An intelligent delivery system for apricot preservation: pH-responsive hollow ZIF-8 loaded with lavender essential oil and its preservation efficacy

Abstract

Lavender essential oil (LEO) has strong antibacterial and antioxidant properties, but its volatility limits practical use in fruit preservation. To address this limitation, we developed a pH-responsive delivery system (LEO@HZIF-8) using tannic acid (TA)-etched hollow ZIF-8 (HZIF-8) to enable controlled release of LEO. Mild etching generated a hollow structure that increased LEO encapsulation efficiency to 68.96%. Results indicate that LEO@HZIF-8 nanoparticles retain a well-defined crystalline structure and release LEO rapidly under acidic conditions, with a release rate of 73.36% at pH 4.5 after 72 h. Furthermore, LEO@HZIF-8 exhibited strong antioxidant activity, with DPPH and ABTS radical-scavenging rates of 85.48% and 78.24%, respectively, and significant antibacterial activity. When incorporated into a sodium alginate/zein (SA/Zein) coating, LEO@HZIF-8 extended the shelf life of white apricots to 63 days and effectively delayed decay, weight loss, and firmness loss. This study demonstrates an efficient pH-responsive delivery strategy for plant essential oils in freshness preservation and provides a basis for developing next-generation smart, eco-friendly active food packaging materials.

Highlights

• •A TA-modified hollow ZIF-8 nanomaterial with high LEO loading (68.96%) was synthesized.
• •Hollow ZIF-8 serves as a carrier for achieving pH-responsive sustained release of LEO.
• •LEO@HZIF-8 exhibits outstanding antioxidant and antibacterial activity.
• •LEO@HZIF-8/Zein/SA coating effectively extends the shelf life of white apricots (63 days).

1. Introduction

Apricots (Prunus armeniaca), a nutrient-rich fruit with a distinctive flavor, are highly susceptible to quality deterioration and microbial contamination during post-harvest storage due to their high water content and tender tissue (Yang et al., 2024), resulting in significant economic losses. Traditional preservation methods often fail to provide precise, durable antimicrobial protection, particularly for climacteric fruits with a postharvest climacteric respiration peak. Consequently, smart packaging systems that respond to changes in the storage microenvironment and enable controlled release of active agents have become a major focus in postharvest preservation research (Fu et al., 2022; Jiang et al., 2024).

Plant essential oils (EOs) serve as natural antimicrobial agents. Lavender essential oil (LEO), derived from Lavandula angustifolia, is rich in bioactive constituents, including linalool, linalyl acetate, and other terpenoids (Sun et al., 2021). LEO is considered a natural alternative to chemical preservatives due to its antibacterial (Hirsch et al., 2024), antifungal (Ding et al., 2023), antioxidant (Liu et al., 2022), and preservative activities. However, its practical use in freshness preservation is limited by high volatility, photo-thermal instability, and poor dispersibility in aqueous solutions. To improve the bioavailability and sustained efficacy of EOs, recent studies have either incorporated EOs directly into natural biopolymer networks (Bhatia et al., 2023; Gómez-Contreras et al., 2021) or encapsulated them in delivery matrices such as nanoparticles (Muthalagu et al., 2024; Zhang et al., 2024), microcapsules (Shi et al., 2021), nanoemulsions (Guo et al., 2022). However, current EO-based systems still face several critical challenges, including uneven dispersion of essential oils, a limited range of antimicrobial and antioxidant strategies, uncontrolled release behavior, low encapsulation efficiency, and limited environmental responsiveness, which greatly restrict their practical applications. Metal-organic frameworks (MOFs) are crystalline porous inorganic-organic materials that exhibit well-defined porosity and chemical stability, and they have been widely applied in gas sensing (Ram Kumar et al., 2020), energy storage (Lu et al., 2023), catalysis (Yang et al., 2020), sensing (Mohanty et al., 2024), and drug delivery (Khafaga et al., 2024). Zeolitic imidazolate framework-8 (ZIF-8) is a widely studied metal-organic framework. Constructed from Zn²⁺ nodes coordinated with 2-methylimidazolate linkers, ZIF-8 provides a high specific surface area, tunable pore size, good biocompatibility, and low reactivity toward typical substrates. These properties make ZIF-8 suitable for food-contact applications and have motivated its use in drug delivery (Nguyen et al., 2026; Rahman et al., 2025) and active food packaging (Yin et al., 2024). Xiao et al. (Xiao et al., 2021) reported that ZIF-8, owing to its alkali-resistant yet acid-sensitive behavior, dissociates rapidly in mildly acidic microenvironments associated with fruit decay or pathogen metabolism, thereby enabling targeted, on-demand release of antimicrobial agents. However, pristine ZIF-8 suffers from extremely low encapsulation efficiency toward bioactive compounds and severe particle agglomeration, which further restrict its loading and controlled-release performance. Pandey et al. (Pandey et al., 2024) reported encapsulation efficiencies of morin, rutin, and rosmarinic acid in one-step, in situ–grown ZIF-8 of 5%, 9%, and 16%, respectively.

To overcome these limitations, surface modification of ZIF-8 is an effective strategy (Nguyen et al., 2025; Nguyen et al., 2026). In recent years, natural plant polyphenols such as tannic acid (TA) have been widely employed to modify ZIF-8 and other metal-organic frameworks because of their universal adhesion, excellent hydrophilicity, and strong metal-ion chelation (Wang et al., 2024). This study employs TA for surface modification of ZIF-8. As a natural polyphenol, TA induces hollow structure formation by mildly etching ZIF-8, thereby increasing encapsulation efficiency (Xu et al., 2025), and it improves the material's hydrophilicity and dispersibility via abundant phenolic hydroxyl groups (Mu et al., 2025). Notably, this study proposes, for the first time, a triple-synergistic preservation system that utilizes TA for the surface functionalization and assisted etching of ZIF-8 to obtain hollow porous nanomaterials. Beyond the aforementioned advantages, this material serves as a wall material to load LEO, thereby exerting multi-component synergistic antimicrobial and antioxidant effects and further enhancing the performance of the preservation system.

Building on this foundation, we developed a composite coating film composed of sodium alginate (SA) and zein to provide sustained release of the nanomaterials on the fruit surface and to reduce the risk of direct contact with the fruit. This film acted as a matrix to immobilize LEO-loaded, TA-modified ZIF-8 (LEO@HZIF-8), enabling controlled release of natural antimicrobial agents. As a hydrophobic protein, zein exhibits excellent film-forming properties (Geng et al., 2023) and effectively disperses and immobilizes TA-modified ZIF-8 nanoparticles, thereby preventing agglomeration. Meanwhile, SA, a hydrophilic polysaccharide, forms a dense three-dimensional network. Owing to its good film-forming ability and gas-barrier properties, the coating modulates fruit respiration rate and thereby delays metabolic senescence (Bhatia et al., 2022). The novelty of this study lies in the fact that cross-linking between SA and zein, two naturally derived biodegradable polymers, enabled successful encapsulation of the triple-antimicrobial and antioxidant LEO@HZIF-8 nanoparticles within the composite coating film. The resulting system provides excellent process controllability, extended long-term freshness preservation, and coordinated release of active agents. Within this system, the outer SA/Zein polymer network functions as a physical barrier. The internal LEO@HZIF-8 nanoparticles respond to pH changes in the fruit microenvironment and release antimicrobial agents in a pH-triggered manner, thereby achieving intelligent synergistic preservation.

In summary, this study aims to develop a pH-responsive composite coating film based on LEO@HZIF-8/SA/Zein. The system enables precise pH-responsive on-demand release of lavender essential oil, significantly improves the encapsulation efficiency and stability of LEO and achieves efficient, intelligent preservation of white apricots by prolonging shelf life and maintaining postharvest quality through multi-component synergies. This strategy not only addresses the stability and release-control challenges encountered in practical applications of plant essential oils but also provides a novel and feasible technical route for developing next-generation MOF-based active food packaging materials.

2. Materials and methods

2.1. Materials and reagents

ZIF-8 (model XFF28-1; particle size 100–400 nm; specific surface area ∼1400 m²/g; pore size 0.8–1.8 nm) was obtained from Jiangsu Xianfeng Nanomaterials Technology Co., Ltd. Sodium alginate (SA; ultra-high viscosity, type II; 1% viscosity 7000–10,000 mPa·s), zein (corn-derived), tannic acid (TA; 95%), and lavender essential oil (LEO; CAS: 8000-28-0) were purchased from Shanghai Aladdin Chemical Co., Ltd. Anhydrous ethanol, glycerol, and other reagents were of analytical grade and were purchased from Tianjin Xinbote Chemical Co., Ltd. The test strains, Escherichia coli (E. coli) and Staphylococcus aureus (S. aureus), were provided by the Microbiology Laboratory of Kashi University. White apricots were purchased from a local farmers' market in Kashi, Xinjiang, China.

2.2. Preparation of LEO@HZIF-8 nanoparticles and coating solution

Using the previously reported method (Maghsoudi et al., 2024), ZIF-8 (0.2 g) was uniformly dispersed in tannic acid aqueous solution (5 g/L, 100 mL) by ultrasonication and aged at room temperature for 10 min. The etched nanoparticles were collected by centrifugation at 8000 rpm for 5 min, the supernatant was removed, and the pellet was washed three times with water and anhydrous ethanol. The final precipitate was collected by centrifugation and freeze-dried under vacuum for 2 h to obtain HZIF-8.

For loading experiments, LEO volumes of 2, 4, 6, 8, and 10 μL (equivalent to 1.764, 3.528, 5.292, 7.056, and 8.82 mg, respectively) were each dispersed in 5 mL of anhydrous ethanol. HZIF-8 (8.82 mg) was added to each dispersion and the mixtures were shaken for 3 h. The resulting LEO@HZIF-8 nanoparticles were collected by centrifugation, washed three times with anhydrous ethanol to remove surface-adsorbed LEO, and collected again by centrifugation. The pellets were then freeze-dried under vacuum for 2 h to obtain LEO@HZIF-8 with different LEO loadings.

Weigh 0.05 g of corn zein and dissolve it in 10 mL of 80% v/v ethanol-water solution. Dissolve 0.15 g of SA in 10 mL of water to prepare the film-forming solution. Add 1.5% w/v LEO@HZIF-8 nanomaterial to the solution, followed by 1.5 g of crosslinking agent glycerol. After magnetic stirring at room temperature for 1 h, the mixture was left to stand for defoaming, yielding the LEO@HZIF-8/SA/Zein film solution (LHSZ). The membrane solution without added nanomaterials, namely the SA/Zein (SZ), was stored at room temperature for later use.

2.3. Feature description

Intermolecular interactions were assessed by Fourier transform infrared spectroscopy (FT-IR) using an IRAffinity spectrometer (Shimadzu, Japan) in the range of 500–4000 cm⁻¹ at a resolution of 4 cm⁻¹. Powder crystal structures were determined by X-ray diffraction (XRD) on an X'Pert Powder diffractometer equipped with an automatic sample changer (PANalytical, The Netherlands). Microstructural features and elemental composition were examined by field-emission scanning electron microscopy (FE-SEM; SU5000, Hitachi High-Tech, Japan) and transmission electron microscopy (TEM; HT7800, Hitachi High-Tech, Japan) coupled with energy-dispersive X-ray spectroscopy (EDS; X-MaxN 80, Oxford Instruments, UK). Thermal stability was evaluated by thermogravimetry–differential scanning calorimetry (TG-DSC) using a STA 449 F3 simultaneous thermal analyzer (NETZSCH, Germany) under a nitrogen atmosphere from 30 °C to 800 °C at a heating rate of 10 °C min⁻¹. Specific surface area and pore characteristics were quantified using an Autosorb 6100 FKM MP-MP surface area and porosity analyzer (Quantachrome Instruments, USA).

2.4. LEO encapsulation efficiency

The encapsulation efficiency and loading efficiency of LEO@HZIF-8 were quantified at a detection wavelength of 213 nm using a Ultraviolet-visible spectrophotometer (UV-1800, Shimadzu, Japan). A calibration curve was constructed for LEO solutions over 0.003–0.015 mg/mL. Encapsulation efficiency (EE, %) and loading efficiency (LE, mg/g) were calculated (Akshaya, Kr & Reshma, 2025).

$$A=54.2667C-0.0492\left(R^2=0.9996\right)$$ (1)

In the formula: A, the absorbance; C, the concentration of LEO (mg/mL).

$$\mathrm{EE}\left(\%\right)=M_t/M_0\times 100$$ (2)

$$\mathrm{LE}\left(\mathrm{mg}/\mathrm{g}\right)=M_t/M_1\times 100$$ (3)

In the formula: M_t, mass of LEO actually loaded in the sample (mg); M₀, initial mass of LEO added during preparation (mg); M₁, total mass of the dried LEO@HZIF-8 sample (mg).

2.5. LEO release capacity

LEO release was investigated in phosphate-buffered saline (PBS) at pH 4.5, 7.4, and 9.0. Briefly, 10 mg of LEO@HZIF-8 was dispersed in 10 mL of PBS and stirred at room temperature. At designated time points (0, 3, 6, 9, 12, 24, 36, 48, and 72 h), 1 mL of suspension was collected and immediately replenished with 1 mL of fresh PBS at the corresponding pH to maintain a constant volume (Chen et al., 2025; Muthalagu & Natarajan, 2024). LEO concentrations were quantified by UV–Vis spectroscopy at 213 nm using the previously established calibration curve, and cumulative release (%) was calculated to evaluate pH-responsive controlled-release behavior. The specific equation is given as follows (Rahman et al., 2025):

$$\text{Cumulative Release}\left(\%\right)=\left(m_1+m_2+m_3+\cdots +m_n\right)/M\times 100$$ (4)

In the formula: M_n (n = 1, 2, 3, …), mass of LEO in PBS at the predetermined time (mg); M, actual mass of lavender essential oil added during the preparation of LEO@HZIF-8 (mg).

To elucidate release kinetics under different pH conditions, cumulative-release profiles were fitted to four kinetic models, including zero-order, first-order, Higuchi, and Ritger-Peppas, according to the equations below (Liu et al., 2025):

$${\mathrm{M}}_{\mathrm{t}}/{\mathrm{M}}_{\infty }=\mathrm{kt}+\mathrm{b}$$ (5)

$${\mathrm{M}}_{\mathrm{t}}/{\mathrm{M}}_{\infty }=1-{\mathrm{e}}^{-\mathrm{kt}}$$ (6)

$${\mathrm{M}}_{\mathrm{t}}/{\mathrm{M}}_{\infty }=\mathrm{k}{\mathrm{t}}^{\frac{1}{2}}+\mathrm{b}$$ (7)

$${\mathrm{M}}_{\mathrm{t}}/{\mathrm{M}}_{\infty }=\mathrm{k}{\mathrm{t}}^{\mathrm{n}}$$ (8)

In the formula: M_t/M_∞, the release rate of LEO@HZIF-8 at time t; k, the release rate constant; n, the diffusion parameter.

2.6. UV stability of LEO

Solutions of free LEO and LEO@HZIF-8 with an LEO concentration of 1 mg/mL were prepared in anhydrous ethanol in equal volumes. Samples were irradiated with a 36 W germicidal ultraviolet lamp (213 nm) positioned 15 cm above the specimens. At predetermined intervals (0, 3, 6, 9, 12, 24, 36, 48, and 72 h), 1 mL aliquots were withdrawn and replaced with an equal volume of ethanol. LEO content was quantified by UV–Vis at 213 nm as described for the loading determinations. Photodegradation kinetics were fitted to a first-order model as follows:

$$\ln {\mathrm{C}}_{\mathrm{t}}/{\mathrm{C}}_0=-\mathrm{kt}$$ (9)

In the formula: C_t, the concentration of LEO at time t; C₀, the initial concentration of LEO; k, the release rate constant.

2.7. Antioxidant performance Evaluation

A 0.1 mM DPPH solution was prepared in ethanol and further diluted with anhydrous ethanol to obtain a working solution with an initial absorbance of 0.80 ± 0.02 at 517 nm. ZIF-8, HZIF-8, LEO@HZIF-8, free LEO, and l-ascorbic acid (Vc; positive control) were prepared at 10 mg in 2 mL of 95% (v/v) ethanol (Vc was dissolved in water). For each measurement, 100 μL of sample solution was mixed with the DPPH working solution and adjusted to a final volume of 3.0 mL. The reaction mixtures were incubated in the dark at room temperature for 60 min, after which the absorbance at 517 nm was measured in a 1 cm path-length cuvette (Qanash et al., 2023).

The ABTS radical cation (ABTS•⁺) working solution was generated by mixing 7.4 mM ABTS with 2.6 mM potassium persulfate at a 1:1 (v/v) ratio and incubating the mixture in the dark at room temperature for 12 h. Prior to analysis, the solution was diluted with PBS (pH 7.4) to an absorbance of 0.70 ± 0.02 at 734 nm. ZIF-8, HZIF-8, LEO@HZIF-8, free LEO, and l-ascorbic acid (Vc; positive control) were prepared at 10 mg in 2 mL of 95% (v/v) ethanol (Vc was dissolved in water). For each assay, 10 μL of the sample solution was added to the ABTS•⁺ solution and adjusted to a final volume of 3.0 mL. Absorbance was measured at 734 nm, and ABTS radical-scavenging activity (%) was calculated as the percentage inhibition relative to the control (ABTS•⁺ solution containing the corresponding solvent) (Akhtar et al., 2025).

$$\text{Radical Scavenging Rate}\left(\%\right)=\left(1-A/A_0\right)\times 100$$ (10)

In the formula: A, the experimental group; A₀, the blank group.

2.8. Antimicrobial performance evaluation

Following the method of Fu et al. with appropriate modifications (Wang et al., 2023), the activated bacterial suspension was diluted to 10⁸ CFU/mL. A 100 μL aliquot of the suspension was spread evenly onto solid LB medium. A sterile filter paper disc (6 mm in diameter) was positioned centrally in the culture dish. The filter paper discs were soaked for 20 min in LEO@HZIF-8 suspensions at concentrations of 0, 5, 10, 15, and 20 mg/mL, as well as 20 mg/mL suspensions of ZIF-8, LEO, and TA, respectively. Plates were inverted and incubated at 37 °C for 24 h. The diameters of the inhibition zones were measured using a vernier caliper to evaluate the antibacterial effects of LEO@HZIF-8 nanomaterials and the individual components against E. coli and S. aureus.

2.9. Fruit preservation test

White apricots with uniform size and 70–80% ripeness were selected and subjected to different coating treatments. After coating and air-drying, the fruits were stored at 4 °C. Untreated fruit served as the control (CK), and the treatment groups were polyethylene film (PE), SZ, and LHSZ. Decay incidence, weight-loss rate, pH, and firmness were measured at 7-day intervals until the LHSZ group reached a decay incidence ≥40% and exhibited complete loss of flavor quality, after which measurements were discontinued. Among these, the decay rate and weight loss rate were assessed using the counting method. pH was measured using a high-precision digital refractometer (Sanryo, Japan). Firmness was measured at two points on opposite sides along the equatorial line using an ENS-ipro FTC texture analyzer (Ensoul Technology, China). A stainless-steel cylindrical probe (2 mm diameter) penetrated 6 mm at a speed of 1 mm/s. The maximum force recorded during penetration was expressed in Newtons (N) and taken as the firmness value (Jieying et al., 2025; Saharika et al., 2021).

2.10. Metal ion migration test

The experiment was carried out according to a previous method (Yin et al., 2024). Fresh white apricots were coated with PE film, SZ, and LHSZ coatings, air-dried, and stored at 4 °C for 63 days. After storage, the coatings were removed and the fruits were homogenized. A 3 g sample was dried at 65 °C for 16 h, then digested with 6 mL nitric acid via microwave for 2 h. The acid was evaporated for about 3 h, and the digest was diluted with deionized water and stored at −4 °C. The Zn²⁺ migration from the composite coating into apricots was quantified by iCAP RQ inductively coupled plasma mass spectrometry (Thermo Fisher Scientific, USA).

2.11. Biosafety test

To evaluate the biosafety of the nanocomposite coating, a zebrafish survival test was performed (Guo et al., 2024). 0, 0.3, and 0.6 g of LHSZ were separately placed into round plastic containers containing 600 mL of tap water, followed by the introduction of 10 zebrafish (3 months old, 3–5 cm in body length). Each treatment was conducted in triplicate. The water was renewed and zebrafish were fed daily. The number of surviving zebrafish was recorded regularly over 14 days.

2.12. Statistical analysis

All experiments were performed in triplicate, and data are expressed as mean ± standard deviation (n = 3). IBM SPSS Statistics 27 (IBM Corporation, Chicago, USA) was used to perform One-way ANOVA and Duncan's multiple range test on the results (*: p < 0.05, **: p < 0.01, ***: p < 0.001). Origin 2018 (OriginLab Corporation, Northampton, MA, USA) was used to draw the figures.

3. Results and discussion

3.1. Chemical structure analysis

Fig. 1A shows the Fourier transform infrared (FT-IR) spectra of ZIF-8, HZIF-8, LEO@HZIF-8, TA, and LEO. Specifically, the FT-IR spectrum of ZIF-8 exhibits characteristic absorption peaks at 1175 cm⁻¹, 753 cm⁻¹, and 435 cm⁻¹, corresponding to the C—N stretching vibration of the imidazole ring, Zn—O stretching vibration, and Zn—N stretching vibration within the zeolitic imidazolate framework, respectively, confirming the successful formation of the ZIF-8 framework. In addition, absorption bands at 3137 cm⁻¹, 1584 cm⁻¹, 1464–1387 cm⁻¹, 1310–992 cm⁻¹, and 753–684 cm⁻¹ are attributed to C—H stretching, C N stretching, overall stretching of the imidazole ring, in-plane bending of the imidazole ring, and aromatic C—H bending vibrations in the ZIF-8 structure, respectively (Nguyen et al., 2026). Compared with ZIF-8, HZIF-8 exhibits a significantly enhanced broad O—H stretching band at 3200–3500 cm⁻¹, mainly ascribable to the abundant phenolic hydroxyl groups in TA. HZIF-8 displays two new bands at 1701 cm⁻¹ and 1349 cm⁻¹, attributable to the C O and C—O—C vibrations of TA, respectively, indicating that TA coordinates with Zn²⁺ centers and exerts an electron-withdrawing effect (Xu et al., 2025). Furthermore, after modification of ZIF-8 with TA, the characteristic C—H absorption peak of the imidazole ligand disappeared, and the characteristic peaks at 753 cm⁻¹ (Zn—O vibration) and 435 cm⁻¹ (Zn—N vibration) were significantly weakened, while no new absorption peaks appeared. These results further confirm that TA is mainly adsorbed on the surface of ZIF-8 and may form internal cavities by decomposing its framework structure (Lu et al., 2023). In the FT-IR spectrum of LEO, characteristic bands are observed at 2829 cm⁻¹ and 1466 cm⁻¹, assigned to aliphatic C—H stretching and bending (scissoring), respectively. In addition, bands at approximately 3328 cm⁻¹ and 1722 cm⁻¹ correspond to O—H stretching of alcohols (e.g., linalool) and ester C O stretching (e.g., linalyl acetate), respectively. In the FT-IR spectrum of LEO@HZIF-8, slight peak shifts can be observed (the O—H and C O absorption bands shift to lower wavenumbers), indicating that LEO interacts with HZIF-8 via hydrogen bonds. Meanwhile, the characteristic absorption peaks of HZIF-8 are well preserved, while the absorption bands at 2829 cm⁻¹ and 1466 cm⁻¹ disappear, suggesting the successful loading of LEO into the HZIF-8 carrier. Moreover, the spectrum of LEO@HZIF-8 closely matches that of HZIF-8, suggesting that the framework remains structurally stable during the loading process. However, the intensities of LEO-characteristic bands decreased substantially after encapsulation. This attenuation is consistent with host-guest interactions, such as hydrogen bonding with TA phenolic groups and hydrophobic confinement within the HZIF-8 framework, which can contribute to sustained-release behavior (Fang et al., 2025).

Fig. 1.

Physicochemical characterization of ZIF-8, HZIF-8, LEO@HZIF-8, TA, and LEO. (A) FT-IR spectra. (B) XRD patterns. (C) TGA curves. (D) DTG curves. (E) N₂ adsorption-desorption isotherms. (F) Pore size distribution curves.

Fig. 1B shows the X-ray diffraction (XRD) patterns of ZIF-8, HZIF-8, and LEO@HZIF-8. The ZIF-8 pattern exhibits characteristic reflections at 2θ = 7.3° (0 1 1), 10.3° (0 0 2), 12.6° (1 1 2), 14.6° (0 2 2), 16.5° (0 1 3), 17.9° (2 2 2), 14.5° (2 3 3) and 26.5° (1 3 4), consistent with the SOD topology reported in the literature (Zhang et al., 2024). After TA etching modification, the as-obtained hollow ZIF-8 displays decreased intensities of characteristic peaks in the low-angle region, which may arise from structural distortion or reduced crystallinity during the surface functionalization with TA, thus disturbing the lattice regularity. Meanwhile, HZIF-8 exhibits broadened diffraction peaks at 2θ values of 30° and 42.5°, corresponding to the amorphous TA structure, which may originate from the shielding effect of the surface TA coating on the internal crystal structure (Gao et al., 2023). In the XRD patterns, LEO@HZIF-8 displays similar crystalline diffraction peaks, indicating that LEO does not affect the crystal structure of ZIF-8. These results verify that LEO was successfully encapsulated within the internal cavities of HZIF-8, while the LEO@HZIF-8 composite largely preserved the structural integrity of the parent ZIF-8 framework. Consistently, SEM and TEM analyses (Fig. 2A, B) revealed nanoparticles with a narrow size distribution and well-defined morphologies, in agreement with the XRD results.

Fig. 2.

(A) Scanning electron microscopy (SEM) images of ZIF-8, HZIF-8, and LEO@HZIF-8 (a–f). (B) Transmission electron microscopy (TEM) images of ZIF-8, HZIF-8, and LEO@HZIF-8 (a–c). (C) Particle size distribution of ZIF-8, HZIF-8, and LEO@HZIF-8. (D) Energy-dispersive X-ray spectroscopy (EDS) spectrum of ZIF-8.

Fig. 1C and D show the thermogravimetric (TGA) and derivative thermogravimetric (DTG) curves of ZIF-8, LEO@HZIF-8, and LEO. In the DTG trace of ZIF-8, two mass-loss peaks appear at ∼234 °C and 640.5 °C; the former is attributed to carbonization/decomposition of the 2-methylimidazolate linker, and the latter to the formation of ZnO residue after complete framework degradation (Guo et al., 2024). As a volatile essential oil, LEO underwent nearly complete mass loss below 120 °C, indicating poor thermal stability and high susceptibility to evaporation and/or thermal degradation. In contrast, the TGA profile of LEO@HZIF-8 displayed a three-stage mass-loss pattern. The first stage (30–120 °C) showed a 6.3% mass loss, consistent with the volatilization range of free LEO. However, the mass-loss rate was lower than that of free LEO, primarily due to evaporation of residual surface solvent and water confined within the pores. The second stage (220–415 °C) showed a substantial mass loss (34.2 wt%), attributable to thermal degradation of TA and to thermal decomposition of LEO encapsulated within the HZIF-8 cavities. This result further supports the successful loading of LEO into the interior of HZIF-8. The third stage (560–800 °C) is attributed to collapse and final decomposition of the HZIF-8 framework, leaving a zinc-containing residue. Overall, LEO@HZIF-8 effectively encapsulates LEO, and the HZIF-8 shell acts as a thermal barrier, thereby enhancing the thermal stability of LEO and slowing its volatilization.

Fig. 1E shows the N₂ adsorption-desorption isotherms of ZIF-8 and LEO@HZIF-8 measured at 77 K, and Fig. 1F presents the corresponding pore-size distribution curves. Both materials exhibit a steep uptake at low relative pressure, indicative of microporosity. At higher relative pressures, the isotherms display H4-type hysteresis loops, consistent with capillary condensation within mesopores. Overall, this aligns with the characteristic features of a typical Type IV adsorption-desorption isotherm (Xia et al., 2022), indicating that the materials predominantly feature a mesoporous structure, potentially supplemented by micropores. The pore size distribution of ZIF-8 and LEO@HZIF-8 was determined from the BJH-Desorption model. The pore diameters of both materials predominantly fall within the range of 3.52 to 6.96 Å, while macropores make a negligible contribution. The pore-size distribution was bimodal; the narrow principal peak indicates precise control of pore formation during synthesis. Specific-surface-area analysis showed a marked decrease from 1450.50 m²/g for ZIF-8 to 651.57 m²/g for LEO@HZIF-8. Consistently, the N₂ uptake of LEO@HZIF-8 at high relative pressure was substantially lower than that of ZIF-8. These results support that encapsulated LEO partially occupies or blocks the pores, reducing accessible surface area and limiting N₂ adsorption.

Microstructure was examined by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The ZIF-8 nanoparticles exhibited a regular hexagonal morphology with sharp edges and smooth surfaces. The particle-size distribution was uniform, with an average size of 287 ± 49.3 nm (Fig. 2Aa, d). After TA etching, the morphology changed markedly. As shown in Fig. 2B, HZIF-8 presented hollow particles with more irregular outlines and reduced dimensions. This transformation is consistent with acid-mediated selective dissolution: protons released from TA preferentially attack high-energy vertices and edges of ZIF-8 and then etch inward, generating internal cavities. Concurrently, TA macromolecules adsorb onto the ZIF-8 surface and coordinate with dissociated Zn²⁺ to form metal-phenolic networks that stabilize the outer framework and limit further dissolution (Luo et al., 2024). Following LEO loading, LEO@HZIF-8 retained its hollow-particle morphology without obvious structural change, while the average particle size increased to 364.1 ± 61.6 nm. This indicates successful encapsulation of LEO within the HZIF-8 cavities and supports the feasibility of the TA-assisted hollowing and loading process. In addition, elemental mapping (Fig. 2D) of ZIF-8, a coordination framework constructed from Zn²⁺ nodes and imidazolate linkers, showed uniform distributions of C, N, and Zn. The O signal is primarily attributed to trace water retained within the ZIF-8 pores.

3.2. pH-responsive release performance

To evaluate the loading performance of HZIF-8 for LEO, the encapsulation efficiency and loading efficiency of LEO@HZIF-8 were determined at different initial LEO dosages. As shown in Fig. 3A, the absorbance of LEO at 213 nm increased with concentration. The encapsulation efficiency of LEO@HZIF-8 across the tested conditions ranged from 54.51% to 68.96%. At an HZIF-8-to-LEO mass ratio of 1:1, the composite achieved the highest EE and a maximum LE of 344.91 ± 5.93 mg/g. Therefore, this mass ratio was selected for the preparation of LEO@HZIF-8 in subsequent experiments.

Fig. 3.

(A) Encapsulation efficiency and loading efficiency of LEO@HZIF-8 at different LEO addition amounts. (B) Release rate of LEO@HZIF-8 in PBS buffer solutions with pH values of 4.5, 7.4, and 9.0. (C) Zero-order kinetic release model. (D) First-order kinetic release model. (E) Higuchi kinetic release model. (F) Ritger-Peppas kinetic release model.

To elucidate the release profile of LEO from LEO@HZIF-8, cumulative release was measured in PBS at pH 4.5, 7.4, and 9.0 (Fig. 3B). As pH increased from 4.5 to 9.0, cumulative release decreased and was lowest at pH 9.0. The cumulative release rate of LEO under acidic conditions was higher than that under neutral and alkaline conditions at all time points. After 72 h, the cumulative release reached 73.36%, 35.44%, and 7.42% at pH 4.5, 7.4, and 9.0, respectively. These results clearly demonstrate that the release behavior of LEO exhibits a significant pH-dependent profile, and the encapsulation of LEO using hollow ZIF-8 as a carrier enables sustained and pH-responsive release of bioactive ingredients. Such pH-responsive release is mainly attributed to the instability and easy decomposition of the ZIF-8 structure under acidic conditions. During fruit storage, organic acids generated from the decomposition of glycogen and other components, as well as acidic metabolites produced by microbial anaerobic respiration, contribute to fruit spoilage. Therefore, the pH-responsive release property of the LEO@HZIF-8 nanomaterial in this study is beneficial for achieving long-term antibacterial activity in food systems. To clarify the release mechanism, cumulative-release profiles were fitted to the zero-order, first-order, Higuchi, and Ritger-Peppas models (Table 1). The results indicated that the Ritger-Peppas model yielded the highest goodness of fit under all pH conditions (Fig. 3F), with the release exponent n being less than 0.45 at all three pH values. According to the model exponent (n), release is characterized as Fickian diffusion when n < 0.45, anomalous (non-Fickian) transport when 0.45 < n < 0.89, and framework dissolution–dominated behavior when n > 0.89 (Jiang et al., 2024). In the present study, the fitted n values indicate that LEO release is primarily governed by Fickian diffusion. Accordingly, release is diffusion-controlled and determined by the effective diffusivity of LEO and the pore architecture of the HZIF-8 matrix; under acidic conditions, partial framework dissociation may additionally contribute to the observed kinetics.

Table 1.

Kinetic model fitting results for sustained-release curves under different pH conditions.

Kinetic model	Conditions	Fitting equation	R2
	pH = 4.5	y = 0.5346x + 45.1083	0.3304
Zero-order	pH = 7.4	y = 0.3852x + 13.9992	0.6480
	pH = 9.0	y = 0.0816x + 2.6042	0.7408
	pH = 4.5	y = 68.2867(1-e-0.4054x)	0.9810
First-order	pH = 7.4	y = 33.0722(1-e-0.1255x)	0.9811
	pH = 9.0	y = 6.5976(1-e-0.1121x)	0.9154
	pH = 4.5	y = 6.4132x0.5 + 31.2470	0.5854
Higuchi	pH = 7.4	y = 4.0387x0.5 + 6.4025	0.8770
	pH = 9.0	y = 0.822x0.5 + 1.1324	0.9257
	pH = 4.5	y = 49.2641x0.0958	0.9913
Ritger-Peppas	pH = 7.4	y = 11.3216x0.2805	0.9626
	pH = 9.0	y = 2.0595x0.3039	0.9807

3.3. UV protection performance

Given its susceptibility to UV-induced degradation, developing UV-shielding delivery systems is essential to preserve the stability of LEO. Accordingly, the UV-protective capacity of LEO@HZIF-8 was evaluated (Fig. 4A). Free LEO and LEO@HZIF-8 were irradiated with 254 nm ultraviolet light at a distance of 15 cm for 0, 3, 6, 9, 12, 24, 36, 48, and 72 h to monitor photodegradation. After 3 h of UV irradiation, free LEO showed 15.34% degradation. By contrast, LEO@HZIF-8 exhibited only 9.86% degradation even after 72 h, indicating markedly improved photostability. These results demonstrate that the HZIF-8 shell provides effective UV shielding, reducing LEO photodegradation and limiting the loss of active components during delivery. In summary, ZIF-8 served as an effective nanocarrier, enabling efficient loading and pH-responsive release of LEO and markedly improving its photostability, thereby supporting its application in active food packaging.

Fig. 4.

Photochemical stability of LEO and LEO@HZIF-8 under UV irradiation. (A) Retention rates of LEO and LEO@HZIF-8. (B) Pseudo-first-order kinetic modeling of the photodegradation process for LEO and LEO@HZIF-8.

3.4. Antioxidant and antibacterial properties

The antioxidant activity of nanomaterials was evaluated using DPPH and ABTS radical scavenging assays (Fig. 5A). Both TA and Vc exhibited remarkable antioxidant capacity, which can be attributed to the phenolic hydroxyl structures in their molecules. They demonstrated scavenging activities exceeding 85% against both DPPH and ABTS radicals. In contrast, ZIF-8 exhibits radical scavenging activity below 5% and shows no significant antioxidant capacity. HZIF-8 exhibited enhanced antioxidant performance compared with ZIF-8, attributable to surface phenolic hydroxyl groups introduced by TA. However, during TA surface modification, many phenolic hydroxyls chelate Zn²⁺, reducing the number of free phenolic sites available for radical quenching; consequently, the antioxidant activity of HZIF-8 remained lower than that of free TA. LEO's primary constituents are linalool and linalyl acetate, with phenolic compounds present at relatively low levels; accordingly, its antioxidant activity is lower than that of TA and Vc. By contrast, LEO@HZIF-8 exhibited DPPH and ABTS radical-scavenging activities of 85.48% and 78.24%, respectively, indicating that encapsulation within the HZIF-8 nanocarrier did not attenuate the antioxidant activity imparted by TA. This high activity is mainly attributed to the synergistic action of TA and LEO. Therefore, LEO@HZIF-8 exhibits outstanding antioxidant properties, effectively delaying oxidative reactions in perishable fruits. It holds promise for inhibiting spoilage and extending shelf life in fruit and vegetable preservation. (See Fig. 6.)

Fig. 5.

(A) Radical scavenging activities of ZIF-8, HZIF-8, LEO@HZIF-8, LEO, TA, and VC measured by the DPPH and ABTS methods. (B) Antibacterial zone radii of LEO@HZIF-8 at different concentrations. (I) Negative control: PBS, (II) 5 mg/mL LEO@ HZIF-8, (III) 10 mg/mL LEO@HZIF-8, (IV) 15 mg/mL LEO@HZIF-8, and (V) 20 mg/mL LEO@HZIF-8 against (C) E. coli and (D) S. aureus. (I) Negative control: PBS, (II) 20 mg/mL ZIF-8, (III) 20 mg/mL LEO, and (IV) 20 mg/mL TA against (E) E. coli and (F) S. aureus. (G) Inhibition zone radii of 20 mg/mL ZIF-8, LEO, and TA.

Fig. 6.

Multifactorial synergistic antibacterial mode and diagram.

Microbial infection is the primary cause of postharvest fruit rot. To evaluate the antibacterial activity of LEO@HZIF-8, E. coli and S. aureus were used as model strains in disk-diffusion assays (Fig. 5B). As shown in Fig. 5C and D, clear inhibition zones were observed around LEO@HZIF-8 at the tested concentrations. At equivalent concentrations, S. aureus exhibited larger inhibition zones than E. coli, which may be attributed to the distinct cell membrane structures between Gram-positive and Gram-negative bacteria. The more complex outer membrane of E. coli hinders the penetration of reactive oxygen species and antibacterial components. At 20 mg/mL, the inhibition-zone radius was 3.49 mm for E. coli and 5.99 mm for S. aureus. This enhanced antibacterial efficiency can be ascribed to the intelligent pH-responsive controlled-release behavior of LEO@HZIF-8, which improves the utilization efficiency of LEO. To further clarify the contribution of individual components, the antibacterial activities of single components (ZIF-8, LEO, and TA) at 20 mg/mL were also investigated (Fig. 5E and F). All individual samples showed certain inhibitory effects, but their inhibition zones were significantly smaller than those of LEO@HZIF-8 at the same concentration, confirming remarkable synergistic antibacterial effects among ZIF-8, TA, and LEO. The Zn²⁺ release from LEO@HZIF-8 was further quantified by ICP-MS, with cumulative release ratios of 2.79% at 24 h and 9.59% at 72 h, demonstrating sustained and long-term Zn²⁺ release. Under acidic conditions, partial dissolution of the ZIF-8 framework releases Zn²⁺ and can induce reactive oxygen species (ROS), which together damage bacterial lipids, proteins, and nucleic acids. TA exhibits high protein affinity and interacts with charged surface components, leading to enzyme inactivation and increased membrane permeability. Meanwhile, hydrophobic components in LEO disrupt the phospholipid bilayer of bacterial membranes and further increase membrane permeability, causing the leakage of intracellular contents and ultimately bacterial death (Pandey et al., 2024). Therefore, LEO@HZIF-8 exhibits strong antibacterial activity that can effectively inhibit microbial infection on fruit surfaces and thereby extend the shelf life of fruits.

3.5. Preservation capability

To assess the preservation performance of the LEO@HZIF-8/SA/Zein composite coating (LHSZ), we monitored fruit appearance and key quality attributes over the storage period. Visual comparison across time points (Fig. 7A) showed that the LHSZ group maintained the best appearance throughout storage. With increasing storage time, decay incidence rose in all treatment groups (Fig. 7B). The trial was terminated when the LHSZ group reached ≥40% decay. At this endpoint, the final decay incidences for PE, SZ, and LHSZ were 82.33%, 64.67%, and 51.00%, respectively. Results showed that the LHSZ coating (LEO@HZIF-8/SA/Zein) was the most effective among treatments at delaying decay and browning in apricots, extending shelf life to 63 days.

Fig. 7.

Photographs of fresh white apricots during storage (A), decay rate (B), weight loss rate (C), pH (D), and hardness (E) for different treatment groups of CK, PE, SZ, and LHSZ.

The uncoated control fruit exhibited 35.44% weight loss over the 63 days storage period (Fig. 7C), consistent with higher respiration and transpiration. By contrast, the SZ and LHSZ coatings reduced weight loss to 18.97% and 16.81%, respectively. The composite film loaded with LEO@HZIF-8 nanoparticles helped retain internal moisture by establishing a stable microenvironment around the fruit, thereby mitigating water loss under fluctuating storage conditions.

Throughout the storage period, the fruit pH showed a consistent increasing trend (Fig. 7D), which is closely related to the metabolic transformation of organic acids during postharvest ripening (Wang et al., 2023). Notably, the LHSZ treatment group exhibited the slowest pH increase, reaching only 4.64 on day 63, indicating that this treatment effectively inhibited the decline in fruit acidity and helped maintain quality stability during storage.

Firmness decreased progressively with storage time (Fig. 7E), primarily because of depolymerization and solubilization of cell-wall polysaccharides (cellulose, hemicellulose, and pectins) and loss of tissue turgor due to water loss (Luo et al., 2024). Among the treatments, fruit firmness was best preserved in the LHSZ group. After 63 days, firmness in LHSZ was 50.72%, 31.45%, and 7.47% higher than in CK, PE, and SZ, respectively. These results indicate that the composite coating effectively delayed softening and helped maintain postharvest texture.

3.6. Metal ion migration test

Considering the potential health risks associated with Zn²⁺ migration from ZIF-8-containing coatings into food, metal ion migration assays were conducted. After 63 days of storage of fresh apricots, the Zn²⁺ concentration in the polyethylene (PE) packaging group was 8.81 μg/L, while the average Zn²⁺ migration values for SZ and LHSZ coating treatments were 8.65 μg/L and 17.77 μg/L, respectively. No significant differences were observed among these values, all of which were far below the EU food safety limit for zinc of 50 μg/L (EU 10/2011). Furthermore, zinc is an essential and beneficial trace element for the human body, so the ion migration during apricot preservation using the composite coating is negligible. Accordingly, the metal ion migration results indicate that ZIF-8 can be used as a food-contact material, and the LHSZ nanocomposite coating is applicable for food preservation.

3.7. Biosafety test

Owing to the particular characteristics of MOFs in food applications, their safety cannot be accurately evaluated solely by metal ion migration. We further investigated the biosafety of LHSZ using a 14-day zebrafish survival test. As shown in the Fig. 8A, the survival rate of zebrafish remained 100% in all LHSZ aqueous solutions at different concentrations within 14 days, indicating that LHSZ possesses favorable biosafety and poses no harm to humans as an active food packaging material.

Fig. 8.

(A) Effects of different concentrations of LHSZ on the survival of zebrafish.

3.8. Comparative study

Overall, the LEO@HZIF-8/SA/Zein system developed in this study exhibits favorable encapsulation efficiency and stable pH-responsive release behavior, with performance comparable or superior to most reported studies (Table 2). In addition, compared with active packaging without MOF carriers, the present system prolongs the shelf life of white apricots by more than 50%, which is significantly higher than values reported in the literature (Ali et al., 2021; Firdaus et al., 2024; Gull et al., 2021). This can be attributed to the TA-modified ZIF-8 carrier that enables on-demand release of essential oils in the weakly acidic microenvironment of decaying fruits. From an industrial perspective, the system shows favorable feasibility: the preparation process is based on simple surface modification and solution casting techniques, and all raw materials are food-contact safe. However, the response condition of this study is relatively single. In the future, intelligent packaging systems with light, temperature, humidity, and enzyme responsiveness can be further developed, and the long-term stability of the composite coating under industrial storage conditions still needs to be verified.

Table 2.

Summary of studies on ZIF-8-based delivery systems for bioactive components in food preservation packaging.

Delivery system	Bioactive component	EE (%)	Stimulus	Preservation system	Application model	Shelf life (d)	Reference
Thy@ZIF-8	Thy	81.47	pH	Thy@ZIF-8/Kc/Zein	Blueberry	9	Li et al., 2025
Thy@ZIF-8	Thy	–	pH	CS-PA/Thy@ZIF-8	Strawberry	7	Chen et al., 2025
Cur-ZIF-8	Cur	52.2	pH	Cur-ZIF-8/CS/Zein	Litchi	8	Geng et al., 2023
CAR@ZIF-8	CAR	–	pH/ enzyme	CAR@ZIF-8/TOCNF/Pec	Raspberry / Mango	5/7	Min et al., 2024
CV@ZIF-8	CV	–	–	FSG/SA/CV@ZIF-8	Strawberry	8	Li et al., 2023
GA@ZIF-8	GA	–	–	CA-GA@ZIF-8	Beef	12	Zhang et al., 2024
QDs@ZIF-8	QDs	–	light	CS/SA/QDs@ZIF-8	Kiwi fruit	15	Wang et al., 2023
TOCNFs@ZIF-8@TA-Cur	Cur	–	pH	TOCNFs@ZIF-8@TA-Cur-SA	Strawberry	7	Xu et al., 2025
CuO@ZIF-8	CuO	–	–	PVA-HACC-CuO@ZIF-8	Tomato	21	Fu et al., 2022
APG@ZIF-8	APG	71.73	pH	APG@ZIF-8-PG	Apple	3	Wang et al., 2023

Note: Thy: Thymol, Kc: κ-carrageenan, CS: Chitosan, PA: Procyanidin, Cur: Curcumin, CAR: Carvacrol, TOCNF: TEMPO-oxidized cellulose nanofibrils, Pec: Pectinase, CV: Carvacrol, FSG: Fish scale gelatin, GA: Gallic acid, CA: Carrageenan, QDs: Quantum dots, PVA: Polyvinyl alcohol, HACC: Hydroxypropyltrimethyl ammonium chloride chitosan, APG: Apigenin.

4. Conclusion

In this study, a pH-responsive delivery system based on LEO@HZIF-8 was developed and incorporated into SA/Zein composite films, yielding a smart active packaging system for preserving white apricots. The system exhibited the following performance. First, material characterization confirmed that TA modification produced a hollow ZIF-8 carrier with a LEO encapsulation efficiency of 68.96%. The carrier showed pH-triggered release with an accelerated and sustained profile under acidic conditions (pH 4.5). Its release kinetics were consistent with Fickian diffusion, enabling precise on-demand delivery in the weakly acidic microenvironment of decaying fruit. Second, LEO@HZIF-8 exhibited strong functional activity, with DPPH and ABTS radical-scavenging activities of 85.48% and 78.24%, respectively, and significant inhibition of E. coli and S. aureus. These effects are attributable to the synergistic antioxidant and antibacterial actions of TA and LEO. Finally, the fruit-preservation trial confirmed the practical effectiveness of the system: compared with the control, the LHSZ coating markedly delayed decay in white apricots and extended shelf life to 63 days. It also effectively limited weight loss, slowed firmness decline, and restrained pH drift, thereby maintaining the postharvest quality of the fruit.

In summary, this study demonstrates the effectiveness of the LEO@HZIF-8/SA/Zein system for preserving perishable fruits and provides a design basis and technical route for the development of MOF-based, stimuli-responsive active food packaging materials.

CRediT authorship contribution statement

Xinyu She: Writing – review & editing. Qian Zhou: Writing – original draft. Jingjing Song: Methodology. Bin Wu: Formal analysis. Na Xia: Supervision, Project administration. Chao Li: Conceptualization. Siming Zhu: Investigation.

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.

Data availability

Data will be made available on request.