Biomacromolecules 2009, 10, 11351139

1135

Plasticization of Zein: A Thermomechanical, FTIR, and

Dielectric Study
Thomas Gillgren, Susan A. Barker, Peter S. Belton, Dominique M. R. Georget, and
Mats Stading*,,
SIK, The Swedish Institute for Food and Biotechnology, PO Box 5401, SE 402 29 Gothenburg, Sweden,
School of Chemical Sciences and Pharmacy, University East Anglia, Norwich NR4 7TJ, United Kingdom,
and Department of Materials and Manufacturing Technology, Chalmers University of Technology,
SE-402 29 Gothenburg, Sweden
Received November 27, 2008; Revised Manuscript Received February 10, 2009

Zein, the main seed storage protein of maize, has been widely studied as a possible source of material for the
production of biodegradable plastic films. Plasticization of zein is critical to make functional films. While there
have been a number of publications which report the behavior of systems with a wide variety of plasticizers,
there have been few which attempt to examine the interactions of protein and plasticizer at the molecular level.
In this paper, we report on the plasticizing effects of water, glycerol, and 2-mercaptoethanol, which were examined
by a combination of spectroscopy (FTIR and dielectric) and thermomechanical methods. The results suggest that
both water and glycerol are adsorbed onto the protein and form hydrogen bonds with the amide groups. The
plasticizer then builds up in patches on the protein surface. 2-Mercaptoethanol only exhibited a weak plasticizing
effect due probably to disulfide bond breaking.

Introduction
Sustainability and the replacement of fossil hydrocarbonbased products is becoming an increasingly urgent issue.
Materials from biopolymers, such as starch, cereal proteins, and
cellulose, are seen as possible successors to materials from
conventional petroleum-based plastics. Biopolymer-based materials have the advantage that they are both renewable and
biodegradable.
Zein and gluten are the only commercially produced prolamins (cereal storage protein), but zein has the advantage that it
can be extracted from the co-products of the brewing and the
biofuel industries.1 If zein films are to be established as a
packaging alternative, plasticization is necessary since unplasticized zein materials are rather brittle. However, according to
Lawton,2 a perfect plasticizer for zein has yet to be found.
Indeed, it is not clear that any single plasticizer will be suitable
for all applications. Currently, little is known about the
molecular details of the plasticization process, and a systematic
approach to plasticization requires a great deal more information
about the molecular processes involved.
Related studies have been conducted on zein,3 kafirin,4 and
gluten,5 although the gluten study was less concerned about
plasticization than about the mechanism of hydration, and the
work on zein3 was an infrared study limited to only two levels
of hydration. A more germane study is the plasticization of the
sorghum prolamin kafirin with glycerol4 since kafirin is highly
homologous with zein. In order to understand more about the
plasticization process, comparative studies of other substances
need to be carried out.
Commercial zein, which was used in this investigation,
consists mainly of R-zein.2,6 This is the most abundant maize
* To whom correspondence should be addressed. Phone: +46-10-516
66 00. E-mail: mats.stading@sik.se.

The Swedish Institute for Food and Biotechnology.

University East Anglia.

Chalmers University of Technology.

prolamin and is about 70% of the total prolamin fraction. In

SDS-PAGE analysis, R-zein shows as two broad bands with
molar masses of 19 and 22 kDa. Glutamine and proline, as well
as alanine and leucine, are abundant in this prolamin, whereas
lysine and tryptophan are absent.7 There are various models
for the structure of R-zein. The latest by Momany et al.8
describes the molecule as being a coiled-coil of R helices with
four residues per turn in the central helical section. Nonpolar
residues form a hydrophobic surface inside a triple superhelix.
The nine helical sections that had been previously postulated7
are modeled by Momany et al.8 as three sets of three interacting
coiled-coil helices.
In this study, the plasticizing effects of water, glycerol, and
2-mercaptoethanol were investigated by a combination of dynamic mechanical analysis (DMA), Fourier transform infrared
spectroscopy (FTIR), and dielectric spectroscopy, respectively.
Glycerol is a relatively small polar molecule that has been
widely used as plasticizer and may have a similar mechanism
of plasticization to water. 2-Mercaptoethanol is less polar, but
since zein contains disulfide bonds7 which affect the internal
mobility of the protein, breaking these bonds may be an alternate
form of plasticization. 2-Mercaptoethanol, which is a reducing
agent, can fulfill this function.
No previous studies have used this combination of techniques
and plasticizers nor have they ensured that water was excluded
from the nonaqueous systems. This work therefore represents
the first truly comparative study of the functionality of plasticizers in zein.

Materials and Methods

Protein Preparation. Commercial zein (Z3625; Sigma-Aldrich,
Schnelldorf, Germany) was defatted in n-hexane as described by Oom
et al.6 The protein content of zein was 92% (w/w), determined using
the Dumas method and multiplying the nitrogen content by 6.25.
Film Casting. Zein films were cast by mixing 1.60 g (on a pure,
dry basis) of zein with 10.0 g of 70% (w/w) aqueous ethanol, glycerol

10.1021/bm801374q CCC: $40.75

2009 American Chemical Society
Published on Web 03/24/2009

1136

Biomacromolecules, Vol. 10, No. 5, 2009

(24387.292; VWR International, Fontenay sous Bois, France), or

2-mercaptoethanol (1.15433; Merck, Darmstadt, Germany) and a
magnetic stirrer in a 100 mL Erlenmeyer flask, which was sealed with
aluminum foil. Total plasticizer (glycerol and 2-mercaptoethanol,
respectively) levels were 8, 16, 32, and 40% (w/w, with respect to
protein). The zein mixture was stirred and heated at 70 C for 10 min.
During heating, a frozen cooling block was put on top of the sealing
foil to reduce the evaporation of the solvent. After heating, 5.0 g of
zein solution was put into a preheated 9 cm diameter polystyrene Petri
dish. The dish was placed, level, in a 50 C oven with a slight forced
draft for 2 h. The typical film thickness for all materials was 100 m
and was determined individually using a Mitutoyo IDC-112CD
micrometer (Mitutoyo Corporation, Kawasaki, Japan).
Conditioning and Water Sorption Measurements. Before spectroscopy analysis, the films without plasticizer were conditioned over
P2O5 and over the different saturated salt solutions. During drying and
conditioning, the films were weighed by microbalance every day until
constant weight was obtained, which normally took place within 1 week.
Equilibrium was then considered to have been reached. P2O5 gave a
dry (0% relative humidity) environment, and the saturated salt solutions
LiCl, MgCl2, NaBr, NaCl, and KCl gave the conditions of 11, 33, 58,
75, and 84% relative humidity (RH, at 25 C), respectively. The films
stored over P2O5 were assumed to be dry when equilibrium was reached.
Water contents were determined by weighing the samples stored over
the different salt solutions, thereafter storing them over P2O5 until
equilibrium was reached, and then weighing them again. Water sorption
is calculated as the weight loss divided by the initial weight and
expressed as a percentage. Plasticizer containing films were dried over
P2O5 until equilibrium was reached before spectroscopic analysis.
Samples for DMA testing were conditioned inside the instrument, due
to the difficulties in mounting fragile, low plasticizer films.
Thermomechanical Analysis. Thermomechanical properties were
measured following the method of Stading9 using DMA analysis using
a Rheometrics RSAII (Rheometrics Scientific, Piscataway, NJ). Samples
containing glycerol or 2-mercaptoethanol were maintained dry by dry
air circulation. Samples to be hydrated were dried or conditioned to
proper moisture content inside the instrument with an air moisture
controlling device before testing. The samples reacted to hydration by
expansion and to dehydration by shrinkage. The changes in sample
dimensions were measured by the instrument. When no size change
could be observed, the samples were considered to be at equilibrium.
After conditioning, the water-containing films were coated with grease
(Stabox 9415, AB Axel Christiernsen, Nol, Sweden) following the
method of Stading, which has been shown to prevent any further change
in moisture content.9 Film strips were 2 or 3 mm wide, and the original
gauge length was approximately 22.5 mm. The testing temperature
range was 30-180 C with a ramping speed of 1 C/min. Tg was
determined as the temperature of the maximum of the loss modulus
curve. Between three and eight samples were analyzed for each plasticizer content.
FTIR Measurements. The measurements were conducted as
described by Almutawah et al.5 Spectra were obtained on a Bio-Rad
FTS165 FTIR spectrometer fitted with a single-reflection diamond
attenuated total reflection (ATR) accessory (SPECAC). A mercurycadmium-telluride (MCT) detector was used. The background spectrum of the ATR cell was recorded at 64 scans and 2 cm-1 resolution.
The samples were placed on the ATR crystal, sufficient pressure was
applied to the sample to enable good contact with the ATR crystal,
and the spectra were recorded under the same conditions as the
background. Three subsamples were taken and scanned separately.
Experiments were carried out at room temperature. The infrared spectra
were analyzed by use of the Omnic software package (version 6.1a,
Thermo Nicolet Corp.). The spectra of three replicate measurements
for each sample were averaged, and a reference water vapor spectrum
was subtracted if required; this was then followed by subtraction of
the water signal from the protein region as described previously where
necessary.10 The resultant spectrum was Fourier self-deconvoluted, by

Gillgren et al.

Figure 1. Water sorption of zein films conditioned at different relative

humidities.

Figure 2. Tg of films with different water (squares), glycerol (triangles),

or 2-mercaptoethanol (diamonds) contents. Error bars indicate the
scatter as the 95% confidence interval.

using parameters: bandwidth ) 30 and enhancement ) 1.3, and

automatically baseline-corrected. The relative heights of the peaks at
1680, 1650, and 1620 cm-1 in the amide I region and at 1540 and
1515 cm-1 in the amide II region were measured.
Dielectric Spectroscopy. The measurements were conducted as
described by Almutawah et al.5 Low-frequency dielectric measurements
were carried out on a BDC-N broadband dielectric converter (Novocontrol GmbH, Germany) linked to a frequency response analyzer
(Solarton-Schlumburger, Germany). A parallel plate stainless steel
electrode system with electrodes of area 254.5 mm2 applied directly
onto the film, with a frequency range of 106-10-2 Hz and an applied
voltage of 0.5 Vrms. Dielectric spectra were obtained from three
subsamples, each measured twice separately.

Results and Discussion

Water Sorption. Conditioning the zein films at 0, 11, 33,
58, 75, and 84% RH (at 25 C) resulted in water contents of 0,
2.5, 4.5, 7.8, 9.9, and 12.3%,, respectively. The results, shown
in Figure 1, match those reported by Beck et al.11
Thermomechanical Analysis. Figure 2 displays the measured
Tg values of the films. Results show a lack of linearity in the
relationship between the plasticizer contents and the Tg values;
increasing the plasticizer content (both water and glycerol) from
zero to a low level had a greater effect on the Tg than increasing
the content from a medium to a high level had. This runs counter
to the results reported by Beck et al.,11 who reported a linear
fit between Tg and the water activity at which the film was
equilibrated. Water was a far more efficient plasticizer than
glycerol when plasticizer content was expressed in weight
percent terms. For example, it took approximately 8% water
content to reduce the Tg from 160 to 60 C, whereas it required
24% of glycerol to have the same effect. However, when the
results are expressed in mols of plasticizer per gram of protein,
the picture is somewhat different, as shown in Figure 3. In molar
terms, glycerol is a slightly more efficient plasticizer than water.

Plasticization of Zein

Biomacromolecules, Vol. 10, No. 5, 2009

Figure 3. A plot of Tg versus plasticizer content expressed in mols

per gram of protein for water (squares) and glycerol (triangles). Error
bars indicate the scatter as the 95% confidence interval.

Tg was virtually independent of added 2-mercaptoethanol

content, so it may be concluded that the only plasticizing effect
of 2-mercaptoethanol was cleavage of disulfide bonds. It appears
unlikely that 40% of any added compound would not affect
the Tg more than the small reduction observed at all levels of
addition. Therefore, it seems likely that the added material did
not remain in the film and that only a small residue remained.
2-Mercaptoethanol is known to break disulfide bonds, and it
would be expected that one of its effects in the film would be
to increase molecular mobility by decreasing inter- and intramolecular disulfide bonding. The small residue remaining in
the film was presumably sufficient to break disulfide bonds and
cause a small change in Tg. However, given the uncertainty about
the amounts in the film, no firm quantitative conclusions about
effects can be made.
Previously, the Gordon Taylor equation has been used to
parametrize the effect of plasticizing on the Tg of cereal
proteins.4,12,13 Gao et al. have expressed the equation as4

T* ) k

W2
W1

1137

Figure 4. Tg data plotted according to the Gordon Taylor equation

using the form given in eq 1. Water (squares) and glycerol (triangles).

Figure 5. Spectra obtained with dry films (A) and those plasticized
with 40% glycerol (B), 12.3% water (C), and 2-mercaptoethanol (D).
Spectra are offset for clarity.

(1)

where

Tg0 - Tg1
Tg0 - Tg2

) T*

Tg0 is the observed Tg, W is the weight fractions, and the

subscripts 1 and 2 refers to the components 1 and 2, respectively;
k is a constant reflecting the plasticizing effect of component
2. Plotting the data of the water and glycerol containing films
using the Gordon Taylor equation as expressed above (Figure
4) shows a lack of linearity. This indicated that the plasticizing
effect of both water and glycerol changed with increasing
content, and that neither glycerol nor water is an ideal plasticizer
for zein. These results for glycerol are consistent with those
observed for glycerol/kafirin systems by Gao et al.4
FTIR Measurements. Figure 5 shows the spectra obtained
with dry film and those plasticized with 12.3% water, 40%
glycerol, and 40% 2-mercaptoethanol. The amide II band in
the glycerol sample has a reduced intensity at 1515 cm-1
compared with that of the dry material. This is also true of the
water sample. Generally, the amide II band responds to
differences in the hydrogen bonding environment.5 In both cases,
the effects of plasticizer are to change the shape of the amide
II band in the same sense. Thus, presumably, both have similar

Figure 6. Ratios of the relative intensities at the amide I and the amide
II band plotted against water content: 1650/1680 (circles); 1650/1620
(diamonds); 1540/1515 (triangles).

effects on the hydrogen bonding of the amide groups in that

amide/amide interactions are reduced by an increase of amide/
plasticizer interactions.
The amide I band changes are similar in size to the amide II
changes. The dry material shows a broader peak with the main
intensity around 1650 cm-1 consistent with an R helical
structure. There are also bands at 1620 and 1680 cm-1, probably
both due to sheet-like structures.10 In the case of glycerol, the
intensity of the bands due to sheets is diminished. These results
indicate that glycerol enhances the formation of R helical forms.
Water, on the other hand, while reducing the 1680 cm-1 band
intensity, mildly enhances the 1620 cm-1 band intensity. This
may be due to differences in the detail of hydrogen bonding
patterns and other factors such as the differences of dielectric
constant of water and glycerol.
The changes in relative intensities of the bands are shown in
Figures 6 and 7. The data are too scattered to be sure that the
linear effect of water observed by Almutawah et al. in gluten
water systems5 is present here, but there is a consistent upward
trend indicating that a maximum effect, as was observed in

1138

Biomacromolecules, Vol. 10, No. 5, 2009

Figure 7. Ratios of the relative intensities at the amide I and the amide
II band plotted against glycerol content: 1650/1680 (circles); 1650/
1620 (diamonds); 1540/1515 (triangles).

gluten by Almutawah et al., has not been reached. However,

since in gluten this effect was observed at around the 35% water
level, the data reported here for zein are consistent with the
gluten data.
Although there is some scatter, it is clear that the effect of
glycerol and water is equally strong on the amide I band
structure and the amide II band. This is not consistent with other
studies4,5 which have shown a primary effect of plasticizer on
hydrogen bonding rather than overall conformation. However,
the changes in the amide II bands reported by Gao et al. for
glycerol/kafirin are considerably larger than those reported here
for glycerol/zein and imply that the interaction of the plasticizers
with zein, in terms of hydrogen bonding, is rather weaker than
that with kafirin.
Notwithstanding this difference on degree of behavior, the
trends may be explained by a model which explains similar
infrared results observed in water gluten systems.5 In this, the
plasticizing molecule initially solvates the amide groups in the
protein, and only after complete solvation of the amide groups
does bulk plasticizer appear in the system. This point is indicted
by a reduction in the rate of change of the amide II intensity
ratios and in gluten happens at about 30% water content. The
data presented here show no evidence of such an effect for the
glycerol and water samples in this study.
In comparison to the effects of water and glycerol, 2-mercaptoethanol showed even smaller effect on conformation with
no significant changes in the spectra with concentration. This
is consistent with a small effect due to the breaking of a few
disulfide linkages but no major effect either on conformation
or hydrogen bonding. As discussed in the Thermomechanical
Analysis section, the reason for the lack of effect on conformation or hydrogen bonding is probably a lack of plasticizer.
During film formation, a small portion of the 2-mercaptoethanol
reacted with the protein by breaking the disulfide linkages, but
most of the excess plasticizer evaporated.
Dielectric Spectroscopy. The dielectric response of the films
containing water and glycerol are shown in Figures 8-11. The
dry material showed only a weak dependence of dielectric
constant on frequency for both the real and imaginary parts,
although there is some evidence of a weak dispersion in the
imaginary part of the spectrum. At low levels of water below
7.8%, both the real and imaginary parts remain separate,
indicating that the cross over point of the low frequency
dispersion has not been reached. At this point, the real and
imaginary parts become equal and the ratio of the imaginary to
real part changes from less than one to greater than one. Both
parts thereafter increase in parallel fashion.14 Although a low
frequency dispersion may exist in samples where the cross over
is not observed, its existence cannot be certain if no such point

Gillgren et al.

Figure 8. Real (filled) and imaginary (empty) dielectric responses of

the dry films (diamonds), and the 2.5% (squares) and 4.5% (triangles)
water containing films.

Figure 9. Real (filled) and imaginary (empty) dielectric responses of

the 7.8% (squares), 9.9% (circles), and the 12.3% (triangles) water
containing films.

Figure 10. Real (filled) and imaginary (empty) dielectric response of

the films containing 0% (diamonds), 8% (squares), and 16% (triangles) glycerol.

is observed. Thus, we limit our consideration of the low

frequency dispersion to those data where such a cross over is
observed. In the water samples, a clear low frequency dispersion
is seen in the 12.3% sample and a crossover occurs in the 9.9%
sample about 1 Hz. The 7.8% sample has an indication of the
cross over at about 10-2 Hz. In the 12.3% sample, there is an
indication at very low frequency of a change in the slope of the
real and imaginary parts that may indicate a further dispersion.
Similar behavior is observed in the glycerol samples. With up
to 16% glycerol, no evidence of a low frequency dispersion is
seen. Such a dispersion is apparent above this level and can be
most clearly seen in the 40% sample.
The low frequency dispersion is attributed to collective
motions of hydrogen-bonded clusters. The dielectric response

Plasticization of Zein

Biomacromolecules, Vol. 10, No. 5, 2009

Figure 11. Real (filled) and imaginary (empty) dielectric response of

the films containing 24% (squares) and 40% (triangles) glycerol.

is not easily related to such simple parameters as cluster size

but to variables such as the number of molecules taking part in
collective motions, the frequency of site motions within the
cluster, as well as intercluster exchange.14 It would be expected
that as the amount of plasticizer was increased the size of these
clusters would increase, but accompanying this are a number
of other changes, including protein conformation, internal
protein mobility, and the mobility of the plasticizer molecules.
The sum of these effects is to increase the frequency of the
local modes of motion and possibly, as a result, decrease the
correlation length of the collective motions. Both of these would
result in an increase in the frequency at which the cross over
takes place.
The behavior observed for both water and glycerol is therefore
of initially isolated molecules being absorbed on the surface of
the proteins at the highest energy absorption sites. These do
not show a low frequency dispersion. More plasticizer then
concentrates at these sites to give clusters which continue to
build up in size as more plasticizer is added.
There is no evidence of Maxwell-Wagner dispersion,5 which
is indicative of the existence of bulk solvent, thus the dielectric
observations are constant with the infrared data.

Conclusion
Thermomechanical data fitted into the Gordon Taylor equation shows that neither water nor glycerol is an ideal plasticizer
for zein. FTIR and dielectric spectroscopy results show that the
plasticizing effect of glycerol and water are similar. Neither
water nor glycerol has any greater effect on the conformation
of the protein, and at the concentrations used, no evidence for
the presence of bulk solvent was found. A model consistent

1139

with the data is that separate plasticizer molecules are first

adsorbed on the surface of the protein. Thereafter, clusters grow
around these sites as more plasticizer molecules bind to the
protein. The lack of bulk plasticizer indicates that the binding
sites are not fully occupied at the maximum concentrations of
plasticizer used in these experiments. Kafirin and zein are highly
homologous. However, kafirin is generally recognized as being
more hydrophobic than zein as it is most readily extracted into
tertiary butanol water rather than ethanol-water, and the
difference in the effects of glycerol may be due to hydrophobic
interactions rather than hydrogen bonding interactions. The
2-mercaptoethanol results were unexpected and fell outside the
scope of this investigation and should therefore be considered
with great caution. It seems though as if 2-mercaptoethanol has
an effect by breaking disulfide bonds, which has an insignificant
effect on the conformation.
Acknowledgment. This study has been carried out with
financial support from the Commission of the European Communities, Framework 6, Priority 5 Food Quality and Safety,
Integrated Project NovelQ FP6-CT-2006-015710. The Swedish
Research Council for Environment, Agricultural Sciences and
Spatial Planning, FORMAS, are also gratefully acknowledged
for financial support.

References and Notes

(1) Core, J. Agric. Res. 2002, 2021.
(2) Lawton, J. W. Cereal Chem. 2002, 79, 118.
(3) Mizutani, Y.; Matsumura, Y.; Imamura, K.; Nakanishi, K.; Mori, T.
J. Agric. Food Chem. 2003, 51, 229235.
(4) Gao, C.; Stading, M.; Wellner, N.; Parker, M. L.; Noel, T. R.; Mills,
E. N. C.; Belton, P. S. J. Agric. Food Chem. 2006, 54, 46114616.
(5) Almutawah, A.; Barker, S. A.; Belton, P. S. Biomacromolecules 2007,
8, 16011606.
(6) Oom, A.; Pettersson, A.; Taylor, J. R. N.; Stading, M. J. Cereal Sci.
2008, 47, 109116.
(7) Coleman, C. E.; Larkins, B. A. The Prolamins of Maize. In Seed
Proteins; Shewry, P. R., Casey, R., Eds.; Kluwer Academic Publishers:
Dordrecht, The Netherlands, 1999; pp 109-139.
(8) Momany, F. A.; Sessa, D. J.; Lawton, J. W.; Selling, G. W.; Hamaker,
S. A. H.; Willett, J. L. J. Agric. Food Chem. 2006, 54, 543547.
(9) Stading, M. Trans. Nord. Rheol. Soc. 1998, 6, 147150.
(10) Georget, D. M. R.; Belton, P. S. Biomacromolecules 2006, 7, 469
475.
(11) Beck, M. I.; Tomka, I.; Waysek, E. Int. J. Pharm. 1996, 141, 137
150.
(12) Micard, V.; Guilbert, S. Int. J. Biol. Macromol. 2000, 27, 229236.
(13) Barreto, P. L. M.; Roeder, J.; Crespo, J. S.; Maciel, G. R.; Terenzi,
H.; Pires, A. T. N.; Soldi, V. Food Chem. 2003, 82, 425431.
(14) Dissado, L. A.; Hill, R. M. J. Chem. Soc., Faraday Trans. 2 1984,
80, 291319.

BM801374Q