Optimization of CMP defect detection schemes [VLSI manufacture]

In this paper, the use of Fourier filtering in dark field defect detection is investigated for the post chemical mechanical polishing (CMP) inspection application using a simulation tool, METRO. Advanced detection techniques have been studied to determine methods of increasing capture rates in densely patterned arrays. Two critical CMP process defects, microscratches and particles, are evaluated. The results illustrate that Fourier filtering will improve the defect signal relative to the background signal. Additionally, color variation, underlying grain and pattern noise will be minimized using Fourier filtering.