Linking calcite surface roughness resulting from dissolution to the saturation state of the bulk solution

It has long been supposed that mineral dissolution results in surface microtopography features that may be related to the weathering conditions. Detection of etch pits on weathered mineral surfaces is indicative of far-from-equilibrium conditions (O?0), whereas their absence points towards close-to-equilibrium conditions (O?1). However, surface microtopography characterizations limited solely to qualitative comparisons may conceal the potential of reconstructing intermediate saturation conditions. To investigate this prospect, we performed flow-through dissolution experiments at room temperature and atmospheric pCO(2), on mechanically-polished {104} calcite surfaces reacting at different saturation states with respect to calcite (0 = O =0.8), under alkaline conditions (pH = 7.9). Time-resolved topography data of the dissolved calcite surface were acquired ex situ using vertical scanning interferometry (VSI). Quantitative comparisons of the ensuing surface topography data relied on surface roughness characterizations, based on a combination of unidimensional descriptors (R-a) and spatial statistics metrics (power spectral density, or PSD, and semi-variogram). Time-resolved surface roughness analyses suggested a temporal stabilization of the calcite surfaces undergoing dissolution for all targeted saturation states, and that the steady-state surface roughness evaluations corresponding to different O values are statistically distinguishable, with larger roughness values corresponding to lower O values, according to a seemingly bijective relationship. We then investigated the atomic-scale mechanisms that might partially explain the empirical O-roughness relationship derived experimentally at the VSI-scale through kMC and Ising dissolution modeling of a Kossel crystal. We found that the Ising model successfully reproduced the O-roughness behavior observed experimentally, and we discussed the features that interpretative stochastic models need to satisfy to agree with the experimental findings. Overall, the present study suggests that, under the investigated conditions, the steady-state calcite surface roughness resulting from dissolution can be used as a proxy to back-estimate the saturation state of the fluid under which the reaction occurred for samples reacted under unknown conditions.