ANALYSIS OF N-NITROSODIETHANOLAMINE 325 both standard and sample to dryness and reconstitute in 100/xl of H20. Inject 25/xl of reconstituted standard and sample into Partisil PXS column (C•8). NDE1A is quantitated using an SP4100 computing integrator (in the external standard method mode). RESULTS AND DISCUSSION Numerous methods for the trace analysis of nitrosamines have been developed and published (4). The primary technique used has been High Pressure Liquid Chromatog- raphy with detection by a Thermal Energy Analyzer. The majority of nitrosamines studied have been volatile and presented little separation problem using gas chroma- tography, TEA detection. However, NDE1A, a polar nitrosamine, is very amenable to separation by reverse phase liquid chromatography using either water/alcohol or water as the mobile phase. The TEA's inability to handle an aqueous mobile phase is thus alleviated by quantitating the nitrosamine with a UV detector. We have found that when reverse phase liquid chromatography is extended to the analysis of commercially available alkanolamide, it is sometimes difficult to separate NDE1A from other early eluting materials, resulting in high values for the detection limit. It is also accepted that low recoveries can occur when NDE1A is isolated by multi-extractions and open column chromatography. We have developed a system by which the NDE1A is []. Z •. •BE [] LL •. •= U9 n, m9 [] z z NF=INI-II-,RF=IM'5 __%F = lIKED Figure 1. Linear regression line of linoleamide DEA spiked at six different levels.

326 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS isolated from the amide by normal phase HPLC, followed by quantitation on a reverse phase HPLC system (3). Figure 1 shows a linear regression line of a sample of Linoleamide DEA spiked at six different levels, while Figure 2 shows the chromatogram of a spiked and unspiked sample. The lowest spiking level was a 1.687 ng spike, which gave 1.685 ng found in the sample. Our absolute limit of detection is 1 ng to the detector which, at a 30% dilution, gives a limit of detection in the sample of 267 ppb. By increasing sample size and final injection volume, the limit of detection for the sample can be increased without deterioration of the chromatography. In conclusion, we have been able to determine NDEIA levels in this cosmetic raw material by a simple sample preparation, followed by micro-handling techniques. Solvent extraction and open column chromatography have been avoided and recover- ies are close to 100%. Figure 2. Reinjected collections of both spiked and unspiked samples of linoleamide DEA.