Ultrasonic microreactor-mediated fabrication of stable W1/O/W2 double emulsions for efficient vitamin C encapsulation

TL;DR

Researchers used an ultrasonic microreactor to create stable double emulsions for encapsulating vitamin C and E, achieving high efficiency and stability for potential use in cosmetics.

Contribution

The novel use of an ultrasonic microreactor enables continuous fabrication of stable W1/O/W2 emulsions with high encapsulation efficiency.

Findings

The fabricated emulsions had uniform droplet sizes (∼1.1–2.5 μm) and excellent physical stability.

Encapsulation efficiency of vitamin C reached up to 80% with minimal degradation during storage.

The emulsions showed nearly 100% radical-scavenging activity, indicating strong antioxidant properties.

Abstract

Water-in-oil-in-water (W1/O/W2) emulsions are promising carriers for the encapsulation and controlled release of bioactive compounds. However, achieving long-term stability of W1/O/W2 emulsions remains a significant challenge. In this study, an ultrasonic microreactor (USMR) was utilized for the fabrication of stable W1/O/W2 emulsions, achieving the simultaneous encapsulation of vitamin C (VC) in the internal W1 phase and vitamin E (VE) in the intermediate oil phase. The precisely controlled acoustic cavitation within the USMR facilitates the uniform fragmentation of the primary W1/O droplets to the nanoscale (∼100 nm), which provides a robust foundation for the subsequent encapsulation into the micron-sized W1/O/W2 system. The resulting W1/O/W2 emulsions displayed a uniform micrometer-scale droplet size, with an average diameter of ∼ 1.1–2.5 μm, depending on ultrasonic power and flow rate. Furthermore, the addition of konjac glucomannan (KGM) induced a weak gel-like network of the inner aqueous phase, providing additional stabilization. The encapsulation efficiency of VC reached up to 80%. The double emulsions exhibited excellent physical stability, showing negligible phase separation after centrifugation for 20 min, and minimal droplet size variation during accelerated thermal storage at 60°C for 12 h. Long-term storage tests further demonstrated stable droplet size and morphology for up to 35 days at room temperature. By encapsulating VE within the oil phase to protect VC, the freshly prepared emulsions exhibited nearly 100% DPPH and ABTS radical-scavenging activity. Overall, the USMR provides an efficient and controllable approach for continuously preparing highly stable double emulsions, promising for cosmetics and related functional delivery systems.

Figures

c](#f0010), increasing PGPR from 1% to 2% sharply reduced interfacial tension from 28.0 mN/m to 2.51 mN/m, with a gradual decline to 1.09 mN/m at 10%, indicating interfacial saturation. Correspondingly, droplet size decreased with rising PGPR content ([Fig. 2b, d](#f0010)). At 1% PGPR, high interfacial tension (∼28 mN/m) led to milky emulsions with large droplets (∼797 nm) and visible coalescence. At 3–5%, tension dropped to 2.51–1.23 mN/m, droplet size fell to 118–262 nm, and emulsions turned semi-transparent blue, indicating a shift from Mie to Rayleigh scattering. At 7–10%, tension reached

a](#f0015), the addition of KGM significantly altered neither the primary W_1_/O droplet (105–110 nm) nor the droplet size distribution. [Fig. 3b](#f0015) shows that the emulsions containing KGM exhibited the same transparent blue appearance as the control, indicating that the droplet size remained at the nanoscale. However, the concentration of KGM affects the rheological properties of the emulsions. The flow curves (shear stress vs. shear rate) in [Fig. 3c](#f0015) indicate that the emulsions with KGM exhibit significant shear thinning behavior, that is, the shear stress and shear rate have

c](#f0025), the droplet size analysis revealed that increasing the W_1_/O nanoemulsions proportion led to a gradual enlargement of droplet size. However, all emulsions remained macroscopically stable without phase separation after 12 h of storage at 60°C, indicating that the ultrasonic microreactor effectively maintained the emulsions’ macroscopic stability. Although high-temperature accelerated aging slightly reduced droplet size, all emulsions maintained good internal stability across different ratios. These findings provide valuable experimental evidence for optimizing the formulation param

b](#f0030)), indicating that VC did not compromise droplet morphology or stability. Simultaneously, Interfacial tension also showed negligible variation (1.232–1.247 mN/m vs. 1.245 mN/m for the blank). CLSM images (Fig. S9) further confirmed that the VC‑loaded emulsions retained the same well-defined W_1_/O/W_2_ structure as the blank, demonstrating excellent compatibility. The antioxidative performance of the W_1_/O/W_2_ emulsions was characterized by scavenging efficiencies of DPPH and ABTS radicals, which assess the total radical-scavenging capacity of the multiphase system. To boost antiox

a](#f0035)). This minor size reduction in acidic media is likely not due to the chemical degradation of KGM, which is known for its excellent acid resistance, but rather to a subtle physical contraction of the weak gel-like network. Under low pH, the high proton concentration can interfere with the hydrogen-bonded junction zones, while the increased ionic strength triggers a transmembrane osmotic pressure gradient, leading to localized water migration and minor droplet refinement. Regarding encapsulation efficiency (EE), the emulsions exhibited superior protective capacity under acidic conditi

Full text

Introduction

1

Vitamin C (ascorbic acid, VC) is widely recognized for its antioxidant properties, its role in promoting collagen synthesis, and its ability to brighten and even skin tone. Owing to these remarkable biological activities, VC is frequently incorporated into cosmetic formulations and pharmaceutical products [1], [2]. However, its inherent instability—particularly in aqueous environments—significantly limits its practical applications. When exposed to light, oxygen, or elevated temperatures, VC readily undergoes oxidative degradation, leading to a rapid decline in its efficacy in various formulations. Therefore, the development of stable and efficient encapsulation and delivery systems for VC is essential to maximize its potential benefits in cosmetic, pharmaceutical, and food applications. In recent years, with the advancement of novel controlled-release technologies and micro/nano-carrier systems in the pharmaceutical, food, and cosmetic industries. Although gelation of the internal aqueous phase is a pivotal strategy for enhancing the stability of nanoparticles, and numerous efforts have been extensively explored to improve vitamin C (VC) stability, these approaches still exhibit notable limitations. For instance, while recent lecithin-cholesterol composite liposomes have shown promise in co-encapsulation, they remain prone to vesicle fusion and compound leakage during long-term storage [3]. Similarly, electrosprayed chitosan nanoparticles demonstrate enhanced thermal stability via polymer matrix shielding, yet their fabrication often involves complex, solvent-intensive processing that complicates industrial scalability [4]. Furthermore, there is a growing interest in developing synergistic co-encapsulation systems of VC and vitamin E (VE) to maximize antioxidant efficacy [5], which is often challenging to achieve efficiently within single-phase carriers. Compared to other encapsulation technologies, W_1_/O/W_2_ double emulsions provide a robust oil barrier with a tunable thickness that physically isolates VC from the external environment. Furthermore, this intermediate oil phase can be tailored or functionalized (e.g., by incorporating lipophilic antioxidants) to significantly enhance its protective efficacy [6]. Specifically, when lipophilic VE is enriched at the oil–water interface, it acts as the “first line of defense” by scavenging reactive oxygen species (ROS) through hydrogen atom donation [7], thereby reducing the flux of radicals penetrating the inner aqueous phase. Moreover, the integrated interfacial layer formed by emulsifiers and antioxidants serves as a physical barrier that significantly inhibits oxygen diffusion, providing a “sacrificial protection” for the encapsulated VC [8]. Consequently, double emulsions have attracted increasing research interest as unique multiphase dispersion systems [9].

Typically structured as water-in-oil-in-water (W_1_/O/W_2_) or oil-in-water-in-oil (O_1_/W/O_2_) systems, double emulsions feature hierarchical inner–outer droplet configurations that enable multi-stage release profiles and enhanced stabilization of encapsulated active compounds [10]. Hu et al. reported that W/O/W double emulsions, followed by freeze-drying to form microcapsules, extended the half-life of encapsulated VC and VE by 3.35 and 1.56 times, respectively [6]. Akhtar et al. demonstrated that encapsulating VC in double emulsions improved its topical performance over a 4-week skin study [11]. Similarly, Farahmand et al. confirmed through release studies that multiple emulsion structures enable sustained and stable release of VC [12]. Collectively, these studies highlight the great potential of double emulsions as carriers for VC encapsulation and delivery. Gelation of the internal aqueous phase is a pivotal strategy for enhancing the stability of double emulsions, with various gelling agents available, such as pectin, sodium alginate, and xanthan gum. Although ionic polysaccharides like sodium alginate and pectin are frequently employed to construct gel networks, they exhibit high environmental sensitivity. Fluctuations in pH or changes in ionic strength often disrupt their electrostatic cross-linking structures, leading to the failure of the protective barrier [13], [14]. In contrast, Konjac Glucomannan (KGM) was selected for this study as a non-ionic, neutral polysaccharide, which exhibits enhanced stability across a broad range of pH and ionic strengths compared to many charged hydrocolloids. Furthermore, KGM facilitates the formation of a viscoelastic physical entanglement network through extensive hydrogen bonding, which serves as an effective mechanical barrier within the aqueous phase. This robust structure provides a more effective shield against thermal degradation and oxygen permeation, thereby overcoming the instability inherent in traditional ionic hydrocolloids under complex processing environments.

However, conventional methods for preparing double emulsions—such as high-speed homogenization or high-pressure emulsification—often result in broad droplet size distributions and unstable interfacial structures, leading to low encapsulation efficiency and poor long-term stability. These drawbacks have limited their scalability and application in industrial formulations [15], [16]. Therefore, developing a novel preparation strategy for stable and controllable double emulsions is crucial to overcoming these challenges and advancing their practical use in product development. In contrast to conventional ultrasonic techniques, such as horn-type and bath-type, the ultrasonic microreactor (USMR) intensified the collapse of cavitation bubbles, a process where acoustic-induced bubbles grow and undergo violent implosion within the liquid, generating localized high-energy micro-explosions. These cavitation events produce intense shock waves and micro-jets that effectively fragment the oil phase into smaller droplets with significantly lower polydispersity indices (PDI) than those achievable by low-energy methods. This high-energy mechanism not only optimizes droplet size and distribution but also increases the interfacial area, thereby reducing the emulsifier concentration required to maintain long-term stability [17].

Moreover, the USMR represents a paradigm shift by merging the high energy density of power ultrasound with the precise control of microfluidics. Unlike conventional horn-type ultrasonic systems, where energy intensity decays exponentially with distance from the probe, leading to 'dead zones' and non-uniform droplet sizes [18]. The USMR employs a standing-wave resonance-based structure. Recent literature has emphasized that the 'confined-space effect' in USMR significantly enhances cavitation efficiency [19]. In a micro-scale environment (ID 4 mm), the collapse of cavitation bubbles generates localized micro-jets and intense shear forces more effectively than in bulk reactors [19], [20]. This configuration allows for rapid and highly efficient emulsification that surpasses conventional ultrasonic systems. The USMR generates a uniform cavitation field due to its resonance-based design. Recent studies have demonstrated that such micro-confined environments facilitate a more homogeneous energy distribution than conventional horn-type transducers, thereby eliminating localized overheating and inconsistent droplet sizes. Furthermore, the USMR excels in controlled internal droplet fragmentation [21], [22]. The intense localized shear forces generated by collapsing cavitation bubbles in the 4 mm channel allow for the effective reduction of W_1_/O primary emulsion droplets to the nanometer range. This precise energy control ensures the preservation of the external W_1_/O/W_2_ interface while maximizing encapsulation efficiency, which is a task that remains challenging for high-pressure homogenization or high-shear mixing. For instance, Xu et al. demonstrated that stable emulsions can be continuously produced using a USMR without the need for a pre-emulsification step [17], [23], [24]. The USMR system can be fabricated in multiple designs and throughput scales, offering excellent adaptability, flexibility, and ease of fabrication. Furthermore, its operational parameters—such as ultrasonic power and residence time—can be precisely controlled, enabling fine-tuned emulsification performance. In recent years, USMR systems have gained increasing attention for their ability to produce monodisperse droplets and significantly enhance the encapsulation efficiency of active compounds.

In this study, a high-throughput continuous-flow USMR was employed to fabricate stable W_1_/O/W_2_ double emulsions encapsulating VC, integrating precisely controlled ultrasonic cavitation with a strategy based on the formation of a weak gel-like network within the internal phase. Specifically, konjac glucomannan (KGM) and glycerol were introduced into the internal aqueous phase to establish a viscoelastic physical entanglement network, while VE was incorporated into the oil phase to provide synergistic antioxidant protection. The study comprehensively investigated the effects of critical formulation variables—including the emulsifier concentration of polyglycerol polyricinolate (PGPR), KGM dosage, and the internal aqueous phase ratio—alongside processing parameters such as flow rate and ultrasonic power. This work aims to validate the double-core structure via gel-assisted transmission electron microscopy (TEM) and confocal laser scanning microscopy (CLSM), ultimately developing a robust co-delivery system with enhanced bioavailability and antioxidant activity in acidic environments.

Materials and methods

2

Materials

2.1

PGPR, caprylic/capric triglyceride (GTCC), polysorbate 80 (Tween 80), VC, glycerol, VE, 1,1-diphenyl-2-picrylhydrazyl (DPPH), and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS) stock solutions were all purchased from Macklin Biochemical Co., Ltd. (Shanghai, China). KGM was purchased from Yingxin Laboratory Equipment Co., Ltd. (Shanghai, China). Nile Red was purchased from Mairuier Chemical Technology Co., Ltd. (Shanghai, China). Glyceryl monostearate (GMS) was provided by Maya Reagent Co., Ltd. (Jiaxing, China). For all experiments, deionized water with an electrical conductivity of 1.8 μS/cm was prepared by a pure water machine, provided by Super-Genie (Shanghai Rephile Bioscience Co., Ltd.).

Preparation of W1/O nanoemulsions

2.2

The preparation processes of W_1_/O nanoemulsions and W_1_/O/W_2_ emulsions, and the synthesis principle within the USMR are shown in Fig. 1. The USMR employed in this study was provided by MoGe um-Flow Technology Co., Ltd. (Shantou, China). The microreactor operates at an ultrasonic frequency of 20 kHz. A schematic diagram of the double emulsions fabrication process is presented in Fig. 1. Featuring a metal 3D-printed aluminum alloy transducer, the microreactor contains four flow channels (4 mm inner diameter). With an overall effective length of 1080 mm, the device maintains a total internal volume of approximately 13.6 mL. The flow path consists of four linear segments that are interconnected via connectors and flexible silicone tubing. The residence time (τ) within the USMR was calculated as the ratio of the microreactor volume to the total volumetric flow rate. For instance, at a total flow rate of 13.6 mL/min, the corresponding residence time was 60 s.Fig. 1(a) Schematic illustration of the preparation processes of W_1_/O nanoemulsions and W_1_/O/W_2_ double emulsions; (b) The synthesis principle within the USMR.

Ultrasonic irradiation was supplied by an integrated piezoelectric generator. When the generator amplitude was adjusted within the 0–25% range, the ultrasonic powers of the USMR measured by calorimetry ranged from 0 to 40 W. According to our previous calibration studies [21], the maximum ultrasonic power actually density delivered into the reaction channel was determined to be 2.18 W/mL (see Supporting Information − S4. Calorimetric method), reflecting the effective acoustic energy acting on the flowing emulsions system. During the preparation of the W_1_/O nanoemulsions, the temperature was maintained at 50°C. For the subsequent formation of the W_1_/O/W_2_ double emulsions, the process temperature was also controlled at 50°C.

A typical run: First, 2.5 g of PGPR and 0.4 g of VE were dispersed in 37.1 g of GTCC oil and stirred at room temperature for 30 min to obtain the oil phase (O phase). Separately, 0.1 g of KGM and 0.06 g of NaCl were dissolved in 6 g of deionized water, followed by the addition of 3.75 g of glycerol and 0.1 g of VC. The mixture was covered with aluminum foil to protect it from light and stirred at room temperature for 30 min to obtain the inner aqueous phase (W_1_ phase). The W_1_ phase was then slowly added dropwise into the O phase and pre-emulsified using a high-speed homogenizer (Fluko FA30, provided by FLUKO Shanghai Equipment Co., Ltd.) at 7000 rpm for 3 min to form pre-emulsions. The resulting pre-emulsions were subsequently injected into the USMR using a syringe pump at a flow rate of 6 mL/min and an ultrasonic power of 40 W, yielding uniformly dispersed W_1_/O nanoemulsions. Finally, the obtained W_1_/O nanoemulsions were stored at 4°C for 1 h to allow the KGM in the W_1_ phase to form a stable gel network.

Preparation of gelled W1/O/W2 emulsions

2.3

A typical run: First, 0.4 g of Tween 80 was dissolved in 39.6 g of deionized water and stirred for 30 min to obtain the outer aqueous phase (W_2_ phase). Then, 10 g of the previously prepared W_1_/O nanoemulsions under the following conditions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 30 W, 0.1% KGM, 0.75% NaCl, 40% glycerol, and 5% PGPR, was slowly added dropwise into the W_2_ phase. Similarly, the mixture was pre-emulsified using a high-speed homogenizer (Fluko FA30, provided by FLUKO Shanghai Equipment Co., Ltd.) at 7000 rpm for 3 min to form coarse double emulsions. Subsequently, the coarse double emulsions were injected into the USMR using a syringe pump at a flow rate of 6 mL/min and an ultrasonic power ranging from 10 to 30 W, resulting in the formation of gelled W_1_/O/W_2_ double emulsions.

Determination of droplet size and distribution

2.4

The structural hierarchy of the emulsions necessitated different particle-sizing techniques tailored to their respective size scales. The droplet size of the primary W_1_/O nanoemulsions was analyzed using a Dynamic Light Scattering (DLS) analyzer (Zetasizer Pro, Malvern Panalytical, Shanghai, China). DLS is optimized for sub-micron particles where Brownian motion is dominant; therefore, 1 mL of the 100-fold diluted sample was transferred into a PS cuvette for analysis at 25°C to determine the nanoscale dimensions of the primary droplets.

In contrast, the overall droplet size of the W_1_/O/W_2_ double emulsions was measured using a Laser Particle Size Analyzer based on laser diffraction (LT3600, Truth Optics, Zhuhai, China). This technique was selected to accommodate the micrometer-scale diameters of the double-emulsion system, which typically fall outside the reliable detection range of DLS. Samples were added to the measurement cell until an obscuration of 10%–15% was reached. The volume-weighted mean diameter and droplet size distribution were calculated using Mie theory to evaluate the macroscopic dispersion uniformity and structural characteristics of the final carrier system.

Observation with a confocal laser scanning microscopy

2.5

The microstructure of emulsions was observed using a CLSM (Zeiss LSM 900, Germany) equipped with a 40 × objective lens. Nile Red staining solution was prepared by dissolving the dye in DMSO at a concentration of 0.01% (w/v) and stored in the dark at 25°C. A 100 μL aliquot of the double emulsions (diluted 10-fold) was thoroughly mixed with 5 μL of Nile Red solution, then placed on a microscope slide. A small amount of mounting medium was added around the sample, covered with a coverslip, and allowed to equilibrate for 2 min. The sample was then examined under the CLSM to observe the internal structure of the double emulsions. Nile Red fluorescence was excited using a 633 nm laser.

Evaluation of emulsions’ stability

2.6

The stability of the emulsions was evaluated through thermal stability, pH stability, and long-term storage tests to assess their physical stability under various conditions.

• (1)Evaluation of centrifugation stability: Double emulsions were centrifuged at 25°C and 2800 xg for 20 min. The percentage of phase separation — defined as the ratio of the height of the transparent phase to the initial height of the emulsions — was used to evaluate the centrifugation instability of the samples.
• (2)Thermal stability test: The emulsions were stored in an oven at 60°C, and samples were collected at 3 h, 6 h, 9 h, and 12 h to observe any phase separation and changes in droplet size.
• (3)The pH stability test: Buffer solutions were prepared at pH 3.0 (citric acid–potassium dihydrogen phosphate buffer), pH 5.0 (citric acid–sodium citrate buffer), and pH 7.0 (phosphate buffer). All buffers were prepared at a 0.05 M concentration to maintain a consistent ionic strength. A pH meter was calibrated with standard buffer solutions of pH 4.0 and pH 7.0.
• (4)Long-term storage test: The emulsions were stored at 25°C, and their droplet size, zeta potential, and encapsulation efficiency were measured on 0, 7, 14, 21, and 28 days to evaluate long-term stability. Additionally, the samples were visually examined for phase separation, droplet aggregation, and color changes.

Evaluation of the encapsulation efficiency of VC

2.7

The encapsulation efficiency (EE) of VC was evaluated as follows. First, an appropriate amount of double emulsions was sampled and diluted 15-fold with isopropanol for disemulsification, thereby releasing both the encapsulated VC in the W_1_ phase and the free VC in the W_2_ phase to obtain the total VC content. The sample was then centrifuged using an ultrafiltration tube (100 kDa cutoff) at 3000 rpm, and 4°C for 10 min, and the filtrate was collected to determine the free VC. Both the total and free VC contents were analysed by an HPLC. The analysis was performed on a C18 reversed-phase column (100 mm). The mobile phase A consisted of 0.05 mol/L potassium dihydrogen phosphate solution (adjusted to pH 2.5 with hydrochloric acid, filtered through a microporous membrane) and 5% chromatographic-grade formic acid in water. The injection volume was 10 µL, the detection wavelength was set at 244 nm, the flow rate was 0.4 mL/min, and the column temperature was maintained at 30°C. The encapsulation efficiency of VC was calculated according to equation (1):

$$[eqn]$$

where EE is the encapsulation efficiency of VC (%), m_total_ is the content of total VC in the system (mg), and m_free_ is the content of free VC in the W_2_ phase (mg).

Other characterization methods

2.8

The methods for determining zeta potential, interfacial tension, antioxidant activity, rheological behavior, and calorimetric method are described in detail in the Supporting Information.

Results and discussion

3

Critical factors influencing W1/O nanoemulsions

3.1

Effect of flow rate on W1/O nanoemulsions

3.1.1

The flow rate, which determines the residence time, is a critical factor determining emulsion quality during the USMR process. It directly controls the duration and intensity of ultrasonic energy exposure, thereby influencing droplet breakup efficiency, droplet size distribution, interfacial characteristics, and storage stability. Insufficient residence time may result in incomplete cavitation and inadequate droplet disruption, producing emulsions with larger and less uniform droplets. Conversely, excessive residence time can cause interfacial film fatigue or local thermal effects, which may compromise emulsion integrity. Therefore, systematically investigating the influence of residence time is essential for optimizing process parameters and improving emulsion performance [25].

As shown in Fig. 2a and Fig. S1, increasing the residence time from 0.91 min (flow rate 15 ml/min) to 2.27 min (flow rate 6 ml/min) led to a stepwise reduction in the average droplet size. Although longer residence times enhanced droplet size reduction, the rate of improvement gradually slowed, indicating that droplet breakup efficiency was limited by interfacial adsorption saturation. Previous studies have reported a similar phenomenon — when residence time exceeds a certain threshold, energy input per unit volume approaches a limit. For instance, in the preparation of chitosan-stabilized W_1_/O nanoemulsions, droplet size reached a minimum (156 nm) at 15 min under 300 W ultrasound, but slightly increased at 20 min due to emulsifier desorption caused by prolonged sonication and heat accumulation [26]. The stepwise reduction in droplet size with increased residence time (Fig. 2a) is fundamentally driven by the increased frequency of cavitation bubble collapse events per unit volume. In the USMR, the aluminum alloy walls facilitate the formation of a stable standing-wave field. At lower flow rates, the fluid experiences more frequent cavitation bubble actions, ensuring that the primary W_1_ droplets are subjected to intense micro-jets and shock waves. This localized energy intensification within the 4 mm channel overcomes the Laplace pressure of the droplets more efficiently than bath-type sonication, leading to the observed droplet size distribution. Overall, residence time plays a decisive role in controlling droplet breakup and structural stability in USMR-prepared W_1_/O nanoemulsions. Within the investigated residence-time window (0.91–2.27 min), extended sonication enhanced droplet size reduction without inducing emulsion breakdown, whereas beyond this window, excessive ultrasonic exposure may compromise interfacial integrity due to acoustic overheating or emulsifier desorption, as reported in previous studies.Fig. 2(a) Effect of different flow rates on the droplet size of W_1_/O nanoemulsions (Preparation conditions: 5% PGPR, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 1% KGM,0.75% NaCl, and 40% glycerol); (b) Effect of different emulsifier concentrations on the droplet size of W_1_/O nanoemulsions and their changes after 14 and 28 days of storage at 25°C; (c) Effect of different emulsifier concentrations on oil–water interfacial tension; (d) Optical images showing the visual appearance of W_1_/O nanoemulsions with different emulsifier concentrations (Preparation conditions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 0.1% KGM, 0.75% NaCl, and 40% glycerol).

Effect of emulsifier concentration on the W1/O nanoemulsions

3.1.2

During the preparation of water-in-oil (W_1_/O) nanoemulsions, the type and concentration of emulsifier play a crucial role in determining emulsion properties. The hydrophilic–lipophilic balance (HLB) value is a key indicator for evaluating emulsifier suitability [27]. Emulsifiers with different HLB values exhibit distinct lipophilicity and interfacial adsorption behaviors, leading to variations in droplet size distribution, interfacial tension, and storage stability.

In this study, PGPR (HLB = 2.5) was selected as a lipophilic emulsifier, which has been widely validated as an effective stabilizer for double emulsions [28]. W_1_/O nanoemulsions were prepared with PGPR concentrations ranging from 1% to 10%. The effects of PGPR concentration on droplet size, interfacial tension, and visual appearance are summarized in Fig. 2b–d. As shown in Fig. 2c, increasing PGPR from 1% to 2% sharply reduced interfacial tension from 28.0 mN/m to 2.51 mN/m, with a gradual decline to 1.09 mN/m at 10%, indicating interfacial saturation. Correspondingly, droplet size decreased with rising PGPR content (Fig. 2b, d). At 1% PGPR, high interfacial tension (∼28 mN/m) led to milky emulsions with large droplets (∼797 nm) and visible coalescence. At 3–5%, tension dropped to 2.51–1.23 mN/m, droplet size fell to 118–262 nm, and emulsions turned semi-transparent blue, indicating a shift from Mie to Rayleigh scattering. At 7–10%, tension reached ∼ 1.09 mN/m, droplet size stabilized around 85 nm, and emulsions appeared clear blue. However, the reduction in droplet size diminished above 5% PGPR, suggesting a concentration limit for further size minimization [29]. Overall, PGPR concentration exerts a significant influence on the droplet size, optical properties, and stability of W_1_/O nanoemulsions. Increasing emulsifier concentration reduces droplet size and enhances transparency.

Effect of KGM on the W1/O nanoemulsions

3.1.3

KGM, a natural polysaccharide, contains abundant hydroxyl and acetyl groups along its molecular chains that can form a three-dimensional network structure through hydrogen bonding, significantly enhancing the viscosity of the internal aqueous phase. Literature reports have also shown that KGM can enhance the stability of W_1_/O nanoemulsions by increasing the viscosity of the W_1_ phase, thereby reducing coalescence caused by Brownian motion [30]. Furthermore, KGM can form hydrogen bonds and hydrophobic interactions when combined with other thickening agents, exposing more hydrophobic groups that enhance droplet adsorption and stabilize Pickering high-internal-phase emulsions with a storage modulus exceeding 200 Pa [31].

As shown in Fig. 3a, the addition of KGM significantly altered neither the primary W_1_/O droplet (105–110 nm) nor the droplet size distribution. Fig. 3b shows that the emulsions containing KGM exhibited the same transparent blue appearance as the control, indicating that the droplet size remained at the nanoscale. However, the concentration of KGM affects the rheological properties of the emulsions. The flow curves (shear stress vs. shear rate) in Fig. 3c indicate that the emulsions with KGM exhibit significant shear thinning behavior, that is, the shear stress and shear rate have a nonlinear relationship. A critical observation is the substantial augmentation in shear stress at any given shear rate with increasing KGM concentration (from 0.1% to 1.0%), underscoring the formation of a more robust and entangled polymeric network. The viscosity shear-thinning nature is immediately apparent, with viscosity decaying dramatically with increasing shear rate (Fig. 3d). Thus, the KGM-thickened system exhibited typical shear-thinning behavior characteristic of non-Newtonian fluids, demonstrating its excellent thickening ability for the internal aqueous phase and suitability for preparing stable W_1_/O nanoemulsions. The primary W_1_/O droplet size was consistently maintained at approximately 100 nm alongside improved internal stability. This suggests that the high-energy cavitation field within the USMR effectively overcomes the viscous resistance of KGM. The presence of KGM leads to freezing point depression and reduced vapor pressure, which suppresses the formation of large “cushioning” vapor bubbles that typically weaken cavitation. Consequently, the high-velocity micro-jets and shock waves from bubble collapse remain energetic enough to facilitate droplet breakup despite increased macro-viscosity.Fig. 3(a) Effect of KGM addition on the droplet size distribution of W_1_/O nanoemulsions; (b) Optical images showing the visual appearance of W_1_/O nanoemulsions with and without KGM (Preparation conditions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, 40% glycerol, and 0.75% NaCl); (c–d) Effect of KGM concentration on the shear stress and apparent viscosity of the W_1_ phase at 25°C; (e) Effect of W_1_ phase proportion on the droplet size of W_1_/O nanoemulsions (Preparation conditions: flow rate of 6 mL/min, ultrasonic power of 40 W, 5% PGPR, 1% KGM, 40% glycerol, and 0.75% NaCl); (f) Effect of glycerol content on the droplet size of W_1_/O nanoemulsions (Preparation conditions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, and 0.75% NaCl).

As shown in Fig. S2, the incorporation of KGM significantly prolonged the lifetime of W_1_/O nanoemulsions during VC encapsulation stability tests at 60°C, enhancing their ability to retain the active compound. These results further indicate that KGM enhances the stability of the VC-loaded emulsions while having no adverse effect on their appearance or droplet size.

The influence of other factors

3.1.4

The W_1_ phase fraction (volume ratio, φ) is a key parameter governing the performance of emulsion systems, as it directly influences droplet breakup efficiency and coalescence behavior by altering the oil–water interfacial area and phase volume ratio. According to Pickering emulsion theory, when the internal aqueous phase volume fraction exceeds the critical threshold (∼74%), the densely packed droplets experience a sharp increase in coalescence risk [32]. As shown in Fig. 3e, the average droplet size increased with increasing internal aqueous phase content, showing a strong positive correlation [33]. Furthermore, emulsions are thermodynamically unstable systems that tend to minimize their total interfacial area to reduce free energy, favoring droplet coarsening. As the W_1_ fraction increases, the droplet volume and interaction forces between droplets also intensify, promoting coalescence and rupture. This destabilization negatively affects both the storage stability and encapsulation efficiency of the emulsion [34].

Glycerol, a trihydroxy alcohol, contains hydroxyl (–OH) groups that can form hydrogen-bond networks with water molecules. These interactions weaken the intrinsic intermolecular forces within the aqueous phase, thereby reducing interfacial tension and enhancing emulsification efficiency [35], [36]. The interfacial tension of aqueous solutions containing different glycerol concentrations was measured using an interfacial tension meter. Unlike previous measurements, the platinum plate method was employed to minimize non-equilibrium errors caused by dynamic stretching in the platinum ring method [37]. As shown in Fig. S3, the interfacial tension of glycerol solutions decreased linearly from 72.0 mN/m (0%) to 53.8 mN/m (80%), corresponding to a 25.3% reduction. The decrease in interfacial tension significantly enhanced the role of ultrasonic cavitation, improving droplet breakup efficiency and reducing the mean primary W_1_/O droplet size from 144 nm to 81.3 nm (a 43% reduction) (Fig. 3f and Fig. S3).

Overall, the internal aqueous phase fraction and glycerol content jointly governed droplet breakup and interfacial stability, where increasing internal aqueous phase fraction enlarged droplet size, while glycerol effectively reduced interfacial tension, enhanced cavitation efficiency, and significantly improved droplet refinement and emulsion stability.

Critical factors influencing W1/O/W2 emulsions

3.2

Microscopic structure of the double emulsions

3.2.1

To characterize the internal microstructure, TEM was employed following GMS-induced oil-phase gelation. Since double emulsions cannot maintain their original liquid state during the vacuum drying process of TEM, GMS was incorporated as a gelling agent to create a solidified oil-phase scaffold. This pretreatment preserves the structural integrity of the internal morphology post-dehydration [38], [39]. As shown in Fig. 4a–c, the TEM images reveal nanoscale, non-stained cavities which represent the templated footprints of internal water droplets (∼100 nm). The observed multicore morphology confirms the formation of a double-emulsion architecture that was robust enough to withstand the gelation and drying process. As the GMS concentration increased, these cavities became more uniformly distributed, demonstrating that the GMS-gelled network effectively mitigated structural collapse during sample preparation, rather than reflecting the exact in situ dimensions in the liquid state.Fig. 4(a–c) TEM images of W_1_/O/W_2_ emulsions prepared with different GMS concentrations; (d–f) CLSM micrographs of W_1_/O/W_2_ emulsions prepared at different ultrasonic powers (Preparation conditions of W_1_/O nanoemulsions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, and 0.75% NaCl, and 40% glycerol; Preparation conditions of W_1_/O/W_2_ emulsions: water-to-oil ratio of 4:1, 2% Tween 80, and flow rate of 6 ml/min); (g–i) CLSM micrographs of W_1_/O/W_2_ emulsions prepared at different flow rates (Preparation conditions of W_1_/O nanoemulsions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, 0.75% NaCl, and 40% glycerol; Preparation conditions of W_1_/O/W_2_ emulsions: water-to-oil ratio of 4:1, 2% Tween 80, ultrasonic power of 20 W).

CLSM images (Fig. 4d–i) further support these findings. In emulsions without GMS, the red-fluorescent regions correspond to the oil phase, while the dark regions represent encapsulated internal water droplets, clearly displaying the characteristic “core–shell” configuration. Under various ultrasonic powers and irradiation times, multiple internal droplets remained uniformly dispersed within larger oil droplets; the red oil droplets became smaller with increasing ultrasonic power and also decreased in size under lower flow rates. Together, the TEM and CLSM results confirm that the structurally stable W_1_/O/W_2_ double emulsions with well-preserved internal architecture were successfully prepared by using the USMR.

Effect of flow rate on double emulsions

3.2.2

The flow rate, i.e., residence time of the emulsions within the ultrasonic field, directly determines the intensity of ultrasonic energy input and the droplet breakage dynamics, thereby influencing the droplet size distribution, microstructure, and long-term stability of the double emulsions [40]. Laser particle size analysis (Fig. 5a, Fig. S4) showed that flow rate significantly affected the initial droplet size and distribution. At a low flow rate of 3 mL/min, the emulsions exhibited the smallest mean droplet size (1.14 μm) with a narrow unimodal distribution. Increasing the flow rate to 6 mL/min resulted in a slightly broader but still unimodal distribution, suggesting that the ultrasonic energy was sufficient to break droplets effectively while maintaining internal phase stability. However, when the flow rate further increased to 9 or 12 mL/min, the average droplet size increased (up to 5.28 μm), and a clear bimodal distribution appeared, indicating insufficient cavitation energy and incomplete droplet disruption due to shorter residence time, along with enhanced coalescence.Fig. 5(a) Droplet size stability of double emulsions prepared at different flow rates during 35 days at room temperature (Conditions preparing double emulsions: water-to-oil ratio of 4:1, 2% Tween 80, ultrasonic power of 20 W); (b) Droplet size stability of double emulsions prepared under different ultrasonic powers during 35 days at room temperature (Conditions preparing double emulsions: water-to-oil ratio of 4:1, 2% Tween 80, and flow rate of 6 ml/min); (c) Droplet size stability of double emulsions with different oil-to-water ratios (W_1_/O nanoemulsions to W_2_ phase volume ratio) after storage at 60°C for 12 h (Conditions preparing double emulsions: flow rate of 6 ml/min, 2% Tween 80, and ultrasonic power of 20 W); (d-g) Droplet size distributions of double emulsions containing different concentrations of KGM after storage at 60°C for 12 h (Conditions preparing W_1_/O/W_2_ emulsions: water-to-oil ratio of 4:1, 2% Tween 80, ultrasonic power of 20 W, and flow rate of 6 ml/min). All W_1_/O nanoemulsions were prepared under the following conditions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, 0.75% NaCl, and 40% glycerol.

The CLSM images (Fig. 4g–i) provide visual evidence of the cavitation-driven fragmentation process. In the USMR, the flow rate fundamentally dictates the cumulative acoustic energy dose delivered to the fluids. At lower flow rates, the extended residence time ensures that a higher number of cavitation bubble collapse events occur per unit volume of the emulsion. The resulting intense micro-jets and localized shock waves provide sufficient mechanical stress to overcome the Laplace pressure of the secondary droplets, yielding finer and more uniform droplets.

In contrast to the large, polydisperse droplets formed at high flow rates, the smaller droplets generated under optimized cavitation conditions (6 mL/min) exhibited superior interfacial stability. While W_1_/O/W_2_ emulsions prepared at high flow rates showed a pseudo-decrease in average droplet size during storage, this phenomenon is not due to true droplet refinement but rather the structural instability of incompletely emulsified macro-droplets. These large droplets, formed under a weak and non-uniform shear field, tend to undergo rapid coalescence or phase separation, eventually disappearing from the detectable range of the particle size analyzer. Conversely, the smaller droplets produced under sufficient and homogeneous cavitation remain structurally intact due to the rapid, cavitation-enhanced adsorption of stabilizers at the interface. Overall, the flow rate in USMR serves as a regulator of cavitation efficiency: low flow rates facilitate the intensification of acoustic energy, whereas high flow rates lead to an “energy deficit,” resulting in incomplete emulsification [23], [41]. Balancing droplet size control, structural integrity, and production efficiency, a flow rate of 6 mL/min was determined as the optimal parameter for this system.

Effect of ultrasonic power on double emulsions

3.2.3

Ultrasonic power is another critical factor in the preparation of double emulsions, as it directly regulates the cavitation intensity, thereby influencing droplet breakup efficiency, interfacial film stability, and the structural integrity of the multilayer double emulsions [42]. During double emulsions formation, excessively high ultrasonic power can cause over-disruption of internal water droplets or rupture of interfacial membranes, while too low power may result in droplet coalescence [43]. Therefore, systematically investigating the effects of three power levels (10 W, 20 W, and 30 W) on droplet size distribution, microstructure, and emulsion stability is essential for optimizing ultrasonic energy input and balancing breakup efficiency with structural preservation.

As illustrated in Fig. 5b, the average droplet size exhibited a monotonic decrease with the increase of ultrasonic power, which is fundamentally governed by the intensification of acoustic cavitation. Higher power inputs elevate the acoustic energy density within the USMR's microchannel, generating more cavitation bubble formation and collapse. At a relatively low power of 10 W, the system likely operates near the cavitation threshold, resulting in a broad and polydisperse droplet distribution (1–10 μm). Under these conditions, the insufficient shear forces fail to uniformly fragment the oil phase, leading to the formation of “weakly stabilized” macro-droplets that are prone to coalescence during accelerated aging at 60°C.

In contrast, increasing the power to 30 W significantly enhances the cavitation-driven micro-streaming and shock wave intensity. This energetic environment not only promotes efficient droplet disruption but also accelerates the interfacial adsorption kinetics of KGM and surfactants. The rapid formation of a dense, viscoelastic interfacial film, driven by the high-velocity micro-jets, effectively “locks” the droplet structure and prevents post-process coalescence. Consequently, the samples emulsified at 30 W exhibited enhanced structural homogeneity, characterized by a narrower unimodal distribution and improved thermal stability compared to samples processed at lower power levels. The CLSM images (Fig. 4d–f) further confirm that the high-intensity cavitation in the USMR facilitates the formation of a robust, gel-like network that preserves the structural integrity of the W_1_/O/W_2_ system even under thermal stress [44], [45].

Effect of oil-to-water ratio on double emulsions

3.2.4

The oil-to-water ratio (W_1_/O nanoemulsions to W_2_ phase volume ratio) is also a critical factor in the preparation of double emulsions. It directly influences interfacial energy, shear force distribution, and phase volume effects, thereby regulating droplet size, microstructure, and storage stability [46]. As shown in Fig. 5c, the droplet size analysis revealed that increasing the W_1_/O nanoemulsions proportion led to a gradual enlargement of droplet size. However, all emulsions remained macroscopically stable without phase separation after 12 h of storage at 60°C, indicating that the ultrasonic microreactor effectively maintained the emulsions’ macroscopic stability. Although high-temperature accelerated aging slightly reduced droplet size, all emulsions maintained good internal stability across different ratios. These findings provide valuable experimental evidence for optimizing the formulation parameters of double emulsions and improving their long-term structural stability.

Effect of KGM content on the stability of double emulsions

3.2.5

The long-term stability of double emulsions remains a major bottleneck limiting their practical applications, primarily governed by the structural strength and anti-coalescence ability of the internal aqueous phase [47], [48]. However, its actual impact on the microstructure and storage stability of double emulsions requires systematic evaluation. Based on the optimized W_1_/O nanoemulsions formulation, different concentrations of KGM (0%, 0.1%, 0.5%, and 1.0%) were added to the inner aqueous phase. The effects of KGM on droplet size distribution, microstructure, and thermal storage stability (Fig. S7) were investigated using laser droplet size analysis, CLSM, and accelerated aging tests at 60°C for 12 h. As shown in Fig. 5d–g, the freshly prepared double emulsions without KGM or with 0.1%–0.5% KGM displayed a distinct bimodal droplet size distribution, indicating that low concentrations of KGM provided limited suppression of droplet coalescence and osmotic migration. The CLSM images (Fig. S8) show that the viscosity of the internal aqueous phase has a significant impact on the structural stability of the double emulsions. When the internal phase viscosity is low (no KGM or ≤ 0.1% KGM), the droplets exhibit high mobility and are prone to collision and coalescence during emulsification, resulting in larger droplets (1–10 μm) with a distinct bimodal size distribution. In contrast, emulsions containing 1.0% KGM maintain a stable unimodal distribution throughout storage. The high-viscosity internal phase effectively suppresses Brownian motion and droplet coalescence, while the three-dimensional KGM network reinforces the oil–water interfacial film and reduces osmotic shrinkage. The droplet size remains concentrated around 1 μm, with negligible change after 12 h, confirming the stabilizing effect of KGM. At 1.0% KGM, the internal water droplets remain well-dispersed even after heating, indicating that increased viscosity and network reinforcement synergistically enhance the structural stability of the double emulsions.

Stability and encapsulation efficiency of VC double emulsions

3.3

In cosmetics and food formulations, VC is widely used as an antioxidant but is prone to instability [49], [50]. Encapsulation of VC in the inner aqueous phase of double emulsions can enhance its stability. However, due to VC’s strong polarity and acidity, its addition may affect the system’s pH and thereby influence emulsion formation and droplet size distribution [51]. The chemical integrity of encapsulated VC is preserved within the USMR through precise control of the ambient-temperature cavitation environment, where the 3D-printed aluminum structure ensures rapid thermal dissipation. Combined with a short residence time (<2.3 min), this approach minimizes thermal degradation, consistent with literature showing over 90% VC retention during sonication below 60°C [52]. Initially, we evaluated the effect of VC encapsulation on the W_1_/O/W_2_ emulsions. As shown in Fig. 6a, the concentration of VC does not significantly affect the zeta potential.Fig. 6(a) Comparison of the Zeta potential, (b) Droplet size and the interfacial tension of double emulsions with different amounts of VC added; (c, d) Changes in the scavenging efficiency of DPPH and ABTS radicals in the double emulsions containing VC, and co-encapsulating VC (W1 phase) and VE (oil phase), respectively. during storage at 60°C; (Preparation conditions of W_1_/O nanoemulsions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, and 0.75% NaCl, and 40% glycerol. Preparation conditions of W_1_/O/W_2_ emulsions: water-to-oil ratio of 4:1, 2% Tween 80, ultrasonic power of 20 W, and flow rate of 6 ml/min).

Additionally, the average droplet size remained around 2.5  µm with no aggregation or phase separation (Fig. 6b), indicating that VC did not compromise droplet morphology or stability. Simultaneously, Interfacial tension also showed negligible variation (1.232–1.247 mN/m vs. 1.245 mN/m for the blank). CLSM images (Fig. S9) further confirmed that the VC‑loaded emulsions retained the same well-defined W_1_/O/W_2_ structure as the blank, demonstrating excellent compatibility. The antioxidative performance of the W_1_/O/W_2_ emulsions was characterized by scavenging efficiencies of DPPH and ABTS radicals, which assess the total radical-scavenging capacity of the multiphase system. To boost antioxidant performance, the VC-loaded emulsions were prepared with 1% VC in the W_1_ phase, simultaneously, 1% VE in the O phase. As shown in Fig. 6c–d, fresh co-loaded emulsions reached nearly 100% scavenging in both DPPH and ABTS assays, outperforming VC-only samples (83% and 90%, respectively). After 12  h at 60°C, the dual system retained 85% (DPPH) and 88% (ABTS) activity, whereas VC-loaded emulsions in the absence of VE in the O phase dropped to 60% and 58%, due to VE at the oil–water interface and retarding degradation of VC [6], [53], [54].

Regarding droplet size stability under different pH conditions (Fig. 7a), all double emulsions showed a significant reduction from ∼ 1.3  µm to 0.86–1.0  µm after 7 days.Fig. 7(a, b) Variations in droplet size and encapsulation efficiency of double emulsions after one week of storage under different pH conditions; (c,d) Scavenging efficiencies of DPPH and ABTS radicals in VC-encapsulated double emulsions in the presence of VE in the O phase after one week of storage under different pH conditions (Preparation conditions of W_1_/O nanoemulsions: flow rate of 6 mL/min, oil-to-water ratio of 3:1, ultrasonic power of 40 W, 5% PGPR, 0.75% NaCl, and 40% glycerol. Preparation conditions of W_1_/O/W_2_ emulsions: water-to-oil ratio of 4:1, 2% Tween 80, ultrasonic power of 20 W, and flow rate of 6 ml/min).

The stability of the W_1_/O/W_2_ double emulsions exhibited a clear pH-dependence, as illustrated in Fig. 7. Interestingly, the mean droplet size showed a slight downward trend at pH 3 (Fig. 7a). This minor size reduction in acidic media is likely not due to the chemical degradation of KGM, which is known for its excellent acid resistance, but rather to a subtle physical contraction of the weak gel-like network. Under low pH, the high proton concentration can interfere with the hydrogen-bonded junction zones, while the increased ionic strength triggers a transmembrane osmotic pressure gradient, leading to localized water migration and minor droplet refinement. Regarding encapsulation efficiency (EE), the emulsions exhibited superior protective capacity under acidic conditions, maintaining an EE > 40% after 7 days at pH 3 (Fig. 7b). In contrast, the EE dropped precipitously under neutral and alkaline conditions, reaching a minimum of 2% at pH 9. This loss at high pH is attributed to the deacetylation of KGM, which triggers intense molecular aggregation and syneresis (water expulsion), effectively collapsing the internal protective matrix. Antioxidant assays (Fig. 7c–d) further corroborated these findings. The W_1_/O/W_2_ system, featuring the synergistic defense of VC in the internal phase and VE in the oil phase, initially achieved nearly 100% radical scavenging. While activity declined over time across all groups, the samples at pH 3–5 retained significantly higher potency (∼93% ABTS scavenging) compared to those at pH 9 (∼22%). The results emphasize that while acidic stress causes minor physical network adjustments via osmotic effects, alkaline stress induces irreversible chemical transitions in the KGM structure, leading to accelerated oxidation and leakage [55].

In summary, the VC-encapsulated W_1_/O/W_2_ double emulsions in the presence of VE in the O phase showed excellent radical-scavenging capacity within pH 3–7, with optimal performance in acidic environments. These results support its suitability for acidic-to-neutral applications, such as in cosmetics and functional foods requiring stable antioxidant delivery.

Conclusions

4

This study successfully constructed a system of W_1_/O/W_2_ double emulsions with a nanogel-structured inner aqueous phase using a USMR. By systematically optimizing key parameters, including ultrasonic power, flow rate, oil-to-water ratio, and emulsifier concentration, the coupling mechanism between ultrasonic energy input and interfacial stability was elucidated. Under the optimal conditions (ultrasonic power of 20 W, oil-to-water ratio of 2:8, flow rate of 6 mL/min), the resulting W_1_/O/W_2_ emulsions exhibited a narrow droplet size distribution with an average diameter of ∼ 1.1–2.5 μm, a high VC encapsulation efficiency of up to 80%, and excellent physical stability under both 60°C storage and centrifugation. The addition of KGM enhanced the weak gel-like network of the internal phase, effectively increasing the viscosity and structural integrity of the inner aqueous phase, thereby suppressing osmotic exchange and coalescence and extending the emulsion's lifespan. As a result, the double emulsions demonstrated excellent physical stability, minimal structural changes during accelerated thermal storage at 60°C for 12 h. Long-term storage experiments further confirmed that the emulsions maintained stable droplet size and morphology for up to 35 days at room temperature. The incorporation of VC did not significantly alter interfacial charge or morphology, while the encapsulation of VC in the W_1_ phase and the presence of VE in the O phase demonstrated synergistic antioxidant effects in acidic-to-neutral environments (DPPH and ABTS radical scavenging efficiencies approaching 100%). In summary, the USMR enables precise and controllable preparation of double emulsions. Combined with nanogel-enhanced inner phase stabilization, this method offers a novel approach for the efficient encapsulation and controlled release of oxidation-sensitive active ingredients such as VC. This study establishes both theoretical and technical foundations for the structural optimization and industrial application of double emulsions in cosmetic and food systems.

CRediT authorship contribution statement

Chengke Zhou: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Yanting Feng: Writing – original draft, Formal analysis, Data curation. Rongjia Chen: Writing – original draft, Methodology, Investigation. Jingjing Li: Writing – original draft, Data curation. Jie Zhang: Methodology. Jingfu Jia: Methodology, Investigation. Jun Yan: Methodology. Xianwu Peng: Conceptualization. Zhengya Dong: Supervision. Zhilin Wu: Writing – review & editing, 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.