Abstract
Key issues involved in implementing Warnock's phenomenological erosion model for chemical‐mechanical polishing (CMP) optimization are addressed. Mathematical details of the model are examined through an enhanced and generalized development. The improved development gives rise to a version of the model that allows for closer fits to experimental data and nonuniform discretization of the wafer features. The model's suitability for use in wafer polishing optimization is determined by examining the model's dependence on its adjustable parameters and by addressing its computational efficiency. Lastly, the minimization of adequate objective functions representing the optimal CMP process is described.