The chain-breaking antioxidant activity of phenolic compounds with different numbers of O-H groups as determined during the oxidation of styrene

Article abstract of DOI:10.1002/kin.20377

The technique based on monitoring oxygen, consumption was applied to test 18 polyphenols (PP) and model phenolics as a chain-breaking antioxidant during the oxidation of styrene initiated by 2,2′-azobis(2,4- dimethylvaleronitril) at 37°C. The chain-breaking capability of PP was characterized by two parameters: the rate constant k₁ for the reaction of antioxidants with the peroxy radical produced from styrene and the stoichiometric coefficient of inhibition, f, which shows how many kinetic chains are terminated by one molecule of PP. Rate constants k₁ × 10⁵ (in M^-1 s^-1) were found to be 10 (catechol), 27 (pyrogallol), 34 (3,6-di-tert-Bucatechol), 4.3 (protocatechic acid), 12 (gallic acid), 15 (caffeic acid), <0.01 (chrysin), 1.3 (kaempferol), 19 (quercetin), 5.3 (baicalein), 16 (epicatechin), 32 (epigallocatechin), 9.0 (dihydroquercetin), 3.3 (resveratrol), and 16 (nordihydroguaiaretic acid). The value of k₁ increases when going from one to two and three adjacent O-H groups in a benzene ring (catechol and pyrogallol derivatives, respectively). At the same time, two O-H. groups in metaposition in a A-ring of flavonoids actually do not participate in the inhibition. For the majority of PP, f is near to 2 independent of the number of OH groups. The correlation of k₁ with the structure of PP and the O-H bond dissociation enthalpy has been discussed.

Full text of DOI:10.1002/kin.20377

The Chain-Breaking
Antioxidant Activity
of Phenolic Compounds
with Different Numbers of
O-H Groups as Determined
During the Oxidation
of Styrene
IVAN TIKHONOV,1 VITALY ROGINSKY,2 EVGENY PLISS1
1Yaroslavl State University, Sovetskaya St. 14, 150000 Yaroslavl, Russia
2N. Semenov Institute of Chemical Physics, Russian Academy of Sciences, Kosygin St. 4, 119991 Moscow, Russia
Received 8 April 2008; revised 28 June 2008; accepted 18 July 2008
DOI 10.1002/kin.20377
Published online in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT: The technique based on monitoring oxygen consumption was applied to test 18
polyphenols (PP) and model phenolics as a chain-breaking antioxidant during the oxidation of
styrene initiated by 2,2ꢀ-azobis(2,4-dimethylvaleronitril) at 37◦C. The chain-breaking capability
of PP was characterized by two parameters: the rate constant k1 for the reaction of antioxidants
with the peroxy radical produced from styrene and the stoichiometric coefficient of inhibition,
f, which shows how many kinetic chains are terminated by one molecule of PP. Rate constants
k1 × 105 (in M−1
s
−1) were found to be 10 (catechol), 27 (pyrogallol), 34 (3,6-di-tert-Bu-
catechol), 4.3 (protocatechic acid), 12 (gallic acid), 15 (caffeic acid), <0.01 (chrysin), 1.3
(kaempferol), 19 (quercetin), 5.3 (baicalein), 16 (epicatechin), 32 (epigallocatechin), 9.0 (di-
hydroquercetin), 3.3 (resveratrol), and 16 (nordihydroguaiaretic acid). The value of k1 increases
when going from one to two and three adjacent O-H groups in a benzene ring (catechol and
pyrogallol derivatives, respectively). At the same time, two O-H groups in metaposition in a
A-ring of flavonoids actually do not participate in the inhibition. For the majority of PP, f is
near to 2 independent of the number of OH groups. The correlation of k1 with the structure
of PP and the O H bond dissociation enthalpy has been discussed. ꢁ 2008 Wiley Periodicals,
C
Inc. Int J Chem Kinet 41: 92–100, 2009
Correspondence to: Vitaly Roginsky; e-mail: rogin@postman.
ru.
c
ꢀ
2008 Wiley Periodicals, Inc.

ANTIOXIDANT ACTIVITY OF POLYPHENOLS
93
MATERIALS AND METHODS
INTRODUCTION
The commercial-grade sample of styrene was purged
from antioxidants by using an alumina adsorption col-
umn with the subsequent distillation. A free radical ini-
tiator, 2,2ꢀ-azobis(2,4-dimethylvaleronitrile) (AMVN)
was purchased from Polysciences Inc. Catechol, gal-
lic acid, caffeic acid, epicatechin, and nordihydrogu-
iaretic acid were purchased from Aldrich, and pyrogal-
lol, 4-hydroxybenzoic acid, p-cumaric acid, chrysin,
baicalein, resveratrol were obtained from Sigma.
Kaempferol was purchased from Fluka, quercetin and
dihydroquercetin from Serva, and myricetin and epi-
gallocatecgin from Carl Roth. 3,6-di-tert-Bu-catechol
was a gift from A. Wasserman. Other reagents used
were of the highest available quality. Antioxidants
were added to the testing system as a stock solution
in chlorobenzene or in a mixture of acetonitrile with
chlorobenzene, depending on the solubility.
Polyphenols (PP), i.e., molecules having several hy-
droxyl groups on aromatic rings, are common con-
stituents of foods of plant origin and are among major
antioxidants of our diet. The main dietary sources of
polyphenols are fruits and beverages such as tea, red
wine, cacao, and coffee. The beneficial health effects
of fruits and above-mentioned beverages against ag-
ing, cancer, and cardiovascular diseases are associated
basically with PP antioxidant activity, i.e., with the ca-
pability of the PP of deactivating active free radicals
[1–4]. It is generally agreed that the most principle
reaction is that of PP (QH2) with the peroxy radicals
LO•2
LO•2 + QH2 → LOOH + QH• k1
(1)
The kinetics of oxygen consumption during styrene
oxidation initiated by AMVN was studied with a glass
capillary microvolumometer of high sensitivity with
a cell construction that allowed addition of required
components without opening the cell [14]. All the runs
Reaction (1) is basically responsible for the antioxidant
activity of PP. The antioxidant activity of PP, first of all
flavonoids, has been the objective of numerous studies
[5–10]. However, the majority of these studies are con-
ducted by using indirect methods and the information
reported is only of a qualitative or semiquantitative in
nature.
were conducted at 37.0
0.1◦C under air. The vol-
ume of the reaction mixture varied within the range
from 0.5 to 1 mL. A kinetic run was started with
measuring the rate of the oxidation in the absence
of antioxidants, R0, and the rate of initiation, RIN.
RIN was varied by alteration of [AMVN] and was
determined by the inhibitor method with 6-hydroxy-
2,2,5,7,8-pentamethylchromane (HPMC) as a refer-
ence inhibitor. RIN was calculated from the induction
period of inhibited oxidation, tIND, using Eq. (1)
The most adequate characteristic of the antioxidant
activity is the rate constant k1. Only for a few PP, reli-
able values of k1 have been reported in the literature.
The most justified approach to k1 determination is that
based on the application of the kinetic model of con-
trolled chain oxidation. The key point of this approach
is a constant and controlled rate of the initiation (active
free radical generation). The latter is achieved, as a rule,
by the application of thermolabile azo-compounds. The
antioxidant activity of the chain-breaking antioxidants
may be characterized by two independent parameters,
k1 and f , the stoichiometric coefficient of inhibition,
which shows how many kinetic chains can be termi-
nated by one molecule of antioxidant. The theoretically
justified procedure for the determination of k1 and f
has been developed originally for monophenols [11–
13], and then it was modified for PP by the examples
of p-hydroquinones [14] and natural PP [15]. In this
work, the subject for study was the antioxidant activity
of the derivatives of catechol and pyrogallol during the
oxidation of styrene. Styrene seems to be the best sub-
strate for determining k1. Along with other things, in
this case reaction (1) is not complicated by a specific
interaction of PP with a solvent and the value of k1 de-
termined may be considered as a genuine measure of
the antioxidant reactivity. The structures of the antiox-
idants studied in this work are presented in Scheme 1.
RIN = 2 · [HPMC]/tIND
(1)
RESULTS AND DISCUSSION
Choosing Solvents
The majority of PP shows a very poor solubility in non-
polar solvents including styrene and chlorobenzene.
Therefore, polar solvents are usually applied to pre-
pare PP stock solutions. This presented a considerable
challenge to the experiments. The fact is that many
polar solvents form H-bonds with PP that results in
significant reduction of the PP activity in reaction (1)
[13,16]. This is a reason why the rate constants k1
for PP reported in the literature are usually underes-
timated [13]. In this work, it was shown that even a
very small addition of DMSO, one of the best solvents
for PP, to the reaction mixture significantly decreased
International Journal of Chemical Kinetics DOI 10.1002/kin

94
TIKHONOV, ROGINSKY, AND PLISS
OH
R1
R1
R3
R1
R2
OH
HOOC
OH
HOOC
R2
R1 = R2 = H
1
R1 = R2 = R3 = H
5
6
7
8
9
R1 = H
2
R1 = OH; R 2 = R3 = H
R1 = R2 = OH; R 3 = H
R1 = OH; R 2 = R3 = t-Bu
R1 = OH; R 2 = H
R1 = R2 = OH
R1 = OH
3
4
R1
10 R1 = R2 = R3 = R4 = R5 = H
R2
R3
11 R1 = R3 = R5 = H; R2 = R4 = OH
12 R1 = R2 = R4 = OH; R 3 = R5 = H
B
O
HO
R5
A
13 R1 = R2 = R3 = R4 = OH; R 5 = H
14 R1 = R2 = R3 = R4 = H; R5 = OH
R4
OH
O
O
OH
OH
OH
B
OH
R1
B
HO
O
O
HO
A
A
OH
OH
OH
OH
15 R1 = H
17
16 R1 = OH
HO
OH
Me
HO
HO
Me
OH
OH
HO
18
19
Scheme 1
the experimental value of k1 as compared with styrene
or the styrene–chlorobenzene mixtures. For example,
in the case of catechol (PP 2) the addition of 0.012
M DMSO decreases k1 by 2.5-fold, with caffeic acid
(PP 9), 0.0056 M DMSO caused a reduction of k1 ca.
5-fold, with quercetin (PP 12), with 0.014 M DMSO
the k1 was reduced by a factor of 3. To overcome this
problem, DMSO in PP stock solutions was changed
to acetonitrile (AN). Acetonitrile was reported to dis-
play a relatively moderate tendency to form H-bonds
with phenolics and to reduce their reactivity [17]. In
this work, the amount of AN in the reaction mixture
never exceeded 0.5% (0.12 M). It was shown that such
a concentration of AN did not reduce k1 as compared
to that measured in styrene.
Scheme 2 [12,13] slightly modified for the case, when
a chain-breaking antioxidant is a hydroquinone (QH2)
[14]. Scheme 2 is presented for a “standard” peroxida-
tion of substrate LH when hydroperoxide is a principle
product that was formed by reaction (2) (for instance,
polyunsaturated lipid). Meanwhile, during the oxida-
tion of styrene chain propagation occurs as the addition
of the peroxy radical to the double bond of styrene, re-
sulting finally in the formation of polyperoxide rather
(0) AMVN + (LH + O2) → LO2• + products
RIN
k2
k3
k1
k4
•
(2) LO2• + LH + (O2) → LOOH + LO2
•
(3) LO2• + LO2
→
products
(1) LO2• + QH2 → LOOH + QH•
•
(4) QH• + LO2
→
LOOH + Q
(5) QH• + QH• → QH2 + Q
k5
Kinetic Scheme
The oxidation of styrene (LH) initiated by AMVN and
inhibited by PP may be described by the classic kinetic
Scheme 2 Oxidation of LH initiated by AMVN and inhib-
ited by QH2.
International Journal of Chemical Kinetics DOI 10.1002/kin

ANTIOXIDANT ACTIVITY OF POLYPHENOLS
95
has been reported elsewhere [14,19]. The coefficient n
in the denominator of Eqs. (2) and (3) (1 or 2) depends
on the fate of QH•. If QH• terminates in reaction (4),
n = 1; when reaction (5) predominates, n = 2. Equa-
tion (2) allows the calculation of k1/k2 directly from
[O2] trace. This way is suitable for rather reactive QH2
when the concentration of QH2 decreases significantly
with time during a kinetic run (the “dynamic” method).
It should be noted that the application of Eq. (2) does
not require the use of RIN and absolute concentration
of QH2. If the reactivity of QH2 is relatively low and
its concentration does not change significantly during
a kinetic run, k1/k2 can be determined from the plot of
the starting value of R vs. [QH2] by using Eq. (3) (the
“static” method). Equation (1) can be applied to deter-
mine the f value for QH2, if [HPMC] is changed for
[QH2] and the coefficient 2 is changed for the variable
f value
O
R
O
R
O
R
OH
R
H
H
+
R
O
R
HO
Scheme 3
than hydroperoxide. However, from point of view of
formal kinetics, this does not matter as all the kinetic
equations are the same for both cases.
Attention should be drawn to the fact that both re-
actions (4) and (5) in Scheme 2 are presented as the
disproportionation. The reason for that is the O–H bond
in hydroxyl-substituted phenoxyls is rather weak [17],
which makes reactions (4) and (5) thermodynamically
profitable. This is in contrast to the reactions with par-
ticipation of the phenoxy radical PhO• formed from a
monophenol PhOH
f = RIN · tIND/[QH2]
(4)
PHO• + LO•2 → Quinolide peroxide
PhO• + PhO• → Products
(4a)
(5a)
The determination of tIND presents some difficulties.
The “graphical” procedure commonly employed to de-
termine tIND from experimental [O2] traces has no the-
oretical basis and is generally incorrect [14,15]. To
circumvent these difficulties, it has been suggested to
determine tIND as the integral [14,15]
While reaction (4a) occurs unambiguously as the re-
combination with formation of quinolide peroxides
[18], reaction (5a) may occur by both the mechanism
of recombination or disproportionation depending on
the PhO• structure [12]. The recombination is typi-
cal of PhO• in which the p-position or at least one of
o-positions in the benzene ring is not substituted. The
recombination occurs originally with formation of ke-
todimers followed by the enolization [12], for instance
(Scheme 3). It should be noted that a final product of
this process, dimer, has active O-H groups that can
participate again in the reaction with LO•2 .
ꢀ
0
ꢁ
ꢂ
tIND
=
1 − (R/R0)2 dt
(5)
∞
Only this procedure was applied in this work.
Kinetic Parameters Characterizing
the Chain-Breaking Capability of PP
and Some Monophenols
Calculation of k1/k2 and f from [O2] Traces
The examples of kinetic runs aimed at determination
of k1/k2 are depicted in Fig. 1 (the dynamic method)
and Fig. 2 (the static method). As seen from Fig. 1, the
plot of F1 vs. time is a straight line as it is predicted
by Eq. (2). The plot of F2 vs. QH2 concentration is
also a straight line (Fig. 2) in accordance with Eq. (3).
The value of k1/k2 can be calculated from the slope of
these lines by using Eq. (2) or (3), respectively. In both
cases, the absolute value of k1 can be calculated from
k1/k2, assuming that k2 is equal to 57 M−1 s−1 [20].
The values of k1 determined in this work are listed in
Table I. In all the cases, the coefficient n in Eq. (2) or
(3) was assumed to be equal to 2 (see below).
The reactivity of QH2 to LO•2 is determined from the
competition between reactions (1) and (2), and it is
originally given as the k1/k2 ratio. The calculation of
k1/k2 can be performed by using Eqs. (2) and (3)
1 + R/R0
1 − R/R0
R0
R
k1R0
F1 = In
−
=
t + constant (2)
nk2[LH]
R0
R
2k1R0
nk2[LH]RIN
F2 =
−
=
[QH2]
(3)
R
R0
where R and R0 are the rates of inhibited and noninhib-
ited oxidation, respectively; [QH2] is the starting con-
centration of QH2; [LH] is the concentration of styrene
(it is commonly very high and remains almost constant
during a kinetic run). The deduction of Eqs. (2) and (3)
In Table I, stoichiometric coefficients f calculated
for the more reactive antioxidants by using Eqs. (5)
and (4) are also presented. As for the less active
International Journal of Chemical Kinetics DOI 10.1002/kin

96
TIKHONOV, ROGINSKY, AND PLISS
3
3.0
1.0
0.8
0.6
0.4
0.2
0.0
0
2.0
4
F
1.0
1
2
0.0
0
20
40
60
80
Time (min)
Figure 1 Kinetics of the oxidation of 8.5 M styrene at 37◦C at RIN = 4 × 10−9 M s−1. Plot 1: [O2] trace for the noninhibited
oxidation; plot 2: [O2] trace for the oxidation inhibited by 1 × 10−5 M pyrogallol (PP 3); plot 3: trace 2 in the axes of Eq. (2);
plot 3: the change of the parameter 1 – (R/R0)2 with time; plot 4: trace 2 in the axes of Eq. (2).
antioxidants studied by the static method, experimen-
tal determination of f was evidently impossible in the
framework of the technique applied. As Table I sug-
gests, f is close to 2 for the majority of antioxidants
studied including flavonoids containing several O-H
groups, for instance 12, 13, 15, and 17. This value is
also typical of numerous monophenols [12,13,21]. It
is indicative that f for PP 19, which contains two cat-
echol moieties, is close to 4 (Table I), i.e., 2 for one
catechol fragment. At the same time, for a few PP, first
of all 6 and 7, the f value is distinctly lower than two.
Earlier this was reported for several p-hydroquinones,
and this phenomenon was explained by the reaction
of the p-hydroxy-substituted phenoxy radicals with
molecular oxygen [14]. Experimentally, this manifests
itself as a significant decrease in f , when RIN is re-
duced [14]. This effect was not observed with the PP
studied in this work (not shown). Starting from these
observations, we can conclude that the contribution
of the reaction of QH• with O2 is very moderate if
any. So the reason for rather low values of f for 6, 7
and some other PP remains unclear. Interestingly, the
f value for some PP (2, 12, 15) determined during the
oxidation of methyl linoleate in an aqueous micellar
solution significantly exceeds 2 [15]. The difference in
f determined in [15] and in the current work is likely
caused by the difference in the nature of the antiox-
idant free radicals participating in the process under
consideration. While in nonpolar styrene we deal with
uncharged free radicals, in aqueous systems at neutral
pH hydroxyl-substituted phenoxy radicals are depro-
tonated converting into a radical-anion.
4
3
2
1
0
6
5
4
3
2
1
0
2
1
Many PP, especially PP containing two or three ad-
justed OH groups, show a very high reactivity in reac-
tion (1). k1 for most active PP, 3, 13, 16, which have
in their structure three adjusted OH groups (Table I),
0
1
2
3
4
5
6
[QH2] (μM)
Figure 2 Kinetics of the oxidation of 8.5 M styrene at 37◦C
is close to k1 for α-tocopherol (3.3 × 106
M
−1s−1
inhibited by p-cumaric acid (PP 8) at RIN = 4 × 10−9 M s−1
.
[13,21]) known as the most active natural chain-
breaking antioxidant.
Plot 1: Plot of the rate of oxidation against [QH2]; plot 2:
plot 1 in the axes of Eq. (3).
International Journal of Chemical Kinetics DOI 10.1002/kin

ANTIOXIDANT ACTIVITY OF POLYPHENOLS
97
Table I Kinetic Parameters Characterizing the Antioxidant Activity of PP and Model Antioxidants
Compound Numbera
PP
Method
k1 (M−1 s−1 b
)
f b
1
2
3
4
5
6
7
8
Phenol
Catechol
Pyrogallol
–
∼2 ×103c
–
Dynamic
Dynamic
Dynamic
Static
Dynamic
Dynamic
Static
(1.0 0.1) × 106
(2.7 0.2) × 106
(3.4 0.3) × 106
(1.3 0.2) × 103
(4.3 0.3) × 105
(1.2 0.1) × 106
(5.0 0.3) × 104
(1.5 0.1) × 106
<103d
2.1 0.2
1.8 0.1
2.0 0.1
nd
1.1 0.1
1.0 0.1
nd
1.5 0.1
nd
>1.4
1.9 0.1
1.6 0.1
2.1 0.1
1.4 0.1
>0.9
1.9 0.1
2.1 0.1
4.1 0.2
3,6-di-tert-Bu-catechol
4-Hydroxybenzoic acid
Protocatechic acid
Gallic acid
p-Cumaric acid
Caffeic acid
Chrysin
Kaempferol
Quercetin
Myricetin
Baicalein
Epicatechin
9
Dynamic
Static
10
11
12
13
14
15
16
17
18
19
Dynamic
Dynamic
Dynamic
Dynamic
Dynamic
Dynamic
Dynamic
Dynamic
Dynamic
(1.3 0.1) × 105
(1.9 0.1) × 106
(2.8 0.2) × 106
(5.3 0.2) × 105
(1.6 0.1) × 106
(3.2 0.2) × 106
(9.0 0.5) × 105
(3.3 0.2) × 105
(1.6 0.1) × 106
Epigallocatechin
Dihydroquercetin
Resveratrol
Nordihydroguaiaretic acid
nd: Not determined.
aSee Scheme 1.
bAveraged from at least four independent experiments.
cReported in [20].
d No inhibition even at 0.2 mM chrysin.
For several PP studied in this work, k1 was earlier
reported in the literature. We shall restrict our con-
sideration to the data obtained during the oxidation in
nonpolar media. The values of k1 during the oxidation
of styrene were reported to be of 5.5 × 105 and
1.5 × 106 M−1 s−1 for catechols 2 and 4, respectively
[22], which are almost equal to k1 determined in this
work (if these are corrected for n = 2). The values of
k1 reported in [23] for 6 (6.5 × 104 M−1s−1) and for
9 (2.9 × 105 M−1s−1) are several fold lower than k1
determined in this work (Table I). Most likely, the rea-
son is that in work [23] the reaction mixture contained
0.1 M methanol that forms H-bonds with phenolics.
Several k1 values were determined during the oxidation
of methyl linoleate (ML) [16,19]. All of them were also
distinctly lower than these determined in the current
work. Most likely, the effect is due to the occurrence
of H-bond between QH2 and the carboxy group of
ML. If this is the case, the effect should become more
pronounced when the ML concentration increases.
The latter is in line with the experimental data. When
the concentration of ML in the reaction mixture was
only 0.24 M, the effect was rather moderate: with PP
12 and PP 15 k1 was ca. 8 × 105 M−1s−1 (corrected for
n = 2) [16] that is only twice as less as compared to
k1 measured in our work. When the reaction mixture
consisted of nearly 100% ML [19], the effective values
of k1 were much lower: ca. 3 × 104 M−1s −1 for 9 and
12 (less than k1 in Table I by nearly two orders of mag-
nitude). A similar effect was also reported with several
synthetic monophenols [24]. In a more general form,
the reducing influence of H-bonds on reactivity of phe-
nol antioxidants has been considered in other works
[16,25].
With some PP studied in this work, very high k1
values were reported during the oxidation of−d1iphenyl-
for 9,
methane (for instance, 1.5 × 107 M−1
s
2.1 × 107 M−1s−1 for 11, 1.0 × 106 M−1s−1 for 10)
[26]. These values are higher than k1 presented in
Table I by nearly one order of magnitude. The differ-
ence in k1 values between those reported by [26] and
those determined in our work may only partly explain
the elevated reactivity of LO•2 from diphenylmethane
[26]. One more reason for this difference could be the
fact that in [26] determinations were conducted at the
rate of initiation of 1.0 × 10−10 Ms−1 that is lower than
typical values of RIN in our work by one–two orders
of magnitude. To rule out this reasoning, the deter-
mination of k1 for catechol 2 was also performed at
RIN = 4.0 × 10−10 Ms−1 instead of 4.0 × 10−9 Ms−1
.
,
In this case, k1 was found to be 0.9 × 106 M−1s−1
which actually does not differ from that in Table I. In
conclusion, the reason for so high values of k1 in the
work of [26] remains unclear.
International Journal of Chemical Kinetics DOI 10.1002/kin

98
TIKHONOV, ROGINSKY, AND PLISS
4. As may be expected, the electron-withdrawing
substituent COOH causes the significant re-
duction of k1 (compare 2 and 6, 3, and 7 in
Table I). The occurrence of two bulky tert-butyl
substituents in 4 results in the significant in-
crease in k1 as compared to the nonsubstituted
catechol 2.
Relationship of k1 with the PP Structure
Let us consider how k1 changes with the number and
the position of OH groups. The following regularities
may be noted:
1. Two meta-OH groups display a very low if any
reactivity to LO•2. For instance, chrysin that has
no OH groups in ring B shows a very low reac-
tivity, undetectable under our conditions. This is
also evident from the fact that antioxidants con-
taining two meta-OH groups along with other
OH groups (flavonoids 11–13, 15–17, as well
as resveratrol 18) display f close to 2 (Table I).
The latter means that meta-OH groups in the
A-ring do not actually participate in the an-
tioxidative action of the mentioned antioxidants.
The low reactivity of meta-OH groups is also
in line with the very moderate reduction in
BDE as compared with the nonsubstituted phenol
[27].
Competition between Reactions (4) and (5).
Coefficient n in Eqs. (2) and (3)
As mentioned above, the coefficient n depends on the
predominant way of QH• termination. The competition
between reactions (4a) and (5a) has been considered
for the case, when the antioxidant is a monophenol
PhOH [12,29]. The contribution of reaction (4a), α, is
given by the following relation:
k42RIN
k12k5[PhOH]2
α2(1 + α) −
(1 − α) = 0
(6)
2. Antioxidants containing only one active OH
group (1, 5, 8, 10) display a rather moderate reac-
tivity to LO•2, varying from 2 × 103 to 1.3 × 105
Equation (6) still works with the change of [PhOH]
for [QH2]. The competition depends on five inde-
pendent values, RIN, [QH2] ([PhOH]), k1, k4, and k5.
Whereas two first values are always known and k1 is
determined in the course of every run, the informa-
tion on k4 and k5 for PP is highly limited (for PhOH
k4 ≈ 3 × 108 M−1 s−1 and actually independent of the
PhOH structure [12]). As for k4 for QH2, this has never
been reported. The values of k5 reported in the litera-
ture for several PP fall within the range between 106
and 109 M−1 s−1 [12,30,31]. The estimations of the α
value by using Eq. (6) show that the situation when
M
−1 s−1. Resveratrol 18 (k1 = 3.3 × 105 M−1 s−1
(Table I)) is a remarkable exception. There are
numerous publications suggesting that resvera-
trol should show an outstanding chain-breaking
antioxidant activity (see [28,29] and references
therein). As a rule, the above speculations are
based on indirect data. The k1 value reported for
18 in this work has been determined directly for
the first time. Despite a rather high value of k1 for
18, it is lower by nearly one order of magnitude
than k1 for many other natural PP, for instance 9,
12, 13, 15, and 16 (Table I).
α
1 (reaction (5) prevails over reaction (4)) is real in
many cases, especially with the most active PP. For ex-
ample, with the following reasonable combination of
3. k1 increases dramatically up to (1–2) × 106 M−1
parameters: k1 = 1 × 106 M−1
s
−1; k4 = k5 = 1 × 108
−1, when going to the antioxidants containing
M
−1 s−1; RIN = 3 × 10−9 Ms−1; [QH2] = 1 × 10−5 M,
s
two o-OH groups (catechol derivatives). This
is evident when the following couples are con-
fronted: 1 and 2, 5 and 6, 8 and 9, and 11 and 12
(Table I). Meanwhile, the increase in the num-
ber of adjacent OH groups from two to three
results in the subsequent increase in k1 by a
factor of 2–3 (compare 2 and 3, 6 and 7, 12
and 13, and 15 and 16 (Table I)). Interestingly,
baicalein 14, a rather exotic flavonoid with three
adjacent OH groups in the A-ring rather than in
the B-ring (see Scheme 1) shows slightly lower
reactivity as compared to myricetin 13, pyro-
gallol 3 (also with three adjacent OH groups),
and even in comparison with antioxidants with
only two adjacent OH groups, 2, 9, 12, and 15
(Table I).
α was found to be 0.032. We may speculate that the
situation, when reaction (5) predominates, is more typ-
ical. This is the reason why the coefficient n in Eqs.
(2) and (3) applied to the calculation of k1 listed in
Table I is equal to 2. However, in the general case, the
coefficient n can vary within the range from 1 to 2. It
means that the values of k1 in Table I are given with
the uncertainty within the range (1–0.5)k1.
Correlation of k1 with the O H Bond
Dissociation Enthalpy
Bond dissociation enthalpy (BDE) for O H bonds is
one of the most significant factor in determining the re-
activity of phenolics to LO•2. So it is very promising to
correlate k1 with BDE. In contrast with monophenols
International Journal of Chemical Kinetics DOI 10.1002/kin

ANTIOXIDANT ACTIVITY OF POLYPHENOLS
99
[12,32,33], the experimental information on BDE for
O H bonds in PP is highly limited [27,35]. It could be
possible to make an attempt to correlate k1 for PP with
BDE calculated by using quantum-chemical methods,
since extensive studies of such a kind have been pub-
lished recently [27,36–41]. Unfortunately, the BDE
values calculated for PP vary over rather a wide range
when going from one work to another. For instance,
the reported calculated values of BDE for catechol
were (in kcal/mol): 72.8 [38], 74.7 [27], and 77.9 [41],
which are visibly less than the experimental value of
80.5 [35]. Under these circumstances, we are forced
to restrict our consideration to a general tendency. As
calculations suggest, BDE decreases when the number
of adjusted O H groups increases from one to three.
For instance, the work [27] reported the following val-
ues of BDE (in kcal/mol): 88 for phenol, ∼77 for PP
with catechol moiety, and 72 for pyrogallol derivatives
(the latter value belongs to the middle OH group [27]).
At the same time, phenolics with two O H groups in
the metaposition (resorcinol derivatives) show a very
small if any decrease in BDE as compared to phenol
[42]. As for derivatives of gallic acid, the calculated
value of BDE is significantly higher than in pyrogal-
lols derivatives and is close to that in catechols [27].
The mentioned values of BDE are generally in line
with the reactivity of PP as reported in Table I.
12. Roginsky, V. A. Phenolic Antioxidants. Efficiency and
Reactivity; Nauka: Moscow, 1988 (in Russian).
13. Barclay, L. R. C.; Vinqvist, M. R. In The Chemistry of
Phenols; Rappoprt, Z. (Ed.); Wiley: New York, 2003;
Part 2, Ch. 12.
14. Loshadkin, D.; Roginsky, V.; Pliss, E. 2002, Int J Chem
Kinet 2002, 34, 162–171.
15. Roginsky, V. Arch Biochem Biophys 2003, 414, 261–
270.
16. Pedrielli, P.; Pedulli, G. F.; Skibsted, L. H. J Agric Food
Chem 2001, 49, 3034–3040.
17. Roginsky, V. A. J Org Chem 1989, 25, 403–413 (trans-
lated from Russian).
18. Roginsky, V. A.; Dubinsky, V. Z.; Shlyapnikova, I. A.;
Miller, V. B. Eur Polym J 1977, 13, 1043–1051.
19. Roginsky, V. A.; Remorova, A. A.; Barsukova, T. K.;
Bors, W. J Am Oil Chem Soc 1996, 73, 777–
786.
20. Landolt-Bo¨rnstein Numerical Data and Relationships in
Science and Technology New Series, Group II; Springer:
Berlin, 1984; Vol. 13d.
21. Ross, A. B.; Mallard, W. G.; Helman, W. P.; Buxton,
G. F.; Huie, R. E.; Nets, P. NDRL-NIST Solution Kinet-
ics Datanase, Version 2; NST Standard Reference Data,
Gaithersburg, MD, 1994.
22. Xi, P.; Barclay, R. C. Can J Chem 1998, 76,
171–182.
23. Amorati, R.; Pedulli, G. F.; Cabriani, L.; Zambonin,
L.; Landi, L. J Agric Food Chem 2006, 54, 2932–
2937.
24. Roginsky, V. A. Kinet Catal 1990, 31, 475–481 (trans-
lated from Russian).
25. Litwinenko, G.; Ingold, K. U. Acc Chem Res 2007, 40,
222–230.
BIBLIOGRAPHY
1. Scalbert, A.; Manach, G.; Morand, C.; Remesy, C.;
Jimenez, L. Crit Rev Food Sci Nutr 2006, 45, 287–
306.
2. Manachi, C.; Donovan, J. L. Free Radical Res 2004, 38,
771–785.
3. Sang, S.; Hou, Z.; Lambert, J. D.; Yang, C. S. Antioxid
Redox Signal 2005, 7, 1704–1714.
4. Lotito, S.; B.; Frei, B. Free Radical Biol Med 2006, 41,
1727–1746.
26. Beljakov, V. A.; Roginsky, V. A.; Bors, W. J Chem Soc,
Perkin Trans 1995, 2, 2319–2326.
27. Wright, J. S.; Johnson, E. R.; DiLabio, G. A. J Am Chem
Soc 2001, 123, 1173–1183.
28. Hao, Y. D.; He, L. R. J Med Food 2004, 7, 290–
298.
29. Fremont, L. Life Sci 2000, 66, 663–673.
30. Land, E. J Chem Soc, Faraday Trans 1993, 89, 803–810.
31. Deeble, D. J.; Parsons, B. J.; Philips, G. O.;
Schuchmann, H.-P.; von Sonntag, C. Int J Radiat Chem
1988, 54, 179–193.
5. Rice-Evans, C.; Miller, N. J.; Paganga, G. Free Radical
Biol Med 1996, 20, 933–956.
6. Roginsky, V.; Lissi, E. A. Food Chem 2005, 92, 235–
254.
32. Mulder, P.; Saastad, O. W.; Griller, D. J Am Chem Soc
1988, 110, 4090–4092.
7. Heim, K. E.; Tagliaferro, A. K.; Bobilya, D. J. J Nutr
Biochem 2002, 13, 572–584.
33. Jonsson, M.; Lind, J.; Eriksen, T. E.; Merenyi, G. J Chem
Soc, Perkin Trans 1993, 2, 1567–1568.
8. Silva, M. M.; Santos, M. R.; Caraco, G.; Rocha, R.;
Justino, G.; Mira, L. Free Radical Res 2002, 36, 1219–
1227.
9. Erlejman, A. G.; Verstraeten, S. V.; Fraga, C. G.; Oteiza,
P. I. Free Radical Res 2004, 38, 1311–1320.
10. Burda, S.; Oleszek, W. J Agric Food Chem 2001, 49,
2774–2779.
11. Burton, G. W.; Doba, T.; Hughes, L.; Lee, E. L.;
Ingold, K. U. J Am Chem Soc 1985, 107, 7053–
7065.
34. Lucarini, M.; Mugnaini, V.; Pedulli, G. F. J Org Chem
2002, 67, 928–931.
35. Suryan, M. M.; Kafafi, S. A.; Stein, S. E. J Am Chem
Soc 1989, 111, 4594–4600.
36. Giacomelli, C.; da Silva Miranda, F.; Gonc¸alves, N. S.;
Spinelli, A. Redox Rep 2004, 9, 263–269.
37. Leopoldini, M.; Marino, T.; Russo, N.; Toscano, M. J
Phys Chem A 2004, 108, 4916–4922.
38. Zhang, H.-Y.; Sun, Y.-M.; Wang, X.-L. Chem Eur J
2003, 9, 502–508.
International Journal of Chemical Kinetics DOI 10.1002/kin

100
TIKHONOV, ROGINSKY, AND PLISS
39. Hussain, H. H.; Babic, G.; Durst, T.; Wright, J. S.;
Flueraru, M.; Chichirau, A.; Chepelev, L. L. J Org Chem
2003, 68, 7023–7032.
41. Nenadis, N.; Wang, L.-F.; Tsimidou, M. Z.; Zhang,
H.-Y.
299.
J
Agric Food Chem 2005, 53, 295–
40. Cren-Olive, C.; Hapiot, P.; Pinson, J.; Rolando, C. J Am
Chem Soc 2002, 124, 14027–14038.
42. Thavasi, V.; Leong, L. P.; Bettens, R. P. A. J Phys Chem
A 2006, 110, 4918–4923.
International Journal of Chemical Kinetics DOI 10.1002/kin