High Performance Center »Functional Integration in Micro- and Nanoelectronics«
The growing demands of artificial intelligence require new approaches to processing data with high energy efficiency and low latency. For edge applications in particular, it is advantageous to process information as close as possible to where it is generated. Neuromorphic computing and compute-in-memory follow a fundamentally different approach: memory and computing functions are closely integrated, eliminating the need to transfer data between separate memory and processing units for every operation.
The High Performance Center Micro/Nano provides expertise across a broad technological chain – from materials and novel devices through their fabrication, characterization, and integration to circuits and architectures. A key foundation is provided by many years of research into ferroelectric and memristive technologies. Ferroelectric devices such as FeFETs enable non-volatile memory functions and provide a basis for neuromorphic and compute-in-memory concepts. In parallel, memristive devices have been developed, among others, based on BiFeO₃ and TiO₂, and have been implemented both as individual devices and in crossbar arrays.
This technological foundation enables the investigation of new concepts for processing AI models and their step-by-step development toward application-oriented hardware. This includes the development and characterization of memory and synaptic devices, the realization of high-density crossbar structures, integration into existing technology concepts, and the design of circuit components for neuromorphic systems. The integration of analog computing cores with digital peripherals and suitable data converters is also an active area of research.
The technological foundation has been built through numerous research activities and projects. These include studies on FeFET-based synapses and neuromorphic circuits, ferroelectric memory arrays, memristive devices and their integration, as well as current research on compute-in-memory and energy-efficient edge AI.
For customers and collaboration partners, this provides access to a development environment in which new materials, devices, and architectures do not have to be considered in isolation. Technologies can be investigated in terms of their electrical properties, scalability, and integration capabilities, and subsequently advanced at the circuit and system levels. This makes it possible to assess new approaches at an early stage with regard to energy efficiency, performance, reliability, and technical feasibility.
Your Benefits / Our Offering
- Devices and technologies: Development and characterization of ferroelectric and memristive devices, including crossbar structures and their integration into existing technology concepts.
- Circuits and architectures: Development of components and concepts for neuromorphic systems and compute-in-memory, including circuit design and the integration of analog and digital processing.
- Application-oriented development: Joint investigation and validation of customer-specific approaches – from materials or devices through integrated technologies to demonstrators.