# US 2019/0378012 A1 · Metrology Apparatus and Method for Determining a Characteristic of One or More Structures on a Substrate

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

- Number: US 2019/0378012 A1
- Document ID: US20190378012A1
- Status: Published application
- Kind: A1
- Application number: 16/424,811
- Filed: 2019-05-29
- Published: 2019-12-12
- Earliest claimed: 2018-06-08
- Assignees: ASML Netherlands B.V.
- Applicants: ASML Netherlands B.V.
- Inventors: Lorenzo Tripodi; Patrick Warnaar; Grzegorz Grzela; Mohammadreza Hajiahmadi; Farzad Farhadzadeh; Patricius Aloysius Jacobus Tinnemans; Scott Anderson Middlebrooks; Adrianus Cornelis Matheus Koopman; Frank Staals; Brennan Peterson; Anton Bernhard Van Oosten
- Web page: https://patentgrep.com/patent/US20190378012A1

## CPC

- G03F 7/70616: {Monitoring the printed patterns}
- G06N 3/08: Learning methods
- G01B 11/02: for measuring length, width or thickness (G01B11/08 takes precedence)
- G01N 21/55: Specular reflectivity
- G03F 7/70625: {Dimensions, e.g. line width, critical dimension [CD], profile, sidewall angle or edge roughness}
- G03F 7/70633: {Overlay, i.e. relative alignment between patterns printed by separate exposures in different layers, or in the same layer in multiple exposures or stitching}
- G03F 7/70641: {Focus}
- G03F 7/7065: {Defects, e.g. optical inspection of patterned layer for defects}
- G03F 7/70683: {Mark designs}
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G06N 3/096: Transfer learning
- G06T 7/0006: {using a design-rule based approach}
- G06T 7/001: {using an image reference approach}
- G01B 2210/56: Measuring geometric parameters of semiconductor structures, e.g. profile, critical dimensions or trench depth
- 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)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G01B: MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- 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)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Disclosed is a method of determining a characteristic of interest relating to a structure on a substrate formed by a lithographic process, the method comprising: obtaining an input image of the structure; and using a trained neural network to determine the characteristic of interest from said input image. Also disclosed is a reticle comprising a target forming feature comprising more than two sub-features each having different sensitivities to a characteristic of interest when imaged onto a substrate to form a corresponding target structure on said substrate. Related methods and apparatuses are also described.

## Claims

### Claim at position 1

1-15. (canceled)

### Claim 16

A method comprising:
obtaining an input image of a structure formed on a substrate formed by a lithographic process; and
using a trained neural network to determine a characteristic of interest relating to the structure based on the input image.

### Claim 17 (depends on 16)

The method of claim 16, comprising:
performing a calibration to train the neural network, the calibration comprising:
obtaining first training images of a calibration structure for different values of the characteristic of interest and corresponding known values of the characteristic of interest, the structure and the calibration structure comprising similar structures; and
using the first training images and corresponding known values of the characteristic of interest to train the neural network to infer values for the characteristic of interest from the first training images.

### Claim 18 (depends on 17)

The method of claim 17, wherein the known values for the characteristic of interest are obtained from second training images which are higher quality than the first training images, the second training images being of a corresponding calibration structure as the first training images.

### Claim 19 (depends on 18)

The method of claim 18, wherein the second training images comprise computationally enhanced training images and the computationally enhanced training images comprise phase retrieved images.

### Claim 20 (depends on 19)

The method of claim 19, wherein the second training images comprise computationally enhanced versions of the first training images.

### Claim 21 (depends on 18)

The method of claim 18, further comprising using a metrology apparatus to obtain the second training images through using a sensor having fewer and/or lesser aberrations than a corresponding sensor within a metrology apparatus used to obtain the first training images.

### Claim 22 (depends on 18)

The method of claim 18, further comprising:
outputting an enhanced output image based on the input image; and
wherein the trained neural network comprises an encoder-decoder neural network.

### Claim 23 (depends on 17)

The method of claim 17, wherein:
the neural network comprises an expanded neural network trained to handle a vector of multiple images corresponding to measurement of the calibration structure for different illumination characteristics of measurement illumination, and
the input image has been obtained using a plurality of illumination characteristics.

### Claim 24 (depends on 16)

The method of claim 16, further comprising performing transfer learning comprising training one or more of the final layer(s) of the neural network for one or more specific lithographic steps relevant to a particular process which is to be monitored.

### Claim 25 (depends on 16)

The method of claim 16, wherein the characteristic of interest is one or more of: overlay, critical dimension, any other reconstruction parameter, focus, dose, line edge roughness, line width roughness, any 2D contour parameter, and an illumination characteristic or aberration of the lithographic apparatus used in the lithographic process.

### Claim 26 (depends on 16)

The method of claim 16, wherein:
the structure comprises a multi-sensitivity target having more than two features with different sensitivities to the characteristic of interest;
the multi-sensitivity target comprises more than five or more than ten features with different sensitivities to the characteristic of interest.

### Claim 27 (depends on 26)

The method of claim 26, wherein:
the features with different sensitivities each comprise a periodic feature with variation in pitch and/or critical dimension; and
each periodic feature comprises first and second pitches, the second pitch being twice the first pitch.

### Claim 28

A reticle comprising:
a target forming feature comprising more than two sub-features each having different sensitivities to a characteristic of interest when imaged onto a substrate to form a corresponding target structure on the substrate.

### Claim 29

A metrology system for determining a characteristic of interest relating to at least one structure on a substrate formed by a lithographic process, the metrology apparatus comprising:
a trained neural network; and
a processor being configured to:
obtain an input image of the structure; and
use the trained neural network to determine the characteristic of interest from the input image.

### Claim 30

A non-transitory computer program product comprising machine-readable instructions for causing a processor to cause operations comprising:
obtaining an input image of a structure formed on a substrate formed by a lithographic process; and
using a trained neural network to determine a characteristic of interest relating to the structure based on the input image.