Substrate transfer robot system and substrate detection method

Abstract

This substrate transfer robot system (100) is provided with: a robot (20) that transfers a substrate (101); an irradiation unit (50) that irradiates detection light (L1) having a predetermined shape to a main surface (101a) of the substrate (101) held by the robot (20); an imaging unit (40) for imaging the detection light (L1) reflected by the main surface (101a) of the substrate (101) by imaging the substrate (101); and a control unit (60) that detects the state of the substrate (101) on the basis of a captured image of the detection light (L1) captured by the imaging unit (40).

Description

Technical Field

[0001] This disclosure relates to a substrate handling robot system and a substrate inspection method. Background Technology

[0002] Previously, measuring devices for measuring the deformation of semiconductor wafers have been disclosed. The measuring device disclosed in Japanese Patent No. 7169994 measures the deformation of a semiconductor wafer by analyzing an image obtained from photographing a geometric illumination pattern reflected on the semiconductor wafer. The measuring device described in Japanese Patent No. 7169994 photographs a semiconductor wafer being processed on a turntable placed within a vacuum chamber. Since the processing within the vacuum chamber is performed while the turntable rotates, a reflected image of the entire semiconductor wafer can be captured even when the semiconductor wafer is large. Furthermore, Japanese Patent No. 7169994 describes an example where a reflected image of the entire semiconductor wafer is captured by moving the semiconductor wafer linearly using a translation mechanism outside the vacuum chamber.

[0003] Patent Document 1: Japanese Patent No. 7169994

[0004] However, as described in Japanese Patent No. 7169994, when photographing a substrate placed on a turntable within a vacuum chamber to detect substrate deformation or other substrate conditions, the turntable only moves in the rotational direction, making it difficult to adjust the position of the substrate being photographed. Furthermore, when photographing a reflected image of the entire substrate by linearly moving the substrate using a translation mechanism located outside the vacuum chamber, the substrate also only moves linearly, making it difficult to adjust the position of the substrate being photographed. Because the position of the substrate being photographed is difficult to adjust, it is difficult to detect the substrate condition with higher precision. Therefore, when detecting the substrate condition by photographing the substrate, it is desirable to adjust the position of the substrate to detect the substrate condition with higher precision. Summary of the Invention

[0005] This disclosure was made to solve the aforementioned problems. One object of this disclosure is to provide a substrate handling robot system and a substrate inspection method that can detect the state of a substrate with higher accuracy.

[0006] The substrate handling robot system based on the first aspect of this disclosure includes: a robot for handling substrates; an illumination unit for illuminating the main surface of the substrate held on the robot with detection light having a predetermined shape; an imaging unit for capturing the detection light reflected on the main surface of the substrate by imaging the substrate; and a control unit for detecting the state of the substrate based on the captured image of the detection light by the imaging unit.

[0007] As described above, the substrate handling robot system based on the first aspect of this disclosure illuminates the main surface of the substrate held on the robot with a detection light of a predetermined shape. Then, the state of the substrate is detected based on the image captured by the imaging unit of the detection light. Since the substrate is held by the robot handling the substrate, the position of the substrate can be easily adjusted by the robot's movements. Therefore, the substrate can be easily positioned in a suitable location for detecting the state of the substrate by the robot's movements, thus enabling higher precision detection of the substrate's state.

[0008] The substrate inspection method based on the second aspect of this disclosure captures a detection light of a predetermined shape reflected on the main surface of the substrate by photographing the substrate held by a robot transporting the substrate, and detects the state of the substrate based on the captured image of the detection light.

[0009] As described above, the substrate inspection method based on the second aspect of this disclosure captures an image of a substrate held by a robot, and detects the state of the substrate based on the captured image of the detection light. Since the robot handling the substrate holds the image, the position of the substrate can be easily adjusted by the robot's movements. Therefore, a substrate inspection method can be provided that allows the substrate to be easily positioned in a suitable location for detecting its state by the robot's movements, thus enabling higher precision detection of the substrate's state.

[0010] According to this disclosure, the state of the substrate can be detected with higher precision. Attached Figure Description

[0011] Figure 1 This is a top view of a substrate processing system equipped with a substrate handling robot system based on the first embodiment of the present disclosure.

[0012] Figure 2 It is a three-dimensional diagram used to illustrate the structure of the substrate.

[0013] Figure 3 This is a block diagram illustrating the structure of a substrate handling robot system based on a first embodiment of the present disclosure.

[0014] Figure 4 This is a perspective view schematically illustrating the structure of a substrate handling robot system.

[0015] Figure 5 This is an example of an image captured by the imaging unit of the detection light.

[0016] Figure 6 This is a diagram illustrating an example of an image captured based on the movement of a substrate along multiple directions of movement.

[0017] Figure 7This is a diagram used to illustrate the learning process of a model after it has been fully learned.

[0018] Figure 8 This is a flowchart illustrating the substrate inspection method based on the first embodiment.

[0019] Figure 9 This is a perspective view schematically illustrating the structure of a substrate handling robot system based on a second embodiment of the present disclosure.

[0020] Figure 10 This is a perspective view schematically showing the structure of a substrate handling robot system based on a third embodiment of the present disclosure. Detailed Implementation

[0021] [First Implementation]

[0022] Hereinafter, a first embodiment of the present disclosure, which embodies the present disclosure, will be described based on the accompanying drawings.

[0023] Reference Figures 1-7 The substrate handling robot system 100 based on the first embodiment will be described.

[0024] (Substrate processing system)

[0025] like Figure 1 As shown, a substrate handling robot system 100 is disposed in a substrate processing system 400. The substrate processing system 400 includes a processing unit 401 for processing substrates 101. The processing unit 401 includes a robot 402, a loading and locking unit 403, multiple processing module units 404, and a transport chamber 405. The robot 402 is a horizontal multi-joint robot for handling substrates 101. The loading and locking unit 403 is a chamber for picking up and removing substrates 101. That is, the loading and locking unit 403 holds the substrate 101 to be transferred to the processing unit 401. The processing unit 401 of the substrate processing system 400 includes two loading and locking units 403. The loading and locking units 403 are connected to the substrate handling robot system 100. The processing module units 404 perform processes such as resist coating and etching on the substrate 101. The robot 402 is disposed in the transport chamber 405. In the transport chamber 405, the robot 402 transports the substrate 101 between the loading locking unit 403 and the processing module unit 404.

[0026] The substrate handling robot system 100 handles the transfer of substrates 101 between a FOUP 110 (Front Opening Unify Pod) and a processing device 401 within a substrate processing system 400 that processes substrates 101. The substrate handling robot system 100 is an EFEM (Equipment Front End Module). The FOUP 110 is a container for holding multiple substrates 101. In the FOUP 110, the multiple substrates 101 are arranged vertically. The FOUP 110 is an example of a substrate storage container.

[0027] like Figure 2 As shown, substrate 101 is, for example, a silicon wafer with a disk shape. Substrate 101 has a mirror-like main surface 101a. In addition, substrate 101 has a notch 101b indicating the orientation of substrate 101. The notch 101b is a V-shaped groove disposed on the outer periphery of substrate 101 to indicate the orientation of crystal orientation in substrate 101.

[0028] (Substrate handling robot system)

[0029] like Figure 3 As shown, the substrate handling robot system 100 includes a handling chamber 10, a robot 20, an alignment device 30, an imaging unit 40, an illumination unit 50, and a control unit 60. The robot 20 includes a horizontally articulated robotic arm 21 and a substrate holding hand 22. Furthermore, the alignment device 30 is an example of an adjustment device.

[0030] like Figure 4 As shown, the transport chamber 10 is a cubic-shaped shell. The transport chamber 10 has a front surface 11 on the Y1 direction side and a back surface 12 on the Y2 direction side. Additionally, the transport chamber 10 has a top surface 13 on the Z1 direction side, i.e., the vertically upward side. The substrate 101 is transported within the transport chamber 10. The transport chamber 10 internally houses a robot 20, a aligner 30, a camera 40, an irradiation unit 50, and a control unit 60. Three loading ports 11a for mounting FOUP 110 are arranged horizontally on the front surface 11 on the Y1 direction side of the transport chamber 10. The loading ports 11a have opening and closing mechanisms. The loading ports 11a open and close the doors of the mounted FOUP 110. Furthermore, two loading locking parts 403 are connected to the back surface 12 on the Y2 direction side of the transport chamber 10. An opening and closing mechanism is also arranged between the loading locking parts 403 and the back surface 12. Furthermore, in Figure 4 The image shows a state in which a FOUP 110 is installed on one of the three loading ports 11a on the X1 direction side.

[0031] Robot 20 is configured in transport chamber 10 for transporting substrate 101 between FOUP 110 and loading locking part 403. Robot 20 performs the transporting operation of substrate 101 in transport chamber 10. Specifically, in robot 20, an end effector, namely substrate holding hand 22, is arranged at the front end of robotic arm 21 to hold substrate 101. Robot 20 performs the transporting operation of substrate 101 between FOUP 110 and loading locking part 403 mounted on loading port 11a. Robot 20 transports substrate 101 while holding substrate 101 along a horizontal plane in substrate holding hand 22. Substrate holding hand 22 is a passive type hand that does not hold substrate 101 in a fixed position.

[0032] Alignment device 30 aligns the substrate 101 transported by robot 20. Alignment device 30 adjusts the orientation of substrate 101 to a predetermined direction. Alignment device 30 is disposed in transport chamber 10 on the side facing left-right direction, i.e., the X direction, i.e., the X2 direction side. Alignment device 30 has a substrate mounting portion 30a for mounting substrate 101. Alignment device 30 detects the notch 101b of substrate 101 from the outer peripheral portion of substrate 101 by rotating substrate 101 placed on substrate mounting portion 30a along a horizontal plane. Then, based on the position of notch 101b detected from the outer peripheral portion, alignment device 30 rotates substrate 101 along the horizontal plane and performs rotational positioning along the circumferential direction of main surface 101a of substrate 101, thereby adjusting the orientation of substrate 101 to a predetermined direction. In addition, alignment device 30 detects the center position of substrate 101.

[0033] In the substrate handling robot system 100, the substrate 101, held by the robot 20 before processing, is transported from the FOUP 110 to the alignment unit 30. Then, the substrate 101, aligned by the alignment unit 30, is again transported by the robot 20 to the loading locking unit 403. Additionally, the substrate 101, after processing in the processing unit 401, is transported by the robot 20 from the loading locking unit 403 to the FOUP 110. The handling actions of the robot 20 on the substrate 101 include both a removal action (taking out each substrate 101 from the FOUP 110 and the loading locking unit 403) and a loading action (placing the substrate 101 in place).

[0034] The imaging unit 40 is, for example, a two-dimensional camera. The imaging unit 40 includes imaging elements such as a CCD (Charge Coupled Device) and a CMOS (Complementary Metal Oxide Semiconductor). The imaging unit 40 is positioned opposite the substrate 101 held in the state of the robot 20. The imaging unit 40 captures images of the substrate 101 held in the state of the robot 20. In the first embodiment, the imaging unit 40 is positioned above the robot 20 within the transport chamber 10. The imaging unit 40 is, for example, positioned on the top surface 13 of the transport chamber 10. The imaging unit 40 is positioned to capture images of the substrate 101 removed from the FOUP 110. Three imaging units 40 are arranged corresponding to the three loading ports 11a of the FOUP 110.

[0035] The irradiation unit 50 irradiates a detection light L1 onto the main surface 101a of the substrate 101, whose position and angle have changed relative to each other. In the first embodiment, the irradiation unit 50 irradiates a detection light L1 having a predetermined shape onto the main surface 101a of the substrate 101, which is held on the robot 20 and moves relative to it along a predetermined moving direction. The irradiation unit 50 irradiates a linear detection light L1 along a direction intersecting the moving direction of the substrate 101 in the change of relative position with respect to the irradiation unit 50. The irradiation unit 50 includes, for example, an LED (Light-Emitting Diode) that irradiates a red linear light as the detection light L1. Furthermore, the term "linear" light here refers to a shape in which the detection light L1 extends linearly on the irradiated plane when it irradiates a plane such as the main surface 101a of the substrate 101. In addition, the irradiation unit 50 irradiates the detection light L1 onto the substrate 101 when the imaging unit 40 performs an imaging operation. The irradiation unit 50 is disposed above the robot 20 within the transport chamber 10. The irradiation unit 50 is disposed, for example, on the top surface 13 of the transport chamber 10. Like the imaging unit 40, three irradiation units 50 are disposed in a manner corresponding to the three loading ports 11a on which the FOUP 110 is mounted.

[0036] like Figure 1As shown, the control unit 60 includes a machine learning unit 61. The control unit 60 controls the operation of each part of the substrate handling robot system 100. In the first embodiment, the control unit 60 also functions as a robot control unit that controls the operation of the robot 20. That is, the control unit 60 is a robot controller. In addition, the control unit 60 controls the operation of the alignment unit 30. Furthermore, the control unit 60 controls the operation of the imaging unit 40 and the illumination unit 50. The control unit 60 includes, for example, an arithmetic unit such as a CPU (Central Processing Unit). In addition, the control unit 60 also includes memory such as RAM (Random Access Memory) and ROM (Read Only Memory), and storage devices such as hard disks. The control unit 60 performs control processing based on the arithmetic unit based on programs and parameters stored in the storage devices.

[0037] (Image generation)

[0038] like Figure 5 As shown, an image 71 is acquired by the imaging unit 40. The image 71 acquired by the imaging unit 40 is sent to the control unit 60. The imaging unit 40 captures the detection light L1 reflected on the main surface 101a of the substrate 101 by capturing the substrate 101. When the imaging unit 40 captures the detection light L1, the robot 20 changes the relative position of the substrate 101 with respect to the irradiation unit 50. Specifically, when the imaging unit 40 captures the detection light L1, the robot 20 moves the substrate 101 relative to the irradiation unit 50 by moving the substrate 101 along a predetermined moving direction. The irradiation unit 50 irradiates the main surface 101a of the substrate 101, which has moved relative to the robot 20, with the detection light L1. Then, the imaging unit 40 captures the detection light L1 reflected on the main surface 101a of the substrate 101 multiple times, as the relative position of the substrate 101 with respect to the irradiation unit 50 changes, that is, based on the relative movement of the substrate 101 by the robot 20. The control unit 60 generates an image 71 by capturing multiple detection lights L1 reflected by the imaging unit 40 through multiple captures. In detail, the control unit 60 generates an image 71 by combining the multiple image 70 captured by the imaging unit 40 through multiple captures, in a manner that combines the detection lights L1 reflected throughout the substrate 101.

[0039] For example, after the substrate 101, which has not yet been processed by the processing device 401, is removed from the FOUP 110 by the action of the robot 20, as... Figure 5As indicated by the dashed arrow, it moves along the Y2 direction. The irradiation unit 50, when removed from the FOUP 110, continuously irradiates the main surface 101a of the substrate 101, which is moving via the robot 20, with a linear detection light L1 extending in the X direction orthogonal to the movement direction, i.e., the Y2 direction. The imaging unit 40, accompanying the movement of the substrate 101 via the robot 20 in the Y2 direction, continuously captures multiple images of the detection light L1 reflected on the main surface 101a of the substrate 101. The control unit 60 generates an image 71 by combining the images 70 continuously captured by the imaging unit 40 during movement in one movement direction, i.e., the Y2 direction. For example, to generate an image 71, the control unit 60 moves the substrate 101 at a constant speed via the robot 20 and combines the images 70 continuously captured by the imaging unit 40 at constant time intervals. To correspond to the movement of the substrate 101, the synthesis of the captured images 70 is performed by overlapping multiple captured images 70 in a state where they are staggered in the Y direction. That is, a captured image 71 is generated in such a way that it is an image obtained by reflecting multiple detection lights L1 across the entire main surface 101a of the substrate 101 along the movement direction of the substrate 101. The interior of the transport chamber 10 does not contain any light source other than the irradiation unit 50, and is therefore relatively dark. Therefore, each of the multiple captured images 70 is captured as an image that extracts only the detection light L1 reflected on the main surface 101a of the substrate 101. In other words, the captured image 71 is an image captured by extracting the detection light L1 reflected in a manner corresponding to the shape of the substrate 101.

[0040] In addition, such as Figure 6 As shown, the control unit 60 moves the substrate 101 along a movement direction different from that used to generate a single image 71 while the imaging unit 40 captures images of the substrate 101, thereby generating an image 72. That is, in the first embodiment, when the imaging unit 40 captures the detection light L1, the robot 20 moves the substrate 101 along multiple different movement directions. Then, the control unit 60 generates an image 71 and an image 72 obtained by the imaging unit 40 capturing the detection light L1 during each movement of the substrate 101 along the multiple movement directions of the robot 20. For example, the control unit 60 generates two images 71 and 72 by moving the substrate 101 along two movement directions: a first movement direction (Y2 direction) for generating the image 71 and a second movement direction (a different direction from Y2 direction) for generating the image 72. Furthermore, the image 72, like the image 71, is generated by combining multiple consecutively captured images 70.

[0041] (Control processing for detecting the state of the substrate)

[0042] like Figure 7 As shown, in the first embodiment, the control unit 60 detects the state of the substrate 101 based on the captured images 71 and 72 of the detection light L1 captured by the imaging unit 40 within the transport chamber 10 using a learned model 62 generated using machine learning. The control unit 60 detects the state of the substrate 101, including its position and orientation, based on the captured images 71 and 72. Furthermore, the state of the substrate 101 detected by the control unit 60 includes abnormalities such as warping, deformation, notches, cracks, and scratches. The learned model 62 is generated by machine learning in a manner that outputs a recognition result 75 identifying the state of the substrate 101 based on the reflected image of the detection light L1 in the captured images 71 and 72. The learned model 62 is pre-generated by the machine learning unit 61 of the control unit 60 and stored in the storage device of the control unit 60.

[0043] <Generation of the Model After Learning>

[0044] The control unit 60 generates a learned model 62 in the machine learning unit 61 based on the input learning data, namely the teacher image 80, through machine learning. The learned model 62 is generated by machine learning using a dataset that takes the teacher image 80 as input and outputs teacher data 85. For example, the learned model 62 is learned to take each of the captured images 71 and 72 as input and output the coordinates representing the position of the substrate 101 and the orientation of the substrate 101 represented by the position of the notch 101b as a recognition result 75. In addition, when the substrate 101 has an abnormality such as warping, the shape of the detection light L1 in each of the captured images 71 and 72 becomes a different shape than normal. In the first embodiment, the learned model 62 is generated by machine learning in a way that identifies the abnormality of the substrate 101 based on the reflected image of the detection light L1 in the captured images 71 and 72. After learning, model 62 is trained to take each of the captured images 71 and 72 as input and output a recognition result 75 that represents the probability of whether an anomaly is contained in the substrate 101.

[0045] The teacher images 80 include captured images 71 and 72 obtained by capturing detection light L1 reflected from the main surface 101a of each of multiple types of substrates 101, depending on their position, orientation, or the presence or absence of abnormalities. That is, captured images 71 and 72 obtained by capturing multiple types of substrates 101, including substrates 101 determined to be normal and without abnormalities, and substrates 101 determined to have abnormalities, in multiple types of position and orientation configurations, are collected as multiple types of teacher images 80. Furthermore, output teacher data 85 representing the state of the substrates 101 for each of the multiple types of teacher images 80 is acquired, and each of the multiple types of teacher images 80 and its corresponding output teacher data 85 are stored in the storage device of the control unit 60 in a linked state. The output teacher data 85 includes information indicating the position of the substrate 101, information indicating the orientation of the substrate 101, and information indicating whether an abnormality has occurred on the substrate 101, as information representing the state of the substrate 101.

[0046] Then, the machine learning unit 61 takes multiple teacher images 80 as input teacher data and multiple output teacher data 85 associated with the multiple teacher images 80 as output teacher data, and performs machine learning to generate a learned model 62. The machine learning unit 61 generates the learned model 62 using deep learning machine learning. For example, the machine learning unit 61 generates the learned model 62 using machine learning based on a fully convolutional network (FCN).

[0047] As described above, the control unit 60 obtains recognition results 75 from each of the captured images 71 and 72 by using a learned model 62 generated using machine learning. The control unit 60 detects the state of the substrate 101, for example, by calculating the average of the recognition results 75 obtained based on the captured image 71 and the recognition results 75 obtained based on the captured image 72.

[0048] <Correction of position offset>

[0049] In the first embodiment, the control unit 60 detects the positional offset of the substrate 101 relative to the substrate holding hand 22 based on the detected state of the substrate 101. For example, the detection light L1 of each of the captured images 71 and 72 is captured based on the control amount of the robot 20, and is performed when the substrate holding hand 22 is in a predetermined position. Therefore, by detecting information representing the position of the substrate 101 as the state of the substrate 101 based on each of the captured images 71 and 72, the magnitude and direction of the positional offset of the substrate 101 held relative to the position of the substrate holding hand 22 of the robot 20 can be detected. The control unit 60 obtains the average of the magnitude and direction of the positional offset detected based on the captured image 71 and the magnitude and direction of the positional offset detected based on the captured image 72 as the magnitude and direction of the detected positional offset, and corrects the handling action of the robot 20 based on the obtained magnitude and direction of the positional offset in order to correct the positional offset of the substrate 101.

[0050] As an example, when images 71 and 72 are acquired by photographing the substrate 101 removed from FOUP 110, and when it is detected based on images 71 and 72 that the substrate 101 has shifted by +2 cm in the X direction and -1 cm in the Y direction relative to the substrate holding hand 22, the control unit 60 corrects the handling action of the substrate 101 so that the substrate 101 is placed at a position shifted by -2 cm in the X direction and +1 cm in the Y direction from the taught position of the substrate mounting part 30a set as the aligner 30. Specifically, the control unit 60 corrects the handling action of the robot 20 by correcting the drive signals output to drive units such as servo motors that serve as drive sources for driving the robotic arm 21 and the substrate holding hand 22.

[0051] <Alignment aid>

[0052] Furthermore, in the first embodiment, the control unit 60 controls the operation of the aligner 30 using the state of the substrate 101 detected based on the captured images 71 and 72. Specifically, the control unit 60 detects the orientation of the substrate 101 held in the substrate holding hand 22 based on the detected state of the substrate 101. The control unit 60 detects the notch 101b of the substrate 101 based on the recognition result 75 obtained from the learned model 62 as the state of the substrate 101 detected in each of the captured images 71 and 72. For example, the control unit 60 obtains the average position of the notch 101b detected based on the captured image 71 and the notch 101b detected based on the captured image 72 as the position of the detected notch 101b. Then, the control unit 60 performs the operation of adjusting the orientation of the substrate 101 by the aligner 30. The control unit 60 rotates the substrate 101 along the horizontal plane based on the position of the notch 101b detected using the captured images 71 and 72, in a manner that the detection target is a portion of the outer periphery of the substrate 101 including the detected notch 101b, rather than the entire circumference of the outer periphery of the substrate 101. That is, after the control unit 60 roughly detects the orientation of the substrate 101 by detecting its state based on the captured images 71 and 72, the alignment unit 30 detects the orientation of the substrate 101 in detail. In other words, the control unit 60 pre-detects the state of the substrate 101 based on the captured images 71 and 72 to assist the operation of the alignment unit 30.

[0053] <Detection of Substrate Abnormalities>

[0054] Furthermore, in the first embodiment, the control unit 60 detects any abnormalities in the substrate 101 removed from the FOUP 110 based on images 71 and 72 of the detection light L1 captured during the handling action of the robot 20 at the moment the substrate 101 is removed from the FOUP 110. If the likelihood of whether the substrate 101 contains an abnormality, included in the recognition result 75 (output as the inference result based on the learned model 62), is greater than a predetermined threshold, the control unit 60 determines that an abnormality has been detected in the substrate 101. For example, if the likelihood included in the recognition result 75 is a value between 0 and 1, 0.8 is set as a predetermined threshold to determine the detection of an abnormality in the substrate 101. For example, the control unit 60 determines the detection of an abnormality in the substrate 101 based on the average of the likelihood included in the recognition result 75 obtained based on the captured image 71 and the likelihood included in the recognition result 75 obtained based on the captured image 72.

[0055] If an abnormality is detected in substrate 101, the control unit 60 outputs abnormality detection information indicating that an abnormality has been detected in substrate 101 to an external device. For example, the control unit 60 outputs the abnormality detection information to a higher-level control device that performs overall control of the substrate processing system 400. As an example, if an abnormality is detected in substrate 101, the control unit 60, based on a signal from the higher-level control device, uses the robot 20 to reposition the abnormal substrate 101 into FOUP 110. If no signal is received from the higher-level control device instructing the substrate 101 to be repositioned into FOUP 110, the control unit 60 stops the handling of substrate 101 and awaits further processing. Alternatively, information indicating that an abnormality detection has been detected can be displayed on a display device, thereby reporting that an abnormality has been detected in substrate 101. For example, when processing the substrates 101 in the processing unit 401 by sequentially transporting each of the multiple substrates 101 stored in the FOUP 110 to the processing unit 401, the control unit 60 controls the transporting action of the robot 20 to transport the substrates 101 one by one to the loading and locking unit 403 via the alignment unit 30. During the transport of each of the multiple substrates 101, the control unit 60 acquires images 71 and 72, and uses a pre-generated and stored learned model 62 to detect anomalies in the acquired images 71 and 72 based on the acquired images 71 and 72. When an anomaly is detected in one of the multiple substrates 101, the control unit 60 outputs anomaly detection information, and based on a signal from the upper control unit, returns the substrate 101 with the detected anomaly to its original position in the FOUP 110 and re-stores the substrate 101. Then, the control unit 60 performs anomaly detection on the next substrate 101 after the one with the detected anomaly and transports it to the loading and locking unit 403.

[0056] (Substrate inspection method)

[0057] Next, refer to Figure 8 The control processing of the substrate inspection method based on the substrate handling robot system 100 of the first embodiment will be described. The control processing of the substrate inspection method of the first embodiment is executed by the control unit 60 of the substrate handling robot system 100.

[0058] First, in step S1, a learned model 62 is generated. Specifically, a teacher image 80 and output teacher data 85 are acquired. Then, using the teacher image 80 and output teacher data 85 as a dataset, machine learning is used to generate the learned model 62, so as to detect the state of the substrate 101 based on the captured images 71 and 72. The generated learned model 62 is stored in the storage device of the control unit 60.

[0059] Next, in step S2, the transfer operation of the substrate 101 is started. Specifically, the operation of the robot 20 is controlled so that the substrate 101 disposed on the FOUP 110 is held by the substrate holding hand 22 of the robot 20.

[0060] Next, in step S3, substrate 101 is removed from FOUP 110. With substrate 101 held in substrate holding hand 22, one substrate 101 before processing is transferred from FOUP 110 into transfer chamber 10.

[0061] Next, in step S4, the imaging unit 40 captures images of the substrate 101 held on the robot 20 and moving relative to it along a predetermined moving direction, thereby obtaining images 71 and 72 of the detection light L1 reflected on the main surface 101a of the substrate 101. Specifically, the irradiation unit 50 irradiates the main surface 101a of the substrate 101 with the detection light L1, and the imaging unit 40 captures images of the detection light L1 reflected on the main surface 101a, thereby obtaining images 71 and 72 of the detection light L1, respectively.

[0062] Next, in step S5, the state of the substrate 101 is detected based on each of the captured images 71 and 72 of the detection light L1 by using the learned model 62 generated through machine learning. For example, the position of the substrate 101 relative to the substrate holding hand 22, the orientation of the substrate 101, and the state of the substrate 101 in case of any abnormalities are detected.

[0063] Then, in step S6, it is determined whether an abnormality is detected in the substrate 101 based on the detected state of the substrate 101. If it is determined that an abnormality is detected in the substrate 101, proceed to step S7. If it is not determined that an abnormality is detected in the substrate 101, proceed to step S8.

[0064] In step S7, as an abnormality is detected in the substrate 101, abnormality detection information indicating that an abnormality has been detected in the substrate 101 is output. For example, if a signal indicating that the abnormal substrate 101 should be placed back into the FOUP 110 is received, the robot 20 is controlled to place the substrate 101 back into the FOUP 110 without processing it. Then, the control processing of the substrate detection method ends.

[0065] In step S8, the positional offset of the substrate 101 in the substrate holder 22 is corrected. Specifically, assuming no abnormality is detected on the substrate 101, the substrate 101 transport operation continues. In step S5, the positional offset of the substrate 101 relative to the substrate holder 22 is detected based on the position of the substrate 101 as detected by the state of the substrate 101. Then, the transport operation of the substrate 101 relative to the alignment device 30 performed by the robot 20 is corrected based on the detected positional offset of the substrate 101.

[0066] Then, in step S9, the orientation of the substrate 101 based on the aligner 30 is adjusted to a predetermined direction. Specifically, the position of the notch 101b on the outer periphery of the substrate 101 is detected based on the state of the substrate 101 detected in step S5. Then, the detection range for detecting the orientation of the substrate 101 in the aligner 30 is set based on the detected position of the notch 101b. Then, the notch 101b is detected within the set detection range using the aligner 30 without using images 71 and 72, and the rotational position of the substrate 101 is adjusted so that the substrate 101 is oriented in the predetermined direction.

[0067] Then, in step S10, in order to perform processing based on the processing device 401, the robot 20 transports the substrate 101 from the alignment device 30 to the loading locking unit 403. Then, the control processing of the substrate detection method ends.

[0068] Furthermore, when the control processing of the substrate detection method is continuously performed on multiple substrates 101, the learning model 62 in step S1 is generated only once, and the control processing of steps S2 to S10 is repeatedly performed on each substrate 101. Alternatively, the control processing in steps S6 to S8 can be performed at any time.

[0069] [Effects of the first embodiment]

[0070] As described above, the substrate handling robot system 100 irradiates the main surface 101a of the substrate 101 held on the robot 20 with a detection light L1 having a predetermined shape. Then, the state of the substrate 101 is detected based on the captured images 71 and 72 of the detection light L1 captured by the imaging unit 40. Since the robot 20, which handles the substrate 101, holds the photographed substrate 101, the position of the substrate 101 can be easily adjusted by the movement of the robot 20. Therefore, the substrate 101 can be easily positioned in a position suitable for detecting the state of the substrate 101 by the movement of the robot 20, thus enabling higher precision detection of the state of the substrate 101.

[0071] Furthermore, in the first embodiment, in order to detect the state of the substrate 101 during the transport operation of the substrate 101 by the robot 20, an image of the substrate 101 held by the robot 20 is captured. This differs from situations where the substrate 101 is placed in a dedicated mounting section separate from the robot 20 for image capture, or where the substrate 101 is moved by a translation mechanism or other moving mechanism separate from the robot 20. Therefore, it is possible to suppress the complication of the device structure in order to detect the state of the substrate 101 with higher accuracy. Additionally, by capturing an image of the substrate 101 for detecting its state during the transport operation of the substrate 101, the workload is reduced compared to situations where the detection of the state of the substrate 101 is performed separately from the transport operation.

[0072] In addition, in the first embodiment, the irradiation unit 50 irradiates the main surface 101a of the substrate 101 with detection light L1, as at least one of its position and angle changes relative to the irradiation unit 50. The imaging unit 40 captures multiple images of the detection light L1 reflected from the main surface 101a of the substrate 101, as the position and angle of the substrate 101 relative to the irradiation unit 50 change. The control unit 60 detects the state of the substrate 101 based on each of an image 71 and an image 72 generated by capturing multiple images of the reflected detection light L1 using the multiple images captured by the imaging unit 40. Thus, even when the detection light L1 is irradiated only on a portion of the main surface 101a of the substrate 101, by performing multiple images as the position and angle of the substrate 101 change, it is possible to capture a reflected image of the detection light L1 reflected from the entire main surface 101a of the substrate 101. Therefore, even when the detection light L1 is irradiated only on a portion of the main surface 101a of the substrate 101, the state of the entire substrate 101 can be detected with high precision.

[0073] Furthermore, in the first embodiment, when the detection light L1 is captured by the imaging unit 40, the robot 20 changes at least one of the position and angle of the substrate 101 relative to the irradiation unit 50. Therefore, by changing at least one of the position and angle of the substrate 101 relative to the imaging unit 40 using the robot 20, it is possible to adjust at least one of the position and angle of the substrate 101 relative to the imaging unit 40 with high precision, and the imaging unit 40 captures images of the substrate 101. As a result, the state of the substrate 101 can be detected with even higher precision.

[0074] Furthermore, in the first embodiment, the irradiation unit 50 irradiates a linear detection light L1 in a direction intersecting the moving direction that changes the relative position of the substrate 101 with respect to the irradiation unit 50. As a result, the linear detection light L1 can be reflected from the main surface 101a of the substrate 101, thus allowing the state of the substrate 101 in a continuous region from one end to the other to be acquired based on the reflected image of the linear detection light L1. Therefore, the state of the substrate 101 can be further detected with high precision based on the captured images 71 and 72 obtained by capturing the detection light L1 reflected from the main surface 101a of the substrate 101.

[0075] Furthermore, in the first embodiment, when the imaging unit 40 captures the detection light L1, the robot 20 moves the substrate 101 along multiple different movement directions. The control unit 60 detects the state of the substrate 101 based on the captured images 71 and 72 of each movement of the substrate 101 along the multiple movement directions of the robot 20. Thus, by acquiring captured images 71 and 72 for each movement of the substrate 101 along the multiple different movement directions, it is possible to acquire multiple captured images 71 and 72 of the substrate 101 with different movement directions during imaging. Therefore, the state of the substrate 101 can be detected based on each of the multiple types of captured images 71 and 72 with different movement directions of the substrate 101, thereby enabling further high-precision detection of the state of the substrate 101.

[0076] Furthermore, in the first embodiment, the control unit 60 detects the state of the substrate 101 based on the captured images 71 and 72, including at least one of the shape, position, orientation, and abnormalities generated in the substrate 101. Therefore, the state of the substrate 101, including at least one of the shape, position, orientation, and abnormalities generated in the substrate 101, can be detected with higher precision based on the captured images 71 and 72.

[0077] Furthermore, in the first embodiment, the control unit 60 detects the notch 101b or orientation plane indicating the orientation of the substrate 101 based on the state of the substrate 101 detected in each of the captured images 71 and 72. Therefore, the notch 101b or orientation plane can be detected based on each of the captured images 71 and 72, and thus the orientation of the substrate 101 can be easily detected based on each of the captured images 71 and 72.

[0078] Furthermore, in the first embodiment, the control unit 60 uses a learned model 62 generated through machine learning to detect the state of the substrate 101 based on the captured images 71 and 72 taken by the imaging unit 40. Here, in machine learning, by learning the states of multiple types of substrates 101, the learned model 62 can be learned in a way that detects the state of substrates 101 with a different shape than the learned state. For example, when using the machine learning-based learned model 62 to detect warping anomalies of the substrate 101, by using the learned model 62, which has learned warping at multiple locations and in multiple directions as anomalies, warping at locations or in directions different from the learned warping can be detected with high precision. Therefore, by using the learned model 62 generated through machine learning to detect the state of the substrate 101 based on each of the captured images 71 and 72, the states of multiple types of substrates 101 can be detected with higher precision.

[0079] In the first embodiment, a robot 20 is disposed in a transport chamber 10, which transports the substrate 101 relative to a FOUP 110, which serves as a substrate storage container for the substrate 101. An irradiation unit 50 is disposed above the robot 20 within the transport chamber 10 and irradiates a detection light L1 onto the main surface 101a of the substrate 101. An imaging unit 40 is disposed above the robot 20 within the transport chamber 10 and captures the detection light L1 reflected from the main surface 101a of the substrate 101. A control unit 60 detects the state of the substrate 101 based on images 71 and 72 captured within the transport chamber 10. Thus, the robot 20 can perform the transport of the substrate 101 relative to the FOUP 110, and the state of the substrate 101 can be detected based on the captured images 71 and 72. Therefore, compared to separately performing the transport of the substrate 101 relative to the FOUP 110 and detecting the state of the substrate 101, the increased operation time required for detecting the state of the substrate 101 can be suppressed.

[0080] Furthermore, in the first embodiment, the substrate handling robot system 100 includes an alignment device 30, which serves as an adjustment device for adjusting the orientation of the substrate 101 to a predetermined direction. The control unit 60 controls the operation of the alignment device 30 using the state of the substrate 101 detected based on captured images 71 and 72. Therefore, the state of the substrate 101 can be detected in advance based on captured images 71 and 72, and thus the orientation of the substrate 101 can be adjusted using the alignment device 30 with the approximate orientation of the substrate 101 detected in advance. Therefore, the time required to adjust the orientation of the substrate 101 using the alignment device 30 can be shortened.

[0081] Furthermore, in the first embodiment, the robot 20 includes a substrate holding hand 22 for holding the substrate 101. The control unit 60 also functions as a robot control unit for controlling the transporting actions of the robot 20. The control unit 60 detects at least one offset of the position and orientation of the substrate 101 relative to the substrate holding hand 22 based on the detected state of the substrate 101, and corrects the detected offset of at least one offset of the position and orientation of the substrate 101 during the transporting actions of the robot 20. Thus, by detecting the state of the substrate 101 based on each of the captured images 71 and 72, the transporting actions of the robot 20 can be corrected, thereby enabling the execution of transporting actions based on the robot 20 with higher precision.

[0082] [Second Implementation]

[0083] Next, refer to Figure 9 The structure of the substrate handling robot system 200 based on the second embodiment will be described. In the second embodiment, the substrate handling robot system 200 differs from the substrate handling robot system 100 of the first embodiment, which includes an irradiation unit 50 that irradiates one detection light L1, and includes an irradiation unit 250 that irradiates two detection lights L11 and detection light L12. Furthermore, the same reference numerals are used for structures identical to those in the first embodiment, and descriptions are omitted.

[0084] like Figure 9 As shown, the substrate handling robot system 200 based on the second embodiment includes an illumination unit 250 and a control unit 260. The illumination unit 250 illuminates the main surface 101a of the substrate 101 with two different linear detection lights L11 and L12. For example, the illumination unit 250 illuminates the main surface 101a of the substrate 101 with two detection lights L11 and L12 that extend in parallel directions and have the same shape, color, thickness, etc. Furthermore, the detection lights L11 and L12 are examples of "first detection light" and "second detection light," respectively.

[0085] Similar to the imaging of detection light L1 in the first embodiment, the imaging unit 40, along with the movement of the substrate 101 based on the robot 20, repeatedly captures the detection light L11 and detection light L12 reflected on the main surface 101a of the substrate 101. The control unit 260, by capturing the multiple images of the reflected detection light L11 and multiple reflected detection light L12 using the multiple images from the imaging unit 40, thereby acquiring a captured image, similar to the captured image 71 and captured image 72 in the first embodiment. That is, the captured image in the second embodiment includes both detection light L11 and detection light L12. Then, similar to the captured image 71 and captured image 72 in the first embodiment, the control unit 260 detects the state of the substrate 101 based on the acquired captured image.

[0086] For example, in the second embodiment, the control unit 260 sets the movement speed of the robot 20 when the imaging unit 40 is taking pictures to twice the movement speed of the robot 20 in the first embodiment. The control unit 260 captures both the detection light L11 and the detection light L12 in one captured image. That is, when the control unit 260 sets the movement speed of the robot 20 to twice the speed, both the detection light L11 and the detection light L12 are also irradiated on the main surface 101a of a substrate 101, and the imaging unit 40 captures the reflected detection light L11 and the detection light L12, thereby making the total number of detection light L11 and detection light L12 contained in one captured image equal to the number of detection light L1 in the first embodiment. The other structures of the second embodiment are the same as those of the first embodiment.

[0087] [Effects of the second embodiment]

[0088] In the substrate handling robot system 200 based on the second embodiment, as described above, the irradiation unit 250 irradiates a detection light including a detection light L11 as a first detection light and a detection light L12 as a second detection light, which are different from each other. The control unit 260 detects the state of the substrate 101 based on an image including the detection light L11 and the detection light L12 generated by multiple shots taken by the imaging unit 40. Thus, by taking a single shot, both the detection light L11 and the detection light L12 reflected on the main surface 101a of the substrate 101 can be captured, thereby reducing the number of shots required to capture the detection light L11 and L12 reflected on the entire main surface 101a of the substrate 101 compared to the case where only one detection light is reflected. Therefore, the time required to detect the state of the substrate 101 can be shortened. Furthermore, other effects based on the second embodiment are the same as those in the first embodiment.

[0089] [Third Implementation]

[0090] Next, refer to Figure 10 The structure of the substrate handling robot system 300 based on the third embodiment will be described. In the substrate handling robot system 300 based on the third embodiment, the robot 320 moves the held substrate 101 by changing the angle relative to the horizontal plane. Furthermore, the same reference numerals are used for the same structures as in the first embodiment described above, and descriptions are omitted.

[0091] like Figure 10As shown, the substrate handling robot system 300 based on the third embodiment includes a robot 320, an illumination unit 350, and a control unit 360. The robot 320 includes a robotic arm 321 and a substrate holding hand 322. The illumination unit 350 illuminates the substrate 101 with a linear detection light L21 along the YZ plane. That is, the illumination unit 350 illuminates the substrate 101 with a linear detection light L21 extending along the Y direction on the main surface 101a of the substrate 101. In the third embodiment, the illumination unit 350 illuminates the substrate 101's main surface 101a, whose angle is relatively changed, with the detection light L21. When the imaging unit 40 captures the detection light L21, the robot 320 changes the angle of the substrate 101 relative to the illumination unit 350. The robot 320 has a rotation drive mechanism that rotates the substrate holding hand 322 along the θ direction with the direction from the base end of the substrate holding hand 322 toward the front end as the rotation axis. For example, at the front end of the horizontal multi-joint robotic arm 321, a rotation drive mechanism such as a servo motor is provided to rotate the substrate holding hand 322, which serves as an end effector. The irradiation unit 350 irradiates a linear detection light L21 along the rotation axis direction of the rotational movement of the substrate 101 relative to the irradiation unit 350 during its relative angle change. "Along the rotation axis direction" includes a direction parallel to the rotation axis and a direction inclined from the direction parallel to the rotation axis. In other words, the irradiation unit 350 irradiates a linear detection light L21 extending in a direction not orthogonal to the rotation axis of the substrate holding hand 322. Specifically, the detection light L21 extends on the main surface 101a of the substrate 101 in a direction parallel to the rotation axis of the substrate holding hand 322's rotational movement. In the third embodiment, similar to the first embodiment, the state of the substrate 101 is detected based on an image captured by the detection light L21 reflected on the main surface 101a of the substrate 101. When the imaging unit 40 captures the detection light L21, the robot 320 moves the substrate 101 along a predetermined movement direction by changing the angle of the substrate 101 held in the substrate holding hand 322 relative to the horizontal plane. That is, in the third embodiment, the θ direction is an example of a predetermined movement direction.

[0092] To detect the state of substrate 101, control unit 360 uses robot 320 to transport substrate 101 to a predetermined position for imaging unit 40. Then, without changing the position of substrate 101 in the XY plane, control unit 360 tilts substrate 101 along the θ direction by changing its angle relative to the horizontal plane, thereby rotating substrate 101. As an example, control unit 360 rotates substrate 101 along the θ direction within a range of -5 degrees to +5 degrees, based on the state of the main surface 101a of substrate 101 along the horizontal plane. Imaging unit 40, accompanying the change in the relative angle of substrate 101 with irradiation unit 350 (i.e., the movement of substrate 101 along the θ direction), repeatedly captures the detection light L21 reflected from the main surface 101a of substrate 101. Similar to the first embodiment, control unit 360 detects the state of substrate 101 based on a single image generated by capturing multiple detection lights L21 reflected through multiple imaging operations by imaging unit 40. The detection of the state of the substrate 101 based on the captured image of the detection light L21 is the same as the detection of the state of the substrate 101 based on the captured images 71 and 72 of the first embodiment. Other structures of the third embodiment are the same as those of the first embodiment. For example, the control unit 360 changes the orientation of the rotation axis by changing the orientation of the substrate holding hand 322, thereby generating multiple captured images. Then, the state of the substrate 101 is detected based on the multiple captured images with different orientations of the rotation axis.

[0093] [Effects of the Third Implementation]

[0094] In the third embodiment, as described above, the irradiation unit 350 irradiates a linear detection light L21 along the rotation axis direction of the rotational action of the substrate 101, which changes its relative angle with respect to the irradiation unit 350. As a result, the linear detection light L21 can be reflected from the main surface 101a of the substrate 101, and thus the state of the substrate 101 in a continuous region from one end to the other can be obtained based on the reflected image of the linear detection light L21. Therefore, the state of the substrate 101 can be further detected with high precision based on the captured image obtained by capturing the detection light L21 reflected from the main surface 101a of the substrate 101.

[0095] Furthermore, in the third embodiment, as described above, when the imaging unit 40 captures the detection light L21, the robot 320 changes the angle of the substrate 101 relative to the horizontal plane. Therefore, by changing the angle of the substrate 101 relative to the horizontal plane, the illumination angle of the detection light L21 illuminating the substrate 101 can be changed three-dimensionally. Thus, detection light L21 from more diverse illumination directions can be reflected on the main surface 101a of the substrate 101, and the state of the substrate 101 can be detected based on the captured image obtained by capturing the reflected detection light L21, thereby enabling further high-precision detection of the state of the substrate 101. In addition, other effects of the third embodiment are the same as those of the first embodiment.

[0096] [Variation Example]

[0097] Furthermore, the embodiments disclosed herein should be considered illustrative rather than restrictive in all respects. The scope of this disclosure is not a description of the above embodiments, but is shown by the scope of protection claimed in this application, and also includes the meaning equivalent to the scope of protection claimed in this application and all modifications (variations) within that scope.

[0098] For example, in the first, second, and third embodiments described above, an example is shown of detecting the state of the substrate 101 based on a single image generated by capturing multiple reflected linear detection lights L1, L11, L12, L21 using the imaging unit 40 through multiple captures. However, this disclosure is not limited to this. In this disclosure, the state of the substrate can also be detected based on an image of the detection light captured in a single capture. For example, when the detection light is a grid of multiple point light sources instead of lines, the state of the substrate is detected by capturing the detection light reflected across the entire main surface of the substrate in a single capture. Furthermore, the shape of the detection light irradiating the main surface of the substrate is not limited to linear or grid-like shapes, but can also have a predetermined geometric shape such as concentric circles. In addition, the irradiation unit can be irradiated with detection light having a predetermined color and shape using a projector device or the like. Furthermore, an image can be captured as a dynamic image by multiple captures by the imaging unit, and the state of the substrate can be detected based on the image as a dynamic image.

[0099] Furthermore, in the second embodiment described above, an example was shown where the state of the substrate 101 was detected using an image captured by the irradiation unit 250 based on a detection light L11 (a first detection light) and a detection light L12 (a second detection light) that are parallel to each other and have the same shape. However, this disclosure is not limited to this. In this disclosure, when two different detection lights, namely the first detection light and the second detection light, are irradiated, the first detection light and the second detection light can have different shapes and different colors. Additionally, when both the first and second detection lights are linear, irradiation can be performed with the first and second detection lights intersecting each other. Furthermore, three or more different linear detection lights can be irradiated. Furthermore, when irradiating multiple detection lights, multiple irradiation units can be individually configured.

[0100] Furthermore, in the first, second, and third embodiments described above, examples were shown of detecting the position, orientation, and abnormalities of the substrate 101 as the state of the substrate 101 based on images captured by detection lights L1, L11, L12, and L21 using the learned model 62. However, this disclosure is not limited to this. In this disclosure, the state of the substrate detected based on the captured images can also be at least one of the substrate's shape, position, orientation, and abnormalities. That is, the shape of the substrate can also be detected as the state of the substrate. For example, the shape of the substrate's outline can be detected based on the captured images, and the coordinates of the substrate's center position and the notch or orientation plane can be detected based on the detected substrate shape, thereby detecting the position and orientation of the substrate. In addition, the abnormalities of the substrate detected as the state of the substrate can also be any one or more of substrate warping, deformation, notches, cracks, and damage. For example, only substrate warping can be detected as an abnormality. Alternatively, the state of the substrate can be detected without using the learned model. For example, based on the captured images obtained by repeatedly shooting linear detection lights accompanying the movement of a substrate, the state of the substrate is detected by using a predetermined algorithm other than the learned model to detect linear structures in the captured images. In this case, for example, the linear detection light can also be detected by extracting the center position of the linear detection light in the width direction contained in the captured image.

[0101] Furthermore, in the first, second, and third embodiments described above, examples were shown where abnormalities in the substrate 101 were detected based on images 71 and 72 taken at the moment the substrate 101 was removed from the FOUP 110, which serves as a substrate storage container, as the substrate 101 was moved by the robots 20 and 320, and the substrate 101 with detected abnormalities was returned to the FOUP 110. However, this disclosure is not limited to this. In this disclosure, the state of the substrate can also be detected based on images obtained by capturing images of the substrate in a stopped state without being moved by the robot. For example, the irradiation unit can be operated by changing at least one of the irradiation direction or position of the irradiation unit, thereby moving the substrate held by the robot relative to the irradiation unit. Abnormalities in the substrate can also be detected based on images obtained by capturing images of the substrate moved to the predetermined detection position by the robot at a predetermined detection position. In addition, when images are captured by the imaging unit, at least one of the relative position and angle of the substrate relative to the irradiation unit can be kept unchanged. For example, detection light irradiated and reflected from the irradiation unit in a stationary state relative to the stationary substrate can be captured by the imaging unit while the substrate is held by the robot. Alternatively, the imaging unit can be moved. Furthermore, the state of the substrate can be detected by capturing images taken when the substrate, after processing by the processing device, is placed in the substrate storage container. Alternatively, substrates with detected abnormalities may not be returned to the substrate storage container. For example, substrates with detected abnormalities may be placed in a different container or placement position than the substrate storage container from which they were removed. In the event of a substrate abnormality, the imaging unit can be used to capture another image, thereby re-performing the process of detecting the state of the substrate using the learned model. When capturing another image, the position, orientation, or direction of movement of the substrate can be changed from the first image capture. Additionally, a signal indicating the detected state of the substrate can be output to the processing device. Furthermore, information indicating the detected state of the substrate can be announced via a notification unit, such as a display device, located on the front surface of the transport chamber, through the display of text information and the output of audio information.

[0102] Furthermore, in the first and second embodiments described above, examples were shown of acquiring an image 71 while the substrate 101 moves along a predetermined moving direction, and acquiring an image 72 while the substrate 101 moves along a separate moving direction different from the predetermined moving direction in which the image 71 was acquired. However, this disclosure is not limited to this. In this disclosure, the state of the substrate can be detected based on an image captured while the substrate moves along one moving direction, and images can be acquired in three or more mutually different moving directions. In addition, when acquiring multiple images, multiple images can be acquired while the substrate moves in a common moving direction. Furthermore, images can be captured when the relative angles between the substrate and the irradiation unit are different.

[0103] Furthermore, in the third embodiment described above, an example was shown of detecting the state of the substrate 101 by repeatedly capturing images obtained from the detection light L21, based on the rotational movement of the substrate 101 relative to the horizontal plane along the θ direction while maintaining its position in the XY plane. However, this disclosure is not limited to this. In this disclosure, multiple images with different angles of the substrate can be obtained by moving the substrate along the horizontal plane when the angles relative to the horizontal plane are different. In this case, the movement directions of the substrate along the horizontal plane can also be different. That is, both the relative position and angle of the substrate relative to the irradiation unit can be changed. In this case, both the relative position and angle of the substrate relative to the irradiation unit can be changed by a robot. The detection light irradiated by the irradiation unit can be a linear detection light along a direction intersecting the movement direction of the substrate relative to the irradiation unit during the change in its relative position, or a linear detection light along the rotation axis direction of the rotational movement of the substrate relative to the irradiation unit during the change in its relative angle.

[0104] Furthermore, in the first, second, and third embodiments described above, examples were shown of detecting the notch 101b indicating the orientation of the substrate 101 by detecting the state of the substrate 101, but this disclosure is not limited to this. In this disclosure, the orientation plane disposed on the substrate can also be detected by detecting the state of the substrate.

[0105] Furthermore, in the first, second, and third embodiments described above, examples were shown of correcting the handling actions of robots 20 and 320 based on the detected state of substrate 101 to correct at least one of the offsets in position and orientation of the detected substrate 101, and of controlling the operation of the aligner 30, which serves as an adjustment device, based on the detected state of substrate 101. However, this disclosure is not limited to this. In this disclosure, the handling actions of robots can also be corrected based on the detected state of substrate to correct at least one of the offsets in position and orientation of the substrate. That is, the handling actions of robots can also be corrected based on the detected state of substrate to correct the offset in orientation of the substrate. For example, if the orientation of the substrate held by the substrate holder hand is different from a predetermined orientation, the robot's actions can be corrected by moving the substrate in a predetermined orientation based on the detected state of substrate. In addition, when the operation of the aligner, which serves as an adjustment device, is controlled based on the detected state of substrate, the adjustment device may only perform position adjustment in the rotational direction of the substrate without detecting the orientation of the substrate. Furthermore, when controlling the operation of the aligner as an adjustment device based on the detected state of the substrate, it is also possible to detect the notch or orientation plane indicating the orientation of the substrate as the state of the substrate based on the image captured by the detection light, and set the rotation direction of the substrate in the aligner according to the detected position of the notch or orientation plane. In this case, the substrate can be rotated in the aligner towards the notch or orientation plane, thus shortening the time required for detecting the orientation of the substrate in the aligner.

[0106] Furthermore, in the first, second, and third embodiments described above, examples of robots 20 and 320 equipped with horizontally multi-jointed robotic arms 21 and 321 were shown, but this disclosure is not limited to this. In this disclosure, a vertically multi-jointed robotic arm may also be used instead of a horizontally multi-jointed one. Additionally, the robot may be equipped with a linear motion mechanism such as a sliding mechanism. Furthermore, the substrate holding hand may be an active hand that uses a vacuum or chuck to fix and hold the substrate. Additionally, while the substrate is fixed, it may be rotated and moved in a manner tilted relative to the horizontal plane, thereby capturing detection light reflected from the main surfaces on both sides of the substrate.

[0107] Furthermore, in the first, second, and third embodiments described above, the substrate handling robot system 100 of this disclosure is shown as an example of an EFEM (Engineering, Procurement, and Equipment) for handling substrates 101 to a processing device 401, but this disclosure is not limited to this. In this disclosure, the substrate handling robot system may also be something other than an EFEM. For example, the substrate handling robot system may also be a storage device or sorter that handles substrates from one substrate storage container to another. Additionally, the substrate handling robot system may also handle substrates that are not handled by a substrate storage container but by a conveyor belt or the like to a substrate storage container or a substrate mounting section. Furthermore, the substrate state detection method of this disclosure can also be applied to systems that handle substrates on the processing device side. That is, when handling substrates between a loading locking section and a processing module section, the substrate state can be detected using the substrate detection method of this disclosure.

[0108] Furthermore, in the first, second, and third embodiments described above, examples were shown where the control units 60, 260, and 360 of the substrate handling robot system 100 performed the generation of the learned model 62, the detection of anomalies in the substrate 101 based on the captured images 71 and 72, and the control of the actions of the robots 20 and 320. However, this disclosure is not limited to these examples. In this disclosure, at least one of the generation of the learned model and the control of the robot's actions may be performed by a control device different from the control unit that performs the detection of anomalies in the substrate based on the captured images. For example, the generation of the learned model may be performed by a control device configured separately from the robot. That is, the learned model may be generated by a control device configured separately from substrate handling robot systems such as EFEM through machine learning. In addition, the control unit that performs the detection of the state of the substrate based on the captured images may also be configured separately from the robot control unit that controls the actions of the robot.

[0109] Furthermore, in the first, second, and third embodiments described above, examples are shown where the imaging unit 40 and the irradiation units 50, 250, and 350 are arranged inside the transport chamber 10, but this disclosure is not limited to this. In this disclosure, any one of the marking member, imaging unit, and irradiation unit may also be arranged outside the transport chamber where the substrate is transported. For example, in the transport chamber of a processing apparatus for processing substrates, in the event of detecting an abnormality in the substrate, the top plate of the transport chamber of the processing apparatus may be made of a transparent component such as glass, and the marking member, imaging unit, and irradiation unit may be arranged above the outside of the transport chamber.

[0110] Furthermore, in the first, second, and third embodiments described above, examples were shown where the irradiation units 50, 250, and 350 irradiate the detection lights L1, L11, L12, and L21 at the timing when the imaging unit 40 captures images; however, this disclosure is not limited to these examples. In this disclosure, the irradiation unit may continuously irradiate with illumination light. Additionally, the irradiation of the illumination light may be controlled by the control unit based on the captured images. For example, the amount and wavelength of the illumination light may be controlled based on the brightness value of the captured images. Furthermore, if an abnormality is detected in the substrate, the abnormality detection of the substrate using the learned model may be performed again by capturing images again while controlling the illumination light to change its amount and wavelength. Additionally, the irradiation unit irradiating the illumination light may be arranged separately from the irradiation unit within the transport chamber.

[0111] Furthermore, in the first, second, and third embodiments described above, examples were shown of irradiation units 50, 250, and 350 equipped with LEDs that illuminate red light as detection lights L1, L11, L12, and L21, but this disclosure is not limited to these. In this disclosure, the irradiation unit may also include a light source device other than an LED. Additionally, the illumination light of the irradiation unit may have a wavelength of a predetermined color other than red. Furthermore, the illumination light may be light with a wavelength other than visible light, such as infrared light. Furthermore, the number of irradiation units may be one, or three or more.

[0112] Furthermore, in the first, second, and third embodiments described above, examples are shown where three FOUPs 110 serving as substrate storage containers are installed in the transport chamber 10 of the substrate handling robot system 100, 200, and 300, and three imaging units 40 and three irradiation units 50, 250, and 350 are each arranged corresponding to the FOUPs 110. However, this disclosure is not limited to this. In this disclosure, the number of substrate storage containers installed in the transport chamber may be two or fewer, or four or more. Alternatively, one imaging unit and one irradiation unit may be configured for multiple substrate storage containers. In this case, the substrates transported by the multiple substrate storage containers are imaged by a common imaging unit. Additionally, the number of imaging units and the number of irradiation units may be different values.

[0113] Furthermore, in the first, second, and third embodiments described above, examples of generating the learned model 62 using machine learning based on deep learning of fully convolutional networks were shown, but this disclosure is not limited to this. In this disclosure, the machine learning used to generate the learned model may not be deep learning. Additionally, the machine learning used to generate the learned model may also be machine learning based on convolutional neural networks other than fully convolutional networks. Furthermore, the machine learning used to generate the learned model may also be machine learning other than convolutional neural networks.

[0114] Furthermore, in the first, second, and third embodiments described above, the substrate 101 under test is shown as an example of a silicon wafer with a disk shape, but this disclosure is not limited to this. In this disclosure, the substrate 101 may also be a component other than silicon. Alternatively, the state of a quadrilateral substrate on which electronic components are mounted may be detected instead of a disk shape. In this case, the state of the substrate, including its orientation and position, may be detected by detecting the edges of the quadrilateral.

[0115] Furthermore, in the first, second, and third embodiments described above, examples were shown where images of detection lights L1, L11, L12, and L21 captured by the imaging unit 40 were used as input, and anomalies in the substrate 101 were detected by using the learned model 62 that outputs recognition results 75. However, this disclosure is not limited to this. In this disclosure, the captured images input to the learned model for detecting substrate anomalies may also undergo image processing such as noise removal, edge enhancement, and contrast adjustment before being input to the learned model.

[0116] The functions of the elements disclosed in this specification can be performed using circuitry or processing circuitry, including general-purpose processors, special-purpose processors, integrated circuits, ASICs (Application Specific Integrated Circuits), existing circuits, and / or combinations thereof, configured or programmed to perform the disclosed functions. Processors include transistors and other circuitry, and are therefore considered processing circuitry or circuitry. In this disclosure, a circuit, unit, or mechanism is hardware that performs or is programmed to perform the listed functions. The hardware can be the hardware disclosed in this specification, or it can be other known hardware programmed or configured to perform the listed functions. Where the hardware is a processor considered a type of circuit, the circuit, mechanism, or unit is a combination of hardware and software used in the configuration of the hardware and / or processor.

[0117] [Way]

[0118] Those skilled in the art will understand that the exemplary embodiments described above are specific examples of the following approaches.

[0119] (Method 1)

[0120] A substrate handling robot system, comprising:

[0121] Robots are used to transport circuit boards.

[0122] The irradiation section irradiates the main surface of the substrate held on the robot with a detection light of a predetermined shape.

[0123] The imaging unit captures the detection light reflected from the main surface of the substrate by photographing the substrate; and

[0124] The control unit detects the state of the substrate based on the image captured by the imaging unit of the detection light.

[0125] (Method 2)

[0126] According to the substrate handling robot system described in Method 1, wherein,

[0127] The aforementioned irradiation unit irradiates the aforementioned main surface of the aforementioned substrate with the detection light, wherein at least one of its position and angle has been changed.

[0128] As the imaging unit changes at least one of the relative position and angle of the substrate with respect to the irradiation unit, it repeatedly captures the detection light reflected from the main surface of the substrate.

[0129] The control unit detects the state of the substrate based on a captured image generated by capturing multiple reflected detection lights through multiple captures performed by the imaging unit.

[0130] (Method 3)

[0131] According to the substrate handling robot system described in Method 2, wherein,

[0132] When the detection light is captured by the imaging unit, the robot changes at least one of the position and angle of the substrate relative to the irradiation unit.

[0133] (Method 4)

[0134] According to the substrate handling robot system described in method 2 or 3, wherein,

[0135] The irradiation unit irradiates at least one of the following: a linear detection light that intersects the direction of movement in the change of the relative position of the substrate with respect to the irradiation unit, and a linear detection light that is along the direction of the rotation axis of the rotational action in the change of the relative angle of the substrate with respect to the irradiation unit.

[0136] (Method 5)

[0137] According to the substrate handling robot system described in Method 3, wherein,

[0138] When the aforementioned detection light is captured by the aforementioned imaging unit, the aforementioned robot moves the aforementioned substrate along multiple mutually different movement directions.

[0139] The control unit detects the state of the substrate based on the captured images taken during the movement of the substrate along the multiple directions of movement of the robot.

[0140] (Method 6)

[0141] According to the substrate handling robot system described in Method 3, wherein,

[0142] When the detection light is captured by the aforementioned imaging unit, the robot changes the angle of the substrate relative to the horizontal plane.

[0143] (Method 7)

[0144] According to any one of methods 1 to 6, the substrate handling robot system described therein,

[0145] The control unit detects the state of the substrate based on the captured images, including the shape, position, orientation of the substrate, and at least one of the abnormalities generated on the substrate.

[0146] (Method 8)

[0147] According to the substrate handling robot system described in Method 7, wherein,

[0148] The control unit detects the state of the substrate based on the captured image to detect the notch or orientation plane indicating the orientation of the substrate.

[0149] (Method 9)

[0150] According to any one of methods 1 to 8, the substrate handling robot system described therein,

[0151] The control unit uses a learned model generated through machine learning to detect the state of the substrate based on the captured image obtained by the imaging unit.

[0152] (Method 10)

[0153] According to any one of methods 2 to 6, the substrate handling robot system described therein,

[0154] The aforementioned irradiation unit irradiates the aforementioned detection light, which includes a first detection light and a second detection light that are different from each other in linear form.

[0155] The control unit detects the state of the substrate based on a captured image containing the first detection light and the second detection light, generated by multiple captures performed by the imaging unit.

[0156] (Method 11)

[0157] According to any one of methods 1 to 10, the substrate handling robot system described therein,

[0158] The robot described above is configured in a transport chamber for transporting the substrate relative to a substrate storage container that holds the substrate.

[0159] The aforementioned irradiation unit is disposed above the robot within the aforementioned transport chamber, and irradiates the aforementioned main surface of the aforementioned substrate with the aforementioned detection light.

[0160] The aforementioned imaging unit is positioned above the robot within the aforementioned transport chamber, and captures the detection light reflected from the aforementioned main surface of the aforementioned substrate.

[0161] The control unit detects the state of the substrate based on the captured images taken in the transport chamber.

[0162] (Method 12)

[0163] According to any one of methods 1 to 11, the substrate handling robot system described therein,

[0164] It also includes an adjustment device that adjusts the orientation of the substrate to a predetermined direction.

[0165] The control unit uses the state of the substrate detected based on the captured image to control the operation of the adjustment device.

[0166] (Method 13)

[0167] According to any one of methods 1 to 12, the substrate handling robot system described therein,

[0168] The robot described above includes a substrate holding hand that holds the substrate.

[0169] The aforementioned control unit also serves as the robot control unit for controlling the robot's transporting actions.

[0170] The control unit detects, based on the detected state of the substrate, an offset of the substrate relative to at least one of the hand's position and orientation.

[0171] Furthermore, during the robot's handling operation, the control unit corrects at least one of the detected offsets in the position and orientation of the substrate.

[0172] (Method 14)

[0173] A substrate inspection method, wherein,

[0174] The detection light with a predetermined shape reflected on the main surface of the substrate is captured by photographing the substrate held by the robot carrying the transport substrate.

[0175] The state of the substrate is detected based on the captured image of the detection light.

Claims

1. A substrate handling robot system, characterized in that, have: Robots are used to transport circuit boards. The irradiation section irradiates the main surface of the substrate held on the robot with a detection light of a predetermined shape; The imaging unit captures the detection light reflected from the main surface of the substrate by photographing the substrate; and The control unit detects the state of the substrate based on the image captured by the imaging unit of the detection light.

2. The substrate handling robot system according to claim 1, characterized in that, The irradiation unit irradiates the main surface of the substrate with the detection light, as at least one of its position and angle changes relative to the substrate. The imaging unit, in conjunction with changes in at least one of the relative position and angle of the substrate relative to the irradiation unit, repeatedly captures the detection light reflected from the main surface of the substrate. The control unit detects the state of the substrate based on a captured image generated by capturing multiple reflected detection lights through multiple captures performed by the imaging unit.

3. The substrate handling robot system according to claim 2, characterized in that, When the detection light is captured by the imaging unit, the robot changes at least one of the position and angle of the substrate relative to the irradiation unit.

4. The substrate handling robot system according to claim 2, characterized in that, The irradiation unit irradiates at least one of the following: a linear detection light that intersects the direction of movement in the change of the relative position of the substrate with respect to the irradiation unit, and a linear detection light that is directed along the axis of rotation in the change of the relative angle of the substrate with respect to the irradiation unit.

5. The substrate handling robot system according to claim 3, characterized in that, When the detection light is captured by the imaging unit, the robot moves the substrate along multiple different directions of movement. The control unit detects the state of the substrate based on the captured images taken during the movement of the substrate along the plurality of movement directions of the robot.

6. The substrate handling robot system according to claim 3, characterized in that, When the detection light is captured by the imaging unit, the robot changes the angle of the substrate relative to the horizontal plane.

7. The substrate handling robot system according to claim 1, characterized in that, The control unit detects the state of the substrate based on the captured image, including the substrate's shape, position, orientation, and at least one of the abnormalities generated on the substrate.

8. The substrate handling robot system according to claim 7, characterized in that, The control unit detects the state of the substrate based on the captured image, thereby detecting a notch or orientation plane indicating the orientation of the substrate.

9. The substrate handling robot system according to claim 1, characterized in that, The control unit uses a learned model generated through machine learning to detect the state of the substrate based on the captured image obtained by the imaging unit.

10. The substrate handling robot system according to claim 2, characterized in that, The irradiation unit irradiates the detection light, which includes a first detection light and a second detection light that are different from each other in linear shape. The control unit detects the state of the substrate based on a captured image generated by multiple captures performed by the imaging unit, including the first detection light and the second detection light.

11. The substrate handling robot system according to claim 1, characterized in that, The robot is configured in a transport chamber for moving the substrate relative to a substrate storage container that holds the substrate. The irradiation unit is positioned above the robot within the transport chamber and irradiates the main surface of the substrate with the detection light. The imaging unit is positioned above the robot within the transport chamber and captures the detection light reflected from the main surface of the substrate. The control unit detects the state of the substrate based on the captured images taken in the transport chamber.

12. The substrate handling robot system according to claim 1, characterized in that, It also includes an adjustment device that adjusts the orientation of the substrate to a predetermined direction. The control unit uses the state of the substrate detected based on the captured image to control the operation of the adjustment device.

13. The substrate handling robot system according to claim 1, characterized in that, The robot includes a substrate-holding hand that holds the substrate. The control unit also serves as the robot control unit for controlling the robot's handling actions. The control unit detects, based on the detected state of the substrate, an offset of the substrate relative to at least one of the position and orientation of the hand. Furthermore, during the robot's handling operation, the control unit corrects at least one of the detected offsets in the position and orientation of the substrate.

14. A substrate inspection method, characterized in that, The detection light of a predetermined shape reflected from the main surface of the substrate is captured by photographing the substrate held by the robot transporting the substrate. The state of the substrate is detected based on the captured image of the detection light.

Images