EFFECT OF SURFACTANT LOCATION ON EMULSIONS 098 Table I Effect of Initial Surfactant Locations on Mean Droplet Diameter (Tween 80-Arlacel 80 at HLB 10, All Arlacel 80 Dissolved in the Oil Phase Prior to Emulsification) Combined Surfactant Concentration in the System (% by weight) Tween 80 Initially in the Aqueous Phase (% of total) Mean Volume Diameter, d• (cm) 1 100 2.8 X 10 -3 1 40 2.8 X 10-3 1 0 4.8 X 10 -3 8 100 2.8 X 10 -3 3 20 2.5 X 10 -3 3 0 3.2 X 10-3 Although the amount of data presented here does not allow generali- zation, for the systems investigated, however, there is a tendency to produce a coarser emulsion as less hydrophilic surfactant is present in the aqueous phase prior to emulsification. A possible explanation for the observed difference in the droplet size distribution may be given from the surfactant migration viewpoint. In order that the freshly formed droplets remain stable, the surfactant molecules must be adsorbed at the oil-water interface. From the dy- namic surface tension measurements made by a number of authors using the oscillating jet method, it is evident that a finite time is required for the system containing surfactants to reach its surface equilibrium (12- 14). Conceivably, the presence of a surfactant in the oil phase or aque- ous phase prior to emulsification can affect the accessibility of the sur- factant at the interface and hence the droplet size distribution. Effect of Emulsion Stability The effect of the initial surfactant location on emulsion stability was studied both by a shaking method as well as by using the emulsification vessd described in Fig. 1. In the shaking experiment, the oil phase and the aqueous phase con- taining various amount of surfactants were shaken together in an en- closed jar using a mechanical shaker for a predetermined length of time. The emulsion thus formed was then poured into a graduated cylinder and the amount of separation was observed as a function of time. However, in many systems, shaking produced a considerable amount of foam which made it difficult to interpret the data obtained. For example, in a series of experiments using Tween 80 and Arlacd 80, the emulsion prepared by

694 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS (5 % TOTAL} 40- TWEEN 80-ARLACEL 80 HLB I0 / - • • •¸ • 20- co 0 20 40 60 80 I00 % TWEEN 80 IN AQUEOUS PHASE Figure 9. Effect of initial hydrophilic sur- factant location on emulsion stability predissolving all Tween 80 in the aqueous phase was less stable than an equivalent emulsion prepared by initially placing Tween 80 in the oil phase. However, the former emulsion produced a considerably greater amount of foam than the latter emulsion during the shaking. The presence of foam during and after emulsification could have affected the emulsions in many ways: (a) Foam could damp the shaking action to reduce the intensity of mixing operation, (b) adsorption of the surfactants at the liquid-air interface could deplete the surfactants needed at the oil-water interface to stabilize the emulsion, (c) retaining of the liquid in the foam laminas could affect true reading of the degree of separation. For this reason, shaking experiments were abandoned and a limited study was made using the turbine mixer in the emulsification vessel. The emulsions prepared in the emulsification vessel were virtually free of air bubbles and the data were fairly consistent. The example shown in Fig. 9 represents a series of emulsions stabilized with Tween 80 and Arlacel 80 at the required HLB value of the mineral oil. The emul- sion stability was determined by measuring the per cent of the volume of water separated in a graduated cylinder. In this system, placing of the surfactants in their respective phases produced a less stable emulsion than the one prepared by placing all surfactants in the oil phase. However, this observation did not apply to others as an exactly opposite effect was noted in many other systems. Since emulsion stability is an extremely complex function of droplet size distribution, rheological properties of each phase, surfactant type, concentration, etc., the limited experimental data obtained here do not