DETERMINATION OF TOTAL LEAD IN LIPSTICK 407 equipment and is subject to matrix absorption errors. AAS with electrothermal atomiza- tion analysis was also considered. However, since sample digestion would be necessary, inductively coupled plasma–mass spectrometry (ICP–MS) was chosen because of its po- tential for better sensitivity and speed. Thus, the effort focused on developing a method using microwave-assisted acid digestion for sample preparation and determination by ICP–MS. EXPERIMENTAL CHEMICALS, REAGENTS, AND REFERENCE MATERIALS Twenty shades of lipstick sold in the United States under ten brand names were pur- chased from retail stores or provided by manufacturers. Multiple samples with the same lot number were obtained for several shades in order to compare analytical techniques. Six RMs were used for comparing different preparation techniques and for quality assurance: wear-metals in lubricating oil, SRM 1084a (NIST, Gaithersburg, MD) whole milk pow- der, SRM 8435 (NIST) estuarine sediment, SRM 1646a (NIST) lead in base oil 20 standard, ORG-PB8-2Y/Z (SPEX CertiPrep, Inc., Metuchen, NJ) base oil 20 standard (SPEX CertiPrep, Inc.) and trace elements in coal, SRM 1632c (NIST). American Society for Testing and Materials (ASTM) type 1 grade water was used to pre- pare reagents, standards, and analytical solutions. Pb standards (0, 0.1, 0.5, 1.0, and 10 ng Pb/ml), Pb stock (0.1 and 1.0 μg Pb/ml), and thallium internal standard (0.1 μg Tl/ml) solutions were prepared from commercial ICP–MS grade single-element analyte solu- tions (High-Purity Standards, Charleston, SC). Trace metals grade (TMG) HNO3 (Fisher Scientifi c, Pittsburgh, PA) was used for cleaning laboratory ware and digestion vessel lin- ers. Optima grade HNO3 and HF (Fisher Scientifi c) were used for sample, stock, and standard solutions. Boric acid (Puratronic grade, Alfa Aesar, Ward Hill, MA) was used to prepare 4% boric acid solution, which was conveniently dispensed with a bottle-top dis- penser. A 0.100-μg Pb/g in base oil 20 stock solution was prepared from 1000 μg Pb/g (SPEX ORG-PB8-2Y/Z organo-metallic standard solution) serially diluted to 10.00 and then to 0.100 μg Pb/g with SPEX base oil 20. Lipsticks were digested using XP-1500 Plus vessels in a MARS microwave digestion oven (CEM Corp., Matthews, NC). Pb determinations were performed on an Agilent 7500c ICP–MS (Agilent Technologies, Inc., Santa Clara, CA) equipped with a Peltier cooled Scott double-pass spray chamber and a MicroMist nebulizer (Glass Expansion, West Melbourne, Victoria, Australia). The built-in peristaltic pump was used to deliver the analytical and thallium internal standard solutions to the nebulizer at 0.17 ml/min and at 0.01 ml/min, respectively. The analytical and internal standard solutions were merged with a Tee fi tting. METHOD DEVELOPMENT Lipstick is a challenging matrix of many ingredients including waxes, oils, dyes, and pigments (11). The pigments may include refractory minerals such as alumina, silica, titanium dioxide, and mica. Preliminary experiments with one lot of lipstick revealed an easily detectable amount of Pb, but quantitative results varied depending on preparation technique. In order to evaluate preparation techniques, a composite was prepared by

JOURNAL OF COSMETIC SCIENCE 408 melting and mixing together eleven tubes of lipstick (same brand and shade, but several lot numbers). These lipsticks were placed in a beaker submerged in a water bath at 85°C and stirred with a propeller-type mixer. Initial attempts to completely dissolve a lipstick sample using typical microwave- assisted HNO3 digestion were unsuccessful. A cloudy, white suspension remained after the treatment. Also, replicate results for a single lot of lipstick showed variations in Pb recovery for different lipstick portion sizes and maximum digestion temperatures (see Table I). The portion size and temperature effects on Pb recovery suggested that microwave-as- sisted digestion with HNO3 was incomplete for Pb and that levels of Pb recovered would vary depending on digestion parameters. Therefore, other preparation procedures were investigated that might achieve total recovery of Pb. A dry ash procedure was tried in which a lipstick sample was heated gradually to 540°C and held at that temperature for an hour. The resulting residue was treated with HNO3 and HCl but would not com- pletely dissolve. The acid leachate was diluted with water and analyzed by ICP–MS. Results were variable and lower compared to values obtained with microwave-assisted acid diges- tion. Similarly, sodium carbonate fusion at 1075°C resulted in a residue that would not completely dissolve in HNO3 or HCl and produced variable values for Pb. The results indicated that some Pb must be associated with the undissolved refractory mineral matter in the lipsticks and suggested that HF would be necessary to break down the minerals. Based on work of other investigators (10), the initial HF dissolution procedure used 0.3-g portions of lipstick, 7 ml HNO3 + 2 ml HF, and sealed Tefl on microwave digestion ves- sels. A two-step procedure was used for digestion. The vessels were heated to 130°C over 15 minutes and held at that temperature for three minutes before ramping to 200°C over 15 minutes and holding for an additional 30 minutes. The vessels were allowed to cool to 50°C and then were vented. Initially 6 ml of 4% boric acid was then added to each vessel and the solutions were heated to 170°C over 15 minutes and held for ten minutes to complex the HF. Boron forms a strong complex with fl uoride according to the reaction shown in equation 1. After cooling and venting, the solutions were diluted to a fi nal vol- ume of 200 ml for ICP–MS analysis. However, the digests were still cloudy, with a ge- latinous precipitate appearing upon centrifugation. Also, results varied depending on the analytical portion. Equation 1 is as follows: 4HF + B(OH)3 → HBF4 + 3H2O (1) Therefore, an excess of boric acid was used to dissolve any insoluble fl uorides. The revised HF digestion procedure used 30 ml of 4% boric acid and resulted in clear solutions for all Table I Analytical Portion and Temperature Effects on Pb Determined After Digestion with Nitric Acid Portion size (g) Maximum digestion temperature (°C) μg Pb/g 0.3 240 1.4 0.1 240 2.1 0.3 200 0.5