# US 12,111,579 B2 · Method and apparatus for evaluating an unknown effect of defects of an element of a photolithography process

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

- Number: US 12,111,579 B2
- Document ID: US12111579B2
- Status: Granted
- Kind: B2
- Application number: 18/241,578
- Filed: 2023-09-01
- Published: 2023-12-21
- Granted: 2024-10-08
- Earliest claimed: 2018-05-18
- Assignees: Carl Zeiss SMT GmbH
- Applicants: Carl Zeiss SMT GmbH
- Inventors: Alexander Freytag; Christoph Husemann; Dirk Seidel; Carsten Schmidt
- Web page: https://patentgrep.com/patent/US12111579B2

## CPC

- G03F 7/705: {Modelling or simulating from physical phenomena up to complete wafer processes or whole workflow in wafer productions}
- G03F 1/84: Inspecting
- G06F 30/20: Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30)
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/0475: Generative networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- G06T 7/0006: {using a design-rule based approach}
- G06T 7/001: {using an image reference approach}
- G06V 10/751: {Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching}
- G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
- G06V 20/13: Satellite images
- G06T 2207/10032: Satellite or aerial image; Remote sensing
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30148: Semiconductor; IC; Wafer
- 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)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

The present invention relates to a method and an apparatus for determining at least one unknown effect of defects of an element of a photolithography process. The method comprises the steps of: (a) providing a model of machine learning for a relationship between an image, design data associated with the image and at least one effect of the defects of the element of the photolithography process arising from the image; (b) training the model of machine learning using a multiplicity of images used for training purposes, design data associated with the images used for training purposes and corresponding effects of the defects; and (c) determining the at least one unknown effect of the defects by applying the trained model to a measured image and the design data associated with the measured image.

## Claims

### Claim 1

A method for determining at least one unknown effect of defects of an element of a photolithography process, wherein the method comprises the steps of:
a. providing a model of machine learning for a relationship between an image, design data associated with the image and at least one effect of the defects of the element of the photolithography process arising from the image;
b. training the model of machine learning using a multiplicity of images used for training purposes, design data associated with the images used for training purposes and corresponding effects of the defects; and
c. determining the at least one unknown effect of the defects by applying the trained model of machine learning to a measured image and the design data associated with the measured image.

### Claim 2 (depends on 1)

The method of claim 1, wherein the image comprises at least one element from the group: an image recorded by an optical imaging system, an image recorded by a scanning particle microscope, or an image recorded by a scanning probe microscope.

### Claim 3 (depends on 1)

The method of claim 1, wherein the model of machine learning comprises at least one element from the group: a parametric mapping, an artificial neural network, a deep neural network, a time delay neural network, a convolutional neural network, a recurrent neural network, a long short-term memory network, or a generative model.

### Claim 4 (depends on 1)

The method of claim 1, wherein the model of machine learning comprises:
a. at least one encoder block for determining information-carrying features of an image and the design data associated with the image; and
b. at least one decoder block for producing at least one effect of the defects from the determined information-carrying features, wherein the at least one effect of the defects shows what an overlay of the image with a reference image looks like.

### Claim 5 (depends on 1)

The method of claim 1, wherein the element of a photolithographic process comprises at least one element of the group: a photolithographic mask, a template for nanoimprint technology, or a wafer.

### Claim 6 (depends on 1)

The method of claim 1, wherein the defects comprise at least one element from the group: placement errors of one or more pattern elements of the element of the photolithography process; critical dimension errors of one or more pattern elements of the element of the photolithography process; and overlay errors of two or more photolithographic masks.

### Claim 7 (depends on 1)

The method of claim 1, wherein determining the at least one unknown effect of a defect comprises at least one of: determining whether the effect of the defect is below a predetermined threshold, determining whether the defect is visible on a wafer, or whether the defect visible on the wafer is correctable.

### Claim 8 (depends on 1)

The method of claim 1, wherein the image recorded by an optical imaging system comprises at least one of: an aerial image or at least two images of an aerial image focus stack.

### Claim 9 (depends on 1)

The method of claim 1, wherein training the model of machine learning comprises assessing the trained model of machine learning based on validation data.

### Claim 10 (depends on 1)

The method of claim 1, wherein training of the model of machine learning comprises: providing the plurality of images used for training purposes and the plurality of design data associated with the images used for training purposes as input data and providing the plurality of effects of the defects corresponding to the images used for training purposes as comparison data for the output data of the model of machine learning.

### Claim 11 (depends on 1)

The method of claim 1, wherein the design data comprises at least one element of the group: layout data or computer-aided design data.

### Claim 12 (depends on 1)

The method of claim 1, wherein the design data comprises at least one element of the group: pixel data, polygons, or polygon chains.

### Claim 13 (depends on 1)

The method of claim 1, further comprising adapting a number of layers of the model of machine learning to a predetermined accuracy of the image.

### Claim 14 (depends on 1)

The method of claim 1, wherein images used for training purposes comprise measured images and/or simulated images.

### Claim 15 (depends on 1)

The method of claim 1, further including the step of: simulating design data and/or modified design data of the element for the photolithography process for the purposes of producing simulated images.

### Claim 16 (depends on 1)

The method of claim 1, wherein producing simulated images comprises at least one element from the group: carrying out a rigorous simulation by numerically solving Maxwell's equations, wherein design data and/or modified design data of the element of the photolithography process are used as input data, carrying out a simulation with the aid of a Kirchhoff model, wherein the design data and/or the modified design data of the element of the photolithography process are used as input data, carrying out a particle-beam-based imaging simulation, wherein design data and/or modified design data of the element of the photolithography process are used as input data and carrying out a scanning-probe-based imaging simulation, wherein design data and/or modified design data of the element of the photolithography process are used as input data.

### Claim 17 (depends on 1)

The method of claim 1, wherein the provision of corresponding effects of the defects for the purposes of training the model of machine learning further comprises the step of: overlaying an image used for training purposes with a reference image for producing the at least one effect of the defects corresponding to the image.

### Claim 18 (depends on 1)

The method of claim 1, wherein the provision of corresponding effects of the defects for the purposes of training the model of machine learning further comprises the step of: determining a reference image by: imaging a substantially defect-free region of the element of the photolithography process, which has the same pattern elements as the region of the measured image, and/or simulating the design data for the region of the measured image of the element of the photolithography process.

### Claim 19 (depends on 17)

The method of claim 17, wherein overlaying the image with the reference image comprises: forming a difference between the image and the reference image.

### Claim 20 (depends on 19)

The method of claim 19, wherein forming the difference comprises at least one element from the group: determining a deviation of a critical dimension, determining a contrast deviation, and determining a placement deviation of one or more pattern elements.

### Claim 21 (depends on 1)

The method of claim 1, wherein the training of the model of machine learning comprises:
a. training the model of machine learning using a first number of simulated images, design data associated with the simulated images with corresponding effects of the defects in a first phase; and
b. training the model of machine learning using a second number of measured images, design data associated with the measured images with corresponding effects of the defects in a second phase, with the first phase being carried out before the second phase.

### Claim 22 (depends on 1)

The method of claim 1, wherein the first number of simulated images is greater than the second number of measured images.

### Claim 23 (depends on 21)

The method of claim 21, wherein steps (a) and (b) are run through at least twice.

### Claim 24 (depends on 1)

A computer program product comprising instructions which, when executed by a computer system, prompt the computer system to carry out the method steps of claim 1.

### Claim 25

An apparatus for determining at least one unknown effect of defects of an element of a photolithography process, comprising:
a. means for providing a model of machine learning for a relationship between an image, design data associated with the image and at least one effect of the defects arising from the image;
b. means for training the model of machine learning using a multiplicity of images used for training purposes, design data associated with the images used for training purposes and the corresponding effects of the defects; and
c. means for determining the unknown effect of the defects by applying the trained model of machine learning to a measured image and the design data associated with the measured image.

### Claim 26 (depends on 25)

The apparatus of claim 25, wherein the apparatus comprises an exposure system for the element of the photolithography process and a magnifying lens that is embodied to image a portion of the element of the photolithography process on a photodetector.