Super-large-scale silicon single crystal growth method based on high-temperature superconducting magnetic control technology

Through high-temperature superconducting magnetron technology and adaptive control methods, the problems of unstable melt flow and uneven impurity distribution in ultra-large-scale silicon single crystal growth have been solved, high-temperature field uniformity and low-energy consumption silicon single crystal growth have been achieved, and crystal quality and production efficiency have been improved.

CN120738760APending Publication Date: 2025-10-03WUXI UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202510677346.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-26
Publication Date
2025-10-03

AI Technical Summary

Technical Problem

Existing silicon single crystal growth technology faces problems such as unstable melt flow, concentrated thermal stress and uneven impurity distribution when growing ultra-large crystals. Traditional magnetic control methods are difficult to achieve high-intensity, uniform and adjustable magnetic field control. The system has high energy consumption and poor stability, making it difficult to meet the growth needs of high-quality large-size silicon single crystals.

Method used

Using high-temperature superconducting magnetron technology, through the method of adaptive control capability, combined with gradient graphite felt, multi-layer thermal conductive materials and multi-objective optimization algorithm, real-time monitoring of temperature and oxygen concentration, dynamic adjustment of magnetic field and temperature field parameters, realization of multi-physics field coupling simulation, optimization of growth parameters, modular non-insulated winding and conduction cooling structure, integrated feedback control system.

Benefits of technology

Significantly reduce oxygen diffusion, improve temperature field uniformity, reduce energy consumption, enhance mechanical stress resistance, improve crystal structure integrity and defect rate, and achieve efficient and stable growth of large-size silicon single crystals.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120738760A_ABST
    Figure CN120738760A_ABST
Patent Text Reader

Abstract

The invention discloses a super-large-scale silicon single crystal growth method based on a high-temperature superconducting magnetic control technology, and relates to the technical field of silicon single crystal growth, and the method comprises the following steps: setting and calibrating an initial magnetic field intensity, and combining a gradient graphite felt heat conductivity coefficient function to optimize temperature field uniformity; a dynamic magnetic field regulation and control mechanism is introduced in the melt stage, melt convection is inhibited through flow velocity-magnetic field joint control, and oxygen concentration monitoring is synchronized; a multi-physics coupling simulation and multi-objective optimization algorithm is adopted in the crystal growth stage, the magnetic field intensity, the heating power and the crystal pulling parameters are adjusted in real time, and the oxygen content and the growth rate are balanced; and the tail stress is relieved through a gradient field reduction technology, and finally, the low-defect and high-uniformity silicon single crystal is obtained. Melt flow and temperature field uniformity are accurately regulated and controlled through a high-temperature superconducting magnetic control technology, dynamic monitoring and a multi-target optimization algorithm are combined, the growth quality of silicon single crystals is remarkably improved, the defect rate and energy consumption are reduced, and meanwhile efficient and stable production of large-size crystals is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of ultra-large-scale silicon single crystal growth, in particular to an ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology. Background Art

[0002] Silicon single crystal materials serve as a critical foundation for fields such as integrated circuits, photovoltaic power generation, and power devices. Their crystal quality and size directly determine the overall level of downstream chip manufacturing and energy conversion. Currently, the two most widely used silicon single crystal growth methods in industry include the melting zone method (FZ method) and the Czochralski method (CZ method). The CZ method is more widely used in actual production due to its higher growth efficiency and crystal size scalability. However, as silicon wafer diameters continue to grow toward 450mm and above, the CZ method faces numerous challenges in growing ultra-large crystals. In particular, during the crystal pulling process, due to significantly enhanced melt thermal convection and radial temperature differences, the flow field is prone to instability, resulting in oxygen content fluctuations, uneven impurity distribution, and abnormal thermal stress, which can induce dislocations, micro-defects, and even crystal warpage. To address this, studies have attempted to introduce static or dynamic magnetic field control technologies (such as the MCZ method and A-MCZ method) to improve melt flow and crystal quality. However, traditional magnetron technology is limited by the physical properties of conventional electromagnet materials. Not only is the magnetic field strength limited in high-temperature environments, resulting in low control precision, but it also requires high energy consumption to maintain a stable magnetic field, resulting in complex cooling systems and poor system stability. Furthermore, the uneven spatial distribution of the magnetic field makes it difficult to achieve the precise and coordinated control of the thermal, flow, and solute fields during silicon single crystal growth, limiting further improvements in overall crystal quality.

[0003] In contrast, high-temperature superconducting materials offer advantages such as zero resistance, high critical current density, and complete anti-magnetism. They can provide a stable, high-intensity, and spatially uniform magnetic field under low energy consumption conditions, making them one of the core technologies for the next generation of magnetron-controlled silicon single crystal growth. By rationally designing the high-temperature superconducting coil structure and incorporating a feedback control mechanism for the magnetron parameter, precise manipulation of the melt flow state and crystal growth dynamics can be achieved. However, a systematic approach is currently lacking that can deeply integrate the magnetron control capabilities of high-temperature superconductors with the requirements for ultra-large-scale silicon single crystal growth. Significant technical gaps remain, particularly in modeling the response mechanism between magnetic field control and physical parameters of the growth process, optimizing the magnetic field distribution, and coupling control pathways for the three fields (flow field, thermal field, and solute field). Therefore, a method for ultra-large-scale silicon single crystal growth that integrates high-temperature superconducting magnetron technology and possesses adaptive control capabilities is urgently needed to effectively improve crystal structural integrity, reduce defect rates, and lower energy consumption, thereby driving the development of high-end semiconductor material preparation processes towards higher quality and larger sizes. Summary of the Invention

[0004] Existing silicon single crystal growth technologies face numerous limitations when growing ultra-large crystals, such as unstable melt flow, concentrated thermal stresses, and uneven impurity distribution, which can easily lead to crystal defects and deformation. Traditional magnetron methods, limited by the performance of conventional electromagnets, struggle to achieve high-intensity, uniform, and adjustable magnetic field control at high temperatures. Furthermore, the system suffers from high energy consumption and poor stability, making it difficult to meet the growth requirements of the next generation of high-quality, large-scale silicon single crystals. This present invention addresses these issues.

[0005] Therefore, the problem to be solved by the present invention is how to provide a method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology, including: self-checking the sensor to ensure accuracy, evacuating and cooling the magnet to the superconducting operating temperature, setting and calibrating the initial magnetic field strength, and establishing a stable environment for the growth process; defining the thermal conductivity function of each layer of gradient graphite felt according to simulation, and optimizing and calculating the temperature field uniformity index using multi-layer thermal conductive materials; real-time monitoring of temperature distribution and dynamic adjustment of heating power; introducing argon gas after the melt reaches the set temperature, adjusting the magnetic field by flow rate monitoring to suppress melt flow, and synchronously recording the initial value of oxygen concentration; contacting the seed crystal with the melt to form a wetted interface, set the initial crystal pulling parameters and gradually enhance the magnetic field to suppress convection to obtain a sufficient Hartmann number; monitor the flow rate and oxygen concentration in real time, dynamically adjust the magnetic field and temperature field parameters to ensure the initial stability of crystal growth; update the growth model through multi-physics field coupling simulation, use the multi-objective optimization algorithm to balance the oxygen content and growth rate, and screen the optimal parameter combination; combine feedback control to adjust the heating and cooling power in real time to maintain the dynamic balance of the growth process; gradually reduce the pulling speed and magnetic field strength to reduce tail stress, maintain the temperature field to promote slow solidification; cool to room temperature after demagnetization, extract the silicon rod and detect the oxygen content, defects and size to ensure that the finished product quality meets the standards.

[0008] As a preferred solution of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology described in the present invention, the thermal conductivity of the gradient graphite felt adopts a layered design, wherein the upper and lower layers have constant thermal conductivity coefficients, and the middle layer adopts a positive temperature coefficient material, which automatically adjusts the thermal conductivity efficiency with temperature changes to maintain temperature field uniformity.

[0009] As a preferred solution of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology described in the present invention, defining the thermal conductivity function of each layer of the gradient graphite felt includes: dividing the gradient graphite felt into multiple functional gradient layers along the thickness direction, and determining the thermal conductivity value range of each layer according to material design requirements; selecting the functional form of the thermal conductivity varying with thickness according to the thermal conductivity performance requirements to describe the gradient change trend of the overall thermal conductivity performance; considering the continuity and physical feasibility of the transition of the thermal conductivity between layers in the function setting process to avoid mutations while meeting the design goals of the overall thermal conductivity performance.

[0010] As a preferred solution of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology described in the present invention, the calculation of the temperature field uniformity index includes: obtaining the thermocouple temperature values ​​of multiple pixel points in the thermal image, and normalizing all temperature values ​​so that they fall into a unified interval range for subsequent analysis; based on the normalized temperature values, calculating the difference between all pixel points to measure the fluctuation of temperature in spatial distribution; based on the statistical results of the temperature distribution difference, outputting a comprehensive uniformity index, and the closer the value is to the ideal value, the more uniform the temperature field distribution is.

[0011] As a preferred solution of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology described in the present invention, obtaining the Hartmann number includes: forming a stable magnetic field environment by setting a high-temperature superconductor, and obtaining the spatial distribution of magnetic induction intensity in the silicon single crystal growth melt zone by magnetic field simulation or actual measurement; constructing a flow-electromagnetic coupling model of the silicon melt under the action of a magnetic field based on known physical parameters such as conductivity and viscosity of the silicon melt to describe the inhibitory effect of the magnetic field on the flow of the fluid; and calculating a dimensionless numerical index representing the stability of the fluid under the action of the magnetic field based on the established model and the obtained parameters to evaluate the influence of the superconducting magnetic field on the stability of the crystal growth interface and the convection suppression ability.

[0012] As a preferred solution of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology described in the present invention, the screening of the optimal parameter combination includes: comprehensively considering multiple performance indicators such as growth efficiency, crystal quality and energy consumption to form an evaluation standard for the parameter combination; collecting corresponding multiple performance indicator data, and comprehensively scoring different indicators through normalization or weighted methods; selecting the parameter set that achieves the optimal balance among various indicators and meets the process requirements as the final growth solution.

[0013] As a preferred solution of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology described in the present invention, the real-time adjustment of heating and cooling power in combination with feedback control includes: real-time collection of key parameters such as the temperature, crystal morphology and magnetron field strength of the silicon single crystal growth area; analysis of the collected data by the control system to determine the deviation and adjustment requirements of the current heating or cooling power; and dynamic adjustment of the heating power and cooling power by the controller according to the analysis results to achieve precise control of temperature and crystal quality.

[0014] Secondly, in order to further solve the problems existing in the growth of silicon single crystals, the present invention provides an ultra-large-scale silicon single crystal growth system based on high-temperature superconducting magnetron technology, which includes: a high-temperature superconducting magnetron module for generating a magnetic field with adjustable intensity and uniform spatial distribution, suppressing melt convection and reducing oxygen diffusion; a melt heating and temperature control module for heating the melt through an induction heater and a gradient insulation layer, and controlling the temperature field uniformity in combination with a temperature sensor and an intelligent algorithm; a melt flow monitoring and control module for real-time monitoring of the melt surface flow velocity and dynamically adjusting the magnetic field intensity or argon flow rate according to the flow velocity threshold. ; Oxygen content monitoring and control module, used to measure the oxygen concentration and gas phase partial pressure of the melt online, and control the oxygen content at the crystal top by coordinated control of the magnetic field, argon flow rate and crucible speed; multi-objective optimization control module, used to use multi-objective optimization algorithm to coordinately optimize the growth parameters and balance the growth rate and crystal quality; tail solidification control unit, used to gradually reduce the pulling speed and magnetic field intensity at the end of crystal growth, and reduce tail stress and defects through the insulation stage; central control platform, used to integrate data acquisition, multi-physics field simulation and real-time feedback control, coordinate the operation of each module to achieve stable growth of silicon single crystals.

[0015] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology as described in the first aspect of the present invention is implemented.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology as described in the first aspect of the present invention is implemented.

[0017] The beneficial effects of the present invention are:

[0018] 1. The present invention ensures that the Hartmann number is much larger than the critical value by designing a high-temperature superconducting magnet, which can significantly attenuate the melt convection velocity and reduce oxygen diffusion. In addition, the oxygen content is monitored in real time. When the oxygen concentration at the crystal top approaches the set threshold, the magnetic field can be immediately increased or the atmosphere parameters can be adjusted to achieve dynamic oxygen control.

[0019] 2. The gradient graphite felt of the present invention has a positive temperature coefficient thermal conductive material in the middle layer, which can automatically adjust the thermal conductivity efficiency with temperature changes. Combined with the DNN model, it can achieve a temperature field uniformity index of less than or equal to 3%. Moreover, while ensuring a uniform temperature field, it can reduce the energy consumption per silicon rod by more than or equal to 30% compared with the traditional single-layer insulation method.

[0020] 3. The present invention uses the NSGA-II multi-objective optimization algorithm to obtain the optimal parameter combination, achieving a balance between growth rate and crystal quality, and avoiding the performance degradation caused by single-objective optimization. The algorithm also uses population evolution and non-dominated sorting to quickly screen a set of Pareto optimal solutions, providing a variety of parameter selection options, and enhancing the flexibility and robustness of the optimization process.

[0021] 4. The high-temperature superconducting magnet of the present invention adopts a modular non-insulated winding and conduction cooling structure, which enhances the mechanical stress resistance and quench protection reliability; and the control system and simulation platform are highly integrated, which facilitates the subsequent function expansion and algorithm iteration;

[0022] 5. The present invention combines the domestic high-temperature superconducting materials and equipment industry chain to fill the gap in large-diameter (≥450mm) silicon single crystal growth technology, which is conducive to large-scale application in the semiconductor and photovoltaic fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0024] Figure 1 This is a flow chart for implementing the present invention in Example 1. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0027] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0028] Example 1

[0029] Reference Figure 1 , which is the first embodiment of the present invention, provides a method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology, comprising the following steps:

[0030] Step S1: Initialization phase, which includes the following sub-steps:

[0031] S1-1: Sensor and actuator self-test, specifically calibrating the temperature, magnetic field, flow field, and oxygen content sensors one by one to ensure that the reading deviation is ≤2%; check the operating status of equipment such as the induction heater, refrigerator, and vacuum pump to confirm that there are no alarms.

[0032] S1-2: Preparation for furnace chamber vacuum and cooling, specifically starting the mechanical pump and molecular pump to pump the furnace chamber to the set vacuum degree P 炉腔 ≤100Pa; start the low-temperature refrigerator to cool the main magnet to the operating temperature range (about 20K–30K).

[0033] S1-3: Excitation of the high-temperature superconducting magnet. Specifically, the initial excitation current I0 is set to maintain the magnetic field strength B0 at the center of the magnet at the designed value (e.g. 8T-12T); the actual field strength B0 at the center hole is measured using a Hall sensor. ms , guarantee |B ms -B0|≤0.01T; confirm that the quench protection system (fiber Bragg grating and acoustic wave detection) is online and sensitive enough to detect temperature rise or local damage.

[0034] Step S2: preheating and melting, which includes the following sub-steps:

[0035] S2-1: Preheating of the induction heater, setting the initial heating power according to the multi-physics simulation recommendations And define the thermal conductivity function k of each layer of gradient graphite felt eff (z,T);

[0036] Specifically, the above parameter k eff This can be expressed by the following formula:

[0037]

[0038] Where z represents the coordinate from the top to the bottom of the insulation layer along the thickness direction; k1 is the thermal conductivity of the upper layer; k2,0 is the reference thermal conductivity of the middle layer; α is the positive temperature coefficient of the middle layer; T0 is the reference temperature of the middle layer; k3 is the thermal conductivity of the lower layer; z1, z2, z3 are the critical coordinates of the layer thickness, which can be measured according to the actual material thickness;

[0039] It should be noted that the above representation describes the thermal conductivity of the insulation system in three layers.

[0040] S2-2: Monitor and adjust the temperature field, specifically read the temperature of multiple thermocouples in real time i , calculate the temperature field uniformity index η T ;

[0041] Specifically, the above parameter η T This can be expressed by the following formula:

[0042]

[0043] Where, T 设定 Indicates the target temperature value that the system expects to maintain during the heating process of the furnace chamber or melt;

[0044] For example, if η T >η T,阈值 (η T,阈值 (set manually), calculate the temperature distribution error ΔT = T i -T 设定 , the output heating power correction value ΔP is controlled by the deep neural network (DNN) model 热 , update heating power:

[0045] P 热 ←P 热 +ΔP 热 ;

[0046] It should be noted that the DNN model is used to control the output heating power correction value by taking the temperature deviation of each thermocouple, the current heating power, the magnetic field strength, the cooling power and other real-time data as input, completing complex nonlinear mapping in the network, and quickly predicting the most appropriate heating power correction value; it is an existing common technology and will not be described in detail.

[0047] S2-3: Melt heating and argon introduction, specifically when the center temperature of the molten pool T 熔 ≥T 熔化 (For example, at 1450℃), stabilize the heating power and open the argon flow Φ Ar and the furnace chamber pressure P 炉腔 Control within the target range; start the LDV system to measure the melt surface velocity v r If v r If the set threshold is exceeded, the magnetic field strength B needs to be increased to enhance convection suppression.

[0048] S2-4: Oxygen content baseline monitoring, specifically measuring the gas phase SiO partial pressure C using an online mass spectrometer sio ; Synchrotron radiation detection device obtains the oxygen concentration distribution C on the surface or inside of the melt 氧 (x,y), record the baseline value for subsequent comparison.

[0049] Step S3: Introducing the seed crystal and starting crystal pulling, including the following sub-steps:

[0050] S3-1: Seed crystal preparation and wetting, specifically to select single crystal silicon seed crystals that meet the specifications, and the orientation can be <100 or <111> ; Place the bottom of the seed crystal on the surface of the melt and keep it in contact for a certain time (such as 10 minutes) to form a smooth wetting interface between the seed crystal and the melt.

[0051] S3-2: Set the initial pull-down and rotation parameters, specifically the initial pull-down speed and crucible rotation speed w 0 At the start of the crystal pulling phase, the magnetic field B is gradually increased to a target value (eg, 12 T) within a predetermined time to obtain a sufficient Hartmann number Ha;

[0052] Specifically, the above parameter Ha can be expressed by the following formula:

[0053]

[0054] Where Ha is a dimensionless parameter that measures the ability of a magnetic field to suppress fluid convection; L is the characteristic length of the melt pool (e.g., radius); σ is the melt conductivity, which can be measured experimentally; ρ is the melt density, which can be measured experimentally; v is the melt kinematic viscosity, v = μ / ρ, and μ is the melt dynamic viscosity;

[0055] It should be noted that the above idea is that when Ha reaches a certain value (such as ≥100), the convection velocity can be significantly attenuated, thereby achieving convection suppression and slowing down oxygen diffusion.

[0056] S3-3: Real-time convection suppression monitoring, specifically using LDV to monitor the melt surface velocity v r If v r If the predetermined threshold is exceeded, the magnetic field strength is dynamically increased:

[0057] B←B+ΔB;

[0058] Where ΔB is the predetermined increment;

[0059] At the same time, monitor the temperature field uniformity index η T If η T If the value exceeds the limit, the heating power or the thermal conductivity of the gradient graphite felt is adjusted. For example, this can be achieved by replacing the middle layer material combination or adjusting the heater power.

[0060] S3-4: Oxygen content monitoring and preliminary adjustment, specifically online measurement of gas phase SiO concentration C sio And the oxygen concentration C on the melt surface 氧 ;

[0061] For example, if the oxygen concentration C 氧 If the limit is exceeded, it is necessary to increase the magnetic field strength B and appropriately increase the argon flow rate Φ Ar , adjust the crucible speed w or introduce a strategy to reduce the crucible temperature to control S 坩埚 Source item size.

[0062] Step S4: mid-growth optimization, including the following sub-steps:

[0063] S4-1: Update the magnetic-thermal-fluid-oxygen coupled simulation. Specifically, the following coupled equations are solved regularly (e.g., every 10–20 minutes) using finite element software:

[0064] Navier–Stokes equations with Lorentz force terms:

[0065]

[0066] Where v represents the melt velocity vector, which can be measured by LDV and reconstructed by interpolation; P is the fluid pressure distribution, which can be obtained by numerical solution; Φ is the melt dynamic viscosity, which can be measured experimentally; represents the local current density; B is the external magnetic field vector, which is given by the Hall sensor data; j×B represents the Lorentz force density term, which is used to suppress convection; ρg represents the gravity term, and g is the gravitational acceleration vector;

[0067] It should be noted that the above formula describes the movement law of the conductive fluid under the action of the magnetic field, taking into account the combined effects of pressure difference, viscosity, magnetism and gravity.

[0068] Energy equation:

[0069]

[0070] Where T is the temperature field, which can be measured by thermocouples and interpolated; c p is the specific heat capacity of the melt, which can be measured by differential scanning calorimetry; Q 感应 is the induction heating power density;

[0071] It should be noted that the above formula describes the evolution of the temperature field of the melt and its surrounding materials under the effects of heating, convection, and conduction;

[0072] Exemplary:

[0073]

[0074] Where, I cl is the induction coil current; R 感应 is the equivalent resistance; V 熔池 is the melt volume.

[0075] Oxygen diffusion and convection equation:

[0076]

[0077] Where D is the diffusion coefficient of oxygen in the melt, which can be measured by thermal diffusion experiments; C is the concentration of oxygen atoms in the melt, which can be obtained by synchrotron radiation detection equipment; It is used to describe the rate at which the quartz crucible dissolves oxygen atoms from the melt. The parameter A is the frequency factor, and E is the a is the activation energy, which can be fitted by the Arrhenius curve, R is the gas constant, T 坩埚 is the crucible temperature, measured by a thermocouple;

[0078] S4-2: Multi-objective optimization, specifically based on the crystal top oxygen content C 晶顶 Minimize and grow rate R gh Maximization is the optimization goal, and the design variables include magnetic field intensity B, pull-down speed v pull , crucible speed w, melt temperature T 熔 , Argon gas flow Φ Ar , furnace chamber pressure P 炉腔 ; Use NSGA-II algorithm to obtain Pareto frontier solution set, and select the solution that satisfies "C 晶顶 ≤3ppma and R gh The optimal parameter combination of ≥0.7mm / min” is used as the benchmark for real-time adjustment;

[0079] It should be noted that the NSGA-II algorithm uses an elite retention strategy and a fast non-dominated sorting mechanism to iteratively evolve a set of Pareto frontier solutions that are not inferior to each other in the target space. This is an existing mature technology and will not be described in detail.

[0080] S4-3: Online feedback control, specifically using the DNN model to predict the temperature field distribution error ΔT every Δt (e.g., 10 minutes), and generating the heating power correction ΔP through the PID controller 热 and cold head power correction ΔP 冷 :

[0081]

[0082] Where K p , K i , K d With K` p , K` i , K`d are PID parameters, obtained through preliminary experiments or closed-loop adjustment.

[0083] Furthermore, when the melt surface velocity v is monitored r When the threshold is exceeded or the crystal top oxygen content fluctuates too much, the following corrections are performed:

[0084] Magnetic field strength: B←B+ΔB, where ΔB is the preset increment;

[0085] Argon flow rate: Φ Ar ←Φ Ar +ΔΦ Ar , where ΔΦ Ar is the preset increment;

[0086] At the same time, fine-tune the pull-down speed v pull Or crucible speed w to balance crystal quality and production efficiency.

[0087] Step S5: solidification and finishing of the tail, including the following sub-steps:

[0088] S5-1: Tail pulling: When the single crystal is pulled down close to the designed length (for example, the entire silicon rod length reaches the corresponding height), the pulling speed is gradually reduced to reduce the stress accumulation at the tail of the crystal: v pull ←v pull -Δv, where Δv is the preset decrement. The rotational speed is simultaneously adjusted to a lower value: w←w-Δw, where Δw is the preset decrement. The magnetic field strength is slowly reduced to the preset tail solidification value (e.g., 8T) to maintain slow solidification of the tail crystal and avoid stress cracks.

[0089] S5-2: Tail solidification and heat preservation. Specifically, after the pull-down is completed, the crystal is kept still and the magnetic field and temperature field are maintained for a period of time (e.g., 10–20 minutes) to ensure that the last part of the crystal solidifies slowly and reduce dislocation deposition.

[0090] S5-3: Demagnetization and cooling. Specifically, before turning off the excitation power supply, ensure that the optical fiber and the acoustic quench protection system are stationary. Slowly reduce the excitation current by ΔI(A) each time until the magnetic field drops to 0T. Turn off the induction heater, keep the cold head running, and wait for the furnace chamber temperature to gradually drop to room temperature.

[0091] S5-4: Silicon rod extraction and testing, including:

[0092] After the furnace chamber cools to room temperature, the silicon rod is taken out by a robot;

[0093] Instrument detection:

[0094] Diameter measurement: Use a laser diameter gauge to confirm the silicon rod diameter (±1mm);

[0095] Oxygen content detection: Use ICP-OES or SIMS to measure the oxygen content of the crystal to ensure C 实际 ≤3ppma;

[0096] Defect density detection: Combining optical microscopy and transmission electron microscopy (TEM) to count the number of defects per unit area to ensure that device-level requirements are met;

[0097] Record the head and tail utilization rate and available yield rate data as a basis for subsequent process improvements.

[0098] In summary, this invention significantly improves the growth quality and efficiency of ultra-large-scale silicon single crystals by combining high-temperature superconducting magnetron technology with a dynamic control mechanism. It employs gradient layered insulation materials and an intelligent temperature field control strategy to effectively reduce melt convection and radial temperature differences, ensuring temperature field uniformity. Furthermore, by dynamically adjusting parameters such as magnetic field intensity and argon flow rate through real-time monitoring of oxygen content and melt flow rate, it precisely suppresses impurity diffusion and optimizes crystal growth dynamics. A multi-objective optimization algorithm is introduced to balance growth rate and crystal quality, avoiding performance losses associated with optimizing a single parameter.

[0099] Example 2

[0100] Embodiment 2 is the second embodiment of the present invention. This embodiment differs from the first embodiment in that it further provides an ultra-large-scale silicon single crystal growth system based on high-temperature superconducting magnetron technology, including:

[0101] High-temperature superconducting magnetron modules, used to generate a magnetic field with adjustable strength and uniform spatial distribution, suppressing melt convection and reducing oxygen diffusion;

[0102] Melt heating and temperature control module, used to achieve melt heating through induction heaters and gradient insulation layers, and to control temperature field uniformity in combination with temperature sensors and intelligent algorithms;

[0103] Melt flow monitoring and control module, used to monitor the melt surface flow velocity in real time and dynamically adjust the magnetic field strength or argon flow rate according to the flow velocity threshold;

[0104] Oxygen content monitoring and control module, used to measure the oxygen concentration and gas phase partial pressure of the melt online, and control the oxygen content at the top of the crystal by coordinating the magnetic field, argon flow rate and crucible speed;

[0105] Multi-objective optimization control module, which uses a multi-objective optimization algorithm to collaboratively optimize growth parameters and balance growth rate and crystal quality;

[0106] The tail solidification control unit is used to gradually reduce the pull-down speed and magnetic field intensity at the end of crystal growth, and reduce tail stress and defects through the insulation stage;

[0107] The central control platform is used to integrate data acquisition, multi-physics field simulation and real-time feedback control, and coordinate the operation of each module to achieve stable growth of silicon single crystals.

[0108] This embodiment also provides a computer device, which is suitable for a method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement a method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology as proposed in the above embodiment.

[0109] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0110] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for growing ultra-large-scale silicon single crystals based on high-temperature superconducting magnetron technology, characterized by: include: S1, perform self-test on the sensor to ensure accuracy, evacuate and cool the magnet to superconducting operating temperature, set and calibrate the initial magnetic field strength, and establish a stable environment for the growth process; S2, based on the simulation, defines the thermal conductivity function of each layer of gradient graphite felt, optimizes and calculates the temperature field uniformity index using multi-layer thermal conductive materials; monitors the temperature distribution in real time and dynamically adjusts the heating power; introduces argon gas after the melt reaches the set temperature, adjusts the magnetic field by monitoring the flow rate to suppress the melt flow, and simultaneously records the initial value of the oxygen concentration; S3, contacting the seed crystal with the melt to form a wetting interface, setting the initial crystal pulling parameters and gradually increasing the magnetic field to suppress convection to obtain a sufficient Hartmann number; Real-time monitoring of flow rate and oxygen concentration, dynamic adjustment of magnetic field and temperature parameters to ensure stability in the initial stage of crystal growth; S4 updates the growth model through multi-physics coupled simulation, uses a multi-objective optimization algorithm to balance oxygen content and growth rate, and screens the optimal parameter combination. Feedback control is combined with real-time adjustment of heating and cooling power to maintain a dynamic balance during the growth process. S5, gradually reduce the pulling speed and magnetic field strength to reduce tail stress, maintain the temperature field to promote slow solidification; after demagnetization, cool to room temperature, extract the silicon rod and test the oxygen content, defects and size to ensure that the finished product quality meets the standards.

2. The ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to claim 1, characterized in that: The thermal conductivity of the gradient graphite felt adopts a layered design, wherein the upper and lower layers have constant thermal conductivity coefficients, and the middle layer adopts a positive temperature coefficient material, which automatically adjusts the thermal conductivity efficiency as the temperature changes to maintain temperature field uniformity.

3. The ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to claim 2, characterized in that: Defining the thermal conductivity function of each layer of gradient graphite felt includes: dividing the gradient graphite felt into multiple functional gradient layers along the thickness direction, and determining the thermal conductivity value range of each layer according to the material design requirements; selecting the function form of the thermal conductivity changing with thickness according to the thermal conductivity performance requirements to describe the gradient change trend of the overall thermal conductivity performance; considering the continuity and physical feasibility of the transition of the thermal conductivity between layers in the function setting process to avoid mutations while meeting the design goals of the overall thermal conductivity performance.

4. The ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to claim 1, characterized in that: Calculating the temperature field uniformity index includes: obtaining the thermocouple temperature values ​​of multiple pixel points in the thermal image and normalizing all temperature values ​​so that they fall into a unified interval for subsequent analysis; based on the normalized temperature values, calculating the difference between all pixel points to measure the fluctuation of temperature in spatial distribution; based on the statistical results of the temperature distribution differences, outputting a comprehensive uniformity index. The closer the value is to the ideal value, the more uniform the temperature field distribution is.

5. The ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to claim 1, characterized in that: Obtaining the Hartmann number includes: forming a stable magnetic field environment by setting up a high-temperature superconductor, and obtaining the spatial distribution of magnetic induction intensity in the silicon single crystal growth melt zone by means of magnetic field simulation or actual measurement; constructing a flow-electromagnetic coupling model of silicon melt under the action of a magnetic field based on the known physical parameters of silicon melt such as conductivity and viscosity, to describe the inhibitory effect of the magnetic field on the flow of the fluid; and calculating a dimensionless numerical index representing the stability of the fluid under the action of a magnetic field based on the established model and the obtained parameters, which is used to evaluate the influence of the superconducting magnetic field on the stability of the crystal growth interface and the convection suppression ability.

6. The ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to claim 1, characterized in that: Screening the optimal parameter combination includes: comprehensively considering multiple performance indicators such as growth efficiency, crystal quality and energy consumption to form an evaluation standard for the parameter combination; collecting corresponding data on multiple performance indicators and comprehensively scoring different indicators through normalization or weighted methods; selecting the parameter set that achieves the optimal balance among various indicators and meets the process requirements as the final growth plan.

7. The ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to claim 1, characterized in that: Real-time adjustment of heating and cooling power in combination with feedback control includes: real-time collection of key parameters such as the temperature, crystal morphology and magnetron field strength of the silicon single crystal growth area; analysis of the collected data by the control system to determine the deviation and adjustment needs of the current heating or cooling power; and dynamic adjustment of the heating and cooling power by the controller based on the analysis results to achieve precise control of temperature and crystal quality.

8. An ultra-large-scale silicon single crystal growth system based on high-temperature superconducting magnetron technology, based on the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to any one of claims 1 to 7, characterized in that: include, High-temperature superconducting magnetron modules, used to generate a magnetic field with adjustable strength and uniform spatial distribution, suppressing melt convection and reducing oxygen diffusion; Melt heating and temperature control module, used to achieve melt heating through induction heaters and gradient insulation layers, and to control temperature field uniformity in combination with temperature sensors and intelligent algorithms; Melt flow monitoring and control module, used to monitor the melt surface flow velocity in real time and dynamically adjust the magnetic field strength or argon flow rate according to the flow velocity threshold; Oxygen content monitoring and control module, used to measure the oxygen concentration and gas phase partial pressure of the melt online, and control the oxygen content at the top of the crystal by coordinating the magnetic field, argon flow rate and crucible speed; Multi-objective optimization control module, which uses a multi-objective optimization algorithm to collaboratively optimize growth parameters and balance growth rate and crystal quality; The tail solidification control unit is used to gradually reduce the pull-down speed and magnetic field intensity at the end of crystal growth, and reduce tail stress and defects through the insulation stage; The central control platform is used to integrate data acquisition, multi-physics field simulation and real-time feedback control, and coordinate the operation of each module to achieve stable growth of silicon single crystals.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the ultra-large-scale silicon single crystal growth method based on high-temperature superconducting magnetron technology according to any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Growth regulation and control method and device for silicon carbide sublimation method crystal and electronic equipment

    CN121065816A

  • Magnetic pollution source prevention and control optimization method in magnetic suspension compressor assembly process

    CN121960071A