Weibull statistics for multiple flaw distributions and its application in silicon fracture prediction

Focus of this paper is the prediction of silicon fracture in microchips of electronic devices. The strength of silicon chips strongly depends on flaw distributions introduced by wafer manufacturing processes. Consequently strength data scatter and a probabilistic approach to failure is required. For this purpose Weibull theory will be used and combined with test procedures and numerical tools for prediction of fracture probability. Multiple active flaw distributions on the surface of the chip and its sawing edges will be considered. Test procedures for strength measurements and Weibull parameter extraction will be presented. Failure prediction will be demonstrated for a microchip showing multiple active flaw distributions which is assembled on a leadframe.