Method for producing single crystal silicon and device for producing single crystal silicon

Abstract

The purpose of the present invention is to early detect leakage due to breakage of a quartz crucible or deformation of the quartz crucible to prevent secondary damage. In a crystal pulling step, a camera (20) is used to capture an image of a melt surface (2a) of a silicon melt (2) in a quartz crucible (11), a gap value (hGAP) between a lower end of a heat shield (17) disposed above the melt surface (2a) and the melt surface (2a) is obtained from the captured image of the camera (20), and height control of the quartz crucible (11) is performed on the basis of a change in the gap value (hGAP). In the height control of the quartz crucible (11), the correction amount of the crucible rising speed is obtained from the difference between the measured value and the target value of the gap value (hGAP), and the presence or absence of a sudden change in the liquid level position is determined on the basis of whether or not the correction amount exceeds a threshold value.

Description

Technical Field

[0001] This invention relates to a method and apparatus for manufacturing monocrystalline silicon. Background Technology

[0002] The Czeklaussky process (CZ process) is a known method for manufacturing single-crystal silicon for semiconductor devices. The CZ process involves immersing a seed crystal in molten silicon within a quartz crucible, then slowly pulling it up while rotating it relative to the seed crystal, thereby allowing a large single crystal to grow at the lower end of the seed crystal. The CZ process can improve the crystal quality and manufacturing yield of large-diameter single-crystal silicon.

[0003] In the CZ process, to improve the yield and quality of single crystals, precise measurement and control of the crystal diameter and liquid level are performed. Regarding methods for measuring the crystal diameter and liquid level, for example, Patent Document 1 describes a method where the crystal diameter and crystal center position are calculated from a high-brightness portion called a fusion ring generated at the solid-liquid interface, and the liquid level is calculated based on the crystal center position. Furthermore, Patent Document 2 describes a method where the liquid level position of the silicon melt relative to the heat shield is calculated based on the interval between the real image of the heat shield containing a circular opening and its mirror image projected onto the molten surface.

[0004] In the CZ process, the quartz crucible is raised in a controlled manner to maintain a constant level of the molten silicon surface as it descends due to the growth of single-crystal silicon. This allows for the stable extraction of single-crystal silicon from the molten silicon, thereby stabilizing the crystal quality and oxygen concentration distribution along the crystal growth direction.

[0005] During the crystal pulling process, damage or deformation of the quartz crucible can sometimes occur. If the crucible breaks, molten silicon leaks, damaging the electrodes at the bottom of the chamber or the heater, and can cause malfunctions in the crucible drive mechanism. Furthermore, there is a risk of steam explosion due to damage to the cooling water piping. If the crucible deforms, it may come into contact with furnace components such as the heat shield or with the single-crystal silicon being pulled, making it difficult to continue the crystal pulling process.

[0006] Previously, leakage from quartz crucibles was detected by visual observation of the furnace interior or by changes in furnace pressure or crucible rotation. Furthermore, deformation of the quartz crucible was visually confirmed during or after the crystal pulling process.

[0007] Regarding methods for detecting leakage of molten silicon in a quartz crucible, for example, Patent Document 3 describes a method for accurately detecting leakage by combining a first detection mechanism and a second detection mechanism. The first detection mechanism uses a CCD camera to capture a high-brightness area called a fusion ring generated at the point where the single crystal silicon contacts the molten silicon during crystal pulling, and performs image processing on the image to detect leakage. The second detection mechanism intermittently applies an AC or DC voltage from a power source between the crucible support shaft and the wire holding the seed crystal, and detects leakage by comparing changes in conductivity state using a comparator.

[0008] Existing technical documents Patent documents Patent Document 1: Japanese Patent Application Publication No. 2019-85299 Patent Document 2: Japanese Patent Application Publication No. 2013-216505 Patent document 3: Japanese Patent Application Publication No. 11-180794. Summary of the Invention

[0009] The technical problem that the invention aims to solve However, the aforementioned conventional methods for detecting quartz crucibles cannot quantitatively detect abnormalities caused by crucible breakage or deformation, nor can they detect abnormalities at an early stage. Furthermore, the leakage detection method described in Patent Document 3 cannot accurately detect leakage in processes where the single-crystal silicon and molten silicon are not in contact. Moreover, it cannot detect changes in the liquid level caused by deformation of the quartz crucible. In the event of leakage or crucible deformation during the crystal pulling process, significant damage can occur to the apparatus or chamber, thus necessitating early detection.

[0010] Therefore, the object of the present invention is to provide a method and apparatus for manufacturing monocrystalline silicon that can detect early damage or deformation of quartz crucibles.

[0011] Solutions for solving technical problems To address the aforementioned problems, the method for manufacturing monocrystalline silicon according to the present invention is characterized in that, during the crystal pulling process, a camera is used to photograph the molten silicon surface in a quartz crucible; the gap value between the lower end of a heat shield disposed above the quartz crucible and the molten surface is determined based on the photographed image; the height control of the quartz crucible is performed based on the change in the gap value; in the height control of the quartz crucible, a correction amount for the crucible rising speed is determined based on the difference between the measured value and the target value of the gap value; and whether there is a sudden change in the liquid surface position is determined based on whether the correction amount exceeds a threshold.

[0012] In the method for manufacturing monocrystalline silicon of the present invention, preferably, the crystal pulling process is interrupted when a sudden change in the liquid level is detected.

[0013] This improves the safety of the crystal pulling process.

[0014] In the method for manufacturing monocrystalline silicon of the present invention, preferably, when a sudden change in the liquid surface position is detected, it is determined whether there is any abnormality in the edge detection result of the furnace structure shown in the photographic image. If there is no abnormality in the edge detection result, the crystal pulling process is interrupted. In this case, preferably, the furnace structure is the heat shield, and the gap value is calculated based on the real image edge of the heat shield and the mirror image edge of the heat shield mapped onto the molten surface. This avoids unnecessary interruptions to the crystal pulling process and improves the reliability of the crystal pulling process.

[0015] Preferably, the threshold has an upper and lower limit for the correction amount, and is set for each crystal pulling process based on past data on crucible rise speed variations in monocrystalline silicon mass production. This improves the accuracy of determining whether there are abrupt changes in the liquid level.

[0016] Furthermore, the monocrystalline silicon manufacturing apparatus of the present invention is characterized by comprising: a chamber; a quartz crucible holding molten silicon within the chamber; a heater disposed around the quartz crucible and heating the molten silicon; a crucible driving mechanism driving the rotation and lifting of the quartz crucible; a crystal pulling mechanism pulling monocrystalline silicon from the molten silicon; a heat shield disposed above the quartz crucible in a manner surrounding the pulling path of the monocrystalline silicon; a camera capturing images of the interior from the outside of the chamber; an image processing unit processing the images captured by the camera; and a control unit controlling the [device / processor] based on the processing results of the image processing unit. The image processing unit of the heater, the crucible driving mechanism, and the crystal pulling mechanism performs the following processing: during the crystal pulling process, the camera is used to photograph the molten silicon surface in the quartz crucible; and the gap value between the lower end of the heat shield and the molten surface is determined based on the photographed image; the control unit performs height control of the quartz crucible based on the change of the gap value; in the height control of the quartz crucible, the correction amount of the crucible rising speed is determined based on the difference between the measured value and the target value of the gap value, and whether there is a sudden change in the liquid surface position is determined based on whether the correction amount exceeds a threshold.

[0017] According to the present invention, abrupt changes in the liquid level caused by damage or deformation of the quartz crucible can be detected at an early stage based on changes in the rising speed of the quartz crucible. Therefore, secondary damage such as device damage caused by damage or deformation of the quartz crucible can be prevented.

[0018] Invention Effects According to the present invention, a method and apparatus for manufacturing single-crystal silicon that can detect early damage or deformation of quartz crucibles are provided. Attached Figure Description

[0019] Figure 1 This is a schematic cross-sectional view showing the structure of a single-crystal silicon manufacturing apparatus according to an embodiment of the present invention.

[0020] Figure 2 This is a flowchart illustrating the manufacturing process of monocrystalline silicon.

[0021] Figure 3 This is a schematic side view showing the shape of a single-crystal silicon ingot.

[0022] Figure 4 This is a flowchart illustrating an example of a method for controlling the height of a quartz crucible.

[0023] Figure 5 This is a flowchart illustrating another example of a method for controlling the height of a quartz crucible.

[0024] Figure 6 This is a schematic diagram illustrating an example of edge detection results. (a) indicates a normal edge detection result, and (b) indicates an abnormal edge detection result.

[0025] Figure 7 It is a graph showing the change in the crucible rising correction speed caused by gap control. Detailed Implementation

[0026] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0027] Figure 1 This is a schematic cross-sectional view showing the structure of a single-crystal silicon manufacturing apparatus according to an embodiment of the present invention.

[0028] like Figure 1 As shown, the single-crystal silicon manufacturing apparatus 1 includes a water-cooled chamber 10, a quartz crucible 11 holding molten silicon 2 in the chamber 10, a graphite crucible 12 holding the quartz crucible 11, a rotating shaft 13 supporting the graphite crucible 12, a crucible drive mechanism 14 driving the quartz crucible 11 to rotate and lift via the rotating shaft 13 and the graphite crucible 12, a heater 15 disposed around the graphite crucible 12, a heat-insulating material 16 disposed outside the heater 15 and along the inner surface of the chamber 10, a heat shield 17 disposed above the quartz crucible 11, a lifting wire 18 disposed above the quartz crucible 11 and coaxial with the rotating shaft 13, a crystal lifting mechanism 19 disposed above the chamber 10, a camera 20 for capturing images inside the chamber 10, an image processing unit 21 for processing images captured by the camera 20, and a control unit 22 for controlling the various parts of the apparatus.

[0029] 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 shield 17 are disposed within the main chamber 10a. A gas inlet 10c is provided in the crystal pulling chamber 10b for introducing inert gases (purge gases) such as argon and dopant gases into the main 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. Furthermore, an observation window 10e is provided at the upper part of the main chamber 10a to observe the growth status of the single-crystal silicon 3.

[0030] The quartz crucible 11 is a container made of quartz glass with cylindrical sidewalls and a curved bottom. The graphite crucible 12 holds the quartz crucible 11 in close contact with and covers its outer surface, thus maintaining the shape of the quartz crucible 11 as it softens upon heating. The quartz crucible 11 and the graphite crucible 12 constitute a double-structured crucible that supports the molten silicon 2 within the chamber 10.

[0031] 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 rotation mechanism and lifting mechanism of the quartz crucible 11. The rotation and lifting movements of the quartz crucible 11 driven by the crucible drive mechanism 14 are controlled by the control unit 22.

[0032] Heater 15 is used to melt the silicon raw material filled in the quartz crucible 11 to generate silicon molten liquid 2 and maintain the molten silicon molten liquid 2 in a molten state. Heater 15 is a carbon-made resistance heater, arranged to surround the quartz crucible 11 inside the graphite crucible 12. Furthermore, heat-insulating material 16 is provided on the outside of heater 15 to surround heater 15, thereby improving the heat preservation within the chamber 10. The output of heater 15 is controlled by control unit 22.

[0033] The heat shield 17 is configured to suppress temperature fluctuations in the molten silicon 2 to provide appropriate heat distribution near the crystal growth interface and to 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 configured to cover the area above the molten silicon 2, excluding the pulling path of the single-crystal silicon 3.

[0034] The diameter of the opening at the lower end of the heat shield 17 is larger than the diameter of the monocrystalline silicon 3, thereby ensuring the pulling path of the monocrystalline silicon 3. Furthermore, since the outer diameter of the lower end of the heat shield 17 is smaller than the diameter of the quartz crucible 11 and the lower end of the heat shield 17 is located inside the quartz crucible 11, even if the upper edge of the quartz crucible 11 is raised to a position higher than the lower end of the heat shield 17, the heat shield 17 will not interfere with the quartz crucible 11.

[0035] Although the amount of molten material in the quartz crucible 11 decreases as the single-crystal silicon 3 grows, raising the quartz crucible 11 increases the distance between the molten surface 2a and the heat shield 17 (gap value h). GAP The flow rate of the gas flowing near the surface 2a of the molten silicon is kept constant, thus controlling the amount of dopant evaporating from the molten silicon. This gap control improves the stability of crystal defect distribution, oxygen concentration distribution, resistivity distribution, and other properties along the pulling axis of the single-crystal silicon 3.

[0036] Above the quartz crucible 11, a pulling shaft (i.e., wire 18) for single-crystal silicon 3 and a crystal pulling mechanism 19 for pulling the single-crystal silicon 3 by winding the wire 18 are arranged. The crystal pulling mechanism 19 has the function of rotating the single-crystal 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 disposed 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 front end of the wire 18 reaching the internal space of the main chamber 10a. Figure 1 The image shows the monocrystalline silicon 3 suspended on the wire 18 during the growth process. When pulling the monocrystalline silicon 3, the wire 18 is slowly pulled up while the quartz crucible 11 and the monocrystalline silicon 3 are rotated, thereby growing the monocrystalline silicon 3.

[0037] A camera 20 is disposed outside 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 has a camera axis (optical axis) that is tilted relative to the lifting axis of the single crystal silicon 3, and the setting angle of the camera 20 forms a predetermined angle with respect to the vertical direction. 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 shield 17 and the liquid surface of the silicon melt 2, from an obliquely upward position.

[0038] Camera 20 is connected to image processing unit 21, and image processing unit 21 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 shown in the photographed image from camera 20. Furthermore, image processing unit 21 calculates the distance (gap value h) from the heat shield 17 to the liquid surface position based on the mirror image of the heat shield 17 mapped onto the molten surface 2a in the photographed image from camera 20.GAP To eliminate the influence of noise, it is preferable to use the moving average of multiple measured values ​​(instantaneous values) as the clearance value h for actual clearance control. GAP .

[0039] The method for calculating the gap value based on the mirror position of the heat shield 17 is not particularly limited. For example, the gap value can be calculated by preparing a conversion table or formula in advance that represents the relationship between the mirror position of the heat shield 17 and the gap value, and then substituting the mirror position of the heat shield 17 into the conversion table or formula during the crystal pulling process. Furthermore, the gap value can also be calculated geometrically based on the positional relationship between the real image and the mirror image of the heat shield 17 shown in the photographic image.

[0040] The control unit 22 controls the crystal diameter by controlling the crystal pulling speed based on crystal diameter data obtained from the photographed image from the camera 20. Specifically, when the measured crystal diameter is larger than the target diameter, the crystal pulling speed is increased; when it is smaller than the target diameter, the crystal pulling speed is decreased. Furthermore, the control unit 22 controls the crystal diameter based on the crystal length data of the single-crystal silicon 3 obtained from the sensor of the crystal pulling mechanism 19 and the gap value h obtained from the photographed image from the camera 20. GAP The movement (cruise rise speed) of the quartz crucible 11 is controlled by the liquid level to achieve a specified gap value. Regarding the movement of the quartz crucible 11, besides adjusting the gap value h... GAP Besides maintaining a constant value for control, as the single crystal is pulled, there is also a possibility that the gap value h will increase. GAP Control it to make it gradually decrease, or conversely, control it to make it increase.

[0041] A cylindrical shield 23 surrounding the crystal pulling axis can be provided above the heat shield 17. This shield 23 can be a structure called a purge tube, or a cooling body that promotes the cooling of the pulled single crystal silicon 3.

[0042] The purge tube is provided to control the flow of the purge gas. To adjust the resistivity of single-crystal silicon to suit the characteristics of semiconductor devices, impurities (dopants) such as arsenic (As) and antimony (Sb) are sometimes doped into the molten silicon. These dopants have low boiling points and are easily evaporated. In conventional Czochralski (CZ) crystal pulling, since purge gases such as Ar are introduced into the Czochralski furnace under reduced pressure, dopants evaporating from the molten silicon 2 diffuse with the purge gas, contaminating the furnace. Furthermore, the heat shield 17 installed in the furnace accelerates the purge gas flowing near the surface of the molten silicon 2, further promoting the evaporation of dopants from the molten silicon 2. However, by providing a purge tube to rectify the purge gas introduced into the Czochralski furnace, a high-pressure state is created within the chamber, and the evaporation of dopants in the molten silicon can be suppressed.

[0043] The cooling element is configured to control the time it takes for the single-crystal silicon pulled from the molten silicon 2 to pass through a specified temperature range. It is known that the type or distribution of crystal defects in single-crystal silicon produced by the CZ method depends on the ratio V / G of the single-crystal silicon growth rate (pulling speed) V to the temperature gradient G along the pulling axis near the crystal growth interface from the melting point to 1300°C. By strictly controlling V / G, it is possible to produce single crystals that do not contain COPs (Crystal Originated Particles) or dislocation clusters.

[0044] Here, if the crystal diameter increases, the center of the crystal becomes more difficult to cool compared to the outer periphery, and the temperature gradient G within the cross-section of the single-crystal silicon orthogonal to the pulling axis easily becomes uneven. Consequently, the allowable width of V / G that enables the entire surface of the single-crystal silicon cross-section orthogonal to the pulling axis to become a defect-free region becomes very narrow, and controlling the crystal pulling speed V becomes drastically difficult. However, when a cylindrical cooling body is provided above the heat shield 17, the allowable width (PVPi tolerance) of the crystal pulling speed V that enables the entire surface of the single-crystal silicon cross-section orthogonal to the pulling axis to become a defect-free region can be expanded, thereby improving the manufacturing yield of large-diameter single-crystal silicon that does not contain COPs and dislocation clusters.

[0045] Figure 2 This is a flowchart illustrating the manufacturing process of monocrystalline silicon. Furthermore, Figure 3 This is a schematic side view showing the shape of a single-crystal silicon ingot.

[0046] like Figure 2 As shown, the manufacturing process of monocrystalline silicon includes a melt generation process S11, in which polycrystalline silicon raw material in quartz crucible 11 is heated by heater 15 to generate silicon melt 2; an immersion process S12, in which seed crystal installed at the front end of wire 18 is lowered and immersed in silicon melt 2; and a crystal pulling process S13, in which monocrystalline silicon 3 is grown by slowly pulling the seed crystal while maintaining contact with silicon melt 2.

[0047] In the crystal pulling process S13, a necking process S14 is performed sequentially to form a necking neck 3a that narrows the crystal diameter to prevent dislocation; a shoulder growth process S15 is performed to form a shoulder 3b that gradually increases in crystal diameter as the crystal grows; a body growth process S16 is performed to form a body 3c that maintains a constant crystal diameter; and a tail growth process S17 is performed to form a tail 3d that gradually decreases in crystal diameter as the crystal grows. Then, a cooling process S18 is performed to separate the single-crystal silicon 3 from the molten surface 2a to promote cooling. Through the above, a single-crystal silicon ingot 3i having a necking neck 3a, a shoulder 3b, a body 3c, and a tail 3d is completed.

[0048] In the crystal pulling process S13, especially the body growth process S16, control is performed to slowly raise the quartz crucible 11 so that the gap (gap value) between the lower end of the heat shield 17 and the molten surface 2a becomes constant.

[0049] Figure 4 This is a flowchart illustrating an example of a method for controlling the height of the quartz crucible 11.

[0050] like Figure 4 As shown, in controlling the height of the quartz crucible 11, a camera 20 is used to photograph the inside of the furnace at a specified sampling period, and the measured value h of the gap is determined based on the photographed images from the camera 20. GAP (Step S21). Then, based on the measured value h of the gap... GAP With the target value h GAPt The difference is used to calculate the gap correction amount Δh. GAP (Step S22). Target value h GAPt The gap correction amount Δh is determined based on a pre-defined gap data profile based on the crystal length. GAP The measured value h of the gap GAP Compared to the target value h GAPt In cases where the value is large, it becomes positive; the measured value h in the gap... GAP Compared to the target value h GAPt In cases where the value is small, it becomes negative.

[0051] Next, based on the gap correction amount Δh GAP Calculate the correction value ΔV for the crucible's rising speed. c (Step S23). Crucible rising speed V c To provide a quantitative value for the crucible's rising speed V cs Add correction amount ΔV c The obtained value is V. c =V cs +ΔV c The quantitative value of the crucible's rising speed V cs The correction value ΔV is determined based on a profile of crucible rising speed data predetermined by the crystal length. c Based on the gap correction amount Δh GAP Find the correction factor ΔV for the crucible's rising speed. c In the gap correction amount Δh GAP When it is positive, it becomes positive again, in the gap correction amount Δh GAP When the value is negative, it becomes negative. In other words, the measured value h of the gap... GAP Compared to the target value h GAPt In cases of large gaps, slightly increase the crucible's lifting speed Vc to reduce the gap size. Furthermore, the measured gap value h... GAP Compared to the target value hGAPt In cases where the size is small, slightly reduce the crucible's rising speed Vc to increase the gap.

[0052] Next, the correction amount ΔV for the crucible's rising speed is calculated. c With threshold ΔV cth The comparison (step S24). Here, regarding the threshold ΔV cth Based on past experience with crucible rise rate variations in monocrystalline silicon mass production, settings are configured for each crystal pulling process. This is because the allowable gap error or crucible rise rate error varies depending on the required quality of the monocrystalline silicon or the equipment environment. Furthermore, regarding the threshold ΔV... cth Set an upper limit threshold ΔV cuth and lower limit threshold ΔV clth Both of these. In other words, not only the correction amount ΔV for judging the crucible's rising speed. c Is it higher than the upper limit threshold ΔV? cuth Furthermore, it determines whether the value is below the lower threshold ΔV. clth This is because, not only can the liquid level drop rapidly due to leakage caused by damage to the quartz crucible 11, the liquid level can also rise rapidly due to deformation of the quartz crucible 11.

[0053] Typically, the crucible's rising speed V c The liquid level remains roughly constant, but in the event of leakage or volume change due to breakage of the quartz crucible 11, the liquid level position (gap value) changes abruptly, thus affecting the crucible's rising speed V. c The changes are also made to offset them. With continuous correction and control relative to the liquid level, it becomes difficult to determine whether the drop in liquid level is due to crystal growth or damage to the quartz crucible 11. In other words, it is difficult to visually detect minute changes in the liquid level and determine if there are any abnormalities.

[0054] Therefore, in this embodiment, the crucible rising speed V is monitored. c The change, and the correction amount ΔV in the crucible rising speed. c (The crucible rise correction speed) exceeds the threshold ΔV cth If the system determines that leakage is caused by damage to the quartz crucible 11 or that a sudden change in volume has occurred due to deformation of the quartz crucible, the crystal pulling process S13 is interrupted (step S24 "Yes", step S25). Furthermore, the correction amount ΔV of the crucible's rising speed is adjusted. c Threshold ΔV cth In the following cases, even based on the determined crucible rising speed V cBy controlling the height of the quartz crucible 11 and continuing the crystal pulling process S13, the possibility of an immediate accident is low (step S24 "No", step S26). Therefore, damage or deformation of the quartz crucible 11 can be detected early, preventing secondary damage.

[0055] Threshold ΔV cth This can be calculated based on past performance data of the crucible rise rate correction during mass production. For example, the average value ±3σ of the correction performance during mass production can be set as the threshold. Here, σ is the standard deviation. If the value exceeds the average ±3σ, an anomaly, including leakage, can be considered to have occurred.

[0056] Furthermore, when determining the threshold (average ± 3σ), it can be calculated using actual values ​​from both scenarios where abnormalities such as leakage or crucible deformation occur during the crystal pulling process, and scenarios where no abnormalities occur. Alternatively, it can be calculated solely based on actual values ​​from scenarios where no abnormalities occur. Calculating the threshold solely based on actual values ​​from scenarios where no abnormalities occur is a safer approach. This is because leakage or crucible deformation rarely occurs in practice, making it difficult to prepare data that includes these abnormalities.

[0057] Threshold ΔV cth It is preferable to set the gap based on the crystal growth length, because the gap correction amount can sometimes vary depending on the crystal growth length.

[0058] Figure 5 This is a flowchart illustrating another example of a method for controlling the height of the quartz crucible 11.

[0059] like Figure 5 As shown, the gap value h is obtained from... GAP Step (S21) involves comparing the correction amount ΔV of the crucible's rising speed. c With threshold ΔV cth Up to step (step S24), with Figure 4 The control method is the same. The characteristic of this control method lies in the correction amount ΔV for the crucible's rising speed. c Exceeding the threshold ΔV cthInstead of immediately interrupting the crystal pulling process S13, the edge detection results of furnace structures such as the heat shield 17 shown in the photographed image are evaluated (step S27). If the edge detection results are normal, the crystal pulling process S13 is interrupted (step S27 "No", step S25). Furthermore, if the edge detection results are abnormal (step S27 "Yes"), the crystal pulling process S13 does not need to be interrupted immediately; it can continue by eliminating the cause of the abnormal edge detection results. Sometimes, the abnormal edge detection results may be due to erroneous detection; therefore, for example, the edge detection processing can be restored to normal by switching to a backup camera.

[0060] In the gap value h GAP In the calculation, edge detection processing is performed on the real image of the heat shield 17 and the mirror image of the heat shield 17 mapped onto the molten surface 2a, and the gap value h is calculated based on the edge detection results. GAP Therefore, in cases where the edge detection results are abnormal, the calculated gap value h... GAP To the actual gap value h GAP The probability of discrepancies in outliers is relatively high, and incorrect gap control will occur when the height of the quartz crucible 11 is controlled based on such a gap value. However, in this embodiment, the correction amount ΔV for determining the crucible's rising speed is used. c Is the increase in gap value h truly due to damage to the quartz crucible 11, etc.? GAP Whether it is caused by changes in the image or by errors in the image processing results, it can prevent erroneous actions.

[0061] Figure 6 This is a schematic diagram illustrating an example of edge detection processing results. (a) indicates a normal edge detection result, and (b) indicates an abnormal edge detection result.

[0062] like Figure 6 As shown in (a) and (b), in the photographed image from the camera, molten silicon 2 is visible through the opening of the heat shield 17, and a mirror image of the heat shield 17 is mapped at the molten surface 2a. By binarizing the photographed image, the real image edge and the mirror image edge of the heat shield 17 can be detected, and the gap value can be calculated geometrically based on the real image edge and the mirror image edge of the heat shield 17.

[0063] Here, for example, Figure 6 As shown in (a), when the edge detection result is normal, for example, an edge line (approximate curve) can be obtained based on the mirrored edge. On the other hand, in cases where the edge detection result is abnormal, for example... Figure 6As shown in (b), a portion of the mirrored edge in the upper left corner of the diagram may be double-detected. If the approximate circle diameter calculated from the two edge lines detected from the mirrored edge exceeds the controllable range, the edge detection result is judged to be abnormal, and an error is reported. In the event of such an error, as mentioned above, even if the correction amount ΔV of the crucible rising speed... c Exceeding the threshold ΔV cth There is no need to interrupt the crystal pulling process S13.

[0064] As explained above, in the method for manufacturing single-crystal silicon in this embodiment, during the crystal pulling process S13, a camera 20 is used to photograph the molten surface 2a of the silicon melt 2 in the quartz crucible 11, and the gap value h is determined based on the photographed image from the camera 20. GAP According to the gap value h GAP The rise of the quartz crucible 11 is controlled by increasing the gap value, and the correction amount ΔV of the rising speed of the quartz crucible 11 is calculated based on the deviation between the measured gap value and the target value. c And based on this correction amount ΔV c Does it exceed the threshold ΔV? cth This allows for the early and automatic detection of damage or deformation of the quartz crucible 11 without relying on visual observation, thus preventing secondary damage such as device damage caused by breakage or deformation of the quartz crucible 11.

[0065] In the event of significant damage to the single-crystal pulling device 1 due to breakage of the quartz crucible 11, the damaged parts or components need to be replaced to restore it to its original state, thereby enabling the single-crystal pulling device 1 to be reused. However, the restoration process of this single-crystal pulling device 1 generates waste, thus placing a significant burden on the environment. According to the present invention, waste generation can be prevented, reducing the environmental burden and contributing to achieving the United Nations Sustainable Development Goals (SDGs) Goal 12, "Creating Responsibility, Using Responsibility" (preventing waste generation).

[0066] 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 also included within the scope of the present invention. Example

[0067] Multiple single-crystal silicon pulling operations were performed using the single-crystal silicon manufacturing apparatus of the present invention. As a result, most pulling batches were normal, but some pulling batches exhibited abnormal crucible rise speeds due to leakage.

[0068] Figure 7This is a graph showing the change in crucible rising correction speed due to gap control. The horizontal axis represents the crystal length (3c length of the body), and the vertical axis represents the crucible rising correction speed.

[0069] like Figure 7 As shown, in normal batches, the crucible rising correction speed does not change significantly, converging at the upper threshold ΔV used for anomaly detection. cuth To the lower limit threshold ΔV clth Within the acceptable range, but in the abnormal batch, the crucible rising correction speed changed drastically. In this abnormal batch, gap measurement results were confirmed at the time of the anomaly. The results showed no abnormalities in edge detection, therefore it was determined that the anomaly was more likely due to leakage than a control anomaly caused by incorrect gap measurement, and the crystal pulling process was interrupted. After turning off the heater power and cooling the chamber, the chamber was opened to the atmosphere and the situation was confirmed. Leakage traces were found from the side wall of the quartz crucible. However, the leakage was minimal, and the damage was contained to a minimum.

[0070] Explanation of reference numerals in the attached figures 1-Single crystal silicon manufacturing apparatus, 2-Silicon melt, 2a-Melt surface, 3-Single crystal silicon, 3a-Neck, 3b-Shoulder, 3c-Body, 3d-Tail, 3i-Single crystal silicon ingot, 10-Cavity, 10a-Main chamber, 10b-Crystal pulling chamber, 10c-Gas inlet, 10d-Gas outlet, 10e-Observation window, 11-Quartz crucible, 12-Graphite crucible, 13-Rotating shaft, 14-Cruise drive mechanism, 15 - Heater, 16 - Insulation material, 17 - Heat shield, 18 - Wire, 19 - Crystal pulling mechanism, 20 - Camera, 21 - Image processing unit, 22 - Control unit, 23 - Shielding, S11 - Melt generation process, S12 - Immersion process, S13 - Crystal pulling process, S14 - Necking process, S15 - Shoulder cultivation process, S16 - Body cultivation process, S17 - Tail cultivation process, S18 - Cooling process.

Claims

1. A method for manufacturing single-crystal silicon, characterized in that, A camera is used to photograph the molten silicon surface inside a quartz crucible during the crystal pulling process. The gap value between the lower end of the heat shield positioned above the quartz crucible and the molten surface is determined based on the photographic image captured by the camera. The height control of the quartz crucible is performed based on the change in the gap value. In the height control of the quartz crucible, the correction amount of the crucible's rising speed is calculated based on the difference between the measured value and the target value of the gap value, and the liquid level position is determined to be abruptly changed based on whether the correction amount exceeds the threshold.

2. The method for manufacturing monocrystalline silicon according to claim 1, wherein, When a sudden change in the liquid level is detected, the crystal pulling process is interrupted.

3. The method for manufacturing monocrystalline silicon according to claim 1, wherein, When a sudden change in the liquid level is detected, it is determined whether there is any abnormality in the edge detection result of the furnace structure shown in the photographic image. If there is no abnormality in the edge detection result, the crystal pulling process is interrupted.

4. The method for manufacturing monocrystalline silicon according to claim 3, wherein, The furnace internal structure is the heat shield, and the gap value is calculated based on the real image edge of the heat shield and the mirror image edge of the heat shield mapped onto the molten surface.

5. The method for manufacturing monocrystalline silicon according to claim 1, wherein, The threshold has an upper limit and a lower limit for the correction amount, and is set for each crystal pulling process based on the past performance of crucible rise speed variation in monocrystalline silicon mass production.

6. A single-crystal silicon manufacturing apparatus, characterized in that, The single-crystal silicon manufacturing apparatus includes: The system comprises: a chamber; a quartz crucible holding molten silicon within the chamber; a heater positioned around the quartz crucible and heating the molten silicon; a crucible drive mechanism driving the rotation and lifting of the quartz crucible; a crystal pulling mechanism pulling single-crystal silicon from the molten silicon; a heat shield positioned above the quartz crucible surrounding the pulling path of the single-crystal silicon; a camera capturing images of the interior from the outside of the chamber; an image processing unit processing the images captured by the camera; and a control unit controlling the heater, the crucible drive mechanism, and the crystal pulling mechanism based on the processing results of the image processing unit. The image processing unit performs the following processing: The camera is used during the crystal pulling process to photograph the molten silicon surface inside the quartz crucible; and The gap value between the lower end of the heat shield and the molten metal surface is determined based on the photographic image taken by the camera. The control unit performs height control of the quartz crucible based on the change in the gap value. In the height control of the quartz crucible, the correction amount of the crucible's rising speed is calculated based on the difference between the measured value and the target value of the gap value, and the liquid level position is determined to be abruptly changed based on whether the correction amount exceeds the threshold.

7. The monocrystalline silicon manufacturing apparatus according to claim 6, wherein, When a sudden change in the liquid level is detected, the control unit interrupts the crystal pulling process.

8. The monocrystalline silicon manufacturing apparatus according to claim 6, wherein, When a sudden change in the liquid level is detected, the control unit determines whether there is any abnormality in the edge detection result of the furnace structure shown in the photographic image. If there is no abnormality in the edge detection result, the crystal pulling process is interrupted.

9. The monocrystalline silicon manufacturing apparatus according to claim 8, wherein, The furnace internal structure is the heat shield, and the gap value is calculated based on the real image edge of the heat shield and the mirror image edge of the heat shield mapped onto the molten surface.

10. The monocrystalline silicon manufacturing apparatus according to claim 6, wherein, The threshold has an upper limit and a lower limit for the correction amount, and is set for each crystal pulling process based on the past performance of crucible rise speed variation in monocrystalline silicon mass production.

Images