Data collection device, data collection method, data collection program and substrate processing method

Abstract

To collect appropriate data in exploring process conditions for substrate processing.SOLUTION: A data collection system includes a first substrate processing apparatus having a first processing space, a second substrate processing apparatus having a second processing space, and a data collection device connected to the first substrate processing apparatus and the second substrate processing apparatus, and further includes a correction amount calculation unit that compares observation data observed when substrates of the same or similar shape are processed in the first processing space and the second processing space under the same processing conditions, and calculates a correction amount for correcting the observation data observed when processed in the second processing space, and a collection unit that, when searching for processing conditions by processing substrates under different processing conditions in the second processing space, corrects the observation data observed when processed in the second processing space on the basis of the correction amount, and collects the corrected observation data.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a data collection system, a data collection device, a data collection method, and a data collection program. [Background technology]

[0002] In the field of substrate processing apparatuses, for example, optimum processing conditions are searched for by repeating experiments while changing the processing conditions so that the shape of a processed substrate matches a target substrate shape. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2008-034877 A [Patent Document 2] JP 2019-207951 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the case of current substrate processing apparatuses, it is difficult to collect detailed data showing the state inside the processing space, and therefore it is difficult to predict how the processing conditions should be changed to approach the target substrate shape. Therefore, in order to efficiently search for optimal processing conditions, it is desirable to collect detailed data showing the state inside the processing space.

[0005] In addition, even if the optimal processing conditions found using the substrate processing apparatus of the experimental machine are applied to the substrate processing apparatus of the mass production machine, the target substrate shape may not be obtained in the mass production machine due to the difference between the two. In such a case, further experiments are required. Therefore, in order to efficiently find the optimal processing conditions, it is desirable to collect data equivalent to that of the mass production machine during the experiment.

[0006] The present disclosure provides a data collection system, a data collection device, a data collection method, and a data collection program for collecting appropriate data in searching for processing conditions in substrate processing. [Means for solving the problem]

[0007] A data collection system according to an aspect of the present disclosure has, for example, the following configuration. A data collection system comprising: a first substrate processing apparatus having a first processing space; a second substrate processing apparatus having a second processing space different from the first processing space; and a data collection apparatus connected to the first substrate processing apparatus and the second substrate processing apparatus, a correction amount calculation unit that compares observed data obtained by processing a substrate having the same or a similar shape in the first processing space and the second processing space under the same processing conditions, and calculates a correction amount for correcting the observed data obtained by processing the substrate in the second processing space; The apparatus further includes a collection unit that, when searching for processing conditions by processing a substrate under different processing conditions in the second processing space, corrects observation data observed during processing in the second processing space based on the correction amount and collects the corrected observation data. Effect of the Invention

[0008] It is possible to provide a data collection system, a data collection device, a data collection method, and a data collection program for collecting appropriate data in searching for processing conditions in substrate processing. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram showing an overview of the configuration and processing of a data collection system according to the first embodiment. [Diagram 2] FIG. 2 is a diagram showing a specific example of the calibration amount calculation process executed by the data collection system according to the first embodiment. [Diagram 3]FIG. 3 is a diagram showing a specific example of the correction amount calculation process executed by the data collection system according to the first embodiment. [Figure 4] FIG. 4 is a diagram showing a specific example of the data collection process and the condition change process executed by the data collection system according to the first embodiment. [Diagram 5] FIG. 5 is a diagram illustrating an example of collected data. [Figure 6] FIG. 6 is a diagram showing a specific example of the virtual measurement model learning process, the shape simulation model learning process, and the condition narrowing down process executed by the data collection system according to the first embodiment. [Figure 7] FIG. 7 is a diagram showing a detailed example of the virtual measurement model learning process. [Figure 8] FIG. 8 is a diagram showing a detailed example of the shape simulation model learning process. [Figure 9] FIG. 9 is a diagram showing a detailed example of the condition narrowing down process. [Figure 10] FIG. 10 is a diagram showing a specific example of the virtual measurement process and the shape simulation process executed by the data collection system according to the first embodiment. [Figure 11] FIG. 11 is a diagram illustrating an example of a hardware configuration of the data collecting device and the data analyzing device. [Figure 12] FIG. 12 is a diagram showing an overview of the configuration and processing of a data collection system according to the second embodiment. [Figure 13] FIG. 13 is a diagram showing a specific example of the correction amount calculation process executed by the data collection system according to the second embodiment. [Figure 14] FIG. 14 is a diagram showing a specific example of the correction amount calculation process executed by the data collection system according to the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, each embodiment will be described with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functions are denoted by the same reference numerals, and redundant description will be omitted.

[0011] [First embodiment] <Overview of data collection system configuration and processing> First, a configuration of a data collection system according to a first embodiment and an overview of processing executed by the data collection system according to the first embodiment will be described. Fig. 1 is a diagram showing the configuration of the data collection system according to the first embodiment and an overview of processing.

[0012] As shown in FIG. 1, the data collection system 100 includes: A substrate processing apparatus 110, which is an example of a reference substrate processing apparatus; A substrate processing apparatus A120, which is an example of a first substrate processing apparatus; Substrate processing apparatus B131, substrate processing apparatus C132, and substrate processing apparatus D133, which are examples of the second substrate processing apparatus; data collection device 140; Data analysis device 150, has.

[0013] The data collection system 100 is applied, for example, in a situation where a substrate processing apparatus manufacturer supports a substrate manufacturer so that the substrate manufacturer can mass-produce substrates having a target substrate shape.

[0014] The substrate processing apparatus 110 is, for example, a so-called mass production machine installed in a substrate manufacturer. The substrate processing apparatus 110 has a reference chamber, which is an example of a reference processing space, and processes substrates under predetermined processing conditions. The substrate processing apparatus 110 is equipped with various observation sensors, and outputs observation data observed by the various observation sensors during substrate processing.

[0015] The substrate processing apparatus A120 is installed, for example, at a substrate processing apparatus manufacturer (same as a mass production machine). The substrate processing apparatus A120 has a first chamber, which is an example of a first processing space identical to the reference processing space, and processes substrates under predetermined processing conditions. The substrate processing apparatus A120 is equipped with various observation sensors, and outputs observation data observed by the various observation sensors during substrate processing.

[0016] The substrate processing apparatus B131 is, for example, a so-called experimental machine installed at a substrate processing apparatus manufacturer. The substrate processing apparatus B131 has a second chamber, which is an example of a second processing space identical to the reference processing space, and processes substrates under predetermined processing conditions. The substrate processing apparatus B131 is equipped with various observation sensors, and outputs observation data observed by the various observation sensors during substrate processing. In addition, a plasma probe is additionally installed in the second chamber, and outputs plasma measurement data measured by the plasma probe during substrate processing.

[0017] The substrate processing apparatus C132 is installed, for example, at a substrate processing apparatus manufacturer (experimental machine). The substrate processing apparatus C132 has a third chamber, which is an example of a second processing space identical to the reference processing space, and processes substrates under predetermined processing conditions. The substrate processing apparatus C132 is equipped with various observation sensors, and outputs observation data observed by the various observation sensors during substrate processing. In addition, a wear amount sensor is additionally installed in the third chamber, and outputs wear amount measurement data indicating the wear amount of parts in the third chamber measured by the wear amount sensor during substrate processing.

[0018] The substrate processing apparatus D133 is installed, for example, at a substrate processing apparatus manufacturer (experimental machine). The substrate processing apparatus D133 has a fourth chamber, which is an example of a second processing space identical to the reference processing space, and processes substrates under predetermined processing conditions. The substrate processing apparatus D133 is equipped with various observation sensors, and outputs observation data observed by the various observation sensors during substrate processing. In addition, a particle sensor is additionally installed in the fourth chamber, and outputs particle measurement data indicating particles in the fourth chamber measured by the particle sensor during substrate processing.

[0019] Although not shown in FIG. 1, the data collection system 100 includes a shape measuring device that measures the substrate shapes of the unprocessed and processed substrates, and is configured to output measured shape data.

[0020] The data collection device 140 is installed, for example, at a manufacturer of the substrate processing apparatuses. The data collection device 140 is connected to the substrate processing apparatuses 110, the substrate processing apparatus A 120, and the substrate processing apparatuses B 131 to D 133, and a shape measuring device (not shown).

[0021] Thus, the data collecting device 140 collects the processing conditions used by each substrate processing apparatus when processing a substrate, and observation data observed by various observation sensors while each substrate processing apparatus is processing a substrate.

[0022] The data collecting device 140 also collects shape data measured on the unprocessed and processed substrates processed by each substrate processing device.

[0023] Furthermore, the data collecting device 140 collects plasma measurement data, wear amount measurement data, and particle measurement data measured by the plasma probe, wear amount sensor, and particle sensor, respectively, while the substrate processing apparatuses B 131 to D 133 are processing substrates.

[0024] The data analysis device 150 is installed, for example, at a manufacturer of the substrate processing apparatus. The data analysis device 150 performs learning processing on various models using data collected by the data collection device 140 (referred to as collected data).

[0025] The following describes an overview of the processing executed by the data collection system 100 (processing executed by each device constituting the data collection system 100 alone or by a plurality of devices working together). As shown in FIG. 1, the processing executed by the data collection system 100 is as follows: "An experimental phase for searching for optimal conditions" "Optimal Condition Processing Phase" and They can be broadly divided into:

[0026] Among these, the experimental phase for searching optimal conditions is a phase for searching optimal processing conditions for realizing a target substrate shape. Also, the processing phase under optimal conditions is a phase for processing substrates under the optimal processing conditions searched for. An overview of the processing in each phase will be described below.

[0027] (1) Experimental phase to find optimal conditions In step S161, the substrate processing apparatus 110, the substrate processing apparatus A 120, and the data collecting apparatus 140 perform a calibration amount calculation process.

[0028] Specifically, the substrate processing apparatus 110 and the substrate processing apparatus A 120 process substrates of the same or similar shapes under the same processing conditions, and the data collector 140 collects observation data observed by various observation sensors.

[0029] Moreover, the data collecting device 140 calculates a calibration amount for calibrating the difference in the observed data due to the "instrumental difference" between the substrate processing device 110 and the substrate processing device A 120.

[0030] The data collection device 140 subsequently calibrates the observation data output from the substrate processing apparatus A 120 based on the calculated calibration amount. This allows the data collection device 140 to subsequently collect from the substrate processing apparatus A 120 the same observation data as the observation data output when a substrate of the same or similar shape is processed in the substrate processing apparatus 110 under the same processing conditions.

[0031] In step S162, the substrate processing apparatus A120, the substrate processing apparatus B131 to the substrate processing apparatus D133, and the data collecting apparatus 140 perform a correction amount calculation process. Specifically, the substrate processing apparatus A120 and the substrate processing apparatus B131 to the substrate processing apparatus D133 process substrates of the same or similar shapes under the same processing conditions, and the data collecting apparatus 140 collects observation data observed by the various observation sensors.

[0032] In addition, the data collection device 140 calculates the difference between various observation data collected from the substrate processing device A 120 and the substrate processing device B 131, and calculates a correction amount to correct the effect of disturbance caused by installing the plasma probe in the second chamber.

[0033] The data collection device 140 subsequently corrects the observation data output from the substrate processing apparatus B 131 based on the calculated correction amount, thereby enabling the data collection device 140 to subsequently collect from the substrate processing apparatus B 131 the same observation data as that collected when a substrate of the same or similar shape is processed in the substrate processing apparatus A 120 under the same processing conditions.

[0034] Similarly, the data collection device 140 calculates the difference between various observation data collected by the substrate processing device A 120 and the substrate processing device C 132, and calculates a correction amount to correct the effect of disturbance caused by installing a wear amount sensor in the third chamber.

[0035] The data collection device 140 subsequently corrects the observation data output from the substrate processing device C 132 based on the calculated correction amount, thereby enabling the data collection device 140 to subsequently collect from the substrate processing device C 132 the same observation data as that collected when a substrate of the same or similar shape is processed in the substrate processing device A 120 under the same processing conditions.

[0036] Similarly, the data collection device 140 calculates the difference between various observation data collected by the substrate processing device A 120 and the substrate processing device D 133, and calculates a correction amount to correct the effect of disturbance caused by installing a particle sensor in the fourth chamber.

[0037] The data collection device 140 subsequently corrects the observation data output from the substrate processing apparatus D133 based on the calculated correction amount. This allows the data collection device 140 to subsequently collect from the substrate processing apparatus D133 the same observation data as that collected when a substrate of the same or similar shape is processed in the substrate processing apparatus A120 under the same processing conditions.

[0038] In step S163, the data collecting device 140 changes the processing conditions used by the substrate processing apparatuses B131 to D133 when processing substrates in order to search for optimal processing conditions for achieving a target substrate shape. The data collecting device 140 also sets the changed processing conditions in the substrate processing apparatuses B131 to D133.

[0039] In step S164, the substrate processing apparatuses B 131 to D 133 and the data collecting apparatus 140 perform a data collecting process.

[0040] Specifically, the substrate processing apparatuses B131 to D133 process substrates of the same or similar shape under the changed processing conditions, and output various observation data, plasma measurement data, wear amount measurement data, and particle measurement data. In addition, a shape measuring device (not shown) outputs shape data of the substrate before processing and shape data of the substrate after processing.

[0041] The data collecting device 140 also corrects various observation data using the correction amount and collects the various observation data after correction. The data collecting device 140 also collects plasma measurement data output from the substrate processing device B 131, wear amount measurement data output from the substrate processing device C 132, and particle measurement data output from the substrate processing device D 133 (these are referred to as "various measurement data"). Furthermore, the data collecting device 140 collects shape data of the unprocessed substrate and shape data of the processed substrate (these are referred to as "various shape data") output from a shape measuring device (not shown).

[0042] Furthermore, the data collecting device 140 displays the collected data (the previous processing conditions, various observation data, various measurement data, and various shape data) to the experimenter. This allows the experimenter to observe the various shape data while referring to not only the various observation data but also the various measurement data, which is detailed data that indicates the state inside the chamber. As a result, the experimenter can set more appropriate processing conditions for the next processing condition to approach the target substrate shape, and can efficiently search for optimal processing conditions.

[0043] The condition changing process in step S163 and the data collecting process in step S164 are repeatedly executed. After being repeated a predetermined number of times, the correction amount calculation process in step S162 is executed. This is because the various measurement data output from the substrate processing apparatuses B131 to D133 change over time, and by periodically updating the correction amount, the influence of disturbances can be constantly eliminated.

[0044] In this way, the calibration amount calculation process (step S161) and the correction amount calculation process (step S162) are performed, and in the search for optimal processing conditions, the data collecting device 140 -Influence of machine differences, - Effects of disturbances, Observation data from which the defects have been removed (observation data equivalent to the observation data output from the substrate processing apparatus 110, which is a mass production machine) can be collected from the substrate processing apparatuses B 131 to D 133.

[0045] When a certain amount of data is collected by repeatedly executing the condition change process in step S163 and the data collection process in step S164, in step S165, the data analysis device 150 performs a learning process on the virtual measurement model using the collected data. The virtual measurement model is a model that predicts various measurement data (plasma measurement data, wear measurement data, particle measurement data) during substrate processing based on various observation data.

[0046] In the data analysis device 150, various observation data are used as input data, and various measurement data are used as correct answer data to perform learning processing. In this way, the data analysis device 150 can generate a learned virtual measurement model.

[0047] The generated trained virtual measurement model is mounted on, for example, a substrate processing apparatus 110 to which no plasma probe, wear sensor, or particle sensor is attached.

[0048] As a result, in the substrate processing apparatus 110, when processing a substrate in the "processing phase under optimal conditions", plasma measurement data, wear measurement data, and particle measurement data can be predicted based on various observation data and displayed to an operator, etc.

[0049] In step S166, the data analysis device 150 performs a learning process for the shape simulation model using the collected data. The shape simulation model is a model for predicting shape data of a processed substrate when the substrate is processed.

[0050] The data analysis device 150 performs a learning process using various observation data, various measurement data, processing conditions, and shape data of the unprocessed substrate as input data and the shape data of the processed substrate as correct answer data. In this way, the data analysis device 150 generates a learned shape simulation model.

[0051] In this way, by performing the learning process using input data that includes not only various observation data but also various measurement data, it is possible to generate a learned shape simulation model with high prediction accuracy.

[0052] The generated trained shape simulation model is loaded, for example, in the substrate processing apparatus 110. This allows the substrate processing apparatus 110 to predict shape data of a processed substrate after the substrate is processed in the "processing phase under optimal conditions."

[0053] If the target substrate shape has not been reached at this point (at the end of step S166) (if the difference from the target substrate shape is equal to or greater than a predetermined threshold), the process continues to change the process conditions and the search for optimal process conditions continues.

[0054] Specifically, first, in step S167, the substrate processing apparatus A 120, the substrate processing apparatus B 131 to the substrate processing apparatus D 133, and the data collecting apparatus 140 perform a correction amount calculation process. Note that the correction amount calculation process in step S167 is the same as the correction amount calculation process in step S162, and therefore a description thereof will be omitted here.

[0055] In step S168, the data collecting device 140 and the data analyzing device 150 perform a condition narrowing process to narrow down the processing conditions using the trained shape simulation model generated in step S166.

[0056] Specifically, the data analysis device 150 uses the learned shape simulation model to predict shape data of the processed substrate, thereby narrowing down the processing conditions for approaching the target substrate shape.

[0057] In parallel, the data analysis device 150 narrows down the processing conditions for which no collected data has been obtained, in order to improve the prediction accuracy of the trained shape simulation model.

[0058] As a result, the data analysis device 150 sets the processing conditions narrowed down from the viewpoint of searching for optimal processing conditions or the processing conditions narrowed down from the viewpoint of improving prediction accuracy as changed processing conditions in the substrate processing devices B131 to D133.

[0059] In step S169, the substrate processing apparatuses B 131 to D 133 and the data collecting apparatus 140 perform data collecting processing. Note that the data collecting processing in step S169 is the same as the data collecting processing in step S164, and therefore a description thereof will be omitted here.

[0060] Also, the condition narrowing down process in step S168 and the data collection process in step S169 are repeatedly executed, as are the condition change process in step S163 and the data collection process in step S164. However, unlike the cases of steps S163 and S164, the processing conditions are appropriately narrowed down in step S168, so that the speed at which the optimal processing conditions are approached is accelerated (i.e., the optimal processing conditions can be searched for more efficiently).

[0061] Also, similar to step S162, when the processing of the substrate under the changed processing conditions has been repeated a predetermined number of times, a correction amount calculation process is executed in step S167 to correct the effects of changes over time in the various observation data output from the substrate processing apparatus B131 to the substrate processing apparatus D133.

[0062] In this way, by repeatedly processing the substrate while changing the processing conditions, when the target substrate shape is reached (when the error from the target substrate shape is less than a predetermined threshold value), the processing conditions at the time of reaching the target substrate shape are determined as the optimal processing conditions, and the experimental phase for searching for the optimal conditions is completed.

[0063] When the experimental phase for searching for optimal conditions is completed, the process moves to an "optimum condition processing phase." In the optimal condition processing phase, a substrate is processed under optimal processing conditions using a substrate processing apparatus 110 equipped with the trained virtual measurement model and trained shape simulation model generated in the experimental phase for searching for optimal conditions.

[0064] 1, in step S171, the substrate processing apparatus 110 executes a virtual measurement process when processing a substrate under optimal processing conditions. As a result, various observation data observed during the processing of the substrate are input to a trained virtual measurement model, and various measurement data (plasma measurement data, wear amount measurement data, particle measurement data) are predicted. In addition, the various predicted measurement data are displayed to an operator, etc.

[0065] This allows an operator or the like to monitor plasma measurement data, wear amount measurement data, and particle measurement data that could not be monitored conventionally when processing a substrate.

[0066] In step S172, the substrate processing apparatus 110 executes a shape simulation process when the substrate is processed under optimal processing conditions. As a result, various observation data observed during the processing of the substrate, various predicted measurement data, processing conditions, and shape data of the substrate before processing are input to the trained shape simulation model. In addition, the trained shape simulation model predicts shape data of the substrate after processing, and displays the predicted shape data to an operator, etc.

[0067] As a result, an operator or the like can determine whether the processed substrate is good or bad without inspecting every single one of the processed substrates.

[0068] <Examples of processing performed by the data collection system> Next, a specific example of each process executed by the data collection system 100 will be described.

[0069] (1) Specific example of calibration amount calculation process (step S161) First, a specific example of the calibration amount calculation process (step S161) executed by the data collection system 100 will be described. Fig. 2 is a diagram showing a specific example of the calibration amount calculation process executed by the data collection system according to the first embodiment.

[0070] The example in Figure 2 shows that the substrate processing apparatus 110 and the substrate processing apparatus A120 process unprocessed substrates of the same or similar shape under processing condition "Condition 0", and output observation data "Observation data M_α0" and "Observation data M_A0", respectively.

[0071] In addition, a data collection program is installed in the data collection device 140, and by executing the program, the data collection device 140: ·Calibration amount calculation unit 210, ·Correction amount calculation unit 220, A collection unit 230, A condition change unit 240, It functions as:

[0072] In the calibration amount calculation process (step 161), the data collecting device 140 operates the calibration amount calculation unit 210 to acquire the observation data "observation data M_α0" output from the substrate processing apparatus 110. The calibration amount calculation unit 210 also acquires the observation data "observation data M_A0" output from the substrate processing apparatus A 120. The calibration amount calculation unit 210 also calculates the calibration amount a for matching the observation data "observation data M_A0" with the observation data "observation data M_α0". Furthermore, the calibration amount calculation unit 210 notifies the correction amount calculation unit 220 of the calculated calibration amount a.

[0073] (2) Specific Example of Correction Amount Calculation Process (Step S162) Next, a specific example of the correction amount calculation process (step S162) executed by the data collection system 100 will be described. Fig. 3 is a diagram showing a specific example of the correction amount calculation process executed by the data collection system according to the first embodiment.

[0074] The example in Figure 3 shows that substrate processing apparatus A120 and substrate processing apparatuses B131 to D133 process unprocessed substrates of the same shape under processing condition "Condition 1," and output observation data "Observation Data M_A1" to "Observation Data M_D1," respectively.

[0075] In the correction amount calculation process (step S162), the correction amount calculation unit 220 of the data collecting device 140 operates to acquire the observation data (=“observation data M_A1” to “observation data M_D1”) output from the substrate processing apparatus A 120 and the substrate processing apparatuses B 131 to D 133.

[0076] Furthermore, the correction amount calculation unit 220 calibrates the acquired observation data = "observation data M_A1" using the calibration amount a, and generates calibrated observation data.

[0077] Furthermore, the correction amount calculation unit 220 calculates a correction amount b for making the observation data="observation data M_B1" coincide with the calibrated observation data, and notifies the collection unit 230 of the correction amount b.

[0078] Similarly, the correction amount calculation unit 220 calculates the correction amount c for matching the observation data="observation data M_C1" with the calibrated observation data, and notifies the collection unit 230 of the correction amount c.

[0079] Similarly, the correction amount calculation unit 220 calculates the correction amount d for matching the observation data="observation data M_D1" with the calibrated observation data, and notifies the collection unit 230 of the correction amount d.

[0080] (3) Specific examples of the condition change process (step S163) and the data collection process (step S164) Next, a specific example of the condition change process (step S163) and the data collection process (step S164) executed by the data collection system 100 will be described. Fig. 4 is a diagram showing a specific example of the data collection process and the condition change process executed by the data collection system according to the first embodiment.

[0081] In the condition change process (step S163), the condition change unit 240 operates in the data collecting device 140. The example in Fig. 4 shows a state in which the condition change unit 240 transmits processing conditions="Condition 2", "Condition 3", ... to the substrate processing apparatus B 131, the substrate processing apparatus C 132, and the substrate processing apparatus D 133 in order.

[0082] In the example of FIG. 4, the substrate processing apparatus B131 processes unprocessed substrates of the same shape under the processing conditions "Condition 2", "Condition 3", . . . , respectively. Observation data = "Observation data M_B2", "Observation data M_B3", ..., · Measurement data = "Measurement data I_B2", "Measurement data I_B3", ..., 4 is output by the substrate processing apparatus B131. Shape data = "Shape data S_B2", "Shape data S_B3", from the unprocessed board of Shape data = "Shape data S_B2'", "Shape data S_B3'", ... processed substrate, It shows how it was generated.

[0083] Similarly, in the example of FIG. 4, the substrate processing apparatus C132 processes unprocessed substrates of the same shape under processing conditions = "Condition 2", "Condition 3", ..., respectively. Observation data = "Observation data M_C2", "Observation data M_C3", ..., · Measurement data = "Measurement data I_C2", "Measurement data I_B3", ..., The example in FIG. 4 shows how the substrate is processed by the substrate processing apparatus C132, Shape data = "Shape data S_C2", "Shape data S_C3", from the unprocessed board of Shape data = "Shape data S_C2'", "Shape data S_C3'", ... processed substrate, It shows how it was generated.

[0084] Similarly, in the example of FIG. 4, the substrate processing apparatus D133 processes unprocessed substrates of the same shape under processing conditions = "Condition 2", "Condition 3", ..., respectively. Observation data = "Observation data M_D2", "Observation data M_D3", ..., · Measurement data = "Measurement data I_D2", "Measurement data I_D3", ..., The example in FIG. 4 is processed by the substrate processing apparatus D133, Shape data = "Shape data S_D2", "Shape data S_D3", from the unprocessed board of ... Shape data = "Shape data S_D2'", "Shape data S_D3'", ... processed substrate, It shows how it was generated.

[0085] In the data collection process (step S164), the collection unit 230 operates in the data collection device 140. As shown in FIG.

[0086] In the example of FIG. 4, the storage unit 401 Shape data of unprocessed substrate = "Shape data S_B2", "Shape data S_B3", ..., Shape data of processed substrate = "Shape data S_B2'", "Shape data S_B3'", ... is stored in the collected data storage unit 250.

[0087] 4 also shows how the correction storage unit 402 corrects the observation data="observation data M_B2", "observation data M_B3", ... using the correction amount b, and stores the corrected observation data in the collected data storage unit 250. Also, the example in Fig. 4 shows how the correction storage unit 402 stores the measurement data="measurement data I_B2", "measurement data I_B3", ... in the collected data storage unit 250.

[0088] Similarly, in the example of FIG. 4, the storage unit 403 Shape data of unprocessed substrate = "Shape data S_C2", "Shape data S_C3", ..., The shape data of the processed substrates="shape data S_C2'", "shape data S_C3'", . . . are stored in the collected data storage unit 250. FIG.

[0089] 4 also shows how the correction storage unit 404 corrects the observation data="observation data M_C2", "observation data M_C3", ... using the correction amount c, and stores the corrected observation data in the collected data storage unit 250. Also, the example in Fig. 4 shows how the correction storage unit 404 stores the measurement data="measurement data I_C2", "measurement data I_C3", ... in the collected data storage unit 250.

[0090] Similarly, in the example of FIG. 4, the storage unit 405 Shape data of unprocessed substrate = "Shape data S_D2", "Shape data S_D3", ..., The shape data of the processed substrates="shape data S_D2'", "shape data S_D3'", . . . are stored in the collected data storage unit 250. FIG.

[0091] 4 also shows how the correction storage unit 406 corrects the observation data="observation data M_D2", "observation data M_D3", ... using the correction amount d, and stores the corrected observation data in the collected data storage unit 250. Also, the example in Fig. 4 shows how the correction storage unit 406 stores the measurement data="measurement data I_D2", "measurement data I_D3", ... in the collected data storage unit 250.

[0092] Next, a description will be given of the collected data stored in the collected data storage unit 250. Fig. 5 is a diagram showing an example of the collected data.

[0093] As shown in FIG. 5, the collected data 500 includes information items such as "processing conditions", "observation data", "measurement data", "pre-processing substrate shape data", and "post-processing substrate shape data".

[0094] The "processing conditions" store, for example, set values ​​such as the power of the high frequency power supply, the power of the low frequency power supply, the pressure in the chamber, the temperature in the chamber, the gas flow rate, the gas type, etc. The processing conditions = "Condition 2", "Condition 3", ... "Condition x" are configured by combinations in which at least one of these set values ​​is different.

[0095] The "observation data" stores, for example, observed values ​​such as DC self-bias voltage, potential difference, reflected wave power, gas flow rate, plasma density, ion energy, ion flow rate, etc. The observation data = "observation data M_B1", "observation data M_C1", and "observation data M_D1" are each composed of a combination of these observed values.

[0096] "Measurement data" stores the measurement values ​​of additionally installed sensors, such as plasma measurement data, wear measurement data, and particle measurement data. Specifically, measurement data = "measurement data I_B1" includes the measurement value of plasma measurement data. Furthermore, measurement data = "measurement data I_C1" includes the measurement value of wear measurement data. Furthermore, measurement data = "measurement data I_D1" includes the measurement value of particle measurement data.

[0097] The "pre-processing substrate shape data" stores shape values ​​such as critical dimensions, depth, taper angle, tilt angle, bowing, etc. Each of the pre-processing substrate shape data = "shape data S_B2" to "shape data S_D2" is composed of a combination of these shape values.

[0098] The "post-processing substrate shape data" stores shape values ​​such as critical dimensions, depth, taper angle, tilt angle, bowing, etc. The post-processing substrate shape data = "shape data S_B2'" to "shape data S_D2'" are each composed of a combination of these shape values.

[0099] (4) Specific examples of the virtual measurement model learning process (step S165), the shape simulation model learning process (step S166), and the condition narrowing process (step S168) (4-1) Specific examples (overview) Next, specific examples (overview) of the virtual measurement model learning process (step S165), the shape simulation model learning process (step S166), and the condition narrowing down process (step S168) executed by the data collection system 100 will be described.

[0100] FIG. 6 is a diagram showing a specific example of the virtual measurement model learning process, the shape simulation model learning process, and the condition narrowing down process executed by the data collection system according to the first embodiment.

[0101] A data analysis program is installed in the data analysis device 150. By executing the program, the data analysis device 150 performs the following: A virtual measurement model learning unit 610; A shape simulation model learning unit 630; - 640 trained shape simulation models, A shape simulation control unit 650 (an example of a control unit and a determination unit), It functions as:

[0102] In the virtual measurement model learning process (step S165), the data analysis device 150 operates the virtual measurement model learning unit 610 to read the collected data 500 stored in the collected data storage unit 250. In addition, the virtual measurement model learning unit 610 performs a learning process on the virtual measurement model using the read collected data 500.

[0103] The trained virtual measurement model 620 generated by the virtual measurement model training unit 610 performing a training process on the virtual measurement model is mounted on the substrate processing apparatus 110.

[0104] On the other hand, in a shape simulation model learning process (step S166), data analysis device 150 operates shape simulation model learning unit 630 to read collected data 500 stored in collected data storage unit 250. Furthermore, shape simulation model learning unit 630 uses read collected data 500 to perform a learning process for the shape simulation model.

[0105] The trained shape simulation model 640 generated by the shape simulation model training unit 630 performing the learning process is loaded onto the substrate processing apparatus 110. The trained shape simulation model 640 is executed by the shape simulation control unit 650 in the condition narrowing down process (step S168).

[0106] Specifically, in the condition narrowing down process (step S168), the data analysis device 150 operates the shape simulation control unit 650. The shape simulation control unit 650 executes the learned shape simulation model under various processing conditions to obtain a prediction result for the shape data of the processed substrate. The shape simulation control unit 650 also compares the shape data of the processed substrate, which is the prediction result, with a target substrate shape, thereby narrowing down the processing conditions capable of realizing the target substrate shape. Furthermore, the narrowed down processing conditions are notified to the condition change unit 240 of the data collection device 140. As a result, the condition change unit 240 can set the processing conditions narrowed down from the viewpoint of searching for optimal processing conditions as changed processing conditions in the substrate processing apparatuses B131 to D133.

[0107] Furthermore, the shape simulation control unit 650 refers to the collected data 500 stored in the collected data storage unit 250, and determines the processing conditions that are not used in the learning process of the shape simulation model. Furthermore, the shape simulation control unit 650 notifies the condition changing unit 240 of the determined processing conditions. As a result, the condition changing unit 240 sets the processing conditions determined by the shape simulation control unit 650 as changed processing conditions in the substrate processing apparatuses B131 to D133. As a result, the data collection process (step S169) is performed in the substrate processing apparatuses B131 to D133 and the data collection apparatus 140, and the collected data 500 in the collected data storage unit 250 includes the following processing conditions that correspond to the determined processing conditions: -Various observation data, -Various measurement data, Various shape data (pre-processing board shape data, post-processing board shape data), will be newly stored.

[0108] The shape simulation control unit 650 performs a re-learning process on the trained shape simulation model using the determined processing conditions and the various observation data, measurement data, and shape data (pre-processing substrate shape data, post-processing substrate shape data) that are newly stored. The shape simulation control unit 650 repeatedly executes the determination of the processing conditions and the re-learning process multiple times. At this time, the shape simulation control unit 650 identifies processing conditions with low prediction accuracy, and notifies the condition change unit 240 of the identified processing conditions with low prediction accuracy (i.e., narrows down the processing conditions for performing the re-learning process on the trained shape simulation model). As a result, the condition change unit 240 can set the processing conditions narrowed down from the viewpoint of improving the prediction accuracy of the trained shape simulation model 640 as changed processing conditions in the substrate processing apparatuses B131 to D133.

[0109] (4-2) Detailed Example (Virtual Measurement Model Learning Process (Step S165)) Next, a more detailed specific example of the virtual measurement model learning process (step S165) will be described. Fig. 7 is a diagram showing a detailed specific example of the virtual measurement model learning process.

[0110] As shown in FIG. 7, the virtual measurement model learning unit 610 includes a virtual measurement model 710 and a comparison / modification unit 720 .

[0111] The setting values ​​stored in the "processing conditions" of the collected data 500 and the observed values ​​stored in the "observed data" are input as input data to the virtual measurement model 710. As a result, the virtual measurement model 710 outputs output data.

[0112] On the other hand, the measurement value stored in the "measurement data" of the collected data 500 is input to the comparison / change unit 720 as the supervised data. The comparison / change unit 720 calculates an error by comparing the output data output from the virtual measurement model 710 with the measurement value input as the supervised data, and updates the model parameters of the virtual measurement model 710 according to the calculated error. This allows the virtual measurement model learning unit 610 to generate the learned virtual measurement model 620.

[0113] (4-2) Detailed Example (Shape Simulation Model Learning Process (Step S166)) Next, a more detailed example of the shape simulation model learning process (step S166) will be described below. Fig. 8 is a diagram showing a detailed example of the shape simulation model learning process.

[0114] As shown in FIG. 8, the shape simulation model learning unit 630 includes a shape simulation model 810 and a comparison / modification unit 820.

[0115] The configuration values ​​stored in the "processing conditions" of the collected data 500, the observation values ​​stored in the "observation data", the measurement values ​​stored in the "measurement data", and the configuration values ​​stored in the "pre-processing substrate configuration data" are input as input data to the shape simulation model 810. As a result, the shape simulation model 810 outputs output data.

[0116] On the other hand, the comparison / change unit 820 receives the shape values ​​stored in the "post-processing substrate shape data" of the collected data 500 as the correct answer data. The comparison / change unit 820 calculates an error by comparing the output data output from the shape simulation model 810 with the shape values ​​input as the correct answer data, and updates the model parameters of the shape simulation model 810 according to the calculated error. This allows the shape simulation model learning unit 630 to generate the learned shape simulation model 640.

[0117] (4-3) Details of the specific example (Details of the condition narrowing down process (step S168)) Next, a more detailed specific example of the condition narrowing down process (step S168) will be described. Fig. 9 is a diagram showing details of a specific example of the condition narrowing down process.

[0118] 9, the shape simulation control unit 650 executes the learned shape simulation model under processing conditions="condition 101", "condition 102", .... In this way, the shape simulation control unit 650 acquires shape data of the processed substrate predicted by the learned shape simulation model 640 ("shape data S101", "shape data S102", ...).

[0119] Furthermore, the shape simulation control unit 650 narrows down the processing conditions capable of realizing the target substrate shape by comparing the acquired shape data ("shape data S101", "shape data S102", ...) with the shape data of the target substrate shape. Furthermore, the narrowed down processing conditions are notified to the condition change unit 240 of the data collecting device 140.

[0120] The example in FIG. 9 shows how the process conditions are narrowed down from the viewpoint of searching for the optimal process condition, and ultimately, the process condition="condition x" is searched for as the optimal process condition.

[0121] Furthermore, the shape simulation control unit 650 refers to the collected data 500 stored in the collected data storage unit 250, and determines the processing conditions that are not used in the learning process of the shape simulation model.

[0122] A dashed area 900 shown in the lower left of Fig. 9 is a schematic representation of a range that can be searched as a processing condition. Meanwhile, a solid area 910 in Fig. 9 is a schematic representation of a range (interpolation range) used in the learning process of the shape simulation model.

[0123] As described above, in the shape simulation control unit 650, from the viewpoint of improving the prediction accuracy of the trained shape simulation model 640, processing conditions with low prediction accuracy are narrowed down from among the processing conditions not used in the learning process of the shape simulation model.

[0124] The example in FIG. 9 shows a state in which the shape simulation control unit 650 narrows down an area within a dashed line area 900 from areas other than a solid line area 910 to a solid line area 920.

[0125] In addition, the example in Figure 9 shows that the shape simulation control unit 650 notifies the condition change unit 240 of “condition x+1”, “condition x+2”, ... as processing conditions included in the solid line area 920, and inputs them into the learned shape simulation model 640.

[0126] As a result, in the trained shape simulation model 640, Processing conditions = "condition x+1", "condition x+2", ... and Various observation data, various measurement data, and pre-processed substrate shape data when the substrate processing apparatus B131 to the substrate processing apparatus D133 process substrates under the processing conditions = "condition x+1", "condition x+2", ... is the input data, Shape data of a processed substrate when the substrate processing apparatus B131 to the substrate processing apparatus D133 process the substrate under the processing conditions = "condition x+1", "condition x+2", ... Using this as correct answer data, a re-learning process can be performed on the trained shape simulation model 640.

[0127] (5) Specific examples of virtual measurement processing (step S171) and shape simulation processing (step S172) Next, a specific example of the virtual measurement process (step S171) and the shape simulation process (step S172) executed by the data collection system 100 will be described. Fig. 10 is a diagram showing a specific example of the virtual measurement process and the shape simulation process executed by the data collection system according to the first embodiment.

[0128] The example of Fig. 10 shows how, when a substrate is processed under processing conditions = "conditions x", observation data = "observation data M_αx" is output. The example of Fig. 10 also shows how, by inputting various observation data into the trained virtual measurement model 620, plasma measurement data, wear amount measurement data, and particle measurement data are predicted, and measurement data = "measurement data I_αx" is output.

[0129] In this way, by outputting "measurement data I_αx" including plasma measurement data, wear measurement data, and particle measurement data, workers and the like can monitor measurement data that they were previously unable to monitor.

[0130] 10 shows a state in which processing conditions="condition x", observation data="observation data M_αx", measurement data="measurement data I_αx", and pre-processing substrate shape data="shape data S_αx" are input to the trained shape simulation model 640. Furthermore, the example of Fig. 10 shows a state in which post-processing substrate shape data="shape data S_αx'" is output from the trained shape simulation model 640.

[0131] In this way, by predicting the post-processing substrate shape data, an operator or the like can determine the quality of the post-processing substrate without inspecting every single one of the post-processing substrates.

[0132] <Hardware configuration of data collection device and data analysis device> Next, a description will be given of the hardware configuration of the data collecting device 140 and the data analyzing device 150. Fig. 11 is a diagram showing an example of the hardware configuration of the data collecting device and the data analyzing device.

[0133] (1) Data collection device 11, the data collecting device 140 includes a processor 1101, a memory 1102, an auxiliary storage device 1103, an I / F (Interface) device 1104, a communication device 1105, and a drive device 1106. The hardware components of the data collecting device 140 are connected to each other via a bus 1107.

[0134] The processor 1101 has various arithmetic devices such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), etc. The processor 1101 reads various programs (for example, a data collection program, etc.) onto the memory 1102 and executes them.

[0135] The memory 1102 has a main storage device such as a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The processor 1101 and the memory 1102 form a so-called computer, and the processor 1101 executes various programs read onto the memory 1102, causing the computer to realize the various functions described above.

[0136] The auxiliary storage device 1103 stores various programs and various data used when the various programs are executed by the processor 1101. The collected data storage unit 250 described above is realized in the auxiliary storage device 1103.

[0137] The I / F unit 1104 is a connection device that connects the display unit 1108 , the operation unit 1109 , and the data collection unit 140 .

[0138] The communication apparatus 1105 is a communication device for communicating with the substrate processing apparatus 110, the substrate processing apparatus A 120, the substrate processing apparatus B 131 to the substrate processing apparatus D 133, the shape measuring apparatus 1110, the data analyzing apparatus 150, and the like via a network.

[0139] The drive device 1106 is a device for setting the recording medium 1111. The recording medium 1111 here includes media that record information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, a magneto-optical disk, etc. The recording medium 1111 may also include semiconductor memories that record information electrically, such as ROM, flash memory, etc.

[0140] The various programs to be installed in the auxiliary storage device 1103 are installed, for example, by setting the distributed recording medium 1111 in the drive device 1106 and reading out the various programs recorded in the recording medium 1111. Alternatively, the various programs to be installed in the auxiliary storage device 1103 may be installed by being downloaded from a network via the communication device 1105.

[0141] (2) Data analysis equipment As shown in 11b of FIG. 11, the hardware configuration of the data analysis device 150 is similar to that of the data collection device 140, so only the differences from the data collection device 140 will be described here.

[0142] In the case of the data analysis device 150, the processor 1121 reads out a data analysis program onto the memory 1122 and executes it.

[0143] In the case of the data analysis device 150 , the communication device 1125 communicates with the substrate processing device 110 and the data collection device 140 .

[0144] <Summary> As is clear from the above description, the collection system according to the first embodiment has The system includes a substrate processing apparatus A having a first chamber, substrate processing apparatuses B to D having second chambers different from the first chamber, and a data collecting apparatus 140 connected to the substrate processing apparatus A and the substrate processing apparatuses B to D. Various observation data observed when substrates of the same or similar shape are processed in the first chamber and the second chamber under the same conditions are compared, and a correction amount for correcting the various observation data observed when processed in the second chamber is calculated. When searching for optimal processing conditions by processing substrates in the second chamber while changing the processing conditions, the observation data observed during processing in the second chamber is corrected based on the correction amount, and the corrected observation data is collected.

[0145] In this way, the collection system according to the first embodiment uses a plurality of substrate processing apparatuses to process substrates of the same or similar shapes under the same conditions, and collects observation data in which the effects of disturbances have been corrected. As a result, according to the first embodiment, it is possible to collect various types of appropriate observation data (in which the effects of disturbances have been eliminated) in searching for processing conditions in substrate processing.

[0146] In addition, the collection system according to the first embodiment includes The first chamber is configured to be identical to a reference chamber of a reference substrate processing apparatus. Substrates of the same or similar shape are processed under the same conditions in the reference chamber and the first chamber, and from various observation data observed, a calibration amount based on an instrumental difference is calculated.

[0147] As a result, according to the first embodiment, various appropriate observation data (with the effects of machine differences eliminated) can be collected in searching for processing conditions in substrate processing.

[0148] In addition, the collection system according to the first embodiment includes A plasma probe, a wear amount sensor, and a particle sensor are additionally installed in each second chamber of substrate processing apparatuses B to D, and various measurement data (plasma measurement data, wear amount measurement data, particle measurement data) during substrate processing are collected and displayed.

[0149] As a result, according to the first embodiment, it becomes possible to change the processing conditions while referring to various measurement data, which are detailed data that represent the conditions inside the chamber, and thus it is possible to efficiently search for optimal processing conditions.

[0150] In addition, the collection system according to the first embodiment includes Generate trained virtual metrology models to predict various measurement data during substrate processing. - Learning processing is performed including various measurement data during substrate processing, and a learned shape simulation model is generated.

[0151] This makes it possible to improve the prediction accuracy of the trained shape simulation model. In addition, it becomes possible to narrow down the processing conditions using the trained shape simulation model, and it becomes possible to efficiently search for optimal processing conditions.

[0152] In addition, the collection system according to the first embodiment includes The generated trained virtual measurement model and trained shape simulation model are installed in the substrate processing equipment of a mass production machine.

[0153] This makes it possible to monitor various measurement data that could not be monitored in the past when processing substrates in a mass production substrate processing apparatus. Also, it becomes possible to predict and display shape data of the processed substrate, making it possible to determine the quality of the processed substrate without inspecting all the processed substrates.

[0154] [Second embodiment] In the data collection system 100 according to the first embodiment, a substrate processing apparatus A (mass production machine) identical to the substrate processing apparatus (reference substrate processing apparatus) installed at the substrate processing apparatus manufacturer is installed at the substrate processing apparatus manufacturer. However, the configuration of the data collection system is not limited to this, and for example, the data collection system may be configured without installing the substrate processing apparatus A at the substrate processing apparatus manufacturer. Hereinafter, a data collection system according to a second embodiment will be described.

[0155] <Overview of data collection system configuration and processing> First, a configuration of a data collection system according to the second embodiment and an overview of processing executed by the data collection system according to the second embodiment will be described. Fig. 12 is a diagram showing the configuration of the data collection system according to the second embodiment and an overview of processing.

[0156] As shown in FIG. 12, the data collection system 1200 includes: A substrate processing apparatus A1210, which is an example of a first substrate processing apparatus; Substrate processing apparatus B131, substrate processing apparatus C132, and substrate processing apparatus D133, which are examples of the second substrate processing apparatus; data collection device 140; Data analysis device 150, In this embodiment, for convenience of explanation, the system is configured such that the substrate processing apparatus A 1210 is installed in a substrate manufacturer instead of the substrate processing apparatus 110.

[0157] The data collection system 1200, like the first embodiment, is applied, for example, in a situation where a substrate processing apparatus manufacturer supports a substrate manufacturer so that the substrate manufacturer can mass-produce substrates having a target substrate shape.

[0158] The substrate processing apparatus A1210 is installed in, for example, a substrate manufacturer (mass production machine). The substrate processing apparatus A1210 has a first chamber and processes substrates under predetermined processing conditions. The substrate processing apparatus A1210 is equipped with various observation sensors and outputs observation data observed by the various observation sensors during substrate processing.

[0159] 12. The substrate processing apparatus B 131 to substrate processing apparatus D 133, the data collecting apparatus 140, and the data analyzing apparatus 150 shown in FIG. 12 are similar to those in FIG. 1, and therefore will not be described here.

[0160] Moreover, each process included in the "experiment phase for searching for optimal conditions" is the same as that in FIG. 1 except that the calibration amount calculation process (step S161) is not included, and therefore a description thereof will be omitted here.

[0161] Furthermore, each process included in the "processing phase under optimal conditions" is similar to that in FIG. 1 except that it is performed by the substrate processing apparatus A1210, and therefore a description thereof will be omitted here.

[0162] <Examples of processing performed by the data collection system> Next, a specific example of the correction amount calculation process (step S162) among the processes executed by the data collection system 1200 will be described.

[0163] (1) Specific example of correction amount calculation process (step S162) FIG. 13 is a diagram showing a specific example of the correction amount calculation process executed by the data collection system according to the second embodiment.

[0164] The example of Fig. 13 shows how the substrate processing apparatus A120 and the substrate processing apparatuses B131-D133 process unprocessed substrates of the same or similar shapes under processing condition "Condition 1." Furthermore, the example of Fig. 13 shows how observation data "observation data M_A1" to "observation data M_D1" are output from the substrate processing apparatus A120 and the substrate processing apparatuses B131-D133, respectively.

[0165] The difference from Figure 3 is that in the case of Figure 13, the correction amount calculation unit 220 calculates correction amounts b', c', and d' (correction amounts to eliminate the effects of both instrumental error and disturbance) as correction amounts to match the acquired observation data = "observation data M_A1".

[0166] <Summary> As is clear from the above description, the data collection system according to the second embodiment includes: A data collection system is configured without installing a substrate processing apparatus A, which is the same as a substrate processing apparatus (reference substrate processing apparatus) installed at a substrate manufacturer, at the substrate processing apparatus manufacturer. Various observation data observed when substrates of the same or similar shape are processed in the first chamber and the second chamber under the same conditions are compared, and a correction amount for correcting the various observation data observed when processed in the second chamber is calculated. When searching for optimal processing conditions by processing substrates in the second chamber while changing the processing conditions, the observation data observed during processing in the second chamber is corrected based on the correction amount, and the corrected observation data is collected.

[0167] As a result, according to the second embodiment, various appropriate observation data (with the effects of machine differences and disturbances eliminated) can be collected in searching for processing conditions in substrate processing.

[0168] [Third embodiment] In the above first and second embodiments, the correction amount calculation process (steps S162, S167) has been described as calculating the correction amount for various observation data. In contrast, in the third embodiment, the correction amount calculation process also calculates the correction amount for various measurement data (specifically, particle measurement data). The third embodiment will be described below, focusing on the differences from the above first and second embodiments.

[0169] Fig. 14 is a diagram showing a specific example of the correction amount calculation process executed by the data collection system according to the third embodiment. The example in Fig. 14 shows how the substrate processing apparatus A1210 processes a pre-processing QC substrate under processing condition = "condition 0", thereby obtaining a post-processing QC substrate, and how particle measurement data = "particle measurement data A0" is measured.

[0170] The example of FIG. 14 shows that the substrate processing apparatus D133 processes an unprocessed substrate and outputs particle measurement data "particle measurement data D0".

[0171] In the data collection system according to the third embodiment, in the correction amount calculation process (steps S162 and S167), the correction amount calculation unit 1410 operates. At that time, the correction amount calculation unit 1410 acquires the particle measurement data output from the substrate processing apparatus A 1210 and the substrate processing apparatus D 133, that is, “particle measurement data A0” and “particle measurement data D0”.

[0172] Furthermore, the correction amount calculation unit 1410 calculates a correction amount p for making the acquired particle measurement data = "particle measurement data D0" coincide with the particle measurement data = "particle measurement data A0", and notifies the collection unit 230. Thereafter, the data collection device 140 can correct the particle measurement data output from the substrate processing device D133 using the correction amount p, and store the corrected data in the collected data storage unit 250.

[0173] As a result, according to the third embodiment, appropriate measurement data (with the effects of machine differences and disturbances eliminated) can be collected in searching for processing conditions in substrate processing.

[0174] [Fourth embodiment] In the above first to third embodiments, the condition narrowing down process (step S168) is described as being able to search for optimal processing conditions that realize a target substrate shape. However, there may be cases where the data analysis device 150 is unable to search for optimal processing conditions.

[0175] In such a case, a determination result may be output to the effect that the target substrate shape cannot be achieved simply by changing the current processing conditions.

[0176] Furthermore, the apparatus may be configured to propose changes to processing conditions other than the current processing conditions, including, for example, pulse modulation, impedance control, and the like.

[0177] [Other embodiments] Although the first to fourth embodiments do not mention specific examples of substrate processing by the substrate processing apparatus, the substrate processing by the substrate processing apparatus may include, for example, a film forming process, an etching process, etc. Although the first to fourth embodiments do not mention specific examples of the substrate processing apparatus, the substrate processing apparatus may include an etching apparatus, a film forming apparatus, an ashing apparatus, an annealing apparatus, a doping apparatus, etc.

[0178] In the above first to fourth embodiments, data related to the appearance of the substrate, such as critical dimensions, depth, taper angle, tilt angle, bowing, etc., are exemplified as the substrate shape data, but the substrate shape data is not limited to data related to the appearance of the substrate. For example, the substrate shape data may include data other than the data related to the appearance of the substrate, such as film thickness, film type, film characteristics, etc.

[0179] Furthermore, the "same processing conditions" in the first to fourth embodiments are not limited to the case where the processing conditions are completely the same, and may include, for example, processing conditions that provide the same effect in terms of changes in substrate shape data before and after processing. Note that the "same effect" mentioned here does not necessarily mean that the changes in substrate shape data before and after processing are completely the same, but refers to changes in substrate shape data that are of the same order of magnitude (within a predetermined range).

[0180] Furthermore, in the first to fourth embodiments, the term "substrates having the same or similar shape" includes those having substrate shape data that is approximately the same (within a predetermined range).

[0181] In addition, in the first embodiment, the various measurement data include plasma measurement data, wear amount measurement data, and particle measurement data, but are not limited to these, and may include other measurement data.

[0182] In the first embodiment, the various measurement data are described as being measured by a sensor installed in the substrate processing apparatus, but the various measurement data may be acquired from outside the substrate processing apparatus. That is, the other measurement data may include, for example, component analysis data by XPS (X-ray Photoelectron Spectroscopy). Alternatively, the other measurement data may include characteristic data obtained by a process performed by another process processing apparatus.

[0183] Furthermore, in the first to fourth embodiments, the data collecting device 140 and the data analyzing device 150 are configured as separate entities, but the data collecting device 140 and the data analyzing device 150 may be configured as an integrated device.

[0184] In the first to fourth embodiments, the data collecting device 140 and the data analyzing device 150 are configured as separate entities from the substrate processing apparatus 110, the substrate processing apparatus A 120, or the substrate processing apparatus B 131 to D 133. However, the data collecting device 140 or the data analyzing device 150 may be provided inside the substrate processing apparatus 110, the substrate processing apparatus A 120, or the substrate processing apparatus B 131 to D 133.

[0185] In the above first to fourth embodiments, the data collection device 140 and the data analysis device 150 have been described as executing a data collection program and a data analysis program, respectively, by themselves. However, the data collection device 140 and the data analysis device 150 may each be configured, for example, with a plurality of computers, and the data collection program or the data analysis program may be installed in each of them, so that the data collection program or the data analysis program may be executed in a distributed computing form.

[0186] In the above first to fourth embodiments, a method of downloading and installing the data collection program to the auxiliary storage device 1103 of the data collection device 140 has been mentioned as an example of a method of installing the data analysis program to the auxiliary storage device 1123 of the data analysis device 150. In this case, no particular reference was made to the download source, but in the case of installing by such a method, the download source may be, for example, a server device that stores the data collection program or the data analysis program in an accessible manner. In addition, the server device may be a device on a cloud that accepts access from the data collection device 140 or the data analysis device 150 via a network and downloads the data collection program or the data analysis program on the condition of a charge. In other words, the server device may be a device on a cloud that provides a service of providing the data collection program or the data analysis program.

[0187] In the first to fourth embodiments, the details of the virtual measurement model are not mentioned, but the virtual measurement model used in the first to fourth embodiments may be, for example, an ARX model. In the first to fourth embodiments, the details of the shape simulation model are not mentioned, but the shape simulation model used in the first to fourth embodiments may be, for example, a model based on a convolutional neural network. However, the various models used in the first to fourth embodiments are not limited to these examples, and may be other machine learning learning models including deep learning, statistical models, or models that combine these models.

[0188] The present invention is not limited to the configurations shown in the above embodiments, combinations with other elements, etc. These points can be changed without departing from the spirit of the present invention, and can be appropriately determined according to the application form.

[0189] This application claims priority based on Japanese Patent Application No. 2021-032890, filed on March 2, 2021, the entire contents of which are incorporated herein by reference. [Explanation of symbols]

[0190] 100: Data collection system 110: Substrate processing apparatus 120: Substrate processing apparatus A 131: Substrate processing device B 132: Substrate processing equipment C 133: Substrate processing device D 140: Data collection device 150: Data analysis equipment 210: Calibration amount calculation section 220: Correction amount calculation section 230: Collection Department 240: Condition change section 500: Collected data 610: Virtual measurement model learning unit 620: Trained Virtual Measurement Model 630: Shape simulation model learning section 640: Learned shape simulation model 650: Shape simulation control unit 1200: Data Acquisition System 1210: Substrate processing equipment A

Claims

1. 1. A data acquisition apparatus in communication with a first substrate processing apparatus having a first processing space and a second substrate processing apparatus having a second processing space different from the first processing space, the data acquisition apparatus comprising: a correction amount calculation unit that compares observed data obtained by processing a substrate having the same or a similar shape in the first processing space and the second processing space under the same processing conditions, and calculates a correction amount for correcting the observed data obtained by processing the substrate in the second processing space; a collection unit that, when searching for processing conditions by processing a substrate under different processing conditions in the second processing space, corrects observation data observed when the substrate is processed in the second processing space based on the correction amount, and collects the corrected observation data as observation data observed when the substrate is processed in the first processing space; A data collection device having:

2. the first processing space is identical to a reference processing space; The data collection device includes: a calibration amount calculation unit that calculates a calibration amount based on an instrumental difference from observation data obtained by processing substrates having the same or a similar shape in the reference processing space and the first processing space under the same processing conditions, 2. The data collection device according to claim 1, wherein the observation data observed by processing in the first processing space and used for comparison when the correction amount calculation unit calculates the correction amount is observation data calibrated based on the calibration amount.

3. a sensor that is not installed in the first substrate processing apparatus is additionally installed in the second substrate processing apparatus; The collecting unit includes: The data collection device according to claim 1 , wherein measurement data measured by the additionally attached sensor is collected when processing substrates under different processing conditions in the second processing space to search for processing conditions.

4. The second substrate processing apparatus includes: a substrate processing apparatus further comprising a sensor for measuring plasma in the second processing space; a substrate processing apparatus further comprising a sensor for measuring a wear amount of a part in the second processing space; a substrate processing apparatus further comprising a sensor for measuring particles in the second processing space; The data collection device according to claim 1 , comprising at least one of the following:

5. The data collection device according to claim 3 , wherein measurement data measured by the additionally attached sensor is displayed when processing substrates under different processing conditions in the second processing space to search for processing conditions.

6. 2. The data collection device according to claim 1, further comprising a virtual measurement model learning unit that learns a virtual measurement model using the corrected observation data collected by the collection unit as input data and the measurement data collected by the collection unit as corrective data.

7. the first substrate processing apparatus has a trained virtual measurement model trained by the virtual measurement model training unit, The data collection device according to claim 6 , which displays predicted measurement data by inputting observation data observed by processing in the first processing space into the trained virtual measurement model.

8. 7. The data collection device according to claim 6, further comprising a shape simulation model learning unit that learns a shape simulation model using the corrected observation data collected by the collection unit, the measurement data collected by the collection unit, and processing conditions for processing a substrate in the second processing space as input data, and shape data of a processed substrate obtained by processing a substrate under different processing conditions in the second processing space as corrective data.

9. a trained shape simulation model trained by the shape simulation model training unit; and a control unit that narrows down the processing conditions so that shape data of a processed substrate predicted by the trained shape simulation model approaches target shape data when searching for processing conditions by changing the processing conditions in the second processing space. The data collection device of claim 8 further comprising:

10. The data collection device according to claim 9, further comprising a control unit that narrows down processing conditions for collecting corrected observation data and measurement data used to re-learn the trained shape simulation model based on the prediction accuracy of the shape data of the processed substrate predicted by the trained shape simulation model.

11. The data collection device of claim 9, further comprising a judgment unit that, when searching for processing conditions by processing a substrate under different processing conditions in the second processing space, determines whether it is possible to search for processing conditions that result in a difference from a target shape data being less than a predetermined threshold by predicting shape data of the processed substrate using the learned shape simulation model.

12. the first substrate processing apparatus has a trained shape simulation model trained by the shape simulation model training unit, A data collection device as described in claim 8, which displays predicted shape data of the processed substrate by inputting observation data observed during processing in the first processing space, measurement data predicted by inputting the observation data into the trained virtual measurement model, and processing conditions for processing the substrate in the first processing space into the trained shape simulation model.

13. 1. A method of collecting data for a data collection apparatus in communication with a first substrate processing apparatus having a first processing space and a second substrate processing apparatus having a second processing space different from the first processing space, the method comprising: a step of comparing observed data obtained by processing a substrate having the same or a similar shape in the first processing space and the second processing space under the same processing conditions, and calculating a correction amount for correcting the observed data obtained by processing the substrate in the second processing space; when searching for processing conditions by processing a substrate under different processing conditions in the second processing space, correcting observation data observed when the substrate is processed in the second processing space based on the correction amount, and collecting the corrected observation data as observation data observed when the substrate is processed in the first processing space; A data collection method having the following features:

14. a computer of a data collection apparatus in communication with a first substrate processing apparatus having a first processing space and a second substrate processing apparatus having a second processing space different from the first processing space; a step of comparing observed data obtained by processing a substrate having the same or a similar shape in the first processing space and the second processing space under the same processing conditions, and calculating a correction amount for correcting the observed data obtained by processing the substrate in the second processing space; when searching for processing conditions by processing a substrate under different processing conditions in the second processing space, correcting observation data observed when the substrate is processed in the second processing space based on the correction amount, and collecting the corrected observation data as observation data observed when the substrate is processed in the first processing space; A data collection program having:

15. A method for processing a substrate using a mass production machine having a first processing space, comprising: A method for manufacturing a semiconductor device comprising: a step of processing a substrate having the same or similar shape as a substrate used in searching for processing conditions in the first processing space under processing conditions searched for using an experimental machine having a second processing space; when performing processing under the searched processing conditions, the mass production machine is capable of acquiring the searched processing conditions, shape data of a processed substrate processed in the second processing space under the searched processing conditions, and corrected observation data observed by processing in the second processing space under the searched processing conditions, the corrected observation data is corrected to be the observation data observed when a substrate having the same or a similar shape is processed in the first processing space by a correction amount calculated based on the observation data observed when the substrate is processed in the first processing space and the second processing space under the same processing conditions. A method for processing a substrate.

16. A trained virtual measurement model, which is trained using observation data observed as a substrate is processed in a mass production machine as input data and measurement data that is not observed when the substrate is processed in the mass production machine and is measured by a sensor additionally attached to the mass production machine when the substrate is processed as correct answer data; a trained shape simulation model which is trained using corrected observation data obtained by correcting observation data observed when a substrate is processed in a mass production machine to which the sensor is additionally attached, measurement data measured by the sensor when the substrate is processed, processing conditions when the substrate is processed, and shape data of the substrate before processing as input data, and using shape data of the substrate after processing as corrected data; A computer of the substrate processing apparatus equipped with the A step of displaying measurement data predicted by the trained virtual measurement model by inputting observation data observed during processing of a substrate; a step of displaying shape data of the substrate after processing predicted by the trained shape simulation model by inputting observation data observed when the substrate is processed, measurement data predicted by the trained virtual measurement model, processing conditions when the substrate is processed, and shape data of the substrate before processing; A substrate processing method comprising the steps of:

Images