Nonplanarity arising from the chemical mechanical polishing of Cu-oxide damascene structures results in the exposure field ͑die-size͒ being partially out of focus in the subsequent lithography process. Thus the corresponding mechanisms of within-die polishing must be determined and the within-die nonplanarity due to polishing needs to be minimized to increase the process yield. In this paper, contact mechanics models were developed to explain the role of pattern geometry on the variation of material removal rate. The effects of Cu linewidth, area fraction, and the elastic properties of the polishing pad on pad displacement into low features were examined to focus on the mechanical aspects of the process. The pressure distribution on the high features was determined and the rate of pattern planarization was quantified. Experiments on patterned Cu wafers were conducted to verify the model. Based on these results, the planarization and polishing behavior and the within-die nonplanarity due to the variation of pattern geometry were discussed.
Test wafers comprising damascene structures were designed and fabricated to investigate Cu dishing and oxide erosion. The mask design covered a wide range of linewidths and pitches, from 0.5 to 100 m, to represent such features as signal and power transmission lines, and probing or wire-bonding pads. Experiments were conducted to investigate the evolution of the pattern profile during polishing and to determine the onset and rates of dishing and erosion. The effects of Cu linewidth and area fraction on the rates of pattern planarization, Cu dishing, and oxide erosion have been quantified. The effect of hardness of the composite surface on dishing and erosion were examined. An optimization scheme, employing particle size, particle hardness, and pad stiffness, to enhance the selectivity between SiO 2 , Ta, and Cu, and to reduce die-scale nonplanarity is proposed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.