104 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS sunscreen agent separately, setting maximum allowable concentrations in products (1). The maximum allowable concentration was determined on the basis of the concentra- tions that had been used in marketed formulations and that were supported by safety testing for those concentrations. The current policy allows use of combinations of ap- proved agents up to the maximum concentration limits for each agent. At the time of the OTC review only titanium dioxide was submitted as a physical sunscreen agent for use in sunscreen products. Titanium dioxide, therefore, appears to be the only acceptable physical sunscreen for use in sunscreen products in the United States (1). In the past several years, the availability of physical sunscreens and their quality has increased. The range of particle sizes, especially smaller sizes, has been extended dramatically. Today there are many additional choices of possible physical sunscreen agents. Moreover, we now know better how to design acceptable products incorporating physical sunscreens. According to Pathak (3), "Physical sunscreens are usually opaque formulations and contain ingredients that are particulate in nature and do not selectively absorb UVR, but when applied as a film, they primarily reflect and scatter UVR and visible radiation because of the size of the particles and thickness of the film." Also in 1986, Kollias et al. (2), while establishing the optical specifications for a new spectrophotometer, exam- ined the reflectance characteristics of titanium dioxide, showing that titanium dioxide exhibited a strong absorption edge characteristic of semiconductors at wavelengths just shorter than 400 nm. This observation provided the inspiration for this examination of the definition of physical sunscreens. METHODS A Varian Cary 2300 UV/vis spectrophotometer with a Harrick Praying Mantis attach- ment was used for obtaining diffuse reflectance spectra of physical sunscreen powders. Barium sulfate was used throughout as the reflectance reference (Aldrich Chemical Company #24,335-3). All spectra were measured relative to the BaSO 4 spectrum. Spectra were measured from 250 to 720 nm and stored on magnetic media using a Varian DS-15 data station. A diffuse reflectance spectrophotometer was also used to obtain similar data from opti- cally thick powder samples (2). This instrument was modified for UV spectrophoto- metric measurements as follows. A 1000-watt xenon arc (Oriel #6140) was coupled to a Jobin Yvon UV-HL monochromator with the slits set at 1.0 mm (band width of 2 nm), which provided input to a slit-shaped bifurcated silica/silica-clad fiber optic bundle. The common end of the bifurcated bundle had randomly mixed fibers. The other end of the bundle terminated in a slit-shaped fiber array that provided input to a Jobin Yvon H10 monochromator whose output was measured with an Oriel #7070 photomultiplier power supply amplifier. Shutters, stepper motors, and order sorting filter were controlled by and data acquired with a Hewlett Packard Model 150 II mi- crocomputer and a data acquisition controller (HP 3497). In vivo tests were performed to determine if titanium dioxide and zinc oxide in a clinical product (Rose Cream) provided the protection predicted by the in vitro study of sun- screen products. A 10-cm X 10-cm square was drawn on five volunteers' backs. A site within this area was selected and scanned with the fiber optic spectrophotometer prior

PHYSICAL SUNSCREENS 105 to product application. 100 mg of the test product was applied (1 mg/cm 2) and the site was rescanned. The physical sunscreens examined were: Titanium dioxide, Teikoku Kato #3296 Titanium dioxide, Degussa #P25 Titanium dioxide, cosmetic grade #8740-40 Titanium dioxide, Guardian Chemical, ultrafine grade Talc, USP Talc, USP, micronized (8190-40/19888) Zinc oxide USP Iron oxide, yellow cosmetic grade Iron oxide, black cosmetic grade Iron oxide, red cosmetic grade Rosa Cream (7.5% zinc oxide q- 7.5% titanium dioxide prepared by the AI-Sabah Hospital Pharmacy) RESULTS The representative reflectance spectra for titanium dioxide, zinc oxide, talc, and the iron oxides obtained using the Cary 2300 are shown in Figure 1. All spectra are relative to the BaSO4 reference. Throughout the visible spectrum, all agents tested except the iron oxide reflect 90 to 100 percent of the light relative to barium sulfate. Talc reflects visible and UV similarly to the barium sulfate reference. At wavelengths shorter than d00 nm, titanium dioxide and zinc oxide behave differently from the barium sulfate reference. Both zinc oxide and titanium dioxide reflect less radiation in the ultraviolet than does barium sulfate. The results for the pigmented iron oxides show that these compounds reflect, scatter, and absorb visible wavelengths and behave similarly in the UV radiation. The reflectance of titanium and zinc oxide powders using the fiber optic remittance spectrometer is shown in Figure 2. The energy gap is shown to be at approximately the same wavelength for both sunscreens. A representative in vivo spectrophotometric test of 1 mg/cm 2 Rosa Cream containing 7.5% zinc oxide and 7.5% titanium dioxide is shown in Figure 3. Note the sharp decrease in reflectance just short of the visible wavelengths. The absorbance of the product is measured by radiation passing twice through the applied product and, there- fore, two times larger. DISCUSSION The results indicate that there are at least two different types of physical sunscreens: those which only scatter radiation and those that also absorb selected wavelengths. Agents that only scatter radiation are represented by barium sulfate (BaSO4) and talc. Titanium dioxide (TiO2), zinc oxide (ZnO), and the colored iron oxides comprise the second group, namely those that scatter some wavelengths and absorb other selected wavelengths of light. This latter group of compounds is potentially very large in number. Both talc and barium sulfate must be visible on the skin's surface if they are to