Identification of micropollutants by combined chromatography and mass spectrometry techniques

Abstract

The presence of micropollutants in South African aquatic systems has emerged as an issue of public health concern. Micropollutants, such as endocrine disrupting chemicals (EDCs) and antiretroviral compounds, have previously been detected in surface water of South Africa. Micropollutants are often present in complex environmental matrices at ultra-trace levels, complicating their detection. In order to address shortcomings with traditional sample preparation methods, an inhouse developed cheap, disposable polydimethylsiloxane (PDMS) sorptive sampler was developed. The validity of the PDMS sampler was established by comparison with a commercial stir bar sorptive sampler (SBSE). The sample introduction process into a gas chromatograph (GC) was also simplified by using thermal desorption of a PDMS sampler directly in the inlet liner of a GC. Direct thermal desorption was validated by comparison to time-consuming thermal desorption using an expensive commercial thermal desorption system (TDS). With the aim of identifying a vast range of micropollutants in aquatic systems comprehensive gas chromatography coupled to time-of-flight mass spectrometry (GC×GC-TOFMS) was employed. The increased selectivity, sensitivity and larger peak capacity of GC×GC-TOFMS allows the identification of more compounds in complex matrices when compared to conventional GC-MS. An initial screening using sorptive extraction techniques and GC×GC-TOFMS tentatively identified various micropollutants, including EDCs, in surface water samples from the Rietvlei Nature Reserve, Gauteng, South Africa. Ultra-high pressure liquid chromatography coupled to mass spectrometry (UHPLCQTOFMS) was used as a complementary analytical technique in conjunction with GC×GC-TOFMS. Solid phase extraction (SPE) and large volume injection (LVI) sample preparation steps preceded analysis by UHPLC-QTOFMS. SPE is more time consuming and uses expensive solvents, however, adds selectivity to the sample preparation step, by reducing possible matrix interferences which can be problematic with LVI. Matrix matched calibration curves were constructed to identify and quantify target analytes in surface water samples. After validation of the analytical methods using chemometric approaches, these methods were employed to detect micropollutants in surface water from a metropolitan area (Rietvlei Nature Reserve, Gauteng) and a rural area (Albasini and Nandoni Dams, Limpopo Province) in South Africa. Limits of detection (LOD) for the GC methods ranged from 1 to 98 pg/L for the PDMS loop and 1 to 190 pg/L for SBSE. The LODs for the LC methods ranged from 1.97 to 135 ng/L for LVI and 73 pg/L to 57.3 ng/L for SPE. The two simplified methods, the in-house developed PDMS loop with GC×GCTOFMS, and LVI with UHPLC-QTOFMS, were used as complementary methods to detect micropollutants, such as EDCs, in surface water. EDCs such as pharmaceuticals, personal care products and pesticides, as well as the antiretroviral compounds, efavirenz and nevirapine, were detected in surface water from South Africa at concentration levels ranging from 0.16 ng/L to 227 ng/L. As they have not been reported in literature before, experimental linear retention indices are provided for the target analytes on the proprietary phase Rtx®- CLPesticides II column. Lastly, the variance between different sampling sites was investigated using principal component analysis (PCA). PCA revealed a difference in micropollutant profile between sampling sites in the metropolitan and the rural area.