TW202405866A - Processing data analysis method and information processing device - Google Patents

Abstract

The invention provides an analysis method for processing data and an information processing apparatus, which can shorten the time required for analyzing the processing data and can stably analyze an appropriate part of the processing data. The processing data analysis method is used for analyzing processing data acquired during operation of a substrate processing apparatus. The processing data analysis method comprises the following steps: acquiring processing data; segmenting the processing data into a plurality of intervals in a time sequence; and diagnosing the state of health of the substrate processing apparatus on the basis of the processing data in a specific section among the divided plurality of sections. In the step of dividing into the plurality of sections, the start time and the end time of the plurality of sections are determined on the basis of the time at which one or more parameters selected from among the parameters of the plurality of sensors become predetermined prescribed values or the time at which the parameters change with respect to the prescribed values.

Description

Analysis methods and information processing devices for processing data

The present disclosure relates to an analysis method and an information processing device for processing data.

The substrate processing apparatus measures various parameters using multiple sensors during substrate processing, and stores processing data (log information) that correlates each of the various parameters with time. Then, when defects occur in substrate processing, the information processing device connected to the substrate processing device will find the cause of the abnormality based on the processing data.

For example, Patent Document 1 discloses a state prediction device (information processing device) having an algorithm for analyzing test data (process data). The state prediction device predicts the state of the plasma processing device using a first characteristic quantity of the processing data measured by a normal device and a second characteristic quantity of the processing data measured by the device to be evaluated.

In recent years, control contents have become increasingly complicated in order to perform more sophisticated substrate processing. Along with this, the number of parameters during the execution of the substrate processing also becomes huge, and it is necessary to diagnose the status of the device using a large number of parameters.

Patent Document 1: Japanese Patent Application Publication No. 2020-31096

The present disclosure provides a technology that can shorten the time required for analysis of processed data and stably analyze and process an appropriate part of the data.

According to an aspect of the present disclosure, a processing data analysis method is provided, which is a method of analyzing processing data obtained when a substrate processing device is operating; the processing data has caused each of multiple sensors of the substrate processing device to Data that establishes a correlation between a measured parameter and a measured time; the analysis method of the processed data has the following steps: a step of obtaining the processed data and storing it in a storage unit; reading the processed data from the storage unit and the process of dividing the processing data into multiple sections in a sequence according to the measured time; and diagnosing the substrate based on the processing data in a specific section among the divided sections. The process of processing the health status of the device; in the process of being divided into multiple sections, one or more of the parameters selected from a plurality of the sensors are used to become the default specified value. Set the start time and end time of multiple sections from the time or the time when the specified value changes.

According to this aspect, the time required for analysis of the processed data can be shortened and an appropriate portion of the data can be stably analyzed and processed.

Hereinafter, modes for implementing the present disclosure will be described with reference to the drawings. In each drawing, the same components may be designated by the same symbols to omit repeated explanations.

<Configuration of information processing system 100> FIG. 1 is a block diagram showing an example of an information processing system 100 having an information processing device according to an embodiment. As shown in FIG. 1 , the information processing system 100 has a plurality of FPD (Flat Panel Display) manufacturing devices 1 and a server 110 connected to each FPD manufacturing device 1 through a network 120 . In addition, the network 120 may also adopt various communication environments (wired LAN, wireless LAN, etc.) that can communicate between the FPD manufacturing device 1 and the server 110 .

The plurality of FPD manufacturing devices 1 is an example of a substrate processing device that processes a substrate. The FPD manufacturing apparatus 1 is provided with a plurality of sensors 8 that measure various parameters (physical quantities) in substrate processing and a device controller 9 that controls the substrate processing.

The FPD manufacturing device 1 may have the device controller 9 mounted on the device body, or may be configured such that the device controller 9 is disposed at another location and communicatively connected to the device body. The device controller 9 outputs a command for controlling the control components of the FPD manufacturing device 1 to the FPD manufacturing device 1 . Furthermore, the device controller 9 acquires the measurement value measured by each sensor 8 of the FPD manufacturing device 1 being controlled as a parameter.

Furthermore, the device controller 9 has a user interface function that provides information related to the FPD manufacturing device 1 to the user and accepts instructions for the FPD manufacturing device 1 from the operator. In addition, the information processing system 100 may also have a function as a user interface for each FPD manufacturing device 1 on the server 110 side. Furthermore, one device controller 9 may also have the function of communicating directly with the device controllers 9 of other FPD manufacturing devices 1 or communicating with the device controllers 9 of other FPD manufacturing devices 1 through the server 110 . Thereby, the device controller 9 can utilize information related to multiple FPD manufacturing devices 1 (parameters of substrate processing according to the same recipe, etc.).

Each device controller 9 communicates information with the server 110 through the network 120 . The server 110 manages the information related to each FPD manufacturing device 1 transmitted from each device controller 9, and sends and receives programs, recipes, etc. executed by each device controller 9.

In addition, the information processing system 100 shown in FIG. 1 is an example, and it goes without saying that there are various system configuration examples depending on the use or purpose. For example, the information processing system 100 may be provided with a controller that integrates the device controllers 9 of multiple FPD manufacturing devices 1 into one device controller, and the server 110 may be applied to this controller.

<Configuration of FPD manufacturing device 1> Next, an example of the FPD manufacturing apparatus 1 to which the above-mentioned information processing system 100 is applied will be described with reference to FIG. 2 . FIG. 2 is a schematic cross-sectional view showing the overall structure of the FPD manufacturing apparatus 1 . The FPD manufacturing apparatus 1 shown in FIG. 2 is an inductively coupled plasma (ICP) processing apparatus that performs various substrate processes on an FPD substrate G that is rectangular in plan view.

The FPD may be any of Liquid Crystal Display (LCD), Electro Luminescence (EL), Plasma Display Panel (PDP), etc. As the material of the substrate G, glass is mainly used, but transparent synthetic resin, etc. may also be used depending on the purpose. The substrate G may be a substrate on which a pattern of an electronic circuit or a light-emitting element is formed, or may be a support substrate. Examples of the substrate processing of the FPD manufacturing apparatus 1 include etching processing, film formation processing using a CVD (Chemical Vapor Deposition) method, and the like.

The FPD manufacturing apparatus 1 includes a rectangular parallelepiped box-shaped processing container 10 , a substrate mounting table 60 that is rectangular in plan view and has a substrate G placed in the processing container 10 , and the above-mentioned device controller 9 .

The processing container 10 is divided up and down by dielectric plates 11 . The antenna chamber as an upper space is formed by the upper chamber container 12 , and the processing chamber S as a lower space is formed by the lower chamber container 13 . The processing container 10 is provided with a rectangular frame-shaped support frame 14 at the boundary between the upper chamber container 12 and the lower chamber container 13, and the dielectric plate 11 is placed on the support frame 14 as a window member. In addition, the processing container 10 is grounded through the ground wire 13e.

The lower chamber container 13 has a loading/unloading port 13b for loading and unloading the substrate G on the side wall 13a, and has a gate valve 20 that opens and closes the loading/unloading port 13b. The lower chamber container 13 is adjacent to a transport module (not shown) equipped with a transport mechanism. The FPD manufacturing apparatus 1 opens the gate valve 20 and carries out the loading and unloading of the substrate G by the transport mechanism through the loading and unloading port 13b.

In addition, the lower chamber container 13 has a plurality of exhaust ports 13f on the bottom plate 13d. The gas exhaust part 50 is connected to the exhaust port 13f. The gas exhaust part 50 has a gas exhaust pipe 51 connected to the exhaust port 13f, and a pressure control valve 52 and an exhaust device 53 provided in the gas exhaust pipe 51. The exhaust device 53 has a turbomolecular pump, a vacuum pump, etc., and depressurizes the lower chamber container 13 during substrate processing.

A shower head 30 is provided at the upper end of the lower chamber container 13 and is composed of a plurality of elongated members and sprays processing gas into the processing chamber S. The shower head 30 also serves as a support beam for supporting the dielectric plate 11 . The shower head 30 is made of metal such as aluminum, and is surface-treated by anodizing. The shower head 30 has a gas flow channel 31 extending in the horizontal direction, and a plurality of gas ejection holes 32 communicating between the gas flow channel 31 and the processing chamber S.

The processing container 10 has a gas introduction pipe 45 connected to the gas flow channel 31 on the upper surface of the dielectric plate 11 . The gas introduction pipe 45 passes through the top 12 a of the upper chamber container 12 in an airtight manner and is connected to the processing gas supply part 40 . The processing gas supply unit 40 includes a gas supply pipe 41 connected to the gas introduction pipe 45, an on-off valve 42 and a flow controller 43 provided at a midway position of the gas supply pipe 41, and a processing gas supply source 44 that supplies the processing gas. The processing gas is supplied from the processing gas supply source 44 to the shower head 30 via the gas supply pipe 41 and the gas introduction pipe 45 , and is then ejected to the processing chamber S via the gas flow channel 31 and the gas ejection hole 32 .

The processing container 10 is provided with a high-frequency antenna 15 in a chamber container 12 forming an antenna chamber. The high-frequency antenna 15 is formed by winding an antenna 15a made of conductive metal such as copper into a ring or spiral shape.

The upper chamber container 12 has a power supply member 16 extending upward from the terminal of the antenna 15a. A power supply line 17 is connected to the upper end of the power supply member 16 . The power supply line 17 is connected to the high-frequency power supply 19 through a matching device 18 that performs impedance matching on the outside of the processing container 10 . The FPD manufacturing apparatus 1 supplies high-frequency power of, for example, 13.56 MHz from the high-frequency power supply 19 to the high-frequency antenna 15, thereby forming an induced electric field in the lower chamber container 13. By this induced electric field, the processing gas supplied from the shower head 30 to the processing chamber S is plasmaized to generate an inductively coupled plasma, and ions or neutral radicals in the plasma are supplied to the substrate G.

On the other hand, the substrate mounting table 60 includes the base material 63 and the electrostatic chuck 66 provided on the upper surface 63 a of the base material 63 . The base material 63 has approximately the same plane size as the substrate G and is formed into a rectangular shape in plan view. The base material 63 is made of stainless steel, aluminum, aluminum alloy, or the like. A meandering temperature-regulating medium flow channel 62a is provided inside the base material 63 . In addition, the temperature regulating medium flow channel 62a can also be provided in the electrostatic clamp 66.

A circulation pipe 62b that allows the temperature control medium to flow in and out of the temperature control medium flow channel 62a is connected to both ends of the temperature control medium flow channel 62a. The circulation pipe 62b is connected to the cooler 62c. The cooler 62c circulates a temperature control medium such as GALDEN (registered trademark) or Fluorinert (registered trademark). In addition, the base material 63 may have a built-in heater or the like and may be configured to perform temperature adjustment using the heater, or may be configured to perform temperature adjustment using both a temperature-adjusting medium and a heater. The base material 63 is provided with a temperature sensor such as a thermocouple, and the measured value (parameter) of the temperature sensor is transmitted to the device controller 9 . The device controller 9 controls the temperature adjustment action of the cooler 62c based on the transmitted parameters.

A plurality of (for example, 12) lift pins 78 (for example, 12 lift pins) are provided inside the substrate placement table 60 for collecting and transferring the substrate G to and from the transport mechanism (two lift pins 78 are representatively shown in FIG. 1 ). The plurality of lift pins 78 penetrate the substrate mounting table 60 and move up and down by the power of the motor transmitted through the connecting member. A base 68 formed of an insulating material and having a stepped portion on the inside is fixed to the bottom plate 13d of the lower chamber container 13. The base material 63 is placed on the stepped portion of the base 68, and an electrostatic jig 66 on which the substrate G is directly placed is provided on the upper surface of the base material 63.

The electrostatic clamp 66 has a ceramic layer 64 which is a dielectric coating film formed by spraying ceramics such as alumina, and an adsorption electrode 65 provided inside the ceramic layer 64 for electrostatic adsorption. The adsorption electrode 65 is connected to the DC power supply 75 through the power supply line 74 and the switch 76 . The device controller 9 applies a DC voltage from the DC power supply 75 to the adsorption electrode 65 by turning on the switch 76 . Thereby, the adsorption electrode 65 generates Coulomb force to electrostatically adsorb the substrate G to the electrostatic clamp 66 . In addition, when the switch 76 is turned off and the switch 77 provided on the ground line branched from the power supply line 74 is turned on, the charge stored in the adsorption electrode 65 will flow to the ground.

A rectangular frame-shaped focusing ring 69 made of ceramics such as alumina or quartz is mounted on the upper surface of the base 68 located on the outer periphery of the electrostatic jig 66 . The upper surface of the focus ring 69 is set lower than the upper surface of the electrostatic clamp 66 .

The power supply member 70 is connected to the lower surface of the base material 63 , and the power supply line 71 is connected to the lower end of the power supply member 70 . The power supply line 71 is connected to a bias source, that is, a high-frequency power supply 73 through a matching device 72 that performs impedance matching. The high-frequency power supply 73 supplies high-frequency power of, for example, 3.2 MHz to the substrate mounting table 60 . Thereby, the base material 63 attracts the ions generated in the processing chamber S to the substrate G.

That is, the high-frequency power supply 19 connected to the high-frequency antenna 15 is a source for generating plasma, and the high-frequency power supply 73 connected to the substrate mounting table 60 is a bias source that attracts generated ions to impart kinetic energy. Thereby, the FPD manufacturing apparatus 1 can independently generate plasma and control ion energy, thereby increasing the degree of freedom in substrate processing.

The sensor 8 of the FPD manufacturing apparatus 1 is installed at an appropriate location in the processing container 10 and is used for measurement when substrate processing or the like is performed. For example, the sensor 8 includes an RF source power sensor 81 for detecting the power supplied from the high-frequency power supply 19 and the power reflected from the inside of the processing container 10 , and an RF source power sensor 81 for detecting the power supplied from the high-frequency power supply 73 . and a power sensor 82 for RF bias of the power reflected from the inside of the processing container 10 . The RF source power sensor 81 is provided in the matching device 18 . The RF bias power sensor 82 is provided in the matching device 72 . The RF source power sensor 81 and the RF bias power sensor 82 measure power at a sampling interval of, for example, 0.1 seconds, and transmit the measured power to the device controller 9 and the like. In addition, the sensor 8 of the FPD manufacturing apparatus 1 includes, for example, a pressure sensor for detecting the pressure in the processing container 10, a temperature sensor for detecting the temperature of the base material 63, a Flow sensors for processing gas flow, etc.

The parameters (measured values) measured by each sensor 8 are transmitted to the device controller 9 and used in the control of each component by the device controller 9 . Furthermore, the parameters are transmitted from the device controller 9 to the server 110 via the network 120, and are used in the server 110 to manage the status of the FPD manufacturing device 1. In addition, the information processing system 100 may also be configured to directly transmit the parameters of each sensor 8 to the server 110 without passing through the device controller 9 .

＜Hardware structure of information processing device＞ The device controller 9 and server 110 of the information processing system 100 shown in Figure 1 are implemented by a computer 500 (information processing device) composed of hardware as shown in Figure 3, for example. FIG. 3 is a hardware configuration diagram showing an example of the computer 500.

The computer 500 in Figure 3 is equipped with an input device 501, an output device 502, an external I/F (interface) 503, a RAM (Random Access Memory) 504, a ROM (Read Only Memory) 505, a CPU (Central Processing Unit) 506, and a communication I /F507 and HDD508, etc., use bus B to connect each component to each other.

The input device 501 and the output device 502 constitute the above-mentioned user interface, for example, a touch panel may be used. Alternatively, the input device 501 may also use a keyboard, a mouse, a loudspeaker, etc., and the output device 502 may also use a display, a speaker, etc. External I/F503 is the interface with external devices. The computer 500 can read or write from external devices such as the external recording medium 503a through the external I/F 503. The communication I/F 507 is an interface for connecting the computer 500 to the network 120 .

RAM 504 is an example of a volatile semiconductor memory (storage device) that temporarily holds programs or data. ROM505 is an example of a non-volatile semiconductor memory (storage device) in which programs or data are installed. HDD508 is an example of a non-volatile storage device that stores programs, data, recipes, etc. The CPU 506 reads programs or data from storage devices such as the ROM 505 or the HDD 508 and performs operations on the RAM 504 to realize the overall control or function of the computer 500 .

The computer 500 receives the parameters measured by each sensor 8 during the substrate processing of the FPD manufacturing device 1, and stores the time measured in the computer 500 in a storage device (such as HDD 508) as the processing data D of the substrate processing. (log information). For example, the computer 500 stores the RF source advancing wave 83 and the RF source reflected wave 84 detected by the RF source power sensor 81 and the RF bias advancing wave 85 detected by the RF bias power sensor 82. Various parameters of the RF bias reflected wave 86.

FIG. 4 is a graph illustrating time changes of parameters of high-frequency power (RF source traveling wave 83, RF source reflected wave 84, RF bias traveling wave 85, RF bias reflected wave 86). In Figure 4, the horizontal axis is time and the vertical axis is high-frequency power.

For example, the RF bias traveling wave 85 shown in FIG. 4 suddenly rises from a state of approximately zero during substrate processing, and then becomes constant at a set power value (specified value). In addition, after a predetermined time has elapsed, the power value of the RF bias traveling wave 85 decreases and becomes constant at a lower power level (specified value). After a predetermined time elapses at this low power, the power value of the RF bias traveling wave 85 decreases and becomes almost zero.

On the other hand, since the RF bias reflected wave 86 is generated due to the impedance mismatch of the high-frequency power supply 19 when supplying high-frequency power, a waveform linked to the change of the RF bias traveling wave 85 is drawn. Specifically, the RF bias reflected wave 86 will cycle in amplitude as the RF bias traveling wave 85 rises, returning to almost zero when the RF bias traveling wave 85 stabilizes at a specified value. In addition, the RF bias reflected wave 86 repeats its amplitude again and returns to zero when the RF bias traveling wave 85 decreases.

In addition, the RF source traveling wave 83 rises sharply after the RF bias reflected wave 86 decreases, and stabilizes at a power value (specified value) that is larger than the RF bias traveling wave. Then, the RF source traveling wave 83 maintains the specified value and then drops sharply after a predetermined period.

On the other hand, since the RF source reflected wave 84 is generated due to the impedance mismatch of the high-frequency power supply 73 when supplying high-frequency power, a waveform linked to the change of the RF source traveling wave 83 is drawn. That is, the RF source reflected wave 84 will exhibit a waveform with repeated amplitude during the rise or fall of the RF source traveling wave.

<Monitoring the measured value of sensor 8> Then, the information processing system 100 performs a processing data analysis method. The processing data analysis method diagnoses the FPD manufacturing device 1 based on the measurement results measured by each sensor 8 during substrate processing and the substrate status of the substrate processing. sound state. Especially in this embodiment, the server 110 connected to multiple FPD manufacturing devices 1 collects the parameters of each sensor 8 in each FPD manufacturing device 1 to implement the processing data analysis method. In addition, the processing data analysis method may be implemented in the device controller 9 of the FPD manufacturing device 1 individually for the FPD manufacturing device 1 .

FIG. 5 is a block diagram showing the functional modules of the server 110 that implements the data analysis method. The CPU 506 (refer to FIG. 3 ) of the server 110 will operate the program stored in the storage device, thereby forming a functional unit as shown in FIG. 5 in the server 110 . Specifically, a processing data acquisition unit 111, a storage area 112, and a data analysis unit 113 are formed inside the server 110.

The processing data acquisition unit 111 obtains the processing data D stored in the device controller 9 through the network 120 and stores it in the storage area 112 . The processed data D are continuously stored in the device controller 9 as timing information that correlates the parameters of each sensor 8 with time. The storage area 112 is configured to have a large storage capacity in order to store the processing data D. The processing data acquisition unit 111 preferably marks the parameters of each sensor 8 in the plurality of FPD manufacturing devices 1 and stores them in the storage area 112 . In addition, the server 110 can also directly receive the parameters of each sensor 8 and form and store the processing data D that has been correlated with time in the server 110 .

The data analysis unit 113 is a functional unit that analyzes the processing data D stored in the storage area 112 and actually diagnoses the health status of the FPD manufacturing apparatus 1 . In particular, when the data analysis unit 113 analyzes and processes the data D, it divides the processed data D into multiple sections, thereby achieving processing efficiency.

That is, during the entire period of substrate processing in each FPD manufacturing apparatus 1, the parameters measured by each sensor 8 will become a huge amount. Assuming that a defect occurs in the substrate G during substrate processing, it would take a lot of time (man-hours) to confirm all parameters of each sensor 8 during the entire substrate processing period to investigate the cause of the defect. Therefore, in the analysis method of processing data D, predetermined rules for dividing the processed data D are set in advance in the recipe to divide the sections of the processed data D, and to divide the specific sections among the multiple sections. Extraction of processing data D (section of interest). Thereby, the server 110 can reduce the number of parameters to be confirmed, thereby shortening the time required for analysis.

Specifically, the data analysis unit 113 includes a segment setting unit 114 and a diagnosis unit 115 internally. The segment setting unit 114 has a function of extracting the processed data D of the segment divided into a plurality of segments in the sequence of processing the data D.

The processing data D can be divided into multiple sections according to the following division rules (A) ~ (C) set in the recipe, and the processing data D (parameters of multiple sensors 8) of specific sections can be extracted. (A) The first condition: Specify the time when the value of the processed data D of any sensor is larger, smaller, or the same as the preset specified value. (B) Section condition (second condition): Combine the first condition to specify the start time of the section and the end time of the section, and define a plurality of sections. (C) Extract processing data D that satisfies the range of the segment start time and segment end time.

Since the segments are divided according to the above-mentioned rules, the automatic segmentation setting unit 114a and the user segmentation setting unit 114b are formed inside the segment setting unit 114. The automatic partition setting unit 114a is a functional unit that automatically sets the partitions for processing the data D. The user division setting unit 114b is a functional unit that sets division sections of the processing data D by the user through the input device 501 and the output device 502 (see FIG. 3 ).

After receiving the trigger condition to start the analysis of the processing data, the automatic partition setting part 114a will gradually automatically set multiple sections of the processing data D according to the rules of the recipe read from the storage device. Examples of the trigger condition include receiving a code for performing analysis from the FPD manufacturing apparatus 1 or receiving a user's instruction for performing analysis through the input device 501 . For example, the device controller 9 of the FPD manufacturing device 1 may request analysis when an abnormal change is recognized through self-diagnosis. In addition, the user can perform analysis when defects occur in the substrate G after substrate processing. Alternatively, the trigger condition may be a time point when a predetermined monitoring period has elapsed or a time point when a predetermined number of substrate processes have been performed.

Hereinafter, as the processing data D, for the setting procedure of the high frequency power (RF source advancing wave 83, RF source reflected wave 84, RF bias advancing wave 85 and RF bias reflected wave 86) in the section, refer to FIG. 6 The processing flow is explained in detail. FIG. 6 is a flowchart showing the segmentation process of processing data D.

In the segment dividing process, the automatic segmentation setting unit 114a first identifies the time when the power value of the RF source advancing wave 83 or the power value of the RF bias advancing wave 85 changes by more than a predetermined value (becomes larger or smaller) (step S11 ). This processing applies to the division rule of (A) above. The "predetermined value" used to obtain the change of the processed data D is a preset value range, which is preferably determined according to the characteristics of each sensor 8, for example, it is better than being able to eliminate the noise generated by each sensor 8. High noise level.

Next, the automatic division setting unit 114a determines whether the power value of the high-frequency power has changed from zero (specified value) at the time specified in step S11 (step S12). Then, when the power value has changed from zero (step S12: YES), the process proceeds to step S13, and when the power value has changed from zero (step S12: NO), the process proceeds to step S17.

In step S13, the automatic division setting unit 114a sets the time when the power value starts to change from zero as the start time of the section. This processing applies to the division rule of (B) above. Thereby, the automatic division setting part 114a can include in one section the moment when the power value of the appropriate parameter (RF source traveling wave 83, RF bias traveling wave 85) rises from zero, that is, the power value changes from zero so that the parameter in a section that has become unstable.

After step S13, the automatic division setting unit 114a sets the time when either the RF source advancing wave 83 or the RF bias advancing wave 85 reaches a specified value and the RF source reflected wave 84 or the RF bias reflected wave 86 becomes zero. time (step S14). This processing also applies to the division rule (B) above. The "specified value" in this step is a value corresponding to the RF source power output by the high-frequency power supply 19 to the processing container 10, or a value corresponding to the RF bias power output by the high-frequency power supply 73. Therefore, it can be determined by the formula. Calculated using the values recorded in . Thereby, the automatic segmentation setting unit 114a can include the change range of the power value of the appropriate parameter from the change of the power value to the time when the power value stabilizes to the specified value in the section where the power value changes from zero causing the parameter to become unstable. .

Furthermore, the automatic division setting unit 114a sets the same time as step S14 as the start time of the next segment (step S15). This processing also applies to the division rule (B) above. Thereby, the automatic segmentation setting unit 114a can easily set the start time of the next segment.

After step S15, the automatic division setting unit 114a sets the time when the RF source advancing wave 83 or the RF bias advancing wave 85 changes to the specified value as the end time (step S16). This processing also applies to the division rule (B) above. Thereby, the automatic segmentation setting unit 114a can include the range of the power value of the appropriate parameter from the time when the power value stabilizes to the designated value to the time when the power value changes, in the section in which the parameter is in a stable state.

On the other hand, when the power value changes from other than zero, the automatic division setting unit 114a sets the time when the power value changes from other than zero as the start time of the section (step S17). This processing also applies to the division rule (B) above. When the power value changes from zero, the RF source reflected wave 84 or the RF bias reflected wave 86 hardly changes and remains in a zero state. Therefore, the automatic division setting unit 114a only needs to monitor the RF source advancing wave 83 or the RF bias advancing wave 85 to acquire the change time of each advancing wave.

After step S17, the automatic division setting unit 114a sets the time when the RF source advancing wave 83 or the RF bias advancing wave 85 reaches the specified value as the end time (step S18). This processing also applies to the division rule (B) above. Thereby, even when the power value changes from zero, the automatic division setting unit 114a can still include the range of the power value until the time when the power value reaches the designated value in one segment.

Then, after step S18, the automatic division setting unit 114a proceeds to step S19 in the same manner as step S16. After step S16 or S18, the automatic segmentation setting unit 114a determines whether the setting of the segments of the processing data D has been performed throughout the entire substrate processing period (step S19). Then, if it has not been performed during the entire substrate processing period (step S19: NO), the process returns to step S12 and the following same processing flow is repeated. On the other hand, if it has been performed during the entire substrate processing (step S19: YES), the process proceeds to step S20.

In step S20, the automatic segmentation setting unit 114a selects a specific segment (segment of interest) for extracting the processing data D among the multiple segments set by repeating steps S12 to S18. This processing applies to the division rule of (C) above. The area of interest can be set automatically or through the user's prior selection according to the content of the substrate processing and the state of the substrate G.

For example, when the substrate processing produces a defect on substantially the entire surface of the substrate G, since it is assumed that the supply of high-frequency power during the substrate processing will be lower than the target value, the automatic division setting unit 114a specifies that the output power value becomes stable at a specified value. section as the focus section. Furthermore, for example, based on receiving an abnormality code when the voltage rises or falls from the FPD manufacturing apparatus 1, the automatic segmentation setting unit 114a may specify an unstable state when the power value rises or falls as a section of interest.

Then, the automatic segmentation setting unit 114a extracts the parameters of each sensor 8 in the area of interest specified in step S20 (step S21). Thereby, the automatic segmentation setting unit 114a only provides automatically segmented segments and processing data D in the focused segment, thereby making the subsequent processing in the diagnosis unit 115 efficient.

Next, an example of the high-frequency power divided by the above-mentioned segment dividing process will be described using FIG. 7 . FIG. 7 is a diagram showing an example of dividing the parameters of high frequency power into multiple sections.

Each parameter of the high-frequency power is divided into sections A to sections G through section division processing. Here, in section A, time t1 is the start time, and time t2 is the end time. Time t1 is the start time when the RF bias traveling wave 85 suddenly rises from zero, and is specified in step S13 of FIG. 6 . Time t2 is the end time when the RF bias traveling wave 85 reaches a predetermined designated value and the RF bias reflected wave 86 becomes zero, and is specified in step S14 of FIG. 6 .

In section B, time t2 is the start time, and time t3 is the end time. Time t2 as the start point of section B is specified in step S15 of FIG. 6 . Time t3 is the end time when the RF bias traveling wave 85 falls, and is specified in step S16 of FIG. 6 .

In section C, time t3 is the start time and time t4 is the end time. Time t3 as the starting point of section C is specified in step S17 of FIG. 6 . Time t4 is specified in step S18 of FIG. 6 .

In section D, time t4 is the start time and time t5 is the end time. Time t4 as the start point of section D is specified in step S16 of FIG. 6 . Time t5 is specified in step S17 of FIG. 6 .

In addition, the RF source traveling wave can also be divided into segments in the same manner. That is, section E is a section for obtaining the change range when the RF source traveling wave 83 rises, time t6 is the start time, and time t7 is the end time. Section F is a section in which the RF source traveling wave 83 is stabilized in a stable range of a specified value. Time t7 is the starting time, and time t8 is the ending time. Section G is a section for obtaining the variation range of the RF source traveling wave 83 when it descends. Time t8 is the start time, and time t9 is the end time.

In this way, the automatic segmentation setting unit 114a can smoothly segment each parameter in the substrate processing into a segment in a stable state and a segment in an unstable state by performing the process flow of the segment segmentation process. In addition, of course, the automatic segmentation setting unit 114a can also segment the processing data D of other sensors 8 into segments of a changing range and a stable range through the same segmentation process.

Returning to FIG. 5 , the diagnosis unit 115 of the data analysis unit 113 calculates the health status of each area of interest based on the processing data D of the area of interest extracted by the area setting unit 114 to diagnose each FPD manufacturing device 1 of sound status. The "health status" of the device means, for example, digitizing the degree to which each parameter of the processing data D deviates from the reference value of the device in a normal state as a health value (an index indicating the health status), or determining based on the health value. Determine normal or abnormal. The "base value" at this time is preset when the device is shipped from the factory through experiments or simulations, or the parameters of a specific amount of processed data D regarded as normal are learned as guidance data during operation of the device. Set to appropriate value.

Therefore, the diagnosis unit 115 internally includes a health value calculation unit 115a and an abnormality determination unit 115b. For example, in order to calculate the health value based on the processing data D of the segment of interest selected by the segment setting unit 114, the health value calculation unit 115a performs the health value calculation process as shown in FIG. 8 . Figure 8 is a flowchart showing the soundness value calculation process.

In the health value calculation process, the health value calculation unit 115a obtains the processing data D of the attention segment selected by the segment setting unit 114 (step S31). At this time, the health value calculation unit 115a can obtain all parameters in the processing data D of the section of interest, and can also extract specific parameters based on the possibility of anomalies. For example, when it is estimated that there is an abnormality in the health status of the power system of the FPD manufacturing apparatus 1, the health value calculation unit 115a obtains parameters of the processing data D related to the power system.

Furthermore, the health value calculation unit 115a standardizes the obtained parameters of each sensor 8 in the area of interest to calculate health values respectively (step S32). In the calculation of this soundness value, for example, the following equation (1) is used and the characteristic amount V of each parameter of the standardized processed data D is applied.

Formula (1): V=(Da-X)/Y

Here, Da in Formula (1) is a parameter of the processed data D before standardization. X is the average of the characteristic quantities of all samples in the normal FPD manufacturing apparatus 1 . Y is the discrete value (or standard deviation) of all samples of the characteristic quantity in the normal FPD manufacturing apparatus 1.

The characteristic quantity V of the processed data D calculated by the above formula (1) corresponds to an index that digitizes the health state of the device indicating the degree of deviation from the normal state, that is, a health value. Therefore, if the health value is larger, it means that the device deviates more from the normal state; conversely, if the value is smaller, it can be considered that the device is closer to the normal state. Therefore, the health value calculation unit 115a calculates all the feature quantities V for each parameter of the processed data D of the section of interest in step S32 and stores them in the storage device. Through the above health value calculation process, the diagnosis unit 115 can obtain the health value (feature amount V) of the area of interest for each sensor 8 .

Furthermore, the abnormality determination unit 115b of the diagnosis unit 115 determines whether the FPD manufacturing apparatus 1 is normal or abnormal using the health value of each sensor 8 in the area of interest. For example, the abnormality determination unit 115 b compares the health value of the area of interest of each sensor 8 with a preset allowable range of the health value of each sensor 8 . Then, if the health value is within the allowable range, the abnormality determination unit 115b determines that there is no abnormality in the FPD manufacturing apparatus 1. On the contrary, if the health value is outside the allowable range, it is determined that the FPD manufacturing apparatus 1 has an abnormality.

In addition, if the health value exceeds the allowable range, the abnormality determination unit 115b will identify the sensor 8, thereby identifying which part of the device has an abnormality. For example, when the calculated health value in the section of interest of the RF bias advancing wave 85 exceeds the allowable range, the abnormality determination unit 115b specifies the structure related to the RF bias advancing wave 85 (power supply member 70, power supply line 71 , matching device 72, high-frequency power supply 73, etc.). Then, the diagnostic unit 115 notifies the user of the specified abnormality through the output device 502, thereby prompting the user to take necessary countermeasures.

In addition, the diagnosis unit 115 is not limited to the configuration of comparing the health value and the allowable range to determine whether the device is normal or abnormal. For example, the diagnosis unit 115 may directly display the health value on the output device 502 as the health state of the FPD manufacturing apparatus 1 . In the display of this health status, the health value can also be notified by making the health value equal to the display information that is easy for the user to understand (such as normal, alert, maintenance required, etc. image information). Furthermore, the information processing system 100 may take countermeasures such as stopping the operation of the FPD manufacturing apparatus 1 determined to be abnormal and not performing substrate processing.

The information processing system 100 related to this embodiment is basically configured as described above, and its operation (analysis method of processing data) will be described below.

The server 110 of the information processing system 100 communicates information with the device controller 9 of each FPD manufacturing device 1 to manage the status of each FPD manufacturing device 1 (whether or not it is operating, whether or not the substrate G is present, the operation content of the substrate processing, processing time, etc. ). Furthermore, the server 110 instructs the transport module installed adjacent to each FPD manufacturing apparatus 1 to carry out the loading and unloading of the substrate G into the processing chamber S of each FPD manufacturing apparatus 1 through the transport mechanism.

After storing the substrate G in the processing container 10, the device controller 9 of each FPD manufacturing apparatus 1 performs substrate processing on the substrate G. During substrate processing, each sensor 8 will continue to measure and transmit the measured values to the device controller 9 , whereby the device controller 9 can appropriately control each component based on the measured values of each sensor 8 . In addition, the device controller 9 will gradually store the processing data D that correlates the measured values (parameters) of each sensor 8 with time in the storage device.

Figure 9 is a flow chart showing an analysis method for processing data. After the substrate is processed by the FPD manufacturing device 1, as shown in FIG. 9, the server 110 will obtain the processing data D stored in the device controller 92 and store it in the storage area 112 of the server 110 (step S1).

Then, the server 110 determines whether the trigger condition for analyzing the processed data D has been established (step S2). As mentioned above, the trigger condition includes, for example, an analysis request from the FPD manufacturing apparatus 1 or an analysis request based on a user's operation.

The server 110 will enter a standby state when the trigger condition is not met. If no analysis is required, since the trigger condition will not be established, the processing will be terminated directly (step S2: NO). On the contrary, when the trigger condition is satisfied (step S2: YES), the server 110 shifts to the process of analyzing the processing data D of the target FPD manufacturing apparatus 1. The same is true when the trigger condition is met in the standby state.

At this time, the data analysis unit 113 performs a subroutine of segmentation processing on the processing data D of each sensor 8 stored in the storage area 112 (step S3). In the segmentation process, the automatic segmentation setting unit 114a divides the processing data D of each sensor 8 into multiple segments by performing the above-mentioned processing flow shown in FIG. 6, and further extracts the processing data of the segment of interest. D.

Next, the diagnosis unit 115 uses the processing data D of the area of interest extracted in step S3 to execute a subroutine of health value calculation processing for calculating the health value of the area of interest for each sensor 8 (step S4) . In the health value calculation process, the diagnosis unit 115 calculates the health value by performing the processing flow shown in FIG. 8 described above.

Then, the diagnosis unit 115 compares the calculated health value of the area of interest of each sensor 8 with the allowable range, and determines whether the health value is within the allowable range (step S5). If the health value is within the allowable range (step S5: YES), the process proceeds to step S6. Otherwise, if the health value exceeds the allowable range (step S5: NO), the process proceeds to step S7.

In step S6, the data analysis unit 113 notifies the FPD manufacturing device 1 that the processing data D has been analyzed through the output device 502 and that the FPD manufacturing device 1 is normal. At this time, the data analysis unit 113 may also perform processing such as displaying the analyzed processed data D or displaying the calculated health value of each sensor 8 .

On the other hand, in step S7, the data analysis unit 113 notifies the FPD manufacturing apparatus 1 that the processing data D has been analyzed through the output device 502 that there is an abnormality. At this time, the data analysis unit 113 may also display the specified abnormal location based on the processed data D of each sensor 8, or display the calculated health value of each sensor 8, etc. handle.

Then, when the processing flow of step S6 or S7 is completed, the server 110 ends the analysis method of processing the data D this time.

In addition, the analysis method and information processing device for processing data D related to the present disclosure are not limited to the above, and various modifications can also be adopted. For example, the analysis method of the processing data D can also analyze parameters such as temperature, pressure, and flow rate of the substrate processing apparatus in the same manner as the above-mentioned analysis of the high-frequency power. For example, the analysis method of the processing data D may also divide the processing data D in advance according to multiple steps in substrate processing, and perform settings of multiple sections and health status diagnosis for the processing data D of predetermined steps. Thereby, the time-consuming analysis of the processing data D can be further shortened. In addition, the order of step S1 and step S2 may be replaced. Specifically, for example, the analysis process in step S2 can be started when the trigger condition for analysis is established as the input condition, and the processing data D is obtained and stored in the storage area 112 . Thereafter, the processing after step S3 is the same as above.

The analysis method of processing data D can also use the time length of the sections divided into processing data D to diagnose the health status of the device. For example, when the time length of the unstable state section in the processing data D is longer than the reference time, it can be determined that an abnormality has occurred in the device. Furthermore, in the analysis method of the processed data D, the processing flow of the health value calculation process of calculating the health value from the divided processed data D is not limited to the example in FIG. 8 , and various processing flows may be adopted. For example, the diagnosis unit 115 may calculate the correlation coefficient between each parameter of the processed data D and the reference value corresponding thereto, and use the calculated correlation coefficient as a sound value.

The technical ideas and effects of the disclosure described in the above implementation modes are described below.

The first aspect of the present invention is an analysis method of processing data D, which is a method of analyzing the processing data D obtained when a substrate processing device (FPD manufacturing device 1) is operating; the processing data D has enabled the multi-function of the substrate processing device. Data that correlates the parameters measured by each sensor 8 with the measured time; the analysis method of the processed data D has the following steps: obtain the processed data D and store it in the storage unit (HDD508) Process; a process of reading the processing data D from the storage unit and dividing the processing data D into a plurality of sections in a sequence according to the measured time; and a process of dividing the processing data D into a plurality of sections according to a specific section ( The process of diagnosing the health status of the substrate processing device by focusing on the processing data D in the section); in the process divided into multiple sections, one or two or more parameters selected from the parameters of the multiple sensors 8 , set the start time and end time of multiple sections by the time when it becomes the default specified value or the time when it changes from the specified value.

According to the above, the analysis method of the processed data D is divided into multiple sections according to the time when the selected parameter becomes a specified value or changes from the specified value. Thereby, the method for analyzing the substrate processing device (FPD) can be simply limited. Processing data D of the health status of the manufacturing equipment 1). Thereby, the analysis method for processing the data D can shorten the time required for analyzing the data D and stably analyze and process an appropriate part of the data D. As a result, the analysis method of the processing data D can diagnose the health status of the substrate processing apparatus with high accuracy.

In addition, in a process divided into multiple sections, when one or more parameters change, the section is divided into sections containing parameters that become unstable due to changes, and sections containing only parameters after changes. A section of parameters that is in a stable state. Thereby, the analysis method of the processing data D can be divided into a section in which the parameters are in an unstable state and a section in which the parameters are in a stable state, so that the health status of the substrate processing apparatus (FPD manufacturing apparatus 1) can be well diagnosed according to the purpose.

In addition, an unstable state occurs when one or more parameters change from zero. In this way, the analysis method for processing the data D can be smoothly divided into a section of the transition phenomenon of the parameters when the parameters change from zero and a section where the parameters have reached stability.

Furthermore, the analysis method of processing data D uses the end time of one section as the start time of the next section following one section. Thereby, the analysis method for processing data D can simply set one section and the next section.

In addition, in the process divided into multiple sections, the time when the parameter becomes larger than the specified value by more than the preset numerical range, or the time when the parameter becomes smaller than the specified value by more than the preset numerical range is specified as a self-specified The moment when the value changes. Thereby, the analysis method of processing the data D can easily eliminate the noise generated by the sensor 8 to identify the moment when the parameter changes.

In addition, in the process of diagnosing the health status of the substrate processing apparatus (FPD manufacturing apparatus 1), a health value indicating the health status of the substrate processing apparatus is calculated based on the processing data D of a specific section (interest section). the value of the indicator. In this way, the health value is calculated based on the processed data D of the section of interest. Through this, the processing data analysis method can allow the user to understand the numerical health status.

Furthermore, in the process of diagnosing the health status of the substrate processing apparatus (FPD manufacturing apparatus 1), whether the calculated health value is within the set allowable range is compared. When the health value is within the allowable range, the substrate processing is determined. The device is normal. When the health value is not within the allowable range, the substrate processing device is determined to be abnormal. Thereby, the processing data analysis method can determine with high accuracy whether the substrate processing device is normal or abnormal based on the calculated health value.

In addition, the process data D are parameters measured by each of the plurality of sensors 8 when the substrate is processed in the substrate processing apparatus (FPD manufacturing apparatus 1). Thereby, the analysis method of the processing data D can use the parameters during substrate processing to effectively monitor the health status of the substrate processing device.

In addition, the substrate processing device (FPD manufacturing device 1) is a plasma processing device that performs plasma processing on the substrate; the plurality of sensors 8 include power sensors for the RF source that measure the RF source traveling wave and the RF source reflected wave as parameters. The detector 81 and the RF bias power sensor 82 that measure the RF bias traveling wave and the RF bias reflected wave as parameters; in the process divided into multiple sections, the RF source traveling wave or the RF bias traveling wave is used. The time when the wave changes from zero is regarded as the start time of the section, and the time when the RF source reflection wave or the RF bias reflection wave stabilizes to zero after the start time of the section is the end time of the section. Thereby, the analysis method of processing the data D can appropriately set the sections that divide the processing data D according to the RF source traveling wave, the RF source reflected wave, the RF bias traveling wave, and the RF bias reflected wave.

In addition, the second aspect of the present disclosure is an information processing device (device controller 9, server 110) that analyzes the processing data D obtained when the substrate processing device (FPD manufacturing device 1) is operating; the processing data D is Data that correlates the parameters measured by each of the plurality of sensors 8 of the substrate processing device with the measured time; the information processing device performs the following steps: obtains the processing data D and stores it in the storage unit The process of (HDD508); the process of reading the processed data D from the storage unit and dividing the processed data D into a plurality of sections in a sequence according to the measured time; and according to the process of dividing the plurality of sections. A process of diagnosing the health status of the substrate processing device by processing data D in a specific section (section of interest); in a process divided into multiple sections, one or two parameters selected from a plurality of sensors 8 are For more than one parameter, the start time and end time of multiple sections are set by the time when it becomes the default specified value or the time when it changes from the specified value. In this case, the information processing device can shorten the time required for analysis of the data D and stably analyze and process an appropriate part of the data D.

The analysis methods and information processing devices for processing data D related to the embodiments disclosed in this specification should be considered to be illustrative in all respects and not intended to limit the scope of the present invention. The implementation may be modified and improved in various forms without departing from the scope and gist of the patent application. The matters described in the plurality of embodiments described above may be constructed in other ways within the scope of non-inconsistency, and may be combined within the scope of non-inconsistency.

Of course, the substrate processing device of the present disclosure can be applied to a substrate processing device other than the FPD manufacturing device 1, such as a semiconductor wafer processing device. In addition, although the inductively coupled plasma (ICP) processing device using a dielectric plate as a window member has been described in this disclosure, an inductively coupled plasma (ICP) processing device using a metal plate as a window member may also be used instead of the dielectric plate. (ICP) processing device. In addition, it is not limited to the inductively coupled plasma (ICP) processing device. Examples of the substrate processing device include an Atomic Layer Deposition (ALD) device, a Capacitively Coupled Plasma (CCP) device, a Radial Line Slot Antenna (RLSA) device, and an Electron Cyclotron Resonance device. Plasma (ECR) device, Helicon Wave Plasma (HWP) device, etc.

1:FPD manufacturing equipment 8: Sensor 9:Device controller 110:Server 508: HDD D: Process data

FIG. 1 is a block diagram showing an example of an information processing system including an information processing device according to an embodiment. FIG. 2 is a schematic cross-sectional view showing the overall structure of the FPD manufacturing device. FIG. 3 is a hardware configuration diagram showing an example of a computer. FIG. 4 is a graph illustrating time changes of parameters of high-frequency power. Figure 5 is a block diagram showing functional modules of a server for implementing a data analysis method. FIG. 6 is a flowchart showing the segmentation process of processing data. FIG. 7 is a diagram showing an example of dividing the parameters of high frequency power into multiple sections. Figure 8 is a flowchart showing the soundness value calculation process. Figure 9 is a flow chart showing an analysis method for processing data.

83: RF source traveling wave

84: RF source reflected wave

85: RF bias traveling wave

86: RF bias reflected wave

Claims (10)

A method of analyzing processing data, which is a method of analyzing processing data obtained when a substrate processing device is operating; The processing data is data that correlates the measured parameters of each of the plurality of sensors of the substrate processing device with the measured time; The analysis method for processing data has the following steps: The process of obtaining the processed data and storing it in the storage department; The process of reading the processing data from the storage unit and dividing the processing data into a plurality of sections in a sequence according to the measured time; and The process of diagnosing the health status of the substrate processing device based on the processing data in a specific section among the divided sections; In a process divided into a plurality of sections, one or more parameters selected from a plurality of sensors are determined by the time at which they become the default specified value or the change from the specified value. time to set the start time and end time of multiple sections. For example, in the analysis method for processing data in Item 1 of the patent application scope, in the process that is divided into multiple sections, when one or more of the parameters change, it is divided into sections that include the changes accompanying the change. The section of that parameter that becomes unstable, and the section that contains only that parameter that is in a stable state after the change. For example, in the data processing analysis method of item 2 of the patent application, the unstable state occurs when one or more of the parameters change from zero. For the analysis method of processing data in any one of items 1 to 3 of the patent application scope, the end time of one such section is used as the starting time of the next section following one such section. If the data processing analysis method of any one of the patent scope items 1 to 3 is applied for, in the process divided into multiple sections, it is specified that the parameter becomes larger than the specified value and exceeds the preset value range. The time of change from the specified value, or the time when it becomes smaller than the specified value by more than a preset numerical range, is regarded as the time of change from the specified value. If the processing data analysis method in any one of items 1 to 3 of the patent scope is applied for, in the process of diagnosing the health status of the substrate processing device, the health value is calculated based on the processing data of the specific section, the The health value is a numerical value that indicates the health status of the substrate processing apparatus. For example, in the process data analysis method of Item 6 of the patent application, in the process of diagnosing the health status of the substrate processing device, whether the calculated health value is within the set allowable range is compared. When the health value is within If the health value is within the allowable range, the substrate processing device is determined to be normal. When the health value is not within the allowable range, the substrate processing device is determined to be abnormal. For example, the processing data analysis method of any one of items 1 to 3 of the patent application scope, wherein the processing data is a parameter measured by each of the plurality of sensors when the substrate processing is performed in the substrate processing device. For example, the processing data analysis method of any one of items 1 to 3 of the patent scope, wherein the substrate processing device is a plasma processing device that performs plasma processing on the substrate; A plurality of the sensors include an RF source power sensor that measures the RF source traveling wave and the RF source reflected wave as the parameter, and an RF bias sensor that measures the RF bias traveling wave and the RF bias reflected wave as the parameter. power sensor; In the process of dividing the section into multiple sections, the time when the RF source traveling wave or the RF bias traveling wave changes from zero is regarded as the starting time of the section, and the RF source time after the starting time of the section is The time when the source reflection wave or the RF bias reflection wave stabilizes to zero is regarded as the end time of this section. An information processing device that analyzes processing data obtained when a substrate processing device is operating; The processing data is data that correlates the measured parameters of each of the plurality of sensors of the substrate processing device with the measured time; The information processing device performs the following processes: The process of obtaining the processed data and storing it in the storage department; The process of reading the processing data from the storage unit and dividing the processing data into a plurality of sections in a sequence according to the measured time; and The process of diagnosing the health status of the substrate processing device based on the processing data in a specific section among the divided sections; In the process divided into multiple sections, one or more of the parameters selected from multiple sensors are determined by the time when they become the default specified value or the change from the specified value. time to set the start time and end time of multiple sections.

Images