Photochemical heterolysis of the metal-metal bond in (Me3P)(OC)4OsW(CO)5

Article abstract of DOI:10.1021/om970275w

The photochemistry of (Me₃P)(OC)₄OsW(CO)₅ (1), a molecule with a dative covalent metal-metal bond, was investigated. The results indicate that the photoprocess is heterolytic cleavage of the metal-metal bond, rather than homolytic cleavage as is found in compounds with nondative covalent metal-metal bonds. The principal evidence for heterolysis comes - from the irradiation (λ > 400 nm) of 1 in benzene in the presence of PPh₃. The major products were Os(CO)₄(PMe₃), Os(CO)₃(PMe₃)(PPh₃), and W(CO)₅(PPh₃). These products are consistent with a pathway that involves initial heterolysis to form Os(CO)₄(PMe₃) and W(CO)₅ followed by trapping of the coordinatively unsaturated W(CO)₅ species with PPh₃. (Control experiments showed that Os(CO)₃(PMe₃)(PPh₃) forms from Os(CO)₄(PMe₃) and PPh₃ under the reaction conditions.) Primary photoprocesses involving either M-CO bond dissociation or Os-W bond homolysis are incompatible with the results of other experiments. For example, when the irradiation of 1 was carried out under CO, the quantum yield for disappearance of 1 increased to 0.28 ± 0.05 from its value of 0.13 ± 0.01 under N₂, a result inconsistent with M-CO dissociation as the primary photoprocess. Likewise, control experiments showed that [W(CO)5^.1] was readily trapped under the reaction conditions, but all attempts to trap [W(CO)5^.1] (one of the expected homolytic primary photoproducts) with the metal-radical traps benzyl chloride, CCl₄, or TMIO (1,1,3,3-tetramethylisoindoline-2-oxyl) during the photolysis of 1 were fruitless. Compound 1 reacted photochemically with CCl₄ to form fac-Os(CO)₃(PMe₃)Cl₂ (fac-2) and W(CO)₆. The fac-Os(CO)₃(PMe₃)Cl₂ product is likely formed by a secondary photolysis of Os(CO)₄(PMe₃). Control experiments showed that irradiation of Os(CO)₄(PMe₃) in CCl₄ initially formed a compound believed to be Os(CO)₃(PMe₃)(CCl₃)-Cl, which in the presence of W(CO)₅ then reacted to form fac-Os(CO)₃(PMe₃)Cl₂. Small amounts of W(CO)₆ (8percent) were also produced in the photochemical reaction of 1 with PPh₃. An experiment in a more viscous solvent system indicated that W(CO)₆ is not formed in a CO abstraction reaction involving the [(Me₃P)(OC)₄Os, W(CO)₅] solvent cage pair. The formation of W(CO)₆ is attributed to a second (minor) heterolytic pathway involving an intermediate with a bridging CO that directly yields W(CO)₆ and Os(CO)₃(PMe₃).

Full text of DOI:10.1021/om970275w

Organometallics 1997, 16, 3431-3438
3431
P h otoch em ica l Heter olysis of th e Meta l-Meta l Bon d in
(Me3P )(OC)4OsW(CO)5
J onathan L. Male and Roland K. Pomeroy
Department of Chemistry, Simon Fraser University,
Burnaby, British Columbia, Canada V5A 1S6
David R. Tyler*
Department of Chemistry, University of Oregon, Eugene, Oregon 97403
Received April 2, 1997X
The photochemistry of (Me3P)(OC)4OsW(CO)5 (1), a molecule with a dative covalent metal-
metal bond, was investigated. The results indicate that the photoprocess is heterolytic
cleavage of the metal-metal bond, rather than homolytic cleavage as is found in compounds
with nondative covalent metal-metal bonds. The principal evidence for heterolysis comes
from the irradiation (λ > 400 nm) of 1 in benzene in the presence of PPh3. The major products
were Os(CO)4(PMe3), Os(CO)3(PMe3)(PPh3), and W(CO)5(PPh3). These products are consistent
with a pathway that involves initial heterolysis to form Os(CO)4(PMe3) and W(CO)5 followed
by trapping of the coordinatively unsaturated W(CO)5 species with PPh3. (Control experi-
ments showed that Os(CO)3(PMe3)(PPh3) forms from Os(CO)4(PMe3) and PPh3 under the
reaction conditions.) Primary photoprocesses involving either M-CO bond dissociation or
Os-W bond homolysis are incompatible with the results of other experiments. For example,
when the irradiation of 1 was carried out under CO, the quantum yield for disappearance
of 1 increased to 0.28 ( 0.05 from its value of 0.13 ( 0.01 under N2, a result inconsistent
with M-CO dissociation as the primary photoprocess. Likewise, control experiments showed
that [W(CO)5•-] was readily trapped under the reaction conditions, but all attempts to trap
[W(CO)5•-] (one of the expected homolytic primary photoproducts) with the metal-radical
traps benzyl chloride, CCl4, or TMIO (1,1,3,3-tetramethylisoindoline-2-oxyl) during the
photolysis of 1 were fruitless. Compound 1 reacted photochemically with CCl4 to form fac-
Os(CO)3(PMe3)Cl2 (fa c-2) and W(CO)6. The fac-Os(CO)3(PMe3)Cl2 product is likely formed
by a secondary photolysis of Os(CO)4(PMe3). Control experiments showed that irradiation
of Os(CO)4(PMe3) in CCl4 initially formed a compound believed to be Os(CO)3(PMe3)(CCl3)-
Cl, which in the presence of W(CO)5 then reacted to form fac-Os(CO)3(PMe3)Cl2. Small
amounts of W(CO)6 (8%) were also produced in the photochemical reaction of 1 with PPh3.
An experiment in a more viscous solvent system indicated that W(CO)6 is not formed in a
CO abstraction reaction involving the [(Me3P)(OC)4Os, W(CO)5] solvent cage pair. The
formation of W(CO)6 is attributed to a second (minor) heterolytic pathway involving an
intermediate with a bridging CO that directly yields W(CO)6 and Os(CO)3(PMe3).
In tr od u ction
metal dative bonds are intriguing because of the pos-
sibility that on photolysis the dative metal-metal bonds
may be cleaved heterolytically. Such cleavage would be
unusual because, in general, metal-metal homolysis
occurs when compounds with metal-metal bonds are
irradiated.9,10 The studies that demonstrated homoly-
sis, however, were all carried out on molecules having
In 1983, Pomeroy and co-workers reported the syn-
thesis and structural characterization of (OC)5OsOs-
(CO)3(GeCl3)Cl, a molecule with an unbridged Os-Os
bond.1 The complex was unusual because Os(CO)5, a
stable 18-electron molecule, acted as a Lewis base to
the 16-electron Os(CO)3(GeCl3)(Cl) fragment to form a
metal-metal dative bond. Since that report, other
neutral complexes with unbridged dative metal-metal
bonds have been reported.2-8 Molecules with metal-
(7) (a) Einstein, F. W. B.; J ones, T.; Pomeroy, R. K.; Rushman, P.
J . Am. Chem. Soc. 1984, 106, 2707. (b) Davis, H. B.; Einstein, F. W.
B.; Glavina, P. G.; J ones, T.; Pomeroy, R. K.; Rushman, P. Organo-
metallics 1989, 8, 1030. (c) Davis, H. B.; Einstein, F. W. B.; J ohnston,
V. J .; Pomeroy, R. K. J . Organomet. Chem. 1987, 319, C25.
(8) Shipley, J . A.; Batchelor, R. J .; Einstein, F. W. B.; Pomeroy, R.
K. Organometallics 1991, 10, 3620.
X Abstract published in Advance ACS Abstracts, J une 15, 1997.
(1) Einstein, F. W. B.; Pomeroy, R. K.; Rushman, P.; Willis, A. C. J .
Chem. Soc., Chem. Commun. 1983, 854.
(2) Fleming, M. M.; Pomeroy, R. K.; Rushman, P. J . Organomet.
Chem. 1984, 273, C33.
(3) Einstein, F. W. B.; Pomeroy, R. K.; Rushman, P.; Willis, A. C.
Organometallics 1985, 3, 250.
(4) Del Paggio, A. A.; Muetterties, E. L.; Heinekey, D. M.; Day, V.
W.; Day, C. S. Organometallics 1986, 5, 575.
(5) Einstein, F. W. B.; J ennings, M. C.; Krentz, R.; Pomeroy, R. K.;
Rushman, P.; Willis, A. C. Inorg. Chem. 1987, 26, 1341.
(6) Uso´n, R.; Fornie´s, J .; Espinet, P.; Fortun˜o, C.; Tomas, M.; Welch,
A. J . J . Chem. Soc., Dalton Trans. 1988, 3005.
(9) Irradiation of metal-metal-bonded organometallic dimers gener-
ally leads to either metal-metal bond homolysis or metal-ligand bond
dissociation. See: (a) Geoffroy, G. L.; Wrighton, M. S. Organometallic
Photochemistry; Academic Press: New York, 1979. (b) Meyer, T. J .;
Casper, J . V. Chem. Rev. 1985, 85, 187.
(10) Photoinduced heterolytic cleavage of metal-metal bonds11 has
been suggested in several systems,12-14 but definitive mechanistic
confirmation is generally lacking. In several cases, proposed heterolyses
were shown to actually involve homolytic fission followed by electron
transfer.15-17
S0276-7333(97)00275-6 CCC: $14.00 © 1997 American Chemical Society

3432 Organometallics, Vol. 16, No. 15, 1997
Male et al.
nondative covalent metal-metal bonds. This distinction
in bond types is important because there are ample
precedents for photochemical heterolyses in reactions
involving dative bonds. For example, metal-ligand
bond cleavage reactions from d-d excited states are
generally heterolytic.18 Because metal-ligand bonds
are dative in nature, it was reasoned that bimetallic
systems with dative metal-metal bonds might also
undergo heterolytic cleavage.
In order to investigate the possibility that the pho-
tolysis of metal-metal dative bonds may be hetero-
lytic,19 we studied the photochemistry of (Me3P)(OC)4-
OsW(CO)5 (1),7a,b a dinuclear complex with an unbridged,
dative metal-metal bond. Herein we describe the
F igu r e 1. Infrared spectra of the photolysis (λ > 400 nm)
of 1 in C6H6 in the presence of PPh3. The time scale of the
reaction was 0-28 min utilizing a 200 W high-pressure
mercury arc lamp. The species are identified as follows:
club ) Os(CO)4(PMe3); spade ) W(CO)5(PPh3); heart ) Os-
(CO)3(PMe3)(PPh3); diamond ) (Me3P)(OC)4OsW(CO)5 (1);
b ) W(CO)6.
results of our study; a preliminary account of our
findings has appeared.23
acceptor metal-metal bond, also exhibit a strong ab-
sorption in the near-UV region (315-330 nm, ꢀ ≈104 L
mol-1 cm-1).26 These absorptions were also ascribed to
the σ f σ* transition by analogy to studies on the
[M2(CO)10]2- (M ) Cr, W) molecules.27
Exp er im en ts Exp lor in g th e P ossibility of Het-
er olytic Clea va ge. Irradiation (λ > 400 nm) of 1 (1.66
× 10-3 M) in the presence of PPh3 (1.91 × 10-1 M) in
deoxygenated benzene gave Os(CO)4(PMe3), Os(CO)3-
(PMe3)(PPh3), W(CO)5PPh3, and W(CO)6 in the yields
shown in eq 1.
Resu lts a n d Discu ssion
Electr on ic Sp ectr u m of 1. The electronic absorp-
tion spectrum of 1 in the region 300-900 nm exhibits a
peak with a maximum at 368 nm (ꢀ ) 2.79 × 103 ( 1.9
× 102 L mol-1 cm-1) and a subtle shoulder at 334 nm (ꢀ
) 2.7 × 103 ( 1.9 × 102 L mol-1 cm-1) in C6H6 at 25 °C.
By analogy to binuclear complexes with unbridged
metal-metal bonds, the band at 368 nm is assigned to
a σ f σ* transition associated with the metal-metal
bond.24,25 For comparison, the [H(OC)4FeM(CO)5]-
complexes (M ) Cr, Mo, W), with a putative donor-
(11) Armentrout, P. B.; Simons, J . J . Am. Chem. Soc. 1992, 114,
8627.
(12) Allen, D. M.; Cox, A.; Kemp, T. J .; Sultana, Q. J .; Pitts, R. B.
J . Chem. Soc., Dalton Trans. 1976, 1189.
(13) Fletcher, S. C.; Poliakoff, M.; Turner, J . J . J . Organomet. Chem.
1984, 268, 259.
(1)
(14) Haynes, A.; Poliakoff, M.; Turner, J . J .; Bender, B. R.; Norton,
J . R. J . Organomet. Chem. 1990, 383, 497.
(15) For example, the disproportionation reactions of many dinuclear
complexes follow a pathway involving homolysis, followed by coordina-
tion of a ligand to form a 19-electron adduct, and then electron transfer.
See: Stiegman, A. E.; Stieglitz, M.; Tyler, D. R. J . Am. Chem. Soc.
1983, 105, 6032-6037.
The reaction was followed by infrared spectroscopy,
which showed the appearance of the product bands and
the disappearance of bands for 1 over a period of 28 min
(Figure 1; Table 1). The products were identified by
comparison of their infrared spectra to the spectra of
compounds prepared independently or to spectra re-
ported in the literature (Table 1). (Examples of the
spectra for all of the reactions followed by infrared
spectroscopy described in this paper are included in the
Supporting Information for this paper. Note that the
percent yields in eq 1 are reported with respect to the
total Os or W.) The photolyses were carried out into
the low-energy tail of the σ f σ* absorption band (λ >
(16) For a review of disproportionation, see: Stiegman, A. E.; Tyler,
D. R. Coord. Chem. Rev. 1985, 63, 217-240.
(17) Stufkens and Oskam reported several instances where hetero-
lytic products are the result of photochemical homolysis followed by
electron transfer between geminate cage radicals. See: (a) van Dijk,
H. K.; van der Haar, J .; Stufkens, D. J .; Oskam, A. Inorg. Chem. 1989,
28, 75-81. (b) Knoll, H.; de Lange, W. J . G.; Hennig, H.; Stufkens, D.
J .; Oskam, A. J . Organomet. Chem. 1992, 430, 123-132.
(18) For a review, see: Zinato, E. In Concepts of Inorganic Photo-
chemistry; Adamson, A. W., Fleischauer, P. D., Eds.; Wiley-Inter-
science: New York, 1975.
(19) Note that thermal heterolytic cleavage of metal-metal bonds
has been demonstrated, for example, in the complexes (Cp)(PMe3)2-
CoA [A ) Zn(Cl)2(PMe3), Cu(Cl)(PMe3)2],20 (CO)4CoRh(CO)(PEt3)2,21
and (Cp)(H)2MM′(CO)5 (M ) Mo, W; M′ ) Cr, Mo, W)22 and in several
(26) (a) Arndt, L. W.; Darensbourg, M. Y.; Delord, T.; Bancroft, B.
T. J . Am. Chem. Soc. 1986, 108, 2617 and references therein. (b) Arndt,
L. W.; Bancroft, B. T.; Darensbourg, M. Y.; J anzen, C. P.; Kim, C. M.;
Reibenspies, J .; Varner, K. E.; Youngdahl, K. A. Organometallics 1988,
7, 1302. (c) Arndt, L. W. Ph. D. Thesis, Texas A&M University, May
1986.
of the neutral unbridged complexes with
a dative metal-metal
bond.2-4,6,7
(20) Dey, K.; Werner, H. Chem. Ber. 1979, 112, 823.
(21) Roberts, D. A.; Mercer, W. C.; Geoffroy, G. L.; Pierpont, C. G.
Inorg. Chem. 1986, 25, 1439 and ref 3 therein.
(22) Deubzer, B.; Kaesz, H. D. J . Am. Chem. Soc. 1968, 90, 3276.
(23) Male, J . L.; Davis, H. B.; Pomeroy, R. K.; Tyler, D. R. J . Am.
Chem. Soc. 1994, 116, 9353.
(24) Abrahamson, H. B.; Wrighton, M. S. Inorg. Chem. 1978, 17,
1003.
(27) No electronic spectroscopic studies have been published on
complexes with metal-metal dative bonds. However, theoretical
studies suggest that the lowest energy transitions in these complexes
are σ f σ* or possibly dπ f σ* transitions. See: Nakatsuji, H.; Hada,
M.; Kawashima, A. Inorg. Chem. 1992, 31, 1740. Rosa, A.; Ricciardi,
G.; Baerends, E. J .; Stufkens, D. J . Inorg. Chem. 1996, 35, 2886-2897.
(25) Davis, H. B. Ph. D. Thesis, Simon Fraser University, J an 1991.

Metal-Metal Bond Heterolysis in (Me3P)(OC)4OsW(CO)5
Organometallics, Vol. 16, No. 15, 1997 3433
Ta ble 1. Selected IR Da ta for
(Me3P )(OC)4OsW(CO)5 (1) a n d All P h otop r od u cts in
C6H6 a t Room Tem p er a tu r e
IR absorption ν(CO) bands,
(3)
compound
cm-1 (ꢀ, L mol-1 cm-1
)
(Me3P)(OC)4OsW(CO)5
2092 (340), 2041 (1970), 2007 (6980),
1912 (7520), 1894 (2490)
absorption in the IR spectrum close to the single ν(CO)
absorption of W(CO)6, it was difficult to measure the
absorbance of the latter band. The concentrations of
W(CO)6 produced in the various reactions are thus
subject to a larger than normal error.) The W(CO)6 is
believed to form by decomposition of some of the
[W(CO)5] to give CO, which then reacts with remaining
Os(CO)4(PMe3)
2058 (2110), 1974 (1360), 1930 (5300)
1881 (6870)
2125 (1390), 2050 (2660), 2004 (2390)
1976 (15500)
2071 (1730), 1981 (423), 1937 (10500)
2098 (s), 2026 (s), 1979 (s)
Os(CO)3(PMe3)(PPh3)
fac-Os(CO)3(PMe3)Cl2
W(CO)6
W(CO)5(PPh3)
fac-
Os(CO)3(PMe3)(CH2C6H5)Cl
fac-Os(CO)3(PMe3)(CH3)Cl
fac-Os(CO)3(PMe3)(CCl3)Cl
2097 (s), 2027 (s), 1982 (s)a
2119 (m), 2055 (m), 2010 (m)
W(CO)5 to give the observed product (e.g., eq 4).34
A
a
In hexane. See ref 38.
(4a)
(4b)
400 nm) in order to reduce the possibility of multiple
excitation pathways. (Multiple pathways generally
become more common as the exciting energy increases
in metal-metal-bonded complexes.9)
The results are most readily interpreted in terms of
a heterolytic cleavage of the metal-metal bond in 1 to
yield Os(CO)4(PMe3) and [W(CO)5] as the initial photo-
products. The unsaturated, 16-electron [W(CO)5] frag-
ment produced then reacts with PPh3 to give W(CO)5-
(PPh3).28 (The mechanism for the formation of the
similar pathway was proposed to account for the forma-
tion of W(CO)6 in the thermal decomposition of
[(Me)(OC)4FeW(CO)5-] (eq 5).26b A maximum yield of
83% of W(CO)6 would result if this were the only
mechanism of formation (see below, however).
(5)
minor product, W(CO)6, will be discussed later.)
A
control experiment showed that Os(CO)4(PMe3) (1.15 ×
10-2 M) is photochemically converted to Os(CO)3(PMe3)-
(PPh3) in the presence of PPh3 (1.52 × 10-1 M) under
the same reaction conditions (benzene solvent; λ > 400
nm) as reaction 1.29 This result suggests that Os(CO)3-
(PMe3)(PPh3) is formed predominantly in a secondary
photochemical reaction (eq 2).
As a control experiment, the thermal reaction of 1 (1.1
× 10-3 M) with PPh3 (6.0 × 10-2 M) was carried out in
deoxygenated C6H6 under N2 in the dark at both room
temperature and 40 °C. No peaks due to products were
observed in the IR spectrum of the control reaction at
room temperature until 27.5 h, and signals due to 1
were still clearly visible in the spectrum after 9 days.
At 40 °C, traces of Os(CO)4(PMe3) and W(CO)5(PPh3)
were detected by IR spectroscopy after 3 h, but the
absorptions of 1 were detectable up to 66 h. Even the
longest photochemical experiments with 1 described in
this paper were complete within 2 h, and thus the
contribution of the thermal decomposition of 1 to the
products can be considered negligible. Note that in the
thermal control W(CO)5PPh3 and Os(CO)4(PMe3) were
the sole carbonyl-containing products. There was no
evidence to suggest the presence of Os(CO)3(PMe3)-
(PPh3) or W(CO)6.
Exp er im en ts Exp lor in g th e P ossibility of Ho-
m olytic Clea va ge. Rea ction s w ith th e Ra d ica l
Tr a p C6H5CH2Cl. The products in eqs 1 and 3 might
possibly be the result of metal-metal bond homolysis
followed by fast electron transfer. To probe this pos-
sibility, a solution of 1 (1.47 × 10-3 M) in deoxygenated
C6H6 was irradiated (λ > 400 nm) in the presence of an
excess of benzyl chloride (1.47 × 10-2 M), an excellent
metalloradical trap, until all of 1 had disappeared (8
min).35 The photoproducts were identical to those in
eq 3 except the yields of W(CO)6 (57%) and Os(CO)4-
(PMe3) (43%) were lower; no other photoproducts were
observed. Furthermore, no bibenzyl was detected.
(Bibenzyl is the expected organic coupling product
resulting from the Cl atom abstraction from benzyl
(2)
Irradiation of W(CO)6 (3.55 × 10-3 M) and PPh3 (6.34
× 10-2 M) in deoxygenated C6H6 at room temperature
with the same wavelengths used in eq 1 (λ > 400 nm)
showed no formation of W(CO)5(PPh3) even after 30 min
of photolysis.30 The W(CO)5PPh3 product in eq 1 is
therefore not the result of a secondary reaction of
W(CO)6.31
The photolysis of 1 (9.08 × 10-4 M) in deoxygenated
benzene without any ligands present went according to
eq 3 over a 12 min period. The reaction was again
monitored by infrared spectroscopy, and the products
were identified by comparison to reported spectra (Table
1). In no instance was any W(CO)5(η2-C6H6) observed
in the IR spectra.33 (Because benzene has an intense
(28) (a) Dennenberg, R. J .; Darensbourg, D. J . Inorg. Chem. 1972,
11, 72 and references therein. (b) Hyde, C. L.; Darensbourg, D. J . Inorg.
Chem. 1973, 12, 1286.
(29) For Os(CO)4PMe3 in C6H6 at 400 nm, ꢀ400nm ) 20 L mol-1 cm-1
.
(30) For W(CO)6 in C6H6, ꢀ400nm ≈ 2 L mol-1 cm-1
.
(31) When the irradiation wavelength was changed to λ > 335 nm,
the following photochemical reactions were observed:32
(34) (a) Reference 9a, p 69. (b) Leventis, N.; Wagner, P. J . J . Am.
Chem. Soc. 1985, 107, 5807. (c) Albano, V. G.; Busetto, L.; Castellari,
C.; Monari, M.; Palazzi, A.; Zanotti, V. J . Chem. Soc., Dalton Trans.
1993, 3661.
(35) (a) Wrighton, M. S.; Ginley, D. S. J . Am. Chem. Soc. 1975, 97,
4246. (b) Laine, R. M.; Ford, P. C. Inorg. Chem. 1977, 16, 388. (c) Scott,
S. L.; Espenson, J . H.; Zhu, Z. J . Am. Chem. Soc. 1993, 115, 1789.
(32) See ref 9a, p 68, and ref 28.
(33) Tyler, D. R.; Petrylak, D. P. J . Organomet. Chem. 1981, 212,
389.

3434 Organometallics, Vol. 16, No. 15, 1997
Male et al.
chloride.) Continued irradiation (1 h) showed the for-
mation of small amounts of fac-Os(CO)3(PMe3)Cl2 (fa c-
2) and the loss of Os(CO)4PMe3. (fa c-2 is a new
molecule, and its characterization will be described in
a subsequent paper.36) A control irradiation (λ > 400
nm) of Os(CO)4PMe3 (2.17 × 10-3 M) in C6H6 in the
presence of an excess of benzyl chloride (2.09 × 10-1
M) for 3 h gave fa c-2 and fac-Os(CO)3(PMe3)(CH2C6H5)-
Cl.36,37
h.39 (A control reaction showed that irradiation of fa c-3
at λ > 335 nm for 2 h gave no photochemical reaction.
A slight thermal reaction did occur to give trace
amounts of fa c-2.)
No products that could be considered the result of
radical fragments containing tungsten were trapped
during the photolysis of 1 (λ > 400 nm) in the presence
of an excess of CCl4. The only photoproduct containing
tungsten detected by IR spectroscopy was W(CO)6.
Species such as [W(CO)5Cl]-,40 [W(CO)4(Cl)2],41 or
[W(CO)5Cl],42 which could result from the formation of
W-containing radical species, were not detected.
Rea ction s w ith CCl4. Complex 1 (2.18 × 10-3 M)
was irradiated (λ > 400 nm) in the presence of an excess
of CCl4 (5.68 × 10-2 M) in deoxygenated C6H6. The
reaction was monitored by infrared spectroscopy, which
showed the disappearance of the starting material and
the appearance of bands attributed to W(CO)6 (56%) and
fac-Os(CO)3(PMe3)Cl2 (fa c-2) (72%) (eq 6; Table 1). The
reaction was complete after 13 min.
Thermal control reactions of 1 with CCl4 in deoxy-
genated C6H6 in the absence of light at room tempera-
ture demonstrated that the same products are produced
as in the photochemical reaction (eq 6). The thermal
pathway, however, is much slower; no change was
detected in 1 by infrared spectroscopy until after 3.5 h.
The products observed when 1 is photolyzed in the
presence of CCl4 are therefore due to a photochemical
process alone and not a thermochemical pathway.43,44
(6)
To summarize, it is proposed that the initial product
produced in the photolysis of Os(CO)4(PMe3) in the
presence of CCl4 is fa c-3, which subsequently converts
to fa c-2. It is known that complexes with the M-CCl3
unit are unstable and readily convert to the correspond-
ing M-Cl derivative.45 That different products are
observed in the irradiation of 1 with CCl4 or Os(CO)4-
(PMe3) with CCl4 under similar conditions appears to
indicate that the [W(CO)5] fragment plays a key role in
the reactivity. J ust what that role is remains uncertain.
It is suggested that the W(CO)5 fragment may interact
with CCl4 to form a chlorinating reagent that reacts
differently from CCl4.46 Finally, there is no evidence
that metal radicals form during the irradiation of 1 in
C6H6 in the presence of an excess of the radical traps
C6H5CH2Cl or CCl4.47
In order to explain the formation of fa c-2, it was
hypothesized that irradiation of Os(CO)4(PMe3) in the
presence of CCl4 gave fa c-2. To test this hypothesis,
Os(CO)4(PMe3) (5.64 × 10-3 M) was irradiated (λ > 400
nm) in the presence of an excess of CCl4 (5.53 × 10-2
M) in deoxygenated C6H6. Infrared spectroscopic moni-
toring of the reaction, however, showed complete con-
version of Os(CO)4(PMe3) to a product thought to be
Os(CO)3(PMe3)(Cl)(CCl3) (fa c-3; ν(CO) ) 2119 (m), 2055
(m), and 2010 (m) cm-1 36
after 32 min of irradiation
)
(eq 7). The spectrum did not change in intensity after
a further 160 min in the dark at room temperature.
(7)
(39) Note that substantial amounts of W(CO)6 still remained after
this time.
(40) (a) Abel, E. W.; Butler, I. S.; Reid, J . G. J . Chem. Soc. 1963,
2068. (b) King, R. B. Adv. Organomet. Chem. 1964, 2, 157.
(41) Borowczak, D.; Szymanska-Buzar, T.; Zio´lkowski, J . J . J . Mol.
Catal. 1984, 27, 355.
(42) (a) Krausz, P.; Garnier, F.; Dubois, J . E. J . Am. Chem. Soc.
1975, 97, 437. (b) Krausz, P.; Garnier, F.; Dubois, J . -E. J . Organomet.
Chem. 1976, 108, 197.
The results above were puzzling because irradiation
of Os(CO)4(PMe3) in the presence of CCl4 led to fa c-3
(eq 7), while the irradiation of 1 (which forms Os(CO)4-
(PMe3)) led to fa c-2 (eq 6). These results prompted the
additional experiment of irradiating Os(CO)4(PMe3)
(2.02 × 10-3 M) in deoxygenated C6H6 with CCl4 (2.36
× 10-1 M) in the presence of W(CO)6 (1.99 × 10-3 M).
(Excitation wavelengths of λ > 335 nm were used in this
experiment to generate [W(CO)5]; recall W(CO)6 does
not react when it is irradiated with λ > 400 nm.) The
experiment was designed to determine if the [W(CO)5]
fragment promotes the formation of fa c-2 from fa c-3.
In fact, this appears to be the case. Infrared spectro-
scopic monitoring of the reaction showed the disappear-
ance of the bands due to Os(CO)4(PMe3) (after 2 min)
and the formation of fa c-3. With continued irradiation,
conversion of fa c-3 to fa c-2 was complete in about 1
(43) Thermal control reactions were carried out in which 1 (1.47 ×
10-3 M) was stirred in the dark with CCl4 (1.45 × 10-1 M) in
deoxygenated C6H6 under N2. No change in the infrared spectrum of
the reaction mixture was detected until 3.5 h, after which time trace
amounts of fa c-2 and W(CO)6 were detected in the solution. Even after
58 h at room temperature in the dark the reaction was far from
complete: bands due to 1 were clearly visible. In another thermal
control reaction, Os(CO)4(PMe3) (1.79 × 10-3 M) was stirred in the
dark with CCl4 (1.45 × 10-1 M) in deoxygenated C6H6 under N2. Bands
attributed to fa c-3 were observed in the IR spectrum of the solution
immediately after mixing. Continued monitoring of the reaction showed
the slow disappearance of Os(CO)4(PMe3), which was complete after
58 h. During this period all of fa c-3 had converted to fa c-2.
(44) That the radical trapping reagents CCl4 and C6H5CH2Cl reacted
differently with Os(CO)4(PMe3) may be attributed to the fact that CCl4
is considerably more oxidizing than benzyl chloride. (See: White, R.
C.; Buckles, R. E. J . Photochem. 1978, 8, 67.) Thus, oxidative addition
to Os(CO)3(PMe3) may simply be more energetically favorable in the
case of CCl4. Another possibility is that the C6H5CH2Cl reaction is less
efficient because two molecules are required to form the product.
(45) (a) Roper, W. R. J . Organomet. Chem. 1986, 300, 167. (b) Clark,
G. R.; Marsden, K.; Roper, W. R.; Wright, L. J . J . Am. Chem. Soc. 1980,
102, 1206. (c) Gallop, M. A.; Roper, W. R. Adv. Organomet. Chem. 1986,
25, 121.
(36) Male, J . L.; Einstein, F. W. B.; Leong, W. K.; Pomeroy, R. K.;
Tyler, D. R., submitted to J . Organomet. Chem.
(37) The fac-Os(CO)3(PMe3)(CH2C6H5)Cl (ν(C≡O) in benzene: 2098
(s), 2026 (s), and 1979 (s) cm-1) was identified by comparison to fac-
Os(CO)3(PMe3)(CH3)Cl38 (ν(C≡O) in hexane: 2097 (s), 2027 (s), and
1982 (s) cm-1).
(46) Irradiation of W(CO)6 in the presence of CCl4 has been proposed
to produce a variety of chlorinating agents, including phosgene. See,
for example, refs 41 and 42.
(38) Bellachioma, G.; Cardaci, G.; Macchioni, A.; Zanazzi, P. Inorg.
Chem. 1993, 32, 547-553.

Metal-Metal Bond Heterolysis in (Me3P)(OC)4OsW(CO)5
Organometallics, Vol. 16, No. 15, 1997 3435
Ta ble 2. Selected IR Da ta for th e Ir r a d ia tion of [Bu n N]2[W2(CO)10] (4) in CH3CN a t Room Tem p er a tu r e
compound
IR absorption ν(CO) bands, cm-1
IR media
ref
[Bun4N]2[W2(CO)10
W(CO)6
[Et4N][W(CO)5Cl]
]
1940 (m), 1891 (vs), 1788 (s)a
1975 (s)
CH3CN
CH3CN
KBr
CHCl3
CH3CN
KBr
this work, 49
this work, 49
40
2061 (w), 1904 (s), 1869 (m)
2061 (w), 1904 (s), 1869 (m)
2064 (w), 1919 (s), 1846 (m)
2064 (w), 1904 (s), 1868 (m)
2066 (w), 1922 (s), 1851 (m)
2085 (w), 1948 (vs), 1931 (s)
2075, 1940
2078 (w), 1939 (s)
2024 (m), 1900 (vs), 1842 (s)
1898 (s)
41
[Bun4N][W(CO)5Cl]
[Et4N][W(CO)5Br]
this work
40
this work
54a
54a
this work
54a
CH3CN
C6H14
W(CO)5(CH3CN)
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
cis-W(CO)4(CH3CN)2
trans-W(CO)4(CH3CN)2
W(CO)4(CH3CN)2
49
2020 (vw), 1898 (m), 1840 (w)
this work
a
Relative intensities of the IR absorption bands: vw ) very weak, w ) weak, m ) medium, s ) strong, vs ) very strong.
converted to W(CO)5(CH3CN) and trace amounts of
W(CO)4(CH3CN)2 (Figure 2). The products were identi-
(8)
fied by IR spectroscopy (Table 2).51 As in complexes
containing nondative metal-metal bonds, it is sug-
gested that irradiation of 455 leads to homolysis of the
W-W bonds to form [W(CO)5•-], and this radical
subsequently reacts with C6H5CH2Cl in an atom ab-
straction reaction to form [W(CO)5Cl-].
Because the secondary photolysis of W(CO)5Cl- is so
fast, it might be suggested that if W(CO)5Cl- were
produced in the reaction of 1 with benzyl chloride, then
it might not be observed. However, Os(CO)4(PMe3)Cl+
or Os(CO)3(PMe3)Cl2 would also be formed in the
reaction.56 Recall that the primary Os-containing prod-
uct was Os(CO)4PMe3, which upon secondary photolysis
slowly yielded fa c-2. These experiments therefore
strongly suggest that homolysis of the Os-W bond in 1
does not occur.
F igu r e 2. Irradiation (λ > 400 nm) of [Bun4N]2[W2(CO)10]
(4) (9.52 × 10-3 M) in CH3CN in the presence of an excess
of C6H5CH2Cl (9.99 × 10-1 M). The time scale of the
reaction was 0-1 min utilizing a 200 W high-pressure
mercury arc lamp. The species are identified as follows: ∆
) [W2CO10]
2-; ( ) [W(CO)5Cl]- (the initial photoproduct);
L ) W(CO)5(CH3CN) (product of secondary photochemis-
try); O ) W(CO)4(CH3CN)2 (product of secondary photo-
chemistry).
The photochemical reaction of 4 (9.8 × 10-3 M) was
also attempted in CH3CN with CCl4 (9.4 × 10-1 M) as
a radical trap. The CCl4, however, immediately reacted
with 4, without exposure to light. Infrared spectroscopy
P h otor ea ction s of [W2(CO)102-]. In order to probe
•-
whether CCl4 or C6H5CH2Cl could capture [W(CO)5
]
if it were to form, [Bun4N]2[W2(CO)10] (4) was irradiated
(λ > 400 nm) in CH3CN in the presence of these traps.48
Acetonitrile was used as a solvent because 4 is not
sufficiently soluble in C6H6 to give a satisfactory IR
spectrum. Irradiation (λ > 400 nm) of 4 (9.52 × 10-3
M) in deoxygenated CH3CN in the presence of an excess
of C6H5CH2Cl (9.99 × 10-1 M) caused the complete
disappearance of 4 within 10 s and the emergence of
W(CO)5Cl- and W(CO)5(CH3CN) (eq 8). Upon further
irradiation for 50 s, the W(CO)5Cl- was completely
(51) The yellow compound [Et4N][W(CO)5Cl] has its lowest energy
electronic absorption band maximum at 417 nm (ꢀ ) 3.10 × 103 L mol-1
cm-1) in methanol,52 and [Bun4N][W(CO)5Cl] would not be expected to
display vastly different absorption characteristics in CH3CN due to
the slight modification of the cation or the solvent. Irradiation at λ >
400 nm, in the aforementioned reaction, would therefore photoexcite
[W(CO)5Cl]-. Irradiation (λ > 436 nm) of [W(CO)5Br]- at 280 K in CO
saturated CHCl3 is known to yield W(CO)6.53 The analogous reaction
where CH3CN is the entering ligand into [W(CO)5] would, therefore,
appear plausible, particularly since the concentration of CH3CN is
greater than that of CO. The complex W(CO)5(CH3CN) could then
undergo further photochemical substitution to yield W(CO)4(CH3CN)2.
It is known that upon photolysis of W(CO)6, at higher energy
wavelengths, a mixture of W(CO)5(CH3CN) and W(CO)4(CH3CN)2 is
formed quickly.54
(52) McLean, R. A. N. J . Chem. Soc., Dalton Trans. 1974, 1568.
(53) (a) Dahlgren, R. M.; Zink, J . I. J . Chem. Soc., Chem. Commun.
1978, 863. (b) Dahlgren, R. M.; Zink, J . I. Inorg. Chem. 1977, 16, 3154-
3161.
(54) (a) Dobson, G. R.; Amr El Sayed, M. F.; Stolz, I. W.; Sheline, R.
K. Inorg. Chem. 1962, 1, 526. (b) Strohmeier, W. Angew. Chem., Int.
Ed. Engl. 1964, 3, 730.
(47) Experiments were also carried out with the radical trap 1,1,3,3-
tetramethylisoindoline-2-oxyl, TMIO (see: Beckwith, A. L. J .; Bowry,
V. W.; Moad, G. J . Org. Chem. 1988, 53, 1632). One of the advantages
of using TMIO is that when it interacts with a metal radical the TMIO
does not generate new radicals that can then undergo further
chemistry, unlike CCl4 or C6H5CH2Cl. Photolysis (λ > 400 nm) of 1
(1.53 × 10-3 M) in deoxygenated C6H6 at room temperature in the
presence of an excess of TMIO (6.86 × 10-3 M) gave W(CO)6 (22%)
and Os(CO)4(PMe3) (75%), which are identical to the products formed
in the photolysis of 1 in the absence of any other reagent and similar
to the initial products formed by irradiation of 1 and C6H5CH2Cl.
(55) Compound 4 has absorption band maxima at 347 nm (ꢀ ) 7.21
× 103 L mol-1 cm-1) and 390 nm (shoulder, ꢀ ) 4.84 × 103 L mol-1
cm-1) in CH3CN.49
•-
(48) It has been previously proposed that 4 cleaves to yield [W(CO)5
]
radicals upon photolysis.49,50
(49) (a) Silavwe, N. D.; Pan, X.; Tyler, D. R. Inorg. Chim. Acta 1988,
144, 123. (b) Silavwe, N. D.; Goldman, A. S.; Ritter, R.; Tyler, D. R.
Inorg. Chem. 1989, 28, 1231.
(56) Although the stability of this complex is unknown, Os(CO)4-
(PMe3)CH3+ 38 and Os(CO)5Cl+ 57 are known and are stable in solution.
(57) Rushman, P.; Van Buuren, G. N.; Shiralian, M.; Pomeroy, R.
K. Organometallics 1983, 2, 693-694.
(50) Phillips, J . R.; Trogler, W. C. Inorg. Chim. Acta 1992, 198, 633.

3436 Organometallics, Vol. 16, No. 15, 1997
Male et al.
Sch em e 1
showed the formation of [W(CO)5Cl]-, W(CO)6, and
W(CO)5(CH3CN) (Table 1).58
1. This section discusses several experiments designed
to probe how W(CO)6 is formed. As indicated in eq 1,
W(CO)6 was formed in about 8% yield even in the
presence of a large excess of PPh3 ([1]:[PPh3] 1:115,
[PPh3] ) 1.91 × 10-1 M).63 A possible mechanism for
the formation of W(CO)6 is that it results from CO
abstraction by W(CO)5 from Os(CO)4(PMe3) in the
geminate caged intermediate [Os(CO)4(PMe3), W(CO)5]
(Scheme 1).64 The high reactivity of W(CO)5 and its
ability to abstract ligands from other moieties has ample
precedent.26b,28,33,34 To probe this possibility, the pho-
tolysis (λ > 400 nm) of 1 (1.22 × 10-3 M) was repeated
in a more viscous medium of biphenyl (2.87 M) in C6H6
at 25 °C (absolute viscosity 1.2 cP as compared to C6H6
absolute viscosity 0.601 cP at 25 °C65,66) in the presence
of an excess of PPh3 (1.36 × 10-1 M). If W(CO)6 is
formed in a geminate cage reaction, then, in a more
viscous medium, cage escape should be hindered and
the yield of W(CO)6 should be increased. However, the
reaction in the more viscous solvent gave the same yield
of W(CO)6 (8%). The reaction was also carried out under
similar conditions but with mineral oil added to C6H6
to create the more viscous solvent (mineral oil 30.3%
v/v in C6H6 at 28 °C (absolute viscosity 1.1 cP)). When
this solvent system was used, the yield of W(CO)6 was
again 8%.67 These experiments strongly suggest that
a CO abstraction mechanism taking place in a caged
intermediate is not responsible for the formation of the
W(CO)6.
Meta l-Liga n d Bon d Dissocia tion . Irradiation (λ
> 400 nm) of 1 (2.96 × 10-3 M) in a CO-saturated C6H6
solution (7.4 × 10-3 M)61 at room temperature gave Os-
(CO)4(PMe3) (81%) and W(CO)6 (74%), analogous to eq
3. Recall that when 1 was irradiated (λ > 400 nm) in
the absence of any ligand, W(CO)6 was formed in 66%
yield (eq 3). The increased yield in the presence of CO
is similar to the increase in the amount of W product
formed in the presence of PPh3, where 89% tungsten-
containing products were detected (eq 1). The quantum
yield62 for disappearance in the photoreaction (λ ) 368
nm) of 1 in deoxygenated benzene under 1 atm of N2 at
23 ( 1 °C was determined to be 0.13 ( 0.01. Under
the same conditions but with 1 atm of CO, the quantum
yield increased to 0.28 ( 0.05. If the initial photoprocess
was loss of a CO or PMe3 ligand, then the initial process
would be inhibited and the quantum yield for the
disappearance of 1 would be reduced in the presence of
an excess of free ligand such as CO or PPh3. That the
quantum yield under CO increases strongly suggests
that the initial step in the photolysis is not M-CO bond
cleavage. Furthermore, irradiation of 1 in deoxygenated
C6H6 in the presence of excess PPh3 (eq 1) did not yield
[OsW(CO)8(PMe3)(PPh3)], which might be expected if
M-CO dissociation occurred. These results further
support the view that heterolysis of the Os-W bond
results from the excitation of 1. In the presence of CO,
the recombination of the fragments from the heterolysis
would be inhibited because the [W(CO)5] fragment
would be trapped by CO rather than recombine with
Os(CO)4PMe3. This would increase the quantum yield
for the disappearance of 1, as observed.
Another possible origin of the W(CO)6 is that it forms
in a reaction of W(CO)5 with the CO that is liberated
during the secondary photolysis of Os(CO)4(PMe3). To
(60) Briegleb, G. Angew. Chem., Int. Ed. Engl. 1964, 3, 617.
(61) Fogg, P. G. T.; Gerrard, W. Solubility of Gases in LiquidssA
Critical Evaluation of Gas/ Liquid Systems in Theory and Practice;
J ohn Wiley & Sons: Chichester, 1991; pp 2, 277.
F or m a tion of W(CO)6. The results discussed to this
point are all consistent with a Os-W heterolytic pho-
toprocess. However, a heterolytic pathway does not
necessarily account for the formation of W(CO)6 in eq
(62) Typical concentrations of 1 for the quantum yield measurements
were 2.90 × 10-4 M. The quantum yields were measured at 368 nm in
order to obtain the maximum absorption possible for species 1 while
seeking to avoid problems with absorptions due to products (for W(CO)6
at 368 nm, ꢀ ) 220 L mol-1 cm-1, for Os(CO)4(PMe3) at 368 nm, ꢀ ) 98
L mol-1 cm-1 in C6H6 at room temperature). Less than 15% of the
bimetallic complex 1 was allowed to react in these measurements. Note
that the quantum yields reported in this paper differ from those
reported in ref 23 by a factor of 2.23. We used an incorrect quantum
yield for the Aberchrome 540 actinometer in our prior calculations.
(63) The yield of W(CO)6 was essentially unchanged at 7% when
the ratio of the concentrations was altered ([1]:[PPh3] ) 1:10; [PPh3]
) 1.83 × 10-2 M).
(58) It has been reported that the electrochemical oxidation of 4 in
THF in the presence of an excess of CH2Cl2 with
a supporting
electrolyte of tetrabutylammonium perchlorate gave W(CO)5(THF) and
[W(CO)5Cl]- as the products.50 The electron affinity of CCl4 (2.12 eV,
first E1/2 ) -0.25 V in DMF with [Bun4N][Br] as supporting electrolyte)
is much greater than that of CH2Cl2 (1.32 eV, first E1/2 ) -2.14 V in
DMF with [Bun4N][Br] as supporting electrolyte).59,60 Thus if there were
an electron transfer reaction and the process involved an oxidation/
reduction couple, CCl4 would be more reactive than CH2Cl2. One
possible explanation for the products observed would be that 4 forms
a charge-transfer complex with CCl4 in the polar solvent CH3CN in
the dark which yields [W(CO)5Cl]- and W(CO)5(CH3CN), analogous
to the electrochemical reaction of 4 and CH2Cl2 mentioned above. It
was shown that, after the photolysis of W(CO)6 in CH3CN, if the
reaction mixture was allowed to remain in the dark at room temper-
ature, then substantial quantities of W(CO)6 were re-formed.54a This
implies that an equilibrium exists between the W(CO)6 and W(CO)5-
(CH3CN) species, and this may explain the presence of W(CO)6 in the
reaction of 4 with CCl4.
(64) Covert, K. J .; Askew, E. F.; Grunkemeier, J .; Koenig, T.; Tyler,
D. R. J . Am. Chem. Soc. 1992, 114, 10446.
(65) Atkins, P. W. Physical Chemistry, 5th ed.; Oxford University
Press: Oxford, 1994; p 833, C27.
(66) Lide, D. R. Handbook of Chemistry and Physics, 71st ed.; CRC
Press: Boston, 1990; pp 6-142.
(67) In the latter reaction, concerns arose over the possibility of
preferential solvation of the photoproducts due to the differences in
polarity and shape of mineral oil and C6H6. This problem was believed
to be less of a problem with the biphenyl/C6H6 solvent system because
of the similar natures of the solvent and the viscosity enhancer.
(59) Wawzonek, S.; Duty, R. C. J . Electrochem. Soc. 1961, 108, 1135.

Metal-Metal Bond Heterolysis in (Me3P)(OC)4OsW(CO)5
Organometallics, Vol. 16, No. 15, 1997 3437
test this possibility, Os(CO)4(PMe3) (1.81 × 10-3 M) and
W(CO)5(PPh3) (1.58 × 10-3 M) were irradiated (λ > 400
nm) in benzene at room temperature in the presence of
PPh3 at a concentration similar to that used in reaction
1 (1.85 × 10-1 M). (W(CO)5(PPh3) was used as the
source of W(CO)5 in this experiment rather than W(CO)6
because W(CO)6 is the potential product of the reaction.
Note that irradiation of W(CO)5(PPh3) at λ ) 405 nm is
claimed to generate W(CO)5.53 Also, recall that W(CO)6
will not react photochemically with PPh3 for λ > 400
nm.) The reaction was monitored by infrared spectros-
copy, which showed the disappearance of the starting
material bands and the appearance of new bands at
1881 cm-1 assigned to Os(CO)3(PMe3)(PPh3) (Table 1)
and at 2018, 1920, and 1898 cm-1 assigned to W(CO)4-
(PPh3)2.68 There was no band at 1976 cm-1, however,
that would indicate the formation of W(CO)6. (If small
amounts of W(CO)6 were formed, it would likely have
been detected because the molar absorption coefficient
of the absorption at 1976 cm-1 is large; Table 1.) This
result suggests that, under the reaction conditions in
eq 1, CO is not competitive with PPh3 in its reactions
with W(CO)5.
Sch em e 2
produce Os(CO)4(PMe3) and [W(CO)5]; that is, it could
also account for the major products formed in the
photolysis.
Pathway b in Scheme 2 cannot account for the 66%
yield of W(CO)6 that is formed in the photolysis of 1 in
the absence of an incoming ligand (eq 3). (This result
can be stated with certainty because W(CO)6 does not
absorb light at the wavelengths used in this experiment.
Thus, W(CO)6 yields would be 66% in the presence of
ligands if pathway b were followed exclusively.) The
major production of W(CO)6 (eq 3) is believed to be from
the reaction of [W(CO)5] with CO arising from partial
decomposition of [W(CO)5]26b (eq 4) or by [W(CO)5]
abstracting a CO from other species present.34
Other possibilities for the formation of W(CO)6 are
based on the supposition of two different excited states.
There are two isomers of (Me3P)(OC)4OsW(CO)5 in
solution,69 one with the phosphine ligand trans to the
metal-metal bond (major isomer) and the other with
the phosphine ligand cis to the metal-metal bond
(minor isomer).7,25 The isomers are in dynamic equi-
The relaxed excited state in Scheme 2 might have two
bridging CO ligands rather than one. A similar transi-
tion state species, although containing an Os-W dative
metal-metal bond, was proposed to account for the
nonrigidity and thermal decomposition products of
(R3P)(OC)4OsM(CO)5 (M ) Cr, Mo, W).7,25 There are
again two decomposition pathways for a species with
two bridging CO ligands: Symmetric cleavage of the Os-
(µ-CO)2W unit (pathway a) would yield Os(CO)4(PMe3)
and [W(CO)5] and asymmetric cleavage (pathway b)
would give W(CO)6 and [Os(CO)3(PMe3)]. As before, the
formation of W(CO)6 by route b would be independent
of the presence of PPh3 and viscosity of the medium.
Note that there are precedents for the asymmetric
cleavage of a bimetallic species with two bridging CO
ligands in both thermal70 and photochemical reactions.71
Con clu sion s. The photochemical reactions of 1 in
C6H6 can be interpreted in terms of the heterolytic
cleavage of the metal-metal bond. Irradiation (λ > 400
nm) of 1 in C6H6 in the presence of an excess of PPh3
yielded predominantly Os(CO)4(PMe3), W(CO)5(PPh3),
and a secondary photochemical product, Os(CO)3(PMe3)-
(PPh3). Attempts to trap radicals that would be pro-
duced by homolytic cleavage of the metal-metal bond
were unsuccessful. Although this latter result does not
eliminate a homolytic route, in conjunction with the
PPh3 experiments, it strongly suggests that homolysis
does not occur. Quantum yield determinations for the
disappearance of 1 in C6H6 under N2 and under CO
demonstrated that the quantum yield increased in the
presence of CO. This result provides evidence that
metal-ligand photodissociation is not the reaction
pathway in the photolysis of 1.
librium in solution at room temperature.7,25 Irradiation
of the two isomers of 1 would produce two slightly
different Franck-Condon excited states. The two
Franck-Condon states could relax differently, one to
give the major products (i.e., Os(CO)4(PMe3) and
[W(CO)5]) and the other to give [(Me3P)(OC)3Os(µ-
CO)W(CO)5] (Scheme 2). The intermediate with the
bridging CO ligand could then cleave to yield W(CO)6
and [Os(CO)3(PMe3)] (route b, Scheme 2). The latter
unsaturated complex would then react to form [Os(CO)3-
(PMe3)(PPh3)]. This mechanism is consistent with the
results from the reaction carried out in a more viscous
medium because a caged pair is not required. It should
be pointed out that the singly bridged intermediate
could cleave the other way (route a, Scheme 2) to
(68) (a) Hillhouse, G. L.; Haymore, B. L. Inorg. Chem. 1987, 26,
1876. (b) Darensbourg, D. J .; Bischoff, C. J . Inorg. Chem. 1993, 32,
47.
(69) The ratio of the isomers was determined by the integration of
the two sets of doublets in the 1H NMR spectrum in C6D6 over the
range 6-26 °C. At 26 °C, the 1H NMR spectrum of 1 showed a minor
isomer with δ ) 0.83 (d, Me, 9H, 2J P-H ) 10.2 Hz) and a major isomer
2
with δ ) 0.63 (d, Me, 9H, J P-H ) 10.3 Hz). The major:minor isomer
Exp er im en ta l Section
ratio was 2.6:1.0. At 6 °C, the major:minor isomer ratio was 2.7:1.0.
There seems, therefore, to be little variation in the ratio of the isomers
in the temperature range 6-26 °C, the temperature range expected
during the photolysis of 1. For this reason, the determination of the
potentially different molar absorption coefficients of the two isomers
of 1 in this solvent was not possible. Carrying out the studies at a
higher temperature would increase the probability of thermal reactions
occurring in conjunction with photochemical reactions.
Gen er a l Meth od s. All manipulations were performed
under nitrogen or argon by using standard Schlenk, drybox,
(70) Leonhard, K.; Werner, H. Angew. Chem., Int. Ed. Engl. 1977,
16, 649.
(71) Zhang, S.; Brown, T. L. J . Am. Chem. Soc. 1993, 115, 1779.

3438 Organometallics, Vol. 16, No. 15, 1997
Male et al.
or vacuum line techniques, unless stated otherwise. Drybox
manipulations were carried out using a Vacuum Atmospheres
HE-493 Dri-Lab with attached Dri-Train.
had previously been calibrated by using both a thermocouple
and a method based on the chemical shifts of methanol.77 The
temperature values were found to be accurate to (1 °C.
Photochemical reactions were carried out with an Oriel 200
W or an Oriel 100 W high-pressure mercury arc lamp with a
10 cm water filter with quartz windows to reduce IR emissions.
Corning glass filters were used for broad band irradiations:
CS 3-73 (λ > 420 nm), CS 3-74 (λ > 400 nm), CS 3-75 (λ > 365
nm), CS 4-72 (λ > 335 nm). Solutions for mechanistic
experiments were scrupulously protected from extraneous
light. Quantum yield determinations were carried out with
an Oriel 200 W high-pressure mercury arc lamp coupled with
a Beckman DU monochromator and a Merlin radiometer
system, model 70100 (Oriel Corp.).78 Light intensity was
determined by actinometry using Aberchrome 540 in toluene
(φ368 ) 0.20).79 The cuvettes used for the quantum yield
determinations were 1 cm in pathlength and had freeze-
pump-thaw side arms. The solutions (of 4.00 mL volume)
were stirred using magnetic stir bars during the irradiation.
The quantum yields at 368 nm (lamp intensities were typically
Ia ) (2.28 ( 0.01) × 10-10 einsteins s-1) were determined by
the initial (<15%) rates of disappearance of the UV-vis
absorption band maximum at 368 nm in complex 1. The
cuvettes were maintained at 23 ( 1 °C; a flow of compressed
air was passed through the cell holder during the photolysis
to prevent warming and thermal reaction.64 All quantum
Ma ter ia ls. Tetrahydrofuran was freshly distilled from
potassium benzophenone ketyl, hexanes and benzene were
distilled from potassium, dichloromethane and acetonitrile
were distilled from calcium hydride, and carbon tetrachloride
was distilled from P2O5 twice and used immediately with
minimum exposure to light. Benzyl chloride was refluxed with
MgSO4 and then fractionally distilled under reduced pressure;
the middle fraction was collected. Triphenylphosphine was
recrystallized from hexanes and stored under nitrogen in the
dark. TMIO was freshly sublimed at 35 °C and 1 × 10-4
mmHg prior to usage.72 Bibenzyl was recrystallized from 30-
60 °C petroleum ether; mineral oil (345/350 viscosity; specific
gravity at 25 °C, 0.865-0.925) was stirred with sodium
overnight and then filtered under N2 and finally subjected to
three freeze-pump-thaw cycles prior to usage. Carbon
monoxide gas (Linde/Union Carbide, C. P. grade) was used as
provided. The complexes (Me3P)(OC)4OsW(CO)5, (1),7 Os(CO)3-
(PMe3)(PPh3),25 and W(CO)5(PPh3)7,26,73 were prepared accord-
ing to the literature and recrystallized from C6H14/CH2Cl2. The
mononuclear complex Os(CO)4(PMe3)74 was prepared according
to the literature method except that a temperature of 360 (
24 °C was employed.74 [Bun4N]2[W2(CO)10] (4) was prepared
by a literature route.75,76
yields were corrected with
a linear correction factor for
In str u m en ta tion . Infrared spectra were measured by
using a Bomem Michelson model 120 FT-IR instrument, a
Nicolet 5DXB FT-IR spectrometer, a Nicolet Magna 550 FT
IR spectrometer, or a Perkin-Elmer 983G IR spectrophotom-
eter. Samples were prepared as solutions usually in NaCl
(path length 0.0965 or 0.442 mm) or CaF2 (path length 0.116
or 0.105 mm) cells. For mechanistic work, a known amount
of solid was dissolved in the deoxygenated solvent containing
known quantities of other reagents, if appropriate, and
transferred to the IR cell via a gastight syringe.
nonabsorption. Kinematic viscosities of the solutions were
measured with a calibrated Cannon-Fenske viscometer and
corrected to absolute viscosities. Yields of the products were
determined from the absorbances of appropriate CO stretches
in the infrared spectra of the solutions. All products were
independently synthesized to obtain their extinction coef-
ficients.
The preparation and characterization of fac-Os(CO)3(PMe3)-
(Cl)2 (fa c-2) and Os(CO)3(PMe3)(CCl3)(Cl) (fa c-3) will be
discussed in a forthcoming paper.36
Solution UV-vis spectra were measured with a Perkin-
Elmer Lambda 6 series UV-vis spectrophotometer, a (Varian)
Cary 210 UV-vis spectrophotometer, or a Hewlett-Packard
8452A diode array spectrophotometer. All NMR data were
recorded at the temperatures specified on a Bruker AMX 400.
Ack n ow led gm en t is made to the U.S. National
Science Foundation and the Natural Sciences and
Engineering Research Council of Canada for financial
support of this research. We also thank Dr. H. B. Davis
for assistance in the initial stages of this study.
1
The variable temperature H NMR spectra of 1 in C6D6 at 6,
11, 16, 21, and 26 °C were allowed 20 min to equilibrate at
the desired temperature in the 5 mm tube prior to collection
of the data. The temperature controller of the spectrometer
Su p p or tin g In for m a tion Ava ila ble: Nine figures show-
ing the infrared spectroscopic changes for the reactions
reported herein (4 pages). Ordering information is given on
any current masthead page.
(72) (a) Griffiths, P. G.; Moad, G.; Rizzardo, E.; Solomon, D. H. Aust.
J . Chem. 1983, 36, 397. (b) See also ref 47.
(73) (a) Matthews, C. N.; Magee, T. A.; Wotiz, J . H. J . Am. Chem.
Soc. 1959, 81, 2273. (b) Magee, T. A.; Matthews, C. N.; Wang, T. S.;
Wotiz, J . H. J . Am. Chem. Soc. 1961, 83, 3200.
(74) Martin, L. R.; Einstein, F. W. B.; Pomeroy, R. K. Inorg. Chem.
1985, 24, 2777.
OM970275W
(77) Van Geet, A. L. Anal. Chem. 1970, 42, 679.
(78) Lindfors, B. E.; Nieckarz, G. F.; Tyler, D. R.; Glenn, A. G. J .
Photochem. Photobiol. 1996, 94, 101.
(75) Ungurenasu, C.; Palie, M. J . Chem. Soc., Chem. Commun. 1975,
388.
(76) Schwindt, M. A.; Lejon, T.; Hegedus, L. S. Organometallics
1990, 9, 2814.
(79) Heller, H. G.; Langan, J . R. J . Chem. Soc., Perkin Trans. 2 1981,
341.