The feasibility of using a pyrite standard to calibrate the sulfur isotope ratio of marcasite during SIMS analysis

Marcasite (FeS2, orthorhombic) is a dimorph of pyrite (FeS2, isometric), which has been reported in many kinds of ore deposits and sedimentary rocks. The sulfur isotopic composition of marcasite could be used to track ore-forming fluid source(s) and ocean acidification events in deep time. However, matrix-matched standards for marcasite are scarce, and in practice a pyrite standard is commonly used to calibrate the sulfur isotope ratios of marcasite. But the feasibility of this calibration regime needs to be carefully evaluated. In this contribution, we investigated this topic by analyzing a natural marcasite (NJUMc-1) with a newly installed Cameca IMS 1300-HR3. The sulfur isotopic composition of NJUMc-1 varies significantly (from similar to-35 parts per thousand to similar to-17 parts per thousand). A total of 135 spots were applied on NJUMc-1 marcasite (4 outliers were rejected), and when the analyzing spot number reaches 70, the average delta S-34 of marcasite becomes invariant, indicating that the number of grains we analyzed is large enough to represent the whole sample. The results of NJUMc-1 are calibrated using the pyrite standard UWPy-1. In this calibration regime, the average delta S-34 value of NJUMc-1 is -25.21 parts per thousand +/- 0.78 parts per thousand (2 SE), which agrees well with the result determined using gas source isotope ratio mass spectrometry (-25.63 parts per thousand +/- 0.24 parts per thousand, 2 SE, n = 2). This consistency indicates that it is feasible to calibrate the sulfur isotopic ratios of marcasite with a pyrite standard and a new marcasite standard is not a prerequisite for sulfur isotope determination of marcasite using SIMS. Besides, this study shows that instrumental mass fractionation (IMF) could be precisely determined for samples with obvious heterogeneity (e.g., HTS4-3 and NJUMc-1), and by inference, the isotopic composition of unknowns could be precisely determined with an inferior "secondary" standard with obvious heterogeneity, as long as enough grains were analyzed.