An upgraded CFA - FLC - MS/MS system for the semi-continuous detection of levoglucosan in ice cores.

A new Continuous Flow Analysis (CFA) system coupled with Fast Liquid Chromatography - tandem Mass Spectrometry (FLC-MS/MS) has been recently developed for determining organic markers in ice cores. In this work we present an upgrade of this innovative technique, optimized for the detection of levoglucosan in ice cores, a crucial tracer for reconstructing past fires. The upgrade involved a specific optimization of the chromatographic and mass spectrometric parameters, allowing for a higher sampling resolution (down to 1 cm) and the simultaneous collection of discrete samples, for off-line analysis of water stable isotopes and additional chemical markers. The robustness and repeatability of the method has been tested by the analysis of multiple sticks of ice cut from the same shallow alpine ice core, and running the system for several hours on different days. The results show similar and comparable trends between the ice sticks. With this upgraded system, a higher sensitivity and a lower limit of detection (LOD) was achieved compared to discrete analysis of alpine samples for levoglucosan measurements. The new LOD was as low as 66 ng L-1, a net improvement over the previous LOD of 600 ng L-1.

First discrete iron(II) records from Dome C (Antarctica) and the Holtedahlfonna glacier (Svalbard).

Fe(II) is more soluble and bioavailable than Fe(III) species, therefore the investigation of their relative abundance and redox processes is relevant to better assess the supply of bioavailable iron to the ocean and its impact on marine productivity. In this context, we present a discrete chemiluminescence-based method for the determination of Fe(II) in firn matrices. The method was applied on discrete samples from a snow pit collected at Dome C (DC, Antarctica) and on a shallow firn core from the Holtedahlfonna glacier (HDF, Svalbard), providing the first Fe(II) record from both Antarctica and Svalbard. The method showed low detection limits (0.006 ng g-1 for DC and 0.003 ng g-1 for the HDF) and a precision ranging from 3% to 20% RSD. Fe(II) concentrations ranged between the LoD and 0.077 ng g-1 and between the LoD and 0.300 ng g-1 for the Antarctic and Arctic samples, respectively. The Fe(II) contribution with respect to the total dissolved Fe was comparable in both sites accounting, on average, for 5% and 3%, respectively. We found that Fe(II) correctly identified the Pinatubo/Cerro Hudson eruption in the DC record, demonstrating its reliability as volcanic tracer, while, on the HDF core, we provided the first preliminary insight on the processes that might influence Fe speciation in firn matrices (i.e. organic ligands and pH influences).