Nondestructive inspection and inline estimation of profiles of copper-filled through-silicon vias with voids by a nano-focus X-ray microscope

We proposed a scheme of nondestructive inspection and simple inline estimation of profiles of copper-filled through-silicon vias by viewing with a nano-focus X-ray microscope with a resolution of 250nm and applying supervised machine learning method. Defects in Φ5μ×50μm copper-pillars were recognized and classified and at the same time estimation of profiles of voids in the copper-pillars was carried out in cases that the voids can be assumed rotationally symmetric. Results of experiments showed good matches in the classification and estimation of geometrical parameters of the profiles of the copper-pillars with and without voids gave a good correlation with results of cross-sectional focused ion beam microscope views.