Structure prediction of aluminum nitride combining data mining and quantum mechanics

Abstract

Aluminum nitride (AlN) is a compound with wide technological applications from optics to electronics. At ambient pressure and temperature, AlN has a hexagonal wurtzite type of structure, while the zinc blende (ZnS) type of structure is found in very thin films. At high pressures, a first-order phase transformation from the wurtzite structure to a rock salt type structure has been observed. This study covers the experimentally observed modifications of AlN, investigates their relations and searches for new possible modifications. Therefore, data mining of over 140=000 structure candidates has been performed, followed by local optimizations at the ab initio level with Hartree-Fock, LDA, and B3LYP functionals. Finally, twelve structure candidates have proven to be the most promising ones. These include the novel metastable AlN polytypes and the 5-5-type as a possible high pressure candidate of AlN.