June 2005 63

Original

Determination and Confirmation of Five Phenolic Antioxidants

in Foods by LC/MS and GC/MS
(Received October 20, 2004)

Sumiko THJ?>῎1, 2, ῏, Maki N6@6CD῎1, 3, Hisaya T:G696῎4,

Yukio T6BJG6῎4 and Yasuhide TDCD<6>῎1, 5
(῎1Division of Food Chemistry, Osaka Branch, National Institute of Health Sciences: Closed;
῎2Present address, National Institute of Health Sciences: 1῍18῍1, Kamiyoga, Setagaya-ku,
Tokyo 158῍8501, Japan; ῎3Present Address, Saito Lab., AnGes MG, Inc.: 4F,
Saito Bio-incubator, 7῍7῍15, Saito-Asagi, Ibaraki-shi, Osaka, 567῍0085 Japan;
῎4 Nagoya City Public Health Research Institute: 1῍11, Hagiyama-cho, Mizuho-ku,
Nagoya 467῍8615, Japan; ῎5Present Address, Test Inspection Center, Osaka Pharmaceutical:
1῍3῍8, Izumi-cho, Chuo-ku, Osaka 540῍0019, Japan; ῏ Corresponding author)

Identification and determination of butylated hydroxyanisole (BHA), butylated hydroxytolu-

ene (BHT), nordihydroguaiaretic acid (NDGA), propyl gallate (PG) and tert-butylhydroquinone
(TBHQ) by means of LC/MS and GC/MS were examined. These five phenolic antioxidants were
detected as their pseudo-molecular ions [MῌH]ῌ by LC/MS using a Shim-pack FC-ODS column
with drying gas. Moreover, BHA, BHT and TBHQ were detected based on their mass fragment
ions by GC/MS. Decomposition of TBHQ, NDGA and PG during analysis could be prevented by
the addition of A-ascorbic acid (AsA) to the extraction solvent. All five antioxidants were
extracted from nikuman, olive oils, peanut butter, pasta sauce and chewing gum with a mixture of
acetonitrile῍2-propanol῍ethanol (2 : 1 : 1) containing 0.1῍ AsA (AsA mixture), which had been
cooled in a freezer and filtered. One part filtrate and 5 parts water were mixed and placed on a
Mega-Bond Elutῌ C18 cartridge, except in the case of chewing gum. Lipids in foods were removed
on a C18 cartridge by washing with 5 mL of 5῍ acetic acid, and antioxidants were eluted with 5
mL of AsA mixture. The antioxidants spiked into nikuman, olive oil, peanut butter, pasta sauce
and chewing gum were successfully identified and their concentrations determined by LC/MS,
and GC/MS with good recoveries.

Key words: antioxidant; LC/MS; butylated hydroxyanisole (BHA); butylated hydroxytolu-

ene (BHT); nordihydroguaiaretic acid (NDGA); propyl gallate (PG); tert-butylhydroquinone
(TBHQ); GC/MS

foods contain unpermitted phenolic antioxidants. Japa-

Introduction
In Japan, four kinds of phenolic antioxidants, butyl-
ated hydroxyanisole [BHA; a mixture of 2- and 3-(2-
methylpropyl)-4-methoxyphenol, CAS No. 25013-16-5],
butylated hydroxytoluene [BHT; 2,6-bis(1,1-dimethy-
lethyl)-4-methylphenol, CAS No. 128-37-0], nordihy-
droguaiaretic acid [NDGA; 4,4῏-(2,3-dimethyltetra-
methylene) dipyrocatechol, CAS No. 500-38-9], and pro-
pyl gallate (PG; propyl 3,4,5-trihydroxybenzoate, CAS
No. 121-79-9), are permitted for use in foods1)῍3). On the
other hand, tert-butylhydroquinone (TBHQ; CAS No.
1948-33-0), which the Joint FAO/WHO Expert Commit-
tee on Food Additives (JECFA)3) and Code of Federal
Regulations (CFR)4) permit for use in foods, is used in
foods in a limited number of countries around the
world5), though not in Japan. Recent importation of
various kinds of foods from various countries into
Japan6), ῎6 has raised the possibility that some of these
nese inspectors have sometimes found TBHQ in such
foods, and these foods have been returned or
discarded6), ῎6; in general, TBHQ is the most frequently
detected of the banned phenolic antioxidants in im-
ported foods in Japan6), ῎6.
Therefore, improved methods of detecting and con-
firming trace amounts of TBHQ in food are needed. In
addition, it is necessary to be able to simultaneously
analyze TBHQ with other antioxidants such as BHT
and BHA, since in many cases they are used together.
Various analytical methods are available to identify
and determine antioxidant additives in foods, including
colorimetry7), 8), GC7), 9)῍11), HPLC7), 8), 11)῍16), GC/MS17), 18),
and HPLC with fluorescence (Flu)19), 20). Recently, HPLC
has been used after appropriate sample preparation,
such as liquid῍liquid partition or frozen separation
῎6 http://www.mhlw.go.jp/topics/yunyu/tp0130-1.html
 64
following extraction from foods7), 12)῍15).
However, for microanalyses in particular, it is neces-
sary to use several di#erent methods to determine var-
ious antioxidants that are added to foods. Therefore, in
the present study, we examined the possible use of LC/
MS and GC/MS for the simultaneous determination of
five kinds of phenolic antioxidants (BHA, BHT, NDGA,
PG and TBHQ) in foods that may contain potentially
interfering compounds. We also developed a method
for cleaning up the sample solution in order to deter-
mine the antioxidants at trace levels.
Materials and Methods
Samples and reagents
Nikuman (a steamed bun with meat filling), olive oils,
pasta sauce, peanut butter and chewing gum were pur-
chased from markets in Osaka and Aichi prefectures.
BHA, BHT, NDGA (made by ICN Pharmaceuticals, Costa
Mesa, CA, USA), PG and TBHQ were purchased from
Wako Pure Chemical Industries, Ltd. (Osaka, Japan)
and used as standards. Water was obtained from a
Milli-Q SP TOC Reagent Water System (Millipore Co.,
Bedford, MA, USA). Anhydrous sodium sulfate of rea-
gent grade was dried at 110ῐ for 16 hr before use. All
other chemicals were of reagent or HPLC grade. A
mixed solvent was prepared using a mixture of ac-
etonitrile῍2-propanol῍ ethanol (2 : 1 : 1, v/v). An AsA
mixture was also prepared daily by adding 1 g of A-
ascorbic acid (AsA) to 1,000 mL of the mixed solvent.
Mega-Bond Elutῌ C18 cartridges (1 g/6 mL, produced
by Varian Inc., Harbor City, CA, USA) were used. One
series of C18 cartridges was pre-tested using a standard
solution of antioxidants. The cartridge used for deter-
mination was conditioned with 10 mL of methanol, 5
mL of 5ῑ acetic acid (AcOH) and 10 mL of water before
use.
Instruments
HPLC was performed with an LC-10 HPLC system
equipped with SPD-M10Avp photo diode array (PDA)
and RF-10AXL fluorescence (Flu) detectors (Shimadzu
Co., Kyoto, Japan). LC/MS was performed with an
LCMS-2010A mass spectrometer coupled with an
LC-VP HPLC series system and an LCMS solution (Shi-
madzu). GC/MS was performed with a GCMS-QP5050
mass spectrometer coupled with a GC-17A GC system,
and a CLASS-5000 work station (Shimadzu).
Preparation of standard solutions
Standard stock solutions of BHA, BHT, NDGA, PG
and TBHQ were produced by weighing out 0.1 g of each
compound and dissolving it in methanol to make a 100
mL solution. All solutions were used within one month.
The standard stock solutions and AsA mixture were
employed to prepare standard working solutions that
were used within one week. Corresponding standard
solutions for the calibration curve were prepared using
the standard working solutions and AsA mixture, and
were used immediately after preparation.
J. Food Hyg. Soc. Japan Vol. 46, No. 3
Preparation of sample solutions
Samples other than chewing gum
Samples were prepared following a modification of
the method described in “Analytical method for food
additives in foods, 2nd ed.” (O$cial method)13), 14). First,
50 mL of AsA mixture was added to the sample (5.0 g)
and dried sodium sulfate (10 g in the case of samples
other than pasta sauce (15 g) and olive oils (none)), and
the solution was homogenized for 3 min. The homo-
genized solution was left to stand for more than 1 hr
in a freezer (῍20ῌ῍5ῐ), then filtered through No. 5C
filter paper (Advantic Toyo, Tokyo, Japan) in a freezer,
or through a glass filter rapidly at room temperature.
Fifteen mL of cooled AsA mixture was added to the
residue, the solution was filtered, and the filtrate was
taken as a primary test solution. A secondary test
solution (test solution) was made by concentrating the
primary test solution to 2 mL in a vacuum with a
rotary evaporator at 40ῐ, and adding AsA mixture to
make 5.0 mL. The test solution was cleaned up on a
C18 cartridge. Two mL of the test solution and 10 mL
of water were mixed, loaded onto the cartridge, and
washed with 5 mL of 5ῑ AcOH. The cartridge was
then eluted with 5 mL of AsA mixture to give 5.0 mL
of sample solution (sample: 400 mg/mL). Each sample
solution was analyzed by GC/MS and filtered through
a 0.5 mm filter before HPLC or LC/MS.
Chewing gum
A primary test solution was prepared as described
above. A sample solution (sample: 400 mg/mL) for GC/
MS was made by drying an aliquot of the primary test
solution in a vacuum with a rotary evaporator at 40ῐ
and dissolving the residue in ethyl acetate to give a 5.0
mL solution. A sample solution (sample: 400 mg/mL)
for LC/MS was made by drying another aliquot of the
test solution in a vacuum with a rotary evaporator at
40ῐ and dissolving the residue in methanol to give a
5.0 mL solution. The sample solution was filtered
through a 0.5 mm filter before HPLC or LC/MS.
HPLC
HPLC was performed under the following conditions:
column, Develosil ODS-UG-5 (4.6 mm i.d.῎250 mm,
Nomura Chemical Co., Aichi, Japan); injection volume,
10 mL; column temp., 40ῐ; PDA, 240῍340 nm (280 nm);
Flu, Ex 280 nm, Em 325 nm; flow rate, 1.0 mL/min;
mobile phase, a mixture of acetonitrile῍methanol (1 : 1)
(A) and 5ῑ AcOH (B), A/B῏4 : 6 (0 min)ΐ9 : 1 (15῍25
min)ΐ4 : 6 (26῍36 min).
LC/MS
LC/MS was performed under the following condi-
tions. LC conditions: column, a Shim-pack FC-ODS (2.0
mm i.d.῎75 mm, Shimadzu); injection volume, 5 mL;
column temp., 40ῐ; PDA, 240῍340 nm; flow rate, 0.2
mL/min; mobile phase, acetonitrile (A) and 0.05ῑ
AcOH (B); A/B῏1 : 9 (0 min)ΐ95 : 5 (15῍20 min)ΐ1 : 9
(20.01῍30 min).
 June 2005 Determination of Five Antioxidants in Foods by LC/MS and GC/MS
Fig. 1. E#ect of AsA on the stability of antioxidants
PG: ῌ; TBHQ: ΐ; NDGA: ῔; BHA: ῍; BHT: ῑ.
(A) Without AsA (B) With AsA (100 mg/mL)
MS conditions: nebulizer gas flow, 1.5 L/min; drying
gas flow, 0.20 MPa; corn voltage, ῌ3.5 kV; curved desol-
vation line (CDL) temp., 250῏; block heater, 200῏; CDL
voltage, ῌ25.0 V; Q-array voltage, auto; ionization
mode, ESI (negative); SCAN mode, m/z ῎50ῌ350 (1.0
sec/scan), or SIM mode (1.0 sec/5 ch); monitor ions, m/z
῎211 (PG), 165 (TBHQ), 301 (NDGA), 179 (BHA) and
219 (BHT).
GC/MS
The GC/MS experiments were performed using a
modified version of the method reported by Ohta et al.18)
GC conditions: column, DB-5MS (0.25 mm film depth,
0.25 mm i.d.῍30 m, J&W Scientific Co., Folam, CA,
USA); injection volume, 1 mL (splitless); column temp.,
60῏ (0ῌ2 min)῕15῏/min῕300῏ (7 min); injection and
interface temp., 280῏; helium carrier gas, 100 kPa; flow
rate, 20 mL/min.
MS conditions: ionization method, EI; ionization volt-
age, 70 eV; detector, 1.3 kV; fragment ions (m/z; num-
bers in parentheses were used for confirmation), BHA,
165 (137, 180); BHT, 205 (220, 145); TBHQ, 123 (151,
166).
Determination
Antioxidant concentrations were determined by mea-
surement of the peak areas and interpolation from cal-
ibration curves. In determining PG and NDGA using
LC/MS, two calibration curves covering di#erent con-
centration ranges were used in each case.
Results and Discussion
Stability of phenolic antioxidants
Phenolic antioxidants, especially TBHQ, PG, and
NGDA, are readily decomposed by oxygen in air. It is
therefore necessary that their concentrations be deter-
mined immediately after preparation of the sample
solution12). If a lot of numbers of food sample are
analyzed, a lot of time is required to prepare sample
solutions and to conduct determination by HPLC, LC/
MS or GC/MS, so that antioxidants in the samples
are easily auto-oxidizaed, especially if they are present
65
only in very low concentrations. Shimidzu et al. add-
ed A-ascorbyl palmitate to the extract solvent to in-
hibit oxidation῍7, but its retention time (Rt) on HPLC
is unsuitable. We used AsA, which did not interfere
with the HPLC determination of phenolic antioxidants.
The stability of the antioxidants in the presence of
AsA was studied. Standard solutions (10 mg/mL) of
antioxidants with and without AsA were left to stand at
room temperature, and then examined by HPLC.
As shown in Fig. 1, BHA and BHT in standard solu-
tions without AsA were both stable, while the other
three antioxidants each decomposed gradually. Ap-
proximately 40ῐ of TBHQ was lost within 24 h. In
standard solutions with AsA (100 mg/mL), however, all
antioxidants were stable. Therefore, we added a rela-
tively large amount of AsA to the extraction solvent
(corresponding to 0.1ῐ AsA solution).
LC/MS of phenolic antioxidants
The LC conditions with PDA were set with reference
to the HPLC conditions. When 0.2ῐ AcOH was used as
solvent B, the peaks of all five antioxidants appeared in
PDA chromatograms using various semi-micro columns
for LC/MS. When 0.02ῐ AcOH was used as solvent B,
however, all five peaks appeared when only a Shim-
pack FC-ODS column was used. The results of our
preliminary experiments indicated that the Rt of PG
should be set at more than 5 min in order to avoid
interference from components of the gum sample
matrix.
In LC/MS without drying gas, the total ion chromat-
ograms (TICs) showed only peaks of PG and NDGA. In
LC/MS with drying gas, on the other hand, all five
antioxidants were detected. The drying gas creates a
finer spray mist that dries the liquid droplets and pro-
motes ionization of the samples, improving the sensitiv-
ity of LC/MS.
῍7 Shimidzu, T., Mizuno, T., Kato, F., Ishibashi, T., Sakai, S.,
Ishibashi, H.: The 76th Meeting of the Food Hygienics
Society of Japan, Abstract, p. 90 (5ῌ6, November, Niigata,
1998).
 66 J. Food Hyg. Soc. Japan Vol. 46, No. 3

Fig. 2. LC/MS chromatograms (TIC) with PDA and mass spectra of the five antioxidants
The six peaks showed the following pseudo-molecular ions (m/z῍[MῌH]ῌ).
1: AsA, 175; 2: PG, 211; 3: TBHQ, 165; 4: NDGA, 301; 5: BHA, 179; 6: BHT, 219
LC conditions: column, a Shim-pack FC-ODS; injection volume, 5 mL; column temp., 40῎; flow rate, 0.2 mL/min;
mobile phase, acetonitrile (A) and 0.05῏ AcOH (B); A/B῍1 : 9 (0 min)ῑ95 : 5 (15ῌ20 min)ῑ1 : 9 (20.01ῌ30 min)
MS conditions: nebulizer gas flow, 1.5 L/min; drying gas flow, 0.20 MPa; corn voltage, ῌ3.5 kV; CDL temp., 250῎;
block heater, 200῎; CDL voltage, ῌ25.0 V, Q-array voltage, auto; ionization mode, ESI (negative); SCAN mode,
m/z῍50ῌ350

As shown in Fig. 2, the mass chromatograms showed GC/MS of phenolic antioxidants

base peaks with m/z῍175 (AsA), 211 (PG), 165 (TBHQ),
301 (NDGA), 179 (BHA), and 219 (BHT), which corre-
spond to the pseudo-molecular ions [MῌH]ῌ, and these
were used for selected ion monitoring (SIM). The PDA
chromatogram at 280 nm showed sharper peaks with
smaller tails than TICs.
The calibration curves showed good linearity. How-
ever, to determine PG and NDGA with good sensitivity
it was necessary to make separate calibration curves in
two ranges (PG: 0.01ῌ0.5 mg/mL and 0.5ῌ5 mg/mL;
NDGA: 0.01ῌ0.5 mg/mL and 0.5ῌ5 mg/mL; TBHQ and
BHA: 0.1ῌ5 mg/mL; BHT: 0.5ῌ5 mg/mL).
GC/MS was performed following the method report-
ed by Ohta et al.18) However, the injection and interface
temperatures were set at 280῎ in order to avoid ab-
sorption of the antioxidants on the injection port and
interface. Using a Shimadzu GCMS-QP5050, BHA, BHT
and TBHQ were detected, and fragmentation patterns
of these three antioxidants are shown in Fig. 3. The
results agree with the data reported by Ohta et al.18)
Nevertheless, no peaks of PG and NDGA were observed,
presumably because of decomposition within the GC
column.
The calibration curves for standard solutions of these
three antioxidants (m/z for determination: BHA, 165;
 June 2005 Determination of Five Antioxidants in Foods by LC/MS and GC/MS 67

Fig. 3. GC/MS chromatograms (TIC) and mass spectra of three antioxidants

The three peaks showed the following fragment ions (m/z; numbers in parentheses were used for confirmation)
3: TBHQ, 123 (151, 166); 5: BHA, 165 (137, 180); 6: BHT, 205 (220, 145).
GC conditions: column, DB-5BH; injection volume, 1 mL (splitless); column temp., 60ῌ (0῍2 min)῏15ῌ/min῏300ῌ (7
min); injection and interface temp., 280ῌ; helium carrier gas, 100 kPa; flow rate, 20 mL/min
MS conditions: ionization method, EI; ionization voltage, 70 eV; detector, 1.3 kV

BHT, 205; TBHQ, 123) showed good linearity from 0.01
῍20 mg/mL. If only one injection of washing solvent
was done after analysis of high-concentration standard
solution, the carry-over amounted to 0.1῍. Therefore,
we carried out several washing injections on GC/MS
after each analysis, especially for sample solutions con-
taining high concentrations of antioxidants.
Cleanup with solid-phase extraction
Food extract solutions with mixed solvent or ace-
tonitrile contain lipids20). BHT is the least polar com-
pound of the five antioxidants and is readily soluble in
oils; its recovery from oily samples was low. Oishi et
al.20) developed a liquid῍liquid extraction method,
which we applied here, though the sample solution
from olive oils contained only about 7῍ lipids. We also
examined a Mega-Bond Elutῌ C18 cartridge for sample
cleanup. When the standard solution (containing 10
mg/mL each of the five antioxidants) mixed with 5 mL
of water was loaded onto the cartridge, and the car-
tridge was washed with 5 mL of water, BHT was not
fully eluted with methanol, though it was with the AsA
mixture. However, as shown in Fig. 4(A), the recovery
of BHT was very low, as was that of TBHQ. A possible
explanation for this is that TBHQ might be decomposed
in the cartridge, while BHT may bind strongly to the
cartridge. The antioxidants eluted easily from a C18
HPLC column under strongly acidic conditions. There-
fore, the standard solution mixed with 5 mL of 5῍
AcOH was loaded onto a C18 cartridge, and the car-
tridge was washed with 5 mL of 5῍ AcOH, but again
the results were unsatisfactory, as shown in Fig. 4(B).
On the other hand, when the standard solution mixed
with 10 mL of water was loaded onto a C18 cartridge,
and the cartridge was washed with 5 mL of 5῍ AcOH,
then eluted with the AsA mixture, the recoveries were
 68
Fig. 4. E#ects of passing the loading and washing
solvents through the Mega-Bond Elutῌ C18 car-
tridge
PG: ῐ, TBHQ: ῑ, NDGA: ῒ, BHA: ῌ, BHT: ῎
Loading solvent and 2 mL of standard solution (10
mg/mL each) were mixed and loaded on the C18
cartridge. The cartridge was washed with
washing solvent, and the antioxidants were then
eluted with an AsA mixture through the C18
cartridge.
(A) Loading and washing solvents: 5 mL of water
each
(B) Loading and washing solvents: 5 mL of 5῍
AcOH each
(C) Loading solvent: 10 mL of water; washing
solvent: 5 mL of 5῍ AcOH.
satisfactory, as shown in Fig. 4(C). Consequently, the
AsA mixture was selected as a suitable extract solvent.
The antioxidants in purified test solutions were deter-
mined by HPLC, LC/MS, and GC/MS. Lipids in the
sample solutions of olive oils were 98.7῍ removed
through the use of the C18 cartridge.
Comparison of recovery tests of the phenolic antioxidants
in nikuman by various methods
The five kinds of antioxidants were simultaneously
spiked into nikuman at concentrations of 1 and 10 mg/
g, and the sample solution was prepared and examined
by HPLC, LC/MS, and GC/MS. The results are shown
in Table 1.
SIM chromatograms obtained by LC/MS and GC/MS
showed symmetrical peaks. Figure 5 shows the LC/MS
chromatograms (SIM) of nikuman spiked with the five
antioxidants.
In LC/MS, the five antioxidants were successfully
identified, though BHT at the 1 mg/g level could not be
detected. Other compounds in the sample may have
J. Food Hyg. Soc. Japan Vol. 46, No. 3
Fig. 5. LC/MS chromatograms (SIM) of the five anti-
oxidants spiked into nikuman
2: PG, 211; 3: TBHQ, 165; 4: NDGA, 301; 5: BHA,
179; 6: BHT, 219
(A) Standard solution of five antioxidants (5 mg/
mL each)
(B) Nikuman only
(C) Nikuman spiked with five antioxidants (10 mg/
g each)
The sesitivity of LC/MS was changed at 15.0 min.
a#ected the ionization of BHT. In GC/MS, only three of
the five antioxidants were detected.
HPLC with UV detection was e#ective at higher con-
centrations of all five antioxidant (Table 1), though
some interfering peaks of food components appeared
around the Rt of PG. HPLC with Flu detection was not
e#ective for PG and BHT.
However, for microanalysis, it is desirable that BHA,
BHT and TBHQ be identified and determined using GC/
MS, and PG and NDGA using LC/MS. Since the Rts of
antioxidants in LC/MS sometimes vary, confirmation
by co-injection of samples and standards is desirable.
Application of LC/MS and GC/MS to various foods
Olive oils, pasta sauce, peanut butter and chewing
gum were used, and the five kinds of antioxidants were
spiked at concentrations of 1 and 10 mg/g. Sample
solutions were prepared and examined by LC/MS and
GC/MS. In the case of chewing gum, cleanup with a
C18 cartridge could not be conducted because of the
gum bases, and the primary test solution was evapora-
ted to dryness as described by Ohta et al.18) The results
of LC/MS are shown in Table 2 and those of GC/MS
in Table 3.
In LC/MS, the five antioxidants could be identified,
though BHT at the 10 mg/g level was only barely ob-
served. However, food components in chewing gums
were incorrectly identified as PG or NDGA, and food
components in olive oils and peanut butter as TBHQ or
NDGA. It was also important to wash the column
thoroughly to avoid carry-over of antioxidants to the
subsequent analysis.
June 2005 Determination of Five Antioxidants in Foods by LC/MS and GC/MS 69

Table 1. Comparison of Recoveries of Five Antioxidants from Nikuman Determined by Various Methodsa)

HPLC
LC/MS GC/MS
Added UV detection Flu detection
mg/g b) b) b)
Found Recovery Found Recovery Found Recovery Foundb) Recovery
mg/g ῎ mg/g ῎ mg/g ῎ mg/g ῎
PG 0 ND ῌ ῏῏῏c) NDd) ῌ ND ῌ
1 0.80ῌ0.07 79.7 ῏῏῏c) 1.00ῌ0.10 100.7 ND 0.0
10 7.65ῌ0.49 76.5 ῏῏῏c) 8.70ῌ0.56 87.0 ND 0.0
TBHQ 0 ND ῌ ND ῌ ND ῌ ND ῌ
1 0.54ῌ0.03 54.3 0.68ῌ0.03 67.7 0.85ῌ0.05 84.9 0.89ῌ0.02 89.2
10 6.18ῌ0.28 61.8 7.62ῌ0.28 76.1 8.91ῌ0.36 89.1 8.89ῌ0.22 88.9
NDGA 0 ND ῌ ῏῏῏c) NDd) ῌ ND ῌ
1 0.62ῌ0.02 61.1 ῏῏῏c) 0.95ῌ0.07 94.6 0.92ῌ0.02 92.2
10 6.64ῌ0.51 66.4 ῏῏῏c) 8.75ῌ0.39 87.5 8.75ῌ0.37 87.5
BHA 0 ND ῌ ND ῌ ND ῌ ND ῌ
1 0.49ῌ0.03 49.3 0.67ῌ0.09 67.1 0.92ῌ0.02 91.6 0.91ῌ0.03 90.8
10 6.08ῌ0.33 60.8 7.32ῌ0.55 73.2 9.00ῌ0.33 90.0 8.96ῌ0.14 89.6
BHT 0 ND ῌ 0.10ῌ0.08 ῌ ND ῌ ND ῌ
1 ND 0.0 0.77ῌ0.07 66.9 0.90ῌ0.08 90.0 ND 0.0
10 1.44ῌ0.04 14.4 7.06ῌ0.83 69.6 8.87ῌ0.20 88.7 10.54ῌ0.53 105.4
a)
Sample was extracted with AsA mixture. The extracted solution was cleaned up through a C18 cartridge.
Each antioxidant in the prepared sample solution was determined by LC/MS, GC/MS and HPLC with PDA (280 nm; UV)
and Flu detection.
b)
Values are the average and standard deviation of three trials.
c)
PG and NDGA could not be determined because no peak was identified in GC.
d)
There were peaks of food components around the retention times of PG and NDGA.
ND in LC/MS: PG῍0.01 mg/g, TBHQ῍0.05 mg/g, NDGA῍0.01 mg/g, BHA῍0.03 mg/g, BHT῍0.5 mg/g
ND in GC/MS: TBHQ῍0.01 mg/g, BHA῍0.01 mg/g, BHT῍0.01 mg/g
ND in HPLC (UV): PG῍0.5 mg/g, TBHQ῍05 mg/g, NDGA῍0.5 mg/g, BHA῍0.5 mg/g, BHT῍0.5 mg/g
ND in HPLC (Flu): PG῍12.5 mg/g, TBHQ῍0.01 mg/g, NDGA῍0.01 mg/g, BHA῍0.01 mg/g, BHT῍2.5 mg/g

Table 2. Recoveries of Five Antioxidants from Various Foods Using LC/MS

Olive oila) Pasta saucea) Peanut buttera) Chewing gumb)

Added
mg/g Foundc) Recovery Foundc) Recovery Foundc) Recovery Foundc) Recovery
mg/g ῎ mg/g ῎ mg/g ῎ mg/g ῎
PG 0 ND ῌ ND ῌ ND ῌ 0.01ῌ0.01 ῌ
1 1.15ῌ0.12 115.2 0.91ῌ0.15 91.2 0.90ῌ0.09 89.9 1.15ῌ0.12 114.1
10 9.19ῌ0.90 91.9 7.28ῌ0.05 72.8 8.24ῌ0.20 82.4 9.19ῌ0.90 91.7
TBHQ 0 0.29d)ῌ0.03 ῌ ND ῌ 0.01ῌ0.00 ῌ ND ῌ
1 0.67ῌ0.01 37.9 0.51ῌ0.13 50.9 0.66ῌ0.01 65.1 0.43ῌ0.04 42.6
10 3.07ῌ0.28 27.8 4.08ῌ0.16 40.8 6.33ῌ0.11 63.2 5.28ῌ0.38 52.8
NDGA 0 ND ῌ ND ῌ ND ῌ ND ῌ
1 1.36ῌ0.08 135.5 0.67ῌ0.14 66.6 0.70ῌ0.01 70.0 0.78ῌ0.11 77.8
10 10.34ῌ0.43 103.4 6.40ῌ0.19 64.0 8.00ῌ0.95 80.0 7.38ῌ0.78 73.8
BHA 0 ND ῌ ND ῌ ND ῌ ND ῌ
1 0.64ῌ0.06 64.3 0.58ῌ0.13 58.0 0.62ῌ0.02 61.8 0.25ῌ0.03 25.1
10 7.25ῌ0.15 72.5 4.79ῌ0.14 47.9 6.97ῌ0.30 69.7 3.78ῌ0.34 37.8
BHT 0 ND ῌ ND ῌ ND ῌ ND ῌ
1 ND 0.0 ND 0.0 ND 0.0 ND 0.0
10 1.06ῌ0.16 10.6 0.55ῌ0.10 5.5 1.31ῌ0.40 13.1 ND 0.0
a)
Sample was extracted with AsA mixture. The extracted solution was cleaned up through a C18 cartridge.
Each antioxidant in the prepared sample solution was determined by LC/MS.
b)
Sample was extracted with AsA mixture. The extracted solution was evaporated, and the residue was dissolved in
methanol.
Each antioxidant in the prepared sample solution was determined by LC/MS.
c)
Values are the average and standard deviation of three or five trials.
d)
There were peaks of food components around retention time of TBHQ.
ND: PG&NDGA῍0.01 mg/g, TBHQ῍0.05 mg/g, BHA῍0.03 mg/g, BHT῍0.5 mg/g
 70 J. Food Hyg. Soc. Japan Vol. 46, No. 3

Table 3. Recoveries of Three Antioxidants from Various Foods Using GC/MS

Olive oila) Pasta saucea) Peanut buttera) Chewing gumb)

Added
mg/g Foundc) Recovery Foundc) Recovery Foundc) Recovery Foundc) Recovery
mg/g ῏ mg/g ῏ mg/g ῏ mg/g ῏

BHA 0 ND ῌ ND ῌ ND ῌ ND ῌ
1 0.71῍0.04 71.2 0.74῍0.22 74.4 0.49῍0.09 48.1 0.90῍0.01 89.7
10 7.55῍0.53 75.5 7.11῍1.88 71.1 5.79῍0.51 57.9 8.66῍0.52 86.6

BHT 0 0.09῍0.08 ῌ 0.14῍0.06 ῌ 0.18῍0.03 ῌ 0.23῍0.27 ῌ

1 0.76῍0.05 66.3 0.90῍0.25 75.6 0.71῍0.05 52.9 1.14῍0.02 91.6
10 8.69῍0.43 68.6 5.85῍2.56 57.1 4.44῍0.37 42.7 8.62῍0.27 83.9

TBHQ 0 ND ῌ ND ῌ 0.02῍0.03 ῌ ND ῌ
1 0.67῍0.01 66.5 0.57῍0.01 57.0 0.67῍0.13 65.1 0.61῍0.06 61.2
10 7.59῍0.80 75.9 7.03῍1.41 70.3 6.03῍0.71 60.1 8.29῍0.29 83.0
a)
Sample was extracted with AsA-mixture. The extracted solution was cleaned up through a C18 cartridge.
Each antioxidant in the prepared sample solution was determined by GC/MS.
b)
Sample was extracted with AsA-mixture. The extracted solution was evaporated, and the residue was dissolved in ethyl
acetate.
Each antioxidant in the prepared sample solution was determined by GC/MS.
c)
Values are the average and standard deviation of three trials.
ND: BHA῎0.01 mg/g, BHT῎0.01 mg/g, TBHQ῎0.01 mg/g

In GC/MS, BHA, BHT and TBHQ were sensitively

identified and their concentrations were determined.
However, carry-over was observed after injection of
sample solutions with high concentrations of antioxi-
dants. In this case, washing by solvent injection was
ine#ective. However, the combination of washing and
aging of the column overcame the problem.
When determining the antioxidants by means of LC/
MS or GC/MS, prior clean-up was found to be indis-
pensable. The detection limits of PG and NDGA in
LC/MS and those of BHA, BHT, TBHQ in GC/MS were
approximately 0.01 mg/g.
Conclusion
Identification and determination of the concentra-
tions of BHA, BHT, NDGA, PG and TBHQ in foods by
LC/MS and GC/MS were examined. A method of clean-
ing up the sample solution was also developed using a C
18 cartridge.
Five antioxidants were identified and their concen-
trations determined by LC/MS based on the [MῌH]ῌ
peaks. BHA, BHT and TBHQ were identified and deter-
mined using their fragment ions by GC/MS, but the
sensitivities were found to di#er.
BHA, BHT, NDGA, PG and TBHQ were spiked into
nikuman, and the solutions were cleaned up with a C18
cartridge, then the antioxidants identified and deter-
mined by LC/MS, GC/MS and HPLC (UV and Flu).
From the viewpoint of sensitivity, BHA, BHT and
TBHQ should be determined by GC/MS, and NDGA and
PG by LC/MS. This method was applied to olive oils,
pasta sauce and peanut butter. Chewing gum, however,
was analyzed without cleanup.
Acknowledgements
The authors would like to express their gratitude to
Mr. Kimihiko Yoshii, Osaka Branch of NIHS, for his
helpful suggestions on the analyses by LC/MS and GC/
MS, and to Mr. Takahiro Goda and Mr. Hirohisa
Mikami, the Analytical Applications Department
(Hadano City, Japan) of Shimadzu Co., for their assis-
tance in LC/MS determination. This study was sup-
ported in part by the Costs for Food Examinations of
the Ministry of Health, Labour and Welfare in fiscal
years 2001 and 2002.
A portion of this study was presented at the 84th
meeting of the Food Hygienics Society of Japan (Osaka,
2002).
References
1) Ministry of Health and Welfare, “Japan’s Specifications
and Standards for Food Additives, 7th Ed.”, English
Version, reproduced by Japan Food Additives Associa-
tion, Japan, 2000, p. M125῍M126, M125, M271῍M272,
U339, U349, U352 (2000).
2) Suzuki, I. et al. eds., “Shokuhintenkabutsu Koteisyo Kai-
setsusho”, Hirokawa Shoten, 1999, p. F-12, F-21, F-25, F-
27 (ISBN4-567-01852-4).
3) The Joint FAO/WHO Expert Committee on Food Addi-
tives, “Compendium of food additive specification
volume 2”, Food and Agriculture Organization of the
United Nations (FAO), Rome, 1992, p. 245, 249, 253,
1311, 1485.
4) Code of federal regulations, “Food and Drug, Title 21”,
National Archives and Records Administration, 2004,
Parts 166, 172, 173, 175῍177, 181, 182, 184.
5) Japan Food Additives Association ed., “Sekai No Shoku-
hintennkabutsu GaisetsuῌJECFA To Shuyokoku No
Ninkahinmoku Risutoῌ”, Japan Food Additives Associ-
ation, Tokyo, 2004.
6) O$ce of Port Health Adoministration, Policy Planning
Division, Department of Food Safety, Pharmaceutical
 June 2005 Determination of Five Antioxidants in Foods by LC/MS and GC/MS
and Food Safety Bureau, Ministry of Health Labour and
Welfare, Statistis of Imported Food in Japan 2000, Sho-
kuhin Eisei Kenkyu (Food Sanitation Research), 51(10),
113῍167 (2001).
7) Tanimura, A. et al. eds., “Shokuhin Chu No Shokuhinten-
kabutsu Bunsekiho Kaisetsusho”, Kodansha, Japan,
1992, p. 94῍101, 111῍120, 887῍895 (ISBN 4-06-153705-9).
8) Association of O$cial Analytical Chemists O$cial
Methods of Analysis (Helrich, K. ed.), 965.28, 15th ed.,
Vol. 2, AOAC, Arlington, 1990, p. 1137῍1139.
9) Yu, L. Z., Inoko, M., Matzuno, T., A gas chromatographic
method for rapid determination of additives in vegeta-
ble oils. J. Agric. Food Chem., 32, 681῍683 (1984).
10) Choong, Y.-M., Lin, H.-J., Gas chromatographic determi-
nation of synthetic antioxidants in edible fats and oilsῌ
a simple methylation method. J. Food Drug Analysis, 9,
20῍25 (2001).
11) Mizuno, T., Tajima, K., Okino, N., Yamamoto, K., Deri-
vatization of butylated hydroxytoluene for gas chro-
matography with electron-capture detection and high
performance liquid chromatography with ultraviolet
detection. J. Chromatogr., 333, 171῍174 (1985).
12) Yamada, M., Miyata, M., Kato, Y., Nakamura, M., Nishi-
jima, M., Shibata, T., Ito, Y., Determination of nine phe-
nolic antioxidants in foods by high performance liquid
chromatography. Shokuhin Eiseigaku Zasshi (J. Food
Hyg. Soc. Japan), 34, 535῍541 (1993).
13) Food Chemistry Division, Environmental Health
Bureau, Ministry of Health and Welfare, ed., “Shokuhin
Chu No Shokuhintenkabutsu Bunsekiho (Analytical
method for food additives in foods), 2nd ed.”, 2000, p. 49῍
51, 56῍59, 351῍353, 371῍372.
14) Department of Food Safety, Pharmaceutical and Food
71
Safety Bureau, Ministry of Health, Labour and Welfare,
ed., “Standard Methods of Analysis in Food Safety Reg-
ulationῌFood Additives, 2003ῌ”, Japan Food Hygienic
Association, Tokyo, 2003, p. 65῍70, 79῍81, 82῍84, 564῍
567, 595῍597.
15) Andrikopoulons, N. K., Brueschweiler, H., Felber, H., Tae-
schler, C., HPLC anaysis of phenolic antioxidants, toco-
phenols and triglycerides. J. Am. Oil Chem. Soc., 68, 359῍
364 (1991).
16) Page, P. D., Liquid chromatographic method for the de-
termination of nine phenolic antioxidants in butter oil:
collaborative study. J AOAC Int., 76, 765῍79 (1993).
17) Hatanaka, H., Akiyama, Y., Unpermitted antioxidant
(TBHQ) detected from import foods. Report of the Hyogo
Prefectural Institute of Public Health, 32, 184῍187
(1997).
18) Ohta, M., Narita, M., Miyoshi, T., Itoyama, T., Kimura, M.,
Kobayashi, M., Ochi, R., Sekiguchi, Y., Koiguchi, S., Hira-
hara, Y., Hasegawa, M., Miyata, M., Kazumasa, K., Maeda,
K., Otomasu, M., Ishiwata, H., Analysis of BHA, BHT and
TBHQ in chewing gum by GC and GC/MS. Shokuhin
Eiseigaku Zasshi (J. Food Hyg. Soc. Japan), 38, 78῍84
(1997).
19) Yankah, V. V., Ushio, H., Ohshima, T., Koizumi, C., Quan-
titative determination of butylated hydroxyanisole, bu-
tylated hydroxytoluene, and tert-butyl hydroquinone in
oils, foods, and biological fluids by high-performance
liquid chromatography with fluorometric detection.
Lipids, 33, 1,139῍1,145 (1998).
20) Oishi, M., Matsuda, T., Nojiri, S., Saito, K., Simultaneous
determination of five antioxidants in foods by HPLC
with fluorescence detection. Shokuhin Eiseigaku Zasshi
(J. Food Hyg. Soc. Japan), 43, 104῍109 (2002).