Force field development from ab-initio data
Sami Tazi (University Pierre and Marie Curie, UnivParis 06, UMR 7195, PECSA, Paris, France)

In environmental science, the sorption of radioactive contaminants (e.g. cesium or strontium) onto clays will have important consequences in the evaluation of these materials for nuclear waste storage. Therefore, in order to understand the transport and retention of radionucleides, several features must be correctly captured: the interaction of water with ions (both radionucleides and naturally occurring species) and the physical chemistry of the solid/liquid interface involved here.

To get the better accuracy for those kind of system, the study would had been done using first principle molecular dynamic (FPMD), since it required a good precision of the chemical specificity and many-body effects to correctly describe the solid/liquid interface and to obtain a transferable potential.
However to date, the use of FPMD simulations is prohibited because of the large time scales required to study such complex systems.

A good way to find a balance between the large time scales and the precision required for this system is to develop a polarisable force-field using force/dipole-matching from FPMD data.

It will be presented the force/dipole matching method we used to develop a polarizable force field for both aqueous ions and clay. The second part will be devoted to the validation of the model, taking into account the structural, dynamical and thermodynamic properties of ions in water.