JOURNAL OF COSMETIC SCIENCE 430 the major colored species formed through the process (6). Because the rate of formation of the Bandrowski’s base 4 is slower than that of the reaction of p-phenylendiamine with the couplers, the literature data exclude the fact that in these conditions, even if more favor- able, this compound could be formed (2–5). The permanent hair-coloring process employs hydrogen peroxide as an oxidant and in- volves the in-fi ber formation of indo dyes. According to Corbett, the hair could infl uence the relative diffusion rates of the reagents inside the fi ber (7). For this reason it has been concluded that knowledge of the kind of colored molecules produced by these reactions in aqueous solution could not be easily extended to hair dyeing (7). On the other hand, the quantization of the colored species formed in the hair appears diffi cult because of the complexity of the reactions. In fact, in some cases the species 3 and 5 may also be trans- formed to high-molecular-weight colored molecules (6,8). For these reasons, the predic- tion of the chromatic outcome performed by a mixture of precursors and couplers in the hair dyeing process is, at present, an unsolved question. Since the solution of this problem is related to the overall characteristics of the color ob- tained at the end of the dyeing process, colorimetric analysis could be a useful tool in this investigation. In fact, the refl ectance of the dyed support can be considered a useful indi- cation of the colored pigments produced in the dyeing process. In this respect, investiga- tion of the color refl ectance variations produced on a support by a couple precursor–coupler under different reaction conditions could be a new and useful approach to the problem. Our study on the color and on the color variations induced on a model substrate by the reaction of p-phenylendiamine with different couplers at different pH values in the pres- ence of hydrogen peroxide constitutes a fi rst step in this direction. Virgin wool, being of a proteic nature, was considered suitable as a model substrate for human hair. Some dif- ferences exist between hair and wool, but we consider this aspect subordinate to providing homogeneous samples without complicating pigments. In addition, wool fabric has been used by others as a model support for human hair for predicting the tanning reaction of dyes (9). Resorcinol, 2-methylresorcinol, 3-aminophenol, and 2-amino-4-(2-hydroxyethylamino) Figure 1. General pathway of reactions of p-phenylendiamine with meta-derivatives of benzene.

REACTION OF PPD AND BENZENE DERIVATIVES 431 anisole sulfate, employed in current hair dye formulations, were chosen as model cou- plers. Seven pH values in the range of 7.5–10.5 were investigated. MATERIALS AND METHODS MATERIALS Technical grade p-phenylendiamine (PPD), resorcinol (RES), 2-methylresorcinol (MeRES), 3-aminophenol (AP), 2-amino-4-(2-hydroxyethyl-amino)anisole sulphate (AHA), and perhydrol were kindly provided as a gift by Tocco Magico (Italy) and utilized without further purifi cation. Perhydrol, stabilized with phosphoric acid, was opportunely diluted to obtain hydrogen peroxide at 20 volumes. TLC was performed on preparative 20 × 20-cm, 1-mm, Silica gel 60 F254 plates from Merck (Germany). METHODS The commercial virgin wool tissue, cut in 5 × 20-cm strips, was washed with a non-ionic detergent aqueous solution, rinsed with water, and dried before the coloration. Each strip was dipped in 100 ml of a solution formed by 50 ml of hydrogen peroxide (20 volumes, 6%) and 50 ml of a solution containing the p-phenylendiamine-coupler (1:1 ratio) and the ammonium hydroxide needed to obtain the fi nal required pH value. After 30 minutes the strip was removed, carefully rinsed with tap water, and dried at 45°C. In each experiment the reagents’ concentration (precursor and coupler) was 2 × 10−2 M. Experiments were carried out at pH 7.5, 8.0, 8.5, 9.0, 9.5, 10.0, and 10.5. pH values were monitored at room temperature with a Metrohm pH-meter 744. The colorimetric measurements were performed with a colorimeter X-Rite SP62 model, with an integration sphere to determine the color refl ectance. The L, a*, and b* values reported are related to the illuminant D65 at 10 degrees. Each reported value is the mean of ten measurements, performed randomly on the surface of the colored strips, with standard deviation never exceeding 1.2%. In these conditions the colorimetric data of virgin wool were: L = 95.12, a* = 0.05, and b* = 6.73. TLC experiments were carried out by spotting about 0.5 ml of the aqueous solution of the performed reaction on a strip of 5 cm and eluting with CHCl3:EtOAc:EtOH 7:2:1. For the weathering experiments, two wool strips were colored with the color mixture obtained with each couple of reagents at the stated pH. One of these was directly exposed to atmo- s pheric agents the other one was weathered, avoiding a direct exposure to sunlight in order to distinguish the color degradation induced in the presence or in the absence of direct sunlight. The weathering experiments were carried out for 24 days. At days 0, 4, 7, 11, 14, 18, and 24 the color of each strip was measured. All the experiments were performed in duplicate. RESULTS AND DISCUSSION Preliminary experiments showed that for each pair of precursor-coupler, the color refl ectance produced on the wool strips was constant (R∞) when the reagents’ concentration was ≥1 × 10−2 M. As a consequence, a 2 × 10−2 M concentration of the two reagents was employed in all the reported experiments. The values of the colorimetric parameters (L, a*, b*), from each pair of p-phenylendiamine-coupler at different pH values, are reported in Table I.