2.3.1. Pseudo-Ternary Phase Diagrams and Emulsion Classification

The water titration method was used to construct phase diagrams to identify the type of structure that resulted following emulsification and to characterize the behavior of mixtures along dilution paths [30]. Preliminary studies were performed making mixtures of flaxseed oil and a surfactant mixture of (Tween® 80 and Span® 20) (1:1) *m*/*m* in ratios of 9:1, 4:1, 7:3, 3:2, 1:1, 2:3, 3:7, 1:4 *m*/*m* and Winsor I-type products with no isotropic regions were produced at 22 ± 2 ◦C and observed after 48 h of incubation Therefore, a surfactant, co-surfactant and co-solvent mixture was considered for further phase behavior investigations. Surfactants solutions of Tween® 80, Span® 20 and ethanol were mixed using a Genie two vortex mixer (Scientific Industries, Inc.™, Bohemia, NY, USA) at 800 rpm for 30 ± 2 s. Three surfactant mixtures *viz*., S1, S2 and S3 comprised of combinations of ethanol: Tween® 80: Span® 20 in; 0.5:1:1, 1:1:1 and 1.5:1:1 *m*/*m* ratios, respectively. The surfactant mixtures were then added to flaxseed oil to produce pseudo-binary solutions in 9:1, 4:1, 7:3, 3:2, 1:1, 2:3, 3:7, 1:4, 1:9 *m*/*m* ratios to produce surfactant and oil mixtures in Kimax® test-tubes (DWK Life Sciences, Hattenbergstr, Mainz, Germany). The pseudo-binary pre-concentrates mixtures contained no additional water. To minimize this effect, 500 mL of >98% ethanol was poured in a Schott Duran bottle (DWK Life Sciences, Hattenbergstr, Mainz, Germany) containing 300 g of 3A and 4A (1:1) *m*/*m* molecular sieve pellets from (B & M Scientific, Cape Town, South Africa), sealed and kept under air-conditioned laboratory room temperature for 7 days for further dehydration [31,32]. The ethanol was then degassed under vacuum with the aid of a Model A-2S Eyela aspirator degasser (Rikakikai Co., Ltd., Tokyo, Japan) and filtered through a 0.45-µm HVLP Durapore® membrane filter (Millipore® Corporation, Bedford, MA, USA) prior to use. Each of the ratios tested represent a dilution line from one to nine on the Gibbs phase triangle depicted in Figure 1. Water was added in 5% ± 1% increments to each pseudo-binary mixture following the titration chart summarized in Table 1 and after a 48-h incubation at room temperature, 22 ± 2 ◦C the regions of the phase diagram were identified and characterized for Winsor behavior visually prior to further characterization of pre-defined and identified formulation attributes. The titration chart and Gibbs triangle plots were developed using Triplot version 4.1.2 software (Todd A. Thompson, LA, USA). Points that were located within the phase diagram were observed and evaluated against the Winsor phase behavior descriptions [33] as depicted in Figure 2 and the data were plotted as a phase diagram, a graphical plot using Triplot software spreadsheet (titration chart), example is shown in Table 1. The resultant pseudo-ternary phase diagrams for surfactant-mixtures S1, S2 and S3 have been reported to show a representative sample of the types of structures that form when different amounts of ethanol as used and these data are shown in the results. The mixtures for the eight pseudo-binary solutions produced were vortexed using a Genie two vortex mixer (Scientific Industries, Inc.™, Bohemia, NY, USA) and placed into Falcon® clear 24-well cell culture microplates with a lid from Corning®, Inc (Corning, NY, USA). The bottom surface was placed onto a Xerox WorkCenter 3655 scanner (Xerox™, Norwalk, CT, USA) with the top surface lid and sides covered with a clean white background paper. The transparency and turbidity of the phase diagram dilution-line ratio mixtures were visualized by the scanner, characterized for droplet size, PDI and zeta potential. The images for each of the pseudo-binary solutions and corresponding droplet size, PDI and ZP elucidated immediately (within 6 min) after vortex mixing are listed in Table S1 in the Supplementary Material, for the of analyses performed in triplicate. Conductivity of points P1 to P20, in Table 1, which represent each 5% *w*/*w* increment point from 0% water to 95% water approaching the water vertex along dilution-line 9 for surfactant-mixture 1 (S1) was measured using a FiveEasy™ F30 conductivity meter (Mettler Toledo, Greifensee, Switzerland). The electrical conductivity was used to classify the microstructure of the emulsions and establish if w/o or o/w emulsions had formed as o/w systems exhibit higher electrical conductivity than w/o emulsions and these data are reported in Table 1.

*Pharmaceutics* **2020**, *12*, x FOR PEER REVIEW 5 of 22

**Figure 1.** Pseudo-ternary phase diagram and scheme used for phase diagram plotting, showing dilution lines and areas in which electrical conductivity was tested. Adapted from [34]. **Figure 1.** Pseudo-ternary phase diagram and scheme used for phase diagram plotting, showing dilution lines and areas in which electrical conductivity was tested. Adapted from [34].


P8 250 2250 1350 3850 58.44 6.49 35.06 11.43 P9 250 2250 1675 4175 53.89 5.98 40.11 25.3 P10 250 2250 2050 4550 49.45 5.49 45.05 119.6 P11 250 2250 2500 5000 45.00 5.00 50.00 142.2 P12 250 2250 3050 5550 40.54 4.504 54.95 147.6 P13 250 2250 3750 6250 36.00 4.00 60.00 173.6 P14 250 2250 4625 7125 31.57 3.50 64.91 173.9 P15 250 2250 5875 8375 26.86 2.98 70.14 202 P16 250 2250 7500 10,000 22.50 2.50 75.00 261 P17 250 2250 9950 12,450 18.07 2.00 79.91 278 P18 250 2250 14,275 16,775 13.41 1.49 85.09 356 P19 250 2250 22,500 25,000 9.00 1.00 90.00 389 P20 250 2250 46,600 49,100 4.58 0.50 94.90 433

**Table 1.** Titration chart for use along each dilution line to plot phase diagrams with proportions of each component in the nanoemulsion with conductivity along dilution-line 9 for surfactant-mixture 1.

dilution lines and areas in which electrical conductivity was tested. Adapted from [34].

**Figure 2.** Winsor phase behavior used to assess phase diagram plots. **Figure 2.** Winsor phase behavior used to assess phase diagram plots.

#### **Table 1.** Titration chart for use along each dilution line to plot phase diagrams with proportions of each 2.3.2. D-Optimal Design and Statistical Optimization

component in the nanoemulsion with conductivity along dilution-line 9 for surfactant-mixture 1. **Water Addition Points on Dilution Line Oil mg S1 mg Water µL Total mg S 1 % Oil % Water % Conductivity µScm−<sup>1</sup>** P1 250 2250 0 2500 90.00 10.00 0.00 0.18 P2 250 2250 133 2633 85.45 9.49 5.05 0.2 P3 250 2250 280 2780 80.93 8.99 10.07 0.28 P4 250 2250 440 2940 76.53 8.50 14.96 0.45 P5 250 2250 625 3125 72.00 8.00 20.00 0.96 P6 250 2250 835 3335 67.46 7.49 25.03 1.67 P7 250 2250 1075 3575 62.93 6.99 30.06 7.76 P8 250 2250 1350 3850 58.44 6.49 35.06 11.43 P9 250 2250 1675 4175 53.89 5.98 40.11 25.3 P10 250 2250 2050 4550 49.45 5.49 45.05 119.6 P11 250 2250 2500 5000 45.00 5.00 50.00 142.2 P12 250 2250 3050 5550 40.54 4.504 54.95 147.6 P13 250 2250 3750 6250 36.00 4.00 60.00 173.6 P14 250 2250 4625 7125 31.57 3.50 64.91 173.9 P15 250 2250 5875 8375 26.86 2.98 70.14 202 P16 250 2250 7500 10,000 22.50 2.50 75.00 261 P17 250 2250 9950 12,450 18.07 2.00 79.91 278 P18 250 2250 14,275 16,775 13.41 1.49 85.09 356 P19 250 2250 22,500 25,000 9.00 1.00 90.00 389 Surfactant-mixture optimization studies were performed using a D-optimal design to elucidate the effect(s) of the proportion of individual components of the surfactant mixture viz. Span® 20, Tween® 80 and ethanol content on droplet size, PDI and ZP. The D-optimal design studies were performed with the aid of Design-Expert version 12.0 software (Stat-Ease, Inc., Minneapolis, MN, USA). The proportion of oil in the nanoemulsion was maintained at 10% *v*/*v* for dilution-line 9, which falls in the nanoemulsion region. Span® 20, Tween® 80 and ethanol were independent variables and droplet size, PDI and ZP were the responses monitored. The oral route of delivery is proposed for these nanoemulsions, therefore the proportion of ethanol used was maintained at the permissible levels for food content with the largest concentrations at 20% *m*/*m* [35–37]. Legislation relating to the use of ethanol as an excipient in pediatric medicines or drugs is different in different countries and the American Academy of Pediatrics recommends that ethanol content in pediatric drugs should not produce a blood concentration >25 mg/100 mL following administration of a single recommended therapeutic dose. The dose size of EFV is 200 mg for pediatric patients and 600 mg for adults and as a nanoemulsion the dose unit is relatively small and is not expected to produce blood ethanol levels above regulated limits [38]. The European Medicines Agency (EMA) recommends a daily limit of 260.5 mg/kg/day of ethanol [39]. The levels for Span® 20 (A), Tween® 80 (B) and ethanol (C) used in the mixture fell in the range between the minimum and maximum levels listed in Table 2 with the sum of the components A, B and C always totaling 100%.


**Table 2.** Constraints for input variables for D-optimal design.

P20 250 2250 46,600 49,100 4.58 0.50 94.90 433
