Search for Ferroelectric Binary Oxides: Chemical and Structural Space Exploration Guided by Group Theory and Computations

The presence of bistable polarization states along with accessible switching capabilities make ferroelectrics an ideal candidate for a variety of applications. Although many conventional ferroelectric (FE) materials, i.e., perovskite-structure-based oxides, have been extensively studied over the past several decades, their success has been limited owing to challenges encountered with scaling down of devices and compatibility with complementary metal-oxide-semiconductor (CMOS) processing. Thus, the surprising discovery of ferroelectricity in thin films (similar to 30 nm) of hafnia (HfO2), the first fluorite-structure-type FE material with CMOS compatibility, revived interest in FE memories, among others. However, the most critical lesson to be learned from the example of hafnia is that even binary oxides can be FE if low-lying metastable (or stable) polar phases are present and accessible. In this contribution, we cast a wider net (to go beyond hafnia), involving 20 other simple binary nonmagnetic oxides, in an attempt to systematically reveal (meta)stable polar phases, if any, in such oxides. We employed a combination of structural search methods, first-principles computations, and group-theoretical considerations to find at least five new simple oxides as potential FE candidates-CaO2 (Pna2(1), 33), SrO2 (Pna2(1), 33), Ga2O3 (Pna2(1), 33), TiO2 (Pca2(1), 29), and Al2O3 (Pna2(1), 33) with space group information included within parentheses. Among them, the thermodynamic stability and ferroelectric properties of a previously unexplored candidate, CaO2, was investigated in detail. Structure refinement from synchrotron X-ray diffraction data confirmed that CaO2 crystallizes in the Pna2(1) phase, with structural parameters in agreement with our theoretical predictions. Finally, we provide an assessment of the potential for the practical realization of the identified candidate oxides in their polar phase(s) based our computations presented within the context of the past empirical/theoretical observations and discuss some challenges that still need to be overcome to successfully realize ferroelectricity in simple oxides beyond HfO2 and ZrO2. More importantly, this work presents a general strategy to search for ferroelectric or functional materials of any class.