ARTICLE pubs.acs.org/JAFC
Antioxidants Modulate Molecular Mobility, Oxygen Permeability, and Microstructure in Zein Films Jun Liang and Richard D. Ludescher* Department of Food Science, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08901, United States ABSTRACT: The effect of octyl gallate and propyl gallate on the molecular mobility, oxygen permeability, and microstructure of zein/glycerol films was studied. Films were cast from 70% ethanol/water containing 20% (w/w) glycerol and different amounts of the antioxidants propyl gallate or octyl gallate. The oxygen permeability and local mobility of these films were measured using phosphorescence from the dispersed triplet probe erythrosin B. Although both antioxidants increased the local mobility of the zein matrix to about the same extent, octyl gallate and to a lesser extent propyl gallate dramatically increased the permeability of the film to oxygen. Atomic force microscopy imaging indicated that propyl gallate induced aggregation of zein complexes, which could lead to a more condensed film. These results indicate that the addition of specific functional ingredients, such as antioxidants, to edible films may significantly affect the physical properties and structure and, thus, functional properties in ways that influence their eventual use. KEYWORDS: Zein, antioxidant, film, octyl gallate, propyl gallate, mobility, oxygen diffusion, microstructure
’ INTRODUCTION An increased consumer demand for fresh, high-quality food products has focused interest on the properties of active packaging materials that change the conditions of the packaged food to extend shelf life, to improve sensory properties, or to inhibit the growth of pathogenic and spoilage microorganisms.1,2 Active packaging involves interactions between the package material and the food or headspace; this usually involves the incorporation of specific additives in the packaging polymer-based film or within the container itself.3,4 Antioxidant packaging is a major category of active packaging and a very promising tool for extending food product shelf life.58 Edible films containing or coated with antioxidant could provide continuous protection for foods against oxidation during storage or after opening while also providing a recyclable package material.911 Proteins are an important source of edible filmforming materials; upon drying from solution, proteins readily form a continuous amorphous solid matrix with appropriate mechanical and barrier properties.12 Despite extensive research about the potential usage and performance of antioxidants in edible films,58 data about the influence of antioxidants on the mobility of the matrix and oxygen permeability in the films have been rarely reported. The transfer of oxygen to food from the environment, which is the primary determinant of oxidation in the food system, has an important effect on food quality. Edible films and coatings can prevent deterioration for many food products because they often possess excellent oxygen barrier properties. Thus, the relation between the addition of the antioxidant and any oxygen permeability change in films is of immediate concern. As small molecule ingredients, antioxidants could act as a plasticizer in films.13 Polar and hydrophobic groups of the antioxidant could influence the formation of the hydrogen-bond network, which could directly influence the molecular mobility of the matrix. Although previous research in our lab has shown that oxygen permeability in some protein films increases r 2011 American Chemical Society
proportionally with an increase in the mobility of the matrix,12,14 the extent to which oxygen permeability is modulated by mobility is still unclear. Here, we investigate the mechanism(s) involved in the relation. Zein, the prolamine protein of maize seeds, has been extensively investigated as a potential commercial material for edible packaging because of its thermoplastic properties and excellent filmforming behavior.1518 Zein, a relatively hydrophobic protein soluble in aqueous alcohol, is the most abundant protein in corn, acting as a reservoir of amino acids for the growing embryo. Zein forms large aggregates in aqueous ethanol,19,20 and the aggregation number does not appear to change during the evaporation of the solvent.21 Research has shown that zein films cast from ethanol solutions showed hydrophilic surfaces and those cast from acetone showed hydrophobic surfaces,22 while films cast from acetic acid were much smoother and more compacted than films cast from ethanol;23 this is apparently due to differences in the degree of aggregation of zein in solution. Changes in the aggregate microstructure as well as interactions with added compounds may have important influences on the properties of zein films. However, the full details of these interactions are waiting to be elucidated. Propyl gallate and octyl gallate, two important antioxidants widely used in keeping foods fresh,2426 possess identical hydrophilic head groups and differ only in the length of their hydrocarbon chain. The trihydroxy benzoyl headgroup and the hydrocarbon chain will interact differently with the hydrogenbond network in the protein film, although both groups should contribute to any plasticizing effects. Gallate can hydrogen bond with polar groups in the protein, thus interrupting the hydrogenbond network among those groups, while the saturated carbon chains could hinder the formation of hydrogen bonds by steric Received: June 27, 2011 Revised: September 21, 2011 Accepted: November 7, 2011 Published: November 07, 2011 13173
dx.doi.org/10.1021/jf202554u | J. Agric. Food Chem. 2011, 59, 13173–13180
Journal of Agricultural and Food Chemistry hindrance. The different lengths of the propy and octyl chains may thus lead to distinct behavior in the films. Oxygen permeability in zein films has been studied recently by experimental techniques that include detecting the oxygen permeability coefficient using a gas permeability tester,27 monitoring the oxygen partial pressure by gas chromatography,28 and using the equal pressure method.29 In this study, we have used phosphorescence of erythrosin B (Ery B), a triplet state molecular probe, to monitor the local molecular properties and mobility of amorphous zein/glycerol films cast from 70% ethanol aqueous solution. Measurements of the Ery B emission energy and excited-state lifetime provide information about how propyl gallate and octyl gallate influence the mobility of the matrix; comparisons of the phosphorescence lifetime in the presence and absence of air monitor the oxygen permeability. Phosphorescence probe techniques have been shown to provide detailed information about interrelations among local molecular mobility, dynamic heterogeneity, and oxygen permeability in amorphous solid biomaterials.12,14,30 Microstructural changes in the zein film were also determined by imaging with atomic force microscopy (AFM). This combination of techniques provides insight into the molecular mechanisms by which antioxidants modulate the physical and functional properties of edible polymer barriers; such insights should improve our ability to engineer edible food films with appropriate functionalities.
’ MATERIALS AND METHODS Reagents. Zein (regular grade), propyl gallate (g99% purity), octyl gallate (g99% purity), and Ery B (sodium salt) were purchased from Sigma Chemical (St. Louis, MO). Zein, provided to Sigma by Freeman Industries, LLC, is 9096% pure. Unpublished studies from our lab indicate that further purification to remove residual oil does not affect the mobility of dry zein films. Ethanol [high-performance liquid chromatography (HPLC) grade] was purchased from Fisher Scientific (Fair Lawn, NJ). Sample Preparation. We prepared amorphous zein films using a slightly modified version of our published method for preparing dried protein films on quartz slides.30 Zein and glycerol were added to 70% (v/v) ethanol in water solution to final concentrations of 0.4% (w/v) for zein and 0.1% (w/v) for glycerol. Ery B was dissolved in deionized water to prepare 10 mM probe stock solutions; aliquots from probe stock solution were added to the zein solution to make sample solutions with a probe/zein mole ratio of about 3:20 (assuming a molecular weight of 2 104 g/mol for zein). Propyl gallate and octyl gallate were dissolved in ethanol at 50 mM as stock solutions. Aliquots from the antioxidant stock solutions were added to zein solution to obtain antioxidant/zein mole ratios of about 2.5, 10, and 20. The amount of antioxidant in the films thus ranged from 2.1 to 14.4 wt % for propyl gallate and from 2.7 to 18.3 wt % for octyl gallate, and the total amount of additives in the films ranged from 20.4 wt % (no antioxidant) to 35.0 wt % (octyl gallate at a 20:1 mol ratio to zein). To make zein films for measurements of luminescence, 15 μL of protein/probe/antioxidant/glycerol solution at room temperature was spread on approximately one-third of a quartz slide (30 13.5 0.6 mm, custom-made by NSG Precision Cells, Farmingdale, NY). Before use, the slides were soaked in Terg-A-Zyme (Alconox, Inc., White Plains, NY) soap solution for >24 h to remove surface impurities, washed with deionized water, rinsed with ethanol, and dried with acetone. The sample slides were stored at room temperature against the desiccants DrieRite and P2O5, to maintain 0% relative humidity, for at least 1 week and protected from light to prevent any photobleaching of Ery B and antioxidants. The desiccants were refreshed as necessary.
ARTICLE
For AFM experiments, films were prepared using the same method used to prepare slides for luminescence, except that the sample solutions were spread on mica chips (around 0.5 0.5 cm). Phosphorescence Measurements and Analysis. All measurements were conducted on a Cary Eclipse spectrophotometer (Varian Instruments, Walnut Creek, CA) equipped with a temperature controller and multicell holder. All measurements were made at least in triplicate. The slides were first heated, and then data were collected from 100 to 10 °C in all experiments. High-purity nitrogen (minus O2) or dry air (plus O2) flowed directly into capped quartz fluorescence cuvettes holding the slides. The cuvette was flushed for at least 30 min at 100 °C before each set of measurements. The cuvette was capped with a lid having inlet and outlet ports for the gas line; therefore, all experiments were performed at a constant pressure. Each phosphorescence intensity decay was an average of 50 cycles. For each cycle, data were collected from a single lamp flash with a delay of 0.04 ms, a 0.05 ms gate, and 10.0 ms total decay time. Phosphorescence and delayed fluorescence emission scans were collected over the range from 540 to 800 nm with an excitation wavelength of 520 nm. The excitation and emission monochromators were both set at 20 nm band pass. For phosphorescence emission scans, each data point (collected at 1 nm intervals with a 0.1 s averaging time) was collected from a single flash with 0.2 ms delay and 5 ms gate time. For lifetime measurements, because intensity decays were nonexponential, a stretched exponential function was selected to analyze the intensity decays IðtÞ ¼ Ið0Þexp½ðt=τÞβ þ constant
ð1Þ
where I(0) is the initial intensity, τ is the stretched exponential lifetime, and β is an exponent varying from 0 to 1 that characterizes the lifetime distribution. The use of a stretched exponential model provides an analysis in terms of a continuous distribution of lifetimes, which is appropriate for describing a complex glass possessing a distribution of relaxation times for dynamic molecular processes. The smaller the β value, the more non-exponential the intensity decays and the broader the distribution of lifetimes.30 The energy of the emission maximum (νp) and the full width at halfmaximum (fwhm) of the emission band were determined using a lognormal line shape function 8 !2 9 < ln½1 þ 2bðν νp Þ=Δ = IðνÞ ¼ I0 exp lnð2Þ ð2Þ : ; b where I0 is the maximum emission intensity, νp is the peak frequency (cm1), Δ is a line width parameter, and b is an asymmetry parameter. The bandwidth Γ (fwhm) is calculated according to the following equation: sinhðbÞ ð3Þ Γ¼Δ b For delayed luminescence spectra collected from 540 to 750 nm, a sum of two log-normal functions with independent parameters was used to separately fit the delayed fluorescence [IDF(ν)] and phosphorescence [IP(ν)] bands. The phosphorescence lifetimes were used to calculate the rate constants associated with the various processes that depopulate the triplet state. Our analysis of the delayed emission is similar to the photophysical scheme for Ery B outlined by Duchowicz and co-workers using slightly different nomenclature.31 The measured phosphorescence lifetime (τ) is the inverse sum of all possible de-excitation rates for the triplet state T1 1=τ ¼ kP ¼ kRP þ kTS1 þ kTS0 þ kQ ½O2
ð4Þ
where kRP is the rate of radiative decay to the ground state, kTS1 is the rate of reverse intersystem crossing to S1, kTS0 is the rate of intersystem 13174
dx.doi.org/10.1021/jf202554u |J. Agric. Food Chem. 2011, 59, 13173–13180
Journal of Agricultural and Food Chemistry
ARTICLE
crossing to the singlet manifold followed by vibrational relaxation to S0, and kQ[O2] is rate of oxygen quenching (assumed negligible in the absence of oxygen). The radiative decay rate has a value of 41 s1 for Ery B.32 Reverse intersystem crossing is a thermally activated process that has an exponential dependence upon the energy gap ΔETS between T1 and S1.33 kTS1 ðTÞ ¼ kO TS1 expðΔETS =RTÞ
ð5Þ
The ratio of intensity of delay fluorescence (IDF) to phosphorescence (IP), where IDF and IP are determined from analysis of emission spectra using the log-normal function (eq 2), is proportional to the rate of reverse intersystem crossing. A plot of ln(IDF/IP) versus 1/T thus has a slope of ΔETS/R. 7 1 in Values of kO TS1 for Ery B reported in the literature vary from 0.3 10 s 7 1 31 7 1 ethanol and 6.5 10 s in water to 111 10 s in solid polyvinyl alcohol.34 We estimated the maximum possible value for kO TS1 in zein with 20% glycerol by assuming that kTS1(T) cannot result in values for kTS0 that decrease with the temperature.35 This procedure thus estimated the 7 1 minimum possible values of kTS0(T). We used a value of kO TS1 = 1 10 s in this research and the measured value of ΔETS to calculate kTS1(T). In the presence of oxygen, the quenching rate kQ[Q] is the product of the rate constant kQ and the oxygen concentration [O2]. For measurements performed while flushing nitrogen over the slides, we assume that no oxygen quenching occurred ([O2] = 0). The value of kQ[Q] can thus
Figure 1. Natural log of the ratio of peak intensity for delayed fluorescence (IDF) to phosphorescence (IP) [ln(IDF/IP)] for Ery B in zein/glycerol films containing octyl gallate with varying mole ratios to zein plotted as a function of the inverse temperature. The inset shows the values for kTS1(T) in these films calculated as described in the text.
be calculated directly using eq 4 by taking the difference between the inverse lifetimes in the presence (air) and absence (N2) of oxygen. A major nonradiative decay route is through intersystem crossing to the ground S0 state. The decay rate is expressed by kTS0, which reflects the rate of quenching of the probe as a result of coupling of the excited T1 state to a highly excited vibration of the ground S0 state followed by dissipation of vibrational energy from the probe into the matrix. This value of kTS0 is thus a measure of matrix molecular mobility.33,35 The temperature dependence of kTS0 can be calculated from measurements of the Ery B lifetime under nitrogen using eq 4 because kRP is known and kTS1(T) can be calculated as described above (eq 5). AFM. Tapping mode AFM images were collected using a NanoScope IIIA Multimode AFM (Veeco Instruments, Inc., Santa Barbara, CA) equipped with a silicon-etched RTESP7 cantilever (Veeco Nanoprobe, Camarillo, CA) under ambient conditions. Before tip engagement, the drive frequency of the silicon tip was tuned with the aid of Nanoscope 5.30 software and fixed at 200250 kHz for further scanning. All of the collected images were flattened before further analysis.
’ RESULTS AND DISCUSSION Delayed Emission Spectra. Because pure zein dry films are very brittle, a plasticizer is often added to improve their mechanical and film properties; zein films were thus made with 20% by weight glycerol as a plasticizer for this research. The delayed emission spectra of Ery B in amorphous zein/glycerol films displayed two distinct bands: a longer wavelength phosphorescence band (maximum of ∼690 nm) because of emission from the triplet T1 state and a shorter wavelength delayed fluorescence band (maximum of ∼555 nm) because of emission from the singlet S1 state repopulated by thermally stimulated reverse intersystem crossing from T1. Delayed emission spectra of zein/glycerol films with various antioxidant ratios collected over the temperature range from 100 to 10 °C showed a decrease in phosphorescence (IP) and an increase in delayed fluorescence (IDF) intensity with an increasing temperature (data not shown). The intensity ratio was analyzed as a van’t Hoff plot of ln(IDF/IP) versus 1/T, and the linear slope was used to estimate the energy gap (ΔETS) between the triplet and singlet states. van’t Hoff plots of ln(IDF/IP) versus 1/T were linear with R2 g 0.995; the slopes for samples with variable amounts of propyl gallate or octyl gallate were essentially identical, as illustrated in Figure 1 for octyl gallate. The tripletsinglet energy gap of Ery B is typically influenced by the surrounding solvent (matrix). The slope of ln(IDF/IP)
Figure 2. Effect of the temperature on the peak frequency for phosphorescence emission from Ery B dispersed in zein/glycerol films containing either octyl gallate (O) or propyl gallate (P) with varying mole ratios to zein. 13175
dx.doi.org/10.1021/jf202554u |J. Agric. Food Chem. 2011, 59, 13173–13180
Journal of Agricultural and Food Chemistry
ARTICLE
Figure 3. Effect of the temperature on the bandwidth (fwhm) for phosphorescence emission from Ery B dispersed in zein/glycerol film containing octyl gallate (O) or propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20.
Figure 4. Effect of the temperature on the stretching exponent (β) obtained from fits to a stretched exponential decay model of the intensity decay of Ery B dispersed in zein/glycerol film containing octyl gallate (O) or propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20; samples purged with N2 (minus O2).
versus 1/T showed negligible differences for all samples tested, indicating that kTS1(T) was not significantly affected by the addition of octyl gallate or propyl gallate. The inset in Figure 1 shows the calculated variartion of kTS1(T) over the temperature range of this study. The peak emission frequency (νP) and bandwidth (Γ) for both delayed fluorescence and phosphorescence emission were determined by fitting emission spectra to a log-normal line shape function (eqs 2 and 3). The values of νP for phosphorescence emission are plotted in Figure 2 for films containing either octyl gallate or propyl gallate. The peak frequency of all samples showed a slight but significant decrease with the temperature in both the absence (control) and presence of the antioxidant. This decrease in emission energy reflects an increase in the average extent of matrix dipolar relaxation around the excited triplet state prior to emission.35 The phosphorescence bandwidth increased gradually at low temperatures and more steeply at high temperatures in all films (Figure 3). The addition of either propyl gallate or octyl gallate decreased the bandwidth over the whole temperature range, causing a significant decrease in the case of octyl gallate at temperatures above 60 °C. The phosphorescence bandwidth reflects the extent of inhomogeneous broadening because of a range of energetic interactions between the matrix and the excited probe;36 it thus reflects the local molecular structure and interactions of the matrix and not large-scale phenomenon, such as phase separation. A decrease in bandwidth thus reflects a
corresponding decrease in the width of the distribution of energetically distinct matrix environments in the presence of the antioxidant over the whole temperature range. Phosphorescence Emission Lifetimes. The phosphorescence emission intensity decay transients from Ery B in films of zein/ glycerol or zein/glycerol/antioxidant under purging with either nitrogen or dry air were collected as a function of the temperature over the range from 100 to 10 °C. All intensity decay transients were well-fit using a stretched exponential decay model (eq 1), in which the lifetime τ and stretching exponent β were the physically relevant fitting parameters; R2 was g0.995 for all fits. The stretching exponent β of all samples in nitrogen decreased slightly with the temperature (Figure 4), indicating that the distribution of lifetimes in the phosphorescence decay broadened with an increase in the temperature. At each temperature, β decreased with an increase in either the octyl gallate concentration or the propyl gallate concentration, indicating that the matrix was more dynamically heterogeneous in the presence of these additives. The phosphorescence lifetime under N2 (minus O2) decreased monotonically with an increasing temperature; the decrease was gradual at low temperatures and steeper at higher temperatures, indicating that the nonradiative decay rates were thermally activated (Figure 5). The lifetime at each temperature decreased with an increasing concentration of either octyl gallate or propyl gallate, indicating that both antioxidants increased the average matrix mobility of the zein film. At the same concentration, 13176
dx.doi.org/10.1021/jf202554u |J. Agric. Food Chem. 2011, 59, 13173–13180
Journal of Agricultural and Food Chemistry
ARTICLE
Figure 5. Effect of the temperature on the lifetimes obtained from fits to a stretched exponential decay model of the intensity decay of Ery B dispersed in zein/glycerol film containing octyl gallate (O) or propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20; samples purged with N2 (minus O2).
Figure 6. Effect of the temperature on the lifetime obtained from fits to a stretched exponential decay model of the intensity decay of Ery B dispersed in zein/glycerol film containing octyl gallate (O) and propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20; samples purged with air (plus O2).
Figure 7. Temperature dependence of the oxygen quenching rate kQ[O2] for Ery B in zein/glycerol film containing octyl gallate (O) or propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20.
the lifetime in films with octyl gallate was comparable to that with propyl gallate, indicating that hydrocarbon tails of different length in these antioxidants did not significantly influence the nonradiative decay rates. To quantify the influence of octyl gallate and propyl gallate on oxygen permeability in zein/glycerol films, we measured the phosphorescence decay of Ery B under dry air. Because the oxygen
quenching constant kQ[O2] is the product of terms proportional to both the rate of oxygen diffusion (kQ) and the thermodynamics of oxygen solubility ([O2]), it reflects the permeability of the zein matrix to oxygen. The lifetimes under dry air in films with octyl gallate and propyl gallate are plotted as a function of the temperature in Figure 6. The phosphorescence lifetime under air (plus O2) was modulated by the presence of antioxidant in a 13177
dx.doi.org/10.1021/jf202554u |J. Agric. Food Chem. 2011, 59, 13173–13180
Journal of Agricultural and Food Chemistry
ARTICLE
Figure 8. Effect of the temperature on the rate constant for nonradiative decay (kTS0) of the triplet T1 state to the ground S0 singlet state; data were calculated from the lifetime data of Figure 5.
Figure 9. Dependence of kQ[O2] upon kTS0 for pure zein/glycerol films and for films with octyl gallate or propyl gallate at an antioxidant/zein mole ratio of 20:1.
dose-dependent fashion, with octyl gallate having a significantly larger effect than propyl gallate. The value of kQ[O2] as a function of the temperature was calculated from the lifetimes in air and nitrogen using eq 4; these data are plotted versus the temperature in Figure 7. The value of kQ[O2] increased gradually at low temperatures and more steeply at higher temperatures for all samples. Propyl gallate had no significant effect on kQ[O2] at 20 °C and below and only a moderate effect at higher temperatures; octyl gallate, however, significantly increased kQ[O2] at all temperatures from 0 °C and above. In the absence of oxygen, the lifetime of Ery B reflects the rate constants for radiative emission, kRP, reverse intersystem crossing, kTS1, and matrix quenching, kTS0 (eq 4). The value of kRP is 41 s1 and constant.30 Because the magnitude of kTS1 can be estimated as described above, it is possible to calculate kTS0(T) from the lifetime data under N2 using eq 4. The calculated values of kTS0(T) for Ery B in zein/glycerol and zein/glycerol/antioxidant films are plotted as ln(kTS0) versus the temperature in Figure 8. The magnitude of these values indicated that the temperature-dependent increase in kTS0 had the largest influence on the Ery B lifetime measured under N2. Both octyl gallate and propyl gallate significantly increased kTS0 and, thus, increased the local mobility of the zein film, although the effect of octyl gallate was more pronounced. As suggested by the analysis below, this increase in matrix mobility directly influenced the oxygen permeability.
Figure 10. Tapping mode AFM height images of zein/glycerol films containing octyl gallate (O) or propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20.
An approximately linear dependence of the oxygen quenching rate (kQ[O2]) upon the matrix quenching rate (kTS0) was seen in pure zein/glycerol films and in films containing either octyl gallate or propyl gallate. The linear dependence of kQ[O2] upon kTS0 for zein/glycerol films as well as for films containing octyl gallate or propyl gallate at a ratio of 20:1 antioxidant/zein is shown in Figure 9. There is clearly a steeper dependence of kQ[O2] upon kTS0 for the film containing octyl gallate. AFM. Early studies of the zein molecular structure estimated an α-helix content between 50 and 60%.37 Wang and Padua19 have concluded on the basis of surface plasmon resonance studies that the zein molecule contains sharply defined hydrophobic and hydrophilic domains on its surface. Interactions of these domains drive zein complex formation. Kim and Xu proposed that zein could form micelle-like structures in ethanol/ water solution with hydrophilic or hydrophobic particles modulated by the content of ethanol.21 The zein complex is sufficiently stable that only the strong detergent sodium dodecyl sulfate (SDS) can dissociate it.38 For this study, films for AFM were cast on a mica slide about 0.5 0.5 cm. Considering the zein concentration in the ethanol solution used to cast films and assuming a partial specific volume of 1.37 cm3/g,39 the protein thickness on the slides was ∼4000 nm. The films thus included many layers of zein protein. Because AFM only provides very local images of the surface of a film (10 μm square in this study), sample homogeneity is very important. Before measurement, each sample chip was carefully 13178
dx.doi.org/10.1021/jf202554u |J. Agric. Food Chem. 2011, 59, 13173–13180
Journal of Agricultural and Food Chemistry
Figure 11. Tapping mode AFM phase images of zein/glycerol films containing octyl gallate (O) or propyl gallate (P) at antioxidant/zein mole ratios of 2.5, 10, and 20.
checked under the microscope associated with the AFM to ensure no crystallization or phase separation occurred in the glycerol-plasticized films. Because previous research has shown that zein films containing varying levels of glycerol are homogeneous throughout the cooling and heating cycles over the temperature range from 100 to 150 °C,40 we are confident that the AFM images shown represent the microstructure of the entire films. AFM images for height and phase of the surface microstructure of zein/glycerol films in the absence and presence of antioxidants are shown in Figures 10 and 11, respectively. The control sample (without antioxidant) showed a relatively flat surface with generally uniform particles and little indication of particle aggregation. However, images of films with either propyl or octyl gallate showed a range of zein globules of different size that were, in general, larger than those seen in the absence of the antioxidant. At similar ratios of antioxidant/zein, films containing propyl gallate appeared to have larger aggregates than those containing octyl gallate. The aggregation of the zein complex thus appears to be modulated by the presence of the antioxidant. Kim and coworkers21 proposed that the zein complex in 70% ethanol/water solution exposes a hydrophilic surface and that the complex does not change during the evaporation of the solvent. Given the low concentrations of antioxidant in the solutions used to cast films (
