New computational and experimental evidence for the mechanism of the Sakurai reaction

Article abstract of DOI:10.1021/ja9723693

Reaction of aldehydes with allyltrimethylsilane promoted by BF₃ (Sakurai reaction) affords fluorotrimethylsilane and a borylated homoallylic alcohol in a noncatalytic fashion. These reaction products have been identified through ¹³C

Full text of DOI:10.1021/ja9723693

J. Am. Chem. Soc. 1997, 119, 12131-12135
12131
New Computational and Experimental Evidence for the
Mechanism of the Sakurai Reaction
Andrea Bottoni,* Anna Luisa Costa, Donata Di Tommaso, Ivan Rossi, and
Emilio Tagliavini*
Contribution from the Dipartimento di Chimica “G. Ciamician”, UniVersita’ di Bologna,
Via Selmi 2, 40126 Bologna, Italy
ReceiVed July 15, 1997X
Abstract: Reaction of aldehydes with allyltrimethylsilane promoted by BF3 (Sakurai reaction) affords fluorotri-
methylsilane and a borylated homoallylic alcohol in a noncatalytic fashion. These reaction products have been
identified through 13C NMR and 29Si NMR. A model system formed by acetaldehyde, allylsilane, and BH2F has
been used to investigate the mechanism of this reaction at the DFT computational level. Four transition states, quite
close in energy, have been found for the model reaction. These transition states, which explain the stereochemical
features of the Sakurai reaction, correspond to eight-membered cyclic structures deriving from C-C (tight) and
Si-F (loosen) bond formation and synclynal (gauche) disposition of the reacting double bonds. No antiperiplanar
or six-membered cyclic transition structures have been located.
Introduction
Scheme 1. The Sakurai Reaction
The mechanism of the addition reactions to aldehydes of
allylic nucleophiles promoted by Lewis acids1 has been the
object of a long-lasting debate in the community of organic
chemists, since the discovery of the syn-convergent diastereo-
selectivity of the allylsilane addition (Sakurai reaction, Scheme
1)2,3 and allylstannane addition.4 Classic chair-like “cyclic”
transition states5 cannot easily account for the observed stereo-
chemistry, and “open” transition state models have been
generally adopted. However, many questions concerning the
mechanism of these reactions remain unclear.
electrophiles behave in the opposite way.6c The same result
was unambiguously obtained by Denmark using a cyclic
bifunctional probe.7 In these reactions the anti-SE′ mechanism
prevents the occurrence of six-membered cyclic transition states
(trans-cyclohexene type structure). However, the rigidity of
Denmark’s probe does not allow the testing of all the possible
orientations of the reagents, while Kumada’s model is strongly
sterically biased.
Open transition state models offer complete freedom to the
relative orientation of the reagents. Among the various pos-
sibilities, the antiperiplanar (anti orientation of the reacting CdC
and CdO bonds) arrangement provides a straightforward
explanation of the syn selectivity of the allylation reaction. It
was originally introduced by Yamamoto4,8 to account for the
results obtained with crotylstannanes; later on this model was
applied to allylsilanes.2 However, a substantial crop of stereo-
chemical findings, obtained for both reagents, shows a better
fit with a synclinal model.9
In acyclic systems Kumada et al. showed that the addition
of allylsilanes to aldehydes promoted by TiCl4 follows an anti-
SE' mechanism6a,b (antharafacial orientation of the trialkylsilyl
group and the incoming aldehyde), although other kinds of
X Abstract published in AdVance ACS Abstracts, November 15, 1997.
(1) For reviews on the aldehyde allylation see: (a) Fleming, I. Org. React.
1989, 37, 57. (b) Fleming, I. In ComprehensiVe Organic Synthesis;
Heathcock, C. H., Ed.; Pergamon Press: Oxford, 1991; Vol. 2, p 563. (c)
Sakurai H. Pure Appl. Chem. 1992, 54, 1. (d) Yamamoto, Y.; Asao, N.
Chem. ReV. 1993, 93, 2207. (e) Nishigaichi, Y.; Takuwa, A.; Naruta, Y.;
Maruyama, K. Tetrahedron 1993, 49, 7395. (f) Masse, C. E.; Panek, J. S.
Chem. ReV. 1995, 95, 1293.
(2) Hosomi, A.; Sakurai, H. Tetrahedron Lett. 1976, 16, 1295.
(3) Hayashi, T.; Kabeta, K.; Hamachi, I.; Kumada, M. Tetrahedron Lett.
1983, 24, 2865.
(4) Yamamoto, Y.; Maruyama, K.; Naruta, Y.; Yatagai, H. J. Am. Chem.
Soc. 1980, 102, 7101.
(5) Cyclic transition states have been proposed for the reactions of
allylsilacyclobutanes (Matsumoto, K.; Oshima, K.; Utimoto, K. J. Org.
Chem. 1994, 59, 7152) and allyltrifluorosilanes (Kobayashi, S.; Nishio, K.
J. Org. Chem. 1994, 59, 6620). These transition structures have recently
been determined with theoretical methods: Fujimoto, H.; Omoto, K.;
Sawada, Y. J. Am. Chem. Soc. 1996, 118, 1750-1755. Also the thermal or
pressurized reaction of allylstannanes with aldehydes has been proposed to
proceed through cyclic transition states: Yamamoto, Y.; Maruyama, K.;
Matsumoto, K. J. Chem. Soc., Chem. Commun. 1983, 489. A theoretical
investigation of the allylation with allyl tetrafluorosilicates has shown that
this reaction involves a cyclyc transition structure: Kira, M.; Sato, K.;
Sakurai, H.; Hada, M.; Izawa, M.; Ushio, J. Chem. Lett. 1991, 387.
(6) (a) Hayashi, T.; Konishi, M.; Ito, H.; Kumada, M. J. Am. Chem.
Soc. 1982, 104, 4962. (b) Hayashi, T.; Konishi, M.; Kumada, M. J. Am.
Chem. Soc. 1982, 104, 4963. (c) Wetter, H.; Scherer, P.; Schweizer, W. B.
HelV. Chim. Acta 1979, 62, 1985.
Since the interest for the Sakurai and related reactions has
greatly grown up in the last years with the development of
efficient catalytic asymmetric methodologies used to carry out
these reactions,10 we have decided to perform a combined
experimental and computational study of the mechanism of the
Sakurai reaction. We wish to emphasize that, to our knowledge,
our work represents the first determination of the structure of
(7) Denmark, S. E.; Almstead, N. G. J. Org. Chem. 1994, 59, 5130.
(8) Yamamoto, Y.; Yatagai, H.; Ishihara, Y.; Maeda, N.; Maruyama, K.
Tetrahedron 1984, 40, 2239-2246.
(9) Allylsilanes: (a) Denmark, S. E.; Weber, E. J. HelV. Chim. Acta 1983,
66, 1655. (b) Denmark, S. E.; Weber, E. J.; Wilson, T. M.; Tetrahedron
1989, 45, 1053. (c) Mikami, K.; Maeda, T.; Kishi, N.; Nakai, T. Tetrahedron
Lett. 1984, 25, 5151. Allylstannanes: (d) Denmark, S. E.; Hosoi, S. J. Org.
Chem. 1994, 59, 5133. (e) Keck, G. E.; Dougherty, S. M.; Savin, K. A. J.
Am. Chem. Soc. 1995, 117, 6210-6223. (f) Gevorgyan, V.; Kadota, I.;
Yamamoto, Y. Tetrahedron Lett. 1993, 34, 1313. See also: Fleming, I
Chemtracts 1991, 4, 21.
S0002-7863(97)02369-X CCC: $14.00 © 1997 American Chemical Society

12132 J. Am. Chem. Soc., Vol. 119, No. 50, 1997
Bottoni et al.
Scheme 2
Chart 1. Reaction Products [a: R ) CH3, R′ ) H, X ) H
(Computational); b: R ) Ph, R′ ) CH3, X ) F
(Experimental); c: R ) n-C7H15, R′ ) CH3, X ) F
(Experimental)]
Table 1. Reaction of Aldehydes with Allyltrimethylsilane
a
Promoted by BF3‚OEt2
3b
products,
aldehyde (equiv)
T (°C)
t (min) conv, %b
(yield %)b
1b
1b
1b
1c
0.2
1
1
0.2
1
1
-80fRT 240c
20
60
50
30
100
100
4b (8), 7b (12)
4b (24), 7b (36)
4b (32), 7b (18)
4c (17)
4c (73)
4c (47)
-80
120
1
RT
-80fRT 240c
Scheme 3
1c
-80
RT
240
60
1cd
a All reactions were carried out with a 1:2 ) 1.0:1.0 molar ratio in
CH2Cl2 and quenched with pH 7 aqueous buffer solution. b Determined
gas chromatographically. c Longer reaction times did not afford sig-
nificant change in the composition of the reaction mixture. d The
reaction was carried out in pentane.
the transition state of a Lewis acid-promoted addition of carbon
nucleophiles to carbonyl compounds. Our results, besides
clarifying the mechanism of the reaction between allylsilanes
and aldehydes, can be a useful tool for understanding related
processes such as the allylstannane addition and the silylenol
ether addition (Mukayiama reaction11).
Investigation by 29Si NMR has demonstrated that fluorotri-
methylsilane (6b) (δSi 34.7 ppm, JSi-F ) 273 Hz) is the only
silylated product formed in the temperature range from -80 to
20 °C in the reactions of 1b or 1c with 2b + 3b. This finding
is in agreement with the production of ClSiMe3 found by
Reetz13a and Denmark13b in the corresponding reactions pro-
moted by TiCl4. The detection of 5 has been less straightfor-
ward, since reference spectra are unavailable. Nevertheless, 13C-
NMR investigation of the reaction mixture of 1c + 2b + 3b in
Results and Discussion
In this work we have experimentally investigated the Sakurai
reaction involving benzaldehyde (1b) or octanal (1c), allyltri-
methylsilane (2b), and BF3‚OEt2 (3b) (Scheme 2). In our
computations we have considered the model reaction between
acetaldehyde (1a), allylsilane (2a), and BH2F (3a).
Experimental Findings. We have carried out the experi-
mental reactions in CH2Cl2 and pentane using 1b and 1c under
both stoichiometric and catalytic conditions (1.0 or 0.2 mol equiv
of 3b). We have found that the conversion of the aldehyde
has never exceeded the amount of the Lewis acid employed
(Table 1).
As expected, the common products of the two reactions, after
quenching with water, are the homoallylic alcohols 4b,c (Chart
1). However, the two substrates differ in the byproducts
afforded: the diastereomeric ethers 7b are provided by benz-
aldehyde in substantial amount, while only minor quantities of
a complex mixture of byproducts are obtained from octanal.
The bis-homoallylic hydrocarbon product, which was found by
Mayr and Gorath12 in a recent study on the same reaction, could
not be detected in our reaction mixtures. In all cases no traces
of the homoallylic silyl ethers 8b,c have been observed.
CH2Cl2 showed the presence of 6b (δC ) 0.06 ppm, JC-F
)
14.9 Hz) and of a new allylic product with a characteristic set
of signals: δC ) 133.0 (C-2), 118.9 (C-1), 78.3 (C-4), 39.3
(C-3) ppm; the same product has been found to form in the
treatment of either the alcohol 4c or the silylether 8c with
BF3‚OEt2; on this basis we attribute this set of signals to 5c.14
In all cases, upon quenching with water, these reactions afforded
4c as the sole or major product. Since the desilylation of 8c
by BF3‚OEt2 is very fast, we cannot rule out the transient
formation of a silylated product 8, which eventually undergoes
exchange with BF3, but the aVailable experimental data support
the hypothesis of a direct formation of the borylated homoallylic
alcohol 5b,c and of 6b from the reaction of aldehydes and
allylsilanes in the presence of BF3 (Scheme 3).
Computational Findings. The model reaction 1a + 2a +
3a has been investigated at the density functional theory (DFT)15
level with use of the Gaussian 9416 series of programs. The
(13) (a) Reetz, M. T.; Raguse, B.; Marth, C. F.; Hugel, H. M.; Bach, T.;
Fox, D. N. A. Tetrahedron 1992 48, 5731. (b) Denmark, S. E.; Almstead,
N. G.; Tetrahedron 1992, 48, 5565.
(14) Also a differrent allylic product 5'c having δC ) 135.3 (C-2), 117.5
(C-1), 73.2 (C-4), 41.5 (C-3) ppm formed in minor amount from either the
allylation reaction or the desilylation of 8c with BF3‚OEt2. In the latter
case the relative amount of 5c and 5′c was found to vary depending on the
relative amount of the reagents; on this basis we tentatively attribute an
(RO)2BF structure to 5′c.
(10) Allylstannanes: (a) Costa, A. L.; Piazza, M. G.; Tagliavini, E.;
Trombini, C.; Umani-Ronchi, A. J. Am. Chem. Soc. 1993, 115, 7001. (b)
Keck, G. E.; Tarbet, K. H.; Geraci, L. S. J. Am. Chem. Soc. 1993, 115,
8467. (c) Keck, G. E.; Krishnamurthy, D.; Grier, M. C. J. Org. Chem. 1993,
58, 6543. (d) Weigand, S.; Brukner, R. Chem. Eur. J. 1996, 2, 1072. (e)
Bedeschi, P.; Casolari, S.; Costa, A. L.; Tagliavini, E.; Umani-Ronchi, A.
Tetrahedron Lett. 1995, 36, 7897. (f) Yanagisawa, A.; Nakashima, H.;
Ishiba, A.; Yamamoto, H. J. Am. Chem. Soc. 1996, 118, 4723. (g)
Yanagisawa, A.; Ishiba, A.; Nakashima, H.; Yamamoto, H. Synlett 1997,
88. Allylsilanes: (h) Hishihara, K.; Mouri, M.; Gao, Q.; Maruyama, T.;
Furuta, K.; Yamamoto, H. J. Am. Chem. Soc. 1993, 115, 11490. (i) Carreira,
E. M.; Gauthier, D. R., Jr. Angew. Chem., Int. Ed. Engl. 1996, 35, 2363.
(11) Mukaiyama, T.; Banno, K.; Narasaka, K. J. Am. Chem. Soc. 1974,
96, 7503. See also: Gennari, C. In ComprehensiVe Organic Synthesis; Trost,
B. M., Flemings, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 2, p 629.
(12) Mayr, H.; Gorath, G. J. Am. Chem. Soc. 1995, 117, 7862.
(15) Parr, R. G.; Yang W. Density-Functional Theory of Atoms and
Molecules; Oxford University Press: New-York, 1989.
(16) Gaussian 94, Revision D.3, Frisch, M. J.; Trucks, G. W.; Schlegel,
H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;
Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.;
Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.;
Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng,
C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E.
S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.;
Baker, J.; Stewart, J. P.; Head-Gordon, M.; Gonzalez, C.; and Pople, J. A.
Gaussian, Inc.: Pittsburgh, PA, 1995.

EVidence for the Mechanism of the Sakurai Reaction
J. Am. Chem. Soc., Vol. 119, No. 50, 1997 12133
Figure 2. Schematic structures of the two transition states associated
with a syn-SE′ attack of the electrophyle (bond lengths are in angstroms
and angles in degrees).
Figure 1. Schematic structures of the two transition states associated
with an anti-SE′ attack of the electrophyle (bond lengths are in
angstroms and angles in degrees).
to a highly distorted chair-chair (CC) conformation of cyclo-
octane while a sort of boat-chair (BC) and boat-half-chair (BHC)
conformation can be recognized in structures B and D,
respectively.21 All these structures are characterized by a
synclinal arrangement of the reacting CdO and CdC double
bonds (the dihedral angle ω is 58.6° in A, 83.5° in B, 26.7° in
C, and 59.8° in D). In spite of accurate investigation of the
potential energy surface no antiperiplanar transition structures
(dihedral angle ω close to 180°) have been located. Even if
the Si-F incipient bond is quite loose, nevertheless this
interaction seems to be crucial in determining the reaction path
toward the products: analysis of the reaction coordinate in the
product direction by computing the intrinsic reaction path (IRC)
shows that the Si-F distance decreases (formation of a new
Si-F bond) and the Si-C and B-F distances increase
(breaking of the two corresponding bonds) to afford 5a and
6a, which correspond to the experimentally obserVed products.
From this point of view the proposed mechanism appears to be
related to the push-pull hypothesis of Yamamoto,1d,9f who says
that both acidic activation of the carbonyl and contemporary
nucleophylic attack to tin (or silicon) are essential for the
accomplishment of the allylation reaction.
Two of the four TS’s, namely the BB structure A and the
BC structure B, derive from an anti-SE′ attack of the electrophile
on the allylsilane, while C and D correspond to a syn-SE′
approach. Since A and B are diastereotopic (the opposite faces
of the prochiral aldehyde are involved in the reaction) they
would lead to opposite stereoisomers if 3-substituted allylsilanes
were in use; the same considerations hold for C and D.
The energy values for the four transition structures and the
products are collected in Table 2. The relative energy values,
computed with respect to reactants, approximately correspond
to the activation energies for the four reaction channels and to
the reaction enthalpy. From the values reported in Table 2 it
nonlocal hybrid three-parameter functional proposed by Becke17
(B3LYP) and available in Gaussian 94 has been used. A DZ
Dunning basis set18a including sp diffuse functions18b on oxygen
(ê ) 0.00845), boron (ê ) 0.00315), and fluorine (ê ) 0.01076)
and d orbitals on silicon (ê ) 0.3247) has been used. The
geometries have been fully optimized with the gradient method.
The vibrational frequencies and the intrinsic reaction path
(IRC)19 have been calculated at the same level of theory.
Four transition structures A, B, C, and D have been located
on the potential energy surface. These structures are schemati-
cally represented in Figure 1 (A and B) and Figure 2 (C and
D) together with the values of the most relevant geometrical
parameters. We must point out that (i) since any significant
Si-O interaction is lacking, all these structures apparently
correspond to “open” transition states and that (ii) in spite of
extensiVe search no six-membered “cyclic” transition structures
haVe been found. However, a deeper inspection of the structures
reported in the figures shows that the silicon and the fluorine
atoms are not too far away: the Si-F distance ranges from
2.661 Å in A to 3.634 Å in B. These distances are probabily
associated with an electrostatic interaction between the two
atoms (the net charges on Si and F in structure A are 0.36 and
-0.40, respectively) and suggest, especially in A and D, an
incipient Si-F bond. Consequently they could be considered
eight-membered cyclic structures:20 structure A resembles the
boat-boat (BB) conformation of cyclooctane and C corresponds
(17) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(18) (a) Dunning, T. H., Jr. J. Chem. Phys. 1970, 53, 2823. (b) Clark,
T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer, P. v. R. J. Comp. Chem.
1993, 4, 294.
(19) (a) Fukui, K. Acc. Chem. Res. 1981, 14, 363. (b) Gonzales, C.;
Schlegel, H. B. J. Chem. Phys. 1991, 95, 5853.
(20) A hypothetical 8-membered transition state has been quite recently
proposed as a rationale for the catalytic asymmetric Sakurai reaction
developed by Carreira.10i Duthaler, R. O.; Hafner, A. Angew Chem., Int.
Ed. Engl. 1997, 36, 43.
(21) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds;
Wiley: New York, 1994.

12134 J. Am. Chem. Soc., Vol. 119, No. 50, 1997
Bottoni et al.
Table 2. Absolute (E)a and Relative (∆E)b Energies of Reactants
and Transition States for the Model Reaction 1a + 2a + 3a
Computed at the B3LYP and MP2 Levels
accurate theoretical methods can give qualitatively different
results, we believe that our finding can shed light on the
mechanism of this important reaction. In particular we have
found that (i) no six-membered cyclic or antiperiplanar transition
states exist on the reaction surface (ii) the addition of allylsilane
on a BF3-complexed aldehyde occurs through eight-membered
cyclic transition states with synclinal arrangement of the reacting
CdC and CdO double bonds. The formation of the new C-C
and Si-F bonds and the breaking of the Si-C and B-F bonds
occur in a highly asynchronous, concerted manner and lead to
the fluorosilane 6 and the homoallylborate 5 experimentally
observed. These products represent the thermodynamically most
favorable state of the system; thus, the lack in regeneration of
the Lewis acid prevents the catalysis.
Even if four transition states have been located on the reaction
surface, we suggest that, in the real experimental conditions,
the two most likely reaction channels are those involving the
anti-SE′ diastereotopic transition states A and B where the
repulsive interaction between the trialkylsilyl group and the
fluorine atoms of the Lewis acid can be minimized. These two
transition states, whose relative stability appears to be deter-
mined by the nature and position of the substituents, easily
explain the stereochemical features experimentally observed for
this reaction
B3LYP
MP2
E
∆E
0.00
8.03
11.48
11.32
7.12
E
∆E
0.00
9.97
19.23
14.09
10.35
reactants
A
B
C
-688.34948
-688.33668
-688.33118
-688.33144
-688.33813
-688.40494
-686.12382
-686.10793
-686.09318
-686.10136
-686.10733
-686.18666
D
products
-34.80
-39.43
a Hartrees. b kcal mol-1
.
is evident that the reaction is quite exothermic (-34.80 kcal
mol-1). Furthermore, structures A and D have the lowest
activation energies (8.03 and 7.12 kcal mol-1, respectively)
while B and C are higher in energy and are almost degenerate
(11.48 and 11.32 kcal mol-1, respectively). In the same table
we have also reported the energy values obtained at the MP2
level on the B3LYP optimized geometries. At the MP2 level
the reaction is still highly exothermic (-39.43 kcal mol-1); even
if the energetic order of the four transition states remains
approximately the same (A and D are still characterized by the
lowest activation energies: 9.97 and 10.35 kcal mol-1 respec-
tively), the activation energy of B becomes much larger (19.23
kcal mol-1). However, these values should be considered with
great care since no geometry optimization was performed at
the MP2 level.
It is reasonable to believe that the energetic order of the four
structures A, B, C, and D can change significantly when
hydrogens on allylsilane are substituted with larger groups. In
particular a critical interaction should occur in the real systems
between the trimethyl(or larger)silyl group SiR3 and the fluorine
atoms of the Lewis acid. This interaction should rise the energy
of both structures C and D with respect to A and B because of
the syn orientation of the silyl group, which makes it more
difficult to accommodate cumbersome substituents on the Si
atom. It is interesting to point out that the most likely geometry
transformation that can minimize the repulsions in C and D is
a rotation around the C-C single bond of allylsilane to reach
an anti orientation of SiR3. On the contrary, in A and B a further
rotation around the C-C bond can lower the repulsive interac-
tions but leaves unchanged the anti orientation of the silyl group.
This hypothetical behavior can be a rationale for the anti
preference found by Kumada6a and Denmark.7 The two anti-
SE′ diastereotopic transition states A and B, which would be
associated to the most likely reaction channels in real systems,
easily explain the syn diastereoconvergence of the Sakurai
reaction: if the reaction proceeds through transition state A the
syn- or anti-homoallylic alcohols can be obtained from the (E)-
or (Z)-crotylsilane, respectively. On the contrary, if transition
state B is involved, the syn or anti products are obtained from
the Z- or E-isomer, respectively. Thus the stereochemical
outcome of the reaction could depend on the relative energy of
A and B, which will be determined by the nature of substituents
in real systems.22
Experimental Section
1H NMR, 13C NMR, and 29Si NMR spectra were obtained with a
Varian Gemini-300 instrument at 300, 75.5, and 59.6 MHz, respectively,
and are reported in δ units, referenced to TMS (internal standard). IR
spectra were recorded with a Nicolet 205 FT instrument and are reported
in cm-1. MS spectra were obtained with the GC-MS technique on a
Hewlett-Packard 5890 analyzer with EI at 70 eV and are given as m/e
(rel intensity). All the reactions were carried out under Ar atmosphere
in anhydrous pentane (Na) or CH2Cl2 (CaH2). Aldehydes were distilled
prior to use; other chemicals were used as received (98% pure or better).
Allylation of Benzaldehyde. A solution of 1b (0.102 mL, 1 mmol)
and 3b (0.127 mL, 1 mmol; or 0.025 mL, 0.2 mmol) in anhydrous
CH2Cl2 was cooled at the desired temperature (Table 1). After 10 min,
2b (0.160 mL, 1 mmol) was added with stirring. When the time
reported in Table 1 was passed, the reaction was quenched with
Na2HPO4/NaH2PO4 aqueous buffer solution (pH 7, 3 mL), extracted
with ether, and dried (Na2SO4). The products obtained were purified
by silica gel chromatography.
Allylation of Octanal. Premixing of 1c and 3b caused the formation
of a complex scarcely soluble in either CH2Cl2 or pentane; therefore
inverse addition was adopted. A solution of 1c (0.156 mL, 1 mmol)
and 2b (0.160 mL, 1 mmol) in CH2Cl2 or pentane (3 mL) was cooled
at the desired temperature (Table 1). Then 3b (0.127 mL, 1 mmol; or
0.025 mL, 0.2 mmol) was added with stirring. Quenching and workup
procedures were the same as for 1a.
r-(2-Propen-1-yl)benzenemethanol (4b) and 1-undecen-4-ol (4c)
showed identical spectroscopic properties to those previously reported.10a
Ethers 7b. First isomer: MS 237 (15), 195 (2), 131 (100), 116
(41), 91 (99), 77 (45). IR 3078, 3029, 2980, 2931, 2910, 1664, 1602,
1489, 1454, 1363, 1307, 1082, 990, 913, 765, 702. 1H NMR 2.38-
2.62 (m, 4H), 4.10 (t, 2H, J ) 6.3 Hz), 4.95 (m, 4H), 5.60 (m, 2H),
7.25 (m, 10H). 13C NMR 42.8, 77.4, 116.6, 126.7, 127.5, 127.9, 134.9,
142.0. Second Isomer: 1H NMR 2.22-2.35 (m, 4H), 4.40 (t, 2H, J
) 6.3 Hz), 4.90-5.10 (m, 4H), 5.80 (m, 2H), 7.25 (m, 10H). 13C NMR
41.6, 77.0, 117.0, 127.1, 128.0, 128.2, 134.7, 142.0.
Conclusions
Spectroscopic Investigations. For 29Si NMR and 13C NMR inves-
tigations the reaction mixture of 1b or 1c, 2b, and 3b was prepared in
an NMR tube under Ar atmosphere at -80 °C in anhydrous CH2Cl2,
containing 10% C6D6 for the lock, immediately before use, and quickly
transferred into the NMR instrument set at the desired temperature.
Recording of 29Si NMR spectra showed the signals of allyltrimethyl-
silane (2b) (0.71, s) that gradualy dropped and that of fluorotrimeth-
ylsilane (6b) (34.26, d, J ) 273 Hz) that gradualy incrased in both
In this paper we have studied from both computational and
experimental points of view the mechanism of the Sakurai
reaction. Even if we cannot completely rule out that other more
(22) For example, the 3-methyl group of a (Z)-crotylsilane, which points
toward the inner part of the cyclic structure, would probably cause a stronger
interaction with the O-BF2 group in the tighter A than in the looser B
structure.

EVidence for the Mechanism of the Sakurai Reaction
J. Am. Chem. Soc., Vol. 119, No. 50, 1997 12135
cases. Recording of 13C NMR spectra of a mixture of 1c, 2b, and 3b
showed 2b (-2.2, 24.8, 112.5, 135.5), 6b (-0.08, d, J ) 14.9 Hz),
free (15.4, 65.9) and boron-complexed (13.3, 69.9) diethyl ether, the
borylated homoallylic product 5c (14.4, 23.2, 25.8, 29.7, 29.9, 32.2,
33.8, 39.3, 78.3, 118.9, 133.0), and another set of signals14 (14.1, 22.9,
25.7, 29.5, 29.7, 32.1, 36.1, 41.2, 73.4, 117.3, 135.1) in a minor amount.
For comparison we report the spectra of 8c (1.1, 14.2, 23.0, 26.0, 29.6,
29.9, 32.2, 37.2, 42.3, 70.9, 117.6, 135.6) and of 4c (14.2, 23.0, 26.0,
29.6, 29.9, 32.1, 37.2, 42.3, 70.9, 117.5, 135.5) obtained under the same
conditions.
Acknowledgment. We thank the M. U. R. S. T. (Rome)
and the University of Bologna (“Founds for Selective Research
Topics”), for financial support.
JA9723693