Remelting, chamfering and bubble repairing process for quartz crucible

Abstract

The invention discloses a quartz crucible remelting, chamfering and bubble repairing process, and relates to the technical field of intelligent manufacturing, and the quartz crucible remelting, chamfering and bubble repairing process comprises the following steps: S1, collecting temperature field distribution, microbubble particle size and density data of fused quartz in a quartz crucible remelting stage in real time, and synchronously collecting edge contour precision and substrate stress distribution information after crucible forming, forming a whole-process detection data set; and S2, based on the whole process detection data set, dynamically matching differential process parameters of remelting, chamfering and bubble repairing processes through an intelligent cooperative regulation and control algorithm. According to the remelting, chamfering and bubble repairing process of the quartz crucible, by integrating an advanced detection technology and an intelligent regulation and control algorithm, the precision and efficiency of the quartz crucible manufacturing process are remarkably improved, multi-dimensional data such as the temperature field, microbubble distribution and base material stress of fused quartz can be collected and analyzed in real time, and the accuracy and accuracy of the fused quartz manufacturing process are improved. And dynamic matching and optimization of process parameters are realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, specifically to the remelting, chamfering, and bubble repair processes for quartz crucibles. Background Technology

[0002] Quartz crucibles are core components in high-end manufacturing fields such as photovoltaics and semiconductors. Their material uniformity, dimensional accuracy, and structural integrity directly impact the production efficiency and quality of downstream products. Remelting, chamfering, and bubble repair are critical processes in quartz crucible manufacturing. Remelting determines the density of the molten quartz and the effectiveness of microbubble control; chamfering relates to the adaptability of the crucible's edge contour and its safety; and bubble repair directly affects the crucible's structural strength and service life. As high-end manufacturing industries continuously raise their performance requirements for quartz crucibles, higher standards are being placed on the process precision, parameter coordination, and defect control capabilities of each process. Therefore, there is an urgent need to build an efficient and stable end-to-end process system.

[0003] Currently, the implementation of processes related to quartz crucibles largely relies on empirical parameter settings, with insufficient data interaction and parameter coordination between processes, making it difficult to dynamically adjust process strategies based on the real-time state of the molten quartz. In terms of data processing, the lack of systematic integration and precise analysis of multi-dimensional detection data results in weak targeting of key aspects such as temperature field uniformity control, precise microbubble removal, contour accuracy compensation, and stress release. Furthermore, existing processes have room for optimization in areas such as cross-process parameter connection and closed-loop defect management, making it difficult to achieve multi-objective synergistic optimization of bubble defect elimination, contour accuracy achievement, and substrate stress release. To address these issues, we propose a remelting, chamfering, and bubble repair process for quartz crucibles. Summary of the Invention

[0004] To address the aforementioned technical issues, this solution provides a process for remelting, chamfering, and bubble repair of quartz crucibles. This technical solution resolves the problems of existing quartz crucible-related processes relying heavily on empirical parameter settings, insufficient data interaction and parameter coordination between processes, difficulty in dynamically adjusting the process according to the real-time state of the molten quartz, lack of systematic integration and precise analysis of multi-dimensional detection data, weak targeted control of key links, and room for optimization in cross-process parameter connection and defect closed-loop management, making it difficult to achieve multi-objective collaborative optimization.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The remelting, chamfering, and bubble repair process for quartz crucibles includes the following steps: S1. Real-time acquisition of temperature field distribution, microbubble particle size and density data of molten quartz during the remelting stage of quartz crucible, and simultaneous acquisition of edge contour accuracy and substrate stress distribution information after crucible forming, forming a full-process inspection dataset. S2. Based on the full-process inspection dataset, the differentiated process parameters of remelting, chamfering and bubble repair processes are dynamically matched through an intelligent collaborative control algorithm to generate control instructions for each process parameter; S3. According to the parameter control instructions, the remelting stage is carried out by first performing graded vacuum treatment and then inert gas gradient replacement and temperature control strategy. The chamfering stage adopts substrate stress adaptive constant force grinding and real-time curvature compensation. The bubble repair stage selects high-purity quartz composite repair material that is compatible with the substrate characteristics and combines it with local precise temperature control melting and gradient annealing process to build a full-process integrated defect control solution.

[0006] Preferably, S1 includes: During the remelting process of the quartz crucible, infrared thermal imaging is used to continuously scan the molten quartz area to obtain temperature field distribution data. The scanning frequency is synchronized with the remelting heating power adjustment cycle. Simultaneously, ultrasonic testing is used to perform non-destructive testing on the interior of fused silica. By analyzing the ultrasonic echo signals, the particle size parameters of microbubbles and the number of bubbles per unit volume are extracted to form microbubble density data. After the quartz crucible is remelted and formed, a laser contour scanning method is used to perform a full-circumference scan of the crucible edge area to obtain the three-dimensional coordinate data of the edge contour and calculate the contour accuracy deviation value. X-ray diffraction stress testing was used to perform multi-point testing on the substrate of the formed crucible, and stress distribution data in different regions were collected. The testing points covered the crucible edge and the main body transition area, and a correlation mapping was formed with the edge contour accuracy data.

[0007] Preferably, S1 further includes: The collected temperature field distribution data, microbubble particle size and density data, edge contour accuracy data, and substrate stress distribution data are classified and archived. Data modules are divided according to the process dimension from remelting stage detection data to post-molding monitoring data, and each data module is associated with the corresponding detection technology type identifier. The data of each module are preprocessed, and the environmental interference noise in the temperature field data is removed by statistical filtering method. The false signal data in microbubble detection is removed by outlier judgment rules. The contour accuracy three-dimensional coordinate data is smoothed and the consistency of repeated detection data is checked for stress distribution data. Standardize the format of various types of data, convert temperature field distribution data into rasterized matrix format, convert microbubble particle size and density data into quantized numerical arrays, associate edge contour accuracy data with coordinate set under a unified coordinate system, bind substrate stress distribution data to spatial coordinate information of detection points, and unify data dimensions and units. Add timestamps and spatial location labels to all preprocessed datasets. The timestamps of the remelting stage data are synchronized with the remelting heating cycle, and the spatial location labels of the formed data correspond uniquely to the region division of the crucible. Establish a spatiotemporal correlation index. The various types of data after association are integrated into a structured dataset, which is stored in an extensible data storage format. The dataset contains basic data attributes, raw detection data, and preprocessed data results, forming a full-process detection dataset.

[0008] Preferably, S2 includes: Multi-dimensional core features are extracted from the full-process inspection dataset. Temperature field data is used to extract the temperature gradient and high temperature holding time of the melting area; microbubble data is used to extract the peak particle size distribution and density exceeding threshold; contour accuracy data is used to extract edge roundness deviation and step height; and stress distribution data is used to extract the maximum stress value and stress concentration area distribution. Based on the extracted core detection features, a correlation model of detection features, process parameters and defect types is constructed to determine the relationship between remelting parameters and bubble defects and temperature field uniformity, the relationship between chamfering parameters and contour accuracy and residual stress, the relationship between repair parameters and bubble healing effect and substrate compatibility, and to establish a multi-dimensional parameter control relationship matrix.

[0009] Preferably, S2 further includes: For the remelting process, a matching model of vacuum degree, inert gas flow rate and temperature control curve is constructed based on temperature field distribution and microbubble data. The pressure gradient and inert gas replacement rate of the staged vacuum treatment are adjusted according to the bubble particle size and density. The temperature control parameters for heating, holding and cooling are set in combination with the temperature difference gradient. For the chamfering process, a grinding parameter model is established based on the contour accuracy deviation and substrate stress distribution data. The stress value in the stress concentration area is matched with the magnitude of the grinding constant force. The arc compensation amount is calculated based on the contour three-dimensional coordinate deviation, and the grinding speed and feed rate are adjusted. For the bubble repair process, a model is constructed based on the substrate characteristic data and bubble defect parameters to determine the amount of repair material, the accuracy of local temperature control, and the annealing process. The amount of high-purity quartz composite repair material is determined according to the bubble volume. The local melting temperature and holding time are set in combination with the stress distribution of the substrate. The temperature steps and holding time of gradient annealing are set according to the stress release requirements.

[0010] Preferably, S2 further includes: A multi-objective optimization algorithm was adopted, with the bubble defect elimination rate, contour accuracy compliance rate and substrate stress release degree as optimization objectives, to perform collaborative iterative optimization of the parameters of the three processes of remelting, chamfering and bubble repair. An embedded real-time feedback adjustment mechanism is used to input dynamic detection data from each process into the optimization model, update the parameter matching results according to the process execution cycle, and dynamically correct the vacuum degree, grinding force, and repair temperature parameters. The optimized parameters are subjected to process feasibility verification. After the verification is passed, parameter control instructions for each process are generated. The process feasibility verification includes: whether the parameter values ​​are within the equipment operating threshold, whether they are within the material characteristic adaptation range, and whether they comply with the safety operation specifications. The parameter control instructions include: parameter type, value range, adjustment sequence, and execution priority.

[0011] Preferably, S3 includes: During the remelting stage, a graded vacuum treatment, inert gas gradient replacement, and temperature control strategy are implemented. During the graded vacuum treatment, the graded pressure gradient is set according to the vacuum degree parameter in the parameter control command. In the first stage, a large volume of air is evacuated and a large volume of air is removed. In the second stage, a vacuum is evacuated to remove residual gas in the micro gaps. The vacuum holding time of each stage matches the high temperature holding time parameter in the command. After vacuum treatment, inert gas gradient replacement is carried out. According to the inert gas flow parameters in the instruction, high-purity argon is first introduced to atmospheric pressure, and then the gas flow rate is gradually increased. At the same time, the gas is evenly distributed through the gas distributor. The temperature control strategy is executed synchronously. Based on the temperature difference gradient parameters in the instruction, the heating phase heats to the target temperature at a set rate, the holding phase adjusts the temperature fluctuation range according to the microbubble density data, and the cooling phase cools at a set rate.

[0012] Preferably, S3 further includes: During the chamfering stage, when performing adaptive constant force grinding of the substrate stress, the stress concentration area distribution and grinding constant force matching value in the parameter control command are read to verify the initial pressure of the grinding head. If a high stress area is detected, the constant force is dynamically adjusted according to the command rules while maintaining the grinding speed. The high stress area is defined as follows: based on the collected substrate stress distribution data, it is compared with the stress judgment threshold set in the parameter control command. When the stress value of the detected area is greater than or equal to the stress judgment threshold. When performing real-time curvature compensation, the edge profile data of the crucible is collected in real time by a laser profile scanner and compared with the allowable value of edge roundness deviation in the command. If the profile deviation is detected to exceed the allowable range, the curvature correction amount is calculated according to the compensation coefficient and the feed speed of the grinding head is adjusted synchronously.

[0013] Preferably, S3 further includes: When using high-purity quartz composite repair material in the bubble repair stage, select the substrate material according to the substrate characteristic parameters in the parameter control command, and match the material particle size with the bubble particle size. When performing localized temperature-controlled melting, according to the local melting temperature parameters in the instruction, a fiber laser is focused on the bubble area, the spot diameter is controlled to match the bubble size, and the heat is heated to the target temperature at the set rate and held for the corresponding holding time. When performing gradient annealing, after melting, the temperature steps are set according to the stress release parameters in the instruction, and the temperature is reduced in stages at the set rate. The holding time of each stage matches the repair time of the stress concentration area in the instruction.

[0014] Preferably, S3 further includes: When performing real-time data synchronization, the three stages of remelting, chamfering, and bubble repair are connected to the full-process inspection dataset. The bubble data after remelting is transferred to the chamfering stage, and the contour accuracy data after chamfering is synchronized to the repair stage. When performing parameter cross-process connection, the cooling end temperature of the remelting stage is consistent with the initial grinding temperature of the chamfering stage, and the gradient annealing end temperature of the repair stage is matched with the initial temperature of the substrate in the remelting stage. When implementing closed-loop quality control, quality inspection is carried out at the end of each stage according to the defect judgment threshold in the parameter control instruction. If the standard is not met, the result is fed back to the parameter control algorithm to regenerate the instruction.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The remelting, chamfering, and bubble repair processes for quartz crucibles proposed in this invention significantly improve the accuracy and efficiency of the quartz crucible manufacturing process by integrating advanced detection technologies and intelligent control algorithms. These processes can collect and analyze multi-dimensional data such as the temperature field of the molten quartz, microbubble distribution, and substrate stress in real time, achieving dynamic matching and optimization of process parameters. This effectively solves the problems of reliance on experience-based settings and insufficient data interaction in traditional methods. By constructing an integrated defect control scheme for the entire process, this process achieves refined operation at each stage of remelting, chamfering, and bubble repair, enhancing the material uniformity, dimensional accuracy, and structural integrity of the crucible. The cross-process parameter connection and quality closed-loop control mechanism ensure the stability and reliability of the production process, further improving the overall performance and service life of the quartz crucible, and providing high-quality basic component support for high-end manufacturing fields such as photovoltaics and semiconductors. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] In existing quartz crucible manufacturing processes, inspections during remelting, forming, and post-processing are typically conducted separately. The acquisition of temperature field, bubble distribution, profile accuracy, and stress data is asynchronous, leading to delayed process parameter adjustments, incomplete defect assessments, and difficulty in establishing a data foundation for collaborative optimization across the entire process. Furthermore, the inconsistent formats and weak spatiotemporal correlation of various inspection data fail to effectively support closed-loop process control and quality traceability. This step, by integrating infrared thermal imaging, ultrasonic testing, laser profile scanning, and X-ray diffraction stress detection, achieves simultaneous acquisition and preprocessing of multiple parameters throughout the entire process from remelting to forming. It also establishes a structured dataset with spatiotemporal correlation, providing complete and reliable data support for subsequent intelligent process control, real-time defect diagnosis, and quality control, thereby improving process consistency and the predictability of crucible product performance. (Refer to...) Figure 1 As shown, the remelting, chamfering, and bubble repair processes for the quartz crucible are implemented as follows: S1. Real-time acquisition of temperature field distribution, microbubble particle size and density data of molten quartz during the remelting stage of quartz crucible, and simultaneous acquisition of edge contour accuracy and substrate stress distribution information after crucible forming, forming a full-process inspection dataset. S1 includes: During the quartz crucible remelting process, infrared thermal imaging is used to continuously scan the molten quartz area to obtain temperature field distribution data. The specific range of the scanning frequency is set from 10Hz to 100Hz. This range is selected based on the millisecond to second response requirements of the remelting heating power adjustment cycle. The scanning frequency and the remelting heating power adjustment cycle are synchronized through a hardware synchronous trigger signal. That is, after each heating power adjustment is completed, the infrared thermal imager is immediately triggered to perform a complete scan to ensure the real-time correspondence between temperature data and power adjustment. Simultaneous ultrasonic testing was used to perform non-destructive probing of the interior of fused silica. The ultrasonic frequency was set to 1MHz to 10MHz, and the power range was 10W to 100W. A focusing probe was selected to enhance the detection resolution. The particle size parameters and the number of bubbles per unit volume of microbubbles were extracted by analyzing the ultrasonic echo signals. The specific calculation method is as follows: Based on the amplitude attenuation and time delay of the echo signal, combined with the known propagation speed of ultrasound in quartz medium, the particle size of microbubbles is inverted using an acoustic scattering model, and microbubble density data is generated by statistically analyzing the number of echo signals per unit volume. After the quartz crucible is remelted and formed, a laser contour scan is used to perform a full-circumference scan of the crucible edge area. The scanning accuracy is set to ±0.01mm, and the point cloud density reaches 100 points per square millimeter. The three-dimensional coordinate data of the edge contour is obtained, and the contour accuracy deviation value is calculated by comparing it with the preset ideal contour model. X-ray diffraction stress testing was used to perform multi-point testing on the substrate of the crucible after molding. The specific testing mode was the tilt method, the target material was a Cr target, the diffraction angle range was set to 130° to 150°, the crystal plane based on the stress calculation was the (311) crystal plane, and the elastic constant was the known value of the quartz material. The testing points covered the edge of the crucible and the main body transition area. The specific spatial distribution rule was: one testing point was set every 10 mm at the edge of the crucible and one testing point was set every 20 mm in the main body transition area. A correlation mapping relationship was formed with the edge contour accuracy data. The specific mapping method was: the spatial coordinates of each stress testing point were matched with the nearest contour accuracy testing point to establish the data correlation between the two.

[0019] S1 further includes: The collected temperature field distribution data, microbubble particle size and density data, edge contour accuracy data, and substrate stress distribution data are classified and archived. Data modules are divided according to the process dimension from remelting stage detection data to post-molding monitoring data, and each data module is associated with the corresponding detection technology type identifier. The data from each module are preprocessed. The specific algorithm used for statistical filtering is median filtering, with a window size of 3×3 pixels, to remove environmental interference noise in the temperature field data. False signal data in microbubble detection is removed through outlier judgment rules. The specific threshold setting method is as follows: with the mean of normal detection data as the center, ±3 times the standard deviation is set as the normal range, and data exceeding this range are judged as outliers. The three-dimensional coordinate data of contour accuracy are smoothed using the B-spline curve fitting algorithm. The consistency of repeated detection data is checked for stress distribution data, with a tolerance standard set at ±5MPa. Standardize the format of various data types, convert temperature field distribution data into a rasterized matrix format, set the pixel size to 0.5mm×0.5mm, and set the temperature mapping relationship to a linear mapping. Convert microbubble particle size and density data into a quantized numerical array with a two-dimensional data structure. The first dimension represents the particle size range, and the second dimension represents the density value. Associate the edge contour accuracy data with a coordinate set under a unified coordinate system. The unified coordinate system is established by taking the crucible center as the origin and establishing a Cartesian coordinate system. Bind the substrate stress distribution data to the spatial coordinate information of the detection points. Specifically, the stress value is attached as an attribute to the corresponding coordinate point, and the data dimensions and units are unified. Add timestamps and spatial location labels to all preprocessed datasets. The timestamps of the remelting stage data are synchronized with the remelting heating cycle. The synchronization logic is as follows: based on the heating power adjustment time, record the precise time of data acquisition. The spatial location labels of the data after molding correspond uniquely to the region division of the crucible. The specific correspondence rule is: divide the crucible into edge area, transition area and main area, assign a unique spatial location code to each area, and establish a spatiotemporal correlation index. The various types of data after association are integrated into a structured dataset, which is stored in an extensible data storage format such as HDF5. The dataset contains basic data attributes, raw detection data, and preprocessed data results, forming a full-process detection dataset.

[0020] S2. Based on the full-process inspection dataset, the differentiated process parameters of remelting, chamfering and bubble repair processes are dynamically matched through an intelligent collaborative control algorithm to generate control instructions for each process parameter; S2 includes: Multi-dimensional core features were extracted from the full-process inspection dataset. Among them, the temperature field data was used to extract the temperature gradient of the melting area. The calculation method was to calculate the ratio of the temperature difference to the distance between adjacent pixels using the finite difference method. The threshold for judging the high temperature holding time was set as follows: the timing started when the temperature was higher than 95% of the target temperature and ended when the temperature was lower than the threshold. The microbubble data was used to extract the particle size distribution peak. The statistical method was to divide the particle size range into multiple small intervals, count the number of bubbles in each interval, and take the peak value of the interval with the most bubbles as the particle size distribution peak. The density exceeding threshold was set to 100 bubbles per cubic centimeter. The contour accuracy data was used to extract the edge roundness deviation and step height. The stress distribution data was used to extract the maximum stress value and the distribution of stress concentration areas. Based on the extracted core detection features, a correlation model of detection features, process parameters, and defect types is constructed. Specifically, the construction method is based on a machine learning model, using a support vector machine algorithm. Historical detection data and process parameters are used as training samples to determine the relationship between remelting parameters and bubble defects and temperature field uniformity. For example, excessively high remelting temperature will lead to increased bubble density, and excessive temperature gradient will affect temperature field uniformity. The relationship between chamfering parameters and contour accuracy and residual stress is also determined, such as excessive grinding force leading to decreased contour accuracy and increased residual stress. The relationship between repair parameters and bubble healing effect and substrate compatibility is also determined, such as insufficient repair material dosage affecting bubble healing effect. A multi-dimensional parameter control relationship matrix is ​​established.

[0021] S2 further includes: For the remelting process, a matching model of vacuum degree, inert gas flow rate and temperature control curve is constructed based on temperature field distribution and microbubble data. The underlying mathematical model is as follows: using bubble density and temperature gradient as input variables, the output values ​​of vacuum degree and inert gas flow rate are determined through fuzzy control rules, and then the parameters of temperature control curve are adjusted. The pressure gradient and inert gas replacement rate of staged vacuum treatment are adjusted according to bubble particle size and density, and the temperature control parameters for heating, holding and cooling are set in combination with the temperature difference gradient. For the chamfering process, a grinding parameter model is established based on the contour accuracy deviation and substrate stress distribution data. The stress value in the stress concentration area is matched with the magnitude of the grinding constant force. The specific control algorithm is a PID control algorithm. The arc compensation amount is calculated based on the contour three-dimensional coordinate deviation, and the grinding speed and feed rate are adjusted. A diamond grinding head is selected as the grinding head type, and the grinding speed range is set from 1000 rpm to 5000 rpm. For the bubble repair process, a model is constructed based on the substrate characteristic data and bubble defect parameters to determine the amount of repair material, the accuracy of local temperature control, and the annealing process. The amount of high-purity quartz composite repair material is determined according to the bubble volume. The chemical composition of the high-purity quartz composite repair material is mainly SiO2, and the particle size distribution ranges from 1μm to 10μm. The matching rule with the bubble particle size is: the particle size of the repair material is less than 1 / 2 of the bubble particle size. The local melting temperature and holding time are set in combination with the stress distribution of the substrate. The temperature step and holding time of the gradient annealing are set according to the stress release requirements.

[0022] S2 further includes: A multi-objective optimization algorithm, specifically a genetic algorithm, is employed, with the optimization objectives being the bubble defect elimination rate, the contour accuracy achievement rate, and the substrate stress release degree. The mathematical expression of the objective function is as follows:

[0023] in , , These are the weighting coefficients for each objective. The bubble defect elimination rate, To achieve the required profile accuracy rate, To determine the stress release degree of the substrate, constraints include equipment operating parameter ranges and material performance limitations. The parameters for the three processes—remelting, chamfering, and bubble repair—are optimized through coordinated iterative processes. The specific workflow is as follows: Initialize the parameter population, calculate the fitness function value, perform selection, crossover and mutation operations to generate a new generation population, and repeat the iteration until the termination condition is met. An embedded real-time feedback adjustment mechanism is used, triggered when the detected data exceeds a preset threshold. The data input cycle is set to once per second, and the calculation formula for dynamic parameter correction is as follows:

[0024] in For parameter correction amount, For correction factor, To monitor data in real time, The preset target value; The optimized parameters are subjected to process feasibility verification. After the verification is passed, parameter control instructions for each process are generated. The process feasibility verification includes: whether the parameter values ​​are within the equipment operating threshold, such as a vacuum degree between 0.01 Pa and 100 Pa, and a temperature between 500°C and 2000°C; whether they are within the material property adaptation range, such as the matching degree between the melting point of the repair material and the melting point of the base material within ±50°C; and whether they comply with safe operation specifications, such as the gas pressure not exceeding the maximum pressure that the equipment can withstand. The parameter control instructions include: parameter type, numerical range, adjustment sequence, and execution priority. The logic for determining the adjustment sequence and execution priority is as follows: it is determined based on the sequence of the process flow and the degree of influence of each parameter on product quality, with parameters having a greater impact being adjusted first.

[0025] S3. According to the parameter control instructions, the remelting stage is carried out by first performing graded vacuum treatment and then inert gas gradient replacement and temperature control strategy. The chamfering stage adopts substrate stress adaptive constant force grinding and real-time curvature compensation. The bubble repair stage selects high-purity quartz composite repair material that is compatible with the substrate characteristics and combines it with local precise temperature control melting and gradient annealing process to build a full-process integrated defect control solution.

[0026] S3 includes: During the remelting stage, a staged vacuum treatment, inert gas gradient replacement, and temperature control strategy are implemented. During the staged vacuum treatment, the staged pressure gradient is set according to the vacuum degree parameter in the parameter control command. The specific target pressure ranges for the two vacuum stages are as follows: The first stage involves evacuating to below 10 Pa to expel a large volume of air. The second stage involves evacuating to below 0.1 Pa to remove residual gas from micro-gaps. The specific settings for the vacuum holding time for each stage are based on calculations using high-temperature holding time parameters and gas diffusion rates to ensure complete gas expulsion. After vacuum treatment, an inert gas gradient replacement is implemented. The purity of the high-purity argon gas is required to reach 99.999%. According to the inert gas flow rate parameters in the instructions, high-purity argon gas is first introduced to atmospheric pressure, and then the gas flow rate is gradually increased. The specific change curve is a linear increase. The gas distributor has a porous structure and is evenly distributed around the crucible. The temperature control strategy is executed synchronously. Based on the temperature difference gradient parameter in the instruction, the heating stage heats to the target temperature at a set rate, which is set to 10℃ / min to 50℃ / min. During the heat preservation stage, the temperature fluctuation range is adjusted according to the microbubble density data. The control method is to adjust the heating power through a PID controller. During the cooling stage, the cooling is performed at a set rate, which is set to 5℃ / min to 20℃ / min.

[0027] S3 further includes: When performing adaptive constant force grinding of the substrate during the chamfering stage, the exact value of the stress judgment threshold is set to 200MPa. The determination method is based on a combination of material strength theory and experimental verification. The initial pressure setting rule of the grinding head is: determined according to the initial stress detection data and the preset grinding force and stress relationship curve. If a high stress area is detected, the constant force is dynamically adjusted according to the instruction rules and the grinding speed is maintained. The specific control algorithm is a fuzzy control algorithm. The high stress area is defined as: based on the collected substrate stress distribution data, it is compared with the stress judgment threshold set in the parameter control instruction. When the stress value of the detected area is greater than or equal to the stress judgment threshold; When performing real-time arc compensation, the specific value of the allowable contour deviation is set to ±0.05mm. The calculation model of the compensation coefficient is as follows: a linear model is established based on the relationship between the contour deviation value and the feed speed of the grinding head. The edge contour data of the crucible is collected in real time by a laser contour scanner and compared with the allowable value of edge roundness deviation in the command. If the contour deviation is detected to exceed the allowable range, the arc correction amount is calculated according to the compensation coefficient and the feed speed of the grinding head is adjusted synchronously.

[0028] S3 further includes: When using high-purity quartz composite repair material in the bubble repair stage, select the substrate material according to the substrate characteristic parameters in the parameter control command, and match the material particle size with the bubble particle size. When performing localized temperature-controlled melting, the wavelength of the fiber laser is set to 1064nm and the power is adjustable from 10W to 500W, according to the localized melting temperature parameters in the instruction. The method of adjusting and controlling the spot diameter is to adjust it through the focusing lens to match the spot diameter with the bubble size. The laser is heated to the target temperature at the set rate and held for the corresponding holding time. The heating rate is set to 50℃ / min to 200℃ / min and the target temperature is set to 1500℃ to 1800℃. When performing gradient annealing, after melting, the temperature step is set according to the stress release parameter in the instruction. The number of temperature steps is set to 3 to 5. The temperature is reduced in stages at the set rate. The cooling rate of each stage is set to 10℃ / min to 50℃ / min. The holding time of each stage matches the stress concentration area repair time in the instruction. The specific setting method is: calculated according to the stress release rate and material properties.

[0029] S3 further includes: When performing real-time data synchronization, the three stages of remelting, chamfering, and bubble repair are connected to the full-process inspection dataset. The data interface protocol used for real-time data synchronization is TCP / IP protocol, and the communication method is wired Ethernet communication. The bubble data after remelting is transmitted to the chamfering stage, and the contour accuracy data after chamfering is synchronized to the repair stage. When connecting parameters across processes, the cooling endpoint temperature of the remelting stage should be consistent with the initial grinding temperature of the chamfering stage. The specific matching criterion is that the temperature difference between the two should not exceed 50℃. The gradient annealing endpoint temperature of the repair stage should be matched with the initial substrate temperature of the remelting stage. The specific matching criterion is that the temperature difference between the two should not exceed 100℃. When implementing closed-loop quality control, quality inspection is conducted at the end of each stage according to the defect judgment threshold in the parameter control instruction. The specific standards for the defect judgment threshold in different processes are as follows: the bubble density in the remelting stage does not exceed 50 per cubic centimeter, the contour accuracy deviation in the chamfering stage does not exceed ±0.02mm, and the bubble healing rate in the repair stage reaches more than 95%. If the standard is not met, the result is fed back to the parameter control algorithm to regenerate the instruction. The trigger condition for regenerating the instruction after feedback is that the test data fails to meet the standard three times in a row. The process is as follows: input the non-compliant data into the parameter control algorithm, recalculate the optimized parameters, and generate a new control instruction.

[0030] In existing segmented quartz crucible processes, remelting, chamfering, and repair are typically performed independently, lacking parameter coordination and quality feedback loops between processes. This can easily lead to derivative defects such as incomplete vacuum treatment, secondary stress caused by grinding, and poor compatibility between repair materials and the substrate, resulting in insufficient overall process continuity and defect control. This step transforms parameter commands into execution actions with time-coordinated and data-driven linkage. In the remelting stage, graded vacuum and gradient displacement are used to enhance degassing. In the chamfering stage, stress-adaptive grinding and real-time curvature compensation are introduced to suppress crack formation. In the repair stage, in-situ defect repair is achieved through material adaptation and precise local temperature control. Relying on full-process data synchronization, parameter coordination, and quality closed-loop management, an integrated defect control scheme is constructed, significantly improving the coordination of process execution at each stage, the thoroughness of defect repair, and the overall quality stability of the process. This represents a leap from single-point processing to systematic management.

[0031] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A process for re-melting, chamfering and bubble repair of quartz crucibles, characterized in that, The method comprises the following steps: S1, real-time acquisition of temperature field distribution, micro-bubble particle size and density data of molten quartz in the quartz crucible remelting stage, synchronous acquisition of edge profile accuracy and substrate stress distribution information after the crucible is formed, and formation of a full-process detection data set; S2, based on the full-process detection data set, dynamic matching of differentiated process parameters of the remelting, chamfering and bubble repairing processes is realized through an intelligent collaborative control algorithm, and parameter control instructions of each process are generated; S3, according to the parameter control instructions, the remelting stage adopts the mode of first-stage vacuum treatment and then inert gas gradient replacement, and the temperature control strategy, the chamfering stage adopts the mode of substrate stress self-adaptive constant force grinding and real-time compensation of arc, the bubble repairing stage selects a high-purity quartz composite repairing material suitable for the substrate characteristics and combines local accurate temperature control melting and gradient annealing process, and a full-process integrated defect control scheme is constructed.

2. The process for remelting, chamfering and bubble repair of quartz crucibles according to claim 1, characterized in that, The S1 comprises: During the quartz crucible remelting process, the infrared thermal imaging detection is used to continuously scan the molten quartz area to obtain the temperature field distribution data, and the scanning frequency is kept synchronous with the remelting heating power adjustment period; The ultrasonic detection is used to non-destructively detect the inside of the molten quartz, the particle size parameters and the number of bubbles per unit volume of micro-bubbles are extracted through ultrasonic echo signal analysis, and the micro-bubble density data are formed; After the quartz crucible is remelted and formed, the laser profile scanning is used to scan the edge area of the crucible all around to obtain the three-dimensional coordinate data of the edge profile, and the profile accuracy deviation value is calculated; The X-ray diffraction stress detection is used to detect the substrate of the formed crucible at multiple points to collect the stress distribution data of different areas, and the detection points cover the edge and the transition area of the main body of the crucible, and are associated with the edge profile accuracy data to form a mapping.

3. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 2, characterized in that, The S1 further comprises: The collected temperature field distribution data, micro-bubble particle size and density data, edge profile accuracy data and substrate stress distribution data are classified and archived, the data modules are divided according to the process dimension from the remelting stage detection data to the post-forming monitoring data, and each data module is associated with the corresponding detection technology type identifier; The pre-processing is performed on the data in each module, the environmental interference noise in the temperature field data is removed by statistical filtering method, the false signal data in the micro-bubble detection is removed by abnormal value judgment rule, the three-dimensional coordinate data of the profile accuracy are smoothed, and the consistency verification of the repeated detection data is performed on the stress distribution data; The format standards of all types of data are unified, the temperature field distribution data are converted into a grid matrix format, the micro-bubble particle size and density data are converted into a quantized numerical array, the edge profile accuracy data are associated with a coordinate set in a unified coordinate system, the substrate stress distribution data are bound with the spatial coordinate information of the detection points, and the data dimension and unit are unified; Time stamps and spatial position labels are added to all pre-processed data sets, the time stamps of the remelting stage data are synchronous with the remelting heating period, the spatial position labels of the post-forming data are uniquely corresponding to the area division of the crucible, and a time and space association index is established. The associated data of various types is integrated into a structured data set, and is stored in an extensible data storage format, the data set including basic data properties, original detection data and pre-processed data results, forming a full-process detection data set.

4. The process for remelting, chamfering and bubble repair of quartz crucibles according to claim 3, characterized in that, The S2 includes: From the full-process detection data set, multi-dimensional core features are extracted, wherein the temperature field data extracts the melting area temperature gradient and the high temperature retention time, the micro-bubble data extracts the particle size distribution peak value and the density threshold value, the contour accuracy data extracts the edge roundness deviation and the step height, and the stress distribution data extracts the maximum stress value and the stress concentration area distribution; Based on the extracted core detection features, an association model of detection features, process parameters and defect types is constructed to determine the relationship between the remelting parameters and the bubble defects and the temperature field uniformity, the relationship between the chamfer parameters and the contour accuracy and the residual stress, and the relationship between the repair parameters and the bubble healing effect and the substrate compatibility, and a multi-dimensional parameter regulation relationship matrix is established.

5. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 4, characterized in that, The S2 also includes: For the remelting process, based on the temperature field distribution and the micro-bubble data, a matching model of vacuum degree, inert gas flow and temperature control curve is constructed, the pressure gradient of the staged vacuum treatment and the inert gas replacement rate are adjusted according to the bubble particle size and density, and the temperature control parameters of heating, holding and cooling are set in combination with the temperature difference gradient; For the chamfering process, a grinding parameter model is established according to the contour accuracy deviation and the substrate stress distribution data, the stress value of the stress concentration area is matched with the constant grinding force, the radian compensation amount is calculated according to the contour three-dimensional coordinate deviation, and the grinding speed and the feed rate are adjusted; For the bubble repair process, based on the substrate characteristic data and the bubble defect parameters, a model of repair material dosage, local temperature control accuracy and annealing process is constructed, the dosage of high-purity quartz composite repair material is determined according to the bubble volume, the local melting temperature and the holding time are set in combination with the substrate stress distribution, and the temperature ladder and the holding time of gradient annealing are set according to the stress release requirement.

6. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 5, characterized in that, The S2 also includes: A multi-objective optimization algorithm is used to optimize the parameters of the three processes of remelting, chamfering and bubble repair in a collaborative and iterative manner, with the bubble defect elimination rate, the contour accuracy compliance rate and the substrate stress release degree as the optimization objectives; A real-time feedback adjustment mechanism is embedded, the dynamic detection data of each process is input into the optimization model, the parameter matching results are updated according to the process execution cycle, and the vacuum degree, the grinding force and the repair temperature parameters are dynamically corrected; The optimized parameters are subjected to process feasibility verification, and parameter regulation instructions are generated after the verification is passed; the process feasibility verification includes whether the parameter value is within the equipment operation threshold, whether it is within the material characteristic adaptation range and whether it conforms to the safety operation specification, and the parameter regulation instructions include the parameter type, the numerical range, the adjustment timing and the execution priority.

7. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 6, characterized in that, The S3 includes: In the remelting stage, the staged vacuum treatment, inert gas gradient replacement and temperature control strategy are executed, in the staged vacuum treatment, the staged pressure gradient is set according to the vacuum degree parameter in the parameter regulation instruction, the first stage is vacuumized to remove large volume air, the second stage is vacuumized to remove micro-gap residual gas, and the holding time of each stage is matched with the high temperature holding time parameter in the instruction; After the vacuum treatment is completed, an inert gas gradient replacement is implemented, high-purity argon is first introduced to normal pressure according to the inert gas flow parameter in the instruction, and then the gas flow is gradually increased, and the gas is uniformly distributed through the gas distributor; The temperature control strategy is executed synchronously, the temperature is raised at a set rate to the target temperature according to the temperature difference gradient parameter in the instruction in the heating stage, the temperature fluctuation range is adjusted according to the micro-bubble density data in the holding stage, and the temperature is cooled at a set rate in the cooling stage.

8. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 7, characterized in that, The S3 further comprises: When the substrate stress self-adaptive constant force grinding is performed in the chamfering stage, the stress concentration region distribution and the grinding constant force matching value in the parameter control instruction are read, the initial pressure of the grinding head is determined, and if the high stress region is detected, the constant force size is dynamically adjusted according to the instruction rule and the grinding speed is maintained; the high stress region is: based on the collected substrate stress distribution data, compared with the stress determination threshold set in the parameter control instruction, when the stress value of the detected region is greater than or equal to the stress determination threshold; When the real-time compensation of the arc is performed, the edge profile data of the crucible is collected in real time through the laser profile scanner, compared with the edge roundness deviation allowable value in the instruction, and if the profile deviation exceeds the allowable range, the arc correction amount is calculated according to the compensation coefficient and the grinding head feed speed is adjusted synchronously.

9. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 8, characterized in that, The S3 further comprises: When the high-purity quartz composite repair material is selected in the bubble repair stage, the substrate material is selected according to the substrate characteristic parameters in the parameter control instruction, and the material particle size matches the bubble particle size; When local temperature control melting is performed, the fiber laser is focused on the bubble area according to the local melting temperature parameter in the instruction, the spot diameter is controlled to match the bubble size, the target temperature is heated at a set rate and the corresponding holding time is maintained; When the gradient annealing process is performed, the temperature steps are set according to the stress release degree parameter in the instruction after melting is completed, the temperature is lowered at a set rate in stages, and the holding time in each stage matches the stress concentration region repair time in the instruction.

10. The remelting, chamfering and bubble repair process for quartz crucibles according to claim 9, characterized in that, The S3 further comprises: When real-time synchronization of data is performed, the remelting, chamfering and bubble repair stages are connected to the full-process detection data set, the bubble data after remelting is transmitted to the chamfering stage, and the profile accuracy data after chamfering is transmitted to the repair stage; When the parameters are connected across processes, the final temperature of the remelting stage and the initial grinding temperature of the chamfering stage are consistent, and the final temperature of the gradient annealing stage and the initial temperature of the remelting stage are matched; When the quality closed-loop control is performed, the quality is detected according to the defect determination threshold in the parameter control instruction after each stage is completed, and if it does not meet the standard, the result is fed back to the parameter control algorithm to regenerate the instruction.

Images