Intrinsic structure of relaxor ferroelectrics from first principles
This paper introduces the FIRE-Swap first-principles framework, which utilizes machine-learning interatomic potentials to reveal that lead magnesium niobate (PMN) possesses a unique rock-salt-like chemical order and interconnected polar nanoregions within Nb clusters, providing a mesoscale explanation for its relaxor ferroelectricity that distinguishes it from PZT and PST.