Electrochemical mechanical planarization (ECMP) of copper

The effect of electrolyte composition on Cu removal rate, planarization efficiency, and surface quality is studied for several different classes of Cu electrochemical mechanical planarization (ECMP) electrolytes. Some of these electrolytes contain benzotriazole (BTA), which has traditionally been employed for passive film formation during Cu chemical mechanical planarization (CMP), while others contain 5-phenyl-1-tetrazole (PTA), which is more effective than BTA at low pH. Experimental results are reported for rotating disc electrode (RDE) studies in a beaker, with and without pad abrasion. These studies predict a 300 mV process window for the best BTA-containing electrolyte in which both adequate Cu removal rate (> 600 nm/min) and adequate planarization efficiency (> 0.90) are achieved. For the best PTA-containing electrolyte, the process window is wider, about 500 mV, even at much lower PTA concentration. The predicted performance of electrolytes using beaker experiments are compared to results in a pilot ECMP tool using commercially available 4" Cu blanket and patterned wafers. Good correlation has been found between the predicted performance of the electrolyte in the beaker experiments using copper RDE and the wafer experiments done on the pilot ECMP tool.;The nature of passive film formed on copper surface is investigated using EIS (electrochemical impedance spectroscopy) and EQCM (electrochemical quartz crystal microbalance). The EQCM studies showed that the passive film formed in PTA-containing electrolytes is more resistant to anodic oxidation than that formed in BTA-containing electrolytes. EIS studies showed that the passive film formed at an applied anodic potential during Cu ECMP is thicker than that which forms during Cu CMP at the open circuit potential. Impedance studies demonstrate that passive film formation is more complete for a glycine complexing agent than for oxalic acid, resulting in lower values for the Cu surface roughness.