Internally consistent thermodynamic data for aqueous species in the system Na-K-Al-Si-O-H-Cl

Abstract A large amount of critically evaluated experimental data on mineral solubility, covering the entire Na–K–Al–Si–O–H–Cl system over wide ranges in temperature and pressure, was used to simultaneously refine the standard state Gibbs energies of aqueous ions and complexes in the framework of the revised Helgeson–Kirkham–Flowers equation of state. The thermodynamic properties of the solubility-controlling minerals were adopted from the internally consistent dataset of Holland and Powell (2002; Thermocalc dataset ds55). The global optimization of Gibbs energies of aqueous species, performed with the GEMSFITS code ( Miron et al., 2015 ), was set up in such a way that the association equilibria for ion pairs and complexes, independently derived from conductance and potentiometric data, are always maintained. This was achieved by introducing reaction constraints into the parameter optimization that adjust Gibbs energies of complexes by their respective Gibbs energy effects of reaction, whenever the Gibbs energies of reactant species (ions) are changed. The optimized thermodynamic dataset is reported with confidence intervals for all parameters evaluated by Monte Carlo trial calculations. The new thermodynamic dataset is shown to reproduce all available fluid-mineral phase equilibria and mineral solubility data with good accuracy and precision over wide ranges in temperature (25–800 °C), pressure (1 bar to 5 kbar) and composition (salt concentrations up to 5 molal). The global data optimization process adopted in this study can be readily repeated any time when extensions to new chemical elements and species are needed, when new experimental data become available, or when a different aqueous activity model or equation of state should be used. This work serves as a proof of concept that our optimization strategy is feasible and successful in generating a thermodynamic dataset reproducing all fluid-mineral and aqueous speciation equilibria in the Na–K–Al–Si–O–H–Cl system within their experimental uncertainties. The new dataset resolves the long-standing discrepancies between thermodynamic data of minerals and those of aqueous ions and complexes, by achieving an astonishing degree of consistency between a large number of fluid-mineral equilibrium data. All of this at the expense of changing the standard state properties of aqueous species, mainly the Gibbs energy of formation. Using the same strategy, the core dataset for the system Na–K–Al–Si–O–H–Cl can be extended with additional rock-forming elements such as Ca, Mg, Fe, Mn, Ti, S, C, B. In future, the standard-state properties of minerals and aqueous species should be simultaneously optimized, to create the next-generation of fully internally consistent data for fluid-mineral equilibria. Although we employ the widely used HKF equations for this study, the same computational approach can be readily applied to any other speciation-based equation of state for multicomponent aqueous solutions.