FLUORESCENT WHITENING AGENTS IN COSMETICS AND LIQUID DETERGENT 285 of the matrix. Figure 3 shows that some impurities which had ultraviolet absorption at 350 nm could not interfere in the determination of FWAs in samples by FLD. However, the retention time was the solely qualitative basis of FLD method, whereas both retention time and ultraviolet spectrum of analytes were the qualitative basis of DAD method. There- fore, these two detectors were combined to use their respective advantages. If impurity peaks were present near the retention time of the analytes in the chromatogram of FLD, they might be distinguished by comparing the ultraviolet spectrum using the DAD. The method was suitable for laboratories that did not have access to HPLC-tandem MS. Com- pared with ordinary HPLC that use DAD or FLD sole, the confi rmatory ability of this method was improved. Moreover, the limits of detection and quantization were close to HPLC-tandem MS method. The distyryl-type FWAs in aqueous solution were in the form of anion, the molecular struc- ture of analyte steric showed weak retention and could not be completely separated on a reversed-phase chromatographic column. Hence, it was necessary to use ion-pairing reagent to enhance the binding ability of the object to be measured and the stationary phase. To use the mobile phase in the further research of HPLC-tandem MS method with slight improvement, a suitable ion-pairing reagent was needed. Ordinary ion-pairing reagents suppressed the ESI signal and cause contamination of the ion source, especially nonvolatile tetraalkylammonium salts (20,21). Volatile di- and trialkylammonium acetates provide similar separation selectivity to that of tetraalkylammonium salts but greatly reduce ESI signal suppression. Among them, DHAA was compatible with MS detection, increasing Figure 3. Chromatography of cosmetic sample which contained fi ve fl uorescent whitening agent (FWA) standards by fl uorescence detector (upper) and diode array detector (below).

JOURNAL OF COSMETIC SCIENCE 286 the retention and improving the separation of FWAs. Thus, DHAA was selected in the present work. Because the ion-pairing reagent caused damage to the chromatographic column, concen- tration of acetonitrile in the mobile phase was decreased to reduce the use of DHAA. Various concentrations of DHAA (1–5 mmol·l-1, in the ammonium acetate solution) and acetonitrile (30–40%, in the mobile phase) were evaluated to determine the mobile phase composition that offers the highest signal-to-noise (S/N) ratio and suffi cient resolution for the FWAs. The highest S/N ratio and good separation of these fi ve analytes were achieved with 2 mmol·l-1 DHAA as the ion-pairing reagent to form neutral ion pairs, which promote the hydrophobic interaction between the analytes and the C18 stationary phase. Under this condition, the best concentration of acetonitrile in the mobile phase was 35%. Moreover, the ratio of ammonium acetate, which was commonly used in the HPLC anal- ysis for improving separation, was optimized. Lower concentrations of ammonium acetate produced excessively wide peaks and long retention times, whereas higher concentrations of ammonium acetate led to insuffi cient separation of target compounds and impurities. Hence, the best composition of ammonium acetate was 10 mmol·l-1. After optimization, the initial mobile phase was determined to be a mixture of 35% ace- tonitrile and 65% 2 mmol·l-1 DHAA in 10 mmol·l-1 ammonium acetate. However, the retention time of FWA351 was too long in this situation. Thus, a gradient elution pro- gram was adopted. The concentration of acetonitrile was linearly increased to 45% from 12 to 24 min, and then decreased back to 35% from 24 to 25 min, at last, maintained the concentration for 10 min. The effects of three C18 columns, including Shiseido MG C18 (Shiseido, Tokyo, Japan), Merck RP18 column (Merck, Darmstadt, Germany), and Agilent SB C18 (Agilent, Palo Alto, CA), on the separation of the fi ve target FWAs were investigated. Chromatographic peaks of some FWAs were tailed in Merck RP18 column and Agilent SB C18 column, whereas good shape of the peak was observed using Shiseido MG C18 column. Ionic surfactant, a commonly used component in liquid detergent, might interact with ion-pairing reagents in the mobile phase and led to the weak combination of the FWAs and ion-pairing reagents. Because of this, the retention time of some analytes in detergent samples became shorter than the retention time of the same FWAs in standard solutions. Figure 4. C hromatography of standard solutions which contained fi ve fl uorescent whitening agents (FWA) standards.