EVOLUTIONARY PROGRAMMING TECHNIQUES FOR PREDICTING INORGANIC CRYSTAL-STRUCTURES

New techniques based on the implementation of genetic algorithms toget her with energy minimisation procedures are able to predict the crysta l structures of complex inorganic solids. A crucial feature of our app roach is the use in the initial stages of the simulation of a sophisti cated cost function based on Pauling's rules, recently extended and qu antified as the bond valence model. Using such functions, we are able to generate candidate structures whose energies may subsequently be mi nimised using standard lattice energy methods employing Born model pot entials. We demonstrate the efficacy of the method in yielding accurat e solutions of complex crystal structures by its application to the pr eviously unsolved structure of the ternary oxide Li3RuO4.