Reactive transport modeling of plug-flow reactor experiments: quartz and tuff dissolution at 240°C

Abstract

Extension of reactive transport modeling to predict the coupled thermal, hydrological, and chemical evolution of complex geological systems is predicated on successful application of the approach to simulate well-constrained physical experiments. In this study, steady-state effluent concentrations and dissolution/precipitation features associated with crushed quartz and tuff dissolution at 240°C have been determined experimentally using a plug-flow reactor (PFR) and scanning electron microscopy (SEM) techniques, then modeled with the reactive transport simulator GIMRT ( Steefel and Yabusaki, 1996) using a linear rate law from transition state theory (TST) . For quartz dissolution, interdependence of the specific surface area ( A m ) and reaction rate constant ( k m ) predicted from the modeling agrees closely with that obtained from an analytical solution to the reaction–transport equation without diffusion/dispersion, verifying the advection-dominant nature of the PFR experiments. Independently-determined A qtz and k qtz from the literature are shown to be internally consistent with respect to the model and analytical interdependence, implying appropriateness of the linear TST rate law and adequacy of BET-determined A m for use in modeling PFR experiments. Applications of this integrated approach for monomineralic dissolution include assessment of internal consistency among independent A m and k m data, estimation of k m from BET-determined A m , and rapid evaluation of alternative rate laws. For tuff dissolution, accurate simulation of the experimental steady-state effluent concentrations (to within 3% for Na, Al and K; to within 15% for Si and Ca) and dearth of alteration phases (<1 vol.% in the model) defines minimum supersaturation thresholds for nucleation of primary minerals (to suppress their secondary precipitation) and the (non-unique) relative k m of hydrous aluminosilicates. For such multicomponent, multiphase experiments, reactive transport simulators provide the only practical means of estimating and constraining these kinetic parameters. Despite the approximations and uncertainties inherent in present-day models, close agreement between the experimental and simulation results reported herein provides an important measure of confidence for extending the modeling approach to address increasingly complex systems for which development of experimental analogs is impractical or impossible.