146 JOURNAL OF COSMETIC SCIENCE and the presence of traces of proton acceptors and/or donors (25). For most of them, the enol form is the major one. The conjugated system of the enol form spreads further than that of the keto form due to the formation of a pseudo-six-membered ring with OH. This induces a bathochromic shift in absorbance. ClO-DBM and BM-DBM were modelled, and the predicted log (P) values were calcu­ lated. Log (P) for BM-DBM was estimated at 4.0, whereas for the enol form of ClO­ DBM, log (P) was estimated at 7 .9 and the keto form at 8.2. It is, of course, the long aliphatic carbon chain of the ClO-DBM that induces its partition toward an apolar environment and hence an original behavior. When ClO-DBM was incorporated in an organized environment, a micellar solution of SDS, absorbance shifted from UVB to UV A. One hypothesis is that the keto-enolic equilibrium was displaced to the enol form (absorbance maximum at 361 nm). The environment has an important influence on this equilibrium: indeed a particular mi­ cellar distribution of ClO-DBM in the solutions examined can result in the modification of the solute-solvent interactions and of the polarity environment. We can easily imag­ ine the insertion of Cl0-DBM into micelles, favoring the enol isomer. Indeed, it seems that the keto-enolic equilibrium is shifted towards the enol form in apolar environ­ ments. Watarai et al. (26) described the tautomerization of benzoylacetone, benzoyltrifluoroac­ etone, and 2-naphthoyltrifluoroacetone in anionic, cationic, and non-ionic microemul­ sions by means of UV spectrophotometry. The ratio of the two absorbance maxima (around 300-350 nm for the enol form and 250 nm for the keto form) reflected the degree of enolization. They studied the effect of dissolving the molecules in SDS micellar media on the absorbance ratio of these 13-diketones. They noted an increase in the absorbance ratio (Aeno/A keto ) when the SDS is above its critical micellar concentration. They explained this increase by a preferential solubilization of the enol form in micelles. In our case, almost all Cl0-DBM molecules were in the enol form in the micellar solution of SDS. ClO-DBM seems to have a different behavior from other 1,3-dicarbonyl compounds. Unlike BM-DBM, in organic solvents, ClO-DBM is in its keto form. Irradiated for a few minutes, the absorbance shifted from UVB to UVA. Spectral changes can be due to the conversion of the diketone form of ClO-DBM ("-max = 263 nm) to the enol form ("-max = 357 nm) or to its photodegradation into UVA-absorbing species. The mechanism of this phenomenon was studied in more detail and is the subject of a second article (27). HPLC-MS indicated the coexistence of two processes: • in organized media, a shift of the keto-enolic equilibrium towards the enol • under irradiation, decomposition of ClO-DBM leading to BM-DBM (Norrish II photoreaction) In water-in-oil preparations, the same phenomenon as in micellar solutions of SDS was observed. We noted weak UVA absorbance of the cream containing 1 % w/w ClO-DBM before irradiation. Irradiation with either xenon lamp or natural sunlight induced an increase in UV A absorbance for five minutes for the xenon lamp or one hour for sunlight (Figures 8 and 10, respectively). Furthermore, UVA absorbance of the preparation containing ClO-DBM is more stable than that of the preparation containing BM-DBM. After three hours under natural sunlight, we noted a 22% decrease in the UVA absor-

NEW LONG-CHAIN UV ABSORBER 147 bance for the preparation containing Cl0-DBM and an 92% decrease for the preparation with BM-DBM. However, the molar absorption coefficient of ClO-DBM is lower than that of the BM-DBM. In the creams, the molar absorption coefficient at 358 nm of C 10-DBM in the enol form was estimated at 1 7 000 M- 1 cm - 1, whereas that of BM-DBM was estimated at 40 000 M- 1 cm- 1 . Thus Cl0-DBM should not be used alone in cosmetic preparation, but in association with BM-DBM. High initial UV A absorbance will be given by the presence of BM-DBM and it will be preserved by the presence of ClO-DBM, which under irradiation will supply BM-DBM. Therefore, a cream containing a mixture of BM-DBM/ClO-DBM (molar ratio 7:3) was prepared. Looking at the triangles in Figures 8 and 10 (preparations Ml and M' 1, respectively, in Tables I and II), we can confirm that the UV A absorption of these creams was more stable than the UVA absorption of creams Pl and P' 1. ClO-DBM, which is a BM-DBM precursor under irradiation, allowed us to preserve the UV A absorption of the prepara­ tion containing the mixture. Moreover, creams Ml and M' 1 showed an initial absor­ bance that was higher than the absorbance of creams C 1 and C' 1. Thus cosmetic preparations containing both ClO-DBM and BM-DBM exhibit UVA absorbance that remains at useful levels for longer than cosmetic preparations containing only BM-DBM. Studying the curves in Figure 8 and 10, we note that the efficiency of the ClO-DBM molecule was greater under natural sunlight irradiation. Indeed, absorbance of creams C' 1 and M' 1 (Figure 10) can be considered stable under natural sunlight for three hours, corresponding to the period of strongest light intensity (12:00 h-15:30 h solar time, in May and June, Toulouse, latitude 43° North, France). Irradiation with the 150-W xenon lamp was performed without any UVC filter. Therefore, the xenon lamp emitted short wavelength rays, hence more energetic than natural sunlight. Thus the photodegradation of the filters was much faster under xenon lamp irradiation than under natural light. CONCLUSION In this paper, we present a new UV absorber, the l-(4-tert-butylphenyl)-2-decanyl-3- (4' -methoxyphenyl)-propane-1,3-dione called ClO-DBM, derived from BM-DBM by grafting a long alkyl chain. The present work reports the absorbance photostabilities of UVA sunscreen preparations containing BM-DBM and ClO-DBM in water-in-oil prepa­ rations similar to those currently used in cosmetics. Although Cl0-DBM shows a weaker molar absorption coefficient in the UV A than BM-DBM, its absorption is more pho­ tostable in this domain, especially under natural sunlight. The use of both filters together, BM-DBM and ClO-DBM, the first one for its high UVA absorption and the second one for its ability to supply BM-DBM under irradiation, results in good stability in UV A protection. ACKNOWLEDGMENTS The authors thank C. Lebreton and M. P. Dulau of Bioderma, Dipta, Aix-en-Provence, France, for the water-in-oil preparation. REFERENCES (1) International Agency for Research on Cancer, !ARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Solar and UV Radiation (IARC, Lyon, Vol. 55, 1992).