Quantifying Ca exchange in gypsum using a ⁴⁵Ca tracer: Implications for interpreting Ca isotopic effects in experimental and natural systems

Mineral-solution exchange rates of Ca were determined using a ⁴⁵Ca tracer technique for three types of gypsum: natural cm-long gypsum needles, natural microcrystalline gypsum, and synthetic microcrystalline gypsum. Batch reactors containing the gypsum crystals were equilibrated with a saturated ⁴⁵Ca spiked solution (Ω_gypsum ≈ 1) for 30 days and were sequentially sacrificed to estimate exchange rates. The percentage of gypsum that exchanged was determined using a time-dependent box model, which simulated mineral dissolution and reprecipitation over time with variable extents of back reaction. Large natural needles exhibited no detectable exchange over 30 days, whereas a significant percentage of the Ca in the synthetic and natural micro-crystals exchanged. According to our model-based calculations, 45% of the synthetic gypsum and 20% of the natural gypsum exchanged; because of our experimental design, these extents are not affected by the model framework. The surface area normalized exchange rates were faster over the first five days (1.04 · 10⁻⁵ ± 5.84 · 10⁻⁶ mol/m²/d in the synthetic micro-crystals and 8.97 · 10⁻⁵ ± 7.45 · 10⁻⁵ mol/m²/d in the natural micro-crystals) than during the remainder of the experiment (2.37 · 10⁻⁶ ± 6.03 · 10⁻⁷ mol/m²/d in the synthetic micro-crystals and 2.69 · 10⁻⁵ ± 1.23 · 10⁻⁵ mol/m²/d in the natural micro-crystals). As the Ca concentration in solution remained constant over the experiment, the observed exchange is thought to be a result of dissolution/reprecipitation at the mineral-solution interface rather than due to net dissolution or net precipitation. The gypsum exchange rates that we constrain are consistent with those of other sulfate minerals, irrespective of their hydration state, but are considerably faster than carbonates over similar time scales. It is not clear if this is a result of the interpretive model framework, which can strongly influence both the derived rate and the extent of reaction, or whether it is a real phenomenon. Because the influence of the model framework is so important, we detail the importance of the solid to fluid mass ratio and the role of a back-reaction flux to the accurate quantification of recrystallization rates and extents. Finally, we apply our numerical model to the reinterpretation of past work and constrain the Ca isotopic fractionation factor associated with extremely slow rates of gypsum precipitation/exchange to be between 0 and −0.1‰.