Analysis of the block-based circuit design in molecular Field-Coupled Nanocomputing

Since it was introduced in the ‘90s, CMOS technology has brought an important improvement to modern electronics, being significantly suitable for the progressive development of ICs in terms of area, number of transistors per chip and channel length. As predicted by Moore’s laws in the ‘70s, the evolution of these features has followed an exponential trend until now. However, this development curve is expected to encounter a saturation soon: in a few years, it will not be possible to scale down devices and raise the number of transistors per chip at the current rate, due to significant short-channel effects (related to the reduced dimensions of the devices) that negatively influence their performances. This technological limit led to the exploration of innovative approaches that could overcome the problem, essentially divided in two main categories: the “More than Moore” approach, which exploits the existing CMOS technology to integrate devices belonging to different domains on the same SoC (System-on-Chip), and the “Beyond CMOS” approach, which considers alternative technologies able to overcome the standard CMOS limits and, at the same time, to realize a logic that works the same way. This thesis project focuses on Field-Coupled Nanocomputing (FCN), one of the most promising realizations of the latter approach, and in particular its molecular implementation, which uses blocks of 2 molecules each (QCA) coupled with electrostatic fields. Differently from standard CMOS logic, where logic states are defined by charge transport, in this configuration logic information is coded depending on where the charge is localized within the molecules. This brings significant advantages, such as low power, higher operating frequencies and the possibility to realise strongly scaled devices. The aim of this thesis project is to develop and optimize a Characterization Tool capable of obtaining the in-out characteristic of logic gates and interconnections realised with the molecular implementation of FCN, in the most generalized and automated way possible, considering the physical effects due to their real implementation. The Characterization Tool exploits SCERPA, a MATLAB algorithm that evaluates information propagation in molecular FCN layouts, using an iterative procedure that considers the contribution of externally applied voltages. It can be divided into two macro-steps: as a first step, it elaborates the data obtained from SCERPA simulations and, depending on the kind of layout, rearranges them in libraries where the inputs and the related outputs are saved; in its second part, it requires the user to insert a combination of inputs and returns the corresponding outputs of the whole structure, consulting the previously created libraries. The main advantage of the Characterization Tool is that it returns accurate output values if compared to the SCERPA process alone, with a significantly lower time overhead, for any generic layout. In addition, giving the possibility of creating a whole library of components, it allows the user to simulate any combination of these as many times as needed, speeding up the simulations of more complex circuits. Chapter 1 contains a brief technology overview with attention to fundamental notions used in the tool implementation; chapter 2 and 3 analyse more in detail the implementation in MATLAB of the Characterisation tool showing the results on basic layouts at first, then chapter 4 illustrates what happens applying them on a more complex circuit like an exclusive OR. F