Analysis of perfluorinated carboxylic acids in soils: Detection and quantitation issues at low concentrations

Abstract

Methods were developed for the extraction from soil, identification, confirmation and quantitation by LC/MS/MS of trace levels of perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA) and perfluorodecanoic acid (PFDA). Whereas PFOA, PFNA and PFDA all can be quantitated using the method of standard additions, PFOA also can be quantitated less laboriously using 13C 4-PFOA as a matrix internal standard. The impact of extract matrices on signal varied between soils and temporally during analytical runs rendering 13C 4-PFOA unsuitable as a matrix internal standard for quantitating perfluorinated carboxylic acids (PFCAs) other than PFOA, which co-elutes with 13C 4-PFOA. In fact, for soil extracts, quantitation of PFCAs based on external calibrations proved about as accurate as use of matrix internal standards for target analytes that do not co-elute with the matrix internal standard. Also, 13C 4-PFOA should be used carefully as a matrix internal standard for trace levels of PFOA because some 13C 4-PFOA standards contain trace impurities of unlabelled PFOA. When the presence of PFCAs in soil extracts is being determined by LC/MS/MS, detection limits are best defined by statistical methods that quantify the significance of contrast between analytical signal and background noise using multiple analyses. Further, when developing a calibration of low concentrations using weighted regression, the central tendency of the calibration line is best fitted using graphical depictions of error. As the MDL for the transition-product quantitation ion is approached in LC/MS/MS, relatively weak signals of transition-product confirmation ions can be used as a rejection criterion by looking for anomalously high values of the ratio of the confirmation to the quantitation ion.