A method for growing a silicon single crystal

Abstract

This invention provides a method for growing silicon single crystals without applying a magnetic field. The method includes: a welding stage, where the welding temperature is controlled within a preset silicon melt wetting temperature range to achieve dislocation-free welding between the seed crystal and the silicon melt; a crystal pulling stage, where the liquid nozzle distance, crucible rotation speed, and crystal rotation speed are maintained within preset constant process parameter ranges; a shoulder formation stage, where the preset liquid nozzle distance and crucible rotation speed are maintained, and the shoulder diameter is monitored in real time. When the shoulder diameter reaches a preset starting value within a preset multiple of the target crystal rod diameter, the crystal rotation speed is reduced from a first rotation speed to a second rotation speed until the shoulder formation ends; and a constant diameter growth stage, where the crystal rotation speed is maintained at a second rotation speed until the growth ends. The first rotation speed is a high speed sufficient to suppress natural convection in the early stages of shoulder formation, and the second rotation speed is no higher than 4 rpm. This application overcomes the technical contradiction of the incompatibility between shoulder stability and defect-free constant diameter growth, achieving high-quality single crystal growth under magnetic field-free conditions.

Description

Technical Field

[0001] This invention relates to the field of semiconductor material preparation technology, and specifically to a method for growing silicon single crystals. Background Technology

[0002] As integrated circuit manufacturing processes advance towards 7nm and below, higher requirements are being placed on the diameter of silicon component consumables. 450mm silicon wafers have become the next-generation mainstream specification, and their single-crystal growth diameter typically needs to reach 455mm or more. Currently, the industry commonly uses magnetic field-assisted Czochralski (MCZ) to suppress melt convection, but magnetic field equipment is expensive and energy-intensive. When growing ultra-large diameter single crystals without a magnetic field, melt thermal convection is intense, solid-liquid interface stability is poor, and problems such as shoulder fracture, dislocation multiplication, and doublet defects in equal-diameter segments are highly likely to occur.

[0003] Existing research often focuses on the influence of single parameters, lacking a systematic analysis of the coupled effects of multiple factors such as fusion temperature, liquid nozzle distance, crucible rotation speed, and crystal rotation speed. Especially for the critical size of 455 mm, even small deviations in each parameter can lead to growth failure.

[0004] In the field of large-diameter single-crystal silicon growth using the Czochralski method without a magnetic field, a directional technical understanding has long been established: to effectively suppress natural thermal convection of the melt and maintain the axisymmetric stability of the solid-liquid interface, the crystal rotation speed during the constant diameter stage should be maintained at a relatively high level, generally considered not to be lower than 5 rpm. For example, in process simulation studies of Ф300mm Czochralski single-crystal silicon without a magnetic field, 9–10 rpm is recommended as the optimal crystal rotation speed; in process studies of larger diameter single-crystal silicon, there are also reports of 12 rpm as the preferred crystal rotation speed for obtaining a stable solid-liquid interface; even under magnetic field-assisted conditions, related studies still regard 12 rpm as the preferred parameter for large-diameter single-crystal silicon.

[0005] For the non-magnetic field growth of 455mm crystals, there is a long-standing technical contradiction in the field: on the one hand, to suppress natural convection and ensure melt stability during the crystal pulling and shoulder formation stages, high crystal rotation and crucible rotation need to be maintained; on the other hand, after entering the constant diameter stage, the centrifugal effect generated by excessive crystal rotation will induce unsteady convection in the melt, leading to periodic oscillations in the solid-liquid interface temperature, and consequently, generating double-line defects that cannot be eliminated. Traditional process schemes using a single constant parameter cannot simultaneously meet the contradictory requirements of shoulder stability and constant diameter integrity, leading to the industry's general belief that it is impossible to stably pull qualified 455mm single crystals under non-magnetic field conditions. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention proposes a silicon single crystal growth method that aims to overcome the technical contradiction between achieving both shoulder stability and uniform diameter without defects, thereby realizing high-quality single crystal growth under magnetic field-free conditions.

[0007] In a first aspect, embodiments of this application provide a method for growing a single silicon crystal, wherein no magnetic field is applied throughout the silicon growth process, and the method for growing a single silicon crystal includes: Welding stage: Control the welding temperature to the preset silicon melt wetting temperature range to achieve dislocation-free welding between the seed crystal and the silicon melt; During the crystal-leading stage: the liquid nozzle distance, crucible rotation speed, and crystal rotation speed are controlled to remain within the preset constant process parameter ranges; Shoulder formation stage: Maintain the preset liquid nozzle distance and crucible rotation speed, monitor the shoulder diameter in real time, and when the shoulder diameter reaches the preset starting value within the preset multiple range of the target crystal rod diameter, start the crystal rotation speed to decrease from the first rotation speed to the second rotation speed until the shoulder formation ends; Constant diameter growth stage: Maintain the preset liquid nozzle distance and crucible rotation speed, control the crystal rotation speed to maintain the second rotation speed until the end; The first rotational speed is a high speed sufficient to suppress natural convection in the early stages of shoulder formation, and the second rotational speed is a low speed not exceeding 4 rpm. The second rotational speed is lower than the minimum rotational speed commonly used to suppress convection in non-magnetic field melts.

[0008] Optionally, during the shoulder-forming stage and the constant-diameter growth stage, the liquid outlet distance is kept constant at 50 mm, and the crucible rotation speed is kept constant at 12 rpm.

[0009] Optionally, the second rotational speed is 3.5 ± 0.2 rpm.

[0010] Optionally, the preset multiple ranges from 0.4 to 0.6 times.

[0011] Optionally, the preset multiple ranges from 0.44 to 0.55.

[0012] Optionally, the diameter of the target crystal rod is 420 mm to 470 mm.

[0013] Optionally, the diameter of the target crystal rod is 455±15mm; the preset starting value for the shoulder formation stage is 200mm~250mm.

[0014] Optionally, the preset silicon melt wetting temperature range is 1454±2℃.

[0015] Optionally, the preset silicon melt wetting temperature range is 1454±1℃.

[0016] Optionally, the preset constant liquid outlet distance process parameter range is 50±2mm; the preset constant crucible rotation speed process parameter range is 11-13rpm; and the preset constant crystal rotation speed process parameter range is 8-9rpm.

[0017] Optionally, the preset constant liquid outlet distance process parameter range is 50 mm; the preset constant crucible rotation speed process parameter range is 12 rpm.

[0018] Optionally, the preset constant crystal rotation speed process parameter is 8 rpm, that is, the first rotation speed is 8 rpm.

[0019] Optionally, the decrease is a smooth decrease.

[0020] Optionally, the smooth decrease is a linear uniform decrease, and the deceleration rate is calculated based on the preset starting value and the end diameter of the shoulder formation, so that the crystal rotation speed reaches the second rotation speed exactly when the shoulder formation ends.

[0021] Optionally, the silicon single crystal growth method is carried out in a magnetic field-free single crystal furnace with a 36-inch hot zone and a charge of 500 kg, and the weight of the blank accounts for more than 85% of the charge.

[0022] Optionally, the target crystal rod has a diameter of 455 mm, and the outer diameter of the grown single crystal blank is 455 ± 2 mm.

[0023] Compared with the prior art, the present invention has the following beneficial effects: (1) This application overcomes the technical prejudice in the field that "high crystal rotation under no magnetic field can achieve stable growth". It reduces the crystal rotation of equal diameter to 3.5±0.2rpm, which is lower than the lowest level of conventional understanding. Instead, it reconstructs the melt flow field and historically eliminates the double-line defect of the 455mm equal diameter section, producing unexpected technical effects.

[0024] (2) This application discovers that the 200-250mm shoulder diameter range is the critical initiation window for gradual crystal transformation, and reveals its physical essence—this window is the inevitable balance point between the contradiction between the geometric stability of the system and the flow field response speed. Initiating gradual crystal transformation within this window successfully resolves the contradiction between the demand for high crystal transformation in the shoulder stage and the demand for low crystal transformation in the constant diameter stage, and realizes the dynamic coordination of the two-stage operating conditions.

[0025] (3) This application, through staged coupling experiments, has clarified the critical boundary relationship between various parameters: only when the welding temperature is 1454±1℃, the liquid outlet distance is 50mm, the crucible rotation is 12rpm, the crystal rotation is 8rpm in the early stage of crystal shoulder formation, and the crystal rotation gradient is started when the shoulder diameter is 200-250mm, can the process threshold be crossed and defect-free growth be achieved. This shows that the parameters are not independently adjustable, but have a strong coupling and synergistic matching relationship.

[0026] (4) No magnetic field is applied throughout the entire process of this application, which saves the purchase and operation energy consumption of superconducting magnetic field equipment, significantly reduces the manufacturing cost of 455mm-level silicon single crystal, and significantly improves the value of industrial application.

[0027] (5) Under the premise that the thermal field remains unchanged, this application can be directly applied to the growth of large-diameter silicon single crystals in the range of 420-470mm, which improves the universality of single crystal growth. Attached Figure Description

[0028] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0029] Figure 1 This is a schematic flowchart of a silicon single crystal growth method provided in an embodiment of this application. Detailed Implementation

[0030] The present invention will be further described in detail below with reference to specific embodiments. The following embodiments are merely descriptive and not limiting, and should not be used to limit the scope of protection of the present invention.

[0031] When a quantity, concentration, or other value or parameter is described as a range, preferred range, or preferred upper and lower limits, it should be understood that it is equivalent to specifically disclosing any range by combining any pair of upper or preferred values ​​with any lower or preferred values, regardless of whether the range is specifically disclosed. Unless otherwise stated, the numerical range values ​​listed herein include the endpoints of the range and all integers and fractions within that range.

[0032] Unless otherwise stated, all percentages, parts, ratios, etc. in this document are by weight.

[0033] The materials, methods, and embodiments described herein are exemplary and should not be construed as limiting unless otherwise stated.

[0034] All the following examples were performed in a single crystal furnace with a 36-inch hot zone, a charge of 500 kg, and no magnetic field applied throughout the process, with a target growth diameter of 455 mm. Each example was repeated at least four times to ensure data reliability.

[0035] This embodiment uses, as follows: Figure 1 The silicon single crystal growth method shown does not apply a magnetic field during silicon growth. The silicon single crystal growth method includes: Welding stage: Control the welding temperature to the preset silicon melt wetting temperature range to achieve dislocation-free welding between the seed crystal and the silicon melt; During the crystal-leading stage: the liquid nozzle distance, crucible rotation speed, and crystal rotation speed are controlled to remain within the preset constant process parameter ranges; Shoulder formation stage: Maintain the preset liquid nozzle distance and crucible rotation speed, monitor the shoulder diameter in real time, and when the shoulder diameter reaches the preset starting value within the preset multiple range of the target crystal rod diameter, start the crystal rotation speed to decrease from the first rotation speed to the second rotation speed until the shoulder formation ends; Constant diameter growth stage: control the crystal rotation speed to maintain the second rotation speed until the end; wherein, the first rotation speed is a high rotation speed sufficient to suppress natural convection in the early stage of shoulder formation, and the second rotation speed is a low rotation speed not higher than 4 rpm.

[0036] Examples 1-8: Optimization of Welding Temperature Objective: To determine the optimal seed crystal fusion temperature for 455mm single crystal growth under magnetic field-free conditions.

[0037] Principle: Seed crystal fusion is a crucial step in dislocation elimination. The fusion temperature determines the wetting state between the seed crystal tip and the molten silicon. At excessively high temperatures (>1465℃), the seed crystal over-melts, reducing the surface tension of the melt and making it easier for polycrystalline nucleation to occur at the contact point. At excessively low temperatures (<1450℃), the seed crystal fails to fully fuse with the melt, leaving residual solid phase at the interface, preventing dislocation elimination. Only at suitable temperatures can an ideal solid-liquid interface be formed between the seed crystal and the melt, allowing dislocations to slide out of the crystal along the slip plane.

[0038] Experimental conditions: Fixed liquid outlet distance of 40 mm, crucible rotation speed of 10 rpm, crystal rotation speed of 8 rpm, and conventional parameters were used for crystal pulling, shoulder formation, and equal diameter setting. The welding temperatures were set at 1464℃, 1462℃, 1460℃, 1458℃, 1456℃, 1454℃, 1452℃, and 1450℃, with each temperature repeated 4 times. The crystal pulling success rate was then calculated.

[0039] Result: Table 1. Experimental results of welding temperature in Examples 1-8

[0040] Conclusion: When the welding temperature is 1454℃, the crystal pulling success rate reaches 100%. Crystal pulling fails at temperatures above 1458℃ or below 1452℃. Therefore, the optimal welding temperature is determined to be 1454±1℃.

[0041] Examples 9-16: Optimization of crucible rotation speed Objective: To determine the optimal crucible speed during the crystal pulling and shoulder formation stage.

[0042] Principle: The crucible rotation drives the entire melt to rotate via the Coriolis force, thus homogenizing the thermal field. Under non-magnetic field conditions, natural convection is highly random. Too low a crucible rotation (<8 rpm) fails to suppress asymmetric thermal plumes, leading to failure during the shoulder formation stage; too high a rotation (>15 rpm) enhances turbulence, causing fluctuations in the liquid level and temperature. A moderate crucible rotation (around 9-13 rpm) ensures an axisymmetric thermal field, providing a stable environment for crystal growth during the shoulder formation stage.

[0043] Experimental conditions: fixed welding temperature 1453℃, liquid outlet distance 40mm, crystal rotation 8rpm. The crucible rotation was set to 8, 9, 10, 11, 12, 13, 14, and 15rpm respectively, with each group repeated 4 times. The symmetry of the shoulder shape and the success rate were observed.

[0044] Experimental results: Table 2. Results of the crucible rotation experiment in Examples 9-16

[0045] Conclusion: Shoulder removal fails when crucible rotation speed is <10 rpm; significant fluctuations occur when crucible rotation speed is >13 rpm. The success rate is relatively high in the range of 11–13 rpm, with 12 rpm showing the best overall performance. The optimal crucible rotation speed for shoulder removal is determined to be 12 ± 1 rpm, but further optimization in conjunction with the nozzle distance is needed.

[0046] Examples 17-23: Optimization of liquid outlet distance Objective: To determine the optimal liquid outlet distance during the crystal pulling and shoulder formation stage.

[0047] Principle: The liquid gate distance determines the axial temperature gradient G at the crystal growth front. The smaller the liquid gate distance, the larger G, resulting in higher thermal stress and a greater likelihood of shoulder cracking; conversely, the larger the liquid gate distance, the smaller G, leading to insufficient growth driving force and difficulty in diameter control. For a large diameter of 455mm, the initial liquid gate distance must balance shoulder stress and constant diameter stability.

[0048] Experimental conditions: fixed welding temperature 1453℃, crucible rotation 12 rpm, crystal rotation 8 rpm. Liquid outlet distances were set to 30, 35, 40, 45, 50, 55, and 60 mm, with each group repeated 4 times. The success rate of shoulder formation was statistically analyzed.

[0049] Experimental results: Table 3. Experimental results of liquid outlet distance in Examples 17-23

[0050] Conclusion: The success rate of shoulder removal was highest (100%) when the nozzle distance was 50 mm. Although it was possible at 55 mm, the stability decreased, and at 60 mm, it failed completely. The optimal nozzle distance for shoulder removal was determined to be 50 ± 1 mm.

[0051] Examples 24-28: Crystal Transformation Optimization in the Equal Diameter Stage Objective: To determine the optimal crystal rotation speed during the constant diameter stage and eliminate doublet defects.

[0052] Principle: Crystal rotation generates forced convection, affecting boundary layer thickness and interface shape. In the initial stages of crystal growth and shoulder formation, a relatively high crystal rotation (8 rpm) is required to suppress natural convection and establish stable initial growth conditions. During the constant-diameter stage, the crystal rotation must be reduced to an appropriate level; otherwise, the centrifugal effect generated by excessively high crystal rotation will induce an unsteady flow field structure similar to Taylor vortex columns, leading to solid-liquid interface temperature oscillations and subsequently the formation of double-line micro-defects in the crystal. This effect is particularly severe under conditions without a magnetic field.

[0053] Experimental conditions: fixed welding temperature of 1454℃, liquid outlet distance of 50mm, crucible rotation of 12rpm, and crystal rotation of 8rpm during the crystal pulling and shoulder formation stage. The crystal rotation during the constant diameter stage was set to 3.0, 3.5, 4.0, 4.5, and 5.0rpm, with each group repeated 4 times. The double-line defect rate was calculated and the product quality was tested.

[0054] Experimental results: Table 4. Results of the isochronous crystal transformation experiments in Examples 24-28

[0055] Conclusion: While a crystal rotation speed of 3.0 rpm eliminates doublet defects, it negatively impacts product quality uniformity. A rotation speed of 3.5 rpm eliminates defects while maintaining product quality, whereas conventionally, a rotation speed of at least 5 rpm is considered to produce severe defects. Therefore, the optimal crystal rotation speed for constant diameter is determined to be 3.5 ± 0.2 rpm. In the field of large-diameter single crystal growth using the non-magnetic field Czochralski method, it is generally believed that a relatively high crystal rotation speed of at least 5 rpm is necessary to suppress melt convection and stabilize the solid-liquid interface. However, in experiments with a 455 mm crystal, this invention unexpectedly found that reducing the constant diameter crystal rotation speed to 3.5 ± 0.2 rpm—a value lower than the minimum crystal rotation speed typically used in the field to suppress melt convection (generally not less than 5 rpm)—allowed for reconfiguration of the melt flow field, effectively suppressing the aforementioned unsteady convection and resulting in defect-free crystals.

[0056] Examples 29-34: Optimization of the timing for initiating the gradual crystal transformation during the shoulder formation stage Objective: To determine the optimal starting diameter for reducing crystal rotation speed from 8 rpm to 3.5 rpm during the shoulder release process.

[0057] Experimental conditions: Optimal parameters were fixed (welding temperature 1454℃, liquid outlet distance 50mm, crucible rotation 12rpm). During the shoulder formation stage, crystal rotation decreased linearly at a uniform rate, with a total decrease of 4.5rpm. Starting diameters were set to 100, 150, 200, 250, 300, and 350mm, with each group repeated four times. The success rate of shoulder formation was statistically analyzed, and multiple replicates were performed for 200mm and 250mm diameters to verify the results.

[0058] Experimental results: Table 5. Experimental results of shoulder-forming crystal transformation in Examples 29-34

[0059] Conclusion: When the starting diameter is 200–250 mm, the success rate of shoulder formation reaches 100%. The physical mechanism is as follows: when the diameter is less than 200 mm, the crystal's geometric stability is poor, and the downward adjustment of crystal rotation leads to uncontrolled natural convection; when the diameter is greater than 350 mm, the remaining stroke is too short, and the excessively rapid rate of crystal rotation change causes transient flow field mismatch. 200–250 mm is precisely the equilibrium range between these two types of failures. Therefore, the optimal starting diameter is determined to be 225 ± 25 mm.

[0060] At the start of shoulder formation, the crystal rotation speed is 8 rpm, and at the same diameter, it is 3.5 rpm. Gradual changes in crystal rotation speed are necessary during the shoulder formation process. A key technical contradiction exists regarding the timing of initiating the crystal rotation change: if the diameter is too small (<200mm), initiating the change occurs while the shoulder formation is still unstable, and the crystal rotation change easily triggers melt disturbances, leading to rejection (NG); if the diameter is too large (>350mm), initiating the change results in an excessively rapid crystal rotation rate, causing violent melt convection and temperature reactions, which is also detrimental to shoulder formation. Through systematic experiments, this invention has discovered a critical initiation window.

[0061] The determination of the critical start window is based on the following physical mechanism: When the diameter of the shoulder is less than 0.44 times the target diameter, the crystal rotation is initiated. The shoulder is in a geometrically unstable stage. Adjusting the crystal rotation leads to forced convection attenuation and natural convection loss control. When the shoulder diameter is greater than 0.55 times the target diameter, the crystal rotation is initiated. The remaining shoulder stroke is insufficient, resulting in an excessively fast crystal rotation rate. The flow field response lag causes temperature oscillation. Only when the crystal transition is initiated within the range of 0.44 to 0.55 times the target diameter, the crystal has sufficient geometric inertia and the remaining stroke is sufficient to ensure a smooth transition, achieving a synergy between shoulder stability and defect-free constant diameter.

[0062] This application, through systematic experiments, discovered that a shoulder diameter of 200-250 mm constitutes the critical initiation window for crystal transition gradient; initiation too early or too late leads to shoulder formation failure. Further analysis reveals the following physical mechanism: Premature initiation: When the shoulder diameter is less than 200mm, the crystal is in a highly unstable stage of rapid diameter expansion, which is a typical geometric stability-controlled failure. During this stage, the crystal rapidly expands from a 10mm neck to a target diameter of 455mm. The crystal cross-sectional area, structural inertia, and disturbance resistance stiffness are all at extremely low levels, resulting in very poor tolerance to external melt disturbances and thermal field fluctuations. The core function of a high crystal rotation speed of 8rpm is to form controllable forced convection through the viscous drag of the crystal wall, suppressing the disordered development of natural thermal convection on the crucible side and dispersing asymmetric thermal plumes. This is the most crucial stabilizing barrier in the early stages of shoulder formation. If the crystal rotation speed is prematurely reduced in the stage of diameter <200mm, the forced convection intensity decays rapidly, natural convection loses its constraint and dominates the flow field, and irregular thermal plumes continuously impact the solid-liquid interface. Simultaneously, the small-diameter shoulder interface has a small radius of curvature and extremely low interfacial heat capacity, resulting in high thermal disturbance transfer efficiency. This easily leads to solid-liquid interface morphology distortion and uncontrolled diameter fluctuations, and in severe cases, directly causes shoulder fracture, dislocation multiplication, and other defects.

[0063] Late Start-up: When the shoulder diameter is greater than 350mm, the crystal geometry is nearly complete, and geometric stability is no longer a limiting factor. The failure mode shifts to a transient mismatch between the melt flow field and temperature field, resulting in a failure controlled by this mismatch. In this process, the total crystal rotation rate reduction is fixed at 4.5 rpm (8 rpm → 3.5 rpm). The larger the start-up diameter, the shorter the remaining shoulder growth path and the shorter the gradient time, significantly increasing the average crystal rotation rate. Large-size single-crystal growth systems utilize 500kg-level melts. Silicon melts possess large momentum and thermal inertia, leading to significant hysteresis in the response and relaxation of the flow and temperature fields. When the crystal rotation rate decreases too quickly, the melt boundary layer shear conditions change abruptly, and the flow velocity gradient and solid-liquid interface temperature gradient cannot be synchronized, resulting in continuous temperature oscillations and flow field disturbances. The interface structure at the tail of the shoulder is nearly complete, leaving very little room for error in the face of transient shocks. Once interface fluctuations exceed the critical stability range, they directly disrupt the previously accumulated steady-state growth, causing batch shoulder failures.

[0064] The physical essence of the 200-250mm critical window: The shoulder diameter of 200-250mm precisely constitutes the equilibrium range between the two types of failure modes mentioned above. At this point, on the one hand, the shoulder has passed the most dangerous initial geometric instability stage, and the crystal has sufficient geometric inertia to withstand moderate flow field adjustments; on the other hand, the remaining shoulder stroke is still long enough, and the crystal rotation can decrease at a sufficiently gradual rate, allowing the melt flow field ample time to respond to the gradual changes in boundary conditions, thereby avoiding transient mismatch. Therefore, 200-250mm is not an accidental empirical value, but rather an inevitable balance point between the contradiction between system geometric stability and flow field response speed.

[0065] Example 35: Validation of the Overall Optimal Process Objective: To verify the overall effect of the above segmented optimization combination.

[0066] The process parameters are: Welding temperature: 1454℃ Crystal pulling: liquid outlet distance 50mm, crucible rotation 12rpm, crystal rotation 8rpm Shoulder removal: Liquid outlet distance 50mm, crucible rotation 12rpm, crystal rotation speed uniformly reduced from 8rpm to 3.5rpm. Start when the shoulder diameter reaches 225mm (within the range of 200-250mm). Constant diameter: liquid outlet distance 50mm, crucible rotation 12rpm, crystal rotation 3.5rpm. Verification results: Six consecutive batches of growth were successfully repeated, all of which produced complete single-crystal blanks with an outer diameter of 455±2mm. The blank weight accounted for more than 85% of the feed weight, and there were no double-line defects in the constant diameter section. After the crystal rods were rounded to a standard finished product of 448mm, the finished product weight accounted for approximately 67% of the feed weight. No shoulder fractures or dislocation multiplications occurred in any batch.

[0067] In the field of large-diameter single-crystal silicon growth using the Czochralski method without a magnetic field, a directional technical understanding has long existed: to effectively suppress natural thermal convection of the melt and maintain the axisymmetric stability of the solid-liquid interface, the crystal rotation speed during the constant diameter stage should be maintained at a high level, generally considered not to be lower than 5 rpm. For example, in process simulation studies of Ф300mm Czochralski single-crystal silicon without a magnetic field, 9-10 rpm is recommended as the optimal crystal rotation speed; in process studies of larger diameter single-crystal silicon, there are also reports of 12 rpm as the preferred crystal rotation speed for obtaining a stable solid-liquid interface; even under magnetic field-assisted conditions, related studies still regard 12 rpm as the preferred parameter for large-diameter single-crystal silicon. Although the above studies differ in specific rotation speed values, their underlying logic is all based on the unidirectional positive correlation of "higher crystal rotation → stronger forced convection → more stable solid-liquid interface," forming a technical inertia of "defaulting to high crystal rotation." Under this conventional thinking, setting the crystal rotation rate to below 5 rpm for constant diameter crystals—especially 3.5 rpm, which is considered "too low" in conventional understanding—is seen as an unconventional operation that violates common sense in the process, and therefore has rarely been studied. However, the aforementioned studies have failed to realize that under the condition of larger diameters (such as 455 mm) without a magnetic field, this conventional thinking of "defaulting to high crystal rotation" may very well become the root cause of defects—the centrifugal effect caused by excessively high crystal rotation can give rise to new unsteady flow field structures, which in turn lead to double-line defects in the constant diameter segment. This systematic neglect of the "low crystal rotation" direction, and the widespread blind following of the one-way logic of "high crystal rotation → more stability," constitutes the technical bias in the field of large-diameter single crystal growth under the condition of no magnetic field.

[0068] The art has long held the conventional wisdom that "higher crystal rotation leads to more stable convection," typically maintaining a crystal rotation of at least 5 rpm during the constant diameter stage. However, this invention, through experiments, reveals that under 455 mm non-magnetic field conditions, this perception is exactly the opposite—higher crystal rotation (≥4 rpm) during the constant diameter stage actually leads to a sharp increase in the doublet defect rate (from 0% to 100%); only by reducing the crystal rotation to 3.5 rpm, a level considered "too low" in the art, are the defects completely eliminated. This contradictory conclusion, derived from measured data, directly and effectively refutes technological biases and breaks with conventional thinking.

[0069] Shoulder formation requires high crystal rotation (8 rpm) to ensure geometric shaping, while constant diameter requires low crystal rotation (3.5 rpm) to eliminate defects; these two requirements are mutually exclusive. This invention demonstrates through systematic experiments that this contradiction is not irreconcilable—initiating a gradual crystal rotation change within the shoulder diameter range of 200–250 mm allows for a smooth transition to low crystal rotation without sacrificing the shoulder formation success rate. Initiating the change too early (<200 mm) or too late (>350 mm) significantly reduces the success rate, while 200–250 mm represents a perfect balance between geometric stability recovery and ample flow field relaxation time, resolving the stage-specific contradiction.

[0070] This application does not rely on the adjustment of a single parameter, but rather on the synergistic breakthrough achieved when multiple parameters, such as the welding temperature of 1454±1℃, the liquid outlet distance of 50mm, the crucible rotation speed of 12rpm, and the crystal transition gradient start-up window of 200-250mm, simultaneously fall within their respective tolerance ranges. If any parameter deviates from this combination, the growth will fail or produce unacceptable defects.

[0071] This invention achieves a leap from "unstable growth" to "silicon yield > 85%" for 455mm single crystals without magnetic field by following a progressive solution path: "overturning the prejudice that high crystal density leads to greater stability → discovering the critical window for gradual transition at 200-250mm → defining the synergistic boundary for simultaneous achievement of multiple parameters". This effect cannot be achieved by conventional optimization of a single parameter, but is a comprehensive product of the synergistic effect of multiple factors. It breaks through the technical contradiction that shoulder stability and uniform diameter without defects are mutually exclusive, and realizes high-quality single crystal growth under magnetic field-free conditions.

[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for growing silicon single crystals, characterized in that, No magnetic field is applied during silicon growth, and the silicon single crystal growth method includes: Welding stage: Control the welding temperature to the preset silicon melt wetting temperature range to achieve dislocation-free welding between the seed crystal and the silicon melt; During the crystal-leading stage: the liquid nozzle distance, crucible rotation speed, and crystal rotation speed are controlled to remain within the preset constant process parameter ranges; Shoulder formation stage: Maintain the preset liquid nozzle distance and crucible rotation speed, monitor the shoulder diameter in real time, and when the shoulder diameter reaches the preset starting value within the preset multiple range of the target crystal rod diameter, start the crystal rotation speed to decrease from the first rotation speed to the second rotation speed until the shoulder formation ends; Constant diameter growth stage: control the crystal rotation speed to maintain the second rotation speed until the end; wherein, the first rotation speed is a high rotation speed sufficient to suppress natural convection in the early stage of shoulder formation, and the second rotation speed is a low rotation speed not higher than 4 rpm.

2. The silicon single crystal growth method according to claim 1, characterized in that, The second rotational speed is 3.5 ± 0.2 rpm.

3. The silicon single crystal growth method according to claim 1, characterized in that, The preset multiple range is 0.4 to 0.6 times.

4. The silicon single crystal growth method according to claim 3, characterized in that, The target crystal rod has a diameter of 455±15mm; the preset starting value for the shoulder formation stage is 200mm~250mm.

5. The silicon single crystal growth method according to claim 1, characterized in that, The preset silicon melt wetting temperature range is 1454±1℃.

6. The silicon single crystal growth method according to claim 1, characterized in that, The preset constant liquid outlet distance process parameter range is 50±2mm; the preset constant crucible rotation speed process parameter range is 11-13rpm; and the preset constant crystal rotation speed process parameter range is 8-9rpm.

7. The silicon single crystal growth method according to claim 6, characterized in that, The preset constant liquid outlet distance process parameter range is 50mm; the preset constant crucible rotation speed process parameter range is 12rpm.

8. The silicon single crystal growth method according to claim 6, characterized in that, The first rotational speed is 8 rpm.

9. The silicon single crystal growth method according to claim 1, characterized in that, The decrease is a linear and uniform decrease, and the deceleration rate is calculated based on the preset starting value and the end diameter of the shoulder formation, so that the crystal rotation speed reaches the second rotation speed exactly when the shoulder formation ends.

10. The silicon single crystal growth method according to claim 1, characterized in that, The silicon single crystal growth method is carried out in a magnetic field-free single crystal furnace with a 36-inch hot zone and a charge of 500 kg, and the weight of the blank accounts for more than 85% of the charge.

Images