Levoglucosan in speleothems: An evaluation of various sample preparation methods.

<p>Secondary mineral deposits in caves, such as stalagmites or flowstones, are valuable paleoclimate archives because they have several advantages over other environmental archives. These include stable in-cave conditions, protecting the speleothems from external influences, and the potential to precisely date samples up to 600,000 years using <sup>230</sup>Th/U-dating. [1] Supplementing established climate proxies, such as stable isotopes and trace elements, organic proxies have been increasingly used in recent years to inform on local vegetation and soil dynamics. [2]</p><p>Biomass burning events are major sources of atmospheric particulate matter that influences global and local climate. [3] Investigating fire proxies in paleoclimate archives may therefore help determine the interactions of climate, hydrology, and fire activity. Levoglucosan, an anhydrosugar, naturally only originates from the combustion of cellulose and thus constitutes a biomass burning marker. Analysis of levoglucosan in sediments has shown high correlation with traditional burning markers, such as black charcoal. [4] Mannosan and galactosan, both stereoisomers of levoglucosan, are formed during combustion of hemicellulose. Previous work suggests that rather than absolute levoglucosan concentrations the ratio of levoglucosan to its isomers should be considered when characterizing burning events. [5] To date, no data on levoglucosan or its isomers in speleothems has been published, whereas the anhydrosugars are already utilised in other paleoclimate archives, such as sediments and ice cores. [2,3]</p><p>We test three approaches (solid phase extraction (SPE), soxhlet extraction and solid/liquid extraction) for the isolation and quantification of anhydrosugars using HILIC-MS instrumentation. As the anhydrosugars are highly polar molecules, extraction from the calcium carbonate matrix and subsequent sample preparation proved challenging. We evaluate the different approaches and compare the resulting concentrations and assumed recoveries. We find that the anhydrosugars do not show significant retention on any of the evaluated SPE materials. While solid/liquid extractions lead to detectable analyte concentrations, soxhlet extractions with methanol or dichloromethane/methanol mixtures are more efficient.</p><p> </p><p>[1] D. Scholz, D. Hoffmann, Quat. Sci. J. 57 (2008) 52–76 [2] A. Blyth et al. Quat. Sci. Rev. 149 (2016) 1-17 [3] P. Yao et al. J. of Glaciology 59 (2013) 599-611 [4] V. O. Elias et al. Geochim. et Cosmochim. Acta 65 (2001) 267-272. [5] D. Fabbri et al. Atmos. Env. 43 (2009) 2286–2295</p>