Energy-Efficient Analog Front-End Design Using Flexible Electronics:and its Application in Wearable Biopotential Readout Systems

This thesis explores how to make flexible electronics more energy-efficient for use in wearable devices that monitor biopotential signals like EEG, ECoG, and EMG. The focus is on designing circuits that reduce noise and boost input impedance, all while keeping power consumption low. These improvements make flexible electronics suitable for high-quality, fully wearable health monitoring systems. Chapter 1 sets the stage by outlining the challenges in current flexible electronics used for biopotential sensing. It highlights the need for circuits that offer both low noise and high input impedance. Chapter 2 introduces a mathematical model for the 1/f noise in TFT circuits that employ noise-cancellation techniques. The model can be used to design high-efficiency low-noise analog front-ends. Chapter 3 presents a new circuit design for reading out biopotential signals without galvanic contact to the skin. Fabricated using a-IGZO TFT technology, it uses a novel autozeroing method in a ping-pong buffer setup to achieve low noise and high input impedance. The circuit shows good performance, with 250MΩ input impedance at 50Hz, 1kHz bandwidth, and just 0.22mW power usage, making it ideal for flexible health patches. Chapter 4 describes a fully wearable prototype for controlling prosthetic limbs. It features a unique multiplexing front-end circuit on a flexible foil, with each channel designed to minimize added noise and maintain high input impedance. The system reaches 841MΩ input impedance at 50Hz, 22µVrms input-referred noise, and minimal interference between channels. It successfully controls a prosthetic arm with performance comparable to current state-of-the-art technologies. Chapter 5 presents a flexible analog front-end for brain-computer interfaces (BCIs), designed to support high electrode counts in implantable systems. Built with a-IGZO TFTs, the circuit achieves low noise (70nV/√Hz) and low power density (3.5mW/cm²), meeting key requirements for safe, scalable neural monitoring. Tested with real BCI data, it enables accurate gesture classification (94% accuracy), showing its potential for next-generation, high-resolution neural interfaces. Chapter 6 introduces an EEG acquisition system based on a-IGZO and Si-Electronics that combines a flexible low-noise front-end with a custom silicon-based analog-to-digital converter (ADC). This setup is designed to record EEG brain activity with 83-dB dynamic range using only 24µW of power per multiplexer of four channels, showing that flexible and silicon technologies can work together effectively for biopotential sensing. In conclusion, the thesis shows that flexible electronics can meet the demanding requirements of bipotential monitoring applications. It lays the groundwork for ultra-thin, comfortable, and integrated wearable and implantable devices that offer high signal quality with minimal power use.