Abstract
Several arrangements are suggested to explain deviations from stoichiometry in LiNbO3 in a recent report by Lerner, Legras, and Dumas. These arrangements are considered to be unlikely, and an alternative mechanism in terms of partial stacking faults is proposed for LiNbO3 and LiTaO3. A microscopic statistical theory of ferroelectricity, which is outlined, can predict the variation of the Curie temperature with changes in the bonding forces and atomic masses. Using this approach, the stacking‐fault model is consistent with the observed variation of Curie temperature with composition.