Detailed analysis of single molecular junctions for novel computing architectures

Molecular electronics, as a concept to embed molecular compounds into electrical circuits, can be traced back to 1950s. Since that time, a lot of investigations were done, including design and development of molecular-based analogs of electronic compounds such as diodes, transistors etc. A long-term vision of the molecular-scale electronics is the development of unconventional computing scheme based on the properties of individual molecules. Turing-style computers is expected to reach the limit, since further scaling down the computing units has a fundamental constrains in the dimensions. In addition, there are still a lot of computational problems which require exponential amount of resources and computing power. Thereby, the development and implementation of new, unconventional computing paradigms is required. One of the most advanced concepts of unconventional computing is brain-inspired approach. Indeed, human brain has unique computing performance with extremely low power consumption. The hypothetical modern computer to simulate human brain behavior requires gigawatts of power, while the brain itself consumes around 20 W. Therefore even at the beginning of computing era in 1950s von Neumann was looking at the brain for the future developments. All the approaches to mimic brain behavior in the conventional devices get the name of neuromorphic engineering. There are two main approaches in this field: first, to simulate the neuron and its synaptic behavior with possible scaling to the network level, and second, to achieve computing from the network of identical objects. This thesis covers a wide range of experimental investigations in the field of molecular electronics from the level of individual molecular junctions to hybrid devices combining self-assembled molecular networks and graphene