Development of an Active Almond Gum/Gelatin Bio‐Composite Film With Rosemary Essential Oil for Extending the Shelf Life of Refrigerated Chicken Fillets

ABSTRACT

This study investigated the development and optimization of a bioactive edible film based on almond gum–gelatin matrix incorporated with rosemary essential oil (REO) and plasticized with glycerol, using a Box–Behnken design (BBD). The optimization process proved high reliability with adjusted coefficients of determination (R ² _adj) ranging consistently between 0.986 and 0.999 across all studied responses. The optimized formulation (5% almond gum, 2% REO, and 24.8% glycerol) achieved a tensile strength of 29.23 MPa and an antioxidant activity (2,2‐diphenyl‐1‐picrylhydrazyl [DPPH]) of 79.012%. FTIR analysis of the optimized film showed a successful cross‐linking and intermolecular hydrogen bonding between biopolymers and REO, while DSC thermograms proved enhanced thermal stability of the composite matrix. Characterization revealed low opacity and high antimicrobial activity with > 16 mm of inhibition zones against Escherichia coli, Salmonella typhimurium, and Staphylococcus aureus, by the agar disk diffusion method. Applied on refrigerated chicken fillets for 12 days, the active film significantly suppressed microbial proliferation (TAB, Pseudomonas spp, and psychrotrophic bacteria) compared to control samples. It also significantly retarded chemical alteration, maintaining total volatile basic nitrogen (TVB‐N: 25.12 mg/100 g) and thiobarbituric acid reactive substances (TBARS: 0.38 mg malondialdehyde [MDA]/kg) within acceptable limits. These results confirm that the optimized active edible film can be an important transition as an active protection system, delaying microbial proliferation, sensory shelf life, and lipid oxidation of refrigerated poultry meat from 6 days (control) to 12 days.

1. Introduction

The global food industry is currently facing a paradigm shift by the urgent need to mitigate the environmental footprint of conventional plastic packaging. Traditional petroleum‐based polymers pose significant ecological threats due to their nonbiodegradability and the release of microplastics into the food chain (Ceballos‐Santos et al. ²⁰²⁴). In response, the development of bio‐based edible films and coatings has emerged as a basis of sustainable food science. These materials, derived from renewable resources such as proteins, polysaccharides, and lipids, offer a dual advantage: They reduce environmental waste and provide a functional matrix for active compounds that can actively interact with the food product to extend its shelf life (Raj et al. 2026).

Among the biopolymers investigated for film‐forming applications, gelatin occupies a noticeable position. As a fibrous protein obtained through the partial hydrolysis of collagen, gelatin exhibits exceptional film‐forming capacity, high transparency, and a natural ability to block oxygen (Sun et al. 2021). However, the inherent hydrophilicity of gelatin often results in poor moisture barrier properties and limited mechanical stability under high‐humidity conditions (Suderman et al. 2018). To overcome these limitations, several research trends have shifted toward the fabrication of composite materials. The blending of gelatin with polysaccharides has been shown to create more robust networks through hydrogen bonding and electrostatic interactions (Shi et al. 2022). Almond gum (AG) (from Prunus dulcis), a natural transparent exudate polysaccharide, represents an underutilized yet highly promising candidate for this purpose. Its primary composition is arabinose, galactose, and uronic acid (Mudasir and Sundaramoorthy 2016). It was successfully used as edible coating to extend the shelf life of tomato (Mahfoudhi et al. 2014), sweet cherry (Mahfoudhi and Hamdi 2015), fresh‐cut pineapple slices (Rasool et al. 2023), and blueberries (Suresh et al. 2024). When incorporated into a gelatin matrix, it acts as a structural reinforcer, potentially filling the voids in the protein network and reducing water vapor permeability (WVP) (Dehghani et al. 2022; Bostar and Hosseini ²⁰²⁰). The functionality of these films can be significantly enhanced through the incorporation of bioactive agents, transforming them into active packaging (Mirsharifi et al. 2023). Essential oils, derived from aromatic plants, are widely recognized for their antioxidant and antimicrobial activities. Rosemary essential oil (REO) (Rosmarinus officinalis L.) is particularly noteworthy due to its high concentration of volatile monoterpenes, especially eucalyptol (1,8‐cineole), alpha‐pinene, and camphor (Ribeiro‐Santos et al. ²⁰¹⁷). These constituents enable the oil to scavenge free radicals and disrupt the cellular membranes of foodborne pathogens (Alves et al. 2023). However, the direct application of essential oils to food is often limited by their volatility and strong sensory profile. Incorporating REO into a bio‐composite film allows for a controlled release mechanism, maintaining a high concentration of active molecules at the food‐package interface while minimizing adverse impacts on the food's organoleptic properties (El Bayomi et al. 2023).

The application of such active films is particularly critical for highly perishable matrices such as poultry meat. Chicken meat is highly susceptible to microbial spoilage and lipid oxidation, which lead to off‐odors, color, deterioration, and a significant reduction in shelf life (Musalem et al. 2025).

While recent studies have explored AG–based composite films for food packaging, notably Mirsharifi et al. (2023), who developed AG–PVA–chitosan films incorporated with thyme essential oil for chicken preservation, the present work introduces a mechanistically distinct approach. The substitution of synthetic PVA and cationic chitosan by gelatin enables the formation of a semi‐interpenetrating polymeric network (semi‐IPN) stabilized by hydrogen bonding between AG chains and gelatin peptide bonds, yielding a structurally more cohesive matrix. Furthermore, REO was selected for its bioactive profile, combining antimicrobial monoterpenes (1,8‐cineol, camphor, α‐pinene) with potent antioxidant phenolic derivatives, providing a dual preservation mechanism targeting both microbial growth and lipid oxidation simultaneously.

Based on these considerations, we hypothesized that (i) the blending of AG and gelatin would create a synergistic macromolecular network, significantly improving the mechanical and barrier properties of the composite matrix and (ii) REO components would inhibit both lipid oxidation and specific spoilage microorganisms on chicken breast fillets during refrigerated storage. Consequently, the objective of the present study is to develop a novel active composite film based on AG and gelatin and REO. The optimization was achieved using a Box–Behnken experimental design. And to evaluate its efficacy, the optimized active film was used as an edible film to control the physicochemical and microbiological quality of chicken breast fillets during storage at 4°C.

2. Materials and Methods

2.1. Materials

The primary biopolymers used in this study consisted of AG and Type B gelatin. AG was collected manually from almond trees (Prunus dulcis) located in the region of Sidi Bouzid, Tunisia, then purified according to the method described by Mahfoudhi et al. (2014), with slight modifications. Briefly, the raw gum was first manually cleaned to remove bark fragment and visible impurities, then dissolved in distilled water (1:10, w/v) under continuous magnetic stirring at room temperature for 24 h to ensure complete solubilization. The resulting solution was filtered through fine cloth to remove insoluble residues and plant debris. The filtrate was then subjected to different precipitations by adding ethanol (1:2, v/v) under continuous stirring, and the resulting precipitate was recovered by centrifugation. The purified gum was subsequently washed with ethanol to remove residual pigments and low‐molecular weight impurities, dried in a ventilated oven at 35°C for 24 h, and ground into a fine powder using a laboratory mill. The resulting purified AG powder was stored in airtight containers at room temperature until further use.

Prior to film formulation, the proximate composition of this purified AG lot was determined in triplicate according to the AOAC methods (AOAC ²⁰²³). Results showed that the purified gum exhibited a moisture content (MC) of 12.16% ± 0.42%, a protein content of 1.12% ± 0.03%, a fat content of 0.21% ± 0.02%, and ash content of 3.24% ± 0.07%.

Type B gelatin (G), derived from bovine skin (Bloom strength: 225 g), was purchased from Sigma–Aldrich (St. Louis, Missouri, USA). REO (Rosmarinus officinalis), with a density of 0.92 ± 0.01 g/cm³ at 20°C (determined by pycnometry), was previously extracted by a standard hydrodistillation method and stored in amber glass vials at 4°C to prevent photooxidation of its phenolic constituents. The chemical composition of the REO essential oil was determined by GC–MS after 2 days of extraction, and the characterized oil was used for film formulation within the following 3 days. All used chemicals were of analytical grade. Glycerol (purity > 99.5%), used as a plasticizer, was obtained from Merck (Darmstadt, Germany). For the assessment of antioxidant capacity, 2,2‐diphenyl‐1‐picrylhydrazyl (DPPH) and methanol, were purchased from Sigma–Aldrich. For microbiological assays, dehydrated culture media, were sourced from Oxoid (Basingstoke, UK). All aqueous solutions were prepared using deionized water. Fresh chicken breast fillets were purchased from a local poultry processing plant within 2 h of slaughter. The samples were immediately placed in a sterile polyethylene (PET) bag and transported to the laboratory in insulated polystyrene boxes containing ice packs to maintain a temperature below 4°C. Upon arrival, the meat was inspected for freshness, and any invisible fat or connective tissue was aseptically removed prior to the coating application.

2.2. Gas Chromatography–Mass Spectrometry (GC–MS)

The volatile chemical profile of REO was analyzed using an Agilent 6890N Network Gas Chromatography system coupled with Agilent 5975B inert mass selective detector (quadruple analyzer). Ionization was achieved via electron impact mode at 70 eV. Volatile constituents were separated on a capillary column coated with 5% phenyl methyl siloxane (HP‐5MS, 30 m × 0.25 mm internal diameter, 0.25 µm film thickness). The oven temperature program was initiated at 50°C and progressively raised to 280°C at a constant heating rate of 7°C/min. Helium served as carrier gas, adjusted to maintain a constant linear velocity of 34 cm/s). The mass spectrometer operated with a scan time of 2.2 s over a mass‐to‐charge ratio scanning range of 50–550 m/z. The identification of individual volatile constituents was achieved comparing their retention indices (RI), relative to a series of n‐alkanes, and by matching their mass spectra fragmentation profiles with data from the Wiley and NIST mass spectral libraries.

2.3. Experimental Design: Optimization of the Formulation of the Active Film

To investigate the effects of formulation variables on the functional properties of the active films, a three‐factor, three‐level Box–Behnken design (BBD) was employed. The three independent variables investigated were AG concentration (X ₁: 3%–5% w/v), REO concentration (X ₂: 0.5%–2% v/v), and glycerol concentration (X ₃: 15%–30% w/w of total biopolymer weight). The concentration of gelatin was maintained constant at 4% (w/v) in all formulations to serve as a stable structural protein matrix. This concentration was chosen based on a preliminary study, where concentrations ranging from 2% to 7% (w/v) were tested. Gelatin concentrations below 3% produced mechanically fragile, difficult‐to‐handle films, while concentrations above 5% resulted in overly viscous solutions with casting defects (air bubble entrapment, nonuniform thickness). A concentration of 4% (w/v) was identified as providing the best overall compromise and was therefore selected as a fixed parameter. Moreover, as AG concentration (X ₁) varied between 3% and 5% (w/v) while gelatin remained constant at 4% (w/v), the total biopolymer mass ranged from 7% to 9% (w/v), and the absolute glycerol mass at a given coded level of X ₃ was calculated accordingly.

The experimental design consisted of 15 runs, including 12 factorial points and three central points to estimate pure experimental error and ensure the reproducibility of the process (Table 1). The experimental runs were performed in the standard run order generated by Design‐Expert software, version 13, under tightly controlled laboratory conditions to minimize time‐dependent systematic variation between runs. For each of the 15 formulations, one independent film batch was prepared, and each response variable was measured in triplicate per batch.

TABLE 1.

Coded and uncoded values of independent variables and their levels used for the Box–Behnken design.

Independent variables
Assays	Coded variables			Uncoded variables
	Almond gum (X 1)	Rosemary essential oil (X 2)	Glycerol (X 3)	Almond gum (% w/v)	Rosemary essential oil (% v/v)	Glycerol (% w/w)
1	−1	0	−1	3	1.25	15
2	−1	0	1	3	1.25	30
3	0	0	0	4	1.25	22.5
4	1	1	0	5	2	22.5
5	0	0	0	4	1.25	22.5
6	0	−1	1	4	0.5	30
7	−1	1	0	3	2	22.5
8	0	−1	−1	4	0.5	15
9	1	0	1	5	1.25	30
10	1	0	−1	5	1.25	15
11	−1	−1	0	3	0.5	22.5
12	0	1	1	4	2	30
13	1	−1	0	5	0.5	22.5
14	0	1	−1	4	2	15
15	0	0	0	4	1.25	22.5

2.4. Mathematical Modeling and Validation

The relationship between the independent factors and the experimental responses (Y) (Table 2), specifically WVP, tensile strength (TS), elongation at break (EAB), and DPPH radical scavenging activity, was modeled using a second‐order polynomial equation:

$$Y={\beta }_0+\sum _{i=1}^3{\beta }_i{X}_i+\sum _{i=1}^3{\beta }_{ii}{X}_i^2+\sum _{i=1}^3\sum _{j=i+1}^3{\beta }_{ij}{X}_i{X}_j$$

where Y represents the predicted response; β ₀ is the intercept; β _i, β _ii, and β _ij are the linear, quadratic, and interaction coefficients, respectively; and X _i, X _j are the coded levels of the independent variables. The adequacy of the generated models was evaluated via analysis of variance (ANOVA) at a 95% confidence level, focusing on the coefficient of determination (R ²) and the significance of the lack‐of‐fit test (Azarifar et al. 2019), with the variability among the three replicated central point runs contributing to the estimation of pure experimental error used in this evaluation.

TABLE 2.

Box–Behnken experimental design with independent variables (uncoded) and observed values of responses variables.

Uncoded variables				Responses
Runs	Almond gum (% w/v)	Rosemary essential oil (% v/v)	Glycerol (% w/w)	WVP (10−9 g/m.s.Pa)	DPPH (%)	TS (MPa)	EAB (%)
1	3	1.25	15	4.15	45.3	32.8	18.4
2	3	1.25	30	6.2	44.1	18.5	55.2
3	4	1.25	22.5	4.25	46.8	28.5	28.4
4	5	2	22.5	3.45	78.2	29.5	25.8
5	4	1.25	22.5	4.22	47.1	28.2	29.1
6	4	0.5	30	5.45	17.5	21.3	48.5
7	3	2	22.5	5.12	75.4	22.1	42.4
8	4	0.5	15	3.42	18.8	36.4	15.2
9	5	1.25	30	4.75	47.8	24.1	35.6
10	5	1.25	15	2.85	48.6	39.7	12.5
11	3	0.5	22.5	4.85	18.2	26.4	38.5
12	4	2	30	5.92	74.2	19.4	52.1
13	5	0.5	22.5	3.12	19.5	34.2	22.1
14	4	2	15	3.78	76.5	31.2	19.8
15	4	1.25	22.5	4.3	46.5	28.7	27.8

Numerical optimization was conducted using the desirability function to identify the specific formulation providing the best compromise between mechanical integrity, moisture barrier efficiency, and maximum antioxidant potency. Specifically, the following individual optimization goals were defined: TS and EAB were set to “maximize,” WVP was set to “minimize,” and DPPH radical scavenging activity was set to “maximize.” The optimized formulation was then fabricated and characterized to validate the model's predictive accuracy before proceeding with the poultry meat preservation study.

2.5. Film Preparation

Active films were prepared using the solvent casting technique. First, the gelatin solution (4% w/v) was prepared by dispersing Type B gelatin in deionized water under constant magnetic stirring at 50°C for 30 min until complete dissolution. Simultaneously, AG solutions were prepared by dissolving the gum powder in deionized water (3%, 4%, or 5% w/v) at room temperature for 12 h to ensure full hydration of the polysaccharide chain. The two solutions were blended in 1:1 ratio under vigorous stirring for 15 min at 45°C. Glycerol (15%, 22.5%, or 30% w/w of total biopolymers) was added as a plasticizer and stirred for an additional 20 min. To incorporate the bioactive phase, REO (0.5%, 1.25%, or 2% v/v) was added to the mixture. To promote the homogeneous dispersion of REO within the hydrophilic film‐forming matrix, the resulting mixture was homogenized using an ultra‐turrax (T25 digital, IKA, Germany) at 15,000 rpm for 2 min. Finally, the film‐forming solutions were degassed under vacuum to remove air bubbles. A fixed volume of each solution (25 mL) was cast onto level Teflon‐coated Petri dishes (90 mm diameter) to ensure uniform film thickness across all formulations. The films were dried at 30°C and 50% relative humidity for 48 h, in a temperature and humidity‐controlled chamber. After drying, they were peeled from the plates and conditioned at 25°C and 50% relative humidity maintained using a saturated magnesium chloride solution (Mg (NO₃)₂) for 48 h prior to analyses (Dong et al. 2023).

2.6. Characterization of Active Films

Films prepared according to the Box–Behnken matrix were characterized for their WVP, TS, EAB, and antioxidant activity. However, the optimized active film, in addition to those properties, was characterized for its visual aspect and color (L ^*, a ^*, b ^*, and opacity), MC, and water solubility (WS), FTIR, DSC, and its antimicrobial properties.

2.6.1. Film Thickness (L)

The thickness of the films was measured using a digital micrometer (model 395‐261‐30, Mitutoyo, Japan) with a precision of 0.01 mm. Measurements were taken at five random positions on each film sample, and the mean value was used for the calculation of mechanical and barrier properties.

2.6.2. Mechanical Properties

TS and EAB were determined using a texture analyzer (stable Micro system, Godalming, UK) according to the ASTM D882 standard method. Film strips (100 × 20 mm) were mounted between the grips with an initial separation of 50 mm and a crosshead speed of 50 mm/min. Prior to mechanical evaluation, all films were strictly equilibrated for 48 h in a hermetic desiccator maintained at 25°C and 50% relative humidity.

2.6.3. WVP

The WVP was determined gravimetrically following the modified ASTM E96 method. Films were sealed over the circular opening of glass permeation cells containing anhydrous calcium chloride (0% RH). The cells were then placed in a desiccator maintained at 25°C and 75% RH (using saturated NaCl solution). The weight gain of the cells was recorded every 2 h for 24 h. The WVP was calculated using the following equation:

$$\mathrm{WVP}=\frac{\mathrm{WVTR}\times L}{\mathrm{\Delta }P}$$

where WVP is the water vapor permeability (g/m.s.Pa), WVTR is the water vapor transmission rate (g.m⁻². s⁻¹), L is the film thickness (m), and ΔP is the partial water vapor pressure difference across the film (1753.55 Pa at 25°C).

2.6.4. Antioxidant Activity (DPPH Scavenging Assay)

The antioxidant potency of the films was determined using the DPPH radical scavenging assay (Wu et al. 2024; Mirsharifi et al. ²⁰²³). Briefly, 50 mg of each film was dissolved in 5 mL of distilled water. Then, 1 mL of the film solution was mixed with 3 mL of a methanolic DPPH solution (0.1 mM). The mixture was incubated in the dark at room temperature for 30 min. The absorbance was measured at 517 nm using a UV–Vis spectrophotometer. The radical scavenging activity was calculated as

$$\mathrm{DPPH}\mathrm{scavenging}\mathrm{activity}\left(\%\right)=\frac{A_{control}-A_{sample}}{A_{control}}\times 100$$

where A _control is the absorbance of the DPPH solution without the film extract and A _sample is the absorbance of the DPPH solution with the film extract.

2.6.5. Visual Aspect and Color Properties

The visual appearance (homogeneity, transparency, and surface smoothness) of the optimized film was qualitatively assessed. The surface color was measured using a colorimeter (CR‐400, Konica Minolta, Japan). The L ^* (lightness), a ^* (redness/greenness), and b ^* (yellowness/blueness) values were recorded. The opacity (O) was calculated according to the following equation:

$$O=\frac{A_{600}}{L}$$

where A ₆₀₀ is the absorbance at 600 nm and L is the film thickness (mm).

2.6.6. MC and WS

MC of the bioplastic film was determined by drying circular film samples (2 cm in diameter) at 105°C until constant weight was reached. MC was calculated as the percentage of water removed from the samples. These dried films were then used to determine the WS. They were immersed in 50 mL of distilled water at 20°C with agitation. Then, the insoluble fragments of the film were collected and dried until constant weight at 105°C. They were used to evaluate the weight of dry residues not soluble in water.

2.6.7. Antimicrobial Activity

The antimicrobial activity of the optimized film was evaluated using the agar disk diffusion method (Seol et al. 2009) against major poultry related pathogens: Escherichia coli, Staphylococcus aureus, and Salmonella typhimurium. Circular disks of the optimized film (6 mm diameter) were placed on Mueller–Hinton agar plates previously inoculated with 0.1 mL of bacterial suspension (10⁵ CFU/mL). This inoculum level was lower than the CLSI standard used in clinical susceptibility testing, as this test has been used, in this study, as indicative of the film's antimicrobial potential rather than as clinically standardized susceptibility data. After incubation at 37°C for 24 h, the diameter of the inhibition zones (mm) was measured using a digital caliper. Films without REO served as negative controls to distinguish the effect of the active agent from the polymer matrix.

2.6.8. Differential Scanning Calorimetry (DSC) Characterization

Thermal properties of the optimized active film were evaluated using DSC. Analyses were performed using a Pyres 6 DSC instrument (PerkinElmer), connected to a computer equipped with dedicated software for data acquisition and analysis. Around 5 mg of the film was weighed into an aluminum crucible and heated. The film sample was heated from an initial temperature of 20°C to an ultimate temperature of 200°C at a standardized, linear heating velocity of 10°C/min. The characterization was performed in triplicate to extract the precise thermal properties of the active film.

2.6.9. Fourier Transform Infrared Spectroscopy (FTIR) Analysis

The FTIR spectrum of the optimized active film was produced using a Fourier transform infrared spectrometer (Model Nicolet iN10, Thermo Fisher Scientific, USA). Spectra were recorded in transmission mode within the wavenumber range of 4000–500 cm⁻¹.

2.7. Application of the Active Edible Film for Chicken Meat Preservation

For the application of the active film, chicken fillets were divided into 50‐g portions using a sterile scalpel under aseptic conditions. They were standardized into uniform rectangular cuboid shapes to ensure homogeneous surface throughout the control and wrapped samples.

The samples were individually wrapped with optimized active films, following the method described by Safari et al. (2025). Samples were cut into pieces providing a film surface area sufficient to completely wrap each fillet portion on all sides. Each fillet was individually wrapped by placing it at the center of a film piece and folding the film edges over the surface of the meat, ensuring direct and uniform contact between the film and all exposed meat surface. The film to meat interface was maintained under slight pressure throughout storage by placing each wrapped sample into a sealed sterile PET bag prior to storage. Control groups were prepared by leaving the chicken fillets unwrapped. Samples were then placed into a sterile PET bag and stored at 4°C for a total duration of 12 days, under normal aerobic conditions. During this period, samples were analyzed to evaluate pH, the microbiological properties, total volatile base nitrogen (TVB‐N), peroxide value (PV), thiobarbituric acid reactive substances (TBARS), and sensorial properties, at Day 0, then after 2, 4, 6, 9, and 12 days. At each targeted storage interval, three independent packaging replicates from each respective batch were randomly gathered and opened for destructive sampling. From each package, physicochemical and microbiological evaluations were executed in triplicates and results were reported as mean ± standard deviation.

To assess the microbial evolution during storage, 10 g of chicken meat samples were aseptically homogenized with 90 mL of sterile 0.1% peptone water using a stomacher for 120 s. Following the preparation of serial decimal dilutions, specific microbial populations, including aerobic plate count (APC) and psychrotrophic bacteria (PB) was enumerated on plate count agar (PCA); Pseudomonas spp. (PSE) was isolated on Pseudomonas isolation agar supplemented with glycerol; lactic acid bacteria (LAB) was determined on de Man, Rogosa and Sharpe (MRS) agar; S. aureus was enumerated on Baird‐Parker agar supplemented with egg yolk tellurite emulsion; Enterobacteriaceae were counted on Violet Red Bile Glucose (VRBG) Agar; and filamentous fungi (yeasts and molds) were determined on Oxytetracycline Glucose Yeast Extract (OGYE) Agar, in accordance with the protocols described by Kamkar et al. (2021).

However, the chemical deterioration of the chicken breast fillets was evaluated through the determination of TVB‐N (ISO ²⁰²⁶), PV (ISO ²⁰¹⁷), and TBARS according to the method described by Nam and Ahn (2003). Moreover, the sensory analysis was evaluated during refrigerated storage by a trained panel of 12 members. Panelists evaluated the raw samples immediately after opening the packages to assess color, odor, and overall acceptability. Evaluation was performed using a nine‐point hedonic scale (ranging from 9 = extremely desirable to 1 = completely unacceptable).

2.8. Statistical Analysis

The optimization of the formulation of the active film was carried out using Design‐Expert Software version 13. Moreover, for all analysis, results are presented as mean ± standard deviation of at least three independent replicates. Statistical analysis was performed using ANOVA using SPSS (version 23). For comparison between the control and coated chicken breast fillets at each storage time, a one‐way ANOVA followed by Tukey's post‐hoc test was applied.

3. Results and Discussion

3.1. Chemical Composition of the REO

The chemical composition of REO used in this study, as determined by GC–MS, is presented in Table 3. A total of 12 compounds were identified, accounting for 93.7% of the total oil composition. The oil was dominated by 1,8‐cineole (42.4%), followed by α‐pinene (11.3%), camphor (10.6%), borneol (3.8%), α‐terpineol (4.6%), and camphene (3.6%), while minor constituents of β‐pinene, myrcene, linalool, terpinene‐4‐ol, and bornyl acetate, each represented less than 2% of the total composition. This composition profile corresponds to a 1,8‐cineol dominant chemotype, which has been consistently identified as one of the major chemotypes of Tunisian Rosmarinus officinalis populations, with cineole content reported to vary widely (from 29% to 60%) depending on geographical origin, bioclimatic conditions, harvest season, and so forth. (Dhouibi et al. 2023). Recent biochemical investigations confirm that 1,8‐cineole acts as a powerful disruptive agent against both Gram‐negative and Gram‐positive foodborne pathogens (Shektaei et al. 2023).

TABLE 3.

Chemical composition of Rosmarinus officinalis essential oil determined by GC–MS.

Components	RI	Relative area (%)
α‐Pinene	940	11.3
Camphene	953	3.6
β‐Pinene	978	0.9
Myrcene	1002	0.7
1,8‐Cineole	1046	42.4
Linalool	1091	1.1
Camphor	1104	9.6
Borneol	1158	3.8
Terpinen‐4‐ol	1175	1.0
α‐Terpinol	1185	4.6
Verbenone	1203	0.0
Bornyl acetate	1224	0.2
Total identified fraction	—	93.7

3.2. Optimization of Active Edible Coating Formulation

The experimental data obtained from the BBD were analyzed using response surface methodology (RSM) to evaluate the individual and interactive effects of AG (X ₁), REO (X ₂), and glycerol (X ₃) on each response. The adequacy of the models was assessed using ANOVA, as summarized in Tables 4 and 5.

TABLE 4.

Coefficients of regression for the models adjusted for water vapor permeability (WVP), antioxidant activity (DPPH), tensile strength (TS), and elongation at break (EAB) of the active edible films.

	WVP	DPPH	TS	EAB
Model	4.26	46.80	28.47	28.43
A	−0.7687	1.39	3.46	−7.31
B	0.1787	28.79	−2.01	1.97
C	1.02	−0.7000	−7.10	15.69
AB	0.0150	0.3750	−0.1000	−0.0500
AC	−0.0375	0.1000	−0.3250	−3.42
BC	0.0275	−0.2500	0.8250	−0.2500
A2	−0.1383	0.3625	0.6417	0.1458
B2	0.0167	0.6625	−1.06	3.62
C2	0.3692	−0.7125	−0.3333	1.85
Lack of fit (p‐value)	0.08	0.1437	0.076	0.104
R 2	0.995	0.999	0.995	0.996
Adjusted R 2	0.986	0.999	0.986	0.986
Predicted R 2	0.924	0.996	0.926	0.929
Adequate precision	37.263	123.13	35.78	36.39
C.V. (%)	2.67	1.29	2.58	4.93

TABLE 5.

p‐values obtained by the analysis of variance (ANOVA) of the film's properties.

	WVP (g/m.s.Pa)	DPPH (%)	TS (MPa)	EAB (%)
Model	< 0.0001	< 0.0001	< 0.0001	< 0.0001
A	< 0.0001	0.0013	< 0.0001	< 0.0001
B	0.0076	< 0.0001	0.0005	0.0154
C	< 0.0001	0.0222	< 0.0001	< 0.0001
AB	0.8083	0.2704	0.7932	0.9510
AC	0.5506	0.7545	0.4099	0.0069
BC	0.6588	0.4465	0.0714	0.7598
A2	0.0727	0.3019	0.1489	0.8635
B2	0.7957	0.0894	0.0374	0.0064
C2	0.0018	0.0732	0.4163	0.0706

For all evaluated responses, WVP, DPPH, TS, and EAB, the quadratic models were found to be highly significant, with p‐values < 0.0001. This indicates that the probability of the observed results being due to noise is less than 0.01%. Moreover, the lack to fit for each model was not significant (p > 0.05), confirming that the second‐order polynomial equations are suitable for predicting the responses within the experimental range (Asfaw et al. 2023; Herrera‐Vázquez et al. ²⁰²²). Furthermore, the determination coefficients (R ²) ranged from 0.995 to 0.999, indicating that the quadratic models can explain over 99.5% of the experimental variability. Moreover, the adjusted coefficients of determination (R ² _adj) fell within a highly tight and elevated range of 0.986–0.999, indicating that more than 98.6% of the experimental variance is accurately explained by the operational factors. This is strongly supported by the predicted coefficients of determination (R ² _pred), which varied from 0.924 to 0.996. Finally, the adequate precision values, measuring the signal‐to‐noise ratio, were profoundly higher than the minimum desirable threshold of 4.0, spanning from 35.78 to 123.13. On the other hand, with remarkably low coefficients of variance (C.V. ≤ 4.93%), it can be confirmed that the developed models possess high reliability, low experimental noise, and absolute precision for navigating the response surface design space.

3.3. Effect of Independent Variables on the Film's Properties

3.3.1. WVP

WVP is an important parameter for evaluating the protective efficacy of edible coatings and films against moisture migration (Nandane et al. 2017). These matrices act as external barriers that significantly decrease the dehydration rates of coated food products, which preserve their quality. The polynomial model for WVP, as derived from the experimental data, is expressed by the following second‐order equation:

$$\begin{array}{ccc}\hfill \mathrm{WVP}& =& 4.26-0.768X_1+0.178X_2+1.02X_3+0.015X_1{X}_2\hfill \\ & & -0.037X_1{X}_3+0.027X_2{X}_3-0.138X_1^2+0.016X_2^2\hfill \\ & & +0.37X_3^2\hfill \end{array}$$

Statistical analysis (Tables 4 and 5) shows that WVP is primarily governed by the antagonistic effect of AG (X ₁) which had a negative linear coefficient (−0.768) proving a significant enhancement of the film's barrier properties. This result indicates that increasing gum concentration effectively enhances the barrier properties of the active film. As supported by the FTIR analysis, this improvement is attributed to the formation of a dense interpenetrating polymer network (IPN) where the gum's polysaccharides fill the voids within the gelatinous protein scaffold (Reichert et al. 2020). This interaction, often driven by hydrogen bonding and electrostatic forces between the amino groups of gelatin and the hydroxyl groups of the gum, creates a tortuous path that significantly hinders the diffusion of water molecules (Ma et al. 2026). Moreover, recent studies confirm that natural gums act as structural fillers that restrict chain mobility and reduce the free volume available for gas and vapor diffusion (Seididamyeh et al. 2024). Interestingly, REO (REO: X ₂) showed a minor positive effect on WVP with a positive coefficient of +0.178. In general, the incorporation of essential oils in the formulations of edible films leads to decrease in the WVP of the film because of the growth of hydrophobic fraction in the film (Liu et al. 2023). However, their incorporation as droplets in a hydrophilic matrix can disrupt the network continuity, creating micro‐voids at the interface that paradoxically allow easier water vapor passage. This structural disruption is strongly corroborated by our FTIR analysis (Figure 4), where the characteristic aliphatic C–H stretching bands of REO (2925 cm⁻¹ and 2975 cm⁻¹) were successfully identified, proving the physical entrapment of lipophilic domains within the continuous polar matrix without altering its chemical core.

FIGURE 4.

FTIR spectra of the optimized almond gum–gelatin film.

Moreover, glycerol (X ₃) showed a strong positive effect (+1.02) on WVP. As a small, hygroscopic molecule, glycerol intercalates between the polymer chains, increasing the free volume and reducing the density of the matrix, which increases WVP (Asfaw et al. 2023; Suderman and Sarbon 2020). This plasticizing effect facilitates the mobility of polymer segments, thereby promoting moisture migration through the film (Asfaw et al. 2023; Suhag et al. ²⁰²⁰). Moreover, the significant quadratic effect of glycerol (C ² = +0.369) further suggests that excessive plasticization leads to a nonlinear loss of barrier efficiency, likely due to phase separation or structural saturation within the biopolymer blend (Cabello‐Santos et al. ²⁰¹⁵; Tarique et al. ²⁰²¹).

3.3.2. Antioxidant Activity

The incorporation of REO changes conventional coatings into functional active packaging systems (Kawhena et al. 2021). The DPPH response quantifies the film's efficiency to neutralize free radicals, which is crucial for inhibiting surface oxidation and preserving nutritional properties of the wrapped products. The active functionality of the film was modeled through the DPPH radical scavenging response, resulting in the following equation:

$$\begin{array}{ccc}\hfill \mathrm{DPPH}& =& 46.8+1.39X_1+28.79X_2-0.7X_3+0.375X_1{X}_2\hfill \\ & & +0.1X_1{X}_3-0.25X_2{X}_3+0.36X_1^2+0.66X_2^2-0.712X_3^2\hfill \end{array}$$

The dominance of factor X ₂ (REO) is statistically great, with a linear coefficient of 28.79 (p < 0.0001) (Tables 4 and 5). This confirms that the REO serves as the primary bioactive agent, effectively donating hydrogen atoms to neutralize DPPH radicals. AG (X ₁) also contributed to a slight positive linear and quadratic effect, which may be due to its endogenous bioactive polysaccharides or a synergistic stabilization of the oil droplets within the matrix, preventing their oxidation during the drying process. This protective retention is strongly shown by the FTIR profile (Figure 4), where the preserved intensity of the lipophilic terpene bands within the hydrocolloid continuous phase confirms that the core antioxidant volatile fractions were successfully protected from thermal degradation.

However, glycerol showed a negative effect on the overall antioxidant property of the films. This can be attributed to the increased molecular mobility and MC in highly plasticized films, which may accelerate the degradation of the oil's sensitive active molecules or increase their migration toward the film's surface, leading to their volatilization. The interaction between AG and REO (X ₁ X ₂ = +0.375) (Table 4) suggests that the emulsifying property of the gum helps in obtaining a uniform distribution of the oil. By creating stable emulsion in the film‐forming solution, gum protects the hydrophobic compounds of the REO from phase separation, ensuring that the final active film maintains a consistent antioxidant potential across its entire surface (Kawhena et al. 2021). This stabilization aligns with the FTIR fingerprint zone, where the blending of the glycosidic C–O–C bands with the amphiphilic protein matrix provide direct structural evidence of a cohesive IPN capable of mechanically fixing the active essential oil domains.

3.3.3. Mechanical Properties

The mechanical performance of the optimized active films was evaluated through TS which showed the mechanical persistence of the biopolymeric film under external stress (Rao et al. 2010; Herrera‐Vázquez et al. 2022) and EAB, which measures its plastic deformation capacity without premature fracture (Suderman and Sarbon 2020).

The predictive mathematical behaviors of these dual responses are represented by the following second‐order polynomial equations:

$$\begin{array}{ccc}\hfill \mathrm{TS}\left(\mathrm{MPa}\right)& =& 28.47+3.46X_1-2.01X_2-7.1X_3-0.1X_1{X}_2\hfill \\ & & -0.325X_1{X}_3+0.825X_2{X}_3+0.641X_1^2-1.06X_2^2\hfill \\ & & -0.33X_3^2\hfill \end{array}$$

$$\begin{array}{ccc}\hfill \mathrm{EAB}\left(\%\right)& =& 28.43-7.31X_1+1.97X_2+15.69X_3-0.05X_1{X}_2\hfill \\ & & -3.42X_1{X}_3-0.25X_2{X}_3+0.1458X_1^2+3.62X_2^2\hfill \\ & & +1.85X_3^2\hfill \end{array}$$

The alternative sign of the linear coefficients establishes a powerful structural conflict within the multicomponent matrix. Statistical analysis indicates that TS is primarily increased by the reinforcing effect of AG (X ₁ = +3.46; p < 0.0001), which induces a strong reduction in film flexibility (X _1(EAB) = −7.31; p < 0.0001). This diametric response suggests that the branched polysaccharide acts as a rigid structural filler within the gelatinous matrix. Mechanistically, the high Bloom strength (225 g) of the gelatin reflects an abundance of intact triple‐helix peptide segments that actively entangle with the polysaccharide chains of the wild gum. This can be proven by FTIR spectrum, where the intense shifting of the Amide A band (3300 cm⁻¹) and the sharp consolidation of the Amide I (1645 cm⁻¹) and Amide II (1540 cm⁻¹) zones provide direct structural evidence of an extensive hydrogen‐bonding network. Conversely, the glycerol plasticizer (X ₃) exerted the most dominant inverse impact, heavily deteriorating TS (−7.1; p < 0.0001) while driving a massive synergistic surge in ductility (+15.69; p < 0.0001). As a low‐molecular‐weight polyol, glycerol intercalates between the biopolymer chains and disrupts native inter‐macromolecular cohesive junctions (Asfaw et al. 2023). On a molecular scale, this plasticization is validated by the broad hydroxyl envelope in the FTIR spectrum, confirming that glycerol shares hydrogen‐bonding sites with the protein–polysaccharide scaffold, thereby acting as a molecular lubricant. This reduces internal friction, and enhances segmental mobility (Suderman and Sarbon 2020). Similar plasticizing softening has been well‐documented for starch films (Muscat et al. 2012) and Persian gum matrices (Pak et al. 2020). Finally, the incorporation of REO (X ₂) also altered the structural network, weakening TS (−2.01; p = 0.0005) but favoring a nonlinear co‐plasticizing flexibility showed by an increase in EAB (X ₂ ² = +18.42; p = 0.0064). Concurrently, the positive interaction between glycerol and REO (X ₂ X ₃ = +0.825) points toward a slight interfacial stabilization effect, where the polyol improves the matrix adhesion around the oil domains (Al‐Harrasi et al. ²⁰²²).

3.4. Response Surface Plots

The 3D surface plots (Figure 1) provide a spatial visualization of the synergistic and antagonistic effects between the three factors: AG (X ₁), REO (X ₂), and glycerol (X ₃).

FIGURE 1.

Response surface plots showing the interaction effects of independent variables (X ₁: almond gum and X ₂: REO) on (a) WVP, (b) DPPH, (c) TS, and (d) EAB. In all subplots, the third process variable (X ₃: glycerol concentration) was held constant at its central point.

For the WVP (Figure 1a), the 3D surfaces illustrate that WVP is most sensitive to the concentrations of AG and glycerol. A steep downward slope is observed toward the higher limits of AG (5%), suggesting that the gum acts as a structural densifier. This is likely due to the formation of a complex interpenetrating network with gelatin, which increases the tortuosity of the path that water molecules must navigate (Herrera‐Vázquez et al. ²⁰²²; Asfaw et al. ²⁰²³). Conversely, the surface rises sharply as glycerol increases, confirming its role in expanding the macromolecular distance and facilitating moisture diffusion (Nwaka et al. 2025).

For the antioxidant activity of edible films (DPPH), the 3D plot (Figure 1b) is primarily dominated by the concentration of REO. The response surface shows a rapid, near‐linear REO content increase from 1% to 2%. The slight curvature suggests a synergistic interaction with the AG, which may be acting as an emulsifying stabilizer, protecting the phenolic volatiles of the oil during the film‐forming process and ensuring a high radical scavenging capacity.

Moreover, the surface plots for TS (Figure 1c) and EAB (Figure 1d) show opposite trends, a classic phenomenon in biopolymer science. The TS plot reaches its zenith at high AG and low glycerol concentrations, where the matrix is most rigid. In contrast, the EAB surface peaks at high glycerol levels (approximately 25%), reflecting the transition from a brittle glass‐like state to a ductile, rubbery state as the plasticizer reduces intermolecular friction (Nandane et al. 2017). Although the quadratic terms for REO (X ₂ ²) are statistically significant for both TS (p = 0.037) and EAB (p = 0.0064), the visual curvature in their 3D plots appears minimal. This behavior is due to the dominance of the steep linear effects and the fact that the multi‐response optimum lies near the boundary of the design space. This boundary positioning also accounts for the localized validation error observed for EAB (9.16%, R ² = 0.996), as prediction variance naturally expands near the geometric limits of the Box–Behnken matrix.

3.5. Statistical Analysis and Model Fitting

The adequacy and statistical validity of the established second‐order polynomial models were confirmed through residual diagnostic analysis, including normal probability and homoscedasticity plots, which are fully presented and discussed in Figure S1.

3.6. Numerical Optimization and Experimental Validation

To identify the optimal film composition and satisfy the target criteria for all investigated parameters, a multi‐response surface optimization strategy was implemented. The numerical optimization was executed by defining a desirability function focused on four primary performance indicators: (i) minimizing WVP, (ii) maximizing DPPH, and (iii) maximizing both TS and EAB. The objective is to find a gold standard formulation: an edible film that is strong yet flexible, with high antioxidant activity and low moisture permeability. The software identified the optimal coordinates as 5% AG, 2% REO, and 24.827% glycerol, achieving a desirability value of 0.99. The optimized active film (5% AG, 2% REO, and 24.827% glycerol) with 52.36 ± 1.5 (µm) of thickness (Table 7), exhibited a cohesive and flexible structure with a slightly opalescent appearance (Figure 2). The estimated properties were validated with a new formulation using the concentrations suggested above. The validation of the model was carried out via the comparison between experimental and predicted values (Table 6), and the absolute residual error of the dependent variables was calculated. The validation data are shown in Table 6. The absolute residual errors are remarkably low: (1.22%, 3.58%, 5.867%, and 9.16% for WVP, DPPH, TS, and EAB respectively). Additionally, the optimization reveals that 5% AG is the saturation point where the biopolymer network reaches maximum density, significantly hindering water vapor migration through pore‐filling effects. Moreover, Table 6 shows that WVP (4.052 × 10⁻⁹ g/m.s.Pa) of the optimized film was higher than the film based on bitter cassava starch/nanocellulose and REO (WVP = 3.36 × 10⁻¹¹ g/m.s.Pa) developed by Araya et al. (2022).

TABLE 7.

Physicochemical, antimicrobial, and antioxidant properties of the optimized active edible film.

Physicochemical properties	Optimal film formulation
Thickness (um)	52.36 ± 1.5
L *	88.5 ± 1.02
a *	−1.15 ± 0.03
b *	+6.8 ± 0.21
Opacity (A/mm)	3.1 ± 0.06
Moisture content (%)	12.87 ± 0.86
Water solubility (%)	31.45 ± 1.23
Antimicrobial activity: inhibition zone (mm)
Escherichia coli	16.6 ± 0.95
Staphylococcus aureus	16.8 ± 0.97
Salmonella typhimurium	18.0 ± 0.89
Antioxidant activity (DPPH, %)	79.012 ± 1.13

FIGURE 2.

Active almond gum–gelatin film produced using the optimized formulation (5% almond gum, 2% REO, and 24.827% glycerol).

TABLE 6.

Predicted and experimental responses of almond gum–gelatin–rosemary essential oil film prepared using the optimized formulation.

Responses	Predicted value a	Experimental value b (n = 3)	Absolute residual error (%) c
WVP (10−9 g/m.s.Pa)	3.907	4.052 ± 0.94	3.58
DPPH (%)	78.043	79.012 ± 1.24	1.22
TS (MPa)	27.320	29.023 ± 1.3	5.867
EAB (%)	30.717	33.814 ± 1.41	9.16

^a Predicted values obtained from the model equations.

^b Experimental values obtained at optimum conditions (almond gum = 5%, rosemary essential oil = 2%, and glycerol = 24.827%).

^c Absolute residual error (%) = [(experimental value − predicted value) / experimental value] × 100. R ² predicted and experimental values correlation (0.99).

However, REO plays a dual role. While its primary function is radical scavenging (as confirmed by the 79.012% DPPH activity), it also acts as a hydrophobic filler. At the optimized 2% concentration, the REO is sufficiently emulsified by AG to prevent large‐scale phase separation. These observations are confirmed by FTIR analysis, which showed shifts in characteristic bands indicating tight intermolecular hydrogen bonding. Furthermore, DSC thermogram demonstrated enhanced thermal stability and homogeneous phase structure, proving the absence of large‐scale phase separation or structural collapse that would otherwise compromise TS (Bassey et al. 2025; Chen et al. ²⁰²⁵).

The glycerol concentration of 24.8% is the critical balance point. Indeed, high concentration of glycerol (typically > 30%) leads to a leaking effect where the barrier properties are compromised, while insufficient plasticization (< 15%) leads to cracking wrapping during food applications (Ma et al. 2026; Suderman and Sarbon 2020; Asfaw et al. ²⁰²³). EAB of 33.8% in the validated formulation (Table 6) ensures that the active film can withstand the mechanical stresses of packaging without losing its integrity. When compared with starch film incorporated with REO developed by Navia‐Porras et al. (2018), which typically suffer from low TS (8.96 MPa), our film exhibits a substantial structural reinforcement. This behavior is proved by the FTIR analysis, indicating that the cross‐linking and strong intermolecular hydrogen bonding between the multifunctional hydroxyl groups of AG and the amino groups of gelatin significantly upregulate the film's resistance. Concurrently, the EAB remains sufficiently high to withstand structural stress during food wrapping, showing more favorable flexibility to strength tradeoff than the cassava starch film developed by Navia‐Porras et al. (2019) and incorporated with REO.

3.7. Physicochemical and Antimicrobial Properties of Optimized Edible Film

3.7.1. DSC Analysis

The DSC profile of the optimized active film is illustrated in Figure 3. Three distinct thermal regions were identified. A broad endothermic region at 25°C–35°C which was attributed to the combined evaporation of residual bound water structurally associated with the biopolymer network and the mobilization of low‐boiling volatile fractions loosely retained within the film matrix. More notably, a major sharp endothermic enthalpy peak was observed at a peak melting temperature (T _m = 179.34°C), attributed to the thermal melting of the residual ordered triple‐helix crystalline micro‐domains of gelatin, onset of AG polysaccharide chain decomposition, and concomitant thermal release of REO constituents encapsulated within the biopolymer network (Dong et al. 2023). The superposition of these events into a single, intense peak rather than discrete transitions further corroborates the existence of specific interactions between the lipophilic constituents of REO and the hydrophobic domains of gelatin peptide chains, which retain the essential oil within the matrix until the network integrity is disrupted by thermal denaturation.

FIGURE 3.

DSC analysis of the optimized almond gum–gelatin film.

3.7.2. FTIR Analysis

The FTIR spectrum of the optimized active film is given in Figure 4. The structural scale shows a complex interpenetrating hydrocolloid arrangement characterized by intense hydrogen‐bonding bridges and chemical compatibility between the protein strands of gelatin and the polysaccharide domains of AG. A broad, intense absorption band centered at approximately 3300 cm⁻¹ was attributed to the overlapping O–H stretching vibrations of AG hydroxyl groups and N–H stretching vibrations (Amide A) of gelatin peptide chains, whose breath and shift toward lower wavenumber relative to individual components are indicative of extensive intermolecular hydrogen bonding between the polysaccharide and protein phases of the composite matrix, consistent with the formation of a cohesive semi‐IPN. Similar results were observed by Hasdar et al. (2024) for gelatin–konjac glucomannan films. Absorption bonds at 2925 and 2975 cm⁻¹ were assigned to asymmetric and symmetric C–H stretching vibrations of methylene groups present in both biopolymers and the aliphatic monoterpene constituents of REO. The two dominant bands at 1645 cm⁻¹ and 1540 cm⁻¹ were explicitly assigned to the Amide I (C = O stretching) and Amide II (N–H bending combined with C–N stretching) bands of gelatin, respectively, confirming the structural integrity of the gelatin peptide backbone within the composite; the position of the Amide I band at 1645 cm⁻¹, associated with random coil conformation, is consisted with values reported for gelatin in protein–polysaccharide semi‐IPN films, where hydrogen bonding with the polysaccharide component induces measurable shifts relative to pure gelatin films, as explained by Wang et al. (2025). A further band at 1150 cm⁻¹ was assigned to C–O–C asymmetric stretching of AG glycosidic linkages and C–O stretching of polysaccharide chain of AG and glycerol, while the intense band at 1040 cm⁻¹ was attributed to the combined contributions of AG C–O–C glycosidic stretching and the C–O ether stretching of 1,8‐cineole, whose characteristic fingerprint at 1040 cm⁻¹ has been reported for 1,8‐cineole rich essential oils incorporated into biopolymer matrices.

3.7.3. Opacity and Color Evaluation

The physicochemical and biological properties are represented in Table 7. The color analysis (L ^* = 88.5, a ^* = −1.15, and b ^* = 6.8) revealed high lightness (L ^* = 88.57) which is essential for consumer acceptance as it allows the product (chicken fillets) to remain visible (Table 7). The slightly green (a ^*) and yellow (b ^*) nuances are standard for films incorporated with natural extracts as REO (Gaba et al. 2022). Crucially, the low b ^* value (6.8) ensures that the film does not impart an undesirable aged or oxidized yellow appearance to the wrapped product. Similar results were observed by Gomaa et al. (2022) for the bio‐composite films from gum arabic and galangal extract as well as by Seididamyeh et al. (2024) when they used gum arabic to produce an active film enriched with aqueous extracts of Syzygium aqueum, Diploglottis bracteata, and Tasmannia lanceolata.

The opacity value (Table 7) of 3.1 A/mm further confirms the superior homogeneity of the optimized formulation. In general, opacity below five is indicative of a highly transparent material with minimal light scattering (Bassey et al. 2025). Moreover, this transparency is vital as it allows the consumer to visually inspect the freshness and quality of the underlying product without visual distortion (Al‐Harrasi et al. ²⁰²²; Bassey et al. ²⁰²⁵).

3.7.4. MC and WS

The optimized film exhibited an MC of 12.87%. This value is notably moderate for a hydrocolloid‐based system, which typically suffers from high water affinity. This suggests that AG and gelatin have formed a highly dense IPN. The inclusion of REO, a hydrophobic phase, likely contributes to this reduced MC by occupying the free volume within the matrix and limiting the availability of hydrogen‐bonding sites for water molecules (Riaz et al. 2020; Herrera‐Vázquez et al. ²⁰²²). Compared to pure gelatin films which can exhibit MC values exceeding 20% (Wu et al. 2023), this 12.87% reflects more stable and less hygroscopic material, essential for maintaining structural integrity in humid food environments (Al‐Harrasi et al. ²⁰²²; Herrera‐Vázquez et al. ²⁰²²).

Simultaneously, the WS of 31.45% (Table 7) indicates a well‐balanced biodegradable profile. While very low solubility is preferred for high‐moisture foods, a value of 31.45% is considered acceptable for active packaging (Herrera‐Vázquez et al. ²⁰²²). This level of solubility facilitates the controlled migration of bioactive compounds (phenolic antioxidants and antimicrobials) from the film to the food surface. The moderate solubility is driven by the hydrophilic polypeptide chains of gelatin, while the presence of AG and REO limits instantaneous dissolution through physical cross‐linking and hydrophobic interactions, as proven by Ninan et al. (2014).

3.7.5. Antimicrobial Activity

The antimicrobial properties of the optimized active film against E. coli, S. aureus, and S. typhimurium are shown in Table 7. For the optimized film without REO (control), the inhibition zones for all the tested germs were 0.00 mm. However, for the optimized active film, the recorded inhibition zones were 16.6 mm against E. coli, 16.8 mm against S. typhimurium, and 18.0 mm against S. aureus, which proved the high bioactive property of the active film. These values are significantly higher than those typically reported for gelatin films, which often show zones below 12 mm (Martucci et al. 2015). This antimicrobial activity (in general > 16 mm for inhibition zone) could be correlated with the FTIR and DSC structural findings. The FTIR analysis demonstrated successful intermolecular hydrogen bonding and cross‐linking between the AG–gelatin backbone and the active functional groups of the REO. Concurrently, the DSC thermogram confirmed that this internal network enhanced the thermal stability and modified the glass transition properties of the composite matrix which prevents the volatilization of the essential oil.

3.8. Conservation of Chicken Fillets With the Active Film

3.8.1. Microbiological Analysis

The TAB count, also called APC, serves as a fundamental indicator of the overall microbiological quality and degree of freshness in poultry meat (Sheerzad et al. 2024). Results in Figure 5a show that both the control and wrapped samples began with an initial load of 3.7 log CFU/g. However, the proliferation rates diverged significantly over the 12‐day storage period. The control samples exhibited a significant increase (p < 0.05) reaching 7.9 log CFU/g by Day 6 and peaking at 9.3 log CFU/g by Day 12. In contrast, the fillets wrapped with the optimized active film showed a remarkably suppressed growth reaching only 6.4 log CFU/g at the end of the storage period. The critical threshold for spoilage in poultry is generally recognized as 7.0 log CFU/g according to the international commission on microbiological specifications for food. The control samples exceeded this limit between Day 4 and Day 6, whereas the coated samples remained below this limit until Day 12. This can be attributed to the presence of REO active compounds, such as 1.8‐cineole and camphor, which disrupt the bacterial cytoplasmic membrane, increasing its permeability and leading to cell death (El Atki et al. 2019).

FIGURE 5.

(a) Total aerobic bacteria, (b) psychrotrophic bacteria, (c) Pseudomonas, (d) lactic acid bacteria, (e) Staphylococcus aureus, (f) Enterobacteriaceae, and (g) mold and yeast evolutions of chicken fillets during refrigerated storage. Data shown are the mean ± standard deviation. Different lowercase letters show significant difference (p < 0.05) between the two treatments (control and wrapped samples) and uppercase letters show significant difference (p < 0.05) between times for each treatment.

For PB, they are the primary agents responsible for meat decay under refrigerated conditions, as they can thrive at 4°C and produce enzymes that degrade proteins and lipids, resulting in off‐odors (Bazargani‐Gilani et al. ²⁰¹⁵). As shown in Figure 5b, the control samples showed a significant (p < 0.05) rise from 3.8 to 9.7 log CFU/g; however, although there was significant increase (p < 0.05), the active film retarded this growth, maintaining the PB count at 6.6 log CFU/g at the end of the storage period. This significant reduction (3.1 log difference between control and wrapped samples) lies in the synergistic effect of the biopolymer matrix and the essential oil. AG and gelatin enhanced the stability of the REO droplets, ensuring gradual release of antimicrobial molecules onto the meat surface (Shahrampour and Razavi 2023; Noori et al. ²⁰¹⁸).

Moreover, Pseudomonas are recognized as the dominant spoilage flora in aerobically stored poultry meat, known for their ability to metabolize amino acids and produce foul‐smelling compounds such as ammonia and biogenic amines (Hosseini et al. 2021). Results (Figure 5c) indicate a significant (p < 0.05) inhibition of these bacteria: The control samples reached 6.5 log CFU/g, while wrapped samples were limited to 4.3 log CFU/g after 12 days. The antimicrobial efficacy against Pseudomonas is largely driven by the hydrophobic nature of REO, which enables it to penetrate the complex outer lipopolysaccharide membrane of Gram‐negative bacteria (Abbasi et al. 2021).

On the other hand, LAB are facultative anaerobes that contribute to the souring and discoloration of meat through the production of organic acids and hydrogen peroxide (Sheerzad et al. 2024). As illustrated in Figure 5d, while LAB growth was observed in all samples, the wrapped fillets showed a significantly (p < 0.05) lower progression (4.4 log CFU/g) compared to the control (6.1 log CFU/g) on Day 12. The reduced proliferation of LAB in the packaged samples suggests that the optimized active film creates an environment that is less conductive to anaerobic fermentation. This was achieved through the antimicrobial action of the lipophilic REO compounds (Badia et al. 2020), which disrupt bacterial cell membrane integrity and function. By maintaining LAB counts at lower levels, the active film preserves the natural pH and enhances the color stability of the chicken breast fillets (Sheerzad et al. 2024).

The presence of S. aureus in poultry is a major food safety concern due to its potential for enterotoxin production. The S. aureus level (1.5 log CFU/g) is approximately 2 log units below the regulatory acceptability threshold established for fresh poultry meat (3.7 log CFU/g) and below average values reported in comparable studies on raw poultry meat from North Africa, where S. aureus is consistently detected as a surface contaminant of human or environmental origin even under good hygienic practices (Cohen et al. 2007).

As illustrated in Figure 5e, the initial count of 1.5 log CFU/g increased significantly (p < 0.05) to 4.7 log CFU/g in the control samples, whereas the wrapped samples only reached 3.3 log CFU/g. S. aureus are typically sensitive to essential oils because they lack an outer membrane (El Atki et al. 2019). The active film's ability to maintain S. aureus counts below 4.0 log CFU/g, a level often associated with the absence of significant enterotoxin risk, highlights its role as a safety barrier.

Moreover, Enterobacteriaceae are frequently used as hygiene indicators in meat processing, reflecting potential fecal contamination or poor handling practices. As shown in Figure 5f, the control samples reached 6.8 log CFU/g, while the wrapped samples were maintained at 4.8 log CFU/g at the end of the storage days. The 2‐log difference confirms the ability of REO to bind to bacterial proteins and inhibit essential enzymes like amino acid decarboxylase.

However, mold and yeast often colonize meat surfaces when bacterial competition is reduced or in high‐moisture environments. As shown in Figure 5g, the initial yeast and mold count for the samples at Day 0 was approximately 2 log CFU/g. This baseline represents a relatively high initial contamination level, which can be attributed to the open‐market sourcing of fresh poultry. A high initial microbial load typically accelerates food spoilage kinetics; however, despite this challenge baseline, the optimized active film successfully restricted further yeast and mold proliferation during storage compared to control. Indeed, the most important difference has been observed here (Figure 5g): The control samples increased to 8.5 log CFU/g, while the packaged fillets stayed at a moderate 4.7 log CFU/g. This significant difference (p < 0.05) with a reduction of 3.8 log CFU/g proved the high antifungal properties of the REO, specifically its components as alpha‐pinene and 1,8‐cineole which disrupt fungal cell walls and cytoplasm (OuYang et al. ²⁰¹⁹). This action was reinforced by the presence of AG, which is less hygroscopic than pure gelatin, and likely reduces the water activity on the surface, further inhibiting fungal germination. The same findings were observed by other authors (Noori et al. 2018; Abbasi et al. ²⁰²¹) who showed a significant decrease in mold and yeast count using edible coatings enriched with various essential oils and used to preserve chicken meat.

3.8.2. pH Evolution

The evolution of meat pH during refrigerated storage is presented in Table 8. At Day 0, no significant difference (p > 0.05) was observed between control and wrapped samples (5.82 ± 0.04) confirming the homogeneity of the initial material. In control samples, pH increased significantly (p < 0.05) throughout storage, rising from 5.82 ± 0.04 at Day 0 to 6.54 ± 0.08 at Day 12, consistent with the accumulation of alkaline volatile metabolites produced by the proliferating spoilage microorganisms, particularly PSE as reflected by the simultaneous increase in TVB‐N values. In contrast, samples wrapped with the active film exhibited a significantly lower (p < 0.05) rate of pH increase from 5.82 ± 0.04 at Day 0 to 5.98 ± 0.06 at Day 12. This attenuation of pH increase can be attributed to the antimicrobial activity of the REO incorporated in the film, which suppressed the proliferation of microorganisms responsible for alkaline metabolite production.

TABLE 8.

pH evolution of chicken fillets during refrigerated storage.

pH
	Control samples	Wrapped samples
Day 0	5.82 ± 0.04aA	5.82 ± 0.04aA
Day 2	5.98 ± 0.03aB	5.85 ± 0.02bA
Day 4	6.14 ± 0.05aC	5.89 ± 0.03cA
Day 6	6.24 ± 0.02aD	5.91 ± 0.04cA
Day 9	6.32 ± 0.06aE	5.94 ± 0.04dA
Day 12	6.54 ± 0.08aF	5.96 ± 0.06dA

Note: Different lowercase letters in the same row indicate a significant difference between treatments; different uppercase letters in the same column indicate a significant difference over storage time (p < 0.05).

3.8.3. TVB‐N Evolution

TVB‐N is a widely accepted chemical spoilage biomarker, representing the accumulation of ammonia, biogenic amines, and other volatile nitrogenous compounds resulting from protein degradation by enzymes and bacteria (Abbasi et al. 2021). Results (Figure 6) showed that control samples exceeded the standard spoilage limit of 25 mg/100 g (Sheerzad et al. 2024) by Day 6 (26.71 mg/100 g) and reached a high amount of 39.3 mg/100 g by Day 12. Remarkably, wrapped samples only reached 25.12 mg/100 g at the end of the storage period. This result indicates that the active film extended the chemical shelf life of the chicken by at least 6 days. This retardation is linked to the massive suppression of Pseudomonas and PB growth, which are the primary drivers of proteolysis (Sheerzad et al. 2024). These results are in line with the results observed by Sheerzad et al. (2024) who reported that chicken meat coated with whey protein isolate–nanochitosan–bacterial nanocellulose and cinnamon essential oil significantly reduced the TVB‐N amount in the coated samples compared to the control.

FIGURE 6.

Evolution of total volatile basic nitrogen (TVB‐N) of control and wrapped chicken fillets during storage period (12 days) at refrigerated temperature. Data shown are the mean ± standard deviation. Different lowercase letters show significant difference (p < 0.05) between the two treatments (control and wrapped samples) and uppercase letters show significant difference (p < 0.05) between times for each treatment.

3.8.4. PV

The PV measures the formation of primary lipid oxidation products (hydroperoxides). According to the results (Figure 7), control samples showed a significant (p < 0.05) increase from 1.91 to 6.72 meq peroxide/kg. However, the packaged samples exhibited a much slower rate, ending at 4.4 meq peroxide/kg. These results confirm the antioxidant efficiency of the REO, especially its phenolic fraction that acts as radical scavengers, deactivating peroxy radicals and chelating metal ions that would otherwise catalyze the initiation phase of oxidation (Aminzare et al. 2019). Many authors reported that active films and coatings, enriched by essential oils, can reduce PV and TVB‐N in chicken meat, and they explained that by the very low permeability to oxygen and carbon dioxide of film coatings and the presence of antioxidants, which stabilize the lipid fraction of the poultry meat (Dalvandi et al. 2020; Zhou et al. ²⁰²¹).

FIGURE 7.

Evolution of peroxide value (PV) of control and wrapped chicken fillets during storage period (12 days) at refrigerated temperature. Data shown are the mean ± standard deviation. Different lowercase letters show significant difference (p < 0.05) between the two treatments (control and wrapped samples) and uppercase letters show significant difference (p < 0.05) between times for each treatment.

3.8.5. TBARS Evolution

TBARS quantify malondialdehyde (MDA), the most prevalent secondary product of lipid oxidation responsible for rancid off flavors, in chicken meat. Results (Figure 8) showed that control samples reached 1.224 mg MDA/kg, surpassing the common consumer detection threshold for off‐flavors (typically 1.0 mg/kg, as reported by Dalvandi et al. (2020)); however, Zhou et al. (2021) reported that values of TBARS more than 0.5 mg/kg in chicken meat may be detectable as off‐flavor by consumers. However, wrapped samples remained exceptionally stable, with a final value of only 0.38 mg MDA/kg. This value is the most impressive chemical outcome in this study. It proves that the optimized active film provides an environment that almost completely halts secondary oxidation. The gradual release of REO antioxidants combined with the structural trapping of MDA by the primary amino groups creates a powerful protective system (Sandhiutami et al. 2023; Dalvandi et al. ²⁰²⁰).

FIGURE 8.

Evolution of thiobarbituric acid reactive substances (TBARS) of control and wrapped chicken fillets during storage period (12 days) at refrigerated temperature. Data shown are the mean ± standard deviation. Different lowercase letters show significant difference (p < 0.05) between the two treatments (control and wrapped samples) and uppercase letters show significant difference (p < 0.05) between times for each treatment.

3.8.6. Sensory Evaluation

Sensory quality scores of the different samples during refrigerated storage are summarized in Table 9. On Days 0, 2, and 4, all samples retained high scores (> 7.0). However, the difference between control and wrapped samples was significant (p < 0.05) during these 4 days of storage. A critical turning point was observed on Day 6, providing direct evidence of the film's efficacy. At this stage, the control samples showed an overall acceptability score, falling below five (4.6 ± 0.3). This significant decline (p < 0.05) was primarily driven by the beginning of an off‐odor and initial surface sliminess resulting from microbial proliferation. Conversely, the fillets packaged with the optimized active film remained entirely acceptable, with mean scores of 7.8 ± 0.2, proving a significant (p < 0.05) preservation of fresh‐like attributes. By Days 9 and 12, the control samples underwent advanced structural and olfactory decay (scores < 4.0).

TABLE 9.

Sensory attribute scores of control and wrapped chicken fillets during refrigerated storage at 4°C.

Storage days	Samples	Color	Odor	Overall acceptability
0	Control	8.8 ± 0.2aA	8.7 ± 0.3aA	8.7 ± 0.2aA
	Wrapped samples	8.6 ± 0.3aA	8.5 ± 0.4aA	8.6 ± 0.3aA
2	Control	8.2 ± 0.3aB	8.0 ± 0.2aB	8.1 ± 0.4aB
	Wrapped samples	8.4 ± 0.2aA	8.3 ± 0.2aA	8.3 ± 0.2aA
4	Control	7.1 ± 0.4bC	6.9 ± 0.2bC	7.1 ± 0.3bC
	Wrapped samples	8.1 ± 0.3aB	8.0 ± 0.3aB	8.1 ± 0.4aB
6	Control	5.2 ± 0.4bD	4.4 ± 0.2cD	4.6 ± 0.2bD
	Wrapped samples	7.9 ± 0.5aB	7.8 ± 0.3aB	7.8 ± 0.4aC
9	Control	4.4 ± 0.2cE	3.8 ± 0.5cE	4.1 ± 0.3cE
	Wrapped samples	7.3 ± 0.3aC	7.1 ± 0.1aC	7.2 ± 0.3aD
12	Control	4.1 ± 0.2cE	3.5 ± 0.4cE	3.9 ± 0.5cE
	Wrapped samples	6.8 ± 0.5aD	6.5 ± 0.5aD	6.7 ± 0.3aE

Note: Values are expressed as mean ± standard error. Different lowercase superscript letters in the same column within the same storage day indicate a significant difference between treatments (p < 0.05). Different uppercase superscript letters in the same column for the same treatment indicate a significant difference across storage time (p < 0.05).

However, the optimized active film successfully extended the sensory shelf life up to 12 days (scores around 6). This performance confirms that the lipophilic REO volatile compounds exert a continuous antimicrobial and protective barrier directly at the meat interface effectively inhibiting off‐odor formation and stabilizing the color of the wrapped samples.

4. Conclusion

In this study, the formulation of an active film using AG and gelatin as matrix, enriched with REO and plasticized with glycerol, was optimized using RSM. Results showed that the optimized film presented a significant moisture barrier and interesting mechanical integrity. FTIR spectroscopy and DSC analyses confirmed successful intermolecular hydrogen bonding, cross‐linking, and enhanced thermal stability within the composite network. Moreover, the successful stabilization of REO within this matrix provided powerful antioxidant and antimicrobial activities, effectively inhibiting lipid oxidation and major foodborne pathogens. The practical application on chicken fillets showed a shelf‐life extension up to 12 days. This preservation window was rigorously cross‐validated by the precise convergence of restricted microbial growth, retarded chemical alteration (TVB‐N and TBARS), and the maintenance of raw meat sensory acceptability scores.

These bioactive films hold significant commercial potential as sustainable alternatives in the clean‐label food packaging sector. The formulation can be transitioned into industrial batch‐mixers and automated continuous casting or industrial spray coating lines. Furthermore, because AG is an abundant, cost‐effective agroforestry byproduct, its exploitation supports regional bioeconomy loops. This active film is highly suited for preformed commercial trays or protective interlayers for fresh, high‐value perishable foods, aligning with modern consumer demands for biodegradable and synthetic‐preservative‐free packaging.

Author Contributions

Nesrine Mahfoudhi: funding acquisition, writing – original draft, conceptualization, investigation, methodology, software, formal analysis, visualization. Hakima Khelifi: project administration. Karima Gharsallah: methodology, writing – review and editing, formal analysis, software. Leila Rezig: project administration, formal analysis, methodology.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Figure S1: jfds71374‐sup‐0001‐FigureS1.docx