Ru chemical mechanical planarization ͑CMP͒ was studied in slurries containing titania and potassium bromate at different pH values, showing that the Ru removal rate is enhanced at pH 2 or less. Potentiodynamic polarization studies indicate that the corrosion current is enhanced in the presence of bromate, while the static etch rate experiments show that the etch rate is low. Potassium bromate increases Ru removal only at anodic potentials or during mechanical abrasion. Studies at different concentrations of abrasive and oxidizing agent reveal that the removal rate enhancement saturates at 0.75 mM bromate and 4 wt % titania.Cu is used as interconnect material in integrated circuits, and a thin barrier layer is used to prevent the diffusion of Cu through the interlayer dielectric. 1,2 During damascene and dual damascene fabrication of the interconnect structures, the excess Cu and barrier are removed by a process of chemical mechanical planarization ͑CMP͒. Currently, Ta and TaN are used as barrier metals 1 and a thin Cu seed layer is deposited by physical vapor deposition ͑PVD͒, followed by electrodeposition of Cu. However, as feature dimensions shrink, the PVD Cu is not scalable and Ru is proposed as a suitable seed layer. It has low resistivity as well as negligible solid solubility with Cu. 3,4 Although the introduction of Ru would likely require the development of CMP techniques, only a few articles have appeared that describe the Ru CMP methods. 5-17 Ru is a noble metal with high chemical resistance and mechanical hardness, which makes Ru CMP challenging. Sodium periodate, 5-9 hydrogen peroxide, 10-12 persulfate, 11 Ce͑IV͒, 10,13-16 and perborate 17 have been suggested as suitable oxidizing agents for Ru CMP. Here, we report potassium bromate ͑KBrO 3 ͒ as an oxidizing agent and titania as an abrasive for Ru CMP. Titania was chosen as candidate abrasive because it was reported to yield a higher Ru removal rate ͑RR͒ when compared to silica-and alumina-based slurries. 18 The effect of abrasive loading, oxidizing agent concentration, and pH is systematically studied.
ExperimentalRu CMP was studied with a laboratory-scale polishing instrument ͑Struers LaboPol-5/LaboForce 3͒ using a 25 mm diameter, 3 mm thick Ru disk ͑99.9% purity, William Gregor, Ltd., U.K.͒. Electrochemical grade copper disk of 25 mm diameter was used for copper CMP. A constant downward pressure of 5.8 psi was applied, the platen was rotated at 150 rpm, the disk was rotated at 250 rpm, and a slurry flow rate of 100 mL/min was maintained. The slurry contained deionized water and titania abrasives ͑Ϫ300 mesh, DCW, India͒, and pH was adjusted using HNO 3 or KOH. Each run was conducted for 3 min, and at least three runs were performed at each polishing condition. The RR was calculated by the weight loss method. The average RRs along with the standard deviation are reported. For static etch rate experiments, the Ru disk was kept in 100 mL solution for 10 min with stirring and the concentration of Ru in the solution was measured using a Perkin-Elmer Opt...