Catalytic ionic hydrogenations of ketones using molybdenum and tungsten complexes

Article abstract of DOI:10.1039/b107754f

Ketone complexes [CpM(CO)₂(PR₃)(η¹-Et₂C=O)] ⁺BAr′₄^- (R = Ph or Me; M = Mo or W) were prepared from hydride transfer from Cp(CO)₂(PR₃)MH to Ph₃C⁺BAr′₄^- [Ar′ = 3,5-bis(trifluoromethyl)phenyl] in the presence of (3-pentanone. These ketone complexes are catalyst precursors for hydrogenation of Et₂C=O under mild conditions (23°C, <4 atm H₂). Analogous catalytic hydrogenations are obtained from reaction of the PCy₃ complexes Cp(CO)₂(PCy₃)MH with Ph₃C⁺BAr′₄^-. The proposed mechanism involves displacement of the ketone by H₂, producing a cationic metal dihydride [CpM(CO)₂(PR₃)(H)₂]⁺. Proton transfer from the dihydride gives a protonated ketone, followed by hydride transfer from the neutral metal hydride CpM(CO)₂(PR₃)H to produce the alcohol complex [CpM(CO)₂(PR₃)(Et₂CHOH)]⁺. The free alcohol product is released from the metal through displacement by H₂ or ketone, completing the catalytic cycle. In most cases, conversion of the ketone or alcohol complexes to the dihydride is the turnover-limiting step of the catalytic cycle, with ketone and alcohol complexes being observed during the reaction. For reactions using the W-PCy₃ system, the dihydride [CpW(CO)₂(PCy₃)(H)₂]⁺ is observed as the resting state of the catalytic process. Proton transfer is slow and becomes turnover-limiting in this case. The Mo catalysts are more active than W, and the dependence on phosphine is PCy₃ > PPh₃ > PMe₃. The turnover rates are slow, with the fastest initial rate of about 2 turnovers per hour found for the Mo-PCy₃ system. This ionic hydrogenation mechanism does not require coordination of the ketone to the metal for the hydrogenation, thus differing from traditional mechanisms where coordination of a ketone to a metal precedes insertion of the ketone into a M-H bond.

Full text of DOI:10.1039/b107754f

View Article Online / Journal Homepage / Table of Contents for this issue
Catalytic ionic hydrogenations of ketones using molybdenum and
tungsten complexes†
Mark H. Voges and R. Morris Bullock*
Chemistry Department, Brookhaven National Laboratory, Upton, New York 11973-5000, USA
Received 28th August 2001, Accepted 2nd October 2001
First published as an Advance Article on the web 21st January 2002
Ketone complexes [CpM(CO) (PR )(η1-Et C᎐O)]ϩBArЈ Ϫ (R = Ph or Me; M = Mo or W) were prepared from
᎐
2
3
2
4
hydride transfer from Cp(CO)2(PR3)MH to Ph3CϩBArЈ4Ϫ [ArЈ = 3,5-bis(trifluoromethyl)phenyl] in the presence of
3-pentanone. These ketone complexes are catalyst precursors for hydrogenation of Et C᎐O under mild conditions
᎐
2
(23 ЊC, <4 atm H2). Analogous catalytic hydrogenations are obtained from reaction of the PCy3 complexes
Ϫ
Cp(CO)2(PCy3)MH with Ph3CϩBArЈ4 . The proposed mechanism involves displacement of the ketone by H2,
producing a cationic metal dihydride [CpM(CO)2(PR3)(H)2]ϩ. Proton transfer from the dihydride gives a protonated
ketone, followed by hydride transfer from the neutral metal hydride CpM(CO)2(PR3)H to produce the alcohol
complex [CpM(CO)2(PR3)(Et2CHOH)]ϩ. The free alcohol product is released from the metal through displacement
by H2 or ketone, completing the catalytic cycle. In most cases, conversion of the ketone or alcohol complexes to
the dihydride is the turnover-limiting step of the catalytic cycle, with ketone and alcohol complexes being observed
during the reaction. For reactions using the W–PCy3 system, the dihydride [CpW(CO)2(PCy3)(H)2]ϩ is observed
as the resting state of the catalytic process. Proton transfer is slow and becomes turnover-limiting in this case. The
Mo catalysts are more active than W, and the dependence on phosphine is PCy3 > PPh3 > PMe3. The turnover
rates are slow, with the fastest initial rate of about 2 turnovers per hour found for the Mo–PCy3 system. This ionic
hydrogenation mechanism does not require coordination of the ketone to the metal for the hydrogenation, thus
differing from traditional mechanisms where coordination of a ketone to a metal precedes insertion of the
ketone into a M–H bond.
of H to the C᎐O bond. In addition to attempting to develop
Introduction
᎐
2
catalytic pathways that rely on different reactivity patterns for
the M–H bonds, this offers the possibility of using less costly
metals such as molybdenum or tungsten as alternatives to the
use of more expensive rhodium or ruthenium in the well-
established pathways cited above.
An example of the potential that alternative hydrogenation
methods offer is provided by the fantastically reactive Ru cat-
alysts developed by Noyori and co-workers.5,6 Some of these
catalyze the highly enantioselective hydrogenation of ketones
using H2, while others operate by transfer hydrogenation, using
isopropanol as the source of the hydrogen. These and related
catalysts that now appear to proceed through mechanisms that
do not require coordination of the ketone to the metal will be
discussed in a later section.
An alternative method for delivery of H2 is through ionic
hydrogenation, in which a proton and a hydride are sequentially
transferred to a ketone or other substrate. Kursanov and co-
workers pioneered the use of CF3CO2H as the Hϩ source and
HSiEt3 as the HϪ donor for use in the stoichiometric ionic
hydrogenations of ketones and numerous other unsaturated
organic compounds.7 The versatile reactivity patterns of metal
hydrides suggest the possibility of using a metal hydride bond
as a source of protons, and a second metal hydride bond as a
hydride donor in ionic hydrogenations. These two types of
heterolytic cleavage have been demonstrated. Metal hydrides
can serve as acids; the kinetic and thermodynamic aspects of
proton transfer from metal hydrides have been established.8
Metal hydrides can function as hydride donors,9 and the
kinetics of hydride transfer reactions from a series of metal
hydrides toward Ph3Cϩ have been reported.10,11
Hydrogenation of ketones to alcohols is a reaction of fund-
amental importance and practical utility, being used in the
synthesis of fine chemicals as well as compounds used in
the pharmaceutical and agricultural industries. Main group
hydride reagents such as NaBH4 and LiAlH4 are often used for
the reduction of C᎐O bonds; these reactions are well-developed
᎐
and immensely useful, but they consume stoichiometric
amounts of the hydride reagent and produce waste by-products.
Many transition metal homogeneous hydrogenation catalysts
have been developed. Rhodium dihydride catalysts for ketone
hydrogenation were reported by Shrock and Osborn over 30
years ago,1 and highly enantioselective hydrogenations using
Rh catalysts have been reported more recently.2 Ruthenium
phosphine complexes have also been shown to be effective cat-
alysts for ketone hydrogenations.3 The traditional mechanism
for these hydrogenations involves coordination of the ketone to
the metal, followed by insertion of the ketone into a metal–
hydrogen bond (eqn. (1)).4 The alcohol product is released from
(1)
the metal through a reductive elimination reaction involving the
second metal hydride bond. This type of mechanism relies on
the ability of the metal hydride bond to insert ketones. If such a
step were not required, then new types of catalysts could poten-
tially be developed based on different mechanisms for delivery
In view of our intention of having both proton and hydride
donors for the metal-catalyzed ionic hydrogenation of ketones,
the next goal was to incorporate both of these reactions in a
† Based on the presentation given at Dalton Discussion No. 4, 10–13th
January 2002, Kloster Banz, Germany.
DOI: 10.1039/b107754f
J. Chem. Soc., Dalton Trans., 2002, 759–770
This journal is © The Royal Society of Chemistry 2002
759

View Article Online
single metal center. Meeting this challenge requires proper
balancing of both steps, since an acidic dihydride as the proton
donor would be coupled to a conjugate base (the metal hydride)
that must be adequately hydridic to carry out the hydride trans-
fer. Each step of the mechanism may be sensitive to the steric
and electronic properties of the ligands, and these effects could
be opposite for the proton and hydride transfers. For example,
a more electron-donating phosphine (e.g., PMe3 > PPh3)
(3)
ϩ
will make a cationic MH2 less acidic,8 but will also make the
neutral MH more hydridic.10,11
We established the use of transition metal hydrides as the
the corresponding band of free Et C᎐O at 1712 cmϪ1. These
hydride donor in stoichiometric ionic hydrogenations. The C᎐C
᎐
2
᎐
spectroscopic characteristics indicate that the ketone in
bonds of some hindered alkenes are hydrogenated at Ϫ50 ЊC
through reaction with HOTf (OTf = OSO2CF3) and metal
[CpW(CO) (PPh )(η1-Et C᎐O)]ϩ is bound to tungsten through
᎐
2
3
2
donation from a lone pair on oxygen, in an η1, or σ, bonding
mode, based on diagnostic NMR and IR criteria established to
distinguish between η1(σ) and η2 (π) bonding modes.23
carbonyl hydrides.12 The C᎐C bonds of certain alkynes are
᎐
᎐
also hydrogenated by HOTf and Cp(CO)3WH,13 and related
reactions can be used for the conversion of acetals to ether
complexes.14 Ionic hydrogenations of aldehydes or ketones by
HOTf and Cp(CO)3WH lead to the formation of kinetically
stabilized alcohol complexes (eqn. (2)).15,16 The viability of
Similar reactions of the molybdenum and tungsten hydrides
Cp(CO)2(PPh3)MoH, Cp(CO)2(PMe3)MoH, and Cp(CO)2-
(PMe3)WH led to the isolation of the corresponding ketone
complexes. These ketone complexes decompose slowly when
left in CD2Cl2 solution for a few days, but this decomposition is
inhibited by the presence of excess free ketone, i.e., conditions
under which the catalytic reactions are carried out. Whereas the
(2)
isolated sample of [CpW(CO) (PPh )(η1-Et C᎐O)]ϩ was the cis
᎐
2
3
2
isomer, the ketone complexes [CpMo(CO) (PPh )(Et C᎐O)]ϩ,
᎐
2
2
3
[CpMo(CO) (PMe )(Et C᎐O)]ϩ,
and
[CpW(CO)2(PMe3)-
᎐
2
2
3
(Et C᎐O)]ϩ were initially observed as mixtures of cis and trans
᎐
2
isomers. Prior studies of neutral and cationic Mo and W com-
plexes with “four-legged piano stool” geometries established 1H
NMR criteria to readily distinguish cis from trans isomers.10,24
The Cp resonance of the trans isomer is a doublet (JPH ≈ 2 Hz)
appearing about 0.2–0.3 ppm upfield of the singlet resonance
due to the Cp of the cis isomer. Isomerization of the isolated
cis/trans mixture occurs readily under mild conditions. For
using metal hydrides as the hydride donor is thus demonstrated,
albeit in reactions that require stoichiometric amounts of both
the acid and the hydride donor. The goal of making such reac-
tions catalytic in the metal would be more interesting, using
hydrogen as the ultimate source of both Hϩ and HϪ. Obtaining
this goal requires two key reactions of the catalyst in addition
to the hydride transfer reactions already demonstrated—it must
be capable of protonating the ketone and capable of reacting
with H2 to regenerate two M–H bonds. Our initial success in
this endeavor was reported in a preliminary communication.17
We report here the preparation of ketone complexes of molyb-
denum and tungsten, and a study of their use in catalytic
hydrogenation of ketones by an ionic mechanism.
example, the trans isomer of [CpMo(CO) (PPh )(Et C᎐O)]ϩ
᎐
2
3
2
dominated in the isolated product, but an NMR spectrum
taken after 16 hours showed nearly complete isomerization to
the cis isomer, with only 2% trans isomer remaining. The
isomerization appears to occur only in solution and not in the
solid state.
Synthesis of alcohol complexes [Cp(CO)2(PR3)M(Et2CHOH)]؉
The alcohol complexes cis-[CpM(CO)2(PPh3)(Et2CHOH)]ϩ
were prepared (eqn. (4)) for both M = Mo and W, from reac-
Results and discussion
Synthesis of ketone complexes [Cp(CO) (PR )M(ꢀ1-Et C᎐O)]؉
(4)
᎐
2
3
2
Ketone complexes are often prepared by addition of a ketone to
a metal complex containing a weakly coordinating ligand. Beck
and co-workers extensively developed the chemistry of metal
complexes with weakly bound BF4 and PF6 ligands.18 They
Ϫ
Ϫ
prepared Mo and W complexes such as Cp(CO)3MoFBF3 by
Ϫ 19,20
reaction of Cp(CO)3MoH with Ph3CϩBF4 .
Reaction of
these M–FBF3 complexes with ketones produced cationic
ketone complexes.19,21 The ketone complexes used in our studies
were synthesized by a closely related method, using BArЈ4
tions similar to those used in the preparation of the ketone
complex (eqn. (3)). These alcohol complexes were isolated and
fully characterized. The trans isomers of these same alcohol
complexes were prepared through hydrogenation of the ketone
complexes, as will be discussed in a later section.
Ϫ
[ArЈ = 3,5-bis(trifluoromethyl)phenyl] as the counter ion.
Ϫ
Hydride transfer from Cp(CO)2(PPh3)WH to Ph3CϩBArЈ4
in the presence of 3-pentanone, produces the ketone com-
plex cis-[CpW(CO) (PPh )(η1-Et C᎐O)]ϩBArЈ Ϫ (eqn. (3)). This
Reaction of CpMo(CO)2(PMe3)H with Ph3CϩBArЈ4 in
Ϫ
᎐
2
3
2
4
complex was isolated as an orange solid in 88% yield and was
CH2Cl2, followed by addition of Et2CHOH, did not produce
1
characterized by H, 13C and 31P NMR, IR, and elemental
[CpMo(CO)2(PMe3)(Et2CHOH)]ϩ, but instead led to the
analysis. The 13C resonance of the bound ketone carbonyl
appears at δ 236.2, which is downfield of the carbonyl reson-
ance of free Et C᎐O at δ 212.5. The CO ligand trans to
isolation of {[CpMo(CO)2(PMe3)]2(µ-H)}ϩBArЈ4 , a bimetallic
Ϫ
complex with a bridging hydride (eqn. (5)). Hydride transfer
from CpMo(CO)2(PMe3)H to Ph3Cϩ would produce [CpMo-
(CO)2(PMe3)(CH2Cl2)]ϩ, in which the 16-electron cation “Cp-
Mo(CO)2(PMe3)ϩ” is weakly solvated by CH2Cl2. Evidence
for this complex was obtained from low-temperature NMR
experiments in hydride abstractions from CpMo(CO)2(PMe3)H
᎐
2
PPh3 appears as a doublet (2JPC = 6 Hz) at 240.1; the CO ligand
cis to PPh3 has a larger coupling constant (2JPC = 21 Hz).22
A
weak band in the IR spectrum at 1632 cmϪ1 is assigned to the
ν(C᎐O) of the bound ketone, at an energy lower than that of
᎐
760
J. Chem. Soc., Dalton Trans., 2002, 759–770

View Article Online
(5)
with Ph3CϩPF6 . Reaction of a second equivalent of CpMo-
(CO)2(PMe3)H with [CpMo(CO)2(PMe3)(CH2Cl2)]ϩ would dis-
place the weakly bound solvent ligand, producing the bridging
hydride complex. This reaction pathway may be favored due to
the higher nucleophilicity of CpMo(CO)2(PMe3)H compared
to CpMo(CO)2(PPh3)H. The Mo–H bond of CpMo(CO)2-
(PMe3)H, with a trialkylphosphine PMe3, is expected to be
less acidic than CpMo(CO)2(PPh3)H, which has a PPh3 ligand.8
Norton and co-workers found examples where the nucleophil-
icity of metal hydrides was largely the opposite of the order of
kinetic acidity.25 Thus our reaction is consistent with their
observations, since CpMo(CO)2(PMe3)H would appear to be
more nucleophilic than CpMo(CO)2(PPh3)H. Reactions in
which a neutral metal hydride reacts with an unsaturated (or
weakly coordinated) metal complex have led to many other
examples of bridging hydrides, as reviewed by Venanzi.26
Another synthetic route to bridging hydride complexes is
through protonation of metal–metal bonded dimers. Nataro
and Angelici previously reported that {[CpMo(CO)2(PMe3)]2-
(µ-H)}ϩOTfϪ is formed by protonation of the neutral dimer
Ϫ 10
Fig.
1
Time profile of the catalytic hydrogenation of Et2C᎐O
᎐
(300 mM) by [CpW(CO)2(PPh3)(Et2C᎐O)]ϩ (30 mM) under H2 (<4 atm)
᎐
in CD2Cl2 at 23 ЊC.
(Et C᎐O)]ϩBArЈ . Slow hydrogenation occurs, producing the
Ϫ
᎐
2
4
with HOTf.27 As expected, our complex with the BArЈ4
Ϫ
alcohol Et2CHOH, giving 1 turnover in 4 days and 4 turnovers
after 24 days. Fig. 2 shows a comparison of the activity of
counter ion has IR and NMR spectra very similar to those
reported27 for Angelici’s complex with the OTfϪ counter ion.
Catalytic hydrogenations with ketone complexes
[Cp(CO) (PPh )M(ꢀ1-Et C᎐O)]؉ (M ؍ Mo, W)
᎐
2
3
2
When solutions of the isolated ketone complexes in CD2Cl2
were reacted under H with Et C᎐O, hydrogenation of the C᎐O
᎐
2
᎐
2
bond occurs, leading to the alcohol product Et2CHOH (eqn.
(6)). These experiments were carried out in NMR tubes sealed
(6)
under 1 atm H2 at liquid nitrogen temperatures. When warmed
to room temperature, this leads to an actual pressure in the tube
of almost 4 atmospheres. Under these conditions the concen-
tration of H2 dissolved in the CD2Cl2 is about 15 mM, as
determined by integration of the resonance for dissolved H2 at
Fig. 2 Comparison of catalytic hydrogenation of Et2C᎐O (300 mM)
᎐
by [CpM(CO)2(PR3)(Et2C᎐O)]ϩ (25–30 mM) under H2 (<4 atm) in
᎐
CD2Cl2 at 23 ЊC.
1
δ 4.60. The amount of H2 in solution determined by H NMR
this and other catalysts. The formation of the hydrogenation
product alcohol is accompanied by smaller amounts of the
ether (Et2CH)2O arising from condensation of two alcohols
(eqn. (7)) After 24 days, the concentration of the alcohol Et2-
CHOH was 102 mM and that of the ether (Et2CH)2O was 6 mM.
integration is less than the actual amount because of the pres-
ence of 25% para-H2 present in H2, so a correction factor has to
be applied to account for this, multiplying the integrated
amount by 1.33 to determine the actual concentration. The
total amount of H2 in the tube is sufficient to hydrogenate all of
the ketone, but most of it is in the gas phase above the solution.
A variety of concentrations and conditions were used to test the
various ketone complexes as catalysts. A standard set of con-
ditions was chosen for comparison of the catalysts: 25–30 mM
(7)
Ϫ
[CpM(CO) (PR )(Et C᎐O)]ϩBArЈ and 300 mM Et C᎐O (10–
᎐
᎐
2
2
3
2
4
As the reaction proceeds, the concentration of the initial
12 equivalents) in CD2Cl2 at 23 ЊC under <4 atmospheres H2.
An internal standard for integration of the 1H spectra was used.
Carrying out these experiments with only about 10 equivalents
of ketone provided a sufficiently low ratio of ketone to metal
such that the organometallic complexes as well as the organic
starting material and products could all be integrated accur-
ately by 1H NMR. In addition, 31P spectra were recorded, pro-
viding an independent measure on the relative amounts of the
organometallic species.
ketone complex [CpW(CO) (PPh )(Et C᎐O)]ϩ decreases, with
᎐
2
3
2
the concomitant appearance of new NMR resonances due to
the alcohol complex trans-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ. We
previously found that alcohol complexes were formed in stoichi-
ometric ionic hydrogenations of ketones (eqn. (2)), and these
new observations show they are involved in catalytic hydro-
genations as well. The concentration of this alcohol complex
surpasses that of the ketone complex as the reaction proceeds.
As indicated in Fig. 1, ketone and alcohol complexes were
the predominant tungsten products throughout the catalytic
Fig. 1 shows the time profile of the concentrations from
the catalytic hydrogenation of Et C᎐O by [CpW(CO) (PPh )-
᎐
2
2
3
J. Chem. Soc., Dalton Trans., 2002, 759–770
761

View Article Online
Ϫ
reaction. Information about a catalyst deactivation pathway
was provided by observation of about 8% of the phosphonium
acetone to a solution of [CpW(CO)3(Me2CHOH)]ϩBArЈ4
caused its OH resonance to shift downfield to δ 6.74, suggesting
that acetone was serving as a hydrogen bond acceptor in an
O–H ؒ ؒ ؒ O hydrogen bond.16 A plausible explanation for the
different chemical shifts of the OH protons of the cis and trans
alcohol complexes is that hydrogen bonding to free ketone is
occurring in trans-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ and other
trans alcohol complexes, which are observed during the hydro-
genations, with excess ketone present. In contrast, the spectra
for the cis alcohol complexes are recorded on isolated samples
with no ketone present to serve as a hydrogen bond acceptor.
When isolated cis-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ was used
as a catalyst precursor for the hydrogenation of Et2CO, the
alcohol complex was promptly converted to the ketone complex
cation HPPh3 , identified by 31P NMR and confirmed by com-
ϩ
parison with an independently prepared sample. The proton-
ated phosphine presumably forms by protonation of free PPh3
by the cationic dihydride [CpW(CO)2(PPh3)(H)2]ϩ, as shown in
eqn. (8). Small amounts of the neutral tungsten hydride
CpW(CO)2(PPh3)H are also observed at the end of the catalytic
reaction, as implied by the stoichiometry of eqn. (8). While the
(8)
cis-[CpW(CO) (PPh )(Et C᎐O)]ϩ, with <3% of the alcohol
᎐
2
3
2
ϩ
formation of HPPh3 according to eqn. (8) appears entirely
complex remaining after 15 minutes. As the hydrogenation
proceeds, the trans alcohol complex trans-[CpW(CO)2(PPh3)-
(Et2CHOH)]ϩ forms. This trans alcohol complex exists in the
presence of ketone, and has a much higher kinetic stability
towards displacement by ketone, compared to the cis alcohol
complex.
plausible, the more pertinent question concerns the origin of
the free PPh3. Our observations do not identify which inter-
mediate(s) undergo loss of phosphine. The formation of the
free phosphine indicates that decomposition of one tungsten
species has occurred, but the problem is compounded by the
subsequent reactivity. Since the protonation of the phosphine
consumes a proton, it removes a second tungsten species from
the functioning catalytic cycle.
Catalysis of ketone hydrogenation using the analogous Mo–
PPh3 complex enabled a comparison of W and Mo complexes
under the same conditions. Catalysis by the Mo complex
Ϫ
As noted above, the decline in the concentration of the
ketone complex during the catalysis is accompanied by
the appearance of trans-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ. This
alcohol complex is produced as the predominant product
[CpMo(CO) (PPh )(Et C᎐O)]ϩBArЈ produces 3 turnovers in
᎐
2
3
2
4
the first day (Fig. 2), which is significantly faster than that
observed with the W analog. The time progression of the metal
complexes observed during catalysis is similar in this Mo case
to the W example discussed above, with trans-[CpMo(CO)2-
(PPh3)(Et2CHOH)]ϩBArЈ4Ϫ becoming the major Mo species as
the catalysis proceeds.
(94%) when a CD Cl solution of [CpW(CO) (PPh )(Et C᎐O)]ϩ
᎐
2
2
2
3
2
(29 mM) is reacted with 65 atm (950 psi) hydrogen at 22 ЊC in
the presence of Et2CO (5 equivalents) for 17 hours. This reac-
tion gives 1.3 turnovers, indicating that the higher pressure of
hydrogen accelerates the catalysis, but only by about a factor
of three. We found earlier that the rate of hydride transfer
from trans-Cp(CO)2(PCy3)MoH to Ph3CϩBF4Ϫ was much faster
than that from the isomeric cis-Cp(CO)2(PCy3)MoH.10 The
observation of the trans alcohol complex under catalytic con-
ditions indicates that the alcohol binds to the metal at the
site from which the hydride was transferred, without loss of
stereochemistry at the metal center. In our earlier studies of
stoichiometric hydrogenation of ketones using HOTf and
Cp(CO)3WH, it was recognized that there must be some W–O
bond formation in the transition state, prior to complete W–H
bond rupture.15,16
An alternative catalyst precursor was shown to produce
catalysis comparable to that described above from the isolated
ketone complex [CpMo(CO) (PPh )(Et C᎐O)]ϩBArЈ . We
Ϫ
᎐
2
3
2
4
reported that hydride transfer from CpMo(CO)2(PPh3)H to
Ph3CϩBArЈ4 produces a complex with an η3-PPh3 ligand, in
Ϫ
which one C᎐C bond of the arene ring is coordinated to the
᎐
molybdenum.28 Use of this complex as a catalyst precursor
for ketone hydrogenations was successful, and NMR spectra
indicated that the ketone readily displaced the C᎐C bond,
᎐
producing cis-[CpMo(CO) (PPh )(Et C᎐O)]ϩBArЈ Ϫ (eqn. (9)).
᎐
2
3
2
4
(9)
The availability of both cis and trans isomers of the alcohol
complex [CpW(CO)2(PPh3)(Et2CHOH)]ϩ invites a compari-
son of their spectroscopic characteristics. The two isomers are
readily distinguished by 1H, 13C or 31P NMR. The methyl
resonances of the Et2CHOH ligand of cis-[CpW(CO)2(PPh3)-
(Et2CHOH)]ϩ are non-equivalent in the both the H and 13C
1
Along with modifying the phosphine ligand bonded to the
metal, we briefly examined the effect of changing from an
unsubstituted Cp ligand to the more bulky and more electron-
donating C5Me5 (Cp*) ligand. Hydride transfer from Cp*-
NMR spectra, while they are equivalent in the trans isomer. The
chemical shifts of the OH resonances are of particular interest.
The OH resonance of trans-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ
is a doublet (JHH = 8.4 Hz) at δ 5.59. In contrast, the chemical
Mo(CO)2(PPh3)H to Ph3CϩBArЈ4 in the presence of Et2CO
Ϫ
Ϫ
shift of the OH of cis-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ BArЈ4
produced evidence for the formation of a ketone complex
analogous to that found in the Cp example. Surprisingly,
appears at δ 0.10, with couplings to both CH (JHH = 8.2 Hz) and
P (JPH = 2.9 Hz). Similar coupling constants but an even more
upfield chemical shift (δ Ϫ0.62) were found for the OH reson-
ance of the molybdenum alcohol complex cis-[CpMo(CO)2-
however, negligible hydrogenation of Et C᎐O was observed at
᎐
2
<4 atm H2 at 23 ЊC.
Ϫ
(PPh3)(Et2CHOH)]ϩ BArЈ4 . The chemical shifts of alcohol
Catalytic hydrogenations with Mo and W complexes with PCy3
and PMe3 ligands
Hydride transfer from Cp(CO)2(PCy3)MoH to Ph3CϩBArЈ4Ϫ in
complexes can be strongly influenced by hydrogen bonding
of the OH. In our studies of alcohol complexes synthesized
by stoichiometric ionic hydrogenations of ketones, we found
Ϫ
that the OH resonance of [CpW(CO)3(Me2CHOH)]ϩBArЈ4
CD Cl in the presence of Et C᎐O gave a single Mo product, as
᎐
2
2
2
appears at
δ
2.41, while [CpW(CO)3(Me2CHOH)]ϩOTfϪ
evidenced by 1H and 31P NMR spectra. Resonances for Et C᎐O
᎐
2
exhibits a resonance for its OH at δ 7.34.16 The triflate complex
was shown by X-ray crystallography to have strong hydrogen
bonding of the OH with an oxygen of the triflate anion, and the
downfield shift of the OH in the triflate complex suggests
hydrogen bonding is maintained in solution. But addition of
were broadened in the H NMR, suggesting exchange of free
1
and bound ketone, and the ketone complex [CpMo(CO)2-
(PCy )(Et C᎐O)]ϩ was not isolated. Hydride abstraction from
᎐
3
2
Cp(CO)2(PCy3)MoH (10 mM) in the presence of a large excess
(100 equiv.) of Et C᎐O was carried out, and H (<4 atm) was
᎐
2
2
762
J. Chem. Soc., Dalton Trans., 2002, 759–770

View Article Online
added. Hydrogenation of the ketone occurred at 23 ЊC, giving
18 turnovers after 5 hours, and increasing to 28 turnovers in
20 hours. Although this Mo complex produces the most active
hydrogenation catalyst of the series we have examined here, it
decomposes as the reaction proceeds. The formation of
HPCy3ϩ as a decomposition product was verified by 31P NMR.
This product accounted for about 65% of the total integrated
31P NMR intensities after 20 hours. The rate of hydrogenation
decreases as the reaction proceeds, offering evidence that
decomposition products play little or no role in the catalysis.
catalysts with PPh3 or PCy3 ligands. cis and trans isomers of the
ketone complex were the dominant species observed during
the catalysis, though other 31P resonances appeared later in the
reaction that may be due to the alcohol complex. Use of the
Ϫ
tungsten analog, [CpW(CO) (PMe )(Et C᎐O)]ϩBArЈ , gave
᎐
2
3
2
4
even slower reactivity, with only 2.3 turnovers being found after
ϩ
16 days. The phosphonium compound HPMe3 was observed
as a decomposition product in both the Mo and W cases, but in
much smaller amounts compared to the catalytic runs with
PPh3 and PCy3.
When hydrogenation of Et C᎐O was carried out under our
᎐
2
standard conditions (30 mM Mo, 300 mM ketone) enabling
comparisons to other catalysts (Fig. 2), Cp(CO)2(PCy3)MoH/
Mechanistic considerations
Ϫ
Scheme 1 shows the proposed mechanism for the catalytic ionic
Ph3CϩBArЈ4 produced about 6 turnovers in the first 4 hours.
Monitoring the progress of the reaction by NMR provided
evidence for the formation of the alcohol complex [CpMo-
(CO)2(PCy3)(Et2CHOH)]ϩ but this complex was not isolated
for definitive characterization.
The use of other counter ions was studied qualitatively for
catalysis with the Mo–PCy3 system. We previously prepared cis-
Cp(CO)2(PCy3)MoFBF3 in our study of the kinetics of hydride
transfer.10 When excess Et C᎐O is added to this complex,
᎐
2
Ϫ
displacement of the weakly bound BF4 ligand occurs, and
hydrogenation of the ketone occurs at room temperature under
ϩ
H2 (<4 atm). The formation of HPCy3 was observed as a
decomposition product in this example as well, and this cat-
alytic reaction was not studied in detail. Hydride transfer to
Ph3CϩPF6Ϫ from Cp(CO)2(PCy3)MoH in the presence of excess
Et C᎐O also gave evidence for a ketone complex, and hydro-
᎐
2
genation of Et C᎐O occurred at room temperature. Along with
᎐
2
ϩ
the formation of HPCy3 as a decomposition product, other
unidentified 31P NMR resonances were observed that may be
due to decomposition of the PF6 counter ion. Very slow cat-
Ϫ
Scheme 1
alysis of ketone hydrogenation was observed using Cp(CO)2-
(PCy3)MoOTf at <4 atm H2. Apparently the triflate anion is
coordinated too strongly to Mo to provide significant catalytic
activity under the mild conditions used for the reactions with
hydrogenation of ketones catalyzed by molybdenum and tung-
Ϫ
sten complexes [CpM(CO) (PR )(Et C᎐O)]ϩBArЈ . Starting
᎐
2
3
2
4
from the catalyst precursor ketone complexes, displacement of
the coordinated ketone by hydrogen is required to produce the
cationic metal dihydride (eqn. (10)). For most of the systems
Ϫ
BArЈ4 , BF4Ϫ and PF6Ϫ anions.
The reactivity of the W complex with a PCy3 ligand provided
an informative contrast to the Mo analog. Following hydride
Ϫ
transfer from Cp(CO)2(PCy3)WH to Ph3CϩBArЈ4 , no defini-
(10)
tive evidence was obtained for the bound ketone complex
[CpW(CO) (PCy )(Et C᎐O)]ϩ. The NMR data do not dis-
᎐
2
3
2
tinguish between several possible products, such as a fluxional
ketone complex, or a complex of constitution [CpW(CO)2-
(PCy3)]ϩ, possibly with an agostic interaction with a CH bond
of the phosphine ligand. An agostic interaction involving a CH
of one PCy3 ligand was shown by X-ray crystallography for
W(CO)3(PCy3)2,29 and kinetics studies by Hoff and co-workers
provide evidence for a role of this agostic interaction in ligand
substitution reactions.30 Further study would be required to
identify the initial product found in our studies, but the focus of
our current efforts is on the reactivity under catalytic con-
ditions. When hydrogen is added to the solution under our
standard conditions (Fig. 2), hydrogenation of the ketone
occurs, with 3 turnovers occurring after 8 days. This W–PCy3
complex produces a slower catalyst than the Mo–PCy3 com-
plex, similar to the trend found in the comparison of W vs. Mo
for the PPh3 ligands. Also the increase in reactivity observed
upon changing from PPh3 to PCy3 is less pronounced for the W
case than for the Mo example. The most conspicuous difference
is found in the metal-containing complexes observed during the
progress of the catalytic reaction. The dihydride31 [CpW(CO)2-
studied here the ketone complexes [CpM(CO) (PR )(Et C᎐O)]ϩ
᎐
2
3
2
are observed during the catalytic reaction as monitored by
NMR (and later in the reaction the alcohol complexes). Thus
conversion of the ketone or alcohol complex to the dihydride is
the turnover-limiting step of the catalytic cycle in most cases.
A complete determination of the dependence of the catalytic
rate on hydrogen pressure has not been made, but as noted
above a comparison of the amount of hydrogenation of Et C᎐O
᎐
2
by [CpW(CO) (PPh )(Et C᎐O)]ϩ under <4 atm H compared to
᎐
2
3
2
2
65 atm H2 shows that the higher pressure makes it only about
three times faster. This suggests that the rate of catalysis is not
strongly dependent on [H2] under these conditions, and that the
mechanism for displacement of the ketone by H2 is largely dis-
sociative. Factors that promote dissociation of the ketone or
alcohol from the metal are expected to enhance the rate of
conversion to the dihydride, and this is proposed as the primary
explanation for complexes with PCy3 ligands producing faster
catalysts than those with the smaller phosphines PPh3 and
Ϫ
(PCy3)(H)2]ϩBArЈ4 is observed as the predominant tungsten
complex during the reaction, differing from the other examples
discussed above where ketone and alcohol complexes persisted
during the catalysis.
Hydrogenation of Et C᎐O is catalyzed by [CpMo(CO) -
᎐
2
2
Ϫ
(PMe )(Et C᎐O)]ϩBArЈ (Fig. 2) but the reaction is much
᎐
3
2
4
slower (2 turnovers in 8 days) than that observed for the Mo
J. Chem. Soc., Dalton Trans., 2002, 759–770
763

View Article Online
PMe3. The rate of these catalytic reactions increases with
increasing size of the phosphines, based on the steric measure
of cone angles32 (θ), with PCy3 (θ = 170Њ) > PPh3 (θ = 145Њ) >
PMe3 (θ = 118Њ). An important mechanistic distinction between
the traditional catalysts (eqn. (1)) and our ionic hydrogenation
catalysts concerns coordination of the ketone. In the con-
ventional mechanism, ketone coordination is required, and
insertion of the ketone into the M–H bond forms the C–H
bond. Ketone binding is observed in most of our cases, but is
not required for the hydrogenation to proceed. Dissociation of
the ketone is the slow step, since displacement of the ketone by
H2 is necessary to form the dihydride complex that carries out
the hydrogenation. The proton and hydride transfer from the
M–H bonds occurs to free rather than bound ketone in our
catalytic reactions.
(11)
Ϫ
ated by [CpW(CO)2(PMe3)(H)2]ϩBArЈ4 , through proton trans-
fer from the dihydride and hydride transfer from the neutral
metal hydride. This reaction is 90% complete in 1 hour, and
produces the trans alcohol complex. Ketone protonation pre-
sumably occurs by a direct tungsten-to-oxygen proton transfer.
This may also be true in hydrogenations of aldehydes and
ketones that we previously reported using [CpW(CO)2(PMe3)-
(H)2]ϩOTfϪ.16 In the case of the hydrogenations by dihydrides
with the OTfϪ counter ion, however, there exists the alternative
possibility that the proton transfer is mediated by OTfϪ as a
kinetically competent proton carrier. Darensbourg and co-
workers found that chloride and other anions can mediate
proton transfers from cationic dihydrides and thereby influence
the rates of proton transfer.36
The catalytic experiments were carried out using [CpW-
(CO) (PPh )(Et C᎐O)]ϩ under H with excess ketone added, but
᎐
2
3
2
2
the existence of the equilibrium shown in eqn. (10) was verified
from a reaction of [CpW(CO) (PPh )(Et C᎐O)]ϩ with hydrogen
᎐
2
3
2
(<4 atm H2) in the absence of added ketone. After 2 hours, 40%
of the ketone complex had been converted to the dihydride
[CpW(CO)2(PPh3)(H)2]ϩ. At t = 18 hours, about 85% con-
version to the dihydride was found, giving an apparent Keq
≈
2 for eqn. (10). By the time this spectrum was recorded, how-
ever, some hydrogenation of the ketone had occurred, as
expected. The equilibrium is established very slowly, contri-
buting to the slow rate of catalysis, which requires the form-
ation of the dihydride complex. Although a reliable value of Keq
was not determined from these experiments, the observation of
the formation of the dihydride documents the viability of this
step as a part of the catalytic cycle. Since alcohol complexes are
formed as the catalysis proceeds, displacement of the alcohol
ligands by H2 is also required to maintain the catalysis. A sol-
ution of cis-[CpW(CO)2(PPh3)(Et2CHOH)]ϩ placed under H2
at 22 ЊC was also shown to release the free alcohol and produce
the dihydride [CpW(CO)2(PPh3)(H)2]ϩ.
In contrast to the resting state of the catalyst being the
ketone and alcohol complexes for most cases, the reactions
carried out starting with Cp(CO)2(PCy3)WH show the dihydride
[CpW(CO)2(PCy3)(H)2]ϩ as the predominant metal complex. In
this case, the proton transfer step is slow and becomes turnover-
limiting. The slower rate here, compared to the W–PPh3 com-
plex, may be due to steric hindrance in the transition state for
proton transfer from [CpW(CO)2(PCy3)(H)2]ϩ to the ketone,
with the bulky PCy3 ligand impeding approach of the ketone.
The PMe3 dihydride [CpW(CO)2(PMe3)(H)2]ϩBArЈ4 , with a
Ϫ
PMe3 ligand that is similar to PCy3 electronically but very
different sterically, is faster at proton transfer. A related
thermodynamic issue affecting the relative stability of the
ketone complex vs. the dihydride in the PCy3 case may be the
destabilization of bonding of the ketone to the metal, again
owing to the size of the PCy3 ligand. The small size of two
hydride ligands allows them to bond to tungsten with less
hindrance from the bulky PCy3 ligand.
The tungsten dihydrides [CpW(CO)2(PR3)(H)2]ϩ were all
previously synthesized (R = Me, Ph, Cy) by protonation of the
neutral hydrides CpW(CO)2(PR3)H.31 The dihydrides can be
isolated, and were fully characterized, including a crystal struc-
ture for [CpW(CO)2(PMe3)(H)2]ϩOTfϪ. Even though they are
dihydrides in their stable form, this does not preclude η2-H2
dihydrogen complexes33 as intermediates in the formation of
the dihydrides. Norton and co-workers recently studied the
kinetics of protonation of CpW(CO)2(PMe3)H, and found that
protonation of the W–H bond to produce [CpW(CO)2(PMe3)-
(η2-H2)]ϩ is faster than protonation of the metal to give the
dihydride directly.34 Analogous dihydrogen complexes could be
unobserved intermediates in the formation of the dihydrides
from the reactions of the ketone (and alcohol) complexes
with H2.
While the steric properties of the phosphines are a major
factor influencing the rates, the thermodynamics are also
important. Third row metals tend to form stronger bonds to
hydrogen (and other ligands) than second row metals; for
Cp(CO)3MH the M–H bond is about 3 kcal molϪ1 stronger for
M = W than for M = Mo,35 so the increased stability of the
tungsten dihydrides is not surprising. The dihydrides are known
for the tungsten complexes but have not been observed for
molybdenum. It is possible that the molybdenum complexes
form dihydrogen complexes rather than dihydrides that are
active (but unobserved) in the catalytic cycle. Whether a dihy-
dride or a dihydrogen complex is formed is not critically
important, as long as the form that exists under catalytic con-
ditions has sufficient acidity to enable the required proton
transfer to the ketone.
The kinetics and thermodynamics of proton transfer from
metal hydrides are well documented.8 Cationic metal dihydrides
and dihydrogen complexes can be especially acidic;33 a dicat-
ionic dihydrogen complex studied by Morris and co-workers
was shown to be more acidic than HOTf.37 Norton and co-
workers reported a pKa of 5.6 for [CpW(CO)2(PMe3)(H)2]ϩ in
CH3CN. Protonated acetone in CH3CN is reported to have pKa
≈ Ϫ0.138,39 These values lead to the conclusion that the proton
transfer from metal to oxygen of the ketone in our reactions is
thermodynamically uphill. This analysis assumes that the rel-
ative pKa values are not greatly different in our solvent CD2Cl2,
compared to the CH3CN in which these pKa measurements
were made. Stoichiometric hydrogenations of ketones by
[CpW(CO)2(PMe3)(H)2]ϩ proceed smoothly when carried out in
CD3CN, supporting the contention that the change in solvent
will not invalidate any of these conclusions. Proton transfers
from the dihydride [CpW(CO)2(PMe3)(H)2]ϩ as the proton
source are likely less favorable thermodynamically compared to
those from the PPh3 analog. The dihydride [CpW(CO)2(PPh3)-
(H)2]ϩ is presumably more acidic, based on trends in pKa values
of neutral metal hydrides when electronic properties of ligands
are changed.8 Similarly, we assume that the unobserved Mo
dihydride [CpMo(CO)2(PMe3)(H)2]ϩ is more acidic than the W
analog. This is also in accord with observed trends in acidities
as metals are varied; Cp(CO)3MoH is a stronger acid than
Cp(CO)3WH thermodynamically (pKa = 13.9 in CH3CN for
Cp(CO)3MoH; pKa = 16.1 in CH3CN for Cp(CO)3WH).40 The
Once the dihydride complex is formed, the next step in the
hydrogenation mechanism is proton transfer to the ketone.
Evidence for the viability of this proton transfer step comes
from eqn. (11), where Et C᎐O is stoichiometrically hydrogen-
᎐
2
764
J. Chem. Soc., Dalton Trans., 2002, 759–770

View Article Online
rates follow the same order as the thermodynamics, with the
kinetics of proton transfer to aniline from Cp(CO)3MoH being
faster than from Cp(CO)3WH.41
unobserved molybdenum dihydride [CpMo(CO)2(PPh3)(H)2]ϩ
(or dihydrogen complex [CpMo(CO)2(PPh3)(η2-H2)]ϩ), fol-
lowed by intermolecular deprotonation by the amine base.
An alternative mechanism to consider for this reaction
would involve an intermediate or transition state having a
Mo ؒ ؒ ؒ H ؒ ؒ ؒ H ؒ ؒ ؒ N interaction. Numerous examples of
M–H ؒ ؒ ؒ H–N interactions have been discovered since 1994,
with examples involving iridium being particularly prevalent.42
Interactions of this type can be involved in the base-promoted
heterolytic cleavage of hydrogen.43 This unconventional bond-
ing to hydrogen involving protic and hydridic hydrogens has
been called a dihydrogen bond, and recent reviews of this topic
have been published.44
The unfavorable equilibrium for proton transfer from the
dihydride to the ketone does not prevent the reaction from
proceeding since the hydride transfer from the neutral metal
hydride to the protonated ketone is fast. Metal hydrides have
long been known to be capable of undergoing cleavage of the
M–H bond as a hydride.9 We reported the kinetics of hydride
Ϫ
transfer from a series of metal carbonyl hydrides to Ph3CϩBF4
in CH2Cl2,10,11 and found that the rate constants for the kinetic
hydricity span a range of over 106 (eqn. (12)). The second-order
(12)
Comparison with other hydrogenation catalysts
Ϫ
rate constant at 25 ЊC for hydride transfer to Ph3CϩBF4 from
Cp(CO)3MoH was kHϪ = 3.8 × 102 MϪ1 sϪ1, compared to kHϪ
=
Most of the ketone hydrogenation catalysts previously reported
use expensive metals such as Rh and Ru. In contrast, our
method uses molybdenum and tungsten, which are much less
expensive. Costs of metal starting materials fluctuate, but
precious metals like Rh can be orders of magnitude more
expensive than Mo or W. The cost of the metal is one of many
issues that can affect the overall economics of conducting
hydrogenation on an industrial scale; specialized ligands can
also add substantially to the cost of the catalyst.
The advantage of inexpensive metals must be balanced
against the activity obtained. At this early stage of develop-
ment, the catalysts described here are orders of magnitude less
reactive than the ruthenium catalysts reported by Noyori and
co-workers.5,6 Our efforts thus far have focused on the rational
design of catalytic systems that could be demonstrated to
hydrogenate ketones under mild conditions of temperature and
pressure. The results described here provide evidence for the
viability of unconventional mechanisms for catalytic hydro-
genations. It is expected that further studies may ultimately
lead to the design of more efficient catalysts, based on the
mechanistic understanding gained thus far.
7.6 × 101 MϪ1 sϪ1 for Cp(CO)3WH. It is thus reasonable to
expect the Mo complexes postulated in Scheme 1 to be both
faster proton donors as well as faster hydride donors in com-
parison to the corresponding W complexes. These steps may
not have a direct influence on the turnover rate of the catalytic
cycle, however, since the turnover-limiting step in most cases is
displacement of the ketone/alcohol ligand by H2.
Substitution of one CO ligand by a phosphine dramatically
increases the kinetic hydricity (eqn. (12)): kHϪ = 4.3 × 105 MϪ1 sϪ1
for trans-Cp(CO)2(PCy3)MoH, kHϪ = 5.7 × 105 MϪ1 sϪ1 for trans-
Cp(CO)2(PPh3)MoH, and kHϪ = 4.6 × 106 MϪ1 sϪ1 for trans-
Cp(CO)2(PMe3)MoH.10 The electronic effect is as expected, with
the electron-donating phosphine ligands making the metal
hydride more hydridic. The observation that electronic effects
predominate over steric influences certainly favors using these
metalhydridesashydridedonors.Eventhemetalhydridewiththe
most bulky phosphine, Cp(CO)2(PCy3)MoH has a rate constant
for hydride transfer that is about three orders of magnitude faster
than that for Cp(CO)3MoH. Our measurements for hydride
transfer used Ph3Cϩ as the hydride acceptor, whereas the hydro-
genationsinthecatalyticreactionsinvolveaprotonatedketoneas
the hydride acceptor. The rate constant for hydride transfer from
Cp(CO)2(PPh3)MoH to protonated acetone was reported to be
1.2 × 104 MϪ1 sϪ1 in CH3CN.39 It was concluded that hydride
transfer from Cp(CO)2(PPh3)MoH to protonated acetone pro-
ceeds by a single-step hydride transfer39 rather than the alterna-
tive mechanism involving an electron transfer followed by hydro-
gen atom transfer. Evidence for a single-step hydride transfer was
While homogeneous catalysis of ketone hydrogenation is
dominated by the use of ruthenium and rhodium complexes,
there is some precedent for the use of chromium, molybdenum
and tungsten. Darensbourg and co-workers extensively devel-
oped the chemistry of anionic metal carbonyl hydrides such as
Ϫ
HCr(CO)5 , with an emphasis on hydride transfer reactivity of
these complexes.45 Stoichiometric hydrogenations of aldehydes
and ketones were accomplished46 using acids together with
metal hydrides such as HM(CO)5Ϫ (M = Cr, W) and phosphite-
substituted derivatives like HW(CO)4[P(OMe)3]Ϫ. In the case of
reactive aldehydes like benzaldehyde, insertion of the aldehyde
into the M–H bond to give a metal alkoxide was observed, and
the alcohol was released by addition of acetic acid. In contrast,
they found that hydrogenation of ketones required acid add-
ition first to activate the carbonyl towards nucleophilic attack
by the metal hydride. Catalytic hydrogenation of cyclohex-
anone was accomplished using (CO)5W(OAc)Ϫ at 125 ЊC and
600 psi H2.47 The mechanism proposed for these hydrogenations
involved conversion of (CO)5W(OAc)Ϫ to the hydride (CO)5-
WHϪ. Then insertion of the aldehyde or ketone into the M–H
bond would produce an anionic alkoxide that could be cleaved
by HOAc to produce the alcohol product and regenerate the
(CO)5W(OAc)Ϫ. A related catalyst for hydrogenation of aceto-
phenone was produced by reaction of M(CO)6 (M = Cr, Mo,
W) with NaOMe in methanol.48 Reaction conditions of 70 ЊC
and 1450 psi H2 pressure were used for the Mo case, with higher
temperatures (100–120 ЊC) being required for catalysts starting
from Cr(CO)6 or W(CO)6. In this case also the intermediacy
also presented10 for hydride transfers to Ph3CϩBF4 , providing
Ϫ
another similarity between these two types of hydride transfer
reactions. Consideration of all of these hydride transfer kinetics
indicates that the hydride transfer reactions producing alcohol in
our reactions (Scheme 1) are sufficiently fast to overcome the
unfavorable thermodynamics of the proton transfer step.
Although no molybdenum dihydrides have been directly
observed, the involvement of molybdenum ketone complexes
in the heterolytic cleavage of H2 has been demonstrated. Add-
ition of H2 (<4 atm) to a CD2Cl2 solution containing the
Ϫ
ketone complex [CpMo(CO) (PPh )(Et C᎐O)]ϩBArЈ
and
᎐
2
3
2
4
the hindered amine base 2,6-di-tert-butyl-4-methylpyridine
gave the neutral molybdenum hydride in 81% yield in 5 minutes
(eqn. (13)). This could proceed through formation of the
(13)
Ϫ
of HM(CO)5 and anionic metal alkoxide intermediates were
proposed.
Brunet and co-workers reported recent developments in the
Ϫ
use of the anionic metal carbonyl hydride HCr(CO)5 as a
transfer hydrogenation catalyst.49 In their experiments, 20%
J. Chem. Soc., Dalton Trans., 2002, 759–770
765

View Article Online
KHCr(CO)5 was used with HCO2H–NEt3 to catalyze the
(15)
hydrogenation of ketones at room temperature. The metal
Ϫ
hydride HCr(CO)5 is regenerated in this reaction through
decarboxylation of [HCO2Cr(CO)5]Ϫ.
Noyori and co-workers developed a family of remarkably
reactive ruthenium catalysts for the asymmetric transfer hydro-
genation of ketones under mild conditions.5 These transfer
hydrogenations use isopropanol or formic acid as the hydrogen
source, and provide very high turnover numbers of the catalyst
and enantiomeric purity of the alcohol products. The mechan-
ism of these metal–ligand bifunctional catalysts has been stud-
ied experimentally and by theoretical studies.50 Eqn. (14) shows
(14)
by Casey and co-workers provide evidence for the concerted
delivery of Hϩ from the OH site from the hydroxycyclopenta-
dienyl ligand, and HϪ from the RuH.55 Eqn. (16) depicts the
mechanism for this polar hydrogenation of an aldehyde.
(16)
the preparation of a ruthenium hydride complex through reac-
tion with isopropanol; Scheme 2 shows the computed transition
Scheme 2
The hydrogenations from the different ruthenium systems
above are all thought to proceed through concerted delivery of
a proton (from an NH or OH site) and a hydride (from a metal
hydride). This concerted mechanism contrasts with our ionic
hydrogenation system, where sequential proton transfer and
hydride transfer steps are proposed. All of these cases are dis-
tinct from traditional mechanisms where coordination of the
ketone or other substrate is required. These new types of cat-
alysts are generally coordinatively saturated, 18-electron com-
plexes that have no readily accessible vacant site for binding of
a ketone.
state for the hydrogenation of an aldehyde. A key issue in the
hydrogenation was found to be hydrogen bonding of the oxygen
of the substrate to the amine NH bond; this activation of
the carbonyl substrate, in conjunction with the hydricity of the
Ru–H bond, results in the smooth hydrogenation of the C᎐O
᎐
bond. A key mechanistic issue relevant to our studies on Mo
and W complexes is that the ruthenium complex is an 18-
electron complex that has no vacant coordination site. This
case, like our ionic hydrogenations, does not require binding of
the ketone to the metal but requires a metal complex capable
of delivering hydrogen to the unsaturated substrate. Andersson
and co-workers reported highly enantioselective transfer hydro-
genation catalysts based on ruthenium(arene)(amino alcohol)
complexes,51 and a computational study supported a six-
membered transition state, with concerted transfer of HϪ from
the metal and Hϩ from the NH bond.52
The mechanism of stepwise proton and hydride transfers
operative in our catalytic ketone hydrogenations is similar to
the mechanism proposed by Magee and Norton for hydrogen-
ation of C᎐N double bonds using a ruthenium catalyst (eqn.
᎐
(17)).56 They found that enantioface selective catalytic hydro-
genation of the C᎐N bonds of iminium cations could be
᎐
accomplished using a ruthenium hydride with chiral diphos-
phine ligands. In our catalysts, displacement of the ketone by
H2 was usually the slow step of the catalytic cycle. In contrast,
hydride transfer was the turnover-limiting and enantioselectivity-
determining step in eqn. (17).
Asymmetric catalytic hydrogenations have also been accom-
plished starting from chiral RuCl2(diphosphine)(1,2-diamine)
complexes as catalyst precursors, with activation being achieved
by addition of a strong base like KOtBu.6 These catalysts also
appear to involve a mechanism similar to that described above
for the (arene)ruthenium complexes. Morris and co-workers
also discovered ruthenium(diphosphine)(1,2-diamine) com-
plexes that exhibit exceptionally high reactivity for ketone
hydrogenations using H2.53
(17)
Shvo and co-workers reported a novel type of ruthenium
complex that was shown to catalyze hydrogenation of ketones,
aldehydes, alkenes and alkynes.54 The hydrogenation of ketones
was carried out at 145 ЊC at 500 psi H2. The bimetallic catalyst
precursor is bridged both by a bridging hydride as well as
through the phenyl-substituted cyclopentadienone ligands.
Under hydrogen pressure, a mononuclear ruthenium active
species is formed (eqn. (15)). Recent mechanistic experiments
We also discovered a ruthenium catalyst that carries out the
ionic hydrogenation of ketones.57 The bimetallic ruthenium
complex with a bridging hydride, {[CpRu(CO)2]2(µ-H)}ϩOTfϪ,
catalyzes the selective deoxygenation of 1,2-propanediol to n-
propanol in the presence of added acid. The mechanism of the
deoxygenation of diols involves generation of an aldehyde as an
766
J. Chem. Soc., Dalton Trans., 2002, 759–770

View Article Online
observable intermediate, and this aldehyde is hydrogenated
under the reaction conditions. Ketones were hydrogenated
by {[CpRu(CO)2]2(µ-H)}ϩOTfϪ in the absence of added acid.
These hydrogenations were proposed to proceed through
the highly acidic dihydrogen complex [Cp*Ru(CO)2(η2-H2)]ϩ
generated from {[CpRu(CO)2]2(µ-H)}ϩOTfϪ under the reaction
conditions.
0.40 (t, JHH = 7.4 Hz, 3H, CH3), 0.10 (dd, J = 8.2 Hz, 2.9 Hz,
1H, OH). 31P NMR (CD2Cl2): δ 41.0 (s, 1JPW = 276 Hz). 13C{1H}
2
NMR (CD2Cl2): δ 241.6 (d, JPC = 4 Hz, CO trans to P); 239.8
(d, 2JPC = 21 Hz, CO cis to P), 162.2 (1 : 1 : 1 : 1 quartet, JCB
=
Ϫ
Ϫ
50 Hz ipso-C of BArЈ4 ), 135.2 (s, ortho-C of BArЈ4 ), 133.9
(d, 2JPC = 11 Hz, ortho-C of PPh3), 133.4 (d, 4JPC = 2 Hz, para-C
3
of PPh3), 131.0 (d, JPC = 10 Hz, meta-C of PPh3), 129.3 (qm,
Ϫ
2JCF = 31 Hz, meta-C of BArЈ4 ), 127.8 (d, 1JPC = 49 Hz, ipso-C
1
3
of PPh3), 125.0 (q, JCF = 272 Hz, CF3), 117.9 (septet, JCF
=
Experimental
Ϫ
3
272 Hz, para-C of BArЈ4 ), 94.8 (s, Cp), 93.6 (d, JCP = 2 Hz,
CHOH), 26.6 (s, CH2), 26.6 (s, CH2), 8.5 (s, CH3), 8.1 (s, CH3).
IR (CH2Cl2): ν(OH) 3433 (w, br); ν(CO) 1974 (s), 1894 (s) cmϪ1.
Found: C, 49.04; H, 2.99. C62H44BF24O3PW requires C, 49.04;
H, 2.92%.
General
All manipulations were carried out under an atmosphere of
argon using Schlenk or vacuum-line techniques, or in a Vacuum
Atmospheres drybox. 1H NMR chemical shifts were referenced
to the residual proton peak of CD2Cl2 at δ 5.32. Elemental
Analyses were carried out by Schwarzkopf Microanalytical
Laboratory (Woodside, NY). NMR spectra were recorded
on a Bruker AM-300 spectrometer (300 MHz for 1H). IR
spectra were recorded on a Mattson Polaris FT-IR. Et2O and
hexane were distilled from Na–benzophenone, and CH2Cl2 was
distilled from P2O5. Cp(CO)2(PPh3)WH,58,59 Cp(CO)2(PPh3)-
MoH,58Cp(CO)2(PCy3)WH,31Cp(CO)2(PCy3)MoH,10Cp(CO)2-
(PMe3)WH,59 Cp(CO)2(PMe3)MoH,59 [CpW(CO)2(PMe3)(H)2]ϩ-
BArЈ4Ϫ,31 Ph3CϩBArЈ4Ϫ,60 [ArЈ = 3,5-bis(trifluoromethyl)phenyl]
؊
trans-[CpW(CO)2(PPh3)(Et2CHOH)]؉BArЈ4
.
cis-[CpW-
(CO) (PPh )(Et C᎐O)]ϩBArЈ (48.4 mg, 0.031 mmol) was dis-
Ϫ
᎐
2
3
2
4
solved in CD2Cl2 (1.1 mL) in a glass vial, and Et2CO (16 µL,
0.15 mmol) was added. The solution was placed in a Parr
high pressure reactor vessel and H2 (65 atm) was added. After
17 hours at 22 ЊC, the pressure was released, and an NMR
spectrum showed 92% conversion to trans-[CpW(CO)2-
Ϫ
(PPh3)(Et2CHOH)]ϩBArЈ4 , along with 8% of [CpW(CO)2-
(PPh )(Et C᎐O)]ϩ remaining. 1H (CD Cl ): δ 7.73 (br, 8H, o-H),
᎐
3
2
2
2
7.57 (br, 4H, p-H), 7.1–7.2, 7.43–7.56 (m, 15 H, PPh3), 5.59 (d,
J = 8.4 Hz, 1H, OH), 5.39 (d, JPH = 2.4 Hz, 5H, Cp), 3.37 (m,
CH, 1H), 1.35–1.55 (m, CH2, 4H), 0.93 (t, JHH = 7.5 Hz, 6H,
CH3). 31P NMR (CD2Cl2): δ 30.6 (s, 1JPW = 197 Hz).
and [H(Et2O)2]ϩBArЈ4 61 were prepared by previously reported
Ϫ
routes.
Syntheses
Synthesis of [CpMo(CO) (PPh )(Et C᎐O)]؉BArЈ ؊. Ph3Cϩ-
Synthesis of cis-[CpW(CO) (PPh )(Et C᎐O)]؉BArЈ ؊. Ph3Cϩ-
᎐
2
3
2
4
᎐
2
3
2
4
Ϫ
BArЈ4Ϫ (288.5 mg, 0.26 mmol) and Et2CO (100 µL, 0.95 mmol)
were combined in a flask. CH2Cl2 (8 mL) was added, giving a
yellow solution. Addition of CpW(CO)2(PPh3)H (148.5 mg,
0.26 mmol) caused the solution to become red-orange. After
5 minutes of stirring, hexane (30 mL) was slowly added until
the solution became cloudy and a reddish-orange oil precip-
itated. Upon standing at room temperature, the oil solidified
into small microcrystals within a few minutes. This orange solid
was collected by filtration, and washed with hexane (3 × 5 mL).
Yield 354 mg (0.23 mmol, 88%). 1H NMR (CD2Cl2: δ 7.73 (br,
8H, o-H), 7.57 (br, 4H, p-H), 7.25–7.31, 7.50–7.54 (m, 15H,
PPh3), 5.76 (s, 5H, Cp), 2.20 (dq, JHH = 18.2 Hz, 7.3 Hz, 2H,
BArЈ4 (224.0 mg, 0.202 mmol) and CpMo(CO)2(PPh3)H
(99.2 mg, 0.206 mmol) were combined in a 50 mL flask. Et2CO
(100 µL, 0.95 mmol) was added, then CH2Cl2 (5 mL) was
added, generating a dark red-orange solution. After 5 minutes
of vigorous stirring, Et2O (20 mL) was slowly added, but no
precipitation ensued. Addition of hexane (20 mL) caused a
dark red solid precipitate to form. The product was collected by
filtration and washed with hexane. Yield 177.8 mg (0.124 mmol,
61%) of a cis–trans mixture. 1H NMR (CD2Cl2) of trans isomer:
δ 7.73 (br, 8H, o-H), 7.56 (br, 4H, p-H), 7.25–7.32, 7.50–7.55
(m, 15H, PPh3), 5.40 (d, JPH = 2.5 Hz, 5H, Cp), 2.50 (q, JHH
=
1
7.3 Hz, 4H, CH2), 1.07 (t, JHH = 7.3 Hz, 6H, CH3). H NMR
(CD2Cl2) of cis isomer: δ 7.73 (br, 8H, o-H), 7.56 (br, 4H, p-H),
7.25–7.32, 7.50–7.55 (m, 15H, PPh3), 5.61 (s, 5H, Cp), 2.09
(dq, JHH = 18.2 Hz, 7.3 Hz, 2H, CH2), 1.62 (dq, JHH = 18.2 Hz,
7.3 Hz, 2H, CH2), 0.73 (t, JHH = 7.3 Hz, 6H, CH3). 31P NMR
(CD2Cl2): δ 59.6 (s, trans isomer); 53.9 (s, cis isomer). 13C{1H}
NMR (CD2Cl2) of cis isomer: δ 248.5 (d, 2JPC = 29 Hz, CO cis to
CH2), 1.59 (dq, JHH = 18.2 Hz, 7.3 Hz, 2H, CH2), 0.73 (t, JHH
=
7.3 Hz, 6H, CH3). 31P NMR (CD2Cl2): δ 37.1 (s, 1JPW = 277 Hz).
13C{1H} NMR (CD2Cl2): δ 240.1 (d, 2JPC = 6 Hz, CO trans to P);
2
239.7 (d, JPC = 21 Hz, CO cis to P), 236.2 (s, Et C᎐O), 162.2
᎐
2
Ϫ
(1 : 1 : 1 : 1 quartet, JCB = 50 Hz, ipso-C of BArЈ4 ), 135.2
(s, ortho-C of BArЈ4 ), 134.1 (d, 2JPC = 11 Hz, ortho-C of PPh3),
Ϫ
2
4
3
P); 247.1 (d, JPC = 3 Hz, CO trans to P), 236.2 (s, Et C᎐O),
132.6 (d, JPC = 2 Hz, para-C of PPh3), 130.0 (d, JPC = 10 Hz,
meta-C of PPh3), 129.6 (d, 1JPC = 49 Hz, ipso-C of PPh3), 129.3
᎐
2
Ϫ
162.2 (1 : 1 : 1 : 1 quartet, JCB = 50 Hz, ipso-C of BArЈ4 ), 135.2
(s, ortho-C of BArЈ4 ), 133.8 (d, 2JPC = 11 Hz, ortho-C of PPh3),
Ϫ
(q, JCF = 29 Hz, meta-C of BArЈ4ϪЈ), 125.0 (q, JCF = 272 Hz,
2
1
4
1
Ϫ
132.4 (d, JPC = 2 Hz, para-C of PPh3), 130.0 (d, JPC = 48 Hz,
CF3), 117.9 (br s, para-C of BArЈ4 ), 95.9 (s, Cp), 37.0 (s, CH2),
ipso-C of PPh3), 130.0 (d, 3JPC = 10 Hz, meta-C of PPh3), 129.3
8.9 (s, CH ). IR (CH Cl ): ν(CO) 1974 (vs), 1895 (s); ν(C᎐O)
᎐
3
2
2
2
Ϫ
1632(w) cmϪ1. Found: C, 48.95; H, 2.98. C62H42BF24O3PW
requires C, 49.10; H, 2.79%.
(qm, JCF = 31 Hz, meta-C of BArЈ4 ), 125.0 (q, JCF = 272 Hz,
CF3), 117.9 (septet, 3JCF = 4 Hz, para-C of BArЈ4 ), 97.5 (s, Cp),
Ϫ
36.8 (s, CH2), 8.9 (s, CH3). IR (CH2Cl2): ν(CO) 1992 (s), 1910
Synthesis of cis-[CpW(CO)2(PPh3)(Et2CHOH)]؉BArЈ4؊. Ph3-
(vs); ν(C᎐O) 1643 (w); 1610 (vw) cmϪ1. Found: C, 51.83; H,
᎐
Ϫ
CϩBArЈ4 (249.0 mg, 0.225 mmol) and CpW(CO)2(PPh3)H
2.95. C62H42BF24MoO3P requires C, 52.12; H, 2.96%.
(128.2 mg, 0.226 mmol) were combined in a 50 mL flask.
CH2Cl2 (8 mL) was added, generating a dark red-orange sol-
ution. Et2CHOH (49 µL, 0.46 mmol) was added, and after
5 minutes of vigorous stirring, Et2O (20 mL) was slowly added,
followed by cyclohexane (20 mL), but no precipitate formed.
Upon addition of heptane (5 mL), the reaction mixture became
cloudy, and a red-orange precipitate began to appear. Add-
itional heptane (5 mL) was added, and a red-orange solid was
؊
Synthesis of cis-[CpMo(CO)2(PPh3)(Et2CHOH)]؉BArЈ4
.
Ph3CϩBArЈ4Ϫ (208.1 mg, 0.188 mmol) and CpMo(CO)2(PPh3)H
(96.2 mg, 0.200 mmol) were combined in a 50 mL flask. CH2Cl2
(5 mL) was added, producing a dark red-orange solution.
Immediately afterwards, Et2CHOH (100 µL, 0.93 mmol) was
added. After 5 minutes of vigorous stirring, Et2O (5 mL) was
added, followed by hexane (10 mL). Addition of more hexane
(20 mL) resulted in precipitation of a dark red-orange solid.
The product was collected by filtration and washed with hex-
1
collected by filtration. Yield 304.4 mg (0.20 mmol, 89%). H
NMR (CD2Cl2): δ 7.74 (br, 8H, o-H), 7.56 (br, 4H, p-H), 7.27–
7.34, 7.56–7.64 (m, PPh3, 15H), 5.74 (s, 5H, Cp), 3.37 (m, CH,
1H), 0.83–1.20 (m, 4H, CH2), 0.58 (t, JHH = 7.4 Hz, 3H, CH3),
1
ane. Yield: 211.7 mg (0.148 mmol, 79%). H (CD2Cl2): δ 7.73
(br, 8H, o-H), 7.56 (br, 4H, p-H), 7.26–7.33, 7.59–7.67 (m, 15H,
J. Chem. Soc., Dalton Trans., 2002, 759–770
767

View Article Online
PPh3), 5.58 (s, 5H, Cp), 3.17 (m, CH, 1H), 0.88–1.20 (m, 4H,
CH2), 0.57 (t, JHH = 7.4 Hz, 3H, CH3), 0.40 (t, JHH = 7.4 Hz, 3H,
CH3), Ϫ0.62 (dd, J = 8.2 Hz, 3.3 Hz, 1H, OH). 31P NMR
Synthesis of [CpW(CO) (PMe )(Et C᎐O)]؉BArЈ ؊. Ph3Cϩ-
᎐
2
3
2
4
Ϫ
BArЈ4 (152.8 mg, 0.138 mmol) and CpW(CO)2(PMe3)H
(57.2 mg, 0.150 mmol) were combined in a 50 mL flask. Et2CO
(100 µL, 0.95 mmol) was syringed onto the solids, followed by
addition of CH2Cl2 (5 mL), which generated a dark orange
solution. Hexane (15 mL) was added, and an oil precipitated
out of solution. After standing for two hours, no solids were
observed, so more hexane (5 mL) was added, and the reaction
mixture was stirred vigorously. After two additional hours,
some solid had formed. The supernatant was decanted, and the
precipitate was triturated with pentane (20 mL), collected by
filtration, and washed with pentane, giving an orange micro-
crystalline solid (123 mg, 0.0924 mmol, 67%). The cis : trans
(CD2Cl2): δ 58.3 (s). 13C{1H} NMR (CD2Cl2): δ 248.6 (d, 2JPC
=
30 Hz, CO cis to P); 248.2 (s, CO trans to P), 162.2 (1 : 1 : 1 : 1
Ϫ
quartet, JCB = 50 Hz, ipso-C of BArЈ4 ), 135.2 (s, ortho-C of
Ϫ
2
BArЈ4 ), 133.7 (d, JPC = 11 Hz, ortho-C of PPh3), 133.2 (d,
4JPC = 2 Hz, para-C of PPh3), 130.4 (d, JPC = 10 Hz, meta-C
3
1
of PPh3), 130.2 (d, JPC = 42 Hz, ipso-C of PPh3), 129.3 (qm,
2JCF = 34 Hz, meta-C of BArЈ4 ), 125.0 (q, 1JCF = 272 Hz, CF3),
Ϫ
117.9 (septet, 3JCF = 4 Hz para-C of BArЈ4 ), 96.6 (s, Cp), 90.9
Ϫ
(s, CHOH), 26.8 (s, CH2), 8.6 (s, CH3), 8.2 (s, CH3). IR
(CH2Cl2): ν(OH) 3459 (w, br); ν(CO) 1986 (vs), 1912 (s) cmϪ1.
Found: C, 52.18; H, 3.06. C62H44BF24MoO3P requires C, 52.05;
H, 3.10%.
1
ratio was about 1.6 : 1. H NMR (CD2Cl2) of trans isomer:
δ 7.73 (br, 8H, o-H), 7.58 (br, 4H, p-H), 5.56 (d, JPH = 2.4 Hz,
5H, Cp), 2.55 (q, JHH = 7.4 Hz, 4H, CH2), 1.60 (d, JPH = 9.7 Hz,
9 H, P(CH3)3), 1.07 (t, JHH = 7.4 Hz, 6H, CH3). H NMR
1
trans-[CpMo(CO)2(PPh3)(Et2CHOH)]؉BArЈ4؊
.
The trans
isomer of this alcohol complex was observed during hydrogen-
(CD2Cl2) of cis isomer: δ 7.73 (br, 8H, o-H), 7.58 (br, 4H, p-H),
5.82 (s, 5H, Cp), 2.55 (q, JHH = 7.4 Hz, 4H, CH2, overlapped
1
ations catalyzed by [CpMo(CO) (PPh )(Et C᎐O)]ϩ. H (CD -
᎐
2
3
2
2
Cl2): δ 7.74 (br, 8H, o-H), 7.56 (br, 4H, p-H), 7.18–7.28, 7.45–
7.55 (m, 15 H, PPh3), 5.31 (d, JPH = 2.4 Hz, 5H, Cp), 4.96 (d, J =
8.6 Hz, 1H, OH), 3.18 (m, CH, 1H), 1.40–1.55 (m, CH2, 4H),
0.88 (t, JHH = 7.5 Hz, 6H, CH3). 31P NMR (CD2Cl2): δ 61.1 (s).
with same resonance for the trans isomer), 1.71 (d, JPH =
10.0 Hz, 9 H, P(CH3)3), 1.10 (t, JHH = 7.4 Hz, 6H, CH3).
1
31P NMR (CD2Cl2): δ Ϫ15.6 (s, JPW = 192 Hz, trans isomer);
Ϫ17.2 (s, 1JPW = 262 Hz, cis isomer). IR (CH2Cl2): ν(CO) 1971
(s), 1884 (s); ν(C᎐O) 1632 (w); 1611 (w) cmϪ1. Found: C, 42.12;
᎐
Synthesis of trans-[CpW(CO)2(PMe3)(Et2CHOH)]؉BArЈ4
.
H, 2.87. C47H36BF24WO3P requires C, 42.43; H, 2.73%.
؊
Et2CO (11 µL, 0.10 mmol) was added to a solution of
[CpW(CO)2(PMe3)(H)2]ϩBArЈ4 (35.3 mg, 0.0283 mmol) in
Ϫ
Synthesis of [{CpMo(CO)2(PMe3)}2(ꢁ-H)]؉BArЈ4؊. Ph3Cϩ-
BArЈ4 (332.0 mg, 0.300 mmol) and CpMo(CO)2(PMe3)H
Ϫ
CD2Cl2 (0.72 mL). The solution turned light orange, and con-
version to trans-[CpW(CO)2(PMe3)(Et2CHOH)]ϩ was about
90% complete in 1 hour. 1H (CD2Cl2): δ 7.73 (br, 8H, o-H), 7.57
(br, 4H, p-H), 5.46 (d, JPH = 2.6 Hz, 5H, Cp), 5.36 (d, J = 8.4 Hz,
1H, OH), 3.28 (m, CH, 1H), 1.67 (d, JPH = 9.9 Hz, 9 H,
P(CH3)3), 1.35–1.51 (m, CH2, 4H), 0.87 (t, JHH = 7.5 Hz, 6H,
CH3). 31P NMR (CD2Cl2): δ Ϫ15.1 (s, 1JPW = 179 Hz).
(91.5 mg, 0.311 mmol) were combined in a 50 mL flask, and
CH2Cl2 (5 mL) was added. Immediately afterwards, Et2CHOH
(100 µL, 0.93 mmol) was added. It appeared that much of the
Ph3CϩBArЈ4 remained unreacted, so additional CpMo(CO)2-
Ϫ
(PMe3)H (25 mg, 0.085 mmol) was added, which caused the
color of the solution to darken. The dark maroon precipitate
that formed upon the addition of hexane was collected by fil-
tration. Yield: 142.0 mg (0.098 mmol, 33%). 1H (CD2Cl2): δ 7.72
Synthesis of [CpMo(CO) (PMe )(Et C᎐O)]؉BArЈ ؊
. Ph3-
᎐
2
3
2
4
CϩBArЈ4 (360.5 mg, 0.326 mmol) and CpMo(CO)2(PMe3)H
(97.4 mg, 0.331 mmol) were combined in a 50 mL flask. Et2CO
(100 µL, 0.95 mmol) was added, then CH2Cl2 (5 mL) was
added, generating a dark red-orange solution. After brief
stirring, hexane (20 mL) was slowly added. The product pre-
cipitated out of solution as an oil. Decanting the supernatant
and triturating with pentane (30 mL) resulted in the conversion
of the oil to a solid. The product was collected by filtration and
washed with pentane. Yield 337.4 mg (0.272 mmol, 83%). The
cis : trans ratio was 0.7 : 1. 1H NMR (CD2Cl2) of trans isomer:
δ 7.72 (br, 8H, o-H), 7.57 (br, 4H, p-H), 5.47 (d, JPH = 2.5 Hz,
5H, Cp), 2.46 (q, JHH = 7.4 Hz, 4H, CH2), 1.61 (d, JPH = 9.9 Hz,
Ϫ
(br, 8H, o-H), 7.56 (br, 4H, p-H), 5.22 (d, JPH = 1.8 Hz, 10H,
2
Cp), 1.68 (d, JPH = 9.9 Hz, 18 H, P(CH3)3), Ϫ19.92 (t, JPH
=
11 Hz, 1 H, hydride). IR (CH2Cl2): ν(CO) 1981 (m), 1955 (m),
1896 (s) cmϪ1. Found: C, 42.84; H, 2.88. C52H41BF24Mo2O4P2
requires C, 43.06; H, 2.85%.
NMR tube kinetics experiments
A standard solution of Et2CO in CD2Cl2 was prepared by
dissolving bibenzyl (136.7 mg, 0.750 mmol; internal standard
1
for H NMR integration) and Et2CO (792 µL, 7.50 mmol)
in a 25 mL volumetric flask. The flask was filled to the mark
1
9 H, P(CH3)3), 1.07 (t, JHH = 7.4 Hz, 6H, CH3). H NMR
with CD2Cl2, giving a 300 mM solution of Et2CO. For
[CpW(CO) (PPh )(Et C᎐O)]ϩBArЈ , [CpW(CO)2(PPh3)(Et2-
Ϫ
(CD2Cl2) of cis isomer: δ 7.72 (br, 8H, o-H), 7.57 (br, 4H, p-H),
5.62 (s, 5H, Cp), 2.46 (q, JHH = 7.4 Hz, 4H, CH2, overlapped
with same resonance for the trans isomer), 1.54 (d, JPH = 9.9 Hz,
9 H, P(CH3)3), 1.09 (t, JHH = 7.4 Hz, 6H, CH3). 31P NMR
(CD2Cl2): δ 20.8 (s, trans isomer); 9.7 (s, cis isomer). 13C{1H}
NMR (CD2Cl2) of cis–trans mixture, in the presence of ≈
᎐
2
3
Ϫ
2
4
CHOH)]ϩBArЈ , [CpMo(CO) (PPh )(Et C᎐O)]ϩBArЈ , [Cp-
Ϫ
᎐
4
2
3
2
4
W(CO) (PMe )(Et C᎐O)]ϩBArЈ , and [CpMo(CO)2(PMe3)-
Ϫ
᎐
2
3
2
4
Ϫ
(Et C᎐O)]ϩBArЈ the following procedure was used: 0.021
᎐
2
4
mmol of the metal complex was weighed and transferred into a
5 mm NMR tube equipped with a Young valve. Then 0.7 mL of
the standard 300 mM solution of Et2CO in CD2Cl2 was added,
giving a catalyst concentration of 30 mM. (If 0.8 mL is added,
then the concentration of the metal will drop to 26 mM, such
that 11.4 equivalents of ketone will be present, rather than the
intended 10 equivalents.) In all cases the solid dissolved and
the solution turned red-orange. The tube was then frozen in
liquid nitrogen, evacuated on a high vacuum line, then filled
with 1 atm H2. The valve was closed, and the tube was warmed
to room temperature. Using this procedure, the pressure of H2
after the solution was warmed to room temperature should be
<4 atm (298/77 = 3.9). The pressure was maintained by period-
ically refilling with H2 as needed. All reactions were carried out
at room temperature (23 ЊC). In an attempt to promote better
diffusion of the H2 gas into the solvent, the tubes were spun
2 equiv. free Et C᎐O: δ 248.3 (d, J = 32 Hz, CO cis to P; cis
᎐
2
PC
isomer); 247.1 (d, JPC = 4 Hz, CO trans to P; cis isomer), 241.0
(d, JPC = 27 Hz, CO; trans isomer), 238.0 (d, JPC = 2 Hz, Et C᎐
᎐
2
O, cis isomer), 235.7 (d, JPC = 2 Hz, Et C᎐O, trans isomer),
᎐
161.7 (1 : 1 : 1 : 1 quartet, ipso-C, JCB2 = 50 Hz), 135.3 (s,
2
ortho-C), 129.3 (q, meta-C, JCF = 28 Hz), 125.1 (q, JCF
=
272 Hz, CF3), 118.0 (s, para-C), 96.4 (Cp), 95.8 (Cp), 37.4
(CH2), 37.1 (CH2), 20.1 (d, JPC = 36 Hz, P(CH3)3), 17.4 (d, JPC
=
30 Hz, P(CH3)3), 9.2 (CH3), 8.9 (CH3). IR (CH2Cl2): ν(CO) 1986
(vs), 1900 (vs); ν(C᎐O) 1646 (w); 1611 (vw) cmϪ1. Found: C,
᎐
45.49; H, 2.80. C47H36BF24MoO3P requires C, 45.43; H, 2.92%.
When an CD2Cl2 solution of this cis–trans mixture was allowed
to stand at room temperature for 2 days, the cis : trans ratio
increased to 1.1 : 1, but some decomposition of the alcohol
complex was also observed.
slowly end-over-end using
a mechanical stirring motor
768
J. Chem. Soc., Dalton Trans., 2002, 759–770

View Article Online
ϩ
mounted sideways. The catalytic reactions of the PCy3 com-
upon partial protonation of PPh3 (31P at δ Ϫ5.0) to give HPPh3
plexes were prepared by in situ reactions of Ph3CϩBArЈ4 with
(31P at δ 8.1), and for protonation of PMe3 (31P at δ Ϫ61) to give
HPMe3ϩ (31P at δ Ϫ4.2, 1JPH = 487 Hz, 2JPH = 15 Hz)
Ϫ
CpW(CO)2(PCy3)H or CpMo(CO)2(PCy3)H (see below); other-
wise they were treated in the same manner as those described
above.
Results shown in Fig. 2 include formation of the alcohol
Et2CHOH and its subsequent conversion to the ether (Et2-
Acknowledgements
We thank the U.S. Department of Energy, Office of Science,
Laboratory Technology Research Program, and the Division
of Chemical Sciences, Office of Basic Energy Sciences, for
support. This research was carried out at Brookhaven
National Laboratory under contract DE-AC02-98CH10886
with the U.S. Department of Energy. We thank DuPont
Central Research for additional support of this work through
a CRADA grant. We gratefully acknowledge Dr Paul Fagan
and Dr Elisabeth Hauptman (DuPont) for many helpful
discussions.
CH)2O. Amounts of these two products are given here. For
Ϫ
catalysis with [CpW(CO) (PPh )(Et C᎐O)]ϩBArЈ , [Et2-
᎐
2
3
2
4
CHOH] = 93 mM and [(Et2CH)2O] = 6 mM, at t = 24 days,
for a total of 3.9 turnovers. For catalysis with [CpW(CO)2-
Ϫ
(PPh3)(Et2CHOH)]ϩBArЈ4 , the concentration of alcohol
formed in the reaction (i.e., after subtracting the alcohol ligand
present at the start of the reaction) [Et2CHOH] = 95 mM and
[(Et2CH)2O] = 86 mM, at t = 25 days, for a total of 4.1
turnovers. For catalysis with [CpMo(CO) (PPh )(Et C᎐O)]ϩ-
᎐
2
3
2
Ϫ
BArЈ4 , [Et2CHOH] = 255 mM and [(Et2CH)2O] = 46 mM, at
t = 6 days, for a total of 11.6 turnovers. For catalysis with
Ϫ
[CpW(CO) (PMe )(Et C᎐O)]ϩBArЈ , [Et CHOH] = 124 mM
᎐
and [(Et2CH)2O] = 3 mM, at t = 424 days, for a total of 4.3
2
3
2
2
References
turnovers. For catalysis with [CpMo(CO) (PMe )(Et C᎐O)]ϩ-
᎐
2
3
2
1 R. R. Schrock and J. A. Osborn, Chem. Commun., 1970, 567.
2 Q. Jiang, Y. Jiang, D. Xiao, P. Cao and X. Zhang, Angew. Chem.,
Int. Ed., 1998, 37, 1100.
3 R. A. Grey, G. P. Pez and A. Wallo, J. Am. Chem. Soc., 1981, 103,
7536; R. A. Sánchez-Delgado, N. Valencia, R.-L. Márquez-Silva,
A. Andriollo and M. Medina, Inorg. Chem., 1986, 25, 1106.
4 P. A. Chaloner, M. A. Esteruelas, F. Joó and L. A. Oro, Homo-
geneous Hydrogenation, Kluwer Academic Publishers, Dordrecht,
1994.
Ϫ
BArЈ4 , [Et2CHOH] = 101 mM and [(Et2CH)2O] = 25 mM, at
t = 24 days, for a total of 5.0 turnovers.
Reaction of CpMo(CO)2(PCy3)H with Ph3C؉BArЈ4؊. 0.8 mL
of the standard 300 mM solution of Et2CO in CD2Cl2 was
added to an NMR tube containing Cp(CO)2(PCy3)MoH
(10.6 mg, 0.021 mmol) and Ph3CϩBArЈ4 (23.2 mg, 0.021
Ϫ
5 S. Hashiguchi, A. Fujii, J. Takehara, T. Ikariya and R. Noyori,
J. Am. Chem. Soc., 1995, 117, 7562; A. Fujii, S. Hashiguchi,
N. Uematsu, T. Ikariya and R. Noyori, J. Am. Chem. Soc., 1996,
118, 2521; K. Matsumura, S. Hashiguchi, T. Ikariya and R. Noyori,
J. Am. Chem. Soc., 1997, 119, 8738; K.-J. Haack, S. Hashiguchi,
A. Fujii, T. Ikariya and R. Noyori, Angew. Chem., Int. Ed. Engl.,
1997, 36, 285; R. Noyori and S. Hashiguchi, Acc. Chem. Res., 1997,
30, 97.
6 T. Ohkuma, H. Doucet, T. Pham, K. Mikami, T. Korenaga,
M. Terada and R. Noyori, J. Am. Chem. Soc., 1998, 120, 1086;
H. Doucet, T. Ohkuma, K. Murata, T. Yokozawa, M. Kozawa,
E. Katayama, A. F. England, T. Ikariya and R. Noyori, Angew.
Chem., Int. Ed., 1998, 37, 1703; T. Ohkuma, M. Koizumi,
H. Doucet, T. Pham, M. Kozawa, K. Murata, E. Katayama, T.
Yokozawa, T. Ikariya and R. Noyori, J. Am. Chem. Soc., 1998, 120,
13529; T. Ohkuma, D. Ishii, H. Takeno and R. Noyori, J. Am. Chem.
Soc., 2000, 122, 6510; R. Noyori and T. Ohkuma, Angew. Chem., Int.
Ed., 2001, 40, 40.
mmol), and H2 (<4 atm) was added as described above. The
hydride transfer proceeded cleanly, as evidenced by the form-
ation of Ph3CH (δ 5.56, s, CH ). A singlet Cp resonance at
δ 5.63 as observed in the 1H NMR and a singlet at δ 49.1 in the
31P NMR are assigned to the ketone complex [CpMo(CO)2-
(PCy )(Et C᎐O)]ϩ. Resonances for Et C᎐O were significantly
᎐
᎐
2
3
2
1
broadened in the H NMR, suggesting exchange of free and
1
bound ketone. These resonances overlap with H NMR reson-
ances due to the PCy3 ligand and were not definitely assigned.
Also observed during the catalytic reaction was a singlet Cp
resonance at δ 5.73 in the 1H NMR and a singlet at δ 48.4 in the
31P NMR, which are assigned to the alcohol complex [Cp-
Mo(CO)2(PCy3)(Et2CHOH)]ϩ. At t = 24 hours, [Et2CHOH] =
272 mM and [(Et2CH)2O] = 34 mM, for a total of 11.3
turnovers.
7 D. N. Kursanov, Z. N. Parnes and N. M. Loim, Synthesis, 1974, 633;
D. N. Kursanov, Z. N. Parnes, M. I. Kalinkin and N. M. Loim, Ionic
Hydrogenation and Related Reactions, Harwood Academic
Publishers, New York, 1985.
8 S. S. Kristjánsdóttir and J. R. Norton, in Transition Metal Hydrides,
ed. A. Dedieu, VCH, New York, 1992 ch. 9.
9 J. A. Labinger, in Transition Metal Hydrides, ed. A. Dedieu, VCH,
New York, 1992, ch. 10.
10 T.-Y. Cheng, B. S. Brunschwig and R. M. Bullock, J. Am. Chem.
Soc., 1998, 120, 13121.
11 T.-Y. Cheng and R. M. Bullock, J. Am. Chem. Soc., 1999, 121, 3150.
12 R. M. Bullock and B. J. Rappoli, J. Chem. Soc., Chem. Commun.,
1989, 1447; R. M. Bullock and J.-S. Song, J. Am. Chem. Soc., 1994,
116, 8602.
13 L. Luan, J.-S. Song and R. M. Bullock, J. Org. Chem., 1995, 60,
7170.
Reaction of CpW(CO)2(PCy3)H with Ph3C؉BArЈ4؊. 0.7 mL
of the standard 300 mM solution of Et2CO in CD2Cl2 was
added to an NMR tube containing Cp(CO)2(PCy3)WH
Ϫ
(12.4 mg, 0.021 mmol) and Ph3CϩBArЈ4 (23.2 mg, 0.021
mmol), and H2 (<4 atm) was added as described above. As in
the Mo analog above, the observation of Ph3CH indicated
a clean hydride transfer. The dihydride [CpW(CO)2(PCy3)-
Ϫ31
(H)2]ϩBArЈ4
was the predominant metal complex observed
throughout the hydrogenation. After 24 days, [Et2CHOH] =
130 mM and [(Et2CH)2O] = 17 mM, for a total of 5.4 turnovers.
ϩ
Only about 2% HPCy3 was observed at t = 8 days, indicating
less decomposition compared to the Mo example.
14 J.-S. Song, D. J. Szalda and R. M. Bullock, J. Am. Chem. Soc., 1996,
118, 11134.
؊
Independent synthesis and identification of HPR3؉BArЈ4
15 J.-S. Song, D. J. Szalda, R. M. Bullock, C. J. C. Lawrie, M. A.
Rodkin and J. R. Norton, Angew. Chem., Int. Ed. Engl., 1992, 31,
1233.
16 J.-S. Song, D. J. Szalda and R. M. Bullock, Organometallics, 2001,
20, 3337.
17 R. M. Bullock and M. H. Voges, J. Am. Chem. Soc., 2000, 122,
12594.
18 W. Beck and K. Sünkel, Chem. Rev., 1988, 88, 1405.
19 W. Beck and K. Schloter, Z. Naturforsch., Teil B, 1978, 33, 1214.
20 K. Sünkel, H. Ernst and W. Beck, Z. Naturforsch., Teil B, 1981, 36,
474.
The phosphonium cations observed as decomposition products
in the catalytic reactions were independently synthesized
Ϫ
and characterized. [H(Et2O)2]ϩBArЈ4 (5.3 mg, 0.0052 mmol,
0.5 equiv.) was added to PCy3 (3.1 mg, 0.011 mmol) in CD2Cl2
(0.6 mL). The 31P NMR resonance due to free PCy3 (δ 11.1)
broadened, and a broadened new resonance appeared at δ 33.6.
This is apparently due to proton transfer exchange between
ϩ
Ϫ
PCy3 and HPCy3 . Addition of 1 equiv. [Hϩ(Et2O)2]ϩBArЈ4
resulted in the disappearance of the resonance for PCy3, and
the resonance for HPCy3ϩ (δ 34.6, 1JPH = 440 Hz) was no longer
broadened. Analogous experiments showed line broadening
21 K. Sünkel, G. Urban and W. Beck, J. Organomet. Chem., 1985, 290,
231.
J. Chem. Soc., Dalton Trans., 2002, 759–770
769

View Article Online
22 These assignments, with the CO cis to the phosphine having a larger
coupling constant, are based on the assignments made by
Wrackmeyer and co-workers in their detailed multinuclear NMR
study of cis and trans isomers of CpW(CO)2(PMe3)H: B.
Wrackmeyer, H. G. Alt and H. E. Maisel, J. Organomet. Chem.,
1990, 399, 125.
23 Y.-H. Huang and J. A. Gladysz, J. Chem. Educ., 1988, 65, 298.
24 J. W. Faller and A. S. Anderson, J. Am. Chem. Soc., 1970, 92, 5852;
P. Kalck and R. Poilblanc, J. Organomet. Chem., 1969, 19, 115;
O. B. Ryan, M. Tilset and V. D. Parker, J. Am. Chem. Soc., 1990, 112,
2618.
25 B. D. Martin, K. E. Warner and J. R. Norton, J. Am. Chem. Soc.,
1986, 108, 33.
26 L. M. Venanzi, Coord. Chem. Rev., 1982, 43, 251.
27 C. Nataro and R. J. Angelici, Inorg. Chem., 1998, 37, 2975.
28 T.-Y. Cheng, D. J. Szalda and R. M. Bullock, Chem. Commun., 1999,
1629.
J. Am. Chem. Soc., 1994, 116, 11014; E. Peris, J. C. Lee, Jr. and
R. H. Crabtree, J. Chem. Soc., Chem. Commun., 1994, 2573;
J. Wessel, J. C. Lee, Jr., E. Peris, G. P. A. Yap, J. B. Fortin, J. S. Ricci,
G. Sini, A. Albinati, T. F. Koetzle, O. Eisenstein, A. L. Rheingold
and R. H. Crabtree, Angew. Chem., Int. Ed. Engl., 1995, 34, 2507;
E. Peris, J. C. Lee, Jr., J. R. Rambo, O. Eisenstein and R. H.
Crabtree, J. Am. Chem. Soc., 1995, 117, 3485.
43 D.-H. Lee, B. P. Patel, E. Clot, O. Eisenstein and R. H. Crabtree,
Chem. Commun., 1999, 297.
44 R. H. Crabtree, P. E. M. Siegbahn, O. Eisenstein, A. L. Rheingold
and T. F. Koetzle, Acc. Chem. Res., 1996, 29, 348; R. H. Crabtree,
J. Organomet. Chem., 1998, 577, 111.
45 M. Y. Darensbourg and C. E. Ash, Adv. Organomet. Chem., 1987,
27, 1.
46 P. L. Gaus, S. C. Kao, K. Youngdahl and M. Y. Darensbourg, J. Am.
Chem. Soc., 1985, 107, 2428.
47 P. A. Tooley, C. Ovalles, S. C. Kao, D. J. Darensbourg and M. Y.
Darensbourg, J. Am. Chem. Soc., 1986, 108, 5465.
48 L. Markó and Z. Nagy-Magos, J. Organomet. Chem., 1985, 285, 193.
49 J.-J. Brunet, R. Chauvin and P. Leglaye, Eur. J. Inorg. Chem., 1999,
713; J.-J. Brunet, Eur. J. Inorg. Chem, 2000, 1377.
29 H. J. Wasserman, G. J. Kubas and R. R. Ryan, J. Am. Chem. Soc.,
1986, 108, 2294.
30 A. A. Gonzalez, K. Zhang, S. P. Nolan, R. L. de la Vega, S. L.
Mukerjee, C. D. Hoff and G. J. Kubas, Organometallics, 1988, 7,
2429; A. A. Gonzalez, K. Zhang and C. D. Hoff, Inorg. Chem., 1989,
28, 4285.
50 M. Yamakawa, H. Ito and R. Noyori, J. Am. Chem. Soc., 2000, 122,
1466.
31 R. M. Bullock, J.-S. Song and D. J. Szalda, Organometallics, 1996,
15, 2504.
51 D. A. Alonso, D. Guijarro, P. Pinho, O. Temme and P. G.
Andersson, J. Org. Chem., 1998, 2749; S. J. M. Nordin, P. Roth, T.
Tarnai, D. A. Alonso, P. Brandt and P. G. Andersson, Chem. Eur. J,
2001, 7, 1431.
52 D. A. Alonso, P. Brandt, S. J. M. Nordin and P. G. Andersson,
J. Am. Chem. Soc., 1999, 121, 9580.
53 K. Abdur-Rashid, A. J. Lough and R. H. Morris, Organometallics,
2000, 19, 2655; K. Abdur-Rashid, A. J. Lough and R. H. Morris,
Organometallics, 2001, 20, 1047; K. Abdur-Rashid, M. Faatz, A. J.
Lough and R. H. Morris, J. Am. Chem. Soc., 2001, 123, 7473.
54 Y. Blum, D. Czarkie, Y. Rahamim and Y. Shvo, Organometallics,
1985, 4, 1459; Y. Shvo, D. Czarkie, Y. Rahamim and D. F. Chodosh,
J. Am. Chem. Soc., 1986, 108, 7400; N. Menashe, E. Salant and
Y. Shvo, J. Organomet. Chem., 1996, 514, 97; Y. Shvo, I. Goldberg,
D. Czerkie, D. Reshef and Z. Stein, Organometallics, 1997, 16, 133.
55 C. P. Casey, S. W. Singer, D. R. Powell, R. K. Hayashi and
M. Kavana, J. Am. Chem. Soc., 2001, 123, 1090.
32 C. A. Tolman, Chem. Rev., 1977, 77, 313.
33 P. G. Jessop and R. H. Morris, Coord. Chem. Rev., 1992, 121, 155;
D. M. Heinekey and W. J. Oldham, Jr., Chem. Rev., 1993, 93, 913;
G. J. Kubas, Metal Dihydrogen and σ-Bond Complexes: Structure,
Theory, and Reactivity, Kluwer Academic/Plenum Publishers, New
York, 2001.
34 E. T. Papish, F. C. Rix, N. Spetseris, J. R. Norton and R. D.
Williams, J. Am. Chem. Soc., 2000, 122, 12235.
35 M. Tilset and V. D. Parker, J. Am. Chem. Soc., 1989, 111, 6711, as
modified in J. Am. Chem. Soc., 1990, 112, 2843.
36 J. M. Hanckel and M. Y. Darensbourg, J. Am. Chem. Soc., 1983,
105, 6979; M. Y. Darensbourg and M. M. Ludvig, Inorg. Chem.,
1986, 25, 2894.
37 S. E. Landau, R. H. Morris and A. J. Lough, Inorg. Chem., 1999, 38,
6060.
38 I. M. Kolthoff and M. K. Chantooni, Jr., J. Am. Chem. Soc., 1973,
95, 8539.
39 K.-T. Smith, J. R. Norton and M. Tilset, Organometallics, 1996, 15,
4515.
56 M. P. Magee and J. R. Norton, J. Am. Chem. Soc., 2001, 123,
1778.
57 M. Schlaf, P. Ghosh,, P. J. Fagan, E. Hauptman and R. M. Bullock,
Angew. Chem., Int. Ed., 2001, 40, 3887.
40 E. J. Moore, J. M. Sullivan and J. R. Norton, J. Am. Chem. Soc.,
1986, 108, 2257.
41 R. T. Edidin, J. M. Sullivan and J. R. Norton, J. Am. Chem. Soc.,
1987, 109, 3945.
42 A. J. Lough, S. Park, R. Ramachandran and R. H. Morris, J. Am.
Chem. Soc., 1994, 116, 8356; S. Park, R. Ramachandran, A. J.
Lough and R. H. Morris, J. Chem. Soc., Chem. Commun., 1994,
2201; J. C. Lee, Jr., E. Peris, A. L. Rheingold and R. H. Crabtree,
58 A. Bainbridge, P. J. Craig and M. Green, J. Chem. Soc. A, 1968,
2715; W. Beck, K. Schloter, K. Sünkel and G. Urban, Inorg. Synth.,
1989, 26, 96.
59 P. Kalck, R. Pince, R. Poilblanc and J. Roussel, J. Organomet.
Chem., 1970, 24, 445.
60 S. R. Bahr and P. Boudjouk, J. Org. Chem., 1992, 57, 5545.
61 M. Brookhart, B. Grant and A. F. Volpe, Jr., Organometallics, 1992,
11, 3920.
770
J. Chem. Soc., Dalton Trans., 2002, 759–770