Demonstration of a phosphazirconocene as a catalyst for the ring opening of epoxides with TMSCI

Article abstract of DOI:10.1021/ol034816r

(Matrix presented) In this study, it was demonstrated for the first time that a phosphazirconocene catalyzes the ring opening of epoxides with TMSCI. This reactivity leads to a facile preparation of chlorohydrins. The late transition metal Fe analogue was found to catalyze the reaction at rates and stereoselectivity comparable to those of the Zr complex.

Full text of DOI:10.1021/ol034816r

ORGANIC
LETTERS
2003
Vol. 5, No. 14
2543-2545
Demonstration of a Phosphazirconocene
as a Catalyst for the Ring Opening of
Epoxides with TMSCl
Li-Sheng Wang† and T. Keith Hollis*
Department of Chemistry, UniVersity of California-RiVerside,
RiVerside, California 92521
keith.hollis@ucr.edu
Received May 12, 2003
ABSTRACT
In this study, it was demonstrated for the first time that a phosphazirconocene catalyzes the ring opening of epoxides with TMSCl. This
reactivity leads to a facile preparation of chlorohydrins. The late transition metal Fe analogue was found to catalyze the reaction at rates and
stereoselectivity comparable to those of the Zr complex.
Upon beginning our investigations of the chemistry of early
transition metal phosphametallocenes, we decided to evaluate
the nucleophilicity of the phosphorus lone pair. Garrett and
Fu had reported recently the ring opening of epoxides in the
presence of TMSCl catalyzed by a phosphaferrocene.1 The
reaction was first reported to be catalyzed by PPh3.2 A study
allowing the direct comparison of a phosphazirconocene and
a phosphaferrocene as a catalyst for this reaction was
designed, and the preliminary results are reported here.
conocene dichloride 1, (TMP)2ZrCl2, would have similar
reactivity. The d0 zirconium(IV) fragment of 1 is significantly
more electron withdrawing than the d6 Fe(II) fragment in
the analogue, 2,2′,3,3′,4,4′,5,5′-octamethyl-1,1′-diphospha-
ferrocene 2, (TMP)2Fe. This difference can be seen in the
145 ppm downfield shift in the 31P resonance of 1 compared
to that of 2, (1,31P NMR 88.0 ppm; 2, 31P NMR -57.0 ppm).3
We report here the catalytic activity of 1 and 2, which
allowed evaluation of the metal on the reactivity of the
phospholyl ligand in the ring opening of epoxides.
There has been much interest in the ring opening of
epoxides, and several examples of enantioselective nucleo-
philic catalysis have been reported recently.4,5 Denmark has
shown that chiral phosphoramides will nucleophilically and
enantioselectively catalyze the ring opening of epoxides in
the presence of SiCl4.6 Jacobsen7 and Nugent8 have each
(3) Nief, F.; Mathey, F.; Richard, L. Organometallics 1988, 7, 921-
926.
We were interested in demonstrating the nucleophilic
nature of the phosphorus atom in the phosphazirconocene
1. Despite Fu’s report, it was not assured that the Zr
analogue, 2,2′,3,3′,4,4′,5,5′-octamethyl-1,1′-diphosphazir-
(4) For recent examples of phosphorous-based nucleophilic catalyses,
see: Zhu, G.; Chen, Z.; Jiang, Q.; Xiao, D.; Cao, P.; Zhang, X. J. Am.
Chem. Soc. 1997, 119, 3836-3837. Vedejs, E.; Dauglis, O. J. Am. Chem.
Soc. 1999, 121, 5813-5814.
(5) Nucleophilic catalysis: Tao, B.; Ruble, C.; Hoic, D. A.; Fu, G. C. J.
Am. Chem. Soc. 1999, 121, 5091-5092.
† Current Address: Maxdem, Inc., 140 E. Arrow Rd., San Dimas, CA
91773.
(1) Garrett, C. E.; Fu, G. C. J. Org. Chem. 1997, 62, 4534-4535.
(2) Andrews, G. C.; Crawford, T. C.; Contillo, L. G., Jr. Tetrahedron
Lett. 1981, 22, 3803-3806.
(6) Denmark, S. E.; Barsanti, P. A.; Wong, K.-T.; Stavenger, R. A. J.
Org. Chem. 1998, 63, 2428-2429.
(7) Ready, J. M.; Jacobsen, E. N. J. Am. Chem. Soc. 1999, 121, 6086-
6087.
10.1021/ol034816r CCC: $25.00 © 2003 American Chemical Society
Published on Web 06/13/2003

Table 1. Ring Opening of Epoxides with TMSCI and 1 mol % Catalysta
a Reaction conditions: epoxide (0.5 M), TMSCl (0.5 M), catalyst (1 mol %), CDCl3 followed by treatment with HCl/Et20. b TMP ) 2,3,4,5-
tetramethylphospholyl. c Percent completion of uncatalyzed reaction at the longest catalytic time. d Major product illustrated first. e Time required for 90%
completion as determined by 1H NMR. f Isolated yield of chlorohydrin. g Where applicable, the stereochemical or regiochemical results obtained with each
catalyst are presented. h Uncatalyzed reaction slows dramatically, reaching only 66% after 6.5 h.
disclosed highly enantioselective catalysts for the ring
opening of epoxides, as well as elegant mechanistic work
elaborating the details of these reactions.
The ring opening of aromatic epoxides gave clean conver-
sion to the chlorohydrin. Styrene oxide showed a strong
preference for the opposite regioselectivity to that observed
in the aliphatic series, suggesting electronic control of the
ring opening.10 Styrene oxide was observed to yield chlo-
rohydrin with the chlorine at the more substituted R-carbon
atom in a ratio >20:1 (entry 3) with both catalysts. To our
knowledge, these results represent the highest regioselectivity
reported to date for nucleophilic catalytic ring opening of
epoxides in the presence of TMSCl.
The stereochemistry of ring opening had several unex-
pected and unique features. The aliphatic epoxides were
observed to yield products with exclusiVe inVersion of
stereochemistry (entries 4-7).1 In contrast to the aliphatic
epoxides, the ring opening of trans-stilbene oxide produced
a single stereoisomer with exclusiVe retention of configura-
tion. Additionally, cis-stilbene oxide produced a mixture of
stereoisomeric chlorohydrins, with net inVersion of config-
uration predominating. This inversion yields a predominant
stereochemistry identical to that derived from trans-stilbene
oxide. The stereochemistry of ring opening was confirmed
Presented in Table 1 are the results of combining 1 mol
% 1 or 2 and 1.0 equiv of TMSCl with various epoxides in
CDCl3 at room temperature followed by cleavage of the
trimethylsilyl group with HCl. We note that 1 and 2 catalyze
the ring opening of epoxides at comparable rates, despite
the large difference in the 31P NMR shift (vide supra). The
largest difference in rate was by a factor of 6 in favor of Fe
catalyst 2 (entry 1). In other cases, the Zr catalyst 1 was
faster (entry 2). Typically, there was a significant rate
enhancement when using catalyst 1 or 2. The exception is
cis-2-butene oxide (entry 4), which in the first hour reacts
almost as fast without catalyst. After the first hour, the
uncatalyzed reaction slows dramatically, reaching only 66%
at 6.5 h. cis-Epoxides reacted notably faster than trans-
epoxides (entries 1 vs 2 and 4 vs 5). The stereoselectivity
and regioselectivity of the ring opening are generally similar
for these catalysts. The exception is the reaction of styrene
oxide where the Fe catalyst 2 offers significantly higher
regioselectivity (49:1 vs 23:1).9
(9) Determined by 1H NMR (500 MHz) calibrated against the 13C
sidebands and authentic samples.
(8) McCleland, B. W.; Nugent, W. A.; Finn, M. G. J. Org. Chem. 1998,
63, 6656-6666.
(10) For an example of similar results, see: Afonso, C. A. M.; Vieira,
N. M. L.; Motherwell, W. B. Synlett 2000, 382-384.
2544
Org. Lett., Vol. 5, No. 14, 2003

by in situ treatment of the silylated chlorohydrin with TBAF
to regenerate the epoxide (>95%). We also noted the
formation of trans-stilbene oxide (<5%) during the reaction
of cis-stilbene oxide with catalysts 1 and 2. While the
formation of trans-stilbene oxide has important implications
for the mechanism of the reaction, it does not account for
the major stereoisomer observed. trans-Stilbene oxide does
not react sufficiently fast to account for the reaction rate
observed for cis-stilbene oxide (compare entries 1 and 2).
Previously, a mechanism was proposed for this reaction that
invoked nucleophilic attack by P at Si with the chloride then
nucleophilically opening the epoxide.1 Such a mechanism
does not account for the stereochemical outcome of the
reaction of stilbene oxides. Therefore, a mechanistic alterna-
tive that accounts for these results but does not necessi-
tate trans-stilbene oxide as the common intermediate is
required. Studies are underway to elucidate this mecha-
nism.
identical conditions. No reaction was observed in these
control experiments (Scheme 1).
Scheme 1
The results presented herein demonstrate that the Zr
phospholyl complex 1 is comparable in reactivity to the Fe
analogue 2 despite the significant difference in 31P NMR
signals (vide supra). The necessity of the P atom for catalysis
was demonstrated (Scheme 1). Several mechanistic questions
arise from this study, and our investigation of the mechanism
will be reported shortly.
We found that ring opening with PPh3 was comparable to
or slightly faster than catalysis with phospholyl complexes
1 or 2.2 The rates of reaction that we observed with our
catalysts were also comparable with those reported by Garrett
and Fu.11
It was established that the P atom of 1 and 2 was
efficacious in these systems by attempting ring opening of
cis-stilbene oxide with Cp2ZrCl2 or FeCp2 under otherwise
Acknowledgment. The University of California-River-
side Department of Chemistry and the The University of
California Cancer Research Coordinating Committee are
acknowledged for financial support of this work. Acknowl-
edgment is made to the donors of the Petroleum Research
Fund, administered by the American Chemical Society, for
partial support of this work. We thank Professor Steven
Weinreb of The Pennsylvania State University for insightful
discussions regarding this work.
(11) Garrett and Fu employed 5 mol % vs our 1 mol % for comparable
rates; see ref 1.
OL034816R
Org. Lett., Vol. 5, No. 14, 2003
2545