Granted Patent B2

US 12,505,524 · App. 17/442,662 · Granted Dec 23, 2025

Inventors: Ziyang Ma (Mountain View, CA); Jin Cheng (Santa Clara, CA); Ya Luo (Saratoga, CA); Leiwu Zheng (San Jose, CA); Xin Guo (Sunnyvale, CA); Jen-Shiang Wang (Sunnyvale, CA); Yongfa Fan (Sunnyvale, CA); Feng Chen (San Jose, CA); Yi-Yin Chen (Santa Clara, CA); Chenji Zhang (San Jose, CA); Yen-Wen Lu (Saratoga, CA)

Assignee: ASML NETHERLANDS B.V.

G06T7/0004G03F7/70525G06T2207/20081G06T2207/20084G06T2207/30148

Abstract

Claims (51)

1 . A method for training a patterning process model, the patterning process model configured to predict a pattern that will be formed by a patterning process, the method comprising:

obtaining (i) an image data associated with a desired pattern, (ii) a measured pattern of a substrate, the measured pattern being associated with the desired pattern, (iii) a first model comprising physical terms and configured to make a prediction of a first physical aspect of the patterning process, the first model comprising a first set of parameters, and (iv) a machine learning model configured to make a prediction of a different second physical aspect of the patterning process, the machine learning model comprising a second set of parameters, wherein the first model is coupled to the machine learning model in a series combination or in a parallel combination; and

performing a neural network training to determine values of the first set of parameters and the second set of parameters, wherein the neural network training comprises:

executing, using the image data, the first model and the machine learning model to cooperatively and in combination predict a pattern as expected to be printed on the substrate; and

modifying the values of the first set of parameters and the second set of parameters based on the measured pattern and the predicted pattern.

2 . The method of claim 1 , wherein the first model and the machine learning model are configured and trained in a deep convolutional neural network framework.

3 . The method of claim 2 , wherein the training involves:

predicting the pattern as expected to be printed on the substrate by forward propagation of outputs of the first model and the machine learning model;

determining a difference between the measured pattern and the predicted pattern;

determining a differential of the difference with respect to the first set of parameters and the second set of parameters; and

determining values of the first set of parameters and the second set of parameters by backward propagation of the outputs of the first model and the machine learning model based on the differential of the difference.

4 . The method of claim 1 , wherein the first model is coupled to the machine learning model in a series combination and the series combination of the first model and the machine learning model comprises providing output of the first model as an input to the machine learning model.

5 . The method of claim 1 , wherein the first model is coupled to the machine learning model in a series combination and the series combination of the first model and the machine learning model comprises providing output of the machine learning model as an input to the first model.

6 . The method of claim 1 , wherein the first model is coupled to the machine learning model in a parallel combination and the parallel combination of the first model and the machine learning model comprises:

providing a same input to the first model and the machine learning model;

combining output of the first model and the machine learning model; and

determining the predicted pattern as expected to be printed on the substrate based on the combined output.

7 . The method of claim 1 , wherein the first model is a resist model, and/or an aerial image model.

8 . The method of claim 7 , wherein the first model is a resist model and the first set of parameters of the resist model correspond to at least one selected from:

an initial acid distribution;

an acid diffusion;

an image contrast;

a long range pattern loading effect;

an acid concentration after neutralization;

a base concentration after neutralization;

a diffusion due to high acid concentration;

a diffusion due to high base concentration;

a resist shrinkage;

a resist development; or

two-dimensional convex curvature effects.

9 . The method of claim 1 , wherein the first model is an empirical model that models physics of an aspect of the patterning process.

10 . The method of claim 9 , wherein the first model relates to an acid-based diffusion after exposure of the substrate.

11 . The method of claim 1 , wherein the machine learning model is a neural network that models an aspect of the patterning process.

12 . The method of claim 11 , wherein the second set of parameters comprises: one or more weights and one or more biases of one or more layers of the neural network.

13 . The method of claim 12 , wherein the patterning process model corresponds to a post-exposure process of the patterning process.

14 . The method of claim 13 , wherein the aspect of the post-exposure process comprises: resist baking, resist development, and/or etching.

15 . A computer program product comprising a non-transitory computer readable medium having instructions therein, the instructions, when executed by a computer system, configured to cause the computer system to at least:

obtain (i) an image data associated with a desired pattern, (ii) a measured pattern of a substrate, the measured pattern being associated with the desired pattern, (iii) a first model comprising physical terms and configured to make a prediction of a first physical aspect of a patterning process, the first model comprising a first set of parameters, and (iv) a machine learning model configured to make a prediction of a different second physical aspect of the patterning process, the machine learning model comprising a second set of parameters, wherein the first model is coupled to the machine learning model in a series combination or in a parallel combination; and

perform a neural network training to determine values of the first set of parameters and the second set of parameters, wherein the neural network training comprises:

execution, using the image data, of the first model and the machine learning model to cooperatively and in combination predict a pattern as expected to be printed on the substrate; and

modification of the values of the first set of parameters and the second set of parameters based on the measured pattern and the predicted pattern.

16 . The computer program product of claim 15 , wherein the first model and the machine learning model are configured and trained in a deep convolutional neural network framework.

17 . The computer program product of claim 15 , wherein the first model is a resist model, and/or an aerial image model.

18 . The computer program product of claim 15 , wherein the first model is an empirical model that models physics of an aspect of the patterning process.

19 . The computer program product of claim 15 , wherein the machine learning model is a neural network that models an aspect of the patterning process.

20 . The computer program product of claim 15 , wherein the instructions configured to perform the training are further configured to cause the computer system to:

predict the pattern as expected to be printed on the substrate by forward propagation of outputs of the first model and the machine learning model;

determine a difference between the measured pattern and the predicted pattern;

determine a differential of the difference with respect to the first set of parameters and the second set of parameters; and

determine values of the first set of parameters and the second set of parameters by backward propagation of the outputs of the first model and the machine learning model based on the differential of the difference.

21 . The computer program product of claim 15 , wherein the first model is coupled to the machine learning model in a series combination.

Assignments (2)

Continuity (3)

References Cited (53)