RESUMO
Two well known derivatization procedures, pentafluorobenzylation and BF(3)/methanol esterification, were compared for their applications to GC analysis of acidic water micropollutants (chloroacetic and phenoxyalkanoic acids). A two-level factorial design was used to determine the influence of different parameters and their interactions on each derivatization process. The studied parameters are the reaction time, the amount of reagent (PFBBr) or catalyst (BF(3)) and the temperature. Considering pentafluorobenzylation, the most influential factors are the concentration of PFBBr and the interaction ;temperature-time', which improve the derivatization efficiency. However, a PFBBr concentration of 250 mg l(-1) in the reaction medium cannot be exceeded because of the increase in interfering by-products in GC/ECD. Moreover, chloroacetic acid derivatives are co-eluted with these compounds. This disadvantage was not observed in the operating conditions of GC/MS. The improved pentafluorobenzylation procedure allows the direct determination of the derivatives in GC/ECD without any purification step. The average detection limits are 1.6 and 80 mug l(-1), respectively in GC/ECD and in GC/MS. The reproducibility is 13%. For the BF(3)/methanol esterification, the interactions ;BF(3) concentration-temperature' and ;BF(3) concentration-reaction time' are significant and have a negative effect on the derivatization yield. A linear model was therefore proposed and validated in the experimental area under study. All the compounds studied were detected in GC/MS, and the average detection limit is 2 mug l(-1). The reproducibility is around 7%. Therefore, after optimization, BF(3)/methanol esterification followed by GC/MS is as sensitive as pentafluorobenzylation used with GC/ECD, and more reproducible.