Radical isomerization of 9,10-dihydro-9,10-disilaanthracene derivatives: Stabilization factor of silyl radicals

Article abstract of DOI:10.1016/S0022-328X(00)00942-6

In order to examine the nature of the silyl radical in the disilaanthracene framework, separation and radical isomerization of the cis and trans isomers of the disilaanthracene derivatives were investigated. When a pentane solution of the disilaanthracene derivative was irradiated with a 400 W mercury lamp in the presence of DTBP for 4 h, both isomers readily isomerized to give a mixture of cis and trans isomers in a ratio of ca. 50:50. However, no such isomerization reaction of 9,10-dihydro-9-silaanthracene derivatives occurred under similar conditions.

Full text of DOI:10.1016/S0022-328X(00)00942-6

Journal of Organometallic Chemistry 626 (2001) 32–36
www.elsevier.nl/locate/jorganchem
Radical isomerization of 9,10-dihydro-9,10-disilaanthracene
derivatives: stabilization factor of silyl radicals
Kozaburo Nishiyama a,*, Makoto Oba a, Hidenori Takagi a, Takeshi Saito a,
Yoko Imai b, Izumi Motoyama b, Shigeru Ikuta c, Hiroshi Hiratsuka d
a Department of Material Science and Technology, Tokai Uni6ersity, Numazu 317, Nishino, Shizuoka 410-0395, Japan
b Department of Engineering and Technology, Kanagawa Uni6ersity, Yokohama, Kanagawa 221-0802, Japan
c Educational Technology, Faculty of Engineering, Tokyo Metropolitan Uni6ersity, Hachioji, Tokyo 192-0397, Japan
d Department of Chemistry, Gunma Uni6ersity, Kiryu, Gunma 376-8515, Japan
Received 20 October 2000; received in revised form 8 December 2000; accepted 9 December 2000
Abstract
In order to examine the nature of the silyl radical in the disilaanthracene framework, separation and radical isomerization of
the cis and trans isomers of the disilaanthracene derivatives were investigated. When a pentane solution of the disilaanthracene
derivative was irradiated with a 400 W mercury lamp in the presence of DTBP for 4 h, both isomers readily isomerized to give
a mixture of cis and trans isomers in a ratio of ca. 50:50. However, no such isomerization reaction of 9,10-dihydro-9-silaan-
thracene derivatives occurred under similar conditions. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Silyl radical; Radical isomerization; 9,10-Disilaanthracenes; Conjugation
1. Introduction
effective and versatile reducing agent, of which the Si
radical is believed to be dramatically stabilized by
hyperconjugation with adjacent trimethylsilyl groups
and has a longer lifetime than the usual silyl radicals.
We previously reported the utilization of 9,10-dihydro-
9,10-disilaanthracene derivatives as reducing agents in
radical-based dehalogenation and deoxygenation reac-
tions, in which the two silicon atoms at the 9- and
10-positions were essential to generate the silyl radical
efficiently [3]. In this paper, we would like to report the
separation and radical isomerization reaction of cis-
and trans-9,10-dimethyl-(3a), 9,10-diphenyl-9,10-dihy-
dro-9,10-disilaanthracene (3b), and 9,10-diphenyl-9,10-
dihydro-9-silaanthracene (7) in order to confirm the
stabilization factor, and a calculation by density func-
tional calculations (B3LYP/cc-pVDZ) for the silyl
radicals.
Although many reductive reactions of organic halides
and alcohols involving free radicals as intermediates
have employed tributyltin hydride as a reducing agent,
the replacement of such toxic and relatively expensive
organotin compounds has been performed with other
hydrogen donors. However, trialkyl- or triarylsilicon
compounds are not appropriate for the purpose under
mild conditions. Formation of silyl radicals from these
compounds requires relatively drastic conditions be-
cause such silyl radicals are relatively unstable. On the
other hand, the nature of these silyl radicals is well
established [1], and they are generally believed to have
a pyramidal configuration, in contrast to the usual
carbon radicals. For example, the silyl radicals derived
from an optically active hydrosilane undergo chlorine-
abstraction reaction mostly with retention of configura-
tion. Recently, some organosilicon compounds with
reduced SiꢀH bond strength were discovered [2]. In
particular, tris(trimethylsilyl)silane has proved to be an
2. Results and discussion
The synthesis and separation of cis- and trans-9,10-
dimethyl-9,10-dihydro-9,10-disilaanthracene (cis-3a and
trans-3a) were reported by Corey et al. [4]. By applying
* Corresponding author. Fax: +81-559-681155.
E-mail address: nishiyam@wing.ncc.u-tokai.ac.jp (K. Nishiyama).
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00942-6

K. Nishiyama et al. / Journal of Organometallic Chemistry 626 (2001) 32–36
33
Scheme 1.
the procedure, we prepared the corresponding diphenyl
derivatives, cis-3b and trans-3b, novel 9,10-dihydro-
9,10-disilaanthracenes, as shown in Scheme 1. 9,10-
Diphenyl-9,10-dihydro-9-silaanthracenes (cis-7 and
trans-7), the monosila analogues of cis-3b and trans-3b,
were prepared by a modified procedure of Winkel et al.
[5] (Scheme 2). The 2,2%-dibromotriphenylmethane (6)
was condensed with phenyldichlorosilane via the corre-
sponding dilithio derivatives to afford 9-silaanthracene
7 in 43% yield as a 51:49 mixture of cis and trans
isomers. The separation of the isomers was performed
by repeated column chromatography on silica gel and
the structure of the trans-7 was unambiguously confi-
rmed by X-ray crystallographic analysis (vide infra).
Recently, Matsumoto et al. reported that a radical
isomerization reaction of cis- and trans-9,10-di-tert-bu-
tyl-9,10-dihydro-9,10-disilaanthracene gave a mixture
of trans and cis isomers [6]. From kinetic studies, they
concluded the cis-isomer was slightly more stable by 0.8
kcal/mol than the trans isomer at 20°C and the isomer-
ization reaction could be explained by the stabilization
of the silyl radicals, due to bulky tert-butyl sub-
stituents. As a part of our interest in the reactivity of
the disilaanthracenes, we tried a similar reaction of
compounds 3a, 3b, and 7. When a pentane solution of
trans-9,10-dimethyl-9,10-dihydro-9,10-disilaanthracene
(3a) and DTBP was irradiated with a 400-W medium-
pressure mercury lamp for 1 h, a mixture of trans- and
cis-3a (91:9) was obtained and the recovery of the total
amounts of 3a was 92%. After an additional 3 h, the
reaction gave a mixture of 52:48 of trans- and cis-3a,
and the recovery was 79%. On the other hand, cis-3a
gave a mixture of trans- and cis-3a (35:65) after 1 h
under similar conditions and the recovery was 82%.
The ratio of trans- and cis-3a became 52:48 after 2 h
and the recovery was 73%. These results show that
trans- and cis-3a isomerize to each other to reach an
equilibrium after 2–4 h. The results are compiled in
Table 1. Phenyl derivatives 3b also gave similar results.
Both isomers gave a 53:47 mixture of trans- and cis-3b
after 4 h. For the methyl and phenyl substituted dihy-
drodisilaanthracene framework, trans- and cis-isomers
are energetically almost equal, in contrast to the tert-
butyl substituted compound [6]. Under similar condi-
tions, trans- and cis-9,10-diphenyl-9,10-dihydro-
9-silaanthracene (7) did not give the isomerized product
at all. The stereochemistry of these compounds was
confirmed by X-ray crystallographic analysis. These
findings show the isomerization reaction is attributed to
the stability of the silyl radicals generated and is not
due to the bulkness of the substituent, indicating the
importance of the presence of two silicon atoms at both
the 9- and 10-positions. In the previous paper, we also
reported that 9,10-dihydro-9-methyl-9-sila-10-thiaan-
thracene had a reductive ability similar to that of the
disilaanthracene analogue under radical conditions [3].
From these facts, it is deduced that neither hyperconju-
gation nor steric hindrance contributes to stabilize the
Scheme 2.

34
K. Nishiyama et al. / Journal of Organometallic Chemistry 626 (2001) 32–36
Table 1
3. Experimental
Isomerization of silaanthracene derivatives a
3.1. General
Silaanthracene
Time (h)
trans/cis
Recovery (%)
trans-3a
0
1
2
4
100:0
91:9
56:44
52:48
–
92
88
79
GC measurements were performed on a Shimadzu
GC-18A and
a GL Sciences GC-380 gas chro-
matograph using a 50 m×0.25 mm methyl silicone
capillary column (Quadrex). TLC was carried out on
Merck silica gel 60 F254. Flash column chromatography
was performed on Merck silica gel 60 (230–400 mesh).
Infrared spectra were recorded on a Perkin–Elmer 1600
cis-3a
0
1
2
4
0:100
35:65
52:48
53:47
–
82
73
65
1
infrared spectrometer. H-, 13C-, and 29Si-NMR spectra
trans-3b
cis-3b
0
1
2
4
96:4
73:27
65:35
53:47
–
85
78
69
were measured in CDCl3 solution on a Varian UNITY-
400 spectrometer. All chemical shifts are reported as
d-values (ppm) relative to residual chloroform (lH
7.26), the central peak of deuteriochloroform (lC 77.0)
or tetramethylsilane (lSi 0.0). J-Values are given in Hz.
GC-MS (EI) was measured with the direct combination
of GC (Hewlett–Packard GC 5890 Series II with a 25
m×0.25 mm methyl silicone capillary column) and a
JEOL JMS-AX-500 spectrometer with DA7000 data
system.
0
1
2
4
2:98
8:92
51:49
53:47
–
53
37
15
trans-7
cis-7
0
4
100:0
100:0
–
91
0
4
0:100
0:100
–
82
3.2. 9,10-Diphenyl-9,10-dihydro-9,10-disilaanthracene
(3b)
a In the absence of DTBP, the isomerization reaction did not occur.
silyl radicals in these reactions, although they would
have a pyramidal or a planar configuration.
To a solution of 1,2-dibromobenzene (1, 11.8 g, 50.0
mmol) in THF (50 ml) and diethyl ether (50 ml) was
added 1.6 M butyllithium in hexane (35 ml, 55 mmol)
at −110°C under argon atmosphere. After being
stirred for an additional 20 minutes, a solution of
phenyldichlorosilane in THF (25 ml) and diethyl ether
(25 ml) was added dropwise to the mixture and stirred
for 2 h. Then the reaction mixture was allowed to warm
up to room temperature (r.t.) and quenched with a
mixture of conc. HCl (5 ml) and ethanol (12.5 ml). The
mixture was diluted with diethyl ether and the organic
layer was washed with brine, dried over MgSO4, and
evaporated. The residue was purified by column chro-
matography on silica gel (hexane) to give the silane 2b
Next, we calculated spin density of the silyl radicals
(8 and 9) by Density Functional Calculations method
(B3LYP/cc-pVDZ). The spin densities of the benzene
rings of 9 are slightly larger than that of 8. It shows
that radical spin of the silyl radical 9, having a conve-
nient ring conformation for conjugation, generated
from disilaanthracene derivatives is conjugated to the
benzene rings, but such a contribution would be rela-
tively small in the monosilaanthracene framework.
1
(7.32 g, 70%) as a colorless solid, m.p. 89–90°C. H-
NMR (CDCl3) l 5.74 (s, 1 H), 7.26 (m, 4 H), 7.30 (m,
2 H), 7.40 (m, 2 H), 7.47 (m, 1 H), 7.55 (m, 2 H), 7.59
(m, 2 H). 13C-NMR (CDCl3) l 126.7, 128.2, 130.0,
131.4, 131.8, 132.6, 135.3, 135.9, 138.7. 29Si-NMR
(CDCl3) l −19.9. IR (KBr) wSiꢀH 2166 cm−1. HRMS
m/z 417.9199 (M+, calcd for C18H Br81BrSi 417.9211).
79
14
To a solution of 2b (2.99 g, 7.15 mmol) in THF (70
While the stabilization factor of the silyl radicals
could not be precisely confirmed, it is proposed such
stabilization arises from conjugation with benzene rings
(delocalization of an electron on the silicon atom), due
to suitable ring conformation of the disilaanthracene
framework. Thus, stabilization by hyperconjugation
and steric hindrance can be ruled out as a less impor-
tant factor in our reaction systems.
ml) was added 1.6 M butyllithium in hexane (10 ml,
16.0 mmol) at −78°C. After being stirred for 1 h,
phenyldichlorosilane (2.50 g, 14.1 mmol) was added to
the solution and the resultant reaction mixture was
allowed to warm to r.t. and refluxed overnight. Then,
the mixture was poured into water and extracted with
ether. The organic layer was washed with water, dried
over MgSO4, and evaporated. The crude product was

K. Nishiyama et al. / Journal of Organometallic Chemistry 626 (2001) 32–36
35
cis-7, m.p. 110–111°C. 1H-NMR (CDCl3) l 5.40 (s, 1
H), 5.49 (s, 1 H), 6.80 (m, 2 H), 6.92 (m, 3 H), 7.15 (m,
2 H), 7.25 (m, 3 H), 7.34 (m, 2 H), 7.43 (m, 2 H), 7.48
(m, 2 H), 7.83 (m, 2 H). 13C-NMR (CDCl3) l 56.8,
125.8, 126.1, 127.7, 128.0, 128.4, 129.4, 130.06, 130.14,
133.5, 135.1, 136.2, 145.7, 149.6, 1 C was overlapped
elsewhere. 29Si-NMR (CDCl3) l −33.6. IR (KBr) wSiꢀH
2127 cm−1. HRMS m/z 348.1371 (M+, Calc. for
C25H20Si 348.1334).
purified by column chromatography on silica gel (hex-
ane) to give 9,10-disilaanthracene derivative 3b as a
54:46 mixture of cis and trans isomers (2.02 g, 78%).
The trans-3b was obtained as a colorless powder by
crystallization of the mixture from EtOH, m.p. 112–
1
114°C. H-NMR (CDCl3) l 5.46 (s, 2 H), 7.39 (m, 10
H), 7.60 (m, 4 H), 7.66 (m, 4 H). 13C-NMR (CDCl3) l
128.2, 129.0, 129.9, 133.9, 135.7, 135.9, 140.6. 29Si-
NMR (CDCl3) l −30.3. IR (KBr) wSiꢀH 2121 cm−1
.
HRMS m/z 364.1126 (M+, calcd for C24H20Si2
364.1104).
1
trans-7, m.p. 130–131°C. H-NMR (CDCl3) l 5.12
(s, 1 H), 5.53 (s, 1 H), 7.04 (m, 2 H), 7.11 (m, 1 H), 7.20
(m, 4 H), 7.41 (m, 6 H), 7.50 (m, 1 H), 7.54 (m, 2 H),
7.63 (m, 2 H). 13C-NMR (CDCl3) l 57.5, 125.9, 126.0,
128.0, 128.19, 128.23, 129.5, 129.9, 130.3, 131.5, 132.1,
135.7, 136.7, 145.1, 149.7. 29Si-NMR (CDCl3) l −34.3.
IR (KBr) wSiꢀH 2140 cm−1. HRMS m/z 348.1340 (M+,
Calc. for C25H20Si 348.1334).
Isolation of the cis isomer was performed according
to the procedure of Corey [4]. Thus, a solution of 3b
(289 mg, 0.794 mmol) in CCl4 (30 ml) in the presence of
benzoyl peroxide (19 mg, 0.079 mmol) was refluxed for
4.5 h. A direct analysis of the reaction mixture by GC
analysis showed completion of the chlorination. To the
cooled solution was added a mixture of cis-2-butene-
1,4-diol (1.05 g, 11.9 mmol) and triethylamine (400 mg,
3.97 mmol) and the reaction mixture was stirred at r.t.
overnight. The mixture was washed by water, dried
over MgSO4, and evaporated. The residue was chro-
matographed on silica gel (hexane:AcOEt=95:5) to
3.4. Isomerization of silaanthracene deri6ati6es
A pentane solution of silaanthracene derivative (ca.
0.02–0.05 M, 3–5 ml) in the presence of DTBP (1.4–
4.9 equivalents) in a Pyrex tube was irradiated with a
400-W medium-pressure mercury lamp. After an appro-
priate reaction time as shown in Table 1, the ratio of
the trans and cis isomers was determined by a direct
1
give the alkoxide 5b (105 mg) in 28% yield. H-NMR
(CDCl3) l 4.31 (m, 4 H), 5.33 (m, 2 H), 7.36 (m, 4 H),
7.47 (m, 10 H), 7.79 (m, 4 H).
The dialkoxy 9,10-dihydro-9,10-disilaanthracene
derivative 5b (891 mg, 1.98 mmol) was reduced by
LiAlH4 (300 mg, 7.92 mmol) in refluxing ether (70 ml)
for 3 h. The cooled reaction mixture was quenched with
water and the insoluble materials were filtered off. The
filtrate was dried over MgSO4 and evaporated. The
residue was chromatographed on silica gel (hexane) to
1
analysis of the reaction mixture using H-NMR spec-
troscopy. The recovered yields were obtained by purifi-
cation of the crude products by column
chromatography on silica gel (hexane).
3.5. X-ray structure determination of trans-7
afford
cis-9,10-diphenyl-9,10-dihydro-9,10-disilaan-
A colorless prismatic crystal of C25H20Si, M=
348.52, having approximate dimensions of 0.30×
0.20×0.50 mm was mounted on a glass fiber. All
measurements were made on a Rigaku AFC6S diffrac-
tometer with graphite monochromated Mo–Ka (u=
thracene (cis-3b, 681 mg, 94%) as a colorless solid, m.p.
100–104°C. H-NMR (CDCl3) l 5.50 (s, 2 H), 7.28 (m,
1
4 H), 7.36 (m, 2 H), 7.42 (m, 4 H), 7.49 (m, 4 H), 7.70
(m, 4 H). 13C-NMR (CDCl3) l 128.0, 129.0, 129.7,
134.0, 135.6, 135.8, 140.5. 29Si-NMR (CDCl3) l −30.7.
IR (KBr) wSiꢀH 2115 cm−1. HRMS m/z 364.1139 (M+,
calcd for C24H20Si2 364.1104).
,
0.71069 A) radiation. Monoclinic, a=10.509(3),
,
b=8.463(2), c=21.860(6) A, i=100.40(2)°, V=
3
,
1912.2(9) A (by least-squares refinement using the
setting angles of 24 carefully centered reflections in the
range 22.48B2qB24.54°), space group P21/n (c14),
Z=4, Dcalc=1.211 g cm−3. The data were collected at
23°C using the ꢀ–2q scan technique to a maximum 2q
value of 40.0°. Omega scans of several intense reflec-
tions, made prior to data collection, had an average
width at half-height of 0.36° with a take-off angle of
6.0°. Scans of (1.00+0.30 tan q)° were made at a speed
of 2.0° min−1 (in ꢀ). Of the 2079 reflections measured,
1942 were unique (Rint=0.009). The intensities of three
representative reflections were measured after every 150
reflections. Over the course of data collection, the stan-
dards increased by 0.1%. The linear absorption coeffi-
cient, v, for Mo–Ka radiation is 1.3 cm−1. An
empirical absorption correction based on azimuthal
3.3. 9,10-Diphenyl-9,10-dihydro-9-silaanthracene (7)
To a solution of 2,2%-dibromotriphenylmethane (6,
894 mg, 2.22 mmol) in ether (40 ml) was added 1.6 M
butyllithium in hexane (3.5 ml, 5.6 mmol) at r.t.. Then,
phenyldichlorosilane (2.6 g, 14.7 mmol) was added and
the mixture was stirred overnight. Direct analysis of the
1
reaction mixture by H-NMR spectroscopy showed the
ratio of the cis and trans isomers was 51:49. The residue
was separated by column chromatography on silica gel
(hexane). The fast fraction contained the cis isomer
(50.0 mg, 6%) and further elution gave a mixture of cis
and trans isomers (271 mg, 35%). Finally, the trans
isomer (17.0 mg, 2%) was obtained from the last frac-
tion. The total yield of the compound 7 was 43%.

36
K. Nishiyama et al. / Journal of Organometallic Chemistry 626 (2001) 32–36
1EZ, UK (Fax: +44-1223-336033; e-mail: deposite@
ccdc.cam.ac.uk or www: http://www.ccdc.cam.ac.uk).
scans of several reflections was applied which resulted
in transmission factors ranging from 0.97 to 1.00. The
data were corrected for Lorentz and polarization ef-
fects. The structure was solved by direct methods and
expanded using Fourier technique. The non-hydrogen
atoms were refined anisotropically. Hydrogen atoms
were included but not refined. The final cycle of full-
matrix least-squares refinement was based on 1419 ob-
served reflections (I\3.00|(I)) and 236 variable
parameters and converged with unweighted and
weighted agreement factors of R=0.041 and Rw=
0.031. The standard deviation of an observation of unit
weight was 3.23. The weighting schemes were based on
counting statistics and included a factor (p=0.004) to
downweight the intense reflections. Plots of Sꢀ(ꢀFoꢀ−
ꢀFcꢀ)2 versus ꢀFoꢀ, reflection order in data collection,
sin q/u and various classes of indices showed no un-
usual trends. The maximum and minimum peaks on the
final difference Fourier map corresponded to 0.13 and
Acknowledgements
One of the authors (K.N.) gratefully acknowledges
financial support of a Grant-in-Aid for Scientific Re-
search (No. 09640646) from the Ministry of Education,
Science, Sports and Culture of Japan.
References
[1] (a) H. Sakurai, in: J.K. Kochi (Ed), Free Radicals, Vol II, Wiley,
New York, 1973, p. 741. (b) C. Chatgilialoglu, Chem. Rev. 95
(1995) 1229. (c) H. Sakurai, M. Murakami, M. Kumada, J. Am.
Chem. Soc. 91 (1969) 519. (d) H. Sakurai, M. Murakami, Bull.
Chem. Soc. Jpn. 50 (1977) 3384. (e) C. Chatgilialoglu, K.U.
Ingold, J.C. Scaiano, J. Am. Chem. Soc. 104 (1982) 5123.
[2] (a) C. Chatgilialoglu, M. Guerra, A. Guerrini, G. Seconi, K.
Clark, D. Griller, J.M. Kanabus-Kaminska, J.A. Martinho-
Simoes, J. Org. Chem. 57 (1992) 2427. (b) C. Chatgilialoglu, A.
Guerrini, M. Lucarini, J. Org. Chem. 57 (1992) 3405. (c) J.M.
Kanabus-Kaminska, J.A. Hawari, D. Griller, C. Chatgilialoglu, J.
Am. Chem. Soc. 109 (1987) 5267. (d) M. Ballestri, C. Chat-
gilialoglu, K.B. Clark, D. Griller, B. Giese, B. Kopping, J. Org.
Chem. 56 (1991) 678.
[3] (a) M. Oba, K. Nishiyama, J. Chem. Soc. Chem. Commun. (1994)
1703. (b) M. Oba, Y. Kawahara, R. Yamada, H. Mizuta, K.
Nishiyama, J. Chem. Soc. Perkin Trans. 2 (1996) 1843.
[4] K.M. Welsh, J.Y. Corey, Organometallics 6 (1987) 1393.
[5] Y.V.D. Winkel, B.L.M.V. Baar, H.M.M. Bastiaans, F. Biekel-
haupt, M. Schenkel, H.B. Stegmann, Tetrahedron 46 (1990) 1009.
[6] S. Kyushin, T. Shinnnai, T. Kubota, H. Matsumoto,
Organometallics 6 (1997) 3800.
−0.23 e−
−3, respectively. Neutral atom scattering
,
A
factors were taken from Cromer and Waber [7].
Anomalous dispersion effects were included in Fcalc [8];
the values for Df % and Df %% were those of Creagh and
McAuley [9]. All calculations were performed using the
TEXSAN crystallographic software package of Molecu-
lar Structure Corporation.
4. Supplementary material
Crystallographic data for the structural analysis have
been deposited with the Cambridge Crystallographic
Data Centre, CCDC no. 139320. Copies of this infor-
mation may be obtained free of charge from The
Director, CCDC, 12 Union Road, Cambridge, CB2
[7] D.T. Cromer, J.T. Waber, International Tables for X-ray Crystal-
lography: Table 2.2 A, vol. IV, Kynoch Press, Birmingham, 1974.
[8] J.A. Ibers, W.C. Hamilton, Acta Crystallogr. 17 (1964) 781.
[9] D.C. Creagh, W.J. McAuley, International Tables for Crystallog-
raphy: Table 4.2.6.8, vol. C, Kluwer Academic Publishers,
Boston, 1992.
.