Manufacturing condition design support system, manufacturing condition design support method, manufacturing condition design support program, and manufacturing method

Abstract

The present invention provides support appropriate for designing manufacturing conditions. This system for supporting the design of manufacturing conditions in a process for manufacturing a SiC epitaxial wafer by a CVD method comprises: a storage unit that stores a trained model that has been trained by using, from among input / output data when a simulation to predict in-furnace states is executed by a simulation device for said process, input data related to specific manufacturing conditions and output data regarding specific items indicating in-furnace states; a read-out unit that reads out, from the storage unit, a trained model corresponding to designated specific items; an execution unit that, when the input data related to the specific manufacturing conditions is input, executes the read-out trained model, and predicts the output data of the specific items indicating in-furnace states; and a display unit that displays the designated specific items, the input data related to the specific manufacturing conditions that have been input, and the predicted output data.

Description

Manufacturing Condition Design Support System, Manufacturing Condition Design Support Method, Manufacturing Condition Design Support Program, and Manufacturing Method

[0001] The present disclosure relates to a manufacturing condition design support system, a manufacturing condition design support method, a manufacturing condition design support program, and a manufacturing method.

[0002] In the process of manufacturing a SiC (silicon carbide) epitaxial wafer by CVD (Chemical Vapor Deposition), since the temperature in the furnace of the manufacturing apparatus becomes high, it is generally difficult to monitor the state inside the furnace. Therefore, in this process, conventionally, simulations using a simulation apparatus have been utilized to understand the image inside the furnace and promote development.

[0003] Japanese Patent Laid-Open No. 2018-169818

[0004] Keiji Wada et al., "Low-Defect 6-inch SiC Epitaxial Substrate 'EpiEra'", July 2018, SEI Technical Review, No. 193, pp. 53-57 Shunta Harada et al., "Design of SiC High-Quality Crystal Growth Conditions Utilizing Machine Learning", Materia, Vol. 59, No. 3 (2020), pp. 145-152

[0005] Here, in the above simulation, in addition to various prerequisite conditions such as parameters indicating the structure of the furnace and values indicating the physical properties of each part inside the furnace, it is necessary to set a large number of manufacturing conditions. Also, the CVD method is a complex process in which chemical reactions occur at high temperatures, and a dedicated analyst is required to perform the simulation.

[0006] On the other hand, the manufacturing conditions that can be adjusted at the manufacturing site are limited. Also, among the predicted states inside the furnace, the items that the user should confirm in designing the manufacturing conditions are only a part of the items that can be predicted by the simulation using the simulation apparatus.

[0007] Thus, the conventional simulation apparatus has not necessarily been configured to be highly convenient for the user from the perspective of designing manufacturing conditions.

[0008] This disclosure provides a manufacturing condition design support system, a manufacturing condition design support method, and a manufacturing condition design support program that provide suitable support for designing manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, as well as a manufacturing method based on the designed manufacturing conditions.

[0009] A first aspect of this disclosure is a manufacturing condition design support system, which supports the design of manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, comprising: a storage unit for storing a trained model learned using input data related to specific manufacturing conditions and output data of specific items indicating the state inside the furnace, from input and output data when a simulation predicting the state inside the furnace is performed by a simulation device for the process; a trained model reading unit for reading a trained model corresponding to a specific item specified by the user from the storage unit; a trained model execution unit for executing the read trained model and predicting output data of a specific item indicating the state inside the furnace when input data related to specific manufacturing conditions is input by the user; and a display unit for displaying the specified specific item, the input input data related to the specific manufacturing conditions, and the predicted output data.

[0010] A second aspect of this disclosure is a manufacturing condition design support system as described in the first aspect, wherein the input data relating to the specific manufacturing conditions includes data for controlling heaters placed in the furnace and data indicating the flow rate of raw material gas flowing into the furnace.

[0011] A third aspect of this disclosure is a manufacturing condition design support system as described in the second aspect, wherein the input data relating to the specific manufacturing conditions further includes data indicating the rotational speed when rotating a wafer placed in a furnace.

[0012] A fourth aspect of this disclosure is a manufacturing condition design support system as described in the first aspect, wherein the output data for a specific item indicating the state inside the furnace includes at least one of the following: data indicating the distribution of raw material gases inside the furnace, data indicating the distribution of decomposition gases inside the furnace, data indicating the distribution of film thickness growth on the surface of a wafer placed inside the furnace, and data indicating the dopant concentration distribution.

[0013] A fifth aspect of this disclosure is a manufacturing condition design support system as described in the fourth aspect, wherein the output data for a specific item indicating the state inside the furnace further includes at least one of data indicating the temperature distribution inside the furnace and data indicating the flow velocity distribution inside the furnace.

[0014] A sixth aspect of this disclosure is a manufacturing condition design support system according to the first aspect, further comprising a manufacturing condition search unit that searches for input data related to specific manufacturing conditions for the trained model execution unit to predict modified output data when the output data in a region specified by the user among the output data displayed by the display unit is changed, and the display unit displays the searched input data.

[0015] A seventh aspect of this disclosure is a manufacturing condition design support system as described in the sixth aspect, wherein the manufacturing condition search unit comprehensively generates input data included within a predetermined constraint range, and inputs the comprehensively generated input data into the trained model execution unit to search for input data that corresponds to output data similar to the modified output data among the output data predicted by the trained model execution unit.

[0016] An eighth aspect of this disclosure is a manufacturing condition design support method, wherein in a process of manufacturing a SiC epitaxial wafer by the CVD method, the computer of a manufacturing condition design support device that supports the design of manufacturing conditions stores in a storage unit a trained model learned using input data related to specific manufacturing conditions and output data of specific items indicating the state inside the furnace, from input and output data when a simulation predicting the state inside the furnace is performed by a simulation device for the process; reads out a trained model corresponding to a specific item specified by the user from the storage unit; when input data related to specific manufacturing conditions is input by the user, executes the read-out trained model and predicts output data of specific items indicating the state inside the furnace; and displays the specified specific item, the input input data related to the specific manufacturing conditions, and the predicted output data.

[0017] A ninth aspect of this disclosure is a manufacturing condition design support program, which, in a process of manufacturing a SiC epitaxial wafer by the CVD method, causes a computer of a manufacturing condition design support device that assists in designing manufacturing conditions to store a trained model in a storage unit, which is trained using input data related to specific manufacturing conditions and output data of specific items indicating the state inside the furnace, from input and output data of specific items indicating the state inside the furnace, when a simulation predicting the state inside the furnace is performed by a simulation device for the process; reads a trained model corresponding to a specific item specified by the user from the storage unit; when input data related to specific manufacturing conditions is input by the user, executes the read trained model and predicts output data of specific items indicating the state inside the furnace; and displays the specified specific item, the input input data related to the specific manufacturing conditions, and the predicted output data.

[0018] A tenth aspect of this disclosure is a manufacturing method for manufacturing a SiC epitaxial wafer by CVD under manufacturing conditions specified by input data searched by a manufacturing condition search unit of a manufacturing condition design support system described in the sixth or seventh aspect.

[0019] According to this disclosure, a manufacturing condition design support system, a manufacturing condition design support method, and a manufacturing condition design support program are provided to provide suitable support for designing manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, and a manufacturing method based on the designed manufacturing conditions is also provided.

[0020] Figure 1 is the first diagram showing an example of the system configuration of a manufacturing condition design support system. Figure 2A is the first diagram showing an example of the overview of a manufacturing device. Figure 2B is the second diagram showing an example of the overview of a manufacturing device. Figure 3 is a diagram showing an example of the hardware configuration of a simulation device, a learning device, and a manufacturing condition design support device. Figure 4 is a diagram showing a specific example of the processing of the simulation device. Figure 5 is a diagram showing an example of the functional configuration of a learning device. Figure 6 is a diagram showing an example of training data. Figure 7 is a flowchart showing the flow of the training data generation process. Figure 8 is a flowchart showing the flow of the learning process. Figure 9 is the first diagram showing an example of the functional configuration of a manufacturing condition design support device. Figure 10 is the first diagram showing an example of a manufacturing condition design support screen. Figure 11 is the first flowchart showing the flow of the manufacturing condition design support process. Figure 12 is the second diagram showing an example of the functional configuration of a manufacturing condition design support device. Figure 13 is the second diagram showing an example of a manufacturing condition design support screen. Figure 14 is the second flowchart showing the flow of the manufacturing condition design support process. Figure 15 is a second diagram showing an example of the system configuration of a manufacturing condition design support system.

[0021] Each embodiment will be described below with reference to the attached drawings. In this specification and the drawings, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant descriptions will be omitted.

[0022] [First Embodiment] <System Configuration of Manufacturing Condition Design Support System> First, the system configuration of the manufacturing condition design support system according to the first embodiment will be described. Figure 1 is a first diagram showing an example of the system configuration of the manufacturing condition design support system.

[0023] As shown in Figure 1, the manufacturing condition design support system 100 according to the first embodiment includes a simulation device 110, a learning device 120, and a manufacturing condition design support device 130.

[0024] The simulation device 110 reproduces the target manufacturing apparatus 10 and performs simulations under various manufacturing conditions to predict the state inside the furnace. In the first embodiment, the target manufacturing apparatus 10 is an apparatus that manufactures SiC epitaxial wafers from bulk wafers using the CVD method.

[0025] The simulation device 110 reproduces the target manufacturing apparatus 10 by setting various preconditions (details of which will be described later), such as parameters indicating the structure of the furnace and values ​​indicating the physical properties of each part inside the furnace.

[0026] Furthermore, the various manufacturing conditions input to the simulation device 110 include the type of gas, the flow rates of various gases, data for controlling the heater, and the wafer rotation speed. Details of the manufacturing conditions will be described later.

[0027] Furthermore, the furnace conditions predicted by the simulation device 110 include the state of the space inside the furnace, the state of the surface of the wafer placed inside the furnace, and the state of the surface of the furnace components. In addition, the state of the space inside the furnace includes gas concentration distribution, flow velocity distribution, temperature distribution, temperature gradient distribution, and pressure distribution, while the state of the surface of the wafer placed inside the furnace includes the thickness growth distribution of the wafer surface, dopant concentration distribution, and temperature distribution. Details of the predicted furnace conditions will be described later.

[0028] The learning device 120 has a learning program installed on it, and when this program is executed, the learning device 120 functions as a learning data generation unit 121 and a learning unit 122.

[0029] The learning data generation unit 121 extracts input data related to specific manufacturing conditions that can be adjusted at the manufacturing site from among the manufacturing conditions input when the simulation device 110 is running a simulation. The learning data generation unit 121 also extracts output data of specific items that the user should check when designing manufacturing conditions from among the furnace state output when the simulation device 110 is running a simulation. Furthermore, the learning data generation unit 121 generates learning data that includes the extracted input data related to specific manufacturing conditions and the output data of specific items indicating the extracted furnace state, and stores it in the learning data storage unit 123.

[0030] The learning unit 122 uses the learning data stored in the learning data storage unit 123 to perform model training and generate a trained model.

[0031] The manufacturing condition design support device 130 has a manufacturing condition design support program installed, and when this program is executed, the manufacturing condition design support device 130 functions as a display unit 131 and a prediction unit 132.

[0032] The display unit 131 provides the user 140 with a manufacturing condition design support screen. In response to providing the manufacturing condition design support screen, the display unit 131 acquires input data related to the manufacturing conditions (specific manufacturing conditions that can be adjusted at the manufacturing site) entered by the user 140 and notifies the prediction unit 132. In response to the notification of the input data related to the specific manufacturing conditions, the display unit 131 acquires output data for specific items indicating the state inside the furnace, predicted by the prediction unit 132, and displays it to the user 140. The output data for specific items indicating the state inside the furnace, output from the trained model, may be output as a CSV file.

[0033] The prediction unit 132 reads a trained model generated by the learning device 120 and stored in a trained model storage unit 133, which is an example of a storage unit, and executes the trained model by inputting input data related to specific manufacturing conditions notified by the display unit 131. The prediction unit 132 also notifies the display unit 131 of output data for specific items indicating the state inside the furnace, which is output from the trained model after it has been executed.

[0034] Thus, in the manufacturing condition design support system 100 according to the first embodiment, in the process of manufacturing SiC epitaxial wafers by the CVD method, a trained model is generated that specializes in inputting and outputting data related to: - input data related to specific manufacturing conditions that can be adjusted at the manufacturing site, and - output data of specific items that the user should check when designing the manufacturing conditions from among the conditions inside the furnace.

[0035] As a result, according to the manufacturing condition design support system 100 of the first embodiment, by using the trained model, it is possible to predict output data for specific items indicating the state inside the furnace that the user should check when designing the manufacturing conditions, simply by inputting input data related to specific manufacturing conditions, without having to set a large number of preconditions such as parameters indicating the structure of the furnace or values ​​indicating the physical properties of each part inside the furnace, and without having to input a wide range of manufacturing conditions.

[0036] <Overview of Manufacturing Equipment> Next, an overview of the manufacturing equipment 10 (equipment for manufacturing SiC epitaxial wafers by the CVD method) reproduced by the simulation device 110 will be described. Figure 2A is the first figure showing an example of the overview of the manufacturing equipment, and shows the structure of the XZ plane when cut at a predetermined Y coordinate in the XYZ space.

[0037] As shown in FIG. 2A, the manufacturing apparatus 10 includes a gas introduction unit 210 that blows a raw material gas, a dopant gas, etc. downward. Further, the manufacturing apparatus 10 includes a heater 220 on a furnace wall that forms a space 240 inside the furnace. The heater 220 is, for example, a resistance heating heater, and raises the temperature of the space 240 inside the furnace to about 1600°C. The heater 220 may be composed of a plurality of heater groups and may be individually temperature-controlled.

[0038] By raising the temperature of the space 240 inside the furnace in this way, the raw material gas (reference numeral 250) among the various gases blown from the gas introduction unit 210 is decomposed, and a decomposition gas (reference numeral 260) is generated.

[0039] Further, the manufacturing apparatus 10 includes a heater 230 on a mounting table on which the wafer W is placed. The heater 230 is, for example, a resistance heating heater, and raises the temperature of the wafer W to about 1600°C.

[0040] With such a configuration, the decomposition gas generated in the space 240 inside the furnace and the dopant gas blown from the gas introduction unit 210 are deposited on the surface of the wafer W to form a SiC epitaxial layer, thereby manufacturing a SiC epitaxial wafer.

[0041] Note that the manufacturing apparatus 10 is not limited to the mode shown in FIG. 2A, and as shown in FIG. 2B, it may be a mode in which a raw material gas, a dopant gas, etc. are blown in the horizontal direction (that is, the raw material gas, the dopant gas, etc. may be supplied from a direction parallel to the wafer surface). FIG. 2B is a second diagram showing an example of the outline of the manufacturing apparatus, and shows a manufacturing apparatus called a so-called horizontal furnace. Among the structures in FIG. 2B, for the structures corresponding to the structures in FIG. 2A, the same reference numerals are used and the description is omitted.

[0042] Further, in the manufacturing apparatus 10 shown in FIGS. 2A and 2B, the case where one wafer W is placed on the mounting table is shown, but a plurality of wafers may be placed on the mounting table. Alternatively, a furnace may be configured by combining a plurality of spaces on which wafers are placed.

[0043] <Hardware Configuration of Simulation Device, Learning Device, and Manufacturing Condition Design Support Device> Next, the hardware configurations of the simulation device 110, learning device 120, and manufacturing condition design support device 130 that constitute the manufacturing condition design support system 100 will be described. Since the simulation device 110, learning device 120, and manufacturing condition design support device 130 generally have similar hardware configurations, they will be collectively described here using FIG. 3.

[0044] FIG. 3 is a diagram showing an example of the hardware configuration of the simulation device, learning device, and manufacturing condition design support device. As shown in FIG. 3, the simulation device 110, learning device 120, and manufacturing condition design support device 130 each have a processor 301, a memory 302, an auxiliary storage device 303, an I / F (Interface) device 304, a communication device 305, and a drive device 306. The hardware of each of the simulation device 110, learning device 120, and manufacturing condition design support device 130 is mutually connected via a bus 307.

[0045] The processor 301 has various arithmetic devices such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). The processor 301 reads out and executes various programs (for example, a learning program, a manufacturing condition design support program, etc.) on the memory 302.

[0046] The memory 302 has main storage devices such as a ROM (Read Only Memory) and a RAM (Random Access Memory). The processor 301 and the memory 302 form a so-called computer, and the computer realizes various functions by the processor 301 executing various programs read out on the memory 302.

[0047] The auxiliary storage device 303 stores various programs and various data used when the various programs are executed by the processor 301. For example, the learning data storage unit 123 and the learned model storage unit 133 are realized in the auxiliary storage device 303.

[0048] The I / F device 304 is a connection device for connecting an operating device 311 and a display device 312, which are examples of user interface devices. The communication device 305 is a communication device for communicating with an external device (not shown) via a network (not shown).

[0049] The drive device 306 is a device for setting the recording medium 313. The recording medium 313 here includes media that record information optically, electrically, or magnetically, such as CD-ROMs, flexible disks, and magneto-optical disks. The recording medium 313 may also include semiconductor memory that records information electrically, such as ROMs and flash memory.

[0050] The various programs to be installed on the auxiliary storage device 303 are installed, for example, when the distributed recording medium 313 is set in the drive device 306 and the various programs recorded on the recording medium 313 are read by the drive device 306. Alternatively, the various programs to be installed on the auxiliary storage device 303 may be installed by downloading them from the network via the communication device 305.

[0051] <Specific Examples of Simulation Device Processing> Next, specific examples of the processing of the simulation device 110 will be explained. Figure 4 is a diagram showing a specific example of the processing of the simulation device.

[0052] As described above, various preconditions are set in the simulation device 110 in order to reproduce the target manufacturing apparatus 10. As shown in Figure 4, the various preconditions set in the simulation device 110 include the CVD furnace structure, material properties, etc. The CVD furnace structure includes the gas introduction section structure, the heater position, etc. The material properties include values ​​that show the physical properties (thermal conductivity, specific heat, emissivity, etc.) of the materials used at each position in the furnace.

[0053] Note that the example in Figure 4 only shows some of the various preconditions set for the simulation device 110, and many other preconditions will be set for the simulation device 110 in addition to the example in Figure 4.

[0054] Furthermore, as mentioned above, various manufacturing conditions are input to the simulation device 110 during the simulation. As shown in Figure 4, the manufacturing conditions input to the simulation device 110 include heater power value, gas type, various flow rates, SiC wafer rotation speed, etc.

[0055] Furthermore, the gas types include Si-based gases, C-based gases, dopant gases, and carrier gases, with multiple types of each being input. The various flow rates include the flow rates of each source gas, the carrier gas, and the dopant gas. The SiC wafer rotation speed refers to the rotation speed of the mounting platform on which the wafer W is placed during the manufacturing process.

[0056] Note that the example in Figure 4 only shows a portion of the manufacturing conditions input to the simulation device 110; the simulation device 110 will accept a wide range of manufacturing conditions in addition to those shown in Figure 4.

[0057] Furthermore, as described above, the simulation device 110 predicts the state inside the furnace through simulation. As shown in Figure 4, the state inside the furnace predicted by the simulation device 110 includes the state of the space inside the furnace and the state of the surface of the wafer placed inside the furnace.

[0058] Furthermore, the conditions of the space inside the furnace include the flow velocity distribution, temperature distribution, temperature gradient distribution, pressure distribution, gas concentration (gas partial pressure) distribution, and C / Si ratio (ratio of the number of C atoms to the number of Si atoms) distribution.

[0059] Furthermore, as shown in Figure 4, the surface condition of the wafer placed in the furnace includes the film thickness growth distribution, dopant concentration distribution, temperature distribution, C / Si ratio distribution, gas concentration distribution, heat flux distribution, etc.

[0060] <Functional Configuration of the Learning Device> Next, the details of the functional configuration of the learning device 120 will be explained. Figure 5 is a diagram showing an example of the functional configuration of the learning device.

[0061] As described above, the learning device 120 functions as a learning data generation unit 121 and a learning unit 122. Of these, the learning data generation unit 121 further includes an input data selection unit 510, a first output data selection unit 520_1 to the nth output data selection unit 520_n, and a combination unit 530 (where n is an integer of 2 or more). The learning unit 122 further includes a learning data reading unit 540, a first model 550_1 to the mth model 550_m, and comparison and modification units 560_1 to 560_m (where m is an integer of 2 or more, provided n ≥ m).

[0062] The input data selection unit 510 selects and extracts input data related to specific manufacturing conditions that can be adjusted at the manufacturing site, from among the input data related to manufacturing conditions entered when the simulation device 110 is running a simulation.

[0063] The first to the nth output data selection units 520_1 to nth output data selection units 520_n each select and extract output data for specific items that the user should check when designing manufacturing conditions, from among the furnace conditions output during the simulation by the simulation device 110.

[0064] The combination unit 530 combines: - Input data related to specific manufacturing conditions extracted by the input data selection unit 510, and - Output data of specific items indicating the state inside the furnace, extracted by the first output data selection unit 520_1 to the nth output data selection unit 520_n. The combination unit 530 also stores the combined input data and output data as learning data in the learning data storage unit 123. In the first embodiment, m combinations are stored in the learning data storage unit 123 as learning data.

[0065] The training data reading unit 540 reads m training data from the training data storage unit 123. The training data reading unit 540 also inputs the input data contained in the m training data that it reads into one of the first model 550_1 to the mth model 550_m. The training data reading unit 540 also inputs the output data contained in the m training data that it reads into one of the comparison and modification units 560_1 to 560_m.

[0066] The first model 550_1 to the mth model 550_m output output data when input data is received by the training data reading unit 540.

[0067] The comparison and modification units 560_1 to 560_m compare the output data output from the first model 550_1 to the mth model 550_m with the output data input by the learning data reading unit 540. The comparison and modification units 560_1 to 560_m also update the model parameters of the first model 550_1 to the mth model 550_m according to the comparison results.

[0068] As a result, the learning unit 122 generates the first trained model to the mth trained model.

[0069] The first to the mth trained model are assumed to be composed of neural networks with the following parameters, for example: • Hidden layers: 128 nodes x 4 layers • Activation function: tanh • Number of training data: 50 • Optimization algorithm: Adam • Number of epochs: 50 However, the above parameters are just an example, and the first to the mth trained model are not limited to neural networks composed of the above parameters. For example, machine learning methods such as linear, generalized linear (Lasso, Ridge, Elasticnet, Logistic), partial least squares, kernel ridge, Gaussian process, k-nearest neighbors, decision tree, random forest, AdaBoost, bagging, gradient boosting, and support vector machine may be used.

[0070] <Specific Examples of Training Data> Next, we will explain specific examples of training data generated by the training data generation unit 121. Figure 6 shows an example of training data.

[0071] In Figure 6, the first training data 610 to the fourth training data 640 show that the input data extracted includes data for controlling multiple heaters, the flow rates of multiple types of raw material gases, and data indicating the rotation speed when rotating the wafer placed in the furnace.

[0072] Furthermore, Figure 6 shows that the first training data 610 extracted data indicating the distribution of raw material gases in the furnace and data indicating the distribution of decomposition gases in the furnace as output data, and combined them with the input data.

[0073] Furthermore, Figure 6 shows that the second training data 620 extracted data indicating the temperature distribution inside the furnace as output data and combined it with the input data.

[0074] Furthermore, Figure 6 shows that the third training data 630 extracted data representing the chemical reaction rate distribution inside the furnace was combined with the input data as output data.

[0075] Furthermore, Figure 6 shows that the fourth training data 640, as output data, extracts data showing the film thickness growth distribution on the surface of a wafer placed in the furnace and combines it with the input data.

[0076] <Flow of Training Data Generation Process> Next, the flow of training data generation process by the training device 120 will be explained. Figure 7 is a flowchart showing the flow of training data generation process.

[0077] In step S701, the learning device 120 acquires input / output data (input data related to manufacturing conditions and output data indicating the state inside the furnace) from the simulation device 110.

[0078] In step S702, the learning device 120 selects and extracts input data related to specific manufacturing conditions from the input data related to manufacturing conditions acquired from the simulation device 110.

[0079] In step S703, the learning device 120 inputs "1" to counter i, which counts the number of output data.

[0080] In step S704, the learning device 120 selects and extracts the output data for a specific i-th item indicating the state inside the furnace from the output data indicating the state inside the furnace acquired from the simulation device 110.

[0081] In step S705, the learning device 120 combines the extracted input data related to specific manufacturing conditions with the extracted output data for a specific i-th item indicating the state inside the furnace to generate the i-th learning data, which is then stored in the learning data storage unit 123.

[0082] In step S706, the learning device 120 determines whether or not to generate other learning data. If it determines to generate other learning data (i.e., if the answer in step S706 is YES), the process proceeds to step S707.

[0083] In step S707, the learning device 120 increments the counter i and returns to step S704.

[0084] On the other hand, if it is determined in step S706 that no other training data will be generated (i.e., the answer in step S706 is NO), the training data generation process is terminated.

[0085] <Learning Process Flow> Next, the learning process flow by the learning device 120 will be explained. Figure 8 is a flowchart of the learning process flow.

[0086] In step S801, the learning device 120 inputs "1" to counter i, which counts the number of learning data.

[0087] In step S802, the learning device 120 reads the i-th learning data from the learning data storage unit 123.

[0088] In step S803, the learning device 120 performs a learning process on the i-th model using the i-th training data.

[0089] In step S804, the learning device 120 generates the i-th trained model and stores it in the trained model storage unit 133 of the manufacturing condition design support device 130.

[0090] In step S805, the learning device 120 determines whether or not to generate another trained model. If it determines in step S805 to generate another trained model (i.e., if the answer is YES in step S805), the device proceeds to step S806. This makes it possible to generate different trained models depending on the training data (i.e., depending on the item).

[0091] In step S806, the learning device 120 increments the counter i and returns to step S802.

[0092] On the other hand, if it is determined in step S805 that no other trained models will be generated (i.e., the answer in step S805 is NO), the training process is terminated.

[0093] <Functional Configuration of the Manufacturing Condition Design Support Device> Next, the details of the functional configuration of the manufacturing condition design support device 130 will be explained. Figure 9 is the first diagram showing an example of the functional configuration of the manufacturing condition design support device.

[0094] As described above, the manufacturing condition design support device 130 functions as a display unit 131 and a prediction unit 132. Of these, the display unit 131 further includes an output item setting unit 910, a manufacturing condition input unit 911, and a prediction result output unit 912. The prediction unit 132 further includes a trained model reading unit 920 and a trained model execution unit 921.

[0095] The output item setting unit 910 sets specific items of the furnace state that the user 140 should check when designing the manufacturing conditions in the trained model readout unit 920. By default, the output item setting unit 910 is configured so that all specific items that the user 140 should check when designing the manufacturing conditions are set in the trained model readout unit 920. However, the output item setting unit 910 may be configured to allow the user 140 to select or discard specific items to set.

[0096] The manufacturing condition input unit 911 accepts input data related to specific manufacturing conditions when the user 140 inputs data in response to the display unit 131 displaying the manufacturing condition design support screen. The manufacturing condition input unit 911 also notifies the trained model execution unit 921 of the input data related to the specific manufacturing conditions that it has received.

[0097] The prediction result output unit 912 acquires output data for specific items indicating the state inside the furnace, which has been predicted by the trained model execution unit 921, in response to the trained model execution unit 921 notifying the trained model execution unit 921 of input data related to specific manufacturing conditions. The prediction result output unit 912 also displays the acquired output data for specific items indicating the state inside the furnace to the user 140 via the manufacturing condition design support screen.

[0098] The trained model reading unit 920 reads the trained model corresponding to a specific item set by the output item setting unit 910 from the trained model storage unit 133. As mentioned above, the trained model storage unit 133 stores different trained models depending on the item.

[0099] Furthermore, the trained model retrieval unit 920 notifies the trained model execution unit 921 of the retrieved trained model.

[0100] The trained model execution unit 921 inputs input data related to specific manufacturing conditions notified by the manufacturing condition input unit 911 to the trained model notified by the trained model readout unit 920. As a result, the trained model execution unit 921 executes the trained model and predicts output data for specific items indicating the state inside the furnace.

[0101] Furthermore, the trained model execution unit 921 notifies the prediction result output unit 912 of the output data predicted by the trained model.

[0102] Thus, according to the manufacturing condition design support system 100 of the first embodiment, by simply inputting input data related to specific manufacturing conditions that can be adjusted at the manufacturing site, and by predicting output data for specific items that the user should check when designing manufacturing conditions among the conditions inside the furnace, the convenience of the user when designing manufacturing conditions can be improved. In other words, according to the first embodiment, suitable support can be provided for designing manufacturing conditions in the process of manufacturing SiC epitaxial wafers by the CVD method.

[0103] <Explanation of the Manufacturing Condition Design Support Screen> Next, the manufacturing condition design support screen displayed by the display unit 131 will be explained. Figure 10 is the first diagram showing an example of the manufacturing condition design support screen.

[0104] As shown in Figure 10, the manufacturing condition design support screen 1000 includes a manufacturing condition input area 1010 and a furnace state display area 1020.

[0105] Furthermore, as shown in Figure 10, the manufacturing condition input area 1010 is provided with input fields for inputting data related to specific manufacturing conditions. The example in Figure 10 shows input fields for inputting data to control the heater and input fields for inputting data indicating the flow rate of the raw material gas.

[0106] Furthermore, as shown in Figure 10, the furnace state display area 1020 is provided with display fields for displaying output data for specific items indicating the state inside the furnace. The example in Figure 10 shows the display fields for displaying data indicating the raw material gas distribution inside the furnace, data indicating the decomposition gas distribution inside the furnace, and data indicating the temperature distribution inside the furnace, displayed in the upper section.

[0107] Specifically, the image displays data from a cross-section of the furnace inside the manufacturing apparatus 10, viewed from the side. The horizontal axis (x-axis) of the image represents the width of the furnace (the X-axis direction in Figure 2), and the vertical axis (z-axis) represents the height of the furnace (the Z-axis direction in Figure 2). The pixel values ​​(R, G, B) at each coordinate (x, z) in the image indicate the color corresponding to the data value of that coordinate. However, in the example in Figure 10, for the sake of explanation, the color corresponding to the data value of each coordinate is represented by the dot density.

[0108] Furthermore, the example in Figure 10 shows a display area at the bottom showing data indicating the film thickness growth distribution on the wafer surface, data indicating the temperature distribution on the wafer surface, and data indicating the dopant concentration distribution on the wafer surface.

[0109] Specifically, the image displays data of the wafer surface as viewed from above inside the furnace of the manufacturing apparatus 10. The horizontal axis (x-axis) of the image represents the width direction inside the furnace (X-axis direction in Figure 2), and the vertical axis (y-axis) represents the depth direction inside the furnace (Y-axis direction in Figure 2). In addition, the pixel values ​​(R, G, B) of each coordinate (x, y) in the image indicate the color corresponding to the data value of each coordinate. However, in the example in Figure 10, for the sake of explanation, the color corresponding to the data value of each coordinate is represented by the dot density.

[0110] <Flow of Manufacturing Condition Design Support Process> Next, the flow of the manufacturing condition design support process by the manufacturing condition design support device 130 will be explained. Figure 11 is the first flowchart showing the flow of the manufacturing condition design support process.

[0111] In step S1101, the manufacturing condition design support device 130 sets specific items that the user should check when designing the manufacturing conditions.

[0112] In step S1102, the manufacturing condition design support device 130 reads out a learned model corresponding to a specific item that has been set.

[0113] In step S1103, the manufacturing condition design support device 130 receives input data related to specific manufacturing conditions.

[0114] In step S1104, the manufacturing condition design support device 130 executes the trained model read in step S1102 by inputting the input data related to the specific manufacturing conditions received in step S1103. As a result, the manufacturing condition design support device 130 predicts output data for specific items that indicate the state inside the furnace.

[0115] In step S1105, the manufacturing condition design support device 130 displays output data for specific items that indicate the predicted state inside the furnace.

[0116] In step S1106, the manufacturing condition design support device 130 determines whether or not there has been an instruction to change a specific item. If it is determined in step S1106 that there has been an instruction to change a specific item (i.e., if the answer is YES in step S1106), the process returns to step S1101.

[0117] On the other hand, if it is determined in step S1106 that there was no instruction to change a specific item (i.e., the answer is NO in step S1106), the process proceeds to step S1107.

[0118] In step S1107, the manufacturing condition design support device 130 determines whether or not there has been an instruction to change the input data related to specific manufacturing conditions. If it is determined in step S1107 that there has been an instruction to change the input data related to specific manufacturing conditions (i.e., if the answer is YES in step S1107), the device returns to step S1103.

[0119] On the other hand, if it is determined in step S1107 that there were no instructions to change the input data related to specific manufacturing conditions (i.e., the answer in step S1107 is NO), the manufacturing condition design support process is terminated.

[0120] <Summary> As is clear from the above explanation, the manufacturing condition design support system 100 according to the first embodiment is a system that supports the design of manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, and: - Acquires input and output data when a simulation predicting the state inside the furnace is performed using a simulation device that reproduces the process. - Extracts input data related to specific manufacturing conditions and output data for specific items indicating the state inside the furnace from the acquired input and output data, and generates training data including the extracted input data and output data. - Generates a trained model by performing a training process using the generated training data and stores it in the trained model storage unit. - Reads out the trained model corresponding to a specific item specified by the user from the trained model storage unit. - When input data related to specific manufacturing conditions is input by the user, reads out the retrieved trained model and executes it to predict output data for specific items indicating the state inside the furnace. - Displays the specified specific item, the input data related to the specific manufacturing conditions that was input, and the predicted output data to the user.

[0121] Thus, according to the manufacturing condition design support system 100 of the first embodiment, - by simply inputting input data related to specific manufacturing conditions that can be adjusted at the manufacturing site, - it is possible to predict output data for specific items that the user should check when designing manufacturing conditions among the conditions inside the furnace, thereby improving the convenience of the user when designing manufacturing conditions.

[0122] In other words, according to the first embodiment, a manufacturing condition design support system 100 can be provided that provides suitable support for designing manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method.

[0123] [Second Embodiment] In the first embodiment described above, a case was described in which, when input data related to specific manufacturing conditions is entered on the manufacturing condition design support screen, output data for a specific item indicating the state inside the furnace is displayed.

[0124] In contrast, in the second embodiment, when output data is displayed on the manufacturing condition design support screen, and the user further inputs a request to change the output data, the system searches for input data related to the specific manufacturing conditions that will result in the changed output data. Thus, according to the second embodiment, it becomes possible to propose to the user input data related to the specific manufacturing conditions necessary to realize the output data requested by the user. The second embodiment will now be described, focusing on the differences from the first embodiment.

[0125] <Functional Configuration of the Manufacturing Condition Design Support Device> First, the functional configuration of the manufacturing condition design support device according to the second embodiment will be described. Figure 12 is a second diagram showing an example of the functional configuration of the manufacturing condition design support device. As shown in Figure 12, the manufacturing condition design support device 1200 according to the second embodiment functions as a display unit 1210 and a prediction unit 1220. Of these, the display unit 1210 further includes an output item setting unit 910, a manufacturing condition input unit 911, a prediction result output unit 912, a change request acquisition unit 1211, and a proposed data output unit 1212. The prediction unit 1220 further includes a trained model reading unit 920, a trained model execution unit 921, a trained model execution unit 1221, an error calculation unit 1222, and a manufacturing condition search unit 1223.

[0126] Of the various parts of the display unit 1210, the output item setting unit 910 to the prediction result output unit 912 have already been explained using Figure 9 in the first embodiment described above, so their explanation will be omitted here.

[0127] The change request acquisition unit 1211 accepts a change request from the user 140 when the user 140 inputs a change request in response to the prediction result output unit 912 displaying output data for specific items indicating the state inside the furnace to the user 140 via the manufacturing condition design support screen. A change request refers to a request to change a part of the output data for specific items indicating the state inside the furnace. The change request acquisition unit 1211 notifies the error calculation unit 1222 of the modified output data that has been changed by the change request.

[0128] The proposed data output unit 1212, upon receiving notification from the change request acquisition unit 1211 that the modified output data has been sent to the error calculation unit 1222, acquires input data (referred to as proposed data) related to the proposed specific manufacturing conditions from the manufacturing condition search unit 1223. The proposed data output unit 1212 also displays the acquired proposed data to the user 140 via the manufacturing condition design support screen.

[0129] The trained model execution unit 1221 is similar to the trained model execution unit 921, and by inputting input data related to specific manufacturing conditions notified by the manufacturing condition search unit 1223, it executes the trained model and predicts output data for specific items that indicate the state inside the furnace.

[0130] The error calculation unit 1222 calculates the error between the output data predicted by the trained model execution unit 1221 and the modified output data notified by the change request acquisition unit 1211, and notifies the manufacturing condition search unit 1223.

[0131] The manufacturing condition search unit 1223 comprehensively generates input data related to a specific manufacturing condition within a predetermined constraint range and notifies the trained model execution unit 1221. The manufacturing condition search unit 1223 also obtains the error corresponding to each of the comprehensively generated input data related to the specific manufacturing condition from the error calculation unit 1222 and searches for input data related to a specific manufacturing condition with a small error (i.e., a specific manufacturing condition with similar output data).

[0132] Furthermore, the manufacturing condition search unit 1223 notifies the proposed data output unit 1212 of the input data related to the specific manufacturing conditions it has searched for as proposed data.

[0133] <Explanation of the Manufacturing Condition Design Support Screen> Next, we will explain the manufacturing condition design support screen (a screen for receiving change requests and displaying proposed data) displayed by the display unit 1210. Figure 13 is a second diagram showing an example of the manufacturing condition design support screen.

[0134] As shown in Figure 13, the manufacturing condition design support screen 1300 includes a change request input area 1310 and a proposed data display area 1320.

[0135] Furthermore, as shown in Figure 13, the change request input area 1310 is provided with input fields for changing the output data of specific items indicating the state inside the furnace. The example in Figure 13 shows input fields for changing data indicating the raw material gas distribution inside the furnace, data indicating the decomposition gas distribution inside the furnace, data indicating the temperature distribution inside the furnace, and data indicating the chemical reaction rate distribution inside the furnace. The example in Figure 13 also shows input fields for changing data indicating the film thickness growth distribution on the wafer surface, data indicating the temperature distribution on the wafer surface, and data indicating the dopant concentration distribution on the wafer surface.

[0136] User 140 modifies data by specifying a desired range in one of the input fields provided in the change request input area 1310. Specifically, when User 140 specifies a desired range in an input field, the coordinate values ​​and data values ​​of each pixel within that range are displayed, and User 140 modifies each data value. When a data value is changed, an image of the pixel values ​​corresponding to the changed data value is displayed in the input field. The method of changing data values ​​is not limited; they may be entered numerically or adjusted using a slider bar.

[0137] In the example shown in Figure 13, the user 140 specified a desired range 1311 in the data showing the temperature distribution on the surface of the wafer, and the coordinate values ​​of each pixel within that range 1311 (x １ , y １ ) ~ (x ２０ , y ２０ )), ・Data value of each pixel (t １ ), is displayed, and user 140 sets the data value of each pixel to t １ →t ２ This shows how it was changed.

[0138] Furthermore, as shown in Figure 13, when a change request is entered into the input field in the change request input area 1310, the proposed data is displayed in the proposed data display area 1320. The example in Figure 13 shows the proposed data displaying data for controlling the heater and the flow rate of the raw gas. The proposed data displayed in the proposed data display area 1320 is displayed, for example, in order of smallest error with the modified output data (in order of similarity to the modified output data).

[0139] <Flow of Manufacturing Condition Design Support Process> Next, the flow of the manufacturing condition design support process by the manufacturing condition design support device 1200 will be explained. Figure 14 is a second flowchart showing the flow of the manufacturing condition design support process. Note that in the second flowchart shown in Figure 14, each step from step S1101 to step S1107 is the same as each step from step S1101 to step S1107 in the first flowchart shown in Figure 11, so the explanation will be omitted here.

[0140] In step S1108, the manufacturing condition design support device 130 determines whether or not a change request has been entered. If it is determined in step S1108 that no change request has been entered (i.e., the answer is NO in step S1108), the manufacturing condition design support process is terminated.

[0141] On the other hand, if it is determined in step S1107 that a change request has been entered (i.e., if the answer is YES in step S1108), the process proceeds to step S1401.

[0142] In step S1401, the manufacturing condition design support device 1200 receives a change request input from the user 140 via the manufacturing condition design support screen 1300.

[0143] In step S1402, the manufacturing condition design support device 1200 comprehensively generates input data related to specific manufacturing conditions within a predetermined constraint range. The manufacturing condition design support device 1200 also searches for input data related to specific manufacturing conditions from among the output data predicted based on each input data, which has a small error with the modified output data changed by the change request.

[0144] In step S1403, the manufacturing condition design support device 1200 displays the input data related to specific manufacturing conditions with small errors, which were searched in step S1402, as proposed data to the user 140 via the manufacturing condition design support screen 1300.

[0145] <Summary> As is clear from the above explanation, the manufacturing condition design support system 100 according to the second embodiment: - When output data within a range specified by the user is changed among the output data displayed on the manufacturing condition design support screen, it searches for input data related to specific manufacturing conditions in order to predict output data with a small error with the changed output data. - It displays the searched input data related to specific manufacturing conditions as proposed data on the manufacturing condition design support screen.

[0146] Thus, according to the manufacturing condition design support system 100 of the second embodiment, it is possible to search for input data related to specific manufacturing conditions in order to make the output data of specific items indicating the state inside the furnace into desired output data, thereby improving the convenience of the user when designing manufacturing conditions.

[0147] In other words, according to the second embodiment, a manufacturing condition design support system 100 can be provided that provides suitable support for designing manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method.

[0148] [Third Embodiment] In the second embodiment described above, the method of using the proposed data output by the manufacturing condition design support device 1200 was not mentioned, but the proposed data output by the manufacturing condition design support device 1200 may be reflected in the manufacturing device 10. The third embodiment will be described below, focusing on the differences from the above embodiments.

[0149] <System Configuration of the Manufacturing Condition Design Support System> First, the system configuration of the manufacturing condition design support system according to the third embodiment will be described. Figure 15 is a second diagram showing an example of the system configuration of the manufacturing condition design support system.

[0150] As shown in Figure 15, the manufacturing condition design support system 1500 according to the third embodiment includes a manufacturing apparatus 10, an evaluation apparatus 20, and a manufacturing condition design support apparatus 1200.

[0151] Of these, the manufacturing apparatus 10 and the manufacturing condition design support apparatus 1200 have already been described in the first or second embodiment described above, so their description will be omitted here.

[0152] The evaluation device 20 is an evaluation device for evaluating SiC epitaxial wafers manufactured by the manufacturing device 10. The evaluation results obtained by the evaluation device 20 are notified to the user 140.

[0153] Based on the evaluation results notified by the evaluation device 20, user 140 considers how to change the state inside the furnace and inputs a change request to the manufacturing condition design support device 1200 via the manufacturing condition design support screen 1300. It should be assumed that, when inputting the change request, the manufacturing condition design support screen 1300 displays output data for specific items indicating the state inside the furnace, predicted based on the input data related to the current manufacturing conditions set in the manufacturing device 10.

[0154] As a result, the proposed data is displayed on the manufacturing condition design support screen 1300 of the manufacturing condition design support device 1200, and the user 140 applies the displayed proposed data to specific manufacturing conditions in the manufacturing device 10. As a result, the manufacturing device 10 can manufacture SiC epitaxial wafers by the CVD method under the manufacturing conditions to which the proposed data is applied. Note that the process of applying the proposed data to specific manufacturing conditions in the manufacturing device 10 may be performed automatically by the manufacturing condition design support device 1200.

[0155] <Summary> As is clear from the above explanation, the manufacturing condition design support system 1500 according to the third embodiment includes: - A manufacturing apparatus 10, an evaluation apparatus 20, and a manufacturing condition design support device 1200. - The manufacturing condition design support device 1200 receives change requests in accordance with the evaluation results from the evaluation apparatus 20, and the proposed data output by the manufacturing condition design support device 1200 is reflected in the specific manufacturing conditions of the manufacturing apparatus 10.

[0156] Thus, according to the manufacturing condition design support system 1500 of the third embodiment, it is possible to design manufacturing conditions in accordance with events occurring at the manufacturing site, thereby improving user convenience when designing manufacturing conditions.

[0157] In other words, according to the third embodiment, a manufacturing condition design support system 1500 is provided that provides suitable support for designing manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, and a manufacturing method based on the designed manufacturing conditions is also provided.

[0158] [Other Embodiments] In the first embodiment described above, when displaying output data for specific items indicating the state inside the furnace on the manufacturing condition design support screen 1000, an image of a color (predetermined color) corresponding to the data value of each coordinate was displayed. However, the relationship between the data value and the color (pixel value) may be configured so that the user 140 can change it as appropriate.

[0159] Specifically, the system may be configured to allow setting a range of data values ​​for each item, and then calculate the pixel value of each coordinate by color-converting the data value of each coordinate according to that range.

[0160] Furthermore, in the first embodiment described above, when displaying data on the surface of the wafer among the output data of specific items indicating the state inside the furnace on the manufacturing condition design support screen 1000, the case in which data is displayed as if the inside of the furnace of the manufacturing apparatus 10 were viewed from above was described. However, the display method when displaying data on the surface of the wafer is not limited to this, and for example, the system may be configured to display data from one cross-section when the wafer is viewed from the side.

[0161] Furthermore, in the first embodiment described above, the case in which the manufacturing condition design support system 100 is configured with the simulation device 110, the learning device 120, and the manufacturing condition design support device 130 as separate devices was explained. However, in the manufacturing condition design support system 100, the simulation device 110, the learning device 120, and the manufacturing condition design support device 130 may be configured as an integrated device. Alternatively, any two of the devices may be configured as an integrated device.

[0162] Furthermore, in the first embodiment described above, the functional configurations of the simulation device 110, the learning device 120, and the manufacturing condition design support device 130 are merely examples, and some of the functions of any one of the devices may be implemented in the other devices.

[0163] Furthermore, in the third embodiment described above, when the user 140 inputs a change request, it was explained that the user 140 has already entered input data related to the current manufacturing conditions set in the manufacturing apparatus 10 into the manufacturing condition design support screen 1300. However, the input data related to the current manufacturing conditions, which is entered into the manufacturing condition design support screen 1300, may be configured to be entered automatically through communication with the manufacturing apparatus 10. As a result, the manufacturing condition design support screen 1300 will always display output data for specific items that indicate the current state inside the furnace of the manufacturing apparatus 10.

[0164] Furthermore, in the third embodiment described above, the user 140 considers how to change the state inside the furnace based on the evaluation results notified by the evaluation device 20, and inputs a change request to the manufacturing condition design support device 1200 via the manufacturing condition design support screen 1300. However, it is also possible to configure the system to automatically make appropriate change requests based on the evaluation results.

[0165] It should be noted that the present invention is not limited to the configurations shown in the above embodiments, including combinations with other elements. These aspects can be modified without departing from the spirit of the present invention and can be appropriately determined according to their application.

[0166] This application claims priority based on Japanese Patent Application No. 2024-168472, filed on 27 September 2024, which is incorporated herein by reference to the entire contents of the said Japanese Patent Application.

[0167] 100: Manufacturing Condition Design Support System 110: Simulation Device 120: Learning Device 121: Learning Data Generation Unit 122: Learning Unit 130: Manufacturing Condition Design Support Device 131: Display Unit 132: Prediction Unit 910: Output Item Setting Unit 911: Manufacturing Condition Input Unit 912: Prediction Result Output Unit 920: Learned Model Retrieval Unit 921: Learned Model Execution Unit 1000: Manufacturing Condition Design Support Screen 1010: Manufacturing Condition Input Area 1020: Furnace Condition Display Area 1200: Manufacturing Condition Design Support Device 1210: Display Unit 1211: Change Request Acquisition Unit 1212: Proposed Data Output Unit 1220: Prediction Unit 1221: Learned Model Execution Unit 1222: Error Calculation Unit 1223: Manufacturing Condition Search Unit 1300: Manufacturing condition design support screen 1310: Change request input area 1320: Proposed data display area 1500: Manufacturing condition design support system

Claims

1. A manufacturing condition design support system for assisting in the design of manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, comprising: a storage unit for storing a trained model learned using input data related to specific manufacturing conditions and output data of specific items indicating the state inside the furnace, from input and output data when a simulation predicting the state inside the furnace is performed by a simulation device for the process; a trained model reading unit for reading a trained model corresponding to a specific item specified by the user from the storage unit; a trained model execution unit for executing the read trained model and predicting output data of a specific item indicating the state inside the furnace when input data related to specific manufacturing conditions is input by the user; and a display unit for displaying the specified specific item, the input input data related to the specific manufacturing conditions, and the predicted output data.

2. The manufacturing condition design support system according to claim 1, wherein the input data relating to the specific manufacturing conditions includes data for controlling heaters placed in the furnace and data indicating the flow rate of raw material gas flowing into the furnace.

3. The manufacturing condition design support system according to claim 2, wherein the input data relating to the specific manufacturing conditions further includes data indicating the rotation speed when rotating the wafer placed in the furnace.

4. The manufacturing condition design support system according to any one of claims 1 to 3, wherein the output data for specific items indicating the state inside the furnace includes at least one of the following: data indicating the distribution of raw material gases inside the furnace, data indicating the distribution of decomposition gases inside the furnace, data indicating the distribution of film thickness growth on the surface of a wafer placed inside the furnace, and data indicating the dopant concentration distribution.

5. The manufacturing condition design support system according to claim 4, wherein the output data of a specific item indicating the state inside the furnace further includes at least one of data indicating the temperature distribution inside the furnace and data indicating the flow velocity distribution inside the furnace.

6. A manufacturing condition design support system according to any one of claims 1 to 5, further comprising a manufacturing condition search unit that searches for input data related to specific manufacturing conditions for the trained model execution unit to predict modified output data when the output data in a region specified by the user among the output data displayed by the display unit is changed, wherein the display unit displays the searched input data.

7. The manufacturing condition search unit comprehensively generates input data included within a predetermined constraint range, and inputs the comprehensively generated input data into the trained model execution unit to search for input data that corresponds to output data similar to the modified output data among the output data predicted by the trained model execution unit, as described in claim 6.

8. A manufacturing condition design support method for a process of manufacturing SiC epitaxial wafers by the CVD method, wherein the computer of a manufacturing condition design support device assists in designing manufacturing conditions, and the computer stores a trained model in a storage unit, which is trained using input data related to specific manufacturing conditions and output data of specific items indicating the state inside the furnace, from input and output data of specific items indicating the state inside the furnace, when a simulation is performed by a simulation device for the process to predict the state inside the furnace; reads a trained model corresponding to a specific item specified by the user from the storage unit; when input data related to specific manufacturing conditions is input by the user, the read trained model is executed to predict output data of specific items indicating the state inside the furnace; and displays the specified specific item, the input input data related to the specific manufacturing conditions, and the predicted output data.

9. A manufacturing condition design support program for a manufacturing condition design support device that assists in the design of manufacturing conditions in a process for manufacturing SiC epitaxial wafers by the CVD method, wherein the computer of the manufacturing condition design support device stores in a storage unit a trained model that has been learned using input data related to specific manufacturing conditions and output data of specific items indicating the state inside the furnace, which are input data from a simulation device for the process that predicts the state inside the furnace; reads out a trained model corresponding to a specific item specified by the user from the storage unit; when input data related to specific manufacturing conditions is input by the user, executes the read-out trained model and predicts output data of specific items indicating the state inside the furnace; and displays the specified specific item, the input input data related to the specific manufacturing conditions, and the predicted output data.

10. A manufacturing method for manufacturing a SiC epitaxial wafer by CVD under manufacturing conditions specified by input data searched by a manufacturing condition search unit of the manufacturing condition design support system described in claim 6 or 7.

Images