44 JOURNAL OF COSMETIC SCIENCE FORMULATING FOR MILDNESS WITH SULFOSUCCINATES Andrea Allardice Witco Corporation While sulfosuccinates have been widely used for years within the personal care industry for formulating mild skin and hair cleaning products, their importance today is greater than ever. Formulating novel products will be discussed. Synergies of sulfosuccinates with other low irritation surfactants will be shown. This will allow the formulators to achieve the performance characteristics the consumer demands while maintaining the low irritation profile required. In addition, the general chemistry and variations of sulfosuccinates will be outlined. The main focus is to be the dermatological properties. These will be demonstrated by the comparison of in-vivo and in-vitro tests, especially the Red Blood Cell Test. A MATHEMATICAL TREATISE OF THE PERCEIVED ATTRIBUTES AND RHEOLOGICAL PROPERTY CORRELATIONS OF THE SKIN CARE CREAMS J. Mendoza, H.Y. Kung, M Wang, I. Dumanli, C. Lee and M.S. Kislalioglu Cosmetics and Personal-Care Products Technology Program, Department of Applied Pharmaceutical Sciences, The University of Rhode Island, Kingston, RI, 02881 INTRODUCTION Functional properties and sensory perceptions are equally important in cosmetic product development. Determination of a universal model that can provide acceptable correlations between the physical and perceived properties of a wide range of emulsions was among ongoing interests of this group. This study seeks mathematical correlations between the rheological properties of 10, 30, 50 % O/W and 10, 30, 50 % W/O skin creams and their perceived attributes. It is continuation of the former studies that is expected to explain the functional and perceived property relationships of such products one step further METHODS Preparation of Emulsions: All emulsions were perpared using castor oil (Ruger, NJ) as the oil phase, sorbitan stearate (ICI, NJ) and ceteareth-6 & stearyl alcohol (BASF, NJ) as surfactants. The oil phase and water containing 5.0% (w/w) surfactant were heated up to 72 o C and mixed together using a Ross Mixer ME 100 LC (Charles Ross & Son Co, NY) at 5,000 rpm for 10 minutes. O/W and W/O emulsions containing 10, 30 and 50% w/w dispersed phase were prepared and characterized. All emulsions were in the cream form and had similar appearences. Determination of Perceived Attributes: Twelwe untrained oriental female adults with healthy skin, whose ages ranged from 24 to 32 years participated in these studies. Emulsions prepared were placed in identical containers and presented to the panel for evaluation. The panel members were asked to rank the pick-up (firmness, stickiness, peaking and cohesiveness), rub-out (wetness, thickness, absorbency,

PREPRINTS OF THE 1997 ANNUAL SCIENTIFIC MEETING 45 amount of residue and spreadability), after-feel (stickiness, gloss and slipperiness) and over-all properties (liking, absorbency and softness ) from 1 (the worst) to 9 (the best) according to ASTM Committee E-18 Guidelines for the Descriptive Analysis of Skin Feel. The means and standard deviations of each attribute was calculated. Determination of Rheologleai Properties: After determination of the optimum stress-sweep range which was within 0.05 -100 Pascals, the theological properties of the emulsions were measured with a CVO 50 Bholin Instruments Rheometer, software version 4.5, at 25 o C by using a 1 mm gap width. The theological components measured were the phase angle, elastic (G') and viscous (G") moduli, complex viscosity and strain. Correlation Studies: Among the theological components measured, the loss tangent ( G'/G" * 100) was found to be the most represantative parameter to describe the flow properties of the creams. They were tabulated from the data obtained and correlated to the sensory attributes by using linear ( Y=AX+B ), logarithmic (Y=A+B*Ln(X)), exponential (Y=A*Exp(B*X)) and power (Y=AX•), equations, where X was either the oil concentration or the loss tangent, Y was the perceived attribute and A and B were the regression conslants. The probable fits were postulated and their resulting correlation coefficients (r2), were used to compare the goodness of fit. The data were fitted to the models by using Hewlett Packard 48G software. Additionally, a linear three-dimensional model which incorporated the weight fraction of oil, perceived attribute, and the loss tangent was tested by the use of Maple version 4.0 software and resulting correlation coefficients were also obtained in order to compare its fit. RESULTS AND DISCUSSION All creams were in the form of singular emulsions. As shown in Table I, the two dimensional models showed scattered correlations with only occasionally acceptable fits. Whereas, the three dimensional model, Z=Co+AX+BX, where Z was G'/G"*100. X was the weight fraction of oil , Y was the perceived attribute and C o, A and B were the regression constants, presented their relationship with most acceptable accuracy,Table II. In this model the correlation coefficients of 0.993 to 0.936 (eigt cases) demonstrate an excellent, and 0.768 to 0.738 (four cases) good correlations. Coefficients that change between 0.600 to 0.256 (three cases) are the measure of poor fit,(Table II) Among all the models tested, the three dimensional linear model was the most satisfying to describe the relationship between the perceived attributes, the weight fractions and the theological properties of the emulsions.