Atmospheric Measurement Techniques (Feb 2019)

Large-volume air sample system for measuring <sup>34</sup>S∕<sup>32</sup>S isotope ratio of carbonyl sulfide

  • K. Kamezaki,
  • S. Hattori,
  • E. Bahlmann,
  • N. Yoshida,
  • N. Yoshida

DOI
https://doi.org/10.5194/amt-12-1141-2019
Journal volume & issue
Vol. 12
pp. 1141 – 1154

Abstract

Read online

Knowledge related to sulfur isotope ratios of carbonyl sulfide (OCS or COS), the most abundant atmospheric sulfur species, remains scarce. An earlier method developed for sulfur isotopic analysis for OCS using S+ fragmentation by an isotope ratio mass spectrometer is inapplicable for ambient air samples because of the large samples required (approx. 500 L of 500 pmol mol−1 OCS). To overcome this difficulty, herein we present a new sampling system for collecting approximately 10 nmol of OCS from ambient air coupled with a purification system. Salient system features are (i) accommodation of samples up to 500 L (approx. 10 nmol) of air at 5 L min−1; (ii) portability of adsorption tubes (1∕4 in. (0.64 cm) outer diameter, 17.5 cm length, approximately 1.4 cm3 volume) for preserving the OCS amount and δ34S(OCS) values at −80 ∘C for up to 90 days and 14 days; and (iii) purification OCS from other compounds such as CO2. We tested the OCS collection efficiency of the systems and sulfur isotopic fractionation during sampling. Results show precision (1σ) of δ34S(OCS) values as 0.4 ‰ for overall procedures during measurements for atmospheric samples. Additionally, this report presents diurnal variation of δ34S(OCS) values collected from ambient air at the Suzukakedai campus of the Tokyo Institute of Technology located in Yokohama, Japan. The observed OCS concentrations and δ34S(OCS) values were, respectively, 447–520 pmol mol−1 and from 10.4 ‰ to 10.7 ‰ with a lack of diurnal variation. The observed δ34S(OCS) values in ambient air differed greatly from previously reported values of δ34S(OCS) = (4.9±0.3) ‰ for compressed air collected at Kawasaki, Japan, presumably because of degradation of OCS in cylinders and collection processes for that sample. The difference of atmospheric δ34S(OCS) values between 10.5 ‰ in Japan (this study) and ∼13 ‰ recently reported in Israel or the Canary Islands indicates that spatial and temporal variation of δ34S(OCS) values is expected due to a link between anthropogenic activities and OCS cycles. The system presented herein is useful for application of δ34S(OCS) for investigation of OCS sources and sinks in the troposphere to elucidate its cycle.