PHREEQC simulation dataset for the SO4/Ca sensitivity study of Atacama-type lithium brines

Thermodynamic database, simulation code, model output and figures for a two-dimensional PHREEQC study of how the sulfate-to-calcium ratio of a feed brine, and its absolute sulfate loading, control lithium losses during solar evaporation of Atacama-type brines. VERSION 1.2 SUPERSEDES 1.0.1 (version 1.1 was internal and never published). The thermodynamic database was replaced, and two conclusions of v1.0.1 are reversed: lithium recovery has no optimum window but a plateau terminated by a threshold, and the lithium sink is Li2SO4.H2O rather than lithium carnallite. Users of v1.0.1 should re-derive their results from this version. The file 01_CHANGELOG.md documents every change against the primary sources. FILE ORGANISATION. Zenodo does not allow folders, so the deposit is distributed as thirteen flat files: the documentation, plus the code, data and figures as ZIP archives. The numeric prefixes give the reading order. Extracting all archives into one directory rebuilds the working layout (database/, code/, data/, figures/), and the scripts then run without changes. See 00_README.md. MODEL. The Harvie-Moller-Weare Pitzer database is extended to lithium with the complete parameter set of Song and Yao (2003), in the PHREEQC formulation of Declercq et al. (2018). Evaporation is simulated with PHREEQC 3.8.6, isothermally at 25 C, with ten phases precipitating at equilibrium. Only published brine analyses are used (Ogawa et al. 2014; Torres et al. 2024). No confidential or operator data were used. CONTENTS- The two-dimensional grid: 66 compositions, crossing ten SO4/Ca molar ratios (R = 0.2-22) with seven absolute sulfate loadings (0.010-0.267 mol/kgw), with the full trajectory of every cell.- A 108-run sensitivity study over the three published parameterisations of the Li-Mg mixing term theta(Li,Mg).- Calcium restoration of a fixed brine by two routes: calcium chloride as a reagent, and blending with a parametrised calcic source.- Comparison with the Atacama- and Uyuni-type brines of Ogawa et al. (2014).- A phase-disabling battery.- External validation against the seawater evaporation series of Shalev et al. (2018).- The figures of the associated manuscript, as vector PDF and 1000-dpi TIFF. PRINCIPAL RESULTS- Dissolved-lithium recovery as a function of R is a plateau terminated by a threshold, not a window with an optimum. Recovery is 100 percent for R <= 1 across a 27-fold range of sulfate loading.- The ratio fixes where the threshold lies. The absolute sulfate loading fixes how much lithium is lost once it is crossed.- Under the three published parameterisations of theta(Li,Mg), the minimum recovery in all 36 cells with R <= 1 is 99.98 percent. R = 1 is therefore a lower bound on the threshold under every published parameterisation, at every sulfate loading examined.- Restoring calcium beyond the stoichiometric point brings no further gain in lithium recovery under either route, while continuing to cost halite on contact or evaporative duty. KNOWN LIMITATIONS, STATED EXPLICITLY- Validation against Shalev et al. (2018) shows the sulfate subsystem degrading above I ~ 15 mol/kgw, while the chloride subsystem holds to within a few percent. The direction of the bias implies that the reported lithium losses are a floor rather than a ceiling.- All runs assume equilibrium, and are isothermal.- No calcium chloride solid precipitates in the phase set.