Free electron transfer mirrors rotational conformers of substituted aromatics: Reaction of benzyltrimethylsilanes with n-butyl chloride parent radical cations

Article abstract of DOI:10.1039/b402052a

The rotation motion of a larger substituent of an aromatic ring is accompanied by the electron density fluctuation of the highest occupied molecular orbitals. For benzyltrimethylsilanes p-R₃-C ₆H₄-○-CR₁-R_2<

Full text of DOI:10.1039/b402052a

View Article Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
Free electron transfer mirrors rotational conformers of substituted
aromatics: Reaction of benzyltrimethylsilanes with n-butyl chloride
parent radical cations
Ortwin Brede,*a Ralf Hermann,a Sergej Naumov,b Antonios K. Zarkadis,c
Gerasimos P. Perdikomatisc and Michael G. Siskosc
a
Interdisciplinary Group Time-Resolved Spectroscopy, University of Leipzig,
Permoserstrasse 15, D-04303 Leipzig, Germany. E-mail: brede@mpgag.uni-leipzig.de
Leibniz Institute of Surface Modification, Permoserstrasse 15, D-04303 Leipzig,
Germany
Department of Chemistry, University of Ioannina, 45110 Ioannina, Greece
b
c
Received11th February 2004, Accepted 2nd March 2004
F|rst published as an AdvanceArticle on the web 5th April 2004
The rotation motion of a larger substituent of an aromatic ring is accompanied by the electron density
fluctuation of the highest occupied molecular orbitals. For benzyltrimethylsilanes p-R3–C6H4– –CR1-
R2(SiMe3) [R3 ¼ H,PhCO; R1 , R2 ¼ H,H; H,Me; Me,Me; H,Ph; Ph,Ph] the silyl containing substituent rotates
on the axis between the aromatic moiety and the benzylic carbon as indicated in the formula. In the course
of this rotation a diversity of various conformers is passed. For the sake of simplicity, we reduce this diversity
to two (extreme) borderline structures, one with the substituent in plane with the aromatic ring (transient),
and one where the substituent is twisted by 90ꢀ (stable structure). In the very rapid and non-hindered electron
transfer from the conformer mixture to n-butyl chloride parent radical cations, the singlet ground state
conformers are converted into two kinds of radical cations. The cations tending to the planar type are stable
whereas those of the twisted type dissociate immediately into benzyl type radicals (p-R3–C6H4–CꢁR1R2)
and trimethylsilyl cations. The above mentioned product pattern and quantum-chemical calculations
on characteristic parameters of the ground state conformers and the product cations, show that the
electron transfer occurs completely unhindered after diffusional encounter, i.e. after each approach of the
reactants. Therefore the statistics of the rotation conformers mixture is reflected by the pattern and ratio
of the product transients such as metastable radical cations and radicals derived from a dissociative
radical cation. Up to now this is the only case where a diffusion-controlled electron transfer reaction
proceeding in the nanosecond scale is influenced by intramolecular rotation motions within the
electron donor.
cations as well as benzyl type radicals,5 both of them in similar
amounts. The electron transfer from benzyltrimethylsilane to
n-butyl chloride radical cations is represented by the following
equation.
Introduction
The photooxidation of alkyl and aryl substituted silanes has
been the subject of various time-resolved and steady state
experimental studies.1,2 The co-sensitized oxidation of benzyl-
trimethylsilanes3 is particularly interesting because the electron
transfer process occurs decoupled of the primary excitation
procedure. It can be analyzed as a proper ion–molecule
reaction between the relaxed aromatic tetracyanobenzene
radical cation (Cꢁþ) and the benzyltrimethylsilane, cf. reaction
(1).
n-C4H9Clꢁþ þ ArCH2 SiMe3 ! n-C4H9Cl þ ðMe3SiþÞ
ꢁ
ꢁ
þ
þ ArCH2 SiMe3e ðꢂ 40%Þ þ Ar CH2 ðꢂ 60%Þ
ð3Þ
A few tenths of nanoseconds after the rapid and diffusion-
controlled process (3), the part of silane radical cations was,
in a delayed manner, converted to radicals according to reac-
tion (2), which agrees well with the findings of Dockery
et al.3 Yet, the electron transfer mechanism (3) seemed to be
quite different from the co-sensitized one given in reaction (1).
A behavior similar to reaction (3) has been recently observed
during the electron transfer from phenols (ArOH), or other
chalkogenols, to alkane and alkyl chloride radical cations
(RXꢁþ).6–8 Comparable parts of phenol radical cations and
phenoxyl radicals were formed simultaneously, and this
unusual phenomenon could be explained using two basic
arguments:
Cꢁþ þ ArCH2 SiR3 ! C þ ArCH2 SiR3e
ð1Þ
ꢁ
þ
It has been shown that in the non-polar surrounding the
decay of the benzyltrimethylsilane radical cation proceeds in
the nanosecond time scale. The C–Si bond cleavage occurs
due to a SN2 reaction with a nucleophile, or under distinct
circumstances, in a monomolecular manner.3,4
ꢁ
ꢁ
ArCH2 SiR3e ðþNuÞ ! Ar CH2 þ R3Siþ ð2Þ
þ
However, in the electron transfer from arylmethyl-trimethyl-
silanes to saturated parent radical cations derived from alkanes
or alkyl chlorides, we found a primary simultaneous formation
of two products such as the corresponding silane radical
(i) the rotation of the phenol group around the C–O axis
generates a mixture of extremely short-lived conformers with
very different electron distributions of the highest molecular
orbitals and
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5
2267

View Article Online
(ii) a rapid non-adiabatic electron transfer takes place in
each diffusive encounter between the reaction partners. That
is why the rotation motion of the C–OH group is reflected
by the formation of a stable phenol radical cation and a disso-
ciative phenol radical cation, see reactions (4a,b), respectively.
These products carry excess energy (*) which is lost in a few of
picoseconds.
benzylsilanes used:
p-R3 C6H4 CR1R2 SiMe3
No.
R1
R2
R3
1a
1b
1c
1d
1e
H
H
H
H
H
H
H
Â
Ã
Me
Me
Ph
Ph
H
RXꢁþ þ ArOHplane ! ArOHꢁþꢃ ! ArOHꢁþ
Me
H
"#
Â
Ãꢃ
RXꢁþ þ ArOHperp ! Ar ꢄ Oꢁþ ꢄ H ! ArOꢁ þ Hþ
Ph
ð4a; bÞ
2a
2b
2c
2d
2e
H
H
PhCO
PhCO
PhCO
PhCO
PhCO
Me
Me
Ph
Ph
H
Me
H
The electron transfer in non-polar media involving parent
solvent radical cations, derived from alkanes or alkyl chlor-
ides, exhibits some peculiarities which are described in detail
in refs. 9 and 11. Because of the high excess energy (0.5 to 2
eV) and the unusual uniform charge distribution over the
whole sigma bonded skeleton of the parent cations,9 this
non-adiabatic electron transfer is very rapid. Therefore, it is
called free electron transfer (FET) and is characterized by a
non-hindered rapid electron jump.
Ph
The time-resolved optical spectroscopy gave time profiles
and optical transient spectra with an overall time resolution
of the pulse radiolysis set up of < 10 ns, which is convenient
for analyzing the bimolecular diffusion-controlled electron
transfer reactions of type (6).
Electron pulse radiolysis is the best method for the genera-
tion and kinetic study of the parent solvent radical cations.
This method has also been used to investigate the kinetic
problem illustrated in the aforementioned reactions (4a,b).
The aim of this paper is to analyze the unusual primary pro-
duct behavior of the free electron transfer (FET) reaction (3)
that involves bezyltrimethylsilanes with a varying substitution
pattern at the benzylic carbon and in a para-position of the
aromatic ring. In order to identify and quantify the product
transients, pulse radiolysis experiments were conducted. More-
over, quantum chemical calculations were used to describe the
dynamics of the substituent rotation.
Kinetics of the free electron transfer (FET)
p-Benzoyl substituted benzylsilanes 2. In the pulse radiolysis
of pure n-butyl chloride, the optical transient absorption spec-
trum exhibits a broad absorption band betweenꢁ350 and 650
þ
nm (lmax ¼ 500 nm) of the parent ion n-C4H9Cl with a life
time t ¼ 140 ns.13,14 Adding to the sample millimolar concen-
trations of the silanes 1 and 2 shortens t due to a pseudo-first
order reaction. In the course of this reaction (6), product tran-
sient spectra appear (see Fig. 1) for the representative example
of a 2 ꢈ 10ꢄ3 mol dmꢄ3 solution of 2e.
The spectrum of the nitrogen purged solution (Fig. 1a) taken
after the 12 ns electron pulse in the maximum of the primary
effect (80 ns), consists of three bands of a short-lived transient
(t ¼ 2.1 ms, lmax ¼ 830, 510 and 380 nm). They are superim-
posed on a long lasting absorption in the near UV (t ¼ 29
ms, lmax ¼ 350 nm). The time profiles given on the top of
Fig. 1 show the decay behavior of the two species which, in
spite of the spectral superpositions, can be well distinguished.
In the oxygen saturated solution, the short-lived transient is
marginally influenced, whereas the long-lasting one seems to
be converted into another long-lived species (see left part of
Fig. 1b). In light of these observations and spectral analogies
to literature data concerning the benzyltrimethylsilane radical
cation,15 the short-lived transient was assigned to the radical
Results and discussion
Experimental approach
Because of the high ionization potentials of the parent solvent
molecules (IPgas ꢅ 10 eV),10 alkane and alkylchloride radical
cations are very good electron acceptors. These cationic species
represent complete s-bond structured species where the charge
is equally distributed over the whole skeleton. It results in
their metastability (t ꢆ 150 ns) in liquid phase at room
temperature.11,12a
Due to the unfavorable optical properties of alkenes and
alkyl chlorides (light absorption in the far UV), electron pulse
radiolysis has to be used for generating the parent ions and the
study of the electron transfer involving appropriate donor
molecules, see reactions (5) and (6). n-Butyl chloride is a useful
solvent and acts as an internal electron scavenger. Hence
the pulse radiolysis of n-butyl chloride generates mainly the
following transients, cf. reaction (5):13,14
ꢁ
þ
cation 2e whereas the more stable species was interpreted
asꢁ a p-benzoyltriphenylmethyl radical16,17 2eꢁ derived from
þ
2e by dissociation (2).
The kinetics of the electron transfer (6) was analyzed in spe-
cified points of the spectrum, i.e. around the transient maxima.
This is shown in Fig. 2 for the short time range, i.e. for the
ꢁ
þ
nanosecond scale. The decay of n-C4H9Cl (620 nm) in tens
of nanoseconds is described by k6 ¼ 5.5 ꢈ 109 dm3 molꢄ1
n-C4H9Cl L n-C4H9Clꢁþ þ Clꢄ þ n-C4H9
ð5Þ
ꢁ
s
ꢄ1, and it corresponds well both with the formation of the
ꢁ
þ
silane radical cation 2e (840, 520 nm) and the rapid part of
the fragment radical formation (350 nm). Judging the kinetics
ofꢁthe latter it should be taken into account that also the cation
n-C4H9Clꢁþ þ D ! Dꢁþ þ n-C4H9Cl; D ¼ electron donor
ð6Þ
þ
2e decays by a delayed bond cleavage analogous to (2) and
Under pulse radiolysis conditions in the nanosecond time
scale reactions other than the diffusion-controlled electron
transfer (6) can be excluded because of the much lower re-
activity of the radiolytically formed radicals. Therefore we
forming the triphenylmethyl type radical 2eꢁ. Hence there are
two channels of radical formation, a rapid and direct one via
electron transfer, and a delayed one via dissociation of the
observed metastable cation 2eꢁþ, taking place with k8 ¼ 4.8 ꢈ
105 sꢄ1. This mechanism is described by the reactions (7a,b)
and (8). These reactions can be better distinguished by
pulsed very diluted solutions of D (c ꢇ 10ꢄ3 mol dmꢄ3
)
D stands for the different
in n-butyl chloride where
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
2268
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5

View Article Online
the electron transfer part (7b) is visualized. This complex
kinetic situation is presented in the following reaction sequence
(7)–(10).
ꢁ
þ
Both the silane radical cation 2e and the fragment radical
are sensitive to oxygen with k9 around 5 ꢈ 107 dm3 molꢄ1 sꢄ1
and k10 in the range of 109 dm3 molꢄ1 ꢄ1, where the latter is
s
difficult to calculate because of spectral superpositions with the
produced peroxyl radical. A similar value of k10 ¼ 3 ꢈ 108
Mꢄ1 ꢄ1, typical for trityl radicals, has been reported for
s
the reaction of the p-benzoyltriphenylmethyl radical with
oxygen.17
A similar reaction behaviour was also found for the other
benzoyl substituted silanes of type 2. Fig. 3 demonstrates spec-
tra and decay kinetics of the silane radical cations and frag-
ment radicals derived from the methyl substituted compound
ꢁ
þ
2b. Here the silane radical cation 2b with bands at 640 and
400 nm as well as the radical signal at 340 nm can be easily dis-
tinguished. It can be seen that the rapidly formed radical part
(analogous to reaction (7b)) is quenched by oxygen already in
an early state. The remaining absorption around 340 nm is
caused by the resulting peroxyl radical. The decay time profiles
taken at 400 and ꢁ340 nm (Fig. 3) demonstrate the relatively
clean decay of 2b þ, whereas the radical time profile follows
a more complex kinetics—rapid (7b) and delayed (8) forma-
tion of the radical, and slightly superimposed with the cation
absorption tail. The kinetic data for 2b and the other
compounds of type 2 are summarized in Table 1.
Fig. 1 Transient absorption spectra taken in the pulse radiolysis of a
solution of 2 ꢈ 10ꢄ3 mol dmꢄ3 2e in n-butyl chloride in the nitrogen
purged sample (upper spectra, a) and in oxygen saturated solution
(lower spectra, b), times after the pulse: (X) 80 ns, (c) 18 ms (for a,
N2) or 3.5 ms (for b, O2). The time profiles on the top of the figure were
taken in N2 bubbledꢁsolution: 380 nm mainly for a radical 2eꢁ, 520 nm
þ
representative for 2e
.
comparing the time profiles of the radical absorption at 350
nm in N2 purged and in O2 saturated solution, see Fig. 2. In
the first case the rapid (7b) and the delayed (7a), (8) radical for-
mations are well observable. In the second one, oxygen
quenches mainly the delayed formation (8) and, therefore,
Fig. 3 Transient absorption spectra taken in the pulse radiolysis of a
solution of 1.2 ꢈ 10ꢄ3 mol dmꢄ3 2b in n-butyl chloride. N2 purged
solution: 100 ns (X), 3 ms (c) and 18 ms (L) after the pulse; O2
saturated solution: after 100 ns (˘). The time profilꢁes demonstrate the
ꢁ
þ
Fig. 2 Time profiles of the n-C4H9Cl (620 nm), tꢁhe silane radical
cation 2eꢁþ (840, 520 nm) and the fragment radical 2e (350 nm) taken
in the pulse radiolysis of a solution of 2 ꢈ 10ꢄ3 mol dmꢄ3 2e in n-butyl
chloride purged with nitrogen (at 350 nm also treated with O2).
ꢁ
þ
typical transient decay behavior: 340 nm mainly 2b , 400 nm for 2b
.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5
2269

View Article Online
Table 1 Kinetic and spectral parameters of the product formation by
free electron transfer from benzyltrimethylsilanes to n-butyl chloride
radical cations
tioned, in the presence of oxygen a marked part of the benzyl
absorption ꢆ320 nm is quenched from the very beginning. It
gives arguments for its rapid and direct formation by the elec-
tron transfer, analogous to the reaction path (7b). The time
profiles shown in Fig. 5 illustrate the FET kiꢁnetics in more
Silane
radical
cations
Benzyl
type
þ
detail. At 640 nm only the decay of n-C4H9Cl is observed,
k7a,b of
FET/
dm3
radicals
ꢁ
þ
k10 , O2
rct./dm3
Product
ratio
at 530 nm a superposition of the n-C4H9Cl and the silane
ꢁ
ꢁ
þ
þ
lmax/
tdecay
ns
/
lmax/
tdecay/
ms
cation 1a appears, and at 310 nm the kinetics of 1a could
be seen clearly. The radical signal at 260 nm shows the rapid
radical formation by (7b) and the delayed one via the cation
molꢄ1 sꢄ1 nm
nm
molꢄ1 sꢄ1 cat./rad.
1a 5 ꢈ 109
530
310
550
315
290
320, 260 12
320, 260 > 30
320, 270 30
109
109
40/60
30/70
ꢁ
þ
1a by reactions (7a) and (8). All silane species were found
to be formed with a rate constant k7a,b ¼ 5 ꢈ 109 dm3 molꢄ1
1b 5 ꢈ 109
140
s
ꢄ1, i.e. with a diffusion-controlled rate.
Analyzing the quenching of the transients by oxygen it
should be mentioned that because of its short lifetime the radi-
1c diff.-contr. 450
330
1d 4 ꢈ 109
<100
100
7 ꢈ 108
30/70
þ
ꢁ
450
330
330
40
cal cation 1a does not show any effect. The benzyl radical is
quenched with k10 of about 109 dm3 molꢄ1 sꢄ1 where, because
of spectral superpositions, the exact value is difficult to
determine.
1e diff.-contr. 480–520 <70
2a 5 ꢈ 109
340
330
40
20
7 ꢈ 108
25/75
40/60
620
380
640
400
670
410
760
525
380
830
510
380
1500
2800
1400
2800
5 ꢈ 108
2b
2c
2d
340
340
345
20
28
38
The other silanes of type 1 show an analogous kinetic beha-
vior. In the case of higher substitution at the benzylic carbon
atom it was observed that the lifetime of the silane radical
cations becomes shorter. For the triphenyl substitution pattern
1e, in the red range a signal peaking different from that of n-
109
109
ꢁ
ꢁ
þ
C4H9Cl was found and it is assigned to 1e þ. However,
because of a lifetime ꢆ50 ns, it could not be kinetically distin-
guished from the solvent parent ion. The spectral and kinetic
data of the transients produced by FET ionization of the
benzyltrimethylsilanes 1 and 2 are summarized in Table 1.
2e 6 ꢈ 109
2100a
350
29
25/75
a
k9 with oxygen ¼ 5 ꢈ 107 dm3 molꢄ1 sꢄ1
.
Interpretation of the FET kinetics and its product distribution.
In the free electron transfer from ꢁtþhe benzyltrimethylsilanes to
the solvent parent ions n-C4H9Cl two kinds of products are
directly observed and formed with the same rate, namely the
benzyltrimethylsilane radical cations and the benzyl type radi-
cals. Subsequently, the metastable silane radical cations are
transformed into benzyl type radicals in a delayed process.
The transformation can be well seen in the radical time profile
(cf. Fig. 2, 350 nm profile; Fig. 5, 260 nm profile). In the cases
when the spectral transient superposition is not too strong, the
ratio between the two reaction channels (7a) and (7b) can be
estimated from the fast and delayed parts in the radical time
profile. In the case of 1a, about 40%silane radical cations (path
(7a)) and 60% radicals (path (7b)) were formed by FET. The
product distribution of the FET varies predominantly between
20 and 40% silane radical cations and hence the benzyl type
radicals dominate with 80 to 60% (see Table 2). Because of
FET producing less stable benzyltrimethylsilane radical
cations. The benzyltrimethylsilanes of type 1 were found to
generate less stable radical cations than the p-benzoyl substi-
tuted ones of type 2. Therefore their reaction is discussed after
the principal studies given in the preceding paragraph. Fig. 4
presents transient absorption spectra obtained in the pulse
radiolysis of a 1.8 ꢈ 10ꢄ3 mol dmꢄ3 solution of the silane 1a
in n-butyl chloride.
Partially in accordance with the literature3 and also because
of the kinetic behavior, the two absorption maxima of the
short-lived transient (t ¼ 290 ns, lmax ¼ 530 and 310 nm)
ꢁ
þ
were assigned to the metastable silane radical cation 1a
.
The longer-lasting absorption bands (t ¼ 12 ms, lmax ¼ 320
anꢁd 260 nm) are interpreted to be caused by the benzyl radical
1a , which is well known from the literature.18 As already men-
ꢁ
þ
Fig. 4 Transient absorption spectra taken in the pulse radiolysis of a
solution of 1.1 ꢈ 10ꢄ3 mol dmꢄ3 1a in n-butyl chloride. N2 purged
solution: 100 ns (X) and 2 ms (L) after the pulse; O2 saturated solution:
after 100 ns (c).
Fig. 5 Time profiles of the n-C4H9Cl (640 nm), ꢁthe silane radical
ꢁ
þ
cation 1a (530, 310 nm) and the benzyl radical 1a (partially at 310
nm, main part 260 nm) taken in the pulse radiolysis of a solution of
2 ꢈ 10ꢄ3 mol dmꢄ3 1a in n-butyl chloride purged with nitrogen.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
2270
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5

View Article Online
Table 2 Calculated (B3LYP/6-31G(d)) parametersa
˚
C–Si/A
Singlet
˚
C–Si/A
Radical cation
R1
H
R2
H
R3
H
H
H
H
Structure
t
Rot/ps
Ea/kcal molꢄ1
3.8(15.7)b
5.7(19.4)b
3.9(20.8)b
2.3
D(C–Si)/A
S(Si)
0.096
˚
1a
1b
1c
1d
Stable
Tr. St.
Stable
Tr. St.
Stable
Tr. St.
Stable
Tr. St.
Stable
Stable
Tr. St.
Stable
Tr. St.
Stable
Tr. St.
Stable
Tr. St.
Stable
Tr. St.
1.33(0.43)b
0.9
1.919
1.910
1.929
1.934
1.946
1.946
1.939
1.948
1.997
1.922
1.913
1.933
1.930
1.950
1.948
1.945
1.953
1.996
1.999
2.082
1.942
2.138
1.963
2.211
1.974
2.164
0.163
0.032
0.209
0.029
0.265
0.028
0.225
ꢂ0.00
0.114
ꢂ0.00
0.134
ꢂ0.00
0.118
H
Me
Me
Ph
Me
H
1.8(0.46)b
0.9
1e
2a
Ph
H
Ph
H
H
ArCO
2.0
1.45(0.84)b
2.365
2.025
1.935
2.080
1.960
2.140
1.967
2.135
0.368
0.103
0.022
0.147
0.030
0.190
0.019
0.190
0.149
0.064
ꢂ0.00
0.085
0.005
0.104
ꢂ0.00
0.105
4.4(11.5)b
5.6(13.5)b
3.5(13.9)b
2.1
2b
2c
2d
2e
H
Me
Me
Ph
ArCO
ArCO
ArCO
ArCO
2.04
Me
H
2.19(1.59)b
2.04
Ph
Ph
2.38
2.5
2.308
0.312
0.135
a
trot and activation energies (Ea) of the trimethylsilyl-group rotation, bond length C–Si for the ground state and the radical cation and bond length
change D(C–Si) in dependence on molecular geometry. S(Si)—atomic spin density on Si atom of radical cation. Tr. St.—transition state structure.
Radical cation value in parentheses.
b
the aforementioned spectral superpositions these ratios repre-
sent very rough measures. No such estimates could be made
in the cases where the UV main radical band (around 260
nm) is covered by the silane own absorption.
—The free electron transfer takes place during the first
approach of the reactants (as a non-adiabatic and longer-range
interaction). The electron transfer of two characteristic kinds
of products indicates the diversity of the short-living confor-
mers that are being continuously formed by the rotation
motion.
Also in the case of the benzyltrimethylsilanes studied here,
the rotation of the substituent around the axis connecting the
aromatic moiety and the benzylic carbon (see also Scheme 1).
is accompanied by electron fluctuation at the HOMO level.
As a consequence, metastable silane radical cations (7a) with
In order to understand the reason for this unusual reaction
behaviour, a former result that takes into account the peculia-
rities of the free electron transfer has to be highlighted. As far
as the parent radical cations are concerned, the even distribu-
tion of the charge over the whole s-bond skeleton of alka-
nes8,12b,c and alkyl chlorides8,12a as well as the relatively
large ionzation potential differences between donor and accep-
tor (DIP ꢅ 0.5 eV) impose a non-hindered electron transfer
step. The non-hindered electron transfer step occurs practically
in each encounter between the reactants. Therefore the elec-
tron transfer works in each encounter geometry of the molecu-
lar electron clouds, i.e. also without a defined encounter state
in the usual sense. Such a picture is typical for a non-adiabatic
electron transfer which can happen also via longer distances. If
each approach of the reactants succeeds in electron transfer,
then the rotation motions within the electron donor molecule
can play a role. This is true in particular for the rotation
motions of substituents at aromatic rings. As already men-
tioned in the Introduction, when the FET involves phenols,
the rotation of the –OH group around the C–O axis is reflected
by two different kinds of products. These products originate
from a metastable phenol radical cation and dissociative
phenol radical cation, respectively, cf. reaction (4a,b).
To explain the molecule dynamic effect reported in this
paper the following points are to be considered:
—The donor substances (benzyltrimethylsilane type com-
pounds) exhibit a rotation motion around the bond between
the aromatic ring and the substituent as indicated by the arrow
p–R3–C6H4––CR1R2(SiMe3) with a frequency around 1012 Hz.
Statistically seen, during the rotation motion all imaginable
twist angles between 90ꢀ (stable state) and 0ꢀ (transition state)
are passed. The manifold of these twisted extremely short-
living structures is difficult to describe by a simple picture.
Therefore we used only the extreme cases (90ꢀ, 0ꢀ) and called
them border line structures. It means that the diversity of all
other twisted structures is contained either in the 90ꢀ confor-
mer or in that of 0ꢀ angle and it is determined by the electron
distribution in the ground state.
Scheme 1
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5
2271

View Article Online
charge at the aromatic ring as well as those with their charge
partially localized also at the benzyl carbon (7b) are formed,
where the latter dissociates immediately. On the other hand
the partially hindered rotation of the benzylic bond seems to
cause the delayed dissociation (8), see the preceeding chapter.
Quantum-chemical calculations
Quantum-chemical calculations with 10 molecules bearing dif-
ferent substituents in the positions R1, R2 and R3 were per-
formed in order to support the proposed mechanism5 of the
generation of two different radical cations in dependence on
the molecular structure. All of the studied molecules behave
similarly. The equilibrium geometries of the singlet ground
state and of the radical cation have a C–Si bond nearly perpen-
dicular to the ring plane (dihedral angle C2–C1–C7–Si nearly
90ꢀ). The frequency analysis of the singlet ground state shows
the existence of a relatively slow (from 0.4 ꢈ 1012 to 1.1 ꢈ 1012
Hz depending on the substituents) rotation motion around the
axis connecting the benzylic and the ipso-aromatic carbon
atom. This leads to a ‘‘planar structure’’ (C–Si bond coplanar
to the ring plane). The frequency analysis of the ‘‘planar struc-
tures’’ produces only one imaginary frequency indicating that
this conformation represents the transition state for rotation.
The activation energies (barrier heights of rotation) Ea calcu-
lated for the molecules in the ground state are between 2.3
and 5.7 kcal molꢄ1 and, hence, at room temperature the inter-
nal rotation of the trimethylsilyl-group should proceed. How-
ever, if the radical cation is generated, the activation energy for
rotation is found to be essentially higher, i.e., the rotation
motion should be strongly hindered.
Fig. 6 Transformation of HOMO of the 1a singlet ground state
(isocontur ¼ 0.08) in dependence on C–Si(Me)3-group rotation and
spin density distribution of the radical cation (isospin ¼ 0.01).
symmetry as the p system of the aromatic ring. Due to the
strong resonance with the p electron of the aromatic ring,
the electron from HOMO is partly shifted from the aromatic
moiety to the C–Si bond. By a rotation around the C–Si bond
the coupling between ring and sp3 hybrid from C–Si bond will
be disturbed.
Therefore the HOMO in the planar structure is entirely loca-
lized on the aromatic ring. It should be mentioned that accord-
ing to the Hammet rules the –CH2SiMe3 group is an electron
donating substituent in electrophilic aromatic substitutions20
with a sp constant of ꢄ0.21.
This is also expected considering imaginary resonance forms
and the extended stabilisation provided by the b-silicon
effect,19 see Scheme 2. The calculated data for the stable
(perpendicular) and transition (planar) structures are given
in Table 2.
The calculated valence C–Si vibrations have a frequency of
about 2.7 ꢈ 1013 Hz corresponding with a time for one vibra-
tion of about 37 fs and being at least 30 times faster than
the rotation of the –CR1R2–SiMe3 group. Yet, both of them
are slow in comparison to the electron transfer jump. Accord-
ing to the Born–Oppenheimer approximation and due to the
relatively slow internal rotation of the substituent –CR1R2–
SiMe3 , the molecular geometry is stiff relative to the electron
relaxation process and the very fast electron transfer event.
In order to test the influence of the molecular geometry on
the electronic structure, we investigated the behaviour of the
highest double occupied MO (HOMO) of the singlet ground
state. This state is involved in the electron transfer process,
in dependence on –CR1R2–Si(Me)3 group rotation (torsion
angle j).
Thus, due to the relatively slow rotation motion, after elec-
tron transfer from the HOMO of Ar–CH2–Si(Me)3 to the sol-
vent radical cation two different radical cation types can be
generated. In the ‘‘perpendicular’’ radical cation the unpaired
electron is partly localized on the C–Si bond. It leads to a
strong weakening of this bond in the radical cation state. In
the planar radical cation, there is no unpaired electron on
the C–Si bond and, consequently, the bond length should
not be influenced coming from the ground to the radical cation
state.
With respect to these results and for a sufficiently short time
scale where the rotation of the critical bond is ‘‘frozen,’’ we
can assume two limiting conformer structures strongly differ-
ing in the tendency to a prompt dissociation, in correspon-
dence with eqns. (7a,b).
The changes of electron distribution from HOMO show a
similar trend (as shown in case of 1a in Fig. 6) in all the studied
molecules. As can be seen from Fig. 6, the s C–Si bond is
formed through an sp3 hybrid from carbon and Si atoms. In
the stable perpendicular structure this s-bond has the same
This different behaviour of the HOMO electrons is strongly
supported by calculations of C–Si bond length changes (D(C–
Si)) between the ground and the radical cation state, in depen-
dence on molecular geometry. While in the planar structure the
C–Si bond is practically unchanged, in the case of the perpen-
˚
dicular structure it becomes longer (up to 0.368 A for 1e), see
Table 2. Thus, the radical cations in the perpendicular struc-
ture seem to be less stable than those in the planar one and,
therefore, they are favoured for a rapid desilylation. Addition-
ally, there is a strong dependence of the D(C–Si) on the type of
the substituents R1, R2 and R3 (Table 2), which correlates
clearly with the atomic spin density on Si calculated for the
perpendicular radical cation (see Fig. 7a). The more the
unpaired electron resides on Si, the larger is the change of
D(C–Si) and, consequently, the weaker is the C–Si bond and
the shorter is the life time of the radical cations.
Scheme 2
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
2272
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5

View Article Online
Fig. 7 (a) Calculated for radical cations: spin densities on Si-atom and C–Si bond length changes (D(S–Si)) between ground state and radical
cation state in dependence on substituent R1, R2 and R3. (b) Correlations between calculated spin densities on Si-atom and the experimental decay
rate constant (1/log tdecay , originating from Table 1) of radical cations in dependence on substituent R1, R2 and R3. Correlations between the spin
densities on the Si-atom and (c) the bond dissociation energy of the C–Si bond of the radical cations 1ꢁþ and 2ꢁþ, or (d) change of the C–Si bond in
these radical cations.
The compatibility of the data can be also seen from the
correlations given in Fig. 7a–ꢁd where the spin densities of
R3 ¼ PhCO is also indicated by the observation that benzylic
substitution within this series does not affect Ea or log(1/t)
significantly.
ꢁ
þ
þ
the different cations 1 and 2 are plotted against the life-
time, cation dissociation energies and bond lengths changes
of the C–Si bond.
Electron transfer involving r- or p-type radical cations
Preliminary concluding the results of the quantum-chemical
analysis, the two types of benzyltrimethylsilane radical cations
should be qualified as
—stable ones (non-dissociative) with planar structure and
charge positioned at the aromatic ring, decaying by statistical
transformation into the perpendicular state, and
—dissociative ones with perpendicular structure and charge
distribution over the whole molecule, decaying during the
C–Si bond vibration.
As far as perpendicular radical cation is concerned, the
above conclusions are in line with the often used tentative
and qualitative resonance structures given by the formulae
(I) to (IV). It is obvious that the structure (IV) of the p-benzoyl
derivatives is responsible for the diminished spin density found
on Si (Fig. 7). The importance of (IV) is also demonstrated in
EPR spectra of the p-benzoyltriphenylmethyl radical21 where
the p-benzoyl group causes an approximately 18% shift of
the spin density to the p-benzoyl substituted ring. On the other
hand within each series (R3 ¼ H or PhCO), an increasing sub-
stitution on the benzylic carbon causes an increase in the spin
density at Si due to increased contribution of (III). Therefore it
is resonable to assume that (III) contributes more to R3 ¼ H
than to R3 ¼ PhCO. It explains (i) the larger bond length
changes D(C–Si), (ii) the lower activation energies of the Ar–
CSi bond rotation and (iii) the corresponding higher decay rate
constants (Tables 1, 2 and Fig. 7b) for R3 ¼ H than for
R3 ¼ PhCO. The diminished importance of (III) for
Comparing the results of the (cosensitized) electron transfer3
(reaction (1)) with those of the FET described in this paper,
the reason for the different product distribution comes into
question. In the case of cosensitization when a p system stabi-
lized acceptor radical cation was used, exclusively benzylsilane
radical cations were formed, cf. reaction (1). The process fol-
lows the usually assumed reaction path proceeding through
the formation of an encounter complex. This results from
encounter of an aromatic p-stabilized radical cation with the
donor molecule. The encounter leads to the formation of an
energetically favored complex which is optimized in the course
of the donor–acceptor interaction. Next a controlled electron
transfer (1) occurs. The transfer can be imagined as a con-
certed electron transition with bond length change in the par-
ticipating molecules, which results in one dominating product,
here the donor radical cation only. A situation has been
reported for the case of the electron transfer from phenols to
benzene radical cations. There a controlled sandwich-like
encounter complex has been described by a quantum chemical
approach.22
ꢁ
ꢁ
C6H6 þ þ Ar OH , ½C6H6 þ ꢉ ꢉ ꢉ Ar OHꢊ ,
ꢁ
ꢁ
½C6H6 ꢉ ꢉ ꢉ Ar OH þꢊ , C6H6 þ ArOH
ð13Þ
þ
On the other hand, s-type parent radical cations derived
from alkanes or alkyl chlorides are metastable species
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5
2273

View Article Online
stabilized only by the equilibration of the charge over the
whole molecule. In the light of the above mentioned results
and taking into account the very high free energy of (up to 2
eV) of the free electron transfer in non-polar systems, it can
be said that every single approach regardless of its geometry
succeeds in an electron transfer, apart from an optimized
encounter complex. Therefore only the direct forward reaction
path of reaction (6) is to be considered. If the donor molecule
contains a defined rotating group, then this motion is accom-
panied by electron fluctuations of the highly occupied orbitals
along the molecule. Consequently, the molecule conformation
is permanently changed with a relatively high frequency and
on the average it results in a mixture of the diverse conformers.
A completely unhindered bimolecular electron transfer reac-
tion may identify the different electron distributions and in dis-
tinct cases can result in different products. It happens during
the free electron transfer from phenols (and other chalcogen-
ols) and from benzyltrimethylsilanes to parent ions. This spe-
cial transfer can be detected when different product ions
show different stability, i.e. if one of them is stable and the
other one dissociates immediately. It is specially true for mole-
cules where good leaving groups exist such as protons (phenol
radical cations) or trimethylsilyl cations (here for the silanes 1
and 2). So the exotic phenomenon is given where a diffusion-
controlled reaction reflects intramolecular rotation motions
within the donor molecule.
experimental conditions another possibility, alternative to the
interpretation through rotation frequencies, exists. In the radi-
olysis of n-butyl chloride a constant amount of the nucleophile
Clꢄ is generated in the non-polar solvent. So a bimolecular
decay reaction of the radical cation derived from the silanes
type 1 and Nu ¼ Clꢄ could also fit the data. It can be justified
by the correlation between spin density at Si and the decay rate
of the cations 1ꢁþ (cf. Table 2), and is further supported by the
evidence from other arylmethyl silane systems.23
Concluding remarks
It is well known that substantial molecule oscillations, e.g.,
rotational motions of functional groups at aromatic moieties,
are accompanied by the electron fluctuations of the upper
occupied orbitals. Consequently, different short-lived confor-
mer states of the molecule come into being and are periodically
passed with the rotation frequency. This scenario was set in the
time range of hundreds of femtoseconds (chalcogenols8) up to
a few picoseconds (benzylsilanes, this paper) and has usually
no chemical consequences in bimolecular reactions. In this
paper the rapid and non-hindered free electron transfer from
benzyltrimethylsilane type molecules to parent radical cations
of n-butyl chloride is demonstrated. It reflects the intramolecu-
lar group rotation and its two characteristic product lines of
a bimolecular reaction that, in the end, yield two different con-
former silane radical cations. The rotation of the substituents
in the benzyltrimethylsilanes was simplified by defining two
ground state conformer structures—a perpendicular stable
ground state and a planar transient state which have consider-
ably different electron distribution of the HOMO. Through
free electron transfer, the aforementioned conformer cate-
gories are transformed into the corresponding radical cations
which occur without nuclear rearrangement. Of the resulting
two types of radical cations, the perpendicular one has
dissociative character, whereas the planar one is stable.
The results reported here are in line with the observations
in the bimolecular phenol ionization by FET.8 Therefore the
described phenomenon is a more general effect and it is valid
for systems with good leaving groups (protons or trimethylsilyl
cations) that enable the identification of the different reaction
channels, see reactions (4) or (7). In light of the bimolecular
chemical reactions philosophy, the described phenomenon
represents the molecule dynamics control instead of the usual
mesomery determined (encounter complex controlled) product
formation.
In conclusion, the electron transfer involving s- or p-type
radical cations according to reaction (13) or Scheme 1 should
be compared. On the basis of the experimental results and the
considerations given above, it caꢁn be stated that the saturated
parent radical cations (n-C4H9Cl þ) react in a manner which is
controlled by the molecule dynamics rather than by the meso-
ꢁ
þ
meric structure whereas aromatic radical cations (C6H6
seem to follow the latter reaction mode.
)
Stability of the silane radical cations
In order to interprete the decay of the benzyltrimethylsilane
type radical cations two reactions are to be considered.
Dinnocenzo et al.3,4 gave convincing arguments for a nucleo-
phile driven bimolecular decay (14).
ꢁ
ArCH2 SiR3 þ þ Nu ! Ar CHꢁ2 þ R3Siþ ꢉ ꢉ ꢉ Nu ð14Þ
Yet, in light of our quantum-chemical calculations, a uni-
molecular decay should be considered. It accompanies a rota-
tional motion of the C–Si bond from the stable planar
radical cation state to the transient perpendicular one. Then
the perpendicular transient immediately fragmentates, see
reaction sequence (15). Of course, the fragmentation depends
strongly on the type and size of the substituents at both
neighboring atoms and their electronic effects.
Experimental section
Materials
ꢁ
ꢁ
þ
þ
n-Butyl chloride has been purified by treating it with molecular
sieve (A4, X13) and distillation under nitrogen as described in
ref.13. The benzyltrimethylsilanes of types 1 and 2 where
synthesized according to literature.24
ArCH2 SiR3
! ½ArCH2 SiR3 ꢊperp:
plane
ꢁ
! Ar CH2 þ R3Siþ ð15Þ
The data for the rotation times (period of rotation) calcu-
lated along the ArC–Si bond both for the ground state and
radical cation state are given in Table 2. The rotation times
are quite comparable in the case of the grouꢁnd state ꢁmolecules,
but they differ significantly for the cations 1 þ and 2 þ (up to a
factor of three), which leads to a different decay mechanism.
As can be seen from Fig. 7(b), the decay of the radical
Pulse radiolysis
The liquid samples purged with nitrogen or oxygen were irra-
diated with high energy electron pulses (1 MeV, 12 ns dura-
tion) generated by a pulse transformer electron accelerator
ELIT (Institute of Nuclear Physics, Novosibrisk, Russia).
The dose delivered per pulse was measured using the absorance
of the solvated electron in slightly alkaline aqueous solution,
and was usually around 100 Gy. Detection of the transient spe-
cies was carried out using an optical absorption technique,
consisting of a pulsed xenon lamp (XBO 450, Osram), a Spec-
traPro-500 monochromator (Acton Research Corporation), a
R955 photomultiplier (Hamamatsu Photonics) and a 1 GHz
ꢁ
þ
cations 1 is well correlated with the calculated spin densities
ꢁ
þ
on the Si atom. Otherwise, the decay of radical cations 2 is
practically independent of the calculated spin densities on the
Si atom. This discrepancy could be explainedꢁby the essentially
slower rotational motion of radical cations 2 þ due to the very
large substituent R3 .
ꢁ
þ
ꢁ It is obvious that the decay of 1 is faster than that of
þ
2
by at least one order in magnitude. Yet, in the defined
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
2274
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5

View Article Online
10 CRC Handbook of Chemistry and Physics, 73rd edn., CRC Press,
Boca Raton, FL, 1992.
11 O. Brede, R. Mehnert and W. Naumann, Chem. Phys., 1987,
115, 279.
digitizing oscilloscope (TDS 640, Tektronix). Further details of
this equipment are given elsewhere.25
Quantum-chemical approach
12 (a) R. Mehnert, in Radical Ionic Systems: Properties in Condensed
Phases, eds. A. Lund and M. Shiotano, Kluver, Dordrecht, The
Netherlands, 1991, p. 231; (b) K. Toriyama, in Radical Ionic
Systems: Properties in Condensed Phases, eds. A. Lund and M.
Shiotano, Kluver, Dordrecht, The Netherlands, 1991, p. 99; (c)
M. Lindgren and M. Shiotaini, in Radical Ionic Systems:
Properties in Condensed Phases, eds. A. Lund and M. Shiotano,
Kluver, Dordrecht, The Netherlands, 1991, p. 125.
13 R. Mehnert, O. Brede and W. Naumann, Ber. Bunsen-Ges. Phys.
Chem., 1982, 86, 525.
Quantum-chemical calculations were performed using the
Gaussian 98, Revision 11 program package26 based on Density
Functional Theory (DFT) Hybrid B3LYP methods27–29 with a
standard 6-31G(d) basis set. The frequency calculations were
used to determine the nature of stationary points found by
geometry optimisation, and to obtain thermochemistry para-
meters such as zero-point energy (ZPE) and activation energy
Ea (height of rotation barrier).
14 A. Kira, Y. Nosaka and M. Imamura, J. Phys. Chem., 1980,
84, 1882.
15 H. Hiratsuka, Y. Kadokura, H. Chida, M. Tanaka, S. Kobayashi,
T. Okutsu, M. Oba and K. Nishiyama, J. Chem. Soc., Faraday
Trans., 1996, 92, 3035.
Acknowledgements
16 D. C. Neckers, S. Rajadurai, O. Valdes-Aguilera, A. Zakrzewski
and S. M. Linden, Tetrahedron Lett., 1988, 29, 5109.
17 (a) D. A. Tasis, M. G. Siskos and A. K. Zarkadis, Macromol.
Chem. Phys., 1998, 199, 1981; (b) V. Georgakilas, PhD Thesis,
University of Ioannina, Greece, 1998.
Financial support of this work by the program IKYDA of the
German and Greece Academic Exchange Services (DAAD,
IKY) is grateful appreciated.
18 (a) R. F. C. Claridge and H. Fischer, J. Phys. Chem., 1983, 87,
1960; (b) D. Meisel, P. K. Das, G. L. Hug, K. Bhattacharya
and R. W. Fessenden, J. Am. Chem. Soc., 1986, 108, 4706.
19 J. B. Lambert, Y. Zhao, R. W. Emblidge, L. A. Salvador, X. Liu,
J.-H. So and E. C. Chelius, Acc. Chem. Res., 1999, 32, 183.
20 C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91, 165.
21 (a) A. K. Zarkadis, PhD Thesis, University of Dortmund,
Germany, 1981; (b) H. Hillgartner, W. P. Neumann, W. Schulten
and A. K. Zarkadis, J. Organomet. Chem., 1980, 201, 197.
22 O. Brede, S. Naumov and R. Hermann, Radiat. Phys. Chem.,
2003, 67, 225.
23 D. Tasis, M. G. Siskos, A. K. Zarkadis, S. Steenken and G.
Pistolis, J. Org. Chem., 2000, 65, 4274.
24 V. Georgakilas, G. P. Perdikomatis, A. S. Triantafyllou, M. G.
Siskos and A. K. Zarkadis, Tetrahedron, 2002, 2441.
25 O. Brede, H. Orthner, V. Zubarev and R. Hermann, J. Phys.
Chem., 1996, 100, 7097.
References
1
2
H. Bock and B. Solouki, Chem. Rev., 1995, 95, 1161.
(a) M. G. Steinmetz, Chem. Rev., 1995, 95, 1527; (b) M. Freccero,
A. Pratt, A. Albini and C. Long, J. Am. Chem. Soc., 1998, 120,
284; (c) E. Baciocchi, M. Bietti and O. Lanzolung, Acc. Chem.
Res., 2000, 33, 243 and citations herein.
3
4
5
6
7
8
9
K. P. Dockery, J. P. Dinnocenzo, S. Farid, J. L. Goodman, I. R.
Gould and W. P. Todd, J. Am. Chem. Soc., 1997, 119, 1876.
H. J. P. de Lijser, D. W. Snelgrove and J. P. Dinnocenzo, J. Am.
Chem. Soc., 2001, 123, 9698.
O. Brede, R. Hermann, S. Naumov, G. P. Perdikomatis, A. K.
Zarkadis and M. G. Siskos, Chem. Phys. Lett., 2003, 376, 370.
O. Brede, G. R. Mahalaxmi, S. Naumov, W. Naumann and R.
Hermann, J. Phys. Chem. A, 2001, 105, 3757.
G. R. Mahalaxmi, R. Hermann, S. Naumov and O. Brede, Phys.
Chem. Chem. Phys., 2000, 2, 4947.
O. Brede, R. Hermann, W. Naumann and S. Naumov, J. Phys.
Chem. A, 2002, 106, 1398.
O. Brede, Res. Chem. Intermed., 2001, 27, 709.
26 M. J. Frisch et al., Gaussian 98, Revision A9, Gaussian Inc.,
Pittsburgh PA, 1998.
27 A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
28 A. D. Becke, J. Chem. Phys., 1996, 104, 1040.
29 Ch. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1987, 37, 785.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 2 2 6 7 – 2 2 7 5
2275