JP2876050B2 - Crystal diameter measurement method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for measuring the diameter of a single crystal with high optical precision when pulling the single crystal by the CZ method (Czochralski method).

[0002]

2. Description of the Related Art One of the methods for producing a single crystal as a semiconductor raw material is pulling by a CZ method. In the CZ method, as shown in FIG. 1, a crucible 2 installed in a furnace body 1 of a CZ furnace is filled with a crystal melt 3, and a single crystal 4 is rotated by the rotating device 6 from the melt 3. While being pulled up by the pulling device 5. At this time, in order to make the heating of the melt 3 by the heater 7 uniform, the crucible 2 is controlled to rise so that the distance from the heater 7 to the liquid surface is kept constant.

It is desirable that the single crystal 4 being pulled has the same shape as the target at the upper end and the lower end, and the uniform diameter equal to the target value in the straight body portion and the seed crystal portion. desired. In addition, single crystal 4
It is also necessary to set the deformation ratio [(maximum diameter−minimum diameter) / minimum diameter] representing the deviation of the cross-sectional shape from the perfect circle to an allowable value or less.

On the other hand, regarding the crystal quality, it is necessary to keep the density of oxidation-induced stacking faults (hereinafter referred to as OSF), which is one of the evaluation items, low. OSF refers to stacking faults caused by a phenomenon in which oxygen dissolved in a crystal precipitates as an oxide during heat treatment for oxidizing the crystal. This OS
If the pulling speed is increased, the density of F can be kept low because the crystal can be rapidly cooled by that amount. Therefore, it is necessary to increase the pulling speed. Increasing the pulling speed is also beneficial in improving production efficiency.

[0005] However, when the pulling speed is increased, the deformation rate increases and exceeds the allowable value, so that the yield decreases. Therefore, it is necessary to set an optimum pulling speed at which the pulling can be performed within an allowable range of the deformation rate, from the viewpoint of improving the yield of the single crystal, improving the productivity, securing quality, and the like. Here, it is important to accurately measure the diameter of the single crystal being pulled and calculate an accurate deformation rate.

As a method of measuring the diameter of a single crystal during pulling by the CZ method, there are a method of calculating the diameter from the weight of the pulled single crystal (hereinafter referred to as a gravimetric method) and a one-dimensional CC method.
There are two known methods for measuring the diameter using an optical device such as a D camera (hereinafter referred to as an optical method).

By the way, in pulling a single crystal by the CZ method, as shown in FIG. 3, protrusions 4a called habit lines are regularly formed on the outer peripheral surface of the single crystal 4 in the circumferential direction. The projection 4a extends in the crystal axis direction and occurs at a circumferential position unique to the crystal orientation of the single crystal 4. When calculating the deformation rate, it is necessary to accurately measure the diameter of this habit line, but in the gravimetric method, only the average diameter is measured because the crystal diameter is calculated from the weight and length of the pulled single crystal. It is not possible to measure the diameter of the habit line part. In this regard, the optical method can measure the diameter of a habit line portion in order to optically measure the diameter of a high-intensity fusion ring generated at the interface between the melt and the single crystal.

In this optical method, as shown in FIG.
The single crystal 4 and the melt 3 are passed through a window 9 provided at the upper end of the furnace body 1.
Is linearly measured by the one-dimensional CCD camera 8. Then, as shown in FIG. 2, the positions of the intersections C and C are determined from the luminance change at the intersections C and C of the fusion ring A generated around the single crystal 4 and the photometric line BB of the one-dimensional CCD camera 8. Then, the diameter of the single crystal 4 is measured.

More specifically, while the single crystal 4 makes one rotation, the position of the intersections C and C is continuously detected, and the distance W (α) between the intersections C and C is obtained by the following equation, whereby the entirety of the single crystal 4 is obtained. Measure its diameter over the circumference. W (α) = L (α) −R (α) L (α), R (α): Position detection data of intersections C, C α: Rotation angle of single crystal

In this case, if the one-dimensional CCD camera 8 is installed so that the photometric line BB of the one-dimensional CCD camera 8 passes through the center O of the crystal, the fusion ring A is shaded by the single crystal 4 when the crystal diameter decreases. This may cause a measurement error or, in some cases, make the diameter measurement impossible. Therefore, usually, the photometric line BB of the one-dimensional CCD camera 8 is set on the camera side (front side) of the crystal center O. In this case, the crystal diameter is calculated by the following equation from the distance W between the intersections C, C measured by the one-dimensional CCD camera 8. D = (W ² + 4a ² ) ^1/2 D: crystal diameter W: distance between intersections C, C a: distance from crystal center O to photometric line BB

[0011]

In the conventional optical method, as described above, the interval W between the intersections C, C is obtained from the difference between the position detection data L (α) and R (α) at the intersections C, C. When the crystal orientation of the single crystal is (100), as shown in FIG. 3, crystal habit lines 4a are formed at intervals of 90 ° on the outer surface of the single crystal 4, and the photometric line BB of the one-dimensional CCD camera 8 is formed. When passing through the crystal center O, the two habit lines 4a, 4a located at the target positions across the crystal center O pass through the photometric line BB at the same time, so that the position detection data L (α) of the intersections C, C , R
From the difference of (α), the diameter of the single crystal 4 including the vicinity of the crystal habit line 4a is measured around the entire circumference with relatively high accuracy.

However, in the actual pulling, the photometric line BB is separated from the crystal center O as described above. In that case, the two habit lines 4a, 4a located at the target positions with the crystal center O interposed therebetween do not pass through the photometric line BB at the same time, but pass through one after the other. Therefore, intersections C and C
From the difference between the position detection data L (α) and R (α)
In the conventional optical method for obtaining the interval W of C, the diameter measurement accuracy is significantly reduced near the habit line 4a.

In the current crucible lifting control, an error occurs in the liquid level position because an accurate liquid level detection method has not been put to practical use. As a result, one-dimensional CCD
The photometric line BB of the camera 8 shifts from the initial set position, and the distance a from the crystal center O to the photometric line BB fluctuates. Therefore, the measured diameter D includes an error.

To solve this problem, a one-dimensional CCD
A method of moving the photometric line BB of the camera 8 in a direction perpendicular to the line and obtaining a true diameter value from the crystal diameter measured before and after the movement and the moving distance of the photometric line BB is as follows.
It is proposed by JP-A-63-256594. However, even with this method, a decrease in the accuracy of diameter measurement near the habit line caused by the photometry line being separated from the crystal center is inevitable.

An object of the present invention is to prevent a decrease in the accuracy of diameter measurement near a habit line caused by a photometry line being separated from a crystal center, thereby preventing the photometry line from being separated from the crystal center. Another object of the present invention is to provide a crystal diameter measuring method capable of accurately measuring the crystal diameter over the entire circumference.

[0016]

As described above, a single crystal pulled by the CZ method has a habit line at a position in the outer circumferential direction specific to the crystal orientation. For example, if the crystal orientation is (10
In the case of (0), habit lines are generated every 90 °. Since the single crystal being pulled is rotating in the circumferential direction, a one-dimensional CCD
When the crystal habit line crosses the photometry line of the camera, the position of the intersection of the fusion ring and the photometry line fluctuates. (10
In the case of 0), the position of the intersection changes every 90 °.

When the photometric line passes through the center of the crystal, the change in the position of the intersection occurs simultaneously at the detection positions on both sides. However, when the photometric line moves away from the center of the crystal, the timing of occurrence of the change in the position of the intersection on both sides is shifted. Further, the timing difference becomes larger as the distance from the crystal center to the photometric line becomes longer.

The method for measuring the diameter of a single crystal according to the present invention is a method for pulling a single crystal from a crystal melt by rotating the single crystal in the circumferential direction by the CZ method. When measuring the position of the interface of the liquid and detecting the positions of the intersections on both sides from the change in brightness at the intersection of the fusion ring generated around the single crystal and the photometry line of the camera, the positions of the intersections on both sides are detected independently, From both the detection data, the timing difference of the intersection position change on both sides due to the habit line intermittently generated with the rotation of the single crystal,
By removing the timing difference and comparing the two detection data, it is possible to prevent a decrease in diameter measurement accuracy near the crystal habit line caused by the photometry line being separated from the crystal center.

Further, by using the timing difference, it is possible to correct the distance from the crystal center to the photometric line of the line sensor.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below.

In the diameter measuring method of the present invention, as shown in FIG. 1, a one-dimensional CCD camera 8 installed outside the furnace body 1 of the CZ method is used. In this example, a one-dimensional CCD camera is used, but a two-dimensional CCD or the like can be used. The camera 8 linearly measures the melt surface in the crucible 1 through a window 9 formed in the furnace body. Then, as shown in FIG. 2, the fusion ring A generated around the single crystal 4 being pulled and the photometric line BB of the one-dimensional CCD camera 8 intersect, and the intersection C is determined based on the luminance change on the photometric line BB. , C
Are independently detected. This detection is a single crystal 4
Is rotated once, and continues at a constant pitch corresponding to the rotation angle.

The single crystal pulled by the CZ method is shown in FIG.
As shown in (1), a habit line 4a is formed at a position in the outer circumferential direction specific to the crystal orientation. The figure shows the case of the crystal orientation (100). In this case, four habit lines 4 are formed every 90 °.
a occurs. Thus, the single crystal 4 being pulled is rotating in the circumferential direction. Therefore, when the crystal habit line 4a crosses the photometric line BB, the position of the intersection C fluctuates.

Now, assuming that the photometric line BB passes through the crystal center O, the positional fluctuation of the intersections C, C on both sides caused by passing the photometric line BB through the crystal habit line 4a is on both sides. It will be the same timing. That is, when a certain habit line 4a passes through the photometric line BB, the crystal center O
The crystal habit line 4a at the target position across the line is also a photometric line BB
Pass through.

However, in actual pulling, the photometric line B
-B is separated from the crystal center O to the one-dimensional CCD camera 8 side (front side). Therefore, the intersections C and C on both sides generated by passing the photometric line BB through the habit line 4a.
Is different timing on both sides. For example, when the single crystal 4 is rotating clockwise, the crystal habit line 4a passes through the photometric line BB at the left measurement position, causing the position of the intersection C to change, and then the photometry line at the right measurement position. B-
The crystal habit line 4a passes through B, and the position of the intersection C changes, and the position of the intersections C and C on both sides changes in timing.

FIG. 4A shows intersection position data L on both sides measured by a one-dimensional CCD camera when a single crystal having a crystal orientation (100) is pulled while rotating clockwise.
(Α) and R (α) are shown over time. With the rotation of the single crystal, the intersection position data L (α) and R (α) on both sides are changed to the habit line 4.
Although it fluctuates due to the passage of “a”, the fluctuation on the right is detected with a delay of θ from the fluctuation on the left.

In the crystal diameter measuring method of the present invention, the positions of the intersections C, C on both sides are independently detected, and the timing of the change of the intersection positions on both sides is determined from the deviation of the respective intersection position data L (α), R (α). Find the difference θ. Then, in order to remove the timing difference θ, an interval W (α) between the intersections C on both sides is obtained using the following equation. W (α) = R (α + θ) -L (α)

When the single crystal 4 is pulled while rotating counterclockwise, the distance W between the intersections C on both sides is calculated by the following equation.
(Α) is obtained. W (α) = R (α) −L (α + θ)

As a result, as shown in FIG. 4B, the timing difference between the intersection positions on both sides is removed. Therefore, the interval W (α) between the intersections C on both sides is accurately measured as shown in FIG. The measured interval W (α) is converted into a crystal diameter D (α) using the following equation. D = (W ² + 4a ² ) ^1/2

Thus, the diameter of the single crystal 4 is accurately measured also in the vicinity of the habit line 4a.

Here, the distance a from the center of the crystal to the photometric line BB changes as the liquid level of the melt 3 moves up and down. The timing difference θ also changes. That is, as the distance a from the crystal center O to the photometric line BB increases, the timing difference θ also increases. Therefore, the distance a is corrected by the following equation using the timing difference θ. a = Wavg / 2 · sin ^-1 (θ / 2) Wavg: average interval of previous measurement

As a result, a measurement error due to a change in the level of the melt surface is also eliminated.

The pitch for detecting the positions of the intersections C, C is desirably 2 ° or less as the rotation angle of the single crystal being pulled in order to accurately grasp the shape near the habit line. This pitch is 1 °
When the present invention was carried out, the diameter passing around the habit line, which could not be accurately measured by the conventional method, can also be accurately measured, and the error in both the diameter and the deformation ratio is less than half that of the conventional method. It was confirmed that. Also,
Since the position of the habit line can be accurately detected from the measured diameter, the monitoring of the polycrystallization of the single crystal being pulled can be automated.

[0033]

As described above, the crystal diameter measuring method of the present invention detects the timing difference of the change of the position of the both-side intersection due to the habit line, and eliminates the measurement error due to the timing difference. The diameter can be measured with high accuracy over the entire circumference of the crystal including the above, and the improvement in the detection accuracy of the deformation rate and the like can greatly improve the yield, the productivity, and the quality.

[Brief description of the drawings]

FIG. 1 is a diagram showing the configuration of an apparatus for pulling a single crystal by the CZ method.

FIG. 2 is a conceptual diagram of a method for measuring the diameter of a single crystal.

FIG. 3 is a schematic plan view showing a positional relationship between a photometric line of a line sensor and a crystal center.

FIG. 4 is a graph showing a change over time of position data at both-side intersections.

FIG. 5 is a graph showing a change with time of an intersection interval.

[Explanation of symbols]

 Reference Signs List 1 furnace body 3 melt 4 single crystal 8 one-dimensional CCD camera A fusion ring B photometric line C intersection of fusion ring and photometric line O crystal center

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int. Cl. ⁶ , DB name) C30B 15/20-15/28 C30B 28/00-35/00

Claims (2)

(57) [Claims] In the pulling of a single crystal by a CZ method, which pulls a single crystal from a crystal melt while rotating the single crystal in a circumferential direction,
When measuring the position of the interface between the single crystal and the melt using a camera installed diagonally above, and detecting the intersection point on both sides from the luminance change at the intersection of the fusion ring around the single crystal and the photometry line of the camera , Independently detect the intersection positions on both sides, and from both the detected data, find the timing difference of the intersection position change on both sides due to the habit line intermittently generated with the rotation of the single crystal, remove the timing difference, And measuring the diameter of the single crystal by comparing the detection data of the above and the distance between the intersections on both sides. 2. The crystal diameter measuring method according to claim 1, wherein a distance from a crystal center to a photometric line of a camera is corrected using the timing difference.