Intercalation of ethylene glycol in smectites: Several molecular simulation models verified by X-ray diffraction data.
Organoclays represent a special challenge for molecular simulations because they require accurate representation of the clay and the organic/aqueous sections of the model system and accurate representation of the interactions between them. Due to the broad range of force-field models available, an important question to ask is which sets of parameters will best suit the molecular modeling of the organo-intercalated smectites? To answer this question, the structure of the ethylene glycol (EG)-smectite complex is used here as a testing model because the intercalation of EG in smectites provides a stable interlayer complex with relatively constant basal spacing.Three smectite samples with substantially different layer charge and charge localization were selected for X-ray diffraction (XRD) measurements. Their molecular models were built and molecular-dynamics simulations performed using various combinations of the organic force fields (CGenFF, GAFF, CVFF, and OPLS-aa) with ClayFF and INTERFACE force fields used to describe smectites. The simulations covered a range of different EG and water contents. For every structure, the density distribution of interlayer species along the direction perpendicular to the layer plane was calculated and then used to optimize the XRD patterns for these simulated models.A comparison of these results with experimental XRD patterns shows very large discrepancies in the structures and basal spacings obtained for different layer charges as well as for different force fields and their combinations. The most significant factor affecting the accuracy of the calculated XRD patterns was the selection of the clay-mineral force-field parameters. The second important conclusion is that a slight modification of the basal oxygen parameters for non-electrostatic interactions (increase of their effective atomic diameters) may be a simple and straightforward way to improve significantly the agreement between the modeled XRD patterns with experiments, especially for high-charge smectites. Generally, among organic force fields, the least accurate results were obtained with CGenFF. For unmodified ClayFF, its combination with GAFF gave the best results, while the two other sets (OPLS-aa and CVFF) gave the best results in combination with ClayFFmod. The INTERFACE and INTERFACEmod produced much better results for low-charge montmorillonite than for high-charge smectites.