Substrate abnormality detection method and substrate abnormality detection apparatus

Abstract

It accurately detects abnormalities such as cracks in the circuit board. [Solution] The control unit of the substrate abnormality detection device acquires a video, including the state in which the substrate is detached from the mounting surface of the mounting table by a lifting pin, which is captured by a camera that photographs the inside of the processing container through an observation window provided on the side wall of the processing container. Among the multiple images that make up the video, the evaluation target range is set to the part corresponding to the mounting surface so that the substrate is not included in the image of the pin-up state in which the substrate is detached from the mounting surface. Adaptive binarization processing is applied to the evaluation target range of the grayscale image obtained by representing the pin-up state image in grayscale. If the proportion of white pixels in the binarized image obtained by applying adaptive binarization processing to the evaluation target range is greater than or equal to a predetermined threshold, it is determined that an abnormality has occurred in the substrate.

Description

Technical Field

[0004] , , ,

[0005] , , , ,

[0001] The present disclosure relates to a substrate abnormality detection method and a substrate abnormality detection device.

Background Art

[0002] There is known a plasma processing apparatus that performs plasma processing on a substrate placed on a mounting table inside a processing container. In such a plasma processing apparatus, an abnormality of the substrate after plasma processing, for example, a crack in the substrate, is detected. For example, in the substrate abnormality detection method of Patent Document 1, while the substrate after plasma processing is carried out from the processing container and transferred to a transfer device, a crack in the substrate is detected by a sensor provided near a gate connecting the processing container and the transfer device.

Prior Art Documents

Patent Documents

[0006] According to the technology disclosed herein, abnormalities such as cracks in the substrate can be accurately detected. [Brief explanation of the drawing]

[0007] [Figure 1] This is a block diagram schematically showing the configuration of a substrate abnormality detection system, including an information processing device as a substrate abnormality detection device, according to one embodiment of the technology described herein. [Figure 2] Figure 1 is a schematic cross-sectional view showing the configuration of the film deposition apparatus. [Figure 3] This figure illustrates the image of the inside of the processing container taken by the camera in Figure 2. [Figure 4] Figure 1 is a block diagram illustrating the hardware configuration of the information processing device. [Figure 5]This diagram illustrates images showing pin-up states with reflections of circuit board fragments, shadows, and shielding mesh. [Figure 6] This is a diagram illustrating the adaptive binarization process. [Figure 7] This diagram illustrates the process of applying adaptive binarization to a pin-up grayscale image. [Figure 8] This is a diagram illustrating the binarized image of the evaluation range. [Figure 9] Figure 1 shows a flowchart illustrating the substrate anomaly detection process as a substrate anomaly detection method executed by the information processing device. [Figure 10] Figure 9 is a diagram illustrating an image of the pin-up state that is the target of the substrate abnormality detection process. [Figure 11] This figure illustrates the binarized image of the evaluation target range in the image of Figure 10. [Figure 12] This diagram illustrates the morphological processing applied to the binarized image within the evaluation range. [Figure 13] This diagram illustrates the contraction and expansion processes in the opening and closing operations of morphological processing. [Modes for carrying out the invention]

[0008] Hereinafter, an embodiment of the technology relating to this disclosure will be described with reference to the drawings. Figure 1 is a schematic block diagram showing the configuration of a substrate abnormality detection system 1, which includes an information processing device 40 as a substrate abnormality detection device according to an embodiment of the technology relating to this disclosure. In Figure 1, the substrate abnormality detection system 1 consists of a film deposition apparatus 10 as a substrate processing apparatus, a camera 22, and an information processing device 40. In the substrate abnormality detection system 1, the information processing device 40 performs the substrate abnormality detection processing shown in Figure 9, which will be described later, on the video transmitted from the camera 22, determines whether or not a crack has occurred in the substrate placed on the mounting table 13 (see Figure 2) inside the processing container of the film deposition apparatus 10, and notifies the film deposition apparatus 10 of the result of the determination.

[0009] Figure 2 is a schematic cross-sectional view showing the configuration of the film deposition apparatus 10 shown in Figure 1. The film deposition apparatus 10 in Figure 2 is an inductively coupled plasma processing apparatus. The film deposition apparatus 10 uses plasma generated from a processing gas to perform a film deposition process to form a protective film or the like on a rectangular substrate, for example, a glass substrate G for an FPD (Flat Panel Display) (hereinafter referred to as "substrate G").

[0010] The film deposition apparatus 10 includes a rectangular tubular processing container 11 made of a conductive material, which houses the substrate G. The top of the processing container 11 is airtightly sealed by a window member 12. Inside the processing container 11, a mounting table 13 for placing the substrate G is positioned at the bottom, and the mounting table 13 faces the window member 12. In addition, a processing space U is formed between the mounting table 13 and the window member 12 in the processing container 11. Plasma is generated from the processing gas in the processing space U, as will be described later.

[0011] An electrostatic chuck (not shown) is provided on the upper surface of the mounting table 13. The substrate G placed on the mounting table 13 is held in place by attraction to the mounting table 13 by the electrostatic chuck to which a DC voltage is applied. Lift pins 38 (lifting and lowering pins) (see Figure 3(B)) are also arranged on the mounting table 13.

[0012] The lift pins 38 can protrude from the mounting surface of the mounting table 13 and can raise and lower the substrate G. The lift pins 38 are mainly used when loading and unloading the substrate G into and out of the processing container 11 through the loading / unloading port 19, which will be described later. More specifically, the lift pins 38 are used to lift the substrate G and detach it from the mounting table 13, and the substrate G is transferred to a loading / unloading device (not shown) that is installed adjacent to the processing container 11 and extends into the interior of the processing container 11. Hereafter, the detachment of the substrate G from the mounting table 13 by the lift pins 38 will be referred to as "pin-up". The lift pins 38 are also used when performing static electricity removal on the substrate G that is electrostatically attached to the mounting table 13.

[0013] In the static charge removal process of the substrate G, first, the application of the DC voltage to the electrostatic chuck is stopped after the film formation process is completed. Next, a gas for static charge removal, such as argon (Ar) gas, nitrogen (N2) gas, oxygen (O2) gas, etc., is supplied into the processing chamber 11 to generate plasma for static charge removal. Then, after generating the plasma for static charge removal, the substrate G is pinned up by the lift pins 38 to enhance the static charge removal effect.

[0014] Also, inside the mounting table 13, a temperature control mechanism such as a chiller and a heat transfer gas supply mechanism (both not shown) are provided to control the temperature of the mounted substrate G. The temperature control mechanism adjusts the temperature of the mounting table 13, and the heat transfer gas supply mechanism supplies heat transfer gas between the back surface of the substrate G and the mounting surface of the mounting table 13. The mounting table 13 is installed on the bottom surface of the processing chamber 11 via an insulator frame 14.

[0015] A metal frame 15 is provided at the upper end of the side wall of the processing chamber 11, and a side wall portion 16 is installed on the upper surface of the metal frame 15. The side wall portion 16 supports the top plate 17, and the top plate 17 covers the window member 12 from above. A seal member 18 such as an O-ring is provided between the side wall of the processing chamber 11 and the metal frame 15 to keep the processing space U airtight.

[0016] On one side wall of the processing container 11, there are provided a loading / unloading port 19 for loading and unloading the substrate G into and out of the processing space U, and a gate valve 20 for opening and closing the loading / unloading port 19. Further, on another side wall of the processing container 11, there is provided a monitoring window 21 (observation window) made of an optically transparent member that is airtightly fitted, and outside the processing container 11, there is provided a camera 22 (imaging device) for color photographing a video inside the processing container 11 through the monitoring window 21. The camera 22 may be provided as a part of the film forming apparatus 10, or may be provided separately from the film forming apparatus 10. Further, on the inner side of the processing container 11 of the monitoring window 21, a shielding net 39 is provided. The shielding net 39 blocks the plasma generated in the processing space U to prevent the plasma from reaching the monitoring window 21. The camera 22 transmits the photographed video inside the processing container 11 to the information processing device 40. In the present embodiment, the image inside the processing container 11 photographed by the camera 22 is a video, but it may also be a still image. Further, the video photographed by the camera 22 includes information regarding the mounting surface of the mounting table 13 from the mounting of the substrate G on the mounting table 13 to the pin-up of the substrate G by the lift pins 38.

[0017] The window member 12 has a rectangular shape and is divided into a plurality of divided pieces 23 made of a conductor, for example, metal, and partition members 24 made of an insulator are arranged between adjacent divided pieces 23. The partition member 24 partitions adjacent divided pieces 23 and electrically insulates them from each other.

[0018] In the film forming apparatus 10, the space surrounded by the window member 12, the side wall portion 16, and the top plate 17 constitutes an antenna chamber 25. In the antenna chamber 25, an inductively coupled antenna 26 is arranged so as to face the substrate G on the mounting table 13 through the window member 12. Further, the inductively coupled antenna 26 is formed in a spiral shape so as to span across the regions facing each divided piece 23 as a whole and to circulate along the circumferential direction of the window member 12, but the shape of the inductively coupled antenna 26 is not limited to the spiral shape.

[0019] Each segmented piece 23 has numerous gas holes 27 that open toward the processing space U. A gas supply pipe 28 is connected to each segmented piece 23, and the gas supply pipe 28 is connected to a gas supply device 29. A gas diffusion chamber 30 is also formed inside each segmented piece 23, and the gas supply device 29 introduces the processing gas into the gas diffusion chamber 30 via the gas supply pipe 28. The processing gas introduced into the gas diffusion chamber 30 is supplied to the processing space U through each gas hole 27.

[0020] In the film deposition apparatus 10, a high-frequency power supply 32 is connected to the inductively coupled antenna 26 via a matching unit 31. The high-frequency power supply 32 supplies, for example, 13.56 MHz high-frequency power for plasma generation to the inductively coupled antenna 26. This induces eddy currents that circulate from the upper surface (inductively coupled antenna 26 side) to the lower surface (processing space U side) in each of the segmented pieces 23 constituting the window member 12, and these eddy currents create an induced electric field in the processing space U. This induced electric field then excites the processing gas supplied to the processing space U, generating plasma. In this embodiment, the case in which the window member 12 is made of a conductor such as a metal is described, but the window member 12 may also be made of a dielectric material. In that case, the electric field formed by the inductively coupled antenna 26 passes through the window member 12 and acts directly on the processing gas to generate plasma. Gas diffusion chambers and gas holes are formed, for example, in the support beams that support the dielectric material.

[0021] Furthermore, a high-frequency power supply 34 is connected to the mounting stage 13 via a matching unit 33. The high-frequency power supply 34 supplies, for example, 3.2 MHz bias high-frequency power to the mounting stage 13. This draws various ions from the plasma in the processing space U onto the substrate G, performing a film deposition process to form various films on the substrate G. In addition, the film deposition apparatus 10 may also perform an etching process on the substrate G by drawing various ions onto the substrate G. In this case, the film deposition apparatus 10 functions as an etching apparatus.

[0022] Furthermore, in the film deposition apparatus 10, an exhaust port 35 is formed on the bottom surface of the processing container 11. An exhaust device 36, such as a turbomolecular pump or a dry pump, is connected to this exhaust port 35. When performing the film deposition process, the exhaust device 36 maintains the processing space U at a predetermined pressure lower than atmospheric pressure. The film deposition apparatus 10 is also provided with a control unit 37. The control unit 37 is a control device consisting of a computer having at least a CPU and memory, and the memory stores a recipe (program) for executing a series of substrate processes, such as film deposition and static elimination of the substrate G. The recipe also includes information for performing pin-up of the substrate G. The control unit 37 controls the execution of various processes applied to the substrate G.

[0023] Figure 3 is a diagram illustrating the image of the inside of the processing container 11 captured by the camera 22 in Figure 2. Figures 3(A) and 3(B) each show one of several images that make up the video captured by the camera 22. More precisely, a video is composed of multiple frames, and each frame contains a captured image (captured image), but for the sake of simplicity in the following explanation, we will assume that a video is composed of multiple images without using the concept of "frame".

[0024] As described above, the camera 22 captures images of the inside of the processing container 11 through the monitoring window 21, specifically from the state in which the substrate G is placed on the mounting table 13 to the state in which the substrate G is pinned up by the lift pins 38 (hereinafter simply referred to as the "pinned-up state"), and acquires a video including the captured images in Figure 3(A) and Figure 3(B). The captured image in Figure 3(A), which shows the state in which the substrate G is placed on the mounting table 13, includes the structural members inside the processing container 11, such as the mounting table 13 and the substrate G placed on the mounting table 13. The captured image in Figure 3(B), which shows the pinned-up state, includes the mounting table 13, the substrate G, and the lift pins 38 that lift the substrate G. The camera 22 then transmits the captured video to the information processing device 40. In this embodiment, the camera 22 directly transmits the video to the information processing device 40, but the camera 22 may also transmit the video to the information processing device 40 via the control unit 37 of the film deposition apparatus 10.

[0025] Figure 4 is a block diagram schematically showing the hardware configuration of the information processing device 40 shown in Figure 1. The information processing device 40 consists of, for example, a personal computer. The information processing device 40 includes a CPU 41 (control unit), ROM 42, RAM 43, HDD 44, communication I / F 45, and input / output I / F 46, which are connected to each other via a system bus 47 so that they can communicate with one another.

[0026] The CPU 41 is a microprocessor that comprehensively controls the information processing unit 40. Specifically, the CPU 41 reads programs stored in the ROM 42 and HDD 44 and executes programs loaded onto the RAM 43. The ROM 42 stores programs and other data that the CPU 41 executes. The RAM 43 is used as the main memory of the CPU 41 and is also used as a temporary storage area for programs and data. The HDD 44 stores programs and various data that the CPU 41 executes. In this embodiment, a configuration in which the information processing unit 40 is equipped with an HDD as an example of a storage device is described, but the information processing unit 40 may be equipped with storage devices other than an HDD, such as an SSD.

[0027] Communication I / F 45 is an interface for the information processing device 40 to communicate with external devices. For example, communication I / F 45 receives video footage captured by camera 22 from camera 22. Communication via communication I / F 45 can be achieved using various communication technologies, whether wireless or wired. Output devices such as displays that show various types of information, and input devices for user use such as mice and keyboards are connected to input / output I / F 46.

[0028] Incidentally, in the substrate abnormality detection method described in Patent Document 1 mentioned above, a sensor (not shown) located near the gate valve 20 connecting the processing container 11 and the transport device (not shown) detects cracks in the substrate G. Therefore, even if a crack occurs in the substrate G during the film deposition process inside the processing container 11, it is not possible to detect that an abnormality such as a crack has occurred in the substrate G until the substrate G is removed from the processing container 11. As a result, fragments of the broken substrate G may be scattered not only on the mounting platform 13 of the processing container 11 but also near the gate valve 20, potentially leading to prolonged downtime for fragment removal.

[0029] Therefore, it is conceivable to detect early on whether or not a crack has occurred in the substrate G from a video taken of the inside of the processing container 11, including the mounting table 13. In this case, the CPU 41 of the information processing device 40 acquires a grayscale image representing the pin-up state in grayscale from the video taken by the camera 22. Furthermore, the CPU 41 binarizes the grayscale image of the pin-up state to acquire a binarized image, and at this time, the fragments of the substrate G remaining on the mounting surface of the mounting table 13 are converted to white. Then, if the proportion of white pixels in this binarized image is above a predetermined threshold, the CPU 41 determines that fragments 49 of the substrate G remain on the mounting surface of the mounting table 13, that is, that an abnormality such as a crack has occurred in the substrate G.

[0030] However, as shown in Figure 5, images taken in the pin-up state may include not only the fragments 49 of the substrate G, but also the shadows 50 of the substrate G caused by the emission of plasma, and the shielding net 39 installed in front of the camera that photographs the inside of the processing container 11. In this case, since the camera 22 is not in focus on the shielding net 39, the shielding net 39 appears as a blurred shadow in the images taken in the pin-up state, as shown in Figure 5.

[0031] Furthermore, when binarizing a grayscale image in a pin-up state and converting pixels with pixel values ​​below a threshold to white, the shadows 50 and shielding nets 39 of the substrate G may also be converted to white. Therefore, if the shadows 50 and shielding nets 39 of the substrate G are captured in an image of the mounting table 13, it may be determined that a crack has occurred in the substrate G even if no fragments 49 of the substrate G remain on the mounting surface of the mounting table 13, making it difficult to accurately detect cracks or other abnormalities in the substrate G. In addition, in a grayscale image in a pin-up state, the size of the small fragments 49 and the shielding nets 39 do not differ significantly, so in the binarized image, it is impossible to distinguish between the small fragments 49 converted to white and the shielding nets 39 converted to white. As a result, it is not possible to detect the small fragments 49 remaining on the mounting table 13.

[0032] In contrast, the technology relating to this disclosure applies adaptive binarization processing to a grayscale image obtained by representing a pin-up image in grayscale from a video captured by the camera 22, extracting the contour rather than the entire fragment 49 of the substrate G, and preventing the shadow 50 and shielding net 39 of the substrate G from being converted to white in the binarized image.

[0033] Figure 6 is a diagram illustrating adaptive binarization. In Figure 6, a grayscale image consisting of 5x6 pixels is shown as an example, where 100, 120, 140, 160, 180, and 200 represent the pixel values ​​of each pixel, and pixels shown at the same intensity have the same pixel value.

[0034] Adaptive binarization is a process in which a binarization calculation region is set in an image, which has multiple pixels surrounding the target pixel (hereinafter referred to as the "central pixel") (for example, in Figure 6, eight pixels adjacent to the central pixel in the top, bottom, left, right, and diagonal directions), and the pixel value of the central pixel is binarized using the average value of the pixel values ​​of each pixel in the binarization calculation region as the binarization threshold. Specifically, if the pixel value of the central pixel is equal to or greater than the binarization threshold, the pixel value of the central pixel is converted to 0 (black), and if the pixel value of the central pixel is less than the binarization threshold, the pixel value of the central pixel is converted to 255 (white).

[0035] Here, there are two methods for calculating the binarization threshold: one is to calculate the average value from the pixel values ​​of all pixels in the binarization calculation region including the central pixel, and the other is to calculate the average value from the pixel values ​​of each pixel in the binarization calculation region other than the central pixel. Below, we will explain the case where the binarization threshold is calculated using the latter method. For example, in the case of Figure 6(A), for central pixel A, which has a pixel value of 200 and is surrounded only by pixels with a pixel value of 200, the binarization threshold is 200, so central pixel A is converted to black. For central pixel B, which has a pixel value of 200 and is surrounded by pixels with a pixel value of 200 and pixels with a pixel value of 100, the binarization threshold is 162.5, so central pixel B is converted to black. Also, for central pixel D, which has a pixel value of 100 and is surrounded only by pixels with a pixel value of 100, the binarization threshold is 100, so central pixel D is converted to black. On the other hand, for the central pixel C, which has a pixel value of 100 and is surrounded by pixels with a pixel value of 200 and pixels with a pixel value of 100, the binarization threshold becomes 137.5, so the central pixel C is converted to white.

[0036] Furthermore, for example, in the case of Figure 6(B), for central pixel A, which has a pixel value of 180 and is surrounded by a pixel with a pixel value of 200, a pixel with a pixel value of 180, and a pixel with a pixel value of 160, the binarization threshold is 180, so central pixel A is converted to black. For central pixel B, which has a pixel value of 160 and is surrounded by a pixel with a pixel value of 180, a pixel with a pixel value of 160, and a pixel with a pixel value of 140, the binarization threshold is 160, so central pixel B is converted to black. For central pixel C, which has a pixel value of 140 and is surrounded by a pixel with a pixel value of 160, a pixel with a pixel value of 140, and a pixel with a pixel value of 120, the binarization threshold is 140, so central pixel C is converted to black. For central pixel D, which has a pixel value of 120 and is surrounded by a pixel with a pixel value of 140, a pixel with a pixel value of 120, and a pixel with a pixel value of 100, the binarization threshold is 120, so central pixel D is converted to black.

[0037] In other words, when an adaptive binarization process is applied to a grayscale image, the central pixel whose pixel value abruptly changes to a small value from any of the surrounding pixels in the binarization calculation region is converted to white. On the other hand, the central pixel whose pixel value does not abruptly change to a small value from any of the surrounding pixels in the binarization calculation region, or the central pixel whose pixel value changes gradually without abrupt change, is converted to black.

[0038] Incidentally, as shown in Figure 5, in the image taken in the pin-up state, the brightness and saturation change abruptly at the boundary between the substrate G fragment 49 and the mounting base 13, so in the grayscale image of the pin-up state, the pixel values ​​change abruptly at the boundary between the substrate G fragment 49 and the mounting base 13. On the other hand, as shown in Figure 5, in the image taken in the pin-up state, the brightness and saturation only change gradually in the shadow 50 of the substrate G and the shielding net 39, so in the grayscale image of the pin-up state, the pixel values ​​only change gradually without abrupt changes in the shadow 50 of the substrate G and the shielding net 39. Therefore, when adaptive binarization processing is applied to the grayscale image of the pin-up state, the boundary between the substrate G fragment 49 and the mounting base 13 is converted to white, while the shadow 50 of the substrate G and the shielding net 39 are converted to black. In this embodiment, the boundary between the substrate G fragment 49 and the mounting base 13, i.e., the outline of the substrate G fragment 49, is converted to white and extracted from the pin-up grayscale image, while preventing the substrate G shadow 50 and the shielding net 39 from being converted to white.

[0039] Figure 7 illustrates the process of applying adaptive binarization to a grayscale image in a pin-up state. Note that Figure 7 assumes that although no fragments 49 of the substrate G remain on the mounting surface of the mounting table 13, the shadow 50 of the substrate G and the shielding net 39 are visible in the image of the mounting table 13.

[0040] First, in this embodiment, in a grayscale image of the pin-up state, the portion corresponding to the mounting surface of the mounting base 13 is set as the evaluation target range. When setting the evaluation target range, four coordinates 51a to 51d are specified to enclose the portion corresponding to the mounting surface of the mounting base 13, but at this time, the user specifies the coordinates so that the pin-up substrate G is not included in the evaluation target range (Figure 7(A)).

[0041] Subsequently, adaptive binarization processing is applied to the evaluation target area. At this time, for example, in area A where the shadow 50 of the substrate G and the shielding net 39 are partially visible, the binarization threshold for the central pixel with a pixel value of 100 in the binarization calculation area, which has pixels with pixel values ​​of 70, pixels with pixel values ​​of 100, and pixels with pixel values ​​of 120, becomes 97.5. Therefore, this central pixel is converted to black (Figure 7(B)). In this way, in the evaluation target area, the central pixels in the parts including the shadow 50 of the substrate G and the shielding net 39 are converted to black, so as shown in Figure 8(A), the shadow 50 of the substrate G and the shielding net 39 are not converted to white in the binarized image of the evaluation target area after adaptive binarization processing. Note that, as shown in Figure 7(A), the evaluation target area includes the lift pins 38, and the pixel values ​​change abruptly at the boundary between the lift pins 38 and the mounting base 13. Therefore, in the binarized image of the evaluation range after adaptive binarization processing, the contour of the lift pin 38 is converted to white, as shown in Figure 8(A).

[0042] On the other hand, if a normal binarization process is applied to the evaluation range instead of adaptive binarization, when the pixel values ​​of the shadow 50 of the substrate G and the shielding net 39 in the grayscale image fall below the binarization threshold, not only the lift pins 38 but also the shadow 50 of the substrate G and the shielding net 39 are converted to white, as shown in Figure 8(B).

[0043] By the way, in adaptive binarization processing, as described above, if the average value of the pixel values ​​of multiple pixels surrounding the central pixel is simply set as the binarization threshold, the range of change in the pixel values ​​of some of the shadows 50 and shielding nets 39 of the substrate G is somewhat large, even if not abrupt, so some of these shadows 50 and shielding nets 39 of the substrate G may be converted to white. In such cases, the binarization threshold is adjusted by the adjustment value C, as shown in equation (1) below.

[0044] Threshold = Average of the pixel values ​​of multiple pixels surrounding the central pixel - C …(1)

[0045] At this time, the user repeatedly performs adaptive binarization processing of the evaluation range while changing the adjustment value C, and sets the adjustment value C to a value at which the shadow 50 and shielding mesh 39 of the substrate G are all converted to black. Alternatively, the user may repeatedly perform adaptive binarization processing of the evaluation range while changing the adjustment value C, and set the adjustment value C to a value at which the outline of the lift pin 38 is clearly converted to white.

[0046] Furthermore, the setting of the adjustment value C may be performed using machine learning with training data created in advance through repeated adaptive binarization processing by the user, rather than through repeated adaptive binarization processing by the user each time. In this case, a grayscale image of the evaluation range is input to a pre-trained model configured as a Convolutional Neural Network, and the adjustment value C is output from the pre-trained model. The pre-trained model is generated by machine learning using the training data and is configured to output the adjustment value C when a grayscale image of the evaluation range is input. The training data here is training data for supervised learning and consists of a combination of a large number of grayscale images of the evaluation range for training, including shadows 50 of the substrate G and reflections of the barrier net 39, and an appropriate adjustment value C corresponding to each of these grayscale images (an adjustment value C that converts all reflections such as shadows 50 of the substrate G and barrier net 39 to black). The appropriate adjustment value C here may be obtained by first obtaining a binarized image from the grayscale image used as training data without using the adjustment value, and then adjusting the adjustment value C so that a binarized image can be obtained from the same grayscale image without reflections of the substrate G's shadow 50 or the shielding net 39. A large number of combinations of the adjustment value C obtained in this way and the grayscale image can then be prepared and used as training data.

[0047] Furthermore, the binarization threshold may be adjusted not by adjustment value C, but by changing the method of calculating the binarization threshold. For example, in the embodiment described above, the binarization calculation region was constructed by multiple pixels that single-fold surround the central pixel, and the average value of the pixel values ​​of each pixel included in the binarization calculation region was used as the binarization threshold. However, the binarization calculation region may be constructed by multiple pixels that double-fold surround the central pixel, and the average value of the pixel values ​​of each pixel included in the binarization calculation region may be used as the binarization threshold. Also, the multiple pixels surrounding the central pixel may, if necessary, be multiple pixels that triple-fold or more surround the central pixel. Furthermore, in the embodiment described above, in the binarization calculation region, the average value of only the pixel values ​​of the multiple pixels surrounding the central pixel was used as the binarization threshold (i.e., the pixel value of the central pixel is not reflected in the binarization threshold). However, in the binarization calculation region, the average value of the pixel values ​​of the multiple pixels surrounding the central pixel and the pixel value of the central pixel may be used as the binarization threshold.

[0048] Figure 9 is a flowchart showing the substrate abnormality detection process as a substrate abnormality detection method executed by the information processing device 40 in Figure 1. The substrate abnormality detection process in Figure 9 is realized by the CPU 41 executing a program stored in the ROM 42 or HDD 44. The substrate abnormality detection process in Figure 9 is executed, for example, when the information processing device 40 receives a user instruction to start the substrate abnormality detection process. In this embodiment, a recipe used by the film deposition apparatus 10 when executing a series of processes such as film deposition and static elimination of the substrate G is pre-stored in the memory device (built-in or external) of the control unit 37. The recipe stored in the memory device is loaded into the memory of the control unit 37 when the process is executed. This recipe includes, for example, the completion time of pin-up of the substrate G.

[0049] In Figure 9, first, the CPU 41 acquires a video captured by the camera 22 from the camera 22 (S91). As described above, this video includes images of the pin-up state that are the target of abnormality detection, that is, images showing the state in which the substrate G has been detached from the mounting surface. The CPU 41 also acquires a recipe stored in the memory of the control unit 37 from the control unit 37. At this time, it is not necessary for the CPU 41 to acquire all the information contained in the recipe; at least information regarding the end time of the pin-up state should be acquired. Note that the timing of acquiring the recipe is not limited to this timing, and the information processing device 40 may acquire the recipe from the control unit 37 in advance before starting this process. Note that in S91, the video may be associated with the time included in the recipe.

[0050] Next, the CPU 41 sets the evaluation range for the acquired video using the coordinates 51a to 51d of four points specified in advance by the user (S92). In this embodiment, before starting this process, the user specifies the coordinates 51a to 51d of four points using an image of the pin-up state (image of the pin-up state) so as to surround the part of the image that corresponds to the mounting surface of the mounting table 13 (see Figure 7(A)).

[0051] Next, the CPU 41 converts the images of the pin-up state to grayscale based on the acquired video and acquires (generates) a grayscale image of the pin-up state (S93). Specifically, the CPU 41 acquires an image from the acquired video that corresponds to the time (for example, the end time of the pin-up) indicating the pin-up state of the board G included in the recipe, and converts this image to grayscale to generate a grayscale image of the pin-up state.

[0052] Next, the CPU 41 performs adaptive binarization on the evaluation target area of ​​the grayscale image in the pin-up state (S94). Specifically, all pixels surrounded by other pixels in the evaluation target area are designated as central pixels, and a binarization calculation area having multiple pixels surrounding each central pixel is set. Adaptive binarization is performed on all binarization calculation areas to obtain a binarized image of the evaluation target area. That is, each pixel surrounded by other pixels in the evaluation target area is both a central pixel and a pixel that constitutes a binarization calculation area for the other central pixels. At this time, as shown in Figure 10, even if the image taken in the pin-up state includes not only the fragment 49 of the substrate G but also the shadow 50 of the substrate G and the shielding net 39, as shown in Figure 11, in the binarized image of the evaluation target area after adaptive binarization, the shadow 50 of the substrate G and the shielding net 39 are not converted to white, and only the outlines of the lift pins 38 and the fragment 49 of the substrate G are converted to white.

[0053] Next, the CPU 41 determines whether the proportion of white pixels in the binarized image of the evaluation range is equal to or greater than a predetermined anomaly detection threshold (S95). Here, the anomaly detection threshold is changed according to the type of substrate G. For example, if the substrate G is a substrate with a pattern, not only the outlines of the lift pins 38 and the fragments 49 of the substrate G, but also the outlines of the pattern may be converted to white. Therefore, the proportion of white pixels in the binarized image of the evaluation range will be larger than that of a substrate G made of plain glass without a pattern. Thus, the anomaly detection threshold for a substrate with a pattern is set higher than that for a substrate G made of plain glass. For example, the anomaly detection threshold for a substrate with a pattern is set to 4%, and the anomaly detection threshold for a substrate G made of plain glass is set to 2%. Note that the anomaly detection thresholds for a substrate with a pattern and a substrate G made of plain glass are not limited to these values.

[0054] Furthermore, while the user sets the anomaly detection threshold through prior experiments, the setting of the anomaly detection threshold may also be performed using machine learning. In this case, the binarized image of the evaluation range may be input to a trained model configured as a convolutional neural network, and the trained model may output the anomaly detection threshold. The trained model is generated by machine learning using training data and is configured to output the anomaly detection threshold when a binarized image of the evaluation range is input. Here, the training data includes a large number of binarized images of the evaluation range, including white pixels of the outlines of the substrate G fragment 49 and the lift pin 38, and each of these images is associated with an appropriate anomaly detection threshold corresponding to the binarized image of the evaluation range.

[0055] If, in S95, it is determined that the proportion of white pixels in the binarized image of the evaluation range is above a threshold, this means that the outline of the fragment 49 of the substrate G is clearly visible in the binarized image of the evaluation range. Therefore, the CPU 41 notifies the film deposition apparatus 10 that an abnormality such as a crack has occurred in the substrate G (S96) and terminates this process. Upon receiving this notification, the film deposition apparatus 10 stops the static elimination process that is currently running in order to prevent the spread of damage caused by the abnormality in the substrate G.

[0056] On the other hand, if in S95 it is determined that the proportion of white pixels in the binarized image of the evaluation range is below the threshold, the CPU 41 notifies the film deposition apparatus 10 that there is no abnormality in the substrate G (S97) and terminates this process. Upon receiving this notification, the film deposition apparatus 10 continues the static elimination process that is currently running. Alternatively, if it is determined that the proportion of white pixels in the binarized image of the evaluation range is below the threshold, the film deposition apparatus 10 may not be notified in any way and may continue the static elimination process that is currently running. In this process, after acquiring the video in S91, the notification in S96 or S97 is made in a relatively short time, for example, about 1 second.

[0057] According to this embodiment, adaptive binarization is applied to the evaluation target range of a pin-up grayscale image. In adaptive binarization, a central pixel whose pixel value abruptly changes to a small value from among the multiple pixels surrounding it is converted to white, while a central pixel whose pixel value does not change from among the multiple pixels surrounding it, or a central pixel whose pixel value does not change abruptly from among the multiple pixels surrounding it but only changes gradually, is converted to black. In addition, in a pin-up grayscale image, the pixel value abruptly changes at the boundary between the substrate G fragment 49 and the mounting base 13, while the pixel value does not change abruptly at the shadow 50 and shielding net 39 of the substrate G but only changes gradually. Therefore, when adaptive binarization is applied to a pin-up grayscale image, the boundary between the substrate G fragment 49 and the mounting base 13 is converted to white, while the shadow 50 and shielding net 39 of the substrate G are converted to black. In other words, the shadow 50 and shielding net 39 of the substrate G are not converted to white. This prevents the shadow 50 or shielding mesh 39 of the substrate G from being converted to white and being mistakenly determined to have cracked in the substrate G, thereby enabling accurate detection of abnormalities such as cracks in the substrate G.

[0058] Furthermore, in this embodiment, as described above, the blocking net 39 in the pin-up grayscale image is not converted to white, so the situation where it is impossible to distinguish between the small fragments 49 converted to white and the blocking net 39 converted to white in the binarized image does not occur. As a result, it is possible to detect the small fragments 49 remaining on the mounting base 13.

[0059] Furthermore, in this embodiment, since abnormalities in the substrate G are detected based on video, abnormalities in the substrate G can be detected in a relatively short time, for example, in about 1 second. That is, abnormalities in the substrate G can be detected almost immediately after the pin-up state of the substrate G is photographed. As a result, abnormalities in the substrate G can be detected after the pin-up state of the substrate G is photographed and before the substrate G is transported to the transport device, preventing fragments 49 of the substrate G from scattering near the gate valve 20.

[0060] Furthermore, in this embodiment, since video recording is performed using plasma emission emitted by plasma generated in the processing space U, video recording for detecting abnormalities in the substrate G can be performed without providing a new light source for recording. Moreover, in this embodiment, since the information processing device 40, which is a substrate abnormality detection device, is a separate device from the film deposition apparatus 10, it can also be used to detect the occurrence of substrate abnormalities in other substrate processing devices having a similar configuration to the film deposition apparatus 10.

[0061] Incidentally, in adaptive binarization, whether or not the central pixel is converted to white depends on the distribution of pixel values ​​from multiple pixels surrounding the central pixel, so the contours of the lift pin 38 and the substrate G fragment 49 are not guaranteed to be converted to white. For example, when the contour of the substrate G fragment 49, which has been converted to white, is magnified in part B of the binarized image of the evaluation range after adaptive binarization (Figure 12(A)), it can be seen that the contour is not smooth, and that there is noise in the shape of chips and protrusions (Figure 12(B)). Therefore, simply applying adaptive binarization to a grayscale image in a pin-up state may not accurately extract the contours of the lift pin 38 and the substrate G fragment 49, and as a result, the accuracy of detecting abnormalities in the substrate G using an abnormality detection threshold may decrease.

[0062] Therefore, it is preferable to perform morphological processing on the binarized image of the evaluation target range after adaptive binarization processing, before detecting abnormalities in the substrate G using an abnormality detection threshold. Specifically, a closing process is applied to the binarized image to fill in missing contours and reshape the contours, and then an opening process is applied to the binarized image that has undergone the closing process to remove protruding noise from the contours. In this embodiment, the closing process and the opening process are performed at least once each in the morphological processing. Note that in the morphological processing, the closing process may be performed after the opening process.

[0063] In the closing process, the binarized image of the evaluation range is subjected to dilation followed by condensation. In the opening process, the binarized image is subjected to condensation followed by dilation. In the dilation process, for example, as shown in Figure 13(A), multiple pixels including pixel 52, which is to be converted, are enclosed in kernel 53. If there is at least one pixel with a pixel value of 255 (white) within kernel 53, the pixel value of pixel 52, which is to be converted, is converted to 255 (white). On the other hand, in the condensation process, for example, as shown in Figure 13(B), multiple pixels including pixel 54, which is to be converted, are enclosed in kernel 55. If there is at least one pixel with a pixel value of 0 (black) within kernel 55, the pixel value of pixel 54, which is to be converted, is converted to 0 (black).

[0064] In this way, by applying morphological processing to the binarized image of the evaluation target range after adaptive binarization, the contours of the substrate G fragments 49 that have been converted to white can be smoothed (Figure 12(C)). Subsequently, abnormalities in the substrate G are detected using an abnormality detection threshold (S95).

[0065] While preferred embodiments of this disclosure have been described above, this disclosure is not limited to the embodiments described above, and various modifications and changes are possible within the scope of its essence.

[0066] In the embodiment described above, whether or not an abnormality has occurred in the substrate G is determined based on whether or not the proportion of white pixels in the binarized image of the evaluation range is equal to or greater than a predetermined threshold for abnormality detection. However, since the outline of the fragment 49 of the substrate G is thought to be composed of multiple consecutive white pixels, even if multiple white pixels are detected in the binarized image of the evaluation range, if these multiple white pixels are not consecutive, they are not considered to be the outline of the fragment 49 of the substrate G, and therefore, it can be considered that no abnormality has occurred in the substrate G. Therefore, in addition to the proportion of white pixels in the binarized image of the evaluation range, whether or not an abnormality has occurred in the substrate G may also be determined based on whether or not the multiple white pixels are consecutive.

[0067] Furthermore, in this embodiment, each camera 22 captures a video of the inside of the processing container 11 in color, but each camera 22 may also capture a video of the inside of the processing container 11 in monochrome. In this case, a pin-up grayscale image can be directly obtained from the captured video without performing grayscale conversion, thereby reducing the processing load required to obtain the grayscale image.

[0068] Furthermore, although this embodiment describes a case where the substrate abnormality detection device is configured as an information processing device 40 separate from the film deposition apparatus 10, the film deposition apparatus 10 may, for example, have the functions of the information processing device 40. In this case, the control unit 37 of the film deposition apparatus 10 executes the substrate abnormality detection process shown in Figure 9, and the same effects as in this embodiment can be achieved even with such a configuration.

[0069] Furthermore, although this embodiment describes the case where there is only one camera 22 for shooting video, the substrate anomaly detection system 1 may also have multiple cameras 22, for example, two cameras 22, that shoot video at different angles of view. In this case, if the proportion of white in either of the two binarized images of the evaluation target range acquired based on the two videos shot by the two cameras 22 is greater than or equal to the threshold for anomaly detection, it may be determined that an anomaly has occurred in the substrate G. This allows for the determination of whether or not an anomaly has occurred in the substrate G based on videos from different angles of view, thereby suppressing the oversight of anomalies in the substrate G and improving the accuracy of detecting anomalies in the substrate G.

[0070] Furthermore, in this embodiment, the camera 22 captured video footage of the inside of the processing container 11, but the camera 22 may also capture still images of the inside of the processing container 11. In this case, the camera 22 captures still images of the inside of the processing container 11 at the time indicating the pin-up state of the substrate G included in the recipe. This eliminates the step of acquiring images from video footage and reduces the processing load required to acquire grayscale images of the pin-up state.

[0071] Furthermore, in this embodiment, the film deposition apparatus 10 includes a rectangular tubular processing container 11, but the shape of the processing container 11 is not limited to a rectangular tubular shape; for example, the processing container 11 may have other shapes such as a cylindrical shape. In this case, the shape of the substrate may be circular.

[0072] In this embodiment, the application of the technology described herein to a film deposition apparatus has been described. However, the substrate processing apparatus to which the technology described herein is applied is not limited to a film deposition apparatus; the technology described herein can be applied to any substrate processing apparatus that performs pin-up. Examples of such substrate processing apparatuses include etching apparatuses, ashing apparatuses, and substrate heating apparatuses. Furthermore, although the film deposition apparatus to which the technology described herein is applied is configured as an inductively coupled plasma processing apparatus, it may also be configured as a plasma processing apparatus of other types, such as a capacitively coupled plasma processing apparatus or a microwave plasma processing apparatus. [Explanation of Symbols]

[0073] G board 1. Substrate abnormality detection system 10. Film deposition apparatus 11 Processing container 13. Mounting platform 21 Surveillance window 22 cameras 38 Lift Pins 39 Barrier net 40 Information Processing Devices 49 fragments 50 Shadows

Claims

1. A substrate abnormality detection method for detecting abnormalities in a substrate being processed inside a processing container of a substrate processing apparatus, The substrate processing apparatus is A mounting platform provided inside the processing container and having a mounting surface on which the substrate is placed, A lifting pin that can protrude from the mounting surface and allows the substrate to be raised and lowered, At least one observation window provided on the side wall of the processing container, The system includes a control device that controls the execution of the process applied to the substrate, The process includes, at least, acquiring a video, captured by at least one camera that photographs the inside of the processing container through the observation window, which includes the state in which the substrate is detached from the aforementioned mounting surface by the lifting pin, A step of setting the evaluation range to the portion corresponding to the aforementioned mounting surface in order that the substrate is not included in the image of the pin-up state in which the substrate is detached from the aforementioned mounting surface, among the multiple images that constitute the aforementioned video, A step of applying adaptive binarization processing to the evaluation target range of a grayscale image obtained by representing the pin-up image in grayscale, A substrate abnormality detection method comprising the step of determining that an abnormality has occurred in the substrate if the proportion of white pixels in the binarized image obtained by applying adaptive binarization processing to the evaluation target range is equal to or greater than a predetermined threshold.

2. The substrate abnormality detection method according to claim 1, wherein in the step of applying adaptive binarization processing to the evaluation target range, the outlines of the substrate fragments remaining on the aforementioned surface included in the evaluation target range are converted to white.

3. The substrate abnormality detection method according to claim 1, wherein in the step of applying adaptive binarization processing to the evaluation target range, all pixels in the evaluation target range that are surrounded by other pixels are defined as central pixels, a binarization calculation region having a plurality of pixels surrounding each of the central pixels is defined for each of the central pixels, the average value of the pixel values ​​of each pixel in the binarization calculation region is defined as the binarization threshold, if the pixel value of the central pixel in the binarization calculation region is equal to or greater than the binarization threshold, the central pixel is converted to black, and if the pixel value of the central pixel is less than the binarization threshold, the central pixel is converted to white.

4. The substrate anomaly detection method according to claim 3, wherein the binarization threshold is the average value of only the pixel values ​​of the plurality of pixels surrounding the central pixel in the binarization calculation region.

5. The substrate anomaly detection method according to claim 3, wherein the binarization threshold is the average value of the pixel values ​​of a plurality of pixels surrounding the central pixel and the pixel value of the central pixel in the binarization calculation region.

6. The substrate abnormality detection method according to claim 3, further comprising the step of setting the binarization threshold by adjusting the average value of the pixel values ​​of each pixel in the binarization calculation region with an adjustment value.

7. The substrate abnormality detection method according to claim 6, wherein the adjustment value is set to a value such that, in the step of applying adaptive binarization processing to the evaluation target range, all shadows of the substrate included in the evaluation target range and reflections of the shielding mesh that blocks the observation window from the plasma are converted to black.

8. In the step of setting the binarization threshold, a grayscale image of the substrate detached from the mounting surface is input to the trained model, and the trained model outputs the adjustment value. The substrate anomaly detection method according to claim 7, wherein the trained model is generated by machine learning using a combination of a plurality of grayscale images of the evaluation target range, including shadows of the substrate and reflections of the shielding net, and adjustment values ​​associated with the plurality of grayscale images that convert the reflections to black as training data.

9. The substrate abnormality detection method according to claim 1, further comprising the step of directly obtaining a grayscale image, which is a grayscale representation of the pin-up state image, from the video.

10. The substrate abnormality detection method according to claim 1, wherein the predetermined threshold is changed according to the type of substrate.

11. A substrate abnormality detection device that detects abnormalities in a substrate being processed inside a processing container of a substrate processing apparatus, The substrate processing apparatus is A mounting platform provided inside the processing container and having a mounting surface on which the substrate is placed, A lifting pin that can protrude from the mounting surface and allows the substrate to be raised and lowered, At least one observation window provided on the side wall of the processing container, The system includes a control device that controls the execution of the process applied to the substrate, The substrate abnormality detection device includes a control unit, The control unit, The process includes, at least, acquiring a video, captured by at least one camera that photographs the inside of the processing container through the observation window, which includes the state in which the substrate is detached from the aforementioned mounting surface by the lifting pin, A step of setting the evaluation range to the portion corresponding to the aforementioned mounting surface in order that the substrate is not included in the image of the pin-up state in which the substrate is detached from the aforementioned mounting surface, among the multiple images that constitute the aforementioned video, A step of applying adaptive binarization processing to the evaluation target range of a grayscale image obtained by representing the pin-up image in grayscale, A substrate abnormality detection device that performs the steps of determining that an abnormality has occurred in the substrate if the proportion of white pixels in the binarized image obtained by applying adaptive binarization processing to the evaluation target range is equal to or greater than a predetermined threshold.

Images