Magnetoelectric flexural gate transistor with nanotesla sensitivity.

摘要

In this dissertation, we have successfully demonstrated the prototype of a chip-scale magnetoelectric (ME) sensor system at room temperature towards biomedical imaging applications, such as magnetocardiography (MCG), magnetoencephalography (MEG). Preliminary results on different approaches to enhance bulk laminate sensor performance have been investigated together with my colleague by Zhao Fang. To achieve sensor integration, an easily controlled deposition process---ion milling sputtering for MetglasRTM (Fe85Si10B5) thin film has been developed and in-situ magnetic domain alignment can be accomplished at room temperature as the film is being deposited. The thin film has been characterized by X-ray Photoelectron Spectroscopy (XPS) for atom composition, SQUID measurement for domain alignment and deflection measurement for magnetostrictive (MS) coefficient.;The first direct integration of thin film MetglasRTM/ Pb(Zr0.52Ti0.48)O3 (PZT) cantilevers shows the quality factor enhancement at the resonant frequency and sensitivity of 1.8 V/T. To improve the integrated sensor performance, a MetglasRTM thin film based magnetoelectric flexural gate transistor (MEFGT) has been realized with the capability of tiny vector magnetic field sensing. The device combines the benefits of high-deflection property of MS cantilever sensors with FET based motion sensing by integrating a magnetostrictive thin film micromechanical cantilever directly atop a sensing and amplifying transistor. Both optical and electrical measurements show the advantage of MEFGT in small magnetic field detection. A sensitivity of 1.5 mV/microT, which corresponds to 150 pT/vHz minimum detectable field (MDF), has been reported as the most sensitive integrated ME sensor to date.;Post simulations of MEFGT have been done to analyze both the mechanical and electrical performance. Device noise modeling and analysis on transistor flicker noise has been established. A strained Si0.5Ge0.5 quantum well FET is introduced as the readout transistor candidate which exhibits the lowest reported flicker noise performance for future low noise sensor devices.;Building on the results of MEFGTs, several readout circuits has been demonstrated. Onboard signal conditioning circuits of charge mode and voltage mode have been realized for both bulk sensors and thin film sensors. Design and simulation for the front-end readout amplifier has been also demonstrated for the on-chip magnetic sensor system. Moreover, post signal processing circuit for real time brain wave signal has been proposed by using a customer designed lock-in amplifier.