Intelligent temperature control system and method for crystal growth furnace

Abstract

The application relates to the technical field of temperature control of crystal growth equipment, in particular to an intelligent temperature control system and method of a crystal growth furnace. The system comprises a data acquisition module, a calculation module, a control module and an optimization module; wherein the data acquisition module acquires physical parameters in the furnace and crystal quality parameters in real time; the calculation module dynamically determines a temperature control precision level and a temperature deviation threshold value by calculating thermal resistance values of non-uniform grid nodes, and outputs a control instruction; the control module adjusts the temperature in response to the control instruction; and the optimization module establishes a correlation model of defect density and temperature fluctuation, and corrects the thermal resistance values of the grid nodes. The application realizes accurate temperature control in the whole cycle of crystal growth, improves temperature regulation precision and control strategy adaptability, and improves crystal quality and yield.

Description

Technical Field

[0001] This application relates to the field of temperature control technology for crystal growth equipment, specifically to an intelligent temperature control system and method for a crystal growth furnace. Background Technology

[0002] As core foundational materials in modern electronic information technology, optoelectronic communication, aerospace, and other fields, the quality of crystal materials is a key factor determining the performance and reliability of downstream devices. During crystal growth, temperature control is a decisive factor affecting the final crystal quality; precise thermal field regulation is crucial for obtaining high-quality crystals.

[0003] Currently, commonly used traditional temperature control techniques in crystal growth mainly include: PID control, which uses a few thermocouples to collect temperature data and employs a proportional-integral-derivative algorithm for feedback regulation; segmented fixed parameters, which preset fixed temperature parameters and adjustment step sizes for different stages of crystal growth and switch control modes according to time nodes; and trial-and-error methods, which rely on operators manually adjusting parameters such as heating power and gas flow rate based on past experience. However, these traditional techniques generally suffer from the following problems: On the one hand, traditional methods mainly rely on limited local point measurements, making it difficult to comprehensively and accurately reflect the complex three-dimensional temperature field distribution within the furnace, especially failing to effectively capture the dynamic temperature changes in the critical area of ​​the crystal growth interface, resulting in a lack of sufficient global information for control decisions; on the other hand, whether existing technologies use a single PID parameter throughout the entire growth process or simply switch between preset parameters in stages, they are unable to dynamically adapt to the differentiated requirements of temperature accuracy, gradient distribution, and rate of change at different stages of crystal growth, affecting crystal quality.

[0004] For example, Chinese patent CN119292384B discloses a method and system for optimizing heat management in the crystal growth process. This includes: constructing multiple heating sub-regions and multiple monitoring points based on crystal processing equipment parameters; generating a predicted crystal growth curve based on the parameters of the crystal to be processed, and constructing heating, growth, and cooling cycles based on the predicted growth curve; sequentially setting control strategies for the heating, growth, and cooling cycles; constructing multiple heating sub-regions within the crystal processing equipment, and performing regional temperature control during crystal growth to ensure the solid-liquid interface of the crystal is in a stable temperature field, ensuring crystal growth efficiency and quality, while simultaneously making targeted adjustments to the remaining heating sub-regions to reduce overall heat loss. By setting control strategies for each cycle based on the predicted crystal growth curve, the temperature control within the crystal processing equipment is improved, reducing overall heat loss.

[0005] For example, patent application CN119336101A discloses a machine learning-based crystal growth process optimization control system and method, including: a difference detection module for acquiring the crystal outline of the current crystal growth image and obtaining the crystal growth difference degree; a model building module for acquiring historical crystal growth data and establishing a growth morphology parameter prediction model based on the historical crystal growth data; a morphology prediction module for inputting the current crystal growth difference degree into the growth morphology parameter prediction model to obtain the predicted growth morphology parameters of the current crystal; and a temperature setting module for determining the required growth temperature based on the predicted growth morphology parameters of the current crystal and correcting the required growth temperature based on historical crystal growth data. By establishing a growth morphology parameter prediction model to predict the crystal growth morphology, the optimal temperature required for crystal growth is obtained, thereby improving the crystal growth quality.

[0006] All of the above patents suffer from the problem described in this background: it is difficult to dynamically adapt to the different requirements of crystal growth at different stages for temperature accuracy, gradient distribution and rate of change.

[0007] The information disclosed in this background section is intended only to enhance the understanding of the overall background of this application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0008] The purpose of this application is to provide an intelligent temperature control system and method for a crystal growth furnace. By precisely controlling the temperature throughout the entire crystal growth cycle, it can solve, to some extent, the problems of limited temperature field perception and insufficient control precision adaptability in traditional temperature control methods during crystal growth, thereby effectively improving the quality and yield of crystal growth.

[0009] To achieve the above objectives, this application provides the following technical solution:

[0010] In a first aspect, this application provides an intelligent temperature control system for a crystal growth furnace, including a data acquisition module, a calculation module, a control module, and an optimization module;

[0011] The data acquisition module is used to collect physical parameters and crystal quality parameters inside the crystal growth furnace in real time; the crystal quality parameters include crystal growth parameters and crystal defect detection data.

[0012] The calculation module includes a thermal resistance network unit and a control precision unit;

[0013] The thermal resistance network unit divides the furnace space into non-uniform grid nodes and calculates the thermal resistance value of the non-uniform grid nodes based on the physical parameters; the thermal resistance network unit also outputs the temperature gradient distribution of the crystal growth interface based on the thermal resistance value of the non-uniform grid nodes and the boundary conditions.

[0014] The control precision unit dynamically determines the temperature control precision level and temperature deviation threshold of the crystal growth interface based on the temperature gradient distribution and crystal growth parameters. When the absolute value of the difference between the actual temperature and the target temperature of the crystal growth interface is greater than the temperature deviation threshold, a control command is output.

[0015] The control module is used to respond to the instructions of the control precision unit and drive the actuator to adjust the temperature according to the adjustment mechanism corresponding to the temperature control precision level;

[0016] The optimization module is used to establish a correlation model between defect density and temperature fluctuation based on the crystal defect detection data and historical temperature fluctuation data. After each batch of crystal growth is completed, the thermal resistance value of the non-uniform grid node is corrected based on the crystal defect data.

[0017] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0018] The thermal resistance network unit is configured with a meshing strategy to divide the furnace space into non-uniform mesh nodes. The meshing strategy specifically includes:

[0019] Based on the three-dimensional structure of the furnace body, a basic grid covering the entire furnace chamber is divided, and a minimum grid size is set.

[0020] Establish a mapping relationship between crystal growth stages and grid density, wherein: the highest grid density is adopted in the seed crystal stage, and the node density of the seed crystal tip, melt surface, solid-liquid interface area and its neighborhood is increased; the node density of the contact area between the crystal sidewall and the melt and its adjacent area is increased in the constant diameter stage; the node density of the finishing stage is only retained in the furnace edge area and heating area, and the node density is less than the basic grid density.

[0021] The current growth stage of the crystal is determined based on the crystal growth parameters, and the grid density is adjusted according to the mapping relationship between the crystal growth stage and the grid density.

[0022] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0023] The control accuracy unit is configured with a control accuracy level classification strategy to determine the temperature control accuracy level of the crystal growth interface and the temperature deviation threshold under the temperature control accuracy level. The control accuracy level classification strategy specifically includes:

[0024] The total risk value is calculated based on the thermal stress factor, growth stability factor, and interface curvature factor.

[0025] The temperature control accuracy level of the crystal growth interface is determined based on the magnitude of the total risk value, and a basic temperature deviation threshold is set for each temperature control accuracy level; the temperature control accuracy levels include ultra-precision mode, high-precision mode, standard mode, and relaxed mode;

[0026] The temperature deviation threshold is dynamically corrected based on the basic temperature deviation threshold by adjusting the factor.

[0027] The adjustment factors include crystal length factor, crystal gradient factor, and crystal defect factor; the temperature deviation threshold is obtained by coupling the basic temperature deviation threshold with the adjustment factors.

[0028] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0029] The control module is equipped with an actuator for adjusting the temperature according to a regulation mechanism corresponding to the temperature control accuracy level. The actuator includes a multi-zone heater, a gas flow valve, and a crucible lifting mechanism. The regulation mechanism specifically includes:

[0030] After receiving the ultra-precision mode command, the control module starts some multi-zone heaters to perform preset minimum power step increment adjustment. When the cumulative power adjustment of the heater reaches the preset maximum range, the gas flow valve is activated to adjust the flow rate with a preset minimum flow step, and the radial gradient is stabilized by adjusting the preset minimum amplitude of the crucible rotation speed and lifting speed.

[0031] When the control module receives the high-precision mode command, it starts some multi-zone heaters to perform stepped power step adjustment. If the absolute value of the difference between the actual temperature and the target temperature value is still not less than the temperature deviation threshold after continuous adjustment of the preset number of steps, the gas flow valve is activated to adjust the flow rate in a coordinated manner with the preset basic flow rate step. At the same time, the crucible lifting mechanism adjusts the rotation speed once every fixed time to control the temperature gradient fluctuation.

[0032] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0033] The regulation mechanism also includes:

[0034] When the control module receives the standard mode command, it drives all multi-zone heaters to adjust their power synchronously according to a preset ratio. At the same time, the gas flow valve adjusts in coordination with a step size greater than the preset base flow rate and less than the preset maximum flow rate. When the crystal diameter fluctuation is greater than the preset first fluctuation threshold, the crystal pulling speed of the crucible lifting mechanism is activated for adjustment.

[0035] If the control module receives a relaxed mode command, it will start all heaters to adjust the preset maximum power step size, and at the same time set the gas flow valve within the preset flow range. When the crystal diameter fluctuation is greater than the preset second fluctuation threshold, the crucible lifting mechanism will be activated to adjust the preset maximum amplitude.

[0036] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0037] The thermal resistance network unit is also configured with a temperature gradient calculation strategy, which is used to output the temperature gradient distribution of the crystal growth interface based on the thermal resistance values ​​between non-uniform mesh nodes and boundary conditions. The temperature gradient calculation strategy specifically includes:

[0038] The physical parameters include temperature data, thermal image data, motion parameters, and gas parameters; the thermal resistance between non-uniform grid nodes is calculated based on the temperature data, motion parameters, and gas parameters.

[0039] Based on the aforementioned thermal resistance value and the law of conservation of energy, an unsteady-state heat conduction equation is constructed, and the heat conduction equation is solved by combining boundary conditions;

[0040] The heat conduction equation is discretized into a system of linear equations. The temperature values ​​of all non-uniform grid nodes are obtained by solving the system of equations using a numerical algorithm, thus obtaining an approximate distribution of the three-dimensional temperature field.

[0041] By taking the spatial derivative of the three-dimensional temperature field at the crystal growth interface, the temperature gradient distribution at the crystal growth interface is obtained.

[0042] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0043] The optimization module is configured with an association model construction strategy for constructing an association model between defect density and temperature fluctuation; the association model construction strategy specifically includes:

[0044] Based on the timeline of crystal growth, the defect detection data and temperature fluctuation data of crystals at the same growth stage in each batch are time-aligned.

[0045] The crystal is divided into different spatial regions, and a coordinate mapping relationship between each spatial region of the crystal and the non-uniform grid of the three-dimensional temperature field is established. The correlation results of the corresponding data of the spatial regions are recorded. The correlation results are classified according to the crystal spatial regions and growth stages to form data subsets. The defect characteristic value and temperature fluctuation characteristic value of each subset are calculated.

[0046] A mapping relationship between defect characteristic values ​​and temperature fluctuation characteristic values ​​is established by using scatter plots and correlation calculations.

[0047] Based on the mapping relationship, a temperature fluctuation and defect density correlation table is generated for each region to form a temperature fluctuation and defect density correlation model.

[0048] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0049] The optimization module is also configured with a thermal resistance correction strategy, used to correct the thermal resistance of the non-uniform mesh nodes based on crystal defect data after each batch of crystal growth is completed; the thermal resistance correction strategy specifically includes:

[0050] Based on offline defect detection data, identify defect-dense regions on the crystal, and locate the set of non-uniform grid nodes corresponding to the defect-dense regions through the coordinate mapping relationship;

[0051] The temperature fluctuation and defect density correlation model is invoked to determine whether the defect-dense region is significantly positively correlated with the temperature fluctuation. If the correlation model shows that the temperature fluctuation and defect density in the defect-dense region are significantly positively correlated, the corresponding grid node is marked as a priority correction node; otherwise, the corresponding node is recorded as a node to be observed.

[0052] The correction amount of the thermal resistance value of the priority correction node is calculated based on the defect influence coefficient, temperature prediction error, and correlation strength coefficient of the region corresponding to the priority correction node; the thermal resistance value of the priority correction node is updated based on the correction amount of the thermal resistance value of the priority correction node.

[0053] As a preferred embodiment of the intelligent temperature control system for a crystal growth furnace described in this application, wherein:

[0054] The data acquisition module is equipped with thermocouples and acquires the temperature data through the thermocouples;

[0055] The data acquisition module is also equipped with an infrared thermal imager, and acquires the thermal image data through the infrared thermal imager;

[0056] The crystal growth parameters include crystal diameter and crystal growth rate;

[0057] The crystal diameter is obtained by scanning the outer diameter of the crystal with a laser beam and outputting the crystal diameter value in real time; the crystal growth rate is calculated by the pulling displacement and time difference; the crystal defect detection data includes online detection data and offline detection data.

[0058] Secondly, this application provides an intelligent temperature control method for a crystal growth furnace, comprising:

[0059] Real-time acquisition of physical parameters and crystal quality parameters within the crystal growth furnace; the crystal quality parameters include crystal growth parameters and crystal defect detection data.

[0060] The furnace space of the crystal growth furnace is divided into non-uniform grid nodes, and the thermal resistance of the non-uniform grid nodes is calculated based on the physical parameters; the temperature gradient distribution of the crystal growth interface is calculated based on the thermal resistance of the non-uniform grid nodes and the boundary conditions.

[0061] Based on the temperature gradient distribution and crystal growth parameters, the temperature control accuracy level and temperature deviation threshold of the crystal growth interface are dynamically determined. When the absolute value of the difference between the actual temperature of the crystal growth interface and the target temperature is greater than the temperature deviation threshold, the temperature is adjusted according to the adjustment mechanism corresponding to the temperature control accuracy level.

[0062] Based on the crystal defect detection data and historical temperature fluctuation data, a correlation model between defect density and temperature fluctuation is established. After each batch of crystal growth is completed, the thermal resistance value of the non-uniform grid nodes is corrected based on the crystal defect data.

[0063] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0064] By dividing the non-uniform grid nodes into thermal resistance network units and combining multi-dimensional physical parameters to calculate thermal resistance values ​​and output the temperature gradient distribution at the crystal growth interface, the limitations of traditional single-point temperature measurement are overcome, enabling precise perception of the three-dimensional temperature field of the furnace. The control precision unit dynamically determines the temperature control precision level and deviation threshold, making temperature adjustment more targeted at different growth stages and significantly reducing the impact of temperature fluctuations on crystal quality. The control module adopts a differentiated adjustment mechanism according to different temperature control precision levels, coordinating with actuators such as multi-zone heaters, gas flow valves, and crucible lifting mechanisms, solving the problem that traditional fixed adjustment modes are difficult to adapt to the complex dynamic process of crystal growth, and improving the system's response speed and stability to temperature changes. Through precise temperature control throughout the entire crystal growth cycle, dynamically adapted control strategies, and a system self-optimization mechanism, crystal defects caused by temperature gradient fluctuations and insufficient control precision are effectively reduced, significantly improving the quality consistency and yield of crystal growth and reducing production costs. Attached Figure Description

[0065] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0066] Figure 1 A schematic diagram of the structure of an intelligent temperature control system for a crystal growth furnace provided in this application;

[0067] Figure 2A flowchart of an intelligent temperature control method for a crystal growth furnace provided in this application. Detailed Implementation

[0068] The technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0069] Example 1

[0070] like Figure 1 As shown in the figure, this embodiment introduces an intelligent temperature control system for a crystal growth furnace, including a data acquisition module, a calculation module, a control module, and an optimization module;

[0071] The data acquisition module is used to collect physical parameters and crystal quality parameters inside the crystal growth furnace in real time;

[0072] The physical parameters include temperature data, thermal imaging data, motion parameters, and gas parameters; wherein, the temperature data is continuously collected at a fixed sampling frequency by thermocouples distributed at different locations within the growth furnace, providing real-time temperature data for each key location within the growth furnace.

[0073] In this embodiment, the distribution locations of the multiple sets of thermocouples include: the melt surface and raw material area, the solid-liquid interface and crystal growth area, the vicinity of the heating element and the inner and outer sides of the insulation layer, the gas passage and the key areas of the furnace cavity; the gas passage includes the gas inlet and the gas outlet; the key areas of the furnace cavity include the top of the furnace cavity, the bottom of the furnace cavity and the side wall.

[0074] The thermal image data acquisition method is as follows: the two-dimensional temperature field of the crystal growth interface is imaged by an infrared thermal imager to obtain the thermal image data, which can compensate for the spatial limitations of thermocouple point temperature measurement; specifically, the infrared thermal imager takes two-dimensional thermal images of the crystal growth interface in real time at a fixed frequency, and extracts the temperature distribution of the solid-liquid interface region in the thermal image after emissivity correction.

[0075] The motion parameters include crucible rotation speed and lifting speed;

[0076] Optionally, the crucible rotation speed is obtained by detecting the frequency of the rotating magnetic field in real time at a fixed frequency using a magnetoelectric speed sensor at the end of the crucible drive motor shaft; the pulling speed is the rate at which the seed crystal or the crystal itself is pulled upward during crystal growth, reflecting the speed of axial crystal growth, and it is acquired by counting the pulse signals at a fixed sampling frequency using a photoelectric encoder on the ball screw in the crystal pulling mechanism to obtain the pulling displacement, and then calculating the real-time pulling speed.

[0077] The gas parameters include the inlet flow rate and the furnace pressure. The inlet flow rate is collected in real time by a mass flow meter in the protective gas pipeline. The furnace pressure is measured in real time by a pressure sensor to measure the absolute pressure inside the furnace, and the pressure is adjusted and stabilized by a pressure balance valve.

[0078] The crystal quality parameters include crystal growth parameters and crystal defect detection data;

[0079] The crystal growth parameters include crystal diameter and crystal growth rate; the crystal diameter is obtained by scanning the outer diameter of the crystal with a laser beam and outputting the crystal diameter value in real time; the crystal growth rate is calculated by the pulling displacement and time difference; optionally, the growth rate can also be obtained by measuring the crystal mass increment in real time with a weighing sensor and converting the crystal growth rate in combination with the crystal density.

[0080] The crystal defect detection data includes online detection data and offline detection data;

[0081] Optionally, the online detection focuses a laser beam onto the crystal surface during the growth process and analyzes the defect density through scattering spectroscopy to capture the defect changes during the crystal growth process; the offline detection involves selecting a crystal sample at the end of crystal growth to perform lattice integrity analysis and generate a defect density distribution map to analyze the types, distribution patterns, and overall density of defects; the defects include, but are not limited to, dislocations, vacancies, cracks, and bubbles.

[0082] The calculation module includes a thermal resistance network unit and a control precision unit;

[0083] The thermal resistance network unit divides the furnace space into non-uniform grid nodes and calculates the thermal resistance value between the non-uniform grid nodes based on the physical parameters; the thermal resistance network unit also outputs the temperature gradient distribution of the crystal growth interface based on the thermal resistance value of the non-uniform grid nodes and the boundary conditions.

[0084] The step of dividing the furnace space into non-uniform grid nodes is as follows:

[0085] Based on the three-dimensional structure of the furnace body, a basic grid covering the entire furnace chamber is pre-divided, and a minimum grid size is set.

[0086] In this embodiment, for a cylindrical furnace, a cylindrical coordinate system is used with the central axis of the furnace as the polar axis. An initial basic grid is divided based on the diameter and height of the furnace body, and the minimum grid size is set according to the crystal growth stage. The crystal initiation stage, as the starting stage of crystal growth, has the highest requirement for solid-liquid interface stability and the smallest grid size. The constant diameter stage is the main growth stage of the crystal, which needs to balance accuracy and computational efficiency, and the grid size is moderate. The closing stage is the final stage of crystal growth, which has lower requirements for detail accuracy and the largest grid size.

[0087] Establish the mapping relationship between crystal growth stages and grid density;

[0088] Specifically, during the seed crystal stage, the seed crystal initially contacts the melt, and the highest grid density is used to increase the node density of the seed crystal tip, the melt surface, the solid-liquid interface area, and its neighboring regions; during the constant diameter stage, the crystal diameter grows steadily, and the node density of the crystal sidewall contact area with the melt and its adjacent areas is increased; during the finishing stage, the crystal gradually detaches from the melt, retaining only the nodes in the furnace edge area and the heating area, and the node density is less than the basic grid density.

[0089] In this embodiment, the mesh density during the crystal initiation stage is set to be more than 3 times that of the base mesh; the mesh density during the constant diameter stage is about 2 times that of the base mesh; and the mesh density during the finishing stage is about half that of the base mesh.

[0090] The current growth stage of the crystal is determined based on the crystal growth parameters, and the grid density is adjusted according to the mapping relationship between the crystal growth stage and the grid density. The specific determination rules are as follows:

[0091] If the crystal diameter is less than or equal to a first threshold value for the target crystal diameter, or if the crystal growth rate is less than or equal to the initial low growth rate and the duration is greater than or equal to a first time threshold value, then it is determined to be the crystal initiation period; if the crystal diameter is stable within a second threshold value range for the target crystal diameter and the duration is greater than or equal to a second time threshold value, then it is determined to be the constant diameter period; if the crystal length is greater than or equal to a target crystal length threshold value, and within a range less than a third time threshold value, the rate of decrease in the crystal growth rate is greater than or equal to a preset speed threshold value, then it is determined to be the finishing period.

[0092] The thermal resistance between the non-uniform grid nodes is calculated based on the temperature data, airflow rate, and crucible rotation speed; the node thermal resistance refers to the thermal resistance between two adjacent discrete nodes, used to quantify the heat transfer resistance between nodes.

[0093] The temperature gradient distribution at the crystal growth interface, based on the thermal resistance between non-uniform grid nodes and boundary conditions, specifically includes:

[0094] An unsteady-state heat conduction equation is constructed based on the thermal resistance between the nodes and the law of conservation of energy. The heat conduction equation includes heat conduction, convection, radiation, and internal heat sources.

[0095] The heat conduction equation is solved using the finite element method, taking into account boundary conditions. These boundary conditions mainly include temperature boundary conditions, heat flux boundary conditions, convection boundary conditions, and radiation boundary conditions.

[0096] The heat conduction equation is discretized on non-uniform grid nodes and transformed into a system of linear equations in matrix form. The temperature values ​​of all non-uniform grid nodes are obtained by solving the system of equations using numerical algorithms, thus obtaining an approximate distribution of the three-dimensional temperature field.

[0097] By taking the spatial derivative of the three-dimensional temperature field at the crystal growth interface, the temperature gradient distribution at the crystal growth interface is obtained.

[0098] In this embodiment, the boundary conditions include: the heater surface temperature boundary or heat flux density boundary determined by the heater power; the heat dissipation boundary of the furnace wall, such as the heat transfer coefficient and cooling water temperature of the water-cooled furnace wall; the gas convection boundary caused by the gas flow determined by the gas velocity and temperature; and the temperature boundary measured at key locations such as the crucible wall or furnace shell.

[0099] Optionally, the finite element method discretizes the continuous physical field into a finite number of interconnected elements, and by solving the system of equations composed of these finite number of elements, the solution to the original problem is approximately obtained.

[0100] The control precision unit dynamically determines the temperature control precision level and temperature deviation threshold of the crystal growth interface based on the temperature gradient distribution and crystal growth parameters. When the absolute value of the difference between the actual temperature and the target temperature of the crystal growth interface is greater than the temperature deviation threshold, a control command is output.

[0101] The temperature control accuracy levels of the crystal growth interface include ultra-precision mode, high-precision mode, standard mode, and relaxed mode.

[0102] The temperature control accuracy level of the crystal growth interface is determined by integrating the total risk value of multiple risk factors. These risk factors include a thermal stress factor, a growth stability factor, and an interface curvature factor. The thermal stress factor is calculated based on the temperature axial gradient modulus and the rate of change of the temperature gradient at the crystal growth interface. The growth stability factor is calculated based on the crystal growth rate and the rate of change of the crystal diameter. The interface curvature factor is calculated based on the curvature of the crystal growth interface.

[0103] Specifically, the curvature calculation steps for the crystal growth interface are as follows:

[0104] The initial outline of the crystal growth interface is formed by connecting all pixels whose temperature value is equal to the melting point of the crystal extracted from the thermal image data.

[0105] The initial contour is subjected to noise reduction processing, such as Gaussian filtering or bilateral filtering, to obtain a smooth and continuous contour curve of the crystal growth interface.

[0106] The contour curve is divided into multiple equal curve segments, and the equation of each curve segment is obtained by polynomial fitting. The curvature of each curve segment is then calculated.

[0107] The average value of the normalized curvature of all curve segments is taken as the final curvature value of the crystal growth interface.

[0108] The total risk value is obtained by weighted summation of the thermal stress factor, growth stability factor, and interface curvature factor. The weighting coefficients are dynamically adjusted according to the current growth stage of the crystal. Specifically, during the crystal initiation stage, the crystal diameter is the smallest, and the temperature gradient has the greatest impact on lattice defects, so the thermal stress factor has the largest weight. During the intermediate period between the crystal initiation stage and the constant diameter stage, the crystal diameter expands, and the temperature gradient and growth stability are equally important, so both the thermal stress factor and the growth stability factor have relatively high weights. During the constant diameter stage, it is necessary to maintain a constant crystal diameter, so the growth stability factor has the highest weight. During the final stage, the crystal growth ends, and stress concentration should be avoided, so the thermal stress factor has the smallest weight.

[0109] The specific method for determining the temperature control accuracy level of the crystal growth interface based on the total risk value is as follows: If the total risk value is greater than the first risk threshold, it is determined to be in ultra-precision mode, and the basic temperature deviation threshold in this mode is the first deviation threshold; if the total risk value is less than or equal to the first risk threshold and greater than the second risk threshold, it is determined to be in high-precision mode, and the basic temperature deviation threshold in this mode is the second deviation threshold; if the total risk value is less than or equal to the second risk threshold and greater than the third risk threshold, it is determined to be in standard mode, and the basic temperature deviation threshold in this mode is the third deviation threshold; if the total risk value is less than or equal to the third risk threshold, it is determined to be in relaxed mode, and the basic temperature deviation threshold in this mode is the fourth deviation threshold.

[0110] Based on the basic temperature deviation threshold, the temperature deviation threshold is dynamically corrected by an adjustment factor, increasing the adaptability of temperature control accuracy to the dynamic characteristics of crystal growth. This allows for strict control to suppress defect generation during critical crystal growth stages, while also appropriately relaxing temperature control accuracy during non-critical stages to reduce ineffective adjustments, thereby reducing system energy consumption while ensuring crystal quality. The adjustment factor includes a crystal length factor, a crystal gradient factor, and a crystal defect factor. The temperature deviation threshold is obtained by multiplying the basic temperature deviation threshold by the adjustment factor.

[0111] Optionally, a crystal length factor is assigned based on the actual length of the crystal, and the crystal length factor is positively correlated with the actual length of the crystal; a crystal gradient factor is assigned based on the temperature gradient of the crystal, and the crystal gradient factor is negatively correlated with the temperature gradient of the crystal; a crystal defect factor is assigned based on the probability of crystal defects, and the crystal defect factor is negatively correlated with the probability of crystal defects.

[0112] Each adjustment factor is normalized and dimensionless; the temperature deviation threshold is calculated based on the adjustment factor and the base temperature deviation threshold; wherein the temperature deviation threshold is the cumulative product of each adjustment factor and the base temperature deviation threshold;

[0113] The absolute value of the difference between the actual temperature and the target temperature of the crystal growth interface is calculated. When the absolute value is greater than the temperature deviation threshold, a control command is output. The target temperature is the target temperature field of the crystal growth interface, which is jointly determined by the crystal material and the crystal growth stage. The control command includes the temperature control accuracy level to be adjusted, the target temperature value, and the corresponding temperature deviation threshold.

[0114] The control module is used to respond to the instructions of the control precision unit and drive the actuator to perform temperature adjustment according to the adjustment mechanism corresponding to the temperature control precision level; the actuator includes a multi-zone heater, a gas flow valve and a crucible lifting mechanism;

[0115] The multi-zone heater divides the furnace into multiple heating zones according to the furnace structure and crystal growth requirements. For example, it is divided into a top heating zone, a middle heating zone, and a bottom heating zone along the furnace axis, and a center heating zone and an edge heating zone along the furnace radial direction. Each heating zone is equipped with an independent heating device, and the heating power can be adjusted individually to achieve independent control of the temperature at different spatial locations, thus meeting the requirements for temperature gradient and uniformity during crystal growth.

[0116] The gas flow valve changes the gas flow rate by adjusting the valve opening, thereby affecting the heat transfer and crystal growth conditions inside the furnace; the gas flow valve includes two independent gas paths for protective gas and doping gas, and the gas flow rate can be adjusted separately.

[0117] The crucible lifting mechanism comprises two independent systems: a lifting drive and a rotation drive. By adjusting the lifting displacement, rotation speed, and crystal pulling speed of the crucible, it indirectly affects the solid-liquid interface morphology, temperature gradient distribution, and crystal growth rate.

[0118] The regulation mechanism includes:

[0119] After receiving the command for ultra-precision mode, the control module starts some multi-zone heaters to perform preset minimum power step increment adjustment. When the cumulative power adjustment of the heater reaches the preset maximum range, the gas flow valve is activated to adjust the flow rate with a preset minimum flow step, and the radial gradient is stabilized by adjusting the preset minimum amplitude of crucible rotation speed and lifting speed.

[0120] When the control module receives the high-precision mode command, it starts some multi-zone heaters to perform stepped power step adjustment. If the absolute value of the difference between the actual temperature and the target temperature value is still not less than the temperature deviation threshold after continuous adjustment of the preset number of steps, the gas flow valve is started to adjust in coordination with the preset basic flow step. At the same time, the crucible lifting mechanism adjusts the speed once every fixed time to control the temperature gradient fluctuation.

[0121] When the control module receives the standard mode command, it drives all multi-zone heaters to adjust their power synchronously according to a preset ratio. At the same time, the gas flow valve adjusts in coordination with a step size greater than the preset base flow rate and a step size less than the preset maximum flow rate. When the crystal diameter fluctuation exceeds the preset first fluctuation threshold, the crystal lifting speed of the crucible lifting mechanism is activated for fine adjustment.

[0122] If the control module receives a relaxed mode command, it will start all heaters to make coarse adjustments with larger power steps, and at the same time set the gas flow valve within the preset flow range. When the crystal diameter fluctuation exceeds the preset second fluctuation threshold, the crucible lifting mechanism will be activated to make preset maximum amplitude adjustments.

[0123] Optionally, the adjustment mechanism corresponding to the ultra-precision mode specifically includes: when the absolute value of the difference between the actual temperature and the target temperature value of a single heating zone or a small area exceeds the temperature deviation threshold, the multi-zone heater in that area is activated with a preset minimum power as the step size, and temperature is adjusted in conjunction with a high-frequency cycle, while the temperature of that area is monitored in real time. When the absolute value of the difference between the actual temperature and the target temperature value is less than half of the temperature deviation threshold, the heater heating is stopped; if the cumulative power adjustment of the heater reaches the preset maximum range, the gas flow valve is activated with a preset minimum flow rate as the step size, and flow rate is adjusted in conjunction with a medium-frequency cycle, while the uniformity of the melt surface temperature is monitored in real time to avoid local temperature being too low due to airflow disturbance; when the radial temperature gradient of the crystal growth interface is close to the preset proportion of the critical gradient, the crucible lifting mechanism is activated with a preset minimum amplitude as the step size, and rotation speed is adjusted in conjunction with a low-frequency cycle, while the lifting speed is adjusted proportionally according to the deviation between the crystal diameter and the target crystal diameter with a preset minimum amplitude as the step size.

[0124] Optionally, the adjustment mechanism corresponding to the high-precision mode specifically includes: when the absolute value of the difference between the actual temperature and the target temperature value of a single heating zone exceeds the temperature deviation threshold, the multi-zone heater of that zone is activated to perform step-by-step temperature adjustment with a preset base power as the step size; if the absolute value of the temperature difference within the preset adjustment time range still exceeds the temperature deviation threshold, the adjustment step size is increased, and the temperature of that zone is periodically monitored using medium and high frequency, until the absolute value of the difference between the actual temperature and the target temperature value is less than the temperature deviation threshold; when the cumulative power adjustment of the heater reaches the preset maximum range, the gas flow valve is activated with a preset base flow rate as the step size, and the protective gas flow rate is adjusted in conjunction with a medium and low frequency cycle, allowing for slight fluctuations in the gas flow rate; when the radial temperature gradient of the crystal growth interface is close to the preset proportion of the critical gradient, the crucible lifting mechanism is activated with a preset minimum amplitude as the step size, and the rotation speed is adjusted in conjunction with a low frequency cycle, while the lifting speed is adjusted proportionally according to the deviation between the crystal diameter and the target crystal diameter with a preset base minimum amplitude as the step size.

[0125] Optionally, the adjustment mechanism corresponding to the standard mode specifically includes: when the absolute value of the difference between the actual temperature and the target temperature value of multiple non-adjacent heating zones exceeds the temperature deviation threshold, all multi-zone heaters are activated to synchronously adjust the temperature according to a preset ratio, while monitoring the average temperature of the entire furnace in conjunction with a medium-low frequency sampling cycle until the absolute value of the difference between the average temperature of the entire furnace and the target temperature value is less than the temperature difference threshold; if the cumulative power adjustment of the heater reaches the preset maximum range, the gas flow valve is activated with a preset basic flow rate as the step size, and the protective gas flow rate is adjusted in conjunction with a medium-low frequency cycle; when the crystal diameter fluctuation exceeds the preset fluctuation first threshold for a continuous preset period, the crucible lifting mechanism is activated with a preset basic amplitude as the step size, and the lifting speed is adjusted proportionally according to the deviation between the crystal diameter and the target crystal diameter, while maintaining the rotation speed within a preset range, without frequent adjustments, to ensure that the melt volume is relatively stable.

[0126] Optionally, the adjustment mechanism corresponding to the relaxed mode specifically includes: when the absolute value of the difference between the actual temperature and the target temperature value of multiple non-adjacent heating zones exceeds the temperature deviation threshold, all multi-zone heaters are activated with a preset maximum power as the fixed step size, and temperature adjustment is performed in conjunction with a low-frequency cycle. At the same time, the heating zone power is allowed to fluctuate within a preset fluctuation range until the absolute value of the difference between the average temperature of the entire furnace and the target temperature value is less than the temperature difference threshold. If the absolute value of the difference between the average temperature of the entire furnace and the target temperature value after heater adjustment still far exceeds the temperature difference threshold or the furnace temperature is too high, the gas flow valve is activated to adjust the flow rate to a preset flow range, and the average temperature of the entire furnace is detected in conjunction with a low-frequency cycle. When the crystal diameter fluctuation exceeds the preset second fluctuation threshold, the crucible lifting mechanism is activated to fix the lifting speed to a stage experience value. When the crystal diameter fluctuation exceeds the preset large fluctuation range, a preset maximum amplitude adjustment is performed, while the rotation speed is maintained within a preset range without strict control.

[0127] The optimization module is used to establish a correlation model between defect density and temperature fluctuation based on the crystal defect detection data and historical temperature fluctuation data. After each batch of crystal growth is completed, the thermal resistance value of the non-uniform grid node is corrected based on the crystal defect data.

[0128] The establishment of the correlation model between defect density and temperature fluctuation specifically includes:

[0129] Based on the timeline of crystal growth, the defect detection data and temperature fluctuation data of crystals at the same growth stage in each batch are time-aligned.

[0130] Optionally, during online detection, a time-stamped defect density distribution map is generated at a fixed frequency; the defect density distribution map marks the defect location and density on the crystal growth surface; during offline detection, after crystal growth is completed, the internal defect distribution is mapped to the time node of the growth stage by cutting and positioning; time-stamped three-dimensional temperature field data is extracted from the calculation module, and the difference and duration between the actual temperature and the target temperature value of the crystal growth interface are marked; the crystal defect detection data and temperature fluctuation data of the same growth stage are bound by the timestamp;

[0131] The crystal is divided into different spatial regions, for example, along the axial direction into a top region, a middle region, and a bottom region, and along the radial direction into a central region and an edge region. In the three-dimensional temperature field, each non-uniform grid node is divided into regions according to the same rules to ensure that the defect data of a certain region of the crystal can be directly mapped to the grid node of the same region in the three-dimensional temperature field. The coordinate mapping relationship between each spatial region of the crystal and the non-uniform grid of the three-dimensional temperature field is established.

[0132] Record the correlation results of the data corresponding to the spatial region. The correlation results include: defect status, temperature fluctuation status, temperature deviation characteristics, and temperature fluctuation time.

[0133] The correlation results are classified according to crystal spatial regions and growth stages to form data subsets, such as edge region data during the constant diameter stage and central region data during the crystallization stage. Defect characteristic values ​​and temperature fluctuation characteristic values ​​are calculated for each subset.

[0134] Optionally, the defect characteristic values ​​include the average defect density and the maximum defect density of the region; the temperature fluctuation characteristic values ​​include the arithmetic mean, the maximum value, and the cumulative duration of the difference between the actual temperature and the target temperature of the region.

[0135] A mapping relationship between defect characteristic values ​​and temperature fluctuation characteristic values ​​is established by using scatter plots and correlation calculations.

[0136] Optionally, a scatter plot can be drawn with the absolute value of the difference between the actual temperature and the target temperature on the horizontal axis and the defect density on the vertical axis to observe the data distribution trend and calculate the correlation coefficient.

[0137] Based on the mapping relationship, a temperature fluctuation and defect density correlation table is generated for each region to form a temperature fluctuation and defect density correlation model; the correlation table includes the absolute value range of the difference between the actual temperature and the target temperature, the range of the difference duration, the corresponding defect density range, and the correlation.

[0138] After each batch of crystal growth is completed, the specific steps for correcting the thermal resistance value of the non-uniform mesh nodes based on crystal defect data are as follows:

[0139] Defect-dense regions on the crystal are identified based on offline defect detection data; these regions refer to areas where the defect density is higher than the average defect density of the entire crystal; and the set of non-uniform grid nodes corresponding to these defect-dense regions is located using the coordinate mapping relationship, denoted as... ;

[0140] The temperature fluctuation and defect density correlation model is invoked to determine whether the defect-dense area is significantly positively correlated with the temperature fluctuation. If the correlation model shows that the temperature fluctuation and defect density in the defect-dense area are significantly positively correlated, the corresponding grid node is marked as a priority correction node; otherwise, the corresponding node is recorded as a node to be observed.

[0141] The correction amount of the thermal resistance value of the priority correction node is calculated based on the defect influence coefficient, temperature prediction error and correlation strength coefficient of the region corresponding to the priority correction node;

[0142] Optionally, the defect influence coefficient is specifically the ratio of the defect density of the crystal region corresponding to the priority correction node to the average defect density of the entire crystal; the larger the ratio, the more significant the influence of temperature fluctuation on the defects in the region; the temperature prediction error is specifically the difference between the temperature of the priority correction node output by the thermal resistance network unit and the measured temperature; the correlation coefficient between the temperature fluctuation and the defect density in the correlation model is used as the correlation strength coefficient between the temperature fluctuation and the defect density in the crystal region corresponding to the priority correction node.

[0143] The correction amount of the thermal resistance value of the priority correction node The calculation formula is as follows:

[0144]

[0145] In the formula, for The current thermal resistance between the node and its nearest neighbor. For the defect impact coefficient, For correlation strength coefficient, This represents the temperature prediction error.

[0146] The thermal resistance value of the priority correction node is updated according to the correction amount of the thermal resistance value of the priority correction node; the thermal resistance value of the priority correction node is the sum of the current thermal resistance value and the correction amount between the node and its nearest neighbor node.

[0147] Example 2

[0148] This embodiment is the second embodiment of this application; it is based on the same inventive concept as Embodiment 1, and refers to... Figure 2 This embodiment introduces an intelligent temperature control method for a crystal growth furnace, including:

[0149] Real-time acquisition of physical parameters and crystal quality parameters within the crystal growth furnace; the crystal quality parameters include crystal growth parameters and crystal defect detection data.

[0150] The furnace space of the crystal growth furnace is divided into non-uniform grid nodes, and the thermal resistance of the non-uniform grid nodes is calculated based on the physical parameters; the temperature gradient distribution of the crystal growth interface is calculated based on the thermal resistance of the non-uniform grid nodes and the boundary conditions.

[0151] Based on the temperature gradient distribution and crystal growth parameters, the temperature control accuracy level of the crystal growth interface is dynamically determined by integrating the total risk value of multiple risk factors, and a basic temperature deviation threshold is set for each temperature control accuracy level. Based on the basic temperature deviation threshold, the temperature deviation threshold is dynamically corrected by adjusting the factor to increase the adaptability of temperature control accuracy to the dynamic characteristics of crystal growth.

[0152] When the absolute value of the difference between the actual temperature of the crystal growth interface and the target temperature is greater than the temperature deviation threshold, the temperature is adjusted according to the adjustment mechanism corresponding to the temperature control accuracy level.

[0153] Based on the crystal defect detection data and historical temperature fluctuation data, a correlation model between defect density and temperature fluctuation is established. After each batch of crystal growth is completed, the defect-dense region on the crystal is identified based on the offline defect detection data. The temperature fluctuation and defect density correlation model is then called to determine whether the defect-dense region is significantly positively correlated with the temperature fluctuation. If the defect-dense region is significantly positively correlated with the temperature fluctuation, the thermal resistance value of the non-uniform grid node is corrected based on the crystal defect data.

[0154] The specific functions of each of the above steps are described in the relevant content of the intelligent temperature control system for a crystal growth furnace described in Example 1, and will not be repeated here.

[0155] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0156] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of protection of this application, and these forms are all within the protection scope of this application.

Claims

1. An intelligent temperature control system for a crystal growth furnace, comprising a data acquisition module, a calculation module, a control module, and an optimization module; The data acquisition module is used to collect physical parameters and crystal quality parameters inside the crystal growth furnace in real time. The crystal quality parameters include crystal growth parameters and crystal defect detection data. The physical parameters include temperature data, thermal imaging data, motion parameters, and gas parameters. The motion parameters include crucible rotation speed and lifting speed. The gas parameters include gas flow rate and furnace pressure. The calculation module includes a thermal resistance network unit and a control precision unit; The thermal resistance network unit divides the furnace space into non-uniform grid nodes and calculates the thermal resistance value of the non-uniform grid nodes based on the physical parameters; the thermal resistance network unit also outputs the temperature gradient distribution of the crystal growth interface based on the thermal resistance value of the non-uniform grid nodes and the boundary conditions. The control precision unit dynamically determines the temperature control precision level and temperature deviation threshold of the crystal growth interface based on the temperature gradient distribution and crystal growth parameters. When the absolute value of the difference between the actual temperature and the target temperature of the crystal growth interface is greater than the temperature deviation threshold, a control command is output. The control module is used to respond to the control command and drive the actuator to adjust the temperature according to the adjustment mechanism corresponding to the temperature control accuracy level; The optimization module is used to establish a correlation model between defect density and temperature fluctuation based on the crystal defect detection data and historical temperature fluctuation data. After each batch of crystal growth is completed, the thermal resistance value of the non-uniform grid node is corrected based on the crystal defect data. The specific method for outputting the temperature gradient distribution of the crystal growth interface based on the thermal resistance values ​​between non-uniform grid nodes and boundary conditions is as follows: An unsteady-state heat conduction equation is constructed based on the aforementioned thermal resistance value and the law of conservation of energy, and the heat conduction equation is solved by combining boundary conditions; the boundary conditions include temperature boundary conditions, heat flow boundary conditions, convection boundary conditions, and radiation boundary conditions. The heat conduction equation is discretized into a system of linear equations. The temperature values ​​of all non-uniform grid nodes are obtained by solving the system of equations using a numerical algorithm, thus obtaining an approximate distribution of the three-dimensional temperature field. By taking the spatial derivative of the three-dimensional temperature field at the crystal growth interface, the temperature gradient distribution at the crystal growth interface is obtained. The specific steps for correcting the thermal resistance value of the non-uniform mesh nodes based on crystal defect data are as follows: Based on offline defect detection data, identify the defect-dense regions on the crystal, and locate the non-uniform grid node set corresponding to the defect-dense regions through coordinate mapping relationships; The temperature fluctuation and defect density correlation model is invoked to determine whether the defect-dense area is significantly positively correlated with temperature fluctuation; If the correlation model shows that the temperature fluctuation in the defect-dense region is significantly positively correlated with the defect density, then the corresponding grid node is marked as a priority correction node; otherwise, the corresponding node is recorded as a node to be observed. The defect influence coefficient is obtained by calculating the ratio of the defect density of the crystal region corresponding to the priority correction node to the average defect density of the whole crystal; the temperature prediction error is obtained by calculating the difference between the temperature of the priority correction node output by the thermal resistance network unit and the measured temperature; and the correlation coefficient of the crystal region corresponding to the priority correction node is used as the correlation strength coefficient between the temperature fluctuation and the defect density in the crystal region corresponding to the priority correction node. The correction amount of the thermal resistance value of the priority correction node is calculated based on the defect influence coefficient, temperature prediction error, and correlation strength coefficient of the region corresponding to the priority correction node; the thermal resistance value of the priority correction node is updated based on the correction amount of the thermal resistance value of the priority correction node.

2. The intelligent temperature control system for a crystal growth furnace as described in claim 1, characterized in that, The thermal resistance network unit is configured with a meshing strategy to divide the furnace space into non-uniform mesh nodes. The meshing strategy specifically includes: Based on the three-dimensional structure of the furnace body, a basic grid covering the entire furnace chamber is divided, and a minimum grid size is set. Establish a mapping relationship between crystal growth stages and grid density, wherein: the highest grid density is adopted in the seed crystal stage, and the node density of the seed crystal tip, melt surface, solid-liquid interface area and its neighborhood is increased; the node density of the contact area between the crystal sidewall and the melt and its adjacent area is increased in the constant diameter stage; the node density of the finishing stage is only retained in the furnace edge area and heating area, and the node density is less than the basic grid density. The current growth stage of the crystal is determined based on the crystal growth parameters, and the grid density is adjusted according to the mapping relationship between the crystal growth stage and the grid density.

3. The intelligent temperature control system for a crystal growth furnace as described in claim 2, characterized in that, The control accuracy unit is configured with a control accuracy level classification strategy to determine the temperature control accuracy level of the crystal growth interface and the temperature deviation threshold under the temperature control accuracy level. The control accuracy level classification strategy specifically includes: The total risk value is calculated based on the thermal stress factor, growth stability factor, and interface curvature factor. The temperature control accuracy level of the crystal growth interface is determined based on the magnitude of the total risk value, and a basic temperature deviation threshold is set for each temperature control accuracy level; the temperature control accuracy levels include ultra-precision mode, high-precision mode, standard mode, and relaxed mode; The temperature deviation threshold is dynamically corrected based on the basic temperature deviation threshold by adjusting the factor. The adjustment factors include crystal length factor, crystal gradient factor, and crystal defect factor; the temperature deviation threshold is obtained by multiplying the basic temperature deviation threshold by the adjustment factors.

4. The intelligent temperature control system for a crystal growth furnace as described in claim 3, characterized in that, The control module is equipped with an actuator for adjusting the temperature according to a regulation mechanism corresponding to the temperature control accuracy level. The actuator includes a multi-zone heater, a gas flow valve, and a crucible lifting mechanism. The regulation mechanism specifically includes: After receiving the ultra-precision mode command, the control module starts some multi-zone heaters to perform preset minimum power step increment adjustment. When the cumulative power adjustment of the heater reaches the preset maximum range, the gas flow valve is activated to adjust the flow rate with a preset minimum flow step, and the radial gradient is stabilized by adjusting the preset minimum amplitude of the crucible rotation speed and lifting speed. When the control module receives the high-precision mode command, it starts some multi-zone heaters to perform stepped power step adjustment. If the absolute value of the difference between the actual temperature and the target temperature value is still not less than the temperature deviation threshold after continuous adjustment of the preset number of steps, the gas flow valve is activated to adjust the flow rate in a coordinated manner with the preset basic flow rate step. At the same time, the crucible lifting mechanism adjusts the rotation speed once every fixed time to control the temperature gradient fluctuation.

5. The intelligent temperature control system for a crystal growth furnace as described in claim 4, characterized in that, The regulation mechanism also includes: When the control module receives the standard mode command, it drives all multi-zone heaters to adjust their power synchronously according to a preset ratio. At the same time, the gas flow valve adjusts in coordination with a step size greater than the preset base flow rate and less than the preset maximum flow rate. When the crystal diameter fluctuation is greater than the preset first fluctuation threshold, the crystal pulling speed of the crucible lifting mechanism is activated for adjustment. If the control module receives a relaxed mode command, it will start all heaters to adjust the preset maximum power step size, and at the same time set the gas flow valve within the preset flow range. When the crystal diameter fluctuation is greater than the preset second fluctuation threshold, the crucible lifting mechanism will be activated to adjust the preset maximum amplitude.

6. The intelligent temperature control system for a crystal growth furnace as described in claim 5, characterized in that, The optimization module is configured with an association model construction strategy for constructing an association model between defect density and temperature fluctuation; the association model construction strategy specifically includes: Based on the timeline of crystal growth, the defect detection data and temperature fluctuation data of crystals at the same growth stage in each batch are time-aligned. The crystal is divided into different spatial regions, and a coordinate mapping relationship is established between each spatial region of the crystal and a non-uniform grid of a three-dimensional temperature field. The correlation results of the data corresponding to the spatial regions are recorded and classified according to the crystal spatial regions and growth stages to form data subsets. The defect feature values ​​and temperature fluctuation feature values ​​of each subset are calculated. The defect feature values ​​include the average defect density and the maximum defect density of the corresponding spatial region. A scatter plot was drawn with the absolute value of the difference between the actual temperature and the target temperature on the horizontal axis and the average defect density on the vertical axis. The correlation coefficient between temperature fluctuation and defect density was calculated by analyzing the data distribution of the scatter plot in order to establish the mapping relationship between defect characteristic values ​​and temperature fluctuation characteristic values. Based on the mapping relationship, a temperature fluctuation and defect density correlation table is generated for each region to form a temperature fluctuation and defect density correlation model.

7. The intelligent temperature control system for a crystal growth furnace as described in claim 6, characterized in that, The data acquisition module is equipped with thermocouples and acquires the temperature data through the thermocouples; The data acquisition module is also equipped with an infrared thermal imager, and acquires the thermal image data through the infrared thermal imager; The crystal growth parameters include crystal diameter and crystal growth rate; The crystal diameter is obtained by scanning the outer diameter of the crystal with a laser beam and outputting the crystal diameter value in real time; the crystal growth rate is calculated by the pulling displacement and time difference; the crystal defect detection data includes online detection data and offline detection data.

8. A smart temperature control method for a crystal growth furnace, implemented based on a smart temperature control system for a crystal growth furnace as described in any one of claims 1-7, characterized in that, Includes the following steps: Real-time acquisition of physical parameters and crystal quality parameters within the crystal growth furnace; the crystal quality parameters include crystal growth parameters and crystal defect detection data. The furnace space of the crystal growth furnace is divided into non-uniform grid nodes, and the thermal resistance of the non-uniform grid nodes is calculated based on the physical parameters; the temperature gradient distribution of the crystal growth interface is calculated based on the thermal resistance of the non-uniform grid nodes and the boundary conditions. Based on the temperature gradient distribution and crystal growth parameters, the temperature control accuracy level and temperature deviation threshold of the crystal growth interface are dynamically determined. When the absolute value of the difference between the actual temperature of the crystal growth interface and the target temperature is greater than the temperature deviation threshold, the temperature is adjusted according to the adjustment mechanism corresponding to the temperature control accuracy level. Based on the crystal defect detection data and historical temperature fluctuation data, a correlation model between defect density and temperature fluctuation is established. After each batch of crystal growth is completed, the thermal resistance value of the non-uniform grid nodes is corrected based on the crystal defect data.

Images