Development of MEMS-based piezoresistive 3D stress/strain sensor using strain technology and smart temperature compensation

Abstract

This paper presents the microfabrication and testing of a membrane-free eight-element single-polarity (n-type) sensing rosette integrated with strained silicon technology over (111) silicon plane to measure the full 3D stress/strain tensor with full temperature compensation. Such n-type piezoresistive (PR) sensor has low sensitivity to the out-of-plane components compared to the in-plane components. To improve the sensitivity of such sensors to the out-of-plane components, a strained silicon technique was integrated into the sensing rosette during the microfabrication process using a highly compressive film produced by plasma enhanced chemical vapor deposition silicon nitride. For experimental verification, a prototype device featuring the proposed sensing rosette was microfabricated using semiconductors fabrication processes. The experimental analysis applied both, in-plane and out-of-plane stresses at different temperatures over a range from −20 °С to 60 °С. In this work, a smart sensing calibration algorithm, utilizing machine learning, is employed to reduce the temperature impact on both sensitivity and resistance of PR coefficients during stress measurement. The developed sensor is capable of accurately extracting the applied stress/strain components with temperature compensation.

Strain Sensing Technology to Enable Next-Generation Industry and Smart Machines for the Factories of the Future: A Review

“…Thanks to the current rapid development in microfabrication and micromachining techniques, MEMS can now be fabricated more efficiently, offering miniaturized and durable solutions. These systems, which combine the features of mechanical and electrical components, have the advantages of compactness, lightweight, superb performance, simplicity in mass production, and competitive prices [233], [234]. Contrary to bulky fiber-optic strain sensors, MEMS devices can be operated wirelessly and instrumented in difficult locations as temperature and load sensors.…”

Development of MEMS composite sensor with temperature compensation for tire pressure monitoring system

“…Temperature drift error can be software-compensated based on real-time information provided by the temperature sensor that is integrated in another signal treatment MCU chip [5]. There are many methods such as polynomial fitting algorithm, support vector machine, neural network and machine learning algorithm [24][25][26][27][28][29]. Compared with other methods, the neural network can fully approximate complex nonlinear relationship between input and output data which makes compensation more effective, and is suitable for later transplantation into microprocessor to run in real time [30].…”