Information processing method, experiment method, and computer program

Abstract

Provided are an information processing method, an experiment method, and a computer program that are capable of reducing the number of experiments for modeling a substrate processing apparatus.　The information processing method includes: acquiring a control target model which includes a plurality of control targets in the substrate processing apparatus and in which each of the control targets interacts with any of the other control targets; acquiring a plurality of control contents to be performed on the control targets in order to perform the experiments in the substrate processing apparatus; classifying the plurality of control targets included in the control target model into a plurality of groups according to a predetermined condition; and creating a plurality of combinations in which the control contents different from each other are assigned respectively to the plurality of groups to create a plurality of experiment conditions.

Description

Information Processing Method, Experiment Method, and Computer Program

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

[0002] A substrate processing apparatus performs substrate processing such as etching or film formation on a substrate such as a semiconductor wafer or a glass substrate. With the expansion of the automation of substrate processing apparatuses, there has been a need to model a substrate processing apparatus as a system composed of a plurality of components. Modeling of a substrate processing apparatus can be used for simulation of substrate processing or control of the substrate processing apparatus. Patent Document 1 discloses a technique for controlling substrate processing using modeling. In order to model a substrate processing apparatus, it is necessary to conduct an experiment to confirm what results will be obtained when different controls are applied to the substrate processing apparatus.

[0003] Japanese Patent Application Laid-Open No. 2023-107441

[0004] In an experiment for modeling a substrate processing apparatus, it is necessary to conduct an experiment in which a plurality of types of controls are executed for each of a plurality of control targets to be controlled for substrate processing. For example, when conducting an experiment to control the temperature of a substrate, for each of the N sites included in the substrate, an experiment is conducted in which n types of temperature controls with different parameters such as target temperature are executed, and about n to the power of N experiments are required. However, there is a problem that the number of necessary experiments is too large.

[0005] The present disclosure provides an information processing method, an experiment method, and a computer program capable of reducing the number of experiments in an experiment for modeling a substrate processing apparatus.

[0006] An information processing method according to one aspect of the present disclosure includes a control object model obtained in which a plurality of control objects in a substrate processing apparatus are each control object interacting with some of the other control objects, a plurality of control contents to be performed on the control objects in order to conduct an experiment in the substrate processing apparatus, a plurality of experimental conditions to be created by classifying the plurality of control objects included in the control object model into a plurality of groups according to predetermined conditions, and creating a plurality of combinations in which different control contents are assigned to each of the plurality of groups.

[0007] According to this disclosure, it is possible to provide an information processing method, an experimental method, and a computer program that can reduce the number of experiments in experiments for modeling substrate processing equipment.

[0008] This is a conceptual diagram showing an example configuration of a substrate processing system. This is a block diagram showing an example of the internal configuration of an information processing device. This is a flowchart showing an example of the procedure for creating experimental conditions by an information processing device. This is a schematic diagram showing an example of multiple control objects. This is a schematic diagram showing a first example of a control object model. This is a graph showing an example of a control range. This is a chart showing an example of multiple control contents. This is a schematic diagram showing a first example of the results of graph coloring. This is a schematic diagram showing a second example of a classified set of control objects. This is a chart showing a first example of multiple experimental conditions. This is a schematic diagram showing an example of multiple control objects being distributed parameter circuits. This is a schematic diagram showing a second example of a control object model. This is a schematic diagram showing a second example of the results of graph coloring. This is a chart showing a second example of multiple experimental conditions. This is a flowchart showing an example of the procedure for conducting an experiment performed by a substrate processing system. This is a block diagram showing an example of the internal functional configuration of a learning device. This is a flowchart showing an example of the procedure for processing performed by a learning device.

[0009] Hereinafter, the present disclosure will be described in detail with reference to the drawings illustrating its embodiments. In this embodiment, experimental conditions are created for experiments on the substrate processing apparatus. In the experiment, multiple control operations are performed for each of the multiple control targets to be controlled by the substrate processing apparatus. In this embodiment, experimental conditions are created for conducting experiments efficiently.

[0010] Figure 1 is a conceptual diagram showing an example configuration of a substrate processing system 100. The substrate processing system 100 includes a substrate processing device 21, a control device 22 that controls the substrate processing device 21, and an information processing device 1 that performs processing to create experimental conditions. The substrate processing device 21 performs substrate processing such as etching or film deposition on substrates such as semiconductor wafers, glass substrates, or substrates for flat panel displays. The control device 22 controls the operation of the substrate processing device 21. The information processing device 1 executes an information processing method that creates multiple types of experimental conditions. The control device 22 controls the substrate processing device 21 according to the experimental conditions created by the information processing device 1, so that the substrate processing device 21 performs experiments according to the experimental conditions.

[0011] Figure 2 is a block diagram showing an example of the internal configuration of the information processing device 1. The information processing device 1 is configured using a computer such as a personal computer or a server device. The information processing device 1 includes an arithmetic unit 11, a memory 12, a storage unit 13, a reading unit 14, an operation unit 15, a display unit 16, and an input / output unit 17. The arithmetic unit 11 is a processor and is configured using, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a multi-core CPU. The arithmetic unit 11 may also be configured using a quantum computer. The memory 12 stores temporary data generated in connection with calculations. The memory 12 is, for example, RAM (Random Access Memory). The storage unit 13 is non-volatile and is, for example, a hard disk or a non-volatile semiconductor memory. The reading unit 14 reads information from a recording medium 10 such as an optical disc or portable memory.

[0012] The operation unit 15 accepts input of information such as text by receiving operations from the user. The operation unit 15 is, for example, a keyboard, a pointing device, or a touch panel. The display unit 16 displays images. The display unit 16 is, for example, a liquid crystal display or an EL display (Electroluminescent Display). The operation unit 15 and the display unit 16 may be integrated. The input / output unit 17 performs data input and output. The input / output unit 17 is, for example, an input / output interface or a communication unit.

[0013] The arithmetic unit 11 causes the reading unit 14 to read the computer program (program product) 131 recorded on the recording medium 10, and stores the read computer program 131 in the storage unit 13. The arithmetic unit 11 executes processing to realize the functions of the information processing device 1 according to the computer program 131. The computer program 131 may be stored in the storage unit 13 in advance, or it may be downloaded from outside the information processing device 1. In this case, the information processing device 1 does not need to have a reading unit 14.

[0014] The computer program 131 can be deployed on a single computer, at a single site, or distributed across multiple sites and run on multiple computers interconnected by a communication network. That is, the information processing device 1 may consist of multiple computers, and the computer program 131 may run on multiple computers connected via a communication network. The information processing device 1 may also be configured using a cloud server.

[0015] The processing steps described below for executing the information processing method can be performed on multiple computers. The processing steps can also be performed on different computers. The data used during processing may be stored on multiple computers. The processing steps can also be performed using a virtual machine. The processing steps may be performed by multiple arithmetic units. The processing steps may also be performed by different arithmetic units. For example, part of the processing may be performed on one computer, and other parts on other computers.

[0016] Figure 3 is a flowchart illustrating an example of the procedure for creating experimental conditions by the information processing device 1. Hereinafter, steps will be abbreviated as S. The information processing device 1 performs the following processes by executing information processing in accordance with the computer program 131 by the calculation unit 11. The information processing device 1 acquires a plurality of control objects to be controlled in the experiment for modeling the substrate processing device 21 (S11). A control object is some mechanism included in the information processing device 1 that operates for substrate processing. Each control object can interact with other control objects.

[0017] Figure 4 is a schematic diagram showing an example of multiple controlled objects. In the example shown in Figure 4, the controlled object is a temperature control mechanism that adjusts the temperature of a substrate, and the multiple controlled objects include controlled objects C1, C2, C3, C4, C5, C6, C7, and C8. The multiple controlled objects are arranged in a planar configuration, and each controlled object adjusts the temperature of each of the multiple areas contained within the substrate. For example, in a temperature control mechanism having a mounting surface for holding a substrate, each controlled object includes a temperature sensor for measuring the temperature of the substrate, a heater for heating the substrate, and a cooler for blowing cooling gas onto the substrate from the back. In Figure 4, the substrate whose temperature is controlled is shown by a dashed line. The boundaries of the controlled objects shown in the figure are the boundaries of the areas of the substrate whose temperature is adjusted by each controlled object.

[0018] The temperature of one area on a substrate affects the temperature of other areas through heat conduction. Therefore, the result of adjusting the temperature of one area on the substrate by one controlled object influences the result of adjusting the temperature of other areas by other controlled objects. In other words, controlled objects interact with each other. However, the influence of heat conduction decreases as the distance between them increases. Therefore, a controlled object can be considered to interact with other controlled objects that are close to it, and not with those that are far away. For example, a controlled object interacts with other adjacent controlled objects, but not with other controlled objects that are not adjacent. In the example shown in Figure 4, controlled object C1 interacts with controlled objects C2, C5, and C6, but does not interact with controlled objects C3, C4, C7, and C8. In this way, each controlled object interacts with some of the other controlled objects, and does not interact with the other controlled objects.

[0019] In S11, the user inputs multiple control objects corresponding to the configuration of the substrate processing device 21 to the information processing device 1 by operating the operation unit 15. For example, the user inputs the arrangement of control objects C1 to C8 as shown in Figure 4. The exact boundaries of each control object are not input; only the approximate position of each control object may be input. In S11, the calculation unit 11 acquires the input multiple control objects and stores the acquired multiple control objects in the storage unit 13.

[0020] The information processing device 1 then generates a controlled object model that includes the acquired control objects and defines the relationship between each control object and other control objects (S12). Figure 5 is a schematic diagram showing a first example of a controlled object model. In the example shown in Figure 5, the controlled object model is represented as a planar graph that includes nodes N1, N2, N3, N4, N5, N6, N7, and N8 corresponding to the control objects C1, C2, C3, C4, C5, C6, C7, and C8, and edges connecting two nodes corresponding to two interacting control objects. Nodes are represented as circles, and edges are represented as line segments connecting two nodes. In S12, the calculation unit 11 generates the controlled object model by replacing each control object with a node and generating edges connecting two nodes corresponding to adjacent control objects.

[0021] If the precise boundaries of each controlled object are not input, the calculation unit 11 may generate boundaries and generate a controlled object model according to the generated boundaries. For example, the calculation unit 11 may generate the boundaries of each controlled object using a spatial division method such as a Voronoi partition. The calculation unit 11 may also generate a controlled object model without generating boundaries. For example, if the distance between two controlled objects is less than a predetermined distance, the calculation unit 11 may generate a controlled object model by generating an edge connecting two nodes corresponding to the two controlled objects.

[0022] By generating a controlled target model in S12, the information processing device 1 acquires a controlled target model. The calculation unit 11 stores the acquired controlled target model in the storage unit 13. The calculation unit 11 may also display the controlled target model on the display unit 16. Alternatively, the information processing device 1 may acquire a controlled target model by receiving input, rather than generating it. For example, a user may input a controlled target model as shown in Figure 5 to the information processing device 1 by operating the operation unit 15. Alternatively, a controlled target model may be input to the information processing device 1 via the input / output unit 17.

[0023] The information processing device 1 then acquires the control range of the controlled object (S13). The control range is the range of control parameters that are adjusted to control the controlled object. For example, the control parameters are the heater temperature and the cooling gas temperature. Figure 6 is a graph showing an example of the control range. In Figure 6, the horizontal axis shows the cooling gas temperature, and the vertical axis shows the heater temperature. The control range is the range in which multiple types of control parameters can be adjusted, and is the range in which combinations of values ​​of multiple types of control parameters can be taken when the substrate processing device 21 conducts an experiment. The control range is determined according to the configuration or settings of the substrate processing device 21. For example, adjusting the control parameters outside the control range is not recommended because it will cause malfunctions in the information processing device 1.

[0024] In S13, the user inputs the control range to the information processing device 1 by operating the operation unit 15. For example, the user inputs a control range as shown in Figure 6. A control range arbitrarily determined regardless of the configuration or settings of the substrate processing device 21 may also be input. In S13, the calculation unit 11 acquires the input control range and stores the acquired control range in the storage unit 13.

[0025] The information processing device 1 then generates multiple control contents according to the control range (S14). The control contents are the control contents that should be performed on each control target when conducting experiments with the substrate processing device 21. Each control contents includes combinations of values ​​for multiple types of control parameters. In S14, the calculation unit 11 generates multiple control contents by generating multiple combinations of values ​​for multiple types of control parameters.

[0026] In S14, the calculation unit 11 extracts multiple combinations of values ​​for multiple types of control parameters included in the control range. For example, the calculation unit 11 extracts combinations in which the value of each control parameter is at the upper or lower limit of the control range, extracts combinations in which the value of any of the control parameters is changed by a predetermined amount within the control range, and repeats the process of changing the control parameter values ​​and extracting combinations. For example, the calculation unit 11 may randomly extract multiple combinations of values ​​for multiple types of control parameters from within the control range. For example, the calculation unit 11 may accept a specification of a combination of values ​​for multiple types of control parameters included in the control range by the user operating the operation unit 15, and extract the accepted combination.

[0027] In S14, the calculation unit 11 ensures that the number of combinations of values ​​for multiple types of extracted control parameters does not exceed a predetermined upper limit. For example, if the number of extracted combinations reaches the upper limit, the calculation unit 11 terminates the extraction of combinations. For example, if the number of extracted combinations exceeds the upper limit, the calculation unit 11 deletes one of the combinations according to a predetermined rule. For example, the calculation unit 11 may accept a specification of a combination to be deleted by the user operating the operation unit 15, and delete the specified combination.

[0028] In this way, the calculation unit 11 determines multiple combinations of values ​​for multiple types of control parameters and generates multiple control contents, each containing a combination of values ​​for multiple types of control parameters. The combinations of values ​​for multiple types of control parameters differ for each control content. Each control content controls the target to adjust the control parameters to predetermined values. Let n be the number of predetermined combinations of values ​​for multiple types of control parameters. Therefore, the calculation unit 11 generates n control contents.

[0029] Figure 7 is a diagram illustrating examples of multiple control actions. Each control action is denoted as Control Action M1, M2, ..., Mn. Each control action is associated with values ​​for multiple types of control parameters. The control parameters, heater temperature and cooling gas temperature, are expressed as relative values ​​to the temperature measured by the temperature sensor. Each control action controls the target system to adjust the control parameters to the associated values. In the example shown in Figure 7, Control Action M1 controls the heater so that the heater temperature is 50 degrees higher than the temperature measured by the temperature sensor, and controls the cooler so that the cooling gas temperature is 50 degrees lower than the temperature measured by the temperature sensor.

[0030] By generating multiple control contents in S14, the information processing device 1 acquires multiple control contents. The calculation unit 11 stores the acquired multiple control contents in the storage unit 13. The calculation unit 11 may also display the multiple control contents on the display unit 16. The information processing device 1 may acquire multiple control contents by receiving multiple control contents as input, rather than generating them. For example, a user may input multiple control contents as shown in Figure 7 to the information processing device 1 by operating the operation unit 15. For example, multiple control contents may be input to the information processing device 1 via the input / output unit 17.

[0031] Note that there may be only one type of control parameter. In this case, the control range is the range of one type of control parameter, and the control content includes the value of one type of control parameter. The value of the control parameter differs depending on the control content. Even in this case, multiple control content can be obtained.

[0032] The information processing device 1 then classifies the multiple control objects included in the control object model into multiple groups (S15). In S15, the calculation unit 11 classifies the multiple control objects included in the control object model such that two interacting control objects are classified into different groups. The calculation unit 11 also classifies the multiple control objects such that the number of groups is kept as small as possible.

[0033] In S15, the calculation unit 11 performs classification by coloring the control object model, which is represented as a planar graph as shown in Figure 5. The calculation unit 11 colors the graph so that two nodes connected by edges corresponding to two interacting control objects do not have the same color, and so that the number of colors used is kept to a minimum. The calculation unit 11 also classifies multiple control objects into multiple groups such that multiple control objects corresponding to multiple nodes colored with the same color are classified into the same group, and multiple control objects corresponding to multiple nodes colored with different colors are classified into different groups. Let the number of groups be k. According to the four-color theorem, for any planar graph, the maximum number of colors required to color two nodes connected by edges so that they do not have the same color is 4. Therefore, k is a natural number less than or equal to 4.

[0034] Figure 8 is a schematic diagram showing a first example of the graph coloring results. Of the eight nodes shown in Figure 5, nodes N1, N3, and N8 are colored red in Figure 8, nodes N2, N5, and N7 are colored blue, and nodes N4 and N6 are colored green. In Figure 8, red-colored nodes are indicated by white circles, blue-colored nodes by black circles, and green-colored nodes by double circles. The arithmetic unit 11 stores the results of classifying the multiple control targets into multiple groups in the storage unit 13.

[0035] The calculation unit 11 may classify multiple control objects into multiple groups such that the number of groups exceeds four. For example, a control object may be able to interact with an adjacent control object, and there may be other control objects that can interact with objects other than adjacent ones. In this case, the number of colors required to color two nodes connected by edges corresponding to two interacting control objects so that they do not have the same color may exceed four. For example, the control object model may be represented in three dimensions instead of a two-dimensional graph. In this case as well, the number of colors required to color two nodes connected by edges so that they do not have the same color may exceed four.

[0036] In S15, the arithmetic unit 11 may perform a process to classify multiple control objects into multiple groups in such a way that the probability of two interacting control objects being classified into the same group is reduced. For example, the arithmetic unit 11 may classify multiple control objects by repeatedly coloring multiple nodes with multiple colors and selecting the coloring that minimizes the number of two nodes connected by edges that are the same color. In this case, although the probability is small, there is still a possibility that two interacting control objects will be classified into the same group.

[0037] The information processing device 1 then displays a plurality of classified control objects (S16). In S16, the calculation unit 11 displays the plurality of control objects on the display unit 16 in different display formats according to the classified group. For example, the calculation unit 11 displays a graph-colored control object model on the display unit 16 as shown in Figure 8. At this time, the calculation unit 11 colors each node in a different color according to the classified group and then displays the control object model on the display unit 16. By displaying the graph-colored control object model, a plurality of classified control objects are displayed.

[0038] Figure 9 is a schematic diagram showing a second display example of multiple classified control objects. It shows multiple control objects C1, C2, C3, C4, C5, C6, C7, and C8, which are areas included in the temperature control mechanism. Each control object is colored differently according to its classified group. In Figure 9, red is shown as blank, and blue and green are shown with different hatching. Multiple control objects may be displayed in other display formats. By displaying each control object by group, the user can understand how each control object is classified.

[0039] The information processing device 1 then creates experimental conditions (S17) by assigning control content to each group into which multiple control targets have been classified. In S17, the calculation unit 11 creates multiple combinations of assigning different control content to each group, thereby creating multiple experimental conditions. The experimental conditions define the conditions for the experiment to be performed by the substrate processing device 21. By assigning control content to each group, the calculation unit 11 associates control content with each group, and the control content is associated with the nodes classified into each group, and with the control targets corresponding to those nodes. The calculation unit 11 may also perform processing to associate control content with edges.

[0040] Figure 10 is a diagram showing a first example of multiple experimental conditions. Assume that multiple control subjects are classified into one of three groups: red, blue, and green. The control subjects classified into each group are shown. Each group is associated with one of the control contents M1, M2, ..., Mn. In Figure 10, multiple experimental conditions with different combinations of which of the multiple control contents are associated with each group are referred to as experimental condition E1, experimental condition E2, experimental condition E3, ...

[0041] In experimental condition E1, control content M1 is associated with group red, control content M2 with group blue, and control content M3 with group green. This means that in the experiment, control according to control content M1 is performed on the control targets classified as group red, control according to control content M2 is performed on the control targets classified as group blue, and control according to control content M3 is performed on the control targets classified as group green. Each experimental condition contains multiple control contents associated with multiple groups, and the combination of multiple groups and multiple control contents differs depending on the experimental condition.

[0042] Since each group is assigned different control parameters, two interacting controlled objects will be assigned different control parameters. For example, in the multiple controlled objects shown in Figure 4, two adjacent controlled objects are assigned different control parameters. Therefore, in the experiment, two interacting controlled objects will be subjected to different controls. Even in a classification system where the probability of two interacting controlled objects being classified into the same group is small, the probability of two interacting controlled objects being assigned the same control parameter is small, and therefore the probability of them being subjected to the same control in the experiment is small.

[0043] When the same control is applied to two interacting controlled objects, the same phenomenon occurs in both, making it easy to predict the outcome. When different controls are applied to two interacting controlled objects, different phenomena occur, and these phenomena influence each other, making it more difficult to predict the outcome. Therefore, applying different controls to two interacting controlled objects yields more information from the experiment and increases its value compared to applying the same control. Accordingly, assigning different control settings to each group can increase the value of the experiment.

[0044] The number of experimental conditions that can be created in S17 is the number of permutations obtained by extracting and arranging several control contents from among a plurality of control contents. Since the number of groups is k and the number of control contents is n, the number of permutations is nPk. That is, the information processing apparatus 1 creates nPk different experimental conditions. Assuming that the number of controlled objects is N, when the experimental conditions are created without omission, the number of experimental conditions is n to the Nth power. Among the n to the Nth power different experimental conditions, there are those in which the same control is performed on two interacting controlled objects. Since the number of groups k is smaller than the number of controlled objects N and is at most 4, nPk is less than n to the Nth power. Therefore, in the present embodiment, the number of experimental conditions is reduced. Even in a form where k exceeds 4, the number of experimental conditions is reduced. The calculation unit 11 stores the created multiple experimental conditions in the storage unit 13.

[0045] Next, the information processing apparatus 1 displays the created multiple experimental conditions (S18). In S18, the calculation unit 11 displays the contents of the created multiple experimental conditions on the display unit 16. For example, the calculation unit 11 displays a chart showing the contents of the multiple experimental conditions on the display unit 16 as shown in FIG. 10. In S18, the calculation unit 11 may simultaneously display the multiple controlled objects classified into multiple groups and the multiple experimental conditions on the display unit 16. By displaying the multiple experimental conditions, the user can know the contents of each experimental condition. When S18 ends, the information processing apparatus 1 ends the process of creating the experimental conditions.

[0046] In the above explanation, an example was shown where the controlled object is a temperature control mechanism that adjusts the temperature of a substrate. However, the controlled object may be any other mechanism that operates for substrate processing. For example, multiple controlled objects may be multiple strain adjustment mechanisms that adjust the strain of each area by individually pressing multiple areas of the substrate. For example, a strain adjustment mechanism includes an actuator that presses the substrate and a sensor that measures the amount of displacement of the pressed substrate in the direction of the pressure. The control content is the content of the method for controlling the strain adjustment mechanism. The control content includes, as a control parameter, the amount of displacement of the substrate caused by the controlled object pressing an area of ​​the substrate. The information processing device 1 can create multiple experimental conditions for conducting an experiment to adjust the strain of a substrate using a strain adjustment mechanism by processing S11 to S18.

[0047] The controlled object may be a circuit. Figure 11 is a schematic diagram showing an example where multiple controlled objects are distributed-parameter circuits. The distributed-parameter circuit shown in Figure 11 is an electrical circuit in which multiple circuits, each containing capacitance and reactance, are connected to each other. The distributed-parameter circuit includes a voltage application and data acquisition unit that applies voltage to each circuit and measures the current and voltage output from each circuit. Multiple circuits included in the distributed-parameter circuit, each forming a part of the distributed-parameter circuit, are connected to each other. Two directly connected circuits interact with each other. Two circuits that are not directly connected can be considered not to interact with each other. Each circuit is a controlled object C1, C2, C3, ... The circuits may contain inductance. Figure 11 shows an example where multiple circuits are connected in one dimension, but the multiple circuits may be connected in two or three dimensions.

[0048] The plurality of circuits included in the distributed constant circuit correspond to the plurality of electric circuits included in the substrate processing apparatus 21. By controlling the plurality of controlled objects included in the distributed constant circuit, an experiment on the operation of the substrate processing apparatus 21 can be performed. For example, by examining the circuit constants included in each circuit, an experiment for examining the characteristics of the operation of the substrate processing apparatus 21 can be performed. In the present embodiment, by the processes of S11 to S18, a plurality of sets of experimental conditions can be created according to the connection relationship of the circuits.

[0049] FIG. 12 is a schematic diagram showing a second example of a controlled object model. In the example shown in FIG. 12, nodes N1, N2, N3, N4, N5, and N6 corresponding to controlled objects C1, C2, C3, C4, C5, and C6 are represented by a plurality of parallel horizontal lines. Also, an edge connecting two nodes is represented by a vertical line. For example, node N1 is connected to node N3 by an edge. This represents that the controlled object C1 corresponding to node N1 is connected to the controlled object C3 corresponding to node N3. Thus, the connection relationship of the plurality of controlled objects included in the distributed constant circuit is represented by the control model.

[0050] FIG. 13 is a schematic diagram showing a second example of the result of graph coloring. In FIG. 13, among the six nodes shown in FIG. 12, nodes N1 and N2 are colored red, nodes N3, N4, and N5 are colored blue, and node N6 is colored green. That is, the controlled objects C1 and C2 are classified into group red, the controlled objects C3, C4, and C5 are classified into group blue, and the controlled object C6 is classified into group green. In FIG. 13, the nodes colored red are shown by solid lines, the nodes colored blue are shown by broken lines, and the nodes colored green are shown by one-dot chain lines.

[0051] Two nodes connected by an edge are classified into different groups. In a classification system where the probability of two interacting controlled objects being classified into the same group is small, it is possible that two nodes connected by an edge may be classified into the same colored group. In S16, the information processing device 1 displays a graph-colored controlled object model as shown in Figure 13. By coloring each node with a different color according to the classified group and then displaying the graph-colored controlled object model, multiple classified controlled objects are displayed.

[0052] Figure 14 is a diagram showing a second example of multiple experimental conditions. Controlled objects are shown, each classified into a group. Each group is associated with a control function. As an example of a control function, Figure 14 shows the voltage applied to the controlled circuit. Step(x,y) represents a voltage that changes in a step-like manner over time, with a value of zero from time 0 to time x and a value of y from time x onward. Sine(x,y) represents an AC voltage with a period of x and a maximum value of y. Each experimental condition includes multiple control functions associated with multiple groups, and the combination of multiple groups and multiple control functions differs depending on the experimental condition. As described above, even when the controlled object is a circuit, multiple experimental conditions can be created by processing S11 to S18.

[0053] The controlled objects in this embodiment may include controls other than the temperature control mechanism, strain adjustment mechanism, or circuits included in the distributed constant circuit. The controlled object may also be an operating mechanism or electrical circuit included in the substrate processing apparatus 21, other than the temperature control mechanism, strain adjustment mechanism, or circuits included in the distributed constant circuit.

[0054] The substrate processing device 21 conducts the experiment using the experimental conditions created by the information processing device 1. Figure 15 is a flowchart showing an example of the procedure for processing an experiment performed by the substrate processing system 100. The information processing device 1 inputs multiple sets of experimental conditions into the control device 22. For example, data representing the experimental conditions is output from the input / output unit 17 and input into the control device 22. The control device 22 stores the multiple sets of experimental conditions that have been input. The control device 22 causes the substrate processing device 21 to execute the experiment according to the multiple sets of experimental conditions (S21).

[0055] In S21, the control device 22 controls the operation of multiple control objects included in the substrate processing apparatus 21 according to the experimental conditions. The experimental conditions include control content associated with each control object, and the control device 22 controls the operation of each control object to execute the control content associated with each control object. For example, the control content includes the value of a control parameter, and the control device 22 controls the operation of the control object so that the value of the control parameter becomes the value included in the control content. In this way, the substrate processing apparatus 21 performs the experiment according to the experimental conditions.

[0056] The control device 22 causes the substrate processing device 21 to perform at least one experiment according to one set of experimental conditions. The control device 22 also causes the substrate processing device 21 to perform experiments according to multiple sets of experimental conditions. In other words, multiple experiments are performed in the substrate processing device 21 according to multiple sets of experimental conditions. In each experiment, the substrate processing device 21 acquires experimental results. For example, experimental results may include the results of adjusting the temperature of multiple areas included in the substrate, the results of adjusting the strain of multiple areas included in the substrate, or the results of examining the circuit constants of multiple circuits included in a distributed constant circuit.

[0057] The control device 22 stores the experimental results acquired by the substrate processing device 21 (S22). The substrate processing device 21 acquires multiple experimental results corresponding to multiple experimental conditions, and in S22, the control device 22 stores multiple experimental results. For example, the control device 22 stores the experimental conditions in association with the results of the experiment performed according to those experimental conditions. The experimental results may be input from the control device 22 to the information processing device 1 and stored in the information processing device 1. After S22 is completed, the substrate processing system 100 terminates the experimental processing.

[0058] From the obtained experimental conditions and results, a trained model can be generated that outputs what results can be obtained when certain controls are applied to the substrate processing apparatus 21. The trained model is generated by training by the learning device 3. Figure 16 is a block diagram showing an example of the internal functional configuration of the learning device 3. The learning device 3 is a computer such as a server or a personal computer. The learning device 3 includes an arithmetic unit 31, a memory 32, a storage unit 33, a reading unit 34, an operation unit 35, and a display unit 36. The arithmetic unit 31 is configured using, for example, a CPU, GPU, or multi-core CPU. The arithmetic unit 31 may also be configured using a quantum computer. The memory 32 stores temporary data generated in conjunction with calculations. The memory 32 is, for example, RAM. The reading unit 34 reads information from a recording medium 30 such as an optical disc or portable memory. The storage unit 33 is non-volatile and is, for example, a hard disk or a non-volatile semiconductor memory.

[0059] The operation unit 35 accepts information input by receiving operations from the user. The operation unit 35 is, for example, a keyboard, a pointing device, or a touch panel. The display unit 36 ​​displays an image. The display unit 36 ​​is, for example, a liquid crystal display or an EL display.

[0060] The arithmetic unit 31 causes the reading unit 34 to read the computer program (program product) 331 recorded on the recording medium 30, and stores the read computer program 331 in the storage unit 33. The arithmetic unit 31 executes processing to realize the functions of the learning device 3 according to the computer program 331. The computer program 331 may be stored in the storage unit 33 in advance, or it may be downloaded from outside the learning device 3. In this case, the learning device 3 does not need to have a reading unit 34.

[0061] The computer program 331 can be deployed on a single computer, at a single site, or distributed across multiple sites and run on multiple computers interconnected by a communication network. That is, the learning device 3 may consist of multiple computers, and the computer program 331 may run on multiple computers connected via a communication network. The learning device 3 may also be configured using a cloud server.

[0062] The processes described below for generating a trained model can be executed on multiple computers. Each step of the process can be executed on different computers. The data used during the process may be stored on multiple computers. Each step of the process can also be executed using a virtual machine. Each step of the process may be executed by multiple processing units. Each step of the process may be executed by different processing units. For example, part of the process may be executed on one computer, and other parts of the process may be executed on other computers.

[0063] The learning device 3 generates a trained model by performing machine learning. The learning device 3 includes a training model 332, which serves as the basis for the trained model. The learning device 3 generates a trained model by training the training model 332. The training model 332 is realized by the arithmetic unit 31 executing processing according to the computer program 331. The training model 332 is realized using a neural network. The training model 332 may also be configured using hardware.

[0064] The memory unit 33 stores training data 333 for training the learning model 332 and generating a trained model. The training data 333 includes experimental conditions created by the information processing device 1 and experimental results obtained from experiments conducted by the substrate processing device 21. The training data 333 contains experimental conditions and experimental results that are related to each other, and includes multiple sets of experimental conditions and experimental results.

[0065] Figure 17 is a flowchart showing an example of the procedure performed by the learning device 3. The learning device 3 performs the following processes by having the arithmetic unit 31 perform information processing according to the computer program 331. The learning device 3 acquires the training data 333 by having the arithmetic unit 31 read the training data 333 stored in the storage unit 33 (S31). In S31, the learning device 3 may acquire the training data 333 by receiving the training data 333 from an external source.

[0066] The learning device 3 then trains the learning model 332 using the training data 333 to generate a trained model that outputs experimental results when experimental conditions are input (S32). In S32, the calculation unit 31 inputs the experimental conditions contained in the training data 333 into the learning model 332, which is the basis of the trained model. The learning model 332 performs calculations according to the input experimental conditions and outputs experimental results.

[0067] The calculation unit 31 adjusts the calculation parameters of the learning model 332 so that the error between the experimental results output by the learning model 332 and the experimental results associated with the input experimental conditions in the training data 333 is minimized. For example, the calculation unit 31 adjusts the parameters using backpropagation. The calculation unit 31 performs machine learning on the learning model 332 by repeatedly processing using multiple sets of experimental conditions and experimental results recorded in the training data 333 and adjusting the parameters of the learning model 332. By adjusting the calculation parameters of the learning model 332 in this way, a trained model is generated that outputs experimental results when experimental conditions are input. The calculation unit 31 stores the final adjusted parameters in the storage unit 33. After S32 is completed, the learning device 3 terminates processing.

[0068] The trained model generated by steps S31 to S32 is stored in the information processing device 1. For example, the final parameters adjusted in S32 are input to the information processing device 1 through the input / output unit 17 and stored in the storage unit 13. The trained model is realized when the calculation unit 11 performs information processing using the stored parameters. The information processing device 1 may also perform processing as a learning device 3.

[0069] The information processing device 1 can model the substrate processing device 21 using the generated trained model. The information processing device 1 inputs arbitrary experimental conditions into the trained model, the trained model outputs experimental results corresponding to the input experimental conditions, and the information processing device 1 acquires the output experimental results. For example, experimental results can be obtained according to experimental conditions that include performing control on an arbitrary control target with arbitrary control content. Therefore, the information processing device 1 can simulate what kind of results can be obtained when arbitrary control is performed on multiple control targets included in the substrate processing device 21.

[0070] As detailed above, in this embodiment, the information processing device 1 acquires a control object model that includes multiple control objects, each of which interacts with a part of the other control objects, and classifies the multiple control objects into multiple groups. Furthermore, the information processing device 1 creates multiple experimental conditions by creating multiple combinations of assigning different control contents to each group. If the number of control objects is N and the number of control contents is n, then if all experimental conditions are created, the number of experimental conditions will be n to the power of N. In this embodiment, the number of experimental conditions is less than n to the power of N, and the number of experiments for modeling the substrate processing device 21 can be reduced.

[0071] In this embodiment, the information processing device 1 classifies multiple control objects such that the probability of two interacting control objects being classified into the same group is small, or that two interacting control objects are classified into different groups. Under experimental conditions, different control contents are assigned to different groups. Therefore, in the experiment, two interacting control objects are controlled with different control contents. Applying different controls to two interacting control objects rather than applying the same control to them yields more information from the experiment, thus increasing the value of the experiment. Consequently, in this embodiment, more information can be obtained from a single experiment, allowing for more efficient experimentation.

[0072] The present invention is not limited to the embodiments described above, and various modifications are possible within the scope of the claims. That is, embodiments obtained by combining technical means that have been appropriately modified within the scope of the claims are also included in the technical scope of the present invention.

[0073] The matters described in each embodiment can be combined with each other. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. Moreover, although the claims do not use a form in which claims referencing two or more other claims (multi-claim form), they are not limited to this. They may be described using a multi-claim form, or a form in which multi-claims referencing at least one multi-claim (multi-multi-claim).

[0074] 100 PCB processing system 1 Information processing device 10 Recording medium 11 Calculation unit 13 Storage unit 131 Computer program 21 PCB processing device 22 Control device 3 Learning device

Claims

1. An information processing method for creating multiple experimental conditions by obtaining a control object model that includes multiple control objects in a substrate processing apparatus, in which each control object interacts with some of the other control objects; obtaining multiple control contents to be performed on the control objects in order to conduct an experiment in the substrate processing apparatus; classifying the multiple control objects included in the control object model into multiple groups according to predetermined conditions; and creating multiple combinations of assigning different control contents to each of the multiple groups.

2. The information processing method according to claim 1, which classifies a plurality of controlled objects included in the controlled object model into a plurality of groups such that the probability of two interacting controlled objects being classified into the same group is small, or that two interacting controlled objects are classified into different groups.

3. The information processing method according to claim 1, wherein the controlled object model is represented by a planar graph including nodes corresponding to controlled objects and edges connecting two nodes corresponding to two interacting controlled objects, and the plurality of controlled objects included in the controlled object model are classified into a plurality of groups by performing graph coloring according to the predetermined conditions.

4. The information processing method according to claim 3, wherein graph coloring is performed so that two nodes connected by edges do not have the same color and the number of colors used is kept to a minimum, and the control objects included in the control object model are classified into multiple groups such that multiple control objects corresponding to multiple nodes colored with the same color are classified into the same group, and multiple control objects corresponding to multiple nodes colored with different colors are classified into different groups.

5. The information processing method according to claim 3, which assigns control content to each group by associating control content with each node or edge.

6. The information processing method according to claim 1, wherein experimental conditions are created in a number of permutations obtained by selecting and rearranging several control contents from the plurality of control contents.

7. The information processing method according to claim 1, wherein each control content includes a combination of values ​​for one or more types of control parameters, and the multiple control contents are obtained by acquiring multiple combinations of values ​​for one or more types of control parameters from within the range in which the substrate processing device can adjust one or more types of control parameters.

8. The information processing method according to claim 1, wherein each controlled object is a temperature control mechanism that adjusts the temperature of each area of ​​the substrate, each control content is the content of a method for controlling each temperature control mechanism, and the value of the control parameter includes a temperature value.

9. The information processing method according to claim 1, wherein each controlled object is a strain adjustment mechanism that adjusts the distortion of each area by pressing each area of ​​the substrate, each control content is the content of the method for controlling each strain adjustment mechanism, and the value of the control parameter includes the displacement amount of the pressed substrate.

10. The information processing method according to claim 1, wherein the plurality of controlled objects included in the controlled object model is a distributed constant circuit including a plurality of circuits corresponding to a plurality of electrical circuits of the substrate processing apparatus, the control content of each is the content of a method for controlling each circuit included in the distributed constant circuit, and the value of the control parameter includes the value of the voltage or current supplied to each circuit.

11. The information processing method according to claim 1, which displays a plurality of control objects included in the control object model in different display formats according to classified groups.

12. The information processing method according to claim 1, which involves obtaining experimental results obtained by conducting experiments in the substrate processing apparatus according to multiple experimental conditions created, and training data including the multiple experimental conditions, and generating a trained model that outputs experimental results when experimental conditions are input by learning using the training data.

13. An experimental method for performing experiments using a substrate processing apparatus according to multiple experimental conditions created by the information processing method described in claim 1.

14. A computer program that causes a computer to perform the following processes: obtain a control object model in which a plurality of control objects in a substrate processing apparatus are included, each control object interacting with some of the other control objects; obtain a plurality of control contents to be performed on the control objects in order to conduct an experiment in the substrate processing apparatus; classify the plurality of control objects included in the control object model into a plurality of groups according to predetermined conditions; and create a plurality of combinations in which different control contents are assigned to each of the plurality of groups, thereby creating a plurality of experimental conditions.

Images