Stereospecific stilbene formation from β-hydroxy-α,β-diphenylethylphosphoranes. Mechanistic proposals based upon stereochemistry

Article abstract of DOI:10.1016/S0040-4020(02)01407-2

All four diastereomers of β-hydroxy-α,β-diphenylethylspirophosphoranes bearing two Martin ligands were prepared from a P-H (equatorial) spirophosphorane. Treatment of these phosphoranes with t-BuOK or t-BuONa at low temperature (ca. -40°C) led to the formation of stilbene with complete stereospecificity. In half of the cases the reaction proceeded quantitatively, whereas the other half gave rise to retro-aldol reaction with the formation of varying amounts of benzaldehyde and benzylphosphorane. A kinetic examination revealed that of the four the quantitatively proceeding reaction that gives cis-stilbene was the fastest. Based on the experimental observations, a mechanistic rationale for the previously reported Wittig-type reaction of exceptionally high Z-selectivity using an ethoxycarbonylmethylspirophosphorane is presented.

Full text of DOI:10.1016/S0040-4020(02)01407-2

Tetrahedron 59 (2003) 255–265
Stereospecific stilbene formation from
b-hydroxy-a,b-diphenylethylphosphoranes.
Mechanistic proposals based upon stereochemistry
*
Satoshi Kojima, Kazuhiro Kawaguchi, Shiro Matsukawa and Kin-ya Akiba
Department of Chemistry, Graduate School of Science, Hiroshima University, Higashi-Hiroshima 739-8526, Japan
Received 9 September 2002; revised 29 October 2002; accepted 30 October 2002
Abstract—All four diastereomers of b-hydroxy-a,b-diphenylethylspirophosphoranes bearing two Martin ligands were prepared from a
P–H (equatorial) spirophosphorane. Treatment of these phosphoranes with t-BuOK or t-BuONa at low temperature (ca. 2408C) led to the
formation of stilbene with complete stereospecificity. In half of the cases the reaction proceeded quantitatively, whereas the other half gave
rise to retro-aldol reaction with the formation of varying amounts of benzaldehyde and benzylphosphorane. A kinetic examination revealed
that of the four the quantitatively proceeding reaction that gives cis-stilbene was the fastest. Based on the experimental observations, a
mechanistic rationale for the previously reported Wittig-type reaction of exceptionally high Z-selectivity using an ethoxycarbonylmethyl-
spirophosphorane is presented. q 2002 Published by Elsevier Science Ltd.
1. Introduction
typically Z:E¼95:5 or better could be achieved with an
ester2b (4a) or amide2d (4b) group as the electron-with-
drawing group regardless of whether the reacting aldehyde
was aromatic or aliphatic. Also found was that 4c, which
bears a cyano2d group, is effective for aliphatic aldehydes
though not for aromatic aldehydes. These compounds are
additions to a rather limited list of reagents that give rise to
the thermodynamically less stable Z-olefins with high
selectivity.8 Mechanistically, we assumed for 4a that the
reaction involved hexacoordinated intermediates, which
have one bond more than the ordinary pentacoordinated
Wittig-type reaction intermediate, and that the high
Z-selectivity was a consequence of the initial aldehyde
addition reaction being the rate determining step and the
relative rate of the reverse reaction (retro-aldol reaction)
being slow. The retro-aldol pathway has been established to
be the cause for the high E-selectivity often observed in
ordinary HWE reactions using reagents bearing electron-
withdrawing groups (7a).9 As for the assumption of
hexacoordination, such intermediates could not be observed
under the reaction conditions for 4a. However, the
quantitative formation of persisting hexacoordinated phos-
phates upon deprotonation of 5 has provided strong
evidence for the intermediacy of such species in general
for this system.2a,10
The Wittig reaction and its variant, the Horner–Wads-
worth–Emmons (HWE) reaction are exceptionally useful
reactions for regioselectively and often stereoselectively
introducing unsaturation with the accompaniment of
carbon–carbon bond formation.1 It is generally assumed
that both of these groups of reagents undergo olefin
formation via 10-P-5 phosphorane intermediates. We2 and
others3–5 have found yet another group of phosphorus
compounds, 10-P-5 phosphoranes, that are already
hypervalent to begin with can also serve as olefination
reagents, thus providing another example of the utility of
hypervalent compounds for organic transformations.6 As
for the limited examples on Wittig type reactions
involving 10-P-5 phosphoranes of ‘stabilized ylide’ and
‘semi-stabilized ylide’ type, those with dative P–N
bonds such as 14 and 25 have the tendency to give high
E-selectivity, whereas neutral phosphoranes such as 33
and 42 show a completely opposite trend favoring Z-
olefins.
We have found that by the use of spirophosphorane 4
bearing Martin ligands,7 extremely high Z-selectivities,
Keywords: hypervalent; phosphorane; pentacoordinated; hexacoordinated;
Wittig reaction; stereospecific; mechanism.
In order to gain evidence on the slow retro-aldol process, we
attempted to prepare 13 and carry out stereochemical
examinations similar to those previously carried out on 7a.9
However, this could not be achieved, so instead, we decided
to look into the closest analogs, i.e. ‘semi-stabilized’ type
*
Corresponding author. Address: Advanced Research Center for Science
and Engineering, Waseda University, 1-4-1 Ohkubo, Shinjuku-ku, Tokyo
169-8555, Japan. Tel./fax: þ81-3-5286-3165;
e-mail: akibaky@waseda.jp
0040–4020/03/$ - see front matter q 2002 Published by Elsevier Science Ltd.
PII: S0040-4020(02)01407-2

256
S. Kojima et al. / Tetrahedron 59 (2003) 255–265
2. Results and discussions
2.1. Preparation and structural determination of all four
diastereomeric b-hydroxy-a,b-diphenylethylspiro-
phosphoranes
Since our initial objective was to gain insight to the
mechanism of the reaction of 4a, we commenced our
investigation by attempting to isolate the aldehyde adduct of
4a by using metals such as Mg2þ which have proven
effective for systems such as 7.9 However, the desired
adducts 13 could not be obtained. We reasoned that this was
because of the low reactivity of the Mg2þ enolate of 4a. We
next turned our attention to the reaction of the Mg2þ salt of
the phosphoranide from 1018,19 (10-Mg2þ) with a suitable
epoxide. However, no reaction occurred upon using ethyl
trans-3-phenyloxiranecarboxylate as the epoxide. With no
alternatives left, we reacted 10-Liþ with the epoxide. No
desired 13-erythro-A or 13-erythro-B could be found in
the mixture. Obtained was a complex mixture and the only
isolable phosphorane species were 11-erythro-A (4%) and
11-erthyro-B (8%) along with hydroxide 12 and a small
amount of ethyl cis-cinnamate (Scheme 1). The stereo-
chemical identity of 11-erythro-B was determined by X-ray
analysis (see Section 4.4) and that of 11-erythro-A in
analogy with 14 (vide infra). Here, 11-erythro-A and 11-
erthyro-B could be viewed as intermediates in the Wittig-
type reaction of 6, a semi-stabilized ylide type reagent and
the fact that these compounds could be isolated implied that
intermediates for semi-stabilized reagents are not as prone
to decomposition as those for stabilized ylide reagents such
as 4, and thus would be more easy to handle. To simplify
matters for further investigations, we decided to examine
analogs of 7 or 8, which give rise to symmetric stilbene.
reagents, 6. Thus, we prepared all four diastereomers of
b-hydroxy-a,b-diphenylethylspirophosphoranes 14 bearing
two Martin ligands and preliminarily reported that olefin is
formed with high stereospecificity upon treatment of
these compounds with sodium or potassium bases.2c In
mechanistic studies involving analogous HWE reagents, it
has been found that treatment of erythro b-hydroxy-a,b-
diphenylethylphosphonates (7b)11 and -phosphine oxides
(7c)12 with strong base lead to high percentages of the
thermodynamically more favorable E-olefin and/or benzyl-
phosphonates (or benzylphosphine oxides). The formation
of the E-olefins was appropriately rationalized to have
arisen via retro-aldol reaction, ensuing recombination to
give the corresponding threo isomer, and consequent
breakdown of the adduct into the observed products.
Although there is a difference in reactivity between ordinary
‘stabilized’ and ‘semi-stabilized’ type HWE reagents, the
predominance of E-olefin formation is much alike.
Exceptions to the E-selectivities have come about by
utilizing of the ‘ring effect’.13 The reactions of 8a14 and
8b15 were found to give Z-stilbene, exclusively or nearly
so. Unlike 7a and 7b, which could be prepared from
HWE reagent and aldehyde in an ordinary manner,
preparation of 8a and 8b required the use of epoxides.
Thus, with semi-stabilized HWE reagents that use
aldehyde as the reactant, it is rather difficult to obtain
Z-olefins with high selectivity and the only satisfactory
methods that have been reported utilize the Wittig reaction
with modifications either in the reagent (9a)16 or reaction
conditions (9b).17 Herein, we provide a full account of the
preliminarily reported results along with a mechanistic
rationale based upon a kinetic examination of the olefination
step.
The preparation of three of the four possible diastereomeric
b-hydroxyethylspirophosphoranes was carried out by treat-
ing 10-Liþ with stilbene oxide at rt in THF. Reaction with
trans-stilbene oxide yielded a pair of anti diastereomers, 14-
erythro-A (SPpS pR p20; y. 49%) and 14-erythro-B (SpPR pS p;
y. 29%), along with a small amount of P–OH phosphorane
Scheme 1.

S. Kojima et al. / Tetrahedron 59 (2003) 255–265
257
Scheme 4.
Scheme 2.
Scheme 5.
harsh conditions, thus indicating that the b-hydroxyethyl-
phosphoranes are sturdy as long as they are not
deprotonated.
Since 14-threo-B could not be obtained in an efficient
manner, an alternative method was sought. The reaction of 6
with n-BuLi followed by the treatment with benzoyl
chloride furnished 15-B (SpPR p) as the sole adduct
(Scheme 5). The relative stereochemistry for this compound
was also established by X-ray analysis (Fig. 2). Reduction of
this compound with LiBH4 was also found to be stereo-
selective, giving only the desired diastereomer 14-threo-B
(Scheme 5).
Other than the furnished isomer, the reduction could have
given rise to the side-chain b-carbon epimer 14-erythro-B,
the diastereomer established by X-ray analysis. As it was
not, chemical correlation allowed the assignment of the
stereochemistry of 14-threo-B to be SpPR pR p as shown and
it followed that the stereochemistry of 14-threo-A was
SpPS pS p. That left 14-erythro-A with the remaining SpPS pR p
configuration, thus permitting the establishment of the
relative stereochemistries of all four stereoisomers.
Figure 1. The ORTEP drawing of 14-erythro-B showing the thermal
ellipsoids at the 30% probability level. All the hydrogens other than C19-
H19, C20-H20 and O3-H3 have been omitted for clarity.
Scheme 3.
12 (y. ca. 5%) and cis-stilbene (y. ca. 10%), the latter two of
which are products of overreaction (Scheme 2). The relative
stereochemistry of 14-erythro-B was established by X-ray
structural analysis (Fig. 1).
On the other hand, the reaction of 10-Liþ with cis-stilbene
oxide (Scheme 3) gave 14-threo-A (SpPS pS p; y. 26%) as the
only isolable adduct, together with trans-stilbene (y. ca.
20%) and oxide 12 (y. ca. 20%).
The remaining diastereomer 14-threo-B (SpPR pR p) was
obtained as a diastereomeric mixture with 14-threo-A
(14-threo-A–14-threo-B¼5:2) upon heating 14-threo-A in
t-butyltoluene at 1908C for 11 days via pseudorotation upon
the phosphorus atom (Scheme 4). No olefination to stilbene
and 12 or reverse reaction to benzaldehyde and benzyl-
spirophosphorane 6 could be observed under the rather
Figure 2. The ORTEP drawing of 15-B showing the thermal ellipsoids at
the 30% probability level. All the hydrogens other than C19-H19 have been
omitted for clarity.

258
S. Kojima et al. / Tetrahedron 59 (2003) 255–265
Scheme 6.
Scheme 7.
Incidentally, treating 15-B with DBU at rt for 2 days
resulted in a ca. 1:1 mixture of 15-B and its side-chain
a-carbon epimer 15-A (SpPS p) (Scheme 6), and as expected,
reduction of 15-A similarly with LiBH4 resulted in the
exclusive formation of 14-threo-A (Scheme 7).
2.2. Rationalization of the stereoselective
transformations
Figure 4. Plausible transition states of the reaction of the benzyl anion 6-
Liþ (generated from 6) with benzoyl chloride.
For the reaction of the phosphoranide 10-Liþ with trans-
stilbene oxide (Scheme 2), since both the oxide and
phosphorane 10 are racemic, not much of a differentiation
was expected, and accordingly a diastereomeric mixture
was obtained. There is an apparent selectivity of
14-erythro-A–14-erythro-B¼2:1, however, it is the latter
of these two, which gives olefin at a higher rate (vide infra).
Thus, the observed ratio could just be a consequence of
advanced decomposition of the latter adduct 14-erythro-B.
Of the two transition states, TS-1A has the remaining
epoxide carbon in the open space between the two
bidentates whereas in TS-1B there is repulsion between
this epoxide carbon and a CF3 group facing it. Thus, A
should be more favored and this coincides with the actual
results. As in the 14-erythro case, there was the possibility
that 14-threo-B was not observed due to decomposition.
However, 14-threo-A has been found to decompose faster
than 14-threo-B (vide infra) so the predominance of 14-
threo-A as a primary reaction product does not change.
cis-Stilbene oxide on the other hand is meso and the
occurrence of selective reaction is possible as observed here
(Scheme 3). The observed selectivity can be rationalized by
assuming that phosphoranide 10-Liþ is high in TBP
character. This is not unreasonable because in an antimony
analog with Sb in the place of P and Et4Nþ in the place of
Liþ in 10-Liþ, we have established approximate TBP
structure by X-ray analysis.21 Upon reaction, bulky
substituents upon an electrophile are expected to be oriented
away from the apical bond since its bond angle is nearly
1808 whereas the corresponding angle involving equatorial
bonds have an angle of about 1208. Thus, in the transition
state, the substituent of primary hindrance, the bulky phenyl
group directly attached to the carbon that undergoes SN2
reaction is expected to have its C(phenyl)–C(epoxide) bond
in the equatorial plane as shown in the drawings in Figure 3.
For the benzoylation reaction (Scheme 5), if we can assume
that chelation is not involved, the approach of the
nucleophilic phosphorus species would be expected to
occur with the bulky phosphorus moiety anti to the carbonyl
oxygen as depicted in Figure 4. In that case four modes of
attack (TS-2A to TS-2D) can be envisioned and of the four,
three will experience either repulsion between the two
phenyl groups (size order Ph.Cl.O) or between the phenyl
group of the acid chloride and a CF3 group of the
phosphorus part, leaving TS-2B as the most likely candidate
to account for the observed selectivity. We cannot strictly
disregard chelation or electrostatic interactions. If these
were in effect, it would be more appropriate to coordinate
the metal to an oxygen atom of the apical bond and to have
the carbonyl oxygen in the coordinating sphere of the metal
as described for the reaction of 4a and aldehyde in the
presence of base (Scheme 12). Of the two plausible
transition states (TS-2E and TS-2F), TS-2E, which is less
prone to steric hindrance, is sufficient for explaining the
observed selectivity.
The reduction processes (Schemes 5 and 7) are rather
straightforward. As shown in Figure 5, the stereochemistry
Figure 3. Plausible transition states of the reaction of phosphoranide
10-Liþ (generated from 10) with cis-stilbene oxide.

S. Kojima et al. / Tetrahedron 59 (2003) 255–265
259
Scheme 9.
Figure 5. Felkin–Anh models for the reaction of LiBH4 with 15-A (TS-3A)
and 15-B (TS-3B).
which was prepared from anti-apicophilic23 phosphorane 16
(Scheme 9).23d Compound 17-Kþ-18-c-6 was found to be
stable at rt and required heating for the formation of olefin.
is consistent with Felkin–Anh selectivity. Similar threo
selectivity (ca. 9:1) has been observed for the reduction of
b-ketophosphine oxides.22
2.4. Kinetic examination of the olefin formation process
2.3. Olefin formation reactions
The fact that the product distribution differed among
diastereomers implied that there was a difference in the
rate of olefin formation. Thus, we decided to carry out rate
measurements on the olefination process at 2408C using
t-BuONa as base. Since it was found that t-BuONa could not
deprotonate 6, and the addition of aldehyde to 6-Mþ was
very slow, the formation of 6-Mþ via retro-aldol reaction at
low temperature could be assumed to irreversibly give 6
upon protonation by t-BuOH. This irreversibility allowed us
to apply first order kinetics.
As anticipated, treatment of 14 with n-BuLi resulted in slow
and rather complex reactions for all four diastereomers.
Upon using NaH, t-BuONa or t-BuOK, 14-erythro-B and
14-threo-B quantitatively and stereospecifically furnished
cis and trans stilbenes, respectively (Scheme 8). 14-
erythro-A and 14-threo-A also furnished cis and trans
stilbenes, respectively, with nearly complete selectivity.
However, varying amounts of retro-aldol product, 6 were
also found to form depending upon reaction conditions.
While the total amount of 6 formation for the latter was less
than 5% and practically negligible in some instances, the
amount of 6 was 30–60% for the former, which gives rise to
cis-stilbene. The fact that all four diastereomers of 14 gave
rise to olefinic product at temperatures as low as 2408C is in
remarkable contrast with hexacoordinated phosphate 17
For 14-erythro-A, the rate of 6 formation was also
measured. The reaction of cis-stilbene-forming 14-erythro-
B was fastest and the measurement had to be carried out at
2608C. As listed in Table 1, the order of olefin formation
was
14-erythro-B.14-threo-B.14-threo-A.14-
erythro-A. At this temperature, for 14-erythro-A, the
retro reaction was found to be almost twice as fast as
the olefin formation reaction. Kinetic parameters for the
reaction of 14-threo-B were calculated to be
DH –¼12.6^0.8 kcal mol21 and DS –¼223.0^3.2 eu
(260 to 2408C). For 17-Kþ-18-c-6, the entropy value
was found to be near zero. The large negative entropy value
here indicates that although the reactant is already in a
hexacoordinated state, even higher order is required for
formation of olefin. This could be due either to solvation of
the countercation to localize charge and weaken the bonds
about the phosphorus atom or to orientation of the phenyl
rings to conformations, which facilitate bond cleavage. The
fact that the presence of 18-c-6 facilitated the reaction of 17
supports the former assumption on bond weakening.
2.5. Attempted observation of hexacoordinated
intermediates
In order to determine whether hexacoordinated species are
actually involved in the reaction, all four diastereomers of
Table 1. Reaction rates
Substrate
Temperature (8C)
k (s21
)
14-erythro-B
14-threo-B
260
250
240
230
240
240
240
(1.44^0.03)£1024
(1.90^0.03)£1025
(7.58^0.14)£1024
(2.02^0.07)£1024
(1.05^0.03)£1024
(1.21^0.02)£1025
a(2.05^0.03)£1025
14-threo-A
14-erythro-A
a
Scheme 8.
Denotes 6 formation reaction.

260
S. Kojima et al. / Tetrahedron 59 (2003) 255–265
Scheme 10.
14 were first treated with n-BuLi in THF at rt only to provide
31P NMR signals assignable to pentacoordinated species.
Upon lowering the temperature to 2308C, however, the
solution containing 14-threo-B showed a sharp signal (dP
(THF) 2112) attributable to a hexacoordinated species in a
4:1 ratio with a pentacoordinated species. Furthermore, the
sodium salt of 14-threo-B (14-threo-B-Naþ) formed upon
treatment with t-BuONa at 2608C lead to complete
conversion to the hexacoordinated species (dP (THF)
2112) while 14-erythro-B (14-erythro-B-Naþ) newly
gave rise to a hexacoordinated species (dP (THF) 2110
(br)) (ca. 1:1) at the same temperature. The 1:1 ratio did not
change upon progression of the olefin forming process, thus
indicating that the 1:1 ratio was the equilibrium ratio
between the pentacoordinated sodium alkoxide of 14-
erythro-B and a hexacoordinated species at the specified
temperature and that equilibration is faster than olefin
formation. Only pentacoordinated species could be
observed for 14-erythro-A and 14-threo-A.
upon the phosphorus atom, ring closure with P–O bond
formation occurs to give hexacoordinated species 18.
Stereomutation upon the phosphorus atom prior to olefin
formation can be disregarded since stereomutation of
similar hexacoordinated phosphates have been determined
to be rather slow even at rt with Kþ or Naþ as the
countercation2a while the actually observed olefin formation
temperature here is well below 08C. Thus, olefin formation
could be regarded to occur directly from these species, and
stereochemical interpretations require only to consider these
configurations. The common feature between 14-erythro-
B-Mþ and 14-threo-B-Mþ, which show no sign of reverse
reaction, is that CF3-a-phenyl interaction is minimized in
their ring closed hexacoordinated isomers 18-erythro-B-
Mþ and 18-threo-B-Mþ, respectively. This absence of
repulsive interaction would be expected to facilitate ring
closure and is in good accord with 31P NMR observations.
In 14-erythro-B-Mþ, 1,2-interaction between the phenyl
groups becomes operative upon ring closure and this serves
as an unfavorable factor for ring-closure compared with 14-
threo-B-Mþ and this is reflected in the fact that the portion
of the hexacoordinated isomer is larger for 14-threo-B-Mþ
than 14-erythro-B-Mþ. The fact that hexacoordinated
species could not be observed for 14-erythro-A-Mþ and
14-threo-A-Mþ indicates that repulsion between CF3 and
the a-phenyl is quite severe. The fact that all the X-ray
structures, i.e., 11-erythro-B, 14-erythro-B, and 15-B,
showed the a-phenyl group and the closest (upper) CF3
groups to be farthest apart also is a good indication that such
interaction is highly unfavorable. In the case of 14-erythro-
A-Mþ, in addition to the CF3-a-phenyl interaction,
2.6. Mechanistic rationalization of the decomposition of
the four diastereomers
The phosphoranes that gave rise to hexacoordinated species
incidentally were the ones that showed no sign of reverse
reaction to generate aldehyde. This implied that facility to
take on hexacoordinated structure was related to ease in
olefin formation. A mechanistic rationale of the decompo-
sition reaction of 14 based on observations is rather
straightforward as shown in Scheme 10. Deprotonation of
14 gives rise to 14-Mþ. Without configurational change

S. Kojima et al. / Tetrahedron 59 (2003) 255–265
261
1,2-phenyl interaction disfavors ring closure and this is
reflected in the fact that the ratio of reverse reaction is
highest for this species. Although hexacoordinated species
are ordinarily expected to be less stable than pentacoordi-
nated species, the use of the unique Martin ligand system
makes the pentacoordinated metal alkoxide less stable than
the hexacoordinated species, in which charge delocalization
by the coordination of the oxygen atom seems to have
overcome the instability caused by the formation of higher
coordination. Thus, for 14-erythro-Mþ and 14-threo-A-
Mþ, it can be said that the favored ring closure has been
inhibited by steric reasons. As for the formation of olefin,
the formation of the hexacoordinated species becomes a
prerequisite, and although this is more favorable for 14-
threo-B, the hexacoordinated species 18-threo-B formed
from 14-threo-B-Mþ is stereochemically stable and other
than the weakness of the P–C bond in this species there are
no other driving forces to help accelerate olefin formation.
On the other hand, in 18-erythro-B, the 1,2-phenyl eclipse
interaction not only serves to facilitate ring opening back to
14-erythro-B-Mþ, but can also serve as a buttress effect to
facilitate P–C bond dissociation. Thus, for 14-erythro-B-
Mþ, the delicate balance in steric hindrance works as an
advantage and as a result olefin formation becomes fastest
for this substrate. Thus, in the reaction of 4a-Mþ2b or
6-Mþ23d with aldehydes, whether or not the retro-aldol
reaction occurs is highly dependent upon which adduct
among 14-Mþ is formed.
from 14 was found to proceed at temperatures as low as
2408C. Olefin formation from 13 is expected to be even
faster. Therefore, it is not unreasonable to say that the rate
determining step for the olefin forming reaction of both
4a-Mþ2b and 6-Mþ23d is definitely the initial addition step.
This is in contrast to HWE reactions where the rates of the
initial aldehyde addition and olefin furnishing decompo-
sition are regarded to be comparable.1 The reasons for the
diverse behavior of the phosphoranes here would be because
of the lower rate of retro-aldol reaction resulting from the
lower thermodynamical acidity of phosphoranes in general,
and the steric hindrance exerted by the extra coordination,
as previously assumed.2b
Although direct verification of the mechanism of the
reaction of 4a could not be carried out, the stereochemical
results gained here provide useful information. Thus,
consideration of the order of olefin formation rate and 31P
NMR observations enables us to make minor revisions of
our tentative rationale on the mechanism involving 4a,
which did not take the stereochemistry upon the chiral (but
in this case not optically active) phosphorus atom in
consideration.2b Enolate 4a-Mþ formed upon deprotonation
of 4a probably has the OM group and the phosphorus
moiety on the same side of the double bond due to either
chelation or electrostatic attraction between the metal and
an oxygen atom of an apical bond (Scheme 12). The
incoming aldehyde should also be involved in the
interaction as shown in the transition state drawings.
Calculations involving HWE reagents also favor similar
spacial arrangement with attractive interaction between the
enolate metal, the phosphoryl oxygen atom and the
aldehyde oxygen.24 Comparing the four plausible transition
states which involve metal–oxygen interaction, TS-tB and
TS-tA are clearly unfavorable due to severe steric
interaction between the phenyl group of the aldehyde and
the phosphorus moiety of the reagent. Of the remaining two
transition states, it is rather difficult to say which is more
favorable. Actually, it would not matter which is preferred
since both lead to the Z-olefin. Since interaction between a
CF3 group and the metal in TS-eB is projected to be
unfavorable, it could be that TS-eA is more likely as the
kinetically favored transition state. In that case, analogy
from isomers of 18 would make 20-erythro-A-Mþ
extremely unfavorable, and both olefin formation and
retro-aldol reaction would be assumed to occur. Upon
regeneration of aldehyde and P enolate via this path,
primary equilibration would be between 13-erythro-B-Mþ
and 13-erythro-A-Mþ, and since the product olefin could
be formed via 13-erythro-B-Mþ with no obstacles,
secondary equilibration involving 13-threo-B-Mþ and 13-
threo-A-Mþ need not be operative, thus accounting for the
observed high Z-selectivity. This again is in contrast to
HWE reactions which have to consider E-olefin precursors
corresponding to 13-threo-B-Mþ and 13-threo-A-Mþ in
pre-equilibrations.
The fact that hexacoordinated phosphates assumed here, in
which two carbon atoms are situated trans to the two
ligating atoms (oxygen and carbon) of the oxaphosphetane
ring as represented by 18-threo-B-Mþ, are less stable than
17-Mþ in which one oxygen atom has become trans to the
oxygen atom of the 4-membered ring indicates that trans
influence is highly related to the stability of the hexa-
coordinated phosphates bearing an oxaphosphetane ring. In
other words, 17 which has a better trans group (oxygen) is
more reluctant to undergo decomposition to olefin. Thus, we
can further anticipate that if we were able to obtain an
isomer in which two oxygen atoms are trans to the atoms of
the oxaphosphetane ring, it would be even more unwilling to
produce olefin (Scheme 11). This thermodynamic propen-
sity is also consistent with 5-Mþ undergoing quantitative
rearrangement to a species analogous to 19-Mþ (differing
only in the position of the phenyl groups on the phosphetane
ring) under equilibration conditions.2a
2.7. Rationalization of the selectivity observed in the
reaction of 4a
The reaction between aldehydes and reagent 4a-Mþ2b or
6-Mþ23d has been found to require temperatures as high as
08C or more. On the other hand, the olefin formation process
The reaction of 6-Kþ with benzaldehyde was found to be
considerably less selective (Z:E¼80:20) compared with the
reaction of 4-Kþ (Z:E¼98:2). Thus, since the retro-aldol
reaction is suppressed, similar considerations would lead to
the conclusion that unlike the reaction of 4-Mþ, the initial
addition of aldehyde to 6-Mþ is not very selective and
Scheme 11.

262
S. Kojima et al. / Tetrahedron 59 (2003) 255–265
Scheme 12.
either 14-threo-B-Mþ or 14-threo-A-Mþ is formed along
with one or two of the precursors to Z-stilbene.
suppressed and this serves to maintain the selectivity due to
a faster protonation reaction. Although reverse reactions are
expected to be more facile for 4a because of the stability of
the anion generated from it, the mechanistic results for 14
allow for a rational plausible mechanism for the reaction of
4a.
3. Conclusion
In summary, we have found that all four diastereomers of
b-hydroxy-a,b-diphenylethylspirophosphoranes bearing
Martin ligands lead to the formation of stilbene with
complete stereospecificity upon treatment with t-BuOK or
t-BuONa at low temperature (ca. 2408C). In half of the
cases the reaction proceeded quantitatively, whereas the
other half gave rise to retro-aldol reaction side reactions and
formation of varying amounts of aldehyde and benzyl-
spirophosphorane. A kinetic examination revealed that of
the four, the quantitatively proceeding reaction that gives
cis-stilbene was the fastest. The reason for the high
specificity can be attributed to the low acidity of the benzyl
hydrogen and to low reactivity of the benzyl anion once
generated due to the sterically hindered environment in its
vicinity. These factors combine to disfavor retro-aldol
reaction and even in cases where retro reaction does
proceed, recombination of aldehyde and benzyl anion is
4. Experimental
4.1. General procedures
Melting points were measured with a Yanaco micro melting
1
point apparatus and are uncorrected. H NMR (400 MHz),
13C NMR (100 MHz), 19F NMR (376 MHz), and 31P NMR
(162 MHz) spectra were recorded on a JEOL EX-400
spectrometer. 1H NMR chemical shifts (d) are given in ppm
from residual chloroform-d (d¼7.26). 13C NMR chemical
shifts (d) are given in ppm from chloroform-d (d¼77.0). 19
F
NMR chemical shifts (d) are given in ppm from external
CFCl3. 31P NMR chemical shifts (d) are given in ppm from
external 85% H3PO4. Elemental analyses were performed
on a Perkin–Elmer 2400 CHN elemental analyzer.

S. Kojima et al. / Tetrahedron 59 (2003) 255–265
263
4
All reactions were carried out under N2. THF was freshly
distilled from Na-benzophenone. Ethanol was distilled from
magnesium. All other solvents were distilled from CaH2.
Preparative thin layer chromatography was carried out on
(CDCl3) d 273.4 (q, JFF¼10.1 Hz, 6F), 274.6 (q,
4JFF¼10.1 Hz, 6F); 31P NMR (CDCl3) d 223.5. Anal.
calcd for C32H21F12O3P: C, 53.95; H, 2.97. Found: C, 54.04;
H, 2.91.
plates of Merck silica gel 60 GF254
.
4.1.3. (SpP,1S p,2S p)-1-(2-Hydroxy-1,2-diphenylethyl)-
3,3,30,30-tetrakis(trifluoromethyl)-1,10-spirobi[3H,2,1l 5-
benzoxaphosphole] (14-threo-A). To P–H (equatorial)
phosphorane 10 (201 mg, 0.389 mmol) in THF (2 mL) was
added n-BuLi (0.27 mL, 0.421 mol) at 08C. To this solution
was added cis-stilbene oxide (86.9 mg, 0.443 mmol) in THF
(2 mL) and the resulting mixture was stirred at rt for 5 h.
The solution was quenched with water, extracted with Et2O
and the combined organic solvents were washed with water
and brine. This was followed by drying over anhydrous
MgSO4 and concentration in vacuo. Separation of the
residue on TLC (hexane–Et2O¼9:1) followed by recrys-
tallization (hexane–CH2Cl2) gave 14-threo-A (72.1 mg,
26%). Mp 196–1978C (dec); 1H NMR (CDCl3) d 8.03 (br s,
2H), 7.75–7.54 (m, 6H), 7.02–6.59 (m, 10H), 5.27 (dd,
4.1.1. (SpP,1S p,2R p)-1-(2-Ethoxycarbonyl-2-hydroxy-1-
phenylethyl)-3,3,30,30-tetrakis(trifluoromethyl)-1,10-
spirobi[3H,2,1l 5-benzoxaphosphole] (11-erythro-A)
and (SpP,1R p,2S p)-1-(2-ethoxycarbonyl-2-hydroxy-1-
phenylethyl)-3,3,30,30-tetrakis(trifluoromethyl)-1,10-
spirobi[3H,2,1l 5-benzoxaphosphole] (11-erythro-B). To
P–H (equatorial) phosphorane 10 (214 mg, 0.415 mmol) in
THF (5 mL) was added n-BuLi (0.27 mL, 0.427 mol) at
08C. To this solution was added ethyl trans-3-phenyloxir-
anecarboxylate (72.0 mg, 0.374 mmol) in THF (2 mL) and
the resulting mixture was stirred at rt for 11 h. The solution
was quenched with 1 M HCl, extracted with Et2O and the
combined organic solvents were washed with water and
brine. This was followed by drying over anhydrous MgSO4
and concentration in vacuo. Separation of the residue on
TLC (hexane–Et2O¼3:1) followed by recrystallization
(hexane–CH2Cl2) gave 11-erythro-A (10.8 mg, 4%) and
11-erythro-B (23.5 mg, 8%). 11-erythro-A: mp 211.5–
2
J¼10.3 Hz, JPH¼8.3 Hz, 1H), 4.02 (dd, J¼10.3 Hz,
2JPH¼17.1 Hz, 1H), 3.72 (s, 1H); 19F NMR (CDCl3) d
273.7 (br s, 6F), 274.5 (br s, 6F); 31P NMR (CDCl3) d
219.3. Anal. calcd for C32H21F12O3P: C, 53.95; H, 2.97.
Found: C, 54.00; H, 2.88.
1
212.58C (dec); H NMR (CDCl3) d 8.04–7.99 (m, 2H),
7.75–7.64 (m, 6H), 7.17–7.01 (m, 5H), 4.81 (s, 1H), 4.25
(d, J¼7.3 Hz, 1H), 4.21–4.12 (m, 1H), 4.01–3.95 (m, 1H),
2.35 (s, 1H), 0.89 (t, J¼7.3 Hz, 3H); 31P NMR (CDCl3) d
221.2. Anal. calcd for C29H21F12O5P: C, 49.17; H, 2.99.
Found: C, 49.13; H, 2.78. 11-erythro-B: mp 156–1578C
4.1.4. (SpP,1R p)-1-(2-Oxo-1,2-diphenylethyl)-3,3,30,30-tet-
rakis(trifluoromethyl)-1,10-spirobi[3H,2,1l 5-benzoxa-
phosphole] (15-B). To benzylphosphorane 6 (1.00 g,
1.65 mmol) in THF (5 mL) was added n-BuLi (1.10 mL,
1.74 mol) at 08C. After stirring at rt for 5 m, benzoyl
chloride (0.200 mL, 1.72 mmol) was added and the
resulting mixture was stirred overnight. The solution was
quenched with water, extracted with Et2O and the combined
organic solvents were washed with water and brine. This
was followed by drying over anhydrous MgSO4 and
concentration in vacuo. Separation of the residue on TLC
(hexane–benzene¼1:1) followed by recrystallization
(hexane–CH2Cl2) gave 15-B (958 mg, 82%). Mp 160–
1618C; 1H NMR (CDCl3) d 8.17–8.12 (m, 2H), 7.76–7.19
(m, 14H), 6.97–6.95 (m, 2H), 5.91 (d, 2JPH¼18.6 Hz, 1H);
1
(dec); H NMR (CDCl3) d 8.26–8.21 (m, 2H), 7.72–7.61
(m, 6H), 7.21–7.00 (m, 5H), 5.34 (dd, J¼3.4 Hz,
2
2JPH¼7.4 Hz, 1H), 4.16 (dd, J¼3.4 Hz, JPH¼23.4 Hz,
1H), 3.96–3.89 (m, 2H), 0.89 (t, J¼7.3 Hz, 3H); 31P
NMR (CDCl3) d 225.0. Anal. calcd for C29H21F12O5P: C,
49.17; H, 2.99. Found: C, 49.10; H, 2.77.
4.1.2. (SpP,1S p,2R p)-1-(2-Hydroxy-1,2-diphenylethyl)-
3,3,30,30-tetrakis(trifluoromethyl)-1,10-spirobi[3H,2,1l 5-
benzoxaphosphole] (14-erythro-A) and (SpP,1R p,2S p)-1-
(2-hydroxy-1,2-diphenylethyl)-3,3,30,30-tetrakis(tri-
fluoromethyl)-1,10-spirobi[3H,2,1l 5-benzoxaphosphole]
(14-erythro-B). To P–H (equatorial) phosphorane 10
(196 mg, 0.380 mmol) in THF (2 mL) was added n-BuLi
(0.25 mL, 0.408 mol) at 08C. To this solution was added
trans- stilbene oxide (75.0 mg, 0.382 mmol) in THF (2 mL)
and the resulting mixture was stirred at rt for 5 h. The
solution was quenched with water, extracted with Et2O and
the combined organic solvents were washed with water and
brine. This was followed by drying over anhydrous MgSO4
and concentration in vacuo. Separation of the residue on
TLC (hexane–benzene¼2:1) followed by recrystallization
(hexane–CH2Cl2) gave 14-erythro-A (132.6 mg, 49%) and
14-erythro-B (78.2 mg, 29%). 14-erythro-A: mp 200.5–
4
19F NMR (CDCl3) d 273.9 (q, JFF¼10.0 Hz, 6F), 274.4
4
(q, JFF¼10.0 Hz, 6F); 31P NMR (CDCl3) d 225.2. Anal.
calcd for C32H19F12O3P: C, 54.10; H, 2.70. Found: C, 53.98;
H, 2.49.
4.1.5. (SpP,1S p)-1-(2-Oxo-1,2-diphenylethyl)-3,3,30,30-
tetrakis(trifluoromethyl)-1,10-spirobi[3H,2,1l 5-benz-
oxaphosphole] (15-A). A mixture of phosphorane 15-B
(300 mg, 0.422 mmol) and DBU (0.13 mL, 0.85 mmol) in
Et2O (3 mL) was stirred at rt for 52 h. After removal of
solvent, the residue was separated on TLC (hexane–
benzene¼1:1) followed by recrystallization (hexane–
1
CH2Cl2) to give 15-A (111 mg, 37%). Mp 189–1908C; H
1
201.58C (dec); H NMR (CDCl3) d 7.60–7.08 (m, 18H),
2
NMR (CDCl3)
d
7.70–7.12 (m, 18H), 5.91 (d,
5.36 (d, J¼10.3 Hz, 1H), 4.29 (dd, J¼10.3 Hz, JPH
¼
2JPH¼26.9 Hz, 1H); 19F NMR (CDCl3) d 273.5 (q,
19.5 Hz, 1H), 1.83 (s, 1H); 19F NMR (CDCl3) d 271.9 (q,
4JFF¼10.1 Hz, 6F), 273.3 (q, 4JFF¼10.1 Hz, 6F); 31P NMR
(CDCl3) d 221.9. Anal. calcd for C32H21F12O3P: C, 53.95;
H, 2.97. Found: C, 54.12; H, 2.87. 14-erythro-B: mp
4JFF¼9.8 Hz, 6F), 274.6 (q, JFF¼9.8 Hz, 6F); 31P NMR
4
(CDCl3) d 224.3. Anal. calcd for C32H19F12O3P: C, 54.10;
H, 2.70. Found: C, 54.08; H, 2.53.
1
203.5–204.58C (dec); H NMR (CDCl3) d 7.62–7.05 (m,
2
4.1.6. (SpP,1R p,2R p)-1-(2-Hydroxy-1,2-diphenylethyl)-
3,3,30,30-tetrakis(trifluoromethyl)-1,10-spirobi[3H,2,1l 5-
benzoxaphosphole] (14-threo-B). A mixture of phosphorane
18H), 5.31 (dd, J¼10.3 Hz, JPH¼4.9 Hz, 1H), 4.47 (dd,
2
J¼10.3 Hz, JPH¼16.1 Hz, 1H), 1.90 (s, 1H); 19F NMR

264
S. Kojima et al. / Tetrahedron 59 (2003) 255–265
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford,
1991; Vol. 1, pp 730–817. (e) Johnson, A. W. Ylides and
Imines of Phosphorus; Wiley: New York, 1993. (f) Vedejs, E.;
Peterson, M. J. Top. Stereochem. 1994, 21, 1–157. (g) Vedejs,
E.; Peterson, M. J. Adv. Carbanion Chem. 1996, 2, 1–85.
(h) Clayden, J.; Warren, S. Angew. Chem., Int. Ed. Engl. 1996,
35, 241–270. (i) Kolodiazhnyi, O. I. Phosphorus Ylides.
Wiley: Weinheim, 1999. (j) Ando, K. J. Synth. Org. Chem.,
Jpn 2000, 58, 869–876.
15-B (30 mg, 0.042 mmol) and LiBH4 (3.0 mg,
0.138 mmol) in THF (3.5 mL) was stirred at rt for 20 h.
The solution was quenched with water, extracted with Et2O
and the combined organic solvents were washed with water
and brine. This was followed by drying over anhydrous
MgSO4 and concentration in vacuo. Separation of the
residue on TLC (hexane–benzene¼1:1) followed by
recrystallization (hexane–CH2Cl2) gave 14-threo-B
(24.1 mg, 80%). Mp 165–165.58C (dec); 1H NMR
(CDCl3) d 8.08 (br s, 2H), 7.68–6.93 (m, 16H), 5.29 (ddd,
2. (a) Kojima, S.; Akiba, K.-y. Tetrahedron Lett. 1997, 38,
547–550. (b) Kojima, S.; Takagi, R.; Akiba, K.-y. J. Am.
Chem. Soc. 1997, 119, 5970–5971. (c) Kojima, S.;
Kawaguchi, K.; Akiba, K.-y. Tetrahedron Lett. 1997, 38,
7753–7756. (d) Kojima, S.; Kawaguchi, K.; Matsukawa, S.;
Uchida, K.; Akiba, K.-y. Chem. Lett. 2002, 170–171.
3. Bojin, M. L.; Barkallah, S.; Evans, Jr. S. A. J. Am. Chem. Soc.
1996, 118, 1549–1550.
2
J¼11.2, 4.9 Hz, JPH¼11.2 Hz, 1H), 4.33 (dd, J¼11.2 Hz,
2JPH¼13.2 Hz, 1H), 1.82 (d, J¼4.9 Hz, 1H); 19F NMR
4
4
(CDCl3) d 273.6 (q, JFF¼8.5 Hz, 6F), 274.2 (q, JFF¼
8.5 Hz, 6F); 31P NMR (CDCl3) d 219.1. Anal. calcd for
C32H21F12O3P: C, 53.95; H, 2.97. Found: C, 54.04; H, 2.91.
4.2. General procedure for the olefination reaction
´ ´
4. (a) Carre, F.; Chauhan, M.; Chuit, C.; Corriu, R. J. P.; Reye, C.
To a mixture of 14 and an excess of t-BuOK (or t-BuONa)
was added a specified solvent at 2788C. After stirring at a
specified temperature, the solution was worked up and the
residue was separated on TLC (hexane) to give the products.
Phosphorus Sulfur Silicon 1997, 123, 181–195. (b) Chuit, C.;
´
Reye, C. Eur. J. Inorg. Chem. 1998, 1847–1857.
5. (a) Wang, Z.; Verkade, J. G. Tetrahedron Lett. 1998, 39,
9331–9334. (b) Wang, Z.; Zhang, G.; Guzei, I.; Verkade, J. G.
J. Org. Chem. 2001, 66, 3521–3524.
4.3. Kinetic measurements of the olefin formation
process
6. Chemistry of Hypervalent Compounds; Akiba, K.-y., Ed.;
Wiley: New York, 1999.
7. (a) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48,
4155–4156. (b) Martin, J. C. Science 1983, 221, 509–514, and
references therein.
Phosphorane 14 (ca. 10 mg) and t-BuONa (2 equiv.) were
placed in NMR tubes and flushed with N2. To the tubes was
added freshly distilled solvent (0.5–0.6 mL) at liquid N2
temperature, and the tubes were then sealed under N2 while
contents were still frozen. The decomposition reactions
were monitored by measuring 19F NMR spectra in a variable
temperature mode, and the specified temperatures were
maintained throughout each set of measurements (error
within ^18C). The data were analyzed by assuming
irreversible first-order kinetics using the equation lnðc0=cÞ ¼
kT; in which c0¼concentration of reactant at t¼0, c¼
concentration of reactant at arbitrary intervals.
8. For examples of the ester group see: (a) Still, W. C.; Gennari,
C. Tetrahedron Lett. 1983, 24, 4405–4408. (b) Ando, K.
Tetrahedron Lett. 1995, 36, 4105–4108. (c) Ando, K. J. Org.
Chem. 1997, 62, 1934–1939. For examples of the amide group
see: (d) Kojima, S.; Inai, H.; Hidaka, T.; Ohkata, K. Chem.
Commun. 2000, 1795–1796. (e) Kojima, S.; Inai, H.; Hidaka,
T.; Fukuzaki, T.; Ohkata, K. J. Org. Chem. 2002, 67,
4093–4099. (f) Ando, K. Synlett 2001, 1272–1274.
(g) Kojima, S.; Hidaka, T.; Ohba, Y.; Ohkata, K. Phosporus,
Sulphur, Silicon 2002, 177, 729–732. For examples of the
keto group see: (h) Yu, W.; Su, M.; Jin, Z. Tetrahedron Lett.
1999, 40, 6725–6728. (i) Yamakado, Y.; Ishiguro, M.; Ikeda,
N.; Yamamoto, H. J. Am. Chem. Soc. 1981, 103, 5568–5570.
(j) Furuta, K.; Ishiguro, M.; Haruta, R.; Mori, A.; Ikeda, N.;
Yamamoto, H. Bull. Chem. Soc. Jpn 1984, 57, 2768–2776.
(k) Sato, Y.; Niinomi, Y. J. Chem. Soc., Chem. Commun.
1982, 56–57. (l) Matsuda, I.; Okada, H.; Izumi, Y. Bull. Chem.
Soc. Jpn 1983, 56, 528–532. (m) Palomo, C.; Aizpurua, J. M.;
Aurrekoetxea, N. Tetrahedron Lett. 1990, 31, 2209–2210.
(n) Zhang, T. Y.; O’Toole, J. C.; Dunigan, J. M. Tetrahedron
Lett. 1998, 39, 1461–1464.
4.4. Supplementary material
ORTEP drawing of 11-erythro-B and crystallographic data
on 11-erythro-B, 14-erythro-B, and 15-B.
Acknowledgements
The authors are grateful to Central Glass Co. Ltd for a
generous gift of hexafluorocumyl alcohol. Partial support of
this work through Grant-in-Aid for Scientific Research
(Nos. 06740488 and 07740501) provided by the Ministry of
Education, Science, and Culture of the Japanese
Government is heartily acknowledged.
9. (a) Deschamps, B.; Lefebvre, G.; Seyden-Penne, J.
Tetrahedron 1972, 28, 4209–4222. (b) Deschamps, B.;
Lefebvre, G.; Redjal, A.; Seyden-Penne, J. Tetrahedron
1973, 29, 2437–2444. (c) Bottin-Strzalko, T. Tetrahedron
1973, 29, 4199–4204.
10. Kawashima, T.; Watanabe, K.; Okazaki, R. Tetrahedron Lett.
1997, 38, 551–554.
References
11. (a) Buss, A. D.; Warren, S.; Leake, J. S.; Whitham, G. H.
J. Chem. Soc., Perkin Trans. 1 1983, 2215–2218. (b) Buss,
A. D.; Warren, S. J. Chem. Soc., Perkin Trans. 1 1983,
2307–2325.
1. For leading references see: (a) Wadsworth, Jr. W. S. Org.
React. 1977, 25, 73–253. (b) Cadogan, J. I. G. Organo-
phosphorus Reagents in Organic Synthesis; Academic: New
York, 1979. (c) Maryanoff, B. E.; Reitz, A. B. Chem. Rev.
1989, 89, 863–927. (d) Kelly, S. E. Comprehensive Organic
12. Bottin-Strzalko, T.; Seyden-Penne, J. Bull. Soc. Chim. Fr.
1984(II), 161–163.

S. Kojima et al. / Tetrahedron 59 (2003) 255–265
265
13. Thatcher, G. R. J.; Kluger, R. Adv. Phys. Org. Chem. 1989, 25,
99–265, and references cited therein.
nated phosphorus atom see: Martin, J. C.; Balthazor, T. M.
J. Am. Chem. Soc. 1977, 99, 152–162.
14. (a) Vedejs, E.; Fuchs, P. L. J. Am. Chem. Soc. 1973, 95,
822. (b) Ward, W.; McEwen, W. E. J. Org. Chem. 1990,
55, 493.
21. Akiba, K.-y.; Nakata, H.; Yamamoto, Y.; Kojima, S. Chem.
Lett. 1992, 1559–1562.
22. Buss, A. D.; Mason, R.; Warren, S. Tetrahedron Lett. 1983,
24, 5293–5296.
15. Roberts, T. G.; Whitham, G. H. J. Chem. Soc., Perkin Trans. 1
1985, 1953–1955.
23. (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.;
Yamamoto, Y.; Akiba, K.-y. Org. Lett. 2001, 4, 1873–1875.
(c) Kojima, S.; Sugino, M.; Nakamoto, M.; Matsukawa, S.;
Akiba, K.-y. J. Am. Chem. Soc. 2002, 124, 7674–7675.
(d) Matsukawa, S.; Kojima, S.; Kajiyama, K.; Yamamoto, Y.;
Akiba, K.-y.; Re, S. Y.; Nagase, S. J. Am. Chem. Soc. 2002,
124, 13154–13170.
16. Tsukamoto, M.; Schlosser, M. Synlett 1990, 605–608.
17. Bellucci, G.; Chiappe, C.; Moro, G. L. Tetrahedron Lett. 1996,
37, 4225–4228.
18. Granoth, I.; Martin, J. C. J. Am. Chem. Soc. 1979, 101,
4623–4626.
19. Kojima, S.; Kajiyama, K.; Akiba, K.-y. Bull. Chem. Soc. Jpn
1995, 68, 1785–1797.
20. The stereochemical denotations correspond (in the written
order) to the phosphorus atom, the carbon atom a to the
phosphorus and the carbon atom b to the phosphorus,
respectively. For proposed nomenclature for the pentacoordi-
24. (a) Brandt, P.; Norrby, P.-O.; Martin, I.; Rein, T. J. Org. Chem.
1998, 63, 1280–1289. (b) Ando, K. J. Org. Chem. 1999, 64,
6815–6821.