The project focuses on the transfer of neurobiological information processing principles and information storage into superconducting memristive systems with the aim of realizing energetically highly efficient microelectronic circuits for self-adapting (neuromorphic) systems with parallel architecture. Technologically, the aim is to combine superconducting microelectronics and neuromorphic memristor electronics. This approach should rigorously push the limits of today's microelectronic system concepts in terms of signal processing speed and energy efficiency. To solve these problems, novel material systems, devices and computing architectures will be developed. Methodologically, the project is based on the physical investigation and description of quantum mechanical effects in order to make them usable for electronic components, the process and technology development for the hardware realization of such components on wafer level by means of thin film technologies and the development of circuit design concepts (system design).
The investments planned in NSME are aimed in particular at research into physical and material science phenomena and process development for device integration. By means of low-temperature scanning tunneling and low-temperature scanning electron microscopy, the necessary analysis techniques are to be procured and operated in a resource-saving manner by means of helium recovery. The defined deposition of oxide layers (Hf02/Al2O3) under reduced temperatures is to be carried out by means of plasma-assisted atomic layer deposition (PEALD), the planarization of multilayer structures with a Chemical Mechanical Polishing Plant (CMP) and the structuring of the metallization (Nb, Al) by a chlorine-based etching process (CI-ICP) with precise process parameter control. In order to ensure the lithographic structure alignment of the multilayer structures to be realized, it is planned to extend the lithography process by an auto-alignment function. The investments will enable the TU Ilmenau to make significant contributions to research into quantum mechanical effects and to develop neuromorphic superconducting memristive electronics. Building on this, the integration methods are to be combined with established CMOS processes with partners from the Research Factory Microelectronics Germany in the future and a cell library for design scaling is to be built up step by step. The series production of neuromorphic superconducting memristic electronics has an enormous economic potential with great relevance for social and environmental policy.
