Researchers from Washington University in St. Louis, in collaboration with scientists from the University of Nebraska–Lincoln, Georgia Tech, and South Korea’s UNIST, have identified ferroelectric domains smaller than 1 nm² in zirconia, a material chemically similar to an insulator used in silicon transistors. These atomic-scale domains could potentially allow for significantly denser data storage in semiconductor memory devices. The team combined atomic-scale simulations with electron microscopy and experimental data to study how these materials could address the growing demand for higher-density data storage driven by expanding AI applications. In a related study, they discovered that hafnia, an insulating material already used in silicon transistors, can exhibit ferroelectric properties in extremely thin films, which is unusual as most ferroelectric materials lose this behavior when thinned below 10 nanometers. The researchers found that hafnia can be grown in a stable antiferroelectric structure and then switched using an electric field during operation, offering a potential pathway for memory applications. Additionally, they identified that oxygen atoms help stabilize tiny zirconia domains, and their calculations suggest that the boundaries between these domains, known as domain walls, could conduct electricity and move under voltage, potentially serving as components in future memory devices. The studies, published in Advanced Materials and Science journals, highlight two different approaches to controlling electrical behavior at the atomic scale: hafnia through switchable crystal structures and zirconia through domain wall movement. The use of hafnia, which is already integrated into silicon transistor technology, provides a familiar material for engineers to explore these memory applications. These findings represent a significant step toward developing next-generation memory technologies that could support the increasing demands of artificial intelligence and data-intensive applications.