Learning apparatus, information processing apparatus, substrate processing apparatus, substrate processing system, learning method, and processing condition determination method

Abstract

The learning apparatus includes: an experimental data acquisition unit that drives a substrate processing apparatus under a processing condition including a variation condition indicating a relative position of a nozzle that varies with respect to a substrate over time, and performs processing of a film formed on the substrate; acquiring a first processing amount indicating the difference between the film thicknesses before and after the processing of the film; and a first conversion unit that converts the variation condition into compressed data indicating an operation amount relating to the nozzle for each of the plurality of movement sections. The plurality of movement sections are movement sections obtained by dividing the movement range of the nozzle in a scanning period from the start of nozzle operation to the end of the nozzle operation into a number less than the number of data of the variation condition; and a predictor generation unit that performs machine learning on learning data including the compressed data and the first processing amount corresponding to the processing condition, and generates a learning model that estimates the second processing amount.

Description

Technical Field

[0001] The present invention relates to a learning device, an information processing device, a substrate processing device, a substrate processing system, a learning method and a processing condition determination method, and also to a learning device that generates a learning model for simulating processing according to the processing conditions of a substrate processing device, an information processing device that uses the learning model to determine the processing conditions, a substrate processing device equipped with the information processing device, a substrate processing system equipped with the learning device and the information processing device, a learning method performed by the learning device, and a processing condition determination method performed by the information processing device. Background Art

[0002] There is a cleaning process in the semiconductor manufacturing process. In the cleaning process, the film thickness of the film formed on the substrate is adjusted by etching treatment in which a chemical solution is applied to the substrate. In this film thickness adjustment, it is important to perform the etching process in a manner that the surface of the substrate becomes uniform, or to make the surface of the substrate flat by etching. When the etching liquid is ejected from the nozzle to a part of the substrate, the nozzle must be moved radially relative to the substrate. However, the etching process is a complex process in which the amount of processing of the film changes depending on the difference in the action of moving the nozzle. Moreover, the amount of processing of the film by etching is clear after processing the substrate. Therefore, the operation of setting the action of moving the nozzle must be repeated by a technician. A lot of cost and time are required until the optimal action of the nozzle is determined.

[0003] Patent Document 1 describes a device that uses a learning model for machine learning using learning data with the "input" set as the processing amount (etching amount) and the "output" set as the scanning speed information to determine the scanning speed information according to the processing amount set as the target. According to this technology, one scanning speed information is determined according to the processing amount set as the target.

[0004] Patent Document 1: Japanese Patent Application Publication No. 2021-108367 Summary of the invention

[0005] Problem that the invention aims to solve

[0006] On the other hand, it is desirable to make the action of moving the nozzle more complex. The action of moving the nozzle is time series data representing a position that changes over time. If the action of moving the nozzle is complex, the sampling interval becomes shorter, so the dimension of the time series data becomes more. Generally speaking, if the dimension of the learning data increases, the number of data required for machine learning increases exponentially. Therefore, since the dimension of the learning data increases, it is difficult to optimize the learning model obtained by machine learning. In addition, since the etching process is a complex process, the action of the nozzle suitable for the target processing amount is not limited to one, and sometimes there are multiple actions.

[0007] One of the objects of the present invention is to provide a learning device, a learning method and a substrate processing system suitable for performing machine learning on conditions that change over time to process a substrate.

[0008] Furthermore, another object of the present invention is to provide an information processing device, a substrate processing device, a substrate processing system, and a processing condition determination method that can indicate a plurality of processing conditions for a processing result of a complex process of processing a substrate.

[0009] Means used to solve problems

[0010] A learning device according to one embodiment of the present invention includes: an experimental data acquisition unit that acquires a first processing amount representing a difference in film thickness before and after the film is processed by driving a substrate processing device that supplies a processing liquid to a substrate on which a film is formed to move a nozzle that supplies a processing liquid to the substrate under a processing condition including a variable condition representing a relative position of the nozzle that varies with time and relative to the substrate; a conversion unit that converts the variable condition into compressed data representing an action amount related to the nozzle for each of a plurality of movement intervals, wherein the plurality of movement intervals are movement ranges obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle action in which the substrate processing device moves the nozzle relative to the substrate into a smaller number of movement intervals than the number of data of the variable condition; and a model generation unit that performs machine learning on learning data including the compressed data and the first processing amount corresponding to the processing condition to generate a learning model, wherein the learning model infers a second processing amount representing a difference in film thickness before and after the film is processed for the film formed on the substrate before the film is processed by the substrate processing device.

[0011] An information processing device according to another embodiment of the present invention is used for managing a substrate processing device, wherein the substrate processing device processes a film formed on a substrate by supplying a processing liquid to a substrate on which a film is formed under a processing condition including a variable condition indicating a relative position of a nozzle that varies with time relative to the substrate. The information processing device includes: a conversion unit that converts the variable condition into compressed data indicating an amount of movement associated with the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range in which the nozzle moves during a scanning period from the start to the end of a nozzle movement in which the substrate processing device moves the nozzle relative to the substrate into a number of movement intervals smaller than the number of data of the variable condition; and a processing condition determination unit that determines a processing condition for driving the substrate processing device using a learning model, the learning model estimating a second processing amount indicating a difference in film thickness before and after the film is processed for the film formed on the substrate before the film is processed by the substrate processing device. The learning model is an inference model that performs machine learning on learning data, and the learning data includes: compressed data obtained by performing the same conversion as the conversion unit on the variable conditions included in the processing conditions for the substrate processing device to process the film formed on the substrate, and a first processing amount representing the difference in film thickness before and after the film is processed by the substrate processing device. The processing condition determination unit determines the processing condition including the temporary variable condition as the processing condition for driving the substrate processing device when the compressed data obtained by converting the temporary variable condition by the conversion unit is given to the learning model and the second processing amount estimated by the learning model satisfies the allowable condition.

[0012] A substrate processing system according to another embodiment of the present invention is used to manage a substrate processing device for processing a substrate, and includes a learning device and an information processing device. The substrate processing device processes a film formed on the substrate by supplying a processing liquid to a substrate on which a film is formed under processing conditions including a changing condition indicating a relative position of a nozzle relative to the substrate that changes over time. The learning device includes: an experimental data acquisition unit, which acquires a first processing amount representing the difference in film thickness before and after the film is processed by driving a substrate processing device under processing conditions to process a film formed on a substrate; a first conversion unit, which converts a variable condition into compressed data, the compressed data representing the same movement amount as the nozzle for each of a plurality of movement intervals whose number is smaller than the number of data of the variable condition, in which a movement range of the nozzle during a scanning period from the start to the end of a nozzle action of the substrate processing device that moves the nozzle relative to the substrate is divided into a smaller number than the number of data of the variable condition; and a model generation unit, which performs machine learning on learning data including the compressed data converted by the first conversion unit from the variable condition and the first processing amount corresponding to the processing conditions, and generates a learning model, the learning model inferring a second processing amount representing the difference in film thickness before and after the film is processed for the film formed on the substrate before the film is processed by the substrate processing device. The information processing device includes: a second conversion unit, which is the same as the first conversion unit; and a processing condition determination unit, which uses a learning model generated by a learning device to determine a processing condition for driving a substrate processing device. When the processing condition determination unit assigns a conversion result obtained by the second conversion unit converting the temporary change condition to the learning model and the second processing amount inferred by the learning model satisfies an allowable condition, the processing condition including the temporary change condition is determined as the processing condition for driving the substrate processing device.

[0013] A learning method in another embodiment of the present invention causes a computer to perform the following processing: after a substrate processing device that supplies a processing liquid to a substrate on which a film is formed is driven to move a nozzle that supplies a processing liquid to the substrate under a processing condition including a variable condition indicating a relative position of the nozzle that changes with time relative to the substrate, and processes the film formed on the substrate, a first processing amount representing the difference in film thickness before and after the film is processed is obtained; the variable condition is converted into compressed data, the compressed data representing an action amount related to the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement ranges that are obtained by dividing the movement range of the nozzle during a scanning period from the start to the end of the nozzle action of the substrate processing device moving the nozzle relative to the substrate into a smaller number of data than the number of data of the variable condition; and machine learning is performed on learning data for the first processing amount including the compressed data and the processing condition, to generate a learning model, the learning model inferring a second processing amount representing the difference in film thickness before and after the film is processed for the film formed on the substrate before the film is processed by the substrate processing device.

[0014] A processing condition determination method according to yet another embodiment of the present invention is executed by a computer managing a substrate processing device. The substrate processing device processes a film formed on a substrate by supplying a processing liquid to a substrate formed with a film under a processing condition including a variable condition indicating a relative position of a nozzle that varies with time relative to the substrate. The processing condition determination method includes the following processing: converting the variable condition into compressed data indicating a nozzle-related motion amount for each of a plurality of moving intervals, wherein the plurality of moving intervals are moving intervals obtained by dividing a moving range of the nozzle during a scanning period from the start to the end of a nozzle motion in which the substrate processing device moves the nozzle relative to the substrate into a number of moving intervals that is smaller than the number of data of the variable condition; and determining a processing condition for driving the substrate processing device using a learning model for a second processing amount that infers a difference in film thickness before and after the film is processed by the substrate processing device for the film formed on the substrate. The learning model is an inference model for machine learning of learning data, the learning data including: compressed data obtained by performing the same conversion process as the conversion process on the variable conditions included in the processing conditions for the substrate processing apparatus to process the film formed on the substrate, and a first processing amount representing the difference in film thickness before and after the film formed on the substrate after the substrate processing apparatus processes the film. The process of determining the processing condition determines the processing condition including the temporary variable condition as the processing condition for driving the substrate processing apparatus when the compressed data obtained by converting the temporary variable condition by the conversion process is given to the learning model and the second processing amount inferred by the learning model satisfies the permissible condition.

[0015] Effects of the Invention

[0016] The present invention can provide a learning device, a learning method and a substrate processing system suitable for performing machine learning on conditions that change over time to process a substrate.

[0017] Furthermore, it is possible to provide an information processing apparatus, a substrate processing apparatus, a substrate processing system, and a processing condition determination method that can present a plurality of processing conditions for a processing result of a complex process of processing a substrate. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a diagram for explaining the configuration of a substrate processing system according to one embodiment of the present invention.

[0019] Figure 2 A diagram showing an example of the configuration of an information processing device.

[0020] Figure 3 A diagram showing an example of the configuration of a learning device.

[0021] Figure 4This is a diagram showing an example of the functional configuration of a substrate processing system according to one embodiment of the present invention.

[0022] Figure 5 A diagram for explaining changes in the relative position of the nozzle with respect to the substrate.

[0023] Figure 6 This is a diagram showing an example of the operation mode of the nozzle.

[0024] Figure 7 This is a diagram showing an example of film thickness characteristics.

[0025] Figure 8 A diagram for explaining segmented regions.

[0026] Fig. 9 This is a diagram showing an example of compressed data.

[0027] Fig.10 A diagram illustrating the predictor.

[0028] Fig.11 This is a flowchart showing an example of the flow of the predictor generation process.

[0029] Fig.12 This is a flowchart showing an example of the flow of the processing condition determination process.

[0030] Fig.13 This is a flowchart showing an example of the flow of compressed data generation processing.

[0031] Fig.14 This is a flowchart showing an example of the flow of the additional learning process.

[0032] Fig.15 This is a diagram showing an example of the supply amount of the processing liquid that changes with the passage of time.

[0033] Fig.16 This is a diagram showing another example of compressed data. DETAILED DESCRIPTION

[0034] Hereinafter, a substrate processing system according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings. In the following description, substrate refers to a semiconductor substrate (semiconductor wafer), a FPD (Flat Panel Display) substrate of a liquid crystal display device or an organic EL (Electro Luminescence) display device, a disc substrate, a magnetic disk substrate, an optical magnetic disk substrate, a photomask substrate, a ceramic substrate, or a solar cell substrate.

[0035] 1. Overall composition of substrate processing system

[0036] Figure 1 It is a diagram for explaining the configuration of a substrate processing system according to one embodiment of the present invention. Figure 1 The substrate processing system 1 includes an information processing apparatus 100, a learning apparatus 200, and a substrate processing apparatus 300. The learning apparatus 200 is, for example, a server, and the information processing apparatus 100 is, for example, a personal computer.

[0037] The learning apparatus 200 and the information processing apparatus 100 are used to manage the substrate processing apparatus 300. In addition, the substrate processing apparatus 300 managed by the learning apparatus 200 and the information processing apparatus 100 is not limited to one, and a plurality of substrate processing apparatuses 300 may be managed.

[0038] In the substrate processing system 1 of the present embodiment, the information processing device 100, the learning device 200 and the substrate processing device 300 are connected to each other through a wired or wireless communication line or communication line network. The information processing device 100, the learning device 200 and the substrate processing device 300 are each connected to a network and can send and receive data to each other. The network uses, for example, a local area network (LAN) or a wide area network (WAN). Moreover, the network can be the Internet. Moreover, the information processing device 100 and the substrate processing device 300 can be connected by a dedicated communication line network. The connection form of the network can be a wired connection or a wireless connection.

[0039] In addition, the learning device 200 is not necessarily connected to the substrate processing device 300 and the information processing device 100 by a communication line or a communication line network. In this case, the data generated by the substrate processing device 300 can be delivered to the learning device 200 via a storage medium. In addition, the data generated by the learning device 200 can be delivered to the information processing device 100 via a storage medium.

[0040] A display device, a sound output device, and an operation unit (not shown) are provided in the substrate processing apparatus 300. The substrate processing apparatus 300 operates according to a default processing condition (processing recipe) of the substrate processing apparatus 300.

[0041] 2. Overview of substrate processing equipment

[0042] The substrate processing device 300 includes a control device 10 and a plurality of substrate processing units WU. The control device 10 controls the plurality of substrate processing units WU. The plurality of substrate processing units WU processes the substrate by supplying a certain flow rate of processing liquid to the substrate W on which a film has been formed. In the present embodiment, the substrate W to be processed has a diameter of 300 mm, but the present invention is not limited thereto. The processing liquid includes an etching liquid, and the substrate processing unit WU performs an etching process. The etching liquid is a chemical liquid. For example, the etching liquid is a mixture of fluorinated nitric acid (fluoric acid (HF) and nitric acid (HNO 3) mixture), fluoric acid, buffered hydrofluoric acid (BHF), ammonium fluoride, HFEG (a mixture of fluoric acid and ethylene glycol), or phosphoric acid (H 3 PO 4 ).

[0043] The substrate processing unit WU includes a rotary chuck SC, a rotary motor SM, a nozzle 311, and a nozzle moving mechanism 301. The rotary chuck SC holds the substrate W horizontally. The substrate W is held by the rotary chuck SC in a manner that the first rotary axis AX1 of the rotary motor SM coincides with the center of the substrate W. The rotary motor SM has a first rotary axis AX1. The first rotary axis AX1 extends in the up-down direction. The rotary chuck SC is mounted on the upper end of the first rotary axis AX1 of the rotary motor SM. When the rotary motor SM rotates, the rotary chuck SC rotates around the first rotary axis AX1. The rotary motor SM is a stepping motor. The substrate W held on the rotary chuck SC rotates around the first rotary axis AX1. Therefore, the rotation speed of the substrate W is the same as the rotation speed of the stepping motor. In addition, in the case where an encoder that generates a rotation speed signal indicating the rotation speed of the rotary motor is provided, the rotation speed of the substrate W can be obtained from the rotation speed signal generated by the encoder. In this case, the rotary motor SM can use a motor other than a stepping motor.

[0044] The nozzle 311 supplies the etching liquid to the substrate W. The nozzle 311 is supplied with the etching liquid from an etching liquid supply unit (not shown) and discharges the etching liquid to the substrate W being rotated.

[0045] The nozzle moving mechanism 301 moves the nozzle 311 in a substantially horizontal direction. Specifically, the nozzle moving mechanism 301 includes: a nozzle motor 303 having a second rotation axis AX2, and a nozzle arm 305. The nozzle motor 303 is configured so that the second rotation axis AX2 is along a substantially vertical direction. The nozzle arm 305 has a longitudinal shape extending in a straight line. One end of the nozzle arm 305 is mounted on the upper end of the second rotation axis AX2 in a manner such that the length direction of the nozzle arm 305 is in a direction different from the second rotation axis AX2. The nozzle 311 is mounted on the other end of the nozzle arm 305 in a manner such that its nozzle is directed downward.

[0046] When the nozzle motor 303 is in operation, the nozzle arm 305 rotates in a horizontal plane around the second rotation axis AX2. Therefore, the nozzle 311 mounted on the other end of the nozzle arm 305 moves (rotates) in a horizontal direction around the second rotation axis AX2. The nozzle 311 moves in a horizontal direction while ejecting etching liquid onto the substrate W. The nozzle motor 303 is, for example, a stepping motor.

[0047] The control device 10 includes a CPU (central processing unit) and a memory, and the CPU executes a program stored in the memory to control the entire substrate processing apparatus 300. The control device 10 controls the rotary motor SM and the nozzle motor 303.

[0048] The learning device 200 receives the experimental data from the substrate processing apparatus 300 , performs machine learning on the learning model using the experimental data, and outputs the learned learning model to the information processing apparatus 100 .

[0049] The information processing apparatus 100 uses the learned learning model to determine the processing conditions for processing a substrate to be processed next by the substrate processing apparatus 300. The information processing apparatus 100 outputs the determined processing conditions to the substrate processing apparatus 300.

[0050] Figure 2 FIG. 1 is a diagram showing an example of the configuration of an information processing device. Figure 2 The information processing device 100 is composed of a CPU 101, a RAM (random access memory) 102, a ROM (read only memory) 103, a storage device 104, an operation unit 105, a display device 106, and an input / output I / F (interface) 107. The CPU 101, the RAM 102, the ROM 103, the storage device 104, the operation unit 105, the display device 106, and the input / output I / F 107 are connected to a bus 108.

[0051] RAM 102 is used as a work area of ​​CPU 101. System programs are stored in ROM 103. Storage device 104 includes a storage medium such as a hard disk or a semiconductor memory, and stores programs. Programs can be stored in ROM 103 or other external storage devices.

[0052] The CD-ROM 109 can be detached from the storage device 104. The storage medium for storing the program executed by the CPU 101 is not limited to the CD-ROM 109, and may be a medium such as an optical disk (MO (Magnetic Optical Disc) / MD (Mini Disc) / DVD (Digital Versatile Disc)), an IC card, an optical card, a mask ROM, an EPROM (Erasable Programmable ROM), or other semiconductor memory. Furthermore, it is feasible that the CPU 101 downloads the program from a computer connected to the network and stores it in the storage device 104, or writes the program to the storage device 104 from the computer connected to the network, loads the program stored in the storage device 104 into the RAM 102, and executes it by the CPU 101. The program mentioned here includes the original program, the compressed program, the encrypted program, etc., and is not only a program that can be directly executed by the CPU 101.

[0053] The operation unit 105 is an input device such as a keyboard, a mouse or a touch panel. The user can give a predetermined instruction to the information processing device 100 by operating the operation unit 105. The display device 106 is a display device such as a liquid crystal display device, and displays a GUI (Graphical User Interface) for accepting user instructions. The input / output I / F 107 is connected to a network.

[0054] Figure 3 FIG. 1 is a diagram showing an example of the configuration of a learning device. Figure 3 The learning device 200 is composed of a CPU 201 , a RAM 202 , a ROM 203 , a storage device 204 , an operation unit 205 , a display device 206 , and an input / output I / F 207 . The CPU 201 , the RAM 202 , the ROM 203 , the storage device 204 , the operation unit 205 , the display device 206 , and the input / output I / F 207 are connected to a bus 208 .

[0055] RAM 202 is used as a work area of ​​CPU 201. ROM 203 stores a system program. Storage device 204 includes a storage medium such as a hard disk or a semiconductor memory, and stores a program. The program can be stored in ROM 203 or other external storage devices. CD-ROM 209 can be detached from storage device 204.

[0056] The operation unit 205 is an input device such as a keyboard, a mouse, or a touch panel. The input / output I / F 207 is connected to a network.

[0057] 3. Functional composition of substrate processing system

[0058] Figure 4 FIG. 1 is a diagram showing an example of the functional configuration of a substrate processing system according to one embodiment of the present invention. Figure 4 The control device 10 of the substrate processing device 300 controls the substrate processing unit WU to process the substrate W according to the processing conditions. The processing conditions are conditions for processing the substrate W during a preset processing time. The processing time is the time determined for processing the substrate. In this embodiment, the processing time is the time between the nozzle 311 spraying the etching solution onto the substrate W.

[0059] The processing conditions include: the temperature of the etching liquid, the concentration of the etching liquid, the flow rate of the etching liquid, the rotation speed of the substrate W, and the relative position of the nozzle 311 and the substrate W. The processing conditions include variable conditions that change with the passage of time. In this embodiment, the variable condition is the relative position of the nozzle 311 and the substrate W. The relative position is represented by the rotation angle of the nozzle motor 303. The processing conditions include fixed conditions that do not change with the passage of time. In this embodiment, the fixed conditions are the temperature of the etching liquid, the concentration of the etching liquid, the flow rate of the etching liquid, and the rotation speed of the substrate W.

[0060] The learning device 200 causes the learning model to learn the learning data, and generates an inference model that estimates the etching profile from the processing conditions. Hereinafter, the inference model generated by the learning device 200 is referred to as a predictor.

[0061] The learning device 200 includes an experimental data acquisition unit 261, a first conversion unit 263, a predictor generation unit 265, and a predictor transmission unit 267. The functions of the learning device 200 are implemented by the CPU 201 of the learning device 200 executing a learning program stored in the RAM 202.

[0062] The experimental data acquisition unit 261 acquires experimental data from the substrate processing apparatus 300. The experimental data includes processing conditions used when the substrate processing apparatus 300 actually processes the substrate W, and film thickness characteristics of a film formed on the substrate W before and after the processing.

[0063] The film thickness characteristics are expressed by the film thickness of the film formed on the substrate W at each of a plurality of different positions in the radial direction of the substrate W.

[0064] The changing condition includes the relative position of the nozzle 311 with respect to the substrate W which changes over time. The substrate W rotates around the first rotation axis AX1, and the nozzle 311 rotates around the second rotation axis AX2. Therefore, the change in the relative position between the nozzle 311 and the substrate W is represented by the change in the position of the nozzle 311. The position of the nozzle 311 is determined by the rotation angle of the nozzle motor 303. Moreover, the range of the rotation angle of the nozzle motor 303 is limited to a specified range. Furthermore, the processing time is a preset period. In the present embodiment, the processing time is 60 seconds.

[0065] Figure 5 : is a diagram for explaining the change in the relative position of the nozzle with respect to the substrate. Figure 5 , showing the change in the relative position of the nozzle 311 with respect to the substrate W held on the rotary chuck SC. The nozzle 311 moves in the area above the substrate W held on the rotary chuck SC. Since the nozzle rotates around the second rotation axis AX2, the trajectory of the movement of the nozzle 311 is an arc. The trajectory of the movement of the nozzle 311 passes through the center of the substrate, that is, the substrate center OP. Therefore, the nozzle 311 moves from the substrate center OP to the entire peripheral portion in the radial direction of the substrate W. Here, with respect to the trajectory of the movement of the nozzle 311, one end is represented by an action end EP1 that is closer to the inner side than the peripheral portion of the substrate W, and the other end is represented by an action end EP2 that is closer to the inner side than the peripheral portion of the substrate W. The scan from the action end EP1 of the nozzle 311 to the center OP of the substrate is indicated by arrow a1, the scan from the center OP of the substrate to the action end EP2 of the nozzle 311 is indicated by arrow a2, the scan from the action end EP2 of the nozzle 311 to the center OP of the substrate is indicated by arrow a3, and the scan from the center OP of the substrate to the action end EP1 of the nozzle 311 is indicated by arrow a4.

[0066] Figure 6 FIG. 1 is a diagram showing an example of the operation mode of the nozzle. Figure 6 In , the vertical axis represents the relative position of the nozzle 311 relative to the substrate W, and the horizontal axis represents the elapsed time (seconds). In the present embodiment, the scanning period from the start to the end of the nozzle action of moving the nozzle 311 relative to the substrate W is equal to the processing time. Since the processing time is set to 60 seconds as mentioned above, the action pattern of the nozzle represents the relative position during the period of 0 to 60 seconds. The relative position of the nozzle is such that the substrate center OP is set to zero, the range from the substrate center OP to the action end EP1 is represented by a negative value, and the range from the substrate center OP to the action end EP2 is represented by a positive value. Since the substrate W has a diameter of 300 mm, the distance from the substrate center OP to the action ends EP1 and EP2 is set to less than ±150 mm. Here, the distance from the substrate center OP to the action end EP1 is set to -147 mm, and the distance from the substrate center OP to the action end EP2 is set to +147 mm. In Figure 6 In the nozzle action mode, the relative position of the nozzle 311 when the nozzle 311 is located at the center OP of the substrate is represented by 0, the relative position of the nozzle 311 when the nozzle 311 is located at the action end EP1 is represented by -147mm, and the relative position of the nozzle 311 when the nozzle 311 is located at the action end EP2 is represented by 147mm.

[0067] Figure 6The nozzle operation mode shown in FIG. 1 shows five reciprocating scans between the operation end EP1 and the operation end EP2. Figure 5 Parts of relative positions corresponding to scans indicated by arrows a1 to a4 are shown with the same reference numerals.

[0068] Figure 7 This is a diagram showing an example of film thickness characteristics. Figure 7 , the horizontal axis represents the position in the radial direction of the substrate, and the vertical axis represents the film thickness. The origin of the horizontal axis represents the center of the substrate. The film thickness of the film formed on the substrate W before being processed by the substrate processing device 300 is represented by a solid line. The substrate processing device 300 performs a process of applying an etching solution according to the processing conditions, thereby adjusting the film thickness of the film formed on the substrate W. The film thickness of the film formed on the substrate W after being processed by the substrate processing device 300 is represented by a dotted line.

[0069] The difference between the film thickness of the film formed on the substrate W before being processed by the substrate processing apparatus 300 and the film thickness of the film formed on the substrate W after being processed by the substrate processing apparatus 300 is the processing amount (etching amount). The processing amount represents the thickness of the film reduced by the processing of applying the etching solution by the substrate processing apparatus 300. The radial distribution of the processing amount is called an etching profile. The etching profile includes the processing amount at each of a plurality of radial positions of the substrate W.

[0070] Moreover, the film thickness formed by the substrate processing apparatus 300 is preferably uniform over the entire surface of the substrate W. Therefore, a target film thickness is determined for the processing performed by the substrate processing apparatus 300. The target film thickness is indicated by a dotted line. The deviation characteristic is the difference between the film thickness formed on the substrate W after processing by the substrate processing apparatus 300 and the target film thickness. The deviation characteristic includes the difference at each of a plurality of positions in the radial direction of the substrate W.

[0071] return Figure 4 The first conversion unit 263 converts the variable conditions included in the processing conditions of the experimental data input from the experimental data acquisition unit 261 into low-dimensional compressed data. Here, the variable conditions are the relative position of the nozzle 311 with respect to the substrate W that changes with the passage of time. The first conversion unit 263 outputs the compressed data to the predictor generation unit 265.

[0072] The compressed data indicates the amount of movement related to the nozzle for each of a plurality of movement intervals, and the plurality of movement intervals are obtained by dividing at least a part of the movement range in which the nozzle 311 moves during the scanning period from the start to the end of the nozzle movement of moving the nozzle 311 relative to the substrate W into a number of movement intervals smaller than the number of data of the variation condition. The amount of movement is the time that the nozzle 311 stays in each of the plurality of movement intervals. Here, the compressed data is described.

[0073] Figure 8 Diagram for explaining segmentation areas. Figure 8 , showing 15 divided areas b1 to b15 divided by a plurality of concentric circles centered on the substrate center OP from the upper surface of the substrate W. The divided area b15 is a circle, and the divided areas b1 to b14 are annular rings. The radial length of the substrate W of each of the plurality of divided areas b1 to b14 is the same. The radial length of the substrate of each of the divided areas b1 to b14 is the difference between the radius of the outer circumference and the radius of the inner circumference. The radius of the divided area b15 is the same as the radial length of the substrate W of each of the plurality of divided areas b1 to b14. Here, the radius of the divided area b15 is 10 mm, and the difference between the radius of the outer circumference and the radius of the inner circumference of each of the divided areas b1 to b14 is 10 mm. The difference between the radius of the outer circumference and the inner circumference of each of the divided areas b1 to b14 and the radius of the divided area b15 are larger than the inner diameter of the nozzle 311. The radial length of the substrate of each of the divided areas b1 to b15 and the radius of the divided area b15 are preferably larger than the inner diameter of the nozzle 311.

[0074] Since the nozzle 311 rotates around the second rotation axis AX2, the rotation center is different from the substrate center OP. The movement range of the nozzle 311 is a trajectory drawn during the movement from the operating end EP1 through the substrate center OP to the operating end EP2, and is an arc.

[0075] The moving range of the nozzle 311 is the trajectory of the nozzle 311 during the processing period (scanning period) in which the nozzle 311 processes the substrate. The moving range is divided into 30 moving intervals d1 to d30 by the divided areas b1 to b15. The moving intervals d1 to d15 are intervals that intersect each of the divided areas b1 to b15 in the trajectory of the nozzle 311 moving between the action end EP1 and the substrate center OP. For example, the moving interval d1 is an interval that intersects the divided area b1 during the period when the nozzle 311 moves between the action end EP1 and the substrate center OP. Moreover, the moving intervals d16 to d30 are intervals that intersect each of the divided areas b1 to b15 in the trajectory during the period when the nozzle 311 moves between the action end EP2 and the substrate center OP. For example, the moving interval d30 is an interval that intersects the divided area b1 during the period when the nozzle 311 moves between the action end EP2 and the substrate center OP. In addition, the number of divided areas b1 to b15 is not limited to 15, and can be set to any value. In this case, the number of divided movement ranges, in other words, the number of movement sections is different.

[0076] Fig. 9 is a diagram showing an example of compressed data. Fig. 9In FIG. 1 , the horizontal axis represents the position on the substrate W. The position of the substrate center OP is represented by 0 mm, one radial end of the substrate W is represented by -150 mm, and the other radial end of the substrate W is represented by 150 mm. The movement intervals d1 to d30 are allocated between -150 mm and +150 mm on the horizontal axis.

[0077] The vertical axis represents the residence time of the nozzle 311 in each of the moving sections d1 to d30. Figure 6 The residence time of each of the moving sections d1 to d30 when the operation mode shown is moved. The residence time is the cumulative time that the nozzle 311 is located in each of the plurality of moving sections d1 to d30. Figure 6 When the nozzle 311 moves in the illustrated operation pattern, the nozzle 311 crosses the movement section d2 ten times. The residence time in the movement section d2 is the cumulative time that the nozzle 311 crosses the movement section d2.

[0078] As described above, the moving range of the nozzle 311 is divided into a plurality of moving intervals d1 to d30. Therefore, the residence time of the nozzle 311 in each of the plurality of divided areas b1 to b15 including information on the radial position of the substrate W is calculated. Therefore, the residence time in each of the plurality of divided areas b1 to b15 is information on the radial position of the substrate W. Moreover, the radial lengths of the portions of the substrate W that are cross-cut by the nozzle 311 in each of the plurality of divided areas b1 to b15 are the same. Therefore, the residence time of the nozzle 311 in each of the plurality of divided areas b1 to b15 can be set to a time in which there is no deviation between different positions in the radial direction of the substrate W with respect to changes in the relative position of the nozzle 311 with respect to the substrate W.

[0079] In addition, in this embodiment, since the upper surface of the substrate W is divided into 15 divided areas b1 to b15, the number of compressed data is 30. The larger the radial length of each of the divided areas b1 to b15, the smaller the number of compressed data. Since the radial length of each of the divided areas b1 to b15 is greater than the inner diameter of the nozzle 311, the maximum value of the number of compressed data is determined by the inner diameter of the nozzle 311.

[0080] return Figure 4 The predictor generation unit 265 receives the compressed data after the variable condition is converted from the first conversion unit 263 and receives the experimental data from the experimental data acquisition unit 261. The predictor generation unit 265 generates the predictor by causing the neural network to perform teaching learning. In addition, the neural network may be a convolutional neural network.

[0081] Specifically, the learning data includes input data and correct answer data. The input data includes: compressed data converted by the first conversion unit 263 from variable conditions, and fixed conditions other than the variable conditions of the processing conditions contained in the experimental data. The correct answer data includes an etching profile. The etching profile is the difference between the film thickness characteristics of the film before processing contained in the experimental data and the film thickness characteristics of the film after processing contained in the experimental data. The etching profile contained in the correct answer data is an example of the first processing amount. The predictor generation unit 265 inputs the input data to the neural network and determines the parameters of the neural network in such a way that the output of the neural network is equal to the correct answer data. The predictor generation unit 265 generates a neural network as a predictor that incorporates the parameters set in the neural network after learning. The predictor is an inference program that incorporates the parameters set in the neural network after learning. The predictor generation unit 265 sends the predictor to the information processing device 100.

[0082] Fig.10 is a diagram illustrating the predictor. Fig.10 , the predictor includes an input layer, an intermediate layer and an output layer, and each layer includes a plurality of nodes represented by 0. In addition, one intermediate layer is shown in the figure, but the number of intermediate layers can be more than one. Moreover, five nodes are shown in the input layer, four nodes are shown in the intermediate layer, and three nodes are shown in the output layer, but the number of nodes is not limited to this. The output of the upper node is connected to the input of the lower node. The parameter includes a coefficient for weighting the output of the upper node. Moreover, the number of intermediate layers is more than one, and its number is unlimited.

[0083] If the compressed data after the variable conditions are converted into a low-dimensional data set and the fixed conditions are input to the predictor, an etching profile is output. The etching profile output by the predictor is an example of the second processing amount. The etching profile is represented by the difference E[n] of the film thickness before and after the processing at each of the plurality of radial positions P[n] (n is an integer greater than 1) of the substrate W. In addition, the number of output nodes of the predictor is shown as 3 in the figure, but the number of output nodes is actually n.

[0084] return Figure 4 The information processing device 100 includes a processing condition determination unit 151, a predictor receiving unit 155, a second conversion unit 157, a prediction unit 159, an evaluation unit 161, and a processing condition sending unit 163. The functions of the information processing device 100 are implemented by the CPU 101 of the information processing device 100 executing a processing condition determination program stored in the RAM 102.

[0085] The predictor receiving unit 155 receives the predictor transmitted from the learning device 200 , and outputs the received predictor to the predicting unit 159 .

[0086] The processing condition determination unit 151 determines the processing condition for the substrate W to be processed by the substrate processing apparatus 300. The processing condition determination unit 151 outputs the variable condition included in the processing condition to the second conversion unit 157, and outputs the fixed condition included in the processing condition to the prediction unit 159. The processing condition determination unit 151 selects one from a plurality of variable conditions prepared in advance using an experimental design method, a pairwise test, or a Bayesian estimation, and determines the processing condition including the selected variable condition and the fixed condition as the processing condition for the prediction unit 159 to perform estimation. The plurality of variable conditions prepared in advance are preferably the plurality of variable conditions for generating a compressor generated by the learning device 200.

[0087] The second converter 157 has the same function as the first converter 263 of the above-mentioned learning device 200. The second converter 157 converts the change condition input from the processing condition determination unit 151 into compressed data. The second converter 157 outputs the converted compressed data to the prediction unit 159.

[0088] The predictor 159 estimates the etching profile based on the compressed data and the fixed conditions using a predictor. Specifically, the predictor 159 inputs the compressed data input from the second conversion unit 157 and the fixed conditions input from the processing condition determination unit 151 to the predictor, and outputs the etching profile output by the predictor to the evaluation unit 161.

[0089] The evaluation unit 161 evaluates the etching profile input from the prediction unit 159, and outputs the evaluation result to the processing condition determination unit 151. In detail, the evaluation unit 161 obtains the film thickness characteristics before processing of the predetermined substrate W as the processing object of the substrate processing device 300. The evaluation unit 161 calculates the film thickness characteristics predicted after the etching process based on the etching profile input from the prediction unit 159 and the film thickness characteristics of the substrate W before processing, and compares it with the film thickness characteristics set as the target. If the result of the comparison meets the evaluation criterion, the processing conditions determined by the processing condition determination unit 151 are output to the processing condition sending unit 163. For example, the evaluation unit 161 calculates the deviation characteristic and determines whether the deviation characteristic meets the evaluation criterion. The deviation characteristic is the difference between the film thickness characteristics of the substrate W after the etching process and the target film thickness characteristics. The evaluation criterion can be determined arbitrarily. For example, the evaluation criterion can be that the maximum value of the difference in the deviation characteristic is below the threshold value, or that the average value of the difference is below the threshold value.

[0090] The processing condition transmitting unit 163 transmits the processing condition determined by the processing condition determining unit 151 to the substrate processing apparatus 300. The substrate processing apparatus 300 processes the substrate W according to the processing condition.

[0091] When the evaluation result does not satisfy the evaluation criterion, the evaluation unit 161 outputs the evaluation result to the processing condition determination unit 151. The evaluation result includes the film thickness characteristic predicted after the etching process or the difference between the film thickness characteristic predicted after the etching process and the target film thickness characteristic.

[0092] The processing condition determination unit 151 determines a new processing condition for the prediction unit 159 to make an inference based on the evaluation result input from the self-evaluation unit 161. The processing condition determination unit 151 selects one from a plurality of variable conditions prepared in advance using an experimental design method, a pairwise test, or a Bayesian inference, and determines a processing condition including the selected variable condition and a fixed condition as a new processing condition for the prediction unit 159 to make an inference.

[0093] The processing condition determination unit 151 may use Bayesian estimation to explore the processing conditions. When the evaluation unit 161 outputs a plurality of evaluation results, there are a plurality of sets of processing conditions and evaluation results. Based on the tendency of the etching profiles of each of the plurality of sets, the processing conditions that make the film thickness uniform or the processing conditions that minimize the difference between the predicted film thickness characteristics after the etching process and the target film thickness characteristics are explored.

[0094] Specifically, the processing condition determination unit 151 explores the processing conditions in a manner that minimizes the objective function. The objective function is a function that represents the uniformity of the film thickness of the film or a function that represents the consistency between the film thickness characteristics of the film and the target film thickness characteristics. For example, the objective function is a function that represents the difference between the film thickness characteristics predicted after the etching process and the target film thickness characteristics by parameters. The parameters here are compressed data obtained by converting the corresponding variable conditions by the second conversion unit 157. The corresponding variable conditions are variable conditions before the compressed data used by the predictor to infer the etching profile is converted. The processing condition determination unit 151 selects a variable condition corresponding to the parameter determined by the exploration, i.e., the compressed data, from a plurality of variable conditions, and determines a new processing condition including the selected variable condition and a fixed condition.

[0095] Fig.11 2 is a flowchart showing an example of the flow of the predictor generation process. The predictor generation process is a process performed by the CPU 201 of the learning device 200 when the CPU 201 executes the predictor generation program stored in the RAM 202. The predictor generation program is a part of the learning program.

[0096] Reference Fig.11, the CPU 201 of the learning device 200 obtains experimental data. The CPU 201 controls the input / output I / F 107 to obtain experimental data from the substrate processing device 300 (step S01). The experimental data can be obtained by reading the experimental data recorded in a storage medium such as CD-ROM 209 using the storage device 104. The experimental data obtained here are plural. The experimental data includes processing conditions and film thickness characteristics of the film formed on the substrate W before and after the processing. The film thickness characteristics are represented by the film thickness of the film formed on the substrate W at different plural positions in the radial direction of the substrate W.

[0097] In the subsequent step S02, experimental data to be processed is selected, and the process proceeds to step S03. In step S03, compressed data generation processing is executed, and the process proceeds to step S04. The compressed data generation processing is described in detail later, and is a process for converting the variation conditions included in the experimental data into compressed data.

[0098] In step S04, the compressed data, the fixed conditions contained in the experimental data, and the etching profile are set as learning data. The etching profile is the difference between the film thickness characteristics of the film before processing contained in the experimental data and the film thickness characteristics of the film after processing contained in the experimental data. The learning data includes input data and correct answer data. The compressed data converted in step S03 and the fixed conditions contained in the experimental data are set as input data. The etching profile is set as the correct answer data.

[0099] In the subsequent step S05, the CPU 201 causes the predictor to perform machine learning, and the process proceeds to step S06. The input data is input to the neural network, i.e., the predictor, and the parameters are determined in such a way that the output of the predictor is equal to the correct answer data. Therefore, the parameters of the predictor are adjusted. The predictor is a neural network having parameters determined by machine learning using learning data. In addition, the neural network may be a convolutional neural network.

[0100] In step S06, it is determined whether the adjustment is completed. The learning data used for the evaluation of the predictor is prepared in advance, and the performance of the predictor is evaluated using the learning data for evaluation. When the evaluation result meets the preset evaluation benchmark, it is determined that the adjustment is completed. When the evaluation result does not meet the evaluation benchmark (No in step S06), the process returns to step S02, and when the evaluation result meets the evaluation benchmark (Yes in step S06), the process proceeds to step S07.

[0101] When the processing returns to step S02, in step S02, experimental data that has not been selected as the processing object is selected from the experimental data obtained in step S01. In the loop of steps S02 to S06, CPU201 uses a plurality of learning data to make the predictor perform machine learning. Therefore, the parameters of the neural network, i.e., the predictor, are adjusted to appropriate values. In step S08, the predictor is sent, and the processing ends. CPU201 controls the input / output I / F107 to send the predictor to the information processing device 100.

[0102] Fig.12 1 is a flowchart showing an example of the flow of the processing condition determination process. The processing condition determination process is a process executed by the CPU 101 of the information processing device 100 when the CPU 101 executes a processing condition determination program stored in the RAM 102 .

[0103] Reference Fig.12 , the CPU 101 of the information processing device 100 selects one of the plurality of change conditions prepared in advance (step S11), and proceeds to step S12. The plurality of change conditions are the plurality of change conditions generated by the learning device 200 to generate a compressor. One of the plurality of change conditions prepared in advance is selected using an experimental design method, a pairwise test, or a Bayesian estimation.

[0104] In step S12, a compressed data generation process is executed, and the process proceeds to step S 13. The details of the compressed data generation process will be described later.

[0105] In step S13, a predictor is used to estimate the etching profile based on the compressed data and fixed conditions, and the process proceeds to step S14. The compressed data and fixed conditions generated in step S12 are input to the predictor, and the etching profile output by the predictor is obtained. In step S14, the film thickness characteristics after processing are compared with the target film thickness characteristics. Based on the film thickness characteristics before processing of the substrate W that becomes the processing object of the substrate processing device 300 and the etching profile estimated in step S13, the film thickness characteristics after processing of the substrate W are calculated. Then, the film thickness characteristics after processing are compared with the target film thickness characteristics. Here, the difference between the film thickness characteristics after processing of the substrate W and the target film thickness characteristics is calculated.

[0106] In step S15, it is determined whether the comparison result satisfies the evaluation criterion. When the comparison result does not satisfy the evaluation criterion (yes in step S15), the process proceeds to step S16, otherwise the process returns to step S11. For example, when the maximum value of the difference is below the threshold value, it is determined that the evaluation criterion is satisfied. In addition, when the average value of the difference is below the threshold value, it is determined that the evaluation criterion is satisfied.

[0107] In step S16, the processing conditions including the variable conditions just selected in step S11 are set as candidates for the processing conditions for driving the substrate processing apparatus 300, and the process proceeds to step S17. In step S17, it is determined whether an end instruction of the search is received. When the end instruction is received by the user operating the information processing apparatus 100, the process proceeds to step S18, otherwise the process returns to step S11. In addition, it can be determined whether a preset number of processing conditions are set as candidates instead of the end instruction input by the user.

[0108] In step S18, one is determined from one or more processing conditions set as candidates, and the processing proceeds to step S19. A user operating the information processing device 100 can select one from one or more processing conditions set as candidates. Therefore, the range of user selection is expanded. Moreover, a variable condition with the simplest nozzle movement can be automatically selected from the variable conditions contained in the plurality of processing conditions. The variable condition with the simplest nozzle movement can be set, for example, as a variable condition with the least number of speed change points. Therefore, a plurality of variable conditions can be prompted for the processing result of a complex nozzle movement for processing the substrate W. If a variable condition that is easy to control the nozzle is selected from the plurality of variable conditions, it is easy to control the substrate processing device 300.

[0109] In step S19, the processing conditions including the variable conditions determined in step S18 are sent to the substrate processing apparatus 300, and the processing ends. The CPU 101 controls the input / output I / F 107 to send the processing conditions to the substrate processing apparatus 300. When the substrate processing apparatus 300 receives the processing conditions from the information processing apparatus 100, it processes the substrate W according to the processing conditions.

[0110] Fig.13 FIG. 1 is a flowchart showing an example of the process of compressed data generation processing. Fig.11 Step S03 or Fig.12 In step S21, time series data representing the time series of the movement of the nozzle 311 is obtained. In step S22, the number of divisions K (K is an integer greater than 2) used to set a plurality of movement intervals is obtained from the time series data. In this embodiment, the number of divisions K is preset. In addition, the number of divisions K can be input by the user via an operation unit, etc. In addition, Figure 6 In the example, the number of divisions K is set to 30.

[0111] In step S23, the variable n is set to 1. In step S24, the residence time of the nozzle 311 in the moving interval d(n) is calculated. The moving interval d(n) represents the nth moving interval in the moving intervals d1 to d30. When calculating the residence time of the nozzle 311 in the moving interval d(n), in step S25, 1 is added to the variable n. At this time, in step S26, it is determined whether the variable n is greater than the number of divisions K. When the variable n is less than the number of divisions K, the processing returns to step S24. When the variable n is greater than the number of divisions K, in step S27, the calculated residence time of the nozzle 311 in the moving intervals d(1) to d(K) is set as the compressed data. By repeating steps S24 to S26, the residence time of the nozzle 311 in each of the moving intervals d(1) to d(K) is calculated.

[0112] 4. Specific examples

[0113] In this embodiment, the change condition is a time series data sampled with a processing time of 60 seconds and a sampling interval of 0.01 seconds for the nozzle action. The change condition is composed of 6001 values. Therefore, the change condition can express complex nozzle actions. In particular, the change condition can correctly express a nozzle action with a large number of speed change points for changing the moving speed of the nozzle. On the contrary, due to the large number of change conditions, when machine learning is performed on the time series data of the change condition, overfitting is sometimes generated.

[0114] The first conversion unit 263 of this embodiment converts the variation condition into compressed data. The compressed data is the residence time of the nozzle 311 in each of the plurality of movement intervals obtained by dividing the movement range of the nozzle 311 by the number of divisions 30. The inventors have found through experiments that even when the variation condition consisting of 6001 values ​​representing complex nozzle movements is converted into compressed data, the etching profile predicted by the predictor can obtain the desired result.

[0115] Therefore, the amount of data input to the predictor can be reduced, so the configuration of the predictor can be simplified, making it easier for the neural network to learn. Furthermore, the parameters of the neural network can be adjusted to appropriate values, which can improve the accuracy of the predictor.

[0116] Furthermore, since the variable conditions with a dimension of 6001 are converted into compressed data with a dimension of 30, there are sometimes multiple variable conditions whose compressed data become equal among the multiple variable conditions. In this case, the etching profiles predicted by the predictor for each of the multiple variable conditions with the same compressed data are the same. In this embodiment, when the processing condition determination unit 151 explores the processing conditions, since the corresponding processing conditions for different etching profiles are explored, the processing conditions corresponding to the multiple different etching profiles are selected. Therefore, the processing condition determination unit 151 can efficiently explore the processing conditions that predict the etching profile that becomes the target from the multiple processing conditions.

[0117] In addition, although the example of setting the sampling interval to 0.01 seconds is described, the sampling interval is not limited to this. The sampling interval may be set to be longer or shorter than this. For example, the sampling interval may be set to 0.1 seconds or 0.005 seconds.

[0118] 5. Other Implementation Methods

[0119] (1) In the above-mentioned embodiment, the learning device 200 generates a predictor based on learning data. The learning device 200 may perform additional learning on the predictor. After generating the predictor, the learning device 200 obtains the film thickness characteristics and processing conditions of each film before and after the substrate W processed by the substrate processing device 300. Then, the learning device 200 generates learning data based on the film thickness characteristics and processing conditions of each film before and after the processing, and performs additional learning on the predictor by causing the predictor to perform machine learning. Through additional learning, the structure of the neural network constituting the predictor is not changed, but the parameters are adjusted.

[0120] Since the predictor is subjected to machine learning using information obtained as a result of the substrate processing apparatus 300 actually processing the substrate W, the accuracy of the predictor can be improved. In addition, the amount of learning data used to generate the predictor can be minimized.

[0121] Fig.14 2 is a flowchart showing an example of the flow of the additional learning process. The additional learning process is a process executed by the CPU 201 when the CPU 201 of the learning device 200 executes an additional learning program stored in the RAM 202. The additional learning program is a part of the learning program.

[0122] Reference Fig.14, the CPU 201 of the learning device 200 obtains the production data (step S31), and the processing proceeds to step S32. The production data includes the processing conditions when the substrate processing device 300 processes the substrate W after the predictor is generated, and the film thickness characteristics of each film before and after the processing. The CPU 201 controls the input and output I / F 107 to obtain the production data from the substrate processing device 300. The production data can be obtained by reading the experimental data recorded in the storage medium such as CD-ROM 209 using the storage device 104.

[0123] In step S32, execute Fig.13 The compressed data generation process shown is performed, and the process proceeds to step S33. By executing the compressed data generation process, the variable conditions are converted into compressed data. In step S33, the compressed data, the fixed conditions contained in the processing conditions of the production data, and the etching profile are set as learning data. The etching profile is the difference between the film thickness characteristics of the film before processing contained in the production data and the film thickness characteristics of the film after processing contained in the production data. The compressed data generated by the first conversion unit 263 and the fixed conditions contained in the processing conditions are set as input data. The etching profile is set as the correct solution data.

[0124] In the subsequent step S34, the CPU 201 performs additional learning on the predictor and advances the process to step S35. The input data is input to the predictor, which is a neural network, and the parameters are determined so that the output of the predictor is equal to the correct answer data. Therefore, the parameters of the predictor are further adjusted.

[0125] In step S35, it is determined whether the adjustment is complete. The performance of the predictor is evaluated using the learning data for evaluation. When the evaluation result meets the preset additional learning evaluation benchmark, the adjustment is determined to be complete. The additional learning evaluation benchmark is a higher benchmark than the evaluation benchmark used when the predictor was generated. When the evaluation result does not meet the additional learning evaluation benchmark (No in step S35), the process returns to step S31. When the evaluation result meets the additional learning evaluation benchmark (Yes in step S35), the process ends.

[0126] (2) The learning device 200 can generate a distillation model for machine learning of a new learning model using the distillation data including the processing conditions determined by the information processing device 100 and the etching profile estimated by the predictor based on the processing conditions. Therefore, it is easy to prepare data for learning the new learning model.

[0127] (3) In the present embodiment, in the learning data used to generate the predictor, the input data includes compressed data after the variable conditions are converted and fixed conditions. The present invention is not limited to this. The input data may include only compressed data after the variable conditions are converted, and may not include fixed conditions.

[0128] (4) In the present embodiment, an example is described in which the first conversion unit 263 and the second conversion unit 157 convert the time series data related to the movement of the nozzle 311 into compressed data representing the residence time of the nozzle 311 in each of a plurality of movement intervals, but the present invention is not limited to this. For example, in the above-mentioned embodiment, an example is shown in which the supply amount of the processing liquid is constant, but the supply amount of the processing liquid may vary with the passage of time. In this case, the variation condition includes the supply amount of the processing liquid that varies with the passage of time. In this case, the compressed data includes the supply amount of the processing liquid in each of the plurality of movement intervals.

[0129] Fig.15 FIG. 1 is a diagram showing an example of the amount of treatment liquid supplied that changes over time. Fig.15 The upper side shows Figure 6 The same time series data of the nozzle action pattern is shown. Fig.15 An example of a change in the amount of the processing liquid ejected from the nozzle 311 (time history of the ejection flow rate of the processing liquid) is shown below. Fig.15 The horizontal axis of the plotted diagram at the bottom of represents time, and the vertical axis represents the flow rate of the processing liquid ejected from the nozzle 311. The supply amount of the processing liquid in each of the plurality of moving intervals d1 to d30 is calculated based on the relative position of the nozzle relative to the substrate that changes with the passage of time, and the flow rate of the processing liquid ejected from the nozzle 311 that changes with the passage of time. Specifically, the supply amount of the processing liquid in each of the plurality of moving intervals is calculated based on the residence time of the nozzle in the moving interval and the flow rate of the processing liquid supplied from the nozzle 311.

[0130] Fig.16 FIG. 1 is a diagram showing another example of compressed data. Fig.16 In FIG. 1 , the horizontal axis represents the position on the substrate W. The substrate center OP is represented by 0 mm, one radial substrate end is represented by -150 mm, and the other radial substrate end is represented by 150 mm. The movement intervals d1 to d30 are allocated between -150 mm and +150 mm on the horizontal axis.

[0131] The vertical axis represents the supply amount of the processing liquid in each of the moving sections d1 to d30. Figure 6 The supply amount of the processing liquid in each of the moving sections d1 to d30 in the case where the action pattern moves as shown. The supply amount is calculated as the accumulation of the processing liquid supplied from the nozzle 311 during the period when the nozzle 311 is located in each of the moving sections d1 to d30. In this case, since the time series data representing the action of the nozzle 311 is converted into compressed data that takes into account the supply amount of the processing liquid to the substrate W, a more accurate predictor can be generated.

[0132] (5) Although the information processing apparatus 100 and the learning apparatus 200 are provided separately from the substrate processing apparatus 300, the present invention is not limited thereto. The information processing apparatus 100 may be incorporated into the substrate processing apparatus 300. Furthermore, the information processing apparatus 100 and the learning apparatus 200 may be incorporated into the substrate processing apparatus 300. Furthermore, the information processing apparatus 100 and the learning apparatus 200 are provided separately, but they may be configured as an integrated apparatus.

[0133] (6) In the above embodiment, the plurality of moving intervals d1 to d30 are each set to have the same length in the radial direction of the substrate, but the plurality of moving intervals d1 to d30 can be set to different lengths. For example, the moving range can be divided in a manner that the trajectory of the movement of the nozzle 311 is evenly divided. For example, when the angle between the action end EP1 and the action end EP2 is 60 degrees with the second rotation axis AX2 as the center, the angle of each of the plurality of moving intervals d1 to d30 is 2 degrees with the second rotation axis AX2 as the center. In this way, the plurality of moving intervals d1 to d30 can be represented by the rotation angle of the nozzle motor 303.

[0134] 6. Effects of implementation methods

[0135] The learning device 200 of this embodiment drives the substrate processing device 300 with the processing conditions including the variable conditions to process the film formed on the substrate W, obtains the processing amount representing the difference in film thickness before and after the film is processed, and causes the neural network to perform machine learning on the learning data including the compressed data converted by the first conversion unit 263 from the variable conditions as input data and the etching profile corresponding to the processing conditions as correct answer data, thereby generating a learning model, i.e., a predictor, for estimating the etching profile. Since the learning data includes the compressed data converted in a manner that reduces the dimension of the variable conditions that change over time as input data, the dimension of the learning data can be reduced. Therefore, a learning device suitable for performing machine learning on the conditions that change over time to process the film formed on the substrate W can be generated.

[0136] Furthermore, the learning apparatus 200 compresses the time that the nozzle 311 stays in each of the plurality of movement sections obtained by dividing the movement range of the nozzle 311 of the substrate processing apparatus 300 , and thus can easily convert the variation conditions into compressed data.

[0137] Furthermore, the processing conditions include variable conditions and fixed conditions that do not change with the passage of time. Therefore, it is possible to cope with processing with different fixed conditions without generating a plurality of learning models with different fixed conditions.

[0138] Furthermore, after generating the predictor, the learning device 200 obtains a processing amount representing a difference in film thickness before and after the film is processed on the substrate W after the substrate processing device 300 processes the film according to the processing conditions, and uses additional learning data including a conversion result obtained by converting the variable conditions by the compressor and the obtained processing amount to learn the learning model. Therefore, the learning model is additionally learned, so that the performance of the learning model can be improved.

[0139] Furthermore, when the learning device 200 assigns the conversion result obtained by converting the temporary change condition by the compressor to the learning model and the processing amount estimated by the learning model satisfies the permissible condition, it generates a new learning model using the distillation data including the conversion result and the processing amount estimated by the learning model. Therefore, it is easy to prepare data for learning the new learning model.

[0140] Furthermore, the substrate processing apparatus 300 includes a nozzle 311 for supplying a processing liquid to the substrate W, and a nozzle moving mechanism 301 for changing the relative position of the nozzle and the substrate W, and the change condition is the relative position of the nozzle 311 and the substrate W changed by the nozzle moving mechanism 301. The relative position of the nozzle 311 and the substrate W is changed, and a learning model is generated for estimating the processing amount of the film processed by supplying the processing liquid to the substrate W from the nozzle 311. Therefore, a learning model for estimating the processing amount in the etching process can be generated.

[0141] Furthermore, when the information processing device 100 assigns compressed data converted from temporary change conditions by a compressor generated by the learning device 200 to a learning model generated by the learning device 200 and the etching profile inferred by the learning model satisfies the permissible condition, the processing condition including the temporary change condition is determined as the processing condition for driving the substrate processing device 300. Therefore, since the etching profile is inferred based on the processing conditions, it is not necessary to conduct experiments, etc. to obtain the influence of the action of the nozzle that moves in a complex manner on the processing result of the etching process. Furthermore, since a plurality of temporary change conditions are determined for the processing amount that satisfies the permissible condition, a plurality of change conditions corresponding to the plurality of etching profiles that satisfy the permissible condition can be determined. Therefore, a plurality of change conditions can be prompted for the processing result of the complex process of processing the substrate. If a processing condition that is easy to control the nozzle action is selected from the plurality of change conditions, the control of the substrate processing device 300 becomes easy.

[0142] Furthermore, since the information processing device 100 determines a plurality of processing conditions for the processing amount that satisfies the permissible condition, it is possible to determine a plurality of processing conditions corresponding to the plurality of etching profiles that satisfy the permissible condition. Furthermore, the fixed condition includes the temperature of the etching liquid. Therefore, it is possible to indicate a plurality of etching liquid temperatures for the processing result of the complex process of processing the substrate. Furthermore, it is possible to select the temperature of the etching liquid that is easily applicable to the etching process from the plurality of etching liquid temperatures. Furthermore, since it is possible to select the temperature of the etching liquid that is easily applicable, it is easy to control the temperature of the etching liquid used in the etching process.

[0143] 7. Correspondence between the components of the claims and the parts of the embodiments

[0144] The substrate W is an example of a substrate, the etching liquid is an example of a processing liquid, the substrate processing device 300 is an example of a substrate processing device, the experimental data acquisition unit 261 is an example of an experimental data acquisition unit, the first conversion unit 263 is an example of a conversion unit and a first conversion unit, the predictor is an example of a learning model, and the predictor generation unit 265 is an example of a model generation unit. In addition, the information processing device 100 is an example of an information processing device, the second conversion unit 157 is an example of a second conversion unit, the nozzle 311 is an example of a nozzle that supplies the processing liquid to the substrate, the nozzle moving mechanism 301 is an example of a moving unit, and the prediction unit 159, the evaluation unit 161 and the processing condition determination unit 151 are examples of a processing condition determination unit.

[0145] 8. Summary of implementation methods

[0146] (Item 1) A learning device comprising:

[0147] an experimental data acquisition unit that acquires a first processing amount indicating a difference in film thickness before and after the film is processed, after the film formed on the substrate is processed by driving the substrate processing device under processing conditions including a change condition indicating a relative position of the nozzle relative to the substrate that changes with time, wherein the substrate processing device moves the nozzle that supplies the processing liquid to the substrate on which the film is formed, and supplies the processing liquid to the substrate;

[0148] a conversion unit that converts the change condition into compressed data, the compressed data indicating the amount of movement associated with the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle movement in which the substrate processing apparatus moves the nozzle relative to the substrate into a smaller number of movement intervals than the number of data of the change condition; and

[0149] A model generating unit performs machine learning on learning data including the compressed data and the first processing amount corresponding to the processing conditions to generate a learning model, wherein the learning model estimates a second processing amount representing a difference in film thickness before and after the processing of the film, for the film formed on the substrate before the film is processed by the substrate processing device.

[0150] According to the learning device described in item 1, the learning data includes compressed data and processing volume converted in a manner that reduces the dimension of the change condition of the relative position of the nozzle relative to the substrate that changes over time. Therefore, the dimension of the learning data can be reduced. As a result, a learning device suitable for performing machine learning on the conditions that change over time to perform processing of a film formed on a substrate can be provided.

[0151] (Item 2) The learning device according to Item 1, wherein the movement amount is a residence time of the nozzle in each of the plurality of movement intervals.

[0152] According to the learning device described in item 2, since the operation amount is the residence time of the nozzle in each of the plurality of movement sections, conversion from the variable condition to the compressed data can be easily performed.

[0153] (Item 3) The learning device according to Item 1 or Item 2, wherein the variable condition further includes a flow rate of the processing liquid ejected from the substrate processing device to the substrate over time;

[0154] The operation amount is a supply amount of the processing liquid calculated based on a residence time of the nozzle in the movement section and a flow rate of the processing liquid supplied from the nozzle in each of the plurality of movement sections.

[0155] According to the learning device described in item 3, since the supply amount of the processing liquid for each of the plurality of movement intervals is considered as the operation amount, it can be converted into compressed data with a dimension smaller than the dimension of the plurality of types of change conditions.

[0156] (Item 4) A learning device as described in any one of Items 1 to 3, wherein the plurality of movement intervals are of the same length.

[0157] According to the learning device described in item 4, since the plurality of movement intervals are each set to the same length, the compressed data represents the movement of the nozzle within the plurality of movement intervals set to the same length. Therefore, it is possible to convert the entire substrate into compressed data in which the deviation of the movement amount of the nozzle relative to the substrate between different positions on the substrate is reduced.

[0158] (Item 5) A learning device as described in Item 4, wherein the length of each of the plurality of movement intervals is the length in the radial direction of an area of ​​the upper surface of the substrate that the nozzle crosses during movement in the movement interval.

[0159] According to the learning device described in Item 5, the length of each of the multiple movement intervals is the radial length of the area where the nozzle crosses the substrate during the movement of the movement interval. Therefore, the movement amount of the nozzle relative to the substrate can be converted into compressed data with reduced deviation between different radial positions of the substrate.

[0160] (Item 6) An information processing device for managing a substrate processing device, wherein:

[0161] The substrate processing apparatus processes the film formed on the substrate by supplying a processing liquid to the substrate on which the film is formed under processing conditions including a change condition indicating a relative position of a nozzle relative to the substrate that changes with time; and the information processing apparatus comprises:

[0162] a conversion unit that converts the aforementioned change condition into compressed data, the compressed data indicating the amount of movement associated with the aforementioned nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the aforementioned nozzle during a scanning period from the start to the end of a nozzle movement in which the aforementioned substrate processing apparatus moves the aforementioned nozzle relative to the aforementioned substrate into a smaller number of movement intervals than the number of data of the aforementioned change condition, and

[0163] a processing condition determination unit that determines a processing condition for driving the substrate processing apparatus using a learning model that estimates a second processing amount indicating a difference in film thickness before and after the film is processed, for the film formed on the substrate before the film is processed by the substrate processing apparatus;

[0164] The learning model is an inference model that performs machine learning on learning data, and the learning data includes: compressed data obtained by performing the same conversion as the conversion unit on the variable conditions included in the processing conditions for the substrate processing apparatus to process the film formed on the substrate, and a first processing amount representing a difference in film thickness before and after the film is formed on the substrate after the substrate processing apparatus processes the film;

[0165] When the compressed data obtained by converting the temporary change condition by the conversion unit is assigned to the learning model and the second processing amount inferred by the learning model satisfies the permissible condition, the processing condition determination unit determines the processing condition including the temporary change condition as the processing condition for driving the substrate processing device.

[0166] According to the information processing device of item 6, when the compressed data converted from the temporary change condition that changes with the passage of time is given to the learning model and the processing amount estimated by the learning model satisfies the allowable condition, the processing condition including the temporary change condition is determined as the processing condition for driving the substrate processing device. Therefore, a plurality of temporary change conditions can be determined for the processing amount that satisfies the allowable condition. As a result, a plurality of processing conditions can be presented for the processing result of a complex process of processing a film formed on a substrate.

[0167] (Item 7) A substrate processing device comprising the information processing device as described in Item 6.

[0168] According to the substrate processing apparatus described in item 7, it is possible to present a plurality of processing conditions with respect to the processing results of a complex process of processing a substrate.

[0169] (Item 8) A substrate processing system for managing a substrate processing device for processing a substrate, wherein:

[0170] It includes a learning device and an information processing device;

[0171] The substrate processing apparatus processes the film formed on the substrate by supplying a processing liquid to the substrate on which the film is formed under processing conditions including a change condition indicating that a relative position of a nozzle relative to the substrate changes with time.

[0172] The aforementioned learning device comprises:

[0173] an experimental data acquisition unit that acquires a first processing amount indicating a difference in film thickness before and after the film is processed after the substrate processing apparatus is driven under the processing conditions to process the film formed on the substrate;

[0174] a first conversion unit that converts the aforementioned change condition into compressed data, the compressed data indicating the amount of movement related to the aforementioned nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the aforementioned nozzle during a scanning period from the start to the end of a nozzle movement in which the aforementioned substrate processing apparatus moves the aforementioned nozzle relative to the aforementioned substrate into a number of movement intervals that is smaller than the number of data of the aforementioned change condition, and

[0175] a model generating unit that performs machine learning on learning data including the compressed data converted by the first converting unit from the variable condition and the first processing amount corresponding to the processing condition, and generates a learning model that estimates a second processing amount indicating a difference in film thickness before and after the processing of the film, for the film formed on the substrate before the film is processed by the substrate processing apparatus;

[0176] The aforementioned information processing device comprises:

[0177] a second conversion unit, which is the same as the first conversion unit, and

[0178] a processing condition determination unit that determines a processing condition for driving the substrate processing device using the learning model generated by the learning device;

[0179] When the processing condition determination unit assigns the conversion result of the temporary change condition by the second conversion unit to the learning model and the second processing amount inferred by the learning model satisfies the permissible condition, the processing condition including the temporary change condition is determined as the processing condition for driving the substrate processing device.

[0180] The substrate processing system described in Item 8 is suitable for performing machine learning on conditions that change over time to process a film formed on a substrate, and can prompt multiple processing conditions for the processing results of a complex process of processing a film formed on a substrate.

[0181] (Item 9) A learning method, wherein a computer is caused to execute the following processing:

[0182] After a film formed on the substrate is processed by driving a substrate processing apparatus under processing conditions including a change condition indicating a relative position of a nozzle that changes with time relative to a substrate, a first processing amount indicating a difference in film thickness before and after the processing of the film is obtained, wherein the substrate processing apparatus is configured to move the nozzle that supplies a processing liquid to the substrate on which the film is formed, so as to supply the processing liquid to the substrate;

[0183] converting the aforementioned change condition into compressed data, the compressed data representing the movement amount related to the aforementioned nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the aforementioned nozzle during a scanning period from the start to the end of a nozzle movement in which the aforementioned substrate processing apparatus moves the aforementioned nozzle relative to the aforementioned substrate into a smaller number of movement intervals than the number of data of the aforementioned change condition; and

[0184] Machine learning is performed on learning data including the compressed data and the first processing amount corresponding to the processing conditions to generate a learning model, which infers a second processing amount representing the difference in film thickness before and after the processing of the film, for the film formed on the substrate before the film is processed by the substrate processing device.

[0185] According to the learning method described in item 9, the learning data includes compressed data and processing volume converted in a manner that reduces the dimension of the changing conditions that change over time. Therefore, the dimension of the learning data can be reduced. As a result, a learning method suitable for performing machine learning on conditions that change over time to perform processing of a film formed on a substrate can be provided.

[0186] (Item 10) A method for determining processing conditions, executed by a computer managing a substrate processing apparatus, wherein:

[0187] The substrate processing apparatus processes the film formed on the substrate by supplying a processing liquid to the substrate on which the film is formed under processing conditions including a change condition indicating that a relative position of a nozzle relative to the substrate changes with time.

[0188] The processing condition determination method includes the following processing:

[0189] converting the aforementioned change condition into compressed data, the compressed data representing the movement amount related to the aforementioned nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the aforementioned nozzle during a scanning period from the start to the end of a nozzle movement in which the aforementioned substrate processing apparatus moves the aforementioned nozzle relative to the aforementioned substrate into a smaller number of movement intervals than the number of data of the aforementioned change condition; and

[0190] determining a processing condition for driving the substrate processing apparatus using a learning model for estimating a second processing amount indicating a difference in film thickness before and after the processing of the film, for the film formed on the substrate before the film is processed by the substrate processing apparatus;

[0191] The learning model is an inference model that performs machine learning on learning data, and the learning data includes: compressed data obtained by performing the same conversion process as the conversion process on the variable conditions included in the processing conditions for the processing of the film formed on the substrate by the substrate processing device, and a first processing amount representing a difference in film thickness before and after the processing of the film formed on the substrate after the processing of the film by the substrate processing device;

[0192] The processing condition for determining the above-mentioned processing condition is determined as the processing condition for driving the above-mentioned substrate processing device when the compressed data obtained by converting the temporary change condition by performing the above-mentioned conversion process is assigned to the above-mentioned learning model and the above-mentioned second processing amount inferred by the above-mentioned learning model satisfies the allowable condition.

[0193] According to the processing condition determination method described in Item 10, it is possible to provide a processing condition determination method capable of presenting a plurality of processing conditions for a processing result of a complex process of processing a film formed on a substrate.

Claims

1. A learning device, wherein: Include: an experimental data acquisition unit that acquires a first processing amount indicating a difference in film thickness before and after the film is processed after driving a substrate processing device under processing conditions including a change condition indicating a relative position of a nozzle with respect to a substrate that changes with time to process a film formed on the substrate, the substrate processing device being configured to move the nozzle that supplies a processing liquid to the substrate on which the film is formed to supply the processing liquid to the substrate; a conversion unit that converts the change condition into compressed data, the compressed data indicating an amount of movement related to the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle movement in which the substrate processing apparatus moves the nozzle relative to the substrate into a smaller number of movement intervals than the number of data of the change condition; and A model generating unit performs machine learning on learning data including the compressed data and the first processing amount corresponding to the processing conditions to generate a learning model, wherein the learning model estimates a second processing amount representing a difference in film thickness before and after the film is processed for the film formed on the substrate before the film is processed by the substrate processing device.

2. The learning device according to claim 1, wherein: The movement amount is the residence time of the nozzle in each of the plurality of movement sections.

3. The learning device according to claim 1 or 2, wherein: The variable condition further includes a flow rate of the processing liquid sprayed onto the substrate by the substrate processing device over time; The operation amount is a supply amount of the processing liquid calculated based on a residence time of the nozzle in the movement section and a flow rate of the processing liquid supplied from the nozzle in each of the plurality of movement sections.

4. The learning device according to any one of claims 1 to 3, wherein: The plurality of movement intervals have the same length.

5. The learning device according to claim 4, wherein: The length of each of the plurality of movement sections is the length in the radial direction of a region of the upper surface of the substrate that the nozzle crosses while moving in the movement section.

6. An information processing device for managing a substrate processing device, wherein: The substrate processing apparatus processes the film formed on the substrate by supplying a processing liquid to the substrate on which the film is formed under processing conditions including a change condition indicating that a relative position of a nozzle with respect to the substrate changes with time. The information processing device comprises: a conversion unit that converts the change condition into compressed data, the compressed data indicating the amount of movement related to the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle movement in which the substrate processing device moves the nozzle relative to the substrate into a smaller number of movement intervals than the number of data of the change condition, and a processing condition determination unit that determines a processing condition for driving the substrate processing apparatus using a learning model that estimates a second processing amount representing a difference in film thickness before and after the film is processed, for the film formed on the substrate before the film is processed by the substrate processing apparatus; The learning model is an inference model for performing machine learning on learning data, the learning data including: compressed data obtained by performing the same conversion as the conversion unit on the variable condition included in the processing condition for the substrate processing apparatus to process the film formed on the substrate, and a first processing amount representing a difference in film thickness before and after the film formed on the substrate after the substrate processing apparatus processes the film; The processing condition determination unit determines the processing condition including the temporary change condition as the processing condition for driving the substrate processing device when the compressed data obtained by converting the temporary change condition by the conversion unit is assigned to the learning model and the second processing amount estimated by the learning model satisfies the permissible condition.

7. A substrate processing device, wherein: An information processing device comprising the information processing device according to claim 6.

8. A substrate processing system for managing a substrate processing device for processing a substrate, wherein: It includes a learning device and an information processing device; The substrate processing apparatus processes the film formed on the substrate by supplying a processing liquid to the substrate on which the film is formed under processing conditions including a change condition indicating that the relative position of the nozzle relative to the substrate changes with time. The learning device comprises: an experimental data acquisition unit that acquires a first processing amount indicating a difference in film thickness before and after the film is processed after the substrate processing apparatus is driven under the processing conditions to process the film formed on the substrate, a first conversion unit that converts the change condition into compressed data, the compressed data indicating an amount of movement related to the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle movement in which the substrate processing apparatus moves the nozzle relative to a substrate into a smaller number of movement intervals than the number of data of the change condition, and a model generating unit that performs machine learning on learning data including the compressed data converted by the first converting unit from the variable condition and the first processing amount corresponding to the processing condition, and generates a learning model that estimates a second processing amount indicating a difference in film thickness before and after the processing of the film, for the film formed on the substrate before the film is processed by the substrate processing apparatus; The information processing device comprises: a second conversion section, which is the same as the first conversion section, and a processing condition determination unit that determines a processing condition for driving the substrate processing apparatus using the learning model generated by the learning apparatus; The processing condition determination unit determines the processing condition including the temporary change condition as the processing condition for driving the substrate processing device when the conversion result of the temporary change condition by the second conversion unit is assigned to the learning model and the second processing amount estimated by the learning model satisfies the permissible condition.

9. A learning method, wherein: Causes the computer to perform the following processing: After driving a substrate processing device under processing conditions including a change condition indicating a relative position of a nozzle that changes with time relative to a substrate to process a film formed on the substrate, a first processing amount indicating a difference in film thickness before and after the film is processed is obtained, wherein the substrate processing device moves the nozzle that supplies a processing liquid to the substrate on which the film is formed to supply the processing liquid to the substrate; converting the change condition into compressed data, the compressed data representing the movement amount related to the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle movement in which the substrate processing device moves the nozzle relative to the substrate into a smaller number of movement intervals than the number of data of the change condition; and Machine learning is performed on learning data including the compressed data and the first processing amount corresponding to the processing conditions to generate a learning model, wherein the learning model infers a second processing amount representing the difference in film thickness before and after the film is processed, for the film formed on the substrate before the film is processed by the substrate processing device.

10. A method for determining a processing condition, wherein: The method is executed by a computer managing a substrate processing device, wherein: The substrate processing apparatus processes the film formed on the substrate by supplying a processing liquid to the substrate on which the film is formed under processing conditions including a change condition indicating that the relative position of the nozzle relative to the substrate changes with time. The processing condition determination method includes the following processing: converting the change condition into compressed data, the compressed data representing the amount of movement related to the nozzle for each of a plurality of movement intervals, the plurality of movement intervals being movement intervals obtained by dividing a movement range of the nozzle during a scanning period from the start to the end of a nozzle movement in which the substrate processing device moves the nozzle relative to the substrate into a smaller number of movement intervals than the number of data of the change condition, and determining a processing condition for driving the substrate processing apparatus using a learning model for estimating a second processing amount indicating a difference in film thickness before and after the film is processed for the film formed on the substrate before the film is processed by the substrate processing apparatus; The learning model is an inference model for performing machine learning on learning data, the learning data including: compressed data obtained by performing the same conversion as the conversion process on the variable condition included in the processing condition of the substrate processing apparatus performing the processing of the film formed on the substrate, and a first processing amount representing a difference in film thickness before and after the processing of the film formed on the substrate after the substrate processing apparatus performs the processing of the film; In the process of determining the processing conditions, when the compressed data obtained by converting the temporary change conditions by performing the conversion process is assigned to the learning model and the second processing amount inferred by the learning model satisfies the allowable conditions, the processing conditions including the temporary change conditions are determined as the processing conditions for driving the substrate processing device.

Images