DETERMINATION OF ZPT IN SHAMPOOS BY HPLC AND HPLC-MS/MS 269 Then, the content of ZPT in water, twice methanol–acetonitrile extract were analyzed and compared. Table IV showed that very little of ZPT was dissolved in the water rinse step and most ZPT was extracted in the fi rst methanol–acetonitrile extraction step. Tiny ZPT was dissolved in the second extraction. Thus, in the routine tests, the samples only need to be extracted by methanol–acetonitrile once. Table III Gradient Elution Program (HPLC-APCI-MS/MS) Time (min) Ammonium acetate solution (%) Acetonitrile (%) Methanol (%) 0.00 70 15 15 5.00 5 45 50 6.00 5 45 50 7.10 70 15 15 10.0 70 15 15 Figure 2. Chromatography of sample with (A) and without (B) water washing.

JOURNAL OF COSMETIC SCIENCE 270 HPLC ANALYSIS Initial work investigated a range of reversed-phase columns for the determination of ZPT in an attempt to identify a stationary phase that was essentially free from metallic impu- rities and also exhibited very low silanol activity. Most C18 column resulted in very short retention time, low column effi ciency, or tailing peak. Fortunately, using a simple gradi- ent elution as depicted in Table I, the peak for ZPT on Shiseido MG C18 column eluted as a small fronted peak because of the high purity monolithic silica substrate and also the superior level of endcapping. Ingredients in the shampoos did not show any interference in the determination of ZPT. Moreover, other common antidandruff agents (salicylic acid, ketoconazole, climbazole, and piroctone olamine) could be analyzed in the same chroma- tography condition. HPLC-MS/MS ANALYSIS Through several tests, the Chromolith Performance RP-18e monolithic column outper- formed all HPLC-MS/MS particulate columns with a large sharp symmetrical peak, whereas many normal column leading to serious leading peak (Figure 3). For this reason, the Chromolith column was used for all further HPLC-MS/MS studies. According to the former reports, the ionization and MS response of ZPT was low thus, the research evaluates both positive and negative ionization modes of APCI. Poor ionization of ZPT was observed in the negative ion mode, whereas positive APCI produced both a quasi- molecular ion (m/z 317.0) and two characteristic fragments (m/z 173.8 and 189.8). These two product ions were chosen to meet the requirement of Identifi cation Points ≥ 4.0 by the Decision 2002/657/EC. The m/z 173.8 was chosen as quantitation ion and the m/z 189.8 as confi rmation ion. The relative ion abundance was measured at various levels spiked in blank shampoo samples. Both relative ion abundance ratios of 317.0/173.8 and 317.0/189.8 met the requirements set by Decision 2002/657/EC. To determine the optimum ionization, a standard ZPT solution was infused into the mass spectrometer to minimize fragmentation and maximize sensitivity. These parameters were automatically fi ne-tuned using the software and the collision energy for each daughter ion Table IV Content of ZPT (%) in Water, Twice Methanol–Acetonitrile Extract Extract Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Sample 6 Sample 7 Water 0.014 0.010 0.019 0.010 0.012 0.013 Not detected First methanol–acetonitrile extraction 1.21 0.47 1.33 0.99 0.85 0.66 0.29 Second methanol–acetonitrile extraction 0.016 0.009 0.019 0.015 0.016 0.010 Not detected Shampoo suppliers said that these samples contained about 1.3%, 0.5%, 1.4%, 1.0%, 0.9%, 0.7%, and 0.3% ZPT, respectively.