Comparison of supercritical fluid extraction and Soxhlet extraction for the analysis of native polycyclic aromatic hydrocarbons in soils

Abstract

Abstract Two supercritical fluid extraction (SFE) methods were compared to Soxhlet extraction for the analysis of native polycyclic aromatic hydrocarbons (PAHs) in soil. Dichloromethane cosolvent was added directly to the soil sample in the “Manual” method, whereas isooctane-dichloromethane (1:3, v/v) was introduced to the CO 2 during dynamic extraction in the “Pump” method. Recoveries of native PAHs from reference marine sediment HS-4 using the Manual and Pump SFE methods were generally equivalent to or better than those using Soxhlet extraction. The recoveries of five deuterated surrogate PAHs spiked onto ten different soils were generally greater than 70% with intra-assay precisions of ±7 and ±4% for the Pump and Manual SFE methods, respectively. For the analysis of native PAHs present in these soils, the average intra-assay precision of the SFE methods was ±16%, which was similar to that for Soxhlet extraction of ±13%. The accuracies of the observed native PAH concentrations in these ten soils using SFE (relative to Soxhlet extraction) were generally low and variable: both the Pump and Manual methods showed average accuracies of 0.6±0.4. Analysis of variance showed that ca. 85% of this accuracy variability was due to an apparent matrix effect (relative to Soxhlet extraction), and that ca. 15% was due to error in the SFE analytical methods. Due to this high apparent matrix effect observed for these ten soils analyzed, the two SFE methods tested were not validated for the analysis of client samples. Additional sample processing, such as cryogenic grinding, may be required for SFE.