amount of catalyst R4X, of the corresponding undesired
R1R2R4P(O).
2 was isolated in 64% (TMSBr) and 75% (TMSI) yield.
NMR analysis of the crude material revealed the pres-
ence of both hydrolysis product diphenylphosphine oxide 3
(estimated as 15 and 12%, respectively) and oxidation
product O-methyl diphenyl phosphonate, related to unre-
acted starting material (estimated as 11 and 2%, respec-
tively).
The proposed reaction mechanism was further confirmed
by the observation that, when treated with methyl iodide,
O-trimethylsilyloxy diphenylphosphinite (prepared from
diphenylphosphine oxide 3 and hexamethyldisilazane)6 led
to phosphine oxide 2 under the same reaction conditions
(87% yield when heated overnight at 80 °C and 83% yield
when stirred for 5 days at room temperature).7 31P NMR
analyses of the crude reaction mixture (in CDCl3) did not
provide evidence of nucleophilic addition of TMSI on
phosphorus (Ph2P-Si phosphonium salts display a 31P NMR
shift at ca. 54 ppm;8 the only detected reaction interme-
diate displayed a 75.2 ppm 31P NMR shift and could be
attributed to the phosphonium salt corresponding to the
alkylation of O-trimethylsilyloxy diphenylphosphinite by
methyl iodide).9
In an effort to take advantage of this “autocatalytic”
rearrangement and to prevent use of an auxiliary but
troublesome and useless alkyl halide that lowers the final
yield, we investigated the use of a catalytic amount of
trimethylsilyl halide TMSI or TMSBr on phosphite in a
sealed reaction vessel. The aim was to form in situ the
required alkyl halide together with a particularly reactive
phosphorus(III) silyl ester (Scheme 2).4
Scheme 2. Proposed Mechanism for TMSX-Catalyzed
Rearrangement
We found, as described in Scheme 3, that when the
reaction was performed with methyl diphenyl phosphinite
Interestingly, such an Arbuzov rearrangement is not limited
to heat-sensitive and reactive diphenylphosphinites.10 Scheme
4 indicates, for instance, yields in O,O-dimethyl methylphos-
Scheme 3. Treatment of Phosphinite 1 with Trimethylsilyl
Halide
Scheme 4. Yield of Isolated Methylphosphonate 5 from
Trimethyl Phosphite 4 Rearrangement Using TMSX as a
Catalyst
1, in sealed tubes and at moderate temperature (80 °C),
such an Arbuzov rearrangement did take place with either
TMSI or TMSBr (20 mol %). No reaction was observed
with TMSCl. After overnight heating at 80 °C, cooling the
mixture to room temperature afforded a white (TMSBr) to
pale brown (TMSI) solid. 31P NMR analysis showed that
the products obtained were mainly phosphine oxide 2,
together with traces of hydrolysis product diphenylphos-
phine oxide 3. After a single recrystallization in cyclohex-
ane, white crystals of phosphine oxide 25 were obtained in
96% (TSMBr) and 92% (TMSI) yields. Interestingly, the
reaction also occurred when performed at room tempera-
ture in dry CH2Cl2 (0.5 M phosphonite 1), with 10 mol %
catalyst, albeit much slower. After 5 days, phosphine oxide
phonate 5 when trimethyl phosphite 4 is placed in the
presence of trimethylsilyl bromide or iodide.
Catalyst purity and quality proved to be a pivotal issue
for higher rearrangement yields. Use of exceedingly sensitive
trimethylsilyl iodide was particularly detrimental when it was
degraded and gave irreproducible results according to the
batch used. Thus, from a practical point of view, although
longer reaction times were required, use of trimethylsilyl
bromide was favored.
To determine the scope of this catalyzed Arbuzov rear-
rangement, various O-alkyl phosphinites were assayed
(6) Hansen, H. I.; Kehler, J. Synthesis 1999, 1925-1930.
(7) See also: Bondarenko, N. A.; Tsvetkov, E. N. J. Gen. Chem. USSR
(Engl. Trans.) 1989, 59, 1361-1364.
(4) (a) Thottathil, J. K.; Przybyla, C. A.; Moniot, J. L. Tetrahedron Lett.
1984, 25, 4737-4740. (b) Thottathil, J. K. Handbook of Organophosphorus
Chemistry; Engeln, R., Ed.; Marcel Dekker, Inc.: New York, 1992. (c)
Matqiari, M.; Georgiadis, D.; Dive, V.; Yiotakis, A. Org. Lett. 2001, 3,
659-662.
(5) Melting point (109-111°C) and spectroscopic data were in perfect
agreement with commercially available phosphine oxide and literature data;
see, for example: Wittig, G.; Maercker, A. Chem. Ber. 1964, 97, 747-
768. Even, L.; Florentin, D.; Marquet, A. Bull. Soc. Chem. Fr. 1990, 6,
758-768. Elementary analysis was also in perfect accordance (calcd for
C13H13OP: C, 72.22; H, 6.06. Found: C, 72.28; H, 6.27).
(8) Gomeyla, N. D.; Feschenko, N. G. J. Gen. Chem. USSR (Engl. Trans.)
1991, 61, 2338-2341.
(9) (a) Romanenko, V. D.; Tovstenko, V. I.; Markovskii, L. N. Synthesis
1980, 823-825. Walther, B.; Scho¨ps; Kolbe, W. Z. Chem. 1980, 20, 189-
190. (c) Laurenc¸o, C.; Villien, L.; Kaufmann, G. J. Chem. Res., Miniprints
1982, 0232-0252. (d) Colle, K. S.; Lewis, E. S. J. Org. Chem. 1978, 43,
571-574.
(10) Arbuzov, A. E.; Nikonorov, K. V. Zh. Obschch. Khim. 1948, 18,
2008.
1662
Org. Lett., Vol. 5, No. 10, 2003