Reversible isomerization of a diphosphine ligand about a triosmium cluster: Synthesis, kinetics, and X-ray structures for the bridging and chelating isomers of Os3(CO)10[(Z)-Ph2PCH=CHPPh 2]

Article abstract of DOI:10.1021/om0505413

Substitution of the MeCN ligands in the activated cluster Os ₃(CO)₁₀(MeCN)₂ (1) by the unsaturated diphosphine ligand (Z)-Ph₂PCH=CHPPh₂ proceeds rapidly at room temperature to furnish the ligand-bridged cluster 1,2-Os ₃(CO)₁₀[(Z)-Ph₂PCH=CHPPh₂] (2b). Heating 2b leads to the formation of the thermodynamically more stable chelating isomer 1,1-Os₃(CO)₁₀-[(Z)-Ph₂PCH=CHPPh ₂] (2c). The molecular structure of each isomer of 2 has been crystallographically determined, and the ³¹P NMR data have been recorded. The kinetics for the ligand isomerization have been investigated by UV-vis and ³¹P NMR spectroscopy in toluene solution over the temperature range of 358-383 K. The reversible nature of the diphosphine isomerization is confirmed by ³¹P NMR measurements, and reported within are the forward (k_-1) and reverse (k_-1) first-order rate constants for the bridge-to-chelate rearrangement. On the basis of the activation parameters and lack of CO inhibition on the reaction rate, a nondissociative, intramolecular mechanism involving the migration of one P group from an adjacent osmium center to the other P-substituted osmium center via a μ₂-bridged phosphine species is presented.

Full text of DOI:10.1021/om0505413

Organometallics 2005, 24, 5431-5439
5431
Reversible Isomerization of a Diphosphine Ligand about
a Triosmium Cluster: Synthesis, Kinetics, and X-ray
Structures for the Bridging and Chelating Isomers of
Os3(CO)10[(Z)-Ph2PCHdCHPPh2]
William H. Watson*
Department of Chemistry, Texas Christian University, Fort Worth, Texas 76129
Guanmin Wu and Michael G. Richmond*
Department of Chemistry, University of North Texas, Denton, Texas 76203
Received June 29, 2005
Substitution of the MeCN ligands in the activated cluster Os3(CO)10(MeCN)2 (1) by the
unsaturated diphosphine ligand (Z)-Ph2PCHdCHPPh2 proceeds rapidly at room temperature
to furnish the ligand-bridged cluster 1,2-Os3(CO)10[(Z)-Ph2PCHdCHPPh2] (2b). Heating 2b
leads to the formation of the thermodynamically more stable chelating isomer 1,1-Os3(CO)10-
[(Z)-Ph2PCHdCHPPh2] (2c). The molecular structure of each isomer of 2 has been
crystallographically determined, and the 31P NMR data have been recorded. The kinetics
for the ligand isomerization have been investigated by UV-vis and 31P NMR spectroscopy
in toluene solution over the temperature range of 358-383 K. The reversible nature of the
diphosphine isomerization is confirmed by 31P NMR measurements, and reported within
are the forward (k1) and reverse (k-1) first-order rate constants for the bridge-to-chelate
rearrangement. On the basis of the activation parameters and lack of CO inhibition on the
reaction rate, a nondissociative, intramolecular mechanism involving the migration of one
P group from an adjacent osmium center to the other P-substituted osmium center via a
µ2-bridged phosphine species is presented.
Introduction
interesting chemistry, due to phosphine-assisted stabi-
lization of cluster unsaturation. For example, the tri-
osmium cluster HOs3(CO)8[PhP(C6H4)CH2PPh2], which
may be isolated from the thermolysis of Os3(CO)10-
(dppm),6 is electronically unsaturated by virtue of its
46-valence-electron count and readily undergoes a wide
variety of ligand addition reactions under mild condi-
tions.7 The inherent unsaturation associated with the
valence-electron count notwithstanding, HOs3(CO)8-
[PhP(C6H4)CH2PPh2] and related clusters also possess
a “masked site” of unsaturation whose origin derives
from a reversible reductive coupling of the bridging
hydride ligand with the µ,η1-carbon atom of the ortho-
metalated aryl ring. Such a rapid and reversible activa-
The reaction of bidentate phosphines with the acti-
vated triosmium clusters Os3(CO)12-n(MeCN)n (n ) 1,
2) and Os3(CO)10L2 (L2 ) 1,5-cod, 1,3-butadiene) has
been investigated over the years with respect to the
bonding mode adopted by the bidentate P ligand upon
coordination.1-3 Simple coordination of diphosphine
ligands at these activated clusters typically proceeds to
furnish a bridged or chelated cluster Os3(CO)10(P-P),
where the P-P ligand is bound to adjacent osmium
centers or attached to a single osmium atom:
(1) (a) Deeming, A. J.; Donovan-Mtunzi, S.; Kabir, S. E. J. Orga-
nomet. Chem. 1984, 276, C65; 1987, 333, 253. (b) Deeming, A. J.; Kabir,
S. E. J. Organomet. Chem. 1988, 340, 359. (c) Deeming, A. J.; Donovan-
Mtunzi, S.; Hardcastle, K. I.; Kabir, S. E.; Henrick, K.; McPartlin, M.
J. Chem. Soc., Dalton Trans. 1988, 579. (d) Deeming, A. J.; Stchedroff,
M. J. Chem. Soc., Dalton Trans. 1998, 3819.
(2) Cartwright, S.; Clucas, J. A.; Dawson, R. H.; Foster, D. F.;
Harding, M. M.; Smith, A. K. J. Organomet. Chem. 1986, 302, 403.
(3) Poe¨, A. J.; Sekhar, V. C. J. Am. Chem. Soc. 1984, 106, 5034.
(4) (a) Shapley, J. R.; Richter, S. I.; Churchill, M. R.; Lashewycz, R.
A. J. Am. Chem. Soc. 1977, 99, 7384. (b) Churchill, M. R.; Lashewycz,
R. A. Inorg. Chem. 1978, 17, 1950. (c) Churchill, M. R.; Lashewycz, R.
A.; Shapley, J. R.; Richter, S. I. Inorg. Chem. 1980, 19, 1277. (d) Puga,
J.; Arce, A.; Braga, D.; Centritto, N.; Grepioni, F.; Castillo, R.; Ascanio,
J. Inorg. Chem. 1987, 26, 867.
While several unequivocal examples exist for the isomer-
ization of a diphosphine ligand at other polynuclear
clusters,4,5 no such studies have been reported for the
family of Os3(CO)10(P-P) clusters. The resulting Os3-
(CO)10(P-P) clusters have been shown to exhibit some
(5) (a) Yang, K.; Smith, J. M.; Bott, S. G.; Richmond, M. G.
Organometallics 1993, 12, 4779. (b) Bott, S. G.; Yang, K.; Richmond,
M. G. J. Organomet. Chem. 2005, 690, 3067.
* To whom correspondence should be addressed. Tel: 817-257-7195
(W.H.W.); 940-565-3548 (M.G.R.). E-mail: w.watson@tcu.edu (W.H.W.);
cobalt@unt.edu (M.G.R.).
(6) Clucas, J. A.; Foster, D. F.; Harding, M. M.; Smith, A. K. J. Chem.
Soc., Chem. Commun. 1984, 949.
10.1021/om0505413 CCC: $30.25 © 2005 American Chemical Society
Publication on Web 09/29/2005

5432 Organometallics, Vol. 24, No. 22, 2005
Watson et al.
The reported infrared data were recorded on a Nicolet 20
SXB FT-IR spectrometer in 0.1 mm amalgamated NaCl cells,
tion of the ancillary diphosphine ligand has the ability
to furnish a transient cluster intermediate capable, in
theory, of multisite substrate activation and cluster-
promoted catalysis.8
1
using PC control and OMNIC software, while the H and 13C
NMR spectra were recorded at 200 and 50 MHz, respectively,
on a Varian Gemini-200 spectrometer. The 31P NMR spectra
were recorded at 121 MHz on a Varian 300-VXR spectrometer.
All 13C and 31P NMR spectra were collected in the proton-
decoupled mode. The reported 31P chemical shift data were
referenced to external H3PO4, whose chemical shift was set
at δ 0.
In comparison to the extensive number of known Os3-
(CO)10(P-P) clusters possessing a diphosphine ligand
with a saturated carbon backbone, fewer paradigms of
diphosphine derivatives that possess an unsaturated
carbon backbone are known,9,10 prompting our explora-
tion of the reaction of Os3(CO)10(MeCN)2 with (Z)-Ph2-
PCHdCHPPh2. This particular ligand, which is the
unsaturated analogue of 1,2-bis(diphenylphosphino)-
ethane (dppe or diphos), has not been examined, to our
knowledge, for its reactivity with simple triosmium
clusters.11 Herein we report our data on the synthesis
of 1,2-Os3(CO)10[(Z)-Ph2PCHdCHPPh2] and, more im-
portantly, its interconversion to the corresponding
chelating isomer 1,2-Os3(CO)10[(Z)-Ph2PCHdCHPPh2].
Both cluster products have been structurally character-
ized, and the kinetics for the ligand isomerization have
been investigated, providing strong evidence for a facile,
nondissociative intramolecular rearrangement of the
ancillary phosphine about the Os3 frame.
Synthesis of 1,2-Os3(CO)10[(Z)-Ph2PCHdCHPPh2] from
Os3(CO)10(MeCN)2. To a Schlenk tube under argon containing
25 mL of CH2Cl2 and Os3(CO)10(MeCN)2, the latter of which
was prepared in situ from 0.50 g (0.55 mmol) of Os3(CO)12 and
91 mg (1.2 mmol) of Me3NO, was added 0.22 g (0.56 mmol) of
(Z)-Ph2PCHdCHPPh2 in 15 mL of CH2Cl2 via syringe. The
reaction mixture was stirred overnight at room temperature,
after which the solvent was removed under vacuum and
purification effected by column chromatography over silia gel
using CH2Cl2/hexanes (1:9) as the eluent. The major product
that was isolated corresponded to 1,2-Os3(CO)10[(Z)-Ph2PCHd
CHPPh2] along with a small amount of yellow Os3(CO)11[η1-
(Z)-Ph2PCHdCHPPh2], whose identity was ascertained spec-
troscopically and confirmed by a sample of the same that was
isolated from the independent reaction of Os3(CO)11(MeCN)
with (Z)-Ph2PCHdCHPPh2. Recrystallization of 1,2-Os3(CO)10-
[(Z)-Ph2PCHdCHPPh2] from CH2Cl2/hexane gave cluster 2b
as an analytically pure yellow-orange solid. Yield: 0.47 g
(67%). IR (CH2Cl2): ν(CO) 2088 (s), 2025 (s), 2013 (vs), 2001
(vs), 1971 (s), 1949 (b, sh) cm-1. 1H NMR (CDCl3): δ 6.29 (2H,
m, vinyl), 6.80-7.45 (20H, m, aryl). 13C NMR (toluene-d8; 243
K): δ 173.92 (s, 2C, equatorial), 178.91 (s, 2C, equatorial),
184.96 (s, 2C, axial), 191.49 (s, 4C, axial). 31P NMR (CDCl3):
δ -6.45. Anal. Calcd (found) for C36H22O10Os3P2: C, 34.64
(34.95); H, 1.76 (1.78). Yield of Os3(CO)11[η1-(Z)-Ph2PCHd
CHPPh2]: 25 mg (3.6%). IR (CH2Cl2): ν(CO) 2107 (m), 2053
Experimental Section
General Considerations. The activated cluster Os3(CO)10-
(MeCN)2 was synthesized from Os3(CO)12 according to the
published procedure,12 while the parent cluster Os3(CO)12 was
prepared from OsO4 and CO using a 500 mL Series 4571 Parr
autoclave.13 The diphosphine ligand (Z)-Ph2PCHdCHPPh2 and
Me3NO‚2H2O were both purchased from Aldrich Chemical Co.,
with the former chemical used as received and the latter dried
by azeotropic distillation from benzene. The 13CO (>99%) used
in the synthesis of the 13CO-enriched Os3(CO)12 was purchased
from Isotec Inc. and used as received. The reaction and NMR
solvents were purified by distillation under argon from an
appropriate drying agent. All distilled solvents were handled
under argon and stored in Schlenk vessels equipped with
Teflon stopcocks.14 The photochemical experiments were per-
formed at room temperature with either GE blacklight bulbs,
having a maximum output of 366 ( 20 nm, or a 200 W Oriel
Hg(Xe) arc lamp. The combustion analyses on clusters 2b,c
were performed by Atlantic Microlab, Norcross, GA.
(s), 2032 (s), 2016 (vs), 2001 (s, sh), 1989 (m), 1974 (m) cm-1
.
1H NMR (CDCl3): δ 6.98-7.60 (22H, m, vinyl and aryl). 31P
NMR (CDCl3): δ -16.37 (1P, d, JP-P ) 21 Hz), -30.65 (1P, d,
JP-P ) 21 Hz).
Synthesis of 1,1-Os3(CO)10[(Z)-Ph2PCHdCHPPh2] from
1,2-Os3(CO)10[(Z)-Ph2PCHdCHPPh2]. To a Carius tube
under argon was added 0.20 g (0.16 mmol) of 1,2-Os3(CO)10-
[(Z)-Ph2PCHdCHPPh2] and 15 mL of toluene, after which CO
was bubbled through the solution for several minutes prior to
sealing the tube. The reaction mixture was heated to ca. 85
°C for 22 h, and then the solution was cooled to room
temperature and the solvent removed. The residue was puri-
fied by chromatography over silica gel using CH2Cl2/hexanes
(4:6) as the eluent to afford a yellow-orange solid that actually
consisted of a 1:7 mixture of 1,2- and 1,1-Os3(CO)10[(Z)-Ph2-
PCHdCHPPh2 (vide infra) by 31P NMR spectroscopy. The
chelating isomer was obtained free from the bridging isomer
by careful recrystallization from a 1:1 mixture of CH2Cl2 and
hexane. Yield: 80 mg (40%). IR (CH2Cl2): ν(CO) 2092 (m), 2042
(vs), 2006 (vs, b), 1973 (m), 1957 (m), 1926 (w, b) cm-1. 1H NMR
(CDCl3): δ 7.30-7.80 (22H, m, vinyl and aryl). 13C NMR
(toluene-d8; 243 K): δ 171.24 (s, 2C, equatorial), 178.59 (s, 2C,
equatorial), 185.34 (s, 4C, axial), 196.39 (t, 2C, axial, JP-C ) 9
Hz). 31P NMR (CDCl3): δ 37.39. Anal. Calcd (found) for
C36H22O10Os3P2: C, 34.64 (34.71); H, 1.76 (1.68).
X-ray Structural Determinations. Single crystals of
clusters 2b were grown from a CH2Cl2 solution containing the
pure bridging isomer that had been layered with hexane, while
single crystals of cluster 2c suitable for diffraction analysis
were grown from an equilibrated solution containing both
isomers (ca. 85:15 mixture of 2c and 2b) using CH2Cl2 and
hexane. X-ray data were collected on a Bruker SMART 1000
CCD-based diffractometer at 297(2) and 213(2) K for samples
2b,c, respectively. The frames were integrated with the
available SAINT software package using a narrow-frame
(7) (a) Deeming, A. J.; Hardcastle, K. I.; Kabir, S. E. J. Chem. Soc.,
Dalton Trans. 1988, 827. (b) Colbran, S. B.; Irele, P. T.; Johnson, B.
F. G.; Kaye, P. T.; Lewis, J.; Raithby, P. R. J. Chem. Soc., Dalton Trans.
1989, 2033. (c) Brown, M. P.; Dolby, P. A.; Harding, M. M.; Mathews,
A. J.; Smith, A. K. J. Chem. Soc., Dalton Trans. 1993, 1671. (d) Abedin,
S. M. T.; Azam, K. A.; Hursthouse, M. B.; Kabir, S. E.; Malik, K. M.
A.; Mottalib, M. A.; Rosenberg, E. J. Cluster Sci. 2001, 12, 5. (e) Azam,
K. A.; Hursthouse, M. B.; Islam, M. R.; Kabir, S. E.; Abdul Malik, K.
M.; Miah, R.; Sudbrake, C.; Vahrenkamp, H. J. Chem. Soc., Dalton
Trans. 1998, 1097.
(8) (a) Lavigne, G. In The Chemistry of Metal Cluster Complexes;
Shriver, D. F., Kaesz, H. D., Adams, R. D., Eds.; VCH: New York,
1990; Chapter 5. (b) Lavigne, G.; de Bonneval, B. In Catalysis by Di-
and Polynuclear Metal Cluster Complexes; Adams, R. D., Cotton, F.
A., Eds.; Wiley-VCH: New York, 1998; Chapter 2.
(9) (a) Clucas, J. A.; Dawson, R. H.; Dolby, P. A.; Harding, M. M.;
Pearson, K.; Smith, A. K. J. Organomet. Chem. 1986, 311, 153. (b)
Ang, H. G.; Ang, S. G.; Wang, X. J. Chem. Soc., Dalton Trans. 2000,
3429.
(10) For a related reaction starting from a triosmium nitrate
complex, see: Hui, B. K.-M.; Wong, W.-T. J. Chem. Soc., Dalton Trans.
1998, 447.
(11) For the reaction between H4Os4(CO)12 and (Z)-Ph2PCHd
CHPPh2, see: Choi, Y.-Y.; Wong, W.-T. J. Organomet. Chem. 1997,
542, 121.
(12) Nicholls, J. N.; Vargas, M. D. Inorg. Synth. 1989, 26, 289.
(13) Drake, S. R.; Loveday, P. A. Inorg. Synth. 1990, 28, 230.
(14) Shriver, D. F. The Manipulation of Air-Sensitive Compounds;
McGraw-Hill: New York, 1969.

Isomerization of a Diphosphine about an Os3 Cluster
Organometallics, Vol. 24, No. 22, 2005 5433
algorithm,15 and each structure was solved and refined using
the SHELXTL program package.16 The Flack parameter for
structure 2b was found to be 0.006(7), confirming the choice
of absolute configuration. The molecular structure was checked
by using PLATON,17 and all non-hydrogen atoms were refined
anisotropically, with the hydrogen atoms assigned calculated
positions and allowed to ride on the attached heavy atom. The
refinement for 2b converged at R ) 0.0274 and Rw ) 0.0548
for 8700 independent reflections with I > 2σ(I) and for 2c at
R ) 0.0448 and Rw ) 0.0585 for 8180 independent reflections
with I > 2σ(I).
Kinetic Studies. The UV-vis studies were carried out in
toluene solution at a cluster concentration of ca. 10-4 M using
1.0 cm quartz UV-visible cells that were equipped with a high-
vacuum Teflon stopcock to facilitate handling on the vacuum
line and glovebox transfers. Stock solutions of 2b in toluene
when not in use were stored in the dark and under argon. All
samples of 2b for UV-vis kinetic experiments were saturated
with CO prior to heating to suppress the formation of the
hydride-bridged cluster HOs3(CO)9[µ-(Z)-PhP(C6H4)CHd
CHPPh2] (vide infra). The Hewlett-Packard 8452A diode array
spectrometer employed in our studies was configured with a
variable-temperature cell holder and was connected to a VWR
constant-temperature circulator, which regulated the reaction
temperature to within (0.5 K. The 31P NMR kinetic and
equilibration studies were conducted in 5 mm NMR tubes that
possessed a J-Young valve for the easy admission of CO gas.
All NMR studies employed toluene-d8 as a solvent with a
cluster concentration of ca. 10-2 M. The NMR samples were
heated in the same VWR temperature bath and quenched in
an external ice bath immediately before NMR analysis.
The UV-vis kinetics were monitored by following the
increase of the 368 nm absorbance band as a function of time
for a minimum of 6 half-lives. The rate constants for the
approach to equilibrium (ke) were determined by nonlinear
regression analysis using the single-exponential function18
vis experiments. The Keq values reported here for the bridge
a chelate equilibrium were determined after the isomerization
reactions had reached 10 half-lives. The quoted activation
parameters were calculated from plots of ln(k/T) versus T-1 21
,
with the error limits representing deviation of the data points
about the least-squares line of the Eyring plot.
Results and Discussion
I. Synthesis and Spectroscopic Data for the
Isomeric Clusters Os3(CO)10[(Z)-Ph2PCHdCHPPh2].
The reaction between Os3(CO)10(MeCN)2 (1) and the
unsaturated diphosphine ligand (Z)-Ph2PCHdCHPPh2
proceeds rapidly and without complications at room
temperature to give the ligand-bridged cluster 1,2-Os3-
(CO)10[(Z)-Ph2PCHdCHPPh2] (2b) as the predominant
product, along with minor amounts of Os3(CO)11[η1-(Z)-
Ph2PCHdCHPPh2], whose origin undoubtedly comes
from the mono(acetonitrile) derivative Os3(CO)11(MeCN),
which accompanies the preparation of 1.22 2b was
purified by chromatography over silica gel and charac-
terized in solution by IR and NMR spectroscopy. Par-
ticularly diagnostic in the characterization of 2b was
the high-field 31P resonance at δ -6.45, which supports
the bridging of adjacent osmium centers by the ancillary
diphosphine ligand,4c,23 with the 1H NMR spectrum
revealing the two vinyl hydrogens of the ethylene bridge
at δ 6.29 as part of an AA′XX′ multiplet due to coupling
with the two phosphorus centers.24 The limiting 13C
NMR spectrum of a 13CO-enriched sample of 2b re-
corded at 243 K exhibits four singlet resonances in a
1:1:1:2 integral ratio consistent with a diphosphine-
bridged cluster having idealized C2v symmetry. The CO
assignments for 2b are
A(t) ) A∞ + ∆A[e(-k t)
]
e
To accurately determine the rates to equilibrium by 31P
NMR spectroscopy for the bridge-to-chelate isomerization
starting from 2b, we first measured the 31P spin-lattice
relaxation times (T1) for the three participant clusters using
the standard inversion-recovery pulse sequence.19 The T1
values found for the clusters 2b (4.8 s), 2c (1.8 s), and HOs3-
(CO)9[µ-(Z)-PhP(C6H4)CHdCHPPh2] (an average of 2.2 s for
the two different 31P groups)20 allowed us to determine an
appropriate acquisition delay for the accurate collection of our
kinetic and thermodynamic data free from saturation effects.
Use of a 41.5° flip angle and an acquisition delay of 20 s
ensured complete relaxation (>99.6%) of all 31P nuclei in
solution. The integrated intensities of the 31P signals for
clusters 2b,c exhibited clean exponential decays and growths,
respectively, and afforded values for ke that were found to be
in excellent agreement with those data obtained from the UV-
Our initial evidence for the transformation of 2b into
the chelating isomer 2c came from a routine thermolysis
study. Heating a sample of 2b in benzene-d6 in a sealed
NMR tube at 75 °C overnight showed a diminution in
the intensity of the vinyl protons of 2b and a new 31P
resonance at δ 37.39, whose large nuclear deshielding
is consistent with the formation of a chelating diphos-
phine ligand.23 Continued heating of this same mixture
for over 1 week led to the slow growth of 2c at the
expense of 2b and establishment of an equilibrium that
(15) Saint Version 6.02; Bruker Analytical X-ray Systems, Inc.,
Madison, WI, 1997-1999.
(16) SHELXTL Version 5.1; Bruker Analytical X-ray Systems, Inc.,
Madison, WI, 1998.
(17) Spek, A. L. PLATON-A Multipurpose Crystallographic Tool;
Utrecht University, Utrecht, The Netherlands, 2001.
(18) All rate calculations were performed with the aid of the
commercially available programs Origin6.0 and KaleidaGraph. Here
the initial (A0) and final (A∞) absorbances and the rate constant to
equilibrium (ke) were floated to give the quoted least-squares value
for the first-order rate constant ke.
(21) Carpenter, B. K. Determination of Organic Reaction Mecha-
nisms; Wiley-Interscience: New York, 1984.
(22) For a study on the substitution kinetics involving MeCN
replacement by group 15 ligands in this genre of cluster, see: Dahl-
inger, K.; Poe¨, A. J.; Sayal, P. K.; Sekhar, V. C. J. Chem. Soc., Dalton
Trans. 1986, 2145.
(23) (a) Garrou, P. E. Chem. Rev. 1981, 81, 229. (b) Richmond, M.
G.; Kochi, J. K. Inorg. Chem. 1987, 6, 254.
(24) (a) Garbisch, E. W., Jr. J. Chem. Educ. 1968, 45, 480. (b) Akitt,
J. W.; Mann, B. E. NMR and Chemistry, 4th ed.; Stanley Thornes:
Cheltenham, U.K., 2000.
(19) Sanders, J. K. M.; Hunter, B. K. Modern NMR Spectroscopy;
Oxford University Press: New York, 1993.
(20) Our T1 values found for the 31P nuclei are in close agreement
with the values determined for other phosphine-substituted inorganic/
organometallic compounds. See: (a) Plourde, F.; Gilbert, K.; Gagnon,
J.; Harvey, P. D. Organometallics 2003, 22, 2862. (b) Socol, S. M.;
Lacelle, S.; Verkade, J. G. Inorg. Chem. 1987, 26, 3221.

5434 Organometallics, Vol. 24, No. 22, 2005
Watson et al.
Scheme 1
Table 1. X-ray Crystallographic Data and
favored the chelating isomer 2c.25 Interestingly, TLC
examination of parallel thermolysis reactions conducted
in Schlenk vessels starting with 2b showed only a single
spot for the equilibrium involving 2b and 2c. Had this
assay been the sole method of analysis, we would have
missed the isomerization reaction. A trace amount of a
third cluster (<1%) was also detected spectroscopically
under these conditions and assigned to that of HOs3-
(CO)9[µ-(Z)-PhP(C6H4)CHdCHPPh2], on the basis of a
high-field triplet at δ -17.00 (2JP-H ) 13 Hz) in the 1H
spectrum and two nonequivalent 31P resonances at δ
43.31 and 54.13.26 The ortho metalation of the ancillary
(Z)-Ph2PCHdCHPPh2 ligand in 2c is a phenomenon
that has been seen in related triosmium clusters con-
taining the diphosphine ligands dppm,6 dppe, and
dppp.7a Thermolysis of 2b in the presence of 1 atm of
CO greatly suppressed the amount of hydride cluster
formed without adversely affecting the formation of 2c.
The origin of HOs3(CO)9[µ-(Z)-PhP(C6H4)CHdCHPPh2]
was unequivocally demonstrated through the irradiation
of samples of pure 2c (vide infra) using near-UV light.
Here the same NMR resonances (1H and 31P) for the
hydride cluster were reproduced. The possibility of 2b
serving as the source of the hydride cluster was studied
by UV-vis spectroscopy. Irradiation of 2b with 366 nm
light led to a slow bleaching of the absorption bands for
2b with no evidence for any new species. Taken col-
lectively, these observations strongly support the in-
volvement of the two equilibria depicted in Scheme 1.27
Processing Parameters for the Isomeric
Triosmium Clusters 2b and 2c
2b
2c
CCDC entry no.
space group
a, Å
274229
274230
monoclinic, P21/n
10.444(1)
24.115(3)
14.370(2)
100.940(2)
3553.6(9)
C36H22O10Os3P2
monoclinic, P21
12.715(3)
b, Å
10.606(2)
c, Å
15.128(3)
â, deg
99.982(4)
V, Å3
2009.2(8)
mol formula
C
36H22O10Os3P2‚
CH2Cl2
fw
1332.00
2
1247.08
4
formula units per cell (Z)
Dcalcd, Mg/m3
λ(Mo KR), Å
abs coeff, mm-1
abs cor
max/min transmissn
total no. of rflns
no. of indep rflns
no. of data/restraints/
params
2.202
2.331
0.71073
9.729
empirical
0.8112/0.2635
16 711
8700
8700/0/488
0.71073
10.848
empirical
0.8804/0.2609
29 748
8180
8180/0/461
R
Rw
0.0274
0.0548
0.945
0.0448
0.0585
0.921
GOF on F2
weights
[0.04F2 + (σF)2]-1 [0.04F2 + (σF)2]-1
quired more care in order to avoid the cocrystallization
of both isomers of cluster 2, which led to poor-quality
X-ray crystals. Slow diffusion of hexane into a CH2Cl2
solution containing an equilibrium mixture of 2b and
2c reliably furnished small quantities of X-ray-quality
crystals of 2c. Single crystals of each cluster were found
to exist as discrete molecules in the unit cell with no
unusually short inter- or intramolecular contacts. Tables
1 and 2 give the pertinent X-ray processing parameters
and selected bond distances and angles, respectively.
The Os-Os bond distances in 2b range from 2.8642-
(6) Å (Os(1)-Os(2)) to 2.8774(6) Å (Os(2)-Os(3)) and
display a mean distance of 2.8717 Å. The Os-Os
distances are normal relative to those in other simple
polynuclear osmium clusters and indicate that the
triosmium frame itself does not experience any signifi-
cantly adverse perturbation due to the coordination of
(Z)-Ph2PCHdCHPPh2.28 The 10 terminal carbonyl groups
may be considered as linear with standard distances.
The observed distortion or twisting of the axial CO
groups in 2b from idealized D3h to D3 symmetry is a
feature common to many structurally characterized
II. X-ray Diffraction Structures for Os3(CO)10-
[(Z)-Ph2PCHdCHPPh2]. The solid-state structures of
the isomeric clusters were determined by X-ray crystal-
lography. Single crystals of cluster 2b could be grown
directly without any problems, but crystals of 2c re-
(25) The rate-limiting 13C NMR spectrum of a 13CO-enriched sample
of cluster 2c obtained from an equilibrated sample of 2b revealed four
carbonyl resonances in a 1:1:2:1 integral ratio, of which the downfield
triplet (JP-C ) 9 Hz) is readily ascribed to the two axial CO groups at
the phosphine-substituted osmium center. The remaining chemical
shifts are easily assigned by making use of the fact that the pairwise
equivalent equatorial carbonyl groups appear to high field of the axial
CO groups.
(26) The depicted stereochemistry for coordination of the ortho-
metalated phenyl ring and the diphosphine ligand in HOs3(CO)9[µ-
(Z)-PhP(C6H4)CHdCHPPh2] is strengthened by a recent X-ray struc-
ture of the related hydride-bridged cluster HOs3(CO)9[µ-PhP(C6H4)Cd
CC(O)C(dCHFc)C(O)PPh2]: Unpublished results.
(27) Additional evidence corroborating the proposed sequence de-
picted in Scheme 1 comes from the ligand isomerization and hydride
formation in the diphosphine-substituted cluster 1,2-Os3(CO)10(bpcd).
The isomerization of the bpcd ligand takes place under milder
conditions than that of 2b a 2c without the formation of the
corresponding hydride cluster HOs3(CO)9[µ-PhP(C6H4)CdCC(O)CH2C-
(O)PPh2], whose origin derives from 1,1-Os3(CO)10(bpcd). Detailed
kinetic studies on Os3(CO)10(bpcd) reveal that the rates of bpcd
isomerization are unaffected by added CO and phosphine trapping
ligands (unpublished results). The effects of phosphine trapping ligands
on the isomerization 2b a 2c could not be investigated, due to the
competitive loss of CO in 2c and formation of HOs3(CO)9[µ-(Z)-PhP-
(C6H4)CHdCHPPh2].
(28) (a) Churchill, M. R.; DeBoer, B. G. Inorg. Chem. 1977, 16, 878.
(b) Green, M.; Orpen, A. G.; Schaverien, C. J. J. Chem. Soc., Dalton
Trans. 1989, 1333. (c) Bruce, M. I.; Pain, G. N.; Hughes, C. A.; Patrick,
J. M.; Skelton, B. W.; White, A. H. J. Organomet. Chem. 1986, 307,
343. (d) Adams, R. D.; Tanner, J. T. Organometallics 1989, 8, 563. (e)
Constable, E. C.; Johnson, B. F. G.; Khan, F. K.; Lewis, J.; Raithby, P.
R.; Mikulcik, P. J. Organomet. Chem. 1991, 403, 15. (f) Biradha, K.;
Hansen, V. M.; Leong, W. K.; Pomeroy, R. K.; Zaworotko, M. J. J.
Cluster Sci. 2000, 11, 285.

Isomerization of a Diphosphine about an Os3 Cluster
Organometallics, Vol. 24, No. 22, 2005 5435
Table 2. Selected Bond Distances (Å) and Angles
mode of coordination of the unsaturated phosphine to
the metal frame in 2b is seen in the C(1)-P(1)-Os(1)
(122.1(3)°) and C(2)-P(2)-Os(2) (121.7(3)°) bond angles,
which exhibit a substantial deviation from the antici-
pated bond angle of ca. 109° expected for a tetracoor-
dinate phosphorus center.
(deg) for the Isomeric Triosmium Clusters
2b and 2ca
Cluster 2b
Bond Distances
Os(1)-Os(2)
Os(2)-Os(3)
Os(1)-P(1)
C(1)-C(2)
2.8642(6)
2.8774(6)
2.307(2)
1.32(1)
Os(1)-Os(3)
P(1)‚‚‚P(2)
Os(2)-P(2)
2.8734(7)
3.863 (4)
2.314(2)
The molecular structure of 2c shown in Figure 1
confirms the attendant migration of the diphosphine
ligand upon heating cluster 2b. The mean Os-Os and
Os-P bond lengths were computed as 2.9139 and 2.296
Å, respectively, and are unremarkable with respect to
2b and other phosphine-substituted osmium clusters.
The 10 ancillary carbonyls are all linear in nature and
reveal a slight twist or canting of the axial CO groups,
albeit much less pronounced than the twist observed
in 2b. The nonbonding P(1)‚‚‚P(2) bond distance of
3.092(6) Å in 2c is nearly 0.77 Å shorter than the
internuclear P(1)‚‚‚P(2) separation in 2b and is in
keeping with the trend reported for the free ligand (vide
supra) and other mono- and polynuclear compounds
containing a chelating (Z)-Ph2PCHdCHPPh2 ligand.10,11,33
Consistent with our premise concerning the greater
stability for chelation of the (Z)-Ph2PCHdCHPPh2
ligand in 2c vis-a´-vis cluster 2b are the angles of 109.0-
(3), 107.8(3), 120.2(6), and 118.1(6)° found for the Os-
(1)-P(1)-C(1), Os(1)-P(2)-C(2), P(1)-C(1)-C(2), and
P(2)-C(2)-C(1) linkages, respectively. These values
show minimal deviation from the idealized bond angles
of 109° for phosphorus and 120° for carbon.
III. Kinetic Data for Diphosphine Ligand Isomer-
ization in Os3(CO)10[(Z)-Ph2PCHdCHPPh2]. The
isomerization reaction involving 2b and 2c was inves-
tigated by both UV-vis and 31P NMR spectroscopy, with
the latter analytical method providing unequivocal
evidence for the reversible equilibrium between 2b and
2c. The approach to equilibrium for the conversion of
2b to 2c was monitored by 31P NMR spectroscopy in
toluene-d8 at 358.0 and 373.0 K, with the data from the
latter temperature displayed in Figure 2. The plot of
the total Os3 concentration versus time reveals smooth
exponential decay and growth curves for the consump-
tion and formation of clusters 2b,c, respectively. Non-
linear regression analysis of these data afforded a first-
order rate constant for the consumption of 2b (ke )
[1.82(4)] × 10-4 s-1) that was nearly identical with the
rate determined for the formation of 2c (ke ) [1.74(2)]
× 10-4 s-1). The rate constants (ke) reported for the two
31P NMR experiments in Table 3 represent an average
of the reaction rates from the decay and growth curves
of cluster 2. The final 31P NMR spectrum in Figure 2
recorded after 11 h of thermolysis (>10 half-lives)
clearly favors the chelating isomer (δ 37.86) and affords
Keq ) 6.9 (at 373 K) for the 2b a 2c isomerization. An
identical Keq value was determined from the 31P NMR
experiment conducted at 358.0 K.34 With the Keq value
in hand, the first-order rate constants k1 and k-1, as
defined in Scheme 1, are readily extracted and are
Bond Angles
159.83(5) P(1)-Os(1)-Os(3)
100.27(6) P(2)-Os(3)-Os(2)
P(1)-Os(1)-Os(2)
P(2)-Os(3)-Os(1)
C(27)-Os(1)-Os(2)
C(36)-Os(1)-Os(2)
C(29)-Os(2)-Os(1)
C(32)-Os(2)-Os(1)
C(34)-Os(3)-Os(1)
C(35)-Os(3)-Os(1)
C(1)-P(1)-Os(1)
C(2)-C(1)-P(1)
101.88(6)
157.10(5)
80.4(3)
93.8(3)
83.0(2)
82.3(3)
95.3(3)
78.6(3)
97.0(3)
122.1(3)
134.2(6)
C(27)-Os(1)-Os(3)
C(36)-Os(1)-Os(3)
C(29)-Os(2)-Os(3)
C(32)-Os(2)-Os(3)
C(34)-Os(3)-Os(2)
C(35)-Os(3)-Os(2)
C(2)-P(2)-Os(3)
C(1)-C(2)-P(2)
94.1(3)
93.7(3)
77.7(3)
97.6(2)
81.9(3)
121.4(3)
134.0(6)
Cluster 2c
Bond Distances
Os(1)-Os(2)
Os(2)-Os(3)
Os(1)-P(1)
C(1)-C(2)
2.9218(5)
Os(1)-Os(3)
P(1)‚‚‚P(2)
Os(1)-P(2)
2.9286(5)
3.092 (6)
2.299(2)
2.8914(5)
2.293(2)
1.322(9)
Bond Angles
P(1)-Os(1)-P(2)
P(2)-Os(1)-Os(2)
P(2)-Os(1)-Os(3)
C(27)-Os(1)-Os(3)
C(28)-Os(1)-Os(3)
C(29)-Os(2)-Os(1)
C(32)-Os(2)-Os(1)
C(34)-Os(3)-Os(1)
C(36)-Os(3)-Os(1)
C(2)-P(2)-Os(1)
C(1)-C(2)-P(2)
84.66(7) P(1)-Os(1)-Os(2)
107.62(5) P(1)-Os(1)-Os(3)
166.48(5) C(27)-Os(1)-Os(2)
167.54(5)
108.61(5)
92.9(2)
85.2(2)
92.4(2)
86.5(2)
90.7(2)
84.8(2)
91.6(2)
107.8(3)
120.2(6)
C(28)-Os(1)-Os(2)
C(29)-Os(2)-Os(3)
C(32)-Os(2)-Os(3)
C(34)-Os(3)-Os(2)
C(36)-Os(3)-Os(2)
C(1)-P(1)-Os(1)
C(2)-C(1)-P(1)
86.4(2)
91.7(2)
84.9(2)
92.2(2)
85.3(2)
109.0(3)
118.1(6)
a Numbers in parentheses are estimated standard deviations
in the least significant digits.
P-substituted Os3 and Ru3 clusters,29 as supported by
molecular mechanics calculations.30 The P(1)-Os(1)
(2.307(2) Å) and P(2)-Os(3) (2.314(2) Å) distances are
in excellent agreement with the average lengths re-
ported for a variety of P-substituted osmium com-
pounds.31 Coordination of the diphosphine ligand across
the Os(1)-Os(3) bond in 2b results in a significant
stretching of the (Z)-Ph2PCHdCHPPh2 ligand, which
leads to ground-state destabilization relative to the
chelating isomer 2c. The internuclear P(1)‚‚‚P(2) dis-
tance of 3.863(4) Å found for 2b is ca. 0.58 Å longer than
the corresponding distance reported for the free ligand
(Z)-Ph2PCHdCHPPh2, while the P(1)-C(1)-C(2) and
P(2)-C(2)-C(1) bond angles of 134.2(6) and 134.0(6)°,
respectively, associated with the unsaturated backbone
of the diphosphine ligand are ca. 11° larger as compared
to the analogous linkages in (Z)-Ph2PCHdCHPPh2.32
Additional evidence supporting the unfavorable bridging
(29) (a) Alex, R. F.; Einstein, F. W. B.; Jones, R. H.; Pomeroy, R. K.
Inorg. Chem. 1987, 26, 3175. (b) Bruce, M. I.; Liddell, M. J.; Hughes,
C. A.; Skelton, B. H.; White, A. H. J. Organomet. Chem. 1988, 347,
157. (c) Bruce, M. I.; Liddell, M. J.; Hughes, C. A.; Patrick, J. M.;
Skelton, B. H.; White, A. H. J. Organomet. Chem. 1988, 347, 181. (d)
Bruce, M. I.; Liddell, M. J.; Skawkataly, O. B.; Hughes, C. A.; Skelton,
B. H.; White, A. H. J. Organomet. Chem. 1988, 347, 207.
(33) (a) Yang, K.; Bott, S. G.; Richmond, M. G. Organometallics 1994,
13, 3788. (b) Watson, W. H.; Ejsmont, K.; Liu, J.; Richmond, M. G. J.
Chem. Crystallogr. 2003, 33, 775.
(34) The equilibrium constant for the 2b a 2c isomerization at each
of the temperatures reported in Table 3 was also determined by 31P
NMR spectroscopy. The calculated Keq values were found to be
relatively temperature insensitive and yielded an average value of 6.9,
which was employed in extraction of the individual rate constants
associated with the isomerization reaction.
(30) Lauher, J. W. J. Am. Chem. Soc. 1986, 108, 1521.
(31) Orpen, A. G.; Brammer, L.; Allen, F. K.; Kennard, O.; Watson,
D. G.; Taylor, R. J. Chem. Soc., Dalton Trans. 1989, S1.
(32) Berners, S. J.; Colquhoun, L. A.; Healy, P. C.; Byriel, K. A.;
Hanna, J. V. J. Chem. Soc., Dalton Trans. 1992, 3357.

5436 Organometallics, Vol. 24, No. 22, 2005
Watson et al.
Figure 1. Thermal ellipsoid plots of the bridging (left) and chelating (right) isomers of Os3(CO)10[(Z)-Ph2PCHdCHPPh2],
showing the thermal ellipsoids at the 50% probability level.
Figure 2. Representative 31P NMR spectra (left) recorded at 373 K in toluene-d8 for 2b a 2c and the resulting plot for
the disappearance of 2b and the appearance of 2c and HOs3(CO)9[µ-(Z)-PhP(C6H4)CHdCHPPh2] (right). The bottom 31P
NMR spectrum was recorded after 11 h of thermolysis.
reported in Table 3.35 Finally, the hydride cluster
observed in Figure 2 at δ 43 and 54 after 11 h of
thermolysis accounted for less than 4% of the total Os3
concentration.
The kinetics for the isomerization reaction were also
studied by UV-vis spectroscopy in toluene solution
under 1 atm of CO over the temperature range of
358.5-381.5 K by following the increase in the absor-
bance of the 368 nm band belonging to 2c. The UV-vis
spectral changes for the thermolysis of 2b in toluene
carried out at 373 K are shown in Figure 3, where
clearly defined isosbestic points at 394 and 437 nm are
seen to accompany the reaction. The excellent fit
between the least-squares regression curve and the
absorbance data (Figure 3) confirms that the isomer-
ization is well-behaved and free from complications.
Unlike the 31P NMR experiments, where the thermally
promoted loss of CO in 2c gives rise to small quantities
of the hydride cluster HOs3(CO)9[µ-(Z)-PhP(C6H4)CHd
(35) (a) The values for k1 and k-1 for a reversible first-order process
were computed with the aid of the following equations: k1 ) ke/[1 +
Keq-1] and k-1 ) ke/[1 + Keq]. (b) For a detailed derivation of the
integrated rate expression of a reversible first-order reaction, see:
Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2nd ed.;
McGraw-Hill: New York, 1995.

Isomerization of a Diphosphine about an Os3 Cluster
Organometallics, Vol. 24, No. 22, 2005 5437
from PtMe4(dppe).38 There the rate-limiting chelate
opening of the diphosphine ligand in PtMe4(dppe)
furnishes a reactive five-coordinate intermediate that
serves as the platform for the evolution of ethane.39
Table 3. Kinetic Data for the Isomerization of 2b
and 2c in Toluenea
b
temp (K) 104ke (s-1
)
104k1 (s-1
)
104k-1 (s-1
)
method
358.0
358.5
364.0
369.0
373.0
373.0
373.0
378.0
381.5
0.37 ( 0.01 0.32 ( 0.01 0.05 ( 0.01 31P NMR
0.40 ( 0.01 0.35 ( 0.01 0.05 ( 0.01 UV-vis
0.66 ( 0.01 0.58 ( 0.01 0.08 ( 0.01 UV-vis
1.16 ( 0.02 1.01 ( 0.02 0.15 ( 0.02 UV-vis
1.78 ( 0.04 1.56 ( 0.04 0.22 ( 0.04 31P NMR
1.76 ( 0.01 1.54 ( 0.01 0.22 ( 0.01 UV-vis
IV. Diphosphine Isomerization Mechanism. Any
acceptable mechanism for the isomerization of the (Z)-
Ph2PCHdCHPPh2 ligand about the triosmium frame
must account for the observed P ligand/CO exchange
that accompanies the reaction. A terminal-bridge-
terminal sequence involving a CO group(s) and a PPh2
moiety across an Os-Os bond(s) would effectively serve
to scramble the participant ligands. The highly ordered
transition state mandated by any such terminal-
bridge-terminal sequence is reconciled with the nega-
tive ∆Sq values found from Eyring analysis of the rate
data. A rate-limiting process involving the dissociation
or release of one of the arms of the bidentate ligand,
followed by scrambling of the diphosphine and subse-
quent ligand ring closure (bridge or chelate), can be
immediately eliminated from consideration, as it would
violate the observed entropies of activation and insen-
sitivity of the reaction to added CO.
The top portion of Scheme 2 depicts a pairwise
terminal-bridge exchange in cluster 2b involving an
axial group (COa) and one P moiety across the diphos-
phine-bridged Os-Os bond that proceeds through the
doubly bridged cluster Os3(CO)9(µ-CO)[(Z)-µ,η1-Ph2-
PCHdCHPPh2]. This species can produce the chelating
isomer 2c ultimately upon opening of the µ-CO and
µ-phosphine bridges. However, the simple opening of
this doubly bridged cluster cannot give 2c directly, as
an additional permutation(s) or turnstile rotation(s) of
the remaining spectator groups associated with the
phosphine-substituted osmium centers must accompany
the bridge-opening reaction. A more direct diphosphine
equilibration that relies on the well-known “merry-go-
round” mechanism is illustrated in the bottom portion
of Scheme 2. Here a terminal-bridge exchange of two
equatorial CO groups and one of the phosphine centers
in 2b gives rise to the triply bridged cluster Os3(CO)8-
(µ-CO)2[(Z)-µ,η1-Ph2PCHdCHPPh2], which directly fur-
nishes 2c upon completion of the bridge-terminal
exchange.40,41 We favor the latter mechanism for diphos-
phine isomerization in Os3(CO)10[(Z)-Ph2PCHdCHPPh2],
on the basis of its simplicity.42 Given that no dynamic
diphosphine isomerization has been observed in Os3-
(CO)10(dppe),1a the saturated analogue of 2, we submit
that the stereochemistry and rigid nature of the ancil-
1.75 ( 0.02
UV-visc
2.92 ( 0.14 2.55 ( 0.14 0.37 ( 0.14 UV-vis
3.82 ( 0.12 3.34 ( 0.12 0.48 ( 0.12 UV-vis
a All kinetic data were collected under 1 atm of CO unless
otherwise noted. The quoted UV-vis kinetic data represent an
average of two independent measurements. b The quoted rate
constants (ke) represent the rate toward equilibrium, as deter-
mined by following the increase in the 368 nm band (UV-vis) and
the changes in the 31P resonances for the bridging and chelating
isomers (31P NMR). c Experiment conducted in a Fischer-Porter
tube under 6.8 atm of CO (constant pressure). Only the determined
value of ke is reported here.
CHPPh2], the amount of hydride cluster present in the
UV-vis experiments is predicted to be negligible. The
equilibrium in Scheme 1 mandates the suppression of
the hydride cluster for reactions carried out under high
concentrations of CO relative to cluster 2. A conservative
10-fold concentration excess of CO relative to the
bridging and chelating cluster 2 is predicted from
available solubility data for CO in toluene.36 The rate
constants for the approach to equilibrium (ke), along
with the values of k1 and k-1, as computed with the
aforementioned Keq value for 2b a 2c, are given in
Table 3.37 The UV-vis data collected at 373 K furnished
values for ke, k1, and k-1 in excellent agreement with
those values determined by the 31P NMR experiment,
underscoring the validity of the data obtained from the
two different experiments. That the presence of added
CO does not affect the equilibrium between 2b and 2c
was demonstrated by the reaction conducted under 6.8
atm of CO. Here the value of [1.75(2)] × 10-4 s-1 found
for ke is within experimental error to the corresponding
reaction run under 1 atm of CO.
Eyring plots of ln(k1/T) and ln(k-1/T) versus T-1
allowed for the determination of the isomerization
activation parameters for the forward and reverse
reaction directions in cluster 2. The calculated values
of ∆Gq
) 28.4(6) kcal/mol, ∆Hq ) 26.5(6) kcal/mol,
373
and ∆Sq ) -5(2) eu for the forward isomerization (2b
f 2c) and ∆Gq373 ) 29.9(6) kcal/mol, ∆Hq ) 26.5(6) kcal/
mol, and ∆Sq ) -9(2) eu for the reverse isomerization
(2c f 2b) reveal that the entropy of activation is small
and negative in both directions and argues against a
dissociative mechanism involving the release of one of
the PPh2 moieties from the cluster frame to generate
an unsaturated Os3 cluster having a pendant η1-diphos-
phine moiety. Such a phosphine ligand release process
that exhibits a positive ∆Sq value has been demon-
strated by Goldberg in the reductive coupling of ethane
(38) Crumpton-Bregel, D. M.; Goldberg, K. I. J. Am. Chem. Soc.
2003, 125, 9442.
(39) For some other examples exhibiting chelate-ring opening, see:
(a) Knebel, W. J.; Angelici, R. J. Inorg. Chem. 1974, 13, 627, 632. (b)
Schultz, L. D.; Dobson, G. R. J. Organomet. Chem. 1976, 124, 19. (c)
Dobson, G. L.; Corte´s, J. E. Inorg. Chem. 1988, 27, 3308. (d) Henderson,
R. A. J. Chem. Soc., Dalton Trans. 1988, 509, 515.
(40) In principle the VT 13C NMR spectra of both 2b and 2c could
provide critical data capable of differentiating the two proposed
diphosphine scrambling mechanisms. Unfortunately, the complete
coalescence of all carbonyl groups in 2b and 2c by 363 K does not allow
the 13C NMR data to distinguish between the two mechanisms depicted
in Scheme 2.
(36) (a) Van Raaij, E. U.; Schmulbach, C. D.; Brintzinger, H. H. J.
Organomet. Chem. 1987, 328, 275. (b) Basickos, L.; Bunn, A. G.;
Wayland, B. B. Can. J. Chem. 2001, 79, 854.
(37) A caveat worth pointing out is that the reversible first-order
behavior for the 2b a 2c equilibrium could have easily been misin-
terpreted as an irreversible first-order conversion involving 2b a 2c
had the UV-vis data been analyzed in the absence of the 31P NMR
data.
(41) The scrambling of the carbonyl groups in 2b and 2c followed
terminal-bridge-terminal merry-go-round and turnstile processes
found in other phosphine-substituted Os3 clusters. For related VT NMR
studies, see: (a) References 1c,d. (b) Deeming, A. J.; Donovan-Mtunzi,
S,; Kabir, S. E. J. Organomet. Chem. 1985, 281, C43. (c) Alex, R. F.;
Pomeroy, R. K. Organometallics 1987, 6, 2437.
(42) The simplicity alluded to stems from the application of Occam’s
razor.

5438 Organometallics, Vol. 24, No. 22, 2005
Watson et al.
Figure 3. UV-vis spectral changes for 2b a 2c recorded at 373 K in toluene (left) and the absorbance versus time curves
for the experimental data (9) and the least-squares fit of ke (right).
Scheme 2
lary (Z)-Ph2PCHdCHPPh2 ligand actually promote the
scrambling of the ligand about the cluster by maintain-
ing constant and intimate contact between the diphos-
phine ligand and the osmium centers.43 We believe that
this isomerization phenomenon is general in nature, at
least for this genre of cluster, provided that the diphos-
phine ligand is rigid and possesses cis stereochemistry,
based on our observations of facile diphosphine isomer-
ization in the cluster compounds Os3(CO)10(P-P) (where
P-P ) bpcd, 3,4-bis(diphenylphosphino)-5-methoxy-
2(5H)-furanone (bmf), 1,2-bis(diphenylphosphino)ben-
zene).44,45
Adams has provided conclusive NMR evidence for the
intramolecular scrambling of phosphine ligands between
the platinum and adjacent ruthenium centers in the
mixed-metal carbide clusters PtRu5(µ6-C)(CO)15P,47 while
phosphine equilibration between the two ruthenium
centers in the dinuclear compounds Cp*2Ru2(µ-H)2P has
been demonstrated by Suzuki.48,49 The postulated trans-
formation of an η1-phosphine ligand to a µ-bound
phosphine in our work and the above examples is
strengthened by reports showing the coordination of a
tertiary P ligand simultaneously to two and three metal
centers.50
The dynamic migration of CO ligands about the metal
centers in polynuclear clusters through terminal a
bridge a terminal CO exchanges has been a phenom-
enologically established fact for over 3 decades.46 Analo-
gous migratory ligand behavior has only recently been
observed for tertiary phosphine ligands. For example,
(46) (a) Cotton, F. A. Inorg. Chem. 1966, 6, 1083. (b) Johnson, B. F.
G.; Lewis, J.; Reichert, B. E.; Schorpp, K. L. J. Chem. Soc., Dalton
Trans. 1976, 1403. (c) Cotton, F. A.; Hanson, B. E. In Rearrangements
in Ground and Excited States; de Mayo, P., Ed.; Academic Press: New
York, 1980; Chapter 12.
(47) Adams, R. D.; Captain, B.; Fu, W.; Pellechia, P. J. Inorg. Chem.
2003, 42, 3111.
(48) Ohki, Y.; Suzuki, H. Angew. Chem., Int. Ed. 2002, 41, 2994.
(49) For additional reports of phosphine-ligand isomerization in
metal clusters, see: (a) Laurenczy, G.; Bondietti, G.; Ros, R.; Roulet,
R. Inorg. Chim. Acta 1996, 247, 65. (b) Braunstein, P.; Boag, N. M.
Angew. Chem., Int. Ed. 2001, 40, 2427. (c) Bradford, A. M.; Douglas,
G.; Manojloviæ-Muir, L.; Muir, K. W.; Puddephatt, R. J. Organome-
tallics 1990, 9, 409.
(50) (a) Balch, A. L.; Davis, B. J.; Olmstead, M. M. J. Am. Chem.
Soc. 1990, 112, 8592; Inorg. Chem. 1993, 32, 3937. (b) Pechmann, T.;
Brandt, C. D.; Werner, H. Angew. Chem., Int. Ed. 2000, 39, 3909;
Chem. Commun. 2003, 1136. (c) Pechmann, T.; Brandt, C. D.; Roger,
C.; Werner, H. Angew. Chem., Int. Ed. 2002, 41, 2301.
(43) In the case of an isomerization reaction involving a dppe ligand,
we can envision an intermediate state where the hydrogen atoms of
the ethane backbone must adopt an eclipsed conformation during the
bridge a chelate transformation. This specific destabilization is avoided
in diphosphines possessing an unsaturated carbon backbone.
(44) Unpublished results.
(45) Our isomerization data for 2b
a
2c parallel those of
Os3(CO)10(Ph2PCH2CH2SMe), where thiophosphine isomerization is
believed to proceed through a bridging thioether moiety. See: Persson,
R.; Monari, M.; Gobetto, R.; Russo, A.; Aime, S.; Calhorda, M. J.;
Nordlander, E. Organometallics 2001, 20, 4150.

Isomerization of a Diphosphine about an Os3 Cluster
Organometallics, Vol. 24, No. 22, 2005 5439
reflects the scrambling mechanism, then the fortuitous
similarity in the values determined for ∆G° and ∆∆Gq
indicate that the barrier heights relating 2b and 2c to
the triply bridged cluster Os3(CO)8(µ-CO)2[(Z)-µ,η1-Ph2-
PCHdCHPPh2] must be small and identical in magni-
tude.
Conclusions
The reaction between Os3(CO)10(MeCN)2 and (Z)-Ph2-
PCHdCHPPh2
gives
1,2-Os3(CO)10[(Z)-Ph2PCHd
CHPPh2] as the kinetic product of ligand substitution.
Isomerization of the bridged cluster into the chelated
cluster 1,1-Os3(CO)10[(Z)-Ph2PCHdCHPPh2] proceeds
readily upon heating. The isomerization kinetics and
thermodynamics have been explored by spectroscopic
methods and both products subjected to structural
analysis. On the basis of the kinetic data, a nondisso-
ciative phosphine migration across one of the Os-Os
bonds has been proposed. The generality of this type of
ligand isomerization as a function of metal cluster and
diphosphine ligand will be probed by our groups, with
our data reported in due course.
Figure 4. Free-energy diagrams for the merry-go-round
isomerization of 2b a 2c at 373 K where the triply bridged
cluster Os3(CO)8(µ-CO)2[(Z)-µ,η1-Ph2PCHdCHPPh2] func-
tions as a transition state (left) and as a discrete interme-
diate of finite lifetime (right).
Finally, Figure 4 shows two free-energy diagrams for
the merry-go-round isomerization for 2b a 2c at 373
K, whose major difference lies in the nature of the
contribution of the putative species Os3(CO)8(µ-CO)2[(Z)-
µ,η1-Ph2PCHdCHPPh2]. The left-hand diagram depicts
a single-step isomerization where the triply bridged
cluster Os3(CO)8(µ-CO)2[(Z)-µ,η1-Ph2PCHdCHPPh2] func-
tions as the transition state for the two isomers of Os3-
(CO)10[(Z)-Ph2PCHdCHPPh2], while the two-step mech-
anism involving Os3(CO)8(µ-CO)2[(Z)-µ,η1-Ph2PCHd
CHPPh2] as a reactive intermediate of finite lifetime in
the 2b a 2c isomerization is shown to the right.
Differentiation between the concerted (left) and two-step
(right) processes is not a trivial matter, but we would
suggest that our data favor the concerted scrambling
of the diphosphine ligand. The ∆G° value of -1.4 kcal/
mol in favor of cluster 2c, as determined from the
magnitude of Keq for 2b a 2c by 31P NMR spectroscopy
(vide supra), is identical with the ∆∆Gq value found from
the kinetic experiments. This is reasonable and expected
for a single-step or concerted reaction, with the triply
bridged cluster serving as the transition state for both
2b and 2c. Theoretical calculations on the bridge-to-
chelate isomerization of the H2PCH2CH2PH2 ligand in
the dinuclear compound Mo2Cl4(H2PCH2CH2PH2)2 re-
veal that a nondissociative migration of the phosphine
moieties across the Mo-Mo quadruple bond, analogous
to that proposed by us in Os3(CO)8(µ-CO)2[(Z)-µ,η1-Ph2-
PCHdCHPPh2], proceeds as a transition state and not
a bona fide intermediate.51,52 If the two-step process
Acknowledgment. We thank Professor Jim Espe-
nson for his advice concerning the nonlinear regression
treatment of our UV-vis data and Professor Paul
Marshall for his shared insights during our infrequent
luncheons. Financial support from the Robert A. Welch
Foundation (Grant No. P-074, W.H.W.; Grant No.
B-1093, M.G.R.) is greatly appreciated.
Supporting Information Available: Tables giving crys-
tallographic data for the bridging and chelating isomers of Os3-
(CO)10[(Z)-Ph2PCHdCHPPh2]. This material is available free
of charge via the Internet at http://pubs.acs.org. Crystal-
lographic data for the structural analyses have also been
deposited with the Cambridge Crystallographic Data Center:
CCDC No. 274229 for 1,2-Os3(CO)10[(Z)-Ph2PCHdCHPPh2]‚CH2-
Cl2 and CCDC No. 274230 for 1,1-Os3(CO)10[(Z)-Ph2PCHd
CHPPh2]. Copies of this information may be obtained free of
charge from the Director, CCDC, 12 Union Road, Cambridge
CB2 1EZ, U.K. (fax, +44(1223)336-033; e-mail, deposit@
ccdc.ac.uk; web, http://www:ccdc.cam.ac.uk).
OM0505413
(52) In a subsequent paper by the same researchers, it was disclosed
that a multistep process involving the dissociative release of one arm
of the diphosphine ligand, followed by phosphine migration and ligand
reattachment, was energetically competitive with the concerted phos-
phine migration process reported in ref 51. Since a dissociative
mechanism is not supported by our kinetic data, we can exclude it from
consideration. See: Demachy, I.; Jean, Y.; Lledos, A. New J. Chem.
2004, 28, 1494.
(51) Blasco, S.; Demachy, I.; Jean, Y.; Lledos, A. New J. Chem. 2002,
26, 1118.