Method for preventing heavy air-doped pore product from flowing out

Abstract

The invention relates to a method for preventing a heavily-doped porous product from flowing out, which belongs to the technical field of silicon wafer detection and comprises the following operation steps: step 1, in an appearance detection station, surface defects are photographed by a front camera, identified and classified into particles and PIT; and 2, taking an image of the defect identified as PIT by a PIT camera, and determining whether the defect is an air hole according to the image calculation logic of the PIT camera. And 3, according to different defects in the PIT image, taking the captured peripheral and central bright spot parts and whether the whole is similar to a circle as a basis for judging the air hole. And 4, identifying the heavily doped pores by starting a corresponding PIT camera identification and judgment function of the appearance detection equipment. The method has the advantages of convenience in operation and good running stability, and solves the problem of air hole defects in the manufacturing process of the monocrystalline silicon wafer. The heavy air doping hole product is effectively prevented from flowing to the next station, and the problem that the final device is abnormal due to unidentified abnormity is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of silicon wafer detection, and particularly relates to a method for preventing heavy-doped air hole products from flowing out. BACKGROUND

[0002] In the manufacturing process of monocrystalline silicon wafers, a crucible is used to draw a single crystal. Ar or SiO in the crystal pulling furnace is adsorbed on the damaged surface or pit of the quartz crucible, and bubbles enter the crystal bar through thermal convection in the crystal growth process, resulting in air holes. On the one hand, it is currently difficult to completely avoid the occurrence of air holes.

[0003] At present, air hole defect inspection is carried out by using infrared light transmission inspection, which is suitable for products with resistivity of 5-6 mohm-cm or more. For products with resistivity of 5-6 mohm-cm or less, most of the infrared light source is absorbed by the product and cannot be transmitted, so it cannot effectively detect air holes. The heavy-doped silicon wafer with air holes flows out to the client, causing abnormal devices. SUMMARY

[0004] The present application mainly solves the problems in the prior art, and provides a method for preventing heavy-doped air hole products from flowing out, which has the advantages of convenient operation and good running stability, solves the problem of air hole defects in the manufacturing process of monocrystalline silicon wafers. Effectively prevent heavy-doped air hole products from flowing out to the next station, avoid the problem of abnormal final devices caused by unrecognized abnormalities.

[0005] The above technical problems of the present application are mainly solved by the following technical scheme:

[0006] A method for preventing heavy-doped air hole products from flowing out, comprising the following operation steps:

[0007] Step 1: In the appearance detection station, the surface defects are identified and classified as particles and PIT by the front camera.

[0008] Step 2: The defects identified as PIT are imaged by the PIT camera, and whether it is an air hole is confirmed according to the image calculation logic of the PIT camera.

[0009] Step 3: The PIT image captures the outer and central bright spots, and whether the whole is similar to a circle as the basis for determining air holes.

[0010] Step 4: Heavy-doped air holes are identified by turning on the PIT camera identification function of the appearance detection equipment.

[0011] PIT (Progressive Image Transmission) is an image transmission technology mainly used to gradually improve the reconstruction quality of images.

[0012] As preferred, the image calculation logic of the PIT camera mainly includes a perspective projection transformation and an iterative optimization process, and the PIT algorithm, i.e., a perspective n-point algorithm, is used to estimate the 3D pose of an object relative to the translation and rotation of the lens, and the core is to calculate the pose parameters of the object through a perspective projection model.

[0013] As preferred, the PIT algorithm estimates the 3D pose of an object through a perspective projection transformation. The mathematical expression of the perspective projection transformation is as follows:

[0014]

[0015] where s is a scale factor, f X and f y are the focal length components of the camera in the x and y directions, (u0, v0) is the optical center coordinates, R is the rotation matrix, T is the translation vector, (X W , Y W , Z W ) is the coordinates of a point on the object in the world coordinate system, and (u, v) is the coordinates of the point on the image plane.

[0016] As preferred, the object is first initialized, assuming that all points of the object are at the same depth, i.e., the Z C values of all points are approximately equal, and a simplified calculation is performed; then the iterative optimization is performed, the rotation matrix R and the translation vector T are iteratively optimized so that the projection coordinates (u, v) are as close as possible to the actual observed image coordinates; in each iteration, the image coordinates are calculated using the perspective projection model, and the parameters are adjusted according to the error; finally, the convergence is judged, and when the iteration result reaches a certain accuracy or the number of iterations reaches a preset value, the iteration is stopped, and the final rotation matrix R and translation vector T are output.

[0017] As preferred, the calculation logic of the PIT camera is based on the fact that the stomata have different characteristics from the conventional PIT, and the stomata are generally circular.

[0018] The present application can achieve the following effects:

[0019] The present application provides a method for preventing the outflow of heavy doped stomata products, which has the advantages of convenient operation and good running stability compared with the prior art, and solves the problem of stomata defects in the manufacturing process of single crystal silicon wafers. The outflow of heavy doped stomata products to the next station is effectively prevented, and the problem of abnormal final devices caused by unidentified abnormalities is avoided. DETAILED DESCRIPTION

[0020] The technical solutions of the present application will be further specifically described below through examples.

[0021] Embodiment: A method for preventing the outflow of heavy doped stomata products, comprising the following operation steps:

[0022] First step: At the appearance detection station, the surface defects are identified and classified as particles and PITs by the front camera.

[0023] Second step: The defects identified as PITs are imaged by the PIT camera, and whether it is a pore is confirmed according to the image calculation logic of the PIT camera.

[0024] The image calculation logic of the PIT camera mainly includes perspective projection transformation and iterative optimization process. The PIT algorithm, also known as perspective n-point algorithm, is used to estimate the 3D pose of an object relative to the lens translation and rotation. The core is to calculate the pose parameters of the object through the perspective projection model.

[0025] The PIT algorithm estimates the 3D pose of an object through perspective projection transformation. The mathematical expression of perspective projection transformation is as follows:

[0026]

[0027] where s is the scale factor, f x and f y are the focal length components in the x and y directions of the camera, (u0, v0) is the optical center coordinates, R is the rotation matrix, T is the translation vector, (X W , Y W , Z W ) is the coordinates of a point on the object in the world coordinate system, and (u, v) is the coordinates of the point on the image plane.

[0028] Third step: The PIT image will capture the outer and central bright spots, and whether the overall shape is similar to a circle as the basis for determining pores.

[0029] First, initialize, assuming that all points of the object are at the same depth, i.e. the Z C values of all points are approximately equal, for simplified calculation; then iterative optimization, by iteratively optimizing the rotation matrix R and the translation vector T, so that the projected coordinates (u, v) are as close as possible to the actual observed image coordinates; in each iteration, use the perspective projection model to calculate the image coordinates, and adjust the parameters according to the error; finally, perform convergence judgment, when the iteration result reaches a certain accuracy or the number of iterations reaches a preset value, stop iteration, and output the final rotation matrix R and translation vector T.

[0030] The calculation logic of the PIT camera is based on the fact that pores have different characteristics from regular PITs, and pores are generally circular.

[0031] Fourth step: Identify the heavy doping pores by enabling the PIT camera recognition and judgment function of the appearance detection device.

[0032] In summary, the method for preventing the outflow of the heavy doped air hole product has the advantages of convenient operation and good running stability, solves the problem of air hole defects of the single crystal silicon wafer in the manufacturing process, effectively prevents the outflow of the heavy doped air hole product to the next station, and avoids the problem of abnormal final device caused by unidentified abnormalities.

[0033] The above merely describes specific embodiments of the present application, but the structural features of the present application are not limited thereto, and any changes or modifications made by those skilled in the art within the scope of the present application are covered by the patent range of the present application.

Claims

1. A method for preventing the outflow of heavily aerated products, characterized in that The steps are as follows: Step 1: At the appearance inspection station, surface defects are photographed by a front-facing camera and classified into particles and PITs. Step 2: The defect identified as PIT is captured by the PIT camera and confirmed to be a pore based on the image calculation logic of the PIT camera; Step 3: The PIT image will capture the peripheral and central bright spots according to the defect, and whether the overall shape is similar to a circle as the basis for determining whether it is a pore; Step 4: Identify the re-injection holes by turning on the corresponding PIT camera recognition and judgment function of the appearance inspection equipment.

2. The method for preventing the outflow of heavily aerated products according to claim 1, characterized in that: The image calculation logic of the PIT camera mainly includes perspective projection transformation and iterative optimization process. The PIT algorithm, also known as the perspective n-point algorithm, is used to estimate the translation and rotation of an object's 3D posture relative to the lens. The core is to calculate the object's posture parameters through a perspective projection model.

3. The method for preventing the outflow of heavily aerated products according to claim 2, characterized in that: The PIT algorithm estimates the 3D pose of an object through perspective projection transformation. The mathematical expression of perspective projection transformation is as follows: Where s is the scale factor, f x and f y is the component of the camera's focal length in the x and y directions, (u0, V0) is the optical center coordinate, R is the rotation matrix, T is the translation vector, (X W ,Y W , Z W ) is the coordinate of a point on the object in the world coordinate system, and (u,v) is the coordinate of the point on the image plane.

4. The method for preventing the outflow of heavily aerated products according to claim 3, characterized in that: Initialize first, assuming that all points of the object are at the same depth, that is, the Z of all points c The values ​​are approximately equal, and the calculation is simplified; then iterative optimization is performed to make the projection coordinates (u, V) as close as possible to the actual observed image coordinates by iteratively optimizing the rotation matrix R and the translation vector T; In each iteration, the image coordinates are calculated using the perspective projection model and the parameters are adjusted according to the error; Finally, a convergence judgment is performed. When the iteration result reaches a certain accuracy or the number of iterations reaches a preset value, the iteration is stopped and the final rotation matrix R and translation vector T are output.

5. The method for preventing the outflow of heavily aerated products according to claim 1, characterized in that: The calculation logic of the PIT camera is based on the fact that pores have characteristics different from those of conventional PIT, and pores are generally circular.

Images