Curvilinear multiferroics

Disruptive development of brain-inspired computing, machine learning and deep neural networks for electronic assistants in smartphones, autonomous driving systems or chatGPT require new logic and storage concepts, which are beyond conventional von Neumann computer architectures. Materials known as multiferroics are identified as key enablers of these technologies.

In contrast to ferromagnets, where the magnetic state is controlled by magnetic fields (energy inefficient due to flowing currents), magnetoelectric multiferroics allow manipulation of magnetic states by electric fields.

The great promise of multiferroic materials towards energy efficient computing is compromised by the limited material portfolio. Literally, after decades of research, we have only one multiferroic (i.e. BiFeO3) potentially promising for room temperature spintronic devices.

Even this most advanced material meets challenges outperforming state-of-the-art CMOS elements.The mainstream approach to design magnetoelectrics is to induce ferroelectric ordering in magnets to control non-volatile magnetic states electrically. Although intuitively understandable, this concept failed providing a technologically-relevant multiferroics.

Here, we propose a paradigm shift in the design of multiferroic materials.

Instead of trying to induce ferroelectricity, we will realize curvilinear multiferroics by imposing ferrotoroidal order in geometrically curved magnetic thin films.Fundamentally, we predict that effects of geometric curvature and inhomogeneous mechanical strain will induce sizeable ferrotoroidal order parameter in any magnetically ordered material, rendering this material multiferroic.

Hence, curvilinear systems will turn rare ferrotoroidic materials in a broadly populated material class, available for material science and technologies similar to the conventional ferromagnets.This project will enable curvilinear multiferroics and validate their application potential for memory and logic devices.