Method for manufacturing a single crystal

Abstract

This invention provides a method for manufacturing a single crystal, which accurately measures the distance between the seed crystal and the molten surface, thereby reducing the deviation of the preheating position of the seed crystal. The method for manufacturing the single crystal includes the following steps: measuring the distance (L) between the lower end of the seed crystal (5) positioned above the molten surface and the molten surface (2a). S Based on the interval, the seed crystal (5) is lowered and placed in the melt; while maintaining contact with the melt, the seed crystal (5) is pulled up to grow a single crystal. At the measurement interval (L) S In the step of ), a camera (20) set obliquely above the molten surface (2a) is used to photograph the seed crystal (5) and the molten surface (2a). The edge pattern of the lower end of the straight part (5a) of the real image (5R) of the seed crystal (5) photographed in the image is approximated by a circle to generate a real image edge approximation circle. Furthermore, the edge pattern of the upper end of the straight part (5a) of the mirror image (5M) of the seed crystal (5) reflected in the molten surface (2a) is approximated by a circle to generate a mirror image edge approximation circle. The distance (L) between the lower end of the seed crystal (5) and the molten surface (2a) is calculated based on the distance from the center coordinates of the real image edge approximation circle to the center coordinates of the mirror image edge approximation circle. s ).

Description

Technical Field

[0001] This invention relates to a method for manufacturing single crystals using the Czeklaussky process (CZ process), and particularly to a method for measuring the distance between the seed crystal and the melt surface, as well as a method for preheating the seed crystal using this method.

[0002] Single-crystal silicon, used as a substrate material for semiconductor devices, is mostly manufactured using the CZ (Crystallization-Zeroing) method. In the CZ method, polycrystalline silicon raw material is heated in a quartz crucible to generate a molten solution. A seed crystal is then attached to the molten solution, and while rotating the seed crystal and the molten solution, the seed crystal is slowly pulled up, thereby growing a large-diameter single crystal at the lower end of the seed crystal. The CZ method can improve the manufacturing yield of large-diameter single crystals.

[0003] Regarding the method for manufacturing single crystals using the CZ method, for example, Patent Document 1 describes: before single crystal growth, an optical camera is used to photograph the seed crystal, the image obtained by the camera is processed and the position of the seed crystal is detected, the leading edge of the seed crystal is stopped at a reference position set above the raw material melt, the distance from the reference position to the surface of the raw material melt is detected, and the crucible containing the raw material melt is moved up and down according to the detected distance.

[0004] Furthermore, Patent Document 2 describes a method where, in order to suppress dislocation of the seed crystal caused by thermal shock during immersion, the seed crystal is preheated above the molten surface to minimize the temperature difference between the two before immersion. It also describes a method for accurately measuring the distance between the molten material surface and the seed crystal by obtaining the position information of a specific point at the lower end of the seed crystal (i.e., the lower end point of the real image) and the position information of the point corresponding to the lower end point of the real image in the mirror image of the seed crystal reflected in the molten surface (i.e., the mirror point). At the point where the position of the lower end point of the real image coincides with the position of the mirror point, the distance between the molten material surface and the lower end of the seed crystal is set to 0, thus determining the distance between the molten material surface and the lower end of the seed crystal. Additionally, Patent Documents 1, 3, and 4 describe the use of a cone-shaped seed crystal with a sharp tip to suppress dislocation of the seed crystal caused by thermal shock during immersion.

[0005] Patent document 5 describes a method for reducing the preheating position deviation of the seed crystal and increasing the dislocation-free liquid deposition rate of the seed crystal. This involves using a camera to capture an image from an oblique angle, containing both a real image of the seed crystal and a mirror image of the seed crystal reflected in the molten surface. The method then determines the position of the lower end of the straight portion of the real image of the seed crystal (i.e., the lower endpoint of the real image) and the position of the upper end of the straight portion of the mirror image of the seed crystal (i.e., the mirror point). At the point where the lower endpoint of the real image and the mirror point coincide, the distance between the molten surface and the lower end of the straight portion of the seed crystal is set to 0. The distance between the molten surface and the lower end of the straight portion of the seed crystal is then calculated. Finally, the length of the cone portion of the seed crystal is subtracted from the distance between the molten surface and the lower end of the straight portion of the seed crystal to determine the distance between the molten surface and the lower end of the seed crystal.

[0006] Existing technical documents

[0007] Patent documents

[0008] Patent Document 1: Japanese Patent Application Publication No. 2005-170773

[0009] Patent Document 2: Japanese Patent Application Publication No. 2016-155729

[0010] Patent Document 3: Japanese Patent Application Publication No. 2000-128691

[0011] Patent Document 4: Japanese Patent Application Publication No. 2009-234889

[0012] Patent Document 5: Japanese Patent Application Publication No. 2019-214486 Summary of the Invention

[0013] The technical problem that the invention aims to solve

[0014] When controlling the preheating position of the seed crystal based on its height direction determined from images captured by the camera, it is preferable to keep the seed crystal as close to the molten surface as possible without contacting the molten metal. However, since the camera captures images of the seed crystal and the molten surface from an oblique angle, when using a cone-shaped seed crystal, the tip of the cone will be hidden inside the seed crystal, making it impossible to capture the tip of the cone shape with the camera. Furthermore, since it is impossible to capture the tip of the cone from the mirror image of the seed crystal reflected on the molten surface, it is also difficult to detect the position of the lower end of the seed crystal based on the mirror image. In addition, since there may be manufacturing deviations in the cone shape of the seed crystal, these manufacturing deviations must also be taken into account when controlling the seed crystal position.

[0015] Patent document 5 describes a method for determining the distance from the tip of a cone-shaped seed crystal to the molten surface. However, the accuracy of the calculated distance from the tip of the seed crystal to the molten surface is insufficient, and more effort must be put into calculating this distance more accurately to further improve the dislocation-free liquid deposition rate of the seed crystal.

[0016] Therefore, the object of the present invention is to provide a method for manufacturing a single crystal that can accurately measure the distance between the lower end of the seed crystal and the molten surface, even when using a seed crystal with a conical shape at the lower end. As a result, when preheating is performed above the molten surface, the deviation of the preheating position can be reduced and the dislocation-free liquid deposition rate of the seed crystal can be improved.

[0017] Solutions for solving technical problems

[0018] To address the aforementioned issues, the single crystal manufacturing method of the present invention utilizes the Czeklaussky method, characterized by comprising the following steps: measuring the distance between the lower end of a seed crystal disposed above a molten surface and the molten surface; lowering the seed crystal based on the distance to contact the molten surface; and while maintaining contact with the molten surface, lifting the seed crystal to grow a single crystal at its lower end. In the step of measuring the distance between the lower end of the seed crystal and the molten surface, a camera positioned obliquely above the molten surface is used to photograph the seed crystal and the molten surface. A circular approximation is performed on the approximately arc-shaped edge pattern of the lower end of the real image of the seed crystal captured in the camera image to generate a real image edge approximation circle. Furthermore, a circular approximation is performed on the approximately arc-shaped edge pattern of the upper end of the mirror image of the seed crystal reflected in the molten surface to generate a mirror image edge approximation circle. The distance between the lower end of the seed crystal and the molten surface is calculated based on the distance from the center coordinates of the real image edge approximation circle to the center coordinates of the mirror image edge approximation circle.

[0019] According to the present invention, even if the lower end of the seed crystal is tapered, the distance between the lower end of the seed crystal and the molten surface can be accurately determined. Therefore, when the seed crystal is preheated above the molten surface, the deviation of the preheating position can be reduced, thereby increasing the dislocation-free liquid deposition rate of the seed crystal.

[0020] In the single-crystal manufacturing method of the present invention, it is preferable to multiply half the number of pixels from the center coordinates of the approximate circle at the edge of the real image to the center coordinates of the approximate circle at the edge of the mirror image by a conversion factor, thereby converting it into a distance unit in real space. This allows for the accurate determination of the distance between the lower end of the seed crystal and the molten surface in real space.

[0021] In the method for manufacturing a single crystal according to the present invention, it is preferable to further include the following step: before measuring the distance between the lower end of the seed crystal and the molten surface, the conversion coefficient is calculated based on the change in the pixel position of the seed crystal in the captured image when the position of the seed crystal is moved a certain distance in the vertical direction. This allows for an accurate determination of the distance between the lower end of the seed crystal and the molten surface in real space.

[0022] In the single crystal manufacturing method of the present invention, the distance between the lower end of the seed crystal and the molten surface can be calculated after transforming the coordinate projections of the captured images of the approximate real image edge circle and the approximate mirror image edge circle into coordinates in real space based on the camera's setting angle and focal length. Thus, the distance between the lower end of the seed crystal and the molten surface can be calculated without prior calibration to determine the conversion factors.

[0023] In the single crystal manufacturing method of the present invention, it is preferable to set a reference plane at the same height position as the lower end of the seed crystal, project and transform the approximate circle of the real image edge and the approximate circle of the mirror image edge onto the reference plane, and calculate the distance between the lower end of the seed crystal and the molten surface based on the center coordinates of the approximate circle of the real image edge, the center coordinates of the approximate circle of the mirror image edge, and the center coordinates of the camera lens. Thus, the distance between the lower end of the seed crystal and the molten surface can be easily calculated using the coordinates after projection transformation.

[0024] In this invention, preferably, in the step of measuring the distance between the lower end of the seed crystal and the molten surface, after determining the approximate circle of the mirror edge based on the captured image, the approximate circle of the real image edge located above the approximate circle of the mirror edge is determined. This allows for efficient and effective determination of the positions of both the real and mirror images of the seed crystal in the captured image.

[0025] In this invention, preferably, the step of determining the approximate circle of the mirror edge includes the following steps: determining the upper position of the image of the seed crystal based on the brightness distribution of a first region predetermined in the captured image; defining a second region containing the upper position of the image of the seed crystal, binarizing the second region and detecting the edge pattern of the straight portion of the image of the seed crystal; and approximating the edge pattern of the straight portion of the image of the seed crystal with a circle to determine the approximate circle of the real image edge includes the following steps: determining the position where the differential value of the brightness exceeds a predetermined threshold when scanning the first region upward from the center coordinates of the approximate circle of the mirror edge as the lower position of the real image of the seed crystal; defining a third region containing the lower position of the real image of the seed crystal, binarizing the third region and detecting the edge pattern of the straight portion of the real image of the seed crystal; and approximating the edge pattern of the straight portion of the real image with a circle. Thus, the center coordinates of the approximate circle of the real image edge and the center coordinates of the approximate circle of the mirror edge in the captured image can be determined efficiently.

[0026] In this invention, preferably, in the step of measuring the distance between the lower end of the seed crystal and the molten surface, multiple real and mirror images of the seed crystal at the same height position are continuously captured. The distance between the lower end of the seed crystal and the molten surface is calculated based on the average value obtained from each of the multiple images. This improves the accuracy of the measurement of the distance between the lower end of the seed crystal and the molten surface.

[0027] In this invention, preferably, the seed crystal further includes a conical portion disposed below the straight portion, the lower end of the seed crystal being the lower end of the conical portion, and the cone angle of the conical portion being larger than the camera's setting angle. Before measuring the distance between the lower end of the seed crystal and the molten surface, a step of pre-measuring the length of the conical portion is included. The step of measuring the distance between the lower end of the seed crystal and the molten surface includes the following step: subtracting the length of the conical portion from the distance between the lower end of the straight portion of the seed crystal and the molten surface. When the cone angle of the seed crystal is greater than the camera's setting angle, it is impossible to photograph the front end of the seed crystal with the camera, and the position of the front end of the seed crystal cannot be directly determined from the photographed image. However, according to this invention, the distance between the seed crystal and the molten surface can be accurately measured even without photographing the front end of the seed crystal.

[0028] In the method for manufacturing a single crystal according to the present invention, it is preferable to further include the following steps: after measuring the distance between the lower end of the seed crystal and the molten surface and before placing the seed crystal in the molten surface, adjusting the height position of at least one of the seed crystal and the crucible supporting the molten surface so that the distance becomes a target value; and preheating the seed crystal by keeping it stationary at the target value position. According to the present invention, when preheating a seed crystal with a tapered front end above the molten surface, the deviation of the preheating position can be reduced, thereby increasing the dislocation-free liquid adhesion rate of the seed crystal.

[0029] Furthermore, the present invention provides a method for measuring the distance between the lower end of a seed crystal disposed above a molten surface and the molten surface. The method is characterized by using a camera positioned obliquely above the molten surface to photograph the seed crystal and the molten surface; approximating the approximately arc-shaped edge pattern of the lower end of the real image of the seed crystal captured in the camera image to generate a real image edge approximation circle; and approximating the approximately arc-shaped edge pattern of the upper end of the mirror image of the seed crystal reflected in the molten surface to generate a mirror image edge approximation circle; and calculating the distance between the lower end of the seed crystal and the molten surface based on the number of pixels from the center coordinates of the real image edge approximation circle to the center coordinates of the mirror image edge approximation circle.

[0030] According to the present invention, even if the lower end of the seed crystal is tapered, the distance between the lower end of the seed crystal and the molten surface can be accurately determined. Therefore, when the seed crystal is preheated above the molten surface, the deviation of the preheating position can be reduced, thereby increasing the dislocation-free liquid deposition rate of the seed crystal.

[0031] Furthermore, the preheating method for the seed crystal of the present invention is characterized by comprising the following steps: measuring the interval between the lower end of the seed crystal and the surface of the molten liquid using the above-described interval measurement method; adjusting the height position of at least one of the seed crystal and the crucible supporting the molten liquid so that the interval becomes a target value; and preheating the seed crystal by keeping it stationary at the target value position.

[0032] According to the present invention, when a seed crystal with a tapered front end is preheated above the molten surface, the deviation of the preheating position can be reduced and the dislocation-free liquid deposition rate of the seed crystal can be increased.

[0033] Furthermore, the single crystal manufacturing apparatus of the present invention is characterized by comprising: a crucible supporting molten metal; a heater for heating the molten metal; a crystal pulling mechanism for lifting and lowering a seed crystal disposed above the molten metal; a camera for capturing images of the seed crystal and the molten metal surface from an obliquely upward position; an image processing unit for processing the images captured by the camera; and a control unit for controlling the crystal pulling mechanism based on the processing results of the image processing unit. The image processing unit performs a circular approximation on the lower end of the real image of the seed crystal captured by the camera to generate a real image edge approximation circle, and performs a circular approximation on the upper end of the mirror image of the seed crystal reflected in the molten metal surface to generate a mirror image edge approximation circle. The distance between the lower end of the seed crystal and the molten metal surface is calculated based on the number of pixels from the center coordinates of the real image edge approximation circle to the center coordinates of the mirror image edge approximation circle.

[0034] According to the present invention, even if the lower end of the seed crystal is tapered, the distance between the lower end of the seed crystal and the molten surface can be accurately determined. Therefore, when the seed crystal is preheated above the molten surface, the deviation of the preheating position can be reduced, thereby increasing the dislocation-free liquid deposition rate of the seed crystal.

[0035] Invention Effects

[0036] According to the present invention, a method for manufacturing a single crystal can be provided, which can accurately determine the distance between the lower end of the seed crystal and the molten surface even if the lower end of the seed crystal has a conical shape. Attached Figure Description

[0037] Figure 1 This is a side cross-sectional view schematically illustrating the structure of a single crystal manufacturing apparatus according to an embodiment of the present invention.

[0038] Figure 2This is a flowchart illustrating the manufacturing process of monocrystalline silicon.

[0039] Figure 3 This is a schematic cross-sectional view showing the shape of a single-crystal silicon ingot.

[0040] Figure 4 It is a side cross-sectional view schematically showing the arrangement of seed crystals in a single crystal manufacturing apparatus during the preheating process.

[0041] Figure 5 This diagram illustrates the method for measuring the distance between the lower end of the seed crystal and the molten surface, specifically showing the positional relationship between the real and mirror images of the seed crystal.

[0042] Figure 6 Figures (a) to (c) are used to illustrate the positional relationship between the seed crystal and the melt surface. Figure 6 (a) is a schematic diagram when the distance from the seed crystal to the melt surface is far. Figure 6 (b) is a schematic diagram when the distance from the seed crystal to the melt surface is close. Figure 6 (c) is a graph showing the relationship between the height position of the seed and the number of pixels between the real and mirror images.

[0043] Figure 7 (a) to (c) are used to describe the number of pixels N between the real image and the mirror image. D The diagram for finding the method.

[0044] Figure 8 It is used to explain in detail the coordinates P of the real image center of seed crystal 5 in the captured image. R and the center coordinates P of the mirror image M The flowchart of the detection method.

[0045] Figure 9 This is a schematic diagram illustrating a method for projecting two-dimensional coordinates of a captured image into coordinates in real space.

[0046] Figure 10 This is a schematic diagram illustrating a method for calculating the seed-molten surface spacing based on the center coordinates of the circular outer periphery pattern at the lower end of the real image and mirror image of the seed crystal. Detailed Implementation

[0047] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0048] Figure 1 This is a side cross-sectional view schematically illustrating the structure of a single crystal manufacturing apparatus according to an embodiment of the present invention.

[0049] like Figure 1As shown, the single crystal manufacturing apparatus 1 includes: a water-cooled chamber 10; a quartz crucible 11 holding molten silicon 2 within the chamber 10; a graphite crucible 12 holding the quartz crucible 11; a rotation shaft 13 supporting the graphite crucible 12; a crucible drive mechanism 14 that drives the quartz crucible 11 to rotate and move up and down via the rotation shaft 13 and the graphite crucible 12; a heater 15 disposed around the graphite crucible 12; and an insulating layer disposed along the inner surface of the chamber 10 on the outside of the heater 15. The apparatus includes a heat-generating material 16, a heat insulation body 17 disposed above the quartz crucible 11, a crystal pulling shaft (i.e., wire) 18 located above the quartz crucible 11 and disposed on the same axis as the rotation axis 13, a crystal pulling mechanism 19 disposed above the chamber 10, a camera 20 disposed outside the chamber 10 and taking pictures of the chamber 10 through an observation window 10e, an image processing unit 21 that processes the pictures taken by the camera 20, and a control unit 22 that controls each part of the single crystal manufacturing apparatus 1.

[0050] The chamber 10 consists of a main chamber 10a and a slender cylindrical crystal pulling chamber 10b connected to the upper opening of the main chamber 10a. A quartz crucible 11, a graphite crucible 12, a heater 15, and a heat insulation body 17 are disposed within the main chamber 10a. The crystal pulling chamber 10b is provided with a gas inlet 10c for introducing inactive gases (purification gases) such as argon or doping gases into the chamber 10. A gas outlet 10d is provided at the lower part of the main chamber 10a for discharging the atmospheric gases from the chamber 10. In addition, an observation window 10e is provided at the upper part of the main chamber 10a, through which the growth status of the single crystal silicon 3 can be observed.

[0051] The quartz crucible 11 is a container made of quartz glass with cylindrical side walls and a curved bottom. The graphite crucible 12, in order to maintain the shape of the quartz crucible 11 after it has softened by heating, is held in place by tightly sealing it against the outer surface of the quartz crucible 11. The quartz crucible 11 and the graphite crucible 12 together form a double-structured crucible within the chamber 10 that supports the molten silicon.

[0052] The graphite crucible 12 is fixed to the upper end of the rotating shaft 13, and the lower end of the rotating shaft 13 passes through the bottom of the chamber 10 and is connected to the crucible drive mechanism 14 located outside the chamber 10. The graphite crucible 12, the rotating shaft 13, and the crucible drive mechanism 14 constitute the rotating mechanism and lifting mechanism of the quartz crucible 11. The rotation and lifting actions of the quartz crucible 11 driven by the crucible drive mechanism 14 are controlled by the control unit 22.

[0053] Heater 15 melts the silicon raw material filled in the quartz crucible 11 to generate molten silicon 2, and also maintains the molten silicon 2 in a molten state. Heater 15 is a carbon-based resistance heater, arranged to surround the quartz crucible 11 within the graphite crucible 12. Furthermore, an insulating material 16 is provided on the outside of heater 15 to surround it, thereby improving the heat insulation within the chamber 10. The output of heater 15 is controlled by control unit 22.

[0054] The heat shield 17 is provided to suppress temperature fluctuations in the molten silicon 2, provide appropriate heat distribution near the crystal growth interface, and prevent heating of the single-crystal silicon 3 caused by radiant heat from the heater 15 and the quartz crucible 11. The heat shield 17 is a generally cylindrical graphite component, provided to cover the area above the molten silicon 2 except for the pulling path of the single-crystal silicon 3.

[0055] The diameter of the opening at the lower end of the heat insulation body 17 is larger than the diameter of the monocrystalline silicon 3, thereby ensuring the pulling path of the monocrystalline silicon 3. In addition, the outer diameter of the lower end of the heat insulation body 17 is smaller than the diameter of the quartz crucible 11. Since the lower end of the heat insulation body 17 is located inside the quartz crucible 11, even if the upper edge of the quartz crucible 11 is raised above the lower end of the heat insulation body 17, the heat insulation body 17 will not interfere with the quartz crucible 11.

[0056] As the monocrystalline silicon 3 grows, the amount of molten silicon in the quartz crucible 11 decreases. However, by raising the quartz crucible 11 to maintain a constant gap (spacing) between the molten silicon surface 2a and the heat insulation body 17, the temperature fluctuation of the molten silicon 2 is suppressed, and the flow rate of the gas flowing near the molten silicon surface 2a is kept constant, thereby controlling the evaporation of dopants from the molten silicon 2. Through such spacing control, the stability of crystal defect distribution, oxygen concentration distribution, resistivity distribution, etc., in the pulling axis direction of the monocrystalline silicon 3 can be improved.

[0057] Above the quartz crucible 11, a wire 18 serving as a pulling shaft for the monocrystalline silicon 3, and a crystal pulling mechanism 19 for pulling the monocrystalline silicon 3 by winding the wire 18, are arranged. The crystal pulling mechanism 19 has the function of rotating the monocrystalline silicon 3 together with the wire 18. The crystal pulling mechanism 19 is controlled by a control unit 22. The crystal pulling mechanism 19 is positioned above the crystal pulling chamber 10b, and the wire 18 extends downward from the crystal pulling mechanism 19 through the crystal pulling chamber 10b, with the leading end of the wire 18 reaching the internal space of the main chamber 10a. Figure 1 The diagram shows the monocrystalline silicon 3 suspended by wire 18 during its growth process. During the pulling of the monocrystalline silicon 3, the wire 18 is slowly pulled while the quartz crucible 11 and the monocrystalline silicon 3 are rotated, thereby growing the monocrystalline silicon 3. The crystal pulling speed is controlled by the control unit 22.

[0058] A camera 20 is disposed on the outside of the chamber 10. The camera 20 is, for example, a CCD camera, which takes pictures of the interior of the chamber 10 through an observation window 10e formed in the chamber 10. The camera 20 is disposed at a predetermined angle to the vertical direction (preferably 20 to 30 degrees), and the camera 20 has an optical axis that is tilted relative to the lifting axis of the single crystal silicon 3. That is, the camera 20 takes pictures of the upper surface area of ​​the quartz crucible 11, which includes the circular opening of the heat insulation body 17 and the liquid surface of the silicon melt 2, from an obliquely upward position.

[0059] Camera 20 is connected to image processing unit 21, which in turn is connected to control unit 22. Image processing unit 21 calculates the crystal diameter near the solid-liquid interface based on the outline pattern of the single crystal captured in the image taken by camera 20. Furthermore, it calculates the distance, i.e., the spacing, from the lower end of heat insulation 17 to the melt surface 2a based on the position of the mirror image of the heat insulation 17 reflected in the captured image. The method for calculating the spacing is not particularly limited; for example, a conversion formula can be prepared beforehand to approximate the relationship between the position of the mirror image of the heat insulation 17 and the spacing as a straight line. During the crystal pulling process, the position of the mirror image of the heat insulation is substituted into this conversion formula to calculate the spacing. Alternatively, the spacing can be calculated geometrically based on the positional relationship between the real image and the mirror image of the heat insulation 17 captured in the captured image.

[0060] The control unit 22 controls the crystal pulling speed based on the crystal diameter data obtained from the image captured by the camera 20, thereby controlling the crystal diameter. Specifically, when the measured crystal diameter is greater than the target diameter, the crystal pulling speed is increased; when the measured crystal diameter is less than the target diameter, the pulling speed is decreased. Furthermore, the control unit 22 controls the movement amount (crucible rising speed) of the quartz crucible 11 based on the crystal length data of the single-crystal silicon 3 obtained from the sensor of the crystal pulling mechanism 19 and the crystal diameter data obtained from the image captured by the camera 20.

[0061] Figure 2 This is a flowchart illustrating the manufacturing process of monocrystalline silicon 3. Furthermore, Figure 3 This is a schematic cross-sectional view showing the shape of a single-crystal silicon ingot.

[0062] As shown in Figure 2, the manufacturing process of the single crystal silicon 3 in this embodiment includes: a raw material melting process S11, in which the silicon raw material in the quartz crucible 11 is heated by the heater 15 to generate silicon melt 2; a preheating process S12, in which the seed crystal mounted on the front end of the wire 18 is preheated before being immersed in the melt; an immersion process S13, in which the seed crystal is lowered and immersed in the silicon melt 2; and a crystal pulling process (S14 to S17), in which the seed crystal is slowly pulled up while maintaining contact with the silicon melt 2 to grow a single crystal.

[0063] In the crystal pulling process (S14~S17), the following steps are performed sequentially: the necking process S14, which forms a neck 3a with a thinned crystal diameter to prevent dislocation; the shoulder growth process S15, which forms a shoulder 3b with a gradually increasing crystal diameter as the crystal grows; the body growth process S16, which forms a body 3c with a constant crystal diameter; and the tail growth process S17, which forms a tail 3d with a gradually decreasing crystal diameter as the crystal grows.

[0064] Next, a cooling process S18 is performed, in which the monocrystalline silicon 3 is cut away from the molten surface to promote cooling. As described above, the process is complete. Figure 3 The single-crystal silicon ingot 3I shown has a neck 3a, a shoulder 3b, a body 3c, and a tail 3d.

[0065] Figure 4 It is a side cross-sectional view schematically showing the arrangement of seed crystals in the single crystal manufacturing apparatus 1 during the preheating process S12.

[0066] like Figure 4 As shown, the preheating process S12 is as follows: the seed crystal 5 is held above the molten silicon surface 2a for a certain period of time and then heated. If the seed crystal 5, which has a large temperature difference relative to the molten silicon 2, is subjected to liquid bonding, thermal shock-induced slip dislocations will occur. However, with the preheating process S12, dislocation formation can be suppressed, thereby increasing the dislocation-free liquid bonding rate of the seed crystal. If the seed crystal 5 is far from the molten silicon surface 2a during the preheating process S12, it cannot be heated sufficiently, and the temperature difference between the seed crystal 5 and the molten silicon 2 cannot be reduced. Therefore, it is necessary to control the distance between the lower end of the seed crystal 5 and the molten silicon surface 2a to less than 5 mm. Therefore, it is necessary to accurately measure the distance between the lower end of the seed crystal 5 and the molten silicon surface 2a.

[0067] Figure 5 This diagram illustrates the method for measuring the distance between the lower end of the seed crystal 5 and the melt surface 2a, specifically showing the positional relationship between the real image 5R and the mirror image 5M of the seed crystal 5.

[0068] like Figure 5 As shown, the seed crystal 5 of this embodiment has a straight body portion 5a of fixed thickness and a tapered portion 5b extending downward from the lower end of the straight body portion 5a. The diameter of the straight body portion 5a is preferably 6 to 12 mm, and the length of the straight body portion 5a is preferably 30 to 80 mm. Furthermore, the length of the tapered portion 5b is preferably 5 to 10 mm. Such a seed crystal 5 is manufactured as follows: after cutting a cylindrical single-crystal silicon rod into a predetermined length, it is ground to make the front end sharp, and then etched to remove the processing damage. The tapered shape can be a cone shape or a frustum shape with a flat front end.

[0069] The cone angle θ1 of the seed crystal 5 is larger than the setting angle θ0 of the camera 20. When the front end of the seed crystal 5 has such a cone shape, the front end P0 (the lower end of the cone 5b) of the seed crystal 5 enters the blind spot of the camera 20, therefore, the camera 20 cannot capture the front end P0 of the seed crystal 5. Therefore, the length L of the cone 5b must be accurately measured in advance. T The distance L between the lower end P1 of the straight part 5a of the seed crystal 5 and the molten surface 2a, which can be captured by the camera 20, is calculated. P1 Then, from that interval L P1 Subtract the length L of the cone portion 5b T Therefore, the distance L between the front end P0 of seed crystal 5 and the melt surface 2a can be determined. S =L P1 -L T .

[0070] Here, the distance L between the lower end P1 of the straight portion 5a of the seed crystal 5 and the molten surface 2a is... P1 L is the distance (distance between real and mirror images) from the lower end P1 of the straight portion 5a of the real image 5R of seed crystal 5 to its mirror image P2. D The half value of . Therefore, by calculating the distance L between the real and mirror images. D It is possible to determine the distance L between the front end P0 of seed crystal 5 and the melt surface 2a. S .

[0071] For example, it is preferable to accurately measure the length L of the cone 5b using a non-contact measurement method. T When the length of the cone 5b is measured in advance, it is preferable to measure the length of the cone 5b at multiple locations in the circumference and use the average of the multiple measurements. This is because, depending on the machining accuracy of the cone 5b, the length of the cone 5b may deviate from its circumferential position.

[0072] Details will be discussed later, but the distance L between the real and mirror images in real space... D It is possible to measure the distance L between the real image and the mirror image D The equivalent number of pixels in a captured image (N pixels between real and mirror images) D The number of pixels N between the real and mirror images is calculated by multiplying the real image by the specified conversion factor α. D The center coordinates (real image center coordinates P) can be used as the center coordinates of the outer periphery of the circle at the lower end of the straight body portion 5a of the real image 5R of the seed crystal 5. R (X R Y R The center coordinates (mirror center coordinates P) of the outer periphery of the circular part 5a at the upper end of the straight body portion 5a, which is the mirror image of seed 5M. M (X M Y M The number of pixels is used to determine this.

[0073] The conversion factor α can be determined using a calibration process, in which the relationship between the relative displacement (mm) in real space and the relative displacement (number of pixels) in the captured image when the seed crystal 5 is moved a certain distance in the vertical direction is calculated. During calibration, firstly, while moving the seed crystal 5 in the vertical direction, images containing the seed crystal 5 and the molten surface 2a are captured by the camera 20 at multiple height positions of the seed crystal 5. These multiple height positions of the seed crystal 5 only need to be relative height positions, not absolute height positions. Therefore, the multiple height positions of the seed crystal 5 relative to the molten surface 2a can also be set by fixing the height position of the seed crystal 5 and changing the height position of the molten surface 2a (the height position of the crucible).

[0074] Figure 6 Figures (a) to (c) are used to illustrate the positional relationship between the seed crystal and the melt surface. Figure 6 (a) is a schematic diagram when the distance from the seed crystal to the melt surface is far. Figure 6 (b) is a schematic diagram when the distance from the seed crystal to the melt surface is close. Figure 6 (c) is a graph showing the relationship between the height position of the seed and the number of pixels between the real and mirror images.

[0075] like Figure 6 As shown in (a) and (b), the real image 5R and mirror image 5M of the seed crystal 5 have a symmetrical positional relationship, sandwiching the molten surface 2a, and this relationship does not change even if the seed crystal 5 is moved in the vertical direction. For example, if the descent amount (wire transport amount) of the seed crystal 5 from the first height position h1 to the second height position h2 is set as Δh = h1 - h2 (mm), then, when the number of pixels between the real and mirror images increases from N... D1 Change to N D2 (N D1 >N D2 When ΔN is reached, the ratio of the change in the number of pixels to the amount of wire transported Δh is used as ΔN. D / Δh=(N D1 -N D2 The value is obtained by using ΔN / (h1-h2). D / Δh is used as the conversion factor α, thereby transforming the change in the height position of seed 5 in the captured image into the change in the height position of seed 5 in real space. For example... Figure 6 As shown in (c), the conversion factor α can be determined based on the multiple height positions h of the seed 5 and the number of pixels N between the corresponding real image mirrors. D The slope of the linear regression line (y = αx + β) is calculated.

[0076] Distance L between real and mirror images DPreferably, the average value of multiple values ​​obtained from continuously capturing multiple images of seed crystal 5 at the same height position is used. The continuous capturing cycle is several hundred ms. This reduces measurement errors caused by fluctuations in the measurement environment and improves measurement accuracy.

[0077] Figure 7 (a) to (c) are used to describe the number of pixels N between the real image and the mirror image. D The diagram for finding the method.

[0078] like Figure 7 As shown in (a) to (c), the distance L between the real and mirror images is D Able to determine the number of pixels N between real and mirror images D Find the number of pixels N between the real image and its mirror image. D The coordinates P of the real image center can be used as the seed crystal 5 in the captured image. R To the center coordinate P of the mirror image M The coordinates of the real image center P of seed 5 are determined by the number of pixels. R and the center coordinates P of the mirror image M It can be used to binarize the captured image and detect the approximately arc-shaped edge pattern of the straight body portion 5a of the real image 5R and mirror image 5M of the seed crystal 5, and to obtain the center coordinates of the approximate circle when approximating the edge pattern as a circle.

[0079] The distance L between the real and mirror images is calculated from the lower end face of the straight portion 5a of the real image 5R of the seed crystal 5 to the upper end face of the straight portion 5a of the mirror image 5M of the seed crystal 5. D Another method is as follows: the distance from the lowest point of the edge pattern of the straight portion 5a of the real image 5R of the seed 5 in the captured image to the lowest point of the edge image of the straight portion 5a of the mirror image 5M of the seed 5 is taken as the distance L between the real and mirror images. D However, in this method, the measurement error is prone to increase due to the influence of changes in the brightness distribution in the captured image, the installation position of the seed crystal 5, or the eccentric rotation caused by tilt deviation. On the other hand, as in this embodiment, the center coordinates of the approximate circle of the edge pattern of the straight part 5a of the real image 5R and the mirror image 5M obtained from the captured image are calculated, and the distance between the center coordinates of the two points is taken as the distance L between the real image and the mirror image. D In this case, it is less affected by changes in brightness distribution and the eccentric rotation of the seed crystal, and can reduce the distance L between real image mirrors. D Measurement error.

[0080] Figure 8 It is used to explain in detail the coordinates P of the real image center of seed crystal 5 in the captured image. R and the center coordinates P of the mirror image M The flowchart of the detection method.

[0081] like Figure 8 As shown, in this embodiment, after the first region containing the real image 5R and the mirror image 5M of the seed crystal 5 is cut out from the image captured by the camera 20 (step S20), the coordinates P of the mirror center are first calculated based on the first region. M (X M Y M (Steps S21-S24) Next, calculate the coordinates P of the real image center. R (X R Y R (Steps S25-S28). The reason for cutting out the first region from the image captured by camera 20 is that the image captured by camera 20 is a large-scale image of the furnace, and the area where seed crystal 5 is captured is only a small part of the captured image, thus limiting the scope of image processing. In addition, the coordinates P of the mirror center are first calculated. M The reason is that the brightness of the image 5M of the seed crystal 5 in the captured image is higher than that of the real image 5R, making it easier to determine its position.

[0082] At the mirror center coordinate P of seed crystal 5 M In the detection (steps S21 to S24), a brightness level of, for example, 99.9% of the brightness peak in the first region is set as a threshold for mirror detection, and the upper position of the mirror 5M of the position of the uppermost white pixel obtained when the first region is binarized using the threshold is determined (step S21).

[0083] Next, using the upper position of the pixel identified as the mirror image 5M as a reference, the image within a certain range below it (e.g., 50 pixels vertically × 100 pixels horizontally) is set as the detailed detection range (second region) of the mirror image 5M (step S22). A brightness level of, for example, 99% of the brightness peak within this second region is set as the threshold for edge detection. Then, the image within the second region is binarized using this threshold, and the black-and-white boundary position of the binary image is determined as the edge of the straight body 5a of the mirror image 5M of the seed crystal 5 (step S23). Afterwards, the roughly arc-shaped edge pattern is approximated as a circle using the least squares method, and the center coordinates of the approximate circle of the mirror edge, i.e., the mirror center coordinates P, are determined. M (Step S24).

[0084] Next, at coordinate P, the center of the real image of seed 5 R In the detection, based on the approximate detection range (region 1) of the real image 5R and its mirror image 5M of seed 5, the detailed detection range (region 3) of the real image 5R is determined. Since it is known that the real image 5R of seed 5 exists above the mirror image 5M, the coordinates P of the mirror image center of seed 5 are used as the basis for the detection. MBased on this, the first region is scanned upwards, and the position where the differential value of the brightness level exceeds a predetermined threshold is determined as the lower position of the real image 5R (step S25). Based on the lower position of the real image 5R thus detected, the image within a certain range above it (e.g., 50 pixels vertically × 100 pixels horizontally) is set as the third region (step S26), and the brightness level of the brightness peak in the third region, for example, 99%, is set as the threshold for edge detection. Then, the image in the third region is binarized using this threshold, and the black and white boundary position of the binary image is obtained as the edge of the straight part 5a of the real image 5R of the seed crystal 5 (step S27). Afterwards, the roughly arc-shaped edge pattern is approximated as a circle using the least squares method, and the center coordinates of the approximate circle of the real image edge, i.e., the center coordinates P of the real image, are obtained. R (Step S28).

[0085] Since the lower end of the seed crystal 5 is close to the molten surface 2a, when the lower end of the real image 5R of the seed crystal 5 overlaps with the upper end of the mirror image 5M, the mirror image 5M of the seed crystal 5 becomes an interference to the real image 5R, making it difficult to separate the real image 5R from the mirror image 5M. In this case, by performing an opening process (dilation / contraction process) on the binary image, it is possible to separate the mirror image 5M from the real image 5R and determine the edge of the real image 5R.

[0086] Thus, the coordinates P of the real image center of seed crystal 5 are obtained respectively. R and the center coordinates P of the mirror image M Then, it is possible to determine the number of pixels N between two points. D Find the distance L between the real and mirror images of seed crystal 5. D .

[0087] In the preheating process S12 of seed crystal 5, the distance L between the melt surface 2a and seed crystal 5 is measured using the method described above. S Next, the seed crystal 5 is moved in the direction that minimizes the difference between the measured value and the target value. That is, the control unit 22 operates the crystal lifting mechanism 19 to adjust the height of the seed crystal 5 or operates the crucible driving mechanism 14 to adjust the height of the quartz crucible 11, thereby eliminating the difference between the measured value and the target value. This allows the actual distance L between the molten surface 2a and the seed crystal 5 to be reduced. S Set the interval as the target.

[0088] In the preheating process S12 of the seed crystal 5, the distance (target value) between the molten surface 2a and the lower end of the seed crystal 5 is preferably 5 mm or less, and more preferably 3 mm or less. This ensures a sufficiently small temperature difference between the seed crystal and the molten liquid, thereby increasing the dislocation-free liquid adhesion rate of the seed crystal. On the other hand, the initial distance between the molten surface 2a and the seed crystal 5 before measuring the distance between them is preferably 5 mm or more. This is because, without accurately knowing the distance between the molten surface 2a and the seed crystal 5, if the seed crystal 5 is placed too close to the molten surface 2a, liquid adhesion may occur.

[0089] In the preheating process S12 of the seed crystal 5, the seed crystal 5 is maintained close to the molten silicon surface 2a for, for example, several minutes to several hours. As a result, the seed crystal 5 is heated by radiant heat from the molten silicon 2. By preheating the seed crystal 5 at a predetermined position above the molten silicon surface 2a each time, the reproducibility of the temperature of the preheated seed crystal 5 can be improved. Furthermore, the target interval is set to be sufficiently small, thereby minimizing the temperature difference between the two and reducing the thermal shock when the seed crystal 5 touches the molten silicon. Therefore, the introduction of dislocations into the seed crystal 5 can be suppressed.

[0090] Furthermore, in existing methods where the operator visually adjusts the distance between the molten surface 2a and the lower end of the seed crystal while preheating the seed crystal 5, sometimes the seed crystal 5 unintentionally comes into contact with the molten silicon 2 when the target distance is set to a very small value, such as 3 mm. However, as in this embodiment, by accurately measuring the distance between the molten surface 2a and the lower end of the seed crystal 5, contact between the seed crystal 5 and the molten surface 2a can be avoided.

[0091] After preheating the seed crystal 5 in this way, the seed crystal 5 is lowered and placed in the molten silicon 2. Then, the seed crystal 5 is raised to grow single crystal silicon 3.

[0092] As explained above, in the single crystal manufacturing method of this embodiment, the edge pattern of the real image and the mirror image of the seed crystal 5 is approximated as a circle, and the center coordinates P of the approximate circle of the real image edge are used as the basis for the method. R The center coordinates P of the circle approximating the mirror edge M Given the interval, calculate the distance L between the real and mirror images of seed crystal 5. D Therefore, even when the seed crystal 5 has a cone portion 5b, the distance between the lower end of the seed crystal 5 and the molten surface 2a can be accurately determined.

[0093] Next, other methods for measuring the distance between the lower end of the seed crystal 5 and the molten surface 2a will be explained. The above-described seed-to-molten surface distance measurement method requires converting the distance between the real and mirror images of the seed crystal in the image to the distance between the real and mirror images of the seed crystal in real space during calibration. Therefore, calibration is required to determine the conversion factor α.

[0094] Therefore, in this embodiment, for the furnace image captured by the camera, the coordinate points in the image are projected and transformed into coordinates in real space. Thus, the seed crystal level interval can be calculated directly without calibration. Specifically, an image of the molten surface 2a inside the furnace is captured, and the coordinates of the real image of the seed crystal 5 in the captured image and its mirror image reflected in the molten surface 2a are detected. The coordinates of both the real image and its mirror image are projected and transformed into coordinates in real space. Based on the real space coordinates of both the real image and its mirror image, the interval ΔG between the seed crystal and the molten surface is calculated. The method for projecting and transforming the two-dimensional coordinates of the captured image into coordinates in real space will be described in detail below.

[0095] Figure 9 This is a schematic diagram illustrating a method for projecting and transforming the two-dimensional coordinates of a captured image into coordinates in real space.

[0096] like Figure 9 As shown in the left-hand diagram, the camera 20 captures an image of the cavity from an obliquely upward angle. Therefore, the circular outer periphery of the lower end of the straight portion of the seed crystal 5 is distorted in the captured image. To accurately calculate the dimensions of both the real and mirror images of the seed crystal 5, image distortion correction is necessary. Therefore, the coordinate projection of the image captured by the camera 20 is transformed to coordinates on a reference plane set at the same height as the lower end of the straight portion of the seed crystal 5 to correct the distortion.

[0097] Figure 9 The diagram on the right illustrates the coordinate system used for image correction. In this coordinate system, the reference plane is set as the xy plane. Furthermore, the origin C0 of the XY coordinates extends from the center position C of the imaging device 20a of the camera 20 to the center position F (0, y) of the lens 20b of the camera 20. f z f The intersection of the straight line (dotted line) drawn in the manner shown is the point where it intersects the reference plane. This straight line is the optical axis of camera 20.

[0098] Furthermore, the pulling direction of the single-crystal silicon 3 is the positive direction of the z-axis, and the center position C(0, y) of the imaging device 20a is... c z c The center position F(0, y) of lens 20b f z f It lies in the yz plane. Figure 9 The coordinates (u, v) in the image shown on the left are represented in pixels of the imaging device 20a, and are related to any point P(x) on the imaging device 20a as shown in the following equation (1). p y p z p )correspond.

[0099] [Mathematical Expression 1]

[0100]

[0101] Here, α u With α v The pixel dimensions in the horizontal and vertical directions of the imaging device 20a are y c With z c Let C be the y-coordinate and z-coordinate of the center position C of the imaging device 20a. Furthermore, as... Figure 9 As shown in the right-hand figure, θ c The angle between the optical axis and the z-axis of camera 20 is the setting angle of camera 20.

[0102] The coordinate origin C0 on the reference plane is used to reach the center position C(0, y) of the imaging device 20a. c z c Let the distance be L. c At that time, y c z c As shown in equation (2) below.

[0103] [Mathematical Expression 2]

[0104]

[0105] When the distance from the origin C0 to the center position F of the lens 20b of the camera 20 is set as 'a' and the distance from the center position F of the lens 20b to the center position C of the imaging device 20a is set as 'b', the distance L from the origin C0 to the center position C of the imaging device 20a is... c As shown in equation (3) below.

[0106] [Mathematical Expression 3]

[0107] L c =L1+L2 (3)

[0108] Furthermore, according to the lens imaging formula, the focal length f l Distances a and b are expressed as shown in equation (4) below.

[0109] [Mathematical Expression 4]

[0110]

[0111] Eliminating distance L2 from equations (3) and (4), if L c With distance L1 and focal length f l To express it, it is as shown in the following formula (5).

[0112] [Mathematical Expression 5]

[0113]

[0114] The value of distance L1 from the origin C0 to the center position F of the lens 20b of camera 20 can be obtained using the center position F(0, y) of the lens 20b of camera 20. f z f It can be represented by the following formula (6).

[0115] [Mathematical Expression 6]

[0116]

[0117] Therefore, the center position F(0, y) of the lens 20b of the camera 20 can be used. f z f ) and the focal length f of lens 20b l The above equation (2) can be represented by the following equation (7).

[0118] [Mathematical Expression 7]

[0119]

[0120] When lens 20b is considered as a pinhole, any point P(x) on imaging device 20a... p y p z p ) through F(0, y f z f The projection point P'(X, Y, 0) is projected onto the reference plane and can be represented by the following equation (8).

[0121] [Mathematical Expression 8]

[0122]

[0123] Therefore, by using equations (1), (7), and (8), it is possible to determine the coordinates of the real image and the mirror image of the seed crystal that has been projected onto the reference plane.

[0124] Next, the method for calculating the center coordinates and radius of the circular outer peripheral edge pattern at the lower end of the straight body of the seed crystal will be explained. As a method for calculating the coordinates (x0, y0) of the center of the seed crystal and the radius r based on the coordinates of the real image and the mirror image projected onto the reference plane, the least squares method can be used. The shape of the outer peripheral edge at the lower end of the straight body of the seed crystal is circular, and the image of the outer peripheral edge satisfies the equation of the circle shown in the following equation (9).

[0125] [Mathematical Expression 9]

[0126] (x-x0) 2 +(y-y0) 2 =r 2 (9)

[0127] Here, the least squares method is used to calculate (x0, y0) and r in equation (9). To simplify the calculation using the least squares method, the following transformation of equation (10) is performed.

[0128] [Mathematical Expression 10]

[0129]

[0130] The variables a, b, and c in equation (10) are obtained by the least squares method. This yields the condition that the sum of squares of the differences between equation (10) and the measured points is minimized, which is obtained by solving the partial differential equation shown in equation (11) below.

[0131] [Mathematical Expression 11]

[0132]

[0133] Then, the solution to this equation (11) can be calculated by the simultaneous equations shown in the following equation (12).

[0134] [Mathematical Expression 12]

[0135]

[0136] By using the least squares method in this way, it is possible to calculate the center coordinates and radius of the outer periphery of the circular lower end of the straight body of the real image 5R and mirror image 5M of the seed crystal projected onto the reference plane.

[0137] Figure 10 This is a schematic diagram illustrating the method of calculating the seed-melt surface interval ΔG based on the center coordinates of the circular outer peripheral edge pattern at the lower end of the straight body of the real image 5R and the mirror image 5M of the seed crystal 5.

[0138] like Figure 10 As shown, the center coordinate C of the lower end of the straight portion of the mirror image 5M of seed crystal 5 is... m Originally, the center coordinate C of the lower end of the straight part 5a of the real image 5R of the seed crystal 5, which was sandwiched between the molten surface 2a, was... h (X hc Y hc On the opposite side of (0) and (0), the straight line connecting these two points passes through C. h (X hc Y hc 0) becomes a straight line parallel to the Z-axis.

[0139] On the other hand, the center coordinate C of the lower end of the straight portion of the seed crystal 5 mirror image 5M on the reference plane is... m '(X mc Y mcThe center coordinates C of the lower end of the straight portion of the real image 5R of seed crystal 5 are 0) h (X hc Y hc The coordinates (X, 0) are projected onto the reference plane. Therefore, the center coordinates (X, 0) of the lower end of the straight part of the mirror image 5M of the seed crystal 5 are... mc Y mc Z gap The center coordinates (X) of the lower end of the straight portion of the seed crystal 5, mirror image 5M, located on the reference plane are as follows: mc Y mc The center position F(0, y) of lens 20b and lens 20b f z f The straight line is on the line. Therefore, the desired seed-liquid surface spacing ΔG is Z. gap The half value can be calculated according to the following formula (13).

[0140] [Mathematical Expression 13]

[0141] -2ΔG=Z gap =Z f -Z f (Y mc -Y f ) / (Y hc -Y f ) ( 13 )

[0142] In this way, the seed-liquid surface spacing ΔG can be determined based on the center coordinates C of the real image of seed 5 on the reference plane. h (X hc Y hc The center coordinates C of the mirror image of seed 5 on the reference plane (0), 0). m '(X mc Y mc The coordinates of the center of the lens 20b of the camera 20 are F(0, y). f z f The projection transformation reference plane is set to the same height as the lower end of the straight portion 5a of the seed crystal 5. However, it can also be set to other height positions. Therefore, for example, the reference plane can be set to the same height as the opening edge at the lower end of the heat insulation body 17.

[0143] The preferred embodiments of the present invention have been described above, but the present invention is not limited to the above embodiments. Various modifications can be made without departing from the spirit of the present invention, and these modifications are of course included within the scope of the present invention.

[0144] For example, in the above embodiment, the real image edge is calculated after the mirror image edge is calculated, but the mirror image edge can also be calculated after the real image edge is calculated.

[0145] Furthermore, in the above embodiment, the case where the front end of the seed crystal 5 has a tapered shape is given as an example. However, the present invention can also be applied to the case where the seed crystal 5 is composed only of a straight body portion 5a, and can be based on the coordinates P of the real image center in the captured image. R To the center coordinate P of the mirror image M Number of pixels N D Accurately measure the distance L between the lower end of the seed crystal 5 and the molten surface 2a. S .

[0146] Furthermore, while the above embodiments exemplify a method for manufacturing single-crystal silicon, the present invention can also be applied to methods for manufacturing various single crystals, such as germanium or sapphire, that can be pulled using the CZ method.

[0147] Explanation of reference numerals in the attached figures

[0148] 1-Single crystal manufacturing apparatus, 2-Silicon melt, 2a-Melt surface, 3-Single crystal silicon, 3I-Single crystal silicon ingot, 3a-Neck, 3b-Shoulder, 3c-Main body, 3d-Tail, 5-Seed crystal, 5M-Mirror image, 5R-Real image, 5a-Straight body, 5b-Conical body, 10-Cavity, 10a-Main chamber, 10b-Drawing chamber, 10c-Gas inlet, 10d-Gas outlet, 10e-Observation window, 11-Quartz crucible, 12-Graphite crucible, 13-Rotation shaft, 14-Cruise drive mechanism, 15-Heater, 16-Insulation material, 17-Insulation body, 18-Wire, 19-Crystal pulling mechanism, 20-Camera, 21-Image processing unit, 22-Control unit, L D - Distance between real and mirror images, L P1 - The distance between the lower end of the straight portion of the seed crystal and the melt surface, L S - The distance between the lower end of the seed crystal and the melt surface, N D - Number of pixels between real and mirror images, P0 - Front end of the seed crystal, P1 - Lower end of the straight portion of the seed crystal, P2 - Mirror image of the lower end of the straight portion of the seed crystal, P M - Mirror center coordinates, P R-Real image center coordinates, S11-Raw material melting process, S12-Preheating process, S13-Liquid coating process, S14-Diameter reduction process, S15-Shoulder growth process, S16-Main body growth process, S17-Tail growth process, S18-Cooling process, S20-First region cutting step, S21-Mirror image position detection step, S22-Second region cutting step, S23-Second region binarization step, S24-Mirror image center coordinate calculation step, S25-Real image position detection step, S26-Third region cutting step, S27-Third region binarization step, S28-Real image center coordinate calculation step, h, h1, h2-Seed height position, Δh-Seed delivery amount, α-Conversion coefficient, θ, θ0-Camera setting angle, θ1-Cone angle.

Claims

1. A method for manufacturing a single crystal, said method being a single crystal manufacturing method utilizing the Czeklaussky method, characterized in that, The following steps are required: Measure the distance between the lower end of the seed crystal placed above the melt and the melt surface; The seed crystal is lowered and deposited in the melt based on the interval; and While maintaining contact with the molten metal, the seed crystal is pulled up to allow a single crystal to grow at its lower end. In the step of measuring the distance between the lower end of the seed crystal and the molten surface, The seed crystal and the melt surface are photographed using a camera positioned obliquely above the melt surface. A circular approximation is generated by approximating the lower arc-shaped edge pattern of the real image of the seed crystal captured in the image taken by the camera, and a circular approximation is also generated by approximating the upper arc-shaped edge pattern of the mirror image of the seed crystal reflected in the molten surface, to produce a mirror image edge approximation. The distance between the lower end of the seed crystal and the molten surface is calculated based on the distance from the center coordinates of the approximate circle at the edge of the real image to the center coordinates of the approximate circle at the edge of the mirror image. In the step of measuring the distance between the lower end of the seed crystal and the molten surface, after determining the approximate circle of the mirror edge based on the captured image, the approximate circle of the real image edge located above the approximate circle of the mirror edge is determined. The steps to determine the approximate circle of the mirrored edge include the following: The upper position of the image of the seed crystal is determined based on the brightness distribution of a first region preset in the captured image. A second region containing the upper position of the mirror image of the seed crystal is defined; the second region is binarized; and the edge pattern of the straight portion of the mirror image of the seed crystal is detected; and The edge pattern of the straight portion of the mirror image is approximated as a circle. The steps to determine the approximate circle of the edge of the real image include the following: The position where the differential value of the brightness exceeds a predetermined threshold when scanning the first region upwards from the center coordinates of the approximately circular edge of the mirror image is determined as the lower end position of the real image of the seed crystal; A third region containing the lower end position of the real image of the seed crystal is defined, the third region is binarized, and the edge pattern of the straight part of the real image of the seed crystal is detected. as well as The edge pattern of the straight portion of the real image is approximated by a circle.

2. The method for manufacturing a single crystal according to claim 1, wherein, The distance in real space is converted by multiplying half the number of pixels from the center coordinates of the approximate circle at the edge of the real image to the center coordinates of the approximate circle at the edge of the mirror image by a conversion factor.

3. The method for manufacturing a single crystal according to claim 2, wherein, It also includes the following steps: Before measuring the distance between the lower end of the seed crystal and the molten surface, the conversion factor is calculated based on the change in the pixel position of the seed crystal in the captured image when the position of the seed crystal is moved a certain distance in the vertical direction.

4. The method for manufacturing a single crystal according to claim 1, wherein, Based on the camera's setting angle and focal length, the coordinates in the captured images of the approximate real image edge circle and the approximate mirror image edge circle are projected and transformed into coordinates in real space, and then the distance between the lower end of the seed crystal and the molten surface is calculated.

5. The method for manufacturing a single crystal according to claim 4, wherein, A reference plane is set at the same height as the lower end of the seed crystal. The approximate circle of the real image edge and the approximate circle of the mirror edge are projected onto the reference plane. The distance between the lower end of the seed crystal and the molten surface is calculated based on the center coordinates of the approximate circle of the real image edge, the center coordinates of the approximate circle of the mirror edge, and the center coordinates of the camera lens.

6. The method for manufacturing a single crystal according to any one of claims 1 to 5, wherein, In the step of measuring the distance between the lower end of the seed crystal and the molten surface, Multiple real and mirror images of the seed crystal at the same height position are continuously captured. The distance between the lower end of the seed crystal and the molten surface is calculated based on the average value obtained from each of the multiple images.

7. The method for manufacturing a single crystal according to any one of claims 1 to 5, wherein, The seed crystal also includes a tapered portion disposed below the straight portion. The lower end of the seed crystal is the lower end of the cone portion, and the cone angle of the cone portion is larger than the camera's setting angle. Before measuring the distance between the lower end of the seed crystal and the melt surface, the process also includes a step of pre-measuring the length of the cone. The step of measuring the distance between the lower end of the seed crystal and the molten surface includes the following steps: subtracting the length of the conical portion from the distance between the lower end of the straight portion of the seed crystal and the molten surface.

8. The method for manufacturing a single crystal according to any one of claims 1 to 5, wherein, It also includes the following steps: After measuring the distance between the lower end of the seed crystal and the surface of the molten liquid, and before placing the seed crystal in the molten liquid, adjust the height of at least one of the seed crystal and the crucible supporting the molten liquid so that the distance becomes the target value. as well as The seed crystal is preheated by being kept stationary at the target value position.

Images