60 JOURNAL OF COSMETIC SCIENCE CHAP•CTERIZATION OF ANTIPERSPIRANT ACTIVES USING SIZE EXCLUSION CHROMATOGRAPHY AND LIGHT SCATTERING DETECTION Allan H. Rosenberg, Ph.D., Walter Carmody, Summit Research Labs, Huguenot, New York 12746 Introduction Size Exclusion Chromatography (SEC) has been successfully applied to characterization of antiperspirant actives in terms of their polymer distributions (1). This information was crucial for fully understanding the chemistry of antiperspirant actives with subsequent development of high performance (enhanced efficacy) actives which were synthesized and manufactured to contain particular polymer distributions which enhanced clinical performance (2, 3, 4). Not withstanding the success of SEC it still remains difficult to obtain accurate absolute molecular weight data from SEC analysis due to the lack of satisfactory "marker" compounds necessary to accurately calibrate SEC columns for the type of polymer species present in antiperspirant actives (highly charged cationic polymers). Since light scattering has been used to obtain molecular weights of polymers in solution, the idea was to combine SEC with light scattering detection to determine on-hne molecular weights of the separated polymer species present in antiperspirant actives. Methods On-line determination of molecular weights of aluminum polymers present in aluminum chlorohydrate (ACH) actives were obtained by separating the aluminum polymers on a Sephadex G-25 (S) column followed by light scattering detection (measuring the scattering intensities) of the separated polymers using the Dawn DSP Multi-angle Laser Light Scattering Photometer (Wyatt Technology) followed by refractive index (RI) detection of the eluent stream exiting from the Dawn Photometer. Molecular weights are calculated from the scattering intensities and RI responses using the Wyatt Technology Astra program. dn/dc values (change 6f refractive index versus change of concentration) of the aluminum polymers being evaluated are required for the molecular weight calculations. In the case of antiperspirant actives it is virtually impossible to isolate and measure dn/dc values for each individual polymer present in the active. Therefore, dn/dc values for the bulk solution of active is used for each of the aluminum polymers. This approximation will perturb the resultant molecular weight value of each polymer, the degree of which is dependent upon the difference between the bulk drddc value and the "true" dn/dc value of the individual polymer. BuLk dn/dc values were determined in separate experiments using an Optilab 903 Refractometer which is specially designed for such measurements. Operation and calibration of the Dawn Photometer was checked by determining the molecular weight of a pullenen standard. For this work a value of 47,000 daltons was obtained which is within 2.5% of the stated value of 45,800 daltons. Results and Discussion Figures 1 and 2 show aluminum polymer distributions for a standard ACH active (Summit ACH-323) using hght scattering and RI detection, respectively. Both profiles are similar with four distinct aluminum polymers observed. This is typical for ACH characterization using Sephadex G-25 (S) columns. The similarity in relative peak heights for both profiles indicates that the molecular weight range between peak I polymer and peak 4 polymer is not very large since scattering intensities are strongly dependent on molecular weight. Table 1 summarizes SEC-light scattering results for ACH-323. Also included are literature values of molecular weight ranges for the aluminum polymers compiled from a variety of experimental techniques (5). TABLE 1 SEC-Light Scattering Characterization of Summit ACH-323 Literature MW From Peak MW Ranges SEC-Light Scattering i 5,000 - 8,000 7,600 2 3,000 - 4,000 6,000 3 1,500 - 3,000 5,000 4 500- 1,500 2,100

PREPRINTS OF THE 1997 ANNUAL SCIENTIFIC MEETING 61 As seen in Table 1 molecular weights obtained from SEC-Light Scattering are somewhat higher than the estimated literature ranges except for the peak i polymer which falls within the range. These differences may be due to individual drddc values differing from the bulk value used in the molecular weight calculations, the small scattering intensities for the aluminum polymers which can impact accuracy and the uncertainty of the literature values themselves. Further work is in progress to address the above issues. References 1. Antiperspirants and Deodorants edited by K. Laden and C. B. Felger Chapter 6, pp 163-178, Marcel Dekker (1989) 2. Markarian, H. and Rosenberg, A. H. U.S. Patent 4,818,512 (1989) 3. Rosenberg, A. H., et al European Patent Application 0653 203A1 (1995) 4. Callaghan, D. T., et al U.S. Patent 5,486,347(1996) 5. Antiperspirants and Debdorants, edited by K. Laden and C. B. Felger, Chapter 6, p 170, Marcel Dekker (1989) Elution Time Figure 1. Aluminum polymer distributions for ACH-323 using light scattering detection. Elution Time Figure 2. Aluminum polymer distributions for ACH-323 using RI detection. RAPID PREDICTION OF EMULSION STABILITY Gerd H. Dahms IFAC (Institute for Applied Colloidtechnology), Duisburg, Germany The prevalent method for testing the physical stability of cosmetic emulsions is storage at temperature. A disadvantage of this method is the long time interval between production of the sample and the observation of any visible phase separation that would indicate instability. This time interval sl•etcbes to several weeks. Methods employing rheological or conductivity measurements can prodict physical stability within 6 or 48 hours respectively after sample production, and with 100% reliability. Such methods thus offer the potential to greatly accelerate this part of the formulation development process. Rheoloeical Methods Physical instabilities such as creaming or coalescence are caused by flow phenomena in emulsions. Droplets which are unable to move through the continuous phase cannot flow together or move to the top or bottom of a container through the operation of Stokes' Law. The viscosity described by Stokes' Law is the "zero shear viscosity", i.e. the viscosity of an emulsion in the state where no sWucture breakdown has taken place. To be stable, oil-in-water emulsions need to display an infinite zero shear viscosity, i.e. to display a yield sl•ess. At very low shear rates, they behave more like an elastic solid than a viscous liquid. The forces applied to the continuous phase by the discontinuous phase due to gravity/buoyancy must be below the system yield stress. The yield sla'ess is due to slxucture set up in the continuous phase by theology modifiers or liquid crystalline gel networks. A cone-and-plate rheometer can be used in oscillation mode to extract the magnitude of the elastic forces operating within the emulsion G' (i.e. the extent to which the emulsion is behaving like a solid), and the magnitude of the viscous forces within the emulsion G" (i.e. the extent to which the emulsion is behaving like a Newtonian liquid). These two parameters can be measured across a range of temperatures. Essentially an oil-in-water emulsion will be stable so long as, at very low shear rates, the viscous forces are less than