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 appropriately predicting reliability information of a polishing end point detection function showing the reliability of an end detection function for detecting that chemical mechanical polishing processing reaches an end point.SOLUTION: An information processing device 5 includes an information acquisition part 500 for acquiring reliability deterioration element state information including at least one of consumption state information showing a consumption state of a component of a substrate processing device 2 in chemical mechanical polishing processing of a substrate performed by the substrate processing device 2 and processing state information showing a processing state during polishing processing, and a state prediction part 501 for predicting the reliability information of the polishing end point detection function to the reliability deterioration element state information by inputting the reliability deterioration element state information acquired by the information acquisition part 500 in a learning model which learns a correlation between the reliability deterioration element state information and reliability information of a polishing end point detection function showing the reliability of an end point detection function for detecting that the chemical mechanical polishing processing reaches an end point.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 a machine learning method. [Background technology]

[0002] One type of substrate processing equipment that performs various processes on substrates such as semiconductor wafers is the chemical mechanical polishing (CMP) processing equipment. In this equipment, for example, a polishing table with a polishing pad is rotated, and a polishing liquid (slurry) is supplied to the polishing pad from a liquid supply nozzle. The substrate is then pressed against the polishing pad by a polishing head called a top ring, thereby chemically and mechanically polishing the substrate. After polishing, foreign matter such as polishing debris adhering to the substrate is removed by supplying a substrate cleaning fluid to the polished substrate and bringing a cleaning tool into contact with it for scrubbing. The substrate is then dried, completing the series of processes and allowing the process to move on to the next substrate.

[0003] In the series of processes described above, if polishing is insufficient, the insulation of the electronic circuits formed on the finished wafer may be inadequate, potentially leading to short circuits. Conversely, if polishing is excessive, the cross-sectional area of ​​the wiring may decrease, leading to increased resistance or even failure to form the circuit itself. Therefore, recent polishing equipment detects changes in frictional force, motor current, or physical quantities of the wafer to determine the end point of polishing. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2021-028099 [Overview of the project] [Problems that the invention aims to solve]

[0005] Patent Document 1 discloses a technology that improves the accuracy of endpoint detection by having a trained model obtained by machine learning from the waveform of measurement data from the start to the end of polishing output from each of several types of endpoint detection sensors provided in a single polishing unit, and by taking the measurement data from the start to the present time output from each of the several types of endpoint detection sensors during a new polishing as input, and estimating whether the present time is the endpoint timing indicating the end of polishing.

[0006] However, conventional polishing equipment with endpoint detection technology lacked the ability to determine whether the endpoint detection results were reliable, and therefore, the only option was to use the detected endpoint results as they were.

[0007] In view of the above problems, the present invention aims to provide an information processing device, an inference device, a machine learning device, an information processing method, an inference method, and a machine learning method that enable appropriate prediction of reliability information of a polishing endpoint detection function, which indicates the reliability of an endpoint detection function that detects when a chemical mechanical polishing process has reached its endpoint. [Means for solving the problem]

[0008] To achieve the above objective, an information processing apparatus according to one aspect of the present invention is: In the chemical mechanical polishing process of a substrate performed by a substrate processing apparatus, wear status information indicating the wear status of the components of the substrate processing apparatus, and processing status information indicating the processing status during the polishing process. An information acquisition unit that acquires reliability degradation factor state information, including at least one of the following, The system includes a state prediction unit that predicts the reliability information of the polishing endpoint detection function in relation to the reliability reduction factor state information by inputting the reliability reduction factor state information acquired by the information acquisition unit into a learning model that has been trained by machine learning to correlate the aforementioned reliability reduction factor state information with the reliability information of the polishing endpoint detection function, which indicates the reliability of the endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint. [Effects of the Invention]

[0009] According to an information processing apparatus according to an aspect of the present invention, reliability degradation factor state information including at least one of wear state information indicating a wear state of a component of the substrate processing apparatus and processing state information indicating a processing state during polishing in chemical mechanical polishing is input to a learning model, so that reliability information of a polishing end point detection function with respect to the reliability degradation factor state information is predicted. Therefore, it is possible to appropriately predict reliability information of a polishing end point detection function indicating the reliability of an end point detection function for detecting that the chemical mechanical polishing process has reached an end point.

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

Brief Description of Drawings

[0011] [Figure 1] It is an overall configuration diagram showing an example of a substrate processing system 1. [Figure 2] It is a plan view showing an example of a substrate processing apparatus 2. [Figure 3] It is a perspective view showing an example of first to fourth polishing units 22A to 22D. [Figure 4] It is a cross-sectional view schematically showing an example of a top ring 221. [Figure 5] It is a block diagram showing an example of a substrate processing apparatus 2. [Figure 6] It is a hardware configuration diagram showing an example of a computer 900. [Figure 7] It is a data configuration diagram showing an example of production history information 30 managed by a database device 3. [Figure 8] It is a data configuration diagram showing an example of polishing test information 31 managed by a database device 3. [Figure 9] It is a block diagram showing an example of a machine learning apparatus 4 according to the first embodiment. [Figure 10] It is a diagram showing an example of a first learning model 10A and first learning data 11A. [Figure 11]It is a flowchart showing an example of a machine learning method by a machine learning device 4. [Figure 12] It is a block diagram showing an example of an information processing device 5 according to the first embodiment. [Figure 13] It is a functional explanatory diagram showing an example of an information processing device 5 according to the first embodiment. [Figure 14] It is a flowchart showing an example of an information processing method by an information processing device 5. [Figure 15] It is a schematic diagram for explaining an optical sensor 226 provided on a polishing table 220 according to the second embodiment. [Figure 16] It is a block diagram showing an example of a machine learning device 4a according to the second embodiment. [Figure 17] It is a diagram showing an example of a second learning model 10B and a first learning data 11B. [Figure 18] It is a block diagram showing an example of an information processing device 5a according to the second embodiment. [Figure 19] It is a functional explanatory diagram showing an example of an information processing device 5a according to the second embodiment.

Embodiments for Carrying Out the Invention

[0012] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. Hereinafter, the scope necessary for explaining for achieving the object of the present invention will be schematically shown, and mainly the scope necessary for explaining the relevant part of the present invention will be described, and the parts where the explanation is omitted shall be based on known techniques.

[0013] (First Embodiment) FIG. 1 is an overall configuration diagram showing an example of a substrate processing system 1. The substrate processing system 1 according to the present embodiment functions as a system that manages a series of substrate processes including a chemical mechanical polishing process (hereinafter referred to as "polishing process") for flattening the surface of a substrate (hereinafter referred to as "wafer") W such as a semiconductor wafer, and a cleaning process for cleaning the wafer W after the polishing process.

[0014] The substrate processing system 1 comprises, as its main components, a substrate processing device 2, a database device 3, a machine learning device 4, an information processing device 5, and a user terminal device 6. Each device 2 to 6 is, for example, composed of a general-purpose or dedicated computer (see Figure 6 below) and connected to a wired or wireless network 7, enabling the mutual transmission and reception of various types of data (in Figure 1, the transmission and reception of some data are illustrated by dashed arrows). Note that the number of each device 2 to 6 and the connection configuration of the network 7 are 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 is a device that performs a series of substrate processing operations on one or more wafers W, such as loading, polishing, cleaning, drying, film thickness measurement, and unloading. In doing so, the substrate processing apparatus 2 controls the operation of each unit while referring to device setting information 265, which consists of multiple device parameters set for each unit, and substrate recipe information 266, which defines the polishing process status information for the polishing process and the cleaning process conditions for the cleaning process.

[0016] The substrate processing device 2 transmits various reports R to the database device 3, user terminal device 6, etc., according to the operation of each unit. The various reports R include, for example, process information that identifies the wafer W to be processed when substrate processing is performed, device status information that shows the status of each unit when each process is performed, event information detected by the substrate processing device 2, and user (operator, production manager, maintenance manager, etc.) operation information for the substrate processing device 2.

[0017] The database device 3 is a device that manages production history information 30 regarding the history of substrate processing performed on wafers W for actual production, and polishing test information 31 regarding the history of polishing tests (hereinafter referred to as "polishing tests") performed on dummy wafers for testing. In addition to the above, the database device 3 may also store device setting information 265 and substrate recipe information 266, in which case the substrate processing device 2 may refer to this information.

[0018] When the substrate processing device 2 performs substrate processing on the wafer W for mass production, the database device 3 receives various reports R from the substrate processing device 2 as needed and registers them in the production history information 30, thereby accumulating reports R related to substrate processing in the production history information 30.

[0019] When the substrate processing device 2 performs a polishing test on a dummy wafer for testing, the database device 3 receives various reports R (including at least device status information) from the substrate processing device 2 as needed, registers them in the polishing test information 31, and also registers the test results of the polishing test in association with them. As a result, the polishing test information 31 stores reports R and test results related to the polishing test.

[0020] A dummy wafer is a jig that simulates a wafer W. Dummy wafer sensors, such as pressure sensors and temperature sensors, are provided on the surface or inside the dummy wafer to measure the state of the wafer W when polishing is performed, and the measured values ​​from the dummy wafer sensors are registered as test results in the polishing test information 31. The dummy wafer sensors may be provided at one or more locations on the substrate surface of the dummy wafer, or they may be provided across the surface. The polishing test may be performed in the substrate processing apparatus 2 for actual production, or in a test polishing test apparatus (not shown) that can reproduce the same polishing process as the substrate processing apparatus 2.

[0021] The machine learning device 4 operates as the main component of the machine learning learning phase. For example, it acquires a portion of the polishing test information 31 from the database device 3 as the first training data 11A, and generates the first learning model 10A used by the information processing device 5 through machine learning. The trained first learning model 10A is provided to the information processing device 5 via the network 7, recording media, etc.

[0022] The information processing device 5 operates as the main component of the machine learning inference phase, and uses the first learning model 10A generated by the machine learning device 4 to predict the state of the wafer W when the substrate processing device 2 performs polishing on the wafer W for mass production, and transmits the reliability information of the polishing endpoint detection function, which is the result of the prediction, to the database device 3, user terminal device 6, etc. The timing at which the information processing device 5 predicts the reliability information of the polishing endpoint detection function may be after the polishing process is performed (post-prediction processing), during the polishing process is performed (real-time prediction processing), or before the polishing process is performed (pre-prediction processing).

[0023] The user terminal device 6 is a terminal device used by the user, and may be a stationary device or a portable device. The user terminal device 6 accepts various input operations via a display screen such as an application program or a web browser, and displays various information (for example, event notifications, reliability information of the polishing endpoint detection function, production history information 30, polishing test information 31, etc.) via the display screen.

[0024] (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 configured to include a load / unload unit 21, a polishing unit 22, a substrate transport unit 23, a cleaning unit 24, a film thickness measuring unit 25, and a control unit 26, all housed inside a housing 20 which is roughly rectangular in plan view. The load / unload unit 21 is separated from the polishing unit 22, the substrate transport unit 23, and the cleaning unit 24 by a first partition wall 200A, and the substrate transport unit 23 is separated from the cleaning unit 24 by a second partition wall 200B.

[0025] (Load / Unload Unit) The load / unload unit 21 includes first to fourth front loading sections 210A to 210D on which wafer cassettes (FOUP, etc.) capable of storing a large number of wafers W in the vertical direction are placed, a transport robot 211 that can move up and down along the storage direction (vertical direction) of the wafers W stored in the wafer cassette, and a horizontal movement mechanism 212 that moves the transport robot 211 along the direction in which the first to fourth front loading sections 210A to 210D are aligned (the short side direction of the housing 20).

[0026] The transport robot 211 is configured to access wafer cassettes, substrate transport units 23 (specifically, the lifter 232 described later), cleaning units 24 (specifically, the drying chamber 241 described later), and film thickness measuring units 25, each of the first to fourth front load sections 210A to 210D, and is equipped with two upper and lower hands (not shown) for transferring wafers W between them. The lower hand is used when transferring wafers W before processing, and the upper hand is used when transferring wafers W after processing. When transferring wafers W to the substrate transport unit 23 or the cleaning unit 24, a shutter (not shown) provided in the first partition wall 200A is opened and closed.

[0027] (Polishing unit) The polishing unit 22 includes first to fourth polishing sections 22A to 22D, each performing polishing (planarization) of the wafer W. The first to fourth polishing sections 22A to 22D are housed in a housing 20 They are arranged and positioned along the longitudinal direction of the object.

[0028] 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 first to fourth polishing sections 22A to 22D are common to all.

[0029] Each of the first to fourth polishing sections 22A to 22D includes a polishing table 220 to which a polishing pad 2200 having a polishing surface is attached, a top ring (polishing head) 221 for holding a wafer W and polishing the wafer W while pressing it against the polishing pad 2200 on the polishing table 220, a polishing fluid supply nozzle 222 as a polishing fluid supply section for supplying polishing fluid to the polishing pad 2200, a dresser 223 for dressing the polishing surface of the polishing pad 2200, an atomizer 224 for spraying cleaning fluid onto the polishing pad 2200, and an environmental sensor 225 for measuring the state of the internal space of the housing 20 where the polishing process is performed.

[0030] The polishing table 220 is supported by a polishing table shaft 220a and includes a rotary moving mechanism 220b that rotates the polishing table 220 around its axis via a polishing table rotary connector 2201, and a polishing pad surface temperature control mechanism 220c that adjusts the surface temperature of the polishing pad 2200. The polishing pad surface temperature control mechanism 220c has a radiation thermometer 220c1 above the polishing table 220 that measures the surface temperature of the polishing pad 2200 or the surface temperature of the grinding wheel.

[0031] Furthermore, the polishing table 220 of this embodiment has an internal polishing table temperature control mechanism 220d that adjusts the temperature of the polishing table 220 by supplying and discharging temperature-controlled water into and out of the polishing table 220. The internal polishing table temperature control mechanism 220d has a temperature-controlled water supply pipe 220d1 that supplies temperature-controlled water into the polishing table 220 and a temperature-controlled water discharge pipe 220d2 that discharges the temperature-controlled water. The internal polishing table temperature control mechanism 220d has a supply temperature-controlled water thermometer 220d3 in the temperature-controlled water supply pipe 220d1 that measures the temperature of the supplied temperature-controlled water, and a discharge temperature-controlled water thermometer 220d4 in the temperature-controlled water discharge pipe 220d2 that measures the temperature of the discharged temperature-controlled water.

[0032] The top ring 221 is supported by a top ring shaft 221a that is movable in the vertical direction and comprises a top ring rotational movement mechanism 221c that rotates the top ring 221 around its axis, a top ring vertical movement mechanism 221d that moves the top ring 221 in the vertical direction, and a top ring oscillating movement mechanism 221e that oscillates (oscillates) the top ring 221 around the top ring oscillating support shaft 221b as the pivot point.

[0033] The top ring swinging movement mechanism 221e includes a top ring swinging support shaft 221b, a top ring swinging arm 221f that swingably connects the top ring shaft 221a to the top ring swinging support shaft 221b, and a top ring swinging shaft motor 221g that rotationally drives the top ring swinging support shaft 221b.

[0034] The top ring swinging mechanism 221e has a top ring swinging torque sensor 221h at the connection point between the top ring swinging arm 221f and the top ring swinging shaft motor 221g, which detects the top ring swinging torque applied to the top ring swinging arm 221f. Specifically, the top ring swinging torque sensor 221h should detect the torque applied to the top ring swinging arm 221f from the current value of the top ring swinging shaft motor 221g. The current value of the top ring swinging shaft motor 221g is a quantity that depends on the torque of the top ring swinging arm 221f at the connection point to the top ring swinging shaft motor 221g. In this embodiment, the current value of the top ring swinging shaft motor 221g may be the current value supplied to the top ring swinging shaft motor 221g, or a current command value generated in a driver (not shown). Note that the top ring swinging torque may be detected by other methods.

[0035] Furthermore, an acceleration sensor 221j and / or an amplitude sensor (not shown) may be attached to the top ring 221 to measure vibrations of the top ring 221 during polishing. In addition, a noise meter 221k may be provided near the top ring 221 to measure noise during polishing.

[0036] The polishing fluid supply nozzle 222 is supported by a support shaft 222a and includes a swinging movement mechanism 222b that pivots the polishing fluid supply nozzle 222 around the support shaft 222a as the pivot point, a polishing fluid flow rate adjustment unit 222c that adjusts the flow rate of the polishing fluid, and a polishing fluid temperature control mechanism 222d that adjusts the temperature of the polishing fluid. The polishing fluid is a polishing liquid (slurry) or pure water, and may also contain a chemical solution, or may be a polishing liquid with a dispersant added.

[0037] The dresser 223 is supported by a dresser shaft 223a that is movable in the vertical direction and includes a dresser rotational movement mechanism 223c that rotates the dresser 223 around its axis, a dresser vertical movement mechanism 223d that moves the dresser 223 in the vertical direction, and a dresser oscillating movement mechanism 223e that pivots the dresser 223 around the dresser oscillating support shaft 223b as the pivot point.

[0038] The dresser swinging movement mechanism 223e includes a dresser swinging support shaft 223b, a dresser swinging arm 223f that swingably connects the dresser shaft 223a to the dresser swinging support shaft 223b, and a dresser swinging shaft motor 223g that rotationally drives the dresser swinging support shaft 223b.

[0039] The dresser oscillating movement mechanism 223e has a dresser oscillating torque sensor 223h at the connection point between the dresser oscillating arm 223f and the dresser oscillating shaft motor 223g, which detects the dresser oscillating torque applied to the dresser oscillating arm 223f. Specifically, the dresser oscillating torque sensor 223h should detect the torque applied to the dresser oscillating arm 223f from the current value of the dresser oscillating shaft motor 223g. The current value of the dresser oscillating shaft motor 223g is a quantity that depends on the torque of the dresser oscillating arm 223f at the connection point to the dresser oscillating shaft motor 223g. In this embodiment, the current value of the dresser oscillating shaft motor 223g may be the current value supplied to the dresser oscillating shaft motor 223g, or a current command value generated in a driver (not shown). Note that the dresser oscillating torque may be detected by other methods.

[0040] The atomizer 224 is supported by a support shaft 224a and includes an atomizer oscillating movement mechanism 224b that rotates the atomizer 224 around the support shaft 224a as the pivot point, and a cleaning fluid flow rate adjustment unit 224c that adjusts the flow rate of the cleaning fluid. The cleaning fluid is a mixed fluid of liquid (e.g., pure water) and gas (e.g., nitrogen gas) or liquid (e.g., pure water).

[0041] The environmental sensor 225 consists of sensors arranged in the internal space of the housing 20, and includes, for example, a temperature sensor 225a for measuring the temperature of the internal space, a humidity sensor 225b for measuring the humidity of the internal space, and a pressure sensor 225c for measuring the atmospheric pressure of the internal space. The environmental sensor 225 may also include a camera (image sensor) capable of photographing the surface of the polishing pad 2200 during or before / after the polishing process.

[0042] Note that Figure 3 omits the specific configurations of each rotational movement mechanism 220b, 221c, 223c, each vertical movement mechanism 221d, 223d, and each oscillating movement mechanism 221e, 222b, 223e, 224b. However, they are configured by appropriately combining, for example, a module for generating driving force such as a motor or actuator, a driving force transmission mechanism such as a linear guide, ball screw, gear, belt, coupling, or bearing, and sensors such as a linear sensor, encoder sensor, limit sensor, or torque sensor.

[0043] Furthermore, although the specific configurations of each flow rate control unit 222c and 224c are omitted in Figure 3, they are configured by appropriately combining, for example, fluid control modules such as pumps, valves, and regulators with sensors such as flow rate sensors, pressure sensors, and liquid level sensors. In addition, although the specific configurations of the polishing table surface temperature control mechanism 220c, the polishing table internal temperature control mechanism 220d, and the polishing fluid temperature control mechanism 222d are omitted in Figure 3, they are configured by appropriately combining, for example, temperature control modules (conduction type, radiation type, convection type) such as heaters and heat exchangers with sensors such as temperature sensors and current sensors.

[0044] Figure 4 is a schematic cross-sectional view showing an example of a top ring 221. The top ring 221 comprises a top ring body 2210 attached to a top ring shaft 221a, a substantially disc-shaped carrier 2211 housed in the top ring body 2210, an elastic film 2212 positioned below the carrier 2211 to press the wafer W against the polishing pad 2200, a substantially annular retainer ring 2213 positioned on the outer circumference of the carrier 2211 and the elastic film 2212 to directly press the polishing pad 2200, and a retainer ring airbag 2214 positioned between the top ring body 2210 and the retainer ring 2213 as a retainer ring pressing mechanism to press the retainer ring 2213 against the polishing pad 2200.

[0045] In this embodiment, a retainer ring airbag 2214 was used as the retainer ring pressing mechanism, but the retainer ring pressing mechanism may be a fluid actuator using air, water, or oil, an electric actuator using a ball screw, or an elastic member including a spring or a bag-shaped bag.

[0046] The elastic membrane 2212 is formed of an elastic membrane and has a plurality of concentric partition walls 2212e inside, thereby having first to fourth elastic membrane pressure chambers 2212a to 2212d arranged concentrically from the center outward toward the outer circumference of the top ring body 2210. The elastic membrane 2212 also has a plurality of holes 2212f for adsorption of the wafer W on its lower surface and functions as a substrate holding surface for holding the wafer W. The retainer ring airbag 2214 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 it may have a pressure chamber that presses the entire carrier 2211, the number and shape of the elastic membrane pressure chambers in the elastic membrane 2212 may be changed as appropriate, and the number and arrangement of the adsorption holes 2212f may be changed as appropriate. The elastic membrane 2212 may also not have adsorption holes 2212f.

[0047] The first to fourth elastic membrane pressure chambers 2212a to 2212d are connected to the first to fourth flow paths 2216A to 2216D, respectively, and the retainer ring pressure chamber 2214a is connected to the fifth flow path 2216E. The first to fifth flow paths 2216A to 2216E communicate with the outside via a top ring rotary connector 2215 provided on the top ring shaft 221a, and branch into the first branch flow paths 2217A to 2217E and the second branch flow paths 2218A to 2218E, respectively. Pressure sensors PA to PE are installed in the first to fifth flow paths 2216A to 2216E, respectively. The first branch flow paths 2217A to 2217E are connected to a gas supply source GS for pressurized fluid (air, nitrogen, etc.) via valves V1A to V1E, flow sensors FA to FE, and pressure regulators RA to RE. The second branch channels 2218A to 2218E are connected to the vacuum source VS via valves V2A to V2E, respectively, and are configured to communicate with the atmosphere via valves V3A to V3E.

[0048] The wafer W is held by suction on the lower surface of the top ring 221 and moved to a predetermined polishing position on the polishing table 220. After that, it is polished by being pressed against the polishing surface of the polishing pad 2200, which is supplied with polishing fluid from the polishing fluid supply nozzle 222, by the top ring 221. At this time, the top ring 221 controls the pressure regulators RA~RE independently. The pressurizing fluid supplied to the first to fourth elastic membrane pressure chambers 2212a to 2212d adjusts the pressing force that presses the wafer W against the polishing pad 2200 for each region of the wafer W, and the pressurizing fluid supplied to the retainer ring pressure chamber 2214a adjusts the pressing force that presses the retainer ring 2213 against the polishing pad 2200. The pressure of the pressurizing fluid supplied to the first to fourth elastic membrane pressure chambers 2212a to 2212d and the retainer ring pressure chamber 2214a is measured by pressure sensors PA to PE, respectively, and the flow rate of the pressurized gas is measured by flow sensors FA to FE, respectively.

[0049] (Circuit board transport unit) As shown in Figure 2, the substrate transport unit 23 includes first and second linear transporters 230A and 230B that are horizontally movable along the direction of alignment of the first to fourth polishing sections 22A to 22D (the longitudinal direction of the housing 20), a swing transporter 231 positioned between the first and second linear transporters 230A and 230B, a lifter 232 positioned on the load / unload unit 21 side, and a temporary wafer W stand 233 positioned on the cleaning unit 24 side.

[0050] The first linear transporter 230A is positioned adjacent to the first and second polishing sections 22A and 22B, and is a mechanism for transporting wafers W between four transport positions (denoted as the first to fourth transport positions TP1 to TP4, in order from the load / unload unit 21 side). The second transport position TP2 is the position where wafers W are handed over to the first polishing section 22A, and the third transport position TP3 is the position where wafers W are handed over to the second polishing section 22B.

[0051] The second linear transporter 230B is positioned adjacent to the third and fourth polishing sections 22C and 22D and is a mechanism for transporting the wafer W between three transport positions (referred to as the fifth to seventh transport positions TP5 to TP7, in order from the load / unload unit 21 side). The sixth transport position TP6 is the position where the wafer W is handed over to the third polishing section 22C, and the seventh transport position TP7 is the position where the wafer W is handed over to the fourth polishing section 22D.

[0052] The swing transporter 231 is positioned adjacent to the fourth and fifth transport positions TP4 and TP5 and has a hand that can move between the fourth and fifth transport positions TP4 and TP5. The swing transporter 231 is a mechanism for transferring wafers W between the first and second linear transporters 230A and 230B and for temporarily placing wafers W on the temporary storage table 233. The lifter 232 is positioned adjacent to the first transport position TP1 and is a mechanism for transferring wafers W between it and the transport robot 211 of the load / unload unit 21. When transferring wafers W, a shutter (not shown) provided in the first partition wall 200A is opened and closed.

[0053] (Washing unit) As shown in Figure 2, the cleaning unit 24 includes first and second cleaning chambers 240A and 240B for cleaning wafers W using cleaning tools, a drying chamber 241 for drying wafers W, and first and second transport chambers 242A and 242B for transporting wafers W. Each chamber of the cleaning unit 24 is partitioned and arranged along the first and second linear transporters 230A and 230B in the order of, for example, the first cleaning chamber 240A, the first transport chamber 242A, the second cleaning chamber 240B, the second transport chamber 242B, and the drying chamber 241 (in order of furthest from the load / unload unit 21). Note that the number and arrangement of the cleaning chambers 240A and 240B, the drying chamber 241, and the transport chambers 242A and 242B are not limited to the example in Figure 2 and may be changed as appropriate.

[0054] (Film thickness measurement unit) The film thickness measurement unit 25 is a measuring instrument that measures the film thickness of the wafer W before or after polishing, and is composed of, for example, an optical film thickness measuring instrument, an eddy current film thickness measuring instrument, etc. The transfer of wafers W to the module is performed by the transport robot 211.

[0055] (Control unit) Figure 5 is a block diagram showing an example of a substrate processing apparatus 2. The control unit 26 is electrically connected to each unit 21-25 and comprehensively controls each unit 21-25. In the following explanation, the control system (module, sensor, sequencer) of the polishing unit 22 will be used as an example, but the other units 21, 23-25 ​​have the same basic configuration and functions, so their explanation will be omitted.

[0056] The polishing unit 22 is positioned in each of its subunits (e.g., polishing table 220, top ring 221, polishing fluid supply nozzle 222, dresser 223, atomizer 224, etc.) and controls multiple modules 2271 to 227 r And multiple modules 2271~227 rThey are arranged in each of the following modules 2271-227 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.

[0057] Polishing unit 22 sensors 2281~228 s These include, for example, a sensor for detecting the flow rate of the polishing fluid, a sensor for detecting the pressing force of the retainer ring pressing mechanism, a sensor for detecting the rotational torque of the top ring 221, a sensor for detecting the rotational torque of the polishing table 220, a timer for measuring the time until the end point is detected, an optical or other sensor for detecting the end point, and an environmental sensor 225. In this embodiment, rotational torque refers to the sliding resistance between the top ring 221 and dresser 223 that are in contact with the polishing surface and the polishing table 220.

[0058] The control unit 26 comprises a substrate processing control unit 260, a communication unit 261, an input unit 262, an output unit 263, and a storage unit 264. The control unit 26 is composed of, for example, a general-purpose or dedicated computer (see Figure 6, described later).

[0059] The communication unit 261 is connected to the network 7 and functions as a communication interface for sending and receiving various types of data. The input unit 262 accepts various input operations, and the output unit 263 functions as a user interface by outputting various types of information via a display screen, signal tower illumination, and buzzer sound.

[0060] The memory unit 264 stores various programs (operating system (OS), application programs, web browser, etc.) and data (device setting information 265, board recipe information 266, etc.) used in the operation of the board processing device 2. The device setting information 265 and board recipe information 266 are data that can be edited by the user via the display screen.

[0061] The substrate processing control unit 260 obtains the detection values of a plurality of sensors 2181 to 218 q , 2281 to 228 s , 2381 to 238 u , 2481 to 248 w , 2581 to 258 y (hereinafter referred to as the "sensor group") through 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 p , 2271 to 227 r , 2371 to 237 t , 2471 to 247 v , 2571 to 257 x (hereinafter referred to as the "module group") to perform a series of substrate processes such as loading, polishing, cleaning, drying, film thickness measurement, and unloading.

[0062] (Hardware Configuration of Each Device) FIG. 6 is a hardware configuration diagram showing an example of the computer 900. Each of the control unit 26, the database device 3, the machine learning device 4, the information processing device 5, and the user terminal device 6 of the substrate processing apparatus 2 is constituted by a general-purpose or dedicated computer 900.

[0063] As shown in FIG. 6, the computer 900 mainly includes a bus 910, a processor 912, a memory 914, an input device 916, an output device 917, a display device 918, a storage device 920, a communication I / F (interface) unit 922, an external device I / F unit 924, an I / O (input / output) device I / F unit 926, and a media input / output unit 928. Note that the above components may be appropriately omitted according to the application for which the computer 900 is used.

[0064] The processor 912 consists of one or more arithmetic processing units (CPU (Central Processing Unit), MPU (Micro-processing unit), DSP (digital signal processor), GPU (Graphics Processing Unit), NPU (Neural Processing Unit), etc.) and controls the entire computer 900. The memory 914 stores various data and programs 930 and consists of volatile memory (DRAM, SRAM, etc.) that functions as main memory, and non-volatile memory (ROM), flash memory, etc.

[0065] The input device 916 consists of, for example, a keyboard, mouse, numeric keypad, or electronic pen, and functions as an input unit. The output device 917 consists of, for example, a sound (voice) output device or a vibration device, and functions as an output unit. The display device 918 consists of, for example, a liquid crystal display, an organic EL display, electronic paper, or a projector, and functions as an output unit. The input device 916 and the display device 918 may be configured as an integrated unit, such as a touch panel display. The storage device 920 consists of, for example, an HDD or SSD (Solid State Drive), and functions as a storage unit. The storage device 920 stores various data necessary for the execution of the operating system and program 930.

[0066] The communication I / F unit 922 is connected by wire or wireless to a network 940 such as the Internet or an intranet (which may be the same as network 7 in Figure 1) and functions as a communication unit that sends and receives data with other computers according to a predetermined communication standard. The external device I / F unit 924 is connected by wire or wireless to external devices 950 such as cameras, printers, scanners, and reader / writers and functions as a communication unit that sends and receives data with external devices 950 according to a predetermined communication standard. The I / O device I / F unit 926 is connected to I / O devices 960 such as various sensors and actuators and functions as a communication unit that sends and receives various signals and data with the I / O devices 960, such as detection signals from sensors and control signals to actuators. The media input / output unit 928 is composed of a drive device such as a DVD drive or CD drive and reads and writes data to media (non-temporary storage medium) 970 such as DVDs and CDs.

[0067] In the computer 900 having the above configuration, the processor 912 calls and executes the program 930 stored in the storage device 920 in the memory 914, and controls various parts of the computer 900 via the bus 910. The program 930 may also be stored in the memory 914 instead of the storage device 920. The program 930 may be recorded on the media 970 in an installable or executable file format and provided to the computer 900 via the media input / output unit 928. The program 930 may also be provided to the computer 900 by downloading it via the network 940 through the communication interface unit 922. Furthermore, the computer 900 may implement the various functions realized by the processor 912 executing the program 930 using hardware such as an FPGA or ASIC.

[0068] Computer 900 is an electronic device of any form, consisting of, for example, a stationary computer or a portable computer. Computer 900 may be a client computer, a server computer, or a cloud computer. Computer 900 may also be applied to devices other than those specified in devices 2 to 6.

[0069] (Production history information 30) Figure 7 is a data configuration diagram showing an example of production history information 30 managed by the database device 3. The production history information 30 includes, for example, a wafer history table 300 for each wafer W and a polishing history table 301 for equipment status information in the polishing process, as tables in which reports R acquired when substrate processing is performed on wafers W for production are classified and registered. In addition to the above, the production history information 30 also includes a cleaning history table for equipment status information in the cleaning process, an event history table for event information, and an operation history table for operation information, but a detailed explanation is omitted.

[0070] Each record in the wafer history table 300 registers, for example, the wafer ID, cassette number, slot number, start time, end time, and unit ID used for each process. Although Figure 7 shows polishing and cleaning processes as examples, other processes are registered in the same manner.

[0071] Each record in the polishing history table 301 contains information such as wafer ID, wear status, and processing status.

[0072] The wear status information is information indicating the wear status of each component of the substrate processing apparatus 2, which can be obtained before polishing. The wear status information includes, for example, information indicating the condition of the polishing pad 2200, information indicating the condition of the top ring rotary connector 2215, information indicating the condition of the polishing table rotary connector 2201, information indicating the condition of the dresser 223, and information indicating the condition of the substrate processing control unit 260. The information indicating each condition includes at least the usage time and the number of wafers processed for each component.

[0073] The processing status information is information indicating the processing status of the substrate processing apparatus 2 that can be acquired during the polishing process. The processing status information is, for example, the detection values ​​of each sensor sampled at predetermined time intervals by a group of sensors such as a flow rate sensor for the polishing fluid, a pressing force sensor for the retainer ring pressing mechanism, a rotational torque sensor for the top ring 221, or a rotational torque sensor for the polishing table 220, which are all part of the substrate processing apparatus 2. Furthermore, the processing status information is, for example, statistical values ​​of the time until the end point is detected for each wafer W, and statistical values ​​of the sensor time series data for each wafer W.

[0074] By referring to the polishing history table 301, time-series data from each sensor can be extracted as the device status of the substrate processing apparatus 2 when polishing was performed on wafer W, which is identified by the wafer ID.

[0075] (Polishing test information 31) Figure 8 is a data configuration diagram showing an example of polishing test information 31 managed by the database device 3. The polishing test information 31 includes a polishing test table 310 in which reports R obtained when a polishing test is performed and test results are classified and registered.

[0076] Each record in the polishing test table 310 contains, for example, a test ID, wear status information, processing status information, and test result information. The wear status information and processing status information in the polishing test table 310 indicate the state of each part during the polishing test, and their data structure is the same as that of the polishing history table 301, so a detailed explanation is omitted.

[0077] The test result information indicates the state of the test polishing apparatus when the polishing process is performed in the polishing test. The test result information may be measured values ​​obtained by the polishing apparatus measuring instrument installed on the test polishing apparatus. The test result information shown in Figure 8 includes the number of wafers processed (1, 2, ..., ...m, ..., n) and the usage times (t1, t2, ..., ...tm, ..., tn) since the start of the polishing process. It also includes the reliability of endpoint detection, signs of reliability degradation, the type of component causing reliability degradation, and the types of processes TR1 to TR4 causing reliability degradation, respectively, at each time point tm1, tm2, ..., ...tmm, ..., tmn included in the polishing process period for the mth wafer.

[0078] By referring to the polishing test table 310, it is possible to extract time-series data from each sensor indicating the state of the polishing unit 22 when the polishing process was performed, as well as the reliability of the endpoint detection function at that time, for the polishing test identified by the test ID. For example, to detect signs of a decrease in the reliability of the endpoint detection function, data from the first target period of the polishing process prior to the point tmm at which the reliability of the endpoint detection function decreased can be used, and to determine whether or not the reliability of the endpoint detection function has decreased, data from the second target period including the point tmm at which the reliability of the endpoint detection function decreased can be used.

[0079] (Machine learning device 4) Figure 9 is a block diagram showing an example of a machine learning apparatus 4 according to the first embodiment. The machine learning apparatus 4 comprises a machine learning control unit 40, a communication unit 41, a training data storage unit 42, and a trained model storage unit 43.

[0080] The machine learning control unit 40 functions as a training data acquisition unit 400 and a machine learning unit 401. The communication unit 41 is connected to external devices (e.g., a substrate processing device 2, a database device 3, an information processing device 5, and a user terminal device 6, a polishing test device (not shown), etc.) via the network 7 and functions as a communication interface for sending and receiving various types of data.

[0081] The learning data acquisition unit 400 is connected to an external device via the communication unit 41 and the network 7, and acquires first learning data 11A, which consists of reliability degradation factor state information as input data and reliability information of the polishing endpoint detection function as output data. The first learning data 11A is used as training data, validation data, and test data in supervised learning. The reliability information of the polishing endpoint detection function is used as the correct label in supervised learning.

[0082] The learning data storage unit 42 is a database that stores multiple sets of the first learning data 11A acquired by the learning data acquisition unit 400. The specific configuration of the database constituting the learning data storage unit 42 can be designed as appropriate.

[0083] The machine learning unit 401 performs machine learning using multiple sets of first training data 11A stored in the training data storage unit 42. Specifically, the machine learning unit 401 inputs multiple sets of the first training data 11A into the first learning model 10A and generates a trained first learning model 10A by having the first learning model 10A learn the correlation between the reliability degradation factor state information contained in the first training data 11A and the reliability information of the polishing endpoint detection function.

[0084] The trained model storage unit 43 is a database that stores the trained first trained model 10A (specifically, the adjusted weight parameter set) generated by the machine learning unit 401. The trained first trained model 10A stored in the trained model storage unit 43 is provided to the actual system (for example, the information processing device 5) via the network 7 or a recording medium. In Figure 9, the training data storage unit 42 and the trained model storage unit 43 are shown as separate storage units, but they may be composed of a single storage unit.

[0085] The number of first learning models 10A stored in the trained model storage unit 43 is not limited to one. For example, multiple learning models with different conditions may be stored, such as machine learning methods, differences in the mechanism and material of the top ring 221, types of elastic membranes 2212, types of retainer rings 2213, types of polishing pads 2200, types of polishing fluids, types of data included in the reliability degradation factor state information, and types of data included in the reliability information of the polishing endpoint detection function. In that case, the learning data storage unit 42 should store multiple types of learning data, each having a data configuration corresponding to the multiple learning models with different conditions.

[0086] 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 consists of reliability degradation factor state information and reliability information of the polishing endpoint detection function.

[0087] The reliability degradation factor state information constituting the first learning data 11A includes wear status information indicating the wear status of the components of the substrate processing apparatus 2, and polishing status information indicating the processing status of the substrate processing apparatus 2 that can be obtained during polishing.

[0088] The wear status information included in the reliability degradation factor status information is information indicating the wear status of the components of the substrate processing apparatus 2. The wear status information includes, for example, the condition of the polishing pad 2200, the condition of the top ring rotary connector 2215, the condition of the polishing table rotary connector 2201, the condition of the dresser 223, and the condition of the substrate processing control unit 260.

[0089] The condition of the polishing pad 2200 includes at least the usage time of the polishing pad 2200 and the number of wafers processed by the polishing pad 2200. The condition of the polishing pad 2200 may be set based on, for example, the cumulative number of rotations of the polishing table 220, the rotation speed of the polishing table 220, the rotation torque of the polishing table 220, the presence or absence of dressing, whether or not it has been replaced, an image of the surface taken, the surface shape, flatness, cleanliness, wetness, etc. The condition of the polishing pad 2200 may also change over time during the polishing process, for example.

[0090] The condition of the top ring rotary connector 2215 includes at least the usage time of the top ring 221 and the number of wafers processed by the top ring 221. The condition of the top ring rotary connector 2215 may also be set based on, for example, the cumulative number of rotations of the top ring 221, the rotational speed of the top ring 221, the rotational torque of the top ring 221, etc. The condition of the top ring 221 may also change over time, for example, during the polishing process.

[0091] The condition of the polishing table rotary connector 2201 includes at least the usage time of the polishing table 220 and the number of wafers processed by the polishing table 220. The condition of the polishing table rotary connector 2201 may also be set based on, for example, the cumulative number of rotations of the polishing table 220, the rotational speed of the polishing table 220, the rotational torque of the polishing table 220, etc. The condition of the polishing table 220 may also change over time during the polishing process.

[0092] The condition of the dresser 223 includes at least the operating time of the dresser 223 and the number of wafers processed by the polishing table 220. The condition of the dresser 223 may also be set based on, for example, the cumulative number of rotations of the dresser 223, the rotation speed of the dresser 223, the rotation torque of the dresser 223, whether or not dressing is performed, whether or not it is replaced, images of the surface taken, surface shape, flatness, cleanliness, wetness, etc. The condition of the dresser 223 may also change over time during the polishing process, for example.

[0093] The condition of the substrate processing control unit 260 includes at least the operating time of the substrate processing control unit 260 and the number of wafers processed by the substrate processing control unit 260. The condition of the substrate processing control unit 260 may change over time, for example, during the polishing process.

[0094] The processing state information included in the reliability degradation factor state information is information indicating the processing state of the substrate processing apparatus 2 that can be obtained during polishing. The processing state information includes, for example, at least one of the following: the flow rate of the polishing fluid, the pressing force of the retainer ring pressing mechanism, the rotational torque of the top ring 221, the rotational torque of the polishing table 220, the oscillation torque of the top ring 221, the vibration of the top ring 221, the oscillation torque of the dresser 223, the noise during polishing, the temperature of the polishing surface, the temperature of the temperature-controlled water for the polishing table 220, statistical values ​​of the time until endpoint detection for each wafer, and statistical values ​​of sensor time-series data for each wafer.

[0095] The flow rate of the polishing fluid may be the flow rate of the polishing fluid supplied from the polishing fluid supply nozzle 222. In this embodiment, the pressing force of the retainer ring pressing mechanism may be the pressure in the retainer ring pressure chamber 2214a of the retainer ring airbag 2214, the flow rate of the pressurized fluid supplied to the retainer ring pressure chamber 2214a, etc. The rotational torque of the top ring 221 may be determined from the motor current driving the top ring 221, etc. The rotational torque of the polishing table 220 may be determined from the motor current driving the polishing table 220, etc.

[0096] The oscillation torque of the top ring 221 may be the top ring oscillation torque applied to the top ring oscillation arm 221f as detected by the top ring oscillation torque sensor 221h. The vibration of the top ring 221 may be the vibration of the top ring 221 during polishing as measured by the acceleration sensor 221j attached to the top ring 221. The oscillation torque of the dresser 223 may be the dresser oscillation torque applied to the dresser oscillation arm 223f as detected by the dresser oscillation torque sensor 223h. The noise during polishing may be the noise during polishing as measured by the noise meter 221k installed near the top ring 221. The temperature of the polishing surface may be the surface temperature of the polishing pad 2200 or the surface temperature of the grinding wheel as measured by the radiation thermometer 220c1 installed above the polishing table 220. The temperature of the temperature-controlled water for the polishing table 220 may be the temperature of the supplied temperature-controlled water measured by the supply temperature-controlled water thermometer 220d3.

[0097] The statistical value of the time to endpoint detection for each wafer can be obtained from the measured time to endpoint detection for each wafer. Alternatively, the measured time to endpoint detection for each wafer can be divided into predetermined ranges, and the statistical value can be obtained from the number of data points in each range. The statistical value of the sensor time-series data for each wafer can be obtained from the time-series data measured by sensors such as flow rate, pressing force, and rotational torque of the components. Alternatively, the time to endpoint detection of the measured time-series data can be divided into predetermined ranges, and the number of data points in each range can be determined.

[0098] The statistical values ​​obtained in this way can be used as is, processed to reduce noise and make them easier for measuring instruments to pick up, or used after statistical processing. In any case, it is sufficient to determine the variability of the results.

[0099] In this embodiment, since a retainer ring airbag 2214 is used as the retainer ring pressing mechanism, the top ring state information included in the polishing process state information includes the pressure in the retainer ring pressure chamber 2214a (retainer ring airbag pressure) and the flow rate of the pressurized fluid supplied to the retainer ring pressure chamber 2214a (retainer ring airbag flow rate).

[0100] When using another pressing mechanism as the retainer ring pressing mechanism, the pressing force of the retainer ring pressing mechanism in the processing state information included in the reliability degradation factor state information may be the amount of the element that adjusts the pressing force of the retainer ring pressing mechanism. For example, an electric actuator can be used as the retainer ring pressing mechanism. When a diode is used, the pressing force of the electric actuator may be the amount of current used to adjust the pressing force of the electric actuator. Also, when an elastic member such as a spring or bag is used as the retainer ring pressing mechanism, the pressing force of the retainer ring pressing mechanism may be the pressing force of the elastic member and the vertical position that adjusts the pressing force of the elastic member.

[0101] The reliability information of the polishing endpoint detection function that constitutes the first learning data 11A is information indicating the reliability of the polishing endpoint detection function of a wafer W that has been polished in a state as shown in the reliability degradation factor state information. In this embodiment, the reliability information of the polishing endpoint detection function is information on the current reliability of endpoint detection, information on signs of reliability degradation, information on the types of components that cause reliability degradation, and information on the types of processes that cause reliability degradation.

[0102] The learning data acquisition unit 400 acquires the first learning data 11A by referring to the polishing test information 31 and, if necessary, accepting user input operations from the user terminal device 6. For example, by referring to the polishing test table 310 of the polishing test information 31, the learning data acquisition unit 400 acquires wear status information indicating the wear status of the components of the substrate processing apparatus 2 when a polishing test identified by the test ID is performed, and processing status information indicating the processing status of the substrate processing apparatus 2 that can be obtained during the polishing process, as reliability degradation factor status information.

[0103] In this embodiment, the case in which reliability degradation factor state information is acquired as time-series data of the sensor group is described, but this may be appropriately changed depending on the configuration of the polishing unit 22 (especially the top ring 221 and the polishing table 220). Furthermore, the reliability degradation factor state information may use command values ​​to the module, parameters converted from sensor detection values ​​or command values ​​to the module, or parameters calculated based on detection values ​​of multiple sensors. In addition, the reliability degradation factor state information may be acquired as time-series data for the entire polishing process period, as time-series data for a target period which is a part of the polishing process period, or as time-series data at a specific target point in time, such as the point in time when the reliability of the polishing endpoint detection function deteriorates. When changing the definition of reliability degradation factor state information as described above, the data configuration of the input data in the first learning model 10A and the first learning data 11A should be appropriately changed.

[0104] Furthermore, the learning data acquisition unit 400 refers to the polishing test table 310 of the polishing test information 31 and acquires test result information when a polishing test identified by the same test ID is performed, as reliability information for the polishing endpoint detection function corresponding to the reliability degradation factor state information mentioned above.

[0105] In this embodiment, we will explain the case where the reliability information of the polishing endpoint detection function includes the current endpoint detection reliability information, the indicator of reliability degradation, the type of component causing the reliability degradation, and the type of process causing the reliability degradation, as shown in Figure 10.

[0106] The current endpoint detection confidence level information indicates the current confidence level of endpoint detection as a percentage. Confidence levels can be classified from 0% to 100%. For example, if the current endpoint detection confidence level is 100%, it can be determined that the current endpoint detection is reliable; if the current endpoint detection confidence level is 0%, it can be determined that the current endpoint detection is unreliable.

[0107] The information indicating a decrease in reliability is information about the time or number of processed boards until the reliability of endpoint detection decreases or endpoint detection fails. For example, if there is a high probability that the reliability of endpoint detection will decrease in the next few hours or after a few boards of polishing, some action can be taken on the substrate processing device 2 before that happens.

[0108] Information on the types of components that cause a decrease in reliability is obtained from the pre-prepared components, including the final check. This information pertains to the types of components that reduce the reliability of the knowledge. For example, if a top ring is a factor that reduces the reliability of endpoint detection, then some action can be taken to address the top ring.

[0109] The information on the types of processes that cause a decrease in reliability refers to the types of processes among the pre-defined processes that cause a decrease in the reliability of endpoint detection. For example, if the factor causing a decrease in the reliability of endpoint detection is the film deposition process, then some measures can be taken to address variations in the film strength, film thickness, etc., of the film deposition process.

[0110] The first learning model 10A employs, for example, a neural network structure and comprises an input layer 100, a hidden layer 101, and an output layer 102. Synapses (not shown) connect each neuron between each layer, and each synapse is associated with a weight. The weight parameter set, consisting of the weights of each synapse, is adjusted by machine learning.

[0111] The input layer 100 has a number of neurons corresponding to the reliability degradation factor state information as input data, and each value of the reliability degradation factor state information is input to each neuron. The output layer 102 has a number of neurons corresponding to the reliability information of the polishing endpoint detection function as output data, and the prediction result (inference result) of the reliability information of the polishing endpoint detection function for the reliability degradation factor state information is output as output data.

[0112] If the first learning model 10A is composed of a regression model, the reliability information of the polishing endpoint detection function is output as a normalized numerical value within a predetermined range (e.g., 0 to 1). If the first learning model 10A is composed of a classification model, the reliability information of the polishing endpoint detection function is output as a score (accuracy) for each class, normalized numerical value within a predetermined range (e.g., 0 to 1).

[0113] The "predetermined range (0 to 1)" is pre-set with inference results corresponding to numerical values. For example, in the case of current endpoint detection confidence information, the "predetermined range (0 to 1)" which is the inference result can be divided into multiple ranges, and the current endpoint detection confidence level (0 to 100%) can be set for each divided range. Similarly, in the case of predictive information on decreased reliability, the "predetermined range (0 to 1)" which is the inference result can be divided into multiple ranges, and the predictive time until decreased reliability can be set for each divided range.

[0114] In the case of information about the types of components that contribute to reduced reliability, a predetermined threshold should be set within a predetermined range (0 to 1) of the inference results for each component. If the output value is below the threshold, the component is considered "not a factor contributing to reduced reliability," and if it exceeds the threshold, it is considered "a factor contributing to reduced reliability." Each component should be prepared in advance, and a threshold should be set for each of them.

[0115] Furthermore, in the case of process type information that is a factor in reducing reliability, a predetermined threshold should be set within the "predetermined range (0 to 1)" of the inference results for each process. If the output value is below the threshold, the process is considered "not a factor in reducing reliability," and if it exceeds the threshold, it is considered "a factor in reducing reliability." Each process should be prepared in advance, and a threshold should be set for each of them.

[0116] (Machine learning methods) Figure 11 is a flowchart showing an example of a machine learning method using machine learning device 4.

[0117] First, in step S100, the training data acquisition unit 400 starts machine learning. As a preliminary step, a desired number of first training data 11A are obtained from polishing test information 31, etc., and these obtained first training data 11A are stored in the training data storage unit 42. The number of first training data 11A prepared here should be set considering the inference accuracy required for the final first learning model 10A.

[0118] Next, in step S110, the machine learning unit 401 prepares a first pre-training model 10A in order to start machine learning. The first pre-training model 10A prepared here consists of the neural network model illustrated in Figure 10, and the weights of each synapse are set to initial values.

[0119] Next, in step S120, the machine learning unit 401 randomly selects, for example, one set of first training data 11A from multiple sets of first training data 11A stored in the training data storage unit 42.

[0120] Next, in step S130, the machine learning unit 401 inputs the polishing process state information (input data) contained in a set of first training data 11A to the input layer 100 of the prepared pre-training (or training) first learning model 10A. As a result, the output layer 102 of the first learning model 10A outputs reliability information (output data) of the polishing endpoint detection function as an inference result, but this output data is generated by the pre-training (or training) first learning model 10A. Therefore, in the pre-training (or training) state, the output data output as an inference result shows information different from the reliability information (ground truth label) of the polishing endpoint detection function contained in the first training data 11A.

[0121] Next, in step S140, the machine learning unit 401 compares the reliability information (ground truth labels) of the polishing endpoint detection function included in the set of first training data 11A acquired in step S120 with the reliability information (output data) of the polishing endpoint detection function output from the output layer as an inference result in step S130, and performs machine learning by adjusting the weight of each synapse (backpropagation). In this way, the machine learning unit 401 trains the first learning model 10A to understand the correlation between the reliability degradation factor state information and the reliability information of the polishing endpoint detection function.

[0122] Next, in step S150, the machine learning unit 401 determines whether predetermined learning termination conditions have been met, for example, based on the evaluation value of the error function, which is based on the reliability information (correct label) of the polishing endpoint detection function included in the first training data 11A and the reliability information (output data) of the polishing endpoint detection function output as an inference result, or based on the remaining number of untrained first training data 11A stored in the training data storage unit 42.

[0123] In step S150, if the machine learning unit 401 determines that the learning termination condition has not been met and that machine learning should continue (No in step S150), it returns to step S120 and performs steps S120 to S140 multiple times on the first learning model 10A that is currently being learned, using the first training data 11A that has not yet been trained. On the other hand, in step S150, if the machine learning unit 401 determines that the learning termination condition has been met and that machine learning should be terminated (Yes in step S150), it proceeds to step S160.

[0124] Then, in step S160, the machine learning unit 401 stores the trained first learning model 10A (set weight parameters) generated by adjusting the weights associated with each synapse in the trained model storage unit 43, thus ending the series of machine learning methods shown in Figure 11. In the machine learning method, step S100 corresponds to the training data storage step, steps S110 to S150 correspond to the machine learning steps, and step S160 corresponds to the trained model storage step.

[0125] As described above, the machine learning apparatus 4 and machine learning method according to this embodiment provide a first learning model 10A that can predict (infer) reliability information of the polishing endpoint detection function indicating the state of the wafer W from reliability degradation factor state information including wear state information and processing state information.

[0126] (Information Processing Device 5) Figure 12 is a block diagram showing an example of an information processing device 5 according to the first embodiment. Figure 13 is a functional diagram showing an example of an information processing device 5 according to the first embodiment. The information processing device 5 comprises an information processing control unit 50, a communication unit 51, and a learned model storage unit 52.

[0127] The information processing 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 external devices (for example, a substrate processing device 2, a database 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 sending and receiving various types of data.

[0128] The information acquisition unit 500 is connected to an external device via the communication unit 51 and the network 7, and acquires reliability degradation factor status information, including wear status information and processing status information.

[0129] For example, when performing "real-time prediction processing" of reliability information for the polishing endpoint detection function for a wafer W that is undergoing polishing, the information acquisition unit 500 continuously receives a report R regarding reliability degradation factor status information from the substrate processing apparatus 2 that is performing the polishing process, and acquires the consumption status information and processing status information of the wafer W during the polishing process as reliability degradation factor status information.

[0130] As described above, the state prediction unit 501 inputs the reliability degradation factor state information acquired by the information acquisition unit 500 as input data to the first learning model 10A, thereby predicting the reliability information of the polishing endpoint detection function for the wafer W that is undergoing polishing, as indicated by the reliability degradation factor state information.

[0131] The trained model storage unit 52 is a database that stores the trained first trained model 10A used in the state prediction unit 501. The number of first trained models 10A stored in the trained model storage unit 52 is not limited to one. For example, multiple trained models with different states may be stored and selectively used, such as machine learning methods, differences in the mechanism and material of the top ring 221, types of elastic films 2212, types of retainer rings 2213, types of polishing pads 2200, types of polishing fluids, types of data included in reliability degradation factor state information, and types of data included in reliability information of the polishing endpoint detection function. Furthermore, the trained model storage unit 52 may be replaced by a storage unit on an external computer (for example, a server-type computer or a cloud-type computer), in which case the state prediction unit 501 only needs to access the external computer.

[0132] The output processing unit 502 performs output processing to output reliability information of the polishing endpoint detection function generated by the state prediction unit 501. For example, the output processing unit 502 may transmit the reliability information of the polishing endpoint detection function to the user terminal device 6 so that a display screen based on the reliability information of the polishing endpoint detection function is displayed on the user terminal device 6, or it may transmit the reliability information of the polishing endpoint detection function to the database device 3 so that the reliability information of the polishing endpoint detection function is registered in the production history information 30.

[0133] (Information processing methods) Figure 14 is a flowchart showing an example of an information processing method by the information processing device 5. Below, we will describe an example of operation when a user operates the user terminal device 6 to perform "predictive processing" of reliability information for the polishing endpoint detection function for a specific wafer W.

[0134] First, in step S200, when the user performs an input operation to the user terminal device 6 to input a wafer ID that identifies the wafer W to be predicted, the user terminal device 6 transmits that wafer ID to the information processing device 5.

[0135] Next, in step S210, the information acquisition unit 500 of the information processing device 5 receives the wafer ID transmitted in step S200. In step S211, the information acquisition unit 500 uses the wafer ID received in step S210 to refer to the polishing history table 301 of the production history information 30 and acquires reliability degradation factor status information when polishing is performed on the wafer W identified by that wafer ID.

[0136] Next, in step S220, the state prediction unit 501 inputs the reliability degradation factor state information acquired in step S211 as input data to the first learning model 10A, thereby generating reliability information of the polishing endpoint detection function for the said reliability degradation factor state information as output data, and predicts the state of the wafer W.

[0137] Next, in step S230, the output processing unit 502 transmits the reliability information of the polishing endpoint detection function generated in step S220 to the user terminal device 6 as an output process. In addition to the user terminal device 6, the database device 3 may also be used as the destination for the reliability information of the polishing endpoint detection function.

[0138] Next, in step S240, when the user terminal device 6 receives the reliability information of the polishing endpoint detection function transmitted in step S230 as a response to the transmission process in step S200, it displays a display screen based on the reliability information of the polishing endpoint detection function, allowing the user to visually confirm the state of the wafer W. In the above information processing method, steps S210 and S211 correspond to the information acquisition process, step S220 to the state prediction process, and step S230 to the output processing process.

[0139] As described above, according to the information processing device 5 and information processing method of this embodiment, when reliability degradation factor information in the polishing process is input to the first learning model 10A, reliability information of the polishing endpoint detection function with respect to the reliability degradation factor state information is predicted, so that reliability information of the polishing endpoint detection function, which indicates the reliability of the endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint, can be appropriately predicted.

[0140] (Second embodiment) The second embodiment differs from the first embodiment in that it uses an optical sensor as the polishing endpoint detection function. Below, the machine learning apparatus 4a and information processing apparatus 5a according to the second embodiment will be described, focusing on the differences from the first embodiment.

[0141] Figure 15 is a schematic diagram illustrating an optical sensor 226 provided on a polishing table 220 according to a second embodiment.

[0142] As shown in Figure 15, an optical sensor 226 for detecting the state of the film on the wafer W is embedded inside the polishing table 220. This optical sensor 226 irradiates the wafer W with light and detects the state of the film on the wafer W (such as film thickness) from the intensity of the reflected light (reflection intensity or reflectance) from the wafer W.

[0143] Furthermore, the polishing pad 2200 is designed to transmit light from the optical sensor 226. A light-transmitting section 2200a is attached. This light-transmitting section 2200a is made of a material with high transmittance, such as quartz glass, glass material, or pure water (with a transparent fluid supply section and flow path not shown). Alternatively, a through hole may be provided in the polishing pad 2200, and a transparent fluid such as pure water may be flowed from below through a transparent fluid supply section while this through hole is blocked by the wafer W, thereby forming the light-transmitting section 2200a. The light-transmitting section 2200a is positioned to pass through the center of the wafer W held by the top ring 221.

[0144] As shown in Figure 15, the optical sensor 226B comprises a light source 226a, a light-emitting optical fiber 226b as a light-emitting unit that irradiates the surface of the wafer W to be polished with light from the light source 226a, a light-receiving optical fiber 226c as a light-receiving unit that receives reflected light from the surface to be polished, a spectrometer unit 226d which has a spectrometer that spectrally analyzes the light received by the light-receiving optical fiber 226c and a plurality of light-receiving elements that store the spectrally analyzed light as electrical information, an operation control unit 226e that controls the timing of turning the light source 226a on and off and starting the reading of the light-receiving elements in the spectrometer unit 226d, and a power supply 226f that supplies power to the operation control unit 226e. Power is supplied to the light source 226a and the spectrometer unit 226d via the operation control unit 226e.

[0145] The light-emitting end of the light-emitting optical fiber 226b and the light-receiving end of the light-receiving optical fiber 226c are configured to be approximately perpendicular to the polished surface of the wafer W. For example, a 128-element photodiode array can be used as the photodetector in the spectrometer unit 226d. The spectrometer unit 226d is connected to the operation control unit 226e. Information from the photodetector in the spectrometer unit 226d is sent to the operation control unit 226e, and spectral data of the reflected light is generated based on this information. That is, the operation control unit 226e reads the electrical information stored in the photodetector and generates spectral data of the reflected light. This spectral data shows the intensity of the reflected light, decomposed according to wavelength, and changes with film thickness.

[0146] The motion control unit 226e is connected to the optical sensor control unit 226g. In this way, the spectral data generated by the motion control unit 226e is transmitted to the optical sensor control unit 226g. The optical sensor control unit 226g calculates characteristic values ​​associated with the film thickness of the wafer W based on the spectral data received from the motion control unit 226e, uses these as monitoring signals, and performs endpoint detection. The optical sensor control unit 226g may be included in the substrate processing control unit 260.

[0147] Figure 16 is a block diagram showing an example of a machine learning device 4a according to the second embodiment. Figure 17 is a diagram showing an example of a second learning model 10B and a second training data 11B. The second training data 11B is used for machine learning of the second learning model 10B.

[0148] The reliability degradation factor state information constituting the second training data 11B includes the condition of the optical sensor 226 and the condition of the transparent liquid supply unit as consumption state information, and the light reflectance intensity information of the optical sensor 226 and the transparent liquid flow rate information as processing state information. The other reliability degradation factor state information constituting the second training data 11B is the same as in the first embodiment, so its explanation is omitted.

[0149] The condition of the optical sensor 226 includes at least the usage time of the optical sensor 226 and the number of wafers processed by the optical sensor 226. The condition of the optical sensor 226 may be set based on, for example, the usage time of the light source 226a such as a lamp, the temperature of the optical sensor 226, etc. The condition of the optical sensor 226 may also change over time, for example, during the polishing process.

[0150] The condition of the pure water clear liquid supply unit is determined by at least the usage time of the pure water supply unit, and the research This includes the number of wafers processed by the polishing unit 22. The condition of the transparent liquid supply unit may be set based on, for example, the cumulative supply flow rate of the transparent liquid supply unit. The condition of the transparent liquid supply unit may also change over time during the polishing process.

[0151] The light reflection intensity information of the optical sensor 226 may be the intensity of the reflected light emitted by the optical sensor 226 and reflected by the wafer W. The transparent liquid information may be the flow rate of the transparent liquid, such as pure water, supplied from the transparent liquid supply unit.

[0152] The learning data acquisition unit 400 acquires second learning data 11B by referring to the polishing test information 31 and, if necessary, accepting user input operations from the user terminal device 6. For example, the learning data acquisition unit 400 refers to the polishing test table 310 of the polishing test information 31 and acquires wear status information and processing status information (time-series data of each sensor that each component has) when a polishing test identified by the test ID is performed, as reliability degradation factor status information.

[0153] Figure 18 is a block diagram showing an example of an information processing device 5a functioning according to the second embodiment. Figure 19 is a functional diagram showing an example of an information processing device 5a according to the second embodiment.

[0154] The information acquisition unit 500 acquires reliability degradation factor status information, including wear status information and processing status information, similar to the first embodiment.

[0155] When performing "post-processing" of reliability information for the polishing endpoint detection function on a wafer W after polishing has already been performed, the information acquisition unit 500 may refer to the polishing history table 301 of the production history information 30 to acquire the wear status information and processing status information at the time the polishing process was performed on the wafer W as reliability degradation factor status information.

[0156] As described above, the state prediction unit 501 inputs the reliability degradation factor state information acquired by the information acquisition unit 500 as input data to the second learning model 10B, thereby predicting the reliability information of the polishing endpoint detection function for the wafer W that is undergoing polishing, as indicated by the reliability degradation factor state information.

[0157] As described above, according to the information processing device 5a and information processing method of this embodiment, when reliability degradation factor state information, including wear status information and processing status information in the polishing process, is input to the second learning model 10B, the reliability of the polishing endpoint detection function with respect to the reliability degradation factor state information is predicted, so that the reliability of the polishing endpoint detection function of the wafer W due to the polishing process can be appropriately predicted.

[0158] In the second embodiment, an optical sensor 226 was used to detect the polishing endpoint of the wafer W, but other polishing endpoint detection functions may be used. For example, an eddy current sensor may be used as another example of a polishing endpoint detection function.

[0159] An eddy current sensor has an excitation coil, and when magnetic field lines generated from the excitation coil, which is connected to a high-frequency AC power supply, pass through a conductive film, eddy currents are generated on the wafer surface W. The magnitude of these eddy currents changes according to the resistance of the metal film, i.e., the thickness of the metal film. On the other hand, when eddy currents flow, magnetic field lines are generated from the eddy currents in the opposite direction to the magnetic field lines generated from the excitation coil. By measuring the intensity of these reverse-directed magnetic field lines with a detection coil, the change in the thickness of the metal film can be measured.

[0160] The eddy current sensor is installed below the polishing table 220 shown in Figure 3, and the polishing table Magnetic field lines are generated in a direction that penetrates 220. As the polishing table 220 rotates, the eddy current sensor rotates with the polishing table 220 and passes beneath the wafer W held by the top ring 221. At this time, if a conductive film exists on the surface of the wafer W, the magnetic field lines pass through the conductive film, allowing the thickness of the conductive film to be measured.

[0161] As the polishing of wafer W progresses, the metal film on the wafer W surface decreases, and consequently, the resistance of the metal film increases. This reduces the eddy currents generated by the magnetic field lines produced by the coil of the eddy current sensor, and therefore the strength of the magnetic field lines generated by the eddy currents also decreases. The eddy current sensor converts this change in magnetic field lines generated by the eddy currents into a voltage corresponding to the change in film thickness using its internal circuitry and software.

[0162] Thus, the eddy current sensor can detect the end of polishing by pre-measuring and storing the detection voltage on a wafer W after polishing is complete, and comparing the stored voltage value with the voltage value during polishing of the wafer W. The analog processing portion of the eddy current sensor may be replaced with digital processing. Using digital processing improves the performance and stability of the eddy current sensor.

[0163] When using an eddy current sensor, the condition of the eddy current sensor can be used as the wear status information, and the magnetic field line strength information of the eddy current sensor can be used as the processing status information.

[0164] The condition of the eddy current sensor includes at least the operating time of the eddy current sensor and the number of wafers processed by the eddy current sensor. The condition of the eddy current sensor may be set based on, for example, the operating time of the excitation coil. The condition of the eddy current sensor may also change over time, for example, during the polishing process.

[0165] The magnetic field line strength information from an eddy current sensor can be the magnetic field line strength generated in the opposite direction to that detected by the eddy current sensor.

[0166] Furthermore, a rotational torque method may be used as another example of a polishing endpoint detection function. The rotational torque method only requires detecting at least one of the following: the rotational torque of the polishing table rotation motor of the rotational movement mechanism 220b that rotates the polishing table 220, or the rotational torque of the top ring rotation motor of the top ring rotational movement mechanism 221c that rotates the top ring 221.

[0167] As the wafer W becomes flattened, the polishing resistance decreases sharply, making it possible to detect when the polishing of the wafer W is complete.

[0168] Furthermore, the detection of the wafer W polishing endpoint may be performed using an optical sensor, an eddy current sensor, or a torque method, or a combination of all of these.

[0169] (Other embodiments) The present invention is not limited to the embodiments described above, and can be implemented with various modifications without departing from the spirit of the invention. All such modifications are included in the technical concept of the present invention.

[0170] In the above embodiment, the database device 3, the machine learning device 4, and the information processing device 5 were described as being composed of separate devices, but these three devices may be composed of a single device, or any two of these three devices may be composed of a single device. Furthermore, at least one of the machine learning device 4 and the information processing device 5 may be incorporated into the control unit 26 or user terminal device 6 of the substrate processing device 2.

[0171] In the above embodiment, the substrate processing apparatus 2 was described as comprising each of the units 21 to 25, but the substrate processing apparatus 2 only needs to include at least the polishing unit 22, and the other units may be omitted.

[0172] In the above embodiment, a case in which a neural network is used as the learning model for realizing machine learning by the machine learning unit 401 was described, but other machine learning models may also be used. Other machine learning models include, for example, tree-type models such as decision trees and regression trees, ensemble learning such as bagging and boosting, recurrent neural networks, convolutional neural networks, and neural network types such as LSTM (including deep learning). Classes such as hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, and k-means clustering. Examples include Taring-type analysis, principal component analysis, factor analysis, multivariate analysis such as logistic regression, and support vector machines.

[0173] In the above embodiment, the test result information is information indicating the state when polishing is performed in a polishing test using a dummy wafer in the test apparatus. However, it may also be possible to continuously acquire information indicating the state when polishing is performed on an actual wafer using an actual polishing unit 22 equipped with sensors to detect the state of each component. The continuously acquired test result information is continuously learned by the machine learning device 4.

[0174] Furthermore, in the polishing unit 22 where no sensors are installed, the test result information may be continuously acquired by having a person judge the decrease in reliability of the polishing endpoint detection function and label the data.

[0175] Furthermore, information continuously acquired using the actual polishing unit 22 may be uploaded to the cloud, machine-learned in the cloud, and then the trained model may be deployed to the substrate processing apparatus 2. Alternatively, the processing method may be learned within the substrate processing apparatus 2 without uploading to the cloud.

[0176] (Machine learning programs and information processing programs) The present invention can also be provided in the form of a program (machine learning program) that causes the computer 900 to function as each part of the machine learning device 4, or a program (machine learning program) that causes the computer 900 to execute each step of the machine learning method. Furthermore, the present invention can also be provided in the form of a program (information processing program) that causes the computer 900 to function as each part of the information processing device 5, or a program (information processing program) that causes the computer 900 to execute each step of the information processing method according to the above embodiment.

[0177] (Inference device, inference method, and inference program) The present invention can be provided not only in the form of the information processing device 5 (information processing method or information processing program) according to the above embodiment, but also in the form of an inference device (inference method or inference program) used to infer reliability information of the polishing endpoint detection function. In that case, the inference device (inference method or inference program) may include a memory and a processor, the processor of which may execute a series of processes. The series of processes includes an information acquisition process (information acquisition step) for acquiring reliability degradation factor state information, and an inference process (inference step) for inferring reliability information of the polishing endpoint detection function, which indicates the reliability of the polishing endpoint detection function of a substrate that has undergone polishing based on the reliability degradation factor state information, once the reliability degradation factor state information has been acquired in the information acquisition process.

[0178] By providing an inference device (inference method or inference program), the information processing device can be implemented. Compared to the case of installation, it becomes easier to apply to various devices. It will be obvious to those skilled in the art that when the inference device (inference method or inference program) infers reliability information of the polishing endpoint detection function, it may apply the inference method performed by the state prediction unit using the trained learning model generated by the machine learning device and machine learning method according to the above embodiment. [Explanation of Symbols]

[0179] 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, 10...Learning model, 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...Cleaning unit, 25...Film thickness measurement unit, 26...Control unit, 30…Production history information, 31…Polishing test information, 40...Machine learning control unit, 41...Communication unit, 42...Training data storage unit, 43...Trained model memory unit, 50... Information processing 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...Board processing 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...Elastic membrane, 2212a~2212d...Elastic membrane pressure chamber, 2213... Retainer ring, 2214... Retainer ring airbag (retainer ring pressing mechanism), 2214a... Retainer ring pressure chamber, 226…Optical sensor

Claims

1. An information acquisition unit that acquires reliability degradation factor state information, which includes at least one of the following: wear status information indicating the wear status of the components of the substrate processing apparatus, and processing status information indicating the processing status during the polishing process, in a chemical mechanical polishing process of a substrate performed by the substrate processing apparatus. The system includes a state prediction unit that predicts the reliability information of the polishing endpoint detection function in relation to the reliability reduction factor state information by inputting the reliability reduction factor state information acquired by the information acquisition unit into a learning model that has been trained by machine learning to correlate the aforementioned reliability reduction factor state information with the reliability information of the polishing endpoint detection function, which indicates the reliability of the endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint. Information processing device.

2. The substrate processing apparatus further comprises a polishing table that rotatably supports a polishing pad, a top ring that presses a substrate against 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 a substrate processing control unit that comprehensively controls the entire substrate processing apparatus. The aforementioned wear status information is, The condition of the aforementioned polishing pad, The condition of the rotary connector that rotatably mounts the top ring and the polishing table, The condition of the dresser, and The conditions of the substrate processing control unit include at least one of the following: The information processing apparatus according to claim 1.

3. The substrate processing apparatus further comprises a transparent liquid supply unit that supplies a transparent liquid to the polishing pad, and an optical sensor that measures the reflection intensity of light from a light-emitting lamp to detect when the chemical mechanical polishing process has reached its endpoint. The aforementioned wear status information is, The condition of the aforementioned optical sensor, The condition of the transparent liquid supply unit, including at least one of the following: The information processing apparatus according to claim 2.

4. The substrate processing apparatus comprises a polishing table that rotatably supports a polishing pad, a top ring that presses the substrate against the polishing pad, and a polishing fluid supply unit that supplies polishing fluid to the polishing pad. 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, An elastic membrane housed in the top ring body presses the substrate against the polishing pad in accordance with the pressure fluid supplied to the elastic membrane pressure chamber, A retainer ring is positioned on the outer circumference of the elastic membrane and presses against the polishing pad, The device includes a retainer ring pressing mechanism for adjusting the pressing force of the retainer ring, The aforementioned processing state information is The flow rate of the polishing fluid, The pressing force of the retainer ring pressing mechanism, The rotational torque of the top ring and the polishing table, The oscillation torque of the top ring, The vibration of the top ring, The oscillation torque of the dresser, Noise during polishing of the top ring, The temperature of the polished surface during polishing of the top ring, The temperature control water temperature of the polishing table, The statistical values ​​of the time until endpoint detection for each of the aforementioned substrates, and The statistical values ​​of the sensor time series data for each of the aforementioned substrates, including at least one of these, The information processing apparatus according to claim 2 or 3.

5. The substrate processing apparatus further comprises a transparent liquid supply unit for supplying a transparent liquid to the polishing pad, and an optical sensor for detecting when the chemical mechanical polishing process has reached its endpoint. The aforementioned processing state information is The light reflectance intensity of the optical sensor, and The flow rate of the transparent liquid, including at least one of the following: The information processing apparatus according to claim 4.

6. The reliability information for the aforementioned polishing endpoint detection function is as follows: Current reliability information for endpoint detection, Warning signs of declining reliability, Information on the types of components of the substrate processing apparatus that cause a decrease in reliability, and This includes at least one of the following: information on the type of substrate processing process that causes a decrease in reliability, The information processing apparatus according to claim 1.

7. A learning data storage unit stores multiple sets of learning data, each set comprising: reliability degradation factor state information, which includes at least one of wear status information indicating the wear status of components of a substrate processing apparatus and processing status information indicating the processing status during polishing, and reliability information of a polishing endpoint detection function indicating the reliability of an endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint; and reliability information of a polishing endpoint detection function indicating the reliability of an endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint. 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 reliability degradation factor state information and the reliability information of the polishing endpoint detection function, 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, Machine learning device.

8. An information acquisition step for acquiring reliability degradation factor state information, which includes at least one of the following: wear status information indicating the wear status of the components of the substrate processing apparatus, and processing status information indicating the processing status during the polishing process, in a chemical mechanical polishing process of a substrate performed by the substrate processing apparatus. The system includes a state prediction step, in which the reliability information of the polishing endpoint detection function is predicted in relation to the reliability reduction factor state information by inputting the reliability reduction factor state information obtained in the information acquisition step into a learning model that has been trained by machine learning to correlate the reliability information of the polishing endpoint detection function, which indicates the reliability of the endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint, with the reliability reduction factor state information obtained in the information acquisition step. Information processing methods.

9. A learning data storage step that stores multiple sets of learning data, each consisting of reliability degradation factor state information, which includes at least one of wear status information indicating the wear status of components of a substrate processing apparatus and processing status information indicating the processing status during polishing, and reliability information of a polishing endpoint detection function indicating the reliability of an endpoint detection function that detects when the chemical mechanical polishing process has reached its endpoint; 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 reliability degradation factor state information and the reliability information of the polishing endpoint detection function, The system includes a trained model storage step for storing the trained model that has learned the correlation relationship through the machine learning step. Machine learning methods.

Images