Czochralski single crystal furnace monocrystalline silicon diameter detection method

Abstract

The invention relates to a method for detecting the diameter of monocrystalline silicon in a straight pull type single crystal furnace, which comprises the following steps of: shooting an original picture of a crystal in the straight pull type single crystal furnace by a CCD (Charge Coupled Device) camera, delimiting a rectangular diameter detection area at the left part or the right part of an aperture on the picture, and calculating an edge point set in the rectangular areas at the two sides; and integrating the edge point set, and performing iterative least square method fitting circle on the set. A stable and high-precision crystal diameter signal can be detected by combining a diameter detection method which is not influenced by brightness change and an iterative least square fitting circle on the basis of a least square fitting circle of an idea of eliminating an abnormal value through iterative fitting.

Description

Technical Field

[0001] The present invention relates to a method for detecting the diameter of single-crystal silicon in a Czochralski single-crystal furnace, and more particularly to a stable method for detecting the diameter of single-crystal silicon in a Czochralski single-crystal furnace that is not affected by brightness changes. Background Art

[0002] The Czochralski single-crystal manufacturing method means putting polycrystalline silicon blocks as raw materials into a quartz crucible, heating and melting them in a single-crystal furnace, and then bringing a seed crystal with a diameter of only 10 mm into contact with the surface of the molten solution. At an appropriate temperature, silicon atoms in the solution will form regular crystals on the solid-liquid interface along the silicon atom arrangement structure of the seed crystal, becoming single crystals. By slightly rotating and lifting the seed crystal, silicon atoms in the solution will continue to crystallize on the previously formed single crystal and continue its regular atomic arrangement structure. If the entire crystallization environment is stable, crystallization can be repeated continuously, and finally a cylindrical silicon single-crystal crystal with a neat atomic arrangement, that is, a silicon single-crystal ingot, is formed.

[0003] When the crystallization speed increases, the diameter of the crystal will become thicker. Increasing the pulling speed can make the diameter thinner, and increasing the temperature can inhibit the crystallization speed. Conversely, if the crystallization speed slows down and the diameter becomes thinner, it is controlled by reducing the pulling speed and lowering the temperature.

[0004] At the beginning of crystal pulling, a thin neck with a certain length and a diameter of 3 - 5 mm is first drawn out to eliminate crystal misalignment. This process is called seeding. Then, the diameter of the single crystal is enlarged to the process requirements and enters the isodiametric stage until most of the silicon solution crystallizes into a single-crystal ingot, leaving only a small amount of remaining material. Controlling the diameter to ensure the isodiametric growth of the crystal is an important link in single-crystal manufacturing.

[0005] The melting point of silicon is about 1420 °C, and the crystal pulling process is always carried out in an environment of high temperature and negative pressure. Diameter measurement must be realized non-contact outside the crystal pulling furnace body through an observation window. During the crystal pulling process, a bright light ring with a very high brightness, called an aperture, will be formed at the junction of the solid crystal and the liquid melt. It is the reflection of the bright light of the crucible wall by the meniscus at the solid-liquid interface. When the crystal becomes thicker, the diameter of the aperture becomes larger, and vice versa. By measuring the change in the diameter of the aperture, the change in the diameter of the single crystal can be reflected.

[0006] In the prior art, there are mainly two diameter detection methods based on image processing: the method of binarizing the image based on image grayscale and detecting edge points, and the method of detecting edge points based on image grayscale gradient. These two detection methods are not applicable to a detection environment with continuously changing brightness, and the detection stability and accuracy are not high. During the crystal pulling process, due to the continuous change of the temperature in the furnace, the brightness of the aperture is not fixed, and a diameter detection method that is not affected by brightness changes needs to be developed. On the other hand, it is also necessary to avoid interference with the detection points caused by the crystal ridges and impurities on the liquid surface. Summary of the Invention

[0007] The present invention proposes a method for detecting the diameter of single crystal silicon in a CZ-type single crystal furnace, the purpose of which is to overcome the above-mentioned deficiencies in the prior art and to provide a diameter detection method that is stable and not affected by brightness changes.

[0008] The technical solution of the present invention is a method for detecting the diameter of single crystal silicon in a CZ-type single crystal furnace, comprising the following steps:

[0009] 1) Use a CCD camera to take an original picture of the crystal in the CZ-type single crystal furnace, and define a rectangular diameter detection area on the left or right part of the aperture on the picture;

[0010] 2) Traverse to obtain the gray value list grayValues ​​of the first column x1 of the rectangular area, calculate the maximum gray value grayMax, set a ratio value ratio between 0 and 1, find the last index position index1 greater than the gray threshold grayThreshold = grayMax*ratio and its corresponding gray value gray1, take index1 as the starting position, traverse grayValues ​​in reverse, find the first index position index2 greater than the gray threshold and its corresponding gray value gray2;

[0011] 3) Customize the three-dimensional coordinate system, with x as the image horizontal coordinate, y as the image vertical coordinate, and z as the grayscale value. Fix the horizontal coordinate x1. On the yz two-dimensional plane, calculate the intersection point [y1,0] of the straight line where the line segment formed by the point [index1, gray1] and the point [index2, gray2] is located and the y-axis. [x1, y1] is the aperture edge point p1 found in the first column of the rectangle.

[0012] 4) Execute steps 2)-3) for the other columns of the rectangle in sequence to obtain the corresponding aperture edge points p2,…,pn, that is, the edge point set P1 in the rectangular area on one side is obtained; the edge point set P2 in the rectangular area on the other side is obtained by the same detection method;

[0013] 5) Integrate P1 and P2 to obtain the edge point set points, and perform iterative least squares fitting of the circle on points.

[0014] Preferably, the specific steps of step 5) are as follows:

[0015] ① Execute the least squares method to fit a circle on points to obtain a fitted circle circle1, calculate the distance from all points to the fitted circle, remove abnormal points whose distance is greater than the distance threshold, i.e., the limit of the distance from the point to the fitted circle distThreshold, and obtain a set of normal points normalPoints1;

[0016] ② Repeat step ① multiple times to obtain the final set of normal points normalPoints and the final fitted circle circle.

[0017] Advantages of the present invention: The method is reasonably designed. The least squares method for fitting a circle based on the idea of iterative fitting to eliminate outliers, combined with a diameter detection method that is not affected by brightness changes and the iterative least squares method for fitting a circle, can detect a stable and high-precision crystal diameter signal. Brief Description of the Drawings

[0018] Figure 1 It is a schematic structural diagram of an embodiment of a Czochralski single crystal furnace.

[0019] Figure 2 It is an original picture of a crystal in a Czochralski single crystal furnace taken by a CCD camera in an embodiment.

[0020] Figure 3 It is Figure 2 A schematic diagram of an embodiment for delimiting a diameter detection area.

[0021] Figure 4 It is Figure 3 The detection effect diagram of the detection area on the left side of the aperture in

[0022] Figure 5 It is Figure 3 The detection effect diagram of the detection area on the right side of the aperture in

[0023] Figure 6 It is Figure 2 The fitting effect diagram of an embodiment of iterative circle fitting. Detailed Embodiment

[0024] The present invention will be further described in detail below in conjunction with embodiments and specific implementation manners.

[0025] A method for detecting the diameter of single crystal silicon in a Czochralski single crystal furnace includes the following steps:

[0026] 1) Take a picture of the crystal in the Czochralski single crystal furnace shown by a CCD camera Figure 1 The original picture of the crystal in the Czochralski single crystal furnace is as shown in Figure 2 , and delimit a rectangular diameter detection area on the left or right part of the aperture in the picture as shown in Figure 3 (hereinafter, the left rectangular detection area is taken as an example, as shown in Figure 4 ).

[0027] 2) Traverse to obtain the gray value list grayValues ​​of the first column x1 of the rectangular area, calculate the maximum gray value grayMax, set a ratio value ratio between 0 and 1, find the last index position index1 greater than the gray threshold grayThreshold = grayMax*ratio and its corresponding gray value gray1. Starting from index1, traverse grayValues ​​in reverse order to find the first index position index2 greater than the gray threshold and its corresponding gray value gray2.

[0028] 3) Customize the three-dimensional coordinate system (x is the image horizontal coordinate, y is the image vertical coordinate, and z is the grayscale value). Fix the horizontal coordinate x1, and calculate the intersection point [y1,0] of the line segment formed by the points [index1, gray1] and [index2, gray2] and the y-axis on the yz two-dimensional plane. [x1, y1] is the aperture edge point p1 found in the first column of the rectangle.

[0029] 4) Perform steps 2)-3) on the other columns of the rectangle in turn to obtain the corresponding aperture edge points p2,…,pn, that is, the edge point set P1 in the rectangular area on one side (left side). Use the same detection method to obtain the edge point set P2 in the rectangular area on the other side (right side).

[0030] 5) Integrate P1 and P2 to get the edge point set points, and perform iterative least squares fitting of the circle on the points. The specific steps are as follows:

[0031] ① Execute the least squares method to fit a circle to points to obtain the fitted circle circle1. Calculate the distance from all points to the fitted circle, remove abnormal points whose distance is greater than the distance threshold (distThreshold) from the point to the fitted circle, and obtain a set of normal points normalPoints1.

[0032] ② Repeat step ① multiple times to obtain the final normal point set normalPoints and the final fitting circle circle.

[0033] The algebraic equation of the circle fitted by the least squares method is as follows:

[0034] A(x 2 +y 2 )+Bx+Cy+D=0

[0035] Where A=(A,B,C,D) T is a 4-parameter vector, A, B, C, D are the four parameters of the equation, and (x, y) are the data points used to fit the circle. By minimizing all points {(x i ,yi )} to solve the parameter vector of the sum of squared algebraic distances, that is, to minimize the following function, subject to A T NA = 1, and the formula is as follows:

[0036]

[0037] Among them, and {(x i , y i )} are data points, and n is the number of data points, then:

[0038]

[0039] Among them, X and M are data matrices, and N is a constraint matrix,

[0040] The solution A of the equation is the generalized eigenvector of the matrix pair (M, N); after obtaining the solution of the algebraic equation, the center and radius of the fitted circle are obtained through the conversion formula, and the conversion formula is Among them, (a, b) is the center of the circle, and R is the radius.

[0041] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the inventive concept of the present invention, several deformations and improvements can be made, and these all belong to the protection scope of the present invention.

Claims

1. A method for detecting the diameter of single crystal silicon in a CZ-type single crystal furnace, characterized in that: The following steps are involved: 1) Use a CCD camera to take an original picture of the crystal in the CZ-type single crystal furnace, and define a rectangular diameter detection area on the left or right part of the aperture on the picture; 2) Traverse to obtain the gray value list grayValues ​​of the first column x1 of the rectangular area, calculate the maximum gray value grayMax, set a ratio value ratio between 0 and 1, find the last index position index1 greater than the gray threshold grayThreshold = grayMax*ratio and its corresponding gray value gray1, take index1 as the starting position, traverse grayValues ​​in reverse, find the first index position index2 greater than the gray threshold and its corresponding gray value gray2; 3) Customize the three-dimensional coordinate system, with x as the image horizontal coordinate, y as the image vertical coordinate, and z as the grayscale value. Fix the horizontal coordinate x1. On the yz two-dimensional plane, calculate the intersection point [y1,0] of the straight line where the line segment formed by the point [index1, gray1] and the point [index2, gray2] is located and the y-axis. [x1, y1] is the aperture edge point p1 found in the first column of the rectangle. 4) Execute steps 2)-3) for the other columns of the rectangle in sequence to obtain the corresponding aperture edge points p2,…,pn, that is, the edge point set P1 in the rectangular area on one side is obtained; the edge point set P2 in the rectangular area on the other side is obtained by the same detection method; 5) Integrate P1 and P2 to obtain the edge point set points, and perform iterative least squares fitting of the circle on points.

2. A method for detecting the diameter of single crystal silicon in a CZ-type single crystal furnace as claimed in claim 1, characterized in that: The specific steps of step 5) are as follows: ① Execute the least squares method to fit a circle on points to obtain a fitted circle circle1, calculate the distance from all points to the fitted circle, remove abnormal points whose distance is greater than the distance threshold, i.e., the limit of the distance from the point to the fitted circle distThreshold, and obtain a set of normal points normalPoints1; ② Repeat step ① multiple times to obtain the final normal point set normalPoints and the final fitting circle circle.

Images