Locally active memristive data processing (LAMP)

• Applicants: Professor Dr.-Ing. Thomas Mikolajick; Professor Dr. Ronald Tetzlaff
• Subject Area: Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering
• Term: from 2015 to 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 273537230

Complex behaviors and rich dynamics may arise in locally active systems only. Local activity is the capability of a nonlinear dynamical system to amplify infinitesimal fluctuations in energy. This property applies to systems of any kind, but it was originally defined in the theory of circuits. The transistors are an example of locally active systems from the field of microelectronics. Their invention, tracing back to 1947, has revolutionized information technology, leading to the appearance of electronic systems which make part of our daily lives. It has been recently shown that particular types of nano-scale memristor may act as locally active circuit elements under suitable biasing. Thanks to its simple structure and small dimension the locally active memristor might lead to the next revolution in the world of electronics, enriching the range of functionalities and/or improving the performance of pre-memristor era circuits. Gaining a full understanding of the peculiar nonlinear dynamics of locally active memristors represents a door opener for the design of circuits based upon them. This first aim of this project lies in an accurate experimental and mathematical characterization of the complex dynamics of a locally active memristor fabricated in house. Secondly, we shall use concepts from nonlinear system theory to derive model representations of the mathematical description of the physical structure for analytical treatment and time efficient simulations. We then plan to investigate the dynamics of the model representations taking the stochastic variability in its nonlinear behavior into account. it is a further goal of our experiments to reduce this variability to a suitable low level as required by the application under investigation through the optimization of the device structure and material choice. Finally, the outcome of the experimental and theoretical investigations shall support the design and fabrication of electrical circuits chosen from two distinct applications - spatio-temporal pattern recognition and digital logic - and based upon the locally-active threshold switching behavior of our device.

• DFG Programme: Research Grants