Predicting LER and LWR in SAQP with 3D virtual fabrication

For the first time, process impact on line-edge roughness (LER) and line-width roughness (LWR) in a back-end-of-line (BEOL) self-aligned quadruple patterning (SAQP) flow has been systematically investigated through predictive 3D virtual fabrication. This frequency dependent LER study shows that both deposition and etching effectively reduce high frequency LER, while deposition is much more effective in reducing low frequency LER. Spacer-assisted patterning technology reduces LWR significantly by creating correlated edges, and further LWR improvement can be achieved by optimizing individual process effects on LER. Our study provides a guideline for the understanding and optimization of LER and LWR in advanced technology nodes.