Highly efficient continuous chemical vapor deposition monitoring system for new energy material preparation

Abstract

This invention discloses a high-efficiency continuous chemical vapor deposition (CVD) monitoring system for the preparation of new energy materials. The system comprises: synchronously uploading component structural parameters during the monitoring layer operation phase; synchronously monitoring CVD environmental information based on the monitoring layer during CVD; acquiring the monitored CVD environmental information from the monitoring layer during the evaluation layer operation; evaluating the uniformity of each region of the CVD spatial field based on the CVD environmental information; receiving the evaluation results of the uniformity of each region of the CVD spatial field from the evaluation layer during the subsequent operation of the identification layer; identifying the detection target based on the evaluation results; and performing CVD uniformity detection on the detection target to estimate the CVD qualification of the component. This invention achieves dynamic quantitative evaluation of the environmental homogeneity of each region during the CVD process by collecting multi-dimensional data such as component structural parameters, thermal imaging images of the CVD environment, and gas source pressure and flow rate, based on a spatial field regional uniformity evaluation model, and integrating parameters such as temperature, pressure, flow rate status, and contour image similarity.

Description

Technical Field

[0001] This invention relates to the field of new energy material preparation technology, specifically to a high-efficiency continuous chemical vapor deposition monitoring system for new energy material preparation. Background Technology

[0002] Chemical vapor deposition (CVD) technology is widely used in the preparation of new energy materials. It can be used to prepare lithium-ion battery electrode materials, fuel cell catalyst supports, and photovoltaic thin films. By precisely controlling temperature, gas pressure, and gas composition, uniform material deposition can be achieved, improving conductivity, stability, and interfacial properties, providing crucial support for the development of high-energy-density batteries, high-efficiency catalysts, and low-loss photovoltaic devices.

[0003] Patent application No. 202410455087.5 discloses a CVD process control system based on physicochemical mechanisms and machine learning, comprising: a physicochemical mechanism research and modeling module, used to establish a thin film growth theoretical model based on the physicochemical process of CVD, and to obtain key physicochemical parameters affecting the CVD thin film growth process based on the thin film growth theoretical model, each key physicochemical parameter corresponding to one or more process parameters; a data acquisition and data preprocessing module, used to acquire and preprocess instrument sensor data containing the key physicochemical parameters; and, in the initial training stage, to acquire CVD deposited thin film thickness measurements corresponding to the instrument sensor data; each instrument sensor data corresponds to one process parameter; a machine learning model selection and training module, used to train a machine learning model based on the data acquired by the data acquisition and data preprocessing module and select the optimal model; and a model interpretability analysis module, used to select the optimal model based on the data acquired by the data acquisition and data preprocessing module. The model identifies key process parameters affecting the final film thickness during CVD film growth, as well as the ways in which these parameters influence the final film thickness and the interactions between them. A real-time monitoring and automatic adjustment module is used to deploy the optimal model on the CVD production line, enabling real-time monitoring and control of the CVD process. This application aims to address the uncontrollability issues arising from the complex physicochemical mechanisms of CVD, particularly the significant differences in film thickness and quality between different batches of wafers under fixed process formulations. Current industry methods accept these differences by sampling and measuring wafers after each batch completes the CVD step to obtain actual film thickness information, and then using this information to adjust the process formulation for the next batch. However, this method, relying on experience-based adjustments and subsequent measurements to attempt to control film thickness, is not only inefficient but also fails to achieve real-time system optimization and automated process formulation adjustment.

[0004] However, for the application of CVD technology in the preparation of new energy materials, the CVD process effect directly affects the performance and lifespan of new energy material components in the later use. Existing detection technologies are often focused on completing the quality inspection of CVD components. The monitoring technology during the CVD process is relatively simple, so it is impossible to predictively assess the quality of the finished processed components or to take timely countermeasures during the process.

[0005] To this end, we propose a high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials. Summary of the Invention

[0006] In view of the above-mentioned shortcomings of the existing technology, the present invention provides a high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials, which can effectively solve the problems of the existing technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions;

[0008] This invention discloses a high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials, comprising: a monitoring layer, an evaluation layer, and an identification layer;

[0009] During the monitoring layer operation phase, the component structure parameters are uploaded synchronously. During the CVD process, the component synchronously monitors the CVD environment information based on the monitoring layer. The evaluation layer operation obtains the CVD environment information monitored in the monitoring layer and evaluates the uniformity of each region of the CVD spatial field based on the CVD environment information. The identification layer operation receives the evaluation results of the uniformity of each region of the CVD spatial field in the evaluation layer, identifies the detection target based on the evaluation results, and performs CVD uniformity detection on the detection target to estimate the CVD qualification of the component.

[0010] The evaluation layer includes a partitioning module, an evaluation module, and a normalization module. The partitioning module receives CVD environmental information and performs partitioning processing on the CVD environmental information. The evaluation module obtains the partitioning processing results of the CVD environmental information in the partitioning module, and evaluates the uniformity of each region of the CVD spatial field based on the partitioning processing results of the CVD environmental information. The normalization module receives the evaluation results of the uniformity of each region of the CVD spatial field in the evaluation module and performs normalization processing on each evaluation result.

[0011] The logic for evaluating the uniformity of each region in the CVD spatial field is expressed as follows:

[0012]

[0013] In the formula: f(a) is the uniformity characterization value of CVD spatial field region a; n and m are the total number of memory metal contour images in the sub-images of set α and set β corresponding to CVD spatial field region a; sim(i,i+1) and sim(j,j+1) are the similarity between the i-th contour image and the (i+1)-th contour image, and the similarity between the j-th contour image and the (j+1)-th contour image. P represents the historical average pressure and historical average flow rate of the gas source in the CVD environment; now H now This represents the average current pressure and average current flow rate at the gas source in the CVD environment. This indicates the operation of retrieving the maximum value within the parentheses;

[0014] Wherein, the larger f(a) is, the better the uniformity of the corresponding CVD spatial field region. f(a) is used to comprehensively represent the uniformity of temperature, pressure, and flow state in the CVD spatial field region. The uniformity characterization value corresponding to each CVD spatial field region is obtained based on each CVD spatial field region. This indicates the operation of calculating the average of the expression below.

[0015] Furthermore, the monitoring layer includes an upload module, a thermal imaging module, and a sensing module. The upload module is used to upload the component structure parameters, construct a three-dimensional model of the component based on the component structure parameters, divide the three-dimensional model of the component into two sub-component three-dimensional models, and analyze the complexity of the two sub-component three-dimensional models respectively. The thermal imaging module is used to acquire thermal imaging images of the component's CVD environment in real time, and the sensing module is used to sense the pressure and flow information of the gas source end in the CVD environment in real time.

[0016] In this process, after the complexity analysis of the sub-component 3D model is completed, the complexity analysis results are simultaneously marked on the corresponding sub-component 3D model. The thermal imaging images collected by the thermal imaging module and the pressure and flow information of the gas source end sensed by the sensing module are all recorded as CVD environmental information.

[0017] Furthermore, the monitoring layer includes an upload module, a thermal imaging module, and a sensing module. The upload module is used to upload the component structure parameters, construct a three-dimensional model of the component based on the component structure parameters, divide the three-dimensional model of the component into two sub-component three-dimensional models, and analyze the complexity of the two sub-component three-dimensional models respectively. The thermal imaging module is used to acquire thermal imaging images of the component's CVD environment in real time, and the sensing module is used to sense the pressure and flow information of the gas source end in the CVD environment in real time.

[0018] In this process, after the complexity analysis of the sub-component 3D model is completed, the complexity analysis results are simultaneously marked on the corresponding sub-component 3D model. The thermal imaging images collected by the thermal imaging module and the pressure and flow information of the gas source end sensed by the sensing module are all recorded as CVD environmental information.

[0019] Furthermore, the complexity analysis logic of the sub-component's 3D model is expressed as follows:

[0020]

[0021] In the formula: θ represents the complexity of the 3D model of the sub-element; q represents the total number of visible corner points of the 3D model of the sub-element relative to the CVD context from a low-angle view; χ represents the Gaussian curvature integral term; E represents the number of model edges; F represents the number of model faces; V 2 / 3 K is on the order of surface area; V is on the order of volume; K(CS) max ) represents the curvature of the maximum surface area on the 3D model of the sub-component; S represents the area of ​​the largest curved surface on the surface of the sub-component's 3D model; all This represents the sum of the areas of all faces of the 3D model of the sub-component;

[0022] The larger the θ value, the more complex the 3D model of the sub-element.

[0023] Furthermore, the formula for calculating the Gaussian curvature integral term χ is as follows:

[0024]

[0025] In the formula: v is the volume of the model bounding box; S is the surface to be integrated; K(x,y,z) is the Gaussian curvature of each point on the surface of the sub-element 3D model, and the absolute value is averaged after integration; dS represents the area element on the surface.

[0026] The model bounding box volume v is defined as the ratio of the component's 3D model bounding box volume to the sub-component's 3D model volume.

[0027] Furthermore, during the segmentation module operation phase, the segmentation processing target is the thermal imaging image in the CVD environment information. The thermal imaging image acquired by the thermal imaging module set at the top is symmetrically segmented. The segmentation line is the midline between the two thermal imaging modules deployed on the outside of the CVD environment, resulting in two sub-images, denoted as A and B. A and B are denoted as the sub-image set α.

[0028] The thermal imaging images acquired from the externally mounted thermal imaging module are horizontally segmented relative to the acquisition viewpoint. During horizontal segmentation, the span of each segment is equal to obtain several sub-images, denoted as X. A (1) X A (2) XA (3), ...; X B (1) X B (2) X B (3), ...;

[0029] With X A (1) For example, X A (1) represents the first sub-image from top to bottom on side A of sub-image A, and so on, X A (1) X A (2) X A (3), ...; X B (1) X B (2) X B (3) ... is denoted as the sub-image set β;

[0030] In this process, one sub-image is selected from each of the two sub-image sets to represent a CVD spatial field. The selection logic follows this rule: select one sub-image from sub-image set β, and denote the selected sub-image and the sub-image in sub-image set α whose name index corresponds to the selected sub-image as a CVD spatial field. This process is repeated to obtain a number of CVD spatial field regions equal to the number of sub-images in sub-image set β.

[0031] Furthermore, the normalization process for the evaluation results of the evaluation module in the normalization module is as follows:

[0032]

[0033] In the formula: f(a)′ is the uniformity representation value of the CVD spatial field region a after normalization; f(θ1, θ2) is the decision function; θ1 and θ2 are the complexity calculation results of the two sub-element three-dimensional models; u is the pixel color value level in the thermal imaging image; k v This is the ratio of the number of pixels at the corresponding level in the two sub-images corresponding to the CVD spatial field region a at the v-th level.

[0034] The decision function takes the value θ1 or θ2, and is the result of the complexity calculation of the 3D model of the sub-element closest to the CVD spatial field region a. v When calculating, always compare the smaller value with the larger value to ensure that each ratio is less than 1.

[0035] Furthermore, the identification layer includes an identification module, a detection module, and an output module. The identification module is used to traverse the normalization results of the uniformity representation values ​​of each CVD spatial field region, identify the value of each result, and sort the results from smallest to largest based on the size of the result value, so that the result with the smaller result value is sorted first and the result with the larger result value is sorted last. The detection module is used to detect whether the components in the CVD spatial field region corresponding to each result are qualified based on the sorting of the normalization results of the uniformity representation values ​​of each CVD spatial field region. When it is detected that all components in the CVD spatial field region corresponding to a result are qualified, the process ends, and the result and all components in the CVD spatial field region corresponding to the result in the sorting queue are judged as qualified. The output module is used to receive the component qualification judgment result from the detection module and output the CVD spatial field region to which the qualified component belongs as the output content.

[0036] Among them, the component qualification judgment result in the detection module and the output content of the output module are the contents of the component CVD qualification estimate.

[0037] Furthermore, the identification layer operates after the component completes the CVD process. In the stage of determining whether the component is qualified, the surface of the component is bombarded by focusing an electron beam to release characteristic X-rays with specific energy, which are then detected by an energy dispersive spectrometer to complete the determination.

[0038] Specifically, when determining the affiliation of components in the CVD spatial field region, the affiliation is determined by the intersection spatial region obtained by stretching two sub-images corresponding to the CVD spatial field region.

[0039] Furthermore, the segmentation module is interconnected with a sensing module via a local area network, the sensing module is interconnected with a thermal imaging module and an uploading module via a local area network, the segmentation module is interconnected with an evaluation module and a normalization module via a local area network, the normalization module is interconnected with an identification module via a local area network, and the identification module is interconnected with a detection module and an output module via a local area network.

[0040] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects:

[0041] This invention provides a high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials. During operation, the system collects multi-dimensional data such as component structural parameters, thermal imaging images of the CVD environment, and gas source pressure and flow rate. Based on a spatial field regional uniformity evaluation model, it integrates parameters such as temperature, pressure, flow rate status, and contour image similarity to achieve dynamic quantitative evaluation of the environmental homogeneity of each region during the CVD process.

[0042] This system can not only analyze the spatial field equilibrium in real time during the deposition process, but also proactively identify key detection areas by normalizing the data and correlating it with the complexity of the components. Combined with electron beam detection technology, it can detect the quality of finished components, thereby simplifying and improving the efficiency of the detection process for mass-produced components. It effectively solves the pain point of existing technologies being unable to predictively assess the quality of finished components and respond to process problems in a timely manner. It provides a full-chain intelligent solution for the preparation of new energy materials, from process monitoring to quality prediction, effectively improving the controllability of CVD processes and the reliability of finished component quality. This lays a technical foundation for extending the service life of new energy material components and optimizing their performance in later use. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0044] Figure 1 A schematic diagram of a high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials;

[0045] Figure 2 This is a schematic diagram showing the positional relationship between the thermal imaging module and the CVD cavity in this invention;

[0046] Figure 3 This is a schematic diagram illustrating the thermal imaging image acquisition direction and segmentation principle in this invention;

[0047] Figure 4 This is a schematic diagram illustrating an example of the logic for determining the CVD spatial field region in this invention.

[0048] The markings in the diagram represent: 1. CVD cavity; 2. Shape memory metal; 3. Thermal imaging module; 4. Example plane markings where the thermal imaging modules are located on the same plane. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0050] The present invention will be further described below with reference to embodiments.

[0051] Example:

[0052] This embodiment presents a high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials, such as... Figure 1 As shown, it includes: a monitoring layer, an evaluation layer, and an identification layer;

[0053] During the monitoring layer operation phase, the component structure parameters are uploaded synchronously. During the CVD process, the component synchronously monitors the CVD environment information based on the monitoring layer. The evaluation layer operation obtains the CVD environment information monitored in the monitoring layer and evaluates the uniformity of each region of the CVD spatial field based on the CVD environment information. The identification layer operation receives the evaluation results of the uniformity of each region of the CVD spatial field in the evaluation layer, identifies the detection target based on the evaluation results, and performs CVD uniformity detection on the detection target to estimate the CVD qualification of the component.

[0054] The monitoring layer includes an upload module, a thermal imaging module, and a sensing module. The upload module is used to upload the component structural parameters, construct a 3D model of the component based on the component structural parameters, divide the 3D model of the component into two sub-component 3D models, and analyze the complexity of the two sub-component 3D models respectively. The thermal imaging module is used to acquire thermal imaging images of the component's CVD environment in real time. The sensing module is used to sense the pressure and flow information of the gas source end in the CVD environment in real time.

[0055] Among them, after the complexity analysis of the sub-component 3D model is completed, the complexity analysis results are simultaneously marked on the corresponding sub-component 3D model. The thermal imaging images collected by the thermal imaging module and the pressure and flow information of the gas source end sensed by the sensing module are all recorded as CVD environmental information.

[0056] There are three thermal imaging modules, which are deployed on the outside and top of the CVD environment respectively. The three thermal imaging modules are on the same plane, with two on the outside and horizontal to each other, and one on the top and perpendicular to the thermal imaging image acquisition direction of the thermal imaging modules on the outside. The CVD environment corresponds to the cavity in which the CVD operation is performed.

[0057] The outer wall of the CVD environment is uniformly covered with shape memory metal. The shape memory metal deployed on the outer wall of the CVD environment is symmetrically distributed when viewed from the top and front directions. Within the preset temperature range that adapts to the temperature changes of CVD operation, the shape memory metal deforms and recovers according to the preset shape based on temperature changes.

[0058] The complexity analysis logic of the sub-component 3D model is expressed as follows:

[0059]

[0060] In the formula: θ represents the complexity of the 3D model of the sub-element; q represents the total number of visible corner points of the 3D model of the sub-element relative to the CVD context from a low-angle view; χ represents the Gaussian curvature integral term; E represents the number of model edges; F represents the number of model faces; V 2 / 3 K is on the order of surface area; V is on the order of volume; K(CS) max ) represents the curvature of the maximum surface area on the 3D model of the sub-component; S represents the area of ​​the largest curved surface on the surface of the sub-component's 3D model; all This represents the sum of the areas of all faces of the 3D model of the sub-component;

[0061] The larger the θ value, the more complex the 3D model of the sub-element;

[0062] The formula for calculating the Gaussian curvature integral term χ is:

[0063]

[0064] In the formula: v is the volume of the model bounding box; S is the surface to be integrated; K(x,y,z) is the Gaussian curvature of each point on the surface of the sub-element 3D model, and the absolute value is averaged after integration; dS represents the area element on the surface.

[0065] Wherein, the model bounding box volume v is: the ratio of the component 3D model bounding box volume to the sub-component 3D model volume to the component 3D model volume;

[0066] The complexity of the three-dimensional model of the sub-component is calculated using the above logical formula, which provides support for the normalization processing of the evaluation results of the evaluation module in this embodiment.

[0067] The evaluation layer includes a partitioning module, an evaluation module, and a normalization module. The partitioning module receives CVD environmental information and performs partitioning processing on the CVD environmental information. The evaluation module obtains the partitioning processing results of the CVD environmental information in the partitioning module, and evaluates the uniformity of each region of the CVD spatial field based on the partitioning processing results of the CVD environmental information. The normalization module receives the evaluation results of the uniformity of each region of the CVD spatial field in the evaluation module and performs normalization processing on each evaluation result.

[0068] The logic for evaluating the isostatic equilibrium of different regions in a CVD spatial field is as follows:

[0069]

[0070] In the formula: f(a) is the uniformity characterization value of CVD spatial field region a; n and m are the total number of memory metal contour images in the sub-images of set α and set β corresponding to CVD spatial field region a; sim(i,i+1) and sim(j,j+1) are the similarity between the i-th contour image and the (i+1)-th contour image, and the similarity between the j-th contour image and the (j+1)-th contour image. P represents the historical average pressure and historical average flow rate of the gas source in the CVD environment; now H now This represents the average current pressure and average current flow rate at the gas source in the CVD environment. This indicates the operation of retrieving the maximum value within the parentheses;

[0071] Wherein, the larger f(a) is, the better the uniformity of the corresponding CVD spatial field region. f(a) is used to comprehensively represent the uniformity of temperature, pressure, and flow state in the CVD spatial field region. The uniformity characterization value corresponding to each CVD spatial field region is obtained based on each CVD spatial field region. This indicates the operation of calculating the average of the expression below;

[0072] The above logical formula is used to calculate and further define the uniformity characterization value of each CVD spatial field region, providing further support for the operation of the identification layer of the system in this embodiment, so as to finally complete the CVD qualification estimation of the component.

[0073] During the module operation phase, the processing target is the thermal imaging image in the CVD environment information. The thermal imaging image acquired by the thermal imaging module set at the top is symmetrically segmented. The segmentation line is the midline between the two thermal imaging modules deployed on the outside of the CVD environment, resulting in two sub-images, denoted as A and B. A and B are denoted as the sub-image set α.

[0074] The thermal imaging images acquired from the externally mounted thermal imaging module are horizontally segmented relative to the acquisition viewpoint. During horizontal segmentation, the span of each segment is equal to obtain several sub-images, denoted as X. A (1) X A (2) X A (3), ...; X B (1) X B (2) X B (3), ...;

[0075] With X A (1) For example, X A (1) represents the first sub-image from top to bottom on side A of sub-image A, and so on, X A (1) X A (2) XA (3), ...; X B (1) X B (2) X B (3) ... is denoted as the sub-image set β;

[0076] In this process, one sub-image is selected from each of the two sub-image sets to represent a CVD spatial field. The selection logic follows this rule: select one sub-image from sub-image set β, and record the selected sub-image and the sub-image in sub-image set α whose name index corresponds to the selected sub-image as a CVD spatial field. This process is repeated to obtain a number of CVD spatial field regions equal to the number of sub-images in sub-image set β.

[0077] The normalization process for the evaluation results in the evaluation module is as follows:

[0078]

[0079] In the formula: f(a)′ is the uniformity representation value of the CVD spatial field region a after normalization; f(θ1, θ2) is the decision function; θ1 and θ2 are the complexity calculation results of the two sub-element three-dimensional models; u is the pixel color value level in the thermal imaging image; k v This is the ratio of the number of pixels at the corresponding level in the two sub-images corresponding to the CVD spatial field region a at the v-th level.

[0080] The decision function takes the value θ1 or θ2, and is the result of the complexity calculation of the 3D model of the sub-element closest to the CVD spatial field region a. v When calculating, always compare the smaller value with the larger value, ensuring that each ratio is less than 1;

[0081] The above formula further defines the operation logic of the normalization module and normalizes the uniformity representation value of the CVD spatial field region so that the detection module in the system identification layer can determine the detection target element.

[0082] The recognition layer includes a recognition module, a detection module, and an output module. The recognition module is used to traverse the normalization results of the uniformity representation values ​​of each CVD spatial field region, identify the value of each result, and sort the results from smallest to largest based on the size of the result value, so that the result with the smaller result value is sorted first and the result with the larger result value is sorted last. The detection module is used to detect whether the components in the CVD spatial field region corresponding to each result are qualified based on the sorting of the normalization results of the uniformity representation values ​​of each CVD spatial field region. When it is detected that all components in the CVD spatial field region corresponding to a result are qualified, the process ends, and the result and the result in the sorting queue are set to indicate that all components in the CVD spatial field region corresponding to the result are qualified. The output module is used to receive the component qualification judgment results from the detection module and output the CVD spatial field region to which the qualified component belongs as the output content.

[0083] Among them, the component qualification judgment result in the detection module and the output content of the output module are the contents of the component CVD qualification estimation.

[0084] The identification layer operates after the component completes the CVD process. In the stage of determining whether the component is qualified, the surface of the component is bombarded by focusing an electron beam to release characteristic X-rays with specific energy, which are then detected by an energy dispersive spectrometer to complete the determination.

[0085] In determining the affiliation of components in the CVD spatial field region, the affiliation is determined by the intersection spatial region obtained by stretching two sub-images corresponding to the CVD spatial field region.

[0086] The segmentation module is interconnected with the perception module via a local area network. The perception module is interconnected with the thermal imaging module and the upload module via a local area network. The segmentation module is interconnected with the evaluation module and the normalization module via a local area network. The normalization module is interconnected with the recognition module via a local area network. The recognition module is interconnected with the detection module and the output module via a local area network.

[0087] In this embodiment, the upload module uploads the component structure parameters, constructs a 3D model of the component based on the component structure parameters, and segments the 3D model to obtain two sub-component 3D models. The complexity of the two sub-component 3D models is analyzed separately. The thermal imaging module acquires real-time thermal imaging images of the component's CVD environment. The sensing module simultaneously senses the pressure and flow information of the gas source end in the CVD environment. The segmentation module receives the CVD environment information and performs segmentation processing on the CVD environment information. The evaluation module further obtains the CVD environment information segmentation processing results from the segmentation module. Based on the CVD environment information segmentation processing results, the CVD environment information is used to evaluate the uniformity of each region of the CVD spatial field. Then, the normalization module receives the evaluation results of the uniformity of each region of the CVD spatial field from the evaluation module and performs normalization processing on each evaluation result. The system iterates through the normalized results of the uniformity characterization values ​​of each CVD spatial field region using the identification module, identifies the value of each result, and sorts the results from smallest to largest based on the value, with smaller results listed first and larger results listed last. Simultaneously, the detection module checks the components corresponding to each CVD spatial field region for compliance based on the sorted results of the normalized uniformity characterization values ​​of each CVD spatial field region. The process ends when all components corresponding to a given CVD spatial field region are found to be compliant, and the system sets the result and all components corresponding to that CVD spatial field region in the sorting queue to be considered compliant. Finally, the output module receives the component compliance determination results from the detection module and outputs the CVD spatial field regions to which the components deemed compliant belong.

[0088] The system described in the above embodiments provides effective monitoring services for new energy material components during the CVD process. By assessing the CVD environment, it effectively assists staff in managing and maintaining the CVD environment, thereby improving the quality of components produced by the CVD process by enhancing the accuracy of the CVD environment. At the same time, combined with monitoring data, it performs targeted testing on the final produced components and quickly distinguishes between qualified and unqualified components using specific testing logic, providing effective support services for the CVD process preparation of new energy material components.

[0089] See Figure 2 As shown in the figure, the positional relationship between the CVD cavity 1, the shape memory metal 2, and the thermal imaging module 3 is illustrated based on the markings in the figure. Furthermore, the example plane markings showing the thermal imaging module 3 being distributed on the same plane further demonstrate the orientation of the thermal imaging module 3.

[0090] See Figure 3As shown, the example demonstrates the segmentation results of the thermal imaging image through the horizontal arrows, and the small arrows on the left side of the horizontal arrows indicate the acquisition direction of the thermal imaging image, corresponding to the acquisition direction of each thermal imaging module 3.

[0091] See Figure 4 As shown in the figure, based on the arrow indications, the CVD spatial field region of the example was finally obtained through the evolution process of the two sub-images;

[0092] In summary, during operation, the system in the above embodiments collects multi-dimensional data such as component structural parameters, thermal imaging images, and gas source pressure and flow rate in the CVD environment. Based on a spatial field regional uniformity evaluation model, and integrating parameters such as temperature, pressure, flow rate, and contour image similarity, it achieves dynamic quantitative evaluation of the environmental homogeneity of each region during the CVD process. This system can not only analyze spatial field uniformity in real time during deposition, but also proactively identify key detection areas by normalizing and correlating with component complexity. Combined with electron beam detection technology, it can detect the quality of finished components, thereby simplifying and improving the efficiency of the detection process for mass-produced components. It effectively solves the pain point of existing technologies being unable to predictively assess the quality of finished components and respond to process problems in a timely manner. It provides a full-chain intelligent solution for the preparation of new energy materials, from process monitoring to quality prediction, effectively improving the controllability of the CVD process and the reliability of finished component quality. This lays a technical foundation for extending the service life of new energy material components and optimizing their performance in later use.

[0093] Application scenario simulation at the implementation level:

[0094] I. Application Scenarios

[0095] Taking the continuous chemical vapor deposition (CVD) preparation of lithium-ion battery cathode materials (such as lithium iron phosphate) as an example, traditional processes for coating nanoscale particles with conductive layers (such as carbon layers) suffer from problems such as difficulty in controlling the uniformity of the deposition environment and lagging quality inspection, resulting in poor product consistency. This system, through real-time monitoring and analysis, accurately assesses the stability of the deposition environment, thereby improving the uniformity of the coating layer and the yield of the finished product.

[0096] II. System Deployment and Core Functions

[0097] (I) Monitoring Layer: Multi-dimensional Data Acquisition

[0098] Component model construction:

[0099] Upload the structural parameters (such as particle size and surface morphology) of lithium iron phosphate particles, construct a 3D model and split it into two sub-models: "core" and "coating layer".

[0100] The complexity of the sub-models (such as the number of surface edges and the curvature of the surface) is analyzed, and the differences in complexity are marked to provide a reference for subsequent regional assessments.

[0101] Environmental parameter monitoring:

[0102] Thermal field monitoring: Two thermal imaging modules are horizontally deployed on the outside of the CVD cavity and one is vertically deployed on the top to collect temperature distribution images (such as horizontal and vertical thermal field differences) from different angles of the cavity in real time.

[0103] Physical perception: Special shape memory metal (such as nickel-titanium alloy) is symmetrically arranged on the outer wall of the cavity. It deforms with temperature changes within a preset temperature range. The uniformity of the temperature field is judged by observing the consistency of the metal deformation (the more consistent the deformation, the more uniform the temperature field).

[0104] Gas monitoring: Real-time sensing of pressure and flow fluctuations in the incoming gas source (such as the stability of methane and carbon dioxide input).

[0105] (II) Evaluation Layer: Spatial Field Equilibrium Analysis

[0106] Regional division:

[0107] The top thermal imaging image is divided into two sub-images along the center line of the cavity, and the outer thermal imaging image is divided into several sub-regions (such as upper, middle and lower). Each sub-image corresponds to a specific spatial field region of the cavity (such as upper left, lower right, etc.).

[0108] Overall assessment:

[0109] By comparing the similarity of shape memory metal deformation in different regions (e.g., whether the metal morphology changes are consistent at adjacent times) and the differences between historical averages and current values ​​of gas parameters, the equilibrium of temperature, pressure, and flow rate in that region can be determined. For example:

[0110] Regions with continuous metal deformation and small fluctuations in gas parameters are judged as having "excellent" equilibrium.

[0111] Regions exhibiting disordered deformation and large parameter fluctuations are classified as having "poor" equilibrium.

[0112] Results normalization:

[0113] By combining the differences in the complexity of the sub-models (such as the fact that a more complex kernel structure may affect the uniformity of deposition) and the characteristics of thermal imaging pixel distribution, the uniformity evaluation results of each region are weighted to generate an intuitive comprehensive score (e.g., 1-10 points, with higher scores indicating a more stable environment).

[0114] (III) Identification Layer: Component Conformity Determination

[0115] Priority sorting:

[0116] The regions are sorted from low to high based on their balance scores (regions with poor environmental stability are prioritized for testing).

[0117] Quality Inspection:

[0118] For the regions ranked first, the surface of the element is bombarded by focusing an electron beam, and an energy dispersive spectroscopy instrument is used to detect indicators such as coating thickness and compositional uniformity to determine whether the element is qualified.

[0119] Qualification logic: If a component in a certain area is qualified, then all components in that area and all subsequent areas are judged as qualified, quickly locking in the qualified production range.

[0120] Output results:

[0121] Mark the cavity location corresponding to the qualified area (such as the lower right or middle) to prompt the process personnel to adjust the parameters of the unqualified area (such as increasing the gas flow rate in the area or optimizing the heating power).

[0122] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A high-efficiency continuous chemical vapor deposition monitoring system for the preparation of new energy materials, characterized in that, include: Monitoring layer, evaluation layer, identification layer; During the monitoring layer operation phase, the component structure parameters are uploaded synchronously. During the CVD process, the component synchronously monitors the CVD environment information based on the monitoring layer. The evaluation layer operation obtains the CVD environment information monitored in the monitoring layer and evaluates the uniformity of each region of the CVD spatial field based on the CVD environment information. The identification layer operation receives the evaluation results of the uniformity of each region of the CVD spatial field in the evaluation layer, identifies the detection target based on the evaluation results, and performs CVD uniformity detection on the detection target to estimate the CVD qualification of the component. The evaluation layer includes a partitioning module, an evaluation module, and a normalization module. The partitioning module receives CVD environmental information and performs partitioning processing on the CVD environmental information. The evaluation module obtains the partitioning processing results of the CVD environmental information in the partitioning module, and evaluates the uniformity of each region of the CVD spatial field based on the partitioning processing results of the CVD environmental information. The normalization module receives the evaluation results of the uniformity of each region of the CVD spatial field in the evaluation module and performs normalization processing on each evaluation result. The logic for evaluating the uniformity of each region in the CVD spatial field is expressed as follows: ； In the formula: This represents the isostatic characterization value of the CVD spatial field region a; The set corresponding to CVD spatial field region a The inner sub-image represents the total amount and set of memory metal contour images. The sub-image represents the total amount of the memory metal contour image; Let i be the similarity between the i-th contour image and the (i+1)-th contour image, and let j be the similarity between the j+1-th contour image. , The historical average pressure and historical average flow rate at the gas source terminal in the CVD environment; This represents the average current pressure and average current flow rate at the gas source in the CVD environment. This indicates the operation of retrieving the maximum value within the parentheses; in, The larger the value, the better the uniformity of the corresponding CVD spatial field region. It is used to comprehensively represent the temperature, pressure, and flow state uniformity of the CVD spatial field region, and to obtain the corresponding uniformity characterization value of each CVD spatial field region based on each CVD spatial field region. The monitoring layer includes an upload module, a thermal imaging module, and a sensing module. The upload module is used to upload the component structure parameters, construct a three-dimensional model of the component based on the component structure parameters, divide the three-dimensional model of the component into two sub-component three-dimensional models, and analyze the complexity of the two sub-component three-dimensional models respectively. The thermal imaging module is used to acquire thermal imaging images of the component's CVD environment in real time. The sensing module is used to sense the pressure and flow information of the gas source end in the CVD environment in real time. Among them, after the complexity analysis of the sub-component 3D model is completed, the complexity analysis results are simultaneously marked on the corresponding sub-component 3D model. The thermal imaging images collected by the thermal imaging module and the pressure and flow information of the gas source end sensed by the sensing module are all recorded as CVD environmental information. The complexity analysis logic of the sub-component's 3D model is expressed as follows: ； In the formula: The complexity of the 3D model of the sub-component; The total number of corner points on the visible surface of the 3D model of the sub-component relative to the CVD context from a low-angle view. For the Gaussian curvature integral term; This represents the number of edges in the model. The number of faces in the model; It is on the order of surface area; On the order of volume; The curvature of the maximum surface curvature of the sub-component's 3D model; The area of ​​the largest curved surface on the surface of the sub-component's 3D model; This represents the sum of the areas of all faces of the 3D model of the sub-component; in, The larger the value, the more complex the 3D model of the sub-component; During the segmentation module's operation phase, the segmentation processing target is the thermal imaging image in the CVD environment information. The thermal imaging image acquired by the top-mounted thermal imaging module is symmetrically segmented. The segmentation line is the midline between the two thermal imaging modules deployed on the outer side of the CVD environment, resulting in two sub-images, denoted as A and B. A and B are then designated as the sub-image set. ; The thermal imaging images acquired from the externally mounted thermal imaging module are horizontally segmented relative to the acquisition viewpoint. During horizontal segmentation, each segment has an equal span to obtain several sub-images, denoted as . ; ; by For example, This represents the first sub-image from top to bottom on side A, and so on. ; denoted as sub-image set ; In this process, one sub-image is selected from each of the two sub-image sets to represent a CVD spatial field, and the selection logic follows: in the sub-image set Select a sub-image from the set of sub-images whose names are indexed. A neutron image is denoted as a CVD spatial field, and so on, to obtain a set of sub-images. The neutron image has an equal number of CVD spatial field regions; The normalization process for the evaluation results from the evaluation module in the normalization module is as follows: ； In the formula: This represents the equilibrium characterization value of the CVD spatial field region a after normalization. Let it be the decision function; The result of the complexity calculation for the two sub-components' 3D models; The pixel color value level in a thermal imaging image; This is the ratio of the number of pixels at the corresponding level in the two sub-images corresponding to the CVD spatial field region a at the v-th level. The decision function takes the value of or The complexity calculation results of the 3D model of the sub-element closest to region a in the CVD spatial field are taken. When calculating, always compare the smaller value with the larger value to ensure that each ratio is less than 1.

2. The high-efficiency continuous chemical vapor deposition monitoring system for new energy material preparation according to claim 1, characterized in that, The thermal imaging module is provided in three parts, which are respectively deployed on the outside and top of the CVD environment. The three thermal imaging modules are on the same plane, with two on the outside and horizontal to each other, and one on the top and perpendicular to the thermal imaging image acquisition direction of the thermal imaging modules on the outside. The CVD environment corresponds to the cavity in which the CVD operation is performed. The outer wall of the CVD environment is uniformly covered with shape memory metal. The shape memory metal on the outer wall of the CVD environment is symmetrically distributed when viewed from both the top and front views. Within a preset temperature range adapted to the temperature changes of CVD operation, the shape memory metal deforms and recovers according to a preset shape based on temperature changes.

3. The high-efficiency continuous chemical vapor deposition monitoring system for new energy material preparation according to claim 2, characterized in that, The Gaussian curvature integral term The calculation formula is: ； In the formula: The bounding box volume of the model; The surface to be integrated; Let be the Gaussian curvature of each point on the surface of the 3D model of the sub-element, and take the average of the absolute values ​​after integration; Represents an area element on a curved surface; Among them, the model bounding box volume The value is: the ratio of the volume of the bounding box of the component's 3D model to the volume of the sub-component's 3D model.

4. The high-efficiency continuous chemical vapor deposition monitoring system for new energy material preparation according to claim 3, characterized in that, The identification layer includes an identification module, a detection module, and an output module. The identification module is used to traverse the normalization results of the uniformity representation values ​​of each CVD spatial field region, identify the value of each result, and sort the results from smallest to largest based on the size of the result value, so that the result with the smaller result value is sorted first and the result with the larger result value is sorted last. The detection module is used to detect whether the components in the CVD spatial field region corresponding to each result are qualified based on the sorting of the normalization results of the uniformity representation values ​​of each CVD spatial field region. When it is detected that all components in the CVD spatial field region corresponding to a result are qualified, the process ends, and the result and all components in the CVD spatial field region corresponding to the result in the sorting queue are judged as qualified. The output module is used to receive the component qualification judgment results from the detection module and output the CVD spatial field region to which the qualified component belongs as the output content. Among them, the component qualification judgment result in the detection module and the output content of the output module are the contents of the component CVD qualification estimate.

5. The high-efficiency continuous chemical vapor deposition monitoring system for new energy material preparation according to claim 4, characterized in that, The identification layer operates after the component completes the CVD process. In the stage of determining whether the component is qualified, the surface of the component is bombarded by an electron beam to release characteristic X-rays with specific energy, which are then detected by an energy dispersive spectrometer to complete the determination. Specifically, when determining the affiliation of components in the CVD spatial field region, the affiliation is determined by the intersection spatial region obtained by stretching two sub-images corresponding to the CVD spatial field region.

6. The high-efficiency continuous chemical vapor deposition monitoring system for new energy material preparation according to claim 5, characterized in that, The segmentation module is interconnected with a sensing module via a local area network. The sensing module is interconnected with a thermal imaging module and an upload module via a local area network. The segmentation module is interconnected with an evaluation module and a normalization module via a local area network. The normalization module is interconnected with a recognition module via a local area network. The recognition module is interconnected with a detection module and an output module via a local area network.

Images