Application of response surface methodology for optimization of Ag-Ag Direct Bonding for Wafer-Level Power Electronics Packaging
Conference: CIPS 2020 - 11th International Conference on Integrated Power Electronics Systems
03/24/2020 - 03/26/2020 at Berlin, Deutschland
Proceedings: ETG-Fb. 161: CIPS 2020
Pages: 8Language: englishTyp: PDF
Yu, Zechun; Zeng, Weijian; Maerz, Martin (Fraunhofer IISB, Schottkystr. 10, 91058 Erlangen, Germany & Friedrich-Alexander University Erlangen-Nuremberg, Cauerstr. 9, 91058 Erlangen, Germany)
Bayer, Christoph Friedrich; Schletz, Andreas (Fraunhofer IISB, Schottkystr. 10, 91058 Erlangen, Germany)
This study evaluated the influence of process parameters on shear strength by the Ag-Ag direct bonding process using response surface methodology (RSM). Central composite design (CCD) based on RSM was applied to determine the effects of the process parameters and their optimum conditions. Silicon dies with two different metallizations (Cr/Ni/Ag and Ti/Ni/Ag) are prepared for this RSM with ten specimens per unit. The effects of isothermal holding temperature (226-294 °C), applied pressure (11.6-28.4 MPa) and bonding time (1.6-18.4 min) were investigated. One-way analysis of variance (ANOVA) was used for statistical data analysis. High value for the coefficient of determination, R2 (91.44% and 94.04 %) and adjusted R2 (80.43 % and 86.39 %) confirm a satisfactory adjustment of the second-order regression model with the achieved data. Further effect of experimental parameters on the shear strength was established by the response surface and contour plots. Results showed that bonding temperature, together with time, affects shear strength significantly. A maximum shear strength of 43 MPa can be achieved with Cr/Ni/Ag metallized dies at 255.6 °C for 18.4 min under 24.2 MPa and with Ti/Ni/Ag metallized dies at 262 °C for 12 min under 23.8 MPa, respectively. Two different failure mechanisms were presented to explain the difference between the two metallizations in shear strength. This study clearly showed that RSM was one of the suitable methods to optimize the process parameters.