Enhanced Bioavailability and Stability of Curcumin in Cosmeceuticals: Exploiting Droplet Microfluidics for Nanoemulsion Development

Abstract

Curcumin (Cur), a natural polyphenolic compound with potent antioxidant and anti-inflammatory properties, faces significant challenges in cosmeceutical applications due to its poor aqueous solubility and low bioavailability. Nanotechnology offers a promising approach to overcome these limitations and enhance the functionality of cosmetic formulations. In this work, Cur-loaded nanoemulsions (NEs) were developed using a droplet microfluidics technique to enhance Cur’s stability, bioavailability, and permeability for advanced cosmeceuticals. Various oils were screened for Cur solubility, with coconut oil demonstrating the highest capacity. Optimal oil-to-water flow ratios were determined to produce monodisperse NEs with controlled droplet sizes. Characterization via dynamic light scattering (DLS) revealed stable NEs with Z-potential values exceeding −30 mV at both room temperature and +4 °C for up to 21 days, indicating strong colloidal stability. Antioxidant activity was evaluated through DPPH assays, while in vitro permeability studies of the drug-loaded NEs after incorporation into suitable hydrogels, using Strat-M^® membranes mimicking human skin, demonstrated significantly enhanced penetration of the encapsulated Cur. In sum, this work highlights the potential of droplet microfluidics as a scalable and precise method for producing high-performance Cur NEs tailored for cosmeceutical applications.

1. Introduction

Cosmeceuticals are topical formulations that bridge the gap between cosmetics and pharmaceuticals, containing biologically active ingredients designed to improve skin health and appearance by providing both aesthetic and therapeutic benefits [1]. Unlike traditional cosmetic products, cosmeceuticals incorporate components such as antioxidants, peptides, herbal extracts, and vitamins, which have demonstrated physiological effects on skin cells [2]. The cosmeceutical industry is experiencing a major shift toward the incorporation of medicinal plants and botanical extracts into skincare and personal care products. This trend is driven by consumer demand for natural, sustainable, and effective alternatives to synthetic cosmetic ingredients, reflecting concerns for both personal health and environmental impact [3].

In this context, Curcumin (Cur), a bioactive polyphenolic compound derived from Curcuma longa, has attracted significant attention due to its potent anti-inflammatory, antioxidant, and anticancer properties [4]. Cur plays a multifaceted role in cosmeceuticals, serving as a potent antioxidant and anti-inflammatory agent that helps protect the skin against oxidative stress and premature aging. It is valued for its skin-brightening properties, ability to reduce hyperpigmentation and redness, and its antimicrobial action, which makes it effective in acne-prone and sensitive skin formulations. It is commonly used in creams, serums, and masks to enhance skin luminosity, improve elasticity, and combat visible signs of aging [5,6]. However, its clinical and industrial applications are hindered by poor aqueous solubility, chemical instability, susceptibility to photodegradation and low bioavailability [7]. To address the aforementioned challenges, numerous research studies have explored alternative technological approaches aimed at improving the solubility, stability, and bioavailability of Cur. These include the development of amorphous solid dispersions [8], the use of deep eutectic solvents [9], and the incorporation of the active compound into natural polymer carriers such as chitosan [10]. Although these methods have yielded encouraging results, nanoencapsulation approaches—particularly the development of Cur-loaded nanoemulsions (NEs)—have emerged as effective strategies to enhance solubility, protect the active compound from degradation, and improve bioaccessibility [11].

NEs are thermodynamically unstable yet kinetically stable nanodispersed systems, typically characterized by droplet sizes below 200 nm. Their small droplet size and large surface area promote the absorption of bioactive components into the skin, supporting tissue repair and thus enhancing the therapeutic efficacy of active ingredient. Various preparation techniques have been employed in the literature for the fabrication of NEs, including ultrasonication [12], high-pressure homogenization [13] and phase inversion methods [14]. However, these approaches present certain challenges related to achieving consistent droplet size, maintaining stability, and ensuring high encapsulation efficiency.

Droplet-based microfluidics technique offers a highly controlled and versatile platform for producing uniform NEs with tunable droplet sizes and narrow size distributions. By enabling precise manipulation of fluid flow at the microscale, this technology facilitates the generation of monodisperse droplets and highly reproducible NE formulations [15]. Compared to conventional high-energy methods like ultrasonication, droplet microfluidics offers advantages in scalability, reduced energy consumption, and enhanced control over physicochemical properties, making it a promising approach for industrial translation of Cur nanoformulations [16].

Building upon advances in encapsulation technologies, the present study aims to develop and characterize Cur-loaded NEs using droplet-based microfluidics. The primary objective is to produce stable, high-loading NEs with small droplet sizes, narrow size distributions, and enhanced physical stability, thereby improving the bioavailability and therapeutic efficacy of Cur. By exploiting the precise control over droplet formation offered by microfluidic technology, this work seeks to overcome key limitations of conventional high-energy methods and advance scalable strategies for producing high-performance Cur nano-formulations.

2. Materials and Methods

2.1. Materials

Cur (from Curcuma longa, turmeric) was used as the active antioxidant compound in powder form and was supplied by Sigma-Aldrich (St. Louis, MO, USA). Soy lecithin was obtained from Serva (Heidelberg, Germany), while the non-ionic emulsifier (hydrogenated castor oil) Kolliphor^® RH 40 (RH40), squalene, jojoba oil, liquid coconut oil, calendula oil, and argan oil were purchased from Sigma-Aldrich (St. Louis, MO, USA). Glycerin, coco glucoside, phenoxyethanol, Calendula officinalis, Carbopol 940 and PEG-40 hydrogenated castor oil were kindly donated from Novita Group (Thessaloniki, Greece). 1,1-Diphenyil-1-picrylhydrazyl (DPPH) free radical (purity > 97.0) was purchased from Tokyo Chemical Industry Co. (Eschborn, Germany). SERVAPOR dialysis tubing, MWCO 12–14,000 with a diameter of 50 mm was purchased from SERVA Electrophoresis GmbH, Heidelberg, Germany. Strat-M^® membranes were purchased from Merck Millipore, Burlington, MA, USA. All other materials and chemical reagents were of analytical or pharmaceutical grade.

2.2. Cur’s Solubility Studies in Various Oils

The solubility of Cur in various oily phases (squalene, jojoba oil, coconut oil, calendula oil, and argan oil) was determined following a modified method based on Mandal et al. [17]. Specifically, an excess amount of Cur (~200 mg) was added to a glass vial containing 2 g of each oil. Τhe mixtures were sealed, heated at 60 °C for 1 h in a water bath to accelerate the dissolution, and subsequently shaken at 200 rpm at 40 °C for 24 h to reach equilibrium. After centrifugation (18,000× g for 10 min using a Microfuge^® 22R centrifuge, Beckman Coulter, Brea, CA, USA), the supernatants were diluted with absolute ethanol. Quantification of Cur was performed using UV-Vis spectrophotometry at 420 nm using a calibration curve prepared with known amounts of Cur dissolved in ethanol.

2.3. Droplet Microfluidics for NEs; Preparation and Pseudo-Ternary Phase Diagram Construction

The blank NEs (i.e., without the addition of Cur) were prepared using predetermined emulsifier/co-emulsifier weight ratios derived from phase diagram studies. Formulations with varying proportions of coconut oil, S_mix (surfactant/co-surfactant mass ratio), and aqueous phase were tested to identify the nanoemulsion region within the pseudo-ternary phase diagrams. The aqueous phase consisted of deionized water containing RH40 (6% w/w), dissolved at 50 °C under magnetic stirring (300 rpm), while the oil phase consisted of coconut oil and lecithin (2% w/w). All solutions were filtered using 0.22 μm PVDF membrane filters (Sigma Aldrich, St. Louis, MO, USA) prior to use. The aqueous and oil phases were each delivered by separate Mitos pressure pumps (Dolomite Microfluidics, Royston, UK), each equipped with an integrated Mitos flow sensor (Dolomite Microfluidics, Royston, UK). Fluid connections were made using 250 μm thick FEP tubing with an internal diameter (ID) of 1/16″ and an outer diameter (OD) of 2 mm, respectively. NEs were generated using a 5-inlet, 150 μm microfluidic chip (Dolomite Microfluidics, United Kingdom) and collected in glass vials. Various flow rates (μL/min) (

Table 1. Different sets of oil phase and aqueous phase flow rates (μL/min) tested in the preparation of the NEs.

Trial Number	Flow Rate (μL/min)
	Aqueous Phase	Oil Phase
Test 1	200	20
Test 2	400	20
Test 3	1000	20
Test 4	1000	10

) of the aqueous and oil phases were tested to evaluate their influence on NE formation. The experimental setup is depicted in Figure 1. The samples were stored at 25 °C in the dark for at least 24 h prior to analysis.

2.4. NEs Characterization

2.4.1. Attenuated Total Reflectance–FTIR Spectroscopy (ATR-FTIR)

The successful synthesis of the new materials was confirmed using ATR-FTIR spectroscopy. Spectra were collected using a Shimadzu IRTracer-100 spectrometer (Kyoto, Japan) equipped with a QATR™ 10 single-reflection diamond ATR accessory. Absorption spectra were recorded over the spectral range of 450–4000 cm⁻¹ with a resolution of 2 cm⁻¹. A total of 32 scans were collected for each sample, and the spectra were normalized and baseline-corrected prior to analysis.

2.4.2. Physical Stability of NEs During Storage

Freshly prepared NEs (without the addition of Cur) were stored in the dark at room temperature (RT) and at +4 °C for 4 weeks without the addition of preservatives. Physical stability of NEs was assessed by monitoring changes in droplet size, Z-potential and polydispersity index (PDI) using a Zetasizer Nano ZS 2000 (Malvern Instruments, Worcestershire, UK), over the storage period.

2.5. Preparation of Cur-Loaded NEs

Cur NEs were prepared using the predetermined emulsifier/co-emulsifier weight ratios obtained from pseudo-ternary phase diagrams. Two distinct Cur loadings, specifically 25 mg and 50 mg, were used by dissolving them (at a 1.25 and 2.5 mg/mL concentration) into the dispersed (oil) phase. The aqueous phase consisted of deionized water (89% w/w) containing RH40 (2% w/w), dissolved at 50 °C under magnetic stirring (300 rpm) while the oil phase consisted of coconut oil (7.94 and 7.85% w/w), lecithin (1% w/w) and dissolved curcumin (0.06 and 0.13% w/w). All solutions were filtered using 0.22 μm PVDF membrane filters (Sigma Aldrich, St. Louis, MO, USA) prior to use. The experimental setup and procedure followed the same method as described for the neat NEs (Section 2.3, Figure 1). After their preparation, Cur NEs were stored at 25 °C in the dark for at least 24 h prior to analysis.

2.5.1. Encapsulation Efficiency (EE) of Cur

The EE of Cur NEs was determined to quantify the amount of Cur incorporated into the formulation. EE (%) was calculated by centrifuging the NEs at 18,000× g for 10 min using a Microfuge^® 22R centrifuge (Beckman Coulter, Brea, CA, USA), collecting an aliquot of the supernatants, diluting the supernatants with absolute ethanol. No clear phase separation and curcumin precipitation were observed after centrifugation due to the small droplet size of NEs. The supernatants collected were akin to the original NEs and Cur content was measured via UV–Vis spectrophotometry at 420 nm, using the following equation:

$$EE \left(\%\right)= {\displaystyle \frac{Amount of curcumin encapsulated}{Amount of curcumin added in preparing nanoemulsion }}\times 100$$

(1)

2.5.2. pH Measurement of CUR ΝΕs

Immediately following preparation, the pH of the emulsions was determined by immersing the pH sensor (Microprocessor, WTW, pH 535, Gemini BV, Apeldoorn, The Netherlands) into the samples. Each pH value mentioned in the manuscript represents the average of three consecutive measurements.

2.5.3. Determination of L*, a* and b* Values

The NEs were analyzed at RT for color variations using a portable MiniScan XE Plus spectrophotometer (HUNTERLAB, Washington, VA, USA), and the results were expressed in the CIE Lab system, where L* represents lightness (0 = black, 100 = white), a* ranges from green (−a*) to red (+a*), and b* ranges from blue (−b*) to yellow (+b*). Positive a* values indicate red, negative a* values indicate green; positive b* values indicate yellow, and negative b* values indicate blue. Samples were placed in a cuvette and analyzed with the spectrophotometer. Color parameters were determined in triplicate. The color intensity (E) of NEs was calculated using the following formula:

$$\mathrm{E}=\sqrt{{\mathsf{\alpha }}^2+{\mathsf{\beta }}^2+{\mathrm{L}}^2}$$

(2)

2.5.4. DPPH Antioxidant Assay

DPPH (1,1-diphenyl-1-picrylhydrazyl) free radical scavenging activity was measured to determine the antioxidant activity of Cur in NEs before and after UV exposure following a method from a previous study [18]. Briefly, the emulsions were placed inside glass vials and underwent UV irradiation for 24 h in a controlled environment (chamber BS-09) at a wavelength of 280 nm, with a constant temperature of 25 ± 2 °C and relative humidity (RH) of 50%. After a continuous 24 h irradiation period, the samples were removed from the chamber for analysis. 1 mL of each emulsion dispersed in ethanol (1% v/v) was added to 3 mL of a 5 × 10⁻³ mg/mL ethanol DPPH solution. The reference sample consisted of 1 mL EtOH and 3 mL of the DPPH/EtOH solution. The samples were placed in a sonication bath and their absorbance was recorded after 30, 60 and 90 min using a UV-Vis spectrometer (UV Probe 1650, Shimadzu, Tokyo, Japan) at 517 nm. The free radical scavenging activity was described according to the following Equation [19]:

$$\mathrm{Free} \mathrm{radical} \mathrm{scavenging} \mathrm{activity}\left(\mathrm{\%}\right)={\displaystyle \frac{\mathrm{Absorbance} \mathrm{of} \mathrm{control}-\mathrm{Absorbance} \mathrm{of} \mathrm{emulsion}}{\mathrm{Absorbance} \mathrm{of} \mathrm{control}}}\times 100$$

(3)

2.5.5. In Vitro Permeation Study of Cur NEs

Vertical Franz diffusion cells (PermeGear Inc., Hellertown, PA, USA) fitted with regenerated cellulose membranes (SERVAPOR, MWCO 12–14,000; SERVA Electrophoresis GmbH, Heidelberg, Germany; diameter 50 mm) were used. Membranes were cut to the appropriate size and positioned securely between the donor and receptor chambers, ensuring that no air bubbles were trapped between the membrane and the receptor fluid. The receptor chamber was filled with phosphate-buffered saline (PBS, pH 5.5) and ethanol in a 50:50 (v/v) ratio to maintain sink conditions, while the donor chamber was loaded with 3 mL of Cur NEs in two different concentrations (0.06% and 0.13% w/w). The receptor chamber was maintained at 32 °C (skin surface temperature) in a water bath and stirred continuously at 300 rpm for 30 min to stabilize the temperature before sample addition. The donor chamber was then sealed with parafilm. At predetermined intervals over 24 h, aliquots (0.5 mL) were withdrawn from the receptor chamber and immediately replaced with an equal volume of fresh buffer solution. Cur release into the receptor chamber was quantified by UV–Vis spectrophotometry at 420 nm, using a calibration curve prepared from known concentrations of Cur in EtOH.

2.5.6. Permeability Study of Cur from Exfoliating Gels Using Strat-M^® Membranes

To evaluate the permeability of Cur NEs incorporated into appropriate hydrogels (

Table 2. Composition of hydrogels in which Cur-loaded NEs were incorporated.

Ingredients	Function	Content (%wt)
Coco-glucoside	Surfactant	1.4
Carbopol 940	Thickener	0.9
Glycerin	Conditioner	4.5
PEG-40 hydrogenated castor oil	Surfactant	1.6
Phenoxyethanol	Fragrance agent	0.45
Calendula officinalis	Conditioner	0.9
Water	Vehicle	90.25%

), Strat-M^® membranes (Merck Millipore, Burlington, MA, USA), which simulate human skin, were employed. The membrane was positioned over the receptor chamber with the shiny side facing upward to correspond to the stratum corneum. Special attention was paid to remove air bubbles between the membrane and the receptor fluid to ensure optimal contact. The receptor chamber was filled with 200 μL of PBS (pH 5.5) and EtOH in a 50:50 (v/v) ratio to maintain sink conditions, while the donor chamber was loaded with 100 mg of Cur NEs mixed with the gel at a 40/60 w/w ratio. The final Cur concentrations in the mixtures are 0.024% w/w for the 0.06% starting concentration, and 0.052% w/w for the 0.13% starting concentration. The receptor chamber was maintained at 32 °C (skin surface temperature) under continuous stirring (100 rpm) and stabilized for 30 min prior to sample addition. The donor chamber was then sealed with parafilm. Aliquots of approximately 0.5 mL were withdrawn at predetermined time points (1, 2, 4, and 6 h) and replaced with an equal volume of fresh buffer solution. The released Cur content was determined using high-performance liquid chromatography (HPLC) (Agilent 1260 Infinity II, Agilent, Santa Clara, CA, USA) employing a validated method [20]. Analysis was performed on a reversed-phase C18 Luna column (250 mm × 4.6 mm i.d., 5 μm particle size) with isocratic elution using a mobile phase of acetonitrile and 2% acetic acid in water (50:50, v/v) at a flow rate of 1.2 mL/min. Standard curcumin solutions (0.01–100 ppm) were used to construct a calibration curve.

2.6. Statistical Analysis

All sample collections and measurements were conducted in triplicate. The average and standard deviation were calculated using Microsoft Excel. Origin Pro 9.0 (OriginLab Corporation, Northampton, MA, USA) was applied to prepare all figures.

3. Results and Discussion

3.1. Selection of Oil Phase

The successful formulation of NEs relies on the ability of oil phase (vehicle) to solubilize the active substance, as solubility directly influences the loading capacity, and, consequently, the pharmaceutical efficiency of the final system. In the case of Cur (a highly hydrophobic polyphenolic compound), selecting an appropriate oil phase is particularly critical for achieving both stability and efficacy [14]. Another significant aspect concerns the thermodynamic stability of the system. When the active compound exhibits high solubility in the oil phase, the likelihood of Cur molecules migrating out of the oil is reduced, thereby helping to prevent destabilization mechanisms such as Ostwald ripening or recrystallization [21]. This is particularly important in cosmeceutical formulations, where the stability of the final product directly affects its efficacy. Therefore, the selection of the optimal oil (the one that can dissolve the highest possible amounts of the bioactive molecule) is vital.

The Cur solubility results in the various tested oils are presented in

Table 3. Solubility values (mg/mL) of Cur in various oils.

Oil	Solubility (mg/mL)
Jojoba	0.155 ± 0.015
Calendula	0.291 ± 0.021
Argan	0.115 ± 0.012
Coconut	0.529 ± 0.030
Squalene	Insoluble

. Significant differences were observed in the solubilization capacity across the tested oils. Coconut oil exhibited the highest solubility (0.529 mg/mL), nearly twice that of calendula oil, while all other oils showed substantially lower values. This superior performance can be attributed to the fatty acid profile of coconut oil, which is rich in short-chain saturated fatty acids—primarily octanoic acid (53.1% w/w) and capric acid (30.8% w/w), corresponding to C8 and C10 chains, respectively. Literature reports indicate that Cur solubility in individual fatty acids generally follows the order: octanoic acid >> linoleic acid > oleic acid [14,22]. This trend is associated with differences in hydrophilic–lipophilic balance (HLB), which reflects the relative proportions of hydrophilic and hydrophobic components and determines the suitability of an oil for stabilizing oil-in-water (O/W) or water-in-oil (W/O) emulsions. Octanoic acid, with a balanced HLB, promotes the diffusion and integration of Cur molecules into both polymeric and fluid lipid matrices, resulting in higher solubility. Calendula oil (HLB 7), which exhibited the second-highest solubility for Cur (0.291 mg/mL), consists mainly of linoleic and palmitic acid. These fatty acids provide modest, yet notable, solubilizing potential [23]. Jojoba oil exhibited an intermediate solubilization ability (0.155 mg/mL), relative to other tested oils. Structurally, jojoba oil consists of long, linear alkyl esters lacking polar groups, a characteristic that limits its capacity to solubilize hydrophobic molecules with high polarity or aromatic character, such as Cur [24]. Additionally, the presence of free hydroxyl or carboxyl groups in the oil may increase the dispersibility of Cur within this system. The low solubility observed for argan oil is likely associated with its high oleic acid (C18) content—reported to be less effective for Cur dissolution than short-chain fatty acids [25]. In contrast, the purely hydrophobic nature of squalene is the likely reason behind its complete inability to dissolve Cur. Due to the absence of polar or semi-polar groups and its linear molecular structure, squalene lacks the capacity to interact with Cur, precluding the formation of a compatible microenvironment for incorporating hydrophobic active molecules.

3.2. Pseudo-Ternary Phase Diagrams

Studying NE formation in the absence of an active ingredient (blank systems) is a critical step in the design of stable systems, as it enables the determination of the stability region and the appropriate preparation conditions. In this context, pseudoternary phase diagrams were constructed based on the three main components of NEs: (1) coconut oil, (2) water and (3) the emulsifier/co-emulsifier (mass ratio) mixture (S_mix). The objective was to examine the effect of the component mass ratios and the flow rates of the aqueous and oil phases on NE formation using microfluidic devices. Figure 2a presents the pseudo-ternary phase diagram. Red circles indicate compositions that yielded NEs with an average diameter < 350 nm and ζ-potential < −30 mV, whereas black squares correspond to compositions that did not form NEs. According to the results, the diagram clearly demonstrates that only S_mix values of 1, 2, and 3, combined with aqueous-to-oil phase flow ratios of 50:1 and 100:1, produced NEs meeting the defined particle size and ζ-potential criteria. In contrast, lower flow ratios (e.g., 10:1 and 20:1) failed to generate NEs within the target specifications.

Visual observation of the nanoemulsified systems immediately after preparation is given in Figure 2b. According to the obtained images, all NEs exhibited a milky appearance and light white color, indicating evident stability at the macroscopic level. This milky character is a typical indicator of successful nanostructure formation, associated with a high refractive index and homogeneous particle distribution.

DLS analysis revealed that the average hydrodynamic droplet diameter ranged between 154 nm and 343 nm (

Table 4. Mean hydrodynamic diameter, Z-potential, and PDI of the blank Nes.

Sample (Smix-Flow Rate Ratio)	Mean Hydrodynamic Diameter (nm)	Z-Potential (mV)	PDI
1-50:1	343 ± 1.19	−13.96 ± 3.57	1.000 ± 5.32
1-100:1	225 ± 0.36	−37.20 ± 0.90	0.579 ± 1.49
2-50:1	154 ± 0.25	−40.79 ± 0.71	0.269 ± 0.27
2-100:1	261 ± 0.42	−35.75 ± 1.02	0.733 ± 1.17
3-50:1	321 ± 0.98	−40.57 ± 0.88	0.473 ± 0.56
3-100:1	328 ± 0.92	−40.10 ± 0.69	0.772 ± 1.89

). Among the tested formulations, sample 2-50:1 (i.e., S_mix = 2 and flow ratio 50:1) exhibited the smallest diameter (154 nm), while sample 1-50:1 (S_mix = 1, flow ratio 50:1) resulted in the largest (343 nm). The ζ-potential values varied from −13.96 to −40.79 mV, with most samples exceeding the absolute threshold of 30 mV, indicative of good kinetic stability in NEs. The polydispersity index (PDI), which reflects size distribution and homogeneity, ranged from 0.269 to 1.000. The lowest value was observed in sample 2-50:1 (0.269), indicating good homogeneity, whereas sample 1-50:1 exhibited the highest PDI (1.000), suggesting size heterogeneity and reduced stability. The differentiation between these two examples underscores the critical influence of both the S_mix ratio and the aqueous-to-oil flow rate on the physicochemical quality of NEs prepared via microfluidics.

The effect of the aqueous-to-oil phase flow ratio on microfluidics production can be attributed to three main mechanisms. First, as reported by Fathordoobady et al. [15] and Amoyav & Benny [26], a higher proportion of aqueous phase promotes immediate dilution of oil droplets, thereby reducing the rate of coalescence. Second, the increased water content facilitates more efficient and rapid distribution of the emulsifier at the oil–water interface, lowering the interfacial tension and enabling the stabilization of droplets at smaller sizes [27]. Third, the increased flow of the continuous phase generates conditions of higher shear stress and inertial resistance, which facilitates the faster and more stable breakup of droplets in flow regions with enhanced size control [28].

3.3. Stability During Storage

Following preparation, ΝΕs were stored in the dark for 21 days at both +4 °C and RT, without the addition of preservatives. Stability during storage was assessed by monitoring changes in droplet size (nm), Z-potential (mV) and PDI. During this storage period, no apparent visual changes or phase separation were observed, indicating the macroscopic stability of the prepared NEs (Figure 3).

It is important to emphasize that most systems exhibited greater stability when stored under refrigeration (+4 °C) compared to RT. This enhanced stability can be attributed to the reduction in droplet kinetic energy, which slows down coalescence and molecular rearrangement processes. In some samples, refrigeration further contributed to smaller droplet diameters and more consistent PDI values, without evidence of phase separation or sediment formation.

Regarding the size of the nanodroplets, the results from the stability study are presented in Figure 4a. The most stable profile was observed for sample 2-50:1. This particular NE maintained an average diameter below 166 nm across all time points and storage conditions, with only a ±12 nm variation between day 0 and day 21. This low variability underscores the structural robustness of the system and the ability of the emulsifiers to preserve nanodroplet integrity. By contrast, sample 2-100:1 displayed more pronounced fluctuations, with its initial diameter (261 nm) decreasing to 137 nm on day 7 at RT, before increasing again to 221 nm by day 21 at +4 °C. Such transient destabilization is likely attributable to temporary interfacial redistributions. Similarly, sample 1-100:1 showed significant size variation. It started at 225 nm, increased to 388 nm by day 14 (RT), and stabilized at 319 nm on day 21 (+4 °C). This instability correlates with the initially high polydispersity index of this sample, suggesting heterogeneity in the droplet population and a tendency toward coalescence. These large changes may also be attributed to insufficient stabilization at the oil–water interfaces.

Z-potential (Figure 4b) hardly changed with time, whereas PDI (Figure 4c) was more ‘vulnerable’ to long-term storage, which corroborates also with the findings of Ye et al. [14]. In brief, the Z-potential provides an indication of the surface charge of the emulsion particles. When its absolute value is high, it creates repulsive forces between the particles, which can enhance the physical stability of multiphase systems [29]. Sample 2-50:1 consistently demonstrated superior thermal and kinetic stability. Its Z-potential ranged from −36.4 to −41.3 mV throughout the storage period, values that are considered highly indicative of good stability due to the electrostatic repulsion between droplets. Relatively stable Z-potentials were also observed in samples 3-50:1 and 3-100:1 (varying from −37.9 to −40.6 mV), likewise indicating the achievement of an acceptable kinetic stability profile. The high negative values likely result from the distribution of ionic groups within the emulsified layers and surface rearrangement during storage. Both RH40 and lecithin can contribute to the overall negative charge due to their molecular structure and ionizable groups at the interface, which reflects in the negative Z-potential measurement [30].

The most monodisperse systems were samples 2-50:1 and 3-50:1, with PDIs ranging from 0.252 to 0.321 and 0.472 to 0.517, respectively. These values fall within the acceptable range for monodisperse NEs (PDI < 0.3–0.5) [31]. Sample 2-100:1 exhibited a higher PDI (~0.8), indicating the presence of mixed populations, despite the relative stability of its droplet size. Such polymodal distributions may be associated with a slower rate of droplet stabilization due to the higher aqueous-to-oil phase flow ratio, thus leading these NEs to a transitional increase in their droplet size.

3.4. Characterization of Cur NEs

The parameters for Cur-loaded NE preparation were based on the optimal conditions identified in the previous section for blank NEs. Specifically, a fixed S_mix ratio (=2) was applied, and two aqueous-to-oil phase flow ratios (50:1 and 100:1) were selected, both of which had already demonstrated the ability to produce blank NEs with the desired properties (small droplet size, satisfactory Z-potential and low PDI). Two different Cur amounts, namely 25 and 50 mg, were dissolved in 20 mL of coconut oil and were further used for encapsulation in order to determine whether increasing the concentration of the active substance affects the morphology and macroscopic appearance of the NEs. Cur-loaded NEs with two different concentrations were thus prepared, namely 0.06% w/w and 0.13% w/w. Figure 5a shows the images of Cur NEs immediately after preparation.

ΝΕs exhibited a milky texture and an intense, bright yellow color, regardless of the flow rate ratio or Cur content. The color homogeneity and the absence of precipitates or sedimentation suggest good distribution of the drug within the core of the nanodroplets. Furthermore, the fact that the optical properties are not significantly affected by the increase in Cur concentration suggests that the system has the capacity to incorporate larger amounts of the API without immediate risk of saturation or phase separation.

In general, NEs are carrier systems consisting of an oil and a water phase stabilized by surfactants and co-surfactants, distinguished from conventional emulsions by having droplets sized at the nanometer scale [32]. The successful encapsulation of the active substance is explained at the molecular level through the proposed droplet formation mechanism, as presented in Figure 5b. Soy lecithin, a natural phospholipid, acts as a co-emulsifier by forming a stable film at the interface between the two phases (aqueous and oil). Its hydrophilic phosphate head orients towards the external aqueous phase, while the hydrophobic chains anchor within the droplet core, along with the oil phase containing the dissolved Cur. RH40, a polyoxyethylene derivative of hydrogenated castor oil, further enhances system stability by also acting as an emulsifier, filling the gaps between lecithin molecules at the interface of the two phases. The PEG moieties within the RH40 structure orient towards the aqueous phase, providing an electrostatic and steric barrier, whereas its saturated C18 chains structurally integrate into the oily core. This mechanism promotes the formation of small-sized droplets by reducing the interfacial tension and enhancing the thermodynamic stability of the nanodroplets, as demonstrated in previous studies [33]. The selection of coconut oil as the lipid phase, primarily composed of medium chain saturated fatty acids, offers a low-viscosity core that facilitates the diffusion of Cur and maintains stability within the nanodroplets.

An additional crucial factor for the successful production of drug-loaded NEs is the use of droplet microfluidics with a flow focusing configuration. This technology allows precise control over flow rates, mixing intensity, and shear energy, providing significant advantages over other conventional NE production methods. As reported, nanodroplets formed under controlled flow conditions exhibited reduced size distribution, greater thermal stability, and higher encapsulation efficiency compared to those produced by probe sonication or simple mechanical stirring methods (such as homogenization [34,35,36]). The preservation of the physicochemical properties of sensitive active compounds, such as Cur, is markedly enhanced by the limited thermal and mechanical stresses imposed by the microfluidics approach. Therefore, these findings support the suitability of microfluidics technology for the production of Cur-loaded NEs suitable for topical application.

The IR spectra of the raw materials and the NEs (both blank and Cur-loaded) are shown in Figure 6. The presence of Cur was confirmed by the characteristic peaks observed in the spectra of the NEs containing the active compound. Specifically, the peak at 1637 cm⁻¹ is attributed to overlapping stretching vibrations of the C=C bonds of alkenes and the C=O bonds of carbonyl groups. This peak further supports the encapsulation of Cur within the NE system. Additionally, the peak at 1250 cm⁻¹ corresponds to the bending vibrations of the C-O bond and is particularly characteristic of Cur’s chemical structure, while the peak at 1540 cm⁻¹ is assigned to in-plane bending vibrations of aliphatic and aromatic systems, including δ(CC–C), δ(CC=O), and ν(C–C) modes of aromatic bonds [37].

The FTIR spectra of RH40 and soy lecithin exhibited distinct absorption patterns. The peak at 1462 cm⁻¹ is associated with the bending vibrations of the C-H bonds in the CH₂ and CH₃ groups, while the peak at 1735 cm⁻¹ corresponds to the stretching vibrations of carbonyl esters. The peaks at 2958, 2925, and 2854 cm⁻¹ arise from successive C-H stretching vibrations characteristic of the alkyl chains in triglycerides. These spectral regions are in full agreement with literature data for plant-derived lipid emulsifiers and confirm the presence of functional groups responsible for interfacial stabilization [38]. The FTIR spectrum of pure coconut oil revealed pronounced double peaks at 2920 and 2852 cm⁻¹, characteristic of the asymmetric and symmetric stretching vibrations of the C-H bonds in CH₂ and CH₃ groups. These peaks are attributed to the high content of saturated fatty acids with linear carbon chains in the oil, such as palmitic, lauric, and stearic acids. Another significant peak was observed at 1739 cm⁻¹, corresponding to the carbonyl (C=O) stretching vibration of ester groups, while the peaks at 1165 and 1105 cm⁻¹ are associated with the C-O stretching vibrations [37,39].

Looking at the obtained spectra of the blank NEs and Cur-loaded NEs (at both drug loadings), the obtained results indicate that each ingredient, especially Cur, was well retained in the NEs after preparation.

The effectiveness of a NE as a carrier for bioactive compounds, such as Cur, depends on the encapsulation efficacy of the respective active substance and the stability of the system. Results, summarized in

Table 5. Characteristic properties of Cur-loaded NEs.

Sample	Cur Amount Added (mg)	Cur Amount Encapsulated (mg)	EE (%)	Size (nm)	Z-Potential (mV)	PDI	pH
25Cur_50_1	1.25	0.63 ± 0.13	50.4 ± 1.2	172 ± 0.41	44.11 ± 0.45	0.332 ± 0.13	5.1 ± 0.11
25Cur_100_1	1.25	0.41 ± 0.07	32.8 ± 1.1	181 ± 0.28	41.34 ± 1.13	0.468 ± 0.41	4.9 ± 0.20
50Cur_50_1	2.5	1.24 ± 0.29	49.6 ± 1.07	240 ± 0.53	42.67 ± 3.01	0.547 ± 0.15	4.8 ± 0.22
50Cur_100_1	2.5	0.78 ± 0.19	31.2 ± 1.13	194 ± 0.39	45.05 ± 0.89	0.486 ± 0.31	5.0 ± 0.21

, present, for each formulation, the amount of Cur initially added and ultimately encapsulated, the EE (%), the average hydrodynamic diameter (size), the Z-potential, the PDI and the pH values of each system.

Starting with the systems containing the lowest initial Cur quantity (25 mg), sample 25Cur_50_1 exhibited an encapsulation efficiency of 50.4%, with a droplet size of 172 nm, Z-potential of −44.11 mV, and PDI of 0.332. This performance is considered particularly satisfactory, as it achieves good incorporation of the active compound with a relatively low PDI and high Z-potential. The corresponding sample with a flow ratio of 100:1 (25Cur_100_1) showed a markedly reduced encapsulation efficiency (32.8%), while the droplet diameter slightly increased to 181 nm and the PDI to 0.468. The pH values of both systems ranged between 4.9 and 5.1, indicating a neutral to mildly acidic environment suitable for dermal applications.

In the case of samples with a higher Cur concentration (50 mg), the EE of the 50Cur_50_1 sample reached 49.6%, with a droplet size of 240 nm and a Z-potential of −42.67 mV. Although the droplet size increased, the encapsulation of the active compound remained high, indicating that the system can satisfactorily accommodate double the amount of the active ingredient without a dramatic loss in stability. The corresponding sample with a flow ratio of 100:1 (50Cur_100_1) showed the lowest EE (31.2%), with a size of 194 nm and a PDI of 0.486. The decrease in EE, despite the increase in initial Cur concentration, is likely attributed to saturation of the oil phase and the limited ability of the emulsifiers to sufficiently stabilize the excess active molecules. The significant difference in EE between samples prepared at aqueous-to-oil flow ratios of 50:1 and 100:1 highlights the critical importance of this parameter. The 50:1 flow ratio appears to provide optimal encapsulation regardless of the initial Cur concentration, possibly due to stronger interfacial stabilization during droplet formation. Conversely, formulations prepared at a 100:1 flow ratio, although yielding smaller droplet sizes, exhibited lower EE—a factor that should be balanced according to the final pharmaceutical objectives.

The Z-potential values in all samples remained below −40 mV, indicating high kinetic stability of the NEs. The maintenance of a high absolute Z-potential value is associated with the presence of the emulsifiers RH40 and lecithin, which create a surface barrier preventing nanodroplet coalescence, as previously noted. Furthermore, the consistency of the Z-potential regardless of the initial Cur amount suggests that the active compound is incorporated within the droplet core without affecting the interfacial structure between the aqueous and oil phases. Regarding PDI, results indicate an increase in the index with higher initial Cur content, attributed to a broader droplet size distribution due to altered stabilization dynamics. Notably, the 50Cur_50_1 sample exhibited a PDI of 0.547, suggesting a relatively wide size distribution, bordering on the upper acceptable limits for pharmaceutical applications. PDI values in systems with an aqueous-to-oil flow ratio of 100:1, were slightly lower, which may be explained by better dispersion and homogenization of the two phases, as well as the formation of a more stable interfacial layer in these more dilute preparations.

Stability studies of the Cur-loaded NEs showed no evidence of macroscopic phase separation in any sample throughout the testing period (Figure 7), providing an initial indication of their long-term stability. Nevertheless, subsequent physicochemical analyses (droplet size, Z-potential, and PDI) revealed notable variations that were dependent on both formulation composition and storage conditions.

Regarding the droplet size (Figure 8a), sample 25Cur_50/1 exhibited the smallest overall size variation, starting from 172 nm (day 1) and reaching 189 nm at RT (21st day). In contrast, 25Cur_100/1 started at 181 nm but showed stronger fluctuations, reaching 285 nm (at RT, 21st day), indicating lower stability at RT. Even more pronounced were the size increases in the samples with 50 mg of Cur. The sample 50Cur_50/1 increased from a droplet size of 240 nm up to 365 nm (at RT, 21st day), while the sample 50Cur_100/1 started at 194 nm and reached 326 nm. This increase in size correlates both with the higher concentration of the active substance and the system’s inability to maintain interfacial stability under higher loading and/or elevated temperature conditions (RT vs. +4 °C).

The results for the Z-potential (Figure 8b) reveal the relative stability of the surface electrostatic balance, despite the variations in size. All samples maintained absolute values greater than 37 mV throughout the storage period. These values are considered sufficient for the kinetic stability of the NEs due to the significant repulsion generated between the formed nanodroplets. Specifically, the Z-potential for sample 25Cur_50/1 decreased from −44.11 mV on the first day to −39.1 mV on the 21st day (in refrigeration), without a significant impact on its size. Similarly, sample 50Cur_100_1 initially exhibited the highest Z-potential (−45.1 mV) and maintained relatively high values (−39.7 mV) until the end of the study, despite the significant increase in diameter.

The small fluctuations in Z-potential are not always accompanied by size stability, indicating that mechanisms such as Ostwald ripening or the redistribution of Cur in the oil phase may be responsible for the increases in size, even when electrostatic stability remains high. The analysis of the stability of the PDI values (Figure 8c) also provided significant indications regarding the systems’ ability to maintain monodispersity. For example, the sample 25Cur_50_1 shifted from a PDI of 0.232 to 0.503, a value indicating a transition from monodispersity to polydispersity. A similar trend was observed in the sample 25Cur_100_1, which showed the highest final PDI value (0.651) and also repeatedly exhibited values greater than 0.6 in intermediate measurements, indicating an unstable droplet size distribution. Conversely, the sample 50Cur_50_1, despite having a larger initial diameter, showed a milder increase in PDI, from 0.547 to 0.511 (in refrigeration, 21st day). The same was observed in the sample 50Cur_100_1 (from 0.446 to 0.559), suggesting that samples with higher amounts of Cur maintained their size distribution more efficiently, possibly due to the reduced mobility of molecules within the oil phase.

Studying the color of emulsions aids in understanding their microstructure. The relative proportions of light transmitted and reflected at different wavelengths depend on the scattering and absorption by each emulsion. Light scattering and absorption are influenced by the size, concentration, and distribution of the droplets, as well as the refractive indices of the dispersion medium and the dispersed substance, and naturally by the wavelengths contained in the incident light [40,41]. The color of emulsions results from the way light is reflected and transmitted. To better determine this change in the color of the systems,

Table 6. L*a*b* color values of Cur-loaded and blank NEs.

Sample	E	L*	a*	b*
25Cur_50_1	124.67 ± 0.56	90.01 ± 0.44	−21.35 ± 0.11	83.6 ± 0.33
25Cur_100_1	128.14 ± 0.23	90.45 ± 1.11	−20.97 ± 0.57	88.33 ± 0.58
50Cur_50_1	143.97 ± 1.26	97.12 ± 0.83	−21.81 ± 0.91	104.03 ± 0.22
50Cur_100_1	137.22 ± 1.48	97.19 ± 0.79	−21.46 ± 0.74	94.48 ± 0.19
Blank NE	66.5 ± 0.79	66.17 ± 2.12	−1.49 ± 0.99	−3.13 ± 0.24

presents the color parameters measured for each NE. NEs containing 50 mg of Cur exhibited a slightly stronger yellow color compared to those containing half the amount (25 mg), while the latter were more transparent, and thus less bright, resulting in a lower intensity of white color. Compared to the initially prepared blank NEs, the Cur-loaded systems showed significantly higher L* values (90.01 to 97.19), indicating a brighter appearance, whereas the blank NEs displayed L* values around 66, corresponding to a much darker appearance. These results indicate that the encapsulation of Cur within the nanoemulsion droplets alters the light scattering and absorption properties, resulting in increased brightness and enhanced yellow coloration. This enhanced color intensity is attributed to the intrinsic yellow pigment of Cur and its uniform dispersion within the nanoscale droplets, which modulates the optical path and interaction with light, thereby improving the overall visual appearance of the emulsions.

3.5. Antioxidant Capacity of Free Radicals

Cur, as a natural polyphenolic compound, is characterized by strong antioxidant activity, primarily based on its ability to neutralize free radicals. However, its stability under light exposure is limited, especially due to UV radiation, which can lead to molecular degradation and loss of bioactivity. In this context, the present study investigated the effect of 24-h UV radiation exposure on the DPPH inhibition activity of Cur-loaded NEs (Figure 9).

The statistical analysis of the data showed that the differences between the measurements before and after irradiation were not significant (p > 0.05), showing the great potential of NEs as topical delivery system of bioactive compounds. The hydrophobic oil phase and the emulsifiers (such as RH40 and soy lecithin) contribute to reducing the amount of Cur available for degradation by UV. Additionally, the retention of active phenolic and carbonyl groups in the degradation products supports the preservation of part of the bioactivity of the compound, even after chemical modification of the original molecule. Antioxidant capacity of Cur was significantly higher (p < 0.05) than that of the blank NE, although it was lower compared to the pure, non-encapsulated active substance. This indicates that the antioxidant capacity of Cur is retained, albeit to a reduced extent, after encapsulation in the nanoemulsions. The reduced antioxidant activity of nano-encapsulated Cur compared to the free form has also been confirmed in other studies. For example, Donsi et al. [42] demonstrated that Cur encapsulated in solid lipid nanoemulsions exhibited significantly lower antioxidant activity than the free substance (0.996 ± 0.07 mM versus 2.504 ± 0.06 mM ascorbic acid equivalents). Similarly, Sari et al. [13] recorded a significant decrease in the activity of the nano-encapsulated Cur form as compared to the pure Cur (3.33 ± 0.02 mM Trolox/mg versus 3.53 ± 0.11 mM Trolox/mg).

3.6. In Vitro Permeation of Cur-Loaded NEs

The permeability of Cur through a synthetic membrane (cellulose-based) was evaluated using Franz diffusion cells in order to investigate the ability of the NEs to facilitate the transdermal permeation of Cur. The parameters studied included various Cur concentration (i.e., 0.06% and 0.13% w/w) and various phase flow ratios during the preparation of the NEs (i.e., 50:1 and 100:1). The five formulations examined were (1) 25Cur_50_1, (2) 25Cur_100_1, (3) 50Cur_50_1, (4) 50Cur_100_1, and (5) the pure active substance (Cur) in the form of an aqueous dispersion. The results are presented in Figure 10.

Initially, at the first time point of 2 h, the highest permeability (although marginally) was recorded for the sample 25Cur_50_1 (with 2851.3 μg/cm²), followed by 25Cur_100_1 (2753.2 μg/cm²) and pure Cur (2100.0 μg/cm²). In contrast, the samples 50Cur_100_1 and 50Cur_50_1 showed lower values (2036.6 and 1629.3 μg/cm², respectively). The pronounced permeation of Cur from sample 25Cur_50_1 at the initial stage suggests possible surface incorporation of Cur into the droplets or a loose NE structure, which allowed faster initial diffusion of the active ingredient. Conversely, the lower performance of systems containing 50 mg indicates a probable gradual diffusion from a more compact or denser oily phase.

At 4 h, the samples 25Cur_50_1 and 25Cur_100_1 maintained high permeation levels. The 50Cur_100_1 sample displayed a remarkable increase (reaching 3665.9 μg/cm²), whereas 50Cur_50_1 exhibited some relative limitation. The pure compound showed no significant change, indicating that although free (i.e., non-encapsulated), the pure Cur initially achieves relatively good permeation, followed by saturation—likely due to precipitation or, more probably, its low solubility in the chosen dissolution medium. This observation further supports that NE can enhance the permeation of poorly soluble active compounds like Cur.

Finally, at 8 and 24 h, the formulations 50Cur_100_1, 25Cur_50_1, and 25Cur_100_1 emerged as the most effective, exhibiting high permeability values. In contrast, the 50Cur_50_1 sample and pure Cur showed significant restriction in the amount of substance capable of permeating through the membrane. From the analysis of these results, it is evident that the phase ratio during NE preparation and the initial Cur concentration exert a complex influence on substance permeation. The 50Cur_100_1 sample combines a high payload with a more diluted aqueous phase, providing a sufficient reservoir of active ingredients and controlled release. Conversely, 50Cur_50_1, despite having the same Cur amount, significantly underperforms in permeation, possibly due to differences in distribution or higher lipid phase density. Additionally, the observation that 25Cur_50_1 outperforms 25Cur_100_1 at most time points suggests that the lower flow ratio (50:1) may enhance faster permeation of the active ingredient. Moreover, the superiority of encapsulated forms within NEs compared to free Cur is clearly demonstrated and further supported by the fact that the permeation profile of the pure substance stabilizes and declines after 4–8 h. In contrast, NEs exhibit an increasing permeation trend even at 24 h, offering significant potential for achieving Cur’s sustained release.

3.7. In Vitro Permeation of Cur from Hydrogels

The ability of Cur NEs to permeate through synthetic membranes that simulate human skin (Strat-M^®) was evaluated in vitro, when incorporated into the suitable hydrogels. The results of the tests, as presented in Figure 11, demonstrate significant differences in the rate and the total amount of Cur that permeated through the membrane within the first 6 h. Specifically, the formulation containing the 50Cur_50_1 NEs exhibited a gradual and steady increase in permeability, reaching a maximum of ~9 μg/cm² at the end of the 6th hour. In contrast, the formulation with the 25Cur_50_1 NEs, although it performed better during the first 2 h (2.25 μg/cm² compared to 1.745 μg/cm² for the 50Cur_50_1 sample), subsequently showed a downward trend and recorded a permeation of 3.125 μg/cm² at the same time point. The superior performance of the system containing the 50Cur_50_1 NEs is likely explained by the higher content of encapsulated Cur, and the greater ability of the oily core to retain the active substance within the nanodroplet. Furthermore, this result confirms that the pharmaceutical form of the final product (i.e., the incorporation into the hydrogel) does not negatively affect the permeability of the NEs but rather can facilitate it through the stabilization of their structure and the targeted, uniform distribution on the application surface.

4. Conclusions

In the current work, the successful encapsulation of antioxidant Cur in NEs was conducted for the first time using droplet microfluidics. NEs loaded with Cur were prepared at two different concentrations (0.06 and 0.13% w/w) using RH40 and soy lecithin as surfactant and co-surfactant. After Cur was encapsulated, the NEs were stored over 4 weeks in +4 °C and RT without the need for pasteurization and preservatives. No evidence of macroscopic phase separation in any NE was observed throughout this period. The tested NEs demonstrated excellent stability (Zeta potential < −40 mV), small droplet size, from 172 to 240 nm, low PDI < 0.55, and EE of up to 50.4%. Furthermore, the incorporation of NEs effectively preserved the antioxidant activity of Cur after 24 h of UV exposure, owing to the protection and retention of active phenolic groups within the encapsulated NE systems. In vitro permeability studies through membranes simulating human skin for the drug-loaded NEs incorporated in suitable hydrogels, revealed a maximum of ~9 μg/cm² at 6 h for the sample containing the 50Cur_50_1 NE, unlike a permeation of 3.125 μg/cm² at the same time point for the 25Cur_50_1-containing hydrogel. This difference was probably due to the higher content of encapsulated Cur and the greater ability of the oil (coconut) to retain Cur within the prepared nanodroplets. Therefore, the results obtained provide critical insights that can guide the formulation and design of Cur NEs based on the use of the microfluidics technology, enhancing their suitability and effectiveness for use in commercial cosmetic products. It is thus concluded that the nanoencapsulation of highly lipophilic and unstable compounds via a modern NE formulation strategy, such as droplet microfluidics proposed by the current study, is an effective way to increase the hydrophilicity, bio-accessibility and to protect them from chemical degradation during storage, a significant first step towards their translation to new and more effective cosmeceutical products.