# US 12,765,304 B2 · Etching control device, etching control method, and etching control system

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

- Number: US 12,765,304 B2
- Document ID: US12765304B2
- Status: Granted
- Kind: B2
- Application number: 18/469,673
- Filed: 2023-09-19
- Published: 2024-03-21
- Granted: 2026-10-13
- Earliest claimed: 2022-09-20
- Assignees: Tokyo Electron Limited
- Applicants: Tokyo Electron Limited
- Inventors: Toyohisa Tsuruda; Hiroshi Marumoto; Suguen Lee; Masashi Enomoto
- Web page: https://patentgrep.com/patent/US12765304B2

## CPC

- H10P 72/0424: {using mainly spraying means, e.g. nozzles}
- G06N 20/00: Machine learning
- H10P 72/0414: {using mainly spraying means, e.g. nozzles}
- H10P 72/0604: {Process monitoring, e.g. flow or thickness monitoring}
- H10P: GENERIC PROCESSES OR APPARATUS FOR THE MANUFACTURE OR TREATMENT OF DEVICES COVERED BY CLASS H10
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An etching control device includes an updating unit configured to update, to optimize a model indicating a relationship between distribution of an etching amount within a surface of a substrate and a process parameter, which is a parameter of controlling operations of multiple nozzles configured to etch the substrate, a parameter of the model; a calculator configured to calculate the process parameter corresponding to distribution of a designated etching amount by using the model whose parameter has been updated by the updating unit; and an operation controller configured to control the operations of the multiple nozzles by using the process parameter.

## Claims

### Claim 1

An etching control device, comprising:
a parameter updating unit configured to update, to optimize a model indicating a relationship between distribution of an etching amount within a surface of a substrate and a process parameter, which is a parameter of controlling operations of multiple nozzles configured to etch the substrate, a parameter of the model;
a calculator configured to calculate the process parameter corresponding to distribution of a designated etching amount by using the model whose parameter has been updated by the parameter updating unit; and
an operation controller configured to control the operations of the multiple nozzles by using the process parameter;
a recipe updating unit configured to update a process recipe defining operations of the multiple nozzles based on the process parameter calculated by the calculator;
wherein the multiple nozzles include a first nozzle configured to discharge a rinse on the substrate being rotated, and a second nozzle configured to discharge a chemical liquid to etch the substrate being rotated, and
where the operation controller is further configured to simultaneously control the discharge of the rinse from the first nozzle and the discharge of the chemical liquid from the second nozzle during an etching process,
wherein the operation controller is further configured to control the operation of the multiple nozzles according to the process recipe updated based on the process parameter,
wherein the model is a neural network, and
the parameter updating unit is further configured to update parameters of the neural network by using an error back-propagation method based on training data including an etching amount distribution and corresponding process parameters.

### Claim 2 (depends on 1)

The etching control device of claim 1,
wherein the parameter updating unit is further configured to update to optimize a model indicating a relationship between distribution of an etching amount and a process parameter, which is a parameter of determining a discharge time, a discharge position, and a moving speed of each of the first nozzle and the second nozzle, a parameter of the model.

### Claim 3

An etching control method performed by an etching control device, the etching control method comprising:
updating, to optimize a model indicating a relationship between distribution of an etching amount within a surface of a substrate and a process parameter, which is a parameter of controlling operations of multiple nozzles configured to etch the substrate, a parameter of the model;
calculating, by using the model whose parameter has been updated in the updating of the parameter of the model, the process parameter corresponding to distribution of a designated etching amount;
updating a process recipe defining operations of the multiple nozzles based on the calculated process parameter; and
controlling the operations of the multiple nozzles by using the process parameter,
wherein the multiple nozzles include a first nozzle configured to discharge a rinse on the substrate being rotated, and a second nozzle configured to discharge a chemical liquid to etch the substrate being rotated,
wherein, in the controlling of the operations of the multiple nozzles, the discharge of the first processing liquid from the first nozzle and the discharge of the second processing liquid from the second nozzle are simultaneously controlled during an etching process,
wherein the controlling of the operations of the multiple nozzles includes controlling the operations of the multiple nozzles according to the updated process recipe,
wherein the model is a neural network, and
the updating includes updating parameters of the neural network by using an error back-propagation method based on training data including an etching amount distribution and corresponding process parameters.

### Claim 4

An etching control system comprising:
a measuring device configured to measure distribution of an etching amount of a substrate; and
an etching control device,
wherein the etching control device comprises:
a parameter updating unit configured to update, to optimize a model indicating a relationship between the distribution of the etching amount within a surface of the substrate and a process parameter, which is a parameter of controlling operations of multiple nozzles configured to etch the substrate, a parameter of the model;
a calculator configured to calculate, by using the model whose parameter has been updated by the parameter updating unit, the process parameter corresponding to distribution of a designated etching amount measured by the measuring device; and
an operation controller configured to control the operations of the multiple nozzles by using the process parameter;
a recipe updating unit configured to update a process recipe defining operations of the multiple nozzles based on the process parameter calculated by the calculator;
wherein the multiple nozzles include a first nozzle configured to discharge a rinse on the substrate being rotated, and a second nozzle configured to discharge a chemical liquid to etch the substrate being rotated, and
wherein the operation controller is further configured to simultaneously control the discharge of the rinse from the first nozzle and the discharge of the chemical liquid from the second nozzle during an etching process,
wherein the operation controller is further configured to control the operations of the multiple nozzles according to the process recipe updated based on the process parameter,
wherein the model is a neural network, and
the parameter updating unit is further configured to update parameters of the neural network by using an error back-propagation method based on training data including an etching amount distribution and corresponding process parameters.