γ-silyl cyclobutyl carbocations

Article abstract of DOI:10.1021/jo901821f

(Graph Presented) A series of 3-trimethylsilyl-1-substituted cyclobutyl trifluoroacetates have been prepared and reacted in CD₃CO ₂D. Rate data indicate that the substrates with the trimethylsilyl group cis to the leaving group react with assistance due to γ-silyl participation. Rate enhancements range from a factor of 209 for α-phenyl-substituted cations to 4.6 × 10⁴ for α-methyl-substituted cations to >10¹⁰ for the unsubstituted γ-trimethylsilylcyclobutyl cation. Acetate substitution products are formed with net retention of stereochemistry. These experimental studies, as well as B3LYP/6-31G* computational studies, are consistent with the involvement of carbocations where the rear lobe of the γ-Si-C bond interacts strongly with the developing cationic center. Solvolytic rate studies, as well as computational studies, suggest that the secondary γ-trimethylsilylcyclobutyl cation is even more stable than the β-trimethylsilylcyclobutyl cation, i.e., the γ-silyl effect actually outweighs the potent β-silyl effect. Although computational studies suggest the existence of certain isomeric cations, where the front lobe of the Si-C bond interacts with the cationic center, solvolytic evidence for the involvement of these front lobe stabilized cations is less compelling.

Full text of DOI:10.1021/jo901821f

pubs.acs.org/joc
γ-Silyl Cyclobutyl Carbocations
Xavier Creary* and Elizabeth D. Kochly
Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, Indiana 46556
creary.1@nd.edu
Received August 24, 2009
A series of 3-trimethylsilyl-1-substituted cyclobutyl trifluoroacetates have been prepared and reacted
in CD3CO2D. Rate data indicate that the substrates with the trimethylsilyl group cis to the leaving
group react with assistance due to γ-silyl participation. Rate enhancements range from a factor of 209
for R-phenyl-substituted cations to 4.6 ꢀ 104 for R-methyl-substituted cations to >1010 for the
unsubstituted γ-trimethylsilylcyclobutyl cation. Acetate substitution products are formed with net
retention of stereochemistry. These experimental studies, as well as B3LYP/6-31G* computational
studies, are consistent with the involvement of carbocations where the rear lobe of the γ-Si-C bond
interacts strongly with the developing cationic center. Solvolytic rate studies, as well as computa-
tional studies, suggest that the secondary γ-trimethylsilylcyclobutyl cation is even more stable than
the β-trimethylsilylcyclobutyl cation, i.e., the γ-silyl effect actually outweighs the potent β-silyl effect.
Although computational studies suggest the existence of certain isomeric cations, where the front
lobe of the Si-C bond interacts with the cationic center, solvolytic evidence for the involvement of
these front lobe stabilized cations is less compelling.
Introduction
been observed in quantitative studies by Lambert.2 These
cations 1 are tremendously stabilized by a very favorable
interaction between the cationic center and the adjacent
carbon-silicon bond. Computational studies,3,4 as well as
a recent X-ray crystallographic study,5 also point to this
mode of stabilization, as opposed to a potential stabilization
mode involving silicon bridging as in 2. Less well studied is
the effect of a silyl group one carbon further removed from
the carbocationic center, i.e., “the γ-silyl effect”. Shiner6
The β-silyl effect on carbocations of type 1 is a well-studied
phenomenon.1 Cations 1 are formed much more readily than
unsilylated analogues and rate enhancements up to 1011 have
(1) For reviews and early studies, see: (a) Colvin, E. W. Silicon in Organic
Synthesis; Butterworths: London, UK, 1981. (b) White, J. M. Aust. J. Chem.
1995, 48, 1227. (c) Ushakav, S. N.; Itenberg, A. M. Zh. Obshch. Khim. 1937, 7,
2495. (d) Sommer, L. H.; Dorfman, E.; Goldberg, G. M.; Whitmore, F. C. J. Am.
Chem. Soc. 1946, 68, 488. (e) Sommer, L. H.; Bailey, D. L.; Whitmore, F. C. J.
Am. Chem. Soc. 1948, 70, 2869. (f) Sommer, L. H.; Baughman, G. A. J. Am.
Chem. Soc. 1961, 83, 3346. (g) Davis, D. D.; Jacocks, H. M. III J. Organomet.
Chem. 1981, 206, 33.
(2) (a) Lambert, J. B.; Wang, G.-t.; Finzel, R. B.; Teramura, D. H. J. Am.
Chem. Soc. 1987, 109, 7838. (b) Lambert, J. B.; Chelius, E. C. J. Am. Chem.
Soc. 1990, 112, 8120. (c) Lambert, J. B. Tetrahedron 1990, 46, 2677. (d)
Lambert, J. B.; Liu, X. J. Organomet. Chem. 1996, 521, 203. (e) Lambert, J.
B.; Zhao, Y.; Emblidge, R. W.; Salvador, L. A.; Liu, X.; So, J. H.; Chelius, E.
C. Acc. Chem. Res. 1999, 32, 183.
(3) (a) Wierschke, S. G.; Chandrasekhar, J.; Jorgensen, W. L. J. Am.
Chem. Soc. 1985, 107, 1496. (b) Ibrahim, M. R.; Jorgensen, W. L. J. Am.
Chem. Soc. 1989, 111, 819.
(4) Creary, X.; Kochly, E. D. J. Org. Chem. 2009, 74, 2134.
€
(5) Klaer, A.; Saak, W.; Haase, D.; Muller, T. J. Am. Chem. Soc. 2008,
130, 14956.
(6) Shiner, V. J. Jr.; Ensinger, M. W. J. Am. Chem. Soc. 1986, 108, 842.
9044 J. Org. Chem. 2009, 74, 9044–9053
Published on Web 11/02/2009
DOI: 10.1021/jo901821f
r
2009 American Chemical Society

Creary and Kochly
JOCArticle
examined this phenomenon in a study of brosylate 5, which
was found to solvolyze 452 times faster than the unsilylated
analogue 3 in 97% trifluoroethanol. A negligible rate en-
hancement was observed for the trans-brosylate 4.
was found that a β-silyl group migrated preferentially to the
carbenic center of 9.10a,d The bicyclo [2.2.1] and [2.2.2]
systems 11 and 13 were used as a probe for γ-trimethylsilyl
interactions.10b The products 12 and 14 formed from these
carbenes indicate that 1,3-trimethylsilyl migration to the
carbenic center occurs much more readily than either 1,2-
or 1,3-hydrogen migration. From these studies it was con-
cluded that the γ-trimethylsilyl group interacts quite effec-
tively with a carbenic center. This γ-interaction of silicon
with carbenes is quite different from the γ-stabilization of
carbocations described by Shiner. Whereas Shiner’s stabili-
zation of carbocations occurs via the “rear lobe” of the
carbon-silicon σ-orbital, our carbene studies demonstrate
“front lobe” interaction of carbenes analogous to the β-silyl
effect seen in carbocations of type 1.
We therefore wanted to study carbocations of type 15.
What kind of stabilization will be involved? We can envisage
two types of stabilization, i.e., “rear lobe” stabilization, 16,
analogous to that previously proposed by Shiner, as well as
“front lobe” stabilization, 17, analogous to that suggested in
our previous carbene studies. Reported here are the results of
these studies.
The major product from solvolysis of 5 was the solvent
capture product 7, with retention of configuration. The
observed rate enhancement, along with the observed pro-
ducts, can be explained by a γ-silicon stabilized intermediate
6 formed by assistance involving the rear lobe of the
carbon-silicon orbital in a “W” fashion. Computational
studies also provided evidence for carbocation stabilization
by a γ-trimethylsilyl group.7 Other studies by Shiner8 and
Grob9 support this type of γ-silyl stabilization of cations.
Our laboratory has extensively studied the effect of both
β- and γ-silyl groups on a variety of carbenes (Scheme 1).10 It
Results and Discussion
The 1-Phenyl-3-(Trimethylsilyl)cyclobutyl Carbocation.
Synthesis. Synthesis of precursors to cation 15 (R = Ph)
(Scheme 2) involved the cyclobutanone 18, which is readily
available by addition of excess diazomethane to trimethylsilyl
SCHEME 1. Migration of Neighboring Silyl Groups to Carbe-
nic Centers
SCHEME 2. Synthesis of Trifluoroacetates 21 and 22
(7) Davidson, E. R.; Shiner, V. J. Jr. J. Am. Chem. Soc. 1986, 108, 3135.
(8) (a) Shiner, V. J. Jr.; Ensinger, M. W.; Rutkowske, R. D. J. Am. Chem.
Soc. 1987, 109, 804. (b) Shiner, V. J. Jr.; Ensinger, M. W.; Huffman, J. C. J.
Am. Chem. Soc. 1989, 111, 7199.
(9) Grob, C. A.; Sawlewicz, P. Tetrahedron Lett. 1987, 28, 951.
(10) (a) Creary, X.; Wang, Y.-X. Tetrahedron Lett. 1989, 19, 2493. (b)
Creary, X.; Wang, Y.-X. J. Org. Chem. 1994, 59, 1604. (c) Creary, X.;
Butchko, M. A. J. Org. Chem. 2001, 66, 1115. (d) Creary, X.; Butchko, M. A.
J. Org. Chem. 2002, 67, 112.
J. Org. Chem. Vol. 74, No. 23, 2009 9045

Creary and Kochly
JOCArticle
ketene.11 Addition of phenylmagnesium bromide to this
silylated cyclobutanone gave a mixture of alcohols 19 and
20. Chromatography of the mixture gave pure alcohol 20.
However, pure 19 could not be completely separated from 20.
Therefore an enriched mixture of 19 and 20 was converted to a
mixture of trifluoroacetates 21 and 22. These trifluoroacetates
have significantly different reactivity in methanol, where 21
converts completely to methyl ether 23, while trifluoroacetate
22 remained intact at room temperature. Addition of sodium
methoxide to the reaction mixture resulted in transesterifica-
tion of unreacted 22. Chromatography of the resulting mix-
ture of 19 and 23 gave pure alcohol 19, which was then
converted to the pure trifluoroacetate 22.
TABLE 1. Solvolysis Rates for Substrates in CD3CO2D
The stereochemistry of alcohol 20 was confirmed by
nuclear Overhauser effect (NOE) studies. Irradiation of the
ortho hydrogens (Hb) on the phenyl ring caused a 4.6%
increase in the signal due to Ha. This confirmed the cis-
orientation of the phenyl group relative to Ha.
Solvolytic Studies. Solvolyses of trifluoroacetates 21 and
22 were carried out in CD3CO2D, where both substrates gave
the same product, the acetate 24 (Scheme 3). The structure of
24 was verified by independent synthesis of the protio
aExtrapolated from data at other temperatures.
SCHEME 3. Solvolyses of Trifluoroacetates 21 and 22
participation by silicon in the ionization step for 22. There-
fore further information was desirable.
Computational Studies. To further understand the carbo-
cation intermediates in the solvolyses of 21 and 22, density
functional calculations were carried out at the B3LYP/6-
31G* level. Two energy minima, corresponding to cations 26
analogue by acetylation of alcohol 20 with acetic anhydride.
The isomeric acetate was also synthesized from 19 to verify
that it was not present as a solvolysis product. This common
product 24 suggests that the same cationic intermediate is
involved in both solvolyses.
Solvolyses of 21 and 22 in CD3CO2D were monitored by
1H NMR and data for these and the unsilylated analogue 25
are summarized in Table 1. As can be seen from the relative
rates, both silylated compounds are somewhat rate enhanced
relative to the unsilylated analogue 25.12 Although the
enhancement in trifluoroacetate 22 is relatively small, the
enhancement seen for trifluoroacetate 21 is more significant.
What is the nature of the cationic intermediate in these
solvolyses? While the factor of 209 for 21 suggests stabiliza-
tion of the intermediate by the neighboring silyl group, the
rate enhancement of 13 is hardly compelling evidence for
(11) Zaitseva, G. S.; Bogdanova, G. S.; Baukov, Yu. I.; Lutsenko, D. F. J.
Organomet. Chem. 1976, 121, C21.
(12) For kinetic data on chloride and p-nitrobenzoate analogues of 25,
see: (a) Tanida, H.; Tsushima, T. J. Am. Chem. Soc. 1970, 92, 3397. (b)
Brown, H. C.; Ravindranathan, M.; Peters, E. N.; Rao, C. G.; Rho, M. M. J.
Am. Chem. Soc. 1977, 99, 5373.
FIGURE 1. B3LYP/6-31G* calculated energies of 1-phenylsilylcy-
clobutyl cations.
9046 J. Org. Chem. Vol. 74, No. 23, 2009

Creary and Kochly
JOCArticle
FIGURE 2. B3LYP/6-31G* calculated structures of 1-phenylsilylcyclobutyl cations.
and 27, were located on the potential energy surface. The
“rear lobe” form 26 is 3.7 kcal/mol more stable than the
“front lobe” form 27 (Figures 1 and 2). Structurally, this can
be seen in the contraction of the distance between the
cationic center and the carbon bonded to the TMS group,
as well as in the pucker angle of the cyclobutyl backbone. The
more stable “rear lobe” conformation 26 is more puckered
stable cation 26, which utilizes rear lobe γ-silyl participation,
is stabilized by 6.8 kcal/mol relative to 32. Computational
studies, as well as rate studies, therefore suggest that the
γ-silyl effect in this system is real, albeit not as large as the
β-silyl effect.
How do these computational studies fit with our solvolytic
studies? It is suggested that the faster reacting trifluoroace-
tate 21 ionizes with assistance from the rear lobe of the
γ-Si-C bond leading directly to the more stabilized cation
26. Cation 26 then captures solvent from the rear of this
delocalized cation, giving net retention of stereochemistry in
the product 24. The slower reacting trifluoroacetate 22 could
ionize with assistance from the front lobe of the Si-C bond
leading to the less stabilized conformation of the cation 27,
but the case is less compelling. The calculated barrier for
inversion of 27 to the more stable 26 is so small (0.45 kcal/
mol) that this inversion should occur much faster than other
potential pathways. Cation 26 (derived from 22) should then
capture solvent, ultimately giving 24 (a product of net
inversion from 22). Alternatively, trifluoroacetate 22 could
solvolyze via a simple unassisted (kC) pathway to give a
benzylic cation not further stabilized by the γ-silyl group.
This cation should rapidly reorganize to 26. The small rate
enhancement factor of 13 does not allow a clear distinction
between these two ionization modes for 22.
˚
and more contracted (1.916 A) than the “front lobe” con-
˚
formation 27 (2.316 A). The structure of the transition state
28 for the interconversion of 26 and 27 was also calculated at
the B3LYP/6-31G* level. This transition state lies 0.45 kcal/
mol above the “front lobe” stabilized form 27.
How do the stabilites of these γ-silyl-stabilized cations 26
and 27 compare with that of the β-silyl cation? The analo-
gous β-silyl cation 29 is 6.9 kcal/mol lower in energy than the
rear lobe stabilized cation 26 at the B3LYP/6-31G* level.
The isodesmic calculation shown in Scheme 4 was used to
further evaluate stabilization of cation 27 relative to the
SCHEME 4. Isodesmic Calculation Evaluating Trimethylsilyl
Stabilization of Cation 27
The 1-Methyl-3-(trimethylsilyl)cyclobutyl Carbocation.
Synthesis. The next objective was to replace the phenyl
groups in 21 and 22 with methyl groups, thereby forcing
the resulting cation to rely more heavily on stabilization
from the γ-trimethylsilyl group. It was anticipated that
larger γ-trimethylsilyl rate enhancements would result.
The required alcohols 33 and 34 were prepared and
purified (Scheme 5) in a similar fashion to their phenyl
analogues. Reaction of 3-trimethylsilylcyclobutanone 18
with methylmagnesium iodide gave a mixture of alcohols
33 and 34, which were not completely separated by column
chromatography. A separation procedure analogous to
that used to prepare pure alcohol 19, based on the higher
reactivity of trifluoroacetate 35 in methanol, was again
employed.
unsilylated analogue 32. The cation 27, which utilizes front
lobe γ-silyl participation, was found to be stabilized by 3.1
kcal/mol relative to the unsilylated cation 32. The more
J. Org. Chem. Vol. 74, No. 23, 2009 9047

Creary and Kochly
SCHEME 6. Solvolysis Product from Trifluoroacetate 35
JOCArticle
SCHEME 5. Synthesis of Trifluoroacetates 35 and 36
SCHEME 7. Solvolysis Products from Trifluoroacetate 36
As in the case of the phenyl analogues, stereochemistry of
alcohols 33 and 34 was again confirmed by NOE studies.
Irradiation of the methyl signal of 34 gave an increase of 11%
in the signal due to cis-Ha. Conversely, when the methyl
signal of 33 was irradiated, no change was seen in the
intensity of the signal due to trans-Ha. These alcohols were
converted to trifluoroacetates 35 and 36 in a straightforward
fashion.
SCHEME 8. Solvolysis of Trifluoroacetate 36: Mechanistic
Analysis
Solvolytic Studies. Solvolyses of trifluoroacetates 35, 36,
and the unsilylated analogue 38 were carried out in CD3CO2D.
The faster reacting substrate 35 gave the completely retained
acetate 39 as the exclusive product. Rate data are summarized
in Table 1. The rate enhancement of 46 000 seen for trifluor-
oacetate 35 is much larger than the value of 209 seen for the
phenyl analogue 21. This is presumably due to the fact that
methyl is a less effective carbocation stabilizing group than
phenyl. The carbocation intermediate demands more stabili-
zation from the γ-trimethylsilyl group, thus leading to a larger
rate enhancement. When one considers the fact that the model
compound, unsilylated 38, is itself “rate enhanced” due to the
nature of the delocalized cyclobutyl carbocation,13,14 the
actual anchimeric assistance in solvolysis of 35 is probably
even greater than the factor of 46 000. The exclusively retained
product, and the large rate enhancement, are completely
consistent with the rear lobe stabilized γ-silyl cation 40 as the
intermediate in this reaction (Scheme 6).
strong evidence for a front lobe stabilized γ-silyl cation.
Instead, it is suggested that 36 enters the cyclobutyl-
cyclopropylcarbinyl-homoallylic cation manifold,15 as
illustrated in Scheme 8. Rearrangement of cyclobutyl
cation 43a to cyclopropylcarbinyl cation 43b to cyclobu-
tyl cation 43c allows access to β-silyl stabilization. The
expected fate of 43c is desilylation to give the observed
major product 41. Capture of a cyclobutyl cation 43a,
where there is no rear lobe or front lobe γ-silyl stabiliza-
tion, accounts for the isomeric mixture of acetates 39
and 42.
Computational Studies. B3LYP/6-31G* calculations again
provide insight into the nature of the cationic intermediates
derived from trifluoroacetates 35 and 36. As with the phenyl-
substituted cations 26 and 27, two energy minima were
located on the potential energy surface (Figures 3 and 4).
These correspond to rear lobe and front lobe γ-trimethylsilyl
cations 40 and 44. Both cations show a larger pucker angle
and a shorter distance between the cationic carbon and the
silicon-substituted carbon than their phenyl counterparts.
This indicates a greater interaction of the cationic center with
the silyl-substituted carbon in 40 and 44 relative to the phenyl
analogues 26 and 27. Cation 40 is 16.0 kcal/mol more stable
than 44. This calculated stabilization is consistent with the
relatively large solvolysis rate of trifluoroacetate 35, which
presumably solvolyzes via rear lobe γ-silyl participation to
give cation 40.
In contrast to the exclusive product formed on solvolysis
of 35, the slower reacting trifluoroacetate 36 gave three
products (Scheme 7). The major product was the desilylated
1-methylcyclobutene, 41. Also formed were acetates 39 (net
inversion) and 42 (net retention). The differing behavior of
trifluoroacetates 35 and 36 indicates that a common cationic
intermediate is not involved in these solvolyses. There is no
(13) Cox, E. F.; Caserio, M. J.; Silver, M. S.; Roberts, J. D. J. Am. Chem.
Soc. 1961, 83, 2719.
(14) Nikoletic, M.; Borc-ic, S.; Sunko, D. E. Tetrahedron 1967, 23, 649.
(15) For reviews, see: (a) Richey, H. G., Jr. Carbonium Ions; Olah, G. A.,
Schleyer, P. v. R., Eds.; John Wiley & Sons: New York, 1972; Vol. III, Chapter 25.
(b) Wiberg, K. B.; Hess, B. A., Jr.; Ashe, A. J., III Carbonium Ions; Olah, G. A.,
Schleyer, P. v. R., Eds.; John Wiley & Sons: New York, 1972; Vol. III, Chapter 26.
9048 J. Org. Chem. Vol. 74, No. 23, 2009

Creary and Kochly
JOCArticle
SCHEME 9. Synthesis of Trifluoroacetate 48 and Mesylate 53
FIGURE 3. B3LYP/6-31G* calculated energies of 1-methylsilylcy-
clobutyl cations.
50 to the acetate 52. This transformation presumably occurs
via protonation of 50 followed by loss of CH3SO2H to
generate a carbocation intermediate.16 Under conditions
where cis-50 reacts readily, the trans-sulfinate 51 is unreac-
tive. This allows for isolation of 51 with no 50 present.
Oxidation of 51 with peracid to the trans-mesylate 53 was
straightforward.
Trifluoroacetate 48 reacted in CD3CO2D at 70 °C to give
the retained acetate 54 exclusively (Scheme 10). No bicyclo-
butane was observed, nor was any cyclopropylcarbinyl
SCHEME 10. Solvolysis of Trifluoroacetate 48: Mechanistic
Analysis
FIGURE 4. B3LYP/6-31G* calculated structures of 1-methylsilyl-
cyclobutyl cations.
The transition state 45 in the inversion of 44 to 40 was also
located, and this barrier is 1.5 kcal/mol. This is somewhat
larger than the barrier to inversion for phenyl analogues 27
and 26. The structure of the β-silyl-stabilized cation 43c was
also calculated and it is only 2.1 kcal/mol more stable than
40. Apparently rear lobe γ-silyl stabilization can be a much
larger effect than was originally anticipated.
The 3-(Trimethylsilyl)cyclobutyl Carbocation. Lithium
aluminum hydride reduction of ketone 18 gave cis- and
trans-3-trimethylsilylcyclobutanols 46 and 47 in a 85:15
ratio, while NaBH4 reduction gave a 93:7 ratio. These
alcohols were converted to the corresponding trifluoroace-
tates 48 and 49 as shown in Scheme 9. The cis-trifluoroace-
tate 48 reacted conveniently in CD3CO2D at 70 °C.
However, the trans-trifluoroacetate 49 was quite unreactive
and the more reactive trans-mesylate derivative 53 was
desired. Since chromatography was not successful in com-
pletely separating a pure sample of the trans-alcohol 47, an
alcohol mixture was converted to a mixture of sulfinate esters
50 and 51 (Scheme 9). Methanesulfonic acid catalyzed reac-
tion in acetic acid led to facile conversion of the cis-sulfinate
acetate formed. This latter product would have formed from
addition of acetic acid to bicyclobutane.17 The retained
acetate product 54 is completely consistent with the involve-
ment of the rear lobe silyl stabilized cation 55.
On the other hand, the behavior of mesylate 53 is quite
different from that of 48. Acetolysis of 53 (Scheme 11) gives
the cyclopropylcarbinyl acetate 56 as the major product
along with smaller amounts of cyclobutene, 57, and the
acetates 58 and 59a, and the mesylate 59b from internal
(16) Creary, X. J. Org. Chem. 1985, 50, 5080.
(17) Wiberg, K. B.; Szeimies, G. J. Am. Chem. Soc. 1970, 92, 571.
J. Org. Chem. Vol. 74, No. 23, 2009 9049

Creary and Kochly
JOCArticle
SCHEME 11. Solvolysis of Mesylate 53: Mechanistic Analysis
TABLE 2. Solvolysis Rates for Substrates in CD3CO2Da
return. Front lobe silyl stabilization of the developing catio-
nic intermediate does not appear to be a factor. Instead,
mesylate 53 appears to enter the cyclobutyl-cyclopropylcar-
binyl-homoallylic cation manifold, where the ionization of
53 is assisted by participation of a cyclobutane σ-bond.
Solvent capture of cyclopropylcarbinyl cation 60 leads to
56, while further rearrangement, involving cations 61 and 62,
leads to the minor products 57-59.
Rate data for trifluoroacetate 48, mesylate 53, cyclobutyl
mesylate, 63, and β-trimethylsilylcyclobutyl mesylate, 64, are
summarized in Table 2. The rate difference between cis-48
and trans-49 can be estimated from the rate of the trans-
mesylate 53. If one assumes a mesylate/trifluoroacetate rate
ratio18 of 105, then the rate enhancement in 48 relative to 49 is
about 103. Comparison of 48 with the unsubstituted cyclo-
butyl mesylate 63 is also informative. Since the solvolysis rate
of cyclobutyl tosylate is enhanced by σ-participation, the
aEstimated relative rate of mesylate assuming kROMs/kROCOCF
=
3
105.
the well-studied β-silyl group, which can offer rate enhance-
ments as large as 1012 in the cyclohexyl system.
Computational Studies. The B3LYP/6-31G* structures
and energies of potential γ-silyl-stabilized cations derived
from 48 and 53 are of interest. Energies of these and related
estimated rate enhancement due to the γ-trimethylsilyl group
in cis-trifluoroacetate 48 is much larger than the rate ratio of
105. The rate enhancement in cyclobutyl systems19 has been
estimated to be as low as 105 and as high as 1010. Hence the
actual rate enhancement in trifluoroacetate 48 relative to an
cations are shown in Figure 3. Calculations have also been
carried out with a larger basis set (B3LYP/6-311þG**) as
well as at the MP2/6-311þG** level. The general trends are
quite similar to those shown in Figure 3.20 The rear lobe
stabilized cation 55 has been located as an energy minimum
at the B3LYP/6-31B* level (Figure 5). This cation 55 appears
to be a corner protonated 1-trimethylsilylbicyclobutane. The
bond between the cationic carbon and the carbon carrying
“unassisted rate” is probably in the range of 1010 to 1015.
Rate data allow for a comparison of the enhancement due
to γ-silyl participation with enhancement due to β-silyl
stabilization. The γ-silyl rate enhancement is actually much
larger, as shown in Table 2. This again points to an enormous
stabilization of the cationic intermediate derived from 48 due
to the γ-silyl group, a stabilization that even exceeds that of
˚
the silyl group is only 1.662 A. In valence bond terms, the
bicyclobutane form 55b is a major resonance contributor.
In contrast to the Ph and CH3 analogues, the analogous
front lobe stabilized cation 63 is not an energy minimum but
is a transition state (4.5 kcal/mol above 64) for the inter-
conversion of cyclopropyl carbinyl cations 64a and 64b.
(18) (a) Noyce, D. S.; Virgilio, J. A. J. Org. Chem. 1972, 37, 2643. (b)
Creary, X. J. Org. Chem. 1979, 44, 3938. (c) Creary, X.; Jiang, Z. J. Org.
Chem. 1996, 61, 3482.
(19) (a) Foote, C. S. J. Am. Chem. Soc. 1964, 86, 1853. (b) Schleyer,
P. V. R. J. Am. Chem. Soc. 1964, 86, 1856.
(20) B3LYP/6-311þG** and MP2/6-311þG** calculated energies of
cations 55, 62, 63, 64, 65, and 66 are shown as Supporting Information.
9050 J. Org. Chem. Vol. 74, No. 23, 2009

Creary and Kochly
JOCArticle
FIGURE 5. B3LYP/6-31G* calculated structure and valence bond
description of cation 55.
A second minimum energy cyclopropylcarbinyl cation, 65, has
also been located on the potential energy surface (Figure 6).
Of note is the relative energy of the γ-silyl cation 55. The
B3LYP/6-31G* energy of this cation is actually 1.0 kcal/mol
lower than that of the β-silyl cation 62 and 5.9 kcal/mol lower
than the cyclopropylcarbinyl cation 64 (Figure 7). This
computational study is in line with the kinetic study, where
the rate enhancing effect in the acetolysis of 48 exceeds the
β-silyl rate enhancing effect in 64. This computational result
contrasts with that of the Ph and CH3 analogues 26 and 40
and demonstrates that indeed, γ-silyl stabilization can be an
enormous cation stabilizing effect. In fact, a planar transi-
tion state 66 has been located, where there is presumably no
cross ring γ-silyl stabilization, and this structure lies 28.8 kcal
above 55.
FIGURE 7. B3LYP/6-31G* calculated energies of silylcyclobutyl
and silylcyclopropylcarbinyl cations.
the γ-carbon-silicon bond. Rate enhancements relative to
unsilylated analogues range from a small factor (209) for the
phenyl-substituted system 21, to a very large factor (105) for
the more demanding unsubstituted system 48. The γ-silyl-
stabilized cation 55 derived from 48 is even more stable that
the β-silyl cation 62, as revealed by solvolytic rate studies as
well as B3LYP/6-31G* computational studies. While com-
putational studies suggest the existence of cations 27 and 44,
which are stabilized by an interaction of the vacant orbital
Conclusions. γ-Trimethylsilylcyclobutyl systems with a
leaving group cis to the silyl group ionize with assistance
due to a cation-stabilizing interaction with the rear lobe of
FIGURE 6. B3LYP/6-31G* calculated structures of cations 63, 64, and 65.
J. Org. Chem. Vol. 74, No. 23, 2009 9051

Creary and Kochly
JOCArticle
with the front lobe of the γ-carbon-silicon bond, solvolytic
studies are less conclusive. trans-Silyl systems 22, 36, and 53
do not solvolyze with significant rate enhancements relative
to unsilylated analogues. The systems 36 and 53 lead to
cations in the cyclobutyl-cyclopropylcarbinyl manifold. The
front lobe silyl stabilized cation 63 is not an energy minimum
at the B3LYP/6-31G* level, and is significantly higher in
energy than the rear lobe γ-silyl stabilized cation 55.
the mixture of 2- and 3-trimethylsilylcyclobutanone.11 Ketone 18
(191 mg) was dissolved in 2 mL of ether and the solution was
cooled to 0 °C. Methylmagnesium iodide (1.5 mL of a 1.0 M
solution in ether) was then added. The solution was warmed to
room temperature for 1 h and then recooled to 0 °C. Aqueous
NH4Cl solution was then added. The ether phase was separated,
washed with water and saturated NaCl solution, and dried over a
mixture of Na2SO4 and MgSO4. After filtration, solvent was
removedwitha rotaryevaporator leaving 192mg(91% yield) ofa
mixture of alcohols 33 and 34 in a 36:64 ratio. This mixture was
chromatographed on 8 g of silica gel and eluted with increasing
amounts of ether in hexanes. Eluting with 8% ether in hexanes
gave a fraction containing 81% alcohol 34. This mixture was used
for further studies. 1H NMR of 34 (CDCl3) δ 2.03 (m, 2 H), 1.84
(m, 2 H), 1.59 (br s, 1 H), 1.42 (t, J = 0.8 Hz, 3 H), 1.11 (t of t, J =
11.1, 8.6 Hz, 1 H), -0.04 (s, 9 H). 13C NMR of 34 (CDCl3) δ 72.7,
39.4, 26.5, 11.7, -3.2. ESI exact mass (M þ Naþ) calculated for
C8H18NaOSi 181.1019, found 181.0998.
Preparation of Alcohol 33. A mixture of alcohols 33 and 34
(38:62 ratio, 249 mg) was dissolved in 5 mL of ether and 294 mg
of 2,6-lutidine was added. The solution was cooled to 0 °C and
435 mg of trifluoroacetic anhydride was added. After 36 min, a
fast, cold aqueous workup ensued with ether extraction. The
ether extract was washed with ice-cold aliquots of water, dilute
HCl solution, water, aqueous NaHCO3 solution, and saturated
NaCl solution. The ether solution was dried over a mixture of
Na2SO4 and MgSO4 and filtered. The solvent was removed by
rotary evaporator to give a mixture of trifluoroacetates 36 and
35 (347 mg, 87% yield) in a 39:61 ratio.
A solution of 72 mg of 2,6-lutidine in 4 mL of methanol was
added to 254 mg of the trifluoroacetate mixture prepared above.
After the mixture was stirred for 16 h at room temperature, 1.6
mL of 0.50 M NaOCH3 in methanol was added. After 20 min at
room temperature, the methanol was removed with a rotary
evaporator and the crude product was chromatographed on 7 g
of silica gel. The column was eluted with increasing amounts of
ether in hexanes. 1-Methoxy-1-methyl-3-trimethylsilylcyclobu-
tane (derived from 35) (58 mg) eluted with 3% ether in hexanes.
The alcohol 33 (60 mg; 88% yield based on the amount of
trifluoroacetate 36 in the starting mixture) eluted with 8-10%
ether in hexanes. 1H NMR of 33 (CDCl3) δ 2.09 (m, 2 H), 1.98
(m, 2 H), 1.78 (br s, 1 H), 1.74 (t of t, J = 10.7, 9.2 Hz, 1 H), 1.26
(s, 3 H), -0.04 (s, 9 H). 13C NMR of 33 (CDCl3) δ 74.9, 37.5,
29.1, 12.2, -3.2. ESI exact mass (M þ Naþ) calculated for
C8H18NaOSi 181.1019, found 181.0983.
Preparation of Alcohol 46. A solution of 42 mg of ketone 18 in
1 mL of methanol was cooled in an ice bath and 15 mg of NaBH4
was added in one portion with stirring. The mixture was warmed
to room temperature for 1 h and about 2/3 of the methanol was
then removed with a rotary evaporator. Dilute NaOH solution
was then added and the mixture was extracted with ether. The
ether extract was washed with saturated NaCl solution then
dried over MgSO4, and the solvent was removed with a rotary
evaporator to give 39 mg (93% yield) of alcohols 46 and 47 in a
93:7 ratio. This mixture was used for conversion to the trifluor-
oacetate 48. 1H NMR of 46 (CDCl3) δ 4.27 (sextet, J = 7 Hz, 1
H), 2.29 (m, 2 H), 1.71 (d, J = 7 Hz, 1 H), 1.65 (m, 2 H), 1.01 (t of
t, J = 11.6, 7.8 Hz, 1 H), -0.04 (s, 9 H). 13C NMR of 46 (CDCl3)
δ 67.2, 35.1, 11.6, -3.3. ESI exact mass (M þ H2O þ Naþ)
calculated for C7H18NaO2Si 185.0968, found 185.1001.
Experimental Section
Preparation of Alcohol 20. A mixture of 2-trimethylsilyl-
cyclobutanone and 3-trimethylsilylcyclobutanone (1811) (1.03
g, 1:1.5 ratio) was dissolved in 8 mL of anhydrous ether and the
solution was cooled to 0 °C. Phenylmagnesium bromide (8.2 mL
of a 1.0 M solution in ether) was then added. The solution was
warmed to room temperature for 2 h and then recooled to 0 °C.
The mixture was then quenched with dilute NH4Cl solution. The
ether phase was separated, washed with saturated NaCl solu-
tion, and dried over a mixture of Na2SO4 and MgSO4. After
filtration, solvent was removed with a rotary evaporator. The
crude product mixture was chromatographed on 15 g of silica
gel and eluted with increasing amounts of ether in hexanes.
1-Phenyl-2-trimethylsilylcyclobutanol (360 mg, 22% yield)
eluted with 4% ether in hexanes. A mixture of alcohols 19 and
20 (717 mg, 45% yield) coeluted with 8% ether in hexanes. This
mixture was rechromatographed on 10 g of silica gel and the
column was eluted with 4% ether in hexanes. A fraction contain-
ing 23 mg of pure 20 eluted first, followed by fractions contain-
ing 20 contaminated with increasing amounts of alcohol 19. The
final fractions contained alcohol 19 as the major component,
1
but contaminated with 20. H NMR of 20 (CDCl3) δ 7.60 (d,
J = 8.3 Hz, 2 H), 7.40 (t, J = 7.8 Hz, 2 H), 7.31 (t, J = 7.3 Hz,
1 H), 2.60 (m, 2 H), 2.22 (m, 2 H), 2.13 (s, 1 H), 1.17 (t of t, J =
11.4, 8.7 Hz, 1 H), 0.02 (s, 9 H). 13C NMR of 20 (CDCl3) δ 145.9,
128.6, 127.6, 125.5, 75.9, 38.6, 11.7, -3.2. ESI exact mass (M þ
Naþ) calculated for C13H20NaOSi 243.1176, found 243.1173.
Preparation of Alcohol 19. A mixture of alcohols 19 and 20
(213 mg of a 55:45 ratio) was dissolved in 5 mL of ether and 285
mg of 2,6-lutidine was added. The solution was cooled to 0 °C
and 373 mg of trifluoroacetic anhydride was added. After 10
min, a cold aqueous workup ensued with ether extraction. The
ether extract was washed with cold aliquots of water, dilute HCl
solution, water, aqueous NaHCO3 solution, and saturated
NaCl solution. The ether solution was then dried over a mixture
of Na2SO4 and MgSO4 and filtered. The solvent was removed
with a rotary evaporator to give a mixture of trifluoroacetates 21
and 22 (291 mg, 95% yield) in a 45:55 ratio.
A solution of 54 mg of 2,6-lutidine in 4 mL of methanol was
added to 271 mg of the trifluoroacetate mixture prepared above.
After the mixture was stirred for exactly 8 min at 26 °C, 1.4 mL
of 0.50 M NaOCH3 in methanol was added. After 10 min at
room temperature, the methanol was removed with a rotary
evaporator and the crude mixture was chromatographed on 7 g
of silica gel. The column was eluted with increasing amounts of
ether in hexanes. Methyl ether 23 (99 mg) eluted with 2% ether in
hexanes. The alcohol 19 (92 mg, 89% yield based on the amount
of trifluoroacetate 22 in the starting mixture) eluted with
6-10% ether in hexanes, mp 86-87 °C. 1H NMR of 19
(CDCl3) δ 7.38 (d, J = 8.4 Hz, 2 H), 7.35 (t, J = 7.7 Hz, 2 H),
7.26 (t, J = 7.2 Hz, 1 H), 2.44 (m, 2 H), 2.32 (m, 2 H), 2.08 (quin,
J = 9.9 Hz, 1 H), 2.02 (s, 1 H), -0.04 (s, 9 H). 13C NMR of 19
(CDCl3) δ 147.0, 128.6, 127.3, 124.9, 78.5, 37.0, 15.0, -3.3.
Exact mass (FAB) calculated for C13H20OSi 220.1283, found
220.1289.
Preparation of Mesylate 53. A solution of 96 mg of a mixture
of alcohols 46 and 47 (82% of 46 and 18% of 47) was dissolved in
2 mL of CH2Cl2 and 118 mg of Et3N in 0.5 mL of CH2Cl2 was
added. The mixture was cooled to -35 °C and 95 mg of
CH3SOCl in 0.5 mL of CH2Cl2 was slowly added. The mixture
was warmed to room temperature for about 5 min and then
transferred to a separatory funnel with ether. The mixture was
then washed successively with cold water, cold dilute HCl, cold
Preparation of Alcohol 34. A pure sample of 3-trimethylsilyl-
cyclobutanone, 18, was isolated by silica gel chromatography of
9052 J. Org. Chem. Vol. 74, No. 23, 2009

Creary and Kochly
JOCArticle
water, and saturated NaCl solution. The organic phase was
dried over a mixture of Na2SO4 and MgSO4. After filtration, the
solvents were removed with a rotary evaporator to give 137 mg
(100% yield) of a mixture of crude sulfinate esters 50 and 51. 1H
NMR of 50 (CDCl3) δ 4.73 (t of t, J = 8.3, 7.1 Hz, 1 H), 2.59 (s, 3
H), 2.33 (m, 2 H), 1.99 (m, 1 H), 1.92 (m, 1 H), 1.15 (t of t, J =
11.7, 7.9 Hz, 1 H), -0.03 (s, 9 H). 13C NMR of 50 (CDCl3)
substrate in 400 mg of CD3CO2D containing approximately
1.5 equiv of 2,6-lutidine. The sample was sealed in an NMR
tube and the tube was placed in the probe of an NMR spectro-
meter at 25.0 °C or at the appropriate temperature. For slower
reactions, the sample was placed in a constant temperature bath
at 25.0 °C between readings. At appropriate time intervals, the
sample was analyzed by 1H NMR and relative areas due to
starting trifluoroacetate or mesylate were measured. For runs
at higher temperatures, the tube was placed in a constant
temperature bath at the appropriate temperature. At appro-
priate time intervals, the tube was then quenched in a water
1
δ 71.6, 44.4, 33.4, 32.8, 12.9, -3.4. H NMR of 51 (CDCl3) δ
4.65 (quintet of doublets, J = 7.3, 1.1 Hz, 1 H), 2.60 (s, 3 H), 2.48
(m, 1 H), 2.40 (m, 1 H), 2.26 (m, 2 H), 1.50 (t of t of d, J = 12.1,
4.3, 1.3 Hz, 1 H), -0.01 (s, 9 H). 13C NMR of 51 (CDCl3) δ 72.4,
44.5, 32.3, 31.9, 12.7, -3.3.
1
bath at 15 °C and rapidly analyzed by H NMR at ambient
temperature.
The mixture of sulfinate esters 50 and 51 obtained above
(132 mg) was dissolved in 3.2 mL of 0.05 M CH3SO3H in acetic
acid. After 90 min at room temperature, the sample was diluted
with 25 mL of pentane and the solution was cooled to 0 °C. The
solution was extracted with 2 portions of cold water and then
with NaHCO3 solution. The pentane extract was washed with
saturated NaCl solution and dried over a mixture of Na2SO4
and MgSO4. Solvent removal by rotary evaporator gave 108 mg
of a mixture of acetate 52 and unreacted sulfinate ester 51.
The mixture of sulfinate ester 51 and acetate 52 obtained
above was dissolved in 2 mL of CDCl3 and the solution was
cooled in a water bath at 15 °C. m-Chloroperbenzoic acid (25 mg
of 86% peracid) was added and the mixture was kept at room
temperature for 3.5 h. The mixture was diluted with 20 mL of
pentane and the solution was cooled in an ice bath. The solution
was extracted with a cold Na2S2O3/NaI/Na2CO3 solution con-
taining about 10 mg of NaOH. The pentane extract was dried
over MgSO4. The solvent was removed with a rotary evaporator
to give 105 mg of a mixture of acetate 52 and mesylate 53.
Further evacuation of this residue at 0.1 mm (to remove the
more volatile acetate 52) gave 24 mg of mesylate 53. 1H NMR
of 53 (CDCl3) δ 4.92 (d of quintets, J = 7.4, 1.3 Hz, 1 H), 2.97 (s,
3 H), 2.56 (m, 2 H), 2.32 (m, 2 H), 1.56 (t of t of d, J = 12.3, 4.3,
1.3 Hz, 1 H), 0.01 (s, 9 H). 13C NMR of 53 (CDCl3) δ 74.3, 38.4,
31.6, 12.8, -3.3. ESI exact mass (M þ Naþ) calculated for
C8H18NaO3SSi 245.0638, found 245.0613.
Preparation of Mesylate 64. The preparation of mesylate 64
was analogous to the preparation of 53. Thus, a mixture of
2-trimethylsilylcyclobutanone and 3-trimethylsilylcyclobuta-
none was reduced with NaBH4 in CH3OH. The resulting alcohol
mixture was converted to a mixture of methanesulfinate esters
by reaction with CH3SOCl and Et3N as described above. This
mixture of sulfinate esters was allowed to react in 0.05 M
CH3SO3H in acetic acid for 90 min, which gave acetate 52, a
small amount of sulfinate 51, as well as the sulfinate derivative of
cis-2-trimethylsilylcyclobutanol. Oxidation of this mixture with
m-chloroperbenzoic acid gave acetate 52, a small amount of
mesylate 53, along with mesylate 64. Kinetic studies on 64 were
carried out with this mixture. Removal of the acetate 52 under
vacuum led to autocatalytic decomposition of the highly reac-
tive mesylate 64.
For trifluoroacetate 21 in CD3CO2D, the rate of disappear-
ance of the TMS singlet at δ 0.03 was monitored. For trifluoro-
acetate 22, the TMS singlet at δ -0.07 was monitored. For
trifluoroacetate 25, the area of the m-H’s of 25 at δ 7.40 was
monitored by using the 2,6-lutidine signal at δ 8.24 as an internal
standard. The reaction of trifluoroacetate 35 was monitored by
measuring the area of the CH3 singlet at δ 1.68. Trifluoroacetate
36 was monitored by measuring the area of the TMS singlet at δ
0.00. Trifluoroacetate 38 was monitored by measuring the
decrease in the area of the multiplet at δ 2.22, using the 2,6-
lutidine signal at δ 7.59 as an internal standard. For trifluoro-
acetate 48, the rate of disappearance of the quintet at δ 5.26 was
monitored.
In the case of mesylates 53, 63, and 64 in CD3CO2D, the rates
of disappearance of the singlets due to the ROSO2CH3 group at
δ 3.00, 2.99, and 2.99, respectively, were monitored with use of
the 2,6-lutidine as an internal standard.
Typical data illustrating the reaction of trifluoroacetates 21,
35, and 48, and mesylate 53 in CD3CO2D are given as Support-
ing Information. First-order rate constants for disappearance of
substrates were calculated by standard least-squares proce-
dures. Correlation coefficients were all greater than 0.9998.
The maximum standard deviation in duplicate runs was (2%.
Computational Studies. Ab initio molecular orbital calcula-
tions were performed using the Gaussian 03 series of pro-
grams.21 All structures were characterized as energy minima
via frequency calculations that showed no negative frequencies,
or as transition states that showed one negative frequency.
Supporting Information Available: Complete ref 21, experi-
mental procedures for preparation of trifluoroacetates, experi-
mental procedures for solvolysis reactions and product
analyses, the B3LYP/6-31G* calculated structures, energies,
and Cartesian coordinates of 26, 27, 28, 29, 40, 43c, 44, 45, 55,
62, 63, 64, 65, and 66, B3LYP/6-311þG** and MP2/6-311þG**
calculated energies of cations 55, 62, 63, 64, 65, and 66, 1H and
13C NMR spectra of compounds 19, 20, 21, 22, 24, 33, 34, 35, 36,
39, 42, 46, 48, 53, 54, 56, 58, 59a, and 59b, as well as evolving 1H
NMR spectra during solvolyses of 21, 35, 48, and 53 in
CD3CO2D. This material is available free of charge via the
Internet at http://pubs.acs.org.
Solvolyses of Trifluoroacetates and Mesylates. Kinetics Pro-
cedures. Rate constants reported in Tables 1 and 2 were all
determined using a 600 MHz 1H NMR spectrometer. A solution
was prepared by dissolving approximately 5 mg of the appropriate
(21) Frisch, M. J. et al. Gaussian 03, Revision C.01; Gaussian, Inc.,
Wallingford, CT, 2004.
J. Org. Chem. Vol. 74, No. 23, 2009 9053