Physically unclonable functions, or PUFs, present a means to securely identify objects, both implicit and attached, alongside several uses in conventional secure communication techniques. Many types of PUF based on varying sources of fingerprint entropy have been suggested, and the higher-level theoretical properties and implications of this primitive have been extensively discussed.

However, each different prospective implementation of PUF typically approaches the practical considerations for the conversion from a unique entropy source to ultimate PUF implementation anew. These studies typically treat the intermediate processing schema, such as response binning, solely as a means to an end rather than a subject of explicit discussion and evaluation. As such, there exist few studies into developing a general framework for the optimisation and simulation of the important elements that lie between the measurement of the particular entropy source and the evaluation of the final device as a whole.

This thesis seeks to outline and validate a generalised schema for the conversion of entropy source to final results, presenting the fundamental design elements and figures of merit for the process at every stage where applicable. Further to this, each stage of the process is expressed analytically, allowing the direct derivation of the ultimate figures of merit based on the measurement outcomes of the initial source of entropy. To validate, this process is applied towards the resonant tunnelling diode (RTD) as the prospective entropic unit cell. This type of semiconductor device has several properties that make it an interesting candidate upon which to base a PUF, and this work additionally seeks to outline these benefits and enumerate the general comparative figures of merit for a PUF derived therefrom.