132 JOURNAL OF COSMETIC SCIENCE Figure 5. Relative efficacy of light-induced spin trapping in suspensions of untreated titanium dioxide and zinc oxide particles. The EPR spectra were obtained with a 9.6-GHz EPR spectrometer using the following instrumental settings incident microwave power, 20 mW scan time, 2 m time constant, 0.1 s modulation amplitude, 1 gauss scan range, 100 gauss modulation frequency, 100 kHz. The aqueous suspension of metal oxide, containing 100 mM DMPO, was exposed to the light source for 30 seconds, and the EPR spectrum was recorded immediately. (A) DMPO, ultrafine zinc oxide, white light (B) DMPO, ultrafine titanium dioxide, white light. oxide. This effect presumably occurred by preventing the reactive species from entering into the solution. At first glance the observed phenomena might appear to have no potential significance in the customary use of these oxides in cosmetics and/or sunscreens. The usual use of these materials is for preparations that are applied to the intact skin. It seems reasonable to assume that the upper layers of the skin, which consist of dead cells and extracellular material, would not be damaged significantly by a flux of strongly oxidizing species such as hydroxyl radicals because these species react very quickly with the first organic molecules that they encounter. This assumption probably is true under most circum- stances (although experimental verification would seem desirable). Even so, there are at least two scenarios in which damage plausibly could occur: (a) if the strongly oxidizing species generated less-reactive products that could migrate significant distances before reacting (e.g. hydrogen peroxide) and (b) if the oxides came into direct contact with functional cells, either because of breaks in the skin or by movement of the oxides into deeper layers of the skin or onto mucous membranes. Under these circumstances the potential for deleterious reactions might be significant. For example, in preliminary experiments with cultured cells we found that light- induced toxicity occurred in the presence of untreated zinc oxide and that this was significantly less if silicone-treated zinc oxide was used. The nature of the reactive species is an interesting question for which the present experiments provide only incomplete evidence. Several control experiments indicate conclusively that the spin adduct that is observed in this system requires the presence of light plus zinc oxide or titanium dioxide. The spectrum that is obtained by spin trapping is characteristic of an adduct of a hydroxyl radical, but because there are several

REACTIVITY OF SUNSCREEN COATINGS 133 Figure 6. Effect of room light on spin trapping in suspensions of untreated titanium dioxide and zinc oxide particles. The EPR spectra were obtained with a 9.6-GHz EPR spectrometer using the following instru- mental settings: incident microwave power, 20 roW scan time, 4 m time constant, 0.2 s modulation amplitude, 2 gauss scan range, 100 gauss modulation frequency, 100 kHz. The aqueous suspension of metal oxide, containing 100 mM DMPO, was laid on the laboratory bench and exposed to room light for 15 minutes, and the EPR spectrum was recorded immediately. (A) DMPO, ultrafine zinc oxide, room light for 15 minutes (B) DMPO, ultrafine titanium dioxide, room light for 15 minutes. other pathways that can lead to the same adduct, this in itself is not conclusive evidence that the reacting species is a hydroxyl radical. The competition experiment (Figure 3) in which formate was added into the reaction system to compete with DMPO for the reacting species provides supporting evidence for our assignment of the adduct as the hydroxyl radical. The observation that this is generated even by a relatively long- wavelength light suggests that the chemistry may be complex and could involve several different intermediates. While taking into account some of the uncertainties noted above, there still appear to be a number of practical considerations that can be derived from the results of these experiments. Most obviously and perhaps most importantly, the effectiveness of the surface treatments with silicone that were used appears to be sufficient to prevent biologically significant damage from the reactive species that were detected by spin trapping. It therefore might be prudent to use such materials wherever possible and to avoid the use ofuntreated zinc oxide or titanium dioxide. The methodology used in these experiments is relatively straightforward and could be applied to other materials, and therefore it might be prudent to consider using this approach more widely to assess cosmetics that might have significant exposure to sunlight or strong room light. It also might be useful to consider the use of a variant of these techniques in which EPR spectroscopy is used directly in living subjects (8,9). In vivo EPR spectroscopy has lower