# US 2018/0293721 A1 · Contour based defect detection

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

- Number: US 2018/0293721 A1
- Document ID: US20180293721A1
- Status: Published application
- Kind: A1
- Application number: 15/896,060
- Filed: 2018-02-14
- Published: 2018-10-11
- Earliest claimed: 2017-04-07
- Assignees: KLA-Tencor Corporation
- Applicants: KLA-Tencor Corporation
- Inventors: Ajay Gupta; Mohan Mahadevan; Sankar Venkataraman; Hedong Yang; Laurent Karsenti; Yair Carmon; Noga Bullkich; Udy Danino
- Web page: https://patentgrep.com/patent/US20180293721A1

## CPC

- G06T 7/001: {using an image reference approach}
- G01N 21/8851: {Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges (G01N21/8806 and G01N21/93 - G01N21/95692 take precedence; optical measurement of dimensions G01B11/00; optical scanning G02B26/10; image transformation G06T3/00; computerised image enhancement G06T5/00; image processing per se for flaw detection G06T7/0002)}
- G01N 21/956: Inspecting patterns on the surface of objects {(contactless testing of electronic circuits G01R31/308; testing currency G07D; manufacturing processes per se of semiconductor devices implementing a measuring step H10P74/20)}
- G01N 21/95607: {using a comparative method}
- G05B 19/41875: {characterised by quality surveillance of production}
- G06N 3/045: Combinations of networks
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/0475: Generative networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 3/094: Adversarial learning
- G06N 20/00: Machine learning
- G06T 7/564: from contours
- G01N 21/9501: {Semiconductor wafers (manufacturing processes per se of semiconductor devices implementing a measuring step H10P74/20)}
- G01N 2021/8867: {using sequentially two or more inspection runs, e.g. coarse and fine, or detecting then analysing}
- G01N 2021/8887: {based on image processing techniques}
- G06T 2207/10061: from scanning electron microscope
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30148: Semiconductor; IC; Wafer
- H10P 74/203: {Structural properties, e.g. testing or measuring thicknesses, line widths, warpage, bond strengths or physical defects}
- H10P 74/23: {characterised by multiple measurements, corrections, marking or sorting processes}
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H10P: GENERIC PROCESSES OR APPARATUS FOR THE MANUFACTURE OR TREATMENT OF DEVICES COVERED BY CLASS H10

## Abstract

Methods and systems for detecting defects in patterns formed on a specimen are provided. One system includes one or more components executed by one or more computer subsystems, and the component(s) include first and second learning based models. The first learning based model generates simulated contours for the patterns based on a design for the specimen, and the simulated contours are expected contours of a defect free version of the patterns in images of the specimen generated by an imaging subsystem. The second learning based model is configured for generating actual contours for the patterns in at least one acquired image of the patterns formed on the specimen. The computer subsystem(s) are configured for comparing the actual contours to the simulated contours and detecting defects in the patterns formed on the specimen based on results of the comparing.

## Claims

### Claim 1

A system configured to detect defects in patterns formed on a specimen, comprising:
an imaging subsystem comprising at least an energy source and a detector, wherein the energy source is configured to generate energy that is directed to a specimen, and wherein the detector is configured to detect energy from the specimen and to generate images responsive to the detected energy; and
one or more computer subsystems configured for acquiring the images of patterns formed on the specimen; and
one or more components executed by the one or more computer subsystems, wherein the one or more components comprise a first learning based model and a second learning based model, wherein the first learning based model is configured for generating simulated contours for the patterns based on a design for the specimen input to the first learning based model by the one or more computer subsystems, wherein the simulated contours are expected contours of a defect free version of the patterns in the images of the specimen generated by the imaging subsystem, and wherein the second learning based model is configured for generating actual contours for the patterns in at least one of the acquired images of the patterns formed on the specimen input to the second learning based model by the one or more computer subsystems; and
wherein the one or more computer subsystems are further configured for:
comparing the actual contours to the simulated contours; and
detecting defects in the patterns formed on the specimen based on results of the comparing.

### Claim 2 (depends on 1)

The system of claim 1, wherein the results of the comparing comprise quantitative differences between dimensions of a first of the patterns in the design and the first of the patterns in the at least one of the acquired images of the patterns formed on the specimen determined based on differences between the actual and simulated contours, and wherein detecting the defects comprises applying a threshold to the quantitative differences between the dimensions.

### Claim 3 (depends on 1)

The system of claim 1, wherein the results of the comparing comprise quantitative differences between the actual and simulated contours for each of the pixels of each of the patterns in the at least one of the acquired images.

### Claim 4 (depends on 1)

The system of claim 1, wherein the one or more computer subsystems are further configured for detecting hot spots in the design based on the detected defects.

### Claim 5 (depends on 1)

The system of claim 1, wherein the design input to the first learning based model does not include features of the design that will not be printed on the specimen.

### Claim 6 (depends on 1)

The system of claim 1, wherein the first and second learning based models are adaptable to different pattern types.

### Claim 7 (depends on 1)

The system of claim 1, wherein the first and second learning based models are adaptable to different patterns densities.

### Claim 8 (depends on 1)

The system of claim 1, wherein the first and second learning based models are adaptable to patterns in different layer types.

### Claim 9 (depends on 1)

The system of claim 1, wherein the first and second learning based models are adaptable to images generated by the imaging subsystem with one or more different imaging parameters.

### Claim 10 (depends on 1)

The system of claim 1, wherein the one or more computer subsystems are further configured for training the first learning based model using a training data set comprising different portions of at least one training design for at least one training specimen and corresponding contour information extracted from training images of the at least one training specimen with a ground truth method.

### Claim 11 (depends on 1)

The system of claim 1, wherein the one or more computer subsystems are further configured for training the second learning based model using a training data set comprising different training images generated by the imaging subsystem for at least one training specimen on which at least one training design was formed and corresponding contour information extracted from the different training images with a ground truth method.

### Claim 12 (depends on 1)

The system of claim 1, wherein the first and second learning based models are deep learning based models.

### Claim 13 (depends on 1)

The system of claim 1, wherein the second learning based model has a VGG network architecture.

### Claim 14 (depends on 1)

The system of claim 1, wherein the second learning based model is further configured as a holistically-nested edge detection model.

### Claim 15 (depends on 1)

The system of claim 1, wherein the first learning based model is further configured as a convolutional neural network.

### Claim 16 (depends on 1)

The system of claim 1, wherein the first learning based model is further configured as a deep generative model using variational auto-encoders.

### Claim 17 (depends on 1)

The system of claim 1, wherein the one or more computer subsystems are further configured for performing an additional comparing step in which the actual contours for the same patterns formed in different dies on the specimen are compared to each other and detecting defects in the same patterns based on results of the additional comparing step.

### Claim 18 (depends on 1)

The system of claim 1, wherein the one or more computer subsystems are further configured for classifying the detected defects.

### Claim 19 (depends on 1)

The system of claim 1, wherein the one or more computer subsystems are further configured for classifying the detected defects using an additional learning based model.

### Claim 20 (depends on 1)

The system of claim 1, wherein the system is further configured as a metrology tool.

### Claim 21 (depends on 1)

The system of claim 1, wherein the system is further configured as an inspection tool.

### Claim 22 (depends on 1)

The system of claim 1, wherein the system is further configured as a defect review tool.

### Claim 23 (depends on 1)

The system of claim 1, wherein the specimen comprises a wafer,

### Claim 24 (depends on 1)

The system of claim 1, wherein the specimen comprises a reticle.

### Claim 25 (depends on 1)

The system of claim 1, wherein the energy directed to the specimen comprises light, and wherein the energy detected from the specimen comprises light.

### Claim 26 (depends on 1)

The system of claim 1, wherein the energy directed to the specimen comprises electrons, and wherein the energy detected from the specimen comprises electrons.

### Claim 27

A non-transitory computer-readable medium, storing program instructions executable on one or more computer subsystems for performing a computer-implemented method for detecting defects in patterns formed on a specimen, wherein the computer-implemented method comprises:
acquiring images of patterns formed on a specimen by an imaging subsystem with the one or more computer subsystems, wherein the imaging subsystem comprises at least an energy source and a detector, wherein the energy source is configured to generate energy that is directed to the specimen, and wherein the detector is configured to detect energy from the specimen and to generate images responsive to the detected energy;
generating simulated contours for the patterns based on a design for the specimen input to a first learning based model by the one or more computer subsystems, wherein the simulated contours are expected contours of a defect free version of the patterns in the images of the specimen generated by the imaging subsystem;
generating actual contours for the patterns in at least one of the acquired images of the patterns formed on the specimen input to a second learning based model by the one or more computer subsystems, wherein one or more components are executed by the one or more computer subsystems, and wherein the one or more components comprise the first and second learning based models;
comparing the actual contours to the simulated contours; and
detecting defects in the patterns formed on the specimen based on results of the comparing.

### Claim 28

A computer-implemented method for detecting defects in patterns formed on a specimen, comprising:
acquiring images of patterns formed on a specimen by an imaging subsystem with one or more computer subsystems, wherein the imaging subsystem comprises at least an energy source and a detector, wherein the energy source is configured to generate energy that is directed to the specimen, and wherein the detector is configured to detect energy from the specimen and to generate images responsive to the detected energy;
generating simulated contours for the patterns based on a design for the specimen input to a first learning based model by the one or more computer subsystems, wherein the simulated contours are expected contours of a defect free version of the patterns in the images of the specimen generated by the imaging subsystem;
generating actual contours for the patterns in at least one of the acquired images of the patterns formed on the specimen input to a second learning based model by the one or more computer subsystems, wherein one or more components are executed by the one or more computer subsystems, and wherein the one or more components comprise the first and second learning based models;
comparing the actual contours to the simulated contours; and
detecting defects in the patterns formed on the specimen based on results of the comparing.
