Characteristic reactions and properties of C-apical O-equatorial (O-cis) spirophosphoranes: Effect of the σ*P-O orbital in the equatorial plane and isolation of a hexacoordinate oxaphosphetane as an intermediate of the wittig type reaction of 10-P-5 phosphoranes

Article abstract of DOI:10.1021/ja026776c

Novel spirophosphoranes (O-cis) that exhibit reversed apicophilicity having an apical carbon-equatorial oxygen array in a five-membered ring showed enhanced reactivity toward nucleophiles such as n-Bu₄N⁺F^- or MeLi in comparison with the corresponding stable isomeric spirophosphoranes (O-trans) having an apical oxygen - equatorial carbon configuration. The enhanced reactivity of the O-cis isomer could be explained by the presence of a lower-lying σ*_{P-O(equatorial)} orbital as the reacting orbital in the equatorial plane, whereas the corresponding orbital is a higher-lying σ*_{P-O(equatorial)} in the O-trans isomer. Density functional theory (DFT) calculation on the actual compounds provided theoretical support for this assumption. In addition, we found that the benzylic anion α to the phosphorus atom in O-0-cis benzyl benzyi phosphorane is much more stable than that generated from the corresponding O-trans compounds. The experimental results were considered to be due to the n_c → σ*_P-O interaction in the O-cis anion, and this was confirmed by DFT calculations. Furthermore, the hexacoordinate anionic species derived from the reaction of the benzylic anion from O-cis benzylphosphorane with an aldehyde was also found to be stabilized as compared with analogous species from the corresponding O-trans isomer. The first X-ray structural characterization of a hexacoordinate phosphate intermediate in the Wittig type reaction using pentacoordinate phosphoranes is reported.

Full text of DOI:10.1021/ja026776c

Published on Web 10/15/2002
Characteristic Reactions and Properties of C-Apical
O-Equatorial (O-cis) Spirophosphoranes: Effect of the σ*P-O
Orbital in the Equatorial Plane and Isolation of a
Hexacoordinate Oxaphosphetane as an Intermediate of the
Wittig Type Reaction of 10-P-5 Phosphoranes
Shiro Matsukawa,† Satoshi Kojima,† Kazumasa Kajiyama,† Yohsuke Yamamoto,†
Kin-ya Akiba,*,†,‡ Suyong Re,§ and Shigeru Nagase§
Contribution from the Department of Chemistry, Graduate School of Science,
Hiroshima UniVersity, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8526, Japan, the AdVanced
Research Center for Science and Engineering, Waseda UniVersity, 3-4-1 Ohkubo,
Tokyo 169-8555, Japan, and the Department of Theoretical Studies, Institute for Molecular
Science, Myodaiji, Okazaki 444-8585, Japan
Received May 3, 2002. Revised Manuscript Received August 16, 2002
Abstract: Novel spirophosphoranes (O-cis) that exhibit reversed apicophilicity having an apical carbon-
equatorial oxygen array in a five-membered ring showed enhanced reactivity toward nucleophiles such as
n-Bu4N+F- or MeLi in comparison with the corresponding stable isomeric spirophosphoranes (O-trans)
having an apical oxygen-equatorial carbon configuration. The enhanced reactivity of the O-cis isomer
could be explained by the presence of a lower-lying σ*P-O(equatorial) orbital as the reacting orbital in the
equatorial plane, whereas the corresponding orbital is a higher-lying σ*P-C(equatorial) in the O-trans isomer.
Density functional theory (DFT) calculation on the actual compounds provided theoretical support for this
assumption. In addition, we found that the benzylic anion R to the phosphorus atom in O-cis benzyl
phosphorane is much more stable than that generated from the corresponding O-trans compounds. The
experimental results were considered to be due to the nC f σ*P-O interaction in the O-cis anion, and this
was confirmed by DFT calculations. Furthermore, the hexacoordinate anionic species derived from the
reaction of the benzylic anion from O-cis benzylphosphorane with an aldehyde was also found to be stabilized
as compared with analogous species from the corresponding O-trans isomer. The first X-ray structural
characterization of a hexacoordinate phosphate intermediate in the Wittig type reaction using pentacoordinate
phosphoranes is reported.
using various model phosphoranes.3 Through these studies, it
was found that two characteristic aspects, apicophilicity4 (a
Introduction
Phosphoryl transfer, an important biological reaction involved
in processes such as energy transfer and DNA formation, is
generally assumed to involve pentavalent phosphorus intermedi-
ates (or transition states) formed by nucleophilic attack upon
tetracoordinate phosphorus atoms, and it is assumed that the
stability and stereochemistry (both steric and electronic effects
combined) of the transient species (or transition state) greatly
influence the outcome of the process.1 Therefore, to deduce a
basic understanding of the process,2 much attention has been
focused on the stereochemistry of the pentacoordinate state by
thermodynamic aspect) and pseudorotation5 (a kinetic aspect),
are important. As for apicophilicity, that is, the relative
preference for substituents to occupy the apical positions as
opposed to the equatorial positions in trigonal bipyramidal (TBP)
structures, a number of experimental results6 and theoretical
calculations7 have indicated that multiple factors are involved.
(3) (a) Uchimaru, T.; Stec, W. J.; Taira, K. J. Org. Chem. 1997, 62, 5793-
5800. (b) Uchimaru, T.; Uebayashi, M.; Hirose, T.; Tsuzuki, S.; Yliniemela,
A.; Tanabe, K.; Taira, K. J. Org. Chem. 1996, 61, 1599-1608. (c) Taira,
K.; Uchimaru, T.; Storer, J. W.; Yliniemela, A.; Uebayashi, M.; Tanabe,
K. J. Org. Chem. 1993, 58, 3009-3017. (d) Dejaegeve, A.; Lim, C.;
Karplus, M. J. Am. Chem. Soc. 1991, 113, 4353-4355. (e) Pyle, A. M.
Science 1993, 261, 709-714.
* Corresponding author. E-mail: akibaky@waseda.jp.
† Hiroshima University.
(4) (a) Akiba, K.-y. Chemistry of HyperValent Compounds; Wiley-VCH: New
York, 1999. (b) Holmes, R. R. Pentacoordinated PhosphorussStructure
and Spectroscopy; ACS Monograph 175, 176, Vol. I, II; American Chemical
Society, Washington, DC, 1980. (c) Corbridge, D. E. C. Phosphorus: An
Outline of Its Chemistry, Biochemistry, and Technology, 4th ed.; Elsevier:
Amsterdam, 1990; Chapter 14, pp 1233-1256. (d) Burgada, R.; Setton,
R. In The Chemistry of Organophosphorus Compounds; Hartley, F. R.,
Ed.; Wiley-Interscience: Chichester, Great Britain, 1994; Vol. 3, pp 185-
272. (e) Martin, J. C. Science 1983, 221, 509-514. (f) Muetterties, E. L.;
Mahler, W.; Schmutzler, R. Inorg. Chem. 1963, 2, 613-618.
‡ Waseda University.
§ Institute for Molecular Science.
(1) (a) Hengge, A. C. Acc. Chem. Res. 2002, 35, 105-112 and references
therein. (b) Sinott, M. ComprehensiVe Biological Catalysis: A Mechanistic
Reference; Academic Press: San Diego, CA, 1998; Vol. 1, pp 517-542.
(c) Engel, R. Handbook of Organophosphorus Chemistry; Marcel Dekker:
New York, 1992. (d) Breslow, R.; Labelle, M. J. Am. Chem. Soc. 1986,
108, 2655-2659.
(2) (a) Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70-78. (b) Thatcher, G.
R. J.; Kluger, R. AdV. Phys. Org. Chem. 1989, 25, 99-265.
(5) Berry, R. S. J. Chem. Phys. 1960, 32, 933-938.
9
13154
J. AM. CHEM. SOC. 2002, 124, 13154-13170
10.1021/ja026776c CCC: $22.00 © 2002 American Chemical Society

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Scheme 1
A general propensity deduced from these results is that
electronegative substituents prefer the apical positions, while
π donative or bulky ligands prefer the equatorial positions. As
for pseudorotation, that is, the mutual positional exchange of a
pair of apical ligands and a pair of equatorial ligands, the barrier
is usually relatively low (calculated to be ca. 2-3 kcal/mol for
PH5), causing rapid pseudorotations in contrast to tetracoordinate
phosphorus species that are ordinarily stereochemically rigid
except in cases in which the substitution at phosphorus facilitates
an edge-inversion process.8,9
Scheme 2
One distinct difference between the more familiar chemistry
of tetracoordinate carbon and that of pentacoordinate compounds
such as phosphoranes is the number of possible stereoisomers.
For simplicity, taking into account only asymmetry at the central
atom, for the former carbon atom only a pair of enantiomers
are possible, and thus in principle there is no difference in
reactivity. On the other hand, for the latter phosphorus in trigonal
bipyramidal structure, up to 10 pairs of diastereoisomers are
possible, and thus a larger variety in reactivity is expected
according to each pair. However, because of the presence of
the pseudorotation process that allows stereoisomers to inter-
convert readily, it is not possible to distinguish the difference
of reactivity among pseudorotamers. In a limited number of
cases, there are reports on the observation of coexisting
stereoisomers in which one apical and one equatorial substituent
are exchanged.10 However, in these cases, the barrier to
pseudorotation is low, and thus these examples are not suitable
for the examination of differences in reactivity. By utilizing the
rigidity of the Martin bidentate ligand, we have successfully
slowed pseudorotation enough to isolate enantiomeric pairs of
optically active 10-P-5 phosphoranes11 and to synthesize spiro-
phosphoranes having an apical carbon-equatorial oxygen array
in a five-membered ring (1: O-cis). The latter represents the
first examples of 10-P-5 phosphorane pseudorotamers that
violate the apicophilicity concept and can still be converted to
their more stable pseudorotamers12,13 of apical oxygen-equato-
rial carbon configuration (2: O-trans) (Scheme 1). This now
provides the opportunity to investigate the difference in reactiv-
ity between isomeric phosphoranes differing only in stereo-
chemistry upon the pentacoordinate phosphorus atom. Nucleo-
philic reactions at pentacoordinate compounds are generally
considered to take place within the equatorial plane, with the
nucleophile attacking one of the equatorial bonds from the rear
side.14 Thus, if this were true, 1 would be expected to be more
reactive toward nucleophiles than its more stable counterpart
2, since a lower-lying σ*P-O orbital is available in 1, whereas
the corresponding orbital is a higher-lying σ*P-C orbital in 2
(Scheme 2). We have found this to be experimentally true, and
a theoretical investigation provided evidence for our rationale
based upon orbital considerations. In addition, we have found
that the benzylic anion R to the phosphorus atom in the O-cis
form is much more stable than that derived from the corre-
sponding O-trans benzyl phosphorane. Furthermore, the hexa-
coordinated anionic species derived from the reaction of an
R-benzylic anion from O-cis benzyl phosphorane with an
aldehyde was also found to be more stabilized as compared with
the hexacoordinate species from the corresponding O-trans
isomer. This has allowed us to carry out the first X-ray structural
analysis of a hexacoordinated phosphate bearing an oxaphos-
phetane ring system, an intermediate in the Wittig type reaction
involving pentacoordinated phosphoranes. Full details are given
herein.
(6) (a) Nakamoto, M.; Kojima, S.; Matsukawa, S.; Yamamoto, Y.; Akiba, K.-
y. J. Organomet. Chem. 2002, 643-644, 441-452. (b) Trippett, S.
Phosphorus Sulfur 1976, 1, 89-98. (c) Trippett, S. Pure Appl. Chem. 1970,
40, 595-604. (d) Buono, G.; Llinas, J. R. J. Am. Chem. Soc. 1981, 103,
4532-4540. (e) Eisenhut, M.; Mitchell, H. L.; Traficante, D. D.; Kaufman,
R. J.; Deutsch, J. M.; Whitesides, G. M. J. Am. Chem. Soc. 1974, 96, 5385-
5397. (f) Moreland, C. G.; Doak, G. O.; Littlefield, L. B.; Walker, N. S.;
Gilje, J. W.; Braun, R. W.; Cowley, A. H. J. Am. Chem. Soc. 1976, 98,
2161-2165. (g) Griend, L. V.; Cavell, R. G. Inorg. Chem. 1983, 22, 1817-
1820.
(7) (a) Hoffmann, R.; Howell, J. M.; Muetterties, E. L. J. Am. Chem. Soc.
1972, 94, 3047-3058. (b) McDowell, R. S.; Streitwieser, A., Jr. J. Am.
Chem. Soc. 1985, 107, 5849-5855. (c) Deiters, J. A.; Holmes, R. R.;
Holmes, J. M. J. Am. Chem. Soc. 1988, 110, 7672-7681. (d) Wang, P.;
Zhang, Y.; Glaser, R.; Reed, A. E.; Schleyer, P. v. R.; Streitwieser, A., Jr.
J. Am. Chem. Soc. 1991, 113, 55-64. (e) Wasada, H.; Hirao, K. J. Am.
Chem. Soc. 1992, 114, 16-27. (f) Thatcher, G. R. J.; Campbell, A. S. J.
Org. Chem. 1993, 58, 2272-2281. (g) Wang, P.; Zhang, Y.; Glaser, R.;
Streitwieser, A.; Schleyer, P. v. R. J. Comput. Chem. 1993, 14, 522-529.
(h) Wladkowski, B. D.; Krauss, M.; Stevens, W. J. J. Phys. Chem. 1995,
99, 4490-4500.
Results and Discussion
Preparation of O-cis Spirophosphoranes and Reactions
with Nucleophiles. O-cis phosphoranes 1a and 1b (1a, R )
Me; 1b, R ) n-Bu) were prepared by procedures recently
(12) (a) Kojima, S.; Kajiyama, K.; Nakamoto, M.; Akiba, K.-y. J. Am. Chem.
Soc. 1996, 118, 12866-12867. (b) Kajiyama, K.; Yoshimune, M.;
Nakamoto, M.; Matsukawa, S.; Kojima, S.; Akiba, K.-y. Org. Lett. 2001,
3, 1873-1875.
(8) Moc, J.; Morokuma, K. J. Am. Chem. Soc. 1995, 117, 11790-11797 and
references therein.
(9) (a) Dixon, D. A.; Arduengo, A. J., III; Fukunaga, T. J. Am. Chem. Soc.
1986, 108, 2461-2462. (b) Dixon, D. A.; Arduengo, A. J., III. J. Phys.
Chem. 1987, 91, 3195-3200. (c) Dixon, D. A.; Arduengo, A. J., III. J.
Am. Chem. Soc. 1987, 109, 338-341.
(13) Some compounds violate the concept of apicophilicity. In these cases some
sort of steric constraints disallowed regular configurations. (a) Timosheva,
N. V.; Prakasha, T. K.; Chandrasekaran, A.; Day, R. O. (b) Timosheva, N.
V.; Chandrasekaran, A.; Prakasha, T. K.; Day, R. O.; Holmes, R. R. Inorg.
Chem. 1996, 35, 6552-6560. (c) Vollbrecht, S.; Vollbrecht, A.; Jeske, J.;
Jones, P. G.; Schmutzler, R.; du Mont, W.-W. Chem. Ber./Recl. 1997, 130,
819-822.
(10) (a) Lo¨sking, O.; Willner, H.; Oberhammer, H.; Grobe, J.; Van, D. L. Inorg.
Chem. 1992, 31, 3423-3427 and references therein. (b) Kawashima, T.;
Soda, T.; Okazaki, R. Angew. Chem., Int. Ed. Engl. 1996, 35, 1096-1098.
(11) (a) Kojima, S.; Kajiyama, K.; Akiba, K.-y. Tetrahedron Lett. 1994, 35,
7037-7040. (b) Kojima, S.; Kajiyama, K.; Akiba, K.-y. Bull. Chem. Soc.
Jpn. 1995, 68, 1785-1797.
(14) For example: Kojima, S.; Nakata, M.; Yamazaki, K.; Akiba, K.-y.
Tetrahedron Lett. 1997, 38, 4107-4110.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13155

A R T I C L E S
Matsukawa et al.
Scheme 3
reported by us from the reaction of P-H phosphorane (3)15 with
3 equiv of RLi, followed by treatment with I2 (Scheme 3).12b
O-cis benzyl phosphorane (1c) was prepared similarly from 3
with 3 equiv of benzylpotassium, which was generated by the
reaction of “Superbase” with toluene. Similarly, a p-fluorophen-
ylmethyl substituent could be introduced (1d). The structures
of 1c and 1d were confirmed by X-ray analysis (Figure 1 (1c)
and Supporting Information (1d)).
Reactions of O-cis 1b and O-trans 2b with nucleophiles were
examined. Using TBAF (tetrabutylammonium fluoride; (n-
Bu)4N+F-) as a nucleophile, the reaction of 1b readily afforded
a hexacoordinate phosphate (4) bearing a P-F bond (δP (THF):
1
-105.0 ppm (d, JP-F ) 706 Hz)) in THF at -30 °C, which
was in equilibrium with 1b (ratio of 4/1b ) 1:1) (Scheme 4).
In contrast, 2b did not react at all. The configuration of the
phosphate 4 could not be determined with certainty because of
rapid decomposition by trace amounts of H2O. However, we
have already characterized the corresponding hexacoordinate
fluoroantimonate bearing two Martin ligands. The antimony
analogue has the fluorine atom anti to one of the Martin ligand
oxygens, which was shown by X-ray analysis. Although
pseudorotation is a very low barrier process, especially in the
presence of nucleophilic species for stiboranes, we determined
the observed complex to be the thermodynamic product suc-
cessfully, and the stabilization factor was attributed to the trans
influence of the F-Sb-O bond.16 Incidentally, this stable
hexacoordinate species could be viewed as the immediate
product of the attack of the fluoride anti to the Sb-O (equatorial)
bond of the O-cis isomer, which in the case of antimony could
not be observed because of facile pseudorotation. By analogy,
in phosphate 4, the fluorine atom is also likely to be located
anti to an oxygen atom, and since equilibrium between 4 and
1b can be observed, 4 should also be the thermodynamically
most stable of all the plausible six isomeric hexacoordinate
phosphates.
Figure 1. ORTEP drawings of 1c and 2c showing the thermal ellipsoids
at the 30% probability level.
conditions (5 equiv of MeLi at room temperature for 6 h) did
2b slowly react with MeLi to give 6b after hydrolysis (19%),
along with a large amount of unreacted O-trans phosphorane
2b. Since O-cis 1a undergoes pseudorotation to give O-trans
2a even at room temperature, the reaction of O-cis 1a with MeLi
was carried out at -23 °C for 3 h to diminish complications
from competing pseudorotation. Phosphorane 6a was obtained
in 9.6% yield. Under the same conditions, O-trans 2a was
completely recovered intact. Hexacoordinate intermediates (5a,
δP (THF) -104 ppm; 5b, -83 and -92 ppm)18 were observed
in both reactions of 1a and b. The adducts 6a and b were
Similarly, O-cis 1b reacted with 1 equiv of MeLi at 0 °C in
90 min to give the corresponding adduct 6b (78% yield) after
treatment with H2O, while O-trans 2b17 was recovered unreacted
under similar conditions (Scheme 5). Only under forcing
(15) Ross, M. R.; Martin, J. C. J. Am. Chem. Soc. 1981, 103, 1234-1235.
(16) Kojima, S.; Doi, Y.; Okuda, M.; Akiba, K.-y. Organometallics 1995, 14,
1928-1937.
(17) Kajiyama, K.; Kojima, S.; Akiba, K.-y. Tetrahedron Lett. 1996, 37, 8409-
8412.
9
13156 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Scheme 4
Scheme 6
equimolar n-BuLi as base, O-cis 1c was completely deprotonated
within 5 min, and similarly, O-trans 2c could also be converted
to the corresponding anion. KH was also effective for 1c, and
deprotonation was achieved within 15 min at room temperature.
However, deprotonation from 2c using only KH was quite slow
and was not complete even after 6 h. In this case, complete
deprotonation required the use of 18-crown-6 ether as an
additive. A clear difference was observed between 1c and 2c
in the reaction with KHMDS (potassium hexamethyldisilazide;
(Me3Si)2N-K+). Deprotonation of 1c with KHMDS occurred
in THF at 0 °C in 30 min and gave R-deuterated product (40%
yield) after quenching the mixture with D2O. On the other hand,
2c could not be deprotonated under the same conditions (Scheme
6). These results indicated that the benzyl protons of O-cis 1c
were more acidic than those of O-trans 2c. However, since there
was the possibility that the distinct difference in reactivity was
of a kinetic nature, an equilibration reaction between preformed
O-trans anion 8c and neutral O-cis 1c was carried out. When
O-cis 1c was added to a solution of O-trans anion 8c-K(18-
crown-6), equilibration took place rapidly, and the equilibrium
completely shifted toward O-cis anion 7c-K(18-crown-6) and
neutral O-trans 2c (Scheme 7). Thus, the higher acidity of the
O-cis 1c in comparison with O-trans 2c was confirmed to be
of a thermodynamic nature. The increased acidity of O-cis 1c
in comparison with O-trans 2c would be due to the nC f σ*P-O
interaction in O-cis anion 7c, which is expected to be much
larger than the nC f σ*P-C interaction in O-trans anion 8c-M
(Figure 2). The higher acidity of R-protons of the O-cis series
as compared with those of the O-trans counterparts was
supported by density functional theory (DFT) calculations (vide
infra) on 1a.
Scheme 5
characterized by spectroscopic data, elemental analysis, and
X-ray analysis (see Supporting Information).
Stabilization of Lone Pair Electrons Adjacent to the
Phosphorus Atom of O-cis 1: Effect of σ*P-O Orbital in the
Equatorial Plane on Stabilizing the r-Anion. Because the
yield of 6a could not be improved even after prolonged reaction
(6 h) of 1a with MeLi at 0 °C, we realized that this was due to
the occurrence of facile deprotonation at the methyl group R to
the phosphorus atom instead of nucleophilic attack. Thus, we
decided to compare the behavior of O-cis and O-trans phos-
phoranes toward base by examining products resulting from the
treatment with D2O instead of H2O after the addition of base.
Because of complications arising from the fast pseudorotation
from O-cis 1a, we decided to look into the behavior of O-cis
1c and O-trans 2c19 toward several bases in THF. Using
Before resorting to theoretical calculations, the spectral
properties of the anions were briefly examined by means of
31P NMR to determine whether the structural factors might assist
in stabilizing the O-cis structures. In the case of O-trans
phosphoranes, chemical resonances shifted to higher-field when
2c (δ -21 ppm) was converted to 8c-M (M ) Li, K; δ -34
ppm), regardless of which metal was the countercation. In
contrast, apart from the 31P NMR chemical shift of the parent
1c (δ -7 ppm), the lithium salt of O-cis 7c-Li showed a
(19) Kojima, S.; Kawaguchi, K.; Akiba, K.-y. Tetrahedron Lett. 1997, 38, 7753-
7756. A full account that includes the kinetic data of olefin formation is
currently in preparation.
(18) The reason two signals were observed is unclear at present.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13157

A R T I C L E S
Matsukawa et al.
Scheme 7
significantly different chemical shift (δ 2 ppm) from those of
potassium (7c-K: δ -15 ppm), potassium 18-crown-6 (7c-K-
(18-crown-6): δ -25, -32 ppm (ratio 6:4)), or potassium
[2.2.2]cryptand (7c-K(cryptand): δ -25, -32 ppm (ratio 2:1))
salt. Therefore, 7c-Li may be stabilized by coordination of the
lithium cation with the two cis oxygen atoms in the O-cis
framework. However, in the cases of 7c-K(18-crown-6) and
7c-K(cryptand), such chelation effects would be expected to
be small based on the similar chemical shifts, although an
aggregation effect may be operative even in these cases because
the compounds showed two different 31P NMR signals. It is
also possible that the two signals correspond to rotational
isomers around the P-C(benzyl) bond having a double bond
character (vide infra), since they were observed only for 7c,
not for 8c.
salt to the corresponding phosphorus ylide, which is generally
considered to be of high double bond character.20
Theoretical Study on σ* Orbitals of O-cis and O-trans
Phosphoranes in the Equatorial Plane and on the Stability
of r-Anions from O-cis and O-trans Phosphoranes. The
structures of O-cis (1a) and O-trans (2a) phosphoranes with a
methyl group as a monodentate ligand were optimized with DFT
at the hybrid B3PW91 level,21 using the Gaussian 98 program.22
The basis sets employed were 6-31G(d)23 for C, N, O, and H
and 6-311G(2d)24 for P. The calculated bond distances around
the phosphorus atom are shown in Table 1, and they agree with
the experimental values of 1b-d12a and 2a-c12a (for experi-
mental details of 2a and c, see Supporting Information), although
the calculated P-O and P-C distances of 1a and 2a are slightly
longer than the experimental values. Since the σ*
orbitals
P-X
In 13C NMR, coupling constants (1JP-C) of the anions (220
Hz in 7c-K and 227 and 228 Hz in 7c-K(18-crown-6), 215 Hz
in 8c-K(18-crown-6)) were almost twice as much as those of
the corresponding neutral phosphoranes (107 Hz in 1c and 112
Hz in 2c). These results indicate the contribution of double bond
character between the central phosphorus atom and the anionic
benzyl carbon (Figure 2). This increase in coupling constants
is similar to that observed for the change from phosphonium
in the equatorial plane are responsible for nucleophilicity, we
looked for such orbitals. These were found as a LUMO+4 for
1a and as a LUMO+5 for 2a, as shown clearly in Figure 3.
Other LUMOs lying lower in energy below σ*P-X (X ) O or
C) were distributed on the aromatic rings of the Martin ligands
as π* orbitals. It should be noted that the σ*P-O level of 1a
(LUMO+4) is 18.7 kcal/mol lower than the σ*P-C level of 2a
(LUMO+5). Apparently, this large energy difference between
σ*P-O and σ*P-C orbitals leads to a great enhancement of the
reactivity of O-cis spirophosphorane 1 as compared with O-trans
isomer 2, as shown in Scheme 2.
Calculations were also carried out for O-cis anion (7a) and
the corresponding O-trans anion (8a). The optimized structures
are shown in Figure 4. Table 2 compares the equatorial distances
in O-cis 1a, O-trans 2a, and their corresponding anions 7a and
8a. For both 7a and 8a, the three atoms constituting the CH2
group make a coplanar (sp2) plane, which is almost perpen-
dicular to the equatorial plane. This implies that the lone pair
on the carbon atom in the equatorial plane can be involved in
a nC f σ*P-X (equatorial) interaction (Figure 2). Accordingly,
both anions (7a and 8a) are stabilized by the nC f σ*P-X
(20) Kalinowski, H.-O.; Berger, S.; Braun, S. Carbon 13 NMR Spectroscopy;
Wiley: Chichester, Great Britain, 1988; Chapter 4, pp 586-594.
(21) (a) Becke, A. D. Phys. ReV. 1988, A38, 3098-3100. (b) Becke, A. D. J.
Chem. Phys. 1993, 98, 5648-5652. (c) Perdew, J. P.; Wang, Y. Phys. ReV.
1992, B45, 13244-13249.
(22) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.5; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(23) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S.;
DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982, 77, 3654-3665.
(24) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639-5648.
Figure 2. Effect of σ*P-O orbital in the equatorial plane in 1 on stabilizing
the R-carbanion.
9
13158 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Table 1. Selected Bond Distances (Å) for 1a, 1b,12a 1c, 1d, 2a, 2b,12a and 2c
Calculated
Experimental
1a
2a
1ba
1c
1d
2a
2ba
2c
P(1)-O(1)
P(1)-O(2)
P(1)-C(1)
P(1)-C(2)
P(1)-C(3)
1.769
1.665
1.828
1.876
1.827
1.769
1.764
1.822
1.822
1.818
1.773(3)
1.658(2)
1.806(4)
1.860(4)
1.830(5)
1.782(2)
1.661(2)
1.812(3)
1.874(3)
1.849(3)
1.766(3)
1.667(2)
1.808(3)
1.876(4)
1.849(4)
1.762(2)
1.757(2)
1.825(2)
1.819(2)
1.814(2)
1.765(2)
1.750(2)
1.822(2)
1.817(2)
1.818(3)
1.765(3)
1.766(3)
1.824(3)
1.817(4)
1.845(4)
a These values are revised ones furnished by resolving the structures.
Table 2. Comparison of Calculated Bond Distances (Å) between
Neutral (1a and 2a) and Anionic Phosphoranes (7a and 8a)
O-cis
1a
7a
increment
equatorial
P(1)-C(1)
P(1)-O(2)
P(1)-C(3)
P(1)-O(1)
P(1)-C(2)
1.828
1.665
1.827
1.769
1.876
1.854
1.758
1.667
1.839
1.887
+0.026 (+1.4%)
+0.093 (+5.6%)
-0.160 (-8.8%)
+0.070 (+4.0%)
+0.011 (+0.6%)
apical
O-trans
2a
8a
increment
equatorial
P(1)-C(1)
P(1)-C(2)
P(1)sC(3)
P(1)sO(1)
P(1)sO(2)
1.822
1.822
1.818
1.769
1.764
1.857
1.856
1.668
1.823
1.823
+0.035 (+1.9%)
+0.034 (+1.9%)
-0.150 (-8.3%)
+0.054 (+3.1%)
+0.059 (+3.3%)
apical
interaction, and this is reflected in the P-C(3) bond distances
that are much shorter than the corresponding distances in 1a
and 2a. The double bond character is consistent with the large
P-C coupling constants in the anions. Although the degrees of
P-C(3) bond shortening are similar in 7a and 8a, the degrees
of elongation in the other equatorial bonds are quite different.
Notably, the equatorial P-O bond is ca. 0.1 Å longer in 7a
than in 1a, whereas the P-C bond in the O-trans anion is only
0.04 Å longer than that in 2a. This suggests that the interaction
is considerably stronger in 7a than in 8a, and thereby the anion
is more stabilized in 7a because of the larger electron delocal-
ization. The stability of 7a and 8a is reflected in the relative
energies. While the energy difference between 1a and 2a is as
large as 14.1 kcal/mol, it is reduced to only 4.7 kcal/mol between
7a and 8a (Figure 3). The difference of 9.4 kcal/mol corresponds
to an increase in the acidity of O-cis phosphorane 1a as
compared with O-trans phosphorane 2a, at least in the gas phase.
This is consistent with the observation that the acidity of the
benzyl protons in O-cis 1c is more enhanced than that of O-trans
2c, as shown in Scheme 7.
Reaction of the r-Anion from O-cis 1c: Stabilization of
Hexacoordinated Anions Derived from the Reaction of the
r-Anion from O-cis 1c (Trans Influence of the P-O Bond).
(i) Bromination and Acylation. With the anions in hand,
several reactions of the anions were carried out, and we found
Figure 3. Calculated molecular orbitals and the energies of σ*P-O orbital
in 1 and σ*P-C orbital in 2, which should be responsible for the difference
in the nucleophilic attack.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13159

A R T I C L E S
Matsukawa et al.
Figure 4. Calculated structures and the energies of 1a and 2a and the conjugated bases (7a and 8a).
Scheme 8
notable differences in the stability of the reaction products and
the stereochemical results of the reactions. For example,
bromination of the anion from O-trans 2c by BrCF2CF2Br
afforded 9-SP*S* as the major product in an ca. 12:1 ratio as a
result of direct bromination of the anion, whereas that from O-cis
1c furnished 9-RP*S*as the predominant product in an ca. 12:1
ratio as a result of bromination followed by pseudorotation
(Scheme 8). Recrystallization from n-hexane provided pure
products for both cases. The relative configurations of 9-RP*S*
and 9-SP*S* were determined by X-ray analysis (see Supporting
Information). In the case of the reaction from O-cis 1c, the
electronegative nature of the bromine substituent at the R-posi-
tion in O-cis phosphorane 10 is probably the reason for the
acceleration of pseudorotation to the corresponding O-trans
isomer 9-RP*S*. In the case of the reaction of O-trans 2c, the
major product 9-SP*S* does not undergo pseudorotation to its
enantiomeric O-trans 9-RP*S* at ambient temperatures. There-
fore, the observed product ratio reflects the selectivity in the
9
13160 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Scheme 9
Scheme 10
Scheme 11
bromination reaction. In the case of the reaction of O-cis 1c,
however, the observed ratio is related to the barriers of
pseudorotation of the two separate pathways from the initial
brominated O-cis products to the two diastereomeric products,
9-RP*S* and 9-SP*S*. The reason is because (although there
could be selectivity in the bromination reaction) this information
is lost immediately, since the one-step pseudorotation process
involving the monodentate alkyl group as the pivot that inverts
the stereochemistry at the phosphorus atom is a low-energy
process12a and rapid epimerization gives rise to an equilibrium
mixture of 10-RP*S* and 10-SP*S* (Scheme 9). Configurations
where a five-membered ring is positioned diequatorially are
regarded to be highly unfavorable because of ring strain and
thus can be considered as (or structurally very close to) global
energy maximums of multistep pseudorotation processes of
spirophosphoranes. Therefore, the fact that 9-RP*S* was the
major isomer for O-cis 1c indicates the preference for [SP*S*]q
over [RP*S*]q, although the reason is unclear.
Differences in the reactivity of the R-anion between O-cis
1c and O-trans 2c were also observed in acylation. Scheme 10
shows reactions with benzoyl chloride. As reported previously
by us, the deprotonation of O-trans 2c with n-BuLi, followed
by reaction with benzoyl chloride, gave R-benzoylated phos-
phorane 11-SP*R* diastereoselectively.19 However, the reaction
of O-cis 1c under similar conditions provided unexpected
phosphorane 12. Obviously, compound 12 was formed via
rearrangement of the benzoylated product followed by pseu-
dorotation. This rearrangement can be rationalized to have been
induced by the enhanced electrophilicity of the σ*P-O orbital.
The structure of 12 was confirmed by X-ray analysis (see
Supporting Information).
Using isobutyl chloroformate as an electrophile, the reaction
proceeded in a similar manner as the reactions in Schemes 8
and 10. O-trans 2c provided only one diastereomer 13-SP*R*
(Scheme 11). On the other hand, the reaction of O-cis 1c gave
O-trans hydroxyphosphorane 14. In this case, the reaction
product 15 was unstable to water and hydrolyzed during aqueous
workup to form 14. The structure of 13-SP*R* was confirmed
by X-ray analysis (see Supporting Information).
(ii) Reaction with Aldehydes. Previously, we have reported
on the olefin formation reaction of all four diastereomers of
O-trans-â-hydroxyethylphosphoranes 16, which were indepen-
dently prepared (Scheme 12). It was found that olefin formation
was stereospecific, that the rate of olefin formation was
countercation dependent (K > Na > Li), and that the reaction
proceeded at temperatures lower than -40 °C in the case of
K+.19 Here, the reaction of O-trans benzylphosphorane 2c with
PhCHO was first examined.
The reaction smoothly proceeded to afford stilbenes and 14
in high yields (Scheme 13). Z-Stilbene was formed predomi-
nantly in the reaction, and the Z/E ratio was up to 80:20 using
KH as the base in the presence of 18-crown-6 ether (total yield
of stilbene was 98%).
Likewise, we carried out reactions of the corresponding anion
of O-cis 1c with PhCHO with KH as the base. To our surprise,
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13161

A R T I C L E S
Matsukawa et al.
Scheme 12
Scheme 13
Figure 5. ORTEP drawing of 18B showing the thermal ellipsoids at the
30% probability level.
Table 3. Total Yield and the Ratio 18A/18B
base
time (h)
yield (%)
18A/18B
Figure 6. Low-temperature 31P NMR signals for a rapidly equilibrating
mixture of 18B-RP*S*S* and 18B-SP*S*S*.
n-BuLi
5
10
20
43
42
45
47:53
31:69
15:85
the monodentate as the pivot is rapid, as described above, 18B
likewise existed as a rapidly equilibrating mixture of RP*S*S*
KH
5
10
20
75
71
73
87:13
83:17
78:22
and SP*S*S*, each of which could be observed as different 31
P
NMR signals (δP -4.5, 0.2 ppm) at low temperatures in CDCl3
(Figure 6).
the adducts 18 (i.e., two diastereomers that are diastereomeric
because of the difference in relative configuration upon vicinal
carbon atoms) were obtained at ambient temperature, and
stilbenes were not formed at all (Scheme 14).
Figure 7 shows 31P NMR signals of the reaction of 7c-Li
with PhCHO using n-BuLi as the base. The high-field chemical
shifts (-100 to ca. -120 ppm) indicate the formation of
hexacoordinate phosphorus species, and pentacoordinate species
could not be observed. Signals B1 and B2 increased with time
at the expense of signal A, resulting in only signals B1 and B2
at the end. Thus, signal A corresponds to the hexacoordinate
phosphate (Int-A in Scheme 16) that affords 18A by acidic
hydrolysis. On the other hand, signals B1 and B2 are for the
hexacoordinate phosphates (Int-B1 and B2 in Scheme 16) that
lead to 18B after hydrolysis.
These two diastereomers (18A and B) could not be separated
by TLC. However, after prolonged reaction time (>24 h) using
n-BuLi as the base, equilibration took place between the
â-alkoxides of 18A and B to provide only 18B after aqueous
workup of the reaction mixture (Table 3). The structure of 18B
was determined by X-ray analysis to assume the relative
configuration of RP*S*S* (Figure 5). Since one-step pseudo-
rotation at the phosphorus atom of O-cis spirophosphoranes with
Scheme 14
9
13162 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Scheme 15
Scheme 16
The reaction behavior of the anion from p-fluorophenylmethyl
derivatives 1d with PhCHO or (p-FC6H4)CHO was similar to
that of 1c and PhCHO. The two obtained pairs of diastereomers
19A and B and 20A and B could not be separated in either
case. Longer reaction times (6 days; not optimized) were
required to obtain single isomers (19B and 20B). These products
were characterized by NMR and elemental and X-ray analyses
(see Supporting Information) (Scheme 15).
31P NMR monitoring of the products of the reaction of the
anion generated from 1c with n-BuCHO showed slower but
similar equilibration. In this case, the diastereomers (21A and
B) could be separated by TLC, and the structure of both 21A
(Figure 8) and B (see Supporting Information) could be
confirmed by X-ray analysis.
A possible mechanism of the reaction of O-cis benzyl anion
7c-M with RCHO is illustrated in Scheme 16. Kinetically, the
addition of RCHO to the anion proceeds in a manner in which
the phenyl (the large phosphorus moiety) and the R groups avoid
Figure 7. Time course of the 31P NMR signals of the reaction of 7c-Li
with PhCHO.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13163

A R T I C L E S
Matsukawa et al.
Figure 9. Time course of the 31P NMR signals of the deprotonation reaction
of 18B with KH.
Isolation and Thermolysis of Hexacoordinate Phosphate
Bearing an Oxaphosphetane Ring. When the thermodynamic
product 18B was deprotonated by KH in THF to form the
hexacoordinate phosphate, three 31P NMR signals (δP -104 to
ca. -110) were observed in the solution at the initial stage
(Figure 9). The ratio of the peaks changed gradually, and finally
only a signal at -110 ppm remained after equilibration for 9
days. The fact that only one isomer, which should be sterically
the most stable one (22B-K), existed in the system indicated
the possibility of isolation and stimulated us to investigate the
structure. We were successful in obtaining single crystals of
the phosphate 22B-K(18-crown-6) by recrystallizing from a
n-hexane and CH2Cl2 mixture under an inert atmosphere
(Scheme 17). Surprisingly, these crystals were stable to air at
room temperature for several months. A practical procedure for
isolation was as follows: 18B was deprotonated with KH in
CH2Cl2 in the presence of 18-crown-6 ether. The remaining KH
was removed by filtration, and the resulting solution was allowed
to stand for 9 days at room temperature to convert 22A-K(18-
crown-6) to 22B-K(18-crown-6) completely. Finally, freshly
distilled n-hexane was carefully added on to the solution, and
the resulting two-phase solution was allowed to stand at room
temperature to form colorless crystals. The isolated phosphate
22B-K(18-crown-6) showed a singlet at -113.3 ppm in 31P
NMR (CD3CN), characteristic of a hexacoordinate phosphate.
In 19F NMR (CD3CN), four distinguishable quartets were
observed. These observations indicated that 22B-K(18-crown-
6) had a hexacoordinate structure also in polar CD3CN solution.
The X-ray structure of 22B-K(18-crown-6) is shown in Figure
10. This presents the first structural characterization of a
hexacoordinate phosphate bearing an oxaphosphetane ring,
which is similar to intermediates in the Wittig reaction involving
a pentacoordinate oxaphosphetane. The structure is a slightly
distorted octahedral, and the two phenyl groups on the four-
membered ring are located trans to each other. The most
important finding is that the structure has a configuration that
could be assumed to have formed by the attack of the alkoxide
anion to the σ*P-O orbital (anti to the equatorial oxygen atom)
of O-cis phosphorane 18B. We25 and Kawashima et al.26 have
reported recently on the NMR observation of hexacoordinate
phosphates (23) having an oxaphosphetane ring unsubstituted
at the position R to phosphorus. These were prepared from
Figure 8. ORTEP drawing of 21A showing the thermal ellipsoids at the
30% probability level.
mutual steric congestion at the aldol condensation stage, and
the resulting phosphate Int-A, where the phenyl and the R
groups become overlapped by forming the oxaphosphetane ring,
is the predominant intermediate. Int-A gives isomer A by
treatment with a proton source. However, Int-A is expected to
be relatively unstable because of the steric repulsion between
the phenyl and the R groups on the four-membered ring, and
therefore, retro-aldol cleavage takes place from Int-A to
regenerate 7c-M. On the other hand, when the addition proceeds
in an opposite manner, intermediates Int-B (B1 + B2) are
generated. Both Int-B (B1 + B2) isomers are sterically less
hindered and should be more stable intermediates than Int-A.
After equilibrium between Int-A and Int-B (B1 + B2) is
established, only Int-B (B1 + B2) is observed in the reaction
solution, and 18B is obtained as the sole adduct after hydrolysis
in this system. Int-B (B1 + B2) should correspond to the signals
B1 and B2 in Figure 7. In principle, for the intermediate
phosphates, four isomers are conceivable for both Int-A and
Int-B, arising from the fast interconversion upon the phosphorus
atom and from the site of ligation of the â-oxide anion (anti to
either carbon or oxygen atoms in the equatorial plane). The
identities of the intermediate phosphates could not be deter-
mined. However, we believe that the observed isomers are the
ones in which the â-oxide anion ligates anti to the equatorial
oxygen atom, since the formation of the strained four-membered
ring should be boosted by the trans influence of the oxygen
atom. As for the relative stereochemistry between the phos-
phorus moiety and the R-substituent, the major product is
probably the one having the apical oxygen and the R-substituent
in a syn relationship, since interaction between CF3 (of the
inverted Martin ligand) and Ph groups can be avoided. For the
phosphate(s) corresponding to signal A, the strongly unfavorable
doubled interaction in the isomer, which is not shown here, must
be the reason only one isomer is observed. The fact that the
relative ratio of the B (B1 + B2) signals was constant throughout
the conversion of Int-A to Int-B coincides with the rapid
interconversion upon the phosphorus atom.
(25) Kojima, S.; Akiba, K.-y. Tetrahedron Lett. 1997, 38, 547-550.
(26) Kawashima, T.; Watanabe, K.; Okazaki, R. Tetrahedron Lett. 1997, 38,
551-554.
9
13164 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Scheme 17
22. In our experimental results, not any formation of O-cis
pentacoordinate compounds was indicated after aqueous treat-
ment of 23.25
A possible interpretation of the observations is illustrated in
Scheme 19. Compound 18B exists as the equilibrium mixture
of RP*S*S* and SP*S*S*, as previously shown in Scheme 14.
The X-ray structure of 18B reveals that the configuration
preferred in the solid state corresponds to RP*S*S, and hydrogen
bonding is observed between the hydroxy hydrogen atom and
the apical oxygen atom. Because of this hydrogen bonding, steric
repulsion between the phenyl group R to phosphorus and the
CF3 group of the inverted Martin ligand can be envisioned in
the SP*S*S isomer. Thus, it seems reasonable to assume that
the equilibrium lies to the side of RP*S*S* also in solution. If
the ratio of the equilibrium RP*S*S*/ SP*S*S* is maintained after
deprotonation to some degree and the resulting alkoxide anions
cyclize rapidly, phosphate 22A-K would be the initial major
product as observed as signal A2 (or A1) in NMR (Figure 9).
However, phosphate 22A-K is unstable because of steric
repulsion between the phenyl and the trifluoromethyl groups;
therefore, ring cleavage reaction of the P-O bond of the
oxaphosphetane ring in 22A-K takes place. On the other hand,
phosphate 22B-K generated from the minor alkoxide SP*S*S*
has a less hindered structure and therefore should be the most
stable isomer. On the basis of these considerations, signal B in
Figure 9 corresponds to the phosphate 22B-K in Scheme 19,
and one of signals A1 and A2 should correspond to 22A-K.
Since the reaction time for the equilibration from 22A-K to
22B-K was quite long and not well-reproducible, trace amounts
of water probably coordinate to the oxygen of the four-
membered ring and accelerate the conversion. The identity of
the species corresponding to signal A1 (or A2) eludes us. One
possibility is that it is a species formed by the oxide attack anti
to the equatorial carbon atom. As experimentally demonstrated
(vide supra), this would be an unfavorable process. However,
because of the constraint imposed by the tether, some other
factors such as sterics might be able to facilitate an intra-
molecular nucleophilic attack of the oxide to the equatorial
carbon.
Figure 10. ORTEP drawing of 22B-K(18-crown-6) showing the thermal
ellipsoids at the 30% probability level.
O-trans 24 (Scheme 18). The observations that all hexa-
coordinate species 23 could be observed at room temperature
and were thermally very stable at higher temperatures except
23d are in contrast with our recent results on hexacoordinate
phosphates derived from O-trans benzylphosphorane 2c with
PhCHO (Schemes 12 and 13) that were found to be unstable
and to easily decompose to give olefin below -40 °C. These
differences should be due to the relatively low energy for the
cleavage of the P-C(benzyl) bond in 17 (in Scheme 12),
showing that substituents are important factors in determining
the stability of oxaphosphetanes. Compound 23d, having
strongly electronegative groups (CF3), gave olefin after heating
to 80-100 °C, indicating that the P-CR bond energy is
important. On the basis of these facts, it is quite noteworthy
that 22A and B having a P-C(benzyl) bond are so stable, clearly
showing that the trans influence of the P-O bond serves to
stabilize the anionic species as a whole and even the otherwise
weak P-C(benzyl) bond, by the inductive electron-withdrawing
nature of the oxygen atom. The thermolysis of 22 will be
discussed later (vide infra).
Selected structural parameters for 18B and 22B-K(18-crown-
6) are summarized in Table 4. Upon conversion of 18B to 22B-
K(18-crown-6), the equatorial bond distances P(1)-O(2), P(1)-
C(1), and P(1)-C(3) increased remarkably. These increments
were due to the transformation of the equatorial (sp2) bonds of
18B to the weaker hypervalent (3c-4e) bonds of 22B-K(18-
crown-6). Especially, the increment of P(1)-O(2) was 0.135
Å (+8.2%). This indicates the large flexibility of the P-O bond
as compared with the P-C bonds (+3.0%) of P(1)-C(1) and
P(1)-C(3). It could be that this nature prevents the P(1)-C(3)
Since single crystals of 23 could not be obtained, the
structures of 23 were estimated based on NMR. We cannot
confirm the previous structural discussions with certainty
because of the difference in substituents between 22 and 23.
However, isomer 3 (in Scheme 18), which was proposed by
Kawashima to be one of the three isomers,26 seems to be
incorrect because isomer 3 is expected to give an O-cis
pentacoordinate adduct after treatment with water just like with
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13165

A R T I C L E S
Matsukawa et al.
Scheme 18
Scheme 19
bond from being weakened enough to cleave and furnish olefins
upon the increase of the coordination number from 5 to 6. It is
noteworthy here that there is almost no change of bond lengths
of the original hypervalent bonds of P(1)-O(1) and P(1)-C(2).
A structural comparison of the oxaphosphetane moiety
between 22B-K(18-crown-6) and pentacoordinate oxaphosphe-
tanes 25,27 26,28 27,29 and 2829 was carried out (Figure 11 and
Table 5). Compounds 25 and 26 were isolated by Kawashima-
9
13166 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
Table 4. Comparison of Bond Distances (Å) between 18B and
22B-K(18-crown-6)
(Scheme 20). Kinetic measurements for the reaction were carried
out by NMR to determine activation parameters using the Eyring
plot (Figure 12). This is in great contrast to the decomposition
of 12-P-6 phosphates with K+ as the countercation generated
from isomeric O-trans phosphoranes (Scheme 21), which
occurred at temperatures even below -40 °C.19 A countercation
effect similar to that observed for O-trans phosphoranes was
observed when the decomposition rate of 22B-K was compared
with that of 22B-K(18-crown-6) at 80 °C (Figure 13). The
decomposition of the latter was 37 times faster than that of the
former. Since crown ether was not used in the decomposition
of O-trans phosphoranes (Na+ salt), the actual difference in the
decomposition rate between the phosphate from O-cis phos-
phorane and that from the O-trans isomer would be even larger.
On the basis of these results, 22B-K(18-crown-6) was revealed
to be thermodynamically quite stable, although the phosphate
was a hexacoordinate compound with elongated weakened
bonds.
18B
22B-K(18-crown-6)
increment (Å)
P(1)sO(1)
P(1)sO(2)
P(1)sO(3)
P(1)sC(1)
P(1)sC(2)
P(1)sC(3)
O(3)sC(4)
C(3)sC(4)
1.793(1)
1.653(1)
1.782(1)
1.788(1)
1.731(1)
1.891(2)
1.872(2)
1.926(2)
1.439(2)
1.533(2)
-0.011(2)
(-0.6%)
(+8.2%)
+0.135(2)
1.835(2)
1.872(2)
1.869(2)
1.425(2)
1.558(2)
+0.056(4)
(0.000(4)
+0.057(4)
+0.014(4)
-0.025(4)
(+3.1%)
((0.0%)
(+3.0%)
(+1.0%)
(-1.6%)
Table 5. Selected Structural Parameters for the Oxaphosphetane
Moiety
PsO (Å)
PsCR (Å)
OsPsCR (deg)
22B-K(18-crown-6)
1.731(1)
1.728(2)
1.781(6)
1.748(3)
1.657(3)
1.926(2)
1.808(4)
1.823(9)
1.823(4)
1.916(5)
75.97(6)
77.4(1)
75.5(3)
77.2(1)
76.7(2)
2527
2628
2729
2829
Kinetic parameters of the decomposition of 22B-K(18-crown-
6) were obtained as ∆H+ ) 27.1 ( 0.8 kcal/mol, ∆S+ ) 0.4 (
Okazaki’s group, and compounds 27 and 28 were isolated in
our laboratory. It should be noted that compound 28 has the
oxygen in the oxaphosphetane ring at the equatorial position,
and therefore 28 is an O-cis phosphorane. The most significant
difference in 22B-K(18-crown-6) as compared with 25-27 is
the P-CR bond distance. That of the former is 1.926(2) Å and
is much longer than those of the latter three, which are ca. 1.82
Å (Table 5). This result reflects the differences of the bonding
nature (i.e., P-CR in 22B-K(18-crown-6) is a weak hypervalent
(3c-4e) bond, while those of 25-27 are essentially sp2 bonds).
On the other hand, the length of P-CR in 28 is comparable
with that of 22B-K(18-crown-6). This observation is quite
reasonable, taking into account the fact that this bond in 28 is
also a hypervalent bond (apical bond).
2.5 eu, and ∆G+ ) 27.0 kcal/mol. The activation entropy
298
(∆S+) for the decomposition was close to zero. In this case, the
P-C bond of the oxaphosphetane ring that is to be cleaved is
already elongated and weakened as a 3c-4e bond; thus,
structural change may not be essential for the decomposition
of 22B-K(18-crown-6). On the other hand, the activation
entropies of the decompositions for 25 (∆S+ ) -4.1 ( 1.2 eu)27
and 26 (∆S+ ) -8.8 ( 1.2 eu)28 were reported as negative
values. In these cases, the negative values can be considered to
be due to considerable change of the structure of the oxaphos-
phetane ring before the formation of olefinic products.
In summary, we have recently discovered a method of
preparing a novel group of anti-apicophilic spirophosphoranes
(O-cis), in which the usually more apicophilic oxygen atom and
the less apicophilic carbon atom have exchanged places as
thermodynamically less stable isomers of phosphoranes of
Finally, we investigated the thermal decomposition of 22B-
K(18-crown-6) in solution. 22B-K(18-crown-6) was heated at
60 °C in THF for 4 days to afford trans-stilbene, quantitatively
Figure 11. Reported oxaphosphetanes and the structure of the oxaphosphetane moiety of 22B-K(18-crown-6).
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13167
9

A R T I C L E S
Matsukawa et al.
Scheme 20
Scheme 21
P-C bond that should be less capable of accepting electrons.
The equatorial P-O delocalization effect was demonstrated here
in two kinds of phenomena. One is in the reaction toward the
phosphorus atom, and the other is in the acidity of the carbon
R to the phosphorus atom. For the former, enhanced reactivity
in nucleophilic reactions toward the phosphorus atom was
observed in the reaction of 1 (O-cis) with fluoride, alkyllithium,
and rearrangement of the carbonyl oxygen of the phenacyl group
compared with 2 (O-trans). Another related observation was
the great stabilization of newly formed hypervalent bonds in
hexacoordinate phosphates formed by intramolecular attack of
the oxide anion anti to the equatorial oxygen in 1 (O-cis),
whereas corresponding phosphates generated from 2 (O-trans)
have been found to decompose much more easily. For the latter,
the difference in the stability of carbanions R to the phosphorus
in 1 and 2 has clearly been demonstrated. Theoretical calcula-
tions on 1, 2 and their conjugated bases 7, 8 supported these
ordinary O-trans configuration. This has paved the way for the
first examination of the difference in reactivity between phos-
phoranes differing in configuration only about the phosphorus
atom. A structural divergence between the two isomers that
catches the eye is the presence of a P-O bond in the equatorial
plane in O-cis phosphoranes. This is expected to provide a rather
low-lying σP-O* orbital capable of accepting electron density,
whereas the corresponding bond in O-trans phosphoranes is a
Figure 12. Eyring plot of the rate constants for the thermal decomposition
of 22B-K(18-crown-6).
Figure 13. Comparison of the rates of the thermal decomposition of 22B-K
and 22B-K(18-crown-6).
9
13168 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002

Reactions and Properties of Spirophosphoranes
A R T I C L E S
benzyl bromide (1.50 mL, 12.6 mmol) was added. The mixture was
stirred for 7 h at room temperature. After the solution was treated with
water (100 mL), the mixture was extracted with Et2O (100 mL × 2),
the ethereal layer was washed with brine (100 mL), and the extract
was dried over MgSO4. The crude white solid was recrystallized from
n-hexane/CH2Cl2 to afford colorless crystals of 2c (4.34 g, 7.15 mmol,
experimental results and gave clearer insights into the present
chemistry. The results presented here show that anti-apicophilic
phosphoranes have provided new insights into a full picture of
the nature of the pentacoordinate state of phosphorus. Here, we
have presented a new, unique, and well-designed example in
the chemistry of stereoelectronic effects. Further investigations
concerning this unique group of compounds are currently
ongoing.
1
90.7%). H NMR (CDCl3, δ) 8.24 (dd, 2H, J ) 10.7, 8.1 Hz), 7.71-
7.57 (m, 6H), 7.15-7.08 (m, 3H), 6.99-6.96 (m, 2H), 3.76 (dd, 1H,
2
2
2JP-H ) 15.6 Hz, JH-H ) 13.7 Hz), 3.71 (dd, 1H, JP-H ) 13.7 Hz,
2JH-H ) 13.7 Hz); 19F NMR (CDCl3, δ) -74.6 (q, 6F, JF-F ) 9.8
4
Experimental Section
Hz), -75.0 (q, 6F, JF-F ) 9.8 Hz); 31P NMR (CDCl3, δ) -22.0; mp
4
General. Melting points were measured with a Yanaco micro melting
128-129 °C. Anal. Calcd for C25H15F12O2P: C, 49.52; H, 2.49.
Found: C, 49.25; H, 2.18.
1
point apparatus and are uncorrected. H NMR (400 MHz), 13C NMR
(100 MHz), 19F NMR (376 MHz), and 31P NMR (162 MHz) spectra
Observation of 12-P-6 Phosphate Bearing a P-F Bond (4). To a
THF (0.6 mL) solution of 1b (25 mg, 0.044 mmol) in an NMR tube
was added TBAF (tetra-n-butylammonium fluoride, 1.0 M solution in
THF, 0.05 mL, 0.05 mmol) at room temperature. Then the NMR tube
was loaded into an NMR spectrometer (JEOL EX-400) at -30 °C.
After 10 min, 31P and 19F NMR spectra were recorded. 19F NMR (THF,
1
were recorded on a JEOL EX-400 spectrometer. H NMR chemical
shifts (δ) are given in ppm downfield from Me4Si, determined by
residual chloroform (δ ) 7.26). 13C NMR chemical shifts (δ) are given
in ppm downfield from Me4Si, determined by chloroform-d (δ ) 77.0).
19F NMR chemical shifts (δ) are given in ppm downfield from external
CFCl3. 31P NMR chemical shifts (δ) are given in ppm downfield from
external 85% H3PO4. Elemental analyses were performed on a Perkin-
Elmer 2400 CHN elemental analyzer.
1
δ) -51.4 (4, br d, 1F, JP-F ) 706 Hz), -72.8 (1b, br s, 6F), -73.2
(4, br s, 3F), -74.1 (4, br s, 9F), -75.3 (1b, br s, 6F); 31P NMR (THF,
1
δ) -0.8 (1b, s), -105.0 (4, d, JP-F ) 706 Hz); 1b/4 ) ca. 1:1.
All reactions were carried out under N2 or Ar. THF and Et2O were
freshly distilled from Na-benzophenone, n-hexane was distilled from
Na, and all other solvents were distilled from CaH2. Preparative thin-
layer chromatography was carried out on plates of Merck silica gel 60
GF254. Merck silica gel 60 was used for column chromatography.
[TBPY-5-12]-1-Phenylmethyl-3,3,3′,3′-tetrakis(trifluoromethyl)-
1,1′-spirobi[3H,2,1,λ5-benzoxaphosphole] (1c). n-BuLi (1.50 M hex-
ane solution, 19.3 mL, 30.3 mmol) was added to a mixture of t-BuOK
(3.41 g, 30.4 mmol) and toluene (3.30 mL, 31.0 mmol) suspended with
n-hexane (15 mL) at room temperature. This mixture was vigorously
stirred for 1 h. At 0 °C, Et2O (35 mL) was added to the mixture, and
then a solution of 3 (5.10 g, 9.88 mmol) in Et2O (60 mL) was transferred
dropwise. The mixture was stirred for 9 h at room temperature. Iodine
(7.90 g, 31.1 mmol) was added at -78 °C, and the resulting mixture
was stirred for 1 h. After the cooling bath was removed, stirring was
continued until the color of the mixture became dark-brown. After the
solution was treated with Na2S2O3 (aq) (100 mL), the mixture was
extracted with Et2O (100 mL × 2), and the ethereal layer was washed
with brine (100 mL) and dried over anhydrous MgSO4. The crude
product was subjected to column chromatography (n-hexane/CH2Cl2
) 2:1). The resulting pale red solid was washed with n-hexane to yield
a crystalline white solid of 1c (4.73 g, 7.81 mmol, 79.0%). Colorless
crystals suitable for X-ray analysis were obtained by recrystallization
from CH3CN. 1H NMR (CDCl3, δ) 7.73-7.71 (m, 2H), 7.66-7.61 (m,
2H), 7.58-7.51 (m, 4H), 7.14-7.11 (m, 3H), 6.98-6.97 (m, 2H), 3.96
(d, 2H, 2JP-H ) 10.7 Hz); 13C NMR (CDCl3, δ) 134.5 (d, 2JC-P ) 16.7
Hz), 134.4 (d, 1JC-P ) 82.5 Hz), 132.4 (d, JC-P ) 12.5 Hz), 132.2 (d,
JC-P ) 2.5 Hz), 131.3 (d, JC-P ) 10.8 Hz), 130.5 (d, JC-P ) 7.5 Hz),
Reaction of the R-Anion of 2c (8a-K(18-crown-6)) with PhCHO.
A THF (1.5 mL) solution of 2c (182 mg, 0.301 mmol) and 18-crown-6
ether (83.0 mg, 0.314 mmol) was added to a THF (0.5 mL) suspension
of KH (excess), and then the mixture was stirred for 20 min at room
temperature. After removal of KH by filtration and washing with THF
(1 mL), benzaldehyde (0.04 mL, 0.39 mmol) was added at 0 °C. The
mixture was stirred for 6 h at room temperature, and then the reaction
was quenched with aqueous NH4Cl. The mixture was extracted with
Et2O (80 mL), and the combined organic solution was washed with
brine (50 mL) and dried over anhydrous MgSO4. After the solvents
were removed by evaporation, the crude product was subjected to TLC
(n-hexane) to afford cis-stilbene (42.1 mg, 0.234 mmol, 77.6%) and
trans-stilbene (10.8 mg, 0.060 mmol, 19.9%). Total yield of stilbene
was 97.5% (cis/trans ) 80:20). cis-Stilbene 1H NMR (CDCl3, δ): 7.29-
7.20 (m, 10H), 6.62 (s, 2H). trans-Stilbene 1H NMR (CDCl3, δ): 7.55
(d, 4H, J ) 7.2 Hz), 7.39 (t, 4H, J ) 7.2 Hz), 7.29 (t, 2H, J ) 7.2 Hz),
7.14 (s, 2H).
Synthesis of 18B. To a THF (15 mL) solution of 1c (912 mg, 1.50
mmol) was added n-BuLi (1.50 M hexane solution, 1.10 mL, 1.65
mmol) at 0 °C. The mixture was stirred for 20 min at 0 °C, and then
benzaldehyde (0.18 mL, 1.77 mmol) was added. The mixture was stirred
for 64 h at room temperature, and then the reaction was quenched with
aqueous NH4Cl. The mixture was extracted with Et2O (150 mL), and
the extract was washed with brine (150 mL) and dried over anhydrous
MgSO4. After the solvents were removed by evaporation, the crude
product was separated by TLC (n-hexane/benzene ) 1:5) to afford a
white solid of 18B (620 mg, 0.870 mmol, 57.8%). Colorless crystals
suitable for X-ray analysis were obtained by recrystallization from
CDCl3. 1H NMR (CDCl3, δ) 7.86-7.84 (m, 1H), 7.71-7.68 (m, 2H),
7.61-7.46 (m, 3H), 7.31-7.27 (m, 1H), 7.13-6.88 (m, 8H), 6.54 (br,
129.8 (d, JC-P ) 10.0 Hz), 128.2 (d, JC-P ) 4.2 Hz), 126.8 (d, JC-P
)
)
1
5.0 Hz), 125.4 (d, JC-P ) 10.0 Hz), 122.1 (dq, J ) 8.3 Hz, JC-F
1
2
286 Hz), 121.8 (q, JC-F ) 287 Hz), 79.8 (septet, JC-F ) 31.7 Hz),
3H), 5.20-5.15 (m, 1H), 4.50-4.00 (br, 1H), 3.72-3.69 (m, 1H); 19
F
47.6 (d, 1JC-P ) 109 Hz); 19F NMR (CDCl3, δ) -75.1 (q, 6F, 4JF-F
)
NMR (CDCl3, δ) -74.3 (br s, 3F), -74.6 (br s, 3F), -75.2 (br s, 3F),
-76.0 (br s, 3F); 31P NMR (CDCl3, δ) -0.6 (br); mp 148-149 °C
(decomp). Anal. Calcd for C32H21F12O3P: C, 53.94; H, 2.97. Found:
C, 53.99; H, 3.08.
8.6 Hz), -76.3 (q, 6F, JF-F ) 8.6 Hz); 31P NMR (CDCl3, δ) -8.0;
mp 134-135 °C (decomp). Anal. Calcd for C25H15F12O2P: C, 49.52;
H, 2.49. Found: C, 49.77; H, 2.39.
4
[TBPY-5-11]-1-Phenylmethyl-3,3,3′,3′-tetrakis(trifluoromethyl)-
1,1′-spirobi[3H,2,1,λ5-benzoxaphosphole] (2c).19 To a Et2O (50 mL)
solution of 3 (4.07 g, 7.88 mmol) was added DBU (1.80 mL, 12.0
mmol) at room temperature. After the mixture was stirred for 30 min,
Isolation of 12-P-6 Phosphate Bearing an Oxaphosphetane Ring
(22B-K(18-crown-6)). 18B (299 mg, 0.420 mmol) and 18-crown-6
ether (113 mg, 0.428 mmol) were dissolved into CH2Cl2 (4 mL), and
then the solution was added to a CH2Cl2 (1 mL) suspension of KH
(excess) at 0 °C. The mixture was stirred for 10 min, and then the
remaining KH was removed by filtration and washed with CH2Cl2 (1
mL). The Schlenk apparatus was sealed, and the resulting pale yellow
filtrate was allowed to stand for 7 days. Dry n-hexane was slowly added
so as to create a two-phase solution, which was allowed to stand for 6
(27) Kawashima, T.; Kato, K.; Okazaki, R. J. Am. Chem. Soc. 1992, 114, 4008-
4010. Correction: J. Am. Chem. Soc. 1998, 120, 6848.
(28) Kawashima, T.; Kato, K.; Okazaki, R. Angew. Chem., Int. Ed. Engl. 1993,
32, 869-870. Correction: Angew. Chem., Int. Ed. Engl. 1998, 37, 1606.
(29) Kojima, S.; Sugino, M.; Matsukawa, S.; Nakamoto, M.; Akiba, K.-y. J.
Am. Chem. Soc. 2002, 124, 7674-7675.
9
J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002 13169

A R T I C L E S
Matsukawa et al.
days at room temperature under a flow of N2. The resulting colorless
crystals of 22B-K(18-crown-6) were collected, washed with dry
n-hexane (10 mL) and dry Et2O (5 mL), and dried in vacuo. Yield:
249 mg, 0.245 mmol, 58.4%. The crystals were stored under N2. H
NMR (CD3CN, δ) 8.16 (t, 1H, JH-H ) 6.7 Hz), 7.64 (m, 2H), 7.49 (d,
1H, JH-H ) 6.7 Hz), 7.45-7.37 (m, 3H), 7.24-7.11 (m, 8H), 7.05-
support of this work through Grant-in-Aid for Scientific
Research (Nos. 09239103, 09440218, 11166248, 11304044,
12304044) provided by the Ministry of Education, Culture,
Science, Sports, and Technology of the Japanese Government
is heartily acknowledged.
1
3
3
7.02 (m, 1H), 6.91 (dd, JP-H ) 12.5 Hz, JH-H ) 7.3 Hz), 5.89 (dd,
3JP-H ) 11.9 Hz, 3JH-H ) 7.6 Hz), 5.28 (dd, 2JP-H ) 2.7 Hz, 3JH-H
)
3
3
Supporting Information Available: Experimental procedures,
details of the kinetic measurements, and details for X-ray
analysis. This material is available free of charge via the Internet
at http://pubs.acs.org.
9.4 Hz), 3.80 (dd, JP-H ) 12.8 Hz, JH-H ) 9.4 Hz), 3.55 (s, 24H);
19F NMR (CD3CN, δ) -74.6 (br s, 3F), -74.7 (m, 6F), -74.9 (br s,
3F); 31P NMR (CD3CN, δ) -113.3; mp >165 °C (decomp). Anal. Calcd
for C44H44F12KO9P: C, 52.07 H, 4.37. Found: C, 51.95; H, 4.24.
Acknowledgment. The authors are grateful to Central Glass
JA026776C
Co. Ltd. for a generous gift of hexafluorocumyl alcohol. Partial
9
13170 J. AM. CHEM. SOC. VOL. 124, NO. 44, 2002