An advanced packaging industry bonding process inspection data automatic association method and system

Abstract

This invention relates to the field of semiconductor manufacturing and data fusion technology, specifically to an automatic association method and system for inspection data in the bonding process of advanced packaging industry. The automatic association method for inspection data in the bonding process of advanced packaging industry includes the following steps: Real-time acquisition of multi-source data: Real-time acquisition of inspection data from multiple brands of testing equipment, while simultaneously acquiring production data from the bonding equipment in the manufacturing execution system; Spatiotemporal association algorithm: Using a time dynamic alignment algorithm and a spatial coordinate transformation model, the inspection data and the corresponding production data are matched in both spatiotemporal dimensions to obtain associated matching data; Data quality verification and storage: The associated matching data is verified for quality, and the associated matching data that meets the integrity verification rules is uploaded to the blockchain for storage.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing and data fusion technology, specifically to an automatic association method and system for inspection data of bonding processes in the advanced packaging industry. Background Technology

[0002] In the field of advanced semiconductor packaging, especially for high-end applications such as 5G communication, artificial intelligence, high-performance computing, and automotive electronics, wafer bonding is a core process that determines the final chip's performance, reliability, and yield. This process involves the three-dimensional integration of multilayer wafers or chips through methods such as thermoforming, eutectic bonding, and welding, with requirements for alignment accuracy, interface uniformity, and residual stress control reaching sub-micron to nanometer levels. Against this backdrop, the precise correlation between pre-shipment inspection data (such as post-bonding offset, adhesive layer thickness, and interface voids) and production data (such as bonding pressure, temperature profiles, and equipment parameters) is not only a basic requirement for process monitoring and quality traceability but also a key enabling technology for process optimization, intelligent diagnostics, and yield improvement. However, current industry-wide correlation management methods face the following severe and deep-seated technical bottlenecks, seriously hindering the development of advanced packaging technologies towards higher density and superior performance. Specifically: Manual data entry has poor reliability: It relies on inspectors to manually record microscope measurement data (such as bonding offset and adhesive layer thickness), which is inefficient and prone to errors. The data format is not uniform, and the error rate can reach 0.8%.

[0003] Low data integration: The test data is stored locally and is isolated from the production data in the MES system. Manual batch matching is required, which is time-consuming and has a high matching error rate (about 1.5%), resulting in a break in the quality traceability chain.

[0004] Spatiotemporal information mismatch: Traditional batch number-based association methods cannot track the specific slot changes of the wafer within the FOUP, resulting in the detection position not accurately corresponding to the actual bonding position, which seriously affects the accuracy of root cause analysis. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic association method for inspection data in the bonding process of advanced packaging industry, so as to solve the problems of poor reliability, low data integration and spatiotemporal information mismatch in the association management of pre-shipment inspection data and production data in the wafer bonding process of advanced semiconductor packaging.

[0006] To address the aforementioned problems, this invention proposes an automatic correlation method for inspection data in the bonding process of the advanced packaging industry. The technical solution adopted is as follows: An automatic correlation method for inspection data of bonding processes in the advanced packaging industry includes the following steps: Step S1, Real-time acquisition of multi-source data: Real-time acquisition of test data from multiple brands of testing equipment, and simultaneous acquisition of production data from the bonding equipment in the manufacturing execution system; Step S2, Spatiotemporal Correlation Algorithm: Using a time dynamic alignment algorithm and a spatial coordinate transformation model, the detection data and the production data of the bonding equipment are subjected to spatiotemporal dual correlation matching to obtain the correlated matching data: Step S3, Data Quality Verification and Storage: Perform quality verification on the associated matching data, and upload the associated matching data that meets the integrity verification rules to the blockchain for permanent storage.

[0007] Furthermore, the real-time acquisition of testing data from multiple brands of testing equipment includes: For different brands of testing equipment, develop dedicated data interfaces or middleware, and connect them with the testing equipment software or general industrial protocols to achieve real-time acquisition of test images and data in a unified format.

[0008] Furthermore, acquiring production data from the bonding equipment in the manufacturing execution system includes: The bonding time, bonding equipment number, and operator information are obtained in real time from the manufacturing execution system. At the same time, in conjunction with the UWB positioning system, a high-precision industrial camera deployed above the exit is used to obtain real-time physical slot information of the wafer in the transfer box.

[0009] Furthermore, the time dynamic alignment algorithm includes: Establish a time matching window to cover a reasonable time interval between bonding time and detection time; A dynamic time warping algorithm is used to calibrate the clock deviation between the bonding device and the detection device so that their time series data can be accurately matched within the time matching window.

[0010] Furthermore, the establishment of the spatial coordinate transformation model includes: Calibration is performed using a high-precision calibration plate; Establish a mapping relationship between the coordinate system of the bonding equipment and the coordinate system of the inspection equipment. The coordinate system of the bonding equipment takes the center of the worktable of the bonding equipment as its origin, and the coordinate system of the inspection equipment takes the center of the field of view of the inspection equipment as its origin. Based on the aforementioned mapping relationship, the coordinate transformation formula implemented through linear mapping is applied to the coordinate transformation between the detection point and the bonding point.

[0011] Furthermore, the coordinate transformation formula is defined as follows:

[0012] in,( ( ) represents the coordinates of the bonding point. Here are the coordinates of the detection point. The conversion factor from pixel to μm. and This is the fixed offset between the bonding equipment and the testing equipment.

[0013] Furthermore, the integrity verification rule includes at least one of the following: Verify the number of missing key parameters. If there is ≤1 missing key parameter, the key process parameter shall include at least one of the following: offset X between the bonding equipment and the testing equipment, offset Y, and adhesive layer thickness. The sharpness of the detected image is verified, and the sharpness is greater than or equal to a preset threshold. The sharpness is obtained by calculating the Laplacian variance of the detected image. Verify whether there is a corresponding keyed record within the time matching window determined by the time dynamic alignment algorithm, and whether the number of successful matches is ≥1.

[0014] Furthermore, the uploading to the blockchain for permanent storage includes: The associated matching data that meets the integrity verification rules is encapsulated according to a preset structured data format; The packaged data is uploaded to a blockchain network for permanent storage. The blockchain network ensures the immutability of the packaged data and that the retention period of the packaged data meets the preset compliance requirements. The evidence must include at least one of the following: wafer ID, bonding parameters, test results, and equipment information.

[0015] Furthermore, the method for automatically associating inspection data of bonding processes in the advanced packaging industry also includes: In step S3, the quality of the associated matching data is checked. If the associated matching data does not meet the integrity check rules, a retest is triggered, and steps S1 to S3 are continued.

[0016] This invention also provides an automatic correlation system for bonding process inspection data in the advanced packaging industry, comprising: The multi-source data real-time acquisition module is used to collect test data from multiple brands of testing equipment in real time, and simultaneously acquire production data from the bonding equipment in the manufacturing execution system. The spatiotemporal correlation algorithm module is used to perform spatiotemporal dual correlation matching between the detection data and the production data of the bonding equipment through a time dynamic alignment algorithm and a spatial coordinate transformation model to obtain correlation matching data; The data quality verification and evidence storage module is used to verify the quality of associated matching data and upload the associated matching data that meets the integrity verification rules to the blockchain for permanent evidence storage.

[0017] Compared with the prior art, this application has the following beneficial effects: This invention completely replaces the traditional, inefficient, and error-prone manual data entry and matching with a closed-loop method of "real-time acquisition of multi-source data → spatiotemporal dual correlation matching → data quality verification and reliable evidence storage." It reduces the data correlation time for single-wafer bonding from minutes to seconds, improves quality traceability efficiency by 87.5%, reduces the data correlation error rate from 0.8% to below 0.01%, and achieves 100% integrity of key parameters, realizing high-precision and high-efficiency automatic data correlation. Specifically, this invention combines dynamic temporal alignment with spatial coordinate mapping, fundamentally solving the problems of temporal deviation and positional mismatch in data matching, and improving the correlation accuracy to over 99.9%.

[0018] This invention presents an automatic data association method for bonding process inspection in the advanced packaging industry. In the semiconductor advanced packaging bonding process, through spatiotemporal dual association matching, it achieves automatic and accurate association between inspection data and production data. The association management boasts high reliability and data integration, making it particularly suitable for resolving data silos and spatiotemporal matching discrepancies between inspection equipment such as optical microscopes and electron microscopes and the Manufacturing Execution System (MES). This provides a reliable data foundation for precise single-wafer-level quality traceability and process root cause analysis. Simultaneously, it creatively incorporates quality verification and blockchain notarization as standard outputs of the method. Quality verification supports precise reverse tracing from inspection anomalies to specific bonding process parameters, achieving single-wafer-level positioning accuracy. Blockchain notarization ensures the authenticity, integrity, immutability, and long-term traceability of every piece of critical process data (bonding parameters, inspection results), directly meeting the stringent requirements of standards such as IATF 16949 for data notarization. This provides digital assurance for product reliability and promotes the automation and standardization of inspection data management.

[0019] The real-time acquisition of test data from multiple brands of testing equipment includes: For different brands of testing equipment, dedicated data interfaces or middleware are developed. These interfaces or middleware connect to the testing equipment software or general industrial protocols, enabling real-time acquisition of test images and data in a unified format. Through adaptation of general industrial protocols to dedicated data interfaces or middleware, unified data access for over ten brands of testing equipment is achieved, breaking down data silos and realizing seamless integration of multi-source equipment. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the process for automatically associating inspection data of the bonding process in the advanced packaging industry according to the present invention; Figure 2 This is a comparison chart showing the accuracy of the advanced packaging industry bonding process inspection data automatic association method of the present invention with traditional manual matching and batch number association. Figure 3 This is a block diagram of the advanced packaging industry bonding process inspection data automatic association system of the present invention. Detailed Implementation

[0021] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] The following describes, with reference to the accompanying drawings, an advanced packaging industry bonding process inspection data automatic association method and control method according to embodiments of this application.

[0024] The following is combined Figure 1 This application provides a detailed description of the method for automatically associating inspection data of bonding processes in the advanced packaging industry.

[0025] The method for automatically associating inspection data of bonding processes in the advanced packaging industry as described in this application includes the following steps: Step S1: Real-time acquisition of multi-source data: Real-time acquisition of test data from multiple brands of testing equipment, and simultaneous acquisition of production data from the bonding equipment in the manufacturing execution system.

[0026] Specifically, it involves real-time acquisition of testing data from various brands of testing equipment. This includes developing dedicated data interfaces or middleware for different brands of testing equipment, and connecting these interfaces or middleware with the testing equipment software or with common industrial protocols to achieve real-time acquisition of testing images and data in a unified format. Here, the different brands of testing equipment could be Zeiss, Olympus, etc.

[0027] In one specific embodiment, dedicated data interfaces or middleware are developed for different brands of testing equipment. The testing data is parsed using the software development kit (SDK) provided by the testing equipment, or the testing images and data are captured based on common industrial protocols (such as GigE Vision), ensuring uniform data format and real-time transmission (transmission latency ≤100ms).

[0028] Production data from the bonding equipment in the Manufacturing Execution System (MES) is acquired, including real-time acquisition of bonding time, bonding equipment, and operator information from the MES. Simultaneously, in conjunction with a UWB positioning system, a high-precision industrial camera deployed above the exit gate acquires real-time physical slot information of the wafers in the transfer box. Here, the bonding time is accurate to milliseconds. The inspection equipment is a microscope.

[0029] Step S2, Spatiotemporal correlation algorithm: By using a time dynamic alignment algorithm and a spatial coordinate transformation model, the detection data and the production data of the bonding equipment are spatiotemporally correlated and matched to obtain correlated matching data.

[0030] Specifically, the dynamic time alignment algorithm includes: establishing a time matching window that covers a reasonable time interval between bonding time and detection time; and using a dynamic time warping algorithm to calibrate the clock deviation between the bonding device and the detection device, so that their time series data can be accurately matched within the time matching window, ensuring consistency between the bonding time and the detection timestamp. Here, the time matching window can be ±180 seconds.

[0031] The establishment of the spatial coordinate transformation model includes: calibration using a high-precision calibration plate; establishing a mapping relationship between the bonding equipment coordinate system and the inspection equipment coordinate system, with the bonding equipment coordinate system taking the center of the bonding equipment worktable as the origin and the inspection equipment coordinate system taking the center of the inspection equipment field of view as the origin; and based on the mapping relationship, applying the coordinate transformation formula achieved through linear mapping to the coordinate transformation between the inspection point and the bonding point.

[0032] In one specific implementation, the spatial coordinate transformation model is established using a coordinate transformation algorithm, specifically: the bonding equipment coordinate system takes the center of the bonding equipment worktable as its origin, and obtains the coordinates of the bonding point (…). The unit is μm; the coordinate system of the detection equipment is based on the center of the field of view of the detection equipment to obtain the coordinates of the detection point. The unit is pixels. The high-precision calibration plate measures 20mm × 20mm and has an accuracy of ±0.1μm. A transformation model between the bonding machine coordinate system (μm) and the microscope coordinate system (pixels) is established using the high-precision calibration plate, achieving sub-pixel-level matching between the detection point and the bonding point.

[0033] In one specific embodiment, the coordinate transformation formula is defined as:

[0034] in,( ( ) represents the coordinates of the bonding point. Here are the coordinates of the detection point. The conversion factor from pixel to μm. and This is a fixed offset between the bonding equipment and the testing equipment, obtained through calibration using a high-precision calibration board. .

[0035] Step S3, Data Quality Verification and Storage: Perform quality verification on the associated matching data, and upload the associated matching data that meets the integrity verification rules to the blockchain for permanent storage.

[0036] Specifically, the integrity verification rules must include at least one of the following: Verify the number of missing key parameters. There should be ≤1 missing key parameter. The key parameters must include at least one of the following: the offset X between the bonding device and the testing device, the offset Y between the bonding device and the testing device, and the adhesive layer thickness. Verify the sharpness of the detected image. The sharpness must be greater than or equal to a preset threshold. The sharpness is obtained by calculating the Laplacian variance of the detected image. The preset threshold can be 1500. Verify whether there is a corresponding keyed record within the time matching window determined by the time dynamic alignment algorithm, and whether the number of successful matches is ≥1.

[0037] Uploading to the blockchain for permanent storage includes: encapsulating the associated matching data that meets the integrity verification rules according to a preset structured data format; uploading the encapsulated data to the blockchain network for permanent storage, wherein the blockchain network ensures the immutability of the encapsulated data and that the retention period of the encapsulated data meets preset compliance requirements; the stored content includes at least one of the following: wafer ID, bonding parameters, test results, and device information.

[0038] In one specific embodiment, the associated matching data that meets the integrity verification rules is uploaded to the blockchain in JSON-LD format to ensure that the data is tamper-proof. This ensures the authenticity and auditability of the data throughout its lifecycle (≥15 years), meeting the requirements of high-standard quality systems such as automotive-grade standards. For example, the evidence includes at least: Wafer ID: WAFER-20250521-001", bonding pressure: 8.5 N, detection offset X: 0.9 μm, association time: 2025-05-21T09:17:45Z, and the device mapping is: bonding machine: Bonding-007, microscope: Zeiss Axio Imager Z2.

[0039] In another specific embodiment, the method for automatically associating inspection data of bonding processes in the advanced packaging industry also includes: In step S3, the quality of the associated matching data is checked. If the associated matching data does not meet the integrity check rules, a retest is triggered, and steps S1 to S3 are continued.

[0040] The accuracy of the above-mentioned automatic correlation method for bonding process inspection data in the advanced packaging industry is compared with that of traditional manual matching and batch number correlation methods. Figure 2 As shown, the accuracy rate of traditional manual matching is 92%, the accuracy rate of batch number association method is 95.5%, and the accuracy rate of this invention is improved by 4.4 percentage points to 99.9% through spatiotemporal dual association; at the same time, this data is based on the verification of 100,000 wafers, covering 8 bonding processes.

[0041] This application also provides an automatic correlation system for inspection data of bonding processes in the advanced packaging industry, including: The multi-source data real-time acquisition module is used to collect test data from multiple brands of testing equipment in real time, and simultaneously acquire production data from the bonding equipment in the manufacturing execution system. The spatiotemporal correlation algorithm module is used to perform spatiotemporal dual correlation matching between the detection data and the production data of the bonding equipment through a time dynamic alignment algorithm and a spatial coordinate transformation model to obtain correlation matching data; The data quality verification and evidence storage module is used to verify the quality of associated matching data and upload the associated matching data that meets the integrity verification rules to the blockchain for permanent evidence storage.

[0042] At the hardware level, such as Figure 3 As shown, the advanced packaging industry's automatic data association system for bonding process inspection includes an image acquisition unit for collecting inspection data, a calibration unit for coordinate system calibration, and a position sensing unit for acquiring real-time wafer position information. Specifically, the hardware layer, as the data acquisition layer, enables real-time and synchronous acquisition of data from the inspection equipment and production system. It includes a microscope and a bonding machine. The microscope, acting as the image acquisition unit for collecting inspection data, acquires the data via a data acquisition card. A PLC controller is used to acquire production data from the bonding machine, including real-time wafer position information.

[0043] At the software level, the advanced packaging industry's automatic correlation system for bonding process inspection data includes: The interface adapter module is used to connect to and adapt to testing equipment from various brands to collect the testing data in a unified format; it is compatible with multiple brands of equipment and supports protocol conversion (such as TCP / IP, USB3Vision) for more than 10 brands of testing equipment. The spatiotemporal correlation engine is used to execute the aforementioned dynamic time alignment algorithm and spatial coordinate transformation model to achieve spatiotemporal correlation matching of data; at the same time, it is used to execute the aforementioned integrity verification rules. The data storage module is used to upload verified data to the blockchain.

[0044] In one specific embodiment, the spatiotemporal correlation engine includes three sub-modules: time alignment, coordinate transformation, and quality verification, with a processing capacity of 500 pieces / minute.

[0045] The advanced packaging industry bonding process inspection data automatic association system also includes: a data platform, which provides multi-dimensional queries (such as filtering wafers with abnormal bonding pressure by offset > 1μm), with a query response time ≤ 2 seconds.

[0046] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. An advanced packaging industry bonding process inspection data automatic correlation method, characterized in that, Includes the following steps: Step S1, Real-time acquisition of multi-source data: Real-time acquisition of test data from multiple brands of testing equipment, and simultaneous acquisition of production data from the bonding equipment in the manufacturing execution system; Step S2, Spatiotemporal correlation algorithm: By using a time dynamic alignment algorithm and a spatial coordinate transformation model, the detection data and the production data of the bonding equipment are spatiotemporally correlated and matched to obtain the correlation matching data; Step S3, Data Quality Verification and Storage: Perform quality verification on the associated matching data, and upload the associated matching data that meets the integrity verification rules to the blockchain for permanent storage.

2. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 1, characterized in that, The real-time acquisition of test data from multiple brands of testing equipment includes: For different brands of testing equipment, develop dedicated data interfaces or middleware, and connect them with the testing equipment software or general industrial protocols to achieve real-time acquisition of test images and data in a unified format.

3. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 2, characterized in that, The acquisition of production data from the bonding equipment in the manufacturing execution system includes: The bonding time, bonding equipment number, and operator information are obtained in real time from the manufacturing execution system. At the same time, in conjunction with the UWB positioning system, a high-precision industrial camera deployed above the exit is used to obtain real-time physical slot information of the wafer in the transfer box.

4. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 3, characterized in that, The time dynamic alignment algorithm includes: Establish a time matching window to cover a reasonable time interval between bonding time and detection time; A dynamic time warping algorithm is used to calibrate the clock deviation between the bonding device and the detection device so that their time series data can be accurately matched within the time matching window.

5. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 4, characterized in that, The establishment of the spatial coordinate transformation model includes: Calibration is performed using a high-precision calibration plate; Establish a mapping relationship between the coordinate system of the bonding equipment and the coordinate system of the inspection equipment. The coordinate system of the bonding equipment takes the center of the worktable of the bonding equipment as its origin, and the coordinate system of the inspection equipment takes the center of the field of view of the inspection equipment as its origin. Based on the aforementioned mapping relationship, the coordinate transformation formula implemented through linear mapping is applied to the coordinate transformation between the detection point and the bonding point.

6. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 5, characterized in that, The coordinate transformation formula is defined as follows: in,( ( ) represents the coordinates of the bonding point. Here are the coordinates of the detection point. The conversion factor from pixel to μm. and This is the fixed offset between the bonding equipment and the testing equipment.

7. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 4, characterized in that, The integrity verification rule shall include at least one of the following: Verify the number of missing key parameters. There must be ≤1 missing key parameter. The key parameter must include at least one of the following: the offset X between the bonding device and the detection device, the offset Y between the bonding device and the detection device, and the adhesive layer thickness. The sharpness of the detected image is verified, and the sharpness is greater than or equal to a preset threshold. The sharpness is obtained by calculating the Laplacian variance of the detected image. Verify whether there is a corresponding keyed record within the time matching window determined by the time dynamic alignment algorithm, and whether the number of successful matches is ≥1.

8. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 7, characterized in that, The process of uploading to the blockchain for permanent storage includes: The associated matching data that meets the integrity verification rules is encapsulated according to a preset structured data format; The packaged data is uploaded to a blockchain network for permanent storage. The blockchain network ensures the immutability of the packaged data and that the retention period of the packaged data meets the preset compliance requirements. The evidence must include at least one of the following: wafer ID, bonding parameters, test results, and equipment information.

9. The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to claim 7, characterized in that, The method for automatically associating inspection data of bonding processes in the advanced packaging industry also includes: In step S3, the quality of the associated matching data is checked. If the associated matching data does not meet the integrity check rules, a retest is triggered, and steps S1 to S3 are continued.

10. An automatic correlation system for inspection data of bonding processes in the advanced packaging industry, characterized in that, The method for automatically associating inspection data of bonding processes in the advanced packaging industry according to any one of claims 1-9 includes: The multi-source data real-time acquisition module is used to collect test data from multiple brands of testing equipment in real time, and simultaneously acquire production data from the bonding equipment in the manufacturing execution system. The spatiotemporal correlation algorithm module is used to perform spatiotemporal dual correlation matching between the detection data and the production data of the bonding equipment through a time dynamic alignment algorithm and a spatial coordinate transformation model to obtain correlation matching data; The data quality verification and evidence storage module is used to verify the quality of associated matching data and upload the associated matching data that meets the integrity verification rules to the blockchain for permanent evidence storage.

Images