Offline analysis techniques for the improvement of defect inspection recipes

Abstract

Yield enhancement techniques for the latest generation of devices need sensitive inspection recipes in order to detect the ever-smaller defects that can result in yield loss. Offline analysis techniques (using MATLAB, for example) for the improvement of bright-field defect-inspection tool recipes are presented. Simple techniques are given for the rapid incorporation or modification of care-areas/don't-care areas into pre-existing recipes. Postprocessing analyses of defect data are presented to show their efficacy in improving the signal-to-noise ratio for defects that might otherwise be hidden in the noise created by 'nuisance' defects. Examples are presented to show how design-databases and reticle inspection data can be harnessed in understanding defect mechanisms.