US 12,093,632 B2
Machine learning based inverse optical proximity correction and process model calibration
Marinus Aart Van Den Brink, Moergestel (NL); Yu Cao, Saratoga, CA (US); and Yi Zou, Foster City, CA (US)
Assigned to ASML NETHERLANDS B.V., Veldhoven (NL)
Filed by ASML NETHERLANDS B.V., Veldhoven (NL)
Filed on Sep. 22, 2022, as Appl. No. 17/950,502.
Application 17/950,502 is a continuation of application No. 15/734,141, granted, now 11,544,440, previously published as PCT/EP2019/063282, filed on May 23, 2019.
Claims priority of provisional application 62/685,749, filed on Jun. 15, 2018.
Prior Publication US 2023/0013919 A1, Jan. 19, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 30/398 (2020.01); G06F 30/392 (2020.01); G06F 119/18 (2020.01)
CPC G06F 30/398 (2020.01) [G06F 30/392 (2020.01); G06F 2119/18 (2020.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
obtaining a first patterning device pattern from simulation of an inverse lithographic process that predicts, from a substrate target layout, an output patterning device pattern, the output patterning device pattern corresponding to a pattern configured to be transferred, by a lithographic apparatus, from a patterning device onto a substrate with the aim to form the substrate target layout;
receiving substrate data corresponding to a substrate exposed using the first patterning device pattern; and
training, by a hardware processor system, an inverse process model configured to predict a second patterning device pattern using the substrate data related to the exposed substrate and the first patterning device pattern,
wherein the training the inverse process model is iterative, an iteration comprising:
determining one or more model parameter values of the inverse process model based on the substrate data and the first patterning device pattern; and
adjusting the one or more model parameter values until a first cost function of the inverse process model is improved.