Method of lithography utilizing orthogonal experiment to improve profile

Abstract

A method of lithography capable of optimizing its process condition more simply and precisely based on a limited number of experiments is provided. A line edge roughness of a resist pattern is obtained as a characteristic value for evaluation. The resist pattern is formed in an orthogonal experiment based on an orthogonal table that includes significant factors, which define a process condition of its lithography. Reference level (benchmark) values are set for each factor in the orthogonal experiment, from which an appropriate reference level value which minimizes the edge roughness is selected as a process condition for proceeding with the lithography. By provision of the edge roughness criterion as the characteristic value which has satisfactory reproducibility, can reflect a change in the process condition very sensitively, and has a linear correlation with respect to the resolution, the influences of respective process conditions on the resolution can be reflected more precisely and with an improved reproducibility.

Dependencies are inferred from the claim text.