216 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table II HPLC Parameter Evaluation for NDELA Determination (•Bondapak CN Column) Eluent Flow NDELA a Rate k' PAF b W0,1 t (min) 75/20/5 Isooctane/isopropanol/methanol 50/40/10 Isooctane/isopropanol/methanol 90/8/2 Isooctane/isopropanol/methanol 82/14.4/3.6 Isooctane/isopropanol/methanol 60/34/4 Isooctane/methylene chloride/methanol 65/26.25/8.75 Isooctane/methytene chloride/ methanol 75/19/6 Isooctane/methylene chloride/methanol 3.0 1.0 1.0 0.6 2.0 1.2 0.7 0.7 1.0 1.0 1.2 0.9 1.0 1.5 0.6 1.3 2.0 5.0 2.3 1.7 2.0 1.6 0.8 0.9 1.0 1.0 1.1 1.1 1.0 1.8 1.8 1.2 1.0 1.0 1.3 1.0 3.0 3.1 0.8 0.6 1.5 2.3 1.5 1.0 1.0 2.3 1.8 1.3 elution volume NDELA--void volume void volume width of back half of peak at 10% height b PAF = peak asymmetry factor = width of front half of peak at 10% height c Wo. 1 = width of peak at 10% height. NDELA confirmation. For those samples with quantifiable levels of NDELA, the HPLC- TEA NDELA peak was confirmed as a nitrosamine by transferring an aliquot of the sample solution to a culture tube and irradiating it with a UV lamp (Hanovia Utility Ultraviolet Quartz Lamp, Engelhard Hanovia, Inc., Newark, N.Y.) at a maximum wavelength of 360 nm. An aliquot of the irradiated sample was reanalyzed by HPLC- TEA and the reduction in the HPLC-TEA NDELA peak measured. After 30 rain of irradiation, approximately 90% reduction in the NDELA peak could be expected. RESULTS AND DISCUSSION Since water is incompatible with the TEA detector, only nonaqueous HPLC systems were investigated. Three columns were evaluated with a variety of eluents: Spherisorb Silica (5 •tm), •tBondapak CN, and Spherisorb CN (5 p•m). As can be seen from Tables I and II, which give HPLC parameter evaluations by column, the nitrile column (•Bondapak CN) gave superior chromatography for most of the test solvent systems. Using the nitrile column, the 75/19/6 isooctane/methylene chloride/methanol eluent gave the largest k' with good peak shape. Subsequent investigations found that the Spherisorb CN, 5 •m column gave superior resolution compared to that of the }xBondapak CN column, and the former column was specified as the column of choice for most applications. The sample workup procedure affects the choice of chromatographic conditions to some extent. If the sample extract is dissolved in a nonpolar solvent such as hexane, a silica column would be preferable, while the nitrile column would be more appropriate for a sample dissolved in ethyl acetate or other polar solvent.

HPLC-TEA ANALYSIS OF NDELA 217 Table III HPLC-TEA Characteristics of Potential Internal Chromatographic Standards • W0.1 d Compound k' •' PAD (min) N-Nitrosodibutylamine --0 0.8 0.6 --0 0.8 0.8 N-Nitrosodiisopropylamine Diphenyl nitrosamine •0 0.7 0.8 N-Methyl-N-nit rosourea 1.0 0.7 0.6 N-Ethyl-N'-nitro-N-nitrosoguanidine •0 0.7 0.7 N-B utyl-N'-nitro-N-nitrosoguanidine 0.5 0.6 0.4 1-[2-Chloroethyl-3-(4-methylcyclo- hexyl)]- 1-nitrosourea •0 0.5 0.6 1-(2-Chloroethyl)- 3-cyclohexyl- 1- nitrosourea •0 0.5 0.6 1-(2-Chloroethyl)-3-(2,6-dioxo-3- piperidyl)- 1-nitrosourea 0.4 -- -- 2-[3-(2-Chloroethyl- 3-nitrosoureido)]- 2-deoxy-D-glucopyranose 10 -- -- N-Nitrosoproline 10 -- -- N-Nitroso-3-hydroxypiperidine f f f N-Nitrosodiisopropanolamine f f f p-Toluenesulfonylmethylnitrosamide --0 -- -- N-Nitrosopyrrolidone 3.2 3.8 2.2 N-Nitrosopyrrolidine 0.8 1.2 0.6 • 1.1 I 0.7 HPLC conditions: Spherisorb silica column, 75/19/6 isooctane/methylene chloride/methanol. elution volume NDELA--void volume k' = void volume width of back half of peak at 10% height PAF = peak asymmetry factor = width of front half of peak at 10% height W0 '• = width of peak at 10% height. PH NDELA wt compound RWR = X PH compound wt NDELA Standard impure multiple peaks. SELECTION OF AN INTERNAL CHROMATOGRAPHiC STANDARD The response of the HPLC-TEA analytical system is not generally reproducible between injections. This problem can easily be compensated for with an internal chromato- graphic standard added to the sample at a known concentration in the step immediately prior to analysis. The ideal internal chromatographic standard is well-separated from both the analyte and potential interferents, elutes in a reasonable time, has good chromatographic properties, and has similar detection response. This last criterion requires that an internal chromatographic standard for HPLC-TEA analysis be a semi- volatile or nonvolatile (to prevent evaporative loss) N-nitrosamine. Sixteen nitrosamines were screened for their chromatographic characteristics and their TEA responses, as shown in Table III. While several compounds listed in the top portion of Table III exhibited small peak asymmetry factor, width, and RWR values, they eluted at or near the void volume and would present reproducibility problems due to potential matrix interferences. The peak height of the internal chromatographic standard is in the denominator of the RWR equation presented above thus, a small RWR in Table