 |
Environment from the Molecular Level A NERC eScience testbed project |
 |
Atomistic simulation of aqueous solutions in the vicinity of Iron hydroxide surfaces
Scientific problem
Heavy metal pollution of soils/sediments has increased significantly. The risks (toxicity, bioaccumulation) caused by this contamination need to be assessed and predicted. One of the key steps that controls the mobility and availability of heavy metals concerns their adsorption from solution onto mineral surfaces. Iron hydroxides are excellent heavy metal adsorbers and prime candidates for reducing the mobility of these dangerous species. Of course, the environmentally relevant processes concern the adsorption from solution. The exact structure of aqueous solutions in contact with surfaces is not yet completely elucidated. The distribution and local concentration of the various species is difficult to observe experimentally. Coarse grained Surface Complexation Models do not include explicitly surfaces effects, but conversely, ultra precise ab initio calculations are unable to cope with the required amount of water. The middle way, atomistic modelling, was also traditionally hampered by the fact that realistic ionic concentrations require the treatment of many water molecules (at a factor of 50 water per ion at the already high concentration of 1 mol.l-1). Developments in computer power have made possible the simulation of both surface and solution at the atomic resolution and in large enough quantity to produce statistically meaningful results.
Scientific results
We carried out many Molecular Dynamics simulations (classical potential, DL_POLY code) of aqueous solution(Na+/Cl-)/goethite interfaces Fig. 1), at different ionic strength and surface charges. Our main observation is that the distribution of ions near the surface is not accurately described by the classical models of the electrical double layer.
Figure 1. Representation of the double layer near the surface of goethite.
It is now admitted that any surface has a structuring effect on liquids it contact, as can be observed in the density oscillations in Fig. 2. This density rippling in turn controls the salt concentrations. There is a direct correlation with the water density up to 10 A from the surface, but relatively unexpected, broad, long ranged oscillations continue further away. The explanation lies in the electrostatic potential, as pictured in Fig 3. The structuring effect of the surface on the electrostatic potential of even pure water has a longer range that what could be inferred from the simpler density curves. The addition of ions in solutions only serves to reinforce this effect.
Figure 2. Salt concentrations and water density dependence on the distance from the surface.
Figure 3. Electrostatic potential (in pure and salted solution) and charge dependence on the distance from the surface
Real surfaces are likely to be charged. But the corresponding simulations show that the charged surface effect on the electrostatic potential in the solvent does not significantly differ from the neutral surface’s.
Additional calculations with different minerals’ surfaces (calcite CaCO3 and hematite Fe2O3) confirm these findings and suggest that the long range oscillatory behaviour of salt concentration is a consequence of the structuring presence of a surface on water electrostatic potential
We conclude that although the traditional double(Stern)-layer models are correct in assuming that the ion distribution is controlled by the electrostatic potential, they fail to reproduce the distribution at medium/high salt concentration because the electrostatic contribution of the structured (layered) solvent is not taken into account.
We show that the explicit treatment of solvent molecules is crucial to capture all the effects of the mineral surface on the liquid phase and demonstrate that high-troughput atomistic simulations can be used to reconsider and extend phenomenological models in order to depict a more comprehensive picture of the solid-liquid interface.
Credits
This work was carried out by Sebastien Kerisit, David Cooke, Arnaud Marmier and Steve Parker (Bath).
 |
 |
 |
 |
 |
 |
 |