KLA Corporation
Process optimization using design of experiments and response surface models

Abstract:

Using measurements from a metrology tool, combinations of tool settings on the metrology tool can be determined. Candidates can then be determined and a response surface model can be generated for each of the candidates. A list of the candidates of the tool settings that provide a maximum response and that are least sensitive to sources of noise can then be determined from the response surface models. The list of the candidates can each be from a denser region of the response surface model.