MOLECULAR DYNAMICS STUDY OF PRESSURE-INDUCED STRUCTURAL PHASE TRANSISTION OF IONIC CRYSTALS

Abstract

Molecular Dynamics allows us to solve the newtonian equations for an assembly of molecules in such a way that the knowledge of positions and velocities of the molecules in the system can help us to follow the time evolution of the system under certain specified external conditions. It actually serves as a powerful microscope to examine the detailed local configurations and atomic trajectories and helps us to resolve the atomistic movements which can not be probed experimentally nor be proved theoretically. Hence in this work we are using two slightly different constant pressure Molecular Dynamics (MD) to study the structural phase transitions, which have been elucidated experimentally, of ionic crystals, NaCl and MgF2. The first constant pressure MD is the modified Andersen (MA) algorithm, which is a modified form of the original Andersen algorithm. In the original Andersen algorithm, the volume is one of the dynamical variables, that means the value of volume will change in magnitude but not the shape of the simulation box. For the MA method, the box edges, instead of volume, are the dynamical variable. Hence the MD cell is now able to fluctuate in brick shape while maintaining the three edges at 90° to each other. The second constant pressure MD is the Parrinello-Rahman (PR) algorithm which generalizes the Andersen's algorithm to let the MD box fluctuate in both shape and size. Our studies of structural phase transitions of both NaCl and MgF2 are summarized in the following paragraphs. NaCl assumes the NaCl-type structure under atmospheric pressure and transforms to CsCl-type structure under very high pressure. For this type of transformation, there are two different models, Watanabe's model and Buerger's model, to explain. The Watanabe's model is a combination of two systematic movements of ions : intralayer movement and interlayer translational motion of planes. Whereas the Buerger's model proposes a contraction along one of the trigonal axes while allowing other ions to relax perpendicularly to this direction. After investigating the differences of these two models, we realize that Watanabe's model is just a distorted version of Buerger's model. For the MgF2, its transformation from rutile structure to CaF2 structure has been observed too at high pressure. Due to its complex and low symmetrical structure, its transformation has not been computer simulated before. We have successfully simulated the transition and proposed a mechanism for this kind of transformation. Two first nearest neighbours and the other two nearest neighbours, of a Mg2+ ion, form a parallelogram with the cation at the centre and anions at the corners. The next nearest neighbours will move towards the first nearest neighbours along the longer sides of the parallelogram to form (011) plane of CaF2 structure. Whereas the (011) (CaF2) is obtained by a rotation of about 45° of a plane formed by the four second nearest neighbours.