Lewis acid catalyzed room-temperature Michaelis-Arbuzov rearrangement

Article abstract of DOI:10.1002/anie.200250270

The taming of the shrew! For the first time, a broadly applicable efficient room-temperature Arbuzov rearrangement is described. This reaction is accomplished through an atom-economical Lewis acid catalyzed process (see scheme, TMSOTf=trimethylsilyl trifluoromethane-sulfonate). The method has been generalized to primary and activated secondary phosphites, phosphinites, and phosphonites.

Full text of DOI:10.1002/anie.200250270

Angewandte
Chemie
Making Carbon-Phosphorus Bonds
Lewis Acid Catalyzed Room-Temperature
Michaelis–Arbuzov Rearrangement**
Pierre-Yves Renard,* Philippe Vayron, Eric Leclerc,
Alain Valleix, and Charles Mioskowski*
The Michaelis–Arbuzov rearrangement is one of the most
extensively investigated reactions in organophosphorus
chemistry, and is widely used in many fields of chemistry,
from organic synthesis to catalyst design, to prepare phos-
[
1]
phonates, phosphinates, and phosphane oxides. This rear-
rangement involves the reaction of an ester of a trivalent
1
2
3
4
phosphorus species R R PꢀOR with alkyl halides R X.
Although pentacoordinate phosphorane intermediates
[
1d]
cannot be completely excluded,
mechanism for this Michaelis–Arbuzov rearrangement
Scheme 1) involves formation of a phosphonium intermedi-
ate through nucleophilic addition of the phosphorus lone pair
the generally accepted
(
[
*] Dr. P.-Y. Renard, P. Vayron, E. Leclerc, A. Valleix
Service de Marquage MolØculaire et de Chimie Bioorganique
CEA Saclay, 91191 Gif sur Yvette Cedex (France)
Fax: (+33)1-69-08-79-91
E-mail: pierre-yves.renard@cea.fr
Dr. C. Mioskowski
Laboratoire de synth›se bioorganique (UMR 7514)
facultØde pharmacie, universitØLouis Pasteur
74, route du Rhin 67401 Illkirch-Graffenstaden (France)
Fax: (+33)3-90-24-43-06
E-mail: mioskow@aspirine.u-strasbg.fr
[**] This work was financially supported by the DØlØgation GØnØrale
pour L'Armement.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2003, 42, 2389 – 2392
DOI: 10.1002/anie.200250270
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2389

Communications
acid catalyst, most of the Lewis acids described as epoxide
ring-opening activators (and thus efficient oxygen-binding
Lewis acids) did catalyze an efficient isomerization of 1 into
methyldiphenylphosphane oxide (2). The best catalysts were
found to be BF ·OEt and trimethylsilyl trifluoromethane
Scheme 1. Mechanism of the Michaelis–Arbuzov rearrangement.
3
2
sulfonate (TMSOTf; 87 and 91% yield, respectively). Mod-
erate heating (608C) resulted in complete transformation
within an hour (see Supporting Information).
to the alkyl halide. This first reversible step is followed by
thermally induced intramolecular reaction of the conjugate
The reaction velocity proved to be moderately sensitive to
nucleophile (halide X-) on the a carbon atom of the ester Pꢀ substrate concentration and required a polar and aprotic
3
OR . The overall process gives the simultaneous formation of
solvent (CHCl > CH Cl ꢁ CH CN @ benzene, or toluene).
3
2
2
3
3
a new alkyl halide R X and a tetracoordinate phosphorus
species.
Moreover, while an increase from 5 to 20 molar% of catalyst
did somewhat speed up the reaction, use of over 50 molar%
of catalyst diminished the reaction rate, and use of excess
catalyst completely stopped the reaction. This observation,
together with the substrate-concentration sensitivity, is a first
indication of a bimolecular process involving both a catalyst-
activated species and a phosphinite that is not coordinated to
a Lewis acid (see below).
Two main flaws have precluded wider use of this reaction.
First, the drastic reactions conditions: apart from a with few
particularly reactive halides, high reaction temperatures are
required. Second, this reaction yields, together with the
tetracoordinate P(¼O) phosphorus target product, another
3
alkyl halide R X. However, when this newly formed alkyl
3
halide R X is more reactive, or less volatile than the initial
alkyl halide R X, a troublesome competition takes place, and
mixtures of phosphorylated products are obtained. When
desired, the higher reactivity of the second
halide R X can be used to perform an
autocatalytic” Michaelis–Arbuzov rear-
rangement, by using catalytic amounts of
the first alkyl halide R X (usually methyl
iodide). This latter “catalytic” reaction is
still limited to a few reactive phosphites,
and consumption of the catalyst yields,
among other by-products, a corresponding
amount of an undesired R R R P(O)
phosphorus (v) product.
While the Arbuzov rearrangement has
long been known, apart from the ther-
mally induced alkyl-halide catalysis[
already mentioned and the related
iodine or alkali-metal–iodide catalysts,
the only true thermal Arbuzov-rearrange-
ment catalyst described to date is nickel(ii)
Optimized reaction conditions were then sought for a
wide array of substrates. As illustrated in Table 1, addition of
5 molar% TMSOTf to phosphinite bearing primary O-alkyl
4
[
2]
3
Table 1: Lewis acid promoted Arbuzov rearrangement of phosphinites.[a]
“
R3
O
P
O
P
4
R1
R1
R2
3 R2
R
[
3]
Entry
R1
R2
OR3
Reaction
conditions[
Reaction
time [h]
Yield[c]
[%]
b]
1
2
3
4
5
6
7
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
OMe
OMe
OMe
OMe
OEt
A
B
C
D
C
A
C
72
72
1
87
91
98
95
85
87
90
1
2
4
1
18
72
3
OBn
OBn
3]
[
4]
[5]
8
Ph
Ph
C
18
86
9
Ph
Ph
C
2
91
[
1d,6]
chloride.
A few intramolecular rear-
1
1
1
0
1
2
Et
Ph
Ph
Et
Me
Me
OBn
OMe
OEt
A
C
C
72
3
18
92
75
63
rangements, limited to highly reactive
phosphinites or requiring very high tem-
perature or pressure, have been descri-
[
7]
13
Ph
Me
C
3
78[d]
bed, and no efficient room-temperature
catalysts have been described so far. Our
idea was to test the use of oxophilic Lewis
acids which could bind to the oxygen
atom, and thus weaken the OꢀC bond,
1
4
Ph
Me
A
72
68
15
Ph
Me
C
18
44
enhancing the reaction rate of the second,
energy-requiring step of the Arbuzov
rearrangement.
1
1
6
7
Ph
Ph
tBu
tBu
OBn
OMe
C
C
72
72
52
47
A first screening was conducted on
methyl diphenylphosphinite (1; see Sup-
porting Information), described as one of
the most isomerization-sensitive phos-
phinites. When the reaction was per-
formed at room temperature, in CDCl3
18
Ph
tBu
C
72
20[d]
1
2
9
0
tBu
cy
tBu
cy
OMe
OMe
E
E
72
72
degradation
traces
[e]
[a] Unless stated, the reaction was performed at a phosphinite concentration of 1.0m in
[
9]
chloroform with 5 mol% catalyst. [b] Reaction conditions: A: TMSOTf, 258C, B: BF .OEt ,
208C, C: TMSOTf, 608C, D: BF OEt , 608C, E: no solvent, TMSOTf, 808C. [c] Yield of purified
phosphane oxide. [d] 2 isomers 1:1. [e] Cyclohexyl.
3
2
3
2
using 3m phosphinite 1 and 5% Lewis
2
390
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 2389 – 2392

Angewandte
Chemie
R3
groups initiated an Arbuzov rearrangement to phosphane
oxide with fairly high yields in a few days at room temper-
ature when the phosphorus(iii) ester was reactive (methyl or
benzyl, entries 1, 6, 10, and 14). Use of BF .OEt also gave
R3
R1
R1 R2
P
Si OTf
Si
O
P
+O
–
+
O
R2 R1
TfO
2
P
R
3
2
R3
satisfactory results (entries 2, 4).
R3
Interestingly, O-ethyl phosphinites, which are far less
reactive in Arbuzov rearrangements, also isomerize when
moderately (608C) heated (entries 5 and 12), as do activated
phosphinites bearing secondary O-alkyl groups (sec-phen-
ethyl esters, entries 8, 13, and 18).
R1
R2
P
O
P
Si
– +
O
P
TfO
O
2
R2 R1
R2 R1
R3
R3
Si OTf
Reaction is as yet limited to the migration of primary alkyl
groups or activated secondary alkyl groups, as is the Arbuzov
rearrangement involving alkyl halides. Sterically hindered
alkyl substituents at the phosphorus center also dramatically
hamper the rearrangement process (entries 18–20, and to a
lesser extent, entries 16 and 17).
The reaction was then extended to less reactive phos-
phonites and phosphites. As illustrated in Table 2, Arbuzov
rearrangement to the corresponding tetracoordinate phos-
phorus esters also readily occurred. As observed for the
bimolecular Michaelis–Arbuzov rearrangement using alkyl
halides, and as described for phosphinites (see Table 1),
deactivation of the a carbon atom of the rearranging ester
Scheme 2. Proposed mechanism for Lewis acid catalyzed Arbuzov rear-
rangement.
nylphosphite (1), corresponding to the oxonium inter-
mediate).
*
*
No alkyl trifluoromethane sulfonate has been observed in
the crude reaction mixtures.
No room-temperature TMSOTf-catalyzed rearrangement
was observed with S-methyl diphenylphosphinithioite.
This observation is in favor of an oxonium intermediate,
since sulfur atoms are less susceptible towards silylation.
An equimolar mixture of phosphinite 3 and 4 gave an
approximately 1:1:1:1 mixture of all possible phosphane
oxides 5–8 (Scheme 3), as indi-
*
cated by both 31P NMR spectro-
scopy and GC/MS experiments.
Obtaining an approximately
1:1:1:1 mixture, although the
kinetic rates for the migration
of individual OMe and OEt
groups are different (compare
Table 1, entries 3 and 5), not
only supports the bimolecular
aspect of the mechanism, but
also indicates that, if this mech-
anism is correct, the rate-limit-
ing step is the second one (sily-
loxyphosphinite addition to the
phosphonium intermediate).
Table 2: TMSOTf-promoted Arbuzov rearrangement of phosphonites and phosphites.
Entry
R1
R2
OR3
Reaction
Yield[b]
[%]
conditions[
a]
1
2
Ph
Ph
OMe
OEt
OMe
OEt
208C, 96 h
608C, 72 h
82
68
3
Ph
608C, 18 h
72
4
5
6
7
8
Ph
OiPr
OEt
OEt
OEt
OMe
OiPr
OMe
OMe
OBn
OMe
608C, 72 h
208C, 96 h
608C, 18 h
608C, 18 h
608C, 18 h
traces
80
96
87
98
OEt
OEt
OEt
OMe
9
OEt
OEt
608C, 3 h
93
*
Addition of a catalytic (20%)
amount of methyl trifluorome-
[
a] The reaction was performed at a concentration of 1.0m in chloroform with 5 mol% catalyst. [b] Yield
of phosphi(o)nates.
drastically lowers the rates of the reaction. A particularly high
selectivity with respect to the migrating ester group was
observed (benzyl > methyl @ ethyl, in the reactions given in
entries 6 and 7, only traces of ethyl-migration products were
detected).
As previously observed, with inhibition of the reaction by
excess catalyst, and contrary to the intramolecular mechanism
we originally proposed, much of the experimental evidence is
strongly in favor of the bimolecular process outlined in
Scheme 2.
*
Whatever the oxophilic Lewis acid used (TMSOTf,
+
TBSOTf (TBS = tert-butyldimethylsilyl, Cl C¼N(Me)
2
2
ꢀ
TfO , Sn(OTf) , and BF OEt ), the reaction occurred in
2
3
2
the same way, and the detected reaction intermediates
give similar 31P NMR spectroscopy data (major peak at
d = 75.2 to 75.4 ppm for the reaction with methyl diphe-
Scheme 3. Lewis acid catalyzed Arbuzov rearrangement of an equimo-
lar amount of phosphinites 3 and 4.
Angew. Chem. Int. Ed. 2003, 42, 2389 – 2392
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2391

Communications
thane sulfonate to 1 gave no rearrangement product 2 at
room temperature. Only formation of the corresponding
[4] a) V.K. Yadav, Synth. Commun. 1990, 20, 239 – 246; b) H.
Brunner, W. Zettlmeier, Bull. Soc. Chim. Belg. 1991, 100, 247 –
2
57.
5] a) V. P. Kukhar', E. I. Sagina, Zh. Obshch. Khim. 1976, 46, 2686 –
689; b) U. Rose, J. Heterocycl. Chem. 1991, 28, 2005 – 2012; c) K.
amounts
of
the
quasi-phosphonium
salt
[
+
ꢀ
31
[
MeO(Ph) PMe] [OTf] was detected ( P NMR: d =
2
2
7
7.8 ppm in CDCl ). Only heating the reaction mixture
3
Toshima, H. Yamagushi, T. Jyojima, Y. Nogushi, M. Nakata, S.
to 1008C, without solvent, gave the formation of 2, about
5%, together with a mixture of starting phosphinite, and
the oxidation product O-methyl diphenylphosphinate.
Matsumura, Tetrahedron Lett. 1996, 37, 1073 – 1076.
[6] W. Flisch, P. Rußkamp, W. Langer, Liebigs Ann. Chem. 1985,
1
1413 – 1421.
[
7] a) T. Janecki, R. Bodalski, Synthesis 1990, 799 – 801; b) R. S.
Macomber, D. E. Rardon, D. M. Ho, J. Org. Chem. 1992, 57,
*
*
Equimolar mixtures of phosphane oxide diastereoisomers
(
(
Table 1, entries 13 and 18) and racemic mixtures
Table 1, entry 8) were obtained, even when enantiomeri-
3874 – 3881; c) S. Demay, K. Harms, P. Knochel, Tetrahedron Lett.
1999, 40, 4981– 4984; d) W. G. Bentrude, S. G. Lee, K. Akuta-
cally pure sec-phenethyl alcohols were used.
gawa, W. Ye, Y. Charbonnel, J. Am. Chem. Soc. 1987, 109, 1577;
e) D. Shukla, C. Lu, N. P. Schepp, W. G. Bentrude, L.J. Johnston,
J. Org. Chem. 2000, 65, 6167 – 6172; f) T. Wada, Y. Sato, F. Honda,
S.-I. Kawahara, M. Sekine, J. Am. Chem. Soc. 1997 119, 1 271 0 –
12721; g) M. R. Banks, R. F. Hudson, J. Chem. Soc. Perkin Trans.
For the Arbuzov rearrangement, a bimolecular process
involving nucleophilic addition of phosphite to phospho-
[
8]
nium has already been described, and confirmed by
kinetic experiments.[
3b]
This is actually the only possible
2
1989, 463 – 467; h) A. E. Arbuzov, K.V. Nikonorov, Zh. Obshch.
Khim. 1948, 18, 2008; h) V. Mark, Mech. Mol. Migr. 1969, 2, 31 9 –
37. Noteworthy is the spontaneous rearrangement of particularly
mechanism when methyl triflate is used as a halide
equivalent, since the nucleophilicity of the corresponding
4
ꢀ
trifluoromethane sulfonate anion TfO is particularly
activated triarylmethyl phosphorus-3-esters, for instance D. V. D.
Hardy, H. H. Hatt, J. Chem. Soc. 1952, 3778 – 3783, and references
therein.
[
9]
low.
In conclusion, we have demonstrated that use of an oxophilic
Lewis acid, such as BF .OEt or TMSOTf, could induce an
[
8] a) P. Rumpf, Bull. Soc. Chim. Fr. 1951, 18, 128; b) T. B. Brill, S. J.
Landon, Chem. Rev. 1984, 84, 577 – 585.
9] a) K. S. Colle, E. S. Lewis, J. Org. Chem. 1978, 43, 571– 574;
b) E. S. Lewis, K. S. Colle, J. Org. Chem. 1981, 46, 4369 – 4372.
3
2
Arbuzov-like rearrangement. The reaction conditions are
particularly mild, since in most cases, room temperature or
moderate heating are sufficient, which has not been described
before for Arbuzov rearrangements. The scope of the reaction
has been extended to phosphites, phosphinites, and phos-
phonites. As in the corresponding Arbuzov rearrangement
using alkyl halides, the migrating groups are limited to both
primary and activated secondary alkyl groups. The reaction
mechanism has also been clarified, and a bimolecular process
is proposed.
[
This room-temperature, catalyzed Arbuzov rearrange-
ment, offers new reaction pathways for the synthesis of
compounds containing carbon–phosphorus bonds, for exam-
ple, key intermediates for the synthesis of both physiologi-
cally active drugs and basic chemical building blocks (via
further Wittig or Horner–Wadsworth–Emmons reactions, for
instance).
Received: October 1, 2002
Revised: December 18, 2002 [Z50270]
Keywords: Lewis acids · michaelis–arbuzov rearrangement ·
.
phosphorus · rearrangement
[
1] a) A. Michaelis, R. Kaehne, Ber. Dtsch. Chem. Ges. 1898, 31,
408; b) A. E. Arbuzov, J. Russ. Phys. Chem. Soc. 1906, 38, 687; c)
1
The Chemistry of Organophosphorus Compounds (Ed.: F. H.
Hartley), Wiley, New York, 1996; d) A. K. Bhattacharya, G.
Thyagarajan, Chem. Rev. 1981, 81, 415 – 430.
[
2] M. Saady, L. Lebeau, C. Mioskowski, Tetrahedron Lett. 1995, 36,
5183 – 5186; M. Saady, L. Lebeau, C. Mioskowski, Helv. Chim.
Acta 1995, 78, 670 – 678.
[
3] a) J. E. Griffiths, A. B. Burg, J. Am. Chem. Soc. 1962, 84, 3442 –
3450; b) E. S. Lewis, D. Hamp, J. Org. Chem. 1983, 48, 2025 –
2029; c) H. J. Woerz, E. Quien, H. P. Latscha, Z. Naturforsch. B
1984, 39, 1706 – 1710.
2
392
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 2389 – 2392