Design Information Improves SEM Defect Review Sampling Efficiency

摘要

Defect Pareto quality is improved using design-based binning, which couples layout information from design data with the relative location of each detected defect. SEM review is improved through the binning of systematic and nuisance defects. Brightfield defect inspection results for leading-edge process technologies frequently have very high defect counts, often caused by process or design marginalities that have not yet been identified. A sample of these defects is selected (usually randomly) for SEM review. Two significant issues can occur with random sampling. First, defects can occur over dummy fill or non-critical patterns, which reduce yield learning and waste SEM resources. Second, there is a risk that systematic defects caused by process or design marginalities may not be SEM sampled, or they may not be sampled at the correct ratio.