High precision MEMS displacement sensors: device techniques and readout circuits

Research Doctorate - Doctor of Philosophy (PhD)Displacement sensors are key components in the closed loop control of nanopositioning systems, producing repeatable motions with nanometer resolution. Microelectromechanical systems (MEMS) offer a miniaturised solution for implementing low cost and high speed nanopositioners with integrated sensors. Among the MEMS displacement sensing techniques, only the electrothermal and piezoresistive sensors can be implemented for this purpose within a reasonable footprint. The performance of the aforementioned sensors is severely influenced by their inherent high noise levels, which limits the displacement resolution of the MEMS nanopositioner. As such, this thesis presents several displacement sensing techniques to improve the performance of the MEMS nanopositioners. In the first approach, a readout circuit is presented to increase the sensitivity of the electrothermal sensors. The sensor is coupled to a ring oscillator power supply by a ratiometric interface in order to convert the small resistive changes (390~400Ω) to the wide frequency variation (350~550kHz). Subsequently, the frequency demodulator circuits are designed to produce a voltage output. The experiments are carried out on a 1-Degree of Freedom (1-DoF) nanopositioner with a thermal actuator. In addition to the high sensitivity achievement, the ring oscillator nonlinearity is designed to cancel the actuator nonlinearity to produce a linear input-output transfer function, which otherwise needs a lookup table for calibration and closed loop control. The MEMS sensors are fabricated in doped silicon that inherently generates the flicker and thermal noise. The second approach in this thesis consists of two distinct solutions to mitigate the noise contribution. The flicker noise is inversely proportional to the applied heating signal frequency. Hence, a new excitation and readout technique is presented to drive the sensor with a high frequency voltage. The experiments are successfully conducted on a 1-DoF MEMS nanopositioner, which demonstrate an 8dB flicker noise reduction, when compared to the conventional dc (direct current) excitation. In order to alleviate the thermal noise, a multiple sensor system is developed based on the averaging theory, according to which a combination of multiple independent signals contaminated with the uncorrelated noise results in a higher signal-to-noise ratio (SNR). The sensors are implemented around a 1-DoF nanopositioner in order to produce independent measurements of the displacement in real time. The experimental results on three sensors demonstrate a 4dB SNR enhancement, which is in close agreement with the theory. Electrothermal sensors can be operated in constant current (CC) or in constant voltage (CV) excitation modes. In the third approach, an analytic comparison of the two methods is presented. It is shown that from the SNR point of view, the benefits of operating a sensor in CC mode are only marginal. The analytical investigation is supported by experiments performed on sensors integrated into a silicon on insulator (SOI) MEMS nanopositioner with low noise read out circuits, which leads to a 0.04nm/√Hz displacement resolution for both excitation modes. A new 2-Degree of Freedom (2-DoF) MEMS nanopositioner with electrothermal sensors is also presented in this dissertation for the scanning stage of the atomic force microscope (AFM), which is a significant achievement towards the miniaturisation of low cost AFMs for high speed operation. The electrothermal sensors are integrated to the stage in order to provide displacement information for a feedback control loop, which should be otherwise provided by an external sensor, such as an interferometer. The images obtained by the controlled nanopositioner demonstrate a great quality enhancement compared to the open loop nanopositioner. Finally, a piezoresistive sensing technique is presented in this thesis as an alternative integrated small area solution. Compared to the existing piezoresistive sensors, it is fabricated in the standard SOI process without any customised fabrication step. The embedded sensor is designed and characterised for a 1-DoF nanopositioner and exploits the differential architecture to achieve higher sensitivity, linearity and common mode interference rejection