Sakurai addition and ring annulation of allylsilanes with α,β-unsaturated esters. Experimental results and ab initio theoretical predictions examining allylsilane reactivity

Article abstract of DOI:10.1021/jo991177i

Trimethylallylsilane has been shown to add to methyl acrylate in good yield when catalyzed by TiCl₄ at room-temperature despite literature reporting to the contrary. Further, even with these small alkyl ligands on the metal, ring annulation occurs to a large extent, in addition to simple allylation (Sakurai addition). The kinetic product is the ((trimethylsily)methyl)cyclobutane derivative which can be isomerized to cyclopentanoid, the thermodynamic product, if left in the presence of the catalyst. Consistent with other literature in this area, increasing the size of the ligands on silicon increases both the rate of product formation and the proportion of ring annulation relative to allylation. To develop a predictive model for allylsilane reactivity, ab initio gas-phase calculations have been made on the parent allylsilane with different ligands on the metal and on the reaction between these allylsilanes and acrolein, acrylic acid, and methyl acrylate. Predictions indicate that as the length of n-alkyl ligands on silicon increase, so does the apparent ability of the Si-Cα bond of the allylsilane to hyperconjugate with developing vacant p orbital on Cβ as the allylsilane begins to attack an electrophile. This is corroborated by a gradually increasing HOMO in the ground-state allylsilane as the ligands are changed from methyl through to n-hexyl and an increasing Si-Cα bond length and decreasing Si-Cα-Cβ bond angle in the protonated species. These results in the gas phase mirror the reactivity of these n-alkyl-substituted allylsilanes in experiment; i.e., as the length of the alkyl chain increases, reactivity increases significantly. Triisopropylallylsilane, a very reactive silane, appears to anomalous in charge distribution and geometrical features compared with other substituted allylsilane systems which is due, presumably, to steric effects. The calculations on the protonated species would indicate that almost no hyperconjugative stabilization can occur on the basis of the bond lengths and angles necessary to promote good orbital overlap between the Si-Cα bond and the empty p orbital on Cβ. However, the gas-phase reaction of the triisopropylallylsilane with acrolein and methyl acrylate led to comparatively low energy barriers of 13.1 and 24.5, respectively, which is consistent with its high experimental reactivity. Together, this computational analysis has produced a useful model for predicting allylsilane reactivity and some possible explanations for this reactivity.

Full text of DOI:10.1021/jo991177i

3666
J . Org. Chem. 2000, 65, 3666-3678
Sa k u r a i Ad d ition a n d Rin g An n u la tion of Allylsila n es w ith
r,â-Un sa tu r a ted Ester s. Exp er im en ta l Resu lts a n d a b In itio
Th eor etica l P r ed iction s Exa m in in g Allylsila n e Rea ctivity
Michael G. Organ,*,† Vladimir Dragan,† Michael Miller,† Robert D. J . Froese,‡ and
J ohn D. Goddard‡
Department of Chemistry, York University, 4700 Keele Street, Toronto, Ontario, Canada M3J 1P3, and
Guelph-Waterloo Centre for Graduate Work in Chemistry, Department of Chemistry and Biochemistry,
University of Guelph, Guelph, Ontario, Canada N1G 2W1
Received J uly 23, 1999
Trimethylallylsilane has been shown to add to methyl acrylate in good yield when catalyzed by
TiCl4 at room-temperature despite literature reporting to the contrary. Further, even with these
small alkyl ligands on the metal, ring annulation occurs to a large extent, in addition to simple
allylation (Sakurai addition). The kinetic product is the ((trimethylsilyl)methyl)cyclobutane
derivative which can be isomerized to cyclopentanoid, the thermodynamic product, if left in the
presence of the catalyst. Consistent with other literature in this area, increasing the size of the
ligands on silicon increases both the rate of product formation and the proportion of ring annulation
relative to allylation. To develop a predictive model for allylsilane reactivity, ab initio gas-phase
calculations have been made on the parent allylsilane with different ligands on the metal and on
the reaction between these allylsilanes and acrolein, acrylic acid, and methyl acrylate. Predictions
indicate that as the length of n-alkyl ligands on silicon increase, so does the apparent ability of the
Si-CR bond of the allylsilane to hyperconjugate with developing vacant p orbital on Câ as the
allylsilane begins to attack an electrophile. This is corroborated by a gradually increasing HOMO
in the ground-state allylsilane as the ligands are changed from methyl through to n-hexyl and an
increasing Si-CR bond length and decreasing Si-CR-Câ bond angle in the protonated species.
These results in the gas phase mirror the reactivity of these n-alkyl-substituted allylsilanes in
experiment; i.e., as the length of the alkyl chain increases, reactivity increases significantly.
Triisopropylallylsilane, a very reactive silane, appears to anomalous in charge distribution and
geometrical features compared with other substituted allylsilane systems which is due, presumably,
to steric effects. The calculations on the protonated species would indicate that almost no
hyperconjugative stabilization can occur on the basis of the bond lengths and angles necessary to
promote good orbital overlap between the Si-CR bond and the empty p orbital on Câ. However,
the gas-phase reaction of the triisopropylallylsilane with acrolein and methyl acrylate led to
comparatively low energy barriers of 13.1 and 24.5, respectively, which is consistent with its high
experimental reactivity. Together, this computational analysis has produced a useful model for
predicting allylsilane reactivity and some possible explanations for this reactivity.
In tr od u ction
Since the seminal work on the allylation reaction
between allyltrimethylsilane and R,â-unsaturated ke-
tones was disclosed by Hosomi and Sakurai,2 there has
been a great deal published expanding the scope and
utility of allylsilanes in such reactions in synthesis.3 This
methodology has been expanded to include ring formation
when the allylsilanes involved possess “bulky groups” on
silicon to provide cyclobutane and cyclopentane products
via net [2 + 2]4 and [3 + 2]5,6 annulation, respectively.
The general addition process for allylsilanes onto unsat-
This report discusses the reactivity of allylsilanes and
the mechanism of addition of these compounds to R,â-
unsaturated ester and carbonyl substrates. In our syn-
thetic program, we are interested in exploring the use of
allylsilanes in multireaction sequences that involve cy-
cloaddition chemistry.1 The formation of four-membered
rings using allylsilane precursors via photocycloaddition
or Lewis acid catalysis, followed by ring cleavage or
expansion, is a useful approach to compounds of consid-
erable complexity. Our approach to these cyclobutane
structures has uncovered information concerning the
reactivity of a variety of allylsilanes in such additions
and the kinetic vs thermodynamic stability of the cy-
cloadducts which result.
(2) Hosomi, A.; Sakurai, H. J . Am. Chem. Soc. 1977, 99, 1673-1675.
(3) For general reviews on the uses of silicon in organic synthesis,
see: (a) Fleming, I. Tildon Lecture. Some uses of silicon compounds
in organic synthesis. Chem. Soc. Rev. 1981, 10, 83-111. (b) Parnes, Z.
N.; Bolestova, G. I. Synthesis 1984, 991-1008. (c) Blumenkopf, T. A.;
Overman, L. E. Chem. Rev. 1986, 86, 857-873. (d) Schinzer, D.
Synthesis 1988, 263-273. (e) Fleming, I.; Dunogues, J .; Smithers, R.
Org. React. 1989, 31, 57-575. (f) Chan, T. H.; Wang, D.; Pellon, P.;
Lamothe, S.; Wie, Z. Y.; Li, L. H.; Chen, L. M. Enantioselective
synthesis using silicon compounds. Frontiers of Organosilicon Chem-
istry; The Royal Society of Chemistry: London, 1991; pp 344-355. (g)
Colvin, E. W.; Loreto, M. A.; Montieth, M.; Tommasini, I. Frontiers of
Organosilicon Chemistry; The Royal Society of Chemistry: London,
1991; pp 356.
*To whom correspondence should be addressed. Phone: (416) 736-
5313. Fax: (416) 736-5936. E-mail: organ@yorku.ca.
† York University.
‡ University of Guelph.
(1) (a) Organ, M. G.; Winkle, D. J . Org. Chem. 1997, 62, 1881-1885.
(b) Organ, M. G.; Winkle, D. D.; Huffmann, J . J . Org. Chem. 1997, 62,
5254-5266.
10.1021/jo991177i CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/20/2000

Sakurai Addition and Ring Annulation of Allylsilanes
J . Org. Chem., Vol. 65, No. 12, 2000 3667
in 1994.10 They were able to show that when “bulky
groups”, i.e., isopropyl or phenyl, were placed on the
silicon of an allylsilane, the reactivity of the allyl moiety
toward conjugate addition to R,â-unsaturated esters (X
) OCH3; TiCl4 catalyst) appeared to be enhanced and
ring annulation adducts were isolated. Further, closure
of the titanium enolate onto the siliranium ion of 3 (in
their study) provided cyclobutane adducts 5 (syn and anti
with respect to silylmethyl and carbonyl moieties) as the
major products, with minor amounts of the anti cyclo-
pentanoid 6.
Sch em e 1
That allyltriisopropylsilane and allyltriphenylsilane
added to the R,â-unsaturated esters in Kno¨lker’s study10
(100% combined cycloadduct yield reported for the addi-
tion of triisopropylallylsilane to methyl acrylate, 48% for
allyltriphenylsilane) while allyltrimethylsilane failed to
add at all to the same electrophile2 is puzzling! The
ability of silicon to stabilize a developing â carbocation
is believed to be responsible for the rate of such reac-
tions.11,12 Presumably, this stabilization would be mainly
an electronic effect. Thus, electron-donating ligands on
silicon such as methyl, have an accelerating effect on the
rate, while electron-withdrawing groups, such as chlo-
rine, decrease the rate. It is well documented that large
alkyl groups on silicon have an even greater accelerating
effect on the overall rate of allylsilane substitution/
addition.13 Mayr and co-workers studied the addition of
a large number of allyltrialkylsilanes to p-anisylphenyl-
carbenium tetrachloroborate and their results show a
definite correlation between increasing size of the alkyl
group on silicon and the corresponding second-order rate
constant. However, if one accepts that the increase in rate
is the sole result of an electronic change in the silicon
ligands,14 it is not clear how the rate constants in Mayr’s
study could increase from 313 for triethylallylsilane to
507 for tri-n-butylallylsilane and increase yet again for
trihexylallylsilane (542). Presumably, electronic stabili-
zation drops off appreciably beyond two carbons on each
ligand on silicon. Further, from this same study the rate
urated substrates is illustrated in Scheme 1. Siliranium
ion intermediate 3 is central to all possible products of
initial conjugate addition. In the case of enone substrates
(X ) alkyl), if Y is large (e.g., isopropyl) the principal
product isolated is the five-membered ring adduct 6
which appears to be both the thermodynamic and kinetic
product.7 When Y is methyl, the allylation product 4, i.e.,
the Sakurai product, is the major adduct isolated.
However, there has been one recent report detailing the
formation of a (trimethylsilyl)methyl-substituted cyclobu-
tane adduct, in addition to some of the trimethylsilyl-
substituted cyclopentanoid, from the reaction of naph-
thaquinone and allyltrimethylsilane in the presence of
the mild Lewis acid Me2AlCl.8
Since the release of Sakurai’s paper (1977),2 it has been
generally accepted that allyltrimethylsilane is unreactive
in such additions to analogous ester substrates.9 There-
fore, it was believed that allylsilanes in general were
unreactive in additions to R,â-unsaturated esters until
the work of Kno¨lker and co-workers was communicated
(4) For approaches to four-membered rings using allylsilanes, see:
(a) Hartman, G. D.; Traylor, T. G. Tetrahedron Lett. 1975, 16, 939-
942. (b) Ochiai, M.; Arimoto, M.; Fujita, E. J . Chem. Soc., Chem.
Commun. 1981, 460-461. (c) Paquette, L. A.; Valpey, R. S.; Annis, G.
D. J . Org. Chem. 1984, 49, 1317-1319. (d) Fujiwara, T.; Suda, A.;
Takeda, T. Chem. Lett. 1992, 1631-1634. (e) Hojo, M.; Tomita, K.;
Hirohara, Y.; Hosomi, A. Tetrahedron Lett. 1993, 34, 8123-8126. (f)
Akiyama, T.; Kirino, M. Chem. Lett. 1995, 723-724.
(5) For approaches to five-membered rings carbocycles using allyl-
silanes, see: (a) Danheiser, R. L.; Dixon, B. R.; Gleason, R. W. J . Org.
Chem. 1992, 57, 6094-6097. (b) Hojo, M.; Tomita, K.; Hirohara, Y.;
Hosomi, A. Tetrahedron Lett. 1993, 34, 8123-8126. (c) Danheiser, R.
L.; Takahashi, T.; Berto´k, B.; Dixon, B. R. Tetrahedron Lett. 1993, 34,
3845-3848. (d) Kno¨lker, H.-J .; Graf, R. Tetrahedron Lett. 1993, 34,
4765-4768. (e) Kiyooka, S.; Shiomi, Y.; Kira, H.; Kaneko, Y.; Tanimori,
S. J . Org. Chem. 1994, 59, 1958-1960. (f) Brengel, G. P.; Rithner, C.;
Meyers, A. I. J . Org. Chem. 1994, 59, 5144-5146. (g) Takahashi, Y.;
Tanino, K.; Kuwajima, I. Tetrahedron Lett. 1996, 37, 5943-5946. (h)
Akiyama, T.; Hoshi, E.; Fujiyoshi, S. J . Chem. Soc., Perkin Trans. 1
1998, 2121-2122. (i) Akiyama, T.; Yamanaka, M. Tetrahedron Lett.
1998, 39, 7885-7888.
(10) Kno¨lker, H.-J .; Baum, G.; Graf, R. Angew. Chem., Int. Ed. Engl.
1994, 33, 1612-1615.
(11) For discussions regarding silyl-stabilized â-carbocations, see:
(a) Hanstein, W.; Berwin, H. J .; Traylor, T. G. J . Am. Chem. Soc. 1970,
92, 829-836. (b) Eaborn, C. J . Chem. Soc., Chem. Commun. 1972,
1255. (c) Fleming, I. Frontier Orbitals and Organic Chemical Reactions;
Wiley: London, 1976; p. 81 (d) Wierschke, S. G.; Chandrasekhar, J .;
J orgensen, W. L. J . Am. Chem. Soc. 1985, 107, 1496-1500. (e)
Lambert, J . B.; Wang, G.-T.; Finzel, R. B.; Teramura, D. H. J . Am.
Chem. Soc. 1987, 109, 7838-7845. (f) Lambert, J . B. Tetrahedron 1990,
46, 2677-2689. (g) Lambert, J . B.; Chelius, E. C. J . Am. Chem. Soc.
1990, 112, 8120-8126. (h) Green, A. J .; Kuan, Y.-L.; White, J . M. J .
Org. Chem. 1995, 60, 2734-2738. (i) Gabelica, V.; Kresge, A. J . J . Am.
Chem. Soc. 1996, 118, 3838-3841. (j) Chan, V. Y.; Clark, C. I.;
Giordano, J .; Green, A. J .; Karalis, A.; White, J . M. J . Org. Chem. 1996,
61, 5227-5233.
(12) For examinations of the â-effect in relation to different groups
on silicon, see: (a) Brook, M. A.; Hadi, M. A.; Neuy, A. J . Chem. Soc.,
Chem. Commun. 1989, 957-958. (b) Brook, M. A.; Neuy, A. J . Org.
Chem. 1990, 55, 3609-3616. (c) Dallaire, C.; Brook, M. A. Organome-
tallics 1990, 9, 2873-2874. (d) McGibbon, G. A.; Brook, M. A.; Terlouw,
J . K. J . Chem. Soc., Chem. Commun. 1992, 360-362. (e) Dallaire, C.;
Brook, M. A. Organometallics 1993, 12, 2332-2338. (f) Brook, M. A.;
Henry, C. Tetrahedron 1996, 52, 861-868.
(6) For approaches to substituted furans using allylsilanes, see: (a)
Panek, J . S.; Yang, M. J . Am. Chem. Soc. 1991, 113, 9868-9870. (b)
Panek, J . S.; Beresis, R. J . Org. Chem. 1993, 58, 809-811. (c) Panek,
J . S.; J ain, N. F. J . Org. Chem. 1993, 58, 2345-2348. (d) Akiyama, T.;
Yasusa, T.; Ishikawa, K.; Ozaki, S. Tetrahedron Lett. 1994, 35, 8401-
8404.
(7) (a) Kno¨lker, H.-J .; J ones, P. G.; Pannek, J .-B. Synlett 1990, 429-
430. (b) Kno¨lker, H.-J .; J ones, P. G.; Pannek, J .-B.; Weinkauf, A. Synlett
1991, 147-150. (c) Kno¨lker, H.-J .; Foitzik, N.; Graf, R.; Pannek, J .-B.
Tetrahedron 1993, 49, 9955-9972. (d) Kno¨lker, H.-J .; Foitzik, N.;
Goesmann, H.; Graf, R. Angew. Chem., Int. Ed. Engl. 1993, 32, 1081-
1083.
(8) Murphy, W. S.; Neville, D. Tetrahedron Lett. 1997, 38, 7933-
7936.
(9) Allyltrimethylsilane has been successfully added to methyl
propiolate, a much more reactive electrophile: Monti, H.; Audran, G.;
Le´andri, G.; Monti, J .-P. Tetrahedron Lett. 1994, 35, 3073-3076.
(13) (a) Mayr, H.; Hagen, G. J . Chem. Soc., Chem. Commun. 1989,
91-92. (b) Mayr, H.; Bartl, J .; Hagen, G. Angew. Chem., Int. Ed. Engl.
1992, 31, 1613-1615. (c) Mayr, H.; Patz, M. Angew. Chem., Int. Ed.
Engl. 1994, 33, 938-957. (d) Mayr, H.; Gorath, G. J . Am. Chem. Soc.
1995, 117, 7862-7868.
(14) The issue of electronic vs steric activation of the olefin of
allylsilanes where silicon is substituted by large alkyl groups has been
investigated by a number of groups. For studies which try to compart-
mentalize these two effects, see: (a) Hagen, G.; Mayr, H. J . Am. Chem.
Soc. 1991, 113, 4954-4961. (b) Panek, J . S.; Prock, A.; Eriks, K.;
Giering, W. P. Organometallics 1990, 9, 2175-2176.

3668 J . Org. Chem., Vol. 65, No. 12, 2000
Organ et al.
Sch em e 2
Ta ble 1. Rea ction Con d ition s for TiCl4-Ca ta lyzed Ad d ition of Allyltr im eth ylsila n e to Meth yl Acr yla te
no. of equiv
product ratios
entry
7
8a
TiCl4
solvent
temp (°C)
time (h)
combined yield
10a -syn
10a -anti
11a
13
1
2
3
4
5
6
1
1
1
1
1
1
1.1
1.1
1.1
1.1
1.2
1.2
1
CH2Cl2
CH2Cl2
CHCl3
CH2Cl2
CH2Cl2
CH2Cl2
-78 to 25
5.5
19
14.5
14.5
5.5
74
60
17
74
74
74
2.5
1.1
2
1.8
2.6
2.8
1.7
0.1
1
0.9
0.9
1.3
1.3
1.2
0.9
0.9
1.7
1.7
1
1
1
1
1
1
1.1
1.1
1.1
2
40
65
25
25
25
1.5
5.5
constant for allyltriphenylsilane was reported to be 3.21,
whereas allyltrimethylsilane was 197. The rate constant
for allyltriisopropylsilane (439) relative to allyltrimeth-
ylsilane is consistent for the apparent reactivity seen in
Kno¨lker’s study,10 but the rate constant for allyltriph-
enylsilane is inconsistent with the fact that allyltrim-
ethylsilane is unreactive in such additions to R,â-
unsaturated esters. The electron-deficient phenyl groups
on silicon should slow the rate because they destabilize
the positive charge forming â to silicon. With this in
mind, it does not seem logical that allyltrimethylsilane
would not be reactive in such additions.
A number of fundamental aspects of allylsilane reac-
tivity have come into question, and we address them
herein. We reinvestigated Sakurai’s study regarding the
reactivity of allyltrimethylsilane toward addition to
methyl acrylate. This single experiment led into a broad
study of the factors that determine allylsilane reactivity
and how this effects the formation of adducts in reactions
with R,â-unsaturated ketone and ester electrophiles. We
have synthesized some new allylsilanes to compare
reactivities with that of the other derivatives discussed
above. All of these studies provide product ratios which,
although perhaps indicative, do not give any indication
of relative reactivity; thus we have also developed a
computational model to evaluate allylsilane reactivity
and product stability.
temperature) at which time a homogeneous yellow solu-
tion was observed and product spots appeared on TLC.
This experiment was also conducted in an NMR tube and
1
the reaction’s progress monitored by H NMR spectros-
1
copy. After 1 h at 25 °C, the only products seen in the H
NMR spectrum were 10a and 13,15 in addition to starting
materials. Interestingly, syn cyclobutane adducts are the
major products derived from ester substrates despite
being less stable than the corresponding anti adducts.16
Compound 13 could be formed by a Claisen-type conden-
sation of 9a onto 12, presumably following silicon loss,
which implies that there must be an internal source of
H
+ to protonate the Ti-enolate of 9a . It is possible that
only the trace of acid present in TiCl4 is necessary to
initiate this side reaction as the acidic proton on 13 could
serve as the proton source once formed. This condensa-
tion must be a very rapid process because we do not see
12 in the NMR spectra of the reaction mixture. After 4
h, the first noticeable amount of 11a began to form.
Structural elucidation was carried out by comparison of
1H and 13C NMR data with the analogous triisopropylsilyl
adducts prepared in this study (vide infra) and by
Kno¨lker who confirmed their connectivity by X-ray
analysis.10 The NMR data are quite diagnostic for all syn
and anti cyclobutane and anti cyclopentane structures
in this series.
We also performed a number of experiments where
reaction conditions were varied. TiCl4 seems to be neces-
sary for this reaction to proceed because analogous
experiments under SnCl4 and AlCl3 catalysis only re-
turned starting materials. Increasing the reaction time
had the general effect of increasing the percentage of 11a
and 13 relative to the cyclobutane adducts 10a (see Table
Resu lts a n d Discu ssion
Following Sakurai’s conditions,2 methyl acrylate was
stirred at -20 °C in CH2Cl2 with freshly distilled TiCl4
(1.0 equiv), the solution further cooled to -78 °C, and
the silane (1.1 equiv) added (see Scheme 2 and Table 1,
entry 1). The initial Lewis acid/ester complex is deep
burgundy in color and upon cooling to -78 °C came out
of solution as a yellowish precipitate. No reactivity was
observed until the suspension was warmed to RT (room
(15) Compound 13 has been prepared previously by another route
and the spectral characterization corresponded well; see: Nakahara,
Y.; Fujita, A.; Ogawa, T. Agric. Biol. Chem. 1987, 51, 1009-1015.
(16) This study; see the Theoretical Considerations section and also
ref 10.

Sakurai Addition and Ring Annulation of Allylsilanes
J . Org. Chem., Vol. 65, No. 12, 2000 3669
1, entry 4). Further, heating the reaction also had the
same effect (see Table 1, entry 2). Compounds 9a and
10a exist in a readily reversible equilibrium, and 10a is
clearly the kinetic product of this reaction. Ring strain
present in 10a is responsible for the facile fragmentation
of the cyclobutane which is not present in the case of the
cyclopentane; thus, formation of 11a is likely nonrevers-
ible. Further, 12 reacts rapidly when formed providing
13 which also serves as another sink in this reaction.
Therefore, time and heat simply promote the formation
of the thermodynamic products 11a and 13. Finally, we
dissolved 10a (both isomers) in CH2Cl2 and added 1 equiv
of TiCl4 and observed quantitative conversion of 10a to
11a and 13 in approximately equal proportions which
further supports this idea.
The fact that the cyclobutane adducts could be com-
pletely isomerized to the cyclopentanoid (in addition to
the formation of 13) prompted us to further investigate
the addition of allyltriisopropylsilane to methyl acrylate.
Kno¨lker et al. performed this transformation under two
sets of reaction conditions.10 In the first case, the tem-
perature was kept at or below 0 °C for 19 h, and the
second reaction was performed at 40 °C for 3 h. The ratios
of 10b-syn:10b-anti:11b were 11.9:9.5:1 and 6.3:1.7:1,
respectively. As with the allyltrimethylsilane addition in
our study, the syn cyclobutane adduct dominates under
both sets of conditions with methyl acrylate electrophile.
Further, higher temperature, more so than time, in-
creases the percentage of the cyclopentane adduct which
is presumably the thermodynamic product. Compounds
7 (1.0 equiv) and 8b (1.2 equiv) were loaded into an NMR
tube along with CD2Cl2 (0.5 mL, 0.34 M in 7), the solution
was cooled to 0 °C, and 2 equiv of TiCl4 was added (see
Scheme 2 for structures). The mixture was then warmed
to 40 °C and the reaction’s progress followed by 1H NMR
spectroscopy (see Figure 1). Formation of adducts 10b is
considerably more rapid than with allyltrimethylsilane
addition. The spectrum taken at 3 h had a ratio of 6:1:1
for 10b0syn:10b-anti:11b, which is consistent with Kno¨lk-
er’s result at 3 h. The remaining spectra demonstrate
smooth conversion of 10b to 11b over the span of 24 h.
In an analogous experiment, 10b was treated with 2
equiv of TiCl4 in CD2Cl2 and the conversion to 11b was
quantitative as well. The bulky nature of the isopropyl
groups prevents nucleophilic displacement of the silyl
moiety at the siliranium ion stage (i.e., 9b) which would
lead to the product of Sakurai addition (i.e. 12).
F igu r e 1. NMR spectra of the reaction between methyl
acrylate (7, 1.0 equiv) and triisopropylallylsilane (8b, 1.2 equiv)
in the presence of TiCl4 (2.0 equiv) in CD2Cl2. The reagents
were combined at 0 °C, the tube was warmed to 40 °C, and
the reaction’s progress was followed over 18 h. Signals
indicated are those of the methyl ester which have been shifted
downfield of 4 ppm due to coordination to the titanium
catalyst. Key: a ) syn-cyclobutane adduct; b ) anti-cyclobu-
tane adduct; c ) anti-cyclopentane adduct; d ) methyl acry-
late.
oxidized easily using Tamao conditions,18 the isolated
yield of 11a is modest, largely due to the formation of
side product 13. We reasoned that we could compromise
and design an allylsilane that was suitably reactive in
the ring annulation reaction and that the resultant silane
product could be readily oxidized. We hypothesized that
dihexylphenylallylsilane would be a reactive partner in
ring annulation and the phenyl group would make
oxidation possible.19 Although the phenyl group would
be expected to decrease reactivity, we reasoned that the
two hexyl substituents would compensate on the basis
of Mayr’s rate studies which showed that trihexylallyl-
silane had the largest second-order rate constant in his
addition studies.13
These results indicated that our proposed ring expan-
sion strategy to provide cyclopentane-containing natural
products from cyclobutane precursors could well be
successful. Allyltrimethylsilane does in fact add to methyl
acrylate via stepwise ring annulation. The overall yield
and apparent reactivity of allyltrimethylsilane in this
reaction were consistent with what we would have
expected. That is, general reactivity and isolated yields
of reactions employing this silane with methyl acrylate
lie between that of allyltriisopropylsilane and allyltriph-
enylsilane as reported by Kno¨lker.10
We began this study using trihexylallylsilane to com-
pare the reactivity of this derivative with allyltriisopro-
pylsilane. On the basis of Mayr’s studies, we expected
(17) (a) Fleming, I.; Henning, R.; Plaut, H. J . Chem. Soc., Chem.
Commun. 1984, 29-31. (b) Fleming, I.; Sanderson, E. J . Tetrahedron
Lett. 1987, 28, 4229-4232. (c) Kno¨lker, H.-J .; Wanzl, G. Synlett 1995,
378-382. (d) J ones, G. R.; Landais, Y. Tetrahedron 1996, 52, 7599-
7662.
With these results in hand, we had hope for our
synthetic studies employing unsaturated ester sub-
strates. However, there were two concerns regarding the
trimethyl and triisopropyl derivatives for synthetic work.
First, the triisopropylsilyl group on 11b cannot be
oxidized readily rendering this product of little use.17
Second, while the trimethylsilyl group on 11a can be
(18) (a) Tamao, K.; Ishida, N.; Tanaka, T.; Kumada, M. Organome-
tallics 1983, 2, 1696-1698. (b) Tamao, K.; Kakui, T.; Akita, M.;
Iwahara, T.; Kanatani, R.; Yoshida, J .; Kumada, M. Tetrahedron 1983,
39, 983-990. (c) Tamao, K.; Kumada, M. Tetrahedron Lett. 1984, 25,
321-324.
(19) J ust following the completion of the work for this manuscript,
two reports were disclosed on the oxidation of diphenyl-tert-butylsilane
and diisopropylphenylsilane; see ref 5i and the following: Kno¨lker, H.-
J .; J ones, P. G.; Wanzel, G. Synlett 1998, 613-616.

3670 J . Org. Chem., Vol. 65, No. 12, 2000
Organ et al.
Ta ble 2. Rea ction Con d ition s for TiCl4-Ca ta lyzed Ad d ition of Tr ih exyla llylsila n e to Meth yl Vin yl Keton e
no. of equiv
time (h)
yield of
entry
14
8c
TiCl4
Me2AlCl
temp (°C)
at each temp
15c (%)
1
2
3
4
5
6
7
1
1
1
1
1
1
1
2.0
1.0
0.5
2.0
2.0
2.0
2.1
2.0
1.0
1.1
1.1
1.05
1.05
1.1
1.3
0
1.2
0.15
0.15
0
-78
16.0
2.0
15.0
15.0
1.0, 0.5
1.5
33
>5
26
7
16
10
0
-70
-78
-78
-78 to -5
-78
0.2
40
5.0
Ta ble 3. Rea ction Con d ition s for TiCl4-Ca ta lyzed Ad d ition of Dih exylp h en yla llylsila n e to Meth yl Vin yl Keton e
no. of equiv
time (h)
yield of
entry
14
8d
TiCl4
Me2AlCl
temp (°C)
at each temp
15c (%)
1
2
3
4
1
1
1
1
0.9
1.0
0.5
2.1
2.0
1.0
2.0
2.0
0
0
1.1
1.3
-78 to 25
-78 to 25
-78 to 25
-78 to 25
2.0, 40.0
1.5, 23.0
5.0, 20.0
20.0, 1.0
27
28
40
60
ation.1b One possible explanation for the poor result using
trihexylallylsilane is that the rates of nucleophilic attack
of both the triisopropyl and trihexyl derivatives on the
enone are comparable but increased susceptibility of
trihexylallylsilane to protodesilylation is responsible for
the lower “observed” or “desired” reactivity with the
electrophile. Of course, another possibility is that allyl-
triisopropylsilane is simply much more reactive toward
electrophilic addition than trihexylallylsilane, despite
Mayr’s conclusions, and that this heightened reactivity
is kinetically much faster than protodesilylation. This
would lead to more product formation before protodesi-
lylation has a chance to occur. Slower addition of the
trihexyl derivative to the electrophile would lead to more
protodesilylation, which may be the favored kinetic
process with that particular silane. We are unable to
suggest which hypothesis, if either, is operative although
we do address it in the computational section (see
Theoretical Considerations).
Sch em e 3
the trihexyl derivative would be considerably more reac-
tive in additions to methyl acrylate and MVK than the
triisopropyl case, but the opposite proved to be true.
Reaction of MVK (14) with 8c was sluggish, and if the
reaction was heated, no product (15c) was isolated
whatsoever (see Scheme 3 and Table 2). The comparable
run with allyltriisopropylsilane provided 68% yield (15b).
It is instructive to note that the best product recoveries
were reported when at least 1 equiv of Me2AlCl cocatalyst
was included in the reaction mixture. We have studied
the effects of using this cocatalyst with TiCl4 during
allylsilane substitution chemistry, and it has shown to
have remarkable effects on reaction yield.1b The principal
role of Me2AlCl is believed to be that of a proton sponge
which helps to prevent protodesilylation of the allylsilane
starting material, thus allowing more of the desired
reaction to take place.20 This usually results in signifi-
cantly higher yields when Me2AlCl is used.
Dihexylphenylallylsilane (8d ) provided consistently
higher yields than the trihexyl case with MVK, even
though it is surely less reactive toward initial attack on
the enone (see Scheme 3 and Table 3). In Mayr’s study
of allylsilane rate constants, he reported that replacing
one methyl of allyltrimethylsilane with one phenyl ring
reduces the reactivity of the allyl moiety by a factor of
5.13 However, allylsilanes possessing at least one phenyl
group on silicon are noticeably less prone to protodesilyl-
The presence of an acidic proton on MVK (from the
methyl ketone, pKa 20-25) can promote enolization in
titanium’s presence thus providing an in situ source of
H+.1b Further, allylsilanes can behave as the Bro¨nsted
base resulting in protodesilylation.21 This would dramati-
cally affect product recovery from these reactions because
such a process removes both MVK and the allylsilane
from the mixture before any desired addition can take
place between them. To confirm that protodesilylation
was occurring during the reaction and not during the
workup, MVK, TiCl4, and 8c were loaded into an NMR
tube (CD2Cl2 solvent), the tube was quickly sealed, and
the reaction’s progress was followed by NMR spectros-
copy. As suspected, protodesilylation occurred to a large
extent during the reaction as evidenced by the significant
presence of propene. The identical NMR experiment with
methyl acrylate showed propene formation initially upon
addition of TiCl4, likely owing to the presence of trace
HCl in the catalyst, and this did not increase over the
course of 2 days. In a control experiment, trihexylallyl-
silane was treated with TiCl4 (1.0 equiv) in an NMR tube
and the result was the same as that which contained
(21) For reports dealing with acid-catalyzed protodesilylation, see:
(a) Coughlin, D. J .; Salomon, R. G. J . Org. Chem. 1979, 44, 3784-
3790. (b) Pennanen, S. I. Synth. Commun. 1980, 10, 373-379. (c)
Hosomi, A.; Iguchi, H.; Sasaki, J .-I.; Sakurai, H. Tetrahedron Lett.
1982, 551-554. (d) Ireland, R. E.; Varney, M. D. J . Am. Chem. Soc.
1984, 106, 3668-3670. (e) Shibasaki, M.; Fukasawa, E.; Ikegami, S.
Tetrahedron Lett. 1983, 24, 3497-3500. (f) Corey, E. J .; Letavic, M.
A. J . Am. Chem. Soc. 1994, 117, 9616-9617.
(20) (a) Snider, B. B.; Rodini, D. J .; Conn, R. S. E.; Sealfon, S. J .
Am. Chem. Soc. 1979, 101, 5285-5293. (b) Snider, B. B.; Rodini, D. J .
Tetrahedron Lett. 1980, 21, 1815-1818.

Sakurai Addition and Ring Annulation of Allylsilanes
J . Org. Chem., Vol. 65, No. 12, 2000 3671
Sch em e 4
tuted allylsilanes, both as neutral and cationic species,
to determine if the observed reactivities can be explained
on the basis of these properties. In addition, we have
studied the formation of intermediate cation 3 using
acrolein, acrylic acid, and methyl acrylate as our elec-
trophile together with trimethyl-, triethyl-, and triiso-
propylallylsilane in order to determine the effect of the
silicon ligands on the activation barrier. While solvation
effects surely must play an important role for these ions
and since ab initio calculations will not properly take
these into account, we are most interested in the trends
and not the absolute numbers. Theoretical predictions
were performed with the GAUSSIAN suite of programs.22
The geometry optimizations of the intermediate cations
were performed without symmetry constraints at the
Hartree-Fock level (HF) using the 3-21G* basis set.23
Calculations on the neutral and cationic allylsilanes were
done at both the Hartree-Fock level and the hybrid
density functional theory level, B3LYP24 using the
6-31G*25 basis set. Some calculations were also tested
using second-order perturbation theory, MP2.26 Mulliken
charges were predicted using the 3-21G* and 6-31G*
basis sets.
Figures 2 and 3 depict important geometrical features
of eight different allylsilanes. The three substituents on
the silicon range from hydrogen, to electron-withdrawing
chlorine (-I), to a variety of electron-donating alkyl
groups (+I, +R) including isopropyl and n-hexyl. The
different geometrical features and Mulliken charges are
shown for both the ground-state species (Figure 2) and
the cationic complex that would be obtained by protona-
tion of the allylsilane (Figure 3) that we are using to
model the reaction intermediate.
Considerable effort has been expended to explain how
substituents on silicon increase or decrease the nucleo-
philicity of the olefin of an allylsilane.12,14 To the extent
that ground-state structure can be used or extrapolated
to predict reactivity of any chemical species, examination
of the data in Figure 2 is interesting. Certainly such
considerations must be viewed critically for molecules
generally react from the most reactive conformation and
not the most stable one. However, if the reactivity of an
allylsilane is dependent on the extent to which the Si-
CR bond is parallel to p orbitals of the olefin as nucleo-
philic attack begins,3 these minimized structures, relative
to each other, should provide useful information for the
lowest energy conformation for all structures are in fact
methyl acrylate, i.e., some initial propene formation upon
catalyst addition only. This result supports our suspicion
that MVK is facilitating protodesilylation and that tri-
hexylallylsilane is especially sensitive to this side reac-
tion.
If that which has been proposed above is true, one
would expect that product recovery from experiments
with methyl acrylate (which lacks the acidic proton) to
be higher than for MVK, despite being a less electrophilic
substrate. Treatment of methyl acrylate (7) with 8c (2.0
equiv) in the presence of TiCl4 (1.1 equiv) at 40 °C
provided a 61% combined yield of cycloadducts 10c and
11c (Scheme 4). At the end of this reaction, which had
been heated for 22 h, some starting allylsilane was also
recovered. The analogous reaction with 8d provided a
50% yield of ring annulated products 10d and 11d . In
light of the fact that starting materials can be recovered
from this reaction, we are confident that the yields can
be increased significantly. As hypothesized, dihexylphe-
nylallylsilane provided promising recovery of ring-annu-
lation products when compared to that obtained with
triphenylallylsilane reported by Kno¨lker.10
As anticipated, methyl acrylate has provided a nicely
controllable system to study the relative reactivity of a
number of allylsilane derivatives. It is interesting to note
that the best results in all reactions attempted are with
the allyltriisopropylsilane. Further, reactions with this
silane appear to be kinetically much faster (when moni-
tored by NMR spectroscopy) than with other allylsilanes,
although the appearance of cycloadduct in the spectra is
not an accurate measure of reaction rate. This has
inspired us to take a look at some of these systems
computationally to see if any trends emerge to assist in
predicting the order of reactivity of different allylsilanes.
Further, we can validate these computational estimates
by comparing them with the experimental results ob-
tained in this study and those of other scientists who
have studied allylsilane reactivity by other methods.
(22) Frisch, M. J .; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.;
Wong, M. W.; Foresman, J . B.; J ohnson, B. G.; Schlegel, H. B.; Robb,
M. A.; Replogle, E. S.; Gomperts, R.; Andres, J . L.; Raghavachari, K.;
Binkley, J . S.; Gonzalez, C.; Martin, R. L.; Fox, D. J .; Defrees, D. J .;
Baker, J .; Stewart, J . J . P.; Pople, J . A. Gaussian92; Gaussian Inc.:
Pittsburgh, PA, 1992.
(23) (a) Binkley, J . S.; Pople, J . A.; Hehre, W. J . J . Am. Chem. Soc.
1980, 102, 939-947. (b) Gordon, M. S.; Binkley, J . S.; Pople, J . A.;
Pietro, W. J .; Hehre, W. J . J . Am. Chem. Soc. 1982, 104, 2797-2803.
(c) Pietro, W. J .; Francl, M. M.; Hehre, W. J .; Defrees, D. J .; Pople, J .
A.; Binkley, J . S. J . Am. Chem. Soc. 1982, 104, 5039-5048.
(24) (a) Becke, A. D. J . Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang,
W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (c) Miehlich, B.; Savin, A.;
Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200.
(25) (a) Ditchfield, R.; Hehre, W. J .; Pople, J . A. J . Chem. Phys. 1971,
54, 724-728. (b) Hehre, W. J .; Ditchfield, R.; Pople, J . A. J . Chem.
Phys. 1972, 56, 2257-2261. (c) Hariharan, P. C.; Pople, J . A. Mol. Phys.
1974, 27, 209-214. (d) Gordon, M. S. Chem. Phys. Lett. 1980, 76, 163-
168. (e) Hariharan, P. C.; Pople, J . A. Theor. Chim. Acta 1973, 28,
213-222.
(26) (a) Head-Gordon, M.; Pople, J . A.; Frisch, M. J . Chem. Phys.
Lett. 1988, 153, 503-506. (b) Frisch, M. J .; Head-Gordon, M.; Pople,
J . A. Chem. Phys. Lett. 1990, 166, 275-280. (c) Frisch, M. J .; Head-
Gordon, M.; Pople, J . A. Chem. Phys. Lett. 1990, 166, 281-289.
Th eor etica l Con sid er a tion s
We have undertaken theoretical studies of some of the
allylsilane systems to determine if gas-phase computa-
tional approaches can give any rationalization for the
reactivities observed. More specifically, we have exam-
ined the charges and geometries of a variety of substi-

3672 J . Org. Chem., Vol. 65, No. 12, 2000
Organ et al.
F igu r e 2. Important geometrical parameters and charges of eight allyl silanes for the ground state. Geometries were optimized
at the B3LYP/6-31G* level of theory.
close to the proposed reacting conformation. Aside from
methyl, the Si-CR-Câ-Cγ torsional angle does not vary
substantially among all other n-alkyl ligands indicating
that there are no pronounced steric effects by elongating
the chain length of the groups on silicon. These observa-
tions lend support to the “cone angle” analysis presented
by Panek and co-workers, who hypothesized that the cone
angle of bulkier silicon ligands does not vary appreciably
from nonbulky ones.14b Further, these results also sup-
port the work by Mayr’s group, who observed a steady
increase in allylsilane reactivity as the ligand chain-
length on silicon is lengthened from methyl through to
n-hexyl and that this increase in reactivity is primarily
the result of electronic considerations.13 The increased
steric requirement of the paddling isopropyl moieties
causes noticeable twisting of the Si-CR bond away from
a reacting conformation, relative to the other alkyl ligand
groups.
For the protonated structures, which are intended to
model the cation intermediate of nucleophilic attack
(Figure 3), R ) H would appear to have the best torsional
geometry for hyperconjugative stabilization (-93.5°).
However, careful inspection of the Si-CR-Câ angle
(97.4°) indicates that the Si-CR bond is also bent away
from the adjacent Câ p orbital. The largest deviants in
this regard are triisopropyl and trichloro at 102.3 and
106.6°, respectively, which suggests CR is closer to sp3
hybridized. These two results are in themselves very
interesting for they demonstrate steric and electronic
extremes that separately lead to similar structural
results. The chloro-substituted allylsilane, where Cl is
smaller than the effective van der Waals radius of a
methyl group, resists hyperconjugation of the Si-CR
bond due to the electron-withdrawing Cl ligands. The
isopropyl group, a pronounced electron donor, is too large
to allow the necessary rehybridization of CR to take place
to pull the Si-CR bond closer to the vacant p orbital on
Câ. The significantly shorter Si-CR bond lengths for R
) H (2.086 Å) and Cl (2.037 Å) and the slightly shorter
triisopropyl case (2.214 Å), relative to the other alkyl-
substituted allylsilanes, further suggests that little hy-
perconjugation of the Si-CR bond appears to be taking
place in the cation. Conversely, the Si-CR-Câ angle gets
progressively smaller and the Si-CR bond length pro-
gressively longer as the n-alkyl series progresses from
methyl through to n-hexyl. This is highly suggestive that
as the length of the ligand chain increases on Si,
hyperconjugation also increases which mirrors measured
allylsilane reactivity for this series.13 Although this
analysis nicely accounts for the increasing reactivity of
the tri-n-alkylallylsilanes, it is anomalous in its apparent
treatment of triisopropyl. If hyperconjugation is an
accurate measure of reactivity, owing from a lowered
energy intermediate (i.e., more stable), triisopropyl would
appear to be a poor candidate, yet it is clearly one of the
most reactive nucleophiles of all the allylsilanes.
In an attempt to better understand the relationship
between sterics and hyperconjugation, a series of calcula-
tions was performed whereby the ligands on the silicon
were replaced with three hydrogen atoms for the other
allylsilanes studied, and the data are listed in Table 4.
The new structures were designed to maintain the same
angular and torsional values possessed by the original
silanes. The Si-H bond lengths were fixed at the

Sakurai Addition and Ring Annulation of Allylsilanes
J . Org. Chem., Vol. 65, No. 12, 2000 3673
F igu r e 3. Important geometrical parameters and charges of eight allyl silanes for the cationic species. Geometries were optimized
at the B3LYP/6-31G* level of theory.
Ta ble 4. Effect of th e C-C-C-Si Tor sion An gle on th e En er gies of th e Gr ou n d -Sta te a n d Ch a r ged Allylsila n esa
R3 group
H
Me
Et
n-Prop
i-Prop
nBu
nHex
Cl
Torsion Angles and Energies
C-C-C-Si (-t)
DE (ground state)
C-C-C-Si (-t)
DE (cations)
-107.5
0.0
-93.5
0.0
-106.3
0.2
-95.5
1.2
-113.4
0.4
-100.7
2.1
-113.7
0.4
-100.3
2.3
-114.9
0.8
-101.4
2.4
-114.0
0.4
-100.2
2.2
-113.8
0.4
-100.1
2.4
-108.8
0.2
-95.6
0.9
Relative Energies of the HOMO (eV)
1.00 1.00 1.00
ground state
0.61
0.98
1.00
1.01
0.00
a
The energies of the eight species are in kcal/mol relative to the R ) H structure as shown in Figures 2 and 3. The B3LYP/6-31G* level
of theory was employed where the R groups on the allylsilane were replaced with hydrogens at a bond length of 1.4890 Å for cations and
1.4738 Å for neutral isomers, and all other bond, angle, and torsion values remained the same as those optimized on the full system at
the B3LYP/6-31G* level. The relative energies of the HOMO for the ground-state species at the B3LYP/6-31G* level are also shown with
the HOMO of the tri-Cl being most stable and the others listed relative to thos system in eV.
optimized bond lengths of 1.4890 and 1.4738 Å for the
neutral and cationic species (R ) H), respectively. The
relative energies (kcal/mol) at the B3LYP/6-31G* level
should tell us what the “energetic cost” is of steric effects
that force the Si-CR bond away from hyperconjugation
with Câ, assuming the most ideal Si-CR-Câ-Cγ tor-
sional value is 90° (or -90°). Since R ) H corresponds to
the minimum and represents our energy zero, the
optimized ground-state torsional angle is -107.5° and
deviations from this value will increase the energy. For
the R ) H cation, the torsion angle of -93.5° is close to
the ideal perpendicular orientation desired for hypercon-
jugation. There is a clear correlation between a torsional
angle τ varying further from 90° and an energy increase.
The torsional angle differing furthest from perpendicular
is the isopropyl case (-101.4°). Here, the deviation from
the ideal hyperconjugative geometry is 7.9° (relative to
R ) H) which leads to a cost of 2.4 kcal/mol. These
numbers clearly indicate that hyperconjugative effects
are significant and energetically stabilizing. For the
n-alkyl ligands on Si, there is a gradual increase in
energy as you proceed from trimethyl to tri-n-hexyl for
the cation that results, presumably, from a loss of
hyperconjugation. The high reactivity of allylsilanes with
larger alkyl ligands suggests that the increasing +R
effect of the longer chains overcomes poorer orbital
overlap that forces orbitals out of ideal overlap alignment
to impart electronic stability.
In addition to charges and geometries, we have also
examined the relative energies of the highest occupied
orbitals (B3LYP/6-31G*) for the ground-state isomers
(Table 4). The chlorine HOMO has the lowest energy

3674 J . Org. Chem., Vol. 65, No. 12, 2000
Organ et al.
Ta ble 5. En er getic Da ta for th e Ad d ition of Th r ee Allylsila n es (a , Tr i-Me; b, Tr i-Et; c, Tr i-i-P r op ) to th e Lith ia ted
Acr olein (15), Acr ylic Acid (16), a n d Meth yl Acr yla te (17) Sp ecies
reactants
tot. energy (Hartree)
ZPVE
∆E
species
15
16
17
a
-196.976 72
-271.455 016
-310.273 146
-521.586 815
-638.027 243
-754.470 112
43.5
47.1
C3H4LiO+
+
C3H4LiO2+
66.2
C4H6LiO2
C6H14Si
C9H20Si
122.0
180.8
238.7
b
c
C12H26Si
15 + a
-718.563 540
-718.540 799
-718.544 060
-793.041 830
-793.002 628
-793.002 686
-831.859 827
a
165.5
167.5
168.5
169.1
170.8
170.8
188.2
0.0
16.3
15.2
0.0
26.3
26.3
0.0
C9H18LiOSi+
transition state
intermediate
16 + a
C9H18LiO2Si+
transition state
intermediate
17 + a
transition state
intermediate
C
10H20LiO2Si+
-831.814 322
190.1
30.5
15 + b
-835.003 967
-834.982 853
-834.987 907
-909.482 257
-909.445 324
-909.446 200
-948.300 388
-948.257 362
-948.257 467
224.3
226.2
227.2
227.9
229.1
229.7
247.0
248.8
248.8
0.0
15.1
13.0
0.0
24.4
24.4
0.0
C
C
C
12H24LiOSi+
12H24LiO2Si+
13H26LiO2Si+
transition state
intermediate
16 + b
transition state
intermediate
17 + b
transition state
intermediate
28.8
28.7
15 + c
-951.446 837
-951.430 627
-951.430 728
-1025.925 127
-1025.888 628
-1025.889 047
-1064.743 258
-1064.706 889
-1064.707 312
282.2
285.1
285.2
285.8
287.5
287.6
304.9
b
0.0
13.1
13.1
0.0
C
C
C
15H30LiOSi+
15H30LiO2Si+
16H32LiO2Si+
transition state
intermediate
16 + c
transition state
intermediate
17 + c
transition state
intermediate
24.6
24.4
0.0
24.5c
24.4c
b
Transition state could not be located. Frequency calculations were not performed due the size of the molecules. c These relative
a
b
energies include zero point corrections estimated from the frequency analysis of 16 + c.
because its electron-withdrawing power creates a HOMO
which is electron-deficient. This makes it more difficult
to remove electrons thus lowering the energy and de-
creasing reactivity. As the n-alkyl ligands on silicon go
from methyl to hexyl, there is a slight, but gradual,
increase in the HOMO energy. This slight destabilization
could play a role in increasing the reactivity for these
systems.
indicate the transition state energies lie only between 0
and 0.2 kcal/mol above the minima indicating again that
this intermediate is weakly bound. The highly endother-
mic nature of this reaction by ca. 25 kcal/mol is helping
lead to this result. Thus, the relative reactivities can be
accurately judged thermodynamically by the energetic
height of the intermediate rather than kinetically by the
activation energies.
We have studied theoretically the addition of three
allylsilanes with acrolein, acrylic acid, and methyl acry-
late. To minimize electron count, we have used a Li cation
in place of TiCl3 cation as the Lewis acid in this study.
The energetic data for the additions to the substituted
allylsilanes, CH2dCH2-CH2-SiR3, where R ) Me, Et,
and i-Prop, are summarized in Table 5.
Although calculations are done at the HF/3-21G* level
of theory, the general trends should give us an idea of
which species are predicted to be most reactive. Figure
4 depicts important geometrical features and charges of
the transition states for the addition of trimethyl-,
triethyl-, and triisopropylallylsilane to acrylic acid in the
presence of Li cation. By examination of the structural
and energetic trends, the effect of the alkyl substitution
of the allylsilane and the acrolein group (H, OH, OCH3)
can be determined.
There are some interesting trends which can be
observed in the energetics from Table 5. First, as the R
group is changed in the allylsilane, there are slight
changes in the transition state barriers. On going from
R ) Me, Et, and i-Prop, the activation barrier typically
decreases for any given reaction. For example, for the
addition of the three allylsilanes to lithiated acrolein, the
barrier (kcal/mol) goes from 16.3 (R ) Me) to 15.1 (R )
Et) to 13.1 (R ) i-Prop). These energy differences
qualitatively support the rate increase reported in Mayr’s
work of 67% on going from R ) Me to R ) Et and 40%
from R ) Et to R ) iProp.13 It is interesting that this
series of calculations indeed predict a decrease in activa-
tion energy when the substitution on Si changes from
Me to Et to i-Prop in agreement with experimental data
in this report and others,13 although these substituents
are some distance from the reacting center.
For the addition of the three allylsilanes to the lithiated
acrylic acid species, the forming C-C bond in the three
transition state structures has lengths of 1.748 (16a ),
1.799 (16b), and 1.786 (16c) Å (see Figure 4 for repre-
sentative structures). The bond which has formed in the
intermediate is extremely long at distances of 1.686,
1.644, and 1.661 Å, respectively, indicative of a very
weakly bound intermediate. The energetics in Table 5
This last series of calculations clearly shows the
potential danger of relying on just one computational
model. The reaction models demonstrate that triisopro-
pylallylsilane has energy differences in both the TS
barriers and intermediates that are consistent with the
apparent high reactivity demonstrated by this reagent
experimentally. However, while the data in Figures 2 and
3 and Table 4 provide a nicely explainable trend to follow

Sakurai Addition and Ring Annulation of Allylsilanes
J . Org. Chem., Vol. 65, No. 12, 2000 3675
F igu r e 4. Important geometrical features and Mulliken charges for the transition state structures for the addition of the lithiated
acid (16) to three allyl silanes, the Me (a ), Et (b), and i-Prop (c) substituted systems. Geometries and charges came from the
HF/3-21G* level of theory.
for the n-alkyl substituted silanes, as well for H and Cl,
isopropyl always sat outside of the trends. These data
would have indicated lower reactivity for isopropyl than
is observed experimentally. We believe that this combi-
nation of computational models has provided a useful
method for analyzing allylsilane reactivity, and we are
continuing our studies on this system.
close onto the carbocation, thereby maintaining the
silicon-containing moiety in the structure, giving rise to
either (10a ) or five-membered-ring (11a ) annulation
products. Products 11a and 12 (which self-condenses
when formed) are thermodynamic sinks and do not react
further. The ring strain in 10a makes it susceptible to
fragmentation in the presence of TiCl4, and equilibration
takes place eventually converting all of 10a to 11a and
12.
Con clu sion s
When larger ligands on Si are employed, such as
triisopropyl- or di-n-hexylphenylallylsilane, ring annu-
lation similarly occurs with MVK to yield the five-
membered-ring product exclusively with no observed
Trimethylallylsilane adds effectively to methyl acrylate
in the presence of TiCl4. The intermediate siliranium ion
thus produced (9a ) can undergo desilylation to give the
Sakurai addition product (12) or the titanium enolate can

3676 J . Org. Chem., Vol. 65, No. 12, 2000
Organ et al.
used over a maximum period of 3 weeks at which time it would
be redistilled.
Sakurai product. However, when tri-n-hexylallylsilane
was reacted under identical conditions, protodesilylation
of the starting material occurred to a major extent in the
presence of MVK and TiCl4 and we believe this to be the
result of heightened sensitivity of this reagent relative
to other silanes used, but the origin of this sensitivity is
not understood. When all of the above silanes were
reacted with methyl acrylate, all added smoothly, the
four-membered ring products dominated kinetically, and
all equilibrated to the more stable 5-membered ring
product over time.
A series of ab initio gas phase predictions has been
made. The parent allylsilane with different substituents
on the silicon was examined, and then the reaction
between the allylsilane and different R,â-unsaturated
carbonyl systems was investigated. It was determined
that the triisopropyl group is anomalous in charge and
geometrical features compared with other alkyl substi-
tuted systems. The bulky i-Prop groups lead to a CdC-
C-Si torsion angle differing from the H-substituted
species by about 8° in both the ground state and cation.
Steric effects appear to be important since the Si-CR-
Câ angle of 116.9° in the ground state structure is also
noticibly larger than the other seven species which have
a small range of 112.2-115.8°.
The charge on the silicon atom for the i-Prop system
is the largest at +0.80 and +0.77 on the ground-state
silane and cationic species, respectively. This result
appears to be consistent with the high reactivity observed
with this silane experimentally but appears to contradict
other computational results. These charge data would
suggest that hyperconjugation occurs to the largest
extent with triisopropylallylsilane which should place the
maximum amount of positive charge on the Si atom.
However, the orbital overlap analysis with this very
electron-rich silane appears to be the worst of all the
silanes studied, despite the charge on Si. This is sup-
ported by the torsional angle analysis discussed above,
the fact that CR for the cationic triisopropyl case (Figure
3) is much closer to sp3 hybridized than the other alkyl
ligands, and the Si-CR bond is comparatively short
indicating higher bond order.
For the reaction of the allylsilanes with the unsatur-
ated aldehyde, acid, and ester, two different trends were
observed from the energetic data. First, as the size alkyl
group of the allylsilane is increased, there is a decrease
in the activation barrier, in agreement with experiment.
In addition, on changing from methyl acrylate to acrolein,
the reactivity of the system is predicted to dramatically
increase which also agrees with experimental data.
Clearly, more study is required here to probe the origin
of reactivity with allylsilanes so that useful predictions
can be made about the practical synthetic application of
this very useful family of nucleophiles.
Allyl-tr i-n -h exylsila n e (8c). To a solution of chloro-tri-n-
hexylsilane (500 mg, 574 mL, 1.57 mmol) in dry THF (3 mL)
at 0 °C was added allylmagnesium bromide (2.56 mL, 1.10 M
in ether, 2.82 mmol). After being stirred for 1 h, the reaction
was quenched by the addition of saturated NaHCO3. The
mixture was diluted with diethyl ether and the organic layer
separated. The aqueous layer was extracted 3× with ether,
and the combined organic layers were dried over anhydrous
MgSO4. Following solvent removal in vacuo, the crude product
was purified by flash chromatography (straight hexanes) to
provide 514.0 mg (99.1%) of 8c as a clear oil: 1H NMR (CDCl3,
300 MHz) δ 5.86-5.68 (m, 1H), 4.82 (d, J ) 15.4 Hz, 1H), 4.78
(d, J ) 8.8 Hz, 1H), 1.52 (d, J ) 8.1 Hz, 2H), 1.39-1.14 (m,
24H), 0.93-0.82 (m, 9H), 0.61-0.42 (m, 6H); 13C NMR (CDCl3,
75 MHz, APT pulse sequence, evens up (+), odds down (-)) δ
135.53 (-), 112.43 (+), 33.55 (+), 31.56 (+), 23.75 (+), 22.65
(+), 20.57 (+), 14.12 (-), 12.17 (+); IR (neat) 3078 cm-1; HRMS
calcd for [C21H44Si - (C3H5)]+ 283.2823, found 283.2830.
Allyl-d i-n -h exylp h en ylsila n e (8d ). To a suspension of
magnesium turnings (67.7 mg, 2.78 mmol) in dry THF (5 mL)
was added bromobenzene (205 µL, 306 mg, 1.95 mmol). The
solution was refluxed 1 h, cooled to 0 °C, and then transferred
by cannula to a solution of dichloro di-n-hexylsilane (520 mL,
500 mg, 1.86 mmol) in dry THF (4 mL) at 0 °C. The solution
was allowed to warm to RT overnight and then heated to 40
°C for 45 min. After cooling of the solution to 0 °C, allylmag-
nesium bromide (4.31 mL, 0.56 M in hexanes, 2.41 mmol) was
added dropwise. After warming of the mixture gradually to
room temperature, saturated NH4Cl was added to quench the
reaction. The mixture was diluted with diethyl ether and the
organic layer separated. The aqueous layer was extracted 3×
with ether, and the combined organic layers were dried over
anhydrous MgSO4. Following solvent removal in vacuo, the
crude product was purified by flash chromatography (straight
pentane) to provide 307.3 mg (52.1%) of 8d as a clear oil: 1H
NMR (CDCl3, 300 MHz) δ 7.61-7.52 (m, 2H), 7.45-7.37 (m,
3H), 5.96-5.79 (m, 1H), 4.96 (d, J ) 15.5 Hz, 1H), 4.92 (d, J
) 11.0 Hz, 1H), 1.90 (d, J ) 8.1 Hz, 2H), 1.49-1.26 (m, 16H),
0.92-0.89 (m, 10H); 13C NMR (CDCl3, 75 MHz, APT pulse
sequence, evens up (+), odds down (-)) δ 137.16 (+), 134.76
(-), 134.10 (-), 128.82 (-), 127.64 (-), 113.35 (+), 33.42 (+),
31.47 (+), 23.59 (+), 22.62 (+), 20.49 (+), 14.11 (-), 12.09 (+);
IR (neat) 3071 cm-1; HRMS calcd for [C21H36Si - (C3H5)]+
275.2197, found 275.2202.
Rea ction of Meth yl Acr yla te w ith Tr im eth yla llylsila n e
(10a -syn a n d -a n ti, 11a , a n d 13). Into an NMR tube was
added 0.5 mL of distilled CD2Cl2 followed by TiCl4 (20.1 µL,
34.8 mg, 0.18 mmol, 1.1 equiv) and the solution cooled to -78
°C. To this was then added methyl acrylate (15.0 µL, 14.3 mg,
0.17 mmol) followed by trimethylallylsilane (26.5 µL, 19.0 mg,
0.17 mmol) and the temperature taken to 40 °C in an oil bath.
The tube was cooled with a stream of air above the level of
the liquid inside the NMR tube. The progress of the reaction
was monitored periodically by 1H NMR spectroscopy. After 19
h, the reaction was then cooled to RT and the contents of the
tube were poured into a flask containing saturated NH4Cl. The
mixture was diluted with diethyl ether and the organic layer
separated. The aqueous layer was extracted 3× with ether,
and the combined organic layers were dried over anhydrous
MgSO4. Following solvent removal in vacuo, the crude product
was purified by flash chromatography (2% ether in pentane)
to provide 20.0 mg (60.0%) of adducts 10a and 11a as pale
yellow oils. All ring annulation products were pooled together
after flash chromatography and the ratio of products deter-
mined by 1H NMR spectroscopy. The ratio of 10a -syn:10a -anti:
11a :13 was 1.1:0.1:1.2:1. C and H analysis was performed on
the mixture to confirm yield. Anal. Calcd for C10H20O2Si: C,
59.95; H, 10.06. Found: C, 59.76; H, 10.20. Compound 13 is
known, and spectra conform with the literature.15
Exp er im en ta l Section
Gen er a l P r oced u r e. All reactions were carried out under
a positive atmosphere of dry argon. Solvents were distilled
prior to use: THF was distilled from sodium benzophenone;
CH2Cl2 was distilled from CaH2. All TiCl4 used in these studies
was drawn from a sample that had been distilled and stored
in an air-free storage flask equipped with a glass 14/20 joint
(sealed with fluorine-based grease) through which needles
were inserted while the flask was under a positive pressure
of dry argon which was supplied through a side arm equipped
with a Teflon needle valve. A distilled sample of TiCl4 was
syn -Met h yl 2-((Tr im et h ylsilyl)m et h yl)cyclob u t a n e-
ca r boxyla te (10a -syn ): 1H NMR (CDCl3, 300 MHz) δ 3.59
(s, 3H), 2.70 (m, 1H), 2.05-1.00 (m, 5H), 0.88 (dd, J ) 14.5,
5.0 Hz, 1H), 0.67 (dd, J ) 14.5, 10.0 Hz, 1H), -0.10 (s, 9H);

Sakurai Addition and Ring Annulation of Allylsilanes
J . Org. Chem., Vol. 65, No. 12, 2000 3677
13C NMR (CDCl3, 75 MHz,) δ 175.31, 51.38, 47.87, 36.98, 27.87,
25.04, 21.59, -1.14.
a n ti-Met h yl 2-((Tr im et h ylsilyl)m et h yl)cyclob u t a n e-
ca r boxyla te (10a -a n ti): 1H NMR (300 MHz, CDCl3) δ 3.62
(s, 3H), 3.12 (m, 1H), 2.14 (m, 1H), 2.05-1.00 (m, 6H), -0.09
(s, 9H); 13C NMR (CDCl3, 75 MHz) δ 174.96, 51.09, 44.49,
34.78, 28.14, 20.08, 19.61, -1.08.
37.00, 34.81, 33.68, 33.60, 33.55, 33.48, 32.11, 31.71, 31.53,
31.50, 31.43, 29.69, 29.25, 28.28, 28.04, 24.02, 23.83, 23.78,
23.70, 22.60, 21.42, 20.74, 20.06, 15.43, 14.08, 12.82, 12.77,
11.75, 11.34 (one peak could not be found due to overlap). IR
(neat): 1738 cm-1. Anal. Calcd for C25H50O2Si: C, 73.10; H,
12.27. Found: C, 72.98; H, 12.19.
R ea ct ion of Met h yl Acr yla t e w it h Allyl-d i-n -h exyl-
p h en ylsila n e (10d a n d 11d ). Into a 10 mL round-bottom
flask were added 2.0 mL CH2Cl2 and 23 µL of methyl acrylate
(0.253 mmol) followed by 33 µL of freshly distilled TiCl4 (0.303
mmol) at room temperature. Allyl-di-n-hexylphenylsilane (200
mg, 0.632 mmol) was then added dropwise. The mixture was
heated to reflux for 16 h and then cooled to room temperature.
A 10 mL volume of saturated NH4Cl was then added and the
organic layer separated. The aqueous layer was then extracted
twice with diethyl ether, and the organic fractions were
combined and dried over anhydrous MgSO4. Following solvent
removal in vacuo, the crude residue was purified by flash
chromatography (straight hexanes) to yield 50.0 mg of 3
isomeric ring annulation products (49% yield). Characteriza-
tion, including C and H analysis, was performed on the
mixture of the three isomers, i.e., 10d syn- and anti-cyclobu-
tanes and 11d anti-cyclopentane. For 1H NMR the number of
protons reported has been normalized to one isomer. 1H NMR
(CDCl3, 300 MHz): δ 7.48-7.40 (m, 2H), 7.35-7.28 (m, 3H),
3.70, 3.63, 3.60 (3S, 3H), 3.12 (tt, J ) 5.2, 4.4 Hz, 1H), 2.82-
2.46 (m, 1H), 2.24-2.09 (m, 1H), 2.07-1.64 (m, 3H), 1.61-
1.40 (m, 1H), 1.38-1.15 (m, 15H), 1.03-0.65 (m, 12H). 13C
NMR (CDCl3, 75 MHz): δ (only signals for 10d -syn and 11d
were strong enough to be detected in 13C NMR spectrum, and
these signals are presented here) 177.29, 175.08, 137.57,
136.66, 134.40, 134.00, 128.74, 128.69, 127.59, 127.58, 51.54,
51.29, 47.79, 44.06, 36.58, 33.58, 33.45, 31.98, 31.55, 31.45,
31.42, 29.67, 29.14, 27.85, 23.84, 23.70, 23.67, 23.62, 22.58,
21.52, 20.76, 14.08, 12.56, 11.33. IR (neat): 3070, 1738, 1732
cm-1. Anal. Calcd for C25H42O2Si: C, 74.57; H, 10.51. Found:
C, 74.61; H, 10.65.
a n ti-Meth yl 3-(tr im eth ylsilyl)cyclopen tan ecar boxylate
(11a ): 1H NMR (CDCl3, 300 MHz) δ 3.59 (s, 3Η), 2.60 (m, 1H),
2.05-1.00 (m, 6H), 0.58 (m, 1H), -0.08 (s, 9H); 13C NMR
(CDCl3, 75 MHz) δ 177.38, 51.56, 44.37, 31.82, 31.71, 28.70,
25.79, -3.08.
Rea ction of Meth yl Acr yla te w ith Tr iisop r op yla llyl-
sila n e (10b-syn a n d -a n ti a n d 11b). Into an NMR tube was
added methyl acrylate (15.3 µL, 14.6 mg, 0.17 mmol), triiso-
propylallylsilane (53.0 µL, 44.0 mg, 0.22 mmol), TiCl4 (37.3
µL, 64.5 mg, 0.34 mmol), and 0.5 mL of distilled CD2Cl2. The
tube was warmed in an oil bath at 40 °C for 18 h while the
tube was cooled with a stream of air above the level of the
liquid in the NMR tube. The progress of the reaction was
monitored periodically by 1H NMR spectroscopy. The reaction
was then cooled to RT, and the contents of the tube were
poured into a flask containing saturated NH4Cl. The mixture
was diluted with diethyl ether and the organic layer separated.
The aqueous layer was extracted 3× with ether, and the
combined organic layers were dried over anhydrous MgSO4.
Following solvent removal in vacuo, the crude product was
purified by flash chromatography (2% ether in pentane) to
provide 31.0 mg (64.0%) of 11b as a pale yellow oil.
The above reaction was performed on the same scale in CH2-
Cl2 and halted after 4 h. It was found that the ratio of syn
and anti cyclobutane adducts to anti-cyclopentane was 6:1:1.
All ring annulation products were pooled together after flash
1
chromatography. The ratio of products was determined by H
NMR spectroscopy. Representative NMR spectra for the cy-
clobutane isomers are given below, and C and H analysis was
performed on the mixture to confirm yield. Anal. Calcd for
C
16H32O2Si: C, 67.55; H, 11.34. Found: C, 67.95; H, 11.06.
syn -Meth yl 2-((Tr iisop r op ylsilyl)m eth yl)cyclobu ta n e-
1-Acetyl-3-(tr i-n -h exylsilyl)cyclop en ta n e (15c). Dim-
ethylaluminum chloride (193 µL, 1 M in hexanes, 0.193 mmol)
and TiCl4 (42.3 µL, 0.386 mmol) were added to dry CH2Cl2 (3
mL), and the resultant solution was stirred for 1 h. After
cooling of the solution to 20 °C, MVK (16.1 µL, 0.193 mmol)
was added and the reaction cooled to -78 °C. In a separate
flask, Me2AlCl (58 mL, 1 M in hexanes, 0.058 mmol) was
dissolved in dry CH2Cl2 (4 mL), to this was added 8c (125 mg,
0.386 mmol), and the mixture was cooled to -78 °C. The silane
solution was transferred to the mixture containing MVK by
cannula. After being stirred for 17 h at -75 °C, the mixture
was allowed to warm to RT at which time it was quenched
with saturated NH4Cl. The mixture was diluted with diethyl
ether and the organic layer separated. The aqueous layer was
extracted 3× with ether, and the combined organic layers were
dried over anhydrous MgSO4. Following solvent removal in
vacuo, the crude product was purified by flash chromatography
(2% ether in pentane) to provide 25.4 mg (33.4%) of 15c as a
pale yellow oil: 1H NMR (CDCl3, 300 MHz) δ 2.85 (tt, J )
12.50, 5.15 Hz, 1H), 2.13 (s, 3H), 2.00-1.76 (m, 3H), 1.74-
1.52 (m, 2H), 1.31-1.18 (m, 24H), 1.12-0.94 (m, 2 H), 0.86 (t,
J ) 6.62 Hz, 9H), 0.53-0.42 (m, 6 H). 13C NMR (CDCl3, 75
MHz) δ 211.10, 52.89, 33.69, 31.51, 30.72, 30.51, 29.50, 28.75,
24.04, 23.83, 22.62, 14.09, 11.76; IR (neat): 1713 cm-1; HRMS
calcd for C25H50OSi 394.3633, found 394.3621. Anal. Calcd for
ca r boxyla te (10b-syn ): 1H NMR (CDCl3, 300 MHz) δ 3.66
(s, 3H), 2.59 (m, 1H), 2.50-1.60 (m, 8H), 1.00 (br s, 18H), 0.74
(m, 2H); 13C NMR (CDCl3, 75 MHz) δ 175.12, 51.24, 48.22,
37.08, 28.28, 21.12, 18.71, 17.27, 11.07.
a n ti-Meth yl 2-((Tr iisop r op ylsilyl)m eth yl)cyclobu ta n e-
ca r boxyla te (10b-a n ti): 1H NMR (CDCl3, 300 MHz) δ 3.69
(s, 3H), 3.17 (m, 1H), 2.50-1.60 (m, 8H), 1.20-0.90 (m, 2H),
1.01 (br s, 18H); 13C NMR (CDCl3, 75 MHz) δ 174.80, 50.92,
44.77, 34.65, 28.52, 19.90, 18.71, 11.94, 11.12.
a n ti-Meth yl 3-(Tr iisop r op ylsilyl)cyclop en ta n eca r box-
yla te (11b): 1H NMR (CDCl3, 300 MHz) δ 3.64 (s, 3H), 2.79
(m, 1H), 2.15-1.18 (m, 10H), 1.07 (br s, 18H); 13C NMR (CDCl3,
75 MHz) δ 177.43, 51.60, 44.04, 32.60, 31.72, 30.13, 22.43,
19.17, 11.30; IR (neat) 1722 cm-1
.
Rea ction of Meth yl Acr yla te w ith Tr ih exyla llylsila n e
(10c a n d 11c). Into a 10 mL round-bottom flask were added
3.0 mL CH2Cl2 and 34.7 µL of methyl acrylate (0.39 mmol)
followed by 47 µL of freshly distilled TiCl4 (0.43 mmol) at room
temperature. Allyl-tri-n-hexylsilane (250 mg, 0.73 mmol) was
then added dropwise. The mixture was heated to reflux for 16
h and then cooled to room temperature. A 10 mL volume of
saturated NH4Cl was then added and the organic layer
separated. The aqueous layer was then extracted twice with
diethyl ether, and the organic fractions were combined and
dried over anhydrous MgSO4. Following solvent removal in
vacuo, the crude residue was purified by flash chromatography
(1% ether in pentane) to yield 97.0 mg of three isomeric ring
annulation products (61% yield). Characterization, including
C and H analysis, was performed on the mixture of the three
isomers, i.e., 10c syn- and anti-cyclobutanes and 11c anti-
C
25H50OSi: C, 76.07; H, 12.76. Found: C, 76.14; H, 12.57.
1-Acetyl-3-(d i-n -h exylp h en ylsilyl)cyclop en ta n e (15d ).
To 8 mL of dry CH2Cl2 was added Me2AlCl (172 µL, 1 M in
hexanes, 0.172 mmol) and TiCl4 (37.7 µL, 65.3 mg, 0.344
mmol), and the solution was stirred 20 min at RT. The solution
was cooled to -20 °C and MVK (14.32 µL, 12.06 mg, 0.172
mmol) added dropwise. In a separate flask, 3 mL of dry CH2-
Cl2 was stirred with Me2AlCl (52 µL, 1 M in hexanes, 0.052
mmol) for 20 min at RT and allyl-di-n-hexylphenylsilane (130
µL, 114 mg, 0.361 mmol) added dropwise. Both flasks were
cooled to -78 °C, and the silane solution was transferred via
cannula (flask rinsed 3 × 4 mL with dry CH2Cl2) to the catalyst
mixture. After being stirred at -50 °C for 21 h, the solution
1
cyclopentane. For H NMR the number of protons reported has
been normalized to one isomer. 1H NMR (CDCl3, 300 MHz):
δ 3.67, 3.65, 3.64 (3 s, 3H), 2.82-2.44 (m, 1H), 2.15-1.37 (m,
6H), 1.36-1.11 (m, 24H), 0.86 (t, J ) 6.6 Hz, 9H), 0.74-0.56
(m, 1H), 0.54-0.34 (m, 6H). 13C NMR (CDCl3, 75 MHz): δ
177.32, 175.21, 174.85, 51.47, 51.26, 50.96, 48.02, 44.67, 44.29,

3678 J . Org. Chem., Vol. 65, No. 12, 2000
Organ et al.
was warmed to RT and quenched by the addition of saturated
NaHCO3. The organic layer was separated and the aqueous
layer extracted 3× with ether. The combined organic layers
were dried over anhydrous MgSO4. Following solvent removal
in vacuo, the crude product was purified by flash chromatog-
raphy (3% ether in hexanes) to provide 40.0 mg (60%) of 15d
as a pale yellow oil: 1H NMR (CDCl3, 300 MHz) δ 7.50-7.42
(m, 2H), 7.37-7.29 (m, 3H), 2.84-2.71 (m, 1H), 2.11 (s, 3H),
2.04-1.93 (m, 1H), 1.93-1.77 (m, 2H), 1.74-1.53 (m, 3H),
1.44-1.09 (m, 17H), 0.96-0.70 (m, 10H); 13C NMR (CDCl3, 75
MHz, APT pulse sequence, evens up (+), odds down (-)) δ
210.99 (+), 136.62 (+), 134.39 (-), 128.74 (-), 127.59 (-), 52.59
(-), 33.56 (+), 32.40 (+), 31.40 (+), 30.27 (+), 29.30 (+), 28.80
(-), 23.83 (+), 23.65 (-), 22.58 (+), 14.08 (-), 11.33 (+); IR
(neat) 3069, 1713 cm-1; HRMS calcd for C25H42OSi 386.3007,
found 386.3011. Anal. Calcd for C25H42OSi: C, 77.65; H, 11.00.
Found: C, 77.80; H, 11.50.
J O991177I