102 JOURNAL OF COSMETIC SCIENCE cream is strongly related to the appreciation by industrial marketing of the best pro- fessional or non-professional handling. The same shades could thus be obtained either from a shampoo or a cream. The analytical challenge is thus to develop a method that allows both identification and quantification of target dye intermediates into existing commercial formulations available on the market, with the aim of implementing the European Union Cosmetic Directives 76/768/ECC and 95/17/EC (1). These directives address both targets of identifying and quantifying the authorized hair dyes and of verifying that banned hair dyes listed in Annex II are not present in the finished cosmetic products. The target hair dyes, listed in the C.T.F.A. International Cosmetic Ingredient Dictionary and Handbook (2), were selected after discussion and agreement with cosmetic industrials and scientists [COLIPA • (3) and L'Or6al data (4)], for the following reasons: ß They are representative of three major classes of oxidative hair dyes: aromatic amines, aminophenols, and phenols. ß They appear in the European directive as restricted in concentration (Annex III) or banned (Annex II) (1) for example, the choice of toluene-2,4-diamine, listed in Annex II, was deliberate in order to evaluate the detection of banned compounds as well as authorized ones. ß They appear regularly in the composition of commercial formulations. In our previous work, an analytical procedure involving a RP-HPLC separation after a three-step extraction by an organic solvent was developed that enables the separation and identification of 52 different dye-forming products according to their retention times and UV spectra, first in standard pure solutions (5,6), then in self-prepared synthetic formulations and commercial non-dyeing shampoos spiked with oxidative hair dye intermediates (7). The whole procedure was then validated to assure confidence in the final results (7). In the present contribution, the application of the methodology for the determination of thirteen oxidative hair dyes in real commercial formulations is shown as well as the validation process for this type of samples. This was carried out using commercial shampoos and creams available on the local market. Qualitative determination was clearly achieved for the twelve hair dyes initially present in the five formulations. Quantitation of each dye before and after spiking of the formulations with (i) the target dyes already present in the mixtures and (ii) additional commonly used hair dye inter- mediates was successfully achieved in the two types of formulations, and figures of merit of the method have been established. EXPERIMENTAL INSTRUMENTATION HPLC was performed using a two-piston HPLC pump with a low-pressure ternary gradient system module (System 325 from Kontron Instruments S.P.A., Milan, Italy), an Autosampler 560 with a loop of 20 lal (Kontron Instruments S.P.A), a Photodiode Array • Comit• de Liaison Europ•en de l'Industrie de la Parfumerie, de Produits Cosm•tiques et de Toilette.

OXIDATIVE HAIR DYES 103 Detector 540 (Kontron Instruments S.P.A.), and a vacuum degassing system, Degasys DG 1300 (Unifiows, Japan). The column temperature was kept constant by means of a Column Thermostat 582 (Kontron Instruments S.P.A.). Data processing was done with the Data System 450-MT2/DAD Series (Kontron Instruments S.P.A). Additional equipment was used to deal with some coelution problems encountered. This equipment was an HPLC system allowing detection both with DAD and ESI-MS, and can be described as follows: a Waters Alliance 2690 quaternary solvent delivery system coupled to a Waters 996 photodiode array detector (Waters Corporation, Milford, MA). ESI-MS experiments were performed using a Quattro LC triple stage quadrupole in- strument from Micromass (Manchester, UK). MaxEnt TM and Transform TM software from Micromass were used for mass spectra deconvolution and molecular mass calculations. The column was a Merck Lichrospher RP 60 Select B (C8), 250 x 4 mm, 5-pm particle size. CHEMICALS "Les Colorants Wackherr" (Saint-Ouen l'Aum6ne, France) kindly provided p- phenylenediamine (ppd), m-aminophenol (map), and resorcinol (res). Hydroquinone (hyd), 0-aminophenol (oap), p-aminophenol (pap), toluene-2,4-diamine (2,4-dat), and o•-naphtol (o•-napht) were obtained from Fluka. Toluene-2,5-diamine sulfate (2,5-dat), 2,4-diaminophenoxyethanol (2,4-dape), 2-methyl-5-hydroxyethylaminophenol (2-CH3- 5-OHEtap), 6-hydroxyindole (6-OHi) and hydroxypropyl-bis-(N-hydroxyethyl-p- phenylenediamine) HCI (HBNPPD) were kindly provided by L'Oreal, (Paris, France). Five commercial formulations consisting of four shampoos (two nuances of brown shades and two nuances of blonde shades) and one cream (dark brown shade) were purchased on the local market. L-Ascorbic acid sodium salt (NaAsc) was obtained from Fluka. Sodium tetraborate decahydrate (p.a.), acetic acid 95% (suprapure), ammonia 25% (suprapure), hydrochloric acid (0.1 M), and n-heptane (p.a.) were obtained from Merck. Methanol (HPLC quality) was obtained from Fluka. Pure water (18.2 Mf•/cm quality) used for the preparation of solutions was obtained from a MilliQ plus 185 system (Millipore, Molsheim, France). N57 quality nitrogen (desolvation and nebulizer gas) was obtained from Air Liquide (Liege, Belgium). PREPARATION OF REAGENTS Solvents and chromatographic mobile phase. For the mobile phase (aqueous phase B), a 0.05 M acetic acid solution was adjusted to a pH of 5.9 with a 10% ammonia solution and filtered through a 0.45-1•m filter. When not in use, the eluent was stored at a tempera- ture of 4øC to prevent microbiological growth. Two different solvents were tested regarding pH for the sample preparation: © mixtures of methanol (MeOH) and mobile phase (pH 5.9) (40%) © mixtures of methanol (MeOH) and Soerensen buffer (40%). The Soerensen buffer (pH 8.1) was prepared as follows: 88 ml of hydrochloric acid (0.1 N) and 2 g/l of L-ascorbic acid sodium salt (NaAsc) as an anti-oxidant agent were added to 112 ml of sodium tetraborate solution.