Digital production supervision method and system

Abstract

The invention provides a digital production supervision method and system, belongs to the field of computers, and aims to solve the problems of high rework rate and low efficiency caused by the fact that optical inspection or chemical analysis often depends on cleaned quartz pieces due to incomplete cleaning of cracks and residual metal and the like and the optimal correction opportunity is missed once the problems are found. The method comprises the steps that in the stage of cleaning a hot-worked quartz workpiece, a cleaning formula is automatically matched based on information of the quartz workpiece, and the cleaning formula comprises a cleaning solution ratio and a cleaning time-temperature control curve; state change parameters of cleaning liquid in the cleaning tank are obtained in real time, wherein the state parameters of the cleaning liquid comprise conductivity and oxidation-reduction potential; and evaluating the quality defect of the quartz workpiece based on the state change parameter of the cleaning fluid.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a digital production supervision method and system. BACKGROUND

[0002] In high-end manufacturing industries such as semiconductors, LEDs, and photovoltaics, quartz materials are widely used in the manufacture of key components such as reactor liners, process boats, and mask supports due to their excellent thermal stability, chemical inertness, and optical transparency. These quartz pieces often undergo multiple processes such as cold working, hot forming, and welding, and may have residual particles, metal ions, and organic matter on the surface. If these contaminants are not completely removed, they will directly affect the stability and yield of the device in high-clean processes.

[0003] Currently, the mainstream cleaning process uses a multi-stage process of ultrasonic cleaning, pickling, and ultrapure water rinsing. However, for cracks, residual metal, and incomplete cleaning, optical inspection or chemical analysis of the cleaned quartz piece is often relied on. Once a problem is found, the best correction opportunity has been missed, resulting in high rework rates and low efficiency. SUMMARY

[0004] The embodiments of the present application provide a digital production supervision method and system, which can solve the problem that for cracks, residual metal, and incomplete cleaning, optical inspection or chemical analysis of the cleaned quartz piece is often relied on. Once a problem is found, the best correction opportunity has been missed, resulting in high rework rates and low efficiency.

[0005] The first aspect of the embodiments of the present application provides a digital production supervision method, comprising: In the stage of cleaning the quartz processing piece after hot working, automatically matching a cleaning formula based on the quartz processing piece information, the cleaning formula including a washing liquid ratio and a cleaning time-temperature control curve; Real-time acquisition of cleaning liquid state change parameters in the cleaning tank, the cleaning liquid state parameters including conductivity and redox potential; Evaluation of the quality defects of the quartz processing piece based on the cleaning liquid state change parameters.

[0006] Optionally, the evaluation of the quality defects of the quartz processing piece based on the cleaning liquid state change parameters comprises: Evaluation of the abnormal ion release of the quartz processing piece based on the cleaning liquid state change parameters; Evaluation of the quality defects of the quartz processing piece according to the abnormal ion release.

[0007] Optionally, it further comprises: Real-time recording of the number and size distribution of particles stripped in the fluid per unit time as a dynamic response curve of the cleanliness of the device surface; The method further comprises: The method further comprises:

[0008] Optionally, the method further comprises: In the rinsing and heating zone, an infrared thermal imager is used to capture the thermal distribution of the quartz workpiece during the heating process; In the rinsing and heating zone, a polarized light is used to irradiate and collect an interference fringe pattern; The method further comprises:

[0009] The method further comprises: Based on the thermal distribution of the quartz workpiece during the heating process, a thermal retention area is determined; Based on the interference fringe pattern, a fringe distortion area is determined; The thermal retention area and the fringe distortion area are detected using an image fusion algorithm to determine whether they are stress concentration areas through a classifier.

[0010] Optionally, the method further comprises: In the drying stage after cleaning, the wake information of the drying gas flow after flowing through the quartz workpiece is obtained, and the drying gas flow is a constant-temperature and constant-pressure gas flow; Based on the wake information, a wake flow field is determined to evaluate the quality defects of the quartz workpiece according to the wake flow field, and the quality defects include a liquid film thickness difference curve and / or a surface foreign object adhesion defect of the quartz workpiece.

[0011] Optionally, the cleaning liquid state change parameters further include pH and surface tension, and the method further comprises: Based on the real-time obtained cleaning liquid state change parameters in the cleaning tank, a current cleaning liquid pollution level is evaluated and determined; According to the current cleaning liquid pollution level, a cleaning intensity parameter is adjusted, and the cleaning intensity parameter includes at least one of ultrasonic frequency, cleaning time, or cleaning temperature.

[0012] The second aspect of the embodiments of the present application provides a digital production supervision system, comprising: A matching unit is configured to automatically match a cleaning formula based on the quartz workpiece information during the cleaning stage of the hot-processed quartz workpiece, and the cleaning formula includes a cleaning liquid ratio and a cleaning time-temperature control curve. A collection unit is configured to acquire a cleaning liquid state change parameter in real time, the cleaning liquid state parameter including conductivity and redox potential; An evaluation unit is configured to evaluate the quality defect of the quartz workpiece based on the cleaning liquid state change parameter.

[0013] The third aspect of the embodiment of the present application provides an electronic device, including a memory and a processor, and the processor is configured to execute the computer program stored in the memory to realize the steps of the digital production supervision method.

[0014] The fourth aspect of the embodiment of the present application provides a computer readable storage medium, and a computer program is stored in the computer readable storage medium, and the computer program is executed by a processor to realize the steps of the digital production supervision method.

[0015] In summary, the digital production supervision method provided by the embodiment of the present application realizes the transition from post-detection to process judgment, greatly reduces the rework cost of problems found after cleaning is completed, intelligently matches the cleaning intensity of each batch of products according to the pollution difficulty, avoids the waste of time and reagents caused by heavy cleaning of light pollution, dynamically monitors the pollution dissolution in the cleaning process, can adjust the cleaning parameters in time, improves the complete removal rate of pollutants, and improves the sensitivity to hidden defects such as micro-cracks and embedded impurities by using the cleaning liquid state change signal as an implicit quality index. Not only does it make the quartz cleaning process get rid of the traditional passive mode of relying on end detection and realize the feed-forward defect recognition based on process characteristics, but also builds a precise matching logic between products and the cleaning process, greatly reduces the rework rate and cleaning liquid use redundancy. It can improve the yield and stability of quartz pieces in high-cleanliness processes, provide traceable and evolvable digital supervision capabilities for production lines, and significantly enhance the intelligent level and controllability of high-end manufacturing processes.

[0016] Correspondingly, the system, the electronic device and the computer readable storage medium provided by the embodiment of the present application also have the above technical effects. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A possible digital production supervision method provided by the embodiment of the present application is shown in the flowchart; Figure 2 A possible digital production supervision system provided by the embodiment of the present application is shown in the schematic structural block diagram; Figure 3 A possible hardware structure schematic diagram of a digital production supervision system provided by an embodiment of the present application; Figure 4 A possible schematic structure block diagram of an electronic device provided by an embodiment of the present application; Figure 5 A possible schematic structure block diagram of a computer readable storage medium provided by an embodiment of the present application. DETAILED DESCRIPTION

[0018] The embodiment of the present application provides a digital production supervision method and related equipment, which can solve the problems of cracks, residual metal and incomplete cleaning, and often relies on optical inspection or chemical analysis on the cleaned quartz piece. Once the problem is found, the best correction opportunity has been missed, resulting in high rework rate and low efficiency.

[0019] The terms "first", "second", "third", "fourth" and the like (if any) in the description, claims and above-mentioned drawings of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices. The technical solutions in the embodiments of the present application will be described in detail below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments.

[0020] Please refer to Figure 1 A flowchart of a digital production supervision method provided by an embodiment of the present application, which can specifically include: S110-S130.

[0021] S110, in the stage of cleaning the quartz processing piece after heat processing, automatically matching a cleaning formula based on the quartz processing piece information, the cleaning formula including a washing liquid ratio and a cleaning time-temperature control curve.

[0022] S120, real-time acquisition of a cleaning liquid state change parameter in a cleaning tank, the cleaning liquid state parameter including conductivity and redox potential.

[0023] S130, evaluating the quality defect of the quartz processing piece based on the cleaning liquid state change parameter.

[0024] It can be understood that after experiencing thermal processing (welding, annealing, etc.), the surface of the quartz piece can be left with different types of contaminants (particles, organic matter, metal ions). The cleaning process is essentially a synergistic process controlled by temperature, chemical reaction rate, interfacial diffusion rate, and cleaning liquid composition. The removal of different contaminants has a specific cleaning window, the boundaries of which depend on the activity state of the cleaning liquid (such as pH, conductivity, ORP, etc. indicators). During the cleaning process, by continuously monitoring the conductivity and oxidation-reduction potential (ORP) of the cleaning liquid, the degree of cleaning reaction and the state of contaminant dissolution can be indirectly reflected. For example: when metal ions are dissolved, the conductivity will rise; when organic matter is oxidized and decomposed, the ORP will undergo a transition. These trends can be predicted in advance by modeling the data and comparing it with the experience database to predict the type and degree of contamination of the quartz piece. Based on the source process path of the quartz piece (such as heat treatment temperature, atmosphere, use site, etc.), structural features, historical defect records, etc., the matching formula in the cleaning database is automatically called. The formula includes the concentration of the chemical cleaning liquid, the cleaning sequence, the cleaning time of each stage, the temperature curve, etc., to achieve precise decontamination. The above method aims to solve the problem of lagging quality defect discovery and uncontrollable process in the cleaning stage of quartz pieces in high-end manufacturing industries. By introducing dynamic parameter monitoring and formula matching mechanism in the cleaning process, the method realizes the feedforward judgment of quality defects and the digital control of the production process, avoiding the high rework rate and low efficiency caused by the traditional cleaning and then inspection mode.

[0025] For example, after the quartz piece completes thermal processing (such as welding, hot forming, or annealing process), the system automatically reads the complete process history information of the batch of quartz pieces through the MES platform, including heat treatment temperature, time, environmental atmosphere type (such as , , argon, etc.), historical defect labels, use site (such as mask frame, process boat, reactor substrate), etc. Based on these parameters, the system automatically matches the most suitable cleaning formula by comparing with the pre-set cleaning process database, which includes cleaning agent types (such as HF, , , ultrapure water, etc.), concentration ratio, cleaning step sequence, cleaning time setting, temperature-time change curve, etc., to ensure that different types of contaminants can be effectively removed within the optimal window.

[0026] For example, during the cleaning process, the system uses online sensing modules deployed in the cleaning tank to obtain real-time cleaning fluid parameters, primarily conductivity and ORP (redox potential). These parameters are collected every 30 seconds to form a status curve for the entire cleaning process. Additionally, auxiliary instruments such as turbidity meters and ultraviolet absorption spectrometers can be integrated as needed to enhance the sensitivity of monitoring organic contamination and colloidal impurities. This real-time data is compared and analyzed against a standard state model constructed from historical cleaning data. If conductivity shows an abnormal increase (e.g., significant metal ion dissolution), the system can automatically infer potential solder overheating or surface coating residue issues. If ORP remains below a set lower limit for a prolonged period, it indicates depletion of the oxidant activity, requiring an increase in oxidant concentration or a prolonged cleaning phase. If the ORP curves show a hysteresis, this may indicate cracks on the device surface or a dense contamination layer that is hindering the cleaning reaction.

[0027] For example, based on these real-time judgments, the system will automatically adjust the cleaning process, such as appropriately extending the pickling time, switching to an alternate recipe, inserting an intermediate rinse, or issuing a warning that subsequent non-destructive testing (such as surface scattering intensity analysis or local XPS testing) is needed to avoid device scrapping due to incomplete cleaning. Furthermore, after the cleaning task is completed, the system correlates the cleaning fluid state curve with the final quartz component cleanliness test data (such as particle counts, surface residual metal concentration analysis, and microcrack detection). The results are fed back to the cleaning recipe database, and machine learning algorithms are used to continuously optimize the multi-dimensional mapping relationship from process path to cleaning response to detection tags.

[0028] In summary, the digital production monitoring method provided by the above-mentioned embodiments of the present application automatically matches a cleaning recipe based on quartz workpiece information during the cleaning phase of heat-processed quartz workpieces. The cleaning recipe includes a cleaning solution ratio and a cleaning time-temperature control curve. The cleaning solution state change parameters in the cleaning tank, including conductivity and redox potential, are acquired in real time. Quality defects of the quartz workpiece are assessed based on the cleaning solution state change parameters. This method achieves a shift from post-detection to process judgment, significantly reducing rework costs for problems discovered only after cleaning is complete. The cleaning intensity of each batch of products is intelligently matched based on the severity of the contamination, avoiding the waste of time and reagents by washing lightly contaminated products with heavy cleaning. Dynamic monitoring of contaminant dissolution during the cleaning process allows for timely adjustment of cleaning parameters to improve the complete removal rate of contaminants. The cleaning solution state change signal serves as an implicit quality indicator, enhancing sensitivity to hidden defects such as microcracks and embedded impurities. This not only enables the quartz cleaning process to break away from the traditional passive mode of relying on end-of-line inspection and implements feedforward defect recognition based on process characteristics, but also establishes precise matching logic between products and cleaning processes, significantly reducing rework rates and redundant cleaning solution usage. It can improve the yield and stability of quartz parts in high-clean processes, while providing traceable and evolvable digital supervision capabilities for the production line, significantly enhancing the intelligence and controllability of high-end manufacturing processes.

[0029] In some examples, evaluating the quality defects of the quartz workpiece based on the cleaning fluid state change parameter includes: evaluating abnormal ion release of the quartz workpiece based on the cleaning fluid state change parameter; The quality defects of the quartz workpiece are evaluated based on the abnormal ion release.

[0030] For example, during the cleaning process, the system continuously obtains the state change parameters of the cleaning liquid, especially the conductivity and oxidation-reduction potential (ORP). Conductivity can reflect the overall concentration of ions in the cleaning liquid, especially the dynamic changes of metal ions, sodium ions, aluminum ions, iron ions, etc. dissolved from the quartz surface; while ORP reflects the oxidants in the cleaning liquid (such as 、 ) and reducing pollutants (such as organic matter, surface carbon residue, etc.).

[0031] For example, the system compares the collected conductivity time series with the "standard release curve" in the database. The standard curve is the parameter trend recorded in the cleaning process of a large number of normal quartz pieces, which has obvious segment characteristics: for example, the conductivity slowly rises in the initial stage, and then rises sharply at a certain time point, indicating that the surface ions begin to release in large quantities, and then tends to be stable. If the following abnormal ion release occurs during the cleaning process of the current quartz processing piece, a warning will be triggered, for example, if the conductivity abnormally rises in the early stage (for example, it exceeds the historical median value within the first 30 seconds), it may indicate that there are residual metals on the surface caused by welding splashes; if the conductivity continues to slowly grow and is unstable within a preset time period, it indicates that the pollutants are not uniformly released, and there may be impurities embedded in the surface microcracks; if the ORP continuously stays in the low value interval and abnormally fluctuates, it indicates that the oxidizing agent is consumed in large quantities but the pollutants are not effectively removed, indicating that there is heavy organic impurity or abnormal surface tension. After the determination of the above ion release conditions is completed, the system further combines the ion release rate and peak value amplitude, and uses rules or trained models (such as clustering analysis, SVM classification, etc.) to divide the current device into four quality level labels: no obvious defects, surface pollution is too heavy, may have microcracks, and needs subsequent physical detection. For example, the rapid release of high-concentration metal ions is often related to overheating during the previous process, tool cross-contamination, etc.; and the slow and continuous growth of the ion release rate may be related to internal doping, structural fatigue or local stress concentration of the quartz. Finally, the system can issue instructions based on the ion release evaluation results before the cleaning is completed, such as automatically extending the current cleaning period, replacing the cleaning liquid or adding oxidizing agent, inserting optical surface scanning, prompting the next detection process to enhance, etc., to realize the feedforward control of potential quality problems and avoid irreversible defects or batch rework. This process can also update the evaluation data to the MES system and the process optimization model to form a closed-loop learning mechanism for the cleaning process path. Through this method of inferring quality defects based on cleaning liquid state parameters, the intelligentization and predictive control capability of the cleaning process are significantly improved, and the hidden quality risk in quartz piece production is reduced.

[0032] In some examples, further comprising: Real-time recording of the number and size distribution of particles released in the fluid per unit time as a dynamic response curve of the cleanliness of the device surface; The evaluation of the quality defect of the quartz processing piece based on the cleaning liquid state change parameter comprises: Using the particle release mode to analyze whether the quartz processing piece has hidden cracks caused by edge stress concentration.

[0033] For example, to further enhance the early detection of microscopic defects in quartz components, the system also includes the following steps: Real-time recording of the number and size distribution of particles released from the cleaning fluid per unit time, using this as a dynamic response curve for the surface cleanliness of the quartz component. Specifically, the system integrates a high-sensitivity particle counting module (such as an online particle monitoring device based on laser scattering or light obstruction) at the cleaning tank outlet or in the circulation line. This module records the concentration and size distribution of released particles in the fluid at a frequency of seconds or minutes (e.g., statistics for segments larger than 0.1–1μm, 1–5μm, and 5–10μm), generating a complete time series of particle release during the cleaning process. Based on this particle release data, the system further analyzes its variation patterns to infer possible microstructural anomalies in the quartz component. A key identification mechanism involves using the particle release pattern to analyze whether the quartz component has hidden cracks caused by edge stress accumulation. This analysis is based on the following principles: Under the normal surface release mode, if the quartz surface has no obvious defects, the particle release rate may be slightly elevated in the early stages of cleaning (to remove surface dust and processing residues), then rapidly decrease and stabilize. Particle sizes are concentrated in the sub-1μm range, with a uniform distribution and no sudden spikes. Under the stress crack release mode, if edge stress is accumulated (e.g., due to insufficient internal stress release from heat treatment or machining), this region is susceptible to latent crack propagation under high temperatures or chemical reactions during the cleaning process, leading to the release of characteristic particles. This release pattern manifests as a sudden jump in particle concentration at a specific moment (e.g., 20 minutes after pickling). The particle size distribution shows an abnormally high proportion of medium- to large-sized particles (e.g., 3–10μm). The duration of the particle release peak is highly correlated with changes in cleaning fluid parameters (e.g., increased conductivity). Similar curves are statistically reproducible across multiple devices, constituting a defect signature sample. Leveraging these characteristics, the system uses pre-trained machine learning models (such as pattern matching based on dynamic time warping (DTW) or an LSTM sequence analysis network) to determine in real time whether the particle release curve conforms to the "stress crack" signature model. If so, the device is identified as potentially susceptible to microcracks. Automatically implementing measures can include prompting for additional high-magnification visual inspection or stress testing (such as polarization interferometry), risk-flagging the batch, or inserting neutralization and secondary rinses to prevent crack expansion. By combining particle release information with cleaning fluid state parameters such as conductivity and ORP, the system not only quantifies the degree of contamination removal but also visualizes the invisible defect of crack stress release as a fluid dynamics signal. This significantly enhances the feedforward control capabilities of quartz cleaning quality monitoring and the accuracy of identifying hidden structural anomalies, further ensuring its stability and reliability in high-cleanliness processes.

[0034] In some examples, this also includes: In the rinsing and heating zone, the thermal distribution of the quartz workpiece during the heating process is captured by an infrared thermal imager; In the rinsing and heating zone, interference fringe patterns are collected by irradiating with polarized light; The quality defects of the quartz workpiece are evaluated based on the thermal distribution and interference fringe pattern of the quartz workpiece during the heating process.

[0035] For example, to further enhance the nondestructive assessment of the structural integrity and microscopic defects of quartz workpieces, the following monitoring steps are included: capturing the thermal distribution of the quartz workpiece during the heating process with an infrared thermal imager in the rinsing and heating zone, and obtaining interference fringe patterns through polarized light illumination. The quality defects of the quartz workpiece are then assessed based on these two combined methods. This method, through dual-channel detection using thermal field response and optical interference, enables visualization and quantitative analysis of internal stress, cracks, and material inhomogeneities.

[0036] For example, during the rinse temperature ramp-up stage of the cleaning process (usually a high-purity water or neutralizing liquid rinse is performed after cleaning, while the temperature gradually increases to 60-90°C), the system sets up an infrared thermal imaging device to capture the surface thermal distribution of each quartz workpiece. Since quartz has good infrared transmittance, its surface and shallow layer thermal response can be effectively observed in a reasonable waveband (such as 8-14 pm). The principle of this step is that the discontinuity of heat conduction and the change of heat capacity in the defect area will cause local temperature rise anomalies or heat conduction hysteresis. For example, if the quartz piece has microcracks, inclusions or stress concentration, these parts will show obvious temperature rise low or temperature rise hysteresis dark spots on the thermal distribution map; the local edge area shows asymmetric heat diffusion characteristics; the temperature rise curve slope is significantly different from that of the normal area. At the same time, in the same temperature ramp-up section, the quartz workpiece is irradiated by a polarized light source at an oblique incidence, and the interference fringe pattern (such as equal inclination interference, equal thickness interference, etc.) is recorded by the interference image receiver set on the back. This method is based on the birefringence characteristics and stress photoelastic effect of quartz material. When there is a stress gradient or structural disturbance (such as edge bending, microcrack tip stress concentration), it will produce abnormal distortion (non-equidistant, offset) or local densification or disappearance or fringe jumping or singular point of the interference fringes. The system extracts the stress distribution information in the interference fringe pattern through image processing algorithms (such as fringe enhancement, phase unwrapping, and pattern comparison), and aligns the space with the infrared thermal distribution map and the feature correlation. If both types of patterns show abnormal signals in a local area (such as edge corners, connection welds, or openings), for example, heat spots combined with distorted fringes, this area will be marked as a suspected area with structural defects with high confidence. Finally, the system integrates these results to form a temperature response and cooperation collection thermal image pattern and optical interference pattern defect evaluation framework, outputs the structural health state evaluation report of each quartz workpiece, marks the abnormal area coordinates and risk level, and can trigger the subsequent high-precision verification link (such as white light interferometer retest or stress relief heat treatment suggestion). Through this integrated method, high-throughput and high-resolution defect positioning and classification can be achieved without physical contact or destructive operation, which is especially suitable for internal micro-stress or hidden crack problems that are difficult to identify by traditional conductivity / ORP monitoring. This technology greatly improves the detection capability of the quartz cleaning section in both cleanliness and structural integrity, and promotes the entire cleaning, detection and process feedback link to data-driven intelligent closed-loop control.

[0037] It's important to emphasize that in conventional experiments or nondestructive testing systems, inducing defect responses or stimulating stress distribution often requires a separate heating source, a set heating rate, and temperature control within a dedicated testing environment. This not only increases equipment cost and operational complexity, but also leads to low testing efficiency and difficulty incorporating into mass production processes. However, this solution eliminates the need for a new heating control module. Instead, it cleverly leverages the rinsing and heating phase of the cleaning process to achieve both thermal excitation and optical interferometry detection of the device, thus enabling embedded stress defect identification. The rinsing phase of the quartz device cleaning process is already designed to use ultrapure water or a neutralizing solution at 60–90°C for thermal rinsing, aiming to reduce surface tension and promote the removal of particles and chemical residues. Quartz material exhibits excellent thermal response characteristics in this temperature range, sufficient to induce heterogeneous thermal conduction behavior and stress photoelastic interferometry responses in localized stress or crack regions.

[0038] In some examples, evaluating the quality defects of the quartz workpiece based on the thermal distribution and interference fringe pattern during the heating process of the quartz workpiece includes: Determine the heat retention area based on the thermal distribution of the quartz workpiece during the heating process; Determine the fringe distortion area based on the interference fringe pattern; An image fusion algorithm is used to detect the heat retention area and the stripe distortion area, so as to determine whether they are stress concentration areas through a classifier.

[0039] For example, during the rinsing and heating process, an infrared thermal imager is used to continuously capture the dynamic changes in the surface temperature of the quartz workpiece and generate a time-series thermal distribution map. The system analyzes the thermal maps at different time points and extracts the heat diffusion velocity field and the local temperature rise curve. If certain areas still significantly lag behind in temperature rise when other parts are close to the stable temperature, or if the local temperature rise slope is significantly lower than the surrounding area, these areas are marked as heat retention areas. The root causes of heat retention usually include: microstructural non-uniformity inside the quartz (such as inclusions, bubbles), local stress concentration affecting the heat conduction path, surface cracks, etc. These phenomena will cause time delays and spatial distortions in heat diffusion, becoming important thermal response characteristics for subsequent defect analysis.

[0040] Simultaneously, a polarized light interferometry system acquires the interference fringe pattern of the quartz component during the heating process. This pattern appears as a series of light intensity fringes, whose distribution patterns directly reflect the changes in the stress field within the material. After performing image processing and analysis on the interference pattern (such as edge detection, phase unwrapping, and fringe feature extraction), the system can identify the following fringe distortion characteristics: localized fringe density or sparseness; fringe distortion, breakage, or closed loop formation; and fringe trajectory deviation from the theoretical stress distribution direction. These abnormal fringe regions are marked as fringe distortion areas, which are typically the result of non-uniform stress within the material or concentrated stress at microcrack tips. Finally, the system spatially aligns (registers) the thermal distribution map with the interference fringe pattern and constructs a fused feature map using image fusion algorithms (such as multi-scale feature fusion, methods based on joint edge-texture analysis, or the attention mechanism fusion module in a convolutional neural network). Within this fused map, the system extracts regions of spatial overlap or high similarity between the two types of patterns—regions exhibiting both heat retention and interference fringe distortion—as suspected stress anomalies. The multi-dimensional features of these regions (such as area, shape, thermal gradient change rate, and fringe offset amplitude) are fed into a pre-trained classifier, which uses a support vector machine (SVM), random forest (RF), or lightweight neural network model to determine whether they are stress concentration areas. The classifier is trained using samples from previously labeled stress defects and non-defective areas from historical cleaning batches, continuously optimizing its accuracy through supervised learning.

[0041] In some examples, this also includes: In the drying stage after cleaning, tail flow information of the drying airflow after flowing through the quartz workpiece is obtained, wherein the drying airflow is a constant temperature and constant pressure airflow; The wake flow field is determined based on the wake information, so as to evaluate quality defects of the quartz workpiece according to the wake flow field. The quality defects include the presence of a liquid film thickness difference curve and / or surface foreign matter attachment defects in the quartz workpiece.

[0042] For example, after cleaning is complete and the drying phase begins, the system introduces a constant temperature and pressure airflow (such as temperature-controlled nitrogen, clean dry air, or high-purity argon) that steadily flows over the surface of the quartz workpiece from above or laterally, creating a controllable laminar drying environment. During the evaporation of the water film on the quartz workpiece, any surface morphological anomalies (such as uneven film residue, microscopic particles, scratches, and stress-induced warping) can disrupt the local boundary layer structure, velocity distribution, and vortex formation of the airflow. These disturbances manifest as pressure perturbations, velocity shear fluctuations, or abnormal temperature and humidity gradients in the wake field after the airflow passes through the quartz device. The system incorporates a high-precision wake monitoring device at the tail of the quartz device, employing one or a combination of the following methods: a laser Doppler velocimeter to capture minute velocity changes in real time; a micro hot wire anemometer to measure the spatial distribution of temperature and velocity in the airflow; particle image velocimetry combined with labeled particles to visualize the wake structure; and an infrared thermal imager / gas refractometer to capture thermal disturbances or optical density distortion patterns in the airflow. The above device collects the spatial velocity distribution and local turbulence structure of the wake, and the system reconstructs the flow field. It can use CFD approximate solutions or neural field modeling algorithms to extract the following defect-related characteristic patterns: Liquid film thickness difference curve identification: If the dry airflow suddenly slows down or generates micro-vortices above a certain area, it is often caused by a thick water film or limited evaporation in that area. Based on this, the system extracts the airflow disturbance contour map and infers the liquid film thickness difference contour curve on the quartz surface to determine whether there is cleaning residue.

[0043] Surface foreign matter attachment identification: If a persistent point vortex or flow field shear distortion zone forms in the wake, it indicates that the airflow is obstructed at that location, which is likely corresponding to particles, debris, or scratches and depressions adhered to the surface. The system spatially aligns the center of the disturbance and outputs the coordinates of the surface defect.

[0044] For example, the system maps and aligns the abnormal wake area with the three-dimensional structural model of the quartz device, and outputs an analysis report including the liquid film thickness distribution curve, local attachment judgment results, and wake disturbance index distribution diagram. It can also superimpose the front-end cleaning liquid status and particle release data to collaboratively decide whether the device needs to enter secondary rinsing, microwave heating and drying compensation, or enter the quality inspection review path.

[0045] As can be seen, inverting the residual state on the surface of quartz devices through tail flow field information not only avoids the additional contamination or resolution limitations associated with contact or imaging interference detection, but also highly sensitively captures drying anomalies that are difficult to discern with the naked eye. This method is particularly suitable for detecting hidden defects such as thick liquid film areas, uneven hydrophobicity distribution, and localized particle redeposition left behind in quartz components after cleaning. Combined with multi-stage process monitoring, this significantly enhances the intelligent closed-loop control of the entire cleaning-drying-testing process, playing a crucial role in ensuring the safety of semiconductor quartz devices that require high cleanliness levels and thermal stability.

[0046] In some examples, the cleaning fluid state change parameter further includes pH and surface tension, and the method further includes: Based on the real-time acquired parameters of the cleaning fluid state change in the cleaning tank, the current cleaning fluid contamination level is evaluated and determined; The cleaning intensity parameter is adjusted according to the current pollution level of the cleaning fluid, wherein the cleaning intensity parameter includes at least one of ultrasonic frequency, cleaning time or cleaning temperature.

[0047] For example, during the cleaning process, the system uses a multi-sensor array to synchronously collect real-time state changes in the cleaning fluid, including but not limited to pH, which reflects the degree of neutralization and reactivity of the acid / base components in the cleaning fluid. A gradual deviation of pH from the initial value (e.g., from a strong acid to neutral) typically indicates that the active components in the cleaning fluid have been consumed by contaminants. Surface tension is measured online using a gravimetric or bubble pressure method. Higher surface tension indicates a reduced ability of the cleaning fluid to remove contaminants, such as residual organic matter or surfactant degradation. Conductivity measures changes in ion concentration and is used to identify metal ion release and particle dissolution. ORP characterizes the remaining capacity of active substances in the redox system and reflects its ability to remove organic contaminants. Based on the comparison of these parameter trends with historical samples, the system uses a multi-indicator fusion model, such as a weighted decision tree, fuzzy rule system, or neural network regression model, to determine the contamination level. The current cleaning fluid state is classified into three levels: clean, moderately contaminated, and severely contaminated, and associated with a pre-set process response logic.

[0048] For example, once the contamination level is determined, the system automatically invokes the cleaning process control module to adaptively adjust the cleaning intensity parameters to compensate for the loss of cleaning efficiency caused by the deterioration of the cleaning fluid's performance. When the cleaning fluid is moderately contaminated, the current cleaning step (e.g., pickling, rinsing) is appropriately extended by 5–15% to ensure sufficient surface reaction time. If the system identifies a "severely contaminated" state accompanied by highly viscous contaminant residue (e.g., organic colloids), the system will increase the ultrasonic frequency or switch to a low-frequency, high-power mode to improve stripping efficiency. If the surface tension increases significantly, the cleaning fluid temperature can be increased (e.g., from 60°C to 75°C) to reduce liquid viscosity and enhance decontamination and diffusion capabilities. This adjustment process is a real-time closed-loop control mechanism that can dynamically compensate before cleaning is complete, thereby preventing incomplete cleaning, residue, or secondary particle adsorption caused by deteriorating cleaning fluid performance.

[0049] It is understandable that by introducing pH and surface tension, two key parameters that characterize the activity and interfacial behavior of the cleaning fluid, the dimension and accuracy of pollution monitoring have been significantly improved. At the same time, combined with the linkage control of pollution level and cleaning intensity parameters, an adaptive cleaning scheduling mechanism based on cleaning fluid state perception is constructed. This mechanism is particularly critical in the quartz workpiece cleaning scenario, because the sources of pollution are complex, such as heat treatment oxide layer, residual flux, particulate adsorption, etc., and fixed cleaning schemes are difficult to cover all variables. Through this method, the cleaning efficiency per unit time can be effectively improved, the probability of abnormal rework can be reduced, and the cleaning quality of quartz devices can always be ensured to be in a highly controllable range. It is one of the core guarantee means to achieve high-yield quartz workpiece production. The wake velocity field can be restored using a CFD approximation model or a neural flow field estimation model , determine the standard deviation of flow velocity disturbance >Threshold, marking anomalies; wake disturbance types can be mapped to liquid film thickness curves / particle attachment areas.

[0050] The above describes the digital production supervision method in the embodiment of the present application. The following describes the digital production supervision system in the embodiment of the present application.

[0051] See also Figure 2 In the embodiments of the present application, an embodiment of a digital production supervision system is described, which may include: Matching unit 201, used for automatically matching a cleaning recipe based on the quartz workpiece information during the cleaning stage of the quartz workpiece after heat treatment, wherein the cleaning recipe includes a cleaning solution ratio and a cleaning time-temperature control curve; The acquisition unit 202 is used to obtain the state change parameters of the cleaning liquid in the cleaning tank in real time, wherein the cleaning liquid state parameters include conductivity and redox potential; The evaluation unit 203 is configured to evaluate the quality defects of the quartz workpiece based on the cleaning liquid state change parameters.

[0052] In summary, the digital production supervision system provided by the above embodiments automatically matches a cleaning formula based on quartz processing piece information in the stage of cleaning the quartz processing piece after hot processing, the cleaning formula including a washing liquid ratio and a cleaning time-temperature control curve; real-time acquisition of a cleaning liquid state change parameter in a cleaning tank, the cleaning liquid state parameter including conductivity and a redox potential; and quality defect evaluation of the quartz processing piece based on the cleaning liquid state change parameter. The transition from post-detection to process judgment is realized, and the rework cost after cleaning is completed is greatly reduced; the cleaning intensity is intelligently matched according to the pollution difficulty of each batch of products, and the time and reagents are avoided to be wasted due to light pollution and heavy washing. In the cleaning process, the dissolution of pollutants is dynamically monitored, the cleaning parameters can be adjusted in a timely manner, and the complete removal rate of pollutants is improved; the cleaning liquid state change signal is used as an implicit quality index to improve the sensitivity to implicit defects such as micro-cracks and embedded impurities. Not only does the quartz cleaning process break away from the traditional passive mode of relying on end detection, but also realizes the feedforward defect recognition based on process characteristics, and the precise matching logic between the product and the cleaning process is built, which greatly reduces the rework rate and cleaning liquid use redundancy. The yield and stability of the quartz piece in the high-cleanliness process can be improved, a traceable and evolvable digital supervision capability is provided for the production line, and the intelligent level and controllability of the high-end manufacturing process are significantly enhanced.

[0053] The above Figure 2 The digital production supervision system in the embodiments of the present application is described from the perspective of modular functional entities, and the digital production supervision system in the embodiments of the present application is described in detail from the perspective of hardware processing. Please refer to Figure 3 The digital production supervision system 300 in the embodiments of the present application includes: An input device 301, an output device 302, a processor 303, and a memory 304, wherein the number of the processor 303 can be one or more, Figure 3 In some embodiments of the present application, the input device 301, the output device 302, the processor 303, and the memory 304 can be connected through a bus or other means, wherein, Figure 3 In some embodiments of the present application, the input device 301, the output device 302, the processor 303, and the memory 304 are connected through a bus.

[0054] The processor 303 is configured to execute the following steps by calling the operation instructions stored in the memory 304: In the stage of cleaning the quartz processing piece after hot processing, a cleaning formula is automatically matched based on quartz processing piece information, the cleaning formula including a washing liquid ratio and a cleaning time-temperature control curve; Real-time acquisition of a cleaning liquid state change parameter in a cleaning tank, the cleaning liquid state parameter including conductivity and a redox potential; The quality defects of the quartz workpiece are evaluated based on the cleaning liquid state change parameters.

[0055] By calling the operation instructions stored in the memory 304, the processor 303 is also used to execute Figure 1 Any method in the corresponding embodiment.

[0056] See also Figure 4 , Figure 4 Schematic diagram of an electronic device according to an embodiment of the present application.

[0057] like Figure 4 As shown, an embodiment of the present application provides an electronic device, including a memory 410, a processor 420, and a computer program 411 stored in the memory 420 and executable on the processor 420. When the processor 420 executes the computer program 411, the following steps are implemented: During the cleaning of the quartz workpiece after heat treatment, a cleaning recipe is automatically matched based on the quartz workpiece information, the cleaning recipe including the ratio of the washing liquid and the cleaning time-temperature control curve; Real-time acquisition of state change parameters of the cleaning liquid in the cleaning tank, wherein the cleaning liquid state parameters include conductivity and redox potential; The quality defects of the quartz workpiece are evaluated based on the cleaning liquid state change parameters.

[0058] In the specific implementation process, when the processor 420 executes the computer program 411, it can achieve Figure 1 Any implementation manner in the corresponding embodiments.

[0059] Since the electronic device introduced in this embodiment is a device used to implement a digital production supervision system in the embodiment of this application, based on the method introduced in the embodiment of this application, technical personnel in this field can understand the specific implementation method of the electronic device of this embodiment and its various variations. Therefore, how the electronic device implements the method in the embodiment of this application will not be introduced in detail here. As long as the equipment used by technical personnel in this field to implement the method in the embodiment of this application falls within the scope of protection of this application.

[0060] See also Figure 5 , Figure 5 A schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present application.

[0061] like Figure 5 As shown, this embodiment provides a computer-readable storage medium 500, on which a computer program 511 is stored. When the computer program 511 is executed by a processor, the following steps are implemented: During the cleaning of the quartz workpiece after heat treatment, a cleaning recipe is automatically matched based on the quartz workpiece information, the cleaning recipe including the ratio of the washing liquid and the cleaning time-temperature control curve; Real-time acquisition of state change parameters of the cleaning liquid in the cleaning tank, wherein the cleaning liquid state parameters include conductivity and redox potential; The quality defects of the quartz workpiece are evaluated based on the cleaning liquid state change parameters.

[0062] By calling the operation instructions stored in the memory 304, the processor 303 is also used to execute Figure 1 Any method in the corresponding embodiment.

[0063] Embodiments of the present application provide a computer program product comprising one or more computer instructions that, when loaded and executed on a computer, fully or partially perform the processes or functions described in accordance with the embodiments of the present application. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium may be any available medium capable of computer storage or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, hard disk, or tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0064] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0065] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0066] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0067] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0068] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0069] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A digital production supervision method, characterized in that: include: During the cleaning of the quartz workpiece after heat treatment, a cleaning recipe is automatically matched based on the quartz workpiece information, the cleaning recipe including the ratio of the washing liquid and the cleaning time-temperature control curve; Real-time acquisition of state change parameters of the cleaning liquid in the cleaning tank, wherein the state change parameters of the cleaning liquid include conductivity and redox potential; The quality defects of the quartz workpiece are evaluated based on the cleaning liquid state change parameters.

2. The method according to claim 1, characterized in that The evaluating the quality defects of the quartz workpiece based on the cleaning liquid state change parameter includes: evaluating abnormal ion release of the quartz workpiece based on the cleaning fluid state change parameter; The quality defects of the quartz workpiece are evaluated based on the abnormal ion release.

3. The method according to claim 1, characterized in that Also includes: The number and size distribution of particles removed from the fluid per unit time are recorded in real time as a dynamic response curve of the device surface cleanliness; The evaluating the quality defects of the quartz workpiece based on the cleaning liquid state change parameter includes: The particle release mode is used to analyze whether the quartz workpiece has hidden cracks caused by edge stress accumulation.

4. The method according to claim 1, wherein Also includes: In the rinsing and heating zone, the thermal distribution of the quartz workpiece during the heating process is captured by an infrared thermal imager; In the rinsing and heating zone, interference fringe patterns are collected by irradiating with polarized light; The quality defects of the quartz workpiece are evaluated based on the thermal distribution and interference fringe pattern of the quartz workpiece during the heating process.

5. The method according to claim 4, characterized in that The method of evaluating the quality defects of the quartz workpiece based on the thermal distribution and interference fringe pattern of the quartz workpiece during the heating process includes: Determine the heat retention area based on the thermal distribution of the quartz workpiece during the heating process; Determine the fringe distortion area based on the interference fringe pattern; An image fusion algorithm is used to detect the heat retention area and the stripe distortion area, so as to determine whether they are stress concentration areas through a classifier.

6. The method according to claim 1, characterized in that Also includes: In the drying stage after cleaning, tail flow information of the drying airflow after flowing through the quartz workpiece is obtained, wherein the drying airflow is a constant temperature and constant pressure airflow; The wake flow field is determined based on the wake information, so as to evaluate quality defects of the quartz workpiece according to the wake flow field. The quality defects include the presence of a liquid film thickness difference curve and / or surface foreign matter attachment defects in the quartz workpiece.

7. The method according to any one of claims 1 to 6, characterized in that The cleaning fluid state change parameters further include pH and surface tension, and the method further includes: Based on the real-time acquired parameters of the cleaning fluid state change in the cleaning tank, the current cleaning fluid contamination level is evaluated and determined; The cleaning intensity parameter is adjusted according to the current pollution level of the cleaning fluid, wherein the cleaning intensity parameter includes at least one of ultrasonic frequency, cleaning time or cleaning temperature.

8. A digital production supervision system, characterized in that: The system comprises: A matching unit, for automatically matching a cleaning recipe based on the quartz workpiece information during the cleaning process of the quartz workpiece after heat treatment, wherein the cleaning recipe includes a cleaning solution ratio and a cleaning time-temperature control curve; An acquisition unit, configured to acquire in real time parameters of a cleaning liquid state change in the cleaning tank, the parameters of the cleaning liquid state change including conductivity and redox potential; An evaluation unit is used to evaluate the quality defects of the quartz workpiece based on the state change parameters of the cleaning liquid.

9. An electronic device, characterized in that: The electronic device includes at least one processor and at least one memory connected to the processor, wherein the processor is used to call program instructions in the memory to execute the digital production supervision method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is running, the device where the storage medium is located is controlled to execute the digital production supervision method according to any one of claims 1 to 7.

Images