254 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS 1•4-0 ' 18'60 ' 1•80 ' 19'00 ' 19'20 ' 19'40 Figure 3. Effect of humidity on the second derivative spectra of skin. subjects' legs, leading to differences in the depth of penetration of near-infrared radia- tion. Within each study, however, the negative correlation with humidity holds. Bulk water should be that which is located below the stratum corneum and is expected to be constant. The apparent decrease can only be attributable to another phenomenon such as decreased penetration of radiation below the SC. Recall that Potts has observed a thickening of the SC with humidity (3). The longer resulting path length for NIR radiation through the SC could leave a shorter path length below the SC. Thus, the SC appears to plump out as the humidity increases. The total path length of radiation, however, does not appear to change significantly. EFFECT OF PRODUCT APPLICATION ON SKIN WATER CONTENT After treatment with all three product types, skin shows a noticeable increase in ap- parent absorbance throughout the wavelength range of 1400-2200 nm, accompanied by an increase in the baseline offset (Figure 8). The effect is most likely due to decreased scattering of NIR radiation, possibly accompanied by an increase in water content. Subjective evaluations under visible light indicate a glossier appearance for treated skin, suggesting that visible light, at least, is less scattered. The MSC correction to the spectra was used to eliminate the apparent increase in absorbance due to scattering and to leave only effects due to changes in water content. Because of the strong effects of humidity on water band intensities, relative water concentration changes produced by different products were compared at single time points (constant humidity). For example, Figure 9 shows relative concentration changes for the three formulations at the four water wavelengths after two weeks of product

MOISTURE MEASUREMENT BY NIR SPECTROSCOPY 255 1200 1100 ? 1879 ß • 1000 '- 900 ß • 800 700 600 500 obsolufe humidify, g/m3 Figure 4. Correlation of free water (1879 nm) and absolute humidity. application. The data were collected at 35% RH at 23øC and were corrected by sub- tracting the control and baseline measurements. The results are thus relative to the baseline measurements that were collected at 30% RH. The plasticizer ingredient in the simple base (A) and in the more complex formulation (C) induce similar effects the levels of free (1879 nm) and protein-bound water (1909 and 1927 nm) decline while that of bulk water (1890 nm) increases. Differences in the net water band intensities between these two formulations are not significant. Propylene glycol (B), which is not expected to act as a moisturizer, shows decreases in free water and protein-bound water comparable to those of the other two formulations, but also results in a decrease in bulk water. The decrease in free water (1879 nm) with all three products occurs in spite of the small increase in humidity relative to the baseline measurements. It may arise from increased flux across the skin surface, which would result in less free water trapped within the SC, or from an overall decrease in SC thickness. The protein-bound water follows the behavior of the free water, as it does when the humidity changes. The increase in bulk water (1890 nm) with the plasticizer may signal an increase in the depth of penetration of reflected radiation into the epidermis as the SC becomes thinner. Product application, therefore, appears to have the opposite effect on stratum corneum thickness than does increasing humidity. EFFECT OF PRODUCT APPLICATION ON SCATTERING All three formulations result in decreased scattering after product application, observed as an increase in apparent absorbance and calculated as the difference between the