354 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS t4 t3 t2 SHEAR STRESS Figure 6. Rate of Shear vs. Shear Stress for a semi-solid emulsion sheared after different time periods of rest: t• t 2 t 3. irreversible breakdown due to major structural changes. In such cases, after the first rheological run we are no longer really dealing with the original emulsion, and subsequent testing may be meaningless. Despite this limitation, however, continuous shear rheometry does allow some interpretation, particularly if the emulsion exhibits a well-defined yield value or "spur" in the rheogram, as shown in Figures 5 and 6. The yield value or spur clearly reflect the basic gel-like structure of an emulsion. Therefore, any change in its value or appearance must be a reflection of a change in structure and instability. Unfortunately, many investigators concentrate their efforts on studying the regions of high shear and do not see the subtle changes which often show up at the lower shear values. I would argue, therefore, that we would probably learn much more about emulsion instability, and even have a chance at predicting failure, if we concentrate on rheological measurement at as low a shearing stress as possible. Hiestand (14), for example, has suggested a relatively simple test of a dispersion's ability to withstand "curdling" or other changes in floc structure which might be brought about by vibration or other forms of gentle movement during the shipment of products. The dispersion is placed in a bottle which rotates slowly, e.g., 4 revolutions per hour. The slow agitation tends to cause breaking of the floc at some points, but allows time for reestablishing flocculation at other points. This apparently results in a totally different floc structure and a change in the characteristics of the dispersion. The type of bond breaking and making just described is characteristic of systems exhibiting the phenomenon of viscoelasticity for which there is a sound theoretical

STABILITY ASSESSMENT OF EMULSIONS 355 basis (11). Recent publications (16-18) have clearly indicated the value of evaluating emulsions by measuring their viscoelastic properties and have provided suggested experimental procedures which could be adapted to stability assessment. Despite this information, however, the application of such techniques does not appear widespread. Consequently, I would like to conclude this discussion of stability assessment by reviewing what has been and what could be done. Basically, many structured systems such as polymers, gels, suspensions, and emulsions under very low levels of stress will exhibit linear elastic deformation because of relatively strong intermolecular or interparticle bonding• In such cases the stress applied is directly proportional to the strain (extent of deformation) and independent of the rate of strain. As with perfectly elastic solids, this linear relationship is described by Hooke's law. In the case of a true elastic body the strain will remain constant under constant stress, and upon release of the stress, the solid will fully recover its original structure. In the case of a viscoelastic substance, under such conditions a slow relaxation or increase in strain occurs over an extended time period. This arises because the structured body undergoes a process of bond breaking and remaking in such a way as to modify the structure. Eventually, with continued application of the stress viscous flow will occur. When the stress is small enough to cause Hookean elasticity and NewtonJan viscosity, we speak of linear viscoelasticity, for which theoretical concepts were well worked out (11, 18). Upon release of the stress in such systems some recovery of structure occurs because of the elasticity, but this recovery cannot be complete because of the relaxation which has taken place. A typical plot of compliance (strain per unit stress) versus time, often called a creep curve, is shown in Figure 7. What makes this approach so attractive for emulsion stability assessment is that we are able to obtain fundamental material constants such as elasticity and viscosity without significantly stressing the system mechanically. Indeed, Barry (18) has referred to this as an evaluation of the "ground state" of an emulsion or suspension. Therefore, we are not only in a position to observe very subtle structural changes which take place upon storage and use, but we also can determine the values of fundamental constants which correspond to acceptable stability. This allows us, for example, to develop quality control specifications translatable from laboratory to laboratory. Because of the high sensitivity of viscoelas- tic measurement, it should also be possible to attempt to predict long-term stability utilizing a much less extreme accelerated condition and, hence, minimizing the possibility of a change in the mechanism of instability. Although the application of viscoelastic measurements to an understanding of polymer solution behavior at a fundamental level is well documented (11), reports on the use of such techniques for disperse systems of interest to cosmetic and pharmaceutical scientists have been extremely limited. Barry (19) has carried out extensive studies on some semi-solid o/w emulsions and has been able to monitor very well the significant changes in viscoelasticity brought about by the effect of emulsifier concentration on emulsion structure. Sherman (20) has used viscoelastic creep tests to study liquid w/o emulsions stabilized by sorbitan monooleate. He has correlated the tendency for coalescence during shelf life for emulsions ranging from 30% to 65% water with their viscoelastic properties. Hiestand (14), using the rising sphere viscometer developed by McVean and Mattocks (21), has applied a very small stress on a 5% solid suspension stabilized with methylcellulose and has measured viscoelastic relaxation. He was able to show that the relaxation process arose from the flocculated structure of the suspension and not the vehicle alone. Thus we see the potential utility of these techniques in