Method for determining health index of substrate processing apparatus, and substrate processing apparatus

Abstract

To provide a substrate processing apparatus with multiple modules that makes it possible to grasp the state of health of the whole substrate processing apparatus.SOLUTION: A method for determining a health index of a substrate processing apparatus is provided. The substrate processing apparatus comprises multiple module groups, and each module group consists of one or more modules. The method comprises the steps of: obtaining multiple operation parameter values relating to each of the modules in the substrate processing apparatus; determining a module state of each of the modules, based on the multiple operation parameter values; and determining a health index of the substrate processing apparatus, based on the module states of the respective modules and weight coefficients for the respective modules.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to a method for determining a health index of a substrate processing apparatus and a substrate processing apparatus.

Background Art

[0002] Conventionally, a method for calculating the health of each motor of a transfer unit provided in a substrate processing apparatus is known (see, for example, Patent Document 1 (Claim 1)).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a substrate processing apparatus including a plurality of modules, it is desirable to grasp the health state of the entire substrate processing apparatus. By grasping the health state of the entire substrate processing apparatus, for example, it is possible to prevent the prolongation of the downtime of the substrate processing apparatus.

Means for Solving the Problems

[0005] [Aspect 1] According to Aspect 1, there is provided a method for determining a health index of a substrate processing apparatus, wherein the substrate processing apparatus includes a plurality of module groups, each module group is composed of one or more modules, and the method includes: for each module in the substrate processing apparatus, acquiring a plurality of operation parameter values respectively; determining a module state of each module based on the plurality of operation parameter values; and determining a health index of the substrate processing apparatus based on the module state of each module and a weight coefficient of each module.

[0006] [Form 2]According to Form 2, in the method of Form 1, the weight coefficient of each module includes a first weight coefficient representing a first type of contribution of the module to the health indicator, and a second weight coefficient representing a second type of contribution of the module to the health indicator.

[0007] [Form 3]According to Form 3, in the method of Form 2, the substrate processing apparatus is configured to perform a series of processes on the substrate by each module group performing its respective unique functional operation, the first weight coefficient is set to a value corresponding to the importance of the functional operation of each module group in the series of processes, and the second weight coefficient is set to a value according to the number of available modules included in each module group.

[0008] [Form 4]According to Form 4, in the method of Form 3, based on the module state of each module, steps of controlling the use or non-use of each module, and when any module is changed from use to non-use or from non-use to use, changing the second weight coefficient according to the change in the number of available modules are further included.

[0009] [Form 5]According to Form 5, in any one of the methods of Forms 2 to 4, the step of determining the health indicator includes, for each module, calculating the product of the value of the module state, the first weight coefficient, and the second weight coefficient.

[0010] [Form 6]According to Form 6, in the method of Form 5, the step of determining the health indicator includes calculating the sum of the products for each module over all modules.

[0011] [Form 7]According to Form 7, in the method of Form 1, steps of controlling the use or non-use of each module based on the module state of each module are further included.

[0012] [Aspect 8] According to Aspect 8, in the method of Aspect 7, it further includes the step of discretizing the plurality of operation parameter values of each module based on a predetermined threshold value respectively, and the step of determining the module state of each module includes calculating the sum of the plurality of discretized operation parameter values of the module or the sum of the plurality of operation parameter values of the module.

[0013] [Aspect 9] According to Aspect 9, in the method of Aspect 7 or 8, for the modules controlled to be unused among the respective modules, it further includes the step of identifying abnormal operation parameters based on the plurality of operation parameter values or the plurality of discretized operation parameter values.

[0014] [Aspect 10] According to Aspect 10, a method for determining a health index of a substrate processing apparatus, the substrate processing apparatus includes a plurality of module groups, each module group is composed of one or more modules, the method includes, for each module in the substrate processing apparatus, the step of respectively acquiring a plurality of operation parameter values; the step of training a learning model by machine learning so as to output a value related to the health index of the substrate processing apparatus when the plurality of operation parameter values of the plurality of modules of the substrate processing apparatus are input; and the step of estimating the current health index of the substrate processing apparatus from the plurality of current operation parameter values of the plurality of modules using the trained learning model. The step of training the learning model by machine learning includes calculating the sum of the plurality of operation parameter values of each module or the sum of the discretized values thereof to calculate the module state of each module; calculating the sum of the products of the module state of each module and the weight coefficient of each module to calculate the health index of the substrate processing apparatus; and training the learning model using the plurality of operation parameter values of the plurality of modules of the substrate processing apparatus and the calculated health index as training data. A method is provided.

[0015] [Embodiment 11] According to Embodiment 11, in the method of Embodiment 10, the trained learning model is installed in the substrate processing apparatus.

[0016] [Embodiment 12] There is provided a substrate processing apparatus including a control unit and a plurality of module groups. Each module group is composed of one or more modules. The control unit acquires a plurality of operation parameter values for each module in the substrate processing apparatus, determines the module state of each module based on the plurality of operation parameter values, and determines the health index of the substrate processing apparatus based on the module state of each module and the weight coefficient of each module.

[0017] [Embodiment 13] There is provided a substrate processing apparatus including a control unit and a plurality of module groups. Each module group is composed of one or more modules. The control unit acquires a plurality of operation parameter values for each module in the substrate processing apparatus, and uses a learning model trained by machine learning to output a value related to the health index of the substrate processing apparatus when the plurality of operation parameter values for the plurality of modules of the substrate processing apparatus are input, so as to estimate the current health index of the substrate processing apparatus from the plurality of operation parameter values of the current plurality of modules. The learning model calculates the module state of each module by calculating the sum of the plurality of operation parameter values for each module or the discretized values thereof, calculates the health index of the substrate processing apparatus by calculating the sum of the products of the module state of each module and the weight coefficient of each module, and trains the learning model using the plurality of operation parameter values for the plurality of modules of the substrate processing apparatus and the calculated health index as training data. The substrate processing apparatus is trained in this way.

[0018] [Form 14] According to Form 14, in the substrate processing apparatus of Form 13, the trained learning model is installed in the substrate processing apparatus.

Brief Description of the Drawings

[0019]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15A

Figure 15B

Figure 16

Mode for Carrying Out the Invention

[0020] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings described below, the same or corresponding components are denoted by the same reference numerals, and redundant descriptions are omitted.

[0021] FIG. 1 is an overall configuration diagram of a plating apparatus 100 according to an embodiment of the present invention. The plating apparatus 100 is an example of a substrate processing apparatus. Hereinafter, embodiments of the present invention will be described with reference to the plating apparatus 100. However, the apparatuses and methods disclosed in this specification can be applied to any substrate processing apparatus other than the plating apparatus. For example, in addition to the plating apparatus, the substrate processing apparatus includes a polishing apparatus (e.g., a CMP (Chemical Mechanical Polishing) apparatus, etc.) for polishing a substrate or a thin film formed on the substrate, a film forming apparatus (e.g., a CVD (Chemical Vapor Deposition) apparatus, an evaporation apparatus, etc.) for forming a thin film on the substrate, a lithography apparatus for transferring a fine pattern onto a thin film on the substrate, an etching apparatus for finely processing a substrate or a thin film on the substrate by etching, an ion implantation apparatus for implanting ions into a substrate or a thin film on the substrate, a dicing apparatus for cutting a substrate into chips, and various inspection apparatuses (or measuring apparatuses) for inspecting a substrate or a thin film on the substrate. The substrate (e.g., a semiconductor substrate, a glass substrate, a printed circuit board, etc.) includes any apparatus for processing or handling the substrate by any method.

[0022] Referring to FIG. 1, the plating apparatus 100 includes a load / unload unit 110 that loads a substrate onto a substrate holder (not shown) or unloads the substrate from the substrate holder, and a processing unit 120 that processes the substrate. The processing unit 120 further includes a pre-processing / post-processing unit 120A for performing pre-processing and post-processing of the substrate, and a plating processing unit 120B for performing plating processing on the substrate.

[0023] The load / unload unit 110 has a handling stage 26, a substrate transfer device 27, a fixing station 29, and a cleaning unit 50a. As an example, in the present embodiment, the load / unload unit 110 has two handling stages 26, namely, a loading handling stage 26A for handling the substrate before processing and an unloading handling stage 26B for handling the substrate after processing. In the present embodiment, the loading handling stage 26A and the unloading handling stage 26B have the same configuration and are arranged with their orientations different by 180° from each other. Note that the handling stage 26 is not limited to being provided with the loading and unloading handling stages 26A and 26B, and may be used without being distinguished as loading or unloading. Also, in the present embodiment, the load / unload unit 110 has two fixing stations 29. The two fixing stations 29 have the same mechanism, and the one that is available (the one not handling the substrate) is used. Note that one, or three or more of the handling stage 26 and the fixing station 29 may be provided according to the space in the plating apparatus 100.

[0024] On the handling stage 26 (loading handling stage 26A), substrates are conveyed from a plurality (three in the example of FIG. 1) of cassette tables 25 through the robot 24. The cassette table 25 includes a cassette 25a that houses the substrates. The cassette is, for example, a FOUP. The handling stage 26 is configured to adjust (align) the position and orientation of the placed substrate. A substrate transfer device 27 for transferring the substrate between these is arranged between the handling stage 26 and the fixing station 29. The substrate transfer device 27 is configured to transfer the substrate between the handling stage 26, the fixing station 29, and the cleaning unit 50a. Further, a stocker 30 for housing the substrate holder is provided in the vicinity of the fixing station 29.

[0025] The cleaning unit 50a has a spin rinse dryer 50 that cleans the substrate after plating and dries it by high-speed rotation. The substrate transfer device 27 is configured to transfer the substrate after plating to the spin rinse dryer 50 and take out the cleaned and dried substrate from the spin rinse dryer 50. Then, the substrate after cleaning and drying is passed by the substrate transfer device 27 to the handling stage 26 (unloading handling stage 26B) and returned to the cassette 25a through the robot 24.

[0026] The pre-treatment / post-treatment unit 120A has a pre-wet tank 32, a pre-soak tank 33, a pre-rinse tank 34, a blow tank 35, and a rinse tank 36. In the pre-wet tank 32, the substrate is immersed in pure water. In the pre-soak tank 33, the oxide film on the surface of the conductive layer such as the seed layer formed on the surface of the substrate is etched away. In the pre-rinse tank 34, the substrate after pre-soaking is cleaned with a cleaning liquid (such as pure water) together with the substrate holder. In the blow tank 35, after cleaning Liquid is drained from the substrate. In the rinse tank 36, the substrate after plating is cleaned with a cleaning liquid together with the substrate holder. Note that the configuration of this pretreatment / post-treatment unit 120A is an example, and the pretreatment / post-treatment unit 120A of the plating apparatus 100 is not limited to this configuration, and other configurations can be adopted.

[0027] The plating processing unit 120B is configured, for example, by housing a plurality of plating tanks 39 inside an overflow tank 38. Each plating tank 39 is configured to accommodate one substrate inside, immerse the substrate in the plating liquid held inside, and apply plating such as copper plating to the substrate surface.

[0028] The plating apparatus 100 has a transporter 37 that is located on the side of the pretreatment / post-treatment unit 120A and the plating processing unit 120B and transports the substrate holder together with the substrate, and adopts, for example, a linear motor system. This transporter 37 is configured to transport the substrate holder between the fixing station 29, the stocker 30, the pre-wet tank 32, the pre-soak tank 33, the pre-rinse tank 34, the blow tank 35, the rinse tank 36, and the plating tank 39.

[0029] An example of a series of plating processes by this plating apparatus 100 will be described. First, one substrate is taken out from the cassette 25a mounted on the cassette table 25 by the robot 24, and the substrate is transported to the handling stage 26 (the handling stage 26A for loading). The handling stage 26 adjusts the position and orientation of the transported substrate to a predetermined position and orientation. The substrate whose position and orientation have been adjusted by this handling stage 26 is transported to the fixing station 29 by the substrate transport device 27.

[0030] On the other hand, the substrate holder accommodated in the stocker 30 is transported to the fixing station 29 by the transporter 37 and horizontally placed on the fixing station 29. Then, the substrate transported by the substrate transport device 27 is placed on the substrate holder in this state, and the substrate and the substrate holder are connected.

[0031] Next, the substrate holder holding the substrate is gripped by the transporter 37 and stored in the pre-wet tank 32. Next, the substrate holder holding the substrate processed in the pre-wet tank 32 is transported by the transporter 37 to the pre-soak tank 33, and the oxide film on the substrate is etched in the pre-soak tank 33. Subsequently, the substrate holder holding this substrate is transported to the pre-rinse tank 34, and the surface of the substrate is washed with pure water stored in this pre-rinse tank 34.

[0032] The substrate holder holding the substrate after the water wash is transported from the pre-rinse tank 34 to the plating processing unit 120B by the transporter 37 and stored in the plating tank 39 filled with the plating solution. The transporter 37 sequentially repeats the above procedure to sequentially store the substrate holder holding the substrate in each plating tank 39 of the plating processing unit 120B.

[0033] In each plating tank 39, plating is performed on the surface of the substrate by applying a plating voltage between an anode (not shown) in the plating tank 39 and the substrate.

[0034] After the plating is completed, the substrate holder holding the plated substrate is gripped by the transporter 37, transported to the rinse tank 36, and immersed in the pure water stored in the rinse tank 36 to wash the surface of the substrate with pure water. Next, the substrate holder is transported to the blow tank 35 by the transporter 37, and water droplets adhering to the substrate holder are removed by blowing air or the like. Thereafter, the substrate holder is transported to the fixing station 29 by the transporter 37.

[0035] At the fixing station 29, the processed substrate is taken out from the substrate holder by the substrate transfer device 27 and transported to the spin rinse dryer 50 of the cleaning unit 50a. The spin rinse dryer 50 cleans the plated substrate and dries it by high-speed rotation. The dried substrate is passed to the handling stage 26 (the unloading handling stage 26B) by the substrate transfer device 27 and returned to the cassette 25a through the robot 24.

[0036] FIG. 2 is a schematic diagram for explaining "modules" and "module groups" in the plating apparatus 100 of the present embodiment. In the example of FIG. 2, the plating apparatus 100 includes a first module group 210, a second module group 220, a third module group 230, and a fourth module group 240. Also, in the example of FIG. 2, the first module group 210 includes one first module 2101, the second module group 220 includes two second modules 2201 to 2202, the third module group 230 includes ten third modules 2301 to 2310, and the fourth module group 240 includes two fourth modules 2401 to 2402. For example, the first module 2101 in the first module group 210 may correspond to the fixing station 29 described with reference to FIG. 1, each of the second modules 2201 to 2202 in the second module group 220 may correspond to the pre-wet tank 32 described with reference to FIG. 1, each of the third modules 2301 to 2310 in the third module group 230 may correspond to the plating tank 39 described with reference to FIG. 1, and each of the fourth modules 2401 to 2402 in the fourth module group 240 may correspond to the spin rinse dryer 50 described with reference to FIG. 1. That is, the example of FIG. 2 shows that the plating apparatus 100 includes a first module group 210 composed of one fixing station 29 (first module), a second module group 220 composed of two pre-wet tanks 32 (second modules), a third module group 230 composed of ten plating tanks 39 (third modules), and a fourth module group 240 composed of two spin rinse dryers 50 (fourth modules).

[0037] Note that FIG. 2 is to be understood as a diagram showing only four module groups 210, 220, 230, and 240 for the sake of convenience of explanation. The number of module groups may be arbitrary, and there may be three or fewer module groups, or five or more module groups. For example, in the plating apparatus 100 of FIG. 1, the pre-soak tank 33, the pre-rinse tank 34, the blow tank 35, the rinse tank 36, or other elements other than these may each constitute an additional fifth module group, a sixth module group, and so on.

[0038] Thus, the plating apparatus 100 includes a plurality of module groups, and each module group is composed of one or more modules. The number of modules included in each module group is not limited to the example of FIG. 2 and may be any number of one or more. Each module group performs a unique functional operation. For example, in the example of FIG. 2, the third module group 230 performs a predetermined plating process on the substrate by each plating tank 39 (the third modules 2301 to 2310) included therein. Also, the second module group 220 performs a process of immersing the substrate in pure water in each pre-wet tank 32 (the second modules 2201 to 2202) included therein.

[0039] In the example of FIG. 2, the third module group 230 includes ten plating tanks 39 (the third modules 2301 to 2310) having the same configuration, and the same process can be performed on the substrate in each tank. That is, the third module group 230 can process up to ten substrates in parallel simultaneously by using the ten modules 2301 to 2310 at the same time. Similarly, in the example of FIG. 2, the second module group 220 can perform the same process on the substrate in each of the two pre-wet tanks 32 (the second modules 2201 to 2202) having the same configuration. The same applies to the fourth module group 240.

[0040] FIG. 3 is a configuration diagram of an exemplary system 300 for implementing the method according to an embodiment of the present invention. The system 300 includes a plating apparatus 100 and a computer 320. The plating apparatus 100 is the plating apparatus described with reference to FIG. 1 or FIG. 2. The plating apparatus 100 and the computer 320 are communicably connected to each other via a network 330 such as a LAN (Local Area Network) or the Internet. Alternatively, the computer 320 may be incorporated into the plating apparatus 100 as part of the configuration of the plating apparatus 100. The computer 320 includes a processor 322 and a memory 324. A program 326 for implementing the method according to an embodiment of the present invention is stored in the memory 324. The processor 322 reads and executes the program 326 from the memory 324. Thereby, the system 300 can implement the method according to an embodiment of the present invention. Although only one computer 320 is shown in FIG. 3, the system 300 may include a plurality of computers 320. In such a configuration, programs corresponding to a part of the method according to an embodiment of the present invention are respectively stored in the memories 324 of the respective computers 320, and the processors 322 of the respective computers 320 individually execute those programs, so that the plurality of computers 320 may cooperate to implement the method according to an embodiment of the present invention as a whole.

[0041] <First Embodiment> FIG. 4 is a flowchart showing the operation of the system 300 for implementing the method according to an embodiment of the present invention. The processing of each step in the flowchart of FIG. 4 is executed by the processor 322 of the computer 320 in the system 300. The method according to the embodiment of FIG. 4 starts at step 402 during the normal operation of the plating apparatus 100.

[0042] First, in step 402, the processor 322 obtains a plurality of operation parameter values from each module of the plating apparatus 100 (for example, from each of the first module 2101, the second modules 2201 to 2202, the third modules 2301 to 2310, and the fourth modules 2401 to 2402 shown in FIG. 2). For example, in each module of the plating apparatus 100, various operation parameter values are measured at predetermined time intervals by various sensors, and they are sequentially provided from each module to the computer 320 of the system 300. The operation parameter value may be a numerical value of any type of parameter related to the operation and operating state of the module.

[0043] FIG. 5 shows an example of operation parameters for each module. Referring to FIG. 5, for example, the operation parameter values provided from the fourth modules 2401 to 2402 (spin rinse dryer 50) include the drive current value of the substrate high-speed rotation motor, the rotation position (angle) of the motor, and the supply rate of the cleaning liquid (pure water). The operation parameter values provided from the first module 2101 (fixing station 29) include the drive current value of the substrate holder lifting / rotating motor, the current speed of lifting / rotating, and the rotation position (angle) of the head. The operation parameter values provided from the second modules 2201 to 2202 (pre-wet tank 32) include the supply rate of the pre-wet injection liquid and the pre-wet injection liquid pressure. The operation parameter values provided from the third modules 2301 to 2310 (plating tank 39) include the plating liquid level, the plating liquid temperature, and the plating liquid stirring speed. It should be noted that these are merely examples of operation parameters, and the type and number thereof may be arbitrary.

[0044] Next, in step 404, the processor 322 discretizes each operation parameter value obtained in step 402 based on a predetermined threshold value. For example, when the obtained operation parameter value is within the first numerical range, the processor 322 converts the operation parameter value to the first value (for example, "1"), and when the obtained operation parameter value is within the second numerical range, the processor 322 converts the operation parameter value to the second value (for example, "2"), and when the obtained operation parameter value is within the third numerical range, the processor 322 converts the operation parameter value to the third value (for example, "3") to perform discretization. The upper and lower limit values of each numerical range, the respective converted values, and the number of numerical ranges to be set can be determined as appropriate, and for example, they may differ for each type of operation parameter.

[0045] FIG. 6 is a diagram for explaining an example of discretization, and shows a specific example of discretization for a plurality of operation parameter values obtained from the spin rinse dryer 50 (the fourth modules 2401 to 2402). In this example, for the first operation parameter value from the spin rinse dryer 50 (for example, the “drive current value of the substrate high-speed rotation motor” described above), when the value is in the range of the lower limit value “40” and the upper limit value “50”, it is converted to the value “2”, when the value is in the range of the lower limit value “50” and the upper limit value “60”, it is converted to the value “5”, when the value is in the range of the lower limit value “60” and the upper limit value “70”, it is converted to the value “4”, and when the value is in the range of the lower limit value “70” and the upper limit value “80”, it is converted to the value “2”. Further, for the second operation parameter value from the spin rinse dryer 50 (for example, the “rotation position of the motor” described above), when the value is in the range of the lower limit value “-30” and the upper limit value “-20”, it is converted to the value “1”, when the value is in the range of the lower limit value “-20” and the upper limit value “0”, it is converted to the value “5”, when the value is in the range of the lower limit value “0” and the upper limit value “10”, it is converted to the value “4”, and when the value is in the range of the lower limit value “10” and the upper limit value “20”, it is converted to the value “3”. Further, for the third operation parameter value from the spin rinse dryer 50 (for example, the “supply rate of the cleaning liquid” described above), when the value is in the range of the lower limit value “600” and the upper limit value “650”, it is converted to the value “2”, when the value is in the range of the lower limit value “650” and the upper limit value “750”, it is converted to the value “5”, and when the value is in the range of the lower limit value “750” and the upper limit value “800”, it is converted to the value “2”.

[0046] Note that the converted value may, for example, indicate that the larger the value, the better the operation parameter value. For example, the converted value “5” may mean that the operation parameter is the best, and the converted value “1” may mean that the operation parameter has deteriorated the most (for example, in an abnormal state).

[0047] FIG. 6 shows an example of discretization of the operation parameter values obtained from the fourth modules 2401 to 2402 (spin rinse dryer 50) of the plating apparatus 100. However, it will be understood that the discretization of the operation parameter values from other modules of the plating apparatus 100, such as the first module 2101 (fixing station 29), the second modules 2201 to 2202 (pre-wet tank 32), and the third modules 2301 to 2310 (plating tank 39), etc., is the same.

[0048] Next, in step 406, the processor 322 determines the module state of each module based on the plurality of operation parameter values of each module obtained in step 402 or based on the plurality of discretized operation parameter values of each module obtained in step 404. The module state is an indicator of whether the module is operating normally. The determination of the module state of each module may be based on, for example, the sum of the plurality of operation parameter values of the module obtained in step 402 or the sum of the plurality of discretized operation parameter values of the module obtained in step 404.

[0049] FIG. 7 is a diagram showing an example of calculation of the module state. This example represents an example of calculation of the module state based on the plurality of operation parameter values of the spin rinse dryer 50 (fourth module). Each row in the table of FIG. 7 corresponds to the operation parameter values obtained at different measurement timings. For example, referring to the data in the first row of FIG. 7, the first, second, and third operation parameter values acquired from the spin rinse dryer 50 at this measurement timing are "62", "2", and "712" respectively, and these are shown to be converted into the discretized operation parameter values "4", "4", and "5" (according to the table in FIG. 6). Then, "1 3", which is the sum of these discretized operation parameter values, is determined as the module state of the spin rinse dryer 50 (either of the fourth modules 2401, 2402) at this measurement timing. The same applies to the data in the second and subsequent rows of FIG. 7.

[0050] Note that, although FIG. 7 is an example regarding the fourth modules 2401 to 2402 (spin rinse dryer 50) of the plating apparatus 100, it will be understood that the same applies to other modules of the plating apparatus 100, such as the first module 2101 (fixing station 29), the second modules 2201 to 2202 (pre-wet tank 32), and the third modules 2301 to 2310 (plating tank 39), etc.

[0051] As shown in the rightmost column of the table in FIG. 7, the processor 322 may further normalize the value of the module state thus calculated by a predetermined method. For example, the normalization may be performed by converting the value such that the minimum value among a plurality of module states obtained for the module within a predetermined measurement period is set to "0" and the maximum value is set to "100".

[0052] Next, in step 408, the processor 322 determines the health index of the plating apparatus 100 based on the module state of each module obtained in step 406 and the weight coefficient of each module. The health index is an index that quantifies how normally the apparatus is operating. For example, the higher the numerical value, the higher the degree of normality of the operating state may be indicated. The plating apparatus 100 (generally a substrate processing apparatus) is composed of a plurality of module groups, and the importance of each module group in a series of processes performed by the plating apparatus 100 is different. Therefore, by assigning a weight coefficient corresponding to the importance of each module group of the plating apparatus 100 to each module and determining the health index in consideration of the weight coefficient, the health state of the plating apparatus 100 can be known with higher accuracy.

[0053] Figures 8 and 9 are diagrams showing an example of weight coefficients. Figure 8 shows an example of the first weight coefficient, and Figure 9 shows an example of the second weight coefficient. In the example of Figure 8, the first weight coefficient of the first module 2101 (i.e., the fixing station 29) belonging to the first module group 210 is set to "2", and the first weight coefficients of the second modules 2201 to 2202 (i.e., the pre-treatment tank 32) belonging to the second module group 220 are set to "3". The first weight coefficients of the third modules 2301 to 2310 (i.e., the plating tank 39) belonging to the third module group 230 are set to "5", and the first weight coefficients of the fourth modules 2401 to 2402 (i.e., the spin rinse dryer 50) belonging to the fourth module group 240 are set to "3". The first weight coefficient may indicate that the greater the value, the higher the importance of the functional operation of the module group in a series of processes performed by the plating apparatus 100. For example, in the plating apparatus 100, the plating process performed in the plating tank 39 is the most important, and since it is most necessary to consider whether each plating tank 39 is operating normally in the health index of the plating apparatus 100, in the example of Figure 8, the maximum value "5" is given to the first weight coefficient of the third modules 2301 to 2310 (plating tank 39).

[0054] On the other hand, the second weight coefficient is set to a value corresponding to the number of available modules included in each module group. For example, the second weight coefficient may be set to a larger value as the number of available modules included in each module group is smaller. This is because in a module group with a smaller number of available modules, the importance of each module relative to the entire module group is relatively higher compared to a module group with a larger number of available modules. Therefore, when calculating the health index, it is desirable to increase the weight of the modules belonging to such a module group with a smaller number of available modules. In the example of Figure 9, also referring to the example of Figure 2, the first module group 210 includes one first module 2101 (i.e., the fixing station 29), and the second module group 220 includes two second modules 2201 to 2202 (i.e., the pre-wet tank 32), the third module group 230 includes ten third modules 2301 to 2310 (i.e., the plating tank 39), and the fourth module group 240 includes two fourth modules 2401 to 2402 (i.e., the spin rinse dryer 50). Therefore, for the first module group 210 (the fixing station 29) with the fewest available modules, the maximum value "5" is assigned to the second weight coefficient, and for the third module group 230 (the plating tank 39) with the most available modules, the minimum value "3" is assigned to the second weight coefficient. Also, for the second module group 220 (the pre-wet tank 32) and the fourth module group 240 (the spin rinse dryer 50) with intermediate numbers of available modules, the intermediate value "4" is assigned to the second weight coefficient.

[0055] Figure 10 shows an example of the module state and the weight coefficient for each module of the plating apparatus 100. In Figure 10, the module state of each module is obtained in step 406 respectively. Also, the first and second weight coefficients shown in Figure 10 are the same as the examples shown in Figures 8 and 9. The processor 322 can calculate the machine health index MHI (Machine Health Index) of the plating apparatus 100, for example, according to the following formula. However, in the following formula, MC i is the module state of the i-th module, W1 i is the first weight coefficient of the i-th module, W2 i is the second weight coefficient of the i-th module, N is the total number of modules included in the plating apparatus 100 (for example, in the example of Figure 2, N = 15), and the sum (Σ) is performed for all the modules included in the plating apparatus 100.

[0056]

Equation

[0057] Note that the formula for the above health index MHI is one exemplary formula for calculating the health index of the plating apparatus 100, and it should be noted that the present invention is not limited thereto. For example, in the above formula, the health index of the plating apparatus 100 may be defined by a formula in which either the first weighting coefficient or the second weighting coefficient is omitted.

[0058] Next, in step 410, the processor 322 controls the use or non - use of each module of the plating apparatus 100 based on the module state of each module obtained in step 406. For example, as described above, the larger the value of the operation parameter value discretized in step 404, the better the operation parameter. Therefore, a module with a high module state value is operating normally, while a module with a low module state value can be considered to have a deteriorated operating state. Thus, the processor 322 controls, for example, a module with a module state value greater than a predetermined threshold to the use mode, and controls a module with a module state value less than the predetermined threshold to the non - use mode. Note that the "use mode" may be a mode in which the use of the module is permitted, while the "non - use mode" may be a mode in which the use of the module is prohibited or a mode in which the operation of the module is stopped.

[0059] FIG. 11 is an example of the determination result of use / non - use for each module of the plating apparatus 100. In this example, the predetermined threshold is set to "50", and two modules, the plating tank #5 and the spin - rinse dryer #1, are controlled to the non - use mode, and the other modules are controlled to the use mode. As a result of this control, the number of available modules in the third module group 230 (plating tank 39) is "9", which is less than when all (10) plating tanks 39 are in the use mode and the number of available modules in the fourth module group 240 (spin - rinse dryer 50) is "1", which is less than when all (2) spin - rinse dryers 50 are in the use mode.

[0060] In step 410, the processor 322 may display, on the monitoring screen of the computer 320, for example, the name of the module controlled to the unused mode and all operation parameter values or abnormal operation parameter values of the module. FIG. 12 shows an example of the monitoring screen. With such a monitoring screen, the user of the system 300 can quickly and accurately grasp which module has what kind of problem.

[0061] Next, in step 412, the processor 322 determines whether there is a module that has been changed from the used mode to the unused mode or from the unused mode to the used mode under the control of step 410. If there is no such module, the process returns to step 402 and the above-described process is repeated.

[0062] On the other hand, if there is a module whose mode has been changed, the process proceeds to step 414, and the processor 322 updates the second weight coefficient described above based on the current number of available modules in each module group in the plating apparatus 100. Thereafter, the processes after step 402 are repeated again. For example, referring to the example of FIG. 9, the number of available modules in the third module group 230 (plating bath 39) is "10", and the corresponding second weight coefficient is "3". However, if the number of available modules in the third module group 230 changes to, for example, "8" under the control of step 410, in step 414, the processor 322 may change the second weight coefficient corresponding to the third module group 230 from "3" to, for example, "4". In this way, by adjusting the second weight coefficient according to the actual number of available modules in each module group, a more accurate health index reflecting the operating status of each module of the plating apparatus 100 can be calculated.

[0063] <Second Embodiment> FIG. 13 and FIG. 14 are flowcharts showing the operation of a system 300 for implementing a method according to another embodiment of the present invention. The processing of each step in the flowcharts of FIGS. 13 and 14 is executed by a processor 322 of a computer 320 in the system 300. The method according to this embodiment uses a learning model trained by machine learning to estimate (determine) the health index of the plating apparatus 100. FIG. 13 shows a flowchart for a training phase of training the learning model, and FIG. 14 shows a flowchart for an operation (inference) phase of estimating the health index using the trained learning model.

[0064] Referring to FIG. 13, the training phase in this embodiment includes steps 1302 to 1310 of creating training data. Among these, steps 1302 to 1308 are the same as steps 402 to 408 in the method of the first embodiment described with reference to FIG. 4. That is, in step 1302, the processor 322 acquires a plurality of operation parameter values from each module of the plating apparatus 100. In step 1304, the processor 322 discretizes each operation parameter value acquired in step 1302 based on a predetermined threshold value. In step 1306, the processor 322 determines the module state of each module based on the plurality of operation parameter values of each module obtained in step 1302, or based on the plurality of discretized operation parameter values of each module obtained in step 1304. In step 1308, the processor 322 determines the health index of the plating apparatus 100 based on the module state of each module obtained in step 1306 and the weight coefficient of each module. These steps 1302 to 1308 may be performed, for example, during the trial operation of the plating apparatus 100 (for example, by introducing a test substrate into the plating apparatus 100 and operating each module), or it may be determined that a certain period during the period of actual operation of the plating apparatus 100 with the substrate for the actual product is the training data creation period, and steps 1302 to 1308 may be implemented within that training data creation period.

[0065] In step 1310 following step 1308, the processor 322 sets the operation parameter values of each module of the plating apparatus 100 acquired in step 1302 (for example, all the operation parameter values of all the modules) and the health index of the plating apparatus 100 calculated in step 1308 from those operation parameter values as a set, and uses it as a set of training data. Then, steps 1302 to 1310 are repeated until a sufficient number of training data is collected (step 1312: No). Thereby, training data consisting of a large number of sets of operation parameter values and health indexes is created.

[0066] Figure 15A shows an example of the created training data. Each row of the table shown in Figure 15A represents a set of training data obtained by performing one cycle of steps 1302 to 1310. Note that the training data may include the number of available modules in each module group, as shown in Figure 15B.

[0067] When a sufficient number of training data is collected by repeating steps 1302 to 1310 (step 1312: Yes), next in step 1314, the processor 322 trains a learning model by machine learning using those training data.

[0068] FIG. 16 is a diagram showing the configuration of an exemplary learning model that can be used in the implementation of the method according to the present embodiment. This exemplary learning model 1600 is composed of a neural network including an input layer 1602 having a plurality of input nodes 1601, a hidden layer (intermediate layer) 1604 composed of one or more layers each having a plurality of nodes 1603, and an output layer 1606 having one output node 1605. Each node is connected to a plurality of nodes in the layer adjacent to the layer to which the node belongs with an intensity characterized by a weighted parameter. To the plurality of nodes 1603 in the input layer 1602, operation parameter values of each module of the plating apparatus 100 (for example, all operation parameter values of all modules) among the training data are input. Further, from the output node 1605 of the output layer 1606, an estimated value for the health index of the plating apparatus 100 is output. The weighted parameters between the nodes of the neural network 1600 are adjusted so that this estimated value matches the value of the health index in the training data (that is, by using the value of the health index in the training data as the correct label for the estimated value). The parameter adjustment of the neural network 1600 is repeatedly performed using a large number of sets of training data (see FIGS. 15A and 15B) obtained in step 1310. Thereby, the training of the learning model 1600 is performed.

[0069] Next, with reference to FIG. 14, the operation (inference) phase in the present embodiment will be described. The operation phase in FIG. 14 starts at step 1402 during the normal operation of the plating apparatus 100.

[0070] In step 1402, the processor 322 acquires a plurality of operation parameter values from each module of the plating apparatus 100 (for example, from each of the first module 2101, the second modules 2201 to 2202, the third modules 2301 to 2310, and the fourth modules 2401 to 2402 shown in FIG. 2). Step 1402 is the same as step 402 in the method of the first embodiment described with reference to FIG. 4, and redundant explanations are omitted.

[0071] In the subsequent step 1404, the processor 322 inputs the plurality of operation parameter values from each module of the plating apparatus 100 obtained in step 1402 into the input layer 1602 of the trained learning model 1600 in the aforementioned training phase.

[0072] In step 1406, the processor 322 operates the learning model 1600 to output an estimated value for the health index of the plating apparatus 100 from the output layer 1606 (output node 1605) of the learning model 1600. Thus, according to the method according to this embodiment, by using a learning model trained by machine learning, the computational load for obtaining the health index of the plating apparatus 100 can be reduced, and a more accurate health index can be obtained more quickly.

[0073] As described above, embodiments of the present invention have been described based on several examples. However, the above-described embodiments of the invention are for facilitating the understanding of the present invention and do not limit the present invention. The present invention can be changed and improved without departing from its gist, and it goes without saying that equivalents of the present invention are included. Also, within the scope of solving at least a part of the above-described problems or achieving at least a part of the effects, any combination or omission of each component described in the claims and the specification is possible.

Description of Reference Numerals

[0074] 24 Robot 25 Cassette Table 25a Cassette 26 Handling Stage 27 Substrate Transfer Device 29 Fixing Station 30 Stocker 32 Pre-wet Tank 33 Presoak Tank 34 Prerinse Tank 35 Blow Tank 36 Rinse Tank 37 Transport​ 38 Overflow tank 39 Plating tank 50 Spin rinse dryer 50a Cleaning unit 100 Plating apparatus 110 Load / unload unit 120 Processing unit 120A Pretreatment / post-treatment unit 120B Plating processing unit 210 First module group 220 Second module group 230 Third module group 240 Fourth module group 2101 First module 2201 - 2202 Second module 2301 - 2310 Third module 2401 - 2402 Fourth module 300 System 320 Computer 322 Processor 324 Memory 326 Program 330 Network 1600 Learning model 1601 Input node 1602 Input layer 1603 Node 1604 Hidden layer 1605 Output node 1606 Output layer

Claims

1. A method for determining a health indicator of a substrate processing apparatus, wherein the substrate processing apparatus comprises a plurality of module groups, each module group is composed of one or more modules, and the method comprises: for each module in the substrate processing apparatus, obtaining a plurality of operation parameter values respectively; determining a module state of each module based on the plurality of operation parameter values; determining a health indicator of the substrate processing apparatus based on the module state of each module and a weight coefficient of each module; A method comprising the above.

2. The weight coefficient of each module includes a first weight coefficient representing a first type of contribution of the module to the health indicator and a second weight coefficient representing a second type of contribution of the module to the health indicator, according to the method described in Claim 1.

3. The substrate processing apparatus is configured such that each module group performs a series of processes on a substrate by performing respective unique functional operations, the first weight coefficient is set to a value corresponding to the importance of the functional operation of each module group in the series of processes, the second weight coefficient is set to a value according to the number of available modules included in each module group, The method described in Claim 2.

4. controlling the use or non-use of each module based on the module state of each module; when any module is changed from use to non-use or from non-use to use, changing the second weight coefficient according to the change in the number of available modules; The method described in Claim 3, further comprising the above.

5. The step of determining the health indicator includes, for each module, calculating the product of the value of the module state, the first weight coefficient, and the second weight coefficient, according to any one of Claims 2 to 4.

6. The step of determining the health indicator includes calculating the sum of the products for each module over all modules, according to the method described in Claim 5.

7. The method described in Claim 1, further comprising controlling the use or non-use of each module based on the module state of each module.

8. The method further includes discretizing the plurality of operation parameter values of each module respectively based on a predetermined threshold value. The step of determining the module state of each of the modules includes calculating the sum of the plurality of discretized operation parameter values of the module or the sum of the plurality of operation parameter values of the module. The method according to claim 7.

9. For the modules controlled to be unused among the respective modules, further including the step of identifying abnormal operation parameters based on the plurality of operation parameter values or the plurality of discretized operation parameter values. The method according to claim 7 or 8.

10. A method for determining a health index of a substrate processing apparatus, the substrate processing apparatus including a plurality of module groups, each module group being composed of one or more modules, the method including: For each module in the substrate processing apparatus, obtaining a plurality of operation parameter values respectively; Training a learning model by machine learning so as to output a value related to the health index of the substrate processing apparatus when the plurality of operation parameter values of the plurality of modules of the substrate processing apparatus are input; Estimating the health index of the current substrate processing apparatus from the plurality of operation parameter values of the current plurality of modules using the trained learning model; Including The step of training the learning model by machine learning: Calculating the module state of each module by calculating the sum of the plurality of operation parameter values of each module or the sum of the discretized values thereof; Calculating the health index of the substrate processing apparatus by calculating the sum of the products of the module state of each module and the weight coefficient of each module; Training the learning model using the plurality of operation parameter values of the plurality of modules of the substrate processing apparatus and the calculated health index as training data; Including, a method.

11. The method according to claim 10, wherein the trained learning model is mounted on the substrate processing apparatus.

12. A substrate processing apparatus including a control unit and a plurality of module groups, each module group being composed of one or more modules, the control unit: For each module in the substrate processing apparatus, obtaining a plurality of operation parameter values respectively; ​ Based on the plurality of operation parameter values, determine the module state of each module, Based on the module state of each module and the weight coefficient of each module, determine the health index of the substrate processing apparatus, A substrate processing apparatus configured as described above.

13. A substrate processing apparatus including a control unit and a plurality of module groups, each module group being composed of one or more modules, and the control unit For each module in the substrate processing apparatus, acquire a plurality of operation parameter values respectively, Using a learning model trained by machine learning to output a value related to the health index of the substrate processing apparatus when the plurality of operation parameter values for the plurality of modules of the substrate processing apparatus are input, estimate the health index of the current substrate processing apparatus from the plurality of current operation parameter values of the plurality of modules, Configured as described above, The learning model Calculate the module state of each module by calculating the sum of the plurality of operation parameter values for each module or the discretized values thereof, Calculate the health index of the substrate processing apparatus by calculating the sum of the products of the module state of each module and the weight coefficient of each module, Use the plurality of operation parameter values for the plurality of modules of the substrate processing apparatus and the calculated health index as training data to train the learning model, A substrate processing apparatus trained by the above.

14. The substrate processing apparatus according to claim 13, wherein the trained learning model is mounted on the substrate processing apparatus.

Images