Chitosan–PVA composite films incorporated with roselle anthocyanins: Colorimetric smart sensor for real-time monitoring of meat and seafood freshness

Abstract

Ensuring freshness in perishable foods like meat and seafood is vital for food safety. This study developed a biodegradable colorimetric film from chitosan–polyvinyl alcohol (Cs/PVA) infused with roselle anthocyanin extract (CPAR) to monitor spoilage. Among variants, CPAR-5 (5 mL extract) showed the highest pH sensitivity (ΔE = 44.69, S% = 42%) and rapid response to ammonia vapor. Characterization via UV–Vis, FTIR, and FESEM confirmed effective pigment integration and film structure. Applied to chicken, shrimp, and tilapia, CPAR-5 visually tracked spoilage in real time under ambient and cold storage, correlating with critical pH thresholds (≥6.8 for chicken, ≥7.4 for shrimp, ≥7.0 for fish). A visual card reader was developed for semi-quantitative freshness evaluation, demonstrating that CPAR-5 is a highly sensitive, robust, user-friendly, and environmentally sustainable smart indicator with strong potential for practical food safety monitoring and reduce food waste.

Highlights

• •Colorimetric film developed using roselle anthocyanin and chitosan–PVA matrix.
• •Film shows distinct color shifts in response to pH and ammonia vapor exposure.
• •Enables real-time freshness monitoring of chicken, shrimp, and tilapia meats.
• •Card reader based on RGB analysis facilitates interpretation of freshness level.

1. Introduction

Ensuring the freshness and safety of perishable food products such as chicken, shrimp, and tilapia remains a global concern, particularly due to the rapid microbial spoilage that leads to undesirable changes in organoleptic characteristics, nutritional degradation, and the formation of harmful volatile amines (Ashiq, Saeed, Li, & Nawaz, 2024). These issues are exacerbated under suboptimal packaging and storage conditions, contributing to food waste and raising public health risks (Siddiqui et al., 2022). Conventional freshness evaluation techniques, such as organoleptic assessment, microbiological assays, or physicochemical measurements, are often time-consuming, labor-intensive, and not suitable for real-time or consumer-level monitoring applications (Zhang et al., 2025).

Recent advances in food sensor freshness have given rise to colorimetric indicator films, which offer an intelligent and visual method for real-time freshness monitoring (Almasi, Forghani, & Moradi, 2022; Mkhari, Adeyemi, & Fawole, 2025). These films function by exhibiting distinct and quantifiable color changes in response to biochemical spoilage markers such as ammonia, trimethylamine, and pH shifts (Palanisamy, Kadirvel, & Ganesan, 2025). Several studies have demonstrated the successful application of such smart films for assessing the freshness of various protein-rich foods, including chicken (Ahmad, Abdullah Lim, & Navaranjan, 2020), shrimp (Wu et al., 2021), and tilapia (Ezati, Tajik, Moradi, & Molaei, 2019).

The freshness of protein-rich foods such as fish, shrimp, and chicken can be reliably monitored through pH changes caused by microbial and enzymatic activity. Fresh tilapia typically exhibits a pH of 6.4–6.6, which increases beyond 7.0 as spoilage progresses due to the accumulation of basic nitrogen compounds like ammonia and trimethylamine (Salgado, Di Giorgio, Musso, & Mauri, 2021). Shrimp also deteriorate rapidly, with pH rising from 6.2 to 6.5 to above 7.0–7.5 within 12 h at ambient temperature (Liang et al., 2025). Similarly, chicken breast meat shows pH elevation from 5.8 to 6.2 to over 7.5 within 24 h without refrigeration (Kim et al., 2023). These pH shifts provide a consistent and sensitive indicator of spoilage across different food matrices.

In this context, pH-responsive sensors offer a distinct advantage, as pH serves as a universal and quantitative indicator of microbial and enzymatic degradation across nearly all types of meat (Xu, Yu, Liu, Cheng, & Xiong, 2025). The consistent pH elevation during spoilage, establishes pH detection as a broadly applicable and integrative approach for food sensing systems (Uysal-Unalan et al., 2024). When designed using natural compound-based colorimetric systems, pH sensors can deliver immediate visual feedback, environmental compatibility, and strong potential for widespread implementation in smart sensor platforms for real-time food quality monitoring (Zhao et al., 2024).

Therefore, a compound capable of detecting pH across a broad range through colorimetric changes is required. The compound needed is anthocyanins, found in flowers and vegetables such as roselle, carrots (Li, Simon, & Wu, 2024), and turmeric (a source of curcumin) (Ezati & Rhim, 2020), are promising candidates for this purpose. Anthocyanins derived from roselle (Hibiscus sabdariffa L.) represent an optimal choice for natural pigment applications, as its calyces contain exceptionally high anthocyanin content-reaching up to 1.5 g per 100 g of dry weight (Zhai et al., 2017). Anthocyanins consist of glycosylated anthocyanidin structures. The most prevalent types in natural environments include delphinidin-3-sambubioside (accounting for ∼70% of anthocyanins), delphinidin-3-glucoside, peonidin, malvidin, pelargonidin, petunidin, cyanidin-3-sambubioside (major pigment), and cyanidin-3-(p-coumaroyl-glucoside) (minor pigment) (Kalt, 2019).

Anthocyanins exhibit high pH-dependent color responsiveness, demonstrating distinct chromatic transitions: red at pH 2 (flavylium cation), pink at pH 3, faded pink at pH 4–5 (carbinol pseudobase), purple at pH 6 (quinoidal base), blue at pH 7–9 (anionic quinoidal base), yellow at pH 10–12, and colorless at pH >12 (chalcone) (Li et al., 2022). The incorporation of anthocyanins into polymer matrices significantly enhances color stability, a critical factor for applications requiring long-term durability such as food packaging. For instance, anthocyanin-polysaccharide complexation has been demonstrated to substantially improve long-term color retention (Veloso, Coelho, Trabulo, & Coimbra, 2022).

Chitosan (Cs), a natural cationic biopolymer with a high molecular weight, is produced via the deacetylation of chitin. Chitosan-based films provide dual advantages of environmental biodegradability and processing versatility for packaging applications, achieved through the conversion of low molecular weight chitosan into water-soluble forms via specific modification techniques (Fiddaroini et al., 2025). As a polymeric matrix, chitosan acts as an effective cross-linker through specific molecular interactions, thereby enhancing mechanical properties and structural stability. At the molecular level, chitosan comprises linear chains of β-(1–4)-linked D-glucosamine units, randomly interspersed with N-acetyl-D-glucosamine moieties as structural modifications (Li et al., 2022).

Chitosan-based films demonstrate pH stabilization capacity due to their weak acid characteristics, enabling versatile complexation with diverse substrates to form highly stable complexes (Li, Bao, et al., 2022). However, pure chitosan colorimetric films often exhibit insufficient compositional performance, necessitating the incorporation of polyvinyl alcohol (PVA) to enhance mechanical robustness and functional stability (Li et al., 2023). PVA is recognized as a non-toxic synthetic polymer exhibiting excellent water solubility, biodegradability, and film-forming capabilities, while significantly enhancing the mechanical stability of composite films (Amalraj, Haponiuk, Thomas, & Gopi, 2020). PVA has garnered significant scientific interest due to its unique physicochemical properties, enabling diverse advanced applications including membrane fabrication, controlled drug delivery systems, and polymer recycling technologies (Lee et al., 2017).

Previous studies have demonstrated the feasibility of anthocyanin-based colorimetric indicators using various polymer matrices; however, their applicability remains limited by material selection, sensing tunability, and testing scope. Zhai et al. (2017) reported starch/PVA-based colorimetric films, while Zhang et al. (2019) incorporated roselle anthocyanins into starch/PVA composites and evaluated their performance exclusively on pork stored at 25 °C. More recently, Latiff et al. (2025) employed gelatin as the primary polymer matrix for anthocyanin-based freshness indicators. Despite these advances, systematic modulation of pigment loading and validation across multiple food matrices and storage conditions remain insufficiently explored. In contrast, the present study employs a chitosan–polyvinyl alcohol (Cs/PVA) hybrid matrix integrated with roselle flower-derived anthocyanins, with extract volumes precisely varied from 1 to 5 mL to tailor colorimetric sensitivity. This compositional strategy enables controlled tuning of pH responsiveness while leveraging the complementary film-forming properties of chitosan and PVA. The resulting films are evaluated for real-time freshness monitoring of tilapia, shrimp, and chicken under both ambient (25 °C) and refrigerated (6 °C) storage conditions, thereby providing a broader and more application-relevant assessment compared to earlier single-material, single-product, or single-temperature studies. This approach advances anthocyanin-based smart sensor toward more versatile and practical food quality monitoring systems.

The freshness-monitoring capability of PVA-based films, which exhibit sensitive detection of quality deterioration in chicken and shrimp. The films show a pronounced response to ammonia vapor (NH₃) at concentrations as low as 25 ppm (Wang, Wang, Sun, Liu, & Wang, 2022). The colorimetric response to NH₃ proceeds through ammonia vapor binding with trapped water molecules in the indicator matrix (forming NH₃·H₂O), followed by hydrolysis to generate OH⁻ and NH₄⁺, which collectively drive the observed color transition (Ezati & Rhim, 2020). Furthermore, the RGB values of the film can be quantitatively correlated with total volatile basic nitrogen (TVB-N) levels and automatically analyzed via smartphone applications to determine meat freshness. The synergistic integration of anthocyanin-based pH-sensitive films with smartphone technology provides an accurate, simple, and reliable method for consumer-level food quality assessment (Zhao, Liu, Zhao, & Wang, 2022).

Ensuring food freshness and safety remains one of the most urgent global challenges, particularly as meat and seafood rapidly deteriorate throughout complex supply chains. Growing demand for sustainable, smart sensors has accelerated the search for materials that not only detect early spoilage signals but also integrate seamlessly into next-generation smart packaging. In response to this need, we developed a biodegradable colorimetric film that harnesses the vivid halochromic behavior of roselle-derived anthocyanins, embedded within a chitosan–polyvinyl alcohol (Cs/PVA) matrix engineered for high sensitivity and structural stability. The film exhibits tunable responsiveness toward alkaline volatiles controlled through graded ammonia exposure and delivers instantaneous, non-destructive freshness readouts for chicken, shrimp, and tilapia under both ambient and refrigerated storage. Its chromogenic transitions are further translated into smartphone-readable RGB outputs and an intuitive color reference card, enabling fully instrument-free consumer interpretation. Comprehensive UV–Vis, FTIR, FESEM, and LC-HRMS analyses reveal the molecular interactions responsible for its robust optical behavior, while CIELAB/RGB quantification confirms reliable sensitivity across pH 5–12 and 0.1–1% ammonia. Together, these features position the film not only as a practical freshness indicator but as a scalable, environmentally responsible platform capable of advancing smart sensor-quality monitoring across modern supply chains.

2. Methods

2.1. Materials and instrumentation

All reagents and solvents used in this study were of analytical grade and utilized without further purification. Ultrapure water (resistivity ≥18.2 MΩ·cm) and 75% ethanol (v/v) were employed as extraction solvents for anthocyanins from fresh Hibiscus sabdariffa (roselle) calyces, serving as the natural pigment source. For pH calibration, two buffer systems were prepared: (1) 0.025 M potassium chloride (KCl, 0.186 g; Sigma-Aldrich) adjusted to pH 1.0 using 37% HCl, and (2) 0.4 M sodium acetate (3.2812 g, Sigma-Aldrich), with pH adjusted to 4.5 using glacial acetic acid and a calibrated pH meter.

The additional reagents used include potassium dihydrogen phosphate (KH₂PO₄) 0.02 M (Sigma-Aldrich, >99.0%), citric acid monohydrate (C₆H₈O₇·H₂O) 0.02 M (Sigma-Aldrich, >98%), disodium hydrogen phosphate (Na₂HPO₄) 0.02 M (Emsure), sodium bicarbonate (NaHCO₃) 0.02 M (Sigma-Aldrich, 99.5–100.5%), sodium hydroxide (NaOH) 0.02 M (Emsure), and sodium carbonate (Na₂CO₃) 0.02 M (Sigma-Aldrich, >99.0%).

Colorimetric films were formulated using low-molecular-weight chitosan (Sigma-Aldrich, degree of deacetylation: 75–85%) dissolved in 2% (v/v) aqueous glacial acetic acid and polyvinyl alcohol (PVA, 5% w/v; Emsure) dissolved in deionized water. These Cs/PVA–anthocyanin composite films were designed to monitor freshness in protein-rich foods, including chicken breast (Gallus gallus domesticus), white shrimp (Litopenaeus vannamei), and tilapia fillet (Oreochromis mossambicus).

Analytical instrumentation employed in this study included a UV–Vis spectrophotometer (Shimadzu UV-1600 series), Fourier-transform infrared (FTIR) spectroscopy (Shimadzu 8400 s), and field emission scanning electron microscopy (FESEM; FEI Quanta FEG 650). Furthermore, phytochemical profiling of the roselle extract was performed using high-resolution liquid chromatography–mass spectrometry (LC-HRMS; Thermo Scientific Q Exactive system) coupled with a Dionex Ultimate 3000 RSLCnano HPLC platform.

2.2. Extraction and characterization of roselle flowers

2.2.1. Anthocyanin extraction from roselle (Hibiscus sabdariffa)

Fresh petals of roselle were washed and ground using a food-grade blender followed by manual homogenization with a mortar and pestle to enhance solvent accessibility. Anthocyanins were extracted via maceration, using 10 g of powdered petals in 100 mL of 75% ethanol (1:10 w/v) under dark conditions at 25 ± 1 °C for 24 h to prevent photodegradation. A longer extraction time of 24 h was used to ensure stable and reproducible anthocyanin performance in the chitosan–PVA matrix. The mixture was centrifuged at 3000 rpm for 20 min, and the supernatant was collected. Ethanol was then removed under reduced pressure at 40 °C using a rotary evaporator, concentrating the extract from 100 mL to 15 mL (Zhang et al., 2019). The final anthocyanin-rich solution was stored in amber-glass vials at 4 °C to preserve stability.

2.2.2. Liquid chromatography–high resolution mass spectrometry (LC-HRMS)

For phytochemical profiling, 300 μL of anthocyanin-rich roselle extract was diluted with 75% ethanol to a final volume of 1.5 mL, vortexed for 2 min, then centrifuged at 6000 rpm for 2 min. The supernatant was filtered through a 0.22 μm PTFE syringe filter (Whatman) before LC-HRMS analysis. Chromatographic separation and compound identification were carried out using a Thermo Scientific Q Exactive Orbitrap LC-HRMS system coupled with a Dionex Ultimate 3000 RSLCnano HPLC. Separation employed a Hypersil GOLD C18 column (50 × 1 mm, 1.9 μm) at 30 μL/min with a binary solvent system: 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B), using a linear gradient. Mass spectrometry was performed in positive ion full-scan mode with resolutions of 70,000 (MS1) and 17,500 (MS2). Compound identification was aided by Compound Discoverer 3.3 software and spectral matching against the mzCloud database.

2.2.3. Quantification of total anthocyanin content

The total anthocyanin content in roselle extract was determined using the pH differential method, which selectively quantifies monomeric anthocyanins based on their structural shift between the colored flavylium form at pH 1.0 and the colorless hemiketal form at pH 4.5 (Wu, Yang, & Chiang, 2018). Standardized buffer systems, 0.025 M KCl (pH 1.0) and 0.4 M sodium acetate (pH 4.5), were prepared using a Mettler Toledo pH meter. A 1 mL extract aliquot was mixed with 9 mL of each buffer, vortexed for 30 s, and left to equilibrate in the dark for 15 min. Absorbance was recorded at 500–520 nm (λ_max of flavylium ion) and 650–700 nm (for haze correction) using a Shimadzu UV-1600 spectrophotometer. The net absorbance (A) was then calculated using Eq. (1):

$$\mathrm{A}=\left[{\left(\mathrm{A}\left({\mathrm{\lambda }}_{\max }\right)–\mathrm{A}\left({\mathrm{\lambda }}_{\text{correction}}\right)\right)}_{\mathrm{pH}1.0}\right]-\left[{\left(\mathrm{A}\left({\mathrm{\lambda }}_{\max }\right)–\mathrm{A}\left({\mathrm{\lambda }}_{\text{correction}}\right)\right)}_{\mathrm{pH}4.5}\right]$$ (1)

The total anthocyanin content (TAC) was subsequently calculated using the eq. (2):

$$\mathrm{TAC}\left(\mathrm{mg}/\mathrm{L}\right)=\frac{A\times \mathit{MW}\times \mathit{DF}\times 1000}{\epsilon \times L}$$ (2)

where:

• •A is the absorbance difference,
• •MW is the molecular weight of the predominant anthocyanin (cyanidin-3-glucoside = 449.2 g/mol),
• •DF is the dilution factor,
• •ε is the molar extinction coefficient (26,900 L/mol·cm),
• •L is the path length of the cuvette (1 cm).

2.2.4. UV–visible spectroscopic characterization of roselle anthocyanin extract

The anthocyanin extract from roselle flowers was characterized by UV–Vis spectrophotometry across the visible spectrum (400–800 nm). Spectral measurements were conducted at varying pH conditions (pH 3–12) to simultaneously monitor absorbance characteristics and visually observable color changes. The maximum absorption wavelength was determined for each pH condition through spectral analysis, with particular attention to bathochromic and hypsochromic shifts corresponding to structural transformations of the anthocyanin chromophores.

2.3. Preparation of colorimetric chitosan–PVA–roselle anthocyanin films

Chitosan and PVA powders were pre-dried in a vacuum oven at 60 °C for 1 h to remove residual moisture and enhance solubility. Chitosan solution was prepared by dissolving 2.0 g chitosan in 100 mL deionized water with 2% v/v glacial acetic acid under stirring for 4 h at room temperature. Separately, 5.0 g PVA was dissolved in 100 mL deionized water by heating to 90 °C with constant stirring, then cooled to room temperature.

Both solutions were combined in a 1:1 ratio (10 mL,10 mL) to form the Cs/PVA base. Roselle anthocyanin extract (as described in Section 2.2.1) was added to the Cs/PVA mixture in five concentrations (1–5 mL per 20 mL polymer solution), resulting in CPAR-1 to CPAR-5 formulations. Each mixture was stirred for 2 h to ensure homogeneity and stable interaction between anthocyanins and the polymer matrix.

The 20 mL of each formulation were cast into sterile 90 mm glass Petri dishes, leveled to remove air bubbles, and dried under vacuum at 40 °C for 24 h (Memmert VO200) to minimize oxidation and pigment degradation (Khanifah, Legowo, Sholihun, & Nugraheni, 2023; Vo, Dang, & Chen, 2019; Zhang et al., 2019; Zhi et al., 2025). Dried films were peeled off and visually documented using an iPhone 13 under standardized lighting. RGB color values were quantified using ImageJ by selecting uniform regions and analyzing histograms, serving as a baseline for chromogenic response studies. The films were then individually sealed in polyethylene pouches and stored at −10 °C in the dark to maintain anthocyanin stability prior to characterization and application testing.

2.4. Characterization of colorimetric films

2.4.1. Fourier transform infrared (FTIR) spectroscopy

FTIR spectroscopy was used to analyze the chemical composition and molecular interactions within the colorimetric films. Measurements were performed using a Shimadzu IRTracer-100 spectrometer with an ATR accessory, operated via OMNIC software. Uniform film strips (1 × 1 cm²) were placed directly on the ATR crystal without pretreatment. Spectra were recorded in %Transmittance mode across 4000–400 cm⁻¹ at 4 cm⁻¹ resolution, with 10 scans per sample to enhance signal quality. Baseline correction and smoothing were applied prior to analysis. Functional group identification confirmed film formation and revealed molecular interactions among chitosan, PVA, and anthocyanins, such as hydrogen bonding and dipole–dipole interactions.

2.4.2. Field emission scanning electron microscopy (FESEM)

The surface morphology and microstructural features of the films were evaluated using field emission scanning electron microscopy (FESEM; FEI Quanta FEG 650, Thermo Fisher Scientific). Imaging was conducted under high vacuum conditions at an accelerating voltage of 10.00 kV, with a 3.0 spot size and a working distance of 10 mm. Micrographs were captured at magnifications of up to 250× to assess film homogeneity, surface smoothness, pore structure, and the degree of phase compatibility between polymeric and pigment components.

2.4.3. Thickness measurement

Film thickness was measured using a digital micrometer (Mitutoyo, 0–10 mm, ±0.001 mm) at 10 random points across each sample, including center and edges. Three replicates were evaluated per CPAR formulation (CPAR-1 to CPAR-5) to ensure reproducibility. Mean thickness and standard deviation were calculated to assess the effect of anthocyanin content on film uniformity and structure.

2.5. Determination of colorimetric film properties and pH sensitivity

To assess the pH responsiveness of the developed colorimetric films, Cs-PVA and CPAR film sample (CPAR-1 through CPAR-5) was individually immersed in 0.02 M phosphate buffer solutions spanning a wide pH range from 3.0 to 12.0. The immersion process was conducted under static conditions at ambient temperature (25 ± 1 °C) for 10 min to ensure uniform interaction between the film matrix and the buffer medium. Post-treatment, the physical appearance of each film was documented using a high-resolution digital camera under standardized illumination and a neutral white background to minimize light-based variability. Quantitative colorimetric analysis was performed using a digital colorimeter application (colorimeter app by Serhii smyk) installed on a calibrated smartphone. For each sample, the CIELAB color space parameters L* (lightness), a* (red-green axis), and b* (yellow-blue axis) were recorded at three different positions across the film surface, and the average values were calculated.

The overall color change induced by pH exposure was expressed as the total color difference (ΔE), calculated using the following standard eq. (3):

$$∆\mathrm{E}=\sqrt{{\mathrm{\Delta L}}^2+a^2+b^2}$$ (3)

where:

• •ΔL* = Lₛₐₘₚₗₑ − L_standard
• •Δa* = aₛₐₘₚₗₑ − a_standard
• •Δb* = bₛₐₘₚₗₑ − b_standard

here, the Lab* values of the film at Cs/PVA were designated as the standard reference. The magnitude of ΔE was used as a quantitative indicator of the film's sensitivity and perceptibility to pH-induced chromatic shifts.

2.6. Quantitative analysis of colorimetric film response to ammonia vapor exposure

To assess vapor-phase sensitivity, anthocyanin-based films (1 × 5 cm) were mounted on sterile Petri dish lids, while the bottoms were filled with 20 mL of aqueous NH₄OH at 0.1%, 0.5%, and 1% (w/v). After sealing, the systems were incubated at 25 ± 1 °C for 3 h. Colorimetric changes were recorded (from 0 to 180 min) every 15 min using an iPhone 13 in a Puluz mini lightbox for consistent imaging conditions.

Images were analyzed via ImageJ software, where RGB (Red, Green, Blue) values were extracted from calibrated ROIs. The film's sensitivity to ammonia was quantitatively expressed as the colorimetric sensitivity index (S%) using the normalized RGB difference method from (Chen, Yan, Huang, Zhou, & Hu, 2021), as defined in Eq. (4):

$$S\%=\left(\frac{\left|R-R\ast \right|+\left|G-G\ast \right|+\left|B-B\ast \right|}{R+G+B}\right)\times 100$$ (4)

where:

• •R, G, B are the initial red, green, and blue intensity values (pre-exposure), and
• •R*, G*, B* are the corresponding values post-exposure.

This sensitivity metric (S%) reflects the relative magnitude of chromatic deviation induced by ammonia vapor interaction and was used to compare the reactivity across different film formulations (CPAR-1 through CPAR-5) and ammonia concentrations. Films demonstrating higher S% values were considered more responsive and thus more suitable for application in freshness sensing of nitrogenous spoilage gases.

2.7. Potential applications of colorimetric films for freshness monitoring of tilapia, shrimp, and chicken

To evaluate the real-time applicability of the developed colorimetric films, controlled spoilage simulations were conducted using protein-rich food models: chicken breast, shrimp, and tilapia fillets. Each food sample (25 g of chicken or tilapia, or two whole shrimp totaling ∼25 g) was placed in sterile Petri dishes, and a 1 cm × 5 cm CPAR film strip was affixed to the inner surface of the lid, allowing indirect exposure to headspace volatiles. Samples were stored under two temperature regimes: room temperature (25 °C), with color changes recorded every 6 h for 24 h; and refrigeration (6 °C), with daily monitoring for 11 days. Film responses were quantified using the total color difference (ΔE) in the CIELAB color space served as the primary quantitative indicator of both film response and food freshness degradation.

3. Results and discussion

3.1. Characterization of roselle extract

3.1.1. Compound profiling via LC–HRMS

LC–HRMS profiling of roselle petal extract (Fig. 1a) revealed a metabolically active and chemically rich matrix comprising 46 compounds (Table S1). Canonical flavylium anthocyanins were notably absent, replaced by structurally reconfigured downstream derivatives, indicating post-biosynthetic remodeling within the flavonoid biosynthetic pathway.

Fig. 1.

Characterization of roselle anthocyanin extract. (a) Total ion chromatogram and identification of anthocyanin derivatives in roselle extract; (b) Visual color and corresponding UV–Vis spectra of roselle extract at pH 1.0 and pH 4.5; (c) pH-dependent color variation and UV–Vis absorbance spectra of roselle extract across pH 2–12.

High levels of primary metabolites (including asparagine, glutamic acid, and betaine) suggest robust nitrogen metabolism and redox homeostasis that support secondary metabolite formation. Notably, anthocyanins appeared predominantly as flavonol-type derivatives. Quercetin (C₁₅H₁₀O₇, RT: 15.74 min, Area: 5.58 × 10⁶) was identified as a major aglycone, arising from deglycosylation—likely enzymatic or spontaneous under aqueous conditions. This process eliminates the flavylium cation and yields a neutral flavonol backbone stabilized via keto-enol tautomerism and π-electron delocalization.

Subsequent reglycosylation at the C3–OH position formed quercitrin (C₂₁H₂₀O₁₁, RT: 12.96 min, Area: 9.06 × 10⁵) and rutin (C₂₇H₃₀O₁₆, RT: 11.10 min, Area: 5.50 × 10⁶), catalyzed by UDP-glycosyltransferases. These sugar conjugations enhance water solubility, antioxidant stability, and chemical robustness. Rutin's disaccharide moiety offers greater steric protection to the C-ring, extending molecular stability under physiological conditions.

Highly glycosylated derivatives, such as 2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-4-oxo-4H-chromen-3-yl 6-O-β-d-xylopyranosyl-β-D-glucopyranoside (rutin derivative, C₂₆H₂₈O₁₆, RT: 3.11 & 9.14 min, Area: up to 1.23 × 10⁸), reflect sequential glycosylation steps forming branched sugar chains. While C3 remains the primary conjugation site, modifications at C7 and C4′ are likely, influencing antioxidant activity, membrane transport, and metabolic fate.

The profile further includes daidzein (C₂₁H₂₀O₉, RT: 14.42 min, Area: 1.80 × 10⁷), a non-anthocyanin isoflavonoid indicating phenylpropanoid pathway cross-talk, and scopoletin (C₁₀H₈O₄, RT: 12.17 min, Area: 2.61 × 10⁶), a coumarin derivative signifying adaptive lactonization responses. Alkaloids, terpenoids, and lipophilic molecules (including piperine, osthol, and cholesterol esters) further enrich the metabolomic diversity.

While LC–HRMS analysis identified 46 compounds in the roselle extract, not all constituents contribute equally to the observed pH-responsive colorimetric behavior. The pH sensitivity of the extract is primarily governed by anthocyanin-related compounds and structurally associated flavonoids, particularly glycosylated flavonols such as rutin, quercitrin, and their derivatives. These compounds possess multiple phenolic hydroxyl groups and conjugated aromatic systems that undergo protonation–deprotonation equilibria, directly influencing chromophore stability and color expression across varying pH conditions.

Anthocyanin derivatives and flavonols share common structural motifs, including hydroxylated B-rings and extended π-electron conjugation, enabling reversible electronic rearrangements under acidic and alkaline environments (Trouillas et al., 2016). Glycosylation further modulates this behavior by enhancing aqueous solubility and molecular stability, thereby sustaining chromatic responsiveness within the polymer matrix. In contrast, non-phenolic constituents such as amino acids, alkaloids, and lipophilic metabolites are not directly involved in pH-induced color transitions, but may indirectly support pigment stability by influencing redox balance and intermolecular interactions within the extract.

The internal correlation among these compound classes suggests a cooperative contribution to pH responsiveness, where anthocyanin-derived chromophores act as the primary colorimetric agents, while co-existing flavonoids and phenolic compounds function as stabilizing auxiliaries. This compositional synergy explains the broad pH sensitivity and sustained optical response observed in the Cs/PVA–anthocyanin films, despite the chemical complexity of the roselle extract.

3.1.2. Determination of total anthocyanin content (TAC)

The TAC of Roselle extract (Fig. 1b) was quantified using the pH differential method, which leverages the pH-dependent structural shift of anthocyanins, from red-colored flavylium cations at pH 1.0 to colorless hemiketals at pH 4.5. UV–Vis measurements were taken at 517 and 700 nm (pH 1.0) and 510 and 673 nm (pH 4.5), with turbidity corrections applied. Using cyanidin-3-glucoside as the reference standard, the TAC was determined to be 128.41 mg/L, indicating high extraction efficiency and consistent with previously reported ranges for Roselle extracts (100–250 mg/L), depending on solvent polarity, plant-to-solvent ratio, extraction temperature, and duration (Pham et al., 2019).

The use of 75% ethanol under ambient conditions effectively maximized anthocyanin yield while preserving pigment stability, highlighting Roselle's potential as a natural pH-sensitive dye for intelligent packaging. UV–Vis analysis showed a strong absorbance at pH 1.0 and significant attenuation at pH 4.5, confirming the pH-dependent structural shift. The corresponding color change from deep red to nearly colorless, further supports its suitability as a visual indicator. These results affirm both the efficiency of the extraction method and the chromogenic responsiveness of Roselle anthocyanins for real-time freshness monitoring.

3.1.3. UV–vis spectral behavior of roselle anthocyanins at different pH

The pH-dependent spectral behavior of Roselle anthocyanin extract was evaluated across pH 3–12 using UV–Vis spectrophotometry (Fig. 1c), revealing stepwise chromatic and structural transformations. In acidic conditions (pH 3–6), the extract showed a pink to red coloration with a prominent absorbance peak at ∼520 nm. The highest intensity occurred at pH 3, attributed to the dominance of the flavylium cation or carbinol pseudobase forms, which are stabilized in low-pH environments (Pereira, De Arruda, & Stefani, 2015).

As the pH increased to 7–9, the color gradually shifted to violet, accompanied by a bathochromic shift in λ_max from 520 nm to ∼580 nm. This transition corresponds to deprotonation of hydroxyl and methoxy groups, promoting the formation of quinoidal base species with extended π-conjugation. Although conjugation increased, a moderate decrease in peak intensity was observed, indicating lower molar absorptivity compared to the highly conjugated flavylium form.

Under alkaline conditions (pH 10–12), the extract's color shifted toward yellow to pale brown. The UV–Vis spectra exhibited significantly reduced intensity and curve flattening, consistent with chalcone formation and partial pigment degradation due to irreversible ring-opening reactions. The progressive hypochromic effect and spectral loss at pH ≥11 suggest diminished pigment stability and limited reversibility at high pH extremes.

The distinct and reversible spectral responses, including bathochromic and hypochromic shifts accompanied by clear visual transitions, highlight roselle anthocyanins as structurally dynamic chromophores. Their responsiveness across a broad pH range supports their application as pH-sensitive indicators in intelligent packaging systems for freshness monitoring.

3.2. Characterization of colorimetric films

3.2.1. Visual and colorimetric characteristics of films with varying anthocyanin concentrations

The visual appearance of CPAR films (CPAR-1 to CPAR-5) showed a gradual increase in color intensity with higher roselle anthocyanin loading (Fig. 2a), yielding deeper reddish-purple tones due to elevated pigment content and enhanced pigment–polymer interactions. This behavior aligns with anthocyanins' known chromogenic properties, supporting their role as pH-sensitive dyes in smart packaging.

Fig. 2.

Characteristic of colorimetric films. (a) Visual appearance and corresponding RGB color intensity values derived from ImageJ analysis; (b) FESEM micrographs at 250× magnification; (c) FTIR spectra; and (d) Thickness measurement.

Prior to drying, the film-forming solutions had a pH of ∼4.0, favoring the red flavylium cation form. After drying, the pH slightly increased to ∼5.0, likely due to partial acetic acid loss and matrix rearrangement. Despite this shift, color remained stable, indicating effective pH buffering in the film matrix.

Quantitative colorimetric analysis was performed using ImageJ, based on standardized images of each film captured under uniform lighting. Mean RGB values were extracted from multiple regions of interest across each film surface to ensure representative data. The RGB trends aligned with visual observations, confirming that increased anthocyanin content enhanced overall color intensity and red–blue balance. These results validate the effective incorporation of roselle anthocyanins into the biopolymer matrix and highlight the tunability of film appearance with extract loading, crucial for precise visual cues in freshness monitoring applications.

3.2.2. Surface morphology characterization via FESEM

FESEM analysis (Fig. 2b) was conducted to elucidate the microstructural evolution of chitosan–PVA (Cs/PVA) films upon incorporation of roselle anthocyanin extract. This characterization provides mechanistic insight into polymer–pigment interactions that govern film uniformity, compatibility, and functional stability.

The control Cs/PVA film exhibited a smooth, compact, and defect-free surface, consistent with strong intermolecular hydrogen bonding and chain entanglement between chitosan and PVA reported in previous polymer-blend studies. Incorporation of low pigment concentrations (CPAR-1 and CPAR-2) preserved this uniform morphology, although fine particulates were detectable. These features suggest that anthocyanins were successfully dispersed within the polymer network, likely forming additional hydrogen bonds with Cs/PVA hydroxyl and amine groups. Such intermolecular interactions stabilize pigment distribution and support consistent optical behavior under external pH stimuli.

At intermediate concentrations (CPAR-3 and CPAR-4), the films displayed increased surface roughness and the emergence of pigment-rich microdomains. These morphological changes indicate the onset of partial phase separation, a phenomenon commonly observed when the solubility limit of polyphenolic compounds in hydrophilic polymer matrices is exceeded. Excess anthocyanins may undergo π–π stacking or self-association, reducing polymer–pigment compatibility and leading to localized aggregation, as suggested by prior anthocyanin–biopolymer studies (Abdelghany, Menazea, & Ismail, 2019).

At the highest loading (CPAR-5), the film structure became markedly coarse, with large, irregular aggregates disrupting surface continuity. This microstructural deterioration reflects pigment oversaturation, which compromises polymer cohesion and may affect mechanical stability and light-transmission properties.

The FESEM data demonstrate that anthocyanin incorporation is structurally favorable at low to moderate concentrations, whereas excessive loading induces aggregation-driven heterogeneity. This mechanistic understanding underscores the importance of optimizing pigment levels to maintain both structural integrity and reliable colorimetric performance in smart sensor systems.

3.2.3. Functional group analysis via FTIR spectroscopy

The chemical interactions and functional group compositions of the fabricated films were evaluated using FTIR spectroscopy. Fig. 2c presents the FTIR spectra of pristine Cs/PVA film (a) and CPAR films with increasing anthocyanin concentrations (b–f). The spectra revealed characteristic absorption bands corresponding to the principal functional groups within the biopolymer matrix.

The Cs/PVA control film exhibited a broad and intense absorption band centered at 3274.58 cm⁻¹, attributable to overlapping O—H and N—H stretching vibrations, which originate from the hydroxyl groups of PVA and the amine/hydroxyl groups of chitosan. This band is indicative of strong hydrogen bonding within the polymer matrix. Additional characteristic peaks were observed at 2938 cm⁻¹ (C—H stretching), 1644.22 cm⁻¹ (amide I, C O stretching), 1557.42 cm⁻¹ (amide II, N—H bending), 1409.10 cm⁻¹ (CH–CH₂ bending from PVA), 1250.39 cm⁻¹ (C—N stretching), and 1069.66 cm⁻¹ (C—O stretching). These bands collectively confirm the presence of hydroxyl, amine, and amide functionalities, key reactive sites responsible for hydrogen bonding and polymer chain interactions essential to film integrity (Sani, Tavassoli, Hamishehkar, & McClements, 2021).

Upon incorporation of roselle anthocyanins (CPAR-1 to CPAR-5), the FTIR spectra remained largely similar to the Cs/PVA baseline, with no appearance of new major peaks. This is expected, as the characteristic anthocyanin vibrational bands—such as phenolic O—H stretching at ∼3342 cm⁻¹, C=O stretching at ∼1636 cm⁻¹, and ester C–O/C–O–C stretching at ∼1026 cm⁻¹—partially overlapped with those already present in the chitosan–PVA network (Bertolo, Martins, Horn, Brenelli, & Plepis, 2020). However, close spectral inspection revealed subtle but significant changes in peak characteristics, particularly within the O–H/N–H and C=O regions.

Notably, the broad O–H/N–H stretching band exhibited both a shift in wavenumber and increased band broadness with rising anthocyanin content, moving from 3274.58 cm⁻¹ in the control film to a range of 3278.86–3294.55 cm⁻¹ in anthocyanin-loaded films. This shift, along with enhanced peak intensity, indicates stronger and more extensive hydrogen bonding between the phenolic groups of anthocyanins and the polymer chains of chitosan and PVA. Furthermore, the amide I peak at 1557.42 cm⁻¹ experienced an upward shift to 1630.16–1637.29 cm⁻¹ in the anthocyanin-loaded films, suggesting perturbation of the C O vibrational environment due to hydrogen bonding with anthocyanin carbonyl or hydroxyl groups.

These spectral modifications, although subtle, provide compelling evidence of molecular-level interactions between roselle-derived anthocyanins and the Cs/PVA matrix. Such interactions are vital in stabilizing the pigment within the film network, thereby enhancing its functional performance as a pH-responsive and color-stable indicator. Collectively, FTIR analysis confirms the successful integration of anthocyanins and elucidates the intermolecular forces, primarily hydrogen bonding, that govern the physicochemical stability and uniformity of the resulting colorimetric films.

3.2.4. Thickness measurement of colorimetric films

Film thickness is a crucial parameter that influences the mechanical integrity, barrier function, and optical response of colorimetric films. As presented in Fig. 2d and Table S2, a gradual increase in thickness was observed with higher volumes of roselle anthocyanin incorporated into the Cs/PVA matrix. This trend is primarily due to enhanced hydrogen bonding and hydrophobic interactions between anthocyanin molecules and the biopolymer chains, which promote tighter structural packing and increase material density.

The lowest thickness was found in CPAR-1 (0.061 ± 0.01 mm), containing only 1 mL anthocyanin, while CPAR-5, with the highest anthocyanin volume, exhibited a significantly thicker structure (0.142 ± 0.02 mm). This increase reflects the higher solid content in the casting solution, leading to the formation of thicker films during solvent evaporation. Additionally, the thickening effect can be attributed to enhanced electrostatic interactions and increased crosslinking frequency between polyphenolic anthocyanins and the amino or hydroxyl functional groups of chitosan and PVA. These interactions contribute to the formation of a denser three-dimensional polymeric network, which retains more extract and slows solvent diffusion during drying, resulting in more compact film formation (Yong et al., 2019).

These results align with previous reports that incorporation of bioactive compounds into biopolymer films generally increases thickness due to both physical and chemical matrix modifications. However, although greater thickness may improve pigment stability and barrier function, overly thick films may reduce transparency and delay optical response, thus affecting sensitivity in freshness monitoring applications.

3.3. Colorimetric properties and pH sensitivity of film formulations

Table 1 shows the pH-dependent colorimetric response of Cs/PVA films, evaluated using CIELAB parameters (L*, a*, b*) and ΔE. The unmodified Cs/PVA film exhibited high color stability, with ΔE values below the human perceptibility threshold, indicating minimal chromogenic activity yet its structural neutrality makes it an ideal carrier matrix.

Table 1.

Colorimetric responses of film formulations across pH values ranging from 5.0 to 12.0, based on CIELAB coordinates and ΔE values.

Note: The notation with different “letters” indicates a significant difference (p < 0.05).

In contrast, incorporating roselle anthocyanins into the polymer matrix significantly improved pH responsiveness. CPAR-1 showed moderate colorimetric activity, with ΔE rising from 11.00 to 19.90. This shift was mainly due to an increase in b* (yellow tones) and a slight rise in a* (red-green component), indicating partial chromophore transformation under mildly alkaline conditions.

Higher anthocyanin concentrations led to enhanced film sensitivity. CPAR-4 showed a significant ΔE of 35.60, mainly due to b* changes, while L* remained stable, indicating clear visual contrast during color transitions. This response is ideal for monitoring freshness in the pH range of 7–10, where amine release occurs in protein-rich foods.

CPAR-5 showed the highest chromogenic response, with a ΔE of 44.69, indicating a strong and noticeable color change. Shifts in a* and b* signified a hue transition from red-purple (acidicl) to gray (neutral), greenish and yellow (alkaline), consistent with anthocyanin transformations from flavylium to quinoidal base and chalcone forms. This high sensitivity is attributed to the elevated anthocyanin content, which enhanced pigment dispersion and interaction within the polymer matrix.

The strong chromogenic response of CPAR-4 and CPAR-5 can be attributed to the hydrophilic nature of anthocyanins, which enhances their interaction with aqueous media. The high water content and lower ethanol concentration in the casting solution facilitated better anthocyanin solubilization and diffusion within the Cs/PVA matrix, intensifying visible color changes (Alizadeh-Sani et al., 2021). These findings confirm the excellent pH sensitivity and visual clarity of the films, supporting their potential as effective, interpretable in

3.4. Colorimetric response of CPAR films to ammonia vapor exposure

Spoilage of protein-rich foods like chicken, shrimp, and tilapia produces volatile basic nitrogen compounds (e.g., dimethylamine, trimethylamine, NH₃) through microbial metabolism, raising pH and creating an alkaline headspace. To assess CPAR films as freshness indicators, their colorimetric responses to ammonia vapor were evaluated (Table 2).

Table 2.

Visual color changes of CPAR films in response to 0.1, 0.5, and 1% ammonia vapor exposure.

The films exhibited clear chromogenic transitions from gray to brown or purplish-red, depending on anthocyanin content and ammonia levels. CPAR-1 showed the lowest sensitivity (∼7–8%) with a slow response plateauing after 120 min (Fig. 3), likely due to limited NH₃ adsorption and a low density of reactive functional groups or pigments. In contrast, CPAR-2 reached ∼12% sensitivity with better differentiation across NH₃ concentrations, attributed to its moderate anthocyanin content, which improved pigment dispersion and enhanced molecular affinity toward basic volatiles.

Fig. 3.

Time-dependent colorimetric %sensitivity of CPAR films under varying ammonia vapor concentrations (n = 3): (a) CPAR-1, (b) CPAR-2, (c) CPAR-3, (d) CPAR-4, and (e) CPAR-5.

A significant increase in performance was observed in CPAR-3, which achieved an S% of 16–17%. The color response stabilized between 90 and 150 min, indicating a balanced interaction between pigment reactivity and matrix stability. This suggests that CPAR-3 approached saturation without degradation, making it suitable for moderate ammonia detection. CPAR-4 demonstrated a marked sensitivity enhancement, with S% peaking at ∼35%. Its response strongly correlated with ammonia concentration, indicating dynamic interaction between the anthocyanin-rich matrix and the volatile base. However, at the highest tested ammonia concentration (1% w/v), sensitivity declined after 120 min, likely due to saturation of binding sites or partial anthocyanin degradation under prolonged alkaline exposure. CPAR-5 exhibited the highest responsiveness, reaching ∼42% within the first 90 min, indicating rapid and efficient interaction with ammonia vapor. Still, a notable decline in response at 1% NH₃ after 120 min suggests a saturation or overexposure effect, potentially disrupting chromogenic stability through pigment breakdown or polymer matrix swelling.

Higher anthocyanin concentrations led to more intense, ammonia-dependent color changes, particularly in CPAR-4 and CPAR-5, which exhibited strong visual responses at elevated NH₃ levels. This enhanced reactivity is attributed to increased pigment availability and improved molecular interactions within the Cs/PVA matrix, supporting the films' potential as tunable, non-invasive freshness indicators.

To evaluate its performance, the CPAR-5 film was compared with other anthocyanin-based biopolymer sensors. As summarized in Table S3, CPAR-5 showed the highest ammonia sensitivity (42%), outperforming starch/PVA (30.77%), starch/chitosan (9.54%) (Zhang et al., 2019), starch–PVA with H. sabdariffa (<5%) (Zhai et al., 2017), and agar–butterfly pea extract (17%) (Hashim et al., 2022). These findings confirm the superior responsiveness of the chitosan–PVA matrix used in this study.

The superior sensitivity of CPAR-5 stems from the synergistic properties of the chitosan–PVA matrix. Chitosan, a biocompatible cationic polysaccharide, possesses abundant amino groups that form strong hydrogen bonds with the hydroxyl and carbonyl groups of anthocyanins, enhancing pigment retention. PVA, a water-soluble synthetic polymer with excellent film-forming and mechanical properties, complements chitosan by improving matrix uniformity and stability. This dual-polymer system stabilizes and disperses anthocyanins efficiently, preserving their pH-responsiveness under ammonia exposure. Overall, CPAR-5's enhanced performance highlights the chitosan–PVA matrix as a promising platform for real-time, non-invasive freshness sensing, outperforming starch- and agar-based systems.

3.5. Application of colorimetric films for monitoring chicken, shrimp, and tilapia freshness

Based on the results of pH responsiveness (Table 1) and ammonia vapor sensitivity (Table 2), the CPAR-5 film was selected as the optimal formulation for real-world application due to its superior colorimetric performance. This formulation demonstrated the highest total color difference (ΔE ≈ 44.69) and the fastest ammonia vapor response (S% ≈ 42% within 75 min), enabling clear, rapid, and visually distinguishable transitions in response to environmental changes associated with food spoilage.

To translate this molecular responsiveness into practical usability, a visual card reader was developed based on the continuous pH-dependent color evolution of the CPAR-5 film. As illustrated in Fig. 4a, the film exhibited a gradual and distinguishable color shift from pinkish-red under acidic conditions to greenish-yellow at alkaline pH, covering the range typically associated with freshness deterioration in protein-based foods. This continuous chromatic evolution enables intuitive visual interpretation without the need for analytical instruments.

Fig. 4.

Visual reference chart (card reader) (a) for interpreting pH-dependent color transitions of the CPAR-5 film and reaction mechanism (b).

When applied to chicken, shrimp, and tilapia under both ambient and refrigerated storage, the CPAR-5 film displayed progressive color changes that reflected spoilage development in real time. The observed color responses were consistent with known pH elevation trends arising from protein degradation and the accumulation of volatile basic compounds during storage. By visually comparing the film color with the reference chart, users could semi-quantitatively evaluate freshness status directly on-site.

As depicted in Fig. 4b, the colorimetric response of the CPAR-5 film arises from pH-driven structural transformations of anthocyanins rather than direct covalent interactions with spoilage volatiles. Under acidic conditions, anthocyanins exist predominantly as flavylium cations, producing a red coloration due to an extended π-conjugated system. As pH increases, hydration and deprotonation induce conversion to carbinol pseudobase and quinoidal base structures, resulting in purple to bluish hues. Further alkalization leads to anionic quinoidal species and eventual chalcone formation, manifested as grayish to yellow color transitions. During spoilage, volatile ammonia diffuses into the Cs/PVA matrix and hydrolyzes, elevating the local pH and shifting the anthocyanin equilibrium toward alkaline forms. The hydroxyl and amino groups within the Cs/PVA matrix facilitate rapid proton exchange and stabilize intermediate anthocyanin species through hydrogen bonding, enhancing sensitivity and response clarity. These reversible, non-covalent interactions underpin the rapid and visually distinguishable chromogenic behavior of CPAR-5, supporting its effectiveness for real-time freshness monitoring.

The integration of CPAR-5 with a visual card reader establishes a low-cost, user-friendly, and scalable strategy for smart sensor applications. By translating laboratory-based colorimetric responses into an intuitive, instrument-free format, this system enhances the practical deployment of smart sensor across retail distribution, cold-chain logistics, and household storage. Importantly, this approach reduces dependence on electronic sensing devices, thereby improving accessibility, sustainability, and real-world applicability for decentralized food quality monitoring.

3.6. Monitoring freshness of chicken, shrimp, and tilapia under ambient conditions

Time-resolved RGB intensity values from CPAR-5 films were plotted against storage duration for tilapia, shrimp, and chicken under ambient and refrigerated conditions (Fig. S1), whereas Table 3 provides an integrated summary of freshness degradation at 25–27 °C over 24 h based on pH shifts and CPAR-5 color transitions. The rising pH reflects the release of spoilage-related volatiles, such as ammonia and biogenic amines, from microbial activity and protein decomposition (Mary et al., 2020).

Table 3.

pH evolution and CPAR-5 visual color changes in chicken, shrimp, and tilapia samples during storage at ambient temperature (25 °C).

Tilapia's pH rose from 5.0 to 6.0 within 6–12 h, then sharply to 7.0 and 11.0 by 18–24 h. As fish is deemed spoiled at pH ≥ 7.0 due to bacterial spoilage and proteolytic degradation (Salgado et al., 2021), this threshold was exceeded at 18 h. The film color shifted from purplish-pink to grayish-green, confirming the sensor's ability to visually detect tilapia spoilage.

For shrimp, the pH increased rapidly from 5.0 at 0 h to 7.0 within 6 h, then reached 9.0 and 11.0 at 12 and 18 h, respectively. As shrimp are highly perishable, pH ≥ 7.4 is considered a spoilage threshold due to elevated volatile bases and microbial activity (Wu et al., 2021). In this study, that threshold was crossed at 12 h, underscoring the need for early detection and demonstrating the effectiveness of CPAR-5 films in monitoring seafood freshness.

During ambient storage, the pH of chicken increased from 5.0 to 7.0 at 18 h and reached 11.0 by 24 h. As reported, chicken is considered microbiologically unacceptable above pH 6.8 due to spoilage bacteria and nitrogenous compound accumulation (Ahmad et al., 2020). In this study, the spoilage threshold was exceeded at 18 h, confirmed by a CPAR-5 film color shift from purplish-red to grayish-green/yellow, demonstrating the film's effectiveness in indicating poultry freshness.

These results confirm the pH-sensitive performance of CPAR-5 films for real-time freshness monitoring. The films showed clear visual changes aligned with spoilage thresholds chicken (≥ pH 6.8, 18 h), shrimp (≥ pH 7.4, 12 h), and tilapia (≥ pH 7.0, 18 h) demonstrating their effectiveness in tracking spoilage kinetics across various food types under ambient conditions.

3.6.1. Monitoring freshness of tilapia, shrimp, and chicken under refrigerated storage

Table 4 presents the freshness profiles of tilapia, shrimp, and chicken during refrigerated storage (6 °C) over 11 days, integrating time-resolved RGB intensity evolution (Fig. S2) with corresponding pH changes. The progressive variations in RGB values as a function of storage duration and pH quantitatively reflect the gradual accumulation of volatile basic nitrogen compounds associated with microbial and enzymatic spoilage, thereby validating the responsiveness of CPAR-5 films under cold-chain conditions.

Table 4.

Changes observed in meat samples during refrigerated storage.

Tilapia maintained a stable pH of 5–6 and a consistent pink–purplish film color during the first five days of refrigeration, indicating minimal microbial activity. From day 7, the pH rose toward 7, with a visible color shift to grayish tones by day 8, marking the onset of spoilage. On days 9–11, pH reached 9, and the film turned gray-green to yellowish, indicating deterioration beyond the acceptable threshold for fish freshness (pH ≥ 7.0) (Salgado et al., 2021). These results confirm the CPAR-5 film's effectiveness in detecting spoilage progression under cold storage.

Shrimp exhibited a faster spoilage rate than tilapia. While pH remained between 6 and 8 until day 3, it rose to pH 9 by day 5, accompanied by a film color change from pink to gray. From day 6 onward, the film darkened as pH increased to 10–12 by days 10–11, indicating advanced microbial spoilage. Based on the spoilage threshold of pH ≥ 7.4 (Wu et al., 2021), shrimp exceeded its freshness limit between days 2 and 3. CPAR-5 films effectively signaled this transition, highlighting their high ammonia sensitivity even under refrigerated conditions.

Chicken samples exhibited a gradual pH increase from 7 to 8 within four days, reaching 9 on day 7 and peaking at 12 by day 11. The CPAR-5 film displayed corresponding color changes from pink to grayish-green/yellow. Since chicken is considered spoiled at pH ≥ 6.8 (Ahmad et al., 2020), spoilage likely began around day 4–5. The reddish-to-olive shift in the film reliably indicated this degradation.

Across all three food matrices, CPAR-5 films consistently showed pH-responsive color changes aligned with spoilage progression across all samples. Their stability under refrigeration confirms anthocyanins' resilience and continued reactivity with volatile bases. The films effectively captured spoilage rates (faster in shrimp and chicken, slower in tilapia) highlighting their suitability for intelligent packaging as low-cost, visual freshness indicators for enhanced food safety and reduced waste.

4. Conclusion

This study presents a biodegradable chitosan–polyvinyl alcohol (Cs/PVA) colorimetric film incorporating roselle-derived anthocyanins as a platform for real-time freshness monitoring of protein-rich foods. Among the investigated formulations, CPAR-5 exhibited the strongest colorimetric response, showing distinct pH- and ammonia-induced color transitions arising from anthocyanin structural transformations within the Cs/PVA matrix. The film enabled visual discrimination of spoilage progression in chicken, shrimp, and tilapia under both ambient and refrigerated storage, consistent with pH increases associated with the accumulation of volatile basic nitrogen compounds. Integration of CPAR-5 into a simple visual card reader converted its continuous chromatic response into an intuitive, instrument-free freshness assessment. The combination of sensitivity, structural stability, and visual interpretability supports the applicability of the Cs/PVA–anthocyanin system as a sustainable indicator for intelligent food freshness monitoring.

While the colorimetric response effectively reflected spoilage progression, the present study did not include direct quantitative validation against total volatile basic nitrogen (TVB-N) levels or microbial counts. Further studies incorporating these reference freshness indicators are required to establish quantitative correlations and strengthen the analytical reliability of the developed sensor for practical food packaging applications.

CRediT authorship contribution statement

Alfianita Nuril Hidayaty: Writing – review & editing, Writing – original draft, Visualization, Software, Formal analysis, Data curation. Saidun Fiddaroini: Writing – review & editing, Writing – original draft, Visualization, Data curation. Ahmad Luthfi Fahmi: Writing – review & editing, Writing – original draft, Visualization, Validation. Stephani Rebecca Amabel Siahaan: Validation, Investigation, Formal analysis, Data curation. Qonitah Fardiyah: Supervision, Methodology, Conceptualization. Arie Srihardyastutie: Supervision, Methodology, Data curation. Akhmad Sabarudin: Writing – review & editing, Writing – original draft, Resources, Funding acquisition, Data curation, Conceptualization.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT to enhance the readability of the English text.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.