Detection and reaction of oxaphosphetanes derived from benzaldehyde and 1-adamantylmethylidene ylide

Article abstract of DOI:10.1021/jo951025x

The Wittig reaction of (1-adamantylmethylidene)triphenylphosphorane (PH₃P=CH(1-Ad)) with benzaldehyde was investigated, and the results were compared with those of other ylides. The substituent effect in the reaction of the ylide with benzaldehydes was determined by competition experiments, which gave a Hammett _p value of 3.2. The _p value is much larger than those reported for analogous reactions of Ph₃P=CH(CH₂)₂CH₃ (p=0.20) and Ph₃P=CH(CH₃)₂ (p = 0.59), indicating that the reaction mechanism differs for Ph₃P=CH(1-Ad) and the other ylides. The cis/trans ratio of the product alkene is 74/26 for the reaction with the parent benzaldehyde and highly depends on the position of the substituent; ortho substituted benzaldehydes gave the trans alkenes up to 90%. Monitoring the reaction by means of ³¹P NMR revealed that both cis and trans oxaphosphetane intermediates were formed and that the formation and decomposition of the cis oxaphosphetane are 7-12 times faster than those of the trans oxaphosphetane. From the comparison of the reaction of Ph₃P=CH(1-Ad) + benzaldehyde with those of Ph₃P=CH(CH₂)₂CH₃ + benzaldehyde and benzophenone, and Ph₃P=CH(CH₃)₂ + benzophenone, it was concluded that all the reactions with these nonstabilized ylides proceed via an electron-transfer mechanism and that the rate-determining step changes from the electron transfer step to that of radical combination when the substrate or ylide becomes more sterically demanding.

Full text of DOI:10.1021/jo951025x

722
J . Org. Chem. 1996, 61, 722-726
Detection a n d Rea ction of Oxa p h osp h eta n es Der ived fr om
Ben za ld eh yd e a n d 1-Ad a m a n tylm eth ylid en e Ylid e
Hiroshi Yamataka,*,† Tsutomu Takatsuka, and Terukiyo Hanafusa
The Institute of Scientific and Industrial Research, Osaka University, Ibaraki,
Osaka 567, J apan and Institute for Fundamental Research of Organic Chemistry, Kyushu University,
Hakozaki, Higashi-ku, Fukuoka 812-81, J apan
Received J une 6, 1995X
The Wittig reaction of (1-adamantylmethylidene)triphenylphosphorane (Ph3PdCH(1-Ad)) with
benzaldehyde was investigated, and the results were compared with those of other ylides. The
substituent effect in the reaction of the ylide with benzaldehydes was determined by competition
experiments, which gave a Hammett F value of 3.2. The F value is much larger than those reported
for analogous reactions of Ph3PdCH(CH2)2CH3 (F ) 0.20) and Ph3PdCH(CH3)2 (F ) 0.59), indicating
that the reaction mechanism differs for Ph3PdCH(1-Ad) and the other ylides. The cis/trans ratio
of the product alkene is 74/26 for the reaction with the parent benzaldehyde and highly depends
on the position of the substituent; ortho substituted benzaldehydes gave the trans alkenes up to
90%. Monitoring the reaction by means of 31P NMR revealed that both cis and trans oxaphosphetane
intermediates were formed and that the formation and decomposition of the cis oxaphosphetane
are 7-12 times faster than those of the trans oxaphosphetane. From the comparison of the reaction
of Ph3PdCH(1-Ad) + benzaldehyde with those of Ph3PdCH(CH2)2CH3 + benzaldehyde and
benzophenone, and Ph3PdCH(CH3)2 + benzophenone, it was concluded that all the reactions with
these nonstabilized ylides proceed via an electron-transfer mechanism and that the rate-determining
step changes from the electron transfer step to that of radical combination when the substrate or
ylide becomes more sterically demanding.
The Wittig reaction of benzaldehyde with a phosphorus
intermediates is different under typical reaction condi-
tions; the lifetime is long for that from benzaldehyde and
short for that from benzophenone.6a,7,8 Furthermore, the
mechanism of the oxaphosphetane formation step is
different for the two substrates; the 14C isotope effects
and the substituent effects showed that the reaction of
benzaldehyde with nonstabilized ylides proceeds via rate-
determining ET to give a radical ion pair that is followed
by fast radical coupling (RC), whereas the oxaphosphe-
tane from benzophenone is formed via fast ET and slow
RC processes.
ylide proceeds via one of two reaction pathways depend-
ing on the nature of the ylide (eq 1).1,2 It has previously
been shown that the reaction with a semistabilized ylide
(benzylidenetriphenylphosphorane, 1) proceeded via the
direct polar addition (PL) mechanism3 and gave a mix-
ture of cis/trans alkenes,4 whereas the reaction with a
nonstabilized ylide (Ph3PdCH(CH2)2CH3, 2) yielded pref-
erentially the cis alkene5,6 via an initial electron transfer
(ET) from the ylide to benzaldehyde.2 The mechanistic
difference is due to a different ability of the ylides to
transfer an electron to benzaldehyde. These conclusions
were based on carbonyl carbon-14 kinetic isotope effects
and substituent effects on the oxaphosphetane formation
step as well as on the enone-isomerization and dehalo-
genation probe experiments.2
The reaction of nonstabilized ylides with benzaldehyde
differ in many mechanistic respects from that with
benzophenone.1,2 Although both reactions proceed through
oxaphosphetane intermediates, the kinetic stability of the
† Kyushu University.
In order to clarify the origin of the mechanistic differ-
ence between the two reactions, and also to understand
more the mechanism of the Wittig reaction of nonstabi-
lized ylides, we have carried out the reaction of benzal-
dehyde with 1-adamantylmethylidene ylide 3 (eq 2). The
aforementioned mechanistic difference may arise from
the large steric bulk of benzophenone which makes the
access of the ylide carbon center to the carbonyl carbon
difficult and thus the RC step becomes rate determining.
If this rationalization is correct, then due to its steric
bulkiness, the RC step for 3 is expected to be slowed
down, which would make the RC step rate determining
even in the reaction with benzaldehyde.
X Abstract published in Advance ACS Abstracts, J anuary 1, 1996.
(1) Review articles from a mechanistic viewpoint: Maryanoff, B. E.;
Reitz, A. B. Chem. Rev. 1989, 89, 863; Vedejs, E. In Topics in
Stereochemistry; Eliel, E. L., Wilson, S. H., Eds.; Wiley: New York,
1994; Vol. 21, pp 1-157.
(2) Yamataka, H.; Nagareda, K.; Takatsuka, T.; Ando, K.; Hanafusa,
T.; Nagase, S. J . Am. Chem. Soc. 1993, 115, 8570.
(3) Yamataka, H.; Nagareda, K.; Ando, K.; Hanafusa, T. J . Org.
Chem. 1992, 57, 2865.
(4) McEwen, W. E.; Beaver, B. D. Phosphorus Sulfur 1985, 24, 259.
Ward, W. J ., J r.; McEwen, W. E. J . Org. Chem. 1990, 55, 493. Allen,
D. W.; Ward, H. Tetrahedron Lett. 1979, 29, 2707. McEwen, W. E.;
Cooney, J . V. J . Org. Chem. 1983, 48, 983. Mylona, A.; Nikokavouras,
J .; Takakis, I. M. J . Org. Chem. 1988, 53, 3838.
(5) Vedejs, E.; Meier, G. P.; Snoble, K. A. J . Am. Chem. Soc. 1981,
103, 2823. Vedejs, E.; Marth, C. F. J . Am. Chem. Soc. 1988, 110, 3948.
Vedejs, E.; Fleck, T. J . Am. Chem. Soc. 1989, 111, 5861.
(6) (a) Maryanoff. B. E.; Reitz, A. B.; Mutter, M. S.; Inners, R. R.;
Almond, H. R., J r.; Whittle, R. R.; Olofson, R. A. J . Am. Chem. Soc.
1986, 108, 7664. (b) Reitz, A. B.; Nortey, S. O.; J ordan, A. D., J r.;
Mutter, M. S.; Maryanoff, B. E. J . Org. Chem. 1986, 51, 3302.
(7) Vedejs, E.; Snoble, K. A. J . J . Am. Chem. Soc. 1973, 95, 5778.
(8) Yamataka, H.; Nagareda, K.; Takai, Y.; Sawada, M.; Hanafusa,
T. J . Org. Chem. 1988, 53, 3877.
0022-3263/96/1961-0722$12.00/0 © 1996 American Chemical Society

Detection and Reaction of Oxaphosphetane
J . Org. Chem., Vol. 61, No. 2, 1996 723
Resu lts a n d Discu ssion
Benzaldehyde reacts with butylidenetriphenylphos-
phorane (2) or isopropylidenetriphenylphosphorane (4)
quickly at -78 °C to form the corresponding oxaphos-
phetane, which then decomposes slowly at 0 °C to give
alkene and triphenylphosphine oxide.1 By contrast, it
was found that the reaction of 3 with benzaldehyde is
relatively slow at -50 °C and that the rates of the
formation and the decomposition of the oxaphosphetane
intermediate are similar. The 31P NMR monitoring of
the reaction under the Li-salt free conditions gave a
stacked plot illustrated in Figure 1. In addition to the
signals of 3 and Ph3PdO, those of cis (-57.3 Hz relative
to Ph3P) and trans (-61.7 Hz) oxaphosphetanes were
observed. The amount of cis oxaphosphetane increases
quickly at the early stage of the reaction and then
decreases, while the amount of trans oxaphosphetane
increases gradually. By assuming a simple reaction
scheme (eq 3, ylide ) 3), it was possible to compute rate
constants of the six processes by matching the observed
and the calculated time dependence of the four species.
Maryanoff and the co-workers have carried out a similar
treatment for the reaction of benzaldehyde with 2 under
both Li-salt free and Li-salt present conditions.6a They
found that although little or no “drift” of the stereochem-
istry of the oxaphosphetane was detected for the Li-salt
free reaction, the reaction in the presence of LiBr brought
about considerable “stereochemical drift”. By following
the concentration changes of a pair of oxaphosphetanes
and the product, the four rate constants were obtained;
in this reaction, the formation of the oxaphosphetane is
too fast to follow and only the rate constant ratio, k1/k3,
could be estimated.6a The reaction of benzophenone with
4 has also been followed previously in a similar manner
but there is no cis/trans discrimination in this reaction.8
The rate constants for the three sets of reactions are
summarized in Table 1.9
F igu r e 1. 31P NMR stacked plot for the reaction of benzal-
dehyde with 3 at -50 °C. The aldehyde and ylide concentra-
tions are 0.15 mol L-1 each.
F igu r e 2. Variations of amounts of 3, cis and trans oxaphos-
phetanes, and Ph3PdO with time in the reaction of benzalde-
hyde with 3 at -50 °C in THF.
both oxaphosphetane reversal and decomposition appear
to occur in similar rates for 3 and 2 if the effect of
temperature is taken into consideration, suggesting that
the steric requirements at the reversal and decomposition
transition states are similar to that at the intermediate.
(2) The cis oxaphosphetane forms 11.8 times faster than
the trans oxaphosphetane, in accord with a general trend
of high cis selectivity in the reaction of nonstabilized
ylides with benzaldehyde. (3) The cis oxaphosphetane
decomposes to give products 11.7 times faster and reverts
to the reactants 6.8 times faster than the trans counter-
part; thus the cis oxaphosphetane is kinetically less
stable than the trans oxaphosphetane. The faster forma-
tion and the faster backward reaction of the cis oxaphos-
phetane compared with the trans oxaphosphetane are the
origin of the “stereochemical drift” observed in the
reaction of nonstabilized ylides with benzaldehyde.6a (4)
Equilibrium constants between the reactants (benzalde-
hyde + 3) and the oxaphosphetanes are 118 L mol-1 and
68 L mol-1 for the cis and the trans oxaphosphetane.
These numbers are in between those for benzaldehyde
+ 2 (too large to determine) and benzophenone + 4 (3.3).
(5) Fractionation at the oxaphosphetane (k5/k2) is 2.2 for
cis and 1.3 (k6/k4) for trans. The values are similar for
the reaction of benzophenone + 4 (1.8).
Several important points are apparent in Table 1. (1)
The oxaphosphetane formation rate constants are of the
order of 10-3-10-4 L mol-1 s-1 for 3. These rates are
much slower than those for 2; the oxaphosphetane
formation is too fast to follow even at -78 °C for 2. The
results indicate that the C-C bond formation step suffers
a large steric effect for 3 as expected. In contrast to this,
(9) Reliability of these rate constants were estimated by carrying
out
a series of simulation procedures with slightly different rate
constants. The use of k1 that is 10% larger than the value listed in
Table 1 gave reasonable matching between the calculated and experi-
mental results, but the use of the rate constant 20% larger than that
in Table 1 resulted in unacceptable matching. Thus, the error limit of
k1 in the present derivation is 20%. Similarly, the error limit of k3 is
20% and that of k2 and k5 is 40%. Since k4 and k6 are very small, values
close to zero did not change the calculated curves very much, but they
cannot be 5 times larger than those listed in Table 1.

724 J . Org. Chem., Vol. 61, No. 2, 1996
Yamataka et al.
Ta ble 1. Estim a ted Ra te Con sta n ts of Som e Elem en ta r y Step s of th e Wittig Rea ction s of Ben za ld eh yd e w ith 2 a n d 3,
a n d of Ben zop h en on e w ith 4 in THF
3 + PhCHOa
-50 °C
2 + PhCHOb
-30 °C
4 + Ph2CdOc
step
rate constant
0 °C
oxaphosphetane
formation
cis
k1 /L mol-1s-1
3.3 × 10-3
2.8 × 10-4
11.8
d
d
3.50
1.3 × 10-3
4.0 × 10-4
7.0 × 10-4
trans k3 /L mol-1s-1
k1/k3
oxaphosphetane
reversal
cis
k2 /s-1
2.8 × 10-5
4.1 × 10-6
6.8
1.39 × 10-4
0.9 × 10-5
15.4
trans k4 /s-1
k2/k4
oxaphosphetane
decomposition
cis
k5 /s-1
6.2 × 10-5
5.3 × 10-6
11.7
4.78 × 10-5
7.90 × 10-5
0.61
trans k6 /s-1
k5/k6
a
b
d
Under Li-salt free conditions. Under Li-salt present conditions. Data taken from reference 3a. c Data taken from reference 8. Too
fast to follow.
Ta ble 2. Cis/Tr a n s P r od u ct Ra tio in th e Wittig
Rea ction of Su bstitu ted Ben za ld eh yd es w ith 3 a n d 1
in THF a t 0 °C
%cis
substituent
3a
1b
p-OMe
m-OMe
o-OMe
p-Me
m-Me
o-Me
H
p-Cl
m-Cl
o-Cl
c
c
c
38.7
46.3
83.7
38.6
38.0
37.4
42.9
43.6
43.9
86.3
c
69.3
67.2
29.0
74.0
70.0
71.7
11.3
67.9
17.7
70.3
71.0
p-F
o-F
m-CF3
p-CF3
c
44.6
c
a
Average of two runs. Reproducibility is better than 0.5%.
Data taken from reference 3. c Not determined.
b
The cis/trans ratios of product alkenes were deter-
1
mined by GLC and H NMR for the reactions of substi-
tuted benzaldehydes and 3, and are summarized in Table
2. The cis/trans ratios for reactions with benzylidene-
triphenylphosphorane 1 were also listed for comparison
purpose. It should be pointed out at first that a consider-
able amount of trans alkene was detected in the reactions
of a nonstabilized ylide, 3, under the Li-salt free condi-
tions. This is in sharp contrast to those for the reaction
of benzaldehyde and 2; here predominantly (>96%) cis
alkene was formed under the Li-salt free conditions, and
“stereochemical drift” occurred only when Li salt was
present.6a Clearly the observed stereochemical drift is
due to isomerization at the oxaphosphetane stage or later
since the initial rate ratio (k1/k3 ) 11.8) requires cis
oxaphosphetane about 92% for X ) H. The oxaphosphe-
tane derived from 3 appears kinetically much less stable
compared to the oxaphosphetane derived from 2.
Substituents at the meta- and para-positions of ben-
zaldehyde have little effect on the %cis value. Compari-
son of the present results with those reported previously
for reactions of substituted benzaldehydes with 1 re-
vealed that the two sets of reactions are similar in the
meta- and para-substituent effect in that substituent
electronic property has little influence on the %cis value.
It is also apparent, however, that the effect of ortho
substituents is quite different for the two series. First
of all, o-Cl and o-Me derivatives gave higher %cis value
for 1 in contrast to lower %cis for 3. Second, o-Me showed
little influence on the %cis value for 1; this is also in
sharp contrast to the o-Me effect in the reaction of 3. The
reason for the different ortho substituent effects for the
F igu r e 3. Variations of reactivity with the σ values for the
reactions of substituted benzaldehydes with 3 at 0 °C in THF.
two systems is not apparent, but the present results
clearly indicate that the nature of the transition states
is different for the two reaction systems.
Substituent effects on the overall reactivity were
measured by competition experiments as reported previ-
ously.8 The relative reactivities obtained are illustrated
in Figure 3. Here the log(kX/kH) values for the ortho
substituted derivatives were plotted against the corre-
sponding σp constants and are indicated by closed circles.
The points of o-Cl and o-F derivatives deviate upward
from the correlation line. Such ortho halogen accelera-
tion has previously been observed in the reaction of
benzaldehyde with organometallic reagents and may be
ascribed to a chelating interaction between the halogen
atom and a reagent in the rate-determining transition
state.3,10 The large Hammett F value (3.23) obtained for
the reaction of benzaldehydes and 3 clearly differs from
the small F values observed previously for the reaction
of benzaldehydes with 2 (0.20) and 4 (0.59) under the
same reaction conditions, indicating that the nature of
the rate-determining transition state is different from
that for 2 or 4.2 The large F value for 3 is consistent with
the polar mechanism as well as the ET-RC reaction
pathway with the RC step rate determining. In the latter
mechanism, the overall rate consists of the ET pre-
(10) Yamataka, H.; Nishikawa, K.; Hanafusa, T. Chem. Lett. 1990,
1711.

Detection and Reaction of Oxaphosphetane
J . Org. Chem., Vol. 61, No. 2, 1996 725
Ta ble 3. E-Z Isom er iza tion d u r in g th e Rea ction of
Z-En on e w ith Va r iou s Ylid es in THF a t 0 °Ca
substituent effects on the reactivity, the carbonyl-carbon
kinetic isotope effect, and the dehalogenation probe
experiment, strongly suggested that the ET process is
indeed involved in the main reaction pathway of the
reactions of benzaldehyde with nonstabilized ylides.2
The combination of the substituent effect on reactivity
and the probe experiment revealed that the reaction of
3 with benzaldehyde proceeds via the ET-RC reaction
route to oxaphosphetane with the RC step rate determin-
ing. Thus, there occurs changeover of the rate-determin-
ing step from the ET step for ylides 2 and 4 to the RC
step for 3, and this mechanistic difference can be ratio-
nalized in terms of different steric requirement at the
RC transition state. For less sterically demanding ylides
(2 and 4) the RC step is lower in energy than the ET
step, whereas for sterically demanding ylides 3, the RC
transition state becomes less stable. The occurrence of
the mechanistic changeover is thus fully consistent with
the Wittig reaction mechanism in which the radical ion
pair species is a meaningful intermediate. Finally, the
observed different effects of ortho substituents on the
product cis/trans ratio in the reactions of benzaldehyde
with 3 and 1 can be considered to arise from different
reaction pathways of the two reactions, RC-ET for 3 and
PL for 1.
ylide
time, min
recovered enone, Z:E
Ph3PdCHCH3
15
30
15
30
15
30
15
60
15
60
60
81.1:18.9
70.4:29.6
76.6:23.4
62.7:37.3
86.2:13.8
68.8:31.2
86.1:13.9
33.2:66.8
98.2:1.8
Ph3PdCHPr (2)
Ph3PdCHAd (3)
Ph3PdCMe2 (4)b
Ph3PdCHPh (1)b
96.7:3.3
98.5:1.5
blank
a
Wittig products were not detected. E:Z ratio was normalized;
b
100% ) E + Z. Data taken from reference 2.
equilibrium and the radical coupling rate, and the F value
on the overall rate constant should be large positive2 since
the F value on the ET equilibrium is large positive.11
One way to distinguish between the two mechanistic
alternatives is to examine a possible existence of the
radical ion pair intermediate. The enone-isomerization
probe is one of such criteria, and its usefulness has been
proved in the literature.2,12 Here, isomerization of the
starting Z-enone (Z-5) to the E-enone (E-5) upon mixing
with a reagent is taken as an indication of the occurrence
of ET from the reagent to the enone (eq 4). Since the
reduction potential of the Z-enone is more negative
(-2.28 V vs SCE) than those of benzaldehyde (-1.84 V)
and benzophenone (-1.82 V),13 a positive response of the
enone-isomerization probe then indicates the possible
occurrence of ET to the aromatic aldehyde and ketone.
Exp er im en ta l Section
Ma ter ia ls. THF was dried over sodium/benzophenone and
distilled immediately before use. All substituted benzalde-
hydes were commercially available and purified by either
distillation or recrystallization. 1-(Adamantylmethyl)triph-
enylphosphonium bromide was obtained by heating a mixture
of triphenylphosphine (5.7 g, 22 mmol) and 1-adamantylmethyl
bromide (5.0 g, 22 mmol) at 200 °C for 23 h. The reaction
mixture was dissolved in CHCl3, and colorless precipitates
were collected by adding benzene. Recrystallization from
EtOH/Et2O gave 1-(adamantylmethyl)triphenylphosphonium
bromide as a complex with a molecule of EtOH: 9.94 g (84%),
mp. 267-71 °C. 1H NMR (360 MHz, CDCl3) δ 8.18-7.65 (m,
15H), 3.82 (d, J ) 12.1 Hz, 2H), 3.71 (q, J ) 7.03 Hz, 2H),
1.81-1.51 (m, 15H), 1.23 (t, J ) 7.03 Hz, 3H). Anal. Calcd
for C31H38BrOP: C, 69.27; H, 7.07; Br, 14.96; P, 5.76. Found:
C, 69.46, H, 7.18, Br, 14.88, P, 5.73. The phosphonium
bromide was then dissolved in CHCl3 again, and benzene was
added to obtain a colorless precipitate. This salt was suggested
1
Table 3 lists the results of enone-isomerization probe
experiment of 3, together with those of other ylides. As
reported previously, there is a clear-cut distinction
between nonstabilized and semistabilized ylides.2 It is
apparent that ylide 3 exhibited a positive response in the
probe experiment as in the other typical nonstabilized
ylides, whereas the semistabilized ylide showed a nega-
tive response. Therefore it is reasonable to assume that
the reaction of benzaldehyde with 3 proceeds through the
same reaction pathway with other nonstabilized ylides,
i.e. the ET-RC route, to give the oxaphosphetane inter-
mediate.
It should be noted that although the enone-isomeriza-
tion probe indicates the possible involvement of an ET
process in the reaction, it does not necessarily mean that
the ET process is on the main reaction coordinate.
However, the combined use of several criteria, which
include, in addition to the enone-isomerization probe,
by H NMR to contain a molecule of benzene instead of EtOH
and was used in preparation of ylide 3. (Z)-2,2,6,6-Tetram-
ethylhept-4-en-3-one was synthesized as described in the
literature.14
Rea ction s. All reactions were carried out under dry
nitrogen using the Schlenk tube technique.15 Ylide solution
was prepared by adding an equimolar amount of sodium
hexamethyldisilazide (NaHMDS, 1.0 M Aldrich) to a suspen-
sion of (1-adamantylmethyl)triphenylphosphonium bromide in
THF at 0 °C. The Wittig reactions were followed by GLC or
31P NMR.
The 31P NMR monitoring of the reaction of benzaldehyde
with 1 was carried out as reported before.8 In a flame-dried
NMR test tube (10 mm L) were placed a capillary that contains
(MeO)3PdO in CDCl3 (215 mg/L, internal reference) and a 2.25
mL THF solution of 3 (0.2 M) that was prepared from
equimolar amount of (1-adamantylmethyl)triphenylphos-
phonium bromide and NaHMDS. The solution was then
solidified at liquid N2 temperature. To this was added a THF
solution of benzaldehyde (0.75 mL, 0.6 M), and the tube was
sealed. The Wittig reaction was followed at -50 °C probe
(11) Zumar, P.; Exner, O.; Rekker, R. F.; Nauta, W. Th. Coll. Czech.
Chem. Commun. 1968, 33, 3213.
(12) House, H. O.; Weeko, P. D. J . Am. Chem. Soc. 1975, 97, 2770.
Ashby, E. C.; Wiesemann, T. L. J . Am. Chem. Soc. 1978, 100, 3101.
(13) Yamataka, H.; Nagareda, K.; Hanafusa, T. Nagase, S. Tetra-
hedron Lett. 1989, 30, 7287.
(14) House, H. O.; Crumrine, D. S.; Teranishi, A. Y.; Olmstead H.
D. J . Am. Chem. Soc. 1973, 95, 3310. House, H. O.; Weeks, P. D. J .
Am. Chem. Soc. 1975, 97, 2770.
(15) Schriver, D. F. Manipulation of Air Sensitive Compounds;
McGraw-Hill: New York, 1969; Chapter 7.

726 J . Org. Chem., Vol. 61, No. 2, 1996
Yamataka et al.
temperature. The spectra were recorded on a Bruker 360M
operating at 146 MHz for phosphorus. The conditions of 31P
measurements were chosen to ensure that the relaxation times
of the species (determined by the T1 null method) were
sufficiently short for accurate integrals. The estimated re-
producibility of the intensity was less than 5%, and the results
from the kinetic treatment should be taken semiquantita-
tively.16
cis), 2.23 (s, 3H, trans), 1.95-1.47 (m, 15H). o-Cl, 7.52-7.07
(m, 4H), 6.62 (d, J trans ) 16.1, 1H), 6.20 (d, J cis ) 12.6, 1H),
6.07 (d, J trans ) 16.1 Hz, 1H), 5.44 (d, J cis ) 12.6, 1H), 2.03-
1.53 (m, 15H). m-Cl, 7.29-7.08 (m, 4H), 6.30 (d, J cis ) 13.0,
1H), 6.18 (d, J trans ) 16.1, 1H), 6.06 (d, J trans ) 16.1 Hz, 1H),
5.37 (d, J cis ) 12.6, 1H), 2.03-1.51 (m, 15H). p-Cl, 7.29-7.08
(m, 4H), 6.30 (d, J cis ) 13.0, 1H), 6.18 (d, J trans ) 16.1, 1H),
6.06 (d, J trans ) 16.1 Hz, 1H), 5.37 (d, J cis ) 13.0, 1H), 2.03-
1.51 (m, 15H). o-F, 7.48-6.96 (m, 4H), 6.41 (d, J trans ) 16.5,
1H), 6.16 (d, J cis ) 12.6, 1H), 6.16 (d, J trans ) 16.5 Hz, 1H),
5.50 (d, J cis ) 12.6, 1H), 2.04-1.54 (m, 15H). p-F, 7.36-6.93
(m, 4H), 6.32 (d, J cis ) 12.8, 1H), 6.20 (d, J trans ) 16.3, 1H),
6.00 (d, J trans ) 16.3 Hz, 1H), 5.36 (d, J cis ) 12.8, 1H), 2.02-
1.54 (m, 15H). m-CF3, 7.60-7.25 (m, 4H), 6.36 (d, J cis ) 12.6,
1H), 6.27 (d, J trans ) 16.3, 1H), 6.16 (d, J trans ) 16.3 Hz, 1H),
5.44 (d, J cis ) 12.6, 1H), 2.05-1.52 (m, 15H). p-CF3, 7.58-
7.25 (m, 4H), 6.27 (d, J trans ) 16.3, 1H), 6.19 (d, J trans ) 16.3
Hz, 1H), 2.05-1.64 (m, 15H). The cis/trans product ratio was
determined as reported before.3 In a standard procedure, 1.0
mL of THF solution (0.1 M) of ylide was added at 0 °C to 1.0
mL of THF solution (0.1 M) of benzaldehyde, and the solution
was allowed to react for 30 min. The reaction mixture was
worked up in a usual manner and subject to GLC analysis
(2-m glass column packed with 3% PEG-HT) to determine the
product ratio. The fraction of reaction was more than 70%
except for less reactive substrates such as Me-substituted
benzaldehydes. We could not find serious dependence of the
product ratio on the fraction of reaction, and thus the qualita-
tive discussion given in text is viable. Enone-isomerization
experiments were carried out as reported previously.2
The relative reactivities of substituted benzaldehydes with
1 were determined at 0.0 ( 0.1 °C in THF by competition
experiments as described previously.8 Each Wittig reaction
gave the expected alkenes, and no other products were
detected. All product alkenes were characterized by 1H NMR.
Except for the parent compound, the products were obtained
as mixtures of cis and trans alkenes. The chemical shifts of
substituted cis- and trans-1-phenyl-2-(1-adamantyl)ethene
derivatives are as follows: cis; H, δ 7.29-7.16 (m, 5H), 6.39
(d, J ) 12.8, 1H), 5.36 (d, J ) 12.8 Hz, 1H), 1.86-1.25 (m,
15H). trans; H, δ ) 7.37-7.14 (m, 5H), 6.24 (d, J ) 16.1, 1H),
6.10 (d, J ) 16.1 Hz, 1H), 2.03-1.68 (m, 15H). o-Me, 7.48-
7.09 (m, 4H), 6.42 (d, J trans ) 16.1, 1H), 6.23 (d, J cis ) 12.6,
1H), 5.96 (d, J trans ) 16.1 Hz, 1H), 5.37 (d, J cis ) 12.6, 1H),
2.34 (s, 3H, cis), 2.33 (s, 3H, trans), 2.04-1.54 (m, 15H). m-Me,
7.24-6.97 (m, 4H), 6.35 (d, J cis ) 12.8, 1H), 6.21 (d, J trans
)
16.1, 1H), 6.08 (d, J trans ) 16.1 Hz, 1H), 5.33 (d, J cis ) 12.8,
1H), 2.33 (s, 3H), 2.02-1.57 (m, 15H). p-Me, 7.18-6.98 (m,
4H), 6.28 (d, J cis ) 12.6, 1H), 6.13 (d, J trans ) 16.1, 1H), 5.97
(d, J trans ) 16.1 Hz, 1H), 5.25 (d, J cis ) 12.6, 1H), 2.25 (s, 3H,
(16) For kinetic experiments by use of 31P NMR, see: Sawada, M.;
Takai, Y.: Chong, C.; Hanafusa, T.; Misumi, S. Tetrahedron Lett. 1985,
26, 5065.
J O951025X