Title:  Measurement of sulfur isotope composition in two high-elevation snowpits by Multiple-Collector Thermal Ionization Mass Spectrometry (MC-TIMS) using a 33S-36S Double Spike

The variability of sulfur isotope ratios, caused by mass dependent fractionation during biogeochemical processing, is commonly used for tracing the various sources of sulfur and for understanding the sulfur cycle. A new analytical technique for the determination of d34S and sulfur concentration that is both highly accurate and precise has been developed for the determination of sulfur isotope composition of low concentration samples.  The technique is based on the production of singularly charged arsenic sulfide molecular ions (AsS+) by thermal ionization using silica gel as an emitter and combines multiple-collector thermal ionization mass spectrometry (MC-TIMS) with a 33S-36S double spike to correct instrumental fractionation.  Because the technique is based on thermal ionization of AsS+, and As is mononuclidic, corrections for interferences or for scale contraction/expansion are not required.  The method has been applied to three international sulfur standards (IAEA-S-1, IAEA-S-2, and IAEA-S-3) and snow and firn samples collected from the Inilchek Glacier, Kyrgyzstan (42.16°N, 80.25°E, 5100 m) and Summit, Greenland (72.58°N, 38.53°W, 3238 m)


The standards were measured to evaluate the precision and accuracy of the new technique and to evaluate the consensus values for these standards.  Two different double spike preparations were used.  The d34S values (reported relative to Vienna Canyon Diablo Troilite (VCDT), (d34S (‰) = [(34S/32S)sample/(34S/32S)VCDT – 1) x 1000]), 34S/32SVCDT = 0.0441626) determined were  –0.32 ± 0.04‰ (1s, n=4) and –0.31 ± 0.13‰ (1s, n=8) for IAEA-S-1, 22.65 ± 0.04‰ (1s, n=7) and 22.60 ± 0.06‰ (1s, n=5) for IAEA-S-2, and –32.47 ± 0.07‰ (1s, n=8) for IAEA-S-3.  The amount of natural sample used for these analyses ranged from 0.40 to 2.35 mmol.  Replicate determinations of each standard showed less than 0.5‰ variability (IAEA-S-1 < 0.4‰, IAEA-S-2 < 0.2‰, and IAEA-S-3 < 0.2‰).  The uncertainties reported are comparable to or better then those obtained by gas source isotope ratio mass spectrometers (IRMS).  The d34S measurements of the snow and firn samples were used to estimate seasonal sulfate sources contributing to precipitation in these regions.  d34S data from the Summit snowpit are the first continuous high-resolution (≈ 7 samples/1 year) data for this site.  The d34S values for the Inilchek ranged from 2.6 ± 0.05‰ (1s) to 7.6 ± 0.06‰ (1s) on sample sizes ranging from 0.3 to 1.8 mmol S.  d34S values for Greenland ranged from 3.6 ± 0.19‰ (1s) to 13.3 ± 2.51‰ (1s) for sample sizes ranging from 0.05 to 0.29 mmol S.  For both the Inilchek and Summit samples the uncertainties are dominated by blank corrections and not by measurement uncertainty.  The SO42- concentrations ranged from 0.92 ± 0.009 mmol L-1 to 10.46 ± 0.08 mmol L-1 (95% C.I.) for the Inilchek and from 0.16 ± 0.09 mmol L-1 to 0.92 ± 0.06 mmol L-1 (95% C.I.)

 for the Greenland snowpit.  Although the Inilchek snowpit does not demonstrate the seasonality seen in the Greenland snowpit, the mass balance results show anthropogenic sulfate dominates (≈ 75%) throughout most of the year at both sites.  Because of the reduction in sample size requirements of this technique, by as much as a factor of 10, accessing the higher-resolution sulfur isotope record of low concentration snow and ice is now possible.