Stereogenic P-trisubstituted phosphorus by crystallization-induced asymmetric transformation: A practical synthesis of phenyl(o-anisyl)methylphosphine borane

Full text of DOI:10.1021/ja970369x

J. Am. Chem. Soc. 1997, 119, 9293-9294
9293
crystalline derivatives were obtained starting from the Whitesell
Stereogenic P-Trisubstituted Phosphorus by
Crystallization-Induced Asymmetric
auxiliary 1a,10 or from (R)-pantolactone (1b). For 1a, conver-
Transformation: A Practical Synthesis of
Phenyl(o-anisyl)methylphosphine Borane
Edwin Vedejs* and Yariv Donde
Chemistry Department, UniVersity of Wisconsin
Madison, Wisconsin 53706
ReceiVed February 4, 1997
ReVised Manuscript ReceiVed July 17, 1997
Chiral phosphines have been prepared by a variety of
methods, most of which involve the displacement of a chiral
leaving group from stereogenic phosphorus by an alkylmetal
reagent.1,2 Control of phosphorus configuration has also been
achieved using a non-covalently bound chiral auxiliary.3 We
report an alternative approach based on the first examples of
crystallization-induced asymmetric transformation (AT) at ste-
reogenic phosphorus having three carbon substituents.4,5 The
potential of this method is illustrated in a practical synthesis of
phenyl(o-anisyl)methylphosphine (PAMP).6
In principle, the AT phenomenon can be used to convert an
equilibrating mixture of chiral phosphine diastereomers R*PPhAr
(R*) chiral alkyl) into a single isomer if pyramidal inversion
is faster than crystallization.4 However, the melting point of
the product phosphine would have to be higher than the
inversion temperature (>100 °C for typical tertiary phosphines;
activation energy >30 kcal/mol)7 to achieve AT. More
convenient inversion barriers of 20-25 kcal/mol are expected
for alkoxycarbonylphosphines.8 Thus, chiral derivatives R*O2-
CPPhAr should undergo AT near room temperature if the
auxiliary R* favors a single phosphorus configuration in the
crystal lattice. The alkoxycarbonyl group also provides a built-
in means to modify one of the phosphorus substituents and to
remove and recycle the chiral auxiliary, as described below.
This is accomplished via P-alkylation followed by hydrolytic
cleavage of an intermediate alkoxycarbonylphosphonium salt,9
resulting in the formation of a tertiary phosphine with excellent
enantiomeric purity.
sion to the chloroformate 2a with COCl2/toluene + 2,6-lutidine
followed by lithium o-anisylphenylphosphide11,12 produced the
alkoxycarbonylphosphines 3a and 4a (ca. 1:1 diastereomer
mixture). A similar procedure from pantolactone 1b afforded
the chloroformate 2b, but the phosphide coupling step gave
undesired byproducts. On the other hand, o-anisylphenylphos-
phine reacted cleanly with 2b in the absence of any base to
produce 3b and 4b (1:1 ratio).13
Crystallization of the mixtures of 3a,4a or of 3b,4b afforded
crystals consisting of a single dominant diastereomer in each
case according to 31P and 1H NMR assay at -20 °C, below the
threshold for pyramidal inversion. The structure of 3a was
proven by X-ray crystallography, while the assignment of 3b
was based on its eventual transformation to (RP)-PAMP as
described later. The pantolactone series was selected for
detailed optimization in subsequent steps (see below) because
1b is inexpensive compared to 1a. However, analogous
transformations in the Whitesell auxiliary-derived series via 3a
are facile and provide an excellent route to (RP)-PAMP (see
the Supporting Information for details). This series has the
advantage that both enantiomers of the starting 1a are equally
available.10,14
Crystallization of the 1:1 mixture of 3b and 4b from ethanol
at room temperature resulted in a dramatic change in the
diastereomer ratio. On a multigram scale, the isolated product
contained 3b and 4b in a ratio of 25-32:1 according to 31P
NMR assay (85% overall from pantolactone; direct crystalliza-
tion from ethanol, two crops). Since the recovery of 3b
exceeded the amount originally present in the oil, equilibration
of the phosphorus diastereomers must have occurred during the
crystallization. This confirms that an AT process was involved
in the product isolation step as expected. However, AT also
occurred in a more surprising way. Thus, a sample containing
3b and 4b in a 25:1 diastereomer ratio was found to improve
over time! This required nothing more than storing the solid
material at room temperature (3b:4b ) 48:1 after 2 weeks; 91:1
Several chiral alcohols R*OH were surveyed for crystallinity
in the corresponding P-alkoxycarbonylphosphine 3 or 4 and
(1) For reviews, see: Pietrusiewicz, K. M.; Zablocka, M. Chem. ReV.
1994, 94, 1375. Kagan, H. B.; Sasaki, M. In The Chemistry of Organo-
phosphorus Compounds; Hartley, F. R., Ed.; Wiley: New York, 1990; Vol.
1, Chapter 3.
(2) (a) Juge, S.; Stephan, M.; Laffitte, J. A.; Genet, J. P. Tetrahedron
Lett. 1990, 31, 6357. (b) Imamoto, T.; Oshiki, T.; Onozawa, T.; Kusumoto,
T.; Sato, K. J. Am. Chem. Soc. 1990, 112, 5244. (c) Imamoto, T.; Matsuo,
M.; Nonomura, T.; Kishikawa, K.; Yanagawa, M. Heteroat. Chem. 1993,
4, 475. (d) Corey, E. J.; Chen, Z.; Tanoury, G. J. J. Am. Chem. Soc. 1993,
115, 11000. (e) Sheehan, S. K.; Jiang, M.; McKinstry, L.; Livinghouse, T.;
Garton, D. Tetrahedron 1994, 50, 6155. (f) Kolodiazhnyi, O. I.; Grishkun,
E. V. Tetrahedron: Asymmetry 1996, 7, 967.
(3) Muci, A. R.; Campos, K. R.; Evans, D. A. J. Am. Chem. Soc. 1995,
117, 9075.
(4) AT is used here as an abbreviation for “second order” asymmetric
transformation (more accurately, asymmetric transformation of the “second
kind”): (a) Eliel, E. L.; Wilen, S. H.; Mander, L. N. in Stereochemistry of
Organic Compounds; Wiley: New York, 1994; p 364. (b) Jaques, J.; Collet,
A.; Wilen, S. H. In Enantiomers, Racemates, and Resolutions; Krieger
Publishing Co.: Malabar, FL, 1994.
(5) Previous examples of AT at stereogenic phosphorus involve secondary
phosphine or halophosphine derivatives: Bader, A.; Pabel, M.; Wild, S. B.
J. Chem. Soc., Chem. Commun. 1994, 1405. Pabel, M.; Willis, A. C.; Wild,
S. B. Inorg. Chem. 1996, 35, 1244. Bader, A.; Pabel, M.; Willis, A. C.;
Wild, S. B. Inorg. Chem. 1996, 35, 3874.
(6) PAMP is the precursor to the renowned diPAMP ligand: Knowles,
W. S. J. Chem. Educ. 1986, 63, 222. Knowles, W. S. Acc. Chem. Res.
1983, 16, 106.
(7) Baechler, R. D.; Mislow, K. J. Am. Chem. Soc. 1970, 92, 3090.
Mislow, K. Trans. NY Acad. Sci., Ser. II 1973, 35, 227.
(8) (a) Chervin, I. I.; Isobaev, M. D.; El’natanov, Y. I.; Shikhaliev, S.
M.; Bystrov, L. V.; Kostyanovskii, R. G. Bull. Acad. Sci. USSR 1981, 1438.
(b) Egan, W.; Mislow, K. J. Am. Chem. Soc. 1971, 93, 1805.
(9) Issleib, K.; Priebe, E. Chem. Ber. 1959, 92, 3183.
(10) Whitesell, J. K.; Chen, H. H.; Lawrence, R. M. J. Org. Chem. 1985,
50, 4664.
(11) Prepared from PhPCl2 by sequential treatment with o-anisylmag-
nesium bromide and LiAlH4 by analogy to ref 2b (see the Supporting
Information).
(12) (a) van Doorn, J. A.; Frijns, J. H. G.; Meijboom, N. Recl. TraV.
Chim. Pays-Bas 1991, 110, 441. (b) Mann, F. G.; Tong, B. P.; Wystrach,
V. P. J. Chem. Soc. 1963, 1155.
(13) A similar sequence from bornyl chloroformate produced a ca. 1:1
diastereomer mixture of alkoxycarbonylphosphines analogous to 3 and 4,
but crystallization gave a quasiracemate (both diastereomers in the unit cell,
1.0:1.0 ratio).
(14) King, S. B.; Sharpless, K. B. Tetrahedron Lett. 1994, 35, 5611.
S0002-7863(97)00369-7 CCC: $14.00 © 1997 American Chemical Society

9294 J. Am. Chem. Soc., Vol. 119, No. 39, 1997
Communications to the Editor
after 6 weeks), but the process could be performed more
conveniently by warming the solid sample. When a 32:1
mixture was kept at 50 °C for 22 h, the sample was found to
consist of a 99:1 ratio of 3b:4b. No further improvement was
detected after an additional 22 h. The diastereomer upgrade
occurs without any cost in terms of yield (100% recovery) and
must take place via the AT phenomenon. We do not know
whether the upgrade involves conversion from one crystalline
phase to another. A more likely scenario is that amorphous 4b
precipitates as a contaminant (3-5%) during the initial crystal-
lization of finely divided 3b and that the amorphous material
is slowly converted to 4b by AT.
In solution, purified 3b is configurationally stable at -20 °C,
but it readily re-equilibrates with 4b at room temperature (t1/2
ca. 30 min). The crude 3b/4b mixture is air- and water-stable
on the time scale of routine aqueous workup and it also survives
brief exposure to silica gel. However, the material slowly
decomposes by air oxidation and the key crystallization
procedure was therefore carried out under nitrogen using
deoxygenated solvents as a precaution to avoid potential
complications that might interfere with efficient crystallization.
Cleavage of the P-acyl group in 3b and conversion to PAMP
was performed by P-alkylation (CH3OTf, CH2Cl2 -78 to
-20°C) to the acyl phosphonium salt 5b, followed by hydroly-
sis. The alkylation step is crucial because pyramidal inversion
only technique found to give acceptable accuracy, and it was
used to deduce the diastereomer ratios of 3b:4b mentioned
earlier. The results were confirmed by ee assay after hydrolysis
and conversion to PAMP borane complex as follows.
The crude acylphosphonium salt was treated with H2O/THF
at room temperature (3.5 h) to give recovered (R)-pantolactone
(96% yield; >99 ee by HPLC assay) and (RP)-PAMP, isolated
as the borane complex in >98% yield. Starting from 3b that
had been “aged” for 2 weeks (48:1 ratio of 5b:6b) the (RP)-
PAMP borane 7 was obtained with 94-95% ee (HPLC assay
on chiral stationary phase). The enantiomeric purity of 7 was
easily upgraded from 95% ee to >99.5% ee after conventional
recrystallization from hexane (84% recovery based on 3b). In
another experiment, 5b was hydrolyzed in the presence of
pyridine. This produced a faster reaction (minutes at rt), and
the PAMP was obtained with 97% ee starting from 3b having
98% de.16 This evidence confirms that the P-alkylation and
hydrolysis sequence occurs with <1% equilibration of phos-
phorus configuration. Prior studies by Imamoto et al. have
established efficient procedures for the conversion of 7 to
diPAMP.2b
The overall conversion from pantolactone to recrystallized
(>99.5% ee) (R)-PAMP borane 7 was achieved in 74% yield.
A similar sequence was performed using the Whitesell auxiliary
via 3a and 5a. This series was not optimized in detail, but the
results were comparable: 80% yield of (R)-PAMP borane 7
(98% ee, not recrystallized) from 1b. These findings establish
the AT-based synthesis as a practical route to PAMP on a
multigram scale. The corresponding borane complex 7 can also
serve as the starting point for synthesis of other phosphines
having stereogenic phosphorus, according to preliminary re-
sults.17 Further examples and applications are under investiga-
tion.
Acknowledgment. This work was supported in part by the National
Science Foundation (CHE 9521355), by the Kodak Fellows Program
(scholarship to Y.D.), and by a grant from NSF (CHE-9310428) for
an X-ray diffractometer. The authors are also grateful to Prof. J.
Whitesell for a generous gift of the auxiliary 1b and to Dr. D. R. Powell
for the X-ray structure determinations.
at phosphorus could re-equilibrate the diastereomers, resulting
in lower diastereomeric excess (de) at the stage of 5b and lower
enantiomeric excess (ee) after hydrolysis. A highly sensitive
assay for the diastereomer ratio was therefore desired, and the
necessary precision was achieved using a phosphorus variant
of the 13C satellite technique that is reported for the accurate
determination of large product ratios based on 1H NMR
spectra.15 Thus, the natural abundance 13C satellites of the
proton-decoupled 31P NMR signal of 5b (dominant diastere-
omer) were compared with the 31P signal of the minor
diastereomer 6b. This method gave reproducible ratios of 5b:
Supporting Information Available: Complete experimental pro-
cedures, characterization, and X-ray structure data for 3a (37 pages).
See any current masthead page for ordering and Internet access
instructions.
JA970369X
(15) Dewey, R. S.; Schoenewaldt, E. F.; Joshua, H.; Paleveda, W. J.,
Jr.; Schwam, H.; Barkemeyer, H.; Arison, B. H.; Veber, D. F.; Denkewalter,
R. G.; Hirschman, R. J. Am. Chem. Soc. 1968, 90, 3254. Freidinger, R.
M.; Hinkle, J. S.; Perlow, D. S.; Arison, B. H. J. Org. Chem. 1983, 48, 77.
(16) A small amount of a symmetrical carbonate byproduct (<5%) was
also formed.
(17) Decomplexation of 7 with Et2NH (4 h, 55 °C), P-alkylation with
Br(CH2)3OTBS and reductive cleavage of the resulting phosphonium salt
gave a 10:1 ratio of Ph(Me)P(CH2)3OTBS and o-MeOC6H4(Me)P(CH2)3-
OTBS (68%). The major product was purified as the borane complex and
was deprotected with Bu4NF to give Ph(Me)P(BH3)(CH2)3OH, 77% yield,
99% ee by HPLC assay. This sequence demonstrates that 7 is useful as a
precursor of other chiral phosphines.
6b ranging from 50:1 to 100:1, depending on the history (the
extent of “aging”) of the starting sample of 3b. This was the