📝 SummarXiv

Abstract

Aluminum nitride is a traditional wide-bandgap semiconductor that has been widely used in high-power electronic and optoelectronic devices. Recently, scandium-doped aluminum nitride (AlScN) was shown to host ferroelectricity with high remnant polarization and excellent thermal stability. However, its practical use is currently limited by its high coercive field, $E_c$. Understanding the atomic-scale switching mechanism is essential to guide strategies for reducing $E_c$. Here, we combine density functional theory and machine-learning molecular dynamics to investigate polarization switching mechanisms in AlScN over various Sc concentrations and applied electric fields. We find that collective switching induces excessive lattice strain and is therefore unlikely to occur. Rather, pre-existing domain walls relieve strain and lead to a distinct switching dynamics, with the associated switching mechanism being field dependent. More precisely, at low electric fields, switching proceeds via gradual domain-wall propagation, well described by the Kolmogorov-Avram-Ishibashi model; meanwhile high fields trigger additional nucleation events, producing rapid and more homogeneous reversal, whose mixed switching process is better described by the simultaneous non-linear nucleation and growth model. These findings highlight the critical role of domain-wall dynamics in nitride ferroelectrics and suggest that domain engineering provides a viable route to control coercive fields and enhance device performance.