Stable Emulsions Constructed From Whey Protein Isolate–Arabinogalactan Conjugates: Preparation, Characterization, and β‐Carotene Loading

ABSTRACT

In this study, the results indicated that whey protein isolate–arabinogalactan (WPI–AG) conjugate emulsion exhibited optimal stability at pH 7.0 with a 25% oil phase volume fraction (φ), demonstrating the highest absolute zeta‐potential value, the lowest interfacial tension, and excellent viscoelastic properties. Confocal laser scanning microscopy (CLSM) verified the development of a compact interfacial layer with high coverage. Furthermore, ultrasonication strengthened the stability of β‐carotene‐loaded WPI–AG oil‐in‐water (O/W) emulsions via mechanical shear, which reduced droplet size, increased the absolute zeta potential, and optimized the overall interfacial structure. These structural improvements led to significant enhancements in both physical stability (resistance to ionic, thermal, and pH stresses) and chemical stability (during storage and against light exposure). The 400 W treatment group indicated the most pronounced effects, with the highest cumulative release rate of β‐carotene during in vitro digestion and a bioaccessibility of 81.98% ± 2.1%, which was markedly higher than that of the control group (49.55% ± 4.31%, p < 0.05). This work offers a theoretical basis for fabricating natural stable emulsions as delivery vehicles for lipophilic bioactive substances.

1. Introduction

β‐Carotene, a significant carotenoid, offers various physiological benefits. It functions as a potent antioxidant, enhances immune function, and contributes to the prevention of cardiovascular and cerebrovascular diseases (Feng et al. 2025). Owing to its unsaturated structure, β‐carotene is prone to isomerization under external stressors—including high salinity, elevated temperatures, and acidic environments—commonly encountered during food processing and storage, potentially leading to a loss of its provitamin A activity (Li et al. 2022). Consequently, significant efforts have been devoted to developing effective delivery systems to enhance the stability, utilization, and bioavailability of carotenoids in food products (Qian et al. 2012).

Currently, protein‐stabilized oil‐in‐water (O/W) emulsions have been demonstrated to be a suitable delivery system for β‐carotene, enhancing its dispersibility, stability, and bioavailability (Han et al. 2022; Qian et al. 2012). However, protein‐stabilized emulsions are sensitive to environmental stresses such as temperature, ionic strength, and pH, which can lead to flocculation and coalescence (Fan et al. 2017; Zhang et al. 2023). Consequently, it is necessary to develop strategies for enhancing the physical and chemical stability of protein‐stabilized β‐carotene emulsions. Among the various available approaches, conjugates formed via the Maillard reaction between proteins and polysaccharides have demonstrated exceptional efficacy, as they concurrently improve both the physical and chemical stability of emulsions. W. Chen, Wang, et al. (2022) reported that conjugates formed between whey protein isolate (WPI) and gum arabic via the Maillard reaction enhanced both the chemical stability of β‐carotene in emulsions and its bioaccessibility. Li et al. (2022) reported that emulsions stabilized with conjugates of arabinoxylan hydrolysates and soy protein isolate demonstrated improved chemical stability and increased β‐carotene bioaccessibility. These results indicate that O/W emulsions stabilized by protein–polysaccharide conjugates prepared via the Maillard reaction are ideal delivery systems for β‐carotene, enhancing its physicochemical stability and bioaccessibility (Li et al. 2022).

WPI, a by‐product generated during casein and cheese production in the dairy industry, is highly prized due to its excellent functional properties and wide‐ranging applications in the food and pharmaceutical industries. Endowed with unique structural and functional features—such as high digestibility, the capability to release bioactive peptides, and surface‐active properties—WPI becomes indispensable for various applications, from sports nutrition to encapsulation technologies (Hu et al. 2022; Yuliarti et al. 2019). Yet, despite its diverse applications, the intrinsic stability and functionality of WPI are commonly limited by its native structure, which can hinder its performance under specific processing conditions or within complex food matrices. A major limiting factor lies in the vulnerability of its functional groups to environmental stresses (including pH, ionic strength, and temperature), thus restricting its more extensive application across different food systems and processing technologies (Fan et al. 2017; Zhang et al. 2023; Zhou et al. 2022). To enhance the physicochemical properties of WPI, our previous study has demonstrated that conjugates of WPI and arabinogalactan (AG), prepared via the wet‐heated Maillard reaction, exhibited improved emulsifying capabilities, foam stability, antioxidant activity, heat resistance, and solubility (Gao et al. 2026). Although whey protein isolate–arabinogalactan (WPI–AG) conjugates exhibited potential for application in emulsion systems, their adsorption behaviors at the oil–water interface, the resulting interfacial architecture, and the underlying mechanisms by which they stabilize emulsions remain inadequately investigated and lack systematic elucidation. Therefore, this study hypothesized that WPI–AG conjugates enhanced the structural stability of O/W emulsions by modifying the interfacial properties of the emulsion droplets.

Hence, the present study initially evaluated the effects of various pH values (5.0, 6.0, 7.0, 8.0, and 9.0) and oil phase volume fraction (φ = 10%, 15%, 20%, 25%, and 30%) on the physicochemical characteristics of WPI–AG conjugate‐stabilized O/W emulsions. The analyzed properties included microstructure, particle size distribution, zeta‐potential, turbidity, interfacial protein adsorption, interfacial tension, and rheological behavior. Subsequently, the influence of different ultrasonic power intensities on WPI–AG conjugate‐stabilized β‐carotene‐loaded emulsions was investigated, with an emphasis on the emulsions’ physicochemical stability during storage and β‐carotene's in vitro release rate and bioaccessibility in simulated digestion. This work provided a theoretical foundation for developing high‐performance encapsulation and delivery systems with enhanced bioaccessibility.

2. Materials and Methods

2.1. Materials

WPI (purity > 90%) and AG (purity > 90%) were purchased from Shanghai Yuanye Bio‐Technology Co. Ltd. (Shanghai, China). Corn oil was procured from Walmart (Fuzhou, China). The BCA protein assay kit, Nile red, and Nile blue were also obtained from Solarbio (Beijing, China). All other reagents were of at least analytical purity.

2.2. Preparation of WPI–AG Conjugate

First, a 10% (w/v) WPI solution and a 5% (w/v) AG solution were prepared. Both solutions were placed in a 40°C water bath and stirred magnetically overnight to ensure complete dissolution of the solutes. Equal volumes of the WPI solution and the AG solution were mixed and treated using an ultrasonic cell disruption instrument (SCIENTZ‐IID; Ningbo Scientz Biotechnology Co. Ltd., China) with a Φ6 probe: the ultrasonic power was 300 W, the working time was 40 min, the pulse mode was 5 s of work followed by 2 s of pause, and the system was kept in an ice‐water bath for temperature control, ensuring that the system temperature did not exceed 35°C. After the ultrasonic treatment, the pH of the mixture was adjusted to 7.0 and then heated at 90°C for 1.5 h to promote the formation of the WPI–AG complex in the aqueous phase. Finally, the resulting sample was subjected to freeze‐drying.

2.3. Preparation of WPI–AG Conjugates Emulsion

2.3.1. Stable Emulsions Under Different pH Conditions

O/W emulsions were stabilized using the synthesized WPI–AG conjugate. The WPI–AG conjugate (2.5% w/w) was dissolved in distilled water and stirred for 2 h. The resulting solution was adjusted to different pH values (5.0–9.0) using 1 M HCl or NaOH, followed by storage at 4°C overnight. The solution was then combined with corn oil (φ = 20%) and subjected to homogenization at 10,000 rpm for 2 min using a high‐speed homogenizer. Lastly, the coarse emulsion underwent sonication at 500 W for 5 min by means of an ultrasonic cell disruptor maintained in an ice‐water bath.

2.3.2. Stable Emulsions Under Different Oil Phase Volume Fraction

Similarly, 2.5% (w/w) of the WPI–AG conjugate was dissolved in distilled water, followed by stirring for 2 h. The resulting solution was adjusted to pH 7.0 using 1 M HCl or NaOH and then stored at 4°C overnight. The hydrated solution was subsequently mixed with varying corn oil fractions (φ = 10%, 15%, 20%, 25%, and 30%), followed by homogenization and ultrasonication.

2.4. Creaming Index (CI)

Briefly, the fresh emulsion was transferred into 10 mL transparent glass bottles, properly sealed, and stored at room temperature in the dark. The upper layer corresponds to the emulsion layer, whereas the lower layer is a transparent layer; the CI was calculated using the following formula:

Among them, H_c corresponds to the height of the clear serum layer, and H_t denotes the total height of the emulsion, which was recorded throughout a 14‐day period.

2.5. Particle Size and Ζeta Potential

The zeta potential and particle size of the samples were measured using a Malvern Zetasizer Nano ZS (Malvern Panalytical, Worcestershire, UK). To preclude multiple scattering effects, deionized water was used to dilute all samples to a concentration of 0.1 mg/mL prior to measurement. Tests were performed at 25°C, where each reported value is the average of 13 repeated runs. The refractive index and equilibration time were set to 1.33 and 120 s, respectively.

2.6. Determination of Centrifugal Precipitation Rate (CPR)

The CPR was determined with modifications according to a method described by a previous study (Li et al. 2019). Centrifugation was performed on a 10 mL sample at 2500 × g for 40 min. The supernatant was removed by decantation, and the tube was inverted for 30 min to allow residual liquid to drain. The precipitate's weight was subsequently measured. The CPR was calculated using the following equation:

$$\mathrm{CPR}(\%)=m_1/m_2\times 100\%$$

Among them, the weight of the precipitate after centrifugation of m ₁ and m ₂ are the weights of 10 mL of emulsion.

2.7. Determination of Stability Coefficient

Centrifugation was performed on a 10 mL emulsion sample at 2000 × g for 15 min. The resulting supernatant was collected, diluted 100‐fold with distilled water, and its absorbance was quantified at 750 nm with a spectrophotometer. Thereafter, the stability coefficient was calculated using the formula provided below:

$$R=A_2/A_1\times 100\%$$

where R is the stability coefficient. A ₁ denotes the absorbance of the original sample at a wavelength of 750 nm, and A ₂ is the absorbance of the supernatant at 750 nm.

2.8. Turbidity Titration

The sample was subjected to 50‐fold dilution with deionized water, and the turbidity of the resulting dispersion was measured at 600 nm via an under ultraviolet (UV)‐visible spectrophotometer (UV‐2550, Shimadzu Corporation, Japan) using a 1 cm pathlength cuvette.

2.9. Adsorbed Protein (AP)

The AP content at the emulsion interface was quantified following the method of a previous study (W. Wang, Wang, Yao, et al. 2022) with minor modifications. A Bradford assay kit (Beijing, China) was used to determine the protein concentration in the aqueous phase. Briefly, 100 µL of the aqueous phase sample (collected after emulsion centrifugation) was mixed with 3.0 mL of Bradford reagent in a cuvette. The mixture was vortexed for 5 s and incubated at room temperature for 10 min to allow color development. The absorbance was then measured at 595 nm using a UV‐visible spectrophotometer (UV‐2550, Shimadzu Corporation, Japan). The protein concentration was calculated from a standard curve prepared with bovine serum albumin (BSA) as the standard. The percentage of AP was calculated as follows:

$$\mathrm{AP}(\%)=\frac{C_0-C_i}{C_0}\times 100\%$$

where C ₀ and C_i , respectively, represent the initial and unadsorbed protein concentrations.

2.10. Interfacial Tension

Interfacial tension was measured using an optical contact angle tensiometer (OCA25, DataPhysics Instruments GmbH, Germany) equipped with a high‐speed camera, based on the pendant drop method. Five microliters of the sample, dispensed as a droplet, were introduced into a cuvette with purified corn oil and incubated for 1.5 h to allow for sufficient protein adsorption.

2.11. Rheological Properties

The rheological behavior of the emulsions was assessed using a rheometer (DHR‐2, TA Instruments, USA) equipped with a 40 mm aluminum parallel‐plate geometry. Oscillatory frequency sweeps were performed across an angular frequency range of 0.1–100 rad/s to measure the storage modulus (G′) and loss modulus (G″). Steady‐state flow experiments were also performed within a shear rate range of 0.1–100 s⁻¹ to quantify the apparent viscosity of the emulsions.

2.12. Confocal Laser Scanning Microscopy (CLSM)

CLSM (LSM880, Carl Zeiss AG, Germany) was utilized for the observation of the emulsions’ microstructure and droplet distribution. For sample preparation, isopropanol was used as the solvent to prepare 1% (w/v) Nile blue solution and 0.1% (w/v) Nile red solution. Next, 25 µL Nile blue solution and 20 µL Nile red solution were introduced into 0.5 mL of the prepared emulsion, followed by homogenization and filtration of the resultant mixture. The treated sample was then allowed to stand for 30 min prior to CLSM observation.

2.13. Preparation of WPI–AG Embedded β‐Carotene Emulsion

β‐Carotene powder was added to corn oil at a concentration of 0.5% (w/w), and the mixture was stirred at 40°C for 1 h to dissolve the powder. Subsequently, φ = 25% oil phase was combined with the WPI–AG conjugate solution (2.5%, w/w, pH 7.0), and pre‐homogenization was performed at 10,000 rpm for 5 min. The pre‐homogenization cycle was duplicated twice, resulting in the formation of a coarse emulsion. The coarse emulsion was subsequently sonicated using an ultrasonic cell disruptor at power levels of 300, 400, 500, and 600 W for 40 min each. During sonication, samples were maintained in an ice‐water bath to keep the temperature below 35°C, using a pulse mode (on/off = 5/2 s). An emulsion without sonication treatment served as the control.

2.14. Physical Stability

2.14.1. Ionic Strength Stability

Different volumes of NaCl solution were added to adjust the ionic strength of the emulsion samples to target final concentrations: 0, 100, 200, and 300 mM. After 24 h of storage, the particle size and zeta potential of the samples were measured.

2.14.2. Thermal Stability

Two different heat treatment conditions were applied to the emulsion samples using a water bath: one at 65°C for 30 min and the other at 90°C for 30 s. After the samples were cooled to approximately 25°C, their particle size and zeta potential were measured.

2.14.3. pH Stability

Each emulsion sample was adjusted to different pH conditions: 3.0, 5.0, 7.0, and 9.0. After 24 h of storage at ambient temperature, the particle size and zeta potential were measured.

2.15. Chemical Stability

2.15.1. Storage Stability

Sealed 10 mL brown glass vials were filled with emulsion samples, which were then stored under three sets of conditions: refrigeration at 4°C and incubation at 25°C and 55°C using an incubator shaker. Over an 8‐day storage period, samples were retrieved every 2 days for analysis. β‐Carotene retention was tracked periodically and quantified as a percentage in accordance with the following equation: Retention (%) = C_t /C ₀ × 100%, where C_t is the β‐carotene concentration at a given storage time and C ₀ is the initial β‐carotene concentration.

2.15.2. Photostability Under Ultraviolet (UV) Radiation

The UV stability of the β‐carotene emulsion was assessed using a modified method adapted from a prior investigation (W. Chen, Wang, et al. 2022). The sample was diluted 5‐fold using deionized water, then transferred into a transparent glass vial and exposed to UV light for an 8‐h period. Aliquots were collected at 2 h intervals for analysis.

2.15.3. Analysis of β‐Carotene Content

Briefly, each 1 mL emulsion sample was extracted three times with 3 mL portions of a 1:2 (v/v) ethanol/hexane mixture. The three extracts were pooled together, and the absorbance of the obtained solution was quantified at 450 nm.

2.16. Rheological Properties

Refer to Section 2.11.

2.17. In Vitro Digestion

To mimic the successive gastric and intestinal digestion processes in humans, a sequential in vitro digestion model was developed. For gastric digestion, 15 mL of emulsion was mixed with an equal volume of simulated gastric fluid (SGF); the pH of the mixture was adjusted to 2.5, and the SGF contained 2 g/L NaCl, 7 mL/L of 37% HCl, and 3.2 g/L pepsin. The mixture was then subjected to incubation at 37°C with continuous shaking at 100 rpm for 2 h. For intestinal digestion, the gastric digest was mixed with an equal volume of simulated intestinal fluid (SIF), and the pH was adjusted to 7.5. The SIF was formulated with 0.30 mmol/L CaCl₂, 30.72 mmol/L NaCl, 5 mg/mL bile salts, and 8 mg/mL pancreatin, and the mixture was incubated at 37°C with shaking at 100 rpm for 2 h.

2.17.1. Droplet Size (Sauter Mean Diameter, D _4,3) and Zeta Potential

Measurements of the emulsion's zeta‐potential and particle size were conducted at every stage of digestion.

2.17.2. CLSM Analysis

Refer to 2.12.

2.17.3. Cumulative Release Rate and Bioaccessibility

As described in Section 2.15.3, the bioaccessibility and cumulative release rate of β‐carotene were determined. Quantification was performed on the amount of β‐carotene released into the digesta at different digestion times, and the cumulative release rate (%) under simulated gastrointestinal conditions was computed with the equations provided below:

$$\mathrm{Cumulative}\mathrm{release}\mathrm{rate}\left(\%\right)=\frac{C_{\mathrm{released}}}{C_{\mathrm{initial}}}\times 100\%$$

After the in vitro digestion procedure, centrifugation was performed on the intestinal samples at 10,000 × g for 10 min at 4°C, aiming to isolate the micellar phase that contained solubilized β‐carotene. The micellar phase was collected, and β‐carotene extraction was performed on both the raw digest and the micellar phase using an ethanol–hexane mixture (volume ratio 1:2). As specified in Section 2.15.3, the concentration of β‐carotene was determined, and the effective bioaccessibility was computed as follows:

$$\mathrm{Bioaccessbility}\left(\%\right)=\frac{C_{\mathrm{micelle}}}{C_{\mathrm{digesta}}}\times 100\%$$

where C _micelle is the β‐carotene concentration in the micellar fraction collected after the intestinal phase, and C _digesta denotes the β‐carotene concentration in the total digest, with the two parameters defined, respectively.

2.18. Statistical Analyses

All measurements were performed in triplicate. Statistical analysis was conducted using one‐way analysis of variance (ANOVA) and Duncan's test in SPSS 23.0, with a significance level set at p < 0.05.

3. Results and Discussion

3.1. Appearance and CI

As shown in Figure 1A, all newly prepared O/W emulsions with pH values ranging from 5.0 to 9.0 exhibited a homogeneous, milk‐white appearance without any creaming or gravitational separation, and their CI was determined to be 0%. This indicated the presence of strong repulsive forces, likely electrostatic and Brownian forces, between the droplets, which effectively inhibited flocculation and coalescence (Yuliarti et al. 2019). After 14 days of storage, emulsions at pH 5.0 and 6.0 exhibited pronounced serum separation and sedimentation, whereas those at pH 7.0–9.0 maintained good homogeneity (Figure 1A). As illustrated in Figure 1C, the creaming indices for the pH 5.0 and 6.0 emulsions after storage were the highest, reaching 28.8% ± 0.3% and 21.8% ± 1.23%, respectively. This indicated a greater extent of droplet aggregation in these samples. The instability observed is presumably attributable to diminished electrostatic repulsion between oil droplets coated with protein near the isoelectric point, where the droplets exhibit minimal net charge density (Nooshkam and Varidi 2021). In contrast, the emulsion at pH 7.0 displayed the lowest CI (12.7% ± 0.63%), confirming its superior stability, which is consistent with its homogeneous visual appearance (Figure 1A). Similarly, φ also influenced the visual stability of the emulsions. As displayed in Figure 1B, all emulsions with oil fractions ranging from 10% to 30% were initially stable. However, after 14 days of storage, phase separation occurred in emulsions with lower oil fractions (10%–20%), whereas those with higher oil fractions (25%–30%) maintained good homogeneity. This visual assessment is corroborated by the quantitative CI data presented in Figure 1D, which showed a gradual decrease in the index with increasing oil phase fraction. Emulsions with φ ≥ 25% had the lowest creaming indices, showing that higher oil content improves storage stability. This finding is consistent with the results of Badar et al. (2024), who reported that emulsions prepared at an oil‐to‐dispersant ratio of 4:1 remained homogeneous without creaming or phase separation.

FIGURE 1.

Effects of pH and oil phase fraction on the appearance (A and B), CI (C and D), particle size (E and F), and particle size distribution (G and H) of WPI–AG‐stabilized O/W emulsions. Different lowercase letters indicate significant differences (p < 0.05). CI, creaming index.

3.2. Particle Size and Zeta Potential

Compared with WPI–AG emulsions at other pH values, those at pH 5.0–6.0 showed significantly larger particle sizes and a bimodal size distribution, which indicated the existence of larger oil droplets (Figure 1E,G, p < 0.05). It was probable that this phenomenon stems from the nearness to WPI's isoelectric point, a state in which weakened electrostatic repulsion promotes the aggregation of droplets. Consistent with this result, similar pH‐induced aggregation was also found in Mesona chinensis polysaccharide–WPI composite systems during acid–alkali shifting treatment (Jiang et al. 2021). It has been reported in prior studies that the experimental isoelectric point of WPI falls within the range of approximately 5.0–5.2 (Fan et al. 2017). This finding aligned with the work of Zhang et al. (2023), who documented that emulsions stabilized by WPI‐polysaccharide conjugates also displayed maximal particle size when the pH was set at 5.0. Notably, at pH 7.0–9.0, the emulsion droplet size was consistently below 10 µm, with a markedly narrower size distribution and a peak shift towards smaller diameters. This indicated enhanced electrostatic and steric stabilization within the system. Furthermore, as shown in Figure 2A, the absolute zeta‐potential value increased progressively with pH, reaching a maximum of −32.97 ± 0.76 mV at pH 7.0, which was significantly higher (p < 0.05) than at other pH levels. At pH 8.0 and 9.0, the absolute zeta potential decreased slightly, with no significant difference observed between these two pH values. This could be attributed to the complete deprotonation of functional groups, such as carboxyl groups, within the WPI–AG conjugate at pH 7.0. This resulted in a peak negative charge density on the molecular surface, which explains the significantly higher absolute zeta potential observed for the emulsion at this pH compared to other levels (p < 0.05).

FIGURE 2.

WPI–AG‐stabilized oil‐in‐water emulsion, (A and B) zeta potential; (C and D) centrifugal sedimentation rate; (E and F) stability coefficient. Different lowercase letters indicate significant differences (p < 0.05). CPR, centrifugal precipitation rate.

The effect of different oil phase fractions (φ) on particle size and zeta potential was presented in Figure 1F,H. With an increase in oil fraction, the droplet size of the WPI–AG emulsion increased markedly, rising from 1.55 ± 0.03 µm when φ = 10% to 13.5 ± 0.44 µm when φ = 30% (p < 0.05). Concurrently, the particle size distribution transformed from monomodal to bimodal (Figure 1F), implying that oil droplet aggregation took place at elevated oil fractions. The observed growth in droplet size was likely caused by a shortage of WPI–AG conjugate available for stabilizing all oil droplets when oil fractions were elevated, resulting in the coalescence of droplets and the generation of larger emulsion droplets (Z. Wang, Ma, et al. 2022). Badar et al. (2024) observed that increasing the diacylglycerol‐to‐soy protein isolate dispersion ratio from 1:4 to 4:1 led to a corresponding increase in emulsion droplet size, which reached its maximum at the 4:1 ratio. As illustrated in Figure 2B, notable differences in zeta potential were detected among the emulsion samples with oil phase fractions of 10%, 15%, 20%, 25%, and 30% (p < 0.05). The absolute zeta‐potential value increased progressively with φ, reaching a maximum of −32.8 ± 3.27 mV at φ = 25%. This trend could be attributed to the increased oil fraction promoting a higher surface coverage of oil droplets by WPI–AG particles, thereby enhancing electrostatic repulsion between them (Badar et al. 2024). Consequently, these results revealed that pH and oil phase fraction exerted a regulatory effect on the droplet size and distribution of the WPI–AG emulsion.

3.3. CPR and Stability Coefficient (R)

A significant negative correlation was observed between CPR and R values; lower CPR values corresponded to higher R values, indicating superior emulsion stability (Li et al. 2019). As shown in Figure 2C,E, the emulsion at pH 5.0 exhibited the highest CPR (57.9% ± 1.61%) and the lowest R value (0.31 ± 0.02), indicating severe phase separation after centrifugation and poor redispersibility, which corresponded to the lowest stability. In contrast, at pH 7.0 and 9.0, the CPR decreased to 11.8% ± 1.23% and 16.37% ± 1.04%, respectively, whereas the R value increased to 0.62 and 0.57 (Figure 2C,E). These results demonstrated that the WPI–AG conjugate formed a more robust emulsion structure with enhanced resistance to centrifugal separation at these pH levels. These phenomena were attributed to the regulation of electrostatic and steric stabilization by the interfacial structure of the WPI–AG conjugate. Under neutral and alkaline conditions, WPI, being distant from its isoelectric point, carried a strong negative charge. This helped establish efficient electrostatic repulsion and steric hindrance effects between WPI and AG, which, in turn, prevented the aggregation of oil droplets (Li et al. 2022). Furthermore, the dense composite film formed at the interface helped resist droplet coalescence induced by centrifugal force, thereby structurally enhancing the R value and reducing the CPR. As the oil fraction increased from 10% to 25%, the CPR gradually decreased to 9.97%, whereas the R value increased to 0.70 (Figure 2D,F). Upon further increasing the oil fraction to 30%, we observed a marginal rise in CPR along with a corresponding decrease in the R value. This suggested that an excessively high oil fraction might induce excessive droplet contact and aggregation, thereby compromising emulsion stability (Z. Wang, Ma, et al. 2022).

3.4. Confocal Laser Scanning Microscopy

As shown in Figure 3A, the complete coverage of oil droplets by the WPI–AG conjugate and its filling of the interstitial spaces between droplets collectively stabilized the emulsion by inhibiting coalescence. In contrast, at pH 5.0 and 6.0, protein aggregates were observed at the oil droplet interface, indicating lower emulsion stability. This instability arises because proteins tend to aggregate near their isoelectric point, thus impairing the structural integrity of the emulsion interface (Xiong et al. 2023). These observations were consistent with the particle size results. In contrast, the WPI–AG emulsions at pH 7.0–9.0 exhibited significantly less flocculation. Specifically, at pH 7.0 and 9.0, the interfacial layer was uniform and stable, forming a dense film with high surface coverage (Figure 3A). Figure 3B shows that as the oil fraction rose from 10% to 30%, the extent of flocculation diminished while the filling effect grew more prominent. At the lower oil fractions of 10% and 15%, the emulsion droplets were larger, which could be attributed to insufficient emulsifier coverage at the oil–water interface due to the low oil density, leading to depletion flocculation (Zhao et al. 2025). As the oil fraction increased, the droplet density progressively rose. At φ = 20%, the emulsion exhibited smaller droplet size, uniform dispersion, and well‐defined spherical droplets with regular shapes (Figure 3B). This indicated that the emulsifier concentration was sufficient to fully cover the dispersed phase and form a complete interfacial layer. At φ = 25%, the emulsion showed larger droplet sizes and partial aggregation, with no obvious deformation. This result indicated that WPI–AG conjugates adsorbed on droplet surfaces to form interfacial films and simultaneously built an entangled structure in the continuous phase to hinder oil droplet movement and prevent coalescence (Feng et al. 2021). At φ = 30%, oil droplets partially interconnected and densely packed, accompanied by extrusion and deformation. This creates a dense dispersed structure featuring bridging flocculation, which further gives rise to a weak gel‐like network (Figure 3B). Consistently, Feng et al. (2021) observed that emulsion gelation at φ ≥ 50% originated from bridging structures formed by shared particle layers between droplets.

FIGURE 3.

(A) CLSM images of WPI–AG emulsions with different pH values (5.0, 6.0, 7.0, 8.0, and 9.0), fixed particle concentration of 2.5%, and fixed oil phase of 20%; (B) CLSM images of WPI–AG emulsions with different oil phases (φ = 10%, 15%, 20%, 25%, and 30%), fixed particle concentration of 2.5%, and fixed pH 7.0.

3.5. Turbidity Titration

Turbidity was influenced not only by droplet size but also by droplet morphology and distribution. Furthermore, we found that there was a negative correlation between the turbidity of the emulsion and its stability (W. Wang, Wang, Yao, et al. 2022). The effects of pH and oil phase volume fraction on emulsion turbidity are presented in Figure 4A,B. The turbidity exhibited significant variations in response to changes in both pH and oil fraction. As shown in Figure 4A, the WPI–AG conjugate‐stabilized emulsion displayed higher turbidity within the pH range of 5.0–6.0, which was associated with protein self‐aggregation near the isoelectric point. Similar findings were reported by Zhang et al. (2020). The emulsion exhibited the lowest turbidity at pH 7.0, indicating a more stable system with improved dispersibility. At pH 8.0 and 9.0, turbidity increased, a phenomenon that suggested that although droplet aggregation was weakened with pH moving away from the isoelectric point, the overall emulsion stability was still moderately impaired (Figure 4A). As shown in Figure 4B, the oil phase volume fraction had a significant effect on turbidity (p < 0.05). The turbidity increased with rising oil fraction, a trend that aligned with the particle size results. No significant difference (p > 0.05) in turbidity was found for WPI–AG emulsions at φ ≤ 20%, whereas significant differences emerged at φ > 20% (p < 0.05). This suggested that at higher oil fractions, the reduced interdroplet spacing might increase the frequency of protein complex aggregation, consequently leading to elevated turbidity (Zhang et al. 2020).

FIGURE 4.

Turbidity (A and B), interfacial protein adsorption and interfacial tension (C and D), shear rate scan (E and F), and frequency scanning (G and H) of WPI–AG stabilized emulsion. Different lowercase letters indicate significant differences (p < 0.05). AP, adsorbed protein.

3.6. Interfacial Adsorption Characteristics

As shown in Figure 4C, pH significantly affected both the AP content and the interfacial tension of the WPI–AG conjugate‐stabilized emulsion (p < 0.05). At pH 5.0, the AP was significantly lower and the interfacial tension was relatively higher compared to other pH levels. This demonstrated that near the protein's isoelectric point, the diminished electrostatic repulsion between protein molecules impaired their mobility and diffusion toward the oil–water interface, thus lowering the adsorption efficiency (X. Zheng, Cheng, et al. 2025). When the pH was elevated to the neutral and alkaline range (pH 7.0–9.0), a significant reduction in interfacial tension was observed (p < 0.05). This suggested that the WPI–AG conjugate diffused and adsorbed more readily at the oil–water interface under these conditions. At neutral and alkaline pH, the increased negative charge on the protein molecules enhanced electrostatic repulsion, promoting molecular chain extension and greater exposure of hydrophobic groups. This consequently improved the adsorption capacity and interfacial activity at the droplet surface (Dong and Hua 2018; Li et al. 2022). Consistently, Dong and Hua (2018) observed that the adsorption behavior of protein–gum arabic complexes at the oil–water interface correlated with the exposure of hydrophobic groups on the complex surface, with elevated surface hydrophobicity facilitating interfacial adsorption.

As shown in Figure 4D, φ also significantly influenced the interfacial adsorption behavior and interfacial tension of the WPI–AG emulsion. As φ increased from 10% to 30%, the AP content exhibited an overall increasing trend, whereas the interfacial tension progressively decreased (p < 0.05). This demonstrated that a higher oil fraction increased the total interfacial area within the system, which required more protein to migrate and adsorb at the oil–water interface to form a stabilizing layer, thereby lowering the interfacial energy (L. Zheng, Chen, et al. 2025). When the oil fraction exceeded 20%, the increase in AP content plateaued, indicating that the system had reached a dynamic equilibrium for interfacial adsorption and the protein coverage at the interface approached saturation (Figure 4D). Furthermore, the combination of the highest AP value and the lowest interfacial tension at φ = 25% suggested the formation of sufficient and complete protein coverage at the oil–water interface, with proteins arranged in a relatively regular structure. This interfacial structure not only efficiently reduced the oil–water interfacial energy but also suppressed the aggregation of droplets through electrostatic and steric repulsive forces, thus enhancing the emulsion's physical stability (Kim et al. 2025). Higher AP content accompanied by lower interfacial tension reflected adequate protein coverage at the oil–water interface. Adequate interfacial protein coverage effectively weakened the interfacial incompatibility between oil and water, while the intact interfacial layer further provided sufficient electrostatic repulsion and steric hindrance (Kim et al. 2025). Such interfacial characteristics could effectively resist droplet collision and coalescence during storage, thereby establishing an inherent correlation between interfacial adsorption performance and the overall physical stability of emulsions (X. Zheng, Cheng, et al. 2025). A rational explanation for this phenomenon lay in the strengthened electrostatic and steric repulsion among oil droplets provided by the protein–polysaccharide conjugate, thereby enhancing the stability of the emulsion (L. Zheng, Chen, et al. 2025).

3.7. Rheological Properties

3.7.1. Viscosity

As illustrated in Figure 4E,F, the apparent viscosity of all emulsions declined as the shear rate rose, which is a hallmark of the intrinsic shear‐thinning property of pseudoplastic substances. This result also implies relatively weak intermolecular interactions between emulsion droplets, and these interactions facilitated the formation of a loose interconnected aggregate structure within the emulsion system (Zhao et al. 2025). Figure 4E demonstrated that the emulsion at pH 7.0 had a higher apparent viscosity than emulsions at all other pH values, a trend that was in agreement with the droplet size and CLSM analysis results. This could be explained by the fact that smaller droplets possess a larger total surface area, which helps strengthen interdroplet interactions within the emulsion system. Furthermore, the AG molecules in the conjugate likely act as bridges, connecting adjacent droplet membranes and thereby reducing coalescence while contributing to the increased viscosity (Y. Chen, Yi, et al. 2022). The elevated viscosity, in turn, restricts droplet mobility and reduces coalescence, thereby enhancing emulsion stability (Xiong et al. 2023). Furthermore, Zhou et al. (2022) reported that oil droplets exhibit greater interfacial elasticity at neutral pH. Concurrently, the emulsion viscosity increased markedly with rising oil fraction (Figure 4E,F). The primary cause of this phenomenon lies in the intrinsic viscosity of the oil phase and the preferential distribution of particles at the oil–water interface, which boosts the intensity of interdroplet interactions (Geng et al. 2024). Zhao et al. (2025) reported that higher oil concentrations usually induce an increase in the initial viscosity of emulsions, reflecting the augmentation of gel strength or hardness corresponding to the rise in oil fraction.

3.7.2. Frequency Scanning

As shown in Figure 4G,H, the G′ and G″ exhibited frequency dependence. The emulsions displayed viscoelastic characteristics, featuring both solid‐like elasticity and fluid‐like viscosity, a property that resulted from strong interdroplet interactions among dispersed droplets. As shown in Figure 4G,H, nearly all emulsions exhibited a higher G′ than G″, indicating elastic‐dominated viscoelastic behavior. That G′ was consistently dominant over G″ in all samples suggested that electrostatic and hydrophobic interactions during protein–polysaccharide complexation promoted strong interdroplet association and elastic behavioral characteristics. As shown in Figure 4G, the emulsion at pH 7.0 exhibited higher G′ and loss G″ moduli than those at other pH levels. This result could be attributed to the WPI–AG aggregate formed at pH 7.0 possessing the strongest surface hydrophobicity compared to those formed at other pH values. Figure 4H demonstrated that both the G′ and G″ moduli of the emulsion increased with an increase in the oil phase volume fraction. This finding revealed that higher oil fractions intensified interdroplet entanglement, improving the emulsion's stability under high‐frequency shear oscillation (Feng et al. 2021). This configuration enhanced resistance to deformation and restricted droplet mobility, consequently leading to higher viscosity and moduli (Dong et al. 2022). Dong et al. (2022) noted that raising the oil phase volume fraction in ovotransferrin–carboxymethyl chitosan nanoparticle‐stabilized Pickering emulsions caused an increase in both G′ and G″, consequently boosting the emulsion's overall viscosity and moduli.

3.8. Physical Stability

3.8.1. Ionic Strength Stability

The presence of salt ions could significantly influence emulsion stability. The ion concentration impacted the droplet size and potential of β‐carotene emulsions stabilized by WPI–AG complexes under different ultrasonic power treatments, as illustrated in Figure 5A,B. The addition of NaCl caused a sharp increase in the average droplet size and a significant reduction in the net charge of droplets in the WPI–AG‐stabilized emulsion. This occurred because the charged salt ions screened the electrostatic repulsion between droplets, disrupting the balance between attractive forces (e.g., van der Waals and hydrophobic interactions) and repulsive forces, thereby inducing droplet aggregation (Li et al. 2022). A prior study has revealed that increasing ionic concentrations cause larger particle sizes and lower net charge in β‐carotene emulsions stabilized by complexes of arabinoxylan hydrolysate and soy protein isolate (Li et al. 2022). Furthermore, notable variations in salt tolerance were detected among samples treated with varying ultrasonic power levels, with statistical significance (p < 0.05). In the WPI–AG conjugate emulsion, droplet size initially declined with increasing ultrasonic power but subsequently rose, with the D _4,3 value reaching its lowest point at 400 W. This outcome was attributed to the shear forces generated by ultrasonication, which reduced the aggregation of WPI–AG conjugates in the emulsion. This contributed to more extensive adsorption of the conjugates at the oil–water interface, thus improving the emulsion's salt ion resistance (T. Wang, Wang, Li, et al. 2022). However, when ultrasonic power was increased further, it only induced a minor rise in the D _4,3 value and a decrease in the absolute zeta potential. This phenomenon was probably attributable to the fact that excessive ultrasonic power triggered the aggregation and coalescence of emulsion droplets, consequently reducing the stability of the system (T. Wang, Wang, Li, et al. 2022). Furthermore, the droplet size of the WPI–AG emulsion increased progressively with salt concentration, indicating reduced stability. This phenomenon could be accounted for by the shielding effect of introduced sodium ions on the electrostatic repulsion between negatively charged droplets, which facilitates the aggregation of the WPI–AG conjugate and leads to a reduction in emulsion stability (Wang et al. 2024). Notably, at a given salt concentration, the absolute zeta‐potential value initially increased and then decreased with rising ultrasonic power, reaching its maximum at 400 W. The increased absolute zeta potential enhanced electrostatic repulsion between proteins, thereby suppressing aggregation and consequently improving emulsion stability (Wang, Wang, et al. 2023). Consequently, these results demonstrated that appropriate ultrasonic processing is capable of enhancing the salt‐ion stability of the emulsion system.

FIGURE 5.

Effects of ionic strength (A and B), heat treatment (C and D), and pH (E and F) on the droplet size and zeta potential of β‐carotene emulsion stabilized by WPI–AG complex treated with different ultrasonic powers. Different superscript letters indicate significant differences (p < 0.05).

3.8.2. Thermal Stability

This study sought to clarify the impacts of varying ultrasonic power levels and treatment temperatures on WPI–AG emulsion stability by comparing two thermal treatment regimes (65°C and 90°C) and determining their influences on the D _4,3 and zeta potential of β‐carotene‐loaded emulsions, thereby evaluating the emulsions’ physical stability. As shown in Figure 5C,D, the D_{4 ,3} value of the WPI–AG conjugate emulsion increased with rising temperature, whereas the net charge decreased. This decline in stability at elevated temperatures was likely due to heat‐induced droplet aggregation (Wang, Wang et al. 2023). Furthermore, heat treatment might induce the denaturation of proteins anchored at the oil–water interface, leading to a loss of surface charge. This reduction in charge weakened electrostatic repulsion, thereby allowing van der Waals attractive forces to become dominant between oil droplets (Li et al. 2022). In addition, as the ultrasonic power increased, the D _4,3 value of the emulsion declined continuously and hit the minimum at an ultrasonic power level of 400 W. This suggested that the shear forces generated by ultrasonication fragmented the WPI–AG conjugate into smaller particles, exposing more charged groups. The resulting enhancement of electrostatic repulsion between droplets prevented aggregation, thereby improving emulsion stability. However, a further increase in ultrasonic power led to an increase in particle size, a phenomenon attributed to excessive power inducing droplet coalescence and consequently impairing the thermal stability of the WPI–AG emulsion (W. Wang, Wang, Yao, et al. 2022). This finding aligned with the research by Ye et al. (2025), who demonstrated that moderate ultrasonication improves the size uniformity and zeta‐potential stability of protein‐based emulsion systems, consequently enhancing their colloidal stability under thermal conditions.

3.8.3. pH Stability

As observed in Figure 5E,F, the WPI–AG emulsion exhibited a maximum D _4,3 value and a minimum absolute zeta potential at pH 5.0. However, further elevating the ultrasonic power resulted in larger particle sizes, which was ascribed to the fact that excessive power triggered droplet coalescence and thus reduced the thermal stability of the WPI–AG emulsion. Zhang et al. found that emulsions stabilized by WPI–polysaccharide complexes exhibited maximum particle size at pH 5.0. They attributed this phenomenon to the aggregation of WPI near its isoelectric point, where the weakened electrostatic repulsion was insufficient to counteract droplet flocculation (Zhang et al. 2023). When the pH deviated from WPI's isoelectric point, the droplet size of the emulsion showed a decreasing trend. Moreover, ultrasonic treatment resulted in a significant reduction in droplet size relative to the untreated control, with the minimum droplet size achieved at 400 W. These results indicated that moderate ultrasonication improved the dispersibility and pH stability of the emulsion (M. Wang, Ma, et al. 2025). Under neutral and alkaline conditions, all emulsions exhibited zeta potentials below 32 mV, indicating good stability. Consistently, Li et al. (2022) reported that emulsions prepared with arabinoxylan hydrolysates–soy protein isolate complexes also showed zeta potentials below 15 mV under these pH conditions, demonstrating their optimal stability.

3.9. Chemical Stability

3.9.1. Storage Stability

The storage stability of β‐carotene in the emulsions over 8 days at 4°C, 25°C, and 55°C was shown in Figure 6A–C. β‐Carotene underwent varying degrees of degradation in the WPI–AG emulsions under all storage conditions. Compared to the untreated control, ultrasonic treatment significantly enhanced storage stability. As shown in Figure 6B,C, the most substantial degradation occurred in the untreated group at 25°C and 55°C, whereas the groups treated at 400 and 500 W maintained higher retention rates throughout the storage period, demonstrating effective protection. Notably, under the 55°C storage condition, the emulsion treated at 400 W maintained a higher β‐carotene retention rate than the other groups after 8 days, indicating superior thermal stability of its encapsulation structure. This trend could be attributed to moderate ultrasonication optimizing the interfacial layer structure and enhancing droplet stability, thereby decelerating the oxidation and isomerization of β‐carotene (Teng et al. 2020).

FIGURE 6.

Storage stability (A–C) and ultraviolet stability (D) of WPI–Ag‐stabilized β‐carotene emulsions treated with different ultrasonic powers, as well as shear rate scanning (E) and frequency scanning (F).

3.9.2. Ultraviolet Light Stability

Figure 6D presented the photostability of emulsions stabilized by WPI–AG conjugates that were treated with varying ultrasonic powers. The β‐carotene content in all emulsions decreased progressively with prolonged UV exposure, with a faster degradation rate observed during the initial 2 h. As shown in Figure 6D, the untreated control exhibited the poorest photostability, showing a significantly lower retention rate after 8 h compared to the sonicated groups. In contrast, the emulsions treated at 400 and 500 W effectively delayed β‐carotene degradation, maintaining consistently higher retention rates. This phenomenon could be explained by moderate ultrasonic processing improving the overall interfacial structure, which better impeded UV light and oxygen, thereby inhibiting oxidative degradation to a greater extent (Teng et al. 2020). A previous study demonstrated that in ultrasonicated emulsion systems, the retention rate of carotenoids remained above 86% across different temperature conditions (Teng et al. 2020). This finding provided strong support for the conclusion that ultrasonic treatment enhanced the structural stability of emulsions.

3.10. Rheological Properties

An emulsion's long‐term stability was largely governed by its interfacial viscosity. In general, elevated viscosity restricted droplet migration, which, in turn, enhanced the emulsion's stability (Ma et al. 2020). As depicted in Figure 6E,F, all WPI–AG emulsions displayed a reduction in apparent viscosity as the shear rate increased, demonstrating a shear‐thinning characteristic. This phenomenon stemmed from conformational alterations of proteins adsorbed at the surface of oil droplets under shear stress. Specifically, the hydrophobic moieties of the proteins tended to migrate toward the oil phase, thus undermining the electrostatic and hydrogen‐bonding interactions between adjacent droplets (H. Wang, Yang, et al. 2025). The restructured interfacial film reduced resistance to droplet movement, facilitating system flow and consequently lowering the apparent viscosity. Furthermore, with rising ultrasonic power, the apparent viscosity of the emulsion first increased and subsequently decreased (Figure 6E). The maximum apparent viscosity was observed at 400 W, indicating that moderate ultrasonication promoted the partial unfolding of WPI and the exposure of its hydrophobic groups, which favored effective combination with AG at the oil–water interface. This interfacial layer not only provided substantial charge stabilization at the oil droplet surface but also generated additional steric hindrance through the long‐chain structure of AG, thereby significantly enhancing the viscosity and flow stability of the emulsion (Huang et al. 2024). Consistently, Wang, Wang et al. (2023) reported that the apparent viscosity of O/W emulsions stabilized by rice bran protein combined with tannic acid hit a maximum at 400 W of ultrasonic power.

As shown in Figure 6F, all samples except the untreated control exhibited a G′ greater than the G″ during the frequency sweep, indicating elasticity‐dominated viscoelastic behavior. Notably, both the G′ and G″ moduli generally increased with higher ultrasonic power, indicating strengthened interdroplet interactions. In particular, the 400 W treatment consistently exhibited higher moduli across the entire frequency range, demonstrating that moderate ultrasonication effectively improved interdroplet interactions and viscoelastic properties (Huang et al. 2024). Wang, Wang et al. (2023) illustrated that ultrasonic treatment effectively improves the network architecture and viscoelastic characteristics of emulsions stabilized by rice bran protein‐tannic acid complexes, consistent with the findings of our research.

3.11. In Vitro Digestion

3.11.1. Particle Size and Potential

The variations in particle size across different stages of in vitro digestion were presented in Figure 7A,B. All emulsions exhibited a similar overall trend. Relative to the initial stage, there was a general increase in particle size post‐simulated gastric digestion. The underlying reason was that the complex gastric environment induces significant changes in protein structures. Specifically, the interfacial architecture was partially disrupted under the low pH and pepsin activity of gastric fluid, leading to droplet aggregation or coalescence (Winuprasith et al. 2018). During the intestinal phase, the particle size decreased compared to the gastric phase. This suggests that the synergistic effect of bile salts and lipases in the small intestine may have induced the redispersion of lipid droplets, generating smaller droplets or forming mixed micelles, thereby enhancing the stability of the system (Hu et al. 2021; Winuprasith et al. 2018). It was worth emphasizing that across all digestion phases, the mean particle size of the non‐sonicated group was markedly larger than that of the 400 W treatment group, which displayed the smallest particle size. This suggested that proper ultrasonic cavitation can efficiently modify protein structures, separate protein subunits, and inhibit aggregate formation (Li et al. 2023).

FIGURE 7.

(A) particle size of β‐carotene embedded in WPI–AG composite emulsion treated with different ultrasonic powers during in vitro digestion; (B) zeta potential; (C) CLSM diagram; (D) release rate; and (E) biological accessibility. Different superscript letters indicate significant differences (p < 0.05).

Figure 7B illustrated the changes in zeta potential across different digestive phases. In the initial phase, all samples exhibited high absolute zeta‐potential values (ranging from −35 to −50 mV), indicating a high degree of electrostatic stabilization. During the gastric phase, the gastric environment's complex composition—marked by low pH and high ionic strength—diminished the absolute zeta‐potential values. The pepsin contained in gastric fluid might hydrolyze the surface proteins of the conjugates, thereby bringing about changes in the interfacial composition and charge characteristics. The concomitant reduction in electrostatic repulsion could consequently lead to droplet aggregation (Winuprasith et al. 2018). During the intestinal phase, the absolute zeta potential increased, which could be attributed to the predominance of various anionic substances (e.g., free fatty acids and bile salts) in the digestive fluid (Hu et al. 2021; Li et al. 2022). On the whole, ultrasonic processing led to the exposure of additional negatively charged groups, which, in turn, elevated the effective surface charge of the proteins. This strengthened surface charge enhances electrostatic repulsion between emulsion droplets, a change that was conducive to droplet dispersion and increased the interfacial layer's susceptibility to digestive enzymes (Zhu et al. 2023).

3.11.2. Confocal Laser Scanning Microscopy

Microstructural changes in the β‐carotene emulsion during digestion were shown in Figure 7C. Upon entering the gastric phase, the droplet size increased. The most pronounced increase in droplet size occurred in the untreated control, which could be attributed to pepsin‐mediated proteolysis of the WPI–AG interfacial layer, thereby disrupting the emulsion droplets and causing oil leakage. In contrast, the emulsions treated with moderate ultrasonication maintained a relatively uniform particle size distribution. This phenomenon likely stemmed from ultrasound‐induced partial unfolding of WPI and AG molecular chains and enhanced cross‐linking, which improved the structural stability of the interfacial layer and its tolerance to the low‐pH environment (Han et al. 2022; Wang, Wang et al. 2023). Upon entering the intestinal phase, the droplet size decreased, which aligned with the particle size measurements. This reduction could be mainly ascribed to the enzymatic action of pancreatin, which decomposed the large droplets formed during the gastric phase into smaller micelles stabilized by bile salts (Han et al. 2022). At this stage, the emulsion treated at 400 W maintained a relatively uniform fluorescence distribution, indicating structural stability even during the later phases of digestion, which was conducive to the controlled release of β‐carotene. In contrast, the untreated control contained relatively large particles, likely originating from undigested protein aggregates.

3.11.3. Cumulative Release Rate

The cumulative release profile of β‐carotene from emulsions prepared with different ultrasonic power levels during simulated gastrointestinal tract (GIT) digestion was presented in Figure 7D. During the gastric phase, the WPI–AG emulsion exhibited a gradual rise in the release rate of β‐carotene. However, in comparison with the non‐sonicated group, the ultrasonically treated emulsions released a lower amount of β‐carotene during the gastric phase. These results indicated that ultrasonic treatment of the emulsions could enhance the protective effect on β‐carotene during gastric digestion. In the intestinal phase, the in vitro release rate of β‐carotene increased rapidly within 30–90 min of digestion, after which the release rate reached a plateau. The absence of significant differences in the release rates among the emulsions at 90 min suggested that β‐carotene was fully released in the intestinal phase. Among the ultrasonicated groups, the emulsion treated at 400 W exhibited a higher β‐carotene release rate than the other treatment levels. This result indicated that emulsions prepared with an appropriate ultrasonic power could effectively encapsulate β‐carotene and modulate its release rate, thereby helping to preserve its bioactivity and enhance its absorption (Wang et al. 2024).

3.11.4. Bioaccessibility

Figure 7E showed that ultrasonic power had a significant effect on the bioaccessibility of β‐carotene (p < 0.05). The untreated control (0 W) and the low‐power treatment (300 W) exhibited lower bioaccessibility, at 49.55% ± 4.31% and 58.63% ± 2.44%, respectively. This suggested that without or with only mild ultrasonication, the emulsion structure formed by the WPI–AG conjugate is relatively loose with weak interfacial adsorption, making it susceptible to disruption in the digestive environment and consequently limiting the release and micellization of the lipophilic active compound (N. Wang et al. 2023). In contrast, after treatment at 400 W, the emulsion exhibited a significantly higher bioaccessibility of 81.98% ± 2.1%. This suggested that moderate ultrasonic treatment facilitates the development of a compact and stable emulsion architecture, a structure that boosts both the β‐carotene encapsulation efficiency and its release efficiency in the intestinal phase (Wang et al. 2024). The underlying reason for this phenomenon was probably the cavitation effects induced by ultrasonication, which were critical to facilitating the interfacial assembly of protein–polysaccharide complexes. This process facilitated the formation of a more homogeneous, small‐droplet structure and enhanced the formation of mixed micelles with bile salts under intestinal conditions, thereby promoting the solubilization and absorption of β‐carotene (N. Wang et al. 2023). However, treatment at the higher power of 600 W resulted in a decrease in bioaccessibility to 62.32% ± 3.3%. This decline was likely due to excessive ultrasonication inducing protein aggregation or disrupting the interfacial film, thereby compromising the structural stability and delivery performance of the emulsion (Wang et al. 2024). Li et al. (2024) reported that among all samples, emulsions prepared with moderate ultrasonic treatment exhibited the highest retention rate of β‐carotene, which was attributed to their superior encapsulation efficiency. Zhang et al. (2021) found that during in vitro simulated digestion, moderate treatment markedly improved the chemical stability and bioaccessibility of β‐carotene in emulsion gels.

4. Conclusion

This study prepared emulsions stabilized by WPI–AG conjugates and systematically characterized their stability under different pH conditions and oil phase volume fractions. The results demonstrated that both pH and oil fraction play crucial roles in modulating emulsion stability. The WPI–AG emulsion achieved optimal stability at pH 7.0 and φ = 25%, with the maximum absolute zeta potential, minimal interfacial tension, and favorable viscoelasticity (Figure 8). We further constructed β‐carotene‐loaded emulsions with ultrasonically modified WPI–AG conjugates and explored the effect of ultrasonic power. Ultrasonic treatment effectively improved the physicochemical stability of the emulsions. In vitro digestion experiments showed that emulsions prepared at 400 W had the highest β‐carotene bioaccessibility (81.98% ± 2.1%). Accordingly, ultrasonic‐treated WPI–AG emulsions can serve as promising delivery vehicles for lipophilic bioactives. This work still has some limitations. Future research will focus on the long‐term physical stability of emulsions and additional comparative in vitro digestion tests. We will further explore the performance of this emulsion system in actual food systems and its stability under pasteurization, spray‐drying, and freeze–thaw treatment. Moreover, in vivo trials will be carried out to verify the nutrient bioaccessibility and confirm the application potential of WPI–AG conjugate emulsions as oral carriers of lipophilic bioactive ingredients.

FIGURE 8.

Diagram of the influence mechanism of β‐carotene‐supported O/W emulsion (with WPI–AG as the emulsifier).

Author Contributions

Tingting Gao: writing – original draft, conceptualization, methodology, software, data curation. Benxiu Xiao: data curation, investigation, validation. Guibin Zheng: formal analysis, supervision, visualization. Jingwen Chen: visualization, project administration, resources, data curation. Aohan Feng: methodology. Qiankun Liu: methodology, validation, visualization. Shaoxiao Zeng: writing – review and editing, investigation, methodology, software. Baodong Zheng: formal analysis, data curation, resources. Haoran Chen: writing – review and editing, writing – original draft, validation, software, data curation.

Conflicts of Interest

The authors declare no conflicts of interest.