Simulation And Reliability Study Of Multi-Field Coupling Of SOI High Temperature Pressure Sensors In Leadless Packages

Engines are subjected to extreme environments for long periods of time,and real-time and accurate measurement of their in-situ pressure can be a good way to improve equipment performance.In this paper,we design and fabricate a SOI high-temperature pressure sensor in a leadless package structure that can be used at high temperatures based on the principle of silicon-on-glass anode bonding and a high-temperature silver paste sintering process,using SOI(silicon on insulator)as the substrate of silicon wafers,and explore the reliability and stability of its structure,and the related research is as follows.(1)The chip-level and package-level sensors are modeled and simulated for temperature shock,random vibration and multi-field coupling.The results show that the stresses generated by the resistors in the package-level sensors during temperature shock and random vibration are smaller than those in the chip-level,and the output drift due to the change in resistance is smaller when combined with the piezoresistive effect,indicating that the package is conducive to reducing the temperature drift.Under the high temperature,high pressure and random vibration coupling,the maximum equivalent force in the chip-level sensor and the package-level sensor is up to 255.87 MPa and 239.33 MPa,which are within the yield strength of the material,and the package structure has good resistance to high temperature,high pressure and shock.(2)The key packaging parameters were investigated based on the shear strength,and the zero-point output size was used as a criterion to produce a stable sample.The results show that the sensor has good static characteristics before and after the experiments,and its zero-point drift also meets the requirements of use,indicating that the packaging structure of the sensor has good reliability,which provides a certain reference significance for the development and application of high-temperature pressure sensors.(3)The hardware compensation layout was developed to reduce the zero-point drift,and the temperature compensation effect of the BP neural network method optimized by segmental interpolation,binary linear regression,BP neural network method,genetic algorithm and particle swarm algorithm was compared,and the results showed that the BP neural network optimized by particle swarm algorithm has good compensation accuracy and iterative convergence speed,which has certain guidance significance.