The effect of ring-size on the electrochemical oxidation of perethylcyclopolysilanes [(Et2Si)n]

Article abstract of DOI:10.1016/S0022-328X(98)00975-9

For evaluating the net ring-size effect on the electrochemical properties of cyclic polysilanes, four perethylpolysilanes, namely octaethylcyclotetrasilane (Et₂Si)₄ (I), decaethylcyclopentasilane (Et₂Si)₅ (II), dodecaethylcyclohexasilane (Et₂Si)₆ (III) and tetradecaethylcycloheptasilane (Et₂Si)₇ (IV), were studied by cyclic voltammetry and controlled potential electrolysis. In addition, the effects of the nature of electrolyte, amount of electricity consumption and anode material on the electrolysis outcome was demonstrated for compound II.

Full text of DOI:10.1016/S0022-328X(98)00975-9

Journal of Organometallic Chemistry 574 (1999) 11–18
The effect of ring-size on the electrochemical oxidation of
ꢀ
perethylcyclopolysilanes [(Et Si) ]
2
n
a
a,
b,1
Zeng-Rong Zhang , James Y. Becker *, Robert West
a
Departrnent of Chemistry, Ben-Gurion Uni6ersity of the Nege6, Beer-She6a 84105, Israel
b
Department of Chemistry, Uni6ersity of Wisconsin, Madison, WI 53706, USA
Received 28 April 1998
Abstract
For evaluating the net ring-size effect on the electrochemical properties of cyclic polysilanes, four perethylpolysilanes, namely
octaethylcyclotetrasilane (Et Si) (I), decaethylcyclopentasilane (Et Si) (II), dodecaethylcyclohexasilane (Et Si) (III) and tetrade-
2
4
2
5
2
6
caethylcycloheptasilane (Et Si) (IV), were studied by cyclic voltammetry and controlled potential electrolysis. In addition, the
2
7
effects of the nature of electrolyte, amount of electricity consumption and anode material on the electrolysis outcome was
demonstrated for compound II. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Cyclic polysilanes; Electrochemical oxidation; Ring-size effect
1. Introduction
hindrance and the nature of the heteroatoms included
in the ring.
Peralkylpolysilanes are of special interest because
Peralkyl derivatives of polysilanes undergo irre-
versible electrochemical oxidation, with oxidation po-
tentials which are linearly related to first ionization
energies [5]. Also, the oxidation potentials decrease with
increasing chain length, and are correlated with UV
spectra and with Sandorfy calculations [6].
they exhibit unique electronic, physical and chemical
properties [1]. In particular, cyclic polysilanes mimic the
behavior of aromatic hydrocarbons, by having elec-
tronic transitions in the UV–vis region, forming anion-
and cation-radicals upon respective reduction and oxi-
dation, in which the odd electron is distributed over the
ring, forming charge-transfer complexes with y-accep-
tors, and exhibit substituent effects. A recent review
describes the synthesis, structure, chemical reactions
and photolysis of cyclic polysilanes [2].
Earlier papers from our laboratory [7–9] described
the outcome of preparative scale electrolyses of differ-
ent cyclic polysilanes (Si –Si ), with different sub-
3
9
stituents (Me, Et, Pr, t-Bu and aryl). In the presence of
tetraalkylammonium tetrafluoroborate salts they under-
Chemical oxidation of cyclic polysilanes by organic
peracids [3,4] lead to the formation of cyclic siloxanes,
due to the sequential insertion of oxygen atoms into the
Si–Si bonds of the cyclic polysilanes. Their reactivity is
determined by various factors, such as ring size, the
nature of the substituents attached to the silicon, steric
went ring opening followed by further Si–Si bond
−
cleavage and reaction with BF , forming h,ꢀ-difl-
4
uorosilanes, F–(SiR ) –F, as the major products [7].
2
n
However, their anodic oxidation in the presence of
other electrolytes which do not contain fluorinated
−
4
−
anions (ClO , HSO , AcO ), led to both oxygen
4
insertion and ring-opening processes to form mainly
cyclic and linear siloxanes [8]. In the case of small rings
ꢀ
Dedicated to the memory of the late Professor Rokuro Okawara.
Corresponding author. Fax: +972-7-472925/472943.
Also corresponding author.
[
9], the anodic oxidation of [t-Bu(Me)Si] led to oxygen
*
4
1
insertion process and gave cyclic siloxanes as the major
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(98)00975-9

12
Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
products, whereas (Mes Si) underwent ring opening
2
3
followed by further Si–Si bond cleavage and a reaction
with the electrolyte to generate mostly linear products
of type: X–(Mes Si) –Y (n=1, 2; X, Y=mesityl, H,
2
n
OH or F).
Since all cyclic polysilanes studied previously con-
tained different substituents, the net effect of ring-size
on their electrochemical behavior is not well under-
stood. To shed some light on this particular issue, the
present paper reports on the electrochemical oxidation
of four different ring-size perethylcyclopolysilanes,
namely octaethylcyclotetrasilane (Et Si) (I), decaethyl-
2
4
cyclopentasilane (Et Si) (II), dodecaethylcyclohexasi-
2
5
lane (Et Si) (V) and tetradecaethylcycloheptasilane
2
6
(
Et Si) (VI), which all share in common the same alkyl
2
7
substituent (an ethyl group). In addition, the effects of
the nature of electrolyte, anode material and electricity
consumption is reported for II, as a representative
example.
Fig. 1. Representative cyclic voltammograms.
Electrolyses were conducted under nitrogen atmo-
sphere at controlled potential of 0.95–1.20 V (vs Ag
wire), on Pt foil or graphite anode. The workup in-
volved stripping of the solvent followed by water addi-
tion and extraction of the product mixture into diethyl
ether, in which the electrolytes used were insoluble. In
general, the mixture of products obtained after phase
2. Experimental details
2.1. Chemicals
Perethylcyclopolysilanes I–IV were synthesized ac-
separation, drying over MgSO , filtration and evapora-
4
cording to published methods [14]. Tetrabutylammo-
tion, was submitted to GC-MS and HRMS analyses. A
Finnigan 4020 Quadrupole mass spectrometer was used
for routine mass spectral measurements and a high
resolution Fison VG AutoSpec for high resolution mea-
surements (HRMS), coupled with HP 5890 series II gas
chromatograph (GC-MS) was used also.
−
nium perchlorate (TBA+ClO , from Aldrich),
4
−
4
tetrabutylammonium hydrogen sulfate (TBA+HSO ,
from Fluka) and tetraethylammonium tetrafluorobo-
−
4
rate (TEA+BF , from Aldrich), were used as sup-
porting electrolytes without further purification.
Analytical grade CH Cl and CH CN (distilled from
2
2
3
CaH and P O , respectively).
2
2
5
3
. Results and discussion
2.2. Electrochemical measurements
3.1. Cyclic 6oltammetry
Cyclic voltammetric experiments were performed us-
ing a conventional three-electrode cell under nitrogen
atmosphere. The working electrode was a platinum disk
electrode (ca. 1 mm diam.). A quasi-reference electrode
Fig. 1 shows representative cyclic voltammograms
−1
for I and II, which are recorded in 0.1 mol l
TBA+
−
HSO4 –CH Cl –CH CN (4:1/v:v) on Pt. The results
2
2
3
(
Ag/AgCl) was used. The auxiliary electrode studied in
for compounds I–IV are summarized in Table 1. While
only one irreversible oxidation wave is observed for
perethylcyclopolysilane II, III and IV, the cyclic polysi-
lane I exhibits two irreversible oxidation waves with a
potential difference between them of about 300 mV. It
is noteworthy that there is a little difference only in the
these was a Pt electrode. The concentration of perethyl-
−
1
cyclopolysilane was 0.5–1 mmol l , while that of the
−
1
supporting electrolyte was 0.1 mol l . An H-type
two-compartment cell equipped with medium glass frit
as a membrane was used for preparative electrolysis.
The anode compartment contained a polished silver
wire quasi-reference electrode, immersed in electrolyte
solution in a glass cylinder with a fine glass frit at its
end. Typically, the anodic compartment contained
Table 1
a
Anodic peak potentials of perethylcyclopolysilanes
0.05–0.1 mmol of perethylcyclopolysilane dissolved in
I
II
III
IV
about 25 ml of solvent/electrolyte solution. For electro-
chemical measurements, Princeton Applied Research
1
.15
1.35
1.40
1.40
(
PAR) potentiostat Model 173, PAR Universal Pro-
a
Anodic peak potentials are in V vs Ag ꢀ AgCI reference electrode;
grammer Model 175 and Yokogawa XY-recorder
Model 3068, were employed.
−
4
working electrode: Pt; electrolyte: TBA+HSO ; solvent CH Cl –
CH CN (4:1/v:v); under N atmosphere.
2
2
3
2

Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
13
Table 2
a
The effect of structure on electrochemical oxidation of perethylcyclopolysilanes
Productsc
Yield (%)b
MS (EI)
I
II
III
IV
ACP (V)
0.95
1.2
1.2
1.2
Starting material
–
87
9
–
–
45
3
10
38
4
–
23
60
8
–
4
–
22
67
6
–
5
+
[
[
[
[
[
[
Et Si] O (1)
360(M ), 331(100%), 303, 273
2
4
+
Et Si] O (2)
376(M ), 347(100%), 289, 87
2
4
2
+
Et Si] O (3)
446(M ), 417, 115, 87(100%)
2
5
+
Et Si] O (4)
478(M ), 449, 172(100%), 87
2
5
3
+
Et Si] O (5)
–
–
–
–
392(M ), 363(100%), 335, 87
2
4
3
+
Et Si] O (6)
5
–
408(M ), 379(100%), 351
2
4
4
a
Electrolyses were conducted under N on Pt; applied controlled potentials (ACP) are in V vs Ag quasi-reversible reference electrode; electrolyte
2
−
1
−
−1
solution: CH Cl –CH CN (4:1/v:v) solution containing 0.1 mol l TBA+HSO4 ; electricity consumption stopped at 1.2 F mol arbitrarily in
2
2
3
order to compare the electrochemical reactivity of cyclopolysilanes.
b
Yields reported are relative and estimated by GLC.
High resolution MS results: found (calc.) for 1 (C16H41OSi ): 361.2339 (361.2309); 2 (C16H41O Si ): 377.2250 (377.2258); 3 (C20H51Osi5)
c
4
2
4
4
47.2891 (447.2879); 4 (C20H51O Si ) 479.2797 (479.2777); 5 (C16H41O Si ) 393.2245 (393.2207); 6 (C16H41O Si ) 409.2168 (409.2156).
3 5 3 4 4 4
oxidation potentials of the five-, six- and seven-mem-
bered rings. However, the first oxidation wave of four-
membered ring (I) is significantly lower compared with
all other perethylcyclopolysilanes studied, presumably
due to its ring strain. This kind of behavior is predicted
from UV–Vis absorption bands. In general, the oxida-
tion potentials are well correlated with electronic ab-
3.2. Controlled potential electrolysis
.2.1. Comparison of anodic oxidation of different
3
perethylcyclopolysilanes
The results obtained by the electrochemical oxidation
of a series of perethylcyclopolysilanes with different
ring sizes (I–IV), in CH Cl –CH CN (4:1 v/v) solution
2
1
2
3
−
−
4
containing 0.1 mol l
TBA+HSO , on platinum
sorption
energies.
However,
unlike
linear
foil, under N atmosphere, are summarized in Table 2.
2
permethylpolysilanes, for which the oxidation poten-
tials decreased and the lowest energy electronic absorp-
tion bands moved to longer wavelengths as the chain
length increases [2,6], cyclic polysilanes with small and
medium size rings showed an opposite trend for their
UV absorption bands: a bathochromic shift was ob-
served as the ring size decreases [2]. The electronic
spectra of cyclic polysilanes actually consist of a series
of overlapping absorptions. The low energy bands,
including the HOMO to LUMO transition, have small
absorptivity, while the higher energy transitions are
more highly permitted. The relative energies of the
HOMOs in cyclic polysilanes can be estimated from
oxidation potentials [10,11]. Rings from Si to Si have
Electrolyses were arbitrary terminated after electricity
consumption of about 1–1.2 F mol , when for all
compounds but one, no starting material was left. The
structural formula of emerged products 1–6 are de-
scribed in Scheme 1. The second column in Table 2
describes the product outcome from the electrochemical
−1
oxidation of (Et Si)4 (I). Apparently the reaction is
2
quite selective, affording two products (1 and 2), five-
and six-membered cyclic siloxanes, with four silicon
atoms and one and two oxygen atoms in their skeleton,
respectively. The major product 1 is obtained in 87%
yield. Both products stem from ring expansion due to
oxygen insertion. It is noteworthy that 2 has two iso-
mers (2a and 2b, Scheme 1), however, no attempt was
made to characterize the actual isomer(s) obtained. In
general, the 2a isomer is favorable because it was found
that the initially inserted oxygen atom in the ring
activates adjacent Si–Si bonds towards further inser-
tion of oxygen atoms [3].
7
5
approximately equal oxidation potentials, so the
bathochromic shift with decreasing ring size is mainly
due to stabilization of the LUMO. Going from Si to
5
Si the oxidation potentials decrease significantly, indi-
4
cating destabilization of the HOMO. This is consistent
with the increase in Si–Si bond length [12]. Quantita-
tively, it has been found that about half of the decrease
in the |–|* excitation energy going from Si to Si
The third column (Table 2) describes the results
obtained by (Et Si) (II). Surprisingly, after electricity
2
5
−1
consumption of 1.2 F mol
only 55% of the starting
5
4
material was consumed. The product mixture contained
four products (1–4), all cyclic siloxanes, five- to eight-
membered rings. The major product (38%) is 3 (which
involves five silicon atoms and one oxygen atom). All
rings could be accounted for in the increased energy of
HOMO, due to ring strain and lengthened Si–Si bonds
in the Si ring.
4

14
Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
Scheme 1.
products maintained the same number of silicon atoms
as in the parent compound except for 1 which lost one
silicon atom. However, cyclic siloxanes 1 and 4 were
minor products and formed in 3 and 4% yield, respec-
tively. Again, it is noteworthy that 4 could have two
isomers (4a and 4b, Scheme 1) but no effort was done
to characterize the actual isomer obtained, although we
prefer 4a for the reasons explained above.
Electrolyses of III and V afford similar spectra of
products (the fourth and fifth columns, Table 2), al-
though no starting material is left this time. Both cyclic
polysilanes undergo oxygen insertion and give cyclic
siloxanes 1–3 that were discussed previously, in addi-
tion to small amounts of a seven-membered product 5,
and an eight-membered 6 (obtained from III only). In
both cases, the major product is the six-membered 2,
involving two oxygen atoms in its skeleton.
(d) The major pathway for III and IV involves ring
chopping followed by two oxygen atoms insertion, at
least. It seems that the larger the ring size in the parent
cyclosilane, the higher the number of oxygen atoms
inserted in the products.
(
e) The highest selectivity in terms of the number of
products obtained was shown by I which yielded two
products only, while all other cyclic polysilanes af-
forded four to five products each. It seems that for the
cyclic polysilanes with the same substituent, the product
selectivity depends on the size of ring, with smaller
rings favoring higher product selectivity. Clearly, the
lower selectivity could stem from the Si–Si bond cleav-
age process. In fact, no cleavage product is formed by
the four-membered ring I. However, electrochemical
oxidation of the five-membered ring II gives about 24%
Based on the above described results obtained by
preparative electrolysis of perethylcyclopolysilanes I–
IV, some interesting observations could be deduced, as
follows:
(
with respect to the total products yield) of cleavage
products, whereas all products from the six- and seven-
membered rings (III and IV) are due to a ring cleavage
process.
(
a) Cyclic siloxanes are formed exclusively (no linear
products were detected).
b) The formation of five- (from I) and six-mem-
(
f) The fact that no starting material was left from I,
III and IV, and as much as 45% from II, may suggest
that the former three compounds have higher electro-
chemical reactivity than II, which is particularly stable
in this series. In fact, it is known that ring stability in
cyclic polysilanes depends on the size of the organic
groups which are appended [2]. For instance, the cyclic
pentamer is also the most stable ring in the perphenyl
(
bered (from II–IV) siloxanes, with one or two oxygen
atoms inserted, is preferred. These were the major
products.
(
c) For I and II, ring expansion by one oxygen atom
insertion is predominant. No products due to Si–Si
bond cleavage are observed.

Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
15
Table 3
Electrochemical oxidationa of II under N2
Productsc
Yield (%)b
A
MS (EI)
B
C
D
E
II
13
20
55
–
4
–
8
–
–
–
–
7
–
3
–
10
38
6
7
7
2
24
–
–
–
–
–
–
–
–
–
+
[
[
[
[
[
[
[
[
[
[
[
Et Si] O (2)
Et Si] O (3)
Et Si] O (4)
Et Si] O (5)
Et Si] O (6)
Et Si] O (7)
Et Si] O (8)
Et Si] O (9)
24
39
–
6
–
30
–
–
–
–
49
26
–
16
–
2
–
–
–
376(M ), 347(100%), 289, 87
2
4
2
+
446(M ), 417, 115, 87(100%)
2
5
+
478(M ), 449, 172(100%), 87
2
5
3
+
392(M ), 363(100%), 335, 87
2
4
3
+
408(M ), 379(100%), 351
2
4
4
+
462(M ), 433(100%), 291, 275
2
5
2
+
494(M ), 465(100%), 379, 87
2
5
4
+
–
510(M ), 481, 337, 115(100%), 87
2
5
5
+
Et Si] O (10)
Et Si] F (11)
Et Si] F (12)
13
42
45
306(M ), 277(100%), 221
2
3
3
+
–
–
–
–
296(M ), 277, 172, 87(100%)
2
3 2
+
–
–
210(M ), 191, 172, 87(100%)
2
2 2
a
Applied controlled potentials (ACP) are in V vs Ag quasi-reversible reference electrode; [II]=2–4 mM; electrolyte solution: CH Cl –CH CN
2
2
3
−
1
(
4:1/v:v) solution containing 0.1 mol l
electrolyte.
b
−
4
−1
Yields reported are relative and estimated by GLC. A electrolyte: TBA+HSO ; ACP=1.2; electricity consumption: 2.5 F mol ; working
−
−1
−
electrode: Pt. B electrolyte: TBA+HSO4 ; ACP=1.2; electricity consumption: 4.0 F mol ; working electrode: Pt. C electrolyte: TBA+HSO4
ACP=1.2; electricity consumption: 4.0 F mol ; working electrode: graphite. D electrolyte: TBA+ClO4 ; ACP=1.05; electricity consumption:
.5 F mol ; working electrode: Pt; the cyclic voltammogram exhibits two irreversible oxidation peaks at 1.25 and 1.6 V. E electrolyte:
TEA+BF4 ; ACP=1.05; electricity consumption: 2.5 F mol ; working electrode: Pt; the cyclic voltammogram exhibits two irreversible
;
−
1
−
−
1
2
−
−1
oxidation peaks at 1.25 and 1.6 V.
c
High resolution MS results: found (calc.) for 7 (C20H51O Si ): 463.2819 (463.2828); 8 (C20H51O Si ): 495.2708 (495.2726); 9 (C20H51O Si ):
2
5
4
5
5
5
5
11.2683 (511.2675); 10 (C12H31O Si ): 307.1641 (307.1637); 11 (C12H31F Si ): 297.1741 (297.1750); 12 (C H F Si ): 211.1185 (211.1180).
3 3 2 3 8 21 2 2
−
1
series [13]. Attempts to produce the equilibrium mixture
electricity consumption to 4 F mol , all starting mate-
rial is consumed. Furthermore, the yield of the formerly
major product (3) decreases, while the yield of a higher
oxygen contaminated product (7) increases. This result
is reminiscent of the outcome from the anodic oxida-
of (Et Si) rings by stirring I in THF in the presence of
2
n
K or Na/K, led to fast rates of redistribution. The
observed composition at equilibrium was 96% II, 2% I
and 2% III at 30°C [14], indicating that the five-mem-
bered ring is the most stable one for the ethyl substi-
tuted series. Apparently, decreasing the ring size in
cyclic polysilanes from five to four introduces ring
tension, and as a consequence, destablization of the
Si–Si bond [15].
An alternative explanation by which one of the prod-
ucts which is formed from II competes favorably with
the oxidation of the parent compound is ruled out on
the basis that in that case one would expect III and IV
to show the same behavior, because they all share
similar products. Instead, the latter two derivatives
afford products containing higher number of oxygen
atoms.
tion of the cyclic peralkylsilanes [(n-Pr) Si]5 and [t-
Bu(Me)Si] which showed similar behavior at different
electricity consumption [8,9]. It is noteworthy that the
ratio of Si /Si products for 4.0 F mol
2
4
−1
electricity
4
5
consumption is slightly higher (0.43) than that for 2.5 F
−1
mol
electricity consumption (0.38). This shows that
the Si–Si bond cleavage becomes a competitive process
with increased electricity consumption.
The effect of anode material on the reaction pathway
is described in column C, in which a graphite was used
instead of Pt anode (column B). It seems that on
graphite the conversion of II to products is incomplete
−
1
even after 4 F mol
of electricity were consumed,
since 7% of it remained unreacted at the end of electrol-
3
.2.2. The effects of electricity consumption and anode
material
Cyclic polysilane (Et Si) (II) was chosen as the rep-
ysis. Furthermore, the ratio of Si /Si5 products in
4
column C is reversed from 0.43 on Pt to 2.32 on
graphite. These results indicate that the anode material
has a marked effect on the products distribution. Ap-
parently, there is a large difference in the actual current
density on the two anode surfaces, which causes to a
different concentration gradient near the electrode sur-
face. As a consequence, the distribution of the products
is changed [16].
2
5
resentative example from this series on which various
effects were investigated. Columns A and B in Table 3
describe the product outcome from its electrochemical
oxidation, in the presence of TBA+HSO , with 2.5
−
4
−
1
and 4 F mol
of electricity consumption, respectively.
Evidently, the starting material was still left even after
consumption of 2.5 F mol . Upon increasing the
−
1

16
Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
Scheme 2.
3
.2.3. The effect of supporting electrolyte
The electrochemical oxidation of II was investigated
from the cyclic polysilane to generate a cadon radical,
which is an electrophile. Therefore, one has to look for
a nucleophile to react with the cation-radical. In spite
of the fact chat all electrochemical oxidation experi-
ments were carried out under nitrogen atmosphere, all
in the presence of three different electrolyte solutions,
leaving all other conditions the same. Let us compare
the results presented in columns A, D and E (Table 3).
Column D describes the results obtained with tetra-
butylammonium perchlorate. Evidently the reaction is
less selective compared with the results in column C,
since eight cyclic siloxanes (2, 4–10) were formed in-
stead of four. However, all of them are cyclic siloxanes.
Among the eight products there are two with large size
rings: nine-membered (8) and ten-membered (9). In
addition, the major products (5 and 10) are now the
ones containing three oxygen atoms rather than one (as
in 2).
Column E (Table 3) describes the results obtained in
the presence of the third electrolyte studied, tetrabuty-
lammonium tetrafluoroborate. Both oxygen insertion
and ring-opening processes take place to provide a
cyclic siloxane 10 (13%) and two linear difluoropolysi-
lanes, 11 and 12, as the major products (42 and 45%
yield, respectively). The fact that the ratio of (11+12)/
products (except for the ones obtained in the presence
−
4
of BF ) were cyclic siloxanes, containing one to three
oxygen atoms in their skeletons. As to the origin of the
oxygen atoms in the products, different reaction path-
ways could be suggested. It seems chat a plausible
mechanism could involve molecular oxygen (contami-
nated in the electrolyte solution or diffused from air)
insertion into the Si–Si bond. Scheme 2 describes the
formation of cyclic siloxanes via consecutive steps of
oxygen insertion into the Si–Si bonds of cyclic polysi-
lanes. The initially formed cation-radical A reacts with
molecular oxygen to generate intermediate B, which
may couple with a neutral cyclic silane molecule to
form C. The latter decomposes to a final product D and
+
’
its cation-radical D , which could also be generated
by direct electrochemical oxidation of the siloxane D.
Further oxygen insertion and Si–Si bond cleavage steps
+
’
1
0 is high (6.7/1) suggests that the ring-opening process
could take place via intermediate E
.
competes much more favorably with the oxygen inser-
tion process, in the presence of BF4 salt.
It is noteworthy that the suggested reaction between
molecular oxygen and the electrochemically generated
cation-radical species is in analogy to the chemical
reaction found between oxygen and radical-cations of
−
3.3. Mechanism
orefins [17,18], dienes [19] and acetylenes [20].
−
It has been suggested that the chemical oxidation of
When TBA+ClO
is used as the electrolyte
cyclic polysilanes with peracids proceeds by an elec-
trophilic oxidation process [3]. In the anodic oxidation
process, the initial step involves an electron uptake
(column D, Table 3), the spectrum of products was
significantly different from that obtained in the pres-
ence of TBA+HSO4 (column A). Not only were more
−

Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
17
Scheme 3.
types of cyclic siloxanes formed, but also more oxy-
gen-contaminated products were produced. Therefore,
one cannot rule out the possibility that the perchlo-
rate anion could also promote the oxygen insertion
reaction. A plausible mechanism describing the in-
volvement of the perchlorate anion is shown in
Scheme 3. The electrogenerated cation-radical A coor-
dinates with the perchlorate anion to afford the inter-
mediate F. The latter could be in an unfavorable
equilibrium with G and the chlorate anion (or with H
and the chlorate radical). As long as the chlorate
species is not reoxidized to the corresponding perchlo-
ring-opening process competes favorably with the oxy-
gen insertion process. Scheme 4 describes a mecha-
nism which involves the participation of the
fluoroborate anion. The initially formed unstable
cyclic cation-radical (A) is a highly reactive intermedi-
−
ate which may undergo an attack by BF4 to form
the radical (J), or alternatively, undergo ring opening
to generate a linear cation-radical (K) which could
then form the same entity, J. The latter intermediate
could undergo further oxidation to form the cation L,
−
which may then react with BF4 to yield the h,ꢀ-difl-
uorosilane M. This compound could undergo further
oxidative cleavage to give the products F–Si–(Si)n−
m–Si–F.
rate, the further oxidation of siloxanes is inhibited.
−
4
We have found that in the presence of BF , the
Scheme 4.

18
Z.-R. Zhang et al. / Journal of Organometallic Chemistry 574 (1999) 11–18
4. Conclusion
Acknowledgements
This work investigated the net effect of ring-size on
The authors are grateful to the Israel Academy of
Sciences and Humanities for financial support.
the electrochemical properties of a series of cyclic
perethylpolysilanes, by both cyclic voltammetry and
preparative scale electrolysis. In general, perethylcy-
clopolysilanes show one irreversible oxidation wave,
except for the four-membered derivative which exhibits
a second irreversible wave. The first oxidation peak
potentials depend upon the ring size of the cyclosilanes.
While there is a little difference among the oxidation
potentials of the five-, six-, and seven-membered rings
References
[1] R. West, E. Carberry, Science 189 (1974) 179, and literature cited
therein.
[
2] R. West, in: E. Abe, F.G.A. Stone, G. Wilkinson (Eds.), Com-
prehensive Organometallic Chemistry 11, vol. 3, Pergamon Press,
Oxford, 1995, p. 77.
(
II, III and IV), that of the four-membered ring (I) is
significantly lower compared with all other derivatives
[
3] G.A. Razuvaev, T.N. Brevnova, V.V. Semenov, J. Organomet.
Chem. 271 (1984) 261.
studied, due to ring tension.
[4] (a) I.S. Alnaimi, W.P. Weber, Organometallics 2(1983) 903. (b)
B.J. Helmer, R. West, Organometallics 1 (1982) 1463.
When a non-fluorinated electrolyte is used, the elec-
trochemical oxidation of perethylcyclopolysilanes leads
to the exclusive formation of cyclic siloxanes, due to
oxygen insertion. The formation of five- and six-mem-
bered cyclic siloxanes was favorable. No linear silox-
anes were observed. Generally, molecular oxygen is
responsible for the insertion reaction of oxygen into the
[
[
[
5] R.J. Klingler, J.K. Kochi, J. Am. Chem. Soc. 102 (1980) 4790.
6] W.G. Boberski, A.L. Allred, J. Organomet. Chem. 88 (1975) 65.
7] J.Y. Becker, E. Shakkour, R. West, Tetrahedron Lett. 33 (1992)
5
633.
[8] (a) J.Y. Becker, M.-Q. Shen, R. West, Electrochim. Acta 40
(1995) 2775. (b) J.Y. Becker, M.-Q. Shen, R. West, J. Elec-
troanal. Chem. 417 (1996) 77.
−
[9] Z.R. Zhang, J.Y. Becker, R. West, Electrochim. Acta 42 (1997)
Si–Si bond. However, when ClO4 salt is the elec-
1
985.
10] F. Shafiee, R. West, Silicon Germanium Tin Lead Compd. 9
1986) 1.
trolyte, this particular anion could be associated with
the oxygen insertion process, as well. In the presence of
[
(
−
BF4 anion the spectrum of products changes dramati-
[11] H. Watanabe, Y. Nagai, in: H. Sakurai (Ed.), Organosilicon and
Bioorganosilicon Chemistry, Ellis Horwood, Chichester, 1985, p.
cally; both ring-opening and oxygen insertion processes
take place, the former being predominant.
1
07.
[
12] H. Watanabe, M. Kato, T. Okawa, Y. Kougo, Y. Nagai, M.
Goto, Appl. Organomet. Chem. 1 (1987) 157.
13] H. Gilman, G.L. Schwebke, J. Am. Chem. Soc. 86 (1964) 2693.
14] C.W. Carlson, R. West, Organometallics 2 (1983) 1792.
15] M. Biernaum, R. West, J. Organomet. Chem. 131 (1977) 179.
[16] L. Eberson, K. Nyberg, Adv. Phys. Org. Chem. 12 (1976) 1.
17] S.F. Nelsen, R. Akaba, J. Am. Chem. Soc. 103 (1981) 2096.
18] E.L. Clennan, W. Simmons, C.W. Almgren, J. Am. Chem. Soc.
Among the cyclic polysilanes studied, the four-, six-
and seven-membered rings (I, III and IV) show a higher
electrochemical reactivity, compared with the five-mem-
bered ring (III). Also, the larger the size of the ring, the
higher the degree of Si–Si bond cleavage and the higher
the number of oxygen atoms in the products. The
product selectivity depends on both the size of the ring
[
[
[
[
[
1
03 (1981) 2098.
(
the smaller the ring the higher the product selectivity)
[
[
19] D.H.R. Barton, G. Leclerc, P.D. Magnus, J.D.I. Menzies, J.
Chem. Soc. Perkin Trans. I (1975) 2055.
20] R. Tang, J.H. Yue, F.J. Wolf, F. Mares, J. Am. Chem. Soc. 100
(1978) 5248.
and the amount of electricity consumption (for the
same ring-size, a higher selectivity is obtained at lower
−
1
F mol
values).
.
.