Catalytic reduction of benzoate esters and lactones in the presence of PhMeSiH2 and a titanocene-based catalyst

Article abstract of DOI:10.1139/v02-047

PhMeSiH₂ reduces ethyl benzoate in the presence of a Cp₂TiMe₂ (Cp = η⁵-cyclopentadienyl) catalyst to give PhMeSi(OEt)(OCH₂Ph) (1). Small to moderate amounts of toluene are also produced, depending on the reaction conditions. Under 40 psi of H₂ the same reaction gives quantitative reduction of the ester to toluene. PhCHO is reduced, under the same conditions as the ester, to a mixture of PhMeSi(OCH₂Ph)₂ and toluene in proportions similar to that observed with the ester. A stoichiometric reaction of PhMeSiH₂, PhCOOEt, and Cp₂TiMe₂ resulted in the formation of known [Cp₂Ti(OEt)]₂ (2) in 75% isolated yield. A mechanism is proposed in which the C=O of the ester inserts into a Ti-H bond. A Ti-mediated transfer of the ethoxy group from the resulting (α-ethoxy)benzyloxytitanium intermediate to silicon produces PhMeSi(OEt)H and PhCHO. 1 is generated via hydrosilation of PhCHO by PhMeSi(OEt)H. Both β- and γ-butyrolactones react with PhMeSiH₂ in the presence of a catalytic amount of Cp₂TiMe₂ to produce copolymers -[O-PhMeSi-O(CH₂)₃CHR]_n- (R = H, 3; R = Me, 4) (M_w ≈ ×10³; M_w/M_n ≈ 1.3). A mechanism analogous to that proposed for the ethyl benzoate reduction is proposed. A reaction of PhMeSiH₂ and γ-butyrolactone with a stoichiometric amount of Cp₂TiMe₂ in the presence of tetrahydofuran (THF) produces Cp₂Ti(μ-H)(μ-η¹,η⁵-C ₅H₄)Ti(OC₄H₈)Cp (5) (65% isolated yield) whose structure is determined by X-ray crystallography.

Full text of DOI:10.1139/v02-047

4
89
Catalytic reduction of benzoate esters and
lactones in the presence of PhMeSiH and
2
a titanocene-based catalyst
Ronghua Shu, John F. Harrod, and Anne-Marie Lebuis
5
Abstract: PhMeSiH reduces ethyl benzoate in the presence of a Cp TiMe (Cp = ꢀ -cyclopentadienyl) catalyst to give
2
2
2
PhMeSi(OEt)(OCH Ph) (1). Small to moderate amounts of toluene are also produced, depending on the reaction condi-
2
tions. Under 40 psi of H the same reaction gives quantitative reduction of the ester to toluene. PhCHO is reduced, un-
2
der the same conditions as the ester, to a mixture of PhMeSi(OCH Ph) and toluene in proportions similar to that
2
2
observed with the ester. A stoichiometric reaction of PhMeSiH , PhCOOEt, and Cp TiMe resulted in the formation of
2
2
2
known [Cp Ti(OEt)] (2) in 75% isolated yield. A mechanism is proposed in which the C=O of the ester inserts into a
2
2
Ti–H bond. A Ti-mediated transfer of the ethoxy group from the resulting (a-ethoxy)benzyloxytitanium intermediate to
silicon produces PhMeSi(OEt)H and PhCHO. 1 is generated via hydrosilation of PhCHO by PhMeSi(OEt)H. Both b-
and ꢁ-butyrolactones react with PhMeSiH in the presence of a catalytic amount of Cp TiMe to produce copolymers
2
2
2
3
-
[O-PhMeSi-O(CH ) CHR] - (R = H, 3; R = Me, 4) (M W 1 × 10 ; M /M W 1.3). A mechanism analogous to that
2 3 n w w n
proposed for the ethyl benzoate reduction is proposed. A reaction of PhMeSiH and ꢁ-butyrolactone with a
2
1
5
stoichiometric amount of Cp TiMe in the presence of tetrahydofuran (THF) produces Cp Ti(ꢂ-H)(ꢂ-ꢀ ,ꢀ -
2
2
2
C H )Ti(OC H )Cp (5) (65% isolated yield) whose structure is determined by X-ray crystallography.
5
4
4
8
Key words: hydrosilation, esters, lactones, titanocene, catalysis.
5
Résumé : En présence d’un catalyseur de Cp TiMe (Cp = ꢀ -cyclopentadiényle), le PhMeSiH réduit le benzoate
2
2
2
d’éthyle en conduisant à la formation de PhMeSi(OEt)(OCH Ph), 1. Suivant les conditions expérimentales utilisées, il y
2
a aussi formation de faibles quantités de toluène. À une pression de H de 40 << psi >>, la même réaction conduit à
2
une réduction quantitative de l’ester en toluène. Dans les mêmes conditions de réduction, le PhCHO fournit un mé-
lange de PhMeSi(OCH Ph) et de toluène dont les proportions sont semblables à celles observées avec l’ester. Une
2
2
réaction stoechiométrique de PhMeSiH , PhCOOEt et de Cp TiMe conduit à la formation du [Cp Ti(OEt)] , 2, déjà
2
2
2
2
2
connu, avec un rendement isolé de 75%. On propose un mécanisme dans lequel le C=O de l’ester s’insère dans une
liaison Ti-H. Un transfert du groupe éthoxy, sous l’influence du Ti, de l’intermédiaire (a-éthoxy)benzoyloxytitane vers
le silicium conduit à la formation de PhMeSi(OEt)H et de PhCHO. Le produit 1 est généré par le biais d’une hydrosi-
lation du PhCHO par le PhMeSi(OEt)H. Les b- et ꢁ-butyrolactones réagissent toutes les deux avec le PhMeSiH , en
2
présence d’une quantité catalytique de Cp TiMe pour conduire à la formation de copolymères -[O-PhMeSi-
2
2
3
O(CH ) CHT] - (R = H, 3; R Me, 4) (M W 10 ; M /M W 1,3). On propose un mécanisme semblable à celui suggéré
2
3
n
w
w
n
pour la réduction du benzoate d’éthyle. Une réaction du PhMeSiH avec la ꢁ-butyrolactone et une quantité stoechiomé-
2
1
5
trique de Cp TiMe dans du tétrahydrofurane (THF) conduit à la formation de Cp Ti(ꢂ-H)(ꢂ-ꢀ ,ꢀ -C H )Ti(OC H )Cp,
2
2
2
5
4
4
8
5
(rendement de 65% en produit isolé) dont on a déterminé la structure par diffraction des rayons X.
Mots clés : hydrosilation, esters, lactones, titanocène, catalyse.
Traduit par la Rédaction] Shu et al. 495
[
Introduction
developed procedures for the reduction of both lactones (3)
and esters (4). A variety of esters, including those containing
bromo-, phenoxy-, amino-, alkenyl- or cyclopropyl-groups,
as well as a,b-unsaturated esters, can be reduced with trieth-
oxysilane and catalytic amounts of titanocene dichloride,
preactivated with butyllithium (4a). This procedure can also
be used for the selective reduction of methyl esters in the
presence of tert-butyl esters. For substrates containing a ter-
minal olefin, or an epoxide, a more hindered titanocene
dichloride species, such as ethylene-1,2-bis(tetrahydroin-
Titanocene derivatives have been widely applied as cata-
lysts for the hydrogenation and hydrosilation of carbonyl
compounds (1). Hydrosilation of ketones (2) has received
the most attention, but Buchwald and co-workers have also
Received 9 August 2001. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 30 April 2002.
1
R. Shu, J.F. Harrod, and A.-M. Lebuis. Chemistry
Department, McGill University, 801 Sherbrooke St. W.
Montreal, QC H3A 2K6, Canada.
denyl)titanium dichloride ([EBTHI]TiCl ), is required. Other
2
procedures involving safer and cheaper reagents, such as
poly(methylhydrosiloxane) instead of (EtO) SiH and
3
1Corresponding author (e-mail: john.harrod@mcgill.ca).
EtMgBr instead of n-BuLi, were also reported (4c).
Can. J. Chem. 80: 489–495 (2002)
DOI: 10.1139/V02-047
© 2002 NRC Canada

4
90
Can. J. Chem. Vol. 80, 2002
Most of the reported work on the hydrosilation of esters
Table 1. Yields of toluene from reactions of ethyl benzoate with
and lactones, catalyzed by titanocene derivatives, has been
focused on its use for synthesis of alcohols, while little is
known of the mechanism or the nature of the intermediates
in the reactions. On the other hand, our interest in
hydrosilation chemistry has been focused on the
organosilicon products and the titanocene intermediates (5).
In the present paper, we describe a detailed analysis of the
intermediates and products for some titanocene catalyzed
hydrosilations of esters and lactones. The results of these
studies provide the first direct evidence for the participation
PhMeSiH and Cp TiMe under various conditions.
2
2
2
Cp TiMe /PhCOOEt
T
Yield of toluene
2
2
a
(
%)
(°C)
(%)
5
20
20
20
20
20
0
20
20
40
40
5
10
19
11
13
2
10
20
30
40
10
20
20
20
20
b
c
of Cp Ti in a catalytic hydrosilation reaction catalyzed by
4
3
2
Cp TiMe . Our interest in these reactions evolved from an
2
2
earlier study of hydrosilation–hydrogenation of pyridine
carboxylic esters (6).
40
100
d
Note: Except where indicated, reactions were run under Ar with 10
mol% catalyst, based on PhMeSiH . All reactants were added to the
reactor at the same time. Conversions were quantitative and the balance of
2
Results and discussion
the product was 1.
a
b
1
Based on H NMR integration.
Reduction of ethyl benzoate under hydrosilation
conditions
Ethyl benzoate was added dropwise after the reaction of PhMeSiH and
2
Cp TiMe was complete.
2
2
To remove the complications arising from the presence of
the ring nitrogen in isonicotinate esters, we re-investigated
the reduction of ethyl benzoate under hydrosilation condi-
tions (eq. [1] — this equation is not balanced but shows the
observed reactants and products). Under most conditions
studied, benzyloxysilane (1) was the major product. Toluene
was generally a minor product, but was the only product
when the reaction was carried out under a hydrogen atmo-
sphere.
c
PhMeSiH (0.10 mL, 0.7 mmol) was added to the mixture of Cp TiMe
2
2
2
(0.088 g, 0.43 mmol) and ethyl benzoate (0.30 mL, 2.13 mmol). After the
solution color changed from orange to dark violet blue an additional 0.50
mL of PhMeSiH was added dropwise to the mixture.
2
d
The reaction was run under 40 psi H2.
O
Ph
Cp TiMe2 5%
2
[2]
[3]
+
PhMeSiH2
-[O-Si-O(CH ) CH ] -
2 3 2 n
O
8
0ºC
Cp TiMe2
Me
2
1] PhCOOEt + PhMeSiH2
MePhSi(OCH Ph)(OEt)]
+
PhMe
[
2
3
1
The results of a number of reactions carried out under a
O
Ph
Me
Cp TiMe2 5%
2
range of conditions are summarized in Table 1. It is difficult
to derive any conclusion from the results under Ar due to the
scatter in the data. Nevertheless, it is reasonable to conclude
that the C–O bond reduction is favored by higher tempera-
+
PhMeSiH2
-[O-Si-O(CH ) CH] -
2 3 n
80ºC
O
Me
4
tures and by the presence of H2.
An isolated sample of 1 did not react with PhMeSiH in
The polymers were recovered in ca. 30% yields, as amor-
phous white solids, by precipitation of the reaction mixtures
2
the presence of Cp TiMe . Reactions of benzaldehyde and
2
2
with excess methanol. Molecular weights, determined by
anisole were also briefly studied as substrates for the reac-
tion with PhMeSiH in the presence of Cp TiMe . A reac-
3
GPC, were in the range of 2–4
×
10 D, with
2
2
2
polydispersities in the range of 1.2–1.3. The balance of the
reaction products was not fully characterized, but from their
IR and NMR spectra appeared to be a mixture of lower mo-
lecular-weight oligomers of the same structure as the iso-
lated solid.
tion of benzaldehyde at room temperature for 1 h, with a
ratio of PhCHO–PhMeSiH –Cp TiMe of 1:2:0.15, gave
2
2
2
roughly equal amounts of PhMeSi(OCH Ph) and PhCH .
2
2
3
Under the same conditions, anisole gave only a trace of
product (PhCH , ca. 0.5%). After an additional 4 h at 60°C,
3
Stoichiometric reactions of Cp TiMe , PhMeSiH and ei-
2
2
2,
the yield of PhCH increased to ca. 10%.
3
ther ꢁ- or b-butyrolactone, in a molar ratio of 1:2:1 in hex-
ane–THF (15:2 v/v), yielded dark grayish-brown crystals.
Given the reaction of ethyl benzoate, described above, and
the known oxophilicity of Ti, a product containing Ti–O
bonds was expected. To our surprise, a structure determina-
tion by X-ray crystallographic analysis showed the product
to be 5 (Fig. 1).
A reaction of PhMeSiH , PhCOOEt, and Cp TiMe in a
2
2
2
molar ratio of 2:1:1 gave a 75% isolated yield of the known
ethoxy titanocene(III) dimer, [Cp Ti(OEt)] (2, ref. 7). This
pale green complex reacts slowly with PhMeSiH at room
temperature to give a blue product that exhibits a catalytic
activity similar to that of Cp TiMe for reduction of
2
2
2
2
2
PhCOOEt to 1, but with no detectable production of toluene.
Molecule 5 contains two inequivalent Ti atoms. One, Ti1,
5
5
is coordinated to a ꢀ -C H ligand, a bridging ꢀ -C H
5
5
5
4
1
Reactions of lactones with PhMeSiH and Cp TiMe
2
ligand, and a THF molecule. The other, Ti2, is ꢀ -bonded to
the bridging C H ligand and to two ꢀ -C H ligands. More
2
2
5
Both ꢁ- and b-butyrolactone reacted with PhMeSiH in the
2
5
4
5
5
presence of a Cp TiMe catalyst to yield the ring-opened co-
polymers 3 (eq. [2]) and 4 (eq. [3]), respectively.
importantly, a hydrogen-atom bridge, located in the differ-
ence map, connects the two Ti atoms. The mean distance
2
2
©
2002 NRC Canada

Shu et al.
491
Table 2. A comparison of some selected molecular parameters
Fig. 1. The structure of 5, showing thermal ellipsoids at the 30%
probability level.
for compounds 5 and 5>.
5
5>
Bond lengths (Å)
Ti1—O1
Ti1—Ti2
Ti2—C1
Ti1—C1
Ti1—C2
Ti1—C3
Ti1—C4
Ti1—C5
2.239(3)
Ti1—O1
Ti1—Ti2
Ti1—C5
Ti1—C1
Ti1—C5
Ti1—C4
Ti1—C3
Ti1—C2
Ti1—(C10–14)
2.26(1)
3.363(2)
2.19(2)
2.34(2)
2.32(2)
2.32(2)
2.37(2)
2.37(2)
2.38
3.329(2)
2.207(5)
2.325(5)
2.307(5)
2.363(5)
2.365(5)
2.333(5)
2.3796a
Ti1—(C6–10)
Bond angles (deg)
Ti2-Ti1-O1
Ti2-C1-C5
Ti2-C1-C2
95.65(10)
125.3(4)
123.3(3)
Ti2-Ti1-O1
Ti2-C5-C1
Ti2-C5-C4
96.1(30)
125.6(12)
124.2(13)
Note: The numbering schemes for the two structures are slightly
between the titaniums and the carbon atoms of the ꢀ-C H
different. Those for 5 are those used in ref. 8.
5
5
>
a
rings (2.388 Å) and the distance between C1 and Ti1
2.207(5) Å) agree with the assignment that three of the C5
These values are the average values for the five Ti–C distances of the
ꢀ5
-C H ring attached to Ti2.
(
5 5
5
rings are coordinating to Ti1 and Ti2 in ꢀ -mode, while the
fourth bridging ring is coordinating ꢀ to Ti1 and ꢀ to Ti2.
The formula of 5 (C H OTi ) corresponds to a dimer of
5
1
ground state, interpreted to be the result of a Ti(III)
coordinated to a bipyridyl radical anion ligand (11). It is not
unreasonable however to assume that a Ti(II)-centered trip-
let ground state could be preferred with a ꢃ-bonding weak-
field ligand such as THF that coordinates through O. Indeed,
both theoretical (12) and experimental (13) results support
2
4
28
2
Cp Ti coordinated to a molecule of THF. The crystal struc-
2
ture also reveals that in the lattice there is half a molecule of
THF per formula unit of 5.
The values of the molecular parameters for 5 (a number of
which are compared in Table 2) are essentially the same as
those reported by Pez (8) for compound 5>, which differed
only in the absence of the hydride bridge. We believe these
compounds to be the same. Small differences in the parame-
ter values are attributed to the fact that our data were col-
lected at –60°C, while those of Pez were collected at 23°C.
In addition, our measurements were carried out with Cu Ka
radiation in order to favour detection of the hydride, while
Pez used Mo Ka radiation. In a preliminary study, we also
failed to detect the bridging hydride using Mo Ka radiation.
Finally, there is a slight difference due to the fact that the
Pez structure is reported in the monoclinic space group I2/a,
which is a non-conventional orientation of the C2/c group
used by us.
the conclusion that the ground state of uncomplexed Cp Ti is
2
a triplet. We are continuing to study the magnetochemistry
and EPR properties of solutions of 5 in order to resolve this
question.
O
TiCp2 + 4 C
H
[4]
Cp
Ti
H
O
2 Cp
Ti(OC H )
4 8 2
2
4
8
2
5
6
The two titanium atoms in 5 can be formally assigned ei-
ther as Ti(II) and Ti(IV), or both as Ti(III). This oxidation-
state ambiguity, and the structural motif of a hydride bridge
accompanying another more-complex bridging group, is now
a common feature of titanocene chemistry (6, 14–16).
Solutions of 5 in THF are strongly paramagnetic. NMR
spectra of such solutions are very broad and strongly shifted,
and give essentially no clue to the structure. Their EPR spec-
tra also consist of a very broad single peak, which gives no
structural information. Pez also concluded that his com-
pound was paramagnetic on the basis of the temperature de-
The mechanism of ester hydrosilation–reduction
Earlier studies of hydrosilation–reduction of esters (17)
and lactones (18) in the presence of titanocene-based cata-
lysts were carried out from the perspective of their useful-
ness in organic synthesis. In no case was the intermediate
organosilicon compound identified or isolated. Nevertheless,
a reasonable mechanism involving addition of a Ti–H bond
across the carbonyl group, followed by release of the alkoxy
silicon product was proposed. A similar mechanism is
shown in Scheme 1.
1
pendence of the H NMR chemical shift. The originally
proposed structure, without a bridging hydrogen atom, is an
odd- electron count system and therefore would be expected
to have one unpaired electron. Addition of the bridging hy-
drogen atom results in an even-electron count and the mole-
cule is expected to be diamagnetic. A simple rationalization
of these observations is that 5 dissociates in the weakly co-
ordinating THF solvent to give a titanocene(II) complex (6),
as in eq. [4]. Although Cp TiL complexes are well known
The cycle shown in Scheme 1 rationalizes all of the fea-
tures of the hydrosilation reaction. There is ample evidence
2
2
for strong-field ligands such as PR (9) and CO (10), the
ground states of such complexes are normally spin singlets.
The bipyridyl complex is unusual, in that it has a spin-triplet
from earlier studies for the generation of Cp TiH by reaction
3
2
of Cp TiMe with a hydrosilane (5a and refs. therein). Step
2
2
a is a plausible insertion of C=O into a Ti–H bond. The
©
2002 NRC Canada

4
92
Can. J. Chem. Vol. 80, 2002
Scheme 1. A catalytic cycle for the hydrosilation of benzoate
esters.
Scheme 2. A cycle for the titanocene-hydride catalyzed co-
polymerization of ꢁ-butyrolactone–PhMeSiH2.
C
p
T
i
M
e
+
P
h
M
e
S
i
H
2
2
2
O
O
OEt
O
Cp TiO(CH ) C
2 2 3
Cp TiH
+
2
-
CH OSi
C
2
Cp TiH
H
2
O
PhMeSiH2
Si-H
d
c
a
Cp TiH
2
O-TiCp2
Cp TiH
2
-CH OTiCp2
O
H
Ph
O
H
O
O
H
2
C-OEt
C(CH ) OSiO(CH ) C
O
HPhMeSiO(CH ) C
2
3
2 3
+
2 3
Me
H
b
Cp TiH
H
2
C
O
e
Cp TiH
Cp TiOEt
2
2
PhMeSiH2
O
Ph
C(CH ) OSiO(CH ) CH OTiCp
2
PhMeSi(OEt)H
2 3
2 3
2
Etc
PhMeSiH2
H
Me
Cp TiH
2
observation that a stoichiometric reaction of Cp TiMe ,
2
2
PhMeSiH , and PhCOOEt gives 2 in high yield is explained
2
with Cp TiMe produce a mixture of CpTiSiR and Cp TiH,
2
2
3
2
by the assumption that the initially formed (1-phenyl-1-
ethoxy)methoxy titanocene is unstable with respect to trans-
fer of the ethoxy group to the Ti, perhaps by an initial
intramolecular coordination of the EtO oxygen to the metal.
The intermediate benzaldehyde was not detected under cata-
lytic conditions, perhaps because of an additional rapid reac-
tion. It was however shown that benzaldehyde is reduced
under these conditions to give PhMeSi(OCH Ph) (presum-
which have been isolated and characterized as bridged
dimers (5b, 15). The balance between the amounts of silyl
and hydride complexes is a function of the silane concentra-
tion, with the silyl species being favored by high silane con-
centration. It is also likely that H is in equilibrium with
2
these species as shown in eq. [5] (although this has never
been unequivocally demonstrated). The fact that
2
2
PhMeSi(OEt)(OCH Ph) does not react under the catalytic
2
ably by step c) and toluene in proportions similar to those
observed in the reaction with PhCOOEt. For the
benzaldehyde reaction, there is no possibility of EtO transfer
to the Si and hence it raises the question as to why the ob-
served product in the ester reaction (step d, Scheme 1) is
PhMeSi(OEt)(OCH Ph) rather than PhMeHSi(OCH Ph), or
reduction conditions excludes it as an intermediate in the
production of toluene. On the other hand, the similarity of
the products from the PhCOOEt and PhCHO reductions sug-
gests that Cp Ti(OCH Ph) is a plausible intermediate in the
2
2
production of toluene.
2
2
PhMeSi(OCH Ph) . A problem is also presented by the fact
Cp TiH R SiH
+
Cp TiSiR3
+
H2
2
2
[5]
2
3
2
that PhMeSiH2 does react with [Cp Ti(OEt)] (step e
2
2
Scheme 1), and the latter is also a precatalyst for the
hydrosilation of PhCOOEt. Neither of these reactions, how-
The mechanism of lactone–silane co-polymerization
A plausible mechanism for the titanocene-hydride medi-
ated co-polymerization of ꢁ-butyrolactone and PhMeSiH is
shown in Scheme 2. This scheme couples two of the steps
proposed in Scheme 1 for the reduction of ethylbenzoate;
ever, are fast enough to account for the rate of the Cp TiMe
2
2
2
catalyzed hydrosilation of PhCOOEt.
Both of these problems would be resolved if the slowness
of the reaction of 2 with PhMeSiH was due to a slow disso-
namely, the addition of Cp TiH to an aldehyde and a ꢃ-bond
2
2
ciation of the dimer, which masks the intrinsic reactivity of
metathesis between Ti–O and Si–H to give Si–O and Ti–H.
In this case the ring-opening step is an intramolecular ver-
sion of steps a and b in Scheme 1. In the lactone reactions
no cleavage of C–O bonds was observed.
monomeric Cp Ti(OR). The assumption of the production of
2
a highly reactive monomeric Cp Ti(OEt) in step c permits
2
the rapid recycling of this intermediate through loop d, to al-
low for the transfer of the EtO group to silicon and regenera-
tion of the catalyst, Cp TiH. Finally, once the reaction is
2
The mechanism of formation of 5
underway, the Si–H in the ꢃ-bond metathesis in step d
would have to be mainly PhMe(EtO)SiH, rather than
PhMeSiH2.
Although the steps in the catalytic loop of Scheme 1 can
explain the features of the hydrosilation reaction of
PhCOOEt, they fail to explain the complete reduction
Compound 5 is expected to be formed from the
5
dimerization of bis( -cyclopentadienyl)Ti. The manner in
ꢀ
which the latter species may be generated by reaction of
Cp TiMe with a hydrosilane has been discussed at length
2
2
elsewhere (5b and refs. therein). It is our belief that
titanocene(II), or its dimer (5) is normally present in reac-
tions involving Si–H and Cp TiMe , but its detection in so-
(
deoxygenation) to toluene. The fact that toluene is the only
2
2
product when the reaction is carried out under moderate
lution is precluded by its electronic structure. (Its
paramagnetism interferes with NMR measurements and a
triplet ground state can be very difficult to observe by EPR.)
pressure of H suggests that high concentrations of Ti–H fa-
2
vour complete deoxygenation. Reactions of hydrosilanes
©
2002 NRC Canada

Shu et al.
493
Scheme 3. In A, one molecule of Cp Ti inserts into a C–H bond
of a second molecule. In B, a second intramolecular insertion of
organic phase was washed with distilled water three times
2
and then dried with MgSO . Pure benzylmethyl ether was
4
Ti into a C–H bond of the Cp Ti unit of 5 occurs, followed by
obtained in 80% isolated yield by distillation.
2
1
coupling of the two C H fragments. In C, a molecule of H re-
H NMR spectra were recorded on a Varian XL-200,
5
4
2
acts with two Cp Ti molecules, perhaps by initial formation of
Varian XL-300, Gemini-200, or Unity-500 spectrometer us-
2
1
3
the Cp TiH , followed by reaction with the second Cp Ti mole-
ing an internal solvent as a reference. C NMR spectra were
recorded on Unity-500 spectrometer using C D
128.0 ppm) as a reference. GPC analyses were run with
2
2
2
cule.
a
6
6
(
THF as solvent and calibrated relative to polystyrene stan-
dards.
O
THF
A
H
Ti
H
H
5
Reaction of ethyl benzoate with PhMeSiH catalyzed by
2
C
H
2
B
Cp TiMe
2
2
PhMeSiH2 (0.20 mL, 1.4 mmol) and ethyl benzoate
0.10 mL, 0.70 mmol) were added to Cp TiMe (14 mg,
H
(
2
2
H
H
0
.07 mmol). After a while, the solution color changed to
6
7
dark blue then brown with vigorous gas evolution. After stir-
ring at room temperature for 1 h, quantitative conversion of
ethyl benzoate to a mixture of toluene and 1 was observed
It has also been postulated that titanocene, as a result of its
carbene-like character, will insert into X–H bonds and, in
the absence of other choices, it will insert into one of its
own C–H bonds (19). Pez (8) recognized this possibility
when he first isolated and characterized 5. A second similar-
intramolecular insertion takes 5 to the hydride-bridged
fulvalenyl dititanium complex (7), originally reported by
Bercaw and co-workers (20) and structurally characterized
by Troyanov et al. (21). The relationship between the vari-
ous titanocene hydrides is shown in Scheme 3.
1
1
by H NMR (molar ratio 1:9). H NMR (300 MHz, C D )
6
6
for 1: 0.34 (s, 3H, Si-CH ), 1.15 (t, 3H, OCH CH ), 3.70 (q,
3
2
3
2
H, OCH CH ), 4.78 (m, 2H, PhCH O-), 6.9–8.0 (m, Ph-H).
2 3 2
All of the experiments to obtain the data in Table 1 were
carried out following the above general procedure, with the
exception of the last entry, which was carried out in a stirred
mini-autoclave rather than a Schlenk tube and with all of the
reactants scaled up by a 10-fold.
Preparation of [Cp Ti(ꢂ-OEt)]
2
2
PhMeSiH2 (0.70 mL, 4.9 mmol) and ethyl benzoate
0.40 mL, 2.5 mmol) were added to a solution of Cp TiMe
Conclusions
(
2
2
The titanocene-catalyzed hydrosilation of esters to the
alkoxide level can be rationalized by a conventional Ti–H
mediated reduction of the ester, followed by alkoxide trans-
fer to Si. Further reduction to the hydrocarbon level occurs
under unusually mild conditions, particularly in the presence
of a hydrogen atmosphere. Although the mechanism is not
known, it is likely that this reaction occurs via X–H (X = H
or Ti) induced C–O cleavage of a benzyloxy-Ti complex.
(0.50 g, 2.4 mmol in 40 mL toluene and 10 mL benzene).
The solution turned dark green with precipitation of green
crystals of [Cp2Ti(ꢂ-OEt)]2, which were filtered, washed
with hexane, and dried in vacuo (0.40 g, 75% isolated yield).
The identity of the product was confirmed by comparison of
its IR spectrum and X-ray diffraction pattern to those of an
authentic sample (7).
Reaction of ethyl benzoate and PhMeSiH in the
2
presence of [Cp Ti(ꢂ-OEt)]
Experimental
2
2
PhMeSiH (0.34 mL, 2.60 mmol) was added to [Cp Ti(ꢂ-
2
2
General manipulations and materials
OEt)]2 (50 mg, 0.11 mmol). The mixture was stirred at room
All manipulations were performed under an atmosphere of
nitrogen or argon using Schlenk techniques. Dry oxygen-
free solvents were employed throughout. Glassware was
flame-dried or oven-dried before use. ꢁ-Butyrolactone and
temperature for 24 h (during this time the formation of some
1
1,2-dimethyl-1,2-diphenyldisilane was observed by
H
NMR). Ethyl benzoate (0.17 mL, 1.06 mmol) was then
added. Gas evolution was observed in 5 min, while the solu-
tion changed from green to yellow-green. Conversion (86%)
of the ester to 1 was obtained in 12 h. No toluene was de-
(
±)-b-butyrolactone were purchased from Aldrich and dis-
tilled prior to use. Ethyl benzoate, benzaldehyde, and benzyl
alcohol were purchased from Aldrich and purified by distil-
lation over calcium hydride before use. Dimethyltitanocene
1
tected by H NMR.
(
22) and phenylmethylsilane (23) were prepared by the pro-
Reaction of benzaldehyde and PhMeSiH in the
2
cedures described in the literature.
presence of Cp TiMe
2
2
Benzylmethyl ether was prepared by a Williamson synthe-
sis using the following procedure: sodium (2.3 g, 0.1 mol)
was added in small pieces to benzyl alcohol (11 g, 0.1 mol)
in ether (50 mL). After the sodium had dissolved, CH3I
PhMeSiH2 (0.40 mL, 3.06 mmol) and benzaldehyde
(0.20 mL, 1.97 mmol) were added to Cp TiMe (55 mg,
2
2
0.28 mmol) at room temperature. After a while, the solution
color changed to purple then dark blue accompanied by gas
evolution. After 1 h toluene was observed in a 50% yield
(
16 g, 0.11 mol) was added dropwise from a syringe. A
white solid precipitated during the addition of CH I. The
1
(based on ester, by H NMR). The reaction mixture was di-
3
suspension was stirred for 1 h and then dilute HCl (5 mL
luted with C D (1.0 mL) and saturated with gaseous HCl
6
6
concentrated HCl in 30 mL H O) was added dropwise. The
to destroy the presumed Cp Ti(OCH Ph). Volatiles were
2
2 2
©
2002 NRC Canada

4
94
Can. J. Chem. Vol. 80, 2002
removed under vacuum and collected in an NMR tube. The
C H O Ti : C 67.86, H 6.96, O 4.35, Ti 20.84; found:
2
6
32 1.5
2
resulting mixture contained toluene and PhCH OH in
C 67.58, H 6.75. The low CH analysis is probably due to
loss of lattice THF during sample isolation. The IR spectrum
of 5 showed the same features as that of the compound de-
scribed by Pez (8). Despite several attempts, we were not
able to assign an IR band due to the bridging hydride of 5,
2
roughly equal amounts, as shown by spiking the mixture
with authentic samples.
Reaction of PhCH OMe and PhMeSiH in the presence
2
2
or to its deuterated analogue, prepared with PhMeSiD2.
of Cp TiMe
2
2
PhMeSiH2 (0.60 mL, 4.20 mmol) and PhCH OMe
2
(
0.30 mL, 2.46 mmol) were added to Cp TiMe (0.10 g,
Collection of X-ray data and structure determination of 5
2
2
0
.48 mmol). The solution color changed from orange to dark
Data were collected on a crystal of 5 of dimensions of
blue, accompanied by gas evolution. After stirring at 60°C
for 4 h, toluene was detected in the solution by H NMR in
0.52 × 0.29 × 0.01 mm mounted in a glass capillary under
1
2
argon. Crystal data: C H O Ti ·1/2C H O ; FW
=
2
4
28
1
2
4
8
1
an amount equivalent to 10% of the reactant ether.
456.299; monoclinic, C2/c; a = 31.232(13), b = 7.998(3),
In the absence of Cp TiMe , reactions of PhCH OMe with
c = 19.603(9) Å, b = 116.78(3)°; Z = 8, Z> = 1; D =
2
2
2
c
–
3
PhMeSiH or (EtO)Me SiH at 60°C for 12 h yield no tolu-
1.387 g cm . Data were collected on a CAD4 diffractometer
using Cu Ka radiation. In all, 29561 reflections were mea-
sured, of which 4145 unique (Rint 14.9%) reflections were
used for structure solution and refinement. Data were cor-
2
2
ene.
Reaction of ꢁ-butyrolactone and PhMeSiH catalyzed
2
–
1
rected for absorption (ꢂ = 63.48 cm ; transmission range:
by Cp TiMe
2
2
0
0
.2708–0.9287). Final agreement factors: R : (obsd./all)
ꢁ
-Butyrolactone (0.30 mL, 3.9 mmol) and PhMeSiH2
1
.059/0.094; wR : 0.143/0.157. The structure was solved by
(
(
0.60 mL, 4.2 mmol) were added to a solution of Cp TiMe2
2
2
SHELXLS-96 (24) and refined in SHELXL-96 (25). All
nonhydrogen atoms are anisotropic. Hydrogen atoms are cal-
culated except the bridging hydride, which was located in
the difference map and refined isotropically with the re-
straint that both Ti–H distances be similar.
0.05 g, 0.24 mmol in 1 mL THF). After thorough mixing,
the solution was stirred for 10 h at 80°C. The product was
added dropwise to 10 mL of well-stirred dry methanol; 0.3 g
of a white precipitate were obtained. The compound was as-
1
13
signed as -[OSiPhMeO(CH ) ] - based on H and C NMR
2
4 n
3
spectra. GPC analysis indicated a M = 4.9 × 10 , M /M =
w
w
n
1
1
(
7
(
.3. H NMR (300 MHz, C D ) d: 0.40 (s, 3H, Si-CH ), 1.7
6 6 3
Acknowledgments
m, 4H, CH (CH ) CH ), 3.75 (m, 4H, CH (CH ) CH ),
2 2 2 2 2 2 2 2
1
3
.0–8.0 (m, 5H, Ph-H). C NMR (500 MHz, C D ) d: –5
Financial support from NSERC of Canada and Fonds
FCAR du Québec are gratefully acknowledged.
6
6
1C, SiCH ), 29 (2C, CH (CH)2CH), 62 (2C, CH (CH ) CH ),
3
2
2
2 2
2
1
30, 134 (6C, C H ).
6
5
References
Reaction of (±)-b-butyrolactone and PhMeSiH in the
2
presence of Cp TiMe
1. N.A. Petasis and Y.H. Hu. Current Org. Chem. 1, 249 (1997).
2. (a) F. Sato, T. Jinbo, and M. Sato. Tetrahedron Lett. 21, 2171
(1980); (b) T. Nakano and Y. Nagai. Chem. Lett. 481 (1988);
(c) R.L. Halterman, T.M. Ramsey, and Z.-L. Chen. J. Org.
Chem. 59, 2642 (1994); (d) S.-X. Xin and J.F. Harrod. Can. J.
Chem. 73, 999 (1995); (e) M.B. Carter, B. Schiøtt, A.
Gutiérrez, and S.L. Buchwald. J. Am. Chem. Soc. 116, 11667
(1994).
2
2
(
±)-b-Butyrolactone (0.30 mL, 3.7 mmol) and PhMeSiH2
(
(
0.55 mL, 3.8 mmol) were added to a solution of Cp TiMe2
0.05 g, 0.24 mmol, in 1 mL THF). After thorough mixing,
2
the solution was stirred for 8 h at 80°C. The product was
added dropwise to 10 mL of well-stirred methanol. A white
precipitate (0.28 g) was obtained. The compound was identi-
1
fied as -[OSiPhMeO(CH ) CH(CH )] - based on its
H
2
2
3 n
3
3. (a) X. Verdaguer, S.C. Berk, and S.L. Buchwald. J. Am.
Chem. Soc. 117, 12641 (1995); (b) X. Verdaguer, M.C.
Hansen, S.C. Berk, and S.L. Buchwald. J. Org. Chem. 62,
NMR spectrum. GPC analysis indicated a M = 1.8 × 10
w
1
with M /M = 1.2. H NMR (300 MHz, C D ) d: 0.40 (s,
w
n
6
6
3
H, Si-CH ), 1.1 (m, 3H, CH CH CH(CH )), 1.4 (m, 2H,
3 2 2 3
8
522 (1997).
CH CH CH(CH )), 3.6 (m, 2H, CH CH CH(CH )), 3.9 (m,
2
2
3
2
2
3
4
5
. (a) S.C. Berk, K.A. Kreutzer, and S.L. Buchwald. J. Am.
Chem. Soc. 113, 5093 (1991); (b) S.C. Berk and S.L.
Buchwald. J. Org. Chem. 57, 3751 (1992); (c) K.J. Barr, S.C.
Berk, and S.L. Buchwald. J. Org. Chem. 59, 4323 (1994);
(d) M.T. Reding and S.L. Buchwald. J. Org. Chem. 60, 7884
(1995).
1
H, CH CH CH(CH )), 7.1–7.9 (m, 5H, Ph-H).
2 2 3
Sythesis of crystalline C H O Ti (5)
2
6
32 1.5
2
ꢁ
-Butyrolactone (0.02 mL, 0.26 mmol) and PhMeSiH2
(
0.07 mL, 0.49 mmol) were added to a Cp TiMe in hex-
2 2
ane–THF solution (50 mg, 0.24 mmol, in 15 mL hexane and
mL THF). After thorough mixing, the flask was left undis-
-
. (a) L. Hao, J.F. Harrod, A.-M. Lebuis, Y. Mu, R. Shu, E. Sam
2
uel, and H.-G. Woo. Angew. Chem. Int. Ed. Engl. 37, 3126
(1998); (b) J.F. Harrod. Coord. Chem. Rev. 206–207, 541
(2000).
turbed for 24 h. Dark gray plates of 5 (33 mg) crystallized
from the solution (65% isolated yield). Anal. calcd. for
2
Supplementary material may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research
Council Canada, Ottawa, ON K1A 0S2, Canada. For information on obtaining material electronically go to
http://www.nrc.ca/cisti/irm/unpub_e.shtml. Crystallographic information has also been deposited with the Cambridge Crystallographic Data
Centre (CCDC No. 169316). Copies of the data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the
Cambridge Crystallographic Data Centre, 12, Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033; or deposit@ccdc.cam.ac.uk).
©
2002 NRC Canada

Shu et al.
495
6
7
. J.F. Harrod, R. Shu, H.-G. Woo, and E. Samuel. Can. J. Chem.
9, 1075 (2001).
. E. Samuel, J.F. Harrod, D. Gourier, Y. Dromzee, F. Robert,
and Y. Jeannin. Inorg. Chem. 31, 3252 (1992).
. G.P. Pez. J. Am. Chem. Soc. 98, 8072 (1976).
. (a) L.B. Kool, M.D. Rausch, M. Alt, M. Heberhold, U.
Thewalt, and B. Wolf. J. Organomet. Chem. 310, 27 (1986);
17. (a) S.C. Berk, K.A. Kreutzer, and S.L. Buchwald. J. Am.
Chem. Soc. 113, 5093 (1991); (b) S.C. Berk and S.L.
Buchwald. J. Org. Chem. 57, 3751 (1992); (c) K.J. Barr, S.C.
Berk, and S.L. Buchwald. J. Org. Chem. 59, 4323 (1994);
(d) M.T. Reding and S.L. Buchwald. J. Org. Chem. 60, 7884
(1995).
18. (a) X. Verdaguer, S.C. Berk, and S.L. Buchwald. J. Am.
Chem. Soc. 117, 12641 (1995); (b) X. Verdaguer, M.C.
Hansen, S.C. Berk, and S.L. Buchwald. J. Org. Chem. 62,
8522 (1997).
7
8
9
(
b) D.J. Sikora, J.W. Macomber, and M.D. Rausch. Adv.
Organomet. Chem. 25, 316 (1986).
1
0. (a) L.P. Battaglia, M. Nardelli, C. Pelizzi, G. Predieri, and G.P.
Chiusoli. J. Organomet. Chem. 259, 301 (1983); (b) M.D.
Fryzuk, T.S. Haddad, and D.J. Berg. Coord. Chem. Rev. 90,
19. J.W. Lauher and R. Hoffmann. J. Am. Chem. Soc. 98, 1729
(1976).
1
37 (1990).
20. (a) J. Bercaw and H.H. Brintzinger. J. Am. Chem. Soc. 91,
7301 (1969); (b) H.H. Brintzinger and J. Bercaw J. Am.
Chem. Soc. 92, 6182 (1970).
1
1
1
1
1
1
1. R. Gyepes, P.T. Witte, M. Horcek, I. Císarová, and K. Mach. J.
Organomet. Chem. 551, 207 (1998).
2. J.F. Harrod, T. Ziegler, and V. Tschinke. Organometallics, 9,
21. S.I. Troyanov, H. Antropiusová, and K.J. Mach. J. Organomet.
Chem. 49, 427 (1992).
22. K. Claus and K. Bestiam. Liebigs Ann. Chem. 8, 654 (1962).
23. R.A. Benkeser, H. Landesman, and D.J. Foster. J. Am. Chem.
Soc. 74, 648. (1952).
24. G.M. Sheldrick. SHELXS-96. Program for the solution of
crystal structures. University of Gottingen, Germany. 1996.
25. G.M. Sheldrick. SHELXL-96. Program for structure analysis.
University of Göttingen, Germany. 1996.
8
90 (1990)
3. P.B. Hitchcock, F.M. Kerton, and F.M. Lawless. J. Am. Chem.
Soc. 120, 10264 (1998).
4. L.J. Guggenberger and F.N. Tebbe. J. Am. Chem. Soc. 95,
7
870 (1973)
5. C. Aitken, J.F. Harrod, and E. Samuel. J. Am. Chem. Soc. 108,
059 (1986).
4
6. S. Xin, H.-G. Woo, J.F. Harrod, E. Samuel, and A.-M. Lebuis.
J. Am Chem. Soc. 119, 5307 (1997).
©
2002 NRC Canada