
for the screening of antioxidant residues in food prod-
ucts. Usually, sample preparation for RP-HPLC in-
volves dilution or suspension with hexane, extraction
with a polar solvent (most frequently acetonitrile
mixed with some water), preconcentration by evap-
oration under reduced pressure, reconstitution (e.g.,
with isopropanol), and, if necessary, clean-up with
unmixable solvents or solid phases. Extraction with
supercritical CO
2
has been used for capillary electro-
separations. RP-HPLC separations are normally per-
formed on C18 columns using gradient elution with
aqueous phases containing methanol or acetonitrile
and acetic acid, followed by spectrophotometric
detection at 280–285 nm. Pentachlorophenol and
2,4,6-trimethylphenol have been employed as in-
ternal standards in the analysis of fats and oils and
bakery products.
0002 A modification of the AOAC official method
applied to butter oil has yielded recoveries higher
than 95% for alkyl gallates, t-butylhydroxyquinone
(TBHQ), butylated hydroxyanisole (BHA), nordihy-
droguaiaretic acid (NDGA), 2,4,5-trihydroxybutyro-
phenone, and 2,6-di-t-butyl-4-hydroxymethylphenol,
and lower for butylated hydroxytoluene (BHT)
(79%). Other reported data indicate recoveries higher
than 90% for BHA and BHT in vegetable oil and
propyl gallate (PG) in lard, and for dodecyl gallate
(DG), BHA, BHT, TBHQ, and NDGA in liver pa
ˆ
te
´
,
but the less hydrophobic PG and octyl gallate (OG)
gave recoveries around 80% in this sample. Using
methanol, recoveries between 85 and 95% have
been reported for PG, OG, DG, 2- and 3-BHA,
NDGA and TBHQ, but 3,5-di-t-butylcresol gave
72%. The intralab relative standard deviations
(RSDs) range typically between 3 and 7%, but values
lower than 2% have also been obtained. Reported
interlab RSDs range from 7% for DG to 26% for
TBHQ. Low or rather disperse recoveries can be due
to oxidation during sample storage and preparation,
which easily occurs with TBHQ.
0003 In one procedure for dry food, the sample was
homogenized with hexane, water, and acetonitrile
(25:5:75) and the combined acetonitrile extracts con-
centrated by evaporation, injected, and detected at
280 nm. Separation was performed on a C18 column
with gradient elution using water–acetonitrile mobile
phases containing 5% acetic acid. The detection
limits (DLs) for PG, BHA, BHT, t-butylquinol,
2,4,5-trihydroxybutyrophenone, 3,5-di-t-butyl-4-
hydroxymethylphenol, and NDGA ranged from 0.4
to 2.7 mgml
1
. For a typical 20-ml injection loop,
these data roughly match with other DLs reported
as absolute values: 25 (OG, DG), 35 (PG) and
100 ng (NDGA). In the determination of nine phen-
olic antioxidants in corn oil, butter oil, butter, niboshi
(dried sardines), and frozen shrimps, the DLs were
1.0 mg per gram of sample (85–102% recoveries).
0004Lower DLs can be achieved with fluorometric,
chemoluminometric, and amperometric detection.
For BHA, BHT, and TBHQ, excitation at 280 nm
yields maximum emission at c. 310 nm. Procedures
based on the chemiluminescence of luminol are
mainly used to measure antioxidant activity, and
these are discussed in detail later in the corresponding
section. t-Butylquinol, 2,6-di-t-butyl-p-cresol, and
2- and 3-t-butyl-4-methoxyphenol have been deter-
mined in oils by HPLC with amperometric detection
on a glassy carbon electrode. With a previous modifi-
cation of the carbon electrode surface by electropoly-
merization of nickel phthalocyanine, DLs ranging
from 0.1 to 0.6 mgml
1
have been achieved. When
used in a flow-injection nonchromatographic system,
the modified electrode provides a DL of 2.7 ng ml
1
with a 1.8% RSD (at 0.5 mgml
1
) for BHA in bis-
cuits. Using 0.3- and 0.7-mm microbore HPLC with
electrochemical detection, DLs as low as 0.1 and
0.6 fmol l
1
have been obtained for BHA, BHT,
and PG.
HPLC with Ion-pairing and Micellar Mobile Phases
0005The selectivity of the HPLC separations can be modi-
fied by ion-pairing agents, such as alkyl sulfates and
carboxylates, and alkyl ammonium salts. These re-
agents form micelles when at least one of the alkyl
chains has 12 or more carbon atoms; however, mi-
celles are disrupted in the presence of a large organic
solvent concentration, thus leading to an ion-pairing
rather than a micellar separation mechanism. PG,
OG, DG, BHA, and BHT can be determined in spiked
olive oil with a mobile phase containing sodium
dodecyl sulfate (SDS) and 30% propanol in 18 min.
0006An aqueous micellar mobile phase, which may
also contain a lower percentage of an organic solvent
(usually 5–15% of a short-chain alcohol), is used in
micellar liquid chromatography (MLC). In MLC,
physiological fluids and tissue extracts can be injected
without deproteinization, because proteins are solved
in the micellar phase and elute with the solvent front.
Moreover, hydrophobic samples can be simply di-
luted with a compatible solvent and injected. Thus,
several oils and fats diluted with n-pentanol (5–30%
solutions) can be injected on a C18 column and eluted
with an SDS/n-propanol mobile phase; PG, OG,
TBHQ, and BHA have been quantified in c. 15 min
with absolute DLs ranging from 0.05 to 0.3 ng.
Avoiding a previous two-phase partitioning step sim-
plifies the procedure and may help in obtaining high
recoveries and in reducing random and systematic
errors, particularly when easily oxidizable solutes
such as TBHQ are quantified.
276 ANTIOXIDANTS/Synthetic Antioxidants, Characterization and Analysis