Multizone uniformity control of a chemical mechanical polishing process utilizing a pre- and postmeasurement strategy

Achieving good uniformity process control in chemical mechanical polishing (CMP) requires a representative uniformity metric and strong models relatin g this metric to process tunable inputs. Previous efforts in CMP uniformity control have yielded acceptable results utilizing a center-to-edge (CTE) f irst order nonuniformity metric. Closer analysis of post CMP process nonuni formity, however, reveals significant higher order nonuniformity components such as the center "dimple" and outer "doughnut" regions. These nonuniform ity characteristics are due in part to upstream chemical vapor deposition ( CVD) processing. Utilizing a multizone approach to uniformity modeling, a m ore accurate mathematical model of CMP uniformity has been identified. The model has been utilized to customize a thickness and uniformity multivariat e run-to-run software control solution for the process. The controller is b ased on the generic cell controller structure, which is a proven enabler fo r run-to-run control for a number of processes including CMP, vapor phase e pitaxy, and etch. The control algorithm is a zeroth order adjustable linear approximation two-stage algorithm with exponentially weighted moving avera ge noise filtering. This algorithm, which supports first order linear and n onlinear models, has been demonstrated to be effective in CMP CTE and thick ness multivariate control. The control solution has been enhanced to utiliz e both pre and post CMP process metrology along with process models to sugg est process recipe modifications on a run-to-run basis. Results indicate im proved control of CMP process nonuniformity qualities of interest. Further, the results quantify the significant benefit of utilizing premetrology (fe edforward) information in addition to traditional postmetrology (feedback) in determining control recipe advice. (C) 2000 American Vacuum Society. [S0 734-2101(00)06104-2].