Information processing device, inference device, machine learning device, information processing method, inference method, and machine learning method

Abstract

To provide an information processing device capable of properly predicting a polishing pad state according to an operating state of a substrate processing device.SOLUTION: An information processing device 5 comprises: an information acquisition unit 500; and a state prediction unit 501. A substrate processing device comprises: a polishing table for rotatably supporting a polishing pad; a top ring for pressing a substrate against the polishing pad; a polishing fluid supply nozzle for supplying a polishing fluid to the polishing pad; and an atomizer for spraying a cleaning fluid to a dresser for dressing the polishing pad and the polishing pad. The substrate processing device performs a chemical-mechanical polishing treatment on the substrate. The information acquisition unit 500 acquires operating state information including top ring state information, polishing table state information, polishing fluid supply nozzle state information, dresser state information, and atomizer state information as operating states of when the substrate processing device is operated. The state prediction unit 501 predicts polishing pad state information for the operating state information by inputting the operating state information acquired by the information acquisition unit into a learning model 10A which has been trained through machine learning to learn correlations between the operating state information and the polishing pad state information.SELECTED DRAWING: Figure 13

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an inference device, a machine learning device, an information processing method, an inference method, and This section discusses machine learning methods. [Background technology]

[0002] Chemical machinery is one type of substrate processing equipment that performs various processes on substrates such as semiconductor wafers. Polishing (CMP: Chemical Mechanical Polishing) treatment A substrate processing apparatus is known that performs this. In a substrate processing apparatus, for example, a polishing pad is used. While the table rotates, polishing fluid (slurry) is supplied to the polishing pad from the polishing fluid supply nozzle. With the substrate in place, a polishing head called a top ring presses the substrate against the polishing pad. The substrate is then polished chemically and mechanically. The polishing pad is then removed by the dresser. A washing process is performed, and then a high-pressure cleaning fluid is supplied from the atomizer to the polishing pad. Once the polishing debris remaining on the polishing pad is removed, the series of processes is complete, and the next substrate is ready. Proceed to processing.

[0003] As the above series of processes are repeated, the wear of the polishing pad gradually progresses. Therefore, the polishing pad will need to be replaced, but the timing for replacing the polishing pad is, for example, the polishing pad It was managed by cumulative usage time (see, for example, Patent Document 1). [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2011-204721 [Overview of the project] [Problems that the invention aims to solve]

[0005] Patent Document 1 states that the cumulative usage time of the polishing pad is measured by the top ring when the substrate is polished by the polishing pad. It is determined by accumulating the time spent pressing the polishing pad against the substrate. However, the polishing pad The condition of the top ring is determined not only by the polishing process, but also by the dressing process. Because it also varies during the period when cleaning is performed by the singe or atomizer, cumulative usage time Simply monitoring the condition of the polishing pads does not allow for a detailed understanding of their state.

[0006] On the other hand, the substrate processing equipment includes a top ring, polishing table, polishing fluid supply nozzle, and drain. The operating state of the sass and atomizer are factors that affect the condition of the polishing pad. It interacts with the polishing pad in a complex and interactive manner. Therefore, each operating state affects the polishing pad. It is difficult to accurately analyze what effect it has on the state.

[0007] In view of the above problems, the present invention provides a method for adjusting the state of the polishing pad according to the operating state of the substrate processing apparatus. Information processing device, inference device, machine learning device, information processing method that enables accurate prediction. The objective is to provide inference methods and machine learning methods. [Means for solving the problem]

[0008] To achieve the above objective, an information processing apparatus according to one aspect of the present invention is: A polishing table that rotatably supports a polishing pad, and a mechanism that presses the substrate against the polishing pad. A plucking ring, a polishing fluid supply nozzle for supplying polishing fluid to the polishing pad, and a dresser disc. The dresser disc is rotatably supported and brought into contact with the polishing pad. A dresser for dressing a polishing pad, and a substrate processing apparatus that includes an atomizer for injecting a cleaning fluid onto the polishing pad, when the substrate processing apparatus operates to perform chemical mechanical polishing of a substrate, the operation state information includes top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and information acquisition unit that acquires atomizer state information indicating the state of the atomizer, a state prediction unit that predicts the polishing pad state information for the operation state information by inputting the operation state information acquired by the information acquisition unit into a learning model that learns the correlation between the operation state information and the polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus operates in the operation state indicated by the operation state information by machine learning, and. Comprising.

Advantages of the Invention

[0009] According to the information processing apparatus according to one aspect of the present invention, by inputting operation state information including top ring state information, polishing table state information, polishing fluid supply nozzle state information, dresser state information, and atomizer state information into a learning model, polishing pad state information for the operation state information is predicted, so that the state of the polishing pad can be appropriately predicted according to the operation state of the substrate processing apparatus. can be predicted.

[0010] Problems, configurations, and effects other than those described above will be clarified in the form for implementing the invention described later. be. [[ID=4X]]

Brief Description of the Drawings

[0011] ​[Figure 1] This is an overall configuration diagram showing an example of the substrate processing system 1. [Figure 2] This is a plan view showing an example of a substrate processing apparatus 2. [Figure 3] This is a perspective view showing an example of the first to fourth polished sections 22A to 22D. [Figure 4] This is a schematic cross-sectional view showing an example of the top ring 221. [Figure 5] This is a list of examples of substrate processing equipment 2. [Figure 6] This is a hardware configuration diagram showing an example of the Computer 900. [Figure 7] This is a data configuration diagram showing an example of production history information 30 managed by the database device 3. [Figure 8] This data configuration diagram shows an example of polishing test information 31 managed by the database device 3. [Figure 9] This is a block diagram showing an example of a machine learning device 4 according to the first embodiment. [Figure 10] This figure shows an example of the first learning model 10A and the first training data 11A. [Figure 11] This flowchart shows an example of a machine learning method using machine learning device 4. [Figure 12] This is a block diagram showing an example of an information processing device 5 according to the first embodiment. [Figure 13] This is a functional diagram showing an example of an information processing device 5 according to the first embodiment. [Figure 14] This flowchart shows an example of an information processing method using the information processing device 5. [Figure 15] This is a block diagram showing an example of a machine learning device 4a according to the second embodiment. [Figure 16] This figure shows an example of the second learning model 10B and the second training data 11B. [Figure 17] This is a block diagram showing an example of an information processing device 5a that functions as an information processing device 5a according to the second embodiment. [Figure 18]This is a functional diagram showing an example of an information processing device 5a according to the second embodiment. [Modes for carrying out the invention]

[0012] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. This section schematically illustrates the scope necessary to explain the objectives of the present invention, and explains the relevant parts of the present invention. This explanation will primarily focus on the necessary details, and any parts that are omitted will be explained using publicly available technology. It shall be considered as such.

[0013] (First embodiment) Figure 1 is an overall configuration diagram showing an example of a substrate processing system 1. System 1 presses a substrate such as a semiconductor wafer (hereinafter referred to as "wafer") W against a polishing pad. Chemical mechanical polishing (hereinafter referred to as "polishing process") is a process that polishes the surface of wafer W to make it flat by applying a special coating. The surface of the wafer W is cleaned by bringing the polished wafer W into contact with a cleaning tool. It functions as a system that manages a series of substrate processing steps, including cleaning.

[0014] The substrate processing system 1 consists of a substrate processing device 2 and a database device as its main components. It comprises 3, a machine learning device 4, an information processing device 5, and a user terminal device 6. Each device 2~ 6 consists of, for example, a general-purpose or dedicated computer (see Figure 6 below), Connected to a wired or wireless network 7, various data (Figure 1 shows the transmission of some data) The devices are configured to be able to send and receive signals (received by dashed arrows). Note that the number of each device from 2 to 6 The connection configuration of Network 7 is not limited to the example in Figure 1 and may be changed as appropriate.

[0015] The substrate processing apparatus 2 is composed of multiple units and processes one or more wafers W in a series For substrate processing, for example, loading, polishing, cleaning, drying, film thickness measurement, unloading, and other processes are performed. It is a device that performs the processing separately. In this case, the substrate processing device 2 is set for each unit. Device setting information 265 consisting of multiple device parameters, and polishing process, cleaning process, drying process The operation of each unit is controlled while referring to the circuit board recipe information 266, which defines the operating state of the circuit board. do.

[0016] The substrate processing device 2 generates various reports R in a database device according to the operation of each unit. 3. Transmit to user terminal device 6, etc. Various reports R include, for example, when board processing is performed. Process information that identifies the target wafer W at the time, and information for each unit when each process is performed. Device status information indicating the status, event information detected by the substrate processing device 2, substrate processing device 2 This includes user (operator, production manager, maintenance manager, etc.) operational information for the system.

[0017] Database device 3 records the history of when substrate processing was performed using polishing pads for production. Production history information 30 related to this, and polishing test using a test polishing pad (hereinafter referred to as "polishing") This device manages polishing test information 31 related to the history of when a polishing test (called a "polishing test") was performed. In addition to the above, the database device 3 also contains device setting information 265 and board recipe information. 266 may be stored, in which case the substrate processing device 2 will refer to this information. You may do so.

[0018] The database device 3 is used by the substrate processing device 2 to process substrates using polishing pads for mass production. At that time, various reports R are received from the substrate processing device 2 and registered in the production history information 30. By recording this information, the production history information 30 will store reports R related to substrate processing.

[0019] The database device 3 is used when the substrate processing device 2 performs a polishing test using a test polishing pad. When this happens, the substrate processing device 2 sends various reports R (including at least device status information) The data is received and registered in the polishing test information 31, and the results of the polishing test are associated with it. By registering, the polishing test information 31 will store reports R and test results related to the polishing test. The polishing test may be performed using the substrate processing apparatus 2 for production, or the substrate processing apparatus 2 The same polishing process may be carried out using a test polishing apparatus (not shown) that can reproduce the same process. For polishing pads and polishing test equipment used for testing, the condition of the polishing pad is, for example, The distribution of polishing debris on the polishing surface of the polishing pad, the flatness of the polishing surface, the surface roughness, the temperature, Various polishing pad measuring devices (not shown) for measuring the degree of wetness and the coefficient of friction are provided. Then, the measurement values ​​from the polishing pad measuring device are registered as test results in the polishing test information 31.

[0020] The machine learning device 4 operates as the main component of the machine learning learning phase, for example, in the database A portion of the polishing test information 31 is acquired from the polishing device 3 as the first learning data 11A, and the information processing A first learning model 10A to be used in the processing device 5 is generated by machine learning. The first learning model 10A is provided to the information processing device 5 via the network 7, recording medium, etc. It will be done.

[0021] The information processing device 5 operates as the main component of the machine learning inference phase, and is controlled by the machine learning device 4. Using the first learning model 10A generated, the polishing process by the substrate processing apparatus 2 is performed in full production. When using a polishing pad, the condition of the polishing pad is predicted, and the predicted The resulting polishing pad status information is transmitted to the database device 3, user terminal device 6, etc. The timing at which the information processing device 5 predicts the polishing pad state information is when the polishing process is performed. This can be done afterward (post-prediction processing), or while the polishing process is in progress (real-time prediction processing). ) is also acceptable, or it may be before the polishing process is performed (preliminary treatment).

[0022] User terminal device 6 is a terminal device used by the user, and may be a stationary device or a portable device. A strip-type device is also acceptable. The user terminal device 6 may, for example, display an application program, web It accepts various input operations via the display screen of a browser, etc., and also via the display screen and various information (for example, event notifications, polishing pad status information, production history information 30, polishing Display test information (e.g., 31).

[0023] (Substrate processing device 2) Figure 2 is a plan view showing an example of a substrate processing apparatus 2. The substrate processing apparatus 2 is approximately rectangular in plan view. Inside the shaped housing 20 are a load / unload unit 21 and a polishing unit 22 The board transport unit 23, the finishing unit 24, the film thickness measurement unit 25, and the control unit It is composed of a knit 26, a load / unload unit 21 and a polishing unit 2 2. The substrate transport unit 23 and the finishing unit 24 are separated by a first partition wall 200A. The substrate transport unit 23 and the finishing unit 24 are separated by a second partition wall 200B. It is more compartmentalized.

[0024] (Load / Unload Unit) The load / unload unit 21 is a wafer carrier capable of storing a large number of wafers W in the vertical direction. The first to fourth front load sections 210A to 210D on which the set (FOUP, etc.) is placed. And, it can move up and down along the storage direction (vertical direction) of the wafer W stored in the wafer cassette. Arrangement of the transport robot 211 and the first to fourth front load sections 210A to 210D A horizontal movement mechanism that moves the transport robot 211 along the direction (the shorter side of the housing 20). It comprises part 212.

[0025] The transport robot 211 has first to fourth front load sections 210A to 210D The wafer cassette placed on it, the substrate transport unit 23 (specifically, the lifter 232 described later), Finishing unit 24 (specifically, the first and second drying sections 24E and 24F described later), and It is configured to be accessible to the film thickness measurement unit 25, and the wafer W is transferred between them. It is equipped with two upper and lower hands (not shown) for handling. The lower hand receives the wafer W before processing. The upper hand is used when handing over the wafer W after processing. When transferring the wafer W to the plate transport unit 23 or the finishing unit 24, the first A shutter (not shown) installed in the bulkhead 200A is opened and closed.

[0026] (Polishing unit) The polishing unit 22 is comprised of first to fourth polishing units that perform polishing (planarization) of the wafer W. It is equipped with polishing sections 22A to 22D. The first to fourth polishing sections 22A to 22D are located in the housing 20 They are arranged and positioned along the longitudinal direction of the object.

[0027] Figure 3 is a perspective view showing an example of the first to fourth polishing sections 22A to 22D. The basic configuration and function of the polishing sections 22A to 22D of section 4 are the same.

[0028] Each of the first to fourth polishing sections 22A to 22D has a polishing pad 2200 having a polishing surface. A rotatably supported polishing table 220 and a device that holds the wafer W and polishes the wafer W. Top ring (polishing head) for polishing while pressing against the polishing pad 2200 on the bull 220. The polishing pad 2200 has a polishing fluid supply nozzle 222 that supplies polishing fluid to the polishing pad 2200. , rotatably supports the dresser disc 2230 and the dresser disc 2230 The dressing is applied to the polishing surface of the polishing pad 2200. The system consists of a 223, an atomizer 224 that sprays cleaning fluid onto the polishing pad 2200, and a polishing process. It includes an environmental sensor 225 that measures the state of the internal space of the housing 20.

[0029] The polishing table 220 is supported by the polishing table shaft 220a, and its axial circumference A rotary moving mechanism 220b that rotates the polishing table 220, and a polishing pad 220 It is equipped with a temperature control mechanism 220c that adjusts the surface temperature of 0.

[0030] The top ring 221 is supported by a top ring shaft 221a that is movable in the vertical direction. The rotational movement mechanism 221c rotates the top ring 221 around its axis, A vertical movement mechanism 221d that moves the top ring 221 in the vertical direction, and a support shaft 2 The oscillating movement mechanism 2 rotates (oscillates) the top ring 221 around 21b as the pivot point. It is equipped with 21e.

[0031] The polishing fluid supply nozzle 222 is supported on the support shaft 222a, and the support shaft 22 The oscillating movement mechanism 222 rotates the polishing fluid supply nozzle 222 around 2a as the pivot point. b, a flow rate adjustment unit 222c that adjusts the flow rate of the polishing fluid, and a temperature control unit that adjusts the temperature of the polishing fluid. It comprises a mechanism 222d. The polishing fluid is a polishing liquid (slurry) or pure water, and further, It may contain a chemical solution, or it may be a polishing solution with a dispersant added.

[0032] The dresser 223 is supported by a dresser shaft 223a that is movable in the vertical direction, A rotational movement mechanism 223c that rotates the dresser 223 around its axis, and the dresser 22 A vertical movement mechanism 223d moves 3 in the vertical direction, and a support shaft 223b pivots around the pivot point. The system also includes a swinging movement mechanism 223e that rotates the dresser 223.

[0033] The atomizer 224 is supported by the support shaft 224a, and rotates the support shaft 224a. The oscillating movement mechanism 224b rotates the atomizer 224 around a pivot point, and the cleaning fluid It is equipped with a flow rate adjustment unit 224c for adjusting the flow rate. The polishing fluid is a polishing liquid (slurry) or pure It is water, and may also contain a chemical solution, or it may be a polishing solution with a dispersant added. .

[0034] The environmental sensor 225 consists of sensors arranged in the internal space of the housing 20, for example A temperature sensor 225a measures the temperature of the internal space, and a humidity sensor measures the humidity of the internal space. The system includes a 225b pressure sensor, a 225c pressure sensor for measuring the air pressure in the internal space, and an oxygen concentration sensor 225 It includes d and a microphone (sound sensor) 225e. The environmental sensor 225 is as follows: During and before / after the polishing process, the surface, temperature distribution, airflow distribution, etc. of the polishing pad 2200 were photographed. It may be equipped with a camera (image sensor) capable of capturing shadows. The subject of the camera's shot is visible light. It is not limited to infrared light, ultraviolet light, etc.

[0035] Note that in Figure 3, the rotational movement mechanism parts 220b, 221c, 223c and the vertical movement mechanism part 22 Specifics of 1d, 223d, and the swinging movement mechanism parts 221e, 222b, 223e, and 224b Although the basic configuration is omitted, for example, modules for generating driving force such as motors and air cylinders Lures and drive force transmission devices such as linear guides, ball screws, gears, belts, couplings, and bearings. The structure and sensors such as linear sensors, encoder sensors, limit sensors, and torque sensors It is configured by combining as appropriate. Figure 3 shows the specific configuration of the flow rate adjustment units 222c and 224c. Although omitted here, for example, modules for fluid control such as pumps, valves, and regulators. And, flow sensor, pressure sensor, liquid level sensor, temperature sensor, fluid concentration sensor, fluid party It is configured by appropriately combining sensors such as a cruise sensor. In Figure 3, the temperature control mechanism 220c The specific configuration of 222d is omitted, but for example, it includes heaters, heat exchangers, etc. for temperature control. A module (contact or non-contact) and sensors such as temperature sensors and current sensors are appropriately combined. It is composed of multiple parts.

[0036] Figure 4 is a schematic cross-sectional view showing an example of the top ring 221. This includes a top ring body 2210 attached to the top ring shaft 221a, and the top A roughly disc-shaped carrier 2211 housed in the puller body 2210, and the carrier 2211 A membrane 2212 is positioned on the lower side and presses the wafer W against the polishing pad 2200. The polishing pad 2200 is positioned on the outer circumference of the carrier 2211 and the membrane 2212. A roughly annular retainer ring 2213 that directly presses against the top ring body 2210 and the Positioned between the retaining rings 2213, the retaining ring 2213 is polished by the polishing pad 2200 It is equipped with a retaining airbag 2214 that applies pressure to it.

[0037] The membrane 2212 is made of an elastic film and has multiple concentric partitions 2 inside. By having 212e, from the center of the top ring body 2210 toward the outer circumference It has first to fourth membrane pressure chambers 2212a to 2212d arranged concentrically. Furthermore, the membrane 2212 has a plurality of holes 2212f on its lower surface for adsorbing the wafer W. It functions as a substrate holding surface for holding the wafer W. Retainer ring airbag 2214 It is formed of an elastic membrane and has a retainer ring pressure chamber 2214a inside. The configuration of the top ring 221 may be changed as appropriate, and the entire carrier 2211 may be pressed down. It may also be equipped with pressure chambers, and the number of membrane pressure chambers that the membrane 2212 has may vary. The shape may be changed as appropriate, and the number and arrangement of the adsorption holes 2212f may also be changed as appropriate. Furthermore, the membrane 2212 may not have pores 2212f for adsorption.

[0038] The first to fourth membrane pressure chambers 2212a to 2212d have the first to fourth flow paths 2 216A to 2216D are connected respectively, and the retainer ring pressure chamber 2214a has the 5th The flow path 2216E is connected. The first to fifth flow paths 2216A to 2216E are top It communicates with the outside via a rotary joint 2215 provided on the ring shaft 221a. The first branching channels 2217A~2217E and the second branching channels 2218A~2218E The channels are then branched into two. The first to fifth channels 2216A to 2216E are connected to pressure sensors. PA~PE are installed respectively. The first branching channel 2217A~2217E has valve V Pressurized fluid via 1A~V1E, flow sensors FA~FE, and pressure regulators RA~RE Connects to a gas supply source (GS) for air, nitrogen, etc. Second branch channel 2218A~2218 E is connected to the vacuum source VS via valves V2A to V2E, and valve It is configured to communicate with the atmosphere via V3A to V3E.

[0039] The wafer W is held by suction on the lower surface of the top ring 221 and placed on the polishing table 220. After being moved to a fixed polishing position, polishing fluid was supplied from the polishing fluid supply nozzle 222. The polishing surface of the pad 2200 is polished by being pressed against it by the top ring 221. In this process, the top ring 221 independently controls the pressure regulators RA~RE. The pressure fluid supplied to the first to fourth membrane pressure chambers 2212a to 2212d is used to... The pressing force applied to the wafer W against the polishing pad 2200 is adjusted for each region of the wafer W, The retainer ring 2213 is ground by the pressurized fluid supplied to the retainer ring pressure chamber 2214a. The pressing force applied to the polishing pad 2200 is adjusted. First to fourth membrane pressure chambers 221 The pressure fluid supplied to 2a~2212d and retainer ring pressure chamber 2214a respectively The pressure is measured by pressure sensors PA~PE, and the flow rate of the pressurized gas is measured by a flow sensor. These are measured by FA, ​​FE, and so on.

[0040] (Circuit board transport unit) As shown in Figure 2, the substrate transport unit 23 has first to fourth polishing sections 22A to 22D First and second linear elements that are horizontally movable along the direction of alignment (the longitudinal direction of the housing 20) Linear transporters 230A, 230B, and the first and second linear transporters 230A, 2 Swing transporter 231 and load / unload unit located between 30B The lifter 232 is positioned on the 21 side, and the wafer W is positioned on the finishing unit 24 side. It is equipped with a base 233.

[0041] The first linear transporter 230A is adjacent to the first and second polishing sections 22A and 22B. They are arranged in such a way that there are four transport positions (in order from the load / unload unit 21 side, the first to This is a mechanism for transporting wafer W between the fourth transport positions (let's call them TP1 to TP4). The transfer position TP2 is the position where the wafer W is transferred to the first polishing section 22A, and the third transfer position The feeding position TP3 is the position where the wafer W is transferred to the second polishing section 22B.

[0042] The second linear transporter 230B is adjacent to the third and fourth polishing sections 22C and 22D. They are arranged in such a way that there are three transport positions (from the load / unload unit 21 side in order to the 5th to This is a mechanism for transporting wafer W between the seventh transport position (TP5 to TP7). The transfer position TP6 is the position where the wafer W is transferred to the third polishing section 22C, and the seventh transfer position The feeding position TP7 is the position where the wafer W is transferred to the fourth polishing section 22D.

[0043] The swing transporter 231 is adjacent to the fourth and fifth transport positions TP4 and TP5. A hand is positioned and is movable between the fourth and fifth transport positions TP4 and TP5. The swing transporter 231 has the first and second linear transporters 230 The wafer W is passed between A and 230B, and the wafer W is temporarily placed on the temporary storage stand 233. This is the mechanism. The lifter 232 is positioned adjacent to the first transport position TP1, and loads / This is a mechanism for transferring wafer W between the unloading unit 21 and the transport robot 211. When the wafer W is transferred, a shutter (not shown) provided in the first partition wall 200A is opened and closed. It can be done.

[0044] (Finishing unit) The finishing unit 24 cleans the substrate using a roll sponge 2400, as shown in Figure 2. The apparatus consists of first and second roll sponge cleaning units 24A and 24B, arranged in two stages, upper and lower. And, as a substrate cleaning device using Pen Sponge 2401, the first and The second pen sponge cleaning section 24C, 24D and the substrate drying unit for drying the wafer W after cleaning. The structure includes first and second drying sections 24E and 24F arranged in upper and lower tiers, and a wafer transporter W. It comprises first and second conveying sections 24G and 24H for transporting. Furthermore, the roll sponge cleaning section 2 4A, 24B, pen sponge washing section 24C, 24D, drying section 24E, 24F, and conveying The number and arrangement of sections 24G and 24H are not limited to the example in Figure 2 and may be changed as appropriate, for example, Swap the positions of the ball sponge cleaning units 24A and 24B and the pen sponge cleaning units 24C and 24D. You may do so.

[0045] Each part 24A to 24H of the finishing unit 24 is partitioned in the first and Along the two linear transporters 230A and 230B, for example, the first and second rolls Sponge cleaning section 24A, 24B, first transport section 24G, first and second pen sponge cleaning Sections 24C, 24D, second conveying section 24H, and first and second drying sections 24E, 24F They are arranged in order (from furthest from the load / unload unit 21). The finishing unit 24 is The first and second roll sponge cleaning sections 24A and 24B are used to clean the wafer W after polishing. Primary cleaning treatment by either or both of the following, first and second pen sponge cleaning sections 24C, 2 A secondary cleaning process using either or both of 4D, and the first and second drying sections 24E, 24 The drying process is carried out sequentially using either or both of F. The order of processing by A~24H may be changed as appropriate, and some parts of the processing may be omitted. For example, the cleaning process by the roll sponge cleaning units 24A and 24B is omitted, and the pen sponge cleaning is performed. The process may begin with cleaning using the cleaning units 24C and 24D. The finishing unit 24 also... Roll sponge cleaning units 24A, 24B, and pen sponge cleaning units 24C, 24D By providing a buffing unit (not shown) in place of or in addition to either of the above, the buffing process is performed. It may also be done in this manner. Furthermore, in this embodiment, each part 24A to 24 of the finishing unit 24 H holds the wafer W in a horizontal position (horizontal holding), but if the wafer W is held vertically or It is also acceptable to hold it at an angle.

[0046] (Film thickness measurement unit) The film thickness measurement unit 25 measures the film thickness of the wafer W before or after polishing. It is a device, and for example, it consists of an optical film thickness measuring instrument, an eddy current film thickness measuring instrument, etc. Each film thickness measurement The transfer of wafers W to the module is performed by the transport robot 211.

[0047] (Control unit) Figure 5 is a block diagram showing an example of a substrate processing apparatus 2. The control unit 26 controls each unit A control unit that is electrically connected to units 21-25 and comprehensively controls each unit 21-25. It functions as follows. Below, the control system of the polishing unit 22 (module, sensor, sequencer) I will explain using (S) as an example, but the other units 21, 23-25 ​​also share the same basic configuration and functions. Therefore, the explanation will be omitted.

[0048] The polishing unit 22 comprises each subunit (for example, the polishing table) Lure 220, top ring 221, polishing fluid supply nozzle 222, dresser 223, atomizer Each of the modules 2271-227 (e.g., 224) is located in a separate unit and is the target of control. r And multiple modules 2271~227 r They are arranged in each of the modules 227 1-227 r Multiple sensors 2281-228 detect the data (detected values) necessary for control. s And each sensor 2281~228 s Based on the detected values, each module 2271~227 r It includes a sequencer 229 that controls the operation of [the device].

[0049] Polishing unit 22 sensors 2281~228 s For example, the number of rotations of the polishing table 220 A sensor to detect rotation speed, a sensor to detect the rotational torque of the polishing table 220, and a polishing pad. A sensor that detects the surface temperature of 2200, a sensor that detects the rotation speed of the top ring 221, A sensor that detects the rotational torque of the top ring 221, and a sensor that detects the oscillation position of the top ring 221. Sensors that emit light, sensors that detect the oscillation speed of the top ring 221, the top ring 221 A sensor that detects the oscillation torque, a sensor that detects the height of the top ring 221, the top ring Sensor for detecting the lifting torque of G221, first to fourth membrane pressure chambers 2212a~ A sensor that detects the pressure (positive and negative pressure) in 2212d and the retainer ring pressure chamber 2214a. The first to fourth membrane pressure chambers 2212a to 2212d and retainer ring pressure Sensor for detecting the flow rate of pressurized fluid supplied to chamber 2214a, polishing fluid supply nozzle 222 A sensor that detects the flow rate of the polishing fluid supplied from the polishing fluid supply nozzle 222 A sensor detects the temperature of the polishing fluid, and the polishing fluid is dispensed by the polishing fluid supply nozzle 222. A sensor that detects the oscillation position of the abrasive fluid supply nozzle 222, which can be converted to a different position, and the concentration of the abrasive fluid Sensors to detect degree, cleanliness of polishing fluid (for example, particles contained in the waste liquid of the polishing fluid) A sensor that detects the concentration, particle size, and number of particles per particle size, and the rotation speed of the dresser 223. Sensors that emit light, sensors that detect the rotational torque of the dresser 223, and the oscillation position of the dresser 223 A sensor that detects the position, a sensor that detects the oscillation speed of the dresser 223, the oscillation of the dresser 223 A sensor for detecting dynamic torque, a sensor for detecting the height of the dresser 223, and a dresser disc. Sensor for detecting the pressing force when 2230 is brought into contact with the polishing pad, atomizer 224 A sensor that detects the flow rate of the cleaning fluid supplied from the atomizer 224, and the cleaning fluid supplied from the atomizer 224. A sensor that detects the temperature of the fluid, and a sensor that detects the pressure of the cleaning fluid supplied from the atomizer 224. A sensor, and an atomizer 224 that can convert the position of the cleaning fluid to be dispensed. This includes sensors for detecting the oscillation position, environmental sensors 225, etc.

[0050] The control unit 26 includes a control unit 260, a communication unit 261, an input unit 262, an output unit 263, and The system also includes a memory unit 264. The control unit 26 is, for example, a general-purpose or dedicated computer. It consists of the components shown in Figure 6 below.

[0051] The communication unit 261 is connected to the network 7 and functions as a communication interface for transmitting and receiving various data. The input unit 262 accepts various input operations and, together with the output unit 263 that outputs various information via a display screen, signal tower lighting, buzzer sound, etc., functions as a user interface. The memory unit 264 stores various programs (operating system (OS), application programs, web browsers, etc.) and data (device setting information 265, substrate recipe information 266, etc.) used in the operation of the substrate processing apparatus 2. The device setting information 265 and the substrate recipe information 266 are data that can be edited by the user via the display screen.

[0052]

[0053] The control unit 260 acquires the detection values of a plurality of sensors 2181 to 218, 2281 to 228, 2381 to 238, 2481 to 248, 2581 to 258 (hereinafter referred to as the "sensor group") via a plurality of sequencers 219, 229, 239, 249, 259 (hereinafter referred to as the "sequencer group"), and operates in cooperation with a plurality of modules 2171 to 217, 2271 to 227, 2371 to 237, 2471 to 247, 2571 to 257 (hereinafter referred to as the "module group") to perform a series of substrate processing such as loading, polishing, cleaning, drying, film thickness measurement, and unloading. q s u w <000001 p r t v x

[0054] (Hardware configuration of each device) ​Figure 6 is a hardware configuration diagram showing an example of computer 900. (Substrate processing apparatus 2) Control unit 26, database device 3, machine learning device 4, information processing device 5, and Each of the terminal devices 6 is comprised of a general-purpose or dedicated computer 900.

[0055] As shown in Figure 6, the computer 900 has a bus 910 as its main component. Processor 912, memory 914, input device 916, output device 917, display device Chair 918, storage device 920, communication I / F (interface) unit 922, external device Instrument I / F section 924, I / O (input / output) device I / F section 926, and media input / output section It is equipped with 928. Note that the above components are selected according to the intended use of the computer 900. It may be omitted as appropriate.

[0056] Processor 912 is one or more arithmetic units (CPU (Central Pro processing unit), MPU (micro-processing unit) , DSP (digital signal processor), GPU (Graph It consists of an ICS Processing Unit, etc., and the entire Computer 900 It operates as a central control unit. Memory 914 stores various data and programs 930. It stores data and, for example, functions as main memory (volatile memory such as DRAM, SRAM, etc.) It consists of non-volatile memory (ROM), flash memory, etc.

[0057] The input device 916 consists of, for example, a keyboard, mouse, numeric keypad, electronic pen, etc. It functions as an input unit. The output device 917 is, for example, a sound (voice) output device, a bi It consists of a pulse device and functions as an output unit. The display device 918 is, for example, It consists of liquid crystal displays, organic EL displays, electronic paper, projectors, etc. It functions as an output unit. The input device 916 and the display device 918 are touch panels. It may be configured as an integrated unit, like a display. The storage device 920 may be, for example For example, it consists of HDDs, SSDs (Solid State Drives), etc., and the storage unit is It functions as follows: The storage device 920 contains the operating system and program 930. It stores various data necessary for execution.

[0058] The communication interface unit 922 connects to networks such as the Internet and intranets 940 (Figure) Connected by wire or wirelessly to the same network 7 as in 1, and communicating It functions as a communication unit that sends and receives data with other computers according to standards. The internal device I / F unit 924 connects to external devices 95 such as cameras, printers, scanners, and reader / writers. It is connected to 0 by wire or wireless, and data is transmitted to and from external device 950 according to a predetermined communication standard. It functions as a communication unit that transmits and receives data. The I / O device I / F unit 926 is a communication unit that transmits and receives various data. It is connected to I / O devices 960 such as actuators, and between them and I / O devices 960 For example, various signals and data such as detection signals from sensors and control signals to actuators. It functions as a communication unit that transmits and receives data. The media input / output unit 928, for example, is a DVD drive. It consists of a drive device such as a live CD drive, and media such as DVDs and CDs (non-temporary). It reads and writes data to the storage medium (970).

[0059] In the computer 900 having the above configuration, the processor 912 is a storage device The program 930 stored in 920 is called and executed in memory 914, and the bus 910 It controls various parts of the computer 900 via the storage device. Note that program 930 is stored on the storage device. Instead of being stored in location 920, it may be stored in memory 914. Program 930 is an inst Recorded on media 970 in a downloadable or executable file format, It may also be provided to the computer 900 via the DIA input / output unit 928. Program 93 0 is downloaded via the network 940 through the communication I / F unit 922. It may also be provided to computer 900. Furthermore, computer 900 may have a processor 9 The various functions that 12 implements by executing program 930 can be, for example, FPGA, A It is also acceptable to implement this using hardware such as SIC.

[0060] Computer 900 consists of, for example, stationary computers and portable computers. , any form of electronic device. Computer 900 is a client computer. Alternatively, a server-type computer or a cloud-based computer would also be acceptable. Computer 90 0 may also apply to devices other than devices 2-6.

[0061] (Production history information 30) Figure 7 shows an example of the data structure of production history information 30 managed by the database device 3. This is a drawing. Production history information 30 indicates that the substrate was processed using a polishing pad for this production. The report R obtained is classified and registered in a table, for example, each wafer W Wafer history table 300 and polishing history related to equipment status information in polishing process It is equipped with a table 301. In addition to the above, the production history information 30 includes the cleaning process. Cleaning history table for device status information, event history table for event information, and The system includes operation history tables and other information related to operations, but a detailed explanation is omitted.

[0062] Each record in the wafer history table 300 contains, for example, wafer ID, cassette number, and The lot number, start time, end time, and unit ID used for each process are registered. In example 7, the polishing and cleaning processes are given, but other processes can be registered in the same way. .

[0063] Each record in the polishing history table 301 contains, for example, a wafer ID and top ring status information. Report, polishing table status information, polishing fluid supply nozzle status information, dresser status information, atomizer Status information, internal device environment information, processing performance information, etc., are registered.

[0064] The top ring condition information is information indicating the state of the top ring 221 during the polishing process. The top ring status information is, for example, the sensor group (or module) of the top ring 221. The detected values ​​of each sensor (or each module) are sampled at predetermined time intervals by the module group. This is the command value to the commander.

[0065] The polishing table status information is information that indicates the state of the polishing table 220 during the polishing process. The polishing table status information is, for example, obtained from the group of sensors (or module) of the polishing table 220. The detected values ​​of each sensor (or each module) are sampled at predetermined time intervals by the module group. This is the command value to the commander.

[0066] The polishing fluid supply nozzle status information indicates the state of the polishing fluid supply nozzle 222 during the polishing process. This is the information to be shown. The polishing fluid supply nozzle status information is, for example, that the polishing fluid supply nozzle 222 is Each sensor sampled at predetermined time intervals by the sensor group (or module group) This is the detected value (or command value to each module) of the module.

[0067] The dresser status information indicates the state of the dresser 223 during the polishing process. The state information is obtained, for example, by the group of sensors (or modules) that the dresser 223 has. The detected values ​​(or command values ​​to each module) of each sensor sampled at predetermined time intervals. That is the case.

[0068] The atomizer status information indicates the state of the atomizer 224 during the polishing process. Atomizer status information is, for example, the group of sensors (or modules) that atomizer 224 has. ) The detected values ​​of each sensor sampled at predetermined time intervals (or to each module) (Command value)

[0069] The internal environment information of the device indicates the state of the internal space of the substrate processing apparatus 2 formed by the housing 20. This information indicates that the internal space of the substrate processing apparatus 2 is the space in which the polishing unit 22 is located. Yes, the internal environment information of the device is sampled at predetermined time intervals by, for example, the environmental sensor 225. These are the detected values ​​from each sensor. Note that the internal space of the substrate processing apparatus 2 is the polishing unit 2 If the first to fourth polishing sections 22A to 22D of 2 are separated, then the environmental sensor The SA225 is installed for each of the first to fourth polishing sections 22A to 22D, and provides information on the internal environment of the device. This is obtained for each of the first to fourth polishing sections 22A to 22D.

[0070] Processing performance information is information that shows the results of the polishing process. Processing performance information is polishing pad 22 When polishing is performed using the polishing pad 2200 after the 00 has been replaced, the wafer This includes the cumulative number of W cards used and the cumulative usage time.

[0071] By referring to the polishing history table 301, for wafer W identified by wafer ID The device status of the substrate processing apparatus 2 when the polishing process is performed is represented by the time-series data of each sensor ( Alternatively, time-series data (or data from each module) can be extracted.

[0072] (Polishing test information 31) Figure 8 shows an example of the data structure of polishing test information 31 managed by the database device 3. This is a drawing. Polishing test information 31 describes polishing tests using polishing pads and polishing equipment for testing. When the polishing test is performed, the report R and test results obtained are classified and registered. It is equipped with a cable 310.

[0073] Each record in the polishing test table 310 contains, for example, a test ID and top ring condition information. Polishing table status information, polishing fluid supply nozzle status information, dresser status information, atomizer Status information, internal device environment information, processing performance information, test result information, etc. are registered. Polishing test data Bull 310 top ring status information, polishing table status information, polishing fluid supply nozzle status information The report, dresser status information, atomizer status information, internal device environment information, and processing performance information are as follows: This information shows the condition of each part during the polishing test, and its data structure is as follows: Polishing History Table 3 Since it is the same as 01, a detailed explanation will be omitted.

[0074] The test result information is the condition of the test polishing pad when the polishing process was performed during the polishing test. This is information indicating the state. Test result information is provided on the polishing pads and polishing test equipment used for the test. These are measurement values ​​sampled at predetermined time intervals using a polishing pad measuring device. (See Figure 8) The test result information is included in each time period from the start to the end of the polishing process. Distribution of polishing debris at steps t1, t2, ..., ...tm, ..., tn, flatness of the polished surface, surface The measurement values ​​V1 to V4 include roughness, temperature, wetness, and friction coefficient. For example, shavings removed from wafer W, consumable parts (polishing pads, retainer rings, This includes shavings from the dresser disc, residue from the polishing fluid, etc. Note that the test results information is above. As noted, the measurement value obtained from the polishing pad measuring device is also acceptable, or the environmental sensor 2 Cameras, optical microscopes, and scanning electron microscopes (SEMs) installed as 25 A camera was used to photograph the test polishing pad at predetermined time intervals, and each of the captured images was then used to... The results may be based on image processing or experimental analysis results performed by the experimenter. Furthermore, the test results information is based on a single continuous polishing process from start to finish. The data may be collected in a polishing test, or from the start of the polishing process until a predetermined time is reached. By repeatedly performing polishing tests while gradually increasing the predetermined time, multiple polishing tests can be performed. It is also acceptable if the data was collected in [location / method].

[0075] By referring to the polishing test table 310, in the polishing test identified by the test ID, This shows the state of the polishing unit 22 when polishing is performed using a test polishing pad. The sequence data (or time-series data for each module) and the condition of the test polishing pad at that time. Time-series data indicating the state can be extracted.

[0076] (Machine learning device 4) Figure 9 is a block diagram showing an example of a machine learning device 4 according to the first embodiment. The training device 4 includes a control unit 40, a communication unit 41, a learning data storage unit 42, and a trained model. It is equipped with a memory unit 43.

[0077] The control unit 40 functions as a training data acquisition unit 400 and a machine learning unit 401. Unit 41 connects to external devices (e.g., substrate processing device 2, database device) via the network 7. The device is connected to the following: 3, the information processing device 5, and the user terminal device 6, the polishing test device (not shown), etc. It functions as a communication interface for sending and receiving various types of data.

[0078] The learning data acquisition unit 400 is connected to an external device via the communication unit 41 and the network 7. Therefore, the operating status information is used as input data, and the polishing pad status information is used as output data. The first training data 11A is obtained. The first training data 11A is composed of teacher data. As training data, validation data, and test data in learning This is the data used. Furthermore, the polishing pad state information is used in supervised learning to identify the correct label. This is data used as a reference.

[0079] The learning data storage unit 42 stores the first learning data acquired by the learning data acquisition unit 400. This is a database that stores multiple sets of 11A. It also constitutes the learning data storage unit 42. The specific configuration of the database can be designed as needed.

[0080] The machine learning unit 401 stores multiple sets of first training data stored in the training data storage unit 42. Machine learning is performed using the 11A. That is, the machine learning unit 401 performs the first learning model Multiple sets of the first training data 11A are input to the 10A, and the first training data 11A contains The correlation between the operating state information and the polishing pad state information is learned by the first learning model 10A. This generates a pre-trained first learning model 10A.

[0081] The trained model memory unit 43 stores the first trained model generated by the machine learning unit 401. In the database that stores the training model 10A (specifically, the adjusted weight parameter set) Yes. The first learned model 10A that has been learned and stored in the learned model memory unit 43 is Provided to the actual system (for example, the information processing device 5) via the network 7 or recording media, etc. Note that in Figure 9, the training data storage unit 42 and the trained model storage unit 43 are displayed separately. Although shown as memory units, these may consist of a single memory unit.

[0082] Note that the number of first learning models 10A stored in the learned model storage unit 43 is limited to one. It is not fixed, for example, the machine learning method, the type of wafer W (size, thickness, film type, etc.), polishing process Types of head 2200, differences in the mechanism and material of top ring 221, and types of membrane 2212. Types of retainer rings 2213, types of polishing fluids, types of dresser discs 2300 , type of cleaning fluid, type of data included in operating status information, polishing pad status information included Multiple learning models with different conditions, such as the type of data being used, may be stored. In this case, the learning data storage unit 42 is configured to accommodate multiple learning models with different conditions. It is sufficient to store multiple types of training data with the same data structure.

[0083] Figure 10 shows an example of the first learning model 10A and the first training data 11A. The first training data 11A used for machine learning of the first learning model 10A is, It consists of condition information and polishing pad condition information.

[0084] The operational state information constituting the first learning data 11A is processed by the substrate processing device 2. Top ring state information showing the state of the top ring 221 during the polishing process of wafer W, Polishing table status information showing the condition of the polishing table 220, and the condition of the polishing fluid supply nozzle 222. Polishing fluid supply nozzle status information, dresser status information indicating the status of dresser 223, and It also includes atomizer status information indicating the state of atomizer 224.

[0085] The top ring status information included in the operating status information is the rotation speed of the top ring 221, and the top Rotational torque of the spring 221, swing position of the top ring 221, swing of the top ring 221 Dynamic torque, height of top ring 221, lifting torque of top ring 221, membrane pressure Pressure within force chambers 2212a to 2212d (membrane pressure), membrane pressure chamber 2212a The flow rate of the pressurized fluid supplied to ~2212d (membrane flow rate), the flow rate of membrane 2212 Condition, pressure inside retainer ring pressure chamber 2214a (retainer ring airbag pressure) (force), flow rate of pressure fluid supplied to retainer ring pressure chamber 2214a (retainer ring Includes at least one of the following: bag flow rate, and condition of retainer ring 2213. The condition of membrane 2212 can be expressed, for example, by surface properties, expansion / contraction state, thickness, etc. Furthermore, information on the usage status of membrane 2212 (usage time, whether it was replaced or not), and the condition of the top ring. The condition of the retainer ring 2213 is set based on polishing table condition information, etc. The condition is expressed, for example, by surface texture, flatness, thickness, cross-sectional shape, and wear and dirt on the inner circumference. Information on the usage status of retainer ring 2213 (usage time, whether it was replaced), and the condition of the top ring. The settings are determined based on polishing table condition information, etc. Membrane 2212 and retainer ring The condition of 2213 may change over time, for example, during the polishing process.

[0086] The polishing table status information included in the operating status information is the rotation speed of the polishing table 220, polishing The rotational torque of the table 220, the surface temperature of the polishing pad 2200, and the polishing pad 220 Includes at least one condition of 0. The condition of the polishing pad 2200 is, The condition of the polishing pad 2200 at a point in time prior to the target point in the polishing pad status information. This indicates conditions such as surface texture, flatness, cleanliness, temperature, wetness, and friction. It is expressed by coefficients, etc., and the usage status of the polishing pad 2200 (usage time, membrane pressure during use, etc.) Retainer ring airbag pressure, presence or absence of dressing, replacement status, polishing pad 2200 Images of the surface, top ring condition information, polishing table condition information, polishing fluid supply It is set based on nozzle status information, dresser status information, atomizer status information, etc. Polishing The condition of pad 2200 may change over time, for example, during the polishing process.

[0087] The polishing fluid supply nozzle status information included in the operating status information includes the flow rate of the polishing fluid and the polishing fluid This includes at least one of the dropping position and the temperature of the polishing fluid. Note that there may be multiple types of polishing fluid. If the abrasive fluid is (e.g., polishing liquid, pure water, chemical solution, dispersant, etc.), the flow rate for each type is specified. It is sufficient if it includes at least one of the dropping position for each type and the temperature for each type, for example If the polishing fluid is polishing liquid and pure water, the flow rate of the polishing liquid, the dropping position of the polishing liquid, and the polishing The polishing fluid temperature, the flow rate of pure water, the dropping position of the pure water, and the temperature of the pure water are all included in the above. That would be fine.

[0088] The dresser status information included in the operating status information is the rotation speed of dresser 223, dresser 22 3 Rotational torque, oscillating position of dresser 223, oscillating speed of dresser 223, dresser 22 3 oscillating torque, height of dresser 223, polishing pad 223 dresser disc 2230 The pressing force when making contact with 0, and the condition of the dresser disc 2230 Includes at least one. The condition of the dresser disc 2230 is, for example, dresser Disk 2230 usage status (usage time, pressing force during use, replacement status, dresser day) The dresser disc 2230 is set based on an image of the surface of the disc 2230. This indicates the degree of wear. The condition of the dresser disc 2230 is, for example, during the polishing process. It is also acceptable if it changes over time.

[0089] The atomizer status information included in the operating status information is the flow rate of the cleaning fluid and the dropping position of the cleaning fluid. and include at least one of the pressures of the cleaning fluid.

[0090] Furthermore, the operating status information includes the internal environment information of the device, which indicates the environment of the space where the polishing process is performed. It may also include, and the internal environment information of the device included in the operating status information is shaped by the housing 20. Temperature, humidity, atmospheric pressure, and airflow in the created internal space (for each of the first to fourth polishing sections 22A to 22D) It includes at least one of the following: oxygen concentration and sound. In addition, the operating state information is the polishing process history. It may further include processing performance information that indicates the operating status information For example, after the polishing pad 2200 is replaced, the polishing process using the polishing pad 2200 This includes at least one of the cumulative number of wafers W used and the cumulative usage time when the process was performed. nothing.

[0091] The polishing pad state information that constitutes the first learning data 11A is the operation indicated by the operation state information. This information indicates the state of the polishing pad 2200 when the substrate processing device 2 is operating. In this embodiment, the polishing pad state information is the condition of the polishing surface of the polishing pad 2200. This is condition information indicating the start of the polishing process. Condition information includes, for example, when the polishing process has started. This is included in the polishing process period from start to finish (the time required for polishing one wafer). The distribution of polishing debris, flatness of the polished surface, surface roughness, temperature, wetness, and at the target time point. , including at least one of the coefficients of friction. During the polishing process, the top ring 221 is on the wafer W The action of pressing the polishing pad 2200 against the polishing pad 220, the polishing fluid supply nozzle 222 The operation involves supplying polishing fluid to the dresser 223 and the dresser disc 2230 polishing pad. The operation of dressing the polishing pad 2200 by bringing it into contact with 2200, and atomizer 2 The operation of 24 involves spraying cleaning fluid onto the polishing pad 2200.

[0092] The learning data acquisition unit 400 refers to the polishing test information 31 and, if necessary, uses - By accepting user input operations via the terminal device 6, the first training data 11A To obtain, for example, the learning data acquisition unit 400 acquires the polishing test table of the polishing test information 31. Refer to 310 to see when the polishing test identified by the test ID was performed on the top ring. Grinding status information, polishing table status information, polishing fluid supply nozzle status information, dresser status information, And atomizer status information (top ring 221, polishing table 220, polishing fluid supply node The time-series data of each sensor in the Zuru 222, Dresser 223, and Atomizer 224 The data (data) is obtained as operating status information.

[0093] In this embodiment, the operating state information is obtained from the time-series data of the sensor group shown in Figure 10. The following explains how to acquire it as such, but polishing unit 22 (especially top ring 221 , polishing table 220, polishing fluid supply nozzle 222, dresser 223 and atomizer 22 4) may be modified as appropriate depending on the configuration. Also, as operating status information, instructions to the module Command values ​​may be used, or parameters may be converted from the sensor's detected value or the command value to the module. You may use a data set, or you may use parameters calculated based on the detection values ​​of multiple sensors. This is also acceptable. Furthermore, the operating status information can be obtained as time-series data for the entire polishing process. Alternatively, the data may be obtained as time-series data for the target period, which is part of the polishing process. It may also be obtained as time point data at a specific point in time. As described above, operating state information If the definition of the report is changed, the first learning model 10A and the first training data 11A will be modified. The data structure of the input data can be changed as appropriate.

[0094] Furthermore, the learning data acquisition unit 400 refers to the polishing test table 310 of the polishing test information 31. By comparing, the test result information (research) of polishing tests identified by the same test ID is obtained. The time-series data of the polishing pad measuring instrument (Figure 8) is used to determine the polishing pad in relation to the above operating state information. The status information is obtained. The polishing pad measuring device measures the polishing surface of the polishing pad 2200. If the measuring instrument is capable of surface measurement, the learning data acquisition unit 400 performs surface measurement The value is obtained as polishing pad status information.

[0095] In this embodiment, the polishing pad condition information is as shown in Figure 10. In the case of a report, the distribution of polishing debris, the flatness of the polished surface, the surface roughness, and the temperature are all important factors. It may include at least one of the following: wetness and coefficient of friction. Also, the condition of the polishing pad. The information may be calculated by substituting the measured values ​​of the polishing pad measuring instrument into a predetermined calculation formula. Furthermore, the operating status information may include, for example, time-series data for the entire polishing process or the polishing process period. If the data is obtained as time-series data for the target period, which is part of the polishing pad condition information, The report may be obtained as time-series data for the entire polishing process or as time-series data for the target period. That's fine, and the data can be obtained as time-series data at the end of the polishing process or as time-series data at the target point in time. Furthermore, operational status information can be acquired, for example, as time-series data at a specific point in time. If so, the polishing pad condition information will be treated as time-series data at that specific point in time. It may be obtained. As described above, if you change the definition of the polishing pad state information, the first The data structure of the output data in the learning model 10A and the first training data 11A is appropriate. You can change it as you see fit.

[0096] The first learning model 10A employs, for example, a neural network structure. It includes an input layer 100, an intermediate layer 101, and an output layer 102. Between each layer, each Synapses (not shown) that connect to each neuron are formed, and each synapse has a weight associated with it. A group of weight parameters consisting of the weights of each synapse is adjusted by machine learning.

[0097] The input layer 100 has a number of neurons corresponding to the operation state information as input data, and each value of the operation state information is input to each neuron. The output layer 102 has a number of neurons corresponding to the polishing pad state information as output data, and the prediction result (inference result) of the polishing pad state information with respect to the operation state information is output as output data. When the first learning model 10A is composed of a regression model, the polishing pad state information is output as a numerical value normalized within a predetermined range (for example, 0 to 1). Also, when the first learning model 12A is composed of a classification model, the polishing pad state information is output as a score (accuracy) for each class, as a numerical value normalized within a predetermined range (for example, 0 to ) [[ID=Incorrect range value in the original text, assuming it should be 0 to 1 as in the previous sentence]] 1. output respectively.

[0098] (Machine Learning Method) FIG. 11 is a flowchart showing an example of a machine learning method by the machine learning device 4.

[0099] First, in step S100, the learning data acquisition unit 400 acquires a desired number of first learning data 11A from the polishing test information 31 and the like as preliminary preparation for starting machine learning, and stores the acquired first learning data 11A in the learning data storage unit 42. Here, the number of the first learning data 11A to be prepared may be set in consideration of the inference accuracy required for the finally obtained first learning [[ID=Model number in the original text seems incorrect, assuming it should be 10A]] model 10A.

[0100] Next, in step S110, the machine learning unit 401 starts learning Prepare the first learning model 10A. Here, we prepare the first learning model 10 before training. A consists of the neural network model illustrated in Figure 10, and each synapse The weights are set to their initial values.

[0101] Next, in step S120, the machine learning unit 401 stores the learning data in the learning data storage unit 42. From multiple sets of first training data 11A that have been stored, for example, one set of first training data is randomly selected. Obtain data 11A for use.

[0102] Next, in step S130, the machine learning unit 401 processes a set of first training data 1 The operational state information (input data) contained in 1A is used in the first pre-training (or training) state. This is input to the input layer 100 of the learning model 10A. As a result, the output of the first learning model 10A The force layer 102 outputs polishing pad state information (output data) as an inference result, but The output data is generated by the first trained model 10A before (or during) training. Therefore, in the pre-training (or training) state, the output data that is output as an inference result is The data is different from the polishing pad state information (correct label) included in the first training data 11A. This will provide the necessary information.

[0103] Next, in step S140, the machine learning unit 401 takes in step S120 Polishing pad state information (correct label) included in the first set of training data 11A obtained. In step S130, the polishing pad state information output from the output layer as an inference result is also available. (The process of comparing with the output data and adjusting the weight of each synapse (backpropagation) Machine learning is performed by executing ). As a result, the machine learning unit 401 is in operation state The first learning model 10A is trained to recognize the correlation between the information and the polishing pad state information.

[0104] Next, in step S150, the machine learning unit 401 will determine if a predetermined learning termination condition is met. Whether or not this was done is determined, for example, by the polishing pad state information included in the first training data 11A (correct answer Error based on the label and the polishing pad state information (output data) output as an inference result. The evaluation value of the difference function and the untrained first training data stored in the training data storage unit 42. The decision is made based on the remaining number of 11A units.

[0105] In step S150, the machine learning unit 401 determined that the learning termination condition had not been met. If it is determined to continue machine learning (No in step S150), return to step S120. Furthermore, steps S120 to S140 are performed on the first learning model 10A that is currently being learned. The first training data 11A is used for the exercise and performed multiple times. Meanwhile, in step S150 When the machine learning unit 401 determines that the learning termination conditions have been met and it is time to terminate machine learning, If successful (Yes in step S150), proceed to step S160.

[0106] Then, in step S160, the machine learning unit 401 is associated with each synapse. The first trained model 10A (adjusted weights) is generated by adjusting the weights. The parameter group is stored in the trained model storage unit 43, and a series of machine learning methods shown in Figure 11 are performed. The process ends. In the machine learning method, step S100 is the training data storage process, step Steps S110 to S150 are the machine learning process, and step S160 is the trained model storage process. Do so.

[0107] As described above, according to the machine learning device 4 and the machine learning method according to this embodiment, the top ring state information, polishing table state information, polishing fluid supply nozzle state information, dresser state information and atomizer state information, a first learning model 10A capable of predicting (inferring) the state of the polishing pad 2200 is provided from the operation state information including the polishing pad state information indicating the state of the polishing pad. This can be done.

[0108] (Information processing device 5) FIG. 12 is a block diagram showing an example of the information processing device 5 according to the first embodiment. FIG. 1 [ 3 is a functional explanatory diagram showing an example of the information processing device 5 according to the first embodiment. The information processing device [ 5 includes a control unit 50, a communication unit 51, and a learned model storage unit 52. [

[0109] [ [ The control unit 50 functions as an information acquisition unit 500, a state prediction unit 501, and an output processing unit 502. [ The communication unit 51 is connected to an external device (for example, a substrate processing device 2, a data [ base device 3, a machine learning device 4, and a user terminal device 6, etc.) via the network 7, and functions as a communication interface for transmitting and receiving various data. [ [

[0110] [​​​​​​​​​​​​​ Refer to 301 to see the top ring when the wafer W is polished. Status information, polishing table status information, polishing fluid supply nozzle status information, dresser status information, and Next, atomizer status information is acquired as operating status information. Polishing process is in progress. When performing "real-time prediction processing" of polishing pad state information for wafer W, The information acquisition unit 500 receives information regarding the status of the apparatus from the substrate processing apparatus 2 that is performing the polishing process. By receiving reports R in real time, the polishing process is being carried out on the wafer W. Top ring status information, polishing table status information, polishing fluid supply nozzle status information, dresser Status information and atomizer status information are acquired as operating status information as needed. Polishing process When performing "pre-prediction processing" of polishing pad state information for wafer W before the polishing is performed, The information acquisition unit 500 obtains substrate recipe information 26 from the substrate processing apparatus 2 which is scheduled to perform the polishing process. The apparatus when it receives a 6 and the polishing unit 22 operates according to the substrate recipe conditions 266. By simulating the state information, when polishing is performed on the wafer W... Top ring status information, polishing table status information, polishing fluid supply nozzle status report, dresser Status information and atomizer status information are acquired as operating status information.

[0112] The state prediction unit 501 receives the operating state information acquired by the information acquisition unit 500 as described above. By inputting this as input data into the first learning model 10A, the operating state information indicated Polishing pad status information: This indicates the state of the polishing pad when the substrate processing device 2 is operating in the operating state. Predicts information (in this embodiment, condition information).

[0113] The trained model storage unit 52 stores the first trained model used in the state prediction unit 501. This is a database for storing model 10A. Note that the trained model is stored in the trained model storage unit 52. The number of first learning models 10A is not limited to one; for example, machine learning methods, wafer Type of W (size, thickness, film type, etc.), type of polishing pad 2200, top ring 221 Differences in mechanism and materials, types of membrane 2212, types of retainer ring 2213, polishing flow The following information is included in the operating status: type of body, type of dresser disc 2300, type of cleaning fluid, and operating status information. Multiple conditions can differ, such as the type of data and the type of data included in the polishing pad condition information. A number of pre-trained models may be stored and made available selectively. The memory unit 52 is connected to an external computer (for example, a server computer or a cloud computer). The memory unit of the external controller may be used as a substitute, in which case the state prediction unit 501 will be the external controller. You just need to access the computer.

[0114] The output processing unit 502 outputs the polishing pad state information generated by the state prediction unit 501. Output processing is performed to output the polishing pad status information to the substrate. For example, the output processing unit 502 outputs the polishing pad status information to the substrate. By transmitting this polishing pad status information to the processing device 2 and the user terminal device 6, the display based on the polishing pad status information is shown. The screen may be displayed on the substrate processing device 2 or the user terminal device 6, and the polishing pad status information may also be displayed. By sending the information to the database device 3, the polishing pad status information is transmitted to the production history information 30. You may register.

[0115] (Information processing methods) Figure 14 is a flowchart showing an example of an information processing method by the information processing device 5. Below, the user operates the user terminal device 6 to control the polishing pad state for a specific wafer W. This section describes an example of how to perform "post-event prediction processing" of information.

[0116] First, in step S200, the user tells the user terminal device 6 the prediction target When an input operation is performed to enter a wafer ID that identifies wafer W, the user terminal device 6 then... The wafer ID is transmitted to the information processing device 5.

[0117] Next, in step S210, the information acquisition unit 500 of the information processing device 5 performs the following steps The wafer ID transmitted at 200 is received. In step S211, the information acquisition unit 5 00 uses the wafer ID received in step S210 to determine the polishing history of the production history information 30. By referring to the cable 301, polishing is performed on the wafer W identified by that wafer ID. Obtain operational status information when the operation is performed.

[0118] Next, in step S220, the state prediction unit 501 obtains the data acquired in step S211. By inputting the recorded operating state information as input data into the first learning model 10A, the operating state information is processed. The polishing pad state information is generated as output data for the working state information, and the polishing pad 22 Predict the state of 00.

[0119] Next, in step S230, the output processing unit 502 generates in step S220 As an output process to output the polishing pad status information, the polishing pad status information is used -Send to the user terminal device 6. Note that the destination for sending the polishing pad status information is the user terminal device 6. Alternatively, database device 3 may also be used.

[0120] Next, in step S240, the user terminal device 6 performs the transmission process in step S200. In response to that, when the polishing pad status information transmitted in step S230 is received, Based on the polishing pad status information, the display screen shows the polishing pad 220 The state of 0 is visible to the user. In the above information processing method, step S210, Step S211 is the information acquisition process, step S220 is the state prediction process, and step S230 is the output process. This corresponds to the engineering process.

[0121] As described above, according to the information processing device 5 and information processing method of this embodiment, polishing process Top ring status information, polishing table status information, polishing fluid supply nozzle status information The first learning model includes operational state information, including dresser state information and atomizer state information. When input to 10A, the polishing pad condition information (condition) for the said operating state information is transmitted. Since (information) is predicted, the polishing pad 2200 will be adjusted according to the operating state of the substrate processing device 2. It is possible to accurately predict the state.

[0122] (Second embodiment) In the second embodiment, the polishing pad status information indicates the remaining life of the polishing pad 2200. Information, and at least one piece of polishing quality information indicating the polishing quality of the polishing pad 2200. This differs from the first embodiment in this respect. Below, the machine learning apparatus 4a and according to the second embodiment will be described. The information processing device 5a will now be described, focusing on the differences from the first embodiment.

[0123] Figure 15 is a block diagram showing an example of a machine learning device 4a according to the second embodiment. Figure 16 shows an example of the second learning model 10B and the second training data 11B. The second training data 11B is used for machine learning of the second training model 10B.

[0124] The polishing pad state information that constitutes the second training data 11B is the remaining polishing pad 2200 The remaining life information, which indicates the lifespan, and the polishing quality information, which indicates the polishing quality of the polishing pad 2200, are limited. There is at least one. The remaining lifespan of the polishing pad 2200 is, for example, when the polishing pad 2200 reaches the end of its lifespan. It is determined by the number of times it can be used or the usable time until it reaches its limit. Polishing quality refers to the degree of polishing of the wafer W, such as the polishing rate, polishing profile, and residual film. Information on the degree of polishing, and defects in wafer W such as scratches and corrosion (diff This is determined by substrate defect information regarding the degree and presence or absence of defects. Furthermore, the second training data 1 The operational state information constituting 1B is the same as in the first embodiment, so its explanation is omitted.

[0125] The learning data acquisition unit 400 refers to the polishing test information 31 and, if necessary, uses - By accepting user input operations via the terminal device 6, the second training data 11B The polishing test information 31 includes, for example, test result information, the polishing pad used for the test. When repeated polishing is performed using a polishing test device, the lifespan of the polishing pad 2200 When the remaining lifespan is reached, the remaining lifespan information is set to "0," and the further back in time you go, the larger the value set. The remaining life information obtained, along with measurements taken with measuring instruments such as optical microscopes and scanning electron microscopes (SEMs), Polishing quality information is registered. The learning data acquisition unit 400 then acquires polishing test information. The polishing tests performed when the polishing tests identified by the test ID from the 31 polishing test tables 310 were conducted. By obtaining the results information, we can acquire information on remaining lifespan and polishing quality.

[0126] The machine learning unit 401 incorporates multiple second training data 11B into the second learning model 10B. The operating state information and polishing pad state information (remaining life) included in the second learning data 11B are used. The second learning model 10B learns the correlation between information and at least one of the polishing quality information. This process generates a pre-trained second model 10B.

[0127] Figure 17 shows the information processing device 5a functioning as an information processing device 5a according to the second embodiment. This is a block diagram showing an example. Figure 18 shows an example of an information processing device 5a according to the second embodiment. This is a functional diagram illustrating the features.

[0128] The information acquisition unit 500, similar to the first embodiment, acquires top ring status information, polishing table information, etc. Status information, polishing fluid supply nozzle status information, dresser status information, and atomizer status information Retrieves operational status information, including that information.

[0129] The state prediction unit 501 receives the operating state information acquired by the information acquisition unit 500 as described above. By inputting this as input data into the second learning model 10B, the operating state information indicated Polishing pad status information: This indicates the state of the polishing pad when the substrate processing device 2 is operating in the operating state. Predicts the report (at least one of the remaining lifespan information and polishing quality information).

[0130] The output processing unit 502, similar to the first embodiment, processes the data generated by the state prediction unit 501. Output for outputting polishing pad condition information (at least one of remaining life information and polishing quality information) Force processing is performed. For example, the output processing unit 502 sends the polishing pad status information to the substrate processing unit 2 By transmitting this information to the user terminal device 6, the display screen based on the polishing pad status information is displayed on the circuit board. The polishing pad status information may be displayed on the processing device 2 or the user terminal device 6, or it may be displayed as data. By transmitting the polishing pad status information to the base device 3, the polishing pad status information is registered in the production history information 30. This is also possible. In that case, the output processing unit 502 will, for example, use the remaining life information to determine the polishing pad 2200. If the remaining lifespan falls below the specified number of predicted cycles or time, or if the polishing quality information indicates that the polishing is not working properly, If the polishing quality falls below the specified standard quality, a notice of replacement and replacement work for the polishing pad 2200 will be given. The circuit board processing unit displays information such as the procedure manual, the time required for the replacement work, and the price of replacement parts. It may also be configured to transmit to the device 2 or the user terminal device 6. In addition, the output processing unit 502 is on the board If the processing device 2 has a function to automatically replace the polishing pad 2200, then the polishing pad 2 Alternatively, a command to automatically replace 200 may be sent to the substrate processing device 2, or the polishing pad may be replaced. An order to order replacement parts for the D2200 was issued to the inventory manager responsible for managing the stock of polishing pads 2200. The data may also be transmitted to a storage management device (not shown).

[0131] As described above, according to the information processing apparatus 5a and information processing method of this embodiment, polishing In the process, top ring status information, polishing table status information, polishing fluid supply nozzle status information Operating status information, including the report, dresser status information, and atomizer status information, is used in the second learning model. By inputting into 10B, the polishing pad status information (remaining life information) for the said operating status information is transmitted. Since at least one of the reports and polishing quality information is predicted, the operating state of the substrate processing apparatus 2 Accordingly, the condition of the polishing pad 2200 can be appropriately predicted.

[0132] (Other embodiments) The present invention is not limited to the embodiments described above, and will proceed within the scope of the spirit of the invention. It is possible to implement various modifications within the system. And all of these are based on the technical ideas of the present invention. It is included in the idea.

[0133] In the above embodiment, the database device 3, the machine learning device 4, and the information processing device 5 are separate. Although it was described as being composed of these devices, these three devices are actually composed of a single device. Alternatively, any two of those three devices may be composed of a single device. It may be included. Also, at least one of the machine learning device 4 and the information processing device 5 is a substrate processing device. It may be incorporated into the control unit 26 or user terminal device 6 in the second location.

[0134] In the above embodiment, the substrate processing apparatus 2 is described as comprising each of the units 21 to 25. However, the substrate processing apparatus 2 only needs to be equipped with at least a polishing unit 22, and other units The "t" can be omitted.

[0135] In the above embodiment, as a learning model that realizes machine learning by the machine learning unit 401, We have explained the case where a neural network is used, but what if other machine learning models are used? You may do so. Other machine learning models include, for example, tree-type models such as decision trees and regression trees, and Ensemble learning such as ging and boosting, recurrent neural networks, convolutional Neural networks, including LSTM and other neural network types (including deep learning) Classes such as hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, and k-means clustering. Taring type, principal component analysis, factor analysis, multivariate analysis such as logistic regression, support vector Examples include machines, etc.

[0136] (Machine learning programs and information processing programs) This invention provides a program that enables the computer 900 to function as one of the components of the machine learning device 4. The computer 900 implements each step of the machine learning program (RAM) and machine learning method. It can also be provided in the form of a program (machine learning program) to perform the task. The present invention provides a way for the computer 900 to function as one of the components of the information processing device 5. Each step of the program (information processing program) or the information processing method according to the above embodiment The present invention is a program (information processing program) designed to be executed by computer 900. It can also be provided.

[0137] (Inference device, inference method, and inference program) The present invention relates to the information processing device 5 (information processing method or information processing program) according to the above embodiment. This is not only due to the manner in which the polishing pad state information is inferred, but also to the inference equipment used to infer the state information of the polishing pad. It can also be provided in the form of an inference device (inference method or inference program). In that case, an inference device The (inference method or inference program) includes memory and a processor, of which A processor may perform a series of operations. This series of operations may include: Information acquisition process (information acquisition step) to acquire operating status information, and the information acquisition process acquires operating status information When this information is obtained, the polishing process occurs when the substrate processing device operates in the operating state indicated by the operating state information. Polishing pad condition information (condition information, remaining life information, or polishing quality) that indicates the condition of the pad. It includes an inference process (inference step) that infers information.

[0138] By providing an inference device (inference method or inference program), the information processing device can be implemented. Compared to the installation method, it becomes easier to apply to various devices. Inference device (inference method or hypothesis When the program infers polishing pad state information, the machine learning apparatus and according to the above embodiment The state prediction unit performs the following using the trained model generated by the machine learning method. It is obvious to those skilled in the art that inference methods may be applied. [Explanation of symbols]

[0139] 1...Substrate processing system, 2...Substrate processing device, 3...Database device, 4, 4a... Machine learning device, 5, 5a... Information processing device, 6...User terminal device, 7...Network, 10A...First learning model, 10B...Second learning model, 11A...First training data, 11B...Second training data, 20...Housing, 21...Load / Unload Unit 22... Polishing unit, 22A~22D... Polishing section, 23... Substrate transport unit, 24...Finishing unit, 25...Film thickness measurement unit, 26...Control unit, 30…Production history information, 31…Polishing test information, 40...Control unit, 41...Communication unit, 42...Learning data storage unit, 43...Trained model memory unit, 50...Control unit, 51...Communication unit, 52...Learned model storage unit, 220... Polishing table, 221... Top ring, 222... Polishing fluid supply nozzle, 223... Dresser, 224... Atomizer, 225... Environmental sensor 260...Control unit, 21...Communication unit, 262...Input unit, 263...Output unit, 264...Storage unit, 300...Wafer history table, 301...Polishing history table, 310...Polishing test table, 400...Training data acquisition unit, 401...Machine learning unit, 500... Information acquisition unit, 501... State prediction unit, 502... Output processing unit, 900... Computer 2200... Polishing pad, 2210... Top ring body, 2211... Carrier, 2212...Membrane, 2212a~2212d...Membrane pressure chamber, 2213...Retainer ring, 2214...Retainer ring airbag, 2214a...Retainer ring pressure chamber, 2230...Dresser disc

Claims

1. The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and an information acquisition unit that acquires operating state information including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and atomizer state information indicating the state of the atomizer, as the operating state when the substrate processing apparatus is in operation, The system includes a state prediction unit that predicts the polishing pad state information in relation to the operating state information by inputting the operating state information acquired by the information acquisition unit into a learning model that has been trained by machine learning to determine the correlation between the operating state information and the polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: Information processing device.

2. The operating state of a substrate processing apparatus that performs chemical mechanical polishing of a substrate, comprising a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, includes operating state information when the apparatus is in operation, including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and an information acquisition unit that acquires atomizer status information indicating the state of the atomizer, The system includes a state prediction unit that predicts the polishing pad state information in relation to the operating state information by inputting the operating state information acquired by the information acquisition unit into a learning model that has been trained by machine learning to determine the correlation between the operating state information and the polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, This includes at least one of the cumulative number of substrates used when polishing is performed using the polishing pad after the polishing pad has been replaced, and the cumulative usage time of the polishing pad. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: Information processing device.

3. The aforementioned top ring is The top ring body is moved by a rotational movement mechanism, a vertical movement mechanism, and a swinging movement mechanism, A membrane housed in the top ring body presses the substrate against the polishing pad in accordance with the pressure fluid supplied to the membrane pressure chamber, The device comprises a retainer ring positioned on the outer circumference of the membrane, which presses the polishing pad in accordance with the pressure fluid supplied to the retainer ring pressure chamber, The top ring status information included in the aforementioned operating status information is, The rotational speed of the top ring, The rotational torque of the top ring, The swing position of the top ring, The oscillation speed of the top ring, The oscillation torque of the top ring, The height of the aforementioned top ring, The lifting torque of the top ring, The pressure inside the membrane pressure chamber, The flow rate of the pressure fluid supplied to the membrane pressure chamber, The condition of the aforementioned membrane, The pressure inside the retainer ring pressure chamber, The flow rate of the pressurized fluid supplied to the retainer ring pressure chamber, and The conditions of the retainer ring include at least one of the conditions of the retainer ring, The information processing apparatus according to claim 1 or claim 2.

4. The polishing table status information included in the operating status information is, The rotation speed of the polishing table, The rotational torque of the aforementioned polishing table, The surface temperature of the polishing pad, and, The conditions of the polishing pad include at least one of the following: The information processing apparatus according to any one of claims 1 to 3.

5. The polishing fluid supply nozzle state information included in the operating state information is, The flow rate of the polishing fluid, The dropping position of the polishing fluid, The temperature of the polishing fluid, The concentration of the polishing fluid, and The polishing fluid includes at least one of the cleanliness levels, The information processing apparatus according to any one of claims 1 to 4.

6. The aforementioned dresser is The dresser disc is moved by a rotational movement mechanism, a vertical movement mechanism, and a swinging movement mechanism, The dresser status information included in the operating status information is, The rotation speed of the dresser, The rotational torque of the dresser, The swinging position of the dresser, The oscillation speed of the dresser, The oscillation torque of the dresser, The height of the dresser, The pressing force when the dresser disc is brought into contact with the polishing pad, and The dresser disc includes at least one of the conditions, The information processing apparatus according to any one of claims 1 to 5.

7. The aforementioned operating status information is, The system further includes information about the environment inside the apparatus, which indicates the environment of the space in which the chemical mechanical polishing treatment is performed. The internal device environment information included in the operating state information is, The temperature of the aforementioned space, Humidity of the aforementioned space, The air pressure in the aforementioned space, Airflow in the aforementioned space, The oxygen concentration in the aforementioned space, and Including at least one sound in the aforementioned space, The information processing apparatus according to any one of claims 1 to 6.

8. The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: The information processing apparatus according to claim 2.

9. The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, The cumulative number of substrates used and the cumulative usage time when polishing is performed using the polishing pad after the polishing pad has been replaced, including at least one of these, The information processing apparatus according to claim 1.

10. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: The information processing apparatus according to claim 1.

11. The aforementioned polishing pad condition information is, Remaining life information indicating the remaining life of the polishing pad, and The polishing quality information includes at least one piece of polishing quality information indicating the polishing quality of the polishing pad, The information processing apparatus according to claim 1.

12. An inference device comprising memory and a processor, The aforementioned processor, The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and the substrate processing apparatus performs chemical mechanical polishing of the substrate, and the information acquisition process acquires operating state information including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and atomizer state information indicating the state of the atomizer, as the operating state when the substrate processing apparatus is in operation. When the operating state information is acquired in the information acquisition process, an inference process is performed to infer polishing pad state information, which indicates the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: Reasoning device.

13. An inference device comprising memory and a processor, The aforementioned processor, The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and the substrate processing apparatus performs chemical mechanical polishing of the substrate, and the information acquisition process acquires operating state information including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and atomizer state information indicating the state of the atomizer, as the operating state when the substrate processing apparatus is in operation. When the operating state information is acquired in the information acquisition process, an inference process is performed to infer polishing pad state information, which indicates the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, This includes at least one of the cumulative number of substrates used when polishing is performed using the polishing pad after the polishing pad has been replaced, and the cumulative usage time of the polishing pad. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: Reasoning device.

14. A learning data storage unit stores multiple sets of learning data, each set comprising: a polishing table that rotatably supports a polishing pad; a top ring that presses a substrate against the polishing pad; a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad; a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad; and an atomizer that sprays cleaning fluid onto the polishing pad, and storing multiple sets of learning data, each set comprising: top ring state information indicating the state of the top ring; polishing table state information indicating the state of the polishing table; polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle; dresser state information indicating the state of the dresser; atomizer state information indicating the state of the atomizer; and polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus is operated in the operating state indicated by the operating state information; A machine learning unit that inputs multiple sets of the aforementioned training data into the learning model to train the learning model to recognize the correlation between the operating state information and the polishing pad state information, The system includes a trained model storage unit that stores the trained model in which the correlation has been learned by the machine learning unit, The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: Machine learning device.

15. A learning data storage unit stores multiple sets of learning data, each set comprising: a polishing table that rotatably supports a polishing pad; a top ring that presses a substrate against the polishing pad; a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad; a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad; and an atomizer that sprays cleaning fluid onto the polishing pad, and storing multiple sets of learning data, each set comprising: top ring state information indicating the state of the top ring; polishing table state information indicating the state of the polishing table; polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle; dresser state information indicating the state of the dresser; atomizer state information indicating the state of the atomizer; and polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus is operated in the operating state indicated by the operating state information; A machine learning unit that inputs multiple sets of the aforementioned training data into the learning model to train the learning model to recognize the correlation between the operating state information and the polishing pad state information, The system includes a trained model storage unit that stores the trained model in which the correlation has been learned by the machine learning unit, The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, This includes at least one of the cumulative number of substrates used when polishing is performed using the polishing pad after the polishing pad has been replaced, and the cumulative usage time of the polishing pad. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: Machine learning device.

16. The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and the substrate processing apparatus performs chemical mechanical polishing of the substrate, and the information acquisition step acquires operating state information including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and atomizer state information indicating the state of the atomizer, as the operating state when the substrate processing apparatus is in operation. The system includes a state prediction step which predicts the polishing pad state information in relation to the operating state information by inputting the operating state information acquired in the information acquisition step into a learning model that has been trained by machine learning to determine the correlation between the operating state information and the polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: Information processing methods.

17. The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and the substrate processing apparatus performs chemical mechanical polishing of the substrate, and the information acquisition step acquires operating state information including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and atomizer state information indicating the state of the atomizer, as the operating state when the substrate processing apparatus is in operation. The correlation between the aforementioned operating state information and the polishing pad state information, which indicates the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the said operating state information, is determined by machine learning. The system includes a state prediction step, which involves inputting the operating state information acquired in the information acquisition step into a more trained learning model to predict the polishing pad state information in relation to the operating state information. The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, This includes at least one of the cumulative number of substrates used when polishing is performed using the polishing pad after the polishing pad has been replaced, and the cumulative usage time of the polishing pad. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: Information processing methods.

18. An inference method performed by an inference device comprising memory and a processor, The aforementioned processor, The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and the substrate processing apparatus performs chemical mechanical polishing of the substrate, and the information acquisition step acquires operating state information including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and atomizer state information indicating the state of the atomizer, as the operating state when the substrate processing apparatus is in operation. When the operating state information is acquired in the information acquisition step, an inference step is performed to infer polishing pad state information that indicates the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: Reasoning method.

19. An inference method performed by an inference device comprising memory and a processor, The aforementioned processor, The operating state of a substrate processing apparatus that performs chemical mechanical polishing of a substrate, comprising a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, includes operating state information when the apparatus is in operation, including top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle, dresser state information indicating the state of the dresser, and Furthermore, an information acquisition step is performed to acquire atomizer status information indicating the state of the atomizer, When the operating state information is acquired in the information acquisition step, an inference step is performed to infer polishing pad state information that indicates the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information, The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, This includes at least one of the cumulative number of substrates used when polishing is performed using the polishing pad after the polishing pad has been replaced, and the cumulative usage time of the polishing pad. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: Reasoning method.

20. A learning data storage step involves storing multiple sets of learning data in a learning data storage unit as an operating state when a substrate processing apparatus is in operation, comprising: a polishing table that rotatably supports a polishing pad; a top ring that presses a substrate against the polishing pad; a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad; a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad; and an atomizer that sprays cleaning fluid onto the polishing pad, wherein the substrate processing apparatus performs chemical mechanical polishing of the substrate, and the operating state when the substrate processing apparatus is in operation includes: operating state information including top ring state information indicating the state of the top ring; polishing table state information indicating the state of the polishing table; polishing fluid supply nozzle state information indicating the state of the polishing fluid supply nozzle; dresser state information indicating the state of the dresser; and atomizer state information indicating the state of the atomizer; and polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus is in operation in the operating state indicated by the operating state information; A machine learning process in which multiple sets of the aforementioned training data are input into the training model to train the training model to learn the correlation between the operating state information and the polishing pad state information, The system includes a trained model storage step, which stores the trained model, which has learned the correlation relationship by the machine learning step, in a trained model storage unit. The atomizer state information included in the operating state information is, The flow rate of the cleaning fluid, The dropping position of the cleaning fluid, and The pressure of the cleaning fluid includes at least one of the following: Machine learning methods.

21. The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against the polishing pad, a polishing fluid supply nozzle that supplies polishing fluid to the polishing pad, a dresser that rotatably supports a dresser disc and dresses the polishing pad by bringing the dresser disc into contact with the polishing pad, and an atomizer that sprays cleaning fluid onto the polishing pad, and when the substrate processing apparatus is in operation, the operating states are top ring state information indicating the state of the top ring, polishing table state information indicating the state of the polishing table, and polishing fluid supply information indicating the state of the polishing fluid supply nozzle. A learning data storage step involves storing multiple sets of learning data in a learning data storage unit, each set comprising: nozzle state information; dresser state information indicating the state of the dresser; atomizer state information indicating the state of the atomizer; and polishing pad state information indicating the state of the polishing pad when the substrate processing apparatus operates in the operating state indicated by the operating state information. A machine learning process in which multiple sets of the aforementioned training data are input into the training model to train the training model to learn the correlation between the operating state information and the polishing pad state information, The system includes a trained model storage step, which stores the trained model, which has learned the correlation relationship by the machine learning step, in a trained model storage unit. The aforementioned operating status information is, This further includes processing performance information showing the results of the aforementioned chemical mechanical polishing treatment. The processing performance information included in the operating status information is, This includes at least one of the cumulative number of substrates used when polishing is performed using the polishing pad after the polishing pad has been replaced, and the cumulative usage time of the polishing pad. The aforementioned polishing pad condition information is, This is condition information indicating the condition of the polishing surface of the aforementioned polishing pad. The aforementioned condition information is, The distribution state of polishing debris on the polished surface, The flatness of the polished surface, The surface roughness of the polished surface, The temperature of the polished surface, The degree of wetness of the polished surface, and The coefficient of friction of the polished surface includes at least one of the following: Machine learning methods.

Images