Local stress measurement of thin films with wafer-level integrated test structures

Abstract

Residual stress remains in a material or structure after manufacturing and processing in the absence of external forces or temperature difference. Further, these residual stresses are one of the most common bottlenecks in the case of the fabrication of novel microelectromechanical systems (MEMS). As stress characterisation techniques utilised at the macroscale are not suitable for measuring the local stress on MEMS devices, various MEMS-applicable alternatives have been researched in recent years. The goal of this work was to find such an alternative with an optical detection scheme. First, a brief overview of the macroscale stress measurement techniques is followed by a comprehensive overview of the existing MEMS residual stress characterisation approaches. Next, an adapted version of the rotating arm test structure described in the first chapter was simulated in COMSOL Multiphysics to find the best design for achieving the targeted measurement range: from -1 GPa to 1 GPa. The optimised design parameters served as the basis for the lithography mask designs presented in the third chapter of this diploma thesis. These designs include measures implemented to reduce the out-of-planedeflection and to ensure easy readability of the measurement results, as well as a successful fabrication of the test structure. Moreover, the detailed dimensions of both a smaller and a larger rotating arm test structure are included, followed by a brief description of their fabrication process. To check if the fabricated rotating arm test structure fulfils its purpose successfully, several a-SiC:H, Pt and AlN, and Cr thin films with different mechanical stress levels and thicknesses were deposited by using PECVD, sputtering, and evaporation, respectively, on top of the rotating arm test structures. The verificationmeasurements confirmed that both larger rotating arm test structures covered with 375 nm thick, PECVD-deposited a-SiC:H thin films function correctly but only reach 57 and 60 % of the simulated displacement. These values are significantly smaller than the 95.9 % reached by the best-performing rotating arm test structure presented in this diploma thesis: the larger device rotated by 45◦, coated with a 250 nm thick, evaporated Cr thin film. On the other side, out-of-plane deflection is still observable for this best-performing test structure. Reducing this out-of-plane deflection may be achieved by depositing thestressed thin films only on specific parts of the rotating arm test structure and an increase of the measured vs expected x-displacement ratio may be achieved by making several design adjustments on the central part of the fabricated rotating arm test structure.Thus, further research is needed to verify if the proposed improvement measures will be able to lead to a better-performing rotating arm test structure design.