A nondestructive raman spectra stress 2D analysis for the pressure sensor sensitive silicon membrane. (November 2019)

• Record Type: Journal Article
• Authors: Wang, Zhenyu; Si, Bingjian; Chen, Su; Jiao, Binbin; Yan, Xiulan
• Is Part Of: Engineering failure analysis. Volume 105(2019)
• Page Start: 1252
• Page End: 1261
• Publication Date: 2019-11
• Subjects: MEMS -- Residual stress -- Raman spectra -- Image mosaic; System failures (Engineering) -- Periodicals; Fracture mechanics -- Periodicals; Reliability (Engineering) -- Periodicals; Pannes -- Périodiques; Rupture, Mécanique de la -- Périodiques; Fiabilité -- Périodiques; Fracture mechanics; Reliability (Engineering); System failures (Engineering); Periodicals; Electronic journals; 620.112
• DOI: 10.1016/j.engfailanal.2019.06.089
• Languages: English
• ISSNs: 1350-6307

Abstract

Abstract: The residual stresses introduced by the bonding process has become a key effect for the MEMS product yield improvements. Many in-situ stress measurements have to destruct the devices. Raman spectroscopy is a nondestructive inelastic scattering measuring. Interpreting the silicon Raman frequency shifting can describe the stress location and intensity quantitatively in the lattice. The confocal microscopic observation can realize micrometer scale resolution. However, the limited visual field cannot provide the device stress distribution overview. In the meantime, the microscope loading platform displacement (only 100 μm) will introduce severe random errors (4 ± 1 μm) in the overlapping movements. Since the local stress intensity could be converted into a RGB gray pixel. Through the image mosaic algorithms, the measured local residual stress distribution maps (750 μm × 160 μm) could be stitched to an enlarged stress distribution map (750 μm × 340 μm). This methodology could not only expand the stress inspection overview(60 ± 2%), but also examine and eliminate the displacement random errors. The analysis results showed that the residual stresses were about 70 ± 5% of the simulation value. The uniform difference between the measured and simulated values indicated a trustable stress analysis. Graphical abstract: Unlabelled Image Highlights: Through the image mosaic algorithms, the Raman spectra measured stress observation field could be expanded. The loading platform displacement random errors could be examined by the Harris feature pixels. The local residue stress details could be more precisely and comprehensive investigated.