Ln3+ doping in CaYAl3O7 and luminescence concentration quenching studied via a new computer modelling strategy

Ln-doped CaYAl3O7 (CYAM) has several applications due to its optical properties. This crystal matrix belongs to the melilite family, where Ca2+/Y3+ ions are randomly distributed at the same crystallographic site keeping a composition ratio of 1:1. This natural disorder represents a problem to traditional modelling strategies because it is not possible to determine experimentally which ion is actually substituted when a Ln3+ ion is incorporated. To overcome this problem and to be able to use a static computer modelling approach based on energy minimisation, a supercell was built and Ca2+/Y3+ ions were distributed randomly in the crystallographic positions keeping the composition ratio at 1:1. The substitution of any Ln3+ at cation sites could be successfully simulated. The energetic cost for the extrinsic defect creation were calculated using two different approaches. The first one was the well-established Mott-Littleton method, where the defect of interest is created in the centre of an explicit region and the position of all species were allowed to relax until a minimum potential energy is reached, with more distant regions being treated as a dielectric continuum. The second approach was based on a direct incorporation into a supercell by just redefining the supercell structure with the defect of interest as part of the supercell itself. The results from both approaches demonstrated that substitution at Ca site with charge compensation by Ca displaced to Y site, forming a kind of CaY’ anti-site defect, was the most probable configuration. Analysis of the defect region showed that presence of mostly Ca ions in the next neighbour cation site to the dopant site reduced the solution energies. This latter result can explain the high Ln doping concentration limit observed for Ln3+ - doped CYAM, as compared to Ln3+ doping in other host materials, without any appreciable photoluminescence concentration quenching.