Testing carbonate-associated sulfate (CAS) extraction methods for sulfur isotope stratigraphy: A case study of a Lower–Middle Ordovician carbonate succession, Shingle Pass, Nevada, USA

Abstract

Sulfur isotopes measured from sedimentary rocks are used to reconstruct the global sulfur cycle and chemostratigraphic correlation. Relative and absolute changes in the sulfur isotopic composition of the ocean (δ34Sseawater) are essential for estimating changes in redox conditions in ancient environments, particularly for geochemical models that utilize isotope mass balance. Thus, accurate estimates of δ34Sseawater must be measured using a laboratory protocol optimized for the extraction of primary sulfate (preserved as carbonate-associated sulfate; CAS) by removing any contaminant secondary sulfate (e.g., from diagenetic pyrite oxidation).Here we use two CAS-extraction protocols on Lower–Middle Ordovician carbonate rocks to test the degree to which similar absolute values and stratigraphic trends in δ34SCAS are produced by different CAS extraction methods. One method processes carbonate powders once with a single rinse of a 10% NaCl solution to remove sulfate minerals or weakly bonded sulfate ions, followed by a rinse in 5–6% bleach to remove organically bound sulfur before dissolution in hydrochloric acid (HCl). The second method treats carbonate powders with three rinses of NaCl, without a bleach rinse, to more aggressively remove secondary sulfate before dissolution in HCl. Isotopic results show that samples treated in three NaCl rinses produce δ34SCAS values that are on average 7‰ more positive than samples treated with a single NaCl rinse, but a similar stratigraphic trend is preserved in samples processed with either method. We interpret the difference in δ34SCAS between methods to reflect the incomplete removal of secondary sulfate derived from the oxidation of 32S-enriched pyrite when employing only a single NaCl rinse. Surprisingly, rocks with low CAS concentrations (≤10 mg/kg) using both methods show no significant difference in δ34SCAS values, as well as no significant difference in pyrite concentrations and pyrite δ34S (δ34SPY) between methods. Samples with low CAS concentrations are likely altered and secondary sulfate sourced from pyrite oxidation has likely become incorporated into recrystallized carbonate minerals or incorporated into pore-filling calcite cement, all of which is captured as CAS. Results suggest that both CAS-extraction protocols can produce broadly similar δ34SCAS trends, but a single NaCl rinse may not completely remove contaminant sulfate sorbed onto carbonate grains and thus will not yield meaningful data for reconstructing δ34Sseawater. This has clear implications for modeling studies aimed at reconstructing paleo-redox conditions using δ34S data assumed to record global δ34Sseawater trends if these δ34S data are generated using a non-optimized CAS protocol.