Computer program, information processing method and information processing device

Abstract

To provide a computer program for predicting a temperature of a substrate fixed to an electrostatic chuck, an information processing method and an information processing device.SOLUTION: A feature amount of a surface shape of an electrostatic chuck for fixing a substrate to be processed is acquired, and recipe information representing a content of a recipe specifying a processing condition of the substrate is acquired. The acquired feature amount of the surface shape of the electrostatic chuck and the acquired recipe information are inputted to a learned model which outputs a prediction temperature predicting a temperature of the substrate fixed to the electrostatic chuck in a case where the feature amount of the surface shape of the electrostatic chuck and the recipe information are inputted, and the prediction temperature outputted by the learned model is acquired, thereby predicting the temperature of the substrate fixed to the electrostatic chuck.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present disclosure relates to a computer program, an information processing method, and an information processing apparatus.

Background Art

[0002] In substrate processing such as etching or film formation on a substrate such as a semiconductor wafer or a glass substrate, an electrostatic chuck may be used to fix the substrate. An electrostatic chuck is a device that fixes a substrate by adsorbing the substrate with an electric force. Etching, film formation, or other processes are performed on the substrate fixed to the electrostatic chuck. Patent Document 1 discloses an example of an electrostatic chuck.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] As the electrostatic chuck is used for a long period of time, the heat conduction between the electrostatic chuck and the fixed substrate changes. Therefore, when using an electrostatic chuck with a long usage period, the temperature distribution of the fixed substrate is different from that when using a new electrostatic chuck. When performing substrate processing, it is necessary to use an electrostatic chuck such that the temperature of the fixed substrate is within a reasonable temperature distribution range. Conventionally, it has been difficult to predict the temperature of the substrate fixed to the electrostatic chuck.

[0005] The present disclosure provides a computer program, an information processing method, and an information processing apparatus for predicting the temperature of a substrate fixed to an electrostatic chuck.

Means for Solving the Problems

[0006] A computer program according to an aspect of the present disclosure acquires a feature amount of the surface shape of an electrostatic chuck for fixing a substrate to be processed, acquires recipe information representing the content of a recipe that defines the processing conditions of the substrate, and inputs the feature amount of the surface shape of the electrostatic chuck and the recipe information to a learned model that outputs a predicted temperature of the substrate fixed to the electrostatic chuck when the feature amount and the recipe information are input, and inputs the acquired feature amount of the surface shape of the electrostatic chuck and the recipe information, and causes a computer to execute a process of predicting the temperature of the substrate fixed to the electrostatic chuck by acquiring the predicted temperature output by the learned model.

Advantages of the Invention

[0007] According to the present disclosure, it is possible to provide a computer program, an information processing method, and an information processing apparatus for predicting the temperature of a substrate fixed to an electrostatic chuck.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Embodiments for Carrying Out the Invention

[0009] Hereinafter, the present disclosure will be specifically described based on the drawings showing its embodiments. The process of manufacturing a substrate such as a semiconductor wafer, a glass substrate, or a flat panel substrate includes a process of performing a process such as etching or film formation on the substrate. Hereinafter, the process performed on the substrate is referred to as substrate processing, and the apparatus that executes the substrate processing is referred to as a processing apparatus. For example, the processing apparatus includes a process chamber, and substrate processing such as plasma etching or CVD (chemical vapor deposition) is performed on the substrate disposed in the process chamber. The processing apparatus processes the substrate according to a predetermined recipe that defines the processing conditions of the substrate. The processing conditions defined in the recipe include the flow rate of the gas supplied to the process chamber for performing the substrate processing, the supplied power, pressure, temperature, and the like.

[0010] The processing apparatus is provided with an electrostatic chuck for fixing the substrate to be processed. The electrostatic chuck fixes the substrate by adsorbing the substrate with an electric force. In the processing apparatus, processing is performed on the substrate fixed to the electrostatic chuck. FIG. 1 is a schematic cross-sectional view showing an example of the electrostatic chuck and the substrate. A plurality of protrusions 11 are formed on the surface of the electrostatic chuck 1. When the substrate 2 is adsorbed to the electrostatic chuck 1, the substrate 2 contacts the top surfaces of the plurality of protrusions 11, and the plurality of protrusions 11 support the substrate 2. In addition, an annular seal band 12 is provided at the edge of the surface of the electrostatic chuck 1. When the substrate 2 is adsorbed to the electrostatic chuck 1, the substrate 2 also contacts the seal band 12.

[0011] When the electrostatic chuck 1 continues to be used in substrate processing, the surface shape of the electrostatic chuck 1 changes over time. For example, due to the expansion and contraction of the substrate 2 caused by heat, the substrate 2 slides with respect to the surface of the electrostatic chuck 1, and the surface of the electrostatic chuck 1 is worn. In particular, the top surface of the protrusion 11 is worn. For example, solids generated from the gas supplied into the process chamber are deposited on the surface of the electrostatic chuck 1. In particular, granular solids adhere to the top surface of the protrusion 11. Or, the electrostatic chuck 1 warps due to changes over time. When the surface shape of the electrostatic chuck 1 changes, the contact state between the electrostatic chuck 1 and the substrate 2 changes, and the heat conduction between the electrostatic chuck 1 and the substrate 2 changes. For this reason, the temperature of the substrate 2 fixed to the electrostatic chuck 1 with a long usage time is different from the temperature of the substrate 2 fixed to the electrostatic chuck 1 with a short usage time. More specifically, as the usage time of the electrostatic chuck 1 becomes longer, the contact area between the electrostatic chuck 1 and the substrate 2 becomes smaller, the heat of the substrate 2 is less likely to be transferred to the electrostatic chuck 1, and the temperature of the substrate 2 rises.

[0012] Since the temperature of the substrate 2 affects substrate processing, in order to perform substrate processing with high accuracy, it is necessary to appropriately adjust the temperature of the substrate 2. When the temperature of the substrate 2 is significantly different compared to the case where the electrostatic chuck 1 with a short usage time is used, it is difficult to appropriately adjust the temperature of the substrate 2. An electrostatic chuck 1 with a usage time of zero or an electrostatic chuck 1 with a usage time of a predetermined time or less is used as the electrostatic chuck 1 in the initial state. For example, when the magnitude of the difference between the temperature of the substrate 2 fixed to the electrostatic chuck 1 and the temperature of the substrate 2 fixed to the electrostatic chuck 1 in the initial state reaches a predetermined magnitude, it is assumed that the electrostatic chuck 1 has reached the end of its life. The electrostatic chuck 1 that has reached the end of its life is replaced.

[0013] If the temperature of the substrate 2 fixed to the electrostatic chuck 1 can be predicted in a state where no substrate processing is being performed, it is possible to determine whether the electrostatic chuck 1 has reached the end of its life. Also, it is possible to estimate how much longer the electrostatic chuck 1 can be used until it reaches the end of its life. In the present embodiment, a process of predicting the temperature of the substrate 2 fixed to the electrostatic chuck 1 is performed.

[0014] FIG. 2 is a conceptual diagram showing a configuration example of the information processing system according to the present embodiment. The information processing system includes a processing apparatus 31 that executes substrate processing, a control apparatus 32 that controls the processing apparatus 31, a surface shape measuring apparatus 33, and an information processing apparatus 4. The processing apparatus 31 performs substrate processing on a substrate. For example, the processing apparatus 31 includes a process chamber and performs plasma etching as the substrate processing. The process chamber includes an electrostatic chuck 1 (not shown in FIG. 2). That is, the processing apparatus 31 includes the electrostatic chuck 1. The control apparatus 32 adjusts the processing conditions of the substrate processing performed in the processing apparatus 31 according to a predetermined recipe.

[0015] The surface shape measuring apparatus 33 measures the shape of the surface of the electrostatic chuck 1 provided in the processing apparatus 31 by 3D scanning the surface of the electrostatic chuck 1 using a laser. The surface shape measuring apparatus 33 measures the positions of a plurality of portions on the surface of the electrostatic chuck 1 by 3D scanning and generates surface shape information including the positions of the plurality of portions on the surface of the electrostatic chuck 1. The surface shape information represents the result of 3D scanning the surface of the electrostatic chuck 1. For example, the surface shape information includes xyz coordinate values representing the positions of respective portions on the surface of the electrostatic chuck 1. The surface of the electrostatic chuck 1 is represented by the positions of the plurality of portions on the surface of the electrostatic chuck 1.

[0016] The information processing apparatus 4 executes an information processing method. The information processing apparatus 4 acquires recipe information representing the content of a recipe from the control apparatus 32 and acquires surface shape information from the surface shape measuring apparatus 33. The information processing apparatus 4 performs a process of predicting the temperature of a substrate fixed to the electrostatic chuck 1 provided in the processing apparatus 31 based on the recipe information and the surface shape information.

[0017] FIG. 3 is a block diagram showing a configuration example inside the information processing apparatus 4. The information processing apparatus 4 executes an information processing method. The information processing apparatus 4 is configured using a computer such as a personal computer or a server apparatus. The information processing apparatus 4 includes an arithmetic unit 41, a memory 42, a storage unit 43, a reading unit 44, an operation unit 45, a display unit 46, and an interface unit 47. The arithmetic unit 41 is configured using, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a multi-core CPU. The arithmetic unit 41 may be configured using a quantum computer. The memory 42 stores temporary data generated during the operation. The memory 42 is, for example, a RAM (Random Access Memory). The storage unit 43 is non-volatile and is, for example, a hard disk or a non-volatile semiconductor memory. The reading unit 44 reads information from a recording medium 40 such as an optical disk or a portable memory.

[0018] The operation unit 45 receives an input of information such as text by receiving an operation from the user. The operation unit 45 is, for example, a keyboard, a pointing device, or a touch panel. The display unit 46 displays an image. The display unit 46 is, for example, a liquid crystal display or an EL display (Electroluminescent Display). The operation unit 45 and the display unit 46 may be integrated. The interface unit 47 receives data input from the outside of the information processing apparatus 4. For example, the interface unit 47 is a communication unit that performs wired or wireless communication. For example, the interface unit 47 is a general-purpose interface such as a USB (Universal Serial Bus).

[0019] The arithmetic unit 41 causes the reading unit 44 to read the computer program 431 recorded on the recording medium 40, and stores the read computer program 431 in the storage unit 43. The arithmetic unit 41 executes information processing for realizing the functions of the information processing apparatus 4 according to the computer program 431. The computer program 431 may be a program product. The computer program 431 may be stored in the storage unit 43 in advance, or may be downloaded from outside the information processing apparatus 4. In this case, the information processing apparatus 4 may not include the reading unit 44.

[0020] The computer program 431 can be deployed to be executed on a single computer, or arranged at one site, or distributed over a plurality of sites and executed on a plurality of computers interconnected by a communication network. That is, the information processing apparatus 4 may be composed of a plurality of computers, and the computer program 431 may be executed on a plurality of computers connected via a communication network. The information processing apparatus 4 may be configured using a cloud server.

[0021] The information processing apparatus 4 includes a learned model 432 used for predicting the temperature of a substrate fixed to the electrostatic chuck 1 provided in the processing apparatus 31. The learned model 432 is realized by the arithmetic unit 41 executing processing according to the computer program 431. The storage unit 43 stores data necessary for realizing the learned model 432. For example, the learned model 432 is realized using a neural network. The learned model 432 may be realized using a model other than a neural network, such as a support vector machine.

[0022] The learned model 432 may be configured using hardware. For example, the learned model 432 may be configured by hardware including a processor and a memory that stores necessary programs and data. Or, the learned model 432 may be realized using a quantum computer. Alternatively, the learned model 432 may be provided outside the information processing apparatus 4, and the information processing apparatus 4 may execute processing using the external learned model 432. For example, the learned model 432 may be realized using the cloud.

[0023] FIG. 4 is a conceptual diagram showing a functional example of the learned model 432. The learned model 432 receives recipe information representing the content of a recipe that defines the processing conditions of the substrate in the processing apparatus 31, and a feature amount of the surface shape of the electrostatic chuck 1 used for substrate processing. The learned model 432 has been pre-learned to output the predicted temperature of the substrate fixed to the electrostatic chuck 1 when the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information are input. The predicted temperature is a value obtained by predicting the temperature of the surface of the substrate fixed to the electrostatic chuck 1 when performing substrate processing according to the recipe specified by the recipe information. The learned model 432 outputs a plurality of predicted temperatures obtained by predicting the temperatures at a plurality of portions of the substrate.

[0024] The feature amount of the surface shape of the electrostatic chuck 1 is various numerical values representing the features of the surface shape of the electrostatic chuck 1. The feature amount of the surface shape of the electrostatic chuck 1 is calculated from the surface shape information. Usually, a plurality of types of feature amounts are obtained from the surface shape information. For example, the plurality of types of feature amounts of the surface shape of the electrostatic chuck 1 include the warp of the electrostatic chuck 1, the height of the protrusion 11, the contact area of the protrusion 11, and the undulation amount of the seal band 12. The warp of the electrostatic chuck 1 is the warp of the surface of the electrostatic chuck 1. For example, the warp of the electrostatic chuck 1 is the difference between the distance from the center of the surface of the electrostatic chuck 1 to a predetermined horizontal plane and the distance from the edge of the surface of the electrostatic chuck 1 to the predetermined horizontal plane. The contact state between the electrostatic chuck 1 and the substrate changes depending on the state of the warp of the electrostatic chuck 1.

[0025] The height of the protrusions 11 is the distance from the base to the tip of each protrusion 11. If the heights of the respective protrusions 11 are different, the contact states between the respective protrusions 11 and the substrate are different. The contact area is the area where each protrusion 11 contacts the substrate. The contact area between each protrusion 11 and the substrate affects the contact area between the entire electrostatic chuck 1 and the substrate. The undulation amount of the seal band 12 is the undulation amount on the surface of the seal band 12. For example, it is the variation amount of the distance from each part of the surface of the seal band 12 to a predetermined horizontal plane. The undulation amount of the seal band 12 affects the contact state between the electrostatic chuck 1 and the substrate. Specifically, the larger the undulation amount, the smaller the contact area between the electrostatic chuck 1 and the substrate. The plurality of types of feature amounts may not include any of the warp of the electrostatic chuck 1, the height of the protrusions 11, the contact area, and the undulation amount of the seal band 12, or may include other amounts.

[0026] FIG. 5 is a conceptual diagram showing an example of the content of recipe information. It represents the content of a recipe that defines the processing conditions of the substrate in the processing apparatus 31. The recipe is defined to sequentially execute a plurality of processing steps. In the example shown in FIG. 5, the processing step numbers are shown in the order in which the processing steps are executed. For each processing step, the processing conditions of the substrate processing executed in the processing apparatus 31 are defined. For example, the substrate processing is etching. In the example shown in FIG. 5, as the processing conditions for each processing step, the processing time which is the length of time for which the processing step is executed, the pressure in the process chamber, the power supplied to the process chamber during etching, the flow rates of a plurality of types of gases supplied to the process chamber, and the temperature in the process chamber are defined. The processing conditions are different for each processing step. The recipe information may not include any of the processing time, pressure, power, gas flow rate, and temperature, or may include other processing conditions.

[0027] The contact state between the electrostatic chuck 1 and the substrate is affected by recipe information. For example, due to differences in recipe information, the sliding of the substrate with respect to the electrostatic chuck 1 may increase, or the deposition of solids generated from the gas may increase. When the sliding of the substrate is large, the electrostatic chuck 1 and the substrate are likely to come into contact, the heat of the substrate is likely to be transferred to the electrostatic chuck 1, and the rise in the temperature of the substrate according to the length of use of the electrostatic chuck 1 can be suppressed. When there is a large amount of solid deposition, the electrostatic chuck 1 and the substrate are more difficult to contact, the heat of the substrate is less likely to be transferred to the electrostatic chuck 1, and the rise in the temperature of the substrate according to the length of use of the electrostatic chuck 1 becomes large.

[0028] The learned model 432 is generated by learning by the learning device 5. FIG. 6 is a block diagram showing an example of the internal functional configuration of the learning device 5. The learning device 5 is a computer such as a server device or a personal computer. The learning device 5 includes an arithmetic unit 51, a memory 52, a storage unit 53, a reading unit 54, an operation unit 55, and a display unit 56. The arithmetic unit 51 is configured using, for example, a CPU, a GPU, or a multi-core CPU. The arithmetic unit 51 may be configured using a quantum computer. The memory 52 stores temporary data generated during the calculation. The memory 52 is, for example, a RAM. The reading unit 54 reads information from a recording medium 50 such as an optical disk or a portable memory. The storage unit 53 is non-volatile and is, for example, a hard disk or a non-volatile semiconductor memory.

[0029] The operation unit 55 receives input of information by receiving an operation from the user. The operation unit 55 is, for example, a keyboard, a pointing device, or a touch panel. The display unit 56 displays an image. The display unit 56 is, for example, a liquid crystal display or an EL display.

[0030] The arithmetic unit 51 causes the reading unit 54 to read the computer program 531 recorded on the recording medium 50, and stores the read computer program 531 in the storage unit 53. The arithmetic unit 51 executes processing for realizing the functions of the learning device 5 according to the computer program 531. The computer program 531 may be a program product. The computer program 531 may be stored in the storage unit 53 in advance, or may be downloaded from outside the learning device 5. In this case, the learning device 5 may not include the reading unit 54.

[0031] The computer program 531 can be deployed to be executed on a single computer, or placed at one site, or distributed over a plurality of sites and executed on a plurality of computers interconnected by a communication network. That is, the learning device 5 may be composed of a plurality of computers, and the computer program 531 may be executed on a plurality of computers connected via a communication network. The learning device 5 may be configured using a cloud server.

[0032] The learning device 5 generates a learned model 432 by performing machine learning. Hereinafter, the process by which the learning device 5 generates the learned model 432 will be described. The learning device 5 includes a learning model 532 that is the basis of the learned model 432. The learning device 5 learns the learning model 532 to generate the learned model 432. The learning model 532 is realized by the arithmetic unit 51 executing processing according to the computer program 531. The learning model 532 is realized using a neural network. The learning model 532 may be configured using hardware.

[0033] The storage unit 53 stores learning data 533 for training the learning model 532 to generate the trained model 432. The learning data 533 records a data set associating recipe information, the feature amount of the surface shape of the electrostatic chuck 1, and the temperature of the substrate fixed to the electrostatic chuck 1. The feature amount of the surface shape of the electrostatic chuck 1 is a value calculated from the surface shape information generated by the surface shape measuring device 33. The data set includes a plurality of types of feature amounts. In the processing device 31, substrate processing according to the recipe specified by the recipe information is performed on the substrate fixed to the electrostatic chuck 1, and the temperature of each part of the surface of the substrate is measured. The plurality of measured temperatures are the temperatures of the substrate included in the learning data 533.

[0034] A plurality of data sets are recorded in the learning data 533. The plurality of data sets include a data set in which recipe information representing the content of the actual recipe is recorded. The plurality of data sets also include a data set in which recipe information with increased substrate sliding is recorded and a data set in which recipe information with increased deposition of solids generated from the gas is recorded. Compared with the recipe information representing the content of the actual recipe, the magnitude of substrate sliding and the amount of solid deposition are represented by relative numerical values. For example, taking the magnitude of substrate sliding by the recipe information representing the content of the actual recipe as 50%, the recipe information with increased substrate sliding stipulates that the magnitude of substrate sliding is 100%. For example, taking the amount of solid deposition by the recipe information representing the content of the actual recipe as 50%, the amount of solid deposition by the recipe information with increased solid deposition stipulates that the amount of solid deposition is 100%. Using the learning data 533 including such a plurality of data sets, the trained model 432 is generated.

[0035] FIG. 7 is a flowchart showing an example of the procedure of the process executed by the learning device 5. Hereinafter, steps are abbreviated as S. By the arithmetic unit 51 executing information processing according to the computer program 531, the learning device 5 executes the following processes. By the arithmetic unit 51 reading out the learning data 533 stored in the storage unit 53, the learning device 5 acquires the learning data 533 (S11). In S11, the learning device 5 may acquire the learning data 533 by the learning data 533 being input from the outside.

[0036] Next, the learning device 5 generates the learned model 432 by causing the learning model 532 to be learned using the learning data 533 (S12). In S12, the arithmetic unit 51 inputs the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information recorded in the learning data 533 into the learning model 532 that is the basis of the learned model 432. At this time, the arithmetic unit 51 inputs a plurality of types of feature amounts into the learning model 532. The learning model 532 performs calculations in response to the input of the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information, and outputs a predicted temperature.

[0037] The arithmetic unit 51 adjusts the calculation parameters of the learning model 532 so that the error between the predicted temperature output by the learning model 532 and the temperature of the substrate associated with the feature amount of the surface shape of the input electrostatic chuck 1 and the recipe information becomes small. For example, the arithmetic unit 51 adjusts the parameters by the error backpropagation method. The arithmetic unit 51 repeats the process using a plurality of data sets recorded in the learning data 533, and performs machine learning of the learning model 532 by adjusting the parameters of the learning model 532. By adjusting the calculation parameters of the learning model 532 in this way, the learned model 432 is generated. The arithmetic unit 51 stores the adjusted final parameters in the storage unit 53. After S12 ends, the learning device 5 ends the process.

[0038] The learned model 432 generated by S11 to S12 is provided in the information processing apparatus 4. For example, the final parameters adjusted in S12 are input into the information processing apparatus 4 through the interface unit 47 and stored in the storage unit 43. The learned model 432 is realized by the arithmetic unit 41 executing information processing using the stored parameters. Note that the information processing apparatus 4 may execute processing as the learning apparatus 5.

[0039] Next, the processing executed by the information processing apparatus 4 will be described. The information processing apparatus 4 performs processing to predict what temperature the substrate 2 fixed to the electrostatic chuck 1 provided in the processing apparatus 31 will reach during substrate processing. FIG. 8 is a flowchart showing an example of the procedure of the temperature prediction processing executed by the information processing apparatus 4. By the arithmetic unit 41 executing information processing according to the computer program 431, the information processing apparatus 4 executes the following processing. The information processing apparatus 4 acquires recipe information, surface shape information, and the usage time of the electrostatic chuck 1 (S21). In S21, recipe information representing the content of a recipe that defines the processing conditions for substrate processing performed by the processing apparatus 31 is input from the control apparatus 32 to the information processing apparatus 4. For example, when the recipe information is input from the control apparatus 32 to the information processing apparatus 4 through the interface unit 47, the arithmetic unit 41 acquires the recipe information.

[0040] The surface shape measuring apparatus 33 uses a laser to 3D scan the surface of the electrostatic chuck 1 provided in the processing apparatus 31 and generates surface shape information representing the positions of a plurality of portions on the surface of the electrostatic chuck 1. In S21, the surface shape information generated by the surface shape measuring apparatus 33 is input from the surface shape measuring apparatus 33 to the information processing apparatus 4. For example, when the surface shape information is input from the surface shape measuring apparatus 33 to the information processing apparatus 4 through the interface unit 47, the arithmetic unit 41 acquires the surface shape information.

[0041] The usage time of the electrostatic chuck 1 is the length of time that the electrostatic chuck 1 has been used in the processing apparatus 31. In S21, for example, the usage time is input from the control apparatus 32 to the information processing apparatus 4. By inputting the usage time through the interface unit 47, the arithmetic unit 41 acquires the usage time of the electrostatic chuck 1. The arithmetic unit 41 may also acquire the usage time of the electrostatic chuck 1 when the user operates the operation unit 45 to input the usage time. The arithmetic unit 41 stores the acquired recipe information, surface shape information, and the usage time of the electrostatic chuck 1 in the storage unit 43.

[0042] Next, the information processing apparatus 4 calculates the feature amount of the surface shape of the electrostatic chuck 1 from the surface shape information (S22). In S22, the arithmetic unit 41 calculates a plurality of types of feature amounts. Based on the surface shape information, the arithmetic unit 41 divides the surface of the electrostatic chuck 1 into a plurality of segments and calculates the warp of the electrostatic chuck 1. The position of each segment is represented by the xyz coordinate values of a plurality of portions on the surface of the electrostatic chuck 1. For example, the warp of the electrostatic chuck 1 is calculated from the distribution of the z coordinate values of each segment. Next, the arithmetic unit 41 calculates the height of the protrusion 11. For example, the height of the protrusion 11 is calculated from the z coordinate value of the top surface and the z coordinate value of the base of each protrusion 11.

[0043] Next, the arithmetic unit 41 calculates the contact area between the protrusion 11 and the substrate. For example, the arithmetic unit 41 calculates the contact area between the plurality of protrusions 11 and the substrate based on the position of each protrusion 11, the overall warp of the electrostatic chuck 1, and the height of each protrusion 11. Further, the arithmetic unit 41 calculates the undulation amount of the seal band 12. For example, the undulation amount of the seal band 12 is calculated from the z coordinate values of each part of the seal band 12. By calculating the feature amount of the surface shape of the electrostatic chuck 1 from the surface shape information, the arithmetic unit 41 acquires the feature amount of the surface shape of the electrostatic chuck 1. The arithmetic unit 41 stores each of the calculated feature amounts of the surface shape of the electrostatic chuck 1 in the storage unit 43.

[0044] The information processing apparatus 4 inputs the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information into the learned model 432 (S23). In S23, the arithmetic unit 41 inputs the acquired feature amount of the surface shape of the electrostatic chuck 1 and the recipe information into the learned model 432. At this time, the arithmetic unit 41 inputs a plurality of types of feature amounts into the learned model 432. The learned model 432 performs calculations in response to the input of the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information, and outputs a plurality of predicted temperatures obtained by predicting the temperatures at a plurality of portions of the substrate.

[0045] Next, the information processing apparatus 4 predicts the temperature of the substrate fixed to the electrostatic chuck 1 (S24). In S24, the arithmetic unit 41 predicts the temperature of the substrate fixed to the electrostatic chuck 1 by acquiring the predicted temperature output by the learned model 432. At this time, the arithmetic unit 41 acquires a plurality of predicted temperatures obtained by predicting the temperatures at a plurality of portions of the substrate. The arithmetic unit 41 stores the acquired predicted temperature in the storage unit 43. Note that the arithmetic unit 41 may display the acquired predicted temperature on the display unit 46. For example, the arithmetic unit 41 may generate a temperature distribution image representing the distribution of a plurality of predicted temperatures on the substrate, and display the temperature distribution image on the display unit 46.

[0046] Next, the information processing apparatus 4 estimates the remaining life of the electrostatic chuck 1 (S25). The remaining life is the length of time during which the electrostatic chuck 1 provided in the processing apparatus 31 can be used, and is the length of time during which the electrostatic chuck 1 can be used from now until it reaches the end of its life. As the usage time of the electrostatic chuck 1 increases, the contact area between the electrostatic chuck 1 and the substrate decreases, and the temperature of the substrate during substrate processing increases. The temperature of the substrate during substrate processing when the electrostatic chuck 1 is in the initial state is defined as the initial temperature. The temperature difference obtained by subtracting the initial temperature from the temperature of the substrate during substrate processing increases as the usage time of the electrostatic chuck 1 increases. When the temperature difference reaches a predetermined upper limit temperature difference, it is assumed that the electrostatic chuck 1 has reached the end of its life.

[0047] The remaining life of the electrostatic chuck 1 is estimated by utilizing the relationship between the usage time of the electrostatic chuck 1 and the temperature difference. FIG. 9 is a schematic graph showing an example of the relationship between the usage time of the electrostatic chuck 1 and the temperature difference. The horizontal axis in the figure indicates time, and the vertical axis indicates the temperature difference. The graph of FIG. 9 shows the relationship between the usage time of the electrostatic chuck 1 and the temperature difference obtained by subtracting the initial temperature from the predicted temperature of the substrate. The temperature difference is proportional to the usage time, and the graph showing the relationship between the usage time and the temperature difference forms a straight line. The time corresponding to the point where the temperature difference reaches the upper limit temperature difference on the straight line is the life of the electrostatic chuck 1. From the usage time of the electrostatic chuck 1 and the temperature difference obtained by subtracting the initial temperature from the predicted temperature of the substrate, the slope of the straight line is obtained, and by extrapolating the straight line, a point where the temperature difference reaches the upper limit temperature difference is obtained, and the life of the electrostatic chuck 1 is obtained. By subtracting the usage time from the life, an estimated value of the remaining life of the electrostatic chuck 1 is calculated.

[0048] In S25, the arithmetic unit 41 specifies the representative value of the predicted temperature. The representative value is a value representing the plurality of acquired predicted temperatures. The representative value of the predicted temperature is, for example, the average value, median value, or mode value of the plurality of predicted temperatures, the predicted temperature of a specific part of the substrate, or the average value of the plurality of predicted temperatures in a specific region of the substrate. The upper limit temperature difference is predetermined and stored in the storage unit 43. Also, the initial temperature is stored in the storage unit 43 in advance. For example, the initial state is actually measured using the electrostatic chuck 1 in the initial state, and the measured initial temperature is stored in the storage unit 43. For example, the representative value of the initial temperature is stored in the storage unit 43, and as the upper limit temperature difference, the temperature difference between the representative value of the temperature of the substrate and the representative value of the initial temperature is stored in the storage unit 43. Note that a plurality of initial temperatures may be stored in the storage unit 43, and the representative value of the initial temperature may be calculated by the arithmetic unit 41.

[0049] In S25, the arithmetic unit 41 calculates the temperature difference obtained by subtracting the initial temperature from the predicted temperature by subtracting the initial temperature from the representative value of the predicted temperature. The arithmetic unit 41 calculates the slope of the straight line showing the relationship between the usage time and the temperature difference as shown in FIG. 9 from the usage time of the electrostatic chuck 1 and the calculated temperature difference. The arithmetic unit 41 calculates the lifetime of the electrostatic chuck 1 by calculating the time at which the temperature difference becomes the upper limit temperature difference on the straight line for which the slope has been calculated. The arithmetic unit 41 calculates the estimated remaining lifetime of the electrostatic chuck 1 by subtracting the usage time from the calculated lifetime. By calculating the estimated remaining lifetime of the electrostatic chuck 1, the arithmetic unit 41 estimates the remaining lifetime of the electrostatic chuck 1.

[0050] The arithmetic unit 41 stores the estimated remaining lifetime in the storage unit 43. The arithmetic unit 41 may display the estimated remaining lifetime on the display unit 46. The arithmetic unit 41 may calculate an estimated value of the remaining lifetime by other methods based on the representative value of the predicted temperature, the usage time, the initial temperature, and the upper limit temperature difference. For example, the arithmetic unit 41 may calculate the lifetime corresponding to the usage time using the ratio between the upper limit temperature difference and the temperature difference obtained by subtracting the initial temperature from the predicted temperature, and calculate the estimated value of the remaining lifetime. After S25 ends, the arithmetic unit 41 ends the temperature prediction process.

[0051] As described in detail above, the information processing apparatus 4 predicts the temperature of the substrate fixed to the electrostatic chuck 1 using the learned model 432 that outputs the predicted temperature of the substrate when the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information are input. Thereby, even if the substrate processing is not actually performed, it is possible to predict to what temperature the temperature of the substrate fixed to the electrostatic chuck 1 will be during the substrate processing based on the surface shape and the recipe of the electrostatic chuck 1. That is, before performing the substrate processing, the temperature of the substrate fixed to the electrostatic chuck 1 can be predicted. For example, it becomes possible to perform the actual substrate processing after confirming that the predicted temperature of the substrate becomes an appropriate temperature.

[0052] Further, the information processing apparatus 4 calculates an estimated remaining life of the electrostatic chuck 1 based on the usage time of the electrostatic chuck 1, the predicted temperature of the substrate, and a predetermined upper limit temperature difference that is the temperature difference obtained by subtracting the initial temperature of the substrate from the temperature of the substrate when the electrostatic chuck 1 reaches the end of its life. The estimated remaining life of the electrostatic chuck 1 can be easily estimated from the predicted temperature of the substrate. Conventionally, it has not been easy to estimate the remaining life of the electrostatic chuck 1, so the electrostatic chuck 1 was replaced long before it reached the end of its life. In the present embodiment, since the estimated remaining life of the electrostatic chuck 1 can be easily estimated, substrate processing using the electrostatic chuck 1 can be performed until immediately before the electrostatic chuck 1 reaches the end of its life. It becomes possible to use the electrostatic chuck 1 for a longer time than in the past, and the cost required for substrate processing is reduced.

[0053] In the processes of S21 to S25, the information processing apparatus 4 acquires surface shape information from the surface shape measuring apparatus 33 and calculates a feature amount of the surface shape of the electrostatic chuck 1. The information processing apparatus 4 may acquire the feature amount of the surface shape of the electrostatic chuck 1 by being input from the outside and perform the processes of S23 to S25 using the acquired feature amount. In the description of S21 to S25, an example in which the surface shape measuring apparatus 33 3D scans the surface of the electrostatic chuck 1 provided in the processing apparatus 31 is shown. The surface shape measuring apparatus 33 may 3D scan the surface of the electrostatic chuck 1 not provided in the processing apparatus 31, and the temperature of the substrate may be predicted based on the surface shape information thus obtained.

[0054] The information processing apparatus 4 can perform additional learning of the learned model 432 using an actual recipe. When performing additional learning, the storage unit 43 stores learning data for performing additional learning of the learned model 432. The learning data records a data set in which recipe information, a feature amount of the surface shape of the electrostatic chuck 1, and the temperature of the substrate fixed to the electrostatic chuck 1 are associated with each other. The recipe information here is information representing the content of the recipe that defines the processing conditions in the substrate processing actually performed by the processing apparatus 31. The feature amount of the surface shape of the electrostatic chuck 1 is a value calculated from the surface shape information generated by the surface shape measuring apparatus 33 for the electrostatic chuck 1 used in the substrate processing. The temperature of the substrate is the temperature measured during the substrate processing. A plurality of data sets are recorded in the learning data.

[0055] FIG. 10 is a flowchart showing an example of the procedure of the additional learning process executed by the information processing apparatus 4. By the arithmetic unit 41 executing information processing according to the computer program 431, the information processing apparatus 4 executes the following processes. By the arithmetic unit 41 reading out the learning data stored in the storage unit 43, the information processing apparatus 4 acquires the learning data (S31). In S31, the information processing apparatus 4 may acquire the learning data by the learning data being input from the outside.

[0056] Next, the information processing apparatus 4 performs additional learning of the learned model 432 using the acquired learning data (S32). In S32, the arithmetic unit 41 inputs the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information recorded in the learning data to the learned model 432. The learned model 432 performs calculations in response to the input of the feature amount of the surface shape of the electrostatic chuck 1 and the recipe information, and outputs a predicted temperature.

[0057] The calculation unit 41 adjusts the calculation parameters of the learned model 432 so that the error between the predicted temperature output by the learned model 432 and the temperature of the substrate associated with the feature amount of the surface shape of the input electrostatic chuck 1 and the recipe information becomes small. For example, the calculation unit 41 adjusts the parameters by the error backpropagation method. The calculation unit 41 repeats the process using a plurality of data sets recorded in the learning data, and performs additional learning of the learned model 432 by adjusting the learned model 432. The calculation unit 41 stores the adjusted final parameters in the storage unit 43. After S32 ends, the information processing device 4 ends the process. The processes of S31 to S32 may be executed by the learning device 5.

[0058] The information processing device 4 performs the processes of S21 to S25 using the learned model 432 that has been additionally learned. Since additional learning of the learned model 432 is performed using the actual recipe, the predicted temperature output by the learned model 432 is a more accurate prediction of the temperature of the substrate when the substrate processing is actually performed. Therefore, the accuracy of predicting the temperature of the substrate fixed to the electrostatic chuck 1 is improved by additional learning of the learned model 432. Further, by estimating the remaining life of the electrostatic chuck 1 using the predicted temperature with improved prediction accuracy, the estimated value of the remaining life becomes a value that more accurately estimates the remaining life. Therefore, the accuracy of estimating the remaining life of the electrostatic chuck 1 is improved.

[0059] The present invention is not limited to the content of the above-described embodiments, and various modifications are possible within the scope shown in the claims. That is, embodiments obtained by combining technical means appropriately modified within the scope shown in the claims are also included in the technical scope of the present invention.

[0060] The matters described in each embodiment can be combined with each other. Also, the independent claims and dependent claims described in the claims can be combined with each other in all possible combinations regardless of the citation form. Further, the claims use a form (multi-claim form) of describing claims that cite two or more other claims, but it is not limited thereto. A form of describing a multi-claim (multi-multi-claim) that cites at least one multi-claim may be used.

Description of Reference Numerals

[0061] 1 Electrostatic chuck 2 Substrate 31 Processing device 32 Control device 33 Surface shape measuring device 4 Information processing device 40 Recording medium 41 Arithmetic unit 43 Storage unit 431 Computer program 432 Learned model 5 Learning device

Claims

1. Obtain a feature amount of the surface shape of an electrostatic chuck for fixing a substrate to be processed, Obtain recipe information representing the content of a recipe that defines the processing conditions of the substrate, Input the obtained feature amount of the surface shape of the electrostatic chuck and the recipe information into a learned model that outputs a predicted temperature of the substrate fixed to the electrostatic chuck when the feature amount of the surface shape of the electrostatic chuck and the recipe information are input, By obtaining the predicted temperature output by the learned model, predict the temperature of the substrate fixed to the electrostatic chuck A computer program that causes a computer to execute the processing.

2. Obtain the usage time of the electrostatic chuck, Based on the obtained usage time, the temperature difference between the initial temperature of the substrate fixed to the electrostatic chuck in the initial state and the obtained predicted temperature, and a predetermined temperature difference between the initial temperature and the temperature of the substrate fixed to the electrostatic chuck that has reached the end of its life, estimate the remaining life of the electrostatic chuck The computer program according to claim 1, which causes a computer to execute the processing.

3. The learned model outputs predicted temperatures of a plurality of portions of the substrate, Obtain the plurality of predicted temperatures output by the learned model, Estimate the remaining life of the electrostatic chuck using a value representing the plurality of predicted temperatures The computer program according to claim 2, which causes a computer to execute the processing.

4. The feature amount is an amount obtained based on the result of 3D scanning the surface of the electrostatic chuck, The computer program according to claim 1.

5. Obtain a feature amount of the surface shape of an electrostatic chuck for fixing a substrate to be processed, Obtain recipe information representing the content of a recipe that defines the processing conditions of the substrate, Input the obtained feature amount of the surface shape of the electrostatic chuck and recipe information into a learned model that outputs a predicted temperature obtained by predicting the temperature of a substrate fixed to the electrostatic chuck when the feature amount of the surface shape of the electrostatic chuck and recipe information are input, Predict the temperature of the substrate fixed to the electrostatic chuck by obtaining the predicted temperature output by the learned model Information processing method.

6. Perform additional learning of the learned model using learning data including the feature amount of the surface shape of the electrostatic chuck used for processing the substrate according to a specific recipe, recipe information indicating the features of the recipe, and the measured value of the temperature of the substrate fixed to the electrostatic chuck The information processing method according to claim 5.

7. Comprising an arithmetic unit, The arithmetic unit is, Obtain the feature amount of the surface shape of the electrostatic chuck for fixing the substrate to be processed, Obtain recipe information representing the content of the recipe that defines the processing conditions of the substrate, Input the obtained feature amount of the surface shape of the electrostatic chuck and recipe information into a learned model that outputs a predicted temperature obtained by predicting the temperature of a substrate fixed to the electrostatic chuck when the feature amount of the surface shape of the electrostatic chuck and recipe information are input, Predict the temperature of the substrate fixed to the electrostatic chuck by obtaining the predicted temperature output by the learned model Information processing apparatus.

Images