New synthetic routes for silaheterocycles: Reactions of a chlorosilylenoid with aldehydes

Article abstract of DOI:10.1021/om9008789

(Tsi)chlorosilylenoid [Tsi = C(SiMe₃)₃] reacted with acetaldehyde and benzaldehyde to give 2,4dioxasilolane and 2,5-dioxasilolane, respectively. The direct addition of an aldehyde to a silaoxirane intermediate is proposed as a plausible mechanism based on a theoretical study. These results show a new synthetic application of the chlorosilylenoid and a new route for the synthesis of silaheterocycles.

Full text of DOI:10.1021/om9008789

Organometallics 2010, 29, 1355–1361 1355
DOI: 10.1021/om9008789
New Synthetic Routes for Silaheterocycles: Reactions of a
Chlorosilylenoid with Aldehydes
Young Mook Lim,† Chang Hee Park,† Soo Jin Yoon,† Hyeon Mo Cho,† Myong Euy Lee,*,†
and Kyoung Koo Baeck*,‡
†Department of Chemistry & Medical Chemistry, College of Science and Technology, Yonsei University,
Wonju, Gangwondo 220-710, Korea and Department of Chemistry, College of Natural Science,
‡
Gangneung-Wonju National University, Gangneung, Gangwondo, 210-702, Korea
Received October 9, 2009
(Tsi)chlorosilylenoid [Tsi = C(SiMe3)3] reacted with acetaldehyde and benzaldehyde to give 2,4-
dioxasilolane and 2,5-dioxasilolane, respectively. The direct addition of an aldehyde to a silaoxirane
intermediate is proposed as a plausible mechanism based on a theoretical study. These results show a
new synthetic application of the chlorosilylenoid and a new route for the synthesis of silaheterocycles.
Introduction
C(SiMe3)3, X = Br, Cl), (Mes)2Si(SMes)Li,7 and (R3Si)2SiFLi
(R3Si=t-Bu2MeSi).8 Such stable silylenoids bearing a halo-
gen atom, which we reported, have been particularly im-
portant in the studies of their various reactivity, such as
reduction, substitution, and transmetalation reactions.6 In
this context, we now report some unprecedented examples
of the reactivity of a halosilylenoid with aldehydes. These
reactions of halosilylenoid can be used as new synthetic
routes for functional silaheterocycles. In general, hetero-
cycles have efficiently been used in biology, pharmacology,
and electronic applications.9 Silicon isosteres of various
heterocycles would be regarded as potential candidates for
drug discovery and development of medicinal chemistry.10
The reactivities of silylenoids are often similar to those of
silylenes. Reactions of silylenes with aldehyde compounds11
have been studied by a few research groups.12 Jutzi and co-
workers reported the reaction of decamethylsilicocene with
aldehydes such as benzaldehyde and cinnamaldehyde to
give the respective dioxasilolane derivatives, in which a
silaoxirane formed from [2þ1] cycloaddition of the silylene
with CdO was suggested to be a key intermediate.11a The
Komatsu group photochemically synthesized dioxasilolanes
Over the past few decades, a number of publications have
appeared on novel silylenoids (R2SiMX), compounds in
which an electropositive metal (M) and a leaving group (X)
are bound to the same silicon atom.1-5 Recently, a few stable
silylenoids have been reported, including TsiSiX2Li6 (Tsi =
*Corresponding authors. (M.E.L.) Tel: 82-33-760-2237. Fax: 82-33-
760-2182. E-mail: melgg@yonsei.ac.kr. (For theoretical part, K.K.B.)
Tel: 82-33-640-2307. Fax: 82-33-640-2264. E-mail: baeck@gwnu.ac.kr.
(1) For other experimental studies on silylenoids, see: (a) Oehme, H.;
Weiss, H. J. Organomet. Chem. 1987, 319, C16. (b) Boudjouk, P.;
Samaraweera, U. Angew. Chem., Int. Ed. Engl. 1988, 27, 1355. (c) Tamao,
K.; Kawachi, A. Organometallics 1995, 14, 3108. (d) Kawachi, A.; Doi, N.;
Tamao, K. J. Am. Chem. Soc. 1997, 119, 233. (e) Tamao, K.; Kawachi, A.;
Asahara, M.; Toshimitsu, A. Pure Appl. Chem. 1999, 71, 393. (f) Wiberg, N.;
Niedermayer, W. J. Organomet. Chem. 2001, 628, 57. (g) Driver, T. G.;
Franz, A. K.; Woerpel, K. A. J. Am. Chem. Soc. 2002, 124, 6524.
(h) Wiberg, N.; Niedermayer, W.; Fischer, G.; Noth, H.; Suter, M. Eur. J.
Inorg. Chem. 2002, 1066. (i) Sekiguchi, A.; Lee, V. Y.; Nanjo, M. Coord.
Chem. Rev. 2000, 210, 11.
(2) For the theoretical studies on silylenoids, see: (a) Feng, S.; Feng,
D. J. Mol. Struct. (THEOCHEM) 2001, 541, 171. (b) Feng, S.; Feng, D.;
Li, J. Chem. Phys. Lett. 2000, 316, 146. (c) Clark, T.; Schleyer, P. v. R.
J. Organomet. Chem. 1980, 191, 347. (d) Tanaka, Y.; Hada, M.; Kawachi,
A.; Tamao, K.; Nakatsuji, H. Organometallics 1998, 17, 4573. (e) Feng, S.;
Zhou, Y.; Feng, D. J. Phys. Chem. A 2003, 107, 4116. (f) Feng, D.; Xie, J.;
Feng, S. Chem. Phys. Lett. 2004, 396, 245. (g) Flock, M.; Marschner, C.
Chem.;Eur. J. 2005, 11, 4635.
(9) Joule, J. A.; Mills, K. Heterocyclic Chemistry; Blackwell Science:
Oxford, 2000.
(3) Tamao, K.; Kawachi, A. Angew. Chem., Int. Ed. Engl. 1995, 34,
818.
(4) (a) Tokitoh, N.; Hatano, K.; Sadahiro, T.; Okazaki, R. Chem.
Lett. 1999, 931. (b) Hatano, K.; Tokitoh, N.; Takagi, N.; Nagase, S. J. Am.
Chem. Soc. 2000, 122, 4829. (c) Tajima, T.; Hatano, K.; Sasaki, T.; Sasamori,
T.; Takeda, N.; Tokitoh, N.; Takagi, N.; Nagase, S. J. Organomet. Chem.
2003, 686, 118.
(5) (a) Tamao, K.; Asahara, M.; Saeki, T.; Toshimitsu, A. Angew.
Chem., Int. Ed. 1999, 38, 3316. (b) Tamao, K.; Asahara, M.; Saeki, T.;
Toshimitsu, A. J. Organomet. Chem. 2000, 600, 118.
(6) (a) Lee, M. E.; Cho, H. M.; Lim, Y. M.; Choi, J. K.; Park, C. H.;
Jeong, S. E.; Lee, U. Chem.;Eur. J. 2004, 10, 377–381. (b) Lim, Y. M.;
Cho, H. M.; Lee, M. E.; Baeck, K. K. Organometallics 2006, 25, 4960.
(c) Lee, M. E.; Lim, Y. M.; Son, J. Y.; Seo, W. G. Chem. Lett. 2008, 680.
(7) Kawachi, A.; Oishi, Y.; Kataoka, T.; Tamao, K. Organometallics
2004, 23, 2949.
(10) Showell, D. A.; Mills, J. S. Drug Discovery Today 2003, 8, 551.
(11) (a) Jutzi, P.; Eikenberg, D.; Bunte, E.; Mohrke, A.; Neumann, B.;
Stammler, H. Organometallics 1996, 15, 1930. (b) Sakai, N.; Fukushima, T.;
Minakata, S.; Ryu, I.;Komatsu, M.Chem. Commun. 1999, 1857. (c) Sakai, N.;
Fukushima, T.; Okada, A.; Ohashi, S.; Minakata, S.; Komatsu, M. J. Orga-
nomet. Chem. 2003, 686, 368.
(12) For the studies on silylenes with ketones, see: (a) Ando, W.;
Ikeno, M.; Sekiguchi, A. J. Am. Chem. Soc. 1977, 99, 6447. (b) Ando, W.;
Ikeno, M.; Sekiuchi, A. J. Am. Chem. Soc. 1978, 100, 3613. (c) Ando, W.;
Ikeno, M. Chem. Lett. 1978, 609. (d) Ando, W.; Ikeno, M. J. Chem. Soc.,
Chem. Commun. 1979, 655. (e) Ishikawa, M.; Nishimura, K.; Sugisawa, H.;
Kumada, M. J. Organomet. Chem. 1980, 194, 147. (f) Ando, W.; Hamada,
Y.; Sekiguchi, A.; Ueno, K. Tetrahedron Lett. 1982, 23, 5323. (g) Ando, W.;
Hamada, Y.; Sekiguchi, A. J. Chem. Soc., Chem. Commun. 1983, 952.
(h) Belzner, J.; Ihmels, H.; Pauletto, L.; Noltemeyer, M. J. Org. Chem. 1996,
61, 3315. (i) Gehrhus, B.; Hitchcock, P. B.; Lappert, M. F. Organometallics
1997, 16, 4861. (j) Becerra, R.; Cannady, J. P.; Walsh, R. J. Phys. Chem. A
2002, 106, 11558.
€
(8) Molev, G.; Bravo-Zhivotovskii, D.; Karni, M.; Tumanskii, B.;
Botoshansky, M.; Apeloig, Y. J. Am. Chem. Soc. 2006, 128, 2784.
r
2010 American Chemical Society
Published on Web 02/18/2010
pubs.acs.org/Organometallics

1356 Organometallics, Vol. 29, No. 6, 2010
Lim et al.
Scheme 1. Possible Intermediates: Silaoxirane (A) and Silacar-
bonyl Ylides (B and C)
Scheme 2. Reactions of a Chlorosilylenoid with Aldehydes
greater hindrance of the two phenyl groups, leading to
formation of a single isomer in the cycloaddition reaction.
Interestingly the Si-Cl bond in 4 did not react with MeOH,
suggesting severe steric hindrance of this compound.
As shown in Scheme 1, the silaoxirane and silacarbonyl
ylide generated from the cleavage of the Si-C bond of the
silaoxirane are thought to be candidates for key intermedi-
ates in these reactions based on the previous works.11,12
However, little theoretical work has been conducted so far
to directly investigate the mechanisms involving the silaox-
irane and silacarbonyl ylide intermediates. We carried out a
theoretical investigation to provide a reasonable mechanism
and further insights on new experimental findings.
Theoretical Computations and Discussion. The molecular
geometry, harmonic vibration frequency, and energy of the
transitions-state (TS) structures as well as reactants and
products are calculated by using two density-functional-
theory (DFT) methods. The B3LYP14 and the M05-2X15
DFT functional forms implemented in the Gaussian03 suite
of programs16 are selected in the present work, because
B3LYP is one of the most widely used functional form,
whereas M05-2X has recently been claimed to be one of
the best functional forms for activation energies.15 Though
the standard 6-311þG(2d) basis sets are employed in some
test calculations to check the dependence of relative energies
on the size of basis sets, the main discussions will be based on
the results with the standard 6-31þG(d) basis sets. As will be
shown below, the results with the 6-311þG(2d) basis sets
did not bring any significant change in both qualitative
and quantitative aspects. The tight criteria options of the
Gaussian03 are applied to both the SCF iterations and
the geometry optimization procedures. Six Cartesian d-type
functions are employed in the 6-31þG(d) sets, whereas five
spherical d-type functions are used in the 6-311þG(2d) sets.
The atomic partial charges are determined by the natural
bond orbital analysis (NBO version 3)17 with the electron
by a trapping reaction of dimesitylsilylene with aldehydes, in
which silacarbonyl ylide was a key intermediate.11b
In the following we describe the reactions of a chlorosily-
lenoid with acetaldehyde and benzaldehyde and also theo-
retical computations to provide reasonable mechanisms
considering silaoxirane (A) and silacarbonyl ylide (B and
C) as intermediates (Scheme 1).
Results and Discussion
Syntheses and Characterizations. Trichloro[tris(trimethyl-
silyl)methyl]silane (1), containing the bulky Tsi group, was
prepared in high yield as described previously.6 LiNp
(lithium naphthalenide) in THF was added slowly to com-
pound 1 in THF at -78 °C. The solution was stirred for 12 h
at the same temperature; all the starting reactants were
consumed to give the chlorosilylenoid 2, which was mon-
itored by GC/MS. To the resulting dark brown solution was
added an excess of acetaldehyde at -78 °C, whereupon the
solution rapidly became light yellow, indicating that the
reaction was complete. Treatment of the reaction mixture
with an excess of MeOH gave the trapping product
2,4-dioxasilolane (3) in 83% GC yield as a mixture of four
isomers (Scheme 2).13 From the crude material one of
the isomers, 3, was isolated by silica gel chromatography
(n-hexane) as a colorless oil in 28% yield and was character-
ized by NMR, GC/MS, EA, and HRMS. Due to the non-
symmetrical structure, 3 showed different resonances for the
nonequivalent CH3 (1H NMR; 1.36, 1.40 ppm; 13C NMR
16.4, 22.9 ppm) and CH (1H NMR 3.41, 5.17 ppm; 13C NMR
67.6, 96.3 ppm) in the NMR spectra.
The reaction of 2 with benzaldehyde was carried out using
similar procedures to those described above. Treatment of
the reaction mixture with excess MeOH gave the product
2,5-dioxasilolane (4) in 80% GC yield. After separation
using column chromatography, 4 was isolated as a colorless
oil in 55% yield and characterized by GC/MS, NMR, EA,
and HRMS.
Although four isomers of 4 were expected to be formed
considering product 3, the NMR spectra of 4 indicated that
only one isomer was produced. This may result from the
(14) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Hermanns, J.;
Schmit, B. J. Chem. Soc., Perkin Trans. 1998, 2209. (c) Hermanns, J.;
Schmit, B. J. Chem. Soc., Perkin Trans. 1999, 81. (d) Chinchila, R.; Nꢀajera,
C.; Yus, M. Chem. Rev. 2004, 104, 2667. (e) Lee, C.; Yang, W.; Parr, R. G.
Phys. Rev. B 1988, 37, 785.
(15) Zhao, Y.; Truhlar, D. G. Acc. Chem. Rev. 2008, 41, 157.
(16) Pople, J. A.; et al. et al. Gaussian03, Revision E.01; Gaussian, Inc.:
Wallingford, CT, 2004 (see Supporting Information for program's authors ).
(17) (a) Carpenter, J. E.; Weinhold, F. J. Mol. Struct. (THEOCHEM)
1998, 169, 41. (b) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988,
88, 889.
(13) The four isomers appeared in the GC/MS due to the different up
and down position of methyl and methoxy substituents in the ring (Tsi-
Me-Me: cis-cis, cis-trans, trans-trans, trans-cis), but unfortunately each
isomer could not be isolated individually.

Article
Organometallics, Vol. 29, No. 6, 2010 1357
Figure 1. The model compounds used in theoretical computations.
Scheme 3. Proposed Direct Mechanism
Figure 2. Relative energy profile for the reaction with acetalde-
hyde.
compounds with a C(SiH3)3 group instead of a Tsi group,
sila-carbonyl ylide is about 20 kcal/mol higher energy than
silaoxirane (more details are given elsewhere).18 Considering
the low temperature (-78 °C) of the present experiments, a
mechanism based on the silaoxirane seems more reasonable
than that based on sila-carbonyl ylide species. Moreover, all
of our efforts to generate computational results supporting
any plausible mechanism starting from sila-carbonyl ylide
were in vain. Therefore, we propose the following simpler
mechanism: the direct addition of aldehyde to silaoxirane, as
shown in Scheme 3.
In order to study the possibility and plausibility of the
direct addition mechanism, the molecular structures and
energies of reactants (silaoxirane and aldehyde), possible
products (2,5-dioxasilolane or 2,4-dioxasilolane), and the
corresponding TS structures were calculated, and the ener-
getic relationship is depicted in Figures 2 and 5 for the
reactions of 2 with acetaldehyde and benzaldehyde, respec-
tively. The characteristics of the TSs are confirmed not only
by observing the one imaginary vibration frequency but also
by obtaining the structures of the initial encounters of
reactants as well as the final products, as will be discussed
in detail below.
As shown in Figure 2, the reaction energy from reactants to
5Me is calculated to be -49.6 (-35.4) kcal/mol by the M05-
2X/6-31þG(d) (B3LYP/6-31þG(d)) method, and the magni-
tude is reduced a little by using a larger basis set: -48.5
(-33.5) kcal/mol by the M05-2X/6-311þG(2d) (B3LYP/6-
311þG(2d)) method. Meanwhile, the exothermic energy to
form 4Me is calculated to be -81.3 (-65.0) kcal/mol by the
M05-2X/6-31þG(d) (B3LYP/6-31þG(d)) method, and the
magnitude is almost not changed even with a larger basis
set: -81.4 (-65.0) kcal/mol by the M05-2X/6-311þG(2d)
(B3LYP/6-311þG(2d)) method. The exothermic energy of
4Me is much larger than (almost twice of) that of 5Me,
which implies that the cleavage of the Si-C bond of silaox-
irane is more favored thermodynamically, as expected.11,12
The experimental results, however, contradict the above;
the 2,5-dioxasilorane corresponding to 4Me was not pro-
duced, but 3 is actually obtained after the reaction of MeOH
with the compound corresponding to 5Me. Because the
density calculated by the M05-2X/6-311þþG(3df,2pd) met-
hod at the geometry optimized by the M05-2X/6-31þG(d)
method. The use of the NBO method as well as the very
large 6-311þþG(3df,2pd) basis sets provides less variation
of the atomic partial charges depending on methodologies.
The zero-point energy (ZPE), included in all of the relative
energies in the present work, is obtained by the calculations
of harmonic frequencies at stationary structures by exactly
the same method used in the geometry optimization, and no
scaling factor was used.
In order to reduce computational resources, the Tsi group
of actual experiments is replaced by a C(SiH3)3 group in all of
the present computations, and the model molecules with the
C(SiH3)3 group will be designated by 5Me, 5Ph, 4Me, and
4Ph, as depicted in Figure 1.
The first thing uncovered by the present computation is
the fact that the two sila-carbonyl ylide structures, B and C in
Scheme 1, are not distinguishable, at least within the level of
DFT methods used in the present work. A more important
fact is that the silaoxirane structure (A in Scheme 1) is more
stable than the sila-carbonyl ylide structure (B and C in
Scheme 1); according to the calculations with the model
(18) The total energy (including ZPE) of silaoxirane, (H3Si)3C-ClSi-O-
CHMe with trans-Tsi-Me configurations, is -1815.2829 (-1815.4540) au
by the M05-2X/6-31þG(d) (B3LYP/6-31þG(d)) method, and the corre-
sponding silacarbonyl ylide has 23.7 (19.8) kcal/mol higher energy. The
energy of silaoxirane, (H3Si)3C-ClSi-O-CHPh with trans-Tsi-Ph config-
uration, is -2006.9620 (-2007.1455) au by the M05-2X/6-31þG(d)
(B3LYP/6-31þG(d)) method, and the corresponding sila-carbonyl ylide
has 23.2 (16.4) kcal/mol higher energy. The energy difference between the
two isomers (cis- and trans-Tsi-R) of silaoxirane is less than 2 kcal/mol for
both R = Me and R = Ph cases, whereas the difference between the two
isomers of sila-carbonyl ylide is about 3 and 6 kcal/mol for R = Me and
R = Ph, respectively. Detailed structures, energies, and atomic coordi-
natesof silaoxiranesandsila-carbonylylides(trans-Tsi-Risomer) aregiven
in the Supporting Information.

1358 Organometallics, Vol. 29, No. 6, 2010
Lim et al.
Figure 3. Interatomic distances (in pm) and atomic charges (in e) of TS-5Me (left) and its initial associate of reactants (right). The
normal mode displacement vectors of the imaginary frequency of TS-5Me are shown by black arrows.
Figure 4. Interatomic distances (in pm) and atomic charges (in e) of TS-4Me (left) and its initial associate of reactants (right).
q
thermodynamic viewpoint is in contrast with actual experi-
mental results, the kinetic aspect, i.e., the reaction barrier, has
to be investigated.
changed by the increase of the basis set size; ΔE0 (TS-4Me) is
10.8 (21.4) kcal/mol by the M05-2X/6-311þG(2d) (B3LYP/
6-311þG(2d)) method. The main features of the TS-5Me
and the TS-4Me, without H atoms for clarity, are shown in
the left part of Figures 3 and 4; more details of the structures
and energies of the TSs and products are given elsewhere.19
q
The activation energy (ΔE0 ) from the reactants (silaoxi-
rane and acetaldehyde) to the transition state (TS-5Me)
leading to 2,4-dioxasilolane 5Me is calculated to be 8.6
(18.6) kcal/mol by the M05-2X/6-31þG(d) (B3LYP/
6-31þG(d)) method. The value is not much changed by
increasing the size of the basis set to a larger triple-ζ quality
q
According to Figure 2, ΔE0 (TS-5Me) is 2-3 kcal/mol lower
than ΔE0 (TS-4Me). The energetic relationship shown in
Figure 2, however, is just one of several possible cases
because there are four isomers of 4Me. If the other stereo-
isomers of the TS-5Me and the TS-4Me shown in Figures 3
and 4 are used in the calculations, the magnitudes of these
ΔE0 are slightly different, and the ΔE0 (TS-5Me) is 3-4
kcal/mol lower than the ΔE0 (TS-4Me). The results sum-
marized in Figure 2 correspond to the stereoisomer giving
the smallest difference between ΔE0 (TS-5Me) and ΔE0 -
(TS-4Me).
The simple sum of the energies of the two separated
reactants is taken as the reference in the calculations of
q
q
set; ΔE0 (TS-5Me) is 9.1 (19.6) kcal/mol by the M05-2X/
6-311þG(2d) (B3LYP/6-311þG(2d)) method. The other
activation energy to TS-4Me leading to 2,5-dioxasilolane
4Me is 10.9 (21.6) kcal/mol by the M05-2X/6-31þG(d)
(B3LYP/6-31þG(d)) method, and the value is almost not
q
q
q
q
q
(19) Total energies (including ZPE) of acetaldehyde, (H3Si)3C-ClSi-
O-HMe (silaoxirane with trans-(H3Si)3C-Me), TS-5Me, 5Me, TS4-Me,
and 4Me are -153.7476 (-153.7841), -1815.2829 (-1815.4540),
-1969.0167 (-1969.2085), -1969.1096 (-1969.2944), -1969.0132
(-1969.2037), and -1969.1600 (-1969.3417), respectively, by the M05-
2X/6-31þG(d) (B3LYP/6-31þG(d)) method. The same values by the
M05-2X/6-311þG(2d) (B3LYP/6-311þG(2d)) method are -153.7894
(-153.8241), -1815.4758 (-1815.6313), -1969.2505 (-1969.4241),
-1969.3419 (-1969.5087), -1969.2482 (-1969.4213), and -1969.3949
(-1969.6043), respectively. Detailed structures are given in the Support-
ing Information.
q
the activation energies, and the difference between ΔE0 of
the same TS by the M05-2X and the B3LYP is 10.0 and
10.7 kcal/mol for Ts-5Me and TS-4Me, respectively. If the
energy of the initial associates of reactants is taken as the
reference of the activation barriers, then the above dependence

Article
Organometallics, Vol. 29, No. 6, 2010 1359
q
of ΔE0 on the DFT functional form is reduced to about
q
5 kcal/mol, but the difference between ΔE0 (TS-5Me) and
q
ΔE0 (TS-4Me) does not depend on the choice of the refer-
ence. However, the configuration of the initial associate for
TS-5Me (the right part of Figure 3) differs from that for
TS-4Me (the right part of Figure 4). The interaction between
the O atom of aldehyde and the Si atom of the silaoxirane
ring seems to be the main interaction forming both of the
initial associates of TS-5Me and TS-4Me. The initial associ-
ate of TS-5Me has an additional interaction between the H
atom of the aldehyde and the O atom of the silaoxirane,
whereas that of TS-4Me has the additional interaction
between the H atom of the aldehyde and the Cl atom of
the silaoxirane. The initial associate of TS-5Me was calcu-
lated to be 2.4 (0.8) kcal/mol lower than the initial associate
of TS-4Me by the M05-2X/6-31þG(d) (B3LYP/6-31þG(d))
method, and the magnitude slightly reduced to 2.1 (0.7)
kcal/mol by using the larger 6-311þG(2d) basis sets. The
lower energy of the initial associate of TS-5Me could be an
additional advantage of the reaction path via TS-5Me, but
several delicate considerations have to be noted about the
degree of the advantage, and such complicated and too
detailed considerations are refrained from here, partially
because all such details do not change any qualitative aspect
given below.
Figure 5. The relative energy profile for the reaction with
benzaldehyde.
In spite of some variations of the activation energies
depending on the functional form of DFT, the size of the
basis sets, and the choice of the reference of the relative
q
energies, the important point is the fact that ΔE0 (TS-5Me)
q
is at least 2 kcal/mol lower than ΔE0 (TS-4Me) regardless of
Figure 6. Structure, atomic charges, and normal mode displa-
cements of TS-5Ph (left) and its initial associate of reactants
(right).
the methodology. This means that 5Me is kinetically more
favorable than 4Me, while 4Me is thermodynamically more
stable than 5Me. Though the magnitude of the difference
(2 kcal/mol) is not large enough to be conclusive evidence for
the domination of the kinetic factor over the thermodynamic
factor, it can be one possible explanation for the actual
experimental results because of the very low reaction tem-
perature, -78 °C. Considering the significant yields (∼80%)
of rapidly completed reactions at such a low reaction tem-
perature, the activation barrier computed by the B3LYP
functional (18.6 kcal/mol) is rather too high, whereas the
magnitude by the M05-2X functional (8.6 kcal/mol) seems
more reasonable.
As an important characteristic of TS structures, only one
imaginary frequency is observed in the TS structures; the
values of the imaginary frequencies of TS-5Me and TS-4Me
by the M05-2X/6-31þG(d) (B3LYP/6-31þG(d)) method are
390i (335i) and 368i (365i) cm-1, respectively. The magni-
tudes of the imaginary frequency are not much changed even
by calculations with a larger size of basis sets: 385i (332i) and
375i (353i) cm-1 by the M05-2X/6-311þG(2d) (B3LYP/
6-311þG(2d)) method for TS-5Me and TS-4Me, respec-
tively. The normal mode displacement vectors of the ima-
ginary frequency, represented by arrows in Figures 3 and 4
for TS-5Me and TS-4Me, respectively, show that the nuclear
motion along the imaginary frequency would lead to the
product structure 5Me and 4Me, respectively. We actually
confirmed that not only the corresponding product 5Me or
4Me but also the initial associate structures of reactants are
obtained by the geometry optimization after a few steps of
forward and backward nuclear movements along the intrin-
sic reaction coordinate (IRC) corresponding to the imagin-
ary frequency. It is now clear that TS-5Me and TS-4Me are
the real transition-state structures leading to 5Me and 4Me,
Figure 7. Structure, atomic charges, and normal mode displa-
cements of TS-4Ph (left) and its initial associate of reactants
(right).
respectively, by a single-step addition of acetaldehyde to
silaoxirane.
The energetic relationship shown in Figure 2 also provides
an explanation why not 5 but 3 is the final product in the
actual experiment; 5Me is thermodynamically not stable
enough compared with 4Me, and therefore 5Me converts to
3 by the reaction with MeOH. The magnitude of the further
stabilization energy due to the reaction of 5Me with MeOH is
not actually computed here due to the difficulties and un-
certainties in the calculation of the solvated states of Hþ and
Cl- ions expected to be generated in actual reaction media.
The thermodynamic and kinetic aspects of the direct
addition mechanism of benzaldehyde to silaoxirane are also

1360 Organometallics, Vol. 29, No. 6, 2010
Lim et al.
Figure 8. Molecular structure (without H atoms) and atomic charges of two silaoxiranes.
q
q
calculated by the same methods as used in the above acet-
aldehyde case, but only the 6-31þG(d) basis sets are used
here, partially because the 6-311þG(2d) sets are too large for
compounds containing two phenyl groups and partially
because the changes caused by using the larger basis sets
are not expected to be significant, as seen in the above case of
acetaldehyde.
is 1.4 kcal/mol lower than ΔE0 (TS-5Me), whereas ΔE0 (TS-
q
4Ph) is 6.1 kcal/mol lower than ΔE0 (TS-4Me). The big
change causes the totally different experimental results for
acetaldehyde and benzaldehyde.
The partial atomic charges in Figures 3, 4, 6, and 7 provide
a clue where the difference stems from. The partial charges of
TS-5Me are not much different from those of TS-5Ph;
the magnitudes of atomic charge (in units of electron) of
O (of aldehyde), C (carbonyl carbon of aldehyde), O (of silaoxi-
rane), and C (of silaoxirane ring) in TS-5Me are -0.80, þ0.60,
-0.94, and -0.35, respectively, whereas the values in TS-5Ph
are -0.80, þ0.59, -0.91, and -0.37, respectively. The changes
from TS-5Me to TS-5Ph are almost negligible in spite of the
big difference between methyl and phenyl groups. Due to this
The energetic relationship shown in Figure 5 uncovers
different features compared with those in Figure 2. In the
q
case of benzaldehyde, the ΔE0 of TS-4Ph is about 2-3
q
kcal/mol lower than the ΔE0 of TS-5Ph, and 4Ph is also
more stable than 5Ph. In other words, the reaction path to
4Ph is both kinetically and thermodynamically more favor-
able than the path to 5Ph. Both the thermodynamic and
the kinetic aspects are in accordance with the fact that 4 is the
major product in actual experiments. The large magni-
tude of the exothermic reaction energy (-77.6 kcal/mol by
the M05-2X/6-31þG(d) method) in the formation of 4Ph
also provides an explanation why 4 does not undergo any
further reaction with MeOH; 4 is stable enough and does not
need a further stabilization, whereas 5 is not stable enough
and needs a conversion to 3 by the reaction with MeOH in
actual experiments.
The main features of the TS-5Ph and the TS-4Ph struc-
tures are shown in Figures 6 and 7. The only imaginary
frequencies of TS-5Ph and TS-4Ph by the M05-2X (B3LYP)
method are 236i (207i) and 405i (369i) cm-1, respectively.
The normal mode displacement vectors of the imaginary
frequency of TS-5Ph and TS-4Ph also show proper nuclear
motions, leading to the product structure 5Ph and 4Ph,
respectively. The geometries of 5Ph, 4Ph, and the initial
associates of reactants are also obtained by the direct
geometry optimization after a few steps of forward and
backward nuclear movements along the intrinsic reaction
coordinate of the imaginary frequency of the corresponding
TS structure. The molecular structures of the initial associ-
ates in Figures 6 and 7 are basically the same as the
corresponding ones in Figures 3 and 4. More details of the
energies and structures are given elsewhere.20 It is clear that
the TS-5Ph and TS-4Ph are the real TS structures leading to
5Ph and 4Ph, respectively, by the mechanism of a single step
addition.
q
small effect, the magnitude of ΔE0 (TS-5) is not much changed,
just 1.4 kcal/mol decreased by substituting Me with Ph.
On the other hand, the partial charges of TS-4Me show a
noticeable difference from those of TS-4Ph; the values are
-0.74, þ0.46, -0.31, and -0.93 in TS-4Me and -0.80, þ0.63,
-0.34, and -0.92 in TS-4Ph. By substituting the methyl with
phenyl, the positive charge on the carbonyl C of aldehyde
increased from þ0.46 to þ0.63, while the negative charge on
the C of the silaoxirane ring is also slightly intensified from
-0.31 to -0.34. The phenyl group seems to stabilize both
the positive and negative charge by the resonance effect of the
phenyl ring, and the activation energy is decreased from 10.9 to
4.8 kcal/mol by the effect. The partial charges of the oxygen
and the carbonyl carbon of an isolated acetaldehyde, calcu-
lated by the NBO-M05-2X/6-311þþG(3df,2pd)//M05-2X/
6-31þG(d) method, are -0.53 and þ0.45, respectively, and
the values remain almost the same even when the methyl is
replaced by phenyl, i.e., -0.54 and þ0.43, respectively, in
benzaldehyde. Figure 8 also shows that the bond lengths and
partial charges on the O and C atoms of the silaoxirane ring
remain almost the same even when the methyl is replaced by
phenyl. Therefore, the difference between the partial charge in
TS-4Me and the corresponding value in TS-4Ph can be
regarded as the result of different effects of the methyl and
the phenyl group on the developed partial charges during the
reaction processes.
According to the above analyses on atomic charges, the
methyl group cannot stabilize the developing partial atomic
charges, whereas the phenyl group can in the reaction path
to 2,5-dioxasilolane (4Me/4Ph). In the reaction path to
2,4-dioxasilolane (5Me/5Ph), on the other hand, the location
and the big negative charge of the O atom of the silaoxirane
(see Figures 3 and 6) seem to play a dominating role in
TS-4Me and TS-4Ph, but the methyl (phenyl) group of the
silaoxirane ring is not directly connected to the reacting O
atom. As a result, the effect of the phenyl group is rather small
One of the most noticeable change from Figure 2 to
q
Figure 5 is the big stabilization of TS-4Ph; ΔE0 (TS-5Ph)
(20) Total energies (including ZPE) of benzaldehyde, (H3Si)3C-ClSi-
O-HPh (silaoxirane with trans-(H3Si)3C-Ph), TS-5Ph, 5Ph, TS-4Ph, and
4Ph are -345.4280 (-345.4788), -2006.9620 (-2007.1455), -2352.
3785 (-2352.5971), -2352.4610 (-2352.6681), -2352.3807 (-2352.
6020), and -2352.5136 (-2351.7181), respectively by the M05-2X/6-
31þG(d) (B3LYP/6-31þG(d)) method. Detailed structures are given in
the Supporting Information.
q
in this case, and ΔE0 (TS-5) is reduced by just 1.4 kcal/mol by

Article
Organometallics, Vol. 29, No. 6, 2010 1361
replacing methyl with phenyl. This implies that the effect of
q
spectrometer and a high-resolution (JEOL JMS-600W-Agilent
6890 Series) instrument. The preparation of trichloro[tris(tri-
methylsilyl)methyl]silane, 1, and chlorosilylenoid 2 was repor-
ted in previous papers.7
any change of alkyl (aryl) group of aldehyde on ΔE0 (TS-5)
q
would be small, whereas the effect on ΔE0 (TS-4) would be
large, which will be studied in future experimental and
theoretical works.
Synthesis of 3. Naphthalene (1.28 g, 10 mmol) dissolved in
THF (40 mL) was added to Li (0.060 g, 8.7 mmol) at room
temperature. After stirring for 3 h at room temperature, LiNp
was obtained as a dark green solution. LiNp in THF was added
slowly to TsiSiCl3 (1.46 g, 4.0 mmol) in THF (40 mL) at -78 °C
using a cannula technique within 10 min. The solution was
stirred for 12 h at the same temperature; all the starting reactants
were consumed to give the chlorosilylenoid 2.7 To the dark
brown reaction mixture was added an excess of acetaldehyde
(2.2 mL, 39 mmol) to the reaction products at -78 °C, where-
upon the solution rapidly became light yellow. Then treatment
of the reaction mixture with an excess of MeOH gave the
trapping product 3 (in 83% GC yields). The reaction mixture
was slowly warmed to room temperature, and then the solvent
was evaporated under reduced pressure. After the sublimation
of naphthalene, to the resulting viscous light yellow oil was
added n-hexane (100 mL), and then precipitated species were
removed by filtration. The crude material was purified by silica
gel chromatography (n-hexane) to afford 3 as a colorless oil in
28% yield (0.42 g, 1.1 mmol). Compound 3 was characterized by
NMR, GC/MS, EA, and HRMS. 1H NMR (400 MHz, chloro-
form-d): δ 0.25 (s, 27H), 1.36 (d, J = 4.8, 3H), 1.40 (d, J = 7.2,
Although the effects of reaction media are not included in
the present calculations, the effects are expected to be very
small because the dielectric constant of THF, the main
reaction medium of the actual experiments, is very small.
The relative energies of actual compounds with a Tsi group
may be a little different from the calculated results with the
model compounds of the present calculations. In spite of
such approximations, our theoretical results depicted in
Figures 2 and 5 strongly suggest that an addition mechanism
of aldehydes to silaoxirane is a very plausible one.
Summary and Conclusions
The reaction of (Tsi)chlorosilylenoid 2 with acetaldehyde
and benzaldehyde gave the corresponding products,
2,4-dioxasilolane (3) and 2,5-dioxasilolane (4), respectively.
This first example of the reaction of the silylenoid with
aldehydes showed a synthetic application of the chlorosily-
lenoid as a new route for the synthesis of functional silahe-
terocycles. To investigate the possible pathway for the
formation of the silaheterocycles, theoretical computations
were carried out by using the B3LYP and the M05-2X DFT
functional forms. From the theoretical results, the direct
addition of aldehyde to the silaoxirane intermediate, shown
in Scheme 3, is proposed as a plausible mechanism to explain
the experimental facts, and the theoretical results supporting
the mechanism are depicted in Figures 2 and 5. The produc-
tion of 2,5-dioxasilolane by the cleavage of the Si-C bond of
silaoxirane seems to be more natural result because the
reaction is both thermodynamically and kinetically favored
over the formation of 2,4-dioxasilolane by the cleavage of the
Si-O bond. In this context, the production of 5Me is a rather
exceptional result in a sense, but we think that the main
reason for the exception seems to be because a single methyl
group of acetaldehyde is not enough to stabilize the partial
charges developed at the reacting atoms in the reaction path
toward 2,5-dioxasilolane. Further experimental and theore-
tical studies to support this mechanism or to suggest another
one surely deserve to be conducted in the near future, but our
mechanistic studies provide experimentally and theoretically
consistent, reasonable, and insightful understandings.
3H), 3.41 (q, J = 7.2, 1H), 3.68 (s, 3H), 5.17 (q, J = 4.8, 1H). 13
C
NMR (100 MHz, chloroform-d): δ 1.0 (C(Si(CH3)3)3), 4.5
(SiCH3), 16.4 (CH3), 22.9 (CH3), 51.1 (SiOCH3), 67.6
(CHCH3), 96.3 (CHCH3). GC/MS: m/z (%) 363 (Mþ - 15,
1.9), 305 (7.6), 275 (100), 261 (37.6), 189 (23.6), 129 (14.3), 73
(35.8). HRMS: C14H35O3Si4 363.1663 (calcd), 363.1662 (found).
Anal. Calcd for C15H38O3Si4: C, 47.56; H, 10.11; O, 12.67.
Found: C, 47.46; H, 10.01; O, 12.74.
Synthesis of 4. The product mixture of chlorosilylenoid was
trapped by benzaldehyde (4.0 mL, 39 mmol) in a similar manner
to those described above to give compound 4 (80%, GC yield).
We carried out the workup procedures in a similar manner
to those used in the trapping experiment described above.
The crude material was purified by silica gel chromatography
(n-hexane) to afford 4 as a colorless oil in 55% yield (1.2 g,
2.2 mmol). Compound 4 was characterized by NMR, GC/MS,
1
EA, and HRMS. H NMR (400 MHz, chloroform-d): δ 0.43
(s, 27H), 4.86 (d, J = 9.3, 1H), 5.07 (d, J = 9.3, 1H), 7.15-7.35
(m, 10H). 13C NMR (100 MHz, chloroform-d): δ 4.5 (SiCH3),
7.0 (C(Si(CH3)3)3), 82.6, 84.3 (CH), 126.6, 127.5, 127.9, 128.2,
128.2 (ph-CH), 137.9, 139.2 (ph-ipso-C). GC/MS: m/z (%) 506
(Mþ, 3.1), 491 (21.2), 295 (100), 187 (37.8), 179(53.0), 90 (29.8),
73 (71.5). HRMS: C24H39ClO2Si4 506.1716 (calcd), 506.1715
(found). Anal. Calcd for C24H39ClO2Si4: C, 56.82; H, 7.75; O,
6.31. Found: C, 56.75; H, 7.56; O, 6.35.
Experimental Section
Acknowledgment. This work was supported by the
Korea Research Foundation Grant funded by the
Korean Government (No. KRF-2007-521-C00149).
The authors thank Professor Robert West, distinguished
professor in the WCU (World Class University) Program
through the National Research Foundation of Korea
funded by the Ministry of Education, Science and Tech-
nology for his valuable discussions.
General Considerations. In all reactions in which air-sensitive
chemicals were used, the reagents and solvents were dried prior
to use. THF was distilled from Na/Ph2CO. Other starting
materials were purchased as reagent grade and used without
further purification. Glassware was flame-dried under nitrogen
or argon flushing prior to use. All manipulations were per-
formed using standard Schlenk techniques under a nitrogen or
argon atmosphere. 1H NMR and 13C NMR spectra were
recorded on a Bruker Avance IIþ BBO 400 MHz S1 spectro-
meter in CDCl3. Analyses of product mixtures were accom-
plished using an HP 5890 II Plus instrument with FID (HP-5,
30 m column) with dried n-decane as an internal standard. Mass
spectra were recorded on a low-resolution (Agilent Technolo-
gies GC/MS: 6890N, 5973N mass selective detector) EI mass
Supporting Information Available: Molecular structures in-
cluding H atoms, total energies including zero-point energy
(ZPE), and atomic xyz-coordinates produced by computations.
These materials are available free of charge via the Internet at
http://pubs.acs.org.