Development of computational methods for studying hydrogen interactions with light metal nanostructures

  • Akin Budi, RMIT University, Australia
  • David Henry, RMIT University, Australia
  • Prof Julian Gale, Curtin University, Australia
  • Prof Irene Yarovsky, RMIT University, Australia
  • Light-metal based hydrides have the potential for high hydrogen storage capacity (up to 10.6 wt%). Despite the high storage capacity, none of the materials exhibited fast kinetics, which are essential for practical applications. Cluster based materials are known to have different properties than bulk and may provide a good candidate for novel nanostructures with high hydrogen storage capacity and good kinetics. Computer modelling studies are useful to understand the thermodynamic and kinetic parameters of such materials and help identify promising candidates for experimental investigations.

    To model systems larger than a few hundred atoms, it is necessary to employ approximations such as empirical potentials. However, it is also important to study dissociation mechanisms of H2 on the cluster surfaces. Therefore, a method able to handle both the long range interactions and the reaction mechanisms is necessary. The Embedded Atom Method (EAM) is known to describe metallic bonding accurately, while the quantum-mechanical based ReaxFF approach is suitable to describe interactions at the reaction sites. However, both EAM and ReaxFF have deficiencies for modelling light-metal clusters interacting with hydrogen adequately.

    The overall aim of this project is to develop a hybrid potential based on the combined EAM/ReaxFF technique, which will be able to simulate the interaction between hydrogen and aluminium-based metal clusters in pure and doped forms. We use DFT calculations on small light metal nanoclusters of aluminium, magnesium, and lithium for developing new EAM/ReaxFF parameters and performing structural comparison and temperature effect studies with the best currently available EAM potential.