ZEOLITE/CHITOSAN/GELATIN FILMS: PREPARATION, SUPERCRITICAL CO 2 PROCESSING, CHARACTERIZATION AND BIOACTIVITY

Chitosan/gelatin and chitosan/gelatin/zeolite films prepared by solvent casting method were impregnated with a mixture of thymol and carvacrol using a green solvent, supercritical carbon dioxide at 35 °C and 30 MPa, during 18 h. Proposed method enabled preparation of biocompatible and biodegradable blends with strong antioxidant and antibacterial activity, whereby amounts of loaded thymol/carvacrol mixture were in the range from 3.3-6%. After initial burst release, both types of films exhibited gradual release of bioactive compounds, with around 72 and 96% of impregnated thymol/carvacrol mixture released in water and PBS (pH 7.4) during tested period of 10 days, respectively. Results of water vapor transmission rate (> 76 gm -2 day -1 ) confirmed that prepared composites are suitable for wound dressing application. Thermal analysis showed superior properties of prepared thymol/carvacrol loaded films compared to control samples. In addition, mechanical and structural properties, as well as solubility and swelling behavior of the obtained films were investigated in detail.


INTRODUCTION
Plants produce a variety of compounds that have significant role in its protection against different pathogens [1]. Some of these natural compounds, that exhibit antimicrobial properties, are usually present in plant essential oils as a mixture of volatile secondary metabolites [2]. Among them, phenols and polyphenols are the most studied as they also express antimicrobial activity against human pathogens [3,4]. Two major phenolic compounds of oregano (Origanum vulgare) and thyme (Thymus vulgaris) oils are thymol  and carvacrol . Both thymol and carvacrol have been used in traditional medicine since ancient times [5]. They possess a variety of biological properties such as antimicrobial, antioxidant, antiinflammatory, antitumor, anti-mutagen, analgesic, anti-parasitic and insecticidal [5,6]. The susceptibility of methicillin resistant Staphylococcus aureus (MRSA) to carvacrol and thymol was also proven [7]. Furthermore, it was reported that thymol and carvacrol in a mixture exhibit synergetic activity enabling desired antimicrobial and antioxidant activity at lower concentrations compared to their individual performances [8,9]. In addition, United States Food and Drug Administration (FDA) has generally recognized thymol and carvacrol as safe [10]. In order to preserve these active compounds from evaporation and environmental effects and to enable their gradual release at This article is protected by copyright. All rights reserved. specific site, loading of thymol/carvacrol mixture into biocompatible and biodegradable polymers was explored in this study.
Loading of active substances into polymer carriers in most cases implies utilization of aqueous or organic solutions, depending on an active substance and polymer solubility, solvent toxicity and capability of solvent to be removed [11]. Despite simplicity of conventional methods for incorporation of desirable compounds into polymer matrixes, some disadvantages must be pointed out such as utilization of toxic organic solvents in some cases, undesired reactions, photochemical and thermal degradation, low incorporation, non-unique dispersion of an active substance, additional step of material drying, solvent residue in final material, etc. [11]. On the other hand, application of supercritical solvent impregnation (SSI) as a technique for loading of desirable compounds into polymer network offers a possibility to overcome aforementioned challenges [12,13]. Supercritical carbon dioxide (scCO 2 ) is the green solvent most commonly used in SSI process as it provides high diffusivity into organic matter due to low viscosity and near zero surface [14,15].
In addition, this process enables final product, without solvent residues, obtained by simple depressurization of the system [13].
However, there is only one report in available literature that describes utilization of SSI process for loading of thymol/carvacrol mixture into poly(D,L-lactic acid)/poly(ε-caprolactone) [24].
In this study, thymol/carvacrol mixture was loaded into composite films composed of chitosan and gelatin using SSI process. Both of this polymers, obtained from natural and renewable resources, are biodegradable, non-toxic and biocompatible [25]. In addition, it was previously reported that chitosan and gelatin promote wound healing which makes them perfect materials for wound dressing application [26,27]. Also, recent reports pointed out improved activity of composites obtained by combination of chitosan and gelatin [25]. In order to improve properties of prepared chitosan/gelatin films and to improve its loading capacity, addition of highly porous zeolite was also proposed. Zeolites themselves are also considered as effective carriers for different compounds as they can exhibit controlled release [28][29][30][31]. Our previous research showed the possibility of impregnation of chitosan/starch/zeolite composites with thymol using scCO 2 [32]. However, to the This article is protected by copyright. All rights reserved. best of our knowledge, this manuscript presents application of SSI process for loading of thymol/carvacrol mixture into chitosan/gelatin and chitosan/gelatin/zeolite films for the first time.
Additionally, antibacterial and antioxidant activity, along with structural, thermal and mechanical properties of the obtained composites were discussed in detail.

Preparation of zeolite sample
Zeolitic tuff from Slanci deposit (Serbia) containing about 70 wt% of natural zeolite -clinoptilolite with particle size of 63-125 µm was used. Specific surface area of the zeolite was enlarged by procedure previously published [33]. In short, the procedure consists of firstly transforming the Z into NH 4 -form (NH 4 -Z) by treatment with a solution of ammonium acetate (1 mol dm -3 ) during 24 h at room temperature. Further, NH 4 -Z was calcinated in the air at 550 ºC for 3 h and then calcined product was treated with 0.6 mol dm -3 of HCl at 70 ºC for 1 h. Subsequently, the obtained solid was treated with fresh 0.05 mol dm -3 of HCl and ultrasound (Bandelin, Sonopuls) three times. Solid was than rinsed in distilled water until the negative reaction to chloride ions and dried at 60 ºC to a constant mass. The final product was denoted as Z.

Preparation of blends
Chitosan was dissolved in a 2% (w/v) acetic acid solution to yield a 2% (w/v) chitosan suspension.
The suspension was stirred using a magnetic stirrer until the clear solution was obtained. The 2% (w/v) gelatin solution was prepared by dispersing the gelatin in distilled water and heating the This article is protected by copyright. All rights reserved.
suspension on a hotplate for 20 min at 80 °C with stirring. Chitosan/gelatin (CG) films were prepared by mixing the chitosan solution (2% w/v) with the gelatin solution (2% w/v) in mass ratio of 1:1.
Glycerol (25% w/w of the total solid weight) was added to the mixture. The mixture was stirred at 80 °C for 40 min using a magnetic stirrer and subsequently cooled to the room temperature. The obtained mixture was casted onto the plastic molds (12 cm diameter). After drying at room temperature for 24 h, the films were additionally dried in an oven over night at 35 °C. Obtained sample was denoted as CG (1-0). CG films containing zeolite were prepared as follows. An amount of 30% (w/w) of zeolite (based on the total solid weight) was added during the mixing step of chitosan and gelatin solutions.
Afterwards, the suspension was vigorously stirred (15000 rpm) for 15 min using an Ika Ultra-Turrax disperser (Staufen, Germany). The subsequent steps of casting the obtained suspension and drying were conducted as previously described. Obtained sample was denoted as CGZ (2-0).

Supercritical impregnation
The prepared CG (1-0) and CGZ (2-0) films were impregnated with thymol/carvacrol (TC) mixture in scCO 2 at 30 MPa, 35 ºC for 18 h. These conditions employed in our previous study [32] were found as optimum for functionalization of starch/chitosan/zeolite films with thymol. The experiments were conducted in a high-pressure view chamber (Eurotechnica GmbH, Germany), using a static method, as previously described [17,32,34]. The initial mass ratio of thymol/carvacrol mixture to films in the view chamber was 10/1, while mass ratio of thymol to carvacrol in the mixture was 1:1. The decompression rate, applied at the end of SSI process, was 1.5 MPa/min. TC impregnated 1-0 and 2-0 samples were denoted as 1-1 and 2-1, respectively.
The amount of loaded thymol/carvacrol mixture in the films was determined after complete disintegration/solubilisation of the samples in the 2% acetic solution (100 ml) by measuring of absorbance at 276 nm [24] using a UV-Vis spectrophotometer (Shimadzu 1700).

Antibacterial activity
The antibacterial activity of prepared films was tested towards Gram-negative bacterium Escherichia coli strain DSM 498 and Gram-positive bacterium Staphylococcus aureus strain ATCC 25923. Bacteria were pre-grown on the Nutrient agar (NA, Torlak, Serbia) for 16 h at 37±0.1 ºC to obtain cultures in a log phase of growth.
The disk diffusion method was used for the qualitative assessment of the antibacterial activity. Prior to test, all prepared films were sterilized by UV light for 30 min. Biomass of each bacterial strain was separately suspended in a sterile physiological solution to obtain bacterial concentration of about 10 9 CFU/mL. Agar diffusion test was performed on Mueller Hinton agar (MH, Torlak, Serbia). First, the bacteria biomass was inoculated on the agar, and then sterile films, cut into pieces of 1 cm 2 , were placed on agar plate. The plates were incubated for 24 h at 37 ºC. After incubation, the clear zones without bacteria growth around films were inspected visually. Antibacterial activity of the TC loaded composites was compared with neat films as control.  (1) where A sample and A control refer to the absorbance of DPPH • radical in the sample and control solutions, respectively.

ABTS free radical scavenging assay
The 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) assay was performed using the literature method [36] with slight modifications. Firstly, ABTS •+ radical cation solution was prepared by mixing ABTS stock solution (7.8 mM) and potassium persulfate (K 2 S 2 O 8 ) solution (2.45 mM). Afterwards, prepared solution was left 12-16 h in a dark for completion of a reaction, and then solution was diluted with methanol to obtain absorbance of 0.700±0.020 at 734 nm. The solutions with film sample were prepared as in DPPH assay. The 0.3 mL of each sample solution was added to 2.8 mL of ABTS •+ solutions. The 0.3 mL of pure distilled water/PBS solution in 2.8 mL ABTS •+ were This article is protected by copyright. All rights reserved. used as controls. As in DPPH assay, antioxidant activity of TC mixture was determined as well. The 0.3 mL of the TC mixture were mixed with 2.8 mL ABTS •+ . The distilled water or PBS without antioxidant as control was also prepared. Absorbance at 734 nm of all sample was determined after incubating in the dark for 30 min at room temperature. Ascorbic acid was used as a reference standard. This test was carried out in triplicate. The scavenging activity was calculated using Eq.1, where A sample and A control refer to the absorbance of ABTS •+ in the sample and control solutions, respectively.

Release kinetics
Kinetic study on TC release from the impregnated films in both deionized water (25 °C) and PBS (pH 7.4) at 37 °C was performed using a UV-Vis spectrophotometer (Shimadzu 1800, Japan) at 276 nm. Film samples (0.0022±0.0005 g) were immersed in 100 ml of distilled water/PBS without stirring. At pre-determined periods, an aliquot (3.5 ml) of the solution was taken, analyzed and then returned into the release medium. Release experiments were carried out in duplicate during 10 days. TC concentration was calculated using a previously determined calibration curve. Experimental results were given as released mass of TC mixture per g of impregnated film plotted against time.
Two kinetic models, Weibull (Eq. 2) and Higuchi (Eq. 3) were used to fit the experimental data: Weibull: where M t is the amount of TC released at time t, M ∞ is the amount of TC loaded in the film; a and b are constants. The nonlinear regression module of Polymath Educational 6.10 software package was used to determine parameters of the models used.

FTIR analysis
Fourier-transform infrared (FTIR) spectra of films impregnated with TC mixture as well as neat films were recorded in the ATR mode using a Nicolet™ iS™ 10 Spectrometer (Thermo This article is protected by copyright. All rights reserved. Fisher SCIENTIFIC) with a resolution of 4 cm -1 at wavenumbers in the range of 4000-500cm -1 .

Scanning electron microscopy (SEM) analysis
The surface morphology of the 1-0 and 2-0 samples before and after the SSI with TC mixture was analyzed by field emission scanning electron microscopy (JEOL JSM-6610LV). The samples were coated with a thin layer of gold before the analysis.

TGA-DTG analysis
Thermal properties of the prepared films were investigated by thermogravimetric and derivative thermogravimetric analysis (TGA-DTG) using a SDT Q600 simultaneous TGA-DTA instrument (TA Instruments). The samples were heated in a standard alumina sample pan from room temperature up to 600 °C at a heating rate of 10 °C min -1 under nitrogen atmosphere with a flow rate of 100 cm 3 min -1 .

Water vapor permeability
Water vapor permeability (WVP) measurements were conducted according to the standardized methodology [37,38]. Measurements were carried out gravimetrically at 25 °C and 75% relative humidity (RH) in the desiccator (Star desiccator, Bela). Saturated sodium chloride was used to provide desired RH in the desiccator. The samples were sealed over a circular opening (exposed area of 12.56 cm 2 ) of aluminum permeation cells filled with anhydrous sodium chloride to provide 0% RH inside the cell and placed inside the desiccator.
Due to the moisture gradient, water vapor penetrates through the tested samples towards the inside of the cells, while anhydrous sodium chloride soaks up water vapor, which results in a weight increase of the cells. The cells' weight change was determined twice a day for seven days, using an analytical scale with an accuracy of ±0. 1 mg. Tests were run in triplicates.
Changes in the weight of the cells were plotted with respect to time, and the linear leastsquare method was used for the calculation of the parameters given by Eqs.5 and 6 [39,40] : where WVTR is the water vapor transmission rate of films (g/s), L is average thickness of the film (m), A is the permeation area (m 2 ), Δp is the difference in water vapor pressure between the two exposed sides of the film (Pa).

Water solubility of films
Solubility of films was determined as follows. Specimens of films (2x2 cm) were dried in an oven at 105 °C during 24 h and initial dry weight (W i ) was determined using analytical scale (±0.0001 g). Afterwards, dried samples were immersed into 30 mL of distilled water and placed inside the incubator (Colo, IN1017, Slovenia) at 25 °C for 24 h. After this period, films were taken out from water and dried once again in an oven at 105 °C during 24 h and remaining mass of samples (W f ) was determined. Solubility of films (S%) was determined using following formula [41,42]: The cross-linking or weight remaining percentage was calculated using following formula [43]: linking Cross (8) All tests are the means of at least 3 measurements.

Swelling test of films
Samples were dried to a constant weight prior swelling analysis in an oven at 105 °C for 24 h. The swelling percentage (SW%) was calculated using the following equation [44]: All tests were performed in triplicates.

Mechanical analysis
Mechanical properties of the films were tested using AGS-X testing machine (Shimadzu, Japan) at room temperature (22 °C). Tensile testing was performed at 2 mm/min crosshead speed. Tensile strength as tensile stress at break and percent elongation at break were calculated from tension data. Young's modulus was calculated as a slope of initial linear portion of stress-strain curve. Three specimens of each film were tested and average values were reported.

SSI of composite films with thymol/carvacrol mixture
SSI process was performed in order to functionalize previously prepared CG and CGZ films with bioactive components. The results showed that SSI method was efficient for delivering of TC mixture into the prepared CG (1-0) and CGZ (2-0) films at selected conditions of 30 MPa and 35 °C during 18 h. The amount of loaded TC in the CG (1-0) was 3.3% (i.e. 36.1 mg TC /g film ). Higher loading capacity of CG films were previously reported when clove oil was used [19]. Namely, depending on the SSI process conditions (40 °C, 10-30 MPa, 2-18 h), amounts of loaded clove oil in the CG films were in the range from 50-130 mg/g [19]. The difference in the loading values between reported findings and results of this study can be explained by the differences in process conditions. Namely, process pressure, temperature, operating time, and decompression rate greatly influence polymer impregnation. In addition, the efficiency of the SSI is greatly influenced by the partition coefficient of solute (active substance) between the polymer and the supercritical fluid phase, plasticizing effect of supercritical solution (CO 2 and the solute), and the existence of specific interactions between scCO 2 and/or solute with the polymer network [19,45,46]. This conclusion is also supported by the findings of Lukic et al. [24] who reported that poly(lactic acid)/poly(ε-caprolactone) film can be loaded with TC mixture in the amount of 215 mg TC /g film at 40 °C and 10 MPa during 5 h [24].
Results of proposed SSI process also revealed that CGZ (2-0) film was impregnated with almost 70% higher amount of TC mixture (6.0% i.e. 54.2 mg TC /g film ) compared to the loading of CG film.
Increase of loading capacity of the film with addition of zeolite is in accordance with the results from our previous study, where incorporation of zeolite into the starch/chitosan blend increased the sorption capacity of the material towards thymol greatly [32]. The reason for this can be found in the high porosity of zeolite and relatively high affinity of zeolite towards thymol and carvacrol in scCO 2 -assisted process [47].
In order to investigate biological activity of prepared materials, antibacterial and antioxidant This article is protected by copyright. All rights reserved. activity was further tested. TC impregnated samples with and without zeolite were denoted as 2-1 and 1-1, respectively.

Antibacterial properties of the obtained films
The antibacterial activity of prepared films, with and without bioactive components, is shown in Fig   1. It is evident that both control samples (with and without zeolite) did not exhibit antibacterial activity towards tested bacteria stains. This result is in accordance with the literature, which reports that neither gelatin/chitosan film [48], nor zeolite [49,50] itself exhibit antibacterial activity. S. aureus = 1.4 · 10 9 Antibacterial effect of 1-1 sample was insignificant and slightly more pronounced in contact with S. aureus (Fig. 1a,c). On the other hand, 2-1 sample exhibited strong antibacterial activity towards both bacterial stains (Fig. 1b,d). The reason for different antimicrobial activity toward tested bacteria can be found in amount of loaded TC mixture. Namely, the higher loadings of natural bioactive compound, the stronger antimicrobial activity will be [17,51]. CGZ impregnated film (2-1) enabled formation of the zones of inhibition that have diameter of approximately 4 mm and 8 mm for E. coli and S. aureus, respectively. It is worth to notice that impregnated films showed higher activity towards Gram-positive bacteria strain. This observation can be explained by the structure of the cell wall of tested bacteria cells. Namely, outer membrane of the Gram-negative bacteria is composed primarily of phospholipids and lipopolysaccharide molecules. This layer forms a hydrophilic permeability barrier which can serve as a protection against hydrophobic materials, such as thymol/carvacrol. In the case of Gram-positive bacteria like S. aureus, cell wall allows hydrophobic molecules to easily penetrate the cell and express its antibacterial effect [52]. The mechanism of the antibacterial activity of the phenols -thymol and carvacrol, is based on the interruption of the cell wall and membranes of the bacteria and it can lead to the cell lysis and the leaching of the cell content [53]. Thymol action is attributed to the integration of polar head-groups of the lipid bilayer which induce the alternation of the cell wall, whereas carvacrol action is linked to the presence of the hydroxyl group which acts as a trans-membrane carrier by carrying the H + ions into the cytoplasm and transporting K + ions back [54]. In addition, synergetic effect of carvacrol and thymol towards E. coli and S. aureus was previously reported [55].

Antioxidant activity of the films
High level reactive oxygen species (ROS) and free radical slow down the wound healing process. Therefore, natural biopolymers dressings possessing antioxidant capacity are highly desirable to reduce harmful effects of ROS [56]. Furthermore, it was proven that chitosan films incorporated with plants oils could be used as a potential wound healing materials [57]. The in vitro antioxidant scavenging activity of the control and TC impregnated films were determined using two free radical assays: DPPH and ABTS.

DPPH antioxidant assay
The results of in vitro antioxidant test for all prepared films are shown in Fig. 2 as a percent of inhibition of DPPH • radical in relation to time. The control 1-0 sample showed negligible antioxidant activity (less than 1% of inhibition), while the activity of around 5% in the case of the control 2-0 film indicates that zeolite acts as possible source of dissolution of ion accelerating inhibition process. Previous study [58] showed that chitosan exhibit certain antioxidant potential depending on the molecular weight and the degree of deacetylation.
Antioxidant efficiency was attributed to the presence of hydroxyl and protonated amino groups in its structure. Kyung et al. reported that large number of inter-and intra-molecular interaction between amino and hydroxyl groups in high molecular weight chitosan contribute to lower DPPH • radicalscavenging activity [59]. Accordingly, the results from this study agree with low activity of control films. Additionally, amino acids, glycine and proline, present in the structure of gelatin, have donor groups that could react with DPPH • free radical ccontributing to its stabilization [60]. Consequently, the antioxidant activity of the control films could be attributed to the antioxidant activity of chitosan and gelatin. The interaction of electron-donating hydroxyl groups in the structure of both components of TC mixture, incorporated in both 1-0 and 2-0 films, with stable DPPH • free radical led to significant increase of the measured antioxidant activity, in comparison to control samples. As can be seen from Fig. 2, the TC impregnated films (1-1, 2-1) showed moderate activity, with approximately up to 50% of inhibition after 60 min, compared to ascorbic acid. The activity slightly increased up to 70 min, and did not change after prolongation of the test to 2 h and Furthermore, it could be assumed that perceived antioxidant activity is a consequence of the quantity of released active molecules (effectiveness of film bonding to TC), and medium properties and electronic structure of antioxidant molecules (carvacrol and thymol) i.e. their electron/proton donating capability [63,64]. This article is protected by copyright. All rights reserved. reported that the antioxidant activity of PLA-PLGA films was higher when CT mixture was used, comparing to films containing only carvacrol or thymol [24].

ABTS •+ antioxidant assay
The antioxidant activities of tested films were evaluated using ABTS •+ free radical method ( Figure 3) as well. The benefit of this free radical process is that the reaction between ABTS •+ and potassium persulfate is stoichiometric and produced color remains stable for more than 2 days when stored in the dark at room temperature [67].

Release study
This article is protected by copyright. All rights reserved.
Acute wounds heal in an orderly and efficient manner, and progress smoothly through the four different, overlapping stages of healing: haemostasis, inflammation, proliferation and remodeling [68]. Wound dressing materials can accelerate the duration of these phases. In order to exhibit antimicrobial and antioxidant activity, it is necessary that impregnated materials release loaded bioactive components. In addition, for material to be used as wound dressing, the release of loaded bioactive compounds should be in a gradual manner. Some of the parameters that determine release of loaded active compound are type of release medium, as well as morphology and chemistry of polymeric carrier [17]. Therefore, besides buffer solution (pH 7.4) at 37 °C that simulated body fluids, distilled water was also used as model medium for testing of TC release from the impregnated samples to evaluate their possible application as wound dressing material.

Kinetics of TC mixture release from the impregnated films
Kinetics of TC mixture release from the impregnated films is presented in Figure 4. It can be seen that both impregnated samples (1-1 and 2-1) expressed initial burst release during first 6 h of the test, whereby the amounts of released TC mixture were around 8 and 37 mg TC /g film in water and 22 and 44 mg TC /g film in PBS medium, respectively. After this period, films exhibited significantly slower release of TC mixture in a gradual manner during tested period of 10 days.
The initial burst release is highly desirable to enable delivery of sufficient amount of bioactive compound to the target site at the beginning of the healing process, since it reduces the risk of infections occurrence greatly [69]. Further, sustained release, that enables maintaining of an amount of bioactive compound at target site, is desirable for a prolonged period of time [25].
After 10 days of release experiment, amounts of TC mixture leached from the 2-1 and 1-1 samples were 72% and 29% in the water and 96% and 48% in the PBS media, respectively.
Obviously, the higher the loading the higher the release rate was. Similarly, Milovanovic et al. [17] determined that thymol desorption from the impregnated cellulose acetate beads was correlated with the amount of loaded thymol. Thymol release lasted between 2 days (4.5% of loaded thymol) and 21 days (63.0%).
It can be also noticed that the amount of TC mixture was higher in the PBS medium than in the water for both tested films. This observation can be explained by the poor solubility of thymol and carvacrol in water [70], as well as the fact that slightly alkaline media accelerate polymer degradation and promote leaching of loaded substance [71]. A similar effect of chemical nature of the medium on the release rate has been reported for thymol release from cellulose acetate [51] and PLGA [72] used as a carriers. Recently, Lukic et al. reported synergetic effect of TC mixture, which slow down the release from the PLA films compared to the desorption of single thymol/carvacrol [24]. The amount of released TC mixture in distilled water after 10 days was around 60 mg TC /g film [24]. Similarly, CGZ (2-1) film released around 55 mg TC /g film after 10 days in PBS at 37 °C.

Modeling
The experimental data of release kinetic were correlated with two most frequently used kinetic models (Eqs. 2,3) in order to describe the release mechanism of bioactive components from prepared films. Thus, Weibull function and Higuchi model [73] were applied to simulate the release kinetics of TC mixture from 1-1 and 2-1 films into distilled water and PBS.
Determined values of the models parameters are presented in Table 1 and simulation curves at   The release of active compounds from polymeric matrix through diffusion mechanisms has been already reported for pure thymol and carvacrol release from different materials [51,74,75], as well as for the release of TC mixture from PLA/PCL film into distilled water [24]. When the compound release follows purely Fickian diffusion, a simple and flexible Weibull model is recommended for the entire release profile [74]. Accordingly, the good agreement between experimental values and model was obtained when Weibull model was applied to describe the kinetics of TC mixture release from 1-0 and 2-0 films. Higuchi model gave a significantly lower agreement with the experimental data for all films, with 0.22 < R 2 < 0.89.

FTIR analysis
To investigate the structural properties of neat and impregnated films and potential interactions between films and TC, FTIR analysis was performed ( Figure 6). This article is protected by copyright. All rights reserved.
wavenumbers 1700 and 900 cm −1 [76,77]. Absence of unique carboxylic band at about 1690 cm −1 might suggest participation of -COO − groups of gelatin in the electrostatic interaction with positively charged amino groups of chitosan, confirmed by the shift of amide II band of chitosan to higher wavenumber [77]. Electrostatic interactions were also evidenced by the band at 1377 cm −1 of chitosan recognized as a shoulder of the band at 1410 cm −1 in the spectra gelatin [77].
After addition of zeolite to 1-0 films, the apparent decrease in the band intensities at 3270 cm −1 as well as the amide I (1632 cm −1 ) and II (1547 cm −1 ) was observed, which could be explained as a result of physical inclusion of zeolite in the blends [79].  showed characteristic bands of thymol and carvacrol, isomers with similar spectra and differences only in the fingerprint region [22,76]. Presence of TC mixture on the surface of 1-1 film with only 3.3% of TC mixture loading was confirmed by band attributed to the out-of-plane CH wagging vibrations from isoprenoids, appeared at 809 cm -1 , arising from the overlapping of thymol and carvacrol bands [24,80]. In addition, band at 2959 cm −1 assigned to stretching vibration of CH 3 group originated from TC mixture appeared in the FTIR spectra of 1-1 film (Fig. 6a). Incorporation of TC mixture into the 2-0 film and interactions of phenolic compounds with the polymer matrix through formation of intermolecular hydrogen bonds are more obvious. Aforementioned bands, at 2959 and This article is protected by copyright. All rights reserved.
20 809 cm −1 , are more intense in the spectra of 2-1 film due to higher amount of TC mixture present in film (Fig. 6b). Beside, shift of the broad band corresponding to stretching vibrations of -OH group to the higher wavenumber (3296 cm −1 ) compared to neat 2-0 film evidenced the establishing of hydrogen bonds between TC mixture and -OH groups of the 2-0 film [19]. The broadening and lowering intensity of the band at around 1522 cm −1 , related to the bending vibration of N-H in amino groups, indicate the participation of these groups in interactions and formation of intermolecular hydrogen bond with -OH groups in phenolic compounds. Furthermore, the bands at 1456, 1418 and 1288 cm -1 corresponding to vibrations in phenolic ring of thymol and carvacrol [80].
Bands at around 945 cm -1 assigned to =CH out-of-plane bending adsorption [22] are also observed in the FTIR spectra of 2-1 film. Lower intensity of characteristic saccharide bands situated in the range of 1180-900 cm −1 and appearance of new bands that correspond to TC mixture also confirmed the incorporation of thymol and carvacrol into the polymer film.

SEM analysis of the films
SEM images of fabricated films before and after the SSI with TC mixture are presented in Fig.   7. Surface and cross section of neat film without zeolite was flat, smooth and homogeneous. Similar images of chitosan gelatin films surface and cross section were already published [19,81]. Addition of zeolite to the chitosan gelatin polymer frame led to the formation of rough structure (Fig. 7c).
According to the cross section of 2-0 sample it can be concluded that zeolite particles were fairly distributed through the chitosan gelatin polymer network. The obtained results were in accordance with the study of Milenkovic et al. [82], where the formation of rocky surface of poly(vinyl chloride) films was reported after micronised silver-exchanged natural zeolite (Ag-NZ) was added to the polymer frame.
This article is protected by copyright. All rights reserved. As can be seen, the SSI did not affect the morphology of the prepared films film remarkably ( Fig. 7b and 7c). The polymer melting was not observed in the case of 1-1 neither for 2-1 sample, which was in agreement with the results of thermal analysis (Section 3.5.3).

TG-DTG analysis
Thermal analysis of the prepared films showed that in the observed temperature range up to 600 °C, weight loss occurs through three steps in the case of control samples, while TC impregnated films expressed two stage weight loss. Thermal degradation of control 1-0 sample (Fig. 8a) fits well the literature data [19]. properties. Indeed, degradation temperatures were shifted to somewhat higher values (Fig. 7).
Increased thermal stability of the films loaded with TC mixture can be explained by formation of hydrogen bonds between thymol/carvacrol and polymer network, which was confirmed by FTIR analysis. Accordingly, DTG curves of the impregnated samples did not displayed maxima characteristics for pure thymol (160 °C) [83] nor for pure carvacrol (170 °C) [84] related to decomposition of these compounds.
This article is protected by copyright. All rights reserved.

Barrier properties of the films
The results of WVP analysis are shown in Table 2. It can be seen that the permeability values of pure 1-0 films were lower compared to the 2-0 samples due to the presence of zeolite.
Namely, it can be assumed that increased gaps between polymer chains inside the matrix of the 2-0 film contributed to formation of the polymer network that become less dense and more open to the adsorption/desorption of water molecules [32].

Solubility of films
Results of solubility test in water are summarized in Table 3. Table 3 Solubility of fabricated films in water This article is protected by copyright. All rights reserved.
Solubility values of tested films were above 30% indicating high degree of gelatin and chitosan cross-linking [43]. Namely, gelatin as water soluble compound can easily lose its fibrous structure in high ambient humidity conditions. However, mixing gelatin with other functional materials, such as chitosan, can induce crosslinking and stabilization of gelatin structure and consequently decrease its solubility in aqueous medium [88]. As expected, presence of zeolite led to the higher solubility of fabricated materials (2-0). The increase in percentage of solubility of around 30% could be induced by physical incorporation of zeolite particles into the film and obstruction of the polymer chains of chitosan and gelatin. This is supported by the reduced value of cross-linking in the case of 2-0 sample compared to the cross-linking degree of film without zeolite. After the SSI solubility of films increased in both tested materials (1-1 and 2-1). Obtained results are in accordance with the literature, since it was already published that solubility of gelatin-chitosan films significantly increased in the presence of essential oil [89]. Furthermore, Kavoosi et al. [43] reported that incorporation of thymol caused an increase in solubility of gelatin films. The amount of thymol added to the gelatin films was in the range of 1-8% and it was noticeable that the higher the thymol amount was the higher the films' solubility was [43]. Similarly, higher solubility was observed in the case of 2-1 sample with higher amount of incorporated bioactive compounds compared to the solubility of 1-1 film. It can be assumed that TC molecules interfere the polymer chain-to-chain interactions and lead to the slight decrease in degree of chitosan-gelatin cross-linking. However, according to results from Table 2 it can be concluded that even after incorporation of TC mixture into the films the increase in solubility was not significant and the films retained cross-linking structure.

Mechanical analysis
Mechanical properties of both neat and impregnated films, namely, tensile strength ( σ) as a measure of maximum stress that material can withstand before breaking, Young's modulus (E) which quantifies film flexibility, and elongation at break (ε) which quantifies the capacity of film to extend before breaking, are shown in Table 4. As can be seen, addition of zeolite to 1-0 films decreased σ and E, while ε remains almost unchanged. A reduction in mechanical strength when added amount of zeolite is higher than 20% is explained by the formation of too many interfacial voids at the inte rface of chitosan and zeolite [92]. The same observation was previously reported for starch/chitosan films that contained different amounts of zeolite [32]. Decreased values of Young's modulus after impregnation of TC mixture into 2-0 film revealed that more flexible film was obtained by loading of active substances due to their possible plasticizing effect [24], while elongation at break only MPa) [94].
Loading of TC mixture into the 1-0 and 2-0 films caused decrease of the values of σ by 32 and 20%, respectively, compared to non-impregnated films. Although opposite could be found in the literature [60], our results are in accordance with many studies, where the addition of active compounds has been reported to reduce the σ of polymer films. After adding of Origanum vulgare and orange peel essential oil to fish gelatin/chitosan film, σ decline of 36-69% and 9-21%, respectively, was reported [95,96], while loading of TC mixture into PLA/PCL film reduced the value of σ by 78% [24]. It was proposed that essential oils affect the microstructure of polymer matrix by rearrangement of polymeric chains. The interactions between chitosan and gelatin molecules are partially replaced by the weaker polymer-oil interaction in the film matrix [40,97].