Mechanism of the Kulinkovich Cyclopropanol Synthesis: Transfer-Epititanation of the Alkene in Generating the Key Titanacyclopropane Intermediate

Article abstract of DOI:10.1002/ejoc.200300588

An investigation of the Kulinkovich cyclopropanol synthesis, the interaction of esters with 2:1 or 3:1 mixtures of alkyl Grignard reagents and Ti(OiPr)₄ at low temperatures, has been conducted, in order to ascertain which reactive intermediates are involved and how they are interconverted. Because of the nature of the ultimate product, one of the most obvious intermediates is the 1,1-diisopropoxy-1-titanacyclopropane stemming from the epititanation of the alkene set free from the alkyl Grignard reagent employed. A search for the formation of such a titanocycle by warming an ethereal solution of either Et₂Ti(OiPr)₂ or iPr ₂Ti(OiPr)₂ between -78 °C and +25 °C was attempted by chemical trapping with either an ester or nitrile. In this manner it was shown that such a titanocycle was formed in the case of Et2Ti(OiPr) ₂ but not with iPr₂Ti(OiPr)₂. As to the role of two other potential intermediates, Ti(OiPr)₂ and R ₂Ti(OiPr)₂, it was demonstrated that preformed Ti(OiPr)₂ in the presence of ethylene and an ester does not form the corresponding cyclopropanol. Thus, under the reaction conditions Ti(OiPr) ₂ cannot perform the direct epimetallation necessary to produce the requisite titanacyclopropane. On the other hand, either iPr ₂Ti(OiPr)₂ or Et₂-Ti(OiPr)₂ can achieve the transfer-epititanation of ethylene at low temperatures and hence with methyl benzoate yield 1-phenyl-1-cyclopropanol. In contrast, neither iPr₂Ti(OiPr)₂ nor Et_{2Ti(OiPr) 2} can at low temperatures transfer-epititanate propylene. This difference in alkene reactivity can be ascribed to steric factors operating in the proposed octahedral transition state for transfer-epimetallation. Finally, by introducing free ethylene into such Kulinkovich reaction mixtures, either by ethylene gas itself or a third equivalent of EtMgX, the isolated yields of cyclopropanols were more than doubled over those obtained with a 1:2 ratio of Ti(OiPr)₄/EtMgX. From this observation one can conclude that free ethylene catalytically initiates the Kulinkovich reaction by coordinating with Et₂Ti(OiPr)₂ and undergoes transfer-epititanation to produce the requisite titanacyclopropane and thereby liberates ethylene, which perpetuates the reaction. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2003.

Full text of DOI:10.1002/ejoc.200300588

SHORT COMMUNICATION
Mechanism of the Kulinkovich Cyclopropanol Synthesis: Transfer-Epititanation
[
‡]
of the Alkene in Generating the Key Titanacyclopropane Intermediate
[a]
[a]
[a]
John J. Eisch,* Adetenu A. Adeosun, and John N. Gitua
Keywords: Titanium / Synthesis design / Alkene ligands
An investigation of the Kulinkovich cyclopropanol synthesis,
the interaction of esters with 2:1 or 3:1 mixtures of alkyl Grig-
clopropane. On the other hand, either iPr
2 2 2
Ti(OiPr) or Et -
Ti(OiPr) can achieve the transfer-epititanation of ethylene at
2
4
nard reagents and Ti(OiPr) at low temperatures, has been low temperatures and hence with methyl benzoate yield 1-
conducted, in order to ascertain which reactive intermediates
are involved and how they are interconverted. Because of
the nature of the ultimate product, one of the most obvious
intermediates is the 1,1-diisopropoxy-1-titanacyclopropane
stemming from the epititanation of the alkene set free from
the alkyl Grignard reagent employed. A search for the forma-
tion of such a titanocycle by warming an ethereal solution of
phenyl-1-cyclopropanol. In contrast, neither iPr
2 2
Ti(OiPr) nor
Et Ti(OiPr) can at low temperatures transfer-epititanate pro-
2
2
pylene. This difference in alkene reactivity can be ascribed
to steric factors operating in the proposed octahedral trans-
ition state for transfer-epimetallation. Finally, by introducing
free ethylene into such Kulinkovich reaction mixtures, either
by ethylene gas itself or a third equivalent of EtMgX, the
isolated yields of cyclopropanols were more than doubled
over those obtained with a 1:2 ratio of Ti(OiPr) /EtMgX.
4
From this observation one can conclude that free ethylene
catalytically initiates the Kulinkovich reaction by coordinat-
either Et
C was attempted by chemical trapping with either an ester
or nitrile. In this manner it was shown that such a titanocycle
was formed in the case of Et Ti(OiPr) but not with
iPr Ti(OiPr)
diates, Ti(OiPr)
preformed Ti(OiPr)
does not form the corresponding cyclopropanol. Thus, under
the reaction conditions Ti(OiPr) cannot perform the direct
2 2 2 2
Ti(OiPr) or iPr Ti(OiPr) between −78 °C and +25
°
2
2
2
2 2 2
. As to the role of two other potential interme- ing with Et Ti(OiPr) and undergoes transfer-epititanation to
2
and R
2
Ti(OiPr)
2
, it was demonstrated that
produce the requisite titanacyclopropane and thereby liber-
ates ethylene, which perpetuates the reaction.
2
in the presence of ethylene and an ester
2
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
epimetallation necessary to produce the requisite titanacy-
Germany, 2003)
Introduction
Soon thereafter, a cyclopropanol synthesis catalyzed by
Ϫ10 mol % of 3 focused attention on this process for the
One of the most remarkable transition-metal mediated
reactions of recent origin has been the titanium-catalyzed
Kulinkovich cyclopropanol synthesis.[2,3] In its prototypical
form, reported by Kulinkovich and co-workers in 1989, the
addition of various methyl esters 1 to an ethereal mixture
of ethylmagnesium bromide (2) and titanium() isopropox-
ide (3) in a 3:1 molar ratio, starting at Ϫ78 °C with hydroly-
sis at 0 °C, led to excellent yields of 1-substituted cyclopro-
panols 4 [Equation (1)].[
[5]
5
elaboration of natural products containing either 1-substi-
tuted- or stereoselective 1,2-disubstituted cyclo-
propanols. The use of N,N-dialkylcarboxamides 5 in
[6Ϫ8]
place of esters, using the original Kulinkovich procedure,
allowed de Meijere and co-workers to obtain N,N-dibenzyl-
cyclopropylamines 6 in moderate to good yields [Equa-
[9,10]
tion (2)].
4]
(
2)
(
1)
The course of the Kulinkovich reaction, when an ethyl
Grignard reagent is employed, has been depicted as pro-
The State University of New York at Binghamton, Department ceeding via diethyltitanium() diisopropoxide (7) to give
[
‡]
a]
Organic Chemistry of Subvalent Transition Metal Complexes,
[
1]
27. Part 26: Ref.
[
of Chemistry,
the titanacyclopropane 9, without specifying whether ti-
tanium() diisopropoxide (8) is ever set free before it recom-
bines to form 9 (path b), or whether 9 forms directly from
P. O. Box 6000, Binghamton, New York 13902-6000, USA.
Fax: (internat.) ϩ 1-607-777-4865
E-mail: jjeisch@binghamton.edu
Eur. J. Org. Chem. 2003, 4721Ϫ4727
DOI: 10.1002/ejoc.200300588
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4721

J. J. Eisch, A. A. Adeosun, J. N. Gitua
SHORT COMMUNICATION
7
with the elimination of ethane (path a) (Scheme 1). The gand transfer between 11 and the unsaturated organic sub-
interaction of 9 with an ester to form the cyclopropanolate strate has remained unproved and hence uncertain.
salt precursor to 4, is viewed as occurring by the generally Our recent studies on the epititanation of unsaturated or-
accepted reaction pathway for vicinal dimetalloethane re- ganic linkages, CϭE or CϵE (E ϭ C, O, N), either directly
[
2,3,4,5]
[11,15,16]
agents.
with TiL (15)
or indirectly by transfer-epititanation
to form three-membered titanocycles
2
[
17]
with R TiL (16)
2
2
17 have increased our understanding of the formation and
reactivity of such metallocyclic intermediates (Scheme 3).
Accordingly, in the present work we sought to employ such
insights for determining the nature of the reactive inter-
mediates in the Kulinkovich reaction and their possible
pathways of generation.
Scheme 1
Scheme 3
In modifications of the Kulinkovich procedure, where the
[
7]
Grignard reagent (or an analogous organolithium re-
agent ) does not serve as the source of the olefin or acety-
Results and Discussion
[
11]
lene but only as a reductant for the Ti(OiPr) , it has been
4
Reactive Intermediates in the Kulinkovich Reaction and Its
Modifications
assumed without proof that the dialkyltitanium() dialkox-
ide formed decomposes analogously to that mentioned
above, namely 10 to 11 as does 7 to 9, but then the added
allylic halide, olefin or acetylene undergoes ligand exchange
with the intermediate titanacyclopropane 11 to form the de-
rived titanium reagents 12, 13 or 14 (Scheme 2).
In the prototypical Kulinkovich reaction the thermal de-
composition of diethyltitanium() diisopropoxide (7) is
thought to generate the titanacyclopropane 9 intermediate,
since capture with the ester reactant provides a fitting
rationale for the observed cyclopropanol ring formation
[Scheme 1 and Equation (1)].
If, however, an ethereal solution of the titanium() inter-
mediate 7 is generated at Ϫ78 °C, in the absence of an ester,
and then brought to 25 °C, the solution turns black, sig-
[
18]
naling the formation of 8. Introduction of the ester 1 to
the foregoing reaction solution and further stirring for 7 h
leads, upon the hydrolysis step, to no detectable amounts of
cyclopropanol 4. These results show that 9 is only stable,
and hence available for cyclopropanol synthesis, at low tem-
peratures (Scheme 1).
Similarly, decomposition of diisopropyltitanium() di-
isopropoxide (10) at low temperatures has been proposed
to occur via the titanacyclopropane 11, the presumed
Scheme 2
source of Ti(OiPr) (8) for various organic synthetic pro-
2
[
7]
cedures used by Sato’s group involving allylic halides, ole-
[
13]
[14]
This viewpoint is exemplified by the work of both the fins
or alkynes
(Scheme 2). In order to determine
[
7]
[8,12]
Sato group and the Cha group.
Sato’s approach uses whether alkyltitanium 10 does decompose at Ϫ78 °C or at
mainly the isopropyl Grignard reagent to generate the puta- higher temperatures to give the 2-methyltitanacyclopropane
2
tive Kulinkovich intermediate 11, via 10, with subsequent 11 [termed (η -propene)Ti(OiPr) by Sato, Scheme 4] two
2
transfer of Ti(OiPr) (8) presumably by ligand transfer to reactions were designed and carried out to chemically trap
2
[
7]
[13]
[14]
an allylic halide, olefin
or to acetylene,
to form the any of the intermediate 11 formed. In the first reaction, a
titanometallics 12, 13 or 14, respectively (Scheme 2). Al- mixture of presumed 11 and benzonitrile 18 was allowed to
though the final titanium-free products formed by the hy- react in the temperature range of Ϫ78 °C to ϩ25 °C; in
drolysis or oxidation of, or the insertion of unsaturated sub- the second reaction, a mixture of presumed 11 and methyl
strates into, the σ(CϪTi) bonds are consistent with the benzoate (1a) was subjected to the same reaction con-
presence of 12, 13 or 14 as precursors, the pathway of their ditions. If any of intermediate 11 were generated in this tem-
formation via intermediates 10 and 11, and subsequent li- perature range, it would have undergone the Kulinkovich
4722
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4721Ϫ4727

Mechanism of the Kulinkovich Cyclopropanol Synthesis
SHORT COMMUNICATION
reaction with 1a to produce cyclopropanol 19 and similarly
would have inserted benzonitrile into the CϪTi bonds to
form the titanocycles 20 and 21 and, upon hydrolysis, the
[
19]
ketones 22 and 23.
(
3)
This result has led us to conclude that at 25 °C 10 does
slowly form 11 by transfer-epititanation of free propylene
and thus undergoes the Kulinkovich reaction.
A further pertinent observation made with respect to this
reaction at 25 °C is that approximately half (52 %) of the
ester 1a was recovered as isopropyl benzoate (25). A pos-
sible route to this unexpected product will be given below.
Modes of Forming Titanacyclopropane Intermediates
Scheme 4
Although the foregoing studies offer an experimental,
chemical-trapping approach for detecting any titanacyclo-
propane or related three-membered titanocyclic intermedi-
ates in the Kulinkovich reaction or its modifications, the
question of the mode of formation of such titanocycles
from R Ti(ORЈ) remains to be addressed. With reference
to the Kulinkovich alkyl-alkoxide intermediate 7, one could
inquire whether 7 spontaneously decomposes into 8, which
then epimetallates ethylene to yield 9 or whether ethylene
coordinates with 7 and thereby induces transfer-epimetall-
ation by elimination of ethyl groups from 7 as suggested
in Scheme 6.
However, from the ester experiment, no cyclopropanol 19
was formed and from the benzonitrile experiment, the ni-
trile was recovered unchanged. Therefore, the unavoidable
conclusion is that 10 does not decompose to generate inter-
mediate 11 under these experimental conditions. As a cor-
ollary it then follows that the reagent responsible for the
transfer of the Ti(OiPr) unit to allylic halides, olefins or
acetylenes in Sato’s experiments (Scheme 2) is the diisopro-
pyltitanium() diisopropoxide (10) itself. With all these un-
2
2
2
saturated substrates, the reaction pathway may rather be the
well-documented transfer-epimetallation,[
1,17]
in which π-
coordination of the ligand (in Scheme 5, the π-bond of the
allylic halide) to titanium forms an octahedral transition
state 24.
Scheme 6
In order to examine the mode of ring formation,
(
iPr) Ti(OiPr) (10) was treated with ethylene gas at Ϫ78
2 2
°
C for 30 min (as shown above, at Ϫ78 °C 10 is stable to
decomposition). Subsequent workup gave 1-phenylcyclo-
propanol (27) in a 70 % yield without any evidence for the
formation of 2-methyl-1-phenylcyclopropanol (19). In a
parallel experiment, Et Ti(OiPr) was treated with propyl-
2
2
ene gas at Ϫ78 °C for the same amount of time but sub-
sequent hydrolytic workup, as before, gave 27 in only a 55 %
yield. This outcome shows that ethylene is able to undergo
transfer-epimetallation with 10 via the transition state 26
Scheme 5
Confirmation of the stability of 10, in the absence of any
π-bonded substrate, in the temperature range of Ϫ78 °C to
ϩ25 °C comes from the persistence of the pale yellow to
light brown color of the solution of 10 in this temperature
range. If such an ethereal solution were to have turned
black, this would indicate the decomposition of 10 to
[Equation (4)]. On the other hand, apparently for steric re-
asons operative in an octahedral transition state like 26,
propylene is unable to undergo a similar transfer-epimetall-
ation by 7 at such low temperatures (cf. infra).
[18]
Ti(OiPr) (8) at room temperature (ca. 25 °C).
Further-
2
more, when an ethereal solution of 10 at 25 °C was treated
with methyl benzoate (1a) and the resulting mixture was
allowed to react for 7 h at this temperature, hydrolytic
workup gave a 48 % yield of cis-2-methyl-1-phenylcyclopro-
panol (19) [Equation (3)].
(
4)
Eur. J. Org. Chem. 2003, 4721Ϫ4727
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4723

J. J. Eisch, A. A. Adeosun, J. N. Gitua
SHORT COMMUNICATION
This finding supports the proposal that the Kulinkovich (8) or by transfer with iPr Ti(OiPr) (10) to form 29, which
2
2
reaction is initiated by the presence of free ethylene neces- by a reversible redox process could lead to 31, via 30 and
sary for the initial transfer epimetallation. Thereafter, the thus to ester interchange (Scheme 7). The importance and
ethylene by-product formed in Equation (1) perpetuates the utility of intermediates like 29 and 30 will become evident
reaction. Several experimental observations support the im- in our future report on reductive dimerizations of esters by
[
21]
portance of ethylene for initiation: 1) higher yields of cyclo- transition metal reagents.
propanols are obtained (Ͼ 80 %) when the ethyl Grignard
reagent and Ti(OiPr) are used in a 3:1 ratio. The extra
4
IV
Conclusions
equivalence of Grignard reagent with a Ti catalyst can
readily generate ethylene;[ 2) when a 2:1 ratio of ethyl
20]
Grignard reagent to Ti(OiPr) is employed, yields of the
When a 3:1 mixture of ethylmagnesium chloride and ti-
4
cyclopropanol are lower (30Ϫ35 %). However, yields in tanium() isopropoxide was warmed from Ϫ78 °C to ϩ25
such experiments can be enhanced (to 60 % and higher) if °C, it turned black, signaling the formation of Ti(OiPr) (8)
2
ethylene gas is introduced both at Ϫ78 °C and during and the usual by-products, ethylene and ethane (the typical
warming. In contrast, propylene is incapable of initiating Kulinkovich procedure, part e; cf. Scheme 1). Since sub-
the Kulinkovich reaction of (iPr) Ti(OiPr) , which as noted sequent addition of methyl benzoate (1a) did not lead to
2
2
above occurs slowly at 25 °C or at higher temperatures the formation of the cyclopropanol 27, we can conclude
Equation (3)]. The yield obtained if propylene gas is intro- that 8 is unable to epititanate directly the ethylene present
duced at 25 °C, 45 %, is not higher than that obtained with- (formed either from 7 or from the excess EtMgCl pre-
[
[
20]
out addition (48 %). Steric hindrance may be the cause of sent ) to form the titanacyclopropane 9.
propylene’s lower reactivity in such transfer epimetallation
Since either a third equivalent of EtMgX (the typical
cf. supra).
Kulinkovich procedure, part a) or addition of ethylene gas
part c) at Ϫ78 °C enhanced the yield of the cyclopropanol
(
(
Ester Interchange as Side Reaction in Kulinkovich
Processes
obtained, in comparison with that produced using a 2:1
mixture of titanium() isopropoxide and EtMgCl (part b),
we conclude that the free ethylene (as such or via
EtMgX ) fosters the formation of the essential titanacy-
clopropane 9 by the transfer-epititanation specifically de-
picted for the conversion of 7 to 9 in Scheme 6 and occur-
ring via a transition state like 26. The ethylene set free by
such epititanation of 7 perpetuates the reaction.
Under experimental conditions where certain dialkylti-
tanium() dialkoxides do not or cannot undergo a rapid
Kulinkovich reaction, it has been found that the starting
methyl benzoate (1a) was converted into about a 1:1 mix-
ture of 1a and isopropyl benzoate (25) (cf. supra). Since
MgBr(OiPr) (28) is the by-product formed in producing
Et Ti(OiPr) according to Scheme 1, we first examined the
[
20]
Diisopropyltitanium() diisopropoxide (10) has been
shown to be kinetically stable in the temperature range of
Ϫ78 °C to ϩ25 °C and hence we conclude that 10 does not
decompose within this temperature range into the titanacy-
clopropane 11, as assumed in the work reported by the Sato
2
2
action of 28 on 1a as the possible path of ester inter-
change [Equation (5)].
[
7,13,14]
group.
Reliable chemical trapping agents for titanacy-
clopropanes, such as esters and nitriles, failed to detect the
presence of 11 when 10 was warmed from Ϫ78 °C to ϩ25
(
5)
°
C.
However, no ester exchange was found to occur between
Although 10 does not undergo the Kulinkovich cyclopro-
1
a and 28 in an ethereal solution at room temperature. But panol reaction with methyl benzoate (1a) below 25 °C, it
when 1a was treated with Ti(OiPr) (8) in diethyl ether at does form the cyclopropanol 19 slowly at room temperature
2
ambient temperatures (ca. 25 °C), ester interchange took (ca. 25°C) and at a rate which remains unchanged on ad-
place readily.
dition of propylene. We conclude that the rate limiting step
Accordingly, as an attractive alternative, we propose a is the transfer-epititanation of propylene by 10. Steric hin-
reversible epimetallation of 1a either directly with Ti(OiPr)2 drance in a transition state similar to 26 (involving propyl-
Scheme 7
4724
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4721Ϫ4727

Mechanism of the Kulinkovich Cyclopropanol Synthesis
SHORT COMMUNICATION
ene instead of ethylene) is the likely barrier retarding the re- c) Two Equivalents of EtMgCl and Added Ethylene: Ethylmag-
nesium chloride (20 mmol) and titanium() isopropoxide
action.
(
10 mmol) in a ratio of 2:1 were mixed in diethyl ether (60 mL) at
A further indication of this crucial steric factor emerges
from the comparative transfer-epitatanations of ethylene
with 10 and of propylene with 7. The former reaction oc-
curs smoothly to produce 9 and subsequently 27 (69 %)
upon workup , without any indication of the formation of
Ϫ78 °C. After 90 min methyl benzoate (10 mmol, 1 equiv.) was ad-
ded, as in part b. Thereafter, ethylene gas at 40 psi was bubbled
through the cold solution for a period between 30 min and 2 h,
before the reaction mixture was warmed to room temperature (ca.
1
2
5°C) and subjected to H NMR spectral analysis. The yield of 1-
19 (the corresponding epititanated product from the reac-
phenyl-1-cyclopropanol (27) now ranged from 60 % (30 min ethyl-
tion with propylene). In the latter reaction, propylene with ene treatment) to over 75 % (2 h).
7
, none of cyclopropanol 19 was detected but only the auto-
d) 2.5 Equivalents of EtMgCl and Added Propylene (attempted
transfer-epimetallation): To a stirred solution of titanium() isop-
ropoxide (10 mmol) and methyl benzoate (10 mmol) in diethyl ether
titanated product of ethylene from 7, namely 9, was pro-
duced and transformed into 27 (55 %). We conclude that
the transfer-epititanation transition state 26 is at a much
lower level in energy than the similar, sterically hindered
transition state involving propylene coordination.
(60 mL) at Ϫ78 °C was added a solution of ethylmagnesium chlo-
ride (25 mmol) in diethyl ether (15 mL). After 10 min at Ϫ78 °C
the red-brown mixture was treated with a current of propylene gas
at 40 psi for 30 min. The resulting reaction mixture was warmed to
room temperature over 3 h and then subjected to the usual hydro-
lytic workup and NMR spectral analysis. 1-Phenyl-1-cyclopropanol
Experimental Section
(
27) was obtained in a 55 % yield, with the remainder being reco-
Instrumentation, Analysis and Starting Reagents: All reactions were
carried out under a positive pressure of anhydrous, oxygen-free ar-
gon. All solvents employed with organometallic compounds were
dried and distilled from a sodium metal-benzophenone ketyl mix-
vered ester.
4
e) Three Equivalents of EtMgCl and One Equivalent of Ti(OiPr) at
25 °C: Formation of Titanium(II) Isopropoxide: A red-brown mix-
ture of ethylmagnesium chloride (30 mmol) and titanium() isop-
ropoxide (10 mmol) in diethyl ether (80 mL) was prepared at Ϫ78
[
22]
ture prior to use.
The IR spectra were recorded with a
PerkinϪElmer instrument (model 457) and samples were measured
°
C. When the mixture was warmed to room temperature (3 h), it
turned black, indicating the decomposition of Et Ti(OiPr) (7) into
Ti(OiPr) (8). Methyl benzoate (10 mmol) was then added and the
mixture was allowed to stirred for 7 h at ambient temperature (ca.
5°C). The usual workup gave only recovered methyl benzoate (no
1
either as mineral oil mulls or as KBr films. The NMR spectra ( H
and 13C) were recorded with a Bruker spectrometer (model EM-
2
2
2
3
60) and tetramethylsilane (Me
ard. The chemical shifts reported are expressed on the δ scale in
parts per million (ppm) from the Me Si reference signal. The GC/
MS measurements and analyses were performed with
4
Si) was used as the internal stand-
2
4
isopropyl benzoate 25: cf. infra for a comparable reaction with
EtMgBr, where 25 was found).
a
HewlettϪPackard GC 5890/HewlettϪPackard 5970 mass-selective-
detector instrument. The gas chromatographic analyses were per-
formed with a HewlettϪPackard instrument (model 5880) provided
Modified Kulinkovich Procedure. Reactions with Diisopropyl-
titanium(IV) Diisopropoxide. a) Methyl Benzoate (stoichiometric
with a 2-m OV-101 packed column or with a HewlettϪPackard mode): To a stirred solution of titanium() isopropoxide (2.84 g,
instrument (model 4890) having a 30 m SE-30 capillary column.
Melting points were determined with a ThomasϪHoover Unimelt
capillary melting point apparatus and are uncorrected.
10 mmol) and methyl benzoate (1.36 g, 10 mmol) in diethyl ether
(80 mL) at Ϫ78 °C was added a solution of isopropylmagnesium
chloride (20 mmol) in diethyl ether (30 mL). The brown mixture
was stirred for 90 min before being quenched at Ϫ78 °C with meth-
anol. Usual hydrolytic workup of the quenched mixture at 25 °C
gave only recovered methyl benzoate. A similar reaction was con-
ducted by slowly addiing a solution of isopropylmagnesium bro-
mide (20 mmol) in diethyl ether (30 mL) to a solutionof ti-
tanium() isopropoxide (6.25 mmol) and methyl benzoate
The Typical Kulinkovich Procedure. a) Three Equivalents of EtMgCl
(usual modus operandi): To a stirred solution of titanium() isopro-
poxide (2.84 g, 10 mmol) in diethyl ether (80 mL) at Ϫ78 °C was
added 15 mL of a 2.0  solution of ethylmagnesium chloride
(
2.67 g, 30 mmol) in diethyl ether. The red-brown mixture was
stirred at Ϫ78 °C for 90 min, before addition of methyl benzoate
1.36 g, 10 mmol). The reaction mixture was warmed to room tem-
(
6.25 mmol) in diethyl ether (30 mL) at 25 °C. The red-brown reac-
(
tion mixture formed was stirred for 7 h at 25 °C; hydrolytic workup
perature (ca. 25°C) and was stirred for a further 10 h. Hydrolytic
workup with 1  aqueous HCl, extraction of the organic products
1
and H NMR spectral analysis revealed the formation of cis-2-
methyl-1-phenyl-1-cyclopropanol (19) (48 %) and isopropyl benzo-
ate (25) (25 %), together with recovered methyl benzoate (25 %).
2 4
using diethyl ether, drying of extracts over Na SO and evaporation
of volatiles gave an organic residue, which by 1H NMR spectral
analysis showed signals for 1-phenyl-1-cyclopropanol (27) (90Ϫ95
b) Methyl Benzoate with Added Propylene: A similar reaction at 25
°C was conducted as in part a, except that propylene gas was slowly
passed into the reaction mixture during the 7 h reaction period.
%
). The only other significant product, present in the crude reac-
tion mixture with this Kulinkovich procedure and all other modifi-
cations leading to 27 was 1-phenyl-1-propanol to the extent of Usual hydrolytic workup provided 19 in a 45 % yield, a yield essen-
[
23]
5
Ϫ10 %. A pure sample of 27 could be obtained by distillation
tially unchanged from that of 19 (48 %) obtained without addition
of propylene. From these results it is clear that propylene does not
promote the formation of the titanacyclopropane 11 by transfer ep-
ititanation.
under reduced pressure: b.p. 93Ϫ95 °C (30 Torr).
b) Two Equivalents of EtMgCl (stoichiometric mode): A trial reac-
tion carried out in an identical manner to that described in part a,
except that two equivalents of EtMgCl (20 mmol) were employed.
H NMR spectral analysis of the organic liquid recovered showed titanium() isopropoxide (10 mmol) and methyl benzoate
c) Methyl Benzoate with Added Ethylene: To a stirred solution of
1
the presence of 1-phenyl-1-cyclopropanol (27) (34 %) and methyl
benzoate (1a) (66 %).
(10 mmol) in diethyl ether (50 mL) at Ϫ78 °C was passed a current
of dry ethylene gas at 40 psi for 30 min. A solution of isopro-
Eur. J. Org. Chem. 2003, 4721Ϫ4727
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4725

J. J. Eisch, A. A. Adeosun, J. N. Gitua
SHORT COMMUNICATION
pylmagnesium chloride (25 mmol) in diethyl ether (10 mL) was
then added slowly by syringe to the reaction mixture maintained
at Ϫ78°C for 30 min. The reaction mixture was warmed to room
temperature (ca. 25 °C) over 1 h, followed by the usual hydrolytic
workup and H and C NMR spectral analyses, which showed the
formation of 1-phenyl-1-cyclopropanol (27) (43 %) and 1-phenyl-
[
[
1]
2]
J. J. Eisch, J. N. Gitua, Organometallics 2003, 22 24Ϫ26.
For an insightful overview of the Kulinkovich reaction for or-
ganic synthesis: B. Breit, J. Prakt. Chem. 2000, 342, 211Ϫ214.
For more extensive reviews, also covering the applications of
1
13
[3]
[3a]
other titanium reagents in organic synthesis:
O. G. Kulinko-
vich, A. de Meijere, Chem. Rev. 2000, 100, 2789. [ F. Sato,
H. Urabe, S. Okamoto, Chem. Rev. 2000, 100, 2835.
O. G. Kulinkovich, S. V. Sviridov, D. A. Vasilevski, T. S. Pri-
tyckaja, Zh. Org. Khim. 1989, 25, 2245.
3b]
1
%
-propanol (12 %), together with recovered methyl benzoate (45
). No 1H or C NMR spectral signals characteristic of cis-2-
13
[
[
[
[
[
[
4]
5]
6]
7]
8]
9]
methyl-1-phenyl-1-cyclopropanol (19) were present.
O. G. Kulinkovich, S. V. Sviridov, D. A. Vasilevski, Synthesis
A reaction similar to that described above (the same scale) was
conducted , except that 5 mmol of methyl benzoate was employed.
The reaction gave an increased yield of 1-phenyl-1-cyclopropanol
1
990, 234.
E. J. Corey, S. A. Rao, M. C. Noe, J. Am. Chem. Soc. 1994,
16, 9345Ϫ9346.
1
(27) (69 %) and of 1-phenyl-1-propanol (21 %), with a lower recov-
A. Kasatkin, T. Nakagawa, S. Okamoto, F. Sato, J. Am. Chem.
Soc. 1995, 117, 3881Ϫ3882.
ery of methyl benzoate (10 %).
2
J. Lee, H. Kim, J. K. Cha, J. Am. Chem. Soc. 1996, 118,
Attempts to Detect the Presence of the (η -Propene)Titanium Diiso-
propoxide 11 in the Thermal Decomposition of Diisopropyl-
titanium(IV) Diisopropoxide (10): To a stirred solution of ti-
tanium() isopropoxide (10 mmol) in diethyl ether (80 mL) at Ϫ78
4198Ϫ4199.
V. Chaplinski, A. de Meijere, Angew. Chem. Int. Ed. Engl. 1996,
35, 413Ϫ414.
[10]
A. de Meijere, C. A. Williams, A. Kourdioukov, S. V. Sviridov,
V. Chaplinski, M. Kordes, A. I. Savachenko, C. Stratmann, M.
Noltemeyer, Chem. Eur. J. 2002, 8, 3789Ϫ3801.
J. J. Eisch, J. N. Gitua, P. O. Otieno, X. Shi, J. Organomet.
Chem. 2001, 624, 229Ϫ238.
J. Lee, C. H. Kang, H. Kim, J. K. Cha, J. Am. Chem. Soc.
1996, 118, 291Ϫ292.
A. Kasatkin, F. Sato, Tetrahedron Lett. 1995, 36, 6079Ϫ6082.
K. Harada, H. Urabe, F. Sato, Tetrahedron Lett. 1995, 36,
3203.
J. J. Eisch, J. Organomet. Chem. 2001, 617Ϫ618, 148Ϫ157.
J. J. Eisch, F. A. Owuor, P. O. Otieno, Organometallics 2001,
20, 4132Ϫ4134.
°
(
°
C
was added
20 mmol) in diethyl ether (30 mL). After 90 min of stirring at Ϫ78
C, benzonitrile (1.03 g, 10 mmol) was introduced. The reaction
a solution of isopropylmagnesium bromide
[11]
mixture was allowed to attain room temperature (ca. 25°C) before
being heated at reflux for 5 h. The usual hydrolytic workup led only
to the recovery of benzonitrile.
[
[
12]
13]
A similar reaction involving the admixture and reaction of isopro-
pylmagnesium bromide (20 mmol) and titanium() isopropoxide
[14]
[
[
15]
16]
(10 mmol) in diethyl ether (80 mL) at 20 °C, followed by the ad-
dition of benzonitrile and heating at reflux for 3 h, likewise led to
the complete recovery of benzonitrile, without trace of any car-
bonyl products.
[
[
17]
18]
J. J. Eisch, J. N. Gitua, Organometallics 2003, 22, 24Ϫ26.
The dialkyltitanium() diisopropoxides, generated in alkanes
or in ethers by the interaction of two equivalents of RLi with
Ester Interchange. a) Attempted Ester Interchange with Magnesium
Chloride Isopropoxide: A solution of magnesium chloride isopro-
poxide (20 mmol) in diethyl ether (50 mL) was generated by treat-
ing a solution of ethylmagnesium chloride (20 mmol) in diethyl
ether (10 mL) with anhydrous isopropyl alcohol (1.20 g, 20 mmol)
as a solution in diethyl ether (40 mL). To the resulting solution at
one equivalent of Ti(OiPr) , decompose above Ϫ78 °C into
4
black Ti(OiPr)
Sect. for the Ti(OiPr)
or EtMgBr and one equivalent of Ti(OiPr)
2
: Ref.[11]; cf. also procedures given in the Exp.
2
formed from two equivalents of EtMgCl
, when such mix-
4
tures were then warmed to room temperature.
[19]
The generation and chemical trapping of such titanacyclopro-
pane intermediates have been achieved recently with two differ-
ent olefins: a) The transfer-epimetallation of 1-butene by Bu2-
Ti(OiPr) , trapping with benzonitrile and hydrolysis gave butyl
2
phenyl ketone in a good yield (ref.[11]); and b) The transfer-
25 °C was then added methyl benzoate (10 mmol). The mixture was
stirred for 10 h at ambient temperature (ca. 25°C) and then treated
as described above. The NMR spectral analyses of the crude or-
ganic product showed only the presence of methyl benzoate with-
out trace of isopropyl benzoate.
epimetallation of propylene by tBu TiCl , followed by anal-
2
2
ogous treatment with benzonitrile, gave upon hydrolytic
workup, a 4:1 ratio of ketones 22:23 (ref.[1]). These precedents
verify that the present test for the possible formation of 11
is valid.
Titanium-catalyzed magnesium hydride halide transfer from
ethyl or isopropyl Grignard reagents to olefins or func-
tionalized olefins leads to the liberation of ethylene or propyl-
ene, respectively. Apparently the titanium catalyst (of unknown
oxidation state) promotes the following equilibrium:
b) Ester Interchange with Titanium(II) Isopropoxide (8): To a stirred
solution of titanium() isopropoxide (12.5 mmol) in of diethyl
ether (15 mL) at Ϫ78 °C was added a solution of ethylmagnesium
bromide (40 mmol) in diethyl ether (40 mL). When the initial red-
brown mixture had warmed to 25 °C over 10 h, it turned black.
Methyl benzoate (12.5 mmol) was then injected into the black mix-
ture and stirring at ambient temperature continued for 7 h. The
usual workup and analyses showed that the crude material con-
tained only methyl benzoate and isopropyl benzoate 25 as approxi-
mately a 50:50 (Ϯ5) mixture.
[20]
[
20a]
G. D. Cooper, H. L. Finkbeiner, J. Org. Chem. 1962, 27,
1493. [
2
1
20b]
H. L. Finkbeiner, G. D. Cooper, J. Org. Chem. 1972,
Acknowledgments
7, 3395. [
60, C8ϪC12.
20c]
J. J. Eisch, J. E. Galle, J. Organomet. Chem. 1978,
This research has been conducted with the financial support of
grants from the Solvay Corporation of Brussels, Belgium and the
Boulder Scientific Company of Mead, Colorado, as well as with a
Senior Scientist Award to JJE from the Alexander von Humboldt
Stiftung of Bonn, Germany. In addition, we are indebted to our
colleagues, Peter O. Otieno and Paul O. Fregene, for significant
experimental and mechanistic insights.
[
[
21]
22]
J. J. Eisch, J. N. Gitua, A. A. Adeosun: in ongoing studies we
have observed that 2:1 mixtures of RLi and Ti(OiPr) convert
methyl benzoate into products hydrolyzable into principally
benzoin with minor amounts of benzil.
A detailed description for conducting organometallic reactions
in a safe and reproducible manner is given in: J. J. Eisch, Or-
4
4726
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4721Ϫ4727

Mechanism of the Kulinkovich Cyclopropanol Synthesis
SHORT COMMUNICATION
ganometallic Syntheses, Vol. 2, Academic Press, New York,
1981, pp. 1Ϫ84.
The origin of the 1-phenyl-1-propanol side-product in this Kul-
inkovoich reaction is still under investigation. However, we can
rule out the reductive scission of the cyclopropane ring in 1.
When a pure sample of 27, free of any 1-phenyl-1-propanol,
was treated with a mixture formed by reacting EtMgCl
4
(3 equiv.) and Ti(OiPr) (1 equiv.) at Ϫ78 °C in diethyl ether
[
23]
and then allowing the resulting system to warm to and react
at 25 °C, no ring cleavage reduction was observed.
Received July 18, 2003
Eur. J. Org. Chem. 2003, 4721Ϫ4727
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4727