Diphenol radical cations and semiquinone radicals as direct products of the free electron transfer from catechol, resorcinol and hydroquinone to parent solvent radical cations

Article abstract of DOI:10.1039/b205771a

In the pulse radiolysis of solutions of catechol, resorcinol and hydroquinone in n-butylchloride, dihydroxybenzene radical cations (40%) as well as semiquinone radicals (60%) are observed as direct synchronously formed products of the electron transfer from the solvent parent ions to the solute. This is explained in terms of free electron transfer succeeding in nearly every encounter of the reactants, which in the case of the studied dihydroxybenzenes involves femtosecond molecular dynamics effects. The rotation of the C-OH bond causes cycling of the molecule through transient conformations also exhibiting different electron distributions. From the more chemical point of view, the diphenol radical cations represent the first and till now unknown intermediates of oxidative semiquinone radical formation.

Full text of DOI:10.1039/b205771a

View Article Online / Journal Homepage / Table of Contents for this issue
Diphenol radical cations and semiquinone radicals as direct products of
the free electron transfer from catechol, resorcinol and hydroquinone
to parent solvent radical cations
a
Ortwin Brede,* Sudhir Kapoor, Tulsi Mukherjee, Ralf Hermann and Sergej Naumov
c
c
a
b
a
Interdisciplinary Group for Time-Resolved Spectroscopy, University of Leipzig, Permoserstr. 15,
D-4303, Leipzig, Germany. E-mail: brede@mpgag.uni-leipzig.de
Institute of Surface Modification, D-4303, Leipzig, Germany
Radiation Chemistry and Chemical Dynamics Division, Bhabha Atomic Research Centre, Trombay,
Mumbai 400085, India
b
c
Received 14th June 2002, Accepted 20th August 2002
First published as an Advance Article on the web 11th September 2002
In the pulse radiolysis of solutions of catechol, resorcinol and hydroquinone in n-butylchloride,
dihydroxybenzene radical cations (40%) as well as semiquinone radicals (60%) are observed as direct
synchronously formed products of the electron transfer from the solvent parent ions to the solute. This is
explained in terms of free electron transfer succeeding in nearly every encounter of the reactants, which in the
case of the studied dihydroxybenzenes involves femtosecond molecular dynamics effects. The rotation of the
C–OH bond causes cycling of the molecule through transient conformations also exhibiting different electron
distributions. From the more chemical point of view, the diphenol radical cations represent the first and till now
unknown intermediates of oxidative semiquinone radical formation.
formed, i.e. a conjugated cation (route (2a)) and an oxygen
localized one (route (2b)), where the latter deprotonates imme-
diately.
Introduction
Parent radical cations derived from non-polar solvents (RX)
such as n-alkanes, cycloalkenes and alkyl chlorides are very
efficient electron acceptors. So they are nearly universal tools
Now the question arose whether or not the mentioned effect
applies also for more functional phenols and, in particular for
dihydroxybenzenes such as catechol, resorcinol and hydroqui-
none. The oxidation of these compounds has been well inves-
tigated for a long time and the first step is found to result in
1
,2
for the generation of solute radical cations (S).
ꢀ
þ
ꢀþ
c-C6H12 ; n-C4H9Cl etc: þ S !
ꢀ
þ
9
–
S
þ c-C6H12; n-C4H9Cl . . . ð1Þ
semiquinone radicals or their corresponding anionic forms.
So aside from details of the electron transfer mechanism
1
3
3
This reaction type, free electron transfer (FET), is well studied
by pulse radiolysis. Generally, it has been found that the rate
of eqn. (1) is governed by diffusion kinetics. But recently it
has seemed probable that the real electron jumping step, on
a microscopic scale, proceeds over times typical of molecular
the existence and properties of possible diphenol radical
cations as the real primary oxidation products are of great
interest.
4
vibrations, i.e. in the range of tenths of femtoseconds. This
statement was derived from electron transfer experiments ana-
Experimental
5
logous to eqn. (1) using phenols, thiophenols and selenols
,6
7
8
Pulse radiolysis
as electron donors. In these cases, using nanosecond pulse
radiolysis, we found solute radical cations as well as hetero-
centered radicals, for phenols (ArOH) in comparable amounts,
see reaction (2).
The liquid samples were irradiated with a high energy electron
pulse (1 MeV, 15 ns duration) generated by a pulse transfor-
mer type accelerator ELIT (Institute of Nuclear Physics,
Novosibirsk, Russia). The dose delivered per pulse was mea-
sured with an electron dosimeter and was usually around 100
Gy unless otherwise stated. Detection of the transient species
was carried out using an optical absorption set-up consisting
of a pulsed xenon lamp (XBO 900, Osram), a Spectra Pro-
500 monochromator (Action Research Corporation), an IP28
photomultiplier (Hamamatsu Photonics) and a 1 GHz digitiz-
ing oscilloscope (TDS 640, Tektronix). Further details are
ꢀ
þ
RX þ Ar y OH
(
ꢀ
þ
ꢀ
ArOHe
!
þ RX ð2Þ
þ
ꢀ
H ! ArO þ Hþ
!
Ar O
In reaction (2), it is implied that FET happens during the rota-
tion of the OH group (symbolized by ‘‘y’’) which has a fre-
quency of around 10 Hz. In the course of this rotation the
molecular situation could be understood in terms of two
boundary cases, a plane structure and a twisted one (OH group
twisted 90 degree out of plane) which should exhibit also dif-
1
3
1
4
given elsewhere.
Unless otherwise stated the pulse radiolysis experiments
were carried out in nitrogen-bubbled solutions of n-butyl
chloride. The solutions were continuously passed through the
sample cell with a path length of 1 cm.
4
ferent electron distributions. Assuming direct and non-hin-
dered FET, two types of intermediate cations should be
5
096
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104
This journal is # The Owner Societies 2002
DOI: 10.1039/b205771a

View Article Online
Data analysis
The lifetimes of the biphenol radical cations were obtained
from simple first order exponential regression of the experi-
mental time profiles.
The corresponding rate constants for various paths involved
in the decay of the radical cations were used to simulate the
decay profiles of biphenols radical cations and its correspond-
ing phenoxyl radicals. Details of the procedure are reported
4
,15
earlier.
Chemicals
All the biphenols studied were commercially available
(
Aldrich). Chlorobutane and 1,2-dichloroethane were obtained
from Merck. Triethylamine (Aldrich), dimethylsulfoxide
Aldrich) and triphenylamine (Aldrich) were used as received.
(
Results
Pulse radiolysis of catechol in non-polar solution
The interaction of ionizing radiation with non-polar solvents
results in the generation of sigma-type radical cations which
are metastable in the nanosecond timescale. So for n-butyl
chloride (n-BuCl) and 1,2-dichloroethane (DCE) cationic
transients are observed with broad absorption maxima in the
visible peaking around 550 nm and exhibiting maximum life-
times up to 150 ns.
2
Fig. 1 Transient spectra of the pulse radiolysis of a N purged solu-
2
1
6
4
ꢁ2
ꢁ3
tion of 10 mol dm catechol in n-butylchloride taken (S) immedi-
ately after the pulse and after (X) 30 and (c) 250 ns. The time
profiles demonstrate the kinetic behavior in the absorption maxima.
n-C4H9Cl A n-C4H9Cl þ; Cl ; n-C4H9 ; etc:
ꢀ
ꢁ
ꢀ
ð3Þ
In contrast, because of its lower ionization potential, TEA
acts as an efficient hole and radical cation scavenger:
As crude information about the driving force of the electron
transfer eqn. (1), its free energy can be estimated from the
ionization potential difference between the solute and sol-
vent. The gas phase ionization potentials of catechol, resor-
cinol and hydroquinone are around 8.2 eV, while for the
solvents 1,2-dichloroethane and for n-butyl chloride the
values are 11.0 eV and 10.7 eV, respectively. Hence, with an
enthalpy difference of around 2.5 eV for the gross elec-
tron transfer reaction (1) involving the dihydroxybenzenes,
diffusion-controlled rate constants are expected and indeed
measured.
HO Ar OHe þ þ TEA ! HO Ar OH þ TEAꢀþ ð5Þ
ꢀ
In the presence of millimolar concentrations of TEA, the time
profiles at 390 nm show also an increase in rate, but also deple-
tion of the catechol radical absorption (not shown). Concomi-
tantly, around 330 nm the formation of TEA species could be
observed. The quenching rate constant was determined as
1
7
On pulse-irradiating samples containing catechol, the
absorption band of the solvent radical cations disappears with
concomitant formation of solute transients with absorption in
the wavelength region between 350 and 450 nm, cf. Fig. 1. In
ꢁ
2
1
k
0
mol solution of catechol, according to the transfer rate
ꢁ1 ꢁ1
s all the product absorption is formed
1
0
3
2
¼ 10 dm mol
practically within the electron pulse, but caused, however, by
different species which can be seen from the lifetimes in differ-
ent spectral positions, cf. the spectra and time profiles given in
Fig. 1. So a short living species with lmax ¼ 400 nm (assigned
as radical cation) and a long-lasting one with two bands at
l
max ¼ 365 and 375 nm (assigned to be the catechol semiqui-
none radical) are distinguishable.
The kinetic identification of the short-living transient was
made by adding typical cation scavengers such as dimethyl-
sulfoxide (DMSO) or triethylamine (TEA).
1
8
In the presence of DMSO, the time profiles at l ¼ 390 nm
show an increase in rate, according to a pseudo first order reac-
1
0
3
ꢁ1 ꢁ1
s , see inset in Fig. 2.
The difference of the spectra of the 10 molar catechol solu-
tion with k ¼ 1.5 ꢂ 10 dm mol
4
ꢁ2
ꢁ
3
ꢁ3
tion in the absence and in the presence of 10 mol dm
DMSO yielded the absorption band of the catechol radical
cation, see Fig. 2. Because DMSO does not cause a marked
reduction of the radical absorption we concluded that DMSO
causes deprotonation of the radical cation, eqn. (4), rather
than electron transfer quenching.
ꢁ2 ꢁ3
2
-purged solution of 10 mol dm
Fig. 2 Transient spectra of a N
catechol in n-butylchloride taken 100 ns after the electron pulse: (S)
ꢁ3
ꢁ3
mol dm
this solution, (N) sample additionally containing 10
DMSO, (c) difference spectrum between (S) and (N). The time profile
demonstrates the influence of DMSO on the cationic time profile. The
Stern–Volmer plot concerns the DMSO quenching reaction, eqn. (4).
HO Ar OHe þ þ DMSO ! HO Ar O þ H ðDMSOÞ ð4Þ
ꢀ
ꢀ
þ
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104
5097

View Article Online
Fig. 3 Effect of different doses per pulse (a: 100 Gy, b: 50 Gy, c: 20
Gy) on the time profile of the catechol radical cation at l ¼ 400 nm
Fig. 4 Example of simulation of the time profiles of superimposed
radical cation and phenoxyl radical absorptions with eqns. (8) and (9).
ꢁ2
ꢁ3
measured in a solution of 10 mol dm catechol in n-butylchloride.
1
0
3
ꢁ1 ꢁ1
s . These experiments confirm the
k
transient assignment as stated above.
5
¼ 2ꢂ10 dm mol
d½ArO ꢃ=dt ¼ k2a½ArOHꢃ½RXꢀþꢃ þ k6½ArOHꢀþꢃ ꢁ k10½ArO ꢃ
ꢀ
ꢀ 2
Analyzing the fate of the observed catechol species, the
decay of the radical cation was found to be first order with a
lifetime of about 700 ns while the phenoxyl radical decays by
second order kinetics. In fact there are various routes by which
phenol radical cations can decay, among them quenching reac-
tions, eqns. (4) and (5), deprotonation, eqn. (6), and neutrali-
zation, eqn. (7). Although reactions (6) and (7) have
probably the same chemical consequence, the charge neutrali-
zation process, eqn. (7), can be affected by varying the dose
rate, and hence the transient concentration. This is demon-
strated in Fig. 3 where the decay profiles of the radical cations
exhibit a marked dose rate dependence.
ð9Þ
ð10Þ
ꢀ
2ArO ! products
Fig. 4 shows a typical example of the fitting with reasonable
parameters taken from the electron transfer gross reaction and
those determined at favorable spectral points where one tran-
sient dominates. By adjusting the molar absorptivity with the
WINSPEC program, the relative yields for catechol radical
cation and its phenoxyl radical formed just after the pulse were
calculated to be about 40% and 60%, respectively.
For getting a more quantitative idea about the formation of
catechol radical cations subsequent electron transfer from tri-
phenylamine (TPA) to the primary formed catechol radical
cations was carried out, where TPA has a marked lower ioni-
ꢀ
þ
ꢀ
þ
HO Ar OH þ solvent ! HO Ar O þ H ðsolventÞ ð6Þ
ꢀ
þ
ꢁ
ꢀ
HO Ar OH þ Cl ! HO Ar O þ HCl
ð7Þ
19
20
zation potential and a known extinction coefficient of
ꢀ
+
4
3
ꢁ1
ꢁ1
TPA of e ¼ 2.9 ꢂ 10 dm mol cm . Fig. 5 shows the for-
Though catechol radical cations and their semiquinone radi-
cals are rapidly formed by electron transfer analogous to eqn.
mation of the TPA radical cation after electron pulse irradiat-
ꢁ3
ꢁ4
ing a solution containing 0.01 mol dm catechol and 5 ꢂ 10
(
2), there is also delayed generation of the semiquinone radi-
ꢁ
3
mol dm TPA in n-BuCl where we can assume complete
transformation, eqn. (11). Then the yield of TPA radical cation
obtained in a solution without catechol under conditions of
complete electron transfer, eqn. (12), was measured (Fig. 5,
upper spectrum). Since the experiments were done under iden-
tical conditions one can estimate the yield of catechol radical
cation formed by electron transfer to the solvent, eqn. (2a)
as well as the extinction coefficients of the different catechol
species, cf. Table 1.
cals according to eqns. (6) and (7), i.e. a delayed and time-
resolved transformation of the radical cations. Therefore, the
measured time profiles represent superimpositions of transients
generated by electron transfer, eqn. (2), with subsequent con-
version of the radical cations into radicals according to eqns.
(
6) and (7). This makes direct judging of the ratio of the tran-
sients very difficult, in particular because of the unknown
extinction coefficients of the transients. But on the bases of
the mentioned kinetic connection, fitting the time profiles
taken at different wavelengths enables to find the ratio between
the primary formed electron transfer products, i.e. a semiquan-
titative analysis of the reaction channels, eqns. (2a) and (2b).
This works with an internal calibration taking into account
that the primarily formed phenol radicals are directly con-
verted into phenoxyl radicals according to reaction (6). The
procedure is described in detail in ref. 4 and based formally
on the following differential equations:
HO Ar OHe þ þ TPA ! TPAꢀþ þ HO Ar OH
ꢀ
ð11Þ
ð12Þ
ꢀ
þ
ꢀþ
RX þ TPA ! TPA þ RX
The estimated yield of the radical cation was found to be
2% which agrees well with the value derived from the fit pro-
cedure mentioned above. Thus, one can conclude that the
remaining amount of the solvent parent radical cations via
eqn. (2b) leads to the formation of catechol phenoxyl radicals,
i.e. about 58%.
4
d½ArOH þꢃ=dt ¼ k2a½ArOHꢃ½RXꢀþꢃ ꢁ k6½ArOHꢀþ
ꢀ
ꢃ
ð8Þ
5
098
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104

View Article Online
ꢀ
ꢀ
HO Ar O þ n-C4H9 ! HO Ar O C4H9;
9
3
ꢁ1 ꢁ1
k14 ¼ 2 ꢂ 10 dm mol
s
ð14Þ
But in the presence of oxygen the butyl radicals are elimi-
nated by formation of less reactive peroxyl radicals (reaction
1
considerably prolonged. Applying the disproportionation rate
5) and, therefore, the lifetime of the semiquinone radicals is
1
1
given by Land, under our conditions, however, the semiqui-
none radical is oxidized by the butylperoxyl radicals according
to reaction eqn. (16). This is demonstrated by the time profile
given in Fig. 6d. The shown growth was found to be indepen-
dent on the dose per pulse which excludes reaction (13) and
could reasonably explain eqn. (16), taking into account a high
excess of the peroxyl radicals.
ꢀ
ꢀ
9
3
ꢁ1 ꢁ1
s
n-C4H9 þ O2 ! n-C4H9OO ; k15 ¼ 2 ꢂ 10 dm mol
ð15Þ
ꢀ
ꢀ
ꢁ
n-C4H9OO þ HO Ar O ! o-quinone þ n-C4H9OO ;
8
3
ꢁ1 ꢁ1
s
k16 ꢄ 10 dm mol
ð16Þ
By the way and as expected, electron transfer, eqn. (2), has
not been found to be influenced by oxygen.
Fig. 5 Calibration of the diphenol radical cation yield: Transient
absorption spectra taken 400 ns after pulsing samples consisting in
ꢁ2
ꢁ3
FET with resorcinol and hydroquinone
(
1
L) 10 mol dm TPA in n-BuCl (direct reaction (12)), and (S)
ꢁ2
ꢁ3
ꢁ4
ꢁ3
0
mol dm catechol and 5 ꢂ 10 mol dm TPA solved in n-BuCl
Fig. 7 shows transient absorption spectra and selected time
(stepwise electron transfer, reactions (2a) and (11)). The time profiles
2
profiles obtained in the pulse radiolysis of N -bubbled n-BuCl
show typical kinetics of the stepwise transfer measured with the latter
sample.
ꢁ
3
ꢁ3
solution containing 2 ꢂ 10 mol dm resorcinol. According
to the relatively low resorcinol concentration, the spectrum
taken immediately after the pulse is dominated by residual n-
Concerning the fate of the catechol semiquinone radicals,
the bimolecular and second order deactivation was found to
be much faster as known for aqueous systems. That can be
explained by a modified mechanism where instead of dispro-
ꢀ
+
BuCl absorption, cf. also time profile taken at 520 nm. After
its decay finished in 50 ns, however, the spectrum exhibits
marked absorption bands of resorcinol transients at l ¼ 405,
4
20 and 465 nm which were formed concomitantly as already
1
1
portionation (eqn. (13), as in aqueous solution ) a recombina-
tion with the radiolytically formed excess of butyl radicals
happens, as described by reaction (14). This is derived from
the different product characteristics: in our experiments no
o-benzoquinone (lmax ¼ 390 nm) was found, however, the visi-
ble bands of the semiquinone at 360 and 380 nm could be seen
described for catechol. Whereas the peak at 465 nm decreases
with time as typical for a radical cation, no significant decrease
of the bands absorbing at 405 and 425 nm could be observed at
identical time scale.
Although resorcinol is a meta-dihydroxybenzene where no
oxidation into a quinone is imaginable, the transient reaction
behavior was observed to be quite similar to that of catechol.
In the presence of TEA, the decay rate of the cation absorp-
tion band at 465 nm was found to increase according to an
electron transfer analogous to eqn. (5), exhibiting a rate con-
(
cf. Fig. 1), which in aqueous solution are covered by the o-
1
1
benzoquinone band. The time profiles given in Fig. 6 show
the radical transformation in the 200 ms time range: at 410 nm
spike absorption partially of the catechol radical cations, at 370
nm decay of the semiquinone radicals under product forma-
tion (background) and at 330 superimposed signal of radical
decay and product absorptions, both in equal part. This seems
to be consistent with a reaction sequence of eqns. (7) and (14).
10
3
ꢁ1 ꢁ1
stant of k
5
¼ 1.4 ꢂ 10 dm mol
s . From this experiment
we could easily get the absorption spectrum of the resorcinol
radical cation by subtracting the absorption spectrum taken
in the presence and absence of TEA. It can be seen from
Fig. 8 that the resorcinol radical cation has an absorption max-
imum at 460 nm while the corresponding phenoxyl radical
shows peaks at 400 and 425 nm which is similar to that
ꢀ
2
HO Ar O ! HO Ar OH þ o-quinone;
7
3
ꢁ1 ꢁ1
s
k13 ¼ 1:1 ꢂ 10 dm mol
ð13Þ
Table 1 Experimental data obtained for the FET from dihydroxybenzenes to solvent parent radical cations: QH
ꢀ
2
means dihydroxybenzene, e cali-
+
brated on that of TPA , ref. 20
Catechol
Resorcinol
Hydroquinone
ꢀ
+
4
3
ꢁ1
ꢁ1
4
3
ꢁ1
ꢁ1
ꢁ1
QH
QH
2
e
e
425 ¼ 1.2 ꢂ 10 dm mol cm
e
e
465 ¼ 1.7 ꢂ 10 dm mol cm
l
max ¼ 425 nm
ꢀ
3
ꢁ1
ꢁ1
3
ꢁ1
365 ¼ 6300, e375 ¼ 6700 dm mol cm
425 ¼ 8600 dm mol cm
l
max ¼ 400 nm
3
ꢁ1 ꢁ1
10
10 , in n-BuCl
10
10 , in n-BuCl
10
9
k
2
of FET/dm mol
Quenching rates: k
s
(DMSO),
5 ꢂ 10 , in DCE
10
10
10
4
1.5 ꢂ 10 2.0 ꢂ 10
1.0 ꢂ 10 1.4 ꢂ 10
Not determined because
of strong spectral
super-imposition
7.94 eV, 7.95 eV
3
(TEA)/dm mol
ꢁ1 ꢁ1
k
5
s
Ionization potentials from
8.18 eV, 8.10 eV
8.30 eV, 8.26 eV
17
the literature, our
calculations (HOMO, vacuum)
ꢀ
+
t of QH
2
/ns
ꢀ
700
40/60, 42/58
700
40/60
500
40/60
ꢀ
+
QH
2
/QH ratio via fit (8,9), via
TPA expression,eqn. (11)
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104
5099

View Article Online
Fig. 8 Transient spectra taken 550 ns after electron pulsing of a solu-
ꢁ
3
ꢁ3
tion of 5 ꢂ 10 mol dm resorcinol in n-BuCl, purged with nitrogen
ꢁ4
ꢁ3
in the absence (S) and in the presence of 5 ꢂ 10 mol dm TEA. (;)
represents the difference between these spectra indicating the shape of
the resorcinol radical cation spectrum.
Adding millimolar concentrations of DMSO, a rapid con-
version of the resorcinol cations into the corresponding phe-
noxyl according to the deprotonation (4) happens. Fig. 9
shows this effect on hand of time profiles taken at l ¼ 425
nm, i.e. within the maximum of the phenoxyl radical. In the
presence of DMSO a time-resolved growth to a higher ampli-
tude can be seen which enables us to estimate a rate constant
Fig. 6 Time profiles indicating the fate of the catechol transients in a
ꢁ3
ꢁ3
solution of 5 ꢂ 10 mol dm catechol in n-BuCl, N
2
purged: (a) dom-
inating radical cation, (b) decay of the semiquinone, (c) stable product,
where (b) and (c) are superimpositions of transient absorptions. (d)
Time profile indicating the conversion of the semiquinone radical in
the presence of oxygen, according to reaction (16). The sample con-
1
0
3
ꢁ1 ꢁ1
ꢁ3
ꢁ3
k4 around 10 dm mol s .
sisted of a solution of 2 ꢂ 10 mol dm catechol in n-BuCl, purged
with oxygen.
The neutralization of the resorcinol radical cations, eqn. (7),
as well as the electron transfer involving TPA, eqn. (11), pro-
ceed in the same manner as described for catechol and, there-
fore, are not documented here (cf. Table 1). Also the
deactivation of the phenoxyl radicals according to reaction
1
3,15
reported in water.
found to be about 700 ns, i.e. the species is more stable than
The lifetime of the radical cation was
those derived from phenols.
(
3
14) proceeds with a diffusion-controlled rate k of about
14
ꢂ 10 dm mol
9
3
ꢁ1 ꢁ1
s .
Because of the low solubility of hydroquinone in n-BuCl of
ꢁ3
ꢁ3
p10 mol dm , the electron transfer experiments with it
were carried out in 1,2-dichoroethane where in principle the
same processes should proceed. Fig. 10 shows transient
absorption spectra obtained in the pulse radiolysis of a deaer-
ꢁ
2
ꢁ3
ated solution of 10
mol dm hydroquinone in 1,2-di-
choloroethane. Although much more superimposed as in the
cases of the other biphenols, we can distinguish a hydroqui-
none radical cation (l ¼ 425 nm) and a phenoxyl radical
(
l ¼ 400 nm) absorption band. The kinetics of these species
2
Fig. 7 Transient spectra of the pulse radiolysis of a N -purged solu-
ꢁ3
ꢁ3
tion of 2 ꢂ 10 mol dm resorcinol in n-butylchloride taken (S)
immediately after the pulse and after (X) 50 and (c)550 ns. The time
profiles demonstrate the kinetic behavior near to the absorption max-
ima. Spectrum (S) contains a considerable contribution of residual sol-
vent radical cation.
Fig. 9 Time profiles taken at 425 nm of (a) N
2
purged solution of
ꢁ3
ꢁ3
ꢁ2
2 ꢂ 10 mol dm resorcinol in n-BuCl and (b) after adding 10
ꢁ
3
mol dm DMSO.
5
100
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104

View Article Online
ꢀ
þ
RX þ HO y Ar y OH
(
ꢀ
þ
planar state ! HOArOHe
!
þ RX
twisted state ! ðHOAr Oꢀ HÞ ! HOArO þ Hþ
þ
ꢀ
ð17Þ
(
y) symbolizes here the rotation of at least one C–y–OH
group around the indicated axis. So in the approach of the
reactants by diffusion, an adiabatic and immediate electron
jump happens over times comparable to or shorter than the
rotational frequency. As indicated in eqn. (17) and by analogy
with the monofunctional phenols, because of the rotation of
the C–OH bond at least two boundary cases could be defined:
one with the molecule in plane and the other one with a twisted
bond (90 degree out of plane). Now it was interesting to reason
this behavior by looking on the dynamics of the diphenol
molecule combined with the changes of the electron distribu-
tion in the three highest MO’s, to consider the stability of
the formed radical cations and to compare it with the mono-
functional phenols. Hence we performed quantum-chemical
calculations on these problems.
Calculations on the molecular dynamics of biphenols
Fig. 10 Transient spectra of the pulse radiolysis of a N
2
-purged solu-
Quantum chemical calculations were done using the Gaussian
98, Revision 11 program package. The equilibrium geome-
ꢁ2
ꢁ3
23
tion of 10 mol dm hydroquinone in 1,2-dichloroethane taken (S)
immediately after the pulse and after (X) 50, (;) 100 and (c) 550 ns.
The time profiles demonstrate the kinetic behavior near to the absorp-
tion maxima.
tries of the structures investigated were optimized without
any restriction. The atomic charges as well as the atomic spin
densities were calculated using the Mulliken population analy-
sis. The Hartree–Fock (HF) and hybrid density functional the-
2
4–26
ory (DFT) B3LYP
methods were used for the
follow the same rules as described for the other dihydroxyben-
zene isomers, i.e. that
investigation of the geometrical parameters and electronic
structures of the dihydroxybenzene ground state singlet and
cation radical. All methods show the similar influence of the
molecular geometry on electronic structure. The frequency cal-
culations were used to locate transition state geometries, to
determine the nature of stationary points found by geometry
optimization, and to obtain thermochemistry parameters such
as zero-point energy (ZPE), thermal correction to energy (TE)
(
i) the hydroquinone radical and radical cation were formed
synchronously by electron transfer, eqn. (2a) and (2b) in a
ratio of 60 : 40 calculated from the TPA reaction, eqn. (11),
and
(ii) the quenching experiments with DMSO, eqn. (4), and
TEA, eqn. (5), gave analogous results (cf. Table 1).
Although for the oxidation of hydroquinone in aqueous
solution not in each case it could be well distinguished between
and activation energy E (height of rotation barrier). The cal-
a
culated data for the frequency analysis and thermochemical
parameters are given in Tables 2. The B3LYP/6-31G(d)
method was used because DFT is superior to other methods
1
0,11,13,22
semiquinone radical and its anionic form,
state that the 400 nm species is really the radical form.
here we can
2
7
in reproducing the vibrational frequencies for phenols. Ea
were calculated as difference between E + ZPE + TE of the
0
stable and transition state (TS) structures of the molecules.
As for phenol the frequency calculations on the hydroqui-
none and resorcinol stable planar structures produce no nega-
tive frequencies indicating that this conformation is at an
energy minimum, whereas structures with one O–H group per-
pendicular to the ring plane (about 90 ) produce only one
negative (imaginary) frequency indicating that these conforma-
tions are first–order saddle points (transition states).
Discussion
As already reported for the ionization of various monofunc-
4
tional phenols by free electron transfer to solvent parent ions,
also for all of the bifunctional hydroxybenzenes such as cate-
chol, resorcinol and hydroquinone the product splitting into
radical cations and phenoxyl radical was observed. In varia-
tion of the introducingly formulated FET reaction, eqn. (2a)
and (2b), this is formulated for the biphenols in reaction
ꢅ
The calculations on the catechol resulted in some different
properties. In the stable planar structure there is a strong
(
17a) and (17b):
Table 2 B3LYP/6-31G(d) calculated thermochemistry parameters for stable and transition states. vmin is the minimal calculated frequency. vrot
and vval are rotation and valence vibrational frequencies of OH-group. E
for ordinary phenol are also given.
a
is the activation energy of the OH group rotation. For comparison data
ꢁ1
ꢁ1
ꢁ1
val/cm
ꢁ1
v
min/cm
Structure
v
rot/cm
v
E
0
+ ZPE + TE/a.u.
a
E /kcal mol
Phenol plane
ꢅ
+234
ꢁ351
+158
ꢁ299
+228
ꢁ368
+188
ꢁ205
Minimum
Transition state
Minimum
359
314
350
427
3751
3758
3753
3717
ꢁ307.3546
ꢁ307.3494
ꢁ382.5628
ꢁ382.5591
ꢁ382.5651
ꢁ382.5604
ꢁ382.5628
ꢁ382.5608
3.2
2.3
2.9
3.3
Ar–OH 90
1,4-dOH–Ar plane
Hydroquinone 90
ꢅ
Transition state
Minimum
1,3-dOH–Ar plane
Resorcinol 90
ꢅ
Transition state
Minimum
Transition state
1,2-dOH–Ar plane
Catechol 40.9
ꢅ
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104
5101

View Article Online
interaction between one oxygen atom and the hydrogen atom
of the neighbour OH group shortening the distance OHꢆ ꢆ ꢆO to
only 0.212 nm. This interaction influences the transition state
transfer occurs between them under the assumption of good
overlap.
As shown in Fig. 12 for the case of hydroquinone, in the pla-
nar structure all three doubly-occupied MO’s have the same p-
symmetry with strongly delocalized electrons. In contrast, if
one of the OH groups is perpendicular to the molecular plane,
only HOMO and HOMO-1 are delocalized whereas HOMO-2
exhibits n-symmetry and is strongly localized on the oxygen
atom. Similar dependence of the MO’s on C–OH group rota-
tion was found for the other biphenols.
ꢅ
geometry by OH twisting out of plane of about 41 .
As can be noted from Table 2, the valence O–H oscillations
are at least 10 times faster than the rotation of the C–OH
group, but both of them are postulated to be slower as the elec-
tron transfer jump. It should be mentioned that our data of the
phenol molecular oscillations are compatible with those given
2
8
by Qin and Wheeler.
Because of the small activation energies (barrier heights) E
a
Considering the MO scheme of a neutral closed shell biphe-
nol molecule (Fig. 12), the ground configuration (a) and the
excited configuration (b) of the corresponding radical cation
(Koopman’s configurations) can be achieved by a one step
for all studied molecules, internal rotation of OH-groups is
possible. But within the Born–Oppenheimer approximation
the molecular geometry is rigid compared to electron relaxa-
tion processes and very fast electron transfer events.
It is known that the molecular orbital (MO) description can
be used also for the interpretation of chemical phenomena.
Though, checking the effect of the molecular geometry on
the electronic structure, we investigated the behaviour of the
highest doubly-occupied MO’s (HOMO) of the dihydroxyben-
zene singlet ground state, which certainly are involved in the
electron transfer process, with respect to OH-group rotation
3
1
electron transfer. By a phenomenological comparison of
the planar and the perpendicular structures it can be seen that
only the long-lasting, stable and delocalized radical cation can
be derived from the planar structure whereas in the perpendi-
cular situation, because of the excess energy and the relatively
small ionization potential difference, electron transfer could
happen also from HOMO-2, i.e. from the localized structure.
Moreover, in the actual case because of the strong localization
of the HOMO-2 electron, the overlap (and the charge transfer
integral) between the MO’s of the reactants is about twice as
large as the overlap involving the HOMO and results in a
favourable interaction energy. Though the hypotheses of the
electron transfer from HOMO-2 combined with the formation
of a oxygen localized radical cation in the perpendicular struc-
ture seems to be justified.
(
measured by the torsion angle), cf. Fig. 11.
4
Following the same principles as phenol, the HOMO is sta-
bilized and drops down with an out-of-plane bond twisting,
the HOMO-1 is not influenced by bond twisting, and concomi-
tantly the HOMO-2 level rises due to the strong electron loca-
lization on the oxygen atom in the perpendicular structure.
Thus, if one of the both OH groups is perpendicular to the aro-
matic ring, the three highest doubly-occupied molecular orbi-
tals are closer to each other, i.e., the ionization potential
differences become smaller.
Mechanistic consequences
Electron transfer from donor (biphenol) to acceptor
29,30
+
BuCl ) can be treated in terms of chemical reactivity
ꢀ
(
as
Relating to these results, within a sufficiently short time scale
where the rotation of the critical bond C–OH is ‘‘frozen’’
(femtosecond range) we can assume two borderline conformer
structures, strongly differing in the tendency towards prompt
dissociation after ionization, corresponding to eqn. (17a) and
(17b). Whereas the heteroatom-localized radical cation (with
O–H vibration frequency) dissociates immediately to a phe-
noxyl radical, eqn. (17b), the delocalized planar radical
cation eqn. (17a) represents a species which is metastable on
the direct interaction between overlapping orbitals. Depending
on the energy difference (De) between the highest occupied
MO’s of the donor and the lowest empty orbital of the accep-
tor, the reaction can be either charge controlled or frontier
orbital controlled. In case of biphenols De is very small for
all three HOMO’s and, therefore, interactions between the
frontier orbitals are predominant. This is the case of a frontier
orbital controlled reaction and strongly coupled electron
Fig. 11 Transformation of HOMO, HOMO-1 and HOMO-2 of the hydroquinone singlet ground state (isocontour ¼ 0.08) with respect to the
deviation of the OH-groups out of the molecular plane as calculated with B3LYP/6-31G(d).
5
102
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104

View Article Online
Fig. 12 Molecular orbital scheme of the electronic configurations of the cation radicals formed by electron transfer from the planar (a) and per-
pendicular (b) hydroquinone singlet ground state molecule to the solvent radical cation.
a timescale of some hundred nanoseconds. Afterwards it
deprotonates in a delayed manner, eqn. (6).
A salient feature of the present study is the observation of
radical cations of the three dihydroxybenzene isomers. Just
for the ortho- and para-derivative these cations represent the
hitherto unobserved first products of the well known stepwise
one electron oxidation of diphenols to quinones, eqn. (18),
where the direct formation of the semiquinone radical has been
claimed till now. So these experiments in non-polar solvents
show that the self-deprotonation of the metastable delocalized
biphenol radical cations is a relatively delayed process. But
adding more polar substances, e.g., DMSO, we could demon-
strate diffusion-controlled deprotonation, eqn. (4), which cer-
tainly happens also in water and other protic solvents as a
very rapid and therefore covered process. Recently, this depro-
tonation behavior has been extensively studied for a variety of
Surprisingly, the metastable dihydroxybenzene cation is
more stable as found for the radical cations derived from most
of the monofunctional phenols. Table 3 and Fig. 13 show cor-
relations of the experimentally determined lifetimes of the phe-
nol radical cations with quantum chemical calculated
parameters such as atom spin density on oxygen S(O) and
the Mulliken charge difference on the OH group between the
singlet ground state and the cation radical. So for S(O) and
DQ(OH) reasonable plots were found exhibiting correlation
coefficients around 0.9. The greater the unpaired electron den-
sity on oxygen, the shorter the experimental lifetime of the
radical cation was found to be. And certainly, for the phenols
the lifetimes depend strongly on the electron-withdrawing or
donating properties of further substituents.
1
8
monofunctional phenols.
Concerning the probability of the appearance of products
according to eqn. (17), under the assumption that only the
unsymmetrical mono-twisted biphenol rotational situation
should lead to the unstable localized cation structure, the yield
of the rapidly formed phenoxyl radicals should be higher, as
for simple phenols, where a value of 50% has been found.
Indeed, in accordance with a simple probability estimation,
the experiments described in this paper give a general product
distribution ratio of 60% phenoxyl and 40% radical cation.
ð18Þ
Table 3 DFT B3LYP/6-31G(d)/SCRF ¼ Dipole(e ¼ 8) calculated
atomic spin density on oxygen S(O) and Mulliken charge differences
on OH group between singlet ground state and cation radical DQ(OH)
different substituted phenols in comparison with the experimentally
obtained lifetimes t/ns, of the phenol radical cations in BuCl.
Phenols
t/ns
S(O)
DQ(OH)
4
4
-CN–ArOH
-NO –ArOH
110
230
270
320
340
500
680
700
700
1160
1270
0.204
0.242
0.201
0.189
0.153
0.162
0.118
0.149
0.106
0.103
0.003
0.204
0.241
0.213
0.203
0.176
0.179
0.136
0.178
0.155
0.133
0.084
2
ArOH
4
4
4
4
2
3
4
3
-Cl–ArOH
-MeO–ArOH
-HO–ArOH
-NH –ArOH
2
Fig. 13 Correlation of the experimentally obtained lifetimes t/ns of
the different substituted phenol radical cations in BuCl with the quan-
tum chemical parameters: (a) atomic spin density on oxygen S(O) (b)
Mulliken charge difference on OH group between singlet ground state
and cation radical NQ(OH) as calculated at DFT B3LYP/6-31G(d)/
SCRF ¼ Dipole(e ¼ 8)
-HO–ArOH
-HO–ArOH
-NMe
-NMe
2
–ArOH
–ArOH
2
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104
5103

View Article Online
6
7
8
9
O. Brede, H. Orthner, V. Zubarev and R. Hermann, J. Phys.
Chem., 1996, 100, 7097.
R. Hermann, G. R. Dey, S. Naumov and O. Brede, Phys. Chem.
Chem. Phys., 2000, 2, 1213.
O. Brede, R. Hermann, S. Naumov and H. S. Mahal, Chem. Phys.
Lett., 2001, 350, 165.
S. Patai, The Chemistry of Quinoid Compounds, Wiley, N.Y., 1974,
vol. I and II.
Conclusion
By analogy with phenols, thiols and selenols, by free electron
transfer from the bifunctional phenols catechol, resorcinol
and hydroquinone to parent radical cations derived from
non-polar solvents (alkanes, alkyl chlorides), diphenol radical
cations as well as phenoxyl radicals in the ratio 2:3 are formed
as directly observable products. Although the detection time
was limited to q10 ns, we are stating that within the electron
transfer gross process eqn. (17), the electron jump proceeds in
times of p10 fs and, therefore, diffusion only determines the
time scale of transient observation in so far that reactants must
be brought together for reaction. The very rapid electron jump
proceeds in times faster than the molecular oscillations of the
electron donor molecule and, hence, identifies the molecular
situation of at least C–OH bond rotation.
10 G. E. Adams and B. D. Michael, Trans. Faraday Soc., 1967, 63,
171.
1
11
12
13
E. J. Land, J. Chem. Soc., Faraday Trans., 1993, 89, 803.
R. E. Huie and P. Neta, J. Phys. Chem., 1985, 89, 3918.
V. A. Roginsky, L. M. Pisarenko, W. Bors, C. Michael and M.
Saran, J. Chem. Soc., Faraday Trans., 1998, 94, 1835.
14 O. Brede, F. David and S. Steenken, J. Chem. Soc., Perkin Trans.
2, 1995, 23.
15
16
17
M. R. Ganapathi, R. Hermann, S. Naumov and O. Brede, Phys.
Chem. Chem. Phys., 2000, 2, 4947.
R. Mehnert, O. Brede and W. Naumann, Ber. Bunsen-Ges. Phys.
Chem., 1982, 86, 525.
M. Gerhards, C. Unteberg and S. Schumm, J. Chem. Phys., 1999,
111, 7965.
The formed biphenol radical cations are converted in the
00 ns time range into phenoxyl radicals, eqn. (18), but they
5
were found to be more stable than those of the monofunctional
phenols (see Tables 1 and 3). As related to the biologically and
electrochemically interesting stepwise oxidation of hydro-
quinone (catechol) into the corresponding quinones by
subsequent one-electron transfers, here we could observe
the primary biphenol radical cation directly which has
been ignored until now.
18 M. R. Ganapathi, S. Naumov, R. Hermann and O. Brede, Chem.
Phys. Lett., 2001, 337, 335.
1
9
S. L. Murov, I. Carmichael, G.L. Hugh, Handbook of Photochem-
istry, 2nd edition, Marcel Dekker, N.Y. 1993.
H. D. Burrows, D. Greatorex and T. J. Kemp, J. Chem. Phys.,
2
0
1
21 S. Steenken and P. Neta, J. Phys. Chem., 1982, 86, 3661.
966, 44, 2369.
22
23
24
K. B. Patel and R. L. Willson, J. Chem. Soc., Faraday Trans. 1,
973, 69, 814.
M.J. Frisch et al., Gaussian 98, Revision A9, Gaussian Inc., Pitts-
burgh PA, 1998.
A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
1
References
1
R. Mehnert, in Radical Ionic Systems – Properties in Condensed
Phases, ed. A. Lund and M. Shiotani, Kluwer, Dordrecht, 1991,
p. 231.
O. Brede, R. Mehnert and W. Naumann, Chem. Phys., 1987, 115,
279.
O. Brede, Res. Chem. Intermed., 2001, 27, 549.
O. Brede, R. Hermann, W. Naumann and S. Naumov, J. Phys.
Chem.A, 2002, 106, 1398.
O. Brede, M. R. Ganapathi, S. Naumov, W. Naumann and
R. Hermann, J. Phys. Chem. A, 2001, 105, 3757.
25 A. D. Becke, J. Chem. Phys., 1996, 104, 1040.
26 Ch. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1987, 37, 785.
27 W. Zierkiewicz, D. Michalska and T. Zeegers-Huykens, J. Chem.
Phys., 2000, 104, 11 685.
28 Y. Quin and R. A. Wheeler, J. Phys. Chem., 1996, 100, 10 554.
29 G. Klopman, J. Am. Chem. Soc., 1968, 90, 223.
30 L. Salem, J. Am. Chem. Soc., 1968, 90, 543.
31 T. Bally, in Radical Ionic Systems – Properties in Condensed
Phases, ed. A. Lund and M. Shiotani, Kluwer Dordrecht, 1991,
p. 4.
2
3
4
5
5
104
Phys. Chem. Chem. Phys., 2002, 4, 5096–5104