# US 12,222,690 B2 · Process recipe search apparatus, etching recipe search method and semiconductor device manufacturing system

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

- Number: US 12,222,690 B2
- Document ID: US12222690B2
- Status: Granted
- Kind: B2
- Application number: 17/370,131
- Filed: 2021-07-08
- Published: 2023-01-12
- Granted: 2025-02-11
- Earliest claimed: 2021-07-08
- Assignees: Hitachi High-Tech Corporation
- Applicants: Hitachi High-Tech Corporation
- Inventors: Takashi Dobashi; Hiroyuki Kobayashi; Takeshi Ohmori
- Web page: https://patentgrep.com/patent/US12222690B2

## CPC

- G05B 13/042: {in which a parameter or coefficient is automatically adjusted to optimise the performance}
- H01J 37/32926: {Software, data control or modelling}
- G06N 20/00: Machine learning
- G06T 7/0006: {using a design-rule based approach}
- G06T 11/00: Two-dimensional [2D] image generation
- H10P 50/242: {of Group IV materials}
- H10P 72/0604: {Process monitoring, e.g. flow or thickness monitoring}
- H10P 74/23: {characterised by multiple measurements, corrections, marking or sorting processes}
- G05B 2219/45031: Manufacturing semiconductor wafers
- G06T 2207/20081: Training; Learning
- 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}
- 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)
- H01J: ELECTRIC DISCHARGE TUBES OR DISCHARGE LAMPS (spark-gaps H01T; electric arc lamps H05B31/00; particle accelerators H05H)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- H10P: GENERIC PROCESSES OR APPARATUS FOR THE MANUFACTURE OR TREATMENT OF DEVICES COVERED BY CLASS H10

## Abstract

To facilitate evaluation of a predicted process shape in process recipe development using machine learning, a process recipe search apparatus that searches for an etching recipe that is a parameter of a plasma processing apparatus set so as to etch a sample into a desired shape displays, on a display device, the predicted process shape of the sample by a candidate etching recipe predicted by using a machine leaning model, by highlighting a difference between the predicted process shape and a target shape.

## Claims

### Claim 1

A plasma processing apparatus that searches for an etching recipe that is a parameter of the plasma processing apparatus set so as to etch a sample into a desired shape, the apparatus comprising:
a processor configured to
determine a target shape that defines the desired shape by a plurality of shape elements;
create a machine learning model that predicts a process shape of the sample processed by the plasma processing apparatus from the parameter of the plasma processing apparatus;
search, using the machine learning model, for a candidate etching recipe that becomes a candidate of the etching recipe;
output, on a display device, a predicted process shape of the sample by the candidate etching recipe predicted by using the machine learning model and decides, as a process recipe that is set to the plasma processing apparatus to cause the sample to be etched, the candidate etching recipe chosen from the displayed candidate etching recipes;
highlighting a difference between the predicted process shape and the target shape to display the predicted process shape on the display device;
output, on the display device, the predicted process shape of the sample predicted by using the machine learning model, by a sensitivity evaluation etching recipe in which at least one parameter included in the candidate etching recipe is changed;
superimpose each of the predicted process shape by the candidate etching recipe and the predicted process shape by the sensitivity evaluation etching recipe in a state where the difference from the target shape is highlighted; and
output the superimposed predicted process shapes on the display device; and
a computer configured to control etching of the sample by the plasma processing apparatus in accordance with the chosen etching recipe.

### Claim 2 (depends on 1)

The plasma processing apparatus according to claim 1, wherein the the processor is further configured to output the predicted process shape on the display device by enlarging the difference between the target shape and the predicted process shape at a highlight magnification set for each of the shape elements.

### Claim 3 (depends on 1)

The plasma processing apparatus according to claim 1,
wherein the processor is further configured to output, on the display device, the difference between the predicted process shape and the target shape so that the difference can be visually recognized on the predicted process shape, and
wherein the difference between the predicted process shape and the target shape is displayed by a color or a line width of the shape element in the predicted process shape or is displayed by overlapping, with the shape element, a vector that represents a deviation amount of the predicted process shape with respect to the target shape.

### Claim 4 (depends on 1)

The plasma processing apparatus according to claim 1, wherein the processor is further configured to output, on the display device, a range of the parameter that can be taken by the candidate etching recipe of the plasma processing apparatus, and when a constraint between the parameters of the plasma processing apparatus is designated, the candidate etching recipe is searched for in a constrained parameter space.

### Claim 5

An etching recipe search method that uses a process recipe search apparatus that searches for an etching recipe that is a parameter of a plasma processing apparatus set so as to etch a sample into a desired shape, the method comprising the steps of:
deciding a target shape that defines the desired shape by a plurality of shape elements;
creating a machine learning model that predicts a process shape of the sample processed by the plasma processing apparatus from the parameter of the plasma processing apparatus;
using the machine learning model to search for a candidate etching recipe that becomes a candidate of the etching recipe;
displaying, on a display device, a predicted process shape of the sample by the candidate etching recipe predicted by using the machine learning model;
highlighting a difference between the predicted process shape and the target shape to display the predicted process shape on the display device; and
etching the sample by the plasma processing apparatus in accordance with the candidate etching recipe,
wherein the machine learning model is used to predict the process shape of the sample by a sensitivity evaluation etching recipe in which at least one parameter included in the candidate etching recipe is changed, and
wherein each of the predicted process shape by the candidate etching recipe and the predicted process shape by the sensitivity evaluation etching recipe is superimposed in a state where the difference from the target shape is highlighted, and is displayed on the display device.

### Claim 6 (depends on 5)

The etching recipe search method according to claim 5, wherein the predicted process shape is displayed on the display device by enlarging the difference between the target shape and the predicted process shape at a highlight magnification set for each of the shape elements.

### Claim 7 (depends on 5)

The etching recipe search method according to claim 5, wherein a range of the parameter that can be taken by the candidate etching recipe of the plasma processing apparatus is displayed on the display device, and when a constraint between the parameters of the plasma processing apparatus is designated, the candidate etching recipe is searched for in a constrained parameter space.

### Claim 8

An etching recipe search method that uses a process recipe search apparatus that searches for an etching recipe that is a parameter of a plasma processing apparatus set so as to etch a sample into a desired shape, the method comprising the steps of:
deciding a target shape that defines the desired shape by a plurality of shape elements;
creating a machine learning model that predicts a process shape of the sample processed by the plasma processing apparatus from the parameter of the plasma processing apparatus;
using the machine learning model to search for a candidate etching recipe that becomes a candidate of the etching recipe;
displaying, on a display device, a predicted process shape of the sample by the candidate etching recipe predicted by using the machine learning model;
highlighting a difference between the predicted process shape and the target shape to display the predicted process shape on the display device; and
etching the sample by the plasma processing apparatus in accordance with the candidate etching recipe,
wherein a plurality of candidate etching recipes chosen from the candidate etching recipes displayed on the display device are decided as process recipes that are set to the plasma processing apparatus to cause the samples to be etched,
wherein when the decided process recipes are set to the plasma processing apparatus and among the process shapes acquired by etching the samples, the process shape by a first process recipe is a reference image, and the process shape by a second process recipe is a comparison image, the comparison image is displayed on the display device by enlarging a difference between the comparison image and the reference image at a highlight magnification set for each of the shape elements,
wherein when the decided process recipe is set to the plasma processing apparatus and the process shape acquired by etching the sample does not coincide with the target shape, the target shape is corrected,
wherein the machine learning model is updated, and
wherein the updated machine learning model is used to search for the candidate etching recipe that becomes the candidate of the etching recipe.

### Claim 9

A semiconductor device manufacturing system that includes a terminal, a plasma processing apparatus, and a platform connected to the terminal and the plasma processing apparatus via a network and on which a processor configured to execute a process recipe search application that searches for an etching recipe that is a parameter of the plasma processing apparatus set so as to etch a sample into a desired shape is implemented,
wherein the process recipe search application, when executed by the processor, causes the platform to perform:
a step of deciding a target shape that defines the desired shape by a plurality of shape elements;
a step of creating a machine learning model that predicts a process shape of the sample processed by the plasma processing apparatus from the parameter of the plasma processing apparatus;
a step of using the machine learning model to search for a candidate etching recipe that becomes the candidate of the etching recipe;
a step of displaying, on the terminal, a predicted process shape of the sample by the candidate etching recipe predicted by using the machine learning model;
a step of using the machine learning model to predict the process shape of the sample by a sensitivity evaluation etching recipe in which at least one parameter included in the candidate etching recipe is changed;
a step of superimposing each of the predicted process shape by the candidate etching recipe and the predicted process shape by the sensitivity evaluation etching recipe in a state where the difference from the target shape is highlighted and displaying the superimposed predicted process shapes on the terminal; and
a step of causing the plasma processing apparatus to etch the sample in accordance with the candidate etching recipe,
wherein for the predicted process shape displayed on the terminal, a difference between the predicted process shape and the target shape is highlighted.

### Claim 10 (depends on 9)

The semiconductor device manufacturing system according to claim 9, wherein the predicted process shape is displayed on the terminal by enlarging the difference between the target shape and the predicted process shape at a highlight magnification set for each of the shape elements.

### Claim 11 (depends on 9)

The semiconductor device manufacturing system according to claim 9,
wherein the process recipe search application executes a step of displaying, on the terminal, a range of the parameter that can be taken by the candidate etching recipe of the plasma processing apparatus, and
wherein when a constraint between the parameters of the plasma processing apparatus is designated from the terminal, the candidate etching recipe is searched for in a constrained parameter space in the step of searching for the candidate etching recipe.

### Claim 12 (depends on 9)

The semiconductor device manufacturing system according to claim 9, wherein the process recipe search application executes:
a step of deciding, as a process recipe that is set to the plasma processing apparatus to cause the sample to be etched, the candidate etching recipe chosen from the candidate etching recipes displayed on the terminal; and
a step of setting the process recipe to the plasma processing apparatus.