Molecular Field-Coupled Nanocomputing cell modelling and characterisation

Moore’s laws have predicted with extreme accuracy the IC development in terms of transistors per chip, channel length and area. Although these were formulated empirically in the early ’70s, the manufacturers worked hard to stay on the laws’ traced tracks with outstanding results. In the last decade, the development curve has slowed down marking the reach of a possible technology limit. Nowadays, the CMOS is still the most used circuital technology: introduced in the early ’90s, it revolutionised modern electronics with amazing improvements but, now, the enhancement margins are going to saturate. Actually, there are two main ideas on which designers rely to build more valuable systems: the "More than Moore" approach deals with a high level of abstraction, looking at the systems as heterogeneous (SoCs) and trying to reorganize them from the architectural point of view. The other possibility is the "More Moore" one which is interested in the low level of abstraction, working on the transistor implementation. If there are no more margins to improve the technology following the discussed approaches, a new and worthful solution that outperform the precedent one must be considered. There are a lot of new emerging "Beyond CMOS" technologies, but one of the most promising is the Field-Coupled Nanocomputing (FCN). The FCN exploits blocks (QCA) coupled with electrostatic fields to ensure information propagation. This approach avoids charge transport’s defined logic states, enabling the technology for ultra low power, high operating frequency and strongly scaled devices depending on the FCN’s implementation technology. The molecular implementation is considered one of the greatest solutions: each QCA is composed of two complex molecules with three Quantum-Dots each able to aggregate the charge and work at room temperature. Through SCERPA, the algorithm implemented in MATLAB that can simulate information propagation in layouts due to electrostatic fields and externally applied voltages, nowadays creation and evaluation of basic cells are approachable. However, SCERPA can be particularly time demanding depending on the layout under test and the input combinations chosen for the simulation, so that too complex layouts could be untreatable. This thesis project aims to develop a Characterisation tool that can work with SCERPA to extract the Vin −Vout behaviour of generic layouts. The tool would be a valid instrument to both validate and characterise a cell permitting the eventual design fine-tuning and the stamp of a library file exploitable in the simulation of complex devices. The role of the mentioned Characterisation tool is everything but trivial: the necessity to prevent the work condition of a circuit in a more complex layout pretend the use of dedicated add-ons to treat with, for instance, the need for a bi-stable simulation got with a kind of termination, a higher fan-out or the adaptative input combinations measured in the SCERPA context, where the time overhead is hardly dependent from these. The entire tool has been developed with a strong bias towards automation with a precise focus on ease of use. After a brief technology overview with attention to fundamental notions used in the tool implementation, this thesis project would analyze the implementation in MATLAB of the Characterisation tool showing the results on basic layouts at first and applying them on a more complex circuit like an exclusive OR.