CN-122028784-A - Integrated packaging system and method based on glass substrate

CN122028784ACN 122028784 ACN122028784 ACN 122028784ACN-122028784-A

Abstract

The invention relates to the technical field of semiconductor packaging, in particular to an integrated packaging system and method based on a glass substrate, comprising the following steps: the method comprises the steps of acquiring a time sequence process parameter set through a sensor array on a packaging production line, generating a process stability index set through feature extraction and abnormal point screening, inputting a pre-built packaging quality association model to obtain a multi-dimensional reliability prediction score, matching a process parameter optimization scheme, dynamically adjusting key parameters of production equipment, and starting a verification packaging flow. And (3) carrying out nondestructive testing on the sample to obtain an internal defect distribution map and interface bonding strength data, combining a process stability index set, sending the obtained product into a packaging process traceable analysis engine, reversely identifying key process links and parameter deviations corresponding to defects, generating a closed-loop process correction instruction and updating a knowledge base. The method realizes dynamic regulation and control of the packaging process and closed-loop tracing of defects, improves the stability and structural reliability of the packaging process, and optimizes the production efficiency.

Inventors

Xiao Suqun
ZENG HAI
CHEN QU
TANG YUEYUN
Bi Pengjun
FANG FULIN

Assignees

深圳市比洋光通信科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260320

Claims (10)

1. An integrated packaging method based on a glass substrate, the method comprising: collecting a time sequence process parameter set in a production flow through a sensor array arranged on a packaging production line, and carrying out feature extraction and abnormal point screening on the time sequence process parameter set to generate a process stability index set; Inputting the process stability index set into a pre-constructed packaging quality association model, and calculating to obtain a multidimensional reliability prediction score of the packaging structure; based on the multidimensional reliability prediction score, matching a corresponding process parameter optimization scheme in a packaging process knowledge base; executing the process parameter optimization scheme, dynamically adjusting key control parameters of production equipment, and starting a round of verification packaging flow; After the verification packaging process is finished, carrying out nondestructive testing on the produced sample to obtain the internal defect distribution diagram and interface bonding strength data of the sample; Sending the internal defect distribution map and interface bonding strength data together with the process stability index set to a packaging process traceability analysis engine; Utilizing the package process retrospective analysis engine to reversely deduce and identify key process links and parameter deviations thereof which cause specific defects; And generating a targeted closed-loop process correction instruction according to the identification result, and updating a corresponding entry in the packaging process knowledge base.
2. The integrated packaging method based on a glass substrate according to claim 1, wherein the feature extraction and the outlier screening are performed on the time series process parameter set to generate a process stability index set, comprising: The time sequence process parameter set comprises substrate positioning data, colloid coating form data, bonding pressure waveform data and curing temperature field data; The process stability index set comprises a positioning offset, glue line uniformity, a pressure fluctuation coefficient and a temperature gradient standard deviation; Sliding window processing is carried out on the substrate positioning data sequence, mass center drift amount and standard deviation of positioning coordinates in each window are calculated, and the positioning offset is obtained through synthesis; Performing image processing on the colloid coating form data, extracting a colloid line width sequence and a colloid line height sequence, and calculating the variation coefficient ratio of the width to the height to be used as the colloid line uniformity; Performing frequency spectrum analysis on the bonding pressure waveform data, extracting the energy duty ratio outside the main frequency band, counting the discrete degree of the pressure peak value, and fusing the energy duty ratio outside the main frequency band and the discrete degree of the pressure peak value to obtain the pressure fluctuation coefficient; And carrying out space gridding analysis on the curing temperature field data, calculating the temperature change rate difference of each gridding point in the curing process, and solving the standard deviation of each gridding point to be used as the standard deviation of the temperature gradient.
3. The integrated packaging method based on a glass substrate according to claim 2, wherein inputting the process stability index set into a pre-built packaging quality association model, calculating a multi-dimensional reliability prediction score of a packaging structure, comprises: performing standardized scaling on the positioning offset, the glue line uniformity, the pressure fluctuation coefficient and the temperature gradient standard deviation to form a feature vector; Inputting the feature vector into the packaging quality association model, wherein the packaging quality association model is a trained deep learning network and comprises a plurality of full-connection layers; The packaging quality association model outputs a multidimensional vector, and each dimension of the multidimensional vector corresponds to a mechanical strength prediction score, an air tightness prediction score, a thermal cycle life prediction score and an electrical connection reliability prediction score respectively to jointly form the multidimensional reliability prediction score.
4. The integrated packaging method based on the glass substrate according to claim 3, wherein the training process of the packaging quality association model comprises: collecting a complete process stability index set in a historical packaging batch and a corresponding final product reliability actual measurement data set to form a training sample pair; Taking the process stability index set as an input characteristic, taking the final product reliability actual measurement data set as a supervision tag, and training the deep learning network; And optimizing the network weight by adopting a back propagation algorithm, and minimizing the mean square error between the multidimensional reliability prediction score output by the model and the supervision label until the model converges.
5. The integrated packaging method based on glass substrates as claimed in claim 4, wherein matching the corresponding process parameter optimization scheme in the packaging process knowledge base based on the multidimensional reliability prediction score comprises: Searching historical case records with similar prediction score modes in the packaging process knowledge base by taking the multidimensional reliability prediction score as a query condition; Extracting a process parameter adjustment strategy which successfully improves each reliability score from the matched historical case records; and carrying out adaptive conversion on the extracted process parameter adjustment strategy according to the specific model and parameter range of the current production equipment, and generating the process parameter optimization scheme which can be immediately executed.
6. The integrated packaging method based on a glass substrate according to claim 5, wherein the process of constructing and updating the packaging technology knowledge base comprises: Recording each complete packaging flow, including an original time sequence process parameter set, a calculated process stability index set, an applied process parameter optimization scheme and internal defect distribution diagram and interface bonding strength data of a final sample; Taking the process stability index set and the multidimensional reliability prediction score as indexes, and storing records of each process as a knowledge item; And after a new closed-loop process correction instruction is generated, correcting the process parameter optimization scheme part in the corresponding knowledge item according to the closed-loop process instruction, and marking the version number.
7. The integrated packaging method based on glass substrates as claimed in claim 6, wherein after the verification packaging process is completed, performing nondestructive testing on the output sample to obtain internal defect distribution map and interface bonding strength data of the sample, comprising: scanning the sample by using X-ray tomography equipment, and reconstructing a three-dimensional structure image of the interior of the sample; identifying and dividing holes, cracks, layering and foreign body defects from the three-dimensional structure image, marking the positions, the sizes and the types of the defects, and generating the internal defect distribution map; and exciting and receiving ultrasonic signals at a sample interface by adopting laser ultrasonic detection equipment, and inversely calculating the interface bonding strength data by analyzing the reflection and transmission characteristics of the ultrasonic signals.
8. The integrated encapsulation method of claim 7, wherein sending the internal defect map and interfacial bond strength data, along with the process stability index set, to an encapsulation process retrospective analysis engine comprises: Quantifying the internal defect distribution map and interface bonding strength data into defect density, defect position clustering characteristics and average bonding strength values; Aligning the quantized defect characteristics with the process stability index set according to time stamps to construct a process-defect associated data table; The process-defect associated data table is input into the packaging process retrospective analysis engine.
9. The integrated packaging method based on glass substrates according to claim 8, wherein the identifying key process links and parameter deviations thereof that lead to specific defects by reverse deduction using the packaging process retrospective analysis engine comprises: The packaging process traceability analysis engine adopts a causal inference method to analyze the statistical correlation between each process stability index and each defect characteristic in the process-defect association data table and the Grandide causal relationship; Screening out process stability indexes with strong causal relation with specific defect characteristics, and positioning process links corresponding to the process stability indexes; And further analyzing the original time sequence process parameters of the process links in the verification process, and identifying parameters deviating from the standard process window and deviation amounts thereof.
10. The integrated packaging method based on the glass substrate according to claim 9, wherein generating a targeted closed-loop process correction instruction according to the recognition result, updating the corresponding entry in the packaging process knowledge base comprises: for each identified parameter deviation, calculating a corresponding parameter compensation value or control curve correction amount based on the deviation amount and the direction of the parameter deviation; Integrating all correction amounts to form a structured closed-loop process correction instruction, wherein the closed-loop process correction instruction explicitly designates target equipment, target parameters and corrected set values; and searching and updating the corresponding part of the related process parameter optimization scheme in the packaging process knowledge base according to the content of the closed-loop process correction instruction, and adding the whole process data of the current retrospective analysis into a knowledge item as a new case.

Description

Integrated packaging system and method based on glass substrate Technical Field The invention relates to the technical field of semiconductor packaging, in particular to an integrated packaging system and method based on a glass substrate. Background In the existing integrated packaging production links of glass substrates, discrete technological parameters are collected by adopting single-point sensors, parameter verification is completed through manual verification or conventional statistics, packaging production operation is carried out according to fixed preset parameters, packaging quality is judged in a finished product post-detection mode, offline analysis is completed by adjusting the technological parameters according to manual experience, the prediction of the packaging quality is realized only by means of single-dimension indexes, and the investigation of packaging defects relies on manual post-verification of the technological links one by one. The traditional packaging regulation and control mode does not form a systematic parameter acquisition and analysis system, the process optimization execution lacks standardized matching logic, and the adjustment of production parameters cannot form real-time linkage with the production flow. The traditional packaging management control mode cannot complete continuous collection and refined characteristic processing of time sequence process parameters, can not convert a process running state into a quantized process stability index set, lacks a packaging quality association model to realize prepositive calculation of multi-dimensional reliability, cannot rely on a packaging process knowledge base to complete automatic matching of a process parameter optimization scheme, and cannot realize dynamic adjustment of production equipment parameters and synchronously start a verification packaging process. The invention aims to solve the technical problems of process regulation and defect tracing in the aspect of the process, and aims to solve the problems that the internal defect distribution and interface bonding strength data obtained by nondestructive testing are only used for defect judgment, linkage analysis cannot be formed with a process stability index set, reverse deduction cannot be realized by a packaging process trace analysis engine, key process links and parameter deviations corresponding to specific defects cannot be positioned, closed-loop process correction instructions cannot be generated, and a packaging process knowledge base cannot be updated in real time. Disclosure of Invention The invention aims to solve the defects in the prior art, and provides an integrated packaging system and method based on a glass substrate. In order to achieve the above purpose, the invention adopts the following technical scheme that the integrated packaging method based on the glass substrate comprises the following steps: collecting a time sequence process parameter set in a production flow through a sensor array arranged on a packaging production line, and carrying out feature extraction and abnormal point screening on the time sequence process parameter set to generate a process stability index set; Inputting the process stability index set into a pre-constructed packaging quality association model, and calculating to obtain a multidimensional reliability prediction score of the packaging structure; based on the multidimensional reliability prediction score, matching a corresponding process parameter optimization scheme in a packaging process knowledge base; executing the process parameter optimization scheme, dynamically adjusting key control parameters of production equipment, and starting a round of verification packaging flow; After the verification packaging process is finished, carrying out nondestructive testing on the produced sample to obtain the internal defect distribution diagram and interface bonding strength data of the sample; Sending the internal defect distribution map and interface bonding strength data together with the process stability index set to a packaging process traceability analysis engine; Utilizing the package process retrospective analysis engine to reversely deduce and identify key process links and parameter deviations thereof which cause specific defects; And generating a targeted closed-loop process correction instruction according to the identification result, and updating a corresponding entry in the packaging process knowledge base. As a further aspect of the present invention, performing feature extraction and outlier screening on the time sequence process parameter set to generate a process stability index set, including: The time sequence process parameter set comprises substrate positioning data, colloid coating form data, bonding pressure waveform data and curing temperature field data; The process stability index set comprises a positioning offset, glue line uniformity, a pressure fluctuation coefficient and a tempe