Microelectromechanical system-based wireless microsystem for brain interfacing.

摘要

This thesis presents the design, fabrication and characterization of a microelectromechanical system (MEMS) based complete wireless microsystem for brain interfacing, with very high quality factor and low power consumption. Components of the neuron sensing system include TiW fixed-fixed bridge resonator, MEMS oscillator based action-potential-to-RF module, and high-efficiency RF coil link for power and data transmissions.;First, TiW fixed-fixed bridge resonator on glass substrate was fabricated and characterized, with resonance frequency of 100 - 500 kHz, and a quality factor up to 2,000 inside 10 mT vacuum. The effect of surface conditions on resonator's quality factor was studied with 10s of nm Al2O 3 layer deposition with ALD (atomic layer deposition). It was found that MEMS resonator's quality factor decreased with increasing surface roughness.;Second, action-potential-to-RF module was realized with MEMS oscillator based on TiW bridge resonator. Oscillation signal with frequency of 442 kHz and phase noise of -84.75 dBc/Hz at 1 kHz offset was obtained. DC biasing of the MEMS oscillator was modulated with neural signal so that the output RF waveform carries the neural signal information.;Third, high-efficiency RF coil link for power and data communications was designed and realized. Based on the coupled mode theory (CMT), intermediate resonance coil was introduced and increased voltage transfer efficiency by up to 5 times.;Finally, a complete neural interfacing system was demonstrated with board-level integration. The system consists of both internal and external systems, with wireless powering, wireless data transfer, artificial neuron signal generation, neural signal modulation and demodulation, and computer interface displaying restored neuron signal.