Method for measuring and calculating diameter of crystal in seeding stage of semiconductor crystal growth furnace

Abstract

The invention discloses a crystal seeding stage diameter measuring and calculating method for a semiconductor crystal growth furnace, and relates to the technical field of semiconductor crystal growth furnaces, and the method comprises the following steps: taking an original crystal seeding image as a first characteristic pattern, and cutting a seeding detection range area of the original crystal seeding image as a second characteristic pattern; the pixel diameter DLTXSmaxDistance of the guide cylinder is calculated according to the first feature map; according to the pixel diameter DLTXSmaxDistance of the guide cylinder, a correction coefficient RFactor is calculated; according to the crystal column pixel diameter JZXSmaxDistance and the correction coefficient RFactor, the crystal column correction diameter RealR is calculated and serves as the crystal seeding stage diameter. According to the invention, the diameter of the crystal in the single crystal furnace at the seeding stage is detected, the initial growth size of the crystal can be accurately mastered, accurate initial parameters are provided for the growth of the subsequent crystal, and meanwhile, the semiconductor crystals produced in the same batch can be ensured to have high consistency in diameter through accurate diameter measurement and calculation, so that the production efficiency is improved. And excessive or insufficient crystal growth can be avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor crystal growth furnaces, and in particular to a method for measuring the diameter of a crystal in a seeding stage of a semiconductor crystal growth furnace. Background Art

[0002] In the process of crystal growth, seeding is a key initial step. The purpose of seeding is to form a small crystal nucleus in the melt and make it grow in the desired crystal direction. This small crystal nucleus will gradually grow into a complete crystal. For example, in the process of growing single crystal silicon by the Czochralski method, the seeding stage determines the direction of crystal growth and the initial quality. Accurately measuring the diameter of the crystal in the seeding stage is crucial to controlling the crystal growth process. If the diameter is not well controlled in the seeding stage, defects such as dislocations and twins may occur in the subsequent crystal growth process, affecting the electrical and optical properties of the crystal.

[0003] At present, in the production process of semiconductor crystal growth furnace, the diameter measurement of crystal seeding stage is carried out through two methods: one is that workers put a caliper with a lens tube on the observation port of the crystal growth furnace and measure the diameter by visual inspection; the other is that workers measure the diameter through a third-party visual acquisition solution, such as Hikvision and Keyence. There are mainly the following disadvantages:

[0004] 1) The obvious disadvantage of the manual visual inspection solution is that it lacks sufficient real-time and accuracy. The worker's measurement speed sometimes cannot keep up with the crystal growth speed. The value measured by the worker in the last second loses a certain degree of real-timeness in the next second due to the growth of the crystal. Since the visual inspection results of each worker are not necessarily the same and the caliper scale cannot be accurate to the decimal point, the measured diameter itself does not have a high accuracy.

[0005] 2) The problems with using third-party visual acquisition solutions are mainly reflected in the following aspects. First, the visual hardware used by each third-party platform is different. Take industrial cameras as an example. Keyence's industrial cameras are generally a set of equipment with the visual detection host computer they provide, and the power and data cables use one cable. In the visual acquisition solution used by Hikvision, the industrial camera and the visual detection host computer are separated. The delisting rate of industrial cameras in the industrial camera market is very high. Almost every year, a large number of industrial camera models are delisted. The non-interoperability of hardware replacement parts and the high delisting rate make the maintenance cost high; the visual detection solution of the third-party platform does not have high adaptability, because the detection programs used in it are of fixed types. For example, a company's MVS uses a variety of basic visual processing functions in machine learning to meet the needs of industrial detection through self-combination. This method does not meet the special requirements of some industrial scenarios. Different industrial enterprises have different process flows. In order to increase industrial output and product yield and improve some product parameters, highly customizable detection programs are required.

[0006] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0007] In view of the problems in the related art, the present invention proposes a method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace to overcome the above-mentioned technical problems existing in the existing related art.

[0008] The technical solution of the present invention is achieved in this way:

[0009] A method for measuring the diameter of a crystal in a seeding stage of a semiconductor crystal growth furnace comprises the following steps:

[0010] Pre-collecting a crystal seeding original image through an industrial camera, and using the crystal seeding original image as a first feature map, and cutting a seeding detection range area of ​​the crystal seeding original image as a second feature map;

[0011] Calculate the pixel diameter DLTXSmaxDistance of the guide tube according to the first feature map;

[0012] According to the pixel diameter of the guide tube DLTXSmaxDistance, the correction coefficient RFactor is calculated and expressed as:

[0013]

[0014] Among them, DLTSJDistance represents the actual diameter of the guide tube;

[0015] According to the second characteristic graph, the crystal column pixel diameter JZXSmaxDistance is calculated;

[0016] According to the crystal column pixel diameter JZXSmaxDistance and the correction coefficient RFactor, the crystal column correction diameter RealR is calculated as the crystal seeding stage diameter, expressed as:

[0017] RealR=JZXSmaxDistance×RFactor.

[0018] Furthermore, the step of collecting the original image of the crystal seeding by an industrial camera in advance comprises the following steps:

[0019] Pre-enumerate all industrial cameras in the subnet;

[0020] Select the specified online device to create a device handle;

[0021] Establish a connection between the industrial computer and the industrial camera, including connecting through the GigEVision industrial protocol;

[0022] Set continuous mode, in which the industrial camera will continuously generate image data streams at a set frequency, rate or trigger source;

[0023] Store the image data in nBuffSize, and allocate memory size to a frame of crystal seeding original image according to nBuffSize;

[0024] Set the input and output parameters of the crystal seeding raw image.

[0025] Furthermore, the step of calculating the guide tube pixel diameter DLTXSmaxDistance according to the first feature map includes the following steps:

[0026] Use the cvtColor function to convert the first feature map into a first grayscale image;

[0027] Reading a first grayscale image, and adjusting sharpness, contrast, and brightness of the first grayscale image;

[0028] The first grayscale image is binarized using a threshold function to convert it into a binary image;

[0029] Find all white pixels in the binary image and put all pixel coordinates into the coordinate set YinJing_contours;

[0030] Traverse the two points YinJing_minPoint and YinJing_maxPoint with the farthest horizontal coordinate distance in the coordinate set YinJing_contours;

[0031] Calculate the pixel distance between YinJing_minPoint and YinJing_maxPoint as the pixel diameter of the guide tube DLTXSmaxDistance, expressed as:

[0032]

[0033] Among them, (x 1 ,y 1 ) represents the coordinate value of point YinJing_minPoint, (x 2 ,y 2 ) represents the coordinate value of point YinJing_maxPoint.

[0034] Furthermore, the method further comprises the following steps:

[0035] Presetting setting parameters of image sharpening, contrast, and brightness, and reading the setting parameters when reading the first grayscale image;

[0036] The binarization parameters are preset and used to read the parameters when converting to a binary image.

[0037] Furthermore, the step of calculating and obtaining the crystal column pixel diameter JZXSmaxDistance according to the second characteristic graph includes the following steps:

[0038] Narrow the field of view of the original crystal seeding image and select the target area as the seeding detection range;

[0039] Cut the original crystal seeding image according to the seeding detection range to obtain a seeding detection range area image;

[0040] The crystal column pixel diameter JZXSmaxDistance is calculated.

[0041] Furthermore, the seeding detection range area image includes: a calibrated aperture where the seed crystal contacts the liquid surface, which is represented as the seeding detection range area.

[0042] Beneficial effects of the present invention:

[0043] The present invention is a method for measuring the diameter of a crystal seeding stage in a semiconductor crystal growth furnace. The method collects an original image of crystal seeding through an industrial camera, and uses the original image of crystal seeding as a first feature map. The original image of crystal seeding is cut into a seeding detection range area as a second feature map. The pixel diameter DLTXSmaxDistance of the guide tube is calculated according to the first feature map. The correction coefficient RFactor is calculated according to the pixel diameter DLTXSmaxDistance of the guide tube. At the same time, the pixel diameter JZXSmaxDistance of the crystal column is calculated according to the second feature map. Distance and correction coefficient RFactor are used to calculate the corrected diameter RealR of the crystal column, and the diameter detection of the crystal in the single crystal furnace during the seeding stage can be realized. This can accurately grasp the size of the initial growth of the crystal, provide accurate starting parameters for the subsequent crystal growth, and help reduce defects such as dislocations and twins during the crystal growth process; at the same time, through accurate diameter measurement, it can ensure that the semiconductor crystals produced in the same batch have a high degree of consistency in diameter, which can avoid excessive or insufficient crystal growth. Through accurate measurement, problems can be discovered at an early stage and corrected in time, thereby reducing material waste and reducing production costs. In addition, accurate diameter measurement helps to reasonably arrange the production tasks of the crystal growth furnace. When the seeding diameter can be accurately controlled, the crystal growth furnace can produce crystals that meet the specifications according to the predetermined plan, reducing equipment idle time caused by crystal growth failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 The present invention is a flow chart of a method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0047] According to an embodiment of the present invention, a method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace is provided.

[0048] like Figure 1 As shown, the method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to an embodiment of the present invention comprises the following steps:

[0049] Step S1, pre-collecting a crystal seeding original image by an industrial camera, and using the crystal seeding original image as a first feature map, and cutting a seeding detection range area of ​​the crystal seeding original image as a second feature map;

[0050] This technical solution, for the above industrial camera to collect the original image of crystal seeding, includes the following steps:

[0051] Step S101, pre-enumerate all industrial camera devices in the subnet: to obtain the device numbered 0 in the subnet;

[0052] In this technical solution, the default number of the first industrial camera connected to the subnet is 0. When a binocular scene is required, there are two industrial cameras, 0 and 1.

[0053] Step S102, select the specified online device to create a device handle: the production uses a certain company's MV series global industrial camera, call the MV_CC_CreateHandle function, and pass in the device information and handle pointer parameters to select the online device numbered 0 and create the corresponding device handle for subsequent communication and operation with the industrial camera.

[0054] Step S103, connecting devices: establishing a connection between the industrial computer and the industrial camera through the GigEVision industrial protocol;

[0055] Step S104, setting the continuous mode: In the continuous mode, the industrial camera will continuously generate image data streams at a certain frequency, rate or trigger source. The continuous mode is adopted because during the growth of single crystal silicon, the crystal column is constantly rotating, and the pot in the crystal growth furnace is also rotating during the crystal rotation process. The operation of collecting images needs to be performed uninterruptedly to maintain the real-time nature of the production situation in the furnace.

[0056] Step S105, turn off the trigger source: after setting the trigger mode to off, the industrial camera will switch to continuous mode and no longer rely on external trigger signals to trigger image acquisition. In this way, the industrial camera will continue to collect image data without external interference, which is suitable for some continuous image acquisition scenarios.

[0057] Step S106, store the image data in nBuffSize: nBufSize in the program is the size used to store a frame of data, that is, the memory space required for the image data transmitted by the industrial camera. By assigning PayloadSize to nBufSize, it can be ensured that the program has enough memory space to store each frame of image data when receiving it, so as to facilitate subsequent processing and display.

[0058] Step S107, allocate memory size for a frame of crystal seeding original image according to nBuffSize: by nTestFrameSize=1;, nTestFrameSize is explicitly initialized to 1 to ensure that it is used as the default value of the frame size. At the same time, memory allocation is performed by pFrameBuf=(unsignedchar*)malloc(nBufSize), and a memory space of size nBufSize is opened for frame data.

[0059] Step S108, setting the input and output parameters of the crystal seeding original image:

[0060] Define the input and output parameters of the crystal seeding original image by setting different members of the MV_SAVE_IMAGE_PARAM_EX structure stParam.

[0061] Specifically, the crystal seeding original image parameter settings include the following:

[0062] 1) stParam.enImageType = MV_Image_Jpeg, set the image type to be saved to JPEG format; 2) stParam.nBufferSize = nBufSize, define the buffer size for storing image data, here use the frame data size nBufSize calculated before; 3) unsigned char* pImage = (unsigned char*)malloc(nBufSize), allocate memory space for the converted image data; 4) stParam.pImageBuffer = pImage, specify the output data buffer for storing the converted image data. By setting the above parameters one by one, the program can accurately define the characteristics of the input and output data for image processing, encoding or saving operations.

[0063] Step S109, select the storage location of the original image of crystal seeding: calibrate the storage location of the original image of crystal seeding as "D:\CNW\image\cam-YJ\cam.jpg", which is used for storing and reading the original image of crystal seeding.

[0064] Step S2, calculating the guide tube pixel diameter DLTXSmaxDistance according to the first feature map, includes the following steps:

[0065] Step S201, converting the first feature map to grayscale: using the cvtColor function to convert the first feature map into a first grayscale image and store it in YinJing_gray_image.

[0066] Step S202, reading the first grayscale image, and adjusting the sharpness, contrast, and brightness of the first grayscale image.

[0067] The technical solution performs image preprocessing on the first grayscale image, which helps to improve the features that need to be identified in the first grayscale image and reduce some interference information. For example, appropriately increasing the sharpness and contrast can increase the clarity of the crystal edges in the image, and appropriately reducing the brightness can reduce the degree of reflection of the liquid surface of the molten material.

[0068] In addition, for adjusting the sharpness, contrast, and brightness of the first grayscale image, setting parameters for image sharpness, contrast, and brightness may be preset.

[0069] Specifically, when applying, in the control interface compiled by QT, find the parameter setting tool of the corresponding stage you want to change in advance through the tool box > image parameter setting button >; after setting the parameters, click the corresponding button to update the parameters. For example, after setting the seeding parameters to 0, 1, 0, click the "seeding parameter update" button to update the screen.

[0070] At the same time, the working principle of the three-line parameter setting tool is exactly the same. Take the seeding image parameter setting as an example. The three data of seeding are 0, 1, 0. The first data 0 represents sharpness, and the setting range of sharpness is 0-10. The second data 1 represents contrast, and the setting range of contrast is 0-10. The third data 0 represents brightness, and the setting range of brightness is -255 to +255. The initialized parameters basically meet the production requirements.

[0071] Step S203, using binarization to convert the first grayscale image into a binary image: using a threshold function to binarize the first grayscale image and convert it into a binary image, and the result is stored in YinJing_binary_image.

[0072] In this technical solution, for the above-mentioned binarization, the binarization parameters can be set in advance, and the binarization parameters can be adjusted according to specific production needs. The binarization threshold is set in the control interface compiled by QT, and the set binarization threshold will be read later. This threshold is an upper limit and a lower limit, and the range is 0-255.

[0073] Step S204, searching for all white pixels in the binary image and putting all pixel coordinates into the coordinate set YinJing_contours.

[0074] This technical solution uses the findContour function to find the white pixel contours in the binary image YinJing_binary_image and saves these contours in the YinJing_contours vector.

[0075] Step S205, traverse the two points YinJing_minPoint and YinJing_maxPoint with the farthest horizontal coordinate distance in the coordinate set YinJing_contours:

[0076] Specifically, initialize YinJing_minX to a larger number and YinJing_maxX to 0. At the same time, initialize YinJing_minPoint and YinJing_maxPoint to store the corresponding points. Traverse each contour and its points, convert the coordinates of each point to the original coordinate system, compare them with YinJing_minX and YinJing_maxX, and update the minimum and maximum x values ​​and the corresponding coordinate points.

[0077] Step S206, calculate the pixel distance between YinJing_minPoint and YinJing_maxPoint as the pixel diameter DLTXSmaxDistance of the guide tube, expressed as:

[0078]

[0079] Among them, (x 1 ,y 1 ) represents the coordinate value of point YinJing_minPoint, (x 2 ,y 2 ) represents the coordinate value of point YinJing_maxPoint.

[0080] Step S3, according to the guide tube pixel diameter DLTXSmaxDistance, calculate the correction coefficient RFactor, which is expressed as:

[0081]

[0082] Wherein, DLTSJDistance represents the actual diameter of the guide tube, which is obtained through actual measurement in production.

[0083] Step S4, similarly, the seeding detection range area is cut out from the original crystal seeding image as the second feature map, and the crystal column pixel diameter JZXSmaxDistance is calculated;

[0084] This technical solution includes the following steps:

[0085] Step S401, narrowing the field of view of the original crystal seeding image to select a target area as a seeding detection range;

[0086] In this technical solution, the selection box is to narrow the field of view to calibrate the target area and enlarge it after selection. The crystal is too thin in the seeding stage. With this operation, the crystal condition can be directly observed (for example, whether the double aperture stage appears). Otherwise, the worker cannot see clearly through the observation port because the furnace body is very deep.

[0087] In this technical solution, the field of view is narrowed to obtain the area where the aperture appears when the seed crystal contacts the liquid surface, and a rectangular detection frame is drawn to frame the aperture part as the seed crystal detection range.

[0088] Step S402, cutting the crystal seeding original image according to the seeding detection range to obtain a seeding detection range area image;

[0089] In addition, since the diameter of the crystal in the seeding stage is very small, generally around 1-15mm, and the final crystal column needs to reach a diameter that is ten or even thirty times that, it is necessary to crop and enlarge the crystal image in the seeding stage to obtain an image of the seeding detection range area.

[0090] Step S403, similarly, calculate and obtain the crystal column pixel diameter JZXSmaxDistance, which will not be elaborated here.

[0091] Step S5, according to the crystal column pixel diameter JZXSmaxDistance and the correction coefficient RFactor, calculate the crystal column correction diameter RealR as the crystal seeding stage diameter, expressed as:

[0092] RealR=JZXSmaxDistance×RFactor.

[0093] To sum up, with the help of the above-mentioned technical scheme of the present invention, the original image of crystal seeding is collected by an industrial camera, and the original image of crystal seeding is used as the first feature map, the seeding detection range area of ​​the original image of crystal seeding is cut as the second feature map, and the pixel diameter DLTXSmaxDistance of the guide tube is calculated according to the first feature map, and the correction coefficient RFactor is calculated according to the pixel diameter DLTXSmaxDistance of the guide tube, and the pixel diameter JZXSmaxDistance of the crystal column is calculated according to the second feature map. According to the pixel diameter JZXSmaxDistance of the crystal column and the correction coefficient RFactor, the crystal column correction diameter RealR is calculated, and the diameter detection of the crystal seeding stage in the single crystal furnace is realized, which can accurately grasp the initial growth size of the crystal, provide accurate starting parameters for the subsequent crystal growth, and help to reduce defects such as dislocations and twins in the crystal growth process; at the same time, through accurate diameter measurement, it can be ensured that the semiconductor crystals produced in the same batch have a high consistency in diameter, which can avoid excessive or insufficient crystal growth. Through accurate measurement, problems can be discovered in the early stage and corrected in time, thereby reducing material waste and reducing production costs. In addition, accurate diameter measurement helps to reasonably arrange the production tasks of the crystal growth furnace. When the seed crystal diameter can be accurately controlled, the crystal growth furnace can produce crystals that meet the specifications according to the predetermined plan, reducing the equipment idle time caused by crystal growth failure.

[0094] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. After considering the disclosure of the specification and the examples, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and the examples are only regarded as exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

[0095] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace, characterized in that: The following steps are involved: Pre-collecting a crystal seeding original image through an industrial camera, and using the crystal seeding original image as a first feature map, and cutting a seeding detection range area of ​​the crystal seeding original image as a second feature map; Calculate the pixel diameter DLTXSmaxDistance of the guide tube according to the first feature map; According to the pixel diameter of the guide tube DLTXSmaxDistance, the correction coefficient RFactor is calculated and expressed as: Among them, DLTSJDistance represents the actual diameter of the guide tube; According to the second characteristic graph, the crystal column pixel diameter JZXSmaxDistance is calculated; According to the crystal column pixel diameter JZXSmaxDistance and the correction coefficient RFactor, the crystal column correction diameter RealR is calculated as the crystal seeding stage diameter, expressed as: RealR=JZXSmaxDistance×RFactor.

2. The method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to claim 1, characterized in that: The step of collecting the original image of the crystal seeding by an industrial camera in advance includes the following steps: Pre-enumerate all industrial cameras in the subnet; Select the specified online device to create a device handle; Establish a connection between the industrial computer and the industrial camera, including connecting through the GigEVision industrial protocol; Set continuous mode, in which the industrial camera will continuously generate image data streams at a set frequency, rate or trigger source; Store the image data in nBuffSize, and allocate memory size to a frame of crystal seeding original image according to nBuffSize; Set the input and output parameters of the crystal seeding raw image.

3. The method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to claim 1, characterized in that: The step of calculating the guide tube pixel diameter DLTXSmaxDistance according to the first feature map includes the following steps: Use the cvtColor function to convert the first feature map into a first grayscale image; Reading a first grayscale image, and adjusting sharpness, contrast, and brightness of the first grayscale image; The first grayscale image is binarized using a threshold function to convert it into a binary image; Find all white pixels in the binary image and put all pixel coordinates into the coordinate set YinJing_contours; Traverse the two points YinJing_minPoint and YinJing_maxPoint with the farthest horizontal coordinate distance in the coordinate set YinJing_contours; Calculate the pixel distance between YinJing_minPoint and YinJing_maxPoint as the pixel diameter of the guide tube DLTXSmaxDistance, expressed as: Among them, (x1, y1) represents the coordinate value of the point YinJing_minPoint, and (x2, y2) represents the coordinate value of the point YinJing_maxPoint.

4. The method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to claim 3, characterized in that: The following steps are also included: Presetting setting parameters of image sharpening, contrast, and brightness, and reading the setting parameters when reading the first grayscale image; The binarization parameters are preset and used to read the parameters when converting to a binary image.

5. The method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to claim 1, characterized in that: The step of calculating and obtaining the crystal column pixel diameter JZXSmaxDistance according to the second characteristic graph includes the following steps: Narrow the field of view of the original crystal seeding image and select the target area as the seeding detection range; Cut the original crystal seeding image according to the seeding detection range to obtain a seeding detection range area image; The crystal column pixel diameter JZXSmaxDistance is calculated.

6. The method for measuring the diameter of a crystal in the seeding stage of a semiconductor crystal growth furnace according to claim 5, characterized in that: The seeding detection range area image includes: a calibrated aperture where the seed crystal contacts the liquid surface, which is represented as the seeding detection range area.

Images