Localized electron transfer in nonpoiar solution: reaction of phenols and thiophenols with free solvent radical cations

Article abstract of DOI:10.1021/jp002701o

Free electron transfer (FET) is understood as the reaction of free and uncorrelated solvent parent radical cations with solutes characterized by a lower ionization potential than those of the solvent. We studied electron transfer from phenols and thiophen

Full text of DOI:10.1021/jp002701o

J. Phys. Chem. A 2001, 105, 3757-3764
3757
Localized Electron Transfer in Nonpolar Solution: Reaction of Phenols and Thiophenols
with Free Solvent Radical Cations
Ortwin Brede,*,† Mahalaxmi R. Ganapathi,† Sergej Naumov,‡ Wolfgang Naumann,‡ and
Ralf Hermann†
Interdisciplinary Group Time-ResolVed Spectroscopy, UniVersity of Leipzig, Permoserstr. 15,
D-04303 Leipzig, Germany, and Institute of Surface Modification, Permoserstr 15, D-04303 Leipzig, Germany
ReceiVed: July 27, 2000; In Final Form: January 25, 2001
Free electron transfer (FET) is understood as the reaction of free and uncorrelated solvent parent radical
cations with solutes characterized by a lower ionization potential than those of the solvent. We studied electron
transfer from phenols and thiophenols (as solutes) to molecular radical cations of some nonpolar solvents
(cyclohexane, n-dodecane, 1,2-dichloroethane, n-butyl chloride) using pulse radiolysis. For phenols (ArOH)
as solutes, along with the expected radical cations ArOH•+ , an unexpectedly comparable amount of phenoxyl
radicals (ArO•) was observed, which evidently arises in a parallel reaction channel of the type shown below
•+
with cyclohexane as the solvent: c-C6H12 + ArOH f c-C6H12 + ArOH •+, ArO•, H+solv. Analogous
observations were also made for thiophenols as solutes, with ArSH•+ and ArS• simultaneously occurring as
reaction products. The appearance of cations and radicals as parallel products can be attributed to two alternative,
locally different electron transfer pathways of FET. For example, in the case of phenol it was assumed that
transfer starts from either the aromatic ring or the hydroxyl group of the solute. The occurrence of ArO• as
a reaction product can then be understood if an efficient transfer barrier prevents rapid charge equilibration
in the ionized solute and, therefore, boosts deprotonation. On the basis of quantum chemical calculations,
this hypothesis is proven by analyzing the molecular oscillations. From the effects observed, general conclusions
about FET are derived which characterize this transfer as unhindered, extremely rapid electron jumps from
the donors to the holelike solvent radical cations taking place within almost the first collision between the
reactants in the solvent cage.
A* + D a [A*...D] a [A•-...D•+] a A•- + D•+ (2)
Introduction
Studying detailed mechanisms of electron transfer processes
in organic systems is a widespread subject of reaction kinetics.
Here we focus attention on ion molecule reactions of saturated
molecular radical cations in nonpolar solution, which exhibit
some surprising peculiarities of electron transfer dynamics.
Radical cations of organic molecules in liquid solution are
often metastable and decay on a nanosecond time scale by
deprotonation, fragmentation, reaction with nucleophiles, etc.
This holds in particular for radical cations of saturated nonpolar
liquid compounds such as alkenes and alkyl chlorides, which
can be generated radiolytically, known as parent ions.1-6
In contrast to this mechanism, the analogous equilibrium system
for the reaction of the parent ions involves only one charged
species in each step,3 where S stands for the solute.
RX•+ + S a [RX•+...S] a [RX...S•+] a RX + S•+ (3)
From reaction 3 it is apparent that this ion-molecule reaction
mechanism can be very efficient if a favorable ionization
potential difference between solvent and solute leads to a
reduced Franck-Condon barrier for the forward transfer step.10
Hence for process 3, the activation barrier should be much less
significant than in process 2, and may not even exist at all. For
this reason, and also by virtue of the involvement of free ions,11a
the ion-molecule reaction 3 has been defined as free electron
transfer (FET).12,13
Using the pulse radiolysis technique,14 FET has been utilized
in the past to generate a variety of solute radical cations, e.g.
refs 15-18. In a rapid process, product radical cations are
formed in (3) in what is assumed to be a very selective manner.
n-Alkanes,2 cyclohexane,1 dichloroethane,5 and in particular
n-butyl chloride3 were found to be useful as basic solvents.
Recently, a surprising product distribution was found for FET
from various phenolic compounds to solvent parent cations.
Instead of exclusively phenol radical cations (ArOH•+), phe-
noxyl radicals (ArO•) were also observed to be produced in a
comparable amount (cf. reaction 4a,b).12,13,19,20
RX ' RX•+ + e-
R ) alkyl, X ) H, Cl (1)
solv
In liquids at room temperature, these species exist for a few
hundred nanoseconds and have excellent electron acceptor
properties owing to the high ionization potential of the ground-
state molecules.
Regarding electron transfer mechanisms, the commonly
known photosensitized electron transfer7 is well-described by
the Rehm-Weller kinetic scheme,8 in which the transfer steps
include the reactants’ approach, the encounter complex forma-
tion, the transfer event, and ion separation9 (cf. eq 2, where *
denotes an electronically excited state):
† University of Leipzig.
‡ Institute of Surface Modification.
10.1021/jp002701o CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/28/2001

3758 J. Phys. Chem. A, Vol. 105, No. 15, 2001
Brede et al.
set26 using Gaussian 98. This enabled vibration analysis of the
phenol and thiol molecules and geometry calculation of frontier
orbitals.
Results
In pulse radiolysis, primarily the solvents are ionized. Then
the solvent parent radical cations react with the solutes (here,
e.g., ArOH) according to reaction 5.
The same phenomenon exists for aromatic thiols as electron
donors, with thiophenol radical cations (ArSH•+) and thiyl
radicals (ArS•) appearing as products of FET.12,13
It was demonstrated that the products ArOH•+ and ArO• were
actually formed by electron transfer4 because the decay of the
solvent parent ions and the formation of the two different
products takes place synchronously, as was confirmed by
analyzing the time profiles of all the species involved. Reaction
(4) proceeds at room temperature with rate constants of about
(1-2) × 1010 dm3 mol-1 s-1, which is typical of diffusion
control and in this case is slightly influenced by the primarily
inhomogeneous distribution of the radiation-generated parent
species.3,11b
In this paper we briefly report the experimental details, give
a detailed interpretation of the product distribution phenomenon
observed in the cases of phenols and thiophenols, and compre-
hensively consider the consequences for free electron transfer
in general.
RX•+ + ArOH f RX + products (ArOH•+, ArO•) (5)
As mentioned in the introduction, all the electron transfer
reactions involving the different phenols or thiophenols showed
a uniform diffusion-controlled rate constant, which was gov-
erned only by the solvent viscosity. In our case, all the rate
constants formally determined from the decay of the solvent
radical cation signals in the presence of solutes are k4,5
)
(1-2) × 1010 dm3 mol-1 s-1, because of the similar viscosities
of the solvents used13,19,20 (not documented here). The same
rate constant values were also observed using the deuterated
phenol 4-Cl-Ar-OD. Hence, typical scavenger concentrations
for taking product transient spectra are g2 × 10-3 mol dm-3
.
Taking this and diffusion control of reaction 5, after about 30
ns the reaction is almost complete. Using 15 ns pulses, the
spectrum of the products can be taken shortly after the pulse.
Although sometimes mentioned elsewhere under the aspect
of radical cation identification and decay kinetics,12,13,27 we
demonstrate and discuss in detail three prominent examples of
localized free electron transfer4 from a phenol, a thiophenol and
4-hydroxythiophenol to parent radical cations of different
nonpolar solvents.
Experimental Section
Pulse radiolysis experiments were performed with high-
energy electron pulses (1 MeV, 15 ns duration) generated by a
pulse transformer accelerator ELIT (Institute of Nuclear Physics,
Novosibirsk, Russia). The absorbed dose per pulse measured
with an electron dosimeter was between 50 and 100 Gy,
corresponding to transient concentrations of around 10-5 mol
dm-3. Details of the pulse radiolysis setup are reported
elsewhere.21,27
The solvents n-butyl chloride,3 1,2-dichloroethane,5 cyclo-
hexane,1 and n-dodecane4 were purified as reported. The
spectrograde substances from Aldrich and Merck were generally
dried using molecular sieve chromatography. The phenols and
thiophenols were of maximum commercial grade (Merck) and
used as received.
Phenol in 1,2-Dichloroethane Solution. Pulsing a solution
of 10-2 mol dm-3 phenol in 1,2-dichloroethane, electron transfer
(5) proceeds rapidly, and the product spectrum of the phenol
transients can be seen immediately after the 15 ns pulse (cf.
Figure 1). The transient spectrum shows three maxima at
wavelengths around 300, 400, and 440 nm. The spectral gap
around 280 nm is caused by the depletion of the parent phenol.
Upon adding 0.1 mol dm-3 ethanol, the 440 nm absorption is
completely quenched, whereas the remaining absorption is
markedly reduced compared to the ethanol-free sample. This
absorption fits those well-known for the phenoxyl radical28
(ArO•). The absorption peaking at 440 nm is classified as being
caused by the phenol radical cation (ArOH•+); see also ref 19.
The situation explaining the spectral behavior of Figure 1 can
be described by electron transfer4 synchronously forming
ArOH•+ and ArO•. The quenching effect of ethanol acts in two
modes, deprotonating the radical cation (reaction 6) and
competing with electron transfer4 by deprotonating the parent
cation (reaction 7)
Phenol and thiophenol radical cations have generally been
found to be much less sensitive to oxygen than radicals.13,20
Whereas phenol samples were purged with nitrogen gas,
thiophenol solutions were purged with pure oxygen; this serves
the transformation of radiolytically formed reactive alkyl radicals
into alkylperoxyls, which in our time range do not react with
thiols.22 Alkyl chlorides are internal scavengers for solvated
electrons formed by solvent ionization. Using the alkanes as
solvents, 0.1 mol dm-3 carbon tetrachloride was added to
scavenge electrons.23 In the nonpolar solutions, an imaginable
aggregate formation of phenols and thiophenols (dimers or
ArOH•+ + C2H5OH f ArO• + C2H5OH2
(6)
+
1
multimers) was ruled out by H NMR measurement of the
ClCH2CH2Cl •+ + C2H5OH f
chemical shifts depending on the solute concentration, using
ring current effects as the probe.
+
ClCH2C•HCl + C2H5OH2 (7)
The kinetic fits of the superimposed transient absorption time
profiles were performed using the programs ACUCHEM and
ACUPLOT,24 which numerically solve the differential equation
for the assumed reaction mechanism. To adjust the concentration
profiles to those of optical absorption, reasonable extinction
coefficients were used which were determined independently,
e.g. by one-electron oxidation in aqueous solution.20,25
Looking at the time profiles given in Figure 1b,c, at 400 nm
a longer lasting species and at 430 nm a short-living transient
can be distinguished, corresponding well to our interpretation
of the transient types ArO• and ArOH•+. Certainly, the two time
profiles represent superimpositions of both transients, with one
dominating.
Quantum chemical calculations were performed with the
density functional theory hybrid B3LYP with 6-31G(d) basis
Whereas the deactivation of phenoxyl radicals (reaction
8) is comparably slow, the phenol radical cation decays

Localized Electron Transfer in Nonpolar Solution
J. Phys. Chem. A, Vol. 105, No. 15, 2001 3759
TABLE 1: Ratio of the Products of the Electron Transfer
(4) Evaluated from the Fit Operation for Different Solvents
ratio
k9 × 106 [s-1], [ArOH•+]/
solvent
generating
RX•+
ArOH•+
[ArO•] in
phenol, ArOH
phenol
deprotonation FET (4a,b)
1.9
2.8
3.5
3.5
3.0
3.1
3.1
3.4
3.0
3.1
2.9
3.0
1.8
3.7
2.1
2.9
0.9
4.0
2.0
1.00
0.92
1.00
0.92
1.00
0.92
1.04
1.00
1.02
1.00
0.92
1.04
1.12
1.00
1.00
1.00
c-C6H12
n-C12H26
n-C4H9Cl
1,2-C2H4Cl2
c-C6H12
4-Cl-
n-C12H26
n-C4H9Cl
1,2-C2H4Cl2
n-C4H9Cl
n-C12H26
n-C4H9Cl
1,2-C2H4Cl2
c-C6H12
4-Cl-Ar-OD
4-MeO-
4-Me-2,6-di-t-Bu-
n-C12H26
n-C4H9Cl
1,2-C2H4Cl2
n-C4H9Cl
n-C4H9Cl
n-C4H9Cl
4-NH3-
4-NO2-
n-octadecyl
1.13
(3,5-di-t-Bu-4-OH-
phenyl)propionate
4-MeO-thiophenol, ArSH
4.5
4.8
4.8
3.2
0.82
0.75
0.82
0.85
c-C6H12
n-C12H26
n-C4H9Cl
1,2-C2H4Cl2
Because the delayed formation9 of the phenoxyl radicals is
directly proportional to the decay of the phenol radical cations,
we have a quantitative measure of the relation between the two
reaction channels of the FET4 divided into the formation of
ArOH•+ (4a) and ArO•‚(4b). In the example given, the portions
are nearly the same; i.e., both species are formed with the same
probability expressed by the ratio [ArOH•+]/[ArO•]. This seems
to be irrespective of the solvent used (cf. Figure 1) and amounts
for phenol to an average of 0.96. Analogous experiments, data
evaluation, and fits were performed with various substituted
phenols in different solvents, with nearly identical results
showing transient ratios around 1. All the resulting parameters
are given in Table 1. Only in the case of a sterically hindered
phenol with a very long tail in the 4-position did the transient
ratio of 1.13 indicate a larger amount of ArOH•+.
Figure 1. (a) Transient optical absorption spectra of a N2-bubbled
solution of 10-2 mol dm-3 phenol in 1,2-dichloroethane taken 30 ns
after the electron pulse (O). After adding 0.1 mol dm-3 ethanol (b),
the cation part is quenched. (b, c) Transient time profiles at λ ) 400
and 430 nm taken from the original sample (bold line). The fit curves
describe the decay of ArOH•+ (s · ·) and the ArO• kinetics (- -)
consisting in a rapid (4a) and delayed (9) part as well as the sum fit
(O).
markedly by a first-order process explained by deprotonation2
(reaction 9).
2ArO• f products
2k8 ) 3 × 108 dm-3 mol-1 s-1 (8)
ArOH•+ f ArO• + H+
k9 ) 3 × 107 s-1 (9)
solv
p-Methoxythiophenol in Cyclohexane. As already men-
tioned, thiophenols exhibit analogous effects in free electron
transfer (4), with the radical cations ArSH•+ and ArS• also being
formed in comparable amounts. The kinetic and spectral
parameters of the two transient types have already been listed
for many thiophenols in a former paper.13 Here we demonstrate
and analyze the situation of FET with respect to the transient
ratio using the new example of p-methoxythiophenol dissolved
in cyclohexane. Figure 2 gives the transient absorption spectra
of a solution of 10-2 mol dm-3 of the thiol in cyclohexane taken
immediately after the electron pulse. Three absorption bands
can be distinguished at λ ) 320, 520, and 580 nm. The latter is
short-lived and identified by the ethanol reaction analogous to
reaction 9 caused by the thiophenol radical cation. The other
two bands can predominantly be explained by the thiyl radical.
Representative time profiles are shown in parts b (510 nm) and
c (570 nm) of Figure 2. To save space, suffice it to state that
data analysis was performed in exactly the same manner as
We then fitted the time profiles numerically using reasonable
kinetic and spectroscopic parameters (parent cation generation
within the pulse followed by product formation with k4; for the
fate of the species using k8 and k9, fitting the time profiles with
the experimental ratio of the extinction coefficients). The kinetic
behavior of ArOH•+ and ArO• is described by the following
differential equations:
d[ArOH•+]/dt ) k4a[ArOH][RX•+] - k9[ArOH•+]
d[ArO•]/dt ) k4b[RX•+][ArOH] + k9[ArOH•+] - k8[ArO•]2
Figure 1b,c shows two examples of fits performed at various
wavelengths (400 and 430 nm) using the same rate constants
and only varying ꢀ values. The nonfitted spike (430 nm) is
explained by residual solvent cation absorption. Apart from this,
the fits excellently describe the decay of ArOH•+ as well as the
fast (due to reaction 4b) and delayed formation (due to reaction
9) of ArO•.
described above for phenol. The time profile at λ ) 570 nm
•+
shows the kinetic behavior of p-MeO-Ar-SH
with good
separation. The transient superimposition was fitted with

3760 J. Phys. Chem. A, Vol. 105, No. 15, 2001
Brede et al.
Figure 2. (a) Transient optical absorption spectra of an O2 bubbled
solution of 10-2 mol dm-3 4-methoxythiophenol in cyclohexane
containing 0.1 mol dm-3 carbon tetrachloride (O, taken after 30 ns).
After adding 0.1 mol dm-3 ethanol (b, taken 30 ns after the pulse),
the cation part is quenched. (b, c) Transient time profiles at λ ) 510
and 570 nm taken from the original sample (bold line). The fit curves
describe the decay of ArSH•+ (s · ·) and the ArS• kinetics (- -)
consisting in a rapid (4a) and delayed (9) part as well as the sum fit
(O). The profile in panel c describes nearly pure ArSH•+ kinetics.
Figure 3. Transient optical absorption spectra of an O2-bubbled
solution of 10-2 mol dm-3 4-hydroxythiophenol in n-butyl chloride
taken immediately after the electron pulse (b) and after 1.5 µs (O).
The unmarked spectra are calculated by adjusting the known spectrum12
of HO-Ar-S• at the 520 nm peak of the spectrum (O). Taking
differences, the spectrum of •OH-Ar-SH (bold faced, λmax) 350 and
•+
450 nm) and of HO-Ar-SH (λmax ) 550 nm) result.
after 1.5 µs (also shown in Figure 3), it can easily be seen that
a further radical transient is hidden in the sum absorption. Taking
the difference spectrum of this, only a typical phenoxyl radical
absorption spectrum appears. This clearly indicates the formation
of three different products formed in parallel in the FET and
formulated in eq 12, where for simplicity’s sake the proton
removed from the polar groups is not shown.
experimental parameters for radical recombination (10) and
radical cation deprotonation (11).
2ArS• f products
2k10 ) 3.5 × 109 dm3 mol-1 s-1
(10)
ArSH•+ f ArS• + H+
k11 ) 5 × 106 s-1 (11)
solv
OH-Ar-SH + n-C4H9Cl•+
f
OH-Ar-SH •+, •O-Ar-SH, OH-Ar-S• + n-C4H9Cl
In this example of free electron transfer analogous to reaction
4, the relation between the channels seems to be shifted toward
the formation of ArS• (55%) instead of ArSH•+ (45%), expressed
by a ratio value of 0.81. For many thiophenols studied, the FET
phenomenon with two synchronous reaction channels (5a,b) has
also been reported13.
FET Involving 4-Hydroxythiophenol Dissolved in n-Butyl
Chloride. When studying encounter geometry control of product
distribution in free electron transfer (4), the question arises as
to how benzene rings with more than one OH or SH group
behave. For this purpose we studied 4-hydroxythiophenol (10-2
mol dm-3) in n-butyl chloride solution. Figure 3 shows
absorption spectra of this sample taken immediately after the
15 ns electron pulse and after 1.5 µs, from which the radical
cation absorption can be taken as a difference spectrum showing
the absorption of HO-Ar-SH •+ peaking at 550 nm. This part
can be completely quenched by a small amount of ethanol as
elucidated above for the other product cations. With the known
spectrum of HO-Ar-S• (ref 12) adjusted at the 520 nm peak
(12)
These results convincingly show the product control of the
reactants’ encounter geometry. Because of the strong superim-
positions of the species (also seen in the inset of Figure 3 from
time profiles), we were unable to quantitatively analyze the
effect in the same way as described above. However, speculating
with extinction coefficients of similar radicals and radical cations
it can be assumed that the three channels (corresponding to the
three products) participate to about the same extent in the FET,
with the radical cation channel forming OH-Ar-SH •+ possibly
being of minor importance.
Discussion
Overall, the phenomenon of the encounter-controlled product
distribution in free electron transfer from phenolic and/or
thiophenolic molecules to parent ions of nonpolar solvents has

Localized Electron Transfer in Nonpolar Solution
J. Phys. Chem. A, Vol. 105, No. 15, 2001 3761
Within the encounter zone, the reactants form something like a
supermolecule, existing only in the momentum of the electron
transfer, which should take place in the subfemtosecond time
range. This is also the time range in which charge equilibration
should take place throughout the aromatic moiety (broken line
area).
Assuming this to be the case, it is unclear why the electron
jump at the hydroxyl group does not also undergo the rapid
charge equilibration. Here we hypothesize that a distinctly short-
lived steric barrier exists. Analysis of the dynamics of molecular
oscillations and in particular of the bonds around the oxygen
reaction center reveals the relatively slow deformation motion
of the bond from the aromatic C-atom to the phenolic oxygen
(a) (cf. Figure 4). Furthermore, there is a very extended rotation
motion of the OH group along the C-O axis (b). Both
oscillations happen in times of g100 fs, making them very slow
compared with intramolecular electron exchange. The above-
mentioned intramolecular oscillations upset the geometry favor-
able for intramolecular electron exchange until about 100 fs.
Whereas (a) brings the OH group out of the aromatic plane, (b)
changes the HOMO orbital interaction of the nonbinding
constellation at the C-O bond dramatically. Statistically speak-
ing, only after about 100 fs is a geometry achieved in which
charge equilibration can efficiently succeed. Table 2 summarizes
the molecular dynamic data as frequencies and times of one
motion.
Figure 4. Schematic illustration of the dynamics of the phenol
molecule involved in free electron transfer (4).
been clearly established by several experimental examples,
which can also be extended to larger aromatic ring systems such
as substituted naphthalenes and biphenyls.13,20 Below we
interpret the phenomenon in more detail.
Analysis of the Electron Transfer Reaction Steps. Con-
sidering the individual steps of the electron transfer according
to the equilibrium system,3 the reactants approach by molecular
diffusion, which is the slowest process in the sequence and,
therefore, rate-determining. Consequently, the FET as the gross
reaction also exhibits rate constants of diffusion control. This
involves any time-resolved measurements being limited to the
nanosecond time scale and being quasi-stationary with respect
to the higher rate of the other FET steps. However, the pulse
radiolysis experiments clearly rule out other conceivable forma-
tion paths for the phenol transients, such as normal radical
reactions or formation via electronically excited states.
If the parent ions and the solute meet by diffusion, different
sterical configurations ought to be possible in the encounter
position. Therefore, if not sterically excluded, electron transfer
from the solute to the solvent ion should be possible from both
the aromatic ring and the hydroxyl group. Hence FET takes
place in nearly every reactant approach, producing parallel
phenol radical cations and a local cation center in the hydroxyl
group. At least two types of encounter situations should occur,
as symbolized in the eq 13 reaction sequence of FET from
phenol to n-butyl chloride radical cations.
Clearly, distinguishing between the two transfer product
formation channels as given in reactions 13 only makes sense
if there is a bottleneck between the aromatic ring and the
hydroxyl group impeding rapid internal electron equilibration
after the electron transfer step. Hence the deprotonation of Ar-
O•+H is favored, rather than its stabilization by charge distribu-
tion.
This interpretation can theoretically be assisted by analyzing
the dynamics of the phenol molecule in ultrashort times (cf.
Figure 4). The phenol molecule is surrounded by an ellipsoid
symbolizing the encounter region. This ellipsoid is arbitrarily
divided into two subspheres each describing the range of highest
probability of local electron transfer, at either the aromatic ring
or the hydroxyl group. This electron transfer is suggested to be
an extremely rapid electron jump, proceeding in each encounter
geometry immediately or after just a few collisions. This is
assumed due to the different local ionization potentials, which
are higher for the hydroxyl than the aromatic group. If several
encounters were needed for successful electron transfer, the
highest enthalpy difference ought to control the process;
however, this is contrary to all the experimental observations.
Compared to this, the vibration motion of the O-H bond (c)
is at least 30 times faster than the above-mentioned fluctuating
electron exchange barrier caused by deformation and rotation
motions (a and b). c brings about additional polarization of the
bond, and so the proton can be set free. Additionally, the escape
of H+ is promoted by the nucleophilic and solvation properties
of the solvent.
Concerning the orbital symmetries, the C-O bond MO, which
results from the negative or out-of-phase combination of two
atomic orbitals, is of an antibonding character and has a node
(value of zero) between the atoms. That the electronic coupling
really changes during the rotation of the OH group around the
C-O axis can be seen from orbital symmetries. Considering
the Klopman-Salem equation28,29 for describing the interaction
energy of two molecules in the ground state, we used the
coefficients for atomic orbital interaction Cc,o to assess the
intramolecular electron exchange situation between the OH
group and the aromatic ring, in particular concerning the O-C
barrier. This was calculated for situations in which the OH group
is twisted out of plane by 30°, 60°, or 90°. Then the quantum
chemically calculated Co and Cc values of the orbitals change
by decreasing the overlap between AOs (bonding character),
which is well-expressed by the product of the squares of the
2
2
C-values [Co × Cc ]. Hence, when using normalized values, a
change from 1.00 (0°), 0.84 (30°), 0.35 (60°), to 0.26 (90°)
results. This clearly indicates that the electronic coupling element
of the molecular electron exchange varies considerably with the
twist of the OH along the C-O axis.
Taking the electron equilibration as intramolecular electron
transfer, the electronic coupling matrix element strongly influ-
ences the transfer rate by modulating the transfer probability
depending on the angle of twist of the OH group. On an

3762 J. Phys. Chem. A, Vol. 105, No. 15, 2001
Brede et al.
TABLE 2: DFT B3LYP/6-31G(d) Calculated Frequency
Analysis of the Molecular Oscillations around the Polar
Group XH in Phenol and Thiophenola
Furthermore, in distinction to the electron transfer processes
(4a,b) discussed in this paper, H atom transfer (14) should
proceed mostly activation-controlled via a defined encounter
state (see also the subsequent paragraphs).
ArOH
ArSH
General Consequences for the FET Mechanism. Whereas
in radiation chemistry the nonrelaxed solvent radical cations
(holes) have been the subject of much attention by virtue of
their extreme mobility (up to 100 times faster than molecular
ions31-34), electron transfer reactions of the relaxed molecular
parent radical cations have been less thoroughly investigated.
Concerning holes, the existence of such unusually mobile species
seems to be a general phenomenon in the radiation-induced
ionization of saturated hydrocarbons,33b largely influenced by
the actual structure of the alkanes, especially alicycles such as
cyclohexane31a-c and decaline,31d which have a longer lifetime.
The high mobility of holes is also reflected by free electron
transfer31a,33a with its extremely high rate constants.
motions
locality ν × 1012 (Hz) t (fs) ν × 1012 (Hz) t (fs)
deformation 1 C-XH
deformation 2 C--XH
rotation of OH C--XH
7.0
23.3
143
43
5.5
14.4
3.5
182
69
10.8
112.4
92
8.9
285
12.3
vibration
C--X--H
81.3
a Deformation of the polar group within the aromatic plane:
3.9°(ArOH), 3.3°(ArSH). Times (t) are those of one motion.
electronic basis, this provides a strong argument for the existence
of the above-mentioned fluctuating barrier for the intramolecular
electron exchange.
Considering the vibration dynamics of molecules, isotope
labeling enables the frequencies of the various molecular modes
to be influenced. Concerning the C-O bond, instead of labeling
O18-H, substituting oxygen by sulfur as in thiolphenols should
be considered. Here, the heavy sulfur reduces the frequencies
of a and b, and the vibration mode c also lasts a bit longer.
Table 2 compares the discussed molecular motions for phenol
and thiophenol. And indeed, the transient product ratio observed
after FET (4) is different in this path (4b) with rapid ArS•
formation seeming to dominate (cf. Figure 1). Speculating about
the reason, the larger volume of the sulfur group could be
responsible for this effect. Also in this case, the different solvents
used do not affect the product ratio.
[c-C6H12]•+* + S f c-C6H12 + S•+
k14 ) 1012 dm3 mol-1 s-1 (ref 33a) (15)
where the asterisk denotes a molecule in the hole state
Ignoring the question over the mechanism of hole motion,
the high rate values of reaction 15 indicate that the electron
transfer step in FET is very rapid, and the approach of the
reactants by diffusion or other forms of motion is generally the
rate-determining step of FET.
Returning to the subject of this paper, pulse radiolysis studies
in the 1970s showed that molecular and relaxed radical cations
of alkanes (gC6), of C6 alicycles such as cyclohexanes, and of
alkyl chlorides are metastable up to 200 ns in liquids and at
room temperature and, therefore, can be used for FET10,35 (4).
In accordance with plausibility considerations and also with low-
temperature EPR measurements,35a quantum chemical calcula-
tions show that in the parent ions the charge is distributed
throughout the entire saturated molecules,35b certainly partially
influenced by the actual molecular structure. Seen as a defect
electron in an exclusive σ bond system, the term “hole” also
has some justification for the molecular parent ion.
As such holelike species in condensed systems are the subject
of this investigation, it is no surprise that these parent radical
cations normally react in a diffusion controlled manner in the
free electron transfer assumed to be a function of the reaction-
enthalpy difference between the partners expressed, e.g., by the
ionization potential difference.10 Compared with other molecular
electron transfers in liquids such as the well-known photosen-
sitized process (2), FET exhibits a number of peculiarities as
listed below.
Labeling the phenolic hydrogen by substituting with deute-
rium does not bring about an effect different from that of normal
phenol, neither for the electron transfer rate k4 nor for the product
distribution. This was studied in detail for 4-chlorophenol and
is documented in Table 1. Then again, this result is not surprising
because the O-D vibration is not expected to differ dramatically
from that of O-H, calculated to be 1.4 times slower.
Hence for phenols and thiols of different electronic structure,
the identical effect in the free electron transfer reaction (5) is
observedsthe synchronous formation of both metastable radical
cations and stable heteroatom localized radicals. The indepen-
dence of the molecule electronics can be understood in terms
of the rapid, free electron transfer, which is only governed by
the geometry of collisions. The same reason seems to hold also
for the case of the long-chain substituted phenol, where the
remaining C18 might have an antenna function, increasing the
yield of ArOH•+.
Concerning reaction (4b), such processes are often considered
as normal one-step deprotonation reactions. In this case,
however, we are convinced that an extremely fast electron
transfer (within about 10-15 s) is taking place, followed by a
delayed fast deprotonation driven by the O-H vibration
oscillation (around 10-13 s); otherwise, for simple deprotonation,
an influence of the electronic structure of the molecules would
be expected. Therefore in this particular case of the phenols
and thiols, the principal reaction paths of electron transfer (4a)
and electron transfer with slightly delayed deprotonation (4b)
can be distinguished, whereas simple deprotonation (cf. 9) in
this system takes much longer. On the basis of our experiments,
we can also rule out an imaginable hydrogen atom transfer (14),
(1) FET only takes place in nonpolar solvents where the
parent radical cations are metastable for kinetically reasonable
times (between 10 and 200 ns). In our experience acetone is
the solvent with the highest polarity which can still be
applied,15,17 characterized by dielectric constant (ꢀ ) 20.56) and
dipolar moment (µ ) 2.70), but best defined by Reichart’s
N
solvent polarity parameter36 ET ) 0.355. By the way, phenol
radical cations generated in acetone exhibit a lifetime of a few
nanoseconds.
(2) Because of the low polarity of the solvents, the majority
of the excess energy of the FET step ought to be dispersed by
molecule oscillation modes of the reactants. The solvent
reorganization energy is only of very minor importance.9
Therefore, years ago we explained the dispersion of -∆G by
the excitation of different reactant molecule vibrations.10 This
model well explains the energetic relations in FET.
RX•+ + HO-Ar f [RXH•+...HO-Ar] f RXH+ + ArO•
(14)
which could compete with reaction 4a instead of reaction 4b.
Such a hydrogen transfer should show a kinetic isotope effect
for OD labeling, which was not found to be the case (see above).

Localized Electron Transfer in Nonpolar Solution
J. Phys. Chem. A, Vol. 105, No. 15, 2001 3763
(3) As typical of most electron transfer types, the change to
the electronic state in the transfer step is rapid compared with
the time necessary for nuclear rearrangements in the reactant
molecules. But as already mentioned, the time of the electron
jump in FET is a few orders of magnitude faster than sensitized
processes (2), the encounter state of which was recently
characterized in a time-resolved manner37 using femtosecond
spectroscopy.
case to three distinct products. Such localized electron jumps
are assumed to also occur for all other scavenger molecules,
but in most cases they cannot be detected. This is generally a
distinct feature of electron transfer reactions, which to the best
of our knowledge is unknown in this localized form precisely
for such small molecules (such phenols) in solution. Up till now,
similar local electron transfer effects were observed only from
aromatic donor groups separated by more or less rigid saturated
hydrocarbon bridge spacers (e.g. by 10 σ bonds corresponding
to 12 Å) to solvent ions.39-41
From the detailed studies reported in this paper, some general
conclusions can be drawn about the mechanism of free electron
transfer. Courageously, it should be stated that transfer as a real
electron jump takes place extremely rapidly in the subfemto-
second time range, as generally unhindered electron motion is
understood to proceed. Because of the holelike nature of the
parent ions, the electron jump proceeds adiabatically in an
electronically strongly coupled supermolecule of the reactants.
This calls for common equilibrium considerations with encounter
and successor complexes as intermediates. In this sense, this
paper presents new information on electron transfer dynamics
in the condensed phase.
(4) Regarding the rapid electron jump in FET suggested to
be comparable with the rate of intramolecular electron exchange
in aromatic systems proceeding in times roughly estimated to
lie around 10-15 s, the term encounter complex becomes more
formal and evolves into a question of philosophy in reaction
dynamics. We consider the electron jumps to comprise an
adiabatic transition in an electronically strongly coupled system.
The term “encounter complex” hence takes on the meaning of
a supermolecule formed by the reactants in which the electron
transition is practically unhindered. Assuming such a super-
molecule, the interaction of the reactants (“collisions”) are
relative motions of the partners in ultrashort times. Hence, the
electron jump can really take place in the subfemtosecond time
range. However, we are aware that this point is highly critical
and should be the subject of further theoretical investigation.
In the Introduction, the formulation used of the individual
steps of the free electron transfer as a system of equilibria (3)
should be also reconsidered. As we stated, FET is the energeti-
cally favored filling of a positive hole (parent ion), and such a
process has very limited reversibility. After electron transfer
(5), the former parent ion becomes part of the weak solvation
shell of the product cation and enters something like an
anonymous situation. The memory of the former reaction partner
is therefore lost. The inherent properties of free electron transfer
discussed in this paragraph justify considering this reaction type
as a very distinct version of the commonly known electron
transfer processes. Here it should be mentioned that this effect
could be also considered as two steps in a resolved special case
of dissociative proton coupled electron transfer theoretically
treated by Cukier.38
References and Notes
(1) Mehnert, R.; Brede, O.; Bo¨s, J. Z. Chem. (Leipzig) 1977, 17, 268.
(2) Brede, O.; Mehnert, R., Naumann, W.; Cserep, Gy. Radiat. Phys.
Chem. 1982, 20, 155.
(3) Mehnert, R.; Brede, O.; Naumann, W. Ber. Bunsen-Ges. Phys.
Chem. 1982, 86, 525.
(4) Mehnert, R.; Brede, O.; Naumann, W. Ber. Bunsen-Ges. Phys.
Chem. 1984, 88, 71.
(5) Shank, N. E.; Dorfman, L. M. J. Chem. Phys. 1970, 52, 4441.
(6) Le Motais, B. C.; Jonah, C. D. Radiat. Phys. Chem. 1989, 33, 505.
(7) Marcus, R. A. Annu. ReV. Phys. Chem. 1964, 15, 155 and
subsequent papers.
(8) Rehm, D.; Weller, A. Ber. Bunsen-Ges. Phys. Chem. 1969, 73, 834.
(9) Survey in: Kavarnos, G. J. Fundamentals of Photoinduced Electron
Transfer, VCH Publ.: New York, 1993; pp 287.
(10) Brede, O.; Mehnert, R.; Naumann, W. Chem. Phys. 1987, 115, 279.
(11) (a) Busi, F. In The Study of Fast Processes and Transient Species
by Electron Pulse Radiolysis;. Baxendale, J. H., Busi, F., D., Eds.; Reidel
Publ. Comp.: Dordrecht, 1982; pp 418-421. (b) Hummel, A. In Radiation
Chemistry, Principles and Application; Fartahaziz, Rodgers, M. A. J., Eds.;
VCH Publ.: Weinheim, 1987; pp 98-135.
Conclusions
(12) Dey, G. R.; Hermann, R.; Naumov, S.; Brede, O. Chem. Phys. Lett.
1999, 310, 137.
(13) Hermann, R.; Dey, G. R.; Naumov, S.; Brede, O. Phys. Chem.
Pulse radiolysis experiments on the free electron transfer from
phenols to solvent parent radical cations of alkanes and alkyl
chlorides resulted in the surprising observation that, in addition
to the expected product radical cations, phenoxyl radicals were
also formed to a comparable extent. We explain this phenom-
enon by local electron jumps involving both molecule functions
as the aromatic ring and the hydroxyl group. Whereas in the
first case the charge is immediately equilibrated throughout the
aromatic moiety (subfemtosecond range), the ionized hydroxyl
group exists locally for up to 100 fs due to a slowly fluctuating
reaction barrier in the form of deformation and rotation
oscillations. Hence the kinetically dominating process is depro-
tonation at the positively charged heteroatom group rather than
charge equilibration. This deprotonation is assumed to proceed
in the time range of the vibration oscillation of the O-H bond,
which is faster than the fluctuating barrier.
Chem. Phys. 2000, 2, 1213.
(14) Sauer, M. C.; Fielden, E. M.; Roffi, G. in The Study of Fast
Processes and Transient Species by Electron Pulse Radiolysis, Eds.
Baxendale, J. H.; Busi, F., D. Reidel Publ. Comp.: Dordrecht, 1982. pp
35-89.
(15) Rodgers, M. A. J. J. Chem. Soc., Faraday 1 1973, 68, 1278, 2036.
(16) Brede, O.; Bo¨s, J.; Helmstreit, W.; Mehnert, R. Radiat. Phys. Chem.
1982, 19, 1.
(17) Lomoth, R.; Naumov, S.; Brede, O. J. Phys. Chem. A 1999, 103,
2641.
(18) Guldi, D. M.; Asmus, K.-D. J. Am. Chem. Soc. 1997, 119, 5744.
(19) Brede, O.; Orthner, H.; Zubarev, V.; Hermann, R. J. Phys. Chem.
1996, 100, 7097.
(20) Mohan, H.; Hermann, R.; Naumov, S.; Mittal, J. P.; Brede, O. J.
Phys. Chem. A, 1998, 102, 5754.
(21) Brede, O.; David, F.; Steenken, S. J. Chem. Soc., Perkin Trans. 2
1995, 2, 23.
(22) Scho¨neich, C.; Dillinger, U.; von Bruchhausen, F.; Asmus, K.-D.
Arch. Biochem. Biophys. 1992, 292, 456.
(23) Mehnert, R.; Brede, O.; Naumann, W. J. Radioanal. Nucl. Chem.
1986, 101, 307.
(24) Braun, W.; Herron, J. T.; Kahaner, D. K. Int. J. Chem. Kinet. 1988,
20, 51.
(25) Alfassi, Z. B.; Schuler, R. H. J. Phys. Chem. 1985, 89, 3359.
(26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648; Becke, A. D. J. Chem.
Phys. 1996, 104, 1040.
(27) Mahalaxmi, G. R.; Hermann, R.; Brede, O. Phys. Chem. Chem.
Phys. 2000, 2, 4947.
This phenomenon was observed for many phenols and
thiophenols as well as hydroxyl- and thiyl-substituted naphtha-
lenes and biphenyls, showing that the reported observations are
due to a general effect of electron donors such as aromatics
substituted with OH and SH groups, which are rather unstable
in ionized form. Hence, even for double-substituted aromatics
such as 4-hydroxythiophenol (HO-Ar-SH), local energy
transfer in a modified form was demonstrated, leading in this

3764 J. Phys. Chem. A, Vol. 105, No. 15, 2001
Brede et al.
(28) Land, E. J.; Porter, G. J. Chem. Soc. Trans. Faraday Soc. 1966,
59, 2027.
(29) Klopman, G. J. Am. Chem. Soc. 1968, 90, 223.
(35) (a) Lindgren, M., Shiotani, M. In Radical Ionic SystemssProperties
in Condensed Phases; Lund, A., Shiotani, M., Eds.; Kluwer Academic
Publ.: Dordrecht, 1991; pp 125 (b) Mehnert, R. In Radical Ionic Systemss
Properties in Condensed Phases Lund, A., Shiotani, M., Eds.; Kluwer
Academic Publ.; Dordrecht, 1991; p 231.
(36) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry;
VCH: New York, 1990.
(37) Zhong, D.; Bernhardt, T. M.; Zewail, A. H. J. Phys. Chem. A 1999,
103, 10093.
(38) Cukier, R. I. J. Phys. Chem. A 1999, 5989
(39) Closs, G. L.; Miller, J. R. Science 1988, 240, 440.
(40) Johnson, M. D.; Miller, J. R.; Green, N. S.; Closs, G. L. J. Phys.
Chem. 1989, 93, 440.
(30) Salem, L. J. Am. Chem. Soc. 1968, 90, 543, 553.
(31) (a) Zador, E.; Warman, J. M.; Hummel, A. Chem. Phys. Lett. 1973,
23, 363; (b) de Haas, M. P.; Warman, J. M.; Infelta, P. P.; Hummel, A.
Chem. Phys. Lett. 1975, 31, 382. (c) Warman, J. M.; Infelta, P. P.; de Haas,
M. P.; Hummel, A. Chem. Phys. Lett. 1976, 43, 321. (d) de Haas, M. P.;
Hummel, A.; Infelta, P. P.; Warman, J. M. Can. J. Chem. 1977, 55, 2249.
(32) Beck, G.; Thomas, J. K. J. Phys. Chem. 1972, 76, 3856.
(33) (a) Brede, O.; Helmstreit, W.; Mehnert, R. Chem. Phys. Lett. 1974,
28, 43; (b) Brede, O.; Bo¨s, J.; Naumann, W.; Mehnert, R. Radiochem.
Radioanal. Lett. 1978, 35, 85.
(34) Review: Trifunac, A. D.; Sauer, M. C., Jr.; Shkrob, I. A.; Werst,
D. W. Acta Chem. Scand. 1997, 51, 158.
(41) Warman, J. M.; Hom, M.; Paddon-Row: M. N.; Oliver, A. M.;
Kroon, J. Chem. Phys. Lett. 1990, 172, 114.