Photo- and radiation-chemical production of radical cations of methylbenzenes and benzyl alcohols and their reactivity in aqueous solution

Article abstract of DOI:10.1039/b109974d

Radical cations of methylated benzenes and benzyl alcohols were generated by photoionization and by reaction SO₄^·- in aqueous solution. The photoionization requires two 248 nm photons. The lifetimes with the oxidant SO4 and absorption spectra of the radical cations produced were determined by time-resolved conductance and optical detection, and the reaction products were measured by GC. As expected, the radical cation lifetimes increase strongly with increasing number of additional methyl groups, and so does the ratio of deprotonation from a methyl or hydroxymethyl group vs. addition (of water) to a ring position. In the case of toluene the radical cation appears to have a chemical lifetime τ of 10-100 ps ≤ τ ≤ 20 ns, i.e., longer than it takes for an ion pair to separate into the free (solvated) ions, and it reacts predominantly by addition of water to the ring rather than by deprotonation from the methyl group. A further observation is that, as compared to methoxylated analogues, the methylated benzyl alcohol radical cations are much more reactive, such that OH^-induction of side-chain fragmentation, as often required with methoxylated benzyl alcohol-type radical cations, is not necessary.

Full text of DOI:10.1039/b109974d

View Article Online / Journal Homepage / Table of Contents for this issue
Photo- and radiation-chemical production of radical cations of
methylbenzenes and benzyl alcohols and their reactivity in
aqueous solution
Claudia Russo-Caia and Steen Steenken
Max-Planck-Institut fu¨r Strahlenchemie, D-45413 Mu¨lheim, Germany
Received 1st November 2001, Accepted 7th January 2002
First published as an Advance Article on the web 14th March 2002
Radical cations of methylated benzenes and benzyl alcohols were generated by photoionization and by reaction
ꢀ
ꢁ
with the oxidant SO4 in aqueous solution. The photoionization requires two 248 nm photons. The lifetimes
and absorption spectra of the radical cations produced were determined by time-resolved conductance and
optical detection, and the reaction products were measured by GC. As expected, the radical cation lifetimes
increase strongly with increasing number of additional methyl groups, and so does the ratio of deprotonation
from a methyl or hydroxymethyl group vs. addition (of water) to a ring position. In the case of toluene the
radical cation appears to have a chemical lifetime t of 10–100 ps ꢂ t ꢂ 20 ns, i.e., longer than it takes for an ion
pair to separate into the free (solvated) ions, and it reacts predominantly by addition of water to the ring rather
than by deprotonation from the methyl group. A further observation is that, as compared to methoxylated
analogues, the methylated benzyl alcohol radical cations are much more reactive, such that OHꢁ-induction of
side-chain fragmentation, as often required with methoxylated benzyl alcohol-type radical cations, is not
necessary.
cations in aqueous solution, to determine their lifetimes and
characteristic reactions and to identify (some of) their pro-
ducts.
Introduction
Whereas information is available on radical cations of alkyl-
benzenes either neat or dissolved in organic solvents or low-
temperature glasses,1–3 rather little is known4–8 about the nat-
ure and reactivities of these species in aqueous solution. The
study of these molecules in water is, of course, difficult due
to their low solubility but it is of interest because it is possible
that in water specific ionic reactions may occur. These may be
of practical interest in connection with the removal of aromatic
impurities in drinking water by photochemical means.
It is probable that in polar solvents the radical cations decay
very rapidly (due to the typical unimolecular decay reactions
of radical cations such as side chain deprotonation (Scheme
1 (b) or C–C-fragmentation, Scheme 3 below).9 If the solvent
is a nucleophilic one such as water, addition to the ring yield-
ing a hydroxycyclohexadienyl radical and H+ (Scheme 1 path
(a))10 will further decrease the lifetime, e.g., in the case of
toluene, see Scheme 1.
Results and discussion
1. Methylbenzenes ArCH3
1.1 Time-resolved experimentsDirect photoionization. The
248 nm laser photolysis (20 ns pulse duration) of methylben-
zenes in aqueous solution in all cases led to the production
of the hydrated electron, eaqꢁ, as concluded from its character-
istic11 absorption spectrum with a maximum at ꢃ700 nm (for
an example, see Fig. 1) and by its reactivity12 with electron sca-
vengers such as O2 , N2O or alkyl chlorides. The production of
the electron, which turned out to require two photons (bipho-
tonic reaction, see Fig. 2) means that a radical cation is also
formed. As shown in the following, radical cations could in
fact be directly observed, but only in the cases of multiply
methylated benzenes. For toluene, the radical cation turned
out to be too short-lived (lifetime < 20 ns). For this reason,
product analysis studies were also performed (with the hope
of detecting the radical cation via its reaction products (see sec-
tion 1.2)).
The aim of this laser flash photolysis study is to rapidly gen-
erate methylated benzene and ring-methylated benzyl alcohol
(as representatives of simple, non-stabilized aromatics) radical
The time-resolved absorption spectra recorded on direct 248
nm photolysis of p-xylene (ꢄp-Xy) in deoxygenated aqueous
solution are shown in Fig. 1. At 0.3 ms after the pulse there
is a strong peak at 268 nm, shoulders at 290 and 440 nm
and a very strong and broad band with a maximum at ꢃ720
nm.
The broad band is due to the hydrated electron, eaqꢁ, as
shown by its reactivity with the electron traps 2-chloroethanol,
dichloromethane, O2 or N2O. The presence of eaqꢁ proves that
ionization has occurred. It is concluded that the radical cation
so formed is responsible for the absorption at 290 nm, while
the strong peak centered at 268 nm and the shoulder at 320
Scheme 1
1478
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485
This journal is # The Owner Societies 2002
DOI: 10.1039/b109974d

View Article Online
ciated with a 248 nm photon (5 eV) plus the free energy of
hydration of an organic cation and of the electron (3–3.5
eV)13 add up to only 8–8.5 eV. A monophotonic ionization
is therefore thermodynamically not possible. On the other
hand, p-xylene, whose IP is 8.44 V, is a limiting case, its mono-
photonic photoionization could be thermodynamically feasi-
ble. It turned out, however, that the ionization is biphotonic.
To obtain information about the lifetimes (via the rates of
decay) of methylbenzene radical cations in aqueous solution,
photolysis experiments with time-resolved conductance detec-
tion were carried out. As seen from Fig. 3(a)–(c), on photolysis
in acidic solution there is, after the pulse, a first-order increase
of ꢀconductance which is explained in terms of replacement of
Ar +–CH3 (produced in the photoionization) by the much
more mobile H+ ion produced by deprotonation from –CH3
or by nucleophilic addition of water (see Scheme 1, steps (a)
and (b)). In basic solution, after the pulse, the conductance
decreases (Fig. 3(d)) since the proton formed is neutralized
by the OHꢁ ion whose concentration is thereby reduced. This
‘‘neutralization’’ proves that the positive ion produced on
interaction with water is the proton.
Fig. 1 Absorption spectra observed on 248 nm photolysis (70 mJ per
pulse) of p-xylene (2 mM) in Ar-saturated H2O 0.3 ms ( ) after the
pulse. The spectrum observed 3 ms after the pulse (X) was recorded
in the presence of 0.15 M 2-chloroethanol as an eaqꢁ scavenger. In inset
(a) is shown the decay of the radical cation, in inset (b) the concomi-
tant formation of the 4-methylbenzyl radical.
As mentioned, the conductance increase traces were expo-
nential and the rate constants measured were the same, inde-
pendent of the method of production (photoionization or
ꢀ
reaction with SO4 ꢁ, see below) as those measured optically
for the decay of the radical cations at their lmax values (this
was possible only for the long-lived radical cations). The
deprotonation rates thus obtained for several methylbenzenes
are collected in Table 1.
nm are attributed to the 4-methylbenzyl radical, which is
formed from the radical cation by deprotonation from a
methyl group. At 3 ms after the pulse the radical cation absorp-
As seen from the data in Table 1, for benzene, toluene, m-
xylene and 1,3,5-trimethylbenzene the lifetime of the radical
cation is shorter than the laser pulse width of 20 ns,15 i.e.,
the rate constant for their decay is ꢆ5 ꢇ 107 sꢁ1. The rate con-
stant for the decay reaction in aqueous solution decreases by
more than three orders of magnitude as the number of methyl
groups goes from one to four. As the electron density increases
in this direction (as concluded from the decrease of ionization
potential)16 an increase in the stability of the radical cation and
a change17 in the reaction path are observed (see section 1.2).
In the case of benzene, whose radical cation has no side-chain
from which to deprotonate, addition of water occurs to pro-
duce the hydroxycyclohexadienyl radical (‘‘OH adduct’’),
which, under oxidizing conditions, leads to phenol as the
non-radical, final product.18
ꢁ
tion at 290 nm and the eaq have decayed leaving a ‘‘clean’’
spectrum of the 4-methylbenzyl radical (see Fig. 1), character-
ized by one strong absorption at 268 nm and two weaker peaks
at 310 and 322 nm.
To obtain information about the mechanism of 248 nm
photoionization of methylbenzenes in aqueous solution, the
dependence of the yield of eaqꢁ on the laser intensity was inves-
tigated by attenuating the laser light with neutral density fil-
ters. Fig. 2 shows the situation for toluene, used as a model.
As is evident, the yield of eaqꢁ, monitored via its absorbance
at 650 nm, increases quadratically with the energy per pulse,
indicating that the process is biphotonic, i.e., a first-formed
excited state (eqn. (1)) is ionized by interaction with a second
photon (eqn. (2)):
Column 4 of Table 1 contains a collection of rate constants
determined previously by pulse radiolysis. In these cases the
radical cations were produced from the OH-adducts (formed
hnð248 nmÞ
ꢅ
ArCH3
ArCH
ð1Þ
ð2Þ
!
3
hnð248 nmÞ
ꢀ
ꢅ
ArCH3
ArCH þ þ eaq
ꢁ
!
3
This result is in line with what is expected on the basis of the
gas-phase ionization potential, IP, of the substrate studied.
For example, the IP of toluene is 8.8 eV while the energy asso-
Fig. 3 Photolysis-induced change of conductance of deoxygenated
aqueous solutions at pH ¼ 4.5 of: (a) 5 mM toluene, (b) 2 mM p-
xylene and (c) 1 mM 1,2,4-trimethylbenzene, containing 1.2% of 2-
chloroethanol to scavenge eꢁaq . For 1,2,4-trimethylbenzene the change
of conductance at pH ¼ 10 is also shown (d).
Fig. 2 Dependence of the absorbance (optical density) due to the
hydrated electron (monitored at 650 nm) on the laser power following
248-nm flash photolysis of a deoxygenated aqueous solution of toluene
(saturated).
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485
1479

View Article Online
Table 1 Rate constants (sꢁ1 a
) for decay of methylbenzene radical cations in aqueous solution
Aromatic
Conductance detectionb c
Optical detectionc
From refs. 4, 5 (optical detection)
Benzene
Toluene
ꢆ5 ꢇ 107
ꢆ5 ꢇ 107
2 ꢇ 107
1.0 ꢇ 107
2.0 ꢇ 106
2.0 ꢇ 106
1.4 ꢇ 106
o-Xylene
m-Xylene
p-Xylene
ꢆ5 ꢇ 107
2.4 ꢇ 106
4.0 ꢇ 102 d
ꢆ5 ꢇ 107
4 ꢇ 106
1.8 ꢇ 106
(p-Methoxytoluene)
1,3,5-Trimethylbenzene
1,2,3-Trimethylbenzene
1,2,4-Trimethylbenzene
1,2,4,5-Tetramethylbenzene
1.5 ꢇ 106
1.5 ꢇ 106
2.0 ꢇ 105
2.7 ꢇ 104
3 ꢇ 105
2.8 ꢇ 105
4.4. ꢇ 104 e
4.1 ꢇ 104 e f
Accuracy: ꢈ10%. b At pH ¼ 4.5. c This work. d From ref. 14 e In water/acetonitrile 3 : 1, for solubility reasons. f Obtained by monitoring the
a
ꢀ
ꢁ
decay of the radical cation (formed by reaction with SO4 ) at its lmax
.
ꢀ
by addition of OH to the methylbenzenes) by H+-induced
O2) is tentatively assigned25 to triplet p-Xy formed by the small
portion of light (ꢃ9%) absorbed by p-Xy directly.
After 35 ms both radical cation absorptions have decayed
and are replaced by a new spectrum characterized by 4 peaks,
a strong one with lmax 268, and three less intense peaks at 298,
310 and 322 nm. These are attributed to the 4-methylbenzylra-
dical, formed by rapid deprotonation of the radical cation
(eqn. (5)):
dehydration (analogous to the reverse of reaction (a) in
Scheme 1), i.e., by two bimolecular reactions, whereby the rate
of the first one (addition of ꢀOH) depends on [methylbenzene],
typically a very low value even in saturated aqueous solution.
Thus, with the less long-lived radical cations, their formation is
likely to be rate-determining (unlike the photoionization
method where radical cation formation occurs in ꢂ20 ns). In
the latter cases, the rate constant for decay of the radical
cations will be underestimated, and this is probably the reason
why many of the values given by Sehested et al.4,5 are smaller
(often by more than of factor of 10, see Table 1) than the
values measured by us.
ꢀ
ꢀ
ArCH3 ꢁ! ArCH2 þ Hþ
ð5Þ
þ
To support this assignment, the p-methylbenzyl radical was
generated independently by H atom abstraction from p-Xy
ꢀ
ꢁ 26
It showed an
by the radiation-chemically produced O
.
As concluded from the product analysis data (see section
1.2), in the case of toluene, addition of water is still the most
important reaction of the radical cation, but deprotonation
from the methyl group (see Scheme 1, path (b)) also takes place
(6%). In strongly acidic solution ([H+] ꢆ 1 M), protonation of
the OH-adduct makes the hydration reaction partly reversible;
since the deprotonation from the methyl group is an irreversi-
ble reaction, under these conditions the dominant final (transi-
ent) product is the benzyl radical.
identical spectrum both at 268 nm and in the 290–320 nm ‘‘fin-
gerprint region’’.27 The effect of the introduction of oxygen
into the solution confirms these assignments: In the presence
of oxygen the benzyl radical bands decayed rapidly, whereas
the radical cation bands (lmax 290, 435 nm) were essentially
unaffected.
It was attempted to determine the bimolecular rate constants
for reaction of methyl-substituted benzenes with SO4 by
ꢀ
ꢁ
ꢀ
Oxidation with SO4 ꢁ. Due to its highly positive reduction
ꢀ
potential,19,20 the SO4 radical is a potent one-electron oxi-
dant, particularly with respect to aromatic compounds.21–24
Aqueous solutions containing potassium peroxydisulfate and
the substrates were photolyzed with 20 ns pulses of 248 nm
light. The concentrations were chosen such that the main part
of the ligꢀht was absorbed by S2O82ꢁ, which homolyzes to yield
ꢁ
ꢀ
ꢁ
ꢁ
two SO4 radicals (eqn. (3)). The SO4 radical then reacts
with the aromatic compound by electron transfer (eqn. (4))
yielding the corresponding radical cation.
hnð248 nmÞ
ꢀ
ꢁ
2ꢁ
S2O8
2SO
ð3Þ
ð4Þ
!
4
ꢀ
ꢀ
SO4 ꢁ þ Ar ꢁꢁꢁ! Ar þ þ SO4
2ꢁ
For p-xylene, the resulting time-resolved absorption spectra
are shown in Fig. 4.
At 50 ns after the pulse, there is negative DOD at l < 290
nm (due to the depletion of S2O8 and, to a lesser extent, of
2ꢁ
p-Xy) and a broad band (positive DOD) centered at 450 nm,
ꢀ
ꢁ
due to SO4 (eqn. (3)). The (weak) negative signal in the
region 550 to 750 nm is assigned to luminescence emission.
After 170 ns the intensity of the band at 450 nm has grown
and during this time there has developed a strong band with
lmax 290 nm; this is explained in terms of the formation of
the radical cation p-Xy via eqn. (4), this assignment being
in good agreement with the characterization of the p-Xy radi-
cal cation by Sehested et al.4,5 using the pulse-radiolysis tech-
nique. The shoulder at 350 nm (which can be scavenged by
Fig. 4 Absorption spectra recorded on 248 nm photolysis of 70 mM
K2S2O8 in the presence of 2 mM p-xylene in Ar-saturated H2O,
pH ¼ 3.6: 50 ns ( ), 175 ns (X), 1 ms ( ), 35 ms (˘) after the pulse.
Insets (a) and (b) show the build up of the radical cation at 290 nm
and 435 nm. In inset (c) is shown the decay of the radical cation at
435 nm. In inset (d) is shown the conductance increaseꢀ induced by
direct photolysis of 2 mM p-Xy at pH ¼ 4.5 by which Xy + is formed
whose decay leads to H+ (eqn. (5)).
ꢀ
+
1480
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485

View Article Online
Table 2 Products from one-electron oxidation of toluene
Products (10ꢁ7 M)
Conditionsa
Cresols (o : m : p) Benzylalcohol Bibenzyl Other products
% addition % deprotonation
b c
LFP (248) in H2O + K3Fe(CN)6
LFP (248) in H2Ob , no additive
5.8 (42:8:50)
3.8 (37:8:55)
—
0.4
0.5
0.2
0.2
0.01
0.04
0.6
Photoadductsd
94
87
—
94
6
Photoadductsd
13
100
6
b
LFP (248) in CH3CN
ꢀ
PhCHO PhCH2CH2CN
PhCHO (0.4)
(SO4 ꢁ + K3Fe(CN)6) (g-radiolysis)e 9.7 (43:7:50)
—
a
In all experiments a saturated solution of toluene (6 mM) was degassed with toluene-saturated Ar. b Solutions were irradiated with 1000 pulses
(20 ns, ꢃ150 mJ per pulse) of 248 nm laser light. c The concentration of the ferricyanide added was 0.25 mM; this value was chosen to ensure that
the absorbance of the substrate at 248 nm (OD(Tol)/cm ¼ 0.4) is higher than that of the oxidant (OD(Fe(CN)63ꢁ) ¼ 0.2). d These products were
also obtained from (steady state) irradiation with a medium pressure Hg lamp. e In the presence of 2 mM K2S2O8 (to scavenge eaqꢁ), 1 M t-BuOH
ꢀ
(to scavenge OH) and 5 mM K3Fe(CN)6 (to scavenge the OH-adducts), Ar-saturated solution pH ¼ 4.3.
measuring kobs , the rate of the optical density increase of the
radical cation, as a function of the concentration of added sub-
strate. However, only for the case of durene was the lifetime of
the radical cation long enough (t ¼ 23 ms) to observe its com-
plete build-up. Inꢀthis case the bimolecular rate constant for
phenols formed from the radical cation clearly depends on
its method of production.10 The latter phenomenon was
explained in terms of the different nature (and therefore life-
time) of the reacting speꢀcies, i.e., ion pair formation in the case
of the reaction with SO4 ꢁ and formation of the ‘‘naked’’ radi-
cal cation in the case of photoionization. The difference in the
behavior of benzyl alcohol and toluene can then be understood
by assuming that the radical cation of toluene is less reactive
than that of benzyl alcohol such that the former undergoes
nucleophilic addition of water when it is in the fully hydrated
state and not already when it is in the (earlier) ion pair state
(see also below). Since solvation times of organic ions in water
appear to be typically in the 10 to 25 ps range, (ref. 31 and
references therein) the chemical lifetime of the toluene radical
cation must be ꢆ this value.
ꢁ
reaction with SO4 (determined from the slope of the linear
plot of kobs vs. [durene]) is 4.6 ꢇ 109 Mꢁ1 sꢁ1
.
1.2 Product analysis results. To substantiate the mechanis-
tic conclusions drawn from the time-resolved experiments, an
attempt was made to identify the most important non-radical
products from the reactions of the radical cations, taking
toluene and p-xylene as examples. Saturated solutions of
toluene (OD/cm ꢃ 0.4 at 248 nm) were photolyzed with 150–
200 mJ pulses of 248 nm laser light. The solutions contained
K3Fe(CN)6 as a mild oxidant in order to convert any photo-
generated hydroxycyclohexadienyl radicals into the corre-
sponding phenols.18,28 The results are presented in Table 2.
As can be seen from the data in Table 2, the major products
found are the isomers of cresol ( > 90% of the total photoioni-
zation products). These are suggested to be formed from the
radical cation via addition of water to the ring followed by oxi-
dation of the isomeric hydroxycyclohexadienyl radicals (OH
adducts) by K3Fe(CN)6 (Scheme 2, path (a)).29 Small amounts
of benzyl alcohol, traces of bibenzyl and of benzaldehyde were
also found. These are suggested to derive from side-chain
deprotonation of the radical cation (path (b)) followed by oxi-
dation of the resulting benzyl radical30 and addition of water
to the (incipient) benzyl cation. On this basis, benzaldehyde
is a secondary product, originating from benzyl alcohol. A
particularly interesting piece of information is that the isomer
distribution of the cresols formed is independent of the way of
generation of the toluene radical cation (biphotonic ionization
In the 248 nm laser photolysis experiments, additional pro-
ducts were found which were also produced on photolysis of
toluene with the low-intensity light (254 nm) of a medium-pres-
sure mercury lamp. On the basis of their mass spectra,32 the
products are identified as isomers of the photoadduct of water
to electronically excited toluene, e.g., eqn. (6).33
ð6Þ
In line with this suggestion, these products were not found on
photolysis of toluene in pure CH3CN or on generating the
radical catꢀion by reaction with the radiation-chemically pro-
ꢁ
duced SO4 (see below).
To additionally test the ideas contained in Schemes 1 and 2,
the toluene radical cation was produced in a non-photochemi-
cal way, i.e. by reaction with radiation-chemically produced
ꢀ
vs. oxidation by SO4 ꢁ). This is in contrast to the situation with
benzyl alcohol, where the distribution of the hydroxymethyl-
Scheme 2
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485
1481

View Article Online
Table 3 Products from 248 nm photolysis of aqueous solutions of p-xylene
Products, (10ꢁ7 M)
2,5-Dimethyl 4-Methyl
phenol
Di-(p-tolyl)
Conditionsa
benzylalcohol Ethane
Other products
% Addition % Deprotonation
b
LFP (248) in H2O + K3Fe(CN)6
1.6
1.5
0.2
4-Methylbenzaldehyde (0.09) 44
Dimer (0.03)
56
LFP (248) in H2O
1.2
—
0.4
0.2
3.5
1.0
4-Methylbenzaldehyde (0.02) 14
86
LFP (248) in CH3CN
4-Methylbenzaldehyde (0.05)
Dimer (0.8)
—
100
a
In all experiments saturated solutions of p-Xy (2 mM) were degassed with p-Xy-saturated Ar and irradiated with 1000 pulses (20 ns, 150 mJ per
b
pulse) of 248 nm laser light. The concentration of ferricyanide was 0.25 mM; this value is such that the absorbance of the substrate at 248 nm
(OD(p-Xy)/cm ¼ 0.3) is higher than that of the oxidant.
ꢀ
ꢀ
SO4 ꢁ. On radiolysis of aqueous solutions, OH and eaq aꢀre
ꢁ
From the data in Table 4 it is evident that additional methyl
groups, especially if present in the para-position, increase the
lifetime of the benzyl alcohol-type radical cation. If compared
with the family of methylbenzenes (see Table 1), it is clear that
the hydroxymethyl group, CH2OH, is considerably less ‘‘stabi-
lizing’’ than the methyl group. This may be due to its weaker
2ꢁ
ꢁ
produced. The latter was reacted with S2O8 to form SO4
(eqn. (7)) which then oxidizes toluene to its radical cation
(eqn. (4)), while OHꢀ is scavenged by excess tBuOH yielding
the unreactive Me2C(OH)CH2 .
ꢀ
ꢀ
2ꢁ
ꢁ
eaqꢁ þ S2O8 ! SO42ꢁ þ SO4
ð7Þ
electron donation power (sCH OH ¼ 0.02) as compared to the
2
methyl group (s
¼ ꢁ0.14) or to the greater tendency for
K3Fe(CN)6 (5 mM) was added to the solutions under radi-
olysis in order to quantitatively oxidize the isomeric methylhy-
droxycyclohexadienyl radicals to the corresponding cresols.34
The distribution of products obtained from the g-radiolysis
experiment (bottom row of Table 2) is in accord with the
results of Eberhardt28 and is equal to the results obtained by
photolysis (top row). This means that ꢀthe species produced
CH3
Ca-deprotonation of the radical cation. Furthermore, it is evi-
dent that replacement of the methyl group by the methoxy
group leads to an enhancement of the lifetime of the radical
cation by up to 4–5 orders of magnitude.37 In addition to
the increase of electron density resulting from the methoxy
group the reason for this is the fact that deprotonation from
this group is not possible.
ꢁ
in the two cases (by reaction with SO4 and by biphotonic
ꢀ
Reaction with SO4 ꢁ. As in the case of the methylbenzenes, the
ionization) have similar properties. As already mentioned
above, this situation ꢀis in contrast to that with benzyl alcohol
attempt was made to produceꢀ radical cations from the benzyl
alcohols by reaction with SO4 ꢁ. For example, photolysis of a
deoxygenated solution of peroxydisulfate (see eqns. (3) and
(4)) in the presence of 4-methylphenethylalcohol resulted in
the absorption spectra shown in Fig. 5 (recorded 50, 150 and
350 ns and 1 ms after the pulse).
At 50 ns after the pulse, the broad peak at 450 nm is mainly
due to the radical cation. At 150 ns the spectrum shows strong
absorption in the region between 250 and 320 nm whereby two
peaks can be distinguished, one with lmax 300 nm, assigned to
where a PhCH2OHe +/SO4 ion pair as the reactive inter-
2ꢁ
mediate was postulated.10 The difference in behavior between
the two similar systems is suggested to be due to the longer life-
time (wiꢀth respect to reaction with the nucleophile water) of
+
PhCH3e
as compared to the (slightly) less electron-rich
ꢀ
+
.
PhCH2OHe
In the case of p-xylene (Table 3), 248 nm photolysis of aqu-
eous solutions afforded 2,5-dimethylphenol, 4-methylbenzylal-
cohol, and 1,2-di(p-tolyl)ethane. (A second dimer was found in
an even smaller amount. This dimer was identified by GC-
MS35 on the basis of the characteristic fragmentation pattern
of its molecular ion.)
Table 4 Decay rates of methylated benzyl alcohol and phenethyl
alcohol radical cations in aqueous solution. Comparison with the
methoxylated counterparts (values in brackets)
From a comparison with the data for toluene (Table 2), it is
evident that, in going from toluene to p-xylene, the deprotona-
tion reaction becomes more important. This is expected on sta-
tistical grounds (in p-xylene, there are two methyl groups).
Furthermore, due to electron donation by the second methyl,
the radical cation of xylene should be less electrophilic than
that of toluene. A third factor influencing the addition/elimi-
nation ( ¼ deprotonation) pattern is steric hindrance of the
water addition by the additional methyl group.
Alcohol
k/sꢁ1 a
Conductanceb
ꢆ5 ꢇ 107 d
ꢆ5 ꢇ 107
Opticalc
—
Benzyl alcohol
4-Methylbenzyl alcohol
4-Methoxybenzyl alcohol
2,5-Dimethylbenzyl alcohol
2,4,6-Trimethylbenzyl alcohol
2-Phenylethanol
—
—
(1.5 ꢇ 104)e
1.4 ꢇ 106
1.6 ꢇ 106
1.8 ꢇ 107
—
—
ꢆ5 ꢇ 107
2. Benzyl alcohols, ArCH2OH
2(4-Methylphenyl)ethanol
2(4-Methoxyphenyl)ethanol
Cumyl alcohol
4.9 ꢇ 106
3.8 ꢇ 106
2.1 Time-resolved experimentsDirect photoionization. The
experiments with ArCH2OH involved the same approach as
(5.2 ꢇ 102)f
ꢆ5 ꢇ 107
—
—
4-Methylcumyl alcohol
4-Methoxycumyl alcohol
1.8 ꢇ 107
employed with the methylbenzenes. In all cases, direct 248
ꢁ
(2.9 ꢇ 102)e
nm photolysis led to the formation of the broad band of eaq
,
in a biphotonic process, as in the case of the methylbenzenes.
The lifetimes of the radical cations were determined using
the dc-conductance technique. The results from the photoioni-
zation experiments are collected in Table 4, together with the
results obtained from the use of SO4 to produce the radical
cations.
a
b
Accuracy: ꢈ10%. Determined at pH ¼ 4.5 in the presence of 1%
(v/v) 2-chloroethanol (to scavenge eaqꢁ). The values in brackets are
c
from the corresponding p-methoxybenzyl alcohols. Obtained by
monꢉiꢁtoring the decay of the radical cation formed by reaction with
ꢀ
ꢁ
d
e
From ref. 10. From ref. 36 From ref. 14
f
SO4 at its lmax
.
1482
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485

View Article Online
Fig. 5 Time-resolved absorption spectra observed on photolyzing a
deoxygenated 70 mM solution of K2S2O8 in water in presence of 5
mM 4-MePhCH2CH2OH: 50 ns ( ); 150 ns (X); 350 ns ( ); 1 ms (˘)
after the pulse. Insets: (a) build up of the benzyl radical, (b) corre-
sponding decay of the radical cation at 450 nm and (c) the resulting
change of conductance at pH ¼ 4.5 in the presence of 1% 2-chlor-
oethanol.
Fig. 6 Absorption spectra recorded 0.5 ( ), 1.5 (X) and 14 ms (
)
after 248 nm flash photolysis of an Ar-saturated aqueous solution of
K2S2O8 (70 mM) and 2,5-dimethylbenzylalcohol (5 mM). Insets:
build-up of the benzyl radical (a), decay of the radical cation at 290
nm (b) and at 450 nm (c); and (d) the conductance change at
pH ¼ 4.5 following direct photolysis. In inset (e) is shown the spec-
trum of the 2,5-dimethyl-a-hydroxybenzylradical produced (see text)
by reaction of 2,5-dimethylbenzaldehyde (0.5 mM) with the radia-
ꢁ
a band of the radical cation, to which also the remaining
absorption at 450 nm belongs, and the other with lmax 268
nm due to the benzyl radicals formed from the radical cation
via Ca–Cb or C–H fragmentation (see Scheme 3). At 350 ns
after the pulse the absorptions due to the radical cation have
to a large extent disappeared while the absorbances at 268,
310 and 322 nm, assigned to the benzyl radical, have increased.
The assignments were confirmed by addition of oxygen to the
solution: In agreement with the general behavior of radical
cations, the bands assigned to the radical cation were not
affected by O2 , whereas the benzyl radicals decayed rapidly.
Another example is 2,5-dimethylbenzyl alcohol. In Fig. 6 is
tion-chemically produced eaq at pH ¼ 4.2 (t-BuOH 0.1 M).
cation via side-chain deprotonation (i.e., the 2,5-dimethyl-a-
hydroxybenzyl radical and the 2- or 3-hydroxymethyl-4-
methylbenzyl radical). Of these, an ‘‘authentic’’ spectrum of
the 2,5-dimethyl-a-hydroxybenzyl radical was obtained by
reacting eaqꢁ with 2,5-dimethylbenzaldehyde at pH ¼ 4.2 with
eꢁ (eqn. (8)), in a pulse radiolysis experiment.
aq
ð8Þ
ꢀ
shown the spectrum recorded on reaction with SO4 ꢁ. Evident
immediately after the pulse (at 0.5 ms) are the peaks of the radi-
cal cation at ꢃ450 and 280 with a shoulder at 330 nm. At 1.5
ms after the pulse the 450 and 280 nm peaks have disappeared
to give rise to a strong peak at 272 nm and a smaller one at 328
nm. At this time there is also a weak and broad absorption in
the region ꢃ380 nm (more visible 14 ms after the pulse). The
bands at 272 and 328 nm are assigned to a mixture of the three
different possible benzyl-type radicals formed from the radical
The spectrum, shown in Fig. 6, inset (e), is characterized by
three bands: a strong one with a lmax below 280 nm and two
less intense peaks at 330 and 404 nm. From the difference of
this spectrum from the one of Fig. 6 obtained 14 ms after the
pulse (lmax ¼ 278 and 320, 330 and 380 nm) it can be con-
cluded that the deprotonation of the radical cation does not
proceed in an exclusive way from the OH-substituted carbon
(giving the radical from eqn. (8)) but that there is also depro-
tonation from the methyl group(s) on the ring.
2.2 Photolysis products. To check the mechanistic conclu-
sions drawn in section 2.1, product analysis experiments were
carried out, with conditions similar to those used for the
‘‘toluenes’’. As a general procedure, 5 mM aqueous solutions
of the substrate containing 0.5 mM K3Fe(CN)6 (to oxidize the
OH-adduct(s) expected to be formed by hydration of the radi-
cal cations) were photolyzed with 1000 pulses (150–200 mJ per
pulse) of unfocussed 248 nm light and the resulting products,
after extraction of the aqueous solutions with ether, were ana-
lyzed by GC/MS.
In the case of 4-methylphenethylalcohol, 4-methylbenzylal-
cohol (1, see Scheme 3) was the major product, showing that
C–C fragmentation [reaction (a)] is an important process.
Two minor products were identified on the basis of their mass
spectra38 as 1,2-dihydroxy-(4-methylphenyl)ethane (2) and 2-
(4-hydroxymethylphenyl)ethanol (3).39 Together with the
time-resolved data reported above and the observation that a
phenol-type product was not formed, these results are inter-
preted in terms of Scheme 3.
Scheme 3
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485
1483

View Article Online
was added in a dropwise fashion (over a period of 30 min) a
solution of 3.3 g of acetone in 40 ml of THF. The basic mag-
nesium bromide was then decomposed by addition of 140 ml
of 40% NH4Cl in water. The solution was extracted with ether
and the combined ethereal samples were dried over anhydrous
Na2SO4 . The ether was evaporated and the crude product pur-
ified by elution through a SiO2 column with a petroleum ether/
ethyl acetate mixture. The yield was 5.3 g (76%); 1H NMR
(CDCl3) d 7.4–7.1 (m, 4 H, ring H), 2.3 (s, 3 H, p-CH3), 1.7
(s, 1 H, OH), 1.6 (s, 6 H, C(CH3)) ppm.
Laser flash photolysis
The laser flash photolysis (LFP) experiments were carried out
using a Lambda Physik EMG103MSC excimer laser which
produces 20 ns pulses (10–180 mJ per pulse) of 248 nm light
(KrF*). The optical absorption or conductance signals was
digitized with Tektronix 7612 and 7912 transient recorders
interfaced with a DEC LSI 11/73+ computer which also pro-
cess-controlled the apparatus and on-line pre-analyzed the
data. Final data analysis was performed on a Microvax I con-
nected to the LSI. The solutions were deoxygenated by bub-
bling with argon and flowed through a 2 mm (in the
direction of the laser beam) by 4 mm (in the direction of the
analyzing light) Suprasil quartz cell. Direct photoionization
was performed using saturated solutions41 of methylbenzenes
(OD/cm 0.1–0.4 at 248 nm) or 2–5 mM solutions of the benzyl
alcohols (OD/cmꢁꢃ 0.6–0.9 at 248 nm). For one-electron oxi-
Scheme 4
Compared to elimination [steps (a)–(c)], the hydration reac-
tion [analogous to step (a), Scheme 2] is apparently unimpor-
tant.
In the case of 4-methylcumyl alcohol the major products are
4-methylacetophenone (4) and 4-hydroxymethylcumyl alcohol
(5) (the latter identified only on the basis of its mass spec-
trum40) showing that, again, hydration of the radical cation
(addition) cannot compete with elimination). The results are
rationalized in Scheme 4.
Summary and conclusions
ꢀ
dation with SO4 the concentration of added K2S2O8 (typi-
cally 70 mM) was such that this component absorbed most
of the 248 nm light.
The lifetimes and, in part, absorption spectra and reaction pro-
ducts of the radical cations produced biphotonically from
methylated benzenes and benzyl alcohols in aqueous solution
were determined by time-resolved laser spectroscopy. As
expected, the lifetimes increase strongly with increasing num-
ber of additional methyl groups, the tendency for deprotona-
tion from a methyl or hydroxymethyl group compared to
addition (of water) to a ring position apparently increasing
in this direction. In the case of toluene the radical cation
appears to react as a free (solvated) ion, an observation which
is in contrast to that10 of benzyl alcohol where the reaction
period is of the order of ion solvation (10–100 ps). The conclu-
sion is that the lifetime t of the toluene radical cation is
between 10–100 ps and 20 ns. It would be interesting to com-
pare this lifetime with that of corresponding carbenium ions.
(With small radical cations, their reactivity has been shown
to be intrinsically cationic rather than ‘‘radical’’ (M. Mohr
and H. Zipse, Phys. Chem. Chem. Phys., 2001, 3, 1246).) A
further conclusion is that, as compared to methoxylated analo-
gues, the methylated benzyl alcohol radical cations are much
more reactive, such that OHꢁ-induction of side-chain fragmen-
tation, as often required with methoxylated benzyl alcohol-
type radical cations,37 is not necessary.
Product analysis
Saturated solutions of the methylbenzenes were deoxygenated
by bubbling with Ar (saturated with the aromatic) and photo-
lyzed with 1000 pulses (20 ns, 150–200 mJ per pulse) of 248 nm
laser light. The solutions contained K3Fe(CN)6 (0.25 mM,
OD/cm ¼ 0.2 at 248 nm) in order to oxidize any photo-gener-
ated hydroxycyclohexadienyl radicals into phenols. After irra-
diation an internal standard was added to the solutions (e.g.
for toluene, 0.1 mM 4-methylbenzaldehyde), they were
extracted with diethyl ether and, after concentration to a small
volume, analyzed by GC with flame ionization detection (HP
5890 gas chromatograph with a 30 m Stabilwax column, H2
as carrier gas, 100–250 ꢉC). The products were identified by
comparison with authentic compounds, except in the case of
some of the lower-yield products where GC-MS was used to
establish their structure. Blank experiments using non-exposed
samples were always carried out. The same procedure was
followed in the case of the benzyl alcohols (5 mM substrate,
0.5 mM K3Fe(CN)6), except for the GC where a HP 5890
gas chromatograph with a 15 m Stabilwax column was
used.
Experimental section
Solutions were also irradiated (2 h) with the light of a med-
ium-pressure mercury vapor lamp (Philips HPK 125 W/L) and
analyzed as above.
Materials
Radiolysis was carried out with a 60Co-g-source at a dose
rate of 0.5 Gy sꢁ1 on the following aqueous solution: 6 mM
toluene, 2 mM K2S2O8 , 1 M tBuOH and 0.5–5 mM
K3Fe(CN)6 . The irradiation time was chosen to obtain a
40% conversion of peroxydisulfate.
All chemicals of the purest grade available (Aldrich and
Merck) were used as supplied except for toluene, o-xylene,
m-xylene and p-xylene (p-Xy) which were purified by chroma-
tography over basic Al2O3 . The water used was from a MilliQ-
Millipore system.
p-Methylcumylalcohol
Pulse radiolysis
p-Methylcumylalcohol (2-(4-Methylphenyl)propan-2-ol) was
prepared via a Grignard reaction involving 10.3 g of p-bromo-
toluene and 1.4 g of magnesium turnings in 30 ml of anhydrous
tetrahydrofuran. To the resulting p-tolylmagnesium bromide
Pulse radiolysis (PR) experiments were performed with a 2.8
MeV van de Graaff electron accelerator delivering 300 ns
pulses42 with doses such that 0.5–3 mM radicals were produced.
1484
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485

View Article Online
25 For excited state spectra of o-xylene see:T. Sugawara and H. Iwa-
mura, ꢀChem. Phys. Lett., 1983, 101, 303.
Acknowledgement
26 The O ꢁ radical was produced from the OH radical by deprotona-
tion in an irradiated aqueous solution saturated with p-xylene and
N2O and containing 0.5 M KOH.
C.R.C. thanks the Max-Planck-Institut for their hospitality,
the Italian Ministry for University and Scientific Research
(MURST) for a fellowship, and Marion Stapper for her help-
ful technical assistance.
27 A similar, but less well-resolved spectrum was obtained on photo-
lysis of 4-methylbenzylchloride in 2,2,2,20,20,20-hexafluoroisopro-
panol.
28 M. K. Eberhardt, J. Org. Chem., 1977, 42, 832–835.
ꢀ
29 The reaction of SO4 ꢁ with toluene was also studied by Eberhardt
References
(ref. 28) in the presence of several oxidizing metal salts: cresols
arising from the conversion of the OH-adduct as well as products
formed by side-chain deprotonation of the radical cation (biben-
zyl, benzylalcohol and benzaldehyde) were observed. Several oxi-
dants were examined of which ferricyanide was the most efficient.
30 The rate constant for this reaction was determined by producing
the benzyl radical by photolysis of phenylacetone in the presence
of K3Fe(CN)6 and monitoring its rate of depletion to be of the
1
2
3
4
J. M. Masnovi, S. Sankararaman and J. K. Kochi, J. Am. Chem.
Soc., 1989, 111, 2263–2276.
T. DelGiacco, E. Baciocchi and S. Steenken, J. Phys. Chem., 1993,
97, 5451–5456.
R. M. Kadam, Y. Itagaki, R. Erickson and A. Lund, J. Phys.
Chem. A, 1999, 103, 1480–1486.
K. Sehested, J. Holcman and E. J. Hart, J. Phys. Chem., 1977, 81,
1363–1367.
order 109 Mꢁ1 sꢁ1
.
31 G. G. Gurzadyan and S. Steenken, Chem. Eur. J., 2001, 7, 1808–
1815.
5
6
K. Sehested and J. Holcman, J. Phys. Chem., 1978, 82, 651–653.
G. Merga, C. T. Aravindakumar, B. S. M. Rao, H. Mohan and J.
P. Mittal, J. Chem. Soc. Faraday Trans., 1994, 90, 597–604.
S. Steenken and R. A. McClelland, J. Am. Chem. Soc., 1989, 111,
4967–4973.
S. Steenken, C. J. Warren and B. C. Gilbert, J. Chem. Soc., Perkin
Trans. 2, 1990, 335–342.
M. Bietti and S. Steenken, in Handbook of Electron Transfer, ed.
V. Balzani, Wiley-VCH, Weinheim, 2001, vol. 2, pp. 497–579.
32 Three peaks with different retention times but with similar mass
spectra were found. For the most intense peak: M ¼ 110, m/z
(relative abundance): 65 (25), 77 (35), 91 (85), 05 (100), 109 (20).
33 This reaction is analogous to the well-studied (L. Kaplan, L. A.
Wendling and K. E. Wilzbach, J. Am. Chem. Soc., 1971, 93,
3821–3822; L. A. Kaplan, D. J. Rausch and K. E. Wilzbach, J.
Am. Chem. Soc., 1972, 94, 8638–8640) photooxidation of aqueous
benzene.
7
8
9
10 S. Steenken and R. Ramaraj, J. Chem. Soc., Perkin Trans. 2, 2001,
1613–1619.
11 G. L. Hug, in Optical Spectra of Nonmetallic Inorganic Transient
Species in Aqueous Solution, Natl. Stand. Ref. Data Ser., Nat.
Bur. Stand., Washington DC, USA, 1981.
12 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross, J.
Phys. Chem. Ref. Data, 1988, 17, 513–886.
13 M. Braun, J. Y. Fan, W. Fuss, K. L. Kompa, G. Muller and W. E.
¨
34 The second-order rate constant for this reaction was determined
in a LFP (248 nm) experiment, monitoring the rate of decay of
Fe(CN)6 at 420 nm, to be 2.0 ꢇ 108 Mꢁ1
s
ꢁ1. That this value
3ꢁ
is considerably higher than that given (V. Madhavan and R. H.
Schuler, Radiat. Phys. Chem., 1980, 16, 139–143) for the unsubsti-
tuted hydroxycyclohexadienyl radical is probably due to the
higher electron density in the methyl-containing radical.
35 M ¼ 210, m/z values (relative abundances): 195 ((100), 180 (29),
165 (26), 118 (63)) as (2,5-dimethyl(4-methylbenzyl)benzene)
and, finally, traces of 4-methylbenzaldehyde were found as pro-
ducts.
Schmid, in UV Laser Ionization Spectroscopy and Ion Photochem-
istry, ed. Z. Prior, A. Ben-Reuven and M. Rosenbluh, Plenum,
New York, London, 1986, pp. 367–378.
14 E. Baciocchi, M. Bietti, L. Manduchi and S. Steenken, J. Am.
Chem. Soc., 1999, 121, 6624–6629.
15 S. Steenken, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 113–124.
16 J. O. Howell, J. M. Goncalves, C. Amatore, L. Klasinc, R. M.
Wightman and J. K. Kochi, J. Am. Chem. Soc., 1984, 106,
3968–3976.
17 An additional aspect is that a further increase in the electron den-
sity of the aromatic compound, such as that resulting from the
introduction to toluene of a methyl group in para position (i.e.,
going to p-xylene) leads to less hydration of the radical cation:
As shown in the time-resolved spectra of p-xylene, the OH-adduct
is not optically detectable and the amount of phenol arising from
the OH-adduct drops to 44% of the total amount of the photoio-
nization products (see section 1.2)..
36 E. Baciocchi, M. Bietti, O. Lanzalunga and S. Steenken, J. Am.
Chem. Soc., 1998, 120, 11516–11517.
37 For a review of reactions of anisole-type radical cations, see E.
Baciocchi, M. Bietti and O. Lanzalunga, Acc. Chem. Res., 2000,
33, 243–251.
38 From the mass spectra: 1,2-dihydroxy(4-methylphenyl)ethane (2):
M ¼ 152, m/z values (relative abundance): 134 (16), 121 (100),
105 (7), 91 (28), 77 (27); 2-(4-hydroxymethylphenyl)ethanol (3):
M ¼ 152, m/z values (relative abundance): 121 (100), 107 (12),
91 (17), 77 (23); 2,3-bis(4-methylphenyl)propane-1-ol: M ¼ 240,
m/z values (relative abundance): 135 (43), 117 (7), 105 (100), 91
(4), 71 (10).
39 2,3-Bis(4-methylphenyl)propane-1-ol, 2,3-bis(4-methylphenyl)bu-
tane-1,2-diol and 4-methylbenzaldehyde were found in traces.
The former two are the dimerization products of the primary radi-
cals formed in steps (a) and (b), whereas 4-methylbenzaldehyde is
a secondary product, derived from 1.
18 V. Madhavan and R. H. Schuler, Radiat. Phys. Chem., 1980, 16,
139–143.
19 L. Eberson, Adv. Phys. Org. Chem., 1982, 18, 79.
20 L. Eberson, Electron Transfer Reactions in Organic Chemistry, ed.
K. Hafner, C. W. Rees, B. M. Trost, J.-M. Lehn, P. Von Rague
Schleyer and R. Zahradnik, Springer-Verlag, Berlin, 1987.
21 K. M. Bansal and R. W. Fessenden, Radiat. Res., 1978, 75, 497–
507.
22 O. P. Chawla and R. W. Fessenden, J. Phys. Chem., 1975, 79,
2693–2700.
23 P. Neta, V. Madhavan, H. Zemel and R. W. Fessenden, J. Am.
Chem. Soc., 1977, 99, 163–164.
40 From the mass spectrum of 4-hydroxymethylcumylalcohol:
M ¼ 166, m/z values (relative abundance): 151 (27), 148 (14),
133 (6), 107 (7), 91 (4), 77 (10).
41 For the solubility of methylbenzenes in neutral and alkaline solu-
tion see:K. Sehested, H. Corfitzen, H. C. Christensen and E. J.
Hart, J. Phys. Chem. Ref. Data, 1975, 4, 310.
42 V. Jagannadham and S. Steenken, J. Am. Chem. Soc., 1984, 106,
6542.
ꢀ
ꢁ
24 With aliphatic compounds, SO4 is able to react by H-abstrac-
tion (H. Eibenberger, S. Steenken, P. O’Neill, D. Schulte-Froh-
linde, J. Phys. Chem., 1978, 82, 749–750).
Phys. Chem. Chem. Phys., 2002, 4, 1478–1485
1485