The bovine collagen fiber was supplied by NovaProm Food Ingredients (São Paulo, SP/Brazil). The synthetic PVA was acquired from Kuraray (São Paulo, SP/Brazil). The oregano and rosemary (pre-dehydrated) were purchased locally (Chapecó, SC/Brazil). The other reagents were of analytical quality and included brain heart infusion (BHI) broth (Kasvi, Spain), BHI Agar (Kasvi, Spain), tryptone soy broth (TSB) (Kasvi, Spain), sodium chloride (Synth, Brazil), acetic acid (Vetec, Brazil), Tween 80 (InduLab, Brazil), glycerol (Dinâmica, Brazil), resazurin (Sigma Aldrich, USA), and potassium nitrate (Dinâmica, Brazil).

The strains were stored under freezing (−18˚C) in TSB and glycerol mixture until the time of use. Before the experiments, the strains were grown twice on tilted BHI agar and incubated in a microbiological oven at 37˚C for 24 h. Bacteria from the American Type Culture Collection (ATCC) were used: Staphylococcus aureus (SA) (ATCC 6538), Escherichia coli (EC) (ATCC 8739), Pseudomonas aeruginosa (PA) (ATCC 9027), Salmonella typhimurium (ST) (ATCC 1408), Listeria monocytogenes (LMB) (biofilm forming) (ATCC 7644), and Listeria monocytogenes (LMSA) (Scott A).

The essential oils were extracted using the steam distillation method [9] .

The chemical characterization of the oregano (OEO) and rosemary (REO) essential oils was performed by gas chromatography coupled to mass spectrometry (GC/MS Shimadzu QP2010 Plus; Shimadzu Corporation, Tokyo, JP) with a ZB – 5 MS capillary column (30 m × 0.25; 0.25 mm, flow rate 1 mL/min). The ramp conditions were 40˚C for 1 min and 230˚C at a rate of 4˚C/min for 3 min and the injector at 250˚C. A series of n-alkanes was analyzed under the same chromatographic conditions to calculate the linear retention index (LRI). Identification was made by comparing with mass spectra from the National Institute of Standards and Technology (NIST) and experimental LRI with those available in the literature.

Minimum inhibitory concentration (MIC) analysis was based on the broth microdilution technique [10] . The essential oils (EOs) were diluted to 25% (w/v) in BHI broth supplemented with 3% Tween 80, and serial dilutions were performed in the microplate for the OEO (240 - 0.12 mg/mL) and REO (229.12 - 0.111 mg/mL). Dilutions were calculated based on the density of the EOs: 0.9618 g/mL for the OEO and 0.9165 g/mL for the REO. The concentration of the bacteria of interest was adjusted to 1.5 × 10⁸ (0.5 McFarland) in 0.85% saline solution. Subsequently, 10 µL of the bacterial suspension was added to each well, followed by homogenizing each well of the microplate. The plates were covered and incubated in a bacteriological oven at 37˚C for 24 h. Afterward, 10 µL of resazurin solution (0.01%) was added to verify if the target bacteria was inhibited, and the color blue characterized bacterial inactivity and red characterized activity [11] .

For the minimum bactericidal concentration (MBC), all concentrations that inhibited the target bacterium were used, and a 10-µL aliquot was taken and deposited in Petri plates containing BHI agar. The plates were incubated in an incubator at 37˚C for 24 h. The lowest concentration at which no microbial growth occurred on the plate was the MBC.

The micro-atmosphere analysis was based on the method of Ghabraie et al. [12] with modifications. The concentration of the bacterium of interest was first adjusted (0.5 McFarland), followed by depositing 100 µL of the solution into Petri dishes containing BHI agar. After the plates dried, a sterile paper disk was inserted into the lid of the Petri dish, and a 30-µL aliquot of the EO was inserted (30 µL of sterile distilled water was used for the control plates). Immediately afterward, the plates were sealed with Parafilm®. The plates were incubated at 37˚C for 24 h. Afterward, the total inhibition halo was measured using a manual pachymeter (Messen, Brazil).

The films were developed according to Paglione et al. [13] with modifications. A mixture of collagen fiber (5.0% w/v), polyvinyl alcohol (5.0% w/v), Tween 80 (1.0%), and glycerol (1.30%) was made. The pH of the solution was set to 3.0 using a 50% acetic acid solution. The solution was homogenized for 3 min with a homogenizer (Turratec, Tecnal, TE 102). The solutions were then heated at 90˚C for 20 min in an ultra-thermostatic bath (Solab, Model SL 152). After the heating time, 2%, 4%, and 6% concentrations of the OEO or REO were added. The control solution did not contain the addition of EO. The solutions were then placed in an ultrasound bath (Ultronique, Model Q3.8/40A) to remove the bubbles formed for 5 min at room temperature. Each solution was inverted in glass plates (29.7 × 21.0 cm) and placed in an oven with air circulation (Marconi, Model MA035) for roughly 24 h. The films were then stored until the respective analyses in desiccators with saturated potassium nitrate solution (RH 60% and 25˚C) controlled by a thermo-hygrometer (Akso-AK28).

Aliquots of 100 µL of the bacteria of interest (0.5 McFarland) were inoculated into Petri plates containing BHI agar. The developed films were aseptically cut into 20-mm disks, positioned in the center of the plate, and incubated in an incubator at 37˚C for 24 h. After the elapsed time, only the inhibition halo was measured with a manual pachymeter (Messen, Brazil).

The analysis was performed as described in item 2.5.1, although the 20-mm disks were deposited on the inner face of the lid of the Petri dish in this assay. The plates were then sealed with Parafilm^®. The total inhibition zone (including the film) was measured with a manual pachymeter (Messen, Brazil).

The films’ thickness was measured using a manual micrometer (Messen, Brazil) by randomly reading at 10 points.

In order to determine the L*, a* and b*, C*, h*, and opacity values of the films, six readings were taken using a spectrophotometer (Model CM-700d, Konica Minolta Sensing Inc., Japan). The films were placed on a standard white plate (L = 88.66, a = 0.3, and b = 3.36). Readings were taken using D65 illuminant, specular component excluded (SCE), 8 mm measurement area (MAV), and a 10˚ observation angle.

The total color difference (ΔE) was calculated using Equation (1):

$\Delta E={\left[{\left(\Delta L^{\ast }\right)}^2+{\left(\Delta a^{\ast }\right)}^2+{\left(\Delta b^{\ast }\right)}^2\right]}^{0.\text{5}}$ (1)

The opacity was calculated by the overlapping ratio of the black standard (S_black) and white standard (S_white) according to Equation (2) [14] :

$\text{Opacity}=S_{\text{black}}/S_{\text{White}}\times 100$ (2)

The films’ surface microstructure images were determined with an electron microscope (JEOL, model JSM-6360) with a 10 kV electron beam. The samples were dried in an air oven at 30˚C for 12 h. The film samples were adhered to the surface of double-sided carbon strips and then coated with a gold layer using a vacuum metallizer for 40 s (Denton Desk V; Denton Vacuum, USA). Surface images were taken at 100× and 1000× magnifications.

The infrared spectrometry analyses were performed in infrared spectroscopy (Prestige, Shimadzu) by the direct transmittance method using the KBr tablet technique. The spectra were obtained in the 400 - 4500 cm⁻¹ range with 45 scans and a resolution of 2.0 cm⁻¹.

For the bacterial susceptibility (micro-atmosphere) of the pure essential oils, a completely randomized factorial experimental design with four replications arranged in a 2 × 6 factorial scheme (2 types of essential oil × 6 microorganisms) was used.

For bacterial susceptibility (diffusion disk and micro-atmosphere) of the films, a completely randomized experimental design with a 2 × 4 × 6 factorial arrangement (2 types of essential oil × 4 levels of essential oil × 6 microorganisms) was used, totaling 48 treatments with four repetitions each, according to the statistical model:

$Y_{ijkl}=\mu +{\alpha }_i+{\beta }_j+{\gamma }_k+{\left(\alpha \beta \right)}_{ij}+{\left(\alpha \gamma \right)}_{ik}+{\left(\beta \gamma \right)}_{jk}+{\left(\alpha \beta \gamma \right)}_{ijk}+{\epsilon }_{ijkl}$ (3)

where, $Y_{ijkl}$ = value observed in the i-th essential oil type, j-th level of essential oil, k-th microorganism, and l-th repetition; μ = overall mean of the response variable; ${\alpha }_i$ = fixed effect of i-th type of essential oil; ${\beta }_j$ = fixed effect of the j-th level of the essential oil; ${\gamma }_k$ = fixed effect of the k-th microorganism; ${\left(\alpha \beta \right)}_{ij}$, ${\left(\alpha \gamma \right)}_{ik}$, and ${\left(\beta \gamma \right)}_{jk}$ are the double interactions; ${\left(\alpha \beta \gamma \right)}_{ijk}$ corresponds to the triple interaction; and ${\epsilon }_{ijk}$ = random effect associated with the observation $Y_{ijkl}$, assumption ${\epsilon }_{ijkl}\stackrel{iid}{~}N\left(0,{\sigma }^2\right)$.

For thickness and color (L*, a*, b*, C, h˚, opacity, and ΔE), a completely randomized design with a 2 × 4 factorial arrangement (2 types of essential oil × 4 essential oil levels) was used, totaling 8 treatments with four repetitions each.

The data were subjected to univariate analysis of variance (ANOVA), their means were adjusted by ordinary least squares and compared by the Student Newman Keuls (SNK) test. For the essential oil levels, linear and quadratic tendency were evaluated using orthogonal contrast analysis from the coefficients for interpolation of orthogonal polynomials. Subsequently, polynomial regression analysis was performed to investigate the changes in the dependent variables as a function of the essential oil level. The values of the determination coefficient (r²) were expressed in relation to the treatment sources (regression + lack of adjustment).

The Fourier Transform Infrared Spectroscopy (FTIR) spectral data in the 4000 - 400 cm⁻¹ region were submitted to chemometric analysis. Initially, the spectra were linearized; for this, the first and second derivatives were tested. Then, the multivariate techniques of cluster analysis (CA) and principal component analysis (PCA) were used to reduce the dimensionality of the data and verify differences between treatments.

The statistical analyses were performed in R and RStudio at 5% significance level [15] [16] .

Twenty compounds were identified for the OEO (oregano essential oil): thymol (29.44%), (E)-sabinene hydrate (24.18%), p-cymene (17.94%), terpinen-4-ol (5.52%), linalool (4.91%), (Z)-sabinene hydrate (2.48%), α-terpineol (2.19%), carvacrol (1.86%), D-limonene (1.65%), sabinene (1.56%), caryophyllene oxide (1.53%), carvacrol methyl ether (1.22%), cis-p-menth-2-en-1-ol (0.98%), myrcene (0.96%), thymol methyl ether (0.72%), α-thujene (0.69%), spathulenol (0.67%), β-bisabolene (0.61%), α-pinene (0.48%), and β-pinene (0.40%). Thymol was the main compound found for OEO, corroborating Barbosa et al. [17] . However, some studies have reported that the main compound in OEO is carvacrol [18] .

In the REO (rosemary essential oil), 28 compounds were identified: 1,8-cineole (46.63%), camphor (13.83%), α-pinene (13.63%), β-pinene (4.08%), camphene (4.08%), β-caryophyllene (3.90%), borneol (3.38%), α-terpineol (2.66%), p-cymene (1.90%), β-myrcene (1.40%), linalool (0.97%), terpinen-4-ol (0.80%), γ-terpinene (0.48%), α-terpinene (0.43%), α-humulene (0.40%), bornyl acetate (0.33%), terpinolene (0.24%), α-phellandrene (0.17%), tricyclene (0.12%), 1-octen-3-ol (0.11%), caryophyllene oxide (0.11%), β-bisabolene (0.10%), α-thujene (0.09%), camphene hydrate (0.05%), σ-cadinene (0.04%), dehydrosabinene (0.03%), 3-octanone (0.02%), and trans-pinocarveol (0.02%). 1,8-cineolee was the major compound in the REO, followed by camphor and α-pinene. Cariri et al. [19] also found the same major compounds.

The MIC (minimum inhibitory concentration) (mg/mL) and MBC (minimum bactericidal concentration) (mg/mL) of the OEO (oregano essential oil) were: 7.51 and 7.51 for Escherichia coli (EC), 30.05 and 30.05 for Listeria monocytogenes (Biofilm) (LMB), 30.05 and 60.11 for Listeria monocytogenes (Scot A), 15.02 and 60.11 Pseudomonas aeruginosa (PA), 15.02 and 30.05 Staphylococcus aureus (SA), and 15.02 and 30.05 Salmonella typhimurium (ST), respectively. In the REO, the MIC (mg/mL) and MBC (mg/mL) were 28.64 and 57.28 for EC, 229.12 and 229.12 for LMB, and 57.28 and 229.12 for ST, respectively. With the analyzed REO (rosemary essential oil) concentrations, for some bacteria, it was only possible to verify the MIC: LMSA (229.12), PA (114.56), and SA (229.12).

The bacteria were more sensitive to the OEO, and similar data were reported by Silva et al. [20] and Alvarez et al. [21] . In addition, no differences in efficacy were observed between Gram-positive and Gram-negative bacteria.

The EC showed higher sensitivity, whereas lower sensitivities were presented by LMSA and LMB for the OEO and REO compounds. The SA bacteria also showed lower sensitivity to the REO. Lower values have also been reported elsewhere for the EO of a Mexican oregano species (0.25 - 0.5 mg/mL, MIC) and 0.5 - 0.75 mg/mL, MBC), Greek oregano (0.25 - 0.5 mg/mL, MIC; 1 - 2 mg/mL, MBC), common oregano (2 - 8 mg/mL, MIC; 8 - 16 mg/mL, MBC) [18] [22] . Nevertheless, for the SA bacteria, we observed lower MIC (15.02 mg/mL) and MBC (30.05 mg/mL) values compared to Kosakowska et al. [18] for a common oregano species (32 mg/mL MIC and 64 mg/mL MBC). Literature data indicate chemical composition near those obtained in our analyses. The concentration of thymol (29.44%) was higher than in other studies (0.58% - 28.31%). In contrast, carvacrol (1.86%) was lower (4.64% - 37.21%). Large amounts of this compound (88.50 and 66.89%) in OEO are responsible for the high antimicrobial activity [23] . Although there is a difference in the location of the functional group in thymol and carvacrol molecules, their effects are similar [24] .

Lower MIC values in Gram-negative (2.75 - 5.50 mg/mL) and Gram-positive (0.70 - 2.25 mg/mL) bacteria have already been reported for REO [23] . The composition of REO in the literature was also close to this study, and differences were found for linalool. 1,8-Cineole was identified as one of the main ones responsible for the activity due to the changes it causes in the shape and size of Gram-positive and Gram-negative bacteria; synergistic action with the other compounds has also been suggested [25] .

The micro-atmosphere analysis assesses the biological activity in the steam phase. Differences were observed when comparing the results between MIC and MBC, and such differences have also been reported elsewhere [12] .

The highest (p < 0.05) sensitivities were shown for LMSA, SA, and ST bacteria for OEO ( Table 1 ). The OEO showed higher (p < 0.05) inhibitions in all bacteria, while the REO showed no activity in the steam phase, albeit only for ST bacteria. Among Gram-negative bacteria, PA is less sensitive to the action of Eos [26] . Nevertheless, the REO showed one of the highest inhibition activities compared to the other bacteria analyzed. López et al. [27] did not find inhibitory activity in the steam phase for pure REO for any bacteria. Differences in the biological activity of EOs are expected, as their chemical composition directly influences their bioactivity [28] .

The developed films showed activity against all pathogenic bacteria analyzed, proving that the matrix CF and PVA is promising for applying EOs. According to Seydim and Sarikus [29] , in addition to the composition of the EO present in the film matrix, the material used in the film is fundamental to promoting biological activity through the release property of the EO constituents.

Means followed by different capital letters within the same column and different small letters within the same row differ (p < 0.05), respectively, between bacteria and essential oil by the Student Newman Keuls test. ¹EC = Escherichia coli; LMB = Listeria monocytogenes (biofilm); LMSA = Listeria monocytogenes (Scott A); PA = Pseudomonas aeruginosa; SA = Staphylococcus aureus; ST = Salmonella typhimurium. ²SEM = Standard error of the mean. ³CV = Coefficient of variation.

Levels of EO were incorporated into the CF plus PVA film, and their inhibition potential was verified by the disk-diffusion method. The control films (without adding the EO) showed no inhibitory effect for all bacteria. The regression equations for each bacterium as a function of OEO or REO concentration showed different behaviors for the disk-diffusion analysis ( Figure 1(A) and Figure 1(B) ). Higher OEO concentrations were necessary to inhibit LMSA, EC, and SA compared to REO. However, the point of maximum inhibition was higher for OEO.

In the OEO, the equations for LMSA, LMB, PA, and ST bacteria fitted linear regression, with LMB, PA, and ST bacteria initiating their inhibitions at lower OEO concentrations. The point of maximum inhibition presented by the OEO for LMSA was similar for EC. Quadratic behavior was observed for EC and SA with the increment of OEO. Liu et al. [30] also found disk-diffusion inhibitions of Gram-positive and Gram-negative bacteria for soybean polysaccharide films added with OEO.

In the REO, all bacteria analyzed started their inhibitions at lower concentrations. The SA and LMB bacteria showed maximum inhibition points at 4.6% and 5.0% REO, respectively. Sani et al. [31] reported that protein isolate and cellulose films added with REO were more effective against Gram-positive bacteria, while chitosan films with REO showed inhibition for both Gram-positive and Gramnegative bacteria [32] .

The antimicrobial activity of the films was also verified by micro-atmosphere analysis ( Figure 1(C) and Figure 1(D) ). We observed that higher concentrations were necessary for inhibition in the steam phase for some bacteria due to the need for volatilization of the compounds in the film matrix; thus, the volatilization properties of the EOs are important. The onset of inhibition for LMB, LMSA, PA, and ST bacteria was from the 3.5% concentration of OEO. The EC and SA showed greater sensitivity because they already had inhibition at 1% OEO. Nedorostova et al. [33] found that the same bacteria showed higher sensitivities in the steam phase. Hyun et al. [34] observed activity in the steam phase of OEO for a broad spectrum of pathogenic bacteria.

Similar behavior was noted for the REO in the micro-atmosphere; EC and SA bacteria also showed higher sensitivities (<1%). The bacteria LMB, LMSA, and PA started inhibition at 3.4% and ST at 3.5% REO. Nonetheless, OEO showed greater inhibitions with its increment.

Differences in results between the direct contact and vapor phase method have also been reported in the literature. The inhibition by direct contact is mainly linked to the more hydrophilic compounds in the EO. In the steam phase, there is a balance between hydrophilic and hydrophobic compounds [35] .

As essential oil levels rise there is a higher concentration of phenolic compounds. Then, it is concluded that the biological activity of the films would be greater. Although, different behaviors were observed for some bacteria as essential oil levels raised, because they not showed difference in inhibition according to regression equations. Depending of analysis method (disk-diffusion or micro-atmosphere) and essential oil compounds, the higher amount may not affect the activity.

No higher or lower susceptibility relationship between Gram-positive and Gram-negative was observed, and differences in antimicrobial activity within the positive and negative groups were found. The EOs analyzed have a broad spectrum of activity; researchers have found that Gram-negative bacteria are less susceptible to EOs due to their outer membrane surrounding much of the cell wall [26] , which was not proven in this study. However, relatively small hydrophilic or hydrophobic molecules can cross the outer membrane and access the periplasm through some proteins present, called porins [36] . Compounds such as carvacrol affect the outer membrane of Gram-negative bacteria [37] . The antimicrobial activity of REO, nevertheless, is due to the presence of phenolic compounds such as 1,8-cineole, α-pinene, and camphor with lipophilic and hydrophilic functional groups, because such compounds can interact with bacterial cell components, disrupting the cytoplasmic membrane and affecting cell wall permeability [31] .

The appearance of the packaging plays a determining role in the purchase of a product, in addition to the suitability for its application. The thicknesses found in CF and PVA films, plus the addition of OEO and REO, ranged from 0.22 to 0.32 mm ( Table 2 ). Thickness increased (p < 0.05) with the addition of EO but had no effect (p > 0.05) on EO type. Chen et al. [38] found this in PVA films with EO lower thicknesses, although the higher thickness shown by the films in this study may be attributed to the joint addition of FC to PVA, generating an increase in solids in the films [39] . Chen et al. [38] also found slightly higher values with the addition of EO.

The color indices L*, a* and b*, C value, hue angle, opacity, and color difference presented alterations. There was an effect (p < 0.05) of the interaction between type and level of EO on all the color attributes of the films. The films showed a tendency towards yellow (b*) and green (−a*), which can be explained by the color of the OEOs and the addition of Tween 80 [40] . With the increase of OEO, the films showed higher b* (yellow) values (p < 0.05), whereas the addition of REO changed b* values only in relation to the control ( Table 2 ).

The opacity property increased (p < 0.05) with the increasing addition of OEO and REO. The inclusion of 6% REO provided a higher (p < 0.05) opacity value. Other researchers have reported similar findings [38] [41] . The higher opacity can be explained by the EO droplets causing light scattering in the film matrix [40] . However, the film matrix itself has relevance, as the control CF + PVA film also showed high values compared to other matrices. Comparing the two EOs, the REO showed higher values (p < 0.05) than the OEO. This increase is interesting for enhancing the film’s barrier properties by blocking the passage of light [40] . Despite the high values found for opacity, the developed films presented transparency.

The addition of the EOs in the matrix of the CF + PVA film did not harm its gloss (L*), and an increase (p < 0.05) in this parameter was obtained with the OEO. Interactions between the type and level of EO and the level were also observed. Shen et al. [40] found no significant differences with increasing OEO in starch films, and Chen et al. [38] reported lower L* values in PVA films with the addition of a clove (Syzygium aromaticum) EO.

The surface images of the films were used to acquire important data on the microstructure of the matrix presented with the mixture of CF (collagen fiber) and PVA (polynivyl alcohol) with the addition of the OEO (oregano essential oil) and REO (rosemary essential oil). The combination of FC and PVA ( Figure 2(A) ) showed a partially homogeneous and smooth surface structure. Interconnected fibers were observed, forming a network in which pH 3.0 has great

Means followed by different capital letters within the same column and different small letters within the same row differ (p < 0.05), respectively, between types and levels of essential oil by the Student Newman Keuls test. ¹ ${\hat{y}}_{Overall}=0.2235+0.0353x-0.0035x^2\left(r^2=0.95\right)$.

relevance due to the fiber’s degree of intumescence, which is higher at this pH; Xu et al. [42] also reported a similar result. The formation of the network is also explained due to the denaturation of a part of the proteins, where there is an interaction between different bonds, such as hydrogen bonds, electrostatic and hydrophobic interactions [43] . Some protruding regions can be observed in Figure 2(B) , in which PVA particles were not fully solubilized in the matrix. This condition may be related to the inhomogeneity in temperature and agitation during the heating of the filmogenic solution. Films with the inclusion of the OEO showed spherical droplets on the surface. The addition of 2% OEO led to greater irregularities, as shown in Figure 2(C) . However, with magnification ( Figure 2(D) ), one can observe the OEO droplets distributed in the film matrix together with CF + PVA. Paglione et al. [13] found similar irregularities and attributed them to the addition of OEO. With the increase in OEO addition ( Figure 2(E) and Figure 2(G) ) in the images at 100× magnification, it is possible to visualize an increase in the size of the EO droplets. At the highest OEO concentration, there was a greater distribution in the film matrix.

The REO-added films showed different structures, and the EO type, hydrophobicity, and size stability have been reported to affect the film matrix [44] . The OEO and REO showed different chemical compositions, and these differences may affect the oils’ solubility in different matrices. The addition of 2% REO ( Figure 2(I) and Figure 2(J) ) showed no changes. With increasing REO (6%), the film showed a rough and discontinuous surface. Unlike the films that added OEO, it was impossible to visualize the REO droplets, which can be attributed to a higher interaction of REO to the matrix.

The CF (collagen fiber) and PVA (polynivyl alcohol) films showed the typical bands characteristic of CF and PVA ( Figure 3 ). Some changes in the position of the bands were seen with the blending for the production of the CF + PVA films and with the addition of the OEO (oregano essential oil) and REO (rosemary essential oil). The amide A band represents N-H stretching vibration [7] . This was shifted at lower frequencies for the control film (3437.30), F (Film) + 2% OEO (3420.90), F + 4% OEO (3420.90), F + 6% OEO (3427.65), F + 2% REO (3437.30), F + 4% REO (3431.41), and F + 6% REO (3421.87) cm⁻¹ compared to CF (3445.98 cm⁻¹).

The PVA also showed a peak at 3447.91, which is related to the OH group [45] . The decrease is related to hydrogen bonds formed with the mixing of the materials [7] [41] . Hydrogen bonds are also involved in the process of network formation, which occurs due to partial denaturation [43] . Amide B is associated with the stretching of CH₂ [7] . We also observed the formation of new bands, and in films with OEO and REO, similar findings were found by Saricaoglu [41] . The amide I (~1700 - 1600 cm⁻¹) is sensitive to the secondary structure of the protein and is mainly known for the stretching vibration = CO [7] [45] . For CF + PVA films, lower values were found (1652.10 - 1642.46 cm⁻¹) compared to CF (1655.00 cm⁻¹). There were some changes in the conformation of collagen molecules due to the temperature used; similar data were reported by Xu et al. [7] . For amide II (~1600 - 1500 cm⁻¹), CF + PVA films also shifted to a higher frequency (1559.51 - 1555.66 cm⁻¹) compared to CF (1546.01 cm⁻¹); this band is characterized by C-N stretching and NH bending vibration [46] .

The amide III band is mainly assigned to CN stretching, NH bending vibrations, CC stretching, and CH bending. This band is also sensitive to the secondary structure of the protein and can provide meaningful answers regarding the structural conformation of collagen [46] . For the CF + PVA films, this band was shifted to higher values compared to CF (1241.25 cm⁻¹). For the control film (1256.68), F + 2% OEO (1257.64), F + 4% OEO (1256.68), F + 6% OEO (1256.68), F + 2% REO (1248.96), F + 4% REO (1254.75), and F + 6% REO (1258.61) cm⁻¹. These results show that despite the temperature used in the manufacture of the films (90˚C/20 min), there were no major modifications of the collagen fiber structure, and a part of the fibers was not denatured ( Figure 2 ).

Peaks referring to the same major compounds that were identified by GC/MS were found in the OEO- and REO-added films. Bands between 1500 - 1400 cm⁻¹ are attributed to the C = C stretching vibration of the aromatic ring and bending vibration of the isopropyl methyl group in the band 1381 cm⁻¹ for the p-cymene compound; for carvacrol and thymol, bands range from 1600 to 1585 cm⁻¹ [47] . A band at 1591.34 cm⁻¹ was only found in CF + PVA films for all added levels of OEO. The band at approximately 862 cm⁻¹ strongly influences OEO [48] . In CF + PVA films with 2%, 4%, and 6% OEO, bands of 854.50, 861.25, and 863.18 were found, respectively. As for the REO-added films, bands attributed to the compound p-cymene were also identified.