Effect of Addition of PVA/PG to Oil-in-Water Nanoemulsion Kojic Monooleate Formulation on Droplet Size: Three-Factors Response Surface Optimization and Characterization

An oil in water (O/W) nanoemulsion formulation containing kojic monooleate (KMO) in thin film system was developed. Response surface methodology (RSM) was used to optimize and analyzed the effect of three variables, namely concentration of polyvinyl alcohol (PVA) (20–30% w/w), concentration of propylene glycol (PG) (1–10% w/w), and shear rate of high shear homogenizer (3000–9000 rpm) on droplet size as a response, while other compositions remained constant such as KMO (10.0% w/w), Tween 80 (3.19% w/w), castor oil (3.74% w/w), xanthan gum (0.70% w/w), and germall plus (0.7% w/w, PG (and) diazolidinyl urea (and) iodopropynyl butylcarbamate). The optimized KMO nanoemulsion formulation with desirable criteria was PVA (27.61% w/w) and PG (1.05% w/w), and shear rate (8656.17 rpm) with a predicted droplet size (110.21 nm) and actual droplet size (105.93 nm) with a residual standard error (RSE) of less than 2.0% was obtained. Analysis of variance (ANOVA) showed that the fitness of the quadratic polynomial fit the experimental data with a F-value of 65.30, p–value of p < 0.0001, and a non-significant lack-of-fit. The optimized KMO formulation shows the desired criteria of the thin film system and the physicochemical properties (Zeta potential −37.37 mV, PDI 0.13, pH 4.74) and stability at four different conditions indicate its suitability for cosmeceutical applications.


Introduction
A derivative of kojic acid, kojic monooleate (KMO) is non-toxic, soluble in oil, and enhances stability to pH, temperature, and storability compared to kojic acid [1]. KMO also possesses higher 2, 2-diphenyl-1-picrylhydrazyl (DPPH) inhibition, better mushroom tyrosinase inhibition, and B16F1 melanoma tyrosinase inhibition as well as less cytotoxicity compared to kojic acid [2], hence a formulation containing KMO was developed for cosmeceutical application in this study.
To achieve a desired effect on skin, penetration of the active compound into the skin after its topical application is required [3]. Hence, the size of the droplet plays a major role in determining the effect of a formulation on skin. Nanoemulsion drug delivery systems such as lipid based formulations

Selection of Solvent
A few solvents (propylene glycol (PG), polyethylene glycol 6000 (PEG 6000), polyethylene glycol 10,000 (PEG 10,000), polyethylene glycol 20,000 (PEG 20,000), isopropyl myristate, and glycerol) were incorporated individually at 1% (w/w) into the formulation to enhance the drying time of the thin film. The stability (centrifugal force), dosage form, film formation, peelability, and drying time of the thin film were investigated.

Formulation of Kojic Monooleate Nanoemulsion Formulation in Thin Film System (TFS)
PVA was dispersed in deionized water (1:1 ratio) and heated at 80 • C to dissolve the PVA. For the oil phase, 10.0% (w/w) KMO was mixed with 3.37% (w/w) castor oil. For the aqueous phase, Tween 80 (3.19% w/w), xanthan gum (0.70% w/w), deionized water (42.74-61.74% w/w), PVA (20-30% w/w), and PG (1-10% w/w) were mixed together. Both aqueous and oil phases were stirred continuously while heated at 30 • C before being subjected to sonication for 20 min at 70 • C using an ultrasonic bath sonicator (Power Sonic 405, Hwashin Technology Co., Seoul, Korea). After sonication, the aqueous phase was homogenized while the oil phase was added drop by drop during the homogenization process, which took place for 15 min at 6000 rpm using a high shear homogenizer (T25 digital; IKA-Werk, GmbH & Co. KG, Staufen im Breisgau, Germany) at room temperature. Then, the mixture was further homogenized for 3 h at 250 rpm using a low-shear homogenizer (RW20 digital; IKA-Werk). Finally, 0.7% (w/w) of liquid germall plus was added into the mixture and further homogenized for 15 min at 250 rpm.

Experimental Design
Three factors were utilized to study the interaction effects of polyvinyl alcohol (A), propylene glycol (B), and shear rate of a high shear homogenizer (C) on the droplet size. Tween 80, KMO, castor oil, xanthan gum, and deionized water were kept constant [22]. Table 1 shows the constraints of the independent variable proportions. The experimental design was developed using Design Expert software (version 7.1.5; Stat Ease Inc., Minneapolis, USA) presuming a quadratic model with interactions among the variables. Following the experimental design, a total of 20 runs were tested and statistically evaluated. Six replicates of the center point were performed to reduce the impact of noise and to provide the number of degrees of freedom needed for an adequate statistical evaluation of the model. Variables resulting in the minimum droplet size were selected as the optimum condition to develop the KMO formulation. Analysis of variance (ANOVA) was done to recognize the effect of single factors, binary interactions, and quadratic terms toward droplet size and to evaluate the significant differences among the independent variables in terms of the lack-of-fit test [23]. The best model can be achieved with the probability value (p-value) that was significant (p < 0.05). The equation to predict the droplet size in terms of coded factors can be calculated as in Equation (1): Cosmetics 2020, 7, 73 4 of 15 where Y is the particle size, and A, B, C represent the independent variables, PVA, PG, and shear rate (rpm), respectively.

Verification of Models
Verification of the model was performed to evaluate the competence of the predicted response values. Five random formulations with a different percentage of composition were used to validate the model. The actual and predicted droplet size values and percentage of the residual standard error (RSE) were calculated.

Film Formation
The films were formed on a glass slide following the previous method [10]. Film formation was studied and rated as complete or incomplete, uniform or non-uniform, and with or without precipitation of the compositions. The evaluation was observed in terms of transparency or opaque, sticky or dry, peelable or non-peelable.

Drying Time
The formulation was applied to the inner sides of the forearm of a volunteer. Then, a glass slide was placed on the film without pressure and the time was recorded. If there was no liquid visible on the glass slide, the film was considered dry [10]. However, if there was remaining liquid on the glass slide, the experiment was repeated with an increase in drying time. A good thin film should have minimum drying times to avoid a long waiting time for the consumer.

Viscosity
Viscosity is defined as the shear stress divided by the rate of shear strain [24]. The viscosity was determined using a rheometer. The viscosity of KMO formulation, control formulation, and KMO formulation without PVA were measured. The measurement was carried out with 4 • C/40 mm cone and plate geometries (gap of 0.100 mm) at 25 • C. The steady rheological behavior of the samples were analyzed at a controlled rate varying from 0.1 to 100 s −1 . The sample was allowed to stand for 10 min to achieve an equilibrium state prior to measurement. The experimental data were fitted to the power-law model as in Equation (2): ï = ky n−1 (2) where ï is the viscosity (Pa.s); y is the shear rate (s −1 ); and k and n are the consistency index and flow behavior index, respectively.

Droplet Size
A droplet size analyzer (Zetasizer Nano ZS90; Malvern Instruments, Malvern, UK) was used to measure the droplet size, stability, and size distribution of molecules in the KMO formulation using a dynamic light scattering method, scattered at an angle of 173 • at 25 • C. The samples was diluted with deionized water to obtain a desired concentration and filled into a folded capillary cell (DTS1070; Malvern Instruments) [22]. The count rate was kept between 100 and 300 kcps. The droplet size (20-200 nm) was analyzed based on the intensity weighted distribution (Z-average), while the Zeta potential was determined based on the measurement of the electrophoretic mobility of dispersed particles in a charged field. The size and morphology of the KMO formulation were investigated using a microscopy technique by transmission electron microscopy (TEM, JEOL JEM-1400Flash; JEOL, Tokyo, Japan). The sample was homogenized in deionized water. A formvar coated copper grid was added on top of a drop of diluted sample and left at room temperature (25 • C) for 3 min. It was then stained using 2% phosphotungstic acid for 2 min and air dried prior to analysis.

pH
The pH of the KMO formulation was determined using a Delta 320 pH meter (Mettler Toledo, Columbus, OH, USA). The measurements of the pH of the formulation were performed in triplicate and the average values were calculated. Calibration with standard buffer solutions at pH 4.00, 7.00, and 10.00 was performed before pH measurements were taken [25].

Stability Study
The stability study of the KMO formulation involved stability under centrifugation and thermal stability under storage for 28 days (at room temperature; 25 ± 1.0 • C with exposure to sunlight, at room temperature; 25 ± 1.0 • C, without exposure to sunlight 45 ± 1.0 • C and 4 ± 1.0 • C). Duplicate samples of the optimized formulation were stored in opaque polyethylene bottles and submitted to the mentioned storage conditions in order to evaluate the preliminary physicochemical stability. Each sample was analyzed weekly according to various parameters, namely organoleptic characteristic, applicability, film-forming performance, and particle size. For stability under centrifugation, the sample was subjected to heating at 50 • C for 5 min before centrifugation for 15 min at 3000 rpm. The stability of the sample was observed.

Combination of Polymer and Solvent to Develop KMO Formulation in TFS
Among the polymers and solvents tested, PVA and PG were chosen to develop the TFS, but further improvement is needed to make the thin film peelable and resistant to pressure. Therefore, PVA and PG were combined in a formulation at various concentrations to find the range of concentration of PVA and PG prior to the optimization process. It was found that the addition of PG was only appropriate at 1-10% (w/w), since at concentrations of 12.5% and 15%, the formulations were not stable under centrifugal force. However, the addition of PG does not affect the peelable and resistance of thin film toward pressure. Then, the concentration of PG was fixed at 1% and 10% to study the effect of the concentration of PVA on TFS. The concentration of PVA does not affect the stability of the formulation, it only affected the TFS characteristics. At a higher concentration of PVA (20-30%), the thin film was peelable and showed resistance to pressure. This finding was in agreement with [26] as they mentioned that PVA plays a major role as a plasticizer and film former. Many studies using PVA as a film former to develop TFS included a charcoal peel off film by using 14% (w/w) PVA and 3% (w/w) glycerine as a film former and solvent [25]. Another example was a green clay and aloe vera peel-off film developed using 13% (w/w) PVA and 10% (w/w) carbomer as the plasticizer [23].

Optimization Using the Response Surface Methodology (RSM)
The optimized droplet size of the KMO nanoemulsion formulation was obtained experimentally using the RSM optimization method as tabulated in Table 2. The table shows that the minimum droplet size obtained was 111.1 nm (run no. 17). The actual droplet size values were mostly in good agreement with the predicted values. The analysis of variance (ANOVA) results for the effect of three variables were performed to investigate the suitability and significance of the model ( Table 3). The model was significant with F-value (65.30) and p-value (<0.0001), and the lack-of-fit was not significant toward Cosmetics 2020, 7, 73 6 of 15 the response (droplet size). There is only a 0.01% chance that a "Model F-Value" could occur due to noise. In this case A, B, C, BC, and C 2 were significant model terms. The R 2 (0.9866) was in reasonable agreement with the adjusted R 2 (0.9715), as shown in Table 4. Adequate precision measures the signal to noise ratio. A ratio of 28.498 (>4 is desirable) indicates an adequate signal. Table 2. Experimental data and the actual and predicted values of the droplet size of the nanoemulsion obtained from the RSM experimental design model.  The final equation to predict the droplet size in terms of actual factors can be calculated as in Equation (3):

Run
where A, B, C represent the independent variables PVA, PG, and shear rate, respectively. In this study, the minimum droplet size was the main consideration as nanosizes with a range between 20 and 200 nm are more desirable for a cosmetic product. Smaller droplet sizes are able to penetrate through the stratum corneum of the skin easily. Therefore, it is vital to maintain the nanosize of the KMO formulation to enhance the penetration of KMO to the skin, without disrupting normal tissues. Therefore, Figures 1-3 show the three-dimensional surface graphs on droplet size. Based on Figure 1, increasing the concentration of PVA will decrease the droplet size. This is because of the ability of PVA to act as a surfactant and capping agent, thus making them capable of modifying the surface of particles and prevent the growth of the particles [27]. The presence of the PVA in the polymerization mixture has strongly influenced the morphology by providing better order and reducing the aggregations of the particles [28]. The final equation to predict the droplet size in terms of actual factors can be calculated as in Equation (3): where A, B, C represent the independent variables PVA, PG, and shear rate, respectively. In this study, the minimum droplet size was the main consideration as nanosizes with a range between 20 and 200 nm are more desirable for a cosmetic product. Smaller droplet sizes are able to penetrate through the stratum corneum of the skin easily. Therefore, it is vital to maintain the nanosize of the KMO formulation to enhance the penetration of KMO to the skin, without disrupting normal tissues. Therefore, Figures 1, 2 and 3 show the three-dimensional surface graphs on droplet size. Based on Figure 1, increasing the concentration of PVA will decrease the droplet size. This is because of the ability of PVA to act as a surfactant and capping agent, thus making them capable of modifying the surface of particles and prevent the growth of the particles [27]. The presence of the PVA in the polymerization mixture has strongly influenced the morphology by providing better order and reducing the aggregations of the particles [28]. Based on Figure 2, the increment of the shear rate caused a decrease in the droplet size. This occurrence has been reported in a previous study that mentioned that the increment of the shear rate speed would decrease the droplet size and increase the size uniformity [29]. At high shear homogenization, the turbulent forces are large, thus able to disrupt the bonds that hold the particles  Based on Figure 2, the increment of the shear rate caused a decrease in the droplet size. This occurrence has been reported in a previous study that mentioned that the increment of the shear rate speed would decrease the droplet size and increase the size uniformity [29]. At high shear homogenization, the turbulent forces are large, thus able to disrupt the bonds that hold the particles together, and thus deform the flocs and eventually disrupt them. At low shear homogenization speed, the turbulent forces are not large enough to disrupt the bonds and so the droplets can act like flocs with a fixed size and shape [30]. together, and thus deform the flocs and eventually disrupt them. At low shear homogenization speed, the turbulent forces are not large enough to disrupt the bonds and so the droplets can act like flocs with a fixed size and shape [30]. In contrast, increasing the concentration of PG will cause the increment of the droplet size ( Figure 3). The addition of PG reduced the formation region of the microemulsion and reduced the isotropic region, in which only a small amount of oil and water can be solubilized [31]. On the other hand, other finding have shown that when the concentration of PG was from 0 to 20%, the droplet size decreased slightly; at 30 to 40% PG, the droplet size decreased steeply; and at 40 to 50% PG, the droplet size was slightly increased [32]. Various findings regarding the effect of PG on droplet size may be due to the property of PG, which have two competitive effects: (i) one dissolved in water, which results in the reduction of interactions between droplets, thus lowering the droplet size and (ii) another dissolved in oil, which resulted in the increment of interactions between droplets, making the droplet size bigger [33]. As the increment of PG caused the droplet size to be bigger in this study, it can be concluded that the second effect dominates the first one under the experimental conditions and compositions used. In contrast, increasing the concentration of PG will cause the increment of the droplet size ( Figure 3). The addition of PG reduced the formation region of the microemulsion and reduced the isotropic region, in which only a small amount of oil and water can be solubilized [31]. On the other hand, other finding have shown that when the concentration of PG was from 0 to 20%, the droplet size decreased slightly; at 30 to 40% PG, the droplet size decreased steeply; and at 40 to 50% PG, the droplet size was slightly increased [32]. Various findings regarding the effect of PG on droplet size may be due to the property of PG, which have two competitive effects: (i) one dissolved in water, which results in the reduction of interactions between droplets, thus lowering the droplet size and (ii) another dissolved in oil, which resulted in the increment of interactions between droplets, making the droplet size bigger [33]. As the increment of PG caused the droplet size to be bigger in this study, it can be concluded that the second effect dominates the first one under the experimental conditions and compositions used.
together, and thus deform the flocs and eventually disrupt them. At low shear homogenization speed, the turbulent forces are not large enough to disrupt the bonds and so the droplets can act like flocs with a fixed size and shape [30]. In contrast, increasing the concentration of PG will cause the increment of the droplet size ( Figure 3). The addition of PG reduced the formation region of the microemulsion and reduced the isotropic region, in which only a small amount of oil and water can be solubilized [31]. On the other hand, other finding have shown that when the concentration of PG was from 0 to 20%, the droplet size decreased slightly; at 30 to 40% PG, the droplet size decreased steeply; and at 40 to 50% PG, the droplet size was slightly increased [32]. Various findings regarding the effect of PG on droplet size may be due to the property of PG, which have two competitive effects: (i) one dissolved in water, which results in the reduction of interactions between droplets, thus lowering the droplet size and (ii) another dissolved in oil, which resulted in the increment of interactions between droplets, making the droplet size bigger [33]. As the increment of PG caused the droplet size to be bigger in this study, it can be concluded that the second effect dominates the first one under the experimental conditions and compositions used.  The model was validated and the actual and predicted response values of five randomized formulations were evaluated in Table 5. No significant difference between the actual and predicted Cosmetics 2020, 7, 73 9 of 15 response values was detected, which indicates that the fitness model was developed. Based on the validation set, % RSE was <2.00, which indicated no significant difference between the actual and predicted values and therefore demonstrated that a good model was obtained. The results of the predicted values were studied by calculating the RSE as shown in Equation (4): The optimized formulation was developed to obtain a minimum droplet size of the nanoemulsion ( Table 6). Based on Table 7, an optimized formulation with the composition of PVA (27.61 w/w), PG (1.05% w/w), and shear rate (8656.17 rpm) was suggested with a predicted droplet size of 110.21 nm and actual droplet size of 105.93 ± 0.21 nm. The desirability of the optimum KMO formulation was 1.00. This showed that the quality of the formulation was agreeable. A formulation with a droplet size of 105.93 will theoretically be able to penetrate the skin effectively to work on skin. At this range of droplet size, it is in good agreement with the previous study. A similar result was obtained by Azhar et al. (2018), with 110.01 nm as the minimum droplet size [22].

Film Formation
A uniform and complete thin film system was formed without the precipitation of the polymer on a glass slide. The KMO formulation had an opaque appearance on a glass slide before drying, but after drying, the film formed was transparent, dry, and non-sticky. The film was peelable and after the film was peeled from the glass slide, there was no residue of film left.

Drying Time
A good TFS should have a minimum drying time to avoid a long waiting time for the patient [10]. Initially, the drying time of the KMO formulation containing 1% (w/w) PVA and 1% (w/w) PG was 15 min, but as the percentage of PVA increased after optimization to 27.61% (w/w), the drying time became approximately 8 min. This indicates that the amount of plasticizer decreased the drying time. In contrast, a formulation containing soybean extract as active and 17% PVA as the plasticizer had a drying time range between 27 to 37 min [26]. This showed that the optimized formulation had a better drying time as the waiting time of patients would be less compared to the previous study.

Viscosity
Shear-dependent behavior is one of the principal characteristics of non-Newtonian behavior observed in heterogeneous systems. There are two types of shear-dependent: shear thinning liquid shows a decrease in viscosity as the shear rate increases, while shear thickening liquid has an increasing viscosity with an increase in the shear rate. A power law index, n > 1 denotes that the sample is shear thickening; while n < 1 denotes that the sample is shear thinning and n = 1 shows a Newtonian, viscous behavior. Table 8 shows that the KMO formulation clearly displayed a shear thinning behavior as the flow index of n < 1 using the power law model. The shear thinning (pseudo-plastic) formulation exhibited a low viscosity fluid behavior (low resistance) to flow when tested under high shear conditions. The cosmetic and pharmaceutical industries usually demand shear thinning behavior for topical formulation products as they are easy to apply.  Table 9, the KMO formulation had a droplet size of 105.93 ± 0.21 nm, and a polydispersity index (PDI) 0.13 ± 4.50 × 10 −3 and Zeta potential value of −37.37 ± 0.86 mV were obtained. The nanoemulsion particle size should be between 20-200 nm for cosmetic applications [7]. The droplet size of 105.93 ± 0.21 nm indicated that the KMO formulation would have better penetration of the KMO into the skin layers. Polydispersity (PDI) analyzes the size distribution, demonstrating the uniformity of the droplet size [34]. The good monodisperse system should have a PDI value lower than 0.5 [22] or lower than 0.25 [34]. As the PDI value of the KMO formulation was 0.13, this indicated a good monodisperse system and is expected to be a stable emulsion. Zeta potential analyzes the electrokinetic potential of a particle and determines the stability of the formulation system. The formulation is considered stable when the value is more than +25 mV or lower than −25 mV. The KMO formulation with a Zeta potential value of −37.37 ± 0.86 mV demonstrated a higher electrostatic repulsion, which indicated that the nanoemulsion system was stable.  Figure 4 represents the TEM image of the KMO formulation in the nanoemulsion size. The nanoemulsion was distributed almost uniformly and homogenously in the system. The droplet size analyzed from the TEM corresponded with the size obtained from the Zetasizer analysis. However, there were some large droplet sizes, as can be seen in Figure 4, which may be due to the Ostwald ripening occurrence, causing the enlargement of the droplet size.

Property
KMO Formulation % RSE Droplet size (nm) 105.93 ± 0.21 0.11 PDI 0.13 ± 4.50 x 10 −3 -Zeta potential (mV) -37.37 ± 0.86 -Conductivity (µS/cm) 7.47 ± 4.05 x 10 −3 -pH 4.74 ± 0.02 - Figure 4 represents the TEM image of the KMO formulation in the nanoemulsion size. The nanoemulsion was distributed almost uniformly and homogenously in the system. The droplet size analyzed from the TEM corresponded with the size obtained from the Zetasizer analysis. However, there were some large droplet sizes, as can be seen in Figure 4, which may be due to the Ostwald ripening occurrence, causing the enlargement of the droplet size.

pH and Conductivity
The pH values and conductivity of the KMO formulation are shown in Table 9. The pH measurement is vital as it needs to be between a range of 4.0 and 7.0 to be compatible with human skin [35]. However, others have reported the pH of skin normally ranges from 4 to 6 [36]. The pH value of the KMO formulation was 4.74 ± 0.02 (Table 9), indicating its suitability for use as a topical formulation.
The conductivity value refers to the number of free ions and water present in the system. From conductivity value, type of nanoemulsions can be investigated as either O/W or W/O as high conductivity indicates that the aqueous phase is the continuous phase of the system, whereas the oil phase is nonconductive. Basically, the external phase determines the emulsion's conductivity. According to Mohamed et al. (2017), they mentioned that the O/W emulsion is conductive when the conductivity is >0.00 S⋅cm −1 , whereas the W/O emulsion is nonconductive when the value is ∼0.00 S⋅cm −1 [37]. The KMO formulation had a conductivity of 7.47 ± 4.05 x 10 −3 , indicating an O/W emulsion.

Stability Study
The KMO formulation remained in the homogenous phase even after 28 days. The actual color of the KMO formulation changed from white to a pale yellowish emulsion at day 28, stored at 45 °C, which may be due to the reaction of any substance or ingredient in a formulation with the container

pH and Conductivity
The pH values and conductivity of the KMO formulation are shown in Table 9. The pH measurement is vital as it needs to be between a range of 4.0 and 7.0 to be compatible with human skin [35]. However, others have reported the pH of skin normally ranges from 4 to 6 [36]. The pH value of the KMO formulation was 4.74 ± 0.02 (Table 9), indicating its suitability for use as a topical formulation.
The conductivity value refers to the number of free ions and water present in the system. From conductivity value, type of nanoemulsions can be investigated as either O/W or W/O as high conductivity indicates that the aqueous phase is the continuous phase of the system, whereas the oil phase is nonconductive. Basically, the external phase determines the emulsion's conductivity. According to Mohamed et al. (2017), they mentioned that the O/W emulsion is conductive when the conductivity is >0.00 µS·cm −1 , whereas the W/O emulsion is nonconductive when the value is ∼0.00 µS·cm −1 [37]. The KMO formulation had a conductivity of 7.47 ± 4.05 × 10 −3 , indicating an O/W emulsion.

Stability Study
The KMO formulation remained in the homogenous phase even after 28 days. The actual color of the KMO formulation changed from white to a pale yellowish emulsion at day 28, stored at 45 • C, which may be due to the reaction of any substance or ingredient in a formulation with the container at high temperature. Changes of color are not likely to happen due to enlargement of the droplet size as the droplet size measured was still below 200 nm.
The evaluation of the optimized formulation on the creaming or sedimentation rate can be examined by the centrifugation test by analyzing the presence of phase separation. The KMO formulation was analyzed and it shows no phase separation after centrifugation at 4000 rpm for 15 min. This indicates the good stability of the KMO formulation. Figure 5 shows that the droplet size of the emulsion slightly increased from 107.35 to 122.95 nm at 45 • C, however, the formulation at all conditions was able to maintain its nanosize range between 20-200 nm. This shows the effectiveness of the xanthan gum in the nanoemulsion system to retain smaller droplet sizes and stability [22].
Cosmetics 2020, 7, x FOR PEER REVIEW 12 of 16 at high temperature. Changes of color are not likely to happen due to enlargement of the droplet size as the droplet size measured was still below 200 nm. The evaluation of the optimized formulation on the creaming or sedimentation rate can be examined by the centrifugation test by analyzing the presence of phase separation. The KMO formulation was analyzed and it shows no phase separation after centrifugation at 4000 rpm for 15 min. This indicates the good stability of the KMO formulation. Figure 5 shows that the droplet size of the emulsion slightly increased from 107.35 to 122.95 nm at 45 °C, however, the formulation at all conditions was able to maintain its nanosize range between 20-200 nm. This shows the effectiveness of the xanthan gum in the nanoemulsion system to retain smaller droplet sizes and stability [22]. Coalescence is where two or more droplets fuse together and result in bigger droplets in the system. Figure 6 shows the plotted graph of 1/r 2 vs. storage time (seconds) to determine the factor that affects the changes in droplet size over time. The graph with a linear relationship pattern suggests that the coalescence effect occurred in the system. However, based on the graph, all the droplet size values did not produce a linear pattern. Thus, the increase in droplet size over time was not created by coalescence.
Ostwald ripening is the product of enlargement of droplet size due to the diffusion of the oil phase through the aqueous phase and can be investigated by plotting r 3 vs. storage time (seconds). The KMO formulation under four different temperatures during 28 days of storage was affected by Ostwald ripening (Figure 7). A low Ostwald ripening effect can be seen during storage at 25 °C as the droplet size increased slowly, while higher Ostwald ripening effect was identified at 45 °C as the droplet size became larger, indicating the low stability of the nanoemulsion at a higher temperature. Coalescence is where two or more droplets fuse together and result in bigger droplets in the system. Figure 6 shows the plotted graph of 1/r 2 vs. storage time (seconds) to determine the factor that affects the changes in droplet size over time. The graph with a linear relationship pattern suggests that the coalescence effect occurred in the system. However, based on the graph, all the droplet size values did not produce a linear pattern. Thus, the increase in droplet size over time was not created by coalescence. Ostwald ripening is the product of enlargement of droplet size due to the diffusion of the oil phase through the aqueous phase and can be investigated by plotting r 3 vs. storage time (seconds). The KMO formulation under four different temperatures during 28 days of storage was affected by Ostwald ripening (Figure 7). A low Ostwald ripening effect can be seen during storage at 25 • C as the droplet size increased slowly, while higher Ostwald ripening effect was identified at 45 • C as the droplet size became larger, indicating the low stability of the nanoemulsion at a higher temperature.

Conclusions
In this study, a nanoemulsion formulation containing KMO in a thin film system was successfully developed. The droplet size of the KMO formulation was optimized using the RSM approach and characterized for its thin film system and physicochemical properties. The optimized KMO nanoemulsion formulation with desirable criteria was PVA (27.61% w/w), PG (1.05% w/w), and shear rate (8656.17 rpm) with predicted droplet size (110.21 nm) and actual droplet size (105.93 nm) with a residual standard error (% RSE) < 2.0% was obtained. Analysis of variance showed that the fitness of the quadratic polynomial fit the experimental data with F-value (65.30), a low p-value (P < 0.0001), and a non-significant lack-of-fit. The optimized KMO formulation showed the desired criteria of a thin film system with flexible and peelable characteristics, is non-sticky, and has a short drying time of 7 min 25 s, which shows its suitability for topical application as a peeling mask. The physicochemical properties (Zeta potential −37.37 mV, PDI 0.13, pH 4.74) and the stability of the KMO formulation under centrifugal force, and a stability study for 28 days at four different conditions of 4

Conclusions
In this study, a nanoemulsion formulation containing KMO in a thin film system was successfully developed. The droplet size of the KMO formulation was optimized using the RSM approach and characterized for its thin film system and physicochemical properties. The optimized KMO nanoemulsion formulation with desirable criteria was PVA (27.61% w/w), PG (1.05% w/w), and shear rate (8656.17 rpm) with predicted droplet size (110.21 nm) and actual droplet size (105.93 nm) with a residual standard error (% RSE) < 2.0% was obtained. Analysis of variance showed that the fitness of the quadratic polynomial fit the experimental data with F-value (65.30), a low p-value (p < 0.0001), and a non-significant lack-of-fit. The optimized KMO formulation showed the desired criteria of a thin film system with flexible and peelable characteristics, is non-sticky, and has a short drying time of 7 min 25 s, which shows its suitability for topical application as a peeling mask. The physicochemical properties (Zeta potential −37.37 mV, PDI 0.13, pH 4.74) and the stability of the KMO formulation under centrifugal force, and a stability study for 28 days at four different conditions of 4 • C, 25 • C exposed to sunlight, 25 • C protected from sunlight, and at 45 • C showed the physical properties for cosmeceutical applications.