Elsevier

Applied Geochemistry

Volume 21, Issue 6, June 2006, Pages 1006-1015
Applied Geochemistry

The applicability of accelerated solvent extraction (ASE) to extract lipid biomarkers from soils

https://doi.org/10.1016/j.apgeochem.2006.02.021Get rights and content

Abstract

We investigated the ability of accelerated solvent extraction (ASE) to extract selected lipid biomarkers (C19–C34 n-alkanes, n-alcohols and n-fatty acids as well as dehydroabietic acid and β-sitosterol) from a sandy soil profile under Corsican pine. Two organic layers (moss and F1) as well as two mineral soil horizons (EA and C1) were sampled and extracted with DCM/MeOH (93:7 v/v) by ASE at 75 °C and a pressure of 6.9 × 106 Pa or 17 × 106 Pa. Soxhlet extractions were used as the established reference method. After clean-up and derivatization with BSTFA, the extracts were analyzed on GC/MS.

Using Soxhlet as a reference, we found ASE to extract all compounds adequately. The n-alkanes, especially, were found to be extracted very efficiently from all horizons studied. Only the n-fatty acids and β-sitosterol from the organic layers seemed to be extracted at a slightly lower efficiency by ASE. In all but two instances the relative abundance of extracted lipids within a component class was the same regardless of the extraction method used.

Using a higher pressure in the ASE extractions significantly increased the extraction efficiency for all component classes in the moss layer, except β-sitosterol. The effect was most pronounced for the n-alkanes. In the EA horizon, a higher pressure slightly reduced the extraction efficiency for dehydroabietic acid. The observed differences between ASE and Soxhlet extractions as well as the pressure effect can be explained by a decrease in polarity of the extractant due to the elevated pressure and temperature applied during ASE extractions as compared to Soxhlet extractions. This would mainly increase the extraction efficiency of the least polar biomarkers: the n-alkanes as was observed. In addition, a better penetration of still partially water-filled micro pores under elevated pressure and temperature may have played a role.

Introduction

Organic matter in soils consists of a wide range of chemical components that originate predominantly from plant litter and the microbial biomass (Kögel-Knabner, 2002). The last decade has seen an increasing scientific interest in the organic chemical composition of soil organic matter (SOM) and changes therein upon soil biogeochemical processes (e.g., Gregorich et al., 1996, Guggenberger and Zech, 1999, Kögel-Knabner, 2002, Nierop et al., 2001, Wiesenberg et al., 2004b). Extractable lipids constitute a class of organic components in soils that has received particular attention. Reasons for this are amongst others the role of lipids in SOM accumulation and SOM stability (e.g., Naafs et al., 2004b, Poulenard et al., 2004, Rumpel et al., 2004), their significance for terrestrial food-web studies (e.g., Balser et al., 2005, Ruess et al., 2005), their influence on the fate of contaminants in soils (e.g., Chilom et al., 2005) and their use as vegetation tracers (e.g., Ficken et al., 1998, Van Bergen et al., 1997).

The use of lipids as vegetation tracers is based on the principle that plant-specific combinations of lipids are preserved in the soil and can serve as biomarkers to identify past vegetation compositions. Following our previous investigation of organic matter in a Dutch sandy soil under Corsican pine (Nierop and Verstraten, 2004), we considered applying this so-called biomarker technique to determine its past vegetation composition. However, we were faced with the challenge of isolating the selected lipid biomarkers from the soil matrix prior to analysis, as there is no unequivocal technique for this purpose. Instead, a wide range of extraction procedures are applied in contemporary practice, including Soxhlet extraction (e.g., Naafs et al., 2004a, Winkler et al., 2005), sonication (e.g., Dalton et al., 2005, Otto et al., 2005) and even simple shaking with solvent (Quenea et al., 2004). Soxhlet extraction has been used for the purpose of extracting lipids from soils for over 25 years (e.g., Jambu et al., 1978) and forms the basis of EPA method 3540C for the extraction of non-volatile organics from solids such as soils (EPA, 1996). As such, it is the most well-established of the methods mentioned.

While being a robust and well-established technique, Soxhlet extraction suffers from three main shortcomings: (i) the necessity of using relatively large extractant volumes of usually 250 mL or more; (ii) long analysis times of typically 16 h per analysis; and (iii) a difficulty to automate. A promising alternative is the relatively new technique of accelerated solvent extraction (ASE) (Richter et al., 1996). In short, ASE extracts samples under elevated temperature, while elevated pressure ensures that volatile extractants remain liquid. ASE can be completely automated, it employs very small extractant volumes (normally 5–30 mL) and has typical extraction times of less than an hour (Richter et al., 1996). As such the technique has the potential to overcome the main shortcoming of Soxhlet extractions. However, while the use of ASE to extract organic contaminants from soils is now reasonably well-established (Giergielewicz-Mozajska et al., 2001), its application to the extraction of soil lipids has received very little attention so far. To our knowledge only two studies have been published to date in which ASE was used to extract lipids from soil samples (Rumpel et al., 2004, Wiesenberg et al., 2004a), and no comparison with other techniques for this purpose has yet been made. Still, such a comparison of ASE with more established techniques is crucial if ASE is to be used in biomarker studies. The reason is that differences in extraction efficiencies for various types of lipids between ASE and other techniques would lead to a difference in the composition of the biomarker signal that is obtained.

Therefore, the purpose of the current study was to examine the efficiency of ASE to extract typical lipid biomarkers from a selection of soil horizons from a Dutch sandy soil under Corsican pine, using Soxhlet extractions as a reference. The biomarkers consisted of a selection from the following component classes: (i) straight-chain lipids; (ii) plant sterols; and (iii) terpenoids.

Section snippets

Description of site, soil profile and sampling procedures

The selected sampling site is a plot in ‘De Schoorlse Duinen’, a sand-dune area in The Netherlands near the village of Schoorl. The current vegetation on the study plot consists almost exclusively of Corsican Pine (Pinus nigra var. maritima) that was planted in 1929 (Nierop and Verstraten, 2004). The only undergrowth present in significant quantities is a moss layer situated between the litter (L) horizon and the F1 horizon. The soil was classified as a Haplic Arenosol (FAO-UNESCO, 1990) and a

Results and discussion

The concentrations of the selected biomarkers in the ASE and Soxhlet extracts represented as μg/g extracted soil material are presented in Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5. The values for the ASE extracts represent the average contents of the various biomarkers in the triplicate extracts with error bars indicating the standard error of the mean. Due to insufficient flushing during the filtering of one of the three low-pressure ASE extracts from the F1 horizon, probably a large fraction of

Conclusions

Altogether, we conclude that when investigating the combination of lipid biomarkers chosen in this study through extraction with the common extractant DCM/MeOH, overall ASE is a viable method to extract lipids from soils. The reasons are: (i) on average better extraction efficiencies, especially for the n-alkanes compared to the reference method of Soxhlet extractions; (ii) the reduced volumes of extractant (10–33 mL vs. 300 mL) as well as shorter extraction time (25 min vs. 16–24 h) compared to

Acknowledgements

We wish to thank Frans van der Wielen for his extensive experimental support. The Netherlands Foundation for the Advancement of Tropical Research (WOTRO) is gratefully acknowledged for their funding of this project under number WAN 75-406. We thank Fjällräven for their generous sponsoring in the form of clothing and gear.

References (33)

Cited by (0)

View full text