While there is a rich history of research revealing how domain wall migration occurs in perovskite ferroelectrics and how domain wall dynamics affects the macroscopic properties, the same cannot yet be said for newer ferroelectric classes including fluorites (e.g. HfO2) and wurtzites (e.g. AlN). These new ferroelectrics still possess classic hysteresis behavior with electric field, yet the dynamical behavior and therefore the atomistic mechanisms governing this behavior are vastly different from classic perovskites and thus require new fundamental models for each new material class. Fluorites, for example, have four order parameters governing the transition from paraelectric to ferroelectric. We will describe the multiple domain walls that are possible by sign changes in these order parameters, and how this leads to various types of domain walls that have different formation energies and velocities, greatly affecting experimental observables. Wurtzites, on the other hand, possess weak temperature dependence of polarization and have hexagonal symmetry that allows for only 180-degree domain walls and highly stable domain wall configurations. We will present recently developed Landau-Ginzburg models for fluorite ferroelectrics to uncover the energetics behind the complex landscape of possible domain walls and discuss the possible impacts of domain wall engineering in hafnia. We will also show how we can use machine learned force fields trained with artificial neural networks in aluminum nitride to understand the macroscopic bulk characteristics of these stable and short-ranged domain walls and their impact on the switching mechanism. We envision that these insights will lead to future engineering of improved devices built on these novel systems through microscopic understanding of domains and reversal mechanisms.

The ever-growing family of wurtzite ferroelectrics exemplified by (Al,Sc)N, (Al,B)N, and (Zn,Mg)O has attracted a great deal of interest for both their device integration potential and their expansion of our understanding of the fundamentals of polarization reversal. The structurally-simplest versions of domain walls in wurtzite-based materials are essentially identical to the growth twins often referred to as inversion domain boundaries in GaN, AlN, and other wurtzite compounds. Mobility of such domain walls would essentially involve lateral motion of stacking faults for domain wall motion normal to domain polarity or a direct transition through an h-BN-type nonpolar intermediate structure that results in extremely large charge densities at head-to-head or tail-to-tail boundaries. Computational and experimental evidence exists for both of these simple domain walls, but a significant amount of evidence also suggests that other types of domain walls can also exist and may be critical to a more complete understanding of know ferroelectric wurtzites in addition to providing valuable guidance for new tetrahedrally-based ferroelectric materials.

We combine computational studies based on density functional theory (DFT) and solid-state nudged elastic band (SS-NEB) to describe additional non-trivial switching pathways in multinary wurtzite-based compounds as well as the common alloy systems such as (Al,Sc)N. The focus of this talk will be on how this computational guidance along with high resolution transmission electron microscopy can improve our interpretation of experimental measurements of domain switching behaviors including dynamics, partial switching, and wake-up, eventually leading to a more nuanced and complete understanding of ferroelectricity in the tetrahedrally-bonded crystals.

Full switching of the polarization in wurtzite ferroelectrics was demonstrated for the first time in Al1-xScxN by a group at the University of Kiel. Since that time, ferroelectricity has been demonstrated in Al1-xBxN, Al1-xYxN, AlN, Ga1-xScxN, and Zn1-xMgxO. Most of these films are unipolar as grown, such that they need to be woken up. Typical activation energies for the wake-up process in nitride wurtzites in ~0.15±0.05 eV. Once woken up, the pseudo-activation energies that describe switching between room temperature and 300°C are ~ 20 – 40 meV. In Al1-xBxN, switching occurs through a non-polar intermediate that bears strong structural similarity to the inversion domain boundaries known in the wurtzites. This paper will discuss the kinetics of switching, and the failure mechanisms on bipolar cycling.