Photochemical hydroxylation of salicylic acid with hydrogen peroxide; mechanistic study of substrate sensitized reaction

Article abstract of DOI:10.1135/cccc19961729

The photochemical hydroxylation of salicylic acid (SA) by hydrogen peroxide proceeds via the same intermediate independently of the excited reactant (i.e. SA or H₂O₂). This conclusion is supported by the similar effect of pH on the quantum yields and on the isomer ratio of formed products: 2,3-dihydroxybenzoic acid (2,3-DHB) and 2,5-dihydroxybenzoic acid (2,5-DHB). The product ratio ([2,3-DHB]/[2,5-DHB]) is not affected by H₂O₂ concentration.

Full text of DOI:10.1135/cccc19961729

Photochemical Hydroxylation
1729
PHOTOCHEMICAL HYDROXYLATION OF SALICYLIC ACID WITH
HYDROGEN PEROXIDE; MECHANISTIC STUDY OF SUBSTRATE
SENSITIZED REACTION
1
2
Kamil LANG , Jirina BRODILOVA and Stanislav LUNAK
Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic, 160 00 Prague 6,
Czech Republic; e-mail: 1 lang@iic.cas.cz, 2 lunak@iic.cas.cz
Received June 28, 1996
Accepted August 28, 1996
The photochemical hydroxylation of salicylic acid (SA) by hydrogen peroxide proceeds via the same
intermediate independently of the excited reactant (i.e. SA or H2O2). This conclusion is supported by
the similar effect of pH on the quantum yields and on the isomer ratio of formed products: 2,3-di-
hydroxybenzoic acid (2,3-DHB) and 2,5-dihydroxybenzoic acid (2,5-DHB). The product ratio ([2,3-
DHB]/[2,5-DHB]) is not affected by H2O2 concentration.
Key words: 2-Hydroxybenzoic acid; Hydrogen peroxide; Photochemical hydroxylation.
In our previous papers we have described very pronounced photocatalytic effects of
Fe(III) on photochemical hydroxylation of salicylic acid by hydrogen peroxide1,2. The
mechanism of the photocatalyzed hydroxylation is evident. It consists in the photo-
reduction of Fe(III) to Fe(II) and therefore in photogeneration of the Fenton reagent.
However, the hydroxylation of salicylic acid (SA) takes place in the absence of added
Fe(III) too, although at significantly lower quantum yields. The position of the new
hydroxyl group on the aromatic ring is directed mainly by the effect of –OH group on
the ring electron density and therefore hydroxylation takes place preferably in ortho-
and para-positions. Thus, hydroxylation of SA by H2O2 yields a mixture of 2,3-di-
hydroxybenzoic acid (2,3-DHB) and 2,5-dihydroxybenzoic acid3 (2,5-DHB).
COOH
OH
COOH
OH
COOH
OH
hν
+
+
H O
2
2
(A)
OH
2,3-DHB
HO
SA
2,5-DHB
The photochemical hydroxylation of aromatic ring by hydrogen peroxide is usually
explained in terms of its photodissociation into two hydroxyl radicals with subsequent
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

1730
Lang, Brodilova, Lunak:
attack of the ring4. This mechanism is well understood for the irradiation with wave-
lengths, which are absorbed by the H2O2 itself.
hν
H2O2
+
2 OH•
(B)
(C)
Ar
OH•
ArOH•
The hydroxylation initiated by excitation of aromatic substrate in the presence of
H2O2 has been proposed by Ogata5 for benzoic acid, but the mechanism has not been
unambiguously resolved so far. The question remains still open whether hydroxylation
occurs at all when only substrate (and not H2O2) is excited and what is the mechanism
of this “substrate sensitized” hydroxylation. Aim of this work is to contribute to under-
standing of mechanism of these reactions. SA seems to be suitable substrate for such
study as it possesses good chemical stability, solubility in water, and it absorbs at
longer wavelengths than H2O2 even at relatively low concentrations.
EXPERIMENTAL
Chemicals
Salicylic acid, AR (Reanal, Budapest) was recrystallized from double distilled water. Stabilizer free
hydrogen peroxide, AR (Chemical Works, Sokolov), and chromatography standards 2,3-DHB (Fluka)
and 2,5-DHB (Merck, Schuchardt) were used as received. Double distilled water was used in all ex-
periments; its quality was checked by comparative experiments with water of high purity6.
Apparatus and Procedure
A reaction 10 mm cell was placed in thermostatted (20 °C) aluminium block, the reaction solution
(2.5 ml) was continuously bubbled with nitrogen to prevent competitive oxidation by dioxygen and
to ensure effective stirring. pH was adjusted by addition of 0.5 M HClO4 or 0.1 M K2CO3 and checked
by a pH meter.
A low pressure mercury arc (Spectral lamp Philips 93109 E) served as a source of 254 nm radia-
tion in the hydroxylations initiated by H2O2 photolysis. In most of these experiments the initial con-
centrations of SA and H2O2 were 0.5 mmol l–1 and 50 mmol l–1, respectively. The major part of
incident 254 nm radiation was absorbed by H2O2 (75%). Under the used conditions SA absorbs only
18% of incident radiation due to its comparably low absorbance (ε254 = 414 mol–1 l cm–1).
SA was selectively excited by a high pressure mercury arc HBO 200 W (Osram) equipped with a
100 mm water filter, 300–400 nm band filter UG 1/1.5 g and interference filter UVKSIF 313 nm
(both Zeiss, Jena). The initial concentrations [SA]0 = 8 mmol l–1 and [H2O2]0 = 9.2 mmol l–1 were
chosen to ensure that practically all incident radiation was absorbed by SA (A313 > 3).
At both studied wavelengths (i.e. 254 nm or 313 nm) the calculated overall quantum yields Φ
were related to the total number of photon absorbed by the reaction mixture at the given wavelength:
Φ = ∆c NA/ t Iv(1 − 10−A) ,
(1)
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

Photochemical Hydroxylation
1731
where ∆c is the concentration change (mol l–1) during irradiation time t (s), NA is the Avogadro’s
number, IV is the radiation intensity (quanta s–1 l–1), and A is the absorbance of the reaction mixture
at the radiation wavelength. The radiation intensities Iv, measured by a ferrioxalate actinometer, were 7.8 .
1016 photons l–1 s–1 and 9.1 . 1017 photons l–1 s–1 for the low (254 nm) and high (313 nm) pressure
mercury arc, respectively.
The concentrations of SA and hydroxylation products (i.e. 2,5-DHB and 2,3-DHB) were deter-
mined by HPLC equipped with Separon SGX C18 7 mm reversed phase column (Tessek, Prague), a
HPP 4001 pump (Laboratorni pristroje, Prague) and PU 4025 UV detector (Philips Unicam, Cam-
bridge). The eluting solution was: MeOH–H2O–1 M H3PO4 (1 mol l–1) 53 : 46 : 1. Other possible
products as benzene, phenol, 1,4-dihydroxybenzene, muconic acid have not been detected.
The UV-VIS absorption spectra were recorded on a PU 8720 spectrophotometer (Philips Unicam,
Cambridge). The SA fluorescence quenching experiments were carried out using a Perkin–Elmer LS 50
luminiscence spectrometer. In fluorescence quenching experiments the SA concentration was 1 mmol l–1
(pH 3) so that excitation radiation 355 nm was absorbed only by SA. The fluorescence intensities
were corrected for a dilution effect after addition of aliquots of H2O2.
A laser flash photolysis set-up with an LPX 200 excimer laser (Lambda Physics) was described
previously2.
RESULTS AND DISCUSSION
The Hydroxylation Evoked by Photolysis of H O
2
2
The concentrations of hydroxylation products (i.e. 2,5-DHB and 2,3-DHB) during 254 nm
irradiation related to the SA uptake shown in Figs 1 and 2 illustrate the influence of pH
on the reaction kinetics. At pH < 3.5 the 2,3-DHB and 2,5-DHB formations proceeded
much slower compared with pH ≥ 4.3. The effect of pH on formation of hydroxylation
products is more profound for 2,5-DHB then for 2,3-DHB. The detectable concentra-
tion of 2,5-DHB at pH 2.3 was reached after 180 min irradiation (Fig. 2, curve 2) and
furthermore its yields are even reduced at pH 3.5 when compared with 2,3-DHB (Figs 1
40
A
, %
2
1
2,3
30
20
3
4
5
FIG. 1
The yields of 2,3-DHB related to the
uptake of SA vs irradiation time. A2,3
10
0
=
100[2,3-DHB]t/([SA]0
–
[SA]t);
[SA]0 = 0.5 mmol l–1; [H2O2]0 = 0.05
mol l–1; Iv = 7.8 . 1016 photons l–1 s–1;
λ
irr 254 nm; pH: 1 1.4; 2 2.3; 3 3.5; 4
0
100
200
300
400
500
t, s
4.3; 5 9.4
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

1732
Lang, Brodilova, Lunak:
and 2, curves 3). However, in the pH range 4.3 to 9.4 the yields do not undergo any pH
influence (Fig. 1, curves 4 and 5; Fig. 2, curves 5 and 6). The maximum quantity of
products is reached after 30 min irradiation. The following drop is a result of sub-
sequent photodegradation of both products (namely 2,5-DHB, Fig. 2, curve 6).
pH affects also the isomer ratio Y = [2,3-DHB]/[2,5-DHB], see Table I. Contrary to
it, the initial concentrations of H2O2, which were varied (at pH 3.5) from 7.1 mmol l–1
(Y = 2.41 ± 0.42) to 1 mol l–1 (Y = 2.22 ±0.48), had no effect. Y was not determined at
pH < 3.5 because only trace concentration of 2,5-DHB were formed. In the pH range
4.3–8.3 both isomers are produced nearly at equimolar concentrations and Y is constant
during 300 min of irradiation (Fig. 3, curve 1; Table I). In more alkaline medium, pH 9.4,
TABLE I
The effect of pH on the isomer ratio Y = [2,3-DHB]/[2,5-DHB] (error ±10%)
λirr 254 nma
λirr 313 nmb
pH
Y
pH
Y
3.5
4.3
6.7
8.3
9.4
2.46
1.36
1.12
1.25
1.55
3.5
–
2.47
–
6.5
7.8
–
1.26
1.29
–
a 7.8 . 1016 photons l–1 s–1, [SA]0 = 0.5 mmol l–1, [H2O2]0 = 50 mmol l–1; b 9.1 . 1017 photons l–1 s–1,
[SA]0 = 8 mmol l–1, [H2O2]0 = 9.2 mmol l–1, irradiation time 30 min
30
A
, %
5
4
2,5
20
10
0
3
FIG. 2
The yields of 2,5-DHB related to the
uptake of SA vs irradiation time. A2,5
6
= 100[2,5-DHB]t/([SA]0
– [SA]t);
1
[SA]0 = 0.5 mmol l–1; [H2O2]0 = 0.05
mol l–1; Iv = 7.8 . 1016 photons l–1 s–1;
2
λ
irr 254 nm; pH: 1 1.4; 2 2.3; 3 3.5;
0
100
200
300
400
500
t, s
4 4.3; 5 8.3; 6 9.4
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

Photochemical Hydroxylation
1733
the ratio gradually increases from 1.42 to 4.33 after 300 min irradiation as 2,5-DHB is
more easily photodegradated compared with 2,3-DHB. This has been proven by experi-
ments, in which 2,3-DHB and 2,5-DHB were used instead of SA.
As yields of 2,3-DHB and 2,5-DHB are affected markedly by acidity of reaction
solution, the respective quantum yields are also governed by pH (Fig. 4) exhibiting the
lowest values in acidic solutions pH < 3.5. The reason of decrease of Φ at pH > 8 is not
so clear. The explanation might be found in other mechanism involved, namely that in
alkaline solution (pH > 8) the degradation of aromatic ring instead of hydroxylation
takes place. The Φ dependence smoothes down with prolongation of irradiation time
(Fig. 4, curve 2). The decrease is caused by a subsequent degradation of both products
as follows from the kinetics of product formations (Figs 1 and 2), although the ratio Y
remains constant up to pH 8.3 (Table I). Minor absorption of incident radiation by SA
4
Y
FIG. 3
The isomer ratio
3
Y
=
[2,3-
DHB]/[2,5-DHB] vs irradiation time
for various radiation wavelengths.
2
3
1 [SA]0 = 0.5 mmol l–1; [H2O2]0
=
2
1
0.05 mol l–1; pH 6.7; λirr 254 nm; Iv =
7.8 . 1016 photons l–1 s–1. 2 [SA]0 =
8 mmol l–1; [H2O2]0 = 9.2 mmol l–1;
pH 6.5; λirr 313 nm; Iv = 9.1 . 1017
photons l–1 s–1. 3 [SA]0 = 0.5 mmol l–1;
[H2O2]0 = 0.05 mol l–1; pH 6.7; λirr
300–410 nm; Iv > 1018 photons l–1 s–1
1
0
100
200
300
t, s
0.3
254
Φ
1
0.2
FIG. 4
The effect of pH on the quantum
yields of hydroxylation products.
Φ254 = Φ2,3-DHB + Φ2,5-DHB; [SA]0 =
0.5 mmol l–1; [H2O2]0 = 0.05 mol l–1;
0.1
2
λ
irr 254 nm; Iv = 7.8 . 1016 photons l–1
s–1; irradiation time: 1 30 min; 2 300
0.0
2
4
6
8
min
pH
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

1734
Lang, Brodilova, Lunak:
does not significantly contribute to the hydroxylation yields since the respective quan-
tum yields are about one order of magnitude lower (vide infra, Table II) in comparison
with direct photolysis of H2O2 (Fig. 4).
Described phenomena could be accounted for the fact that the nondissociated SA and
salicylate (dissociated carboxyl group) differ in the reactivity with OH•. At pH < 3,
where nondissociated SA prevails (pKa = 2.97, ref.7), the hydroxylation yields are con-
siderably reduced. Since the reactions of OH• with SA and salicylate are diffusion-con-
trolled8, the only difference has to emerge from the stabilization of an intermediate
radical adduct.
Hydroxylation Evoked by SA Excitation
The hydroxylation of SA by H2O2 initiated by 313 nm radiation (which is absorbed
exclusively by SA) occurs, however, the corresponding quantum yields are signifi-
cantly lower compared with H2O2 excitation (Table II). The effect of pH on Φ313 re-
sembles to the one observed for Φ254. Furthermore, the isomer ratio Y at 313 nm
irradiation matches the one observed for 254 nm radiation (photolysis of H2O2), Table I.
It should be also stated that no hydroxylation products were detected in the absence
of H2O2 at pH 3.5. In neutral solutions, which are more favourable for hydroxylation,
only very low quantum yields of 2,3-DHB and 2,5-DHB (in H2O2 absence) were ob-
served (e.g. Φ313 = (1.4 ± 0.4) . 10–3 at pH 6.5). From this follows that H2O2 acts as a
quencher of SA excited states as only in the presence of H2O2 hydroxylation occurs at
reasonable rate. The low quantum yields of an aromatic ring hydroxylation in the
presence of dissolved oxygen has already been described9. The trapping of liberated
electrons and subsequent formation of OH• or other reactive oxygen species accounts
for this observation.
We have carried out laser flash photolysis and steady-state fluorescence experiments
to clarify which excited states of SA initiate the hydroxylation. The former method,
suitable for monitoring of excited triplet states, have not shown any transient quenched
by H2O2 and dissolved O2. Thus, these states should be excluded. On the other hand,
TABLE II
Quantum yields Φ313 = Φ2,3DHB + Φ2,5DHB in case of SA excitation. [SA]0 = 8 mmol l–1; [H2O2]0 =
9.2 mmol l–1; λirr 313 nm, 9.1 . 1017 photons l–1 s–1, irradiation time 30 min (during first 30 min Φ313
remain fairly constant)
pH
3.5
6.5
7.8
Φ313
(8.2± 1.9) . 10–3
(1.9 ± 0.8) . 10–2
(2.3 ± 0.2) . 10–2
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

Photochemical Hydroxylation
1735
the fluorescence quenching of SA after addition of H2O2 can be processed to the Stern–
Volmer treatment:
F0/F = 1 + KSV[Q] ,
(2)
where F0 and F are the fluorescence intensities in the absence and in the presence of a
quencher concentration [Q]. As follows from Fig. 5 the experimental data fairly fit the
Stern–Volmer equation and the corresponding constant is KSV = 0.45 ± 0.01 mol–1 l. Taking
the SA fluorescence lifetime 0.4 ns (ref.10) into account the quenching rate constant of
1.1 . 109 mol l–1 s–1 can be evaluated. This constant does not approach the diffusion-
controlled rate constant k0 = 7.0 . 109 mol l–1 s–1 given by Smoluchowski relation11. The
dynamic quenching of tryptophan fluorescence by H2O2 has been interpreted by elec-
tron transfer from excited indol ring to H2O2 (ref.12). However, the fluorescence
quenching of SA can be attributed to the homolytic splitting of the peroxo bond –O–O–
producing OH•. The assumption of the same hydroxylation agent, i.e. OH•, at both
irradiation wavelength (254 nm absorbed by H2O2 and 313 nm absorbed by SA) is
supported further by the similar effect of pH on Φ and on the isomer ratio Y. Following
Scheme 1 summarizes the mechanisms of 2,3-DHB formation at different excitation
wavelengths. (The reaction scheme for 2,5-DHB formation can be drawn by analogy with
2,3-DHB.)
The bond energy of HO–OH is 213 kJ mol–1 (ref.13). The energy of 1SA* is at least 272
kJ mol–1 (evaluated from the fluorescence peak of SA in water, λ = 440 nm), so the
salicylic acid sensitized splitting of HO–OH bond is energetically favourable. More-
over, the value (1 – F/F0), which expresses the theoretical quantum yield as the fraction
of quenching events related to the total number of excited singlet states, provides fur-
1.4
F /F
0
1.3
1.2
1.1
FIG. 5
Quenching of SA fluorescence by
H2O2; Stern–Volmer plot. [SA]0 = 1
1.0
mmol l–1;
0.0
0.2
0.4
0.6
0.8
[H O ]
2
2
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

1736
Lang, Brodilova, Lunak:
ther support for the suggested mechanism (Scheme 1). For this estimation the concen-
tration of hydrogen peroxide was considered as constant and equal to the initial one,
because its uptake during irradiation was comparable low. Within experimental error
the theoretical quantum yield 4.2 . 10–3 is in accordance with the overall quantum yield
Φ313 = (8.2 ± 1.9) . 10–3, which was found experimentally at pH 3.5.
254 nm
H O
2 OH
2
2
COOH
OH
COOH
OH
COOH
OH
OH
+
+
OH
+
H O
2
OH
H
OH
H
COOH
OH
COOH
OH
COOH
OH
313 nm
+
H O
2 2
OH OH
COOH
COOH
OH
COOH
OH
OH
+
H O
2
+
OH
OH OH
OH
H
OH
H
SCHEME 1
The authors thank Dr P. Kubat for his valuable help in solutions of problems concerning laser flash
photolysis measurement, and the Grant Agency of the Czech Republic for financial support (Grant No.
203/93/0463).
REFERENCES
1. Lunak S., Muzart J., Brodilova J.: Collect. Czech. Chem. Commun. 59, 905 (1994).
2. Lang K., Lunak S., Sedlak P., Kubat P., Brodilova J.: Z. Phys. Chem. 190, 203 (1995).
3. Grinstead R. R.: J. Am. Chem. Soc. 82, 3472 (1960).
4. Lunak S., Sedlak P.: J. Photochem. Photobiol., A 68, 1 (1992).
5. Ogata Y., Tomizawa K., Yamashita Y.: J. Chem. Soc., Perkin Trans. 2 1980, 616.
6. Lunak S., Sedlak P., Lederer P.: J. Photochem. 39, 249 (1987).
7. CRC Handbook of Chemistry and Physics (R. C. Weast, M. J. Astle and W. H. Beyer, Eds), 68th
ed., p. D-162. CRC Press, Boca Raton 1987.
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

Photochemical Hydroxylation
1737
8. Amphlett C. B., Adams G. E., Michael B. D.: Adv. Chem., Ser. 81, 231 (1968).
9. Joschek H. I., Miller S. I.: J. Am. Chem. Soc. 88, 3273 (1966).
10. Pant D D., Joshi H. C., Bisht P. B., Tripathi H. B.: Chem. Phys. 185, 137 (1994).
11. Lakowicz J. R.: Principles of Fluorescence Spectroscopy, p. 257. Plenum Press, New York 1986.
12. Cavatorta P., Favilla R., Mazzini A.: Biochem. Biophys. Acta. 578, 541 (1979).
13. Murov S. L., Carmichael I., Hug G. L.: Handbook of Photochemistry, 2nd ed. Dekker 1993.
Collect. Czech. Chem. Commun. (Vol. 61) (1996)