Thermolysis of the osmium-antimony clusters Os3(CO) 11(SbMe2Ar): Higher nuclearity clusters and arrested ortho metalation

Article abstract of DOI:10.1021/om0508132

Thermolysis of the clusters Os₃(CO)₁₁(SbMe ₂Ar) (1, where Ar = Ph (b), o-tolyl (c), p-tolyl (d)), in refluxing octane gave the clusters Os₃(μ-SbMe₂)(μ-H) (μ₃η²-C₆H

Full text of DOI:10.1021/om0508132

250
Organometallics 2006, 25, 250-259
Thermolysis of the Osmium-Antimony Clusters
Os3(CO)11(SbMe2Ar): Higher Nuclearity Clusters and Arrested
Ortho Metalation
Kiat Hwa Chan, Weng Kee Leong,* and Kar Hang Garvin Mak
Department of Chemistry, National UniVersity of Singapore, Kent Ridge, Singapore 119260
ReceiVed September 20, 2005
Thermolysis of the clusters Os3(CO)11(SbMe2Ar) (1, where Ar ) Ph (b), o-tolyl (c), p-tolyl (d)), in
refluxing octane gave the clusters Os3(µ-SbMe2)(µ-H)(µ3,η2-C6H3R)(CO)9 (2, where R ) H (b), Me (c)),
in good yields; both tolyl derivatives gave the same 4-substituted phenylene cluster. Among the minor
products also isolated from the thermolysis of 1b were the higher nuclearity clusters Os6(µ3-SbMe)-
(µ3,η2-C6H4)(CO)20 (4) and Os6(µ4-Sb)(µ-SbMe2)(µ-H)(µ3,η2-C6H4)2(CH3)(CO)16 (5), which contains an
Os-CH3 moiety. Thermolysis of the 3,5-xylyl analogue Os3(CO)11{SbMe2(3,5-Me2C6H3)} (1f), on the
other hand, gave the unusual species Os3(µ-SbMe2)(µ3,η1:η2:η2-C6H3Me2)(CO)9 (6), in which ortho
metalation of the xylyl ring has apparently been arrested.
C6H4)(CO)9 (2a) in fairly good yield,7 and this has allowed us
to investigate the novel reactivity of this cluster.8 As part of
our investigations into the chemistry of this cluster, we had
sought to understand its mode of formation and fluxional
behavior in comparison with those of the phosphorus and arsenic
analogues. We wish to report the results of our investigations
here.
Introduction
Triosmium clusters containing a µ3-phenylene ligand have
been known from the very earliest work of Nyholm.1 These
clusters are commonly synthesized via thermolysis of clusters
containing ligands with P-Ph,1,2 As-Ph,3 or S-Ph bonds.4 The
mode of coordination of the µ3-phenylene group to the metal
framework often depends on the other ligands that are on the
cluster. The most common bonding mode is still the µ3,η2 mode,
in which the phenylene acts as a four-electron donor.5 These
phenylene clusters have also been shown to be fluxional via
ring flipping and hydride migration.6 In addition, the mech-
anism for the formation of the phenylene clusters for the
phosphine and arsine systems has been thoroughly investi-
gated by Deeming.2a,6 More recently, we have reported that
thermolysis of the cluster Os3(CO)11(SbPh3) (1a) led initially
to Sb-C cleavage to afford the cluster Os3(µ-H)(µ-SbPh2)(µ3,η2-
Results and Discussion
Deeming et al. have demonstrated the use of alkyl substituents
on phosphine or arsine to control the thermolysis of the clusters
Os3(CO)11(EPh2R) and Os3(CO)11(EPhR2), where E ) P, As
and R ) alkyl.6,9 The central idea was that the P-alkyl or As-
alkyl bonds were less susceptible to thermolytic cleavage. We
thus adopted the same idea in the hope of minimizing the
formation of higher nuclearity clusters and thereby increase the
yield of the initial benzyne product. Toward this goal, we have
synthesized a number of stibine monosubstituted clusters
Os3(CO)11(SbR3) (1, where R3 ) Me2Ph (b), MePh2 (c), Me2[o-
tolyl] (d), Me2[p-tolyl] (e), SbMe2[3,5-Me2C6H3] (f), Me3 (g)).
The general reaction scheme employed is depicted in Scheme
1. The syntheses of the stibines were adapted from earlier
published methods10 and utilized Grignard reactions and a solid-
state metathesis between SbCl3 and SbAr3.11 The final step
involved the use of the lightly stabilized cluster Os3(CO)11-
(NCCH3).12 The identities of 1b and 1d have also been
confirmed by a single-crystal X-ray crystallographic study; the
ORTEP plot showing the molecular structure of 1b is given in
Figure 1.
(1) (a) Bradford, C. W.; Nyholm, R. S. J. Chem. Soc., Chem. Commun.
1967, 384. (b) Bradford, C. W.; Nyholm, R. S. J. Chem. Soc., Dalton Trans.
1973, 529.
(2) (a) Brown, S. C.; Evans, J.; Smart, L. E. J. Chem. Soc., Chem.
Commun. 1980, 1021. (b) Deeming, A. J.; Kabir, S. E.; Powell, N. I.; Bates,
P. A.; Hursthouse, M. B. J. Chem, Soc., Dalton Trans. 1987, 1529. (c)
Cullen, W. R.; Chacon, S. T.; Bruce, M. I.; Einstein, F. W. B.; Jones, R.
H. Organometallics 1988, 7, 2273. (d) Deeming, A. J.; Marshall, J. E.;
Nuel, D.; O’Brien, G.; Powell, N. I. J. Organomet. Chem. 1990, 384, 347.
(e) Cullen, W. R.; Rettig, S. T.; Zhang, H. Organometallics 1991, 10, 2965.
(f) Cullen, W. R.; Rettig, S. T.; Zhang, H. Organometallics 1992, 11, 928.
(g) Cullen, W. R.; Rettig, S. T.; Zhang, H. Organometallics 1992, 11, 1000.
(3) (a) Arce, A. J.; Deeming, A. J. J. Chem. Soc., Dalton Trans. 1982,
1155. (b) Johnson, B. F. G.; Lewis, J.; Massey, A. D.; Braga, D.; Grepioni,
F. J. Organomet. Chem. 1983, 251, 261. (c) Jackson, P. A.; Johnson, B. F.
G.; Lewis, J.; Massey, A. D.; Braga, D.; Gradella, C.; Grepioni, F. J.
Organomet. Chem. 1990, 391, 225. (d) Cullen, W. R.; Rettig, S. T.; Zhang,
H. Organometallics 1993, 12, 1964.
(4) Adams, R. D.; Katahira, D. A.; Yang, L. W. Organometallics 1982,
1, 235.
(5) Johnson, B. F. G.; Lewis, J.; Whitton, A. J.; Bott, S. J. J. Organomet.
Chem. 1990, 389, 129.
(7) Leong, W. K.; Chen, G. Organometallics 2001, 20, 2280.
(8) (a) Deng, M.; Leong, W. K. Organometallics 2002, 21, 1221. (b)
Chen, G.; Deng, M.; Lee, C. K.; Leong, W. K. Organometallics 2002, 21,
1227.
(6) (a) Deeming, A. J.; Nyholm, R. S.; Underhill, M. J. Chem. Soc.,
Chem. Commun. 1972, 224. (b) Deeming, A. J.; Kimber, R. E.; Underhill,
M. J. Chem., Soc., Dalton Trans. 1973, 2589. (c) Deeming, A. J.; Rothwell,
I. P.; Hursthouse, M. B.; Backer-Dirks, J. D. J. J. Chem, Soc., Dalton Trans.
1981, 1879. (d) Arce, A. J.; Deeming, A. J. J. Chem. Soc., Dalton Trans.
1982, 1155. (e) Deeming, A. J.; Kabir, S. E.; Powell, N. I.; Bates, P. A.;
Hursthouse, M. B. J. Chem, Soc., Dalton Trans. 1987, 1529.
(9) Cooksey, C. J.; Deeming, A. J.; Rothwell, I. P. J. Chem. Soc., Dalton
Trans. 1981, 1718.
(10) Pfeiffer, P.; Heller, I.; Pietsch, H. Chem. Ber. 1904, 37, 4620.
(11) (a) Nunn, M.; Sowerby, D. B.; Wesolek, D. M. J. Organomet. Chem.
1983, 251, C45. (b) Breunig, H. J.; Althaus, H.; Ro¨sler, R.; Lork, E. Z.
Anorg. Allg. Chem. 2000, 626, 1137.
(12) Nicholls, J. N.; Vargas, M. D. Inorg. Synth. 1989, 28, 232.
10.1021/om0508132 CCC: $33.50 © 2006 American Chemical Society
Publication on Web 11/17/2005

Thermolysis of Osmium-Antimony Clusters
Organometallics, Vol. 25, No. 1, 2006 251
Figure 1. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 1b.
Scheme 1
SbCl3 + 3ArMgBr f SbAr3
Figure 2. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 2b.
2SbCl3 + SbAr3 f 3SbArCl2
3SbArCl2 + MeMgBr f 3SbArMe2
°C) for a longer period of time in the hope that the production
of 4 and 5, which required Sb-Me bond cleavage, would be
decreased and thereby the yield of 2b increased. Unfortunately,
Os3(CO)11(NCCH3) + SbR3 f Os3(CO)11(SbR3)
1
IR and H NMR monitoring indicated that a steady state was
Thermolysis of 1b in refluxing octane indeed afforded the
phenylene cluster Os3(µ-SbMe2)(µ-H)(µ3,η2-C6H4)(CO)9 (2b)
in a yield of 70%. Another cluster, Os3(µ-SbMe2)(µ-H)(µ,η2-
C6H4)(CO)10 (3), and two higher nuclearity clusters, Os6(µ3-
SbMe)(µ3,η2-C6H4)(CO)20 (4) and Os6(µ4-Sb)(µ-SbMe2)(µ-
H)(µ3,η2-C6H4)2(CH3)(CO)16 (5), were also isolated together
with Os3(CO)12 (7) (Scheme 2). All the new clusters, viz. 2b
and 3-5, have been characterized completely, including by
single-crystal X-ray crystallographic studies. The ORTEP plots
showing their molecular structures are given in Figures 2-5,
respectively.
Cluster 3 is a CO adduct of 2b; it is structurally reminiscent
of the group 15 and isonitrile adducts of 2a.13 Thermolysis of
3 gave 2b, albeit in low yield, together with other as yet
unidentified products, thus pointing to 3 as the precursor to 2b.
Cluster 4 is the methyl analogue of the previously reported
Os6(µ3-SbPh)(µ3,η2-C6H4)(CO)20 (4a), obtained from the ther-
molysis of 1a. Similarly, 5 is closely related to the previously
reported cluster Os6(µ4-Sb)(µ-SbPh2)(µ-H)(µ3,η2-C6H4)2(Ph)-
(CO)16.7 Cluster 2b is stable up to about 125 °C, whereas the
Sb-Ph bond in 2a will cleave at about 90 °C. We therefore
attempted the thermolysis of 1b at a lower temperature (100
reached, i.e., the signals due to 1b and 2b remained the same,
after 1 h. That 7 was also formed suggested that it could
represent the final stage of some eliminated species, and indeed
when 1b was thermolyzed with a small amount of 7 added,
much of 1b remained unreacted. The yields with respect to
reacted 1b were about 85% and 12% for 2b and 7, respectively;
formation of 4 and 5 was not observed. This observation
indicated that the formation of the clusters 2-5 probably
required decarbonylation of 1b as the rate-determining step.
The formation of 4 and 5 required Sb-CH3 bond cleavage.
This is in contrast to that observed for the phosphorus and
arsenic systems; thermolysis of Os3(CO)11(PR3) (where R )
Me, Et) afforded products from C-H bond activation,14 while
thermolysis of arsenic analogues such as Os3(CO)11(AsMe2Ph)
afforded products resulting from metal-metal bond cleavage.6b
This difference in their behavior is probably ascribable to the
weaker Sb-C bond. As in the case of 1a, the clusters 4 and 5
are probably derived from further reaction of 2b. Thus, 4 is
formally derived from 2b via the loss of CH4 and addition of
an “Os3(CO)11” unit, while 5 is formally derived from the
coupling of two units of 2b via the loss of CH4 and two carbonyl
Scheme 2a
a Percent yields are with respect to reacted 1b.

252 Organometallics, Vol. 25, No. 1, 2006
Chan et al.
Figure 6. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 2c.
was very slow. Analysis of the gas space above the solution by
IR and GC indicated that CH4 was indeed liberated; no C2H6
was detectable by IR, GC, or EI-MS.
Figure 3. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for molecule A of 3. Selected bond
lengths (Å) and angles (deg) (molecules A, B): Os(1)-Os(2) )
3.1614(7), 3.1611(7); Os(2)-Os(3) ) 2.9548(7), 2.9649(7); Os(1)-
Sb(4) ) 2.6430(10), 2.6820(12); Os(3)-Sb(4) ) 2.6817(9),
2.6773(10); Os(1)-C(1) ) 2.152(12), 2.116(13); Os(2)-C(2) )
2.142(12), 2.139(13); C(1)-C(2) ) 1.384(18), 1.39(2); Os(2)-
Os(1)-Sb(4) ) 77.32(2), 78.10(3); Os(1)-Os(2)-Os(3) )
89.130(18), 89.39(2); Os(2)-Os(3)-Sb(4) ) 80.52(2), 81.78(3);
Os(1)-Sb(4)-Os(3) ) 107.51(3), 107.11(4).
The mean bond dissociation energies of the Sb-R bond in
SbPh3 and SbMe3 (58.3 and 51.5 kcal mol-1, respectively)15
suggest that if the Sb-R bond cleavage was through a radical
pathway, then the thermolysis of 1a should have been less facile
and that for 1g the most facile. Instead, thermolysis of 1g was
difficult; only two products in extremely low yields were
obtained, one of which has been characterized with the
formulation “Os3SbMe3(CO)10” (8), which is presumably related
structurally to the previously reported cluster Os3(µ-SbPh2)(Ph)-
(CO)9(PPh3).8b That no C2H6 was detected in the thermolysis
also pointed to a nonradical pathway for the elimination of CH4.
This was also further corroborated by the failure to detect
TEMPO-CH3 when 2b was thermolyzed in the presence of the
radical trap TEMPO (2,2,6,6-tetramethyl-1-piperidinyloxy radi-
cal) in d20-nonane.16
Working on the hypothesis that the formation of 4 therefore
involved concerted CH4 elimination, coupled with addition of
an “Os3(CO)11” unit, which may possibly result from loss of
SbMe2Ph from 1b either simultaneously or sequentially (Scheme
3), we thermolyzed 1c to obtain the pair of isomeric products
1
2c and 2c′. The H NMR spectrum of the reaction mixture
showed a 1:1 ratio of the two isomers. The isomers were not
separable by TLC, but crystallization tended to give crystals of
two slightly different colors (yellow and orange), which were
separated by hand. Unfortunately, this did not turn out to be an
absolutely foolproof means to distinguish between the two; at
best, we managed to obtain samples that were either more of
one or the other and hence allowed segregation of the 1H
resonances into two sets. Crystallographic analysis of one of
the orange crystals (2c) established its molecular structure as
that shown in Figure 6.
Figure 4. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 4.
As mentioned above, there was ambiguity as to the assign-
ments of the 1H resonances to 2c (that with the crystal structure
established) and 2c′. Nevertheless, we have made tentative
assignments for the 1H NMR resonances of 2a, 2b, 2c, and 2c′,
particularly with regard to the last two, as shown in Figure 7.
The SbMe2 assignments are made in comparison with those for
6 (see below), for which the resonances were assigned on the
(13) (a) Chen, G.; Deng, M.; Lee, C. K.; Leong, W. K. Organometallics
2002, 21, 1227. (b) Chen, G.; Deng, M.; Lee, C. K. D.; Leong, W. K.;
Tan, J.; Tay, C. T. J. Organomet. Chem., in press.
(14) Deeming, A. J.; Underhill, M. J. Chem. Soc., Dalton Trans. 1973,
2727.
Figure 5. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 5.
(15) Skinner, H. A. AdV. Organomet. Chem. 1964, 2, 49.
(16) (a) Bowry, V.; Lusztyk, J.; Ingold, K. U. Pure Appl. Chem. 1990,
62, 213. (b) Skene, W. G.; Connolly, T. J.; Scaiano, J. C. Int. J. Chem.
Kinet. 2000, 32, 238. (c) Bevington, J. C.; Warburton, J.; Hunt, B. J. J.
Macromol. Sci. 2002, A39, 1295.
ligands. Indeed, the thermolysis of 2b in a Carius tube afforded
5 as the major product (29% isolated yield); refluxing decane
was employed, as the reaction at refluxing octane temperature

Thermolysis of Osmium-Antimony Clusters
Organometallics, Vol. 25, No. 1, 2006 253
Scheme 3
Table 1. Kinetic Parameters for the Aromatic Proton
basis of a NOESY spectrum; the methyl group on the same
side of the Os3Sb plane as the phenylene resonated at higher
field. The trend in the hydride resonances is clear and supports
the structure assigned to 2c′. We have also noticed that the
resonances were temperature as well as solvent dependent. At
room temperature, 2c and 2c′ did not interconvert, even after
standing in solution for several days. However, on mild heating,
a solution of 2c isomerized to afford a mixture containing equal
amounts of 2c and 2c′. Such an isomerization was quite
unexpected, and at this point we are unable to say what the
mechanism may be.
The clusters 2c/2c′ also allowed us to make a direct
comparison of the ease of elimination of a methyl vs a phenyl
group from the antimony. Since the 2c-2c′ isomerization was
also fast at elevated temperature, it was expected that 4 would
be formed selectively over 4a. An ESI-MS analysis of the
reaction mixture when 2c/2c′ was thermolyzed with 1g indeed
showed that 4 was formed selectively over the phenyl analogue
4a. This is thus consistent with the Sb-Ph bond being cleaved
much more readily than the Sb-Me bond and also corroborated
the pathway represented in Scheme 3.
Similarly, for the formation of 5, we worked on the hypothesis
that the same intermediate as that in Scheme 3 was involved.
In this case, the nucleophilic attack was presumably on a second
molecule of 2b; we have already established such reactivity,
for instance, in 2a, which can undergo nucleophilic addition
via a µ3 f µ2 rearrangement of the phenylene followed by
decarbonylation.8b Indeed, the position of substitution in 2a, an
axial site on the osmium across from the antimony, corresponds
to where the intermediate “Os3(µ3-SbMe)(µ3,η2-C6H4)(CO)9”
will need to attack in order to form 5. In this event, we observed
Exchanges in 2a-c
∆Hq
∆Sq
∆Gq
300 K
compd
(kcal mol-1
)
(cal mol-1 K-1
)
(kcal mol-1
)
2a
2b
2c
14.0(9)
15.5(4)
15.0(7)
-3(3)
2.1(12)
0(3)
14.8(19)
14.9(7)
15.1(14)
that thermolysis of 2b alone did afford 5. The mode of transfer
of the methyl group onto the second molecule of 2b is uncertain.
As in the phosphorus and arsenic analogues, 2b is fluxional.
An EXSY spectrum taken at 300 K showed no exchange
between the aromatic protons and the metal hydride. Simulation
1
of the VT H NMR behavior of the aromatic protons of the
phenylene ring yielded a ∆Gq(300 K) value of 14.9(7) kcal
mol-1, which is similar to that reported for the phosphorus
analogue.6a It may thus be concluded that the fluxionality
exhibited by 2b is similar to that for the phosphorus analogue:
viz., a combination of ring rotation and hydride migration.
Similar analyses have also been carried out for 2a and 2c, and
the kinetic parameters are summarized in Table 1. As an
independent check, the ∆Gq values at the coalescence temper-
atures (∼300-310 K) were also estimated, and the values for
all three compounds were ∼14.1 kcal mol-1
.
Thermolysis of both the o- and p-tolyl analogues of 1b, viz.,
Os3(CO)11{SbMe2(o-tol)} (1d) and Os3(CO)11{SbMe2(p-tol)}
(1e), afforded the 4-substituted phenylene cluster Os3(µ-SbMe2)-
(µ-H)(µ3,η2-C6H3-4-Me)(CO)9 (2d) (Scheme 4). This behavior
has also been observed with the arsenic analogue.6d Compound
2d has been characterized completely, including by a single-
1
crystal X-ray crystallographic study (Figure 8). The H NMR
1
Figure 7. Tentative H NMR (d8-toluene, 238 K) assignments for 2a, 2b, 2c, and 2c′.
Scheme 4

254 Organometallics, Vol. 25, No. 1, 2006
Chan et al.
Figure 10. Thermolysis of 1f to form 6 and the 1H NMR
assignments for 6.
Scheme 5
Figure 8. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 2d.
Cluster 6 represents the first example in which a phenyl ring
adopts what appears to be an η4 bonding mode to the triosmium
cluster. A very similar structure has been proposed to be
involved in the aromatic proton-hydride exchange in Os3(CO)9-
(µ-H)2(µ3,η2-C6H4),17 although a variable-temperature 1H NMR
study (d8-toluene) on 6 showed it to be nonfluxional. Cluster 6
may also represent an intermediate in the formation of clusters
2; ortho metalation at position 2 would yield the hypothetical
o-phenylene cluster directly, whereas ortho metalation at position
6 would yield an isomer that can subsequently isomerize to the
o-phenylene cluster itself (Scheme 5); the isomerization step
has been demonstrated to occur in a PPh3 derivative.13a The
isolation of 6 may thus be a consequence of its inability to ortho
metalate due to the sterically bulky Me groups at the two meta
positions.
Structural Discussion. A common atomic numbering scheme
and selected bond parameters for 1b and both crystallographi-
cally independent molecules of 1d are given in Table 2. It is
interesting to note that the difference in bond parameters
between the two crystallographically independent molecules of
1d can be quite large; for example, the Os(2)-Os(3) and Os(1)-
Sb(4) bond lengths differ by about 8σ and 7σ, respectively. This
indicates that the effects of crystal-packing forces on structural
parameters can be quite significant. As has been observed in
numerous compounds of this type, the Os-Os bond cis to the
stibine is lengthened relative to the others.18 Likewise, the sums
of Os-C and C-O bond distances19 are noticeably longer for
the axial than for the equatorial carbonyls (generally >3.06 Å
for the former and <3.05 Å for the latter).
Figure 9. ORTEP diagram (organic hydrogens omitted, 50%
probability thermal ellipsoids) for 6.
spectrum corresponds to the presence of two rapidly exchanging
isomers. This has been tentatively ascribed to two diastereomeric
forms which differ in the relative orientations of the hydride
and the methyl group on the ring. At 215 K, the ratio of the
diastereomers is 2.3:1.0, which is comparable to that reported
for the arsenic analogue (2.1:1.0).6d That thermolyses of both
1d and 1e afforded 2d meant that isomerization of the phenyl
ring must have occurred at some point. However, as stated
above, an EXSY experiment on 2b showed that there was no
aromatic proton-hydride exchange; a similar EXSY experiment
on 2a (90 °C, d8-toluene) also gave similar results. These thus
suggested that isomerization of the phenyl ring after formation
of the o-phenylene cluster was not likely.
In an attempt to force the formation of a 3-substituted
phenylene cluster, and to gain an understanding of the isomer-
ization process, the xylyl analogue Os3(CO)11{SbMe2(3,5-
Me2C6H3)} (1f) was thermolyzed. The result was an unexpected
xylyl-capping cluster, Os3(µ-SbMe2)(µ3,η1:η2:η2-C6H3Me2)-
(CO)9 (6). The molecular structure of 6 has been confirmed by
a single-crystal X-ray crystallographic study (Figure 9), and the
1H resonances have been assigned with the aid of NOE
correlations (Figure 10); NOESY cross-peaks were observed
between the aromatic resonance at δH 6.46 and the methyl
resonances at δH 2.89 and 2.11 ppm, between the aromatic
resonance at δH 6.24 and the methyl resonances at δH 2.89 and
1.43 ppm, and between the resonances at δH 4.30 and 2.11 ppm.
(17) Kneuper, H.-J.; Shapley, J. R. Organometallics 1987, 6, 2455.
(18) For example: (a) Biradha, K.; Hansen, V. M.; Leong, W. K.;
Pomeroy, R. K.; Zaworotko, M. J. J. Cluster Sci. 2000, 11, 285. (b) Leong,
W. K.; Liu, Y. J. Organomet. Chem. 1999, 584, 5267. (c) Bruce, M. I.;
Liddell, M. J.; Hughes, C. A.; Skelton, B. W.; White, A. H. J. Organomet.
Chem. 1988, 347, 157. (d) Bruce, M. I.; Liddell, M. J.; Hughes, C. A.;
Patrick, J. M.; Skelton, B. W.; White, A. H. J. Organomet. Chem. 1988,
347, 181. (e) Hansen, V. M.; Ma, A. K.; Biradha, K.; Pomeroy, R. K.;
Zaworotko, M. J. Organometallics 1998, 17, 5267.
(19) Leong, W. K.; Einstein, F. W. B.; Pomeroy, R. K. J. Cluster Sci.
1996, 7, 121.

Thermolysis of Osmium-Antimony Clusters
Organometallics, Vol. 25, No. 1, 2006 255
Table 2. Common Atomic Numbering Scheme and Selected
Bond Parameters for 1b and 1d
Table 3. Common Atomic Numbering Scheme and Selected
Bond Parameters for 2a-d
1b
1d (molecule A) 1d (molecule B)
Bond Lengths (Å)
Os(1)-Os(2)
Os(2)-Os(3)
Os(1)-Os(3)
Os(1)-Sb(4)
2.8463(3)
2.8494(8)
2.8902(8)
2.9062(7)
2.5897(11)
2.8500(8)
2.8835(8)
2.9065(8)
2.5980(12)
2.8898(3)
2.8981(3)
2.5925(4)
2a
2b
2c
2d
Bond Lengths (Å)
Bond Angle (deg)
Os(1)-Os(2)
Os(2)-Os(3)
Os(1)-Sb(4)
Os(3)-Sb(4)
Os(1)-C(1)
Os(3)-C(2)
Os(2)-C(1)
Os(2)-C(2)
C(1)-C(2)
2.9493(8) 2.9631(9) 2.9563(4) 2.9597(7)
2.8517(8) 2.8449(9) 2.8356(4) 2.8547(7)
2.6571(10) 2.6324(13) 2.6437(5) 2.6341(9)
2.6403(10) 2.6272(13) 2.6365(5) 2.6257(9)
2.172(15) 2.174(16) 2.177(7)
2.217(13) 2.140(16) 2.167(7)
2.325(15) 2.309(16) 2.352(6)
2.388(14) 2.357(16) 2.369(6)
Os(2)-Os(1)-Os(3) 60.399(7)
Os(1)-Os(2)-Os(3) 60.687(7)
Os(2)-Os(3)-Os(1) 59.914(7)
Sb(4)-Os(1)-Os(3) 98.196(12)
60.274(19)
60.837(19)
58.889(18)
98.64(3)
60.11(2)
60.91(2)
58.973(19)
100.40(3)
2.175(12)
2.150(13)
2.321(13)
2.379(11)
1.489(18)
A common atomic numbering scheme and selected bond
parameters for 2a-d are given in Table 3. The structure of 2a
has been reported previously;7 that in this study is the di-
chloromethane solvate. As is to be expected, the longer Os-
Os bond is that bridged by an hydride. This apparently also
has further effects on the structures, such as the asymmetry in
the Os-Sb and Os(2)-phenylene bond lengths. The Sb atom
is tilted away from the Os3 plane; the dihedral angle between
the Os(1)Os(2)Os(3) and Os(1)Sb(4)Os(3) planes varies from
about 22 to 28°. The phenylene ring is almost perpendicular to
the Os3 plane, the dihedral angle between the Os(1)Os(2)Os(3)
plane and the phenylene ring being 70-75° but leaning slightly
toward Os(2).
The crystal of 3 contained two crystallographically indepen-
dent molecules; a common atomic numbering scheme and
selected bond parameters are given in Table 4. Cluster 3 is of
interest, as it may be regarded as the parent carbonyl (or more
precisely, the dimethylantimony analogue of it) of the many
derivatives which we have reported,8b,13b all of which contained
an unusually long Os-Os bond. The derivatives which have
been structurally characterized corresponded to substitution at
CO(13) (PPh3 and AsPh3), CO(21) (PPh3, P(p-tolyl)3, and
tBuNC), and CO(32) (tBuNC). As in those derivatives, the Os-
Os bond bridged by the µ,η2-phenylene is very long, although
within this class of compounds that for 3 (3.1614(7) and
3.1611(7) Å, respectively, for the two molecules) is the shortest.
This thus corroborated our contention that the elongation of this
Os-Os bond is partly due to steric repulsion of any ligand
substituted at Os(1) or Os(2) with carbonyls on the adjacent
osmium. Similarly, the Os(2)-Os(3) bond length in 3 is also
shorter than any among the substituted derivatives, in agreement
with ligand effects in the latter. Finally, it is noted that, as for
1d, chemically equivalent bond parameters can differ very
significantly. For instance, even the Os(2)-Os(3) bond lengths
differ by 14σ between the two molecules.
1.431(19) 1.45(2)
1.444(9)
Bond Angles (deg)
Os(2)-Os(1)-Sb(4) 83.69(3)
Os(1)-Os(2)-Os(3) 87.79(2)
Os(2)-Os(3)-Sb(4) 85.94(3)
Os(1)-Sb(4)-Os(3) 98.82(3)
82.78(3)
87.57(2)
85.22(3)
99.68(4)
82.895(13) 81.76(2)
87.853(10) 87.721(19)
85.410(13) 83.95(2)
99.128(16) 100.01(3)
Dihedral Angles (deg)
21.5
Os(1)Os(2)Os(3)/
Os(1)Sb(4)Os(3)
Os(1)Os(2)Os(3)/
phenylene
24.3
24.2
74.8
28.7
70.7
72.3
70.2
these tend to undergo C-H activation.20 A common atomic
numbering scheme and selected bond parameters for all four
of these clusters are given in Table 4.
Structurally, the clusters 4 and 5 can be regarded as being
derived from an Os3(µ3,η2-C6H4)(CO)9(u-SbR2) fragment; in the
case of 4 and 4a, one R group on the antimony has been replaced
by an Os3(CO)11 unit, while in 5 and 5a, both R groups have
been replaced by another Os3(µ3,η2-C6H4)(CO)9(u-SbR2) frag-
ment, together with further substitution of a carbonyl by a methyl
or phenyl. Clusters 4 and 4a are structurally very similar. It
appears that the bond lengths involving the heavy atoms (but
not the light atoms, such as in the phenylene ring) are all shorter
in 4 than in 4a, although there is no obvious reason why this
should be so. In all four clusters, within the phenylene-bridged
Os3 units, the Os(1)-Os(2) bond is the longest. In the case of
5 and 5a, this bond length is quite similar but the other Os-Os
bonds in that fragment are longer in 5a, presumably due to the
greater σ donation from the phenyl on Os(3).
A schematic diagram of 6, together with selected bond
parameters, is given in Figure 11. The interpretation of the
bonding between the xylyl ring and the cluster is interesting.
The Os(1)‚‚‚Os(2) distance of ∼4.044 Å is clearly nonbonding,
and both the structural refinement and 1H NMR spectrum
indicate the absence of any metal hydride. The valence electron
count therefore requires the xylyl to act as a five-electron donor.
We have therefore chosen this to result from a C(2)-Os(2) σ
bond and donation of the C(2)-C(3) and C(1)-C(6) π electrons.
As mentioned above, the diphenylantimony analogue of 4,
viz., Os6(µ3-SbPh)(µ3,η2-C6H4)(CO)20 (4a), has been reported
by us earlier,7 while 5 is closely related to the previously
reported cluster Os6(µ4-Sb)(µ-SbPh2)(µ-H)(µ3,η2-C6H4)2(Ph)-
(CO)16 (5a). Cluster 5 differs from 5a in that a dimethylantimony
moiety in 5 replaces the diphenylantimony moiety in 5a, and
also there is a methyl group on Os(4) in the former instead of
a phenyl group on Os(3) in the latter. Cluster 5 is a very rare
example of an osmium cluster containing an alkyl ligand, as
(20) (a) Cree-Uchiyama, M.; Shapley, J. R.; St. George, G. M. J. Am.
Chem. Soc. 1986, 108, 1316. (b) Boyar, E.; Deeming, A. J.; Arce, A. J.;
De Sanctis, Y. J. Organomet. Chem. 1984, 276, C45. (c) Keister, J. B.;
Shapley, J. R. J. Am. Chem. Soc. 1976, 98, 1056.

256 Organometallics, Vol. 25, No. 1, 2006
Chan et al.
Table 4. Common Atomic Numbering Scheme and Selected Bond Parameters for 4 and 5 and Their Analogues
4
4aa
5
5a
R
Me
Ph
Me on Os(4)
Ph on Os(3)
Bond Lengths (Å)
Os(1)-Os(2)
Os(2)-Os(3)
Os(3)-Os(1)
Os(4)-Os(5)
Os(5)-Os(6)
Os(6)-Os(4)
Os(1)-Sb(7)
Os(2)-Sb(7)
Os(4)-Sb(7)
Os(5)-Sb(7)
Os(4)-Sb(8)
Os(6)-Sb(8)
Sb(7)-R
2.9019(5)
2.7831(5)
2.7524(6)
2.9056(5)
2.8746(5)
2.8470(5)
2.7171(7)
2.6902(7)
2.6032(7)
2.9095(11), 2.9101(11)
2.7939(12), 2.8024(10)
2.7739(10), 2.7659(12)
2.9133(12), 2.9174(10)
2.8910(12), 2.8824(11)
2.8704(11), 2.8794(10)
2.7240(16), 2.7263(13)
2.7170(14), 2.7201(15)
2.6327(15), 2.6344(14)
2.9057(6)
2.7653(5)
2.7696(5)
2.9333(5)
2.9481(5)
2.8943(9)
2.8188(9)
2.8466(9)
2.9422(8)
2.9113(9)
2.6966(7)
2.7066(7)
2.6589(7)
2.5515(7)
2.6524(7)
2.6553(8)
2.6306(11)
2.6474(12)
2.7664(11)
2.5376(11)
2.6621(12)
2.6310(11)
2.156(9)
2.180(18)
Os-R
2.205(9)
2.133(17)
Bond Angles (deg)
58.83(3), 59.11(3)
58.16(3), 57.88(3)
63.01(3), 63.01(3)
59.98(3), 59.63(3)
59.28(3), 59.53(2)
60.75(3), 60.84(3)
98.56(4), 97.64(4)
Os(2)-Os(1)-Os(3)
Os(1)-Os(2)-Os(3)
Os(1)-Os(3)-Os(2)
Os(5)-Os(4)-Os(6)
Os(4)-Os(5)-Os(6)
Os(4)-Os(6)-Os(5)
Os(5)-Os(4)-Sb(7)
58.900(13)
57.870(13)
63.231(13)
59.950(13)
59.014(12)
61.036(13)
96.363(17)
58.260(13)
58.406(13)
63.333(14)
-
58.81(2)
59.75(2)
61.44(2)
-
87.408(14)
88.52(2)
54.022(16)
52.68(2)
a Contains two crystallographically independent molecules.
alternation to the C-C bond lengths around the xylyl ring which
suggests that the double bonds are fairly localized. The
interpretation that there is an interaction of the C(1)-C(6) π
bond with Os(1) is a bit more problematic. As the bond
parameters indicate, this last interaction is at best a very weak
one (the next closest Os‚‚‚C(xylyl) distance is that of Os(1)‚‚
‚C(2) at 2.960(5) Å). This is also apparent from the 1H
resonances; δH for H(1) is 6.24 ppm, compared to 4.30 ppm
for H(3); this is clearly indicative of the aromatic nature of the
former and the more alkenic nature of the latter. It is therefore
possible that 6 may have some degree of electron deficiency.
This is consistent with the red color of 6, as electron-precise
triosmium clusters are usually yellow; if they are electron
deficient, their colors shift to a darker shade.
Figure 11. Schematic diagram of 6, with selected bond lengths
(Å) and angles (deg).
Concluding Remarks
The Os(2)-C(2) bond length of 2.111(5) Å is comparable to,
for instance, the Os-phenyl bond length in 5a, and so to regard
the xylyl ring as being σ bonded to Os(2) appears reasonable.
This σ donation from the aromatic ring is also manifested in its
effect on CO(22); the sum of the Os(2)-C(22) and C(22)-
O(22) bond lengths is 3.07 Å,19 significantly longer than for
the other carbonyls, which are all below about 3.06 Å. The
Os(3)-C lengths are fairly close to the ∼2.30 Å observed for
the π-bond donation to Os observed in, for example, all four of
the clusters 4 and 5 above, although there is asymmetry here.
This indicates that there is an equivalent donation of the π
electrons along the C(2)-C(3) bond to Os(3). This is also
corroborated by the C(2)-C(3) bond length, which is longer
than that in a normal aromatic ring. Interestingly, there is an
In this report, we have examined the thermolyses of a number
of clusters of the formulas Os3(CO)11(SbMe2Ar), to control the
product distribution. The fluxional behaviors of the clusters
containing the o-phenylene moiety have been examined and their
kinetic parameters determined. Higher nuclearity clusters have
also been obtained, including one which contains a rare Os-
alkyl bond. Studies directed at understanding the mechanistic
pathways of these thermolytic reactions have also been carried
out, and we have determined that the Sb-C bond cleavage
probably does not proceed via a radical mechanism. An unusual
cluster containing an η4-bonded aromatic ring has also been
obtained, which we believe is an intermediate that has been
arrested in the process of ortho metalation.

Thermolysis of Osmium-Antimony Clusters
Organometallics, Vol. 25, No. 1, 2006 257
under constant nitrogen bubbling until the color of the solution
turned from yellow to orange (∼1.5 h). The solvent was then
removed in vacuo, and the residue was redissolved in the minimal
volume of dichloromethane for separation by TLC on silica
(eluant: hexane).
Experimental Section
General Procedures. All reactions and manipulations were
carried out under nitrogen using standard Schlenk techniques.21
Solvents were purified, dried, distilled, and stored under nitrogen
prior to use. The products were generally separated by column
chromatography on silica gel 60 (230-430 mesh ASTM) or by
thin-layer chromatography (TLC), using plates coated with silica
gel 60 F254 of 0.25 mm or 0.5 mm thickness and extracted with
hexane or dichloromethane. Solution IR spectra were recorded in
hexane unless otherwise indicated. Gaseous state IR spectra were
acquired in a gas cell (15 cm path length) with KBr windows.
Routine NMR spectra were recorded as CDCl3 solutions, unless
otherwise stated, on a Bruker ACF-300 FT NMR spectrometer. 1H
chemical shifts reported are referenced against the residual proton
signals of the solvents. Selective decoupling experiments, spin
saturation transfer, variable-temperature, and 2D spectra (EXSY,
NOESY) were acquired on a Bruker Avance DRX500 or Bruker
AMX500 machine. NMR simulations were carried out with the
gNMR program.22 Mass spectra were obtained on a Finnigan
MAT95XL-T spectrometer in an m-nitrobenzyl alcohol matrix. GC
analyses were carried out on a Hewlett-Packard 5890 Series 2 PLUS
gas chromatograph with a flame ionization detector and a Chrompack
CP-Silica fused silica column (30 m length × 0.32 mm i.d.).
Microanalyses were carried out by the microanalytical laboratory
at the National University of Singapore. The cluster Os3(CO)11-
(NCCH3) was prepared according to the literature method.12 All
other reagents were from commercial sources and used as supplied.
Preparation of SbMePh2. To a suspension of CH3I (0.51 mL,
8.24 mmol) and Mg turnings (0.241 g, 9.897 mmol) in tetra-
hydrofuran (30 mL) was added dropwise a solution of SbPh2Cl
(2.276 g, 7.311 mmol) in tetrahydrofuran (20 mL). The reaction
mixture was stirred at room temperature for 12 h, whereupon a
dirty green suspension was obtained. The solvent was removed in
vacuo, and the liquid residue was extracted with 80 mL of hexane.
The colorless supernatant was transferred via cannula into a Schlenk
flask and stored under nitrogen.
Band 1 (Rf ) 0.67): Os3(CO)12 (7). Yield: 12.0 mg, 8.2%.
Identified by IR spectroscopy.
Band 2 (Rf ) 0.45): yellow crystals of Os3(CO)9(µ-H)(µ3,η2-
C6H4)(µ-SbMe2) (2b). Yield: 80.1 mg, 47.2%. IR: ν(CO) 2089
(w), 2068 (vs), 2040 (vs), 2009 (vs), 2003 (m, sh), 1988 (w), 1976
1
3
(w), 1968 (w) cm-1. H NMR (233 K): δ 9.15 (d, 1H, JHH
)
8.25 Hz, aromatic), 6.76 (dd, 1H, aromatic), 7.00 (dd, 1H, aromatic),
8.67 (d, 1H, aromatic), 1.57 (s, 3H, Me), 1.11 (s, 3H, Me), -17.71
(s, 1H, OsHOs) ppm. FAB-MS: m/z 1051.8 (calcd for [M+]
1051.6). Anal. Calcd for C17H11O9Os3Sb‚1/4C6H14: C, 20.70; H,
1
1.36. Found: C, 20.52; H, 1.17. H NMR confirms the presence
of hexane in the sample.
Band 3 (Rf ) 0.43): yellow. Unidentified mixture. Yield: <2
mg.
Band 4 (Rf ) 0.42): yellow crystalline solids of Os3(µ-SbMe2)-
(µ-H)(µ,η2-C6H4)(CO)10 (3). Yield: <2 mg. IR: ν(CO) 2105 (w),
2083 (vs), 2064 (s), 2034 (m), 2024 (vs), 2015 (m), 2011 (s), 2002
1
2
(w), 1991 (w), 1989 (sh) cm-1. H NMR: δ 7.46 (d, 1H, JHH
)
7.4 Hz, aromatic), 7.27 (d, 1H, 2JHH ) 7.4 Hz, aromatic), 6.93 (dd,
1H, aromatic), 6.77 (dd, 1H, aromatic), 1.54 (s, 3H, Me), 0.61 (s,
3H, Me), -20.43 (s, 1H, OsHOs). Anal. Calcd for C18H11O10-
Os3Sb: C, 20.02; H, 1.03. Found: C, 20.25; H, 0.96.
Band 5 (Rf ) 0.40): unreacted 1b. Yield: 59.1 mg. Identified
by IR spectroscopy.
Band 6 (Rf ) 0.20): red crystals of Os6(CO)20(µ3,η2-C6H4)(µ3-
SbMe) (4). Yield: 5.4 mg, 3.5%. IR: ν(CO) 2094 (m), 2065 (m,
sh), 2058 (s), 2022 (vs) cm-1. 1H NMR (CD2Cl2) δ: 7.88 (dd, 3JHH
) 6.60 Hz, 4JHH ) 3.30 Hz, 1H, aromatic), 7.78 (dd, 1H, aromatic),
6.78 (dd, 3JHH ) 6.60 Hz, 4JHH ) 3.30 Hz, 2H, aromatic), 2.25 (s,
3H, Me) ppm. FAB-MS: m/z 1914.4 (calcd for [M+] 1914.3). Anal.
Calcd for C27H7O20Os6Sb‚1/4C6H14: C, 17.68; H, 0.55. Found: C,
1
17.52; H, 0.58. H NMR confirms the presence of hexane in the
1
Spectroscopic data are as follows. H NMR: δ 7.56-7.47 (m,
sample.
4H), 7.36-7.33 (m, 6H), 0.96 (s, 3H). EIMS: m/z (% relative
intensity, ion) 291 (71, M), 275 (80, M - Me), 211 (14, M - Ph),
198 (62, M - Ph - Me), 136 (5, M - 2Ph). HR-EI MS: m/z
[M]+ calcd for C13H13121Sb, 290.0055; C13H13123Sb, 292.0059; found
290.0057, 292.0060 (respectively).
Band 7 (Rf ) 0.15): dark red crystals of Os6(CO)16(CH3)(µ-H)-
(µ3,η2-C6H4)2(µ4-Sb)(µ-SbMe2) (5). Yield: 6.6 mg, 4.0%. IR
(CH2Cl2): ν(CO) 2099 (m), 2078 (m), 2067 (s), 2054 (vs), 2026
(vs), 2014 (vs), 2011 (vs, sh), 1962 (w) cm-1. 1H NMR (CD2Cl2):
3
3
δ 9.06 (d, 1H, Jab ) 7.44 Hz, Ha), 8.49 (d, 1H, Jcd ) 7.44 Hz,
Preparation of Os3(CO)11(SbMe2Ph) (1b). To Os3(CO)11(CH3-
CN) (190 mg, 0.206 mmol) was added SbMe2Ph (50 mg, 0.220
mmol) in dichloromethane (40 mL). The reaction mixture was
stirred at 60 °C for 6 h. Silica gel was then added and the solvent
was removed in vacuo, followed by column chromatographic
separation, to give Os3(CO)12 (6.4 mg, 3.4%, identified by IR
spectroscopy), yellow crystalline 1b (182.6 mg, 80.1%), and yellow
oily Os3(CO)10(SbMe2Ph)2 (28.0 mg, 10.4%, identified by compar-
ing its IR spectrum with those of analogous bis-substituted
triosmium clusters).
Hd), 7.93 (d, 1H, 3Jef ) 9.06 Hz, He), 7.55 (dd, 1H, 3Jfg ) 9.06 Hz,
Hf), 7.46 (dd, 1H, Jgh ) Jfg ) 9.06 Hz, Hg), 7.33 (d, 1H, Hh),
6.94 (dd, 1H, Jbc ) 7.44 Hz, Hc), 6.82 (dd, 1H, Hb), 1.27 (s, 3H,
3
3
3
Me), 0.91 (s, 3H, Me), 0.81 (s, 3H, Me), -16.30 (s, 1H, OsHOs).
Anal. Calcd for C31H18O16Os6Sb2‚2C6H14: C, 23.44; H, 2.10.
1
Found: C, 23.38; H, 1.89. The H NMR spectrum of the sample
confirms the presence of hexane.
Cluster 3 was obtained in a yield of 19.3 mg (8.0%) from a
reaction in which a larger quantity (300 mg) of 1b was used.
A similar thermolysis of 1b (5.2 mg, 0.0047 mmol) at 100 °C
was monitored by IR at 30 min intervals until no further change in
the reaction spectrum was observed. Workup afforded unreacted
1b (4.8 mg) and 2b (0.3 mg, 6.1%).
Thermolysis of 1b in a Carius Tube. Cluster 1b (6.3 mg, 0.0057
mmol) was placed in a Carius tube with distilled n-heptane (10
mL), degassed by three freeze-pump-thaw cycles, and then heated
at 125 °C for 15 h. The color of the solution turned from yellow to
brown. The solution was then cooled to -78 °C in an acetone-
dry ice bath before the gas above the solution was transferred into
a gas sample IR cell for analysis. Detection of the Q band at 3017
cm-1, a strong absorption peak at 1306 cm-1, and accompanying
rotational fine structures confirmed the presence of CH4. Compari-
son of the GC retention time of 1.665 min with the standard sample
of CH4 (tR ) 1.656 min) corroborated the presence of CH4. Neither
Data for 1b are as follows. IR: ν(CO) 2108 (w), 2056 (s), 2032
(m), 2021 (vs), 2004 (vw), 1991 (w), 1971 (w), 1955 (vw) cm-1
.
1H NMR: δ 7.58-7.47 (m, 5H, Ph), 1.63 (s, 6H, Me) ppm. Anal.
Calcd for C19H11O11Os3Sb: C, 20.60; H, 1.00. Found: C, 20.83;
H, 0.97.
Details of the preparation and spectroscopic characterization of
1c-g and their organostibine precursors are given in the Supporting
Information.
Thermolysis of 1b in Refluxing n-Octane. Cluster 1b (180.0
mg, 0.162 mmol) was thermolyzed in refluxing n-octane (150 mL)
(21) Shriver, D. F.; Drzedzon, M. A. The Manipulation of Air-SensitiVe
Compounds, 2nd ed.; Wiley: New York, 1986.
(22) Budzelaar, P. H. M. gNMR; Adept Scientific plc, Letchworth, U.K.,
1999.

258 Organometallics, Vol. 25, No. 1, 2006
Chan et al.
vibrational absorption bands centered at 1450 and 730 cm-1 nor a
GC retention time of 1.880 min due to ethane were detected. The
gas was also analyzed by EI-MS, which showed methane but no
ethane.
Thermolysis of 1b with TEMPO. Cluster 1b (3.3 mg, 0.0030
mmol) and TEMPO (0.2 mg, 0.0013 mmol) were placed in an NMR
tube fitted with a Teflon valve, with d20-nonane (0.5 mL). The
solution was degassed by three freeze-pump-thaw cycles and then
heated at 125 °C for 10 h, whereupon the color of the solution
turned from yellow to red. No TEMPO-CH3 was detected in the
1H NMR spectrum.
Thermolysis of 1b with 7. Clusters 1b (190 mg, 0.172 mmol)
and 7 (25 mg, 0.0276 mmol) were dissolved in n-octane (100 mL)
and the solution was deoxygenated and then brought to reflux for
2 h, during which time the color of the solution turned from yellow
to light orange. Removal of the solvent in vacuo and separation by
TLC afforded unreacted 7 (28.4 mg), 2b (27.2 mg, 15.1%), and
unreacted 1b (156.4 mg).
Thermolysis of 1c in Refluxing n-Octane. Cluster 1c (193.0
mg, 0.165 mmol) was thermolyzed and worked up in a manner
similar to that above to yield four bands on the TLC. In order of
elution they are 7 (23.5 mg, 15.7%), Os3(CO)9(µ-H)(µ3,η2-C6H4)-
(µ-SbPhMe) (2c/2c′; 77.5 mg, 42.2%), 2a (15.7 mg, 8.1%), and
unreacted 1c (65.5 mg). Clusters 2c/2c′ and 2a were separated by
repeated TLC until the 1H NMR showed that both bands were free
from each other. Cluster 2a probably arose from a small quantity
of 1a present in the sample of 1c; these could not be separated by
TLC. This is supported by the detection of SbPh3 in the EI-MS of
the SbMePh2 solution used to prepare 1c. The structure of the
known cluster 2a was established by a single-crystal X-ray
crystallographic study as a solvate of that reported (Supporting
Information).23
C6H3Me2), 2.89 (3H, s, C6H3Me2), 2.11 (3H, s, C6H3Me2), 1.80
(3H, s, SbMe2), 1.43 (3H, s, SbMe2). FAB-MS: m/z 1078.8 (calcd
for [M+] 1078.1).
Thermolysis of 1g in Refluxing n-Octane. A similar thermolysis
and workup with 1g (200 mg, 0.191 mmol) afforded 7 (2.9 mg,
1.7%), “Os3SbMe3(CO)10” (8; 1.8 mg), unreacted 1g (187.7 mg),
and an unidentified band (0.3 mg; IR ν(CO) 2062 (m), 2055 (m),
2021 (vs), 2003 (s), 1991 (m), 1957 (m) cm-1).
Data for 8 are as follows. IR: ν(CO) 2101 (vw), 2075 (m), 2064
(s), 2059 (vw), 2053 (vw), 2039 (s), 2019 (s), 2006 (w), 1996 (m),
1977 (m) cm-1. ESI-MS (negative mode): 1049 (calcd for [M +
OMe]- 1048.8).
Thermolyses of 2b. A sample of 2b (1.5 mg, 0.0014 mmol)
was placed in an NMR tube with d8-toluene (0.5 mL), degassed
by three freeze-pump-thaw cycles, and then heated at 90 °C for
1
4 h. The color of the solution remained yellow. The H NMR
spectrum in the hydride region revealed a single peak at -17.58
ppm (CDCl3, 300 K). TLC analysis revealed the presence of
unreacted 2b only.
A similar thermolysis of 2b (8.4 mg, 0.0080 mmol) with decane
(10 mL) in a Carius tube at 150 °C was monitored by IR
spectroscopy at 15 min intervals for 1 h. Cluster 5 was detected as
the methoxide adduct in the ESI-MS spectrum. At the end of the
thermolysis, the color of the solution was brown. TLC separation
with hexane/dichloromethane (90:10, v/v) yielded 5 as the only
identifiable product (2.3 mg, 29%).
A similar thermolysis of 2b (10.1 mg, 0.0094 mmol) in refluxing
n-octane (10 mL) under a constant nitrogen purge was monitored
by IR spectroscopy at 30 min intervals for 4 h. At the end of the
thermolysis, the color of the solution remained yellow. The IR
spectrum and TLC analysis revealed 2b as the only product present.
Thermolyses of 3. A sample of 3 (17.8 mg, 0.0165 mmol) in
octane (10 mL) was refluxed in a 25 mL round-bottom flask. The
solution turned from yellow to dark orange in 60 min. After removal
of the octane in vacuo, the dark orange residue was dissolved in a
minimal amount of dichloromethane and subjected to TLC separa-
tion with hexane as eluant. The first band (yield 0.9 mg, 5.2%)
was identified as 2b; four other unidentified bands were also
obtained, together with a great deal of material remaining at the
bottom.
Thermolysis of 1g and 2c/2c′. Clusters 1g (2.1 mg, 0.0020
mmol) and 2c/2c′ (2.0 mg, 0.0018 mmol) were placed in a Carius
tube with n-octane (10 mL), degassed by three freeze-pump-thaw
cycles, and then heated at 140 °C for 30 min, whereupon the color
of the solution turned from yellow to red. The ESI-MS spectrum
showed peaks at m/z 1918.4 and 1945.3 only, corresponding to M+
and [M + OMe]+ for 4.
X-ray Crystallographic Studies. Crystals were grown from
dichloromethane/hexane solutions and mounted on quartz fibers.
X-ray data were collected on a Bruker AXS APEX system, using
Mo KR radiation, at 223 K with the SMART suite of programs.24
Data were processed and corrected for Lorentz and polarization
effects with SAINT25 and for absorption effects with SADABS.26
Structural solution and refinement were carried out with the
SHELXTL suite of programs.27 Crystal and refinement data are
summarized in Tables 5 and 6.
The structures were solved by direct methods to locate the heavy
atoms, followed by difference maps for the light, non-hydrogen
atoms. There were two crystallographically independent molecules
in the asymmetric units of 1d and 3. A dichloromethane solvent
molecule was found in 2a and a hexane molecule (half occupancy)
in 6. The locations of the hydrides were either located in a low
angle difference map (2a, 2c, 2d, and 5) or placed by potential
Data for 2c/2c′ are as follows. IR: ν(CO) 2090 (w), 2070 (vs),
2042 (vs), 2016 (m), 2009 (m), 2003 (m), 1988 (w), 1977 (w),
1
3
1971 (w) cm-1. H NMR for 2c (CD2Cl2): δ 9.17 (d, 1H, JHH
)
8.25 Hz, C6H4), 6.74 (dd, 1H, C6H4), 7.05 (dd, 1H, C6H4), 8.64 (d,
1H, C6H4), 7.22 - 7.19 (m, 3H, Ph), 6.93-6.90 (m, 2H, Ph), 1.83
1
(s, 3H, Me), -17.45 (s, 1H, OsHOs). H NMR for 2c′ (CD2Cl2,
218 K): δ 9.17 (d, 1H, 3JHH ) 8.25 Hz, C6H4), 6.74 (dd, 1H, C6H4),
7.05 (dd, 1H, C6H4), 8.64 (d, 1H, C6H4), 7.35-7.42 (m, 5H, Ph),
1.33 (s, 3H, Me), -17.38 (s, 1H, OsHOs). The ratio of 2c to 2c′
was determined from 1H NMR to be 1:1. Anal. Calcd for C17H11O9-
Os3Sb: C, 23.72; H, 1.18. Found: C, 23.86; H, 1.05.
Thermolysis of 1d in Refluxing n-Octane. A similar thermoly-
sis and workup with 1d (150.0 mg, 0.134 mmol) yielded, after TLC
separation, 7 (8.5 mg, 7.0%), Os3(µ-SbMe2)(µ-H)(µ3,η2-C6H3-4-
Me)(CO)9 (2d; 71.5 mg, 50.1%), and unreacted 1d (42.9 mg).
Data for 2d are as follows. IR: ν(CO) 2089 (w), 2068 (vs), 2040
(vs), 2009 (vs), 2003 (m, sh), 1988 (w), 1976 (w), 1968 (w) cm-1
.
1H NMR: δ 9.03 (d, JHH ) 8.25 Hz, C6H3, major), 8.87 (s, JHH
) 8.25 Hz, C6H3, minor), 8.84 (s, C6H3, major), 8.58 (d, C6H3,
minor), 6.84 (d, C6H3, minor), 6.62 (d, C6H3, major), 2.34 (s,
C6H3Me, minor), 2.28 (s, C6H3Me, major), 1.57 (s, 3H, SbMe2),
1.11 (s, 3H, SbMe2), -17.58 (s, 1H, OsHOs). Anal. Calcd for
C18H13O9Os3Sb‚1/4C6H14: C, 21.54; H: 1.53. Found: C, 21.34; H,
3
3
1
1.52. H NMR confirmed the presence of hexane in the sample.
A similar thermolysis of 1e also afforded 2d, identified by the
1
low-temperature H NMR spectrum. Yield: 81.8 mg, 43.0%.
Thermolysis of 1f in Refluxing n-Octane. A similar thermolysis
and workup with 1f (200 mg, 0.175 mmol) afforded Os3(µ-SbMe2)-
(µ3,η1:η2:η2-C6H3Me2)(CO)9 (6), which eluted together with un-
reacted 1f. Fractional crystallization afforded 6 as red crystals, but
these were invariably contaminated with 1f. Yield: ∼68.0 mg,
38.5%. IR: ν(CO) 2069 (w), 2044 (vs), 2022 (s), 1995 (m), 1985
1
(w), 1976 (w), 1968 (m), 1956 (w) cm-1. H NMR (CD2Cl2): δ
(24) SMART, version 5.628; Bruker AXS Inc., Madison, WI, 2001.
(25) SAINT+, version 6.22a; Bruker AXS Inc., Madison, WI, 2001.
(26) Sheldrick, G. M. SADABS, 1996.
6.46 (1H, s, C6H3Me2), 6.24 (1H, s, C6H3Me2), 4.30 (1H, s,
(23) Leong, W. K.; Chen, G. Organometallics 2001, 20, 2280.
(27) SHELXTL, version 5.1; Bruker AXS Inc., Madison, WI, 1997.

Thermolysis of Osmium-Antimony Clusters
Organometallics, Vol. 25, No. 1, 2006 259
Table 5. Crystal and Refinement Data for 1b,d and 2b-d
1b
1d
2b
2c
2d
empirical formula
formula wt
temp, K
C19H11O11Os3Sb
1107.63
223(2)
C20H13O11Os3Sb
1121.65
223(2)
C17H11O9Os3Sb
1051.61
223(2)
C22H13O9Os3Sb
1113.67
298(2)
C18H13O9Os3Sb
1065.63
223(2)
cryst syst
space group
a, Å
b, Å
c, Å
monoclinic
C2/c
monoclinic
P21/c
monoclinic
P21/c
monoclinic
P21/n
monoclinic
C2/c
21.8809(9)
8.3103(4)
27.3697(12)
90.1010(10)
4976.8(4)
8
2.957
16.396
3936
0.30 × 0.24 × 0.18
8.4121(5)
35.499(2)
17.5220(11)
96.161(3)
5202.2(6)
8
2.864
15.688
4000
16.3554(14)
10.2881(9)
13.2426(11)
96.605(2)
2213.5(3)
4
3.156
18.417
1856
8.4185(2)
16.8901(5)
18.1308(5)
96.8350(10)
2559.68(12)
4
2.890
15.935
1984
9.6232(5)
16.1907(8)
29.7325(15)
96.2140(10)
4605.3(4)
8
3.074
17.706
3776
â, deg
V, Å3
Z
density (calcd), Mg/m3
abs coeff, mm-1
F(000)
cryst size, mm3
0.16 × 0.14 × 0.03
2.08-26.37
39 232
0.16 × 0.14 × 0.08
2.34-24.71
23 020
0.32 × 0.30 × 0.20
2.26-29.52
19 891
0.12 × 0.08 × 0.06
2.47-26.37
30 314
θ range for data collecn, deg 2.38-30.01
no. of rflns collected
no. of indep rflns
38 510
7257 (R(int) ) 0.0497) 10640 (R(int) ) 0.0684) 3771 (R(int) ) 0.0374) 6436 (R(int) ) 0.0374) 4712 (R(int) ) 0.0381)
0.1563, 0.0838
max, min transmissn
0.6504, 0.1880
10 640/0/637
1.137
0.3205, 0.1566
3771/0/183
1.227
0.1428, 0.0801
6436/0/321
1.029
0.4164, 0.2251
4712/0/286
1.358
no. of data/restraints/params 7257/0/309
goodness of fit on F2
1.173
final R indices (I > 2σ(I))
R1 ) 0.0326,
wR2 ) 0.0574
R1 ) 0.0404,
wR2 ) 0.0591
R1 ) 0.0585,
wR2 ) 0.1047
R1 ) 0.0840,
wR2 ) 0.1123
2.171, -1.280
R1 ) 0.0498,
wR2 ) 0.1155
R1 ) 0.0552,
wR2 ) 0.1176
3.703, -1.409
R1 ) 0.0333,
wR2 ) 0.0709
R1 ) 0.0469,
wR2 ) 0.0754
1.639, -1.520
R1 ) 0.0533,
wR2 ) 0.1003
R1 ) 0.0569,
wR2 ) 0.1014
2.257, -1.478
R indices (all data)
largest diff peak, hole, e Å-3 1.253, -1.144
Table 6. Crystal and Refinement Data for 3-6
3
4
5
6
empirical formula
formula wt
C18H11O10Os3Sb
1079.62
C27H7O20Os6Sb
1914.28
C31H18O16Os6Sb2
2031.15
C22H22O9Os3Sb
1122.75
temp, K
223(2)
298(2)
223(2)
223(2)
cryst syst
space group
monoclinic
Pn
monoclinic
P21/n
monoclinic
P21/c
monoclinic
P21/n
a, Å
b, Å
c, Å
9.2722(4)
9.6092(3)
15.7364(10)
13.2779(8)
19.5943(12)
91.037(2)
4093.5(4)
4
3.296
19.908
3560
9.8025(6)
15.0937(7)
17.2511(8)
90.7010(10)
2414.14(19)
4
2.970
16.894
1912
27.4384(8)
14.3351(4)
92.5710(10)
3775.80(19)
4
3.367
20.889
3344
18.0300(10)
15.8402(9)
103.4880(10)
2722.4(3)
â, deg
V, Å3
Z
4
density (calcd), Mg/m3
abs coeff, mm-1
F(000)
2.739
14.984
2020
cryst size, mm3
θ range for data collecn, deg
no. of rflns collected
no. of indep rflns
max, min transmissn
no. of data/restraints/params
goodness of fit on F2
final R indices (I > 2σ(I))
0.40 × 0.34 × 0.10
2.36-29.93
19 935
6429 (R(int) ) 0.0512)
0.2829, 0.0565
6429/2/581
1.032
R1 ) 0.0325,
wR2 ) 0.0708
R1 ) 0.0358,
wR2 ) 0.0721
3.017, -1.757
0.20 × 0.08 × 0.02
2.06-28.28
49 800
9046 (R(int) ) 0.0469)
0.6801, 0.1027
9046/0/488
1.096
R1 ) 0.0444,
wR2 ) 0.0781
R1 ) 0.0608,
wR2 ) 0.0832
1.917, -0.973
0.16 × 0.09 × 0.04
2.01-24.71
40 328
6982 (R(int) ) 0.0304)
0.5031, 0.1429
6982/0/502
1.193
R1 ) 0.0342,
wR2 ) 0.0685
R1 ) 0.0382,
wR2 ) 0.0700
1.490, -1.065
0.18 × 0.12 × 0.10
2.24-29.55
38 009
7137 (R(int) ) 0.0275)
0.3157, 0.1734
7137/66/369
1.059
R1 ) 0.0303,
wR2 ) 0.0644
R1 ) 0.0362,
wR2 ) 0.0667
1.699, -0.924
R indices (all data)
largest diff peak, hole, e Å-3
energy calculations (2b and 3) with the program XHYDEX.28 They
were refined fully (2c), refined on position only (2a, 2d, and 5), or
refined riding on one of the osmium atoms they are attached to
(2b and 3) with fixed isotropic thermal parameters. Organic
hydrogen atoms were placed in calculated positions and refined
with a riding model. All non-hydrogen atoms were generally given
anisotropic displacement parameters in the final model, except for
the carbon atoms in 2a and 2b, where there appeared to be problems
due to extreme absorption differences which could not be adequately
corrected for. The structure of 3 was refined as a racemic twin.
Acknowledgment. This work was supported by the National
University of Singapore (Research Grant No. R143-000-190-
112) and one of us (K.H.C.) thanks A*STAR for a Pre-Graduate
Award.
Supporting Information Available: Crystallographic data in
CIF format and text and figures giving experimental details
on the syntheses and characterization of the stibines and clusters
1. This material is available free of charge via the Internet at
http://pubs.acs.org.
(28) Orpen, A. G. XHYDEX; School of Chemistry, University of Bristol,
Bristol, U.K., 1997.
OM0508132