Search device, search program, and plasma processing apparatus

Patent metadata

Field	Value
Publication number	US-11747774-B2
Application number	US-201917272479-A
Country	US
Kind code	B2
Filing date	Dec 3, 2019
Priority date	Dec 3, 2019
Publication date	Sep 5, 2023
Grant date	Sep 5, 2023

Abstract

A parameter compression unit compresses first input parameter values so that a parameter restoration unit can restore the first input parameter values, and generates first compressed input parameter values in which the number of control parameters is reduced, a model learning unit learns a prediction model from learning data that is a set of the first compressed input parameter values and first output parameter values that are processing results obtained by giving the first input parameter values, as a plurality of control parameters, to a processing device, and a processing condition search unit estimates a second compressed input parameter values corresponding to target output parameter values by using the prediction model.

First claim

The invention claimed is:

1. A search device configured to search for an input parameter value to be given to a plurality of control parameters set in a processing apparatus, so that a processing result of a predetermined process performed by the processing apparatus satisfies a target output parameter value, the search device comprising:
a processor;
a memory;
a search program stored in the memory and configured to be executed by the processor to search for the input parameter value satisfying the target output parameter value; and
a parameter compression unit, a model learning unit, a processing condition search unit, a parameter restoration unit, and a convergence determination unit,
wherein the parameter compression unit compresses first input parameter values so that the parameter restoration unit is capable of restoring the first input parameter values, and generates first compressed input parameter values in which the number of control parameters is reduced,
wherein the model learning unit learns a prediction model from learning data that is a set of the first compressed input parameter values and first output parameter values that are processing results obtained by giving the first input parameter values, as the plurality of control parameters, to the processing device,
wherein the processing condition search unit estimates second compressed input parameter values corresponding to the target output parameter values by using the prediction model,
wherein the parameter restoration unit generates second input parameter values by adding control parameter values deleted by the parameter compression unit from the second compressed input parameter values, and
wherein the convergence determination unit determines whether second output parameter values, which are processing results obtained by giving the second input parameter values, as the plurality of control parameters, to the processing device, converges to a predetermined range of the target output parameter values.

2. The search device according to claim 1 , wherein
when the convergence determination unit determines that the second output parameter values does not converge to the predetermined range of the target output parameter values, the search program updates the prediction model by adding set of the second input parameter values and the second output parameter values to a set of the first input parameter values and the first output parameter values.

3. The search device according to claim 1 , wherein
the predetermined processing includes a plurality of steps in which values given to the plurality of control parameters are different from each other, and
the search program searches for values of the plurality of control parameters to be set in the plurality of steps as the input parameter values.

4. The search device according to claim 1 , wherein
the processing condition search unit estimates a plurality of the second compressed input parameter values by using the prediction model.

5. The search device according to claim 1 , wherein
the parameter compression unit deletes a part of the values of the control parameters of the first input parameter values such that the number of control parameters of the first input parameter values is equal to or less than the target number of parameters.

6. The search device according to claim 5 , wherein
the parameter compression unit deletes a value of a control parameter which is not used or is a fixed value among the first input parameter values, and stores the deleted value of the control parameter.

7. The search device according to claim 5 , wherein
the parameter compression unit deletes a value w of the second control parameter while leaving a value v of the first control parameter among the first input parameter values, and stores values of a coefficient a and an intercept b of w=av+b.

8. The search device according to claim 7 , wherein
the value w of the second control parameter is proportional to the value v of the first control parameter, and the value v of the first control parameter and the value w of the second control parameter have a correlation coefficient larger than a predetermined threshold value.

9. A plasma processing apparatus configured to perform plasma processing on a sample by using a plasma, the plasma processing apparatus comprising:
a processing chamber;
a plasma generation unit configured to generate the plasma inside the processing chamber; and
a control device configured to execute a search program to search for an input parameter value to be given to the plurality of control parameters set in the plasma processing apparatus satisfying a target output parameter value, and to control the plasma generation apparatus to perform the plasma processing on the sample placed in the processing chamber based on the searched input parameter value,
wherein the search program is configured to cause the control device to search for input parameter values to be given to the plurality of control parameters set in a processing device, so that a processing result of a predetermined process performed by the processing apparatus satisfies target output parameter values, by performing
a first step of compressing first input parameter values so that the first input parameter values can be restored, and generating first compressed input parameter values in which the number of control parameters is reduced;
a second step of learning a prediction model from learning data that is a set of the first compressed input parameter values and first output parameter values that are processing results obtained by giving the first input parameter values, as the plurality of control parameters, to the processing device;
a third step of estimating a second compressed input parameter values corresponding to the target output parameter values by using the prediction model;
a fourth step of generating second input parameter values by adding a control parameter values deleted by the first step from the second compressed input parameter values, and
a fifth step of determining whether second output parameter values, which are processing results obtained by giving the second input parameter values, as the plurality of control parameters, to the processing device, converges to a predetermined range of the target output parameter values.

10. The plasma processing apparatus according to claim 9 , wherein
in the fifth step, when it is determined that the second output parameter values does not converge to the predetermined range of the target output parameter values, the prediction model is updated by adding a set of the second input parameter values and the second output parameter values to a set of the first input parameter values and the first output parameter values.

11. The plasma processing apparatus according to claim 9 , wherein
in the first step, a part of the values of the control parameters of the first input parameter value is deleted such that the number of control parameters of the first input parameter value is equal to or less than the target number of parameters.

12. The plasma processing apparatus according to claim 9 , wherein
in the third step, a value of a part of the control parameter of the first compressed output parameter values is changed, and a value of another control parameter is fixed to a value of a control parameter of learning data that gives an optimum solution among the learning data to search for an approximate solution.

Assignees

• Hitachi High Tech Corp

Inventors

• Okuyama Yutaka
• Ohmori Takeshi

Classifications

• G05B13/0265 Primary

  the criterion being a learning criterion · CPC title

• G05B13/048 Primary

  using a predictor · CPC title

• G05B13/042

  in which a parameter or coefficient is automatically adjusted to optimise the performance · CPC title

• H05H1/24

  Generating plasma {(nuclear fusion reactors G21B1/00; gas-filled discharge reactors H01J37/32)} · CPC title

• G06N99/00

  Subject matter not provided for in other groups of this subclass · CPC title