Absolute acidity of clay edge sites from ab-initio simulations

Abstract

We provide a microscopic understanding of the solvation structure and reactivity of the edges of neutral clays. In particular we address the tendency to deprotonation of the different reactive groups on the (0 1 0) face of pyrophyllite. Such information cannot be inferred directly from titration experiments, which do not discriminate between different sites and whose interpretation resorts to macroscopic models. The determination of the corresponding pK a then usually relies on bond valence models, sometimes improved by incorporating some structural information from ab-initio simulations. Here we use density functional theory based molecular dynamics simulations, combined with thermodynamic integration, to compute the free energy of the reactions of water with the different surface groups, leading to a deprotonated site and an aqueous hydronium ion. Our approach consistently describes the clay and water sides of the interface and includes naturally electronic polarization effects. It also allows to investigate the structure and solvation of all sites separately. We find that the most acidic group is SiOH, due to its ability to establish strong hydrogen bonds with adsorbed water, as it also happens on the quartz and amorphous silica surfaces. The acidity constant of AlOH 2 is only 1 pK a unit larger. Finally, the pK a of AlOH is outside the possible range in water and this site should not deprotonate in aqueous solution. We show that the solvation of surface sites and hence their acidity is strongly affected by the proximity of other sites, in particular for AlOH and AlOH 2 which share the same Al. We discuss the implications of our findings on the applicability of bond valence models to predict the acidity of edge sites of clays.

Acidity of edge surface sites of montmorillonite and kaolinite

“…7.0 and 8.3, are in good consistence with the experimental range. These two values also 235 agree with the pKas of their counterpart edge sites (6.8 and 7.6 for ≡Si-OH and ≡Al-OH 2 OH 236 respectively) reported in a recent study (Tazi et al, 2012), where the neutral clay framework (i.e. 237 without substitution) is investigated.…”

“…is not possible and therefore previous studies made comparisons with results from bond valence theory 265 rather than experiments (Bickmore et al, 2003;Tazi et al, 2012). However, as discussed above, one 266 finds good agreement between the pKas of ≡Si-OH and ≡Al-OHOH 2 sites and experimental values.…”

Molecular Dynamic Simulations of Montmorillonite–Organic Interactions under Varying Salinity: An Insight into Enhanced Oil Recovery

“…Basal {001} surfaces of montmorillonite (the periodic models have no edge sites) are relatively pH stable and unlikely to vary within the pH range encountered within a typical oil reservoir. 24,25 All organic molecules used in this study were created using the Avogadro molecular editing suite. 26 Periodically replicated supercells contained one layer of montmorillonite composed of 84 unit cells (12 x 7 x 1), dimensions of approximately 6 x 6 x 6 nm, and a d -spacing of approximately 5 nm.…”

Na-Montmorillonite Edge Structure and Surface Complexes: An Atomistic Perspective

“…This pH condition imposes some limitations on our findings; however, the results should be considered in light of the calculated pK a s of the B edge surface groups. The calculated pK a s for the B edge are~7 for silanol groups and 8 for aluminol but are subject to relatively large uncertainties that result in the overlap of the two values [25][26][27]. On the AC edge, the pK a for the amphoteric Al site (i.e., ≡Al-OH 2 ) is 5.5 [26].…”

“…These edge structures have been the initial configuration for atomistic simulations of the 2:1 clay edge at both the quantum mechanical (QM) and molecular mechanical (MM) levels of theory. Many QM simulations have been used for a detailed characterization of the 2:1 edge structures, surface energies, acid-base reactivity, and cationic surface complexes [18][19][20][21][22][23][24][25][26][27]. These simulations are the most theoretically rigorous available, but the theoretical rigor comes with high computational costs that limit the model size and the duration of any QM-based molecular dynamics (MD) simulations.…”