PRODUCT STABILITY--PART II 319 emulsion-water boundary from the rotor center is plotted rs. time ior each revolutions per minute. An earlier example of the use of centrifuge studies is the work of Merrill (39). This investigator studied the mechanical stability of O/W and W/O emulsions by measuring the rate of separation of internal phases when the emulsions were subjected to a constant centrifugal force. The quantitative index of stability used was the reciprocal of the initial rate of separation at a certain centrifugal speed. Plots of per cent oil separated rs. time of centrifuging were linear. Suspensions and Selected Solutions Martin (40) discussed the physical-chemical principles involved in suspensions. Such parameters as sedimentation height were noted here one determines the ratio of the ultimate height of the sediment to the initial height of the suspended material. Naturally, the larger this fraction is the better is suspendability. He also noted the applicability of electrophoresis which employed a microelectrophoresis apparatus. Such instrumentation could measure migration velocity with respect to the surface electric charge or the familiar zeta potential, which has units of viscosity times electrophoretic mobility, or, more familiarly, voltage. Stanko and DeKay (41) evaluated suspensions by electrokinetic methods. Since pH, sedimentation rate, and viscosity would not neces- sarily indicate everything that is happening in the system, it was felt that the electrokinetic properties, that is, the zeta potential and critical potential (zeta potential at which flocculation occurs) would be illumi- nating. The zeta potential was obtained in a microelectrophoresis apparatus from which the rate of movement of particles could be ob- tained. These workers plotted the weight of sediment rs. time and obtained gently curving lines. Although no kinetic plots were made, it was shown that the zeta potential changes upon the addition of additives and that it is related to stability. Haines and Martin (42, 43) studied some formulation factors which influence the stability of suspensions. They correlated zeta potential to the measure of visually observed caking. The zeta potential was de- termined by microscopic electrophoresis, and it was found that certain zeta potentials produced more stable suspensions because flocculation was controlled. Foernzler, Martin, and Banker (4•) studied the effect of thixotropy on suspension stability. They obtained straight line graphs by plotting the sedimentation velocity in a centrifuge at a particular RPM in ml./

320 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS min. against the reciprocal of the thixotropic area. The sedimentation velocity was obtained by plotting the sedimentation volume rs. time. The thixotropic area is the area of the hysteresis loop formed when the viscosity of a non-Newtonian liquid is taken continuously over a range of stress-strains and back the area was measured with a planimeter on the plots of rate of shear rs. shearing stress. The authors thus at- tempted to predict physical stability by the rheologic evaluation of thixo- tropy. Incidentally, Wood (45) used these workers' data to show that a similar or better correlation exists if the reciprocal of the yield value is plotted against the sedimentation velocity. Wood, Catacalos, and Lieberman (46) studied aging magnesium aluminum silicate suspensions. They found new, interesting, logarith- mic, atypical kinetic relationships involving the time and temperature of storage and shear rate and shear stress. Plots of log shear stress rs. log shear rate produced a family of straight lines for a particular storage temperature in which the slopes decreased with sample age, although the older samples' lines "started higher," i.e., required more stress to get the same shear rate as the younger. Thus plots of log apparent viscosity rs. log age were linear for each particular shear rate. These workers also pointed out that plots of log apparent viscosity at a single shear rate rs. either log age at a particular temperature or I/T at a particular age were linear the latter indicates that viscosity build-up follows an Arrhenius type of relationship. Thus plots of log age to attain a given viscosity at a particular shear rate vs. 1IT were also linear. For practical purposes it is noted that work such as this illustrates that clay and gum hydration is not attained instantaneously--it is part of the aging process and is not necessarily completed even by the stress a formulation undergoes in a manufacturing procedure. Garrett and Carper (47) studied the color stability of a multisulfa suspension after it was exposed to accelerated temperature storage condi- tions. They followed the UV absorbance of the supernatant liquid after the suspension was centrifuged. They observed that the absorbance rs. time at the different temperatures provided straight lines from which they could calculate first order rate constants. Their data also fit an Arrhenius plot from which they could predict after a four-week study the rate of deterioration of the color after one year at room temperature. Levy (48) studied the stability of the viscosity of various molecular weight sodium carboxymethylcelluloses and sodium alginates in aqueous solutions stored at an elevated temperature. When the log per cent of the initial viscosity was plotted aginst time, a linear first order plot