Computer program, analysis method, and analysis device

Abstract

Provided are a computer program, an analysis method, and an analysis device with which it is possible to investigate the state of processing for each processing step included in a processing recipe. A computer program causes a computer to execute: time-series data comprising a plurality of measurement values measured by a sensor provided in a processing device that processes a substrate in accordance with one or more processing steps; an index value calculation unit that calculates an index value for each processing step on the basis of a plurality of pieces of time-series data acquired when processing a plurality of substrates, the index value indicating a deviation in measured values between the plurality of substrates during a period in which each processing step is performed; and outputting the relationship between each processing step and the index value.

Description

Technical Field

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

[0002] In processes such as etching of substrates such as semiconductor wafers, it is desirable to maintain a stable processing state in order to maintain substrate quality. Conventionally, processing equipment has been equipped with sensors that acquire data related to the processing state, such as temperature, and manage the processing state based on this sensor data. Patent Document 1 discloses a technique that accumulates data related to the processing state, calculates a coefficient of variation for the data, and controls data accumulation based on the coefficient of variation.

[0003] Patent Document 1: Japanese Patent Application Laid-Open No. 2014-116453

[0004] The substrate treatment process is performed according to a process recipe that specifies the process details. A process recipe consists of multiple process steps in a specified order, with each process step specifying the process details. Generally, the process details vary for each process step, so the process status may vary from process step to process step. Therefore, it is preferable to monitor the process status for each process step. Summary of the Invention

[0005] The present disclosure provides a computer program, an analysis method, and an analysis device capable of investigating the status of each process step included in a process recipe.

[0006] According to a technical solution of the present disclosure, a computer program causes a computer to perform the following processing: obtaining time series data consisting of multiple measurement values ​​measured by a sensor provided in a processing device, wherein the processing device processes a substrate according to one or more processing steps; calculating an index value based on the multiple time series data obtained when processing multiple substrates, and according to the processing steps, wherein the index value represents the deviation of the measurement values ​​between the multiple substrates during the execution of each processing step; and outputting the relationship between each processing step and the index value.

[0007] According to the present disclosure, it is possible to provide a computer program, an analysis method, and an analysis device that can investigate the status of processing in each processing step included in a processing recipe. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 This is a conceptual diagram showing a configuration example of the analysis system according to the first embodiment.

[0009] Figure 2 This is a conceptual diagram showing an example of the content of a process recipe.

[0010] Figure 3 This is a block diagram showing an example of the internal structure of an analysis device.

[0011] Figure 4 This is a flowchart showing an example of the procedure of information processing executed by the analysis device according to the first embodiment.

[0012] Figure 5 is a schematic graph showing an example of normalization.

[0013] Figure 6 This is a flowchart showing an example of the processing steps of the subroutine of the F-value average calculation process.

[0014] Figure 7 This is a schematic graph showing an example of the relationship between the first period and the second period.

[0015] Figure 8 This is a schematic diagram showing a first example of a graph indicating the relationship between the processing steps and the F-value average.

[0016] Figure 9 This is a schematic diagram showing a second example of a graph indicating the relationship between the processing steps and the F-value average.

[0017] Figure 10 This is a graph showing an example of temporal changes in the F value.

[0018] Figure 11 This is a flowchart showing an example of the processing steps of the subroutine of the substrate group processing.

[0019] Figure 12 This is a graph showing an example of changes in the average F value according to the order in which each substrate group is processed.

[0020] Figure 13 This is a conceptual diagram showing a configuration example of an analysis system according to the second embodiment.

[0021] Figure 14 This is a conceptual diagram for explaining batches and time slots.

[0022] Figure 15 This is a conceptual diagram for explaining batches and time slots.

[0023] Figure 16 This is a flowchart showing an example of the procedure of information processing executed by the analysis device according to the third embodiment.

[0024] Figure 17 This is a schematic diagram showing an example of a graph indicating the relationship between the substrate group and the F-value average.

[0025] Figure 18 This is a schematic diagram showing an example of a graph indicating the relationship between the substrate group and the F-value average.

[0026] Figure 19 This is a graph showing an example of the calculation results of the deviation of the values ​​obtained by standardizing or normalizing the measured values ​​for each substrate group.

[0027] Figure 20 This is a conceptual diagram showing an example of the content of data preprocessing.

[0028] Figure 21 This is a conceptual diagram showing an example of the result of principal component analysis.

[0029] Figure 22 Schematic diagram showing an example of an image representing a two-dimensional distribution of a plurality of substrate groups.

[0030] Figure 23 It is a schematic diagram showing an example of a graph showing time changes of values ​​obtained by standardizing or normalizing measured values. DETAILED DESCRIPTION

[0031] Hereinafter, the embodiments will be described in detail based on the drawings.

[0032] <Implementation Method 1>

[0033] The process of manufacturing substrates such as semiconductor wafers includes a processing process for performing treatments such as etching on the substrates. A device that processes a substrate is called a processing device. For example, the processing device is a process chamber that performs treatments such as etching on the substrate disposed in the process chamber. In addition, the processing device processes multiple substrates in sequence. A substrate is placed in the processing device, the substrate is processed, and after the processing is completed, the substrate is removed from the processing device, and the next substrate is placed in the processing device and the same treatment is performed. The processing of the substrate is repeated. In order to stabilize the quality of the substrate, the processing state is preferably stable. In this embodiment, the state of the processing performed by the processing device is analyzed.

[0034] Figure 1This is a conceptual diagram illustrating an example configuration of an analysis system 100 according to Embodiment 1. Analysis system 100 includes a processing device 2, a sensor 3 provided in processing device 2, and an analysis device 1 that analyzes the status of processing performed by processing device 2. Processing device 2 is, for example, a process chamber included in semiconductor manufacturing equipment 20. Processing device 2 sequentially processes multiple substrates. Sensor 3 measures a physical quantity representing the status of processing performed by processing device 2. For example, processing device 2 is a device that performs plasma etching, and sensor 3 is a sensor using an OES (Optical Emission Spectrometer) that detects light generated from the plasma. Sensor 3 is connected to analysis device 1. Sensor 3 repeatedly performs measurements and inputs the measured values ​​into analysis device 1. For example, sensor 3 performs measurements and inputs the measured values ​​every predetermined unit time. Sensor 3 measures multiple physical quantities and inputs the multiple measured values ​​into analysis device 1. For example, sensor 3 measures the intensity of light of multiple different wavelengths and inputs the multiple measured values ​​representing the intensity of the multiple wavelengths into analysis device 1. Analysis device 1 executes the analysis method.

[0035] The processing device 2 processes the substrate according to a predetermined processing recipe. A processing recipe consists of multiple processing steps in a predetermined order. A processing step is the smallest unit of a time-series processing step for a substrate. Each processing step specifies the processing content for the substrate. Figure 2 This is a conceptual diagram showing an example of the content of a processing recipe. The processing recipe includes multiple processing steps such as a first processing step and a second processing step. In each processing step, the content of the processing performed by the processing device 2, such as the temperature in the processing device 2 and the applied voltage, is specified. The processing content includes processing conditions. Generally, the processing content is different for each processing step. The processing recipe may also include multiple processing steps with the same processing content. Since the order of the multiple processing steps is specified, the processing according to each processing step is performed in the specified order. For example, the processing according to the first processing step is performed first, followed by the processing according to the second processing step, and then the processing according to the other processing steps. In addition, the processing recipe may also be composed of a single processing step.

[0036] Figure 3This is a block diagram showing an example of the internal structure of an analysis device 1. The analysis device 1 is configured using a computer such as a personal computer or a server device. The analysis device 1 includes a computing unit 11, a memory 12 that stores temporary data generated by the computation, a reader 13, a storage unit 14, an operating unit 15, a display unit 16, and an interface unit 17. The computing unit 11 may be configured using, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a multi-core CPU. The computing unit 11 may also be configured using a quantum computer. The memory 12 stores temporary data generated by the computation. The memory 12 is, for example, RAM (Random Access Memory). The reader 13 reads information from a recording medium 10, such as an optical disk or portable memory device. The storage unit 14 is non-volatile, such as a hard disk or non-volatile semiconductor memory.

[0037] The operation unit 15 receives input of information such as text by operating the unit 15 from the user. The operation unit 15 is, for example, a keyboard, a pointing device, or a touch panel. The display unit 16 displays images. For example, the display unit 16 is a liquid crystal display or an electroluminescent display (EL). The operation unit 15 and the display unit 16 may be integrated. The sensor 3 is connected to the interface unit 17. The interface unit 17 receives measurement values ​​input by the sensor 3.

[0038] The computing unit 11 causes the reading unit 13 to read a computer program (program product) 141 recorded on the recording medium 10, and stores the read computer program 141 in the storage unit 14. The computing unit 11 executes processing to realize the functions of the analyzing device 1 according to the computer program 141. The computer program 141 causes the analyzing device 1 to perform information processing for analyzing the status of processing performed by the processing device 2. The computer program 141 may be pre-stored in the storage unit 14 or downloaded from outside the analyzing device 1. In this case, the analyzing device 1 may not include the reading unit 13.

[0039] Computer program 141 can be deployed on a single computer or a single website, or it can be distributed across multiple websites and deployed to be executed on multiple computers interconnected via a communication network. In other words, analysis device 1 can be configured with multiple computers, and computer program 141 can be executed on multiple computers connected via a communication network. Analysis device 1 can also be configured using a cloud server.

[0040] The following describes processing performed by the analysis system 100 . The processing device 2 performs processing such as etching on a substrate, and the analysis device 1 analyzes the status of the processing performed by the processing device 2 . Figure 4 This is a flowchart showing an example of information processing steps executed by the analysis device 1 according to Embodiment 1. Hereinafter, the steps of information processing executed by the analysis device 1 are abbreviated as S. The analysis device 1 executes the following processing by the calculation unit 11 executing information processing according to the computer program 141.

[0041] The processing device 2 processes a plurality of substrates. More specifically, the processing device 2 sequentially performs processing according to each processing step included in a prescribed processing recipe on one substrate. After the processing according to the processing recipe is completed, the processing device 2 performs processing according to the processing recipe on the next substrate from the beginning, and similarly repeats the processing on a plurality of substrates. The sensor 3 repeatedly measures and inputs a plurality of types of measurement values ​​to the analysis device 1. The analysis device 1 receives a plurality of types of measurement values ​​input from the sensor 3 using the interface unit 17, and the operation unit 11 stores the received plurality of types of measurement values ​​in the storage unit 14. The analysis device 1 obtains time series data (S1) by storing time series data consisting of a plurality of measurement values ​​measured by the sensor 3 in a time series in the storage unit 14.

[0042] In S1, the calculation unit 11 stores the time series data in the storage unit 14 according to the type of measurement value. That is, multiple types of time series data are stored. One time series data set includes measurement values ​​of the same type. The time series data is associated with information indicating the type of measurement value. For example, information indicating the wavelength is associated with the time series data. In addition, the measurement values ​​included in the time series data are ordered. The order of the measurement values ​​included in the time series data is the order in which they were measured by the sensor 3. For example, the time when the measurement value was measured or the time when the measurement value was input into the analysis device 1 is associated with each measurement value included in the time series data. In addition, information indicating the period during which processing according to a processing step was performed is associated with each measurement value included in the time series data. Here, the period during which processing according to a processing step is performed is defined as the first period. Since the processing recipe includes multiple processing steps, the period during which a substrate is processed includes multiple first periods. That is, information indicating the period during which the measurement was performed is associated with each measurement value.

[0043] Since time-series data is acquired by type of measured value, multiple types of time-series data are acquired depending on the processing of a single substrate. Multiple types of time-series data are acquired each time a substrate is processed. Information representing the substrate is associated with the time-series data. Furthermore, the input device that inputs measured values ​​from sensor 3 and analyzer 1 may be separate devices. Analyzer 1 may also execute the process of S1 by reading time-series data from the input device.

[0044] Next, the analysis device 1 standardizes the measurement values ​​included in the time series data according to the processing steps (S2). In S2, the operation unit 11 standardizes the multiple measurement values ​​involved in each processing step included in the time series data. The multiple measurement values ​​involved in each processing step are multiple measurement values ​​measured during the period when the processing device 2 performs the processing according to each processing step, that is, multiple measurement values ​​obtained in each first period. The operation unit 11 calculates the average value and standard deviation of the multiple measurement values ​​obtained in a first period included in a time series data, and divides the value obtained by subtracting the average value from each measurement value by the standard deviation to perform standardization. Let the measurement value obtained in a first period included in a time series data be x, the average value be x ∘ (mark x with a ∘ sign), and the standard deviation be σ. The standardized measurement value x st It is represented by the following formula (1).

[0045] [Mathematical formula 1]

[0046]

[0047] Figure 5 is a schematic graph showing an example of normalization. Figure 5 The horizontal axis represents time, and the vertical axis represents the measurement value. In the stage before standardization, the measurement value may vary greatly with the passage of time. The multiple measurement values ​​involved in a processing step are standardized in such a way that the average value of the multiple measurement values ​​is 0 and the variance is 1. The operation unit 11 uses formula (1) to standardize the multiple measurement values ​​involved in the multiple processing steps contained in a time series data. The operation unit 11 similarly standardizes the multiple types of time series data obtained for one substrate. In addition, the operation unit 11 similarly standardizes the multiple types of time series data obtained for each substrate of the multiple substrates. Thereafter, the analysis device 1 uses the standardized measurement values ​​to perform information processing.

[0048] The measured values ​​included in time series data may include the effects of an offset added to a certain value or a gain multiplied by a certain value. Therefore, it is difficult to compare the original time series data. Different processing steps result in different processing contents, so the offset and gain may be different. Since multiple substrates are not processed simultaneously in the same processing device 2, the offset and gain may be different depending on the substrate. By standardizing by processing step, the effects of offset and gain are removed from the measured values ​​involved in each processing step. The mean and variance of the measured values ​​involved in each processing step are the same across the time series data, making it easy to compare multiple measured values ​​involved in each processing step across the time series data.

[0049] Furthermore, in S2, the analysis device 1 may normalize the measured values ​​by processing step instead of performing normalization. In this case, the calculation unit 11 performs normalization by determining the maximum and minimum values ​​of a plurality of measured values ​​obtained during a first period included in a set of time series data, and dividing the value obtained by subtracting the minimum value from each measured value by the value obtained by subtracting the minimum value from the maximum value. The plurality of measured values ​​involved in a processing step are normalized so that the minimum value is 0 and the maximum value is 1. The analysis device 1 then uses the normalized measured values ​​for information processing. Even when normalization is performed, the effects of offset and gain are removed from the measured values ​​involved in each processing step, making it easier to compare the plurality of measured values ​​involved in each processing step between time series data. The analysis device 1 may also perform normalization so that the mean value is a value other than 0 or the variance is a value other than 1, or may perform normalization so that the minimum value is a value other than 0 or the maximum value is a value other than 1.

[0050] Next, the analyzer 1 performs an average F value calculation process for calculating the average F value of the variance analysis in each process step (S3). The average F value of the variance analysis is an index value indicating the variation of the measured values ​​in the first period related to each process step among a plurality of substrates. Figure 6 This is a flowchart showing an example of the processing steps of the F-value average calculation subroutine. Analyzer 1 selects a processing step (S31). In S31, calculation unit 11 selects a processing step from the multiple processing steps included in the processing recipe. Next, analyzer 1 selects a type of measurement value (S32). In S32, calculation unit 11 selects a type from the multiple types of measurement values ​​measured by sensor 3. Next, analyzer 1 selects a second period included in the first period associated with the selected processing step (S33).

[0051] Figure 7 This is a schematic graph showing an example of the relationship between the first period and the second period. Figure 7 The horizontal axis represents time, and the vertical axis represents measurement values. Figure 7 The multiple line graphs shown represent the temporal changes in the measured values ​​of the selected type during the first period associated with the selected processing step when processing multiple substrates using processing apparatus 2. Although the multiple substrates are not processed simultaneously, the first period associated with the same processing step is defined as the same first period. Let t be a natural number, and the time point t, which is the time point after t times the unit time has elapsed since the start of the first period. "Time point" here refers to a relative time point within the first period and is the same for multiple substrates. Let T be a natural number, and the period from time point (tT) to time point t is defined as the second period. The length of the second period is T times the unit time. If the unit time is 0.01 seconds, the length of the second period can be expressed as T × 0.01 seconds. The second period is shorter than the first period. In S33, the calculation unit 11 selects a second period from the multiple second periods. The sensor 3 performs a measurement once per unit time (e.g., 0.01 seconds). Therefore, T measured values ​​are obtained for all substrates during the second period. The unit time is not limited to 0.01 seconds. The length of the unit time can be appropriately set based on the performance of the sensor 3 or the processing capability of the analysis device 1 .

[0052] Next, the analysis device 1 calculates the F value of the variance analysis for the plurality of measurement values ​​obtained during the selected second period (S34). In S34, the calculation unit 11 calculates the F value for the plurality of measurement values ​​of the selected type during the selected second period for the plurality of substrates. Furthermore, the calculation unit 11 calculates the F value using the measurement values ​​standardized or normalized by the processing in S2.

[0053] For one substrate, in the selected second period, multiple measurement values ​​of the selected type are aggregated into one group. T measurement values ​​are obtained in the second period, so T measurement values ​​are included in one group. One group is obtained corresponding to one substrate, so multiple groups corresponding to multiple substrates are obtained. The F value is calculated using these multiple groups as objects. The F value is the ratio of the deviation of the measurement values ​​between groups to the deviation of the measurement values ​​within the group. If the number of substrates processed by the processing device 2 is set to N, the number of groups is N. The F value of the variance analysis is the value obtained by dividing the mean square between groups by the mean square within the group. That is, the F value is expressed by the following formula (2).

[0054] F value = mean square between groups / mean square within groups (2)

[0055] The inter-group mean square is the sum of squares between groups divided by the degrees of freedom between groups. The inter-group sum of squares is the square of the difference between the average of the measured values ​​included in each group and the average of the measured values ​​included in all groups, multiplied by the number of measured values ​​included in each group, and summed across multiple groups. As mentioned above, the number of measured values ​​included in each group is T. The degrees of freedom between groups are (N-1). Let the average of the measured values ​​included in the i-th group be x i Pull (in x i The average of the measured values ​​included in all groups is set as X (with a pull symbol on X), and the sum of squares between groups is set as MS a . Between-group sum of squares MS a It is expressed by the following formula (3).

[0056] [Mathematical formula 2]

[0057]

[0058] The intra-group mean square is the sum of squares within the group divided by the degrees of freedom within the group. The intra-group sum of squares is the sum of the squares of the differences between the measured values ​​in each group and the mean, added together across multiple groups. The degrees of freedom within a group are the number of measured values ​​in all groups minus the number of groups, which is (NT-N). Let the jth measured value in the i-th group be x ij , set the mean square within the group to MS w . Group mean square MS w It is expressed by the following formula (4).

[0059] [Mathematical formula 3]

[0060]

[0061] In S34, the calculation unit 11 calculates the inter-group mean square using formula (3), calculates the intra-group mean square using formula (4), and calculates the F value using formula (2). The calculation unit 11 stores the calculated F value in the storage unit 14. The inter-group mean square represents the variation in the measured value between substrates, significantly reflecting the differences between substrates. The intra-group mean square represents the variation in the measured value when processing a single substrate, reflecting the magnitude of the noise. It is clear that the larger the F value, the greater the variation in the measured value between substrates compared to the magnitude of the noise. Therefore, by calculating the F value, the magnitude of the deviation in the measured values ​​between multiple substrates is clearly understood.

[0062] Next, the analysis device 1 determines whether there are any unselected second periods (S35). In S35, the calculation unit 11 determines whether there are any unselected second periods and no F-value calculated among the multiple second periods included in the first period related to the selected processing step. If there are any unselected second periods (S35: Yes), the analysis device 1 returns the process to S33. In S33, the calculation unit 11 selects a second period from the unselected second periods. By repeating S33 to S35, the F-value is calculated for each second period.

[0063] If there are no unselected second periods (S35: No), the analysis device 1 calculates the F-value average (S36). In S36, the calculation unit 11 calculates the F-value average by averaging the multiple F-values ​​calculated for the multiple second periods. The calculation unit 11 stores the calculated F-value average in the storage unit 14. By calculating the F-value average, the magnitude of the non-instantaneous deviation in the measured values ​​generated during the first period of processing according to the processing step is indicated, rather than the instantaneous deviation in the measured values ​​between the multiple substrates. For example, the F-value average indicates the deviation in the measured values ​​between the multiple substrates that persists during the first period.

[0064] Next, the analyzer 1 determines whether there are any unselected measurement value types (S37). In S37, the calculation unit 11 determines whether there are any measurement value types that have not been selected and for which the F-value average has not been calculated. If there are any unselected measurement value types (S37: Yes), the analyzer 1 returns the process to S32. In S32, the calculation unit 11 selects a type from the unselected measurement value types. By repeating S32 to S37, the F-value average is calculated for each type of measurement value.

[0065] If there are no unselected measurement value types (S37: No), the analyzer 1 determines whether there are any unselected processing steps (S38). In S38, the calculation unit 11 determines whether there are any unselected processing steps and no F-value average calculations among the multiple processing steps included in the processing recipe. If there are any unselected processing steps (S38: Yes), the analyzer 1 returns the process to S31. In S31, the calculation unit 11 selects one of the unselected processing steps. By repeating S31 to S38, the F-value average is calculated for each processing step and each type of measurement value. If there are no unselected processing steps (S38: No), the analyzer 1 ends the F-value average calculation process in S3 and returns the process to the main process.

[0066] After S3 is completed, the analyzing device 1 outputs the relationship between the processing steps and the calculated F-value average ( S4 ). Figure 8This is a schematic diagram showing a first example of a graph showing the relationship between processing steps and F-value averages. The "***" included in the graph indicates the F-value averages. Multiple F-value averages created for multiple processing steps and multiple types of measured values ​​are arranged in descending order. In S4, the calculation unit 11 creates a graph that associates the processing steps with the types of measured values ​​and arranges the F-value averages in descending order, and displays the created graph on the display unit 16.

[0067] exist Figure 8 The larger the average F value, the higher the ranking. The combination of the processing steps and the type of measurement value is associated with the average F value. Figure 8 In the example shown, the types of measured values ​​are distinguished according to the wavelength of the measured light. By arranging the F-value averages in descending order, a table showing the combinations of processing steps and types of measured values ​​arranged in descending order of the F-value averages is displayed. By arranging the processing steps in descending order of the F-value averages, processing steps with greater deviations in the measured values ​​between multiple substrates and more unstable processing states are extracted. In addition, the types of measured values ​​with greater deviations in the measured values ​​between multiple substrates are extracted. By confirming Figure 8 The relationship between the process steps and the average F value shown can identify relatively unstable process steps. Identifying unstable process steps can help detect abnormal process steps, identify the cause of failures, or improve process recipes.

[0068] Figure 9 This is a schematic diagram showing a second example of a graph showing the relationship between a processing step and an F-value average. The "***" contained in the graph indicates the value of the F-value average. The values ​​of the F-value average associated with each processing step and each type of measurement value are listed. The F-value average is associated with each of the multiple processing steps arranged in one direction and each of the multiple types of measurement values ​​arranged in a direction intersecting the one direction, and is displayed in a two-dimensional arrangement. In addition, the display color of the F-value average varies depending on the size of the numerical value. The larger the numerical value, the more the display color emphasizes the F-value average. Therefore, Figure 9 The chart shown is called a heat map. Figure 9In S4, the display color is represented by the density of the smear. In S4, the operation unit 11 creates a chart that lists the F-value averages by associating the processing steps and the types of measured values, determines the display color according to the numerical value of the F-value average, and displays the created chart on the display unit 16. For example, a table that establishes a correspondence between the numerical range of the F-value average and the display color is pre-stored in the storage unit 14, and the operation unit 11 refers to the table to determine the display color. In addition, the display method of the F-value average other than the display color may be different according to the size of the numerical value, thereby emphasizing the F-value average with a large numerical value. For example, the concentration of the display color, the size or thickness of the text, the font of the text or the thickness of the frame may also be different according to the size of the numerical value. The F-value averages may also be numbered in the order of the size of the numerical value.

[0069] exist Figure 9 In the process, the processing steps and the types of measured values ​​are associated with each F-value average, and the F-value average with a large value is emphasized in the F-value average list. By confirming the processing steps and the types of measured values ​​associated with the emphasized F-value average, the processing steps and the types of measured values ​​with a large F-value average are clearly identified. The processing steps with a large deviation in the measured values ​​between multiple substrates and a more unstable processing state are extracted. In addition, the types of measured values ​​with a large deviation in the measured values ​​are extracted. Figure 8 Similarly to the case of the graph shown, a relatively unstable processing step can be identified. The analyzing device 1 may also output a warning when the F-value average exceeds a predetermined threshold value.

[0070] In S4 , the analyzing device 1 may also output the temporal change of the F value. Figure 10 This is a graph showing an example of temporal changes in the F value. Figure 10 The horizontal axis represents time, and the vertical axis represents F value. The calculation unit 11 creates a graph showing the time change of the F value based on the F value calculated for each time in S3 , and displays the graph on the display unit 16 . Figure 10 The graph shown shows the temporal variation of the F-value for a single processing step and a single type of measured value. The analyzer 1 can change the processing step and the type of measured value for which the F-value is displayed. For example, the analyzer 1 accepts user designation of a processing step and a type of measured value by operating the operating unit 15, and outputs the temporal variation of the F-value for the designated processing step and the designated type of measured value. Graphs showing the temporal variation of the F-value for multiple processing steps or multiple types of measured values ​​can also be displayed in a superimposed manner. By outputting the temporal variation of the F-value, the temporal variation of the deviation of the measured values ​​can be made clearer and more detailed.

[0071] After S4 is completed, the analyzing apparatus 1 performs substrate group processing of calculating an F value average for each substrate group obtained by dividing the plurality of substrates processed by the processing apparatus 2 ( S5 ). Figure 11 It is a flowchart showing an example of the steps of the processing subroutine for substrate group processing. The analysis device 1 generates a plurality of substrate groups obtained by dividing a plurality of substrates processed by the processing device 2 (S51). In S51, the operation unit 11 divides the plurality of substrates processed by the processing device 2 into a plurality of substrate groups consisting of a plurality of substrates processed continuously. In addition, the operation unit 11 generates a plurality of substrate groups in a manner that repeats a part of the plurality of substrates included in the substrate group. Preferably, the number of substrates included in each substrate group is the same. For example, it is assumed that eight substrates are processed by the processing device 2 in sequence. A first substrate group consisting of the first to fourth substrates, a second substrate group consisting of the second to fifth substrates, a third substrate group consisting of the third to sixth substrates, a fourth substrate group consisting of the fourth to seventh substrates, and a fifth substrate group consisting of the fifth to eighth substrates are generated.

[0072] Next, the analysis device 1 selects a substrate group (S52). In S52, the calculation unit 11 selects a substrate group from the generated plurality of substrate groups. Next, the analysis device 1 performs an F-value average calculation process (S53). In S53, the calculation unit 11 performs the same process as S3 to calculate the F-value average for the plurality of substrates included in the selected substrate group. Next, the analysis device 1 determines whether there are any unselected substrate groups (S54). In S54, the calculation unit 11 determines whether there are any substrate groups among the plurality of substrate groups that have not been selected and for which the F-value average has not been calculated.

[0073] If there are any unselected substrate groups (S54: Yes), the analyzer 1 returns to S52. In S52, the calculation unit 11 selects one substrate group from the unselected substrate groups. By repeating S52 to S54, the F-value average is calculated for each substrate group. If there are no unselected substrate groups (S54: No), the analyzer 1 ends the substrate group processing in S5 and returns to the main process.

[0074] After S5 is completed, the analyzing device 1 outputs the change in the average F value corresponding to the order in which each substrate group was processed (S6). In S6, the computing unit 11 generates a graph showing the change in the average F value corresponding to the order in which each substrate group was processed by the processing device 2 based on the processing results in S5, and displays the generated graph on the display unit 16. Figure 12 This is a graph showing an example of changes in the average F value according to the order in which each substrate group is processed. Figure 12 The horizontal axis represents the classification of substrate groups, and the vertical axis represents the average F value.

[0075] Figure 12 The graph shown shows the change in the average F-value for one processing step and one type of measured value. The analyzer 1 can change the processing step and type of measured value for which the average F-value is displayed. For example, the analyzer 1 accepts user designation of a processing step and type of measured value by operating the operating unit 15, and outputs the change in the average F-value for the designated processing step and type of measured value. Graphs showing the change in the average F-value for multiple processing steps or multiple types of measured values ​​can also be displayed in an overlapping manner.

[0076] The timing at which the multiple substrates included in each substrate group are processed by processing apparatus 2 partially overlaps and is slightly staggered between substrate groups. The second substrate group is processed later than the first substrate group. Furthermore, during the processing period by processing apparatus 2, the first substrate group is processed earliest, while the other substrate groups are processed later. In other words, the change in the average F-value corresponding to each substrate group represents the change in the average F-value corresponding to the duration of processing in processing apparatus 2.

[0077] exist Figure 12 In the example shown, the average F-value associated with the first substrate group is large, while the average F-value associated with the second substrate group is smaller. Furthermore, the average F-value associated with the third, fourth, and fifth substrate groups is even smaller and barely fluctuates. This shows that at the beginning of processing in the processing device 2, the deviation in the measured values ​​is large, but as processing continues, the deviation in the measured values ​​decreases. Therefore, it is clear that the longer the processing is continued, the more stable the processing state becomes. In this way, by outputting the change in the average F-value corresponding to the order in which each substrate group is processed, the change in the processing state in each processing step corresponding to the time the processing in the processing device 2 continues is obtained. The analysis device 1 can also perform processing in S5 to S6 so that the substrates included in the substrate group are not repeated.

[0078] After S6 is completed, the analysis device 1 ends the processing. When a certain number of substrates have been processed by the processing device 2 and time-series data related to each substrate has been obtained, the analysis device 1 performs the processing of S1 to S6. For example, the analysis device 1 performs the processing of S1 to S6 periodically or each time a predetermined number of substrates are processed by the processing device 2. The analysis device 1 may also perform S4 and S5 in the reverse order. When the processing recipe consists of a single processing step, the analysis device 1 also performs the processing of S1 to S6. In addition, the processing of S5 to S6 may be omitted.

[0079] As described above, the analysis device 1 acquires time series data related to multiple substrates processed by the processing device 2, calculates the F-value average according to the processing step, and outputs the relationship between each processing step and the above-mentioned index value. The F-value average does not represent the instantaneous deviation of the measured values ​​between multiple substrates, but rather an index value that represents the magnitude of the non-instantaneous deviation of the measured values ​​generated during the processing according to the processing step. Therefore, the F-value average reflects the overall state of the processing performed according to the processing step. The state of the processing performed according to the processing step can be investigated based on the calculated F-value average. In addition, by calculating the F-value average for each processing step, the state of the processing performed according to the processing step can be investigated for each of the multiple processing steps included in the processing recipe. For example, the greater the deviation of the measured values ​​represented by the F-value average, the more unstable the processing performed according to the processing step. By adjusting the content of the processing step to reduce the F-value average, the stability of the processing performed according to the processing step is improved, and the processing recipe is improved.

[0080] <Implementation Method 2>

[0081] In the second embodiment, the analysis apparatus 1 analyzes substrate processing performed by a plurality of processing apparatuses 2 . Figure 13 This is a conceptual diagram illustrating an example configuration of an analysis system 100 according to Embodiment 2. Analysis system 100 includes multiple processing devices 2. For example, multiple processing devices 2 are multiple process chambers included in a single semiconductor manufacturing apparatus 20. For example, multiple processing devices 2 are process chambers included in each of multiple semiconductor manufacturing apparatuses 20. Sensors 3 provided in each processing device 2 are connected to analysis device 1. Each processing device 2 independently processes a substrate, and each sensor 3 inputs its measured value into analysis device 1.

[0082] The analyzing device 1 receives measurement values ​​input from a plurality of sensors 3 provided in a plurality of processing devices 2 using the interface unit 17, and the operation unit 11 stores the received measurement values ​​in the storage unit 14. The analyzing device 1 performs the processes of S1 to S4. In S1, the analyzing device 1 obtains time series data by storing a plurality of types of time series data consisting of a plurality of types of measurement values ​​measured by each sensor 3 in each processing device 2 in the storage unit 14. The time series data related to the substrates processed by different processing devices 2 is obtained by the analyzing device 1. The input device that inputs the measurement from the sensor 3 and the analyzing device 1 may also be different devices, and the analyzing device 1 may also perform the process of S1 by reading the time series data from the input device. The analyzing device 1 may also read the time series data from a plurality of input devices.

[0083] In S2, the analysis device 1 normalizes the measurement values ​​included in the time series data by processing step. When different processing devices 2 process substrates, offsets and gains may differ. Normalizing by processing step eliminates the effects of offsets and gains from the measurement values ​​associated with each processing step. Therefore, even when the time series data includes multiple substrates processed by different processing devices 2, multiple measurement values ​​associated with each processing step can be easily compared across the time series data. Instead of normalizing, the analysis device 1 can also normalize the measurement values ​​by processing step.

[0084] In S3, the analysis device 1 calculates, by processing step, an average F value representing the deviation in measurement values ​​between multiple substrates processed by different processing devices 2. In S5, the analysis device 1 performs substrate group processing to calculate the average F value on a substrate group comprising multiple substrates processed simultaneously in parallel by multiple processing devices 2. The substrate group in Embodiment 2 differs from the substrate group in Embodiment 1. In S51, the analysis device 1 creates a substrate group comprising multiple substrates processed simultaneously in parallel by multiple processing devices 2. In this case, the analysis device 1 creates multiple substrate groups processed at different times. Multiple substrates processed continuously by a processing device 2 are defined as a batch, and the multiple processing devices 2 sequentially process the multiple batches. For example, a single substrate included in a specific batch is aggregated across multiple processing devices 2 to create a single substrate group, and a substrate group is created for each batch, thereby creating multiple substrate groups. For example, a single substrate included in a specific batch is aggregated across multiple processing devices 2 to create a single substrate group, and a substrate group is created for each substrate, thereby creating multiple substrate groups. In S53, the analysis device 1 calculates the average F value for each substrate group. In S6, the analysis device 1 outputs the change in the average F value corresponding to the order in which each substrate group is processed. After S6 is completed, the analysis device 1 ends the processing. In addition, the processing of S5 to S6 can also be omitted.

[0085] As described above, the analysis device 1 acquires time-series data related to multiple substrates processed by multiple processing devices 2, calculates the average F-value associated with multiple substrates processed by different processing devices 2, and outputs the relationship between each processing step and the aforementioned index value. In the second embodiment, the status of the processing performed according to the processing step can also be investigated based on the F-value average. For example, since the F-value average is calculated based on time-series data related to multiple substrates processed by different processing devices 2, the influence of the processing device 2 on the F-value average becomes smaller, and the influence of the content of the processing step on the F-value average becomes larger. Therefore, in the second embodiment, the F-value average value further reflects the instability of the processing performed according to the processing step. Based on the F-value average, it is possible to reliably identify processing steps with low stability and effectively improve the processing recipe.

[0086] <Implementation Method 3>

[0087] Figure 14 as well as Figure 15 This is a conceptual diagram for explaining batches and time slots. The processed substrates are represented by circles. Figure 14 In FIG, a plurality of substrates processed are arranged in the processing order. A plurality of substrates processed in sequence by one or more processing devices 2 is defined as a batch. After the processing of one batch is completed, the processing of the next batch is performed. For example, Figure 14 As shown, after the first batch is processed, the second batch is processed, and then the third batch is processed. Between the completion of processing one batch and the start of processing the next batch, environmental changes may occur, such as a predetermined period of vacancy, operator rotation, or cleaning or maintenance of the processing apparatus 2. A single batch may include substrates processed by different processing apparatuses 2.

[0088] In order to distinguish each substrate in a batch, a time slot is used as an index value assigned to each substrate in the batch. A time slot is a number assigned in the order of processing, regardless of whether there are multiple substrates that can be included in a batch or whether there are actually processed substrates. Figure 15 As shown, each batch includes a substrate with time slot 1, a substrate with time slot 2, a substrate with time slot 3, .... Figure 15 An example in which m substrates are included in one batch is shown in . In the third embodiment, a process for verifying the stability of a process is performed between a plurality of substrates in different batches but in the same time slot, or between a plurality of substrates in the same batch but in different time slots.

[0089] Furthermore, in Embodiment 3, values ​​other than time slots may be used as index values. Time slots are also assigned to virtual substrates that are not actually processed within a batch. For example, consider the following example: within a batch, the first substrate is processed, the second substrate is not actually processed at the time it should be processed, and the third substrate is processed at the time it should be processed. Time slot 1 is assigned to the first substrate, time slot 2 is assigned to the virtual substrate that is not actually processed at the time it should be processed, and time slot 3 is assigned to the substrate that is processed at the time it should be processed. The index value may also be the actual order of processing. In the above example, processing order 1 is assigned to the first substrate, and processing order 2 is assigned to the substrate that is processed at the time it should be processed.

[0090] As the processing order, the "processing order within the batch" or the "processing order within the chamber" can be used. The processing order within the batch is a number assigned to each substrate in the order in which it is actually processed within the batch. The processing order within the chamber is a number assigned to each substrate in the order in which each process chamber is processed within the batch when substrates are processed by multiple process chambers (processing devices 2) within the batch. For example, consider the following example: in one batch, the first substrate is processed by the first process chamber, the second and third substrates are processed by the second process chamber, and the fourth substrate is processed by the first process chamber. In the order of the first, second, third, and fourth substrates, the "processing order within the batch" is 1, 2, 3, and 4. The "processing order within the chamber" associated with the first process chamber of the first substrate is 1, the "processing order within the chamber" associated with the second process chamber of the second and third substrates is 1 and 2, and the "processing order within the chamber" associated with the first process chamber of the fourth substrate is 2.

[0091] The configuration of the analysis system 100 is the same as that of the first or second embodiment. Figure 16 This is a flowchart showing an example of the steps of information processing performed by the analysis device 1 of embodiment 3. For example, after completing the processing of S1 to S6, or after completing the processing of S1 to S4, the analysis device 1 performs the following processing. The analysis device 1 selects the processing step and the type of measurement value (S71). In S71, the operation unit 11 accepts the designation of the processing step and the type of measurement value by the user operating the operation unit 15, selects the designated processing step from a plurality of processing steps, and selects the designated type of measurement value from a plurality of types of measurement values. For example, Figure 8 As shown, a graph showing the relationship between the processing step and the F-value average is displayed, and the user operates the operation unit 15 to input the designation of the processing step and the type of measurement value into the analyzer 1 .

[0092] Next, the analyzer 1 calculates the average F-value associated with the selected processing step and measurement type for each substrate group consisting of a specific plurality of substrates (S72). In S72, the calculation unit 11 identifies substrate groups consisting of substrates from different batches but the same time slot, and substrate groups consisting of substrates from the same batch but different time slots. Each batch is assigned a batch number according to the order in which it was processed. For example, the first batch is batch numbered 1, and the second batch is batch numbered 2.

[0093] The calculation unit 11 determines a substrate group by selecting multiple substrates with the same time slot from multiple batches with consecutive batch numbers. Alternatively, the calculation unit 11 determines a substrate group by selecting multiple substrates with consecutive time slots from the same batch. For example, each substrate group may include two substrates. The calculation unit 11 may also determine a substrate group including three or more substrates. The calculation unit 11 determines multiple substrate groups. For example, the calculation unit 11 determines a substrate group for all combinations of consecutive batch numbers and time slots, and for all combinations of batch numbers and consecutive time slots.

[0094] In S72, the calculation unit 11 calculates the average F value for the multiple substrates included in the substrate group. The data used for the calculation is a standardized or normalized value of the measurement value of the selected type obtained when the selected processing step is performed on each of the multiple substrates included in the substrate group. The calculation unit 11 calculates the average F value by performing the same processing as S3. The calculation unit 11 calculates the average F value for all substrate groups. The calculation unit 11 stores the average F value calculated for each substrate group in the storage unit 14. In S72, the calculation unit 11 can also calculate the average F value for a substrate group consisting of multiple substrates from different batches but with the same processing order, or a substrate group consisting of multiple substrates from the same batch but with different processing orders.

[0095] The analyzing device 1 outputs the relationship between the substrate group and the F value average ( S73 ). Figure 17 as well as Figure 18 This is a schematic diagram showing an example of a graph showing the relationship between a substrate group and an average F value. "***" included in the graph represents the value of the average F value. Figure 17 This table lists the average F-value values, associated with each substrate group consisting of multiple substrates from different batches but with the same time slot. In the figure, at the intersection of multiple batch number combinations and time slots, the average F-value values ​​for the substrate groups consisting of multiple substrates identified by these multiple batch numbers and time slots are displayed. Figure 18 This table lists the average F-value values, associated with each substrate group consisting of multiple substrates from the same batch but different time slots. In the figure, at the intersection of a batch number and multiple time slot combinations, the average F-value value for the substrate group consisting of multiple substrates identified by that batch number and multiple time slots is displayed.

[0096] In S71, if Figure 17 as well as Figure 18 As shown in FIG. 1 , the calculation unit 11 creates a table in which the average F-values ​​are arranged in association with each substrate group, and displays the table on the display unit 16. The calculation unit 11 creates two types of graphs. Figure 17 As shown in FIG, one type of graph is a graph showing the relationship between a substrate group consisting of a plurality of substrates from different batches and the same time slot and the average F value. Figure 18As shown, the other graph is a graph showing the relationship between a substrate group consisting of multiple substrates from the same batch but with different time slots and the average F value. For example, the calculation unit 11 displays both graphs simultaneously on the display unit 16. The calculation unit 11 may also display one graph on the display unit 16 and change the displayed graph to the other graph in response to a graph change instruction input by the user through the operation unit 15.

[0097] exist Figure 17 as well as Figure 18 In the graph shown, the color of the F-value average is different depending on the value. The larger the value, the more the color of the F-value average is emphasized. Figure 17 as well as Figure 18 The display color is represented by the density of the smear. The calculation unit 11 determines the display color based on the F-value average value and adjusts the display color. The method for determining the display color is the same as in Embodiment 1. The numerical value may vary depending on the density of the display color, the size or thickness of the text, the font of the text, or the thickness of the frame. The display method of the F-value average other than the display color may also vary depending on the numerical value. A table is displayed that associates substrate groups with F-value average values. The display method of the F-value average varies depending on the numerical value, making it easy to identify substrate groups with large F-value average values ​​and unstable processing.

[0098] exist Figure 17 In the figure, the F-value average is shown, which indicates the deviation of the measured values ​​between multiple substrates of different batches and the same time slot. The stability or instability of the processing across multiple batches is visualized. The measured values ​​are compared for the same time slot, so the influence caused by the difference in the time slot does not appear, and the instability of the processing caused by the difference in the batch is clear. The combination of consecutive batches in which the F-value average becomes large is clear, and the combination of batches in which the processing status of the substrate is unstable between batches is clear. For example, when the F-value average is large in the combination of batches with small batch numbers and the F-value average is smaller in the combination of batches with larger batch numbers, it is presumed that the processing status stabilizes as the processing of the substrate continues. For example, when the F-value average suddenly becomes large, it is presumed that some environmental changes have occurred between batches.

[0099] exist Figure 18In the figure, the F-value average, which represents the deviation in measured values ​​between multiple substrates from the same batch but different time slots, is shown. The stability or instability of processing across multiple time slots within a batch is visualized. By comparing measured values ​​for the same batch, the influence caused by batch differences is not apparent, and the instability of processing caused by time slot differences is clearly visible. Combinations of consecutive time slots in which the F-value average increases within the batch are clearly visible, and combinations of time slots in which the processing status of substrates between time slots is unstable are clearly visible. For example, if the F-value average is large in a combination of smaller time slots and small in a combination of larger time slots, it is inferred that the processing status stabilized as the processing of substrates within the batch continued. Furthermore, the time slot in which processing suddenly became unstable can be identified. In S73, the calculation unit 11 may also display on the display unit 16 a table in which the F-value average values ​​are arranged in association with a substrate group consisting of multiple substrates from different batches but with the same processing order, or a substrate group consisting of multiple substrates from the same batch but with different processing orders.

[0100] The analyzing device 1 calculates the deviation of the values ​​obtained by standardizing or normalizing the measured values ​​between the multiple substrates included in the substrate group according to the substrate group (S74). During the first period of executing the processing according to the processing step, the sensor 3 repeatedly performs measurement to obtain the measured values ​​at a predetermined interval. For each of the multiple measured values ​​obtained in the first period, a measurement number is assigned according to the order in which the measurements are made. In the first period, the measured value of measurement number 1 is first obtained, and then the measured value of measurement number 2 is obtained, and the measured values ​​are obtained successively. The increase in the measurement number corresponds to the time elapsed in the first period. In S74, the operation unit 11 calculates the deviation of the values ​​obtained by standardizing or normalizing the measured values ​​by calculating the difference between the values ​​obtained by standardizing or normalizing the measured values ​​with the same measurement number between the multiple substrates.

[0101] Figure 19 This is a graph showing an example of the results of calculating the deviation of the values ​​obtained by standardizing or normalizing the measured values ​​by substrate group. In one substrate group, the difference of the values ​​obtained by standardizing or normalizing the measured values ​​is calculated for each measurement number, and the difference of the values ​​obtained by standardizing or normalizing the measured values ​​is calculated for each substrate group. The "***" included in the figure represents the value of the difference of the values ​​obtained by standardizing or normalizing the measured values. Figure 19 In the example, the difference value is represented by associating the batch number, time slot, and measurement number that identify the substrate group. The calculation unit 11 calculates the difference between the values ​​obtained by standardizing or normalizing the measurement values ​​for all substrate groups and all measurement numbers. The larger the absolute value of the difference, the greater the deviation of the values ​​obtained by standardizing or normalizing the measurement values ​​between multiple substrates. Figure 19 As shown, the calculation unit 11 stores data indicating the calculated deviation in the storage unit 14 .

[0102] Next, the analysis device 1 clusters the plurality of substrate groups based on the calculated deviations ( S75 ). In S75 , the calculation unit 11 pre-processes the data representing the deviations to emphasize the measurement value numbers having large deviations in the measurement values. Figure 20 This is a conceptual diagram showing an example of data preprocessing. The calculation unit 11 aggregates the difference values ​​calculated for multiple substrate groups for each measurement number. In the diagram, the values ​​obtained by aggregated differential values ​​calculated for measurement numbers 1, 2, ..., across multiple substrate groups are represented as aggregated values ​​1, 2, ..., etc.

[0103] The calculation unit 11 normalizes the calculated total value so that the average is 0 and the standard deviation is 1. For example, the calculation unit 11 performs the same calculation as the formula (1) on the total value calculated for a plurality of measurement numbers, thereby normalizing the total value. In this calculation, x in the formula (1) is the total value, x is the average of the total value, σ is the standard deviation of the total value, and x is the standard deviation of the total value. st In the figure, the values ​​obtained by normalizing the total value 1, the total value 2, ... are represented as normalized value 1, normalized value 2, ...

[0104] The calculation unit 11 converts the calculated normalized value into a weight. For example, the calculation unit 11 converts the normalized value into a weight using a normalized exponential function. Let the number of measurement values ​​obtained in the first period be M, and the normalized value for measurement number i be x. i , set the weight to W i , the calculation unit 11 calculates the weight W according to the following formula (5) i .

[0105] [Formula 4]

[0106]

[0107] exist Figure 20 In the example, the weights obtained by converting the normalized value 1, the normalized value 2, etc. are expressed as weight 1, weight 2, etc. By using the normalized exponential function, the weights are positive values ​​and the sum of the weights is 1. The calculation unit 11 multiplies the weight calculated for each measurement number by the difference between the values ​​obtained by normalizing or normalizing the measurement values. Figure 20 As shown, each difference value calculated for measurement number 1 is multiplied by weight 1, each difference value calculated for measurement number 2 is multiplied by weight 2, and the same applies to measurement numbers 3 and onwards. By performing such preprocessing, values ​​indicating large deviations in measurement values ​​are converted to relatively larger values.

[0108] Next, the calculation unit 11 extracts a predetermined number of substrate groups from the plurality of substrate groups in descending order of average F values, for example, 100. The calculation unit 11 performs principal component analysis on the pre-processed data related to the plurality of extracted substrate groups. Figure 21 This is a conceptual diagram showing an example of the results of principal component analysis. The figure shows the extracted substrate groups as Substrate Group 1, Substrate Group 2, and so on. For each substrate group, the preprocessed values ​​obtained for multiple measurement numbers were converted into multiple principal components. The "**" in the figure indicates the principal component values ​​(principal component scores).

[0109] The calculation unit 11 clusters the substrate groups using the principal component values. For example, the calculation unit 11 performs clustering using the k-means method. In this case, the calculation unit 11 performs clustering using a predetermined number of principal components, such as the first principal component, the second principal component, and the third principal component. In this way, the calculation unit 11 performs clustering after performing dimensionality reduction. The calculation unit 11 may also perform clustering by performing dimensionality reduction using the UMAP (Uniform Manifold Approximation and Projection) method. Through clustering, the plurality of substrate groups are classified into a plurality of clusters, each of which includes substrate groups with similar deviations in measured values.

[0110] Next, the calculation unit 11 displays the distribution of the plurality of clustered substrate groups ( S76 ). In S76 , the calculation unit 11 creates an image representing the two-dimensional distribution of the plurality of clustered substrate groups and displays the created image on the display unit 16 . Figure 22 This is a schematic diagram showing an example of an image representing the two-dimensional distribution of multiple substrate groups. The horizontal axis represents the value of the first principal component, and the vertical axis represents the value of the second principal component. Circles in the figure represent substrate groups with first and second principal component values ​​corresponding to the two-dimensional coordinates. Clusters are distinguished by enclosing substrate groups within the same cluster with dotted lines. The calculation unit 11 may also display substrate groups within different clusters in different colors.

[0111] The calculation unit 11 changes the display size of the substrate group based on the average F-value. For example, the larger the average F-value, the larger the display size. In addition to the display size, the display format of the substrate group can also be changed based on the average F-value. For example, the display color density can be changed based on the average F-value, and the shape of the marking representing the substrate group can also be changed based on the average F-value. By changing the display format, substrate groups with large average F-values, i.e., those with large variations in measured values, can be clearly identified.

[0112] By displaying the two-dimensional distribution of clustered substrate groups, it becomes clear which groups exhibit the most measured value deviations. By changing the display format of the substrate groups based on the F-value average, it becomes clear which clusters of measured value deviations contain substrate groups with large measured value deviations. Furthermore, it becomes clear how many other substrate groups exist whose measured value deviations are similar to those of a specific substrate group. For example, the frequency of occurrence of specific anomalies can be investigated. In S76, the computing unit 11 may also display an image representing the three-dimensional distribution of the clustered plurality of substrate groups on the display unit 16.

[0113] Furthermore, during steps S74 to S76, the analyzer 1 may calculate the deviation of the values ​​obtained by standardizing or normalizing the measured values ​​using methods other than calculating differences. For example, the calculation unit 11 may calculate the deviation by calculating the variance of the values ​​obtained by standardizing or normalizing the measured values ​​for multiple substrates. Other methods may be used as data preprocessing methods. Alternatively, principal component analysis may be performed without data preprocessing. Furthermore, methods other than principal component analysis may be used as dimensionality reduction methods.

[0114] The analyzing device 1 selects one substrate group from the plurality of substrate groups (S77). In S77, the operation unit 11 receives an instruction to select one substrate group by operating the operation unit 15 by the user, and selects the designated substrate group from the plurality of substrate groups. For example, through the processing of S73, as shown in FIG. Figure 17 or Figure 18 As shown, when the table in which the average F-value values ​​are arranged in association with each substrate group is displayed on the display unit 16, the user operates the operation unit 15 to select any substrate group. For example, any substrate group with a large average F-value value is selected. Alternatively, through the processing of S76, as shown in FIG. Figure 22 As shown, in a state where the distribution of a plurality of clustered substrate groups is displayed, the user operates the operation unit 15 to select any substrate group, for example, any substrate group included in a specific cluster.

[0115] Next, the analyzer 1 displays the time-varying values ​​obtained by normalizing or standardizing the measured values ​​for the selected substrate group (S78). In S78, a graph showing the time-varying values ​​obtained by normalizing or standardizing the measured values ​​obtained during the first period for the plurality of substrates included in the substrate group is displayed. The calculation unit 11 creates the graph and displays the created graph on the display unit 16.

[0116] Figure 23: This is a schematic diagram showing an example of a graph showing time changes of values ​​obtained by standardizing or normalizing measured values. The horizontal axis in the graph represents the measurement number. The measurement number corresponds to the time elapsed in the first period. The vertical axis in the graph represents the value obtained by standardizing or normalizing the measured values. In the graph, the value obtained by standardizing or normalizing the measured values ​​is recorded only as the measured value. Figure 23 In the figure, white circles represent the standardized or normalized values ​​for one of the multiple substrates included in the substrate group, while black circles represent the standardized or normalized values ​​for the other substrates. This specifically illustrates the temporal changes in the measured values ​​during the first period, making it clear how the measured values ​​vary across the multiple substrates.

[0117] For a selected substrate group, the time-varying values, obtained by normalizing or standardizing the measured values, are displayed, making it clear at which instant in time the measured values ​​differ. Furthermore, for a selected substrate group, the time-varying values ​​for the measured values ​​across multiple substrates are clearly displayed. This information allows users to easily verify and contributes to process improvement.

[0118] After S78 is completed, the analysis device 1 ends the process. The analysis device 1 may repeatedly execute the processes of S77 and S78. For example, the analysis device 1 may select various substrate groups and display the time-varying values ​​obtained by normalizing or standardizing the measured values ​​for each substrate group. The processes of S71 to S78 do not necessarily need to be performed in their entirety. The processes of S73 may be omitted, the processes of S74 to S76 may be omitted, and the processes of S71 to S78 may be omitted.

[0119] In Embodiment 3, an example is shown in which both a substrate group consisting of multiple substrates from different batches but with the same index value and a substrate group consisting of multiple substrates from the same batch but with different index values ​​are used as substrate groups. The analyzer 1 may also process only one of the substrate groups.

[0120] In Embodiments 1 to 3, an example was shown in which the F-value average was used as an indicator value representing the deviation in measured values ​​between multiple substrates during the period of processing according to each processing step. The analyzing device 1 may also calculate an indicator value other than the F-value average. For example, the analyzing device 1 may calculate the variance, standard deviation, or coefficient of variation at multiple time points included in the first period and calculate the average of the variance, standard deviation, or coefficient of variation as the indicator value. The processing device 2 may also be a device that processes substrates other than semiconductor wafers, such as glass substrates or flat plate substrates.

[0121] In Embodiments 1 to 3, a method for acquiring multiple types of measurement values ​​using a single sensor 3 is described. Alternatively, the analysis system 100 may include multiple sensors 3 in the processing device 2 that measure different types of physical quantities, and use these multiple sensors 3 to acquire multiple types of measurement values. While Embodiments 1 to 3 illustrate examples in which the sensor 3 measures light intensity, the sensor 3 may also measure a physical quantity other than light intensity, such as temperature or pressure. Alternatively, the analysis system 100 may include multiple sensors 3 in the processing device 2 that measure the same type of physical quantity, and acquire the measurement values ​​measured by these multiple sensors 3 as multiple types of measurement values. For example, multiple sensors 3 that measure temperature at multiple locations within the processing device 2 may be installed in the processing device 2, and the temperatures at these multiple locations may be acquired as multiple types of measurement values. While Embodiments 1 to 3 illustrate an embodiment in which multiple types of measurement values ​​are acquired, a single sensor 3 that measures a single physical quantity may also be installed in the processing device 2, and the analysis device 1 may acquire a single type of measurement value.

[0122] The embodiments disclosed herein are to be considered in all respects as illustrative and non-restrictive. The embodiments described above may be omitted, replaced, or modified in various ways without departing from the scope of the claims and the spirit thereof.

[0123] The matters described in the various embodiments can be combined with each other. Furthermore, independent claims and dependent claims described in the claims can be combined with each other in all combinations, regardless of the format in which they are cited. Furthermore, the claims may use a format in which claims refer to two or more other claims (multiple claim format), but the present invention is not limited to this. Multiple claims may also be described in a format in which multiple claims refer to at least one multiple claim (multiple-reference-multiple claim).

[0124] Description of Reference Signs

[0125] 100 ...analysis system; 1 ...analysis device; 11 ...calculation unit; 14 ...storage unit; 141 ...computer program; 16 ...display unit; 20 ...semiconductor manufacturing apparatus; 2 ...processing device; 3 ...sensor.

Claims

1. A computer program that causes a computer to execute the following processing: Time series data consisting of a plurality of measurement values ​​measured by a sensor installed in a processing device is acquired, wherein The processing device processes the substrate according to one or more processing steps; calculating an index value for each processing step based on the plurality of time-series data acquired when processing the plurality of substrates, wherein the index value represents a deviation in measurement values ​​between the plurality of substrates during execution of each processing step; and Output the relationship between each processing step and the above indicator value.

2. The computer program according to claim 1, wherein Make the computer execute the following processing: For each processing step, a plurality of measurement values ​​included in the time series data related to each substrate and measured during execution of each processing step are standardized or normalized; The index values ​​are calculated using a plurality of standardized or normalized measured values.

3. The computer program according to claim 1, wherein Make the computer execute the following processing: A plurality of measurement values ​​measured in each of a plurality of second periods included in a first period during which each processing step is performed is defined as a group, and a ratio of a deviation of the measurement values ​​between the groups to a deviation of the measurement values ​​within the group is calculated for the plurality of groups regarding the plurality of substrates. A value obtained by averaging the plurality of ratios calculated in the plurality of second periods is calculated as the index value.

4. The computer program according to claim 1, wherein Make the computer execute the following processing: Acquiring multiple types of time series data consisting of multiple types of measurement values ​​measured by one or more sensors provided in the processing device, calculating the index value for each of the plurality of types of measurement values, The relationship between each of the plurality of types of measurement values, the respective processing steps, and the index value is output.

5. The computer program according to claim 4, wherein The aforementioned plurality of types of measured values ​​are values ​​obtained by measuring the intensities of light having different wavelengths. The computer program according to claim 4 , wherein: Make the computer execute the following processing: A plurality of combinations of processing steps and types of measurement values ​​are arranged and displayed in order of magnitude of the index values.

7. The computer program according to claim 4, wherein Make the computer execute the following processing: Displaying a table in which the index values ​​are arranged in association with each processing step and the measurement value of each of the plurality of types of measurement values, The display method of the above-mentioned indicator value is different according to the size of the above-mentioned indicator value.

8. The computer program according to claim 1, wherein Make the computer execute the following processing: Dividing a plurality of substrates processed by the same processing device into a plurality of substrate groups consisting of a plurality of substrates to be processed continuously, Calculate the above index values ​​according to substrate group. Output the relationship between each substrate group and the above index value.

9. The computer program according to claim 8, wherein Make the computer execute the following processing: A graph showing changes in the index value corresponding to the order in which the substrate groups are processed by the processing apparatus is displayed.

10. The computer program according to claim 8, wherein The plurality of substrates included in each substrate group partially overlap with those in other substrate groups.

11. The computer program according to claim 1, wherein Make the computer execute the following processing: The index value is calculated for each substrate group, wherein the substrate group is composed of a plurality of substrates from different batches with the same index value assigned to each substrate within the batch, or a plurality of substrates from the same batch with different index values. A table is displayed in which the index values ​​are arranged in association with each of the plurality of substrate groups. The display method of the above-mentioned indicator value is different according to the size of the above-mentioned indicator value.

12. The computer program according to claim 2, wherein Make the computer execute the following processing: calculating deviations of values ​​obtained by standardizing or normalizing a plurality of measurement values ​​measured during a first period of execution of one processing step between a plurality of substrates from different batches but with the same index value, or between a plurality of substrates from the same batch but with different index values; Based on the above deviation, a plurality of substrate groups consisting of a plurality of substrates from different batches but with the same index value, or a plurality of substrates from the same batch but with different index values ​​are clustered. Displays the distribution of multiple substrate groups after clustering.

13. The computer program according to claim 2, wherein: Make the computer execute the following processing: Select a processing step, Selecting one substrate group from each of a plurality of substrate groups consisting of a plurality of substrates from different batches but with the same index value, or a plurality of substrates from the same batch but with different index values, Temporal changes in values ​​obtained by normalizing or standardizing a plurality of measurement values ​​measured during a first period in which a selected process step is performed on a plurality of substrates included in a selected substrate group are displayed.

14. An analytical method, Time series data consisting of a plurality of measurement values ​​measured by a sensor installed in a processing device is acquired, wherein The processing device processes the substrate according to one or more processing steps; calculating an index value for each processing step based on the plurality of time-series data acquired when processing the plurality of substrates, wherein the index value represents a deviation in measurement values ​​between the plurality of substrates during execution of each processing step; and Output the relationship between each processing step and the above indicator value.

15. An analysis device, Equipped with a computing unit, The above-mentioned operation unit performs the following processing: Time series data consisting of a plurality of measurement values ​​measured by a sensor installed in a processing device is acquired, wherein The processing device processes the substrate according to one or more processing steps; calculating an index value for each processing step based on the plurality of time-series data acquired when processing the plurality of substrates, wherein the index value represents a deviation in measurement values ​​between the plurality of substrates during execution of each processing step; and Output the relationship between each processing step and the above indicator value.

Images