Asymmetric allylation of aldehydes and glyoxylates through 'C-centered' chiral pentacoordinate allylsilicates or promoted by Lewis acid

Article abstract of DOI:10.1016/S0957-4166(99)00004-X

One-pot asymmetric allylation of aldehydes and glyoxylates with 'C- centered' chiral pentacoordinate allylsilicates generated from a chiral diol- modified allyltrichlorosilane 8 in the presence of Lewis bases, gave optically active homoallylic alcohols 4 with relatively high enantioselectivity (up to 81% ee). The reactions proceed via a six-membered cyclic transition state. In contrast, the allylation reactions of glyoxylate with allylalkoxysilanes promoted by TiCl₄ proceed through an acyclic transition state. The chiral auxiliaries residing at different positions on the molecules exhibited different abilities for asymmetric induction, depending on the reaction pathway and the stereochemistry of the transition state.

Full text of DOI:10.1016/S0957-4166(99)00004-X

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 10 (1999) 327–338
Asymmetric allylation of aldehydes and glyoxylates through ‘C-
centered’ chiral pentacoordinate allylsilicates or promoted by
Lewis acid1
Dong Wang,∗ Zhi Gang Wang, Ming Wen Wang, Yong Jun Chen, Li Liu and Yi Zhu
Institute of Chemistry, The Chinese Academy of Sciences, Beijing 100080, People’s Republic of China
Received 16 November 1998; accepted 15 December 1998
Abstract
One-pot asymmetric allylation of aldehydes and glyoxylates with ‘C-centered’ chiral pentacoordinate allylsi-
licates generated from a chiral diol-modified allyltrichlorosilane 8 in the presence of Lewis bases, gave optically
active homoallylic alcohols 4 with relatively high enantioselectivity (up to 81% ee). The reactions proceed via
a six-membered cyclic transition state. In contrast, the allylation reactions of glyoxylate with allylalkoxysilanes
promoted by TiCl4 proceed through an acyclic transition state. The chiral auxiliaries residing at different positions
on the molecules exhibited different abilities for asymmetric induction, depending on the reaction pathway and the
stereochemistry of the transition state. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
The allylation of carbonyl compounds with allylsilanes under Lewis acid conditions, first described
by Hosomi and Sakurai,2 has been extensively used in synthesis for the formation of C–C bonds
(Scheme 1).3 The possibility of using this reaction for the asymmetric synthesis of optically active
homoallylic alcohols, which can be converted to many important building blocks for optically active
natural product synthesis, has attracted considerable attention.4 Recently, it has been found that the
substituents on silicon can offer stereocontrol in the reaction, even though the reaction center is remote
from the silicon atom.5 However, for the reaction of aldehydes with both ‘Si-centered’ (chirality on
silicon) and ‘C-centered’ (chirality on the carbon of one of the substituents) chiral allylsilanes, only
modest enantioselectivity has been observed.4 This was attributed to the reaction (under Lewis acid
conditions) proceeding through a less rigid, open transition structure (acyclic transition state). It was
hoped that if the reaction could be made to proceed through a rigid cyclic transition state, higher asym-
metric induction could be obtained. Sakurai6 demonstrated that allylation of aldehydes with hydroxy
∗
Corresponding author. E-mail: g210@mimi.cnc.ac.cn
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00004-X

328
D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
compounds and allyltrifluorosilane in the presence of triethylamine, without a Lewis acid promoter,
gave the corresponding homoallylic alcohols in a stereospecific manner. Pentacoordinate allylsilicate
intermediates, which have been proved to have Lewis acid characters, were considered as possible
promoters.7 The reaction of pentacoordinate allylsilicate with carbonyl compounds was suggested to
proceed through a cyclic transition state. On the other hand, Kobayashi8 reported that allyltrichlorosilane
could also react with aldehydes in DMF or some co-solvent mixture via a pentacoordinate silicate
intermediate and six-membered cyclic transition state. Recently, chiral Lewis bases were employed to
catalyze asymmetric allylation, improving the enantioselectivity (65–92% ee).9,10 Herein, we wish to
report the investigation on the asymmetric allylation of aldehydes and glyoxylates via ‘C-centered’
chiral pentacoordinate allylsilicates generated from chiral alcohol-modified allyltrichlorosilane in the
presence of Lewis bases.11 It is also interesting to examine the pathway of the reaction of glyoxylates
with chiral allylalkoxysilanes under Lewis acid conditions, which should be through a Lewis acid
coordinated intermediate. Comparing pentacoordinate silicate intermediates with Lewis acid coordinated
intermediates, two different reaction pathways and related stereochemistry could be described. It is
helpful to understand the roles of the chiral auxiliaries resident on different parts of the molecule for
asymmetric induction in the allylation reactions.
Scheme 1.
2. Results and discussion
2.1. Allylation through a pentacoordinate allylsilicate intermediate
Several groups have described the reaction of pentacoordinate allylsilicates with carbonyl
compounds,7 but very few reported the asymmetric reaction using ‘C-centered’ chiral pentacoordinate
allylsilicate. As Hosomi suggested,12 from a 1:2 mixture of allyltrimethoxysilane 1 and (+)-tartaric acid
2, a ‘C-centered’ chiral pentacoordinate silicate intermediate 3 was generated in situ, which subsequently
reacted with aldehydes to offer chiral homoallylic alcohols 4 with low enantioselectivity (10–17% ee).
For comparison, we prepared optically active pentacoordinate silicates, (+)-triethylammonium bis[(R,R)-
tartrato(2-)]silicates [(+)-5a–b], from 1 and (+)-(2R,3R)-tartrates (6a–b), in 55 to 68% yield (Scheme 2).
By using optically active pentacoordinate silicates 5a–b, the asymmetric allylation of either aromatic
or aliphatic aldehydes was carried out without a catalyst to give optically active homoallylic alcohols 4
with 7–25% ee. It is suggested that a ‘chiral dilution effect’13 caused by bis-chiral auxiliaries on one
silicon atom was responsible for the poor enantioselectivity.
Although attempts to prepare ‘C-centered’ mono-chirally substituted pentacoordinate allylsilicates
have not been successful, we are still interested in generating them in situ. As mentioned above,
allyltrichlorosilane 7 can react with aldehydes in the presence of Lewis bases without a catalyst, giving
homoallylic alcohols. We hoped that by using a chiral diol to modify 7 in the Kobayashi reaction system,8
and thus generating in situ a ‘C-centered’ mono-chirally substituted pentacoordinate allylsilicate, the
enantioselectivity of the subsequent asymmetric allylation could be improved. (+)-Diisopropyl tartrate,
(+)-6b, was treated with 7 in dichloromethane for 3 h at room temperature and then the mixture was
concentrated to generate crude product 8b. Based on the known reaction of dichlorodimethylsilane with

D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
329
Scheme 2.
a tartrate species to give a cyclic silane,14 8b should be a cyclic species. A one-pot procedure for the
subsequent allylation reaction in the presence of Lewis base has been recommended.8,9 If a small quantity
of 7 remained in the crude product 8b, it might participate in the subsequent reaction with the aldehyde,
thus reducing the enantioselectivity. Therefore, (+)-6b was used in a slight excess. Prolonging the reaction
time ensured that the formation of cyclic 8b proceeded completely. In order to avoid the formation of
a bis-chirally substituted pentacoordinate allylsilicate 5, for 6b and 7, parallel addition of two reagents
must be employed and the addition speed of the starting materials must be strictly controlled. Moreover,
it was important not to use a base to absorb HCl produced in the course of the reaction. Without further
purification, the reactions of 8b with aldehyde 9a–c were carried out in the presence of Lewis bases, such
as Et3N, DMF and HMPA, giving the corresponding optically active homoallylic alcohols 4a–c15–17 in
a reasonable overall yield (Scheme 3). Chiral diols, (+)-1-phenylethanediol 6c and (+)-2-binaphthol 6d,
were also selected as modifiers, which could react with 7 to produce the corresponding cyclic silanes 8c
and 8d under the same conditions and be applied in the allylation of aldehydes.
Scheme 3.
The 29Si NMR spectrum of 8b showed only one upfield peak at δ −170.3 ppm in DMF, while a

330
D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
downfield peak at δ 1.96 ppm was observed in CDCl3. This provides evidence for the formation of a
pentacoordinate silicate intermediate 10 in the solution of 8b in DMF. On the other hand, there was no
7 present in the crude product 8b (29Si NMR of 7, δ 8.0 ppm in CDCl3).8b Apparently, a molecule of
DMF, which plays the role of a Lewis base, participates in the coordination of the silicon atom of 8b,
generating a chiral pentacoordinate silicate 10. Similarly, since Et3N and HMPA also have the ability to
coordinate with the silicon atom, chiral pentacoordinate silicates 11 and 12 can also be considered as
intermediates in the reaction. The chiral pentacoordinate silicates 10–12 are of sufficient Lewis acidity
to undergo nucleophilic addition to aldehydes, giving the optically active homoallylic alcohols 4a–c. All
three bases can promote the allylation reaction, but exhibit different promoting ability, the best for Et3N
and the worse for HMPA. As shown in Table 1, the enantioselectivities of the products 4 were as high as
81%. The ees of the reactions promoted by DMF were better than those promoted by Et3N. However, in
the case of DMF, although the reactions proceeded smoothly, the yield was relatively low. Lower reaction
temperatures were found to enhance the ee slightly. Compared with the reaction of aromatic aldehydes,
aliphatic aldehydes reacted sluggishly but with higher ee. Among the selected chiral diols (6b–d), the
ability of a tartrate moiety (6b) for asymmetric induction is much larger. In the case of 2-binaphthol
6d, the reaction was slowed down markedly, probably due to the steric effect. Despite strong reaction
conditions being employed (refluxing in chloroform for 48 h), the yield of the homoallylic alcohol 4a
was only 14% with 29% ee. It is interesting to compare the above with the reactions of ‘C-centered’ chiral
allylalkoxysilane with aldehydes promoted by a Lewis acid (∼20% ee), which are suggested to proceed
through acyclic transition states.18 However, the reaction of a chiral diol-modified allyltrichlorosilane 8,
an analogue of a ‘C-centered’ chiral allylalkoxysilane, with aldehydes in the presence of Lewis bases gave
higher enantioselectivity (up to 81% ee). This is due to the reaction proceeding through a pentacoordinate
silicate intermediate and subsequently a more rigid six-membered cyclic transition state (see later).
In order to obtain more stereochemical information on the transition state of the reaction of aldehydes
with ‘C-centered’ chiral pentacoordinated allylsilicates, we carried out crotylations of benzaldehyde with
chiral pentacoordinate 14 generated in situ from (+)-6b and crotyltrichlorosilanes, (E)-19 and (Z)-13,20 in
the presence of DMF (Scheme 4). The results showed that when (E)-13 (E:Z=4:1) was used, the product
was mainly anti-158b (anti:syn=4:1, yield: 72%), while for (Z)-13, syn-158b was obtained exclusively
(yield 76%). The sense of internal stereoinduction [(E)-13 to anti-15, (Z)-13 to syn-15] clearly supports
reaction via a hexacoordinate silicon species and a six-membered cyclic transition state. The absolute
configuration of the newly created stereogenic center in the homoallylic alcohols 4a–c produced was
deduced by comparison with the literature rotation (Table 1). The (R)-configurations of aliphatic 4b and
4c and (S)-configurations of aromatic 4a are consistent with the stereochemical outcome of the proposed
cyclic hexacoordinate silicon species having a chair conformation.
It is worthy to examine which chiral moieties in either chiral substrate or chiral pentacoordinate
silicate, mainly contribute to asymmetric induction in the double asymmetric allylation of chiral α-
ester aldehydes, with (+)-8b in the presence of DMF (Scheme 5 and Table 2). It was observed that only
the aldehyde carbonyl group in the molecule of an α-ester aldehyde, (1R,2S,3R)-(−)-menthyl glyoxylate
[(−)-16]21 is reactive to the pentacoordinate allylsilicate, giving the α-homoallylic hydroxy ester, menthyl
2-hydroxy-4-pentenoate 17, in yields of 83–87%. There are two doublets at δ 0.78 and 0.82 ppm in
1
the H NMR spectrum of product 17, which are assigned to the 100-H of the menthyl group of the
two diastereomers. Thus, the diastereomeric excess (de) of 17 can be determined as 33–38%. The
configuration of the newly created stereogenic center was deduced as S [cf. Chen et al.22]. If (−)-8b
was used in the reaction with (−)-16 instead of (+)-8b, the de of product 17 decreased to 13% and the
configuration of the new chiral center changed to R. If one chiral compound in this bis-chiral reactant
system, either chiral glyoxylate substrate or chiral silicon reagent, was replaced by an achiral analogue,

D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
331
Table 1
Asymmetric allylation of aldehydes 9 with diol-modified 7 under Lewis base
Scheme 4.
the effect of the chiral auxiliary on stereochemical outcome of the reaction would be more clear. For
the reaction of achiral 7 and chiral (−)-16, only a 5% de of product 17 was observed. In contrast, the
optically active product, (−)-butyl 2-hydroxy-4-pentenoate 19, produced from allylation of achiral butyl
glyoxylate 18 with (+)-6b in DMF had 41% de with S-configuration, which can be determined by treating
with lithium (−)-menthoxide, leading to ester-exchange completely, to give the known compound 17
(Scheme 5). These results indicate that the asymmetric induction is mainly from the chiral auxiliary
on the silicon atom, while the chiral moiety on the glyoxylate has little effect on the stereocontrol
of the reaction. The crotylation reaction of butyl glyoxylate 18 with (E)- and (Z)-13 modified by diol
6b in the presence of DMF was investigated, (E)-13 was transformed to anti-2023 (anti:syn=5:1, yield:
87%) and (Z)-13 to syn-2023 exclusively (yield: 86%), which also provides evidence for reaction via
a hexacoordinate silicon six-membered cyclic species with a chair conformation (Scheme 5). In the
transition state, the chiral menthoxy group of the glyoxylate is remote to the reaction center providing

332
D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
very weak asymmetric induction, while the chiral tartrate moiety is close to the six-membered ring,
resulting in better asymmetric induction.
Scheme 5.
2.2. Allylation of glyoxylate promoted by TiCl4
In the past decades, a lot of work has focused on Lewis acid promoted reaction of aldehydes with
allylsilanes. For comparison, we carried out the reactions of allylalkoxydimethylsilane 21–2518 with
glyoxylate (−)-16 in the presence of TiCl4 to obtain more information about the stereochemistry of the
reaction (Scheme 6 and Table 3). From the work of Kumada24 and Fleming,25 it has been concluded that
the Sakurai–Hosomi reaction proceeds through an acyclic antiperiplanar transition state 26. Denmark26
suggested that the synclinal transition state 27 is also operative under certain conditions, for example,
in intramolecular reactions. Obviously, in a synclinal transition state the silyl group may have a greater
effect on the stereochemistry of the reaction. As in our previous work,18 the chiral alkoxysilyl group
could bind with the Lewis acid, which may induce the reaction to proceed through synclinal transition
state 28,18,27 resulting in asymmetric induction in the course of reaction. However, it is surprising to
note that any change in alkoxy substituents on silicon, either chiral allylalkoxydimethylsilanes as 21–23
or achiral as allylisopropoxydimethylsilane 24 and allyltrimethylsilane 25, did not effect the de of the
product 17 to any significant extent. Moreover, if either (−)-21 or (+)-21 with different configuration, as
well as racemic (ꢀ)-23, were used as a reactant in the reaction with (−)-16, all the product 17 had S-
Table 2
Asymmetric allylation of glyoxylate 16 and 18 with 8 under DMFa

D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
333
configuration (Table 3). In contrast to the results from the reaction via pentacoordinate allylsilicates, the
chiral moiety on the glyoxylate plays the main role in asymmetric induction in the Lewis acid promoted
allylation reaction. This was attributed to their different reaction pathways and transition states.
Scheme 6.
It is also interesting that the addition order of the reagents could influence the de of the product 17
(Table 3). Table 3 shows that if (−)-16 was first mixed with TiCl4, followed by addition of allylsilanes
(method A), the reaction gave a higher de (30–42%). The formation of the chelate complex 29 was
assumed, leading to better stereochemical outcome. Nevertheless, if the addition order was changed
to first mixing 16 with allylsilane, then adding TiCl4 (method B), the de decreased to 11–14%. This
is presumably due to the formation of the mono-complex 30, but not the chelate complex 29, and
subsequent reaction with silane immediately. Apparently, the stereocontrol from the mono-complex is
weaker than that from the chelate complex. At the same time, we tried the crotylation of butyl glyoxylate
18 with crotyltrimethylsilane 31 promoted by TiCl4 with the addition order as method A (Scheme 7).
The transformation of (E)-3120 (E:Z=4:1) to syn-20 (syn:anti=4.5:1, yield: 91%) and (Z)-3120 to syn-
20 (syn:anti=>95:5, yield: 87%), respectively, provided an understanding of the stereochemistry of
the transition state, in which the carbonyl groups chelated complex with TiCl4 was formed. Hence,
there is no longer a coordinate site on Ti atom available to bind silylalkoxy group, thus an open
antiperiplanar arrangement of the alkoxysilyl groups could be accomplished. It is reasonable that syn-
Table 3
a
Allylation of (−)-16 with allylalkoxysilanes 21–25 promoted by TiC4

334
D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
selectivity is favorable for both (E)- and (Z)-crotylsilane in the reaction with glyoxylate promoted by
TiCl4 (Scheme 7).
Scheme 7.
In summary, the Lewis acid promoted allylation reaction of glyoxylate may proceed through two steps.
Firstly, Lewis acid coordinates to both aldehyde and ester carbonyl groups to form a chelate complex,
activating them at the same time. Secondly, the allylsilane selectively attacks the Si face of the aldehyde
carbonyl group, controlled by the chiral auxiliary of glyoxylate to provide the S-configuration of the
newly formed stereogenic center in the reaction. The alkoxy group is disposed anti to the carbonyl
groups chelated with TiCl4 (Scheme 7). The chirality of the alkoxysilyl moiety does not bring about
the asymmetric induction in the course of the reactions at all.
3. Conclusion
One-pot allylation reactions of aldehydes and glyoxylates via a ‘C-centered’ chiral pentacoordinate
allylsilicate intermediate generated in situ from chiral alcohol-modified allyltrichlorosilane 8 in the pre-
sence of Lewis bases gives optically active homoallylic alcohols 4 with relatively high enantioselectivities
(up to 81% ee). It was confirmed that the reactions proceed through a hexacoordinate six-membered
cyclic transition state. The chiral tartrate moiety is close to the six-membered ring of the transition state,
resulting in stronger asymmetric induction. In contrast, the allylation reactions of the glyoxylate with
allylsilanes promoted by TiCl4 proceed through an acyclic transition state, in which the alkoxy group
on silicon is disposed anti to the carbonyl groups chelated with TiCl4. The chirality of ester moiety of
glyoxylate plays the main role in the asymmetric induction in the course of the reaction.

D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
335
4. Experimental
Melting points were measured on a NAGEMA PHMK 05 apparatus and were uncorrected. IR spectra
1
were recorded on a Perkin–Elmer 782 infrared spectrometer. H, 13C and 29Si NMR spectra were
measured on Varian XL-200 and XL-300 spectrometers with tetramethylsilane as an internal standard.
Mass spectra (MS) were recorded on an MS-50/PS30 spectrometer. Optical rotations were measured on a
Perkin–Elmer 241 spectrometer at 589 nm. Flash chromatography was performed on a silica gel column.
All chemicals were reagent grade and used without further purification. The solvents were treated for use
in anhydrous reaction conditions before being employed.
4.1. General procedure for preparation of pentacoordinate allylsilicates 5a–b
A solution of allyltriethoxysilane (1b) (0.96 g, 6.0 mmol) and (+)-6b (2.8 g, 12.0 mmol) in 30 mL of
triethylamine was refluxed for 2 h and then cooled to room temperature. The resulting precipitate was
filtered and washed with hexane, then dried under vacuum to give a white solid 5b, 2.1 g (yield 55%).
4.1.1. Triethylammonium bis[(R,R)-diethyl tartrato(2-)]allylsilicate 5a
A white solid (yield 56%); mp 55–58°C; [α]D=+14.07 (c 2.9, CHCl3); νmax/cm−1 3430 (NH), 2900,
1740 (C_O); δH (CDCl3) 1.0–1.4 (m, 21H, CH3), 1.6–1.7 (m, 2H, CH2Si), 3.1–3.4 (m, 6H, NCH2),
4.1–4.4 (m, 12H, CH2, CH), 4.5–5.0 (m, 2H, CH2_), 5.8–6.0 (m, 1H, CH_), 9.6 (bs, 1H, NH); δH
(CDCl3–D2O) 1.02 (t, 9H, J=7.2 Hz, NCH2CH3), 1.32 (t, 12H, J=7.0 Hz, OCH2CH3), 1.62 (bs, 2H,
CH2Si), 2.53 (q, 6H, J=7.0 Hz, NCH2), 4.30 (q, 8H, J=7.0 Hz, OCH2), 4.54 (s, 4H, CH), 4.8–5.1 (m,
2H, CH2_), 5.6–5.9 (bs, 1H, CH_); δC (CDCl3) 8.1, 13.8, 24.9, 45.1, 60.2, 60.1, 71.9, 71.1, 109.6, 139.1,
173.2, 173.1; δSi −89.8. HR-MS m/z (FAB, negative ion) 477.1440 (M−Et3NH)−, calcd for C19H29O12Si:
477.1434.
4.1.2. Triethylammonium bis[(R,R)-diisopropyl tartrato(2-)]allylsilicate 5b
A white solid (yield 55%); mp (tube-sealing) 45–47°C; [α]D=+9.85 (c 1.5, CHCl3); νmax/cm−1 3460
(NH), 3030 (C_CH), 1730 (C_O), 1620 (C_C); δH (CDCl3) 1.1–1.4 (m, 33H, CH3), 1.6–1.7 (m, 2H,
CH2Si), 3.0 (bs, 6H, NCH2), 4.5 (s, 4H, CH), 4.6–5.2 (m, 6H, CH2_, CH), 5.9–6.2 (m, 1H, C_CH-
), 9.8 (bs, 1H, NH); δH (CDCl3–D2O) 1.02 (t, 9H, J=7.2 Hz, NCH2CH3), 1.32 (d, 24H, J=6.1 Hz,
CHCH3), 1.62 (bs, 2H, CH2Si), 2.54 (q, 6H, J=7.2 Hz, NCH2), 4.85 (s, 4H, CH), 5.18 (m, 4H, J=6.35
Hz, CHCH3), 4.82–5.85 (m, 3H, CH2_CH); δC (CDCl3) 8.3, 21.6, 25.1, 45.3, 67.9, 67.8, 72.4, 71.3,
109.7, 139.6, 173.1, 172.9; δSi (CDCl3) −89.2; HR-MS m/z (FAB, negative ion) 533.2042 (M−Et3NH)−,
calcd for C23H37O12Si: 533.2060.
4.2. General procedure for the reaction of 5 with aldehydes
A mixture of chiral pentacoordinate allylsilicate 5b (1.7 mmol) and aldehyde (1.5 mmol) in hexane (4
mL) was stirred at 60°C for 48 h. Then ethyl ether and hydrochloric acid (10%) were added to quench
the reaction, followed by stirring for 1 h. The aqueous phase was extracted with ether (3×20 mL).
The combined organic phases were washed with water (20 mL), aqueous KOH (1 M, 20 mL), water
(3×20 mL) and brine (20 mL), dried over MgSO4 and concentrated. The residue was purified by flash
chromatography (ethyl acetate–petroleum ether) to afford the optically active homoallylic alcohol 4.

336
D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
4.3. General procedure of the allylation of aldehydes 9 with chiral diol-modified allyltrichlorosilane 8
or crotyltrichlorosilane 13 promoted by DMF
To a flask charged with CH2Cl2 (60 mL) a solution of trichlorosilane 7 or crotyltrichlorosilane 13 (3
mmol) in CH2Cl2 (20 mL) and a solution of chiral diols 6 (3.3 mmol) in CH2Cl2 (20 mL) were added
dropwise in a parallel addition mode at room temperature. The mixture was stirred for 12 h at room
temperature and refluxed for 4 h, and then concentrated to ca. 3 mL, giving a solution of crude product
8 or 14. DMF (3 mL) was added to the solution, followed by stirring for 1 h. The aldehyde 9 (2 mmol)
was added, followed by stirring for 24 h at room temperature. Ether (40 mL) and saturated aqueous
sodium bicarbonate (20 mL) were added to quench the reaction. The aqueous phase was extracted with
ether (3×20 mL). The combined organic phases were washed with water (20 mL), aqueous potassium
hydroxide (1 M, 20 mL), water (3×20 mL) and brine (20 mL), dried over MgSO4 and concentrated. The
residue was purified by flash chromatography (ethyl acetate–petroleum ether) to afford the homoallylic
alcohol 4.
4.3.1. 1-Phenyl-3-buten-1-ol 4a15
A colorless oil; νmax/cm−1 3380 (OH), 3060 (_C-H), 1635 (C_C); δH (CDCl3) 2.10 (s, 1H, OH),
2.50 (t, 2H, J=7.0 Hz, CH2), 4.71 (t, 1H, J=7.0 Hz, OCH), 5.10–5.20 (m, 2H, CH2_), 5.64–5.86 (m, 1H,
CH_), 7.20–7.40 (m, 5H, PhH); δC (CDCl3) 43.8, 73.3, 118.3, 126.8, 127.5, 128.4, 134.4, 143.8.
4.3.2. 1-Nonene-4-ol 4b16
A colorless oil; νmax/cm−1 3350 (OH), 3060 (_CH-), 1630 (C_C); δH (CDCl3) 0.89 (m, 3H, CH3),
1.20–1.58 (m, 8H, CH2), 1.87 (s, 1H, OH), 2.00–2.40 (m, 2H, CH2), 3.64 (m, 1H, OCH), 5.05–5.20 (m,
2H, CH2_), 5.73–5.98 (m, 1H, CH_); δC 14.0, 22.6, 25.3, 31.8, 36.7, 41.8, 70.6, 117.8, 134.8.
4.3.3. 1-Dodecene-4-ol 4c17
A colorless oil; νmax/cm−1 3350 (OH), 3060, 1635 (C_C); δH (CDCl3) 0.88 (m, 3H, CH3), 1.20–1.50
(m, 14H, CH2), 2.12 (s, 1H, OH), 2.05–2.40 (m, 2H, CH2), 3.62 (m, 1H, OCH), 5.00–5.20 (m, 2H,
CH2_), 5.70–5.90 (m, 1H, _CH); δC (CDCl3) 14.0, 22.6, 25.6, 29.1, 29.5, 29.6, 31.8, 36.7, 41.8, 70.6,
117.7, 134.9.
4.4. anti-2-Methyl-1-phenyl-3-buten-1-ol anti-158b
A colorless oil, yield: 72%; νmax/cm−1 3402 (OH), 1635 (C_C), 1450; δH (CDCl3) 0.89 (d, 3H, J=6.9
Hz), 2.40–2.59 (m, 1H), 4.35 (d, 1H, J=7.1 Hz), 5.17–5.23 (m, 2H), 5.75–5.87 (m, 1H), 7.27–7.37 (m,
5H).
4.5. syn-2-Methyl-1-phenyl-3-buten-1-ol syn-158b
A colorless oil, yield: 76%; νmax/cm−1 3400 (OH), 1635 (C_C), 1450; δH (CDCl3) 1.02 (d, 3H, J=6.8
Hz), 2.54–2.65 (m, 1H), 4.60 (d, 1H, J=5.5 Hz), 5.04–5.08 (m, 2H), 5.71–5.83 (m, 1H), 7.26–7.38 (m,
5H).
The asymmetric allylation of glyoxylates 16 or 18 with trichlorosilanes 8 or 14 promoted by DMF
were carried out under the conditions described above to give the corresponding α-homoallylic hydroxy
ester.

D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
337
4.6. (−)-Menthyl 2-hydroxy-4-pentenoate 17
A colorless liquid; yield: 80%; νmax/cm−1 3450 (OH), 2900, 1720 (C_O), 1620, 1210; δH (CDCl3)
0.71–0.78 (2d, 3H), 0.80–0.95 (m, 6H), 0.95–2.10 (m, 9H), 2.30–2.64 (m, 2H), 4.24 (dd, 1H, J=2.0, 6.4
Hz), 4.80 (dt, 1H, J=4.6, 6.4 Hz), 5.10–5.25 (m, 2H), 5.70–5.90 (m, 1H); δC (CDCl3) (15.7, 16.2), (20.6,
20.8), 21.9, (22.8, 23.3), (25.8, 26.2), 31.3, 34.0, (38.6, 38.7), (40.6, 40.8), (46.8, 46.9), (69.6, 70.0),
(75.9, 76.1), (118.6, 118.7), (132.3, 132.4), 174.0. The data in parentheses are from two diastereoisomers.
Anal. calcd for C15H26O3: C, 70.83; H, 10.30. Found: C, 70.76; H, 10.28.
4.7. Butyl 2-hydroxy-4-pentenoate 19
A colorless liquid; [α]D −1.0 (c 4.4, CHCl3); νmax/cm−1 3480 (OH), 3070, 1760 (C_O), 1635; δH
(CDCl3) 0.87 (3H, t, J=7.3 Hz), 1.20–1.50 (m, 2H), 1.55–1.70 (m, 2H), 2.18(s, 1H, OH), 2.35–2.80 (m,
2H), 4.12 (dt, 2H, J=2.0, 6.6 Hz), 4.19 (dd, 1H, J=4.8, 6.4 Hz), 5.05–5.22 (m, 2H), 5.7–5.9 (m, 1H); δC
(CDCl3) 13.6, 19.0, 30.6, 38.7, 65.5, 69.9, 118.7, 132.5, 174.5; m/z (EI) 172 (M+, 1.9), 57 (100). Anal.
calcd for C9H16O3: C, 62.76; H, 9.36. Found: C, 62.65; H, 9.27.
The procedure for ester-exchange reaction: To a solution of (−)-menthol (0.16 g, 1.0 mmol) in CHCl3
(10 mL) n-butyl lithium (0.6 mL, 1.6 M in hexane, 1.0 mmol) was added at room temperature, followed
by stirring for 30 min. A solution of 19 in CHCl3 (5 mL) was added. The mixture was stirred for 12 h at
room temperature and refluxed for 16 h, until the starting material 19 disappeared by TLC. The mixture
was treated with saturated aqueous solution of sodium bicarbonate. The water phase was extracted by
ether (3×30 mL). The combined organic phase was washed with brine (3×30 mL), dried over MgSO4
and evaporated to give crude product. The crude product was purified with flash chromatography (silica
gel, pet. ether:ethyl acetate as eluent) to give product 17, 0.21 g, yield 85%.
4.8. syn-Butyl 2-hydroxy-3-methyl-4-pentenoate syn-2023
A colorless liquid, yield: 86%; νmax/cm−1 3498 (OH), 1730 (C_O), 1640; δH (CDCl3) 0.87 (t, 3H,
J=7.4 Hz), 0.94 (d, 3H, J=6.9 Hz), 1.32 (m, 2H), 1.57 (m, 2H), 2.61 (m, 1H), 2.64 (bs, 1H), 4.11 (m,
2H), 4.13 (d, 1H, J=3.2 Hz), 5.01–5.78 (m, 3H).
4.9. anti-Butyl 2-hydroxy-3-methyl-4-pentenoate anti-2023
A colorless liquid, yield: 87%; νmax/cm−1 3498 (OH), 1730 (C_O), 1640; δH (CDCl3) 0.87 (t, 3H,
J=7.4 Hz), 1.09 (d, 3H, J=6.9 Hz), 1.32 (m, 2H), 1.57 (m, 2H), 2.58 (m, 1H), 2.64 (bs, 1H), 4.04 (d, 1H,
J=3.2 Hz), 4.11 (m, 2H), 5.01–5.78 (m, 3H).
4.10. General procedure for allylation of glyoxylates promoted by TiCl4
To a solution of glyoxylate (2.5 mmol) in CH2Cl2 (15 mL) at −78°C was added dropwise the
solution of TiCl4 (2.5 mmol, 1.45 mL, 1.3 M in CH2Cl2). After stirring for 15 min, a solution of
allylalkoxydimethylsilane 21–25 or crotyltrimethylsilane 31 in 3 mL of CH2Cl2 was added dropwise at
the same temperature. The mixture was stirred at −78°C for 4 h, followed by hydrolysis with a saturated
aqueous solution of NH4Cl. The water phase was extracted by ether three times. The organic phase was
washed with brine (3×10 mL), dried over MgSO4 and filtered. The filtrate was evaporated and residue

338
D. Wang et al. / Tetrahedron: Asymmetry 10 (1999) 327–338
was purified by flash chromatography using hexane–ethyl acetate as eluent to give the corresponding
product, α-homoallylic hydroxy ester.
Acknowledgements
We gratefully acknowledge financial support by the National Natural Science Foundation of China and
thank Professor T. H. Chan for helpful suggestions.
References
1. A preliminary account of this work has been reported: Wang, Z. G.; Wang, D.; Sui, X. M. Chem. Commun. 1996, 2261.
2. Hosomi, A.; Sakurai, H. Tetrahedron Lett. 1976, 17, 1295.
3. (a) Majetich, G. In Organic Synthesis, Theory and Application; Hudlicky, T., Ed. JAI press: Greenwich, CT, 1989; (b)
Yamamoto, Y.; Asao, N. Chem. Rev. 1993, 93, 2207.
4. (a) Chan, T. H.; Wang, D. Chem. Rev. 1992, 92, 995; (b) Masse, C. E.; Panek, J. S. Chem. Rev. 1995, 95, 1293.
5. (a) Chan, T. H.; Koumaglo, K.; Horvath, R.; Wang, D.; Wei, Z. Y.; Yi, G. L.; Li, J. S. In Silicon Chemistry; Corey, Y.
J.; Corey, E. R.; Gaspar, P. P., Eds. Ellis Horwood: Chichester, 1988; Chapter 5; (b) Fleming, I.; Barbero, A.; Walter, D.
Chem. Rev. 1997, 97, 2063.
6. Kira, M.; Sato, K.; Sakurai, H. J. Am. Chem. Soc. 1990, 112, 257.
7. For review on the hypercoordinate silicates and their application in organic synthesis, see: (a) Sakurai, H. Synlett 1989,
1–8; (b) Sakurai, H. Selectivities in Lewis Acid Promoted Reactions; Schinzer, D., Ed. NATO ASI Series 289, 1989;
Chapter 11; (c) Chuit, C.; Corriu, R. J. P.; Reye, C.; Young, J. C. Chem. Rev. 1993, 93, 1371; (d) Kira, M. Yuki Gosei
Kagaku Kyokashi 1994, 52, 510 (Jpn); Chem. Abstr. 1994, 121, 134172; (e) Holmes, R. R. Chem. Rev. 1996, 96, 927.
8. (a) Kobayashi, S.; Nishio, K. Tetrahedron Lett. 1993, 34, 3453; (b) Kobayashi, S.; Nishio, K. J. Org. Chem. 1994, 59,
6620.
9. Denmark, S. E.; Coe, D. M.; Pratt, N. E.; Griedel, B. D. J. Org. Chem. 1994, 59, 6161.
10. (a) Iseki, K.; Kuroki, Y.; Takahashi, M.; Kishimoto, S.; Kobayashi, Y. Tetrahedron 1997, 53, 3513; (b) Iseki, K.; Mizuno,
S.; Kuroki, Y.; Kobayashi, Y. Tetrahedron Lett. 1998, 39, 2767; (c) Nakajima, M.; Saito, M.; Shiro, M.; Hashimoto, S. J.
Am. Chem. Soc. 1998, 120, 6419.
11. Recently similar work was reported: Zhang, L. C.; Sakurai, H.; Kira, M. Chem. Lett. 1997, 129.
12. Hosomi, A.; Kohra, S.; Ogata, K.; Yanagi, T.; Tominaga, Y. J. Org. Chem. 1990, 55, 2415.
13. Blaser, H.-U. Chem. Rev. 1992, 92, 935.
14. Jorge, C.; Orlando, L.; Ruben, C.; Edgard, V. Acta Cient, Venez. 1988, 39, 9; Chem. Abstr. 1989, 111, 7527.
15. Jadhav, P. K.; Bhat, K. S.; Perumal, P. T.; Brown, H. C. J. Org. Chem. 1986, 51, 432.
16. Tamao, K.; Kanatani, R.; Kumada, M. Tetrahedron Lett. 1984, 25, 1913.
17. Boldrini, G. P.; Lodi, L.; Tagliavini, E.; Tarasco, C.; Trombini, C.; Umani-Ronchi, A. J. Org. Chem. 1987, 52, 5447.
18. Wei, Z. Y.; Wang, D.; Li, J. S.; Chan, T. H. J. Org. Chem. 1989, 54, 5768.
19. Kira, M.; Hino, T.; Sakurai, H. Tetrahedron Lett. 1989, 30, 1099.
20. Aoki, S.; Mikami, K.; Terada, M.; Nakai, T. Tetrahedron 1993, 49, 1783.
21. Kornblum, N.; Frazier, H. W. J. Am. Chem. Soc. 1966, 88, 365.
22. Determined by derivation to a known compound through hydrogenation and ester-exchange: Chen, Y. J.; Huang, L.; Wang,
D.; Li, J. S. Chin. J. Chem. 1994, 12, 440.
23. Boldrim, G. P.; Lodi, L.; Tagliavini, E.; Tarasco, C.; Tromblini, C.; Umami-Ronchi, A. J. Org. Chem. 1987, 52, 5447.
24. Hayashi, T.; Konishi, M.; Ito, H.; Kumada, M. J. Am Chem. Soc. 1982, 104, 4962.
25. Fleming, I.; Terrett, N. V. Tetrahedron Lett. 1983, 24, 4153.
26. Denmark, S. E.; Henke, B. R.; Weber, E. J. Am. Chem. Soc. 1987, 109, 2512.
27. Chan, T. H.; Wang, D.; Pellon, P.; Lamothe, S.; Wei, Z. Y.; Li, L. H.; Chen, L. M. In Frontiers of Organosilicon Chemistry;
Bassindale, A. R.; Gaspar, P. P., Eds. The Royal Society of Chemistry: Cambridge, 1991, p. 344.