Information processing device

Abstract

The information processing apparatus (5) of the present invention includes: an information acquisition unit (500) that acquires damage state information, the aforementioned damage state information including damage state information showing the damage state when damage occurs on the substrate, and device state information showing the state of the polishing unit when performing a substrate processing process on the substrate; and a damage process determination unit (501) that inputs the damage state information acquired by the information acquisition unit (500) based on the substrate damage into a learning model (11) to determine the process that causes the substrate damage. The aforementioned learning model learns the damage state information and the correlation between the damage and damage process information through machine learning. The aforementioned damage process information shows the process that causes the substrate damage among the various processes included in the substrate processing process.

Description

Technical Field

[0001] This 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. Prior Technology

[0002] As a substrate processing apparatus for performing various processing on substrates such as semiconductor wafers, a substrate processing apparatus for performing chemical mechanical polishing (CMP) is known. Because the substrates processed by the substrate processing apparatus are formed into thin plates, the substrates may still be damaged not only during the actual polishing process but also during the process of transferring substrates between units (for example, see Patent Document 1 and Patent Document 2). [Previous Technical Documents] [Patent Literature]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2020-188233 [Patent Document 2] Japanese Patent Application Publication No. 2000-223380 Summary of the Invention

[0004] [The problem the invention aims to solve]

[0005] When substrate damage occurs during the substrate processing steps performed by a substrate processing apparatus, it is necessary to inspect and repair each unit of the substrate processing apparatus and confirm various device parameters referenced during the operation of the substrate processing apparatus. Especially in the series of steps involving the polishing unit and the substrate transport unit that transfers substrates between them, if it is unclear which step caused the damage, the above-mentioned work cannot be effectively performed, and substrate damage may occur again. Furthermore, analyzing the cause of substrate damage largely relies on the experience and insights of the substrate processing apparatus user; inappropriate analysis may lead to more serious defects and reduced productivity.

[0006] In view of the above-mentioned problems, the present invention aims to provide an information processing apparatus, inference apparatus, machine learning apparatus, information processing method, inference method, and machine learning method that can analyze the causes of substrate damage in a substrate processing apparatus, without relying on the user's experience and insights. [Solutions]

[0007] To achieve the above objectives, one aspect of the information processing apparatus of the present invention includes: The information acquisition unit acquires damage status information during a substrate processing process performed by a substrate processing apparatus consisting of a polishing unit that performs substrate polishing and a substrate transport unit that transfers the substrate between the polishing unit and the polishing unit. This damage status information includes damage status information showing the damage state when damage occurs on the substrate, and apparatus status information showing the state of the polishing unit during the substrate processing process. The damage process identification unit identifies the process that causes the substrate damage by inputting the damage state information obtained by the aforementioned information acquisition unit into a learning model. The learning model learns the relationship between the damage state information and the damage process information through machine learning. The relationship between the damage process information shows the process that causes the substrate damage among the various processes included in the substrate processing process. [Effects of the Invention]

[0008] When using one aspect of the information processing device of the present invention, since the cause of the substrate damage is determined by inputting the damage state information, which includes damage state information and device state information obtained from the damage of the substrate, into the learning model, the damage of the substrate can be responded to quickly and appropriately without relying on the user's experience and insights.

[0009] Other problems, structures, and effects not described above can be understood by referring to the forms used to implement the invention, which will be discussed later. Simple Explanation of the Diagram

[0010] Figure 1 is an overall configuration diagram of an example of a display substrate processing system 1. Figure 2 is a top view of an example of a substrate processing device 2. Figure 3 is a perspective view showing one example of the first to fourth grinding sections 22A to 22D. Figure 4 is a schematic cross-sectional view showing one example of the top ring 221. Figure 5 is a top view schematically showing one of the first and second linear transmission machines 230A and 230B. Figure 6 shows a schematic front view of one of the first and second linear transmission machines 230A and 230B. Figure 7A is a schematic diagram of one example of the display substrate receiving process. Figure 7B is a schematic diagram of one example of the display substrate receiving process. Figure 7C is a schematic diagram of an example of the substrate receiving process in the display industry. Figure 8A is a schematic diagram of one example of the display substrate delivery process. Figure 8B is a schematic diagram of one example of the display substrate delivery process. Figure 8C is a schematic diagram of one example of the delivery process of a display substrate. Figure 9 is a block diagram of an example of the display substrate processing apparatus 2. Figure 10 is a timeline diagram of an example of the substrate processing process performed by the substrate processing apparatus 2. Figure 11 shows a hardware configuration diagram of an example of a computer 900. Figure 12 is a data structure diagram showing an example of resume information 30 managed by database device 3. Figure 13 is a block diagram showing an example of a machine learning device 4. Figure 14 shows an example of the learning model 10 and the learning data 11. Figure 15 is a flowchart showing an example of a machine learning method implemented by the machine learning device 4. Figure 16 is a block diagram showing an example of an information processing device 5. Figure 17 is a functional illustration of an example of the information processing device 5. Figure 18 is a flowchart showing an example of an information processing method implemented by the information processing device 5. Implementation

[0011] The following describes embodiments of the present invention with reference to drawings. The necessary scope is schematically shown in the description used to achieve the purpose of the present invention. This description mainly explains the necessary scope, while omissions are based on prior art.

[0012] Figure 1 is an overall configuration diagram of an example of the substrate processing system 1. The substrate processing system 1 of this embodiment functions as a system for managing the chemical mechanical polishing (hereinafter referred to as "polishing process") of a substrate (hereinafter referred to as "wafer") W such as a semiconductor wafer.

[0013] The substrate processing system 1 mainly comprises: 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 2-6 is, for example, a general-purpose or special-purpose computer (see Figure 11 below), and is configured to be connected to a wired or wireless network 7, capable of transmitting and receiving various types of data (a portion of the data transmission and reception is illustrated by dashed arrows in Figure 1). Furthermore, the number of devices 2-6 and the connection configuration of the network 7 are not limited to the example in Figure 1 and can be appropriately modified.

[0014] The substrate processing apparatus 2 is an apparatus for performing a polishing process to flatten the surface of a wafer W. The substrate processing apparatus 2 is composed of multiple units and performs a series of operations on one or more wafers W, such as loading, polishing, cleaning, drying, film thickness measurement, and unloading. At this time, the substrate processing apparatus 2 refers to apparatus setting information 265, which is composed of multiple apparatus parameters set in each unit, and substrate processing scheme information 266, which determines the polishing conditions during the polishing process, and controls the operation of each unit. Furthermore, the substrate processing apparatus 2 includes a camera 201 mounted at a position capable of photographing the wafer W.

[0015] The substrate processing apparatus 2 transmits various reports R to the database apparatus 3, user terminal apparatus 6, etc., according to the operation of each unit. The various reports R include, for example: process information of the wafer W that is identified as the object during each process, device status information of each unit that displays the status during each process, image information captured by the camera 201, event information detected by the substrate processing apparatus 2, and operation information of the substrate processing apparatus 2 by users (operators, production managers, maintenance managers, etc.).

[0016] The database device 3 manages the history information 30 during the operation of the substrate processing device 2. The database device 3 receives various reports R from the substrate processing device 2 at any time and records them in the history information 30 outside the substrate processing device 2. The content of the reports R is stored in the history information 30 along with date and time information. In addition to the history information 30, the database device 3 can also store device setting information 265 and substrate processing scheme information 266, which can then be made available for the substrate processing device 2 to refer to.

[0017] The machine learning device 4 operates as the main body of the learning stage of machine learning. For example, it obtains a portion of the resume information 30 from the database device 3 as learning data 11, and generates a learning model 10 for use by the information processing device 5 through machine learning. The completed learning model 10 is provided to the information processing device 5 via the network 7 and recording media. In this embodiment, the machine learning method is explained in the case of teacher-led learning.

[0018] The information processing device 5 operates as the main body of the inference stage of machine learning. In the substrate processing process implemented by the substrate processing device 2, when damage occurs on the wafer W (including damage such as defects and cracks), the learning model 10 generated by the machine learning device 4 identifies the process that is the cause of the damage to the wafer W (the damage occurrence process), and transmits the damage occurrence process information to the database device 3, the user terminal device 6, etc.

[0019] User terminal device 6 is a terminal device for user use, and can be a fixed device or a portable device. User terminal device 6 accepts various input operations through the display screen of applications, web browsers, etc., and displays various information (e.g., notification events, information on damage processes, history information 30, etc.) through the display screen. In addition, user terminal device 6 has a built-in or externally connected camera 60.

[0020] The camera 201 of the substrate processing apparatus 2 and the camera 60 of the user terminal device 6 function as imaging devices when wafer W is damaged, for example, to photograph the damaged wafer W (hereinafter referred to as "damaged wafer") and generate image information. The imaging device, such as the camera 201 of the substrate processing apparatus 2, can automatically photograph the damaged wafer W during a series of actions of the substrate processing apparatus 2, or, like the camera 60 of the user terminal device 6, can manually photograph the damaged wafer W based on the user's photographing operation. Furthermore, the imaging device can be either the camera 201 of the substrate processing apparatus 2 or the camera 60 of the user terminal device 6, or it can be an external device such as an appearance inspection device, replacing or adding to it. Moreover, the camera 201 of the substrate processing apparatus 2 can automatically photograph the entire wafer W, regardless of whether the wafer W is damaged.

[0021] (Substrate processing apparatus 2) Figure 2 is a top view showing an example of a substrate processing apparatus 2. The substrate processing apparatus 2, when viewed from above, comprises, within a roughly rectangular frame 20, a loading / unloading unit 21, a polishing unit 22, a substrate conveying unit 23, a cleaning unit 24, a film thickness measurement unit 25, and a control unit 26. The loading / unloading unit 21 is separated from the polishing unit 22, the substrate conveying unit 23, and the cleaning unit 24 by a first partition wall 200A, and the substrate conveying unit 23 is separated from the cleaning unit 24 by a second partition wall 200B.

[0022] (Loading / Unloading Unit) The loading / unloading unit 21 includes: first to fourth front loading sections 210A to 210D for loading wafer cassettes (FOUP, etc.) that can hold multiple wafers W in the vertical direction; a transport robot 211 that can move up and down along the storage direction (vertical direction) of the wafers W stored in the wafer cassettes; and a horizontal moving mechanism 212 that moves the transport robot 211 along the arrangement direction (short side direction of the frame 20) of the first to fourth front loading sections 210A to 210D.

[0023] The transfer robot 211 is configured to access the wafer cassettes, substrate transfer unit 23 (specifically, the elevator 232 described later), cleaning unit 24 (specifically, the drying chamber 241 described later), and film thickness measurement unit 25 respectively loaded in the first to fourth front loading sections 210A-210D, and has upper and lower arms (not shown) for transferring wafers W between them. The lower arm is used for transferring wafers W before processing, and the upper arm is used for transferring wafers W after processing. When transferring wafers W to the substrate transfer unit 23 and the cleaning unit 24, a gate (not shown) provided on the first partition wall 200A is opened and closed.

[0024] (Grinding unit) The polishing unit 22 has first to fourth polishing sections 22A to 22D that perform polishing (planarization) on the wafer W respectively. The first to fourth polishing sections 22A to 22D are arranged along the length of the frame 20.

[0025] Figure 3 is a perspective view showing an example of one of the first to fourth grinding sections 22A to 22D. The first to fourth grinding sections 22A to 22D have the same basic structure and function.

[0026] The first to fourth grinding sections 22A to 22D each include: a grinding table 220 on which a grinding pad 2200 with a grinding surface is mounted; a top ring (grinding head) 221 for holding the wafer W and grinding it while pressing the wafer W onto the grinding pad 2200 on the grinding table 220; a grinding fluid supply nozzle 222 for supplying grinding fluid (slurry) or dressing fluid (e.g., pure water) to the grinding pad 2200; a dressing device 223 for dressing the grinding surface of the grinding pad 2200; and an atomizer 224 for spraying a mixture of liquid (e.g., pure water) and gas (e.g., nitrogen) or liquid (e.g., pure water) into a mist onto the grinding surface.

[0027] The grinding table 220 includes a rotary movement mechanism 220b that is supported by a grinding table shaft 220a and drives the grinding table 220 to rotate around its axis. The top ring 221 includes: a rotary movement mechanism 221c that is supported by a top ring shaft 221a that can move in the vertical direction and drives the top ring 221 to rotate around its axis; an up-down movement mechanism 221d that moves the top ring 221 in the vertical direction; and a rotary movement mechanism 221e that forms the support shaft 221b as the center of rotation and causes the top ring 221 to rotate (rock). The grinding slurry supply nozzle 222 includes a rotary movement mechanism 222b that is supported by a support shaft 222a and forms the support shaft 222a as the center of rotation and causes the grinding slurry supply nozzle 222 to rotate. The trimmer 223 includes: a rotary movement mechanism 223c supported by a trimmer shaft 223a movable in the vertical direction and driven to rotate the trimmer 223 around its axis; an up-down movement mechanism 223d that moves the trimmer 223 in the vertical direction; and a rotary movement mechanism 223e that forms a rotation center with the support shaft 223b and causes the trimmer 223 to rotate. The atomizer 224 includes a rotary movement mechanism 224b supported by a support shaft 224a that forms a rotation center with the support shaft 224a and causes the atomizer 224 to rotate.

[0028] In addition, Figure 3 omits the specific configuration of the rotary movement mechanism 220b, 221c, 223c, the up and down movement mechanism 221d, 223d, and the rotary movement mechanism 221e, 222b, 223e, 224b, but is composed of appropriate combinations of actuators such as motors and air cylinders; driving force transmission mechanisms such as linear guides, ball screws, gears, belts, couplers, and bearings; and detectors such as linear detectors, code detectors, and limit detectors.

[0029] Figure 4 is a schematic cross-sectional view showing one example of the top ring 221. The top ring 221 includes: a top ring body 2210 mounted on the top ring shaft 221a; a roughly disc-shaped carrier 2211 disposed inside the top ring body 2210; a diaphragm 2212 disposed below the carrier 2211 to press the wafer W against the polishing pad 2200; a roughly annular retaining ring 2213 disposed on the outer periphery of the carrier 2211 to directly press the polishing pad 2200; and a retaining ring airbag 2214 disposed between the top ring body 2210 and the retaining ring 2213 to press the retaining ring 2213 against the polishing pad 2200.

[0030] The diaphragm 2212 is formed of an elastic membrane and has a plurality of concentric partition walls 2212e inside, forming first to fourth diaphragm pressure chambers 2212a to 2212d arranged concentrically from the center of the top ring body 2210 outwards. Furthermore, the diaphragm 2212 has a plurality of holes 2212f on its underside for adsorbing the wafer W, thus functioning as a substrate holding surface for holding the wafer W. The retaining ring airbag 2214 is formed of an elastic membrane and has a retaining ring pressure chamber 2214a inside. Additionally, the configuration of the top ring 221 can be adapted to include a pressure chamber that presses down on the entire carrier 2211, and the number and shape of the diaphragm pressure chambers in the diaphragm 2212 can be adapted to include the number and arrangement of the adsorption holes 2212f. Alternatively, the diaphragm 2212 may not have adsorption holes 2212f.

[0031] The first to fourth flow paths 2216A~2216D are connected to the first to fourth diaphragm pressure chambers 2212a~2212d, respectively, and the fifth flow path 2216E is connected to the retaining ring pressure chamber 2214a. The first to fifth flow paths 2216A~2216E are connected to the outside via a rotary joint 2215 located on the top ring shaft 221a, and branch into first branch flow paths 2217A~2217E and second branch flow paths 2218A~2218E, respectively. Pressure detectors PA~PE are respectively installed in the first to fifth flow paths 2216A~2216E. The first branch flow paths 2217A~2217E are connected to the gas supply source GS of the pressurized fluid (air, nitrogen, etc.) via valves V1A~V1E, flow detectors FA~FE, and pressure regulators RA~RE. The second branching paths 2218A~2218E are connected to the vacuum source VS via valves V2A~V2E respectively, and are configured to be open to the atmosphere via valves V3A~V3E.

[0032] The wafer W is held under the top ring 221 and moved to the designated polishing position on the polishing table 220. Polishing is then performed by pressing the polishing surface of the polishing pad 2200, which has been supplied with polishing slurry from the polishing slurry supply nozzle 222, against the surface of the polishing pad 2200. At this time, the top ring 221, using independently controlled pressure regulators RA~RE, adjusts the pressing pressure of the wafer W against the polishing pad 2200 by the pressurized fluid supplied to the first to fourth diaphragm pressure chambers 2212a~2212d according to each region of the wafer W. It also adjusts the pressing pressure of the retaining ring 2213 against the polishing pad 2200 by the pressurized fluid supplied to the retaining ring pressure chamber 2214a. The pressure of the pressurized fluid supplied to the first to fourth diaphragm pressure chambers 2212a~2212d and the retaining ring pressure chamber 2214a is measured by pressure detectors PA~PE, and the flow rate of the pressurized gas is measured by flow detectors FA~FE.

[0033] (Substrate transport unit) As shown in FIG2, the substrate transport unit 23 includes: first and second linear conveyors 230A and 230B that can move horizontally along the arrangement direction of the first to fourth grinding sections 22A to 22D (the length direction of the frame 20); a swing conveyor 231 disposed between the first and second linear conveyors 230A and 230B; a lift 232 disposed on the side of the loading / unloading unit 21; and a temporary stage 233 for the wafer W disposed on the side of the cleaning unit 24. In addition, cameras 201 are respectively disposed at positions that can photograph the wafers W transported by the first and second linear conveyors 230A and 230B.

[0034] The first linear conveyor 230A is configured adjacent to the first and second polishing sections 22A and 22B, and is a mechanism for conveying the wafer W between four conveying positions (numbered sequentially from the loading / unloading unit 21 to the fourth conveying positions TP1 to TP4). The second conveying position TP2 is the position where the wafer W is transferred to the first polishing section 22A, and the top ring 221 of the first polishing section 22A is configured to move between the second conveying position TP2 and the polishing position by a rocking motion. The third conveying position TP3 is the position where the wafer W is transferred to the second polishing section 22B, and the top ring 221 of the second polishing section 22B is configured to move between the third conveying position TP3 and the polishing position by a rocking motion.

[0035] The second linear conveyor 230B is configured adjacent to the third and fourth polishing sections 22C and 22D, and is a mechanism for transporting the wafer W between three transport positions (numbered fifth to seventh transport positions TP5 to TP7 sequentially from the loading / unloading unit 21 side). The sixth transport position TP6 is the position where the wafer W is transferred to the third polishing section 22C, and the top ring 221 of the third polishing section 22C is configured to move between the sixth transport position TP6 and the polishing position by a rocking motion. The seventh transport position TP7 is the position where the wafer W is transferred to the fourth polishing section 22D, and the top ring 221 of the fourth polishing section 22D is configured to move between the seventh transport position TP7 and the polishing position by a rocking motion.

[0036] The swing conveyor 231 is configured adjacent to the fourth and fifth transport positions TP4 and TP5, and has an arm that can move between the fourth and fifth transport positions TP4 and TP5. The swing conveyor 231 is a mechanism for transferring wafer W between the first and second linear conveyors 230A and 230B, and temporarily placing wafer W on the temporary stage 233.

[0037] The elevator 232 is configured adjacent to the first transport position TP1 and is a mechanism for transferring the wafer W between the elevator and the transport robot 211 of the loading / unloading unit 21. During the transfer of the wafer W, the shutter (not shown) located on the first partition wall 200A is opened and closed.

[0038] Figure 5 is a schematic top view showing one example of the first and second linear transmitters 230A and 230B. Figure 6 is a schematic front view showing one example of the first and second linear transmitters 230A and 230B. Figure 7 is a schematic diagram showing one example of the substrate receiving process. Figure 8 is a schematic diagram showing one example of the substrate delivery process.

[0039] The first and second linear conveyors 230A and 230B each include: a conveyor arm 2300 for holding the wafer W; an up-and-down movement mechanism 2301 for moving the conveyor arm 2300 in the up-and-down direction; a horizontal movement mechanism 2302 for moving the conveyor arm 2300 and the up-and-down movement mechanism 2301 along the arrangement direction of the first to fourth grinding sections 22A to 22D (the length direction of the frame 20); and retaining ring stations 2303 respectively provided at each position (second conveyor position TP2, third conveyor position TP3, sixth conveyor position TP6, and seventh conveyor position TP7) where the wafer W is transferred between the conveyor arm and the top ring 221. The first and second linear conveyors 230A and 230B are configured with the conveyor arm 2300, the up-and-down movement mechanism 2301, and the horizontal movement mechanism 2302 as a group, and thus have such multiple groups.

[0040] The transport arm 2300 has a shape that supports a portion of the lower outer periphery of the wafer W. The retaining ring station 2303 has a plurality of lifting pins 2303a positioned opposite the retaining ring 2213 of the top ring 221 and lifting the retaining ring 2213. When the transport arm 2300 is positioned below the retaining ring station 2303 by the horizontal movement mechanism 2302, the retaining ring station 2303 is positioned in a position that does not interfere with the transport arm 2300 when it is raised by the vertical movement mechanism 2301. Additionally, the retaining ring station 2303 may also have a release nozzle for supplying fluid for releasing the wafer W.

[0041] In Figures 5 and 6, the specific configurations of the vertical moving mechanism 2301 and the horizontal moving mechanism 2302 are omitted. However, the vertical moving mechanism 2301 and the horizontal moving mechanism 2302 are, for example, appropriately combined actuators such as motors and air cylinders; driving force transmission mechanisms such as linear guides, ball screws, gears, couplers, belts, and bearings; and detectors such as linear detectors, code detectors, and limit detectors.

[0042] As shown in Figure 7, during the substrate receiving process of the wafer W before the grinding unit 22 receives the wafer W from the substrate transfer unit 23 for grinding processing, the top ring 221 of the wafer W is not kept down, and the transfer arm 2300 of the wafer W is kept up. With the top ring 221 down and the retaining ring 2213 lifted by the lifting pin 801, when the transfer arm 2300 rises further, the top surface of the wafer W contacts the bottom surface of the diaphragm 2212. At this time, for example, in the second diaphragm pressure chamber 2212b corresponding to the position where the adsorption hole 2212f is formed, the wafer W is adsorbed and held in the retaining ring 2213 by vacuum attraction from the vacuum source VS. Then, the top ring 221 of the wafer W, which is held in adsorption, rises, and the transfer arm 2300 that delivers the wafer W descends.

[0043] As shown in Figure 8, during the substrate delivery process where the polishing unit 22 delivers the polished wafer W to the substrate transport unit 23, the top ring 221 holding the wafer W is lowered while the transport arm 2300 is not raised. With the top ring 221 lowered and the retaining ring 2213 raised by the lifting pin 801, as the transport arm 2300 rises further, the bottom of the wafer W approaches the transport arm 2300. At this point, for example, the vacuum suction on the second diaphragm pressure chamber 2212b corresponding to the location where the suction hole 2212f is formed is stopped, and pressurized fluid is supplied to the third diaphragm pressure chamber 2212c located outside the second diaphragm pressure chamber 2212b, releasing the wafer W from the diaphragm 2212. Then, the top ring 221, having released the wafer W, rises, and the transport arm 2300 receiving the wafer W descends.

[0044] (Cleaning Unit) As shown in FIG2, the cleaning unit 24 includes: first and second cleaning chambers 240A and 240B for cleaning wafer W with cleaning fluid; a drying chamber 241 for drying wafer W; and first and second transport chambers 242A and 242B for transporting wafer W. Each chamber of the cleaning unit 24 is arranged along the first and second linear conveyors 230A and 230B in a divided state, for example, in the following order: first cleaning chamber 240A, first transport chamber 242A, second cleaning chamber 240B, second transport chamber 242B, and drying chamber 241 (in the order away from loading / unloading unit 21).

[0045] (Film thickness measurement unit) The film thickness measurement unit 25 is a measuring instrument for measuring the film thickness of the wafer W before or after polishing, and is composed of, for example, an optical film thickness measuring instrument or an eddy current film thickness measuring instrument. The transfer of wafers W between each film thickness measurement module is carried out by a transfer robot 211.

[0046] (Control unit) Figure 9 is a block diagram of an example of the display substrate processing apparatus 2. The control unit 26 is electrically connected to each of the units 21-25 and the camera 201, and functions as a control unit that oversees and controls each of the units 21-25 and the camera 201.

[0047] The loading / unloading unit 21 includes: a plurality of modules 2171-217p (e.g., a transport robot 211) composed of various actuators; a plurality of detectors 2181-218q respectively disposed in the plurality of modules 2171-217p to detect and control the data (detection values) required by each module 2171-217p; and a sequencer 219 for controlling the operation of each module 2171-217p based on the detection values ​​of each detector 2181-218q.

[0048] The grinding unit 22 includes: a plurality of modules 2271-227r composed of various actuators (e.g., grinding table 220, top ring 221, grinding fluid supply nozzle 222, dressing device 223, atomizer 224, etc.); a plurality of detectors 2281-228s respectively disposed in the plurality of modules 2271-227r to detect and control the data (detection values) required by each module 2271-227r; and a sequencer 229 that controls the operation of each module 2271-227r based on the detection values ​​of each detector 2281-228s.

[0049] The detectors 2281-228s of the grinding unit 22 include, for example: a detector for detecting the number of rotations of the grinding table 220; a detector for detecting the rotational torque of the grinding table 220; a detector for detecting the number of rotations of the top ring 221; a detector for detecting the rotational torque of the top ring 221; a detector for detecting the rocking position of the top ring 221; a detector for detecting the rocking torque of the top ring 221; a detector for detecting the height of the top ring 221; a detector for detecting the lifting torque of the top ring 221; a detector for detecting the pressure (positive and negative pressure) in the first to fourth diaphragm pressure chambers 2212a-2212d and the retaining ring pressure chamber 2214a; a detector for detecting the flow rate of the pressurized fluid supplied to the first to fourth diaphragm pressure chambers 2212a-2212d and the retaining ring pressure chamber 2214a; a detector for detecting the flow rate of the grinding fluid supplied from the grinding fluid supply nozzle 222; and a detector for detecting the dripping position of the grinding fluid supply nozzle 222, etc.

[0050] The substrate transport unit 23 includes: a plurality of modules 2371-237t composed of various actuators (e.g., first and second linear conveyors 230A, 230B, swing conveyor 231, elevator 232, etc.); a plurality of detectors 2381-238u respectively disposed in the plurality of modules 2371-237t to detect and control the data (detection values) required by each module 2371-237t; and a sequencer 239 that controls the operation of each module 2371-237t based on the detection values ​​of each detector 2381-238u.

[0051] The detectors 2381 to 238u of the substrate transfer unit 23 include, for example, a detector for detecting the position of the transfer arm 2300; a detector for detecting the height of the transfer arm 2300; and a detector for detecting whether there is a wafer W on the transfer arm 2300.

[0052] The cleaning unit 24 includes: a plurality of modules 2471-247v composed of various actuators (e.g., a first cleaning chamber 240A, a second cleaning chamber 240B, a drying chamber 241, etc.); a plurality of detectors 2481-248w respectively disposed in the plurality of modules 2471-247v to detect and control the data (detection values) required by each module 2471-247v; and a sequencer 249 that controls the operation of each module 2471-247v based on the detection values ​​of each detector 2481-248w.

[0053] The film thickness measurement unit 25 includes: a plurality of modules 2571 to 257x (e.g., film thickness measurement modules) composed of various actuators; a plurality of detectors 2581 to 258y respectively disposed in the plurality of modules 2571 to 257x to detect and control the data (detection values) required by each module 2571 to 257x; and a sequencer 259 that controls the operation of each module 2571 to 257x based on the detection values ​​of each detector 2581 to 258y.

[0054] The control unit 26 includes a control unit 260, a communication unit 261, an input unit 262, an output unit 263, and a memory unit 264. The control unit 26 may be configured as, for example, a general-purpose or special-purpose computer (see Figure 11 described later).

[0055] The communication unit 261 is connected to the network 7 and functions as a communication interface for transmitting and receiving various data. The input unit 262 accepts various input operations, and the output unit 263 outputs various information through display screens, signal tower lights, and buzzer sounds, thus functioning as a user interface.

[0056] The memory unit 264 stores various programs (operating system (OS), applications, web browsers, etc.) and data (device setting information 265, substrate processing scheme information 266, etc.) used when the substrate processing device 2 is in operation. The device setting information 265 and substrate processing scheme information 266 are data that can be edited by the user through the display screen.

[0057] The control unit 260 obtains the detection values ​​of a plurality of detectors 2181~218q, 2281~228s, 2381~238u, 2481~248w, 2581~258y (hereinafter referred to as the "detector group") via a plurality of sequencers 219, 229, 239, 249, 259 (hereinafter referred to as the "detector group"), and performs a series of processes such as loading, grinding, cleaning, drying, film thickness measurement, and unloading by causing a plurality of modules 2171~217p, 2271~227r, 2371~237t, 2471~247v, 2571~257x (hereinafter referred to as the "module group") to operate in a coordinated manner.

[0058] Figure 10 is a timeline diagram showing an example of a substrate processing step performed by the substrate processing apparatus 2. The substrate processing step shown in Figure 10 represents the process up to the point in the above-mentioned series of steps, where the top ring 221 of the polishing unit 22 receives the wafer W before polishing from the first or second linear conveyor 230A, 230B of the substrate transfer unit 23, polishes the wafer W, and then sends the polished wafer W to the first or second linear conveyor 230A, 230B.

[0059] The substrate processing steps include: a substrate receiving step S1 (Fig. 7) in which the top ring 221 receives the wafer W before polishing from the first or second linear conveyor 230A, 230B; a pre-polishing shaking step S2 in which the top ring 221 moves the wafer W before polishing to the polishing position on the polishing table 220; a pre-polishing lowering step S3 in which the top ring 221 lowers the wafer W before polishing to the polishing height; a polishing step S4 in which the top ring 221 polishes the wafer W before polishing; a post-polishing raising step S5 in which the top ring 221 raises the wafer W after polishing to the moving height; a post-polishing shaking step S6 in which the top ring 221 moves the wafer W after polishing to the handover position on the retaining ring station 2303; and a substrate delivery step S7 (Fig. 8) in which the top ring 221 delivers the wafer W after polishing to the first or second linear conveyor 230A, 230B.

[0060] Furthermore, in this embodiment, the aforementioned substrate processing steps are performed between the first and second polishing units 22A and 22B and the first linear transmission machine 230A, and between the third and fourth polishing units 22C and 22D and the second linear transmission machine 230B.

[0061] (Hardware configuration of each device) Figure 11 is a hardware configuration diagram showing an example of a computer 900. The control unit 26 of the board processing device 2, the database device 3, the machine learning device 4, the information processing device 5, and the user terminal device 6 are respectively configured by a general-purpose or special-purpose computer 900.

[0062] As shown in Figure 11, the computer 900 includes the following main components: bus 910, processor 912, memory 914, input device 916, output device 917, display device 918, storage device 920, communication I / F (interface) unit 922, external device I / F unit 924, I / O (input / output) device I / F unit 926, and media input / output unit 928. Furthermore, the above-mentioned components may be appropriately omitted depending on the intended use of the computer 900.

[0063] The processor 912 consists of one or more arithmetic processing units (CPU (Central Processing Unit)), MPU (Micro-processing Unit)), DSP (Digital Signal Processor) and GPU (Graphics Processing Unit), and operates as the control unit overseeing the entire computer 900. The memory 914 stores various data and programs 930, and is composed of volatile memory (DRAM, SRAM, etc.) that performs the functions of main memory, and non-volatile memory (ROM), flash memory, etc.

[0064] Input device 916, for example, comprises a keyboard, mouse, numeric keypad, electronic pen, etc., and functions as an input unit. Output device 917, for example, comprises a voice (sound) output device, vibration device, etc., and functions as an output unit. Display device 918, for example, comprises a liquid crystal display, organic EL display, electronic paper, projector, etc., and functions as a display device. Input device 916 and display device 918, such as a touch panel display, can also be integrated. Storage device 920, for example, comprises an HDD, SSD (Solid State Drive), etc., and functions as a memory unit. Storage device 920 stores various data required for the operation system and program 930 to execute.

[0065] The communication I / F unit 922 connects to a network 940, such as the Internet or an enterprise network (or the same as network 7 in Figure 1), via wired or wireless means, and functions as a communication unit for transmitting and receiving data between other computers according to specified communication standards. The external device I / F unit 924 connects to external devices 950, such as cameras, printers, scanners, and readers / writers, via wired or wireless means, and functions as a communication unit for transmitting and receiving data between itself and the external devices 950 according to specified communication standards. The I / O device I / F unit 926 connects to I / O devices 960, such as various detectors and actuators, and functions as a communication unit for transmitting and receiving various signals and data, such as detector detection signals and actuator control signals, between itself and the I / O devices 960. The media input / output unit 928 is composed of a drive device such as a digital optical disc drive (DVD drive) or an optical disc drive (CD drive), and reads and writes data to media (non-temporary memory media) 970 such as DVDs and CDs.

[0066] In the computer 900 with the above-described configuration, the processor 912 calls the memory 914 to execute the program 930 stored in the storage device 920, and controls various parts of the computer 900 via the bus 910. Alternatively, the program 930 can also be stored in the memory 914, replacing the storage device 920. The program 930 can also be recorded on the media 970 as an installable or executable file and provided to the computer 900 via the media input / output unit 928. The program 930 can also be provided to the computer 900 via the communication I / F (interface) unit 922 and downloaded via the network 940. Furthermore, the computer 900 can also implement the various functions achieved by the processor 912 executing the program 930 using hardware such as an FPGA (FPGA, programmable gate array) or ASIC (Application-Specific Integrated Circuit).

[0067] Computer 900 can be any type of electronic device, such as a stationary computer or a portable computer. Computer 900 can also be a client computer, a server computer, or a cloud computer. Computer 900 can also be used with devices other than devices 2 to 6.

[0068] (Resume Information 30) Figure 12 is a data structure diagram showing an example of the history information 30 managed by the database device 3. The history information 30 is a table that classifies and records various reports R from the substrate processing device 2, and includes, for example: a process history table 300 for process information; a device status history table 301 for device status information (detector detection values, actuator command values, etc.); and a damage occurrence history table 302 for image information and sketch information (described later). In addition to the above, the history information 30 also includes an event history table for event information and an operation history table for operation information, but detailed explanations are omitted.

[0069] The process history table 300 records entries such as wafer ID, box number, batch number, start time of each process S1~S7, end time of each process S1~S7, and user unit ID of each process S1~S7. Additionally, the process history table 300 can also record information about processes other than S1~S7.

[0070] The records in Device Status History Table 301 include, for example, the login unit ID, detector ID (or actuator ID), and time series data. The time series data consists of detector detection values ​​(or actuator command values) sampled at specified time intervals.

[0071] The records in the damage history form 302 include, for example, the wafer ID, acquisition time, damage status information, and damage process information. The damage status information includes image information of the damaged wafer W captured by either the camera 201 of the substrate processing device 2 or the camera 60 of the user terminal device 6, or drawing information of the damaged wafer W drawn by the user. The damage process information includes the damage process identified by the user or the information processing device 5.

[0072] By referring to the process history table 300 and the device status history table 301, the time series data of each detector (or the time series data of each actuator) can be extracted as device status information for each process S1 to S7 included in the substrate processing process of the wafer W identified by the wafer ID. Furthermore, by further referring to the damage occurrence history table 302, damage occurrence status information and damage occurrence process information of the wafer W identified by the wafer ID can be extracted.

[0073] (Machine learning device 4) Figure 13 is a block diagram showing an example of a machine learning device 4. The machine learning device 4 includes: a control unit 40, a communication unit 41, a learning data memory unit 42, and a learning completed model memory unit 43.

[0074] The control unit 40 functions as the learning data acquisition unit 400 and the machine learning unit 401. The communication unit 41 connects to external devices (e.g., board processing device 2, database device 3, information processing device 5, and user terminal device 6) via the network 7, and functions as a communication interface for transmitting and receiving various types of data.

[0075] The learning data acquisition unit 400 connects to an external device via the communication unit 41 and the network 7 to acquire learning data 11, which consists of damage status information as input data and damage process information as output data. Learning data 11 includes data used by teachers as teacher training data, inspection data, and test data during learning. Furthermore, the damage process information is used by teachers as answer labels during learning.

[0076] The learning data memory unit 42 is a database that stores multiple sets of learning data 11 acquired by the learning data acquisition unit 400. Furthermore, the specific structure of the database constituting the learning data memory unit 42 can be appropriately designed.

[0077] The machine learning unit 401 uses a complex set of learning data 11 stored in the learning data memory unit 42 to perform machine learning. That is, the machine learning unit 401 inputs the learning data 11 into the learning model 10 in a complex set, so that the learning model 10 learns the relationship between the information on the damage state and the information on the damage process contained in the learning data 11, and generates a completed learning model 10.

[0078] The learning completion model memory unit 43 is a database that stores the learning model 10 (specifically, the adjusted weighted parameter group) generated by the machine learning unit 401. The learning model 10 stored in the learning completion model memory unit 43 is provided to a physical system (e.g., an information processing device 5) via a network 7 and a recording medium. In Figure 13, the learning data memory unit 42 and the learning completion model memory unit 43 are shown as separate memory units, but they can also be configured as a single memory unit.

[0079] Figure 14 is an example diagram showing the learning model 10 and the learning data 11. The learning data 11 used by the machine learning of the learning model 10 consists of damage state information, which includes damage state information and device state information, and damage process information.

[0080] The damage status information included in the damage status information is information showing the damage status of the wafer W when damage occurs during the substrate processing process performed by the substrate processing apparatus 2. The damage status information is either an image of the damaged wafer W or a depiction of the damaged wafer W.

[0081] The image information is obtained by viewing the damaged wafer W from above using the substrate processing device 2 (camera 201) or the user terminal device 6 (camera 60). For example, if the wafer W is broken into multiple fragments, it should be photographed clearly in its original arrangement to show the extent of damage. Furthermore, the image information can be either a monochrome or color image, or a two-dimensional or three-dimensional image.

[0082] The depiction information includes, for example, the outline of a wafer W with grooves, depicting the cracks when the wafer W is damaged. The depiction information can also be generated by using a drawing application on the user terminal device 6, where the user draws lines to represent the damaged state of the wafer W, or by using an image reading device such as a scanner to read a drawing of a piece of paper in which the user has hand-drawn the damaged state of the wafer W.

[0083] The device status information included in the damage status information displays the status of the polishing unit 22 during the substrate processing process of the damaged wafer W.

[0084] As for the state of the grinding unit 22, the device state information includes at least one of the following: the rocking position of the top ring 221, the height of the top ring 221, the pressure (diaphragm pressure) in the first to fourth diaphragm pressure chambers 2212a~2212d, the flow rate (diaphragm flow rate) of the pressurized fluid supplied to the first to fourth diaphragm pressure chambers 2212a~2212d, the pressure (clasp airbag pressure) in the retaining ring pressure chamber 2214a, and the flow rate (clasp airbag flow rate) of the pressurized fluid supplied to the retaining ring pressure chamber 2214a. Furthermore, the pressure in the first to fourth diaphragm pressure chambers 2212a~2212d and the retaining ring pressure chamber 2214a includes: the pressure when pressurized fluid is supplied, the pressure when the atmosphere is open, and the pressure when vacuum suction is applied.

[0085] In addition to the status of the polishing unit 22, the device status information may also include information on the status of the display substrate transport unit 23. In this case, as the status of the substrate transport unit 23, the device status information includes at least one of the following: the position of the transport arm 2300 (LTP horizontal position), the height of the transport arm 2300 (LTP height), and the presence or absence of a wafer W on the transport arm 2300 (whether or not an LTP substrate is present).

[0086] The information on the damaged process displays information about the processes (damaged processes) among the substrate processing processes S1 to S7 that caused the wafer W to be damaged. The processes S1 to S7 included in the substrate processing process are: substrate receiving process S1 as shown in Figure 10, pre-grinding shaking process S2, pre-grinding lowering process S3, grinding process S4, post-grinding rising process S5, post-grinding shaking process S6, and substrate delivery process S7.

[0087] The learning materials acquisition unit 400 refers to the resume information 30 and, as needed, acquires the learning materials 11 by accepting the input operations of the user on the user terminal device 6.

[0088] For example, when damage occurs on wafer W, the learning data acquisition unit 400 uses the wafer ID of the damaged wafer W and refers to the process history table 300 and device status history table 301 in the history information 30 to obtain the time sequence data of the detector group during the substrate processing process of the damaged wafer W as device status information. Furthermore, the learning data acquisition unit 400 uses the wafer ID and refers to the damage history table 302 in the history information 30 to obtain damage status information of the damaged wafer W. Additionally, the learning data acquisition unit 400 can also obtain image information captured by the user's photography operation and drawing information generated by the user's drawing and reading operations, instead of obtaining damage status information from the history information 30.

[0089] Furthermore, the learning data acquisition unit 400, based on the user's analysis of the state of the polishing unit 22 and the substrate transport unit 23 according to the damage method (characteristics such as the direction and number of cracks) and device status information of the damaged wafer W, obtains the damage process information by accepting the user's specified operation for the damaged wafer W that indicates the damage process. In addition, the learning data acquisition unit 400 can also obtain the damage process information by using the wafer ID of the identified damaged wafer W and referring to the damage process history table 302 of the history information 30, instead of accepting the user's specified operation, when the damage process information obtained by the user in advance analysis is registered in the damage history table 302 of the history information 30.

[0090] The learning model 10, for example, is constructed using a stacked neural network (CNN) and includes: an input layer 100, an intermediate layer 101, and an output layer 102. Synapses (not shown) connecting each neuron are laid between each layer, and each synapse corresponds to a weight. The weighted parameter group, composed of the weights of each synapse, is adjusted by machine learning.

[0091] The input layer 100 has a number of neurons corresponding to the number of pixels in the damage state information (image data or depiction data) used as input data, and the pixel value of each pixel is input to each neuron. The intermediate layer 101 is composed, for example, of a stacked layer 101a, a traction layer 101b, and a fully combined layer 101c. Not only the traction layer 101b, but the fully combined layer 101c also has inputs corresponding to the number of device state information used as input data, and respectively inputs the feature values ​​of the image information of the traction layer 101b and the variable values ​​of the display device state information (e.g., time series data of the detector group) to each neuron. The output layer 102 has neurons corresponding to the number of damage processes (each process S1 to S7) in the damage process information used as output data, and outputs the judgment result (inference result) of each process S1 to S7 as output data. That is, the learning model 10 is a multi-level classification model, and outputs the score (reliability) for each process S1 to S7 with a specified range (e.g., 0 to 1).

[0092] Furthermore, this embodiment describes the acquisition of device status information as time-series data of the detector group shown in FIG14, but it can also be appropriately changed depending on the configuration of the polishing unit 22 (especially the top ring 221) and the substrate transport unit 23. In addition, the device status information can also use command values ​​to the actuator, parameters converted from detector detection values ​​or command values ​​to the actuator, or parameters calculated based on the detection values ​​of multiple detectors. As described above, when changing the definition of device status information, it is only necessary to appropriately change the data composition of the input data in the learning model 10 and the learning data 11.

[0093] Furthermore, in this embodiment, as shown in Figures 10 and 14, the substrate processing steps performed by the substrate processing apparatus 2 are divided into seven steps S1 to S7. However, the positions of the substrate processing steps can be appropriately changed, and they can be divided in more detail or more generally. In addition, the substrate processing steps may further include steps preceding the substrate receiving step S1, or steps following the substrate delivery step S7. Moreover, for example, if the first polishing section 22A is followed by the second polishing section 22B, and the third polishing section 22C is followed by the fourth polishing section 22D, and the polishing process is performed in two stages, or in three or more stages, the substrate processing steps may also be a series of such steps, i.e., a substrate processing step with multiple stages. As described above, when changing the definition of the substrate processing steps, it is only necessary to appropriately change the data composition of the output data in the learning model 10 and the learning data 11.

[0094] (Mechanical learning methods) Figure 15 is a flowchart showing an example of a machine learning method implemented by the machine learning device 4.

[0095] First, in step S100, the learning data acquisition unit 400 acquires a desired amount of learning data 11 from the resume information 30, etc., and stores the acquired learning data 11 in the learning data memory unit 42 as preparation for starting machine learning. The amount of learning data 11 prepared here can be set by considering the inference accuracy required by the final obtained learning model 10.

[0096] Next, in step S110, the machine learning unit 401 prepares a learning model 10 for machine learning before it begins. The learning model 10 prepared at this time is composed of a neural network model as illustrated in Figure 12, and the weights of each synapse are set to initial values.

[0097] Next, in step S120, the mechanical learning unit 401 randomly selects one set of learning data 11 from the multiple sets of learning data 11 stored in the learning data memory unit 42.

[0098] Next, in step S130, the machine learning unit 401 inputs the damage state information (input data) contained in a set of learning data 11 into the input layer of the prepared pre-learning (or learning in progress) learning model 10. As a result, damage process information (output data) is output from the output layer of the learning model 10 as an inference result, but this output data is generated by the pre-learning (or learning in progress) learning model 10. Therefore, the output data output as an inference result from the pre-learning (or learning in progress) state will display information different from the damage process information (answer label) contained in the learning data 11.

[0099] Next, in step S140, the machine learning unit 401 compares the damage process information (answer tags) included in the set of learning data 11 obtained in step S120 with the damage process information (output data) output from the output layer as an inference result in step S130, and performs machine learning by adjusting the weights of each synapse (backpropagation). In this way, the machine learning unit 401 enables the learning model 10 to learn the correlation between the damage state information and the damage process information.

[0100] Secondly, in step S150, the machine learning unit 401 determines whether the specified learning termination condition is met based on, for example, the evaluation value of the error function between the information on the failure process (answer label) contained in the learning data 11 and the information on the failure process (output data) output as a result of inference; and the remaining quantity of unlearned learning data 11 stored in the learning data memory unit 42.

[0101] In step S150, if the machine learning unit 401 determines that the learning termination condition is not met and continues machine learning (No in step S150), the process returns to step S120, and the steps of S120 to S140 are performed multiple times on the learning model 10 using the unlearned learning data 11. Alternatively, if in step S150 the machine learning unit 401 determines that the learning termination condition is met and ends machine learning (Yes in step S150), the process proceeds to step S160.

[0102] Then, in step S160, the machine learning unit 401 stores the completed learning model 10 (adjusted weighted parameter group), generated by adjusting the weights corresponding to each synapse, in the completed learning model memory unit 43, and ends the series of machine learning methods shown in FIG15. In the machine learning method, step S100 is equivalent to the learning data memory process, steps S110 to S150 are equivalent to the machine learning process, and step S160 is equivalent to the completed learning model memory process.

[0103] As described above, when the information processing device 5 and information processing method of this embodiment are adopted, a learning model 10 can be provided from the damage state information, which includes the damage state information and device state information obtained from the damage state information obtained from the damage of wafer W, to determine (infer) the cause of the damage of wafer W.

[0104] (Information Processing Device 5) Figure 16 is a block diagram showing an example of the information processing device 5. Figure 17 is a functional diagram showing an example of the information processing device 5. The information processing device 5 includes: a control unit 50, a communication unit 51, and a learning completion model memory unit 52.

[0105] The control unit 50 functions as the information acquisition unit 500, the damage process identification unit 501, and the output processing unit 502. The communication unit 51 connects to external devices (e.g., the board processing device 2, the database device 3, the machine learning device 4, and the user terminal device 6) via the network 7, and functions as a communication interface for transmitting and receiving various types of data.

[0106] The information acquisition unit 500 connects to an external device via the communication unit 51 and the network 7 to obtain damage status information, including damage status information and device status information. For example, when the information acquisition unit 500 receives the wafer ID of a identified damaged wafer W from the user terminal device 6, it uses the wafer ID to refer to the process history table 300 and device status history table 301 in the history information 30 to obtain the time sequence data of the detector group during the substrate processing process of the damaged wafer W, as device status information. In addition, the information acquisition unit 500 uses the wafer ID to refer to the damage occurrence history table 302 in the history information 30 to obtain damage status information of the damaged wafer W. Furthermore, the information acquisition unit 500 can also obtain image information captured by the user's photography operation and drawing information generated by the user's drawing and reading operations, instead of obtaining damage status information from the history information 30.

[0107] As described above, the damage process identification unit 501 identifies the process (damage process) that is the cause of the wafer W's damage by inputting the damage status information obtained by the information acquisition unit 500 based on the damage to the wafer W into the learning model 10. At this time, the damage process identification unit 501 may, for example, obtain the scores for each process S1 to S7 as output data for the learning model 10, but it generates damage process information by identifying the process corresponding to the maximum score among the scores (in the example of Figure 17, the substrate receiving process S1).

[0108] The learning completion model memory unit 52 is a database that stores the learned learning models 10 used by the damage process identification unit 501. Furthermore, the number of learned models 10 stored in the learning completion model memory unit 52 is not limited to one; it can store multiple learned models with different conditions, such as machine learning methods, types of data contained in damage state information, and types of data contained in damage process information, and can selectively utilize them. In addition, the learning completion model memory unit 52 can also be replaced by the memory unit of an external computer (e.g., a server computer or a cloud computer). In this case, the damage process identification unit 501 only needs to access the external computer.

[0109] The output processing unit 502 performs output processing for outputting the damage process information generated by the damage process identification unit 501. For example, the output processing unit 502 can also transmit the damage process information to the user terminal device 6 and display a screen based on the damage process information on the user terminal device 6, or it can transmit the damage process information to the database device 3 and register the damage process information in the history information 30.

[0110] (Information Processing Methods) Figure 18 is a flowchart showing an example of the information processing method implemented by the information processing device 5. The following is an analysis of the actions taken by the user when a wafer W is damaged, and the user terminal device 6 is used to analyze the cause of the wafer W's damage.

[0111] First, in step S200, when the user inputs the wafer ID of the identified damaged wafer W into the user terminal device 6, and takes a picture of the damaged wafer W with the camera 60, the user terminal device 6 transmits the wafer ID and the image data taken by the camera 60 as damage status information to the information processing device 5.

[0112] Next, in step S210, the information acquisition unit 500 of the information processing apparatus 5 receives the wafer ID and damage status information (image information) transmitted in step S200. In step S211, the information acquisition unit 500 uses the wafer ID received in step S210 and refers to the process history table 300 and device status history table 301 of the history information 30 to obtain device status information when performing a substrate processing process on the damaged wafer W. As a result, in step S212, the information acquisition unit 500 obtains damage status information, which includes image information showing the damage status of the wafer W identified by the wafer ID as damage status information; and device status information when performing a substrate processing process on the wafer W.

[0113] Secondly, in step S220, the damage process identification unit 501 generates damage process information based on the damage state information obtained in step S210 as output data by inputting the damage state information into the learning model 10, and identifies the process in which the wafer W has been damaged.

[0114] Next, in step S230, the output processing unit 502 transmits the damage process information to the user terminal device 6 as output processing for outputting the damage process information generated in step S220. Furthermore, the recipient of the damage process information transmission can be either the user terminal device 6 or, alternatively, the database device 3.

[0115] Secondly, in step S240, the user terminal device 6, in response to the transmission processing in step S200, upon receiving the damage process information transmitted in step S230, displays a screen based on the damage process information, allowing the user to understand the process in which the damaged wafer W was damaged. In the above information processing method, steps S210-S212 correspond to the information acquisition process, step S220 corresponds to the damage process identification process, and step S230 corresponds to the output processing process.

[0116] As described above, when the information processing device 5 and information processing method of this embodiment are used, the cause of the wafer W’s damage is determined by inputting the damage state information, which includes damage state information and device state information, obtained from the damage of the wafer W into the learning model 10. Therefore, the damage of the wafer W can be responded to quickly and appropriately without relying on the user’s experience and insights.

[0117] (Other implementation forms) This invention is not limited to the above-described embodiments, and various modifications can be made to implement it without departing from the spirit of the invention. All of these modifications are included in the technical concept of this invention.

[0118] In the above embodiment, it is explained that the database device 3, the machine learning device 4, and the information processing device 5 are each composed of different devices. However, these three devices can also be composed of a single device, or any two of these three devices can be composed of a single device. In addition, at least one of the machine learning device 4 and the information processing device 5 can also be integrated into the control unit 26 of the substrate processing device 2 or the user terminal device 6.

[0119] In the above implementation, the learning model 10 used to implement the machine learning of the machine learning unit 401 is described in the context of using a neural network-like system, but other machine learning models can also be used. Other machine learning models include, for example: tree-type systems such as decision trees and regression trees; integrated learning such as bagging and boosting; recurrent neural networks; stacked neural networks; neural network types such as LSTM (including deep learning); hierarchical clustering; non-hierarchical clustering; clustering types such as k-nearest neighbor and k-means; multivariate analysis such as principal component analysis, factor analysis, and logistic regression; and support vector machines.

[0120] (Machine learning programs and information processing programs) The present invention can also be provided as a program (machine learning program) that enables the computer 900 to perform the various functions of the machine learning device 4; and as a program (machine learning program) that enables the computer 900 to execute the various steps of the machine learning method. Furthermore, the present invention can also be provided as a program (information processing program) that enables the computer 900 to perform the various functions of the information processing device 5; and as a program (information processing program) that enables the computer 900 to execute the various steps of the information processing method of the above-described embodiment.

[0121] (Inference apparatus, inference method, and inference procedure) This invention is not only provided in the form of the information processing apparatus 5 (information processing method or information processing program) of the above-described embodiment, but also in the form of an inference apparatus (inference method or inference program) used to infer information about a breakage process. In this case, the inference apparatus (inference method or inference program) includes: a memory and a processor, wherein the processor can be an executioner of a series of processes. This series of processes includes: an information acquisition process (information acquisition step), which, during the substrate processing process performed by the substrate processing apparatus 2, acquires breakage state information including breakage state information showing the breakage state when breakage occurs on the wafer W; and device state information showing the state of the polishing unit 22 during the substrate processing process on the wafer W; and an inference process (inference step), which, when the breakage state information is acquired through the information acquisition process based on the breakage on the wafer W, infers the process (breakage process) among the various processes S1 to S7 included in the substrate processing process that is the cause of the breakage of the wafer W.

[0122] By providing the inference device (inference method or inference program) in the form of an inference device, it can be easily applied to various devices compared to installing an information processing device. Those skilled in the art will naturally understand that when the inference device (inference method or inference program) infers a failure process, the learning model 10 generated by the machine learning device 4 and machine learning method described above can also be used to apply the inference method implemented by the failure process identification department.

[0123] [Industrial Applicability] This invention can be applied to information processing devices, inference devices, machine learning devices, information processing methods, inference methods, and machine learning methods.

[0124] 1: Substrate processing system 2: Substrate processing device 3: Database device 4: Machine learning device 5: Information processing device 6: User terminal device 7: Internet 10: Learning Model 11: Study materials 20: Rack 21: Loading / Unloading Unit 22: Grinding Unit 22A~22D: First to Fourth Grinding Sections 23: Substrate transfer unit 24: Cleaning Unit 25: Film thickness measurement unit 26: Control Unit 30: Resume Information 40: Control Department 41:Ministry of Communications 42: Study Materials Memory Section 43: Learning to complete the model memory section 50: Control Department 51: Ministry of Communications 52: Learning to complete the model memory section 60,201: Camera 100: Input Layer 101: Intermediate Layer 101a: Stacked Layers 101b: Traction Layer 101c: Fully bonded layer 102: Output Layer 200A: First partition wall 200B: Second partition wall 210A~210D: First to fourth front loading sections 211: Transport Robot 212: Horizontal Movement Mechanism 219,229,239,249,259: Sequencer 220: Grinding table 220a: Grinding table shaft 220b: Rotary moving mechanism section 221: Top Ring 221a: Top ring shaft 221b: Support shaft 221c: Rotary Moving Mechanism 221d: Up and down moving mechanism 221e: Rotary Moving Mechanism Section 222: Grinding fluid supply nozzle 222a: Support shaft 222b: Rotary Moving Mechanism Section 223: Dresser 223a: Dresser shaft 223b: Support shaft 223c: Rotary moving mechanism 223d: Up and down moving mechanism 223e: Rotary Moving Mechanism Section 224: Atomizer 224a: Support shaft 224b: Rotary Moving Mechanism Section 230A, 230B: First and second linear transmission machines 231: Swinging Conveyor 232: Elevator 233: Temporary Storage Table 240A, 240B: First and second cleaning chambers 241: Drying Chamber 242A, 242B: First and Second Transport Rooms 260: Control Department 261: Ministry of Communications 262: Input Section 263: Output Section 264: Memory Department 265: Device Setting Information 266: Information on substrate processing solutions 2171~217p, 2271~227r, 2371~237t, 2471~247v, 2571~257x: Modules 2181~218q, 2281~228s, 2381~238u, 2481~248w, 2581~258y: Detectors 300: Process History Form 301: Equipment Status History Table 302: Damage Record 400: Learning Materials Acquisition Department 401: Mechanical Engineering Department 500: Information Acquisition Department 501: Damage Inspection Department 502: Output Processing Unit 900: Computer 910: Busbar 912: Processor 914: Memory 916: Input device 917: Output device 918: Display device 920: Storage device 922: Communication I / F Department 924: External Equipment I / F Section 926:I / O device I / F section 928: Media Input / Output Department 930: Program 940: Internet 950: External devices 960:I / O device 970: Media 2200: Grinding pad 2210:Top ring body 2211: Carrier 2212: Diaphragm 2212a~2212d: First to fourth diaphragm pressure chambers 2213: Buckle 2214: Buckle Airbag 2214a: Snap ring pressure chamber 2300: Transporting Arm 2301: Up and Down Moving Mechanism 2302: Horizontal Moving Mechanism Department 2303: Buckle Station W: Wafer

Claims

1. An information processing apparatus comprising: an information acquisition unit that acquires damage state information during a substrate processing process performed by a substrate processing apparatus comprising a polishing unit for polishing a substrate and a substrate transport unit for transferring the substrate between the polishing unit and the polishing unit, the damage state information including damage state information showing the damage state when damage occurs on the substrate and apparatus state information showing the state of the polishing unit when the substrate processing process is performed on the substrate; and a damage process determination unit that determines the preceding process that is the cause of damage to the substrate by inputting the aforementioned damage state information acquired by the information acquisition unit based on the damage to the substrate into a learning model, the learning model being learned by machine learning of the correlation between the aforementioned damage state information and the damage process information, the correlation between the aforementioned damage process information showing the aforementioned process that is the cause of damage to the substrate among the various processes included in the aforementioned substrate processing process.

2. The information processing apparatus of claim 1, wherein the aforementioned damage state information included in the aforementioned damage state information is either image information of the substrate before the damage occurred, or depiction information of the substrate before the damage occurred.

3. The information processing apparatus of claim 1 or 2, wherein the aforementioned apparatus status information included in the aforementioned damage status information as the state of the aforementioned grinding unit includes at least one of the following: the position of the top ring of the aforementioned grinding unit; the height of the aforementioned top ring; the pressure in the pressure chamber of the aforementioned top ring; and the flow rate of the pressure fluid supplied to the aforementioned pressure chamber.

4. The information processing apparatus according to any one of claims 1 to 3, wherein the aforementioned device status information included in the aforementioned damage status information is, in addition to the status of the aforementioned polishing unit, a display of the status of the aforementioned substrate transport unit, and the status of the aforementioned substrate transport unit includes at least one of the following: the position of the aforementioned substrate transport unit; the height of the aforementioned substrate transport unit; and whether or not the aforementioned substrate is present in the aforementioned substrate transport unit.

5. An information processing apparatus according to any one of claims 1 to 4, wherein the information on the aforementioned damage occurrence process, as part of each step in the aforementioned substrate processing step, includes: a substrate receiving step, in which the aforementioned polishing unit receives the substrate before the aforementioned polishing process from the aforementioned substrate transport unit; a pre-polishing shaking step, in which the aforementioned polishing unit moves the substrate before the aforementioned polishing process to the polishing position; a pre-polishing lowering step, in which the aforementioned polishing unit lowers the substrate before the aforementioned polishing process to the polishing height; a polishing step, in which the aforementioned polishing unit performs the aforementioned polishing process on the substrate before the aforementioned polishing process; a post-polishing raising step, in which the aforementioned polishing unit raises the substrate after the aforementioned polishing process to the moving height; a post-polishing shaking step, in which the aforementioned polishing unit moves the substrate after the aforementioned polishing process to the handover position; and a substrate delivery step, in which the aforementioned polishing unit delivers the substrate after the aforementioned polishing process to the aforementioned substrate transport unit.

6. An inference device comprising: a memory and a processor, wherein the processor performs: an information acquisition process, wherein during a substrate processing step performed by a substrate processing apparatus comprising a polishing unit for polishing a substrate and a substrate transport unit for transferring the substrate between the polishing unit and the polishing unit, the information acquisition process acquires damage state information, the damage state information comprising damage state information displaying the damage state when damage occurs on the substrate; and apparatus state information displaying the state of the polishing unit when the substrate processing step is performed on the substrate; and an inference process, wherein when the information acquisition process acquires the damage state information based on damage occurring on the substrate, the inference process deduces the step included in the substrate processing step that is the cause of damage to the substrate.

7. A machine learning device comprising: a learning data storage unit, which stores multiple sets of learning data during a substrate processing process performed by a substrate processing apparatus comprising a polishing unit for polishing a substrate and a substrate transport unit for transferring the substrate between the polishing unit and the polishing unit, the learning data comprising damage state information and damage process information, the damage state information comprising: damage state information displaying the damage state when damage occurs on the substrate; and apparatus state information displaying the state of the polishing unit when the substrate processing process is performed on the substrate, the damage process information displaying the process among the processes included in the substrate processing process that causes damage to the substrate; a machine learning unit, which learns the correlation between the damage state information and the damage process information by inputting multiple sets of the learning data into a learning model; and a learning completion model storage unit, which stores the learning model in which the machine learning unit has learned the correlation.

8. An information processing method comprising: an information acquisition step, wherein in a substrate processing apparatus comprising a polishing unit for polishing a substrate and a substrate transport unit for transferring the substrate between the polishing unit and the polishing unit, damage state information is acquired, the damage state information including damage state information showing the damage state when damage occurs on the substrate and apparatus state information showing the state of the polishing unit when the substrate processing is performed on the substrate; and a damage process determination step, wherein the aforementioned process that is the cause of damage to the substrate is determined by inputting the aforementioned damage state information acquired by the information acquisition step based on the damage to the substrate into a learning model, the learning model being learned by machine learning of the correlation between the aforementioned damage state information and the damage process information, the aforementioned damage process information showing the aforementioned process that is the cause of damage to the substrate among the various processes included in the aforementioned substrate processing.

9. A reasoning method, executed by a reasoning device comprising: memory and a processor, wherein the processor executes: an information acquisition step, wherein in a substrate processing step performed by a substrate processing apparatus comprising a polishing unit for polishing a substrate and a substrate transport unit for transferring the substrate between the polishing unit and the polishing unit, the information on the damage state includes damage state information displaying the damage state when damage occurs on the substrate; and apparatus state information displaying the state of the polishing unit when the substrate processing step is performed on the substrate; and a reasoning step, wherein when the information on the damage state is obtained through the information acquisition step based on damage occurring on the substrate, the reasoning is inferred from the various steps included in the substrate processing step that is the cause of damage to the substrate.

10. A machine learning method comprising: a learning data memory step, wherein during a substrate processing process performed by a substrate processing apparatus comprising a polishing unit for polishing a substrate and a substrate transport unit for transferring the substrate between the polishing unit and the polishing unit, a learning data array is stored in a learning data memory unit, the learning data comprising damage state information and damage process information, the damage state information comprising damage state information displaying the damage state when damage occurs on the substrate; and apparatus state information displaying the state of the polishing unit during the substrate processing process, the damage process information comprising the process that causes damage to the substrate; a machine learning step, wherein the learning model learns the correlation between the damage state information and the damage process information by inputting the learning data array into a learning model; and a learning completion model memory step, wherein the learning model, having learned the correlation through the machine learning step, is stored in a learning completion model memory unit.