Pressurised liquid extraction–comprehensive two-dimensional gas chromatography for fast-screening of polycyclic aromatic hydrocarbons in soil

Abstract

Pressurised liquid extraction (PLE) was applied to the extraction of polycyclic aromatic hydrocarbons (PAHs) from contaminated soils from Husarviken in Stockholm, Sweden. The extraction step was followed by conventional gas chromatography (GC), comprehensive two-dimensional gas chromatography (GC×GC) (both with flame ionisation detection) and gas chromatography–quadrupole mass spectrometry (GC–MS) analysis. Qualitative and quantitative aspects of the results are considered. Qualitatively, results from all chromatographic analyses are in good agreement, and PLE provides a reliable extraction technique with all PAHs extracted in one extraction step; no carry over was observed. With respect to PAH quantification, some variability in results was noted, with better agreement in PAH concentrations for GC and GC×GC measurement, as compared to GC–MS. GC analysis compares favourably with GC×GC apart from the few exceptions where peaks are not fully resolved from other co-extracted analytes, which compromises GC-FID measurement. For example, acenaphthene shows a much higher concentration when measured by GC-FID, demonstrating the superior separating powers of GC×GC; the latter is the preferred technique if precise and accurate quantification of analytes are required. GC–MS results compare reasonably with GC×GC for low-molecular mass PAHs but not for high-molecular mass PAHs; results for GC–MS are consistently higher than those for GC×GC for high-molecular mass PAHs. Since PLE–GC×GC is proposed as a broad screening tool, the demand for precise quantification may be relaxed in the present situation. GC×GC has the added advantage of providing chemical structural information within the two-dimensional contour presentation. Reproducibilities for GC×GC results (peak area) and 2t R were acceptable with relative standard deviations (R.S.D.) of 8 and 1%, respectively (at the mg/kg level), and good repeatability within samples was achieved.