JOURNAL OF COSMETIC SCIENCE 160 Analytical techniques requiring a solubilized analyte fall into two preparation catego- ries: (a) acid extraction of the analyte from the sample matrix, or (b) total destruction of the matrix. We chose to use a total matrix destruction technique for the validated method. Nitric acid has the desirable properties that at high temperatures can break down organic (carbon-based) molecules such as the oils and waxes in lipstick, and it also solubilizes lead as lead nitrate. Nitric acid, however, will not break down certain inor- ganic minerals, such as talc, mica, or titanium dioxide, which are commonly used in lipstick. In developing a method, we found that not all the lead was extracted from lip- stick with nitric acid alone. After trying several total matrix destruction methods, we modifi ed a method developed by scientists at the Cosmetics, Toiletries, and Fragrances Association (currently the Personal Care Products Council) that included a hydrofl uoric acid step for breaking down the inorganic minerals (7). We then found it necessary to add an excess of boric acid to dissolve the resultant insoluble fl uorides. Analyses of the resulting clear solutions by inductively coupled plasma-mass spectrometry (ICP-MS) gave reproducible lead values even under changing preparation conditions, such as using higher maximum digestion temperatures or reduced portion sizes. We used the method to determine lead in the initial survey of 20 lipsticks. Because the initial survey was small and from a limited segment of products, the FDA decided that an expanded survey of the U.S. lipstick market was needed. Since the FDA had limited time and personnel for the task, a contract laboratory was solicited to perform the study using the validated method. EXPERIMENTAL SELECTION OF CONTRACTOR The contractor was awarded a contract for the expanded survey on September 15, 2009, based on its demonstrated experience with lead analyses in matrices similar to lipstick and the availability of appropriate facilities, equipment, and personnel to meet technical, time, and budget criteria. Prior to analyzing the lipstick samples, the contractor was asked to demonstrate its ability to obtain values that agreed with those for three reference lipsticks, as well as to meet specifi c performance criteria regarding limit of detection, quality control, and precision. LIPSTICK SAMPLING Lipsticks for the expanded survey were selected by choosing manufacturers, brands, and shades that refl ected the products currently on the market. This was accomplished by proportioning the number of lipsticks chosen from each brand to its market share using data from the 2008 Euromonitor Report (8). A few additional lipsticks were selected from niche markets in an effort to capture lipsticks with unusual characteristics. A total of 400 lipsticks with a variety of shades, with no particular targeted colors, were selected from 24 parent companies. One pair of lipsticks per manufacturer with the same lot num- ber (and therefore the same shade) was collected as an additional test of lot homogeneity. Each of the 400 lipsticks was tested in duplicate.

DETERMINATION OF LEAD IN LIPSTICKS 161 APPARATUS Lipsticks were digested in CEM EasyPrepTM vessels on a CEM MARS microwave diges- tion system (CEM Corporation, Matthews, NC). The analytical solutions were stored in metal-free grade polypropylene centrifuge tubes, with additional acid cleaning to remove contamination (50 ml, Tyco Healthcare Group, Mansfi eld, MA, and 15 ml, VWR Inter- national, Radnor, PA). Lead analyses were performed on a Perkin Elmer Elan 6000 with DRC II ICP-MS (Perkin Elmer, Shelton, CT). REAGENTS High purity (18 MΩ) deionized water was used throughout. Lead standard (0, 0.01, 0.02, 0.04, 0.08, 1.25, 2.50, 5.00, 12.50, and 25.00 μg Pb/l), lead stock (0.100 and 1.00 mg Pb/l), and platinum internal standard (0.4 mg Pt/l) solutions were prepared from com- mercial ICP-MS grade single-element analyte solutions (CPI International, Santa Rosa, CA). Trace metal grade nitric acid (EMD Chemicals, Gibbstown, NJ) was used for sam- ple, stock, and standard solutions. Hydrofl uoric acid (Fisher Scientifi c, Pittsburgh, PA) was used in the sample solutions. Boric acid (Puratronic grade, Alfa Aesar, Ward Hill, MA) was used to prepare 4% boric acid solution. A 0.100 mg Pb/kg in oil stock solution was prepared from 1000 mg Pb/kg organo-metallic standard serially diluted to 10.00 and then to 0.100 mg Pb/kg with base oil 20 (organo-metallic lead standard and base oil, both from Spex CertiPrep, Inc., Metuchen, NJ). Nitric acid (1%, v/v) was used for preparing calibration standards and analytical solutions. Water standard reference material NIST SRM 1643e (National Institute of Standards and Technology, Gaithersburg, MD) con- taining 19.15 ± 0.20 mg Pb/kg, diluted 1/10 with 1% nitric acid, was used as a secondary standard to check ICP-MS performance. METHOD The protocol was consistent with the FDA’s validated method (1), and continued reliability of the results was verifi ed by using specifi c quality control procedures, as follows: • Weigh and transfer 0.3 ± 0.1 g of lipstick to a digestion vessel liner. • Add 7 ml of nitric acid and 2 ml of hydrofl uoric acid to the lipstick in the liner. • Seal and heat the vessels in the microwave digestion oven. Ramp the temperature over 20 minutes to 130°C and hold for three minutes, then continue to ramp the temperature over 20 minutes to 200°C and hold for 30 minutes. Allow the samples to cool to less than 50°C. • Vent and open the vessels in a fume exhaust hood. • Add 30 ml of 4% boric acid solution to each. • Reseal and heat the vessels in the microwave digestion system. Ramp the temperature over 20 minutes to 170°C and hold for ten minutes. Cool, vent, and open the vessels in a fume exhaust hood. • Quantitatively transfer the contents of the digestion vessels to 50-ml centrifuge tubes and dilute to 50 ± 0.25 ml with water. • Include two method blanks in each digestion batch to assess contamination.