We are trying to validate a quantitative limit test for 2-chloroethyl amine and hydroxylamine as a mixture employing GC-MS. These compounds are derivatized in Acetone/NaOH solution prior to analysis. We are facing 2 problems:

- We observe high RSD for 2-chloroethylamine (>10%) but not for hydroxylamine (<5%)
- 2-chloroethylamine shows poor correlation (R<0.98) but Hydroxylamine shows a good linear correlation (R> 0.99)

We have tried several different GC liners (split vs splitless etc, with/out glass wool) but for 2-chloroethylamine we always obtain higher RSD (>10%) while hydroxylamine compound has no issues. For SST, 2 ug/mL concentration of both compounds were used and a peak area of 15000 was obtained for 2-chloroethylamine (n=6) but when linearity sample of 2.5 ug/mL was analyzed a peak area of 350000 was obtained for 2-chloroethlyamine (about 20-25 fold). We did not observe this for hydroxylamine product.. We have also used FID as a 2nd detector. We have seen that no linear correlation was obtained for 2-chloroethylamine with MS (SIM mode) results but FID results showed linear correlation. We thought that the problem was fixed but once again we had the same problems: poor injector reproducibility (RSD>15%) and poor correlation (R<0.97). It seems like a product related problem since good results were obtained for hydroxylamine product. The stability of the derivatized 2-chloroethylamine might not be stable during the analysis. But also can we say that significant peak area differences might be caused by GC system. I would appreciate all replies and helps. Thanks