Structural characterisation of the elusive isomer of Os3(CO)10(PPh3)2

Article abstract of DOI:10.1016/S0022-328X(99)00133-3

The X-ray crystal structure of the elusive isomer of Os₃(CO)₁₀(PPh₃)₂ in which the two phosphine ligands are cis and trans with respect to the phosphine-substituted Os-Os edge has been determined; the first dire

Full text of DOI:10.1016/S0022-328X(99)00133-3

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 584 (1999) 174–178
Structural characterisation of the elusive isomer of
Os3(CO)10(PPh3)2
Weng Kee Leong *, Yao Liu
Department of Chemistry, National Uni6ersity of Singapore, Kent Ridge, Singapore 119260, Singapore
Received 7 January 1999
Abstract
The X-ray crystal structure of the elusive isomer of Os3(CO)10(PPh3)2 in which the two phosphine ligands are cis and trans with
respect to the phosphine-substituted Os–Os edge has been determined; the first direct confirmation for the existence of this isomer
in Os3(CO)10(PR3)2 clusters. In contrast to the trans,trans isomer, the cis,trans isomer shows little twisting of the Os(CO)3(PPh3)
groupings with respect to the Os(CO)4 unit. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Osmium; Phosphine; Clusters; Isomers
gave red crystals (1b) while the latter yielded yellow
crystals (1c) by rapid cooling of a concentrated solution
in a CH2Cl2/hexane mixture.1 Recrystallisation of 1c
invariably gives crystals of 1b, indicating the more
favourable thermodynamics for crystallisation of the
latter.
1. Introduction
Phosphine-substituted triosmium clusters are among
the most studied derivatives of osmium carbonyl clus-
ters. One interesting feature of these clusters is the
almost-exclusive tendency for the phosphine ligands to
occupy an equatorial site with respect to the triosmium
plane [1]. In the particular case of the disubstituted
derivatives, Os3(CO)10(PR3)2, there are in principle four
different isomers possible which are depicted in Fig. 1.
Of these, some examples with structure A have been
isolated, clusters with structure D have only been ob-
served for diphosphine ligands, while clusters with
structural types B and C frequently occur as mixtures in
rapid equilibrium. Although many of the Os3(CO)10-
(PR3)2 clusters are known to contain structural types B
and C in rapid equilibrium, all these clusters appear to
crystallise out in the B form; this appears to be the case
even when it has been demonstrated that the major
isomer in solution is that with the structure C [2,3].
In the course of our investigations into the bromina-
tion reaction of Os3(CO)10(PR3)2 clusters [4], we have
found that the cluster Os3(CO)10(PPh3)2 (1) can be
crystallised out in both the B and C forms; the former
The solid-state IR spectra in the carbonyl region of
these two isomers show subtle differences (Fig. 2), but
just as has been observed for Os3(CO)10(PMe2Ph)2 and
Os3(CO)10[P(OMe)3]2, there is a rapid equilibrium be-
tween these two isomers in solution; at −80°C, the
31P{1H}-NMR spectra of either crystalline form in
CD2Cl2 showed two equally intense signals at −2.50
and −5.64 ppm, attributable to 1c, and a lower inten-
sity signal at −1.91 ppm attributed to 1b. The major
isomer in solution is thus 1c, in an approximately 3.4:1
ratio with respect to 1b, from integration of the 31P-
NMR signals. These signals coalesce completely at
about 30°C; the isomerization mechanism in these sys-
tems has been fairly well studied, and involves restricted
turnstile rotation at one of the phosphine-bearing osmi-
ums [2,3].
1 The crystal structure of 1b has been reported earlier by Bruce et
al. [1b]. We have chosen to make comparisons with our own study;
such a comparison would be less likely to suffer from diffractometer-
dependent systematic errors. As one of the two crystallographically
distinct molecules was found to be disordered (ca. 3% for Os(5) and
Os(6)), comparisons have been made with the ordered molecule only.
* Corresponding author. Fax: +65-7791691.
E-mail address: chmlwk@nus.edu.sg (W. Kee Leong)
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(99)00133-3

W. Kee Leong, Y. Liu / Journal of Organometallic Chemistry 584 (1999) 174–178
175
The molecular structures of both 1b and 1c are
shown in Fig. 3, and selected bond parameters are
collected in Table 1. The X-ray crystal structure of 1c
reported here represents the first direct confirmation for
structural type C. In 1c, the Os(1)–Os(2) and Os(2)–
Os(3) bond lengths are much longer (0.0194 and 0.0170
This observation may be rationalised if we consider the
steric interactions involving the PPh3 ligands (Fig. 4).
In 1b, the steric effects of both PPh3 ligands impinge on
Os(3), while in 1c this is not the case. It may thus be
expected that 1b will have a greater tendency to distort
towards the S conformation in order to relieve the
strain, while the steric effects are ‘relayed’ round the
equatorial plane in 1c; the latter is evident in the bond
angles about the equatorial plane for Os(3), which show
that the equatorial carbonyls (CO(32) and CO(34)) are
bent away from Os(2) and towards Os(1).
If our discussions above were correct, then it would
appear that the relative orientation of ligands on clus-
ters can have very significant effects on both the gross
ligand sphere structure of the cluster as well as its
chemistry. For instance, we may expect that an
Os3(CO)10(PR3)2 cluster which exists in solution solely
as the B form will undergo halogenation less fa-
vourably; this is currently under investigation.
,
A, respectively) than the Os(1)–Os(3) bond, consistent
with the observation that a PR3 ligand tends to
lengthen the cis Os–Os bond [5,6]; the longer Os(1)–
Os(2) bond may be the consequence of a combination
of steric repulsions between CO(22) with P(1) and P(2),
which is also reflected in the Os(1)–P(1) bond being
longer than the Os(2)–P(2) bond by a 10| difference
(|=e.s.d.), as compared with a 5| difference for the
two Os–P bond lengths in 1b. It is also noteworthy that
the Os(1)–Os(2) bond length in 1c is the longest
,
at 2.9204(3) A; this is consistent with our earlier
suggestion that in the bromination reaction of
Os3(CO)10(PPh3)2, the isomer 1c is the one that under-
goes bromination and it involves cleavage of this bond
[4]. The most significant structural difference between
1b and 1c, however, is the greater degree of twisting of
the Os(CO)3(PPh3) units with respect to the Os(CO)4
unit in the former (Table 1).
2. Experimental
All reactions and manipulations were carried out
under nitrogen by using standard Schlenk techniques.
Infrared spectra were recorded as KBr disks. 31P-NMR
spectra were recorded on a Bruker ACF-300 FT-NMR
spectrometer and referenced to 85% aqueous H3PO4.
Microanalyses were carried out by the microanalytical
laboratory at the National University of Singapore.
In a recent exhaustive study on Os3(CO)11(PR3) clus-
ters, Pomeroy et al. have found that the nature of the
PR3 ligand can have profound structural implications
[6]. Essentially, their argument was that a better donor
PR3 ligand increases electron density at the Os atom of
the Os(CO)3(PR3) unit, leading to expansion of its filled
5d orbitals. This in turn causes an increase in repulsive
interaction with filled 5d orbitals on the other Os atoms
and hence a reduction in the bonding contribution of
the edge-bridging MOs. Steric effects from the PR3
ligand will then lead to lengthening of the cis Os–Os
bond as well as a propensity towards greater twisting of
the Os(CO)3(PR3) group, i.e. the adoption of what
Pomeroy et al. termed the S conformation, in order to
alleviate the 5d orbitals repulsive interaction.
Cluster
1 was prepared from the reaction of
Os3(CO)10(CH3CN)2 with two equivalents of PPh3.
Crystallisation from hexane/dichloromethane solution
gave a mixture of red crystals (1b) (Found: C, 39.86; H,
2.12. Calc. for C46H30O10Os3P2: C, 40.17; H, 2.18) and
yellow crystals (1c) (Found: C, 39.99; H, 2.14).
2.1. Crystal data for 1b
C46H30O10Os3P2, M=1375.24, monoclinic, space
From the electronic viewpoint, 1b and 1c should have
the same degree of expansion of the 5d orbitals at both
the Os(CO)3(PPh3) units. As measured by the Cax–Os–
Os–Cax torsion angles (Table 1), there is much less
twisting or distortion towards the S conformation in 1c.
group P21/n, a=14.8216(2), b=34.4785(2), c=
3
,
,
17.1059(3) A, i=92.373(1)°; U=8734.1(2) A ; Z=8;
zc=2.092 Mg m−3; v(Mo–Ka)=8.838 mm−1, T=
295 K, 56 987 reflections collected, 21 737 unique reflec-
tions, final R=11.13%, wR=8.44% for all data, 1106
parameters and three restraints.
2.2. Crystal data for 1c
C46H30O10Os3P2, M=1375.24, triclinic, space group
(
,
P1, a=12.5013(5), b=13.0248(5), c=14.2326(5) A,
h=71.784(1), i=79.384(1), k=89.220(1)°; U=
2161.23(14) A ; Z=2; zc=2.113 Mg m−3; v(Mo–
3
,
Ka)=8.929 mm−1
, T=295 K, 18 290 reflections
collected, 10 275 unique reflections, final R=4.11%,
wR=7.58% for all data and 550 parameters.
Fig. 1. Possible isomers for eq,eq-Os3(CO)10(PR3)2.

176
W. Kee Leong, Y. Liu / Journal of Organometallic Chemistry 584 (1999) 174–178
Fig. 2. Solid-state IR spectra (KBr disks, carbonyl region) for 1b and 1c.
2.3. Summary of structural determination
absorption effects with SADABS [8]. The final unit cell
parameters were obtained by least squares on 8192 (1b)
or 6272 (1c) strong reflections. Structural solution and
refinement were carried out with the SHELXTL suite of
programs [9].
The structure was solved by direct methods to locate
the heavy atoms, followed by difference maps for the
light, non-hydrogen atoms. The terminal metal hydride
was located by a low angle difference map and refined;
The crystals were sealed in glass capillaries. X-ray
data were collected on a Siemens SMART diffractome-
ter, equipped with a CCD detector, using Mo–Ka
radiation at ambient temperature (298 K). Ten second
exposures were used.
Data was corrected for Lorentz and polarisation
effects with the SMART suite of programs [7], and for

W. Kee Leong, Y. Liu / Journal of Organometallic Chemistry 584 (1999) 174–178
177
Fig. 4. Schematic view down the Os3 planes showing selected bond
angles (°) and steric interactions of the PPh3 ligands, in 1b and 1c.
given the same anisotropic thermal parameters as those
of the corresponding major components; the corre-
sponding Os–Os bond distances were restrained to be
the same, and the occupancies of major and minor
components summed to unity.
3. Supplementary material
Crystallographic data (excluding structure factors)
for the structures in this paper have been deposited
with the Cambridge Crystallographic Data Centre as
supplementary publication numbers CCDC 114318 and
114319. Copies of the data can be obtained, free of
charge, on application to CCDC, 12 Union Road,
Cambridge CB2 1EZ, UK, (fax: +44-1223-336033 or
e-mail: deposit@ccdc.cam.ac.uk).
Fig. 3. ORTEP diagrams of 1b and 1c (50% probability thermal
ellipsoids).
the phenyl hydrogens were placed in calculated posi-
tions. The final refinement was with an all non hydro-
gen atoms anisotropic model.
There were two independent molecules in the asym-
metric unit for 1b, one of which showed fairly large
residuals near two Os atoms (Os(5) and Os(6)) which
were modelled as disorder of the Os3 triangle. The
minor component of the disordered Os atoms were
Acknowledgements
This work was supported by the National University
of Singapore (Research Grant No. RP 960677) and one
of us (Y.L.) thanks the University for a Research
Scholarship.
Table 1
,
Selected bond lengths (A) and torsion angles (°) for 1b and 1c
Bond parameter
1b
2.9065(4)
1c
2.9204(3)
References
Os(1)–Os(2)
Os(1)–Os(3)
Os(2)–Os(3)
Os(1)–P(1)
Os(2)–P(2)
C(11)–Os(1)–Os(2)–C(21)
C(13)–Os(1)–Os(2)–C(23)
C(11)–Os(1)–Os(3)–C(31)
C(13)–Os(1)–Os(3)–C(33)
C(21)–Os(2)–Os(3)–C(31)
C(23)–Os(2)–Os(3)–C(33)
2.9095(4)
2.8988(5)
2.351(2)
2.361(2)
24.8(4)
19.9(4)
21.5(4)
19.3(4)
18.7(4)
2.9010(3)
2.9180(3)
2.3607(12)
2.3478(13)
8.1 (3)
10.5(2)
4.7(3)
8.7(3)
3.1(3)
[1] See for example, (a) A.J. Deeming, in: E.W. Abel, F.G.A.
Stone, G. Wilkinson, G. (Eds.) Comprehensive Organometallic
Chemistry II, vol. 7, Chap. 12, Elsevier, Oxford, 1995. (b) M.I.
Bruce, M.J. Liddell, C.A. Hughes, J.M. Patrick, B.W. Skelton,
A.H. White, J. Organomet. Chem. 347 (1988) 181.
[2] (a) A.J. Deeming, S Donovan-Mtunzi, S.E. Kabir, P.J. Man-
ning, J. Chem. Soc. Dalton Trans. (1996) 1037. (b) A.J. Deem-
ing, S. Donovan-Mtunzi, S.E. Kabir, J. Organometal. Chem.
281 (1985) C43.
22.1(4)
10.0(3)
[3] R.F. Alex, R.K. Pomeroy, Organometallics 6 (1987) 2437.

178
W. Kee Leong, Y. Liu / Journal of Organometallic Chemistry 584 (1999) 174–178
[4] W.K. Leong, Y. Liu, Organometallics 18 (1999) 800.
[5] M.I. Bruce, M.J. Liddell, C.A. Hughes, B.W. Skelton, A.H.
White, J. Organometal. Chem. 347 (1988) 157.
[6] V.M. Hansen, A.K. Ma, K. Biradha, R.K. Pomeroy, M.J. Za-
worotko, Organometallics 17 (1998) 5267.
[7] SMART version 4.05, Siemens Energy
Madison, WI, USA.
[8] SADABS, G.M. Sheldrick, 1996.
[9] SHELXTL version 5.03, Siemens Energy
Madison, WI, USA.
&
Automation Inc.,
& Automation Inc.,
.