Ionescu, Mihai AdrianRosca, Teodor2022-06-302022-06-302022-06-30202210.5075/epfl-thesis-8122https://infoscience.epfl.ch/handle/20.500.14299/188794Today, we are witnessing the Internet of Things (IoT) revolution, which facilitates and improves our lives in many aspects, but comes with several challenges related to the technology deployment at large scales. Handling ever growing amounts of information that needs to be sensed, stored, transmitted and processed requires severe improvements in energy efficiency and smart distribution of computational power spanning from Cloud systems (handling Big Data in a massively parallel fashion) all the way to Edge devices (interfaces to the real world), where co-integration of sensing and computation plays a big role. Innovations in this field require the development of new device principles in existing technology platforms and/or new abundant and non-toxic materials that can enable electronic functions beyond the classical semiconductors, such as the field of oxide electronics that holds promise for both classical electronics functions and for future neuromorphic implementations. In this thesis, we explore both these aspects, having as a common denominator steep slope devices, which have the merit of offering a path for improved energy efficiency via voltage scaling. We particularly focus our work on their ability to serve energy efficient sensing functions that can be integrated with the computational platforms. The first part of the thesis focuses on Tunnel Field Effect Transistors (TFETs) and how they can be used to perform similar tasks to Single Electron Transistors for qubit readout and also for serving as interfacing electronics. Such applications rely on cryogenic operation where conventional CMOS technology shows performance degradation due to low temperature effects such as dopant deactivation and carrier freeze-out. Our study shows that state-of-the-art heterostructure nanowire TFET arrays maintain excellent figures of merit over wide temperature ranges, down to the Kelvin regime, while simultaneously showing reduced temperature dependence once Trap Assisted Tunnelling mechanisms are removed below 150K. Leveraging such properties, we suggest that TFETs are promising candidates as charge sensing devices for qubit readout architectures with high sensitivity to single or few elementary charges. In the second part of the thesis we focus towards sensing architectures more suitable for Edge-of-Cloud (EoC) applications, by exploring phase-transition materials such as Vanadium Dioxide (VO2). In this context, we explore the optimization of a Pulsed Laser Deposition (PLD) process in order to achieve high quality VO2 thin films grown on CMOS compatible substrates, followed by electrical characterization of fabricated VO2 two-terminal devices, which provides valuable data that aid us in developing compact SPICE-compatible device models. Built on top of the VO2 resistor elements, we propose a novel Spiking Voltage-Controlled Oscillator (VCO) architecture that exhibits low device count (1 Transistor 1 Resistor - 1T1R) while at the same time providing frequency tuning capabilities in excess of 400% in the 10s of kHz range. We experimentally validate that the VCO cell can be used as a power-to-frequency transducer in a wide spectrum, ranging from near-UV, throughout the entirety of the visible domain, and as far as the Mid-Infrared and mmWave ranges, suggesting a new class of sensors capable of responding to a broad range of stimuli.enSteep-slope devicestunnel field-effect transistor (TFET)band-to-band tunneling (BTBT)cryogeniccharge sensingmetal-insulator transition (MIT)vanadium dioxide (VO2)pulsed laser deposition (PLD)voltage controlled oscillator (VCO)sensing platformthesis::doctoral thesis