CHARACTERIZATION AND MODELING OF III-V TRANSISTORS FOR MICROWAVE CIRCUIT DESIGN

New mobile communication technologies have given a boost to innova-tions in electronic for telecommunications and microwave electronics. It’s clear that the increasing request for mobile data availability, as proved by the growth of 69% of mobile data traffic in 2014, poses great challenges to indus-tries and researchers in this field. From this point of view a rapid diffusion of wireless mobile broadband network data standards, like LTE/4G, should be seen, which requests a state-of-the-art transceiver (i.e., transmitter/receiver) electronics. It will be mandato-ry to use higher frequencies, with wider bandwidth and excellent efficiency, to improve battery duration of mobile phones and reduce the energy consump-tion of the network infrastructures (i.e. base stations). Moreover, the microwave electronics is ubiquitous in satellite systems. As an example the GPS-GLONASS systems, developed respectively by United-States and Russian Federation for geo-spatial positioning, now are commonly used as navigation support for planes, ships, trains, automobiles, and even people. Other interesting applications are the earth-observation satellites, like the Italian system COSMO-SkyMed: a constellation of four satellites developed for the observation of the entire planet. These systems are able to produce a detailed image of the earth surface exploiting a microwave synthetic aperture radar, with the possibility to observe an area even by night or with bad weather conditions. Clearly these features are impossible for traditional opti-cal systems. Even if a lot of electronic applications are focused on the system architec-ture, in microwave electronics the single transistor still plays a key role. In-deed, the number of transistors in high-frequency circuits is low and wide ar-eas are occupied by numerous passive elements, required to optimize the sys-tem performance. There is a lot of interest in finding the optimum transistor operating condition for the application of interest, because the high-frequency electron-device technologies are relatively young and often still in develop-ment, so the transistor performance is generally poor. As a matter of fact, transistor characterization plays a very important role: various measurement systems, developed for this purpose, have been pro-posed in literature, with different approaches and application fields. Moreover, a meticulous characterization of the transistor is the basis for the identification of accurate models. These models, allowing to predict the tran-sistor response under very different operating conditions, represent a funda-mental tool for microwave circuit designers. This thesis will resume three years of research in microwave electronics, where I have collaborated in research activities on transistor characterization and modelling oriented to microwave amplifier design. As various kinds of amplifiers (i.e., low-noise amplifier, power amplifier) have been developed, various characterization techniques have been exploited. In the first chapter, after a presentation of the most common large-signal characterization systems, a low-frequency large-signal characterization setup, oriented to transistor low-frequency dispersion analysis and power amplifier design, will be described as well as the development of the control algorithm of the measurement system and its application to the design of a Gallium-Nitride class-F power-amplifier, operating at 2.4 GHz with 5.5 W of output power and 81% efficiency. Another application of the proposed setup for fast-trap characterization in III-V devices is then reported. Successively, an exten-sion of the setup to very low frequencies will be presented. In the second chapter, small-signal characterization techniques will be dealt with, focusing on noise measurement systems and their applications. Af-ter a brief introduction on the most relevant small-signal measurement system (i.e., the vector network analyzer), an innovative formulation will be intro-duced which is useful to analyze the small-signal response of Gallium-Arsenide and Gallium-Nitride transistors at very low frequencies. Successive-ly, the application of neural network to model the low-frequency small signal response of a Gallium-Arsenide HEMT will be investigated. The third and last chapter will deal with the EM-based characterization of Gallium-Nitride transistor parasitic structures and its usage, combined with small-signal and noise measurements, for developing a transistor model ori-ented to low-noise amplifiers design. In particular, the design of a three stages low-noise amplifier with more than 20 dB of gain and less than 1.8 dB of noise figure operating in Ku-band will be described.