A matter of extraction - Extraction yields and ratios of faecal lipid biomarker from archaeological soils using Soxhlet, microwave-assisted and accelerated-solvent extraction

The performance of Soxhlet (SOX), microwave-assisted (MAE) and accelerated-solvent extraction (ASE) in the analysis of faecal lipid biomarkers (FLB, Delta(5)-sterols, stanols, stanones) from archaeological soils was investigated to assess effectiveness and reproducibility of the extraction methods. Results from two Anthrosols that were analysed in six replicates show that SOX achieves significantly higher extraction yields for individual substances and steroid sums than MAE, while ASE produces the lowest lipid yields. Regarding the FLB ratios, which are used for sourcing of faeces, the three extraction methods show comparable values with three out of five ratios differing significantly between different soil samples. The reproducibility of extraction yields decreases with SOX > ASE > MAE, as well as for concentrations <100 ng g(-1) sediment. Analyses of six different soils indicate a weak influence of soil properties (pH, texture, total organic carbon and cation exchange capacity) on the effectiveness of extraction methods. From our study we conclude that the classical SOX is still the preferred extraction approach when reliably higher FLB yields are of foremost interest or low concentrations are expected, as it is most effective and reproducible. However, considering the drawbacks of SOX (high extraction times and high solvent consumption), MAE and ASE appear to be comparably attractive for extracting FLBs in archaeological contexts. In addition to comparable FLB ratios, MAE and ASE are economically more efficient, as they reach a higher sample throughput and waste lower amounts of extractant.