Characterization and Optimization of Single Wafer Processing Equipment with Advanced Aqueous Formulations to Achieve Lower Cost of Ownership for Advanced Cu Interconnect Cleaning Processes

Abstract

The drive to minimize the RC delay in logic copper / low-ɤ interconnect requires new material strategies, especially for porous low-ɤ films for 45 nm and beyond. Incumbent cleaning offerings leave a gap in unmet needs. This gap has been filled with the introduction of advanced aqueous cleaning formulations.These highly aqueous formulations meet customer technical-roadmap requirements while also providing environmental sustainability. Advanced aqueous chemistries demonstrate excellent film compatibility, particularly on porous low-ɤ materials. From a process standpoint, they enable faster chemical dispense coverage on the wafer, readily penetrate very small, high aspect-ratio features, and require significantly shorter de-ionized water (DI) rinse times than more traditional, solvent-based mixtures. These cleaning attributes lead to process cycle-time reductions, higher wafer throughput, and contribute to a lower cost-ofownership (CoO). This study focuses on characterizing the interaction of an advanced aqueous chemical in a single-wafer cleaning tool, and subsequently optimizing the single-wafer equipment setup to minimize chemical usage. Segmentation of where chemical loss occurs and proposed mechanisms for the chemical loss are discussed. The segmentation data are then translated into an optimized best-known method (BKM) for the equipment setup, which includes both equipment facilities settings and specific process recipe parameters. The optimized setup is then validated through a series of marathon wafer runs to calculate process performance and chemical usage. Marathon testing results yielded > 20% reduction in chemical loss while maintaining equal or better cleaning performance compared to the previous process-of-record (POR) for solvents. This emphasizes the importance of matching process and chemical to provide optimal performance and lower CoO.