抄録
In the flip chip junction, void in the solder bump has serious influences on reliability and electric characteristic. In this paper, we propose an automated detection technique of micro voids in solder bump. The void makes hollow on the bump shape in X-ray image. We developed subtraction-based method by means of combination of morphology filter and image subtraction to detect the voids. Furthermore, we introduced noise elimination method using some void feature values and discriminant analysis. We evaluated this technique using simulation model and real bump. As a result, a correct rate reached 99.2%, which demonstrates this technique has very high detection ability.
