Analytical device, analytical method and analytical program product

Abstract

The present invention provides an analysis device, an analysis method and an analysis program for quantitatively evaluating the environment of a processing space of a manufacturing process based on a time series data set. The analysis device includes: a learning unit that uses a time series data set measured when an object is processed in the processing space to perform machine learning and calculates a value representing the correlation of time series data in a corresponding time range between each measurement item; and an evaluation unit that uses a time series data set measured when an object is processed in a known environment in the processing space to evaluate the unknown environment of the processing space based on the value representing the correlation calculated by the learning unit through machine learning.

Description

Technical Field

[0001] The present invention relates to an analysis device, an analysis method and an analysis program. Background Art

[0002] Generally speaking, when the environment (condition) in the processing space of the manufacturing process changes, it will affect the quality of the product when the object is processed in the processing space. Therefore, it is particularly important to know the environment of the processing space in advance when processing the object in order to maintain the quality of the product.

[0003] On the other hand, in the manufacturing process, various data sets (data sets of multiple time series data, hereinafter referred to as time series data sets) are acquired along with the processing of the object. In addition, the acquired time series data sets also include time series data related to the environment of the processing space.

[0004] Prior art literature

[0005] Patent Literature

[0006] Patent Document 1: Japanese Patent Application Publication No. 2010-219263 Summary of the invention

[0007] Technical problem to be solved by the invention

[0008] The present invention provides an analysis device, an analysis method and an analysis program for quantitatively evaluating the environment of a processing space in a manufacturing process based on a time series data group.

[0009] Technical solutions for solving technical problems

[0010] An analysis device according to one embodiment of the present invention has, for example, the following configuration. That is, it includes:

[0011] a learning unit that performs machine learning using a time series data group measured in conjunction with processing of an object in a processing space, and calculates a value representing a correlation between the time series data in a corresponding time range between each measurement item; and

[0012] An evaluation unit that uses a time series data group measured by processing an object in a known environment of a processing space to evaluate an unknown environment of the processing space based on a value indicating the above correlation calculated by machine learning by the above learning unit.

[0013] Effects of the Invention

[0014] According to the present invention, it is possible to provide an analysis device, an analysis method, and an analysis program for quantitatively evaluating the environment of a processing space in a manufacturing process based on a time series data group. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 FIG. 1 is a first diagram showing an example of a system configuration of an environmental adjustment system.

[0016] Figure 2 This is a diagram showing an example of a semiconductor manufacturing process.

[0017] Figure 3 This is a diagram showing an example of the hardware configuration of the analysis device.

[0018] Figure 4 This is a diagram showing an example of learning data.

[0019] Figure 5 This is a diagram showing an example of a time series data group.

[0020] Figure 6 FIG. 1 is a first diagram showing a specific example of processing performed by the learning unit.

[0021] Figure 7 FIG. 2 is a specific example of the processing performed by the learning unit.

[0022] Figure 8 FIG. 1 is a first diagram showing a specific example of processing performed by the evaluation unit.

[0023] Fig. 9 FIG. 2 is a second diagram showing a specific example of the processing performed by the evaluation unit.

[0024] Fig. 10A This is the first flowchart showing the flow of the environment adjustment process.

[0025] Fig. 10B This is the second flow chart showing the flow of the environment adjustment process.

[0026] Fig.11 This is a diagram showing an example of the system configuration of an environment adjustment system when using OES data.

[0027] Fig.12 This is a diagram showing an example of OES data.

[0028] Fig.13 It is a diagram showing a specific example of the processing performed by the evaluation unit when using OES data.

[0029] Fig.14 This is a diagram showing the relationship between the total value of the intensity values ​​indicating the correlation between the wavelengths of OES data and the emission intensity of each wavelength.

[0030] Fig.15 This is a diagram showing a specific example of an environmental adjustment method when evaluating the environment using OES data.

[0031] Fig.16This is a diagram showing an example of the system configuration of the environment adjustment system when the process data group is used.

[0032] Fig.17 This is a diagram showing an example of a process data group.

[0033] Fig.18 This is a diagram showing a specific example of processing performed by the evaluation unit when process data is used.

[0034] Fig.19 This is a diagram showing a specific example of an environment adjustment method when evaluating the environment using process data.

[0035] Fig. 20 FIG. 2 is a second diagram showing an example of a system configuration of an environmental adjustment system.

[0036] Fig.21 FIG. 3 is a specific example of the processing performed by the learning unit.

[0037] Fig. 22 FIG3 is a third diagram showing a specific example of the processing performed by the evaluation unit.

[0038] Fig.23 This is the third flow chart showing the flow of the environment adjustment process.

[0039] Fig.24 This is a diagram showing an example of the system configuration of an end point detection system.

[0040] Fig.25 FIG. 4 is a specific example of the processing performed by the learning unit.

[0041] Fig.26 It is a diagram showing a specific example of the processing performed by the end point detection unit.

[0042] Fig. 27 : is a flowchart showing the flow of the end point detection process.

[0043] Description of Reference Numerals

[0044] 100, 100', 100": Environmental adjustment system

[0045] 110: Wafer before processing

[0046] 120: Processing space

[0047] 130: Wafer after processing

[0048] 140_1~140_n: Time series data acquisition device

[0049] 160: Analytical device

[0050] 161: Learning Department

[0051] 162: Evaluation Department

[0052] 170: Control device

[0053] 610: Regression model generation department

[0054] 620: Regression Model

[0055] 621-623: Node

[0056] 801: Regression model generation department

[0057] 802: Similarity calculation unit

[0058] 1140: Emission spectroscopy device

[0059] 1500: Environmental adjustment parameter determination table

[0060] 1640_1~1640_n:Process data acquisition device

[0061] 1900: Environmental Adjustment Parameters Determination Table

[0062] 2000: Environmental Adjustment System

[0063] 2010: Analytical Devices

[0064] 2011: Learning Department

[0065] 2012: Evaluation Department

[0066] 2101: Regression model generation department

[0067] 2210: Regression Model Execution Department

[0068] 2220: Count value calculation unit

[0069] 2400: End point detection system

[0070] 2410: Analytical Device

[0071] 2411: Learning Department

[0072] 2412: End point detection unit

[0073] 2420: Control Device

[0074] 2501: Regression model generation department

[0075] 2610: Regression Model Execution

[0076] 2620: Count value calculation unit. DETAILED DESCRIPTION

[0077] Hereinafter, each embodiment will be described with reference to the drawings. In addition, in this specification and the drawings, components having substantially the same functional structure are denoted by the same reference numerals to omit repeated descriptions.

[0078] [First embodiment]

[0079] <System Configuration of Environmental Adjustment System>

[0080] First, the system configuration of the environment adjustment system will be described. Figure 1 FIG. 1 is a first diagram showing an example of a system configuration of an environmental adjustment system. Figure 1 As shown, the environment adjustment system 100 includes a semiconductor manufacturing process as an example of a manufacturing process, time series data acquisition devices 140_1 to 140_n, an analysis device 160 and a control device 170.

[0081] In the semiconductor manufacturing process, an object (pre-processed wafer 110) is processed in a predetermined processing space 120 to generate a product (processed wafer 130). The pre-processed wafer 110 mentioned here refers to a wafer (substrate) before being processed in the processing space 120, and the processed wafer 130 refers to a wafer (substrate) after being processed in the processing space 120.

[0082] The timing data acquisition devices 140_1 to 140_n measure timing data respectively as the pre-processed wafer 110 is processed in the processing space 120. The timing data acquisition devices 140_1 to 140_n are devices that measure different types of measurement items. In addition, the number of measurement items measured by each timing data acquisition device 140_1 to 140_n can be one or more.

[0083] The time series data group measured by the time series data acquisition devices 140_1 to 140_n is stored as learning data in the learning data storage unit 163 of the analysis device 160.

[0084] An analysis program is installed in the analysis device 160 , and by executing the program, the analysis device 160 functions as a learning unit 161 and an evaluation unit 162 .

[0085] The learning unit 161 uses the data measured by the time series data acquisition devices 140_1 to 140_n.

[0086] The processing space 120 is a normal environment, and

[0087] Use a standard plan (a specific plan that is pre-set)

[0088] Machine learning is performed on a time series data group (first learning data) measured when the pre-processed wafer 110 is processed. Thus, the learning unit 161 generates “first evaluation data” for quantitatively evaluating the environment of the processing space 120 .

[0089] In addition, the learning unit 161 performs machine learning using each time series data group when the pre-processed wafer 110 is processed in a plurality of known environments (all of which are in a normal environment) in the processing space 120 to generate first evaluation data. In addition, the learning unit 161 stores each generated first evaluation data in the evaluation data storage unit 164 as information indicating the corresponding environment.

[0090] The evaluation unit 162 uses the data measured by the time series data acquisition devices 140_1 to 140_n.

[0091] The processing space 120 is an unknown environment, and

[0092] Use standard protocols

[0093] Machine learning is performed on a time series data group (second learning data) measured when the unprocessed wafer 110 is processed, thereby generating “second evaluation data”.

[0094] Furthermore, the evaluation unit 162 compares the second evaluation data with each first evaluation data stored in the evaluation data storage unit 164, and determines which first evaluation data the second evaluation data is similar to. Thus, the evaluation unit 162 evaluates the unknown environment of the processing space 120. Then, the evaluation unit 162 notifies the control device 170 of the evaluated environment.

[0095] The control device 170 adjusts the environment of the processing space 120 based on the environment evaluated by the evaluation unit 162 of the analysis device 160 .

[0096] <Processing Space in Semiconductor Manufacturing Process>

[0097] Next, a predetermined processing space 120 in the semiconductor manufacturing process will be described. Figure 2 is a diagram showing an example of a semiconductor manufacturing process. Figure 2 As shown, the semiconductor manufacturing process 200 has a plurality of chambers as an example of a processing space. Figure 2 In the example of FIG. 2 , the semiconductor manufacturing process 200 includes chambers 121 (name = “chamber A”) to 123 (name = “chamber C”), and a pre-processed wafer is processed in each chamber.

[0098] In the semiconductor manufacturing process 200, each chamber has the timing data acquisition device 140_1 to 140_n, and the timing data group is measured in each chamber. Therefore, for example, the first evaluation data generated using the timing data group measured in chamber A and the second evaluation data generated using the timing data group measured in chamber B can be compared to evaluate the environment of chamber B.

[0099] However, in order to simplify the description, the following description will be given of a case where the environment is evaluated using the first and second evaluation data generated for the same chamber. In addition, the following description will be given of the chamber A as the chamber to be evaluated for the environment.

[0100] <Hardware Configuration of Analyzer>

[0101] Next, the hardware configuration of the analyzing device 160 will be described. Figure 3 FIG. 1 is a diagram showing an example of the hardware configuration of an analysis device. Figure 3 As shown, the analysis device 160 has a CPU (Central Processing Unit) 301, a ROM (Read Only Memory) 302, and a RAM (Random Access Memory) 303. In addition, the analysis device 160 has a GPU (Graphics Processing Unit) 304. In addition, the processors (processing circuits, processing circuits, processing circuitry) such as the CPU 301 and GPU 304 and the memories such as the ROM 302 and RAM 303 form a so-called computer.

[0102] The analyzing device 160 further includes an auxiliary storage device 305, a display device 306, an operation device 307, an I / F (Interface) device 308, and a drive device 309. In addition, the hardware components of the analyzing device 160 are connected to each other via a bus 310.

[0103] The CPU 301 is a computing device that executes various programs (for example, analysis programs, etc.) installed in the auxiliary storage device 305 .

[0104] ROM 302 is a nonvolatile memory that functions as a main storage device. ROM 302 stores various programs and data required for CPU 301 to execute various programs installed in auxiliary storage device 305. Specifically, ROM 302 stores boot programs such as BIOS (Basic Input / Output System) and EFI (Extensible Firmware Interface).

[0105] RAM 303 is a volatile memory such as DRAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory), and functions as a main storage device. RAM 303 provides a work area for CPU 301 to execute various programs installed in auxiliary storage device 305 .

[0106] The GPU 304 is a computing device for image processing. In this embodiment, when the analysis program is executed by the CPU 301, the GPU 304 performs high-speed computing based on parallel processing on the time series data group. In addition, the GPU 304 is equipped with an internal memory (GPU memory) to temporarily store the information required for parallel processing of various time series data groups.

[0107] The auxiliary storage device 305 stores various programs and various data used when the CPU 301 executes the various programs. For example, the learning data storage unit 163 and the evaluation data storage unit 164 are implemented in the auxiliary storage device 305 .

[0108] The display device 306 is a display device for displaying the internal state of the analyzer 160. The operation device 307 is an input device used by the administrator of the analyzer 160 to input various instructions to the analyzer 160. The I / F device 308 is a connection device for communication with a network (not shown).

[0109] The drive device 309 is a device for setting the storage medium 320. The storage medium 320 mentioned here includes a medium for optical storage, electrical storage or magnetic storage information such as a CD-ROM, a floppy disk, a magneto-optical disk, etc. In addition, the storage medium 320 may also include a semiconductor memory for electrical storage information such as a ROM, a flash memory, etc.

[0110] In addition, various programs installed in the auxiliary storage device 305 are installed by, for example, placing the storage medium 320 in which the programs are located in the drive device 309 and reading the various programs stored in the storage medium 320 by the drive device 309. Alternatively, various programs installed in the auxiliary storage device 305 may be installed by downloading via a network (not shown).

[0111] <Specific example of learning data>

[0112] Next, the learning data read out by the learning data storage unit 163 when the learning unit 161 or the evaluation unit 162 performs machine learning will be described. Figure 4 This is a diagram showing an example of learning data.

[0113] like Figure 4 As shown, the learning data includes information items such as “apparatus”, “lot number”, “recipe type”, and “series data group”. The chamber name is stored in “apparatus”, and the lot number of each wafer before processing is stored in “lot number”.

[0114] In addition, the name for specifying the scheme is stored in "Scheme Type". As described above, the learning data is a time series data set when processed using a standard scheme, so "Standard Scheme" is stored in "Scheme Type". In addition, the measured time series data set is stored in "Time Series Data Set".

[0115] in, Figure 4 (a) shows an example of the first learning data. Figure 4 As shown in (a), the first learning data 410_1, 410_2, 410_3, ... include time series data groups measured when the environment of chamber A is environment 1, environment 2, environment 3, ..., respectively.

[0116] on the other hand, Figure 4 (b) represents the second learning data. Figure 4 As shown in (b) of FIG. 4 , the second learning data 420 includes a time series data group measured to evaluate the unknown environment of chamber A.

[0117] <Specific example of time series data group>

[0118] Next, a specific example of the time series data group measured by the time series data acquisition devices 140_1 to 140_n will be described. Figure 5 is a diagram showing an example of a time series data set. Figure 5 In the example, in order to simplify the description, the time series data acquisition devices 140_1 to 140_n are used to measure one-dimensional data respectively, but one time series data acquisition device can also measure two-dimensional data (a data set of multiple one-dimensional data).

[0119] Figure 5 (a) represents a time series data group consisting of time series data measured by the time series data acquisition devices 140_1 to 140_n in the same time range.

[0120] on the other hand, Figure 5 (b) represents a time series data group consisting of time series data measured by the time series data acquisition devices 140_1 to 140_n within a corresponding time range. Figure 5 As shown in (b), the learning data used for machine learning can include not only a time series data group consisting of time series data measured within the same time range, but also a time series data group consisting of time series data measured within corresponding time ranges.

[0121] <Specific example of processing performed by the learning unit>

[0122] Next, a specific example of the processing performed by the learning unit 161 of the analysis device 160 will be described. Figure 6 FIG. 1 is a first diagram showing a specific example of processing performed by the learning unit. Figure 6 As shown, the learning unit 161 includes a regression model generating unit 610 .

[0123] The regression model mentioned here is a machine learning model that captures and extracts the correlation between multiple time series data at high speed. It is a model that uses a linear regression formula or a nonlinear regression formula to represent the correlation between multiple time series data. As an example of a regression model, a cross-correlation model can be cited. In the case of a cross-correlation model, a time delay term that takes into account the time difference between multiple time series data is also included.

[0124] In addition, generally speaking, regression models suitable for manufacturing processes monitor the timing and location of changes in the correlation of time series data to detect anomalies occurring in the manufacturing process.

[0125] On the other hand, in the analysis device 160 of the present embodiment, a regression model is used to quantitatively evaluate the environment of the processing space.

[0126] Specifically, the regression model generation unit 610 performs machine learning on the regression model using the time series data group included in the first learning data stored in the learning data storage unit 163. Thus, the regression model generation unit 610 calculates a value indicating the strength of the correlation between the time series data of the measurement items measured by each time series data acquisition device 140_1 to 140_n (an example of a value indicating the correlation).

[0127] Figure 6In the example of FIG. 6 , for simplicity of explanation, a case is shown where a value representing the strength of the correlation between the time series data within the same time range of the three measurement items is calculated. Specifically, the value is shown by inputting

[0128] The time series data 1 measured by the time series data acquisition device 140_1,

[0129] Time series data 2 measured by the time series data acquisition device 140_2,

[0130] Time series data 3 measured by the time series data acquisition device 140_3

[0131] On the other hand, machine learning is performed on the regression model 620 to calculate a value representing the strength of the correlation of the time series data.

[0132] In the regression model 620 , the node 621 corresponds to the time series data acquisition device 140_1 , the node 622 corresponds to the time series data acquisition device 140_2 , and the node 623 corresponds to the time series data acquisition device 140_3 .

[0133] according to Figure 6 For example, the value representing the strength of the association between time series data 1 and time series data 2 is "F12", and the value representing the strength of the association between time series data 1 and time series data 3 is "F13". In addition, the value representing the strength of the association between time series data 2 and time series data 3 is "F23".

[0134] In addition, in the regression model generation unit 610 , for example, in the chamber A, the first learning data stored for each environment is used to perform the same machine learning on each regression model for each environment.

[0135] Figure 7 FIG. 2 is a second diagram showing a specific example of the processing performed by the learning unit, and shows a situation where the first evaluation data is generated by the regression model generation unit 610. Figure 7 As shown, the first evaluation data includes information items such as “first node” (first measurement item), “second node” (second measurement item), and “strength of correlation”.

[0136] The “first node” and the “second node” store the sequential data used to calculate the value indicating the strength of the association among the sequential data group included in the first learning data.

[0137] The “strength of the relationship” stores a value indicating the strength of the relationship between the sequence data stored in the “first node” and the sequence data stored in the “second node”.

[0138] In addition, if Figure 7As shown in FIG. 1 , the first evaluation data is generated for each environment. Figure 7 In the example of , the first evaluation data 710_1 is evaluation data indicated as environment 1, and the first evaluation data 710_2 is evaluation data indicated as environment 2. Furthermore, the first evaluation data 710_3 is evaluation data indicated as environment 3.

[0139] The first evaluation data 710_1, 710_2, 710_3, ... generated by the regression model generation unit 610 are stored in the evaluation data storage unit 164 as information indicating environments different from each other.

[0140] <Specific Example 1 of Processing by the Evaluation Department>

[0141] Next, a specific example 1 of the processing performed by the evaluation unit 162 of the analysis device 160 will be described. Figure 8 FIG. 1 is a first diagram showing a specific example of processing performed by the evaluation unit. Figure 8 As shown, the evaluation unit 162 includes a regression model generation unit 801 and a similarity calculation unit 802 .

[0142] The regression model generation unit 801 performs machine learning on the regression model using the time series data group included in the second learning data stored in the learning data storage unit 163. Thus, the regression model generation unit 801 generates a regression model and calculates a value representing the strength of correlation between the plurality of time series data measured by each time series data acquisition device 140_1 to 140_n and between each measurement item.

[0143] As a result, the regression model generation unit 801 generates the second evaluation data 820. Figure 8 As shown, the second evaluation data 820 has the same information items as the first evaluation data 710_1, 710_2, 710_3, ...

[0144] The similarity calculation unit 802 calculates the similarity between the second evaluation data 820 generated by the regression model generation unit 801 and the first evaluation data 710_1 , 710_2 , 710_3 . . . stored in the evaluation data storage unit 164 .

[0145] Specifically, the similarity calculation unit 802 calculates the similarity by comparing the values ​​of the strength of the association when all the measurement items of the time series data representing the “first node” and the “second node” are the same between the first evaluation data and the second evaluation data.

[0146] For example, the similarity calculation unit 802 compares:

[0147] The value “f12” indicating the strength of the association when the measurement item of the time series data of the “first node” of the second evaluation data 820 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 2” is equal to

[0148] A value “F12” indicating the strength of association when the measurement item of the time series data of the “first node” of the first evaluation data 710_1 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 2”.

[0149] Similarly, the similarity calculation unit 802 compares:

[0150] The value “f13” indicating the strength of the correlation when the measurement item of the time series data of the “first node” of the second evaluation data 820 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 3” is equal to

[0151] A value “F13” indicating the strength of association when the measurement item of the time series data of the “first node” of the first evaluation data 710_1 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 3”.

[0152] The similarity calculation unit 802 compares the values ​​indicating the strength of the association for all the combinations in the second evaluation data 820 , thereby calculating the similarities with the first evaluation data 710_1 , 710_2 , 710_3 . . . . , respectively.

[0153] Furthermore, the similarity calculation unit 802 evaluates the first evaluation data determined to have the largest calculated similarity as the environment of the chamber A when the time series data group included in the second learning data 420 was measured.

[0154] For example, when the similarity with the first evaluation data 710_1 is the highest, the similarity calculation unit 802 evaluates the environment of the chamber A when the time series data group included in the second learning data 420 is measured as “environment 1”.

[0155] As described above, in the analysis device 160 of the present embodiment, instead of analyzing the features of the time series data individually, a value indicating the strength of the correlation of the time series data is calculated to evaluate the environment.

[0156] Thus, the analysis device 160 of this embodiment can appropriately capture subtle changes in the time series data accompanying changes in the chamber environment. As a result, the analysis device 160 of this embodiment can accurately evaluate the chamber environment based on the time series data group.

[0157] <Specific Example 2 of Processing by the Evaluation Department>

[0158] Next, a specific example 2 of the processing performed by the evaluation unit 162 of the analysis device 160 will be described. Fig. 9 FIG. 2 is a second diagram showing a specific example of the processing performed by the evaluation unit. Fig. 9 In the case of , the similarity calculation unit 802 calculates the similarity after summing up the first evaluation data and the second evaluation data according to the measurement items of each time series data of the first node, and evaluates the environment.

[0159] exist Fig. 9 In FIG. 9 , the graph 910_1 is a graph in which the first evaluation data 710_1 is aggregated for each measurement item of the first node. The horizontal axis represents the measurement item of the time series data of the first node, and the vertical axis represents the total value of the value indicating the strength of the correlation.

[0160] For example, the total value of the values ​​indicating the strength of the association corresponding to "series data 1" of graph 910_1 is

[0161] A value "F12" indicating the strength of association when the measurement item of the time series data of the "first node" of the first evaluation data 710_1 is "time series data 1" and the measurement item of the time series data of the "second node" is "time series data 2", and

[0162] A value “F13” indicating the strength of association when the measurement item of the time series data of the “first node” of the first evaluation data 710_1 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 3”…

[0163] The total value obtained.

[0164] Similarly, in Fig. 9 In the figure, chart 920 is a chart obtained by summing up the measurement items of each time series data of the first node for the second evaluation data, wherein the horizontal axis represents the measurement items of the time series data of the first node, and the vertical axis represents the total value of the values ​​representing the strength of the association (another example of the value representing the association).

[0165] For example, the total value of the values ​​indicating the strength of the association corresponding to "series data 1" of graph 920 is

[0166] A value “f12” indicating the strength of the association when the measurement item of the time series data of the “first node” of the second evaluation data 820 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 2”,

[0167] A value “f13” indicating the strength of the association when the measurement item of the time series data of the “first node” of the second evaluation data 820 is “time series data 1” and the measurement item of the time series data of the “second node” is “time series data 3”…

[0168] The total value obtained.

[0169] Moreover, in Fig. 9 In the case of, the similarity calculation unit 802 calculates the similarity by comparing the graph 920 with the graphs 910_1, 910_2, 910_3, ... In addition, Fig. 9 In the case of , the similarity calculation unit 802 evaluates the environment corresponding to the graph having the largest calculated similarity as the environment of chamber A when the time series data group included in the second learning data 420 is measured.

[0170] For example, when the similarity with the graph 910_1 is the highest, the similarity calculation unit 802 evaluates the environment of the chamber A when the time series data group included in the second learning data 420 is measured as “environment 1”.

[0171] As described above, in the analysis device 160 of the present embodiment, instead of analyzing the features of the time series data individually, the total value of the values ​​indicating the strength of the correlation of the time series data is calculated to evaluate the environment.

[0172] Thus, the analysis device 160 of this embodiment can appropriately capture subtle changes in the time series data accompanying changes in the chamber environment. As a result, the analysis device 160 of this embodiment can accurately evaluate the chamber environment based on the time series data group.

[0173] <Process of environmental adjustment>

[0174] Next, the overall flow of the environmental adjustment process performed by the environmental adjustment system 100 will be described. In the environmental adjustment process performed by the environmental adjustment system 100, as an adjustment method when the control device 170 adjusts the environment of the chamber, any of the following two adjustment methods are included.

[0175] A method of adjusting a recipe in real time based on an environment evaluated in real time during processing of a wafer before processing.

[0176] A method in which the environment is evaluated after the processing of the pre-processed wafer is completed and adjusted to a certain environment through cleaning, etc.

[0177] Therefore, the flow of the environment adjustment process including each adjustment method will be described below.

[0178] (1) Environmental adjustment processing including methods for adjusting solutions in real time

[0179] Fig. 10A The first flowchart shows the flow of the environment adjustment process, and is a flowchart showing the flow of the environment adjustment process including a method of adjusting a plan in real time based on an evaluated environment.

[0180] In step S1001A, the timing data acquisition devices 140_1 to 140_n measure timing data groups accompanying the processing of the pre-processed wafer in the chamber A, and store them in the learning data storage unit 163. In addition, the timing data acquisition devices 140_1 to 140_n store the timing data groups measured accompanying the processing of the chamber A under a plurality of known environments as the first learning data 410_1, 410_2, 410_3, ...

[0181] In step S1002A, the learning unit 161 of the analysis device 160 performs machine learning on each regression model using the first learning data 410_1, 410_2, 410_3, ... stored in the learning data storage unit 163. In addition, the learning unit 161 of the analysis device 160 generates first evaluation data 710_1, 710_2, 710_3, ... using the value indicating the strength of association calculated when performing machine learning on each regression model.

[0182] In step S1003A, the timing data acquisition devices 140_1 to 140_n measure the timing data group (within a predetermined time range) accompanying the processing of the pre-processed wafer in the chamber A, and store it in the learning data storage unit 163. In addition, the timing data acquisition devices 140_1 to 140_n store the timing data group (within a predetermined time range) measured accompanying the processing of the chamber A in an unknown environment as the second learning data 420.

[0183] In step S1004A, the time series data acquisition devices 140_1 to 140_n determine whether a prescribed adjustment period (e.g., 1 second) has passed. If it is determined in step S1004A that the prescribed adjustment period has not passed (if the answer is No in step S1004A), the device waits until the prescribed adjustment period has passed. On the other hand, if it is determined in step S1004A that the prescribed adjustment period has passed (if the answer is Yes in step S1004A), the device proceeds to step S1005A.

[0184] In step S1005A, the evaluation unit 162 of the analysis device 160 performs machine learning on the regression model using the second learning data 420 (data within a predetermined time range before the time point at which the predetermined adjustment period has passed) stored in the learning data storage unit 163. In addition, the evaluation unit 162 of the analysis device 160 generates the second evaluation data 820 (within a predetermined time range) using the value indicating the strength of the association calculated when performing machine learning on the regression model.

[0185] In step S1006A, the evaluation unit 162 of the analysis device 160 determines the first evaluation data having the highest similarity to the second evaluation data 820 among the plurality of first evaluation data 710_1, 710_2, 710_3, ... Or the evaluation unit 162 of the analysis device 160 determines the graph having the highest similarity to the graph 920 calculated based on the second evaluation data 820 among the graphs 910_1, 910_2, 910_3, ... calculated based on the plurality of first evaluation data. Thus, the evaluation unit 162 of the analysis device 160 evaluates the unknown environment of the chamber A.

[0186] In step S1007A, the control device 170 processes the pre-processed wafer using a recipe corresponding to the evaluated environment.

[0187] In step S1008A, it is determined whether the processing of the wafer before processing is completed. If it is determined that it is not completed (if it is No in step S1008A), the process returns to step S1003A. On the other hand, if it is determined that it is completed in step S1008A (if it is Yes in step S1008A), the environmental adjustment process is terminated.

[0188] (2) Environmental adjustment processing including methods for adjusting to a certain environment

[0189] Fig. 10B This is a second flowchart showing the flow of the environment adjustment process, and is a flowchart showing the flow of the environment adjustment process including a method of adjusting the evaluated environment to a certain environment by cleaning or the like.

[0190] In step S1001B, the timing data acquisition devices 140_1 to 140_n measure timing data groups accompanying the processing of the pre-processed wafer in the chamber A, and store them in the learning data storage unit 163. In addition, the timing data acquisition devices 140_1 to 140_n store the timing data groups measured accompanying the processing of the chamber A under a plurality of known environments as the first learning data 410_1, 410_2, 410_3, ...

[0191] In step S1002B, the learning unit 161 of the analysis device 160 performs machine learning on each regression model using the first learning data 410_1, 410_2, 410_3, ... stored in the learning data storage unit 163. In addition, the learning unit 161 of the analysis device 160 generates first evaluation data 710_1, 710_2, 710_3, ... using the values ​​indicating the strength of association calculated when performing machine learning on each regression model.

[0192] In step S1003B, the timing data acquisition devices 140_1 to 140_n measure the timing data group accompanying the processing of the pre-processed wafer in the chamber A, and store it in the learning data storage unit 163. In addition, the timing data acquisition devices 140_1 to 140_n store the timing data group measured accompanying the processing of the chamber A in the unknown environment (the timing data group until the end of the processing) as the second learning data 420.

[0193] In step S1004B, the evaluation unit 162 of the analysis device 160 performs machine learning on the regression model using the second learning data 420 stored in the learning data storage unit 163. In addition, the evaluation unit 162 of the analysis device 160 generates the second evaluation data 820 using the value indicating the strength of association calculated when performing machine learning on the regression model.

[0194] In step S1005B, the evaluation unit 162 of the analysis device 160 determines the first evaluation data having the greatest similarity to the second evaluation data among the plurality of first evaluation data 710_1, 710_2, 710_3, ... Or the evaluation unit 162 of the analysis device 160 determines the graph having the greatest similarity to the graph 920 calculated based on the second evaluation data among the graphs 910_1, 910_2, 910_3, ... calculated based on the plurality of first evaluation data. Thus, the evaluation unit 162 of the analysis device 160 evaluates the unknown environment of the chamber A.

[0195] In step S1006B, control device 170 adjusts the environment to a certain level by cleaning or the like.

[0196] <Conclusion>

[0197] As can be seen from the above description, in the analysis device of the first embodiment,

[0198] Machine learning is performed using a time series data group measured while processing a pre-processed wafer in a chamber, and a value indicating the strength of correlation between the time series data in the corresponding time range between the measurement items is calculated.

[0199] · Using each set of timing data measured by processing a wafer before processing in multiple known environments using an accompanying chamber, evaluate the unknown environment of the chamber based on the value of the strength of the correlation calculated by performing machine learning.

[0200] Thus, the analysis device according to the first embodiment can quantitatively evaluate the environment of the chamber in the semiconductor manufacturing process based on the set of timing data.

[0201] [Second Embodiment]

[0202] In the above first embodiment, specific examples of the timing data acquisition device and the set of timing data are not mentioned. In this regard, in the second embodiment, the case where the timing data acquisition device is an emission spectroscopy analysis device and the set of timing data is OES (Optical Emission Spectroscopy) data will be described. In addition, the OES data is a data set including timing data of the emission intensity corresponding to the number of types of wavelengths.

[0203] Here, it is known that the OES data is related to the type and amount of deposits attached to the inside of the chamber. Therefore, by using the OES data as the set of timing data, the environment of the chamber can be evaluated from the viewpoints of the type and amount of deposits attached to the inside of the chamber. Hereinafter, the second embodiment will be described centering on the differences from the above first embodiment.

[0204] [System Configuration of the Environment Adjustment System]

[0205] First, the system configuration of the environment adjustment system will be described. Fig.11 is a diagram showing an example of the system configuration of the environment adjustment system when using OES data. The difference from Figure 1 is that an emission spectroscopy analysis device 1140 is configured as the timing data acquisition device; the OES data is stored in the learning data storage unit 163 as the set of timing data; and the environment is adjusted to a certain state using a cleaning protocol.

[0206] The emission spectroscopy analysis device 1140 measures the OES data by emission spectroscopy technology while processing the wafer 110 before processing in the chamber A. The OES data is, for example, timing data representing the emission intensity of each wavelength at each time included in the visible light wavelength range.

[0207] The cleaning protocol is a protocol used when cleaning the inside of the chamber A, and is a protocol for adjusting the environment of the chamber A to a certain state from the viewpoints of the type and amount of deposits attached to the inside of the chamber A.

[0208] [Specific Example of OES Data]

[0209] Next, a specific example of OES data measured by the emission spectrometer 1140 will be described. Fig.12 This is a diagram showing an example of OES data, showing a data set of luminous intensity when each wavelength included in the wavelength range of visible light (400 [nm] to 800 [nm]) is measured at 1 [nm] scale. Fig.12 In the figure, the horizontal axis represents time and the vertical axis represents the luminous intensity of each wavelength.

[0210] exist Fig.12 In the case of, for example, the top graph represents the luminous intensity data at each time when the wavelength is 400 [nm], and the second graph represents the luminous intensity data at each time when the wavelength is 401 [nm]. In addition, Fig.12 The third graph of shows the luminous intensity data at each time when the wavelength is 402 [nm].

[0211] <Specific example of processing by the evaluation department>

[0212] Next, a specific example of the processing performed by the evaluation unit 162 when using OES data will be described. Fig.13 It is a diagram showing a specific example of the processing performed by the evaluation unit when using OES data.

[0213] like Fig.13 As shown, when OES data is used as a time series data group, the emission intensity data of each wavelength is stored in the time series data of "first node" and "second node" of the first evaluation data 710_1', 710_2', 710_3', ...

[0214] In addition, the first evaluation data 710_1' shows:

[0215] The value of the intensity indicating the correlation between the luminous intensity data at each time of the wavelength = 400 [nm] and the luminous intensity data at each time of the wavelength 401 [nm] is "F12",

[0216] The value of the intensity indicating the correlation between the luminous intensity data at each time of the wavelength = 400 [nm] and the luminous intensity data at each time of the wavelength 402 [nm] is "F13",

[0217] The value of the intensity indicating the correlation between the luminous intensity data at each time of the wavelength = 401 [nm] and the luminous intensity data at each time of the wavelength = 402 [nm] is “F23”.

[0218] In addition, if Fig.13As shown, when OES data is used as a time series data group, the horizontal axes of graphs 910_1 ′, 910_2 ′, 910_3 ′, . . . represent the wavelengths included in the wavelength range of visible light (400 [nm] to 800 [nm]).

[0219] Similarly, when OES data is used as the time series data group, the emission intensity data of each wavelength is stored in the time series data of “the first node” and “the second node” of the second evaluation data 820 ′.

[0220] In addition, the second evaluation data 820' shows:

[0221] The value of the intensity indicating the correlation between the luminous intensity data at each time of the wavelength = 400 [nm] and the luminous intensity data at each time of the wavelength 401 [nm] is "f12",

[0222] The value of the intensity indicating the correlation between the luminous intensity data at each time of the wavelength = 400 [nm] and the luminous intensity data at each time of the wavelength = 402 [nm] is "f13",

[0223] The value of the intensity indicating the correlation between the luminous intensity data at each time of the wavelength = 401 [nm] and the luminous intensity data at each time of the wavelength = 402 [nm] is “f23”.

[0224] In addition, when OES data is used as the time series data group, the horizontal axis of the graph 920 ′ represents each wavelength included in the wavelength range of visible light (400 [nm] to 800 [nm]).

[0225] In the similarity calculation unit 802, the similarity is calculated by the same calculation method as in the first embodiment, and the environment is evaluated by the same evaluation method.

[0226] However, although not mentioned in the first embodiment, the second evaluation data 820' may have low similarity with the first evaluation data 710_1', 710_2', 710_3', ... or the graph 920' may have low similarity with the graphs 910_1', 910_2', 910_3', ...

[0227] In such a case, the evaluation unit 162 of the analysis device 160 determines that the environment of the chamber A is abnormal from the viewpoint of the type and amount of the deposits attached to the chamber A. That is, the evaluation unit 162 of the analysis device 160 can determine not only which of the preset environments the environment of the chamber A corresponds to, but also whether it is normal.

[0228] <Relationship between the total value of the intensity values ​​indicating the correlation between the wavelengths of OES data and the luminescence intensity of each wavelength>

[0229] Next, the relationship between the total value of the intensity values ​​indicating the correlation between the wavelengths of OES data and the emission intensity of each wavelength will be described. Fig.14 This is a diagram showing the relationship between the total value of the intensity values ​​indicating the correlation between the wavelengths of OES data and the emission intensity of each wavelength.

[0230] exist Fig.14 Graph 1410 shows the total value of the intensity values ​​indicating the correlation between the wavelengths. On the other hand, graph 1420 shows the maximum emission intensity at each wavelength.

[0231] As can be seen from the comparison between the graph 1410 and the graph 1420 , at a wavelength where the emission intensity reaches a peak, the total value of the intensity values ​​indicating the correlation between the wavelengths becomes low.

[0232] Here, in the learning unit 161, when evaluating the environment of the chamber A, the luminous intensity data of the wavelengths having the largest total value of the intensity values ​​indicating the correlation between the wavelengths are used. In other words, in the learning unit 161, the luminous intensity data of the wavelengths whose luminous intensity is not the peak are used to evaluate the environment. This is a general evaluation method for evaluating the environment of the chamber using OES data, and is greatly different from the evaluation method using the luminous intensity data of the wavelengths whose luminous intensity is the peak.

[0233] That is, when the analyzer 160 of the present embodiment evaluates the environment of the chamber A using OES data, it is possible to perform the evaluation using an evaluation method different from that in the related art.

[0234] <Specific example of environmental adjustment method>

[0235] Next, a specific example of the environment adjustment method will be described. Fig.15 1 is a diagram showing a specific example of an environment adjustment method when evaluating the environment using OES data. As described above, when evaluating the environment using OES data, the control device 170 uses a cleaning recipe to adjust the environment to a certain level.

[0236] At this time, the control device 170 refers to Fig.15 The given environmental adjustment parameter determination table 1500. Fig.15 As shown, in the environment adjustment parameter determination table 1500, the items of information include "current environment", "plan according to the environment", and "cleaning plan".

[0237] The “current environment” stores information indicating the current environment of the chamber A output by the evaluation unit 162 of the analysis device 160 .

[0238] The “plan according to the environment” stores a plan according to the current environment of the chamber A, which is used when adjusting the environment of the chamber A.

[0239] The “cleaning recipe” stores a predetermined cleaning recipe used when cleaning the interior of the chamber A.

[0240] The control device 170 adjusts the environment of the chamber A using a recipe corresponding to the evaluated current environment, and then cleans the inside of the chamber A using a predetermined cleaning recipe, thereby adjusting the inside of the chamber A to a constant environment.

[0241] However, the method of adjusting to a certain environment using a cleaning recipe is not limited to this. For example, the process of cleaning the inside of chamber A using a predetermined cleaning recipe may also be the process of adjusting the environment of chamber A using a recipe corresponding to the environment.

[0242] Specifically, instead of performing the process of adjusting the environment of the chamber A using a recipe corresponding to the environment, the environment may be adjusted to a constant level by adjusting the process time of the process of cleaning the chamber A using a predetermined cleaning recipe.

[0243] <Conclusion>

[0244] According to the above description, it can be seen that in the analysis device of the second embodiment,

[0245] Machine learning is performed using OES data measured while processing a pre-processed wafer in a chamber to calculate an intensity value representing correlation between emission intensity data of corresponding time ranges at respective wavelengths.

[0246] Using each OES data measured while processing a pre-processed wafer under a plurality of known environments of the chamber, an unknown environment of the chamber is evaluated based on a value indicating the strength of correlation calculated by performing machine learning.

[0247] Thus, according to the analysis device of the second embodiment, the environment of the chamber in the semiconductor manufacturing process can be quantitatively evaluated from the viewpoints of the types and amounts of deposits attached to the chamber based on OES data.

[0248] In the second embodiment, the time series data acquisition device is an emission spectrometer and the time series data group is OES data. However, the time series data acquisition device may be a mass spectrometer (e.g., a quadrupole mass spectrometer). In this case, the time series data group is a data set of time series data (mass spectrometer data) of measured intensities corresponding to the number of types of mass values ​​(m / z values).

[0249] [Third Embodiment]

[0250] In the second embodiment, the time series data set is described as OES data, but the time series data set is not limited to OES data, and may be, for example, a process data set (RF power data, pressure data, temperature data, etc.) measured by various process sensors.

[0251] Here, it is known that there is a correlation between the process data and the consumption (or degradation) of each component in the chamber. Therefore, by using the process data group as the time series data group, the environment of the chamber can be evaluated from the perspective of the consumption (or degradation) of each component in the chamber. The third embodiment is described below, focusing on the differences from the first or second embodiment described above.

[0252] <System Configuration of Environmental Adjustment System>

[0253] First, the system configuration of the environment adjustment system will be described. Fig.16 FIG. 1 is a diagram showing an example of a system configuration of an environment adjustment system when a process data set is used. Figure 1 The difference is that the process data acquisition devices 1640_1, 1640_2, ..., 1640_n are configured as the timing data acquisition device. In addition, the process data group is stored in the learning data storage unit 163 as the timing data group, and the environment is adjusted to a certain level using the position data of the focus ring.

[0254] The process data acquisition devices 1640_1, 1640_2, ..., 1640_n measure process data sets as the pre-processed wafer 110 is processed in the corresponding chamber A. The process data sets include, for example, RF power data, pressure data, gas flow data, current data, GAP length data, temperature data, etc. at each time.

[0255] The focus ring position data is position data after the height direction position of the focus ring is changed based on the wear degree of the focus ring as an example of a component in the chamber A. The focus ring position data is data for adjusting the environment in the chamber A to a constant environment.

[0256] <Specific example of process data group>

[0257] Next, a specific example of a process data group measured by the process data acquisition devices 1640_1, 1640_2, ..., 1640_n will be described. Fig.17 This is a diagram showing an example of a process data group. Fig.17The example shows a case where the process data acquisition device 1640_1 measures RF power data as process data 1, and the process data acquisition device 1640_2 measures pressure data as process data 2. In addition, Fig.17 The example shows a situation where the process data acquisition device 1640_3 measures gas flow data as process data 3.

[0258] Likewise, Fig.17 The example shows a situation where the process data acquisition device 1640_n-2 measures current data as process data n-2, and the process data acquisition device 1640_n-1 measures GAP length data as process data n-2. In addition, Fig.17 The example shows a case where the process data acquisition device 1640_n measures temperature data as process data n.

[0259] <Specific example of processing by the evaluation department>

[0260] Next, a specific example of the processing performed by the evaluation unit 162 when the process data group is used will be described. Fig.18 This is a diagram showing a specific example of processing performed by the evaluation unit when using a process data group.

[0261] like Fig.18 As shown, when the process data group is used as the time series data group, the process data of each measurement item is stored in the time series data of the "first node" and the "second node" of the first evaluation data 710_1", 710_2", 710_3", ...

[0262] In addition, the first evaluation data 710_1″ shows:

[0263] The value indicating the strength of the association between process data 1 and process data 2 is "F12",

[0264] The value indicating the strength of the correlation between process data 1 and process data 3 is "F13",

[0265] The value indicating the strength of the correlation between the process data 2 and the process data 3 is “F23”.

[0266] In addition, if Fig.18 As shown, when the process data group is used as the time series data group, the horizontal axes of the charts 910_1", 910_2", 910_3", ... are the various measurement items.

[0267] Similarly, when the process data group is used as the time series data group, the process data of each measurement item is stored in the time series data of the "first node" and the "second node" of the second evaluation data 820".

[0268] In addition, the second evaluation data 820" shows:

[0269] The value indicating the strength of the association between process data 1 and process data 2 is "f12",

[0270] The value indicating the strength of the association between process data 1 and process data 3 is "f13",

[0271] The value indicating the strength of the relationship between the process data 2 and the process data 3 is “f23”.

[0272] In addition, if Fig.18 As shown, when the process data group is used as the time series data group, the horizontal axis of the chart 920" is each measurement item.

[0273] In the similarity calculation unit 802, the similarity is calculated by the same calculation method as in the first embodiment, and the environment is evaluated by the same evaluation method.

[0274] However, although not mentioned in the first embodiment, the second evaluation data 820" may have low similarity with the first evaluation data 710_1", 710_2", 710_3", ... Or the graph 920" may have low similarity with the graphs 910_1", 910_2", 910_3", ...

[0275] In such a case, the evaluation unit 162 of the analysis device 160 determines that the environment of the chamber A is abnormal from the viewpoint of the degree of consumption of each component in the chamber A. In other words, the evaluation unit 162 of the analysis device 160 can determine not only whether the environment of the chamber A corresponds to any of the preset environments, but also whether it is normal.

[0276] <Specific example of environmental adjustment method>

[0277] Next, a specific example of the environment adjustment method will be described. Fig.19 1 is a diagram showing a specific example of an environment adjustment method when evaluating the environment using a process data set. As described above, when evaluating the environment using a process data set, the control device 170 uses the position data of the focus ring and the like to adjust to a constant environment.

[0278] At this time, the control device 170 refers to Fig.19 The environmental adjustment parameter determination table 1900 is shown. Fig.19 As shown, in the environment adjustment parameter determination table 1900, the items of information include "current environment", "focus ring position", and "applied voltage".

[0279] The “current environment” stores information indicating the current environment of the chamber A output by the evaluation unit 162 of the analysis device 160 .

[0280] The “focus ring position” stores position data after the change in the height direction position of the focus ring when the position is changed according to the evaluated environment (the degree of wear of each component).

[0281] In “applied voltage”, applied voltage data when a voltage is applied instead of changing the position of the focus ring is stored.

[0282] When the information indicating the current environment is notified from the analyzing device 160, the control device 170 refers to the environment adjustment parameter determination table 1900 and determines the position data of the focus ring when the position of the focus ring can be changed. In addition, when the position of the focus ring cannot be changed, the control device 170 determines the applied voltage data. Furthermore, the control device 170 adjusts the chamber A to a certain environment using the determined position data or voltage data.

[0283] In addition, according to Fig.19 For example, when the current environment = "environment 1", the control device 170 determines that the focus ring position = "position 1" or the applied voltage = "DC1".

[0284] <Conclusion>

[0285] According to the above description, it can be seen that in the analysis device of the third embodiment,

[0286] Machine learning is performed using a process data group measured while processing a pre-processed wafer in a chamber, and a value indicating the strength of correlation between process data in a corresponding time range between measurement items is calculated.

[0287] Using each process data set measured when processing a pre-processed wafer under a plurality of known environments of the chamber, an unknown environment of the chamber is evaluated based on a value indicating the strength of correlation calculated by performing machine learning.

[0288] Thus, according to the analysis device of the third embodiment, it is possible to quantitatively evaluate the environment of the chamber in the semiconductor manufacturing process from the viewpoint of the degree of consumption of each component in the chamber based on the process data group.

[0289] [Fourth embodiment]

[0290] In the second and third embodiments, when the similarity of the second evaluation data to the plurality of first evaluation data is low, it is described that the environment of the chamber is judged to be abnormal. That is, in the second and third embodiments, it is described that whether the environment deviates from the normal environment is judged based on the value indicating the strength of the correlation of the time series data.

[0291] However, the value used to determine whether the chamber environment is normal is not limited to the value indicating the strength of the correlation of the time series data. For example, it can also be a predetermined count value that can be calculated by executing a regression model. The following describes the fourth embodiment with the differences from the first to third embodiments as the center.

[0292] <System Configuration of Environmental Adjustment System>

[0293] First, the system configuration of the environment adjustment system will be described. Fig. 20 FIG. 2 is a second diagram showing an example of a system configuration of an environmental adjustment system. Figure 1 The differences are that, in the case of the environment adjustment system 2000, the function of the learning unit 211 of the analysis device 2010 is different from that of the learning unit 161, and the function of the evaluation unit 212 is different from that of the evaluation unit 162. In addition, in the case of the environment adjustment system 2000, the analysis device 2010 does not have an evaluation data storage unit.

[0294] The learning unit 2011 performs machine learning on the regression model using the learning data.

[0295] The evaluation unit 2012 calculates a predetermined count value by inputting a time series data group (estimation data) measured in an unknown environment into the regression model generated by the learning unit 161 through machine learning using the learning data.

[0296] In addition, the number of combinations of the first node and the second node in which the first node and the second node have a predetermined association is counted in the evaluation unit 2012. The predetermined count value refers to the number of combinations in which the first node and the second node have a predetermined association in which the predetermined association is destroyed (the value indicating the strength of the association is below a predetermined threshold value).

[0297] When calculating a predetermined count value, the evaluation unit 2012 first obtains a regression model generated by the learning unit 2011 by machine learning using a time series data group (learning data) measured under a normal environment. Next, the evaluation unit 2012 inputs a time series data group (inference data) measured under an unknown environment into the obtained regression model to calculate a predetermined count value. Thus, the evaluation unit 2012 can determine whether the environment of chamber A is normal (or the degree of abnormality of chamber A).

[0298] In addition, based on the information indicating the environment (whether chamber A is normal (or the degree of abnormality of chamber A)) output by the evaluation unit 2012, it is determined that, for example:

[0299] Whether chamber A needs maintenance, whether parts of chamber A need maintenance, whether parts that affect chamber A need maintenance, or

[0300] The timing of maintenance of chamber A, the timing of maintenance of components of chamber A, the timing of maintenance of components affecting chamber A, etc.

[0301] <Specific example of processing performed by the learning unit>

[0302] Next, a specific example of the processing performed by the learning unit 2011 of the analysis device 2010 will be described. Fig.21 FIG. 3 is a specific example of the processing performed by the learning unit. Fig.21 As shown, the learning unit 2011 includes a regression model generating unit 2101 .

[0303] The regression model generator 2101 performs machine learning on the regression model using the time series data group included in the learning data stored in the learning data storage unit 163. Thus, the regression model generator 2101 defines the relationship between the time series data of the measurement items measured by each time series data acquisition device 140_1 to 140_n using the mathematical formula shown by reference numeral 2110.

[0304] Specifically, the regression model generation unit 2101 calculates each parameter of the mathematical formula shown in the figure 2110 in such a manner that the time series data of the second node is derived by inputting the time series data of the first node into the mathematical formula shown in the figure 2110 .

[0305] In the mathematical formula shown by reference numeral 2110, it is expressed as

[0306] t: time,

[0307] m: autocorrelation (a parameter indicating whether there is periodicity),

[0308] n: cross-correlation (a parameter indicating whether they are correlated),

[0309] k: time delay,

[0310] β, α, and C represent predetermined coefficients.

[0311] exist Fig.21 , the learning result 2120 shows the parameters of the mathematical formula shown by reference numeral 2110, which are calculated by performing machine learning on the regression model. Specifically, in the learning result 2120, the items as information include "1st node", "2nd node", "autocorrelation", "cross-correlation", and "time delay" as other examples of values ​​representing correlation.

[0312] In the learning result 2120 , the “first node” and the “second node” respectively store the time series data used to derive the mathematical expression indicated by reference numeral 2110 in the time series data group included in the learning data.

[0313] In addition, in the learning result 2120, the parameters m, n, and k calculated by inputting the time series data of the second node into the mathematical formula shown by the figure mark 2110 to derive the time series data of the second node are stored in "Autocorrelation", "Cross-correlation", and "Time Delay".

[0314] In addition, if Fig.21 As shown, only one learning result 2120 is generated for the learning data including the time series data group measured under the normal environment.

[0315] <Specific example of processing by the evaluation department>

[0316] Next, a specific example of the processing performed by the evaluation unit 2012 of the analysis device 2010 will be described. Fig. 22 FIG. 3 is a specific example of the processing performed by the evaluation unit. Fig. 22 As shown, the evaluation unit 2012 includes a regression model execution unit 2210 and a count value calculation unit 2220 .

[0317] The regression model execution unit 2210 extracts the time series data (measured value 2211) of the first node from the time series data group (inference data) measured in the unknown environment of chamber A. In addition, the regression model execution unit 2210 infers the time series data (inferred value 2212) of the second node by inputting the extracted time series data of the first node into the mathematical formula shown by reference numeral 2110.

[0318] At this time, the regression model execution unit 2210 reads out the parameters m, n, k corresponding to the time series data input to the mathematical formula shown in the figure mark 2110 from the learning result 2120, and after setting the mathematical formula shown in the figure mark 2110, infers the time series data of the second node.

[0319] exist Fig. 22 , the measured value 2211 represents the time series data of the first node input to the mathematical formula shown by reference numeral 2110 in the time series data group (inference data) measured in the unknown environment of chamber A. In addition, the inferred value 2212 represents the time series data of the second node inferred by inputting the measured value 2211.

[0320] On the other hand, the count value calculation unit 2220 includes a difference calculation unit 2221 and a counting unit 2222 .

[0321] The difference calculation unit 2221 extracts the time series data (measured value 2223) of the second node from the time series data group (inference data) measured in the unknown environment of chamber A. In addition, the difference calculation unit 2221 obtains the inferred value 2212 from the regression model execution unit 2210. Furthermore, the difference calculation unit 2221 calculates the difference between the measured value 2223 and the inferred value 2212.

[0322] The counting unit 2222 counts the number of first nodes (i.e., a predetermined count value) whose difference calculated by the difference calculating unit 2221 is greater than a predetermined threshold value. In addition, the counting unit 2222 outputs the counted predetermined count value as information indicating the environment of the chamber A (information indicating whether it is normal (or the degree of abnormality)).

[0323] exist Fig. 22 In the graph 2230, the horizontal axis represents time, and the vertical axis represents a predetermined count value output by the counting unit 2222. As shown in the graph 2230, when the predetermined count value output by the counting unit 2222 is less than the level (abnormality determination level) indicated by the dotted line, the predetermined count value can be said to be information indicating that the chamber A is in a normal environment.

[0324] On the other hand, when the count value outputted from the counting unit 2222 reaches the abnormality determination level, the predetermined count value can be regarded as information indicating that the chamber A is not a normal environment.

[0325] In addition, the predetermined count value output by the counting unit 2222 can also be understood as information indicating the degree of abnormality of the environment of the chamber A by comparing it with the abnormality judgment level. Alternatively, the predetermined count value output by the counting unit 2222 can also be understood as information predicting the time when the environment of the chamber A becomes abnormal by using it to predict the time when the abnormality judgment level is reached.

[0326] <Process of environmental adjustment>

[0327] Next, the overall flow of the environment adjustment process performed by the environment adjustment system 2000 will be described. Fig.23 This is the third flow chart showing the flow of the environment adjustment process.

[0328] In step S2301, the timing data acquisition devices 140_1 to 140_n measure timing data groups while processing the pre-processed wafer in chamber A, and store them in the learning data storage unit 163. In addition, the timing data acquisition devices 140_1 to 140_n store timing data groups measured while processing in chamber A under normal conditions as learning data.

[0329] In step S2302 , the learning unit 2011 of the analysis device 2010 performs machine learning on the regression model using the learning data stored in the learning data storage unit 163 .

[0330] In step S2303 , the time series data acquisition devices 140_1 to 140_n measure a time series data group as the pre-processed wafer is processed in the chamber A. In addition, the time series data acquisition devices 140_1 to 140_n measure a time series data group (inference data) as the chamber A processes in an unknown environment.

[0331] In step S2304 , the evaluation unit 2012 of the analysis device 2010 inputs the time series data group (estimation data) measured in step S2303 into the regression model, and calculates a predetermined count value.

[0332] In step S2305 , the evaluation unit 2012 of the analysis device 2010 outputs the calculated predetermined count value as information indicating the environment of the chamber A to the control device 170 .

[0333] In step S2306, control device 170 determines whether maintenance is required or when maintenance is required, etc., based on information indicating the environment.

[0334] <Conclusion>

[0335] According to the above description, it can be seen that in the analysis device of the fourth embodiment,

[0336] Using a time series data set measured while processing a pre-processed wafer in a chamber, machine learning is performed on a regression model to calculate another example of a value representing the correlation of the time series data in a corresponding time range between each measurement item, namely, autocorrelation, cross-correlation, time delay, etc.

[0337] By inputting a time series data group (estimation data) measured while processing a pre-processed wafer in an unprocessed environment of the chamber into a regression model, a predetermined count value is calculated and output as information representing the environment of the chamber.

[0338] Thus, according to the analysis device of the fourth embodiment, it is possible to quantitatively evaluate whether the environment of the chamber in the semiconductor manufacturing process is normal (or the degree of abnormality) based on the time series data group.

[0339] [Fifth embodiment]

[0340] In the fourth embodiment described above, a scheme has been described in which a time series data group measured while processing a pre-processed wafer is input into a regression model to calculate a predetermined count value and output information indicating the environment of a chamber.

[0341] In contrast, in the fifth embodiment, the time series data set is input into the regression model to calculate a predetermined count value, and the change of the predetermined count value is monitored. Thus, in the fifth embodiment, the change of the chamber environment can be grasped to detect the end point of the etching process and the cleaning process.

[0342] <System Configuration of End Point Detection System>

[0343] First, the system configuration of the endpoint detection system will be described. Fig.24 FIG. 1 is a diagram showing an example of a system configuration of an endpoint detection system. Figure 1 The difference of the illustrated environment adjustment system 100 lies in that, in the case of the end point detection system 2400 , the function of the analysis device 2410 is different and the function of the control device 2420 is different.

[0344] like Fig.24 As shown, the analysis device 2410 functions as a learning unit 2411 and an end point detection unit 2412. The learning unit 2411 and the end point detection unit 2412 have two functions, and the end point detection is performed by either function.

[0345] (i) Description of the First Function of the Learning Unit and the End Point Detection Unit

[0346] The learning unit 2411 uses the time series data acquired by the time series data acquisition devices 140_1 to 140_n.

[0347] The time point when the etching process in the processing space 120 ends, or

[0348] The time point when the cleaning process in the processing space 120 is completed

[0349] The measured time series data set (learning data) is used to perform machine learning on the regression model.

[0350] The endpoint detection unit 2412 inputs the regression model generated by the learning unit 2411 through machine learning using the learning data.

[0351] A timing data group (detection data) measured during the etching process in the processing space 120, or

[0352] A time series data group (detection data) measured during the cleaning process in the process space 120,

[0353] The "predetermined count value" is calculated thereby.

[0354] Then, the end point detection unit 2412 detects the time point when the predetermined count value changes to below a preset threshold value as the end point of the etching process or the end point of the cleaning process. In addition, the end point detection unit 2412 sends the detected end point information such as the end point of the etching process or the end point of the cleaning process to the control device 2420.

[0355] In addition, the number of combinations of the first node and the second node in which the first node and the second node have a predetermined correlation is counted in the end point detection unit 2412. The above-mentioned "predetermined count value" refers to the number of combinations in which the first node and the second node have a predetermined correlation and the predetermined correlation is destroyed (the value indicating the strength of the correlation becomes less than a predetermined threshold value).

[0356] (ii) Description of the Second Function of the Learning Unit and the End Point Detection Unit

[0357] The learning unit 2411 uses the time series data acquired by the time series data acquisition devices 140_1 to 140_n.

[0358] The time point when the etching process starts in the processing space 120, or

[0359] The time point when the cleaning process starts in the processing space 120

[0360] The measured time series data set (learning data) is used to perform machine learning on the regression model.

[0361] The endpoint detection unit 2412 inputs the regression model generated by the learning unit 2411 through machine learning using the learning data.

[0362] A timing data group (detection data) measured during the etching process in the processing space 120, or

[0363] Time series data group (detection data) measured during the cleaning process in the process space 120

[0364] To calculate the "specified count value".

[0365] Then, the end point detection unit 2412 detects the time point when the predetermined count value changes to a predetermined threshold value or more as the end point of the etching process or the end point of the cleaning process. In addition, the end point detection unit 2412 sends the detected end point information such as the end point of the etching process or the end point of the cleaning process to the control device 2420.

[0366] (iii) Description of the functions of the control device

[0367] The control device 2420 adjusts, for example, etching time, etching pattern, or cleaning time, cleaning pattern, etc. based on the endpoint information output by the endpoint detection unit 2412 of the analysis device 2410 .

[0368] <Specific example of processing performed by the learning unit>

[0369] Next, a specific example of the processing performed by the learning unit 2411 of the analysis device 2410 will be described. Fig.25 FIG. 4 is a specific example of the processing performed by the learning unit. Fig.25 As shown, the learning unit 2411 has a regression model generating unit 2501.

[0370] The regression model generation unit 2501 uses the time series data group stored in the data storage unit 2413 to perform machine learning on the regression model. Fig.25 In FIG. 1 , a machine learning is performed using a time series data set measured at the time point when the etching process is finished or the time point when the cleaning process is finished. Fig.25 Although not shown in the figure, machine learning can also be performed using a time series data set measured at the time point when the etching process starts or the time point when the cleaning process starts.

[0371] Thus, the regression model generation unit 2501 defines the relationship between the time series data of each measurement item measured by each time series data acquisition device 140_1 to 140_n using the mathematical formula indicated by reference numeral 2110 .

[0372] The method of defining the relationship between the time series data of each measurement item using the mathematical expression shown by reference numeral 2110 has already been used in the fourth embodiment described above. Fig.21 It has been explained before, so the explanation is omitted here.

[0373] In addition, Fig.25 In the example, the learning result 2520 shows the parameters of the mathematical formula shown by reference numeral 2110 calculated by performing machine learning on the regression model. The details of the learning result 2520 have also been used in the fourth embodiment described above. Fig.21 It has been explained before, so the explanation is omitted here.

[0374] <Specific Example of Processing by End Point Detection Section>

[0375] Next, a specific example of the process performed by the endpoint detection unit of the analysis device 2410 will be described. Fig.26 is a diagram showing a specific example of the processing performed by the endpoint detection unit. Fig.26 As shown, the endpoint detection unit 2412 includes a regression model execution unit 2610 and a count value calculation unit 2620 .

[0376] The regression model execution unit 2610 extracts the time series data (measured value 2211) of the first node from the time series data group (detection data) measured during the etching process or the cleaning process. In addition, the regression model execution unit 2610 inputs the extracted time series data of the first node into the mathematical formula shown by the reference numeral 2110, thereby inferring the time series data of the second node (inferred value 2212).

[0377] At this time, the regression model execution unit 2610 reads out the parameters m, n, k corresponding to the time series data input to the mathematical formula shown in the figure mark 2110 from the learning result 2520, sets the mathematical formula shown in the figure mark 2110, and then infers the time series data of the second node.

[0378] exist Fig.26 , the measured value 2211 represents the timing data of the first node in the timing data group (detection data) measured in the etching process or the cleaning process and input to the mathematical formula shown by the reference numeral 2110. In addition, the estimated value 2212 represents the timing data of the second node estimated by inputting the measured value 2211.

[0379] On the other hand, the count value calculation unit 2620 includes a difference calculation unit 2621 , a counting unit 2622 , and a determination unit 2623 .

[0380] The difference calculation unit 2621 extracts the time series data (measured value 2223) of the second node from the time series data group (detection data) measured during the etching process or the cleaning process. In addition, the difference calculation unit 2621 obtains the estimated value 2212 through the regression model execution unit 2610. The difference calculation unit 2621 also calculates the difference between the measured value 2223 and the estimated value 2212.

[0381] The counting unit 2622 counts the number of first nodes (that is, a predetermined count value) whose differences calculated by the difference calculating unit 2621 are equal to or greater than a predetermined threshold.

[0382] When the determination unit 2623 performs machine learning using the time series data set at the time when the etching process is completed or the time when the cleaning process is completed,

[0383] The time point when a predetermined count value counted by the counting unit 2622 becomes less than a preset threshold value is determined as the end point of the etching process or the end point of the cleaning process, and the end point information is output.

[0384] In addition, when the determination unit 2623 performs machine learning using the time series data set at the time point when the etching process starts or the time point when the cleaning process starts,

[0385] The time point when a predetermined count value counted by the counting unit 2622 becomes equal to or greater than a preset threshold value is determined as the end point of the etching process or the end point of the cleaning process, and end point information is output.

[0386] <Flow of Endpoint Detection Processing>

[0387] Next, the overall flow of the endpoint detection process performed by the endpoint detection system 2400 will be described. Fig. 27 : is a flowchart showing the flow of the end point detection process.

[0388] In step S2701, the time series data acquisition devices 140_1 to 140_n save

[0389] At the time when the etching process (or cleaning process) is completed, or

[0390] A time series data group (learning data) measured at the time point when the etching process (or cleaning process) starts.

[0391] In step S2702 , the learning unit 2411 of the analysis device 2410 performs machine learning on the regression model using the time series data group (learning data) stored in the data storage unit 2413 .

[0392] In step S2703 , the timing data acquisition devices 140_1 to 140_n measure a timing data group in the etching process or a timing data group in the cleaning process (detection data).

[0393] In step S2704, the endpoint detection unit 2412 of the analysis device 2410 inputs the time series data group (detection data) measured in step S2703 into the regression model, and calculates a predetermined count value.

[0394] In step S2705, the endpoint detection unit 2412 of the analysis device 2410 determines whether the calculated predetermined count value satisfies a preset condition. The preset condition here is:

[0395] When machine learning is performed using a time series data set measured at the time point when the etching process (or cleaning process) is completed, it means that the value is below a preset threshold value;

[0396] When machine learning is performed using a time series data group measured at the start time of an etching process (or a cleaning process), this means a case where the value exceeds a preset threshold value.

[0397] In step S2705, if it is determined that the preset condition is not satisfied (if NO in step S2705), the process returns to step S2703.

[0398] On the other hand, when it is determined in step S2705 that the preset condition is satisfied (if yes in step S2705), the process proceeds to step S2706.

[0399] In step S2706 , the end point detection unit 2412 of the analysis device 2410 determines that the end point of the etching process or the end point of the cleaning process has been detected, and outputs the end point information.

[0400] <Conclusion>

[0401] According to the above description, it can be seen that in the analysis device of the fifth embodiment,

[0402] Using a time series data set measured at the end of the etching process or at the cleaning process, machine learning for a regression model, or

[0403] Using a time series data set measured at a time point when an etching process starts or a time point when a cleaning process starts, machine learning is performed on a regression model.

[0404] Based on the estimated value when a time series data group (detection data) measured during etching or cleaning is input to a regression model subjected to machine learning, the number of combinations of time series data having a predetermined correlation between measurement items is counted.

[0405] When the count value satisfies a preset condition, it is determined that the end point of the etching process or the end point of the cleaning process is detected, and the end point information is output.

[0406] Therefore, according to the analysis device of the fifth embodiment, it is possible to determine the end point of the etching process or the cleaning process with high accuracy.

[0407] [Sixth embodiment]

[0408] In the fifth embodiment, no specific examples of the time series data acquisition device and the time series data group are mentioned. However, similarly to the second and third embodiments, the time series data acquisition device may be, for example,

[0409] Emission spectrum analysis device;

[0410] Quadrupole mass spectrometer;

[0411] Various process sensors.

[0412] In addition, the time series data set may be, for example,

[0413] OES data;

[0414] Mass spectrometry data;

[0415] Process data set.

[0416] In addition, in the fifth embodiment, in order to detect the end point of the etching process or the end point of the cleaning process, the time series data group measured at the time point when the etching process ends or the time point when the cleaning process ends is used to perform machine learning on the regression model. However, in order to detect a specific state of the etching process or a specific state of the cleaning process, the time series data group measured at a specific time point of the etching process or a specific time point of the cleaning process may also be used to perform machine learning on the regression model.

[0417] [Other embodiments]

[0418] In the above embodiments, the learning unit is described as a device that performs machine learning on a regression model. However, the model that the learning unit performs machine learning on is not limited to a regression model, and any model that can calculate the correlation (association) of time series data may be another model.

[0419] In addition, in the second embodiment described above, the contents of generating the first learning data and the second learning data are described with the luminous intensity data of each wavelength included in the wavelength range of visible light as the object. However, the luminous intensity data used for generating the first learning data and the second learning data may also be luminous intensity data of a specific wavelength. In addition, it may also be luminous intensity data of a wavelength outside the wavelength range of visible light.

[0420] In addition, in the fourth embodiment described above, the counting unit 2222 calculates the predetermined count value by counting the number of first nodes whose differences calculated by the difference calculation unit 2221 are greater than the predetermined threshold. However, the counting method of the predetermined count value is not limited to this. For example, the predetermined count value may be calculated by counting the number of first nodes that are preset among the first nodes whose differences calculated by the difference calculation unit 2221 are greater than the predetermined threshold.

[0421] In the second embodiment, OES data (or mass spectrometry data) is cited as a specific example of the time series data set, and in the third embodiment, process data is cited as a specific example of the time series data set, but the time series data set is not limited thereto. For example, it may be a time series data set representing a plasma physical quantity measured by a plasma device.

[0422] Furthermore, in each of the above-described embodiments, the analyzing device and the controlling device are configured as separate bodies, but the analyzing device and the controlling device may be configured as an integral body.

[0423] The present invention is not limited to the configurations given in the above embodiments, combinations with other elements, etc. These aspects can be changed within the scope of the present invention and can be appropriately determined according to the application.

Claims

1. An analysis device, characterized in that: include: a learning unit that performs machine learning using a time series data group measured in conjunction with processing of an object in a processing space, and calculates a value indicating a correlation between the time series data in a corresponding time range between the measurement items; and an evaluation unit that uses a time series data group measured by processing an object in a known environment of the processing space to evaluate an unknown environment of the processing space based on a value indicating the correlation calculated by machine learning by the learning unit, The analysis device further includes a storage unit that stores, as information indicating the corresponding environment, a value indicating the correlation calculated by machine learning by the learning unit using a plurality of time series data sets measured by processing an object in a plurality of known environments with a processing space, The evaluation unit evaluates the unknown environment of the processing space by judging whether a value representing association calculated by machine learning of a time series data group measured when processing an object in an unknown environment in the processing space is similar to a value representing association stored in the storage unit.

2. The analysis device according to claim 1, characterized in that: The storage unit stores the value representing the correlation calculated by machine learning performed by the learning unit using multiple time series data groups measured by processing an object based on a specific scheme under multiple normal known environments accompanying the processing space as information representing the corresponding environment.

3. The analysis device according to claim 2, characterized in that: The storage unit will use multiple time series data groups measured when processing an object under multiple known environments with different types or amounts of deposits attached to the processing space to save the value representing the correlation calculated by machine learning performed by the learning unit as information representing the corresponding environment.

4. The analysis device according to claim 2, characterized in that: The storage unit stores the value representing the correlation calculated by machine learning performed by the learning unit using multiple time series data groups measured by processing an object in multiple known environments with different consumption levels of each component in a processing space as information representing the corresponding environment.

5. The analysis device according to claim 3, characterized in that: The plurality of time-series data groups are OES data measured by an emission spectrometer or mass spectrometer data measured by a mass spectrometer.

6. The analysis device according to claim 4, characterized in that: The plurality of time series data groups are process data groups measured by a process data acquisition device.

7. The analysis device according to claim 4, characterized in that: The plurality of time series data sets are time series data sets of plasma physical quantities measured by the plasma device.

8. The analysis device according to any one of claims 1 to 7, characterized in that: The value indicating the correlation is a value indicating the strength of correlation between the time series data in the corresponding time range between the measurement items.

9. The analysis device according to any one of claims 1 to 7, characterized in that: The value indicating the correlation is a value obtained by summing up, for each measurement item, values ​​indicating the strength of correlation of time series data in corresponding time ranges between the measurement items.

10. The analysis device according to any one of claims 1 to 7, characterized in that: Based on the environment evaluated by the evaluation unit, an environmental adjustment parameter for changing the environment of the processing space is determined.

11. The analysis device according to claim 1, characterized in that: The value indicating the correlation includes a parameter indicating autocorrelation, cross-correlation, or time delay, which is calculated so that the time series data of the second measurement item can be derived by inputting the time series data of the first measurement item into a predetermined formula.

12. The analysis device according to claim 1, characterized in that: Based on the environment evaluated by the evaluation department, it is decided that: Whether maintenance of the processing space, maintenance of components of the processing space, or maintenance of components that affect the processing space is required; or The timing of maintenance of the processing space, the timing of maintenance of components of the processing space, and the timing of maintenance of components that affect the processing space.

13. An analysis method, characterized in that: include: A learning step of performing machine learning using a time series data group measured in conjunction with processing of an object in a processing space to calculate a value representing a correlation between the time series data in a corresponding time range between the measurement items; and an evaluation step of evaluating the unknown environment of the processing space based on the value representing the correlation calculated by machine learning in the learning step using a time series data group measured by processing an object in a known environment accompanying the processing space, The analysis method further includes a storage step, which stores the value representing the correlation calculated by machine learning using a plurality of time series data sets measured by processing an object in a plurality of known environments with a processing space as information representing the corresponding environment, The evaluation step evaluates the unknown environment of the processing space by determining which of the values ​​representing the association calculated by machine learning using a time series data set measured when processing an object in an unknown environment in the processing space is similar to the stored values ​​representing the association.

14. A computer program product, characterized in that: The computer program product includes an analysis program, and when the analysis program is executed by a computer, the following steps are implemented: A learning step of performing machine learning using a time series data group measured in conjunction with processing of an object in a processing space to calculate a value representing a correlation between the time series data in a corresponding time range between the measurement items; and an evaluation step of evaluating the unknown environment of the processing space based on the value representing the correlation calculated by machine learning in the learning step using a time series data group measured by processing an object in a known environment accompanying the processing space, When the analysis program is executed by a computer, a saving step is also implemented, wherein the saving step saves the value representing the correlation calculated by machine learning using a plurality of time series data sets measured by processing an object in a plurality of known environments with a processing space as information representing the corresponding environment, The evaluation step evaluates the unknown environment of the processing space by determining which of the values ​​representing the association calculated by machine learning using a time series data set measured when processing an object in an unknown environment in the processing space is similar to the stored values ​​representing the association.

Images