Photohydroxylation of 1,4-Benzoquinone in Aqueous Solution Revisited

Article abstract of DOI:10.1002/chem.200305136

In water, photolysis of 1,4-benzoquinone, Q gives rise to equal amounts of 2-hydroxy-1,4-benzoquinone HOQ and hydroquinone QH₂ which are formed with a quantum yield of ψ=0.42, independent of pH and Q concentration. By contrast, the rate of decay of the triplet (λ_max=282 and ～ 410 nm) which is the precursor of these products increases nonlinearly (k= (2→3.8)×10⁶ s^-1) with increasing Q concentration ((0.2→10) mM). The free-radical yield detected by laser flash photolysis after the decay of the triplet also increases with increasing Q concentration but follows a different functional form. These observations are explained by a rapid equilibrium of a monomeric triplet Q* and an exciplex Q₂* (K=5500±1000m^-1). While Q* adds water and subsequent enolizes into 1,2,4-trihydroxybenzene Ph(OH) ₃, Q₂* decays by electron transfer and water addition yielding benzosemiquinone ^.QH and ^.OH adduct radicals ^.QOH. The latter enolizes to the 2-hydroxy-1,4-semiquinone radical ^.Q(OH)H within the time scale of the triplet decay and is subsequently rapidly (microsecond time scale) oxidized by Q to HOQ with the concomitant formation of ^.QH. On the post-millisecond time scale, that is, when ^.QH has decayed, Ph(OH)₃ is oxidized by Q yielding HOQ and QH₂ as followed by laser flash photolysis with diode array detection. The rate of this pH- and Q concentration-dependent reaction was independently determined by stopped-flow. This shows that there are two pathways to photohydroxylation; a free-radical pathway at high and a nonradical one at low Q concentration. In agreement with this, the yield of Ph(OH)₃ is most pronounced at low Q concentration. In the presence of phosphate buffer, Q* reacts with H₂PO₄-giving rise to an adduct which is subsequently oxidized by Q to 2-phosphato-1,4-benzoquinone QP. The current view that ^.OH is an intermediate in the photohydroxylation of Q has been overturned. This view had been based on the observation of the ^.OH adduct of DMPO when Q is photolyzed in the presence of this spin trap. It is now shown that Q*/Q₂* oxidizes DMPO (k ≈1×10⁸M ^-1S^-1) to its radical cation which subsequently reacts with water. Q*/Q₂* react with alcohols by H abstraction (rates in units of M^-1S^-1): methanol (4.2×10 ⁷), ethanol (6.7×10⁷), 2-propanol (13×10 ⁷) and tertiary butyl alcohol (～0.2×10⁷). DMSO (2.7×10⁹) and O₂ (～2×10⁹) act as physical quenchers.

Full text of DOI:10.1002/chem.200305136

FULL PAPER
Photohydroxylation of 1,4-Benzoquinone in Aqueous Solution Revisited
Justus von Sonntag,*[a] Eino Mvula,[a, b] Knut Hildenbrand,[b] and Clemens von Sonntag[a, b]
Abstract: In water, photolysis of 1,4-
benzoquinone, Q gives rise to equal
amounts of 2-hydroxy-1,4-benzoqui-
none HOQ and hydroquinone QH2
which are formed with a quantum yield
of F=0.42, independent of pH and Q
concentration. By contrast, the rate of
decay of the triplet (lmax =282 and ~
410 nm) which is the precursor of these
products increases nonlinearly (k=
none CQH and COH adduct radicals
CQOH. The latter enolizes to the 2-hy-
agreement with this, the yield of
Ph(OH)3 is most pronounced at low Q
concentration. In the presence of phos-
droxy-1,4-semiquinone
radical
ꢀ
CQ(OH)H within the time scale of the
triplet decay and is subsequently rapid-
ly (microsecond time scale) oxidized by
Q to HOQ with the concomitant for-
mation of CQH. On the post-millisec-
ond time scale, that is, when CQH has
decayed, Ph(OH)3 is oxidized by Q
yielding HOQ and QH2 as followed by
laser flash photolysis with diode array
detection. The rate of this pH- and Q
concentration-dependent reaction was
independently determined by stopped-
flow. This shows that there are two
phate buffer, Q* reacts with H2PO4
giving rise to an adduct which is subse-
quently oxidized by Q to 2-phosphato-
1,4-benzoquinone QP. The current
view that COH is an intermediate in the
photohydroxylation of
Q has been
overturned. This view had been based
on the observation of the COH adduct
of DMPO when Q is photolyzed in the
presence of this spin trap. It is now
shown that Q*/Q2* oxidizes DMPO (k
ꢂ1î108 mꢀ1 sꢀ1) to its radical cation
which subsequently reacts with water.
Q*/Q2* react with alcohols by H ab-
straction (rates in units of mꢀ1 sꢀ1):
methanol (4.2î107), ethanol (6.7î107),
2-propanol (13î107) and tertiary butyl
alcohol (~0.2î107). DMSO (2.7î109)
and O2 (~2î109) act as physical
quenchers.
(2!3.8)î106 sꢀ1) with increasing
Q
concentration ((0.2!10) mm). The
free-radical yield detected by laser
flash photolysis after the decay of the
triplet also increases with increasing Q
concentration but follows a different
functional form. These observations
are explained by a rapid equilibrium of
a monomeric triplet Q* and an exci-
plex Q2* (K=5500ꢁ1000mꢀ1). While
Q* adds water and subsequent enolizes
into 1,2,4-trihydroxybenzene Ph(OH)3,
Q2* decays by electron transfer and
water addition yielding benzosemiqui-
pathways to photohydroxylation;
a
free-radical pathway at high and a non-
radical one at low Q concentration. In
Keywords: excimer
hydroxylation photochemistry
quinones ¥ reaction mechanism
¥
¥
¥
Introduction
induced by the abundance of quinoid structures in nature
and in technical products. For example, some p-quinone-
based dyes show considerable photodegradation of the
fabric while others do not.[6] This photo-tendering action is
due to an H abstraction from the substrate, and therefore
this type of reaction has been studied in detail in organic
solvents and in aqueous solutions, not only with 1,4-benzo-
quinone, but also with substituted 1,4-benzoquinones includ-
In 1886 and 1888, Klinger published the first, and up until
today still worth-reading, reports on the chemical effect of
light on a quinone in solution.[1,2] Ever since, there has been
a continuing interest in the photochemistry of quinones,[3
6]
[a]Dr. J. von Sonntag, Dr. E. Mvula, Prof. Dr. C. von Sonntag
Leibniz-Institut f¸r Oberfl‰chenmodifizierung (IOM)
Permoserstrasse 15
49]
ing 1,4-naphtho- and 9,10-anthraquinones.[7
1,4-Benzoquinone, Q, the parent of the quinones investi-
gated, has three absorption bands at 250, 300 and 424 nm
(in water, see Figure 1). The short-wavelength bands have
been attributed to a p!p* transitions, while the long-wave-
length band is due to an n!p* transition.[50] In a nonpolar
solvent, the n!p* transition shows a vibrational fine struc-
ture.[18] In water, the n!p* transition is blue-shifted, and
the vibrational fine structure is no longer observed (see
04318 Leipzig (Germany)
Fax : (+49)341-235-2584
E-mail: Justus.von.Sonntag@iom-leipzig.de
[b]Dr. E. Mvula, Dr. K. Hildenbrand, Prof. Dr. C. von Sonntag
Max-Planck-Institut f¸r Strahlenchemie
Stiftstrasse 34 36, P.O. Box 101365
45470 M¸lheim an der Ruhr(Germany)
E-mail: Justus.von.Sonntag@iom-leipzig.de
440
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305136
Chem. Eur. J. 2004, 10, 440 451

440 451
Figure 1). Q does not show fluorescence of any importance,
and it has been concluded that rapid intersystem crossing
occurs, that is, the n!p* triplet Q* is rapidly populated and
is the photochemically active species.[50] In fact, picosecond
time-resolved Raman scattering experiments indicate that
the triplet is present within 20 ps.[51] In the presence of a hy-
drogen donor, for example, 2-propanol, photoreduction
occurs with a quantum yield close to unity [1 mol QH2/Ein-
stein, reactions (1) (3)],[7,11,18] and the analogy to ketone
photochemistry has been addressed.
The view that COH are formed in the reaction of excited
quinones with water is shared also by other au-
thors.[32,52,53,58,64,70] However, it has been calculated that the
reduction potential of the relaxed triplet state of Q is insuf-
ficient to oxidize water, and it has been suggested that the
water oxidation (formation of 1,4-benzosemiquinone CQH
observed by time-resolved Raman spectroscopy) must be
due to the reaction of an unrelaxed triplet state.[35] As an-
other potential route to COH, electron transfer from a
ground-state quinone to an excited quinone and oxidation
of OHꢀ or water by the ensuing quinone radical cation has
been considered.[4,54,55] In a recent study, the formation of a
free COH has been rejected, and instead a kind of ™krypto
COH∫ (COH complexed to CQH) has been favoured.[60]
If COH were formed their addition to Q would indeed
result in the formation of the observed products, HOQ and
QH2 {reactions (5) (7) and (3); for details of the complex
kinetics see ref. [71]}.
In the solid state and in inert organic solvents, quinones
undergo cycloaddition reactions.[5] In water and in the ab-
sence of hydrogen donors, however, Q is efficiently photo-
hydroxylated (quantum yield of photodecomposition F=
0.31,[19] F=0.5,[37]), and as resulting products 2-hydroxy-1,4-
benzoquinone, HOQ, and hydroquinone, QH2, are formed
in equal amounts (F=0.25).[37] This type of reaction is also
given by other quinones, for example, anthraquinonesulfo-
59]
nates,[21,52
2-methyl-1,4-benzoquinone,[60] 2-methoxy-1,4-
benzoquinone[47] and 2,6-dimethylbenzoquinone.[61]
Photohydroxylation is suppressed by the addition of
halide ions.[37,59,62,63] As short-lived intermediates CT-com-
plexes have been suggested that can either decay to the
ground state or, at elevated halide concentrations, form di-
halide radical anions and a semiquinone radical anion,
QCꢀ.[62] Oxidation of carbonate to CO3Cꢀ has also been re-
ported.[64,65] Electron transfer processes are also made re-
sponsible for the quinone-sensitized photooxidation of nu-
The observation of the typical four-line COH adduct to
DMPO, when Q is photolyzed in the presence of this spin
trap, has been taken as strong support for the H abstraction
reaction (4).[39] However, it is well known that this COH-
adduct radical can have other adduct radicals as precursor,
for example, the superoxide radical, O2Cꢀ.[72] Moreover, ex-
cited states can oxidize spin traps,[73] and the ensuing spin
trap radical cation may react with water resulting in an
COH adduct. Thus, this evidence is not as strong as believed.[39]
There are, however, observations that are not compatible
with free COH as intermediates. The addition of an excess of
tertiary butyl alcohol that would have scavenged all free
COH does not affect the spectral developments after a laser
flash up to 200 ms.[32] Also a kind of warning is the fact that
acetone in its triplet state does not abstract a hydrogen
69]
cleobases.[66] This view is supported by EPR studies.[67
QCꢀ (F=0.47) is reported to be formed as a transient in
the photolysis Q in water.[38] This is in contrast to data
which indicate that in this system 1,4-benzosemiquinone-
type radicals must be minor contributors to photohydroxyla-
tion (<10%) compared with a non-radical pathway.[37]
Based on spin-trapping experiments, it has been suggest-
ed[39,44,45] that the first step is an H-abstraction from water
[reaction (4)}, that is, that the reaction is analogous to reac-
tion (1).
441
Chem. Eur. J. 2004, 10, 440 451
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. von Sonntag et al.
FULL PAPER
atom from water, although its triplet energy (322 kJmolꢀ1)
254 Da), m/z (%): 254 (84), 239 (100), 223 (4), 112 (8), 73 (82); 2-hydroxy-
methylhydroquinone-tris-TMS (MW 356 Da), m/z (%): 356 (44), 341 (13),
267 (15), 253 (9), 194 (33), 147 (14), 73 (100). 2-Hydroxymethyl-1,4-
benzoquinone was reduced with 1,2,4-trihydroxybenzene to 2-hydroxy-
methylhydroquinone. While the HPLC peak at 36.6 min disappeared, the
peak at 18.9 min increased accordingly.
is higher than that of Q (224 kJmolꢀ1).[74]
In the present paper, we will show that neither a free COH
radical nor an COH radical complexed to CQH are intermedi-
ates in the photohydroxylation reaction, and evidence for an
alternative mechanism will be given.
For the stopped-flow experiments, the Biologic SFM-3 setup equipped
with a diode array detector (Tidas, J&M, Aalen, Germany) was used.
Methanesulfinic acid was readily detected by ion chromatography
(Dionex DX100; column: AG14/AS14, eluent: water containing 4.5î
10ꢀ4 m Na2CO3 and 4.25î10ꢀ4 m NaHCO3) as described elsewhere.[78, 79]
Experimental Section
For the EPR experiments, a stock solution of 5,5-dimethylpyrroline-N-
oxide (DMPO, Aldrich; 250 mg) in CH3CN (5 mL) was kept under N2 at
ꢀ188C. Spin-trapping experiments were carried out on 3.5 mm solutions
of DMPO in 50 mm phosphate buffer, pH 7.2, containing either H2O2 or
Q in the absence or presence of tertiary butyl alcohol. EPR spectra were
recorded on a Varian E-9 X-band spectrometer equipped with an inter-
face from Stelar s.n.c, Mede (PV), Italy and a PC. The spin adducts were
generated by UV irradiation of the solutions in the cavity of the EPR in-
strument with unfiltered light from a Xenon short arc lamp (LX300UV,
Cermax, ICL Technology, Sunnyvale, CA).
1,4-Benzoquinone (Aldrich) was purified by sublimation. Hydroquinone
and 1,2,4-trihydroxybenzene, Ph(OH)3 (both Aldrich) were available as
reference material. Solutions were prepared in Milli-Q-filtered (Milli-
pore) water and stored in the dark. To minimize any exposure to light,
all glassware was covered with black cloth or aluminum foil. For product
studies, the aqueous Q solutions were irradiated at 254 nm using a low-
pressure Hg-arc, while kinetic studies employed the 308 nm light from an
XeCl excimer laser (for the UV spectrum of Q in water see Figure 1;
e(254 nm)=1500 m2 molꢀ1 e(308 nm)=25 m2 molꢀ1 in agreement with
,
The laser photolysis set-up comprised a 308 nm XeCl-excimer laser
(MINex, LTB Berlin, pulse train of three pulses (70, 20 and 10% of total
energy, respectively, each with 5 ns half width, all three within 70 ns, total
pulse train energy up to 15 mJ) as excitation source and a pulsed xenon
short-arc lamp (XBO 450, Osram, power supply LPS 1200, lamp pulser
MCP 2010, both Photon Technology International) supplying the ana-
lyzinglight. The transient recording electronics, a photomultiplier (1P28,
Hamamatsu, operated at 900 V, power supply: PS310, Stanford Research
ref. [19])
Systems) and
a
500 MHz, 2.5 GSsꢀ1 digitizing storage oscilloscope
(TDS620b, Tektronix) guarantee a time resolution within the limits set
by the excitation source. Further details have been published.[80]
The same 308 nm laser was also used for experiments in the minute time
range (denoted as laser flash photolysis long time range detection LFP-
LTRD). In this case, the laser beam was directed onto a Suprasil 10 mm
fluorescence cell (Hellma, Germany) placed into a thermostated and stir-
red cell holder (Flash 100, QNW, USA). Transient spectra were recorded
orthogonal to the laser beam with a fast diode array spectrometer (Tidas
II, J&M, Aalen, Germany) via fiber optics and a home-made coupling to
the cell holder. The trigger system of the laser flash photolysis[81] was
modified and reprogrammed both to synchronize the spectrophotometer
and the laser and to allow a burst of several laser pulses within a very
short time to produce a stronger ™pulse∫.
Figure 1. UV absorption spectrum of 1,4-benzoquinone in aqueous solu-
tion. The excitation wavelengths used in this study (low-pressure Hg-arc
at 254 nm and XeCl excimer at 308 nm) are marked.
For UV irradiation at 254 nm, a low-pressure mercury lamp which does
not emit 185 nm radiation (Heraeus Sterisol NN 30/89) was used. Actino-
metry was done using the ferrioxalate system.[75,76] The contribution to
the bleaching of the ferrioxalate by wavelengths longer than 254 nm was
determined by a separate measurement, blocking the UV light with a
plate of ordinary glass. The effect of the 254 nm radiation was then ob-
tained from the difference (5î1017 photonssꢀ1 dmꢀ3). Irradiations were
carried out in Suprasil cells (Hellma), 4î1 cm2 (face)î1 cm, and lasted
typically less than a minute requiring an automatic shutter (rise time
10 ms).
For pulse radiolysis the 11 MeV linear electron accelerator Electronika
003 (Thorium, Moscow) delivering 43 Gy pulses of 7 ns duration was
used. Its recording optics and electronics were essentially identical to the
laser flash photolysis set-up.
Simulations were performed on the complete 4D data matrix (time,
wavelength, absorbance, concentration) with the specialized 4D data fit-
ting software Pro-K II (Applied Photophysics, UK), see Figure 7. The
models were double-checked with the Chemical Kinetics Simulator soft-
ware, version 1.01, developed by IBM at the Almaden Research Center.
The products were identified by HPLC on a Nucleosil 5C18 column with
water as eluent; retention times (min): HOQ (5.9), Ph(OH)3 (10.1), QH2
(15.5), 2-hydroxymethylhydroquinone (18.9), 2-hydroxymethyl-1,4-benzo-
quinone (36.6), Q (40.7). Under these conditions, HOQ was deprotonat-
ed. When an acidic eluent with 10% methanol was used, HOQ (14 min)
eluted close to Q (15 min), and the separation was not as good. However,
this eluent allowed us to separate well HOQ from 1,4-benzoquinone-2-
phosphate QP (5 min) and also from QH2 (8.5 min). The reaction of
Ph(OH)3 with Q led to a quantitative formation of HOQ (for details see
below), and this method was used for the calibration of HOQ.
Results and Discussion
Final products and their quantum yields: In the preceding
studies on the photolysis Q in aqueous solution, it has been
suggested by a comparison of the UV spectra that the final
products are HOQ and QH2.[19] From this, a quantum yield
of F(Q consumption)=0.31 was obtained.[19] The quantum
yield was found to be independent of pH in the range of
pH 0.4 8. We have recently shown that HOQ is a fairly
strong acid (pKa 4.2).[71] Its anion is characterized by a
strong absorption at 482 nm (e=200 m2 molꢀ1), while the
free acid has an absorption maximum at 380 nm
(e=140 m2 molꢀ1). We had used the electrochemical oxida-
The formation of 2-hydroxymethylhydroquinone as well as 2-hydroxy-
methyl-1,4-benzoquinone in the photoreduction of Q by methanol (10%
of the QH2 yield, confirming earlier pulse radiolysis data[77] on the reac-
tion of CCH2OH with Q) was ascertained by gas chromatography com-
bined with mass spectrometry after rotary evaporation of the irradiated
solution and trimethylsilylation of the residue. The two hydroquinones
were characterized by their mass spectra. Hydroquinone-bis-TMS (MW
442
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 440 451

Photohydroxylation of 1,4-Benzoquinone
440 451
tion of 1,2,4-trihydroxybenzene (Ph(OH)3) to generate
HOQ,[71] but its oxidation by Q[19] (reaction 8; for the kinet-
ics of this reaction see below) is an alternative and more
convenient procedure that was now used for calibration.
This product is characterized by absorption maxima at
~254 and 361 nm (HPLC, diode array, in agreement with
the transient difference spectrum from LFP-LTRD experi-
ments, see Figure 13). QP is unstable and fades away in neu-
tral solution. It was thus impossible to collect sufficient ma-
terial by preparative HPLC for its characterization by NMR
spectroscopy.
Based on redox balance considerations, the quantum yield
of QH2 must match that of the sum of HOQ and QP. The
yields of the former two can be determined and thus the
yield of the latter calculated. As can be seen from Figure 3,
the quantum yield of HOQ is independent of the phosphate
buffer concentration, while those of QH2 and QP rise with
increasing buffer concentration.
With Q in some excess, one mol Ph(OH)3 yields one mol
HOQ and one mol QH2. This quantitative conversion was
ascertained by HPLC (see Experimental Section). The com-
plete consumption of Ph(OH)3 is balanced by an equivalent
formation of QH2. Based on this calibration, it has now
been shown that in the photolysis of Q in aqueous solution
QH2 and HOQ are formed in equal yields (inset in
Figure 2). No further products were detected by HPLC.
From Figure 2 (main graph), it can be seen that the yield
of HOQ is neither dependent on pH (pH 2.1 and 7; confirm-
ing earlier observations[19]) nor on the concentration of Q
(0.1 mm and 1 mm). Based on actinometry, quantum yields
were determined at F=0.42 for the two products, QH2 and
HOQ. This is in fair agreement with an earlier value[19] of
F=0.31.
Figure 3. Photolysis of aqueous solutions of 1,4-benzoquinone at 254 nm.
*
Quantum yields of the formation of 2-hydroxy-1,4-benzoquinone ( ), hy-
*
droquinone ( ) and 1,4-benzoquinonone-2-phosphate as a function of
~
~
the phosphate buffer concentration pH 4.6 ( ) and pH 3.44 ( ).
Identical data were obtained at pH 4.6 and 3.44. At these
two pH values, the phosphate buffer is mainly present as
2ꢀ
H2PO4ꢀ. Even at pH 4.66, the HPO4 concentration is only
0.2% of the total buffer concentration, and at the lower pH
2ꢀ
it drops to 0.016%. If HPO4 were the reactive species in
the experiments shown in Figure 3 the QP yield at the high-
est phosphate buffer concentration (0.8m) should be less
than that found for 0.1m phosphate buffer at pH 4.6. This
not being the case, we conclude that the phosphate buffer
Figure 2. UV Photolysis (254 nm) of 1,4-benzoquinone in Ar-saturated
aqueous solution. Formation of 2-hydroxy-1,4-benzoquinone as a func-
~
,
~
*
,
*
ꢀ
tion of time at pH 2.1 (
) and pH 7 (
) at two different 1,4-benzo-
effect observed here is given by the H2PO4 ions. Experi-
~
,
~
,
*
*
: 1 mm). Inset: Formation of
quinone concentrations (
: 0.1 mm;
ments at pH values noticeably above the ones studied and
at high phosphate buffer concentrations (i.e., high concen-
trations of HPO42ꢀ) were not feasible, since not only QP but
also Q (to a lesser extent) are not sufficiently stable under
these conditions.
It is noted that in contrast to phosphate, the quantum
yield of photodecomposition of Q is greatly diminished by
the addition of halide ions, chloride and bromide.[19]
~
*
hydroquinone ( ) and 2-hydroxy-1,4-benzoquinone ( : as a function of
the irradiation time ([1,4-benzoquinone]=1 mm).
To keep the pH constant, typically ™inert∫ buffers such as
phosphate are used. In the presence of high concentrations
of phosphate buffer pH 4.6, however, a new product is ob-
served that is attributed to 1,4-benzoquinone-2-phosphate,
QP.
Spin-trapping studies: That COH may be an intermediate in
the photohydroxylation of Q is continuously suggested since
the early work on this subject. The ™proof∫ came from spin-
trapping studies. In the presence of DMPO, a four-line EPR
spectrum, typical for the DMPO COH adduct, has been ob-
tained.[39,44] Competition with ethanol[39] and with formate[44]
yielded data that were close to those calculated on the basis
443
Chem. Eur. J. 2004, 10, 440 451
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. von Sonntag et al.
FULL PAPER
of the known[82] rate constants of COH with DMPO and
these substrates. Considering that these two substrates un-
dergo ready photoreduction and thus this reasonable agree-
ment could be due to an artefact, we set-up a competition
with tertiary butyl alcohol, a substrate that is a very poor hy-
drogen donor, that is, not very reactive towards carbonyl ex-
cited states, but nevertheless readily reacts with COH (k=
6î108 mꢀ1 sꢀ1).[82] As can be seen from Figure 4, we also ob-
serve a four-line signal, identical to that obtained upon pho-
tolysis of H2O2. In such competitions, tertiary butyl alcohol
is much more effective in reducing the four-line signal when
it was generated via COH formed in the photolysis of H2O2
than when formed upon Q photolysis (Figure 4).
Table 1. Compilation of rate constants [mꢀ1 sꢀ1]relevant for the present
study.
Reaction
Rate constant
Reference
Q* + O2
~2î109
2.4î109
>5î109
1.3î108
6.5î107
4.2î107
2î106
this work
[32]
Q* + Q
Q* + 2-propanol
Q* + ethanol
Q* + methanol
Q* + tBuOH
Q* + DMSO
Q* + DMPO
Q + Ph(OH)3
this work
this work
this work
this work
this work
this work
this work
this work
2.7î109
~1î108
19 (pH 4.6)[a]
8î105
Q + Ph(OH)2Oꢀ
Q + COH !products
tBuOH + COH !products
DMSO + COH !products
CQOH !CQ(OH)H
CQ(OH)H + Q !HOQ + CQH
CQ(OH)ꢀ + Q !HOQ + QCꢀ
this work
6î109
[71]
6î108
[82]
[82]
[71]
[71]
[71]
7î109
2.5î106 sꢀ1
ꢃ 2.4î107
2î109
[a]Upper limit for the uncatalyzed reaction. For proton and buffer catal-
ysis see text.
The 1,4-benzoquinone triplet state(s): The first transient ob-
served by laser flash photolysis of Q has been assigned to its
triplet state Q* (for the likely involvement of an excimer,
Q2*, see below).[32] In this study,[32] where the UV absorption
spectrum of the triplet was given in the range between 350
Figure 4. Normalized yields of the four-line EPR spectrum (cf. inset) in
*
(
the photolysis of 1,4-benzoquinone
:
0.5 mm) in the presence of
*
DMPO (17.5 mm), and in the photolysis of H2O2 ( : 0.4%) as a function
of the tertiary butyl alcohol concentration. Inset: photolysis of H2O2 in
the absence (a) and in the presence of tBuOH (b, 40 mm) and in the pho-
tolysis of 1,4-benzoquinone in the presence of 120 mm tertiary butyl alco-
hol (c).
and 500 nm, the transient displayed
a maximum at
~410 nm, but the much stronger one at 282 nm (cf.
Figure 5) had been missed, and also in the more recent
study on 2-methyl-1,4-benzoquinone,[60] no attempt has been
made to measure the triplet spectrum below 330 nm. The
shapes of the short-wavelength absorption band when put
on an energy scale shows a good Gaussian distribution, and
based on this analysis its maximum is located at 282 nm.
Upon the decay of Q* (and Q2*, see below), species with
maxima at ~310 and 410 nm are formed (Figure 5). The rate
of triplet decay has been reported at k=1.9î106 sꢀ1.[32] In
We thus conclude that in the case of the photolysis of Q
the four-line signal must have another precursor. It has been
estimated that the oxidation potential of the relaxed Q*
may be around 2.4 V,[35] that is, very high but below the re-
duction potential of COH (E7 =2.65 V).[83] The oxidation po-
tential of DMPO is around 1.7 V.[73] In analogy to experi-
ments where photoexcited chloranil was reacted with
DMPO in the presence of various acids as nucleophiles,[73]
we suggest that an electron transfer from the spin trap to
Q* occurs [reaction (9)]. This is followed by nucleophilic ad-
dition of water to the DMPO radical cation with the con-
comitant release of a proton [reaction (10)]. The rate con-
stant for the reaction of Q* with tertiary butyl alcohol has
been measured here (see below) at ~2î106 mꢀ1 sꢀ1 (for a
compilation of rate constants see Table 1). From this value
and the data shown in Figure 4, we estimate a rate constant
in the order of 108 mꢀ1 sꢀ1 for the reaction of Q* with
DMPO.
Figure 5. 308 nm laser flash photolysis of 1,4-benzoquinone (4 mm) in N2-
*
saturated aqueous solution. UV/Vis spectra recorded at 100 ns ( ) and
*
1.4 ms ( ) after the flash. The wavelengths close to the exciting wave-
length had to be omitted.
444
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 440 451

Photohydroxylation of 1,4-Benzoquinone
440 451
this experiment, the Q concentration was 0.5 mm. We have
now investigated the dependence of the lifetime of this spe-
cies as a function of Q concentration which was varied in
the range of (0.2!10) mm. As can be seen from Figure 6, at
the lowest concentration used the rate of decay is 2î106 sꢀ1
and increases with increasing Q concentration, approaching
at high Q concentration a constant value of k=3.8î106 sꢀ1.
the radical yield is also a function of the Q concentration
(Figure 6, lower curve, left scale).
In a flash photolysis study on duroquinone (2,3,5,6-tetra-
methyl-1,4-benzoquinone) in aqueous solution,[84] free-radi-
cal formation was only observed upon triplet triplet annihi-
lation, that is, at high triplet concentrations. Here, this pro-
cess is excluded because the yield per dose was found to be
independent of laser flash intensity, in agreement with the
weak intensities of our laser flashes. The much shorter life-
time of Q*/Q2* as compared to that of the duroquinone trip-
let further reduces a potential contribution of triplet triplet
reactions.
The disappearance of free-radical intermediates at vanish-
ing Q concentration together with the independence of
F(photohydroxylation) on the Q concentration suggests
that two processes occur side by side. We conclude that Q*
reacts with ground state Q forming an exciplex, Q2*. While
Q* hydrates without giving rise to a product that absorbs
strongly at >300 nm, Q2* gives rise to radical formation and
hence to intermediates absorbing at long wavelengths. With
Q* and Q2* in rapid equilibrium (k11 ꢂ5î109 mꢀ1 sꢀ1) and
decay rates of 2î106 sꢀ1 for Q* and 3.8î106 sꢀ1 for Q2*, the
stability constant K11 =5500ꢁ1000mꢀ1 is calculated from the
data shown in Figure 6. The solid lines drawn through the
data points of these figures were calculated on this basis.
With such a rapid equilibrium and fast decay rates, the
differences in the UV absorption spectra of Q* and Q2* can
only be obtained by global analysis. The result is shown in
Figure 7. According to this analysis, the absorption maxima
of Q* and Q2* do not differ much. However, compared with
Q*, Q2* shows a lower absorption at the short-wavelength
maximum.
Dioxygen quenches Q*/Q2* (see also ref. [32]), and conse-
quently the ensuing free-radical yield is reduced. The rate
constant derived from plotting kobs versus the dioxygen con-
centration (1.0î109 mꢀ1 sꢀ1 at 2.2 mm Q, 1.1î109 mꢀ1 sꢀ1 at
4.0 mm Q) does not agree with the rate constant derived
from a Stern-Volmer plot based on the final free-radical
yield using the same experimental data (2.6î109 mꢀ1 sꢀ1 at
2.2 mm Q, 1.9î109 mꢀ1 sꢀ1 at 4.0 mm Q). This shows, that the
mechanistic assumption behind the Stern-Volmer plot does
Figure 6. Laser flash photolysis at 308 nm of 1,4-benzoquinone in N2-satu-
*
rated aqueous solution at pH 2. Right scale ( ): Rate of decay of the 1,4-
benzoquinone triplet(s) at 285 nm as a function of the 1,4-benzoquinone
concentration. The value obtained in ref. [32]has been included (trian-
*
gle). Left scale ( ): Ratio of the initial absorption at 415 nm (assigned to
the triple) to the final absorption at this wavelength (assigned to prod-
ucts, for example, free radicals) as a function of the 1,4-benzoquinone
concentration. The solid lines have been calculated on the basis of the
harmonic-mean approach to equilibrium kinetics.
Evidently, the Q concentration-dependence of the triplet
lifetime is not compatible with a simple reaction of Q* with
water, and at some point ground-state Q has to be involved
as well. Without the yield data (lower curve in Figure 6),
one would be tempted to ignore the curvature of the triplet
decay rates (upper curve in Figure 6) and draw a straight
line. This would result in a quenching rate of Q* by Q of
2î108 mꢀ1 sꢀ1. However, fitting the yield data by convention-
al first-order competition kinetics results in a 20 times
higher rate constant. Apparently, a more complex reaction
Scheme is required.
It will be discussed below that Q* is most likely in rapid
equilibrium with its dimer, Q2*, an exciplex consisting of
two Q molecules [equilibrium (11)]. This can account for
the data shown in Figure 6 and the other data to be reported
next.
Figure 7. Global analysis of the transients. c: 1,4-benzoquinone triplet,
Q*; g: triplet exciplex, Q2*; a: decay product of Q2*, semiquinone
radical CQH products plus 2-hydroxy-1,4-benzoquinone, QOH; c:
decay product of Q*, 1,2,4-trihydroxybenzene Ph(OH)3 and/or its precur-
sors.
When the triplet has decayed, a noticeable absorption re-
mains (cf. open circles in Figure 5). This has been attributed
to the formation of free radicals.[32] Since only one Q con-
centration had been studied, it escaped attention[32,38] that
445
Chem. Eur. J. 2004, 10, 440 451
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. von Sonntag et al.
FULL PAPER
not hold, that is, simple competition kinetics do not take
place. These marked difference indicate complex kinetics in-
volving the quenching of more than one triplet state by O2.
Alcohols also quench the 282 nm absorption ([Q]=(2!
4) mM in these experiments). The rate constants increase
with increasing H-donating property of the alcohol from
~0.2î107 mꢀ1 sꢀ1 (tertiary butyl alcohol, cf. Figure 8) to 13î
107 mꢀ1 sꢀ1 (2-propanol), cf. Table 1.
Our more stringent data suggest that DMSO acts over-
whelmingly as a physical quencher. Such processes have
been observed before with many other systems, and with op-
tically active sulfoxides an isomerization is observed; an ex-
ciplex as a very short-lived intermediate has been postulat-
ed.[87,88]
The free-radical intermediates: The formation of free-radi-
cal intermediates has been noticed before and has been at-
tributed to the formation of CQH and CQ(OH)H,[32] or only
to CQH,[38] and, depending on pH, to their corresponding
radical anions. From a flash photolysis study of a 0.4 mm Q
solution at pH 7, it was concluded that F(CQH)=0.47.[38] In
this calculation, the formation of HOQ that must have been
formed upon the oxidation of CQ(OH)H by Q [cf. reac-
tion (7)]at the time scale of these experiments (cf. ref. [71)]
has not been accounted for, and thus F(CQH) is lower. Radi-
cal formation is now suggested to be due to a reaction of
Q2* with water (reaction 12) followed by a rapid conversion
of CQOH into CQ(OH)H (reaction 6, k=2.6î106 sꢀ1).
Figure 8. 308 nm laser flash photolysis UV of 1,4-benzoquinone (4 mm) in
*
*
aqueous solution in the absence ( ) and presence ( ) of tBuOH (0.4m).
The spectra were recorded 1.4 ms after the flash. Inset: Observed rate
constant of the decay at 285 nm of the 1,4-benzoquinone triplet as a func-
tion of the tertiary butyl alcohol concentration. The regression line yields
a value of k=(1.94ꢁ0.2)î106 mꢀ1 sꢀ1
.
The same species can be generated in a pulse radiolysis
experiment. There, solvated electrons and COH are formed
in about equal yields with a contribution of some 10% hy-
drogen atoms. In their reaction with Q, solvated electrons
(and potentially also CH) give rise to CQH, while COH yield
CQ(OH)H (via CQOH) (cf. reactions (5) (7), ref. [71], and
references therein).
In Figure 9, the spectra obtained by pulse radiolysis of N2-
saturated aqueous Q solutions at neutral and low pH are
compared with corresponding laser experiments. The pKa
value of CQH is at 4.0,[89] (4.1),[90] and that of CQ(OH)H is at
4.9.[71] In view of the fast oxidation of CQ(OH)ꢀ by Q at
pH 7 (k=2î109 mꢀ1 sꢀ1),[71] 1 2 ms after the laser/electron
pulse we deal with a 2:1 mixture of CQH and HOQ. HOQ
also absorbs in the same wavelength region as CQH, and its
spectrum also changes with pH (pKa(HOQ)=4.1).[71] At low
pH, the oxidation of the now no longer deprotonated
CQ(OH)H is much slower (kꢃ2.4î107 mꢀ1 sꢀ1). The rate of
reaction is governed by the rate of oxidation of CQ(OH)ꢀ in
equilibrium,[71] and at the pH and Q concentrations of
Figure 9 HOQ formation is negligible under acidic condi-
tions at 1 ms.
Because in the present system we deal with two products
(CQH and HOQ), the spectral differences between low and
high pH are less pronounced here than in systems contain-
ing only[89,90] CQH. However, the good agreement of these
two data sets clearly shows that free radicals are indeed
formed upon the decay of Q2*. In contrast to the earlier
pulse radiolysis study,[71] the intermediate CQOH cannot be
detected in the laser flash experiment, because its conver-
sion into CQ(OH)H takes place at the same rate as the
decay of Q*/Q2*.
In an earlier study,[32] experiments have also been carried
out in the presence of 0.1m tertiary butyl alcohol. The
rather scattering data have been interpreted as being com-
patible with a decrease of the total semiquinone radical
yield, and it has been suggested that tertiary butyl alcohol
reacts with the COH radicals formed according to reac-
tion (4). The effect of tertiary butyl alcohol on the total
semiquinone radical yield has been re-investigated here in
somewhat more detail, and it is now clear that in fact the
total semiquinone radical yield increases with increasing ter-
tiary butyl alcohol concentration (cf. Figure 8 which shows
two out of five data sets).
Dimethylsulfoxide (DMSO) is a much better triplet
quencher than the alcohols (Table 1). This effective quench-
ing cannot be attributed to hydrogen donation, since DMSO
is a poor hydrogen donor. In agreement with this, free-radi-
cal formation is suppressed by DMSO. The lack of radical
products also excludes an efficient electron transfer, that is,
the formation of QCꢀ and a DMSO radical cation. The latter
would be expected to react with water yielding the COH ad-
duct, which has been shown[85] to give rise to methanesulfin-
ic acid and a methyl radical. Methanesulfinic acid (or its oxi-
dation product methanesulfonic acid) is not formed, not
even in low yields (F<0.03). Moreover, the methyl radical
is known to react readily with Q yielding 2-methyl-1,4-ben-
zoquinone.[86] We have also looked for this product, and did
not find any indication for its formation. This is in contrast
to a recent study on the photolysis of 2-methyl-1,4-benzo-
quinone in aqueous DMSO solutions.[60] There, the forma-
tion of methyl radicals was concluded based on fluorescent
material formed in the presence of a nitroxyl radical trap.[60]
446
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 440 451

Photohydroxylation of 1,4-Benzoquinone
440 451
in free-radical disproportionation reactions, for example,
with CQH or itself.
Formation of 1,2,4-trihydroxybenzene and its oxidation by
1,4-benzoquinone: Although Ph(OH)3 cannot be formed via
the free-radical pathway, it is a likely intermediate, when
the monomeric triplet, Q*, reacts with water [reaction (14)
followed by the enolization reaction (15)].
Figure 9. UV photolysis (symbols) and pulse radiolysis (lines) of 1,4-ben-
zoquinone (Q) in deoxygenated aqueous solutions. UV/Vis absorption
spectra of the intermediates present 1 ms after the laser flash/electron
pulse. Solid line ([Q]=1.8 mm) and triangles (4 mm) neutral solutions,
dotted line (pH 3.3, [Q]=0.1 mm), circles (pH 2, [Q]=4 mm).
The oxidation of Ph(OH)3 by Q gives rise to the final
products, HOQ and QH2 [reaction (8)]. An oxidation of the
precursor of Ph(OH)3, Q(H2O), by Q [reaction (16)]must
also be envisaged (for discussion see below).
At pH 7, the decay of Q*/Q2* occurs at a similar time
scale as the deprotonation of CQH. Optically, this is reflected
as an increase in the absorption at 315 nm (Figure 10), the
only wavelength where the products (semiquinone-type rad-
icals) show a stronger absorption than the triplets (cf.
Figure 5). As expected, at low pH, where this deprotonation
no longer takes place, no increase at 315 nm is observed on
this time scale.
Using electrochemical oxidation of Ph(OH)3 to generate
HOQ, we have recently shown that HOQ is a fairly strong
acid (pKa 4.2) and that its anion is characterized by a strong
absorption at 482 nm (e=200 m2 molꢀ1), while the free acid
has an absorption maximum at 380 nm (e=140 m2 molꢀ1).[71]
The oxidation of Ph(OH)3 by Q has been used before as a
spectral reference for the formation of HOQ.[19] With the
help of the stopped-flow technique (spectrophotometrical
detection), we now show that the rate of reaction depends
on pH and at a given pH on the phosphate buffer concen-
tration. With Ph(OH)3 in large excess over Q, the kinetics
of the formation of HOQ is of (pseudo) first order. The
rapid autoxidation of the anion of Ph(OH)3 prevented us
from extending our studies to pH values above those given
in Figure 11. These data are compatible with a slow oxida-
tion of the fully protonated Ph(OH)3 (at low pH) and a
much faster one of its monoanion (pKa 9.1)[91] From the data
one estimates that Ph(OH)3 is oxidized with k=14mꢀ1 sꢀ1
and its anion with k=8î105 mꢀ1 sꢀ1 (for a compilation of
rate constants see Table 1). However, the situation is more
complex. In acid solution, the rate of reaction increases
Figure 10. 308 nm laser flash photolysis of 1,4-benzoquinone (4 mm) in
aqueous solution at pH ~7. Normalized absorptions at 288 nm and at
315 nm as a function of time.
As discussed above, after a few ms only CQH remains. It
decays by second-order kinetics into Q and QH2 [reac-
tion (13)].
ꢀ
again, and upon addition of H2PO4 (pH 4.66) the rate in-
creases with increasing phosphate buffer concentration
(inset in Figure 11). Under these conditions, the rate of reac-
ꢀ
tion is described by kobs =(19+170î[H2PO4 ]) mꢀ1 sꢀ1. A
maximum in the rate of reaction between pH 3 and 5.5, as
has been reported before,[19] is not apparent from our data.
The increase in the rate of the oxidation of Ph(OH)3 by
Q at low pH is somewhat unexpected. Two potential inter-
pretations are envisaged. At low pH, Q and its protonated
form, QH+, are in equilibrium, and QH+ is a stronger oxi-
The self-termination rate constant of CQH is 1.09î
109 mꢀ1 sꢀ1, while QCꢀ self-terminates somewhat more slowly
(1.7î108 mꢀ1 sꢀ1).[89] As a consequence of the rapid oxidation
of CQ(OH)H by Q, CQ(OH)H is not reduced into Ph(OH)3
447
Chem. Eur. J. 2004, 10, 440 451
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. von Sonntag et al.
FULL PAPER
that the free-radical pathway requires a reaction of Q* with
Q. The fast component also gains in importance upon lower-
ing the pH. Since this change in mechanism is not connected
with a change in the rate of the decay of Q*/Q2*, it cannot
be due to an increase of the rate of radical formation upon
protonation of Q2*. We, therefore, tentatively suggest that
Q(H2O) is oxidized by Q [reaction (16)]in competition with
its rearrangement to Ph(OH)3 [reaction (15)]. Being of
higher energy than Ph(OH)3, Q(H2O) may be oxidized by
Q (H+-catalyzed, again via a CT complex) much faster than
Ph(OH)3. In case this process occurred at times <1 s, it
would be recorded together with the free-radical pathway.
Our setups do not allow us to monitor times between 1 ms
and 1 s, and testing this mechanistic proposal experimentally
was not possible.
Figure 11. Observed second-order rate constant of the oxidation of 1,2,4-
*
trihydroxybenzene by 1,4-benzoquinone as a function of pH ( ) and in
~
the presence of 1m NaH2PO4 ( ). Data by stopped-flow. The solid line is
calculated on the basis on the values given in the text. Inset: rate of the
phosphate catalyzed oxidation as a function of the phosphate concentra-
*
tion at pH 4.66, laser flash photolysis LFP-LTRD ( ) and stopped-
Reactions involving phosphate: In the presence of phos-
phate buffer, the overall product quantum yield is increased,
and a new product, attributed to 2-phosphato-1,4-benzoqui-
none QP is formed (see above). Its yield increases with in-
creasing phosphate concentration. Concomitantly, the Q
triplets as monitored at 285 nm decay faster with ~4î
106 mꢀ1 sꢀ1 at 0.3 mm Q and ~5î106 mꢀ1 sꢀ1 at 8.3 mm Q (data
not shown). At both Q concentrations, the addition of phos-
phate seems to increase only little if at all the free radical
yield. We hence suggest, that phosphate mainly reacts with
Q*, and a reaction with Q2* must be minor in comparison.
In analogy to reactions (14) and (15), the reaction of Q*
~
flow ( ).
dant than Q. The proton affinity of Q has not yet been de-
termined, but the pKa of acetoneH+ which may serve as a
guide is at ꢀ3.06.[92] Alternatively, Q and Ph(OH)3 could be
in equilibrium with a CT complex, a kind of quinhydrone.
Protonation of this ™quinhydrone∫ would speed-up the oxi-
dation of Ph(OH)3 by Q. We favor the latter mechanistic
concept, because at pH 4.66 phosphate buffer (cf. triangle in
Figure 11) would not increase the QH+ concentration in
equilibrium but could protonate the ™quinhydrone∫.
Above, it has been shown that in the free-radical pathway
a few ms after the flash CQH and HOQ are the products (in
the absence of phosphate buffer; for the effect of phosphate
buffer see below). After this, radicals CQH decay by dispro-
portionation into QH2 and Q [reaction (10)]. This reaction
is completed within a few ms.
ꢀ
with H2PO4 is given by reactions (17) and (18).
Thus, any formation of HOQ observed in the post-milli-
second time range must be due to another process, that is,
the non-radical pathway [reactions (14) (16), (8)]. As shown
in Figure 12, a biphasic buildup of HOQ (for its UV spec-
trum see inset) is indeed observed. The slow part follows
the very characteristic pH dependent kinetics of the oxida-
tion of Ph(OH)3 by Q as discussed above.
The subsequent oxidation of QH2P by Q to QP [reac-
tion (19)]is analogous to that of Ph(OH)3 to HOQ [reac-
tion (8)].
The ratio of the fast/slow components increases with in-
creasing Q concentration (data not shown), substantiating
In the presence of 1 molar phosphate at pH 4.66, follow-
ing the buildup of HOQ, a further buildup between 300 and
450 nm is observed (Figure 13). Subtraction of the spectrum
measured after 5 s from that after 98 s yields a spectrum
with lmax =361 nm (when put on an energy scale, this peak
shows a perfect Gaussian distribution allowing the exact de-
termination of its maximum). It agrees with the spectrum of
QP observed by HPLC.
Figure 12. 308 nm laser flash photolysis of 1,4-benzoquinone (0.5 mm) at
pH 2.5. The laser flash was directed into a cuvette of a diode array spec-
trophotometer, and spectra were recorded every 0.2 s. Main graph: kinet-
ics at 390 nm. Inset: Spectra at 1 s (g) and at 39 s (c).
The oxidation of QH2P by Q is notably slower (kobs
=
7mꢀ2 sꢀ1) than that of Ph(OH)3 (kobs =190mꢀ2 sꢀ1) under
equal conditions (1m phosphate).
448
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 440 451

Photohydroxylation of 1,4-Benzoquinone
440 451
ciplex with water may proceed directly to QCꢀ and CQOH.
CQOH isomerizes rapidly (k=2.5î106 sꢀ1) into CQ(OH)H.[71]
In the presence of phosphate buffer, Q* decays faster
without enhancing the free-radical yield [reaction (17)],
eventually giving rise to QP [reactions (18) and (19)].
Figure 13. Laser flash photolysis (LFP-LTRD) at 308 nm of 1,4-benzoqui-
none (5.5 mm, air-saturated) in the presence of 1m H2PO4ꢀ. a) 5 s after
the flash (assigned to 2-hydroxybenzoquinone); b) 98 s after the flash;
c) difference spectrum (assigned to 2-phosphato-1,4-benzoquinone).
Summary of the photohydroxylation mechanism: It has
been shown above that no free COH is formed. The interme-
diacy of an COH complexed to QCꢀ as has been proposed re-
cently[60] has likewise to be rejected not only on the basis of
the above experiments that exclude the formation of signifi-
cant amounts of methyl radicals in the presence of DMSO
but also as a potential concept in general. p-Complexes of
radicals to aromatic compounds are known, and even for a
radical as reactive as COH it has been shown that such a
complex has to be postulated as an intermediate in its reac-
tion with benzene to the hydroxycyclohexadienyl radical.[93]
However, a p-complex of a strongly oxidizing radical such
as COH with a strongly reducing radical such as CQH/QCꢀ (as
postulated in ref. [60]) will immediately collapse either by
electron transfer of by recombination. We therefore have to
come up with another mechanism.
Scheme 1.
This mechanism has to accommodate two important as-
pects: i) the increase in rate of the decay of the triplet with
increasing Q concentration and ii) the observation that the
free-radical yield increases with increasing Q concentration
despite the fact that the product quantum yields are inde-
pendent of the Q concentration. Furthermore, the incorpo-
ration of phosphate in the presence of phosphate buffer
with a concomitant increase in the total quantum yield of
product formation has to be taken into account.
Conclusion
The formation of an exciplex, Q2*, en route to QH2 and
HOQ via free radicals has been investigated in the present
paper. Exciplexes with Q* seem to be more general. The
photoinduced hydroxylation of benzene by Q in aqueous
solution[64] will have a Q* benzene exciplex and upon its
decay a hydroxycyclohexadienyl radical as the most likely
precursors. The quenching of Q* by halide ions, Xꢀ, is
thought to pass through an exciplex, Q*ꢀXꢀ, which must
have a certain lifetime, since it gives rise to radical forma-
tion at high Clꢀ concentrations (Q*ꢀXꢀ+Xꢀ!QCꢀ+X2Cꢀ).[62]
Exciplexes are postulated for the quenching of Q* by dioxy-
gen and by DMSO, but these reactions do not lead to ob-
servable products such as quenching of Q* by Xꢀ at low Xꢀ
concentrations. Exciplexes are, of course, not restricted to
quinones but an abundant phenomenon in photochemistry.
In analogy to the present system, we have recently shown
that in photochemistry of maleimide in aqueous solution the
maleimide exciplex gives not only rise to a dimer biradical
but also to the formation of free radicals via electron trans-
fer.[96]
Observation ii) requires that excited Q has two pathways
for its decay towards the products QH2 and HOQ, one at
low Q concentrations without the intermediacy of free radi-
cals and one at high Q concentrations via free-radical inter-
mediates. This can be accounted for if Q* reacts with water
yielding an adduct, Q(H2O) [reaction (14); see Scheme 1,
below], which is oxidized by Q straight away [reaction (16)],
or re-arranges into Ph(OH)3 [reaction (15)]and is then oxi-
dized by Q to HOQ [reaction (8)]. In competition with this,
Q* reacts with ground-state Q and the ensuing exciplex Q2*
+
+
decays by electron transfer into QCꢀ and QC . QC is expect-
ed to react ™immediately∫ with water. Benzene-derived radi-
cal cations have lifetimes >1 ns only when substituted by
more than two (electron donating) methyl substituents,[94,95]
and QC+ with its two electron-withdrawing carbonyl groups
may be even shorter-lived. One also may consider that the
radical cation never materializes, and the reaction of the ex-
449
Chem. Eur. J. 2004, 10, 440 451
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. von Sonntag et al.
FULL PAPER
[38]A. I. Ononye, J. R. Bolton, J. Phys. Chem. 1986, 90, 6270 6274.
[39]A. I. Ononye, A. R. McIntosh, J. R. Bolton, J. Phys. Chem. 1986, 90,
6266 6270.
[40]I. Loeff, S. Goldstein, A. Treinin, H. Linschitz, J. Phys. Chem. 1991,
95, 4423 4430.
Acknowledgement
This paper is dedicated to D. Schulte-Frohlinde, mentor and friend, who
raised our interest in the photochemistry of quinones. The continuing
support by R. Mehnert, vice director at the IOM, is highly appreciated.
We thank our colleague W. Knolle for his assistance with the pulse radi-
olysis experiments.
[41]E. Lipczynska-Kochany, J. R. Bolton, Environ. Sci. Technol. 1992,
26, 2524 2527.
[42]R. Das, P. K. Radha, B. Venkataraman, Proc. Indian Acad. Sci.
Chem. Sci. 1993, 105, 273 277.
[43]I. Amada, M. Yamaji, M. Sase, H. Shizuka, J. Chem. Soc. Faraday
Trans. 1995, 91, 2751 2759.
[44]A. E. Alegria, A. Ferrer, E. Sepulveda, Photochem. Photobiol. 1997,
66, 436 442.
[45]A. E. Alegria, A. Ferrer, G. Santiago, E. Sepulveda, W. Flores, J.
Photochem. Photobiol. A 1999, 127, 57 65.
[1]H. Klinger, Ber. Dtsch. Chem. Ges. 1868, 19, 1862 1870.
[2]H. Klinger, Liebigs Ann. Chem. 1888, 249, 137 146.
[3]M. C. Depew, J. K. S. Wan in The Chemistry of Quinoid Compounds
Vol. II (Eds.: S. Patai, Z. Rappoport), Wiley, New York, 1988,
pp. 963 1018.
[4]A. V. Et×sov, O. P. Studzinskii, V. M. Grebenkina, Russ. Chem. Rev.
1977, 46, 93 114.
[46]T. Yoshihara, M. Yamaji, T. Itoh, H. Shizuka, T. Shimokage, S. Tero-
Kubta, Phys. Chem. Chem. Phys. 2000, 2, 993 1000.
[47]S. Bÿarnais-Barbry, R. Bonneau, A. Castellan, Photochem. Photo-
biol. 2001, 74, 542 548.
[48]M. J‰ger, J. R. Norris, J. Magn. Reson. 2001, 150, 26 34.
[49]H. Gˆrner, Photochem. Photobiol. 2003, 77, 171 179.
[50]F. Dˆrr, Z. Elektrochem. 1960, 64, 580 582.
[51]F. D. Lewis, X. Liu, J. Liu, S. E. Miller, R. T. Hayes, M. R. Wasielew-
ski, Nature 2000, 406, 51 53.
[52]H. I. Stonehill, B. Mooney, Chem. Ind. 1961, 1309 1310.
[53]A. D. Broadbent, J. Chem. Soc. Chem. Commun. 1967, 382 383.
[54]G. O. Phillips, N. W. Worthington, J. F. McKellar, R. R. Sharpe, J.
Chem. Soc. Chem. Commun. 1967, 835.
[55]G. O. Phillips, N. W. Worthington, J. F. McKellar, R. R. Sharpe, J.
Chem. Soc. A 1969, 767 773.
[5]J. M. Bruce in The Chemistry of the Quinoid Compounds–Part 1
(Ed.: S. Patai), Wiley, London, 1974, pp. 465 538.
[6]I. H. Leaver, in Photochemistry of Dyed and Pigmented Polymers
(Eds.: N. S. Allen, J. F. McKellar), Applied Science Publishers,
London 1980, pp. 161 245.
[7]A. Berthoud, D. Porret, Helv. Chim. Acta 1934, 17, 694 701.
[8]P. A. Leighton, G. S. Forbes, J. Am. Chem. Soc. 1929, 51, 3549 3561.
[9]F. Poupe, Coll. Czech. Chem. Commun. 1949, 12, 225 236.
[10]J. L. Bolland, H. R. Cooper, Proc. R. Soc. London Ser. A 1954, 225,
405 426.
[11]B. Atkinson, M. Di, Trans. Faraday Soc. 1958, 54, 1331 1339.
[12]N. K. Bridge, G. Porter, Proc. R. Soc. London Ser. A 1958, 225,
259 275.
[13]N. K. Bridge, G. Porter, Proc. R. Soc. London Ser. A 1958, 225,
276 288.
[56]K. P. Clark, H. I. Stonehill, J. Chem. Soc. Faraday Trans. 1 1972, 68,
577 590.
[14]C. F. Wells, Disc. Faraday Soc. 1960, 29, 219 234.
[15]C. F. Wells, Trans. Faraday Soc. 1961, 57, 1719 1731.
[16]C. F. Wells, Trans. Faraday Soc. 1961, 57, 1703 1718.
[17]G. O. Phillips, P. Barber, T. Rickards, J. Chem. Soc. 1964, 3443
3450.
[18]D. Schulte-Frohlinde, C. von Sonntag, Z. Phys. Chem. (Frankfurt/
Main, Germany) 1965, 45, 314 327.
[19]K. C. Kurien, P. A. Robins, J. Chem. Soc. B 1970, 855 859.
[20]D. R. Kemp, G. Porter, Proc. R. Soc. London Ser. A 1971, 326, 117
130.
[57]K. P. Clark, H. I. Stonehill, J. Chem. Soc. Faraday Trans. 1 1972, 68,
1677 1686.
[58]A. D. Broadbent, H. B. Matheson, R. P. Newton, Can. J. Chem.
1975, 53, 828 830.
[59]I. Loeff, A. Treinin, H. Linschitz, J. Phys. Chem. 1983, 87, 2536
2544.
[60]A. Pochon, P. P. Vaughan, D. Gan, P. Vath, N. V. Blough, D. E.
Falvey, J. Phys. Chem. A 2002, 106, 2889 2894.
[61]P. Mazellier, M. Bolte, J. Photochem. Photobiol. A 1996, 98, 141
147.
[21]A. D. Broadbent, R. P. Newton, Can. J. Chem. 1972, 50, 381 387.
[22]B. E. Hulme, E. J. Land, G. O. Phillips, J. Chem. Soc. Faraday Trans.
1 1972, 68, 2003 2012.
[23]S. Hashimoto, K. Kano, H. Okamoto, Bull. Chem. Soc. Jpn. 1972,
46, 966 967.
[24]K. Kano, T. Matsuo, Chem. Lett. 1973, 1127 1132.
[25]K. Ikemizu, M. Orita, M. Sagiike, S. Morooka, Y. Kato, J. Chem.
Eng. Jpn. 1987, 20, 369 374.
[26]H. M. Vyas, S. K. Wong, B. B. Adeleke, J. K. S. Wan, J. Am. Chem.
Soc. 1975, 97, 1385 1387.
[27]J. Boiden Pedersen, C. E. M. Hansen, H. Parbo, L. T. Muus, J.
Chem. Phys. 1975, 63, 2389 2405.
[28]M. Shirai, T. Awatsuji, M. Tanaka, Bull. Chem. Soc. Jpn. 1975, 48,
1329 1330.
[29]G. G. Wubbels, W. J. Monaco, D. E. Johnson, R. S. Meredith, J. Am.
Chem. Soc. 1976, 98, 1036 1037.
[30]G. Bonhomme, J. Lemaire, C. R. Acad. Sci. Ser. IIc 1978, 287, 353
356.
[31]S. Noda, T. Doba, T. Mizuta, M. Miura, H. Yoshida, J. Chem. Soc.
Perkin Trans. 2 1980, 61 64.
[32]J.-C. Ronfard-Haret, R. V. Bensasson, E. Amouyal, J. Chem. Soc.
Faraday Trans. 1 1980, 76, 2432 2436.
[33]I. Forsskahl, R. Gustafsson, B. Nyberg, Acta Chem. Scand. 1981,
B35, 389 394.
[34]S. M. Beck, L. E. Brus, J. Am. Chem. Soc. 1982, 104, 4789 4789.
[35]S. M. Beck, L. E. Brus, J. Am. Chem. Soc. 1982, 104, 1103 1104.
[36]M. C. Depew, L. Zongli, J. K. S. Wan, J. Am. Chem. Soc. 1983, 105,
2480 2481.
[62]I. Loeff, J. Rabani, A. Treinin, H. Linschitz, J. Am. Chem. Soc. 1993,
115, 8933 8942.
[63]I. Loeff, A. Treinin, H. Linschitz, J. Phys. Chem. 1992, 96, 5264
5272.
[64]L. A. Blyumenfel×d, A. V. Karyakin, V. A. Kuz ’min, G. V. Fomin,
A. K. Chibisov, S. I. Sholina, Izv. Akad. Nauk SSSR Ser. Khim. 1969,
2617.
[65]V. A. Kuzmin, A. K. Chibisov, A. V. Karyakin, Int. J. Chem. Kinet.
1972, 4, 639 644.
[66]J. R. Wagner, J. E. van Lier, L. J. Johnston, Photochem. Photobiol.
1990, 52, 333 343.
[67]J. Geimer, O. Brede, D. Beckert, Chem. Phys. Lett. 1997, 276, 411
417.
[68]J. Geimer, D. Beckert, Chem. Phys. Lett. 1998, 288, 449 458.
[69]J. Geimer, D. Beckert, J. Phys. Chem. A 1999, 103, 3991 3998.
[70]S. I. Sholina, G. V. Fomin, L. A. Blyumenfel×d, Russ. J. Phys. Chem.
1969, 43, 447 449.
[71]M. N. Schuchmann, E. Bothe, J. von Sonntag, C. von Sonntag, J.
Chem. Soc. Perkin Trans. 2 1998, 791 796.
[72]S. Pou, D. J. Hassett, B. E. Britigan, M. S. Cohen, G. M. Rosen,
Anal. Biochem. 1989, 177, 1 6.
[73]L. Eberson, O. Persson, J. Chem. Soc. Perkin Trans. 2 1997, 1689
1696.
[74]S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photochemistry,
2 ed., Marcel Dekker, New York 1993.
[75]J. G. Calvert, J. N. Pitts, Jr., Photochemistry, Wiley, New York 1966.
[76]H. J. Kuhn, S. E. Braslavsky, R. Schmidt, Pure Appl. Chem. 1989, 61,
187 210.
[37]A. Rossi, G. Guyot, P. Boule, C. R. Acad. Sci. Ser. II 1986, 303,
1179 1182.
[77]M. Simic, E. Hayon, Biochem. Biophys. Res. Commun. 1973, 50,
364 369.
450
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 440 451

Photohydroxylation of 1,4-Benzoquinone
440 451
[78]R. Flyunt, O. Makogon, M. N. Schuchmann, K.-D. Asmus, C. von -
Sonntag, J. Chem. Soc. Perkin Trans. 2 2001, 787 792.
[79]R. Flyunt, A. Leitzke, G. Mark, E. Mvula, E. Reisz, J. A. Theruva-
thu, C. von Sonntag, J. Phys. Chem. B 2003, 107, 7242 7253.
[80]J. von Sonntag, J. Photochem. Photobiol. A 1999, 126, 1 5.
[81]J. von Sonntag, W. Knolle, R. Mehnert, J. Photochem. Photobiol. A
2000, 132, 25 27.
[82]G. V. Buxton, C. L. Greenstock, W. P. Helman, A. B. Ross, J. Phys.
Chem. Ref. Data 1988, 17, 513 886.
[83]P. Wardman, J. Phys. Chem. Ref. Data 1989, 18, 1637 1755.
[84]E. Amouyal, R. Bensasson, J. Chem. Soc. Faraday Trans. 1 1976, 72,
1274 1287.
[88]W. S. Jenks, The Spectrum 2001, 14(3), 1 7.
[89]G. E. Adams, B. D. Michael, Trans. Faraday Soc. 1967, 63, 1171
1180.
[90]R. L. Willson, Trans. Faraday Soc. 1971, 67, 3020 3029.
[91]J. Sunkel, H. Staude, Ber. Bunsenges. Phys. Chem. 1968, 72, 567
573.
[92]A. Bagno, R. L. Boso, N. Ferrari, G. Scorrano, J. Chem. Soc. Chem.
Commun. 1995, 2053 2054.
[93]L. Ashton, G. V. Buxton, C. R. Stuart, J. Chem. Soc. Faraday Trans.
1995, 91, 1631 1633.
[94]K. Sehested, J. Holcman, J. Phys. Chem. 1978, 82, 651 653.
[95]C. Russo-Caia, S. Steenken, Phys. Chem. Chem. Phys. 2002, 4,
1478 1485.
[85]D. Veltwisch, E. Janata, K.-D. Asmus, J. Chem. Soc. Perkin Trans. 2
1980, 146 153.
[96]J. von Sonntag, W. Knolle, S. Naumov, R. Mehnert, Chem. Eur. J.
2002, 8, 4199 4209.
[86]D. Veltwisch, K.-D. Asmus, J. Chem. Soc. Perkin Trans. 2 1982,
1147 1152.
Received: May 13, 2003
Revised: July 16, 2003 [F5136]
[87]R. S. Cooke, G. S. Hammond, J. Am. Chem. Soc. 1970, 92, 2739
2745.
451
Chem. Eur. J. 2004, 10, 440 451
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim