Photolysis of α-Tocopherol in Olive Oils and Model Systems

Article abstract of DOI:10.1021/jf980239n

The photolysis of α-tocopherol (I) in olive oil (O) and in some model systems (n-hexane = H; anhydrous n-hexane = HA, and triolein = T) was studied under sunlight and under artificial light (λ > 290 nm) by HPLC and GC/MS. In O and T, I disappeared linearly to 50% of the starting concentration, reached a constant value, and finally disappeared rapidly from the medium. In the model system, photolysis followed a pseudo-first-order kinetics. Although no peaks attributable to photoproducts were found in O, a main product identified by ¹H and ¹³C NMR and GC/MS as 5-formyltocopherol (II) was found in the model systems. Irradiation of compound II led to species undetectable by HPLC in agreement with a slower consecutive kinetic process than that of I. In the HA and T systems, the formation of II occurred at lower levels than in H. The possible behavior of photodegradation is discussed.

Full text of DOI:10.1021/jf980239n

J. Agric. Food Chem. 1998, 46, 4529−4533
4529
P h otolysis of r-Tocop h er ol in Olive Oils a n d Mod el System s
Filippo M. Pirisi,*,† Alberto Angioni,† Giovanni Bandino,‡ Paolo Cabras,† Claude Guillou,§
Elisabetta Maccioni,† and Fabiano Reniero§
Dipartimento di Tossicologia dell’Universita` di Cagliari, viale Diaz 182, 09126 Cagliari, Italy,
Commission of the European Communities, J .R.C. Environment Institute, 21020 Ispra, Italy, and
Consorzio Interprovinciale per la Frutticoltura, via Mameli 106, 09123 Cagliari, Italy
The photolysis of R-tocopherol (I) in olive oil (O) and in some model systems (n-hexane ) H;
anhydrous n-hexane ) HA, and triolein ) T) was studied under sunlight and under artificial light
(λ > 290 nm) by HPLC and GC/MS. In O and T, I disappeared linearly to 50% of the starting
concentration, reached a constant value, and finally disappeared rapidly from the medium. In the
model system, photolysis followed a pseudo-first-order kinetics. Although no peaks attributable to
photoproducts were found in O, a main product identified by 1H and 13C NMR and GC/MS as
5-formyltocopherol (II) was found in the model systems. Irradiation of compound II led to species
undetectable by HPLC in agreement with a slower consecutive kinetic process than that of I. In
the HA and T systems, the formation of II occurred at lower levels than in H. The possible behavior
of photodegradation is discussed.
Keyw or d s: R-Tocopherol; photolysis; olive oil
INTRODUCTION
Vitamin E is a well-known mixture of phenolic
compounds with high antioxidative properties. Among
these phenolic compounds, R-tocopherol (I, Figure 1) is
the main component (Burton and Ingold, 1986).
Many foodstuffs (cereals, late leafs vegetables, oil-
seeds, and fruits, and, of course, their oils) contain I in
significant amount for the human diet. The original
amount of I in foodstuffs could be reduced significantly
by a reforming process, whereas cooking and freezing
have only a small influence (Encyclopedia of Food
Science and Technology, 1992).
The reduction of I mainly by reaction with hydro-
peroxides generated from polyunsaturated fatty acids
(PUFA) has been well studied in model systems (Yamau-
chi et al., 1995). Its reactivity and that of some of its
model compounds toward chemical oxidants have also
been reported (Suarna et al., 1988). The influence of
other physical factors on I, such as γ-rays, was studied
by several authors (Knapp and Tappel, 1961; J ore and
Ferradini, 1985; Molnar and Koswig, 1992). Surpris-
ingly, data on the photolysis of I under domestic
conditions (room temperature, sunlight) are not reported
in the literature, although some foodstuffs (e.g., oils)
could easily be injured by light. In past years the only
studies on this topic were a dye-sensitized photooxida-
tion (Grams et al., 1972, which reported that I was
“insensitive to incandescent light. . .at 25 °C for 16 h”)
and a photolysis at -30 °C (Clough et al., 1979).
Recently, the photolysis of I in liposomes by UVB rays
(ultraviolet beta, 290-320 nm) was reported by Kramer
and Liebler (1997), and Servili et al. (1996) have studied
F igu r e 1. Formulas of I, its rearrangement compounds, and
II.
under sunlight the interaction of I with phenolic com-
pounds of antioxidant activity. In this paper we report
data on the photolysis of I under sunlight and under
artificial light (λ > 290 nm) in virgin olive oils and in
model systems of n-hexane (H), anhydrous n-hexane
(HA), and triolein (T).
MATERIALS AND METHODS
Ch em icals. n-Hexane, chloroform, dichloromethane, metha-
nol, and acetonitrile were of HPLC solvent grade (Merck,
Darmstadt, Germany). R- and γ-tocopherol, triolein (Aldrich
or Sigma, Milan, Italy), K2CrO4, Na2CO3 (Carlo Erba, Milan),
and anhydrous Na2SO4 (Merck) were analytical grade reagents
(>98%).
* Author to whom correspondence should be addressed (fax
++ 39 070 30 07 40; e-mail pcabras@unica.it).
† Universita` di Cagliari.
‡ Consorzio Interprovinciale per la Frutticoltura.
§ Commission of the European Communities.
10.1021/jf980239n CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/09/1998

4530 J. Agric. Food Chem., Vol. 46, No. 11, 1998
Pirisi et al.
γ-Tocopherol was used as an internal standard (i.s.) at 200
ppm in CHCl3 or in hexane. The n-hexane was anhydrified
with metallic Na.
Two olive oils were employed. They were obtained from cv.
Bosana olives by mill hammer crushing, beating, and cold
extraction with continuous oil centrifuge. The sample em-
ployed in sunlight experiments shows the following param-
eters: free acidity, 0.33% (as oleic acid); peroxides, 6.66 mequiv
of O2/kg; K270 ) 0.180; K232 ) 2.100; R-tocopherol, 188 ppm.
The sample used in artificial light had the following param-
eters: free acidity, 0.28% (as oleic acid); peroxides, 6.5 mequiv
of O2/kg; K270 ) 0.173; K232 ) 1.630; R-tocopherol, 281 ppm.
The latter sample was also used in other lamp experiments
after saturation with water.
Ir r a d ia tion of I a n d Its P h otop r od u cts. In all experi-
ments nonirradiated samples were held in the dark as control.
The photolysis performed experiments were as follows: olive
oil under sunlight (OS) and artificial light (OL); n-hexane under
sunlight (HS) and artificial light (HL); triolein (TL) and anhy-
drous n-hexane under artificial light (HAL). This latter system
was employed to determine the water content effect on the
process. HPLC and/or GC/MS were performed, as reported
above, and irradiation behavior was checked. With the latter
technique the samples were injected directly without any
further preparation.
OS Conditions. Eighteen 100 mL screw-capped bottles, 9
dark (DBS, dark bottles sunlight) and 9 white (WBS), were
filled with a virgin olive oil, hermetically sealed, and exposed
to sunlight at latitude 39° 12′ north and longitude 9° 07′ east
from Greenwich, between February 1 and November 12, 1996.
Another 18 glass bottles, 9 dark (DBI, dark bottle) and 9 white
(WBI), were exposed to indirect sunlight in our laboratory. On
May 25, September 2, and November 11, 1996, three replicates
from each series were collected and analyzed by HPLC.
OL Conditions. The oil was placed into a 500-mL cylindrical
flask containing a mercury lamp. At selected times three
replicates (each of 200 µL of oil) were withdrawn, dissolved in
1.0 mL of chloroform containing the internal standard, and
analyzed.
HS Conditions. In these experiments the compounds I and
its derivative (II) were dissolved in n-hexane in 2 mL screw-
capped borosilicate vials at the concentration of ≈200 ppm and
exposed to direct sunlight. At selected times three samples
were withdrawn and analyzed. Each experiment was repli-
cated four times.
F igu r e 2. Histogram of the concentration of I in the oil
exposed to sunlight during long-time experiment.
from 9 to 20 min. In the OS and OL experiments the injected
samples were made by dissolving 200 µL of oil in 1.0 mL of
CHCl3 containing the i.s. In HS, HL, and HAL experiments
the samples were prepared as follows: 200 µL of the hexane
solution was evaporated under N2, dissolved with an equal
amount of eluent mixture containing the i.s.
Calculation of the compound concentration in the chromato-
grams was made by the i.s. method.
GC/ MS. An HP-5890 GC (Hewlett-Packard) equipped with
an HP-5971 MS detector (Hewlett-Packard) was used. The
column was a Durabond fused silica (30 m × 0.25 mm i.d.,
J &W Scientific, Folsom, CA) with DB-5 liquid phase (5%
phenyl, 95% dimethylpolysiloxane; film thickness, 0.25 µm).
The sample (2 µL) was injected in the split mode (1:50).
Detector and injector operating conditions were, respectively,
280 and 250 °C; the oven temperature was programmed as
follows: 80 °C (1 min) raised to 250 °C (10 °C/min), and held
for 25 min. Helium was the carrier gas at 1.8 mL/min (45
kPa). Mass spectrometer operating conditions were as fol-
lows: electron ionization, 70 eV; ion source 180 °C; scan mass,
range 50-550; scan interval, 1.5 s; solvent delay, 6 min.
IR. Infrared spectrophotometric analysis was performed on
a Perkin-Elmer 1310 infrared spectrophotometer, using a scan
time of 12 min in liquid phase with chloroform as a solvent.
NMR. Proton and 13C NMR spectra were recorded at 500
and 125 MHz, respectively, in CDCl3 with a Bruker spectrom-
eter. The temperature was fixed at 300 K. All values are
quoted in δ (ppm) with respect to the internal standard
tetramethylsilane (TMS).
Sep a r a tion of P h otop r od u cts. The separation of photo-
products was performed by column chromatography (glass
column, 400 × 25 mm i.d.). The column was filled with silica
gel (particle size ) 0.05-0.20 mm) activated in a stove (12 h
at 110 °C). The mobile phase was a mixture of hexane and
dichloromethane (60:40, v/v). A 1.2 g aliquot of I was irradi-
ated to sunlight in hexane solution; the crude photodegrada-
tion mixture was evaporated to dryness under reduced pres-
sure and chromatographed. Fractions of 10 mL were collected
and analyzed by TLC (silica gel plates 60-F254, 0.2 µm; Merck)
using the same eluents as in chromatography column. The
fractions containing the spots with the same Rf were collected
in six samples (0.410 g of II; 0.260 g of unreacted I; and four
fractions of a total weight of 0.200 g) and analyzed by HPLC
and/or GC/MS. The total final weight was 0.870 g, corre-
sponding to 73% of starting vitamin E.
HL, HAL, and TL Conditions. Solutions containing ∼200
ppm of I (about the same concentration of the oil samples) were
placed in cylindrical flasks as described above. At selected
times samples were withdrawn and analyzed. The experi-
ments in TL were carried out as follows: an aliquot of stock
solution (solvent anhydrous n-hexane) of I was placed in each
of 10 2 mL screw-capped borosilicate vials to reach the final
concentration of ≈200 ppm (0.46 M) in 0.2 mL. After the
solvent had been evaporated with a gentle nitrogen stream,
0.2 mL of a solution obtained by dissolving 1.0 g of T (0.56 M)
in 2.0 mL of anhydrous n-hexane was added to each vial. The
vials were sonicated for 5 min, capped, and irradiated with
the lamp. At selected times the vials were withdrawn, frozen
at -25 °C for 5 min, evaporated under N2, and recovered with
0.2 mL of CHCl3 containing the i.s. The solution was finally
injected and analyzed.
RESULTS AND DISCUSSION
Exp er im en ts in Olive Oil. Sunlight. In the long-
time sunlight experiment, the content of I showed a
peculiar behavior within time (Figure 2), which could
have been enhanced by the transmittance features of
the glass. From a comparison of the white bottles with
the dark ones, the following was observed:
Ar tificia l Ligh t. A water-cooled high-pressure mercury
lamp with a wavelength emittance of λ > 290 nm with a
maximum wavelength emittance at 360 nm (125 W, Iλ ) 3.97
× 10-7 E L-1 s-1, Helios Italquartz, Milan) was used.
Ap p a r a tu s a n d Ch r om a togr a p h y. HPLC. A Varian
5020 HPLC pump (Varian, Palo Alto, CA) was employed. The
pump was equipped with a Hewlett-Packard 1050 automatic
injector (Hewlett-Packard, Avondale, PA), an LC-235 diode
array, and an LCI-100 integrator (Perkin-Elmer, Norwalk,
CT). The samples (25-100 µL) were injected in Spherisorb
ODS2 and C8 columns (250 × 4.0 mm i.d., 5 µm, Waddinxveen,
The Netherlands). The operating conditions were as follows:
eluent mixture, methanol/acetonitrile (50:50, v/v); flow, 1.5 mL/
min; program of λ, at 295 nm from 0 to 9 min and λ ) 286 nm
1. In the sampling of May and September the content
of I in DBS was ≈26% higher than in WBS, whereas
that in DBI was ≈16% higher than in WBI.
2. In the WBS sample the maximum percent reduc-
tion occurred at the first sampling time.
3. Indirectly exposed bottles always showed a lower
reduction than directly exposed ones.

Photolysis of R-Tocopherols
Ta ble 1. Ch em ica l P a r a m eter s of th e Oil Exp osed to Su n ligh t d u r in g Lon g-Tim e Exp er im en t
Feb May Sept
J. Agric. Food Chem., Vol. 46, No. 11, 1998 4531
Nov
bottle
Aa
Pb
K270
K232
A
P
K270
K232
A
P
K270
K232
A
P
K270
K232
DBS
WBS
DBI
WBI
control
0.33
0.33
0.33
0.33
0.33
6.66
6.66
6.66
6.66
6.66
0.18
0.18
0.18
0.18
0.18
2.10
2.10
2.10
2.10
2.10
0.34
0.34
0.34
0.34
0.33
9.8
8.7
10.7
10.2
11.6
0.24
0.22
0.20
0.22
0.19
2.05
2.16
2.10
2.23
2.08
0.39
0.36
0.37
0.35
0.37
8.22
7.0
10.4
9.8
13.3
0.24
0.18
0.18
0.22
0.20
2.45
2.14
2.15
2.20
2.45
0.38
0.39
0.36
0.35
0.37
8.5
7.2
11.1
9.8
0.29
0.30
0.22
0.29
0.20
2.25
2.16
2.16
2.26
2.46
16.4
a
b
A ) acidity as percent oleic acid. P ) peroxides as mequiv of O2/L.
F igu r e 4. Variation of the concentration of I and peroxides
versus the irradiance received by an olive oil irradiated under
lamp (OL) and sunlight (WBS).
F igu r e 3. Variation of the concentration of I and of peroxides
in olive oil irradiated with λ > 290 nm.
4. The sample stored in the dark (cylindrical stainless
steel tank of 50 L) shows a high reduction of the content
of I only in the last sampling. This could be a conse-
quence of the headspace in the tank (∼15 cm).
Thus, in the period February-September, the differ-
ences among the samples exposed directly to sunlight,
those exposed indirectly in the laboratory, and the
control show that light plays a important role in the
disappearance of I.
The main chemical parameters of the oil during the
experiment are reported in Table 1. From these data
it was clear that at the first sampling time the K270
parameters of the oil sample already exceeded the limits
for the virgin olive oil (Directive CEE 2568/91).
The number of peroxides increased up to May and
later remained constant.
The coefficients of variation (CV) showed values
<10%.
F igu r e 5. HPLC chromatograms of R-tocopherol (A) and
photodegradation mixture (B).
Artificial Light. The same experiment was repeated
by lamp irradiation (λ > 290 nm). In these conditions
no degradation was observed in the blanks stored in the
dark. In both cases the concentration of I in the oil
decreases to 50% rapidly and linearly (r ) -0.9788) in
500 min; it then remained constant for about the same
time and later decreased to 17% of the starting concen-
tration in 2500 min. The peroxides showed an increase
up to I and a decrease to 50%, then remained constant
for another 500 min, and later progressively increased
(Figure 3). When the irradiance emitted from the lamp
was compared to the oil with solar actinic irradiance
received under sunlight by the WBS, similar behaviors
were found (Figure 4). During these experiments new
peaks never appeared in the HPLC chromatograms
when I disappeared from the oil. The peroxide behav-
iors were similar in the two experiments.
formed in the semipreparative scale under sunlight and
artificial light showed the same behaviors. I disap-
peared from the solution, yielding a main photoproduct
(2, tR ) 12.88 min; λmax ) 286 nm). After the appear-
ance of this signal, other signals at 6.23 min (3, λmax
)
266) nm, 6.95 min (4, λmax ) 300 nm), 8.04 min (5, λmax
) 300 nm), and 9.29 min (6, λmax ) 300 nm) were
present in the HPLC chromatograms (Figure 5). The
HPLC analysis of the six main fractions obtained by
column chromatography showed the following: unre-
acted I mixed with some minor unidentified peaks; in
the other four fractions, mixtures (in different ratios)
of signals 3-6 (I and signal 2 were present as impuri-
ties); the sixth fraction contains peak 2 alone. Prelimi-
nary UV and GC/MS data of the main peak of each
fraction seem to suggest the structure of the tocopheryl
quinone III (Figure 1) for peak 3 (M+ ) 430; UV
Ir r a d ia tion in Mod el System s a n d Str u ctu r e of
P h otop r od u cts. H Conditions. Experiments per-

4532 J. Agric. Food Chem., Vol. 46, No. 11, 1998
Pirisi et al.
Ta ble 2. 13C NMR Da ta of IIa
Photodegradation experiments of II alone were car-
ried out to understand the following: (1) whether
signals 3-6 could have II as a parent compound and
(2) how the global reaction proceeded kinetically.
The photodegradation of II was very slow (Table 3).
Because the HPLC chromatograms of II did not give
signals 3-6, a parallel kinetic process could be hypoth-
esized, which gave five species contemporaneously.
However, according to Frost and Pearson (1961), for
such a hypothesis to be valid, the formed products must
be in constant ratio with each other irrespective of the
time. Because this condition did not occur, such a
hypothesis must be rejected.
11.0 (CH3); 13.1 (CH3); 18.4 (CH2); 19.6 (CH3); 19.7 (CH3); 20.9 (CH2);
22.6 (CH3); 22.7 (CH3); 23.6 (CH3); 24.4 (CH2); 24.8 (CH2); 28 (CH);
29.7 (CH2); 32.6 (CH); 32.7 (CH); bs37.3-37.4 (4 × CH2); 39.4 (CH2);
39.6 (CH2); 68.9 (C); 114.5 (C); 117.5 (C); 122.2 (C); 138.4 (C); 144 (C);
155.8 (C); 193.9 (CH)
a
δ; the multiplicity was from DEPT experiments.
Ta ble 3. P seu d o-F ir st-Or d er Con sta n ts of th e
P h otod egr a d a tion of I a n d II in Differ en t Con d ition s
I
II
k
t1/2
(min)
k
t1/2
(min)
-1
-1
conditions 10-4
s
r2
10-4
s
r2
HS
HL
HAL
TL
1.03
1.86
1.41
0.98
112 -0.96711
62 -0.9985
82 -0.9719
117 -0.9836
0.03
0.49
4119 -0.9792
393 -0.9446
We should therefore assume that I was photode-
graded with a pseudo-first-order consecutive kinetics (eq
1), where II disappeared with K2 < K1 to give com-
pounds undetectable by HPLC.
spectrum totally overlapping with that of an authentic
specimen). For the peaks 4-6 we were unable to
propose a structure, because only few milligrams (0.5-
0.8) of each was isolated and the spectroscopic data were
scanty. Compound 2 was similar to a yellow oil, and
its elemental analysis agreed with the formula C29H48O3.
The IR spectrum showed the presence of a carbonyl
group at 1680 cm-1, and the GC/MS analysis showed
the molecular ion at m/z 444, the base peak at m/z 201,
and another two important ions at m/z 220 and 179.
The GC/MS spectrum of I shows the molecular ion at
m/z 430 and other important peaks at m/z 165 and 205.
The mass spectra of I and peak 2 exhibit similar
behaviors. The molecular ion at m/z 444 of peak 2 could
correspond to an R-tocopherol-like structure, with a
mass higher than the molecular ion of I by ≈14.
Moreover, the ions at m/z 220 and 179 of peak 2 could
be considered “derivatives” of the ions at m/z 205 and
165 of I, plus a m/z ) 14, respectively. The ion at m/z
201 does not have a corresponding ion in I. 1H and 13C
NMR analyses have provided the necessary information
to identify compound 2 as 5-formyl-R-tocopherol (II)
(Figure 1).
K1
9
K2
9
Ι
8 II
8 undetectable compounds
(1)
This hypothesis has been confirmed by a long-time
photodegradation of I, where, when the concentrations
of other photoproducts were considered to be negligible,
the concentrations of I and II were experimentally found
to be in good accordance with those calculated by eq 2
of the consecutive process
[I0]k1
t
(e-k t - e-k
)
(2)
1
2
[II] )
k2 - k1
where Kobs ) k1 and [Ι0] is the starting concentration of
the R-tochopherol (Frost and Pearson, 1961).
HAL Conditions. Also in this case the disappearance
of I followed a pseudo-first-order kinetics. Some differ-
ences occurred compared to the behaviors in HS and HL
conditions. When the content of unreacted I was of the
same order in both experiments, the Kobs values were
lower and II was formed in very small concentrations
(<5%). Furthermore, the HPLC chromatograms showed
neither signals of 3-6 nor other compounds.
The 1H spectrum of compound II is different from that
of compound I in the following: (a) one singlet at δ
10.12, indicating the presence of a CHO group with a
strong intramolecular hydrogen bonding with an OH (δ
12.05, exchangeable hydrogen with D2O); (b) only two
singlets attributable to CH3 on the aromatic ring of I.
The above results and NOE experiment (strong nega-
tive Overhauser effect of the CH2 group in position 4
on the aldehyde signal) suggest that the CH3 group in
the 5-position (in I) of the aromatic ring has been
oxidized into a CHO group. Further 13C NMR data
(Table 2) confirm the identification.
Several authors report the formation of formyl de-
rivatives from I during the autoxidation of vitamin E
(Fujimaki et al., 1970 and, after irradiation with γ-rays
in two different solvents, n-hexane and CHCl3 (Molnar
et al., 1992).
TL Conditions. Under these conditions the disappear-
ance of I followed a pseudo-first-order kinetics with a
slower constant rate than in HS, HL, and HAL conditions.
Moreover, the appearance of II also occurred as in HAL.
In all experiments the Kobs values of the disappear-
ance of I show CV ranging between 8.4 and 11.0%.
CONCLUSIONS
From these data it can be concluded that in the model
systems employed the photodegradation of I followed
the consecutive kinetic process summarized in eq 2,
where II was the main photoproduct. Under HS and HL
conditions, the unknown species present in very low
amounts might be due to secondary reactions of the
radical species, I•, the intermediate product that led to
all photocompounds. Thus, eq 1 can be rewritten as
follows:
Kin etic Tr ia ls. HS and HL Conditions. Under
sunlight and artificial light II and I showed similar
behaviors. I disappeared from the irradiated solutions
according to a pseudo-first-order kinetics (Table 3) and
yielded photoproducts. Experiments over a long period
(> 3t1/2) showed an increase of compound II for up to
2t1/2. Compound II then slowly disappeared, and no new
signals appeared in the chromatograms. The other
photoproducts, peaks 3-6, reached a constant value
soon after they appeared and were stable until I
disappeared completely.
Ki
K1
K2
[I]
9 98 [II] 98 undetectable compounds
8 [I•]
Of course Ki . K1, and the overall kinetic process
depends on K1 (Burton and Ingold, 1986). Thus, from
I•, via rearrangements, the quinone III could be gener-
ated and, from it, compound VII. The latter species
produced II by Michael addition of H2O (Fujimaki et
al., 1970; Suarna et al., 1988). Under HS and HL
conditions, II was produced because n-hexane contains
0.01% water (corresponding to ∼5 × 10-3 M), which is

Photolysis of R-Tocopherols
J. Agric. Food Chem., Vol. 46, No. 11, 1998 4533
Fujimaki, M.; Knamaru, K.; Kurata, T.; Igarashi, O. Agric.
Biol. Chem. 1970, 34, 1781-1787.
Grams, G. W.; Eskins, K.; Inglett, G. E. Dye-sensitized
photooxidation of R-Tocopherol. J . Am. Chem. Soc. 1972, 94,
866-868.
J ore, D.; Ferradini, C. Radiolytic study of R-Tocopherol oxida-
tion in ethanolic solution. FEBS Lett. 1985, 183 (2), 299-
303.
Knapp, F. W.; Tappel, A. L. Comparison of the radiosensitivi-
ties of the fat-soluble Vitamins by gamma irradiation. J .
Agric. Food Chem. 1961, 9, 430-433.
Kramer K. A.; Liebler, D. C. UVB induced photooxidation of
Vitamin E Chem. Res. Toxicol. 1997, 10, 219-224.
Molnar, I.; Koswig, S. Investigation of γ-irradiation of R-To-
copherol and its related derivatives by high-performance
liquid chromatography using a rapid scanning spectropho-
tometer. J . Chromatogr. 1992, 605, 49-62.
Servili, M.; Baldioli, M; Miniati, E.; Montedoro, G. F. Antioxi-
dant activity of new phenolic compounds extracted from
virgin olive oil and their interaction with R-Tocopherol and
â-carotene. Riv. Ital. Sostanze Grasse 1996, 72, 55-59.
Suarna, C.; Craig, D. C.; Cross, K. J .; Southwell-Keely, P.
Oxidation of Vitamin E (R-Tocopherol) and its model com-
pound 2,2,5,7,8-pentamethyl-6-hydroxy-chroman. A new
dimer. J . Org. Chem. 1988, 53, 1281-1284.
of the same order as that of I. The nature of the further
degraded photoderivatives is unknown. The presence
of II in HAL and TL experiments in low concentrations
(<5%) could be due to impurities of water contained in
the hexane and in triolein. On the contrary, the absence
of II in OS and OL experiments could be explained by
assuming that I• mostly reacts with the peroxyl radicals
generated from the fatty acids during irradiation. This
should explain the constancy in milliequivalents of O2
per kilogram of the oil during the time. Thus, when
the concentration of I was <20 mg/kg, the number of
peroxides increased rapidly. This finding agrees with
the data of Servili et al. (1996).
LITERATURE CITED
Burton, G. W.; Ingold, K. U. Vitamin E: Application of the
Principles of Physical Organic Chemistry to the exploration
of its Structure and Function. Acc. Chem. Res. 1986, 19,
194-201.
Choudhry, G. G.; Webster, B. G. R. Protocol guidelines for the
photolytic study of pesticides in water air and soils. Residue
Rev. 1985, 96, 79-136.
Clough, R. L.; Yee, B. G.; Foote, C. S. Chemistry of singlet
oxygen. 30. The unstable primary product of tocopherol
photooxidation J . Am. Chem. Soc. 1979, 101, 683-686.
Directive CEE 2568/91. Gazzetta Ufficiale CEE L 248, Sept
5, 1991; pp 1-83.
Yamauchi, R.; Kato, K.; Ueno, Y. Free radical scavening
Reactions of R-Tocopherol during the autoxidation of Methyl
Linoleate in Bulk Phase. J . Agric. Food Chem. 1995, 43,
1455-1461 and references cited therein.
Encyclopedia of Food Science and Technology; Hui, Y. H., Ed.;
Wiley: New York, 1992; Vol. 4, pp 2718-2722.
Frost, A. A.; Pearson, R. G. Kinetics and Mechanism; Wiley:
New York, 1961; pp 166-178.
Received for review March 18, 1998. Revised manuscript
received August 3, 1998. Accepted August 20, 1998.
J F980239N