Learning processor, learning processing method, production method of compound semiconductor, and recording medium

1. A learning processor comprising:

a control condition acquiring section configured to acquire control condition data indicating a control condition of a film deposition apparatus;

a film characteristic acquiring section configured to acquire film characteristic data indicating a characteristic of a film deposited by the film deposition apparatus that has been caused to operate according to the control condition indicated by the control condition data; and

a learning processing section configured to perform learning processing of a model that outputs recommended control condition data indicating the control condition of the film deposition apparatus that is recommended in response to input of target film characteristic data indicating a target film characteristic, using learning data that includes the acquired control condition data and film characteristic data.

2. The learning processor according to claim 1 , wherein

the film deposition apparatus is a molecular beam epitaxy apparatus, and

the control condition data includes data concerning at least one of a flux amount of raw material to be radiated to a substrate, a temperature of a cell of the molecular beam epitaxy apparatus, power supplied to a heater of the cell, an opening/closing condition of a shutter of the cell, a temperature of the substrate, power supplied to a substrate heater, vacuum degree of a chamber, type of gas present inside the chamber, amount of the gas, a temperature of a cryopanel inside the chamber, an amount of liquid nitrogen supplied to the cryopanel, time needed for film deposition, and gain in feedback control of the temperature of the cell.

3. The learning processor according to claim 1 , further comprising:

a state acquiring section configured to acquire state data indicating a state of the film deposition apparatus, wherein

the learning processing section performs the learning processing of the model that outputs the recommended control condition data in response to the input of the target film characteristic data and the state data, using the learning data that further includes the acquired state data.

4. The learning processor according to claim 3 , wherein

the state data includes operating history data indicating an operating history of the film deposition apparatus.

5. The learning processor according to claim 4 , wherein

the operating history data includes at least one of data concerning at least one of the number of times maintenance of the film deposition apparatus has been performed and content of the maintenance, data concerning the number of times that at least one component of the film deposition apparatus has been used, data concerning the number of times the film deposition apparatus has performed film deposition, and data concerning a film formed in the past.

6. The learning processor according to claim 1 , wherein

the film characteristic data includes data concerning at least one of film thickness, composition, flatness, an electrical characteristic, an optical characteristic, crystallinity, surface information, and dislocation density of the deposited film, and

the surface information includes information obtained from an optical micrograph of the deposited film.

7. The learning processor according to claim 1 , wherein

the film deposition apparatus deposits the film including at least one of Al, Ga, In, As, Sb, Si, Te, Sn, Zn, and Be.

8. The learning processor according to claim 1 , further comprising:

a target film characteristic acquiring section configured to acquire the target film characteristic data indicating the target film characteristic;

a target film characteristic supplying section configured to supply the target film characteristic data to the model;

a recommended control condition acquiring section configured to acquire the recommended control condition data output by the model in response to the target film characteristic data being supplied to the model; and

a control section configured to cause the film deposition apparatus to operate according to the control condition indicated by the recommended control condition data.

9. A learning processing method comprising:

acquiring control condition data indicating a control condition of a film deposition apparatus;

acquiring film characteristic data indicating a characteristic of a film deposited by the film deposition apparatus that has been caused to operate according to the control condition indicated by the control condition data; and

performing learning processing of a model that outputs recommended control condition data indicating the control condition of the film deposition apparatus that is recommended in response to input of target film characteristic data indicating a target film characteristic, using learning data that includes the acquired control condition data and film characteristic data.

10. The learning processing method according to claim 9 , wherein

the film deposition apparatus is a molecular beam epitaxy apparatus, and

the control condition data includes data concerning at least one of a flux amount of raw material to be radiated to a substrate, a temperature of a cell of the molecular beam epitaxy apparatus, power supplied to a heater of the cell, an opening/closing condition of a shutter of the cell, a temperature of the substrate, power supplied to a substrate heater, vacuum degree of a chamber, type of gas present inside the chamber, amount of the gas, a temperature of a cryopanel inside the chamber, an amount of liquid nitrogen supplied to the cryopanel, time needed for film deposition, and gain in feedback control of the temperature of the cell.

11. The learning processing method according to claim 9 , further comprising:

acquiring the target film characteristic data indicating the target film characteristic;

supplying the target film characteristic data to the model;

acquiring the recommended control condition data output by the model in response to the target film characteristic data being supplied to the model; and

causing the film deposition apparatus to operate according to the control condition indicated by the recommended control condition data.

12. A compound semiconductor production method comprising:

preparing a substrate; and

stacking a plurality of films to be included in a compound semiconductor on the substrate, wherein

the stacking includes causing the film deposition apparatus to operate using the learning processing method according to claim 11 , to deposit at least one of the plurality of films.

13. A non-transitory computer readable storage medium storing thereon a program that, when executed by a computer, causes the computer to perform operations including:

acquiring control condition data indicating a control condition of a film deposition apparatus;

acquiring film characteristic data indicating a characteristic of a film deposited by the film deposition apparatus that has been caused to operate according to the control condition indicated by the control condition data; and

performing learning processing of a model that outputs recommended control condition data indicating the control condition of the film deposition apparatus that is recommended in response to input of target film characteristic data indicating a target film characteristic, using learning data that includes the acquired control condition data and film characteristic data.