Evaluation and reduction of the analytical uncertainties in GC-MS analysis using a boundary regression model

Abstract

The uncertainties in analysis of trace environmental pollutants may come from sample matrix and sample pretreatment process. In this study, a boundary model was developed based on the visualized data of mass spectrometry to evaluate the influences from sample matrix and pretreatment process. The factors affecting the pretreatment procedures, such as the solvents, the extraction sorbents and the extraction process had limited influences compared with matrix effects. Using such boundary model, we found that selecting suitable qualitative and quantitative ions for MS detector is more important for reducing the matrix effect in GC-MS analysis than the traditionally used methods of optimizing the pretreatment process since some clean up sorbents might be useless to reduce the matrix effects. As for 2,2′,4,4′,5-pentabromodiphenyl ether (BDE-99), the fragmental ions were usually used for qualitative and quantitative analysis, which however was easily affected by the matrix effects. While, molecular ions would eliminate the influences from the sample matrix. Such a model could be used to decrease the uncertainty and increase the accuracy of environmental trace analysis.