Description of Group/Project:
This theoretical Thesis project will focus on developing a workflow to accelerate the discovery and assessment of novel material functionalities based on a combined use of machine learning algorithms, first principles methods, and in-house linear scaling quantum (spin) transport simulations applied to new van der Waals heterostructures based on 2D materials. The project objectives include discovering materials combinations that optimizes spin-orbit torque (SOT) for ultralow-power memory applications, and developing quantum descriptors for material assessment using high-throughput and machine learning approaches.
Main Tasks and responsibilities:
The project's primary goal will be to identify the most promising heterostructures displaying technologically relevant spin-to-charge conversion and SOT efficiencies. The candidate will first be trained to advanced quantum transport physics and spin dynamics in topological quantum matter, and assign preliminary theoretical work on model systems. Then he/she will learn, use and participate in the further development of novel in-house methodological approaches combining high-throughput scanning, guided by a symmetry-driven classification process to search for the upper limit of charge-to-spin conversion efficiency in 2D materials-based heterostructures. The search will be done on a large set of stable 2D materials such as graphene, transition metal dichalcogenides, and 2D magnets like FeGeTe. A technical task/objective will be to optimize the conditions for zero-field SOT-driven switching, which is currently one of the major limiting factors for realizing ultralow-power non-volatile memories based on spin torque physics. Such optimization will be carried by combining machine learning algorithms based on reinforcement and self-supervised learning with the group in-house quantum transport tools, a methodology that allows for the discovery of hitherto unknown novel conditions.