ABSORPTION IN HUMAN SKIN 227 4.0 3.2 - 2,4 - 1,6 - 0.8 - 0.0 I I I I I I I I 0.0 0.8 1.6 2.4 3.2 4.0 calculated log Poct Figure 1. Log/log plot of experimental versus calculated n-octanol/water partition coefficients (Poct) for 33 fragrance chemicals (r 2 = 0.97, p 0.00l). etration of the skin barrier. When Kp values of a homologous series of n-alkanols from an aqueous medium through human skin in vitro were plotted against log Poct, the permeability coefficient was seen to rise linearly with increasing chain length. As lipophilicity increased for homologues beyond octanol, however, i.e., at log P values of 4.26 (nonanol) and 4.57 (decanol), Kp levels off, and the correlation curve assumes a sigmoidal character (11). Linearity thus appears to end where intercellular stratum corneum lipids alone are the limiting factor for passive diffusion, a requisite on which the Potts-Guy model is also based. Towards the high end of log P values, the Potts-Guy model also nears the limit of its predictive accuracy for the diffusivity of permeants. It is for that reason in the selection of compounds for this study that those with high measured lipophilicity values (log P 4) were omitted albeit available. This apparently arbitrary cut-off based on lipophilicity is further justified by the phys- icochemical properties characteristic of fragrance chemicals. The polarity range into which the preponderant majority of these chemicals is seen to fit spans log Poct values of -1 and + 4, and structures that exceed that level of lipophilicity are exceptional (12). With fragrance compounds in particular as objective for this discussion, the limited list of compounds with available measured penetration values in Table I was expanded, covering more diverse structures, as they are used in that specialty, so that C log P values could be matched with measured data. That expanded list includes some of the most

228 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS X 6,0 - 4,8 - 3.6- 2.4- 1.2- 0.0 -- 0.0 1.2 2.4 3.6 4.8 6.0 calculated Kp x e-02 (cm/hr} Figure 2. Correlation of experimental versus predicted permeation coefficients, Kp, for 20 fragrance compounds through human epidermis. Predicted values are based on the Potts-Guy algorithm using measured log Poet data (r 2 = 0.86, p 0.001). important fragrance chemicals, used frequently and in large amounts, e.g., benzyl benzoate, phenylethyl alcohol, and cinnamaldehyde, making the list more representa- tive of materials used in fragrance compositions (12). CONCLUSIONS It is demonstrated that the values for the steady-state permeability coefficient Kp of 20 fragrance chemicals across human skin predicted by the Potts-Guy algorithm correlate well with the corresponding experimental data (r 2 = 0.86, p 0.001). Also the calculated octanol-water partition coefficients log Pact for 33 chemical structures of intermediate polarity typical of fragrance chemicals (0.8 log P 4.0) show remark- ably good correlation with measured values (r 2 = 0.97, p 0.001). Since log Pact values demonstrated as reliable are also readily accessible, the prediction method de- veloped by Potts and Guy, based on octanol/water partition and molecular weight for permeation coefficients of otherwise untested compounds through human skin, becomes the obvious method of choice, particularly in the range of polarity typical for fragrance compounds.