Talanta 72 (2007) 1157–1165

Comparative evaluation of Fe(III) reducing power-based antioxidant

capacity assays in the presence of phenanthroline, batho-phenanthroline,
tripyridyltriazine (FRAP), and ferricyanide reagents
Kadriye Işıl Berker, Kubilay Güçlü, İzzet Tor, Reşat Apak ∗
Istanbul University, Faculty of Engineering, Department of Chemistry, Avcilar, 34320 Istanbul, Turkey
Received 11 September 2006; received in revised form 28 December 2006; accepted 9 January 2007
Available online 16 January 2007

Abstract
The chemical diversity of antioxidants in complex matrices such as plant extracts makes it difficult to separate and quantify antioxidants
from these solutions. Therefore it is desirable to establish methods that can measure the total antioxidant capacity (TAC) levels directly from
plant extracts. Iron(III)-based TAC assays, especially the most widely used FRAP (ferric-reducing antioxidant power), play an important role
in this regard. However, many problems have been reported in the application of the FRAP assay, the most serious one being the incomplete
oxidation of a number of antioxidants during the time protocol of the assay. Thus, six different ferric ion-based total antioxidant capacity (TAC)
assays have been comparatively tested, modified, and improved so as to obtain more sensitive and precise results for complex mixtures, namely:
1,10-phenanthroline (o-phen) method (with incubation), batho-phenanthroline method (with incubation), original FRAP method, modified FRAP
method (with incubation), original ferricyanide method, and modified ferricyanide method (with incubation). Two new assays in this regard (i.e.,
o-phen and batho-phen) have been established, and the existing assays (FRAP and ferricyanide) have been modified so as to let the oxidation
reactions of antioxidants reach completion. The molar absorptivity for a variety of antioxidants was highest for modified FRAP, batho-phen, and
original FRAP methods. The absorption maximum wavelength shifted batochromically to a higher extent for modified ferricyanide, FRAP, and
batho-phen procedures, decreasing the possibility of interferences due to organics absorbing in the near-UV range of the visible spectrum where
most antioxidant assays are performed. The linear concentration ranges were shown to be further extended and linear correlation coefficients
improved with respect to the most widely used ferric-based assay, FRAP. Of the six assays tested and developed, only the modified ferricyanide
procedure gave high intercept values and low addivitity of TAC values of constituents in complex mixtures, requiring further attention of method
optimization. Thus, it was shown that the most widely used FRAP could be effectively modified, and o-phen, batho-phen, and ferricyanide methods
constitute cheaper alternatives to FRAP under certain conditions, with partly improved molar absorptivity (and thus sensitivity) for antioxidants,
lower intercept values (and higher precision), broader linear range (and higher flexibility), and better additivity of TAC values of antioxidant
constituents in mixtures.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Total antioxidant capacity (TAC) assay; Fe(III) reducing power; Phenanthroline; Batho-phenanthroline; Tripyridyltriazine; FRAP; Ferricyanide

1. Introduction The existing antioxidant activity/capacity assay methods

Antioxidants, being non-enzymatic defenses of the organism
against reactive O,N-species, are essential for human health.
Most antioxidant compounds are introduced to the organism
through diet. Therefore it is desirable to establish methods that
can directly measure the total antioxidant capacity of food plant
extracts [1].
∗ Corresponding author. Tel.: +90 212 4737028; fax: +90 212 4737180.
E-mail address: rapak@istanbul.edu.tr (R. Apak).
in literature depending on the consumption of chromogenic
radicals, i.e., ABTS [2] and DPPH [3], oxygen radical absorp-
tion capacity: ORAC [4], or ferric-reducing/antioxidant power:
FRAP [5] have been extensively criticized for their inadequa-
cies in our recent publications introducing the novel CUPRAC
(cupric ion reducing antioxidant capacity) method of total
antioxidant assay [6,7]. In general, antioxidant measurements
performed with the above assays correlate poorly among each
other.
From a mechanistic standpoint, FRAP is an electron trans-
fer (ET)-based assay like Folin, ABTS/TEAC, and CUPRAC

0039-9140/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.talanta.2007.01.019
1158 K.I. Berker et al. / Talanta 72 (2007) 1157–1165
in the sense that the oxidant probe accepts an electron from
the antioxidant analyte to be converted into the reduced probe
which is colored [8]. In the presence of the chromogenic ligand
like tripyridyltriazine (TPTZ) that is rather selective for Fe(II),
Fe(III) acts as an oxidant toward the antioxidants in the sample,
and is itself reduced to Fe(II) which readily chelates with the
chromogenic ligand to form a colored species. The increase in
absorbance at 593 nm (A) – due to Fe(II)–TPTZ complex for-
mation – is proportional to the combined (total) ferric-reducing
antioxidant power (FRAP) of the antioxidants in the sample [9].
The FRAP assay has been claimed to be a robust and poten-
tially useful test using inexpensive reagents and equipment and
a speedy reaction applicable over a wide concentration range
[5].
A number of problems arise in decision making for appli-
cation of Fe(III) reduction-based antioxidant assays to plant
extracts and biological fluids. In spite of the fact that ferrozine
yielded the higher molar absorptivity [10] and therefore the
higher sensitivity for ferrous ions, TPTZ was chosen as the suit-
able ligand in FRAP. Pulido et al. [11] examined the FRAP assay
of dietary polyphenols in water and in MeOH, and found out that
the absorbance of Fe(II)–TPTZ formed from certain polyphe-
nols such as caffeic acid, ferulic acid, tannic acid, and quercetin
was slowly increasing after several hours of reaction time. Both
carotenoids (e.g., ␤-carotene, zeaxanthin) [11] and thiol-type
antioxidants (e.g, glutathione, lipoic acid) [12,13] did not con-
tribute to the ferric-reducing ability of plasma, and therefore, the
protective action of carotenoids and thiols to copper-mediated
LDL (low-density lipoprotein) oxidation was concluded not to
result from their ferric-reducing ability in the FRAP assay [11].
Another slowly reacting physiological antioxidant in the FRAP
assay is bilirubin, and also the optimal wavelength of 593 nm
may not be suitable for this antioxidant as its oxidation prod-
uct, beliverdin, has considerable absorbance at this wavelength
[1]. The total antioxidant capacity of non-enzymatic and non-
chelating antioxidants can be measured via their Fe(III) reducing
capability in the presence of substoichiometric amounts of chro-
mogenic ligands complexing with Fe(II) [5], but in this case, the
Fe(II) reaction product can react with H2 O2 to produce • OH rad-
icals [13], which may yield erroneous results due to the redox
cycling of this unbound iron. Although Fe(III)–Fe(II) standard
reduction potential is 0.77 V causing non-specific oxidation of
any species having a redox potential smaller than this value [1],
suitable selection of Fe(II)-stabilizing chromogenic ligands may
bring this value close to the range of physiologically important
antioxidants and render the iron-based assay selective for a group
of analytes, e.g., the Fe(III)–phen/Fe(II)–phen redox couple has
a much greater standard potential than ferri-ferrocyanide. The
Fe(III) reducing capability of a sample is also related to the pH of
the redox reaction involved, and thus selective oxidation reac-
tions for certain antioxidants can be designed by carrying out
the redox reaction at the acidic pH of FRAP or at the neutral
pH of [Fe(CN)6 ]3− . In order to overcome all these analytical
problems and render iron-based assays selective and sensitive
for a group of important antioxidants, the existing assays were
modified in this work (so as to force the oxidation equilib-
ria to completion with elevated temperature incubation) and
improved by changing the reaction conditions. Specifically, o-
phenanthroline, batho-phenanthroline, tripyridyltriazine (FRAP
reagent), and ferricyanide reagents, each in the presence of
Fe(III), were tested at standard protocol conditions (room tem-
perature and short reaction time) versus 30 min incubation
at 50 ◦ C for the antioxidant compounds of divergent classes,
namely trolox, ascorbic acid, caffeic acid, ferulic acid, gallic
acid, quercetin, and rutin. The molar absorptivity, linear range,
and TEAC coefficients of these antioxidants were found using
each method, and were compared among themselves for use-
ful method selection. The linearity of the calibration curves of
antioxidants was also tested in green tea infusions as the real
matrix medium.
2. Materials and methods
2.1. Chemicals
TPTZ (2,4,6-tri(2-pyridyl)-s-triazine), quercetin dihydrate,
gallic acid, batho-phenanthroline (4,7-diphenyl-1,10-
phenanthroline) were purchased from Fluka; caffeic acid, ferulic
acid, trolox, and ascorbic acid from Aldrich; rutin, and sodium
dodecyl sulfate from Sigma Chemical Co.; 1,10-phenanthroline
monohydrate, trichloroacetic acid (TCA), and HCl from Riedel-
de Haën; acetic acid, Na2 HPO4 ·2H2 O, NaH2 PO4 ·2H2 O from
J.T. Baker; FeCl3 ·6H2 O, K3 [Fe(CN)6 ], NH4 Fe(SO4 )2 ·12H2 O,
(NH4 )2 Fe(SO4 )2 ·6H2 O, CH3 COONa·3H2 O, ethanol (96%, by
vol.) from E. Merck; and finally green tea bags (as real matrix
for standard additions) from Doğadan Bitkisel Urunler Co.
2.2. Instruments
A Varian CARY 1E UV–vis spectrophotometer was used
for absorbance readings and spectra recording, using a pair of
matched quartz cuvettes (Hellma). The pH measurements were
made with an E512 Metrohm Herisau pH-meter equipped with
a combined glass electrode. The samples in the original ferri-
cyanide method were centrifuged – when necessary – through
a MSE Mistral 2000 centrifuge apparatus, and sample equili-
bration and incubation operations were made with the aid of an
Elektromag vortex stirrer and a Clifton water bath.
2.3. Procedures
The following standard and modified procedures were used
in iron(III)-based antioxidant assays:
2.3.1. 1,10-Phenanthroline (o-phen) method (with incubation).
2.3.2. Batho-phenanthroline (batho-phen) method (with incu-
bation).
2.3.3. Original FRAP method1 (without incubation).
1 Only the original FRAP method was carried out at room temperature. All
other procedures with incubation were conducted at 50 ◦ C on a water bath for
20 or 30 min, except for the modified FRAP method which was performed at
37 ◦ C for 20 min.
 K.I. Berker et al. / Talanta 72 (2007) 1157–1165
2.3.4. Modified FRAP method (with incubation).
2.3.5. Original ferricyanide method (with incubation).
2.3.6. Modified ferricyanide method (with incubation).
The hypothesis of additivity of absorbances due to the tested
antioxidants was tested for each method using synthetic binary,
ternary, and quaternary mixtures of antioxidants. Also any pos-
sible deviation from Beer’s law due to interaction of the tested
antioxidants with the natural antioxidants of a complex real
matrix was tested by standard addition of three selected antiox-
idants (trolox, gallic acid, and caffeic acid) to properly diluted
green tea infusions using each method. The procedures are
briefly described below.
2.3.1. 1,10-Phenanthroline (o-phen) method (with
incubation)
2.3.1.1. Preparation of solutions. To 0.160 g of NH4 Fe
(SO4 )2 ·12H2 O was added 2 mL of 1 M HCl; a suitable mass of
o-phenanthroline was dissolved in water so as to make its final
concentration 1.0 × 10−2 M. These two solutions were mixed
and diluted to 100 mL with distilled water. This reagent, when
kept in a dark, stoppered bottle and protected from sunlight, was
shown to be stable for several weeks [14]. The stock solutions
of the tested antioxidants were prepared in 96% EtOH in the
concentration range of 5.0 × 10−5 to 1.0 × 10−4 M.
2.3.1.2. Procedure. Antioxidant testing. To (x) mL antioxidant
solution in a 25 mL-flask, add 1 mL of 1,10-phen reagent solu-
tion, (5 − x) mL EtOH (96%), and dilute to the mark (25 mL)
with H2 O. Incubate in a water bath at 50 ◦ C for 30 min, let cool
to room temperature, and measure the absorbance at 510 nm
(A510 ) against a reagent blank (the color remains stable for at
least 1 h). The procedure was originally defined in literature for
ascorbic acid only (without incubation) [14], and not for other
antioxidants. This procedure was modified with incubation so
as to cover the slower-reacting antioxidants.
Antioxidant standard addition to green tea infusion. A green
tea bag was immersed in 250 mL of freshly boiled hot water and
stirred for 2 min; the tea bag was let to remain in the same solu-
tion for three more min without heating (total time of preparation
of tea infusion being 5 min). The infusion was filtered through
a black band (Whatman) filter paper after cooling, and diluted
with H2 O at a ratio of 1:50. To 2.5 mL of the diluted infusion
(i.e., x = 2.5) were added increasing amounts of a given antioxi-
dant solution such that the initial absorbance of the infusion was
about 0.2 with respect to the phen method. The calibration line
(abs. versus concn.) of a given antioxidant was redrawn in the
presence of 2.5 mL diluted green tea infusion to observe if there
was any deviation from Beer’s law.
2.3.2. Batho-phenanthroline (batho-phen) method (with
incubation)
2.3.2.1. Preparation of solutions. Section 2.3.1.1 was repeated
with the exception that batho-phenanthroline alcoholic solution
(in 96% EtOH) was substituted for o-phenanthroline aqueous
solution, and dilution to a final volume of 100 mL was made
with 96% EtOH. The Fe(III)–batho-phen reagent thus prepared
1159
could be used 1 h after its preparation, and was stable for at least
a week.
2.3.2.2. Procedure. To (x) mL antioxidant solution in a 25 mL-
flask, add 1 mL of batho-phen reagent solution, (5 − x) mL EtOH
(96%), and dilute to the mark (25 mL) with H2 O. Incubate in a
water bath at 50 ◦ C for 30 min, let cool to room temperature, and
measure the absorbance at 533 nm (A533 ) against a reagent blank
(the color remains stable for at least 1 h). This procedure was
adopted from the one concerning o-phen (Section 2.3.1.2) with
incubation so as to enable full oxidation of the slower-reacting
antioxidants.
To 1.5 mL of the diluted infusion (i.e., x = 1.5) were added
increasing amounts of a given antioxidant solution such that the
initial absorbance of the infusion was about 0.2 with respect to
the batho-phen method. The calibration line (abs. versus concn.)
of a given antioxidant was redrawn in the presence of 1.5 mL
diluted green tea infusion to observe if there was any deviation
from Beer’s law.
2.3.3. Original FRAP method (without incubation)
2.3.3.1. Preparation of solutions. A suitable mass of
FeCl3 ·6H2 O was weighed so that the final concn. of Fe(III) in
solution would be 2.0 × 10−2 M; 1 mL of 1 M HCl solution was
added, dissolved in some water and diluted to 50 mL with H2 O.
A suitable mass of TPTZ was weighed such that its final concn.
would be 1.0 × 10−2 M, dissolved in 96% EtOH, and diluted
to 50 mL. In order to prepare 0.3 M CH3 COOH/CH3 COONa
buffer solution at pH 3.6, 3.1 g of CH3 COONa·3H2 O was
weighed and 16 mL glacial acetic acid was added, diluted with
water to 1 L [5,9,11,15].
The FRAP reagent was prepared as follows: the pH 3.6 acetic
acid buffer, 1.0 × 10−2 M TPTZ solution, and 2.0 × 10−2 M
FeCl3 ·6H2 O solution were mixed in this order at a volume ratio
of 10:1:1. The FRAP reagent was prepared and used freshly. The
tested antioxidant solutions were prepared in 96% EtOH in the
concn. range of 3.0 × 10−4 to 1.0 × 10−3 M.
2.3.3.2. Procedure. To 3 mL of the FRAP reagent, add
x mL antioxidant solution, (0.1 − x) mL of 96% EtOH, and
0.3 mL H2 O such that the final volume is 3.4 mL. Read the
absorbance at 595 nm (A595 ) against a reagent blank at the end
of 6 min.
The original green tea infusion after filtration was diluted at
a ratio of 1:10 with water, an aliquot of 100 ␮L was taken with
a micropipette such that its final absorbance (without addition
of the tested antioxidants) would be around 0.2 with respect to
the FRAP method, and the technique of standard antioxidant
addition was applied as described earlier.
2.3.4. Modified FRAP method (with incubation)
2.3.4.1. Preparation of solutions. Section 2.3.3.1 as described
for the original FRAP method was repeated with the exception
that the tested antioxidant solutions were prepared in 96% EtOH
in the concn. range of 2.0 × 10−4 to 1.0 × 10−3 M.
1160 K.I. Berker et al. / Talanta 72 (2007) 1157–1165
2.3.4.2. Procedure. To 3 mL of the FRAP reagent, add x mL
antioxidant solution, (0.1 − x) mL of 96% EtOH, and 0.3 mL
H2 O such that the final volume is 3.4 mL. Incubate the mix-
ture at 37 ◦ C for 20 min to enable more complete oxidation
of the tested antioxidants. Read the absorbance of the room
temperature-cooled analyte solutions at 595 nm (A595 ) against a
reagent blank. The color of the final mixture solutions was stable
for at least 30 min.
The original green tea infusion after filtration was diluted at
a ratio of 1:20 with water, an aliquot of 100 ␮L was taken with
a micropipette such that its final absorbance (without addition
of the tested antioxidants) would be around 0.25 with respect
to the modified FRAP method, and the technique of standard
antioxidant addition was applied as described earlier.
2.3.5. Original ferricyanide method (with incubation)
2.3.5.1. Preparation of solutions. To prepare 0.2 M phosphate
buffer at pH 6.6, 7.80 g of NaH2 PO4 ·2H2 O was dissolved in
water and diluted to 250 mL with H2 O such that its final concn.
would be 0.2 M; 8.89 g of Na2 HPO4 ·2H2 O was dissolved in
water and diluted to 250 mL such that its final concn. would be
0.2 M. A volume of 68.5 mL of the NaH2 PO4 ·2H2 O solution
was mixed with 31.5 mL of the Na2 HPO4 ·2H2 O to prepare the
0.2 M phosphate pH 6.6 buffer. Potassium ferricyanide solution
(1%, w/v) was prepared daily by dissolving 1 g K3 Fe(CN)6 in
1 mL of 1 M HCl and some water, and diluting to 100 mL with
water. Ferric chloride solution (0.1%, w/v) was prepared daily
by dissolving 0.1 g of FeCl3 ·6H2 O in 1 mL of 1 M HCl and
some water, and diluting to 100 mL with water. Trichloroacetic
acid (TCA) solution (10%, w/v) was prepared by dissolv-
ing 10 g of TCA in water and diluting to 100 mL with H2 O
[16].
2.3.5.2. Procedure. To x mL of antioxidant solution were added
(1 − x) mL of EtOH (96%), 2.5 mL of 0.2 M phosphate buffer
(pH 6.6) and 2.5 mL of K3 Fe(CN)6 solution (1%); the mixture
was incubated at 50 ◦ C on a water bath for 20 min. The incu-
bated mixture was let to cool to room temperature, and 2.5 mL
of TCA (10%) was added. The solution was thoroughly mixed,
an aliquot of 2.5 mL was withdrawn, and 2.5 mL water followed
by 0.5 mL of FeCl3 ·6H2 O solution (0.1%) was added so that the
final volume was 5.5 mL. The absorbance of the resulting Prus-
sian blue solution at 700 nm (A700 ) was measured after 2 min
against a reagent blank. It should be noted here that although
the exact time period of the original ferricyanide method was
not clearly stated in literature, certain fast-reacting antioxidants
such as trolox caused precipitation of the Prussian blue com-
plex salt if a measurement time longer than 2 min was selected.
Ascorbic acid, being one of the antioxidants routinely tested in
other procedures, was not examined in this procedure due to its
possible thermal degradation during the preliminary incubation
at 50 ◦ C. The tested antioxidants were taken from their stock
solutions in the concn. range of 3.0 × 10−4 to 1.0 × 10−3 M.
For antioxidant standard addition to green tea infusion, the
original infusion was diluted at a ratio of 1:5 after filtration, and
0.2 mL was taken for standard additions such that A700 of the
infusion – without added antioxidant – would be 0.18.
2.3.6. Modified ferricyanide method (with incubation)
2.3.6.1. Preparation of solutions. K3 Fe(CN)6 solution (1%,
w/v) was prepared as described in Section 2.3.5.1. Ferric chlo-
ride solution (0.2%, w/v) was prepared daily by dissolving 0.2 g
of FeCl3 ·6H2 O in 1 mL of 1 M HCl and some water, and dilut-
ing to 100 mL with water. Sodium dodecyl sulfate solution (1%,
w/v) was prepared by dissolving 1 g of SDS in water and dilut-
ing to 100 mL with water. There is no pH 6.6 phosphate buffer
used in the modified ferricyanide method.
2.3.6.2. Procedure. To x mL antioxidant solution were added
(1 − x) mL EtOH (96%), 5 mL H2 O, 1.5 mL of 1 M HCl, 1.5 mL
of ferricyanide solution (1%), 0.5 mL of SDS (1%), and finally
0.5 mL of FeCl3 ·6H2 O (0.2%) so that the final volume would be
10 mL. The mixture was incubated at 50 ◦ C on a water bath
for 20 min, let cool to room temperature, and the resulting
absorbance at 750 nm (A750 ) was measured against a reagent
blank. The color of the final solution was stable for at least
30 min. The Prussian blue formed in this method did not precip-
itate (as opposed to that in the original ferricyanide method) due
to the stabilizing effect of SDS. Another difference worthy of
mention is the addition of ferric ions from the start along with
ferricyanide so that both oxidants (i.e., Fe(III) and Fe(CN)3 3− )
could participate in antioxidant oxidation carried out in acidic
solution. Also there was a batochromic shift of the maximum
absorption wavelength to 750 nm in the modified ferricyanide
method. The tested antioxidants were taken from their stock
solutions in the concn. range 5.0 × 10−4 to 1.0 × 10−4 M.
For antioxidant standard addition to green tea infusion, the
original infusion was diluted with water at a ratio of 1:25 after
filtration, and 0.6 mL was taken for standard additions such that
A750 of the infusion – without added antioxidant – would be
around 0.20.
3. Results and discussion
The chemistry of iron-based assays may be summarized with
the following reaction equation:
Fe(III)–L + antioxidant  Fe(II)–L + oxidizedantioxidant
where L is the ferrous-selective chromogenic ligand produc-
ing the colored species Fe(II)–L as a result of the concerned
redox reaction. Since each mole of n electron-reductants would
produce n moles of Fe(II)–L, the molar absorptivity of such an
antioxidant in the selected iron-based method would be expected
to be n-times that of Fe(II)–L. This reasoning applies for L: o-
phen, batho-phen, and TPTZ ligands. For example, the redox
reaction with tris(phen)Fe(III) of a polyphenolic compound
Ar(OH)n may be represented by the following equation:
nFe(phen)3 3+ + Ar(OH)n  nFe(phen)3 2+ + Ar( O)n + nH+
When the oxidant species is either Fe(III) or Fe(CN)6 3− in
the composite ferricyanide reagent, either Fe(II) or Fe(CN)6 4−
forms, respectively, as the reduction product with the antioxi-
dant, and combines with the other reagent component to produce
Prussian blue, KFe[Fe(CN)6 ], as the colored product. In other
 K.I. Berker et al. / Talanta 72 (2007) 1157–1165
words, when Fe3+ is used along with Fe(CN)6 3− as the oxidiz-
ing agent (in the modified ferricyanide assay), either one of the
two reaction pairs occur, which end up with the same product
(Prussian blue):
Fe3+ + antioxidant  Fe2+ + oxidizedantioxidant,
Fe2+ + Fe(CN)6 3−  Fe[Fe(CN)6 ]−
or
Fe(CN)6 3− + antioxidant  Fe(CN)6 4− +oxidizedantioxidant,
Fe(CN)6 4− + Fe3+  Fe[Fe(CN)6 ]−
Naturally under these conditions, the standard redox potential of
the reagent is not E0 = 0.36 V due to Fe(CN)6 3− alone, because
the reduction product Fe(CN)6 4− is stabilized by the formation
of Prussian blue with Fe(III) existing in the reagent. The redox
potential of the Fe3+ /Fe2+ couple is 0.77 V, but since in either
o-phen or batho-phen procedures, Fe(III) forms a 1:3 chelate
with the phenanthroline ligand (and the mixtures were actually
prepared to contain three phenanthroline molecules per each
Fe(III) ion, see Sections 2.3.1 and 2.3.2), this redox potential is
shifted to higher values due to selective stabilization of ferrous
ion with these ligands (e.g., E0 = 1.06 V for the Fe(III,II)–phen
couple). Thus, since Fe(phenanthroline)3 2+ has a higher condi-
tional stability constant than Fe(phenanthroline)3 3+ for both o-
and batho-phen ligands, these reagents have higher potentials
than 0.77 V. On the other hand, the FRAP reagent was prepared
to contain two molecules of TPTZ ligand per Fe(III) ion, because
TPTZ forms a 2:1 chelate with ferrous ions due to steric reasons
[15]. A stoichiometric excess of Fe(III) in the FRAP reagent, as
formulated by Benzie and Strain [15], may be expected to drive
the major redox reaction (caused by Fe(III)–TPTZ) to the right
and ensure the prevention of any chelation interference when
large amounts of iron chelators are present in complex samples
[5,15]. Thus, the redox potential of the FRAP reagent is about
0.7 V [8], close to the values of ABTS (0.68 V) [8] and CUPRAC
(0.6 V) [6,7], as required by most antioxidant assays for the oxi-
dation of physiologically important non-enzymatic antioxidants
with the help of electron transfer-based oxidizing reagents.
All the iron-based assays (except the original ferricyanide
assay) were performed in distinctly acidic solution due to the
hydrolysis of ferric ion at slightly acidic-neutral pH. Since
Fe(III) would be stabilized with hydroxyl ligands by the for-
mation of hydrolysis complexes, the conditional potential of the
Fe(III)–Fe(II) redox couple would shift to lower values in weakly
acidic-to-neutral solution that would inactivate the reagent for
oxidizing the antioxidants tested. The incubation phase of the
original ferricyanide method was carried out at pH 6.6 using a
(H2 PO4 − + HPO4 2− ) buffer with subsequent color development
using Fe(III) in acidic medium (see Section 2.3.5.), because fer-
ricyanide is a stable ferric-complex without a risk of hydrolysis
of the bound Fe(III). On the other hand, the modified ferricyanide
method was carried out in acidic solution to enable the participa-
tion of Fe(III) along with ferricyanide in antioxidant oxidation
(see Section 2.3.6.). It should be noted here that acidic pH values
may not reflect the true behaviour of physiologically important
1161
Fig. 1. The UV–visible spectra of reduced Fe(III)-ligand reagents of iron-
based methods as a result of oxidation of 1.0 × 10−3 M trolox solution. (a)
1,10-Phenanthroline (o-phen) method; (b) batho-phenanthroline (batho-phen)
method; (c) original FRAP method; (d) modified FRAP method; (e) original
ferricyanide method; (f) modified ferricyanide method.
antioxidants in vivo, because most iron-based assays (though all
are in vitro tests) are performed at a pH far below physiological
pH. This criticism was previously directed to the FRAP method
[6,12].
The spectra of colored products obtained with differ-
ent iron-based methods is shown in Fig. 1. As shown in
this figure for trolox as the standard antioxidant, the max-
imum absorption wavelength (λmax ) shifts to longer values
as the iron-based procedure is selected as o-phen < batho-
phen < FRAP < ferricyanide. The batochromic shift in λmax was
highest for ‘Prussian blue’ in the ferricyanide method, because
the bonding and antibonding energy levels was closest in
Fe[Fe(CN)6 ]− due to the interchanging oxidation states of iron
centers in this complex salt. Longer wavelengths almost always
constitute an important advantage in spectrophotometric method
selection, because most plant pigments as well as some antiox-
idants show significant absorption at shorter wavelengths close
to the UV range of the visible spectrum.
The trolox-equivalent antioxidant capacity (TEAC) is defined
as the millimolar concentration of a trolox solution having the
antioxidant capacity equivalent to a 1.0 mM solution of the sub-
stance under investigation. Since FRAP value was defined as
the Fe(II)-equivalent reducing capacity of the sample [5], and
trolox, a 2e-reductant, has a stoichiometric factor of 2 (meaning
that 1 trolox molecule is equivalent to 2 Fe(II) ions in reducing
power), it can be inferred that all FRAP relative activities can be
converted to TEAC coefficients simply by dividing the relative
FRAP activity by 2, e.g., ascorbic acid having a relative FRAP
activity of 2 [5] should show a TEAC coefficient of 1 in the
FRAP assay. The employed iron-based total antioxidant assay
(TAC) methods correctly report high antioxidant capacities for
3 ,4 -dihydroxy substituted flavonoids with conjugated structure
like quercetin (TEAC coefficient varied between 3 and 5), and
polyhydroxy phenolic acids like gallic acid (TEAC coefficient
varied between 2 and 3) (see Table 1), as the theory predicts
[17]. Since quercetin meets one more criterium for antioxidant
potency than rutin (i.e., the presence of 3- and 5-hydroxyl groups
in the A ring together with a 4-oxo function in the A and C rings
of flavonoids) [18], quercetin showed a higher TEAC than rutin
in all the iron-based assays tested (Table 1). The antioxidant
1162 K.I. Berker et al. / Talanta 72 (2007) 1157–1165

potency of flavonoids of similar conjugation level is roughly

TEAC

a The slope (Sl. × 10−4 ) gives the molar absorptivity (ε) of the method for a given antioxidant, e.g., the ε values for trolox of different methods were—o-phen: 2.14 × 104 ; batho-phen: 4.27 × 104 ; original FRAP: 4.62 × 104 ; modified FRAP:
4.68 × 104 ; original ferricyanide: 1.77 × 104 ; modified ferricyanide: 1.88 × 104 L mol−1 cm−1 . The relative standard deviation (R.S.D.) of the slope was less than 2% for all procedures except ferricyanide, original and modified versions (where

b The (mean ± standard deviation) of the linear correlation coefficients of absorbance vs. concentration equations for different methods (r ± σ) were—o-phen: 0.9997 ± 0.00018; batho-phen: 0.9997 ± 0.00028; original FRAP: 0.9996 ± 0.00024;
1.00
0.99
2.16
1.28
2.30
4.78
1.85
proportional to the total number of –OH groups, and is pos-
itively affected by the presence of o-dihydroxy moiety in the

0.5–5.0
0.5–5.0
0.3–3.0
0.4–4.0
0.2–2.0
0.1–1.0
0.3–3.0
Modified ferricyanide

LCRb
B-ring [17]. Since both ascorbic acid [19] and trolox [6] act as 2
e-reductants in these electron transfer-based assays, the TEAC
4.6
0.2
3.0
12.0
0.7
2.5
3.6
coefficient of ascorbic acid was close to 1 in each assay (Table 1).
Int.

Likewise, the molar absorptivity of ascorbic acid in the ferroin

1.88
1.86
4.06
2.42
4.32
8.98
3.48
Sl.a

method [14], when ascorbic acid was oxidized with tris-(1,10-

phenanthroline) iron(III), was shown to be twice that of ferroin
TEAC
1.00

2.48
0.70
2.78
3.27
2.07
(Fe(phen)3 2+ ), indicating 2 e-oxidation of ascorbic acid possibly
–
The slope (Sl. × 10−4 ), intercept (Int. × 102 ), linear concentration range (LCR × 105 ), and TEAC coefficients of the tested antioxidants with respect to a given iron-based assay

to the dehydroascorbic acid product in the ferroin method. On the

1.0–5.3

0.3–2.7
0.8–8.0
0.4–2.1
0.3–1.6
0.5–2.7

other hand, both caffeic and ferulic acids are hydroxycinnamic

LCRb
Original ferricyanide

acids (i.e., having the aromatic substituent CH CH COOH),

but caffeic acid, being an o-dihydroxy cinnamic acid, has one
−2.6

0.4
8.3
2.9
6.6
2.6
Int.

–OH substituent more than ferulic acid which has one OCH3
1.77

4.39
1.24
4.93
5.78
3.67

substituent in o-position to the OH group, and therefore all

Sl.a

iron-based methods yielded higher TEAC coefficients for caf-

TEAC

feic acid (TEAC: 1.1–2.5) than for ferulic acid (TEAC: 0.7–1.3)
1.00
1.14
1.84
1.12
3.03
4.32
2.20

(Table 1).
The TEAC coefficients of various antioxidants found accord-
0.3–2.1
0.3–2.1
0.1–1.5
0.3–1.8
0.2–0.9
0.1–0.5
0.2–1.0
LCRb

ing to the original ABTS method [2,20], FRAP method [5,11],

Modified FRAP

CUPRAC method [6], and calculated with respect to the

−0.8
−0.1
−5.5
−0.4
−1.6
−1.8
−2.0
Int.

employed iron-based methods (Table 1) were generally in accord

with each other as regards the order of antioxidant potency of
4.68
5.33
8.60
5.28
14.20
20.20
10.30
Sl.a

selected compounds. The TEAC coefficients of various antiox-

idants using iron-based methods were simply calculated by
TEAC
1.00
1.01
1.13
0.87
1.85
2.92
1.12

dividing the molar absorptivity (ε) of the species under investi-

gation by that of trolox for a given method (e.g., the ε values of
0.6–2.9
0.6–2.9
0.3–1.8
0.6–2.9
0.3–1.5
0.2–0.9
0.3–2.1

trolox for different methods were—o-phen: 2.14 × 104 , batho-

LCRb

phen: 4.27 × 104 , original FRAP: 4.62 × 104 , modified FRAP:

modified FRAP: 0.9997 ± 0.00021; original ferricyanide: 0.9993 ± 0.00026; modified ferricyanide: 0.9995 ± 0.00025.
Original FRAP

4.68 × 104 , original ferricyanide: 1.77 × 104 , modified ferri-

−5.0
−2.4
−2.8
4.6
−5.4
−1.9
−6.3
Int.

cyanide: 1.88 × 104 L mol−1 cm−1 , respectively). The molar

4.62
4.66
5.24
4.02
8.57
13.50
5.17

absorptivity (ε) for a given antioxidant was highest for modi-

Sl.a

fied FRAP > batho-phen > original FRAP (Table 1). Naturally,
one can detect lower concentrations of antioxidants as the ε
TEAC
1.00
1.06
1.59
1.28
3.23
3.65
2.08

for a given antioxidant assumes higher values, producing an

advantage in sensitivity. It was also shown (see Table 1) that
0.4–2.0
0.4–2.0
0.2–1.2
0.3–1.6
0.1–1.0
0.1–0.6
0.2–1.0
LCRb

the linear concentration ranges (LCR) of iron-based methods

could be further extended and linear correlation coefficients (r)
−0.05
0.1
0.8

3.1
−0.6
4.2
−0.8

improved with respect to FRAP. The methods of o-phen, batho-

Int.
Batho-phen

phen, and modified FRAP gave lower intercept values (of the
4.27
4.54
6.81
5.50

8.92

absorbance versus concentration equations) than those of orig-

13.8
15.6
Sl.a

inal FRAP (Table 1) which should have a crucial effect on the

TEAC

additivity of TAC values of antioxidant components in mixtures

1.00
1.01
1.94
1.18
3.86
4.16
2.35

(see Table 2). On the other hand, the intercept values for the
modified ferricyanide method were generally high (Table 1),
0.8–4.0
0.4–4.0
0.3–3.0
0.4–4.0
0.1–1.2
0.2–1.0
0.2–2.0
LCRb

possibly adversely affecting the additivity of TAC values of syn-

thetic mixtures assayed by the modified ferricyanide method
−0.6
−2.5
1.2
0.8
3.2
0.5
1.5
Int.

(Table 2).
o-Phen

The rate-limiting step of the FRAP assay was reported as

2.14
2.17
4.17
2.54
8.25
8.92
5.03
Sl.a

R.S.D. reached approx. 5%).

antioxidant oxidation with the Fe(III)–TPTZ reagent [5]. In

separate tests using the FRAP method, oxidation of certain
Antioxidant compound

Ascorbic acid (AA)

antioxidants such as tannic acid, rutin, catechin [11] as well

Caffeic acid (CF)
Ferulic acid (FR)
Gallic acid (GA)
Quercetin (QR)

as of caffeic and ferulic acids [13] were shown to be incom-

Trolox (TR)

plete within the time period of the conventional FRAP protocol.

Rutin (RT)
Table 1

In this work, increase in TEAC coefficients of quercetin, gallic

acid, and rutin upon incubation was typical with most methods
Table 2
Expected and found total antioxidant capacities (TAC, as mM trolox-equivalent) of binary, ternary and quaternary mixtures of antioxidants using iron-based assays
Mixture componentsa o-Phen Batho-phen Original FRAP Modified FRAP Original ferricyanide Modified ferricyanide

Expected Found Expected Found Expected Found Expected Found Expectedb Found Expected Found

RT + FR 1.88 × 10−2 2.04 × 10−2 8.26 × 10−3 9.04 × 10−3 1.17 × 10−2 1.14 × 10−2 8.48 × 10−3 8.34 × 10−3 2.22 × 10−2 2.30 × 10−2 2.13 × 10−2 2.94 × 10−2
CF + FR 2.10 × 10−2 2.19 × 10−2 7.92 × 10−3 8.66 × 10−3 9.10 × 10−3 9.34 × 10−3 9.36 × 10−3 8.80 × 10−3 2.44 × 10−2 2.80 × 10−2 2.32 × 10−2 2.72 × 10−2
FR + TR 1.74 × 10−2 1.99 × 10−2 8.10 × 10−3 9.86 × 10−3 1.10 × 10−2 1.06 × 10−2 9.83 × 10−3 9.53 × 10−3 2.19 × 10−2 2.46 × 10−2 2.02 × 10−2 2.72 × 10−2
RT + TR 1.74 × 10−2 1.78 × 10−2 8.16 × 10−3 7.82 × 10−3 1.24 × 10−2 1.18 × 10−2 1.04 × 10−2 1.02 × 10−2 2.17 × 10−2 2.49 × 10−2 2.11 × 10−2 2.26 × 10−2
QR + RT 1.77 × 10−2 1.73 × 10−2 1.14 × 10−2 1.21 × 10−2 1.17 × 10−2 1.17 × 10−2 9.61 × 10−3 9.52 × 10−3 2.15 × 10−2 2.29 × 10−2 2.06 × 10−2 2.04 × 10−2
TR + RT + QR 2.57 × 10−2 2.54 × 10−2 1.54 × 10−2 1.58 × 10−2 1.76 × 10−2 1.67 × 10−2 1.54 × 10−2 1.52 × 10−2 3.22 × 10−2 3.42 × 10−2 3.06 × 10−2 3.51 × 10−2
TR + RT + GA 2.66 × 10−2 2.83 × 10−2 1.46 × 10−2 1.62 × 10−2 1.79 × 10−2 1.80 × 10−2 1.58 × 10−2 1.56 × 10−2 3.36 × 10−2 3.46 × 10−2 3.03 × 10−2 3.92 × 10−2
2.68 × 10−2 2.92 × 10−2 1.22 × 10−2 1.26 × 10−2 1.76 × 10−2 1.69 × 10−2 1.44 × 10−2 1.38 × 10−2 3.29 × 10−2 3.24 × 10−2 3.13 × 10−2 4.19 × 10−2

K.I. Berker et al. / Talanta 72 (2007) 1157–1165

TR + RT + FR
CF + FR + GA 3.03 × 10−2 4.16 × 10−2 1.44 × 10−2 1.98 × 10−2 1.45 × 10−2 1.60 × 10−2 1.47 × 10−2 1.82 × 10−2 3.62 × 10−2 3.94 × 10−2 3.24 × 10−2 4.18 × 10−2
CF + FR + TR 2.90 × 10−2 2.90 × 10−2 1.19 × 10−2 1.41 × 10−2 1.49 × 10−2 1.50 × 10−2 1.52 × 10−2 1.37 × 10−2 3.51 × 10−2 3.68 × 10−2 3.32 × 10−2 3.71 × 10−2
CF + FR + QR 2.93 × 10−2 3.09 × 10−2 1.52 × 10−2 1.56 × 10−2 1.42 × 10−2 1.42 × 10−2 1.44 × 10−2 1.32 × 10−2 3.48 × 10−2 3.76 × 10−2 3.28 × 10−2 3.68 × 10−2
GA + QR + RT 2.70 × 10−2 2.96 × 10−2 1.79 × 10−2 1.95 × 10−2 1.72 × 10−2 1.79 × 10−2 1.50 × 10−2 1.49 × 10−2 3.34 × 10−2 3.33 × 10−2 2.98 × 10−2 3.42 × 10−2
FR + QR + RT 2.71 × 10−2 2.97 × 10−2 1.56 × 10−2 1.63 × 10−2 1.68 × 10−2 1.62 × 10−2 1.36 × 10−2 1.31 × 10−2 3.27 × 10−2 3.22 × 10−2 3.09 × 10−2 3.53 × 10−2
FR + CF + RT 3.04 × 10−2 3.33 × 10−2 1.20 × 10−2 1.34 × 10−2 1.57 × 10−2 1.52 × 10−2 1.39 × 10−2 1.31 × 10−2 3.54 × 10−2 3.44 × 10−2 3.43 × 10−2 4.04 × 10−2
CF + TR + GA 2.88 × 10−2 4.00 × 10−2 1.42 × 10−2 2.10 × 10−2 1.53 × 10−2 1.62 × 10−2 1.66 × 10−2 1.98 × 10−2 3.58 × 10−2 3.58 × 10−2 3.22 × 10−2 3.75 × 10−2
AA + RT + QR 2.58 × 10−2 2.70 × 10−2 1.57 × 10−2 1.47 × 10−2 1.76 × 10−2 1.56 × 10−2 1.63 × 10−2 1.63 × 10−2 – – 3.06 × 10−2 4.26 × 10−2
GA + AA + CF 2.89 × 10−2 4.32 × 10−2 1.45 × 10−2 2.38 × 10−2 1.54 × 10−2 1.33 × 10−2 1.74 × 10−2 2.04 × 10−2 – – 3.21 × 10−2 5.06 × 10−2
QR + TR + GA 2.56 × 10−2 2.79 × 10−2 1.78 × 10−2 1.78 × 10−2 1.64 × 10−2 1.62 × 10−2 1.63 × 10−2 1.61 × 10−2 3.30 × 10−2 3.27 × 10−2 2.88 × 10−2 3.18 × 10−2
QR + AA + GA 2.57 × 10−2 2.64 × 10−2 1.80 × 10−2 1.83 × 10−2 1.65 × 10−2 1.47 × 10−2 1.71 × 10−2 1.75 × 10−2 – – 2.86 × 10−2 3.26 × 10−2
TR + GA + AA 2.53 × 10−2 2.60 × 10−2 1.47 × 10−2 1.55 × 10−2 1.72 × 10−2 1.62 × 10−2 1.79 × 10−2 1.73 × 10−2 – – 2.91 × 10−2 3.11 × 10−2
TR + QR + AA 2.44 × 10−2 2.59 × 10−2 1.55 × 10−2 1.58 × 10−2 1.70 × 10−2 1.54 × 10−2 1.77 × 10−2 1.70 × 10−2 – – 2.94 × 10−2 3.47 × 10−2
TR + AA + GA + QR 3.36 × 10−2 3.32 × 10−2 2.20 × 10−2 2.57 × 10−2 2.24 × 10−2 2.02 × 10−2 2.30 × 10−2 1.74 × 10−2 – – 3.86 × 10−2 5.42 × 10−2
TR + RT + FR + AA 3.49 × 10−2 3.57 × 10−2 1.65 × 10−2 2.06 × 10−2 2.35 × 10−2 2.08 × 10−2 2.10 × 10−2 1.60 × 10−2 – – 4.12 × 10−2 5.99 × 10−2
GA + AA + QR + RT 3.50 × 10−2 3.75 × 10−2 2.22 × 10−2 2.60 × 10−2 2.31 × 10−2 2.19 × 10−2 2.17 × 10−2 2.13 × 10−2 – – 3.98 × 10−2 5.68 × 10−2
FR + CF + QR + RT 3.87 × 10−2 4.26 × 10−2 1.94 × 10−2 1.99 × 10−2 2.08 × 10−2 1.78 × 10−2 1.90 × 10−2 1.87 × 10−2 4.58 × 10−2 4.58 × 10−2 4.39 × 10−2 4.57 × 10−2
FR + CF + QR + TR 3.73 × 10−2 3.80 × 10−2 1.92 × 10−2 2.02 × 10−2 2.01 × 10−2 1.77 × 10−2 2.03 × 10−2 1.91 × 10−2 4.56 × 10−2 4.72 × 10−2 4.28 × 10−2 4.88 × 10−2
n
TAC(expected) = c (TEAC)i , where TAC(expected) is the total antioxidant capacity in mM trolox (TR)-equivalents, Ci is the final concentration of antioxidant component (i) and (TEAC)i is its TEAC coefficient.
i=1 i
Thus, TAC(expected) = 0.0535 (0.2 × 1.0 × 10−3 × 1.00 + 0.2 × 5.0 × 10−4 × 2.07 + 0.2 × 4.0 × 10−4 × 2.78) = 3.36 × 10−5 M TR = 3.36 × 10−2 mM TR. TAC(found) was simply found by dividing the experimentally
recorded absorbance by εTR = 1.77 × 104 L mol−1 cm−1 , and making the dilution correction (i.e., 3.46 × 10−2 mM TR).
a The synthetic antioxidant mixtures were not prepared at identical concentrations for every iron-based assay employed, because the molar absorptivity (ε) of each antioxidant for a given assay was different from

that for another. The absorbances of the synthetic mixtures at the prespecified wavelengths were adjusted to roughly fall in the range 0.2 ≤ A ≤ 0.8 so as to minimize relative photometric error.
b A typical expected total antioxidant capacity (TAC) with respect to the original ferricyanide method was computed as follows: for a ternary mixture of TR (0.2 mL of 1.0 × 10−3 M solution) + RT (0.2 mL of

5.0 × 10−4 M solution) + GA (0.2 mL of 4.0 × 10−4 M solution), an aliquot of 2.5 mL was drawn from a 8.5 mL-incubated mixture, and color development was made after the addition of Fe(III) in a total volume of
5.5 mL. Therefore, the final concentrations of antioxidants could be calculated from their initial values by multiplying with the volume ratio of 1/8.5 × 2.5/5.5 = 0.0535.

1163
1164 K.I. Berker et al. / Talanta 72 (2007) 1157–1165

(Table 1). With the exception of fast-reacting antioxidants such

as trolox and ascorbic acid, the molar absorptivity (i.e., slope of
the calibration line in Table 1) of all antioxidants tested increased
upon incubation, pointing out to an increase in sensitivity in
determining these relatively slower-reacting antioxidants with
modified iron-based assays. A rule-of-thumb states that every
10 ◦ C elevation of temperature increases the reaction rate by a
factor of 2. Thus, an extended reaction period of 20–30 min at
50 ◦ C (in modified procedures) enabled more complete oxida-
tion of antioxidants with the tested Fe(III)-based reagents (see
Table 1). It should be noted here that high-spin iron(III) has
a d5 -electronic configuration with stable half-filled d-orbitals,
attributing a chemical inertness to it. It was previously shown
by many researchers that the FRAP method is incapable of assay-
ing either thiol- or carotenoid-type antioxidants [12], probably
due to this chemical inertness. In contrast, Apak et al. developed
the CUPRAC method [6,7] to enable the assay of thiols in aque-
ous solution or ␤-carotene in dichloromethane, since cupric ion
with neocuproine is a faster oxidant than ferric ion with TPTZ
or other ligands.
Possible binary, ternary, and quaternary mixtures of the
antioxidants (TR, AA, CF, FR, GA, QR, and RT) were syntheti- Fig. 2. The calibration line of trolox in pure aqueous solution and in green tea
cally prepared, and the suitably diluted solutions were analyzed infusion with respect to the 1,10-phenanthroline method.
for total antioxidant capacity (TAC) using the iron-based meth-
ods (see Table 2). The experimentally measured capacities were tea infusion were in accord with each other within a relative
generally in compliance with the theoretically computed values difference of 7% indicating parallelism, and the correlation
using the formula: coefficients (r) were higher than 0.9997. The two exceptional
deviations from parallelism of linear curves was notable for caf-
TACexpected = (TEAC)1 (concn.)1 + (TEAC)2 (concn.)2 feic acid in the modified FRAP method (relative slope difference
+ (TEAC)3 (concn.)3 + . . . + (TEAC)n (concn.)n 12.6%) and in the ferricyanide method (relative slope difference
14.8%, see Fig. 3), in compliance with the fact that hydroxycin-
where 1,2, . . ., n denote the corresponding constituents of the namic acids like caffeic and ferulic acid showed high intercept
synthetic mixture. The comparison of expected and experimen- values in this method (see Table 1). The slow reaction of caf-
tally found antioxidant capacities of synthetic mixture solutions
(as mM trolox-equivalents) were made, and depicted in Table 2.
The expected and experimentally found Fe(III)-based antioxi-
dant capacities were generally in accord with each other with the
exception of those found by the modified ferricyanide method,
which was shown to have high intercept values in Table 1.
The accordance of theoretical and experimental findings, com-
bined with the parallelism of the linear calibration curves of
each antioxidant compound tested in the presence of the other
compound (figures not shown), effectively demonstrated that
there were no chemical interactions that would cause appar-
ent deviations from Beer’s Law among the synthetic solution
constituents, and that the antioxidant capacities of the tested
antioxidants were additive. This reasoning was also applied to
green tea infusions as the real complex mixture containing natu-
ral antioxidants, and standard calibration curves of three selected
antioxidant compounds (TR, GA, and CF) were redrawn in a
solution of green tea infusion (see Section 2.3), showing good
parallelism of linear curves in pure aqueous solution and in
a real complex mixture having an initial non-zero absorbance
with the iron(III)-based reagent (e.g., see Fig. 2 for parallel cal-
ibration lines of trolox alone and in green tea infusion, drawn
with respect to Fe(III)–phen method). In general, the slopes of Fig. 3. The calibration line of caffeic acid in pure aqueous solution and in green
the calibration curves of TR, GA, and CF alone and in green tea infusion with respect to the modified ferricyanide method.
 K.I. Berker et al. / Talanta 72 (2007) 1157–1165
feic and ferulic acids [13] in the FRAP assay was established
before, hinting to the possibility of different levels of oxida-
tion of these hydroxycinnamic acids with the modified FRAP
and ferricyanide methods. The lower additivity of TAC values
in the modified ferricyanide method (Table 2), combined with
the loss of parallelism of linear curves in Fig. 3, implies that
further attention is needed to optimize the modified ferricyanide
method in spite of the advantages achieved over the original
ferricyanide method as regards the increased molar absorptivi-
ties and batochromic shift of maximum absorbance wavelength.
Data collected for the five Fe(III)-based assays (out of six assays)
definitely showed that the constituents of a real matrix solu-
tion did not chemically interact with selected pure antioxidants,
and that the antioxidant capacities of complex mixtures were
additive. Thus the tested standard and modified iron(III)-based
methods may be effectively used for the antioxidant capacity
assay of synthetic mixtures and real solutions such as plant
extracts.
4. Conclusions
In this work, ferric ion-based total antioxidant capacity (TAC)
assays have been comparatively tested, modified, and improved
so as to obtain more sensitive and precise results for complex
mixtures. Two new assays in this regard (i.e., o-phen and batho-
phen) have been established, and the existing assays (FRAP
and ferricyanide) have been modified so as to let the oxidation
reactions of antioxidants reach completion. The molar absorp-
tivity for a variety of antioxidants was highest for modified
FRAP, batho-phen, and original FRAP methods. The absorp-
tion maximum wavelength shifted batochromically to a higher
extent for modified ferricyanide, FRAP, and batho-phen proce-
dures, decreasing the possibility of interferences due to organics
absorbing in the near-UV range of the visible spectrum where
most antioxidant assays are performed. It was also shown that
the linear concentration ranges may be further extended and lin-
ear correlation coefficients improved with respect to the most
widely used ferric-based assay, FRAP. The procedures o-phen,
batho-phen, and modified FRAP gave low intercept values of
linear absorbance versus concentration equations, increasing
precision, while the modified ferricyanide procedure gave high
intercept values and low addivitity of TAC values of constituents
in complex mixtures, requiring further attention of method opti-
mization. Thus, it may be concluded that the most widely used
FRAP may be effectively modified, and o-phen, batho-phen,
and ferricyanide methods may constitute cheaper alternatives
to FRAP under certain conditions, with partly improved molar
absorptivity (and thus sensitivity) for antioxidants, lower inter-
cept values, broader linear range, and better additivity of TAC
values of antioxidant constituents in mixtures. It is natural that
the results obtained with a variety of electron transfer-based
antioxidant assays would give comparable but not identical
results for antioxidants due to the diversity of reaction conditions
such as redox potential, pH, and kinetics of these assays [8]. For
1165
iron(III) complex reagents where the coordination number of
ferric ion is saturated (i.e., o-phen, batho-phen, and ferricyanide
methods reagents, where there are exactly three bis-chelating
ligands or six monodentate ligands per Fe(III) ion in the reagent
solution), the undesired possibility of redox cycling of iron (e.g.,
the criticism directed to FRAP [13]) is not likely, as free Fe(III)
does not take part in the main redox reaction with the antioxidant,
and the reduced iron(II) is always bound in a stable complex.
Acknowledgements
One of the authors (K. Isil Berker) would like to thank
Istanbul University Research Fund, Bilimsel Arastirma Pro-
jeleri Yurutucu Sekreterligi, for the support given to her M.Sc.
thesis Project T-455/25062004. The present work was sup-
ported by the Research Fund of Istanbul University, Project No.
UDP-825/19072006. The authors would like to extend their grat-
itude to Istanbul University Research Fund, Bilimsel Arastirma
Projeleri Yurutucu Sekreterligi, for the funding of Project YOP-
4/27052004, to TUBITAK (Turkish Scientific and Technical
Research Council) for the research Project 105T402, and to State
Planning Organization of Turkey for the Advanced Research
Project of Istanbul University (2005K120430).
References
[1] B. Ou, D. Huang, M. Hampsch-Woodill, J.A. Flanagan, E.K. Deemer, J.
Agric. Food Chem. 50 (2002) 3122.
[2] N.J. Miller, C.A. Rice-Evans, M.J. Davies, V. Gopinathan, A. Milner, Clin.
Sci. 84 (1993) 407.
[3] C. Sanchez-Moreno, J.A. Larrauri, F. Saura-Calixto, J. Sci. Food Agric. 76
(1998) 270.
[4] G. Cao, C.P. Verdon, A.H.B. Wu, H. Wang, R.L. Prior, Clin. Chem. 41
(1995) 1738.
[5] I.F.F. Benzie, J.J. Strain, Anal. Biochem. 239 (1996) 70.
[6] R. Apak, K. Güçlü, M. Özyürek, S.E. Karademir, J. Agric. Food Chem. 52
(2004) 7970.
[7] R. Apak, K. Güçlü, M. Özyürek, S.E. Karademir, M. Altun, Free Radic.
Res. 39 (2005) 949.
[8] D. Huang, B. Ou, R.L. Prior, J. Agric. Food Chem. 53 (2005) 1841.
[9] I.F.F. Benzie, Y.T. Szeto, J. Agric. Food Chem. 47 (1999) 633.
[10] L.L. Stookey, Anal. Chem. 42 (1970) 779.
[11] R. Pulido, L. Bravo, F. Saura-Calixto, J. Agric. Food Chem. 48 (2000)
3396.
[12] G. Cao, R.L. Prior, Clin. Chem. 44 (1998) 1309.
[13] B. Ou, M. Hampsch-Woodill, R.L. Prior, J. Agric. Food Chem. 49 (2001)
4619.
[14] A. Besada, Talanta 34 (1987) 731.
[15] I.F.F. Benzie, J.J. Strain, Meth. Enzymol. 299 (1999) 15.
[16] A. Lugasi, J. Hóvári, K.V. Sági, L. Bı́ró, Acta Biol. Szegediensis 47 (2003)
119.
[17] K. Robards, P.D. Prenzler, G. Tucker, P. Swatsitang, W. Glover, Food Chem.
66 (1999) 401.
[18] C.A. Rice-Evans, N.J. Miller, G. Paganga, Free Radic. Biol. Med. 20 (1996)
933.
[19] K. Güçlü, K. Sözgen, E. Tütem, M. Özyürek, R. Apak, Talanta 65 (2005)
1226.
[20] N. Salah, N.J. Miller, G. Paganga, L. Tijburg, G.P. Bolwell, C.A. Rice-
Evans, Arch. Biochem. Biophys. 322 (1995) 339.