Hafnium dioxide for memristive applications

The ever-increasing energy demand of modern society, further reinforced by emerging technologies such as artificial intelligence, presents us with a significant challenge. A solution to this energy dilemma is given by neuromorphic engineering, which has as its objective the construction of an electrical equivalent of different brain and nerve branches, which should enable analog and parallel data processing utilizing oscillating spikes. This technology has the potential to be several orders of magnitude more energy efficient than existing digital Complementary Metal Oxide Semiconductor technology. Hafnium dioxide (HfO${}_2$) is a promising candidate for application in such technologies due to its compatibility with existing technologies and its versatile electrical properties. This work deals with comprehensive analyses of the electrical properties of HfO${}_2$ concerning an application as a memristive device. The first three publications cover the recently discovered ferroelectric properties of HfO${}_2$ and the superconducting properties of niobium nitride electrodes. Ferroelectric HfO${}_2$ films have a scalability down to the single-digit nanometer range, have versatile applications and high potential in both digital and analog technologies. A particular focus is on the training effect known from other ferroelectric films, which is particularly strong for HfO${}_2$, as well as the domain wall behavior and the compatible deposition of superconducting electrodes and ferroelectric thin films. In the two following publications, the memristive behavior and the underlying structure of HfO${}_2$, which are already integrated into a fully functional memristive device, are investigated. In the course of the analysis, it is shown that memristive switching is enabled by a manipulation of a space charge region due to the charging and discharging of electron traps in the HfO${}_2$ layer.