Necessity of Chemical Edge Bead Removal in Modern Day Lithographic Processing

摘要

Some form of edge bead removal (EBR) is one of the standard requirements for a lithographic process. Without any intervention, resist may accumulate at the edge of the wafer at up to several times the nominal thickness of the resist. In addition to this edge bead, the resist is likely to wrap around the wafer contaminating the backside of the wafer as well. It's needless to say that such a condition would present a significant contamination risk not only for the resist track and the exposure tool but for process equipment outside of lithography as well. Two not necessarily exclusive strategies have been used in the past for edge bead removal. One is topside chemical EBR where solvent is dispensed on the edge of the wafer as the wafer is rotated immediately after coating, and the other method is where a ring of exposed resist is formed by subjecting the resist on the outer edges of the wafer to a broadband exposure; also know as wafer-edge exposure (WEE). The advantage of the chemical method is that it will remove the photo resist but also the organic anti-reflective coating (ARC), which is not photosensitive. The disadvantage of this method is obvious as any latitude in tool tolerances or imperfections on the wafer will result in solvent dispense to the undesirable areas of the wafer. While the optical method is much cleaner, its main disadvantage is that it will not remove ARC. As the feature size and die size shrink, there is less and less repairable redundancy on modern semiconductor chips. An observed effect in our manufacturing facility has been an increased sensitivity to tool imperfections and a quantifiable level of yield loss due to solvent splashing for the 140 nm generation. Accounting for the fact that the ARC layer is generally an order of magnitude thinner than the resist layer, yield-maximizing setup of edge bead removal for one lithographic layer and complete removal of topside chemical EBR is discussed in detail in this paper as well as the extension of the same principle to maximize yield at other layers.