Facile C(sp2)/O2CR bond cleavage by Ru or Os

Article abstract of DOI:10.1039/b306111f

[RuHC1L₂]₂, L = PⁱPr₃, reacts with H₂C=CH(O₂CR) (R = CH₃, CF₃, C₆H₅) during mixing at 20 °C, via two observable intermediates, to give RuCl(O

Full text of DOI:10.1039/b306111f

View Article Online / Journal Homepage / Table of Contents for this issue
Facile C(sp2)/O2CR bond cleavage by Ru or Osy
a
German Ferrando, Joseph N. Coalter, III, Helene Gerard, Dejian Huang, Odile Eisenstein*
a
b
a
b
´ `
´
and Kenneth G. Caulton*a
a
Department of Chemistry, Indiana University, Bloomington IN 47405-7102, USA.
E-mail: caulton@indiana.edu
b
`
´
Laboratoire de Structure et Dynamique des Systemes Moleculaires et Solides, UMR 5636,
CC14, Universite de Montpellier 2, France. E-mail: eisenst@lsd.univ-montp2.fr
´
Received (in London, UK) 30th May 2003, Accepted 31st July 2003
First published as an Advance Article on the web 9th September 2003
i
ꢀ
=
[RuHClL2]2 , L ¼ P Pr3 , reacts with H2C CH(O2CR) (R ¼ CH3 , CF3 , C6H5) during mixing at 20 C, via two
observable intermediates, to give RuCl(O2CR)(CHMe)L2 ; this carbene complex then redistributes the Cl and
O2CR groups. Vinyl tosylate gives RuCl(OTs)(CHMe)L2 already at ꢁ60 ꢀC. Vinyl chloroformate,
=
H2C CH(O2CCl) reacts rapidly with [RuHClL2]2 to give the olefin metathesis catalyst RuCl2(CHMe)L2 and
i
t
=
CO2 . Os(H)3ClL2 (L ¼ P Pr3 or P Bu2Me) reacts with vinyl esters H2C CHE (E ¼ O2CR) to form first an
Z2-olefin adduct. This is followed by C/O bond cleavage, giving the carbyne OsHCl(O2CCF3)(CMe)L2 .
Vinyl chloroformate and Os(H)3ClL2 gives OsHCl2(CMe)L2 and CO2 . RuHCl(PPh3)3 reacts with vinyl
chloroformate, via RuCl(O2CCl)(CHMe)(PPh3)2 , to give RuCl2(CHMe)(PPh3)2 while OsHCl(PPh3)3 reacts
analogously, through observable OsCl2(CHMe)(PPh3)2 , to form OsHCl2(CMe)(PPh3)2 . Vinyl trifluoroacetate
+
converts OsHCl(PPh3)3 , to OsHCl(O2CCF3)(CMe)(PPh3)2 . The less p-basic metal in OsH(CO)(PtBu2Me)2
reacts with vinyl esters to give only an olefin adduct; detectable binding of the ester oxygen to Os in this adduct
suggests a mechanism for carboxylate migration from carbene carbon to metal. The mechanisms of these
reactions are explored, and the thermodynamic disparity between Ru and Os is discussed. DFT (B3PW91)
calculations have been carried out to establish the energy pattern of possible products. The thermodynamic
preference for cleaving the C–O2CR bond is shown to have a thermodynamic origin associated with the energy
of the formed Ru–O2CR bond. The calculations also indicate the very large thermodynamic driving force for
=
loss of CO2 in the case of H2C CH(O2CCl). The corresponding loss of CO2 is shown to be thermodynamically
=
unfavorable in the case of H2C CH(O2CR). The energy of the Ru-R bond is a key factor.
synthesis depends in part on p-donation of an OR lone pair
=
to the carbene carbon. When M Os, the reaction proceeds
Introduction
[RuHClL2]2 and Os(H)3ClL2 (L ¼ bulky phosphine) have
both been shown1 to be sources of the reactive, nonplanar
fragment MHClL2 , which react (Scheme 1) with vinyl ethers
at ꢂ20 ꢀC to form first an alkyl, then, by a-H migration to
M, a hydrido carbene. This is effectively an isomerization of
a vinyl ether to a heteroatom-substituted carbene, and the
favorable thermodynamics for this remarkably easy carbene
further (Scheme 1) because this metal has a preference (vs.
Ru) for higher oxidation states and for an 18-electron config-
uration; OR migrates to Os, to leave behind a carbyne ligand.
=
Vinyl esters, H2C CH[OC(O)R], have a better leaving
group as substituent on the vinyl carbon. We also report
=
here that chloroformates, H2C CHOC(O)Cl, furnish a very
convenient synthesis of one Grubbs olefin metathesis catalyst,
and have studied the Os analog of this reaction. Finally, we
have studied the reaction of several vinyl esters with an osmium
analog reagent, OsH(CO)(PtBu2Me)2+, where the p-acidity of
Os is much reduced by formal replacement of a p-donor Clꢁ
by the p-acid CO. This provides a useful mechanistic insight.
Results
1. Reaction of [RuHCl(PiPr3)2]2 with vinyl esters
A) Products. Reaction of [RuHCl(PiPr3)2]2 with equimolar
=
(ester : Ru), CH2 CH(O2CR) (R ¼ Me, CF3 , Ph) at room
temperature in arene solvent forms the ethylidene complex,
=
i
RuCl(O2CR)(P Pr3)2( CHMe)inthetimeofmixing(Scheme2).
Scheme 1
After 1 hour, however, anionic ligand redistribution leads to
=
i
50 mole % conversion to RuCl2(P Pr3)2( CHMe), with the bal-
y Electronic supplementary information (ESI) available: 1H NMR
spectra in toluene-d8 for the reaction of compound IV and vinyl
trifluoroacetate mixed at ꢁ80 ꢀC and 1H, 31P(1H) and 13C(1H)
NMR spectra of the two isomers of compound IX. See http://
www.rsc.org/suppdata/nj/b3/b306111f/
ance degrading to intractable materials. Though complexes
similar to the second expected redistribution product, Ru(O2-
i
CR)2(P Pr3)2( CHMe) have been reported,2 we find no
spectroscopic evidence for their persistence here.
=
DOI: 10.1039/b306111f
New J. Chem., 2003, 27, 1451–1462
1451
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2003

View Article Online
ð2Þ
mechanism is less attractive here since we begin with Ru(II) and
with a paucity (14) of valence electrons. Concerning the
mechanism of the transformation in Scheme 2, we have shown
earlier that Ru–H addition to a triple bond is faster than oxi-
Scheme 2
dative addition of the (acidic) CH bond of terminal alkynes
6
(i.e., the reaction of RuH(H2)XL2 with RC CH ), so this is
=
=
more consistent with the occurrence of eqn. 2 vs. eqn. 1. b-alk-
oxide migration analogous to (eqn. 2a) has been demon-
strated7 on Pt(II). Nevertheless, because the reaction of vinyl
ethers goes via the regiochemistry of eqn. 2b, the simplest
(i.e., unified) logic would propose the analogous alkyl inter-
mediate for vinyl carboxylate reactions as for vinyl ethers.
B) Mechanism. Combining [RuHCl(PiPr3)2]2 and vinyl
acetate in toluene-d8 at ꢁ60 ꢀC and warming to room tem-
perature with periodic spectroscopic monitoring reveals
initial
conversion
to
an
adduct,
(I)
C) A faster leaving group: vinyl tosylate. Use of vinyl tosy-
late in place of vinyl acetate for this isomerization allows a
with inequivalent
more detailed examination of the decomposition of RuCl(X)
=
i
(P Pr3)2( CHMe), since with this olefin, carbene formation is
much more rapid than anionic ligand redistribution. When
these reagents are combined at ꢁ60 ꢀC in toluene-d8 , quantita-
2
phosphines (32 and 54 ppm; JP–P ¼ 283 Hz) by 31P{1H}
NMR and a hydride doublet of doublets at ꢁ18.3 ppm
(1H NMR). Based on the correlation of hydride chemical shift
with the donor power of the ligand trans to itself, the similar
i
tive formation of RuCl(OTs)(P Pr3)2( CHMe) is observed
0
=
with no redistribution to the dichloro ethylidene complex.
However, as the temperature approaches ꢁ20 ꢀC, the forma-
i
tion of RuCl2(P Pr3)2( CHMe) (i.e., halide redistribution) is
=
hydride chemical shift of this adduct to that of [RuHCl-
1
substantial, and again, the resulting tosylate complexes were
not identified.
(PiPr3)2]2 with CH2 CH(OEt) suggests that the adduct
=
A recent report8 shows, in an only partially relevant
example, how a carboxylate ligand can be responsible for a
significant change in product type. Reaction of RuCl2-
involves coordination through only the olefinic portion of
the molecule (coordination Z1 through oxygen would render
the phosphines equivalent). Warming the sample to ꢁ30 ꢀC
shows the appearance of a second intermediate with inequiva-
lent phosphines at 34 and 48 ppm (31P{1H} NMR; signals too
(CHPh)(PCy3)2 with 2 AgO2CCF3 does not give Ru-
(O2CCF3)2(CHPh)(PCy3)2 , but instead III.
0
broad to determine JP–P ) and no corresponding hydride peak.
We assign this species as having olefin inserted in the Ru–H
bond, which could be stabilized by intramolecular coordi-
nation of the acetate carbonyl oxygen to the otherwise 14eꢁ
ruthenium center (II). The asymmetric carbon renders the
phosphines inequivalent (this would not be true if the insertion
created a Ru(CH2CH2O2CMe) ligand), and the bulk of the
secondary alkyl and/or Z2 binding mode causes hindered rota-
tion around Ru–P bonds3 and the broad phosphine reso-
nances. Above this temperature, two carbene products form
with 31P{1H} singlets at 46.4 and 35.5 ppm and correspon-
Thus, one phosphine is
1
ding H NMR quartets in the 19–20 ppm range. The first to
appear we attribute to the new, mixed anionic species,
lost, and the bidentate potential of carboxylate is realized, by
bridging two metals. Adventitious water then occupies a brid-
ging site, and hydrogen bonds to two Z1-carboxylates. It is
noteworthy that the carbene ligands persist, and in terminal
i
RuCl(O2CR)(P Pr3)2( CHMe), and the second is RuCl2-
=
i
4
=
(P Pr3)2( CHMe), by comparison to literature data. The
propensity for coordinated acetates to adopt multiple coordi-
nation modes (e.g. Z1, Z2, m2–4) is likely the mechanism for
the decomposition of the mono/bis(acetato) ruthenium species
via acetate aggregates. No evidence for butenes (i.e., coupled
ethylidenes) was seen, although a small amount of ethylene
in addition to free phosphine was observed in several of the
samples.
sites. There have also been brief references to RuCl(X)-
=
9
( CHCH2R)(PCy3)2 for X ¼ CF3CO2 or CN and to bis(tri-
fluoroacetato) ruthenium carbenes bearing triaryl phosphines.2
D) A vinyl ester also containing a reactive C–Cl bond. The
ability of a pendant donor to bind in the intermediate of addi-
tion of Ru–H across the olefin double bond can also be seen in
reaction of [RuHCl(PiPr3)2]2 with vinyl chloroformate,
Possible reaction mechanisms of vinyl carboxylates with
metal hydrides have been discussed previously5 as C–O oxi-
dative addition (eqn. 1) or M–H addition to olefin, followed
by migration of X to M from either Ca or Cb (eqn. 2). The
oxidative addition
ð1Þ
ð3Þ
1452
New J. Chem., 2003, 27, 1451–1462

View Article Online
=
CH2 CH(O2CCl). The combination of these reagents at room
The alkyl complex. The alkyl complex 2Ru with the acetate
group on the a carbon is calculated to be 2.5 kcal molꢁ1 more
stable than 1Ru. The geometry is similar to that of other
Ru(alk)ClL2 complexes in that it has an angle of Cl–Ru–
C ¼ 105.9ꢀ. In contrast to the ethyl analog where only an
agostic C–H interaction could provide some additional stabili-
zation to the highly unsaturated metal center, the O2CMe
group can give additional electron density to the metal via
the terminal oxygen. Thus, the Ru...O distance is short,
temperature (RT) results (eqn. 3) in the immediate, quantita-
=
i
tive formation of dichloro carbene, RuCl2(P Pr3)2( CHMe)
and vigorous CO2 evolution. We believe this represents a
convenient synthesis of this molecule.
Computational studies
˚
˚
2.135 A. The Ca–O bond is significantly elongated (1.519 A).
The two C–O bonds of the acetate group are not equal,
˚
1.244 A for the oxygen bonded to the metal and 1.301 for
The variety of compounds that can be obtained when reacting
=
substituted alkenes CH2 CHG with [RuHClL2]2 or OsH3ClL2
can be better understood through a computational study of the
relative energies of possible products. Such study is thus infor-
mative of the thermodynamic control of the reactions but not
of the mechanisms that could lead to the products. The
mechanism described in Scheme 1 has been suggested for
RO-substituted olefins1 and can be reasonably applied for sub-
stituted systems. For these reasons, olefin, alkyl, and carbene
structures with the O2CR group in various possible sites have
been calculated. As in previous studies, MHCl(PiPr3)2
(M ¼ Ru, Os) is modeled by MHCl(PH3)2 which will be
denoted as [M]. The relative energy schemes of most relevant
minima are shown in Fig. 1 (O2CMe) and Fig. 2 (O2CCl).
the oxygen bonded to Ca . The angles at the Ca center are also
somewhat distorted (Ru–C–C ¼ 118.9ꢀ and Ru–C–H ¼
111.3ꢀ) but not sufficiently to suggest that Ca is already an
sp2 center. The acetate group is bridging across the Ru–C
bond and the complex is best viewed as a square-pyramidal
16-electron complex with an apical alkyl group. The stability
of 2Ru relative to the olefin complex 1Ru is however not
entirely due to the Ru–O interaction. Previous studies on sub-
stitution at the alkyl group of the same metallic fragment have
shown the stabilizing influence of an electron acceptor group
on Ca . The same effect applies here. The complex with the
acetate group on the b carbon was not calculated. Previous
studies for the vinyl ether with the same fragment have shown
a preference for the OMe group to be on Ca .10
=
CH2 CH(O2CMe) and RuHCl(PH3)2
The olefin adducts. The geometry of selected complexes is
=
given in Fig. 3. The coordination of CH2 CH(O2CMe) to
The carbene complexes. The alkyl complex can give several
products depending on which atom or group migrates to Ru.
From the a substituted alkyl complex, migration of H gives
the (MeCO2)CMe complex. Several structures have been
found for this complex. The most stable structure, 3Ru, 4.6
kcal molꢁ1 above the most stable olefin complex, 1Ru, has a
square-pyramidal geometry with the carbene group coplanar
with Ru–H and the acetate group towards the Ru vacant coor-
dination site to form an Ru...O bond. It is interesting to com-
pare the metric parameters of the Ru...O interaction in 3Ru to
that11 of the cationic [RuCl(PiPr3)2(C(CH2Ph)OC(O)R)]+. The
[Ru] occurs preferably trans to Cl with the acetate group
towards the empty site, 1Ru. The metal is in a square-pyrami-
dal environment as found for previous olefin complexes of the
same [Ru] fragment.1,10 The complex, 1’Ru, in which the acet-
ate group is towards the hydride and thus with no possible
interaction with Ru is 4.0 kcal molꢁ1 higher in energy. In
1Ru, an interaction is established between Ru and the carbonyl
˚
oxygen of O2CMe as shown by the Ru...O distance, 2.387 A.
˚
This distance is longer by 0.17 A in comparison to the Ru–
ether bond distance but is still compatible with weak bonding.
A proof of this interaction is given by the change in the orien-
=
tation of the olefin. In 1’Ru, the C C bond is parallel to Ru–
˚
Ru...O distance (2.346 A) in 3Ru is significantly longer than in
˚
the cationic complex (2.108A). In the cationic complex, the
incipient Ru...O bond is trans to a chloride ligand. In 3Ru,
the same bond is trans to a hydride ligand and in addition
the system is neutral. These conditions are responsible for a
longer Ru...O distance. This additional Ru...O bond in 3Ru
brings some significant stabilization, since the minimum
deprived of the Ru...O interaction (acetate group away from
the empty coordination site) is 11.1 kcal molꢁ1 higher in
energy, and it is thus mechanistically irrelevant.
Migration of the acetate group from 2Ru leads to a CHMe
complex 4Ru, which is calculated to be more stable than 3Ru
and is 1.1 kcal molꢁ1 below the olefin complex 1Ru. In 4Ru,the
acetate is bidentate with one O pseudo trans to the carbene
group. This causes the acetate group to have unequal Ru–O
H. The geometrical constraint of the Ru...O interaction in 1Ru
=
results in the C C bond being neither parallel to Ru–H nor to
Ru–P (dihedral angle HRuCC ꢃ40ꢀ). Another consequence of
the Ru...O interaction in 1Ru is that the plane of the acetate
group is essentially perpendicular to the plane of the coordi-
nated olefin. However the Ru...O bonding interaction does
not greatly influence the binding energy of the olefin. The
=
binding energy of CH2 CH(O2CMe) in 1Ru is equal to 39.3
kcal molꢁ1 which is similar to that calculated1 for various
CH2 CHG (G H 39.2 kcal molꢁ1, OMe 36.6 kcal molꢁ1
N(H)C(O)Me 39.5 kcal molꢁ1).
,
=
=
˚
˚
bonds, 2.125 A for O trans to Cl and 2.398 A for O pseudo
trans to the carbene (the group with the larger trans influence).
The two CO bond lengths are not very different (1.256 and
˚
1.284 A). The carbene group is parallel to the Ru–P bond.
The Z1–O2CMe isomer 5Ru lies only 3.4 kcal molꢁ1 above
4Ru. In 5Ru, the acetate group plays the role of a strong p
donor as shown by the shortening of the Ru–O bond by 0.09
˚
A. In going from the dihapto to the monohapto geometry,
the acetate has compensated for the loss of the Ru–O bond
by being a strong p donor through the remaining Ru–O bond.
=
b) CH2 CH(O2CCl)
The olefin adducts. The structure of selected products
=
obtained from CH2 CH(O2CCl) are given in Fig. 4. Many
Fig. 1 Calculated (DFT) energies (kcal molꢁ1) of isomeric structures
results are analogous to those obtained with the acetate group.
The vinyl chloroformate coordinates preferably trans to Cl
=
derived from RuHCl (PH3)2 and H2C CH(O2CMe). For clarity, the
two PH3 groups equidistant from the RuHCl plane are not drawn.
New J. Chem., 2003, 27, 1451–1462
1453

View Article Online
Fig. 2 Calculated (DFT) energies (kcal molꢁ1) of isomeric structures derived from RuHCl (PH3)2 and H2C CH(O2CCl). For clarity, the two PH3
groups equidistant from the RuHCl plane are not drawn.
=
with the oxygen towards the empty coordination site of Ru
as expected from an oxygen with lesser donating ability due
to the presence of the electron withdrawing Cl atom. The binꢁd1-
0
˚
(Ru...O ¼ 2.518 A), 6Ru. However the complex 6 Ru deprived
of the Ru...O interaction is only 1.1 kcal molꢁ1 higher in
energy. The Ru...O interaction is weaker than with O2CMe
=
ing energy of CH2 CH(O2CCl) to [Ru] is 40.8 kcal mol
which is very similar to the values given above for several dif-
ferent G groups. This reiterates that the G group has only a
modest influence on the binding ability of the starting olefin
to the metal fragment and also that the G group does not cre-
ate significant Ru..G interaction that would turn a 16-electron
in a 18-electron complex.
The alkyl complexes. The alkyl complex with the chlorofor-
mate group on the a carbon needs to be stabilized by addi-
tional bonding from the chloroformate group in order not
to be a highly unsaturated 14 electron complex. As expected,
˚
the alkyl complex with an Ru...O bond (2.165 A), 7Ru, is pre-
ferred (5.9 kcal molꢁ1 below 6Ru) followed by a complex with
ꢁ1
˚
an Ru..Cl interaction (2.451 A), 8Ru (0.5 kcal mol below
6Ru) and followed by a complex (not illustrated) with a b
C–H agostic bond from the Me group, (7.6 kcal molꢁ1 above
6Ru). This order follows the electron donating ability of the
atom or group interacting with the metal in this formal 14-
electron complex. In the most stable alkyl complex, 7Ru,
˚
the Ca–O bond is significantly elongated (1.551 A), which is
even slightly longer than in the case of the acetate (1.519
˚
A). The overall geometry of 7Ru is similar to that obtained
for the acetate group. The slightly longer Ru...O and C–O dis-
tances for O2CCl compared to O2CMe reflects the electron
withdrawing property of the chloroformate group compared
to the acetate.
The carbene complexes. The carbene complexes could result
from the migration of H or the migration of the chlorofor-
mate group to the metal. The carbene complexes in which
H has migrated to the metal are above 6Ru in energy and
the preferred such complex 9Ru, 5.4 kcal molꢁ1 above 6Ru,
has an Ru...O bonding interaction. The Ru...O distance in
˚
9Ru is 2.421 A, which is longer than that obtained in the
˚
acetate analog 3Ru (2.346 A) in agreement with chemical
intuition. The carbene complex 10Ru with an Ru..Cl interac-
˚
tion (2.621 A) is as expected even higher in energy, 9 kcal
molꢁ1 above 6Ru.
˚
Fig. 3 Detailed geometry (A and degrees) of species from Fig. 1.
1454
New J. Chem., 2003, 27, 1451–1462

View Article Online
˚
Fig. 4 Detailed geometry (A and degrees) of species in Fig. 2.
All complexes resulting from the cleavage of the Ca–O bond
are energetically more stable than those resulting from the
migration of H. The most stable carbene complex, 11Ru, 27
kcal below 6Ru is best viewed as a weak complex of
RuCl2(PH3)2(CHMe) and fully formed CO2 . In 11Ru,the
square-pyramidal RuCl2(PH3)2(CHMe) complex with an api-
metal center and the decarboxylation of the chloroformate.
In order to better understand the factors that control this
result, we have calculated the relative energies of products
which would result from the migration of H (formation of a
CGMe complex) or the migration of G (formation of a CHMe
=
complex) for various CH2 CHG (G ¼ OMe, NHC(O)Me,
cal carbene coordinates CO2 with
O
adjacent to Ru
O2CMe), in which G represents groups with variable p donat-
ing ability. The difference in energy between RuHCl-
(PH3)2(CGMe) and RuClG(PH3)2(CHMe) decreases in the
order OMe (12.8 kcal molꢁ1), N(H)C(O)Me (9 kcal molꢁ1),
OC(O)Me (ꢁ5.7 kcal molꢁ1). Thus only the acetate clearly
favors migration to the metal center. All other ligands favor
the migration of H over that of G in agreement with experi-
ment. The chloroformate favors it even more as evidenced
by the relative energy of 13Ru versus 9Ru (DE ¼ ꢁ12.5 kcal
molꢁ1) but it should be kept in mind that another product
of reaction is even more preferred (contrast Figs. 1 and 2).
An interpretation of the results can be obtained from the
thermodynamic cycles shown in Fig. 5. The H/G exchange
˚
˚
(Ru...O ¼ 2.889 A) and C adjacent to Cl (3.199 A). The linear
geometry of the CO2 unit (O–C–O ¼ 177.5ꢀ) confirms that the
chloroformate group has been dismantled. The CO2 group is in
fact very weakly bonded (2.3 kcal molꢁ1) to Ru. Inclusion of
the entropy factor, which favors dissociation of CO2 , should
account for the observed spontaneous loss of CO2 . The other
isomeric carbene complexes are significantly higher in energy
than 11Ru. We have located another complex of CO2 and
RuCl2(PH3)2(CHMe), 12Ru, in which Cl occupies the apical
site of the square pyramidal structure (16.2 kcal molꢁ1 below
6Ru). An even less stable species, 13Ru, is the 18-electron com-
plex in which the chloroformate is dihapto through both its
oxygens and is therefore comparable structurally to 4Ru.
13Ru is 19.9 kcal molꢁ1 above 11Ru. As in 4 Ru, the two
˚
Ru–O bond lengths are different (2.145 and 2.500 A), the
longer one being trans to the carbene ligand. It should be kept
in mind that 13Ru is 7.1 kcal molꢁ1 below the corresponding
olefin complex while in the case of the acetate group the corres-
ponding complex 4Ru is only 1.1 kcal molꢁ1 below its corres-
ponding olefin complex 1Ru. Thus chloroformate favors even
more than acetate the coordination of the O2CCl group to the
metal fragment. However, there is an even more stable product
that results from the loss of CO2 . The extreme stability of CO2
is most likely the determining parameter for this strong ther-
modynamic preference. The factors favoring loss of CO2 in
the chloroformate and not in the acetate case will be addressed
later in this work, but the calculations are in agreement with
the observed loss of CO2 .
c) The H/G migrating ability: cleaving a C–H versus a C–G
The calculations reproduce well the experimental observations,
i.e. the preference for the migration of the acetate group to the
Fig. 5 Thermodynamic cycle and DFT energies for isomerization of
variable groups G from carbene carbon to Ru.
New J. Chem., 2003, 27, 1451–1462
1455

View Article Online
(step a) is decomposed into a step b which exchanges G
between the sp2 C of the carbene complex and the sp3 C of
CH3–CH3 and a step c which exchanges G between the sp3 car-
bon of CH3CH2G and the metal center. In this manner the
donating ability of G (vs. H) towards an unsaturated carbon
(step b) and the unsaturated Ru center (step c) is approxi-
mately separated. The energy associated with each step for
each group G is given in Fig. 5.The positive character of the
values for step b show that all groups G are more favorable
on the sp2 carbon. Although the values vary by 3–4 kcal
molꢁ1, such variation is small compared to the overall trend
in step a. The negative energies for step c indicate that G is also
preferred on Ru vs. on the sp3 carbon but the variation is large,
with the largest value for the acetate group. The ability to
become a dihapto coordinated G group is not determining
since our studies (above) have shown that the Z2 ! Z1 change
of the acetate is energetically easy.
The calculations clearly show that the strength of the Ru–G
bond (E(c)) is an important criterion in the ability for the
G group to migrate to the metal. It is remarkable that the
Ru–O binding energies are significantly different for the meth-
oxy and acetate ligands. This can reflect the larger electron affi-
nity of the acetate group that better stabilizes a partially ionic
M–O bond.
decarboxylation reaction can only be thermodynamically
favorable if the R group of O2CR also makes a stable Ru–R
bond (Cl vs. CH3).
Since we have not conducted a study of the reaction path, we
can only discuss the decarboxylation mechanism through the
energy and structures of energy minima. The loss of CO2 can
in principle occur from the alkyl complexes or from the car-
=
bene complexes. In the Ru CHMe species, the calculations
for both O2CMe and O2CCl show that the stabilization
derived from taking Z1-carboxylate to the Z2- alternative is
only 2–3 kcal molꢁ1. This additional Ru...O bond is thus not
very strong, perhaps because it is trans to the carbene. In
any event, it shows that an unsaturated complex with Z1-car-
boxylate is thermally accessible; a pendant carboxylate is also
a candidate for the carboxylate/chloride ligand redistribution
observed experimentally and reported above. In the alkyl spe-
cies, the Ru..Cl interaction provides significant stabilization
and the Cl bonded chloroformate (8Ru) is energetically acces-
sible (about 5 kcal molꢁ1 above the Ru...O bonded 7Ru spe-
cies). The alkyl and the carbene complexes are all within 5
kcal molꢁ1 of the olefin adduct and this provides intermediates
to departure of CO2 . As shown by the energy minima 11Ru
and 12Ru, which are very separated-product-like, the forma-
tion of a CO2 ligand is so very exothermic compared to all
other products calculated (16 and 27 kcal molꢁ1 more stable
than the olefin adduct) that the energy gain must rest mainly
on nearly complete formation of the very stable CO2 molecule.
A very similar binding energy and geometry pattern has been
calculated12 for the interaction of RuH2(PH3)3 with CO2 .
d) Discussion of the decarboxylation thermodynamics and
mechanism
The thermodynamics of decarboxylation can be addressed
through the thermodynamic cycle (Fig. 6) that compares the
acetate and chloroformate groups. Step a illustrates the strong
preference for the chloroformate to lose CO2 and form
RuCl2L2(CHMe) from the olefin complex. The analogous
reaction with the acetate group, leading to RuCl(Me)L2(CH-
Me) + CO2 , is calculated to be endothermic. The formation
of the very stable CO2 molecule (which was responsible for
the great stability of 11Ru and 12Ru over 13Ru) is not suffi-
cient to balance other factors in the acetate case. The decom-
position of step a into several consecutive steps shows (Fig.
6) that the single step which dominates the overall reaction
involves exchange of H and Cl between an sp2 carbon and
the metal center. This step (d) is strongly exothermic for the
transformation of Ru–H and C–Cl bonds into Ru–Cl and
C–H bonds while it is endothermic for the transformation of
Ru–H and C–Me bonds into Ru–Me and C–H bonds. The
Experimental study of Os analogs. Since OsHClL2 is
unknown, we have employed Os(H)3ClL2 , with the hope that
two H can be removed, either as H2 , or by hydrogenation of
equimolar vinyl ester.
a) Os(H)3CIL2 + vinyl trifluoroacetate. Reaction of
Os(H)3Cl(PiPr3)2 and vinyl trifluoroacetate in a 1 : 1 mole ratio
in toluene-d8 at 25 ꢀC gives an immediate color change from
brown to green-yellow and gas evolution (presumably H2),
but the 31P NMR spectrum indicates a variety of products,
only a minor one of which is a carbyne; OsH5Cl(PiPr3)2 ,13
a
product of Os(H)3Cl(PiPr3)2 scavenging released H2 , is
observed, as is free ethylene. The reaction was therefore stu-
died with excess vinyl trifluoroacetate, to more effectively cap-
ture all Os(H)3Cl(PiPr3)2 with the vinyl ester rather than be
confused with the reaction of ester and OsH5Cl(PiPr3)2 . Com-
=
bining Os(H)3ClL2 with 4 moles of H2C CH(O2CCF3) gives,
within short time of observation in toluene-d8 , OsH-
Cl(O2CCF3)(CCH3)L2 .
To elucidate the mechanism of this reaction, we combined
Os(H)3Cl(PiPr3)2 (IV) and vinyl trifluoroacetate at ꢁ80 ꢀC in
1 : 1 stoichiometry in toluene-d8 and slowly warmed the mix-
1
ture to room temperature to identify the intermediates by H
and 31P {1H} NMR. At ꢁ80 ꢀC, we observe (Scheme 3, where
‘‘2H’’ indicates either dihydride or H2) the presence of the
reagent Os(H)3Cl(PiPr3)2 (IV), the final product OsH-
i
Cl(O CCF )( CCH )(P Pr ) (IX, two isomers), free H2 (1H
=
=
2
3
3
3 2
NMR), and several intermediates. The intermediates observed
=
i
are the olefin adduct, OsH3Cl(H2C CH(O2CCF3))(P Pr3)2
(V), the chiral alkyl intermediate OsCl(CH(CH3)(O2CCF3))-
(PiPr3)2 (VII, with or without an Os–O bond), and the ethyli-
dene intermediate due to carboxylate migration to the metal
i
=
OsCl(O2CCF3)( CH(CH3))(P Pr3)2 (VIII). The olefin adduct
1
is characterized in the H NMR spectrum by the two signals
(intensity 1 : 2) of H attached to Os. Analogous decoalesced
hydrides have been observed previously with vinyl ether
adducts of Os(H)3ClL2 . The hydride signals resonate as a tri-
plet at ꢁ4.05 ppm with J(H–P) ¼ 20 Hz and a broad peak at
ꢁ14.78 ppm in 1 : 2 intensity ratio. The signals due to the
Fig. 6 DFT energies for the reactions shown, contrasting R ¼ Cl
with R ¼ CH3 (in parentheses).
1456
New J. Chem., 2003, 27, 1451–1462

View Article Online
Scheme 3
coordinated olefin are a broad AB pattern at 2.75, 2.87 and a
broad peak at 3.97 ppm. Surprisingly, the 31P {1H} NMR
spectrum of V displays a sharp singlet at 37.8 ppm, instead
of the AB pattern expected in the presence of a prochiral olefin.
two CF3 carbons, each a quartet. We suggest that the two
isomers are IXa and IXb.
0
This can only be explained if the value of JP–P is much larger
than the value of Dn. The olefin adduct liberates H2 leading
These undergo no change (e.g., population) or redistribution
to OsCl2H(CMe)L2 and Os(O2CCF3)2H(CMe)L2 upon heating
at 70 ꢀC for 24 h in toluene.
i
to the undetected 16 eꢁ OsHCl(H2C CH(O2CCF3))(P Pr3)2
=
(VI), which undergoes insertion of the olefin into the Os–H
bond to give the chiral 14 eꢁ alkyl OsCl(CH(CH3)(O2CCF3))-
(PiPr3)2 (VII). Given the calculated structure of the ruthenium
analog, 2Ru, it is likely that VII has the carboxylate oxygen
bound to Os. This alkyl compound displays two characteristic
multiplets in the 1H NMR: a methyl doublet at 1.60 ppm with
J(H–H) ¼ 6.6 Hz, and a quartet at 0.77 ppm with J(H–H) ¼ 6.6
b) Vinyl chloroformate. Vinyl chloroformate is a bifunc-
tional substrate, and this is evident in its reaction with an
osmium reagent. Reaction of Os(H)3Cl(PtBu2Me)2 with vinyl
chloroformate (1 : 2 mole ratio) in toluene proceeds to com-
plete consumption of the trihydride (eqn. 4) in less than 5
min at 20 ꢀC to give OsHCl2(CCH3)(PtBu2Me)2 (80% yield),
H2 (1H NMR evidence) and CO2 (visible gas evolution). Some
ethylene and some vinyl chloride are also detected (1H NMR).
This reaction is apparently fast enough that
0
Hz, corresponding to the a proton. Again, the JP–P must be
much larger than the value of Dn, which explains the sharp
31P {1H} NMR singlet at 28.9 ppm. The H2 liberated by V is
trapped by unreacted Os(H)3Cl(PiPr3)2 (IV) forming the
known OsH5Cl(PiPr3)2 . At this point, the C–O bond is cleaved
and the acetate group migrates to the metal, forming the 16 eꢁ
i
mixed-halide ethylidene OsCl(O2CCF3)( CH(CH3))(P Pr3)2
OsðHÞ3ClL2 þ H2C
B
CHðO2CClÞ !
OsHCl2ðCCH3ÞL2 þ CO2 þ H2 ð4Þ
=
(VIII), best detected by a broad signal far downfield in the
1H NMR spectrum (18.5 ppm). In addition, another broad
peak is observed at 1.68 ppm, which corresponds to the methyl
substituent on the carbene carbon. The 31P {1H} NMR spec-
trum displays a sharp singlet for VIII at 20.2 ppm.
=
the released hydrogen does not hydrogenate the C C bond of
the vinyl reagent, or of ethylene; there is no evidence of ethane
or ethyl groups. If the reaction stoichiometry is changed to
1:1, accompanying the carbyne product (within 5 min) are
OsH(H2)Cl(C2H4)(PtBu2Me)2 , which is the product of
Os(H)3Cl(PtBu2Me)2 reacting with ethylene, and free C2H4
(vinyl chloride is absent); all vinyl chloroformate is gone at this
time. The production of ethylene from vinyl chloroformate in
this reaction depletes the Os(H)3ClL2 reagent of one H and
adds to it one Cl, in effect creating OsH2Cl2L2 . In fact, some
of this dichloride complex is observed within 5 min when the
reaction stoichiometry is 1 : 1; we have shown independently
that OsH2Cl2L2 reacts promptly with vinyl chloroformate to
give some OsHCl2(CCH3)L2 , but also an unappealing array
of other phosphine complexes. Because some vinyl chloro-
formate is converted to ethylene, and ethylene binds to
Os(H)3Cl(PtBu2Me)2 , thereby diminishing its reactivity, con-
sumption of vinyl chloroformate slows, and its complete con-
sumption depends on equilibrium 5 shifting to the right. In
general, using Os(H)3ClL2 as a substitute for the unknown
OsHClL2 demands consideration of the consequences of the
released H2 . The side reactions which begin with ethylene
At ꢁ60 ꢀC all Os(H)3Cl(PiPr3)2 (IV) has already reacted.
Formation of more hydrido-carbyne is observed when warm-
ing the sample to ꢁ40 ꢀC. At this temperature, the major spe-
cies are the two hydrido-carbyne isomers, as well as some
olefin adduct V. Minor quantities of the carbene VIII are still
present, whereas the alkyl species VII has already disappeared.
The carbene disappears at ꢁ30 ꢀC, but the olefin adduct does
not disappear until room temperature is achieved, at which
point only the two isomeric carbynes remain.
The hydrido carbyne species IX exists as two isomers in a
1
3 : 1 ratio, as seen by H and 31P{1H} NMR spectroscopies.
Observed are hydrides (ꢁ8.9 and ꢁ5.8 ppm, both triplets)
i
and OsC–CH3 (singlets at 0.7 and 0.4 ppm) and two Pr
methine hydrogens (3 : 1 ratio), as well as two 31P{1H}
NMR singlets (31.0 and 32.6 ppm). Two 13C{1H} carbyne car-
bon triplets (266.7 and 276.5 ppm) of intensity 3:1 have similar
JC–P values (11 Hz); two carbyne methyl carbons are also seen.
The O2CCF3 carbonyl carbons are seen at 161.7 and 159.9
ppm; each is a quartet (C/F coupling). Also observed are
New J. Chem., 2003, 27, 1451–1462
1457

View Article Online
show essentially free rotation around the Ru C bond16 on the
NMR timescale. However, in these dihalide compounds there
is no possibility of Z2 ligand coordination, which we assume
enhances a push-pull interaction between the carboxylic group
and the carbene, increasing back donation from Ru to the
empty p-orbital of the carbene carbon. As a consequence,
=
production thus deplete Os(H)3ClL2
=
the bond order of the Ru C bond increases along with its
rotational barrier. Compound X is the major product of the
reaction 2 h after mixing the reagents, but it slowly disappears
and disproportionates (eqn. 7) to the known dichloride
carbene.15 After 24 h at 20 ꢀC the
2RuClðO2CClÞðCHMeÞðPPh3Þ2 ! RuCl2ðCHMeÞðPPh3Þ2
ð5Þ
X
consumption according to the primary reaction (eqn. 4),
and lead to the slower reaction completion and inexact stoi-
chiometry.
The bifunctional nature of vinyl chloroformate permits a
second, competitive reaction, oxidative addition of the C–Cl
bond to some osmium species (e.g., eqn. 6).
þ RuðO2CClÞ2ðCHMeÞðPPh3Þ2 ð7Þ
reaction is not yet complete; small quantities of unreacted
vinylchloroformate and species X are still present. At this time,
a third product appears in the 1H NMR spectrum, displaying a
multiplet at 17.12 ppm thus confirming its ethylidene nature.
=
This species is assigned as Ru(O2CCl)2( CHMe)L2 , the corres-
ponding disproportionation product.
17
=
b) M Os. The analogous reaction with OsHCl(PPh3)3
was studied in benzene-d6 at 20 ꢀC. Upon combining the
reagents, strong effervescence is observed, indicative of CO2
release. Only one ethylidene fragment is observed in the 1H
NMR spectrum. The signal corresponding to the carbenic pro-
ton appears at 19.76 ppm, displaying the characteristic triplet
of quartets with 3JH–P ¼ 11 Hz and 3JH–H ¼ 5.5 Hz. The ethy-
lidene formulation is confirmed by the appearance of a doublet
at 1.70 ppm with 3JH–H ¼ 5.5 Hz, corresponding to the methyl
substituent on the carbene carbon. The 31P {1H} NMR spec-
trum is a sharp singlet at 8.1 ppm. We attribute this to the car-
ð6Þ
Reductive elimination of ethylene (a) or vinyl chloride (b) then
completes this side reaction.
=
bene OsCl2( CHMe)L2(XI). This carbene is metastable and
over a period of 2 days isomerizes to the carbyne OsHCl2-
Influence of phosphine identity: reactions of vinyl chloroformate
with MHCl(PPh3)3 (M ¼ Ru, Os)
a) M ¼ Ru. These reagents were employed in the hope
that one PPh3 will serve as a leaving group, and thus provide
a functional source of MHCl(PPh3)2 . This would diminish
the side reactions resulting from the H2 released by
Os(H)3ClL2 . Valuable information regarding the mechanism
of carbene delivery was indeed obtained by reaction of vinyl-
(CMe)L2 (XII).
14
At this time the isomerization is not yet complete. Carbyne
chloroformate with the monohydride RuHCl(PPh3)3 in 1:1
stoichiometry in benzene-d6 at 20 ꢀC. The first product
observed is a carbene, characterized by a downfield resonance
2
XII displays a triplet at ꢁ5.60 ppm with JH–P ¼ 15 Hz and
3
1
a triplet at ꢁ0.59 ppm with JH–P ¼ 2.4 Hz in the H NMR
in the 1H NMR at 16.7 ppm, due to the carbenic proton. This
spectrum, corresponding to the hydride and the methyl
carbyne ligands respectively. The 13C{1H} NMR spectrum
3
signal displays an apparent triplet of quartets with JH–P
¼
2
3
displays
a
triplet at 278.2 ppm with JC–P ¼ 12 Hz,
13.5 Hz and JH–H ¼ 6 Hz. The triplet indicates that there
are only two phosphines bound to the metal center, whereas
the quartet indicates that the other substituent in the carbene
carbon is a methyl group. This assignment is confirmed by
the presence of a doublet in the 1H NMR spectrum, corre-
sponding to the methyl substituent on the carbene carbon, at
2.12 ppm with the same coupling constant. Free triphenyl-
phosphine is observed by 31P {1H} NMR. Comparison of
corresponding to the carbyne carbon, and a singlet at 34.8
ppm corresponding to the methyl carbon, thus confirming the
hydrido-carbyne formulation. The hydrido-carbyne displays a
sharp singlet in the 31P{1H} NMR at 4.6 ppm.
c) Reaction of OsHCl(PPh3)3 with vinyl trifluoroacetate. In
an attempt to strengthen the C–X bond (thus avoiding a subse-
quent chloroformate decarboxylation process) in the carboxy-
late substituent XCO2 , we chose vinyl trifluoroacetate as the
vinyl reagent. The products first observed (2.5 h after com-
bining these reagents in C6D6) are the dichloride ethylidene
XI, the mixed chloride-carboxylate carbene OsCl(O2CCF3)-
=
these values with those in the literature for RuCl2L2( CHMe)
(L ¼ PPh3)15 shows that the product is not this one. We
propose this carbene to be the mixed halide-carboxylate six
=
2
coordinate compound RuCl(Z –O2CCl)( CHMe)L2(X). The
corresponding 31P{1H} NMR displays an AB pattern with
2
=
( CHMe)L2 (XIII), and its isomeric hydrido-carbyne as a
0
JP–P ¼ 430 Hz. The AB pattern can be due to either mutually
cis phosphines or orientation of the carbene substituents along
the P–Ru–P vector. A cis disposition of the phosphines is
eliminated by the large value of the coupling constant. There-
fore, the inequivalent phosphines must be due to the configura-
tion of the carbene ligand. This is particularly intriguing since
most of the related 16 electron dihalide ethylidene compounds
result of a–H migration OsHCl(O2CCF3)(CMe)L2 (XIV), all
of them present in very small quantities. The carbene XIII is
identified in the H NMR spectrum by a downfield multiplet
1
3
at 21.1 ppm and a doublet at 1.65 ppm with JH–H ¼ 6.6 Hz,
whereas its isomeric hydrido-carbyne XIVa displays a triplet at
2
ꢁ4.24 ppm with JH–P ¼ 15 Hz and a triplet at ꢁ0.13 ppm
1458
New J. Chem., 2003, 27, 1451–1462

View Article Online
4
with JH–P ¼ 2.4 Hz, corresponding to the hydride ligand and
the methyl substituent on the carbyne carbon
respectively. The 31P {1H} NMR signal for the hydrido-carbyne
XIVa is a singlet at 18.1 ppm.
has
E on Ru (XVII).
Finally, a CO ligand in place of Clꢁ decreases the reducing
power of osmium to the point where olefin binding, but neither
carbene nor carbyne ligand formation occurs.
These studies are the first to observe, at a detectable
rate, the transformation from osmium carbene to hydrido
osmium carbyne by a formal a-H migration (for Ru, this
reaction is not observed because it is endergonic). What is
remarkable is that this reaction is quite fast (minutes at
ꢁ30 ꢀC) for L ¼ PiPr3 yet slow (hours at 25 ꢀC) for
The dichloro-carbene species XI spontaneously transforms
to its isomeric hydrido-carbyne XII analogously to the
behavior observed in the reaction between OsHCl(PPh3)3
and vinylchloroformate. The carbyne species XIVa then
transforms partially to isomeric XIVb. The isomer XIVb is
characterized by a triplet in the hydride region at ꢁ7.45 ppm
L ¼ PPh3 .
While a
2
with a JH–P ¼ 16 Hz and a singlet for the carbyne methyl at
ꢁ0.28 ppm.
unimolecular 1,2-H migration from C to Os seems an
‘‘obvious’’ mechanism, DFT calculations showed this
mechanism to be precluded by a large activation energy (27.2
kcal molꢁ1) in the case where X ¼ Y ¼ Cl.17
The influence of CO on Os. The influence of metal p-basicity
on the C/O bond cleavage reaction is also shown by employ-
+
ing OsH(CO)(PtBu2Me)2
as the unsaturated monohy-
dride.18 The cationic charge and the carbonyl ligand (cf.
Clꢁ in OsHClL2) greatly diminish the reducing ability in
spite of the general reducing character of a 5d metal. This
cation was synthesized by abstraction of triflate from
OsH(OTf)(CO)L2 using NaBArF (ArF ¼ 3,5(CF3)2C6H3) in
CH2Cl2 . This cation reacts with equimolar vinyl trifluoroace-
tate within 5 min in CD2Cl2 at 25 ꢀC to give an olefin adduct
The experimental observations, together with the energies
calculated for observed and postulated intermediates have
defined the general features by which the O–C bond of vinyl
esters is cleaved easily by (electron-rich) Ru and Os com-
plexes devoid of p-acid ligands. As shown in Fig. 1, all neces-
sary intermediates lie within easy energetic reach. Because of
the energetic proximity of both the alkyl (2Ru) and the car-
bene (3Ru) with the ester oxygen bonded to Ru, each is a
viable intermediate for intramolecular O2CMe transfer from
carbon to metal. This contrasts to vinyl ethers, where obser-
vations are consistent with an acid-catalyzed mechanism. The
greater ‘‘reach’’ of an ester than an ether functionality (five-
vs. three-membered ring) accounts for this difference.
with inequivalent 31P nuclei, whose JPP ¼ 112 Hz indicates
0
extreme P–Os–P bending, to increase p donation to the ole-
fin.19–21 The hydride resonance is a doublet of doublets
(JPH ¼ 28 and 33 Hz), consistent with inequivalent phos-
phines, and its chemical shift, ꢁ1.9 ppm, suggests a strong
ligand trans to itself: olefin (see XV).
This
chemical shift contrasts to the values ꢁ23.7 ppm for the
water adduct OsH(OTf)(H2O)(CO)(PtBu2Me)2 and ꢁ27.5
Experimental
+
ppm for18 OsH(CD2Cl2)(CO)(PtBu2Me)2
. A
vinylic 1H
General procedure
NMR signal at 7.09 ppm shows coupling to phosphines, con-
firming olefin/Os binding. Vinyl acetate gives an analogous
adduct under the same conditions. Evidence that the keto
oxygen also binds to Os is the large (136 cmꢁ1) reduction
All manipulations were performed using standard Schlenk
techniques or in an argon filled glovebox. Solvents were dried,
degassed or distilled under argon from Na, Na/benzophenone,
for coordinated vinyl acetate (to 1619 cmꢁ1) from
˚
of nC
P2O5 , CaH2 , and/or 4 A molecular sieves and stored in air-
=
O
its value (1755 cmꢁ1) for free vinyl acetate.
tight solvent bulbs with Teflon closures. All NMR solvents
were dried, vacuum-transferred, and stored in a glovebox.
Complexes OsH3ClL2 (L ¼ PtBu2Me, PiPr3) were synthesized
according to published procedures.13 The synthesis of
OsH3Cl(PiPr3)2 from OsH2Cl2(PiPr3)2 and NEt3 can leave vari-
able small amounts of [HNEt3]Cl as an impurity which is dif-
ficult to remove by recrystallization. When present, this can
lead to the production of OsH2Cl2L2 as a by product of the
dehydrogenation reactions reported here; this product is thus
not derived from the ester reaction, but from the available
‘‘HCl’’. [RuHCl(PiPr3)2]2 ,1,22 was prepared according to pub-
lished procedures. Commercially available vinyl esters were
used as received after drying and degassing when applicable.
Chemical shifts are referenced to residual protio solvent peaks
(1H), external H3PO4 (31P), external CFCl3 (19F), or natural
abundance 13C peaks of the solvent (13C). NMR spectra were
obtained on a Varian Gemini 2000 (300 MHz 1H; 121.4
MHz 31P, 75 MHz 13C, 282 MHz 19F), a Varian Unity Inova
The vinyl trifluoroacetate adduct is unchanged after 2 h at
80 ꢀC in benzene. This less-reducing osmium thus shows no
tendency for C–O bond cleavage.
Discussion
The present work reveals the previously-reported tendency of
ruthenium to prefer the carbene form, while osmium prefers
the isomeric carbyne, formed by transfer a carbene substituent
to the metal. Osmium thus ‘‘benefits’’ from higher coordina-
tion number and oxidation state, as well as from 18-valence
electrons.
This report also shows the relative stability of isomeric
ruthenium carbenes, XVI vs. XVII. For E ¼ alkoxide, the final
product has E on carbon (XVI). As E becomes less electron
donating (e.g., E ¼ OTs or O2CR), the thermodynamic isomer
New J. Chem., 2003, 27, 1451–1462
1459

View Article Online
instrument (400 MHz 1H; 162 MHz 31P), or a Bruker AM
and filtering in air, the ether solution was added to a separating
funnel, washed with dilute aqueous K2CO3 (6 ꢄ 50 mL), and
dried over MgSO4 . After removal of the solvent in vacuo,
the resulting brown liquid was vacuum distilled, collecting
the colorless fraction boiling at 96–97 ꢀC (0.3 torr). Yield:
1
spectrometer (500 MHz H, 125.6 MHz 13C). ‘‘N’’ is the spa-
cing (Hz) of the sharp outer liner in a virtual triplet (‘‘vt’’).
i
RuCl(O2CCH3)(P Pr3)2( CHMe)
=
1
approximately 3 g (26%). H NMR (25 ꢀC, 400 MHz, C6D6):
15.0 mg (0.016 mmol) [RuHCl(PiPr3)2]2 was dissolved in 0.5
mL C6D6 and added to an NMR tube with a septum cap.
Via syringe, 3.0 mL (0.033 mmol) vinyl acetate was added
3
d 1.77 (s, 3H, OSO2C6H4CH3), d 4.08 (dd, JH–H ¼ 6 Hz,
2
3
=
JH–H ¼ 2 Hz, 1H, CH2 CH(OTs)), d 4.52 (dd, JH–H ¼ 11
2
=
Hz, JH–H ¼ 2 Hz, 1H, CH2 CH(OTs)),
d 6.47 (dd,
1
and the sample mixed. H and 31P NMR spectra taken imme-
3
3
=
JH–H ¼ 11 Hz, JH–H ¼ 6 Hz, 1H, CH2 CH(OTs)), d 6.61
3
3
diately reveal signals of the title compound, in addition to
=
(d, JH–H ¼ 8 Hz, 2H, OSO2C6H4CH3), d 7.63 (d, JH–H
8 Hz, 2H, OSO2C6H4CH3).
¼
i
those of RuCl2(P Pr3)2( CHMe). Selected spectroscopic data
4
1
for the title compound: H NMR (25 ꢀC, 300 MHz, C6D6): d
3
1.76 (s, 3H, Ru(O2CCH3)), d 2.48 (d, JH–H ¼ 7 Hz, 3H,
i
RuCl(OSO2C6H4CH3)(P Pr3)2( CHMe)
=
3
=
=
Ru CH(CH3)), d 19.67 (q, JH–H ¼ 7 Hz, 1H, Ru CHMe).
31P NMR (25 ꢀC, 121 MHz, C6D6): d 37.2 (s).
15.0 mg (0.016 mmol) [RuHCl(PiPr3)2]2 was dissolved in 0.5
mL C7D8 and added to an NMR tube equipped with a Teflon
seal. Via syringe, 5.4 mL (0.032 mmol) vinyl tosylate was added
so that the reagents did not mix and the tube was sealed. The
sample was cooled in a dry ice–acetone bath, shaken vigor-
ously, and then placed in a pre-cooled NMR probe (ꢁ60 ꢀC).
1H and 31P NMR spectra taken at this temperature reveal
quantitative conversion to the title compound. Selected spec-
i
RuCl(O2CCF3)(P Pr3)2( CHMe)
=
15.0 mg (0.016 mmol) [RuHCl(PiPr3)2]2 was dissolved in 0.5
mL C6D6 and added to an NMR tube with a septum cap.
Via syringe, 3.7 mL (0.032 mmol) vinyl trifluoroacetate was
added and the tube shaken. 1H, 31P, and 19F NMR spectra
taken immediately reveal predominately signals of the title
1
troscopic data follows: H NMR (ꢁ60 ꢀC, 400 MHz, C7D8):
i
4
=
compound, in addition to those of RuCl2(P Pr3)2( CHMe).
3
d 1.97 (s, 3H, Ru(OSO2C6H4CH3)), d 2.59 (d, JH–H ¼ 4 Hz,
1
Selected spectroscopic data follows: H NMR (25 ꢀC, 400
3
=
3H, Ru CH(CH3)), d 6.78 (d, JH–H ¼ 8 Hz, 2H, Ru(O-
3
=
MHz, C6D6): d 2.49 (d, JH–H ¼ 5 Hz, 3H, Ru CH(CH3)), d
3
SO2C6H4CH3)),
d
7.87 (d, JH–H ¼ 8 Hz, 2H, Ru(O-
3
31
19.88 (q, JH–H ¼ 5 Hz, 1H, Ru CHMe). P NMR (25 ꢀC,
=
19
3
=
SO2C6H4CH3)), d 20.08 (q, JH–H ¼ 4 Hz, 1H, Ru CHMe).
162 MHz, C6D6): d 44.3 (s). F NMR (25 ꢀC, 121 MHz,
31P NMR (ꢁ60 ꢀC, 162 MHz, C7D8): d 51.3 (s).
C6D6): d ꢁ76.7 (s).
Synthesis of OsHCl(Z1-O2CCF3)(CCH3)(PiPr3)2 , IX
i
RuC(O2CPh)(P Pr3)2( CHMe)
=
In an NMR tube, OsH3Cl(PiPr3)2 (0.0100 g, 0.018 mmol) was
dissolved in 0.8 ml of toluene-d8 and vinyl trifluoroacetate
(8.48 mL, 0.072 mmol) was added to the solution. The reaction
proceeds with strong effervescence and is complete in 10 min.
The volatiles were removed under vacuo and the yellowish resi-
due recovered. The solid residue consists of a mixture of two
carbyne products in a 3:1 intensity ratio. Major isomer will
be noted as isomer A, whereas the minor isomer will be noted
as B. 1H NMR (300 MHz, C6D6 , 20 ꢀC): ꢁ8.90 (t, J(H–P) ¼ 16
Hz, Os–H, A), ꢁ5.88 (t, J(HP) ¼ 15 Hz, Os–H, B), 0.39 (s,
15.0 mg (0.016 mmol) [RuHCl(PiPr3)2]2 was dissolved in 0.5
mL C6D6 and added to an NMR tube with a septum cap.
Via syringe, 4.4 mL (0.032 mmol) vinyl benzoate was added
1
and the sample mixed. H and 31P NMR spectra taken imme-
diately reveal signals of the title compound, in addition to
=
i
those of RuCl2(P Pr3)2( CHMe). Selected spectroscopic data
4
1
for the title compound: H NMR (25 ꢀC, 300 MHz, C6D6): d
2.61 (d, 3JH–H ¼ 6 Hz, 3H, Ru CH(CH3); overlaps with those
=
i
3
=
of RuCl2(P Pr3)2( CHMe)), d 7.01 (t, JH–H ¼ 5 Hz, 1H,
3
Ru(O2CC6H5)), d 7.12 (apparent t, JH–H ¼ 5 Hz, 2H,
=
=
=
Os C–CH , B), 0.72 (s, Os C–CH , A), 1.12 (dvt, N ¼ 13.5
3
=
3
3
Ru(O2CC6H5)), d 8.25 (d, JH–H ¼ 5 Hz, 2H, Ru(O2CC6H5)),
Hz, Os–P(CH(CH3)2), A + B overlapped), 1.22 (dvt,
N ¼ 13.5 Hz, Os–P(CH(CH3)2), A + B overlapped), 2.25 (m,
Os–P(CH(CH3)2), B), 2.44 (m, Os–P(CH(CH3)2), A).
31P{1H} NMR (121.4 MHz, C6D6 , 20 ꢀC): 31.0 (A), 32.6
3
31
d 19.75 (q, JH–H ¼ 7 Hz, 1H, Ru CHMe). P NMR (25 ꢀC,
=
121 MHz, C6D6): d 51.3 (s).
Reaction of [RuHCl(PiPr3)2]2 with CH2 CH(O2CCl)–vinyl
=
chloroformate
13
(B). C{1H} NMR (125.6 MHz, C6D6 , 25 ꢀC): 19.1 (s, Os–
P(CH(CH3)2), A), 19.2 (s, Os–P(CH(CH)3)2), B), 19.6 (s, Os–
P(CH(CH3)2), A), 19.7 (s, Os–P(CH(CH3)2), B), 24.3 (t,
J(C–P) ¼ 12 Hz, Os–P(CH(CH3)2), B), 25.3 (t, J(C–P) ¼ 13 Hz,
15.0 mg (0.016 mmol) [RuHCl(PiPr3)2]2 was dissolved in 0.5
mL C6D6 and added to an NMR tube with a septum cap.
Via syringe, 2.8 mL (0.033 mmol) vinyl chloroformate was
added and the sample mixed. Gas evolution was observed
=
Os–P(CH(CH ) ), A), 38.1 (s, Os C–CH , A), 39.4 (s,
=
3 2
3
=
Os C–CH , B), 116.1 (q, J
3
¼ 291 Hz, Os–OCO(CF3),
=
A + B overlapped), 159.9 (q, J(C–F) ¼ 45 Hz, Os–OCO(CF3),
(C–F)
1
immediately and the solution turned deep purple. H and 31P
B), 161.7 (q, J(C–F) ¼ 45 Hz, Os–OCO(CF3), A), 266.7 (t,
NMR spectra taken immediately show quantitative conversion
=
i
to RuCl2(P Pr3)2( CHMe).
4
=
J(C–P) ¼ 11 Hz, Os C–CH , A), 276.5 (t, J
=
¼ 11 Hz,
3
(C–P)
=
Os C–CH , B).
=
3
=
CH2 CH(OSO2C6H4CH3)–vinyl tosylate
Reaction of OsH3Cl(PiPr3)2 and vinyl trifluoroacetate at
low temperatures
Vinyl tosylate was prepared with a slightly modified procedure
from the original literature reference.23 Anhydrous p-tolylsul-
fonic acid was prepared by heating the monohydrate in vacuo
(60 ꢀC, 0.01 torr) for 12 hours. 10.0 g (58 mmol) anhydrous
p-tolylsulfonic acid and 0.67 g (3.1 mmol) yellow HgO were
charged in a glass pressure reaction vessel. 20 mL ether was
added, the vessel sealed, cooled to ꢁ50 ꢀC, and the headspace
gasses evacuated. The flask was filled to 80 psi with acetylene
and re-pressurized as needed. After initial gas uptake had
ceased (2 hours), the mixture was heated at 50 ꢀC for 1 hour
(vessel pressure increased to 120 psi -caution!). After cooling
Vinyl trifluoroacetate (2.12 mL, 0.018 mmol) was dissolved in
0.8 ml of toluene-d8. The solution was vacuum transferred into
an NMR tube charged with OsH3Cl(PiPr3)2 (0.0100 g, 0.018
mmol) and frozen at 78 K. The NMR tube was thawed and
shaken for one second prior inserting it into a precooled
NMR probe at ꢁ80 ꢀC. The temperature of the probe was
raised to room temperature in 10 ꢀC intervals and the solution
was allowed to react for 5 min prior to acquiring the 1H NMR
and 31P {1H} NMR spectra. Only diagnostic data is provided
1460
New J. Chem., 2003, 27, 1451–1462

View Article Online
for the identified compounds. Data for the compound
OsH3Cl(H2C CH(O2CCF3))(P Pr3)2 , V: 1H NMR (300
(d, J(H–H) ¼ 6 Hz, Ru ¼ CH(CH3), 3H). 31P {1H} NMR
i
(121.4 MHz, C6D6 , 20 ꢀC): 37.7, 48.1 (AB, J(P–P ) ¼ 430 Hz)
=
0
MHz, C7D8 , ꢁ70 ꢀC): ꢁ4.05 (t, J(H–P) ¼ 20 Hz, Os–H, H),
Synthesis of OsHCl2(CMe)(PtBu2Me)2
=
ꢁ14.78 (br Os–H, 2H), 2.75, 2.87 (br AB, Os(H2C CH-
(O2CCF3), 2H), 3.97 (br AB, Os(H2C CH(O2CCF3), H). 31P
=
In an NMR tube OsH3Cl(PtBu2Me)2 (0.0100 g, 0.018 mmol)
was dissolved in 0.8 mL of benzene-d6 and vinyl chloroformate
(3.25 mL, 0.036 mmol) was added to the solution. The reaction
was allowed to proceed for 15 min at room temperature. At
this point the hydrido-carbyne is the major species. Some ethy-
lene and vinyl chloride are observed in the solution. The vola-
tiles were removed under vacuo and a brown solid is obtained.
{1H} NMR (121.4 MHz, C7D8 , ꢁ70 ꢀC): 37.8. Data for the
compound OsCl(CH(CH3)(O2CCF3))(PiPr3)2 , VII: 1H NMR
(300 MHz, C7D8 , ꢁ70 ꢀC): 1.60 (d, J(H–H) ¼ 6.6 Hz, Os–
CH(CH3)(O2CCF3), 3H), 0.77 (q, J(H–H) ¼ 6.6 Hz, Os–
CH(CH3)(O2CCF3), H). 31P {1H} NMR (121.4 MHz, C7D8 ,
ꢁ70 ꢀC): 28.9. Data for the compound OsCl(O2-
i
CCF3)( CH(CH3))(P Pr3)2 , VIII: 1H NMR (300 MHz,
=
1
ꢀ
The solid is redissolved in benzene-d6 . H NMR (300 MHz,
C6D6 , 20 ꢀC): ꢁ8.96 (t, J(H–P) ¼ 15 Hz, Os–H, H), 0.78 (s,
=
=
C7D8 , ꢁ60 C): 18.5 (br Os CH(CH3), H), 1.68 (br, Os CH-
(CH3), 3H). P {1H} NMR (121.4 MHz, C7D8 , ꢁ60 ꢀC):
31
=
Os C–CH , 3H), 1.26 (vt, J
¼ 6.3 Hz, Os–(PCH3-
=
3
(H–P)
20.2.
(C(CH3)3)), 1.30 (vt, J(H–P) ¼ 5.5 Hz, Os–(PCH3(C(CH3)3)),
1.86 (vt, J(H–P) ¼ 11.1 Hz, Os–(PCH3(C(CH3)3)). 31P {1H}
NMR (121.4 MHz, C6D6 , 20 ꢀC): 28.6. 13C {1H} NMR (75
MHz, C6D6 , 20 ꢀC): 3.7 (t, J(C–P) ¼ 14 Hz, Os–(PCH3-
(C(CH3)3)), 29.5 (s, Os–(PCH3(C(CH3)3)), 31.1 (s, Os–(PCH3-
(C(CH3)3)), 36.7 (t, J(C–P) ¼ 11 Hz, Os–(PCH3(C(CH3)3)), 38.3
Synthesis of OsHCl2(CMe)(PPh3)2 , XII
In an NMR tube OsHCl(PPh3)3 (0.0100 g, 0.01 mmol) was dis-
solved in 0.8 mL of benzene-d6 and vinyl chloroformate (0.9
mL, 0.01 mmol) was added to the solution. The reaction was
allowed to proceed for 2 days at room temperature. At this
point the hydrido-carbyne is the major species. The volatiles
were removed under vacuo and a brown solid is obtained.
The solid is redissolved in benzene-d6 . Only diagnostic data
is provided for the carbene and the title product. Data for
=
(t, J(C–P) ¼ 11 Hz, Os–(PCH (C(CH ) )), 40.8 (s, Os C–CH ),
=
3
3 3
3
=
271.2 (t, J(H–P) ¼ 13 Hz, Os C).
=
OsH(OTf)(CO)(PtBu2Me)2
Me3SiOTf (70 ml, 0.36 mmol) was added dropwise to benzene
(5 mL) solution of OsHF(CO)L2 (200 mg, 0.36 mmol;
L ¼ PtBu2Me). The solution color changed to dark red after
stirring for 10 min at 20 ꢀC. after evaporation of volatiles,
the residue was dissolved in 2 mL toluene and cooled to
ꢁ40 ꢀC for 24 h to afford orange crystals, which were filtered
at ꢁ78 ꢀC, washed with pentane and dried. Yield 100 mg
(40%). Anal. Calcd for C20H43F3O4OsP2S: C, 34.89; H, 6.29.
Found: C, 34.93; HHHHh, 5.99. 1H NMR (360 MHz,
20 ꢀC): 1.60 (br s, 6H, PCH3), 1.15 (vt, N ¼ 13.3 Hz, 18H,
PC(CH3)3), 1.04 (vt, N ¼ 12.6 Hz, PC(CH3)3), ꢁ35.6 (t,
JPH ¼ 13.7 Hz, 1H, Os–H). 31P{1H} NMR (146 MHz): 49.5
(s). IR (C6D6 , cmꢁ1): 1908 (n(CO))
1
the carbene, XI: H NMR (300 MHz, C6D6 , 20 ꢀC): 19.76
=
(tq, J(H–P) ¼ 11 Hz, J(H–H) ¼ 5 Hz, Os CH(CH3), H), 1.70
(d, J(H–H) ¼ 5.5 Hz, Os CH(CH3), 3H). 31P {1H} NMR
=
(121.4 MHz, C6D6 , 20 ꢀC): 8.1. Data for the title compound:
1H NMR (300 MHz, C6D6 , 20 ꢀC): ꢁ5.60 (t, J(H–P) ¼ 15 Hz,
Os–H, H), ꢁ0.59 (t, J(H–P) ¼ 2.4 Hz, Os C–CH , 3H). 31P
=
=
3
{1H} NMR (121.4 MHz, C6D6 , 20 ꢀC): 4.6. 13C {1H}
ꢀ
=
=
NMR (100 MHz, C D , 20 C): 34.8 (s, Os C–CH ), 278.2
6
6
3
=
(t, J(H–P) ¼ 12 Hz, Os C).
=
Synthesis of OsHCl(O2CCF3)(CMe)(PPh3)2 isomers, XIV
In an NMR tube OsHCl(PPh3)3 (0.0100 g, 0.01 mmol) was dis-
solved in 0.8 mL of benzene-d6 and vinyl trifluoroacetate (1.1
mL, 0.01 mmol) was added to the solution. The reaction was
allowed to proceed for 2 days at room temperature. At this
point the two hydrido-carbyne isomers are the major species.
The volatiles were removed under vacuo and a brown solid is
obtained. The solid is redissolved in benzene-d6 . The major
isomer will be noted as A, whereas the minor isomer will be
noted as B. Only diagnostic data is provided for the carbene
intermediate and the title compound. Data for the carbene
OsH(OTf)(OH2)(CO)(PtBu2Me)2
Water (ca. 0.3 mL) was added to OsH(OTf)(CO)L2 (10 mg,
0.018 mmol) in CD2Cl2 (0.5 mL) to yield a light yellow solu-
1
tion. H NMR (300 MHz, 20 ꢀC): 2.87 (br s, coordinated
H2O), 1.37 (vt, N ¼ 12.6 Hz, 24 H, PC(CH3)3 and overlapping
with PCH3), 1.30 (vt, N ¼ 12.9, 18 H, PC(CH3)3), ꢁ23.7 (t,
J ¼ 14.4, Os–H). 31P{1H} NMR: 39.7 (s). IR (CD2Cl2): 1896
(n(CO)).
1
intermediate, XIII: H NMR (300 MHz, C6D6 , 20 ꢀC): 21.1
=
(m, Os CH(CH3), H), 1.65 (d, J(H–H) ¼ 6.6 Hz,
[OsH(Z2–CD2Cl2)(CO)(PtBu2Me)2]BAr04
Os CH(CH3), 3H). 31P {1H} NMR (121.4 MHz, C6D6 ,
=
20 ꢀC): 5.5. Data for the carbyne isomers: 1H NMR (300
MHz, C6D6 , 20 ꢀC): ꢁ4.24 (t, J(H–P) ¼ 15 Hz, Os–H, H,A),
ꢁ7.45 (t, J(H–P) ¼ 16 Hz, Os–H, H, B), ꢁ0.13 (t, J(H–P) ¼ 2.4
OsH(OTf)(CO)L2 (10 mg, 0.014 mmol) and NaBAr04 (12.9 mg)
was mixed in CD2Cl2 (0.5 ml). Thirty minutes after the mixing,
NMR spectra reveals clean formation of a complex. 1H NMR
(300 MHz, 20 ꢀC): 1.61 (br, 6H, PCH3), 1.32 (vt, N ¼ 13.5 Hz,
18H, PC(CH3)3), 1.18 (vt, N ¼ 13.2 Hz, 18H, PC(CH3)3),
ꢁ27.5 (br, 1 H, w1/2 ¼ 486 Hz, Os–H). 19F NMR (282
MHz): ꢁ61.9 (s, BAr04). 31P{1H} NMR: 42.9 (s)
Hz, Os C–CH , 3H, A), ꢁ0.28 (s, Os C–CH , 3H, B). 31P
=
=
=
=
3
3
{1H} NMR (121.4 MHz, C6D6 , 20 ꢀC): 18.1 (A), 11.7(B).
=
Synthesis of RuCl2( CHMe)(PPh3)2
In an NMR tube, RuHCl(PPh3)3 (0.0100 g, 0.011 mmol) was
dissolved in 0.8 ml of benzene-d6 and vinyl chloroformate (1
ml, 0.011 mmol) was added to the solution. The reaction was
allowed to proceed for 1 day at room temperature. The vola-
tiles were removed under vacuo and a brown solid is obtained.
The solid is redissolved in benzene-d6 . Its spectroscopic data
(1H NMR and 31P {1H} NMR) compares to that reported in
the literature15 within experimental error. only diagnostic data
t
[OsH(Z3–CH2 CHOC(O)CF3)(CO)(P Bu2Me)2]BAr 4 , XV
0
=
OsH(OTf)(CO)L2 (10 mg, 0.018 mmol) and NaBAr04 (12.9 mg,
0.018 mmol) was dissolved in CD2Cl2 (0.5 mL). To the solu-
tion, CH2 CHOC(O)CF3 (1.5 mL) was added. The solution
color changed to light yellow immediately. NMR spectral ana-
=
3
=
lysis show exclusive formation of OsH(Z –CH2 CHOC(O)-
CF3)(CO)L2]BAr04
.
1H NMR (300 MHz, 20 ꢀC): 7.09 (dtd,
=
=
for the intermediate RuCl(O2CCl)( CHMe)(PPh3)2 is
JHH ¼ 13.5 Hz, JPH ¼ 5.1 Hz, JHH ¼ 2.1 Hz, 1H, CH–O),
1
provided: H NMR (300 MHz, C6D6 , 20 ꢀC): 16.7 (tq,
2.58 (m, 1H, CH2 ), 2.51 (m, 1H, CH2 ), 1.84 (d, JPH ¼ 8.4
Hz, 3H, PCH3), 1.49 (d, JPH ¼ 10 Hz, 3H, PCH3), 1.45 (d,
= =
J(H–P) ¼ 13.5 Hz, J(H–H) ¼ 6 Hz, Ru ¼ CH(CH3), H), 2.12
New J. Chem., 2003, 27, 1451–1462
1461

View Article Online
5
(a) S. Komiya, R. S. Srivastava, A. Yamamoto and T.
Yamamoto, Organometallics, 1985, 4, 1504; (b) S. Komiya, J.
Suzuki, K. Miki and N. Kasai, Chem. Lett., 1987, 1287; (c) Y.
Hayashi, T. Yamamoto, A. Yamamoto, S. Komiya and Y. Kushi,
J. Am. Chem. Soc., 1986, 108, 385; (d ) A. Yamamoto, Adv. Orga-
nometal. Chem., 1992, 34, 111.
JPH ¼ 14.4 Hz, 9H, PC(CH3)3), 1.34 (d, JPH ¼ 14.1 Hz, 18H,
PC(CH3)3), 1.23 (d, JPH ¼ 13.5 Hz, 9H, PC(CH3)3), ꢁ1.92
(dd, 1 H, JPP ¼ 27.6, 33 Hz, Os–H). 31P{1H} NMR (121
MHz): 27.6 (d, JPP ¼ 112 Hz, Os–P), 25.6 (d, JPP ¼ 112 Hz,
Os–P).
6
7
8
9
M. Olivan, E. Clot, O. Eisenstein and K. G. Caulton, Organo-
metallics, 1998, 17, 897.
S. Komiya and T. Shindo, J. Chem. Soc., Chem. Commun., 1984,
1672.
W. Buchowicz, J. C. Mol, M. Lutz and A. L. Spek, J. Organo-
metal. Chem., 1999, 588, 205.
3
t
0
=
[OsH(Z –CH2 CHOC(O)CH3)(CO)(P Bu2Me)2]BAr 4
=
Same procedure as above was followed except CH2 CHO-
1
C(O)CH3 was used. H NMR (300 MHz, 20 ꢀC): 6.73 (ddt,
JPH ¼ 5.4 Hz, JHH ¼ 6 Hz, JHH ¼ 11.7 Hz, 1H, CHO), 2.48
(m, CH2), 2.36 (t, J ¼ 5.7 Hz, 1H, CH2), 2.17 (s, 3H,
CH3CO2), 1.77 (vt, 3H, N ¼ 6.9 Hz, PCH3), 1.48 (vt, 3H,
N ¼ 6.9 Hz, PCH3), 1.41 (vt, 9H, N ¼ 14.4 Hz, PC(CH3)3),
1.34 (vt, 9H, N ¼ 14.4 Hz, PC(CH3)3), 1.24 (vt, 9H,
N ¼ 14.1 Hz, PC(CH3)3), 1.26 (vt, 9H, N ¼ 14.1 Hz,
PC(CH3)3), -2.39 (t, 1 H, JPH ¼ 30 Hz, Os–H). 31P{1H}
NMR (121 MHz): 25.9 (s). IR (CD2Cl2): 1959 (n(CO), 1619
J. Wolf, W. Stuer, C. Grunwald, H. Werner, P. Schwab and M.
¨
Schulz, Angew. Chem., Int. Ed. Engl., 1998, 37, 1124.
¨
´
10 H. Gerard, E. Clot, C. Giessner-Prette, K. G. Caulton, E. R.
Davidson and O. Eisenstein, Organometallics, 2000, 19, 2291.
11 P. Gonzalez-Herrero, B. Weberndo¨rfer, K. Ilg, J. Wolf and H.
Werner, Angew. Chem., Int. Ed. Engl., 2000, 39, 3266.
12 Y. Musashi and S. Sakaki, J. Am. Chem. Soc., 2000, 122, 3867.
13 R. Kuhlman, E. Clot, C. Leforestier, W. E. Streib, O. Eisenstein
and K. G. Caulton, J. Am. Chem. Society, 1997, 119, 10 153.
14 P. S. Hallman, B. R. McGarvey and G. Wilkinson, J. Chem. Soc.
(A), 1968, 3143.
´
=
(n(C O)).
15 P. Schwab, R. H. Grubbs and J. W. Ziller, J. Am. Chem. Soc.,
1996, 118, 100.
16 T. M. Trnka and R. H. Grubbs, Acc. Chem. Res., 2001, 34, 18.
17 (a) G. Ferrando and K. G. Caulton, Inorg. Chem., 1999, 38, 4168;
(b) G. J. Spivak, J. N. Coalter III, M. Olivan, O. Eisenstein and
K. G. Caulton, Organometallics, 1998, 17, 999.
18 D. Compare: Huang, J. C. Bollinger, W. E. Streib, K. Folting,
V. Young Jr., O. Eisenstein and K. G. Caulton, Organometallics,
2000, 19, 2281.
19 A. J. Edwards, S. Elipe, M. A. Esteruelas, F. J. Lahoz, L. A. Oro
and C. Valero, Organometallics, 1997, 16, 3828.
20 K. B. Renkema, J. C. Huffman and K. G. Caulton, Polyhedron,
1999, 18, 2575.
21 D. V. Yandulov, D. Huang, J. C. Huffman and K. G. Caulton,
Inorg. Chem., 2000, 39, 1919.
22 J. N. Coalter, W. E. Streib and K. G. Caulton, Inorg. Chem.,
2000, 39, 3749.
Computational details
The calculations were carried out using the Gaussian 98 set of
programs24 within the framework of DFT at the B3PW91
level.25,26 LANL2DQ effective core potentials (quasi-relativis-
tic for the metal centers) were used to replace the 28 innermost
electrons of Ru27 and the ten core electrons of Cl, and P.28 The
associated double-basis set was used27,28 and was augmented
by a d polarization function for Cl, and P.29 The other atoms
were represented by a 6-31 (d,p) basis set (5d).30 Full geometry
optimization was performed with no symmetry restriction, and
the nature of the minima was assigned by analytical frequency
calculations.
23 J. Sauer and J. Wilson, J Am. Chem. Soc., 1955, 77, 3793.
24 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery Jr.,
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D.
Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V.
Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C.
Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G.
Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle and J. A.
Pople, Gaussian 98 Revision A.7, Gaussian, Inc., Pittsburgh,
PA, 1998.
Acknowledgements
This work was supported by the U.S. National Science Foun-
dation, the French CNRS, and the University of Montpellier
2. The authors are grateful to Indiana University Computing
Center for a generous donation of computational time.
References
1
(a) J. N. Coalter, J. C. Bollinger, J. C. Huffman, U. Werner-
Zwanziger, K. G. Caulton, E. R. Davidson, H. Gerard, E. Clot
and O. Eisenstein, New J. Chem., 2000, 24, 9; (b) G. Ferrando,
´
H. Gerard, G. J. Spivak, J. N. Coalter, III, J. C. Huffman, O.
25 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
Eisenstein and K. G. Caulton, Inorg. Chem., 2001, 40, 6610.
Bis(trifluoroacetato) ruthenium carbenes bearing triaryl phos-
phines have been reported Z. Wu, S. T. Nguyen, R. H. Grubbs
and J. W. Ziller, J. Am. Chem. Soc., 1995, 117, 5503.
J. U. Notheis, R. H. Heyn and K. G. Caulton, Inorg. Chim. Acta,
1995, 229, 187.
26 J. P. Perdew and Y. Wang, Phys. Rev. B, 1992, 45, 13 244.
27 P. G. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299.
28 W. R. Wadt and P. J. Hay, J. Chem. Phys., 1985, 82, 284.
29 A. H. Ho¨llwarth, M. B. Bo¨hme, S. Dapprich, A. W. Ehlers, A.
Gobbi, V. Jonas, K. F. Ko¨hler, R. Stegmann, A. Veldkamp and
G. Frenking, Chem. Phys. Lett., 1993, 208, 237.
2
3
4
´
C. Grunwald, O. Gevert, F. Wolf, P. Gonzalez-Herrero and H.
¨
Werner, Organometallics, 1996, 15, 1960.
30 P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973,
28, 213.
1462
New J. Chem., 2003, 27, 1451–1462