journal of functional foods 14 (2015) 111–125

Available online at www.sciencedirect.com

ScienceDirect

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j ff

Hurdles and pitfalls in measuring antioxidant

efficacy: A critical evaluation of ABTS, DPPH, and
ORAC assays

K.M. Schaich *, X. Tian, J. Xie

Department of Food Science, Rutgers University, 65 Dudley Rd., New Brunswick, NJ 08901-8520, USA

A R T I C L E I N F O A B S T R A C T

Article history: Assays developed to measure radical scavenging ability of natural compounds have been
Received 1 October 2014 used as a basis for ranking and recommending best foods for consumption. However, assays
Received in revised form 1 January often were adapted for screening assays with inadequate consideration of reaction chem-
2015 istry, particularly kinetics. Recent research results raise serious questions about the chemistry,
Accepted 23 January 2015 execution, and application of these assays. This paper critically evaluates conceptual and
Available online technical issues that limit use and compromise validity of three commonly-used assays –
TEAC/ABTS•+, DPPH, and ORAC. Recommendations are made for discontinuing use of
Keywords: ABTS•+ and DPPH radicals for measuring radical quenching, redirecting them instead to dis-
Antioxidant efficacy assays tinguishing electron transfer reaction mechanisms. Conditions required for accurate results
Limitations in ORAC are reviewed, and recommendations are made for redirecting this assay to distin-
TEAC – Trolox equivalent guishing compounds that quench radicals by hydrogen atom transfer. The mechanistic
antioxidant capacity assay information so gained can be then applied to understanding how natural antioxidants can
•+
ABTS – 2,2′-azino-bis(3- be used most effectively in foods.
ethylbenzothiazoline-6-sulfonic © 2015 Elsevier Ltd. All rights reserved.
acid) assay
ORAC – Oxygen Radical Absorbance
Capacity assay
Radical quenching kinetics

Contents

1. Introduction ...................................................................................................................................................................................... 112

2. Hurdles and pitfalls in three common antioxidant assays ....................................................................................................... 113
2.1. TEAC/ABTS•+ – Trolox equivalent antioxidant capacity/2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) assay . 113
2.1.1. TEAC I ........................................................................................................................................................................... 114
2.1.2. TEAC II .......................................................................................................................................................................... 114
2.1.3. Limitations of the TEAC/ABTS•+ assay .................................................................................................................... 114

* Corresponding author. Department of Food Science, Rutgers University, 65 Dudley Rd., New Brunswick, NJ 08901-8520, USA. Tel.: +1 848
932 5454; fax: 732 932-6776.
E-mail address: schaich@aesop.rutgers.edu (K.M. Schaich).
Abbreviations: SET, single electron transfer; HAT, hydrogen atom transfer; TEAC, Trolox equivalent antioxidant capacity; ABTS•+, 2,2′-
azino-bis(3-ethylbenzothiazoline-6-sulfonic acid); ORAC, Oxygen Radical Absorbance Capacity; AAPH, 2,2′-azobis(2-amidinopropane)
dihydrochloride; IU, International Units; CuPRAC, Cupric ion reducing antioxidant capacity; FRAP, Ferric reducing antioxidant power
http://dx.doi.org/10.1016/j.jff.2015.01.043
1756-4646/© 2015 Elsevier Ltd. All rights reserved.
112 journal of functional foods 14 (2015) 111–125

2.1.4. Summary and recommendations – TEAC/ABTS•+ assay ...................................................................................... 116

2.2. DPPH (2,2′-diphenylpicryl hydrazyl free radical) assay ................................................................................................... 116
2.2.1. Limitations of the DPPH assay ................................................................................................................................. 116
2.2.2. Summary and recommendations – DPPH assay ................................................................................................... 118
2.3. ORAC – Oxygen Radical Absorbance Capacity ................................................................................................................. 119
2.3.1. Advantages of the ORAC assay ................................................................................................................................ 119
2.3.2. Limitations of the ORAC assay ................................................................................................................................. 119
2.3.3. Summary and recommendations – ORAC assay ................................................................................................... 123
3. Summary and recommendations ................................................................................................................................................. 123
References ......................................................................................................................................................................................... 123

limitation is that, although antioxidant assays were designed

1. Introduction
Recognition of potential health effects of antioxidants more
than twenty years ago stimulated what could be called the “An-
tioxidant Bandwagon” of research seeking to determine which
natural materials contain the highest levels of the most active
antioxidants, addition of antioxidants to beverages and all forms
of foods, prophylactic and therapeutic medical applications,
and hyper marketing. At the same time, numerous in vitro an-
tioxidant assays were developed to measure radical scavenging
and to compare and rank antioxidants in different foods
and natural materials. Antioxidant assays have measured three
values:
• antioxidant capacity – the total number of electrons donated
or target molecules converted per mol of antioxidant at
full reaction under given conditions. This usually approxi-
mates the number of phenolic —OH groups, or two electrons
per —OH group, though not always. Unfortunately, the re-
quirement for full reaction ignores reaction rate and creates
a kind of “tortoise and hare” situation where slow react-
ing antioxidants with many phenol groups receive highest
rankings while fast reacting antioxidants with few phenol
groups can be greatly underestimated or even overlooked.
• antioxidant activity – the antioxidant concentration re-
quired to provide a specified rate or extent of reaction
• antioxidant potential – a nebulous general term used to
describe the expectation that an antioxidant can quench
radicals under specific conditions. This term can easily be
confused with thermodynamic potential, so its applica-
tion to antioxidant assays has been controversial.
Assays based on these measures have been applied so ex-
tensively to evaluation of fruits, vegetables, leaves and stems,
herbs and spices that TEAC, DPPH, ORAC, CuPRAC, and FRAP
are now household words in scientific publications, health food
publications, and internet websites. Thousands of research
papers have been published comparing values of different foods,
and these results have provided the basis for recommending
the best foods to eat, for labeling Super Foods, and for mar-
keting products.
Despite extensive use of in vitro antioxidant assays to dis-
tinguish which antioxidants and foods are most efficacious,
time has shown that the assays have both conceptual and tech-
nical limitations (Apak et al., 2013). One major conceptual
to identify which antioxidants should be expected to provide
greatest protective effects against free radicals in vivo, the radical
scavenging observed in test tubes probably does not occur in
vivo. Most antioxidants are poorly absorbed or are rapidly con-
jugated and eliminated in the urine, so circulating phenol
concentrations reach trace levels at best. It is unlikely that
such minute amounts can mediate physiological actions of
antioxidants merely by scavenging radicals. Reviews of the
antioxidant literature suggest that microbial or physiological
metabolites of antioxidants may actually be more active in vivo
(Frankel & Finley, 2008), but such metabolites remain largely
unidentified and they are not the compounds tested in in vitro
assays. Cell culture assays attempt at least in part to include
effects of metabolites, but most cell types used in assay cul-
tures never see antioxidants in vivo. Phenols bind readily to
proteins in cells and foods, so they may be inactivated for clas-
sical radical quenching or activated for some other physiological
action not tested in the assays. Finally, original assumptions
regarding radical scavenging being the mode of antioxidant
action in vivo may well be erroneous in light of new evidence
regarding signal transduction and regulation of gene expres-
sion of redox controlling and detoxifying enzymes (Frankel,
2007; Halliwell, Rafter, & Jenner, 2005). Indeed, the U.S. De-
partment of Agriculture recently withdrew the compiled tables
of ORAC values of many foods, herbs, and spices which had
been posted on their website for some time, leaving behind
the following statement (USDA, 2010):
“There is no evidence that the beneficial effects of
polyphenol-rich foods can be attributed to the antioxi-
dant properties of these foods. The data for antioxidant
capacity of foods generated by in vitro (test-tube) methods
cannot be extrapolated to in vivo (human) effects and the
clinical trials to test benefits of dietary antioxidants have
produced mixed results. We know now that antioxidant mol-
ecules in food have a wide range of functions, many of which
are unrelated to the ability to absorb free radicals. For these
reasons the ORAC table, previously available on this web site
has been withdrawn.”
A second conceptual limitation is that the chemistry and
molecular targets of most in vitro assays are not relevant to in
vivo conditions. Antioxidant concentrations used in these assays
are orders of magnitude higher than would ever be found in
vivo, the assays use sterically-hindered stable radicals (e.g.
 journal of functional foods 14 (2015) 111–125
ABTS•+ or DPPH) as targets rather than small, readily acces-
sible but short-lived radicals such as HO•, O2−•, or lipid oxyl
radicals that are active in vivo, and they follow reaction times
much longer than radical lifetimes in vivo (ranging from 10 s
for peroxyl radicals to 10−9 s for hydroxyl radicals). If an anti-
oxidant requires many minutes to hours to quench radicals,
its action as a radical scavenger must be irrelevant in vivo in
cells or even in situ in foods.
A third conceptual limitation is that the assays in common
use do not address radical reactions in lipids, when mem-
brane lipids are clearly involved in oxidation of both membranes
and foods. A few lipophilic versions of assays have been de-
veloped (Alcolea, Cano, Acosta, & Arnao, 2003; Arnao, Cano,
& Acosta, 2001; Huang, Ou, Hampsch-Woodill, Flanagan, &
Deemer, 2002; Jimenez-Alvarez et al., 2008; Prior et al., 2003),
but these approaches mostly solubilize lipophilic antioxi-
dants for reaction in aqueous phase rather than testing
reactions in lipid phases. As a result, relatively little informa-
tion is available about how natural antioxidants partition into
and interact with lipids.
Technical limitations are no less problematic. First and fore-
most, the assays have unclear chemistry for quantitative
analyses. They were developed for rapid screening without de-
tailed investigations of underlying initiators, targets, antioxidant
interactions, kinetics, solvent effects, concentration effects, etc.
Lacking full understanding of the chemistry, control issues such
as oxygen (dissolved and atmospheric), light, temperature,
reagent concentrations, solvent, pH, and running blanks are rou-
tinely ignored. Most assays measure stoichiometry (capacity)
rather than kinetics of antioxidant reactions, i.e. how much reacts
rather than how fast. Ignoring the chemistry leads to incorrect
assessment of antioxidant activity, both quantitatively and quali-
tatively. In addition, each laboratory has adapted procedures to
work with instrumentation and capabilities available. As a result,
there is currently almost total lack of standardization in ex-
perimental procedures in the assays, particularly in expressing
results where methods are quite varied and often meaning-
less in terms of action in situ. Finally, most “screening” assays
and natural antioxidants are water-based and ignore lipid phases.
After more than twenty years of antioxidant assays, it is clear
that fruits and vegetables, herbs, and spices have antioxi-
dants and that antioxidant activity generally correlates with
content of total phenols in natural products. What remains
unclear is which fruits and vegetables are “best”, whether and
how antioxidants interact with each other in assays and with
other compounds in vivo and in foods, and how in vitro assay
results correlate with multiple in situ actions of antioxidants
in foods and biological tissues. Thus, when antioxidant assays
are used for screening, results must be taken with a prover-
bial grain of salt.
Recent attempts to standardize assays have raised new ques-
tions about reliability and appropriate applications of antioxidant
assays:
1. Do the assays accurately reflect antioxidant chemistry
occurring in vivo? Increasing evidence suggests that radical
scavenging may occur in the gastro-intestinal tract but other
actions of antioxidants may be even more important there
and systemically. Thus, there are critical needs to move
beyond these simplistic assays to find accurate models that
113
test signaling processes of antioxidants, to conduct de-
tailed pharmacokinetic analyses in animals – where do
antioxidants go, what do they do, and how do they break
down? – and to identify other actions of antioxidants.
2. Do the assays accurately reflect radical quenching capabil-
ity of antioxidants with different structures or in mixtures?
3. How can the assays be used to predict effectiveness or op-
timize usage of natural antioxidants in foods?
This critical evaluation focuses on Question 2, with em-
phasis on ABTS•+, DPPH, and ORAC assays.
2. Hurdles and pitfalls in three common
antioxidant assays
Radicals are most commonly quenched by two mechanisms,
transfer of either a hydrogen atom or an electron to convert
the radical to a stable species:
HAT – hydrogen atom transfer (H atom transferred to target
radical, possible secondary quenching by radical recombina-
tions).
ROO• + AH PheOH → ROOH + A • Phe-O• (1)
ROO• + A • → ROOA (2)
SET – single electron transfer (one or more electrons
transferred to reduce target compounds).
ROO• + AH PheOH
+ H2O (3)
→ ROO− + AH+ PheOH+ ⎯⎯⎯ → A • Phe-O• + H3O+
ROO− + H3O+ → ROOH + H2O (4)
M (III) + AH ArOH → AH+ ArOH+ + M (II) (5)
where AH = any antioxidant with donatable H, PheOH = phenol
or polyphenol, M = redox-active metal.
While the end result of the two mechanisms may be the
same, kinetics and dependence on system conditions, espe-
cially solvent and pH, vary tremendously. Electron transfer is
very fast (femtoseconds) (Ganapathi, Hermann, Naumov, &
Brede, 2000), it is not diffusion-controlled, and it increases with
pH as ionization increases availability of electrons. Hydrogen
atom transfer, in contrast, is considerably slowed by diffu-
sion; it is independent of pH but strongly enhanced by water
and inhibited by hydrogen bonding solvents such as alco-
hols. These differences actually have critical impacts on how
antioxidants act in complex, multiphase systems.
Effects of electron versus hydrogen atom transfer can be
seen in the three most commonly-used assays.
2.1. TEAC/ABTS•+ – Trolox equivalent antioxidant
capacity/2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic
acid) assay
TEAC measures antioxidant capacity as the ability of test an-
tioxidants (AH) to decrease ABTS•+ color by (a) intercepting initial
114 journal of functional foods 14 (2015) 111–125
oxidation and preventing ABTS•+ production, or (b) reacting
directly with the ABTS•+. Two versions of the assay are in use,
based on these two approaches.
2.1.1. TEAC I
TEAC 1 utilizes metmyoglobin-H 2O2 to generate hydroxyl
radicals, which oxidize ABTS to its colored free radical form,
ABTS•+ (Rxn 6a); subsequent reaction with antioxidants (AH)
results in loss of the green color (Rxn 6b) (Miller & Rice-Evans,
1997; Miller, Rice-Evans, Davies, Gopinathan, & Milner, 1993).
This version is available in commercial kits. However, the
reaction is ambiguous because antioxidants (AH) can react with
the original HO• radical oxidant, the metmyoglobin, and the
ABTS•+, causing an overestimation of antioxidant activity
(Strube, Haenen, van den Berg, & Bast, 1997). Furthermore, the
antioxidant reaction mechanism cannot be determined. Are
different mechanisms active with each radical, e.g. HAT with
HO• and SET with ABTS•+? Do active mechanisms vary with
phenol structure? It is not possible to accurately compare
antioxidants without this information.
2.1.2. TEAC II
To circumvent the problems of double reaction, ABTS•+ can be
generated directly in high yield using potassium persulphate
as the oxidizing agent (Re et al., 1999). Antioxidants then
react only with ABTS•+ and color is diminished by only one
reaction (Rxn 7):
For reaction, aliquots of ABTS•+ solution are diluted to an
absorbance of ~1.0 at 734 nm, which is recorded as the start-
ing point. The antioxidant is added and mixed, and the drop
in absorbance (A0–Af) is typically measured after reaction periods
varying from minutes to hours. Antioxidant action is re-
ported as Trolox equivalents by comparing (A0–Af) of the test
antioxidant to (A0–Af) of Trolox standards, or the antioxidant
concentration that gives the same response as 1 mM Trolox
(Rxn 8):
(A 0 − A f )sample − blank
TEAC ( ABTS) = (8)
(A 0 − A f )1 mM Trolox − blank
2.1.3. Limitations of the TEAC/ABTS•+ assay
Despite the extensive use of the ABTS•+ reaction to screen
antioxidant activity in a wide range of materials, serious flaws
in design and performance compromise the usefulness and
even the validity of this assay. First, the most important chem-
istry in this reaction – the antioxidant reaction rate and kinetic
pattern – is totally ignored. The absorbance drop (A0–Af) upon
which TEAC calculations are based provides only reaction stoi-
chiometry (total mols ABTS•+ reduced per mol antioxidant after
full reaction). Usually two electrons can be donated per phenol
—OH group, so antioxidant capacity as measured in TEAC
should be closely predictable from antioxidant structure. Indeed,
under these conditions, polyphenols are highly favored and
other structural effects on antioxidant reactivity are ob-
scured. TEAC correlation with number of phenol groups have
been claimed, but we found the relation held primarily within
common structural groups and that other factors have
stronger control over antioxidant reaction with ABTS •+
(Tian & Schaich, 2013).
Recording the ABTS•+ absorbance continuously during re-
action rather than at the beginning and the end generates
curves from which reaction kinetics can be calculated (Fig. 1).
These curves also reveal marked differences between antioxi-
dant behaviors that reflect whether SET or HAT reaction
mechanisms are dominant. Some antioxidants react com-
pletely in less than mixing time, some react slowly and
gradually, and others have mixed fast and slow initial reac-
tions (Tian & Schaich, 2013). These differences in antioxidant
reaction rates are critical because they reflect ability to quench
short-lived reactive radicals that are present in biological
systems and in foods. Such capability is missed entirely in
normal assay procedures.
What molecular properties account for differences in initial
reaction rates and patterns? Only key points will be reviewed
here; further details are available in Tian and Schaich (2013).
Two factors control reaction rates and shapes of response
curves: SET vs HAT mechanisms and steric access to the
hindered ABTS•+ radical site. Antioxidants acting by electron
transfer and having full access to ABTS•+ radical site react within
milliseconds. Reactions are slowed by the presence of mul-
tiple —OH and rings, bulky ring adducts, high antioxidant
concentrations, and hydrogen atom transfer. Thus, almost in-
stantaneous initial reaction (Group 1, Fig. 1) arises from electron
transfer. The reaction is so fast that accurate results can be ob-
tained only with rapid mixing methods such as autodispensing
plate readers or stopped-flow mixers, which take the assay out
of routine laboratory capabilities. Continued reaction after rapid
initial absorbance drop (Group 2, Fig. 1) results from phenol
structure (ring adducts) that interferes with access to ABTS•+.
Smaller initial absorbance drop with continued reaction af-
terwards (Group 3, Fig. 1) is still mostly electron transfer, but
now the phenol has bulky side groups or multiple rings that
impede diffusion and orientation toward ABTS•+; these com-
pounds may also act partially by slower hydrogen atom transfer.
No initial fast reaction but gradual drop from starting absor-
bance (Group 4, Fig. 1) arises from dominant HAT.
Critical controlling effects of steric accessibility are shown
quite dramatically in effects of antioxidant concentration on
reaction rate (Fig. 2). Small molecules and single phenols reduce
~1:1 ABTS•+ per phenolic —OH and reaction response remains
linear as antioxidant concentration increases (Group 1).
However, the reaction becomes increasingly impeded at higher
antioxidant concentrations as the structural complexity and
 journal of functional foods 14 (2015) 111–125 115

1.4 Group 1, inst & complete, SET mM 1.4 Group 2, very fast, SET 1.4 Group 3, fast SET, some steric
AOX hindrance or HAT
1.2 0 1.2 1.2
Absorbance, 734 nm

0.001
1.0 0.005 1.0 1.0
0.05
0.8 0.10 0.8 0.8
0.6 0.25 0.6 0.6
0.4 0.50 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0
0 2 4 6 0 2 4 6 0 2 4 6
Reaction time (min) Reaction time (min) Reaction time (min)

1.4 Group 4, hindered SET, HAT, or mixed 1.4 Group 5, HAT at low conc, SET at high Group 6, unreactive
1.4
1.2 1.2 1.2
Absorbance, 734 nm

1.0 1.0 1.0

0.8 0.8 0.8
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0
0 2 4 6 0 2 4 6 0 2 4 6
Reaction time (min) Reaction time (min) Reaction time (min)

Fig. 1 – Kinetic response curves showing different types of reactivity of phenolic compounds with ABTS+•. Adapted from
Tian and Schaich (2013), used with permission.

size of the phenol increases because the additional ring adducts reduced/mol antioxidant and per —OH group decrease, par-
(Group 2, Group 3-simple phenols) and especially secondary ticularly for polyphenols where reaction efficiency decreases
rings (Group 3-polyphenols) interfere with phenol access to the with each additional —OH. Such marked differences in rate pat-
hindered ABTS•+. As a result, rates slow and mols ABTS•+ terns mean that this assay must be conducted over a range

Group 1 0.8 Group 2 1.0 Group 3 - simple phenols

0.8
Hydroquinone Pyrogallol Gallic acid
Ascorbic
acid 3-Methylcatechol
0.8 2-Propyl gallate
0.6 Chlorogenic 0.6
Coniferyl alcohol
Initial ΔA

acid Catechol
0.6 Caffeic acid
Trolox
0.4 0.4
Protocatechuic
0.4 Ferulic acid
acid

0.2 0.2
0.2

0.0 0.0 0.0

0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
AOX Concentration (mM) AOX Concentration (mM) AOX Concentration (mM)

1.4 Group 3 - polyphenols 0.8 Group 4 0.8 Group 5

Resorcinol
Gallocatechin gallate
1.2
Catechin gallate p-Coumaric acid
Quercetin
Final ΔA

1.0 0.6 0.6

Gallocatechin
0.8 Glutathione
Catechin 0.4 0.4
0.6
0.4 Curcumin
0.2 0.2
0.2 Rutin

0.0 0.0 0.0

0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
AOX Concentration (mM) AOX Concentration (mM) AOX Concentration (mM)

Fig. 2 – Effects of antioxidant concentrations on reaction rates with ABTS+•. Adapted from Tian and Schaich (2013), used
with permission.
116 journal of functional foods 14 (2015) 111–125
of known phenol concentrations to obtain accurate activity
comparisons of compounds or extracts with different phenol
compositions.
Rate dependence on steric accessibility is further sup-
ported by lack of correlation with initial reaction rate and
stoichiometry or antioxidant properties such as redox poten-
tials, with only weak correlations (p = 0.75–.86) between TEAC
values and reaction rates with number of phenol —OH groups
(Tian & Schaich, 2013). TEAC stoichiometry generally in-
creased while the rate per —OH decreased with the number
of phenol —OH groups.
2.1.4. Summary and recommendations – TEAC/ABTS•+ assay
As a large nitrogen-centered and sterically-hindered radical,
ABTS•+ is not a good model for small highly-reactive radicals
[e.g. OH•, NO•, O2−•. LO(O)•] that are active in biological tissues
and foods. That steric accessibility rather than chemical prop-
erties of antioxidants controls the reaction invalidates the
chemical accuracy of this assay and its ability to predict
which antioxidants will be active in real situations. Negli-
gible hydrogen atom transfer rates with ABTS •+ further
compromise the usefulness of this assay because it totally
misses activity of compounds such as glutathione, thiols,
and proteins that are HAT-dominant and very important
antioxidants in vivo.
Van den Berg recognized problems with the TEAC fifteen
years ago, stating “Quantitative evaluation of antioxidant
capacity using the TEAC assay can be troublesome or even
impossible” (van den Berg, Haenen, van den Berg, & Bast, 1999).
We concur with his conclusion and, further, consider the
limitations of the TEAC/ABTS•+ assay fatal. We thus recom-
mend that the use of this assay to screen and compare
antioxidants of different structure be discontinued. This
assay does not provide information that is sufficiently
accurate on a chemical basis to be used for ranking of anti-
oxidants or for construction of SARS (structure–activity
relationships).
Nevertheless, the TEAC/ABTS•+ assay can be used to assess
whether antioxidants are HAT or SET dominant in their reac-
tions, and it can be used to compare changes in the same
antioxidant during processing or storage. For example, ABTS•+
has been used to monitor changes in tocopherol activity after
heat exposure in frying oils and in extrusion of packaging films
(Lang, 2009) and to determine loss of antioxidant activity in
strawberries dried with different methods (Wojdyło, Figiel, &
Oszmiański, 2009). In such applications, limitations of the assay
remain but antioxidant components are constant, and varia-
tions in ABTS•+ accessibility are not the prime determinant of
reactivity. Rather, alterations in a single antioxidant struc-
ture modify radical quenching capabilities.
2.2. DPPH (2,2′-diphenylpicryl hydrazyl free radical)
assay
DPPH is a stable radical with a deep purple color. Reaction
with other radicals, electrons, or hydrogen atoms leads to
loss of color at 515 nm and loss of the EPR free radical signal
(Rxns 9a–c).
NO2 NO2
O2N
.
N N O2 N
H
N
. N
NO2 NO2
Purple Colorless
DPPH• + Phe-OH → DPPH (H) + Phe-O• (9a)
DPPH• + Phe-O− → DPPH− + Phe-O• (9b)
DPPH• + Phe-OH → DPPH− + Phe-O+ (9c)
Like ABTS•+, DPPH is sterically-hindered with the radical center
highly protected (Foti, Daquino, Mackie, DiLabio, & Ingold, 2008;
Williams, 1966). In contrast to water-soluble ABTS•+, DPPH is
hydrophobic so its reactions must be run in organic solvents.
Although literature reports attribute DPPH reactions mostly at-
tributed to HAT, reactions in strong hydrogen-bonding solvents
such as methanol interfere with release of hydrogen atoms and
thus strongly enhance SET over HAT (Barclay, Edwards, & Vinqvist,
1999; Foti, Daquino, & Geraci, 2004; Foti et al., 2008). Results out-
lined in the Limitations discussion suggest that electron transfer
dominates in the organic solvents required to dissolve DPPH.
DPPH assays are performed very much like ABTS•+ reac-
tions.A stock solution of concentrated DPPH in methanol is diluted
to an absorbance of ~1 at 515 nm, the starting absorbance (A0)
is measured, and the antioxidant is added to start the reaction.
Traditional methodology then measures final absorbance (Af)
at 515 nm after reaction times that range from 4 minutes to
many hours, or even days. The quantity (A0–Af) is used to
generate a value for ranking tested materials, e.g. EC50 (antioxi-
dant concentration required to reduce DPPH 50%) (Bondet,
Brand-Williams, & Berset, 1997; Brand-Williams, Cuvelier, & Berset,
1995), %DPPH loss or %DPPH remaining (Burda & Oleszek, 2001;
Stratil, Klejdus, & Kuban, 2006), or Antiradical Efficiency
[1/[EC50(Sample) × EC50(Trolox)] (Awika, Rooney, Wu, Prior, &
Cisneros-Zevallos, 2003; Butkovic, Klasinc, & Bors, 2004).
2.2.1. Limitations of the DPPH assay
The DPPH reaction has been used as if it is a simplistic chemi-
cal “black box” – reagents are mixed and a number is generated,
and the chemistry occurring between is ignored. The calcula-
tion methods noted in the previous paragraph provide relative
rather than absolute values, and carry no direct chemical
information unless absorbance change is converted to mols DPPH
lost using Beer’s Law and extinction coefficient ε values of 10,900–
12,500 (Lebeau et al., 2000). Even more importantly, this assay
does not measure reaction rates and it misses important in-
formation contained in reaction curves. Hence, deleting reaction
rates can lead to erroneous assessment of antioxidant efficacy.
That antioxidant reactions with DPPH are actually rather
complex becomes obvious when absorbance changes are fol-
lowed continuously. Reaction curves show multiple reactivity
patterns (Fig. 3) that are similar to ABTS•+ but comparisons
of 28 phenolic compounds showed some key differences
 journal of functional foods 14 (2015) 111–125 117

Group 1 - Extremely fast (SET) mM AOX Group 2 - Fast (hindered SET)

1.0 Blank 1.0
1

Absorbance 515 nm
0.8 2.5
0.8
5

0.6 0.6

0.4 0.4
10

0.2 25 0.2
Ø
1000
0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
Reaction time (min) Reaction time (min)

Group 3 - Moderate, delayed (mixed) Group 4 - Slow (HAT)

1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
Reaction time (min) Reaction time (min)

Fig. 3 – Kinetic patterns of phenol reactions with DPPH. From Xie and Schaich (2014), used with permission. Examples for
each group are shown. Antioxidants showing similar reaction patterns: Group 1 (instantaneous initial absorbance drop) –
n-propyl gallate, pyrogallol, and ascorbic acid; Group 2 (rapid initial absorbance drop, then reaction continues more slowly)
– catechol, 3-methylcatechol, α-tocopherol, caffeic acid, gallic acid, quercetin, epicatechin; Group 3 (moderate, delayed initial
reaction) – hydroquinone, protocatechuic acid, rosmarinic acid, Trolox, ferulic acid, chlorogenic acid; Group 4 (absorbance
drop very slow and almost linear over time) – glutathione, resorcinol, and vanillic acid.

(Xie & Schaich, 2014). Antioxidants react with DPPH by very
rapid electron transfer and by slow hydrogen atom transfer. Initial
reactions (electron transfer) are quite rapid but slower than ABTS•+
reactions due to difficult phenol access to the hindered DPPH
radical site. Hindered access impedes all reactions but espe-
cially hydrogen atom transfer for which formation of a hydrogen
bonded complex between the radical α-C—H and the nitrogen
lone pair is required (Salamone, DiLabio, & Bietti, 2011; Salamone,
Martella, & Bietti, 2012). Compounds reacting rapidly (domi-
antioxidant and extract responses over a reasonable phenol
concentration range must be determined to understand
what is happening in the reactions, especially when phenol
structures are different or unknown.
DPPH reactions are highly sensitive to reaction environ-
ment, i.e. water and solvent (even more than ABTS•+), pH, oxygen,
1
Pyrogallol
Initial rate, nmol DPPH/sec

nant electron transfer) are ascorbic acid and simple phenols

with no ring adducts; reactions slow with addition of acid groups 0.8
Catechol
or other side chains to aromatic rings, and there are some
inductive effects of ring adducts within phenol classes.
0.6
Effects of steric hindrance to DPPH are exhibited quite dis-
tinctively in plots of initial rates as a function of antioxidant
concentration (Fig. 4) (Xie & Schaich, 2014). Similar to ABTS•+, 0.4
Hydroquinone
small compounds with fast initial rates (Group 1, e.g. pyrogal-
lol) show a linear increase at low concentrations, but the linear 0.2
reaction range shrinks and saturation concentrations de- Resorcinol
crease with addition of bulky ring adducts and multiple rings. 0
Molecules with complex structures more readily get in the way 0 0.2 0.4 0.6 0.8 1
of each other and impede access to DPPH at low concentra- [AOX], mM
tions and they strongly block reaction at high concentrations.
These results show that single antioxidant concentrations are Fig. 4 – Effects of phenol concentrations on rates of phenol
clearly inadequate to quantitate and compare antioxidant re- reactions with DPPH. Only samples from each group are
activity with DPPH, that phenol concentrations of extracts must shown. Full details for all tested compounds are available
be known for accurate comparisons, and that the patterns of in Xie and Schaich (2014).
118 journal of functional foods 14 (2015) 111–125
light exposure (Xie & Schaich, 2014). The presence of both elec-
tron and hydrogen atom transfer mechanisms is particularly
evident in variability of DPPH reactions with solvent. Methanol,
the solvent most commonly used for the DPPH assay, strongly
binds hydrogen atoms and inhibits HAT processes (Barclay et al.,
1999; Foti et al., 2004; Hogg, Lohmann, & Russell, 1961). However,
water disrupts the binding and facilitates H atom transfer, so
when it is added to the reaction medium, any test compounds
with H atom transfer capabilities increase their reaction rate.
Similarly, electron transfer is pH-dependent, increasing in rate
with pH and degree of ionization, while HAT is pH-independent.
The dominant mechanism of a test compound can thus be as-
sessed by reacting it with DPPH in methanol and in 50% methanol
in which the aqueous phase is buffered to a range of pH from
acid to strongly alkaline (Xie & Schaich, 2014). Reaction rates of
compounds that are HAT-dominant (e.g. glutathione) increase
markedly in 50% water but not with pH. In contrast, SET-
dominant compounds such as hydroquinone will react extremely
fast in methanol, show small change in 50% water, then in-
crease markedly with pH of the water phase. The strength of
SET processes for individual antioxidants can thus be evalu-
ated by plotting DPPH initial reaction rates as a function of pH.
High slopes indicate dominant electron transfer (Xie & Schaich,
2014).
DPPH reactions also vary with dissolved oxygen in the re-
action mixtures (Xie & Schaich, 2014). Compounds that autoxidize,
e.g. ascorbic acid (with metal catalysts) and some phenols, reduce
oxygen to O2•− which then enters the reaction mix:
O2−• reacts instantaneously with ABTS•+, leading to aberrantly
high antioxidant activity, while it reacts poorly with DPPH•,
resulting in aberrantly low antioxidant reactivity because
oxidation of phenols to less reactive quinones or unreactive
products competitively removes them from the reaction and
reduces detected reactivity. Thus, a good working rule of thumb
to eliminate potential artifacts is to sparge all solutions with
argon before running the DPPH and ABTS•+ reactions.
An additional issue with DPPH assays is what happens in
mixtures of compounds. Given problems with steric inaccessi-
bility of the DPPH radical site, do DPPH reactions accurately reflect
radical quenching capabilities of mixtures of antioxidants, such
as would be found in fruit and vegetable extracts? Can they detect
synergisms that have been reported from in vivo feeding studies,
or do multiple compounds in mixtures interfere with each other
reaching the DPPH radical? Studies in our laboratory show that
reaction of phenol mixtures with DPPH is nearly always less than
the sum of individual phenol components because steric factors
control the reactions (G.Wu and K.M. Schaich, unpublished data).
The response shifts with the proportion of components and
balance between HAT and SET, synergisms and antagonisms are
difficult to detect, and ascorbic acid, when present, dominates
reactions. This is a critical limitation when comparing mix-
tures of compounds in extracts and infusions.
Finally, for any quantitative assay to be valid, reactivity
must be directly related to chemical properties of the test com-
pounds, and the reaction mechanism must be constant.
Unfortunately, our results suggest that neither of these re-
quirements can be met in the DPPH assay. Steric accessibility
to the DPPH radical site is the overall most important driver
of reaction rate, with secondary contributions from balance
between electron and hydrogen atom transfer quenching
mechanisms. In contrast, both initial reaction rate and final
stoichiometry are poorly correlated with molecular structure
and chemical characteristics, including number of phenolic
—OH groups and redox potential (Xie & Schaich, 2014). These
observations raise serious questions about using the DPPH assay
to rank antioxidant efficacy of foods and extracts.
2.2.2. Summary and recommendations – DPPH assay
Like ABTS•+, DPPH is a poor model for radical quenching in vivo
and in foods because it is nitrogen-centered and the radical
site is highly hindered. Questions must be raised regarding use
of DPPH in antioxidant screening and quantitation of radical
quenching in extracts of natural materials for the following
reasons (Xie & Schaich, 2014):
1. The main reaction for fast reactors is complete in <10
seconds. Long reaction times to allow for “full reaction”
merely count the phenolic groups, overestimate polyphe-
nols and underestimate small phenols, and are irrelevant
to fast radical reactions in situ.
2. Recording initial fast reaction rates accurately requires use
of fast mixing techniques such as dispensing plate readers
and continuous or stopped flow mixers. Normal hand mixing
can miss much of the early reaction.
3. Rate calculations are not straightforward because the re-
sponse curves lack linear regions, change with antioxidant
concentration and reaction conditions, and do not exhibit
a constant reaction order.
4. DPPH reaction rates are more closely related to steric ac-
cessibility than chemical characteristics of test antioxidants.
5. DPPH reacts by mixed SET and HAT mechanisms, the
proportion of which changes with the antioxidant.
6. DPPH reaction mechanisms and quantitative responses are
altered by many environmental factors.
7. Reaction of antioxidant mixtures with DPPH is less than the
total reactivity of individual components, suggesting that
radical quenching by mixed extracts is actually depressed
in the DPPH assay. The assay thus may greatly underesti-
mate potential radical quenching in cells or foods. Steric
interferences do not account for all antagonisms or syner-
gisms in the assay.
All these factors invalidate the chemical accuracy of this
assay and its ability to predict which antioxidants will be active
in real situations. We thus recommend that the DPPH assay
no longer be used to compare “activity” of antioxidants with
different or unknown structures, especially as occurs in ex-
tracts of natural materials. At the same time, although there
is still much to be elucidated about the fundamental chem-
istry of this reaction, we feel the DPPH assay may be usefully
redirected to gain information about active quenching mecha-
nisms of antioxidants being tested.
Such redirection requires a nearly total revamping of the assay.
Fast mixing methods must be used to capture early reactions.
Reactions must be recorded continuously from the point of mixing
 journal of functional foods 14 (2015) 111–125
to determine initial reaction rate and shape of reaction re-
sponse curves. Kinetics (rate constants) of initial fast response
(mixing to point of diversion from linear, usually a few seconds
to about thirty seconds) must be calculated. When initial fast
absorbance drop is absent, rates should be calculated from the
slope of the first thirty seconds of response curves. Reaction rates
must be calculated over a range of antioxidant concentrations,
then plotted versus concentration to calculate second order rate
constants and determine concentration effects on reactivity. For
comparability between laboratories, standard protocols for
calculating rates and first and second order rate constants must
be adopted. Currently, nearly every study that reports kinetics
uses a different procedure.Antioxidant reactions with DPPH must
be tested in different solvents and pHs over a range of antioxi-
dant or phenol concentrations to determine dominant radical
quenching mechanisms. Extracts, juices, infusions, etc. should
be normalized to identical phenol and ascorbic acid concen-
trations before assays. Finally, DPPH blanks must always be
subtracted from test samples. This step seems to be ignored in
many current research reports.
2.3. ORAC – Oxygen Radical Absorbance Capacity
The ORAC assay originally developed by Glazer and Ghiselli
(Ghiselli, Serafini, Maiani, Azzini, & Ferro-Luzzi, 1995; Glazer, 1990)
was refined and applied to extensive analyses of hundreds
of foods by Prior and colleagues in U.S. Department of Agricul-
ture research laboratories (Cao, Sofic, & Prior, 1996; Huang, Ou,
Hampsch-Woodill, Flanagan, & Prior, 2002; Ou, Hampsch-Woodill,
& Prior, 2001; Prior et al., 2003; Wu et al., 2004) and commer-
cialized by Brunswick Laboratories (Wareham, MA, USA). Perhaps
most widely recognized of all the antioxidant assays, the reac-
tion is simple in concept but complex in practice. Radicals are
generated by heating an azide compound. The azide decom-
poses, eliminating nitrogen gas and leaving behind two carbon-
centered radicals, R• (Rxn 11). In the presence of oxygen, R• are
converted almost instantaneously to reactive peroxyl radicals,
ROO• (Rxn 11) which can either attack target molecules that have
color or fluorescence (Rxn. 12) or react with antioxidants (Rxns.
13, 14).
R − N = N-R ⎯⎯⎯
Heat
→ N2 + 2 R • ⎯⎯
O2
→ ROO•
⎯ (11)
ROO• + target → ROOH + oxidized target (product ) (12)
ROO• + AH → ROOH + A • (13)
ROO• + A • → ROO-A (14)
Competition between reaction of targets and antioxidants
with the ROO• forms the basis of the assay. Fluorescein, the most
common target in current use, is intensely fluorescent in its native
form.When attacked by peroxyl radicals, the fluorescence is lost.
An antioxidant slows the loss of fluorescence by quenching the
peroxyl radicals via hydrogen atom transfer (Rxn 13) or radical
addition (Rxn 14) (Prior, Wu, & Schaich, 2005). The reaction is
followed by recording fluorescence over time (Prior et al., 2005).
Results are calculated as the total area under the reaction curves
(AUC) for each antioxidant sample minus the reaction blank area
119
(no antioxidant) to account for increased induction period (lag
time) and total extent of inhibition, including secondary reac-
tions, induced by antioxidants (Fig. 5A). Converting these unitless
areas into chemically-meaningful values then requires that AUCs
from test samples be compared to AUCs from Trolox standards,
and results are reported as Trolox equivalents. This practice,
however, does not measure rates and it can underestimate fast
reactors.
2.3.1. Advantages of the ORAC assay
The ORAC assay offers several advantages over ABTS•+ and DPPH
assays. It uses peroxyl radicals that are better models of an-
tioxidant reactions with oxidizing lipids and reactive oxygen
species (ROS) in foods and in vivo, and it provides continuous
generation of radicals on a realistic time scale (more like actual
reactions in situ). The completely hydrogen atom transfer radical
quenching mechanism presents a contrast and comparison to
electron transfer in ABTS•+ and DPPH assays. The assay can be
adapted to detect both hydrophilic and hydrophobic antioxi-
dants generally or specifically by altering the radical source,
solvent, and target molecule, and has been routinely automated.
2.3.2. Limitations of the ORAC assay
Nonetheless, the complexity of the ORAC reaction means there
are many points to cause difficulties for researchers using this
assay. The issues outlined below pose significant limitations
and problems for ORAC and can compromise the accuracy and
use of this assay. The purpose in raising these issues is not to
argue that one set of concentrations is correct or optimum, but
to call attention to problems in the assay and the critical need
for more systematic investigations of all ORAC reaction con-
ditions. More detailed studies of the chemistry are essential
to ensure optimization, understand the reactions better for
general use, and guide decisions about standardization of
protocols. More details may be found in Apak et al. (2013).
Very careful control of ORAC reaction temperature is critical!
Consistent generation of radicals in the azide reaction re-
quires accurate and reproducible temperature control to ensure
timely and complete azide decomposition. The temperature
required depends on the azide used, e.g. for AAPH (2,2′-azobis-
2-amidinopropane dihydrochloride), it is 37 °C. When required
temperatures are not reached, reactions are slow and in-
complete, and results are poorly reproducible. Even more
problematic, slow reaction can be misinterpreted as in-
creased radical scavenging, leading to an overestimation of
antioxidant activity.
Never assume that a reaction temperature is achieved just because
it is set on an instrument. Temperatures must be measured. The most
accurate control is obtained by reactions in single cells in heated
holding blocks and fluorescence sample blocks but this elimi-
nates rapid throughput. Obtaining required temperatures with
plate readers is difficult. Plastic plates (e.g. 96 wells) are good
insulators, so temperatures of most ovens must be set at some
higher point to ensure that azide decomposition tempera-
ture in the wells is reached. We tested a number of different
plate readers and found that each unit had to be evaluated in-
dependently to determine the required set temperature. We also
found that plate readers in which the oven and optical reader
were separate gave highly irreproducible results across the
plate and from plate to plate because the temperature
120 journal of functional foods 14 (2015) 111–125

50 µM Trolox (A) (B) (C)

uorescence IIntensity

Net area under curve Oregano

[[AUCAOX AUCno AOX]
Trolox

Blank
Flu

Blank

0 5 10 15 20 25 0
30 10 20 30 40 50 1 0.79 0.50 0.26 0.13
Reaction time (min) Reaction time (min) Fluorescein dilution

(D) (E) Excitation Emission (F)

1:1000 dil. 100 mM Trolox 3.6e7 auc units
oregano extract Fluorescein
1000x dil oregano
1:5000 dil. 4.1e7 auc units
39101087.500
41441937.500 RFU.s
47485687.500 RFU.s

10000x dil oregano Fluorescein

1:10000 dil. 4.7e7 auc units + oregano

Oregano
Blank
00:00 0:10:00
1 0:20:00
2 0:30:00
3
Time
0:40:00
4 0:50:00
7 1:00:00
8 0 20 40 60 450 500 550 600 650
Reaction
Blank time (hr) Reaction time (min) wavelength (nm)

Fig. 5 – Standard ORAC response curve and some aberrations that compromise accuracy of results. (A) Standard ORAC
reaction curves showing area under curve used to quantitate reactions. Fluorescence quenching at high fluorescein
concentrations can be seen in temporary increases in fluorescence immediately after mixing with antioxidants (B) and in
maintenance of fluorescence levels even after dilution (C). Non-radical antioxidant interactions with fluorescein can prevent
fluorescence decay (D). Samples must be diluted until reactions run in reasonable time. Dilution required for each extract
must be determined independently. (E) Effects of non-radical antioxidant interactions with fluorescein can be seen in ORAC
values (area under curve) that increase as antioxidant concentrations decrease. Fluorescein and azide concentrations
different from (D). (E) Overlap of fluorescein and antioxidant fluorescence spectra can dissipate fluorescence in absence of
azide. Fluorescein–antioxidant extract mixture has same reagent concentrations as for individual spectra. Possibility of
multiple effects means that each antioxidant effect must be tested independently with fluorescein alone. Y-axis was
fluorescence intensity in counts per second for A–E; Absorbance or fluorescence units in (F). All graphs: K.M. Schaich,
unpublished data.

varied too much as the plate was moved back and forth between thoroughly with oxygen just before mixing in both manual and
oven and reader. Most accurate, stable, and consistent plate reader testing.
temperatures are obtained in plate readers that are heated from Very careful control of reagent concentrations is critical! Abso-
above and below the plate, not from the sides, and the optical lute and relative concentrations of azide, fluorescein, and samples
system rotates over the plates to take a reading. In this con- control the reaction. Because fluorescein emission intensity is
figuration, the plates never move and the temperature is so intense, only nM or lower concentrations are needed to follow
maintained more accurately. the reaction. However, the reaction is diffusion controlled. At
Very careful control of oxygen is critical! Full and reproducible typical nM concentrations, fluorescein molecules are greatly scat-
oxygenation is required for the azide reaction to efficiently tered in the medium. At the same time, the peroxyl radicals
generate radicals. However, solubility of oxygen declines as tem- generated from the azide are very short-lived and must en-
perature increases, so oxygen becomes depleted when solutions counter a fluorescein molecule before decaying.Thus, very high
are pre-warmed before runs, during thermal equilibration in concentrations of peroxyl radicals are needed to surround the
ovens, and when reaction must be heated over long times due fluorescein and ensure efficient reaction. Prior and colleagues
to strong antioxidant inhibition. Variations in handling and investigated azide levels up to 2000 times greater than fluores-
heating times between samples cause considerable variabil- cein concentrations (Prior et al., 2003) yet adopted 1.6 µmol AAPH
ity in dissolved pO2, hence inconsistent results. With insufficient and 7.5 nmols fluorescein per reaction in their standard pro-
oxygen, reactions are slow, variable, and do not run to comple- cedure (Wu, 2005). We found that azide concentrations in their
tion. In our laboratory, most robust reactions and most higher range provide more consistent reactions and are nec-
reproducible results are obtained when all solutions are sparged essary to keep reaction times for controls at about twenty minutes.
 journal of functional foods 14 (2015) 111–125
Imbalance in reagent proportions – e.g. fluorescein too high
(leading to quenching) and azide too low (leading to insuffi-
cient ROO• for rapid reaction) results in very slow reactions that
do not always go to completion. We find consistent reactions
running to completion in about 15 minutes for controls when
half Prior’s fluorescein concentration (1.9 nM) and 2–10 times
his AAPH concentration (1.6 µM) are used. High fluorescein con-
centrations are a problem because extensive H bonding between
the phenol groups and pi stacking of the rings decrease net
fluorescence, a process called self-quenching (Schauenstein,
Schauenstein, & Wick, 1978). When fluorescence quenching
occurs, emissions do not diminish with dilution – they stay the
same or even increase, as can be seen as a rise at the begin-
ning of some reaction curves when consumption of fluorescein
in reaction releases fluorescence quenching (Fig. 5B) or when
emission intensity does not decrease with dilution (Fig. 5C). Delay
in fluorescence decay leads to loss of resolution between samples
and overestimation of antioxidant action. We find fluores-
cence quenching at the 1.9 nM fluorescein concentrations
recommended in USDA and Brunswick Labs protocols (Prior et al.,
2003; Wu, 2005); using half this fluorescein concentration or less
eliminates initial fluorescence increases and speeds reactions.
Sample concentrations are no less important. When working
with isolated compounds or Trolox standards, the recom-
mended µM solutions can be prepared with certainty. However,
phenol identity and concentrations in extracts from foods and
natural materials are seldom measured before analysis. Most
commonly, a single standard aliquot (e.g. 20 µl) of extract is ana-
lyzed without standardization to phenol or ascorbic acid content
and without testing a range of dilutions. This practice can lead
to distorted if not outright incorrect results, due in part to varying
concentration response patterns with different antioxidants as
well as to non-radical interactions with fluorescein. Phenolic,
quinone, and aromatic rings in the fluorescein structure provide
multiple sites and modes for non-radical associations that block
normal reactions with antioxidants. Polyphenol binding may block
reactive —OH groups and stabilize fluorescein during reac-
tion. Such interactions between antioxidants and fluorescein
mostly increase apparent antioxidant reactivity with ROO•, as
can be seen in exceptionally long reaction times, implausibly
high ORAC values (Fig. 5D), and higher ORAC values with more
dilute antioxidants (Fig. 5E). Alternatively, if antioxidant and
fluorescein excitation or emission manifolds overlap, fluorescein–
antioxidant binding can either stabilize or reduce fluorescence
emissions (Fig. 5F). Interaction interferences must be tested for
in appropriate blanks with extracts plus fluorescein (no azide)
and in overlaps of fluorescein and phenol excitation and emis-
sion manifolds in fluorescence spectra.
Thus, considering all these concentration effects, as with
the ABTS•+ and DPPH assays, it is necessary to assay a range
of antioxidant or extract concentrations to determine linear
response ranges or concentration ranges for valid measure-
ments, and also the cross-over concentration at which anti-
oxidants may become pro-oxidant. To eliminate non-radical
interferences, extracts should be diluted to a concentration that
gives complete reaction in less than one hour. Longer times
increase complicating side reactions and likelihood that the
system will become oxygen limited. Furthermore, the occur-
rence of very long reaction times without substantial fluorescein
decay often signals extract interactions with the fluorescein.
121
Variations in calculation methods as well as reporting and
use of results lead to huge inconsistencies in ORAC values
reported by different laboratories and on internet websites.
This is perhaps the greatest problem with ORAC. Simply
calculating the area under the curve leads to differences and
inaccuracies imposed by timing of data recording and equa-
tions used for integration. One equation proposed for calculating
the AUCs integrates successive slices of the curve deter-
mined by interval sampling time when point values are
recorded (Cao, Alessio, & Cutler, 1993) or a selected time in-
terval when absorbance is recorded continuously:
AUC = ( 0.5 + f 5 f 4 + f 6 f 4 + f 7 f 4 + … + f i f 4) × CT (15)
where fi = fluorescence reading at cycle i (i.e. f4 = initial fluo-
rescence reading at cycle 4), and CT = cycle time in minutes.
It is easy to see from this equation that the time interval
matters. The smaller the time interval in data recording, the
more closely the calculated AUC approaches actual area, or
conversely, the longer the sampling interval, the more the
calculated AUC underestimates actual area. Variations in cal-
culation method contribute to inconsistent quantitative results
between laboratories.
Converting AUC to Trolox equivalents (ORAC units) is a
source of additional vagaries and inconsistencies. Early methods
recommended the following equation for calculating ORAC units
(Ou et al., 2001):
This equation includes a factor (circled in equation) for nor-
malizing to a known Trolox standard concentration (usually
1 mM) that takes the place of a standard curve. Unfortu-
nately, phenol concentrations in extract samples are seldom
determined before analyses, and the equation is often applied
without sample normalization, or with just comparison to 1 mM
Trolox. This practice can lead to huge errors in ORAC values.
Taking problems of ORAC normalization one step further,
the practice of reporting ORAC units as micromolar equiva-
lents of Trolox generates very large numbers which give the
erroneous impression that extracts of foods and other natural
materials have components that are more reactive than to-
copherols by many orders of magnitude. In fact, ORAC values
of 1,000,000 micromolar equivalents are needed to provide the
same protection as 1 M α-tocopherol. Thus, care needs to be
taken so that expressions of ORAC values do not psychologi-
cally or actually overestimate antioxidant action.
Perhaps the greatest problem in calculating ORAC values
is what to do with extracts and edible materials. Calculations
are straightforward for solutions of pure phenols where con-
centrations are known, but phenol concentrations in extracts
are seldom known or measured. What normalization base
should be used in these conditions – mmol Trolox equiva-
lents per ml extract? L of extract? Liter of starting material (e.g.
fruit juice)? g or kg fresh wt? g or kg dried extract? mol phenol?
per 100 g? per serving? unspecified? All of these bases are in
common use, and the variable units and normalization bases
122 journal of functional foods 14 (2015) 111–125

for expressing ORAC values lead to considerable confusion,

Table 1 – ORAC values for foods, selected to illustrate
uncertainty, and obfuscation of results as researchers and mar-
effects of food moisture content and serving size on
keters alike try to enlarge the numbers to amplify psychological numerical values reported.
response to results. So which reporting basis is preferable?
Serving ORAC per
It cannot be overstressed that ORAC units must be speci- sizea serving
fied and reported on a common and reasonable basis for useful (µmol TE/
comparisons. Reporting without normalization to a concen- 100 g)
tration standard and without units, as has occurred often in Spicesb
reports on the internet, is meaningless and often reflects intent Cloves, ground 1 cup 314,446
to deceive or over-exaggerate results. However, even in valid Sumac, bran, raw 312,400
research, determining a basis that is reasonable for practical Cinnamon, ground 1 cup 267,536
understanding and still not misleading is not straightfor- Sorghum, bran, hi-tannin 240,000
Oregano, dried 2 cups 200,129
ward. Compare the ORAC values presented in Table 1, for
Acai berry, freeze-dried 161,400
example, where ORAC values were determined for 100 g of ma-
Turmeric, ground 0.7 cup 159,277
terial. ORAC values varied by several orders of magnitude, and Sorghum, bran, black 100,800
taking the numbers at face value might lead one to think spices Sumac, grain, raw 86,800
had much higher antioxidant power than fruits and veg- Cocoa, dry powder, 0.85 cup 80,933
etables. However, with further consideration several factors that unsweetened
strongly distort interpretation of the ORAC values immedi- Cumin seed 76,800
Maqui berry, 75,000
ately become apparent. First is moisture content when the
concentrated powder
comparison basis is weight. The spices listed are dry, but the
Parsley, dried 4 cups 74,349
foods have moisture contents typically of 90% or greater, which Sorghum, bran, red 71,000
markedly dilute the antioxidants. The second issue is, what Basil, dried 3 cups 67,553
does 100 g mean in terms of how much of a food is normally Foodsc
consumed? One hundred grams of cinnamon, for example, is Aronia black chokeberry 100 g 16,062
about 1 cup, a volume very far from consumable. In contrast, Small red bean 1⁄2 c. dried beans (112 g) 13,727
Wild blueberry 1 cup (145 g) 13,427
100 g strawberries is only half a normal serving, and a medium-
Red kidney bean 1⁄2 c. dried beans 13,259
size apple is about 145 g rather than 100 g. Thus, for the two Pinto bean 1⁄2 cup 11,864
fruits, and indeed most foods in the list, normalizing to 100 g Blueberry 1 cup berries (145 g) 9019
underestimates to variable degrees the antioxidant potential Cranberry 1 cup berries (120 g) 8983
that would actually be consumed. Reporting on a fresh weight Artichoke hearts 1 cup, cooked 7904
basis that ignores moisture contents distorts and underreports Blackberry 1 cup berries (145 g) 7701
Prune 1⁄2 cup (75 g) 7291
antioxidant potential even more. This example demonstrates
Raspberry 1 cup (123 g) 6058
that ORAC values cannot be accepted as just numbers for com-
Strawberry 1 cup (200 g) 5938
parison, but must be expressed on bases that are meaningful Red Delicious apple 1 apple (154 g, medium) 5900
to consumption and include normalization to dry solids. Granny Smith apple 1 apple (154 g, medium) 5381
In addition, to ascertain whether an extract has higher Pecan 1 oz. 5095
phenol content or higher phenol activity, normalization at least Sweet cherry 1 cup (140 g) 4873
to total phenols should be included. Black plum 1 plum (78 g) 4844
Russet potato 1, cooked (148 g) 4649
One further limitation in using and interpreting ORAC values
Black bean 1⁄2 cup dried beans 4181
is understanding what they mean chemically, physiologically,
Plum 1 plum (78 g) 4118
or nutritionally. Chemically, although phenols are assumed to Gala apple 1 apple (154 g, medium) 3903
be the key reactants, correlations of phenol contents to ORAC a
Serving sizes for spices and herbs is the approximate volume for
values have not been established and there is little to no other 100 g, to provide perspective. Weights taken from Aqua-Calc (2014).
documentation of the compounds responsible for extract re- Cups are U.S. volumes.
sponse in ORAC assays. Since some phenols have known toxicity b
ORACValues, 2014.
c
(Bermúdez-Saldaña, Escuder-Gilabert, Medina-Hernández, Wu et al., 2004.
Villanueva-Camañas, & Sagrado, 2007; Hansch, McKarns, Smith,
& Doolittle, 2000; Moridani, Siraki, & O’Brien, 2003; Selassie,
DeSoyza, Rosario, Gao, & Hansch, 1998), especially at higher
concentrations, using such a blind assay to make recommen- but few people would consider that either of these food quan-
dations or decisions about best foods to eat is dangerous. tities contains sufficient antioxidant levels for health. Taking
Physiologically, lack of substantive documentation of radical a different approach, for ORAC expressed as µmoles Trolox
quenching by phenols in vivo has led to removal of the ORAC equivalent, 5000 ORAC units ≅ 5 mmol Vit E = 2153 mg. If it is
Values of Foods list from the USDA website (USDA, assumed that tocopherol reaction with fluorescein is compa-
2010), as noted in the Introduction. rable to that of Trolox, 5000 ORAC units is equivalent to about
Nutritionally, ORAC values are very confusing. At one point, 3230 IU of tocopherol daily. Compared to the 30 IU Toc needed
the USDA recommended that 3000–5000 ORAC units be con- to prevent deficiency and the 400 IU max recommended
sumed per day for health. Five thousand ORAC units can be supplement level, 5000 ORAC units totally absorbed could be
obtained from one Granny Smith apple or one ounce of pecans, a potentially dangerous overdose.
 journal of functional foods 14 (2015) 111–125
2.3.3. Summary and recommendations – ORAC assay
The ORAC assay provides a means of evaluating ability of
an antioxidant to quench radicals by hydrogen atom trans-
fer independently of electron transfer. It also provides a
means of testing reactivity with a number of different
types of radicals, e.g. by changing initiators (Ou et al., 2002).
Nevertheless, the radical reactions employed in the ORAC
assay are complex, even with non-azide initiators, they require
rigid control over reaction conditions to ensure reproducible
and accurate results, and they are subject to numerous
complications and limitations that can compromise use of
this assay.
We recognize that assays such as ORAC provide important
tools for preliminary evaluation and screening of antioxi-
dants. However, ORAC values have been used more as political
and marketing tools than as chemical tools. Despite the exten-
sive use of ORAC assays to analyze a very wide range of materials
and current focus on standardizing this assay as an official
method, many glitches remain in the way the assay is handled.
ORAC reaction chemistry is not well enough defined in either
reaction progress or connection to specific molecular proper-
ties to warrant standardization of any existing protocols. Key
problem areas in chemistry include:
• Fluorescein needs to be replaced with a target that has no
side reactions and does not interact with itself or phenols
by non-radical mechanisms.
• Detailed investigations of the effects of absolute and rela-
tive reagent concentrations on reaction rates and extent are
urgently needed before standardization.
• Non-radical target self-associations and interactions with
phenols in the absence of radicals are insufficiently docu-
mented and may contribute to erroneously high ORAC
values.
• Alternate methods for following reactions that will clearly
determine kinetics as well as stoichiometry need to be
developed.
• To provide chemical logic relating ORAC reactivity to phenol
structures, structural features that control reactivity of
antioxidants with fluorescein or other target molecules must
be identified and ORAC values need to be connected di-
rectly to levels and types of antioxidants in extracts rather
than analyzing extracts blindly.
Critical problems exist also with use and applications of
ORAC values. Although correlations between ORAC values
of consumed foods and elevated plasma antioxidant capac-
ity or beneficial anti-disease actions have been claimed, the
appearance of components of tested materials in the blood-
stream have not been demonstrated. Phenols are poorly
absorbed and rapidly conjugated and excreted so actual intake
from consumed foods cannot be predicted from ORAC values,
and toxicity of high levels of phenols in general and some
specific phenols is well documented (Bermúdez-Saldaña et al.,
2007; Hansch et al., 2000; Moridani et al., 2003; Selassie et al.,
1998; Yen, Duh, & Tsai, 2002). Thus, high ORAC values should
not be used to rank foods, sell foods, or determine food choices.
Much more research is needed to establish clear causal con-
nections between molecules detected in ORAC assays and
physiological actions.
123
3. Summary and recommendations
Twenty five years of antioxidant screening in natural materi-
als have shown that polyphenols are good antioxidants; that
most plant materials have high contents of phenols which
exhibit excellent antioxidant activity; that within classes of
antioxidants, the number and position of phenolic —OH groups
affects antioxidant efficacy; and that non-phenol antioxidant
structures are poorly detected in most existing antioxidant
assays. Twenty five years of antioxidant screening have NOT
resolved issues of assay chemistry, standardization, and re-
porting; provided significant insight into chemical mechanisms
and factors controlling antioxidant action; clearly connected
in vitro assay chemistry to in vivo actions; established rate con-
stants for reaction of antioxidants with radicals that are relevant
in foods and biological tissues; shown how antioxidants with
negligible absorption exert systemic effects; or addressed
counter issues of (poly)phenol toxicity.
The early need to merely identify which natural materials
had strong antioxidant efficacy has been replaced by the need
to elucidate accurately how natural antioxidants act. The
reaction patterns discussed in this paper demonstrate the
chemical inaccuracies of three common in vitro antioxidant
assays and argue for redirecting their orientation and empha-
sis. ABTS•+, DPPH, and ORAC assays conducted with more
detailed and revised protocols can still be useful for reveal-
ing dominant quenching mechanisms. However, for quantitative
antioxidant analysis of natural materials, these assays should
be replaced with new antioxidant assays, including lipid oxi-
dation assays, with clearly identifiable reaction mechanisms,
fully tested reaction conditions, and kinetic as well as stoi-
chiometric analyses.
Finally, it is essential that all assays used to investigate
antioxidant efficacy emphasize fundamental chemistry rather
than easy or rapid analysis in order to elucidate radical quench-
ing mechanisms for different classes of phenols, determine rate
constants and specificity for reactions of antioxidants with dif-
ferent radicals (will require rapid-mix techniques), identify
synergisms and antagonisms among classes of antioxidant
compounds, and then learn how to exploit this information
for more effective application of natural antioxidants in medical
therapies and food stabilization.
REFERENCES
Alcolea, J. F., Cano, A., Acosta, M., & Arnao, M. B. (2003).
Determination of the hydrophilic and lipophilic antioxidant
activity of white and red wines during the wine-making
process.Italian Journal of Food Science, 15, 207–214.
Apak, R., Gorinstein, S., Böhm, V., Schaich, K. M., Özyürek, M., &
Güçlü, K. (2013). Methods of measurement and evaluation of
natural antioxidant capacity/activity (IUPAC Technical
Report). Pure and Applied Chemistry, 85, 957–998.
Aqua-Calc. (2014). Conversions and calculations.
<http://www.aqua-calc.com/page/density-table>. Accessed
14.03.01.
Arnao, M. B., Cano, A., & Acosta, M. (2001). The hydrophilic and
lipophilic contribution to total antioxidant activity. Food
Chemistry, 73, 239–244.
124 journal of functional foods 14 (2015) 111–125
Awika, J. M., Rooney, L. W., Wu, X., Prior, R. L., &
Cisneros-Zevallos, L. (2003). Screening methods to measure
antioxidant activity of sorghum (Sorghum bicolour) and
sorghum products. Journal of Agricultural and Food Chemistry,
51, 6657–6662.
Barclay, L. R. C., Edwards, C. E., & Vinqvist, M. R. (1999). Media
effects on antioxidant activities of phenols and catechols.
Journal of the American Chemical Society, 121, 6226–6231.
Bermúdez-Saldaña, J. M., Escuder-Gilabert, L.,
Medina-Hernández, M. J., Villanueva-Camañas, R. M., &
Sagrado, S. (2007). Chromatographic retention–activity
relationships for prediction of the toxicity pH-dependence of
phenols. Chemosphere, 69, 108–117.
Bondet, V., Brand-Williams, W., & Berset, C. (1997). Kinetics and
mechanisms of antioxidant activity using the DPPH free
radical method. LWT - Food Science and Technology, 30,
609–615.
Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995).
Use of a free radical method to evaluate antioxidant activity.
LWT – Food Science and Technology, 28, 25–30.
Burda, S., & Oleszek, W. (2001). Antioxidant and antiradical
activities of flavonoids. Journal of Agricultural and Food
Chemistry, 49, 2774–2779.
Butkovic, V., Klasinc, L., & Bors, W. (2004). Kinetic study of
flavonoid reactions with stable radicals. Journal of Agricultural
and Food Chemistry, 52, 2816–2820.
Cao, G., Alessio, H. M., & Cutler, R. G. (1993). Oxygen-radical
absorbance capacity assay for antioxidants. Free Radical
Biology and Medicine, 14, 303–311.
Cao, G., Sofic, E., & Prior, R. L. (1996). Antioxidant capacity of tea
and common vegetables. Journal of Agricultural and Food
Chemistry, 44, 3426–3431.
Foti, M. C., Daquino, C., & Geraci, C. (2004). Electron-transfer
reaction of cinnamic acids and their methyl esters with the
DPPH radical in alcoholic solutions. The Journal of Organic
Chemistry, 69, 2309–2314.
Foti, M. C., Daquino, C., Mackie, I. D., DiLabio, G. A., & Ingold, K. U.
(2008). Reaction of phenols with the 2,2-diphenyl-1-
picrylhydrazyl radical. Kinetics and DFT calculations applied
to determine ArO-H bond dissociation enthalpies and
reaction mechanism. The Journal of Organic Chemistry, 73,
9270–9282.
Frankel, E. N. (2007). Antioxidants in food and biology: Facts and
fiction. Bridgwater, England: The Oily Press.
Frankel, E. N., & Finley, J. (2008). How to standardize the
multiplicity of methods to evaluate natural antioxidants.
Journal of Agricultural and Food Chemistry, 56, 4901–4908.
Ganapathi, M. R., Hermann, R., Naumov, S., & Brede, O. (2000).
Free electron transfer from several phenols to radical cations
of non-polar solvents. Physical Chemistry Chemical Physics, 2,
4947–4955.
Ghiselli, A., Serafini, M., Maiani, G., Azzini, E., & Ferro-Luzzi, A.
(1995). A fluorescence-based method for measuring total
plasma antioxidant capability. Free Radical Biology and
Medicine, 18, 29–36.
Glazer, A. N. (1990). Phycoerythrin fluorescence-based assay for
reactive oxygen species. Methods in Enzymology, 186, 161–168.
Halliwell, B., Rafter, J., & Jenner, A. (2005). Antioxidant or not.
Health promotion by flavonoids, tocopherols, tocotrienols,
and other phenols: Direct or indirect effects? The American
Journal of Clinical Nutrition, 81, 268S–276S.
Hansch, C., McKarns, S. C., Smith, C. J., & Doolittle, D. J. (2000).
Comparative QSAR evidence for a free-radical mechanism of
phenol-induced toxicity. Chemico-Biological Interactions, 127,
61–72.
Hogg, J. S., Lohmann, D. H., & Russell, K. S. (1961). The kinetics of
reaction of 2,2-diphenyl-1-picrylhydrazyl with phenols.
Canadian Journal of Chemistry, 39, 1588–1594.
Huang, D., Ou, B., Hampsch-Woodill, M., Flanagan, J. A., &
Deemer, E. K. (2002). Development and validation of oxygen
radical absorbance capacity assay for lipophilic antioxidants
using randomly methylated b-cyclodextrin as the solubility
enhancer. Journal of Agricultural and Food Chemistry, 50,
1815–1821.
Huang, D., Ou, B., Hampsch-Woodill, M., Flanagan, J. A., & Prior, R.
L. (2002). High-throughput assay of oxygen radical absorbance
capacity (ORAC) using a multichannel liquid handling system
coupled with a microplate fluorescence reader in 96-well
format. Journal of Agricultural and Food Chemistry, 50,
4437–4444.
Jimenez-Alvarez, D., Giuffrida, F., Vanrobaeys, F., Golay, P. A.,
Cotting, C., Lardeau, A., & Keely, B. J. (2008). High-throughput
methods to assess lipophilic and hydrophilic antioxidant
capacity of food extracts in vitro. Journal of Agricultural and
Food Chemistry, 56, 3470–3477.
Lang, J. (2009). Tocopherol stability in controlled release packaging
films. M.S. thesis. New Brunswick, NJ: Rutgers University.
Lebeau, L., Furman, C., Bernier, J.-L., Duriez, P., Teissier, E., &
Cotelle, N. (2000). Antioxidant properties of
di-tert-butylhydroxylated flavonoids. Free Radical Biology and
Medicine, 29, 900–912.
Miller, N. J., & Rice-Evans, C. A. (1997). Factors influencing the
antioxidant activity determined by the ABTS+ radical cation
assay. Free Radical Research, 26, 195–199.
Miller, N. J., Rice-Evans, C. A., Davies, M. J., Gopinathan, V., &
Milner, A. (1993). A novel method for measuring antioxidant
capacity and its application to monitoring the antioxidant
status in premature neonates. Clinical Science, 84, 407–412.
Moridani, M. Y., Siraki, A., & O’Brien, P. J. (2003). Quantitative
structure toxicity relationships for phenols in isolated rat
hepatocytes. Chemico-Biological Interactions, 145, 213–223.
ORACValues. (2014). Antioxidant values of foods sorted by ORAC
value. <http://www.oracvalues.com/sort/orac-value>.
Accessed 14.02.04.
Ou, B., Hampsch-Woodill, M., Flanagan, J. A., Deemer, E. K., Prior,
R. L., & Huang, D. (2002). Novel fluorometric assay for
hydroxyl radical prevention capacity using fluorescein as the
probe. Journal of Agricultural and Food Chemistry, 50, 2772–2777.
Ou, B., Hampsch-Woodill, M., & Prior, R. L. (2001). Development
and validation of an improved oxygen radical absorbance
capacity assay using fluorescein as the fluorescent probe.
Journal of Agricultural and Food Chemistry, 49, 4619–4626.
Prior, R. L., Hoang, A., Gu, L., Wu, X., Bacchiocca, M., Howard, L.,
Hampsch-Woodill, M., Huang, D., Ou, B., & Jacob, R. (2003).
Assays for hydrophilic and lipophilic antioxidant capacity
(oxygen radical absorbance capacity (ORACFL) of plasma and
other biological and food samples. Journal of Agricultural and
Food Chemistry, 51, 3273–3279.
Prior, R. L., Wu, X., & Schaich, K. M. (2005). Standard methods for
the determination of antioxidant capacity and phenolics in
foods and dietary supplements. Journal of Agricultural and Food
Chemistry, 53, 4290–4302.
Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., &
Rice-Evans, C. A. (1999). Antioxidant activity applying an
improved ABTS radical cation decolourization assay. Free
Radical Biology and Medicine, 26, 1231–1237.
Salamone, M., DiLabio, G. A., & Bietti, M. (2011). Hydrogen atom
abstraction reactions from tertiary amines by benzyloxyl and
cumyloxyl radicals: Influence of structure on the rate-
determining formation of a hydrogen-bonded prereaction
complex. The Journal of Organic Chemistry, 76, 6264–6270.
Salamone, M., Martella, R., & Bietti, M. (2012). Hydrogen
abstraction from cyclic amines by the cumyloxyl and
benzyloxyl radicals. The role of stereoelectronic effects
and of substrate/radical hydrogen bonding. The Journal of
Organic Chemistry, 77, 8556–8561.
 journal of functional foods 14 (2015) 111–125
Schauenstein, K., Schauenstein, E., & Wick, G. (1978).
Fluorescence properties of free and protein bound fluorescein
dyes. I. Macrospectrofluorometric measurements. The Journal
of Histochemistry and Cytochemistry, 26, 277–283.
Selassie, C. D., DeSoyza, T. V., Rosario, M., Gao, H., & Hansch, C.
(1998). Phenol toxicity in leukemia cells: A radical process?
Chemico-Biological Interactions, 113, 175–190.
Stratil, P., Klejdus, B., & Kuban, V. (2006). Determination of total
content of phenolic compounds and their antioxidant
activity in vegetables – Evaluation of spectrophotometric
methods. Journal of Agricultural and Food Chemistry, 54,
607–616.
Strube, M., Haenen, G. R. M. M., van den Berg, H., & Bast, A. (1997).
Pitfalls in a method for assessment of total antioxidant
capacity. Free Radical Research, 26, 515–521.
Tian, X., & Schaich, K. M. (2013). Effects of molecular structure on
kinetics and dynamics of the Trolox equivalent antioxidant
capacity assay with ABTS. Journal of Agricultural and Food
Chemistry, 61, 5511–5519.
USDA. (2010). Oxygen Radical Absorbance Capacity (ORAC) of
selected foods, Release 2 (2010).
<http://www.ars.usda.gov/Services/docs.htm?docid=15866>
Accessed 14.02.04.
van den Berg, R., Haenen, G. R. M. M., van den Berg, H., & Bast, A.
(1999). Applicability of an improved Trolox equivalent
125
antioxidant capacity (TEAC) assay for evaluation of
antioxidant capacity measurements of mixtures. Food
Chemistry, 66, 511–517.
Williams, D. E. (1966). Structure of 2,2-diphenyl-1-picrylhydrazyl
free radical. Journal of the American Chemical Society, 88,
5665–5666.
Wojdyło, A., Figiel, A., & Oszmiański, J. (2009). Effect of drying
methods with the application of vacuum microwaves on the
bioactive compounds, colour, and antioxidant activity of
strawberry fruits. Journal of Agricultural and Food Chemistry, 57,
1337–1343.
Wu, X. (2005). Standard operating procedure of oxygen radical
absorbance capacity (ORACFL). Provided by R. Prior, Univ.
Arkansas.
Wu, X., Beecher, G. R., Holden, J. M., Haytowitz, D. B., Gebhardt, S.
E., & Prior, R. L. (2004). Lipophilic and hydrophilic antioxidant
capacities of common foods in the United States. Journal of
Agricultural and Food Chemistry, 52, 4026–4037.
Xie, J., & Schaich, K. M. (2014). Re-evaluation of the 2,2-diphenyl-
1-picrylhydrazyl free radical (DPPH) assay for antioxidant
activity. Journal of Agricultural and Food Chemistry, 62,
4251–4260.
Yen, G.-C., Duh, P.-D., & Tsai, H.-L. (2002). Antioxidant and
pro-oxidant properties of ascorbic acid and gallic acid.
Food Chemistry, 79, 307–313.