Ab initio multiscale modelling of multiferroic and magnetic systems for memory applications

NAISS 2023/5-411


NAISS Medium Compute

Principal Investigator:

Vladislav Borisov


Uppsala universitet

Start Date:


End Date:


Primary Classification:

10304: Condensed Matter Physics




Since the discovery of skyrmions in the magnetic B20 compounds MnSi and Co-doped FeSi, the idea of using topologically protected magnetic textures for storing information on the nanometer length scale has been widely discussed and further materials hosting such textures, also going beyond skyrmions, have been identified. Most of the known systems are, however, metallic and do not support the multiferroic behavior, meaning that the only way to control magnetism is by applying external magnetic field or by electric current. A few exceptions are Cu2OSe2O3 and especially lacunar spinels GaV4S8 and GaV4Se8 which show a coexistence of magnetism and ferroelectricity and, for that reason, are rare examples of multiferroic single-phase systems with skyrmions. Due to the rich phase diagram and complex physical behavior, lacunar spinels are being actively investigated and one of the promising future directions would be to study how electric field can be used to switch the magnetic properties and skyrmions through coupling to the crystal lattice. Such electric-field control of magnetic textures is expected to be more energy-efficient than the magnetic-field- or current-control and would have a strong impact on the field of spintronics, finding important applications in the nanoscale memory devices. The proposed project will focus on the theoretical modelling of the magnetic and multiferroic properties of different topologically non-trivial systems, such as lacunar spinels, B20 compounds and layered nanostructures. Multiscale approach will allow to model the system behavior on different length scales, starting with a few Angstrom and proceeding to several hundred nanometers, based on the first principles of quantum mechanics. State-of-the-art electronic structure methods will address the electronic properties which correspond to the shortest length scale. The electronic structure will be calculated using density functional theory and whenever necessary the description of electronic correlations will be improved by means of the dynamical mean-field theory. In the second step, the magnetic interactions of Heisenberg and Dzyaloshinskii-Moriya types will be calculated based on the electronic properties obtained in step 1 and will provide the necessary information for large-scale modelling of magnetic textures. The relative ratio of these interactions is a good indicator of whether a magnetic system can be topologically non-trivial. Finally, the spin dynamics and micromagnetic simulations based on the calculated magnetic interactions will address temperature-dependent magnetic properties and possibility of stabilizing topological textures in external magnetic field. For multiferroic systems, the effect of external electric field will be modelled by considering the magnetic properties for different directions or magnitude of the ferroelectric polarization. A common problem of skyrmionic systems is a low maximal temperature for skyrmion stability which limits their applications. If a multiferroic system with skyrmions at room temperature is discovered within the project, it can lead to important technological advances. Possible strategies for inducing such robust magnetic skyrmions will be the application of mechanical pressure and chemical substitution which will be modelled in this project from first principles. Machine-learning analysis will be applied to the obtained numerical results, in order to reveal important factors for skyrmionic functionality and multiferroic performance.