Abstract
We study a local-structure model suggested recently by Laguta et al. for the K+ site in KTaO3:Fe3+. It is shown to be unreasonable for explaining the unusually large value of the EPR parameter D from calculations on D. A new assumption on the local structure is proposed.