CN112997274A - Substrate processing condition setting assisting method, substrate processing system, storage medium, and learning model - Google Patents

Abstract

The invention provides a method for assisting condition setting of substrate processing, which comprises the following steps: a step of inputting a data set to a machine learning apparatus, wherein the data set includes process conditions of a substrate process and actual data on quality of the substrate process, the substrate process being performed by the substrate processing apparatus and including a process of supplying a process liquid to a substrate; and a step of deriving a recommended processing condition for the substrate processing based on a learning model, wherein the learning model is a model generated by machine learning by the machine learning device based on the plurality of sets of data, and prediction data on the quality of the substrate processing can be output in response to an input of the processing condition.

Description

Substrate processing condition setting assisting method, substrate processing system, storage medium, and learning model

Technical Field

The invention relates to a condition setting assisting method for substrate processing, a substrate processing system, a storage medium, and a learning model.

Background

Patent document 1 discloses an apparatus for forming a photosensitive coating film on a surface of a substrate, exposing the photosensitive coating film, and then developing the photosensitive coating film.

Documents of the prior art

Patent document 1: japanese patent laid-open publication No. 2017-73522

Disclosure of Invention

Technical problem to be solved by the invention

The invention provides a condition setting assisting method which can effectively simplify the operation of setting the processing conditions of substrate processing.

Means for solving the problems

A method for assisting condition setting in substrate processing according to an aspect of the present invention includes: a step of inputting a data set to a machine learning apparatus, wherein the data set includes process conditions of a substrate process and actual data on quality of the substrate process, the substrate process being performed by the substrate processing apparatus and including a process of supplying a process liquid to a substrate; and a step of deriving a recommended processing condition for the substrate processing based on a learning model, wherein the learning model is a model generated by machine learning by a machine learning device based on a plurality of sets of the data sets, and prediction data on the quality of the substrate processing can be output in response to an input of the processing condition.

Effects of the invention

The invention can provide a condition setting assisting method which can effectively simplify the operation of setting the processing condition of the substrate processing.

Drawings

Fig. 1 is a schematic view showing the structure of a substrate processing system of an exemplary embodiment.

Fig. 2 is a schematic view showing a schematic configuration of the coating unit by way of example.

Fig. 3 is a schematic diagram illustrating a schematic configuration of the developing unit.

Fig. 4 is a schematic diagram showing a schematic configuration of the inspection apparatus after the treatment.

Fig. 5 is a schematic diagram showing a schematic configuration of an inspection apparatus in processing, by way of example.

Fig. 6 is a block diagram illustrating a functional configuration of the control device and the machine learning device.

Fig. 7 is a block diagram illustrating a hardware configuration of the control device and the machine learning device.

Fig. 8 is a flowchart illustrating a condition setting support procedure executed by the control device.

Fig. 9 is a flowchart illustrating a condition setting support procedure further executed by the control device.

Fig. 10 is a flowchart illustrating a condition setting support procedure executed by the machine learning apparatus.

Fig. 11 is a flowchart illustrating a condition setting support procedure further executed by the machine learning apparatus.

Detailed Description

Various exemplary embodiments are described below. In the description, the same elements or elements having the same function are denoted by the same reference numerals, and redundant description thereof is omitted.

[ substrate processing System ]

The substrate processing system 1 is a system that forms a photosensitive coating film on the surface of a substrate and performs a developing process on the photosensitive coating film after the exposure process. The substrate to be processed is, for example, a semiconductor wafer W. The photosensitive coating film is, for example, a resist film.

As shown in fig. 1, a substrate processing system 1 includes a coating and developing apparatus 2 and a control apparatus 100. The coating and developing apparatus 2 includes a carrier block 4, a process block 5, and an interface block 6.

The carrier block 4 is used for introducing the wafer W (substrate) into the coating and developing apparatus 2 and for taking the wafer W out of the coating and developing apparatus 2. For example, the carrier block 4 has a transfer arm a1 built therein, which can support a plurality of carriers C for wafers W. The carrier C receives a plurality of circular wafers W, for example. The transfer arm a1 takes an unprocessed wafer W out of the carrier C and returns a processed wafer W to the carrier C.

The processing block 5 comprises a plurality of processing modules 11, 12, 13, 14. The process modules 11, 12, and 13 (process units) perform a film formation process, and apply a deposition solution (a processing solution for film formation) to the surface Wa of the wafer W to form a film. For example, the process modules 11, 12, and 13 include a coating unit U1, a heat treatment unit U2, and a transfer arm A3, and the transfer arm A3 transfers the wafer W to the units.

The process module 11 forms an underlying film on the surface of the wafer W using the coating unit U1 and the heat treatment unit U2. The coating unit U1 of the process module 11 coats the wafer W with the treatment liquid for forming the lower layer film. The heat treatment unit U2 of the process module 11 performs various heat treatments required in the formation of the underlying film.

The process module 12 forms a resist film on the lower layer film using the coating unit U1 and the heat treatment unit U2. The coating unit U1 of the process module 12 coats the treatment liquid for forming the resist film on the underlayer film. The heat treatment unit U2 of the process module 12 performs various heat treatments required in the formation of the resist film.

The process module 13 forms an upper layer film on the resist film using the coating unit U1 and the heat treatment unit U2. The coating unit U1 of the process module 13 coats the resist film with the liquid for forming the upper layer film. The heat treatment unit U2 of the process module 13 performs various heat treatments required in the formation of the upper layer film.

As shown in FIG. 2, the coating unit U1 includes a spin holding section 50 and a deposition-solution supplying section 60. The rotation holding portion 50 holds and rotates the wafer W. For example, the rotation holding portion 50 includes a holding portion 51 and a rotation driving portion 52. The holding portion 51 is used to support the horizontally arranged wafer W and holds the wafer W by, for example, vacuum suction. The rotation driving unit 52 rotates the holding unit 51 about a vertical axis using, for example, a motor or the like as a power source. Thereby, the wafer W held by the holding portion 51 also rotates.

The deposition-solution supplying section 60 supplies a deposition solution to the surface Wa of the wafer W held by the holding section 51. For example, the deposition-solution supplying section 60 includes a nozzle 61 and a liquid source 62. The nozzle 61 is disposed above the wafer W held by the holding portion 51, and discharges the processing liquid downward. The liquid source 62 pumps the treatment liquid to the nozzle 61.

Returning to fig. 1, the process module 14 (process unit) performs a developing process, and supplies a developing process liquid to the resist film (photosensitive film) subjected to the exposure process on the surface Wa of the wafer W. For example, the process module 14 incorporates a developing unit U3, a heat treatment unit U4, and a transfer arm A3 for transferring the wafer W to the units. The process module 14 performs a developing process of the exposed resist film using the developing unit U3 and the heat treatment unit U4. The developing unit U3 applies a developing solution (a processing solution for development) to the surface of the exposed wafer W, and then washes the surface with a rinse solution (a processing solution for rinse), thereby performing a developing process of the resist film. The heat treatment unit U4 performs various heat treatments required during the developing process. Specific examples of the heat treatment include heat treatment before development treatment (PEB), heat treatment after development treatment (PB: Post Bake), and the like.

As shown in fig. 3, the developing unit U3 includes a rotary holding portion 20, a developer supply portion 30, and a rinse liquid supply portion 40. The rotation holding unit 20 holds and rotates the wafer W. For example, the rotary holding portion 20 includes a holding portion 21 and a rotary driving portion 22. The holding portion 21 is used to support the horizontally arranged wafer W, and holds the wafer W by, for example, vacuum suction. The rotation driving unit 22 rotates the holding unit 21 about a vertical axis line using, for example, a motor or the like as a power source. Thereby, the wafer W held by the holding portion 21 also rotates.

The developer supply unit 30 supplies the developer to the surface Wa of the wafer W held by the holding unit 21. For example, the developer supply unit 30 includes a nozzle 31, a nozzle transfer unit 32, and a liquid source 33. The nozzle 31 is disposed above the wafer W held by the holding portion 21, and discharges the developer downward. The nozzle transfer unit 32 moves the nozzle 31 in the horizontal direction using a motor or the like as a power source. The liquid source 33 pumps the developer to the nozzle 31.

The rinse liquid supply unit 40 supplies a rinse liquid to the front surface Wa of the wafer W held by the holding unit 21. For example, the rinse liquid supply unit 40 includes a nozzle 41, a nozzle transfer unit 42, and a liquid source 43. The nozzle 41 is disposed above the wafer W held by the holding portion 21, and discharges the rinse liquid downward. The nozzle transfer unit 42 uses a motor or the like as a power source to move the nozzle 41 in the horizontal direction. The liquid source 43 pumps the rinse liquid to the nozzle 41.

Returning to fig. 1, the interface block 6 transfers the wafer W to and from an exposure apparatus (not shown) for performing an exposure process on a resist film formed on the wafer W. For example, the interface block 6 has a built-in interface arm A8 and is connected to the exposure apparatus. The transfer arm A8 transfers the wafer W before exposure processing to the exposure device, and receives the wafer W after exposure processing from the exposure device.

A receiving portion U10 is provided between the processing block 5 and the carrier block 4. The housing unit U10 is divided into a plurality of cells (cells) arranged in the vertical direction, and wafers W can be housed in each cell. The receiving portion U10 is used for transferring the wafers W between the carrier block 4 and the processing block 5. A lift arm a7 is provided near the storage unit U10. The lift arm a7 lifts and lowers the wafer W between the cells of the storage unit U10. A housing portion U11 is provided between the processing block 5 and the interface block 6. The housing unit U11 is also divided into a plurality of cells arranged in the vertical direction, and the wafers W can be housed in each cell. The receiving unit U11 is used for transferring the wafers W between the processing block 5 and the interface block 6.

The control device 100 controls the coating and developing device 2, for example, so that the coating and developing process is performed in the following order. First, the controller 100 controls the transfer arm a1 to transfer the wafer W in the carrier C to the storage unit U10, and controls the lift arm a7 to place the wafer W in the chamber for the process module 11.

Next, the controller 100 controls the transfer arm a3 to transfer the wafers W in the storage unit U10 to the coating unit U1 and the heat treatment unit U2 in the process module 11. The controller 100 controls the coating unit U1 and the heat treatment unit U2 to form an underlayer film on the surface of the wafer W. Thereafter, the controller 100 controls the transfer arm A3 to return the wafer W with the underlying film formed thereon to the storage unit U10, and controls the lift arm a7 to place the wafer W in the chamber for the process module 12.

Then, the controller 100 controls the transfer arm a3 to transfer the wafers W in the storage unit U10 to the coating unit U1 and the heat treatment unit U2 in the process module 12. The controller 100 controls the coating unit U1 and the heat treatment unit U2 to form a resist film on the lower layer film of the wafer W. Thereafter, the controller 100 controls the transfer arm A3 to return the wafer W to the storage unit U10, and controls the lift arm a7 to place the wafer W in the chamber for the process module 13.

Next, the controller 100 controls the transfer arm a3 to transfer the wafers W in the storage unit U10 to the respective units in the process module 13. The controller 100 controls the coating unit U1 and the heat treatment unit U2 to form an upper layer film on the resist film on the wafer W. Thereafter, the controller 100 controls the transfer arm a3 to transfer the wafer W to the storage unit U11.

Then, the controller 100 controls the transfer arm A8 to send out the wafer W in the storage unit U11 to the exposure apparatus 3. Thereafter, the controller 100 controls the transfer arm A8 to receive the wafer W subjected to the exposure process from the exposure apparatus 3 and dispose the wafer W in the chamber for the process module 14 in the housing unit U11.

Next, the controller 100 controls the transfer arm a3 to transfer the wafer W in the storage unit U11 to each unit in the process module 14, and controls the developing unit U3 and the heat treatment unit U4 to perform a developing process on the resist film of the wafer W. Thereafter, the controller 100 controls the transfer arm A3 to return the wafer W to the storage unit U10, and controls the lift arm a7 and the transfer arm a1 to return the wafer W to the carrier C. In the above manner, the coating and developing process is completed.

The specific structure of the substrate processing system is not limited to the above-illustrated structure. The substrate processing system may have any configuration as long as it includes a processing unit that performs a substrate process including a process of supplying a processing liquid or the like to a substrate, and a control device 100 that can control the processing unit.

(Condition setting Assist System)

The substrate processing system 1 further comprises a condition setting system 7. The condition setting system 7 includes a quality inspection device 70. At least a part of the condition setting system 7 is constituted by the control device 100. That is, the condition setting system 7 includes a quality inspection device 70 and a control device 100. The quality inspection device 70 detects information on the quality of the substrate processing performed by the coating and developing device 2.

The control device 100 executes the following steps: a step of causing the coating and developing apparatus 2 (substrate processing apparatus) to perform a substrate process including supplying a processing liquid to the wafer W in accordance with a predetermined process condition; a step of acquiring actual data (i.e., actual result data) regarding the quality of substrate processing performed in accordance with the processing conditions from the quality inspection device 70; a step of inputting a data set including a process condition of a substrate process and actual data of the substrate process to the machine learning device 200; and a step of deriving a recommended processing condition for the substrate processing based on a learning model generated by machine learning by the machine learning device 200 based on the plurality of sets of data sets, which is capable of outputting prediction data on the quality of the substrate processing in response to an input of the processing condition. The prediction data is, for example, data for predicting the actual data. The actual data may be any data as long as it relates to the quality of the substrate processing. The data of the quality of the substrate after substrate processing is related to the quality of the substrate processing. The supply state of the processing liquid during the substrate processing is also related to the quality of the substrate processing.

The condition setting system 7 may further include a machine learning device 200. The machine learning device 200 executes the following steps: a step of acquiring the data set; and performing machine learning based on the plurality of sets of data to generate the learning model. The machine learning device 200 may be housed in the same housing as the control device 100, or may be provided at a position remote from the control device 100. When installed at a location remote from the control device 100, the machine learning device 200 is connected to the control device 100 via a local area network, for example. The machine learning apparatus 200 may be connected to the control apparatus 100 via a wide area network such as the internet. The structure of each part is described in detail below.

(quality data detecting device)

The quality inspection apparatus 70 includes, for example, a post-processing inspection unit 80 shown in fig. 4. The post-process inspection portion 80 detects information on the quality of the substrate after the substrate process. For example, the post-process inspection portion 80 detects information on the line width of the resist pattern formed on the surface of the wafer W after the development process. For example, the post-processing inspection portion 80 detects image information from which a difference in line width of the resist pattern can be recognized as a difference in at least any one of hue, lightness, and purity.

Specifically, the post-processing inspection unit 80 includes a holding unit 83, a linear driving unit 84, an imaging unit 81, and a light projection reflection unit 82. The holding portion 83 holds the wafer W horizontally. The linear driving unit 84 moves the holding unit 83 along a horizontal linear path using, for example, a motor or the like as a power source. The imaging unit 81 acquires image data of the front surface of the wafer W. The imaging unit 81 is provided on one end side in the post-treatment inspection unit 80 in the moving direction of the holding unit 83, and faces the other end side in the moving direction.

The light projecting reflector 82 projects light toward the imaging range and guides the reflected light from the imaging range to the imaging unit 81 side. For example, the projection reflector 82 includes a half mirror 86 and a light source 87. The half mirror 86 is provided at a position higher than the holding portion 83 in the middle of the movement range of the holding portion 83, and reflects light from below toward the imaging portion 81. The light source 87 is provided above the half mirror 86, and irradiates illumination light downward through the half mirror 86.

The post-process inspection unit 80 operates as follows to acquire image data of the surface of the wafer W. First, the linear driving unit 84 moves the holding unit 83. Thereby, the wafer W passes under the half mirror 86. In this process, the reflected light from each part of the front surface of the wafer W is sequentially sent to the imaging unit 81. The imaging unit 81 images the reflected light from each part of the front surface of the wafer W to acquire image data of the front surface of the wafer W. This enables detection of image information of the resist pattern.

The post-process inspection unit 80 may detect information on the film thickness of the coating film formed on the surface of the wafer W after the film formation process. For example, the post-processing inspection unit 80 detects image information from which a difference in film thickness of the coating film can be recognized as a difference in at least one of hue, lightness, and purity. This image information can also be detected by the structure exemplified in fig. 4.

The quality inspection apparatus 70 may further have an in-process inspection section 90 shown in fig. 5. The in-process inspection section 90 detects information on the supply state of the processing liquid during the substrate processing. For example, the in-process check unit 90 detects information on the supply state of the developer during the developing process. For example, the in-process check unit 90 includes a liquid splash detection unit 91, a liquid accumulation detection unit 92, and a liquid drop detection unit 93.

The liquid splash detecting section 91 detects information on the generation state of liquid splash during the developer supply process. For example, the liquid splash detecting unit 91 includes an irradiation unit 94 and an imaging unit 95. The irradiation unit 94 is fixed to, for example, the nozzle 31, and irradiates the wafer W with laser light in a horizontal direction. The installation height of the irradiation portion 94 is set to a height that can be reached by the droplets splashed from the surface Wa. The imaging unit 95 acquires image data of the irradiation range of the laser beam from the irradiation unit 94. When the liquid splash occurs, scattering of laser light or the like occurs due to the splashed liquid, and image data acquired by the imaging unit 95 changes. Therefore, the image data acquired by the imaging unit 95 includes information on the generation state of the liquid droplets.

The liquid storage detecting section 92 detects information on the formation state of the developing liquid film on the surface Wa. For example, the liquid storage detecting unit 92 includes an imaging unit 96. The imaging unit 96 acquires image data of the front surface Wa of the wafer W held by the holding unit 21. The image data acquired by the imaging unit 96 includes information on the formation state of the liquid film.

The liquid drop detection portion 93 detects information on the generation state of the drop of the developing liquid from the nozzle 31. The liquid dropping is a phenomenon in which the developer drops from the nozzle 31 except for a predetermined supply period of the developer. For example, the liquid drop detection unit 93 includes an imaging unit 97. The imaging unit 97 acquires image data of the nozzle 31 and the lower side thereof. The image data acquired by the imaging unit 97 includes information on the state of occurrence of liquid droplets.

The in-process inspection section 90 may also detect information on the supply state of the deposition solution during the deposition process. In this case, the same configurations as those of the liquid splash detecting section 91, the liquid storage detecting section 92, the liquid drip detecting section 93, and the like described above can detect information on the supply state of the deposition solution in the coating unit U1.

(control device and machine learning device)

As shown in fig. 6, the functional configuration of the control device 100 (hereinafter referred to as "functional block") includes a processing condition holding unit 111, a processing control unit 112, a data acquisition unit 113, a data input unit 114, and a recommendation condition derivation unit 115.

The processing condition holding unit 111 stores preset processing conditions. For example, the process condition holding section 111 stores the development process conditions of the process module 14. The developing process conditions included the heat treatment conditions of the heat treatment unit U4 and the liquid treatment conditions of the developing unit U3. The liquid processing conditions of the developing unit U3 include flows of supply of the developer, supply and drying of the rinse liquid (spin drying by rotation), and the like. The liquid processing conditions of the developing unit U3 include the rotation speed of the wafer W, the supply amount of the developer, the supply time of the developer, the supply amount of the rinse liquid, the discharge time of the rinse liquid, the spin-drying time, and the like in each flow. In the case where the developer is supplied while the nozzle 31 is moved by the nozzle transfer portion 32, the liquid processing conditions of the developing unit U3 may further include a movement start position, a movement speed, a movement end position, and the like of the nozzle 31 during the supply of the developer.

The process condition holding unit 111 may store the film formation process conditions of the process modules 11, 12, and 13. The film formation process conditions included the liquid process conditions of the coating unit U1 and the heat process conditions of the heat treatment unit U2. The liquid processing conditions in the coating unit U1 include flows such as the supply of deposition solution. The liquid processing conditions of the coating unit U1 include the number of revolutions of the wafer W, the amount of deposition solution to be supplied, and the time for supplying the deposition solution in each flow.

The process control unit 112 causes the processing unit to execute the substrate processing in accordance with the processing conditions stored in the processing condition holding unit 111. For example, the process control unit 112 causes the process module 14 to execute the development process according to the development process condition stored in the process condition holding unit 111. For example, the process controller 112 controls the heat treatment unit U4 to perform a heat treatment (e.g., the PEB) on the wafer W after the exposure process under predetermined heat treatment conditions. Thereafter, the process controller 112 controls the developing unit U3 to perform a developing process on the wafer W under the predetermined liquid processing conditions. Thereafter, the process controller 112 controls the developing unit U4 to perform a heat treatment on the wafer W (for example, PB described above) under the predetermined heat treatment conditions.

The process control unit 112 may cause the process modules 11, 12, and 13 to execute the film formation process according to the film formation process conditions stored in the process condition holding unit 111. For example, the process controller 112 controls the coating unit U1 to apply the deposition solution to the surface Wa of the wafer W under predetermined liquid processing conditions. Thereafter, the process controller 112 controls the heat treatment unit U2 to perform the heat treatment on the wafer W under the predetermined heat treatment conditions.

The data acquisition unit 113 acquires actual data on the quality of substrate processing performed according to the processing conditions. The data acquisition unit 113 can acquire actual data including actual values of a plurality of items. The actual values of the plurality of items may include a post-process item indicating the quality of the wafer W after the substrate processing and an actual value of a process-in-process item indicating the supply state of the processing liquid in the middle of the substrate processing. As the actual values of the plurality of items, actual data including a plurality of actual values of the same kind can be acquired. The plurality of actual values of the same kind means a plurality of actual values which should ideally be the same value. As a specific example of the plurality of actual values of the same kind, a plurality of actual values obtained at a plurality of locations can be cited.

For example, as an example of the post-process item, the data acquisition unit 113 acquires an actual value (hereinafter, referred to as "actual line width value") indicating the line width of the resist pattern formed on the surface Wa of the wafer W by the development process. Specifically, the data acquisition unit 113 acquires the actual value of the line width based on the information detected by the post-processing inspection unit 80. The data acquisition unit 113 may acquire actual line widths of a plurality of portions on the surface Wa based on the information detected by the post-processing inspection unit 80.

As an example of the in-process item, the data acquisition unit 113 acquires an actual value indicating a supply state of the developer during the development process. Specifically, the data acquisition unit 113 acquires actual values of the liquid splashing of the developer, the poor formation of the liquid film, and the presence or absence of the liquid dripping based on the information detected by the inspection unit 90 during the processing.

As an example of the post-process item, the data acquisition unit 113 may acquire an actual value (hereinafter, referred to as "actual film thickness value") indicating the film thickness of the film formed on the surface Wa of the wafer W by the film formation process. Specifically, the data acquisition unit 113 may acquire the actual film thickness value based on the information detected by the post-process inspection unit 80. The data acquisition unit 113 can acquire actual film thickness values at a plurality of locations on the surface Wa based on the information detected by the post-processing inspection unit 80.

As an example of the in-process item, the data acquisition unit 113 may acquire an actual value indicating a supply state of the deposition solution during the deposition process. Specifically, the data acquisition unit 113 can acquire actual values of the deposition solution for liquid splashing, liquid film formation failure, and the presence or absence of liquid dripping, based on the information detected by the inspection unit 90 during the processing.

The data input unit 114 inputs a data set including the processing conditions and actual data corresponding to the processing conditions to a model generation unit 214 (described later) of the machine learning device 200. The data input section 114 may select a data set to be input to the model generation section 214 based on the actual value of the item in the above-described processing. For example, the data input unit 114 may exclude a data set in which the supply state of the processing liquid is defective from the input target to be input to the model generation unit 214. As a specific example of the state of supply of the treatment liquid being poor, at least one of the above-described liquid splashing, poor formation of a liquid film, and liquid dripping may be generated.

The recommended condition deriving section 115 derives the recommended processing condition for the substrate processing based on a learning model generated by the model generating section 214 through machine learning based on the plurality of sets of data sets. As described later, the learning model is generated so that it is possible to output the prediction data on the quality of the substrate processing in response to the input of the processing condition (i.e., the learning model is generated with the processing condition as input and the prediction data on the quality of the substrate processing as output). The recommended processing condition is a processing condition that is determined to be recommended based on a predetermined evaluation condition of the learning model and the prediction data.

For example, more detailed functional blocks of the recommendation condition derivation section 115 include an evaluation condition input section 121 and a search result acquisition section 122. The evaluation condition input unit 121 inputs the evaluation condition of the prediction data to a condition search unit 216 (described later) of the machine learning device 200. The evaluation condition is a condition for judging whether or not the predicted data is at an allowable level (allowable level).

The evaluation condition input unit 121 may input an evaluation condition for evaluating the predicted values of the plurality of items to the condition search unit 216. The evaluation condition input to the condition search unit 216 by the evaluation condition input unit 121 may include a condition relating to fluctuation (dispersion) of the predicted values of at least a part of the plurality of items. For example, the evaluation condition includes a derivation method of an evaluation score (score) of the prediction data and a tolerance level of the evaluation score.

For example, the evaluation condition input unit 121 inputs an evaluation condition for evaluating the predicted values of the line widths (hereinafter referred to as "predicted values of line widths") at a plurality of locations on the surface Wa to the condition search unit 216. The evaluation condition includes, as an example of a method of deriving the evaluation score, a calculation formula (for example, a standard deviation calculation formula) for calculating a fluctuation of predicted line width values of at least a part (for example, all parts) of the plurality of parts. The evaluation condition includes an allowable upper limit value of the fluctuation calculated by the calculation formula as an allowable level of the evaluation score.

The evaluation condition input unit 121 may input, to the condition search unit 216, evaluation conditions for evaluating the predicted values of the film thicknesses at a plurality of locations on the surface Wa. The evaluation conditions include a calculation formula (for example, a calculation formula of standard deviation) for calculating the fluctuation of the film thickness predicted values of at least a part (for example, all the parts) of the plurality of parts, as an example of the method for deriving the evaluation score. The evaluation condition includes an allowable upper limit value of the fluctuation calculated by the calculation formula as an allowable level of the evaluation score.

The search result acquisition unit 122 acquires the recommended processing conditions derived by the condition search unit 216 and stores the acquired recommended processing conditions in the processing condition holding unit 111. As described later, the recommended processing conditions are derived based on the plurality of sets of data sets, the learning model, and the evaluation conditions input by the evaluation condition input unit 121.

Here, the process control section 112 may further cause the processing section to execute the substrate processing in accordance with the recommended processing conditions. The data acquisition unit 113 may further acquire additional actual data regarding the quality of the substrate processing performed under the recommended processing conditions. The data input unit 114 may further input an additional data set including the recommended processing conditions and additional actual data to the model generating unit 214. The model generation unit 214 may update the learning model based on the additional data set, and the recommended condition derivation unit 115 may update the recommended processing condition based on the updated learning model. Updating the learning model means that a new learning model is generated based on the plurality of sets of data sets to which the additional data set is added. The update of the recommended processing condition refers to deriving a new recommended processing condition based on the learning model updated by the model generation unit 214.

In this case, the control device 100 may further include a condition evaluation unit 116 and a repetition management unit 117. The condition evaluation unit 116 evaluates whether or not recommended processing conditions can be adopted. The repetition management unit 117 repeats at least the following processing until the evaluation result of the condition evaluation unit 116 is acceptable.

i) The process control unit 112 further causes the processing unit to execute the substrate processing in accordance with the recommended processing conditions.

ii) the data acquisition unit 113 further acquires additional actual data.

iii) the data input unit 114 further inputs the additional data set to the model generation unit 214.

iv) the model generation unit 214 updates the learning model based on the additional data set, and the recommendation condition derivation unit 115 updates the recommendation processing condition based on the updated learning model.

The method of evaluating the recommended processing conditions by the condition evaluation unit 116 is not particularly limited. For example, the condition evaluation unit 116 evaluates whether or not the recommended processing condition can be adopted based on the evaluation result of the additional actual data obtained based on the predetermined evaluation condition. The evaluation conditions may be the same as those of the above-described prediction data. For example, the evaluation conditions include a method of deriving an evaluation score by adding actual data and an allowable level of the evaluation score.

For example, the condition evaluation unit 116 evaluates the line width actual values at a plurality of locations on the surface Wa based on a predetermined evaluation condition. The evaluation condition includes, as an example of a method of deriving the evaluation score, a calculation formula (for example, a calculation formula of a standard deviation) for calculating a fluctuation of a line width actual value of at least a part (for example, all parts) of the plurality of parts. The evaluation condition includes an allowable upper limit value of the fluctuation calculated by the calculation formula as an allowable level of the evaluation score.

The evaluation condition input unit 121 may evaluate the film thickness actual values at a plurality of locations on the surface Wa based on predetermined evaluation conditions. The evaluation conditions include a calculation formula (for example, a calculation formula of standard deviation) for calculating the fluctuation of the film thickness actual value of at least a part (for example, all the parts) of the plurality of parts as an example of the method for deriving the evaluation score. The evaluation condition includes an allowable upper limit value of the fluctuation calculated by the calculation formula as an allowable level of the evaluation score.

The condition evaluation unit 116 may evaluate whether or not the latest recommended processing condition can be adopted based on whether or not a difference between the latest recommended processing condition and a past recommended processing condition (for example, the last recommended processing condition) is an allowable level. It is assumed that the recommended processing condition can gradually converge to one condition in the iterative processing performed by the repetition management unit 117. By reducing the difference between the latest recommended processing condition and the past recommended processing condition to an allowable level, it is possible to adopt a recommended processing condition close to the convergence result.

The condition evaluation unit 116 may evaluate whether or not the latest recommended processing condition can be adopted based on whether or not the difference between the latest additional actual data and the past additional actual data is an allowable level. The condition evaluation unit 116 may evaluate whether or not the latest recommended processing condition can be adopted based on whether or not the difference between the evaluation score of the latest additional actual data and the evaluation score of the past additional actual data is an allowable level.

The control device 100 may further include an actual data correcting unit 118. The actual data correcting unit 118 removes a component due to a factor other than the substrate process performed by the processing unit of the coating and developing apparatus 2 from the actual data of the data set before the data input unit 114 inputs the data set to the model generating unit 214. For example, the actual data correcting unit 118 removes a fluctuation component caused by the exposure processing from the actual line width values of the plurality of portions. Specifically, the actual data correction unit 118 removes a fluctuation pattern (discrete pattern) unique to the exposure process, which has been examined in advance, from the actual line width values of the plurality of portions.

The functional blocks of the machine learning device 200 include a search operation unit 211, a data acquisition unit 212, a data holding unit 213, a model generation unit 214, a model holding unit 215, and a condition search unit 216. The search arithmetic unit 211 is an engine of machine learning in the machine learning apparatus 200. For example, the search arithmetic unit 211 searches for an optimal solution by a genetic algorithm based on a preset learning condition. The learning conditions include individuals of the first generation, a method of deriving an evaluation score of an individual, and a tolerance level of the evaluation score.

The search calculation unit 211 acquires a plurality of individuals of the first generation, and calculates an evaluation score of each individual. Then, the search calculation unit 211 eliminates individuals whose evaluation score is far from the tolerance level, and evolves a plurality of individuals into a plurality of individuals of the next generation by calculation such as crossover, inversion, and mutation. Thereafter, the search calculation unit 211 repeatedly derives individuals whose evaluation scores are at an acceptable level by repeating the derivation of the evaluation scores of the individuals, the elimination of the individuals, and the evolution of the individuals.

The data acquisition unit 212 acquires the data set and the additional data set from the data input unit 114. The data holding unit 213 stores the data set acquired by the data acquisition unit 212 as a database for learning.

The model generation unit 214 performs machine learning to generate the learning model based on the plurality of sets of data stored in the data storage unit 213. The model generation unit 214 may generate a learning model by machine learning including an arithmetic process of searching for the learning model by a genetic program. For example, the model generation unit 214 generates a learning model including a plurality of model equations that can output predicted values of a plurality of items in response to input of processing conditions. When each model equation is generated, the model generation unit 214 sets the learning condition for deriving the model equation, and requests the search operation unit 211 to derive a model equation that complies with the learning condition.

For example, the model generation unit 214 generates a plurality of provisional model equations capable of generating predicted values in response to input of processing conditions, and uses them as the plurality of first-generation individuals. The temporary model formula takes various operators and random numerical values as elements, and a calculation formula is represented by a tree structure. The model generation unit 214 determines a derivation method of the deviation score representing the deviation between the actual value and the predicted value obtained based on the provisional model formula as the evaluation score in the learning condition. For example, the model generation unit 214 specifies a derivation method including at least the following steps.

a1) And inputting the processing conditions of the multiple groups of data sets into a temporary model formula to derive multiple predicted values.

a2) A step of deriving a deviation score representing a deviation between the plurality of predicted values and actual values of the plurality of sets of data.

The deviation score may be any value as long as it can represent the deviation between the plurality of predicted values and the actual values of the plurality of sets of data sets. Specific examples of the deviation score include a square of a difference between the predicted value and the actual value and/or a square root of the sum of squares. The model generation unit 214 uses an upper limit value preset for the deviation score as the allowable level of the evaluation score in the learning condition.

The search calculation unit 211 repeatedly derives a model equation having a deviation score equal to or less than the upper limit value by repeating derivation of the deviation score of the temporary model equation, elimination of the temporary model equation, and evolution of the temporary model equation. The model generation unit 214 acquires the model equation derived by the search operation unit 211 and stores the model equation in the model holding unit 215. Through the above steps, the model generation unit 214 stores each model equation in the model holding unit 215, and then the model holding unit 215 generates a learning model including a plurality of model equations.

The condition search unit 216 derives recommended processing conditions based on the plurality of sets of data stored in the data holding unit 213, the learning model stored in the model holding unit 215, and the evaluation conditions input by the evaluation condition input unit 121. The condition search section 216 may derive the recommended processing condition through a search process including an arithmetic process of searching for the recommended processing condition by a genetic algorithm. For example, the condition search unit 216 sets the learning condition for deriving the recommended processing condition, and requests the search operation unit 211 to derive the recommended processing condition that complies with the learning condition.

For example, the condition search unit 216 uses the processing conditions of the plurality of sets of data stored in the data holding unit 213 as the plurality of individuals of the first generation. Each processing condition represents a condition of a plurality of items in a tree structure.

The condition search unit 216 determines a method of deriving the evaluation score in the learning condition, including at least the following steps.

b1) And a step of inputting the processing conditions for the plurality of sets of data sets to the learning model stored in the model holding unit 215 to derive the prediction data.

b2) And deriving an evaluation score of the prediction data according to the derivation method of the evaluation condition input by the evaluation condition input unit 121.

The condition search unit 216 uses the tolerance level in the evaluation condition input by the evaluation condition input unit 121 as the tolerance level of the evaluation score in the learning condition.

The search calculation unit 211 repeatedly derives recommended processing conditions for which the evaluation score is at a tolerance level by repeating derivation of the evaluation score of the processing conditions, elimination of the processing conditions, and evolution of the processing conditions. The condition search unit 216 acquires the recommended processing condition derived by the search operation unit 211 and outputs the acquired recommended processing condition to the search result acquisition unit 122.

Fig. 7 is a block diagram illustrating a hardware configuration of the control device 100 and the machine learning device 200. The control device 100 includes a circuit 190. The circuit 190 includes at least one processor 191, memory 192, storage 193, display device 194, input device 195, input-output port 196, and communication port 197. The storage device 193 is a non-volatile storage medium (e.g., flash memory) readable by a computer. For example, the storage device 193 stores a program for causing the control device 100 to execute the steps of: a step of causing the coating and developing device 2 to perform a substrate process in accordance with a predetermined process condition; a step of acquiring actual data on the quality of substrate processing performed according to the processing conditions from the quality inspection apparatus 70; a step of inputting a data set including a process condition of a substrate process and actual data of the substrate process to the machine learning device 200; and a step of deriving recommended processing conditions for the substrate processing based on the learning model, wherein the learning model is generated by the machine learning device 200 based on the plurality of sets of data. For example, the storage device 193 includes a storage area for storing a program constituting the above-described functional module, and a storage area allocated to the processing condition holding unit 111.

The display device 194 is used to display recommended processing conditions and the like. The display device 194 and the input device 195 function as a user interface of the control apparatus 100. The display device 194 includes, for example, a liquid crystal monitor or the like for displaying information to a user. The input device 195 is, for example, a keyboard or the like, and acquires information input by the user. The display device 194 and the input device 195 may be integrally configured as a so-called touch panel. The input device 195 can be used to input processing conditions, evaluation conditions, and the like.

The memory 192 temporarily stores programs loaded from the storage device 193, operation results of the processor 191, and the like. The processor 191 controls the coating and developing apparatus 2 by executing the program in conjunction with the memory 192. The input/output port 196 inputs and outputs an electric signal between the display device 194 and the input device 195, according to an instruction from the processor 191. The communication port 197 performs network communication with the machine learning device 200 according to an instruction from the processor 191.

Machine learning apparatus 200 includes circuitry 290. Circuitry 290 includes a processor 291, memory 292, storage 293, and communication ports 294. The storage device 293 is a non-volatile storage medium (e.g., flash memory) readable by a computer. For example, the storage device 293 stores a program for causing the machine learning device 200 to execute the following steps: a step of acquiring the data set; and generating the learning model by machine learning based on the plurality of sets of data. For example, the storage device 293 includes a storage area for storing programs for constituting the above-described functional modules, and storage areas assigned to the data holding unit 213 and the model holding unit 215.

The memory 292 temporarily stores programs loaded from the storage device 293, operation results of the processor 291, and the like. The processor 291 executes the program in conjunction with the memory 292 to generate the learning model. The communication port 294 performs network communication with the machine learning apparatus 100 according to an instruction from the processor 291.

[ Condition setting Assist procedure ]

Next, as an example of the condition setting support method, a condition setting support procedure executed by each of the control device 100 and the machine learning device 200 will be described. The condition setting assisting step executed by the control apparatus 100 includes a derivation step of recommended processing conditions and a modification (flush up) step of recommended processing conditions. The condition setting assisting step executed by the machine learning device 200 includes a learning model generation step and a recommended processing condition search step. The following specifically exemplifies each step.

(derivation procedure of recommended processing conditions)

The step of deriving the recommended processing condition by the control device 100 includes: a step of causing the coating and developing apparatus 2 to perform a substrate process including supplying a process liquid to the wafer W, in accordance with a predetermined process condition; a step of acquiring actual data on the quality of substrate processing according to the processing conditions; a step of inputting a data set including a process condition of a substrate process and actual data of the substrate process to the machine learning device 200; and a step of deriving recommended processing conditions based on the learning model generated by the machine learning device 200 based on the plurality of sets of data. Wherein the step of deriving the recommended processing condition may include: inputting the evaluation condition of the prediction data to the machine learning device 200; and a step of acquiring recommended processing conditions derived by the machine learning device 200 based on the plurality of sets of data sets, the learning model, and the evaluation conditions.

As shown in fig. 8, the control device 100 first executes steps S01, S02, S03. In step S01, the process control unit 112 causes the coating and developing apparatus 2 to start the substrate processing in accordance with the processing conditions stored in the processing condition holding unit 111. In step S02, the data acquisition unit 113 acquires the actual value of the item in the above-described processing. The data acquisition unit 113 can acquire actual values of a plurality of items under processing. For example, the data acquisition unit 113 acquires actual values of the presence or absence of liquid splashing of the developer, poor liquid film formation, and liquid dripping based on the information detected by the inspection unit 90 during the processing. The data acquisition unit 113 may acquire actual values of the deposition solution for liquid splashing, liquid film formation failure, and the presence or absence of liquid dripping, based on the information detected by the inspection unit 90 during the processing. In step S03, the process control unit 112 checks whether or not the substrate processing according to the processing conditions is completed.

If it is determined in step S03 that the substrate processing has not been completed, the control device 100 returns the process to step S02. And then continuously acquiring the actual value of the item in process until the substrate processing is finished. In the case where it is determined in step S03 that the substrate processing is ended, the control device 100 executes step S04. In step S04, the data input unit 114 checks whether or not a failure has occurred in the supply state of the processing liquid based on the actual value of the item under processing.

If it is determined in step S04 that the supply state of the processing liquid is not defective, the control device 100 executes steps S05, S06, and S07. In step S05, the data acquisition unit 113 acquires the actual value of the processed item. The data acquisition unit 113 can acquire actual values of a plurality of processed items. For example, the data acquisition unit 113 acquires the actual line widths of a plurality of portions on the surface Wa based on the information detected by the post-processing inspection unit 80. The data acquisition unit 113 may acquire the actual film thickness values at a plurality of locations on the surface Wa based on the information detected by the post-processing inspection unit 80. In step S06, the actual data correction section 118 removes components due to causes other than the substrate processing from the actual values of the plurality of post-processing items. In step S07, the data input unit 114 inputs a data set including the processing conditions and the actual data (the actual values of the plurality of post-processing items) corresponding to the processing conditions to the machine learning device 200.

Next, control device 100 executes step S08. If it is determined in step S04 that a failure has occurred in the supply state of the processing liquid, the control device 100 does not execute steps S05, S06, and S07 and directly executes step S08. In step S08, the data input unit 114 confirms whether or not the input of the data sets of the number necessary for machine learning in the machine learning device 200 has been completed.

If it is determined in step S08 that the input of the necessary number of data sets for machine learning has not been completed, control device 100 executes step S09. In step S09, the process control section 112 changes the process conditions. For example, the processing control section 112 changes the processing conditions based on an input to the input device 195 by a user or the like. Then, control device 100 returns the process to step S01. Thereafter, the change of the processing conditions, the execution of the substrate processing, and the input of the data set are repeated until the input of the data set of the number necessary for performing the machine learning is completed.

In step S08, if it is determined that the input of the necessary number of data sets for machine learning has been completed, the control device 100 executes steps S11, S12, S13, and S14. In step S11, the evaluation condition input unit 121 waits for a learning completion notification to be received from the machine learning device 200. In step S12, the evaluation condition input unit 121 sets an evaluation condition of the prediction data. For example, the evaluation condition input unit 121 sets the evaluation condition of the prediction data based on the input of the user to the input device 195 or the like. In step S13, the evaluation condition input unit 121 inputs the evaluation condition set in step S12 to the machine learning device 200. In step S14, the search result acquisition unit 122 acquires recommended processing conditions derived by the machine learning device 200 based on the plurality of sets of data sets, the learning model, and the evaluation conditions input by the evaluation condition input unit 121, and stores the recommended processing conditions in the processing module 11. At this point, the derivation of the recommended processing conditions is completed.

(improvement of treatment conditions is recommended)

The improvement of the recommended processing condition by the control device 100 includes: a step of causing the coating and developing device 2 to further perform substrate processing in accordance with the recommended processing conditions; further acquiring additional actual data on the quality of the substrate processing performed under the recommended processing conditions; a step of inputting an additional data set including the recommended processing conditions and additional actual data to the machine learning device 200; and a step of updating the recommended processing conditions based on the updated learning model, which is updated by the machine learning device 200 based on the additional data set. The improving step may further include a step of evaluating the recommended processing condition, wherein the following steps are repeatedly performed until an evaluation result of the recommended processing condition reaches a prescribed level: a step of causing the coating and developing device 2 to further perform substrate processing in accordance with the recommended processing conditions; further acquiring additional actual data; a step of inputting the additional data set to the machine learning device 200; and a step of updating the recommended processing condition based on the learning model updated by the machine learning device 200 according to the additional data set.

As shown in fig. 9, the control device 100 first executes steps S21, S22, S23, S24, S25. In step S21, the process control unit 112 causes the coating and developing apparatus 2 to execute the substrate processing in accordance with the recommended processing conditions stored in the processing condition holding unit 111. In step S22, the data acquisition unit 113 acquires the additional actual value of the post-processing item. The data acquisition unit 113 can acquire additional actual values of a plurality of processed items. In step S23, the actual data correction unit 118 removes the components due to the factors other than the substrate processing from the added actual values of the plurality of post-processing items. In step S24, the condition evaluation unit 116 evaluates the recommended processing condition. In step S25, the repetition management unit 117 confirms whether or not the recommended processing condition can be adopted based on the evaluation result in step S24.

In the case where it is determined in step S25 that the recommended processing conditions cannot be adopted, the control device 100 executes steps S26, S27, and S28. In step S26, the data input unit 114 inputs an additional data set including the processing conditions and additional actual data (additional actual values of a plurality of post-processing items) corresponding to the processing conditions to the machine learning device 200. In step S27, the search result acquisition unit 122 waits for the reception of a learning model update completion notification from the machine learning device 200. In step S28, the search result acquisition unit 122 acquires the recommended processing conditions updated by the machine learning device 200 based on the additional data set, and stores the recommended processing conditions in the processing module 11. Then, control device 100 returns the process to step S21. Thereafter, acquisition of additional actual data and update of the recommended processing condition are repeated until the recommended processing condition can be adopted.

If it is determined in step S25 that the recommended processing condition can be adopted, control device 100 ends the processing. By this point, the step of recommending improvement of the processing conditions is completed.

(learning model creation step)

The step of generating the learning model by the machine learning device 200 includes: a step of acquiring the data set; a step of generating a learning model by machine learning based on the plurality of sets of data. The step of generating the learning model by machine learning may include an arithmetic process of searching the learning model by a genetic program. A learning model may be generated that contains a plurality of model formulas capable of outputting predicted values of a plurality of items, respectively, in response to input of a processing condition.

As shown in fig. 10, the machine learning device 200 first executes steps S31, S32, and S33. In step S31, the data acquisition unit 212 waits for the input of the data set from the data input unit 114. In step S32, the data acquisition unit 212 stores the input data set in the data storage unit 213. In step S33, the data holding unit 213 checks whether or not the number of data sets stored in the data holding unit 213 reaches the number necessary for machine learning.

If it is determined in step S33 that the number of stored data sets has not reached the number required for machine learning, the control device 100 returns the process to step S31. Thereafter, the acquisition of the data sets is repeated until the number of data sets necessary for machine learning is stored.

In the case where it is determined in step S33 that the number of stored data sets reaches the number required for machine learning, the control device 100 executes steps S34, S35, and S36. In step S34, the model generation unit 214 sets the learning condition for deriving the model equation corresponding to any one of the predicted values, and requests the search operation unit 211 for the model equation that conforms to the learning condition. For example, the model generation unit 214 generates a plurality of provisional model equations capable of generating predicted values in response to input of processing conditions, and uses them as the plurality of first-generation individuals. The model generation unit 214 determines the derivation method using the deviation score as the evaluation score, and uses the upper limit value of the deviation score as the allowable level of the evaluation score. In step S35, the search operation unit 211 calculates the deviation score of each provisional model equation according to the learning condition. In step S36, the search operation unit 211 checks whether or not there is a temporary model equation having a deviation score equal to or less than the upper limit value according to the learning condition.

If it is determined in step S36 that there is no provisional model equation having a deviation score equal to or less than the upper limit value, the machine learning device 200 executes step S37. In step S37, the search operation unit 211 eliminates the temporary model equations with a large deviation score exceeding the upper limit value, and evolves the temporary model equations into the next-generation temporary model equations by operations such as intersection, inversion, and mutation. Then, the machine learning device 200 returns the process to step S35. Then, derivation of the deviation score of the temporary model formula, elimination of the temporary model formula, and evolution of the temporary model formula are repeated until the temporary model formula having the deviation score of the upper limit value or less can be derived.

If it is determined in step S36 that there is a temporary model equation having a deviation score equal to or less than the upper limit value, the machine learning device 200 executes steps S38 and S39. In step S38, the search operation unit 211 selects a provisional model formula having the best (smallest) deviation score, and stores the provisional model formula in the model holding unit 215 as one model formula of the learning model. In step S39, the model generation unit 214 checks whether or not the derivation of all the model equations (that is, all the model equations necessary for deriving the predicted values of the plurality of items) constituting the learning model has been completed.

If it is determined in step S39 that the derivation of all the model equations is not completed, the machine learning device 200 executes step S41. In step S41, the model generation unit 214 changes the model equation of the object of derivation. In other words, the model generation unit 214 changes the item to be predicted of the model formula. Then, the machine learning device 200 returns the process to step S34. After that, the setting of the learning condition and the derivation of the model formula based on the learning condition are repeated until the derivation of all the model formulas is completed.

When determining in step S39 that the derivation of all the model equations has been completed, the machine learning device 200 ends the generation of the learning model. By this, the generation step of the learning model is completed.

(search step of recommended processing conditions)

The step of searching for recommended processing conditions by the machine learning device 200 includes a step of deriving recommended processing conditions for substrate processing based on the plurality of sets of data sets, the learning model, and the evaluation conditions of the prediction data. The step of deriving the recommended processing condition may include an arithmetic process of searching for the recommended processing condition by a genetic algorithm. The recommended processing condition may be derived based on the plurality of sets of data, the plurality of model formulas, and an evaluation condition that evaluates predicted values of the plurality of items. For example, the recommendation processing condition may be derived based on an evaluation condition containing a condition on fluctuation of predicted values of a plurality of items.

As shown in fig. 11, the machine learning device 200 first executes steps S51 and S52. In step S51, the condition search unit 216 waits for the input of the evaluation condition from the evaluation condition input unit 121. In step S52, the condition search unit 216 sets the learning condition for deriving the recommended processing condition, and requests the search operation unit 211 to derive the recommended processing condition that complies with the learning condition. For example, the condition search unit 216 uses the processing conditions of the plurality of sets of data stored in the data holding unit 213 as the plurality of individuals of the first generation. The condition search unit 216 determines a method of deriving the evaluation score and a permissible level of the evaluation score based on the evaluation condition input by the evaluation condition input unit 121.

Next, the machine learning device 200 executes steps S53, S54, and S55. In step S53, the search operation unit 211 inputs each processing condition to the learning model stored in the model holding unit 215, and derives prediction data. In step S54, the search operation unit 211 derives an evaluation score of the prediction data. In step S55, the search operation unit 211 checks whether or not there is a processing condition in which the evaluation score is at the allowable level.

If it is determined in step S55 that there is no processing condition for which the evaluation score is at the acceptable level, the machine learning device 200 executes step S56. In step S56, the search operation unit 211 eliminates the processing conditions having the evaluation scores far from the allowable level, and progresses the plurality of processing conditions to the plurality of processing conditions of the next generation by operations such as intersection, inversion, and mutation. Then, the machine learning device 200 returns the process to step S53. Then, the derivation of the evaluation score of the processing condition, the elimination of the processing condition, and the evolution of the processing condition are repeated until the processing condition having the evaluation score at the allowable level can be derived.

If it is determined in step S55 that there is a processing condition in which the evaluation score is at the acceptable level, the machine learning device 200 executes steps S57 and S58. In step S57, the search operation unit 211 uses the processing condition with the evaluation score of the optimal value as the recommended processing condition. In step S58, the condition search unit 216 acquires the recommended processing condition derived by the search operation unit 211 and outputs the acquired recommended processing condition to the search result acquisition unit 122. At this point, the search step of recommending the processing conditions is completed.

The derivation of the recommended processing condition is not limited to the arithmetic processing for searching the recommended processing condition by the above-described genetic algorithm. For example, in step S55, the change of the processing conditions and the derivation of the evaluation score may be repeated until the evaluation score reaches the allowable level, and the recommended processing conditions may be derived by such arithmetic processing.

(concrete examples)

As an example, a setting assisting step of the process conditions of the developing process in the developing unit U3 is specifically illustrated. The processing conditions of the developing process in the developing unit U3 include, for example, the rotation speed of the wafer W, the supply amount of the developer, the supply time of the developer, the supply amount of the rinse liquid, the discharge time of the rinse liquid, the spin-drying time, the movement start position of the nozzle 31, the movement speed of the nozzle 31, the movement end position of the nozzle 31, and the like. Among the items requiring the recommended processing conditions are, for example, the rotation speed of the wafer W during the supply of the developer and the moving speed of the nozzle 31. In this case, in steps S01 to S09, the operation of inputting the data set to the machine learning device 200 while changing the rotation speed of the wafer W and the movement speed of the nozzle 31 is repeated.

For example, in steps S01 to S09, the movement speed of the nozzle 31 is set to 15mm/S, 20mm/S, and 25mm/S with the wafer W rotating at 200rpm, the movement speed of the nozzle 31 is set to 15mm/S, 20mm/S, and 25mm/S with the wafer W rotating at 250rpm, and the movement speed of the nozzle 31 is set to 15mm/S, 20mm/S, and 25mm/S with the wafer W rotating at 300 rpm. In step S04 in which any one of the above-described processing conditions is executed, when it is determined that a failure has occurred in the supply state of the processing liquid, the data set corresponding to the processing condition is excluded from the input objects to be input to the machine learning device 200. In this case, the processing conditions can be further changed in step S09 in order to obtain the number of data sets necessary for machine learning. For example, when it is determined that liquid scattering of the developer has occurred under the processing conditions of the rotation speed of 300rpm and the movement speed of 25mm/s, the rotation speed is changed to 290rpm, and actual data under the conditions of the rotation speed of 290rpm and the movement speed of 25mm/s is again acquired.

In step S05, for example, for the wafer W divided into n parts, the average line width of each of the divided regions is acquired as n actual line widths. An example of a data set in this case is as follows.

The treatment conditions are as follows: the rotation speed of the wafer W was 200rpm, and the moving speed of the nozzle was 15mm/s

Actual data: w1-23 nm, W2-28 nm, W3-31 nm, … Wn-24 nm (Wi: average line width in the divided region i)

The learning model generated in the machine learning device 200 based on the data set is, for example, capable of outputting a predicted value of the average line width of the n divided regions in response to the input of the rotation speed of the wafer W and the movement speed of the nozzle. In step S12, for example, a formula for calculating a standard deviation of n predicted values of line width is set as the formula for calculating the evaluation score, and a permissible value of the standard deviation is set as the permissible level. Based on the evaluation conditions set as described above, the machine learning device 200 can derive recommended values of the rotation speed of the wafer W and the movement speed of the nozzle 31 as the recommended processing conditions (for example, the rotation speed of the wafer W is 234rpm, and the movement speed of the nozzle 31 is 22 rpm).

(Effect of the present embodiment)

As described above, the condition setting assisting method for substrate processing according to the present embodiment includes a step of inputting a data set including processing conditions of substrate processing including processing of supplying a processing liquid to a wafer W and actual data on the quality of the substrate processing performed by the coating and developing apparatus 2 to the machine learning apparatus 200, and a step of deriving recommended processing conditions for the substrate processing based on a learning model generated by the machine learning apparatus 200 based on a plurality of sets of data sets by machine learning, which is capable of outputting predicted data on the quality of the substrate processing in response to the input of the processing conditions.

In this condition setting support method, since the recommended processing condition is derived based on a learning model generated by a machine learning method, it is possible to efficiently search for an appropriate processing condition. Therefore, the operation of setting the processing conditions for the substrate processing can be effectively simplified.

The condition setting assisting method for substrate processing may include: a step of causing the coating and developing device 2 to further perform substrate processing in accordance with the recommended processing conditions; further acquiring additional actual data on the quality of the substrate processing performed under the recommended processing conditions; a step of inputting an additional data set including the recommended processing conditions and additional actual data to the machine learning device 200; and a step of updating the recommended processing conditions based on the updated learning model, which is updated by the machine learning device 200 based on the additional data set. In this case, the recommended processing conditions are updated by feedback of the recommended processing conditions and the additional actual data. Therefore, more appropriate processing conditions can be efficiently searched.

The substrate processing condition setting support method may further include a step of evaluating recommended processing conditions, wherein the following steps are repeatedly performed until the evaluation result of the recommended processing conditions reaches a predetermined level: a step of causing the coating and developing device 2 to further perform substrate processing in accordance with the recommended processing conditions; further acquiring additional actual data; a step of inputting the additional data set to the machine learning device 200; and a step of updating the recommended processing condition based on the learning model updated by the machine learning device 200 from the additional data set. In this case, the iterative process can efficiently search for a more suitable process condition.

The step of deriving the recommended processing condition may comprise: inputting the evaluation condition of the prediction data to the machine learning device 200; and a step of acquiring recommended processing conditions derived by the machine learning device 200 based on the plurality of sets of data sets, the learning model, and the evaluation conditions. In this case, the machine learning device 200 also searches for recommended processing conditions, and more appropriate processing conditions can be more efficiently searched for.

In the condition setting support method for substrate processing, actual data including actual values of a plurality of items may be acquired and input to the machine learning device 200 which generates a learning model including a plurality of model equations capable of outputting predicted values of the plurality of items in response to input of processing conditions, respectively, and evaluation conditions for evaluating the predicted values of the plurality of items may be input to the machine learning device 200. In this case, by expanding the evaluation condition into a plurality of items, the quality of the process can be evaluated more appropriately, and a more appropriate process condition can be searched for.

In the condition setting assistance method of substrate processing, the evaluation condition input to the machine learning device 200 may include a condition on fluctuation of predicted values of at least a part of the plurality of items. In this case, the plurality of items can be efficiently evaluated, and more appropriate processing conditions can be efficiently searched for.

In the condition setting support method for substrate processing, actual data including a post-processing item indicating the quality of the wafer W after substrate processing and an actual value of the item during substrate processing may be acquired, and a data set to be input to the machine learning device 200 may be selected based on the actual value of the item during processing. In this case, the abnormality in the processing can be directly grasped using the data in the processing, and the search range for searching for the recommended processing condition based on the quality after the processing can be narrowed. Therefore, the appropriate processing conditions can be searched more efficiently.

The condition setting assisting method for substrate processing may further include: a step of removing components caused by causes other than substrate processing from actual data of the data set before inputting the data set to the machine learning device 200. In this case, more appropriate processing conditions can be searched for.

The substrate processing may include a developing process of supplying a developer to the photosensitive coating film subjected to the exposure process on the surface Wa of the wafer W, and the acquired actual data may include an actual value of a line width of a pattern formed on the surface Wa of the wafer W by the developing process. When the substrate processing includes a developing process, it tends to require a lot of labor to derive appropriate processing conditions. Therefore, the condition setting support method described above can efficiently search for appropriate processing conditions, and is significantly effective.

The substrate processing may include a film formation process of applying the deposition solution to the surface Wa of the wafer W to form a coating film, and the acquired actual data may include an actual value of a film thickness of the coating film formed on the surface Wa of the wafer W by the film formation process. In the case where the substrate processing includes the film formation processing, the quality of the substrate processing is very sensitive to the processing conditions, and therefore, in order to derive appropriate processing conditions, much labor tends to be required. Therefore, the condition setting support method described above can efficiently search for appropriate processing conditions, and is significantly effective.

The embodiments have been described above, but the present invention is not limited to the above embodiments and various modifications can be made without departing from the scope of the invention. For example, the substrate to be processed is not limited to a semiconductor wafer, and may be, for example, a glass substrate, a mask substrate, an fpd (flat Panel display), or the like.

Description of the reference numerals

2 … coating and developing apparatus (substrate processing apparatus), 11, 12, 13, 14 … processing modules (processing units), 112 … processing control unit, 113 … data acquisition unit, 114 … data input unit, 115 … recommended condition derivation unit, 121 … evaluation condition input unit, 122 … search result acquisition unit, 214 … model generation unit, W … wafer, Wa … surface.

Claims (21)

1. A method for assisting condition setting in substrate processing, comprising:

a step of inputting a data set to a machine learning apparatus, wherein the data set includes process conditions of a substrate process and actual data on quality of the substrate process, the substrate process being performed by the substrate processing apparatus and including a process of supplying a process liquid to a substrate; and

a step of deriving a recommended processing condition for the substrate processing based on a learning model, wherein the learning model is a model generated by machine learning by the machine learning device based on a plurality of sets of the data sets, and prediction data on the quality of the substrate processing can be output in response to an input of the processing condition.

2. The method for assisting in setting conditions for substrate processing according to claim 1, comprising:

a step of causing the substrate processing apparatus to execute the substrate processing again in accordance with the recommended processing condition;

further acquiring additional actual data regarding the quality of the substrate processing performed under the recommended processing conditions;

a step of inputting an additional data set including the recommended processing condition and the additional actual data to the machine learning device; and

updating the recommended processing condition based on the updated learning model, which is updated by the machine learning device based on the additional data set.

3. The substrate processing condition setting assisting method according to claim 2, wherein:

further comprising the step of evaluating the recommended processing condition, and,

until the evaluation result of the recommended processing condition reaches a predetermined level, repeating the steps of:

a step of causing the substrate processing apparatus to execute the substrate processing again in accordance with the recommended processing condition;

further acquiring the additional actual data;

a step of inputting the additional data set to the machine learning device; and

updating the recommended processing condition based on the learning model updated by the machine learning device according to the additional data set.

4. A substrate processing condition setting assistance method according to any one of claims 1 to 3, characterized in that:

the step of deriving the recommended processing condition includes:

inputting an evaluation condition of the prediction data to the machine learning device; and

a step of acquiring the recommended processing condition derived by the machine learning device based on the plurality of sets of data sets, the learning model, and the evaluation condition.

5. The substrate processing condition setting assisting method according to claim 4, wherein:

retrieving the actual data comprising actual values of a plurality of items,

inputting the data set to the machine learning device that generates the learning model including a plurality of model formulas capable of outputting predicted values of the plurality of items, respectively, in response to input of the processing condition,

inputting the evaluation condition capable of evaluating the predicted values of the plurality of items to the machine learning device.

6. The substrate processing condition setting assisting method according to claim 5, wherein:

inputting the evaluation condition including a condition on fluctuation of predicted values of at least a part of the plurality of items to the machine learning apparatus.

7. The method for assisting in setting conditions for substrate processing according to any one of claims 1 to 6, wherein:

acquiring the actual data including an actual value of a post-process item indicating a quality of the substrate after the substrate is processed and an actual value of an in-process item indicating a supply state of the processing liquid to the substrate during the substrate processing,

a data set to be input to the machine learning device is selected based on actual values of the in-process items.

8. The method for assisting in setting conditions for substrate processing according to any one of claims 1 to 7, further comprising:

a step of removing components caused by causes other than the substrate processing from the actual data of the data set before inputting the data set to the machine learning device.

9. The method for assisting in setting conditions for substrate processing according to any one of claims 1 to 8, wherein:

the substrate processing includes a developing process of supplying a developing solution to the photosensitive coating film subjected to the exposure process on the surface of the substrate,

the actual data obtained includes an actual value of a line width of a pattern formed on the surface of the substrate by the developing process.

10. The method for assisting in setting conditions for substrate processing according to any one of claims 1 to 8, wherein:

the substrate processing includes a film-forming process of applying a film-forming solution on a surface of the substrate to form a coating film,

the acquired actual data includes an actual value of a film thickness of the coating film formed on the surface of the substrate by the film formation process.

11. A method for assisting condition setting in substrate processing, comprising:

a step of acquiring a data set including process conditions set for a substrate process including a process of supplying a process liquid to a substrate and actual data on quality of the substrate process performed according to the process conditions; and

a step of generating a learning model by machine learning based on the plurality of sets of the data, wherein the learning model is capable of outputting prediction data on the quality of the substrate processing in response to the input of the processing condition.

12. The substrate processing condition setting assisting method according to claim 11, further comprising:

and deriving a recommended process condition for the substrate process based on the evaluation conditions of the plurality of sets of data, the learning model, and the prediction data.

13. The substrate processing condition setting assisting method according to claim 12, wherein:

the step of generating the learning model by the machine learning includes an arithmetic process of searching the learning model by a genetic program,

the step of deriving the recommended processing condition includes an arithmetic process of searching for the recommended processing condition by a genetic algorithm.

14. The substrate processing condition setting assisting method according to claim 12 or 13, wherein:

taking the dataset with the actual data comprising actual values for a plurality of items,

generating the learning model including a plurality of model formulas capable of outputting predicted values of the plurality of items, respectively, in response to input of the processing condition,

deriving the recommended processing condition based on the plurality of sets of data, the plurality of model formulas, and the evaluation condition capable of evaluating predicted values of the plurality of items.

15. A substrate processing condition setting assisting method according to claim 14, wherein:

the recommendation processing condition is derived based on the evaluation condition including a condition that is a fluctuation condition with respect to the predicted values of the plurality of items.

16. A substrate processing system, comprising:

a processing unit for performing a substrate process including a process of supplying a processing liquid to a substrate;

a process control unit that causes the processing unit to execute the substrate process in accordance with a preset process condition;

a data acquisition unit that acquires actual data on the quality of the substrate processing performed according to the processing conditions;

a data input unit that inputs a data set including the processing conditions and the actual data to a model generation unit; and

a recommended condition deriving section that derives a recommended processing condition for the substrate processing based on a learning model generated by the model generating section through machine learning based on the plurality of sets of the data sets, the learning model being capable of outputting prediction data on quality of the substrate processing in response to an input of the processing condition.

17. The substrate processing system of claim 16, wherein:

the model generation part is also included.

18. The substrate processing system of claim 16 or 17, wherein:

the recommendation condition derivation unit includes:

an evaluation condition input unit that inputs an evaluation condition of the prediction data to a condition search unit; and

a search result acquisition unit that acquires the recommended processing condition derived by the condition search unit based on the plurality of sets of data sets, the learning model, and the evaluation condition.

19. The substrate processing system of claim 18, wherein:

the condition search section is also included.

20. A computer-readable storage medium characterized by:

a program for causing an apparatus to execute the condition setting assistance method for substrate processing according to any one of claims 1 to 15 is stored.

21. A learning model, characterized by:

the learning model is generated by machine learning based on a plurality of sets of data, causing an apparatus to perform: outputting prediction data regarding a quality of a substrate process in response to an input of a process condition of the substrate process,

wherein the plurality of sets of data each include a process condition of the substrate process performed by the substrate processing apparatus and including a process of supplying a process liquid to a substrate and actual data on a quality of the substrate process performed in accordance with the process condition.

Images