Synthesis, characterization, and dynamic behavior of Os2Pt(CO)8(PPh3)2: A trinuclear osmium-platinum cluster with flexible metal framework

Article abstract of DOI:10.1021/om950938e

Reaction of Os₂(CO)₈(μ-η¹,η¹-C ₂H₄) with (η²-C₂H₄)Pt(PPh₃)₂ gives predominantly the heterotrinuclear cluster Os₂Pt(CO)₈(PPh₃)₂, 1. The reaction is accompanied by redistribution of the PPh₃ ligand, and as a result, compound 1 exists in solution as three isomers (1a-c) due to the presence and disposition of a phosphine ligand on both Os and Pt. The nature of the isomers was deduced from a combination of ¹³C, ³¹P, and ¹⁹⁵Pt NMR spectroscopies, and the solid-state structure of the major isomer was corroborated by a single-crystal X-ray analysis. The isomers interconvert on the NMR time scale, and the mechanisms and energetics of the isomerization processes were determined by ³¹P NMR selective inversion magnetization transfer experiments. Two pathways for isomer exchange were identified. Isomers 1a and 1b exchange via an olefin type rotation of the diosmium fragment about the platinum center, whereas interconversion between 1a and 1c is accomplished by a restricted trigonal twist at the phosphine-substituted osmium center; the energetics for the processes are Δ^H? = 10.1(2) kcal·mol^-1, ΔS^? = -13.1(11) eu and ΔH^? = 10.7(3) kcal·mol^-1, ΔS^? = -5.2(12) eu, respectively. The mechanism of the 1a ? 1c exchange was confirmed by ¹³C NMR spin saturation transfer experiments.

Full text of DOI:10.1021/om950938e

Organometallics 1996, 15, 4459-4468
4459
Syn th esis, Ch a r a cter iza tion , a n d Dyn a m ic Beh a vior of
Os2P t(CO)8(P P h 3)2: A Tr in u clea r Osm iu m -P la tin u m
Clu ster w ith F lexible Meta l F r a m ew or k
J ason Cooke,† R. E. D. McClung, and J osef Takats*
Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
Robin D. Rogers
Department of Chemistry, Northern Illinois University, DeKalb, Illinois 60115
Received December 5, 1995X
Reaction of Os2(CO)8(µ-η1,η1-C2H4) with (η2-C2H4)Pt(PPh3)2 gives predominantly the
heterotrinuclear cluster Os2Pt(CO)8(PPh3)2, 1. The reaction is accompanied by redistribution
of the PPh3 ligand, and as a result, compound 1 exists in solution as three isomers (1a -c)
due to the presence and disposition of a phosphine ligand on both Os and Pt. The nature of
the isomers was deduced from a combination of 13C, 31P, and 195Pt NMR spectroscopies, and
the solid-state structure of the major isomer was corroborated by a single-crystal X-ray
analysis. The isomers interconvert on the NMR time scale, and the mechanisms and
energetics of the isomerization processes were determined by 31P NMR selective inversion
magnetization transfer experiments. Two pathways for isomer exchange were identified.
Isomers 1a and 1b exchange via an “olefin type” rotation of the diosmium fragment about
the platinum center, whereas interconversion between 1a and 1c is accomplished by a
restricted trigonal twist at the phosphine-substituted osmium center; the energetics for the
processes are ∆Hq ) 10.1(2) kcal‚mol-1, ∆Sq ) -13.1(11) eu and ∆Hq ) 10.7(3) kcal‚mol-1
,
∆Sq ) -5.2(12) eu, respectively. The mechanism of the 1a a 1c exchange was confirmed
by 13C NMR spin saturation transfer experiments.
Ch a r t 1
In tr od u ction
Recent interest in heteronuclear cluster compounds
of platinum,1 in particular those in combination with
group eight metals (Fe, Ru, Os),2 is the direct result of
the recognized connection between such clusters and
bimetallic alloy catalysts.3 Despite the wealth of on-
going research, the simple trinuclear carbonyl clusters
of iron or ruthenium and platinum are as yet unknown.
In the case of iron, the related phosphine-substituted
trinuclear compounds Fe2Pt(CO)9L and Fe2Pt(CO)8L2 (L
) PR3, L2 ) Ph2PCH2CH2PPh2 (dppe)),4 and the cyclo-
octa-1,5-diene adduct Fe2Pt(CO)8(COD),5 have been
reported (Chart 1). For ruthenium, clusters of the type
Ru2Pt(CO)7(PR3)3 and RuPt2(CO)5(PR3)3 are known,6 but
to date, the only species bearing structural similarity
to the iron-platinum clusters is Ru2Pt(CO)8(dppe),
obtained by phosphine-induced cleavage of the hexa-
nuclear Pt-Ru cluster (eq 1)2b or as one of two products
† NSERC Undergraduate Research Awardee.
X Abstract published in Advance ACS Abstracts, August 15, 1996.
(1) Farrugia, L. J . Adv. Organomet. Chem. 1990, 31, 301.
(2) For example, see: (a) Adams, R. D.; Arafa, I.; Chen, G.; Lii, J .;
Wang, J . Organometallics 1990, 9, 2350 (M ) Fe). (b) Adams, R. D.;
Chen, G.; Wang, J .; Wu, W. Organometallics 1990, 9, 1339 (M ) Ru).
(c) Adams, R. D.; Chen, G.; Lii, J .; Wu, W. Inorg. Chem. 1991, 30, 1007
(M ) Os).
(3) The Chemistry of Heteronuclear Clusters and Multimetallic
Catalysts; Proceedings of the Seminar Held at Albertus Magnus
College, Ko¨nigstein, FRG, Sept 7-11, 1987; Adams, R. D., Herrmann,
W. A., Eds. Polyhedron 1988, 7, 2251-2462.
from the reaction of Ru3(CO)12 with Pt(dppe)2.6 In the
case of osmium, the parent decacarbonyl Os2Pt(CO)10
has been characterized,7 but owing to its thermal
instability and ready dimerization to Os4Pt2(CO)18 (eq
(4) (a) Bruce, M. I.; Shaw, G.; Stone, F. G. A. J . Chem. Soc., Dalton
Trans. 1972, 1082. (b) Mason, R.; Zubieta, J . J . Chem. Soc., Chem.
Commun. 1972, 200.
(5) Farrugia, L. J .; Howard, J . A. K.; Mitrprachachon, P.; Stone, F.
G. A.; Woodward, P. J . Chem. Soc., Dalton Trans. 1981, 1134.
(6) Bruce, M. I.; Shaw, G.; Stone, F. G. A. J . Chem. Soc., Dalton
Trans. 1972, 1781.
(7) Sundberg, P. J . Chem. Soc., Chem. Commun. 1987, 1307.
S0276-7333(95)00938-1 CCC: $12.00 © 1996 American Chemical Society

4460 Organometallics, Vol. 15, No. 21, 1996
Cooke et al.
2), only recently has a high-quality X-ray crystal struc-
ture determination on the complex been obtained.2c
metallic complexes Os2Rh(CO)9(η5-C5R5) (R ) H, Me).18
A recent extension of this methodology to the reaction
of Os2(CO)8(µ-η1,η1-C2H4) with (η5-C5H5)Rh(CO)PR′3 (R′
) Me, Ph) introduced complications arising from redis-
tribution of the phosphine ligand from rhodium to an
osmium center.19
With these established precedents, it was of interest
to investigate the reaction of octacarbonyldiosmacyclo-
butane with an unsaturated platinum fragment. The
greater stability of phosphine-substituted platinum
complexes7 and the convenient synthesis of (η2-C2H4)-
A prevalent difficulty in targeted cluster synthesis is
the occasional lack of simple building blocks from which
to generate more complex structures in a rational
manner. Notable exceptions are the metal-carbyne
synthon (LnMtCR), pioneered by Stone and his co-
workers,8 and the bridging ligand promoted agglomer-
ization of metal-containing fragments.9 In the present
context, Stone and co-workers had some success in the
preparation of Fe2Pt(CO)8L2 complexes by reacting
Fe2(CO)9 with 1 equiv of PtL4 but were hampered by
low yields and formation of mononuclear Fe(CO)4L and
Fe(CO)3L2 byproducts.4a The only convenient syntheses
of Os2Pt(CO)10 involve cleavage of either the hexa-
nuclear cluster Os4Pt2(CO)18 under 50 atm of CO
pressure7 or the pentanuclear cluster Os3Pt2(CO)10-
(COD)2 by CO purge at 25 °C.2c
20
Pt(PPh3)2 made this the reagent of choice for such a
study. Previously, Beringhelli and co-workers success-
fully reacted (η2-C2H4)Pt(PPh3)2 with the formally un-
saturated bimetallic complex Re2(µ-H)2(CO)8 and ob-
tained Re2Pt(µ-H)2(CO)8(PPh3)2.21
Resu lts a n d Discu ssion
Syn th esis a n d Ch a r a cter iza tion . As indicated in
the introduction, octacarbonyldiosmacyclobutane is a
useful synthetic source of the reactive Os2(CO)8 frag-
ment. While previous cases10b,12,18,19 required either
mild heating or photolysis to initiate reaction, (η2-C2H4)-
Pt(PPh3)2 reacts readily at ambient temperature to form
predominantly an orange complex in good yield (eq 3).
Bimetallic osmium carbonyl complexes, which could
act as suitable precursors for the generation of higher
nuclearity clusters, became available with the facile
photochemical synthesis of diosmacyclobutane, Os2-
(CO)8(µ-η1,η1-C2H4),10 and the related, but more labile,
propylene analogue.11 Although the ready loss of eth-
ylene from Os2(CO)8(µ-η1,η1-C2H4) was initially believed
to proceed via a concerted cycloreversion,12a recent
synthetic13 and kinetic14 studies by Norton have dem-
onstrated that the fragmentation of diosmacyclobutane
does not involve direct generation of (OC)4Os)Os(CO)4
but proceeds rather by an intermediate in which the
bridging olefin has slipped to η2 coordination at one
osmium center. This latter point of view is in accord
with matrix isolation11,15 and solution flash photolysis15
studies on these diosmacycles. Whatever the mecha-
nistic intricacies, the feature of synthetic import is that
these diosmacyclobutanes generally act as sources of the
unsaturated Os2(CO)8 fragment. This has been dem-
onstrated in olefin12 and alkyne10b exchange reactions
and, following Stone’s insightful application of the
isolobal analogy16 to the rational synthesis of trimetalla-
cyclopropanes,17 in the preparation of the heterotri-
The mass spectrum shows the molecular ion, followed
by the loss of eight carbonyl groups. The product is
isolated by chromatography and is thermally stable, in
contrast to the parent Os2Pt(CO)10 complex which, in
the absence of a CO atmosphere, readily dimerizes to
form Os4Pt2(CO)18 (eq 1).7 Clearly, the presence of the
two phosphine ligands in 1 stabilizes the trinuclear
cluster and disfavors ligand loss which would lead to
the formation of clusters of higher nuclearity. A similar
example of this type of stabilization by a phosphine
ligand is the dppe-induced cleavage of Ru4Pt2(CO)18 (eq
1).2b
The infrared spectrum of the product in pentane
shows 10 terminal carbonyl stretching bands; their
number and complexity indicate the presence of isomers
and hints at an asymmetric cluster. For instance,
symmetric complexes of the type Fe2Pt(CO)8L2 (L ) PR3)
show only five or six terminal carbonyl stretching bands
in hydrocarbon solvents.4a The presence of isomers was
also established in Re2Pt(µ-H)2(CO)8(PPh3)2.21
The NMR spectra are temperature dependent. The
ambient-temperature 31P NMR spectrum, Figure 1,
shows two well-separated resonances; the downfield
signal is quite broad. Cooling the sample to -70 °C
reveals a mixture of three distinct isomers in a 4.7:3.4:
1.0 (1a :1b:1c) ratio. Each isomer has two phosphine
ligands with clearly dissimilar chemical environments.
The downfield signal in each case (Pa, Pc, Pe) has
platinum satellites separated by ca. 3000 Hz and is thus
(8) Stone, F. G. A. Pure Appl. Chem. 1986, 58, 529.
(9) (a) Adams, R. D. Polyhedron 1985, 83, 203. (b) Adams, R. D.
Coord. Chem. Rev. 1985, 65, 219. (c) Shaw, B. L.; Smith, M. J .; Stretton,
G. N.; Thornton-Pett, M. J . Chem. Soc., Dalton Trans. 1988, 2099 and
references therein. (d) Braunstein, P.; Knorr, M.; Sta¨hrfeldt, T. J .
Chem. Soc., Chem. Commun. 1994, 1913 and references therein.
(10) (a) Burke, M. R.; Takats, J .; Grevels, F. W.; Reuvers, J . G. A.
J . Am. Chem. Soc. 1983, 105, 4092. (b) Burke, M. R.; Seils, F.; Takats,
J . Organometallics 1994, 13, 1445.
(11) Haynes, A.; Poliakoff, M.; Turner, J . T.; Bender, B. R.; Norton,
J . R. J . Organomet. Chem. 1990, 383, 497.
(12) (a) Hembre, R. T.; Scott, C. P.; Norton, J . R. J . Am. Chem. Soc.
1987, 109, 3468. (b) Hembre, R. T.; Ramage, D. L.; Scott, C. P.; Norton,
J . R. Organometallics 1994, 13, 2995.
(13) Spetseris, N.; Norton, J . R.; Rithner, C. D. Organometallics
1995, 14, 603.
(14) Norton, J . R. Personal communication.
(15) Grevels, F. W.; Klotzbu¨cher, W. E.; Seils, F.; Schaffner, K.;
Takats, J . J . Am. Chem. Soc. 1990, 112, 1995.
(16) Hoffman, R. Angew. Chem., Int. Ed. Engl. 1982, 21, 711.
(17) Stone, F. G. A. Angew. Chem., Int. Ed. Engl. 1984, 23, 89.
(18) Washington, J .; Takats, J . Organometallics 1990, 9, 925.
(19) Cooke, J .; Takats, J . Organometallics 1995, 14, 698.
(20) Cook, C. D.; J auhal, G. S. J . Am. Chem. Soc. 1968, 90, 1464.
(21) Beringhelli, T.; Ceriotti, A.; D’Alfonso, G.; Della Pergola, R.;
Ciani, G.; Moret, M.; Sironi, A. Organometallics 1990, 9, 1053.

Os2Pt(CO)8(PPh3)2
Organometallics, Vol. 15, No. 21, 1996 4461
between Pb and Pt. Isomer 1b displays a very small
phosphorus-phosphorus coupling of 2 Hz but a large
2J Pt-P coupling of 85 Hz for Pd, indicating that the
osmium-bonded phosphine is on the extension of the
Pt-Os bond, trans to platinum but not trans to the other
phosphine (Pc). In the structurally related minor isomer
of Re2Pt(µ-H)2(CO)8(PPh3)2, the corresponding value of
2J Pt-P is 130 Hz and there is no observable phosphorus-
phosphorus coupling.21 The upfield signal of isomer 1c
shows a much smaller two-bond platinum-phosphorus
coupling of 25 Hz. The decrease in the magnitude of
2
the J Pt-P coupling constant indicates a departure from
the roughly linear disposition of the phosphorus, os-
mium, and platinum centers found in the other two
isomers;22 Pf is thus assigned a position on the Os-Os
bond extension. The observed 5 Hz phosphorus-
phosphorus coupling negates the possibility of Pf being
syn to Pe and suggests the geometry indicated in Figure
1. Steric congestion between the bulky PPh3 ligands
apparently disfavors generation of the syn-isomer.24
Low-temperature 195Pt NMR spectroscopy (Figure 2)
provides further structural evidence. Figure 2a shows
the spectrum obtained with a sample containing natural
abundance 13CO ligands. The observation of three
doublets is in accord with the presence of one platinum-
1
F igu r e 1. Variable-temperature 31P NMR spectra of 1 at
162 MHz in CD2Cl2. Coupling data at -70 °C are as
bonded phosphine ligand in each isomer; the J Pt-P
coupling constants match those observed in the -70 °C
2
3
follows: 1a (1J Pt-P ) 2913 Hz; J Pt-P ) 60 Hz; J P
)
a-Pb
a
b
31P NMR spectrum. The 1a and 1b resonances are
49 Hz); 1b (1J Pt-P ) 2990 Hz; J Pt-P ) 85 Hz; J P -P ) 2
2
3
2
c
d
c
Hz); 1c (1J
) 2756 Hz; 2J
) 25 Hz; 3J
) 5dHz).
further split by the expected J Pt-P values, 60 Hz for
Pt-Pe
Pt-Pf
Pe-Pf
2
1a and 75 Hz for 1b. In 1c, the J Pt-P coupling of 25
Hz is less than the natural 195Pt line width (ca. 50 Hz)
and the anticipated splitting is not resolved. More
interesting, though, is the -70 °C 195Pt NMR spectrum
of a sample enriched greater than 50% in 13CO. Each
195Pt resonance is now flanked by 13C satellites with
1J Pt-C of 1530, 1490, and 1520 Hz, respectively, for 1a -
c. The coupling constants are of the correct magntitude
for carbonyl ligands which are terminally bonded to
platinum,25 therefore also confirming the presence of a
carbonyl ligand at the platinum center in each isomer.
Thus, the low-temperature 31P and 195Pt NMR spectra
clearly point toward a common metal framework from
which the geometrical isomers 1a -c are derived. The
platinum center is square planar4b and 16 electron,
while each osmium center adopts the expected 18
electron pseudo-octahedral geometry. The equatorial
placement of the phosphine ligands about the trinuclear
metal framework is in accord with the normal trend
observed for trinuclear clusters of osmium.26 The
migration of a triphenylphosphine group to osmium in
exchange for a carbonyl during the formation of the
complex is perhaps not unexpected; Stone has reported
that this is a common reactivity pattern of (η2-C2H4)-
Pt(PPh3)2 with carbonyl-bearing species.27 Further-
more, a similar result was noted in the reaction of (η2-
C2H4)Pt(PPh3)2 with Re2(µ-H)2(CO)8. Interestingly in
this case, the intermediate complex, with both phos-
due to a platinum-bonded phosphorus.22,23 The overlap-
ping upfield signals (Pb, Pd, Pf) are flanked by platinum
satellites with much smaller 195Pt-31P coupling con-
stants and are accordingly assigned to osmium-bonded
PPh3 ligands. At +95 °C, the spectrum shows only two
averaged signals; the observed platinum-phosphorus
and phosphorus-phosphorus couplings are in reason-
able agreement with the calculated average values. The
maintenance of these couplings over the full range of
temperatures and the reversible nature of the line shape
changes indicate that the exchange processes averaging
the Pa, Pc, and Pe signals, and the Pb, Pd, and Pf
resonances, are intramolecular. Full disclosure of the
rearrangement processes responsible for this is deferred
for later discussion.
With the presence of a PPh3 ligand on both Pt and
Os centers having been established, and with the
previous structural precedents shown in Chart 1, the
presence of isomers must be due to the different disposi-
tions of the phosphine ligand on the osmium atoms. The
relative orientation of the phosphine ligands in each
isomer is determined from the magnitude of the phos-
phorus-phosphorus and two-bond platinum-phospho-
rus coupling constants (refer to Figure 1). The major
3
isomer, 1a , exhibits a large J P-P coupling of 49 Hz and
is thus the isomer in which the phosphines are along
the extension of the Os-Pt bond, trans to one another.
A two-bond phosphorus-platinum coupling of 60 Hz is
also observed for the osmium-bonded PPh3, Pb, which
is consistent with the proposed trans relationship
(24) For example, significant steric repulsion between syn-phos-
phites is reported in Os3(CO)6[P(OMe)3]6 (Alex, R. F.; Einstein, F. W.
B.; J ones, R. H.; Pomeroy, R. K. Inorg. Chem. 1987, 26, 3175).
(25) Mann, B. E.; Taylor, B. F. 13C NMR Data for Organometallic
Compounds; Academic: New York, 1981; p 181.
(26) (a) Alex, R. F.; Pomeroy, R. K. Organometallics 1987, 6, 2437.
(b) Deeming, A. J . Adv. Organomet. Chem. 1986, 26, 1 and references
therein.
(22) Verkade, J . G.; Quin, L. D. 31P NMR in Stereochemical Analysis;
VCH: Deerfield Beach, FL, 1987; Chapters 6, 13, and 14.
(23) Pregosin, P. S.; Kunz, R. W. 31P and 13C NMR of Transition
Metal Complexes; NMR Basic Principles and Progress 16;
Diehl, P., Fluck, E., Kosfeld, R., Eds.; Springer-Verlag: New York,
1979.
(27) Farrugia, L. J .; Howard, J . A. K.; Mitrprachachon, P.; Stone,
F. G. A.; Woodward, P. J . Chem. Soc., Dalton Trans. 1981, 1274.

4462 Organometallics, Vol. 15, No. 21, 1996
Cooke et al.
F igu r e 2. 195Pt NMR spectra of 1 at -70 °C in CD2Cl2 with (a) natural abundance 13CO and (b) g50% 13CO enrichment.
phines still on platinum, was observed also.28 The latter
undergoes irreversible CO/PPh3 exchange between Re
and Pt to give a mixture of two isomers of Re2Pt(µ-H)2-
(CO)8(PPh3)2. Contrary to 1a , the major isomer has the
PPh3 ligand on Re perpendicular to the trimetallic
framework.21,28
X-r a y Cr ysta l Str u ctu r e of 1a . To gain further
confirmation of the molecular geometry of compound 1
and to obtain precise metrical parameters, a single-
crystal X-ray analysis was performed. Unfortunately,
the small size and poor quality of the crystals, coupled
with pseudosymmetry disorder, led to a poorly refined
structure (see Experimental Section). Nonetheless, the
basic structural features of the cluster are clear (Figure
3). As deduced from the spectroscopic data, the plati-
num center is square planar and bears one triphenyl-
F igu r e 3. ORTEP drawing of 1a with relevant atom
phosphine and one carbonyl ligand, while the osmium
centers are distorted octahedrons. One osmium sup-
ports four terminal carbonyl ligands; the other is bonded
to one triphenylphosphine and three carbonyl ligands.
In particular, the phosphine ligands are disposed in a
roughly linear arrangement along the extensions of an
osmium-platinum bond, thus confirming the antici-
pated geometry for the major isomer 1a . As noted
previously, the equatorial placement of the PPh3 ligands
is in accord with the normal trend for trinuclear osmium
complexes.26 The axial orientation of the PPh3 ligand
numbering. The phenyl groups on phosphorus have been
omitted for clarity. Selected bond distances (Å): Pt-Os(1)
2.691(3), Pt-Os(2) 2.680(4), Os(1)-Os(2) 2.916(4), Pt-P(1)
2.40(2), Os(1)-P(2) 2.34(2). The P(1)-Pt-Os(1)-P(2) tor-
sion angle is 27.4°.
than that in Os2(CO)8(µ-η1,η1-C2H4) (2.883(1) Å),29
Os3(CO)12 (2.887(3) Å),30 and Os2Pt(CO)10 (2.864(1) Å).2c
Isom er In ter con ver sion . The temperature depend-
ence of the 31P NMR spectra clearly suggests that the
three distinct isomers observed at -70 °C are inter-
converting at +95 °C to give an averaged spectrum.
Since the line shape changes with temperature are
reversible, and the spin-spin couplings are maintained
at all temperatures, the isomer interconversion must
be intramolecular. At -20 °C, the resonances of isomers
1a and 1c are much broader than that of 1b. This
indicates that exchange between isomers 1a and 1c is
occurring at this temperature, as yet without significant
21
in the major isomer of Re2Pt(µ-H)2(CO)8(PPh3)2 is
probably due to the greater steric restraints imposed
by the presence of the two bridging hydride ligands in
the equatorial plane of this complex.
The Pt-Os(1) and Pt-Os(2) bond lengths are es-
sentially equal, 2.691(3) and 2.680(4) Å, respectively and
are similar to the Pt-Os bond lengths in the parent
compound Os2Pt(CO)10 (2.689(1) and 2.678(1) Å).2c The
Os(1)-Os(2) distance of 2.916(4) Å is slightly longer
(29) Motyl, K. M.; Norton, J . R.; Schauer, C. K.; Anderson, O. P. J .
Am. Chem. Soc. 1982, 104, 7325.
(30) Churchill, M. R.; De Boer, B. G. Inorg. Chem. 1977, 16, 878.
(28) Beringhelli, T.; D’Alfonso, G.; Minoja, A. P. Organometallics
1991, 10, 394.

Os2Pt(CO)8(PPh3)2
Organometallics, Vol. 15, No. 21, 1996 4463
mental Section) and give ∆Gq
) 11.8(4) kcal‚mol-1
215
for 1a a 1c and 13.0(4) kcal‚mol-1 for 1a a 1b
isomerizations.
On the basis of the isolobal relation16
(OC)4OsdOs(CO)3PPh3 @; H2CdCHR
the exchange between isomers 1a and 1b can be likened
to the well-known rotation of an olefin at a transition
metal center, i.e. rotation of the diosmium fragment
about an axis extending from the Pt(CO)PPh3 moiety,
Scheme 1a. The thesis that the isomerization occurs
as a result of the flexibility of the trinuclear metal core
has precedence in the literature.32-34 Rotation of the
homobimetallic portion of the trinuclear clusters MRh2(µ-
CO)2(CO)5(η5-C5Me5)2 (M ) Mo, W) was initially pro-
posed as one explanation for their variable-temperature
13C NMR features.33a The synthesis of the unsymmetric
analog MoCoRh(µ-CO)2(CO)5(η5-C5Me5)2 confirmed this
proposal; rotation of the CoRh(µ-CO)2(C5Me5)2 fragment
about the Mo(CO)5 group was directly observed by
variable-temperature 13C NMR spectroscopy.33b A simi-
lar motion has been proposed to account for the fluxional
behavior of Os2Cr(CO)11[P(OMe)3]2.34 In this case, the
Os2(CO)6[P(OMe)3]2 group rotates about the Cr(CO)5
unit. Further, Stone has described in detail the analogy
of the rotation of a bimetallic fragment, within a
triangular metal framework, to that of an olefin rotating
about an MLn fragment.32
F igu r e 4. Plots showing the evolution of magnetization
vs delay time (τ) of the spin inversion magnetization
transfer experiments in the 31P NMR at -50.5(5) °C: (a)
results from selective inversion of the signal belonging to
isomer 1a ; (b) results from selective inversion of the signal
belonging to isomer 1b.
involvement of 1b. However, the 1b resonance broad-
ens at 0 °C and all three signals coalesce near +20 °C;
this, coupled with the narrowing to an average spectrum
at +95 °C, indicates that isomer 1b must also be
exchanging with isomer 1a and/or isomer 1c. Line
shape simulations based on the exchange pathways
The free energy of activation of 13.0(4) kcal‚mol-1 at
215 K for the 1a /1b exchange can be compared with that
of ∆Gq258 ) 12.5(3) kcal‚mol-1 in MoCoRh(µ-CO)2(CO)5(η5-
C5Me5)2.33b It is interesting that, in the isolobal PtRh2(µ-
CO)2(CO)(PPh3)(η5-C5Me5)2, rotation of the dirhodium
fragment is not observed.35 In this case, however, the
interaction of the µ-CO ligands with the square-planar
platinum center acts to anchor the Rh2 portion of the
molecule and inhibits its rotation.32 Since no such
interactions are present in complex 1, the diosmium
fragment is presumably free to rotate about the plati-
num center without much hindrance.
(4 )
gave reasonable agreement between calculated and
observed spectra, but the precision of the rate constants
obtained was low.
In order to delineate the nature of the exchange
pathways more precisely, 31P NMR selective inversion
magnetization transfer experiments31 were undertaken.
The data obtained for the experiments at -50.5(5) °C
are given in Figure 4. Qualitatively, these results
confirm the proposed mechanism of exchange. Inver-
sion of the signal belonging to 1a is followed by initial
rapid decreases in the intensities of the signals for 1b
and 1c due to direct exchange between 1a and these
isomers, while selective inversion of the signal belonging
to 1b results in an initial rapid decrease in the intensity
of the signal for isomer 1a only, followed by a later
decrease in the 1c signal due to the two-step 1b f 1a
f 1c magnetization transfer pathway. The magnitudes
of the initial slopes of the magnetization vs time curves
in Figure 4 indicate the relative magnitudes of the
corresponding exchange rate constants and show clearly
that the rate of the 1a f 1c transfers is about 15 times
that for the 1a f 1b transfers at -50.5(5) °C. The
selective inversion magnetization transfer results do not
exclude the possibility of direct 1b f 1c exchange, only
that the rate of direct transfer is far slower than the
rate of the indirect 1b f 1a f 1c process over the
temperature range studied. The magnetization transfer
results afford accurate determination of the activation
parameters for the two exchange processes (see Experi-
Interconversion of 1a and 1c can be envisaged to
proceed via two possible mechanisms. As the 1a /1b
isomerization, exchange of 1a and 1c could involve a
rotation of the (OC)4Os)Pt(CO)PPh3 fragment about the
Os(CO)3PPh3 moiety, Scheme 1c. On the other hand,
the 1a /1c interconversion could also occur via a re-
stricted trigonal twist26a at the phosphine-substituted
osmium center, Scheme 1b. As detailed in the scheme,
if the latter mechanism were operative, the restricted
trigonal twist at the phosphine-substituted osmium
center would result in exchange of an axial carbonyl of
1a with an equatorial carbonyl of 1c (1 a 18), and
exchange of an equatorial carbonyl in 1a with the axial
carbonyls in 1c (6 a 13). Clearly, in order to probe this
possibility, complete assignment and a clear under-
standing of the variable-temperature features of the 13
NMR spectra were required.
C
(32) Barr, R. D.; Green, M.; Howard, J . A. K.; Marder, T. B.; Orpen,
A. G.; Stone, F. G. A. J . Chem. Soc., Dalton Trans. 1984, 2757.
(33) (a) Barr, R. D.; Green, M.; Marsden, K.; Stone, F. G. A.;
Woodward, P. J . Chem. Soc., Dalton Trans. 1983, 507. (b) Barr, R. D.;
Green, M.; Howard, J . A. K.; Marder, T. B.; Stone, F. G. A. J . Chem.
Soc., Chem. Commun. 1983, 759.
(34) Davis, H. B.; Einstein, F. W. B.; J ohnston, V. J .; Pomeroy, R.
K. J . Am. Chem. Soc. 1988, 110, 4451.
(31) Muhandiram, D. R.; McClung, R. E. D. J . Magn. Reson. 1987,
71, 187.
(35) Green, M.; Mills, R. M.; Pain, G. N.; Stone, F. G. A.; Woodward,
P. J . Chem. Soc., Dalton Trans. 1982, 1309.

4464 Organometallics, Vol. 15, No. 21, 1996
Sch em e 1. P ossible Mech a n ism s for Isom er Rea r r a gem en t in Com p ou n d 1
Cooke et al.
Va r ia ble-Tem p er a tu r e 13C NMR Sp ectr oscop ic
Stu d ies. The -70 °C spectrum (Figure 5) is under-
standably complex, but reliable integration of the
resonances allowed the identification of the signals
belonging to each of the three isomers. Within each set,
six resonances are observed in a 2:2:1:1:0.66:1 ratio for
the two pairs of axial carbonyls and the four inequiva-
lent equatorial carbonyls; the axial carbonyls at each
osmium center are rendered equivalent by the mirror
plane through the three metals, but the two pairs are
themselves inequivalent. The three signals of relative
integration 0.66 (3, 1a ; 9, 1b; 15, 1c) are due to the
platinum-bonded carbonyl. The relative integration of
0.66 for the main peaks results from the 33.8% natural
abundance of 195Pt which causes platinum satellites,
each with relative integration 0.17, to flank the main
bonyl signals are found downfield of the equatorial
carbonyl signals in each isomer is also consistent with
the normal trend in trinuclear clusters of osmium.26
The assignment of the remaining three osmium-
bonded equatorial carbonyl ligands in each isomer was
3
aided by the observation of J P-C coupling for the
carbonyls trans to a triphenylphosphine ligand across
an Os-Pt bond. For instance, in isomer 1b, the
resonance for carbonyl 9, which is unequivocally as-
signed to the carbonyl ligand attached to platinum, is
split into a doublet (3J P-C ) 20 Hz). It is noteworthy
that no such coupling was observed through homo-
nuclear Os-Os bonds in phosphine-/phosphite-substi-
tuted triosmium carbonyl clusters.26
Following this line of reasoning, the signal due to
carbonyl 11, the one trans to Pc, is also expected to be a
doublet. Its doublet appearance (3J P-C ) 14 Hz) is
evident at -50 °C but is obscured by overlap with
carbonyl 5 at -70 °C. That the doublets are due to
three-bond phosphorus-carbon coupling is further cor-
roborated by the fact that, at -70 °C, the 31P NMR
spectrum of the 13CO-enriched sample shows 13C satel-
lites for both signals of isomer 1b (3J P-C ) 22 Hz for Pd
1
signal. The J Pt-C coupling constants match those
observed in the 195Pt NMR spectrum of Figure 2 (1J Pt-C
) 1540 (1a ), 1490 (1b), 1540 Hz (1c)). The overlap of
the platinum satellites of 1a and 1c at δ 185.3 and those
of 1a -c at δ 170.0 is due to the accidental superposition
of signals 3 and 15, the similar chemical shift of
1
resonance 9, and the similar magnitude of the J Pt-C
3
coupling constants.
and J P-C ) 14 Hz for Pc). The remaining carbonyls of
Within each isomer, the lowest field carbonyl signal
of intensity two (1, 1a ; 7, 1b; 13, 1c) is assigned to the
axial carbonyl ligands at the phosphine-substituted
osmium center. This is in accord with well-established
trends that substitution by a σ-donor ligand causes
downfield shift of the 13CO resonance.26 The upfield
carbonyl singlets of intensity two (2, 1a ; 8, 1b; 14, 1c)
are thus assigned to the axial carbonyls at the unsub-
stituted osmium centers. The fact that the axial car-
isomer 1b, 10 and 12, are widely separated and are
confidently assigned on the basis of the trend of down-
field shift of the carbonyl resonance with phosphine
substitution.26 Accordingly, carbonyl 10 is assigned to
the upfield signal at δ 172.6 and carbonyl 12 to the
downfield signal at δ 181.1.
In isomer 1c, the signal at δ 181.7 is a doublet (3J P-C
) 15 Hz) and consequently is attributed to carbonyl 18.
It should be noted also that this is the farthest downfield

Os2Pt(CO)8(PPh3)2
Organometallics, Vol. 15, No. 21, 1996 4465
F igu r e 5. Variable-temperature 13C NMR spectra of 1 at 100.6 MHz in CD2Cl2.
equatorial carbonyl signal in isomer 1c, which is in
accord with its attachment to the triphenylphosphine-
substituted osmium center, and thus further corrobo-
rates the structural assignment of the isomer. The
signals due to carbonyls 16 and 17 were identified by
analogy to those assigned more rigorously in 1b; that
is, the higher field signal of the two is assigned to
carbonyl 16, the one along the extension of the homo-
nuclear Os-Os bond.
For the signals of isomer 1a , there are no distinct
NMR features, such as doublets due to phosphorus
splitting, to aid in the structural assignment of the
equatorial carbonyls. Carbonyl 4, on the extension of
the Os-Os bond, can be confidently assigned to the
furthest upfield signal at δ 175.4 on the basis of the
assignments established for isomer 1b. However, un-
equivocal assignment of the signals at δ 180.2 and 179.2
to equatorial carbonyls 5 and 6 is not possible on the
basis of chemical shift arguments alone.
ments36 were carried out at -60 °C. Irradiation of the
axial carbonyl resonance 1 of 1a , at δ 194.2, was
accompanied by a decrease in intensity of the equatorial
carbonyl signal 18 of 1c; corresponding intensity de-
creases were also observed in the axial carbonyls 13 of
1c and the equatorial carbonyl of isomer 1a at δ 179.2.
A small decrease in intensity at axial site 7 of isomer
1b was also present and is consistent with the known
slow exchange between isomers 1a and 1b. Irradiation
of the equatorial signal at δ 179.2 (1a ) resulted in a loss
of intensity in the axial carbonyl site 13 of isomer 1c
with concomitant intensity decreases at sites 1 (1a ) and
18 (1c). Clearly, these results implicate a restricted
trigonal twist rotation at the phosphine-substituted
osmium center as being responsible for the 1a a 1c
exchange process. Finally, we are also in a position to
complete the assignment of the carbonyl signals in
isomer 1a ; the resonance at δ 179.2 is due to the
exchanging carbonyl 6, and the sole remaining carbonyl
5 is assigned to the signal at δ 180.2.
Delineation of the 1a /1c exchange process presents
an opportunity for both determining the assignment of
the 5/ 6 carbonyls of isomer 1a and for establishing the
mechanism of the 1a /1c isomerization. As mentioned
before, a restricted trigonal twist at the phosphine-
substituted osmium center would result in exchange of
axial carbonyls 1 in 1a with equatorial carbonyl 18 of
1c and exchange of the equatorial carbonyl 6 in 1a with
axial carbonyls 13 in 1c, Scheme 1b. In effect, the
process will lead to exchange between all carbonyls
bonded to the phosphine-substituted osmium centers of
isomers 1a ,c (1, 6; 13, 18). On the other hand, if the
isomerization was to occur by a metal-olefin rotation as
in Scheme 1c, there would be no exchange of the axial
and equatorial sites; instead, equatorial carbonyls 6 and
18, and axial carbonyls 1 and 13, would be exchanging.
In either case, equatorial carbonyl 6 of isomer 1a could
be positively assigned either to the signal at δ 180.2 or
179.2 on the basis of the observed exchange behavior.
To unambiguously identify the exchange related to
carbonyl sites, 13C NMR spin saturation transfer experi-
The action of a restricted trigonal twist mechanism
in 1 at such low temperature was unexpected, since
similar processes in phosphine-/phosphite-substituted
trinuclear osmium clusters typically occur at higher
temperature.26 For instance, the rates in Os3(CO)10-
[P(OMe)3]2 and Os3(CO)9[P(OMe)3]3 are 31(1) s-1 at +17
°C and 40(2) s-1 at -2 °C, respectively.26a The corre-
sponding rate in 1 is already 11.7(9) s-1 at -50 °C. Also,
isomer interconversion by the same mechanism in
Os2Rh(CO)8(η5-C5H5)PMe3 was reported to occur only
above -20 °C and a comparable rate constant of 18(1)
s-1 was observed only at +10 °C.19
The greater facility of this process in 1 is reflected by
the free energy of activation of 11.8(4) kcal‚mol-1 at 215
K when compared with ∆Gq ) 15.1(4) kcal‚mol-1 for
303
Os2Rh(CO)8(η5-C5H5)PMe3,19 ∆Gq290 ) 15.0(4) kcal‚mol-1
for Os3(CO)10[P(OMe)3]2, and ∆Gq271 ) 13.8(4) kcal‚mol-1
(36) Sandstro¨m, J . Dynamic NMR Spectroscopy; Academic Press:
London, 1982; p 53.

4466 Organometallics, Vol. 15, No. 21, 1996
Cooke et al.
100 MHz exactly.38 For the present samples, spectrometer,
and solvent, this corresponds to 86.629 MHz. The NMR
sample tubes were flame-sealed under vacuum. Fast atom
bombardment mass spectra (FAB-MS) were recorded on an
AEI-MS9 mass spectrometer with positive xenon ionization
(+FAB). Elemental analyses were performed by the Micro-
analytical Laboratory of this department.
for Os3(CO)9[P(OMe)3]3.26a Interestingly, a comparably
low value of 10.6(4) kcal‚mol-1 at 228 K was found for
the trigonal twist process which brings about isomer
interconversion in Os3(CO)8[P(OMe)3]4.26a The source
of the steady decrease in activation energy in the
Os3(CO)12-x[P(OMe)3]x (x ) 2, 3, 4) series was attributed
to the activating effect of the phosphite ligands through
electron donation. However, it is not immediately
apparent why the restricted trigonal twist mechanism
in 1 is so facile.37
Rea ction of Os2(CO)8(µ-C2H4) w ith (η2-C2H4)P t(P P h 3)2.
Os2(CO)8(µ-C2H4) (62.4 mg, 0.0987 mmol) and 15 mL of
benzene were placed in a 3-necked 100 mL flask, and 15 mL
of a 5.0 mg/mL benzene solution of (η2-C2H4)Pt(PPh3)2 (75 mg,
0.100 mmol) was added to the stirred Os2(CO)8(µ-C2H4) solu-
tion via a cannula at room temperature. An immediate color
change from pale yellow to orange was observed. The solution
was stirred overnight (ca. 16 h) and the solvent was removed
in vacuo, leaving an orange residue. The residue was ex-
tracted with 3 × 1 mL of CH2Cl2 and was loaded, under argon,
onto a 20 × 4 cm silica-gel column packed in hexane. The
column was eluted with 5:1 hexane/CH2Cl2. Two mobile bands
(yellow and orange) separated cleanly. The solvent was
removed in vacuo from each fraction. The yellow residue was
crystallized from pentane at -80 °C to yield a yellow powder
of uncertain composition (5.3 mg) (see below for explanation),
and the residue from the orange band was crystallized from
CH2Cl2/pentane to give Os2Pt(CO)8(PPh3)2 (1) as an orange
powder (95.2 mg, 73%). An analytically pure sample was
obtained by recrystallizing the orange powder from diethyl
ether/pentane. Anal. Calcd for C44H30O8P2PtOs2: C, 39.91;
H, 2.28. Found: C, 40.04; H, 2.16. IR: (pentane, νCO) 2074
w, 2031 s, 2025 sh, 2012 w, 1989 s, 1983 s, 1967 m, 1956 w,
1948 sh, 1942 sh, cm-1; (CH2Cl2, νCO) 2072 w, 2025 s, 2008 sh,
Con clu sion s
Octacarbonyldiosmacyclobutane, Os2(CO)8(µ-η1,η1-
C2H4), reacts readily with (η2-C2H4)Pt(PPh3)2 to gener-
ate the triangulo-heterotrinuclear cluster Os2Pt(CO)8-
(PPh3)2 (1) and, as such, may prove to be a useful
synthon in the production of metal frameworks of
greater complexity.
In solution, compound 1 exists as a system of three
interconverting isomers which were characterized by a
combination of 13C, 31P, and 195Pt NMR spectroscopies.
Of particular spectroscopic interest was the observation
3
of J P-C coupling when the triphenylphosphine and
carbonyl ligands were disposed in a trans arrangement
along the extensions of an osmium-platinum bond. This
type of long-range coupling between a carbonyl and
phosphine is not commonly observed.
1984 s, 1964 sh, 1951 sh, 1939 sh, cm-1 1H NMR (360 MHz,
.
Magnetization transfer experiments were conducted
at low temperature to establish the mechanisms and
energetics of the isomerization processes. One pathway
involved a low-energy restricted trigonal twist motion
at the phosphine-substituted osmium center of isomers
1a and 1c. Of greater interest was the 1a /1b isomer-
ization which occurred via a rearrangement of the
flexible triangular metal framework and involved rota-
tion of the diosmium fragment, in a metal-olefin
fashion, about an axis extending from the platinum
center.
CDCl3): δ 7.4-7.5 (m, P(C6H5)3). 31P{H} NMR (162 MHz):
3
CD2Cl2, -70 °C, δ 22.18 (d, J P-P ) 49 Hz, with Pt satellites,
3
1J Pt-P ) 2913 Hz, Pt-P, 1a ), 19.54 (d, J P-P ) 2 Hz, with Pt
satellites, 1J Pt-P ) 2990 Hz, Pt-P, 1b), 19.00 (d, 3J P-P ) 5 Hz,
1
3
with Pt satellites, J Pt-P ) 2756 Hz, Pt-P, 1c), 3.29 (d, J P-P
2
) 2 Hz, with Pt satellites, J Pt-P ) 85 Hz, Os-P, 1b), 3.19 (d,
3J P-P ) 49 Hz, with Pt satellites, J Pt-P ) 60 Hz, Os-P, 1a ),
2
3
2
2.48 (d, J P-P ) 5 Hz, with Pt satellites, J Pt-P ) 25 Hz, Os-
P, 1c), 1a :1b:1c ) 4.7:3.4:1.0; toluene-d8, +95 °C, δ 20.02 (d,
3J P-P ) 30 Hz, with Pt satellites, J Pt-P ) 2930 Hz, Pt-P),
1
3.50 (d, 3J P-P ) 30 Hz, with Pt satellites, J Pt-P ) 48 Hz, Os-
2
P). 195Pt NMR (85.6 MHz, CD2Cl2, -70 °C): δ 794.7 (dd, 1J Pt-P
2
1
) 2980 Hz, J Pt-P ) 75 Hz, 1b), 766.3 (dd, J Pt-P ) 2910 Hz,
Exp er im en ta l Section
2J Pt-P ) 60 Hz, 1a ), 696.5 (d, J Pt-P ) 2750 Hz, 1c); 13CO
1
enriched (g50%), δ 794.7 (d, J Pt-P ) 3000 Hz, with 13C
1
Gen er a l P r oced u r es. All manipulations were performed
under a static atmosphere of purified nitrogen or argon using
standard Schlenk techniques. Solvents were dried by refluxing
under nitrogen with the appropriate drying agent and were
distilled just prior to use. CD2Cl2 was dried over P2O5 and
vacuum distilled, while CDCl3 and toluene-d8 were dried over
molecular sieves prior to NMR sample preparation. Os2(CO)8-
(µ-η1,η1-C2H4),10 13CO-enriched Os2(CO)8(µ-η1,η1-C2H4),18 and
1
1
satellites, J Pt-C ) 1490 Hz, 1b), 766.3 (d, J Pt-P ) 2910 Hz,
with 13C satellites, J Pt-C ) 1530 Hz, 1a ), 696.5 (d, J Pt-P
)
1
1
2770 Hz, with 13C satellites, J Pt-C ) 1520 Hz, 1c). 13C NMR
(100.6 MHz, CD2Cl2, CO region only): -70 °C, δ 195.9 (s, 2C,
1b), 195.3 (s, 2C, 1c), 194.2 (s, 2C, 1a ), 190.7 (s, 2C, 1a ), 189.0
1
3
(s, 2C, 1b), 188.8 (s, 2C, 1c), 181.7 (d, J P-C ) 15 Hz, 1C, 1c),
181.1 (s, 1C, 1b), 180.5 (s, 1C, 1c), 180.2 (s, 1C, 1a ), 180.1 (d,
20
(η2-C2H4)Pt(PPh3)2 were prepared by published procedures.
3J P-C ) 14 Hz, 1C, 1b),39 179.2 (s, 1C, 1a ), 177.7 (s, with Pt
satellites, 1J Pt-C ) 1540 Hz, Pt-C, 1a + 1c), 177.3 (d, 3J P-C
)
Infrared spectra were recorded on a Bomem MB-100 FT-IR
spectrometer, and NMR spectra were obtained on Bruker WM-
360 (1H) and AM-400 (13C, 31P, 195Pt) spectrometers. 1H and
13C NMR chemical shifts (δ) were internally referenced to
solvent and are reported in ppm relative to tetramethylsilane
(TMS), while 31P NMR chemical shifts were externally refer-
enced to 85% H3PO4. 195Pt NMR chemical shifts are reported
relative to δ ) 0.0 ppm at the standard frequency of 21.4 MHz
as related to the proton resonance of TMS at a frequency of
1
20 Hz, with Pt satellites, J Pt-C ) 1490 Hz, Pt-C, 1b), 176.5
(s, 1C, 1c), 175.4 (s, 1C, 1a ), 172.7 (s, 1C, 1b). 31P{H} NMR
on a 13CO-enriched (g50%) sample (162 MHz, CD2Cl2, -70
°C): δ 22.18 (d, 3J P-P ) 49 Hz, with Pt satellites, 1J Pt-P ) 2913
Hz, Pt-P, 1a ), 19.54 (d, 3J P-P ) 2 Hz, with Pt satellites, 1J Pt-P
) 2984 Hz, with 13C satellites, 3J P-C ) 14 Hz, Pt-P, 1b), 19.00
(d, 3J P-P ) 5 Hz, with Pt satellites, 1J Pt-P ) 2748 Hz, with 13C
3
3
satellites, J P-C ) 15 Hz, Pt-P, 1c), 3.29 (d, J P-P ) 2 Hz,
28
(37) Isomerization of Re2Pt(µ-H)2(CO)8(PPh3)2 also occurs by a
(38) (a) Kidd, R. G.; Goodfellow, R. J . In NMR and the Periodic
Table; Harris, R. K., Mann, B. E., Eds.; Academic: London, 1978; p
250. (b) Goodfellow, R. J . In Multinuclear NMR; Mason, J ., Ed.;
Plenum: New York, 1987; p 533.
(39) The resonance at δ 180.1 is overlapped with the singlet at δ
180.2 at -70 °C; the doublet feature is ascertained from the -50 °C
spectrum where a combination of line broadening and temperature
dependance of the chemical shifts has separated the resonances (see
Figure 5).
restricted trigonal twist involving the phosphine ligand, but in this
case, the exchange involves a fac to mer rearrangement of the carbonyl
ligands (i.e. an axial to equatorial phosphine migration) whereas, in
Os2Pt(CO)8(PPh3)2, the carbonyls remain in a mer configuration (i.e.
the phosphines remain equatorial). In Re2Pt(µ-H)2(CO)8(PPh3)2, the
activation energy for the isomerization process was determined to be
17.7(4) kcal/mol, which is much larger than that for the 1a a 1c
exchange in Os2Pt(CO)8(PPh3)2.

Os2Pt(CO)8(PPh3)2
Organometallics, Vol. 15, No. 21, 1996 4467
3
3
with 13C satellites, J P-C ) 22 Hz, Os-P, 1b), 3.19 (d, J P-P
)
Ta ble 1. Cr ysta l Da ta a n d Su m m a r y of In ten sity
Da ta Collection a n d Str u ctu r e Refin em en t for
Os2P t(CO)8(P P h 3)2 (1a )
3
49 Hz, Os-P, 1a ), 2.48 (d, J P-P ) 5 Hz, Os-P, 1c). FAB-MS
(m/ e): M+ - nCO, n ) 0-4, 6, 8.
The first yellow fraction appears to contain predominantly
a species of empirical formula Os2Pt(CO)9PPh3 judging from
formula
fw
C44H30O8Os2P2Pt
1324.15
orthorhombic
Pn21a
1
the 31P NMR (δ 22.72 (s, with Pt satellites, J Pt-P ) 3019 Hz,
cryst syst
space group
temp, °C
cell constantsa
a, Å
1
Pt-P)), 195Pt NMR (δ 708.5 (d, J Pt-P ) 3010 Hz); 13CO
18
1
enriched (g50%) δ 708.5 (d, J Pt-P ) 3040 Hz, with 13C
1
satellites, J Pt-C ) 1500 Hz)), and 13C NMR in the carbonyl
16.316(9)
14.502(9)
17.605(9)
4165.6
4
2.11
101.0
region (-70 °C, δ 187.4 (s, 2C, COax), 185.3 (s, 2C, COax), 176.6
b, Å
1
(s, with Pt satellites, J Pt-C ) 1520 Hz, Pt-CO), 175.4 (3C,
c, Å
overlapped, COeq), 173.1 (s, 1C, COeq)). In particular, the 195Pt
NMR spectrum establishes that the platinum center has both
phosphine and carbonyl ligands, while the 31P NMR spectrum
indicates the presence of a species with only one type of
phosphine ligand. However, due to the small amount of
material recovered and the presence of inseparable impurities,
an analytically pure sample could not be isolated for complete
characterization by mass spectrometry and elemental analysis.
V, Å3
Z
D
µ
calc, g cm-3
calc, cm-1
diffractometer/scan
Enraf-Nonius CAD-4/ω-2θ
58/100
range of rel transm factors, %
radiation, graphite monochromator Mo KR (λ ) 0.710 73 Å)
max cryst dimens, mm
scan width
0.05 × 0.18 × 0.25
0.80 + 0.35 tan θ
800; 0,10,0; 008
(1.5
31P NMR Sp in In ver sion Ma gn etiza tion Tr a n sfer Ex-
p er im en ts. Temperatures were monitored by the use of an
external Sensotek BAT-10 copper thermocouple inserted into
a toluene solution in a 5 mm o.d. NMR tube and are considered
accurate to (0.5 K. The temperature at the probe was allowed
to equilibrate for 15 min and was monitored until a steady
reading was obtained over a 10 min interval. The temperature
was remeasured following the completion of the experiments
in order to ascertain any thermal drift. The mean temperature
was then used for subsequent calculations.
std reflcns
decay of stds, %
reflcns measd
2θ range, deg
4102
2 e 2θ e 50
+19,+17,+20
1434
range of h,k,l
reflcns obs [Fo g 5σ(Fo)]b
computer programsc
struct solution
no. of params varied
weights
SHELX44
SHELXS46
171
[σ(Fo)2 + 0.0004Fo
]
2
-1
GOF
2.50
R ) ∑||Fo| - |Fc||/∑|Fo|
0.068
0.077
0.068
The selective inversion magnetization transfer experiments
were carried out by selectively inverting the 31P NMR reso-
nance for the platinum-bonded phosphorus in either isomer
1a or isomer 1b and monitoring the temporal behavior of the
31P magnetizations of the platinum-bonded phosphorus atoms
in all three isomers by sampling the longitudinal magnetiza-
tions at time τ after the inversion. Selective inversion was
accomplished using DANTE,40 spectra at 15-20 values of τ
were collected, and the magnetizations (integrated peak
intensities) were determined using a common set of spectral
processing parameters for each series of experiments. The set
of magnetization data for all values of τ (for a particular
inversion and temperature) were analyzed by least squares31
to obtain the values of the rate constants k1 and k2 and
estimates of their uncertainties. In order to obtain more
reliable estimates of the rate constants, the longitudinal 31P
relaxation times (T1) for each of the platinum-bound phospho-
rus nuclei were determined by the standard inversion-recovery
method41 at each temperature so that the relaxation param-
eters in the least-squares iterations could be fixed at the
corresponding observed values as suggested by Muhandiram.31
The equilibrium constants, K1 and K2, were obtained from the
peak intensities, and simple linear fits of the raw equilibrium
constant-temperature data gave improved estimates of the
equilibrium constants used in the exchange matrices for the
least-squares analysis.42 The equilibrium constants and the
values of the rate constants from the least-squares analyses
at their respective temperatures are as follows: 208.4(5) K,
k1 ) 0.14(1) s-1, k2 ) 2.1(2) s-1, K1 ) 0.70(1), K2 ) 0.21(1);
213.5(5) K, k1 ) 0.26(2) s-1, k2 ) 4.0(2) s-1, K1 ) 0.71(1), K2 )
Rw
R inverse configuration
largest feature final diff map, e Å-3 0.3
a
Least-squares refinement of ((sin θ)/λ)2 values for 17 reflections
b
θ > 14°. Corrections: Lorentz-polarization and absorption
(empirical, ψ scan). c Neutral scattering factors and anomalous
dispersion corrections from ref 45.
activation parameters are the nonlinear errors43 as determined
from the least-squares regression.
13C NMR Sp in Sa tu r a tion Tr a n sfer Exp er im en ts. Con-
ventional double-resonance experiments36 were employed to
selectively irradiate specific resonances in the carbonyl region
of the 13C NMR spectrum at -60 °C. Irradiation of the signal
at δ 194.2 (COax, 1a ) resulted in a 39% decrease in the signal
at δ 195.3 (COax, 1c), a 36% loss of intensity in the doublet at
δ 181.7 (COeq, 1c), a 30% decrease in the signal at δ 179.2
(COeq, 1a ), and a 6% decrease in the signal at δ 195.9 (COax
,
1b). Spin saturation of the signal at δ 179.2 (COeq, 1a ) resulted
in a 32% decrease in the signal at δ 195.3 (COax, 1c), a 27%
loss of intensity in the doublet at δ 181.7 (COeq, 1c), a 19%
decrease in the signal at δ 194.2 (COax, 1a ), and a 3% decrease
in the signal at δ 195.9 (COax, 1b).
X-r a y Cr ysta l Str u ctu r e Deter m in a tion of Os2P t(CO)8-
(P P h 3)2 (1a ). An orange single crystal of the title compound
was mounted in a thin-walled glass capillary flushed with
argon and transferred to the goniometer. The space group was
determined to be either the centric Pnma or the acentric Pn21a
from the systematic absences. The subsequent solution and
refinement of the structure was carried out in the acentric
space group Pn21a. A summary of the data collection param-
eters is given in Table 1.
0.20(1); 217.6(5) K, k1 ) 0.42(2) s-1, k2 ) 6.5(4) s-1, K1
)
0.71(1), K2 ) 0.18(1); 222.7(5) K, k1 ) 0.74(2) s-1, k2 ) 11.7(9)
s-1, K1 ) 0.72(1), K2 ) 0.17(1).
The rate constants were fit to the linear form of the Eyring
equation by a least-squares linear regression routine. The
activation parameters obtained for the two processes were as
follows: k1 (1a /1b exchange), ∆Hq ) 10.1(2) kcal‚mol-1 and
∆Sq ) -13.1(11) eu; k2 (1a /1c exchange), ∆Hq ) 10.7(3)
kcal‚mol-1 and ∆Sq ) -5.2(12) eu. The errors reported in the
The available crystals of the compound were small, and the
highest observed to measured ratio was only 35%. This,
(42) The exchange matrices for the 1a /1b and 1a /1c exchange
processes are as follows:
1/K1
-1/K1
0
1/K2
-1
1
0
0
0
-1
0
0
0
0
Π(1b/1b) )
Π(1a /1c) )
0
(
)
(
)
(40) Morris, G. A.; Freeman, R. J . Magn. Reson. 1978, 29, 433.
(41) Farrar, T. C.; Becker, E. D. Pulse and Fourier Transform NMR;
Academic: New York, 1971; pp 20-22.
-1/K2
0
1
(43) Marquardt, D. W.; Bennett, R. G.; Burrell, E. J . J . Mol.
Spectrosc. 1961, 7, 269.

4468 Organometallics, Vol. 15, No. 21, 1996
Cooke et al.
coupled with the pseudosymmetry problem outlined below,
above led to the final values of R ) 0.068 and Rw ) 0.077. The
final values of the positional and thermal parameters are given
in the Supporting Information.
resulted in
a poorly refined structure. While the gross
structural features are quite clear, the overall refinement is
poor.
While the space group choice was not unambiguous, the
molecule cannot support a mirror plane perpendicular to the
Os2Pt triangle and so the acentric Pn21a was attempted first.
Pseudomirror correlations became immediately obvious with
Os(2) residing on what would be a mirror plane in Pnma.
Closer examination revealed that if the mirror were present,
Pt and Os(1) would be disordered, but only C(3)-O(3) and
C(4)-O(4) would not be close to a mirror position. The mirror
symmetry is also broken by the staggered nature of the phenyl
groups near the Pt-Os(1) axis, but this is not enough to
prevent the pseudomirror correlations which play havoc with
the refinement.
The current refinement is in Pn21a. The phenyl groups were
refined as rigid groups although the individual thermal
parameters were allowed to refine freely. Only the Pt and Os
atoms were refined anisotropically. The hydrogen atoms were
not included in the final refinement. Refinement as described
Ack n ow led gm en t. We thank the Natural Sciences
and Engineering Research Council of Canada (NSERC)
for funding and for an Undergraduate Summer Re-
search Award to J .C. Financial support from University
of Alberta is also gratefully acknowledged. The expert
assistance of Mrs. G. Aarts and Mr. G. Bigam of the
University of Alberta High Field NMR Laboratory in
carrying out the 31P NMR spin inversion magnetization
transfer and 13C NMR spin saturation transfer experi-
ments is sincerely appreciated. We are also grateful to
Dr. J ohn Washington for assistance with the variable-
temperature NMR experiments and Dr. Frank Seils for
important initial work on this project.
Su p p or tin g In for m a tion Ava ila ble: Tables of fractional
coordinates with Be values, thermal parameters, and bond
distances and angles and an ORTEP diagram for isomer 1a
(6 pages). Ordering information is given on any current
masthead page.
(44) Sheldrick, G. M. SHELX76, a system of computer programs for
X-ray structure determination as locally modified, University of
Cambridge, England, 1976.
(45) Sheldrick, G. M. SHELXS. Acta Crystallogr. 1990, A46, 467.
(46) International Tables for X-ray Crystallography; Kynoch: Bir-
mingham, England, 1974; Vol. IV, pp 72, 99, 149.
OM950938E