A mechanistic study of reactions of stable disilenes with haloalkanes

Article abstract of DOI:10.1021/ja002798s

Mechanisms of the reactions of three tetrakis(trialkylsilyl)disilenes and a tetraaryldisilene with various haloalkanes such as carbon tetrachloride, chloroform, dichloromethane, which gave the corresponding 1-alkyl-2-chlorodisilanes and/or 1,2-dichlorodisilanes, were investigated in detail. As evidenced by an ESR observation of an intermediate radical, these reactions were quite unusual, forming neutral radical pairs from two closed shell molecules at the first step; no similar reactions have been observed between alkenes and haloalkanes. Low oxidation potentials of these disilenes, large negative activation entropies, and solvent effects for the rates are in good accord with the direct halogen abstraction of disilenes from haloalkanes instead of single-electron transfer at the rate-determining first step. The structure-reactivity relationship of the reactions and the Hammond postulate suggest that the transition state structures for the first step are similar to those for the halogen abstraction by silyl radicals, but more product-like.

Full text of DOI:10.1021/ja002798s

1676
J. Am. Chem. Soc. 2001, 123, 1676-1682
A Mechanistic Study of Reactions of Stable Disilenes with
Haloalkanes
Mitsuo Kira,* Tai Ishima, Takeaki Iwamoto, and Masaaki Ichinohe
Contribution from the Department of Chemistry, Graduate School of Science, Tohoku UniVersity,
Aoba-ku, Sendai 980-8578, Japan
ReceiVed July 28, 2000. ReVised Manuscript ReceiVed NoVember 30, 2000
Abstract: Mechanisms of the reactions of three tetrakis(trialkylsilyl)disilenes and a tetraaryldisilene with various
haloalkanes such as carbon tetrachloride, chloroform, dichloromethane, which gave the corresponding 1-alkyl-
2-chlorodisilanes and/or 1,2-dichlorodisilanes, were investigated in detail. As evidenced by an ESR observation
of an intermediate radical, these reactions were quite unusual, forming neutral radical pairs from two closed
shell molecules at the first step; no similar reactions have been observed between alkenes and haloalkanes.
Low oxidation potentials of these disilenes, large negative activation entropies, and solvent effects for the
rates are in good accord with the direct halogen abstraction of disilenes from haloalkanes instead of single-
electron transfer at the rate-determining first step. The structure-reactivity relationship of the reactions and
the Hammond postulate suggest that the transition state structures for the first step are similar to those for the
halogen abstraction by silyl radicals, but more product-like.
Introduction
Scheme 1
Since the isolation of tetramesityldisilene as the first stable
disilene in 1981,1 there have been reported a number of
distinctive reactions of stable disilenes with various reagents
such as hydroxylic compounds (water, alcohols, and phenols)
and carbon-carbon multiply bonded compounds (alkenes,
alkynes, and 1,3-dienes) giving various types of adducts.2
Whereas the mechanisms of these reactions have not been
investigated yet in detail, the reaction modes and reactivity of
disilenes appear to be remarkably different from those of
alkenes. Typically, in contrast to CdC double bonds, SidSi
double bonds react with water and various alcohols very
smoothly without catalysts to give the corresponding adducts,
but the mechanistic details of the reactions including the origin
of the diverse stereochemical outcome are still under active
discussion.3,4
types of products as shown in Scheme 1. The formation of these
products is apparently suggestive of the radical nature of the
reactions. These reactions are rationalized if a pair of the
corresponding halodisilanyl radical and alkyl radical is formed
in the first step (Scheme 2); in the second step, various types
Recently, West et al.5 and our group6 have found that disilenes
react rather unusually with various haloalkanes giving several
(1) West, R.; Fink, M. J.; Michl, J. Science (Washington) 1981, 214,
1343.
Scheme 2
(2) For reviews on disilenes, see: (a) West, R. Pure Appl. Chem. 1984,
56, 163. (b) Raabe, G.; Michl, J. Chem. ReV. 1985, 85, 419. (c) West, R.
Angew. Chem., Int. Ed. Engl. 1987, 26, 1201. (d) Raabe, G.; Michl, J. In
The Chemistry of Organic Silicon Compounds; Patai, S., Rappoport, Z.,
Eds.; John Wiley & Sons: New York, 1989; Part 2, Chapter 17. (e)
Tsumuraya, T.; Batcheller, S. A.; Masamune, S. Angew. Chem., Int. Ed.
Engl. 1991, 30, 902. (f) Grev, R. S. AdV. Organomet. Chem. 1991, 33,
125. (g) Okazaki, R.; West, R. AdV. Organomet. Chem. 1996, 39, 231. (h)
Kira, M.; Iwamoto, T. J. Organomet. Chem. 2000, 610, 236.
(3) Sakurai, H. In The Chemistry of Organic Silicon Compounds;
Rappoport, Z., Apeloig, Y., Eds.; John Wiley: New York, 1998; Vol. 2,
Part 1, Chapter 15.
(4) (a) Apeloig, Y.; Nakash, M. J. Am. Chem. Soc. 1996, 118, 9798. (b)
Apeloig, Y.; Nakash, M. Organometallics 1998, 17, 1260. (c) Apeloig, Y.;
Nakash, M. Organometallics 1998, 17, 2307. See also: Takahashi, M.;
Veszpre´mi, T.; Hajgato´, B.; Kira, M. Organometallics 2000, 19, 4660.
(5) Fanta, A. D.; Belzner, J.; Powell, D. R.; West, R. Organometallics
1993, 12, 2177.
of products will be formed via recombination, disproportionation
of the radical pair in the cage, or abstraction of the second
halogen by the halodisilanyl radical out of the cage.
These reactions are quite unusual because a neutral radical
pair is formed from two neutral closed shell molecules;7 no
similar reactions have been observed between alkenes and
haloalkanes. Elucidation of the mechanisms of these reactions
may contribute to understanding the characteristics of double
(7) As a rather exceptional case, reactions of a dioxirane with olefins
etc. have been proposed to proceed via a similar radical pair formation,8
while a concerted mechanism is claimed as an alternative mechanism.9
(6) Iwamoto, T.; Sakurai, H.; Kira, M. Bull. Chem. Soc. Jpn. 1998, 71,
2741.
10.1021/ja002798s CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/02/2001

Reactions of Stable Disilenes with Haloalkanes
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1677
Table 1. Reactions of Various Disilenes with Haloalkanes in
Hexanea
yields (%)b
disilene
R′X
CCl4
CHCl3
CH2Cl2
PhCH2Cl
CCl4
CHCl3
CH2Cl2
CCl4
CHCl3
t-BuBr
CCl4
2
3
1a (R ) i-Pr2MeSi)
83 (2a)
26 (2a)
0
0
70 (3a)
93 (3a′)
51 (3a′′)
0
1b (R ) t-BuMe2Si)c
1c (R ) i-Pr3Si)
1d (R ) Tip)
83 (2b)
0
0
78 (3b)
85 (3b′)
0
70 (2c)
62 (2c)
81 (2c′)
84 (2d)
94 (2d)
85 (2d)
0
0
0
0
0
0
CHCl3
CH2Cl2
Figure 1. An ESR spectrum of disilanyl radical 4c observed during
the reaction of 1c with tert-butyl chloride in benzene-d6 at room
temperature.
a See Experimental Section for detailed reaction conditions. b De-
termined by NMR. c Reference 6.
that the two types of products are produced via the two
competitive reactions from the radical pair shown in Scheme
2.
2. ESR Spectra of Intermediates. If the reaction of a disilene
with a haloalkane proceeds via the corresponding 2-halodisilanyl
radical and alkyl radical, the 2-halodisilanyl radical as a key
intermediate might be observed by ESR. During the reaction
of 1c with excess tert-butyl chloride in benzene-d6 at room
temperature (eq 2), an intense singlet ESR signal with several
satellites due to 29Si nuclei was observed (Figure 1).
bonds of group-14 elements. In this paper, we discuss the
detailed mechanisms of the reactions of several stable disilenes
(1a-d) with various haloalkanes.
Chart 1
Results and Discussion
A. Overall Profile of Reaction. 1. Product Study. Tetrasi-
lyldisilenes 1a-c11 and a tetraaryldisilene 1d12 reacted with
various haloalkanes in the dark to give the corresponding
dihalodisilanes 2 and 1-alkyl-2-halodisilanes 3 in high yields
(eq 1), while the reaction rates and product distribution depended
The ESR spectrum is assignable to the corresponding
oligosilanyl radical 4c on the basis of the g value and splitting
pattern of the satellite signals: g ) 2.0054, a(SiR) ) 5.73 mT,
and hfs values of five satellites with 1.03, 0.73, 0.59, 0.46, and
0.27 mT due to other 29Si and 13C nuclei. The g and a(SiR)
values for 4c are similar to those of tris(trialkylsilyl)silyl radi-
cals:13 a(SiR) ) 5.56 and 5.71mT, g ) 2.0061 and 2.0055 for
(i-Pr3Si)3Si• 13b and (t-BuMe2Si)3Si•,13c respectively. The reaction
of 1c with tert-butyl chloride was very slow and only 10% of
1c were consumed after 1 month at room temperature; the ESR
signal intensity continued slowly to increase for more than half
a year. Disilanyl radical 4c was exceptionally long-lived,
probably because 4c is well protected sterically to cause neither
the chlorine abstraction from tert-butyl chloride nor recombina-
tion with tert-butyl radical. Isobutene and isobutane were
detected in the reaction mixture by NMR, indicative of the
formation of tert-butyl radicals during the reaction.
In contrast, disilenes 1a and 1b react with tert-butyl chloride
rather smoothly to show no ESR signals due to silyl radicals
during the reactions. The major products were 2a and 2b, but
several other products involving ene-addition products of
isobutene with the disilenes were also formed.14 Very weak ESR
signals were detected but unidentified during the reaction of
1d with tert-butyl chloride, although the reaction was very slow
similar to the corresponding reaction of 1c; the major product
strongly on the types of disilenes and haloalkanes as shown in
Table 1. The reactions of sterically very bulky disilenes 1c and
1d gave only the corresponding 1,2-dihalodisilanes 2c (or 2c′)
and 2d irrespective of the structure of haloalkanes. The reactions
of less bulky disilene 1a with dichloromethane and benzyl
chloride gave only the simple addition products 3a′ and 3a′′,
while the reaction of 1a with chloroform produced a mixture
of 2a and 3a. Carbon tetrachloride was the most reactive among
investigated haloalkanes toward these disilenes to afford only
the corresponding 1,2-dichlorodisilanes. These results suggest
(8) (a) Bravo, A.; Fontana, F.; Fronza, G.; Minisci, F. Tetrahedron Lett.
1995, 38, 6945. (b) Bravo, A.; Fontana, F.; Fronza, G.; Minisci, F.; Zhao,
L. J. Org. Chem. 1998, 63, 254.
(9) Adam, W.; Curci, R.; D’Accolti, L.; Dinoi, A.; Fusco, C.; Gasparrini,
F.; Kluge, R.; Paredes, R.; Schulz, M.; Smerz, A. K.; Veloza, L. A.;
Weinko¨tz, S.; Winde, R. Chem. Eur. J. 1997, 3, 105.
(10) Dixon, C. E.; Hughes, D. W.; Bains, K. J. Am. Chem. Soc. 1998,
120, 11049.
(11) Kira, M.; Maruyama, T.; Kabuto, C.; Ebata, K.; Sakurai, H. Angew.
Chem., Int. Ed. Engl. 1994, 33, 1489.
(12) Watanabe, H.; Takeuchi, K.; Fukata, N.; Kato, M.; Goto, M.; Nagai,
Y. Chem. Lett. 1987, 1341.
(13) (a) Kyushin, S.; Sakurai, H.; Betsuyaku, T.; Matsumoto, H.
Organometallics 1997, 16, 5386. (b) Kyushin, S.; Sakurai, H.; Matsumoto,
H. Chem. Lett. 1998, 107. (c) Kira, M.; Obata, T.; Kon, I.; Hashimoto, H.;
Ichinohe, M.; Sakurai, H.; Kyushin, S.; Matsumoto, H. Chem. Lett. 1998,
1097. (d) Cooper, J.; Hudson, A.; Jackson, R. A. Mol. Phys. 1972, 23, 209.
(14) The ene-addition reactions of disilenes 1b are reported to occur very
easily via the six-membered cyclic transition states.6

1678 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Kira et al.
Scheme 3
was again dichlorodisilane 2d. The intermediate silyl radicals
formed from disilenes 1a, 1b, and 1d with tert-butyl chloride
should be more reactive than 4c due to less severe steric
bulkiness of the substituents for 1a and 1b, and due to possible
homolytic aromatic addition of silyl radicals15 for 1d.
estimated by using the following eqs 3 and 4, where Z, λ, C,
and ∆G are the frequency factor, reorganization energy,
Coulomb energy, and free energy difference, respectively.
B. The First Step: Single-Electron Transfer or Halogen
Abstraction. As shown in Scheme 3, the first step may proceed
either through (1) the single-electron transfer (SET) from
disilene to haloalkane followed by the coupling of the disilene
cation radicals and chloride anion or through (2) the direct
halogen abstraction of disilene from haloalkane, where the
disilene is supposed to serve as a biradical.
The ∆G value for the SET reaction between 1c and
chloroform is estimated to be larger than 34 kcal‚mol-1, because
Eox of 1c and Ered of chloroform are +0.70 and <-1.0 V,18b
respectively; the Coulomb energy is estimated to be -5
kcal‚mol-1 by assuming the interatomic distance of 10 Å. The
rate constant k is suggested to be <10-14 M-1‚s-1, if we assume
that the λ value is ca. 50 kcal‚mol-1 according to the evaluation
by Eberson19 for reduction of haloalkanes, and the Z value is
ca. 1011 M-1‚s-1. As shown in the next section, k is measured
to be 1.2 × 10-2 M-1‚s-1 for the reaction of 1c with chloroform
at 298 K. The observed k is 1012 times larger than the above
estimated k, and hence incompatible with the SET process
between 1c and chloroform. Since disilenes 1a-d react with
those haloalkanes which have even lower Ered than chloroform,
the reaction mechanisms should not involve the initial SET
process.
1. Cyclic Voltammetry of Disilenes. To distinguish the above
two possible mechanisms, the assessment of the SET mechanism
using the Marcus theory17 will be effective. For this purpose,
oxidation (Eox) as well as reduction potentials (Ered) of various
disilenes have been investigated in THF by cyclic voltammetry
by using a conventional three-electrode cell design; an AgCl-
coated silver wire, a glassy carbon rod, and a platinum metal
wire were used as a working electrode, a reference electrode,
1
and a counter electrode, respectively. It was confirmed by H
NMR spectroscopy that disilenes 1a and 1d did not react for
several hours with supporting electrolyte [Bu4N][PF6] in THF-
d8. Since both reduction and oxidation of these disilenes were
irreversible, the peak potentials were compared in Table 2.
Meanwhile, interesting substituent effects on both oxidation
and reduction potentials of disilene were observed. The oxidation
potentials of tetrasilyldisilenes 1a and 1c were 0.14-0.44 V
higher than those of tetramesityldisilene and 1d, and the first
reduction potentials of tetrasilyldisilenes 1a and 1c were ca.
0.3-0.9 V lower than those of tetraaryldisilenes. The results
suggest that both HOMO and LUMO of tetrasilyldisilenes are
significantly lower in energy than those of tetraaryldisilenes as
shown in Figure 2; tetrasilyldisilenes are more electron-accepting
and less electron-donating than tetraaryldisilenes. Significant
differences in the various reactions of tetrasilyldisilenes from
those of tetraaryldisilenes may be explained by the major
electronic effects of the substituents, while the steric effects
should also be taken into account; for example, tetrasilyldisilenes
1a and 1b react with 2,3-dimethylbutadiene to produce the
corresponding Diels-Alder adducts, while no [4+2] addition
of tetramesityldisilene to 1,3-butadienes has been reported.6 The
more electron-accepting tetrasilyldisilenes 1a and 1c afforded
Table 2. Peak Potentials of Various Disilenes in THFa
disilene
Ep,Ox (V)
Ep,Red (V)
1a
1c
1d
+0.88
-1.82, -2.41
-1.70, -2.56
-2.66
+0.70
+0.56, +1.32
Mes2SidSiMes2
+0.44 (+0.38)b
-2.12 (-2.12)b
a Conditions: Potentials in V vs ferrocene/ferrocenium (+0.61 V),
supporting electrolyte 0.1 M [NBu4][PF6], reference electrode Ag/Ag+,
at ambient temperatures in THF at 0.2 V/s. b The data in parentheses
are taken from ref 18a.
The oxidation potentials of the investigated tetraaryldisilenes
and tetrasilyldisilenes were in the range of +0.4 to +0.9 V.
The relatively high oxidation potentials of these disilenes
eliminate the possible SET mechanism for the initial reactions
of disilenes with haloalkanes. According to the Marcus theory,17
the second-order rate constant (k) for an SET reaction is
(15) For reviews of silyl radicals, see: (a) Sakurai, H. In Free Radicals;
Kochi, J. K., Ed.; John Wiley & Son: New York, 1973; Vol. II. (b)
Chatgilialoglu, C. Chem. ReV. 1995, 95, 1229.
(16) Recently, Bains et al. have found the reactions of stable disilenes
with ketones to proceed via biradical adducts, while the detailed discussion
on the transition states for the step was not given: Dixon, C. E.; Hughes,
D. W.; Bains, K. J. Am. Chem. Soc. 1998, 120, 11049.
(17) (a) Marcus, R. Annu. ReV. Phys. Chem. 1964, 15, 155. (b) Marcus,
R. J. Chem. Phys. 1965, 43, 679. (c) Marcus, R. J. Phys. Chem. 1968, 72,
891.
(18) (a) Shepherd, B. D.; West, R. Chem. Lett. 1988, 183. (b) Eox,p
-
(chloroform) ) -1.67 V (SCE) in a dioxane-water (75/25) mixture: Mann,
C. K.; Barnes, K. K. Electrochmical Reactions in the Nonaqueous Systems;
Marcel Dekker Inc: New York 1970.
Figure 2. Schematic representation of the relative energy levels of
the frontier orbitals of tetrasilyldisilenes and tetraaryldisilenes.
(19) Eberson, L. Acta Chem. Scand. 1982, B36, 533.

Reactions of Stable Disilenes with Haloalkanes
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1679
Table 3. Rate Constants of the Reactions of Disilenes with
Chloroform at 303 K in Various Solvents
Table 4. Comparison of Second-Order Rate Constants of Halogen
Abstractions of Disilenes (at 303 K in hexane) and Silyl Radicals
(at 300 K) from Various Haloalkanes
104k (M-1‚s-1
)
k (M-1‚s-1
Et3Si• a
)
a
solvent
hexane
toluene
benzene
THF
1c
1d
ET (kcal‚mol-1
)
haloalkane
CCl4
CHCl3
1c
1d
(Me3Si)3Si• b
160
210
200
110
330
d
1.6
0.82
2.2
b
30.9
33.9
34.5
37.4
37.5
41.1
1.7 × 10-1 4.3 × 109 1.7 × 108
1.6 × 10-2 1.6 × 10-4 2.5 × 108 6.8 × 106
5.2 × 10-5 1.4 × 10-7 7.1 × 107
2.6 × 10-3 4.1 × 109
CH2Cl2
8.0c
4.9c
Cl3CCCl3
Cl2CHCHCl2
PhCH2Cl
t-BuCl
chlorobenzene
dichloromethane
2.0 × 10-3 2.2 × 108
1.4 × 10-3 1.6 × 10-6 2.0 × 107 4.6 × 106
a Data were taken from ref 23. b A complex mixture of products was
formed. c Reactions of 1d with these solvents were assumed to be too
slow to affect the rate constants. d Not determined.
7.4 × 10-7
8.8 × 10-3
2.2 × 10-4
1.7 × 10-7
2.5 × 106 4.0 × 105
1.1 × 109 1.2 × 108
4.6 × 107
t-BuBr
i-PrBr
CH3(CH2)4Br 1.0 × 10-5
5.4 × 108 2.0 × 107
PhBr
1.1 × 108 4.6 × 106
the second reduction potentials at around -1.8 and -2.5 V,
while 1d and tetramesityldisilene showed only one reduction
peak at >-2.5 V. The results coincide with the fact that
tetrasilyldisilenes 1a and 1b are reduced by lithium metal in
THF to give the corresponding dianions,20 while a similar
lithium reduction of 1d is known to give the corresponding
disilenyllithium via the Si-Ar bond cleavage.21
2. Kinetic Order, Rates, Activation Parameters, and
Solvent Effects. The reactions of disilenes with haloalkanes in
hexane were followed by UV spectroscopy, using the absorbance
at the ππ*band maxima of the disilenes. When the initial
concentrations of a disilene and haloalkane were ca. 10-6 and
10-3 M, pseudo-first-order kinetics was observed (eq 5). Several
kinetic runs showed that the apparent first-order rate constants
(kobs) were proportional to the concentrations of the haloalkane
(eq 6); the second-order rate constants k were determined as
a Reference 24b. In di-tert-butylperoxide-triethylsilane. b Reference
24d. In hexadecane.
significantly solvent dependent. Polar solvents slightly acceler-
ated the reactions but the extent is smaller than that expected
for the SET reactions. All the features of the reactions, i.e.,
relatively low oxidation potentials of these disilenes, large
negative activation entropies, and solvent effects for the rates,
are incompatible with the SET reactions at the rate-determining
initial step.
The reactions of disilenes with haloalkanes are judged to
proceed via the rate-determining halogen abstraction of disilenes,
where the disilenes serve as biradicals. Since the role of group-
14 element unsaturated bonds as biradicals is rather exceptional,
we have investigated further in detail the structure-reactivity
relationship of the reactions to elucidate the transition state
structures.
C. Transition-State Structure for the First Step. 1.
Disilenes vs Silyl Radicals. To determine the electronic and
steric requirement for the first-step reactions, the second-order
rate constants were compared with those for the halogen
abstraction by silyl radicals.24-26 In Table 4, the rate constants
for the halogen abstractions of disilenes 1c and 1d are shown
together with those reported for silyl radicals.
the slope. The reaction was thus characterized as obeying an
overall second-order rate law, first order in disilene and first
order in haloalkane. All the reactions investigated here showed
similar kinetics, indicating that the rate-determining step of these
reactions is the first step to form radical pairs.
The absolute rates for disilenes are 108-1014 times slower
than the corresponding reactions for silyl radicals. These
differences are understood on the basis of the reaction enthalpies
(∆H0). The reactions of silyl radicals are well-known to be
exothermic by more than 30 kcal‚mol-1, irrespective of chlorine
or bromine abstraction.27 On the other hand, the reactions of
the disilenes with haloalkanes will be less exothermic (∆H0 )
-5 to -15 kcal‚mol-1), because a Si-Si π bond is cleaved
during the reactions.30 The higher reactivity of tetrasilyldisilenes
Typically, k’s for the reactions of 1c and 1d with chloroform
were thus determined to be (1.58 ( 0.005) × 10-2 and (1.62 (
0.06) × 10-4 M-1‚s-1, respectively, in hexane at 303 K. From
the temperature dependence of k, the activation parameters were
determined as ∆Hq ) 7.0 ( 0.1 kcal‚mol-1 and ∆Sq ) -43.9
( 0.05 cal‚mol-1‚K-1 for 1c and ∆Hq ) 7.4 ( 0.1 kcal‚mol-1
and ∆Sq ) -43.1 ( 0.07 cal‚mol-1‚K-1 for 1d. The very large
negative activation entropies of these reactions support the
bimolecular chlorine abstraction mechanism for the first step
of the reactions; large negative activation entropies around -40
cal‚mol-1‚K-1 have also been observed for the reactions of
tetramesityldisilene with phenols.4c The SET mechanism is again
eliminated because typical SET reactions should show relatively
small and even positive activation entropy.22
(23) Kosower, E. M.; Mohammad, M. J. Am. Chem. Soc. 1968, 90, 3271.
(24) (a) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C.; Woynar, H. J.
Am. Chem. Soc. 1981, 103, 3231. (b) Chatgilialoglu, C.; Ingold, K. U.;
Scaiano, J. C. J. Am. Chem. Soc. 1982, 104, 5123. (c) Lusztyk, J.; Maillard,
B.; Ingold, K. U. J. Org. Chem. 1986, 51, 2457. (d) Chatgilialoglu, C.;
Griller, D.; Lesage, M. J. Org. Chem. 1989, 54, 2492. (e) Ito, O.; Hoteiya,
K.; Watanabe, A.; Matsuda, M. Bull. Chem. Soc. Jpn. 1991, 64, 962. (f)
Sluggett, G. W.; Leigh, W. J. Organometallics 1992, 11, 3731.
(25) Ingold, K. U.; Lusztyk, J.; Scaiano, J. C. J. Am. Chem. Soc. 1984,
106, 343.
As shown in Table 3, while rates for disilene 1c are much
larger than those for disilene 1d, these rate constants were not
(26) (a) Tamblyn, W. H.; Vogler, E. A.; Kochi, J. K. J. Org. Chem.
1980, 45, 3912. (b) Blackburn, E. V.; Tannner, D. D. J. Am. Chem. Soc.
1980, 102, 692.
(20) Iwamoto, T.; Yin, T.; Maruyama, T.; Kabuto, C.; Kira, M. To be
submitted for publication. See also: 10th International Symposium on
Organosilicon Chemistry; Pozna´n, Poland, 1993; Abstracts, p 263.
(21) (a) Scott, J. P.; Batchller, A.; Masamune, S. J. Organomet. Chem.
1989, 367, 39. (b) Weidenbruch, M.; Willms, S.; Saak, W.; Henkel, G.
Angew. Chem., Int. Ed. Engl. 1997, 36, 2503.
(27) For this discussion, bond dissociation energies (BDE) of Si-Cl,
Si-Br, C-Cl, and C-Br were taken to be 113, 96, 80, and 64 kcal‚mol-1 28
.
The BDE values depend significantly on the substituents. For further detailed
discussion in this paper, the BDE values of C-Cl are taken to be 83.6,
82.8, 79.2, and 70.3 kcal‚mol-1 for CH3-Cl, ClCH2-Cl, Cl2CH-Cl, and
Cl3C-Cl, respectively.29
(22) Typically, the activation entropy for the electron transfer from
anthracene anion to butyl bromide was reported to be +5 cal‚mol-1‚K-1
:
Lexa, D.; Sa´veant, J.-M.; Su, K.-B.; Wang, D.-L. J. Am. Chem. Soc. 1988,
110, 7617.
(28) Walsh, R. Acc. Chem. Res. 1981, 14, 246.
(29) Weissman, M.; Benson, S. W. J. Phys. Chem. 1983, 87, 243.

1680 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Kira et al.
Figure 3. A log-log plot of the rate constants of the chlorine
abstraction between disilene 1d (k1) and triethylsilyl radical (k3).
Figure 5. Comparison of the reaction profile of the halogen abstraction
of a silyl radical with that of a disilene.
abstraction than that for chlorine abstraction.24b For disilene 1c,
the reactivity difference between t-BuCl and t-BuBr was
observed to be much larger in extent in accord with the
reactivity-selectivity principle; k[t-BuBr]/k[t-BuCl] ) 1.2 ×
104 at 303 K.
The fact that the reactivities of disilenes toward various
haloalkanes are parallel to those of silyl radicals with larger
selectivity for the former suggests the similarity of the transition
states between these two types of halogen abstractions. Since,
like other atom abstraction reactions, the transition states for
halogen abstraction by a silyl radical are considered to have a
linear arrangement of (halide carbon)-halogen-silicon atoms,
similar linear transition states will be applicable to the corre-
sponding reactions of disilenes as shown in Figure 5. According
to the Hammond postulate,33 the transition state for a silyl radical
should be early and reagent-like. On the other hand, the
transition state for a disilene should be later, where the carbon-
halogen bond cleavage is significantly developed.
2. Hammett-Type Relationship. Chatgilialoglu and Ingold34
have investigated the substituent effects on the rate for chlorine
abstraction of triethylsilyl radical from benzyl chlorides, and
found a linear plot of logk vs σp with F of +0.64, indicative of
significant polar effects, i.e., significant development of negative
charge on the benzylic carbon in the transition state for this
reaction. In radical reactions, however, spin delocalization effects
due to substituents may also be important. On this basis, Jiang
et al. have recently proposed a two-parameter relationship for
radical reactions using various combinations of polar and spin
delocalization substituent constants.35 Substituent effects on the
rates for the bromine abstraction of tris(trimethylsilyl)silyl
radical from substituted benzyl bromide (eq 7) were thus
Figure 4. A log-log plot of the rate constants for the reactions of
disilene 1c (k2) vs (Me3Si)3Si• (k4).
than tetraarydisilenes may be attributed to the smaller π-bond
energy for the former.
Despite the significant difference of the reaction enthalpies
between disilenes and silyl radicals, the structure-reactivity
relationships of these reactions are rather parallel to each other.
As shown in Figure 3, logarithms of rate constants for the
reactions of 1d with various haloalkanes correlated roughly
linearly with those for triethylsilyl radical, with a correlation
coefficient of 0.854. The reason for the rather poor correlation
may be ascribed in part to the significant difference in the
electronic and steric effects of the substituents at the radical
centers between 1d and triethylsilyl radical.
A plot of log k for the reactions of chloro- and bromoalkanes
with 1c against those with tris(trialkylsilyl)silyl radical gave two
different linear lines for chloro- and bromoalkanes with almost
the same slope as shown in Figure 4; the slopes and correlation
coefficients were 3.37 (r ) 0.994) and 3.36 (r ) 0.996) for
chloro- and bromoalkanes, respectively. For halogen abstraction
of silyl radicals, it is well-recognized that the reaction with tert-
butyl chloride is much slower than the reaction with tert-butyl
bromide, while the reaction enthalpies are almost the same; ∆H0-
(t-BuX) ) -30.7 and -29.7 kcal‚mol-1, for X ) Cl and Br,
respectively, but k[t-BuBr]/k[t-BuCl] ) 4.4 × 102 and 3.0 ×
102 for Et3Si• and (Me3Si)3Si•, respectively. The reason for the
rate difference is usually ascribed to the stabilization of the
transition states by polar effects, which is larger for bromine
(31) (a) Michalczyk, M. J.; West, R.; Michl, J. Organometallics 1985,
4, 826. (b) Shephered, B. D.; Powell, D. R.; West, R. Organometallics
1989, 8, 2664. (c) Batcheller, S. A.; Tsumuraya, T.; Tempkin, O.; Davis,
W. M.; Masamune, S. J. Am. Chem. Soc. 1990, 112, 9394. See also ref 1
g.
(32) Kira, M.; Ohya, S.; Iwamoto, T.; Ichinohe, M.; Kabuto, C.
Organometallics 2000, 19, 1817.
(33) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334.
(34) Chatgilialoglu, C.; Ingold, K. U.; Scaiano, J. C. J. Org. Chem. 1987,
52, 928.
(35) (a) Jiang, X. K.; Ji, G. Z.; Xie, J. R. Y. Tetrahedron 1997, 53, 8479.
(b) Jiang, X. K. Acc. Chem. Res. 1997, 30, 283. (c) Jiang, X. K.; Ji, G. Z.
J. Org. Chem. 1992, 57, 6051.
(30) The π bond energies of disilenes may be estimated roughly from
the rotational barrier around the SidSi bonds, which are reported to be ca.
25 and 15 kcal‚mol-1 for several 1,2-diaryldisilenes31 and a tetrasilyldi-
silene,32 respectively. Since the barrier for a tetrasilyldisilene is lowered
by the effective σ-π conjugation at the transition state, the barrier of 15
kcal‚mol-1 for tetrasilyldisilenes should be taken as the lower limit of the
π bond energy.

Reactions of Stable Disilenes with Haloalkanes
J. Am. Chem. Soc., Vol. 123, No. 8, 2001 1681
Table 5. Second-Order Rate Constants for the Reaction of
Disilene 1c with Substituted Benzyl Chlorides and Substituent
Constants at 303 K in Hexane
disilene was calculated to be 13% for the CASSCF/6-31G**
structure.37 This biradical character may be related to the π bond
energy, because the smaller the energy required to break the π
bond is, the more favorable the biradical formation. Since the
π bond energies for disilenes (25 kcal‚mol-1) are much smaller
than for the usual alkenes (65 kcal‚mol-1), disilenes should have
larger biradical character. The reason for the small π bond
energies of disilenes would be attributed to the large positive
singlet-triplet energy differences of the component silylenes
[∑∆EST; ∆EST ) E(triplet) - E(singlet)] on the basis of CGMT
theory.38,39 The extension of the above discussion to alkenes
suggests that those alkenes that are composed of singlet carbenes
may work as biradicals in their reactions. Actually, tetrafluo-
roethylene, whose π bond energy is significantly smaller than
those of the normal alkenes in relation to the large ∆EST of
difluorocarbene,38,40 has been known to work as a biradical in
its reactions with 1,3-dienes.41
103kX (M-1‚s-1
)
σp
σJJ
a
b
X
t-Bu
Me
H
F
Cl
CF3
CN
1.7
1.2
1.4
1.5
7.7
4.1
13.0
-0.20
-0.17
0
0.06
0.23
0.54
0.66
0.26
0.15
0
-0.02
0.22
-0.01
0.42
a Reference 36. b Reference 35.
expressed by eq 8,36 typically, using σp for polar substituent
•
constants and σJJ for spin delocalization constants.
Experimental Section
General Methods. ESR spectra were recorded on a Bruker ESP-
300E spectrometer (9.5-GHz frequency, 3300-G magnetic field, and
field modulation at 100 kHz) and were calibrated with perylene anion
radical in DME. 1H NMR (300 MHz) spectra were obtained on a Bruker
AC-300 FT spectrometer. 13C and 29Si NMR spectra were collected on
the machine at 75.5 and 59.6 MHz, respectively. Chemical shifts are
based on the resonances of the deuterated solvents. Mass spectra and
high-resolution mass spectra were obtained on a JEOL JMS 600W mass
spectrometer. Electronic spectra were recorded on a Milton-Roy SP3000
spectrometer. All solvents were dried and degassed over a potassium
mirror in vacuo prior to use.
To elucidate further in detail the electronic nature of the
transition state for the reactions of disilenes with haloalkanes,
we investigated the substituent effects on the reaction rate of
disilene 1c with benzyl chlorides (eq 9). The rate constants for
various benzyl chlorides are shown in Table 5 together with
pertinent substituent constants. A dual parameter correlation
•
using σp and σJJ constants (eq 10) was clearly better than the
single parameter correlation.
Electrochemistry. Voltammetric experiments were performed with
a Hokuto Denko Model HA-501 Potentiostat with a HA-104 Function
Generator in a glovebox (VAC MO-40-M). A conventional three-
electrode cell design was used, the experimental reference electrode
being an AgCl-coated silver wire. Potentials in this paper are referenced,
however, to the ferrocene/ferrocenium couple whose potential was fixed
to +0.61 V in THF.18a The working electrode was a Pt wire polished
by a standard method.
Substituent effects on these two reactions are parallel, while
the F values for the reaction of eq 9 are 3-4 times larger than
those for the reaction of eq 8, in accord with the reactivity-
selectivity principle. The ratio of polar and spin effects, Fp/FJJ,
for the reactions of 1c (eq 9) was 0.79, which was a little smaller
than the 1.09 ratio for the reactions of (Me3Si)3Si• (eq 7).35 As
shown in Chart 2, the resonance structures A-C will contribute
Materials. Tetrakis(2,4,6-triisopropylphenyl)disilene (1d),12 tet-
ramesityldisilene,42 and three tetrakis(trialkylsilyl)disilenes (1a-c)11
were prepared accoring to the literature.
Reactions of Disilenes with Haloalkanes. All the yields of products
2 and 3 were determined NMR spectroscopically after the complete
consumption of the disilenes.
The reactions of 1a-d with carbon tetrachloride gave the corre-
sponding 1,2-dichlorodisilanes 2a-d. In a typical procedure, to a
Schlenk tube (50 mL) containing 1d (30 mg, 3.45 × 10-5 mol) was
introduced a mixture of CCl4 (2.0 mL) and hexane (2.0 mL) through
a vacuum line at -50 °C. With increasing temperatures to room
temperature, the mixture turned from yellow to colorless rapidly. After
completion of the reaction, a constant amount of 1,3,5-trimethylbenzene
was added to the reaction mixture as an internal standard. The 1H NMR
analysis showed formation of the corresponding 1,2-dichlorodisilane
2d in 84% yield. Pure 2d was obtained after removal of solvent and
recrystallization. 2d: colorless crystals; mp 245 °C; 1H NMR (CDCl3)
δ -0.07 (d, J ) 6.7 Hz, 6 H), -0.06 (d, J ) 6.7 Hz, 6 H), 0.11 (d, J
Chart 2
to the transition states for both two-halogen-abstraction reactions.
The comparison of the Fp/FJJ values suggests that the relative
importance of spin structure A to the polar structure C in the
reaction of 1c is a little larger than that in the reactions of (Me3-
Si)3Si•.
D. Concluding Remarks. The origin of the biradical
character of disilenes can be discussed based on several grounds.
Teramae has shown theoretically that disilene has weak but
significant singlet biradical character as a result of the existence
of the UHF instability; the percent biradical character of parent
(37) (a) Teramae, H. J. Am. Chem. Soc. 1987, 109, 4140. (b) Teramae,
H.; Maxika, J. Private communication. Whereas the author has proposed
the biradical character as the origin of the trans-bent geometry of disilene,
the biradical nature may be taken as a consequence of the CGMT theory.38,39
(38) Carter, E. A.; Goddard, W. A., III J. Phys. Chem. 1986, 90, 998.
Carter, E. A.; Goddard, W. A., III J. Am. Chem. Soc. 1988, 110, 4077.
(39) (a) Trinquier, G.; Malrieu, J.-P. J. Am. Chem. Soc. 1987, 109, 5303.
(b) Malrieu, J.-P.; Trinquier, G. J. Am. Chem. Soc. 1989, 111, 5916. (c)
Trinquir, G. J. Am. Chem. Soc. 1990, 112, 2130.
(40) Wang, S. Y.; Borden, W. T. J. Am. Chem. Soc. 1989, 111, 7282.
(41) (a) Bartlett, P. D. Science 1968, 159, 833. (b) Bartlett, P. D.; Mallet,
J. J.-B. J. Am. Chem. Soc. 1976, 98, 143.
(42) Fink, M. J.; Michalczyk, M. J.; Haller, K. J.; West, R.; Michl, J.
Organometallics 1984, 3, 793.
(36) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165.

1682 J. Am. Chem. Soc., Vol. 123, No. 8, 2001
Kira et al.
) 6.7 Hz, 6 H), 1.07-1.20 (m, 42H), 1.31 (d, J ) 6.7 Hz, 6 H), 1.36
(d, J ) 6.7 Hz, 6 H), 2.68 (sept, J ) 6.7 Hz, 2 H), 2.78 (sept, J ) 6.7
Hz, 2 H), 2.92 (sept, J ) 6.7 Hz, 2 H), 3.21 (sept, J ) 6.7 Hz, 2 H),
3.29 (sept, J ) 6.7 Hz, 2 H), 3.55 (sept, J ) 6.7 Hz, 2 H), 6.74 (s, 4
H), 6.80 (s, 2 H), 7.06 (s, 2 H); 13C NMR (CDCl3) δ 22.5, 23.0, 23.1,
23.7, 23.96, 23.99, 24.1, 25.0, 25.4, 25.5, 26.8, 29.5, 33.7, 33.9, 34.2,
34.4, 36.5, 37.5, 122.7, 123.5, 124.3, 124.4, 130.1, 136.9, 150.2, 150.9,
152.8, 154.7, 155.8, 155.9; 29Si NMR (CDCl3) δ -4.45; MS (EI, 14
eV) m/z (%) 938 (0.5, M+), 903 (0.3, M+ - 35), 735 (1.8, M+ - 203),
469 (100). Anal. Calcd for C60H92Si2Cl2: C, 76.63; H, 9.86. Found:
C, 76.13; H, 9.95.
15.48, 15.7, 15.8 (SiCH(CH3)2), 20.1, 20.3, 20.5, 20.65, 20.69, 20.8
(SiCH(CH3)2), 31.4 (CH2Cl); 29Si NMR (C6D6) δ -46.5 (SiCH2Cl),
2.8 (SiMe(i-Pr)2), 4.1 (SiMe(i-Pr)2), 5.9 (SiCl); MS (EI, 14 eV) 656
(1.8, M+), 623 (100), 613 (59, M+ - 43).
1
Adduct of 1a with benzyl chloride (3a′′): 51%; a colorless oil; H
NMR (C6D6) δ 0.36 (s, 6 H), 0.43 (s, 6 H), 1.07-1.40 (m, 52 H), 1.52
(sept, J ) 6.7 Hz, 4 H), 3.13 (s, 2 H), 7.00 (m, 2 H), 7.12 (m, 1 H),
7.40 (d, J ) 7.4 Hz, 2 H); 13C NMR (C6D6) δ -5.2, -4.4 (SiCH3),
15.1, 15.3, 15.6, 15.7 (SiCH(CH3)2), 20.03, 20.09, 20.26, 20.29, 20.53,
20.60, 20.64, 20.68 (SiCH(CH3)2), 22.5 (SiCH2Ph), 125.0, 128.3, 129.8,
142.3; 29Si NMR (C6D6) δ 49.0 (SiCH2Ph), 2.3, 3.3 (SiMe(i-Pr)2), 10.1
(SiCl); MS (EI, 14 eV) 683 (4, M+ - Me), 663 (13, M+ - Cl), 656 (6,
M+ - (i-Pr)), 607 (6, M+ - (CH2Ph)), 59 (100).
Reaction of 1c with tert-Butyl Chloride Monitored by ESR. In a
glass tube (5 mm φ) containing 1c (10 mg, 1.46 × 10-5 mol) were
introduced degassed tert-butyl chloride (1.5 µL, 1.38 × 10-5 mol) and
benzene-d6 (0.5 mL) by distillation through a vaccum line. The tube
was sealed and left to react at room temperature in the dark and then
the ESR spectra were measured. The NMR spectroscopy revealed that
only 10% of 1c was consumed after reaction for 1 month. In addition
to a small amount of 2c, equal amounts of isobutene and isobutane
were detected by the NMR. Signal intensities of the ESR spectra
increased slowly for 6 months.
1
2a: 83%; colorless oil; H NMR (C6D6) δ 0.42 (s, 12 H), 1.19 (d,
J ) 7.2 Hz, 24 H), 1.23 (d, J ) 7.2 Hz, 24 H), 1.52 (sept, J ) 7.5 Hz,
8 H); 13C NMR (C6D6) δ -6.0 (SiCH3), 15.2, 15.4 (SiCH(CH3)2), 19.8,
20.1, 20.3, 20.5 (SiCH(CH3)2); 29Si NMR (C6D6) δ 3.9 (SiCl), 9.4
(SiMe(i-Pr)2); MS (EI, 14 eV) m/z (%) 642 (1.7, M+), 627 (47, M+
-
15), 607 (100, M+ - 35), 599 (52, M+ - 43), 513 (9.9, M+ - 119).
2c: 70%; colorless crystals; mp 170 °C; 1H NMR (C6D6) δ 1.35 (d,
J ) 6.7 Hz, 36 H), 1.37 (d, J ) 6.7 Hz, 36 H), 1.77 (sept, J ) 6.7 Hz,
12 H); 13C NMR (C6D6) δ 16.4 (SiCH(CH3)2), 22.0, 22.2 (SiCH(CH3)2);
29Si NMR (C6D6) δ 11.9 (Si(i-Pr)3), 21.3 (SiCl); MS (EI, 14 eV) m/z
(%) 711 (16, M+ - 43 (i-Pr)), 562 (100, M+ - 192), 377 (53). Anal.
Calcd for C36H84Si6Cl2: C, 57.15; H, 11.19. Found: C, 57.11;, H, 11.06.
1
2c′: 81%; a colorless oil; H NMR (C6D6) δ 1.37 (d, J ) 7.5 Hz,
36 H), 1.38 (d, J ) 7.5 Hz, 36 H), 1.84 (sept, J ) 7.5 Hz, 12 H); 13C
NMR (C6D6) δ 16.8 (SiCH(CH3)2), 22.2, 22.4 (SiCH(CH3)2); 29Si NMR
(C6D6) δ 5.2 (SiBr), 12.2 (Si(i-Pr)3).
Acknowledgment. This work was supported by the Ministry
of Education, Science, Sports, and Culture of Japan (Grant-in-
Aids for Scientific Research (A) No. 08404042, Scientific
Research (B) No.11440185, and Scientific Research on Priority
Areas (A), “The Chemistry of Inter-element Linkage” (No.
09239101)).
Typical procedures for the reactions of disilenes 1b with chloroform
and dichloromethane are presented in ref 6. A similar reaction of 1a
with chloroform gave an adduct 3a (70%) in addition to 2a (26%). 3a:
1
colorless oil; H NMR (C6D6) δ 0.42 (s, 6 H), 0.48 (s, 6 H), 1.13-
1.28 (m, 48 H), 1.54 (sept, J ) 6.8 Hz, 4 H), 1.56 (sept, J ) 6.8 Hz,
4 H), 6.18 (s, 1 H); 13C NMR (C6D6) δ -5.0, -3.8 (SiCH3), 15.2,
15.4, 15.5, 15.6 (SiCH(CH3)2), 20.0, 20.2, 20.57, 20.62, 20.70, 20.77
(SiCH(CH3)2), 65.1 (SiCHCl2); 29Si NMR (C6D6) δ 18.8 (SiCHCl2),
2.9, 3.8 (SiMe(i-Pr)2), 5.4 (SiCl); MS (EI, 14 eV) 675 (22, M+ - 15),
655 (74, M+ - 35), 647 (16, M+ - 57), 607 (10, M+ - 83), 73 (100).
Product of reaction of 1a with dichloromethane. 3a′: 93%; a
Supporting Information Available: Cyclic voltagrams of
1a-d and tetramesityldisilene in THF and plots of ln A vs t for
the reaction of 1d with chloroform in hexane and the dependence
of the pseudo-first-order rates for the reaction of 1d with
chloroform in hexane on concentrations of chloroform and
temperatures (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
1
colorless oil; H NMR (C6D6) δ 0.39 (s, 6 H), 0.42 (s, 6 H), 1.15-
1.28 (m, 48 H), 1.47 (sept, J ) 6.8 Hz, 4 H), 1.48 (sept, J ) 6.8 Hz,
4 H), 3.68 (s, 2 H); 13C NMR (C6D6) δ -5.5, -4.3 (SiCH3), 15.42,
JA002798S