A fundamental study of chemical mechanical polishing of copper thin films

This thesis describes an investigation into the chemical mechanical polishing (CMP) of copper films for the purpose of delineating and planarizing inlaid copper interconnections for multilevel metallization in silicon integrated circuits. Copper CMP has been shown to be an effective method of patterning interconnections and for providing the global planarity required to build multilevel structures. However, CMP processes in general remain poorly understood and unoptimized. The task of optimizing a process is hampered by the lack of a fundamental understanding of the removal mechanisms at work in CMP. A fundamental understanding of these mechanisms will allow greater control of the CMP process in the manufacturing environment.;Experimental results of copper CMP are obtained using metal films deposited on 125 mm silicon wafers and 4 cm;Copper CMP involves a number of electrochemical processes; this thesis relates electrochemistry principles to the understanding and process control of copper CMP. Dissolution of abraded material is an electrochemical process as is the formation of a surface film on the copper. In situ electrochemical potential measurements indicate that polish rate is proportional to dissolution rate, while the surface film is important for obtaining a high degree of planarity. Surface film formation during polishing is evaluated via electrochemical potential measurements, X-ray photoelectron spectroscopy, and etch rate measurements. In addition, electrochemical interactions between copper and other exposed metals on the wafer such as the barrier film, are discussed.;The experimental data is considered in the formulation of a model for copper CMP. The proposed model suggests that removal of copper during CMP is the result of a two step process of (1) mechanical abrasion of the surface followed by (2) removal of the abraded material from the vicinity of the surface. Indeed, the polish rate appears to show both abrasion rate (step 1) and dissolution rate (step 2) limited regimes. The degree to which the experimental data fits the model is discussed.