The more, the better: Al1− <i>x</i> (Mg,Hf)xN wurtzite ferroelectrics through co-alloying

Abstract

Wurtzite ferroelectrics (FEs) are promising candidates for next-generation memory and computing devices due to their compatibility with semiconductor processing. However, their adoption is limited by large coercive electric fields (Ec), often approaching the dielectric breakdown field (Eb), raising concerns about energy efficiency and device reliability. The discovery of wurtzite FEs has also been constrained by a narrow chemical design space, with few known examples to date. Ferroelectricity in AlN-based alloys, particularly Al1−xMxN with trivalent M3+ cations, has been a focal point of recent work. Building on studies of co-alloyed AlN for enhanced piezoelectricity, we computationally investigate ferroelectricity in Al1−x(M1,M2)xN alloys, where M12+ and M24+ are non-trivalent cations. Using density functional theory, solid-state nudged elastic band method, and structural analysis, we predict switchable polarization in Al1−x(Mg,Hf)xN. Compared to the prototypical Al1−xScxN, this co-alloy exhibits a more rapid decrease in both the switching barrier and bandgap with increasing x, suggesting a simultaneous reduction in Ec and Eb. This reduction in Ec is attributed to enhanced structural distortions introduced by co-alloying. By using bandgap and distortion as design metrics, we identify several other promising M1–M2 combinations, and highlight Al1−x(Ca,Si)xN as a strong candidate for experimental validation. Our work introduces a co-alloying strategy to access new wurtzite FEs and expands the design space to include earth-abundant elements.