Montmorillonite dissolution kinetics: Experimental and reactive transport modeling interpretation

The dissolution kinetics of K-montmorillonite was studied at 25 degrees C, acidic pH (2-4) and 0.01M ionic strength by means of well-mixed flow-through experiments. The variations of Si, Al and Mg over time resulted in high releases of Si and Mg and Al deficit, which yielded long periods of incongruent dissolution before reaching stoichiometric steady state. This behavior was caused by simultaneous dissolution of nanoparticles and cation exchange between the interlayer K and released Ca, Mg and Al and H. Since Si was only involved in the dissolution reaction, it was used to calculate steady-state dissolution rates, RSi, over a wide solution saturation state (Delta G(r) ranged from -5 to -40 kcal mol(-1)). The effects of pH and the degree of undersaturation (Delta G(r)) on the K-montmorillonite dissolution rate were determined using R-Si. Employing dissolution rates farthest from equilibrium, the catalytic pH effect on the K-montmorillonite dissolution rate was expressed as R-diss = k.a(H)(0.56 +/- 0.05) whereas using all dissolution rates, the Delta G(r) effect was expressed as a non-linear f(Delta G(r)) function R-diss = k . [1 - exp(-3.8 x 10(-4) . (vertical bar Delta G(r)vertical bar/RT)(2.13))] The functionality of this expression is similar to the equations reported for dissolution of Na-montmorillonite at pH 3 and 50 degrees C (Metz, 2001) and Na-K-Ca-montmorillonite at pH 9 and 80 degrees C (Cama et al., 2000; Marty et al., 2011), which lends support to the use of a single f(Delta G(r)) term to calculate the rate over the pH range 0-14. Thus, we propose a rate law that also accounts for the effect of pOH and temperature by using the pOH-rate dependence and the apparent activation energy proposed by Rozalen et al. (2008) and Amram and Ganor (2005), respectively, and normalizing the dissolution rate constant with the edge surface area of the K-montmorillonite. 1D reactive transport simulations of the experimental data were performed using the Crunchflow code (Steefel et al., 2015) to quantitatively interpret the evolution of the released cations and to elucidate the stoichiometry of the reaction. After the implementation of (i) the obtained f(Delta G(r)) term in the K-montmorillonte dissolution rate law, (ii) a fraction of highly reactive particles and surfaces and (iii) the cation exchange reactions between the interlayer K+ and the released Al3+, Mg2+, Ca2+ and H+, the simulations agreed with the experimental concentrations at the outlet. This match indicates that fast dissolution of fine particles and highly reactive sites and exchange between the interlayer K and dissolved structural cations (Al and Mg) and protons are responsible for the temporary incongruency of the K-montmorillonite dissolution reaction. As long as dissolution of the bulk sample predominates, the reaction is stoichiometric. (C) 2018 Elsevier Ltd. All rights reserved.