Ultrasound-fabricated thyme essential oil nanoemulsions stabilized by clean-label emulsifiers: Physicochemical properties, antibacterial activity, and food decontamination applications

Graphical abstract

Highlights

• •Ultrasonic fabricated stable thyme essential oil nanoemulsions (120–220  nm)
• •Tea saponin and soy lecithin outperform Tween 60 in nanoemulsion stability.
• •TS nanoemulsion has lowest MIC (156.3  µg/mL) against E. coli O157:H7 and S. aureus.
• •Emulsifier interfacial properties regulate nanoemulsion antibacterial mechanisms.
• •TS nanoemulsion effectively decontaminates cherry tomatoes and milk matrices.

Abstract

In this study, thyme essential oil nanoemulsions stabilized by clean-label emulsifiers (tea saponin, TS; soy lecithin, SL; soy protein isolate, SPI), with Tween 60 (TW) as a synthetic benchmark, were fabricated via ultrasonic emulsification (573 W/cm², 5 min). Ultrasonication effectively reduced droplet sizes to 120–220 nm, producing narrow size distributions (PDI < 0.22) and high encapsulation efficiencies, which were associated with cavitation-induced droplet disruption and enhanced emulsifier adsorption at newly generated interfaces. Among the formulations, TS- and SL-stabilized nanoemulsions exhibited improved physicochemical stability under centrifugal, ionic, thermal, pH, and storage stresses, suggesting the formation of more stable interfacial layers. All nanoemulsions showed strong antibacterial activity against Escherichia coli O157:H7 and S. aureus, with NE-TS demonstrating the highest potency (MIC = 156.3 µg/mL; MBC = 156.3–312.5 µg/mL) and achieving > 6 log CFU/mL reductions within 30 min. Mechanistic analyses indicated that NE-TS exerted its effect through severe membrane disruption, membrane depolarization, and intracellular reactive oxygen species accumulation, resulting in extensive cell collapse. In contrast, TW mainly induced membrane permeabilization, SL may promote membrane fusion with bacterial envelopes, and SPI may cause electrostatic aggregation with limited lytic activity. In food applications, TS-stabilized nanoemulsions efficiently decontaminated cherry tomato surfaces and significantly reduced pathogens in milk matrices. These findings demonstrate that emulsifier-mediated interfacial architecture regulates nanoemulsion physicochemical properties and antibacterial performance of ultrasound-fabricated nanoemulsions, providing insights for the development of food-grade essential oil nanoemulsions for potential food preservation applications.

1. Introduction

Foodborne diseases remain a major challenge to global food safety, particularly in minimally processed foods such as fresh produce and dairy products [1], [2]. Among the most important foodborne pathogens, Escherichia coli O157 and S. aureus are frequently associated with food contamination and outbreaks, posing significant risks to public health [3], [4]. Meanwhile, increasing consumer demand for natural preservatives has stimulated the development of effective antimicrobial systems derived from natural sources [5], [6].

Plant essential oils (EOs) have emerged as promising natural antimicrobial agents because of their broad-spectrum antibacterial activity and low tendency to induce microbial resistance [7]. Among them, thyme essential oil (TEO) derived from Thymus vulgaris L. is of particular interest due to its high concentration of bioactive phenolic monoterpenes, particularly thymol and carvacrol, which exhibit potent antibacterial activity against a wide range of foodborne pathogens [8], [9]. Furthermore, its Generally Recognized as Safe (GRAS) status [10] and cost-effective availability make it an ideal candidate for developing natural preservatives in the food industry. However, the practical application of TEO in food systems is often limited by high hydrophobicity, volatility, chemical instability, and poor water dispersibility, which reduce their bioavailability and antimicrobial efficacy [11]. As a result, relatively high dosages are often required to achieve effective microbial inhibition, which may adversely affect food sensory quality [12]. To address these challenges, nanoemulsion technology has emerged as an effective delivery strategy to enhance the dispersibility, stability, and bioactivity of hydrophobic compounds in complex food matrices [13], [14].

Recent literature-mapping studies have highlighted the rapid development of emerging food processing technologies and natural-product-based approaches, including ultrasound-assisted processing and bioactive compound delivery systems [15], [16]. Ultrasound technology has been widely applied in the food industry for various applications, including extraction, emulsification, and encapsulation, wing to its efficiency, low cost, and environmental friendliness [17]. For example, ultrasound-assisted extraction significantly enhanced the recovery of phenolic compounds from bergamot juice [18], whereas direct ultrasound treatment has also been shown to facilitate the incorporation of natural antioxidants from saffron by-products into sunflower oil, thereby improving its oxidative stability [19]. These studies demonstrate the versatility of ultrasound as an emerging green processing technology for developing functional food systems.

Among the available fabrication techniques, ultrasonic emulsification is particularly attractive because acoustic cavitation can efficiently generate nanosized droplets with narrow size distributions [20], [21]. The rapid collapse of cavitation bubbles produces localized high-energy microenvironments characterized by intense shear forces, microjets, and turbulence, which continuously disrupt coarse oil droplets and create fresh oil–water interfaces [22]. Beyond droplet size reduction, ultrasonication also accelerates emulsifier adsorption onto newly generated interfaces, thereby influencing the formation and organization of interfacial layers [23], [24]. Recent studies have demonstrated that the physicochemical stability and biological functionality of essential oil nanoemulsions are governed not only by droplet size but also by the characteristics of the interfacial layer surrounding the droplets [25]. Therefore, understanding how ultrasonication interacts with different emulsifiers to regulate interfacial architecture has become increasingly important for the rational design of functional nanoemulsions.

The interfacial layer plays a central role in determining colloidal stability, environmental tolerance, antimicrobial activity, and practical performance of nanoemulsions [26]. Because different emulsifiers possess distinct molecular structures and adsorption behaviors, they can generate fundamentally different interfacial architectures and stabilization mechanisms during ultrasonic emulsification [27]. Meanwhile, increasing consumer demand for clean-label foods has stimulated growing interest in replacing synthetic surfactants with naturally derived emulsifiers [28]. Tea saponin (TS), soy lecithin (SL), and soy protein isolate (SPI) are promising food-grade alternatives that stabilized oil–water interfaces through distinct mechanisms, whereas Tween 60 (TW) is widely used as a benchmark synthetic surfactant due to its excellent emulsifying efficiency [29], [30], [31], [32]. Therefore, emulsifier selection is expected to influence not only nanoemulsion formation during ultrasonication, but also the resulting stability, antibacterial behavior, and application performance.

Despite substantial progress in essential oil nanoemulsions, the role of emulsifier-mediated interfacial architectures in ultrasound-fabricated systems remains insufficiently understood. In particular, the relationship between emulsifier-dependent physicochemical characteristics (e.g., droplet size, charge, dispersion stability, and encapsulation behavior), antibacterial responses, and performance in complex food matrices has not been systematically established. A better understanding of these relationships is important for rational selection of food-grade emulsifiers and optimization of ultrasound-assisted nanoemulsion systems.

Therefore, the objective of this study was to establish the interfacial structure-stability-activity relationships of ultrasound-fabricated thyme essential oil nanoemulsions stabilized by different emulsifiers. Three food-grade natural emulsifiers, namely tea saponin (TS), soy lecithin (SL), and soy protein isolate (SPI), were systematically investigated and benchmarked against the conventional surfactant Tween 60 (TW). By integrating physicochemical characterization, environmental and storage stability evaluation, antibacterial assessment against E. coli O157:H7 and S. aureus, and validation in representative food matrices, this study aimed to elucidate the relationship between emulsifier-mediated stabilization behavior and antimicrobial functionality. The findings provide insights into the rational design of clean-label essential oil nanoemulsions and their potential application in food decontamination and preservation.

2. Materials and methods

2.1. Materials

Thyme (Thymus vulgaris L.) essential oil (TEO) was purchased from Sigma-Aldrich (St. Louis, MO, USA; product W306509). According to the manufacturer’s product specification, the essential oil is derived from Thymus vulgaris and/or Thymus zygis, and is of FG (Food Grade), Fragrance grade, Halal, and Kosher quality. The TEO was stored in a tightly closed amber glass container at room temperature in a dry place, protected from light, and was used within 12 months of purchase to ensure compositional stability. To verify the chemical composition of the specific batch used in this study, the TEO was subjected to gas chromatography-mass spectrometry (GC–MS) analysis, following the method described by Dong et al., (2023) with minor modifications [33]. The analysis was performed on an Agilent 7890A-5975C system equipped with an HP-5MS column (30.0 m × 250 μm, 0.25 μm). The oven temperature was programmed from 60 °C (held for 2 min) to 280 °C, with helium as the carrier gas at 1.0 mL/min. Compounds were identified by comparison of mass spectra with the NIST library and retention indices against n-alkane standards (C8-C40). The detailed compositional data are presented in Supplementary Table S1. Four emulsifiers were used in this study. Tween 60 (TW), soy lecithin (SL), tea saponin (TS), and soy protein isolate (SPI), all obtained from Aladdin Bio-Chem Technology Co., Ltd. (Shanghai, China). Tween 60 (polysorbate 60, CAS 9005–67-8) is a nonionic surfactant with a purity of ≥ 98%. Soy lecithin (CAS 8002–43-5) is a mixture of phospholipids with a total phospholipid content of ≥ 98%. Tea saponin (CAS 8047–15-2) is a natural triterpenoid saponin extracted from tea seeds (Camellia oleifera), with a sapogenin content of ≥ 60%. Soy protein isolate (CAS 9010–10-0) is a highly refined soy protein with a protein content of ≥ 90%. All other reagents were of analytical grade. Deionized water was used throughout the experiments.

2.2. Preparation of thyme essential oil nanoemulsions (NE)

Four thyme essential oil nanoemulsions (NEs) were prepared using different emulsifiers. Briefly, 2 g of TEO was mixed with 1 g of emulsifier (TW, SL, TS, or SPI). The mixture was then combined with 48 mL of deionized water. Both the oil-emulsifier blend and the aqueous phase were preheated to 50 ℃ separately. The two phases were then mixed and homogenized using a high-speed homogenizer (FJ200-S, Lichen Instrument Technology Co., Ltd., Hunan, China) equipped with a Φ 12 mm dispersion head (stator diameter 12 mm, rotor diameter 8 mm) at 7000 rpm for 4 min to obtain a coarse emulsion (CE). The homogenization was performed at room temperature, and the temperature increase during this step was monitored using a thermocouple, showing a rise of approximately 0.5 ℃ due to the short processing time (4 min).

The coarse emulsion (30 mL) was transferred into a 75-mL cylindrical glass tube (inner diameter 28 mm) and further subjected to ultrasonication using a probe-type ultrasonic cell disruptor (Scientz-IID, Ningbo Scientz, Zhejiang, China) equipped with a Φ 10 mm tapered titanium probe (tip diameter 10 mm) operating at a fixed frequency of 20 kHz. The probe was immersed 20 mm below the liquid surface during sonication. The ultrasonic treatments were performed under three different ultrasonication conditions: (1) 83% amplitude (nominal intensity 951 W/cm²) for 3 min (NE-1), (2) 50% amplitude (nominal intensity 573 W/cm²) for 5 min (NE-2), and (3) 25% amplitude (nominal intensity 287 W/cm²) for 10 min (NE-3). The actual acoustic power delivered to the liquid was determined by calorimetric measurements following the method described in the literature [34], with measured values of 233.7 W, 188.3 W, and 91.7 W for the three amplitude settings, respectively. The three amplitude levels were selected based on preliminary optimization experiments to represent low-, medium-, and high-energy input conditions, allowing systematic evaluation of the effect of acoustic energy on nanoemulsion properties. A duty cycle of 50% (5 s on, 5 s off) was used for all treatments. During sonication, the glass tube was immersed in an ice-water bath to prevent temperature rise, and the temperature was monitored to ensure it remained below 25 ℃. The resulting nanoemulsions had a final TEO concentration of 40 mg/mL. The four nanoemulsions were designated as NE-TW, NE-SL, NE-TS, and NE-SPI based on the emulsifier used. All samples were stored in transparent glass vials at 4 ℃ until further analysis.

2.3. Characterization of nanoemulsions

2.3.1. Droplet size, polydispersity index (PDI), and zeta potential

The average droplet size, PDI, and zeta potential of both coarse and nanoemulsions were determined using dynamic light scattering (DLS) on a Zetasizer Nano ZS90 (Malvern Instruments, Worcestershire, UK). Prior to measurement, samples were diluted 1000-fold with deionized water to avoid multiple scattering effects. All measurements were performed at 25 ℃ in triplicate.

2.3.2. Encapsulation efficiency (EE)

The encapsulation efficiency of TEO in the nanoemulsions was determined using a solvent extraction method coupled with UV–Vis spectrophotometry. A standard calibration curve for TEO was first established by measuring the absorbance of TEO-ethanol solutions at 255 nm, corresponding to the characteristic absorption wavelength of TEO. A mixture of ethanol, hexane, and nanoemulsion (2:3:1, v/v/v) was gently shaken and then centrifuged at 10,000 × g for 15 min at 4 ℃. The supernatant containing unencapsulated TEO was collected, diluted with ethanol, and its absorbance was measured at 255 nm. The encapsulation efficiency (EE) was calculated as:

$$\text{EE (\%)}=\frac{{\text{Total}}_{\text{TEO}}\text{-}{\text{Free}}_{\text{TEO}}}{{\text{Total}}_{\text{TEO}}}\text{× 100}$$

As TEO is a complex mixture of volatile constituents, the EE values reported in this study represent an apparent encapsulation efficiency based on UV absorbance at 255 nm rather than the absolute encapsulation of individual chemical components.

2.3.3. Turbidity measurement

Each nanoemulsion was diluted 40-fold with phosphate-buffered saline (PBS, pH 7.4). The absorbance of the diluted sample was measured at 600 nm using a UV–Vis spectrophotometer, with PBS as a blank [35]. Turbidity (T) was calculated according to the following equation:

$T=\frac{1.302\times A\times V}{I}$

where A is the absorbance at 600 nm, V is the dilution factor (40), and I is the path length (0.01 m). All measurements were conducted at room temperature.

2.3.4. Viscosity measurement

The viscosity of coarse and nanoemulsions was measured using a digital rotational viscometer (Lichen LC-NDJ-5S, Changsha, China) equipped with a No. 1 rotor. The measurements were performed at a rotational speed of 12 rpm. Each sample was measured in triplicate, and the viscosity values were recorded in mPa·s, ensuring that the torque percentage remained within the reliable range of 10–90%.

2.4. Stability studies

2.4.1. Centrifugation stability

The centrifugal stability constant (Ke) was evaluated by measuring the absorbance of diluted nanoemulsions (1000-fold) at 500 nm before and after centrifugation at 4500 × g for 15 min. The Ke value was calculated as:

$$\text{Ke (\%)}=\frac{{\text{A}}_{\text{1}}\text{-}{\text{A}}_2}{{\text{A}}_{\text{1}}}\text{× 100}$$

where A₁ and A₂ are the absorbances before and after centrifugation, respectively.

2.4.2. pH stability

The pH of each nanoemulsion was adjusted to 5, 6, 7, 8, and 9 using 0.1 M HCl or 0.1 M NaOH. Samples were stored at 4 ℃ for 24h, after which droplet size, polydispersity index (PDI), turbidity, and encapsulation efficiency were measured.

2.4.3. Ionic stability

Different volumes of 5 M NaCl stock solution were added to each nanoemulsion to achieve final NaCl concentrations of 100, 200, 300, 400, and 500 mM. The samples were stored at 4 ℃ for 24h before droplet size, polydispersity index (PDI), turbidity, and encapsulation efficiency were measured.

2.4.4. Thermal stability

Nanoemulsions were incubated in a water bath at 60, 70, 80, or 90 ℃ for 30 min. After heating, samples were cooled to room temperature, and droplet size, polydispersity index (PDI), turbidity, and encapsulation efficiency were subsequently measured.

2.4.5. Storage stability

Each nanoemulsion was stored at 4 ℃ (refrigeration) and 25 ℃ (room temperature) for 15 days. Droplet size and PDI were monitored every 3 days (0, 3, 6, 9, 12, and 15 days).

2.5. Antimicrobial activity assays

2.5.1. Bacterial strains and culture conditions

Escherichia coli (E. coli) O157:H7 ATCC 35150 and S. aureus (S. aureus) ATCC 25923 were used in this study. Bacterial cultures were grown in nutrient broth (NB) at 37 ℃ for 17  h with shaking (180 rpm). Cells were harvested by centrifugation (4000 × g, 10  min, 4 ℃), washed twice, and resuspended in sterile saline to achieve a final concentration of approximately 10⁹ CFU/mL. The bacterial suspension was stored at 4 ℃ until use.

2.5.2. Agar well diffusion method

The antibacterial activity of nanoemulsions was evaluated using an agar well diffusion assay following the method reported by Joshi et al. (2025) with slight modification [36]. Briefly, Petri dishes were poured with 25  mL of plate count agar (PCA) to obtain a uniform thickness of approximately 4  mm. After solidification, 100  µL of bacterial suspension (10⁶ CFU/mL) was spread evenly over the agar surface. Wells (6  mm diameter) were then prepared in the agar, and 100 µL of each nanoemulsion sample was added to the respective wells. Sterile saline was used as a negative control. Three replicate wells were prepared for each sample, and the experiment was repeated three times. Plates were incubated upright at 37 ℃ for 17  h, after which the diameters of the inhibition zones were measured using Image J software (cross–cross method).

2.5.3. Minimum inhibitory concentration (MIC) determination

The MIC of each nanoemulsion against E. coli and S. aureus was determined using the standard broth microdilution method in 96–well microplates. Nanoemulsions were sterilized by filtration through 0.22 μm membranes and serially diluted in sterile NB. Bacterial suspensions were added to achieve a final inoculum of approximately 10⁶ CFU/mL. Wells containing NB with bacterial suspension and NB with sterile saline served as positive and negative controls, respectively. After incubation at 37 ℃ for 24 h, the MIC was defined as the lowest nanoemulsion concentration that completely inhibited visible bacterial growth.

2.5.4. Minimum bactericidal concentration (MBC) determination

To determine the MBC, 100  µL aliquots were taken from wells corresponding to the MIC and four higher concentration levels. Each aliquot was spread onto PCA plates and incubated at 37 ℃ for 24  h. The MBC was defined as the lowest concentration of nanoemulsion that resulted in no visible colonies on the agar plates. Each experiment was performed with two replicates per concentration and repeated three times. All MIC and MBC values reported in this study are expressed as TEO-equivalent concentrations (µg/mL), based on the initial TEO concentration of 40 mg/mL in the original nanoemulsion formulation.

2.5.5. Time–kill assay

Bacterial suspensions (10⁷ CFU/mL) were mixed with nanoemulsions at a final concentration of 2 × MIC and incubated at 37 ℃. At selected time intervals (0, 10, 30, 60, 120, and 240  min), 1 mL aliquots were subjected to tenfold serial dilution. From each dilution, 100  µL was plated onto PCA plates and spread evenly. Based on preliminary results, two to three appropriate dilutions were selected, and two parallel plates were prepared for each dilution. The experiment was performed in duplicate. After incubation at 37 ℃ for 24  h, colonies were counted. Viable bacterial counts were expressed as log CFU/mL and plotted as a function of incubation time. The limit of detection (LOD) was 0 log CFU/mL.

2.6. Antibacterial mechanism analysis

2.6.1. Scanning electron microscopy (SEM)

The morphological changes of E. coli O157:H7 and S. aureus after nanoemulsion treatment were observed by SEM. Bacterial suspensions were treated with each nanoemulsion at 2 × MIC for 10  min, with untreated cells as control. Cells were harvested by centrifugation (8000 × g, 10  min, 4 ℃), washed twice with phosphate–buffered saline, and fixed in 2.5% glutaraldehyde overnight at 4 ℃, followed by 1% osmium tetroxide for 1  h. After dehydration through a graded ethanol series (30–100%), samples were critical–point dried, sputter–coated with gold, and visualized using a field–emission SEM (GeminiSEM 300, Carl Zeiss, Germany) at 3.0  kV.

2.6.2. Measurement of nucleic acid and protein leakage

The leakage of intracellular nucleic acids and protein was determined by measuring the absorbance of the supernatant at 260 nm and 280 nm, respectively. Bacterial suspensions of E. coli O157:H7 and S. aureus were treated with each nanoemulsion (TW, SL, TS, SPI) at 2 × MIC for 10 min. Untreated bacterial suspensions served as the control. After centrifugation at 8000 × g for 10 min at 4 ℃, the supernatants were collected, and the absorbances were measured using a UV–Vis spectrophotometer (UV-2600, Shimadzu, Kyoto, Japan). The increase in OD₂₆₀ and OD₂₈₀ was expressed as an indicator of nucleic acid and protein leakage, respectively. All experiments were performed in triplicate.

2.6.3. Outer membrane permeability assay

The outer membrane permeability of E. coli O157:H7 was assessed using the hydrophobic fluorescent probe N–phenyl–1–naphthylamine (NPN) [37]. Bacterial suspensions were treated with nanoemulsions at 2 × MIC for 10 min. After treatment, the cells were harvested by centrifugation, washed twice with sterile saline, and resuspended in the same volume of saline. NPN was added to a final concentration of 10 µM, and the mixtures were incubated in the dark for 15 min at room temperature. Fluorescence intensity was measured using a microplate reader (Tecan Spark 10 M, Mannedorf, Switzerland) with excitation at 350 nm and emission at 420 nm.

2.6.4. Inner membrane permeability assay

The inner membrane permeability of E. coli O157:H7 and S. aureus was evaluated using o–nitrophenyl–β–D–galactopyranoside (ONPG) as a substrate. Bacterial suspensions were treated with nanoemulsions at 2 × MIC for 10 min, followed by centrifugation and washing twice with sterile saline. The pellets were resuspended in sterile saline, and ONPG was added to a final concentration of 1.5 mM. The mixtures were incubated at 37 ℃ for 30 min. The absorbance was measured at 420 nm using a UV–Vis spectrophotometer.

2.6.5. Membrane potential assay

Changes in bacterial membrane potential were monitored using the fluorescent probe Rhodamine 123. Bacterial suspensions were treated with nanoemulsions and processed as described in Section 2.6.4. After treatment and washing, the cell pellets were resuspended in sterile saline to an OD₆₀₀ of approximately 0.5. Rhodamine 123 was added to a final concentration of 2 µg/mL, and the mixtures were incubated in the dark at 37 ℃ for 30 min. The cells were then washed twice with saline to remove excess probe, resuspended in the same volume of saline, and transferred to a black 96–well plate. Fluorescence intensity was measured using a microplate reader (Tecan Spark 10 M, Mannedorf, Switzerland) with excitation at 480 nm and emission at 530 nm.

2.6.6. Intracellular reactive oxygen species (ROS) assay

Intracellular ROS levels were determined using the fluorescent probe 2′,7′–dichlorodihydrofluorescein diacetate (DCFH–DA). Bacterial suspensions were treated with nanoemulsions and processed as described in Section 2.6.4. DCFH–DA was added to a final concentration of 10 µM, and the mixtures were incubated in the dark at 37 ℃ for 30 min with gentle shaking every 5 min. The cells were then washed three times with saline, resuspended in the same volume of saline, and transferred to a black 96–well plate. Fluorescence intensity was measured using a microplate reader (Tecan Spark 10 M) with excitation at 488 nm and emission at 525 nm.

2.7. Decontamination of artificially contaminated cherry tomatoes

The decontamination efficacy of the nanoemulsions on cherry tomatoes was evaluated according to our previous study with slight modifications [38]. Cherry tomatoes of uniform size, color, firmness, and free from mechanical damage were selected. They were rinsed with tap water, wiped with 75% ethanol–soaked sterile cotton wool, and further washed with sterile deionized water. The cleaned tomatoes were placed on stainless steel racks in a laminar flow cabinet and air–dried for 1  h. A 1  cm × 1 cm area on each fruit surface was marked using a waterproof marker. A mixed bacterial suspension containing equal volumes of E. coli and S. aureus at 10⁶ CFU/mL was prepared. Fifty microliters of the mixed suspension was spotted onto the marked area. The inoculated tomatoes were left to dry in the cabinet for approximately 1  h to allow bacterial attachment. Based on the MIC results, nanoemulsion treatments were applied at a 2 × MIC. Briefly, inoculated cherry tomatoes were immersed in the treatment solutions and treated under gentle agitation (500 rpm) for 3 min. Untreated inoculated tomatoes served as the control. After treatment, the tomatoes were removed and air-dried under aseptic conditions.

To determine the bacterial population remaining on tomato surfaces, the inoculated peel sections were aseptically excised, homogenized in sterile saline, serially diluted, and plated on PCA. After incubation at 37 ℃ for 24 h, viable bacterial counts were enumerated and expressed as log CFU/g. To evaluate the potential for cross-contamination during washing, aliquots of the processing water were collected immediately after treatment, serially diluted, and plated using the same procedure. The microbial load in the processing water was expressed as log CFU/mL.

2.8. Antimicrobial activity in UHT milk

The antibacterial efficacy of the nanoemulsions in complex food matrices was evaluated using a milk model with different fat contents. UHT milk with different fat contents (whole milk, 4.0% fat; low–fat milk, 2.0% fat; skim milk, 0.1% fat) was purchased from a local supermarket and used to minimize interference from background microflora. Overnight cultures of E. coli and S. aureus were centrifuged and resuspended in sterile saline to obtain a final concentration of 10⁹ CFU/mL. The two bacterial suspensions were then diluted to 10⁷ CFU/mL and mixed in a 1:1 ratio to form a bacterial cocktail, and then inoculated into fresh milk to a final concentration of approximately 10⁶ CFU/mL. Nanoemulsions stabilized by TW, TS, SL, or SPI were then added to achieve a final concentration corresponding to 32 × MIC (TEO-equivalent concentration) against the respective pathogen. Taking the NE-TS as an example, this corresponded to a final TEO concentration of 5.0 mg/mL (0.5%, w/v) and an emulsifier concentration of 2.5 mg/mL (0.25%, w/v). Samples were incubated at 25 ℃ for 4h, while inoculated milk without nanoemulsion treatment served as the control.

After incubation, surviving bacterial populations were enumerated by serial dilution and plating on selective media. E. coli O157:H7 was quantified on MacConkey agar, whereas S. aureus was quantified on Baird-Parker agar. Plates were incubated at 37 ℃ for 24 h before colony counting, and results were expressed as log CFU/mL. For the viable count assays, the LOD was 0 log CFU/g or CFU/mL (i.e., no colonies detected on agar plates after incubation). All experiments were conducted in triplicate.

2.9. Statistical analysis

All experiments were independently performed in triplicate using three independent biological replicates (n = 3), and the data were presented as mean ± standard deviation (SD). Statistical analyses were performed using SPSS software (version 25.0, IBM Corp., Armonk, NY, USA). Each experimental dataset was analyzed independently according to its corresponding experimental design. One–way analysis of variance (ANOVA) followed by Duncan’s multiple range test was used to compare means within each dataset. A significance level of P < 0.05 was considered statistically significant.

3. Results and discussion

3.1. Ultrasonic fabrication and physicochemical properties of nanoemulsions

Prior to ultrasonic emulsification, the chemical composition of the thyme essential oil (TEO) used in this study was characterized by GC–MS. As summarized in Supplementary Table S1, thymol (45.57%), carvacrol (15.14%), and p-cymene (13.91%) were identified as the predominant constituents, which is consistent with the characteristic composition of Thymus vulgaris L. essential oil reported in previous studies [39]. For instance, reports have found broad variations in thymol (12.4–79.74%) and carvacrol (4.37–42.14%) among thyme essential oils [40]. Our results fall well within these reported ranges, confirming the phenolic monoterpene-rich nature of the TEO used, which is known to be primarily responsible for its antimicrobial activity.

It should be noted that ultrasonication was applied to the emulsion system rather than the neat oil. While direct ultrasonication may cause compositional changes in essential oils, such effects could be minimized in emulsifier-stabilized nanoemulsions where the oil is confined within droplets. The present study focused on emulsifier effects on nanoemulsion properties; the potential compositional evolution of TEO during ultrasonication warrants further investigation.

3.1.1. Droplet size, PDI, zeta potential and encapsulation efficiency

The effects of ultrasonication conditions and emulsifier type on the droplet size, PDI, zeta potential, and encapsulation efficiency (EE) of thyme essential oil emulsions are shown in Fig. 1A-D. Compared with the corresponding coarse emulsions (CE), ultrasonication significantly reduced droplet size and PDI while increasing EE values (P < 0.05), demonstrating the effectiveness of ultrasonic emulsification in producing finer and more homogeneous nanoemulsions.

Fig. 1.

Physicochemical properties characteristics of coarse emulsions (CE) and nanoemulsions (NE-1, NE-2, and NE-3) stabilized by different emulsifiers (TW, SL, TS, and SPI). (A) Particle size. (B) Polydispersity index (PDI). (C) Zeta potential. (D) Encapsulation efficiency. (E) Turbidity. (F) Viscosity. Data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters indicate significant differences (P < 0.05).

The reduction in droplet size can be attributed to acoustic cavitation generated during ultrasonication. The formation and collapse of cavitation bubbles produce intense local shear forces, microjets, and turbulence, which disrupt large oil droplets and promote the formation of nanoscale droplets [21], [41]. Similar effects have been widely reported in ultrasonic preparation of essential-oil nanoemulsions [42]. Among the three ultrasonic treatments, NE-2 (50% amplitude, 573 W/cm² nominal intensity, 5 min) generally produced smaller droplet sizes and lower PDI values than NE-1 and NE-3. A comparative analysis of energy density (actual acoustic energy per unit volume), calculated from calorimetrically measured power and processing time, revealed that NE-1, NE-2, and NE-3 delivered 1402, 1883, and 1834 J/mL, respectively. Although NE-1 applied a higher acoustic intensity, its total energy density was lower (1402 J/mL) due to the shorter processing time, and excessive instantaneous cavitation energy may increase droplet collisions and recoalescence before sufficient interfacial stabilization occurs. This is consistent with previous report showing that higher energy inputs can increase polydispersity due to droplet recoalescence in the absence of sufficient emulsifier [43]. Similarly, a study on ultrasonic emulsification in a micro-packed bed reactor observed a “plateau region” in emulsion size due to the competition between droplet disruption and coalescence when increasing energy input [44]. In contrast, NE-3 delivered a comparable energy density (1834 J/mL) but at a substantially lower instantaneous intensity, which generated weaker cavitation effects and less efficient droplet disruption. These results suggest that both the instantaneous acoustic intensity and the total energy density determine the final droplet size, and that an optimal ultrasonic energy density window exists, in which droplet breakup predominates over recoalescence, leading to improved nanoemulsion quality.

The emulsifier type also strongly influenced nanoemulsion characteristics. Under the fixed formulation conditions used in this study (thyme essential oil/emulsifier mass ratio of 2:1), different emulsifiers resulted in distinct nanoemulsion properties, reflecting their different interfacial stabilization capacities. TW, as a low-molecular-weight nonionic surfactant, may facilitate droplet disruption by rapidly reducing oil–water interfacial tension during emulsification. TS-stabilized emulsions exhibited the smallest droplet sizes and lowest PDI values, particularly under NE-2 conditions, suggesting that tea saponin formed a dense and elastic interfacial film that efficiently prevented droplet recoalescence. SL showed similar but slightly weaker stabilization effects, which may be related to the formation of phospholipid-based interfacial layers during emulsification. In contrast, SPI-stabilized emulsions consistently exhibited larger droplet sizes and higher PDI values, even after ultrasonication, possibly due to the relatively slow adsorption kinetics and partial aggregation of protein molecules under intense ultrasonic treatment.

The zeta potential results further supported the differences in colloidal stability among formulations. Nanoemulsions stabilized by TS and SL generally exhibited higher absolute zeta potential values than those stabilized by TW, indicating stronger electrostatic repulsion between droplets and enhanced resistance to aggregation. Correspondingly, the increase in EE after ultrasonication were closely associated with the reduction in droplet size and the formation of more stable interfacial layers. Smaller droplets provide a larger interfacial area for emulsifier adsorption, which may reduce the loss of essential oil during emulsification and sample preparation. The relatively high apparent EE values observed for the TS- and SL-stabilized nanoemulsions may therefore be associated with their smaller droplet sizes, narrower size distributions, and superior physicochemical stability. Further investigations integrating advanced interfacial characterization techniques (e.g., interfacial rheology, adsorption kinetics, and spectroscopic analysis) together with GC–MS analysis of individual thyme essential oil constituents would provide deeper insights into emulsifier–oil interactions, interfacial organization, and the preservation of bioactive volatile compounds during nanoemulsion formation.

3.1.2. Turbidity and viscosity of nanoemulsions

The turbidity and apparent viscosity of emulsions prepared with different emulsifiers and ultrasonication conditions are presented in Fig. 1E-F. Ultrasonication markedly reduced turbidity compared with the corresponding coarse emulsions. This reduction can be attributed to the formation of smaller droplets through acoustic cavitation, which decreases light scattering intensity and improves optical uniformity [45]. Among all formulations, NE-TS-2 exhibited the lowest turbidity (51.4 NTU), consistent with its smallest droplet size and lowest PDI. These results suggest that the TS-stabilized nanoemulsion possessed a more homogeneous droplet dispersion under the investigated conditions. In contrast, CE-SL exhibited the highest turbidity (4638.6 NTU), indicating the presence of large droplets and extensive light scattering. Following ultrasonication under NE-2 conditions, the turbidity of the SL-stabilized emulsion decreased dramatically to 291.6 NTU, which was accompanied by a significant reduction in droplet size and PDI, indicating improved dispersion uniformity. It should be noted that turbidity was used here as a complementary optical parameter rather than direct evidence of structural changes.

A similar trend was observed for viscosity. All nanoemulsions exhibited significantly lower apparent viscosity than their corresponding coarse emulsions (P < 0.05). Furthermore, the apparent viscosity of all nanoemulsions decreased with increasing rotational speed (data no shown), indicating shear-dependent flow behavior under the experimental conditions. The viscosity reduction may be associated with the disruption of droplet aggregates and weak three-dimensional networks present in coarse emulsions. During ultrasonication, acoustic cavitation generates intense shear forces and microstreaming effects that break down flocculated structures and promote the formation of uniformly dispersed droplets, thereby reducing flow resistance and enhancing fluidity [46], [47].

The magnitude of viscosity reduction depended on the emulsifier type. TW-stabilized emulsions exhibited a substantial decrease in viscosity after ultrasonication, reflecting the effective disruption of loosely associated droplet clusters. More importantly, the natural emulsifiers TS and SL also showed pronounced viscosity reductions (approximately 86% and 82%, respectively), indicating that ultrasound-assisted emulsification efficiently eliminated droplet flocculation and generated highly dispersed nanoemulsion systems. The excellent performance of TS may be attributed to its amphiphilic saponin structure, which enables rapid adsorption at newly formed interfaces and efficient stabilization of small droplets. Similarly, SL can self-assemble into interfacial phospholipid layers that enhance droplet dispersion and reduce intermolecular interactions within the continuous phase. Although SPI-stabilized emulsions retained the highest viscosity among the nanoemulsions, their viscosity still decreased significantly from 46.7 to 10.2 mPa·s. This behavior is likely related to protein-induced bridging flocculation in the coarse emulsion, which was partially eliminated following ultrasonic treatment [48]. These results collectively demonstrate that ultrasound-assisted emulsification effectively reduces droplet aggregation and structural entanglement, thereby improving the flow properties and physical uniformity of both synthetic- and natural-emulsifier-based systems.

Overall, the physicochemical characteristics of the nanoemulsions were governed by the combined effects of ultrasonic cavitation and emulsifier properties. While the synthetic surfactant TW served as an effective reference emulsifier, the natural emulsifiers TS and SL exhibited comparable or even superior performance in terms of droplet size distribution, colloidal stability, encapsulation efficiency, and optical properties. In contrast, SPI showed relatively limited emulsification efficiency under the investigated conditions. Considering the combined results of droplet size, PDI, zeta potential, encapsulation efficiency, turbidity, and viscosity, the NE-2 condition (50% amplitude, 573 W/cm², 5 min) consistently produced nanoemulsions with the most desirable physicochemical characteristics. Therefore, this ultrasonication condition was selected for the preparation of NE-TW, NE-SL, NE-TS, and NE-SPI used in all subsequent stability, antibacterial, and application studies.

3.2. Interfacial stability of nanoemulsions under environmental stresses

The stability of nanoemulsions is governed not only by droplet size but also by the physicochemical properties of the interfacial layer surrounding the oil droplets. Ultrasonication under the optimized NE-2 condition generated nano-sized droplets with narrow size distributions, thereby reducing gravitational separation and increasing the effectiveness of interfacial stabilization. Previous studies have shown that ultrasound-induced cavitation promotes rapid emulsifier adsorption onto newly formed interfaces, resulting in compact interfacial films that improve resistance to environmental stresses such as centrifugation, ionic strength, temperature, and pH fluctuations [49], [50].

As shown in Fig. 2A, significant differences in centrifugal stability were observed among the nanoemulsions. NE-TS and NE-SL exhibited the lowest centrifugal stability constants (Ke), indicating superior resistance to droplet aggregation and phase separation under accelerated gravitational stress. The excellent stability of TS-stabilized nanoemulsions can be attributed to the amphiphilic triterpenoid saponin molecules, which rapidly adsorb at the oil–water interface and form a dense interfacial film providing both steric and electrostatic stabilization [51]. Similarly, SL forms phospholipid-based interfacial layers that can self-organize into ordered structures, enhancing interfacial rigidity and mechanical resistance. In contrast, SPI-stabilized nanoemulsions exhibited the highest Ke values, likely due to the relatively slow adsorption kinetics of protein molecules and their tendency to undergo bridging flocculation. TW showed intermediate stability, indicating that natural emulsifiers, particularly TS and SL, can provide comparable or even superior protection against centrifugal stress.

Fig. 2.

Stability evaluation of nanoemulsions stabilized by different emulsifiers (TW, SL, TS, and SPI). (A) Centrifugal stability of nanoemulsions after centrifugation at 4500 × g for 15 min. (B) Ionic stability of nanoemulsions in the presence of different NaCl concentrations (100, 200, 300, 400, and 500 mM). (C) Thermal stability of nanoemulsions at different temperatures (RT, 50, 60, 70, 80, and 90 ℃). (D) pH stability of nanoemulsions under different pH conditions (5, 6, 7, 8, and 9). Data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters indicate significant differences (P < 0.05).

The ionic stability results (Fig. 2B, Supplementary Fig. S1) further highlighted differences in stabilization mechanisms. The combined results of droplet size, PDI, turbidity, and encapsulation efficiency indicated that NE-TS maintained relatively stable physicochemical properties across the tested NaCl concentrations, suggesting excellent tolerance to charge screening. This behavior has been attributed to the combined electrostatic and steric stabilization provided by saponin molecules, which remains effective even when ionic strength increases. TW-stabilized nanoemulsions also showed high salt tolerance because their stabilization relies predominantly on steric repulsion rather than electrostatic interactions. In contrast, SL- and SPI-stabilized nanoemulsions exhibited larger increases in droplet size, PDI, and turbidity, accompanied by reduced encapsulation efficiency, indicating decreased stability with increasing NaCl concentration. Charge shielding may reduce electrostatic repulsion between phospholipid-coated or protein-coated droplets, thereby promoting droplet aggregation. This effect was particularly pronounced for SPI, where salt-induced screening may expose hydrophobic regions and enhanced protein-mediated flocculation [52].

Thermal stability tests (Fig. 2C, Supplementary Fig. S1) revealed that NE-TS retained high encapsulation efficiency and exhibited minimal changes in droplet size throughout the investigated temperature range, demonstrating excellent resistance to thermal stress. The superior thermal stability of TS may be related to its robust interfacial layer and the absence of thermally sensitive structural transitions. TW exhibited comparable thermal stability and served as a useful benchmark for comparison. In contrast, SL-stabilized nanoemulsions showed moderate thermal sensitivity, which may be associated with partial oxidation or structural rearrangement of phospholipids at elevated temperatures. SPI-stabilized nanoemulsions were the least thermally stable, showing more pronounced increases in droplet size and turbidity after heating. Heating can induce protein unfolding, intermolecular association, and subsequent droplet aggregation. Similar temperature-dependent destabilization of protein-stabilized nanoemulsions has been widely reported in food colloid systems.

The pH stability profiles (Fig. 2D, Supplementary Fig. S1) reflected the ionization behavior of the emulsifiers. As expected, TW remained largely unaffected across the tested pH range due to its nonionic nature. Among the natural emulsifiers, TS exhibited the broadest pH tolerance, maintaining relatively stable droplet size, PDI, turbidity, and encapsulation efficiency over a wide pH range because the ionizable carboxyl groups on saponin molecules maintained sufficient electrostatic repulsion between droplets. SL showed good stability under neutral conditions but became less stable under acidic environments, where protonation of phospholipid head groups reduced surface charge and weakened droplet repulsion. SPI exhibited the strongest pH dependence, with marked increases in droplet size, PDI, and turbidity under acidic conditions, consistent with flocculation occurring near the isoelectric point of soy proteins, where electrostatic repulsion is minimized. At alkaline pH, structural changes in adsorbed protein layers may further contribute to instability.

Overall, the stability results demonstrate that the optimized ultrasonic process produced nanoemulsions with strong resistance to multiple environmental stresses. The measurements of droplet size, PDI, and turbidity (Supplementary Fig. S1) were consistent with the encapsulation efficiency results, providing complementary evidence for the environmental stability of the nanoemulsions. Among the natural emulsifiers investigated, TS consistently exhibited the best overall stability, followed by SL, whereas SPI showed greater susceptibility to ionic, thermal, and pH-induced destabilization. Importantly, both TS- and SL-stabilized nanoemulsions displayed stability comparable to or exceeding that of the synthetic surfactant TW, highlighting their potential as clean-label alternatives for ultrasound-fabricated thyme essential oil nanoemulsions intended for food decontamination applications.

3.3. Storage stability of nanoemulsions under different temperatures

Building on the centrifugal, ionic, thermal, and pH stability results, the storage stability of nanoemulsions stabilized by different emulsifiers was further evaluated during 15 days of storage at 4 ℃ and 25 ℃ (Fig. 3). These combined stability tests provide a comprehensive assessment of the resistance of ultrasound-fabricated nanoemulsions to both immediate and long-term destabilization processes. Overall, all formulations maintained acceptable physical stability at 4 ℃, whereas clearer differences among emulsifiers emerged at 25 ℃, highlighting the role of interfacial composition in governing long-term stability.

Fig. 3.

Storage stability of nanoemulsions stabilized by different emulsifiers (TW, SL, TS, and SPI) during 15 days of storage at 4 ℃ and 25 ℃. (A) Representative photographs of nanoemulsions on day 0 (upper), after 15 days of storage at 4 ℃ (middle), and after 15 days of storage at 25 ℃ (lower). (B-C) Changes in particle size and PDI of nanoemulsions during storage at 4 ℃. (D-E) Changes in particle size and PDI of nanoemulsions during storage at 25 ℃. Representative photographs from three independent biological replicates are shown. Data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters indicate significant differences (P < 0.05).

The good initial stability of all nanoemulsions is attributed to ultrasonication under the optimized NE-2 condition, which produced uniformly nanosized droplets and promoted rapid emulsifier adsorption at the oil–water interface. The resulting compact interfacial layers effectively suppressed droplet coalescence and slowed Ostwald ripening during storage [53], [54].

Among all formulations, NE-TS exhibited the best storage stability, with the smallest increase in droplet size and consistently low PDI at both temperatures. This behavior is attributed to the amphiphilic saponin structure, which forms a dense interfacial layer providing both electrostatic repulsion and steric hindrance [51]. NE-SL also showed good stability, likely due to the formation of ordered phospholipid assemblies at the interface that enhance interfacial strength. In contrast, NE-SPI was the least stable system, particularly at 25 ℃, where noticeable droplet growth and sedimentation occurred due to protein aggregation and partial interfacial desorption. NE-TW showed moderate storage stability but a more pronounced increase in droplet size over time, especially at elevated temperature, suggesting limited resistance to long-term Ostwald ripening despite effective initial steric stabilization.

Together, the storage stability results are consistent with the centrifugal, ionic, thermal, and pH stability analyses. Across all stress conditions, NE-TS demonstrated the most robust and comprehensive stability, followed by NE-SL, whereas NE-SPI was the most susceptible to destabilization. Importantly, both TS- and SL-stabilized nanoemulsions exhibited stability comparable to or superior to the synthetic TW system, confirming the strong potential of these natural emulsifiers as clean-label alternatives for ultrasound-fabricated essential oil nanoemulsions in food preservation applications.

3.4. Antimicrobial activity

The antibacterial activities of coarse emulsions (CE) and nanoemulsions (NE) prepared under the optimized ultrasonication condition (NE-2) were evaluated against Escherichia coli O157:H7 and S. aureus (Table 1, Table 2, Fig. 4). Overall, ultrasonication markedly enhanced antimicrobial efficacy, as evidenced by reduced MIC and MBC values, larger inhibition zones, and faster bactericidal kinetics. This improvement is mainly attributed to acoustic cavitation-induced droplet size reduction, which increases interfacial area and facilitates closer interaction between essential oil components and bacterial membranes [13]. In addition, nano-sized droplets improve dispersion stability and may enhance membrane permeability through more efficient delivery of active compounds.

Table 1.

MIC (µg/mL) and MBC (µg/mL) for the emulsions against S. aureus, expressed as thyme essential oil (TEO)-equivalent concentrations (µg/mL). Values are representative of three independent experiments (n = 3).

	Coarse emulsion		Nanoemulsion
	MIC	MBC	MIC	MBC
TW	625	1250	312.5	1250
SL	312.5	1250	312.5	625
TS	156.3	312.5	156.3	156.3
SPI	625	625	312.5	625

Table 2.

MIC (µg/mL) and MBC (µg/mL) for the emulsions against E. coli, expressed as thyme essential oil (TEO)-equivalent concentrations (µg/mL). Values are representative of three independent experiments (n = 3).

	Coarse emulsion		Nanoemulsion
	MIC	MBC	MIC	MBC
TW	312.5	625	312.5	312.5
SL	312.5	312.5	156.3	156.3
TS	156.3	156.3	156.3	312.5
SPI	312.5	625	156.3	312.5

Fig. 4.

Antibacterial activity of nanoemulsions stabilized by different emulsifiers against Escherichia coli (E. coli) O157:H7 and S. aureus (S. aureus). (A) Representative images of inhibition zones produced by different nanoemulsions against E. coli O157:H7 and S. aureus. (B) Diameter of inhibition zones. (C-D) Time-kill kinetics of E. coli O157:H7 (C) and S. aureus (D) treated with nanoemulsions at 2 × MIC. Representative photographs from three independent biological replicates are shown. Data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters indicate significant differences (P < 0.05).

Among the four formulations, NE-TS exhibited the strongest antibacterial performance against both strains. For S. aureus, NE-TS showed MIC and MBC values as low as 156.3 µg/mL, with MBC equal to MIC, indicating a clear bactericidal effect. A similar trend was observed for E. coli, although a higher MBC than MIC suggested reduced susceptibility due to the outer membrane barrier of Gram-negative bacteria. The superior activity of TS-stabilized nanoemulsions is attributed to the synergistic effect between thyme essential oil and tea saponin, where the triterpenoid saponin structure can interact with membrane sterols and destabilize lipid bilayers, enhancing intracellular penetration.

NE-SL also showed strong antibacterial activity, particularly against E. coli, where MIC and MBC both reached 156.3 µg/mL. This may be related to the membrane-fusogenic nature of phospholipids, which facilitates transport of hydrophobic compounds across bacterial membranes [55]. NE-TW, used as a synthetic surfactant control, showed moderate antibacterial activity with improved MIC compared to its coarse counterpart but relatively higher MBC values, indicating mainly bacteriostatic effects. NE-SPI exhibited the weakest performance, consistent with its larger droplet size and limited membrane-disruptive capability even after ultrasonication.

The agar diffusion results served as a qualitative screening method and showed inhibition zones following the order TS > SL > TW > SPI for S. aureus and TS > SL ≈ TW > SPI for E. coli. Because inhibition-zone diameter is influenced not only by antibacterial activity but also by formulation-dependent diffusion through the agar matrix, these results were interpreted only as complementary evidence and considered together with the quantitative antibacterial assays. Time-kill assays demonstrated that NE-TS and NE-SL achieved rapid bacterial inactivation, with > 4–5 log reductions within 20–30 min, while NE-TW and NE-SPI showed slower and less complete killing over 60 min. The rapid bactericidal action of TS and SL systems can be attributed to the combined effects of nanoscale droplet delivery and intrinsic membrane-active properties of the natural emulsifiers, which enhance essential oil-membrane interactions and accelerate cell disruption. It should be noted that the time-kill assay was performed without formal neutralization validation; while samples were diluted below the MIC before plating, residual antimicrobial activity could potentially affect the enumerated viable counts. Therefore, the time-kill data should be interpreted as supportive evidence of bactericidal kinetics rather than as absolute quantification of immediate killing.

All ultrasound-fabricated nanoemulsions exhibited effective antibacterial activity against both S. aureus and E. coli, although their efficacy varied with the emulsifier used. The enhanced antibacterial performance after ultrasonication was closely associated with the reduction in droplet size, increased interfacial area, and improved dispersion stability, which collectively facilitated the interaction between thyme essential oil components and bacterial cells. Among the formulations, TS- and SL-stabilized nanoemulsions generally showed stronger activity, which corresponded well with their smaller droplet sizes, lower PDI values, and better environmental stability. In contrast, the relatively weaker antibacterial performance of SPI-stabilized nanoemulsions was consistent with their larger droplet size and less stable interfacial characteristics. TW- and SPI-stabilized systems also demonstrated substantial antibacterial effects compared with their corresponding coarse emulsions.

As all four nanoemulsions displayed pronounced antibacterial activity but differed in interfacial characteristics and stabilization mechanisms, they were further investigated to elucidate the relationship between emulsion structure and antibacterial performance and to evaluate their potential in food decontamination applications.

3.5. Antibacterial mechanisms of nanoemulsions

To elucidate how the physicochemical characteristics of ultrasound-fabricated nanoemulsions contribute to antibacterial activity, a series of membrane integrity and oxidative stress assays were performed using E. coli O157:H7 and S. aureus treated with NE-TW, NE-SL, NE-TS, and NE-SPI at 2 × MIC concentrations (Fig. 5). Because the mechanistic assays were performed at formulation-specific 2 × MIC concentrations, the results primarily reflect the antibacterial responses of each nanoemulsion under its effective inhibitory condition rather than a direct comparison at identical thyme essential oil or emulsifier concentrations. Since the four nanoemulsions differed in droplet size, physicochemical properties, and environmental stability, these experiments also provide insight into the possible relationship between nanoemulsion structure and antibacterial efficacy.

Fig. 5.

Mechanistic insights into the antibacterial activity of nanoemulsions stabilized by different emulsifiers against E. coli O157:H7 and S. aureus. (A-B) Scanning electron microscopy (SEM) images of E. coli O157:H7 (A) and S. aureus (B) before and after nanoemulsion treatment. (C-D) Leakage of intracellular nucleic acids (C) and proteins (D). (E) Changes in outer membrane permeability of E. coli O157:H7. (F) Changes in inner membrane permeability of E. coli O157:H7 and S. aureus. (G) Changes in membrane potential of E. coli O157:H7 and S. aureus. (H) Intracellular reactive oxygen species (ROS) production following nanoemulsion treatment. Representative SEM images from three independent biological replicates are shown. Quantitative data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters demonstrated a significant difference at P < 0.05.

SEM observations suggested marked morphological alterations in both bacterial species following nanoemulsion treatment (Fig. 5A-B). Untreated cells exhibited smooth and intact surfaces with well-defined morphology, whereas all treated cells showed varying degrees of deformation, shrinkage, surface collapse, and aggregation. More pronounced morphological changes were observed in the NE-TS- and NE-SL- treated cells than in other treatment groups. Similar membrane-destructive effects have recently been reported for essential oil nanoemulsions, where nanoscale droplets were suggested to facilitate interactions with bacterial envelopes and accelerated membrane destabilization [14]. The pronounced morphological alterations observed for NE-TS are consistent with its favorable physicochemical characteristics, including smaller droplet size, lower PDI, and superior stability, which may increase the probability of droplet-cell interactions and facilitate the delivery of active compounds to bacterial surfaces.

The disruption of membrane integrity was further confirmed by the leakage of intracellular nucleic acids and proteins (Fig. 5C-D). All nanoemulsions significantly increased extracellular OD₂₆₀ and OD₂₈₀ values compared with the untreated control, indicating loss of membrane barrier function. Among the formulations, NE-TS induced the highest leakage of both nucleic acids and proteins in E. coli, suggesting severe membrane permeabilization and cytoplasmic leakage. NE-TW and NE-SPI also caused substantial leakage, whereas NE-SL showed comparatively moderate release despite its strong antibacterial activity. This observation suggests that SL-stabilized nanoemulsions may act through membrane fusion and permeability modulation rather than extensive membrane rupture. Similar leakage phenomena have been reported for nanoemulsion-based antimicrobial systems, where disruption of membrane integrity was identified as a primary cause of bacterial inactivation.

To further evaluate membrane damage, outer membrane (OM) and inner membrane (IM) permeability were assessed using NPN uptake and ONPG hydrolysis assays (Fig. 5E-F). All nanoemulsions significantly increased OM permeability in E. coli, indicating disruption of the lipopolysaccharide barrier. NE-TS produced the highest NPN fluorescence intensity, suggesting the most effective permeabilization of the outer membrane. Consistent trends were observed for IM permeability, where NE-TS generated the highest ONPG hydrolysis in both bacterial species, followed by NE-TW and NE-SL. The enhanced membrane permeabilization induced by NE-TS may be attributed to the amphiphilic structure of tea saponin, which can interact with membrane lipids and facilitate penetration of thyme essential oil into bacterial cells. The relatively small droplet size and high surface charge of NE-TS may further promote interactions with bacterial envelopes, resulting in more extensive membrane disruption.

Membrane depolarization assays demonstrated that all nanoemulsions significantly affected bacterial membrane potential (Fig. 5G). Compared with untreated cells, fluorescence intensity decreased markedly after treatment, indicating dissipation of the transmembrane electrochemical gradient. NE-TS again induced the strongest depolarization in both E. coli and S. aureus, while NE-TW, NE-SL, and NE-SPI produced moderate but significant effects. Since membrane potential is essential for ATP synthesis, nutrient transport, and cellular homeostasis, its collapse indicates severe impairment of membrane function and energy metabolism [37]. The observed depolarization is consistent with the membrane leakage and permeability results described above, suggesting that disruption of membrane integrity is a major antibacterial target of all nanoemulsions.

In addition to membrane damage, intracellular oxidative stress contributed substantially to bacterial inactivation (Fig. 5H). All nanoemulsion treatments significantly increased ROS production relative to the control, with NE-TS inducing the highest ROS accumulation in both bacterial species. Excessive ROS can oxidize membrane lipids, proteins, and nucleic acids, thereby amplifying cellular damage initiated by membrane disruption. The elevated ROS levels observed for NE-TS may be related to its enhanced cellular penetration and more efficient delivery of active essential oil components, resulting in stronger interference with intracellular redox balance. Similar synergistic effects between membrane disruption and ROS generation have been widely reported for essential oil nanoemulsions and other nanostructured antimicrobial systems [13], [56].

The mechanistic results from the fluorescence-based assays (NPN, ONPG, Rhodamine 123, and DCFH-DA) suggest that membrane disruption is a primary antibacterial target of the four nanoemulsions. Ultrasonication-generated nanodroplets substantially increased the interfacial area available for interaction with bacterial cells, facilitating the delivery of thyme essential oil to the cell envelope. Consistent with this interpretation, all formulations were associated with increased membrane permeabilization, intracellular leakage, membrane depolarization, and ROS accumulation. Together with the antibacterial activity assays (MIC, MBC, and time-kill kinetics) and SEM observations, these findings provide supportive evidence for the proposed antibacterial mechanism. However, additional confirmatory approaches (such as flow cytometry, confocal microscopy, or ATP leakage assays) will be required to further validate the mechanistic findings.

Despite these common responses, the four nanoemulsions exhibited distinct antibacterial modes that were closely associated with their physicochemical characteristics and interfacial composition. Owing to its smallest droplet size (124.8 nm), lowest PDI (0.127), highest zeta potential, and superior environmental and storage stability, NE-TS displayed the strongest membrane-disruptive activity, as evidenced by severe morphological collapse, extensive leakage of intracellular constituents, high membrane permeability, and pronounced ROS accumulation. These results suggest that NE-TS acts predominantly through direct membrane destruction coupled with oxidative stress amplification. The amphiphilic structure of tea saponin likely facilitates insertion into bacterial membranes and promotes the penetration of thyme essential oil into the lipid bilayer.

In contrast, NE-SL exhibited strong antibacterial activity despite causing comparatively lower leakage of intracellular materials. It is hypothesized that the phospholipid-rich interfacial layer formed by soybean lecithin may facilitate fusion-like interactions with bacterial membranes or promote the transfer of hydrophobic essential oil compounds into cells through lipid mixing or membrane perturbation, without causing extensive membrane rupture. Similar membrane interactions have been reported for phospholipid-based liposomal and lipid nanoparticle systems, in which lipid exchange or membrane fusion may contribute to cargo delivery [57], [58]. This hypothesis is based on the combination of strong antibacterial activity and the relatively low leakage of intracellular materials observed for NE-SL; however, direct experimental evidence (e.g., lipid-mixing assays, FRET-based fusion measurements, or cryo-TEM visualization) will be included in our future studies to determine whether a fusion-mediated interaction is involved and to further elucidate the interaction between phospholipid-stabilized nanoemulsions and bacterial membranes.

NE-TW primarily exerted its antibacterial effect through surfactant-mediated membrane perturbation. The small droplet size generated by ultrasonication enhanced bacterial contact, while the polyoxyethylene chains of Tween 60 destabilized membrane organization and increased permeability. However, unlike TS and SL, TW lacks specific membrane-active biological functionalities, resulting in moderate membrane damage and oxidative stress responses.

For NE-SPI, the larger droplet size and thicker protein-based interfacial layer likely reduced direct interactions between nanoemulsion droplets and bacterial membranes. Nevertheless, significant membrane permeability changes, depolarization, and ROS generation were still observed, indicating that antibacterial activity was achieved through the gradual release of thyme essential oil and subsequent membrane stress. In addition, protein-coated droplets may promote bacterial aggregation and surface adsorption, contributing to growth inhibition.

Taken together, these findings suggest that the antibacterial efficacy of thyme essential oil nanoemulsions is associated with both ultrasonication-induced droplet refinement and emulsifier-dependent physicochemical characteristics. The integration of particle size, stability, antibacterial activity, membrane integrity assays, and oxidative stress analysis indicates that interfacial properties play an important role in regulating the biological performance of these nanoemulsions. While all four formulations shared a membrane-targeting antibacterial pathway, TS-, SL-, TW-, and SPI-stabilized nanoemulsions may exhibited patterns consistent with membrane destruction-, membrane fusion-, surfactant perturbation-, and protein-interface-mediated modes of action, respectively. These observations highlight the critical role of interfacial design in regulating the biological performance of essential oil nanoemulsions and provides a theoretical basis for their application in subsequent food decontamination studies. However, additional mechanistic studies are required before distinct emulsifier-specific modes of action can be established.

3.6. Application of nanoemulsions in food decontamination

The practical applicability of the ultrasound-fabricated nanoemulsions was evaluated using two representative food systems, including artificially contaminated cherry tomatoes and milk with different fat contents (Fig. 6, Fig. 7). These experiments were designed to assess whether the physicochemical advantages and antibacterial mechanisms identified in previous sections could be translated into effective microbial control in real food matrices [53], [59].

Fig. 6.

Decontamination efficacy of nanoemulsions stabilized by different emulsifiers on cherry tomato surfaces and their effects on microbial contamination in washing solutions. (A-B) Reduction of E. coli O157:H7 (A) and S. aureus (B) on cherry tomato surfaces after washing with different nanoemulsions, with representative PCA plate images. (C-D) Viable counts of E. coli O157:H7 (C) and S. aureus (D) in the washing solutions after treatment, with representative PCA plate images. Representative plate images from three independent biological replicates are shown. Quantitative data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters demonstrated a significant difference at P < 0.05.

Fig. 7.

Antibacterial efficacy of nanoemulsions stabilized by different emulsifiers in milk with varying fat contents. Viable counts of E. coli O157:H7 (A) and S. aureus (B) in whole, low-fat, and skim milk following nanoemulsion treatment. (C) Representative images of selective plates: MacConkey agar for E. coli O157:H7 (red colonies) and Baird Parker agar for S. aureus (black colonies), showing the growth from milk samples after treatment with different nanoemulsions. Representative plate images from three independent biological replicates are shown. Quantitative data are presented as mean ± SD from three independent biological replicates (n = 3). Different letters demonstrated a significant difference at P < 0.05.

On cherry tomato surfaces, all nanoemulsions significantly reduced the populations of both E. coli O157:H7 and S. aureus after 3 min of washing (Fig. 6A-B). The reductions ranged from 2.17 to 3.91 log CFU/g for E. coli and 2.35–4.59 log CFU/g for S. aureus, providing preliminary evidence that ultrasound-fabricated thyme essential oil nanoemulsions have potential for reducing bacterial contamination on fresh produce under laboratory conditions. Among the formulations, NE-TS achieved the greatest reductions for both pathogens, followed by NE-SPI and NE-SL, whereas NE-TW showed comparatively lower efficacy. These results are consistent with the antimicrobial activity and mechanistic analyses presented in 3.4, 3.5, where NE-TS exhibited the strongest membrane-disruptive effects and highest ROS generation. The superior decontamination performance of NE-TS may also be related to its smaller droplet size, lower PDI, and enhanced interfacial stability, which may facilitate more efficient delivery of TEO to bacteria attached to the produce surface.

In addition to surface decontamination, the microbial load in the washing water was significantly reduced by all nanoemulsion treatments (Fig. 6C-D). Notably, no viable E. coli or S. aureus cells were detected in the wash water after treatment with NE-TS, while NE-SL and NE-SPI also reduced bacterial counts to very low levels. In contrast, substantial bacterial populations remained in the control washing solution, confirming that conventional washing primarily transfers microorganisms from the produce surface into the washing medium rather than eliminating them. The simultaneous reduction of microbial contamination on both produce surfaces and wash water is particularly important for industrial washing operations, where cross-contamination through recycled water remains a major food safety concern.

The antibacterial efficacy of the nanoemulsions was further evaluated in milk with different fat contents (Fig. 7). Compared with the cherry tomato model, the antimicrobial performance was more strongly influenced by food composition, particularly fat content. For E. coli, all nanoemulsions exhibited higher antibacterial activity in skim and low-fat milk than in whole milk. The reduced efficacy observed in whole milk is likely due to the partitioning of hydrophobic essential oil components into milk fat globules, which decreases their availability in the aqueous phase where bacterial cells are located [60], [61]. Similar matrix-dependent effects have been widely reported for essential oil-based antimicrobial systems [62].

Despite the inhibitory effect of milk fat, the natural emulsifier-based nanoemulsions maintained strong antibacterial activity. NE-TS consistently achieved the greatest reductions against E. coli across all milk systems, while NE-SL and NE-SPI also produced substantial bacterial inactivation. Against S. aureus, NE-TS, NE-SL, and NE-SPI completely eliminated the pathogen in skim and low-fat milk and remained highly effective even in whole milk. Their superior performance in dairy systems may be attributed to the robust interfacial layers formed by tea saponin, phospholipids, and soy proteins, which help retain nanoemulsion stability and facilitate the transport of thyme essential oil in complex food matrices [52]. In comparison, the antibacterial efficacy of NE-TW decreased more markedly in milk, suggesting that the synthetic surfactant-based system was more susceptible to matrix interference. It should be noted that these experiments were designed as laboratory-scale challenge tests to compare the antibacterial performance of different nanoemulsion formulations in dairy matrices. Further studies are required to optimize treatment concentrations and to evaluate sensory properties, residual essential oil levels, and interactions with milk proteins and fat globules before practical application in dairy preservation can be established.

The food application results corroborate the physicochemical and mechanistic findings. Nanoemulsions exhibiting smaller droplet size, stronger interfacial stability, and greater membrane-targeting activity generally achieved better antimicrobial performance in food systems. Among the formulations, TS-stabilized nanoemulsions displayed the most consistent efficacy across different application scenarios, while SL- and SPI-stabilized nanoemulsions also maintained strong antibacterial activity, particularly in dairy matrices. These findings demonstrate that the interfacial properties imparted by different emulsifiers play a crucial role in determining the practical performance of thyme essential oil nanoemulsions and support their potential application as natural antimicrobial systems for food decontamination and preservation.

While the emulsifiers and thyme essential oil used in this study are derived from natural sources and are generally recognized as safe for use in food systems, their application at antimicrobial concentrations may present sensory, technological, and regulatory constraints that should not be overlooked [63]. The concept of “clean label” is primarily consumer-driven and generally refers to food products formulated with short ingredient lists, familiar and naturally derived ingredients, and minimal artificial additives, although no universally accepted regulatory definition currently exists [28], [64]. Therefore, the term “clean-label” in the present study refers only to the use of recognizable, naturally derived ingredients and should not be interpreted as evidence of environmental sustainability, regulatory compliance, or consumer acceptability at the concentrations investigated. Accordingly, future studies should evaluate residual exposure, cytotoxicity, consumer acceptance, and environmental sustainability via life cycle assessment (LCA) to better assess practical feasibility. Further work should also combine droplet evolution monitoring with kinetic modeling, thermodynamic analysis of interfacial behavior, and techno-economic analysis (TEA) to support scale-up and industrial application of ultrasound-fabricated nanoemulsions.

4. Conclusion

In this study, thyme essential oil nanoemulsions stabilized by tea saponin (TS), soybean lecithin (SL), soy protein isolate (SPI), and Tween 60 (TW) were successfully fabricated using ultrasonic emulsification. Ultrasonication markedly reduced droplet size, improved size uniformity, and enhanced encapsulation efficiency, resulting in nanoemulsions with improved physicochemical properties and antibacterial performance. Among the tested formulations, TS- and SL-stabilized nanoemulsions exhibited better resistance to environmental and storage stresses, indicating their potential as effective food-grade emulsifier alternatives for ultrasound-fabricated nanoemulsion systems. All nanoemulsions showed significant antibacterial activity against E. coli O157 and S. aureus, with NE-TS generally exhibiting the strongest inhibitory and bactericidal effects. The combined antibacterial and mechanistic results suggested that bacterial inactivation may be associated with membrane damage, increased membrane permeability, intracellular leakage, membrane depolarization, and ROS accumulation. However, these findings mainly provide supportive evidence for the antibacterial actions, and further advanced techniques (such as TEM, flow cytometry, or confocal microscopy) are required to fully elucidate the molecular interactions between nanoemulsions and bacterial cells. The antibacterial performance was also influenced by the physicochemical characteristics and emulsifier-dependent properties of the nanoemulsions. Importantly, the nanoemulsions maintained strong antimicrobial efficacy in practical food systems, including contaminated cherry tomatoes and milk with different fat contents, demonstrating their potential for food decontamination applications. The results highlight the critical role of interfacial design in determining the stability, antibacterial activity, and application performance of essential oil nanoemulsions. Nevertheless, further investigations involving sensory evaluation, safety assessment, residual active compound analysis, and scale-up feasibility are required before practical application in real food systems. Overall, tea saponin and soybean lecithin represent promising food-grade emulsifiers for developing ultrasound-fabricated essential oil nanoemulsions for food decontamination and preservation applications.

CRediT authorship contribution statement

Qiao He: Writing – original draft, Visualization, Investigation. Shuting Yuan: Writing – original draft, Software, Methodology. Yong Bian: Visualization, Investigation. Zixin Zhuang: Visualization, Investigation. Can Ding: Visualization, Investigation. Can Zhang: Visualization, Investigation. Chengxu Tu: Investigation, Visualization. Chenze Lu: Writing – review & editing, Validation, Supervision. Han Jiang: Supervision, Funding acquisition, Conceptualization.

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

The following are the Supplementary data to this article:

Supplementary Data 1

Fig. S1. Physicochemical characterization of nanoemulsions after environmental stability treatments. Table S1. Chemical composition of thyme essential oil via GC-MS.