REEVALUATING GRAVIMETRIC DETERMINATION OF CARBONATE ASSOCIATED SULFATE CONCENTRATIONS

The changes in seawater sulfate isotopic value and concentration through Earth’s history have been determined by measuring the carbonate associate sulfate (CAS). Sulfate ions can substitute for carbonate ions during carbonate mineral precipitation; the amount (in ppm) of sulfate incorporation is thought to be proportional to the amount of sulfate in the water, the rate at which the carbonate precipitates, and the original mineralogy. CAS is commonly determined by first eliminating any organic sulfur species in the powdered sample, dissolving the carbonate powder in acid, and then adding barium chloride to foster the precipitation of the highly insoluble mineral barite (BaSO₄). The barite can then be weighed to produce a gravimetric quantification of the amount of sulfate in the sample and/or analyzed for δ³⁴S. However, while the CAS-extraction procedure may not impart an additional isotopic fractionation to the sulfate sulfur, the gravimetric determination of sulfate concentration may be problematic. Using this method, we extracted the CAS 15g carbonate samples spiked with 2000ppm of sulfate. With samples of this size, the percent of sulfate recovered varied substantially (60%-120%) and unsystematically. Likely sources of error included mass loss/gain during handling, hydration of the filter and/or precipitant, or incomplete reaction. We also spiked some samples with 0.02 g crushed pyrite to determine if sulfur from pyrite may be liberated in the form of sulfate during the acidification and thus add non-structural sulfate into the end-result. As with CAS alone, the error in samples measured gravimetrically varied unsystematically, but did not yield average CAS any different than the average unspiked average CAS. Therefore, if any sulfate was liberated by the acidification of the pyrite, it was in a concentration too low to detect by the gravimetric method. While analyzing CAS is a potentially exciting and fruitful source of new data, the methods may still need further refinement, especially if precision better than 50% is needed to interpret the geologic record.