Resonance-Based Temperature Sensors using a Wafer Level Vacuum Packaged SOI MEMS Process

Gulsah Demirhan Aydin1,*, Tayfun Akin1,2

1METU MEMS Centre, Middle East Technical University, Ankara, 06510, Turkey

2Electrical and Electronics Eng. Department, Middle East Technical University, Ankara, 06800, Turkey

Adv. Mater. Lett., 2020, 11 (1), 20011462 (1-8)

DOI: 10.5185/amlett.2020.011462

Publication Date (Web): Jan 06, 2020

E-mail: gdemirhan@mems.metu.edu.tr

Abstract


This paper reports the development of resonance-based temperature sensors using a wafer level vacuum packaged SOI MEMS process which is normally used to implement various MEMS sensors, including MEMS gyroscopes and accelerometers. Implementing MEMS temperature sensors in such a MEMS process together with sensitive MEMS sensors allows obtaining temperature data, which is very useful for the compensation of a number of parameters of these MEMS sensors for obtaining improved performance from these sensors. Four different types of temperature sensors are designed considering two types of actuation mechanisms (varying gap and varying overlap) and two different mass types (H-shaped single mass and tuning fork double mass), and their design and model analysis are verified using finite element modelling (FEM) simulations. All of the sensors are fabricated in the same die by using the advanced MEMS (aMEMS) process. The fabricated sensors are combined with necessary readout electronics for each structure in LT Spice environment, and their proper operations are verified in MATLAB Simulink. The temperature sensing technique is based on the frequency variations due to the thermal expansion coefficient mismatch between the glass substrate and the silicon that causes a mechanical strain on the resonator and to a smaller extent, by the temperature variation of Si Young modulus, which influences the resonance frequency. The performance of each sensor is measured using the real time data acquisition from the resonators where resonance frequency and resonator controller outputs are monitored for different temperatures. The best performance is obtained with the tuning fork double mass together with varying gap structures, where the temperature coefficient of frequency (TCF) values are measured as ‑128 ppm/K in the measurement range in the hot plate and as ‑114 ppm/K in the measurement range in the oven.

Keywords

MEMS based resonator, resonance-based temperature sensor, resonance frequency.

Upcoming Congress

Knowledge Experience at Sea TM