ligand pairs
are mutually cis (ccc) is readily observed.10
Experimental Section
The synthesis of the dichloro complexes of ruthenium Ru-
(CO)2Cl2L2 (where L ) PMe3, PMe2Ph, and AsMe2Ph) was
achieved according to literature methods.16,17 These complexes were
then converted to dihydrido analogues Ru(CO)2(H)2L2 via NaBH4
reduction in methanol.9 Ru(CO)2Cl2(PMe3)(PMe2Ph) was prepared
according to literature methods by adding the appropriate quantity
of PMe3 to an ethanolic solution of [Ru(CO)2Cl2(PMe2Ph)]2 and
stirring the solution for 1 h. White crystals of Ru(CO)2Cl2(PMe3)-
(PMe2Ph) were obtained on removal of the solvent, and these were
converted to Ru(CO)2(H)2(PMe3)(PMe2Ph) by reaction with NaBH4
in methanol under a hydrogen atmosphere. It should be emphasized
that the product Ru(CO)2(H)2(PMe3)(PMe2Ph) is contaminated by
both Ru(CO)2(H)2(PMe3)2 (5%) and Ru(CO)2(H)2(PMe2Ph)2 (15%)
after this process. The complex Ru(CO)2(H)2(dppe) was obtained
by UV irradiation of a toluene solution of Ru(CO)3(dppe)2 under 3
atm of H2. A 31P-NMR spectrum recorded after 2 h of irradiation
showed that the resonance for Ru(CO)3(dppe) had almost vanished,
and that new doublets for Ru(CO)2(H)2(dppe) 1 had appeared at δ
71.5 and 64.6. A quantitative conversion to 1 was achieved after
the removal of the mixture of CO and H2 by the freeze/pump/thaw
Detailed studies of the reactions of Ru(CO)2(H)2(PMe2-
Ph)2 with alkynes and alkenes have been reported.11,12 This
dihydrido complex converts alkynes RCtCH (where R )
Ph or CMe3) to alkenes in a reaction sequence which also
involves the formation of alkynyl complexes Ru(CO)2(Ct
CR)H(PMe2Ph)2. A sequence consistent with these observa-
tions would involve initial CO loss to generate a site for
(5) (a) Murai, S.; Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.;
Sonoda, M.; Chatani, N. Nature 1993, 366, 529. (b) Murai, S. S.;
Chatani, N.; Kakiuchi, F. Pure Appl. Chem. 1997, 69, 589. (c) Jazzar,
R. F. R.; Mahon, M. F.; Whittlesey, M. K. Organometallics 2001,
20, 3745.
(6) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40. Noyori,
R. Asymmetric Catalysis in Organic Synthesis; Wiley: New York,
1994.
(13) Ogasawara, M.; Macgregor, S. A.; Streib, W. E.; Folting, K.;
Eisenstein, O.; Caulton, K. G. J. Am. Chem. Soc. 1995, 117, 8869;
1996, 118, 10189.
(14) Macgregor, S. A.; Eisenstein, O.; Whittlesey, M. K.; Perutz, R. N. J.
Chem. Soc., Dalton Trans. 1998, 291.
(15) Duckett, S. B.; Mawby, R. J.; Partridge, M. G. J. Chem. Soc., Chem.
Commun. 1996, 383.
(16) Jenkins, J. M.; Lupin, M. S.; Shaw, B. L. J. Chem. Soc. A 1966,
1787.
(7) (a) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18. (b)
Furstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012. (c) Grubbs, R.
H.; Chang, S. Tetrahedron 1998, 54, 4413.
(8) Brown, J. M. Chem. Soc. ReV. 1993, 22, 25.
(9) Bray, J. M.; Mawby, R. J. J. Chem. Soc., Dalton Trans. 1987, 2989.
(10) Alternative IUPAC stereochemical descriptions equate the cct-P
description with OC-6-13, and the cct-CO OC-6-33 and ccc isomer
with OC-6-32.
(11) Bray, J. M.; Mawby, R. J. J. Chem. Soc., Dalton Trans. 1989, 589.
(12) Vessey, J. D.; Mawby, R. J. J. Chem. Soc., Dalton Trans. 1993, 51.
(17) Barnard, C. F. J.; Daniels, J. A.; Jeffery, J.; Mawby, R. J. J. Chem.
Soc., Dalton Trans. 1976, 953.
Inorganic Chemistry, Vol. 41, No. 11, 2002 2961

Schott et al.
method, followed by repressurization with H2, and irradiation for
an additional 2 h. The same technique, using D2 instead of H2,
was employed to obtain Ru(CO)2(D)2(dppe) 1-D2. A similar
approach has been used to prepare Ru(CO)2(H)2(dmpe).18
ized to 100%. A kinetic model was then constructed that allowed
the peak intensities to be explicitly calculated as a function of the
exchange rate constant and reaction time. The calculated peak
intensities were compared to the experimental values, noting the
squares of the differences.21 The rate constants were varied until
the sum of the squares of the differences between measured and
simulated points was minimized. This was carried out using
Microsoft Excel 97. Rate constants obtained in this way were
multiplied by a factor of 2 to take into account the analysis
method.22
A sample of 13CO-enriched Ru(CO)2(H)2(PMe3)2 was obtained
as follows. First, a sample of Ru(CO)x(13CO)3-x(PMe3)2 (where
x ) 3, 2, 1, or 0) was prepared by photolysis of Ru(CO)3(PMe3)2
under 13CO. The solution was then placed under an atmosphere of
1
hydrogen and photolyzed for 15 min. The H-NMR spectrum of
the solution now contained a triplet at δ -7.48 for the hydride
resonance of the product with fine structure due to the coupling to
one or two 13C nuclei.
Results
According to NMR spectroscopy, complexes 1-4 were free from
any impurities at the outset of the measurements. In the case of
the NMR studies on 5, where a mixture of species was present
originally, and 4, where slight decomposition was observed at higher
temperatures, 2-D monitoring of the dynamic behavior of the system
demonstrated that none of the extra species were involved in the
reported exchange processes.
Characterization and Dynamic Behavior of Ru(CO)2-
(H)2(dppe) 1. NMR spectra of 1 were recorded in a
d8-toluene solution at 296 K. As described in the previous
section, the 31P-NMR spectrum of 1 contained a pair of
doublet resonances at δ 71.5 and 64.5. In the H-NMR
spectrum of 1, the resonances for the two hydride ligands
1
Exchange Process Monitoring. To monitor the exchange
pathways involving complexes 1-4, a series of phase-sensitive
EXSY spectra were recorded at selected temperatures with different
mixing times τ (usually 25, 50, 75, 200, 400, 600, and 800 ms).
For processes involving the exchange of nuclei between two sites,
exact rate constants can be determined via analysis of the intensity
versus the mixing time profile.19 The addition of a third spin,
however, complicates matters. Several methods have been applied
to the determination of rate constants in multispin systems: of these
methods, a comprehensive matrix method is the most thorough.19
In this work, we have employed Jeener’s simplification of the matrix
method for two sites,19 the comprehensive matrix approach,20 and
direct simulation. For the studies of 3 and 4 involving parahydrogen
enhancement, we have used Jeener’s approach alone. Here, integra-
tion of the diagonal-peak and cross-peak resonance volumes
provides the data necessary for the calculation of the rate constant
for exchange according to eq 1
appeared as doublets of doublets of doublets centered at δ
2
-7.55 and -6.32, and the H-H coupling constant, JHH
,
was -4.5 Hz. For the resonance at δ -7.55, the magnitudes
2
of the coupling constants to phosphorus, | JPH|, 21.3 and 26.4
Hz, revealed that the hydride ligand concerned was cis to
both phosphorus atoms (and, therefore, trans to CO). For
2
the resonance at δ -6.32, the values of | JPH| were 21.7 and
73.8 Hz, indicating that this hydride ligand was cis to one
phosphorus atom and trans to the other. Selective irradiation
at the frequency corresponding to δ 71.5 in the 31P spectrum
removed the smaller of the two doublet splittings for the
resonance at δ -6.32 and one of the two doublet splittings
for the resonance at δ -7.55, while irradiation at the
frequency corresponding to δ 64.5 removed the larger of
the two doublet splittings for the resonance at δ -6.32 and,
again, one doublet splitting for the other resonance. This
established that the 31P nucleus giving the resonance at δ
71.5 was cis to both hydride ligands, whereas that responsible
for the resonance at δ 64.5 was trans to one hydride and cis
to the other. In the 13C{1H}-NMR spectrum, the resonances
aA
+ 1
aAB
ln
) kextm
(1)
aA
(
)
- 1
aAB
2
for the carbonyl ligands were a triplet (| JPC| ) 4.7 Hz) at δ
2
201.3 and a doublet of doublets (| JPC| ) 7.1 and 79.5 Hz)
where aA and aAB are respectively the integrals of diagonal-peak
and cross-peak volume, kex is the exchange rate constant, and tm is
the mixing time. This process is straightforward for the exchange
of hydrides within the ccc isomer of complexes Ru(CO)2(H)2L2.
Since the intra-ccc interchange process occurs independently of
the ccc-cct interconversion process, rate constants for the former
process were calculated using an “effective diagonal” made up of
the sum of volume elements due to the ccc isomer. It should also
be noted that the equilibrium concentrations of the cct and ccc
isomers of 2 and 3 revealed that kcctfccc , kcccfcct. To check the
validity of this approximation, we analyzed the data for 4 using
the simplified approach and then demonstrated that the matrix
method yielded the same solution.
at δ 203.8, placing one carbonyl ligand cis to both phos-
phorus atoms and the other cis to one phosphorus atom and
trans to the other. Clearly Ru(CO)2(H)2(dppe) 1 possesses
the ccc structure shown in Figure 1.
Assignments of resonances for individual carbon and
hydrogen atoms within the dppe ligand were made on the
basis of 13C{1H, 31P}, COSY, HMQC, and nOe spectra.23
These studies were performed at 295 K, a temperature at
(21) Jones, W. D.; Rosini, G. P.; Maguire, J. A. Organometallics 1999,
18, 1754.
(22) Green, M. L. H.; Wong, L. L.; Sella, A. Organometallics 1992, 11,
2660.
(23) (a) Aue, W. P.; Bartholdi, E.; Ernst, R. R. J. Chem. Phys. 1976, 64,
2229. Nagayama, K.; Kumar, A.; Wuthrich, K.; Ernst, R. R. J. Magn.
Reson. 1980, 40, 3210. (b) Sorensen, W.; Eich, G. W.; Levitt, M. H.;
Bodenhausen, G.; Ernst, R. R. Prog. Nucl. Magn. Reson. Spectrosc.
1983, 16, 163. Wagner, G.; Bodenhasuen, G.; Muller, N.; Rance, M.;
Sorensen, O. W.; Ernst, R. R.; Wuthrich, K. J. Am. Chem. Soc. 1985,
107, 6440. (c) Kessler, H.; Gehrke, M.; Griesinger, C. Angew. Chem.,
Int. Ed. Engl. 1988, 27, 490.
For the least-mean-squares method, the integrals of the diagonal
peaks and corresponding exchange cross-peaks were first normal-
(18) Whittlesey, M. K.; Perutz, R. N.; Virrels, I. G.; George, M. W.
Organometallics 1997, 16, 266.
(19) Perrin, C. L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935.
(20) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem. Phys.
1979, 71, 4546.
2962 Inorganic Chemistry, Vol. 41, No. 11, 2002

Ruthenium Dihydride Complexes
Figure 1. Schematic view of the structure of 1. The phenyl rings are
indicated by the labels A-D, and the chemical shifts of the ortho protons
and hydride ligands are indicated.
which it was clear that rotation about all the phosphorus-
phenyl bonds was rapid on the NMR time scale. The data
obtained from these spectra are listed in Table 1a. It should
be noted that separate sets of proton and carbon signals were
detected for each of the four phenyl rings (labels A-D in
Figure 1). In the nOe spectrum (measured in phase-sensitive
mode) the signal at δ -6.32, due to the hydride ligand trans
to phosphorus, connected strongly to the phenyl ortho proton
resonances at δ 8.14 and 7.62, thereby identifying these
resonances as belonging to the rings (A and B) attached to
the right-hand phosphorus atom PA in Figure 1. Since the
resonance at δ -7.55 connected strongly with ortho protons
at δ 8.14 and 7.78, it was clear that the δ 8.14 resonance
must belong to ring A and δ 7.78 to ring C. Data from
selectively 31P-decoupled 13C{1H} spectra confirmed that
rings A and B were attached to the right-hand phosphorus
atom (PA resonance at δ 71.5) and C and D to the left-hand
phosphorus atom (PB resonance at δ 64.5).
Figure 2. Two-dimensional 1H-1H-nOe spectrum (positive contours) of
complex 1, showing resonance connections in the ortho proton region that
arise due to phenyl ring interchange in 1 at 296 K. The spectrum was
recorded with a mixing time of 1 s. Transfer of magnetization from the
ortho proton resonance of ring A into the corresponding resonance of ring
C (see Figure 1 for ring labels) is specifically highlighted.
The nOe spectra (normally 31P-decoupled) also contained
positiVe cross-peaks connecting the two hydride resonances.
This observation indicated that the hydride ligands inter-
changed positions, and analysis according to the methods
outlined in the Experimental Section yielded the rate
constants and activation parameters listed in Table 2
(observations between 295 and 328 K). Interestingly, positive
cross-peaks also connected specific ortho-phenyl protons in
different phenyl rings, as illustrated by Figure 2. It was
evident, for example, that phenyl ring A must exchange with
ring C (responsible for the resonances at δ 8.14 and 7.78,
respectively). Similarly, it could be seen that ring B must
exchange positions with ring D. At 306.5 K, the rate constant
for hydride interchange was determined to be 1.20 s-1, while
that for ortho-phenyl proton exchange was 1.28 s-1. Figure
3 illustrates the match between the experimental data and
the simulated data at 317 K.
Figure 3. Simulation/experimental trace showing data mapping the
exchange of ring A (9) into ring C (b) of 1 in toluene-d8 at 317 K. The
defined points represent experimental data, and the solid line indicates the
simulated values obtained when kex is 3.28 s-1
.
To examine the fluxional process in more detail, a high-
resolution 2-D EXSY measurement was performed in which
couplings between hydride ligands and 31P nuclei were
retained (Figure 4). This spectrum revealed that the four
components of the hydride resonance at δ -6.32, which
correspond to the four different combinations of phosphorus
spin states, connected to specific components within the δ
-7.55 resonance. Mann and co-workers have previously used
this approach to study the hydride ligand interchange
Figure 4. 1H{31P}-EXSY spectrum (positive contours) of a sample
containing 1 in C6D5CD3 at 305 K with a mixing time of 0.5 s. Exchange
peaks connect the hydride resonances of ccc-1.
pathway of Ru(CO)(H)2(PPh3)3.24 The observed connections
matched those required for an intra-molecular process
(24) Ball, G. E.; Mann, B. E. J. Chem. Soc., Chem. Commun. 1992, 561.
Inorganic Chemistry, Vol. 41, No. 11, 2002 2963

Schott et al.
Table 1
a. NMR Data for Complex 1 in Toluene-d8 at 296 K
coupling
constant (Hz)
assignment
Ru(CO)2-
(H)2(dppe)
1
coupling
constant (Hz)
assignment
Ru(CO)2-
δ
δ
nucleus (H)2(dppe) 1 (multiplicity)
assignment
nucleus
13C
(multiplicity)
assignment
1H
hydride
-7.55 (ddd) Ru-H
-4.5 2JHH
,
dppe, C1
136.3 (d)
137.2 (d)
137.6 (d)
137.6 (d)
133.7 (d)
132.2 (d)
132.3 (d)
131.7 (d)
127.8 (d)
128.2 (d)
128.1 (d)
128.3 (d)
129.8 (s)
129.3 (s)
129.4 (s)
129.5 (s)
30.5 (dd)
ring A
ring B
ring C
ring D
ring A
ring B
ring C
ring D
ring A
ring B
ring C
ring D
ring A
ring B
ring C
ring D
15.0 | JP(A)C
|
1
2
1
21.3 | JPH|,
14.8 | JP(A)C
|
2
1
26.4 | JPH
|
14.2 | JP(B)C
|
1
-6.32 (ddd) Ru-H
-4.5 2JHH
,
14.6 | JP(B)C
|
2
2
21.7 | JP(A)H
|
dppe, C2 and C6
dppe, C3 and C5
dppe, C4
12.5 | JP(A)C
|
2
2
73.8 | JP(B)H
|
13.1 | JP(A)C
|
3
2
dppe, ortho
dppe, meta
dppe, para
8.14 (ma)
7.62 (ma)
7.78 (ma)
7.71 (ma
7.14 (ma)
7.09 (ma)
7.12 (ma)
7.15 (ma)
7.10 (ma)
7.04 (ma)
7.07 (ma)
7.10 (ma)
2.09 (m)
2.24 (m)
1.97 (m)
2.26 (m)
71.5 (d)
ring A
ring B
ring C
ring D
ring A
ring B
ring C
ring D
ring A
ring B
ring C
ring D
8.4 | JP(B)H
|
12.6 | JP(B)C
|
3
2
8.3 | JP(B)H
|
11.4 | JP(B)C
|
3
3
8.3 | JP(A)H
|
9.1 | JP(A)C
|
3
3
8.4 | JP(A)H
|
10.2 | JP(A)C
|
3
10.8 | JP(B)C
|
3
10.0 | JP(B)C
|
2
dppe, ethane
bridge
close to 31P(B) 16.0 | JCP
|
1
27.0 | JCP
|
2
dppe, ethane
bridge
close to 31P(B)
close to 31P(B)
close to 31P(A)
close to 31P(A)
dppe, 31P(A)
dppe, 31P(B)
29.5 (dd)
203.8 (dd)
201.3 (t)
close to 31P(A) 20.8 | JCP
|
1
28.3 | JCP
|
2
CO
trans to 31P(A) 7.1 | JP(B)C
|
2
79.5 | JP(A)C
|
2
2
31P
12.7 | JPP
|
CO cis to
4.7 | JPC
|
2
64.5 (d)
12.7 | JPP
|
both 31
P
b. NMR Data for Complexes 2-6 in Toluene-d8
coupling
coupling
complex,
temp
δ
constant (Hz)
assignment
complex,
temp
δ
constant (Hz)
assignment
nucleus (multiplicity) assignment
nucleus (multiplicity) assignment
Ru(CO)2(H)2- 1H-NMR
7.64 (m) C6H5 (o-H)
Ru(CO)2(H)2- 1H-NMR
7.66 (m) C6H5 (o-H)
(PMe2Ph)2
cct-2
296 K
(PMe3)
(PMe2Ph)
cct-5
2
1.60 (t)
-7.13 (t)
12.0 (s)
CH3
Ru-H
PMe2Ph
6.5 | JPH + 4JPH
|
2
2
4
26.0 | JPH
|
1.58 (dd) P(CH3)2Ph 8.5 | JPH|, 2.0 | JPH
|
2
4
31P-NMR
298 K
1.22 (dd) P(CH3)3
9.3 | JPH|, 2.0 | JPH
|
2
2
2
ccc-2
1H-NMR -7.08 (ddd) Ru-H
27 | JPH|, 21 | JPH|,
-5.9 2JHH
-7.35 (t)
Ru-H
PMe2Ph
PMe3
27 | JPH
|
31P-NMR 11.3 (d)
222 | JPP
|
2
2
2
2
333 K
-7.66 (ddd) Ru-H
73 | JPH|, 32 | JPH|,
-1.5 (d)
222 | JPP
|
-5.9 2JHH
ccc-5
isomer 1
1H-NMR -7.33 (ddd) Ru-H
-7.87 (ddd) Ru-H
-5.7 2JHH
31P-NMR
9.1 (d)
0.2 (d)
1.24 (t)
PMe2Ph
PMe2Ph
CH3
28 | JPP
|
74.3 | JPH|, 32.5 | JPH|,
2
2
2
2
28 | JPP
|
-5.7 2JHH
Ru(CO)2(H)2- 1H-NMR
(PMe3)2
cct-3
296 K
7.0 | JPH + 4JPH
|
333 K
31P-NMR
1.4 (d)
-5.6 (d)
PMe2Ph
PMe3
22 | JPP
|
2
2
2
22 | JPP
|
2
-7.48 (t)
Ru-H
PMe3
CO
26 | JPH
|
ccc-5
isomer 2
1H-NMR -7.41 (ddd) Ru-H
-7.78 (ddd) Ru-H
-6.5 2JHH
31P-NMR -2.2 (s)
13C-NMR 202.4 (t)
24.8 (t)
71.8 | JPH|, 33.9 | JPH|,
2
2
2
9.3 | JCP
|
-6.5 2JHH
2
2
CH3
16.4 | JCP + 4JCP
|
333 K
31P-NMR
9.9 (d)
-15.4 (d)
PMe2Ph
PMe3
28 | JPP
|
2
2
2
ccc-3
1H-NMR -7.51 (ddd) Ru-H
30 | JPH|, 22 | JPH|,
-6.0 2JHH
28 | JPP
|
Ru(CO)(H)2- 1H-NMR
(PMe3)3
6
1.27 (m) [P(CH3)3]2
2
2
333 K
-7.77 (ddd) Ru-H
73 | JPH|, 33 | JPH|,
-6.0 2JHH
1.34 (m) P(CH3)3
-7.77 (qdd) Ru-H
2
2
31P-NMR -6.8 (d)
-16.4 (d)
PMe3
PMe3
CH3
Ru-H
CH3
20 | JPP
|
333 K
29.1 | JPH|, -7.3 2JHH
,
2
2
20 | JPP
|
20.0 | JCH
|
2
2
cct-4
1H-NMR
1.46 (s)
-7.42 (s)
1.20 (s)
-8.44 (dtd) Ru-H
79.5 | JPH|, 30.1 | JPH|,
-7.3 2JHH
298 K
Ru(CO)2(H)2- 1H-NMR
31P-NMR
0.3 (d)
-10.8 (t)
PMe3
PMe3
(AsMe2Ph)2
ccc-4
298 K
1.17 (s)
1.16 (s)
CH3
CH3
1.13 (s)
-6.86 (d)
-8.70 (d)
CH3
Ru-H
Ru-H
-5.8 2JHH
-5.8 2JHH
a Resonance belongs to X[AE]2M spin system.
involving simultaneous interchange of hydride and phosphine
positions. Phosphine interchange has also been monitored
directly via heteronuclear 1-D 31P-EXSY spectra,25 and at
306.5 K the rate of this process was determined to be 1.34
s-1. Similarly, 13C{1H}-EXSY spectra demonstrated that the
two carbonyl ligands interchanged positions at a rate of 1.28
s-1 at 306.5 K. Therefore, it was clear that the processes of
hydride, carbonyl, and phosphine interchange had essentially
(25) Stott, K.; Stonehouse, J.; Keeler, J.; Hwang, T. L.; Shaka, A. J. J.
Am. Chem. Soc. 1995, 117, 4199.
2964 Inorganic Chemistry, Vol. 41, No. 11, 2002

Ruthenium Dihydride Complexes
Table 2. Rate Constants and Activation Parameters (error) for Mutual Hydride Exchange in ccc-1-4
complex
ligand
dppe
process
H T H
temp, K
rate, s-1
∆G350q, kJ mol-1
∆Hq, kJ mol-1
∆Sq, J K-1 mol-1
1
327.7
306.4
306.4
306.4
306.4
350.3
350.2
351.1
8.74
1.20
1.28
1.34
1.28
0.77
0.96
1.42
73.7 (0.7)
85.5 (2)
34 (7)
H T H
Ph(A) T Ph(C)
31Pa T 31Pb
13COa T 13COb
H T H
H T H
H T H
2
3
4
PMe2Ph
PMe3
AsMe2Ph
86.8 (1)
86.7 (1)
85.8 (0.5)
76 (10)
68 (6)
90 (5)
-30 (28)
-80 (16)
11 (11)
the same rate constants and were connected by a common
pathway. The exact details associated with this type of
analysis can be found in the Supporting Information.
To determine the kinetic isotope effect for this process,
the interchanges of rings A and C and of the two phosphorus
atoms were monitored for both Ru(CO)2(D)2(dppe), 1-D2,
and Ru(CO)2H2(dppe) at 310 K. Rate constants of 1.78
(A f C, which is equivalent to C f A) and 1.72 s-1 (P f
P) were obtained for Ru(CO)2(D)2(dppe), while Ru(CO)2-
(H)2(dppe) yielded values of 1.84 (A f C) and 1.80 s-1
(P f P). Using the average of the two rate constants for
Ru(CO)2(D)2(dppe) and the two rate constants for Ru(CO)2-
(H2)(dppe), the ratio kHH/kDD was determined to be 1.04.
Reaction of Ru(CO)2(H)2(dppe), 1, with p-H2: Evidence
for H2 Reductive Elimination. When a d6-benzene solution
of 1 was treated at 323 K with parahydrogen (3 atm pressure),
PHIP-enhanced hydride resonances at δ -6.33 and -7.55
were observed in the 1H-NMR spectrum. The enhancement
decayed over a period of 10 min as the p-H2 enrichment
was depleted, with the reappearance of the regular nonen-
hanced signals of 1. This suggested that 1 underwent
reductive elimination of H2 at a significant rate at 323 K.
Furthermore, the fact that p-H2-enhanced signals were
observed required H2 addition to proceed in a spin-correlated
process via a diamagnetic intermediate. No enhancement
would be expected if the intermediate had a triplet electronic
spin state, since rapid quenching of the spin coherence would
result due to strong interaction with the magnetic anisotropy
associated with the presence of unpaired electrons. The
validity of this deduction was tested by irradiating C6D6
solutions of Ru(CO)3dppe and 13CO-labeled Fe(CO)5 under
Figure 5. 1H-NMR spectra obtained with p-H2 in C6D5CD3 showing the
hydride region only. The antiphase components arise in transitions involving
protons that originate from p-H2: (a) spectrum of 2 at 333 K; (b) spectrum
of 3 at 333 K; (c) spectrum of 13CO-labeled 3 at 333 K; (d) spectrum of
Ru(CO)(H)2(PMe3)3 observed via the warming of 3 with PMe3 in the
presence of p-H2 at 333 K.
2
hydride resonance (δ -7.13, | JPH| ) 26 Hz) was initially
visible. The associated 31P-NMR signal appeared at δ 12.0.
As expected, these signals matched those previously reported
for the cct-P isomer of 2.9 When the data accumulation was
continued for a period of several hours, however, an
additional pair of hydride resonances was detected at δ -7.16
and -7.77. These signals were assigned to the ccc isomer
of 2. Integration of the hydride resonances gave a value of
24 for the equilibrium constant K ) [cct-P]/[ccc]. When a
d8-toluene solution of 2 was exposed to p-H2 (3 atm) at 333
K, the two hydride resonances for the ccc isomer, now at δ
-7.08 and -7.66, were detected as enhanced signals (Figure
5a). A 1H-1H-COSY experiment confirmed their connectiv-
ity. The resonance at δ -7.08 showed splitting by two
inequivalent 31P nuclei of 27 and 21 Hz, values characteristic
of a hydride ligand cis to both 31P nuclei. In contrast, the
hydride resonance at δ -7.66 showed splittings of 73 and
32 Hz, indicating that the hydride ligand associated with this
resonance was trans to one 31P nucleus and cis to the other.
The spectral characteristics of 2 are summarized in Table
1b. Comparison of the spectral characteristics of 1 and 2
reveals fairly similar chemical shifts (δ -7.55 and -7.16,
1
p-H2 directly in the NMR probe while recording the H-
NMR spectra. The sample containing Ru(CO)3dppe yielded
p-H2-enhanced signals for the hydride resonances of 1, while
the signal for Fe(13CO)4(H)2 showed no enhancement.
Reductive elimination of H2 from 1 was monitored over
the temperature range 343-373 K via the mapping of
hydride/dihydrogen connections using the 1-D nOe sequence
of Keeler.25 Subsequent analysis provided values of ∆Hq )
97 ( 10 kJ mol-1 and ∆Sq ) 2 ( 2 J K-1 mol-1 for this
reaction. At 369 K the rate of reductive elimination of H2
from 1 was 0.253 s-1, while the corresponding rate of
elimination of D2 from 1-D2 was 0.56 s-1. The ratio kHH/kDD
for reductive elimination of hydrogen from 1 is, therefore,
approximately 0.5.
NMR Examination of Ru(CO)2(H)2(PMe2Ph)2, 2. When
1
the H-NMR spectrum of 2 in d8-toluene was recorded at
296 K under an atmosphere of normal hydrogen, only one
Inorganic Chemistry, Vol. 41, No. 11, 2002 2965

Schott et al.
Table 3. Rate Constants and Activation Parameters (error) for Isomer Interchange in 2-4 and Reductive Elimination of H2 for 1
complex
ligand
process
ccc f cct-P
ccc f cct-P
ccc f cct-As
cct-As f ccc
reductive elimination
temp, K
rate, s-1
∆G350q, kJ mol-1
∆Hq, kJ mol-1
∆Sq, J K-1 mol-1
2
3
4
4
1
PMe2Ph
PMe3
AsMe2Ph
AsMe2Ph
dppe
350.3
350.2
351.1
351.1
350
0.70
0.32
0.39
0.39
0.022
87.4 (1)
89.4 (1)
86.8 (1)
85.8 (1)
97.7 (1)
112 (10)
107 (6)
92 (4)
88 (5)
97 (10)
70 (25)
50 (18)
15 (10)
-3 (13)
2 (2)
respectively) for the hydride ligand trans to CO but a bigger
gap (δ -6.32 and -7.77) between the values for the hydride
ligand trans to phosphorus.
drogen, it was possible to obtain EXSY spectra of the ccc
isomers of 2 and 3. For each complex, the two hydride
resonances for the ccc isomer showed cross-peaks to one
another and to the hydride resonances for the cct-P isomers.
However, since the ccc isomer in these complexes could only
be detected by virtue of the parahydrogen enhancement, the
signals for the cct-P isomer failed in each case to yield cross-
peaks to hydrides of the ccc isomer, and consequently, the
spectra were not symmetrical about the diagonal. Rate
constants for the interchange of hydride ligands in the ccc
isomers of 2 and 3 and for the conversion of the ccc isomers
to the corresponding cct-P isomers are given in Table 2. The
associated enthalpies and entropies of activation, determined
from 16 rate constants for 2 and 25 values for 3 between
350 and 366 K, are listed in Table 3. Interchange of hydride
ligands in ccc-3 was also monitored by means of 31P-coupled
1H-1H-EXSY spectra (as described above for complex 1,
Figure 4 shows a typical spectrum obtained for 1). The
pattern of connections again matched that required for an
intra-molecular process in which exchanges of the two
hydride ligands and of the two phosphorus ligands occurred
concurrently. The exact details of the analysis of the spectra
can be found in the Supporting Information.
NMR Examination of Ru(CO)2(H)2(PMe3)2, 3. The 1H-
NMR spectrum of 3 in a d6-benzene solution at 296 K
indicated the presence of only one isomer which was
2
characterized by a triplet resonance at δ -7.48 (| JPH| ) 26
Hz) for the hydride ligands and a virtual triplet26 at δ 1.24
for the PMe3 ligand. The 31P-NMR spectrum of 3 consisted
of a singlet at δ -2.2, and the 13C{1H}-NMR spectrum
contained only one resonance in the region characteristic of
2
carbonyl ligands, a triplet (| JCP| ) 9.3 Hz) at δ 202.4.27
Even after overnight accumulation of the 1H-NMR spectra,
no signal attributable to another isomer of 3 could be
detected. However, when the p-H2 experiment described
above was repeated for 3 (again at 333 K), two new hydride
signals, with splitting patterns similar to those for the hydride
resonances of ccc-2, were seen at δ -7.51 and -7.77 (Figure
5b) and were assigned to the ccc isomer of 3. The spectral
characteristics of ccc-3 are listed in Table 1b. Since the two
hydride signals of this isomer of Ru(CO)2(H)2(PMe3)2 were
only detected at and above 333 K, we concluded that
reductive elimination of H2 from Ru(CO)2(H)2(PMe3)2 was
relatively slow at lower temperatures. In the p-H2 experiment,
it was also noticed that the hydride resonance for the cct-P
isomer was distorted, with both emission and absorption
To determine whether ligand exchange within the ccc
isomer involved CO dissociation, a sample of 3 was dissolved
in CO-saturated d8-toluene before being treated with p-H2
in the normal way. The ratios of the exchange-peak to
diagonal-peak volumes in the associated EXSY spectrum
(recorded at 358.6 K) were not altered by the presence of
the CO. This ruled out CO loss at any stage up to, and
including, the rate-limiting step. When the process was
repeated with PMe3 instead of CO, the aim of the experiment
was defeated by the appearance of two new and enhanced
hydride resonances at δ -7.77 and -8.44 (Figure 4d). The
1
features clearly visible in the 31P-decoupled H spectrum,
despite the fact that the two hydride ligands in this isomer
are equivalent.28 An elegant explanation of this effect has
been reported by Aime et al.,29 who studied the addition of
p-H2 to the cluster Os3(CO)10(NCCH3)2. Enhanced hydride
signals were observed for Os3(CO)10(H)2, a species which,
like cct-P-3, contains magnetically equivalent hydrides and,
therefore, should not show enhancement. This reaction,
however, proceeds by way of the unsymmetrical Os3-
(CO)10H(µ-H)(NCCH3), and the enhancement is passed on
from the intermediate to the product. Similarly, the enhance-
ment of the hydride signal for cct-P-3 is partially due to the
fact that it can be formed via ccc-3. A similar parahydrogen
experiment with 13CO-enriched Ru(CO)2(H)2(PMe3)2 showed
a more substantial enhancement of the hydride resonances
for the cct-P isomer, as shown in Figure 5c. The additional
enhancement is due to the formation of a second-order spin
system containing the hydride ligands.
2
former was a quartet of antiphase doublets (| JPH| ) 29.1
Hz, 2JHH ) -7.3 Hz), and the latter was a doublet of triplets
2
2
of antiphase doublets (| JPH| ) 79.5 Hz, | JPH| ) 30.1 Hz,
2
and JHH ) -7.3 Hz). Evidently the two hydride ligands
were in the same complex, the former trans to CO and the
latter trans to PMe3. When the corresponding 13CO-labeled
complex was examined, the resonance at δ -7.77 exhibited
an additional 20.0 Hz doublet splitting, confirming that it
represented the hydride ligand trans to CO. The complex
was identified as mer-cis Ru(CO)(H)2(PMe3)3, 6.30
Exchange Studies Facilitated by Parahydrogen. Thanks
to the signal enhancement provided by the use of parahy-
Unambiguous confirmation of the retention of the phos-
phine ligands within the metal’s coordination sphere during
hydride interchange came from the 1H-1H-EXSY observa-
tion of a mixture of Ru(CO)2(H)2(PMe3)(PMe2Ph) 5, Ru-
(CO)2(H)2(PMe3)2 3, and Ru(CO)2(H)2(PMe2Ph)2 2. No
(26) Jenkins, J. M.; Shaw, B. L. Proc. Chem. Soc., London 1963, 279.
(27) Gill, D. F.; Mann, B. E.; Shaw, B. L. J. Chem. Soc., Dalton Trans.
1973, 311.
(28) We note that similar, weaker, distortions were visible in the corre-
sponding resonances of 2.
(30) Kohlmann, W.; Werner, H. Z. Naturfosch., B: Chem. Sci. 1993, 48,
(29) Aime, S.; Gobetto, R.; Canet, D. J. Am. Chem. Soc. 1998, 120, 6770.
1499.
2966 Inorganic Chemistry, Vol. 41, No. 11, 2002

Ruthenium Dihydride Complexes
Scheme 2
Figure 6. Eyring plot for ccc interchange within 4. Experimental points
indicated with 9 correspond to measurements using normal H2 and the least-
mean squares simulation, 2 correspond to points determined using p-H2,
and b correspond to points determined using normal H2 and the full matrix
analysis.
cross-peaks connecting the different complexes were ob-
served, whereas the expected connections for hydride
exchange within the ccc isomers and for the interconversion
of ccc and cct-P isomers of the individual complexes were
visible. For cct-P-5, the hydride resonance was a triplet
above 323 K, the hydride resonances for ccc-4 at δ -6.86
and -8.70 showed emission-absorption characteristics in-
dicating that the hydride ligands were exchanging with free
p-H2. At 343 K, the enhanced antiphase character indicated
a 1035-fold increase in signal strength for the ccc-4 hydride
resonances over those observed in the absence of p-H2. Rate
constants for the hydride exchange within the ccc isomer
and interconversion of ccc and cct-As isomers were deter-
mined at temperatures between 323 and 369 K using both
normal and p-H2 according to the methods outlined in the
Experimental Section. The values of ∆Hq and ∆Sq for these
processes were calculated and are summarized in Tables 2
and 3. Figure 6 illustrates the Eyring plot for the exchange
of hydride ligands within ccc-4, showing the excellent
agreement between data obtained by the different methods.
2
(| JPcisH| ) 26.5 Hz) at δ -7.35, a chemical shift between
those for cct-P-3 and cct-P-2. The 31P{1H}-NMR spectrum
of cct-P-5 consisted of doublet resonances at δ 11.3 and
2
-1.5, both with | JPP| ) 222 Hz which is indicative of
mutually trans phosphine ligands.
Characterization and Dynamic Behavior of Ru(CO)2-
(H)2(AsMe2Ph)2, 4. When a 1H-NMR spectrum of Ru(CO)2-
(H)2(AsMe2Ph)2, 4, was recorded in d8-toluene at 296 K,
three hydride resonances were detected at δ -7.60, -6.86,
and -8.70. These have been shown to arise from the cct-As
and ccc isomers of 4; both have been previously character-
ized.9 It has also been reported previously that these two
isomers exist in equilibrium in solution, and that the cct-As
isomer reacts preferentially with the alkyne MeO2CCtCCO2-
Me; subsequent reequilibration of the ccc and cct-As isomers
takes around 15 min at room temperature.11 We examined
this system by both 1- and 2-D EXSY methods in conjunc-
tion with the use of both normal and p-H2. No NMR evidence
was obtained for isomer interconversion at room temperature,
confirming the slowness of the reequilibration at this
temperature. However, a new hydride resonance was detected
Discussion
Our studies of the three complexes of Ru(CO)2(H)2L2 (3,
L ) PMe3; 2, L ) PMe2Ph; 4, L ) AsMe2Ph) have revealed
a strong dependence of the relative stabilities of the cct-L
and ccc isomers on the nature of the ligand. For L ) PMe3,
the cct-L isomer is the more stable of the two by such a
wide margin that the ccc isomer cannot be detected without
the use of p-H2. The gap narrows for L ) PMe2Ph, where
both isomers can be detected without resorting to p-H2. At
295 K the [cct-L]:[ccc] ratio is 24, corresponding to a free
energy difference of 7.8 kJ mol-1 in favor of the cct-L form.
This difference is further reduced to only 0.7 kJ mol-1 for
AsMe2Ph, but here matters are complicated by the presence
of the third (cct-CO) isomer. All three isomers can be
detected at 295 K in a cct-L:ccc:cct-CO ratio of 56.4:42.7:
0.9. Evidently, the change of donor atom in L has a major
effect on the isomer equilibrium. This can be readily
rationalized since the donor ability of the phosphines
described here can be ranked as PMe3 > PMe2Ph > dppe
on the basis of Tolman’s study.31 The greater the σ-donor
ability of the phosphine, the greater the preference of the
hydride (which can only make a σ-bond with a metal) for a
position trans to the π-acceptor CO rather than trans to
phosphine. Thanks to this electronic effect, the cct-L isomer,
1
at δ -9.35 in the H-NMR spectrum, and exchange peaks
were observed between this resonance and the cct-As isomer
of 4 at 318 K. The species responsible for the additional
hydride resonance is believed to be the cct-CO isomer of 4.
Analysis of the exchange-peak intensity as a function of
mixing time yielded values for the rate constants for cct-
As-4 f cct-CO-4 and the reverse process of 0.02 and 1.31
s-1, respectively. At this temperature the solution was found
to contain the cct-As and ccc and cct-CO isomers in a 53.9:
45.2:0.9 ratio. At 327 K, EXSY analysis indicated that all
three isomers of 4 were interconverting. The rate constants
for the interconversion between isomers at 327 K are given
in Scheme 2 (estimated error (1 in last significant figure).
Detailed study of the isomerization process was hampered
by the fact that at temperatures above 327 K, 4 begins to
slowly decompose. Nonetheless, it was possible to obtain
values of ∆HL and ∆SL for the conversion of cct-As into
ccc as 1.6 ( 0.7 kJ mol-1 and 4.1 ( 2.0 J K-1 mol-1,
respectively. When 4 was treated with p-H2 at temperatures
(31) Tolman, C. A. Chem. ReV. 1977, 77, 313.
Inorganic Chemistry, Vol. 41, No. 11, 2002 2967

Schott et al.
in which both hydrides are trans to CO, is increasingly
favored. Steric effects can be deduced to play a less
significant role in controlling the equilibrium position, since
increasing the size of either the donor atom or the substituents
would be expected to favor the cct-L isomer. This is the
opposite of what is observed experimentally.
formation of an intermediate formyl complex by combination
of carbonyl and hydride ligands; however, this appears to
be uncommon for ruthenium,37-39 and we saw no evidence
in our p-H2 studies for the formation of a species containing
a formyl ligand. We believe that the most likely pathway
involves the reversible formation of a transition state
featuring significant shortening of the H-H distance or a
thermally accessible η2-H2 intermediate with a trigonal
bipyramidal shape. Species containing η2-bonded H2 ligands
have featured extensively in the literature as intermediates
in the reverse process of H2 addition to transition metal
centers to give dihydride complexes.40 For the dihydride Re-
(CO)(H)2(PR3)2(NO), an intramolecular exchange mechanism
involving a trigonal bipyramidal η2-H2 complex has been
proposed.41 The observation that the activation enthalpies
decrease with increased electron donating ability of L,
requires special comment, therefore, because such a trend
would be expected to parallel an increase in the stability of
the dihydride form over its η2 counterpart. Such a constraint
can, however, be readily rationalized as a ground-state effect
where the inherent instability of the ccc form or the
associated lengthening of the Ru-H distance due to the trans
influence of the phosphine ligands controls the overall
activation enthalpy. The process of hydride exchange is,
however, further complicated by the fact that in 1 it is
accompanied by simultaneous exchange of the L and CO
pairs. Therefore, rotation about the resultant H-H bond must
be restricted. A barrier to rotation about the M-H2 axis can
arise from electronic interactions between the η2-bonded H2
and the transition metal center and nonbonded interactions
between the hydrogen atoms and the ancillary ligands.
Barriers to rotation for octahedral d6 complexes, such as
W(CO)3(PR3)2(η2-H2), are generally low because the dπ-
σ* interaction does not change significantly with rotation.
However, higher barriers have been observed for complexes
such as OsX{NHdC(Ph)C6H4}(H)2(PiPr3)2 (X ) Cl, Br, and
I), where the formation of an elongated dihydrogen ligand
has been proposed to account for the barrier.42 The direct
correlation between increased rotational barriers and longer
H-H distance has been discussed for [Os(H2)Cl(H2PCH2-
CH2PH2)2]+, where an elongated H-H bond has been
predicted.43 Our observation that the ratio kHH/kDD was 1.04
for the dppe complex could be interpreted to indicate that
the exchange in 1 proceeds with limited H-H bonding in
reaching the transition state (see later). If this situation
corresponds to an elongated H-H bond, the degree of bond
making and breaking during the approach from the ground
state must be about equal.
The ccc isomers of 2-4 and ccc-Ru(CO)2(H)2(dppe), 1,
all undergo intramolecular exchange of their two hydride
ligands, with 1 having the fastest rate; Table 2 contains the
corresponding values of ∆Gq350, ∆Hq, and ∆Sq. The values
of ∆Hq fall across the series AsMe2Ph > dppe > PMe2Ph
> PMe3 from 90 to 85.5 to 76 to 68 kJ mol-1. This decrease
signifies that the activation enthalpy is reduced by an increase
in the electron density at the metal in accordance with the
electron donating ability of L (see later). The corresponding
values of ∆Sq at 11 (11), 34 (7), -30 (28), and -80 (16) J
K-1 mol-1, respectively, reveal that the enhanced rate of
exchange in the dppe system arises from a favorable entropy
change on moving to the transition state. This difference
might, therefore, signify that the ethane bridge connecting
the phosphorus centers in 1 modifies the ground-state
geometry in a way which reduces the activation barrier to
exchange. This theory receives support from the fact that
the couplings between the hydride ligand that is trans to L
and the 31P nucleus that is cis to it vary from 21.7 Hz in 1
to 32 and 33 Hz in 2 and 3, respectively. A narrowing of
the P-Ru-P bond angle to 84.7°, as a consequence of the
ethane bridge connecting the two phosphorus centers of the
dppe ligand in the related complex Ru(dppe)(CO)2(OSO2-
CF3)2, has been reported.32 In contrast, the corresponding
P-Ru-P bond angles in mer-Ru(CO)(H)2(PMe2Ph)3 are
100.8° and 98.5°.33 These data, therefore, support conclu-
sively that dppe complex 1 undergoes the most rapid
exchange as a direct consequence of the steric influence of
the bulky chelating phosphine. However, despite the fairly
wide variation, none of these values of ∆Sq are sufficiently
positive to suggest a dissociative mechanism for the exchange
process.
In addition to the ∆Sq changes, evidence was presented
earlier which demonstrated that the exchange does not
involve the loss of CO or ligand L. Various pathways have
been suggested for intramolecular exchange of hydride
ligands. Rearrangement by a tunneling mechanism,34-36
originally described by Muetterties for complexes Fe(H)2L4
and Ru(H)2L4 (where L is a phosphorus ligand), and the
related trigonal twist mechanism described by Mann 24 were
both candidates; however, both can be ruled out because the
exchanges of all three pairs of ligands (hydrides, carbonyls,
and the ligands L) occur simultaneously and have the same
rate constant. Another potential route would involve the
(37) Pearson, R. G.; Walker, H. W.; Mauermann, H.; Ford, P. C. Inorg.
Chem. 1981, 20, 2743.
(38) Brougham, D. F.; Brown, D. A.; Fitzpatrick, N. J.; Glass, W. K.
Organometallics 1995, 14, 151.
(39) Dedieu, A.; Nakamura, S. J. Organomet. Chem. 1984, 260, C63.
(40) Heinekey, D. M.; Oldham, W. J. Chem. ReV. 1993, 93, 913. Maseras,
F.; Lledos, A.; Clot, E.; Eisenstein, O. Chem. ReV. 2000, 100, 601.
(41) Bakhmutov, V.; Burgi, T.; Burger, P.; Ruppli, U.; Berke, H.
Organometallics 1994, 13, 4203.
(32) Mahon, M. F.; Whittlesey, M. K.; Wood, P. T. Organometallics 1999,
18, 4068.
(33) Bruno, J. W.; Huffman, J. C.; Caulton, K. G. Inorg. Chim. Acta 1984,
89, 167.
(34) Meakin, P.; Muetterties, E. L.; Jesson, J. P. J. Am. Chem. Soc. 1973,
95, 75.
(35) Meakin, P.; Muetterties, E. L.; Tebbe, F. N.; Jesson, J. P. J. Am.
Chem. Soc. 1971, 93, 4701.
(36) Jesson, J. P.; Muetterties, E. L.; Meakin, P. J. Am. Chem. Soc. 1971,
93, 5261.
(42) Barea, G.; Esteruelas, M. A.; Lledos, A.; Lopez, A.; Onate, E.; Tolosa,
J. I. Organometallics 1998, 17, 4065.
(43) Gelabert, R.; Moreno, M.; Lluch, J. M.; Lledos, A. J. Am. Chem.
Soc. 1998, 120, 8168.
2968 Inorganic Chemistry, Vol. 41, No. 11, 2002

Ruthenium Dihydride Complexes
Scheme 3
Interconversion of the ccc and the cct-L isomers of
complexes 2-4 was also observed, but on a slower time scale
than ligand interchange within the ccc isomers. This is
consistent with the higher ∆Hq values for this process: 2,
112 (10) kJ mol-1; 3, 107 (6) kJ mol-1; and 4, 92 (4) kJ
mol-1. Unfortunately, the large errors for the ∆Sq values
make them unreliable as an indicator of the molecularity of
the inter-isomer rearrangement process, but there is other
evidence to rule out the loss of a carbonyl ligand or a
phosphine ligand. Structure B in Scheme 3 would allow the
ccc isomer to equilibrate with the cct-L isomer if the η2-H2
ligand completed an anticlockwise rotation to place the H-H
bond along the Ru-CO bond before reestablishing the two
Ru-H bonds. Neither structure A nor B, however, allows
access to a form with trans CO’s. Significantly, access to
structure C would enable interconversion of the ccc isomer
into both the cct-CO and cct-L without hydride interchange
within the ccc form. The rates and activation parameters
measured for these processes indicate that these three
exchange pathways have substantially different barriers. We
also note that structure B would serve as an intermediate
(or transition state) that, via clockwise rotation of the η2-H2
ligand and reestablishment of the two Ru-H bonds, would
lead to interchange of the hydrides, carbonyls, and phos-
phines within the ccc isomer. We suggest for Ru(CO)2(H)2L2,
containing 2 π-acceptor ligands, that the trigonal bipyramidal
form, corresponding to situation A where the η2-H2 ligand
is located in the equatorial plane, accounts for the lower-
energy pathway, while access to the higher-energy form B
accounts for the slower ccc-cct isomer interconversion.
Values of ∆Hq ) 97 ( 10 kJ mol-1 and ∆Sq ) 2 ( 2 J
K-1 mol-1 obtained for the reductive elimination of H2 from
1 indicated that there is a substantially higher barrier to
reductive elimination. The observed kinetic isotope effect,
kHH/kDD, for this process of 0.5 requires a substantial amount
of dihydrogen bond formation in the associated transition
state; this contrasts with the evidence that the amount of
H-H bond formation in the transition state for a simple
hydride ligand exchange is relatively small. The value of
0.5 also contrasts with the kinetic isotope effects for the
reductive elimination of hydrogen from the eight-coordinate
W(H)2(I)2(PMe3)4, which has been estimated to be 2 at 333,47
and from Ir(H2)(H)2Cl(PBut2Me)2, which is 1.8 at 293 K.48
However, H2 elimination from the classical dihydride
(fulvalene)Cr2(CO)6(H)2 has been found to be 1.1 times
slower than the corresponding elimination of D2.49 Given
the enhanced stability of the η2-D2 form over the corre-
sponding η2-H2 isotopomer reported for W(CO)3(PCy3)2(η2-
H2), the observation of an inverse kinetic isotope effect for
reductive elimination from 1 suggests that access to the more
stable η2-D2 state surpasses the increased stability associated
with M-D versus M-H bonds in the initial dihydride
complexes.
Furthermore, there is evidence that, since H2 is a single-
face π-acceptor, its rotation couples with the movement of
other single-face π-acceptor ligands during the isomerization
process. The complexes RuH2(H2)2(PH3)2 and [Os(NH3)4-
(H2)(CH2)]+ provide examples of such coupled rotations.44
Such a linking has also been observed for the more closely
related iron(0) system Fe(CO)4(alkene), where NMR obser-
vations require synchronous Berry pseudorotation (exchang-
ing the carbonyl ligands) and alkene rotation.45 The similarity
in bonding between the single-face π-acceptor ligand H2 and
an alkene is, therefore, especially relevant in this regard.
Albright et al. have modeled the hydride exchange pathway
in the complexes Ru(H)2(PMe3)4 and Fe(H)2(CO)4.46 In their
scheme, structures containing η2-H2 ligands with both square
pyramidal and trigonal bipyramidal ligand arrangements are
considered as possible transition states for the interchange
process. Therefore, it would seem sensible to suggest that
the three structures containing η2-H2 ligands shown in
Scheme 3 might be featured as either the transition state for
hydride exchange in Ru(CO)2(H)2L2 or intermediates in-
volved in the pathway. Their studies reveal that structure A
in Scheme 3 is likely to be higher in energy than either B or
C when the four additional ligands are strong σ-donors, but
that this difference is reduced by the addition of π-acceptor
carbonyls so that in Fe(CO)4(H)2 form A yields the most
stable transition state.
If it is assumed that the η2-H2 ligand occupies one of the
equatorial sites, as shown in structure A, a Berry pseudoro-
tation achieves the required interchange of both the pair of
carbonyl ligands and the pair of ligands L for the cis-cis-cis
isomer (Scheme 4). Since the exchanges of hydride ligands,
carbonyl ligands, phosphorus atoms, and the pairs of phenyl
rings within the dppe ligand all occur at the same rate for 1,
the Berry pseudorotation must be linked to the rotation of
the dihydrogen ligand. Such a situation is well-known for
Fe(CO)4(alkene) and has been predicted for Fe(CO)4(H)2.45,46
Interestingly, the direction of the rotation of the H2 ligand
must also be controlled, because only one direction (that
shown in Scheme 4) achieves interchange of the identities
of the hydride ligands. A random choice of direction would
result in a hydride ligand exchange rate only half that of the
carbonyl ligands and phosphorus centers. It seems likely that
the interactions between the H2 ligand with the substituents
on the L ligands determine the preferred direction of rotation.
(44) (a) Rodrigues, V.; Sabo-Etienne, S.; Chaudret, B.; Thorburn, J.; Ulrich,
S.; Limbach, H. H.; Eckert, J.; Barthelat, J.-C.; Hussein, K.; Marsden,
C. J. Inorg. Chem. 1998, 37, 3475. (b) Jarid, A.; Lledos, A.;
Lauvergnat, D.; Jean, Y. New J. Chem. 1997, 21, 953.
(45) Kreiter, C. G.; Lang, M. J. Organomet. Chem. 1973, 55, C27.
(46) Soubra, C.; Oishi, Y.; Albright, T. A.; Fujimoto, H. Inorg. Chem.
2001, 40, 620.
(47) Rabinovich, D.; Parkin, G. J. Am. Chem. Soc. 1993, 115, 353.
(48) Hauger, B. E.; Gusev, D.; Caulton, K. G. J. Am. Chem. Soc. 1994,
116, 208.
(49) Vollhardt, K. P. C.; Cammack, J. K.; Matzger, A. J.; Bauer, A.; Capps,
K. B.; Hoff, C. D. Inorg. Chem. 1999, 38, 2624.
Inorganic Chemistry, Vol. 41, No. 11, 2002 2969

Schott et al.
Scheme 4
These studies also reveal information about the spin states
involve rearrangements of the ligand sphere. Interchange of
the hydride positions within 1, which exists solely as the
ccc isomer, has been shown to be accompanied by synchro-
nized exchange of the two carbonyl ligands and the dppe
phosphorus atoms. We suggest that this process involves the
accessing of a five-coordinate complex containing an η2-H2
ligand which, in view of the small normal kinetic isotope
effect and restricted rotation, must have only a weak H-H
interaction. The associated transition state is most likely to
place the η2-H2 ligand equatorial as shown in structure A in
Scheme 3. Access to this intermediate is facilitated by an
increase in the electron density available to the metal center
in accordance with a ground-state effect where placing the
hydride trans to a good σ-donor is less favored.
In contrast to the dihydride exchange pathway, the kinetic
isotope effect, kHH/kDD, of 0.5 observed for the reductive
elimination of dihydrogen from Ru(CO)2(H)2dppe suggests
that substantial H-H bond formation occurs before the
dihydrogen is released. Our studies also indicate that the key
intermediates Ru(CO)2L2 involved in the elimination and
addition of hydrogen must have singlet electron configura-
tions, in agreement with previous studies.11,42,43
of the associated 16-electron intermediates Ru(CO)2L2. In
view of the observation that we can monitor the exchange
process in 1-4 by NMR and that the formation of 1 from
Ru(CO)2(dppe) proceeds with the observation of p-H2-
enhanced NMR signals for 1, we can state unequivocally
that Ru(CO)2(dppe) is diamagnetic in nature. Theoretical and
experimental studies of the reaction of H2 with Fe(CO)4
suggest that the latter has a triplet ground state in agreement
with the failure to obtain p-H2-enhanced signals for Fe-
(13CO)4(H)2 at 296 K.50,51 Singlet ground states have,
however, been suggested for Ru(CO)4,52 Ru(CO)2L253 (L )
PBut2Me), and Ru(PH3)4,14 in agreement with our observation
of p-H2-enhanced ccc-Ru(CO)2(H)2L2. Studies with p-H2,
therefore, provide a ready means of studying spin state effects
in reaction mechanisms.
Conclusions
Our NMR studies of complexes Ru(CO)2(H)2L2 (L )
PMe3, PMe2Ph, and AsMe2Ph) enable us to conclude that
they can have three possible geometries, ccc, cct-L, and cct-
CO. The equilibrium ratio of these three isomers is dramati-
cally influenced by the electronic properties of L, with better
donors favoring the cct-L form. Consequently the ccc form
for L ) PMe3 is only visible when p-H2 is used to amplify
its spectral features. In contrast, when L ) AsMe2Ph, the
ccc and cct-L forms have similar free energies, and a new
cct-CO species becomes visible as a minor isomer.
Acknowledgment. We are grateful to the EPSRC (C.J.S.
and spectrometer), the University of York (D.S., J.P.L.,
M.G.P.), JREI (spectrometer), Bruker UK (spectrometer), and
the Royal Society for financial support. We appreciated the
helpful discussions with Prof. R. N. Perutz, Dr. M. K.
Whittlesey, and Dr. J. M. Lynam. A generous loan of
ruthenium trichloride from Johnson Matthey is also gratefully
acknowledged.
These complexes, and also Ru(CO)2(H)2(dppe), have been
shown to undergo a number of dynamic processes which
(50) Wang, W.; Weitz, E. J. Phys. Chem. A 1997, 101, 2358.
(51) For example see: (a) Poliakoff, M. Chem. Soc ReV. 1978, 7, 527. (b)
Snee, P. T.; Payne, C. K.; Kotz, K. T.; Yang, H.; Harris, C. B. J. Am.
Chem. Soc. 2001, 123, 2255.
(52) Bogdan, P. L.; Witz, E. J. Am. Chem. Soc. 1989, 111, 3163.
(53) Ogasawara, M.; Macgregor, S. A.; Strieb, W. E.; Folting, K.;
Eisenstein, O.; Caulton, K. G. J. Am. Chem. Soc. 1995, 117, 8869;
1996, 118, 10189.
Supporting Information Available: Analysis of hydride
interchange in Ru(CO)2(H)2(dppe) 1 and ccc-Ru(CO)2(H)2(PMe3)2
3. This material is available free of charge via the Internet at
http://pubs.acs.org.
IC011255W
2970 Inorganic Chemistry, Vol. 41, No. 11, 2002