106 JOURNAL OF COSMETIC SCIENCE for a large number of matrix components while not affecting the active compounds (6) except ot-naphtol. Identification of oxidative hair dye intermediates. A spectra library containing 200•400 nm spectra of the target substances and their HPLC retention times has been recorded for the identification using the acquisition software. The active compounds contained in the different formulation samples can thus be identified by means of comparison of both retention time and UV spectra. Determination of concentration of dyes. The quantitation was performed for nine of the thirteen identified oxidative hair dyes, namely p-phenylenediamine, toluene-2,4- diamine, toluene-2,5-diamine sulfate, m-aminophenol, resorcinol, hydroquinone, ot-naphtol, 2,4-diaminophenoxyethanol, and hydroxypropyl-bis-(N-hydroxyethyl-p- phenylenediamine) HC1. It was based on calibration lines separately established for each of them in pure standard solutions prepared in a mixture of mobile phase pH 5.9 (40%) and MeOH. Each sample in a pure solution was injected in the HPLC system at concentrations ranging from 0 g/1 to 0.8 g/1 (p-phenylenediamine), 0 g/1 to 2 g/1 (toluene-2,4-diamine and toluene-2,5-diamine sulfate), 0 g/1 to 0.3 g/1 (m-aminophenol), 0 g/1 to 0.8 g/1 (resorcinol), 0 g/1 to 0.3 g/1 (hydroquinone), 0 g/1 to 0.08 g/1 (ot-naphtol and hydroxypropyl-bis-(N-hydroxyethyl-p-phenylenediamine) HC1, and 0 g/1 to 0.1 g/1 (2,4-diaminophenoxyethanol). The analysis was repeated 15 times in order to test the repeatability. Statistical evaluations and quality assurance. For the chromatography, the quality of the separations was evaluated according to different criteria classically considered in HPLC. The repeatability of the separation was calculated for each hair dye and was based on both the retention time and the peak area. Three other criteria were taken into account to evaluate the quality of the column, namely the resolution R, the peak asymmetry factor s, and the capacity factor k' (7). This capacity factor enables a comparison between columns of the separation of selected compounds. For the calibration, the regression type used was as described previously (7), i.e., the "trendline with intercept set at zero" regression. The pipettes were tested both by gravimetry and colorimetry, and the cal- culation of the uncertainties was performed according to international standards (8). RESULTS AND DISCUSSION An analytical procedure was optimized and validated for the analysis of commonly used oxidative hair dye intermediates. Five commercial formulations were tested using this procedure, each of them containing initially four to six active hair dyes according to the ingredient labeling. For two of these formulations [one brown shade (DBRS) and one blonde shade (DBLS), shampoos], quantification was performed for target hair dyes before or after spiking. CHROMATOGRAPHIC SEPARATION AND IDENTIFICATION OF THE PEAKS IN NON-SPIKED FORMULATIONS Each sample prepared in mixtures of MeOH and Soerensen buffer (pH 8.1) was sub- mitted to chromatography using the gradient 1 after extraction. Each analysis was repeated three times in order to test the repeatability of both the extraction and the

OXIDATIVE HAIR DYES 107 chromatographic separation. As an example, the chromatographic separations obtained for the dark brown shade cream (DBRC) and the light blonde shade shampoo (LBLS) are depicted in Figure 1. Similar types of chromatograms were obtained for the three other formulations. Obviously, most of the matrix compounds are 100% extracted, and the chromatograms show only seven (brown shade cream, DBRC) to nine (blonde shade shampoo, LBLS) peaks. The resolution, calculated according to reference 7, was excellent for all two consecutive peaks except for res and oap (peaks 7 and 8 in Figure la), which were coeluted in these conditions in the dark brown shade cream (DBRC). All the calculated resolutions were greater than R6, 7 = 2.30 (see peak labeling in Figure 1). The identification of the hair dye intermediates could be easily performed using indi- vidual retention times and UV spectra. Indeed, these two characteristics having been previously recorded in a database containing 68 hair dyes and matrix compounds in pure standard solutions commonly found in oxidative hair dye formulations (9), the identi- fication could then be performed by comparison between the experimental data obtained and the recorded data. Table II shows the identification obtained using the retention time. Peaks 3, 4, 5, 7, 11, and 12 in Figure lb are attributed to p-aminophenol, m-aminophenol, toluene-2,5-diamine sulfate, resorcinol, 6-hydroxyindole, and 2-meth- yl-5-hydroxyethylaminophenol in the light blonde shade shampoo (LBLS), while peaks 1, 4, 6, 7, and 8 in Figure la are attributed to p-phenylenediamine, m-aminophenol, 2,4-diaminophenoxyethanol, resorcinol, and 0-aminophenol in the dark brown shade cream (DBRC). Peaks 15 or 16 in this cream may be attributed to hydroxybenzomor- pholine, a hair dye intermediate that appears in the ingredient labeling but that we did not consider for this validation study. Peaks 0 (attributed to sodium ascorbate) and M are due to remaining matrix compounds not interfering with the dye separation in accordance with previous observations (6,7). The repeatability of the analysis was also investigated, in terms of retention times and peak areas for the ten hair dyes in the five formulations. On the whole, the repeatability of the separation is high for both the retention times (0.12%RSD 1.91%) in the five formulations and the peak response intensity (0.90 % RSD 3.35 %) (data not reported). The quality of the column was also evaluated by calculating the peak asymmetry factor s and the factor capacity k' of the peaks attributed to the hair dyes according to the formulas described in reference 7. Results are given in Table III. The peak asymmetry factor is very close to the optimum value of 1 (10) for the peaks ofp-aminophenol (peak 3 in the light blonde shade shampoo, LBLS), 2-methyl-5-hydroxyethylaminophenol (peak 12 in the light blonde shade shampoo, LBLS), m-aminophenol (peak 4 in all five formulations), and 6-hydroxyindole (peak ! ! in the light blonde shade shampoo, LBLS). For p-phenylenediamine (peak ! in the dark brown shade cream, DBRC), the peak asymmetry factor is 0.8, showing a light fronting of the peak. For resorcinol (peak 7 in the light blonde shade shampoo, LBLS) and toluene-2,5-diamine sulfate (peak 5 in the light blonde shade shampoo, LBLS), the s value is greater than 1, what commonly occurs in HPLC. Moreover, it was noticed that the peak asymmetry factor is very close to the optimum value of 1 for hydroxypropyl-bis-(N-hydroxyethyl-p-phenylenediamine) HCI in all three formulations where present [dark brown shampoo (DBRS), light brown shampoo (LBRS), and dark blonde shampoo (DBLS)] and for resorcinol in the dark brown shampoo (DBRS) and in the light brown shampoo (LBRS). Concerning the k' factor, the values calculated for each hair dye peak range from 1. ! !