The increasing ubiquity of information and communications technologies (ICT) has an underlying and ever-increasing energy cost that is pointing to a scenario of unsustainability. This is particularly acute for Artificial Intelligence (AI) systems, like Large Language Models, whose fast rise in popularity is putting pressure on the energy supply chain due to the large amounts of energy required for their training and operation. These processes relay on data centres, which are expected to account for up to 3.1% of the electricity demand in Europe by 2030. This led to the recognition by the European Commission of the need to achieve climate-neutral data centres in its strategy on “Shaping Europe’s digital future”. The strategy recommends increasing energy efficiency, reusing waste energy, and using more renewable energy sources. One fundamental way of tackling the problem of energy efficiency is to radically change the paradigm of computing supporting AI, which in turn needs a new set of electronic devices and materials to overcome the limitations of current solutions. Topological insulators (TIs), a new class of emerging quantum materials with peculiar properties, could play a role in this electronics revolution. However, due to the challenges associated with the observation of their topological properties, TI research has been mostly restricted to cryogenic temperatures.
With this in mind, project TopRooT targeted the application of TIs for room-temperature electronic devices, with the objective of contributing to establish a clear path for creating and designing such devices. To this end, a crucial technology is the integration of TIs with topologically-trivial materials, such as ferromagnets (FM), which enables applications in spintronics and beyond. Therefore, the project focused on optimising TI/FM heterostructures for spin-orbit torque magnetic random access memory (SOT-MRAM), one of the possible candidates to support energy-efficient computing inspired by the brain, such as neuromorphic computing. TIs promise even higher energy efficiency than that obtained with conventional SOT materials. The objective is to find strategies to achieve an efficient interfacing of MBE-grown Bi2Se3 with relevant FM layers using conventional micro- and nanofabrication technologies and explore the integration of Bi2Se3 with nanoscale magnetic tunnel junctions.
At the end of the project, several fabrication routes to obtain Bi2Se3/CoFeB heterostructures and nanoscale magnetic tunnel junctions for integration with emerging SOT materials were explored and established. The project showed progress in the integration of TIs with ferromagnetic materials and conventional microfabrication techniques, but the transfer methods of the Bi2Se3 from the growth chamber to the subsequent deposition of the ferromagnetic layer need to be further optimized to harness the real potential of the Bi2Se3 for SOT-MRAM.
