Artificial intelligence data analysis method and system for intelligent gold bonding wire
Through the analysis of the spectrum after processing of bonded wire, combined with deformation resistance model and impurity detection, a quantitative understanding of the resistivity changes of bonded wires was solved, and the problem of uncertain resistivity in the existing technology was achieved, and the precise classification and optimized use of bonded wires were achieved.
Patent Information
- Application Number
- CN202510375335.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-27
- Publication Date
- 2025-07-29
AI Technical Summary
The lack of understanding of the resistivity changes during the processing of bonded wires in the prior art, resulting in differences in the resistivity of the processed bonded wires, affecting the performance of the integrated circuit.
The processing spectra of bonded alloy wire is obtained through the detection device, the defect analysis model is established based on the spectra, the crystal defect data is calculated, and the deformation resistance model and annealing and impurity detection are combined to quantitatively analyze the impact of deformation, annealing and impurities on resistivity, and finally the total resistivity is obtained.
A quantitative understanding of the resistivity changes during bonded wire processing is achieved, and the appropriate bonded wires can be accurately classified and selected for different chip packaging, improving the conductive effect.
Smart Images
Figure CN120388657A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of semiconductor technology, and particularly to an artificial intelligence data analysis method and system for intelligent bonding wires. Background Art
[0002] A bonding wire is a fine metal wire inner lead used to achieve electrical connection between the input / output bonding points of the internal circuit of a chip and the internal contact points of a lead frame (or substrate) during the assembly of semiconductor devices and integrated circuits. The quality of the bonding effect directly affects the performance of the integrated circuit.
[0003] Currently, in the production process of bonding wires, processes such as wire drawing, cleaning, and annealing are usually involved. During the processing of bonding wires, the resistivity of the bonding wires will be affected. However, there is currently a lack of understanding of the change in resistivity during the processing of bonding wires. Summary of the Invention
[0004] In view of the above technical problems, embodiments of the present application propose an artificial intelligence data analysis method and system for intelligent bonding wires, which can solve the problem of the current lack of understanding of the change in resistivity during the processing of bonding wires.
[0005] Embodiments of the present application provide an artificial intelligence data analysis method for intelligent bonding wires, including:
[0006] Obtaining a processed spectrum of the bonding wire through a detection device;
[0007] Obtaining crystal defect data of the bonding wire based on the processed spectrum;
[0008] Inputting the crystal defect data into a deformation resistance model to obtain the deformation resistivity of the bonding wire after processing;
[0009] Obtaining a deformation-induced resistivity based on the deformation resistivity and an ideal resistivity.
[0010] In some embodiments, the obtaining crystal defect data of the bonding wire based on the processed spectrum includes:
[0011] Establishing a spectrum defect analysis model for analyzing the processed spectrum, where the spectrum defect analysis model includes a point defect analysis sub-model, a line defect analysis sub-model, and a surface defect analysis sub-model;
[0012] Inputting the processed spectrum into the point defect analysis sub-model, the line defect analysis sub-model, and the surface defect analysis sub-model respectively to obtain the point defect data, line defect data, and surface defect data of the bonding wire;
[0013] The crystal defect data is obtained based on the point defect data, the line defect data, and the surface defect data.
[0014] In some embodiments, the obtaining of the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity further includes:
[0015] Annealing the processed bonding wire;
[0016] Obtaining the post-annealing defect data of the bonding wire after the annealing treatment;
[0017] Inputting the post-annealing defect data into the deformation resistance model to obtain the post-annealing resistivity of the bonding wire;
[0018] Obtaining the annealing-influenced resistivity based on the deformation resistivity and the post-annealing resistivity.
[0019] In some embodiments, the obtaining of the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity further includes:
[0020] Using an impurity detector to photograph the impurity spectrum of the processed bonding wire;
[0021] Inputting the impurity spectrum into an impurity resistance influence model to obtain the impurity-influenced resistivity of the processed bonding wire.
[0022] In some embodiments, the inputting of the impurity spectrum into the impurity resistance influence model to obtain the impurity-influenced resistivity of the processed bonding wire includes:
[0023] Inputting the impurity spectrum into an impurity spectrum analysis model to obtain the impurity types and the impurity ratios;
[0024] Inputting the impurity types and the impurity ratios into the impurity resistance influence model to obtain the impurity-influenced resistivity of the processed bonding wire.
[0025] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire further includes:
[0026] Based on the deformation-influenced resistivity, the annealing-influenced resistivity, the impurity-influenced resistivity, and the ideal resistivity, obtaining the total resistivity of the bonding wire.
[0027] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire further includes:
[0028] Classifying and labeling the bonding wire based on the total resistivity to obtain a class-labeled bonding wire.
[0029] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire alloy is characterized by further comprising:
[0030] Obtain the drawing speed, drawing temperature, vibration level, and die aperture when drawing the bonding wire alloy;
[0031] Use the drawing speed, drawing temperature, the vibration level, the die aperture, and the crystal defect data to train the model to be trained, and obtain a crystal processing defect model.
[0032] In some embodiments, the step of using the drawing speed, drawing temperature, the vibration level, the die aperture, and the crystal defect data to train the model to be trained and obtain a crystal processing defect model further comprises:
[0033] Obtain the real-time drawing speed, real-time drawing temperature, real-time vibration level, and real-time die aperture when drawing the bonding wire alloy;
[0034] Set the target defect level after drawing the bonding wire alloy;
[0035] Based on the target defect level and the crystal processing defect model, adjust the real-time drawing speed, the real-time drawing temperature, and the real-time vibration level so that the drawn bonding wire alloy reaches the target defect level.
[0036] In a second aspect, an artificial intelligence data analysis system for an intelligent bonding wire alloy provided by an embodiment of the present application includes:
[0037] A spectrogram detection module for obtaining the spectrogram after processing the bonding wire alloy through a detection device;
[0038] The defect analysis module is used to obtain the crystal defect data of the bonding wire alloy based on the spectrogram after processing;
[0039] A resistance calculation module for inputting the crystal defect data into a deformation resistance model to obtain the deformation resistivity after processing the bonding wire alloy;
[0040] An influence calculation module for obtaining a deformation influence resistivity based on the deformation resistivity and the ideal resistivity.
[0041] The present application provides an artificial intelligence data analysis method for intelligent bonding wire alloys, including: obtaining the processed spectrogram of the bonding wire alloy through a detection device; obtaining the crystal defect data of the bonding wire alloy based on the processed spectrogram; inputting the crystal defect data into a deformation resistance model to obtain the deformation resistivity of the bonding wire alloy after processing; and obtaining the deformation-induced resistivity based on the deformation resistivity and the ideal resistivity, which can solve the problem of the lack of understanding of the change in resistivity during the processing of bonding wire alloys, and quantitatively understand the influence of the deformation generated during the processing of bonding wire alloys on the resistivity of the bonding wire alloys. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Hereinafter, the present invention will be described in more detail based on embodiments and with reference to the accompanying drawings.
[0043] Figure 1 is a flowchart of an artificial intelligence data analysis method for intelligent bonding wire alloys provided by an embodiment of the present invention;
[0044] Figure 2 is a schematic diagram of an artificial intelligence data analysis system for intelligent bonding wire alloys provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0045] The present invention will be further described below with reference to the accompanying drawings.
[0046] Wire bonding is a connection method that uses a thin metal wire and utilizes heat, pressure, and ultrasonic energy to tightly bond the metal lead to the substrate pad to achieve electrical interconnection between the chip and the substrate and information communication between chips.
[0047] Existing bonding wires (or bonding wire alloys) are all obtained by drawing a metal wire with a relatively large outer diameter through a wire drawing machine to the required diameter, and then through tempering (annealing) and winding operations. The bonding wire alloy equipment includes a wire drawing machine for drawing the metal wire, an annealing furnace for annealing the metal wire, and a winding device for winding the metal wire. During the wire drawing operation, the bonding wire alloy is forced through a die under an external force, compressing the cross-sectional area of the metal and obtaining the required cross-sectional area shape and size. During the processing of the bonding wire alloy, it will have an impact on the resistivity of the bonding wire alloy. However, currently, there is a lack of understanding of the change in resistivity during the processing of the bonding wire alloy, resulting in differences in the resistivity of the processed bonding wire alloys, which is not conducive to classifying and using the bonding wire alloys to achieve a better conductive effect.
[0048] As Figure 1 shown, in view of the above technical problems, the embodiments of the present application provide an artificial intelligence data analysis method for intelligent bonding wire alloys, including:
[0049] S101: Obtain the processed spectrogram of the gold bonding wire through a detection device;
[0050] S102: Obtain the crystal defect data of the gold bonding wire based on the processed spectrogram;
[0051] S103: Input the crystal defect data into a deformation resistance model to obtain the deformation resistivity after processing the gold bonding wire;
[0052] S104: Obtain the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity.
[0053] It should be noted that the detection device can be an electron microscope, an optical microscopic observation instrument, etc. The processed spectrogram is the spectrogram after the wire drawing process. The deformation resistance model can be a neural network model, a machine learning model, etc. The deformation resistance model is obtained through training and is used to calculate the change in the resistivity of the gold bonding wire caused by the generation of crystal defects according to the crystal defect data, that is, the deformation-influenced resistivity. Among them, obtaining the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity specifically means: subtracting the ideal resistivity from the deformation resistivity to obtain the deformation-influenced resistivity, and the ideal resistivity is the ideal resistivity of the gold bonding wire.
[0054] In some embodiments, obtaining the crystal defect data of the gold bonding wire based on the processed spectrogram includes:
[0055] Establish a spectrogram defect analysis model for analyzing the processed spectrogram, where the spectrogram defect analysis model includes a point defect analysis sub-model, a line defect analysis sub-model, and a plane defect analysis sub-model;
[0056] Input the processed spectrogram into the point defect analysis sub-model, the line defect analysis sub-model, and the plane defect analysis sub-model respectively to obtain the point defect data, line defect data, and plane defect data of the gold bonding wire;
[0057] Obtain the crystal defect data based on the point defect data, the line defect data, and the plane defect data.
[0058] It should be noted that crystal defects usually include point defects, line defects, and plane defects. Generally, the more the number of defects, the greater the resistivity. By establishing the spectrogram defect analysis model, the crystal defect data can be quickly and accurately obtained from the processed spectrogram. By dividing the spectrogram defect analysis model into the point defect analysis sub-model, the line defect analysis sub-model, and the plane defect analysis sub-model, the scale of the model can be reduced, and the calculation speed and accuracy can be improved. Among them, adding the point defect data, the line defect data, and the plane defect data can obtain the crystal defect data.
[0059] In some embodiments, obtaining the resistivity affected by deformation based on the resistivity of deformation and the ideal resistivity further includes:
[0060] Annealing the processed bonding wire;
[0061] Obtaining the defect data after annealing of the bonding wire after annealing treatment;
[0062] Inputting the defect data after annealing into the deformation resistance model to obtain the resistivity of the bonding wire after annealing;
[0063] Obtaining the resistivity affected by annealing based on the resistivity of deformation and the resistivity after annealing.
[0064] It should be noted that after drawing the bonding wire, the bonding wire usually needs to be annealed. The annealing treatment is usually carried out by an annealing furnace or an annealing device. The annealing treatment can repair the crystal defects of the bonding wire and restore the resistivity of the bonding wire to a certain extent. However, it is not clear to what extent the crystal defects (or resistivity) of the bonding wire can be restored after annealing treatment. By obtaining the defect data after annealing and obtaining the resistivity affected by annealing, the restoration effect of the annealing treatment can be quantitatively understood.
[0065] It should be noted that to obtain the defect data after annealing of the bonding wire after annealing treatment, an annealed spectrogram can be obtained first by a detection device, and then the defect data after annealing can be obtained by analyzing the annealed spectrogram through a neural network model. The defect data after annealing is input into the deformation resistance model to obtain the resistivity after annealing, where the resistivity affected by annealing is the restoration effect of the annealing treatment.
[0066] In some embodiments, obtaining the resistivity affected by deformation based on the resistivity of deformation and the ideal resistivity further includes:
[0067] Using an impurity detector to photograph the impurity spectrogram of the processed bonding wire;
[0068] Inputting the impurity spectrogram into an impurity resistance influence model to obtain the impurity influence resistivity of the processed bonding wire.
[0069] In some embodiments, inputting the impurity spectrogram into an impurity resistance influence model to obtain the impurity influence resistivity of the processed bonding wire includes:
[0070] Inputting the impurity spectrogram into an impurity spectrogram analysis model to obtain the types of impurities and the proportion of impurities;
[0071] Input the types of the impurities and the proportion of the impurities into the impurity resistance influence model to obtain the resistivity affected by the impurities after the bonding wire is processed.
[0072] It should be noted that during the processing of the bonding wire, not only cold deformation (i.e., generation of point defects, line defects and surface defects) will occur, but also some impurities (such as other metals, sweat, etc.) will be mixed into the bonding wire, thus affecting the resistivity of the bonding wire. By obtaining the impurity spectrogram and the resistivity affected by the impurities, the influence of the impurities on the resistivity of the bonding wire can be quantitatively understood. Among them, the impurity detector can be a microscope, an ultrasonic detector, etc., and the impurity resistance influence model can be a neural network model, a machine learning model, etc. The present application does not make specific limitations on this.
[0073] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire further includes:
[0074] Based on the resistivity affected by deformation, the resistivity affected by annealing, the resistivity affected by impurities and the ideal resistivity, obtain the total resistivity of the bonding wire.
[0075] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire further includes:
[0076] Classify and label the bonding wire based on the total resistivity to obtain a class-labeled bonding wire.
[0077] It should be noted that the obtaining of the total resistivity of the bonding wire based on the resistivity affected by deformation, the resistivity affected by annealing, the resistivity affected by impurities and the ideal resistivity is specifically: ideal resistivity + resistivity affected by deformation (which can be positive or negative) + resistivity affected by annealing (which can be positive or negative) + resistivity affected by impurities (which can be positive or negative) = total resistivity.
[0078] It should be noted that after obtaining various factors (i.e., crystal defects, impurities, annealing treatment) that may affect the resistivity of the bonding wire, the final resistivity (i.e., the total resistivity) of the bonding wire can be obtained. By classifying the bonding wire according to the total resistivity, bonding wires with different resistivity levels can be obtained, that is, the class-labeled bonding wires. Bonding wires with different resistivity levels can meet the chip packaging requirements of different uses, and bonding wires with corresponding resistivity levels can be selected according to the chip packaging use requirements for different chips.
[0079] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire further includes:
[0080] Obtain the drawing speed, drawing temperature, vibration level and die aperture when drawing the bonding wire.
[0081] Use the wire drawing speed, wire drawing temperature, the vibration level, the die aperture, and the crystal defect data to train the model to be trained, and obtain a crystal processing defect model.
[0082] It should be noted that the influencing factors of the crystal defects of the bonding wire usually include the wire drawing speed, temperature, the vibration level (i.e., the magnitude of vibration) it bears, and the die aperture. That is, the faster the wire drawing speed, the greater the cold deformation amount and the more crystal defects of the bonding wire; the higher the temperature, the more crystal defects the bonding wire generates; the higher the vibration level, the more crystal defects the bonding wire generates; the die aperture mainly determines the deformation amount of the bonding wire; by establishing the crystal processing defect model, it is possible to calculate the crystal defect data (i.e., including the defect type and the number of defects) of the bonding wire after the wire drawing process according to the wire drawing speed, wire drawing temperature, vibration level, and die aperture.
[0083] In some embodiments, the step of using the wire drawing speed, wire drawing temperature, the vibration level, the die aperture, and the crystal defect data to train the model to be trained and obtain a crystal processing defect model further includes:
[0084] Obtain the real-time wire drawing speed, real-time wire drawing temperature, real-time vibration level, and real-time die aperture for wire drawing the bonding wire.
[0085] Set the target defect level after wire drawing the bonding wire.
[0086] Based on the target defect level and the crystal processing defect model, adjust the real-time wire drawing speed, real-time wire drawing temperature, and real-time vibration level so that the bonding wire after wire drawing reaches the target defect level.
[0087] It should be noted that due to the influence of various uncertain factors, such as too high temperature during wire drawing, abnormal wire drawing speed, deterioration of the bonding wire processing equipment, abnormal vibration, etc., all will affect the crystal defects of the bonding wire. And when processing the bonding wire, it is necessary to ensure that the defect level (or resistivity) of the bonding wire meets the set requirements (i.e., the target defect level). Therefore, based on the target defect level, the real-time wire drawing temperature, and / or the real-time vibration level, and / or the real-time wire drawing speed can be adjusted through the crystal construction defect model to ensure the defect level of the bonding wire.
[0088] It should be noted that the real-time drawing speed of the bonding wire can be obtained through a speed sensor, the real-time drawing temperature can be obtained through a temperature sensor, and the real-time vibration level can be obtained through a vibration sensor. The die aperture is usually determined in advance. The crystal processing defect model is deployed in the bonding wire processing equipment (drawing equipment, cooling equipment, annealing equipment, etc.). Generally, the smaller the real-time drawing speed, the lower the real-time drawing temperature, and the lower the real-time vibration level, the fewer crystal defects of the bonding wire. When making adjustments, the real-time vibration level is usually difficult to control. Therefore, it can be set to first adjust the real-time drawing temperature. Adjusting the real-time drawing speed temperature is usually achieved by increasing or decreasing the cooling capacity of the cooling device (or cooling equipment). In this way, both the drawing speed can be ensured and the crystal defects of the bonding wire will not increase. When the target defect level cannot be achieved by adjusting the real-time drawing temperature, the real-time drawing speed can be adjusted to reduce the crystal defects of the bonding wire and ensure that the target defect level can be achieved.
[0089] It should be noted that the real-time vibration level needs to be monitored in real time. If the real-time vibration level is greater than the vibration threshold, at this time, adjusting the real-time drawing speed and real-time drawing temperature cannot ensure that the bonding wire can reach the target defect level. This situation is usually due to equipment failure or serious deterioration. At this time, a vibration abnormality reminder can be issued to remind the staff to check and repair to reduce the real-time vibration level, so as to ensure that the target defect level can be achieved. It can also increase the number of dies to slow down the reduction speed of the die aperture to reach the target defect level.
[0090] In summary, the present application provides an artificial intelligence data analysis method for intelligent bonding wires, including: obtaining the processed spectrum of the bonding wire through a detection device; obtaining the crystal defect data of the bonding wire based on the processed spectrum; inputting the crystal defect data into a deformation resistance model to obtain the deformation resistivity of the bonding wire after processing; and obtaining the deformation influence resistivity based on the deformation resistivity and the ideal resistivity, which can solve the problem of the lack of understanding of the resistivity change during the processing of the bonding wire and quantitatively understand the influence of the deformation generated during the processing of the bonding wire on the resistivity of the bonding wire.
[0091] Second, as Figure 2 shown, an artificial intelligence data analysis system for intelligent bonding wires provided by an embodiment of the present application includes:
[0092] A spectrum detection module 210, configured to obtain the processed spectrum of the bonding wire through a detection device;
[0093] A defect analysis module 220, configured to obtain crystal defect data of the bonding wire based on the processed spectrogram;
[0094] A resistance calculation module 230, configured to input the crystal defect data into a deformation resistance model to obtain the deformation resistivity of the bonding wire after processing;
[0095] An influence calculation module 240, configured to obtain a deformation influence resistivity based on the deformation resistivity and the ideal resistivity.
[0096] In some embodiments, obtaining the crystal defect data of the bonding wire based on the processed spectrogram includes:
[0097] Establishing a spectrogram defect analysis model for analyzing the processed spectrogram, where the spectrogram defect analysis model includes a point defect analysis sub-model, a line defect analysis sub-model, and a surface defect analysis sub-model;
[0098] Inputting the processed spectrogram into the point defect analysis sub-model, the line defect analysis sub-model, and the surface defect analysis sub-model respectively to obtain the point defect data, line defect data, and surface defect data of the bonding wire;
[0099] Obtaining the crystal defect data based on the point defect data, the line defect data, and the surface defect data.
[0100] In some embodiments, obtaining the deformation influence resistivity based on the deformation resistivity and the ideal resistivity further includes:
[0101] Annealing the processed bonding wire;
[0102] Obtaining post-annealing defect data of the bonding wire after the annealing treatment;
[0103] Inputting the post-annealing defect data into the deformation resistance model to obtain the post-annealing resistivity of the bonding wire;
[0104] Obtaining an annealing influence resistivity based on the deformation resistivity and the post-annealing resistivity.
[0105] In some embodiments, obtaining the deformation influence resistivity based on the deformation resistivity and the ideal resistivity further includes:
[0106] Using an impurity detector to photograph an impurity spectrogram of the bonding wire after processing;
[0107] Inputting the impurity spectrogram into an impurity resistance influence model to obtain the impurity influence resistivity of the bonding wire after processing.
[0108] In some embodiments, inputting the impurity spectrum into the impurity resistance influence model to obtain the resistivity affected by impurities after the bonding wire is processed includes:
[0109] Inputting the impurity spectrum into an impurity spectrum analysis model to obtain the types and proportions of impurities;
[0110] Inputting the types and proportions of impurities into the impurity resistance influence model to obtain the resistivity affected by impurities after the bonding wire is processed.
[0111] In some embodiments, the influence calculation module 240 is further configured to:
[0112] Based on the resistivity affected by deformation, the resistivity affected by annealing, the resistivity affected by impurities, and the ideal resistivity, obtain the total resistivity of the bonding wire.
[0113] In some embodiments, the influence calculation module 240 is further configured to:
[0114] Classify and label the bonding wire based on the total resistivity to obtain a classified and labeled bonding wire.
[0115] In some embodiments, the artificial intelligence data analysis method for the intelligent bonding wire is further characterized by including:
[0116] Obtain the drawing speed, drawing temperature, vibration level, and die aperture when drawing the bonding wire;
[0117] Use the drawing speed, drawing temperature, vibration level, die aperture, and the crystal defect data to train a to-be-trained model to obtain a crystal processing defect model.
[0118] In some embodiments, the step of using the drawing speed, drawing temperature, vibration level, die aperture, and the crystal defect data to train a to-be-trained model to obtain a crystal processing defect model further includes:
[0119] Obtain the real-time drawing speed, real-time drawing temperature, real-time vibration level, and real-time die aperture when drawing the bonding wire;
[0120] Set a target defect level after drawing the bonding wire;
[0121] Based on the target defect level and the crystal processing defect model, adjust the real-time drawing speed, real-time drawing temperature, and real-time vibration level so that the bonding wire after drawing reaches the target defect level.
[0122] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (devices, systems), and / or computer program products according to embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.
[0123] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.
[0124] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.
[0125] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present application. It should be understood that the above are only specific embodiments of the present application and are not used to limit the protection scope of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
Claims
1. An artificial intelligence data analysis method for intelligent bonding alloy wires, characterized in that, Including: Obtaining the processed spectrogram of the gold bonding wire through a detection device; Obtaining the crystal defect data of the gold bonding wire based on the processed spectrogram; Inputting the crystal defect data into a deformation resistance model to obtain the deformation resistivity of the gold bonding wire after processing; Obtaining the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity.
2. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 1, wherein The obtaining the crystal defect data of the gold bonding wire based on the processed spectrogram includes: Establishing a spectrogram defect analysis model for analyzing the processed spectrogram, wherein the spectrogram defect analysis model includes a point defect analysis sub-model, a line defect analysis sub-model, and a surface defect analysis sub-model; Inputting the processed spectrogram into the point defect analysis sub-model, the line defect analysis sub-model, and the surface defect analysis sub-model respectively to obtain the point defect data, the line defect data, and the surface defect data of the gold bonding wire; Obtaining the crystal defect data based on the point defect data, the line defect data, and the surface defect data.
3. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 1, characterized in that, The obtaining the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity further includes: Performing annealing treatment on the processed gold bonding wire; Obtaining the post-annealing defect data of the gold bonding wire after the annealing treatment; Inputting the post-annealing defect data into the deformation resistance model to obtain the post-annealing resistivity of the gold bonding wire; Obtaining the annealing-influenced resistivity based on the deformation resistivity and the post-annealing resistivity.
4. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 3, characterized in that, The obtaining the deformation-influenced resistivity based on the deformation resistivity and the ideal resistivity further includes: Using an impurity detector to photograph the impurity spectrogram of the gold bonding wire after processing; Inputting the impurity spectrogram into an impurity resistance influence model to obtain the impurity-influenced resistivity of the gold bonding wire after processing.
5. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 4, characterized in that, The inputting the impurity spectrogram into an impurity resistance influence model to obtain the impurity-influenced resistivity of the gold bonding wire after processing includes: Inputting the impurity spectrogram into an impurity spectrogram analysis model to obtain the impurity types and impurity ratios; Inputting the impurity types and the impurity ratios into the impurity resistance influence model to obtain the impurity-influenced resistivity of the gold bonding wire after processing.
6. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 4, wherein Further including: Obtaining the total resistivity of the gold bonding wire based on the deformation-influenced resistivity, the annealing-influenced resistivity, the impurity-influenced resistivity, and the ideal resistivity.
7. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 6, characterized in that, Further including: Classifying and marking the gold bonding wire based on the total resistivity to obtain a classified and marked gold bonding wire.
8. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 1, characterized in that, Further including: Obtaining the drawing speed, drawing temperature, vibration level, and die aperture when drawing the gold bonding wire; Using the drawing speed, drawing temperature, vibration level, die aperture, and the crystal defect data to train a model to be trained to obtain a crystal processing defect model.
9. The artificial intelligence data analysis method of the intelligent key alloy wire according to claim 8, characterized in that, The using the drawing speed, drawing temperature, vibration level, die aperture, and the crystal defect data to train a model to be trained to obtain a crystal processing defect model further includes: Obtaining the real-time drawing speed, real-time drawing temperature, real-time vibration level, and real-time die aperture when drawing the gold bonding wire; Setting a target defect level after drawing the gold bonding wire; Adjust the real-time wire drawing speed, the real-time wire drawing temperature, and the real-time vibration level based on the target defect level and the crystal processing defect model, so that the bonded wire after wire drawing reaches the target defect level.
10. An artificial intelligence data analysis system for intelligent bonding alloy wires, characterized in that, Including: A spectrogram detection module for obtaining the processed spectrogram of the bonded wire through a detection device; A defect analysis module for obtaining the crystal defect data of the bonded wire based on the processed spectrogram; A resistance calculation module for inputting the crystal defect data into a deformation resistance model to obtain the deformation resistivity of the bonded wire after processing; An influence calculation module for obtaining a deformation influence resistivity based on the deformation resistivity and the ideal resistivity.