The reaction of dialkoxytitanacyclopropanes and dialkoxytitanacyclopropenes with carbon dioxide

Article abstract of DOI:10.1039/b201527g

Dialkoxytitanacyclopropanes and dialkoxytitanacyclopropenes react with one molecule of carbon dioxide to afford diversely substituted carboxylic acids after hydrolysis.

Full text of DOI:10.1039/b201527g

View Article Online / Journal Homepage / Table of Contents for this issue
The reaction of dialkoxytitanacyclopropanes and
dialkoxytitanacyclopropenes with carbon dioxide
Yvan Six
1
Institut de Chimie des Substances Naturelles, U.P.R. 2301 du C.N.R.S., Avenue de la Terrasse,
91198 Gif-sur-Yvette, France. E-mail: Yvan.Six@icsn.cnrs-gif.fr; Fax: 33 (1) 69 07 72 47;
Tel: 33 (1) 69 82 30 83
Received (in Cambridge, UK) 11th February 2002, Accepted 21st March 2002
First published as an Advance Article on the web 3rd April 2002
Dialkoxytitanacyclopropanes and dialkoxytitanacyclo-
propenes react with one molecule of carbon dioxide to
afford diversely substituted carboxylic acids after
hydrolysis.
of Ϫ30 ЊC did not improve the yield, but made the reaction
slower. The presence of non-deuterated cyclopentylcarboxylic
acid probably stemmed from the reaction of remaining
cyclopentylmagnesium chloride with CO2.
Intermediate complex 3 probably reacts with one molecule of
carbon dioxide to yield the new species 4, which delivers 2 upon
hydrolysis or alternatively, the β-hydroxy acid 5 when oxygen is
added as the second reagent (Scheme 3).
Dialkoxytitanacyclopropanes are putative intermediates in the
Kulinkovich cyclopropanation reaction.1 They are most con-
veniently generated from titanium alkoxides2 and Grignard
reagents (Scheme 1, A). When this procedure is run in the
Scheme 1
Scheme 3
Solutions of complex 3 can thus be prepared in about 50%
yield by simply mixing cyclopentylmagnesium chloride and
titanium tetraisopropoxide in diethyl ether at Ϫ70 ЊC, warming
to Ϫ30 ЊC in 5 minutes, then cooling again to Ϫ70 ЊC. † We
believe that the formation of 3 is faster than its decomposition
at Ϫ30 ЊC if the concentration of Grignard reagent is sufficient,
although both processes seem to occur simultaneously. The
choice of diethyl ether as the solvent proved to be essential:
in diisopropyl ether, the process was much less efficient, while in
THF it did not proceed.10
The reaction was repeated in the presence of oct-4-yne
under otherwise identical conditions. The major product was
still deuterated acid 2, while vinyl carboxylic acid 6 arising from
the reaction of diisopropyloxy(η2-octyne)titanium 7 with CO2
was produced in low yield (Scheme 4).
presence of internal alkynes, dialkoxytitanacyclopropenes are
formed3 (Scheme 1, B). The chemistry of both types of com-
plex is very rich.1,4 However, to the best of our knowledge, their
reaction with carbon dioxide has not been investigated so far,
although examples involving bis(cyclopentadienyl)titanacyclo-
propenes are known.5
Dialkoxytitanacyclopropanes are normally generated in the
presence of the requisite electrophilic trap, a carboxylic ester (in
the original Kulinkovich reaction) or amide (in the extension
discovered by de Meijere et al.1). Because of the well-known
reaction of Grignard reagents with CO2,6 this standard pro-
cedure could not be applied, and a pre-formed solution of
titanacyclopropane was required.
Marek et al.7 have already reported the preparation of diiso-
propyloxy(η2-propene)titanium 1 in diethyl ether at Ϫ50 ЊC,8
albeit with unsatisfactory reproducibility. We chose to work
with the cyclopentylmagnesium chloride–titanium tetraiso-
propoxide system9 because we thought side-reactions might be
slower than in the case of 1.
When a mixture of cyclopentylmagnesium chloride and
titanium tetraisopropoxide in diethyl ether was rapidly warmed
from Ϫ70 ЊC to Ϫ30 ЊC, then treated with carbon dioxide
followed by D2O, β-deuterated cyclopentylcarboxylic acid 2
was produced (Scheme 2). Best results were obtained with
2.4 equivalents of Grignard reagent. Working at Ϫ40 ЊC instead
Scheme 4
The ligand exchange reaction appeared to be slower than
the formation of intermediate 3 at Ϫ30 ЊC, and the reaction
time was thus extended to 15 minutes. Using 2.0 equivalents
of cyclopentylmagnesium chloride instead of 2.4 relative to
Ti(OiPr)4, the reaction was found to be much cleaner, and best
Scheme 2
DOI: 10.1039/b201527g
J. Chem. Soc., Perkin Trans. 1, 2002, 1159–1160
This journal is © The Royal Society of Chemistry 2002
1159

View Article Online
Table 1
and the reaction with CO2 is observed again. Improvement of
the efficacy of the carbon dioxide addition step is currently
being studied.
The extension of the process to the hydrocarboxylation of
terminal alkenes via ligand exchange from 3 was also investi-
gated. Preliminary results show that the reaction is feasible, but
further work is needed to obtain useful yields of alkylcarboxylic
acids.‡
In conclusion, dialkoxytitanacyclopropanes and dialkoxy-
titanacyclopropenes react with one molecule of carbon dioxide,
leading to carboxylic acids after hydrolysis. We hope the present
method will complement the existing procedures for the hydro-
carboxylation of alkynes,11 only a few of which use carbon
dioxide under standard pressure as the carbonyl source.12 The
possibility of extending this result to the use of carbon
disulfide, as well as trapping the intermediate five-membered
cyclic titanium complexes with other electrophiles is currently
under study in our laboratory and will be reported in due
course.
Acknowledgements
This work was funded by the C.N.R.S. We also wish to thank
Professor S. Z. Zard, Mr A. Parenty and Dr J.-M. Campagne
for the gift of several reagents and alkynes.
Notes and references
† 3 appears to be reasonably stable at Ϫ70 ЊC: when the solution was
stirred for one hour at that temperature before the CO2 addition, the
yield in deuterated 2 was 43%.
‡ For instance, 5-benzyloxypentanoic acid was obtained from 4-benzyl-
oxybut-1-ene as a single isomer, but in 8% yield only.
1 Review: O. G. Kulinkovich and A. de Meijere, Chem. Rev., 2000,
100, 2789 . Theoretical study: Y.-D. Wu and Z.-X. Yu, J. Am. Chem.
Soc., 2001, 123, 5777.
2 For an evaluation of the titanium partner in the Kulinkovich reac-
tion, see: J. C. Lee, M. J. Sung and J. K. Cha, Tetrahedron Lett.,
2001, 42, 2059.
3 K. Harada, H. Urabe and F. Sato, Tetrahedron Lett., 1995, 36, 3203.
4 F. Sato, H. Urabe and S. Okamoto, Synlett, 2000, 753; J. J. Eisch,
J. Organomet. Chem., 2001, 617–618, 148.
5 I. S. Kolomnikov, T. S. Lobeeva, V. V. Gorbachevskaya, G. G. Alek-
sandrov, Y. T. Struchkov and M. E. Vol’pin, Chem. Commun., 1971,
972; B. Demerseman, R. Mahé and P. H. Dixneuf, Chem. Commun.,
1984, 1394; H. G. Alt, G. S. Herrmann, M. D. Rausch and
D. T. Mallin, J. Organomet. Chem., 1988, 356, C53; V. V. Burkalov,
U. Rosenthal, A. I. Yanovskii, Y. T. Struchkov, O. G. Ellert,
V. B. Shur and M. E. Vol’pin, Organomet. Chem. USSR, 1989, 2,
633.
6 J. March, Advanced Organic Chemistry: reactions, mechanisms, and
structure, 5th edn., Wiley-Interscience, New York, 2001, 1215.
7 C. Averbuj, J. Kaftanov and I. Marek, Synlett, 1999, 1939.
8 The group of Sato has described the successful synthesis of silylated
diisopropyloxytitanacyclopropanes, but these are not suitable for
further ligand exchange with alkynes: R. Mizojiri, H. Urabe and
F. Sato, J. Org. Chem., 2000, 65, 6217. More recently, Eisch et al.
have reported the preparation of solutions of diisopropyloxy-
titanacyclopropanes, using a novel method: J. J. Eisch, J. N. Gitua,
P. O. Otieno and X. Shi, J. Organomet. Chem., 2001, 624, 229.
9 J. Lee, Y. G. Kim, J. G. Bae and J. K. Cha, J. Org. Chem., 1996, 61,
4878.
a Isolated yield after flash column chromatography. b Isolated by crystal-
lisation as a single isomer. c The other regioisomer was also detected in
the crude (selectivity ≈ 86 : 14). d The reaction was quenched with D2O
(deuterium incorporation ≈ 86%). e Two minor isomers were also
detected by 1H NMR of the crude (Ratio ≈ 3 : 1 : 1).
results were obtained when a deficiency of alkyne was used
(Table 1).
10 Similarly, Cha et al. had observed that the formation of a titana-
cyclopropane from cyclohexylmagnesium chloride and ClTi(OiPr)3
did not occur below Ϫ5 ЊC in THF: see Ref. 9.
The method appears to be quite general, highly stereoselec-
tive, and moderately (entry 7) to highly (entries 2 and 8) regio-
selective. However, it is poorly efficient in several cases (entries
4, 5, 8 and 9). The variable amounts of triple-bond reduction
products isolated indicate that the intermediate titanacyclo-
propenes are formed but react incompletely with carbon
dioxide or do not react at all. To explain this fact, steric
hindrance may be invoked for the examples shown in entries 5
and 8. In the case of 1-benzyloxyhex-2-yne (entry 4), we believe
the titanacyclopropene is stabilised by coordination with the
oxygen atom. When the benzyl group is replaced with a bulky
TBS group (entry 5), the oxygen loses its coordination ability
11 Via transition metal-mediated hydroformylation with CO:
J. K. Stille, in Comprehensive Organic Synthesis, eds.B. M. Trost
and I. Fleming, Pergamon Press, 1991, vol. 4, 913. Via intra-
molecular nucleophilic acyl substitution: S. Okamoto, A. Kasatkin,
P. K. Zubaidha and F. Sato, J. Am. Chem. Soc., 1996, 118, 2208.
12 For instance, Ni-catalysed electrochemical carboxylation: S. Dérien,
E. Duñach and J. Périchon, J. Am. Chem. Soc., 1991, 113, 8447.
Ni-promoted alkylative carboxylation: M. Takimoto, K. Shimizu
and M. Mori, Org. Lett., 2001, 3, 3345. Ti-catalysed hydromagnesi-
ation: F. Sato, H. Ishikawa and M. Sato, Tetrahedron Lett., 1981,
22, 85; P. J. Kocienski, C. J. Love, R. J. Whitby, G. Costello and
D. A. Roberts, Tetrahedron, 1989, 45, 3839.
1160
J. Chem. Soc., Perkin Trans. 1, 2002, 1159–1160