Desilylation versus elimination reactions of β-hydroxysilanes: Effect of substituents on silicon

Article abstract of DOI:10.1016/S0040-4039(03)00540-9

β-Hydroxy(trimethyl)silanes undergo stereospecific desilylation reactions to give alcohols using KOt-Bu in DMSO:H₂O, but the reactions are very slow. In this work, desilylation reactions are facilitated using β-hydroxysilanes with (i-PrO)Me₂Si, HMe₂Si, or HOMe₂Si groups (e.g. 9→11). The (HO)Me₂Si compound (10) (or alkoxide) was shown to be an intermediate, and undergoes β elimination less readily than the Me₃Si compound (7).

Full text of DOI:10.1016/S0040-4039(03)00540-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 3409–3412
Desilylation versus elimination reactions of b-hydroxysilanes:
effect of substituents on silicon
Paul F. Hudrlik,* Petros Gebreselassie, Laykea Tafesse and Anne M. Hudrlik
Department of Chemistry, Howard University, Washington, DC 20059, USA
Received 21 February 2003; accepted 24 February 2003
Abstract—b-Hydroxy(trimethyl)silanes undergo stereospecific desilylation reactions to give alcohols using KOt-Bu in
DMSO:H2O, but the reactions are very slow. In this work, desilylation reactions are facilitated using b-hydroxysilanes with
(i-PrO)Me2Si, HMe2Si, or HOMe2Si groups (e.g. 9“11). The (HO)Me2Si compound (10) (or alkoxide) was shown to be an
intermediate, and undergoes b elimination less readily than the Me3Si compound (7). © 2003 Elsevier Science Ltd. All rights
reserved.
Silicon compounds are widely used in organic synthesis,
partly because of the stability of many types of
organosilicon compounds to a variety of reagents,
enabling the silicon group to be carried through a
multi-step synthesis, and because of the availability of
conditions to remove the silicon group when it is no
longer needed. For example, arylsilanes, vinylsilanes,
and epoxysilanes are easily desilylated, and these reac-
tions have been widely used. In some cases, however,
removal of silicon is not easy. For example, b-func-
tional organosilicon compounds readily undergo elimi-
nation reactions, but desilylation (without elimination)
is difficult.1 In this work, we find that desilylation of
b-hydroxysilanes is greatly facilitated by changing the
silyl group.
A number of methods for preparing b-hydroxysilanes
exist,2 some with impressive stereochemical control.
The olefin-forming b-elimination reactions of b-hydroxy-
silanes are stereospecific, and take place under mild
conditions and in high yield, so have been useful for the
synthesis of olefins and heteroatom-substituted olefins
of defined stereochemistry.3 Some time ago we showed
that b-hydroxy(trimethyl)silanes undergo desilylation
reactions to give simple alcohols (e.g. 1“2),4 that the
reactions were stereospecific,4 and could be used for
introduction of deuterium.4b However, the reactions
were very slow in unactivated compounds unless the
silicon was at a primary carbon. For example, treat-
ment of b-hydroxysilane 7 with 5% KOt-Bu in 19:1
DMSO-d6:D2O containing 18-crown-6 at rt gave a 2:3
mixture of starting material (7) and product (cis-2-de-
uteriocyclohexanol) after 15 days.4b
These reactions are believed to take place via a pathway
that can be considered as
a
homo-Brook
rearrangement5 followed by hydrolysis (see Scheme 1).
An anionic four-membered ring species was proposed
(such as that generally accepted for the b-elimination
reactions,2d,6 which are normally carried out under
anhydrous conditions such as KH/THF).
If these reactions are to be useful, they need to be made
faster. (The reactions can be made faster by decreasing
the water concentration, but this is accompanied by
increasing amounts of olefin from b-elimination.) We
have therefore studied the effect of substituents on the
silicon on the course of the reaction. Initially we
attempted to facilitate the desilylation reactions by
replacing the Me3Si group by PhMe2Si or Ph2MeSi
groups. [Phenyl groups on silicon have been shown to
Keywords: desilylation; elimination; b-hydroxysilane.
* Corresponding author. Tel.: +1-202-806-4245; fax: +1-202-806-5442;
e-mail: phudrlik@fac.howard.edu
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00540-9

3410
P. F. Hudrlik et al. / Tetrahedron Letters 44 (2003) 3409–3412
We have now found that the desilylations of b-hydroxy-
silanes having the silicon at a secondary carbon can be
carried out effectively if the (i-PrO)Me2Si, HMe2Si, or
(HO)Me2Si groups are used (i.e. 4“5; 8, 9, or 10“11).
Furthermore, b-hydroxysilanes 10 and 7, having the
(HO)Me2Si and Me3Si groups, respectively, undergo a
dramatic change in relative reactivity at low water
concentrations.
Preparation of starting materials. The b-hydroxysilanes
were prepared as shown in Scheme 2. The reaction
conditions were not optimized. All the reaction path-
ways take advantage of the fact that a,b-epoxysilanes
generally undergo ring opening at the a position.2b,c,3
To prepare b-hydroxysilane 4, we treated the (i-
PrO)Me2Si epoxysilane 139 with MeLi/CuCN (2:1).
[Tamao and co-workers9 earlier found that 13 under-
went the normal a opening characteristic of reactions of
epoxysilanes with organocopper reagents2b without
affecting the i-PrO group.]
Scheme 1.
The most direct pathway to the cyclic b-hydroxysilanes
8–10 involved epoxidation of vinylsilane 16 followed by
ring opening of the epoxide with LiAlH4. Treatment of
16 with MCPBA resulted in oxidation of the SiꢀH as
well as the double bond: Epoxysilane 17 was isolated
(after chromatography) in 65% yield. [Epoxysilane 18
was obtained (24% yield) in less pure form, but was
identified by GC/MS (171 (M+−H, 6), 157 (M+−Me, 8),
129 (17), 77 (60), 75 (Me2SiOH+, 100)) and by 1H NMR
(includes l 0.158 (s), 3.11 (appears as d, J=3 Hz).]
Treatment of 17 with LiAlH4 resulted in the HMe2Si
b-hydroxysilane 8, the expected product of a opening.
[A more efficient way to 8, which we intend to explore,
might be LiAlH4 reduction of the crude product from
the epoxidation.] Small scale reactions of the HMe2Si
b-hydroxysilane 8 with NaOi-Pr in i-PrOH appeared to
give fairly pure (i-PrO)Me2Si b-hydroxysilane 9, with
only small amounts of hydrolysis product, silanol 10.
However, in a larger scale reaction 37% of 9 and 23%
of 10 were isolated after chromatography.
As an alternative entry to the cyclic b-hydroxysilanes 9
and 10, we also looked at reduction of the (i-
PrO)Me2Si epoxysilane 20. Compound 20 was prepared
by treatment of vinylsilane 16 with i-PrOH/H2PtCl6 to
give vinylsilane 19, followed by treatment with
MCPBA. Epoxide 20 was normally used in crude form,
although some was chromatographed for spectra. Reac-
tion of epoxide 20 with LiAlH4 gave HMe2Si b-hydroxy-
silane 8, product of both a opening of the epoxide and
reduction of the i-PrO group, even when the reaction
was carried out at low temperature (−78°C). The (i-
PrO)Me2Si b-hydroxysilane 9 was prepared in 30%
yield by treatment of 8 with H2PtCl6 in i-PrOH.
Scheme 2. Reagents and conditions: (a) MeLi/CuCN (2:1),
ether, −70°C for 4 h, warm to rt (84%); (b) From 14: Li
powder, naphthalene, THF, −78°C for 40 min; add 14, −78°C
for 4 h, add ClMe2SiH, −78°C for 10 h (54%). From 15:
t-BuLi, THF–ether–hexane, −105 to −80°C, 45 min, then
ClMe2SiH, allow to warm to 0°C (82%); (c) MCPBA,
Na2HPO4, CH2Cl2, 0°C, 6 h (65% of 17); (d) LiAlH4, THF,
rt, 8 h (78%); (e) NaOi-Pr, i-PrOH, rt, 3 h (37% of 9 and 23%
of 10); (f) i-PrOH, H2PtCl6, 0°C, 2 h (43%); (g) MCPBA,
CH2Cl2, 0°C to rt, 18 h; (h) LiAlH4, THF, rt, 5 h (62%); (i)
i-PrOH, H2PtCl6, 0°C, 4 h (30%).
accelerate the Brook rearrangement,7 presumably by
stabilizing an anionic pentacoordinated intermediate.]
Unfortunately, preliminary experiments showed the
PhMe2Si and Ph2MeSi groups facilitate both desilyla-
tion and elimination reactions.8
Desilylation reactions. For the desilylations we used 5%
KOt-Bu (5% by weight of the DMSO) in varying ratios
of DMSO/H2O with 5 mol% of 18-crown-6 at rt.
We initially studied the (i-PrO)Me2Si b-hydroxysilane 4
and found that it undergoes desilylation using 50:1

P. F. Hudrlik et al. / Tetrahedron Letters 44 (2003) 3409–3412
3411
DMSO:H2O (20 h, rt) to form 3-nonanol (5) and
2-nonene (6) in a GC ratio of 92:8.10 The 2-nonene was
identified by GC/MS, and the stereochemistry was not
determined; the 3-nonanol was isolated by chromatog-
raphy in 76% yield.
We also subjected HMe2Si b-hydroxysilane 8 and
silanol 10 to the above desilylation conditions. Com-
pound 8 disappeared by the first aliquot (within 5–10
min) to form silanol 10; subsequent reactions were
comparable to those above.
In agreement with our earlier work,4 for b-hydroxysi-
lanes having the silicon at a secondary center, we found
the cyclic system to be more sluggish than the acyclic.
The (i-PrO)Me2Si b-hydroxysilane 9 however, under-
went smooth desilylation using 200:1 DMSO:H2O (20
h, rt) to form cyclohexanol (11) with cyclohexene (12)
in 74 and 9.5% GC yields, respectively (using butylben-
zene as internal standard). In a preparative experiment,
67% yield of cyclohexanol (11) was isolated. When the
above reactions were followed by GC analysis (with
aqueous work-up) of aliquots, the (i-PrO)Me2Si b-
hydroxysilanes (4 and 9) were found to disappear by
the first aliquot (within 5–10 min at rt) with formation
of a lower retention time compound, believed to be the
corresponding silanol (e.g. 10 from 9).
We were interested to compare the reactivity of the
Me3Si compound 7 with that of silanol 10 (or its
precursors). The comparison could only be approxi-
mate for the following reasons: The possible formation
of disiloxane or oligomers from 10 means that reactiv-
ity comparisons represent an upper limit for the reactiv-
ity of 10. In addition, because of the small amounts of
water used, the conditions were difficult to reproduce
exactly, and small amounts of water made a large
difference in the reactivity, especially with 7. However,
several comparisons were carried out on a mixture of
the (i-PrO)Me2Si compound 9 and the Me3Si com-
pound 7.
The Me3Si compound 7 and silanol 10 (from hydrolysis
of 9) disappeared at comparable rates under conditions
with higher water concentrations, but under conditions
with very little water, 7 disappeared much more quickly
than 10, forming predominately cyclohexene (while 10
formed predominantly cyclohexanol). At the ratio of
96:4 DMSO:H2O, the Me3Si compound 7 and silanol
10 disappeared at comparable rates: after 48 h, about
60% of each remained. At the ratio of 98:2
DMSO:H2O, the Me3Si compound 7 disappeared a
little more quickly: after 24 h, 46% of 10 and 15–20% of
7 remained. At the ratio of 99.5:0.5 DMSO:H2O, the
Me3Si compound 7 was completely gone within 5–10
min, and the product was predominantly cyclohexene
(12). For comparison, 61% of silanol 10 remained after
3 h under these conditions, and the major product was
cyclohexanol (11) (see Table 1).
We also carried out a series of experiments in which the
ratios of DMSO:H2O were varied from 96:4 to 100:0.11
The (i-PrO)Me2Si b-hydroxysilane (9) was initially used
as a substrate. GC yields (and % remaining starting
material) were determined using butylbenzene as an in-
ternal standard. The results are summarized in Table 1.
Under all the reaction conditions, the (i-PrO)Me2Si
b-hydroxysilane 9 disappeared by the first aliquot
(within 5–10 min) with appearance (after aqueous
work-up) of the corresponding silanol (10) (identified
by GC/MS). Silanol 10 disappeared more slowly to
form cyclohexanol (11) as the major product. Very little
cyclohexene (12) was observed until the ratio of
DMSO:H2O reached 100:0. The reactions became
somewhat faster as the water concentration decreased.
Table 1. Reaction of the (i-PrO)Me2Si compound 9 with KOt-Bu in DMSO:H2O
DMSO:H2O
Time (h)
% silanol 10a
GC yield of 11 (%)
GC yield of 12 (%)
b
b
b
b
b
b
b
b
b
96:4
96:4
97:3
97:3
97:3
98:2
98:2
98:2
2
120
9
45
120
2
98
55
76
1
49
20
53
77
9
36
74
88
25
42
53
55
27
48
53
32
b
88
60
9
48
120
3
6.5
12
36
3
13
b
98:2
99.5:0.5
99.5:0.5
99.5:0.5
99.5:0.5
100:0
100:0
100:0
61
37
24
7
32
10
0
1
2
2
2
44
33
37
12
21
a The percentage remaining 10 was taken to be 100% at the first aliquot (5–10 min). (It should be noted that silanol 10 may have been
accompanied by the corresponding disiloxane or oligomers, which might not have been detected under our GC conditions.)
b Little or none.

3412
P. F. Hudrlik et al. / Tetrahedron Letters 44 (2003) 3409–3412
Discussion. With all silyl groups studied, when the
water concentration was sufficiently high, the reac-
tions of b-hydroxysilanes with KOt-Bu in DMSO/
H2O gave mostly alcohol from desilylation, with little
or no olefin from b elimination. However, in some
cases the reactions are too slow to be practical.
Under the reaction conditions, the (i-PrO)Me2Si and
HMe2Si groups were quickly converted to the
(HO)Me2Si group, and the (HO)Me2Si compound was
more slowly converted to the products. As the water
concentration was decreased, both the Me3Si and
(HO)Me2Si compounds disappeared more quickly,
and the percentage of olefin in the product mixtures
increased slightly. At very low water concentrations
(ꢀ0.5%), the Me3Si compound disappeared very
quickly, giving mainly olefin from b elimination. The
(HO)Me2Si compound disappeared more slowly, giv-
ing mainly alcohol from desilylation.
Acknowledgements
Financial support from the National Science Founda-
tion (CHE9505465) is gratefully acknowledged.
References
1. For example, see: Adam, W.; Saha-Mo¨ller, C. R.;
Schmid, K. S. J. Org. Chem. 2001, 66, 7365–7371.
2. (a) Hudrlik, P. F.; Peterson, D. J. Am. Chem. Soc. 1975,
97, 1464–1468; (b) Hudrlik, P. F.; Peterson, D.; Rona, R.
J. J. Org. Chem. 1975, 40, 2263–2264; (c) Hudrlik, P. F.;
Hudrlik, A. M.; Rona, R. J.; Misra, R. N.; Withers, G.
P. J. Am. Chem. Soc. 1977, 99, 1993–1996; (d) See also:
Ager, D. J. Org. React. 1990, 38, 1–223.
3. Hudrlik, P. F.; Hudrlik, A. M. In Advances in Silicon
Chemistry; Larson, G. L., Ed.; JAI Press: Greenwich,
CT, 1993; Vol. 2, pp. 1–89.
4. (a) Hudrlik, P. F.; Hudrlik, A. M.; Kulkarni, A. K. J.
Am. Chem. Soc. 1982, 104, 6809–6811; (b) Hudrlik, P. F.;
Holmes, P. E.; Hudrlik, A. M. Tetrahedron Lett. 1988,
29, 6395–6398.
5. For a review of the Brook rearrangement, see: Brook, A.
G. Acc. Chem. Res. 1974, 7, 77–84.
6. Kawashima, T.; Okazaki, R. Synlett 1996, 600–608.
7. Brook, A. G.; LeGrow, G. E.; MacRae, D. M. Can. J.
Chem. 1967, 45, 239–253.
The difference between the Me3Si and (HO)Me2Si
compounds at low water concentrations is remark-
able. Under the same conditions (with 0.5% H2O),
the Me3Si compound undergoes rapid elimination
(e.g. 7“12) while the (HO)Me2Si compound under-
goes slow desilylation (e.g. 10“11). The presence of
an OH group on the silicon, which may exist as O−
in the presence of base, seems to slow down the elim-
ination reaction compared to that with the analogous
silyl group having a Me in place of the OH. The low
reactivity of the (HO)Me2Si compounds may possibly
result from the formation of dianions under basic
conditions.
8. Bhamidipati, R. S., Ph.D. Dissertation, Howard Univer-
sity, 1991.
9. Tamao, K.; Najako, E.; Ito, Y. J. Org. Chem. 1987, 52,
4412–4414.
10. Spectral grade DMSO was used.
11. DMSO was freshly distilled from CaH2.
12. It should be noted that silanols, alkoxysilanes, and
halosilanes can be prepared from a variety of precursors,
including arylsilanes. See: (a) Fleming, I. Chemtracts Org.
Chem. 1996, 9, 1–64; (b) Jones, G. R.; Landais, Y.
Tetrahedron 1996, 52, 7599–7662.
This work suggests that b-hydroxysilanes could find
use as precursors to alcohols if the silicon has a lig-
and which will be converted to OH (or O−) under the
reaction conditions, and that it should be possible to
fine-tune the reaction conditions to suit the substrate
by adjusting the water concentration.12