Reactions of Phosphate Radicals with Monosubstituted Benzenes. A Mechanistic Investigation

Article abstract of DOI:10.1002/hlca.200390203

The kinetics and reaction mechanism of phosphate radicals with substituted benzenes, PhX (X = OH, Me, H, Cl, MeO, and CHO), were studied by flash photolysis and continuous irradiation of aqueous solutions containing potassium peroxodiphosphate (K4[(PO3)2O2]). The rate constants for the reactions of phosphate radicals with the aromatic contaminants PhX were measured and the reaction intermediates and products detected. A correlation between the logarithm of the rate constant for every substrate with H2PO4 radical, HPO4(1-) radical, and PO4(2-) radical and the logarithm of the dissociation constant K^a for H3PO4, H2PO4(1-), and HPO4(2-) was found. A general mechanism is proposed (Scheme 1), which accounts for the experimental results.

Full text of DOI:10.1002/hlca.200390203

Helvetica Chimica Acta Vol. 86 (2003)
2509
Reactions of Phosphate Radicals with Monosubstituted Benzenes.
A Mechanistic Investigation
by Janina A. Rosso, Paula Caregnato, Vero¬nica C. Mora, Mo¬nica C. Gonzalez*, and Daniel O. Ma¬rtire*
¬
Instituto de Investigaciones FisicoquÌmicas Teoricas y Aplicadas (INIFTA), Facultad de Ciencias Exactas,
Universidad Nacional de La Plata, Casilla de Correo 16, Sucursal 4, (1900) La Plata, Argentina
(fax: 542214254642; e-mail: gonzalez@inifta.unlp.edu.ar, dmartire@inifta.unlp.edu.ar)
Dedicated to Prof. Dr. Silvia E. Braslavsky on the occasion of her 60th birthday
The kinetics and reaction mechanism of phosphate radicals with substituted benzenes, PhX (X ˆ OH, Me,
H, Cl, MeO, and CHO), were studied by flash photolysis and continuous irradiation of aqueous solutions
containing potassium peroxodiphosphate (K4[(PO3)2O2]). The rate constants for the reactions of phosphate
radicals with the aromatic contaminants PhX were measured and the reaction intermediates and products
.
.
detected. A correlation between the logarithm of the rate constant for every substrate with H2PO4 , HPO4 À, and
.
PO4 2À and the logarithm of the dissociation constant Ka for H3PO4, H2PO4À, and HPO42À was found. A general
mechanism is proposed (Scheme 1), which accounts for the experimental results.
1. Introduction.
Numerous constituents of natural and atmospheric waters
contribute to the photochemical and/or thermal production of highly reactive species,
.
.
such as HO and SO4 À radicals, which contribute to the self-cleansing of the ecosphere.
Moreover, aqueous-phase procedures for waste-water treatment are frequently based
.
.
on these radicals as initiators of the oxidation of organic substrates. HO and SO4 À are
able to oxidize most inorganic anions to secondary, less-reactive radicals [1]. In
particular, the reactions with phosphate ions present in the aqueous matrixof
contaminated waters yield phosphate radicals that may have unexpected consequences
in the overall oxidation [2].
In previous articles, we investigated the reactivity of phosphate radicals towards
organic substrates [3 5]. The bimolecular rate constants obtained for the reactions of
.
.
.
2À
several aromatic substrates with the phosphate radicals H2PO4 , HPO4 À, and PO4
correlate with the electron-withdrawing ability of the substituent, as expected from the
electrophilicity of these radicals [3][4]. The proposed mechanism [5] includes both
electron transfer from the aromatics to the phosphate radicals and formation of an
adduct between phosphate radicals and the aromatics, which may eliminate phosphate
ions leading to hydroxycyclohexadienyl radicals or phosphite ions yielding phenoxyl
radicals.
Following this research line in the present work, we measure the rate constants for
the reactions of phosphate radicals with aromatic contaminants PhX (X ˆ OH, Me, H,
Cl, MeO, and CHO) and study in detail the reaction intermediates of these reactions.
This information is of importance for evaluating the potential use of phosphate radicals
as initiators for the destruction of pollutants in contaminated waters and for an in-depth
analysis of the reaction mechanism involved.

2510
Helvetica Chimica Acta Vol. 86 (2003)
For this purpose, we generate phosphate radicals by UV irradiation of peroxodi-
phosphate (ˆ hexaoxo-m-peroxodiphosphate(4À)) anions P2O48À. The three acid-base
forms of these phosphate radicals in water are related by fast equilibria [1 5]. To study
the reactions of each of these forms, experiments were performed under controlled pH
.
.
.
conditions, i.e., pH 4, 7, or 10, where H2PO4 , HPO4 À, or PO4 2À, respectively, are the
main radicals present
.
2À
P2O48À ! 2PO4
‡
‡
ÀH
ÀH
.
.
.
2À
À
H2PO4
HPO4
PO4
„
„
pKa5:7
pKa8:9
2. Results and Discussion. 2.1. Reaction Kinetics. The reactions of the phosphate
.
.
.
2À
radicals H2PO4 , HPO4 À, and PO4
with substituted benzenes (Eqn. 1) are
conveniently studied by flash-photolysis by following the phosphate radical decay rate
as a function of added solute concentration.
0
@
1
A
.
H2PO4
k1
.
À
HPO4
‡ PhX !Products
(1)
.
2À
PO4
Photolysis experiments on K4P2O8 solutions at pH 4.0, 7.1, and 10.0 in the presence of
low concentrations of substrate ([PhX] < 2 ¥ 10À4 m) showed absorption traces at l >
400 nm, recorded immediately after the flash of light, that agree with the spectra of the
phosphate radicals [3 5]. The experimental traces could be well fitted to DA ˆ DA0
exp (À kapp ¥ t) ‡ C (Fig. 1, inset). The very small constant term C (lower than 3% of the
maximum absorbance) is associated with the absorption of longer-lived species, mainly
organic radicals, formed following phosphate radical depletion [3 5]. The calculated
wavelength-independent constant kapp linearly increases with the substrate concen-
tration (see Fig. 1). The slopes of the linear plots in Fig. 1 yield the bimolecular rate
constants k1 [3 5]. Table 1 shows the values of k1 for the different substituted benzenes
studied in this work and for those aromatics for which the rate constants with the three
phosphate radicals are known. All the rate constants determined here fall in the
‡
reported correlation between log k1 and the substituent Hammett parameter s [3 5],
thus implying that the transition state of the reaction may have either significant polar
character or important resonance interaction with the substituents.
A linear correlation between the logarithm of the rate constant k1 for every
.
.
.
2À
substrate given in Table 1 with H2PO4 , HPO4 À, and PO4 and the logarithm of the
dissociation constant Ka for H3PO4, H2POÀ4 , and HPO42À was also found (Fig. 2).
Table 1 also includes the slopes a of the plots in Fig. 2 for the aromatic compounds. The
observed a values lie within the range 0.1 0.2, showing an average value of 0.15 with a
standard deviation of 0.03. The intercepts b of the plots of log k1 vs. log Ka (Fig. 2)
strongly depend on the nature of the substrate. The plot of the intercepts b vs. the
‡
Hammett parameters s yields the straight line shown in the inset of Fig. 2.

Helvetica Chimica Acta Vol. 86 (2003)
2511
&
*
Fig. 1. Plot of kapp vs. [PhX] for anisole (.), benzaldehyde ( ), and chlorobenzene ( ) obtained at pH 10. Inset:
Absorbance profiles obtained at l ˆ 500 nm for 8 ¥ 10À4 m K4P2O8 air-saturated solutions at pH 10 containing 8 ¥
10À6 m (upper trace) and 1.5 ¥ 10À5 m (lower trace) anisole. The solid lines show the first-order-fitting curves.
.
.
À
Table 1. Bimolecular Rate Constants k1 [mÀ1sÀ1] of the Reactions of the Phosphate Radicals H2PO4 À, HPO4
,
.
2À
‡
and PO4 with Substituted Benzenes PhX, Substituent s
Values, and Slopes a (see text)
min
a
‡
PhX
k1
s
)
a
min
.
.
.
2À
À
H2PO4
ca. 1010 b
HPO4
PO4
Aniline
Phenol
Anisole
Benzyl alcohol
Toluene
)
1.47 ¥ 109 b
)
7.1 ¥ 108 b
)
1.30
0.92
0.78
0.108
e
(6.9 Æ 0.2) ¥ 108 c
(8.5 Æ 1.6) ¥ 109 c
(1.0 Æ 0.1) ¥ 109 g
)
)
)
)
)
)
)
)
(5.3 Æ 0.4) ¥ 108 d
)
)
)
)
)
)
)
)
)
(1.9 Æ 0.1) ¥ 108 f
(1.3 Æ 0.1) ¥ 108 g
(1.4 Æ 0.1) ¥ 107 g
(4.0 Æ 1.0) ¥ 107 c
(1.7 Æ 0.2) ¥ 107 g
(5.9 Æ 0.6) ¥ 107 c
(8.3 Æ 0.6) ¥ 107 g
)
)
)
)
)
)
)
0.203
0.102
0.147
0.114
0.124
0.115
0.149
(5.2 Æ 0.5) ¥ 108 g
(2.1 Æ 0.8) ¥ 108 c
(1.4 Æ 0.1) ¥ 107 g
(1.3 Æ 0.1) ¥ 107 c
0.31
Benzaldehyde
Benzene
(8.9 Æ 0.5 ) ¥ 107 g
(4.3 Æ 0.3) ¥ 106 g
0.00
0.11
0.57
Chlorobenzene
(3.3 Æ 0.4) ¥ 107 c
(6.9 Æ 0.7) ¥ 106 d
(2.0 Æ 0.1) ¥ 106 c
h
TFT
)
(3.5 Æ 0.5) ¥ 107 g
(2.7 Æ 0.5) ¥ 106 g
(9.0 Æ 1.0) ¥ 105 g
a) From [6]. b) From [7]. c) This work. d) From [3]. e) At pH 10, phenolate is the main species. f) The value
.
À
from [3] (4.6 ¥ 107 mÀ1 sÀ1) was obtained from measurements of the decay rate constant of HPO4 in the
presence of solutions obtained by dilution of the anisole-saturated aqueous solution. The value reported here
was corrected for the solubility of anisole at 258 determined in the present work (2.28 ¥ 10À2 m). g) From [4].
h) a,a,a-Trifluorotoluene.
Phosphate radicals behave towards aromatics as oxidants. The efficiency of an
oxidant does not necessarily depend solely on its redox potential. For instance, the
hydroxyl radical, in spite of being a powerful one-electron oxidant, usually does not
react by electron transfer but by addition [8]. This behavior is caused by the

2512
Helvetica Chimica Acta Vol. 86 (2003)
&
*
Fig. 2. Plot of log k1 vs. log Ka for aniline (~), benzaldehyde ( ), benzene ( ), and chlorobenzene (.). Inset:
‡
Plot of b vs. s .
.
stabilization of the transition state for addition by bond formation. The addition of OH
radicals to substituted benzenes PhX yields the hydroxycyclohexadienyl radicals
.
(HOchdX ). To yield from this adduct the electron-transfer products (the radical
cation and hydroxy anions), heterolysis of the CÀOH bond must occur. However, since
OHÀ is a very poor leaving group, the rate of spontaneous heterolysis is very low. Thus,
the reaction products are typically those obtained by dimerization or disproportiona-
.
tion of the HOchdX radicals. One-electron oxidation products are observed only at
low pH, where an acid catalysis may take place [9][10], i.e., protonation of the OH
group converts the poor leaving group OHÀ (pKa of its conjugate acid H2O, 15.7) to the
‡
excellent leaving group H2O (pKa of its conjugate acid H3O , À 1.7) [8].
Sulfate radicals were proposed to react with aromatics by outer-sphere electron
transfer [11], leading to electron-transfer products. However, Steenken [8], on the basis
of a Marcus treatment and considering the redoxpotentials of the reactants, proposed
that the electron-transfer products could be the result of an addition/elimination
sequence, involving the formation of an adduct followed by sulfate ion elimination.
Assuming Steenken×s mechanism, the different reactivity of hydroxyl and sulfate
radicals lies in the leaving-group abilities of their redoxpartners, SO 24À and HOÀ ions,
respectively. The difference between the pKa values of the conjugate acids, HSOÀ4 (pKa
1.9) and H2O (pKa 15.7) translates into a difference in the lifetimes of both adducts. In
fact, the lifetime of the sulfate radical adduct with benzene was reported to be ꢀ 100 ns
[12][13], while the OH adduct (hydroxycyclohexadienyl radical) lives for milliseconds.
This behavior can be interpreted in terms of the Br˘nsted catalysis law (Eqn. 2),
which expresses the relationship between the acidity constant Ka for an acid HA and its

Helvetica Chimica Acta Vol. 86 (2003)
2513
catalytic constant kcat for a given acid-catalyzed reaction [14]. The values of a for the
reactions studied in the literature are found to be almost less than unity [14]. The plots
shown in Fig. 2 are Br˘nsted correlations (with kcat ˆ k1 in Eqn. 2) that indicate that the
observed rate constant k1 depends on the leaving-group ability of each phosphate ion.
This behavior can be rationalized assuming the formation of a s adduct between the
phosphate radicals and the aromatics and further elimination of phosphate ions.
.
Therefore, the scheme proposed in [3] modifies to Scheme 1 for HPO4 À, where
formation of the substrate radical cation 3 due to an addition/elimination route from
.
substrate 1 via 2 is considered. Similar reactions can be written for the radicals H2PO4
.
and PO4 2À, with elimination of H2PO4À and PO34À ions, respectively, from the
substrateÀradical adduct corresponding to 2.
log kcat ˆ a ¥ log Ka ‡ b
(2)
Scheme 1
Since the phosphate ions are much better leaving groups than OHÀ, a short lifetime
.
.
2À
is also expected for the s adduct 2 or the corresponding adducts of H2PO4 and PO4
radicals, as for those of the sulfate radicals, in agreement with the lack of observation of
these radicals in our experiments. This scheme considers phosphate (HPO24À, H2PO4À,
or PO34À ) loss yielding the radical cation 3 of the aromatics as proposed for the sulfate
.
adducts [8] and HOcdhX 5 as also suggested for the sulfate radical cation [15b].
Phosphite (H2POÀ3 , H3PO3, or HPO32À) elimination yielding phenoxyl radicals 4 was
proposed in our previous investigations [3], in agreement with a similar pathway
reported by Merga et al. [15b] for the sulfate radical.
A kinetics analysis of the mechanism proposed in Scheme 1 assuming the steady-
state condition for the phosphate radical adduct 2, yields Eqn. 3 for k1. This equation
shows the proportionality between k1 and k2, as expected from the Hammett
correlation found for different substrates (vide supra). However, k1 is also affected
by the decomposition rate of the s adduct (through the rate constants kÀ2, k3, k4, and
k5). Moreover, if kÀ2 > (k3 ‡ k4 ‡ k5), k1 is also proportional to k3 ‡ k4 ‡ k5, as expected
.
.
from the observed Br˘nsted correlation, as for each radical H2PO4 , HPO4 À, and

2514
Helvetica Chimica Acta Vol. 86 (2003)
.
2À
PO4
both k4 and k5 depend on the elimination of H2POÀ4 , HPO42À, and PO43À,
respectively.
ꢀ
ꢁ
k3 ‡ k4 ‡ k5
k1 ˆ k2
ꢁ3†
kÀ2 ‡ k3 ‡ k4 ‡ k5
The intercept b of the plots shown in Fig. 2 can be interpreted as the value of log k1 for a
much better leaving group (Ka ˆ 1) than the three phosphate ions. The observed
correlation between b and the Hammett parameters (Fig. 2, inset) further reflects the
dependence of k1 on k2.
2.2. Reaction Intermediates and Products. Flash-photolysis experiments of perox-
odiphosphate solutions in water containing almost saturated concentrations of the
.
.
.
organic substrates show depletion of H2PO4 , HPO4 À, and PO4 2À radicals within less
than 100 ms and the formation of transient species absorbing in the wavelength region
250 500 nm due to the organic radicals formed according to Eqn. 1. As the reactions of
.
2À
the radical PO4 with the aromatics are not very efficient, the traces obtained from
.
solutions at pH 10 show important absorbance contributions due to PO4 2À, thus
hindering the identification of the organic radicals formed at this pH. Consequently,
except for benzaldehyde, most of the absorption spectra of the organic radicals are not
analyzed at pH 10. The observed intermediates and reaction products obtained for each
substrate will be discussed separately. Table 2 depicts the products observed in each
case.
Table 2. Observed Reaction Products of the Reaction of Dihydrogen and Hydrogen Phosphate Radicals with
Substituted Benzenes in Aqueous Solution
PhX
Phosphate radical
Products
.
Benzene
Toluene
H2PO4
phenol (15%)a), CHOCHˆCHÀCOÀCH2OH, biphenylb)
benzaldehyde (11%), p-cresol
.
H2PO4
.
HPO4
benzaldehyde (4%), benzoate (41%)c), benzyl alcohol, bibenzyls
dihydroxybiphenylsd)
À
.
Phenol
Anisole
Chlorobenzene
Benzaldehyde
H2PO4
H2PO4
H2PO4
.
2,2'-dimethoxybiphenyl
.
phenol (3%), 2-chlorophenol, 4-chlorophenol
.
HPO4
benzoate (100%)c)
À
a) The percentages were calculated from a calibration curve with respect to the initial molar concentration of the
substrate PhX. b) Detected only for conversions higher than 44%. c) Isolated as sodium salt. d) Distinction
between the different isomers was not possible.
.
.
Benzene. The intermediates observed for the reactions of H2PO4 and HPO4 À with
benzene (1; X ˆ H) show a very similar absorption around 310 nm, indicating that the
same organic transient is formed for both phosphate radicals. The traces show a fast
decay with a complexkinetics and formation of stable products. The transient spectrum
obtained with N2-saturated aqueous solutions is assigned to the hydroxycyclohex-
.
adienyl radical 5 (X ˆ H) of benzene (OHchdH , lmax 315 nm). Radical 5 (X ˆ H)
reversibly reacts with molecular oxygen yielding peroxyl radical 6 (X ˆ H) (Scheme 2).
Lower concentrations of 5 and faster decay kinetics are observed in air- than in N2-

Helvetica Chimica Acta Vol. 86 (2003)
2515
saturated solutions. This behavior is rationalized by the existence of a reversible
reaction between 5 and O2, with k6 ˆ 3.1 ¥ 108 mÀ1 sÀ1 and kÀ6 ˆ 1.2 ¥ 104 sÀ1 [16].
Scheme 2
To support the presence of these reactions and neglect any possible contribution of
the phosphate adduct to the observed spectrum, computer simulations of the traces
obtained at 310 nm for the air-saturated solutions at pH 4 were done (see Exper. Part
for details). The program considers a set of well established reactions involving
reactions of phosphate radicals, 5 (X ˆ H) and 6 (X ˆ H): the reaction of Eqn. 1 for the
.
.
H2PO4 radical (see Table 1 for k1); the H2PO4 decay reactions in the absence of
benzene [4]; the reactions 5 > 6 (k6 and kÀ6); the recombination reactions of 5 to 7 (k7)
and to 1/8 (k8) (see Scheme 2), with 2(k7 ‡ k8) ˆ 7.3 ¥ 108 mÀ1 sÀ1 [17], in agreement with
the second-order kinetics observed for the decay of 5 in N2-saturated solutions;
recombination of 6, with 2kREC ˆ 9 ¥ 108 mÀ1 sÀ1 [2]; and reactions of 6 to 9 and to 8 with
k9 ‡ k10 ˆ 800 sÀ1 [16]. The simulated concentration profiles of 5 (X ˆ H) and 6 (X ˆ H)
were converted into the corresponding absorbances taking e310 ˆ 2270 [17] and
300 mÀ1 cm-1 for 5 (X ˆ H) and 6 (X ˆ H), respectively. The e310 value for 6 (X ˆ H)
was estimated from the spectrum reported in [18] and e300 ˆ 500 mÀ1cmÀ1 taken from
[16]. Since the radicals 5 (X ˆ H) and 6 (X ˆ H) are the main species contributing to
the absorption traces at 310 nm, the addition of the absorbance contributions of 5 (X ˆ
H) and 6 (X ˆ H) were compared with the experimental traces at this wavelength. A
good agreement between experimental and simulated traces is found, as shown in the
inset of Fig. 3, further supporting the assignment of the traces to the species 5 (X ˆ H)
and 6 (X ˆ H).
Recombination of radicals 5 (X ˆ H) leads to formation of biphenyl 10 (X ˆ H)
observed as a reaction product (see Scheme 2), or disproportionation of 5 (X ˆ H)
yields phenol (8; X ˆ H; also detected) and benzene (1; X ˆ H) with water elimination

2516
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 3. Transient absorption spectra obtained at 400 ms after the flash irradiation of 9.2 ¥ 10À4 m K4P2O8 solutions
containing 5.6 ¥ 10À5 m toluene at pH 4: air-saturated (.) and N2-saturated ( ). Inset: Experimental and simulated
*
(solid line) absorbance profile obtained at 310 nm with 2.36 ¥ 10À3 m K4P2O8 air-saturated solution at pH 4
containing 4.8 ¥ 10À3 m benzene.
[19]. The peroxyl radical 6 (X ˆ H) may rearrange to endoperoxide 9 (X ˆ H) leading
to the formation of open-chain oxidized products in agreement with the detection of
CHOCHˆCHÀCOÀCH2OH (see Scheme 2). Otherwise, peroxyl radical 6 (X ˆ H)
.
eliminates HO2 yielding phenol (8; X ˆ H).
.
Toluene: The intermediates of the reaction of H2PO4 with toluene (1; X ˆ Me)
were analyzed in air- and N2-saturated aqueous solutions. Both in the presence and
absence of molecular oxygen, the absorption spectrum taken 400 ms after the flash of
light shows an absorption band around 280 300 nm and a less-intense one in the 370
430 nm region, as shown in Fig. 3. The traces in the 280 300 nm region decay faster and
are of lower amplitude in the presence of O2. The traces in the 370 430 nm region
show a second-order decay with 2k/e400 ˆ (3.2 Æ 1.5) ¥ 106 cm sÀ1, independently of the
presence of O2.
.
According to Scheme 1, the HOchdMe radical 5 (X ˆ Me) and the phenoxyl
radical 4 (X ˆ Me) of toluene could be formed. Phenoxyl radicals exhibit an absorption
maximum at ca. 300 nm and a less-intense one at around 400 nm, decay by second-
.
order kinetics, and do not react with molecular oxygen [15]. The HOchdMe radical 5
(X ˆ Me) shows an absorption maximum at 315 nm and reversibly reacts with
molecular oxygen, yielding peroxyl radical 6 (X ˆ Me), as already discussed for the
.
HOchdH radical [16]. The absorption band in the 370 430 nm range is assigned to the
phenoxyl radical 4 (X ˆ Me) of toluene. Moreover, taking e400 ꢂ 1600 mÀ1 cmÀ1 as for
other phenoxyl radicals, we obtain 2k ˆ 5.1 ¥ 109 mÀ1 sÀ1, which is of the same order as
the value reported for the unsubstituted phenoxyl radical 4 (X ˆ H; 2.3 ¥ 109 mÀ1 sÀ1
[21]).

Helvetica Chimica Acta Vol. 86 (2003)
2517
The benzyl radical 11 absorbing in the 250 290 nm range [20] can be formed by
‡
dehydration of 5 (X ˆ Me) or by H loss from the radical cation 3 (X ˆ Me) [9] and
reacts with molecular oxygen with a rate constant k ˆ 2.8 ¥ 109 mÀ1 sÀ1 leading to the
peroxyl radicals 12 (Scheme 3). Since the rate constants for the reactions of the three
phosphate radicals with toluene fall in the corresponding Hammett correlations and, for
this substrate, the straight line of the Br˘nsted correlation has a slope similar to those
.
for the other PhX, formation of the benzyl radical 11 through H abstraction from the
methyl group of toluene must not be a relevant reaction pathway.
Scheme 3
The ratio of the initial absorbances observed at 300 and 400 nm in N2-saturated
solutions, DAo300/DAo400, is higher than that expected for the phenoxyl raxical 4 (X ˆ
Me) [15], thus indicating that other radicals, such as 5 (X ˆ Me) and benzyl radical 11,
contribute to the 280 330 nm absorption band. In fact, formation of benzaldehyde and
p-cresol (ˆ4-methylphenol) as reaction products support the participation of these
radicals. Benzaldehyde (1; X ˆ CHO) can arise from peroxyl radical 12 generated by
reaction of the benzyl radical 11 with molecular oxygen (see Scheme 3). Formation of
p-cresol (8; X ˆ Me) is explained by disproportionation of 5 (X ˆ Me) to 1/8 (X ˆ
Me), as shown in Scheme 2.
.
À
A similar absorption spectrum was obtained for the reaction of HPO4 with
toluene. The reaction products benzyl alcohol (13), benzaldehyde (1; X ˆ CHO),
benzoic acid, and bibenzyl 14 are all formed either by the oxidation of the benzyl
radical 11 or from its recombination (Scheme 3).
.
Phenol. The intermediate formed by the reaction between HPO4 À and phenol (1;
X ˆ OH) with lmax at 300 and 400 nm and a second-order-decay rate constant 2k ˆ 1.7 ¥
109 mÀ1 sÀ1 was assigned to the phenoxyl radical 4 (X ˆ H) [3][22]. A similar spectrum
.
was found for the intermediate of the reaction of H2PO4 with phenol in air-saturated
solutions. This unsubstituted phenoxyl radical 4 (X ˆ H) may be formed either by
dihydrogen phosphate elimination from the s adduct 2 (X ˆ OH), as shown in
.
À
Scheme 4 for the HPO4 reaction, or by the fast phosphate-catalyzed dehydration of
.
the dihydroxycyclohexadienyl radical 5 (HOchdOH ) [23]. Formation of the dihy-

2518
Helvetica Chimica Acta Vol. 86 (2003)
.
droxybiphenyl 10 (X ˆ OH) observed as products for the reactions of H2PO4 with
phenol can be explained by C,C coupling of the unsubstituted phenoxyl radical4 (X ˆ H)
and tautomerization (see Scheme 4) [24].
Scheme 4
The hydroxyphenoxyl radical 4 (X ˆ OH) with lmax around 320 and 420 nm [25],
which could be formed from the s adduct 2 according to Scheme 1, is not observed, in
agreement with the lack of detection of quinone and hydroquinone, formed after
hydroxyphenoxyl radical disproportionation [21].
.
.
À
Anisole: The intermediates formed by the reaction of H2PO4 and HPO4 with
anisole (1; X ˆ MeO), 400 ms after the flash of light, show two absorption bands at
280 290 nm and 360 450 nm and a shoulder at 320 nm, shown in Fig. 4 for the radical
.
obtained from HPO4 À. Due to the electron-donating character of the MeO group,
stabilization of the radical cation 3 (X ˆ MeO) of anisole is expected. The radical
cation 3 (X ˆ MeO) shows two absorption bands with maxima at 280 nm (e280
ˆ
7400 mÀ1 cmÀ1) and at 430 nm (e430 ˆ 3200 mÀ1 cmÀ1) [9]. The traces taken at 280 nm
Fig. 4. Transient absorption spectrum obtained at 400 ms after the flash irradiation of air-saturated 9.2 ¥ 10À4
m
K4P2O8 solutions containing 1.8 ¥ 10À5 m anisole at pH 7. Inset: Corresponding absortion spectra of the products
at pH 4 ( ) and pH 7 (- - - -).

Helvetica Chimica Acta Vol. 86 (2003)
2519
and in the 360 450 nm region follow a second-order decay both at pH 4 and 7. Taking
the reported values of e280 and e430 for the radical cation 3 (X ˆ MeO) [9], 2k ˆ (2.5 Æ
0.8) ¥ 109 mÀ1 sÀ1 is estimated in line with the value reported in [9] (2k ˆ 4 ¥ 109 mÀ1 sÀ1).
The traces at 320 nm show different decay kinetics, thus supporting the formation of
.
a second intermediate. Formation of the HOchdOMe radical 5 (X ˆ MeO) also
absorbing at this wavelength (lmax ˆ 320 nm and e320 ˆ 3400 mÀ1 cmÀ1 [9]) must not be
.
neglected. In acid medium, the HOchdOMe radical 5 (X ˆ MeO) reversibly
dehydrates, yielding the radical cation 3 (see Scheme 1) and, consequently, a higher
contribution of this radical is expected at lower pH [26]. In fact, the traces obtained at
pH 4 show higher initial absorbances than those at pH 7. The absorption spectra of the
products formed at pH 4 and 7 are coincident (inset of Fig. 4). This spectrum
corresponds to 2,2'-dimethoxybiphenyl, which was detected at pH 4 by GC/MS.
Formation of dimethoxybiphenyls 10 (X ˆ MeO) was reported to be due to
recombination of the radical cation 3 (X ˆ MeO) [9].
.
À
Chlorobenzene. The reaction of HPO4 with chlorobenzene (1; X ˆ Cl) leads to
formation of phenoxyl radical 4 (X ˆ Cl) [3] showing two absorption bands around
290 300 nm and in the 390 400 nm region, as shown in Fig. 5. Formation of the
phenoxyl radical 4 (X ˆ Cl) means that the reaction 2 ! 4 in Scheme 1 is an important
reaction route for this substrate.
Fig. 5. Transient absorption spectra obtained at 400 ms after the flash irradiation of air-saturated 9.2 ¥ 10À4
m
)
K4P2O8 solutions containing 8.9 ¥ 10À4 m chlorobenzene at pH 4 (.) and 1.7 ¥ 10À3 m chlorobenzene at pH 7 (
*
[3]
.
The reaction of H2PO4 with chlorobenzene shows an intermediate with absorption
maximum at 315 nm (also shown in Fig. 5) whose decay kinetics and spectrum in air-
.
saturated aqueous solutions are coincident with those reported [3] for the HOchdCl
radical 5 (X ˆ Cl) under identical experimental conditions and supported by formation
of phenol and 2- and 4-chlorophenol as reaction products. Disproportionation of
radical 5 (X ˆ Cl) with loss of water yields chlorophenol 8 (X ˆ Cl) and chlorobenzene
(1; X ˆ Cl) (for the o-, m-, and p-chloro-substituted radicals 5, see Scheme 2). Loss of

2520
Helvetica Chimica Acta Vol. 86 (2003)
HCl from the ipso-chloro-substituted isomer 5 yields phenol (8; X ˆ H), as shown in
Scheme 5 [27]. The absence of the band at 400 nm indicates that formation of the
phenoxyl radical 4 (X ˆ Cl) is of no importance at this pH.
Scheme 5
Benzaldehyde. The transient absorption spectrum obtained for the reaction of
.
H2PO4 with benzaldehyde (1; X ˆ CHO) in the absence of molecular oxygen in
aqueous solution, shows absorption maxima around 370 and 425 nm (Fig. 6) in
agreement with that reported for the a,a-dihydroxybenzyl radical 15 (−hydrated×
benzoyl radical) [19b][28]. Moreover, from a second-order-decay kinetic analysis and
the e value reported for the hydrated benzoyl radical 15, 2k ˆ (1.2 Æ 0.1) ¥ 109 mÀ1 sÀ 1 is
obtained, in line with [29] (> 8 ¥ 108 mÀ1 sÀ1). Any contribution of the naked benzoyl
radical to the observed spectrum is of no significance because it shows a different
absorption spectrum with a very low e value (e ˆ 150 mÀ1 cmÀ1 at 370 nm [30]).
Fig. 6. Transient absorption spectra obtained at 400 ms after the flash irradiation of Ar-saturated 9.2 ¥ 10À4
m
P2OÀ8 4 solutions containing 2.5 ¥ 10À5 m benzaldehyde at pH 4 (.) and 3.8 ¥ 10À4 m P2OÀ8 4 solutions containing
3.1 ¥ 10À5 m benzaldehyde at pH 7 ( ). Published absorption spectrum of the benzoyl radical (solid line) and
*
benzoyl radical anion (dotted line) [28].
In aqueous solutions aldehydes easily yield a 1,1-diol (benzaldehyde hydrate),
.
which is more easily oxidized than the parent aldehyde. Since H2PO4 reacts with
.
aliphatic alcohols mainly by H-abstraction [29], H2PO4 could abstract a H atom from

Helvetica Chimica Acta Vol. 86 (2003)
2521
the hydrated form of benzaldehyde (which has a CÀH bond dissociation energy of
88 kcal/mol) yielding an a,a-dihydroxybenzyl radical 15 [31] in equilibrium with its
basic form, the corresponding radical anion 16, as shown in Scheme 6. However, since
the rate constants for the reactions of benzaldehyde with the three phosphate radicals
fall in the correlation shown in Eqn. 2 with a slope similar to those found for the other
substituted benzenes (Fig. 2 and Table 1), the mechanism shown in Scheme 1 is also
expected for this substrate. Formation of the hydrated benzoyl radical 15 is very likely
due to phosphate elimination from the 2-(phosphatoxy)cyclohexadienyl radical 2
(X ˆ CHO) to yield a radical cation 3 (X ˆ CHO; see Scheme 1), which further
deprotonates, yielding the hydrated benzoyl radical 15.
Scheme 6
.
.
2À
À
The transient of the reaction of HPO4 and PO4
with benzaldehyde in Ar-
saturated solutions shows a broad absorption maximum around 430 nm coincident with
that of the anion 16 of the a,a-dihydroxybenzyl radical [19b][28] obtained by
deprotonation. From the second-order-decay kinetics and the reported e values for the
anion 16, 2k ˆ (1.4 Æ 0.5) ¥ 109 mÀ1 sÀ1 and (1.2 Æ 0.1) ¥ 109 mÀ1 sÀ1 are obtained at pH 7
and 10, respectively. These values are higher than that reported (7 ¥ 108 mÀ1 sÀ1 [28]) at
lower ionic strength, as expected from the ionic-strength effect on the recombination
rate of two negatively charged ions.
Further disproportionation of the radical anion 16 leads to benzaldehyde (1; X ˆ
CHO) and benzoate (17); the latter, isolated as sodium salt, was the sole reaction
product (Scheme 6).
3. Conclusions. The presented results support Scheme 1. An addition pathway
yields a phosphate-radical adduct 2 of the aromatic substrate. Such radical adducts are
expected to be unstable (as is the case of the sulfate radical adducts [13], vide supra)
.
and rapidly eliminate phosphate ions, yielding the radical cation 3 and the HOchdX
radicals 5, respectively. The radical cations 3 are stabilized in acid medium as observed
for anisole (1; X ˆ MeO). Otherwise, they decay to a more-stable radical, as may be the
case of the formation of the −hydrated× benzoyl radical 15 for the reaction with
benzaldehyde (1; X ˆ CHO). Radicals 5 can also be formed by reversible hydration of

2522
Helvetica Chimica Acta Vol. 86 (2003)
.
the radical cations 3 [1][10] (Scheme 1). Evidence for the formation of the HOchdX
.
.
À
radical 5 from the reactions of PhX with H2PO4 and HPO4 was found here for the
substrates benzene (X ˆ H), toluene (X ˆ Me), chlorobenzene (X ˆ Cl; only for the
.
reaction with H2PO4 ), and anisole (X ˆ MeO). The formation of phenoxyl radicals 4
.
of toluene and chlorobenzene (only for the reaction with HPO4 À) can be explained by
phosphite loss from the phosphate adducts 2.
.
The correlation shown in Fig. 2 supports that the different k1 values for H2PO4 ,
.
.
2À
HPO4 À, and PO4
with every substrate could lie in the different leaving-group
abilities of their redoxpartners, i.e., H2POÀ4 , HPO42À, and PO34À, respectively. This
.
argument is further supported by the observation of the HOchdCl radical 5 (X ˆ Cl)
and the phenoxyl radical 4 (X ˆ Cl) as intermediates for the reactions of chlorobenzene
.
.
with the radicals H2PO4 and HPO4 À, respectively. Formation of 5 (X ˆ Cl) in the
.
reaction of H2PO4 with chlorobenzene indicates the preferential loss of H2POÀ4 , which
is a better leaving group than HPO42À
.
Experimental Part
General. Tetrapotassium peroxodiphosphate (ˆtetrapotassium hexaoxo-m-peroxodiphosphate(4À)) was
»
electrochemically obtained as described in [4]. Chlorobenzene (J. T. Baker, p.a.), toluene (Riedel-de-Haen,
p.a.), sodium benzoate (Fluka, p.a.), CHCl3 (Roth, p.a.), benzene (Merck, p.a.), benzaldehyde (Merck, p.a.),
phenol (Fluka, p.a.), anisole (Fluka, p.a.), KOH (Riedel-de-Hae»n, p.a.), Na2HPO4 ¥ 2 H2O (Merck, p.a.),
NaH2PO4 ¥ H2O (Aldrich, p.a.), and H3PO4 (Merck, p.a.) were used as received. Distilled water (>18 MW cmÀ1
,
< 20 ppb of org. C) was obtained from a Millipore system. The pH was adjusted with mixtures H3PO4/NaH2PO4,
.
.
.
NaH2PO4/Na2HPO4, or K2HPO4/KOH to study the reactions of H2PO4 , HPO4 À, or PO4 2À. The ionic strength
of the solns. was within the range 0.1 0.2m.
Time-Resolved Experiments. Flash-photolysis experiments were carried out in a conventional apparatus
(Xenon Co. model 720C, 10 ms fwhm pulse) with modified optics and electronics [32]. The optical-path length
was 20 cm. The emission of the flash lamps was filtered with sat. aq. solns. of the substrates in the water jacket of
the cell to avoid photolysis of the aromatic compounds. The analysis source was a high-pressure mercury lamp
(Osram HBO-100 W). To avoid product accumulation, each soln. was irradiated only once. When necessary,
signals arising from single shots were averaged.
Steady-State Experiments and Product Analysis. Product analysis was performed in samples irradiated
continuously. In the steady-state experiments, phosphate radicals were produced by UV photolysis of the
peroxodiphosphate anion. A cylindrical low-pressure Hg lamp (Heraeus, MNNI 35/20, Germany) emitting at
254 nm was used. The photochemical reactor was of annular geometry (volume: 260 ml). The experimental
Â
Ã
conditions ( P2O84À ˆ 1 m m and 0.2 m m ꢀ [PhX]o ꢀ 0.6 m m) were chosen to avoid possible PhX photolysis.
0
Solns. of the aromatics were prepared by dilution of sat. aq. solns. at 258.
Samples were taken at constant time intervals, and the total volume of sample withdrawn was always less
than 5% the volume of the reactor. To avoid secondary reactions, products were analyzed at irradiation times
lower than 6 min. Under these conditions, substrate consumption was kept below 15% in all cases. The reaction
products formed after the reaction of phosphate radicals with the substituted benzenes were determined either
by GC/MS or IC (ion chromatography).
For GC/MS analyses, org. substrates were extracted from the aq. solns. with a fixed volume of CHCl3. The
extracts were stored at 48 in glass vials with poly(tetrafluoroethene)/silicone septum-lined screw caps and
minimized head space. GC/MS: HP-6890 chromatograph equipped with a fused silica HP5-MS GC capillary
column and coupled to an HP5973 mass-selective detector; temp. program from 308 to 3208; He as carrier gas,
flow rate 29 ml sÀ1; injection volume, 2 ml. Aliphatic compounds of low molecular mass, i.e., containing four C-
atoms or less, are eluted with the solvent in GC/MS determinations and could not be detected with our
chromatographic set-up.
Anions evolution was performed by IC (Dionex DX 500) with suppressed conductivity detection: 4 Â
250 mm Ionpac AS11 (Dionex) column, 4 Â 50 mm Ionpac AG11 (Dionex) guard column, and 5 Â 35 mm

Helvetica Chimica Acta Vol. 86 (2003)
2523
(5 mm) Ionpac NG1 (Dionex) guard column; eluent gradient program, flow rate, 1 ml minÀ1; injection volume,
10 ml.
Kinetics Computer Simulations. A computer program based on the numerical resolution of the differential
equations system by the third-order Runge Kutta method [32] was used to simulate the experimental absorbance
profiles at 310 nm for experiments at pH 4 in the presence of 4.8 Â 10À3 m benzene. The program considers the
.
.
pulsed excitation as a delta function, producing H2PO4 radicals. The concentration of H2PO4 immediately after
.
excitation, [H2PO4 ]o ˆ 1.3 ¥ 10À6 m, was taken as an input parameter, estimated from experiments under
identical experimental conditions but in the absence of benzene.
¬
¬
¬
This research was supported by the Agencia Nacional de Promocion Cientifica y Tecnologica (ANPCyT,
¬
¬
Argentina). J. A. R. thanks the Consejo Nacional de Investigaciones Cientificas y Tecnicas (CONICET,
Argentina), for a graduate studentship. M. C. G. is a research member of CONICET. D. O. M. is a research
member of the Comision de Investigaciones Cientificas de la Provincia de Buenos Aires (CIC), Argentina.
¬ ¬
¬
REFERENCES
[1] A. B. Ross, W. G. Mallard, W. P. Helman, G. V. Buxton, R. E. Huie, P. Neta, −NDRL-NIST Solution
Kinetics Database: Ver. 3.0×, Notre Dame Radiation Laboratory, Notre Dame, IN and National Institute of
Standards and Technology, Gaithersburg, MD, 1998; http://www.rcdc.nd.edu.
[2] P. Maruthamuthu, P. Neta, J. Phys. Chem. 1978, 82, 710.
¬
[3] S. S. Cencione, M. C. Gonzalez, D. O. Martire, J. Chem. Soc., Faraday Trans. 1998, 94, 2933.
¬
[4] J. A. Rosso, F. J. RodrÌguez Nieto, M. C. Gonzalez, D. O. Martire, J. Photochem. Photobiol., A: Chemistry
1998, 116, 21.
¬
[5] D. O. Martire, M. C. Gonzalez, Prog. React. Kinet. Mechan. 2001, 26, 201.
[6] J. March, in −Advanced Organic Chemistry, Reactions, Mechanisms, and Structure×, Wiley-Interscience,
New York, 1991, p. 278.
[7] P. Subramanian, V. Ramakrishnan, J. Rajaram, J. C. Kuriacose, Proc. Indian Acad. Sci. (Chem. Soc.) 1986,
97, 573.
[8] S. Steenken, −One Electron RedoxReactions between Radicals and Organic Molecules. An Addition/
Elimination (Inner-Spere) Path×, in Top. Curr. Chem., Vol. 177, −Electron Transfer II×, Ed. J. Mattay,
Springer, Berlin 1996, p. 125.
[9] J. Holcman, K. Sehested, J. Phys. Chem. 1976, 80, 1642.
[10] H. Mohan, J. P. Mittal, J. Phys. Chem. 1995, 99, 6519; H. Mohan, J. P. Mittal, J. Phys. Chem. A, 1999, 103,
379; H. Mohan, J. P. Mittal, Chem. Phys. Lett. 1995, 235, 444.
[11] P. Neta, V. Madhavan, H. Zemel, R. W. Fessenden, J. Am. Chem. Soc. 1977, 99, 163.
[12] S. Steenken, in −Free Radicals in Synthesis and Biology×, NATO ASI Series C 260, Ed. F. Minisci, Kluwer
Academic Publishers, Dordrecht, 1989, p. 220.
[13] S. Steenken, J. Chem. Soc., Faraday Trans. 1 1987, 83, 113.
[14] E. S. Gould, −Mechanism and Structure in Organic Chemistry×, Holt, Rinehart & Winston, New York, 1959,
p. 113.
[15] a) J. Dzengel, J. Theurich, D. W. Bahnemann, Environ. Sci. Technol. 1999, 33, 294; b) G. Merga, C. T.
Aravindakumar, B. S. M. Rao, H. Mohan, J. P. Mittal, J. Chem. Soc., Faraday Trans. 1994, 90, 597.
[16] C. von Sonntag, H.-P. Schuchmann, in −The Chemistry of Free Radicals: Peroxyl Radicals in Aqueous
Solutions×, Ed. Z. B. Alfassi, John Wiley & Sons Ltd., Chichester, 1997, p. 173.
[17] G. V. Buxton, J. R. Langan, J. R. Lindsay Smith, J. Phys. Chem., 1986, 90, 6309.
[18] L. M. Dorfman, I. A. Taub, R. E. B¸hler, J. Chem. Phys. 1962, 36, 3051.
[19] a) M. K. Eberhardt, J. Phys. Chem. 1974, 78, 1795; b) J. A. Rosso, S. G. Bertolotti, A. M. Braun, D. O.
¬
Martire, M. C. Gonzalez, J. Phys. Org. Chem. 2001, 14, 300.
[20] K. Sehested, H. Corfitzen, H. C. Christensen, E. J. Hart, J. Phys. Chem. 1975, 79, 310.
[21] G. N. R. Tripathy, R. H. Schuler, Chem. Phys. Lett. 1982, 88, 253.
[22] M. N. Schuchmann, E. Bothe, J. von Sonntag, C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 1998, 791.
[23] E. J. Land, M. Ebert, Trans. Faraday Soc. 1967, 63, 1181.
[24] F. Berho, R. Lesclaux, Chem. Phys. Lett. 1997, 279, 289.
[25] E. J. Land, J. Chem. Soc., Faraday Trans. 1993, 89, 803.
[26] P. O×Neill, S. Steenken, K. Sehested, D. Schulte-Frohlinde, J. Phys. Chem.1975, 79, 2773 and ref. cit. therein.

2524
Helvetica Chimica Acta Vol. 86 (2003)
[27] H. Mohan, M. Mudaliar, C. T. Aravindakumar, B. S. Madhav Rao, J. P. Mittal, J. Chem. Soc., Perkin Trans.
2 1991, 1387.
[28] M. Simic, M. Z. Hoffman, J. Phys. Chem. 1972, 76, 1398.
[29] P. Maruthamuthu, P. Neta, J. Phys. Chem. 1977, 81, 1622.
[30] H. Fischer, R. Baer, R. Hany, I. Verhoolen, M. Walbiner, J. Chem. Soc., Perkin Trans.2 1990, 787.
[31] K. B. Wiberg, F. Freeman, J. Org. Chem. 2000, 65, 573.
¬
¬
[32] M. L. Alegre, M. Gerones, J. A. Rosso, S. G. Bertolotti, A. M. Braun, D. O. Martire, M. C. Gonzalez, J.
Phys. Chem. A 2000, 104, 3117.
Received December 3, 2002