Impact of different slurry and polishing pad choices on the planarization efficiency of a copper CMP process

The planarization efficiency of four different copper CMP processes was studied and compared. All processes under consideration were tuned to have the same removal rate of 300 nm/min on a blanket copper wafer. In this way, the planarization efficiency of a CMP process depends primarily on the used consumables (slurry and polishing pad) and not on the process conditions such as polishing pressure and speed. The results showed that a CMP process using a fixed-abrasives pad has the best planarization behavior: copper features of different sizes and pattern densities were homogeneously planarized at the highest rate. Conventional CMP and abrasive-less CMP produced similar copper surface topography after 90 s of polishing. Abrasive-less CMP exhibited a slow start to the planarization process. In addition, there was a fairly large difference between planarization rates of small and large copper features. Explanations were proposed for different planarization behaviors of the studied CMP processes.