A method and system for defect detection of an integrated circuit package

Abstract

The present invention discloses a method and system for defect detection of integrated circuit packaging. The method includes: obtaining a scanned image of an integrated circuit packaging sample, analyzing the scanned image to determine the defect area of the integrated circuit packaging sample; collecting multi-modal data of the defect area and performing fusion to obtain comprehensive defect information of the integrated circuit packaging sample; statistically analyzing the defect types of the integrated circuit packaging sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information; screening out defects to be repaired and negligible defects according to the defect types and the performance impact parameters caused by each type of defect and generating a visual report. By locating the defect area and detecting the multi-modal fusion data of the defect area to determine the comprehensive defect information, various types of defects of the integrated circuit packaging sample can be determined in all directions, ensuring the accuracy, reliability and detection efficiency of defect detection.

Description

Technical Field

[0001] The present invention relates to the technical field of circuit defect detection, and particularly to a method and system for detecting defects in integrated circuit packaging. Background Art

[0002] At present, integrated circuit packaging is an important link in the integrated circuit manufacturing process, and its quality directly affects the performance and reliability of integrated circuits. With the development of integrated circuits towards high density, high performance, and miniaturization, the packaging process has become increasingly complex, and it has become more and more difficult to detect packaging defects. Existing defect detection methods are all based on single-site visual inspection by industrial cameras and defect detection and positioning based on single-modal data. Their detection capabilities for different types of defects are limited, and high-precision defect detection and filtering cannot be achieved, reducing the detection efficiency. Summary of the Invention

[0003] In view of the problems shown above, the present invention provides a method and system for detecting defects in integrated circuit packaging to solve the problems in the background art that single-site visual inspection is carried out by industrial cameras and defect detection and positioning are based on single-modal data, resulting in limited detection capabilities for different types of defects, inability to achieve high-precision defect detection and filtering, and reduced detection efficiency.

[0004] A method for detecting defects in integrated circuit packaging includes the following steps:

[0005] Obtain a scanned image of an integrated circuit packaging sample, and analyze the scanned image to determine the defect area of the integrated circuit packaging sample;

[0006] Collect multi-modal data of the defect area and perform fusion to obtain comprehensive defect information of the integrated circuit packaging sample;

[0007] According to the comprehensive defect information, count the defect types of the integrated circuit packaging sample and the performance impact parameters caused by each type of defect;

[0008] Screen out defects to be repaired and negligible defects according to the defect types and the performance impact parameters caused by each type of defect, and generate a visualization report.

[0009] Preferably, the step of obtaining a scanned image of an integrated circuit packaging sample and analyzing the scanned image to determine the defect area of the integrated circuit packaging sample includes:

[0010] Determine the structural parameters of the integrated circuit packaging sample, determine the layering information of the integrated circuit packaging sample according to the structural parameters, and determine the scanning method according to the layering information. The scanning methods include: internal scanning, surface scanning, and layered scanning;

[0011] Scan the integrated circuit package sample in a scanning manner to obtain a scanning image, and perform image enhancement preprocessing on the scanning image;

[0012] Obtain the defect characteristics of different types of defects, determine the recognition elements of different types of defects according to the defect characteristics, and determine the recognition and positioning strategies of different types of defects according to the recognition elements;

[0013] Analyze the scanning image through image processing software based on the recognition and positioning strategies of different types of defects, and identify and locate the defect area of the integrated circuit package sample according to the analysis results.

[0014] Preferably, collect and fuse the multi-modal data of the defect area to obtain the comprehensive defect information of the integrated circuit package sample, including:

[0015] Determine the regional attributes of the defect area, where the regional attributes include: internal area, surface area, and delamination area, and determine the data collection method according to the regional attributes, where:

[0016] Collect the surface topography of the surface defect area through a high-resolution optical microscope, and determine the two-dimensional optical surface image of the defect area according to the surface topography;

[0017] Scan the internal defect area with an ultrasonic probe to obtain the ultrasonic defect reflection data of the internal defect area;

[0018] Perform X-ray imaging on the delamination defect area to obtain the three-dimensional X-ray image of the delamination defect area;

[0019] Spatially align and register the two-dimensional optical image, the ultrasonic defect reflection data, and the three-dimensional X-ray image to generate multi-modal fusion data;

[0020] Extract the current defect characteristics from the multi-modal fusion data through a feature extraction algorithm, and generate the comprehensive defect information of the integrated circuit package sample based on the current defect characteristics using data fusion technology.

[0021] Preferably, count the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information, including:

[0022] Determine the defect type distribution of the integrated circuit package sample according to the comprehensive defect information, and determine the target defect type corresponding to each defect area according to the defect type distribution;

[0023] Determine the defect depth information and defect size corresponding to each defect area based on the multi-modal fusion data through a defect depth analysis model based on a deep learning algorithm;

[0024] Determine the defect morphology and position parameters of each defect area based on the defect depth information and defect size, obtain the packaging function information of the integrated circuit package sample, and determine the performance parameters to be analyzed according to the packaging function information;

[0025] Obtain experimental defect analysis data according to the performance parameters to be analyzed, and construct a defect-performance relationship model based on the experimental defect analysis data;

[0026] Based on the defect-performance relationship model, determine the performance impact correlation parameters and performance impact descriptive statistical parameters of the target defect type for each defect area based on the defect morphology and position parameters of each defect area.

[0027] Preferably, screen out the defects to be repaired and negligible defects according to the defect type and the performance impact parameters caused by each type of defect, and generate a visualization report, including:

[0028] Determine the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect according to the defect type and the performance impact parameters caused by each type of defect;

[0029] Determine the screening criteria for the defects to be repaired and negligible defects respectively, and determine the reasonable parameter intervals and reasonable parameter descriptions of each performance index according to the screening criteria;

[0030] Divide each type of defect into defects to be repaired and negligible defects according to the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect, as well as the reasonable parameter intervals and reasonable parameter descriptions of the performance indexes of the defects to be repaired and negligible defects respectively;

[0031] Generate repair suggestions according to the division result and the target defect morphology of the defect to be repaired, and generate a visualization report according to the repair suggestions.

[0032] A defect detection system for integrated circuit packaging, the system includes:

[0033] An analysis module, configured to obtain a scanned image of an integrated circuit package sample, and analyze the scanned image to determine the defect area of the integrated circuit package sample;

[0034] An acquisition module, configured to collect and fuse multi-modal data of the defect area to obtain comprehensive defect information of the integrated circuit package sample;

[0035] A statistics module, configured to count the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information;

[0036] A generation module, configured to screen out the defects to be repaired and negligible defects according to the defect type and the performance impact parameters caused by each type of defect, and generate a visualization report.

[0037] Preferably, the analysis module includes:

[0038] A first determination sub-module, configured to determine the structural parameters of the integrated circuit package sample, determine the layering information of the integrated circuit package sample according to the structural parameters, and determine the scanning method according to the layering information, where the scanning method includes: internal scanning, surface scanning, and layer scanning;

[0039] A scanning sub-module, configured to scan the integrated circuit package sample by the scanning method, obtain a scanned image, and perform image enhancement preprocessing on the scanned image;

[0040] A second determination sub-module, configured to obtain the defect characteristics of different types of defects, determine the identification elements of different types of defects according to the defect characteristics, and determine the identification and positioning strategies of different types of defects according to the identification elements;

[0041] An analysis sub-module, configured to analyze the scanned image through image processing software based on the identification and positioning strategies of different types of defects, and identify and locate the defect area of the integrated circuit package sample according to the analysis result.

[0042] Preferably, the acquisition module includes:

[0043] A third determination sub-module, configured to determine the regional attributes of the defect area, where the regional attributes include: internal area, surface area, and layer area, and determine the data acquisition method according to the regional attributes, where:

[0044] Collect the surface topography of the surface defect area through a high-resolution optical microscope, and determine the optical surface two-dimensional image of the defect area according to the surface topography;

[0045] Scan the internal defect area with an ultrasonic probe to obtain the ultrasonic defect reflection data of the internal defect area;

[0046] Perform X-ray imaging on the layer defect area to obtain the X-ray three-dimensional image of the layer defect area;

[0047] A first generation sub-module, configured to perform spatial alignment and registration on the optical two-dimensional image, the ultrasonic defect reflection data, and the X-ray three-dimensional image to generate multi-modal fusion data;

[0048] A second generation sub-module, configured to extract the current defect characteristics from the multi-modal fusion data through a feature extraction algorithm, and generate the comprehensive defect information of the integrated circuit package sample based on the current defect characteristics by using data fusion technology.

[0049] Preferably, the statistics module includes:

[0050] The fourth determination sub-module is configured to determine the defect type distribution of the integrated circuit package sample according to the comprehensive defect information, and determine the target defect type corresponding to each defect area according to the defect type distribution;

[0051] The fifth determination sub-module is configured to determine the defect depth information and defect size corresponding to each defect area based on the multi-modal fusion data through a defect depth analysis model based on a deep learning algorithm;

[0052] The sixth determination sub-module is configured to determine the defect morphology and position parameters of each defect area according to the defect depth information and defect size, obtain the package function information of the integrated circuit package sample, and determine the performance parameters to be analyzed according to the package function information;

[0053] The construction sub-module is configured to obtain experimental defect analysis data according to the performance parameters to be analyzed, and construct a defect-performance relationship model according to the experimental defect analysis data;

[0054] The seventh determination sub-module is configured to determine the performance impact correlation parameters and performance impact descriptive statistical parameters of the target defect type of each defect area based on the defect morphology and position parameters of each defect area through the defect-performance relationship model.

[0055] Preferably, the generation module includes:

[0056] The eighth determination sub-module is configured to determine the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect according to the defect type and the performance impact parameters caused by each type of defect;

[0057] The ninth determination sub-module is configured to determine the screening criteria for the defects to be repaired and the negligible defects respectively, and determine the reasonable parameter interval and reasonable parameter description of each performance index according to the screening criteria;

[0058] The division sub-module is configured to divide each type of defect into defects to be repaired and negligible defects according to the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect and the reasonable parameter interval and reasonable parameter description of the performance indexes of the defects to be repaired and the negligible defects respectively;

[0059] The third generation sub-module is configured to generate a repair suggestion according to the division result and the target defect morphology of the defect to be repaired, and generate a visualization report according to the repair suggestion.

[0060] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in the written specification and the drawings.

[0061] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0062] The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, but do not constitute a limitation to the present invention.

[0063] Figure 1 It is a flowchart of the working process of a method for defect detection of an integrated circuit package provided by the present invention;

[0064] Figure 2 It is another flowchart of the working process of a method for defect detection of an integrated circuit package provided by the present invention;

[0065] Figure 3 It is a schematic structural diagram of a system for defect detection of an integrated circuit package provided by the present invention;

[0066] Figure 4 It is a schematic structural diagram of an analysis module in a system for defect detection of an integrated circuit package provided by the present invention. Detailed Embodiments

[0067] Here, the exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0068] Currently, integrated circuit packaging is an important link in the integrated circuit manufacturing process, and its quality directly affects the performance and reliability of integrated circuits. With the development of integrated circuits towards high density, high performance, and miniaturization, the packaging process has become increasingly complex, and it has become more and more difficult to detect packaging defects. Existing defect detection methods all perform single-site visual inspection through industrial cameras and perform defect detection and positioning based on single-modal data. Their detection capabilities for different types of defects are limited, and they cannot achieve high-precision defect detection and filtering, reducing the detection efficiency. To solve the above problems, this embodiment discloses a method for defect detection of an integrated circuit package.

[0069] A method for defect detection of an integrated circuit package, as Figure 1 shown, includes the following steps:

[0070] Step S101, obtain a scanned image of an integrated circuit package sample, and analyze the scanned image to determine the defect area of the integrated circuit package sample;

[0071] Step S102: Collect multimodal data of the defective area and fuse them to obtain the comprehensive defect information of the integrated circuit package sample;

[0072] Step S103: According to the comprehensive defect information, count the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect;

[0073] Step S104: Screen out the defects to be repaired and the ignorable defects according to the defect types and the performance impact parameters caused by each type of defect, and generate a visualization report;

[0074] In this embodiment, the defect types include: packaging defects, assembly defects, and material defects. Among them, the packaging defects include: poor wire bonding, chip cracking, voids in packaging materials, etc.; the assembly defects include solder joint voids, solder ball bridging, component misalignment, etc.; the material defects include substrate cracks, solder impurities, delamination of encapsulant, etc.

[0075] The working principle of the above technical solution is as follows: Obtain the scanned image of the integrated circuit package sample, analyze the scanned image to determine the defective area of the integrated circuit package sample; collect multimodal data of the defective area and fuse them to obtain the comprehensive defect information of the integrated circuit package sample; according to the comprehensive defect information, count the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect; screen out the defects to be repaired and the ignorable defects according to the defect types and the performance impact parameters caused by each type of defect, and generate a visualization report.

[0076] The beneficial effects of the above technical solution are as follows: By locating the defective area and detecting the multimodal fusion data of the defective area to determine the comprehensive defect information, various types of defects of the integrated circuit package sample can be determined in all aspects, ensuring the accuracy, reliability, and detection efficiency of defect detection. Further, by evaluating the performance impact of each type of detected defect and then screening out the ignorable defects, defects with negligible performance impact can be filtered out, improving the defect repair efficiency, and solving the problem mentioned in the prior art that through a single-station vision inspection by an industrial camera and defect detection and positioning based on single-modal data, its detection ability for different types of defects is limited, unable to achieve high-precision defect detection and filtering, and reducing the detection efficiency.

[0077] In one embodiment, as Figure 2 shown, the obtaining the scanned image of the integrated circuit package sample and analyzing the scanned image to determine the defective area of the integrated circuit package sample includes:

[0078] Step S201: Determine the structural parameters of the integrated circuit package sample, determine the layering information of the integrated circuit package sample according to the structural parameters, and determine the scanning method according to the layering information. The scanning methods include: internal scanning, surface scanning, and layer-by-layer scanning;

[0079] Step S202: Scan the integrated circuit package sample by the scanning method, obtain the scanned image, and perform image enhancement preprocessing on the scanned image;

[0080] Step S203: Obtain the defect characteristics of different types of defects, determine the recognition elements of different types of defects according to the defect characteristics, and determine the recognition and positioning strategies of different types of defects according to the recognition elements;

[0081] Step S204: Analyze the scanned image by image processing software based on the recognition and positioning strategies of different types of defects, and identify and locate the defect area of the integrated circuit package sample according to the analysis results.

[0082] The beneficial effects of the above technical solutions are as follows: By collecting different scanning methods to perform systematic scanning on the integrated circuit package sample, defect scanning can be achieved comprehensively, improving the scanning efficiency and comprehensiveness. Further, by positioning the defect area based on the recognition and positioning strategies of different types of defects, accurate defect area positioning can be quickly performed through reference features, improving the positioning accuracy and positioning efficiency.

[0083] In this embodiment, obtaining the defect characteristics of different types of defects, determining the recognition elements of different types of defects according to the defect characteristics, and determining the recognition and positioning strategies of different types of defects according to the recognition elements includes:

[0084] Obtain the defect image dominant characteristics and defect text description characteristics of different types of defects, and determine the defect characteristic point distribution map according to the defect image dominant characteristics and defect text description characteristics;

[0085] Determine the affine abnormal frequency component according to the defect characteristic point distribution map through the affine transformation matrix, compare the affine abnormal frequency component with the standard frequency component, and obtain the comparison result;

[0086] Determine the recognition elements of different types of defects according to the comparison result and the intensification elements of the abnormal frequency component, and obtain the shallow detection expression attribute and deep detection expression attribute of the recognition elements;

[0087] Respectively, the characteristic change arrays of the shallow detection expression attribute and the deep detection expression attribute, and determine the low-frequency scanning parameter and high-matching parameter for the recognition elements based on the characteristic change arrays;

[0088] Determine the scanning form according to the low-frequency scanning parameter, and determine the low-frequency parameter recognition and positioning strategy for the recognition elements based on the scanning form;

[0089] Obtain the parametric change rules of high-frequency pairing parameters, determine the pairing mechanism according to the parametric change rules, and determine the high-frequency parameter identification and positioning strategy for the identification elements based on the pairing mechanism;

[0090] Integrate the low-frequency parameter identification and positioning strategy and the high-frequency parameter identification and positioning strategy for the identification elements to generate the identification and positioning strategies for different types of defects.

[0091] The beneficial effects of the above technical solution are: By determining the shallow-level and deep-level parametric identification and positioning strategies of the identification elements of different types of defects for comprehensive defect identification and positioning, the identification accuracy and reliability for various types of defects can be comprehensively ensured, improving the accuracy and reliability of defect positioning and identification.

[0092] In one embodiment, collect and fuse the multi-modal data of the defect area to obtain the comprehensive defect information of the integrated circuit package sample, including:

[0093] Determine the regional attributes of the defect area, where the regional attributes include: internal area, surface area, and delamination area, and determine the data collection method according to the regional attributes, where:

[0094] Collect the surface topography of the surface defect area through a high-resolution optical microscope, and determine the two-dimensional optical surface image of the defect area according to the surface topography;

[0095] Scan the internal defect area with an ultrasonic probe to obtain the ultrasonic defect reflection data of the internal defect area;

[0096] Perform X-ray imaging on the delamination defect area to obtain the three-dimensional X-ray image of the delamination defect area;

[0097] Perform spatial alignment and registration on the two-dimensional optical image, the ultrasonic defect reflection data, and the three-dimensional X-ray image to generate multi-modal fusion data;

[0098] Extract the current defect features from the multi-modal fusion data through a feature extraction algorithm, and generate the comprehensive defect information of the integrated circuit package sample based on the current defect features using data fusion technology.

[0099] The beneficial effects of the above technical solution are: By performing multi-modal data collection and fusion, the structural parameters of each structure of the circuit package sample can be comprehensively obtained, thereby detecting the defect information, improving the defect detection accuracy and efficiency.

[0100] In one embodiment, according to the comprehensive defect information, count the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect, including:

[0101] Determine the defect type distribution of the integrated circuit package sample according to the comprehensive defect information, and determine the target defect type corresponding to each defect area according to the defect type distribution;

[0102] Based on the defect depth analysis model based on the deep learning algorithm, determine the defect depth information and defect size corresponding to each defect area based on the multi-modal fusion data;

[0103] Determine the defect morphology and position parameters of each defect area according to the defect depth information and defect size, obtain the package function information of the integrated circuit package sample, and determine the performance parameters to be analyzed according to the package function information;

[0104] Obtain the experimental defect analysis data according to the performance parameters to be analyzed, and construct a defect-performance relationship model according to the experimental defect analysis data;

[0105] Based on the defect-performance relationship model, determine the performance impact correlation parameters and performance impact descriptive statistical parameters of the target defect type of each defect area based on the defect morphology and position parameters of each defect area.

[0106] The beneficial effects of the above technical solutions are as follows: By constructing a defect-performance relationship model, the performance impact evaluation process can be quickly carried out according to the defect forms of various types of defects, improving the evaluation efficiency and accuracy. Further, by determining the performance impact correlation parameters and performance impact descriptive statistical parameters of the target defect type of each defect area, the performance impact description parameters and correlation parameter impact description parameters of various types of defects on the operation performance of the integrated circuit package sample can be accurately determined, comprehensively determining the impact of performance, laying a reference condition for subsequent defect screening and repair suggestion generation, and improving the practicability.

[0107] In one embodiment, screen the defects to be repaired and negligible defects according to the defect type and the performance impact parameters caused by each type of defect, and generate a visualization report, including:

[0108] Determine the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect according to the defect type and the performance impact parameters caused by each type of defect;

[0109] Determine the screening criteria for the defects to be repaired and negligible defects respectively, and determine the reasonable parameter range and reasonable parameter description of each performance index according to the screening criteria;

[0110] Divide each type of defect into defects to be repaired and negligible defects according to the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect and the reasonable parameter range and reasonable parameter description of the performance indexes of the defects to be repaired and negligible defects respectively;

[0111] Generate repair suggestions based on the division results and the target defect morphology of the defects to be repaired, and generate a visualization report according to the repair suggestions.

[0112] The beneficial effects of the above technical solution are as follows: The detected defects in each defect area can be divided intuitively and quickly, and then repair suggestions can be generated quickly, improving work efficiency.

[0113] In one embodiment, this embodiment also discloses a defect detection system for integrated circuit packaging, as Figure 3 shown. The system includes:

[0114] An analysis module 301, configured to obtain a scanned image of an integrated circuit packaging sample, and analyze the scanned image to determine the defect area of the integrated circuit packaging sample;

[0115] An acquisition module 302, configured to collect multi-modal data of the defect area and perform fusion to obtain comprehensive defect information of the integrated circuit packaging sample;

[0116] A statistics module 303, configured to count the defect types of the integrated circuit packaging sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information;

[0117] A generation module 304, configured to screen out defects to be repaired and negligible defects according to the defect types and the performance impact parameters caused by each type of defect, and generate a visualization report.

[0118] The working principle and beneficial effects of the above technical solution have been described in the method embodiment, and will not be repeated here.

[0119] In one embodiment, as Figure 4 shown, the analysis module 301 includes:

[0120] A first determination sub-module 3011, configured to determine the structural parameters of the integrated circuit packaging sample, determine the layering information of the integrated circuit packaging sample according to the structural parameters, and determine the scanning method according to the layering information. The scanning method includes: internal scanning, surface scanning, and layer-by-layer scanning;

[0121] A scanning sub-module 3012, configured to scan the integrated circuit packaging sample by the scanning method, obtain a scanned image, and perform image enhancement preprocessing on the scanned image;

[0122] A second determination sub-module 3013, configured to obtain the defect characteristics of different types of defects, determine the recognition elements of different types of defects according to the defect characteristics, and determine the recognition and positioning strategies of different types of defects according to the recognition elements;

[0123] The analysis sub-module 3014 is used to analyze the scanned image through image processing software based on the identification and location strategies for different types of defects, and identify and locate the defect areas of the integrated circuit package sample according to the analysis results.

[0124] In one embodiment, the acquisition module includes:

[0125] The third determination sub-module is used to determine the regional attributes of the defect area. The regional attributes include: internal area, surface area, and delamination area, and determine the data acquisition method according to the regional attributes, where:

[0126] Collect the surface topography of the surface defect area through a high-resolution optical microscope, and determine the two-dimensional optical surface image of the defect area according to the surface topography;

[0127] Scan the internal defect area with an ultrasonic probe to obtain the ultrasonic defect reflection data of the internal defect area;

[0128] Perform X-ray imaging on the delamination defect area to obtain the three-dimensional X-ray image of the delamination defect area;

[0129] The first generation sub-module is used to spatially align and register the two-dimensional optical image, the ultrasonic defect reflection data, and the three-dimensional X-ray image to generate multi-modal fusion data;

[0130] The second generation sub-module is used to extract the current defect features from the multi-modal fusion data through a feature extraction algorithm, and generate the comprehensive defect information of the integrated circuit package sample based on the current defect features using data fusion technology.

[0131] In one embodiment, the statistics module includes:

[0132] The fourth determination sub-module is used to determine the defect type distribution of the integrated circuit package sample according to the comprehensive defect information, and determine the target defect type corresponding to each defect area according to the defect type distribution;

[0133] The fifth determination sub-module is used to determine the defect depth information and defect size corresponding to each defect area based on the multi-modal fusion data through a defect depth analysis model based on a deep learning algorithm;

[0134] The sixth determination sub-module is used to determine the defect morphology and position parameters of each defect area according to the defect depth information and defect size, obtain the package function information of the integrated circuit package sample, and determine the performance parameters to be analyzed according to the package function information;

[0135] The construction sub-module is used to obtain experimental defect analysis data according to the performance parameters to be analyzed, and construct a defect-performance relationship model according to the experimental defect analysis data;

[0136] A seventh determination sub-module, configured to determine, based on a defect-performance relationship model, a target defect type performance impact correlation parameter and a performance impact descriptive statistical parameter for each defect area according to the defect morphology and position parameters of each defect area.

[0137] In one embodiment, the generation module includes:

[0138] An eighth determination sub-module, configured to determine a quantitative impact evaluation parameter and a qualitative impact evaluation parameter for each type of defect according to the defect type and the performance impact parameters caused by each type of defect;

[0139] A ninth determination sub-module, configured to determine respective screening criteria for defects to be repaired and negligible defects, and determine a reasonable parameter range and a reasonable parameter description for each performance index according to the screening criteria;

[0140] A division sub-module, configured to divide each type of defect into defects to be repaired and negligible defects according to the quantitative impact evaluation parameter and the qualitative impact evaluation parameter of each type of defect, and the reasonable parameter range and the reasonable parameter description of the performance indexes of the defects to be repaired and the negligible defects respectively;

[0141] A third generation sub-module, configured to generate a repair suggestion according to the division result and the target defect morphology of the defect to be repaired, and generate a visualization report according to the repair suggestion.

[0142] Those skilled in the art should understand that the first and second in the present invention refer to different application stages.

[0143] After considering the specification and practicing the disclosure herein, those skilled in the art will readily conceive of other embodiments of the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0144] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A method for defect detection of an integrated circuit package, characterized in that, Including the following steps: Obtain a scanned image of an integrated circuit package sample, and analyze the scanned image to determine the defective area of the integrated circuit package sample; Collect multi-modal data of the defective area and perform fusion to obtain comprehensive defect information of the integrated circuit package sample; Statistically analyze the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information; Screen out defects to be repaired and negligible defects according to the defect types and the performance impact parameters caused by each type of defect, and generate a visualization report; The obtaining of the scanned image of the integrated circuit package sample and the analysis of the scanned image to determine the defective area of the integrated circuit package sample include: Determine the structural parameters of the integrated circuit package sample, determine the layering information of the integrated circuit package sample according to the structural parameters, and determine the scanning method according to the layering information. The scanning methods include: internal scanning, surface scanning, and layer-by-layer scanning; Scan the integrated circuit package sample by the scanning method to obtain a scanned image, and perform image enhancement preprocessing on the scanned image; Obtain the defect characteristics of different types of defects, determine the recognition elements of different types of defects according to the defect characteristics, and determine the recognition and positioning strategies of different types of defects according to the recognition elements; Analyze the scanned image through image processing software based on the recognition and positioning strategies of different types of defects, and identify and locate the defective area of the integrated circuit package sample according to the analysis results; Obtain the defect characteristics of different types of defects, determine the recognition elements of different types of defects according to the defect characteristics, and determine the recognition and positioning strategies of different types of defects according to the recognition elements, including: Obtain the defect image dominant characteristics and defect text description characteristics of different types of defects, and determine the defect characteristic point distribution map according to the defect image dominant characteristics and defect text description characteristics; Determine the affine anomaly frequency component according to the defect characteristic point distribution map through the affine transformation matrix, compare the affine anomaly frequency component with the standard frequency component, and obtain the comparison result; Determine the recognition elements of different types of defects according to the comparison result and the intensification elements of the affine anomaly frequency component, and obtain the shallow detection expression attribute and the depth detection expression attribute of the recognition elements; Respectively obtain the characteristic change arrays of the shallow detection expression attribute and the depth detection expression attribute, and determine the low-frequency scanning parameters and high-frequency pairing parameters for the recognition elements based on the characteristic change arrays; Determine the scanning form according to the low-frequency scanning parameters, and determine the low-frequency parameter recognition and positioning strategy for the recognition elements based on the scanning form; Obtain the parameter change rule of the high-frequency pairing parameters, determine the pairing mechanism according to the parameter change rule, and determine the high-frequency parameter recognition and positioning strategy for the recognition elements based on the pairing mechanism; Integrate the low-frequency parameter recognition and positioning strategy and the high-frequency parameter recognition and positioning strategy for the recognition elements to generate the recognition and positioning strategy of different types of defects.

2. The defect detection method for the integrated circuit package according to claim 1, wherein The collecting of multi-modal data of the defective area and the fusion to obtain comprehensive defect information of the integrated circuit package sample include: Determine the area attribute of the defective area. The area attributes include: internal area, surface area, and layer-by-layer area. Determine the data collection method according to the area attribute, where: Collect the surface topography of the surface defect area through a high-resolution optical microscope, and determine the two-dimensional optical surface image of the defect area according to the surface topography; Scan the internal defect area with an ultrasonic probe to obtain the ultrasonic defect reflection data of the internal defect area; Perform X-ray imaging on the delamination defect area to obtain the three-dimensional X-ray image of the delamination defect area; Perform spatial alignment and registration on the two-dimensional optical surface image, the ultrasonic defect reflection data, and the three-dimensional X-ray image to generate multi-modal fusion data; Extract the current defect features from the multi-modal fusion data through a feature extraction algorithm, and use data fusion technology to generate comprehensive defect information of the integrated circuit package sample based on the current defect features.

3. The method for defect detection of the integrated circuit package according to claim 2, wherein, Statistically analyze the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information, including: Determine the defect type distribution of the integrated circuit package sample according to the comprehensive defect information, and determine the target defect type corresponding to each defect area according to the defect type distribution; Determine the defect depth information and defect size corresponding to each defect area based on the multi-modal fusion data through a defect depth analysis model based on a deep learning algorithm; Determine the defect morphology and position parameters of each defect area according to the defect depth information and defect size, obtain the package function information of the integrated circuit package sample, and determine the performance parameters to be analyzed according to the package function information; Obtain experimental defect analysis data according to the performance parameters to be analyzed, and construct a defect-performance relationship model according to the experimental defect analysis data; Determine the target defect type, its performance impact correlation parameters, and performance impact descriptive statistical parameters of each defect area based on the defect morphology and position parameters of each defect area through the defect-performance relationship model.

4. The defect detection method for the integrated circuit package according to claim 3, characterized in that, Screen out the defects to be repaired and the negligible defects according to the defect type and the performance impact parameters caused by each type of defect, and generate a visualization report, including: Determine the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect according to the defect type and the performance impact parameters caused by each type of defect; Determine the screening criteria for the defects to be repaired and the negligible defects respectively, and determine the reasonable parameter range and reasonable parameter description of each performance index according to the screening criteria; Divide each type of defect into defects to be repaired and negligible defects according to the quantitative impact evaluation parameters and qualitative impact evaluation parameters of each type of defect, and the reasonable parameter range and reasonable parameter description of the performance indexes of the defects to be repaired and the negligible defects respectively; Generate repair suggestions according to the division results and the target defect morphology of the defects to be repaired, and generate a visualization report according to the repair suggestions.

5. A defect detection system for an integrated circuit package, characterized in that, The system includes: An analysis module for obtaining the scanned image of the integrated circuit package sample and analyzing the scanned image to determine the defect area of the integrated circuit package sample; An acquisition module for collecting multi-modal data of the defect area and performing fusion to obtain comprehensive defect information of the integrated circuit package sample; A statistics module for statistically analyzing the defect types of the integrated circuit package sample and the performance impact parameters caused by each type of defect according to the comprehensive defect information; A generation module, configured to screen out defects to be repaired and ignorable defects according to the defect type and the performance impact parameters caused by each type of defect, and generate a visualization report; The analysis module includes: A first determination sub-module, configured to determine the structural parameters of the integrated circuit package sample, determine the delamination information of the integrated circuit package sample according to the structural parameters, and determine the scanning method according to the delamination information. The scanning methods include: internal scanning, surface scanning, and delamination scanning; A scanning sub-module, configured to scan the integrated circuit package sample by the scanning method, obtain a scanned image, and perform image enhancement preprocessing on the scanned image; A second determination sub-module, configured to obtain the defect characteristics of different types of defects, determine the recognition elements of different types of defects according to the defect characteristics, and determine the recognition and localization strategies of different types of defects according to the recognition elements; An analysis sub-module, configured to analyze the scanned image through image processing software based on the recognition and localization strategies of different types of defects, and identify and locate the defect area of the integrated circuit package sample according to the analysis result; Obtaining the defect characteristics of different types of defects, determining the recognition elements of different types of defects according to the defect characteristics, and determining the recognition and localization strategies of different types of defects, including: Obtaining the defect image dominant characteristics and defect text description characteristics of different types of defects, and determining the defect characteristic point distribution map according to the defect image dominant characteristics and defect text description characteristics; Determining the affine abnormal frequency component according to the defect characteristic point distribution map through an affine transformation matrix, comparing the affine abnormal frequency component with the standard frequency component, and obtaining a comparison result; Determining the recognition elements of different types of defects according to the comparison result and the intensification elements of the affine abnormal frequency component, and obtaining the shallow detection expression attribute and the depth detection expression attribute of the recognition elements; Respectively obtaining the characteristic change arrays of the shallow detection expression attribute and the depth detection expression attribute, and determining the low-frequency scanning parameter and the high-frequency pairing parameter for the recognition elements based on the characteristic change arrays; Determining the scanning form according to the low-frequency scanning parameter, and determining the low-frequency parameter recognition and localization strategy for the recognition elements based on the scanning form; Obtaining the parameter change rule of the high-frequency pairing parameter, determining the pairing mechanism according to the parameter change rule, and determining the high-frequency parameter recognition and localization strategy for the recognition elements based on the pairing mechanism; Integrating the low-frequency parameter recognition and localization strategy and the high-frequency parameter recognition and localization strategy for the recognition elements to generate the recognition and localization strategy of different types of defects.

6. The defect detection system for the integrated circuit package according to claim 5, wherein, The acquisition module includes: A third determination sub-module, configured to determine the region attribute of the defect area. The region attributes include: internal region, surface region, and delamination region, and determine the data acquisition method according to the region attribute, where: Collecting the surface topography of the surface defect area through a high-resolution optical microscope, and determining the optical surface two-dimensional image of the defect area according to the surface topography; Scanning the internal defect area through an ultrasonic probe to obtain ultrasonic defect reflection data of the internal defect area; Performing X-ray imaging on the delamination defect area to obtain the X-ray three-dimensional image of the delamination defect area; The first generation sub-module is used to perform spatial alignment and registration on the two-dimensional optical surface image, ultrasonic defect reflection data, and three-dimensional X-ray image to generate multi-modal fusion data; The second generation sub-module is used to extract the current defect features from the multi-modal fusion data through a feature extraction algorithm, and generate comprehensive defect information of the integrated circuit package sample based on the current defect features by using data fusion technology.

7. The defect detection system for the integrated circuit package according to claim 6, wherein The statistics module includes: The fourth determination sub-module is used to determine the defect type distribution of the integrated circuit package sample according to the comprehensive defect information, and determine the target defect type corresponding to each defect area according to the defect type distribution; The fifth determination sub-module is used to determine the defect depth information and defect size corresponding to each defect area based on the multi-modal fusion data through a defect depth analysis model based on a deep learning algorithm; The sixth determination sub-module is used to determine the defect morphology and position parameters of each defect area according to the defect depth information and defect size, obtain the package function information of the integrated circuit package sample, and determine the performance parameters to be analyzed according to the package function information; The construction sub-module is used to obtain experimental defect analysis data according to the performance parameters to be analyzed, and construct a defect-performance relationship model according to the experimental defect analysis data; The seventh determination sub-module is used to determine the target defect type, its performance impact correlation parameter, and performance impact descriptive statistical parameter of each defect area based on the defect morphology and position parameters of each defect area through the defect-performance relationship model.

8. The defect detection system for the integrated circuit package according to claim 7, wherein The generation module includes: The eighth determination sub-module is used to determine the quantitative impact evaluation parameter and qualitative impact evaluation parameter of each type of defect according to the defect type and the performance impact parameter caused by each type of defect; The ninth determination sub-module is used to determine the screening criteria for the defects to be repaired and the defects that can be ignored respectively, and determine the reasonable parameter range and reasonable parameter description of each performance index according to the screening criteria; The division sub-module is used to divide each type of defect into defects to be repaired and defects that can be ignored according to the quantitative impact evaluation parameter and qualitative impact evaluation parameter of each type of defect, and the reasonable parameter range and reasonable parameter description of the performance indexes of the defects to be repaired and the defects that can be ignored respectively; The third generation sub-module is used to generate a repair suggestion according to the division result and the target defect morphology of the defects to be repaired, and generate a visualization report according to the repair suggestion.

Images