This thesis investigates the sources of the significant discrepancy that exists between simulated and the measured results for the temperature dependence of bandgap voltage references. Experiments have been conducted to characterize the diodes which are the main factors contributing to the nonlinearity in the temperature dependence of bandgap reference circuits. Challenges involved in the modelling of the diodes suitable for predicting the performance of circuits which extract the bandgap voltage are discussed. Model parameters such as the ideality factor n and the saturation current temperature exponent m (or xti) are the major terms contributing to the temperature-dependent nonlinearity in existing device models. Extraction of the parameter m, in particular, from measured characteristics of a diode is heavily dependent upon the extraction algorithm used which suggests that there are limitations in the functional form of existing device models when used in applications where the precise relationship between current, voltage, and temperature is critical.

A two level classification of unauthentic ICs into counterfeits and Trojan bearing ICs is discussed that separates perpetuators driven by financial incentives for selling commodity components from those driven by a disruptive payload goal. A technique is discussed for practically introducing identifiers into any IC that eliminates financial incentives for the much more prevalent counterfeit IC problem.