Comparison of reactions of radical cations of 1-phenylalkanols produced by photoionization and by one-electron oxidation in aqueous solution

Article abstract of DOI:10.1039/b102515p

Benzyl alcohols in aqueous solution react with photo- and radiation-chemically produced ^.OH and SO₄^.- radicals with diffusion-controlled rates to yield OH-adducts and benzyl alcohol radical cations, respectively. The former can be converted to the radical cations by H⁺-induced (heterolytic) dehydroxylation, whereas the latter decay by a) electrophilic reaction with water (= reverse of the dehydroxylation reaction) giving rise to C_nucleus-OH-adducts and by b) side chain C-H deprotonation yielding α-hydroxybenzyl-type radicals. If, however, the radical cation is produced by biphotonic ionization of the benzyl alcohol, the pattern of C_nucleus-OH bond formation and side chain C-H bond breakage is different from that in the reaction with SO₄^.-. It is concluded that, at least in this reaction, it is not the free, solvated radical cation that reacts with water but the ion pair [radical cation-SO₄^2-].

Full text of DOI:10.1039/b102515p

View Article Online / Journal Homepage / Table of Contents for this issue
Comparison of reactions of radical cations of 1-phenylalkanols
produced by photoionization and by one-electron oxidation in
aqueous solution
2
Steen Steenken* and Ramasamy Ramaraj
Max-Planck-Institut für Strahlenchemie, D-45413, Mülheim, Germany
Received (in Cambridge, UK) 19th March 2001, Accepted 30th April 2001
First published as an Advance Article on the web 31st May 2001
Benzyl alcohols in aqueous solution react with photo- and radiation-chemically produced OH and SO4 Ϫ radicals
with diffusion-controlled rates to yield OH-adducts and benzyl alcohol radical cations, respectively. The former can
be converted to the radical cations by Hϩ-induced (heterolytic) dehydroxylation, whereas the latter decay by a)
electrophilic reaction with water (= reverse of the dehydroxylation reaction) giving rise to Cnucleus–OH-adducts and by
b) side chain C–H deprotonation yielding α-hydroxybenzyl-type radicals. If, however, the radical cation is produced
by biphotonic ionization of the benzyl alcohol, the pattern of Cnucleus–OH bond formation and side chain C–H bond
ؒ
ؒ
breakage is different from that in the reaction with SO4 Ϫ. It is concluded that, at least in this reaction, it is not the
ؒ
free, solvated radical cation that reacts with water but the ion pair [radical cation–SO42Ϫ].
method has great potential for studying highly reactive radical
Introduction
cations, since in practice their formation rate is limited only by
Radical cations of aromatics whose gas-phase ionization
the temporal width of the exciting light pulse. A particularly
potential (IP) is close to that of benzene (9.25 eV) tend to have a
interesting aspect is that under these conditions the radical
high degree of reactivity in solution, and even with nanosecond
cations may have a very special ‘counter-anion’, namely the
detection techniques it is typically not possible to observe these
(solvated) electron.
species directly. The nature of the one-electron oxidized species
So far the method of producing radical cations by biphotonic
therefore has to be deduced from their reaction products,
ionization, and the nature of the reaction products, have
and this sometimes introduces ambiguities. Due to the high
not been systematically studied. We felt this was, therefore,
reactivity of these radical cations, even their sufficiently rapid
necessary, and decided to take a look at benzyl alcohol and
formation becomes a problem. Obviously, in order to obtain a
some of its derivatives. This family of compounds has been
radical cation, the process of electron removal from the parent
thoroughly studied under one-electron oxidizing conditions
using product analysis13–15 and also the method of electron
must be shorter than the radical cation’s lifetime. This requires
powerful one-electron acceptors (oxidants) and polar solvents.1
paramagnetic resonance has been applied, and, as a result, the
Furthermore, radical cations of electron-poor systems are
nature of the radicals that exist on the ms time scale is well
established.16–21
usually strong electrophiles. On the other hand, polar solvents,
which are an essential prerequisite for electron transfer pro-
In the very polar and quite nucleophilic solvent water, radical
cesses to occur rapidly, are often nucleophilic and thus tend
cations of mono-substituted benzenes have been observed only
with the electron-rich systems anisole,22,23 thioanisole24 or N,N-
to destroy the radical cations by chemical reaction. A good
example of such a solvent is water. Nucleophilic attack of
a radical cation by the solvent can easily be faster than its
dimethylaniline.25 With toluene, however, only products of the
hypothetical radical cation were seen, i.e., the benzyl radical in
strongly acidic solutions26,27 (produced by deprotonation of the
formation, particularly if the latter is a bimolecular process
whose rate is of course determined by the concentration
radical cation from the methyl group) and (isomeric) methyl-
of substrate, which is usually smaller than that of solvent. A
hydroxycyclohexadienyl radicals (formed by hydroxylation by
water of the radical cation) in neutral solutions.28–31
(quasi)unimolecular method of producing a radical cation is
therefore desirable, and photoionization (PI) is, in principle,
a method to achieve this goal.
Photoionization has, in fact, been extensively used in the past
to produce one-electron oxidized species. With electron-rich
molecules such as, e.g., N,N,N Ј,N Ј-tetramethyl-p-phenylene-
On the basis of the Hammett σ-constant of hydroxymethyl
as compared to methyl, the radical cation of benzyl alcohol32
should be more difficult to produce and more reactive than that
of toluene. Nonetheless, benzyl alcohol has the advantage of
possessing good solubility in water, permitting experiments to
be performed in a more controlled way. We therefore decided
diamine, PI is typically a mono-photonic process with quantum
yields often >10%,2–4 and these types of radical cations5 are
to take a look at benzyl alcohol with the aim of studying its
usually long-lived. As the electron density of the molecules
radical cation.
is decreased, mono-photonic ionization tends to become less
efficient, if not thermodynamically impossible, unless the
Results and discussion
photon energy of the exciting light is very high (e.g., >200 kcal
molϪ1 (which corresponds to a photon of 143 nm) for molecules
with gas-phase IP of ca. 9 eV).
A way to photoionize molecules with a “normal” IP of
ca. 9 eV (such as benzene) without using vacuum UV light is
by biphotonic excitation,6,7 for instance with 248 nm quanta.8–12
If ionization is the dominant photochemical process, this
1. Pulse radiolysis and laserflash photolysis results
ؒ
a) The reaction of OH. This was investigated since it yields
᎐
the same hydroxycyclohexadienyl (᎐ HCHD) radicals (although
in a different proportion, see later) as those from water addition
to the radical cation. N2O-Saturated aqueous solutions at
᎐
DOI: 10.1039/b102515p
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619
This journal is © The Royal Society of Chemistry 2001
1613

View Article Online
pH ≈ 6 that contained 0.1–0.5 mM benzyl alcohol or its side-
chain substituted derivatives PhCR1(R2)OH (R1 = H, R2 = H,
Me, Et, CH2OH, CO2H; R1 = R2 = Me) were irradiated with
100 ns pulses from a 3 MeV electron accelerator. Irradiation
of water [eqn. (1)] produces OH-radicals and the hydrated
electron, eϪaq, which can be converted to an additional OH by
ؒ
reaction with N2O [eqn. (2)].
Ϫ
ؒ
H2O
OH ϩ e aq ϩ Hϩ
(1)
(2)
eϪaq ϩ N2O ϩ H2O → N2 ϩ OH ϩ OH
Ϫ
ؒ
In the presence of 0.1–2 mM of the substrates, the OH-
radicals produced by the pulse are scavenged in ≤2.5 µs to
mainly yield (by addition to the ring) isomeric HCHD radicals
and, to a minor extent (by H-abstraction from the benzylic
position), α-hydroxybenzyl radicals (see later). The observed
spectra (not shown) have intense bands with λ(max) ≈320 nm
and, except for the case of α,α-dimethylbenzyl alcohol, weaker
bands at ∼270 nm. The 320 nm bands are assigned to the
isomeric (hydroxymethyl)hydroxycyclohexadienyl radicals
Fig. 1 Absorption spectra of the transients [recorded at 250 ns (᭺),
1 µs (᭹), 1.5 µs (᭝), and 3 µs (ꢀ) after the 248 nm pulse] from the
reaction of OH (generated by photolysis of 0.1 M H2O2) with 1 mM
ؒ
ؒ
(“OH adducts”) formed by addition of OH to the ring, and
the 270 nm band to the α-hydroxybenzyl radicals produced
by H-abstraction from the CH2OH group (see Scheme 1 and
benzyl alcohol in 1 M HClO4. In the left inset is shown the increase
ؒ
of absorption at 268 nm due to formation of PhC (Me)OH from the
“OH-adduct” in a 2 mM solution of 1-phenylethanol containing 0.2 M
H2O2 and 2 M HClO4. From the inset to the right it is evident that the
ؒ
rate of formation of PhC (Me)OH from the “OH-adduct” increases
with increasing acidity of the aqueous solution.
the 248 nm light is predominantly absorbed by the H2O2 which
undergoes homolysis yielding two OH radicals [eqn. (3)] which
then react with the organic substrate.
(3)
ؒ
The advantage of this method is that the yield of OH is pH-
independent below pH ≈ 3, which is not the case with radiation-
ؒ
chemically produced OH in N2O-saturated aqueous solutions,
ؒ
where the OH yield decreases due to competition between
ϩ
N2O and Hϩ for eϪ (eϪaq ϩ H → H ). As a result of this
ؒ
aq
reaction, in the radiation-chemical experiment, below pH ≈ 3
ؒ
ؒ
ؒ
the yield of H becomes comparable to that of OH. Like OH,
H-atoms react with aromatics by addition and the so-formed
cyclohexadienyl radicals have very similar absorption spectra
to those of the OH adducts.37 However, the H adducts cannot
be converted to α-hydroxybenzyl radicals by Hϩ-induced
dehydroxylation and deprotonation [see eqn. (4)]. Therefore,
in order to study, in the presence of large Hϩ concentrations,
the conversion of OH adducts without the interference from
Scheme 1
ؒ
H adducts it is necessary to produce OH without the inter-
section 2). The 320 nm assignment is supported by the obser-
mediacy of eϪ
.
ؒ
vation of similar spectra on reaction of OH with other simple
aq
benzene derivatives including benzene itself,33–37 and the identi-
fication of the 270 nm band by its similarity with the spectra
observed on photoreduction of benzaldehyde in propan-2-ol,
on photocleavage of benzoin, on oxidative decarboxylation of
mandelic acid, or on Hϩ induced dehydroxylation of the OH
adducts (see later). The observation that the 270 nm band
is absent in the case of α,α-dimethylbenzyl alcohol (there is
a minimum at 270 nm) is in agreement with this picture, since
with this compound a benzyl radical can not be formed by
H-abstraction.
In Fig. 1 are shown the time-dependent absorption spectra
measured on reacting 1 mM benzyl alcohol with OH, pro-
duced by photolysis of 0.1 M H2O2, in the presence of 1 M
HClO4.
ؒ
The spectrum recorded at 250 ns after the 20 ns, 248 nm
pulse (᭺) shows a peak at 320 nm, due to the OH adducts, and
a larger one at 270 nm, from the α-hydroxybenzyl radical.38 In
the acidic solution, after about 3 µs the α-hydroxybenzyl radical
peak at 270 nm has considerably increased, at the expense of
the OH adducts absorbing at 320 nm. There is a well-defined
isosbestic point at 285 nm, indicating that the conversion of the
two species is a 1:1 process. From the left inset (showing the
case of 1-phenylethanol) it is evident that the increase of optical
density (OD) at 268 nm, reflecting this conversion, follows
exponential kinetics.
The conversion of the OH adducts to the α-hydroxybenzyl
radicals is explained by Hϩ-induced heterolytic dehydroxylation
with subsequent (or synchronous) deprotonation (or decarb-
oxylation, C–C fragmentation)39 of the (incipient) radical
cation, e.g., in the case of deprotonation, eqn. (4):
ؒ
The rate constants for the reaction of OH with benzyl
alcohol and its derivatives were measured by monitoring the
rate of buildup of the optical density (OD) at 320 nm as a
function of the concentrations of the substrates in the range
0.05–0.5 mM. The second order rate constants so obtained are
listed in Table 1. The values are all (4–6) × 109 MϪ1 Ϫ1, which
s
means the reactions are essentially diffusion controlled.
The transients described above could also be produced by
248 nm photolysis of H2O2 (e.g., 0.1 M) in the presence of,
e.g., 0.1–1 mM of the benzyl alcohols. Under these conditions,
1614
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619

View Article Online
Ϫ
Table 1 Rate constants k OH and kSO Ϫ for reaction of OH and SO with 1-phenylalkanols in aqueous solution, a 20 ЊC
ؒ
ؒ
ؒ
4
ؒ
4
Ϫ9
1-Phenylalkanol
λobs b/nm
10Ϫ9k OH/MϪ1 sϪ1
λobs b/nm
10
k
Ϫ/MϪ1 sϪ1
SO
ؒ
ؒ
4
Benzyl alcohol
320
320
315
320
320
320
320
6.4
6.4
5.7
4.3
4.5
6.4
4.4
270
270
270
320
270
270
270
3.2
3.3
3.1
2.1
2.6
3.1
1.5
1-Phenylethanol
1-Phenylpropanol
2-Phenylpropan-2-ol
Phenylethyleneglycol
Benzyl methyl ether
Mandelic acid
a At pH 5–7. b Wavelength of observation.
Table 2 Sensitivities for Hϩ-induced conversion of the OH-adducts of
Cα-substituted benzyl alcohols into α-hydroxybenzyl-type radicals a
(4)
Benzyl alcohol
Sensitivity (log kobs/H0)
PhCH(OH)CO2H
PhCH2OH
PhCH(OH)CH2OH
PhCH(Me)OH
PhCH(Et)OH
0.41
0.53
0.54
0.58
0.58
a 1 mM Benzyl alcohol, 0.1 M H2O2, 0–5 M HClO4, 248 nm photolysis,
20 ЊC.
the ether (saturated). The O Ϫ radical is known41 to react by H-
ؒ
abstraction from the side chain, and not by addition to the ring.
The resulting spectrum [Fig. 2 (᭝)] is almost identical to that
(᭹) observed on reaction with SO4 Ϫ [(see section b), below] or
ؒ
ϩ
ؒ
on OH addition/H -induced dehydroxylation in 1 M HClO4
(not shown) and is thus identified in terms of the α-methoxy-
benzyl radical. The spectrum of this species has a shoulder at
310–320 nm, the region where the OH adducts also absorb. In
ؒ
this region, other benzyl radicals, including PhCH2 , also have
bands which are, however, considerably weaker than their main
bands at 260–270 nm.
؊
Ϫ
ؒ
ؒ
b) The reaction of SO4 . The radical SO4 was produced
radiation-chemically, eqn. (5), and by photolysis, eqn. (6), in a
manner analogous to OH generation:
Fig. 2 Absorption spectra recorded at 20 µs on reaction of 1 mM
ؒ
Ϫ
ؒ
ؒ
ؒ
PhCH2OMe with OH (ᮀ), SO4 (᭹), pH 6, or O Ϫ(᭝), 0.5 M KOH.
The extinction coefficients are based on G( OH) = 6, G(SO4 Ϫ) = 3, and
ؒ
ؒ
G(O Ϫ) = 6.6.
ؒ
(5)
hv
2Ϫ
Ϫ
ؒ
In all cases (R = H, Me, Et, CO2H, CH2OH), the rate of
S2O8
2 SO4
(6)
ؒ
PhC ROH formation increased monotonously with increasing
Hϩ activity up to 5 M HClO4, where it was ∼5 × 107 sϪ1. The
rate constant for C–H deprotonation of the intermediately
produced radical cations, kdeprot, must therefore be greater than
this value. The sensitivity towards Hϩ, as measured by the slope
In the photochemical experiments the concentrations of
the organic substrates were high enough to enable SO4 to be
rapidly scavenged, but low enough to ensure that >;90% of the
248 nm light was absorbed by the S2O8 ion. In Fig. 3 are
shown the spectra measured after completion of the reaction
of SO4 with benzyl alcohol, benzyl methyl ether, and α,α-
dimethylbenzyl alcohol at pH 6–7. In the first two cases, there
are pronounced peaks at 260–270 nm and quite weak bands (or
shoulders) at 320 nm. In the last case only a band at 320 nm
is visible, whereas at 270 nm there is a minimum. As pointed out
in the previous paragraph, the 270 nm band is due to α-
hydroxy- or methoxy-benzyl radical and the 320 nm absorption
to OH adducts. With the exception of α,α-dimethylbenzyl alco-
Ϫ
ؒ
2Ϫ
ؒ
of the log k(formation of PhC (OH)R) vs. H0 dependence,
Ϫ
ؒ
increases in the series R = CO2H, H, CH2OH, CH3, CH2CH3,
i.e. with increasing electron donating power of R, as shown in
Table 2.40
Benzyl methyl ether. The OH radical (produced radiation-
or photo-chemically) was also reacted with benzyl methyl ether,
for comparison with the behavior of the benzyl alcohols. The
spectrum observed at pH 6 (see Fig. 2) consists of two peaks:
one at 320 nm and a second one, which is only slightly less
intense, at 270 nm.
hol (DMBA), the amplitude of the 320 nm peak can be reduced
Ϫ
by reacting SO4 with the substrates in the presence of Hϩ.
ؒ
The former is assigned to the isomeric (methoxymethyl)-
hydroxycyclohexadienyl radicals formed by addition of OH to
the ring, the latter to the α-methoxybenzyl radical, produced
by H-abstraction from the benzylic carbon. This assignment is
corroborated by the changes that occur on reacting the OH-
adduct radicals with Hϩ: as in the case of benzyl alcohol, at 1 M
HClO4 the peak at 320 nm disappears rapidly, giving rise to
more of the 270 nm species. That this is the α-methoxybenzyl
The reduction in the OD at 320 nm led to an increase in that
at 270 nm, indicating that the reaction of the OH adducts
with Hϩ gives rise to benzyl radicals, obviously by heterolytic
dehydroxylation, eqn. (4). At, e.g., 0.1 M Hϩ the conversion of
the OH adducts took place in a few µs.
In the case of DMBA, the 320 nm band was found to
disappear in an Hϩ catalyzed process without giving rise to a
product band. This observation is in agreement with the con-
version of the aromatic OH adduct radical to the aliphatic
radical was shown also by reacting O Ϫ, produced by pulse
ؒ
radiolysis of an aqueous solution containing 0.5 M KOH, with
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619
1615

View Article Online
Ϫ
ؒ
Fig. 3 Absorption spectra recorded at 10 µs after production of SO4
Fig. 4 Absorption spectrum measured on 248 nm photolysis of a
1 mM aqueous solution of benzyl alcohol, pH 6, ᮀ. For comparison
are shown the spectra obtained on addition of 0.2 M H2O2 and 3 M
ClCH2CH2OH and 0.5 M HClO4 (ꢀ) or on addition of 0.2 M
(NH4)2S2O8 (᭺). All spectra have been normalized to the peak at
270 nm.
[via eqn. (5)] in the presence of 1 mM benzyl alcohol (᭹), benzyl
alcohol methyl ether (ᮀ), or α,α-dimethylbenzyl alcohol (᭝) at pH 6–7.
The solution contained 2 mM K2S2O8 and 0.1 M tert-butyl alcohol. To
obtain the extinction coefficients, G(radical) was assumed to be equal to
G(SO4 Ϫ) = 3.0.
ؒ
methyl radical, a proton, and acetophenone, [eqn. (7)], as con-
cluded from EPR21 and product13 measurements.
(7)
c) Direct 248 nm photolysis. The substrates were also
photolyzed directly, i.e., with 248 nm light and in the absence
of H2O2 or S2O82Ϫ. Photolysis of, e.g., benzyl alcohol at pH 6–7
42
led to the production of eϪ
,
aq
as evident (monitored) in the
wavelength region 600 to 750 nm (see Fig. 4, squares. All the
spectra are normalized to the peak at 260–270 nm). In addition
Fig. 5 Dependence of the yield of eϪ (as monitored at 650 nm) on
the laser (λ = 248 nm) pulse power inaqa 1 mM aqueous solution of
benzyl alcohol at pH 6.
to the broad band of eϪ with a maximum at ∼720 nm, there
aq
is a shoulder at ∼310 nm, which is assigned to the isomeric
(hydroxymethyl)hydroxycyclohexadienyl
radicals
(“OH-
adducts”), and a sharp peak at 260 nm, which is mainly due
to the α-hydroxybenzyl radical (see later). On scavenging the
electron with 3 M 2-chloroethanol and adding 0.5 M HClO4,
the shoulder at 310 nm becomes considerably smaller, and
the 260 nm band (spectrum with full circles) is increased. This
was directly photolyzed in solutions additionally containing
1–2 mM Fe(CN)6 (to oxidize hydroxycyclohexadienyl and
benzyl-type radicals formed as a result of the photolysis).
Under these conditions, the percentage of the 248 nm light
absorbed by the benzyl alcohol was >95%. These results are
presented in Table 3.
3Ϫ
Ϫ
ؒ
spectrum is very similar to that recorded on reaction of SO4
with benzyl alcohol (open circles). That the shoulder at 310 nm
is due to the OH-adducts is evident from comparison with the
ؒ
a) The reaction of OH with benzyl alcohol. The results are
ؒ
spectrum recorded on reaction of OH with benzyl alcohol
ؒ
explained by reactions (1) and (2) (production of OH) and
(triangles).
The yield of eϪaq increased with the intensity of the laser light
(measured as mJ per pulse) in a quadratic fashion (see Fig. 5),
which demonstrates the involvement of two light quanta in the
ionization process.
Scheme 1 (formation and oxidation of the organic radicals).
Among the products, the hydroxybenzyl alcohols account for
about 85% of the initiating OH radicals (Scheme 1, reactions
b,c,d). Phenol is produced with a yield of 4%, and its precursor
ؒ
is suggested to be the radical formed by addition of OH at
the ipso-position (reaction a).43 In order to give phenol, the
carbocation formed by one-electron oxidation must undergo
C–C fragmentation with formaldehyde as the other product.
Attempts to detect H2CO were, however, unsuccessful in the
2. Product analysis results
Aqueous solutions (N2O saturated) containing 10 mM benzyl
alcohol and 1–8 mM K3Fe(CN)6 were 60Co-γ-irradiated with
doses such that conversion was kept below 1%. GC with FID
was used for qualitative, and HPLC with optical and electro-
chemical detection for quantitative product analysis. o-, m-,
and p-Hydroxybenzyl alcohol, phenol, and benzaldehyde were
found as reaction products (see Table 3 and Scheme 1). Similar
3Ϫ
presence of Fe(CN)6 (probably because H2CO is oxidized by
ferricyanide), and in its absence phenol was not formed.
ؒ
Thus, the dominant reaction of OH is addition to the phenyl
ring (reactions a–d, Scheme 1), in agreement with previous
studies.33–36 The so-formed isomeric hydroxycyclohexadienyl
radicals are then converted to the corresponding phenols by
(formal) one-electron oxidation by Fe(CN)63Ϫ and C–H depro-
2Ϫ
γ-radiolysis experiments were performed with S2O8 as a
source of SO4 Ϫ radicals [eqn. (5)]. Furthermore, benzyl alcohol
ؒ
1616
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619

View Article Online
Table 3 Yields of products formed in the reaction of benzyl alcohol with OH and SO4 Ϫ. Comparison with products from the 248 nm photolysis
ؒ
ؒ
Yield of product a
2-Hydroxybenzyl
alcohol b
3-Hydroxybenzyl
alcohol b
4-Hydroxybenzyl
alcohol b
Total
Initiating radical
Benzaldehyde c
Phenol d
yield e
ؒ
᎐
OH
32.6 (0.74)
16.9 (0.49)
7 f, g (0.07)
26.7 (0.61)
18.4 (0.53)
8 f, g (0.08)
22.0 (᎐ 1)
14.6
42.3
31g
3.7
≤3.4
99.6
≤98.3
᎐
Ϫ
ؒ
᎐
SO4
17.3 (᎐ 1)
᎐
53 f, g (᎐ 1)
᎐
248 nm Photolysis
᎐
a
Ϫ
ؒ
ؒ
ؒ
ؒ
In percent of the OH or SO4 initially produced. G( OH) was taken to be 6 for the case of N2O saturated solutions and G(SO4 ) = 2.6 for the argon
saturated solutions which contained 1 mM K2S2O8 and 50 mM tert-butyl alcohol. The error in the yields is estimated to be 10%. The pH was 3–8 for
the case of OH and 6 for the case of SO4 Ϫ. The numbers in parentheses are the relative yields per position on the benzene ring of benzyl alcohol.
b Determined by optical and electrochemical detection. c Determined by optical detection. d Determined by electrochemical detection. e Relative to
the depletion of the parent. f In the case of 248 nm photolysis, the concentration of products depends on the number of pulses (typically 1000 pulses)
and the dose per pulse (typically 150 mJ per pulse). Thus, relevant here is only the ratio of the products. Signals from products other than those
identified were ≥10 times weaker. g Molar ratio of products when conversion was ≤1%.
ؒ
ؒ
tonation of the resulting carbocation, a reaction type that is
well documented.36,44–46
Corrected for the number of available positions, there is a
ؒ
clear preference for the addition of OH at the para-position,
and even the ortho is favored over the meta-position. On this
ؒ
basis, with respect to the electrophilic OH, the HOCH2
substituent apparently behaves as a weakly electron-donating
group.
The sum of the yields of phenols is 85%. The remaining 15%
of the OH radicals react by H-abstraction from the benzylic
carbon, and the resulting α-hydroxybenzyl radical is oxidized
by Fe(CN)63Ϫ to give benzaldehyde (Scheme 1, reaction e). This
relatively high percentage of H-abstraction as compared to the
toluene case (H-abstraction <10%)27,31 is in agreement with the
analogous observations made on the basis of EPR measure-
ments.21 It indicates activation of the benzylic hydrogens by the
OH group, an activation that is also observed with analogous
aliphatic systems (see ref. 47 and references therein) and that is
probably caused by stabilization of positive charge by the α-OH
group in the ionic transition state of the H-abstraction process
ؒ
of OH. Formation of the α-hydroxybenzyl radical is directly
Scheme 2
seen in the pulse radiolysis and laser photolysis experiments
described in Section 1.
the electron-seeking OH group. Stabilization of the incipient
radical by the α-OH group, as recently suggested,51 goes in the
؊
ؒ
b) The reaction of SO4 with benzyl alcohol. Deoxygenated
same direction.
Also, the ortho-, meta-, para-isomer ratio is different from
aqueous solutions containing 1 mM benzyl alcohol, 1 mM
K2S2O8 [to yield SO4 Ϫ, eqn. (5)], 0.2 M tert-butyl alcohol (to
ؒ
ؒ
that in the case of addition of OH to the ring. Addition of
ؒ
scavenge OH ), 2 mM K3Fe(CN)6 (to oxidize any hydroxy-
H2O to the para-carbon of the radical cation is twice as likely
cyclohexadienyl radicals formed), and 1 mM phosphate at pH 6
as that to an ortho- or meta-position, compared to a factor of
were γ-irradiated with doses such that conversion remained
<10%. The products were identified (using GC and HPLC with
optical and electrochemical detection) as the isomeric hydroxy-
benzyl alcohols, benzaldehyde and a small amount of phenol,
ؒ
only ∼1.5 in the case of OH addition to the benzyl alcohol
parent.
The question as to what extent the radical cation is “free”,
i.e. uninfluenced by the counter-ion SO42Ϫ, will be addressed
below.
ؒ
i.e. the same products as in the case of the OH reaction (see
Table 3). However, the distribution of the products is different;
specifically, there is much more benzaldehyde, showing that the
α-hydroxybenzyl radical is formed in much higher yield than in
c) Products from the 248 nm laser photolysis of benzyl alcohol.
Samples (3 mL) of deoxygenated aqueous solutions containing
ؒ
the case of the OH reaction.
3Ϫ
10 mM benzyl alcohol and 1–2 mM FeIII(CN)6 in 1 × 1 cm
The results are explained by Scheme 2, in which it is
assumed48 that the SO4 Ϫ radical reacts with benzyl alcohol by
quartz cuvettes were photolyzed with the unfocused 248 nm
light (KrF*) of the Lambdaphysik 103 EMG ESC excimer
laser, using 200 to 2000 “shots”. The power per pulse (“shot”)
was typically 150 mJ. The irradiated samples were subjected
to HPLC analysis, using the same conditions as described in
sections a, b). With ≤400 shots, the conversion was ≤1%, and
the main photolysis products were the isomeric hydroxybenzyl
alcohols and benzaldehyde. Additional products were seen but
were not identified.52 The distribution of products identified is
presented in row 3 of Table 3.
ؒ
electron transfer49 to give a radical cation–SO42Ϫ ion pair which
reacts further via two channels: a) nucleophilic addition of
water to the radical cation (“hydration”) at all ring positions
including ipso (total yield 56%) followed by deprotonation
from OH and b) deprotonation of one of benzylic H’s (42%).
As shown in section 2a, the “hydration” reaction is reversible,
whereas the deprotonation from the benzylic carbon is not.
Addition and deprotonation are approximately equally fast, a
situation quite different from the case of the toluene radical
cation, which shows a pronounced preference for addition of
water rather than C–H deprotonation.19,27,28,30,31,50 The higher
tendency to deprotonate is probably the result of a higher
acidity of the benzyl carbon in the case of its substitution by
d) Mechanistic conclusions. From these data it is evident that
the ratio of ring addition products (the phenols) to benzalde-
hyde (derived from oxidation of the side chain) is not very
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619
1617

View Article Online
different from that in the case of the reaction of SO4 Ϫ. How-
ؒ
9 S. Steenken and R. A. McClelland, J. Am. Chem. Soc., 1989, 111,
4967.
10 S. Steenken, C. J. Warren and B. C. Gilbert, J. Chem. Soc., Perkin
Trans. 2, 1990, 335.
11 H. Görner, Photochem. Photobiol., 1990, 52, 935.
12 J. L. Faria, R. A. McClelland and S. Steenken, Chem. Eur. J., 1998,
4, 1275.
13 M. E. Snook and G. A. Hamilton, J. Am. Chem. Soc., 1974, 96, 860.
14 C. Walling, D. M. Camaioni and S. S. Kim, J. Am. Chem. Soc., 1978,
100, 4814.
15 C. Walling, G. M. El-Taliawi and C. Zhao, J. Org. Chem., 1983, 48,
4914.
16 R. O. C. Norman, Spec. Publ. Chem. Soc., 1970, 117.
17 R. O. C. Norman, P. M. Storey and P. R. West, J. Chem. Soc. B,
1970, 1087.
18 M. J. Davies, B. C. Gilbert, C. W. McCleland, C. B. Thomas and
J. Young, J. Chem. Soc., Chem. Commun., 1984, 966.
19 B. C. Gilbert, C. J. Scarratt, C. B. Thomas and J. Young, J. Chem.
Soc., Perkin Trans. 2, 1987, 371.
20 B. C. Gilbert and C. W. McCleland, in Cyclisation of γ-Arylalkanols
via Aryl Radical-Cation and Alkoxyl Radical Intermediates,
ed. H. Fischer and H. Heimgartner, Berlin-Heidelberg, 1988.
21 B. C. Gilbert and C. J. Warren, Res. Chem. Intermed., 1989, 11, 1.
22 P. O’Neill, S. Steenken and D. Schulte-Frohlinde, J. Phys. Chem.,
1975, 79, 2773.
ever, there is a large difference with respect to the ratio of the
isomers of hydroxybenzyl alcohols, the para-position being
much more favored in the case of 248 nm photolysis than in
the case of reaction of SO4 Ϫ. The conclusion is thus that the
ؒ
Ϫ
ؒ
species formed by SO4 reaction and by photolysis are not
the same. The ion pair produced by electron removal by SO4
Ϫ
ؒ
2Ϫ 53,54
[radical cation–SO4
]
should have a reactivity different
from the [radical cation–electron] pair. The latter is a “special”
ion pair, due to the very rapid (300 fs55,56) solvation of the
electron,57 by which the bulky and only weakly solvated radical
cation is left behind “naked”. Since in water (even “con-
ventional”) ion pairs are typically very short-lived (≤20 ps)53,57
the addition of water to the one-electron-oxidized aromatic
ring in the ion pair must take place on a similarly short
time scale in order to compete with the ion-pair separation
(= formation of free ions). This implies
a very high
electrophilic/deprotonation reactivity (k1.o. ≥ 5 × 1010
s
Ϫ1) of
2Ϫ
the benzyl alcohol radical cation in the [radical cation–SO4
]
pair.
Experimental
23 P. O’Neill, S. Steenken and D. Schulte-Frohlinde, J. Phys. Chem.,
1977, 81, 31.
The benzyl alcohols and ether(s) were from commercial sources
and they were fractionally distilled to a purity ≥99.8%. Water
was from a Millipore system, acetonitrile was of spectroscopic
grade. The solutions, after removing air by bubbling with
argon, were pumped through the 2 by 4 mm Suprasil quartz
cell (flow rates ca. 0.2–0.5 mL minϪ1) and photolyzed with
unfocused 20 ns pulses of 248 nm light (total energy ca. 5–40
mJ per pulse) from a Lambda-Physik EMG ESC excimer laser.
The light-induced optical transmission changes were digitized
by Tektronix 7612 and 7912 transient recorders interfaced with
a DEC LSI11/73ϩ computer which also process-controlled the
apparatus and on-line preanalyzed the data. Final data analysis
was performed on a Microvax I connected to the LSI. The error
in the quantum yields and extinction coefficients is estimated as
10%.
24 M. Jonsson, J. Lind, G. Merényi and T. E. Eriksen, J. Chem. Soc.,
Perkin Trans. 2, 1995, 67.
25 J. Holcman and K. Sehested, J. Phys. Chem., 1977, 81, 1963.
26 K. Sehested, J. Holcman and E. J. Hart, J. Phys. Chem., 1977, 81,
1363.
27 K. Sehested and J. Holcman, J. Phys. Chem., 1978, 82, 651.
28 M. K. Eberhardt and M. Yoshida, J. Phys. Chem., 1973, 77, 589.
29 C. Walling and D. M. Camaioni, J. Am. Chem. Soc., 1975, 97,
1603.
30 M. K. Eberhardt and M. I. Martinez, J. Phys. Chem., 1975, 79, 1917.
31 M. K. Eberhardt, J. Org. Chem., 1977, 42, 832.
32 For product analysis studies on benzyl alcohols, see M. E. Snook
and G. A. Hamilton, J. Am. Chem. Soc., 1974, 96, 860.
33 L. M. Dorfman, I. A. Taub and R. E. Bühler, J. Chem. Phys., 1962,
36, 3051.
34 R. O. C. Norman and B. C. Gilbert, Adv. Phys. Org. Chem., 1967, 5,
53.
For identification of photo- or radiation-chemical products
(the conversion was always ≤1%), a ∼1 mM solution of the
substrates in aqueous solution in the absence or presence of
additives such as K2S2O8 or tert-butyl alcohol (the latter to
35 K. Eiben and R. W. Fessenden, J. Phys. Chem., 1971, 75, 1186.
36 V. Madhavan and R. H. Schuler, Radiat. Phys. Chem., 1980, 16, 139.
37 G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross,
J. Phys. Chem. Ref. Data, 1988, 17, 513.
38 This peak apparently overlaps the (broader) second band of the
OH-adducts.
ؒ
scavenge radiation-chemically produced OH) was photolyzed
(in a 1 cm quartz cuvette) or γ-irradiated (after saturation of the
solution with N2O) and analyzed by HPLC, using a 150 × 4.6
mm Luna-5-C18 column (from Phenomenex) with 0.8 mL minϪ1
MeOH–H2O 1:2 as eluent (containing 0.01 M NaClO4 in the
case of electrochemical detection) and optical (Shimadzu SPD-
M6A diode array detector) and electrochemical (Perkin-Elmer
LC 17, electrode potential 0.8 V) detection. Identification and
quantitation of the photo- or radiation products was by com-
parison with authentic samples.
39 R. O. C. Norman and P. M. Storey, J. Chem. Soc. B, 1970, 1099.
40 In the case of R = CO2H and CH2OH, a major part of the α-
hydroxybenzyl radicals formed are produced by Cα–Cβ frag-
mentation rather than by deprotonation from Cα (see B. C. Gilbert,
C. J. Scarratt, C. B. Thomas and J. Young, J. Chem. Soc., Perkin
Trans. 2, 1987, 371).
41 P. Neta and R. H. Schuler, Radiat. Res., 1975, 64, 233.
42 As concluded from experiments with ps time resolution, the
formation of eϪaq is complete in ≤20 ps.
43 Ipso-addition by the OH radical is by no means exceptional. It has
been observed with benzenes substituted by OH, OMe, CO2H, NO2,
and halogen.
44 K. Bhatia and R. H. Schuler, J. Phys. Chem., 1974, 78, 2335.
45 G. W. Klein, K. Bhatia, V. Madhavan and R. H. Schuler, J. Phys.
Chem., 1975, 79, 1767.
46 S. Steenken and N. V. Raghavan, J. Phys. Chem., 1979, 83, 3101.
47 H. Eibenberger, S. Steenken, P. O’Neill and D. Schulte-Frohlinde,
J. Phys. Chem., 1978, 82, 749.
48 O. P. Chawla and R. W. Fessenden, J. Phys. Chem., 1975, 79, 2693.
49 A possible alternative to an electron transfer reaction is addition to
the ring [followed by “hydrolysis” of the SO4Ϫ-adduct(s)] and/or
H-abstraction from the CH2OH group. Concerning the latter
reaction, rate constants for abstraction of H from >CHOH are
typically (H. Eibenberger, S. Steenken, P. O’Neill and D. Schulte-
Frohlinde, J. Phys. Chem., 1978, 82, 749) of the order 106–107 as
compared to the measured (Table 1) 3 × 109 MϪ1 sϪ1, indicating that
H-abstraction should be negligible. With the putative addition
reaction, strong steric effects are expected as a result of the bulky
SO4•Ϫ. On the basis of the data of Table 3, however, such steric
effects (i.e. low yield of 2-hydroxy- as compared to 4-hydroxybenzyl
alcohol) are not visible.
References
1 However, even if the driving force for electron transfer is very large
(e.g., >1 eV), the mechanism may still be an inner-sphere one, see,
e.g., (a) S. Steenken, in Free Radicals in Synthesis and Biology,
ed. F. Minisci, NATO ASI 260, Kluwer, Dordrecht, 1989; (b)
L. Eberson, Electron Transfer Reactions in Organic Chemistry,
Springer, Berlin, 1987.
2 Y. Hirata, N. Murata, Y. Tanioka and N. Mataga, J. Phys. Chem.,
1989, 93, 4527.
3 Y. Hirata, M. Ichikawa and N. Mataga, J. Phys. Chem., 1990, 94,
3872.
4 L. P. Candeias and S. Steenken, J. Am. Chem. Soc., 1992, 114, 699.
5 G. Trampe, J. Mattay and S. Steenken, J. Phys. Chem., 1989, 93,
7157.
6 L. I. Grossweiner and H. I. Joschek, Adv. Chem. Ser., 1965, 50, 279.
7 A. Reuther, D. N. Nikogosyan and A. Laubereau, J. Phys. Chem.,
1996, 100, 5570.
8 M. P. Irion, W. Fuβ and K.-L. Kompa, Appl. Phys. B; Photophys.
Laser Chem., 1982, 27, 191.
50 S. Steenken, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 113.
1618
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619

View Article Online
51 B. C. Gilbert, J. R. L. Smith, P. Taylor, S. Ward and A. C. Whitwood,
J. Chem. Soc., Perkin Trans. 2, 1999, 1631.
55 J. M. Wiesenfeld and E. P. Ippen, Chem. Phys. Lett., 1980, 73,
47.
52 These low-yield products are probably derived from methyl-
benzvalene, see L. Kaplan, L. A. Wendling and K. E. Wilzbach,
J. Am. Chem. Soc., 1971, 93, 3821.
56 Y. Gauduel, H. Gelabert and M. Ashokkumar, Chem. Phys., 1995,
197, 167.
57 In the case of thymine in aqueous solution the lifetime of the ion
pair (radical cation–electron) has been measured to be 120 fs:
A. Reuther, D. N. Nikogosyan and A. Laubereau, J. Phys. Chem.,
53 The lifetime of the Naϩ/ClϪ ion pair in water has been estimated to
be 20 ps: P. L. Geissler, C. Dellago and D. Chandler, J. Phys. Chem.
B, 1999, 103, 3706.
ؒ
1996, 100, 5570 In the case of the (Cl –electron) pair in water the
2Ϫ
54 An [SO4 radical cation] pair has been suggested to be involved
lifetime is 770 fs: M. Ashokkumar, H. Gelabert, A. Antonetti and
Y. Gauduel, in Ultrafast Reaction Dynamics and Solvent Effects,
ed. Y. Gauduel and P. J. Rossky, AIP Conference Proceedings 298,
American Institute of Physics, New York, 1994, pp. 107–118.
•Ϫ
in the reaction of SO4 with alkenes in aqueous solution:
G. Koltzenburg, E. Bastian and S. Steenken, Angew. Chem., Int. Ed.
Engl., 1988, 27, 1066.
J. Chem. Soc., Perkin Trans. 2, 2001, 1613–1619
1619