# US 2026/0073504 A1 · Method for training or using a process model for determining a pattern in a patterning process

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0073504 A1
- Document ID: US20260073504A1
- Status: Published application
- Kind: A1
- Application number: 19/393,193
- Filed: 2025-11-18
- Published: 2026-03-12
- Earliest claimed: 2019-03-25
- Assignees: ASML Netherlands B.V.
- Applicants: Asml Netherlands B.V.
- Inventors: Ziyang MA; Jin CHENG; Ya Luo; Leiwu ZHENG; Xin GUO; Jen-Shiang WANG; Yongfa FAN; Feng CHEN; Yi-Yin CHEN; Chenji ZHANG; Yen-Wen LU
- Web page: https://patentgrep.com/patent/US20260073504A1

## CPC

- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06T 7/0004: {Industrial image inspection}
- G03F 7/70441: {Optical proximity correction [OPC]}
- G03F 7/705: {Modelling or simulating from physical phenomena up to complete wafer processes or whole workflow in wafer productions}
- G03F 7/70525: {Controlling normal operating mode, e.g. matching different apparatus, remote control or prediction of failure}
- G03F 7/70625: {Dimensions, e.g. line width, critical dimension [CD], profile, sidewall angle or edge roughness}
- G03F 7/70666: {Aerial image, i.e. measuring the image of the patterned exposure light at the image plane of the projection system}
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30148: Semiconductor; IC; Wafer
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G03F: PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR; (phototypographic composing devices B41B; photosensitive materials or processes for photographic purposes G03C; electrophotography, sensitive layers or processes therefor G03G)

## Abstract

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 involves obtaining an image data associated with a desired pattern, a measured pattern of the substrate, a first model including a first set of parameters, and a machine learning model including a second set of parameters; and iteratively determining values of the first set of parameters and the second set of parameters to train the patterning process model. An iteration involves executing, using the image data, the first model and the machine learning model to cooperatively predict a printed pattern of the substrate; and modifying the values of the first set of parameters and the second set of parameters such that a difference between the measured pattern and the predicted pattern is reduced.

## Claims

### Claim at position 1

-15. (canceled)

### Claim 16

A method comprising:
obtaining (i) resist pattern data associated with a target pattern to be printed on a substrate, (ii) physical effect data characterizing effects of an etching process on the target pattern, and (iii) measured bias between a resist pattern and an etched pattern formed on a printed substrate; and
training, based on the resist pattern data, the physical effect data, and the measured bias, a machine learning model configured to predict an etch bias related to the etching process to reduce a difference between the measured bias and the predicted etch bias.

### Claim 17 (depends on 16)

The method of claim 16, wherein the machine learning model is configured to receive the resist pattern data at one layer of the machine learning model, and the physical effect data is received at a different layer of the machine learning model.

### Claim 18 (depends on 17)

The method of claim 17, wherein the different layer is a last layer of the machine learning model, wherein an output of the last layer is a linear combination of (i) the etch bias predicted by executing the machine learning model using the resist pattern data as an input, and (ii) another etch bias determined based on the physical effect data related to the etching process.

### Claim 19 (depends on 18)

The method of claim 18, wherein the output of the last layer is an etch bias map from which the etch bias is extracted, wherein the etch bias map is generated via:
executing the machine learning model using the resist pattern data as the input to output an etch bias map, wherein the etch bias map includes a biased resist pattern; and
combining the etch bias map with the physical effect data.

### Claim 20 (depends on 16)

The method of claim 16, wherein the machine learning model is configured to receive the resist pattern data and the physical effect data at the first layer of the machine learning model.

### Claim 21 (depends on 16)

The method of claim 16, wherein the training of the machine learning model is an iterative process comprising:
predicting, via executing the machine learning model using the resist pattern data and the physical effect data as inputs, the etch bias;
determining the difference between the measured bias and the predicted etch bias;
determining a gradient of the difference with respect to one or more model parameters of the machine learning model;
adjusting, using the gradient as a guide, one or more model parameter values such that the difference between the measured bias and the predicted etch bias is reduced;
determining whether the difference is minimized or breaches a training threshold; and
responsive to the difference not minimized or the training threshold not breached, repeating the predicting with the machine learning model having the adjusted one or more model parameter values, the determining the difference with that new prediction, the determining a gradient using that new difference, the adjusting one or more model parameters using that new gradient and the determining whether that new difference is minimized or breaches the training threshold.

### Claim 22 (depends on 16)

The method of claim 16, wherein the obtaining of the resist pattern data comprises executing, using the target pattern to be printed on the substrate, one or more process models including a resist model of the patterning process.

### Claim 23 (depends on 16)

The method of claim 16, wherein the resist pattern data is represented as a resist image, the resist image being a pixelated image.

### Claim 24 (depends on 16)

The method of claim 16, wherein the physical effect data is data related to an etch term characterizing an etching effect, the etch term comprising at least one selected from:
a concentration of plasma within a trench of the resist pattern associated with the target pattern;
a concentration of plasma on top of a resist layer of the substrate; a loading effect determined by convolving the resist pattern with a Gaussian kernel having a specified model parameter;
a change in the loading effect on the resist pattern during the etching process;
a relative position of the resist pattern with respect to neighboring patterns on the substrate;
an aspect ratio of the resist pattern; or
a term related to a combined effect of two or more etching process parameters.

### Claim 25 (depends on 16)

The method of claim 16, wherein the obtaining of the physical effect data comprises executing a physical effect model including one or more etch terms and a Gaussian kernel specified for the respective one or more etch terms.

### Claim 26 (depends on 16)

The method of claim 16, wherein the physical effect data is represented as a pixelated image, wherein each pixel intensity is indicative of the physical effect on the resist pattern associated with the target pattern.

### Claim 27 (depends on 16)

The method of claim 16, further comprising:
obtaining a resist contour of the resist pattern; and
generating an etch contour by applying the etch bias to the resist contour.

### Claim 28 (depends on 16)

The method of claim 16, further comprising executing the trained machine learning model using the resist pattern and the physical effect data as the input to determine the etch bias.

### Claim 29 (depends on 16)

A non-transitory computer-readable medium comprising instructions therein, the instructions, when executed by a computer system, configured to cause the computer system to at least perform the method of claim 16.

### Claim 30

A method comprising:
determining, via executing a physical effect model, physical effect data characterizing an effect of an etching process on a substrate;
executing a trained machine learning model using a resist pattern and the physical effect data as input to determine an etch bias; and
controlling a semiconductor apparatus or process based on the etch bias.

### Claim 31 (depends on 30)

The method of claim 30, wherein the trained machine learning model is a convolutional neural network (CNN) including specific weights and biases, wherein the weights and biases of the CNN are determined via a training process employing a plurality of resist patterns, physical effect data associated with each resist pattern of the resist patterns, and measured bias associated with each resist pattern of the resist patterns, so that a difference between the measured bias and the determined etch bias is minimized.

### Claim 32 (depends on 30)

The method of claim 30, further comprising executing, using a target pattern to be printed on the substrate, one or more process models including a resist model of the patterning process to obtain the resist pattern.

### Claim 33 (depends on 30)

The method of claim 30, wherein the machine learning model is configured to receive the resist pattern data at one layer of the machine learning model, and the physical effect data is received at a different layer of the machine learning model.

### Claim 34 (depends on 30)

The method of claim 30, further comprising:
obtaining a resist contour of the resist pattern; and
generating an etch contour by applying the etch bias to the resist contour.

### Claim 35 (depends on 30)

A non-transitory computer-readable medium comprising instructions therein, the instructions, when executed by a computer system, configured to cause the computer system to at least perform the method of claim 30.
