Thermally Stimulated Luminescence in Powdered

Soy Proteins
Dereje Abdi, Muhammad S. Jahan, William L. Boatright, Benjamin M. Walters, and Qingxin Lei

C: Food Chemistry
Abstract: Heating powder isolated soy proteins (ISPs) in a N2 environment produced thermally stimulated lumi-
nescence (TSL), in 2 major temperature regions, 50 to 250◦ C (region R1) and 250 to 350◦ C (region R2). In soy
protein 7S fraction, strong TSL was detected in both regions with glow peak maximum (Tm ) at 150 ± 15◦ C and at
300 ± 10◦ C. Two additional satellite or shoulder peaks were detected from the ISP and 7S protein fraction within region
R1 at Tm = 90◦ C and Tm = 210◦ C. The soy protein 11S fraction produced a broad, poorly defined TSL peak in the
low-temperature region. Electron paramagnetic resonance spectroscopy data from the control ISP sample, deuterium
sulfide-treated ISP, ISP stored in either N2 or O2 , and defatted soy flour, indicated that the trapped radicals present in
ISP is associated with the production of the primary TSL peak at 150 ± 15◦ C. Activation energies required to release
the trapped charges (for luminescence to occur) are approximately 0.70, 0.78, 1.50, and 1.8 eV for TSL at Tm = 100,
150, 200, and 300◦ C, respectively. The reaction mechanism that leads to the release of the trapped charges for TSL to
occur followed a mixed order kinetic, between 1.5 and 1.8. The frequency factor varied between 107 /s and 1017 /s.

Keywords: electron paramagnetic resonance spectroscopy, free-radicals, soy protein, thermally stimulated luminescence

Practical Application: Free radicals are capable of catalyzing oxidative degradation of food components, and powdered
soy proteins typically contain from 10 to 100 times more metastable radicals than other protein sources. The research
described in this paper provides novel information about the nature of these radicals that can be used to develop processes
that can minimize the content of free radicals in foods containing soy proteins.

Introduction
Thermally stimulated luminescence (TSL), also known as ther-
moluminescence (TL), has been described as a radiation-specific
phenomenon in which a material emits light or luminescence
(UV-visible region) when radiation-induced trapped electrons are
thermally released from the traps and are recombined with holes
or luminescent centers. In inorganic crystalline solids, TSL has
been extensively used to study defects or impurities and color
centers. Readers are referred to an introductory chapter on TSL
by Horowitz (1984). In organic solids and polymers, TSL is at-
tributed to the presence of impurities, molecular imperfections,
voids within the polymer matrix, and a host of other factors in-
cluding free radicals (Fleming 1990). Because of this, TSL can
also be described as a defect or impurity-specific phenomenon. In
1989, a U.S. patent was issued for detection of surface impurities
in high-temperature superconducting materials using TSL tech-
nique (Cooke and Jahan 1989; Jahan and others 1991), and, by
employing the same technique, Jahan and others (1991) detected
the surface impurity phases in high-temperature superconducting
materials.
In foods, herbs, and spices, TSL has been attributed to impurity
contents, and it has been used as a diagnostic tool for deter-
mination of the radiation history or radiation processing results
MS 20130979 Submitted 7/16/2013, Accepted 11/4/2013. Authors Abdi,
Jahan, and Walters are with Dept. of Physics, Univ. of Memphis, 216 Manning
Hall, Memphis, TN 38152, U.S.A. Authors Boatright and Lei are with Dept.
of Animal and Food Sciences,Univ. of Kentucky, 412 W.P. Garrigus Building, Lex-
ington, KY 40546-0215, U.S.A. Direct inquiries to author Boatright (E-mail: wl-
boat1@uky.edu).
(Glidewell and others 1993; Kiss and Kispeter 1995; Ziegelmann
and others 1999; Carmichael and others 2000; Raffi and oth-
ers 2000). Since radiation-induced radicals can remain trapped in
a solid (organic material, polymer, or food), one would expect
TSL from such a solid when radicals are thermally recombined
or thermally released electrons recombine with radicals or radical
ions. In the 1960s, Charlesby and Partridge (1963a, 1963b, 1965a,
1965b) proposed that upon capturing a thermally detrapped elec-
tron, a carbonyl group (present in irradiated hydrocarbons) moved
to excited state, and subsequent de-excitation of which emitted
a photon or luminescence (Charlesby and Partridge 1963a; Par-
tridge 1972). The 2nd model, chemiluminescence (CL), suggests
that luminescence is the result of peroxide or hydroperoxide de-
composition into excited carbonyl groups which de-excite by
emitting photons (Broska 2001; Corrales 2002). Jacobson (2001)
used TSL technique to record luminescence by recording TSL at
a constant temperature or time and attributed the observed lumi-
nescence (defined as chemiluminescence, CL) to the breakdown
of hydroperoxy followed by de-excitation of the excited RC = O
species. Thus, the TSL or CL was directly correlated to the pres-
ence of ROOH rather than to its precursor, the trapped radicals.
In low-temperature studies by Serpi and others (1975) , Charlesby
and Partridge (1963a) and Mele and others (1968), TSL could not
be correlated to radical–radical recombination because the radi-
cals were found to remain frozen (inactive) during (active) TSL
emission.
Kiss and Kispeter (1995) examined the TSL curves of intact
gamma-irradiated milk protein, and reported activation energies
of electron traps to be 0.65 and 1.6 eV for the glow peaks at
425 K (152◦ C) and 500 K (227◦ C), respectively. Without any



C 2013 Institute of Food Technologists
R

doi: 10.1111/1750-3841.12325 Vol. 79, Nr. 1, 2014 r Journal of Food Science C25
Further reproduction without permission is prohibited
 Thermally stimulated luminescence . . .
electron paramagnetic resonance (EPR) data, however, the role
of milk protein radical in the production of thermoluminescence
dosimetry (TSI) could not be ascertained. Mamoon (1995) re-
ported glow curves of intact cornstarch, “Semilac” milk powder,
and Klim milk powder both before and after irradiation treatment.
Only the irradiated samples produced broad glow peaks around
175 to 200◦ C. Ahn and others (2013) compared the glow curves
C: Food Chemistry
from irradiated and nonirradiated intact garlic powder.
In other studies, TSL was accomplished on extracted inorganic
components as an indicator or irradiation treatment. Raffi and
others (2000) compared EPR and TSL results of extracted minerals
of irradiated aromatic herbs, spices, and fruits, and Sanyal and
others (2009) examined gamma-irradiated Basmati rice, which
was found to produce strong TSL and EPR signals. In current
work, we observed TSL in nonirradiated commercial soy protein
containing metastable radicals.
Isolated soybean proteins (ISPs) typically are not treated with
ionizing radiation; however, the free radicals trapped in “dry”
soy proteins produce EPR spectra similar to protein exposed to
ionizing radiation (Boatright and others 2008; Boatright and oth-
ers 2009). The free radical contents in commercial food products
containing soy proteins ranged from 6.12 × 1014 to 9.10 × 1015
spins/g of soy protein, which is from 10 to 98 times greater than
other food protein sources examined. Tritiated hydrogen sulfide
has been used to specifically label the location of carbon-radicals
within dry proteins (Riesz and others 1966; White and others
1967; Riesz and others 1968; Riesz and White 1970). Lei and
others employed deuterium sulfide (D2 S) to quench the primary
EPR signal at g = 2.005 in powdered soy proteins. The resulting
isotopic labeling revealed that the metastable radicals were located
on Ala, Gly, Leu, Ile, Asx (Asp+Asn), Glx (Glu+Gln), and Trp,
while the results for tyrosine were inconclusive due to nonspecific
deuterium—hydrogen exchange, which occurred during subse-
quent sample processing (Lei and others 2010). In the current
investigation, TSL results were used in conjunction with EPR
data to learn more about metastable (trapped) charges and radicals
in soy proteins.
Materials and Methods
ISPs were obtained from local markets or prepared in the lab-
oratory according to the procedure of Stine and others (2004).
Small-scale preparation of 7S and 11S protein fractions (hence-
forth, they are labeled as 7S and 11S) from defatted soy flour (DF)
was achieved by the protocol described by Nagano and others
(1992) with modification. DF (20 g), obtained from the Archer
Daniels Midland Co. (Decatur, Ill., U.S.A.), was suspended with
300 mL of deionized water at 23◦ C. The pH of the slurry was
adjusted to 7.5 with 1.0 N NaOH, followed by being stirred for
1 h. Then the slurry was centrifuged at 3000 × g for 20 min at
20◦ C and the supernatant was filtered with layers of cheese clothes.
The filtrate was further centrifuged at 9000 × g for 30 min at 4◦ C.
The pH was adjusted to 6.4 with 2 N HCl, and the mixture was
kept in ice bath overnight. The 11S globulin fraction was ob-
tained as slightly yellowish pellet from the centrifugation of above
mixture at 6500 × g for 20 min at 4◦ C, which was resuspended
with 50 mL of deionized water and the pH of the suspension was
adjusted to 7.0 before being freeze dried. To the supernatant from
6500 × g centrifugation, solid NaCl was added to a final con-
centration of 0.25 M, followed by adjusting the pH to 5.0 with
2 N HCl. While keeping the temperature at 4◦ C with ice-water
bath, the mixture was stirred for 1 h before being centrifuged at
9000 × g for 30 min at 4◦ C to remove the insoluble fraction.
C26 Journal of Food Science r Vol. 79, Nr. 1, 2014
The supernatant was diluted 2-fold with ice-cold water and kept
at ice-water bath for 30 min, then its pH was adjusted to 4.8 with
2 N HCl. The 7S globulin fraction, obtained as precipitate from
centrifugation (6500 × g for 20 min at 4◦ C), was washed twice
with distilled water, and resuspended with 50 mL of deionized
water. After being adjusted pH to 7.0 with 2 N NaOH, the slurry
was lyophilized.
Quenching the metastable radicals in ISP with D2 S
Soy protein samples (approximately 1 g) were loosely wrapped
with filter paper and introduced into the sample chamber. Af-
ter being sealed, air was removed from the system with a high
vacuum (<0.02 Torr) for 25 min, followed by introducing D2 S
and maintaining its pressure at 50 psi for 12 h at 23◦ C. The D2 S
was removed from the chamber by vacuum treatment before the
treated ISP samples were removed.
Measurement techniques
TSL. TSL measurements were performed using a Harshaw-
Bicron 3500 commercial dosimeter. For each measurement, ap-
proximately 2.1 mg ISP powder was evenly distributed on an
open (without any lid) differential scanning calorimeter (DSC)
pan. The DSC-pan-with-ISP powder was then placed on the
heating planchet (sample chamber) of the TLD reader, and it was
then heated from 25 to 350◦ C at a rate of 1◦ C/s. To avoid any
oxygen-induced reaction during heating, the sample chamber was
continuously purged with dry nitrogen during TSL detection. A
gas filter (0.003 μm, model# TEM-811-P, TEM Filter Co., Santa
Clara, Calif., U.S.A.) was used to remove particulates and oil from
the nitrogen gas. The basic components of a TSL unit are the sam-
ple chamber fitted with a heating planchet, temperature controller,
photomultiplier tube (PMT), current amplifier, and a computer.
Before any sample run, system calibration was performed using
an internal light source, and a background test was performed
with no sample. Care was taken to maintain the sample mass and
its distribution on the DSC pan the same for all tests. In addi-
tion, to check reproducibility, at least 3 TSL measurements were
performed for each type of sample, and an average glow curve
was produced for analysis purpose (as shown in the figures). The
TSL output (glow curve) was recorded by a dedicated computer.
(Harshaw/Bicron Radiation Measurement Products, Solon, Ohio,
U.S.A).
EPR spectroscopy. EPR measurements were performed at
23◦ C (room temperature) using a Bruker EMX series spectrometer
(EMX 300) with the following parameters: microwave frequency
of 9.42 GHz; receiver gain 2 × 104 ; modulation amplitude 5.0 G;
modulation and detection frequency 100 kHz; microwave power
1 mW; time constant 327.68 ms, conversion time 183.84 ms; and
field sweep of 800 Gauss. To produce 1st-derivative (also known
as derivative) spectrum, as presented in the figures, EPR signals
were detected at the modulation frequency (100 kHz). Soy protein
samples were packed into a 4-mm quartz sample tube (Wilmad,
Buena, N.J., U.S.A.) to a density of 0.03 ± 0.002 g/cm of tube
length. The g-factor axis was calibrated relative to crystalline 1,1-
diphenyl-2-picryl hydrazyl (DPPH) using Bruker WINEPR Sys-
tem software. The g-factor axis was calibrated relative to a crys-
talline DPPH standard at 2.0036 using Bruker WINEPR System
software. Spin concentration was determined by performing dou-
ble integral of the 1st-derivative EPR spectrum of a sample and
comparing it with that of a Ruby standard (SRM 2061, NIST,
Bethesda, Md., U.S.A.) containing a known spin concentration.
Spin concentration (free radical concentration, FRC) within the
Thermally stimulated luminescence . . .

EPR cavity (calculated from the standard curve) was divided by

the mass of the sample in the cavity to obtain the “spins/g” or the
“radicals/g” of the sample.

TSL analysis
Although glow-curve intensity (y-axis) was recorded in arbi-

C: Food Chemistry
trary units as a function of temperature, the total light output was
given as a product of the PMT current in nA and time in s, nAt =
nC (charge). Each experimental glow curve was deconvoluted to
individual or component glow curves using PeakFit program, ver-
sion 4.12 (SeaSolve Software Inc. 1999 to 2003). TSL parameters
characterizing each individual glow curve were then determined
using a TSL simulation program known as TSL CurveFit, version
1.5.0 (Los Alamos Natl. Laboratory, 2000). This program requires
users to enter the heating rate (b), start temperature (T0 ), end
temperature (T ) and peak temperature (Tm ), and the magnitudes
of the lower half (t) and the upper half (d ) at full–width at half
Figure 1–TSL from commercial ISP control. The individual (deconvoluted)
maximum (FWHM) (see the glow curve in Region 2 in Figure 2, glow curves of ISP powder are labeled as, from left (lowest temperature) to
for example). When best fit is achieved (theoretical curve matches right (highest temperature), Curve 1 (Tm = 90◦ C), Curve 2 (Tm = 140◦ C),
the experimental curve) TSL parameters such as shape parameter Curve 3 (Tm = 210◦ C), and Curve 4 (Tm = 318◦ C).
(μg ), kinetics order (b), activation energy (E) and frequency factor
(s) are produced.
The basic formula used for calculation of activation energy is formulation is needed to explain TSL due to multiple trapping
given by Eq. 1 (Chen 1981): sites and recombination centers. Most materials have traps of dif-
fering depths, which imply that for a given temperature electrons
  in deeper traps are released after the shallower-trap electrons. In
k B Tm
Et = [2.52 + 10.2(μg − 0.42)] − 2k B Tm (1) the single trap/recombination-center formulation, the number of
ω
trapped electrons is equal to the number of holes, but in the case of
multiple trapping sites, the numbers can be different and the deep
where ω = δ + τ is the FWHM, δ and τ are the high and low
traps may capture electrons released from shallower traps at Et . A
temperature half widths, respectively; μγ = δ/ω is an asymmetric
combination of these factors can have influence on the number of
parameter or shape parameter of a glow curve which determines
glow peaks and/or glow curve shape in a given material. It has to
the kinetic order, and Tm is the maximum peak temperature. In
be emphasized that TSL from nontransparent solids is primarily a
1st order kinetics μg = 0.42, and Eq. 1 reduces to
surface phenomenon. Fortunately, powder particles provide a large
  surface area and, as a consequence, a small amount of ISP powder,
k B Tm approximately 2 mg, produces significantly high TSL. Although
Et = 2.52 − 2k B Tm (2)
ω care was taken to distribute the powder uniformly on the sample
holder (DSC pan) in order to minimize the temperature gradient
With known μg for each well-defined glow peak (deconvo-
luted), kinetics order (b) of the corresponding TSL process can
be found from Chen’s μg versus b plot (Chen 1977; Chen and
McKeever 1997). The activation energy is the depth of the po-
tential well in which the electron is trapped. Using the glow-peak
parameters (glow-peak temperature Tm , heating rate b, activation
energy Et , frequency factor s, and kinetic order b) several iterative
calculation of the TSL intensity Eq. 3 yields best-fitting parameters
that characterize each individual glow curve:

⎡ ⎤− b −1
b
  T  
Et ⎣ s Et
I T L = n 0 s exp 1 + (b − 1) exp − dT  ⎦
kB T β kB T
T0
(3)
In Eq. 3, n0 is the concentration of trapped charges at time t = 0,
Et is the trap activation energy, T is the absolute temperature and
kB is Boltzmann’s constant. The frequency factor s is the attempt
to escape frequency associated the transition from the trapping
level to the excitation level.
Figure 2–TSL from soy protein 7S fraction. The individual (deconvoluted)
Although the above formula is quite sufficient to explain TSL glow curves of 7S powder are labeled as, from left (lowest temperature) to
phenomena in materials containing only a single trapping level as right (highest temperature), Curve 1 (Tm = 90◦ C), Curve 2 (Tm = 140◦ C),
well as a single type of recombination center, a more complex Curve 3 (Tm = 210◦ C), and Curve 4 (Tm = 318◦ C).

Vol. 79, Nr. 1, 2014 r Journal of Food Science C27

 Thermally stimulated luminescence . . .

Table 1–Tabulation of calculated glow curve parameters of iso- Results and Discussion
lated soy protein (ISP).
When nonirradiated ISP powder is heated from room temper-
Peak Tm (◦ C) τ δ μg b E (eV) S (per s) ature (23◦ C) to 350◦ C in N2 environment in the heating cham-
1 104 26 26 0.50 1.60 0.75 6 E07 ber of the commercial TSL detector, used in this study and as
2 157 34 34 0.50 1.76 0.73 5 E07 described earlier, broad-band light (UV-visible) or luminescence
3 213 23 23 0.49 1.54 1.42 2 E15 (also known as glow) are observed or detected in 2 major temper-
4 309 24 24 0.49 1.64 2.00 3 E17 ature regions, Region 1 (50 to 250◦ C) and Region 2 (250 to
C: Food Chemistry

350◦ C; Figure 1). Region 1 is defined or labeled as a low-

between the bottom and the top of the powder particles, slight
variation in distribution could shift the glow-peak temperature
and/or shape of the glow curve (see “Results and Discussion”
section).
Figure 3–TSL from soy protein 11S fraction. The individual (deconvoluted)
glow curves of 11S powder are labeled as, from left (lowest tempera-
ture) to right (highest temperature), Curve 1 (Tm = 90◦ C), Curve 2 (Tm =
140◦ C), Curve 3 (Tm = 210◦ C), and Curve 4 (Tm = 318◦ C).
temperature region in which the peak (maximum) of the lu-
minescence or glow appears around 150 ± 15◦ C, which is by
definition the peak temperature Tm (see Eq. 1 to 3). The high-
temperature region with Tm = 300 ± 10◦ C is labeled as Region
2. A large variation in Tm (approximately, ±15◦ C in Region 1
and ± 10◦ C in Region 2) could be attributed to nonhomogeneity
within the ISP powder, and/or to the differences resulting from
uniform/nonuniform distribution of a very small amount of pow-
der (approximately 2 mg) on the heating pan. A typical glow curve
(TSL intensity as a function of temperature) for 7S fraction of ISP
powder is shown in Figure 2. Because the ability of powdered
soy proteins to trap metastable radicals is likely related to protein
structure, TSL was use to examine the 7S and 11S soy protein
fractions. These 2 protein fractions are primarily comprised of
the 2 major, and structurally distinct, soy proteins; β-conglycinin
and glycinin, respectively. It is clear from the figure that the TSL
output in Region 1 is broad and that it is narrow and relatively
intense in Region 2. TSL analysis, as described above, revealed 2
additional satellite or shoulder peaks within Region 1, while the
glow in the Region 2 remained as a single glow curve. Analysis
program also yielded TSL parameters such as shape parameter (μg ),
kinetics order (b), activation energy (E), and frequency factor (S)
for each one of these glow curves (see Table 1). The method for
computing glow-peak shape parameter, μg = δ/ω, is illustrated
in Figure 2 using the 300◦ C glow curve, where the FWHM is
given by ω = δ + τ . Similar analyses of TSL from ISP 11S fraction

Figure 4–TSL from commercial ISP treated

with deuterium sulfide. The individual
(deconvoluted) glow curves are labeled
as, from left (lowest temperature) to right
(highest temperature), Curve 1 (Tm =
90◦ C), Curve 2 (Tm = 140◦ C), Curve 3 (Tm
= 210◦ C), and Curve 4 (Tm = 318◦ C).

C28 Journal of Food Science r Vol. 79, Nr. 1, 2014

Thermally stimulated luminescence . . .
produced an intense glow in Region 2, Curve 4 with Tm = 310◦ C,
μg = 0.5, b = 1.9, E = 1.9 eV, and S = 4 × 1017 /s (Figure 3). The
TSL glow of this particular fraction of ISP in Region 1 was found
to be very weak (remained slightly above the background) and, as
a result, Curves 1 to 3 could not be resolved. Further discussion
will follow. The corresponding EPR spectra from the 7S and 11S
soy protein fractions both produced a primary peak at g = 2.0049
that was indistinguishable from the spectrum from ISP in both
Figure 5–First-derivative EPR spectra of commercial ISP control and the
same ISP treated with D2 S.
Figure 6–TSL from laboratory ISP samples stored in either oxygen at strong manganese-II signal in the defatted flour that is not clearly
◦ ◦
40 C, or in nitrogen at 23 C.
location and shape (spectra not shown). The corresponding EPR
peak area from the 7S fraction was 1.092 × 106 and the EPR peak
area from the 11S fraction was 6.094 × 105 .
A comparison of TSL and EPR results were made for 3 dif-
ferent ISP treatments/preparations: (1) D2 S-treated ISP compared
with the control ISP; (2) ISP stored in N2 compared with ISP
stored in O2 ; and (3) ISP compared with DF. The purpose of
C: Food Chemistry
these comparisons was to determine if the production of TSL,
total or any individual glow (Curves 1 to 4), is associated with
radical contents (FRC) in a given preparation. It has to be further
emphasized that the samples tested in this study produced identical
glow curves (within experimental variations), and that the analyses
of total glow curve and the TSL parameters for each glow curve
(deconvoluted) are presented in the previous section (see, for ex-
ample, Figure 2 and Table 1). Commercial ISP powder containing
trapped radicals of 2.09 × 1015 spins/g ISP produced stronger TSL
glow in Region 1 (Figure 1). Note, on intensity scale, the Curves
1, 3, and 4 are within the same range whereas the Curve 2 is about
4.5 times more intense. ISP D2 S produced the same glow curves
(deconvoluted); however, the intensity of glow in Region 1 (see
intensity axes in Figure 1 and 4) is approximately 10 times weaker
than that of the commercial ISP (Figure 4). It is also clear from
the intensity axes of the Figure 1 and 4 that the glow Curve 4 in
Region 2 was not affected by the variation in the materials radical
content. The EPR spectra of these 2 powder samples are shown
in Figure 5. Evidently, the TSL glow with Tm near 150◦ C, which
is represented by the theoretical (deconvoluted) glow Curve 2, is
dependent on the radical content (FRC) in the soy powders.
TSL outputs comparing laboratory prepared ISP powdered
stored in N2 and O2 is shown in Figure 6. Strikingly, the ISP
powder stored in O2 at 40◦ C, which contains high FRC, also
produced high TSL as shown by Curve 2 with Tm = 150◦ C. It is
also illustrated by Figure 6 that these 2 samples produced identical
glow in Region 2 (Curve 4 with Tm = 300◦ C), and similar glows
at 100◦ C (Curve 1) and 200◦ C (Curve 3). The ISP sample stored
in nitrogen exhibited an EPR peak area of 2.513 × 105 and the
ISP sample stored in oxygen exhibited a peak area of 1.014 × 106
(approximately a 4-fold higher radical content in the sample stored
in oxygen).
Figure 7(A) and 7(B) shows the comparison between the TSL
and EPR data from a sample of commericial soy protein isolate
and defatted soy protein. The most striking difference between
the 2 is that the glow in the high-temperature region, Curve 4
(Tm = 300◦ C) in Region 2, typically found in ISP, is not detectable
in defatted flour. Also, as reported previously, defatted flour has a
Figure 7–Comparison of TSL (A) and EPR
(B) from commercial soy protein isolate
(ISP) and defatted soybean flour.
Vol. 79, Nr. 1, 2014 r Journal of Food Science C29
 Thermally stimulated luminescence . . .
detectable in commercial ISP samples (Boatright and others 2008;
Boatright and Jahan 2013).
On the basis of the TSL and EPR data obtained from the mea-
surements of the control compared with the D2 S-treated ISP sam-
ples (Figure 1, 4, and 5) and the ISP stored in the O2 /N2 environ-
ment (Figure 6), it is apparent that the primary luminescence near
150◦ C in region R1 results from trapped (metastable) radicals. TSL
C: Food Chemistry
and EPR data from the 7S and 11S protein fractions indicate that
the majority of the metastable radicals are located in the soy pro-
tein β-conglycinin. The satellite or secondary glows in the same
region, Curve 1 (Tm = 100◦ C) and Curve 3 (Tm = 200◦ C), could
be associated with impurities and/or imperfections. Notice that
these latter glow curves, which appear as weak shoulders within the
primary TSL in region R1, can only be revealed via deconvolu-
tion method. Therefore, one could argue that, like the glow Curve
2 (Tm = 150◦ C), these weak glows are also associated with the
trapped radicals because they appear within the same TSL region
(R1). But in the total glow from the ISP sample stored in N2 at
23◦ C in Figure 6, these glows stand out clear (though weak)
whereas the main glow Curve 2 (Tm = 150◦ C) disappear. This
result further suggests that the TSL glows near 100◦ C (Curve 1)
and 200◦ C (Curve 3) may be associated with species other than
the trapped radicals.
The TSL in ISP in the high-temperature region, R2 (250 to
400◦ C), does not appear to be associated with trapped radicals.
It is likely to be associated with impurities and/or molecular im-
perfections or defects present in the ISP. In defatted flour (DF),
however, the 300◦ C peak disappeared completely and the TSL
intensity in region R1 was reduced significantly. Without ad-
ditional work, TSL in DF may not be understood; however,
one may speculate that the paramagnetic Mn2+ present in soy
proteins, as revealed by EPR measurements (Boatright and oth-
ers 2008; Boatright and Jahan 2013), could play an important
role.
Raffi and others (2000) observed EPR singlet at g = 2.0050 in
nonirradiated herbs, fruits, and spices, and they attributed the cor-
responding TSL to the mineral contents, primarily silicates. Very
weak TSL (relative to those detected after gamma irradiation)
was detected in nonirradiated shellfish species (Nephrops norvegicus)
following extraction of silicate minerals (Carmichael and others
2000). The TSL glow peak appeared between 300 and 400◦ C
when extraction was performed manually and between 200 and
300◦ C when chemical hydrolysis was used. A broad glow rang-
ing from 150 to 350◦ C with a peak near 240◦ C was detected in
nonirradiated mollusc shells (Ziegelmann and others 1999). The
corresponding EPR singlet at g = 2.0055 was attributed to unde-
termined organic paramagnetic species.
In gamma-irradiated milk protein concentrate, Kiss and Kispeter
(1995) observed a broad TSL glow in temperature region 50 to
300◦ C with peaks near 154 and 227◦ C. The corresponding acti-
vation energies were reported to be 0.7 and 1.6 eV, respectively. As
computed in this study, the activation energies required to release
the trapped charges (for luminescence to occur) are approximately
0.70, 0.78, 1.50, and 1.8 eV for TSL at Tm = 100, 150, 200, and
300◦ C, respectively. The kinetic orders (b in Eq. 3) for the TSL
processes at these temperatures are found to vary between 1.5 and
1.8 indicating the chance of retrapping of the released charges is
moderate to high. According to Kiss and Kispeter (1995), TSL
in milk protein concentrate follows a 2nd order (b = 2) process.
Although physical implication may not be apparent, the frequency
factors associated with the charge release processes vary between
107 and 1015 /s.
C30 Journal of Food Science r Vol. 79, Nr. 1, 2014
Conclusion
In conjunction with EPR data, TSL results of this study find that
the trapped radical present in nonirradiated ISP is associated with
the production of the primary TSL near 150 ± 15◦ C. Additional
weak luminescence near 100 and 200◦ C, and strong luminescence
at 300◦ C could be associated with mineral impurities or defects.
The reaction mechanism that leads to the release of trapped charges
and subsequently produces TSL follows a mixed order kinetic,
between 1.5 and 1.8. Additional work is needed to elucidate TSL
mechanism at 300◦ C.
Acknowledgments
This project was supported by the Natl. Research Initiative of
the USDA Cooperative State Research, Education and Extension
Service, grant number 2009-35503-05190. The information re-
ported in this paper (Nr. 13-07-121) is part of a project of the
Kentucky Agricultural Experiment Station and is published with
the approval of the Director.
References
Ahn J-J, Akram K, Baek J-Y, Lee J, Jeong I-Y, Kwon J-H. 2013. Identification of gamma-
irradiated ingredients in liquid seasonings by thermoluminescence analysis: and interlaboratory
blind trial. Eur Food Res Technol 236:771–6.
Boatright WL, Jahan MS. 2013. Effect of sequestering intrinsic iron on the electron paramagnetic
resonance signals in powdered soy proteins. J Food Sci 78(5):C660–6.
Boatright WL, Jahan MS, Walters BM, MIiller AF, Cui D, Hustedt EJ, Lei Q. 2008. Carbon-
centered radicals in isolated soy proteins. J Food Sci 73(3):C222–6.
Boatright WL, Lei Q, Jahan MS. 2009. Effect of storage conditions on carbon-centered radicals
in soy protein products. J Agric Food Chem 57(17):7969–73.
Broska R, Rychly J. 2001. Double stage oxidation of polyethylene as measured by chemilumi-
nescence. Polym Degrad Stab 72(2):271–8.
Carmichael LA, Sanderson DCW. 2000. The use of acid hydrolysis for extracting minerals from
shellfish for thermoluminescence detection of irradiation. Food Chem 68(2):233–8.
Charlesby A, Partridge RH. 1963a. The thermoluminescence of irradiated polyethylene and
other polymers. Proc Roy Soc A 271:170–87.
Charlesby A, Partridge RH. 1963b. The effect of oxygen on the thermoluminescence of irradi-
ated polyethylene. Proc Roy Soc A 271:188–206.
Charlesby A, Partridge RH. 1965a. The identification of luminescence centers in polyethylene
and other polymers. Proc Roy Soc A 283:312–28.
Charlesby A, Partridge RH. 1965b. Thermoluminescence and phosphorescence in polyethylene
under ultraviolet irradiation. Proc Roy Soc A 283:329–42.
Chen R. 1977. On the analysis of thermally stimulated processes. J Electrostat 3:15–24.
Chen R, Kirsh Y. 1981. Analysis of thermally stimulated processes. Oxford: Pergamon Press.
Chen R, McKeever SWS. 1997. Theory of thermoluminescence and related phenomena. River
Edge, N.J.: World Scientific.
Cooke DW, Jahan MS. 1989. Detection of surface impurity phases in high-temperature super-
conductors using thermally stimulated luminescence. U.S. Patent No. 4, 873,444.
Corrales T, Catalina F, Peinado C, Allen NS, Fontan E. 2002. Photooxidative and thermal
degradation of polyethylenes: interrelationship by chemiluminescence, thermal gravimetric
analysis and FTIR data. J Photochem Photobiol A: Chem 147(3):213–24.
Fleming RJ. 1990. Charge trapping in organic polymers. Radiat Phy Chem 36(1):59–68.
Glidewell SM, Deighton N, Goodman BA, Hillman JR. 1993. Detection of irradiated food: a
review. J Sci Food Agric 61:281–300.
Horowitz YS. 1984. Thermoluminescence and thermoluminescent dosimetry. Chapter 1, Vol.
1. Boca Raton, Fla.: CRC Press, Inc.
Jacobson K, Costa L, Bracco P, Augustsson N, Stenberg B. 2001. Effects of microtoming on
oxidation of ultra high molecular weight polyethylene (UHMWPE). Polym Degrad Stab
73(1):141–50.
Jahan MS, Cooke DW, Gray ER, Smith JL, Mayer A, Hults WL, Bennett BL, Maez MA.
1991. Thermally stimulated luminescence from high-temperature superconducting films and
substrates. J Luminescence 48 & 49:823–7.
Kiss LI, Kispeter J. 1995. Interpretation of the thermoluminescence phenonemon in milk protein
concentrate powder. Acta Alimentaria 24(3):251–6.
Lei Q, Leibold CM, Boatright WL, Jahan MS. 2010. Distribution of stable free radicals among
amino acids of isolated soy proteins. J Food Sci 75(7):C633–40.
Mamoon AM. 1995. Thermoluminescence of irradiated milk powders. Appl Radiat Isot
46(6/7):733–4.
Mele, A, Delle Site, D, Bettinali, C, Di Domenico, A, 1968. Thermoluminescence and phase
transitions of irradiated fluorinated polymers. J Chem Phys 49(7):3297–3302.
Nagano T, Hirosuta M, Mori H, Kohyama K, Nishinar K. 1992. Dynamic viscoelastic study on
the gelation of 7S globulin from soybean. J Agric Food Chem. 40(6):941–4.
Partridge RH. 1972. Thermoluminescence in polymers in the radiation chemistry of macro-
molecules. Vol. 1. New York, N.Y.: Academic Press. p 193.
Raffi J, Yordanov ND, Chabane S, Douifi L, Gancheva V, Ivanova S. 2000. Identification of
irradiated treatment of aromatic herbs, spices and fruits by electron paramagnetic resonance
and thermoluminescence. Spectrochimica Acta Part A 56:409–16.
Riesz P, White Jr FH. 1968. Tritiated free radical scavengers in the study of the irradiated
protein molecule. In: Hart EJ, editor. Advances in chemistry series no. 81, radiation chemistry,
American Chemistry Society. Washington DC. p 496–20.
Riesz P, White Jr FH. 1970. Radical distribution in gamma-irradiated dry proteins at 195 ◦ K.
Radiat Res 44(1):24–34.
Thermally stimulated luminescence . . .
Riesz P, White Jr FH, Kon H. 1966. Free radical distribution in the gamma-irradiated ribonu-
clease. J Am Chem Soc 88:872–7.
Sanyal B, Chawla SP, Sharma A. 2009. An improved method to identify irradiated rice by EPR
spectroscopy and thermoluminescence measurements. Food Chem 116(2):526–34.
Serpi, PM, Rucci, A, Serpi, A. 1975. Thermoluminescence and isothermal decay of lumines-
cence in x-irradiated urea and thiourea. J Lumin 9(6):488–501.
Stine, CJ, Boatright WL, Lu G. 2004. Intrinsic sulfite content of isolated soy proteins. J Am
Chem Soc 81(9):829–33.
White Jr FH, Riesz P, Kon H. 1967. Free-radical distribution in several gamma-irradiated
dry proteins as determined by free radical interceptor techniques. Radiat Res 32:744–
59.
Ziegelmann B, Bogl KW, Schreiber GA. 1999. TL and ESR signals of mullusc shells—
correlations and suitability for the detection of irradiated foods. Radiat Phys Chem 54:413–
23.
C: Food Chemistry
Vol. 79, Nr. 1, 2014 r Journal of Food Science C31