Solvent effects on the steric course of the [2,3]-wittig rearrangement of (S,E)-[3-(Allyloxy)prop-1-ene-1,3-diyl]dibenzene and derivatives

Article abstract of DOI:10.1002/ejoc.201001061

The effect of solvents and additives on the steric course of [2,3]-Wittig rearrangement of the chiral 1,3-diphenyl-1-propenyloxy-2-propen-1-yl carbanion and its derivatives wasexamined on the basis of chirality transfer in intramolecular trapping. Copyrig

Full text of DOI:10.1002/ejoc.201001061

FULL PAPER
DOI: 10.1002/ejoc.201001061
Solvent Effects on the Steric Course of the [2,3]-Wittig Rearrangement of (S,E)-
[3-(Allyloxy)prop-1-ene-1,3-diyl]dibenzene and Derivatives
Hidaka Ikemoto,[a] Michiko Sasaki,[a] and Kei Takeda*[a]
Keywords: Carbanions / Chirality / Configuration determination / Rearrangement / Wittig reactions
The effect of solvents and additives on the steric course of
[2,3]-Wittig rearrangement of the chiral 1,3-diphenyl-1-prop-
enyloxy-2-propen-1-yl carbanion and its derivatives was
examined on the basis of chirality transfer in intramolecular
trapping.
enantioselectively using sBuLi/(–)-spartein, the extent of the
enantiomeric induction was highly dependent on the sol-
vent, affording the product with an enantiomeric ratio rang-
ing from 52:48 (THF) to 96:2 (toluene).[5] In the former
case, the comparison was only made between a hydro-
carbon solvent and Et2O, whereas in the latter case the ob-
served trend can be interpreted in terms of competitive
binding between the chiral ligand and the solvent toward
the organolithium species. Publications by the groups of
Hoppe[6] and Gawley[7] have also referred to solvent effects
on the configurational stability of benzyl and allyl carban-
ions, and Kapeller and Hammerschmidt have recently dis-
cussed the configurational stability of aroyloxy[D1]meth-
yllithium in some solvent systems over a range of tempera-
tures.[8]
One of the most elegant and widely used methods for
evaluating configurational stability of a chiral carbanion is
the Hoffmann test,[9] which is based on kinetic resolution
during an electrophilic substitution reaction when using ra-
cemic and enantioenriched aldehydes as electrophiles. The
test is usually conducted in THF, occasionally in Et2O, but,
to the best of our knowledge, solvent effects in the
Hoffmann test have not been explicitly discussed. The
reason for this, in addition to the qualitative nature of the
test, is probably that the rate of the reaction with an electro-
phile is not fast enough to reflect the differences between
solvents.
In the above [2,3]-Wittig rearrangement, the extent of ra-
cemization should be controlled by the relative rates of the
racemization of a chiral carbanion and the [2,3]-Wittig re-
arrangement; these processes are associated with the config-
urational stability and chemical reactivity of the carban-
ions, respectively, which may be interrelated. Consequently,
solvents and coordinating additives such as tetramethyleth-
ylenediamine (TMEDA) can influence the outcome of both
processes in either the same or opposite manners. Al-
though, in general, polar solvents or the additives promote
the formation of separated ion pairs (SIPs), which facilitates
Introduction
Recently, we have reported on the effect of conjugative
electron-withdrawing groups and α-anion-stabilizing
heteroatom substituents that distribute charge through a
double bond, on the configurational stability of chiral carb-
anions.[1] The effects can be estimated on the basis of the
extent of chirality transfer induced by intramolecular trap-
ping during [2,3]-Wittig rearrangement[2] of chiral 3-substi-
tuted 1-phenyl-1-propenyloxy-2-propen-1-yl carbanions of
type 1 (Scheme 1).[3] In that paper, we showed that the con-
figurational stability obtained in the system is highly sensi-
tive to a change in solvent. For example, when the reaction
was performed in tetrahydrofuran (THF), which is one of
the most common solvents used for reactions involving
carbanions, racemization occurred with almost all substitu-
ents; in contrast, 1,4-dioxane and diethyl ether (Et2O) gave
much better enantiomeric ratios (er).
Scheme 1. [2,3]-Wittig rearrangement of 1.
There are some literature precedents regarding solvent
effects on the configurational stability of carbanions. Curtin
and Koehl reported that racemization of enantioenriched
sec-butyllithium in hydrocarbon solvents is greatly ac-
celerated by addition of a small amount of Et2O.[4] Beak
and co-workers also reported that when pyrrolidine deriva-
tives were formed by cyclization of an anion generated
[a] Graduate School of Medical Sciences, Hiroshima University,
1-2-3 Kasumi, Minami-Ku, Hiroshima 734-8553, Japan
Fax: +81-82-257-5184
E-mail: takedak@hiroshima-u.ac.jp
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201001061.
Eur. J. Org. Chem. 2010, 6643–6650
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6643

K. Takeda et al.
FULL PAPER
racemization by decreasing the covalent character of the other than 1,4-dioxane (Table 1, Entry 10) resulted in either
carbon–metal interaction, to the best of our knowledge, no poor enantiomeric ratios or complete racemization, whereas
systematic studies on the effects of solvents on the configu- less polar hydrocarbon solvents afforded relatively high en-
rational stability of chiral carbanions have been under- antiomeric ratios (Table 1, Entries 12 and 13).
taken.
We became interested in examining the influence of sol-
When lithium diisopropylamide (LDA) was used as a
base in a range of solvents, a similar trend was observed,
but slightly higher enantiomeric ratios were achieved in
most cases (Table 1, Entries 1–3, 10, and 13). Changing the
counter cation from lithium to potassium, however, resulted
in complete racemization, except when the reactions were
conducted in less polar hydrocarbon solvents in which low
enantiomeric ratios were observed. These results suggest
that racemization is accelerated more by an increased ionic
character of the counter cation[10] than the acceleration of
the [2,3]-Wittig rearrangement due to the enhanced reacti-
vity of the carbanion.
vents and additives on the preservation of the optical purity
at the stereogenic centers in [2,3]-Wittig rearrangement of
3, which should reflect the effects on the entire processes
involving deprotonation, racemization, and [2,3]-Wittig re-
arrangement. We decided to use styryl derivatives 3 (Fig-
ure 1) to study these processes because the enantiomeric ra-
tios of their [2,3]-Wittig rearrangement products in a
number of solvents were in the appropriate range for com-
parison.
Because the reactions afforded high enantiomeric ratios
even in nonpolar solvents such as hexane, we cannot ex-
clude the possibility that the rearrangement product 5 stems
from a [1,2]-Wittig rearrangement[11] proceeding through a
radical-pair mechanism with inversion of the lithium-bear-
ing terminus in most cases. This led us to examine the reac-
tion using 1,1-dideuterioallyl derivative [D2]-4[12] in selected
solvents (Table 2). Under these conditions, only a trace
amount of [1,2]-Wittig rearrangement product [1,1-D2]-5
was detected in solvents other than toluene and hexane, in
which ca. 15% of [1,1-D2]-5 to [2,3]-Wittig rearrangement
product [3,3-D2]-5 was formed, suggesting that [1,2]-Wittig
rearrangement is not the major reaction pathway even in
nonpolar solvents.
Figure 1. Compound 3.
Results
The results obtained when (S)-4 was treated with nBuLi
at 30 °C are shown in Table 1. Reactions in acyclic ethers,
except those containing more than one oxygen atom, af-
forded moderate to good enantiomeric ratios and showed a
general increase of this ratio with increasing steric bulk of
the alkyl groups (Table 1, Entries 1–3). With solvents con-
taining more than one oxygen atom, almost complete race-
We then examined the influence of temperature on the
mization and recovery of the starting material were ob- reaction, because we had obtained preliminary results that
served, probably because of their higher donor ability indicated that, in contrast to normal trends, performing the
(Table 1, Entries 4 and 5).[1g] The reactions of cyclic ethers reactions at higher temperatures gives higher enantiomeric
Table 1. [2,3]-Wittig rearrangement of (S)-4 in various solvents.
Entry
Solvent[a,b]
nBuLi
LDA
KHMDS[c]
(R)-5
Yield [%]
(S)-4[d]
Yield [%]
(R)-5
Yield [%]
(S)-4[d]
Yield [%]
(R)-5
(S)-4[d]
Yield [%]
er
er
Yield [%]
er
1
2
3
4
5
6
7
8
9
10
11
12
13
Et2O
CPME
MTBE
DME
diglyme
THF
76
76
69
18
17
95
89
0
96
70
91
38
41
67:33
73:27
80:20
51:49
50:50
50:50
55:45
–
50:50
86:14
74:26
86:14
91:9
–
–
–
74
79
–
–
38
–
–
–
71
67
71:29
82:18
–
8
61
59
50:50
50:50
22
12
74
92:8
–
63
50:50
14
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
76
50:50
–
28
50:50
58
[e]
[e]
[e]
[e]
[e]
[e]
THP
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
[e]
1,3-dioxolane
1,3-dioxane
1,4-dioxane
NMM
70
83
52
41
87:13
74:26
82:18
92:8
–
–
20
19
41
47
66
67
50:50
50:50
52:58
54:46
35
35
0
toluene
hexane
52
44
0
[a] Cyclopentyl methyl ether (CPME), tert-butyl methyl ether (MTBE), dimethoxyethane (DME), tetrahydropyran (THP), N-methyl-
morpholine (NMM). [b] In the reactions of nBuLi and LDA, 7.5% of n-hexane originating from nBuLi solution was also present. [c]
Potassium hexamethyldisilazane. [d] Compound 4 was recovered without any loss of optical purity. [e] Not conducted.
6644
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6643–6650

Solvent Effects on a [2,3]-Wittig Rearrangement
Table 2. Wittig rearrangements of [D2]-4.
We then examined the influence of coordinating additives
capable of solvating the lithium ion, such as TMEDA,
Me2NEt, hexamethylphosphoramide (HMPA), and
THF.[14] Addition of TMEDA resulted in lower enantio-
meric ratios in all solvents (Table 4, Entries 1, 5, 9, 13, 16,
20, 24, and 28), but was particularly noticeable in hydro-
carbon solvents. In contrast, the enantiomeric ratios ob-
tained with Me2NEt, a non-chelating analogue of TMEDA,
were either comparable or better than those observed in the
absence of an additive in solvents other than the hydro-
carbon solvents (Table 4, Entries 2, 6, 10, 14, 17, 21, 25,
and 29). The lower enantiomeric ratios observed in the
presence of TMEDA relative to those with Me2NEt in
ethereal solvents might indicate the importance of the bi-
dentate interaction of the former additive with lithium ions
in the solvent. In contrast, all the reactions conducted in
THF and those with HMPA resulted in complete racemiza-
tion (Table 4, Entries 3, 7, 11, 13–15, 18, 22, 26, and 30).
The reduction in the enantiomeric ratio observed when
THF was used as an additive seemed to be most pro-
nounced in acyclic ethereal solvents (Table 4, Entries 4, 8,
Solvent[a]
Yield [%]
[1,1-D2]-5
[3,3-D2]-5
[D2]-4
Et2O
CPME
MTBE
THF
1,4-Dioxane
NMM
80
83
78
81
82
82
29
33
2
2
2
2
3
3
4
5
–
–
–
–
–
–
53
36
Toluene
Hexane
[a] 7.5% of n-hexane originating from nBuLi solution was also
present.
ratios in Et2O. The results in some selected solvents are
shown in Table 3.[13] In almost all cases, other than the reac-
tions in THF in which complete racemization was observed,
the same trend as observed in the previous case was found.
These results suggest that acceleration of the rate with in-
creasing temperature is greater for the [2,3]-Wittig re-
arrangement than for racemization, with the latter rate ex-
ceeding that of the former at lower temperatures. Thus, the
rates of racemization and [2,3]-Wittig rearrangement seem
to be of the same order of magnitude. This allowed the
above system, by using [2,3]-Wittig rearrangement of 1, to
be used as a tool for evaluating the effect of substituents on
the configurational stability of chiral carbanions.[3]
Table 4. Effect of additives on the chirality transfer in the [2,3]-
Wittig rearrangement of (S)-4.
Entry
Solvent
Additive
(R)-5
Yield [%]
er
1
2
3
4
5
6
7
8
Et2O
Et2O
Et2O
TMEDA
Me2NEt
HMPA
THF
TMEDA
Me2NEt
HMPA
THF
TMEDA
Me2NEt
HMPA
THF
TMEDA
Me2NEt
HMPA
TMEDA
Me2NEt
HMPA
THF
TMEDA
Me2NEt
HMPA
THF
TMEDA
Me2NEt
HMPA
THF
TMEDA
Me2NEt
HMPA
THF
74
77
79
87
75
74
90
85
80
71
83
94
95
81
85
82
87
50
89
80
85
85
87
79
47
79
90
74
44
77
92
63:37
67:33
50:50
57:43
56:44
77:23
50:50
58:42
66:34
81:19
50:50
58:42
50:50
50:50
50:50
78:22
84:16
50:50
71:29
66:34
74:26
50:50
63:37
65:35
63:37
50:50
62:38
77:23
63:37
50:50
73:27
Et2O
CPME
CPME
CPME
CPME
MTBE
MTBE
MTBE
MTBE
THF
THF
THF
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
NMM
Table 3. [2,3]-Wittig rearrangement of (S)-4 at lower temperatures.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Solvent[a]
T [°C]
(R)-5
Yield [%]
(S)-4
Yield [%]
er
er
Et2O
Et2O
Et2O
MTBE
MTBE
MTBE
THF
–60
–25
30
–60
–25
30
–60
–25
30
–60
–25
30
46
30
30
74
44
32
87
88
88
90
86
40
10
17
12
51:49
59:41
68:32
77:23
87:13
90:10
50:50
50:50
50:50
54:46
59:41
71:29
69:31
79:21
79:21
0
0
0
15
0
0
0
0
0
0
0
0
64
28
0
–
–
–
100:0
–
–
–
–
–
–
–
–
NMM
NMM
NMM
THF
THF
NMM
NMM
NMM
Hexane
Hexane
Hexane
toluene
toluene
toluene
toluene
hexane
hexane
hexane
hexane
–60
–25
30
99:1
98:2
–
[a] 7.5% of pentane originating from tBuLi solution was also pres-
ent.
Eur. J. Org. Chem. 2010, 6643–6650
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6645

K. Takeda et al.
FULL PAPER
and 12), whereas those in cyclic ethers and hydrocarbon slightly higher enantiomeric ratios (Table 5, Entries 2–4)
solvents (Table 4, Entries 19, 23, 27, and 31) were more than those observed in the case of the parent compound
modest.
(S)-4, which afforded lower enantiomeric ratios (Table 1,
To obtain information on the influence of the solvents Entries 1–3 and 10–13). In contrast, the enantiomeric ratios
on the inductive and resonance stabilization of carbanions, obtained with the cyano derivatives (S)-6e–g depended on
we focused on substrates (S)-6a–f, which bear either an in- the solvent (Table 5, Entries 5–7). Thus, a lowering of the
ductively anion-stabilizing halogen atom or a cyano group enantiomeric ratio, which was ascribed to the resonance ef-
at the meta- or para-positions, and (S)-6g, a vinylogous p- fect of the para-cyano group relative to its meta-substituted
cyanophenyl derivative. The results are shown in counterpart, was observed in solvents other than hexane, in
Table 5.[15,16] The halogen derivatives (S)-6b–d, other than which the enantiomeric ratio was slightly increased (Table 5,
the meta-fluoro derivative (S)-6a, resulted in similar or Entries 5 and 6). The lowering was remarkable in 1,4-di-
Table 5. [2,3]-Wittig rearrangement of (S)-6a–g.
Entry
X
Y
Et2O
Yield
[%]
CPME
MTBE
1,4-Dioxane
NMM
Toluene
Hexane
er
Yield
er
Yield
er
Yield
[%]
er
Yield
er
Yield
er
Yield
er
[%]
[%]
[%]
[%]
[%]
1
2
3
4
5
6
7
6a
m-F
m-Cl
m-Br
p-Cl
H
H
H
H
H
H
CN
80
74
77
78
30
14
33
71:29
72:28
75:25
75:25
67:33
66:34
70:30
75
75
74
81
41
32
32
82:18
85:15
86:14
85:15
75:25
70:30
77:23
78
77
76
86
40
40
35
90:10
93:7
93:7
75
81
79
82
41
37
38
82:18
87:13
88:12
88:12
65:35
56:44
68:32
63
75
78
82
48
38
45
66:34
67:33
68:32
68:32
60:40
53:47
64:36
72
76
64
77
14
29
20
82:18
84:16
86:14
82:18
90:10
70:30
80:20
67
80
70
77
13
33
10
90:10
92:8
92:8
89:11
81:19
83:17
82:18
6b
6c
6d
93:7
6e m-CN
6f
6g
87:13
76:24
82:18
p-CN
H
Table 6. Hoffmann test by using the reaction of 8 with 9.
Entry
Solvent
9
T [°C]
Ratio
Yield [%]
Conclusion
stable
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1,4-dioxane[a]
1,4-dioxane[a]
NMM[a]
NMM[a]
NMM[a]
NMM[a]
Et2O[a]
rac-9
(S)-9
rac-9
(S)-9
rac-9
(S)-9
rac-9
(S)-9
rac-9
(S)-9
rac-9
(S)-9
rac-9
(S)-9
rac-9
(S)-9
5
5
5
5
42:58
48:52
56:44
50:50
68:32
68:32
78:22
76:24
74:26
75:25
76:24
76:24
76:24
76:24
70:30
70:30
91
88
87
88
83
79
84
86
85
84
80
84
75
stable
–50
–50
–78
–78
–78
–78
–78
–78
–78
–78
–78
–78
unstable
unstable
unstable
unstable
unstable
unstable
Et2O[a]
CPME[a]
CPME[a]
toluene[a]
toluene[a]
hexane[a]
hexane[a]
THF
75
91[b]
93[b]
THF
[a] Contains 9% of THF. [b] See Hoffmann et al.[9c]
6646
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6643–6650

Solvent Effects on a [2,3]-Wittig Rearrangement
oxane (65:35 to 56:44) and NMM (60:40 to 53:47). The used to trap chiral carbanions is that a carbanion can be
resonance effect of the para-cyano group through a double generated in the presence of a trapping agent in the former
bond was smaller than that of the same group on the 1- case. Consequently, processes involving concerted depro-
phenyl moiety (Table 5, Entry 7).
tonation and trapping of the generated carbanion were also
Because it is possible that the effects of solvents and ad- considered. The reaction in hexane is a case in which such
ditives could operate solely on the rate of the [2,3]-Wittig a concerted deprotonation/[2,3]-Wittig rearrangement may
rearrangement, we became interested in determining be a major pathway (Scheme 2, 11).[20] This is supported by
whether the effects could also be observed in carbanion- the findings that the enantiomeric ratios were relatively
trapping reactions other than [2,3]-Wittig rearrangement. high, that the starting material was recovered without any
Thus, we decided to conduct the Hoffmann test[9] by using loss of optical purity at lower temperatures (Table 1, En-
[phenyl(trimethylsilyl)methyl]lithium (8) and Reetz aldehyde tries 12 and 13), and that planarization due to resonance
9[17] in the above solvents (Table 6). The substrate 8 has stabilization by the para-cyano group shows a less pro-
been reported to be configurationally unstable in THF nounced effect on the enantiomeric ratio (Table 5, En-
(Table 6, Entries 15 and 16).[9c] The diastereomeric ratio try 6).[21] Because THF, 1,2-dimethoxyethane (DME), and
was determined by using the stereospecificity of the Pe- diglyme can operate as strong solvating agents, albeit in-
terson elimination to give (Z)- or (E)-10. In acyclic ethers ferior to HMPA (Table 1, Entries 4–6; Table 4 Entry 3, 7,
and hydrocarbon solvents, reactions using rac-9 proceeded 11, 13–15, 18, 22, 26, and 30), this suggests that these sol-
smoothly to give 10 in ratios of 70–78% rac-(E)-10 to 22– vents can drive the conversion of a contact ion pair (CIP)
30% rac-(Z)-10 (Table 6, Entries 7, 9, 11, and 13). Experi- 12 into a separated ion pair 13. Cyclic ethers, 1,4-dioxane,
ments using (S)-9 resulted in almost the same dia- and NMM seem to behave differently to acyclic ethers, par-
stereomeric ratios as those obtained from the first set of ticularly regarding the resonance effect of the para-cyano
experiments described above (Table 6, Entries 8, 10, 12, and group (Table 5, Entry 6). Thus, the enantiomeric ratios in
14), indicating that the carbanion is configurationally un- the former solvents were much lower than those in the acy-
stable on the timescale of the reaction with aldehyde. Reac- clic ethers. In similar reactions with the corresponding cro-
tions of 8 with rac-9 in cyclic ethers, 1,4-dioxane, or NMM, tyl derivatives 1 (X = Me), only slight racemization was
at 5 °C, gave (E)-10 and (Z)-10 in ratios of 42:58 and 56:44, observed in THF, Et2O, and 1,4-dioxane (98–97:2–3 er in
respectively (Table 6, Entries 1 and 3), indicating the low THF; 100–98:0–2 er in Et2O; 99–97:1–3 er in 1,4-diox-
intrinsic diastereoselectivity under these conditions, proba- ane).[3] Consequently, marked differences in the enantio-
bly because of the relatively high reaction temperature. The meric ratios obtained with different solvents in the reaction
ratios in the reactions with (S)-9 changed to 48:52 and of (S)-4 can be ascribed to the extent of ion separation,
50:50 (Table 6, Entries 2 and 4), respectively. Because it was which depends on the nature of solvent and/or additive.
difficult to draw definitive conclusions from the subtle Thus, conversion of CIP into SIP caused by trapping of
changes in the ratios, we carried out the reaction in NMM lithium ions by the solvent and/or additive, allows charge
at –50 °C, expecting to observe an enhancement in the dia- delocalization into the double bond and the phenyl ring to
stereoselectivity.[18] Identical ratios (68:32) were obtained in occur, which can lead to racemization.[6,22,23]
both experiments, showing that the carbanion is configura-
tionally unstable at lower temperatures.[19] These results
suggest that the ratios obtained at 5 °C in the solvents are
significant, even if the differences are small, and that the
behavior of the cyclic ethers are different to those of the
other solvents with respect to the configurational stability
of the carbanions. Consequently, the solvent effects can, at
least in part, determine the configurational stability of chi-
ral carbanions.
Discussion
Although the influence of the solvent on the relative rates
of racemization and rearrangement, and on the aggregation
state of the bases and the carbanion in the system, are un-
Scheme 2. Process of chirality transfer in the [2,3]-Wittig rearrange-
known at present, we are tempted to assume that the sol-
vents and additives would not significantly affect the rela-
tive rate of the [2,3]-Wittig rearrangement, based on the re-
ment of 4.
How can the fact that reactions provide higher enantio-
sults of the low-temperature experiments (Table 3) and of meric ratios in bulkier acyclic ethereal solvents and in 1,4-
the Hoffmann test (Table 6). dioxane be explained? We propose the following speculative
The clearest difference between the [2,3]-Wittig re- hypothesis that accommodates the above observations. Sol-
arrangement approach and other systems that have been vation of the lithium cation, which loosens the C–Li coordi-
Eur. J. Org. Chem. 2010, 6643–6650
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6647

K. Takeda et al.
FULL PAPER
General Procedure for [2,3]-Wittig Rearrangement of 4: To a solu-
tion of 4 (100:0 er, 50.1 mg, 0.20 mmol) in dioxane (4 mL), was
added a solution of nBuLi (2.17 m in hexane, 276 μL, 0.60 mmol)
at 30 °C. The reaction mixture was stirred at room temperature for
10 min, and a few drops of saturated aqueous NH4Cl solution was
added. The mixture was diluted with Et2O, dried, and concen-
trated. The residual oil was subjected to column chromatography
(silica gel, 5 g; hexane/Et2O, 5:1) to give 5 (35.3 mg, 70%); 86:14 er
[Chiralcel OD; hexane/iPrOH, 9:1; flow rate 1.0 mL/min; detection
at 254 nm; tR = 7.35 min (major) and 8.38 min (minor)].
nation and increases the fraction of SIP, becomes less effec-
tive in bulkier solvents such as MTBE and CPME due to
the steric repulsion with 13.[24] Although a similar type of
repulsion can also exist in 12, this is reduced when the sol-
vent molecule is replaced with the allyl ether oxygen, as
found in 14. This is consistent with a reactant-like (early)
transition structure, such as 16, for the [2,3]-Wittig re-
arrangement of allyl lithiomethyl ether, as calculated by ab
initio methods by Wu, Houk and Marshall in which a lith-
ium atom coordinates with the ether oxygen atom.[25] With
bulkier ethers, the concerted mechanism proceeding
through 11 may also operate.
The fact that the use of 1,4-dioxane led to higher enan-
tiomeric ratios than the use of acyclic ethers in the reactions
of (S)-4 with nBuLi, can be explained by assuming that the
former solvent acts as a bidentate ligand to form a strained
complex 15 (X, Y = O),[26] which is more stable than 14
(Solv = acyclic ether). Although there is no report on the
General Procedure for the Preparation of Compounds (S)-6a–g:
Compounds (S)-6a, (S)-6b, (S)-6e, (S)-6f, and (S)-6g were prepared
starting from known or commercially available enones through Lu-
che reduction, Sharpless kinetic resolution, and allylation. Com-
pounds (S)-6c and (S)-6d were prepared from known alcohol deriv-
atives.[27] It was assumed that the allylation of the alcohols pro-
ceeded without any loss in the enantiomeric ratio.
(S,E)-1-[1-(Allyloxy)-3-phenylallyl]-3-fluorobenzene [(S)-6a]: To a
cooled (ice/water) solution of (E)-1-(3-fluorophenyl)-3-phenylprop-
use of N-methylmorpholine as an additive, to the best of 2-en-1-one[28] (2.00 g, 8.24 mmol) and CeCl3·7H2O (3.07 g,
8.24 mmol) in MeOH (66 mL) and CH2Cl2 (16.5 mL), was added
NaBH4 (312 mg, 8.24 mmol). The reaction mixture was warmed to
room temperature, and, after being stirred at the same temperature
for 30 min, the mixture was diluted with Et2O (80 mL) and 10%
aqueous NH4Cl solution (80 mL). The mixture was separated, and
the aqueous phase was extracted with Et2O (3ϫ80 mL). The com-
bined organic phases were washed with saturated brine (30 mL),
dried, and concentrated. The residual oil was subjected to column
chromatography (silica gel, 40 g; hexane/AcOEt, 5:1) to give (E)-1-
(3-fluorophenyl)-3-phenylprop-2-en-1-ol (1.93 g, 96%).
our knowledge, it may act in a manner similar to that of
1,4-dioxane, albeit with lower coordinating ability presum-
ably because of steric repulsion of the methyl group. The
differing behavior of 1,4-dioxane and N-methylmorpholine
to the other solvents assessed in the Hoffmann test, also
suggests their participation as a bidentate ligand. Conse-
quently, the solvent effect operates mainly on the configura-
tional stability of chiral carbanions and not on the [2,3]-
Wittig rearrangement.
To a cooled (–20 °C) solution of (E)-1-(3-fluorophenyl)-3-phen-
ylprop-2-en-1-ol (1.84 g, 7.5 mmol), l-(+)-diisopropyl tartrate [l-
(+)-DIPT; 703 mg, 3.00 mmol] and molecular sieves (4 Å; 703 mg)
in CH2Cl2 (30 mL), were added Ti(OiPr)4 (619 μL, 2.25 mmol) and
tert-butyl hydroperoxide (TBHP; 5.0–6.0 m in decane, 1.05 mL).
After the mixture had been stirred at the same temperature for
2.5 h, a solution of FeSO4·7H2O (6.81 g) and tartaric acid (2.27 g)
in water (31 mL) was added. After stirring for 10 min, the mixture
was extracted with Et2O (2ϫ46 mL), and the combined organic
phases were washed with water (31 mL) and saturated brine
(31 mL) and then dried and concentrated. The residual oil was sub-
jected to column chromatography (silica gel, 170 g; hexane/CH2Cl2,
1:3) to give (S,E)-1-(3-fluorophenyl)-3-phenylprop-2-en-1-ol
(710 mg, 39%); 100:0 er [Chiralcel OD-H; hexane/iPrOH, 7:1; flow
rate 1.0 mL/min; detection at 254 nm; tR = 11.7 min (major) and
16.9 min (minor)]; colorless oil; Rf = 0.34 (hexane/AcOEt, 3:1);
Conclusions
We have demonstrated that the steric course of the [2,3]-
Wittig rearrangement of 1,3-diphenyl-1-propenyloxy-2-pro-
pen-1-yl carbanion (S)-4 and its derivatives (S)-6a–g is
greatly affected by the solvent and the additive. We have
proposed that the differences can be attributable to the con-
figurational stability of the chiral carbanions, which de-
pends on the solvent and which reflects the ratio of CIP
and SIP associated with their solvated structures.
Experimental Section
General: Infrared spectra were recorded with an FTIR spectrome-
[α]2D1 = –13.7 (c = 1.04, CHCl ). IR (film): ν = 3334, 3064, 3031,
˜
3
1
ter. H NMR spectra were recorded with a 500 MHz spectrometer
2925, 2861 cm–1. H NMR (500 MHz, CDCl3, 25 °C): δ = 2.06 (d,
1
with samples in CDCl3 referenced to CHCl3 (δ = 7.26 ppm). 13C
NMR spectra were measured with a 125 MHz spectrometer with
samples in CDCl3 referenced to the CDCl3 triplet (δ = 77.2 ppm).
Resonance multiplicities are described as singlet (s), doublet (d),
triplet (t), multiplet (m), and broad (br.). Mass spectra were ob-
tained either in the EI mode or in the FAB mode either with NBA
as the matrix or without any matrix. For routine chromatography,
the following adsorbents were used: silica gel 60N, particle size 63–
210 μm, for column chromatography; precoated silica gel 60 F-254
plates for analytical thin-layer chromatography. All moisture-sensi-
tive reactions were performed under a positive pressure of nitrogen.
Anhydrous MgSO4 was used to dry all organic solvent extracts
during workup, and the removal of the solvents was performed
with a rotary evaporator. Anhydrous solvents and reagents were
obtained according to standard procedures.
3JH,H = 3.5 Hz, OH), 5.39 (dd, 3JH,H = 6.4, 3.5 Hz, 1 H, 1-H), 6.34
3
3
(dd, JH,H = 15.8, 6.4 Hz, 2-H), 6.70 (d, JH,H = 15.8 Hz, 1 H, 3-
H), 6.97–7.40 (m, 9 H, PhH) ppm. 13C NMR (125 MHz, CDCl3,
25 °C): δ = 74.7, 113.3 (d, JC,F = 21.9 Hz), 114.6 (d, JC,F
=
21.0 Hz), 122.0, 126.8, 128.2, 128.8, 130.2, 130.3, 131.1, 131.4,
136.4, 145.5, 145.6, 162.2 (d, JC,F = 244.2 Hz) ppm. HRMS: calcd.
for C15H13FO 228.0950; found 228.0950.
To a cooled (ice/water) solution of (S,E)-1-(3-fluorophenyl)-3-phen-
ylprop-2-en-1-ol (667 mg, 2.73 mmol) and allyl bromide (666 mg,
5.45 mmol) in DMF (12.6 mL) was added NaH (60% oil suspen-
sion, 218 mg, 5.45 mmol). The reaction mixture was warmed to
room temperature, and, after stirring at the same temperature for
40 min, the mixture was diluted with saturated aqueous NH4Cl
solution (10 mL), and extracted with Et2O (2ϫ20 mL). The com-
6648
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6643–6650

Solvent Effects on a [2,3]-Wittig Rearrangement
Mechanisms, American Chemical Society, Washington DC,
2003, pp. 119–162; h) R. W. Hoffmann, Stereochemical Aspects
of Organolithium Compounds, Topics in Stereochemistry (Eds.:
R. E. Gawley, J. Siegel), Wiley, New York, 2010, vol. 26, chap-
ter 5.
For reviews on [2,3]-Wittig rearrangements, see: a) T. Nakai,
K. Mikami, Chem. Rev. 1986, 86, 885–902; b) J. A. Marshall,
Comprehensive Organic Synthesis (Eds.: B. M. Trost, I. Flem-
ing), Pergamon, Oxford, 1991, vol. 3, pp. 975–1014; c) K. Mi-
kami, T. Nakai, Synthesis 1991, 594–604; d) T. Nakai, K. To-
mooka, Pure Appl. Chem. 1997, 69, 595–600; e) K. Tomooka,
The Chemistry of Organolithium Compounds, vol. 1 (Eds.: Z.
Rappoport, I. Marek), Wiley, Chichester, 2004, pp. 749–828.
M. Sasaki, H. Ikemoto, M. Kawahata, K. Yamaguchi, K.
Takeda, Chem. Eur. J. 2009, 15, 4663–4666.
bined organic phases were successively washed with water
(2ϫ10 mL) and saturated brine (10 mL), dried, and concentrated.
The residual oil was subjected to column chromatography (silica
gel, 30 g; hexane/AcOEt, 22:1) to give (S)-6a (640 mg, 82%); color-
less oil; Rf = 0.65 (hexane/AcOEt, 3:1); [α]2D8 = 9.04 (c = 1.00,
[2]
CHCl ). IR (film): ν = 3082, 3060, 3025, 2981, 2923, 2857 cm–1.
˜
3
1H NMR (500 MHz, CDCl3, 25 °C): δ = 4.00–4.11 (m, 2 H, 1-HЈЈ),
3
3
4.98 (d, JH,H = 7.3 Hz, 1 H, 1-HЈ), 5.22 (ddd, JH,H = 10.3, 3.0,
3
1.4 Hz, 1 H, 3-HЈЈ), 5.32 (ddd, JH,H = 17.2, 3.2, 1.6 Hz, 1 H, 3-
HЈЈ), 5.93–6.01 (m, 1 H, 2-HЈЈ), 6.25 (dd, JH,H = 16.0, 7.3 Hz, 1
H, 2-HЈ), 6.64 (d, JH,H = 16.0 Hz, 1 H, 3-HЈ), 6.96–7.40 (m, 9 H,
3
3
PhH) ppm. 13C NMR (125 MHz, CDCl3, 25 °C): δ = 69.6, 81.3,
113.8, 114.0 (d, JC,F = 22 Hz), 114.6 (d, JC,F = 21 Hz), 114.8, 117.3,
122.6, 122.6, 126.8, 128.1, 128.8, 129.8, 130.1, 130.2, 132.2, 134.8,
136.6, 144.1, 144.2, 162.3 (d, JC,F = 244.2 Hz) ppm. HRMS: calcd.
for C18H17FO [M – H]+ 267.1180; found 267.1191.
[3]
[4]
D. Y. Curtin, W. J. Koehl Jr., J. Am. Chem. Soc. 1962, 84, 1967–
1973.
[5]
[6]
[7]
[8]
S. Wu, S. Lee, P. Beak, J. Am. Chem. Soc. 1996, 118, 715–721.
A. Carstens, D. Hoppe, Tetrahedron 1994, 50, 6097–6108.
R. E. Gawley, Q. Zhang, Tetrahedron 1994, 50, 6077–6088.
D. C. Kapeller, F. Hammerschmidt, J. Org. Chem. 2009, 74,
2380–2388.
a) R. W. Hoffmann, J. Lanz, R. Metternich, G. Tarara, D.
Hoppe, Angew. Chem. Int. Ed. Engl. 1987, 26, 1145–1146; b)
R. Hirsch, R. W. Hoffmann, Chem. Ber. 1992, 125, 975–982; c)
R. W. Hoffmann, T. Rühl, J. Harbach, Liebigs Ann. Chem.
1992, 725–730; d) R. W. Hoffmann, M. Julius, F. Chemla, T.
Ruhland, G. Frenzen, Tetrahedron 1994, 50, 6049–6060; e) A.
Basu, D. Gallagher, P. Beak, J. Org. Chem. 1996, 61, 5718–
5719; f) P. Beak, A. Basu, D. J. Gallagher, Y. S. Park, S. Thayu-
manavan, Acc. Chem. Res. 1996, 29, 552–560; g) see also ref.[1h]
P. R. Peoples, J. B. Grutzner, J. Am. Chem. Soc. 1980, 102,
4709–4715.
a) E. J. Verner, T. Cohen, J. Am. Chem. Soc. 1992, 114, 375–
377; b) E. J. Verner, T. Cohen, J. Org. Chem. 1992, 57, 1072–
1073; c) R. Hoffmann, R. Brückner, Chem. Ber. 1992, 125,
1957–1963; d) K. Tomooka, T. Igarashi, T. Nakai, Tetrahedron
1994, 50, 5927–5932; e) R. Hoffmann, T. Rückert, R. Brückner,
Tetrahedron Lett. 1993, 34, 297–300; f) H. Schäfer, U.
Schöllkopf, D. Walter, Tetrahedron Lett. 1968, 9, 2809–2814; g)
D. A. Evans, D. J. Baillargeon, Tetrahedron Lett. 1978, 19,
3315–3322; h) U. Azzena, T. Denurra, G. Melloni, A. M. Pir-
oddi, J. Org. Chem. 1990, 55, 5532–5535; i) K. Tomooka, H.
Yamamoto, T. Nakai, J. Am. Chem. Soc. 1996, 118, 3317–3318;
j) K. Tomooka, H. Yamamoto, T. Nakai, Liebigs Ann./Recueil
1997, 1275–1281.
General Procedure for the Hoffmann Test with 8 and 9: To a solution
of benzyltrimethylsilane (114 mg, 0.25 mmol) in anhydrous THF
(750 μL) was added a solution of nBuLi (2.38 m in hexane, 116 μL,
0.275 mmol) at 0 °C. After stirring at the same temperature for
10 min, a solution of (S)-2-(dibenzylamino)-3-phenylpropionalde-
hyde (9; 165 mg, 0.50 mmol) in 1,4-dioxane (7.5 mL) was added by
using a cannula over 2 min. After stirring at 5 °C for 10 min, acetic
acid (1 m in 1,4-dioxane, 0.275 mL) was added dropwise, and the
mixture was diluted with Et2O (5 mL) and 10% aqueous NH4Cl
(5 mL). The phases were separated, and the aqueous phase was
extracted with Et2O (2ϫ5 mL). The combined organic phases were
washed with saturated brine (30 mL), dried, and concentrated. The
residual oil was subjected to column chromatography (silica gel,
30 g; hexane/CH2Cl2, 1:1) to give alkene 10 (88.6 mg, 88%) in an
(E)/(Z) ratio of 48:52. The alkenes were characterized by their
NMR spectra.
[9]
[10]
[11]
For the reaction of 8 with racemic aldehyde 9, the same procedure
was performed as described above.
Supporting Information (see footnote on the first page of this arti-
cle): Experimental procedures, spectroscopic data, and copies of 1H
and 13C NMR spectra for all new compounds.
Acknowledgments
[12]
Compound [D2]-4 was prepared by treatment of (E)-1,3-di-
phenylprop-2-en-1-ol with hydrochloric acid in [1,1-D2]allyl
alcohol. For the preparation of [1,1-D2]allyl alcohol, see: M.
Solomon, W. Hoekstra, G. Zima, D. Liotta, J. Org. Chem.
1988, 53, 5058–5062.
To secure complete deprotonation at lower temperatures, tBuLi
was used.
a) D. B. Collum, A. J. McNeil, A. Ramirez, Angew. Chem. Int.
Ed. 2007, 46, 3002–3017; b) D. B. Collum, Acc. Chem. Res.
1992, 25, 448–454.
Use of nBuLi as a base resulted in the formation of a butyl
ketone derivative as a byproduct in the case of (S)-6e and (S)-
6f.
This research was partially supported by a Grant-in-Aid for Scien-
tific Research (B) 19390006 (K. T.), a Grant-in-Aid for Young Sci-
entists (B) 20790011 (M. S.), a Grant-in-Aid for Young Scientists
(Start-up) 60549004 (H. I.) from the Ministry of Education, Cul-
ture, Sports, Science and Technology (MEXT). We thank the Re-
search Center for Molecular Medicine, Faculty of Medicine, Hiro-
shima University and the Natural Science Center for Basic Re-
search and Development (N-BARD), Hiroshima University for the
use of their facilities.
[13]
[14]
[15]
[1] a) D. Hoppe, The Chemistry of Organolithium Compounds, vol.
1 (Eds.: Z. Rappoport, I. Marek), Wiley, Chichester, 2004, pp.
1055–1164; b) D. Hoppe, F. Marr, M. Brüggemann, Organo-
lithiums in Enantioselective Synthesis (Ed.: D. M. Hodgson),
Springer, New York, 2003, pp. 61–137; c) D. M. Hodgson, K.
Tomooka, E. Gras, Organolithiums in Enantioselective Synthe-
sis (Ed.: D. M. Hodgson), Springer, New York, 2003, pp. 217–
250; d) J. Clayden, Organolithiums: Selectivity for Synthesis,
Pergamon, Oxford, 2002, pp. 169–240; e) A. Basu, S. Thayum-
anavan, Angew. Chem. Int. Ed. 2002, 41, 717–738; f) V. K. Ag-
garwal, Angew. Chem. Int. Ed. Engl. 1994, 33, 175–177; g) E.
Buncel, J. M. Dust, Carbanion Chemistry: Structures and
[16]
[17]
The absolute stereochemistry of 7a–g was assigned by analogy
with 5. See ref.[20a]
M. T. Reetz, M. W. Drewes, A. Schmitz, Angew. Chem. Int. Ed.
Engl. 1987, 26, 1141–1143.
The m.p. of 1,4-dioxane is 11 °C.
Although the origin of the different temperature dependencies
of the rates for the organolithium racemization and the reac-
tion with aldehyde is not clear, the same trend is observed in
the [2,3]-Wittig rearrangement (Table 3).
We reported that [2,3]-Wittig rearrangement of the type ob-
served with 4 proceeds with inversion of configuration at the
[18]
[19]
[20]
Eur. J. Org. Chem. 2010, 6643–6650
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6649

K. Takeda et al.
FULL PAPER
lithium-bearing carbon atom in the same way as reported for
the alkyl counterparts, see: a) M. Sasaki, M. Higashi, H. Masu,
K. Yamaguchi, K. Takeda, Org. Lett. 2005, 7, 5913–5915. For
systems other than chiral benzyl anions, see: b) R. Hoffmann,
R. Brückner, Angew. Chem. Int. Ed. Engl. 1992, 31, 647–649;
c) K. Tomooka, T. Igarashi, M. Watanabe, T. Nakai, Tetrahe-
dron Lett. 1992, 33, 5795–5798; d) D. C. Kapeller, L. Brecker,
F. Hammerschmidt, Chem. Eur. J. 2007, 13, 9582–9588.
[21] Although the hypothesis is not in accord with the widely ac-
cepted mechanism for [2,3]-Wittig rearrangement, there is no
report in the literature to the best of our knowledge on the
enantioselective [2,3]-Wittig rearrangement by using relatively
stable chiral carbanions such as that generated from 4 in non-
polar solvents.
[23] N. Deora, P. R. Carlier, J. Org. Chem. 2010, 75, 1061–1069 and
references cited therein.
[24] H. Ott, C. Däschlein, D. Leusser, D. Schildbach, T. Seibel, D.
Stalke, C. Strohmann, J. Am. Chem. Soc. 2008, 130, 11901–
11911 and references cited therein.
[25] Y.-D. Wu, K. N. Houk, J. A. Marshall, J. Org. Chem. 1990, 55,
1421–1423.
[26] For participation of dioxane as a bidentate ligand with lithium
ions, see: L. Pasumansky, C. J. Collins, L. M. Pratt, N. V.
Nguyên, B. Ramachandran, B. Singaram, J. Org. Chem. 2007,
˜
72, 971–976.
[27] Z. Chai, X.-Y. Liu, J.-K. Zhang, G. Zhao, Tetrahedron: Asym-
metry 2007, 18, 724–728.
[28] M. Verma, A. F. Chaudhry, C. J. Fahrni, Org. Biomol. Chem.
2009, 7, 1536–1536.
[22] H. J. Reich, W. H. Sikorski, J. L. Thompson, A. W. Sanders,
A. C. Jones, Org. Lett. 2006, 8, 4003–4006 and references cited
therein.
Received: July 27, 2010
Published Online: October 15, 2010
6650
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 6643–6650