JP4984091B2 - Single crystal diameter detection method and single crystal pulling apparatus - Google Patents

Description

  The present invention relates to a method for detecting the diameter of a single crystal grown by the Czochralski method and a single crystal pulling apparatus.

Czochralski method (Czochralski method) as a method for growing semiconductor silicon single crystals
Method (hereinafter, CZ method) is known. In this method, a single crystal is grown by attaching a seed crystal to a melt and slowly pulling it upward while rotating. Single crystals are manufactured with a certain diameter. For example, if the final product is an 8-inch (200 mm) wafer, it is common to produce crystals at 202 to 210 mm, which is slightly larger than the diameter. Thereafter, the outer periphery of the crystal is ground into a cylindrical shape, sliced into a wafer, and then subjected to a chamfering, a polishing process, and the like to reach a final target wafer diameter. Therefore, the target diameter in single crystal manufacturing must be larger than the final product wafer diameter. However, if it is too large, the amount of grinding and polishing increases, which is not economical. Therefore, a single crystal having a diameter larger than that of the wafer and as small as possible is required.

  There are two main methods for controlling the diameter in the CZ method: an optical method (camera method) and a weight method (load cell method). In the optical system, the grown crystal inside the furnace is observed through quartz glass by a camera attached outside the furnace. The image captured by the camera is processed, the position of the end of the crystal is determined, the position is coordinated and converted into a diameter. Optical methods include a method of measuring both ends of a crystal, a method of measuring one side of a crystal, and a method of determining a diameter from the curvature of an arc.

  However, the method of measuring both ends of the crystal with a camera makes it difficult to capture the entire diameter D as shown in FIG. 11 as the diameter of the crystal increases. Moreover, even if the whole is caught, there exists a problem that the resolution deteriorates. In addition, for example, as shown in (Patent Document 1), there is a method in which two cameras are installed in a pulling device and both ends are viewed using each camera. However, there is a problem of an error due to a relative position shift of the cameras. Become.

  Further, as a method of measuring one side of the crystal, there is a method of calculating the diameter by the distance R from the virtual center point as shown in FIG. 12, but a measurement error occurs due to the shift of the virtual point due to the positional deviation of the camera. End up.

  Further, as shown in FIG. 13, the optical method includes a method of calculating the distance R from the center point from the curvature of the arc and determining the diameter, but this method also has a curvature as the crystal becomes larger in diameter. There is a problem that the measurement error is increased and the measurement error is increased.

  As described above, the optical single crystal diameter detection method has a problem that a measurement error occurs due to an increase in the diameter of the crystal or a shift of the detection camera. For example, when the crystal diameter is deviated, problems such as production of a defective product due to insufficient diameter and a decrease in yield due to an increase in machining allowance due to excessive diameter occur. In addition, the crystal growth conditions achieve uniform quality while changing the conditions in the crystal growth direction, but the amount of silicon melt in the crucible deviates from the target due to the crystal diameter deviating from the target. As a result, it also causes the problem of quality deviation.

  On the other hand, in the weight method, for example, a method (load cell method) in which a weight meter called a load cell as shown in (Patent Document 2) is attached to the upper shaft and the weight of a growing crystal is measured (load cell method) is common. The load cell method is a method for calculating the diameter of a crystal from an increase in weight per unit length. This method does not generate an error as in the optical type, and if the error of a single load cell is grasped, it should be possible to measure the diameter. However, due to the higher weight of the crystal, it is necessary to increase the maximum allowable weight of the load cell. At that time, the measurement error increases or the sensitivity decreases, and the diameter cannot be calculated in a short time. When the diameter increases in a short time, it is necessary to increase the crystal growth rate and return the diameter to the target. However, there is a problem in that it cannot be controlled in a short time and an uneven crystal is produced. In addition, when the single crystal to be grown is an uneven crystal, there has been a problem that the unevenness of quality at the uneven portion becomes large, or that a defective product is produced due to insufficient diameter at the recessed portion.

JP 2004-35352 A JP-A-9-175893

  The present invention provides a single crystal diameter detection method and a single crystal pulling apparatus capable of improving the measurement accuracy of large diameter and high weight crystal diameters, and achieving improvement in single crystal yield and reduction in quality variation. The purpose is that.

  In order to solve the above problems, the present invention is a method for detecting the diameter of a single crystal grown by the Czochralski method, which detects the diameter of a single crystal by both a camera and a load cell, The difference between the diameter calculated by the load cell, the correction coefficient α obtained in advance according to the growth rate of the single crystal, the parameter representing the shape of the cone portion of the single crystal, and the straight body portion of the single crystal grown in advance. Provided is a method for detecting a diameter of a single crystal, wherein the camera-detected diameter is corrected by cone shape correction obtained from a relationship with the diameter, and the value obtained by the correction is used as the diameter of the single crystal. (Claim 1).

  Thus, when growing a single crystal by the Czochralski method, the diameter of the single crystal is detected by both the camera and the load cell, and the difference between the camera detected diameter and the diameter calculated by the load cell is determined. Camera detection diameter by correction coefficient α obtained in advance according to speed, cone shape correction obtained from the relationship between the parameter representing the shape of the cone part of the single crystal and the diameter of the straight body of the single crystal grown in advance. Is corrected, and the value obtained by the correction is used as the diameter of the single crystal, whereby the single crystal diameter detection method can be improved with improved measurement accuracy of the diameter of the large diameter and high weight crystal. Therefore, it is possible to suppress the variation in diameter in the first half portion of the straight body portion of the single crystal, and to improve the yield of the single crystal and reduce the variation in quality.

  In the detection method of the present invention, the correction is performed by multiplying the difference between the camera detection diameter and the diameter calculated by the load cell by the correction coefficient α or adding the value and the cone shape correction. It is preferable to carry out by adding to the diameter (claim 2).

  As described above, the difference between the camera detection diameter and the diameter calculated by the load cell is multiplied by the correction coefficient α, or the added value and the cone shape correction are added to the camera detection diameter to correct the camera detection diameter. Thus, the accuracy of the absolute value of the diameter can be improved, and the single crystal diameter detection method can be improved by improving the measurement accuracy of the diameter of the large diameter and heavy crystal. Therefore, it is possible to suppress the variation in diameter in the first half of the straight body portion of the single crystal, and to effectively improve the yield of the single crystal and reduce the quality variation.

In the detection method of the present invention, the parameter may be any one of a length of the cone portion of the single crystal, a weight of the cone portion of the single crystal, a length of the cone portion of the single crystal, and a shape constant obtained from the weight. It is preferable to use one (Claim 3).
Thus, any one of the shape constants determined from the length of the cone portion of the single crystal, the weight of the cone portion of the single crystal, the length of the cone portion of the single crystal, and the weight is used as the shape of the cone portion of the single crystal. By using the parameters to represent, the shape of the cone portion of the single crystal can be easily defined. Therefore, cone shape correction can be easily derived using this parameter.

In this case, the shape constant (C) can be obtained by the following formula (1) using the length (Lc) of the cone portion of the single crystal and the weight (Wc) of the cone portion of the single crystal (claim) Item 4).
C = Lc ³ / Wc (1)
By obtaining the shape constant by such an expression (1), the shape of the cone can be defined without being influenced by the size of the cone of the single crystal. The shape of the crystal interface can be reflected.

In the detection method of the present invention, the cone shape correction is obtained in advance according to the difference between the camera detection diameter and the diameter calculated by the load cell and the growth rate of the single crystal in the single crystal diameter detection method. The camera-detected diameter is corrected only by the correction coefficient α obtained, and the single crystal is grown in advance while detecting the diameter of the single crystal using the value obtained by the correction as the diameter of the single crystal, and then the pre-grown It is preferable to obtain by using the least square method from the relationship between the actual diameter at an arbitrary point of the straight body portion of 0 mm to 120 mm of the single crystal and the parameter representing the shape of the cone portion of the single crystal.
Thus, the camera detection diameter is corrected only by the correction coefficient α obtained in advance according to the difference between the camera detection diameter and the diameter calculated by the load cell and the growth rate of the single crystal, and the value obtained by the correction is obtained. Detecting the diameter of a single crystal, growing a single crystal in advance, and the relationship between the actual diameter at an arbitrary point of the straight body portion of the grown single crystal from 0 mm to 120 mm and the parameter representing the shape of the cone portion of the single crystal From the above, by obtaining the cone shape correction using the least square method, it is possible to perform the correction that surely reflects the shape of the cone portion. Therefore, the variation in diameter in the first half portion of the straight body portion of the single crystal can be reliably suppressed, and the yield of the single crystal can be improved and the quality variation can be reduced.

  The present invention is also a single crystal pulling apparatus for growing a single crystal by the Czochralski method, comprising at least both a camera and a load cell for detecting the diameter of the single crystal to be pulled, and any of the above A single crystal pulling apparatus is provided, wherein the single crystal diameter is detected by the method for detecting a single crystal diameter described in (6).

  Thus, a single crystal pulling apparatus for growing a single crystal by the Czochralski method includes at least both a camera and a load cell for detecting the diameter of the single crystal to be pulled. By detecting the diameter of the single crystal by the method for detecting the single crystal diameter as described above, the advantages of the single crystal diameter detection methods of the camera and the load cell method are utilized, and the mutual disadvantages are A single crystal pulling apparatus that can be supplemented can be provided. Therefore, the measurement accuracy of the diameter of large diameter and heavy crystal is effectively improved, and the variation in diameter in the first half of the straight body of the single crystal is suppressed, thereby improving the yield of single crystal and reducing the quality variation. The single crystal pulling apparatus can achieve the above.

  As described above, the single crystal diameter detection method of the present invention detects the diameter of a single crystal by both the camera and the load cell, the difference between the camera detected diameter and the diameter calculated by the load cell, and the growth rate of the single crystal. By correcting the camera detection diameter with the correction coefficient α and the cone shape correction determined in advance, and using the obtained value as the diameter of the single crystal, the measurement accuracy of the diameter of the large diameter and heavy crystal is improved. It is possible to improve the diameter variation in the first half of the straight body of the single crystal. Then, by detecting the diameter of the single crystal by this detection method, it is possible to provide a single crystal pulling apparatus that can achieve improvement in yield of single crystals and reduction in quality variation.

Hereinafter, the present invention will be described more specifically.
As described above, there are two methods for controlling the diameter of a single crystal in the CZ method: a method using a camera and a method using a load cell. Each of the camera methods has a low absolute value accuracy. There has been a problem that it is difficult to control short-time fluctuations.

  Therefore, the diameter of the single crystal is detected by both the camera and the load cell, and the camera detected diameter is determined by the difference between the camera detected diameter and the diameter calculated by the load cell, and the correction coefficient α determined in advance according to the growth rate of the single crystal. Attempts were made to detect the diameter of a single crystal using the value obtained by correcting the diameter as a single crystal, but there is a problem that the diameter of the crystal deviates from the target, particularly in the first half of the straight body of the single crystal. I understood.

  Therefore, the present inventors pay attention to the fact that the problem that the diameter of the crystal deviates from the target in the first half of the straight body portion of the single crystal is caused by the influence of the shape of the cone portion. Tried to correct the error.

  Specifically, the diameter of the single crystal is detected by both the camera and the load cell, the difference between the camera detected diameter and the diameter calculated by the load cell, and the correction coefficient α determined in advance according to the growth rate of the single crystal, Then, the camera detection diameter is corrected by the cone shape correction obtained from the relationship between the parameter representing the shape of the cone portion of the single crystal and the diameter of the straight body portion of the single crystal grown in advance, and the value obtained by the correction is corrected. The single crystal was grown while detecting the diameter of the single crystal as the diameter of the crystal.

  As a result, as shown in FIG. 1, the grown single crystal has improved variation in diameter in the first half of the straight body of the single crystal, compared to the case where the conventional single crystal diameter detection method is used. I found out. In addition, as shown in FIG. 2, it can be seen from the actual diameter in the first half (80 mm) of the straight body of the grown single crystal that the variation in diameter is improved as compared with the conventional method for detecting the diameter of single crystal. It was. FIG. 1 is a diagram showing improvement in diameter variation in the first half of the straight body of a single crystal in the method for detecting a single crystal diameter according to the present invention, and FIG. 2 shows the method for detecting a single crystal diameter according to the present invention. It is a figure which shows the actual diameter in the first half part (80 mm) of the straight body part of the single crystal grown using it.

  Further, the correction at this time is performed by multiplying the difference between the camera detection diameter and the diameter calculated by the load cell by the correction coefficient α or adding the value and the cone shape correction to the camera detection diameter. It was found that the accuracy of the absolute value of the diameter can be improved.

The parameter representing the shape of the cone portion of the single crystal is any one of the shape constant determined from the length of the cone portion of the single crystal, the weight of the cone portion of the single crystal, and the length and weight of the cone portion of the single crystal. Therefore, the shape of the cone portion of the single crystal can be easily defined. In particular, the shape constant (C) obtained from the length and weight of the cone portion of the single crystal is the length of the cone portion of the single crystal. (Lc), it was found that it can be obtained by the following formula (1) using the weight (Wc) of the cone portion of the single crystal. In addition, since the unit dimension is the reciprocal of the density (g / mm ³ ), this shape constant is a dimensionless amount considering that the density of the silicon crystal is 2.33 g / cm ^3. It was found to be independent of the crystal size.
C = Lc ³ / Wc (1)

  Here, FIG. 3 is a diagram showing the shape of the crystal interface immediately after the formation of the cone portion of the single crystal. In general, since the center part of the crystal is kept warm in the straight body part of the single crystal, the shape of the crystal interface is convex upward, but since the heat is removed from the upper part of the cone immediately after forming the cone part, as shown in FIG. In addition, the interface is convex downward. The downward convex amount shown in FIG. 3 is the difference in height in the growth axis direction from the shoulder height of the cone portion to the center of the crystal. FIG. 4 is a diagram showing the relationship between the downward convex amount and the shape constant. FIG. 4 indicates that there is a correlation between the downward convex amount and the shape constant. Therefore, it was found that the shape constant defined by the above formula (1) can be used as a parameter representing the downward convex amount of the crystal interface immediately after the cone portion formation is completed.

  The cone shape correction can be obtained by using the least square method from the relationship between the parameters representing the shape of the cone portion of the single crystal and the diameter of the straight body portion of the single crystal grown in advance, and the shape of the cone portion is reliably determined. It was also found that the correction reflected in can be performed.

The present invention has been completed based on the above findings and discoveries. Hereinafter, the present invention will be described in more detail with reference to the drawings, but the present invention is not limited thereto.
FIG. 5 is a schematic view of the single crystal pulling apparatus of the present invention.

  The single crystal pulling apparatus 20 includes a hollow cylindrical chamber 1, and a crucible 5 is disposed at the center thereof. This crucible has a double structure, and is an inner holding container made of quartz having a bottomed cylindrical shape (hereinafter simply referred to as “quartz crucible 5a”), and also has a bottomed structure adapted to hold the outside of the quartz crucible 5a. It is composed of a cylindrical graphite outer holding container (“graphite crucible 5b”).

  These crucibles 5 are fixed to the upper end of the support shaft 6 so as to be able to rotate and move up and down, and a resistance heating heater 8 is arranged substantially concentrically outside the crucible 5. Further, a heat insulating material 9 is concentrically arranged around the outside of the heater 8. And the silicon raw material melt 2 which melt | dissolved the silicon raw material with the heater 8 is accommodated in the quartz crucible 5a.

  At the center axis of the quartz crucible 5a filled with the silicon raw material melt 2, the pulling wire (or pulling shaft, which is the same as the support shaft 6) is rotated in the reverse direction or in the same direction at the predetermined speed. And the seed crystal 4 is held at the lower end of the pulling shaft 7. A silicon single crystal 3 including a cone portion 13 and a straight body portion 14 as shown in FIG. 3 is formed on the lower end surface of the seed crystal 4.

  The single crystal pulling device 20 includes a load cell 12 attached to a pulling shaft 7 for measuring the weight of the single crystal 3 during pulling, a camera 11 for detecting the diameter of the single crystal 3 from the outside of the furnace, and measurement of the load cell 12. The diameter of the single crystal 3 is calculated from the value, the difference between the detected diameter of the camera 11 and the diameter calculated by the load cell 12, the correction coefficient α determined in advance according to the growth rate of the single crystal, and the cone of the single crystal An arithmetic unit 10 that corrects the diameter of the single crystal 3 detected by the camera 11 by correcting the cone shape obtained from the relationship between the parameter representing the shape of the portion 13 and the diameter of the straight body portion 14 of the single crystal grown in advance. It has. The arithmetic unit 10 outputs signals to the support shaft 6 and the pulling shaft 7 or the heater 8 to control the crucible position, the crucible rising speed, the seed crystal position, the pulling speed, or the heater power.

  Such a single crystal pulling apparatus 20 of the present invention is controlled according to the single crystal diameter detection method of the present invention described below, thereby taking advantage of the single crystal diameter detection method of the camera and the load cell system, And it can be set as the single crystal pulling apparatus which can supplement a mutual shortcoming. Further, by correcting the diameter of the single crystal 3 detected by the camera 11 by the arithmetic unit 10, the measurement accuracy of the diameter of the large diameter and heavy crystal is effectively improved, and in particular, the first half of the straight body portion of the single crystal. It is possible to provide a single crystal pulling apparatus capable of suppressing the variation in diameter in the portion and improving the yield of single crystal and reducing the quality variation.

In the present invention, for example, using such a single crystal pulling apparatus, the diameter of the single crystal is detected as follows.
The method for detecting the single crystal diameter in the present invention will be described below.

  FIG. 6 is a diagram showing a flow of the method for detecting a single crystal diameter according to the present invention. As shown in FIG. 6, the method for detecting a single crystal diameter according to the present invention can be broadly divided into calculation of cone shape correction and correction of the diameter of the straight body portion. Then, before calculating the cone shape correction, a single crystal is grown in advance in order to obtain a correction formula used for calculating the cone shape correction. Thereafter, the actual diameter of the grown straight body portion is measured, and a correction formula used for cone shape correction is obtained from the relationship with the parameter representing the shape of the single crystal cone portion. For example, when the length of the cone portion is used as a parameter representing the shape of the cone portion, the length of the cone portion is determined from the amount of winding of the pulling wire during the formation of the cone portion and the amount of change in the crucible position with respect to the molten metal surface. The length converted to how much the crystal was pulled up is used.

  Thereafter, after the correction formula used for the cone shape correction obtained in advance is incorporated into the arithmetic unit 10 as shown in FIG. 5, the formation of the single crystal cone portion is started as shown in the flowchart of FIG. Is calculated. At this time, the weight (Wc1) of the single crystal when the formation of the cone portion of the single crystal is started is measured with the load cell 12 as shown in FIG. 5 (FIG. 6 (a)). Next, the weight (Wc2) and the length (Lc) of the single crystal immediately after completion of the formation of the single crystal cone are measured (FIG. 6B). Thereafter, the cone shape correction is obtained from the relationship between the parameter representing the shape of the cone portion of the single crystal and the diameter of the straight body portion of the single crystal grown in advance (FIG. 6C). Specifically, when using the length of the cone part as a parameter representing the shape of the cone part, the correction formula obtained from the relationship between the actual diameter of the straight body of the single crystal grown in advance and the length of the cone part, The cone shape correction is obtained by substituting the length (Lc) of the cone portion measured in FIG.

At this time, any one of the shape constants obtained from the length (Lc) of the cone part, the weight (Wc2-Wc1) of the cone part, the length (Lc) and the weight (Wc2-Wc1) of the cone part of the single crystal Can be used as a parameter representing the shape of the cone portion of the single crystal.
Thus, the shape of the cone portion of the single crystal can be easily defined, and cone shape correction can be easily derived using this parameter.

Further, the shape constant can be obtained by Lc ³ / (Wc2-Wc1) using the length of the cone part (Lc) and the weight of the cone part (Wc2-Wc1).
As a result, the shape of the cone portion can be defined without being influenced by the size of the cone portion of the single crystal, and the shape of the crystal interface immediately after the end of the formation of the cone portion of the single crystal can be reflected. . The shape constant is a reciprocal of the density (g / mm ³ ). Therefore, considering that the density of the silicon crystal is 2.33 g / cm ³ , this constant is a dimensionless quantity. Yes, regardless of crystal size. That is, when the shape of the cone portion is expressed only by the length of the cone portion (Lc) and the weight of the cone portion (Wc2-Wc1), the downward convex amount of the crystal interface as shown in FIG. 3 may not be reflected. However, this shape constant is proportional to the downward convex amount as shown in FIG. Therefore, it can be introduced as a parameter representing the shape of the cone portion.

The cone shape correction is obtained in advance according to the difference between the camera detection diameter and the diameter calculated by the load cell and the growth rate of the single crystal using a method of detecting the diameter of the single crystal by both the camera and the load cell. The diameter detected by the camera is corrected only with the correction coefficient α, and the single crystal is grown in advance while detecting the diameter of the single crystal using the value obtained by the correction as the diameter of the single crystal. From the relationship between the actual diameter at an arbitrary point of the body part 0 mm to 120 mm and the parameter (Lc etc.) representing the shape of the cone part of the single crystal, using the correction formula shown in FIG. Can be sought.
This makes it possible to perform correction that reliably reflects the shape of the cone portion. Therefore, the variation in diameter in the first half portion of the straight body portion of the single crystal can be reliably suppressed, and the yield of the single crystal can be improved and the quality variation can be reduced.

  The correction coefficient α is a numerical value obtained in advance according to the growth rate of the single crystal, and the weight measured by the load cell is a weight that can be measured when the single crystal is suspended in a stationary state. Here, in actual crystal production, there is a discrepancy between the converted diameter from the weight by the load cell and the actual diameter due to the shape of the crystal growth interface and the presence of surface tension. That is, the shape of the growth interface becomes larger as the crystal growth rate increases. This error can be obtained by manufacturing an actual crystal and measuring it for each pulling device, but if this upward convex shape can be predicted, the discrepancy between the converted diameter from the weight and the actual diameter should be predicted to some extent. Is possible. Therefore, the relationship between the growth rate and the correction coefficient is obtained in advance for each crucible diameter of the single crystal pulling apparatus and for each machine, and the correction coefficient α is calculated.

Next, the diameter of the straight body portion is corrected to detect the single crystal diameter. As shown in FIG. 6, the diameter of the straight body portion is corrected by first forming the straight body portion of the single crystal and starting the calculation with the straight body length (L1) of the single crystal (FIG. 6D). Then, the weight (Wt1) of the single crystal is measured with a load cell (FIG. 6E). Thereafter, the single crystal is pulled up to the straight body length (L2) (FIG. 6 (f)), and the weight (Wt2) of the single crystal is measured again by the load cell (FIG. 6 (g)). Next, the average (Dw) of the diameter in the average section from the single crystal weight (Wt1, Wt2) measured by the load cell to the straight body length (L2-L1) of the single crystal is calculated (FIG. 6 (h)). ). At this time, the calculation formula for calculating the converted diameter (Dw) from the load cell weight is Dw = 2√ ((Wt2−Wt1) / (π × (L2−L1) × 2.33)). (Wt1, Wt2 are g, L2, L1 are cm, Dw is cm, 2.33 is silicon specific gravity, g / cm ³ ).

  Further, in parallel with the weight measurement of the single crystal by the load cell, the diameter of the straight body length of the single crystal is detected several times by the camera, the obtained diameters are integrated, and an integrated value (T1) is calculated (FIG. 6). (I)). At this time, the number of times of integration is counted (FIG. 6 (j)), and the camera detection diameter (Do) is calculated by dividing the calculated integration value (T1) by the counted number of integrations (C1). (FIG. 6 (k)).

  Next, the difference between the camera detection diameter (Do) and the diameter (Dw) calculated by the load cell is obtained, and the difference is multiplied or added by a correction coefficient α obtained in advance according to the growth rate of the single crystal. The cone shape correction obtained as described above is added to the camera detection diameter (FIG. 6 (l)), and the diameter is controlled to the value obtained by this correction, and the single crystal is pulled up. The second and subsequent corrections are performed by repeating the same calculation when the next correction calculation start length is reached (FIG. 6 (m)).

  Thus, as shown in the flowchart of FIG. 6, by detecting the diameter of the single crystal, it is possible to provide a single crystal diameter detection method that improves the measurement accuracy of the diameters of large diameter and heavy crystal. Therefore, it is possible to suppress the variation in diameter in the first half portion of the straight body portion of the single crystal, and to improve the yield of the single crystal and reduce the variation in quality.

Next, the present invention will be described more specifically with reference to Examples and Comparative Examples of the present invention, but the present invention is not limited to these.
Example 1
First, in order to obtain cone shape correction, a silicon single crystal having a target diameter of 306 mm was repeatedly grown in advance by the following procedure. Using a single crystal pulling apparatus 20 as shown in FIG. 5, 410 kg of polycrystalline silicon is filled in a quartz crucible 5a having a diameter of 800 mm, the heater 8 is energized to melt the polycrystalline silicon, and the silicon raw material melt 2 Formed.

  Thereafter, the seed crystal 4 was immersed in the silicon raw material melt 2, and the single crystal 3 was pulled up by rotating the pulling shaft 7 in the direction opposite to the rotation direction of the crucible 5 to grow a single crystal in advance. At this time, the single crystal diameter is detected by both the camera 11 and the load cell 12, and in the step of correcting the diameter of the straight body part in FIG. A correction coefficient α = 6.5 mm was used as L2, and the diameter of the straight body portion was corrected without correcting the cone shape, and the single crystal diameter was detected to form the straight body portion. Then, the actual diameter in the straight body part 80mm of the grown single crystal was measured.

  In the single crystal pulling apparatus and pulling conditions used at this time, it was found that the correlation between the length of the cone portion and the diameter of the straight body portion of the single crystal grown in advance was very strong. The length of the cone was used as a parameter representing the shape. FIG. 7 shows an error from the target values of the length of the cone portion and the diameter of the straight body portion of the single crystal grown in advance for obtaining the cone shape correction. As shown in FIG. 7, an approximate straight line of R = −0.0229Lc + 5.24 was obtained as a correction formula used for cone shape correction.

  Thereafter, after the obtained correction formula was incorporated into the arithmetic unit 10 of the single crystal pulling apparatus 20, the single crystal diameter was detected and the single crystal was grown as shown in the flowchart of FIG. At this time, under the same single crystal pulling apparatus and pulling conditions as in the above single crystal growth, formation of the cone portion is started, the length Lc of the cone portion is measured, and the Lc is substituted into the correction equation. The cone shape was corrected, the diameter of the straight body was corrected, and a single crystal having a target diameter of 306 mm was grown 15 times. Moreover, the actual diameter in the straight body part 100mm of a single crystal was measured, and the dispersion | variation was calculated | required.

(Comparative Example 1)
In the growth of the single crystal of Example 1, the single crystal having the target diameter of 306 mm was grown 15 times without correcting the cone shape, and the actual diameter of the single crystal in the straight body portion of 100 mm was measured to obtain variation. .

  Here, Table 1 shows the measurement results of the actual diameter in the straight body portion 100 mm of the single crystal of Example 1 and Comparative Example 1. FIG. 8 is a diagram showing the length of the cone portion of Example 1 and Comparative Example 1 and the actual diameter of the single crystal straight body portion 100 mm.

  As can be seen from Table 1, since the standard deviation of Example 1 is smaller than that of the comparative example, the variation of the straight body portion of the single crystal of Example 1 is smaller than that of Comparative Example 1. Further, it can be seen from FIG. 8 that the diameter of the straight body of the single crystal of Example 1 varies in a range closer to the target diameter than that of Comparative Example 1.

(Example 2)
In the growth of the single crystal of Example 1, the cone shape of the single crystal was changed using a single crystal pulling apparatus of a different machine from that of Example 1 equipped with a quartz crucible having the same crucible diameter as in Example 1. The single crystal was grown 23 times using the weight of the cone part (Wc2-Wc1) as a parameter to be expressed. At this time, R = −0.0005 (Wc2−Wc1) +308.52 was used as a correction formula, and used when obtaining cone shape correction. Wc1 is the weight of the single crystal at the start of cone portion formation, and Wc2 is the weight of the single crystal immediately after the cone portion formation is completed. Moreover, the actual diameter in the straight body part 100mm of the grown single crystal was measured, and the dispersion | variation was calculated | required.

(Comparative Example 2)
In the growth of the single crystal of Example 2, the single crystal was grown 23 times without performing cone shape correction. Moreover, the actual diameter in the straight body part 100mm of the grown single crystal was measured, and the dispersion | variation was calculated | required.

  Here, Table 2 shows the result of measuring the error of the actual diameter with respect to the target diameter in the straight body 100 mm of the single crystal of Example 2 and Comparative Example 2. FIG. 9 is a diagram showing the error of the actual diameter of the cone part weight and the actual diameter of the straight crystal part 100 mm of Example 2 and Comparative Example 2 with respect to the target diameter.

  As can be seen from Table 2, since the standard deviation of Example 2 is much smaller than that of Comparative Example 2, the variation of the straight body of the single crystal of Example 2 is smaller than that of Comparative Example 2. Further, FIG. 9 shows that the diameter of the straight body portion of the single crystal of Example 2 has a small error with respect to the target diameter, and is closer to the target diameter than Comparative Example 2.

Example 3
In the growth of the single crystal of Example 1, the cone shape of the single crystal was changed using a single crystal pulling apparatus of a different machine from that of Example 1 equipped with a quartz crucible having the same crucible diameter as in Example 1. A single crystal was grown 24 times using a shape constant (Lc ³ / (Wc 2 −Wc 1)) determined from the length and weight of the cone part as a parameter to be expressed.
At this time, R = −0.0009 (Lc ³ / (Wc2−Wc1)) + 307 was used as a correction formula, and this was used when obtaining cone shape correction. Moreover, the actual diameter in the straight body part 100mm of the grown single crystal was measured, and the dispersion | variation was calculated | required.

(Comparative Example 3)
In the growth of the single crystal of Example 3, the single crystal was grown 24 times without performing cone shape correction. Moreover, the actual diameter in the straight body part 100mm of the grown single crystal was measured, and the dispersion | variation was calculated | required.

  Here, Table 3 shows the result of measuring the error of the actual diameter with respect to the target diameter in the straight body 100 mm of the single crystal of Example 3 and Comparative Example 3. FIG. 10 is a diagram showing the cone shape constant of Example 3 and Comparative Example 3 and the error of the actual diameter in the single crystal straight body 100 mm with respect to the target diameter.

  From Table 3, it can be seen that the standard deviation of Example 3 is half that of Comparative Example 3, and the variation of the straight body portion of the single crystal of Example 3 is smaller than that of Comparative Example 3. Further, it can be seen from FIG. 10 that the diameter of the straight body portion of the single crystal of Example 3 has a small error with respect to the target diameter and is in the range of the target diameter ± 0.5 mm.

  Thus, according to the method for detecting the diameter of a single crystal of the present invention, it is possible to provide a method for detecting the diameter of a single crystal with improved accuracy in measuring the diameter of a large diameter and heavy crystal. According to the single crystal pulling apparatus of the present invention, it is possible to improve the measurement accuracy of the diameter of the large diameter and high weight crystal, and to suppress the variation in diameter in the first half portion of the straight body of the single crystal. Improvement of crystal yield and reduction of quality variation can be achieved.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has substantially the same configuration as the technical idea described in the claims of the present invention, and any device that exhibits the same function and effect is the present invention. It is included in the technical scope of the invention.

It is a figure which shows the improvement of the dispersion | variation in the diameter in the first half part of the straight body part of a single crystal in the detection method of the single crystal diameter of this invention. It is a figure which shows the actual diameter in the front half part (80 mm) of the straight body part of the single crystal grown using the detection method of the single crystal diameter of this invention. It is a figure which shows the shape of the crystal | crystallization interface immediately after completion | finish of cone part formation of a single crystal. It is a figure which shows the relationship between a downward convex amount and a shape constant. It is the schematic of the single crystal pulling apparatus of this invention. It is a figure which shows the flow of the detection method of the single crystal diameter of this invention. It is a figure which shows the difference | error from the target value of the length of the cone part of a single crystal and the diameter of a straight body part which were grown beforehand in order to obtain | require cone shape correction | amendment. It is a figure which shows the length of the cone part of Example 1 and the comparative example 1, and the actual diameter in 100 mm of straight body parts of a single crystal. It is a figure which shows the difference | error with respect to the target diameter of the actual diameter in the weight of the cone part of Example 2 and the comparative example 2, and the straight body part of 100 mm of a single crystal. It is a figure which shows the difference | error with respect to the target diameter of the cone diameter of Example 3 and the comparative example 3, and the actual diameter in the straight body part 100mm of a single crystal. It is a figure which shows the method of measuring the both ends of a crystal | crystallization with the camera by the conventional optical system. It is a figure which shows the method of measuring one side of a crystal | crystallization with the camera by the conventional optical system. It is a figure which shows the method of calculating | requiring a diameter from the curvature of the circular arc by the conventional optical system.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Chamber, 2 ... Silicon raw material melt, 3 ... Silicon single crystal, 4 ... Seed crystal, 5 ... Crucible, 5a ... Quartz crucible, 5b ... Graphite crucible, 6 ... Supporting shaft, 7 ... Pulling up shaft, 8 ... Heater , 9 ... heat insulating material, 10 ... arithmetic device, 11 ... camera, 12 ... load cell, 13 ... cone part, 14 ... straight body part, 20 ... single crystal manufacturing apparatus.

Claims (6)

  A method for detecting the diameter of a single crystal grown by the Czochralski method, wherein the diameter of the single crystal is detected by both the camera and the load cell, respectively, and the difference between the camera detection diameter and the diameter calculated by the load cell, Cone shape correction determined from the relationship between the correction coefficient α determined in advance according to the growth rate of the single crystal, the parameter representing the shape of the cone portion of the single crystal, and the diameter of the straight body portion of the single crystal grown in advance. And correcting the detected diameter by the camera, and using the value obtained by the correction as the diameter of the single crystal.   The correction is performed by multiplying the difference between the camera detection diameter and the diameter calculated by the load cell by the correction coefficient α or adding the value and the cone shape correction to the camera detection diameter. The method for detecting a single crystal diameter according to claim 1.   The parameter is any one of a length of the cone portion of the single crystal, a weight of the cone portion of the single crystal, and a shape constant obtained from the length and weight of the cone portion of the single crystal. The method for detecting a single crystal diameter according to claim 1 or 2. The shape constant (C) is obtained by the following formula (1) using the length (Lc) of the cone portion of the single crystal and the weight (Wc) of the cone portion of the single crystal. The method for detecting a single crystal diameter according to 1.
C = Lc ³ / Wc (1)   The cone shape correction is performed by using only the correction coefficient α obtained in advance according to the difference between the camera detection diameter and the diameter calculated by the load cell and the growth rate of the single crystal in the single crystal diameter detection method. The detected diameter is corrected, and the single crystal is grown in advance while detecting the diameter of the single crystal using the value obtained by the correction as the diameter of the single crystal. 5. The method according to claim 1, wherein the distance is obtained using a least square method from a relationship between an actual diameter at an arbitrary point of 120 mm and a parameter representing a shape of a cone portion of the single crystal. For detecting the diameter of a single crystal.   6. A single crystal pulling apparatus for growing a single crystal by the Czochralski method, comprising at least both a camera and a load cell for detecting the diameter of the single crystal to be pulled, and any one of claims 1 to 5. A single crystal pulling apparatus, wherein the single crystal diameter is detected by the method for detecting a single crystal diameter according to item 1.

Images